Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-04-03

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

All times shown according to UTC.

Time Nick Message
00:14 subsigna_ joined #salt
00:15 dude051 joined #salt
00:17 baweaver joined #salt
00:19 dendazen joined #salt
00:35 scbunn joined #salt
00:37 jerrcs joined #salt
00:37 Pixionus joined #salt
00:38 subsigna_ joined #salt
00:40 markm joined #salt
00:44 dalexander joined #salt
00:44 JDiPierro joined #salt
00:45 sk_0 i keep getting conflicting id errors in my init.sls. i comment out the offending block and the next block gets the error. i do a grep -r <id> /srv/salt and there is only one instance of that id
00:47 subsignal joined #salt
00:47 blast_hardcheese joined #salt
00:58 debian112 joined #salt
00:59 wvds-nl joined #salt
01:01 MatthewsFace joined #salt
01:02 rickflare joined #salt
01:02 mariusv joined #salt
01:02 mariusv joined #salt
01:02 Vye salt 2014.7 fork bomb: "import salt.client; c = salt.client.Caller()" inside a module.
01:02 CryptoMer joined #salt
01:02 wolog joined #salt
01:03 Corey joined #salt
01:04 murrdoc evil
01:04 rome_390 joined #salt
01:05 Vynce joined #salt
01:06 nich0s joined #salt
01:06 eightyeight joined #salt
01:06 manfred joined #salt
01:06 Vye Maybe lazy loading in 2015.2 won't allow that. Believe it or not, someone broke our dev environment with one of those.
01:07 Vye o_O
01:07 rocket joined #salt
01:08 hardwire joined #salt
01:08 murrdoc should put that on github
01:08 murrdoc maybe saltstack wil make a 'dont do this' page
01:09 aqua^mac joined #salt
01:09 teskew joined #salt
01:10 Vye think so? I thought it was a silly mistake. Why would you need to initialize a client within module...?
01:11 otter768 joined #salt
01:11 _JZ_ joined #salt
01:16 Vynce remember that every single one of the users is a newb at some point, and they're often the ones reading the docs.
01:17 Vynce speaking of which — is it permissble to break a jinja tag over multiple lines?  eg. {% if (long_cond_one) \n and (long_cond_two) %}  ?
01:25 I3olle joined #salt
01:32 MindDrive Wow, it looks like salt.client.Caller.function() is completely, utterly broken in 2014.7.1.  Nice!
01:33 nahamu Is that an actual fork bomb or just an infinite loop?
01:41 echo joined #salt
01:48 ilbot3 joined #salt
01:48 Topic for #salt is now Welcome to #salt | 2014.7.2 is the latest | Please use https://gist.github.com for code, don't paste directly into the channel | Please be patient when asking questions as we are volunteers and may not have immediate answers | Channel logs are available at http://irclog.perlgeek.de/salt/
01:52 ajw0100 joined #salt
01:52 cheus joined #salt
01:54 subsignal joined #salt
01:55 ajw0100_ joined #salt
01:57 rocket joined #salt
01:57 rocket_ joined #salt
02:03 Vye Vynce: True
02:04 Vye nahamu: Didn't check whether it was forking processes, just assumed it was. It is at least an infinite loop.
02:08 markm joined #salt
02:08 Furao joined #salt
02:21 JDiPierro joined #salt
02:25 ajw0100 joined #salt
02:25 malinoff joined #salt
02:28 evle joined #salt
02:31 michelangelo joined #salt
02:32 jerematic joined #salt
02:37 otter768 joined #salt
02:42 zwi joined #salt
02:44 dendazen joined #salt
02:55 fusionx86 joined #salt
02:55 favadi joined #salt
02:58 CryptoMer Does anyone know if Jinja has a conditional 'unless'?
02:59 CryptoMer I've got a state I want to apply all types of my servers, with the exception of one...Not having any luck finding a great way of doing it without that unless.
03:01 jeffspeff so,i've moved my master to another system. i copied all the configs, etc. on to a new server, changed the bind address in the master conf file. i use autosign to automatically accept minion keys. i copied over the previous master keys to use on the new setup. at first all the minions were working fine, now the minions are showing the error that authentication failed. i tried deleting tall of keys, but that hasn't helped. any ideas?
03:02 CryptoMer turn up logging and identify which host you're actually connecting to. That'd be my first step.
03:02 dendazen joined #salt
03:09 murrdoc joined #salt
03:13 jeffspeff the minions all register to a public IP address. I've changed the NAT setting to point the public IP to the new private IP.
03:14 babilen CryptoMer: Why do you target that box in the first place?
03:14 timoguin joined #salt
03:15 Vynce CryptoMer: if not ?
03:17 Vynce (by which i mean, where you would type "unless" type "if not")
03:17 CryptoMer Oh, there's an 'if not' conditional?! perfect!
03:19 CryptoMer babilen: I've got the same services in multiple environments. The specific state in question installs a 3rd party application, for which I have to pay per host. I want the states to be functional in all environments, I just don't want to pay for the 3rd party app in my dev env. :)
03:22 babilen So, target that one specifically?
03:23 babilen Or, put it in a {% if grains['host'] != "foo" %} block or whatnot
03:24 babilen I mean why don't you simply have my_super_expensive_application.sls and target exactly the boxes you want?
03:25 babilen Or make the dev environment explicit. There are multiple ways to do that (naming scheme, custom grains, lists in pillars/external pillars) and then exclude grains['dev'] hosts
03:26 babilen http://jinja.pocoo.org/docs/dev/templates/ is probably a good read at this point
03:27 catpig joined #salt
03:31 germs_1 joined #salt
03:31 beneggett joined #salt
03:32 CryptoMer I'll check it out, thanks :D
03:39 yohnnyjoe joined #salt
03:42 markm joined #salt
03:57 writtenoff joined #salt
04:01 bhosmer joined #salt
04:10 mdln joined #salt
04:11 AlexStraunoff joined #salt
04:13 jasonrm joined #salt
04:14 SheetiS joined #salt
04:18 jasonrm joined #salt
04:25 hasue1 left #salt
04:36 Vye Anyone knows if Thomas' saltconf talk on raet was recorded and posted?
04:46 dude051 joined #salt
04:46 germs_ joined #salt
04:47 dopesong joined #salt
04:48 zipkid joined #salt
04:48 manytrees joined #salt
04:49 mikepea_ joined #salt
04:49 joeyparsons joined #salt
04:49 Zachary_DuBois joined #salt
04:50 antonw joined #salt
04:50 munhitsu joined #salt
04:50 whiteinge joined #salt
04:50 kwmiebach joined #salt
04:50 doriftoshoes joined #salt
04:50 markm joined #salt
04:53 abele joined #salt
04:53 supermike_ joined #salt
04:53 imanc_ joined #salt
05:00 mdln joined #salt
05:12 joshin The fun part about coming up the salt learning curve is that every time I get stuck, I ultimately find a better way to do things than I first intended.
05:19 beneggett joined #salt
05:22 hemebond joined #salt
05:23 flebel joined #salt
05:23 stoogenmeyer__ joined #salt
05:31 Norbell_ joined #salt
05:31 skullone hrm nice my openstack environment is now automated into AD, puppet/salt, DNS
05:33 catpigger joined #salt
05:34 dopesong_ joined #salt
05:36 mdln joined #salt
05:39 yohnnyjoe joined #salt
05:48 pravka joined #salt
05:53 mdln joined #salt
05:54 otter768 joined #salt
05:54 dopesong joined #salt
05:57 dopesong joined #salt
06:02 mosen joined #salt
06:04 flebel_ joined #salt
06:06 stoogenmeyer__ joined #salt
06:11 badon joined #salt
06:14 beneggett joined #salt
06:16 flebel joined #salt
06:19 subsignal joined #salt
06:19 JlRd joined #salt
06:20 badon joined #salt
06:22 Furao http://www.rce-cast.com/Podcast/rce-96-salt-stack.html
06:27 flyboy joined #salt
06:34 beneggett joined #salt
06:35 dopesong joined #salt
06:41 stoogenmeyer__ joined #salt
06:41 retr0h joined #salt
06:49 subsignal joined #salt
06:54 jalbretsen joined #salt
06:56 trikke joined #salt
07:00 otter768 joined #salt
07:02 eseyman joined #salt
07:06 _ether_ joined #salt
07:07 pravka joined #salt
07:21 Auroch joined #salt
07:25 pravka joined #salt
07:50 subsignal joined #salt
07:53 dRiN joined #salt
07:55 beneggett joined #salt
07:57 MatthewsFace joined #salt
07:59 o5k joined #salt
08:07 gwmngilfen joined #salt
08:08 __gotcha joined #salt
08:10 marnom joined #salt
08:11 marnom Hi guys, having an issue setting an iptables rule for protocol 'vrrp' - Salt keeps appending '-m vrrp' to the iptables line, which results in an error... If I remove this manually, it works. Any ideas?
08:12 TyrfingMjolnir joined #salt
08:15 flyboy82 in the state you have for adding the line, put a only_if or not_if argument checking if the rule is present
08:15 kawa2014 joined #salt
08:15 flyboy82 or was it "unless"?
08:16 marnom flyboy82, was that directed at me? If so, shouldn't Salt check that itself? My other iptable states don't have any manual onlyif/unless constructions...
08:16 marnom flyboy82, It seems Salt thinks it needs to add '-m vrrp' as well, but that is not a module/match...
08:18 ckao joined #salt
08:20 _ether_ joined #salt
08:21 flyboy82 mornom: check http://linux.die.net/man/8/iptables  the "Match Extensions" section
08:21 flyboy82 marnom*
08:25 marnom flyboy82, thanks I will take a look!
08:26 linjan joined #salt
08:28 dewdrop joined #salt
08:46 N-Mi_ joined #salt
08:46 N-Mi_ joined #salt
08:48 glyf joined #salt
08:49 KermitTheFragger joined #salt
08:51 subsignal joined #salt
08:52 teddyalf joined #salt
09:01 otter768 joined #salt
09:06 favadi joined #salt
09:17 gdm85 joined #salt
09:18 gdm85 hi there. is there a way in YAML/jinja to wrap definition lines, like one does in shell scripts with '\'?
09:19 markm joined #salt
09:23 wicope joined #salt
09:24 wicope joined #salt
09:37 dopesong_ joined #salt
09:38 losh joined #salt
09:38 gfa joined #salt
09:43 evle joined #salt
09:48 echo joined #salt
09:48 losh hello, does rallytime frequent this channel?
09:51 subsignal joined #salt
09:51 babilen I don;t think so, why?
09:57 _1_chopper3 joined #salt
09:59 __gotcha joined #salt
10:01 agend joined #salt
10:04 gdm85 pillars should work also in master-less mode, right?
10:05 toddnni joined #salt
10:15 ethX joined #salt
10:17 toddnni_ joined #salt
10:20 bluenemo joined #salt
10:33 giantlock joined #salt
10:34 badon_ joined #salt
10:45 andreyst joined #salt
10:48 aquassaut joined #salt
10:52 subsignal joined #salt
11:02 otter768 joined #salt
11:06 andreyst joined #salt
11:06 toddnni_ joined #salt
11:18 hemebond left #salt
11:18 evle1 joined #salt
11:21 peters-tx joined #salt
11:21 amcorreia joined #salt
11:24 amcorreia_ joined #salt
11:25 toddnni joined #salt
11:28 __gotcha joined #salt
11:34 MatthewsFace joined #salt
11:40 denys joined #salt
11:41 andreyst joined #salt
11:48 faust joined #salt
11:53 subsignal joined #salt
11:55 andreyst1 joined #salt
11:56 wado joined #salt
12:04 antani joined #salt
12:04 __gotcha joined #salt
12:04 bhosmer joined #salt
12:07 pdayton joined #salt
12:11 dopesong joined #salt
12:12 KennethWilke joined #salt
12:16 jerematic joined #salt
12:22 dopesong_ joined #salt
12:27 vstoniest joined #salt
12:30 cberndt joined #salt
12:30 pviktori joined #salt
12:31 cmcmacken joined #salt
12:34 emaninpa joined #salt
12:37 TyrfingMjolnir joined #salt
12:38 murrdoc joined #salt
12:40 edrocks joined #salt
12:40 dendazen joined #salt
12:49 agend joined #salt
12:54 toastedpenguin joined #salt
12:55 JDiPierro joined #salt
12:57 zwi joined #salt
13:00 jeremyr joined #salt
13:03 murrdoc joined #salt
13:04 niseak joined #salt
13:05 murrdoc joined #salt
13:08 amcorreia_ joined #salt
13:09 cpowell joined #salt
13:10 cpowell joined #salt
13:13 elfixit joined #salt
13:13 racooper joined #salt
13:15 slav0nic joined #salt
13:15 chandankumar joined #salt
13:20 markm joined #salt
13:23 sunkist joined #salt
13:25 favadi joined #salt
13:26 favadi left #salt
13:26 Tecnico1931 joined #salt
13:29 subsignal joined #salt
13:32 giantlock joined #salt
13:32 mpanetta joined #salt
13:35 jdesilet joined #salt
13:38 faust left #salt
13:39 sunkist joined #salt
13:43 mpanetta joined #salt
13:47 timoguin joined #salt
13:50 debian112 joined #salt
13:55 gfa joined #salt
13:55 __aphor__ joined #salt
13:56 learning joined #salt
13:56 JDiPierro joined #salt
13:57 perfectsine joined #salt
13:57 fusionx86 joined #salt
14:00 CeBe joined #salt
14:00 faust joined #salt
14:01 faust left #salt
14:01 learning salt 'develop' branch seems to be broken at the moment - at least for some things. command "salt-run fileserver.update" causes an invalid syntax error for example
14:01 badon_ joined #salt
14:02 sunkist1 joined #salt
14:04 I3olle_ joined #salt
14:08 badon_ joined #salt
14:10 andrew_v joined #salt
14:13 murrdoc learning:  the salt devs would appreciate if you doc the bug and put it in a github
14:14 learning murrdoc: will do
14:16 murrdoc normally all merges to develop go through a ton of tests in jenkins
14:17 murrdoc so its surprising that a syntax error gets through
14:20 markm joined #salt
14:20 teskew joined #salt
14:22 dude051 joined #salt
14:23 ek6 joined #salt
14:23 masterkorp joined #salt
14:25 gdm85 would it be possible to make a state dependent on another? e.g. if state X fails then don't even bother trying to apply Y
14:25 mpanetta gdm85: That is what requires is for
14:26 gdm85 mpanetta: require with 'sls' type, nice! thanks for the pointer
14:27 zwi joined #salt
14:27 mpanetta gdm85: Yw :)
14:27 jalbretsen joined #salt
14:27 gdm85 actually when I started using salt I was expecting that the previous state is always a requirement of the next. but as far as I understand there's no way to enforce it without requires?
14:28 mpanetta I believe that is correct
14:29 mpanetta They just run in file order if you don't use requires or watch
14:29 dimeshake you can also set a hard-fail mode, where it'll stop as soon as any state fails
14:29 mpanetta I always forget about that one
14:35 ek6 joined #salt
14:35 gdm85 dimeshake: which one is it?
14:36 gdm85 not --hard-crash, I believe
14:37 dimeshake http://docs.saltstack.com/en/latest/ref/states/failhard.html
14:39 TyrfingMjolnir joined #salt
14:39 viq joined #salt
14:40 badon joined #salt
14:40 gdm85 dimeshake: thanks!
14:44 AlexStraunoff joined #salt
14:46 QuintinAdam joined #salt
14:48 beneggett joined #salt
14:50 pravka joined #salt
14:50 badon_ joined #salt
14:56 TyrfingMjolnir joined #salt
15:02 sandah joined #salt
15:02 TyrfingMjolnir joined #salt
15:02 dude051 joined #salt
15:03 __gotcha joined #salt
15:04 otter768 joined #salt
15:05 beneggett joined #salt
15:07 badon_ joined #salt
15:07 thedodd joined #salt
15:08 murrdoc joined #salt
15:10 scbunn joined #salt
15:11 mohae joined #salt
15:12 mohae joined #salt
15:12 StDiluted joined #salt
15:16 chandankumar joined #salt
15:17 iggy Vye: I don't think he did a talk specicially about raet. His talk on advanced salt topics (part of the paid training) did cover raet a lot, and I don't know if that was recorded
15:19 iggy gdm85: there is failhard
15:19 hobakill joined #salt
15:19 murrdoc there was the talk about raet
15:19 murrdoc went over libnacl and other such things
15:20 iggy was there? I must have blocked it out since I did the training
15:20 murrdoc it was at 2 or something on the last day, so its possible u blacked out
15:20 iggy that is possible too
15:21 eseyman joined #salt
15:21 gdm85 iggy: yes I am using that now. helpful as I just started with salt
15:22 yawniek joined #salt
15:22 iggy it's also helpful if you have more minions x states than will fit into a scrollback
15:24 v0rtex iggy: the talk he did about raet on the last day was a sort of super condensed overview of some of what he did during the training day
15:24 dude051 joined #salt
15:25 iggy ahhh
15:25 murrdoc still a great talk, dude is ridiculously smart
15:25 v0rtex unfortunately I don't see his training day session listed on the tech track playlist though so I'm not sure if it was recorded
15:27 v0rtex murrdoc: agreed...if I didn't love my current job so much I could see myself trying to get hired with them so I could work with Thomas :)
15:27 murrdoc excet u d have to work in slc
15:27 murrdoc which isnt all bad
15:28 clintberry joined #salt
15:28 UberUser joined #salt
15:28 v0rtex I like working with really smart people (unfortunately not a common occurrence at the jobs I've worked at)
15:28 UberUser hey,
15:28 murrdoc i dont even know how it would be working in slc
15:28 v0rtex murrdoc: I already do :)
15:28 murrdoc what else is in slc
15:28 murrdoc like what company do u work at
15:28 UberUser I was wondering if there is a way with salt to pull instead of push. The minion needs pull from the master. Something like: http://docs.ansible.com/playbooks_intro.html#ansible-pull
15:29 v0rtex murrdoc: I actually work in downtown SLC for Deseret Digital Media
15:29 iggy I saw them
15:29 iggy my hotel was down the road
15:29 v0rtex we do a bunch of media sites that are local and national
15:29 murrdoc hows life in slc
15:30 murrdoc UberUser:  no clue, you could do standalone minions with gtifs
15:30 \ask joined #salt
15:30 v0rtex I like it - don't have much to compare with because I've always lived in Utah
15:31 iggy UberUser: that's not enough of a description to go off of, but there's masterless and/or salt-call
15:31 iggy amongst other things
15:31 iggy your best bet is to explain what you're trying to achieve, not how you're trying to achieve it
15:32 UberUser alright.
15:32 v0rtex murrdoc: I actually live about 45 minutes southwest of SLC proper so I don't spend a ton of time downtown aside from work
15:32 UberUser I have application servers running. I want to put their configuration management onto git.
15:32 dude051 joined #salt
15:33 UberUser It is not possible for the configurations to be changed while the applications servers are running, this needs to happen on a restart
15:33 UberUser therefore I would like to pull instead of push.
15:33 chandankumar joined #salt
15:33 bluenemo_ joined #salt
15:33 iggy have startup_state: highstate in the minion config and just don't ever use "salt '*' ..."
15:33 iggy problem solved
15:34 murrdoc or
15:34 murrdoc no what iggy said
15:34 mpanetta UberUser: Um, why don't you just watch the config file and when it changes restart the app?
15:34 mpanetta No pull needed
15:34 UberUser The app scheduled to only restart and specific times.
15:34 mpanetta Unless you are changing it outside of salt
15:34 UberUser *at
15:35 UberUser the app restart also overwrites all configs. :(
15:35 mpanetta Then why not stage the change (put the changes in place) and then the app will pick them up on next scheduled restart?
15:35 mpanetta Oh
15:35 mpanetta SHitty app...
15:35 UberUser yeah :D
15:36 UberUser I could pull into an exchange directory and let the app catch it once it restarts.
15:36 UberUser but thats a little dirty.
15:36 mpanetta Not any more dirty then an app that overwrites its own config on restart :P
15:36 UberUser *let the master psuh
15:36 UberUser true that.
15:37 murrdoc or you could look into salt reactor
15:37 murrdoc let the app notify the master that its stopped
15:37 mpanetta Hmm
15:37 mpanetta That would work
15:37 murrdoc and the master can verify that the app server isnt lying
15:37 murrdoc do the necessary work
15:37 murrdoc and on the way out start the app
15:37 drz2k joined #salt
15:37 murrdoc then disable salt-call on the minion
15:37 murrdoc and ssh on the master
15:38 denys joined #salt
15:38 UberUser ahh ok. so: app -> notify master -> master do work -> master verify -> master notify minion to start app again ?
15:38 murrdoc seems the least involved
15:38 murrdoc lock humans out of the salt master so you dont have accidental salt runs
15:39 mpanetta That is scary :P
15:39 murrdoc its not actually
15:39 murrdoc think of an ssh free world
15:39 murrdoc where machines decide stuff on metrics
15:39 murrdoc and not gut feelings
15:39 mpanetta How do you get back in if something fails and SSH never gets re-enabled?
15:40 murrdoc i mean ssh is enabled for power users
15:40 murrdoc not actually shut down
15:40 mpanetta Meh that is an ideal world with no failures ;)
15:40 mpanetta Ah ok
15:40 mpanetta Then it is not scary ;)
15:40 mkropinack joined #salt
15:41 mpanetta As soon as salt gets true distributed master support then that would be a grad idea.  Until then the master is still a single POF :(
15:41 mpanetta *grand
15:41 pravka joined #salt
15:41 murrdoc pssh use nfs across every single master you have
15:42 murrdoc it will be fine
15:42 murrdoc #yolo
15:42 mpanetta lol
15:42 ek6 murrdoc: afs or nothing...slacker
15:42 murrdoc nah man s3
15:43 murrdoc truly distributed object storage
15:43 murrdoc amirite
15:43 mpanetta swiftfs
15:43 murrdoc hawt
15:43 ek6 guess there is always sneakernet
15:43 murrdoc how about setup consul
15:44 murrdoc put all the keys in consul and use consul templates to publish keys down to every master
15:44 mpanetta MMm consul
15:44 Fiber^ joined #salt
15:44 murrdoc yeah etcd to consul to consul template to disk
15:44 murrdoc that seems the best way
15:45 mpanetta etcd and consul both do the same thing here (KV store) so don't need both
15:45 murrdoc dont u try to make sense right now
15:45 murrdoc guy
15:46 murrdoc :D
15:46 cheus joined #salt
15:46 mpanetta haha
15:48 johntron joined #salt
15:51 lude joined #salt
15:53 lude i have an odd issue, dual master setup, one of them behaves normally, the other i can't run commands on '*', can't even salt-run manage.down, it just hangs
15:53 dopesong joined #salt
15:53 lude the '*' commands return from a bunch of minions, but never actually finishes
15:54 hasues joined #salt
15:54 hasues left #salt
15:56 smcquay joined #salt
16:02 SheetiS joined #salt
16:02 aparsons joined #salt
16:03 murrdoc joined #salt
16:05 iggy think there are some issues open for multi-master
16:05 iggy but I'd say... upgrade to 2015.2 when it's out
16:06 murrdoc 2 weeks
16:06 mpanetta twooooooo weeeeekkksssss
16:06 lude you think for now, just pull one of the masters from the config?
16:07 mpanetta Sorry, had a Total Recall moment...
16:08 renoirb joined #salt
16:08 johntron I created a custom state to download artifacts reliably from Jenkins jobs: https://github.com/johntron/saltstack-jenkins
16:09 murrdoc thats not yaml
16:10 __gotcha joined #salt
16:11 johntron *state module
16:11 murrdoc +1 johntron
16:11 iggy lude: I would do that anyway (until true multi-master is ready)
16:11 lude yah guess i jumped the gun
16:11 v0rtex 2 weeks??! do we have an actual release date?
16:11 lude i upgraded my infrastructure and was like "might as well go multi-master"
16:11 murrdoc btw if you could have used salt[cp.file] for the retrieveal u are doing johntron
16:11 iggy v0rtex: no, that's a running joke
16:12 murrdoc in the IT industry, not jus this chatroom
16:12 v0rtex iggy: thanks :(
16:12 johntron murrdoc: nah, I tried that, but the filename changes with each build, because it includes a version number.
16:12 badon_ joined #salt
16:12 iggy we're coming up on Feb 65th though
16:13 murrdoc sorry, not in your yaml state johntron , in your python code
16:13 v0rtex yep
16:13 murrdoc johntron:  in this area https://github.com/johntron/saltstack-jenkins/blob/master/_states/jenkins.py#L45-L89
16:14 johntron murrdoc: oh, hmm. I tried __salt__[cp.file], but realized __salt__ wasn't defined yet.
16:15 murrdoc yeah
16:15 andreyst joined #salt
16:15 badon_ joined #salt
16:16 murrdoc you would have to import salt.modules.cp and then __salt__ = { cp.file : salt.modules.cp.file }
16:16 murrdoc only reason i mention btw johntron, if salt removes requests as a requirement and moves to urllib2, your state will work the same
16:16 beneggett joined #salt
16:17 johntron murrdoc: yeah, I get it - thanks for the tip. I just couldn't get cp.file working, but I'll try importing like you recommend.
16:18 murrdoc mah bad
16:18 zwi joined #salt
16:19 johntron murrdoc: should it be cp.get_url?
16:19 johntron http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.cp.html#salt.modules.cp.get_url
16:19 murrdoc one second
16:20 elfixit joined #salt
16:22 quintinadam joined #salt
16:23 murrdoc k my bad johntron , https://github.com/saltstack/salt/blob/v2014.7.2/salt/fileclient.py#L497 would be better
16:24 murrdoc cause u need to pass in auth too
16:24 murrdoc and it uses the same code you use with requests and stuff
16:24 bfoxwell joined #salt
16:24 SheetiS joined #salt
16:25 \ask joined #salt
16:26 chandankumar joined #salt
16:27 bhosmer joined #salt
16:27 thedodd joined #salt
16:31 hackel joined #salt
16:35 iggy salt.modules.http?
16:36 theologian joined #salt
16:36 murrdoc thats 2015.2 only
16:38 emaninpa joined #salt
16:38 crdarwin Anyone around to respond to a PR on the docker formula? https://github.com/saltstack-formulas/docker-formula/pull/18
16:38 iggy oh right
16:39 iggy for once, I'm ahead of everybody else
16:39 iggy crdarwin: someone will get to it
16:41 andrew_v joined #salt
16:41 Auroch joined #salt
16:44 rocket joined #salt
16:46 bhosmer_ joined #salt
16:50 MTecknology This is interesting... I can clone a repo just fine until I try to have salt-master do it... Then I get this - 2015-04-03 11:49:19,736 [salt.pillar.git_pillar ][ERROR   ] Unable to fetch the latest changes from remote git@git.domain.tld:moses/moses-pillar.git: 'git fetch' returned exit status 128: fatal: protocol error: bad line length character: No s
16:50 dimeshake iggy: took your advice and have a PR in to use map.jinja with the epel formula, too
16:50 forrest joined #salt
16:52 MatthewsFace joined #salt
16:53 MatthewsFace joined #salt
17:05 otter768 joined #salt
17:09 badon_ joined #salt
17:09 captine joined #salt
17:10 |Fiber^| joined #salt
17:11 eth2 joined #salt
17:19 druonysus joined #salt
17:20 notnotpeter joined #salt
17:22 seev what is the best way to check if a pillar is defined?
17:22 robawt seev: use salt-call pillar.raw
17:22 robawt from the target host
17:22 seev I have a key: value pair and I've tried {% if salt['pillar.get']('key') == 'blah' %}
17:23 seev it's not giving an error, but it bypasses that part of my state when I execute it on the minion
17:26 ajw0100 joined #salt
17:29 murrdoc from the cli run
17:29 murrdoc salt 'minionname' pillar.get key
17:31 seev that works, I have tried running the state with debug enabled and it's just not rendering the {% if %} {% endif %} block - outputs blank whitespace
17:32 murrdoc it works on the minion u are running the state against
17:32 murrdoc ?
17:33 seev I can run the salt state with a 'salt-call state.sls statename' without any errors
17:34 Eureka_ @seev I have had that happen as well. You may not be matching anything with the if statement and its just inserting whitespace rather than what you are expecting.
17:34 murrdoc salt-call ?
17:34 seev https://gist.github.com/anonymous/8549c02f3d4726c66e64
17:34 murrdoc try state.show_sls
17:34 seev here's the state snippet
17:34 murrdoc oh
17:34 murrdoc you dont need the 'True'
17:35 murrdoc {% if userdata['admin']} will suffice
17:35 dude051 joined #salt
17:35 seev that did it, thank you
17:41 echo joined #salt
17:41 beneggett joined #salt
17:44 ajw0100 joined #salt
17:45 bhosmer joined #salt
17:46 iggy == 'True' is string comparison
17:47 ajw0100_ joined #salt
17:49 markm joined #salt
17:51 elektrix_ quick question...am following best practices by having my default variables set in map.jinja and then values in pillar if I want something other than the default. Issue is that I want the default value to be a blank string, but instead it returns "None"
17:51 jtang joined #salt
17:54 dalexander joined #salt
17:57 murrdoc joined #salt
17:59 JlRd joined #salt
18:03 JDiPierro joined #salt
18:05 druonysuse joined #salt
18:09 cpowell joined #salt
18:10 drawsmcgraw1 joined #salt
18:13 ajw0100 joined #salt
18:20 ingwaem joined #salt
18:23 seev I finally got my users stored in pillar, I had defined them in an ad-hoc list in a state file
18:23 seev it was pretty terrible
18:27 murrdoc joined #salt
18:27 elektrix_ in short, is there a way to set a variable as a blank string and not have it return "None"?
18:28 JlRd joined #salt
18:28 miscreant joined #salt
18:29 badon_ joined #salt
18:29 markm_ joined #salt
18:30 seev you can use salt['pillar.get']('variable', '') to call it with a default value of ''
18:31 spicyWith joined #salt
18:40 yohnnyjoe left #salt
18:40 spookah joined #salt
18:41 ajw0100 joined #salt
18:44 miscreant I'm learning salt and was curious, do we *have* to store our states in /srv/salt? Can I store them in my $HOME instead and pass some cli param to salt to tell it where my states are?
18:46 beneggett joined #salt
18:47 denys joined #salt
18:47 murrdoc yes
18:47 khaije1 joined #salt
18:48 elektrix_ you can set the default store directory in the master file (e.g. /srv/salt/master)
18:49 elektrix_ oh...not setting the default, but wanting to call the directory from the command line...my bad
18:51 JupiterFea joined #salt
18:51 JupiterFea hey
18:51 miscreant yeah, so I can setup "per project" stacks.. though it looks like we could achieve the same thing through host naming schemes, ie. projectname-dev projectname-prod
18:51 JupiterFea I was wondering if there is some web ui dashoard for the job status? (something active and non enterprise only)
18:54 khaije1 JupiterFea: http://docs.saltstack.com/en/latest/topics/tutorials/halite.html may be what you're looking for
18:54 mpanetta I think Halite is dead
18:54 JupiterFea unfortunately it is abandoned :(
18:54 khaije1 oops, I didn't know
18:54 rocket elektrix_: it doesnt work that way from what I can tell..., but if you operate in masterless mode you can specify a config to load instead ..
18:55 badon_ joined #salt
18:55 quintinadam joined #salt
18:56 mpanetta JupiterFea: Try https://github.com/tinyclues/saltpad
18:56 JupiterFea mpanetta: thanks
18:57 mpanetta YW
18:57 mpanetta I have not personally used it, but it looks pretty spiffy
18:58 ajw0100 joined #salt
19:01 hackel joined #salt
19:04 timoguin joined #salt
19:04 druonysus joined #salt
19:05 otter768 joined #salt
19:08 ajw0100 joined #salt
19:10 ajw0100_ joined #salt
19:12 eagen joined #salt
19:13 miscreant left #salt
19:13 miscreant joined #salt
19:26 ek6 iggy: you running off dev right now or did you bump to rc2 tag?
19:28 elektrix_ rocket: thanks...good to know
19:29 sunkist joined #salt
19:32 catpig joined #salt
19:37 smcquay joined #salt
19:40 learning joined #salt
19:41 learning I seeing jinja written as {%- for pkg in packages %} I can't find any reference to what the dash after the {% would mean. Any ideas?
19:44 murrdoc remove whitespace
19:44 MindDrive learning: http://jinja.pocoo.org/docs/dev/templates/ - it mentions a bit down the page that the '-' is for whitespace stripping
19:45 learning thank you both
19:45 bhosmer joined #salt
19:53 MindDrive Urgh... so I have a custom module that does things that take a while... using salt.client.Caller.function() with 'publish.full_data' sadly has a timeout of 5 seconds and from what I could determine, I can't actually raise the timeout value from within the 'function()' call, so...
19:53 MindDrive Is there a way to do a call asynchronously instead, then poll for the results?
19:54 JayFK joined #salt
20:06 druonysuse joined #salt
20:06 beneggett joined #salt
20:13 MindDrive No suggestions? :)
20:13 beneggett joined #salt
20:14 murrdoc MindDrive:  no clue
20:15 MindDrive murrdoc: Had a feeling no one was sure... it's frustrating because I'm *so* close to getting Salt fully integrated into my application, but this one little thing is holding me up!
20:15 murrdoc so publish is not waiting ?
20:18 Setsuna666 joined #salt
20:20 MindDrive publish.full_data, but yes.
20:20 MindDrive Default timeout of 5 seconds, which it strictly adheres to. :)
20:20 Setsuna666 We are exploring Salt Syndic and would like to confirm that you cannot run states from the master of masters to the minions of a syndic ?
20:21 MindDrive I'd call it directly, but I'm not sure that's actually possible - I'm using Caller.function(...) so I can get the right client configuration read in.
20:21 murrdoc is it a state
20:22 murrdoc cos u can call state.high within python to render any state
20:23 jonher937 Setsuna666: We are doing just that with syndics in our environment, master of master will push highstate to all minions owned by syndics as well
20:25 Setsuna666 jonher937: But the actual files are not stored on the master of master, right ?
20:25 Setsuna666 jonher937: But on the syndic master
20:25 perfectsine_ joined #salt
20:26 jonher937 Setsuna666: We have git configured on the syndic as well if that's what you are asking
20:27 Setsuna666 jonher937: What I'm trying to understand is that you actually need to transfer the states and or pillars from the master of masters to the syndic master before being able to run the state.highstate from the master of masters
20:28 Setsuna666 jonher937: You cannot run a state from the master of masters directly on minions from a syndic, you need to sync them to the syndic, right ?
20:28 amcorreia_ joined #salt
20:29 jonher937 Setsuna666: You need to have all state and pillar files in git or similar so that the synic access the same data as the master of master
20:30 jonher937 Setsuna666: So yes, the states would have to be "synced"
20:30 Setsuna666 jonher937: Ok, great, this is what I tought. Thanks for the help
20:33 MatthewsFace joined #salt
20:33 toddnni joined #salt
20:39 bhosmer joined #salt
20:41 dendazen I have this error
20:41 dendazen Detected conflicting IDs, SLS IDs need to be globally unique.
20:41 dendazen The conflicting ID is 'boost-static' and is found in SLS 'base:packages.boost-static' and SLS 'base:packages.boost-devel'
20:41 dendazen what could that be?
20:42 murrdoc u are using environments
20:42 dendazen yes
20:46 dendazen so i have role development
20:47 dendazen and in there i have in init.sls - packages.boost
20:47 dendazen i also have packages/boost-static.sls
20:48 dendazen oh i found the issue
20:56 druonysus joined #salt
20:56 druonysus joined #salt
21:04 druonysuse joined #salt
21:06 otter768 joined #salt
21:07 niseak joined #salt
21:09 TyrfingMjolnir joined #salt
21:13 minum joined #salt
21:15 minum I'm using a minion to fire an event to the salt master. Is there any way to have the master return the results (or some success/failure signal) to the minion's console?
21:15 iggy ek6: 2015.2 branch (so not dev per se, but not a tag)
21:15 minum (using a reactor btw)
21:16 ek6 iggy: and i havent heard you 'fall down/go boom' yet so things relatively smooth?
21:19 debian112 can you do compound matching in a state file?
21:19 debian112 or is it just grains
21:19 iggy ek6: much better than 2014.1 (and our failed attempts at 2014.7)
21:20 iggy we never even got 2014.7 into dev... we skipped it
21:20 ek6 iggy: you stuck with zmq though didnt you (makes more chicken sounds) ?
21:21 iggy MindDrive: can you leverage the reactor? i.e. call whatever you need and instead of waiting for it to finish, it fires off a custom event when it's done (polling is bad... mkay?)
21:21 MindDrive Iggy: That would be much more preferrable than polling, yes!  Just not sure how that would work from within another Python application. :)  I'll do some reading up.
21:22 iggy ek6: we did so far... we only have ~100 minions and just the 2915.2 update resolved most of our "network related" issues
21:22 MindDrive Whoo, the future!  (2915)
21:23 iggy MindDrive: you are already calling publish, so it shouldn't be hard to have it do an event.fire_master after it's done
21:24 ek6 no no mindrive, i think iggy is amending his guess of which day in feb that 2015.2 will be out for real
21:24 dimeshake debian112: you can in the top.sls
21:24 dimeshake for pillar / states
21:25 iggy minum: no
21:25 Dev0n joined #salt
21:25 debian112 dimeshake: I can there
21:26 debian112 but I have need to do it in the state file
21:26 sunkist joined #salt
21:26 iggy that probably means you are doing it wrong
21:26 debian112 top file will be super long
21:27 iggy gist code so we can see what you're trying (even pseudo code works)
21:27 mpanetta joined #salt
21:27 iggy I mean most of the code I write is pseudo code
21:27 debian112 hold on crap, i got to run.
21:27 debian112 I will explain on Monday
21:28 mpanetta joined #salt
21:29 subsignal joined #salt
21:31 minum iggy: Thanks!
21:31 perfectsine joined #salt
21:36 murrdoc joined #salt
21:42 scbunn joined #salt
21:53 yidhra joined #salt
21:53 bhosmer_ joined #salt
21:56 forrest joined #salt
21:57 notnotpeter joined #salt
22:05 seev is there a way to insert an entire pillar tree
22:05 seev without doing iteration or individual items
22:07 jonlangemak joined #salt
22:09 jonlangemak Hi guys - quick question.  Im rather new to salt so this might be an easy one…  Is it possible to do jinja based templating on files that you copy down to the host (config files or others) through a state?  Ideally I’d like to pull pillar data in the configs but it seems like the pillar data is more for state files.
22:10 JDiPierro joined #salt
22:11 jonlangemak So I guess the short question is can I use pillar data in files I pull down through things like file.managed?
22:12 o5k joined #salt
22:13 druonysuse joined #salt
22:13 druonysuse joined #salt
22:13 scbunn joined #salt
22:14 ajw0100 joined #salt
22:14 forrest left #salt
22:17 quintinadam joined #salt
22:20 alynpost left #salt
22:23 TyrfingMjolnir joined #salt
22:29 jonlangemak joined #salt
22:31 CeBe joined #salt
22:36 timoguin joined #salt
22:39 ajw0100 joined #salt
22:41 jonlangemak joined #salt
22:44 forrest joined #salt
22:47 TyrfingMjolnir joined #salt
22:47 aquassaut joined #salt
23:01 jerematic joined #salt
23:04 forrest joined #salt
23:07 otter768 joined #salt
23:09 ek6 joined #salt
23:12 notnotpeter joined #salt
23:13 miscreant left #salt
23:20 Eugene Yes, this is exactly what templating does
23:22 sunkist joined #salt
23:30 MatthewsFace joined #salt
23:30 MindDrive Iggy: someone on the salt-users mailing list actually gave me a solution to my earlier issue that was ridiculously simple and worked like a charm: result = caller.sminion.functions['publish.full_data'](host, cmd, args, timeout=30)
23:34 Gareth whiteinge: ping
23:36 JDiPierro joined #salt
23:38 bhosmer_ joined #salt
23:39 dendazen joined #salt
23:40 markm joined #salt
23:42 captine joined #salt
23:45 wolfpackmars2 joined #salt
23:47 ahammond I want to run salt foo --batch-size 10% state.sls my_state via a reactor. How do I get the --batch-size 10% stuff in there?
23:48 markm joined #salt
23:49 captine joined #salt
23:49 teskew joined #salt
23:58 renoirb joined #salt
23:58 dude051 joined #salt

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