Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-06-24

| Channels | #salt index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:04 oz_akan_ joined #salt
00:05 possibilities joined #salt
00:07 xcbt joined #salt
00:08 joehoyle joined #salt
00:09 Guest84025 joined #salt
00:10 arnoldB argh. this will be a long night :D
00:10 arnoldB 2am here..
00:11 joehoyle joined #salt
00:12 ecdhe does anyone have a syntactic example of a map.jinja that sets different values based on oscodename?
00:12 agliodbs is there any way to use a match: grains to require *multiple* grains?
00:13 arnoldB ecdhe: https://github.com/bechtoldt?tab=repositories there are more than 50 examples for this
00:13 manfred agliodbs: match: 'G@os_family:Debian and G@osarch:x86_64' ?
00:13 arnoldB ecdhe: you can also check https://github.com/saltstack-formulas even though I don't like these very much
00:13 agliodbs manfred: ah, I can use "and"?  keen
00:14 manfred agliodbs: yes
00:14 manfred agliodbs: you checked and I was right that it was doing compound?
00:14 manfred before i push this PR fixing the documentation
00:14 arnoldB agliodbs: http://docs.saltstack.com/en/latest/topics/targeting/index.html
00:14 Guest6143 joined #salt
00:14 agliodbs manfred: the docs imply that if I try to match a grain which doesn't exist, I'll get an error.  Is that correct?
00:15 manfred if the grain doesn't exist, i don't believe so
00:15 manfred it should always check with a default of None
00:15 agliodbs manfred: unfortunately, we found something else wrong with their FQDNs.  Which is why I'm thinking of giving up and moving to custom grains in the minion config
00:16 ecdhe arnoldB, bechtoldt's formulae seem to favor defaults.yaml over map.jinja
00:16 ecdhe Doesn't saltstack recommend map.jinja instead?
00:16 agliodbs manfred: so I could do a match against "postgrescluster:role:master" and it won't error if "postgrescluster" doesn't exist?
00:16 manfred agliodbs: it should default to None if it doesn't exist iirc
00:17 manfred double check it, but i am 90% sure
00:17 agliodbs ok
00:18 ecdhe arnoldB, do you have a formula that specifically discriminates between 'trusty' and 'precise' or something similar?
00:18 ecdhe I already have examples for Debian vs Redhat.
00:18 arnoldB ecdhe: if you need map.jinja examples, check https://github.com/saltstack-formulas
00:19 arnoldB ecdhe: as soon as you write lots of code, you will recognize a huge difference between map.jinja and defaults.yaml: the syntax
00:19 ecdhe is jinja too verbose?
00:19 Guest6143 joined #salt
00:19 agliodbs manfred: just to confirm, we were not able to test the compound matching in the overstate because our salt config blew up for unrelated reasons
00:20 bhosmer joined #salt
00:20 arnoldB ecdhe: both are using jinja. the latter is yaml. map.jinja is json
00:20 agliodbs manfred: but I should be able to test it soon
00:20 manfred agliodbs: i just double checked it, it works
00:20 agliodbs manfred: thanks!
00:20 arnoldB ecdhe: https://github.com/bechtoldt/httpd-formula/blob/master/httpd/defaults.yaml you can use arbitrary salt commands here (like accessing any grains)
00:21 manfred well, it at least accepts it, so i assume it works, cause I am using G@os_family:Debian and something else
00:23 arnoldB ecdhe: simply add an if: if codename == trusty than output this YAML code, if codename == precise than another. it's a arbitrary jinja template
00:23 garthk joined #salt
00:23 Tekni joined #salt
00:23 nineteen1ightd joined #salt
00:23 ecdhe thanks arnoldB
00:23 brain5ide_ joined #salt
00:24 bhosmer_ joined #salt
00:24 arnoldB ecdhe: https://github.com/bechtoldt/opennebula-formula/blob/master/opennebula/defaults.yaml#L47
00:24 Guest6143 joined #salt
00:24 Ryan_Lane hm. is there any way to make vagrant display the log output of the state.highstate during the run?
00:24 Ryan_Lane I set log_level=info, but it doesn't seem to be showing me anything
00:24 Ryan_Lane 20 minutes with no output isn't wonderful :(
00:25 Ryan_Lane ah. I bet it's because the file has the log_level set
00:26 ajprog_laptop joined #salt
00:27 manfred agliodbs: yup, it works
00:27 manfred test:
00:27 manfred match: G@os:Fedora and G@test:something and salt*
00:27 agliodbs manfred: awesome
00:28 manfred agliodbs: https://github.com/saltstack/salt/pull/13661
00:28 Ryan_Lane when the release candidate is released, will there be a PPA for it? or maybe a pip target?
00:28 Ryan_Lane anything other than source?
00:32 agliodbs manfred: looks good
00:33 agliodbs manfred: now, to get complicated: I have a custom pillar written in python.  in an ideal world, I'd like to use the pillar to pull some information from the grains of a different minion than the minion I'm currently provisioning.  is that even possible?
00:33 Guest6143 joined #salt
00:33 Ryan_Lane hm. no, even with the config file having a log_level of info, I don't see anything
00:34 manfred agliodbs: i have no idea for that.
00:34 agliodbs as an example, I'd like to pull the fqdn of a minion who is a database server while provisioning the proxy, since the proxy needs to know where the database server is
00:34 mateoconfeugo joined #salt
00:35 agliodbs right now we're doing that via a careful URL system, but that's already broken a couple of times
00:36 agliodbs and I can actually find the right database server by looking through the grains of all minions
00:38 manfred agliodbs: why not use the salt mine?
00:38 manfred http://docs.saltstack.com/en/latest/topics/mine/#mine-functions
00:40 gamedna joined #salt
00:40 nmistry joined #salt
00:41 Hell_Fire joined #salt
00:41 UtahDave Ryan_Lane: Hey, what do you think of this solution?  https://github.com/saltstack/salt/issues/13657
00:41 Ryan_Lane ah, figured it out. salt.verbose needs to be set in vagrant file
00:41 Ryan_Lane UtahDave: would listen_in also be there?
00:42 Ryan_Lane I almost always use the _in requisites
00:42 Ryan_Lane since it makes a lot more sense to put it on the config files than the service
00:42 UtahDave It's not a requisite
00:43 manfred Ryan_Lane: looke like listen is a state, not a requisite
00:43 manfred UtahDave: kind of like the events state?
00:43 Ryan_Lane having to add all the config files on the one state isn't really going to work so well
00:44 Ryan_Lane what if we have config files in 30 sls files that reference a service defined in one sls file?
00:44 Ryan_Lane it isn't a good separation of concerns
00:44 UtahDave hmm.  good point
00:44 Ryan_Lane it makes way more sense to have the 30 states say "I need to restart this thing"
00:44 Ryan_Lane you could even if it: "if this thing exists, restart it"
00:45 Ryan_Lane it would be nicer to have a requisite that did no reordering that acted like watch
00:46 kermit joined #salt
00:46 UtahDave OK, we're rehashing this.
00:47 UtahDave Ryan_Lane: so if there is a change that would require an apache restart, when should that restart happen?  Always at the very end of the entire state run?
00:47 Ryan_Lane yes
00:48 Ryan_Lane if someone wants it to happen in order in the file, they can explicitly have that occur at that point
00:48 otter768 joined #salt
00:48 Guest6143 joined #salt
00:49 arnoldB there was an  update of the salt debian packages to 2014.1.5+ds-5~bpo70+1. anyone knows why?
00:49 arnoldB 2014.1.5.x -> 2014.1.5.y
00:49 manfred arnoldB: joehh is the packager
00:49 Ryan_Lane UtahDave: I updated the issue with a synopsis of what I said in here.
00:50 TheThing joined #salt
00:50 * UtahDave reading
00:50 arnoldB state.sls is suddenly broken =)
00:51 arnoldB traceback: https://np.adm.sx/view/38289419
00:52 bhosmer joined #salt
00:53 Shenril joined #salt
00:53 agliodbs is there a way to test a complex pillar without running the highstate?
00:53 arnoldB agliodbs: pillar.get you:pillar
00:53 agliodbs thanks!
00:55 arnoldB well I forgot to enable the file_roots definition in the master config
00:55 Luke_ joined #salt
00:55 arnoldB then this exception should be catched earlier with an human readable error message
00:58 ajolo__ joined #salt
00:58 UtahDave Ryan_Lane: OK, so rehashing is done.
00:59 UtahDave listen and listen_in will be called just like watch and watch_in, but they won't execute the mod_watch function until the end of the state run and won't change the order of anything
01:00 UtahDave So they are not requisites.
01:00 Sway|Away joined #salt
01:00 Guest6143 joined #salt
01:01 Ryan_Lane UtahDave: yeah, that sounds like a good approach
01:01 UtahDave Yeah, I think it's pretty straightforward.
01:01 bhosmer joined #salt
01:02 UtahDave I think the listen.changes state module was premature optimization.
01:02 Guest6143 joined #salt
01:04 Guest6143 joined #salt
01:06 ajolo_ joined #salt
01:06 bhosmer joined #salt
01:08 Guest6143 joined #salt
01:09 meteorfox joined #salt
01:10 Blunt joined #salt
01:16 fllr joined #salt
01:19 davet joined #salt
01:19 joehh arnoldB: it is in the changelog, but latest was to renable connection with 2014.1.4 and earlier minions
01:19 mgarfias i'm having trouble getting pillar data to show up now that i just moved to aws (from rack)
01:20 mgarfias master and minion on 2014.1.5
01:21 Blunt Hello,gays,I have a question about install Windows .exe Program,and I know salt can do it silent by add install_flags: ' /S' or someelse.But,my .exe still need input something,what should I do?Thank guys.
01:22 fllr joined #salt
01:24 joehh arnoldB: http://paste.debian.net/106465/
01:24 arnoldB joehh: alright, sorry for my lazyness. I was having to much Salt today :)
01:24 l0x3py joined #salt
01:25 mgw joined #salt
01:25 joehh no worries - not trying to say you should have looked there first - very happy for discussion here re updates
01:26 l0x3py joined #salt
01:27 joehh there were 3 patches applied since 2014.1.5 to the debian and ubuntu packages, mainly based on feedback here
01:27 joehh and usability issues (iptable --help)
01:27 aw110f when using ext_pillar with git, will it conflict with pillar_roots?
01:27 aquinas joined #salt
01:28 shaggy_surfer joined #salt
01:29 joehh actual patches: http://paste.debian.net/106466/
01:31 joehh I wonder what the best way to communicate this is?
01:32 joehh maybe something on debian.saltstack.com
01:33 bhosmer_ joined #salt
01:33 arnoldB joehh: If I would have tried it, I would have found the changelog. :)
01:35 catpig joined #salt
01:38 tkharju2 joined #salt
01:41 ajprog_laptop joined #salt
01:47 mgw joined #salt
01:47 googolhash joined #salt
01:53 fllr joined #salt
01:54 brain5ide__ joined #salt
01:55 TaiSHi Does sls matches basic regexp ?
02:01 bhosmer joined #salt
02:02 HeadAIX joined #salt
02:07 googolhash joined #salt
02:11 mgw joined #salt
02:14 CeBe2 joined #salt
02:14 jalaziz joined #salt
02:15 fllr joined #salt
02:15 googolhash joined #salt
02:22 jalaziz joined #salt
02:22 Guest6143 joined #salt
02:23 ipalreadytaken joined #salt
02:24 Ryan_Lane joined #salt
02:24 mgw joined #salt
02:25 Guest6143 joined #salt
02:26 ipalreadytaken joined #salt
02:26 Guest6143 joined #salt
02:28 ajolo__ joined #salt
02:28 jalaziz joined #salt
02:30 MatthewsFace joined #salt
02:31 HeadAIX joined #salt
02:31 jcsp1 joined #salt
02:31 Guest6143 joined #salt
02:31 CeBe joined #salt
02:33 Guest6143 joined #salt
02:35 jalaziz joined #salt
02:37 savvy-lizard joined #salt
02:38 anuvrat joined #salt
02:45 savvy-lizard joined #salt
02:47 jalaziz joined #salt
02:49 druonysus joined #salt
02:50 ajolo__ joined #salt
02:53 Guest6143 joined #salt
02:54 jalaziz joined #salt
02:56 baffle joined #salt
02:57 Guest6143 joined #salt
02:58 jerrcs joined #salt
02:59 hvn joined #salt
02:59 zirpu joined #salt
02:59 Ahlee joined #salt
03:00 ipalreadytaken joined #salt
03:00 JPaul joined #salt
03:06 googolhash joined #salt
03:09 catpigger joined #salt
03:09 marnom joined #salt
03:23 yomilk joined #salt
03:23 ipalreadytaken joined #salt
03:27 ramishra joined #salt
03:29 jalaziz joined #salt
03:35 dude051 joined #salt
03:37 ramishra_ joined #salt
03:41 jalaziz joined #salt
03:43 obimod joined #salt
03:44 obimod ozmegod - "Failed to render "/srv/reactor/created.sls"
03:44 obimod before that it sayd "Jinja search path: ['/var/cache/salt/master/files/base']
03:45 obimod salt-cloud -m /etc/salt/cloud.map deploy (to digital ocean) i'm trying to react with a state.highstate call and I'm getting these errors
03:48 obimod i've added the update_cachedir to True inside /etc/salt/cloud ... i'm wondering if that's messing it up by looking elsewhere for the /var state file
03:48 obimod (also diff_cache_events: True)
03:48 obimod it's pretty much the same thing as in here... http://docs.saltstack.com/en/latest/topics/reactor/index.html ... so i'm stumped. i'll try to remove those cache vars and try again.
03:49 obimod good talk, good talk
03:50 bhosmer joined #salt
03:50 manfred can you share your created.sls?
03:51 notbmatt good news, everyone!
03:51 obimod http://pastebin.com/BGts8h7s
03:51 obimod what?!
03:51 notbmatt my work's legal dept has given me permission to contribute my bug- er, my code to salt!
03:51 vbabiy joined #salt
03:51 obimod nice!
03:51 mosen joined #salt
03:52 manfred obimod: why not just use startup_states: highstate
03:52 obimod in /etc/salt/minion?
03:52 manfred yeah
03:52 obimod doh
03:52 manfred or just in your profile
03:52 obimod #facepalm
03:52 manfred obimod: in the cloud profile
03:52 manfred minion:
03:52 manfred startup_states: highstate
03:52 obimod cloud.map, right?
03:52 obimod damn
03:52 manfred or in cloud map
03:52 obimod arrgg
03:53 manfred could just put it in /etc/salt/cloud though
03:53 obimod thank you
03:53 obimod my head is only bleeding a little bit
03:53 manfred obimod: i think you also want data['data']['id'] but i have to go check my reactor states
03:53 obimod if that's the case some docs need to be updated
03:53 manfred hrm... nope
03:53 obimod oh okay
03:53 cyrusdavid joined #salt
03:53 obimod cool
03:53 manfred obimod: data['name'] maybe?
03:54 obimod that sounds right... it's what i had there to start with. i'll go on that for now
03:54 manfred it is what I have back when I was trying to delete the mine before adding flush_mine_on_destroy:
03:55 manfred but yeah, that should work fine
03:55 manfred obimod: this is the data passed to the reactor for digital_ocean https://github.com/saltstack/salt/blob/develop/salt/cloud/clouds/digital_ocean.py#L476
03:56 manfred data[] will have the name, profile and provider
03:56 manfred args= https://github.com/saltstack/salt/blob/develop/salt/utils/cloud.py#L1314
03:57 mgw joined #salt
03:57 obimod and we're off! launching le cloud
03:57 manfred obimod: the cache portion is more usable when doing something like this one https://github.com/saltstack-formulas/salt-cloud-reactor/blob/master/reactor/autoscale.sls
03:57 obimod wow, that sounds like a bigger pickle
03:58 obimod dill, even
03:58 manfred bread and butter
03:58 possibilities joined #salt
03:58 obimod mmm
03:58 obimod that actually sounds really good right now
03:58 obimod brb
03:58 manfred heh
03:58 ajolo joined #salt
04:03 obimod hokay bak
04:05 n8n joined #salt
04:07 mgw joined #salt
04:15 Chi_ joined #salt
04:15 Chi_ How to mamange many file in folder
04:16 notbmatt file.recurse!
04:16 Chi_ could you recommend me the best way
04:20 manfred Chi_: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.recurse
04:21 obimod hmmm salt '*' state.show_top aint showing shit
04:22 obimod looks like i'm going to have a long night ahead of me : D
04:22 jcsp1 joined #salt
04:23 Guest6143 joined #salt
04:23 obimod i wonder if using a virtualenv is blocking some stuff out
04:26 Guest6143 joined #salt
04:27 ajolo_ joined #salt
04:30 MZAWeb joined #salt
04:31 Guest6143 joined #salt
04:36 renothing joined #salt
04:37 renothing hi,there, how can I config minion_id on minion by module or function or extral script?
04:37 Guest6143 joined #salt
04:37 manfred renothing: i use
04:37 manfred /etc/salt/minion.d/id.conf:
04:37 manfred file.managed:
04:38 manfred - content: minion_id='something'
04:38 manfred but there isn't a module
04:38 obimod manfred: might you know why my cloud minion isn't returning anything for salt '*' state.show_top ?
04:39 obimod it's not setting things up, iptables in particular
04:39 manfred obimod: can you do salt \* test.ping?
04:39 obimod yup i'm able to send commands
04:39 manfred obimod: you shouldn't have to touch the firewall on the minion
04:39 obimod checking logs and what not
04:39 obimod i'm a paranoid one
04:39 manfred obimod: if test.ping returns, but show_top doesn't it sounds like that minion isn't getting assigned correctly
04:40 obimod id?
04:40 manfred obimod: you can shut off all INPUT rules on the minion, and still have it work
04:40 obimod until the zmq has to reconnect
04:40 manfred obimod: zmq only runs on the master
04:41 manfred you can shut off all ports on the minion
04:41 manfred it listens across the network to the master
04:41 obimod persistent connection?
04:41 manfred yes
04:41 obimod keepalive type of thing?
04:41 obimod cool
04:41 manfred no idea, but it doesn't require the minion to have any ports open
04:41 manfred only the master
04:41 manfred 4506
04:41 obimod i'm more worried about closing them : )
04:42 obimod mailserver got hacked a year ago
04:42 obimod i've learned a lot since then! : )
04:42 manfred then shut them all on the minion
04:42 manfred the only thing that is required to be open is 4506 on the master
04:42 manfred or whatever your port on the master is
04:42 Hell_Lap Hmm.. think I'm doing this wrong.. thinking I should write a script to do all my generated keys/passwords out to a pillar file and just read those rather than have the nodes export what they need to share out through grains
04:42 obimod gotchya
04:42 obimod so there's an assignment problem
04:42 obimod hmmm
04:42 l0x3py joined #salt
04:43 manfred obimod: does salt \* test.ping work?
04:43 renothing I didn't config minion_id on minion. but when I see Unaccepted Keys on the master, the minion didn't use its hostname as minion_id, but it used the client ip reserved record name.
04:43 obimod manfred: yup
04:43 manfred obimod: and where is your top.sls file?
04:44 obimod it's in /srv/salt/top.sls
04:44 manfred please share it as well as your minions name
04:45 felskrone joined #salt
04:46 renothing and I don't think using salt to manage minion_id is an good idea. because when the system first boot there is no salt master communitication at all.
04:47 renothing can I config minion_id by script results ? just like config master address , it can be config as func type.
04:47 ramteid joined #salt
04:49 Chi_ I have salt source with many file and permistion. How to keep it on minion like source
04:49 manfred obimod: does it work if you run it as salt-call from the minion?
04:50 obimod oooo
04:50 Chi_ I have problem with owner and permission because it very complicated
04:50 obimod Function state.highstate is not available
04:50 obimod salt '*' cmd.run 'salt-call state.hightstate'
04:50 Chi_ can not define step file by file
04:50 manfred obimod: add -l debug
04:51 manfred sounds like something is erroring out on the minion for some reason.
04:51 manfred But i am done for the evening, gotta get some sleep
04:51 obimod but of "Loaded <mod_name> as virtual <shorter_mod_name>" debug messages
04:51 obimod bunch*
04:52 obimod aww... you should have some bread and butter!
04:52 obimod i've upgraded
04:52 obimod cheese and egg
04:53 obimod egg with fried cheese on it to be exact
04:53 Chi_ I have salt source with many file and permission. How to keep it on minion like source
04:54 Chi_ I have problem with owner and permission because it very complicated
04:54 Chi_ can not define step file by file
04:54 Chi_ any one have solution for this
04:54 Chi_ pls help me
04:54 obimod file.managed
04:54 obimod google "salt file managed state"
04:54 Chi_ you mean define step file by file
04:55 obimod there is also file.recurse
04:55 schimmy joined #salt
04:55 obimod which is probably what you're looking for
04:55 obimod http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html
04:57 Chi_ thank you so much
04:59 schimmy1 joined #salt
05:00 obimod manfred: wow, i was executing state.hightstate
05:01 obimod sleep deprivation ftw
05:09 malinoff joined #salt
05:13 wendall911 joined #salt
05:14 DaveQB joined #salt
05:22 borgstrom joined #salt
05:25 wendall911 joined #salt
05:36 Katafalkas joined #salt
05:38 borgstrom joined #salt
05:39 bhosmer joined #salt
05:39 MZAWeb joined #salt
05:41 tinuva joined #salt
05:53 borgstrom joined #salt
05:56 thayne joined #salt
05:59 picker joined #salt
06:08 borgstrom joined #salt
06:10 chuffpdx joined #salt
06:18 borgstrom joined #salt
06:27 TheThing joined #salt
06:30 Katafalkas joined #salt
06:35 bhosmer joined #salt
06:38 schimmy joined #salt
06:42 HeadAIX joined #salt
06:42 MZAWeb joined #salt
06:43 schimmy joined #salt
06:46 ramishra joined #salt
06:56 slav0nic joined #salt
06:57 jhauser joined #salt
07:02 Katafalk_ joined #salt
07:04 Katafalkas joined #salt
07:05 possibilities joined #salt
07:07 jeddi joined #salt
07:09 mastrolinux joined #salt
07:10 garthk joined #salt
07:11 Katafalkas joined #salt
07:13 alanpearce joined #salt
07:23 joehh joined #salt
07:24 artificialexit joined #salt
07:27 ghartz joined #salt
07:28 bhosmer joined #salt
07:34 jhauser joined #salt
07:54 Lomithrani joined #salt
07:55 chiui joined #salt
07:57 yomilk joined #salt
08:01 darkelda joined #salt
08:02 darkelda joined #salt
08:03 babilen Hi all. I am seeing strange behaviour in regard to testing presense events locally with vagrant. Specifically I can stop minions, but I will *not* see a salt/presense/change event when i do that and salt/presense/present still contains the "offline" minion.
08:05 babilen I can't say that I completely understand CkMinions.connected_ids yet, but there seems to be some special cases for minions running on localhost and 127.0.0.1 is indeed *one* address in grains['ipv4'] ...
08:05 possibilities joined #salt
08:05 babilen Do you know how I can get around this and get presence (well, presense) events working with vagrant?
08:07 babilen connected_ids seems to check against a list of *remote* addresses returned by utils.network.local_port_tcp() and it might be that that function does *not* return the (bridged) IPs of the hosted VMs, but that is just guesswork.
08:13 wigit Sorry if this is obvious, but I am still a noob at this. I want to do a "salt['pillar.get']('<hardcoded value>:<content of variable>:<hardcoded value>', <default>)". Apparently I'm too dense to see how to get the variable in there.
08:18 Lomithrani joined #salt
08:18 babilen wigit: You *might* be able to use jinja's string formatting to get it in there, but I am not sure if that is early enough in the template rendering
08:20 babilen wigit: http://jinja.pocoo.org/docs/templates/#list-of-builtin-filters → format()
08:20 babilen wigit: If you have to do other fancy things in that state you might want to consider writing it in Python or pydsl directly.
08:21 wigit Thanks - I'll check it out. If all else fails I can fall back to pillar['string'][variable]['string']. But it would be nice to be able to set a default value.
08:22 agend joined #salt
08:22 babilen wigit: ('foo:%s:bar'|format(variablethingmajig)) should work
08:30 babilen Hmm, local_port_tcp works as expected. I am now trying to debug CkMinions.connected_ids(), but would need self.opts for that to work. Where do I get these options from?
08:32 ramishra joined #salt
08:32 Kenzor joined #salt
08:36 wigit babilen: sweet - that worked like a charm. Thanks :)
08:37 londo_ joined #salt
08:38 babilen wigit: enjoy :)
08:43 thart joined #salt
08:44 thart Is it possible to use cp.get_file or file.copy with salt-ssh yet? It is not working for me. maybe I have the path structure wrong?
08:46 thart salt-ssh '192.168.1.4' cp.get_file "salt:///root/test.txt" /root/thart/
08:47 ml_1 joined #salt
08:48 kiorky joined #salt
08:50 Shish joined #salt
08:51 Ruben joined #salt
08:54 Lomithrani is there a grains for minion's main ip ?
08:54 babilen Lomithrani: What is the main IP?
08:55 babilen Lomithrani: But no, there isn't. You can get ipv4 and ipv6 but both are lists
08:55 Lomithrani or the IP anyway
08:55 babilen Lomithrani: Well, out of those lists: Which one is it and how do you determine that?
08:55 malinoff Lomithrani, how can you determine the main IP without using salt?
08:56 Lomithrani well not main , i didnt use the right word
08:56 Lomithrani but the ip on eth0 that is public
08:56 elfixit joined #salt
08:56 antonw joined #salt
08:56 babilen Lomithrani: I mean you could filter that list and remove local IPs (127.0.* for example, then other private ones (192.168.*, 10.10, ...)) ... but in the end you have to write some logic. So, what is *your* logic?
08:57 malinoff Lomithrani, again, how can you determine that ip assigned to eth0 is public?
08:57 babilen Lomithrani: Do you use aliased interfaces? But you can get the IP on eth0 from your grains.
08:57 Lomithrani well I Can filter like babilen said , that seems fair as I know the first numbers wont change
08:58 babilen Lomithrani: You can get a list of IPs on eth0 from ip_interfaces:eth0
08:59 babilen But in the end you will have to decide which address you want to use there. I essentially filtered all http://en.wikipedia.org/wiki/List_of_assigned_/8_IPv4_address_blocks which typically results in a single unique IP, but then you might use multiple aliased ones on the same interface
09:00 Lomithrani ok thanks :)
09:01 the_lalelu Lomithrani: if you do logic in python you can use IPy and get the info from "IPy.IP(<your_ip>).iptype()" - it is one of "PRIVATE" or "PUBLIC"
09:02 babilen Yeah, that is exactly what I did
09:05 possibilities joined #salt
09:06 the_lalelu babilen: "There should be one-- and preferably only one --obvious way to do it." ;)
09:07 babilen Well, IPy is the obvious choice. I pondered that decision for all but five minutes as I wasn sure that I want to introduce the dependency, but then ...
09:07 malinoff Do you really need to write that login in python, and use external library for that? {% if not ip.startswith('127') or not ip.startswith('10') or not ip.startswith('192') %} public {% else %} private {% endif %}
09:08 giantlock joined #salt
09:08 malinoff and* of course
09:09 babilen malinoff: You cannot easily filter lists in jinja and writing that in Python is *much* easier than the horrible mess jinja turns into as soon as you have to introduce some logic
09:09 malinoff It is much easier if you're a python programmer, yes
09:10 babilen indeed
09:11 babilen Does anybody know under which circumstances minions will be removed from the minion cachedir ?
09:11 babilen I tracked down my earlier issue and the reason why I see faulty results is that the /var/cache/salt/master/minions/$ID/data.p is still present even though I stopped the minion.
09:12 the_lalelu malinoff: afaik there a public addresses starting with "192" - so your filter logic fails.
09:12 babilen So either the code shouldn't rely on cached information or the minions should have been removed from the cachedir
09:12 malinoff the_lalelu, whatever - it's just an example
09:12 babilen the_lalelu: There are lots of corner cases and I much rather rely on some externally maintained library that does the right thing that trying to shoehorn that myself.
09:13 babilen :)
09:13 babilen There always are ..
09:14 Sp00n joined #salt
09:17 bhosmer joined #salt
09:17 TheThing joined #salt
09:20 Sypher_ joined #salt
09:20 babilen So, nobody knows when minions are removed from the cachedir?
09:20 the_lalelu i've neveer seen a minion removed from cachedir. ;)
09:23 ramishra joined #salt
09:23 babilen the_lalelu: Please elaborate. How is that data supposed to reflect the presence of minions in that case?
09:23 happytux joined #salt
09:27 Lomithrani {% set ip = salt['network.ip_addrs']('eth0')[0] %}    what's wrong with this ? :)
09:29 eofs is this valid? - require: - file: /foo - file: /bar
09:29 eofs (with correct newlines ofc)
09:29 TheThing joined #salt
09:29 student_ joined #salt
09:32 student_ Hi all, I have some suggestion. What do You think about using git submodules in gitfs? Using git submodules in gitfs would be possible to separate access between independent states for different groups of minions. In this case We have one repo with the common states and some submodules for user-dependent states and We can allowing access to the submodule without access to all states. As far as I could see this feature can be impleme
09:33 student_ t-python part. I can do it, but have a one question. Is submodules pulled in all cases or if defined special variable in config (like pull_with_submodules)?
09:34 student_ sorry for participation
09:37 ninkotech_ joined #salt
09:38 briner Ahlee, you advise me to run the state.sls written in #!py outside of salt to better debug it. How could I do to get the ability to have such variable as __pillar__ working out of salt.
09:41 the_lalelu maeh - i dont like the "magic" __salt__, __pillar__, etc. pp stuff.
09:42 malinoff the_lalelu, +1
09:42 Lomithrani {{ salt['network.ip_addrs']('eth0')[0] }} just throws me a jinja error  Unable to manage file: Jinja variable function object has no element 0; line 4    , but if I type salt'*' network.ip_address I do have results ... :(
09:42 the_lalelu i prefer explicit imports (and my IDE too) ;)
09:42 malinoff the_lalelu, __opts__ is the most painful dunder dict :)
09:43 rjc joined #salt
09:46 TheThing_ joined #salt
09:48 babilen Lomithrani: {{ salt['network.ip_addrs']('eth0')|first() }} ? Or: {{ salt['pillar.get']('ip_interfaces:eth0')|first() }} ?
09:49 Lomithrani actually is the eth0 I don't know why I switched to salt ['network.ip_addrs']()[0]
09:49 Lomithrani and it works fine
09:49 Lomithrani thanks
09:50 babilen yeah, network.ip_addrs() doesn't take that argument.
09:50 babilen Hmm, no it should.
09:54 picker joined #salt
09:56 babilen Lomithrani: What does "salt 'FOO' network.ip_addrs" give you?
09:57 babilen (and what about "salt 'FOO' network.ip_addrs eth0"?)
09:58 anuvrat joined #salt
09:58 Lomithrani babilen no argument = works eth0  => no results
09:58 Lomithrani as expected seeing the resulst in the jinja file
10:00 babilen Lomithrani: What does "salt '*' grains.item ip_interfaces" give you?
10:01 babilen http://paste.debian.net
10:01 Lomithrani "net0"
10:01 Lomithrani is the name of the interface
10:01 babilen Which is obviously not eth0
10:02 Lomithrani yep , seems like smartos is net0 , my debian is eth0 didn't notice that
10:03 babilen .oO( If only we had a system to do distribution specific things? )
10:03 Shish joined #salt
10:05 possibilities joined #salt
10:06 stephanbuys joined #salt
10:07 sdebot joined #salt
10:14 luette joined #salt
10:15 the_lalelu malinoff: if you find a way to import __salt__ and so on from the salt module explicitly, please let me know.
10:15 malinoff the_lalelu, you simply can't
10:15 the_lalelu *sigh*
10:16 malinoff It is because how salt works. Loader injects these dunder dicts here: https://github.com/saltstack/salt/blob/develop/salt/loader.py#L573
10:22 N-Mi hi, I have a question regarding file.append : I added a state to add stuff in /etc/rc.local, which worked fine. I tried to remove the line, and re-run state.highstate, and Salt says that /etc/rc.local is in correct state
10:22 vu_ joined #salt
10:22 N-Mi is file.append the right state to use to be sure a line is present in a file ?
10:23 vu_ joined #salt
10:24 vu_ joined #salt
10:25 N-Mi here is my state : http://dpaste.com/043ZBJA
10:31 vu joined #salt
10:35 TheThing joined #salt
10:41 MZAWeb joined #salt
10:53 bhosmer joined #salt
10:59 anuvrat joined #salt
11:01 luette is there a way to debootstrap salt on a client and preserve already distributed keys while doing that?
11:05 possibilities joined #salt
11:20 Lomithrani joined #salt
11:22 jrdx joined #salt
11:36 hotbox joined #salt
11:37 bhosmer_ joined #salt
11:39 masterkorp hello
11:40 masterkorp [WARNING ] TypeError encountered executing state.highstate: 'bool' object is not iterable. See debug log for more info.  Possibly a missing arguments issue:  ArgSpec(args=['test', 'queue'], varargs=None, keywords='kwargs', defaults=(None, False))
11:40 masterkorp Traceback (most recent call last):
11:40 malinoff masterkorp, http://pastie.org
11:40 masterkorp i am having this problem runing highstate on just one minion that i just recently bootrapped
11:40 masterkorp malinoff: its just a line :)
11:41 masterkorp full pastie is comming
11:41 malinoff masterkorp, thought you wanted to paste the whole traceback :)
11:42 masterkorp heh, thanks for the warning
11:43 masterkorp http://pastie.org/private/fdrtx619wca0iceytvuiq
11:43 masterkorp the traceback
11:43 masterkorp this happens on the minion
11:44 masterkorp salt 2014.1.0-8555-gb46f08c (Hydrogen)
11:44 masterkorp salt version
11:44 masterkorp salt 2014.1.5
11:45 masterkorp everyone else is this one
11:46 babilen basepi: Could you cherry-pick 5d497d69 from 2014.1 into develop?
11:47 MZAWeb joined #salt
11:47 masterkorp was that for me ?
11:48 babilen My comment?
11:48 masterkorp yeah
11:49 babilen No, I would have addressed you if that had been the case
11:50 masterkorp hehe sorry
11:50 babilen No problem :)
11:53 diegows joined #salt
11:57 viq I have a couple backup servers, and config needs to go both on backup servers, and servers to be backed up - and different config depending which backup server the "backupee" is using. How would you go about tying the two together?
11:58 babilen backup.server + backup.client states don't cut it?
11:59 ramishra joined #salt
11:59 viq Using backuppc if that matters. So on backup.server I would have definitions of per-client config; and on client I would need the ssh key of appropriate backup server, and possibly IP of the backup server allowed on firewall - but let's say for now all I need is the ssh key of the appropriate backup server
12:01 viq So I guess I could do in backup.server pass a pillar with 'here's a list of backup clients, generate configs', and to backup.client pass 'here's the key of the backup server that will be serving you'
12:01 viq I was just pondering, whether there's more automated way
12:02 babilen Do you have *some* invariant that you could use to filter your minions? I mean what you describe is exactly what I'd do.
12:02 bhosmer joined #salt
12:03 chiui joined #salt
12:03 viq invariant? As in grain or similiar?
12:04 babilen Yeah, something ... I mean you obviously want "list of minions that you should backup" and "list of servers that will add as backup server" -- Is there any data that you could use to differentiate between these two groups?
12:04 arnoldB viq: without reading the whole text: are you looking for something like this https://github.com/bechtoldt/opennebula-formula/blob/master/opennebula/_oneuser.sls#L89 ?
12:04 babilen I mean this doesn't have to be set in grains, but could be an invariant in their minion-id, the presence of some package or whatever ...
12:05 viq sorry, need to make a call, brb
12:05 possibilities joined #salt
12:08 viq back
12:09 viq babilen: thinking about it, and it's more "those are our machines" vs "those are machines we manage for clients" so nothing obvious jumps out at me at the moment. Maybe a pillar, but then I wouldn't be able to use it in pillars...
12:09 babilen viq: As soon as you have that you could write some code that generates you those two lists (when generating the respective config) and it might even allow you to target the minions in question
12:10 viq arnoldB: "couple backup servers, how to tie them to the machines to be backed up in sensible way"
12:11 babilen My intuition is "Set a grain on all customer boxes" and then write a pillar that uses salt['test.ping'
12:11 viq babilen: yeah, that's what I was thinking right now as well, seems to make most sense...
12:11 babilen and the presence of that grain to generate the "boxes to backup" list and then test.ping + absence of that grain for the "backup server" group
12:11 babilen Or make both explicit
12:11 viq And preferably using states, so "list of boxes": set grain, every other box: make sure grain is unset
12:12 babilen Well, if you want to maintain that grain in a decent way (e.g. via salt-formula) you would have the same initial problem again in that you would then need some invariant for targetting that config
12:13 babilen Could you maybe filter by netblock or something like that?
12:13 alanpearce joined #salt
12:14 bhosmer joined #salt
12:14 viq I'd have to take a closer look, but I think not
12:14 viq But even a formula that would set a grain on an explicit list of hosts could give an advantage, as then you could just operate on the grain in multiple places
12:15 babilen https://github.com/saltstack-formulas/salt-formula allows you to do that
12:16 blarghmatey joined #salt
12:16 viq Hm, pillar that sets a grain that I can later work with, that has potential.
12:17 babilen If only that formula wouldn't contain all those comments in its files it would be perfect, but that allows you to generate minion configs from pillar data (and you would then only have to target your customer boxes in the pillar top.sls)
12:17 babilen But then you would *still* have to maintain an explicit list somewhere.
12:17 babilen Which is fine, but cumbersome
12:18 viq Aye
12:18 babilen In this case you would simply shift it from state's to pillars ;)
12:18 babilen But then such a grain might come in handy for a lot of other things.
12:18 viq Yeah, that's what I was thinking
12:19 babilen And there really is *nothing* that differentiates customer boxes from your own?
12:20 viq I'm not able to answer without having a closer look.
12:20 babilen If there were you could easily write a custom grain to set the "customer: True" grain (or whatever is appropriate)
12:21 viq I'm more thinking owner:Ours or owner:CustomerA or owner:CustomerB and so on
12:21 babilen yeah, that make perfect sense
12:21 babilen *makes
12:21 viq Some hostnames may reflect that, but not all
12:24 babilen We use netblock + minion_id (as fallback, as we tend to use CUSTOMERID-RANDOM), but you might have something else. Hell you could simply even write /etc/customer_id when you bootstrap a new box for a customer and return that as external grain.
12:24 viq And by the same account, I can set backupserver:backup1
12:24 babilen But then the best approach really depends on your infrastructure ...
12:25 viq "bootstrap a new box", I like your optimism ;)
12:25 babilen sure (and you can then use that grain to target backuppc.server to those minions)
12:25 babilen viq: s/bootstrap a new box/get the intern to painstakingly install it by hand/ (better?)
12:26 arnoldB we really need https://github.com/saltstack/salt/issues/6955 get implemented
12:26 viq intern, I like your optimism ;)
12:26 babilen s/intern/viq/ ?
12:26 xmj ah
12:26 xmj viq: ping!
12:26 viq aye, sadly. And a bunch of old and _old_ boxes...
12:26 viq xmj: pong
12:26 babilen arnoldB: yeah
12:26 babilen viq: Well, you get the idea :)
12:28 xmj viq: thx :)
12:30 vbabiy joined #salt
12:36 nebuchadnezzar joined #salt
12:44 viq babilen: pondering if for managing grains it's better to use salt-formula or http://docs.saltstack.com/en/latest/ref/states/all/salt.states.grains.html
12:45 babilen "New in version Helium."
12:45 babilen But then you could just use grains.present in this case
12:45 viq absent and append are, present is available
12:46 babilen Hmm, that seems to be more appropriate yeah
12:46 babilen I wasn't aware of that, but it doesn't really come as a surprise that these states exist
12:47 viq Right
12:47 viq OK, that gives me some framework to start thinking about this
12:47 viq xmj: have you seen? There are comments, that gitlab 7.0 doesn't work with ruby < 2
12:47 jcsp1 joined #salt
12:50 Guest6143 joined #salt
12:56 Guest6143 joined #salt
12:57 jaimed joined #salt
12:58 to_json joined #salt
13:02 nebuchadnezzar joined #salt
13:05 possibilities joined #salt
13:05 happytux joined #salt
13:08 Guest6143 joined #salt
13:10 Guest030 joined #salt
13:15 mikekunze joined #salt
13:16 mVla joined #salt
13:16 mVla Hello everybody
13:17 babilen o/
13:19 mVla I'm struggling pkgrepo module in my state
13:19 mVla http://pastebin.com/Ge8w3HWV -> if someone has a clue.
13:20 Guest030 joined #salt
13:21 happytux joined #salt
13:21 mVla && each time I highstate a minion, it adds another line in apt sourcelist file.
13:22 babilen mVla: Your "name" is wrong. It should be "deb http://cran.irsn.fr/bin/linux/debian wheezy-cran3" shouldn't it?
13:24 vejdmn joined #salt
13:24 thayne joined #salt
13:25 laubosslink joined #salt
13:25 mVla babilen: the trailing / has to be present in the file
13:28 manfred salt-key -A accepts the minions that are trying to be accepted.
13:28 babilen mVla: And python-apt doesn't add it?
13:28 manfred you can preseed the keys to /etc/salt/pki, but i don't know if there is a utility to do it for you like that
13:28 mVla babilen: I'm just testing ....
13:29 babilen mVla: Wait, are you saying that you need that in your sources.list? That shouldn't be necessary for a Debian repository.
13:29 babilen But then who knows what they've done to their repo :)
13:29 mVla babilen: you know .... r-cran is for "R" ....
13:30 babilen yes, I am well aware of that
13:30 babilen It is still a third-party repository that might not be maintained to the same standard as official Debian repositories.
13:30 mVla I mean .... R people are wired
13:31 dude051 joined #salt
13:32 gq45uaethdj26jw6 joined #salt
13:33 kermit joined #salt
13:34 Guest030 joined #salt
13:35 babilen laubosslink: Another option (to putting the keys in /etc/salt/pki) would be to accept certain clients automatically ... See http://docs.saltstack.com/en/latest/topics/reactor/index.html#a-complete-example for an example
13:36 babilen laubosslink: http://docs.saltstack.com/en/latest/topics/tutorials/preseed_key.html + http://docs.saltstack.com/en/latest/ref/configuration/master.html#std:conf_master-autosign_file + http://docs.saltstack.com/en/latest/ref/configuration/master.html#std:conf_master-auto_accept
13:36 srage_ joined #salt
13:36 babilen laubosslink: You are probably looking for "autosign_file" there, but the best solution really depends on your infrastructure and wishes.
13:37 Guest030 joined #salt
13:38 mgw joined #salt
13:39 ramishra joined #salt
13:39 ramishra joined #salt
13:40 ajprog_laptop joined #salt
13:40 Guest030 joined #salt
13:40 mVla babilen: I thing I got it !
13:40 or1gb1u3 joined #salt
13:41 Guest030 joined #salt
13:42 babilen mVla: Hmm?
13:43 mVla babilen: - name : without the trailing slash and "- dist: wheezy-cran3/". The final result is OK.
13:43 vejdmn joined #salt
13:43 Guest030 joined #salt
13:44 babilen mVla: The final slash shouldn't be necessary. What happens if you run "apt-get update" without that? → http://paste.debian.net
13:44 babilen but: yay!
13:44 wigit Hmm - any tips on how to create a bunch of files on the minion from a single template file? My first thought was to define all the variables in a loop in the state file and then use these variables in the template. That doesn't seem to be possible. Is there an easy way to see the name of the file that is currently being processed from within the template? I could move much of the logic there in that case.
13:44 mVla babilen: but it still adds a newline in a file each time I highstate a minion.
13:44 babilen mVla: bastard!
13:46 mVla babilen: when I run apt-get update with "deb http://cran.irsn.fr/bin/linux/debian wheezy-cran3" (without a /) I have : E: Ligne 1 mal formée dans la liste des sources /etc/apt/sources.list.d/r-cran.list (analyse de distribution)
13:46 mVla I'll translate : the line is malformed ...
13:47 aquinas joined #salt
13:47 tristianc|Alt joined #salt
13:47 jas-_ joined #salt
13:48 Guest030 joined #salt
13:48 babilen mVla: Try "deb http://cran.irsn.fr/bin/linux/debian/ wheezy-cran3" and please include the *complete* output of "LANG=C apt-get update" and "head -v -n -0 /etc/apt/sources.list{,.d/*}" on http://paste.debian.net
13:49 sandbenderca joined #salt
13:49 wb024 joined #salt
13:49 mVla babilen: OK ...
13:54 brain5ide_ joined #salt
13:54 Guest030 joined #salt
13:55 quickdry21 joined #salt
13:57 mVla babilen: http://paste.debian.net/106523/
13:57 wb024 hello, i'm trying to run salt on centos 5, does anyone know if there are any performance hits for altering the buffer sizes per the salt troubleshooting docs?
13:59 wb024 or if there are newer noarch or 32 bit rpms for python-zmq and zeromq?
14:00 gq45uaethdj26jw6 joined #salt
14:03 Lomithrani joined #salt
14:04 babilen mVla: Ah, they "forgot" to define a suite ("main", "contrib", "non-free", ...)
14:04 babilen mVla: Shitty repo, sorry for that.
14:05 possibilities joined #salt
14:07 HeadAIX joined #salt
14:07 mVla babilen: thanks
14:08 jcristau babilen: component not suite :)
14:09 gq45uaethdj26jw6 anyone know how to install a package through salt-cloud BEFORE it attempts to send the salt boostrap script? the target image does not have SCP installed by default, so the transfer fails
14:09 gq45uaethdj26jw6 otherwise, any idea on where to add the ssh command so i can add it myself?
14:10 manfred gq45uaethdj26jw6: out of curiosity, does it have ssh?
14:10 gq45uaethdj26jw6 yeah, but ssh does not imply scp
14:11 gq45uaethdj26jw6 ssh works, scp fails. i even know the missing package
14:11 HeadAIX joined #salt
14:12 timoguin then it should definitely be possible to execute a remote command beforehand, but i'm not sure where that would fit in
14:12 gq45uaethdj26jw6 yeah, need to find where those ssh commands are issued... i'm not the greatest dev
14:12 gq45uaethdj26jw6 scp is a shadow dependency that isn't checked for. some fedora and centos images do not have the required packages for scp on certain hosting providers
14:13 timoguin that warrants throwing things, IMO
14:13 gq45uaethdj26jw6 you have no idea
14:13 TheThing joined #salt
14:13 HeadAIX joined #salt
14:14 alanpearce joined #salt
14:14 gq45uaethdj26jw6 anyone have an idea where i can issue the command to install the package through ssh? don't know where to find that
14:16 babilen jcristau: Well, the lingo is different everywhere, but meh ... Policy refers to them as "archive areas"
14:16 anuvrat joined #salt
14:16 babilen But component is probably better, yeah
14:18 wendall911 joined #salt
14:18 babilen err, suite naturally not
14:18 viq gq45uaethdj26jw6: https://github.com/saltstack/salt-bootstrap/blob/develop/bootstrap-salt.sh#L227
14:18 babilen Argh, bloody R people
14:19 timoguin viq: don't think that will help him because that script is copied over to the host first
14:19 gq45uaethdj26jw6 viq: that actually is a parameter that gets passed to the script after it is uploaded
14:19 gq45uaethdj26jw6 viq: uploading the script is where the error occurrs. that function never gets reached if scp is not installed
14:19 viq gq45uaethdj26jw6: http://www.packer.io/ and build your own images?
14:20 gq45uaethdj26jw6 viq: not reasonable for my environment
14:21 gq45uaethdj26jw6 viq: some of my providers do not allow for custom images, i believe
14:22 viq https://github.com/saltstack/salt/blob/c36544472304163ad3e401ebe35034b6ef5741cb/doc/topics/cloud/deploy.rst
14:23 jnials joined #salt
14:23 viq there's --no-deploy, and then you could ssh in and "do stuff"...
14:25 tyler-baker joined #salt
14:25 gq45uaethdj26jw6 would be alright if I could automatically run an ssh command on all non-deployed boxes, then deploy on them. i'm basically running a freaking stress test with hundreds of vps's and like 8 providers
14:26 viq https://github.com/saltstack/salt/blob/3c85ffaa1803d05fafa7205e101137ee5ecb2af4/salt/cloud/clouds/saltify.py
14:26 gq45uaethdj26jw6 anything i have to do manually is prohibitive
14:26 viq this seems to be what uploads the scripts
14:26 manfred gq45uaethdj26jw6: i was just thinking about looking into salt.utils.cloud.bootstrap() and switching it to use sftp instead since ssh does imply sftp
14:26 gq45uaethdj26jw6 manfred: that is probably a good design decision anyway, and it would completely fix my problem
14:28 manfred yar
14:28 manfred gq45uaethdj26jw6: i was thinking about it the other day
14:28 gq45uaethdj26jw6 manfred: i was also notified of some changes to the linode API. i have accounts on I think every providers salt-cloud supports. would it be helpful for me to try and let you guys know when I see changes like that?
14:29 manfred yessir
14:29 manfred issue tracker that shit
14:29 gq45uaethdj26jw6 cool
14:30 manfred (unfortunately i only have a rackspace account, but techhat is the true salt-cloud maintainer, and thus has all the salt cloud accounts to look at that)
14:30 gq45uaethdj26jw6 rgr
14:30 taterbase joined #salt
14:30 Lomithrani joined #salt
14:30 briner how can we test a sls in "#!py" outside of salt.
14:31 gq45uaethdj26jw6 i'm currently running 200 spread evenly over all providers, so i get a shit ton of email spam about changes to service levels and all that
14:31 viq briner: salt-kitchen ?
14:31 CheKoLyN joined #salt
14:32 viq briner: so kind of within salt, but outside of 'production'
14:32 alanpearce joined #salt
14:35 gamedna joined #salt
14:35 nmistry joined #salt
14:36 briner sorry, I did not express myself well. I'm wanting to debug a state in #!py. And the only way I've got to debug it, is to do something as open»("/tmp/salt").write( stuff_to_debug ) and to log on the minion to see the output
14:36 briner Is there a way to do at least some prints or to get the debugger.
14:38 danielbachhuber joined #salt
14:38 jaimed joined #salt
14:39 jalbretsen joined #salt
14:39 briner viq, any idea
14:40 tedski if a schedule is in the minion's pillar, does the scheduled job run independently of the master?
14:40 tedski in other words, is the master involved with scheduled jobs?
14:41 tedski i'm assuming a scheduled job results in salt-call or equivalent thereof
14:41 gq45uaethdj26jw6 manfred: as far as switching to sftp, if that's gonna happen, I can drop the issue until the patch. is there any way I can help?
14:45 viq briner: sorry, outside of my knowledge
14:46 TyrfingMjolnir joined #salt
14:47 smcquay joined #salt
14:48 briner viq. thanks anyway
14:48 manfred gq45uaethdj26jw6: create the issue and tag me in it, @gtmanfred
14:48 ecdhe Can I have multiple cmd.runs under a single state name?
14:49 viq ecdhe: no
14:49 Lomithrani what in your opinion would be the best way to count the number of minion responding to a certain cpre for instance "Db*"
14:49 ecdhe thanks viq.
14:49 manfred gq45uaethdj26jw6: i am slammed at work right now figuring out why these two RHCS clusters died over the weekend, but I should have time tonight to look at it
14:49 kaptk2 joined #salt
14:49 gq45uaethdj26jw6 manfred: no worries!
14:49 viq ecdhe: generally it's module.function, and under a single state name you can have only one instance of module
14:50 ecdhe viq, I've got my desktop salted.  I've slowly been automating everything I do, inlcuding git config user.email and user.name.
14:51 ecdhe I thought I might do two cmd.runs under a single state called configure-git.
14:51 ecdhe Sounds like I could do cmm && cmd
14:51 ecdhe But I'll just do two states instead.
14:51 Lomithrani I was thinking about using mine but I thing there might me a better solution with salt-key
14:51 Lomithrani or test.ping
14:52 viq ecdhe: you are aware that if you have it in cmd.run those commands will run every single time you run this state?
14:52 ecdhe viq, I've got an "unless" that watches my .gitconfig
14:53 viq Or that, yeah ;)
14:54 borgstrom joined #salt
14:55 ecdhe Thanks again viq!
14:55 jas-_ joined #salt
14:57 miqui_ joined #salt
15:00 mgw joined #salt
15:05 mgw Are there any known issues with master_tops on 2014.1.5? My module is loading on startup, per the debug log, but the top function is never being invoked. This was working fine on another system under 2014.1.4.
15:06 jcsp1 joined #salt
15:10 Guest030 joined #salt
15:12 darkelda joined #salt
15:12 darkelda joined #salt
15:12 rlarkin lxc-list
15:12 xt No such file or directory
15:12 viq Make me!
15:12 rlarkin set FocusFollowsEyes=True
15:14 mgw rlarkin: what version of lxc do you have installed?
15:14 mgw usertools that is
15:15 rlarkin 0.8.0-rc1 on wheezy
15:15 rlarkin haven't looked at that newfangled Docker thing yet.
15:15 mgw Is there an issue with it?
15:15 rlarkin no, no issues, but I'm using my own template
15:16 mgw oh, ok, i saw xt say "No such file" and got confused.
15:16 bhosmer_ joined #salt
15:16 xt :D
15:16 mgw lxc-list does not exist with usertools 1.0
15:16 rlarkin oh really
15:16 mgw lxc-ls --fancy is better anyway
15:16 Ahlee let's talk targeting on dynamic data from a minion.
15:17 mgw rlarking: lxc is much more robust under 1.0
15:17 mgw but I don't know if 1.0 packages are available for wheezy
15:17 Ahlee I want to be able to target hosts based on their multicast membership (output of netstat -g) - thoughts?
15:18 rlarkin thanks mgw.  I'll look at updating
15:18 Lomithrani joined #salt
15:18 viq Ahlee: custom grain that does that?
15:18 viq Ahlee: unless you already can see it in grains
15:18 Ahlee viq: do custom grains fire ever time?
15:18 Ahlee as multicast memership isn't exactly static
15:18 viq Ahlee: no, they're cached, AFAIK for configurable time
15:19 Ahlee interesting
15:19 Ahlee that might be fine grain enough
15:21 Guest030 joined #salt
15:22 stevednd is there any way to import all variables declared in a jinja template without having to specify each one?
15:22 Guest030 joined #salt
15:23 mgw rlarkin: you might have to wait for the next major release of salt
15:23 mgw to upgrade
15:23 mgw unless you want to use backported lxc functionality
15:23 mgw I don't think 2014.1 works very well with lxc 1.0, but I'm not 100% on that
15:23 mgw I backported lxc modules from develop
15:24 rlarkin i'm not in a hurry. we built up our environment so fast, I have a universe of prototypes and proof-of-concepts to fix / replace.
15:24 Guest030 joined #salt
15:25 xmj viq: in that case - don't run gitlab on platforms where ruby < 2.0
15:25 xmj :)
15:25 viq xmj: "don't run gitlab on released distros" ;P
15:25 * xmj chuckles
15:26 Guest030 joined #salt
15:26 xmj viq: *cough* FreeBSD ships with ruby 1.9, 2.0 and 2.1
15:27 borgstrom joined #salt
15:28 viq xmj: shame it isn't on the list of supported systems ;P
15:29 bemehow joined #salt
15:29 xmj viq: that's actually something _they_ need to fix, I'll neg them.
15:30 xmj thanks for reporting
15:30 ramishra joined #salt
15:30 mgw is anybody here using master_tops on 2015.4.5?
15:30 mgw s/2015/2014/
15:30 mgw lol
15:30 mgw 2014.1.5
15:31 Guest030 joined #salt
15:31 * ecdhe *gasps* -- is this online evidence of time travel?
15:31 mgw ecdhe: no, just typing too fast
15:31 mgw sorry to disappoint
15:32 bemehow_ joined #salt
15:34 borgstrom joined #salt
15:35 ksk hey. anyone in here who knows about halite? i mean whats its purpose? you can fire a command and see its responses. close the browserwindow and everything is lost...
15:36 ksk has everyone thought about logging? or maybe using salt-eventsd as a backend for the GUI?
15:36 wb024 anyone running salt in rhel 5?
15:36 bemehow joined #salt
15:38 manfred wb024: dimeshake was for some things iirc
15:38 dimeshake aye
15:40 borgstrom joined #salt
15:40 thayne joined #salt
15:40 wb024 dimeshake, are you using rpms to install?
15:41 Guest030 joined #salt
15:41 chiui joined #salt
15:41 wb024 manfred, thanks :)
15:41 manfred np
15:42 dimeshake wb024: yes, from epel
15:43 martoss joined #salt
15:43 Ahlee wb024: same, epel source rpms recompiled locally
15:43 Ahlee along with python26 and pyzmq, and zeromq to pull in the 3.x branch to cent5
15:44 wb024 dimeshake, are you modifying the net buffer to bypass the 2.x dc bug?
15:44 wb024 Ahlee, are u installing with pip or yum?
15:44 dimeshake I'm not, I'm running it damn near stock - and haven't hit that bug, that I know of.
15:44 dimeshake have hit others, though
15:44 dimeshake like insane memory usage on certain rhel5 minions
15:44 Ahlee wb024: yum.  I do not use pip anywhere in my environment
15:44 wb024 i'm having trouble building an updated replacement for python26-zmq :/ any pointers?
15:45 wb024 dimeshake, really? all my rhel5 minions dc in ~5min
15:45 wb024 after restart
15:46 Ahlee yeah, that's why I upgraded, let me see if i can find the source rpm I used as the basis
15:46 dimeshake I have 15 boxes reporting fine on CentOS/RHEL5  mix
15:46 meteorfox joined #salt
15:47 wb024 that would be amazing, i found some rpms from salt for rhel5 but they were the wrong arch
15:47 bemehow_ joined #salt
15:47 benturner joined #salt
15:47 wb024 dimeshake, do u know what kind of nic cards?
15:47 dimeshake they're all VMs on rackspace cloud, mostly
15:49 tligda joined #salt
15:49 Ahlee my cent5 hosts are mostly broadcom 10g cards
15:50 Guest030 joined #salt
15:50 ramishra joined #salt
15:50 Ahlee wb024: i've petitioned compliance to allow me to post my modified source RPMs for python26, pyzmq, and zeromq3, but i don't think they meet until July :(
15:51 wb024 we're almost there :p
15:52 Ahlee true!
15:52 wb024 also, all my boxes are broadcom cards too, do u know if u had connection issues with the epel rpms?
15:52 Guest030 joined #salt
15:53 Ahlee I recall having them, yes
15:53 Ahlee hence the rebuild
15:53 wb024 dimeshake, guess u missed out on some fun being in the cloud ;)
15:53 dimeshake :D
15:56 jonbrefe joined #salt
15:56 Guest030 joined #salt
15:57 felskrone joined #salt
15:57 dotplus what's the difference between a state module and an execution module?
15:58 viq dotplus: for one, execution is "do it now" and state is "this host is to look like this"; state modules I believe are usually built on top of execution modules
15:59 dotplus aha, thanks
15:59 Guest030 joined #salt
15:59 viq Also state modules try to be idempotent ("doesn't make any difference how many times you run it"), while execution modules don't
16:00 smcquay so I started seeing msgpack.exceptions.UnpackValueError: Unpack failed: error = 0
16:00 smcquay sorry, I meant to link to https://github.com/saltstack/salt/issues/11690 and then ask what a proc folder is ...
16:01 smcquay wait, found it under job management.
16:01 ramishra joined #salt
16:01 Guest030 joined #salt
16:02 cwright joined #salt
16:02 smcquay I am trying to accept a bunch of vms that have been renamed and am running into issues adding their keys. Turns out my cleanse was incomplete (was just clearing out /etc/salt/minion_id, but it seems there is more scrubbing needed)
16:05 ajolo joined #salt
16:06 rallytime joined #salt
16:07 KyleG joined #salt
16:07 KyleG joined #salt
16:08 possibilities joined #salt
16:08 Heartsbane joined #salt
16:08 Heartsbane joined #salt
16:09 MatthewsFace joined #salt
16:09 dotplus viq: so what's the difference between a state module and a formula then? I was under the impression that a formula was the collection of code used to enforce a declared state?
16:09 dotplus sorry to be uninformed about the jargon, I'm coming from a background of various other cfg mgmt tools but new to salt
16:10 tligda1 joined #salt
16:10 timoguin a state module is the actual python function that enforces some change of state
16:10 timoguin a formula is a collection of SLS files that call various state modules
16:10 viq dotplus: state module is python code that instructs salt how to do something, eg how to add a user or install a package. Formula is same as a state file, a structured data file telling salt _what_ to do, eg what packages to install or what users to create
16:11 dotplus ok, got it, I think. thanks to both.
16:11 mgw Is nobody using the master_tops system? To me it appears totally broken on 2014.1.5: https://github.com/saltstack/salt/issues/13673
16:11 tristianc|Alt joined #salt
16:12 ramishra Hey guys.. I'm trying to run a state with caller.sminion.functions['state.sls']('top')  and it just gets stuck.. Is the syntax correct?
16:13 ramishra running from the minion..
16:13 mateoconfeugo joined #salt
16:14 mateocon_ joined #salt
16:15 ramishra anyone can help tell me if this is correct? complete code I'm using https://gist.github.com/anonymous/5fedd672066099a76a97
16:18 Guest030 joined #salt
16:19 ckao joined #salt
16:19 jcsp1 joined #salt
16:20 mgw ramishra: what error are you getting?
16:20 mgw anything in the minion logs?
16:20 mgw with debug level
16:21 mastrolinux joined #salt
16:22 Guest030 joined #salt
16:23 mastrolinux joined #salt
16:23 ramishra mgw: nothing in the logs... let me set the debug level and check.. I had the understanding caller does not use the minion deamon
16:23 ramishra mgw:it can work without it
16:24 joehillen joined #salt
16:25 Guest030 joined #salt
16:25 jcsp1 joined #salt
16:25 oz_akan_ joined #salt
16:26 Guest030 joined #salt
16:27 ramishra mgw: nothing in the logs with log_level: debug
16:27 bhosmer joined #salt
16:27 ramishra mgw: Is the syntanx correct?
16:27 ipalreadytaken joined #salt
16:28 mgw ramishra, it looks right to me, but I've not used that exact config
16:29 vejdmn joined #salt
16:31 Lomithrani I'm trying to so something that would do something like that ::   {% test= (salt['test.ping']('App*', expr_form='pcre')|length) %}  any idea how I could ?
16:31 basepi babilen: done, sorry for the delay:  https://github.com/saltstack/salt/pull/13676
16:31 Gareth morning
16:32 ramishra mgw: i can call caller.sminion.functions['grains.items']() without any issues.. wanted to know  whoe caller is used with states
16:32 babilen basepi: ta!
16:32 ramishra whoe/how
16:33 Guest030 joined #salt
16:34 ramishra mgw: salt-call 'state.sls' db --local  .. also runs fine
16:35 ramishra from the cli
16:37 Guest030 joined #salt
16:38 dlam joined #salt
16:38 mateoconfeugo joined #salt
16:39 hopefulstranger joined #salt
16:39 hopefulstranger Is there a common pattern for running salt commands without having to SSH to the master?
16:40 Guest030 joined #salt
16:40 Ahlee hopefulstranger: I use the rest API
16:40 kballou joined #salt
16:40 Ahlee formerly salt-api, now netapi
16:40 viq hopefulstranger: salt-api and/or halite and/or rundeck
16:40 schimmy joined #salt
16:41 Ahlee rundeck?
16:41 hopefulstranger thanks guys
16:41 Ahlee you've posted that before, I believe
16:41 Ahlee but it didn't stick
16:41 viq http://rundeck.org/
16:41 Ahlee yeah
16:41 viq specifically http://rundeck.org/plugins/2013/07/26/salt-step.html
16:41 Guest030 joined #salt
16:42 schimmy1 joined #salt
16:42 forrest joined #salt
16:42 mVla left #salt
16:44 ramishra anyone?
16:48 Guest030 joined #salt
16:50 repl1cant hepoful: i use lighttpd + salt-api on two masters behind a single vip
16:50 ksalman joined #salt
16:50 Ahlee I should probably look into putting something in front of cherry-py
16:51 repl1cant i'm using rest_wsgi, works great behind lighttpd
16:51 ksalman Can I not use windows path in jinja template? I get the couple of different errors when I try to do that  - gistfile2.txt and gistfile3.txt https://gist.github.com/anonymous/52910396b19825950bf4
16:52 ksalman It works fine with using linux style paths, gistfile4.txt
16:52 Ahlee repl1cant: thanks
16:52 ksalman somehome it maps /salt/conf/foo.txt to c:\salt\conf\foo.txt?
16:52 shaggy_surfer joined #salt
16:53 viq ksalman: try single quotes instead of double?
16:53 ipmb joined #salt
16:54 mastrolinux joined #salt
16:54 bmatt I expect that's because of how jinja templating works
16:54 Guest030 joined #salt
16:54 bmatt {{ foo }} means "print the literal value of $foo here"
16:55 bmatt so that expands to c:\salt\conf\foo.txt:\n etc
16:55 ksalman viq: same error
16:55 viq It looks like \f means something special
16:55 bmatt try {% set foo = "\'c:/salt/conf/foo.txt\'" %}
16:55 mastrolinux1 joined #salt
16:56 viq ksalman: or "c:\\salt\\conf\\foo.txt"
16:56 Guest030 joined #salt
16:56 bmatt I still think the colon might be a problem
16:56 ramishra joined #salt
16:57 gq45uaethdj26jw6 joined #salt
16:57 shaggy_surfer good morning.
16:57 vejdmn joined #salt
16:58 ksalman it looks like both suggestions, bmatt and viq, works
16:58 shaggy_surfer anyone have a good example of a multi-environment setup for salt… I read the documentation and tried implementing but am still getting an error of "Data failed to compile:   Malformed topfile (state declarations not formed as a list)"  .  I am trying to implement a dev/qa/staging/prod env setup for different state trees.
16:59 timoguin shaggy_surfer: paste your top.sls somewhere
16:59 shaggy_surfer I read this document:  http://docs.saltstack.com/en/latest/ref/states/top.html  and followed the example.. but I can't seem to get it to work and maybe I am not understanding it properly.  Do you need a top.sls in all env's or just the base?  I just have one in the base right now
16:59 ksalman thanks guys
17:00 mgw is there a way to get state.sls to run with the env from tops?
17:00 mgw In other words, as if it were being run as part of a highstate
17:00 mgw it defaults to base unless saltenv is passed as an arg
17:00 earloop joined #salt
17:01 earloop howdy,
17:01 mgw which is fine except then "salt \* state.sls foo" will clobber any system that should not be getting the base sls
17:01 shaggy_surfer here is my top.sls:  https://gist.github.com/mf-collinhayden/7e3cccda67e75472bc56
17:01 troyready joined #salt
17:01 shaggy_surfer right now I am just using base and a test env called trial
17:02 chrisjones joined #salt
17:02 timoguin shaggy_surfer: that looks okay except the last match doesn't have any states defined
17:03 shaggy_surfer does it have to have them?
17:03 shaggy_surfer my conf has the two file roots defined.  base and trial
17:04 timoguin yea it's expecting a list of states for that match, i believe
17:04 shaggy_surfer let me take it out then and test.
17:04 tpaul joined #salt
17:04 shaggy_surfer I would figure it would have given a differnet error if that was the case
17:04 earloop quick question on supervisor
17:04 earloop I have this statement on my sls file
17:04 earloop http://pastebin.com/Ntb5tCSM
17:05 earloop however; salt only starts foo and not bar
17:06 gq45uaethdj26jw6 joined #salt
17:06 earloop maybe I misunderstood supervisord, do I have to place a statement per item that I want supervisor to watch?
17:06 bmatt is supervisord a service (started with init/upstart/etc)?
17:07 bmatt or is it a state module I'm unfamiliar with
17:07 bmatt at, a state
17:07 bmatt er, ah. a state.
17:08 bmatt so each top-level object definition (in your pastebin, you define an object "foo") gets managed and monitored individually
17:08 earloop ok
17:08 bmatt so you're starting foo, and it's watching two configuration files, but salt doesn't know about bar
17:08 tpaul I'm trying to use 'grains.get_or_set_hash' in a state just like the example in the docs, but I get an "object has no attribute 'get_or_set_hash';" error
17:10 earloop ah that clear my suspicions
17:10 earloop thanks bmatt
17:10 Guest030 joined #salt
17:10 bmatt np
17:10 davet joined #salt
17:11 bmatt tpaul: can you share a snippet?
17:15 kermit joined #salt
17:16 jcsp1 joined #salt
17:17 Guest030 joined #salt
17:18 n8n joined #salt
17:19 Guest030 joined #salt
17:19 aw110f joined #salt
17:19 MatthewsFace joined #salt
17:21 shaggy_surfer ok, one step closer , now I am getting this error:  The state "htop" in sls system_tools.htop is not formed as a list
17:21 shaggy_surfer what does "not formed as a list mean?"
17:22 forrest it isn't a list
17:22 forrest can you gist it?
17:22 forrest the state
17:23 shaggy_surfer https://gist.github.com/mf-collinhayden/d6e3c4e07c652d0d2553
17:24 Guest030 joined #salt
17:26 timoguin shaggy_surfer: that colon at the end of pkg.installed tells salt that it's a dictionary and it expects to find elements prepended with a dash
17:26 timoguin aka, a list
17:26 timoguin remove the colon
17:26 anuvrat joined #salt
17:26 timoguin normally you're defining a dictionary and passing all the arguments as a list
17:27 forrest timoguin, are you going to hang out during the sprint?
17:27 shaggy_surfer oh, ic.  It's because it's so simple and has no others that I need ot remove it
17:28 Ryan_Lane joined #salt
17:28 shaggy_surfer thanks… usually my salt states aren't this short… this was just for testing and apparently I don't understand all the syntax yet… thank you.
17:29 timoguin forrest: yea, i just rsvpd on the eventbrite page
17:29 possibilities joined #salt
17:29 druonysus joined #salt
17:29 druonysus joined #salt
17:29 forrest cool, I believe only the salt guys are getting the details on there.
17:29 forrest manfred,
17:29 forrest manfred, check if you can host an event
17:29 forrest I actually remembered to remind you
17:30 timoguin but who's gonna scan my eventbrite ticket!? :)
17:32 forrest You can hold it up to your webcam if you want I guess
17:33 ekristen is there any way with salt cloud to have an instance in ec2 provision with a public IP already attached?
17:33 manfred forrest: hooray, cause i completely forgot, been running around all day
17:34 forrest manfred, cool
17:35 babilen salt's testsuite takes a while to run :)
17:36 Guest030 joined #salt
17:42 arnoldB hm how can I build salt packages using fakeroot debian/rules binary without build the doc (which fails)
17:42 arnoldB ?
17:42 arnoldB dh binary dh_installman
17:42 arnoldB doc/man/salt-api.7: No such file or directory at /usr/bin/dh_installman line 128, <IN> line 36.
17:42 Guest030 joined #salt
17:43 babilen arnoldB: you have to install debhelper and other build dependencies
17:44 * Heartsbane blames whiteinge.
17:44 babilen (or is doc/man/salt-api.7 missing?)
17:44 babilen brb
17:45 arnoldB => https://np.adm.sx/view/2972bffb
17:46 arnoldB actually I'm missing something: ImportError: No module named ext.pygmentplugin
17:46 arnoldB running wheezy here
17:46 forrest it's too early for hate Heartsbane
17:47 AdamSewell joined #salt
17:49 azylman joined #salt
17:49 abe_music joined #salt
17:50 arnoldB I hacked the docs away..
17:50 arnoldB ^^
17:50 dlam is there a salt-call way to test what pillar things will be loaded on a minion?
17:50 dlam i got a staging computer and wanna debug why it aint getting my staging stuff =\
17:52 timoguin salt-call pillar.items
17:52 timoguin or pillar.get for specific keys
17:54 mastrolinux joined #salt
17:54 Guest030 joined #salt
17:55 abe_music hey all, anyone know how to use pkg.installed and prevent the installed packages from automatically starting their services? e.g, hbase?
17:56 TyrfingMjolnir joined #salt
17:56 abe_music hey timoguin, how's it going?
17:56 Guest030 joined #salt
17:56 arnoldB abe_music: this mosrtlydepends on packages scripts within the packages itself
17:57 arnoldB s/mosrtlydepends/mostly depends/
17:57 timoguin abe_music: hey, going good. you?
17:57 abe_music arnoldB, right...i'm finding that when installing hbase via deb files it wants to start automatically, but i haven't written any config files yet, so the service fails which then fails the salt provision
17:58 abe_music timoguin: not too bad, just working hah
17:59 Guest030 joined #salt
17:59 viq abe_music: Ithink there's a debian setting for that
18:00 abe_music viq: cool, googling for something like that
18:01 Guest030 joined #salt
18:01 tkharju3 joined #salt
18:01 viq abe_music: http://ask.debian.net/questions/how-to-prevent-daemons-from-starting-at-installation-time
18:01 timoguin that's the one i was reading. :)
18:01 viq brought by 'debian start service package install' search :P
18:02 abe_music viq: was just reading through that :)
18:02 viq ;)
18:02 abe_music my search "prevent deb from starting services"
18:02 timoguin mine: "debian don't start service on install"
18:03 abe_music heh
18:03 yomilk joined #salt
18:03 timoguin funny how everyone thinks just a little differently
18:03 arnoldB hm I guess you can disable dpkg-configure in the APT config
18:03 Guest030 joined #salt
18:03 arnoldB this will prevent APT from invoking configure/prerm/preinstall/postinstall etc.
18:03 arnoldB not sure you want this
18:03 abe_music could be dangerous :)
18:04 arnoldB True
18:04 abe_music another useful thread if anyone is interested: http://askubuntu.com/questions/74061/install-packages-without-starting-background-processes-and-services
18:09 ramteid joined #salt
18:09 TheThing joined #salt
18:10 tpaul joined #salt
18:11 smcquay joined #salt
18:11 earloop d
18:11 earloop question on salt and lvm
18:12 tpaul Bleh lost network connection or a while! I am having an issue with grains.get_or_set_hash, I have the state and error here: http://hastebin.com/ukopaxejey can anyone point out what I'm doing wrong?
18:12 earloop can you add several devices to a logical volume group?
18:12 bmatt earloop: yes
18:13 earloop http://pastebin.com/60bpDcsK
18:13 earloop is my syntax correct?
18:14 bmatt I'd expect that to work, yes
18:14 bmatt your indentation might be off? not sure how much yaml cares
18:15 bmatt tpaul: is the grains dunder object congruent with the grains module?
18:16 bmatt that is to say, should that perhaps be salt['grains.get_or_set_hash'](...) or something similar?
18:16 cperkins joined #salt
18:16 smurfy__ joined #salt
18:16 bmatt or is there another way to specifically invoke a method from a module?
18:17 tpaul bmatt: Does look like I have some weird formatting in that state, I'll check that and see if it helps
18:20 earloop joined #salt
18:20 ml_1 joined #salt
18:23 joehoyle joined #salt
18:23 mastrolinux joined #salt
18:23 earloop sorry seems that I got kicked out
18:24 earloop I get an exception when trying to add device to my vgs
18:24 earloop http://pastebin.com/XwDKjscf
18:24 kermit joined #salt
18:25 whiteinge babilen: there' shouldn't be a salt-api (7) manpage now that it's merged into salt. there should only be salt (7)
18:25 druonysus joined #salt
18:26 whiteinge arnoldB: ^^
18:26 racooper joined #salt
18:27 rjc joined #salt
18:28 earloop ah corrected my syntax...this is how it should look. Works
18:28 earloop http://pastebin.com/bfHfS1pr
18:29 anuvrat joined #salt
18:31 vu_ joined #salt
18:33 azylman joined #salt
18:35 arnoldB whiteinge: ok
18:38 Guest030 joined #salt
18:41 whiteinge jacksontj: pm?
18:41 Guest030 joined #salt
18:45 Guest030 joined #salt
18:45 jhauser joined #salt
18:45 jacksontj whiteinge: ?
18:46 babilen Hmm, running the entire testsuite takes quite some time. What are sensible ways to cut down on that, while still getting reasonable feedback if what one is currently working on is breaking everything?
18:47 jacksontj you can run specific tests
18:47 jacksontj or groups of tests (unit, integration, etc)
18:47 mprymek joined #salt
18:47 Guest030 joined #salt
18:47 babilen jacksontj: Sure. Keep in mind that I am completely unfamiliar with the testsuite.
18:48 jacksontj ah, yea so if you look in tests there is a "runtests.py" script
18:48 jacksontj its -h is very helpful :)
18:48 babilen But then I can take a look myself and find reasonable subsets of tests to run. How would I specify them when I want to run them in docker?
18:48 jacksontj its something like --docked ubuntu or something like that, the help page on that runtests has the details
18:49 arnoldB what's the best state function to empty a file?
18:49 babilen I'm running it already, just wasn't sure if I can run a subset of tests in docker. But I'll take a look ... But first I'll wait another ten minutes for this run to complete :)
18:49 arnoldB I've used file.managed so far, Salt Helium is complaining that the contents arg is ''
18:50 jacksontj babilen: i think you can run all the same commands with docked -- since it basically just shells out to the same thing inside docker
18:50 jacksontj arnoldB: want it to be empty? or non existant?
18:50 babilen jacksontj: You would expect that. I'll test it.
18:50 arnoldB switched from 2014.1.5 to develop, executed state.highstate and got no errors. *a miracle* :D
18:50 tpaul bmatt: After much fiddling, It appears that grains.get_or_set_hash doesn't like setting a grain with the name "backup"...
18:51 arnoldB jacksontj: empty as empty
18:51 marco_en_voyage joined #salt
18:52 jacksontj hmm, yea i'd do that empty string contents. I'd create an issue if thats complainingn since thats a valid value
18:52 chamunks joined #salt
18:53 arnoldB jacksontj: I guess pyaml/jinja is converting '' to None . I found another solution (simply adding a comment to the file)
18:53 forrest joined #salt
18:54 jacksontj hmm, might need to quote it asusming you have something like 'foo: ' change it to 'foo: ""'
18:56 Guest030 joined #salt
18:57 bmatt tpaul: oh, I suspect there's a name collision there
18:58 Guest030 joined #salt
18:59 shel3over left #salt
19:01 bmatt oh. hm. that's weird. I can't figure out from the code what the issue might be
19:02 tpaul bmatt: same here, it seems to work  when run from the cli too
19:05 bmatt do you already have a backup grain?
19:05 tpaul Okay, disregard. I have something weird going on, now the grain "backup_user" is doing the same thing (it wasn't a second ago) I must have some weird characters sneaking in somewhere
19:06 bmatt rgr
19:07 babilen jacksontj: Is 30 minutes runtime normal?
19:09 jacksontj babilen: yea, for the full test suite
19:09 jacksontj if you look at the jenkins builds its 30-40m usually
19:10 Guest030 joined #salt
19:11 babilen ack, ta
19:11 InAnimaTe joined #salt
19:12 Guest84025 joined #salt
19:12 babilen My personal testuites ran their hundreds of tests within seconds/a minute, but then I guess that salt requires a lot more setting up, mocking and other boilerplate.
19:12 tkharju3 joined #salt
19:15 felskrone joined #salt
19:15 UtahDave joined #salt
19:16 ksalman why does one of the windows minions not show up on master when I do something like salt -G 'os:windows' grains.get id
19:17 forrest ksalman, did you already query that machine for all of it's grains?
19:17 ksalman or salt <minion> test.ping
19:17 forrest lol
19:17 forrest is it even connected to the master?
19:17 arnoldB hmm is Salt setting umask 0077 or so in salt.modules.ssh? trying to figure out why newly created /etc/ssh/ssh_known_hosts (salt.states.ssh_known_hosts) gets 0600
19:18 ksalman forrest: yea, I can do think on the salt-minion itsellf, like deploy a state, using salt-call
19:18 tkharju4 joined #salt
19:18 ksalman s/think/things
19:19 kermit joined #salt
19:20 ksalman deploying state on the salt-minion, I can see it talking to the master, using tcpdump
19:20 alainv hm, if I define different items of a dictionary in two pillar files, do they merge or conflict?
19:20 FeatherKing joined #salt
19:20 forrest salt-call --local? or salt-call? Also can you see the connection from the minion to the master on the master itself?
19:20 forrest seems odd
19:20 alainv e.g. generic instructions for nginx logging setup and node-specific vhost config
19:20 ksalman i can see the minion connection on the master itself
19:21 alainv all under nginx:lookup
19:21 forrest ksalman, strange, is the master a different version than the minion?
19:21 ksalman forrest: no, they are both 2014.1.1
19:22 forrest ksalman, weird, did you already restart the master/minion services?
19:22 FeatherKing is it possible to grab the second item in a particular grain where that grain has more than one value? something like this:  {{ salt['pillar.get']('spacewalk:grains['roles'][1]', {}) }}
19:22 hotbox joined #salt
19:22 Ryan_Lane joined #salt
19:23 mastrolinux joined #salt
19:23 tkharju joined #salt
19:24 FeatherKing where the roles are dev, resin
19:25 bmatt FeatherKing: yes, but you have to be careful about how you structure the data
19:26 FeatherKing jinja does not seem to render the grains directive in that example my syntax must be wrong?
19:26 bmatt salt['pillar.get'](('spacewalk',salt['grains.get']('roles')[1])|join,{})
19:26 FeatherKing bmatt: in our case these grains are custom and brought in from a central web server so the structure we control totally
19:27 hotbox joined #salt
19:27 bmatt you have to construct the argument to pass to pillar.get()
19:27 bmatt think of it like
19:27 FeatherKing that join, is that a jinja thing or python
19:27 bmatt that particular construction is a jinja thing
19:27 ksalman forrest: i did restart the services. would it matter that zmq version on salt-minion is 4.0.4 and zma version on salt-master is 3.2.3?
19:28 bmatt param = 'spacewalk:{0}'.format(salt['grains.get']('roles')[1]))
19:28 bmatt pillar.get(param)
19:28 InAnimaTe joined #salt
19:28 forrest ksalman, yea that could be a problem for sure
19:28 forrest ksalman, what version is running on the other minions? 3.2.3 most likely>?
19:29 bmatt anyway, there are a few ways to do it
19:29 jergerber joined #salt
19:29 ksalman forrest: yes. This is my first Windows minion
19:29 forrest ksalman, ahh ok, yea can you try downgrading it real quick?
19:29 FeatherKing bmatt: let me try this
19:29 ksalman downgrading the minion?
19:29 FeatherKing i felt like i was in the ballpark
19:30 smcquay I am having trouble getting test.ping back from all our machines (only about 400) and have tried increasing worker_threads, as well as specified a long timeout. any thoughts on how to improve the situation?
19:30 ksalman i used the salt package from http://docs.saltstack.com/en/latest/topics/installation/windows.html
19:30 bmatt smcquay: are you confident that your minions are running?
19:30 smcquay yes.
19:31 smcquay I get "did not return in time"
19:31 smcquay and then i check the logs on those machines, and they say:
19:31 oz_akan_ joined #salt
19:32 smcquay ooo, this is different. they say SaltReqTimeoutError, but before it had a message warning me to increase worker_threads.
19:32 smcquay I recycled the master, I wonder if the minions need recycling?
19:32 hotbox joined #salt
19:33 bmatt they shouldn't unless you reissued keys
19:33 smcquay I have not reissued keys.
19:33 forrest ksalman, LET US SUMMON THE ALMIGHTY WINDOWS OVERLORD, MASTER OF THE BITS AND BYTES, UtahDave
19:33 bmatt hm.
19:33 ksalman lol
19:33 UtahDave you rang????
19:33 UtahDave lol
19:33 bmatt UtahDave: Please pick up the blue courtesy phone. UtahDave, please pick up the blue courtesy phone.
19:34 forrest :P
19:34 ksalman hi!
19:34 bmatt O_o
19:34 UtahDave smcquay: did you delete the cached master key on the minions?
19:34 bmatt forrest: that was surprisingly effective
19:34 forrest UtahDave, ksalman is being a horrible human being and trying to run different versions of zeromq
19:34 ksalman =(
19:34 srage joined #salt
19:34 forrest bmatt, hah, he would have noticed at some point and responded anyways, but I have to take advantage of it to be funny
19:34 FeatherKing bmatt: +1 points for you thank you
19:35 * bmatt tallies his score
19:35 bmatt -1029!
19:35 FeatherKing lol
19:35 ksalman UtahDave: I have installed salt-minion (2014.1.1) on windows and it's having trouble talking to salt-master (2014.1.1). master has zmq 3.2.3 and miniona has zmq 4.0.4
19:35 smcquay hmm. UtahDave I may have done that.
19:35 forrest We need like a spare change gittip jar we can shake
19:35 forrest ksalman, that zmq came with the package you said though right
19:35 * bmatt passes the hat
19:35 gq45uaethdj26jw6 joined #salt
19:35 bmatt "Reach deep, my brothers, and help little Timmy pay for some leet dox."
19:35 ksalman forrest: yes, I got the package from http://docs.saltstack.com/en/latest/topics/installation/windows.html
19:36 UtahDave forrest: yeah zmq is included in the windows bundle
19:36 forrest ksalman, cool, just wanted to confirm so Dave doesn't have to scroll back up
19:36 forrest UtahDave, WELL THEN THIS IS ALL YOUR FAULT I BELIEVE GOOD SIR, unless the issue isn't zmq
19:36 forrest in which case, windows
19:36 ksalman so newer zmq on minion could be the issue?
19:36 bmatt something something developers!
19:36 xzarth joined #salt
19:36 ksalman yea, windows..
19:37 UtahDave ksalman: try adding    -t 30   to your command
19:37 oz_akan_ joined #salt
19:37 stevednd has anyone experienced jinja reporting compilation errors on incorrect lines? It's not even off a consistent number of lines between various errors.
19:38 forrest stevednd, can you gist the state and the error?
19:38 ksalman okay I'll try that. Though just to test I removed 2014.1.1 on the minion and installed 2014.1.0 (zmq 4.0.3) and that is behaving okay
19:39 UtahDave ksalman: try 2014.1.5
19:39 joehoyle joined #salt
19:39 stevednd forrest: not really. It's a very large statefile, and I don't want to sanitize it. It's just the typical compilation error that a variable is undefined, or dict object has no attribute
19:40 ksalman UtahDave: on the minion?
19:40 stevednd Rendering SLS "base:app.events.web" failed: Jinja variable 'dict object' has no attribute 'shell'
19:40 stevednd /var/cache/salt/minion/files/base/app/base/init.sls(80):
19:40 stevednd the real error is actually on line 26
19:40 UtahDave ksalman: yeah.
19:40 forrest stevednd, ahh yea sometimes that happens
19:40 stevednd Rendering SLS "base:app.events.web" failed: Jinja variable 'source_cache' is undefined
19:40 stevednd /var/cache/salt/minion/files/base/app/base/init.sls(192)
19:40 stevednd that one is actually line 131
19:40 oz_akan_ joined #salt
19:41 stevednd it all started when I started messing around with import, and import_yaml
19:41 jhauser joined #salt
19:42 stevednd forrest: have you experienced it under similar circumstances?
19:42 forrest stevednd, a lot of people have, it seems to be random sometimes
19:43 smcquay UtahDave: not sure if that was meant for me, but I am using -t120 and still get minion failed to return job info
19:44 UtahDave smcquay: What version of Salt are you on and which OS?  (sorry if someone has already asked you that.  :)   )
19:44 bmatt stevednd: I suspect it has to do with how jinja does imports
19:44 smcquay no worries. debian, 2014.1.5 (just ran package manager updates this morning)
19:44 bmatt stevednd: jinja processes the entire file a bit like the C preprocessor works
19:44 bmatt so you end up with a monolithic file that's passed into the parser
19:45 bmatt and I think there's a disconnect between line identification in the error output and the preprocessed file
19:45 bmatt but that's just a theory
19:45 UtahDave smcquay: ok, good
19:45 smcquay minions are on os x 2014.1.4 ... hmm.
19:46 smcquay I just pip installed the minions, though I would've been okay.
19:46 bmatt stevednd: I'm willing to bet that you include a 56 line file
19:46 bmatt er, 54
19:46 UtahDave smcquay: what are the specs on your master?
19:47 smcquay single core 2.53 GHz xeon, virtualized. 1G rams :/
19:47 smcquay nothing on the system (salt is the only thing on this box) seems to be cheweing up more than half the ram, though, and cpu seems pretty idle.
19:47 forrest ksalman, I'm going to get some food, can you let me know if that works?
19:48 vejdmn joined #salt
19:48 ksalman forrest: willd o
19:48 ksalman will do
19:48 ksalman thanks
19:49 stevednd bmatt: no. That was my first guess, but if you look at both of the errors I posted, and the lines they actually appear on it's a different number for both
19:50 UtahDave smcquay: so all masters and minions are 2014.1.5?
19:50 stevednd also, the file I'm bringing in is 45 lines
19:50 bmatt do you do any looping?
19:50 bmatt {% for ... %}?
19:50 smcquay no, the minions are on 2014.1.4; checking to see why this is the case. I recently installed from pip. I failed to check versions
19:52 smcquay yeah, will start by upgrading my minions. brb.
19:52 UtahDave smcquay: OK, let me know how that goes.
19:53 stevednd bmatt: yes, but not before these errors
19:53 bmatt intradestin.
19:53 stevednd it's not the end of the world, just a pita
19:53 obimod joined #salt
19:54 obimod hey ya'll!
19:54 LordMike91 Ohai
19:54 LordMike91 Anyone know if file.recurse can remove existing content?
19:54 obimod yup
19:54 bmatt see "clean"
19:54 obimod wait
19:54 LordMike91 I tried adding "clean" but it's apperantly a function
19:54 obimod hmm
19:54 hotbox joined #salt
19:55 bmatt foo:\n  file.recurse:\n    - clean: True
19:55 UtahDave LordMike91: what bmatt said
19:55 obimod n1
19:55 LordMike91 oh
19:55 LordMike91 Prob. forgot ": True"
19:55 LordMike91 Sec.
19:55 bmatt =]
19:55 obimod I'm getting a "Data failed to compile: Rendering SLS "base:gunicorn" failed: Conflicting ID "file"
19:55 obimod here's the state... http://pastebin.com/qQfN5Kkj
19:56 obimod i can't see a conflict!
19:56 LordMike91 Awesum, thx bmatt & obimod
19:56 bmatt np
19:56 obimod hehe, my contribution "hmmm"
19:56 bmatt -1028.5!
19:58 bmatt obimod: I'm willing to bet your pillar data is corrupted and you're building things named "gunicorn-logger--.conf"
19:58 patarr joined #salt
19:58 eykd joined #salt
19:58 bmatt also, I'm not sure python will iterate over a tuple
19:58 bmatt (line 41)
19:59 possibilities joined #salt
19:59 obimod i got that iteration via the nginx formula
19:59 bmatt hm, I guess it will
19:59 obimod hmm
19:59 obimod how can i check the pillar data?
19:59 ahammond joined #salt
19:59 bmatt pillar.items
19:59 bmatt salt-call pillar.items
20:00 hotbox joined #salt
20:01 eykd Is anyone here familiar with the docker states? e.g. http://docs.saltstack.com/en/latest/ref/states/all/salt.states.dockerio.html
20:01 eykd Those should be available on 2014.1.5, right?
20:02 hellome joined #salt
20:02 zach Is it possible to restrict salt-cloud with the ACL function? I'd like to have another user spin up VMs
20:03 obimod bmatt: nope pillar looks good
20:03 bmatt not possible to construct a duplicate ID in your loops?
20:03 obimod nope, only listing everything once
20:04 obimod i wonder if there's a way i can check
20:04 obimod like, a rendered state file
20:05 bmatt hm, i wonder if
20:05 bmatt foo:\n  state:\n    - function\n    - name: foo\n
20:05 bmatt causes a problem
20:05 mway joined #salt
20:06 bmatt I've never tried it - the top level declaration becomes the state name implicitly
20:06 bmatt try cleaning your names out and defining them implicitly?
20:06 racooper howdy. I'm getting an error trying to set up pillars in gitfs: Unable to fetch the latest changes from remote git+ssh://git@github.tamu.edu/racooper/ocss-salt-pillars.git: 'git fetch' returned exit status 128: fatal: 'git@github.tamu.edu/racooper/ocss-salt-pillars.git' does not appear to be a git repository\fatal: The remote end hung up unexpectedly
20:06 obimod i've tried contracting all the - <function> items into their parent... like file.managed
20:06 racooper relevant config and the error are in https://gist.github.com/racooper/373a3a06d2ecaca8c545
20:06 obimod that's a good idea
20:07 hotbox joined #salt
20:07 mway Hey guys, anyone have any idea why salt-minion (installed via yum) wouldn't attempt to connect to the 'salt' hostname and just attempts to connect to a local interface? Tried explicitly setting 'master: salt' in minion/minion.d config as well.
20:08 bmatt can you define 'connect to a local interface'? specifically what you mean by 'interface'?
20:08 mway The minion tries authenticating to the master on the minion's eth0, instead of resolving 'salt'
20:09 LordMike91 could you ping salt from the minion?
20:09 mway Yep. It resolves properly
20:09 joehoyle joined #salt
20:09 LordMike91 What's the ip of your master, minion?
20:10 LordMike91 And how do you know it's connection to 127.0.0.1?
20:10 LordMike91 connecting*
20:10 Guest84025 joined #salt
20:10 ksalman UtahDave: I updated the Windows to 2014.1.5. The master can see talk to the minion for couple of minutes and then i don't see the minion anymore
20:11 jslatts joined #salt
20:11 ksalman i.e. master does not see the minion
20:11 mway Running salt-minion -l debug: [DEBUG   ] Attempting to authenticate with the Salt Master at 172.17.0.5 (minion eth0)
20:11 mway ping salt: PING salt (172.17.42.1) 56(84) bytes of data.
20:11 mgw joined #salt
20:11 obimod oes anyone know how to output a rendered a state file via salt to check if there are any conflicting keys?
20:12 LordMike91 Oh, odd.. mway, you're beyond basic dns so I'm out .. :P
20:12 UtahDave ksalman: try this.  Stop the salt-minion service from the service manager.   Then open up a terminal with elevated privileges  (right-click open as administrator or something like that)
20:12 UtahDave ksalman: then run     cd c:\salt        then      salt-minion -l debug -c c:\salt\conf
20:13 MindDrive joined #salt
20:13 mway FWIW this is in a docker container, where the DNS server is the gateway (the docker host)
20:13 hotbox joined #salt
20:14 Guest84025 joined #salt
20:14 mway Was working on ubuntu precise, I can't see why it would still just arbitrarily pick an IP (eth0) instead of attempting to resolve 'salt'
20:14 vejdmn joined #salt
20:14 UtahDave mway: it should just resolve   salt   to whatever IP  dns says it is
20:14 bmatt mway: what does socket.gethostbyname("salt") return?
20:15 Guest84025 joined #salt
20:15 bmatt in python, on the minion
20:15 mway Ha. eth0's addr
20:15 bmatt I think I've found the problem
20:15 bmatt :)
20:15 mway am I just derping or is this a bug?
20:15 bmatt what's in /etc/hosts?
20:15 bmatt (and what's in DNS? etc)
20:16 bmatt it sounds like it's a host configuration problem, since python agrees with salt
20:16 bmatt (unless it's a python library bug)
20:17 ksalman UtahDave: https://gist.github.com/anonymous/61c84b4d40ee04547c74
20:17 bmatt hm, I have a file managed by a formula, but I need to do other things to that file
20:17 mway I just found the issue also, the host is running dnsmasq and manages itself via a minion itself (has master/minion running), so it has salt defined as well - solution will be to explicitly have the master use a different hostname than 'salt' in its minion config, and then have 'salt' resolve to the gateway addr
20:17 mway thanks a bunch!
20:17 bmatt np
20:18 bmatt what's the best way to do both a file.managed and a file.replace on a single file?
20:18 bmatt hm, two different IDs, I suppose
20:18 mway yeah, and have replace have managed as a requirement
20:18 blarghmatey joined #salt
20:18 mway if you feel like being explicit
20:19 hotbox joined #salt
20:20 obimod How can I output the rendered state file?
20:20 bmatt obimod: I'm not sure there's a way, but salt-util might have something
20:20 bmatt I'd start looking there
20:20 obimod hokay, thx  : )
20:21 eykd Is anyone here familiar with the docker states? e.g. http://docs.saltstack.com/en/latest/ref/states/all/salt.states.dockerio.html Those should be available on 2014.1.5, right? I’m getting the following error: “State docker.running found in sls mystate.docker is unavailable”
20:23 mastrolinux joined #salt
20:23 timoguin obimod: state.show_sls will show the rendered state
20:23 obimod ooo ty
20:24 timoguin state.show_low_sls will give you the lower level rendering as well.
20:24 obimod damn! i cant because of a "Conflicting ID 'file'"
20:24 bmatt neat! thanks timoguin
20:24 obimod i don't see any conflict!
20:24 obimod http://pastebin.com/qQfN5Kkj
20:24 obimod arggg
20:25 timoguin there's also show_highstate and show_lowstate as well
20:25 hotbox joined #salt
20:25 yekta joined #salt
20:26 obimod oo
20:28 seldon419 joined #salt
20:28 blarghmatey joined #salt
20:30 anuvrat joined #salt
20:31 pssblts joined #salt
20:31 hotbox joined #salt
20:32 pssblts joined #salt
20:33 manfred and all your show_sls thingies
20:33 manfred obimod: U
20:34 manfred http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.state.html#salt.modules.state.show_highstate
20:35 bmatt yeah, but obimod can't get his sls to compile
20:35 bmatt so show_* won't work since the compilation step bails
20:35 bmatt he wants a post-preprocessor, pre-compilation output
20:35 manfred ti should spit out the errors if the thing won't compile.
20:35 manfred ¯\(°_o)/¯
20:35 manfred i am off to another meeting
20:36 Guest84025 joined #salt
20:36 manfred forrest: what date is the sprint again?
20:37 forrest manfred, july 10th
20:37 forrest manfred, https://www.eventbrite.com/e/saltstack-documentation-sprint-tickets-12010895913?aff=eorg
20:37 forrest that's the link to the event
20:37 manfred cool
20:37 forrest from 1-5 PDT
20:37 Guest84025 joined #salt
20:37 forrest yar
20:39 Guest84025 joined #salt
20:39 alainv Does grains.setval set the grain permanently on the minion (ideally in /etc/salt/grains)?
20:40 obimod manfred: you've gotten me hooked on bread and butter... it's might be a problem
20:40 manfred alainv: should
20:40 Guest84025 joined #salt
20:40 manfred obimod: they are the best
20:40 obimod it is might
20:40 obimod yes
20:40 alainv manfred: perfect, thanks
20:40 Guest84025 joined #salt
20:40 bhosmer joined #salt
20:41 obimod manfred: if your meeting gets boring, feel free to play "find the conflict for ID 'file'" here: http://pastebin.com/qQfN5Kkj
20:41 obimod i've been at it all morning, it's pretty fun... excitment, let downs, surprises, etc.
20:43 bmatt obimod: without the pillar data, it's not really possible to find the issue
20:43 timoguin yea i'd check the projects pillar first
20:44 pdayton joined #salt
20:45 kedo39 joined #salt
20:47 obimod bmatt: timoguin: http://pastebin.com/2DpY2kum
20:47 obimod pretty simple
20:47 Guest84025 joined #salt
20:47 bmatt what happens when a host that gets that sls file but isn't server1 interprets it?
20:47 obimod OMG
20:47 obimod i think i've found it
20:47 obimod {{  proj_home }} is a recursive statement in projects/init.sls
20:48 obimod my head is bleeding again
20:48 obimod jk, jk
20:49 obimod so one can't have a recursive file.managed and subsequent child file.managed statements?
20:49 obimod one can't recursively manage a directory and have sub-files which are also managed seperately?
20:49 babilen You can only have on file.managed per ID iirc
20:50 obimod but what about ......  {{ folder }}: \n\t file.managed: \n\t recurse: True
20:50 obimod and...
20:50 manfred you can only have one file.x in any stateid
20:50 obimod {{ folder }}/file.txt: \n\t file.managed
20:50 Guest84025 joined #salt
20:50 manfred obimod: you can do a recursive file.managed
20:51 obimod does that file.txt have to be in the folder's source rather than being managed independently?
20:51 manfred hrm?
20:51 MK_FG joined #salt
20:51 obimod i guess it depends on how the folder is recusively managed, like, if it removes anything that isn't in it's source...
20:52 manfred it shouldn't
20:52 manfred well
20:52 manfred i am not certain of that
20:52 obimod this is like an rsync question
20:52 manfred but i am pretty certain there is at least an option to not do that
20:52 obimod rsync --help = AHHHH
20:53 obimod i think that might b where i'm getting the conflict.. i should look at code
20:53 obimod and get more milk
20:53 obimod and take a shower
20:53 peters-tx joined #salt
20:53 obimod and do healthy things in general besides looking at a screen
20:53 manfred obimod: file.recurse might be what you want http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.recurse
20:54 obimod manfred: sweet digs
20:54 obimod : )
20:55 forrest ksalman, any luck?
20:56 ksalman forrest: UtahDave, I think the problem is that i am rewriting minion.conf in the initial deploy and not setting "ipc_mode: tcp"
20:57 forrest ahh
20:57 ksalman since this is my first windows minion, I didn't realize that needed to be done
20:57 dkfjhs joined #salt
20:57 forrest I didn't know that either
20:57 ksalman "Windows platforms lack POSIX IPC and must rely on slower TCP based inter- process communications. Set ipc_mode to tcp on such systems"
20:57 ksalman from http://docs.saltstack.com/en/latest/ref/configuration/minion.html
20:57 forrest so did that work on 2014.1.4?
20:58 ksalman i just found that out, and testing it now =)
20:58 dvogt joined #salt
20:59 Guest84025 joined #salt
21:01 UtahDave ksalman: that would be totally it.
21:01 gothix does anyone else have issues installing packages on rhel 6 for soem reason repoquery cant find anything
21:01 ksalman UtahDave: thanks!
21:01 or1gb1u3 joined #salt
21:02 gothix i wonder if this isse has raised itself again https://github.com/saltstack/salt/issues/12820
21:02 oz_akan_ joined #salt
21:03 smcquay hmm, UtahDave after the upgrade the test.version does indeed peg that little vm's cpu something fierce
21:03 smcquay my guess is that with multiprocessing that getting more cores might help?
21:04 smcquay (although it slows down, the salt command is still running, but no cpu used; I find this odd.)
21:04 Guest84025 joined #salt
21:05 UtahDave smcquay: Yeah, I'd recommend a few more cores, if possible.
21:06 smcquay on it. thanks.
21:07 UtahDave if it's not a big deal, I'd add another gig of ram while you're at it.
21:08 Guest84025 joined #salt
21:09 babilen Is there a list of minimum specs for a master that handles FOONUM minions somewhere for BARTASK? (or anything along those lines?)
21:10 Guest84025 joined #salt
21:11 kermit joined #salt
21:15 UtahDave babilen: Not really. There are many factors involved.  cpus, ram, storage media, etc.  Plus some config options.  You kind of have to test in your environment under your workload
21:15 ksalman must I have leave the default of utc to true when setting timezone, on a windows host? I get the error "Failed to set UTC to True"
21:15 Guest84025 joined #salt
21:16 jhauser joined #salt
21:16 UtahDave ksalman: where are you setting this?
21:17 whidbeywalker joined #salt
21:18 ksalman UtahDave: I just have a timezone/init.sls state with the following https://gist.github.com/anonymous/a78a3e5ff795d7255e0c
21:19 whidbeywalker joined #salt
21:19 Guest84025 joined #salt
21:20 UtahDave ksalman: I believe you have to use a timezone name in this format:  https://github.com/saltstack/salt/blob/develop/salt/modules/win_timezone.py
21:21 ksalman UtahDave: there is a PST8PDT though, like 418
21:21 smcquay UtahDave: alright, upped to 4vCPU, 4G rams. still same behavior :(
21:21 ksalman s/like/line
21:21 babilen UtahDave: It's just that out master(s) sometimes take quite some time and I wonder if upping their specs would be a good idea. They are single core with 1G for around 200-300 minions each and I have no idea if that is severely underpowered or not.
21:21 babilen Hence the: Are there some guidelines somewhere question :)
21:23 Bosch[] Good question. :)
21:23 mastrolinux joined #salt
21:24 shaggy_surfer joined #salt
21:25 KyleG joined #salt
21:25 KyleG joined #salt
21:25 UtahDave smcquay: Hm. can you pastebin the output you're getting?
21:25 Guest84025 joined #salt
21:26 kermit joined #salt
21:27 smcquay sure thing.
21:28 smcquay so output from minion?
21:28 arnoldB any ideas how to get ZMQ 3.2 on wheezy?
21:28 smcquay UtahDave: http://pastebin.com/NmA2ka1v
21:28 UtahDave Well, from the master when you call the command.  Also if you have output from a minion, that would be great too.  Anything that will help troubleshoot
21:28 smcquay sure thing.
21:29 UtahDave smcquay: also if you could give the output of    salt-master --versions-report        and from a minion     salt-call test.versions_report
21:30 possibilities joined #salt
21:30 aw110f do you have to restart the minion service whenever you update the schedule to run highstate in pillar?  I see that the schedule pillar got propagated to the minion and "pillar.item schedule" return the correct new value
21:30 anuvrat joined #salt
21:30 UtahDave aw110f: I think you have to at least run    salt minionid saltutil.refresh_pillar
21:31 aw110f i did as well
21:31 UtahDave hm.  Well, I know that if you modify the schedule using the salt master's config file you have to restart the master.   Let me ask about the minion
21:32 Bosch[] UtahDave: Are there some ballpark figures or cases where system requirements are given for particular setups?
21:32 Bosch[] [performance wise]
21:33 aw110f can you also find out about setting the schedule in the salt minion config, whether is requires a restart of the minion service or not?  I'm trying to find a way to control the schedule without having to restart any services
21:34 UtahDave smcquay: well, one good thing.  Your pastebin led me to fix a misspelling:  https://github.com/saltstack/salt/pull/13692
21:34 smcquay :P
21:34 smcquay glad I could help
21:35 gothix is there anyone else having issues isntalling packages on RHEL 6 ??
21:35 UtahDave gothix: I haven't heard of any, what are you seeing?
21:35 racooper that's vague gothix. are you getting an error?
21:36 smcquay so the machines that aren't yet showing up in output are trying to send information back to the master.
21:37 smcquay and oddly (this may be normaly chatter) I keep getting auth req, auth accepted in the master logs
21:37 UtahDave smcquay: have you made any changes to your master and minion configs?
21:38 gothix racooper, UtahDave http://pastebin.com/YZkbqujM i think this may be similar to a prrevious issue https://github.com/saltstack/salt/issues/12820
21:39 gothix I can log in to the server and yum the packages
21:40 gothix that works but the repoquery fails
21:40 UtahDave gothix: what's the output of   yum search tomcat6    ?
21:41 UtahDave smcquay: is your master's cpu pegged?
21:41 aw110f UtahDave: looks like setting the scheduler directly in the minion config does not require any restart.  Next time highstate runs salt.config rereads my file /etc/salt/minion.d/scheduler.conf
21:42 UtahDave aw110f: ok, good to know.  I'm waiting for someone to get out of a meeting to find out more details
21:43 smcquay only pointed the minions to the appropriate master; checking master (specified file_roots, gitfs_remotes, pillar_roots, pillar_opts=False, debug logging)
21:43 gothix UtahDave, http://pastebin.com/ei88PscP
21:45 dstokes does the reactor system have built-in events like minion start / stop etc?
21:45 dstokes can't find confirmation in the docs
21:45 whiteinge dstokes: look at the event docs instead of the reactor docs
21:46 dstokes whiteinge: nice! thx
21:46 UtahDave smcquay: OK, we think you're hitting an auth flood problem that has been fixed in develop.   You can workaround it by adding these options to your minion config
21:46 whiteinge !events | dstokes
21:46 wm-bot4 dstokes: http://docs.saltstack.com/en/latest/topics/event/index.html#events
21:46 UtahDave smcquay: https://github.com/saltstack/salt/issues/11781#issuecomment-44878982
21:47 UtahDave smcquay: that's an example of a state that will update that for you
21:48 smcquay ooo, thanks. I'll try it out.
21:48 dstokes so there's a `start` event, but no `stop` or `destroyed` event for minions. is the idea to watch `presence/change` to catch disconnected minions?
21:50 TaiSHi I'm using nginx.ng with pillars
21:50 racooper having an issue with gitfs for pillars. https://gist.github.com/racooper/373a3a06d2ecaca8c545 the error is "Unable to fetch the latest changes from remote git+ssh://git@github.tamu.edu/racooper/ocss-salt-pillars.git: 'git fetch' returned exit status 128: fatal: 'git+ssh@github.tamu.edu/racooper/ocss-salt-pillars.git' does not appear to be a git repository"
21:50 TaiSHi I want to define nginx: ng: in one file
21:50 TaiSHi and nginx: vhosts: in another
21:50 TaiSHi Can I do that?
21:50 racooper but the repo is available and the URI is the same (format) as the states tree gitfs, which does work.
21:51 mFacenet joined #salt
21:52 racooper (oh, the error URI and the ext_pillar URI are different because I was testing several varients. the error was the same on all of them, however.)
21:57 bhosmer joined #salt
21:58 wendall911 joined #salt
22:01 gothix UtahDave, so yum finds the package and repoquery doesent so my state fails
22:02 gothix it looks similar to the other issue
22:02 UtahDave gothix: Hm.  could you open an issue on this? I'll have one of our resident yum experts(TM) take a look at it.
22:02 Guest84025 joined #salt
22:02 gothix UtahDave, sure
22:03 UtahDave thanks!
22:04 kermit joined #salt
22:04 MZAWeb joined #salt
22:07 Guest84025 joined #salt
22:08 lude does anyone use pkg.removed on gentoo?
22:08 lude i think there may be a little bug
22:08 lude [INFO    ] Executing state pkg.removed for sys-process/vixie-cron
22:08 lude [INFO    ] Executing command 'emerge --unmerge --quiet --quiet-unmerge-warn --ask nsys-process/vixie-cron' in directory '/root'
22:08 lude seems to be missing a space after --ask n
22:09 aw110f UtahDave: I re-tested, updating the schedule in the minion config, and it also needs a restart of the minion service actually
22:09 aw110f http://pastebin.com/6gQfPtUK
22:09 Guest84025 joined #salt
22:10 UtahDave lude: ah, could you open an issue on that?
22:10 lude UtahDave: yes sir
22:10 UtahDave thanks, a ton!
22:13 lude np, submitted :)
22:13 Guest84025 joined #salt
22:14 oz_akan_ joined #salt
22:14 TaiSHi UtahDave: mind if I bug you with something ?
22:14 UtahDave TaiSHi: what's up?
22:15 TaiSHi I'm working with a series of servers that should include the same base configuration but different vhosts
22:15 TaiSHi As per DRY, I'd like to maintain base config in a single file
22:16 TaiSHi Can I do a pillar call, say, nginx: ng: on one file, and nginx: vhosts: on the other?
22:16 TaiSHi Sorry if I'm unclear, I'm still feverish and at the office
22:22 felskrone joined #salt
22:23 mastrolinux joined #salt
22:24 forrest TaiSHi, ahh the worst kind of employee, 'I am sick, but I gotta go to work' :P
22:24 forrest I'm sure your coworkers will thank you
22:24 bmatt when my coworkers show up sick, I go home
22:24 bmatt I figure *someone* should be at home; I'm fine with that person being me
22:24 forrest lol
22:24 LordMike91 In a file.managed "source: salt:// ....", can I use jinja templating in the source file?
22:25 nickg LordMike91: yes
22:25 forrest LordMike91, yea
22:25 LordMike91 mode: jinja or something?
22:25 forrest you don't have to mark it as a mode for state files
22:25 forrest LordMike91, https://github.com/gravyboat/hungryadmin-sls/blob/master/salt/hungryadmin/app.sls
22:25 forrest I'm not using it for source there
22:25 forrest but it's the same principle
22:26 jalbretsen joined #salt
22:26 LordMike91 You specify "template: jinja" at your source'd file
22:26 LordMike91 But it's default?
22:26 forrest yes you need to do that for a managed file
22:26 LordMike91 Ah
22:26 LordMike91 Awesum
22:26 forrest in a STATE file it automatically parses jinja
22:27 DaveQB joined #salt
22:27 UtahDave TaiSHi: You can use the same pillar key for both.   you use the pillar top.sls to specify which hosts get which pillar sls files applied to them.
22:27 forrest LordMike91, if you want to read more on the renderer, you can do so here: http://docs.saltstack.com/en/latest/ref/renderers/
22:27 bmatt all sls files are automatically processed by jinja; managed files aren't unless the state object gets the "template" argument
22:27 TaiSHi forrest: hey, I -am- sick :P
22:27 TaiSHi of my boss
22:27 LordMike91 That's the part I need, i have a source file (the contents thing), that needs to be templated
22:27 Guest84025 joined #salt
22:27 LordMike91 But col
22:27 LordMike91 cool*
22:27 forrest LordMike91, then just use template: jinja
22:28 TaiSHi UtahDave: but the thing is, same server will match both SLS
22:28 TaiSHi With same key, but different subkey
22:28 TaiSHi (key was the word, thank you)
22:29 UtahDave TaiSHi: can you pastebin what you have so far?  your (sanitized)  /srv/pillar/top.sls  and the relevant pillar sls files?
22:29 forrest LordMike91, let me know if that file and the associated repo doesn't provide enough detail, but I think it should cover most of it
22:29 LordMike91 Another question. It's possible to do targeting, but I can do negative targeting?
22:29 TaiSHi UtahDave: let me do it, no data yet, just starting to test
22:29 LordMike91 Ie.. all that 'don't' have a certain role
22:29 LordMike91 I've tried  '* and not G@roles:postfix':
22:30 LordMike91 Oh wait
22:30 forrest LordMike91, yes, you need to mark that as a compound match though
22:30 LordMike91 Forgot the "match: grain"
22:30 LordMike91 Oh.. compound?
22:30 forrest salt -C  '* and not G@roles:postfix' test.ping
22:30 LordMike91 It'd be wrong anyways :P
22:30 forrest if you're running it off the command line
22:30 dimeshake silly question - can I specify a salt master and minion_id on the command line when starting the minion daemon?
22:31 forrest dimeshake, uhh you know I don't think so, but I'm not sure, I've never tried it
22:31 dimeshake i'm playing with the idea of a bare bones docker instance that I can launch salt-minion on and specify those two for arbitrary purposes
22:32 forrest dimeshake, is it not possible for you to have different masters in the config per instance?
22:32 LordMike91 Behold - Postfix on all servers, using a local SMTP relay!... Thx folks :)
22:32 forrest LordMike91, cool
22:32 dimeshake you sure could, but i essentially want a bare image with nothing but salt installed on it - and the ability to easily pass a different master and minion_id so taht I don't have to shuffle around images should i want to use a different master, etc
22:33 alainv Hey, how does everyone distribute SSL certs securely? I found https://github.com/saltstack/salt/issues/1543, but it doesn't clarify the best method of going about it in the end
22:33 forrest dimeshake, ahh yea, I'm not sure then
22:33 TaiSHi UtahDave: http://pastebin.com/MDkxLJXy
22:33 LordMike91 @dimeshake, I have a templated debian, where I log in as root initially, and call "./setup.sh <newname>"
22:33 dimeshake yeah i could do that, too, but i'm hoping to go as "vanilla" as possible on this :)
22:33 LordMike91 Preseed-like ? :P
22:34 dimeshake my own wrapper scripts may get ugly and unmaintainable and i'd rather use some standard method, if available.
22:34 dimeshake seems it may not be
22:34 Guest84025 joined #salt
22:35 UtahDave TaiSHi: OK, so basically you're going to want to make sure that any key ina pillar sls is not found in another pillar sls file that might both potentially be applied to the same minion
22:36 UtahDave TaiSHi: so I would probably have an ng key and an ng-hosts key, just to make sure there isn't a collision.
22:37 whiteinge smcquay: smoother sailing?
22:38 Ryan_Lane joined #salt
22:38 MatthewsFace joined #salt
22:38 smcquay still waiting for the deploys.
22:39 TaiSHi UtahDave: I'm using nginx-formula, it all goes there
22:39 smcquay it's looking hung
22:39 TaiSHi Yet, I should give it a try
22:40 Hell_Lap joined #salt
22:41 druonysuse joined #salt
22:41 Guest84025 joined #salt
22:42 whiteinge smcquay: are you deploying that fix via salt?
22:42 smcquay yes :P
22:42 smcquay I guess I could've thought that one through a few times.
22:43 smcquay I thought about using fabric. alas. So let's say I want to "clear things out"; the scheduler still seems to know about "jobs" that are in flight. If I want to clear that out I stop master/minions and clear cache directories?
22:43 UtahDave smcquay: yeah.
22:44 dvogt joined #salt
22:44 UtahDave smcquay: I've also used salt-ssh to help manage minions in the situation.  I'm sorry you've hit this bug. Usually managing 400+ minions is a breeze
22:45 smcquay I figured so much. I remember Tom speaking of 1000s in the demos he'd run in UT
22:45 smcquay and honestly, when it works it's startlingly fast. very refreshing.
22:46 UtahDave Yeah, seriously.  I was down in CA with a customer a month or so ago when this bug popped up. Once we fixed it with those changes we gave you 3,400 minions was nothing.
22:48 Ch3LL_ joined #salt
22:49 Guest84025 joined #salt
22:49 Luke_ joined #salt
22:52 Ch3LL_ is this the best resource for current state files. https://github.com/saltstack/salt-states
22:52 TaiSHi Rendering SLS "base:nginx.ng.config" failed: Jinja error: get() got an unexpected keyword argument 'merge'
22:52 yomilk joined #salt
22:52 TaiSHi Hmm
22:53 joehoyle joined #salt
22:55 TaiSHi Helium release != 2014.1.x?
22:55 eliasp hmm, just getting here a 'image "pctue/mysqld" does not exist' for a docker.installed ( image: pctue/mysqld ) state… "docker images | grep pctue/mysqld" clearly shows the image
22:55 eliasp running 2014.5 on Ubuntu 13.10 … any idea what could be wrong?
22:55 eliasp TaiSHi: no, Helium will be the one after 2014.1
22:55 eliasp probably 2014.6, I'm not sure about the exact numbering scheme
22:56 TaiSHi Damn, can't use ng
22:56 TaiSHi Configured everything for it :(
23:02 mosen joined #salt
23:02 possibilities joined #salt
23:02 smcquay hmm, UtahDave: this is odd; it still takes forever for test.ping to run. still getting through about 371/384 machines and hanging
23:05 seventy3 joined #salt
23:06 TaiSHi Damn, no dev version for ubuntu
23:06 yomilk joined #salt
23:06 manfred eliasp: looking more and more like 2014.7
23:06 vejdmn joined #salt
23:07 eliasp manfred: as long as it's worth the result, it could also be a 2014.12 :)
23:07 manfred heh
23:07 manfred 2014.13
23:08 manfred worth the effort to make a new month
23:08 eliasp TaiSHi: if you need specific modules or states in older releases, just deploy them locally using: http://docs.saltstack.com/en/latest/ref/file_server/dynamic-modules.html
23:08 TaiSHi eliasp: exactly what I was doing <3
23:08 eliasp ok
23:09 TaiSHi Thanks, I actually spent a little bit thinking on that
23:09 TaiSHi God I want to go home!!!
23:09 TaiSHi My boss is still here (8 pm)
23:09 eliasp meh… I'm giving up for today (1:08AM)… Docker doesn't like me today
23:10 eliasp TaiSHi: 8PM? I'd be happy about leaving that early ;)
23:10 eliasp have a good night… see you tomorrow/later
23:10 TaiSHi Sleep well
23:11 TaiSHi I leave at 00
23:11 TaiSHi I should*
23:11 whiteinge smcquay: the CLI will wait for unresponsive minions if they are expected to return. that's a UI choice (i.e., the master workers aren't hanging).
23:11 whiteinge smcquay: any chance the salt-minion daemon on those problem minions wasn't fully restarted?
23:11 smcquay yeah, I am prodding around.
23:12 whiteinge how's the master performance otherwise? CPU, memory, disk IO?
23:12 smcquay it seems as though my deploy of the fix was incomplete
23:12 smcquay oh, just fine.
23:13 Guest84025 joined #salt
23:15 obimod manfred: 12 months getting to short for you ? : )
23:15 TaiSHi Any reason why a module isn't syncing ?
23:15 TaiSHi No error, no output at all
23:15 manfred obimod: i got 99 problems and 12 months isn't enough to fix them all
23:16 manfred TaiSHi: did you do saltutils.sync_modules ?
23:16 TaiSHi Yeah
23:16 obimod manfred: 99? lucky..
23:16 manfred saltutil*
23:16 TaiSHi No output at all
23:16 manfred obimod: heh
23:16 TaiSHi I mean, it outputs modules:
23:16 manfred TaiSHi: and the module is in /srv/salt/_modules?
23:16 manfred TaiSHi: it should output the module it is syncing too
23:16 TaiSHi ls /srv/salt/_modules/ -> pillar.py
23:16 TaiSHi Yeah, I did work with modules before
23:17 Guest84025 joined #salt
23:17 manfred hrm...
23:17 TaiSHi Hmm
23:17 TaiSHi Perhaps it's the name ?
23:17 manfred it shouldn't be
23:17 manfred TaiSHi: it might be, but i don't know if it should be
23:17 TaiSHi I just renamed it
23:17 TaiSHi to pillar_dev
23:17 manfred TaiSHi: if you cahnge the name, but pass a __virtualname__ and return the __virtualname__ in the __virtual__() does it work?
23:18 smcquay whiteinge: I finally got the config rolled out appropriately and the master is definitely churning. i am guessing there is a backlog it's trying to get through.
23:18 manfred it shouldn't need to be renamed, because i know that i have had it work with the copied name before
23:18 TaiSHi And it synced
23:18 manfred TaiSHi: what if you change it back to pillar.py?
23:18 dvogt joined #salt
23:18 TaiSHi I mean, I renamed it and worked, but still loading old function
23:18 TaiSHi Well fuck me, it synced.
23:18 manfred lol
23:19 TaiSHi I submitted an issue for next release
23:19 manfred TaiSHi: if it happens again, i would be interested if you ran salt-call saltutil.sync_modules -l debug on the minion
23:20 whiteinge smcquay: anything suspicious in the master log?
23:20 smcquay nothing that I can tell.
23:20 TaiSHi manfred: I can try moving it out & in
23:20 * smcquay still looks
23:20 seventy3 joined #salt
23:20 ipalreadytaken joined #salt
23:20 whiteinge smcquay: you can watch the event bus to see what exactly it's churning on
23:20 TaiSHi manfred: mind if I PM you ?
23:21 smcquay I'm still a bit green; what is the event bus?
23:21 manfred TaiSHi: nope, go ahead
23:22 whiteinge smcquay: download and run this: https://github.com/saltstack/salt/blob/develop/tests/eventlisten.py
23:23 whiteinge smcquay: you'll see events for minion authentications, new jobs, job returns, etc
23:23 smcquay cool, cool.
23:23 smcquay it's pretty quiet right now.
23:24 whiteinge oh? but you're seeing churn still? is that CPU usage or what?
23:25 alainv hey, why would i (recently) be seeing a lot of spam in highstate with [INFO    ] Executing command 'iptables --help' in directory '/root'
23:25 smcquay actually cpu is down now.
23:26 manfred alainv: it is fixed in develop
23:26 alainv poop. ok
23:26 alainv thanks manfred
23:26 manfred alainv: it is also backported to 2014.1 branch, and will be fixed in .6, which is hopefully being tagged this week
23:26 renothing joined #salt
23:26 alainv any relatively simple way to suppress that in the meantime?
23:26 smcquay just posting these brief updates is likely unhelpful; I wonder if it's better for me to reset state across the cluster to diagnose things. It seems like there is still old stuff trying to be run.
23:26 manfred alainv: grab the iptables.py module from the 2014.1 branch
23:27 manfred alainv: curl -o /srv/salt/_modules/iptables.py https://raw.githubusercontent.com/saltstack/salt/2014.1/salt/modules/iptables.py
23:27 alainv manfred: Awesome! Thank you.
23:27 manfred then salt \* saltutil.sync_modules
23:27 forrest digitalocean api 2.0 is in beta
23:27 manfred alainv: fun story
23:27 manfred alainv: that was actually a bug caused by me :/
23:28 alainv :o
23:28 smcquay whiteinge: the odd thing is I see nothing coming across the bus. very odd.
23:28 alainv Well, these things happen + you helped me fix it personally, so definitely no hard feelings here ;)
23:28 whiteinge smcquay: the master won't queue up jobs to be run, say, after a daemon restart. since you've restarted both master and minion i wouldn't expect any jobs to be running (unless you've configured minions to run things on startup)
23:28 manfred :)
23:29 smcquay whiteinge: yeah, I haven't.
23:29 whiteinge smcquay: what are you seeing / trying right now?
23:30 TheThing joined #salt
23:31 Hell_Lap joined #salt
23:32 smcquay still pings. I can do about 150 at a time, but then past that it freezes. Crazy though, if I c-c, then salt-run jobs.lookup_jid 20140624163058592315 it seems to go, but skips a few nodes.
23:34 smcquay well, I will try to get a more complete failure story together for you tomorrow. my ride leaves in 1 min :/ ttyl. thanks for the help, UtahDave and whiteinge
23:35 whiteinge ttyl
23:35 TaiSHi Is there any way to install -dev in ubuntu easily ?
23:35 TaiSHi (and bootstrap it to minions)
23:35 UtahDave the bootstrap script allows you to specify the git branch/tag/rev you want.  are you using salt-cloud?
23:36 whitepaws hi all .. how do you escape the colon in a cmd.run argument? for example
23:36 whitepaws salt -v host cmd.run 'echo "this is a thing\: thing"'
23:36 whitepaws i have used template=jinja but that seems like a cheat
23:36 TaiSHi UtahDave: I am
23:37 UtahDave cool, let me pastebin what to put in your profile.  just a sec
23:38 TaiSHi UtahDave: and what should I do on the master? I mean, how do I update it ?
23:38 UtahDave TaiSHi: http://pastebin.com/9XUc74CU
23:38 UtahDave TaiSHi: the easiest way to update your master is with packages.
23:38 dstokes guise, i confuse. installing develop from bootstrap gets me 2014.1.0, but installing stable gets me 2014.1.5?
23:39 Guest84025 joined #salt
23:39 UtahDave dstokes: what's the output of   salt <minin id> test.version
23:39 dstokes testing install on a master / minion setup. not sure how to target 'self'
23:39 UtahDave TaiSHi: otherwise remove the salt package, clone the git repo, checkout the branch or tag you want, then run   sudo python setup.py install --force
23:39 TaiSHi UtahDave: but wouldn't the master be falling behind? I have it installed with ppa
23:40 mgw joined #salt
23:40 dstokes both salt & salt call report 2014.1.0-3338-g572cb60 tho
23:41 dstokes just confused as to why unstable is behind stable (re: versions)
23:41 UtahDave dstokes: so that means you are on the develop branch. The last branch is 2014.1.0, you are 3338 git commits since that branch and g572cb60 is your current git revision
23:41 UtahDave unstable isn't behind stable.  you're 3338 commits past stable
23:42 ajprog_laptop joined #salt
23:42 dstokes ok, got it. i always assumed releases were date based (from the format) is there semver in there somewhere? 2014 major rev?
23:43 dstokes (idea of date based releases is where my confusion comes from i think)
23:43 UtahDave dstokes: Yeah, it isn't the clearest.
23:44 dstokes like, do we know what the target helium release will be? (assuming it's _not_ 2014.1.5)
23:44 manfred dstokes: 2014.1 is january 2014, helium release will be this month or next, so 2014.6 or 2014.7
23:45 schimmy joined #salt
23:46 dstokes k, thx guys (utahdave, manfred)
23:47 schimmy1 joined #salt
23:48 mastrolinux joined #salt
23:50 otter768 joined #salt
23:52 MoonSweep joined #salt
23:52 TaiSHi Anyone running into issues with salt-cloud returning [ERROR   ] Authentication failed: status code 255 ?
23:52 chiui joined #salt
23:53 TaiSHi Happened after I added script_args: '-- git develop'
23:53 UtahDave can you pastebin your config where you added script_args?
23:54 jonbrefe joined #salt
23:54 TaiSHi http://dpaste.com/0ZDQ4ZR
23:54 manfred is it ubuntu?
23:54 manfred err
23:54 manfred not ubuntu
23:54 manfred digital ocean?
23:55 TaiSHi Yes
23:55 TaiSHi It was working before
23:55 manfred TaiSHi: where do you specify your ssh_key_file?
23:55 manfred digital ocean doesn't pass the password back through the api anymore (if they ever did at all) so you need to specify an ssh key to use
23:55 TaiSHi I do, throug cloud.providers.d
23:56 manfred you specify the private key in ssh_key_file? and the ssh_key_name should be the key name in their system for them to put on the server.
23:56 anuvrat joined #salt
23:56 TaiSHi Yes, it was working perfectly before
23:56 TaiSHi The instance WAS created actually
23:56 forrest hmm, does anyone know how Ryan set it up so he had a link for his name at http://bots.wmflabs.org/~wm-bot/dump/%23salt.htm ?
23:57 TaiSHi But it didn't bootstrap salt
23:57 MoonSweep Hi, I'd like to report a problem. The Release file provided by the debian repository for up-to-date Salt packages (debian.saltstack.com) doesn't provide an "Archive" field. APT accepts this but python-apt (on which some tools, like unattended-upgrades, are based) don't. This problem is solved in python-apt for the next release of debian (jessie) but not for the current release (wheezy), which means that until the Release file on debian.saltstack.com is
23:57 manfred forrest: i bet it is in the freenode settings
23:58 forrest manfred, could be.
23:58 manfred TaiSHi: if you build again, does the same error happen? I was having some race conditions in salt-cloud with the bootstrap on other providers
23:58 TaiSHi Well, let me check
23:58 Luke_ joined #salt
23:59 TaiSHi http://dpaste.com/1DE1XMM
23:59 TaiSHi This was the output on the first one
23:59 manfred forrest: i am a chatty person apparently
23:59 forrest manfred, the chattiest
23:59 MoonSweep Should I send a mail directly to Joe Healy or is there a proper way to report this bug ?
23:59 UtahDave MoonSweep: joehh is our debian maintainer.  If he's not online right now, if you send an email to the salt-users mailing list I'll make sure he sees it

| Channels | #salt index | Today | | Search | Google Search | Plain-Text | summary