Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-05-20

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

All times shown according to UTC.

Time Nick Message
00:00 KyleG joined #salt
00:00 KyleG joined #salt
00:00 MatthewsFace joined #salt
00:00 tmclaugh[work] joined #salt
00:01 lnr joined #salt
00:02 lnr joined #salt
00:02 lnr joined #salt
00:05 lnr joined #salt
00:05 lnr joined #salt
00:05 mens joined #salt
00:05 lnr joined #salt
00:06 lnr joined #salt
00:06 lnr joined #salt
00:06 lnr joined #salt
00:06 neogenix I think I may have a bug at hand.
00:06 neogenix anyone feel like validating it with me ?
00:07 lnr joined #salt
00:07 lnr joined #salt
00:07 lnr joined #salt
00:08 aurynn basepi, UtahDave, can you block lnr for a while?
00:08 lnr joined #salt
00:08 lnr joined #salt
00:08 forrest neogenix: I'm headed out soon, but I'd check the issues page and confirm someone didn't already report it if you're running the latest release.
00:08 lnr joined #salt
00:09 lnr joined #salt
00:09 lnr joined #salt
00:09 lnr joined #salt
00:09 lnr joined #salt
00:10 lnr joined #salt
00:10 lnr joined #salt
00:10 lnr joined #salt
00:11 lnr joined #salt
00:11 lnr joined #salt
00:11 aurynn la la la
00:13 Tyrm joined #salt
00:13 mrbigglesworth joined #salt
00:13 UtahDave Hm. did I do that correctly?
00:13 aurynn looks like, yes
00:14 UtahDave @bansearch lnr
00:14 lnr joined #salt
00:14 UtahDave crap
00:14 lnr joined #salt
00:14 forrest UtahDave: I think we need a freenode admin to drop him if I remember correctly
00:14 lnr joined #salt
00:15 forrest Usually I'd have Corey do it but I don't see him around and the other two staff members are away
00:15 lnr joined #salt
00:15 KyleG UtahDave: /mode +b Inr!*@aim.engr.arizona.edu
00:15 aurynn Someone could camp the nick and keep the IRC connection from establishing
00:15 lnr joined #salt
00:15 lnr joined #salt
00:15 UtahDave thanks, KyleG!
00:15 mosen its a lowercase L
00:16 neogenix forrest: yessir. Checked, and can't find anything. I think it's related to my other bug, but I'll check back with you in the morning.
00:16 KyleG wtf he still made it in?
00:16 forrest neogenix: Sounds good
00:16 forrest KyleG: Yep
00:16 forrest I'm trying to find who he is.
00:16 mosen there we go :)
00:16 KyleG Did I mixx up l and I
00:16 forrest or she, whoever
00:16 mosen gotta make a proxy minion for bans
00:16 KyleG eye and ell
00:16 KyleG lol
00:16 forrest haha
00:17 ajw0100 joined #salt
00:17 UtahDave forrest: did you get lnr's info before we figured out the ban syntax?
00:19 forrest no
00:19 forrest but I'm fine with a ban
00:19 forrest looks like that user does the same thing in several other channels
00:19 UtahDave :)   I'll remove it tomorrow and see if his/her client is still acting up
00:19 forrest cool, lots of IRC loggers show them having this issue
00:20 UtahDave ok.
00:20 UtahDave well, time for me to head home.  catch you all tomorrow
00:21 forrest later
00:22 KyleG he/she is in another room I idle
00:22 KyleG and there's no ops
00:22 KyleG the pain, lol
00:23 hybridpollo joined #salt
00:26 dendazen joined #salt
00:30 Tyrm joined #salt
00:39 saysay joined #salt
00:39 saysay Hi, sll
00:40 mapu joined #salt
00:41 saysay can anyone tell me when the first keygen is generated?
00:41 saysay At install or on boot
00:44 c10b10 joined #salt
00:51 julez joined #salt
00:56 ip` joined #salt
00:59 ALLmightySPIFF joined #salt
01:00 ageorgop joined #salt
01:03 Ryan_Lane saysay: for salt?
01:03 Ryan_Lane I think when the minion starts
01:07 mrbigglesworth joined #salt
01:14 yexingok joined #salt
01:14 mdupont joined #salt
01:14 ip` left #salt
01:15 saysay Yes: Ryan_Lane
01:16 saysay During install does it also create a key
01:22 aparsons joined #salt
01:26 beauby joined #salt
01:27 ITChap joined #salt
01:27 saysay nevermind
01:36 otter768 joined #salt
01:45 cromark joined #salt
01:49 big_area joined #salt
01:49 kusams joined #salt
01:55 Tyrm joined #salt
01:58 ageorgop joined #salt
02:00 timoguin_ joined #salt
02:08 beauby joined #salt
02:23 beauby joined #salt
02:23 TyrfingMjolnir joined #salt
02:25 mrbigglesworth joined #salt
02:27 MatthewsFace joined #salt
02:31 CeBe joined #salt
02:31 writtenoff joined #salt
02:33 c10b10 joined #salt
02:33 desposo joined #salt
02:40 julez joined #salt
02:40 VR-Jack2 joined #salt
02:42 Nazca__ joined #salt
02:53 ageorgop joined #salt
02:57 michelangelo joined #salt
03:00 rhodgin joined #salt
03:03 ITChap joined #salt
03:04 enarciso joined #salt
03:04 jalaziz joined #salt
03:08 beauby joined #salt
03:17 favadi joined #salt
03:27 mrbigglesworth joined #salt
03:28 dendazen joined #salt
03:32 cromark joined #salt
03:35 soren_ joined #salt
03:37 otter768 joined #salt
03:40 cberndt joined #salt
03:41 ageorgop joined #salt
03:42 mosen joined #salt
03:51 enarciso joined #salt
03:55 UForgotten_ joined #salt
03:57 joeto joined #salt
04:01 hal58th__ joined #salt
04:02 hal58th_1 joined #salt
04:03 UForgotten joined #salt
04:06 Singularo joined #salt
04:09 joeto1 joined #salt
04:17 mrbigglesworth joined #salt
04:20 cberndt joined #salt
04:21 echo joined #salt
04:22 c10b10 joined #salt
04:25 clintberry joined #salt
04:36 mrbigglesworth joined #salt
04:37 golodhrim|work joined #salt
04:43 elfixit joined #salt
04:44 cromark joined #salt
04:47 ramteid joined #salt
04:50 sinenitore joined #salt
04:50 hal58th joined #salt
05:00 ALLmightySPIFF joined #salt
05:05 cromark_ joined #salt
05:07 soren joined #salt
05:08 enarciso joined #salt
05:18 ITChap joined #salt
05:23 elfixit joined #salt
05:30 thehaven joined #salt
05:32 catpigger joined #salt
05:36 enarciso joined #salt
05:38 otter768 joined #salt
05:39 Berty_ joined #salt
05:40 zhiwliu joined #salt
05:42 rdas joined #salt
05:42 cromark joined #salt
05:55 TyrfingMjolnir joined #salt
05:56 dopesong joined #salt
05:59 mrbigglesworth joined #salt
06:01 tmclaugh[work] joined #salt
06:02 colttt joined #salt
06:03 impi joined #salt
06:04 ithello joined #salt
06:10 c10b10 joined #salt
06:10 samed joined #salt
06:10 samed joined #salt
06:18 clintber_ joined #salt
06:21 flyboy joined #salt
06:23 JayFK joined #salt
06:23 dopesong joined #salt
06:26 viq joined #salt
06:30 KermitTheFragger joined #salt
06:39 Grokzen joined #salt
06:44 stephanbuys joined #salt
06:47 cromark joined #salt
06:48 kawa2014 joined #salt
06:48 c10b10 joined #salt
06:50 Auroch joined #salt
06:52 markm joined #salt
06:54 VR-Jack joined #salt
06:54 jhauser joined #salt
07:00 viq joined #salt
07:08 David_an11 joined #salt
07:08 kusams_ joined #salt
07:09 dopesong joined #salt
07:10 c10b10 joined #salt
07:10 david_an11 joined #salt
07:12 CeBe joined #salt
07:15 eseyman joined #salt
07:17 CeBe joined #salt
07:18 c10b10 joined #salt
07:18 dRiN joined #salt
07:19 zer0def joined #salt
07:20 david_an11 [CRITICAL] Could not deserialize msgpack message: Unexpected Error in MworkerThis often happens when trying to read a file not in binary mode. Please open an issue and include the following error: unpack(b) received extra data.
07:20 david_an11 Hi Is there a solution for this bug
07:22 CeBe1 joined #salt
07:28 thalleralexander joined #salt
07:32 dimeshake joined #salt
07:32 chiui joined #salt
07:35 joeto joined #salt
07:35 babilen joined #salt
07:38 wm-bot4 joined #salt
07:39 otter768 joined #salt
07:41 jeanneret joined #salt
07:45 TheHelmsMan joined #salt
07:47 supersheep joined #salt
07:48 lb1a joined #salt
07:48 jeanneret Hi I have a problem for installing Windows Software I want to install only on os Windows but it doesn't work... https://gist.github.com/jeanneret/57ff5180290ec89bff0a thank you in advance for your help
07:50 AndreasLutro jeanneret: http://docs.saltstack.com/en/latest/ref/states/top.html#other-ways-of-targeting-minions you're missing the - match: ...
07:51 clintberry joined #salt
07:51 supersheep joined #salt
07:54 jeanneret Thank you I think it work sorry I'm tired
08:00 stephanbuys joined #salt
08:04 keimlink joined #salt
08:08 TyrfingMjolnir joined #salt
08:17 julez joined #salt
08:19 Grokzen joined #salt
08:26 impi joined #salt
08:31 robot_hands is there a way to query the host_cache pre-5.6?
08:32 robot_hands query/show
08:32 robot_hands oh wrong channel, my bad bro's
08:34 colttt joined #salt
08:38 mrbigglesworth joined #salt
08:41 Berty_ joined #salt
08:52 supersheep joined #salt
08:57 HarryUP joined #salt
08:57 HarryUP hipper channel
09:00 badon joined #salt
09:03 adelcast joined #salt
09:08 is_null hi all, why does salt-cloud fail, reporting that i should check minion logs, and there's no minion logs ? salt-cloud debug out: https://dpaste.de/gRzX
09:09 is_null that debug log doesn't look good: Error loading runners.nacl: libnacl import error, perhaps missing python libnacl package
09:10 is_null humn, instance has no network, that's probably not good, sry for the noise ^^
09:11 phx don't worry, we're professional rubber ducks here
09:12 Sathya_ joined #salt
09:12 Sathya_ Hi All
09:13 Sathya_ yesterday i upgrade my salt master from 2014 to salt 2015.5.0 (Lithium)
09:13 Sathya_ after this upgrade, when i try to execute a command targetting my nodegroups i get the below mentioned error
09:14 Sathya_ No minions matched the target. No command was sent, no jid was assigned.
09:14 is_null try with -l trace and paste the output online
09:14 linjan joined #salt
09:14 Sathya_ sure
09:15 kawa2014 joined #salt
09:16 Sathya_ https://www.refheap.com/101323
09:17 denys joined #salt
09:19 Sathya_ Please refer the above mentioned link for the output of trace command
09:24 Sathya_ M seeing some missing files
09:24 Sathya_ [DEBUG   ] Missing configuration file: ~/.saltrc [TRACE   ] None of the required configuration sections, 'logstash_udp_handler' and 'logstash_zmq_handler', were found the in the configuration. Not loading the Logstash logging handlers module.
09:25 TyrfingMjolnir joined #salt
09:29 stephanbuys joined #salt
09:31 mrbigglesworth joined #salt
09:31 impi joined #salt
09:34 Sathya_ Any guess ?
09:35 is_null Sathya_: not really, i'm a noob withsalt, but could you try downgrading salt to the last version that worked for you, run the same command (again with trace) and paste the output ?
09:36 is_null we could maybe find a clue in the difference between both outputs ;)
09:40 otter768 joined #salt
09:40 Sathya_ downgraded to 2014 and the same command works just fine
09:41 Sathya_ guess this might be a bug or something which needs to be fixed in the latest version
09:43 lictor36 joined #salt
09:44 is_null could you paste online the trace of the same command on 2014 please ?
09:44 is_null a diff would probably give some clu
09:49 fbretel joined #salt
09:51 fredvd joined #salt
09:52 Sathya_ https://www.refheap.com/101324 ------> working version - salt 2014.7.5 (Helium)
09:53 fbretel hi all, I get "Command 'npm --version' failed with return code: 127" errors since upgrading to 2015.5.0, and although I don't seem to have calls to the npm module/state. Am I the only one ?
09:54 duaeo joined #salt
09:57 is_null Sathya_: what's pretty interresting is that with 2014 you get func get_cli_event_returns() and then -[DEBUG   ] get_iter_returns for jid 20150520151234855784 sent to set(['XXXXX078.xx.xxx.xxx', 'XXXXX049.csscorp.com',...), but in 2015 you got sending event, get_event() and nodegroup_comp("Ambit") => ( S@10.9.XX.XX/24,S@10.XX.XX.XX/24 )
09:57 is_null did you read release notes for 2015-05 ? such a change should be mentioned
09:59 codehotter What's the right word to google if I want to automate a management task with salt?
09:59 codehotter Not a state a server needs to be in, but, let's say a set of instructions for creating an amazon ami
10:01 is_null codehotter: maybe salt-cloud or orchestration, not sure what an ami is
10:01 codehotter image
10:01 is_null i see, probably orchestration then
10:04 renat joined #salt
10:06 sgate1 joined #salt
10:08 ramaseshan joined #salt
10:08 c10b10 joined #salt
10:09 renat Hello! Who knows, how correctly set start and end parameter in percent in module.parted (partition) mkpart? One significant moment - this module running from sls, where is wrapped into module.run
10:11 renat Or, maybe not in percents. What I need is the one partition, that use entire disc.
10:14 renat Docs says that start and end in megabytes by default
10:24 is_null kiorky: hi, is there a way to set the value of the -P switch of the lxc command anyhow, anywhere ?
10:24 is_null or maybe i would try to add it to your PR which I haven't yet fully reviewed
10:25 klocek joined #salt
10:25 Sathya_ is_null: i havent gone through the release notes yet
10:26 bluenemo joined #salt
10:26 is_null i can't find any way by reading the source code of 2015-05 which contains a docstring "ask kiorky", but the question is open to anyone
10:26 julez joined #salt
10:26 is_null Sathya_: http://docs.saltstack.com/en/latest/topics/releases/2015.5.0.html
10:26 kiorky is_null: which lxc command
10:27 is_null kiorky: if you use -P with lxc-create, then you'll need the same -P for lxc-info
10:27 is_null and lxc-start, etc, etc ... this is useful to use lxc containers in tmpfs for CI
10:27 kiorky is_null: what are you using ? the lxc runner ? salt cloud ? directly the exec module ?
10:27 c10b10 joined #salt
10:27 is_null i'm currently using salt-cloud but I'd use any way that lets me set the -P option
10:28 kiorky is_null: /var/lib/lxc is hardcoded a lot if i can recall well, maybe the easier for you is for such envs to directly plkace your /var/lib/lxc under tmpfs
10:28 kiorky is_null: salt_cloud use in first place lxc.init
10:28 kiorky not lxc.create
10:29 kiorky is_null: so you have to traverse more layers to give your creation argument.
10:29 is_null it should traverse along with the -n option because they are tied
10:29 kiorky is_null: saltcloud -> runner lxc -> module lxc.cloudinitinterfface  -> runnerlxc -> module lxc.inir
10:29 kiorky is_null: saltcloud -> runner lxc -> module lxc.cloudinitinterfface  -> runnerlxc -> module lxc.init
10:29 kiorky is_null: and in lxc.init, there is the final call to lxc.create
10:30 kiorky is_null: i know, but for know, the -P isnt supported
10:30 kiorky it is not that hard to implement, through
10:30 is_null that's why i was wondering if it would help that i add it to your PR
10:30 kiorky is_null: but you ll have to transport the options from the cloud init interface up to the lxcreate call
10:30 giantlock joined #salt
10:30 kiorky is_null: i ll have to go to lunch
10:30 is_null bonap ;)
10:31 kiorky is_null: when are you avalaible in the next hours ?
10:31 is_null absolutely kiorky
10:31 kiorky is_null: you can add it, but i doubt that it will be accepted for 2015.5
10:31 kiorky is_null: as it is a new feature
10:31 kiorky is_null: you can ask basepi for that :)
10:32 kiorky is_null: but for develop, i dont see any blocker
10:32 kiorky is_null: i can help you to review or tell you where to do stuff, no problem
10:32 kiorky i just need now to run
10:32 is_null thanks a lot, bonap
10:32 kiorky is_null: but order of operation: support option in lxc.create
10:32 kiorky is_null: then in lxc.init
10:32 kiorky is_null: then in lxc.cloud_init_interface
10:32 kiorky from there, you are done
10:33 is_null i'll have lunch first too, we're in the same timezone, thanks a heap
10:33 kiorky fine :)
10:35 Sathya_ Now that i was having one more problem with salt was the autocomplete in RHEL
10:36 Sathya_ I can see the autocomplete working in RHEL, but its dam slow,
10:36 Sathya_ as soon as i hit a command and hit tab, i have to wait for atlest 3 min to list out the options
10:36 Sathya_ does any one have this problem ?
10:37 david_an11 joined #salt
10:38 julez joined #salt
10:39 N-Mi joined #salt
10:40 Berty_ joined #salt
10:53 CeBe joined #salt
10:54 kawa2014 joined #salt
10:57 wnkz joined #salt
10:57 mrbigglesworth joined #salt
10:58 kiorky is_null: i have a pr nearly finished
10:58 fbretel left #salt
10:59 pf_moore joined #salt
10:59 HarryUP joined #salt
10:59 salt-n00b joined #salt
11:00 HarryUP joined #salt
11:09 amcorreia joined #salt
11:10 jespada joined #salt
11:12 Sathya_ Is there a way to grep particular services using salt sls file and then to kill those services ?
11:12 evle1 joined #salt
11:12 Sathya_ or how can we instruct salt to execute ps -ef and gerep those ID's and then to kill it in one single command
11:12 c10b10 joined #salt
11:13 AndreasLutro Sathya_: probably not, just write a script and execute it with cmd.run or cmd.wait
11:13 Sathya_ ok
11:17 joehh is_null, babilen: I'll fix that up, thanks for the report
11:17 babilen +1
11:19 joehh I believe that should be fixed now
11:20 quist joined #salt
11:21 HarryUP joined #salt
11:25 dubey joined #salt
11:25 dubey Hi
11:25 Setsuna666 joined #salt
11:32 favadi left #salt
11:35 cromark joined #salt
11:35 kiorky is_null: ubuntu-vivid
11:35 kiorky [INFO    ] Executing command 'lxc-create -n foobar -P /srv/mytest -f /tmp/tmpG4ighB -t ubuntu -- --release trusty --arch amd64' in directory '/root'
11:35 kiorky is_null: :)
11:36 cromark_ joined #salt
11:36 saltuser joined #salt
11:41 otter768 joined #salt
11:41 dimeshake joined #salt
11:43 multani joined #salt
11:43 bmcorser joined #salt
11:47 Chronojam joined #salt
11:49 cheus_away joined #salt
12:01 Berty_ joined #salt
12:01 kiorky is_null: well
12:01 kiorky there is a lot of change, in fact
12:01 kiorky but im on it
12:02 kiorky is_null: https://github.com/makinacorpus/salt/commit/1a992cf4720cc925057f0543b7b765fcc3772e34
12:02 kiorky current changeset
12:06 thehaven_ joined #salt
12:07 flyboy82 joined #salt
12:08 mpanetta joined #salt
12:17 Auroch joined #salt
12:19 rhodgin joined #salt
12:21 quist joined #salt
12:21 overyander joined #salt
12:28 is_null kiorky: i hope i'll rock like you when i grow up xD
12:28 matthew-1arlette joined #salt
12:28 is_null i'll proofread carefuly asap and let you know if i find something
12:29 is_null then i can test it ;)
12:32 Hazelesque Hey, so....
12:32 Hazelesque is salt-syndic any good? and how does it work with salt-key etc?
12:32 Hazelesque currently we use CFEngine, and we have separate "policy hubs" for dev, test, prod and our overseas (georedundancy) environments
12:33 Hazelesque we're using CFEngine to deploy salt-minion (everywhere) and salt-master (on policy hubs), and to configure the salt minions to use the IP address of their CFEngine policy hub as their salt master too
12:34 Hazelesque thus, we have multiple salt-masters at the moment
12:34 Hazelesque and it would be nice to be able to see everything from the prod one
12:34 big_area joined #salt
12:35 Hazelesque will salt-syndic do the job? I've seen reports that it's slow...
12:36 nyx__ joined #salt
12:40 TheHelmsMan joined #salt
12:42 is_null kiorky: why can we specify a path for lxc_config in https://github.com/makinacorpus/salt/commit/1a992cf4720cc925057f0543b7b765fcc3772e34#diff-f86c714cf791c98b23d05f8c20375087R2185 ?
12:42 is_null isn't it always $PATH/$NAME/config ?
12:43 jeanneret joined #salt
12:44 is_null what would happen if lxc_config argument gets a value that's different from $PATH/$NAME/config ? wouldn't that potentially lead to obscure user-misconfiguration ?
12:45 mage_ does the git state work with digest auth ?
12:45 jeanneret Hi I'm again here. I want to do Windows Update on my computer I had done that https://gist.github.com/jeanneret/eee98469c381f2ec8b24 but when I do that it apply one or 2 updates miss I anything? Thanks in advance for your help
12:45 kiorky is_null: it is juist a convention
12:46 kiorky is_null: but in absolute, it can also be different
12:46 fredvd joined #salt
12:46 kiorky is_null: https://github.com/makinacorpus/salt/commit/83e3f292f8d04c759eebd595234c487826e73bd9
12:46 kiorky is_null: you can refresh
12:46 kiorky im always on it ;)
12:46 is_null kiorky: will any official lxc userland tool support using an alternative config path at all ?
12:47 kiorky is_null: lxcutils does
12:47 _JZ_ joined #salt
12:47 is_null ok, makes sense then
12:49 bhosmer joined #salt
12:49 is_null kiorky: just a detail, but why do you prefer `if not path: path = DEFAULT_PATH` to `if path is None: path = DEFAULT_PATH` ?
12:51 kiorky is_null: it is the usual pythonic way to test
12:52 is_null a one liner would be path = path or DEFAULT_PATH
12:52 is_null i think that would be the same
12:53 is_null nope, that was wrong, sry
12:53 markm_ joined #salt
12:53 mschiff hmmm contents_pillar in file.managed seems broken in recent versions... is this a known problem? (file.py throws exception that contents has no attribute endswidth...
12:53 is_null kiorky: my bad, it's correct, path = path or DEFAULT_PATH is the same as `if not path: path = DEFAULT_PATH`, but it's a line less
12:53 Setsuna666 joined #salt
12:54 is_null just a detail, nothing important though
12:54 jeanneret AndreasLutro: Hi I'm again here. I want to do Windows Update on my computer I had done that https://gist.github.com/jeanneret/eee98469c381f2ec8b24 but when I do that it apply one or 2 updates miss I anything? Thanks in advance for your help
12:58 is_null time to test that branch now ;)
12:59 iggy mschiff: what version?
13:02 codehotter What's the idea between actions and functions?
13:02 kiorky is_null: it is still not testable
13:02 codehotter Creating an image, is that an action or a function?
13:02 codehotter I create an image FROM a virtual machine, so it's an action? But I create the image separately, from the virtual machine, it's not really an action performend on the virtual machine. So it's a function?
13:02 codehotter What's the reasoning behind the action/function distinction?
13:03 XenophonF joined #salt
13:04 Sathya_ guys any body can tell me how to configure and start salt-api to make it work with jenkins ?
13:05 is_null kiorky: right, so is there anything i can do that would help, ie. working on the tests or docs for that change or something ?
13:05 Sathya_ sorry, Any one can tell me how to configure salt-api and make it to work with jenkins ?
13:06 kiorky is_null: testing thoroughly once done
13:06 kiorky is_null: i hope this to be done in the hour
13:06 Tecnico1931 joined #salt
13:07 numkem_ joined #salt
13:07 is_null kiorky: awesome, take you're time I'm not going anywhere xD
13:09 primechuck joined #salt
13:09 JDiPierro joined #salt
13:09 jdesilet joined #salt
13:10 mapu joined #salt
13:12 dork joined #salt
13:15 hunmaat does anyone use salt-cloud with provider/profile config coming from pillar?
13:15 julez joined #salt
13:16 hunmaat docs aren't really helpful, but i couldn't get it work
13:16 dork any ideas on what's going on here? seems like it's handling the list in jinja strangely? http://www.pastebin.ca/3006112
13:16 is_null got some tracebacks when using salt-cloud: http://dpaste.com/1R3EJ9V does this ring a bell to anyone ?
13:17 c10b10 joined #salt
13:19 dyasny joined #salt
13:20 KennethWilke joined #salt
13:21 cpowell joined #salt
13:22 dendazen joined #salt
13:22 dork guess it's because there was only one item in the list
13:22 dork weird
13:23 is_null dork: what's that supposed to do ? {% for usr in 'rev' %}
13:23 XenophonF Sathya_: are you using salt-formula?
13:23 is_null dork: that's not supposed to work to my knowledge, or maybe have the loop run three times with respectively 'r', 'e', 'v'
13:24 Sathya_ Sorry XenophoneF M new to this saltstack, can you tell me what is salt-formula is ?
13:24 iggy is_null is correct
13:24 is_null dork: if rev is a variable it should be `for user in rev`, not `for user in 'rev'`
13:25 Sathya_ I just googled and found http://docs.saltstack.com/en/latest/topics/development/conventions/formulas.html
13:25 iggy or for usr in ['rev']
13:25 XenophonF Sathya_: take a look at https://github.com/saltstack-formulas/salt-formula and http://docs.saltstack.com/en/latest/topics/development/conventions/formulas.html
13:25 Sathya_ I have to go through that link
13:25 is_null dork: iggy is correct
13:25 XenophonF Sathya_: you can use it to enable the REST API
13:25 is_null so if you're in for an infinite loop i think we're ready xD
13:25 XenophonF Sathya_: or you can look through those states to see how to enable it manually yourself
13:26 Sathya_ ok let me check that our
13:26 Sathya_ out *
13:26 emaninpa joined #salt
13:26 Sathya_ thanks
13:26 XenophonF Sathya_: as for wiring it up to Jenkins, that's not something I can help you with (yet!)
13:26 vovcia is_null: im using salt-cloud
13:28 vovcia is_null: and someone on this # was pasting similar error yesterday
13:29 murrdoc joined #salt
13:31 Berty_ joined #salt
13:32 vovcia how can write SLS for this command: "salt '*' zpool.create pool0 /dev/sdd log /dev/sdc cache /dev/xvdb" ?
13:32 vovcia im beginner with salt :)
13:34 dork is_null: these are supposed to be usernames
13:34 dork http://docs.saltstack.com/en/latest/topics/tutorials/states_pt3.html
13:35 dork i've tried with both having the list in brackets and with the second example as well
13:35 Tyrm joined #salt
13:37 LyndsySimon joined #salt
13:39 LyndsySimon I looked at Salt about a year ago for work, but ended up going with Docker+Ansible. Now I'm looking to automate some personal stuff, and would like to use Salt is possible.
13:40 mage_ it's me or the https_user and https_pass options of the git state file don't work ?
13:40 LyndsySimon Where should I start? I'll be doing 90% WordPress stuff, and would like to be able to provision and deprovision WordPress sites with a single command if possible.
13:41 murrdoc https://github.com/saltstack-formulas/wordpress-formula
13:41 LyndsySimon The other 10% would likely be simple static site hosting, perhaps with an additional build step consisting of pulling a git repo down, build a Python virtualenv, and running a script to generate the content.
13:42 LyndsySimon murrdoc: Excellent, thanks.
13:42 otter768 joined #salt
13:42 LyndsySimon I feel bad asking such a specific and "handhold-y" question, but I promise it won't be a trend from me. :) I just wanted to avoid the first few hours of digging if possible.
13:44 nyx__ joined #salt
13:45 JDiPierro joined #salt
13:46 subsignal joined #salt
13:46 elfixit joined #salt
13:46 perfectsine joined #salt
13:49 hunmaat alias salt-cloud='salt-call cloud.profile' ...
13:49 is_null vovcia: i found out that the minion created by salt-cloud have no access to internet
13:49 is_null so that's likely to be a problem
13:50 is_null dork: right, so it'd be {% for user in ['rev'] %} like iggy suggested
13:50 dork yeah i get the same results
13:51 dork but even in that tutorial it shows the original way i had it as a possibility
13:52 rm_jorge joined #salt
13:53 mage_ https://dpaste.de/JAKx what's wrong with this ? https_user and https_pass don't work at all ..
13:53 Setsuna666 joined #salt
13:54 kiorky is_null: from the salt cloud driver, the path can be hard to give
13:54 kiorky im figuring on how to do it
13:54 kiorky is_null: (the chain is now wired up to the lxc runner)
13:56 kusams joined #salt
13:57 zero joined #salt
13:57 cromark joined #salt
13:59 Morbus joined #salt
14:00 arknix joined #salt
14:03 cowpunk21 joined #salt
14:03 dork it was that home: True tag
14:03 mschiff iggy: The problem is, when pillar data is a list for example (2014.7.1)
14:04 andrew_v joined #salt
14:10 drawsmcgraw1 How does the 'include' statement work, again?  I had assumed that if I put an 'include' statement in the middle of a statefile, then the included states would be injected right where the include statement is.
14:10 drawsmcgraw1 Which has a handy side effect of guaranteeing the order of execution
14:10 s_kunk joined #salt
14:10 MagiCrazy joined #salt
14:10 drawsmcgraw1 i.e. The "install some packages" statefile contains states that install the software, then at the bottom just includes the 'configure' statefile.
14:11 drawsmcgraw1 But... that doesn't seem to be the case... I have my 'configure' states being run before the 'install' states :/
14:11 iggy mschiff: yeah, it needs to be a key with a single string
14:12 sybix joined #salt
14:12 iggy mschiff: but maybe file a bug anyway, that's probably a thing that should be caught and reported rather than throwing an exception
14:13 Morbus left #salt
14:16 nyx__ joined #salt
14:18 mschiff iggy:
14:18 mschiff iggy: ok
14:20 zer0def drawsmcgraw1: is that an exercise? you'd ordinarily make requisites for included states to be dependencies
14:21 drawsmcgraw1 zer0def: I think I partially understand.... do you have an example to illustrate?
14:21 drawsmcgraw1 I'm afraid I'm over-thinking things again
14:22 GrueMaster joined #salt
14:22 zer0def can it be as simple as packages and services?
14:22 drawsmcgraw1 sure.
14:22 drawsmcgraw1 I have states separated out into two files
14:22 drawsmcgraw1 So that I can hook up the 'configure' states to a schedule or Highstate to ensure the configs are correct
14:22 drawsmcgraw1 (just to be sure I'm clear about my intent)
14:23 iggy that's not how include: works
14:24 iggy if you want to do that, use jinja includes
14:24 drawsmcgraw1 iggy: I figured as much...
14:24 drawsmcgraw1 ?
14:24 iggy {% include 'path/to/other.sls' %}
14:25 drawsmcgraw1 !
14:25 drawsmcgraw1 And that *will* inject the states where I put that Jinja line?
14:25 iggy http://jinja.pocoo.org/docs/dev/templates/#include
14:25 iggy yes
14:25 zer0def yeah, it will, forgot about jinja injects
14:25 iggy it's injected as part of the jinja rendering
14:25 iggy vs include: which is actually a special case handled by salt
14:25 zer0def i've dug up something like this: http://codepad.org/L2zo47jU
14:26 slav0nic joined #salt
14:26 zer0def and, as far as requisites go: http://docs.saltstack.com/en/latest/ref/states/requisites.html
14:26 drawsmcgraw1 zer0def: Ah, I understand what you mean with that example
14:27 favadi joined #salt
14:27 timoguin joined #salt
14:27 zer0def i'd go this way, mainly because you then don't rely on an assumed order of execution
14:28 drawsmcgraw1 Yeah. For some reason I wanted to be able to run the 'configure' state file without needing to re-run the 'install' statefile
14:28 drawsmcgraw1 But... I'm having difficulty coming up with a reason why
14:28 zer0def yeah, states sometimes get real minimalistic
14:28 iggy most people just shove them together... pacakage managers are pretty quick to check if a package is installed
14:29 iggy well, decent ones are anyway
14:29 zer0def yeah, i prefer to keep package and state states seperate, so that i can do some shenanigans in-between package being installed and service being (re)started, like modding config files
14:29 zer0def s/state states/service states/
14:30 drawsmcgraw1 Noted. Yeah I misunderstood the Salt 'include' statement. I'll go that route zer0def.
14:30 drawsmcgraw1 Thanks zer0def and iggy !
14:31 iggy drawsmcgraw1: you could always file a bug to clarify the documentation on that (or even submit a PR)
14:31 iggy https://github.com/saltstack/salt/blob/develop/doc/ref/states/include.rst
14:31 drawsmcgraw1 Good point
14:31 is_null hi all, is there another way than hard-coding an ip in each host of a mapfile to make the lxc driver create containers with an eth0 interface on my host's bridge interface network ?
14:31 mage_ any idea for             - 'https_pass' and 'https_user' are invalid keyword arguments for git.latest. If you were trying to pass additional data to be used in a template context, please populate 'context' with 'key: value' pairs. Your approach will work until Salt Lithium is out. Please update your state files.
14:31 mage_ it is explicitely mentionned in the doc that it should work ..?
14:32 mage_ this is with py27-salt-2014.7.5
14:33 jamin joined #salt
14:33 iggy mage_: that was added in 2015.5
14:33 iggy (and not documented as such grrrrrr)
14:34 mage_ mmh
14:35 mage_ ok it's my fault, I was browsing doc for version 2015 ..
14:35 iggy I love commenting on 9 month old commits
14:35 rhodgin joined #salt
14:35 drawsmcgraw1 iggy: I'm thinking of a quick note in "include.rst" along the lines of "Hey, this doesn't actually inject states where you put me. In fact, it will run the included states before the statefile you're in"
14:36 drawsmcgraw1 Thoughts?
14:36 iggy something like that
14:36 iggy it comes up often enough, it's not a terrible idea to add some sort of note
14:36 Norrland I'm having some trouble getting debconf.set_file to render my jinja2 templated files. Runn
14:37 Norrland Running 2015.5 on both master/minion.
14:37 iggy I could swear, include: used to have to be at the top of the file, but I can't find any docs to support that
14:37 drawsmcgraw1 will do. Thanks
14:37 big_area joined #salt
14:38 drawsmcgraw1 iggy: Agreed. In times that I've used it in the past, it's always been at the top
14:38 drawsmcgraw1 Plus, I feel really dirty for wanting to put an include stanza in the middle of a statefile
14:38 drawsmcgraw1 Kind of a "Surprise!!!" for anyone who has to read it later.
14:38 iggy Norrland: can you paste the files you're working with?
14:39 big_area hey man
14:39 big_area you up?>
14:39 big_area s/*/ /
14:39 Norrland iggy: sure, 2sec
14:39 Brew joined #salt
14:40 favadi left #salt
14:40 ageorgop joined #salt
14:40 kaptk2 joined #salt
14:41 cowpunk21 joined #salt
14:42 smcquay joined #salt
14:42 andrew_v joined #salt
14:42 Norrland iggy: http://589806150d3333aa.paste.se/
14:42 c10b10 joined #salt
14:44 bhosmer joined #salt
14:46 kiorky is_null: ive got, some client issues, sorry
14:46 kiorky is_null: ill ping you, anyway whenever i ve something tstable, no worry
14:47 favadi joined #salt
14:47 favadi left #salt
14:48 ageorgop joined #salt
14:50 drawsmcgraw1 PR made: https://github.com/saltstack/salt/pull/23974
14:50 mage_ is there a good practice to avoid storing password in sls files ?
14:50 giantlock joined #salt
14:51 Ph-x joined #salt
14:51 conan_the_destro joined #salt
14:52 iggy mage_: pillars can be targeted (so not just any minion can see the data), if that's not good enough, there is the gpg renderer
14:52 iggy or an ext_pillar that talks to Vault (still in development)
14:52 VR-Jack I'd use both, since you probably backup the pillar
14:52 bhosmer joined #salt
14:53 mage_ ah, I'll take a look at the gpg renderer, thanks
14:54 Norrland iggy: any suggestions?
14:54 iggy Norrland: what version of salt?
14:55 murrdoc joined #salt
14:55 Norrland iggy: salt-master                    2015.5.0+ds-1~bpo8+1
14:55 iggy minion
14:56 peters-tx joined #salt
14:56 Norrland salt-minion                    2015.5.0+ds-1~bpo8+1
14:56 iggy hmm, then I got nothing
14:56 Norrland debian jessie, using the saltstack repos
14:57 iggy if it was older, I was going to say they didn't have jinja support for that file
14:57 supersheep joined #salt
14:58 Norrland read somewhere that it wasn't sure if debconf.set_file used file.managed or not.
15:02 iggy effectively
15:03 druonysus joined #salt
15:03 druonysus joined #salt
15:04 murrdoc no need to read somewhere
15:04 murrdoc open the code for debconf state
15:04 murrdoc check out the funciton
15:08 Norrland https://github.com/saltstack/salt/issues/936
15:10 jalbretsen joined #salt
15:10 ageorgop joined #salt
15:12 hasues joined #salt
15:13 dopesong_ joined #salt
15:17 smcquay joined #salt
15:18 cpaclat joined #salt
15:21 Norrland iggy: if i set "template: jinja" instead of jinja2 I got around the error at least...
15:22 sdm24 joined #salt
15:22 bhosmer joined #salt
15:29 TheHelmsMan1 joined #salt
15:29 lothiraldan joined #salt
15:34 apergos joined #salt
15:38 Sathya_ rest_cherrypy: in salt configration is not working
15:38 Sathya_ the port is not listening
15:38 Sathya_ pls help
15:38 bradthurber joined #salt
15:42 otter768 joined #salt
15:43 c10b10 joined #salt
15:44 MatthewsFace joined #salt
15:46 impi joined #salt
15:47 rhodgin joined #salt
15:49 iggy Norrland: yeah, wonder how I missed that, where did you see template: jinja2?
15:50 Norrland http://docs.saltstack.com/en/latest/ref/states/all/salt.states.debconfmod.html#salt.states.debconfmod.set_file
15:51 Norrland perhaps the docs are wrong..
15:51 iggy very
15:51 iggy mind opening an issue on that?
15:53 Norrland iggy: sure, where do I go about doing that?
15:54 iggy https://github.com/saltstack/salt/issues
15:56 Norrland ah sure.
16:00 peters-tx I see salt.config reconfiguring salt like every minute; is this normal?
16:01 iggy peters-tx: gitfs backends?
16:01 peters-tx iggy, Not that I know of
16:01 iggy well, the base Salt time is 1 minute, so it could be a lot of things
16:01 iggy fileserver updates are the most common
16:02 iggy Norrland: https://github.com/saltstack/salt/pull/23980 don't worry about opening the issue
16:02 peters-tx iggy, Hmm, well I'm not aware of anything going and touching or updating stuff in Salt
16:03 Norrland iggy: ah. too late :)
16:03 Norrland https://github.com/saltstack/salt/issues/23979
16:04 dopesong joined #salt
16:09 neogenix joined #salt
16:11 dkrae1 joined #salt
16:11 kiorky is_null: ping
16:11 kiorky is_null: whats your githubid
16:13 codehotter How do I run a runner from an orchestration? Am I doing something strange?
16:14 codehotter basically, in the examples, you always have to give a target, but I just want to execute this part of the orchestration on the master itself, like I can do with a runner
16:15 desposo joined #salt
16:17 neogenix_ joined #salt
16:19 impi joined #salt
16:21 lothiraldan joined #salt
16:22 spookah joined #salt
16:24 saifi joined #salt
16:24 smcquay joined #salt
16:25 bhosmer joined #salt
16:26 HarryUP joined #salt
16:26 writtenoff joined #salt
16:28 Ahlee any reports of mongo return not working in 2015.5? Just upgraded a test env and it looks like site-packages/pymongo/mongo_client.py fails to connect
16:31 zhiwliu_ joined #salt
16:32 markm joined #salt
16:32 Pixionus joined #salt
16:33 zhiwliu__ joined #salt
16:34 aparsons joined #salt
16:35 nicolerenee_ joined #salt
16:35 ALLmightySPIFF joined #salt
16:38 KyleG joined #salt
16:38 KyleG joined #salt
16:39 zhiwliu_ joined #salt
16:39 zhiwliu__ joined #salt
16:42 zhiwliu_ joined #salt
16:43 toastedpenguin joined #salt
16:44 hasues left #salt
16:46 codehotter is_null: how do I use orchestration to call cloud functions?
16:47 codehotter is_null: I have my create_image function working with salt-cloud -f or salt-run cloud.action, but I don't know how to use that from orchestrate
16:49 zhiwliu__ joined #salt
16:53 zhiwliu joined #salt
16:56 iggy that's kind of not the point of orchestration
16:57 ashp joined #salt
16:58 ashp Hey guys, quick question - I'm currently an ansible user (because it makes life easy being masterless in aws) but I'm running into issues/limitations while building out automated testing environments.  Is there a framework in the salt community that is commonly used for things like "run salt, then make various assertions about the state of the box."  I'm thinking of things like "beaker/serverspec" in the puppet
16:58 ashp world.
16:58 tiadobatima joined #salt
17:00 ksj hi, I'm unable to run a state.highstate, but can run individual states fine. It was all working fine, but since I created a clean set of vms, I'm getting the following weird debug output on the minion http://dpaste.com/3YKPEM3.txt and the minions fail to return. any ideas?
17:00 Tahm joined #salt
17:01 theologian joined #salt
17:03 forrest joined #salt
17:04 ksj it seems to be related to LazyLoad....maybe it's a versioning thing
17:07 ksj nope...master and minions all on 2015.5
17:07 iggy ashp: you should read pretty much everything on http://ryandlane.com/blog/ and check all of his videos from saltconf
17:08 kiorky is_null: im on my way to push saltcloud support
17:10 moderation joined #salt
17:11 codehotter iggy: OK, so I have a series of steps, first I need to configure a server, then I need to create an image. How do I capture these steps in something that I can run?
17:11 codehotter iggy: I now have a runner "create_image" that I added to cloud/clouds/ec2.py, it works when using salt-run or salt-cloud
17:11 codehotter but how do I specify more such steps?
17:11 ksj any ideas? or can anyone point me at info about lazyload? it's a new feature in 2015.5 right?
17:11 codehotter I just want to have a file somewhere that contains all the steps. Ideally not a bash script with "salt-run, salt-run, salt-run", but some kind of salt file?
17:12 iggy codehotter: you could... just target the create_image at the master...
17:12 codehotter I don't have a minion running on the master, is that OK?
17:13 iggy then that won't work
17:13 iggy ... and why not?
17:13 manfred http://docs.saltstack.com/en/latest/ref/clients/#runnerclient
17:13 manfred i just use the runnerclient
17:13 manfred python ftw
17:13 codehotter so an orchestration file is a file I can put multiple steps in, right?
17:14 codehotter should I make a python script that uses runnerclient, rather than trying to do orchestration?
17:14 manfred sure, but i do not believe it can do runners
17:14 codehotter Should I try to make it be able to do runners, or is that not the philosophy of orchestration?
17:14 codehotter I'm really new to salt and I don't really understand all the concepts yet
17:14 iggy it's not the philosophy
17:15 manfred overstate used to do runners, but they just stripped our overstates the other day
17:15 codehotter Basically, I have states I want to apply to servers, but sometimes I also have management tasks I want to automate
17:15 codehotter how do I capture those management tasks?
17:15 manfred http://docs.saltstack.com/en/latest/topics/tutorials/states_pt5.html#states-overstate
17:15 codehotter what's the 'best' or 'salt' way to do that?
17:15 manfred the way that works
17:15 ksj running an sls file directly gives me  'LazyLoaded highstate.output' , but highstate gives me  LazyLoaded no_return.output
17:15 ashp iggy: Perfect, I'll have a read :)
17:16 ashp I used salt once before but it was several years ago and I just tinkered, so catching up on the current capabilities is a lot of effort :)
17:16 codehotter iggy: Is automating management tasks not a niche orchestrate wants to fill? What's the philosophy behind orchestrate?
17:16 iggy do things on different hosts in order
17:17 codehotter Yea! So if I have a database and I need to add it to the backup, then first I do something on the database server, then I do something on the backup server
17:17 codehotter that's something I can use orchestrate for, right?
17:17 iggy I do
17:17 manfred yes
17:17 iggy (postgres freeze on db servers -> snapshot volumes on master -> postgres unfreeze
17:18 codehotter but, now, let's say I want to migrate a database from one server to another. That's not really a 'state', it's more a 'task' right? It's a run once thing~ so, first, I want the loadbalancer to hold requests, then I migrate the database, then I tell the loadbalancer to continue requests. If I want to minimize user downtime, this should happen automated. I can capture all these individual tasks as salt, perhaps
17:18 codehotter as states perhaps as runners, but how do I capture the overall 'migration'?
17:19 JDiPierro joined #salt
17:19 iggy you still haven't mentioned a case that requires a runner/master context
17:19 manfred i mean, all those could be made into states.
17:19 manfred on minions, and could use file_recv, to cp.push the files to the master, and then later pull them down to the minion using minionfs
17:19 dkrae joined #salt
17:19 codehotter OK, currently I'm working on creating an image from a server. I want to deploy centos.org standard image, run some states on it, then take the image.
17:20 codehotter that's something I'll be doing more often, how do I capture that in salt?
17:20 manfred https://github.com/gtmanfred/servermill/blob/master/salt/mysql/dump.sls#L9
17:20 codehotter There's no state "create image"
17:20 gladiatr joined #salt
17:20 manfred https://github.com/gtmanfred/servermill/blob/master/salt/mysql/replication.sls#L13
17:20 manfred codehotter:  is there any reason you couldn't have a minion on your master to do those states?
17:20 codehotter no, I can make am inion on my masdter, if that's the way to do it
17:20 irctc341 joined #salt
17:20 codehotter but it's notn a state, creating an image is not a state, is it?
17:21 iggy put a minion on the master, tell the master to take the image
17:21 manfred it could be?
17:21 ksj ok, I think I figured it out - I had a duplicate line in my top.sls...no idea why it worked before, but seems to be going well now
17:21 manfred useing the cloud.py state module
17:21 codehotter manfred: yea, but conceptually, it's not a state, is it? The state "this image exists and was created from this source instance" - it's not something that's continually true. The source instance might not even exist anymore.
17:21 manfred sure it is
17:21 iggy ksj: if you think that's something someone else could run into, file a bug
17:22 manfred making sure an image of cloud-image-%{date}
17:22 manfred exists
17:22 codehotter and if it doesn't exist? What then? You can't recreate it, because the source instance is long gone
17:22 iggy or just run it as an execution module
17:22 codehotter it doesn't make sense as a state
17:22 codehotter yea! Execution module makes more sense, right?
17:22 manfred module.run state
17:22 codehotter What's the difference between a runner and an execution module?
17:22 manfred runner runs on the master
17:23 manfred execution module rusn on the minion
17:23 iggy a runner is in the master context, a module runs in a minion
17:23 codehotter other than that it's the same thing?
17:23 codehotter Can I orchestrate execution modules?
17:23 manfred it could be considered similar
17:23 iggy yes
17:23 manfred yes
17:23 manfred salt.function:
17:24 manfred orchestrate does salt.sls, and salt.function, sls does individual states, or a highstate, and function does execution moduels
17:24 codehotter so, there's a cloud.action execution module like there's a runner? Currently I do salt-run cloud.action create_image blablabla
17:24 impi joined #salt
17:24 manfred yes
17:24 manfred same thing
17:25 codehotter OK, how do I write that? Sorry to need so much handholding. I've been struggling with this for 3 hours. I install minion on my master, target my master for salt_function, use cloud.action as the name? Where do I put "create_image" as the cloud action to take?
17:26 UtahDave joined #salt
17:27 primechuck joined #salt
17:27 jespada joined #salt
17:27 ALLmightySPIFF joined #salt
17:27 irctc341 hi guys... Question about formulas from a newbie here: the doc implies we should install each formula in it's own directory under /srv/formulas, and have each one of these formulas "registered" in the config file... Is there a better way of doing this? It's not possible to have all the formulas under /srv/salt like any other state?
17:27 irctc341 http://docs.saltstack.com/en/latest/topics/development/conventions/formulas.html#adding-a-formula-directory-manually
17:30 forrest irctc341: You can just put then in /srv/salt if you want, or modify the file roots to just look for /srv/formulas. Then you just do the referencing like apache-formula.fcgid or whatever.
17:30 dfinn joined #salt
17:30 catpig joined #salt
17:31 iggy just hope the formulas you use don't have anything extra
17:31 iggy i.e. every script/one-liner I've seen people use to manage formulas like that doesn't take into account _* dirs in the formula
17:32 forrest rightg
17:32 c10b10 joined #salt
17:32 forrest ahh, so looking at the apache formula in actuallity, it would have to be apache-formula.apache.fcgid ;)
17:32 irctc341 Hi iggy... what do you mean by "anything extra" :)
17:33 forrest I imagine that's why the docs were written like that irctc341, just so people didn't have an extra directory to reference.
17:35 irctc341 forrest: I see what you mean.
17:35 soren joined #salt
17:39 cpaclat seem to be having an issue with syndicates and executing as the non-root user
17:39 irctc341 we'll run salt standalone, and plan to have each different teams handling their own stuff, but we want the base salt configs to be fairly static if possible... so from what you said I think just adding /srv/formulas then reference each formula as apache-formula.apache.fcgid would be the way to go, right?
17:40 cpaclat when I do salt '*' test.ping  I get the full hierarchy of minions
17:40 cpaclat when I do.  salt -a pam --username=saltdev --password=saltdev '*' test.ping
17:40 cpaclat I only get the directly connected ones
17:40 forrest irctc341: That's how I'd do it if you want content to be static (after you fork the repo obviously).
17:41 conan_the_destro joined #salt
17:41 jasonrm joined #salt
17:42 cpaclat offending line in master.py seems to be     elif clear_load.pop('key') != self.key[self.opts.get('user', 'root')]:                     log.warning(                         'Authentication failure of type "user" occurred.'                     )
17:43 druonysus joined #salt
17:43 otter768 joined #salt
17:44 debian112 joined #salt
17:45 Setsuna666_ joined #salt
17:46 cpaclat anyone?
17:46 cpaclat not much response on salt-users
17:46 cpaclat syndicates seem to be a bit on the not so popular side?
17:47 fishdust joined #salt
17:47 irctc341 Awesome! Thanks forrest and iggy
17:47 forrest np
17:48 iggy cpaclat: they fulfill a function that not many people need
17:48 cpaclat i think they are pretty cool as well
17:48 cpaclat just trying to figure out some of the subtleties
17:48 Berty_ joined #salt
17:49 cpaclat seems like there is an issue there with identity as you cross boundaries of masters
17:49 cpaclat ?
17:49 ajw0100 joined #salt
17:50 cromark joined #salt
17:50 iggy irctc341: if you use a formula with _modules/_states/_grains/etc you'll have to manually copy them to $file_roots/_\1 (f.ex. the postgres formula has _modules and _states and will fail to function without those modules/functions)
17:51 codehotter I installed a minion on the master and now I can do this: salt 'mgmt' cloud.action create_image blablabla. But from orchestrate, it still doesn't work. I now get "no minions match the target" regardless of what I enter as tgt
17:52 ALLmightySPIFF joined #salt
17:52 kiorky is_null: https://github.com/saltstack/salt/pull/23982/files
17:52 kiorky is_null: fully implemented, enjoy.
17:53 kiorky but it's for develop, im afraid this wont hit 2015.5.
17:53 kiorky this is a way too large change.
17:53 cpaclat does anyone have any suggestions on where to look for informaation on managing these identities
17:55 ekristen joined #salt
17:56 tomh- joined #salt
17:57 clintberry joined #salt
17:57 MK_FG joined #salt
17:57 ALLmightySPIFF joined #salt
18:00 JayFK joined #salt
18:02 irctc341 hrmmm... interesting iggy... this is not that nice. I hope there aren't many formulas like that... Am I wrong in thinking that requiring the config file to be changed for every formula we add is weird way of dealing with this?
18:02 iggy we do it
18:03 iggy I guess it depends how many formulas you're planning on using
18:04 abe_music joined #salt
18:04 wiqd joined #salt
18:04 iggy (we actually use pillar data to get the salt-formula to setup our gitfs remotes with formulas)
18:05 forrest wow that is a huge change kiorky, did someone already write tests for all the lxc stuff?
18:05 jhauser joined #salt
18:06 dopesong_ joined #salt
18:08 bootstrappm joined #salt
18:08 bootstrappm hello all! About to transfer a production server from AWS to DigitalOcean and using the opportunity to deploy it with salt
18:09 bootstrappm I've used Salt w/ Vagrant for my local dev machines, what do I need to know about how doing this on DigitalOcean is different from doing it in a vagrant?
18:09 irctc341 yeah.... I guess iggy... ideally we'd like to people not to change the config, as it can get out of hand quickly. But worst case, if a formula doesn't work we can write our own state (or formula). Thx!
18:09 iggy bootstrappm: it doesn't have to be any different... but it can
18:10 bootstrappm okay ...
18:10 codehotter iggy: manfred: I got it to work! Thanks for your pointers!!
18:10 codehotter iggy: Wouldn'
18:10 forrest bootstrappm: Not much different probably, but since we don't know what your setup looks like we won't be able to provide a lot of guidance there.
18:10 codehotter it be useful to be able to do runners from orchestrate though? If I make a pull request will that be rejected?
18:11 codehotter If overstate had it and orchestrate doesn't, maybe it was removed for a reason
18:11 murrdoc this is weird
18:11 murrdoc after salt runs highstate
18:11 iggy bootstrappm: vagrant uses what's called "masterless", you can continue using that (some large companies rely solely on masterless)
18:11 murrdoc vm cant ping no more
18:11 bootstrappm just an apache w/ a PHP Laravel app forrest. Source code in a bitbucket private git, previous deploy script in shell script
18:11 linjan joined #salt
18:11 iggy codehotter: orchestrate is a runner... it could be an issue of calling a runner from a runner
18:11 murrdoc oh for fucks sake
18:12 iggy codehotter: it sucks, but try and find out
18:12 Ahlee i love it when you talk dirty to me murrdoc
18:12 irctc341 now a question about pillars for formulas: in the case I use forrest's solution, eg /srv/formulas in FILE_ROOTS and reference the formula as in apache-formula.apache.fcgid... do the pillars for the formula also have to be apache-formula.apache.fcgid? or apache.fcgi would work?
18:12 iggy irctc341: apache.fcgi should work
18:13 forrest irctc341: Assuming you are putting them in /srv/pillar, your second would work.
18:13 denys joined #salt
18:13 forrest bootstrappm: Shouldn't be too problematic then. I believe digitalocean's API is pretty good now, so you should be able to provision no problem with them if you want. I'd say just spin an instance and start working on it, they cost almost nothing.
18:13 iggy irctc341: you should test this a little before you get too far down this road... every formula I've seen is going to fail miserably doing it that way (includes are almost always absolute not relative)
18:13 kiorky forrest: nope, but frankly, this will be hard to test.
18:13 codehotter iggy: if I get it to work, is that something that other people would find useful ?
18:13 codehotter so I know to waste time implementing it or not
18:13 bootstrappm thanks forrest
18:13 younqcass joined #salt
18:13 iggy codehotter: me, no... can't speak for anyone other than me though
18:13 forrest bootstrappm: np
18:13 kiorky forrest: but at least, it is "real" battle tested now.
18:14 fxhp joined #salt
18:14 forrest I guess.
18:14 forrest has anyone been having issues running bundle with salt in 2015?
18:15 clintberry joined #salt
18:15 forrest It fails with some missing libs when run with salt, but completes without issue afterwards, and this worked fine on previous releases
18:15 forrest wait iggy didn't you say there was a cmd.run bug?
18:15 irctc341 interesting iggy... this is not what I expected. I guess I need to do more reading then :D
18:15 iggy forrest: the default python_shell changed
18:16 iggy and some of the env handling changed
18:16 julez joined #salt
18:16 iggy (i.e. I had to add env: HOME to all of the aptly stuff)
18:17 dharper_ii Hi all. Has anyone seen on Windows Server 2012 when running "salt-call state.highstate", that right after 'Re-using SAuth', it seems like salt is internally issuing a Ctrl-c?
18:18 irctc341 roger that iggy!
18:18 forrest iggy: Hmm, it's pretty simplistic, just a cmd.run with bundle install; bundle --binstubs, and a cwd value.
18:18 dharper_ii it is weird. The highstate just stops. with the output: "Exiting gracefully on Ctrl-c"
18:18 forrest and it runs as a specific user
18:18 forrest but I don't see how that would negatively be impacted according to the release notes
18:19 forrest am I missing something iggy?
18:19 iggy I hit other issues than what were mentioned in the release notes... don't know if that's because I pushed out 2015.2 before it was ready or if those issues still exist
18:20 forrest Yeah this is problematic, because it claims it can't find postgres config stuff.
18:20 forrest I'll modify it to the old style and see what happens.
18:21 iggy try explicitly setting env: HOME
18:21 iggy see if that fixes it
18:21 forrest iggy: Will do, trying the cmd_safe: False option first just to rule that out
18:22 forrest I'm honestly going to be a bit annoyed if such a simple command fails because of either of these things
18:22 iggy welcome to my world :/
18:22 forrest because that's going to break almost every implementation of salt anyone has.
18:23 tmclaugh[work] joined #salt
18:24 forrest Did you create an issue iggy?
18:24 iggy for cmd.run python_shell changes, yes
18:24 buxy left #salt
18:25 iggy not for the env thing (I guess I had already worked around it so it didn't annoy the piss out of me)
18:26 forrest the env thing (if it works) is total bullshit
18:26 dkrae joined #salt
18:26 iggy I never tested the cmd_safe: False thing (it was added in response to my bug report)
18:26 forrest why should I have to set that? Makes no sense.
18:26 forrest Yeah that's why I am testing it now
18:26 iggy they were fully intending to ship with cmd.run way more broken than it is currently
18:28 forrest cmd.run is essential and one of the best features of salt
18:28 forrest ugh
18:28 forrest so annoyed
18:28 forrest just wastes my time. Now I've got to go through every one of my repos, reprovision, and confirm that things work, if not modify them
18:29 iggy hopefully that setting fixes it
18:29 iggy actually, if it does, I'll probably enable it too ;)
18:29 forrest I'm provisioning now
18:30 forrest we'll see
18:31 impi joined #salt
18:32 forrest nope, still failed with with a pg_config/lib problem
18:32 forrest iggy: did you set the env: value to the home of the user?
18:32 iggy yes
18:32 forrest iggy: That you wanted the command run as
18:33 forrest Alright, I'll give that a shot now
18:33 iggy see aptly-formula for example
18:33 big_area joined #salt
18:35 forrest iggy: Thanks, provisioning now, we'll see what happens.
18:37 forrest iggy: Where is your issue specifically for cmd.run?
18:37 forrest Not catching it on a search for open issue with you as the author
18:38 bhosmer joined #salt
18:39 forrest EPEL connections have been failing a lot lately...
18:39 abe_music left #salt
18:41 tiadobatima joined #salt
18:41 spookah joined #salt
18:42 iggy closed
18:43 iggy https://github.com/saltstack/salt/issues/22052
18:43 wt joined #salt
18:43 toastedpenguin joined #salt
18:44 wt is it expected the 2015.2.0 minions should work with 2014.7.5 masters?
18:44 forrest wt: no
18:44 twork joined #salt
18:44 wt Is the converse expected to work?
18:44 wt inverse, I mean
18:45 soren joined #salt
18:45 mrbigglesworth joined #salt
18:46 forrest wt: nope
18:46 baweaver joined #salt
18:46 forrest wt: Occasionally you'll get lucky and it will work between minor releases, but it's pretty rare, especially when security issues are addressed between releases.
18:47 wt forrest, that's certainly been my experience.
18:47 forrest wt: That it worked between minor releases?
18:48 twork i think i've seen, but cannot currently find... a document that talks about having minions on non-routable IP addresses behind a (routable, presumably) firewall. i can think of ways to do that but if it's a Solved Problem<tm> i could use the url.
18:48 wt forrest, no, what you said
18:48 forrest gotcha
18:49 wt forrest, I was just wondering if something changed with 2015.2.0
18:49 forrest wt: Not as far as I know. I haven't tested it myself, but after so many releases where it didn't work, or something specific didn't work, I jus always say it doesn't work.
18:50 twork maybe the only thing i found (but didn't read at the time) was http://docs.saltstack.com/en/latest/topics/tutorials/firewall.html ...which obviously isn't what i'm talking about
18:50 RedundancyD joined #salt
18:51 forrest still failed iggy, this is bullshit
18:52 wt forrest, what's failing for you?
18:52 forrest local dev envs are bombing with bundle commands on the newest release.
18:53 forrest running bundle with cmd.run makes it fail with being unable to find some postgres stuff, but if you just run bundle manually it works fine
18:53 forrest It might be an order of operations issue that changed on this release though
18:55 iggy if/when you figure it out, I'm kind of curious now
18:56 forrest I'm thinking it might be order of operations, I just realized that for some reason the devel package is getting installed at the incorrect time
18:56 baweaver joined #salt
18:56 forrest I modified the top file to re-order how it's done
18:56 forrest looking back at my commit logs now to see what is up
18:56 irctc341 Hi iggy... out of curiosity, I've browser around over 30 the formulas that I expect to be the most popular, such as apache, php, mysql, haproxy, graphite, docker, nginx, etc... I found that just postgres has those _modules, _states directories
18:56 forrest Yeah looks like this one is my fault
18:57 forrest I moved something a while back, not sure how it provisioned fine then, we'll see if it works this time.
18:57 iggy irctc341: okay... so you're safe I guess
18:58 wt s3fs appears to be broken in the epel6 build
18:59 wt another upgrade surprise
19:00 forrest wt: Hmm, Ryan_Lane have you messed with s3fs on 2015?
19:01 druonysuse joined #salt
19:01 tiadobatima iggy: I really want to avoid getting the main config file for the standalone client changed everytime a new formula is added
19:01 tiadobatima thank you very much for your help
19:02 cromark joined #salt
19:02 tiadobatima btw... tiadobatima = irctc341 :)
19:03 cowpunk21 joined #salt
19:04 perfectsine joined #salt
19:04 wt forrest, Ryan_Lane: https://github.com/saltstack/salt/issues/23992
19:04 wt There is a backtrace this time.
19:04 mapu joined #salt
19:06 wt FWIW, the s3fs was broken in 2014.7.* in other ways. I was just tryting to see it it had been fixed.
19:07 _prime_ joined #salt
19:09 s_kunk joined #salt
19:10 s_kunk joined #salt
19:10 forrest wt: For sure, good to note
19:12 litwol joined #salt
19:12 litwol Hello
19:13 litwol How can i chain different minion selectors without having to suplicate state file inclusions for each?
19:13 litwol for example i ahve "- common" state included for my minions
19:13 litwol would i need to do
19:13 litwol prod: \n 'minion1': \n - common, 'minion2': \n -common ...
19:13 litwol etc
19:14 litwol or can i do
19:14 RedundancyD state.highstate from 2015.5.0 master to 2014.7.x minions = blowed up minions… glad I cought it after only 5 minons… update all minions = all happy
19:14 manfred you can do prod:\n 'minion*': \n -common
19:14 litwol 'minion1': \n 'minion2': \n 'minion3...': \n - common
19:14 litwol manfred: unfortunately my minion names are non-uniform.
19:14 litwol i'll be looking into fixing naming uniformity
19:15 litwol by the way, i dont have 'base' state in my top.sls
19:15 wt litwol, I have a common state that is just included on *
19:15 litwol i could add it and it'll fix my issue.
19:15 cberndt joined #salt
19:15 wt I then have other states that are applied based on (in my case) aws tags.
19:15 cedwards any salt or community devs around that could tell me if there is an example state that takes an arbitrary list of arguments?
19:15 jespada joined #salt
19:15 kent joined #salt
19:16 wt cedwards, you can look at cmd.wait maybe?
19:16 cedwards I can't figure out how to describe it in yaml properly
19:16 wt You could have an argument that is a list?
19:16 wt or a dict
19:16 wt cedwards, I assume you mean a custom state?
19:17 cedwards yeah. I'm working on a custom state and I'm getting stumped on describing my list of kwargs in yaml
19:17 litwol k. i went with adding base: env to my top.sls.. don't like it. i want explicit dev/qa/prod-only definition.
19:17 litwol will need to look into renaming my minions
19:18 seev I always put the env as part of the sub-domain namespace
19:18 seev ie, web01.prod.tld
19:18 cedwards wt: this is essentially what I'm trying to pass through - http://pastebin.com/2S1tZ02K
19:19 seev it makes targeting groups of servers and roles easy without a bunch of weird tricks
19:19 cedwards wt: but when I run the state it only appears to be setting the first list item and ignoring the rest
19:19 wt cedwards, that would be passed as keyword args, IIRC
19:19 litwol seev: that is what i'm doing now. i have a few minions named prior to this convention taking place.
19:20 seev I also have a 'custom hosts' portion of the top file that has one-off declarations
19:20 wt so it would call update(key="value", key1="value1", key2="value2")
19:20 cedwards wt: I think I'm on the right track. I'm using **kwargs as the input and then iterate over the k,v in .items()
19:20 c10b10 joined #salt
19:20 wt try logging kwargs.keys()
19:21 cedwards ok
19:21 MrThrax left #salt
19:21 wt However, I don't think that's quite what you want
19:21 wt I think you'd want to add explicit keyword args for required arguments.
19:21 wt use kwargs for optional args
19:22 wt What does your state ultimately do?
19:22 wt write a file?
19:22 cedwards that's the tricky part. what I'm trying to pass in is going to have arbitrary keys and values
19:22 wt something else?
19:23 cedwards i wrote an execution module that connects to our cmdb API to update values based on gathered grains
19:23 cedwards then i wrote a state on top of that that would allow me to define all the key:value i wanted to update and templatize it in jinja
19:23 wt That's like what I did with AWS. :)
19:24 cedwards so a team here could simply populate a state with teh list of attributes they wanted updated and the state calls the module and updates it.
19:24 wt So, are the keys all names of cmdb attributes?
19:24 cedwards the module part works fine taking in key=value pairs as arguments, i'm just stuck on the state now.
19:24 cedwards the keys should be names of cmdb attributes, yes.. but there are a lot and didn't want to try hard-coding all of them.
19:24 wt Should the attributes become grains?
19:25 cedwards the attributes are populated from the grains, ie; salt['grains.get']('kernelrelease') is the value for the 'kernel' cmdb attribute
19:25 wt okay, so it will issue some update command to your cmdb?
19:26 cedwards yes
19:26 wt ok, so you might want to consider a dict arg rather than having the attrs at the top level
19:26 rhodgin joined #salt
19:27 cedwards how would I describe that in proper yaml that Salt wont bark at?
19:27 murrdoc dickts for life!
19:27 murrdoc wait
19:27 murrdoc dicts!
19:28 wt cedwards, maybe this? https://gist.github.com/wt/10870ed8cf47db2aceef
19:28 wt you'd have an attrs arg that will be populated with the dict
19:28 sfoger joined #salt
19:29 wt cedwards, btw, take special note of the 4 space indents for the dict
19:29 wt that is important
19:29 wt Try it here: http://yaml-online-parser.appspot.com/
19:30 sfoger Does anyone have an example of how to setup an inotify beacon, both on the minion (the beacon itself), and on the master (the reactor - i.e. tag and reactor sls declaration)?
19:30 wt http://yaml-online-parser.appspot.com/?url=https%3A%2F%2Fgist.githubusercontent.com%2Fwt%2F10870ed8cf47db2aceef%2Fraw%2Ffb477a3f3d4bf9e55ee36424d2cc8c9364ed89d5%2Fgistfile1.yml
19:30 wt If you make them 2-space indents, you'll see a nice corner case of yaml
19:31 cedwards strange
19:31 cedwards i wonder if that's what tripped me up earlier
19:31 sfoger I've setup a beacon on a minion in /etc/salt/minion.d/beacons.conf, and in my /etc/salt/master I have a reactor section where I specify a tag of 'salt/beacon/*/inotify/*', but I never see an event come across on the event bus after I modify a file being 'watched' by inotify on the minion
19:31 wt That also will make sure that you can add other args without conflicting with your attrs.
19:32 wt cedwards, This corner case of YAML is simply awful.
19:32 wt I don't know anything better there.
19:32 wt s/there/though.
19:33 jeffspeff joined #salt
19:34 cedwards I think that's given me something to focus on. Hopefully I can wrap this up here soon.
19:34 cedwards appreciate the help
19:35 wt cedwards, NP. Good luck.
19:35 wt cedwards, I am going to remove the gist now.
19:35 cedwards supposed to present on this module to the dept tomorrow. hope i can get it done by then lol
19:42 markm joined #salt
19:43 ckao joined #salt
19:44 otter768 joined #salt
19:45 racooper joined #salt
19:48 Pixionus cp.push seems to be giving me a FALSE.  what am I doing wrong?
19:50 hybridpollo joined #salt
19:50 wt Pixionus, do you allow cp.push in your master config?
19:50 wt I think that has to be turned on .
19:53 Pixionus I had just set it to True
19:53 Pixionus Salt doesn't need to be restarted or some such does it?...
19:53 VR-Jack joined #salt
19:53 iggy the master has to be restarted when you make config file changes
19:54 Pixionus oh maybe there's a way to reload config without restarting.
19:54 bbhoss joined #salt
19:56 murrdoc reboot the server
19:58 iggy redeploy
19:59 ALLmight_ joined #salt
20:00 Pixionus noooo
20:01 Pixionus come on iggy, murrdoc, you're telling me I have to take the Master down to reload a modification to the config?
20:01 Pixionus Even if momentary, i don't know what all other folks are doing with it atm and this is production..
20:02 murrdoc i can go ahead and not tell u
20:02 murrdoc that
20:02 forrest iggy: Issue ended up being my fault, due to the way we have some stuff coded I moved things around, and the top file was catching some values I had set.
20:02 druonysus joined #salt
20:02 * Pixionus slumps in his chair
20:02 cpowell why doesn't the user.present state support the --no-user-group flag from useradd
20:03 Pixionus Thanks anyway guys
20:03 forrest Pixionus: https://github.com/saltstack/salt/issues/570 if you want to track it/submit a PR
20:03 Pixionus Thought I had used this on the master before without having to restart
20:03 slav0nic joined #salt
20:04 mrbigglesworth joined #salt
20:07 Pixionus hmm..  I thin it'll compress smaller than a MB.. I think Ill try and gzip it then cat | to file
20:08 TaiSHi joehh: are you still in charge of ubuntu packages?
20:12 sfoger if python-inotify isn't installed, inotify beacon will fail virtually silently and give no indication that this package must be installed for this beacon to work. The only error salt-minion -l debug throws is 'Could not LazyLoad inotify.beacon' which isn't terribly helpful.
20:13 wt Are wheel modules broken in 2015.5.0?
20:13 sfoger Looked at https://github.com/saltstack/salt/blob/develop/salt/beacons/inotify.py to determine pyinotify (provided by python-inotify) was required for this beacon to work
20:13 murrdoc its not in the docs ?
20:13 iggy sfoger: in general people don't need their logs clogged up with errro messages about stuff they don't use
20:16 markm_ joined #salt
20:16 sfoger in general, people don't run in debug mode, but when they do, they'd like to get more information to help resolve an issue
20:17 TaiSHi Also they wont run inotify beacon unless they want to
20:17 sfoger Is there a more verbose debug mode I should use to catch mode detail?
20:17 TaiSHi I agree that the error should at least include a comment that points that python-inotify isn't installed
20:17 iggy feel free to file a bug, I was speaking in generalities
20:19 wt apparently, opts is no longer set in the WheelClient constructor
20:20 giantlock joined #salt
20:21 c10b10 joined #salt
20:22 ajw0100 joined #salt
20:26 tmclaugh[work]_ joined #salt
20:33 jespada joined #salt
20:35 wt Why does a WheelClient have output when I set the opts['output'] = 'no_out'?
20:38 wt Also, the docs for the WheelClient are creating RunnerClient instances.
20:40 c10b10 joined #salt
20:40 GR8EV1L joined #salt
20:41 wt The there an API to just get the salt-master config from the default file instead of having to manually specify "/etc/salt/master" as the config file to salt.config.master_config?
20:42 wt kinda weird that I can't just do "opts = salt.config.master_config()"
20:43 jhauser joined #salt
20:43 baweaver joined #salt
20:44 JayFK joined #salt
20:45 florinandrei joined #salt
20:46 rsimpkins joined #salt
20:47 JayFK joined #salt
20:50 jhauser joined #salt
20:51 Pixionus obviously the cat-ing of a binary didn't work so had to jump through a few hoops to get my logs.  Lesson learned. Impliment rotate from the start.
20:51 litwol gosh pkg.installed is killing me.
20:52 litwol specially ebuild pkg module
20:52 tmclaugh[work]_ joined #salt
20:52 mrbigglesworth joined #salt
20:52 litwol *every* pkg.installed, unless explicitly defined with "refresh: False", will trigger repository refresh despite in documentation being defined as not default behavior
20:52 litwol ugh.
20:53 tmclaugh[work] joined #salt
20:55 iggy litwol: check for/file a bug
20:55 litwol iggy: i am not sure where i would do that.. ?
20:56 iggy https://github.com/saltstack/salt/issues
21:00 forrest litwol: Yeah I had to set a bunch of our stuff to skip_suggestions: True
21:00 forrest I don't like that refresh/search behavior
21:01 litwol forrest: were you able to disable refresh globally ?
21:01 forrest litwol: Nope.
21:01 litwol forrest: i dont know what minion config i can set to prevent pkg repo refresh globally
21:01 litwol ah darn
21:01 Berty_ joined #salt
21:02 tmclaugh[work]_ joined #salt
21:02 forrest Yeah just disable it, unless someone else knows of a way. I don't know why it does that
21:02 litwol forrest: by "just disable it" do you mean on per pkg.installed state, or more globally?
21:02 forrest per pkg.installed
21:02 litwol k
21:03 litwol yeah that's what i'm doing
21:03 litwol super annoying
21:03 forrest yep
21:03 litwol i don't know python well enough to dig into salt implementation to correct the behavior
21:03 litwol i did find where it happens..
21:03 litwol just dont know how to undo it
21:03 forrest I feel like it's been discussed. iggy do you remember?
21:03 rodrigo_ joined #salt
21:03 rodrigo_ howdy
21:03 litwol i discussed it :)
21:04 litwol relatively long time ago
21:04 litwol i just didnt submit bug report last time
21:04 forrest ahh
21:04 forrest should submit a bug repor
21:04 forrest *t
21:04 rodrigo_ I have a question on regarding pkg.installed using it in a state and trying to install specific version
21:04 rodrigo_ http://paste.debian.net/181277/
21:04 iggy it's been brought up a few times (specific to Gentoo as I don't see it happening in Debian)
21:05 rodrigo_ this is my statement, but it fails to install it. I just want to make sure my syntax is correct
21:05 hal58th_ joined #salt
21:05 iggy rodrigo_: is that what your pkg manager sees as the version?
21:06 hal58th_1 joined #salt
21:06 rodrigo_ yes
21:06 rodrigo_ if I do it on the minion directly it finds it
21:06 rodrigo_ ..
21:06 iggy at the very least, that's not going to work on RHEL or Debian
21:07 forrest I use skip_suggestions for centos
21:07 forrest which skips the lookup
21:07 forrest pretty sure the actual 'search every time' was removed from RHEL styles a while back
21:07 rodrigo_ if I run apt-get install cassandra=2.1.5 the package is found
21:07 neogenix forrest: forgot to mention this morning, found the bug and put in the PR to fix it (re: last night's note)
21:08 forrest neogenix: Ahh cool.
21:08 neogenix forrest: #23950, if you're interested.
21:08 rodrigo_ nvm it was something from my end nothing to do with salt
21:08 rodrigo_ sorry about that
21:08 rodrigo_ :P
21:09 arknix joined #salt
21:10 amcorreia joined #salt
21:11 litwol oh wait a second
21:12 litwol maybe... maybe pkg.installed defaults to ebuild.latest_version instead of something alternative?
21:12 markm_ joined #salt
21:12 tmclaugh[work]_ joined #salt
21:13 rideh joined #salt
21:16 litwol will test before i submit issue. need to make sure refresh auto-triggers when only package name without version number is specified.
21:18 clintberry joined #salt
21:20 Ahlee is there a way to load _grains after saltutil.sync_grains short of restarting the minion?
21:21 iggy they should
21:21 iggy automatically
21:25 dimeshake joined #salt
21:26 Tyrm joined #salt
21:28 arknix joined #salt
21:28 mrbigglesworth joined #salt
21:29 meeteOrite joined #salt
21:29 Ahlee helps to return the dictionary you built
21:30 meeteOrite I am looking at implementing a returner to push data to a remote system however I am not finding any thing that can push data to a web service.  Is there a basic http returner that can push the job data in JSON format to a URI?
21:32 baweaver joined #salt
21:32 Ahlee I don't see a stock one, but shouldn't be hard to implement
21:33 meeteOrite Ahlee: It's super simple to do with a custom returner, however I didn't want to reinvent the wheel.  Seems like something that SS should do out of the box
21:34 Ahlee yeah, probably is
21:34 Ahlee other than how do you configure the return value, i.e. do you just dump everything to endpoint_uri
21:38 jalaziz joined #salt
21:39 cberndt joined #salt
21:40 ajw0100 joined #salt
21:41 lumtnman joined #salt
21:45 otter768 joined #salt
21:47 dendazen joined #salt
21:49 enarciso joined #salt
21:49 Setsuna666_ joined #salt
21:49 cromark joined #salt
21:53 Tyrm joined #salt
21:54 julez joined #salt
21:58 supersheep joined #salt
22:02 markm_ joined #salt
22:04 c10b10 joined #salt
22:08 preachermanx p
22:09 conan_the_destro joined #salt
22:09 murrdoc r
22:12 iggy you forgot the q
22:14 gyre007 joined #salt
22:15 druonysus joined #salt
22:15 druonysus joined #salt
22:16 sine_nitore joined #salt
22:17 c10b10 joined #salt
22:17 cromark_ joined #salt
22:19 cromark__ joined #salt
22:19 keimlink joined #salt
22:20 sinenitore joined #salt
22:21 c10b10 joined #salt
22:21 c10b10 joined #salt
22:22 baweaver joined #salt
22:22 shaggy_surfer joined #salt
22:26 viq joined #salt
22:26 c10b10 joined #salt
22:30 mrbigglesworth joined #salt
22:31 lionel joined #salt
22:35 spookah joined #salt
22:37 hasues joined #salt
22:37 hasues left #salt
22:38 elfixit joined #salt
22:41 druonysuse joined #salt
22:44 jespada joined #salt
22:45 shaggy_surfer left #salt
22:45 cromark joined #salt
22:52 nethershaw joined #salt
22:54 simonmcc joined #salt
22:57 andrej is there an elegant salty way to list all minions who have e.g. apache installed?  Short of pkg.get_selections or cmd.run "dpkg -l apache2", the output of which isn't particularly parse-friendly
22:57 lionel joined #salt
22:58 mosen joined #salt
22:58 murrdoc not that i know off
22:59 mosen hiya murrdoc
22:59 mosen hey salties is there any public formula that uses Salt Mine to form a cluster?
22:59 murrdoc sup mosen
22:59 murrdoc i think the hostfile formula does
23:00 mosen ahh yeah i see its got salt mine
23:01 baweaver joined #salt
23:02 ALLmightySPIFF joined #salt
23:03 iggy andrej: how did you install apache (i.e. if grains.get roles:web = install apache, use that same lookup)
23:04 iggy mosen: you can use the mine in pillars (which is what I expect most formulas kind of expect if you want to do that)
23:04 mosen iggy: ahh ok.. that makes more sense
23:05 murrdoc andrej:  you could do salt '*' aptpkg.list_pkgs
23:05 rhodgin joined #salt
23:06 murrdoc or
23:06 iggy aptpkg.version and check for a value
23:06 murrdoc salt '*' pkg.version apache2
23:06 iggy #ssceapproved
23:08 sergutie joined #salt
23:11 andrej iggy, no, apache was manually installed on these machines before salt came along...
23:11 murrdoc andrej:  write a grain
23:12 andrej it might make sense to define a state and retro-fit it
23:12 andrej murrdoc, how do you mean? List all machines that have apache installed and add a custom grain to them?
23:12 murrdoc u can write a python script
23:13 murrdoc that can run and figure out if apache was installed
23:13 andrej Heh. I'll have to do a lot of reading for that to happen :}
23:13 mosen I have no idea what im doing
23:13 andrej But yes, I could
23:13 murrdoc in your salt code
23:14 murrdoc under the file root make a _state directory
23:14 Aidan_ joined #salt
23:14 murrdoc put something like this in it
23:14 murrdoc https://github.com/saltstack/salt-contrib/blob/master/grains/has_battery.py
23:14 murrdoc change the has_batter to has_apache
23:14 murrdoc result = salt.modules.cmdmod._run_quiet('acpi -b')
23:14 murrdoc change that to a command that will tell u true/false
23:14 andrej cheers, I'll look into that straight away
23:14 murrdoc of whether the server has apache installed
23:15 murrdoc once u write that fiel
23:15 murrdoc then do a salt '*' saltutil.sync_all
23:15 murrdoc and a salt '*' saltutil.sync_grains ?
23:15 murrdoc or clear_cache
23:15 murrdoc then u can do salt -G 'has_apache:1' test.ping
23:16 murrdoc and that will give u all servers with apache installed
23:16 andrej sweet
23:16 scbunn joined #salt
23:17 iggy _grains (not _states)
23:17 andrej is _grains something with a special meaning, or just your preference?
23:18 pppingme joined #salt
23:18 nethershaw joined #salt
23:20 rwaterbury joined #salt
23:28 tr_h left #salt
23:28 cberndt joined #salt
23:32 ajw0100 joined #salt
23:33 druonysus joined #salt
23:34 keimlink_ joined #salt
23:35 hasues joined #salt
23:37 murrdoc it tells salt to sync the dirctory over
23:37 murrdoc and it tells salt that its a custom grain
23:37 murrdoc and the lazyloader will look for grains stuff there
23:41 zz_ashmckenzie joined #salt
23:43 mosen iggy: the salt mine doco explains salt mine using a template to run mine.get but would you say that running mine.get in the pillar is a better method of organisation/abstraction?
23:44 iggy jinja works in pillars too
23:44 scbunn joined #salt
23:44 iggy so the same kind of examples you see in states also work in pillars
23:45 mosen so everything gets rendered with jinja?
23:45 iggy by default
23:45 mosen heh, I always thought pillars were an exception somehow
23:45 iggy you just can't do {{ pillar['foo'] }} in pillars
23:46 iggy pillarception
23:46 otter768 joined #salt
23:47 mosen can you run salt modules in pillars? :)
23:47 mosen not that i have a use case
23:47 iggy yeah
23:47 iggy one other thing to remember is that pillars are rendered on the master
23:48 mosen ahh right, so that would be pretty horrible
23:48 iggy well it depends
23:48 iggy instead of using salt environments, we have separate masters for each "env"
23:48 iggy and they have a grain set based on GCE project metadata
23:49 iggy so {{ grains['ocsenv'] }} is going to be the same on everything in that env/project
23:49 rwaterbury Hi, I am seriously ignorant as far as salt goes.  And I need to create an api request to our salt server to remove a node.  I have some code in our app that already logs in a retreives a token and then runs some commands on a server.  So I have that as a blueprint, but the api docs are a little confusing to me and I am not sure how to get a list of nodes regardless of removing one of them from Salt’s set of administrable machines.
23:49 iggy so that's safe for us to reference
23:49 mosen I see
23:49 iggy but some things are not obviously
23:49 mosen yep
23:50 iggy rwaterbury: you just want the master to stop managing the minion? or you actually want to get rid of the minion (i.e. destroy the instance)?
23:51 rwaterbury I just want it to remove it’s idea of the machine.
23:51 rwaterbury The machine is terminated seperately.
23:51 iggy use the key wheel module
23:52 scbunn joined #salt
23:52 kusams joined #salt
23:53 cowpunk21 joined #salt
23:54 rwaterbury We have Tornado installed on the Salt server… if that helps
23:55 rwaterbury Like I said, I know almost nothing.
23:55 murrdoc the api lets u run any module
23:55 rwaterbury Where you responding to me when you mentioned the key wheel module?
23:55 jespada joined #salt
23:55 murrdoc he wuz
23:55 rwaterbury ah, ok
23:56 murrdoc http://docs.saltstack.com/en/latest/ref/cli/salt-api.html
23:57 rwaterbury Ah, thanks :)
23:57 iggy yeah, I'm terrible about targeting my replies
23:57 murrdoc actually http://docs.saltstack.com/en/latest/ref/netapi/all/index.html#all-netapi-modules
23:57 gazarsgo joined #salt
23:58 * baweaver grabs popcorn
23:58 rwaterbury ok, I’ll poke around a bit.  Thanks so much!
23:59 * baweaver wrote aforementioned connection code
23:59 murrdoc the doc has some noobage
23:59 murrdoc alias of <Mock object at 0x116e42910>
23:59 murrdoc someone didnt read rst
23:59 baweaver Yeah, kinda fun getting through some of that when I made the connection gem

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