Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-01-26

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

All times shown according to UTC.

Time Nick Message
00:00 JDiPierro joined #salt
00:04 jerematic joined #salt
00:05 GabLeRoux joined #salt
00:16 robinsmidsrod joined #salt
00:19 PovAddict in a cmd.run, is the 'unless' command executed in the same environment (cwd, user...) as the main command?
00:20 GabLeRoux joined #salt
00:22 viq joined #salt
00:30 bluenemo_ joined #salt
00:34 yetAnotherZero joined #salt
00:36 GabLeRoux joined #salt
00:43 cberndt joined #salt
00:44 johtso joined #salt
00:44 wolfpackmars2 joined #salt
00:56 glyf joined #salt
00:57 aqua^mac joined #salt
00:58 brendanashworth joined #salt
01:00 nettus joined #salt
01:01 signull joined #salt
01:09 laax joined #salt
01:12 brendanashworth left #salt
01:12 wolfpackmars2 is there a way to sync a directory from the master to the minion?  I'm using file.recurse, but it doesn't seem to propagate deletions to the minion
01:12 wolfpackmars2 if I delete a file from the master, the minion doesn't reflect the deletion
01:14 pdayton joined #salt
01:22 nettus joined #salt
01:28 bash1234123123 joined #salt
01:29 iggy wolfpackmars2: clean: True
01:29 wolfpackmars2 i'll try it
01:36 GabLeRoux joined #salt
01:39 AnswerGuy joined #salt
01:44 pdayton joined #salt
01:45 imaack joined #salt
01:45 bhosmer__ joined #salt
01:47 NightMonkey joined #salt
01:48 aldeb joined #salt
01:50 aldeb joined #salt
01:50 vbabiy joined #salt
01:51 out180 joined #salt
01:54 aqua^^ joined #salt
01:58 ec2-user_ joined #salt
02:01 ec2-user_ hi
02:02 ec2-user_ What's mco equivalent in salt?
02:07 whatevsz joined #salt
02:08 iggy ec2-user_: what's mco?
02:08 clubanderson joined #salt
02:11 ckao joined #salt
02:13 ajw0100 joined #salt
02:13 clubanderson_ joined #salt
02:16 clubanderson joined #salt
02:17 ec2-user_ iggy: Like fabric
02:17 Yoda-BZH joined #salt
02:17 Yoda-BZH joined #salt
02:17 GabLeRou_ joined #salt
02:17 clubanderson joined #salt
02:18 ec2-user_ iggy: parallel execution of commands, its a puppet thing.
02:19 pdayton joined #salt
02:19 favadi joined #salt
02:22 imaack joined #salt
02:22 imaack I'm having some trouble using a formula on a masterless minion... does this look Right:
02:22 imaack #top.sls:  file_roots:   base:     - /srv/salt/     - /srv/salt/formulas/apache-formula/  base:   ‘*’:     - webserver   #webserver.sls: include:   - apache  apache:   pkg.installed:[]
02:23 imaack With the output I'm getting it seems like salt is ignoring file_roots settings I declare.
02:25 irctc394 joined #salt
02:25 clubanderson joined #salt
02:25 imaack Here's a gist that should be more readable: https://gist.github.com/imaack/17192829f1656ff25ebd
02:38 bhosmer_ joined #salt
02:45 PovAddict ec2-user_: to answer "what's the equivalent of puppet's mco" we'd have to know about puppet :P
02:45 PovAddict explain what it does
02:47 ilbot3 joined #salt
02:47 Topic for #salt is now Welcome to #salt | SaltConf 2015 is Mar 3-5! http://saltconf.com | 2014.7.1 is the latest | 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/
02:48 ec2-user_ PovAddict: Well like check if httpd is running on 100 intsances and do it in parallel
02:49 elfixit joined #salt
02:50 dwfreed joined #salt
02:51 tkharju joined #salt
02:54 voxxit joined #salt
02:58 malinoff joined #salt
02:59 laax joined #salt
02:59 jonatas_oliveira joined #salt
03:02 bhosmer joined #salt
03:06 iggy ec2-user_: that's salt
03:12 MugginsM joined #salt
03:20 iggy ec2-user_: for example: salt '*' service.status httpd
03:24 bhosmer joined #salt
03:28 ec2-user_ iggy: I sse so its built in
03:30 Zachary_DuBois joined #salt
03:31 twellspring joined #salt
03:31 jalbretsen joined #salt
03:38 Zachary_DuBois joined #salt
03:42 iggy yes
03:42 iggy anything in salt.modules.* can be run the same way
03:43 iggy anything in salt.states.* is meant to be in sls files
03:46 ajw0100 joined #salt
03:51 twellspring joined #salt
03:53 pdayton joined #salt
03:59 urtokk joined #salt
04:09 TheThing joined #salt
04:14 brendanashworth joined #salt
04:14 brendanashworth left #salt
04:15 kermit joined #salt
04:21 pdayton joined #salt
04:35 jeddi joined #salt
04:48 laax joined #salt
04:48 jonatas_oliveira joined #salt
04:50 brendanashworth joined #salt
04:51 brendanashworth left #salt
04:56 Zachary_DuBois joined #salt
05:18 tkharju joined #salt
05:22 Furao joined #salt
05:36 aquinas joined #salt
05:37 felskrone joined #salt
05:51 TyrfingMjolnir joined #salt
05:56 ramteid joined #salt
06:03 I3olle joined #salt
06:06 ramteid joined #salt
06:18 Furao joined #salt
06:19 mikeywaites joined #salt
06:25 robawt joined #salt
06:26 calvinh joined #salt
06:36 mtanski joined #salt
06:37 laax joined #salt
06:37 linjan joined #salt
06:37 jonatas_oliveira joined #salt
06:48 jhauser joined #salt
06:49 s51itxsyc joined #salt
06:50 glyf joined #salt
06:59 stoogenmeyer__ joined #salt
06:59 colttt joined #salt
07:05 AndreasLutro joined #salt
07:05 ajw0100 joined #salt
07:16 mtanski joined #salt
07:18 flyboy joined #salt
07:18 otter768 joined #salt
07:22 sienna joined #salt
07:25 Grokzen joined #salt
07:27 aldeb_ joined #salt
07:29 mikkn joined #salt
07:33 bluenemo joined #salt
07:33 bluenemo joined #salt
07:37 monkey66 left #salt
07:37 JK_ joined #salt
07:37 JK_ hi all
07:38 Auroch joined #salt
07:39 JK_ Is it correct that the module state.sls_id does not call the watch states? From the doc: "Call a single ID from the named module(s) and handle all requisites"
07:51 laax joined #salt
07:53 mbersoult joined #salt
07:53 mbersoult Hello
07:53 seshan joined #salt
07:54 seshan_ joined #salt
07:56 slafs joined #salt
07:56 elfixit joined #salt
07:56 slafs left #salt
07:59 tomspur joined #salt
08:00 hebz0rl joined #salt
08:01 twellspring joined #salt
08:02 twellspring joined #salt
08:09 bash1234123123 joined #salt
08:09 twellspring joined #salt
08:14 ajw0100 joined #salt
08:18 kawa2014 joined #salt
08:19 dunz0r Can I make salt only output failed/changed states instead of all the states it has run?
08:23 esogas_` joined #salt
08:26 jonatas_oliveira joined #salt
08:30 dunz0r When I run a state.hightstate that is
08:30 KermitTheFragger joined #salt
08:38 I3olle joined #salt
08:38 calvinh joined #salt
08:45 dkrae joined #salt
08:45 karimb joined #salt
08:47 intellix joined #salt
08:48 lb1a joined #salt
08:49 xsteadfastx joined #salt
08:50 ptinkler joined #salt
08:51 Grokzen joined #salt
08:53 calvinh_ joined #salt
09:00 Grokzen joined #salt
09:00 laax joined #salt
09:01 lionel joined #salt
09:05 bhosmer__ joined #salt
09:05 cberndt joined #salt
09:05 ptinkler hey guys, is there anyway for a salt state to watch a local file? so for example with my current setup:  http://pastebin.com/iAFv4s9b
09:06 bash1234123123 joined #salt
09:07 ptinkler so /mnt/bootstrap/ is a folder shared from my host to a vagrant guest box that has salt on it, but I'd like the `mysite-reqs` state to watch the requirements file and run again if it changes (I'd be editing the file on the host)
09:08 trikke joined #salt
09:15 xsteadfastx joined #salt
09:16 xsteadfastx joined #salt
09:17 pipeep joined #salt
09:19 otter768 joined #salt
09:23 laax_ joined #salt
09:25 pcdummy ptinkler: i think you can add that file as "my-state-id: \n file: \n - name: /mnt/bootstrap/my-file"  "mysite-reqs: \n ..... - watch: - file: my-state-id" you get the idear?
09:25 jrluis joined #salt
09:26 forrest joined #salt
09:30 ptinkler ah ok, so I don't actually need the "- source: " bit to declare a file? Cos I tried that but as it wasn't hosted by salt or elsewhere I couldn't get it to work
09:35 pcdummy ptinkler: hmm think i'm wrong. :/
09:38 ptinkler " No function declared in state 'file' in SLS 'python.python-pip'" :)
09:38 pcdummy ptinkler: hmm https://github.com/saltstack/salt/issues/663 <-- there should be something like that.
09:39 ptinkler hum I thought I'd tried that as well, and it complained that the file doesn't exist
09:39 ptinkler I'll try it again
09:39 phx how can i get the pillar top.sls, what it renders to?
09:43 intellix joined #salt
09:47 ptinkler yeah pcdummy : this is what I get when I do it that way - http://pastebin.com/2KMgXC9K
09:47 ptinkler despite being sure that the file exists. maybe it hasn't been mounted at that point?
09:47 keyser joined #salt
09:52 bash1234123123 joined #salt
09:53 karimb joined #salt
09:55 TyrfingMjolnir joined #salt
09:56 bluenemo hi guys. I'm trying to set my PS1 in my .bashrc via salt. I'd like to use file.replace for that as I'd like to look where 'export PS1' is present and replace it. I cant figure out why my state is being executed twice before the regex is happy: http://paste.debian.net/142282/  when you compare line 61 and 62, at the end of it at the  \\$  there is only a  \$ in the second version. is there something I can do to not have those interpreted? And why
09:56 bluenemo is it interpreting those incorrectly only in the second run (after the string was not present before)?
09:56 aqua^mac joined #salt
09:57 viq pcdummy: ptinkler: salt can only watch states. You cannot watch for a file outside of salt's control
09:58 wincyj joined #salt
09:58 viq The best you could do is depend on file's existence (file.present), but that doesn't notice changes in file.
09:58 bluenemo pcdummy, you can use unless or only_if with bash commands like test -f
09:58 bluenemo or use file.present, yes
09:59 ptinkler so there's no way to create a state out of a local (or in this case mounted) file?
09:59 bluenemo what do you mean by create state?
09:59 viq ptinkler: no
09:59 bluenemo you can manage the mount via salt..
10:00 ptinkler in what way bluenemo?
10:00 bluenemo http://docs.saltstack.com/en/latest/ref/states/all/salt.states.mount.html
10:00 fe92 joined #salt
10:00 ptinkler oh I see, well I shouldn't need to do that as vagrant manages the mount for me
10:01 bluenemo but you can use it to depend on
10:01 bluenemo argh I cant figure out my regex.
10:01 ptinkler it wasn't so much "depends on" that I wanted, I wanted it to watch the requirements file and reinstall pip dependencies when requirements changed
10:03 MorbusIff joined #salt
10:03 ptinkler in a similar vein, I'd want the webserver to restart itself whenever the files changed
10:03 pcdummy ptinkler: think you need to copy that file and compare it with cmd.run and test
10:03 pcdummy copy with cmd.run and test with "unless: cmp <old> <new>"
10:04 pcdummy tought its easier, but  :/
10:05 ptinkler yeah it seems quite odd to me that there isn't a way to watch a specific file. i'd have thought that would be right up there
10:05 Rawkode joined #salt
10:05 bluenemo ptinkler, if you dont manage that file by salt, how should salt know?
10:06 ptinkler well that's the point: I'd like to manage the file by salt :) but there's no way to do it
10:06 ptinkler unless a file is served by salt:// or http[s]:// you can't manage it
10:07 bluenemo how do I disable all escapes for values I give to file.append or file.replace? I dont want it to interpret the string I give it to use as replacement
10:09 lothiraldan joined #salt
10:09 bluenemo when I use http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html (For complex regex patterns it can be useful to avoid the need for complex quoting and escape sequences by making use of YAML's multiline string syntax.) I get the same result as posted above
10:11 pcdummy ptinkler: untested: http://pastebin.com/62hpzieJ
10:14 gngsk joined #salt
10:15 jonatas_oliveira joined #salt
10:15 ptinkler thanks pcdummy , I'll give that a try in a bit
10:16 TyrfingMjolnir joined #salt
10:18 viq ptinkler: erm, and how otherwise would you manage a file?
10:18 viq regarding < ptinkler> unless a file is served by salt:// or http[s]:// you can't manage it
10:19 Mindfart joined #salt
10:20 ptinkler I guess I don't understand the restriction, why can't I do:-     "file.managed: \n   - source: /mnt/bootstrap/myfile.txt" etc? Surely salt can find that file and manage it, why do I have to  serve it through the salt master
10:20 viq ptinkler: because it's the master that serves files
10:22 ptinkler sure, and i "could" store the requirements.txt file on the salt master and get it from there. but the requirements are tied to the project code- so it's madness to store them outside of the project repository
10:23 ptinkler I get that "this is how it works" I just don't fully grasp WHY "this is how it works" :)
10:23 TyrfingMjolnir joined #salt
10:24 viq ptinkler: how about git.latest and watch that?
10:25 kim0 joined #salt
10:25 che-arne joined #salt
10:25 jonatas_oliveira joined #salt
10:25 kim0 Hi folks .. I want to give users the ability to run some repititive salt orchestration tasks .. Ideally through an easy web UI (but cli works too) .. Any recommendation how to do that
10:26 buxy left #salt
10:26 viq kim0: rundeck maybe ?
10:26 kim0 yeah .. looking at that .. just wondering if there's some other way (because java :)
10:27 kim0 is this within scope for halite ?
10:27 viq possibly, yeah, though they'd have to remember what to call - but they would need to as well with CLI
10:28 viq https://github.com/saltstack/pepper  maybe
10:28 kim0 viq: the main problem with just salt .. is it able to restrict users to running certain commands on certain nodes only ?
10:28 ralala joined #salt
10:30 viq kim0: well, there's http://docs.saltstack.com/en/latest/ref/clientacl.html and http://docs.saltstack.com/en/latest/topics/eauth/access_control.html
10:31 wnkz joined #salt
10:31 kim0 oh thanks! didn't know about that
10:33 bash1234123123 joined #salt
10:33 Gilou joined #salt
10:34 CeBe joined #salt
10:37 giantlock joined #salt
10:49 jtang joined #salt
10:52 zadock joined #salt
10:52 Grokzen joined #salt
10:54 mndo joined #salt
10:55 ptinkler viq: git.latest doesn't really help because while it is a git repository (on my local host) it's mounted to the vagrant guest where it technically isn't a git repo, and i wouldn't want it acting like one
10:56 viq ptinkler: how about making salt check it out elsewhere on host, and work on that?
10:57 viq mount it in vagrant in eg /vagrant/myrepo and have a salt state that checks it out from there to /src/myproject and works on that?
10:58 ptinkler that still means that I'd have to commit every change immediately for it to show up and be pulled down by git on the guest, which is a bit awful when what I'm trying to do is set up a dev environment that response quickly to changes :)
11:00 viq ptinkler: or mount the repo in /srv/salt/something
11:00 viq Or add it to file_roots
11:01 viq Work with salt.not around it ;)
11:02 ptinkler I could mount the entire project  repo in /srv/salt/
11:03 ptinkler the only issue is that the project repo contains /salt/root/.. etc directories (cos it's setup to be masterless right now)
11:03 ptinkler is that going to mess it up?
11:04 ptinkler and how responsive is the salt file server to changes? cos I'll be editing files on the host and wanting the change to show up immediately on the guest
11:05 viq If it's on file system, I think it's immediate. For gitfs the default refresh is 60 seconds
11:06 viq You can always do salt-run fileserver.update to be certain
11:08 ptinkler so basically, my vagrant file would have 'config.vm.synced_folder ".", "/srv/salt"' and any paths in my .sls files that are currently "salt://python/whatever.py" (f.ex) would become "salt://salt/roots/python/whatever.py" ?
11:11 ptinkler is that not going to confuse it when it looks for top.sls ?
11:11 ptinkler i'm not sure how smart it is at detecting where it is
11:12 istram joined #salt
11:12 viq if you have ./salt/roots/python/whatever.py then yes
11:12 viq and yes, it will
11:13 phx anyone tried to install pygit2 for gitfs on centos6? it needs a newer version of libgit2 than what's available in base/epel
11:13 TyrfingMjolnir joined #salt
11:14 calvinh joined #salt
11:15 lothiraldan joined #salt
11:16 support joined #salt
11:16 Guest27379 left #salt
11:20 otter768 joined #salt
11:25 seanz joined #salt
11:28 jtang joined #salt
11:35 Mindfart joined #salt
11:35 shiin joined #salt
11:36 calvinh joined #salt
11:36 shiin can anyone confirm salt-minion works on raspbian ?
11:40 istram joined #salt
11:45 aqua^mac joined #salt
11:46 toanju joined #salt
11:56 ecdhe joined #salt
11:58 hebz0rl joined #salt
11:59 bhosmer joined #salt
12:03 Furao joined #salt
12:03 hebz0rl joined #salt
12:08 jonatas_oliveira joined #salt
12:13 RobertFach joined #salt
12:20 lothiraldan joined #salt
12:25 bash1234_ joined #salt
12:29 mikkn joined #salt
12:30 bash1234123123 joined #salt
12:32 hobakill joined #salt
12:33 numkem joined #salt
12:42 xsteadfastx joined #salt
12:44 markm_ joined #salt
12:44 fredvd joined #salt
12:47 ptinkler what's the best way to say "restart service x" once a state completes?
12:47 ptinkler so my pip install state does its stuff, then after it's done I want to always restart the server, but the pip install state watches a file so it might run multiple times
12:49 viq ptinkler: watch the pip state in service state
12:49 viq or watch_in the service state in pip state, same difference
12:50 ptinkler viq: ah right, does that trigger when the pip install state starts or when it finishes though? or both
12:51 viq finishes
12:51 ptinkler ah, perfect :)
12:51 viq "when this state is done, step over to the one over there"
12:52 gfa joined #salt
12:52 __number5__ joined #salt
12:52 wm-bot4118 joined #salt
12:53 pviktori joined #salt
12:53 keekz joined #salt
12:53 bxz joined #salt
12:53 smkelly joined #salt
12:53 rome_390 joined #salt
12:54 hobakill what is the procedure for contributing to the development branch? i have a few nit-picky yet relevant documentation changes i'd like to see but I don't exactly know how to go about doing it. do i need to fork it then do a pull request?
12:55 viq hobakill: I believe so, yes
12:55 hobakill thanks viq
12:55 calvinh_ joined #salt
12:58 bash1234123123 joined #salt
13:03 xsteadfastx joined #salt
13:04 moderation joined #salt
13:08 giantlock joined #salt
13:16 bash1234123123 joined #salt
13:20 ughloru joined #salt
13:20 ughloru Hello! Is this a good place to ask a question about salt-cloud?
13:21 otter768 joined #salt
13:21 ughloru I want to configure regular snapshots of AWS EC2 instances, and it looks like salt-cloud might be a good way of doing this.
13:21 ughloru But I'm a bit confused by some of the documentation.
13:22 ughloru If I was connecting to AWS EC2 using the python module boto, or the command line tools, I could do so using an AWS secret access key, which is a string.
13:23 ughloru But salt-cloud seems to require a .pem file as well
13:24 ughloru Actually, never mind. I see it discussed in the documentation.
13:24 ughloru I just had to scroll down a little more.
13:24 ughloru :-/
13:25 babilen :D
13:26 bash1234123123 joined #salt
13:26 epcim joined #salt
13:32 fredvd joined #salt
13:33 linjan joined #salt
13:33 wiqd_ joined #salt
13:34 aqua^mac joined #salt
13:37 toanju joined #salt
13:41 kwmiebach joined #salt
13:43 `ph8 joined #salt
13:47 _prime_ joined #salt
13:49 mndo joined #salt
13:50 kawa2014 joined #salt
13:53 bash1234123123 joined #salt
13:57 twellspring joined #salt
13:59 karimb joined #salt
14:00 JDiPierro joined #salt
14:00 ptinkler pcdummy: trying out your cp script, but it doesn't seem to be working and I can't work out how to check if the cmd is actually running correctly
14:00 Cottser|away joined #salt
14:01 timoguin joined #salt
14:01 signull joined #salt
14:02 jonatas_oliveira joined #salt
14:03 signull joined #salt
14:04 intellix joined #salt
14:05 PepperIndustries joined #salt
14:06 numkem Has anyone worked on salt with coreos? I understand it's probably not necessary but at least for reporting it would be useful
14:08 seanz joined #salt
14:08 murrdoc joined #salt
14:09 GabLeRou_ joined #salt
14:09 gngsk joined #salt
14:10 jY numkem: if you package it up in a virtualenv and ship a tarball to it it might work
14:10 numkem jY: coreos doesn't have a python interpreter but it might be possible to "fake" it with a python inside a container
14:12 PepperIndustries joined #salt
14:13 GabLeRou_ joined #salt
14:17 cpowell joined #salt
14:19 jtang joined #salt
14:19 numkem watching things like this http://thenewstack.io/saltstacks-big-role-in-google-kubernetes-and-why-immutable-infrastructure-makes-the-cloud-a-giant-computer/ makes me wonder if something like coreos is really necessary.
14:20 numkem i have a big problem with the current version because you can't pull from a repo that has a self-signed certificate but beside that the only problem I have is service discovery/ha
14:21 jbub joined #salt
14:22 jeremyr joined #salt
14:23 favadi joined #salt
14:23 perfectsine joined #salt
14:24 Grokzen joined #salt
14:25 redzaku joined #salt
14:26 msheiny joined #salt
14:28 Guest27379 joined #salt
14:30 mpanetta joined #salt
14:31 TheoSLC joined #salt
14:33 elfixit joined #salt
14:36 z3uS joined #salt
14:37 patrek joined #salt
14:37 keeth joined #salt
14:39 Furao joined #salt
14:41 phx when using gitfs for pillars (ext_pillar), does it also treat branches/tags as environments?
14:44 Ahrotahntee phx: looks like it. http://docs.saltstack.com/en/latest/topics/tutorials/gitfs.html section 3.4.8.8
14:47 phx Ahrotahntee, if you check 3.4.8.12: Note that this is different from gitfs, and is not yet at feature parity with it.
14:50 shiin left #salt
14:51 jtang_ joined #salt
14:53 phx Git-based ext_pillar is enabled in configuration but could not be loaded, is GitPython installed?
14:53 phx mhm, shouldn't that work throught pygit2?
14:54 Ahrotahntee phx: sorry, my mistake
14:56 mdasilva joined #salt
14:56 ALLmightySPIFF joined #salt
14:57 rnts joined #salt
14:57 phx http://docs.saltstack.com/en/latest/ref/pillar/all/salt.pillar.git_pillar.html
14:57 phx ahh it works entirely differently
15:02 housl joined #salt
15:03 cpowell_ joined #salt
15:03 andrew_v joined #salt
15:03 jeddi joined #salt
15:03 TheoSLC joined #salt
15:05 krelo joined #salt
15:06 bhosmer_ joined #salt
15:06 Cidan joined #salt
15:08 karimb joined #salt
15:10 twellspring joined #salt
15:10 Deevolution Anyone see the salt scheduler not running jobs?  I have 2 nodes (2014.7.0; CentOS) with identical jobs in pillar.  One ran them, the other didn't.  Both are showing the jobs in a schedule.list.
15:15 _prime joined #salt
15:15 nmadhok joined #salt
15:19 kaptk2 joined #salt
15:21 dude051 joined #salt
15:22 otter768 joined #salt
15:23 aqua^mac joined #salt
15:24 johtso joined #salt
15:26 rojem joined #salt
15:30 bash1234123123 joined #salt
15:32 _JZ_ joined #salt
15:37 Furao joined #salt
15:38 jalbretsen joined #salt
15:38 nmadhok joined #salt
15:39 JDiPierro joined #salt
15:42 calvinh joined #salt
15:43 elfixit1 joined #salt
15:44 _mel_ joined #salt
15:44 cleme1mp joined #salt
15:46 matti joined #salt
15:46 murrdoc morning
15:46 gyre007 joined #salt
15:46 hobakill has anyone reported centos 6.6 issues and staying connected to a salt-master?  when i run a salt-run manage.status the only ones that show up as down are centos 6.6 boxes. i find that...... peculiar
15:46 _mel_ Hi. i try to manage salt-minion on windows. i have two sates defined. one for a config file (scheduler.conf) and one service state to restart salt-minion. it works in linux but in windows the services isn't restartet.
15:46 hobakill murrdoc, morning
15:48 bluenemo joined #salt
15:48 bluenemo joined #salt
15:49 calvinh_ joined #salt
15:49 jrluis1 joined #salt
15:51 Deevolution _mel_:  You essentially can't restart the minion-service 'directly' via salt.  You'll need to schedule a job to run after your state runs.
15:52 mdasilva does --state-output=mixed work with the smtp_returner?
15:53 Deevolution _mel_: see http://docs.saltstack.com/en/latest/faq.html#id14
15:53 _mel_ Deevolution: ok. how would i do that?
15:53 Deevolution _mel_:  I do something similar to the above link to handle the restarts
15:54 _mel_ Deevolution: AAh i see, Thx :-)
15:56 penguin_dan joined #salt
15:56 murrdoc does AT exist in windows
15:57 phx not really, windows has a difference facility for it
16:00 mdasilva joined #salt
16:02 Deevolution murrdoc:  It does exist, but it ends up creating scheduled tasks for it.
16:02 Ahlee if you equate scheduled tasks to cron, then windows has no direct 'atd' equivalent
16:02 murrdoc interesting
16:03 Deevolution I'd also point out, that we stopped using Windows 'at' command because of the idiosyncrasies of it
16:04 simonmcc joined #salt
16:04 Ahlee Deevolution: oh?  Any direct examples?
16:04 racooper joined #salt
16:04 Ahlee we just set up the windows minions to salt-call state.highstate
16:05 SheetiS joined #salt
16:05 Deevolution Ahlee:  The scheduled tasks it creates only run once (as you would expect), but end up littering the task scheduler in windows with task labeled 'at1', 'at2', etc...  They don't get cleaned up.
16:06 Ahlee Deevolution: ah, yeah.
16:06 nmadhok joined #salt
16:06 Ahlee Deevolution: we have states up states for "remove_scheduled_task_win_1", etc
16:06 Deevolution LOL
16:06 Deevolution We found a way of doing the same thing without leaving them in place.
16:08 Furao joined #salt
16:09 Ahlee oh?  On the todo list is to write some simple PS to look for jobs that aren't scheduled to run again and loop over them removing them
16:09 clintberry joined #salt
16:10 Ahlee but, given we try to not schedule stuff through salt that's so far back burnered it'll never happen
16:16 hasues joined #salt
16:16 hasues left #salt
16:17 Deevolution Ahlee:  One of my co-workers wrote the following that we call in a state when we need to restart a Windows salt-minion
16:17 Deevolution schtasks /Query /TN RestartSaltMinion & if ERRORLEVEL 1 (schtasks /create /RU SYSTEM /tn \"RestartSaltMinion\" /tr \"net stop salt-minion && net start salt-minion\" /sc once /st {0} /sd {1} /V1 /Z
16:18 Brew joined #salt
16:19 Ahlee Deevolution: similar: https://gist.github.com/jalons/62577acb88d54fdb64b4
16:20 Deevolution Very similar.
16:20 Deevolution Makes me think we're (both) on to a more or less correct solution.
16:21 jtang joined #salt
16:22 manytrees joined #salt
16:28 tov-tov_ joined #salt
16:28 pdayton joined #salt
16:31 StDiluted joined #salt
16:33 schristensen joined #salt
16:36 chrism__ joined #salt
16:36 chrism__ morning
16:37 chrism__ anyone have any specific experience with deploying minions in the DMz?  and how they can be leveraged to copy files/etc into the same?
16:37 repl1cant joined #salt
16:37 PepperIndustries joined #salt
16:44 schristensen joined #salt
16:45 nitti joined #salt
16:49 tligda joined #salt
16:49 PovAddict joined #salt
16:55 tligda1 joined #salt
16:56 neogenix joined #salt
16:56 elfixit joined #salt
17:00 viq chrism__: you have to keep in mind that minions make connections to master, not the other way around. Not sure what you mean by the second part of the question
17:01 linjan joined #salt
17:03 tzero joined #salt
17:05 conan_the_destro joined #salt
17:05 chrism__ Morning viq
17:06 toanju joined #salt
17:06 implicitnewt joined #salt
17:06 chrism__ here is my usecase: salt as scp replacement in DMZ.
17:07 mount-a joined #salt
17:07 viq not enough info
17:07 ralalala joined #salt
17:08 viq time to run, cya
17:08 chrism__ have a great day
17:08 chrism__ we build WL domains, and are having issues building and deploying into the dmz hosts
17:09 iggy so long as the 2 ports mentioned in the docs are open from minion to master, you should have full salt funcitonality
17:09 keeth joined #salt
17:09 neogenix joined #salt
17:09 chrism__ 4505/4506?
17:09 peters-tx0 joined #salt
17:09 chrism__ have you tried this yourself iggy?
17:10 signull_ joined #salt
17:11 iggy no, I haven't
17:11 mount-a joined #salt
17:11 chrism__ our security team wouln't permit ports being opened in the dmz
17:12 implicitnewt Setup a salt master, currently have about 50 minions setup on it and managing them.  I've noticed a few times when I run a salt '*' state.sls <somestate> not all the minions will be hit unless I issue multiple runs.  Even targeting a single minion will sometimes return no response then if I run it again it will respond.  Is there any configuration I can tweak or timeouts somewhere that would cause this?  Pretty vanilla config,
17:12 implicitnewt the only things I've changed are file and pillar roots
17:12 chrism__ thx for your honestly
17:12 aqua^mac joined #salt
17:12 chrism__ honesty <spelling>
17:13 iggy implicitnewt: sadly, that's something we've all seen... it seems to be worse with certain network configs (/me glares at GCE)
17:13 iggy implicitnewt: I usually run a few test.ping's before I start doing any salt stuff if I haven't touched anything in a while
17:14 felskrone joined #salt
17:14 mount-a joined #salt
17:14 smcquay joined #salt
17:16 implicitnewt iggy: good enough, just wanted to make sure I wasn't missing something simple :)  Not a huge deal, just a bit of an annoyance.
17:16 nmadhok joined #salt
17:18 JDiPierro joined #salt
17:20 apookH joined #salt
17:22 StDiluted joined #salt
17:23 otter768 joined #salt
17:23 seanz joined #salt
17:28 iggy yeah, I'm pretty sure everybody hits it at some point
17:29 rojem joined #salt
17:30 hal58th joined #salt
17:31 mohae joined #salt
17:31 aparsons joined #salt
17:31 ajw0100 joined #salt
17:32 KyleG joined #salt
17:32 KyleG joined #salt
17:34 wendall911 joined #salt
17:35 codekobe joined #salt
17:35 arnoldB pcdummy: #salt-de might be interesting for you
17:35 arnoldB felskrone: ^
17:39 thedodd joined #salt
17:40 jeremyr joined #salt
17:40 signull joined #salt
17:42 mschiff what can be considered "better practice" for a jinja template source:'ed from within a file.managed state: putting data via pillar.get with defaults: and context: OR loading via pillar.get within the template?
17:42 GabLeRou_ joined #salt
17:42 nickdew joined #salt
17:45 Ryan_Lane joined #salt
17:45 arnoldB mschiff: examples?
17:47 _JZ_ joined #salt
17:48 mschiff arnoldB: well say you want to assemble a config file
17:48 scarcry joined #salt
17:50 schlueter joined #salt
17:50 patrek joined #salt
17:51 hebz0rl joined #salt
17:52 druonysus joined #salt
17:53 _JZ_ joined #salt
17:53 desposo joined #salt
17:56 clintber_ joined #salt
17:56 ngolnik joined #salt
17:57 spookah joined #salt
17:58 ngolnik FYI: looks like epel retired python26-distribute which is required by python26-jinja2 in rhel5
17:59 iggy ngolnik: probably file a ticket in the salt-bootstrap project
18:02 kitp joined #salt
18:02 nmadhok joined #salt
18:02 nmadhok1 joined #salt
18:03 ngolnik iggy: no problem
18:05 clintberry joined #salt
18:05 forrest joined #salt
18:05 arnoldB mschiff: I use both methods. https://github.com/bechtoldt/saltstack-binddns-formula/blob/master/binddns/files/named.conf.zones vs https://github.com/bechtoldt/saltstack-elasticsearch-formula/blob/master/elasticsearch/init.sls#L58 vs https://github.com/bechtoldt/saltstack-postfix-formula/blob/master/postfix/files/main.cf
18:06 arnoldB mschiff: it also depends whether you need to prepare the data (in the SLS file) or it can be directly source'd from the pillars
18:07 arnoldB mschiff: sometimes I load large data structures directly instead of using the defaults or context param in the state file
18:07 snuffychi joined #salt
18:08 arnoldB mschiff: and sometimes I don't care and do what my brain/ clock says :)
18:15 cpowell joined #salt
18:18 elfixit joined #salt
18:20 felskrone joined #salt
18:21 bash1234123123 joined #salt
18:21 clintberry joined #salt
18:22 mschiff arnoldB: ok ;-)  and thank you for the examples
18:22 clintberry joined #salt
18:25 ajw0100 joined #salt
18:25 clintber_ joined #salt
18:27 desposo1 joined #salt
18:27 arnoldB l
18:27 jalbretsen joined #salt
18:27 arnoldB ups
18:28 clintberry joined #salt
18:30 murrdoc joined #salt
18:32 twellspring joined #salt
18:34 clintber_ joined #salt
18:41 shaggy_surfer joined #salt
18:42 Grokzen joined #salt
18:42 shaggy_surfer1 joined #salt
18:47 shaggy_surfer joined #salt
18:49 saffe joined #salt
18:51 chrism_ joined #salt
18:51 vbabiy joined #salt
18:51 shanedlee joined #salt
18:52 EWDurbin joined #salt
18:52 ipmb joined #salt
18:53 UtahDave joined #salt
18:53 jrluis joined #salt
18:57 OnTheRock joined #salt
18:57 epcim joined #salt
18:58 chiui joined #salt
18:58 echtish joined #salt
19:00 aqua^mac joined #salt
19:01 clintberry joined #salt
19:02 favadi left #salt
19:09 bhosmer joined #salt
19:10 clintberry joined #salt
19:15 paha_ joined #salt
19:15 linjan joined #salt
19:16 paha joined #salt
19:17 bhosmer joined #salt
19:19 mdasilva_ joined #salt
19:21 bhosmer_ joined #salt
19:24 otter768 joined #salt
19:24 jhauser joined #salt
19:24 mdasilva joined #salt
19:31 chrism_ afternoon.  has anyone successfully leveraged a minion as a replacement to scp in the DMz network?
19:36 rojem joined #salt
19:37 murrdoc need more context
19:39 chrism_ one thing I want to investigate is using salt as scp replacement in DMZ if we can get minion in DMZ then at very least we can mimic what we do outside DMZ which is to use minion as proxy for scp requests to other DMZ hosts then if my scripts need scp and the host is in DMZ, then fire request to minion to proxy the request
19:41 murrdoc you want a minion to be able to push files to other minions ?
19:41 iggy as we said earlier, you're going to beed ports open for master/minion communications
19:41 chrism_ if we can get a minion in the DMZ then at least we can mimic what we do outside DMz (which is to use minion as proxy for scp requests to other DMz hosts).  then, our scrits need scrip and the host is in the dmz, fire request to minion to proxy
19:41 iggy and you said that wasn't possible, so no, salt isn't going to work for you
19:41 murrdoc what iggy said
19:41 chrism_ iggy your back :)
19:41 thedodd joined #salt
19:42 theologian joined #salt
19:42 iggy *you're
19:42 chrism_ can I use salt syndic?
19:42 iggy still need ports open
19:43 chrism_ but ports would only need to be open from the master? no?
19:43 chrism_ iggy - I didn't know you were a teacher. sorry for the bad grammar :(
19:43 iggy from the minions to the master
19:43 iggy (as was also stated earlier)
19:44 chrism_ from the docs -> e Salt master communicates with the minions using an AES-encrypted ZeroMQ connection. These communications are done over TCP ports 4505 and 4506, which need to be accessible on the master only
19:45 Ozack-work joined #salt
19:47 iggy basepi: ! I think I have something for you to look at
19:47 * basepi hides
19:48 iggy basepi: https://www.refheap.com/96501
19:49 sine_nitore joined #salt
19:50 basepi How'd you get that one?
19:50 basepi Mine I assume?
19:51 iggy it's in the logs, I'm not sure yet what that equates to
19:51 murrdoc yup definitely your bug
19:51 murrdoc not that that was your question :D
19:51 iggy oh, yeah, I think it's from a mine call in an orchestrate call (or a scheduler call)
19:51 iggy not sure exactly which one yet
19:52 paha joined #salt
19:52 iggy jeez, 2014.1 -> 2015.2 is going to be a painful migration
19:53 thedodd joined #salt
19:54 nahamu iggy: skipping 2014.7?
19:54 iggy at this point, yes
19:54 nahamu where will the biggest pain be?
19:54 basepi iggy: shouldnt' be more painful than 2014.1 to 2014.7
19:55 basepi As far as I can tell
19:55 iggy considering most of this code was running on 2014.7, I'm guessing it's changes made after that
19:55 basepi iggy: but ya, this looks like a bug, matcher_args isn't being used, and is where delimiter is being handled
19:55 basepi I'll investigate this more after I get some lunch
19:55 iggy is Utah mountain or pacific time?
19:55 basepi Doesn't matter, because I'll get to it when I get to it. ;)
19:56 basepi However, I am in mountain
19:57 iggy yeah, well, as I said, I doubt we'll be on Salt much longer unless we find something that's just too painful to run in kubernetes... boss has huge kuboner
19:59 iggy I'm just trying to add some hell to your lives before I leave
20:00 murrdoc woah losing iggy would suck
20:02 andrej joined #salt
20:04 andrej Can I mix and match yaml2 and python in a state file? I'd like to get rid of my current manual method of creating splunk password files by scripting it
20:08 mdasilva i thought iggy was part of the salt dev team
20:08 mdasilva or a bot
20:08 iggy closer to a bot
20:09 mdasilva boss has a docker boner huh?
20:09 iggy andrej: you should be able to include a python one from a yaml one (totally untested)
20:10 iggy mdasilva: I heard someone say clocker in the hallway the other day and heard something hit the underside of his desk with pretty decent force
20:11 mdasilva lol
20:13 signull joined #salt
20:14 kitp joined #salt
20:14 signull joined #salt
20:15 kitp any pointer at how to get a mined 'network.ip_addrs' for a specific machine in jinja?  i know how to get a list for a given role...
20:16 mpanetta kitp: USe minion ID instead of role?
20:16 mikeywaites joined #salt
20:18 spiette joined #salt
20:21 kitp mpanetta: what's the format?  just replace 'roles:role' with 'minion_id'?
20:21 iggy and remove the "grains" argument
20:22 iggy but if you know that much, why are you using mine anyway? Just to get the consistent network.ip_addrs dict returned?
20:23 kitp that's kinda what i'm asking.  i think.  :)  how do i get that?
20:27 twellspring joined #salt
20:29 laax joined #salt
20:34 linjan joined #salt
20:37 signull_ joined #salt
20:42 lothiraldan joined #salt
20:42 iggy I'm not sure we're the best people to answer that
20:44 jhauser joined #salt
20:45 twellspring joined #salt
20:46 xsteadfastx joined #salt
20:47 glyf joined #salt
20:49 clintberry When I set full_restart: True on a service in salt, is it supposed to do a stop and start command, or just a restart?
20:49 aqua^mac joined #salt
20:50 jerematic joined #salt
20:54 Grokzen joined #salt
20:54 kim0 joined #salt
20:55 basepi iggy: Sad to hear it. At least kubernetes has some salt. But I'm definitely going to get that fixed this afternoon.
20:56 basepi (The grains issue, not the fact that kubernetes has some salt. ;) )
20:57 linjan joined #salt
20:57 mdasilva the data variable available to reactor, its values are just the grains for that minion correct?
20:58 iggy basepi: so that was enough to track it down?
20:58 basepi Yep, I see exactly what's going on. Need to figure out why it regressed
20:59 iggy mdasilva: it's the "event" data (for the most part you'll have to watch the event bus to grab all the info)
20:59 basepi The fix should be relatively easy, I think
20:59 mdasilva roger
21:00 cpowell quick question, how does salt handle package installations that have an ncurses prompt
21:00 cpowell like postfix, mysql, proftpd
21:00 cpowell at least in ubuntu
21:01 iggy there's a module for that
21:01 iggy http://docs.saltstack.com/en/latest/ref/states/all/salt.states.debconfmod.html#module-salt.states.debconfmod
21:05 yorjo joined #salt
21:09 xsteadfastx joined #salt
21:10 yorjo Is it possible to access the salt job_id in a state? I would like to reuse this unique identifier in the name of an output directory (where data produced by commands runned by the state will be written)
21:11 glyf_ joined #salt
21:11 shanedlee joined #salt
21:14 nikogonzo irccloud outage also serves as a guerilla ad
21:14 nikogonzo noice
21:15 iggy yorjo: I'm pretty sure no, but if you find out different, let us know
21:15 shanedlee left #salt
21:15 shanedlee joined #salt
21:18 andrej thanks iggy ... does anyone have examples of pure python state files flying about?  Can I pass (in including a pthon state from a yaml one) parameters to it? I need to process the fqdn for what I'm trying to do.
21:21 cpowell thanks iggy
21:22 murrdoc andrej:  salt-contrib in github.com/saltstack
21:23 andrej murrdoc - ta
21:23 modafinil_ joined #salt
21:24 thunderbolt_ joined #salt
21:25 basepi iggy: haven't actually tested this yet, but 95% sure that https://github.com/saltstack/salt/pull/20075 will fix your issue.
21:25 otter768 joined #salt
21:25 octarine joined #salt
21:25 iggy basepi: Do you want me to test?
21:26 basepi iggy: would be useful. I'm testing before and after now, regardless.
21:26 natewalck joined #salt
21:26 pdayton joined #salt
21:27 antonw joined #salt
21:27 badon joined #salt
21:27 joeyparsons joined #salt
21:29 munhitsu___ joined #salt
21:30 gladiatr joined #salt
21:30 gladiatr joined #salt
21:31 mattl joined #salt
21:32 copelco joined #salt
21:32 glyf joined #salt
21:34 akoumjian joined #salt
21:35 TheoSLC joined #salt
21:35 beneggett joined #salt
21:36 wavis joined #salt
21:36 iggy basepi: well, it's working(ish)
21:36 JonGretar joined #salt
21:36 iggy now I reproduceably get #20065
21:36 glyf joined #salt
21:37 mihait joined #salt
21:37 iggy before it was sometimes the error I told you and sometimes the other one in the bug
21:37 grepory joined #salt
21:38 davedash joined #salt
21:38 basepi Yep, basically the bug I just fixed would trigger on any job targeted with grains/pillar targeting within a compound match. Now they're all failing on #20065, so that makes sense.
21:39 basepi I'll get a high severity on 20065 and get some eyes on it. May have to ping tom, my state-compiler-foo is lacking.
21:40 iggy it looks like it's some sort of nesting thing (scheduler -> orchestrate -> state)
21:40 whiteinge joined #salt
21:41 iggy let me see if I can actually paste some code
21:41 basepi iggy: do the orchestrate runs run successfully if run manually?
21:42 basepi via `salt-run state.orchestrate`?
21:42 iggy yes
21:42 basepi Interesting
21:42 abeym joined #salt
21:42 iggy or not
21:42 m0nky joined #salt
21:43 TheoSLC Whats the difference between module.mod_watch and module.wait?
21:44 iggy it is running, but reporting failure: Running function postgres.psql_query failed on minions: dev-arcee01.c.oncenter-oasis-dns-project.oncenter.com.internal Function postgres.psql_query ran on dev-arcee01.c.oncenter-oasis-dns-project.oncenter.com.internal.
21:44 iggy failed and then ran
21:44 iggy weird
21:44 iggy I'll put that output in there
21:44 basepi iggy: Now that I look closely at that one, it may be related to the currently in-process rewrite of the runners to be event-driven
21:44 epcim joined #salt
21:45 basepi cachedout originally rewrote it, then jacksontj rewrote that rewrite, then rewrote it again with feedback from kiorky, and apparently it's still kind of broken so they're fixing it again. So runners are kind of in flux at the moment. =P
21:45 JDiPierro joined #salt
21:45 basepi Which is why we don't have an official RC out for 2015.2 yet
21:46 kiorky we are on it
21:46 kiorky ^^
21:46 kiorky iggy: i'm bitten with this also since 2 weeks :p
21:46 kiorky iggy: but its finalizing now.
21:46 johtso joined #salt
21:47 kiorky and way.more.better(tm)
21:47 jacksontj basepi: yea, cachedout made some backwards incompat changes, but we are trying to add that back in ;)
21:48 basepi jacksontj: No biggie, just wanted to keep iggy apprised. =)
21:48 basepi iggy: The takeaway is that 2015.2 is really not ready for migration testing yet. =\
21:48 twellspring joined #salt
21:49 jacksontj for runners, everything else *should* be good (hopefully ;) )
21:51 istram joined #salt
21:56 ALLmightySPIFF joined #salt
21:56 paha joined #salt
21:56 iggy these changes are in 2015.2 branch already?
21:58 iggy I think I'm using about a week old checkout
22:00 hal58th1 joined #salt
22:00 giantlock joined #salt
22:01 druonysuse joined #salt
22:04 cpowell iggy: does the depconf.set need to be required by a pkg.installed or does simply setting the name field enough to glue it together
22:05 iggy cpowell: I try to rely on the topdown processing as much as possible
22:05 iggy so just put it above the pkg.installed in the sls file
22:05 cpowell gotcha
22:05 GabLeRoux joined #salt
22:05 iggy when you start using requires/etc ordering changes a lot more than most people expect
22:06 WatfordJohn joined #salt
22:06 cpowell i know, I am just used to older Salt where state order wasn't guarenteed. its a habit i haven't broken 100% yet
22:09 mosen joined #salt
22:11 fllr joined #salt
22:11 ralala joined #salt
22:11 fllr Hey guys. What is this use_vt option I started seeing around the documentation?
22:12 iggy not as useful as you'd expect (at least in my experience)
22:14 basepi iggy: yes, they've been ongoing for a few weeks
22:15 _JZ_ joined #salt
22:17 alexhayes joined #salt
22:18 abe_music joined #salt
22:19 nitti_ joined #salt
22:23 mariner joined #salt
22:23 basepi iggy: for the record, I can reproduce your compound matcher issue on 2015.2, and it appears to be fixed with my fix in 20075. (Fixed another bug along the way, that's why it took so long to test)
22:24 iggy yeah, either way, it looks correct
22:24 Edgan Is salt autosigning known broken in 2014.7.0 as https://github.com/saltstack/salt/issues/18705 implies?
22:28 murrdoc yeah its merged into develop
22:28 iggy looks like a big change to be still working on T-5 days till 2015.2 (at least the first day of said month)
22:28 murrdoc Edgan:  how did u install your salt package
22:28 murrdoc cause its been back ported in
22:28 scoates joined #salt
22:29 twellspring joined #salt
22:30 Nazca joined #salt
22:30 Edgan murrdoc: I installed it via rpms from epel
22:31 murrdoc ouch
22:31 Edgan murrdoc: The supposedly fix seems to work.
22:31 murrdoc yay
22:31 scoates joined #salt
22:31 Edgan murrdoc: I guess I need to make my own rpms, and bug epel about fixing theirs.
22:32 murrdoc i guess
22:32 murrdoc making your rpm should move your forward
22:32 Edgan 2015.2 is going to be the next release?
22:33 murrdoc there is a .1 releas for 2014.7
22:33 Edgan ah
22:33 murrdoc and yes thats the next release
22:33 murrdoc basepi:  can answer more questions about that one
22:33 Edgan murrdoc: Looks like github hides 2014.7.1 :\
22:34 basepi Github is hiding it? o.O
22:34 basepi Oh, it's possible it's not on the github release page, I haven't been maintaining that
22:34 iggy tags
22:34 Edgan It calls 2014.7.0 latest, and has 2014.7.1 behind a click of "show newer tags"
22:34 Edgan murrdoc: Looks like 2014.7.1 is in epel-testing
22:35 murrdoc haha
22:35 Edgan Seems like 2014.7.1 should be latest
22:35 murrdoc try it out Edgan
22:35 Edgan murrdoc: Going to :)
22:35 murrdoc i assume all fixes in .0 will be in .1
22:35 Vye Has anyone successfully used eventlisten.py w/ the minion event bus? It works on my master, but not on my minion.
22:35 murrdoc if not holler at basepi
22:36 basepi 2014.7.1 is latest. I'm working on the Github release page
22:36 basepi Keep forgetting that exists
22:36 iggy Vye: I've never tried, but I wouldn't expect it to work on the minion
22:36 Vye iggy, Why?
22:36 basepi (PS: 2014.7.1 is pretty awesome, I think. Try it!)
22:37 Edgan It seems to have the auto-sign fix in it :)
22:37 iggy Vye: no clue, it's just not something I'd expect
22:37 iggy (and it seems kind of pointless because reactors are master side)
22:38 murrdoc i didnt thing event_listener worked on minions
22:38 linjan joined #salt
22:38 iggy looking at the source, it looks like it does
22:38 aqua^mac joined #salt
22:38 Vye Well, the script is configured with minion/master options. There are sockets in /var/run/salt/minion so, I should be able to connect and listen.
22:39 basepi Release is up on Github now: https://github.com/saltstack/salt/releases/tag/v2014.7.1
22:39 iggy the question still kind of remains... to what end?
22:40 Vye I want a monitor running on the minion to tap into the event bus. There may be something there I can use to monitor the health of the minion.
22:40 markmarine joined #salt
22:41 Vye Mainly want to detect if it ever gets hung.
22:41 markmarine anyone have helpful suggestions for how to debug jinja templates?
22:41 esogas_ joined #salt
22:42 iggy Vye: as a few of us have mentioned we didn't even know it would do that... there's no telling how long it's been broken... try downloading some older version and trying that
22:44 jacksontj basepi: iggy if you want to give it a spin with our current version i just made a 2015.2 PR https://github.com/saltstack/salt/pull/20086
22:44 jacksontj not ready to merge, but feel free to check it out
22:44 jacksontj hopefully we found all the things that were broken ;)
22:47 markmarine so, I'm trying to figure out what my template renders into before sending it to a machine, is there any way to do this? I tried to import salt in python then pass my template to the jinja.render function, but that doesn't seem to work
22:47 markmarine and the error message coming from the jinja stacktrace isn't extra helpful
22:48 markmarine so, any tips or suggestions? What do you all do to test these during development?
22:48 iggy markmarine: use salt-call -l debug and it should show you the rendered output
22:48 jer__ joined #salt
22:49 Edgan basepi: :)
22:49 iggy holy crap that's a lot of commits
22:50 markmarine @iggy  Thank you I'll try that. if there is a jinja error (right now it's telling me that 'dict' doesn't have the variable "user")
22:51 iggy markmarine: that can mean a few different things, you're better off posting the state that's causing it for us to look at
22:51 kermit joined #salt
22:53 SheetiS joined #salt
22:55 spookah joined #salt
22:56 shaggy_surfer joined #salt
22:56 jakesyl joined #salt
22:57 markmarine iggy: http://pastebin.com/MzrNRCvK
22:58 markmarine iggy: I pulled out most of my logic while I was troubleshooting, I know in this iteration I don't need the map.jinja, I should just be able to get some of that data from the pillar, but I will need it later
22:58 murrdoc basepi:  thanks for release page
22:58 ajw0100 joined #salt
22:59 iggy markmarine: did you verify that the minion is getting the proper pillar data?
22:59 markmarine no, I haven't yet. let me do that
23:02 tligda joined #salt
23:02 druonysus joined #salt
23:02 druonysus joined #salt
23:03 nitti joined #salt
23:05 markmarine iggy: ok, I checked and the pillar data is there
23:06 iggy can you paste the full error message too?
23:06 markmarine sure
23:06 iggy ahhh
23:07 iggy pillars aren't namespaced by file name
23:07 iggy you are looking up uwsgi_test:user but the pillar is actually just user
23:08 markmarine http://pastebin.com/H0Uj8BQ9
23:08 markmarine ohhh
23:08 markmarine ok awesome
23:08 iggy I almost feel like that deserves a mention in the docs... it comes up often enough
23:08 linjan joined #salt
23:09 jer___ joined #salt
23:10 markmarine iggy: I'm still getting the same error after I change that
23:10 markmarine {% set user = salt['pillar.get']('user') %}
23:10 markmarine that should work right?
23:10 iggy assuming you didn't leave anything out
23:10 markmarine then I can call out {{ user.app_name }} or whatever in jinja
23:11 REdOG joined #salt
23:11 markmarine iggy: haha. big assumption :) I
23:11 markmarine 'll go triple check
23:12 iggy repaste what you've got now and the new error
23:12 markmarine SaltRenderError: Jinja variable 'None' has no attribute 'app_class'
23:13 bhosmer joined #salt
23:13 iggy I don't see anywhere in there a lookup for app_Class
23:13 iggy (aside from {{ app_class }}
23:13 markmarine it's in the uwsgi.conf
23:14 iggy and how are you getting it in there?
23:14 markmarine I'm pulling the pillar data in through the map.jinja
23:15 markmarine which is called in the top of init.sls
23:15 iggy you are pulling user in
23:15 markmarine oh, I fixed the namespacing in that pastebin
23:15 iggy how are you mapping out user:app_class to the flat jinja variable {{ app_class }}
23:15 markmarine now they say {{ user.app_class }}
23:16 iggy ahh
23:17 iggy see, this is why updated pastes are awesome
23:17 markmarine I pull in user with map.jinja from pillar as user, then dot notation the variables in the .conf   .... this thing is so simple at this point, I can't figure out what's going wrong
23:17 markmarine roger, I'll update them all  :)
23:17 iggy you aren't passing user to the conf file
23:17 iggy - context: - user: {{ user }}
23:17 iggy or something
23:19 markmarine http://pastebin.com/LUdPM0ir
23:20 iggy you need the to use context in the file.managed stanza
23:20 iggy that gets the user dict into the conf file
23:21 iggy (the other way being to just grab user directly in the conf file)
23:22 rypeck joined #salt
23:23 murrdoc (dont do that)
23:24 iggy (I do it all the time)
23:24 mndo joined #salt
23:25 otter768 joined #salt
23:27 smcquay joined #salt
23:28 murrdoc (u is l33t)
23:28 mosen joined #salt
23:28 markmarine no it's complaining that the context isn't a dictionary. wtf
23:29 iggy did you follow the docs
23:29 murrdoc http://yaml-online-parser.appspot.com/
23:29 markmarine yeps, 4 spaces if there is context and default, 2 spaces if there is just context, right?
23:29 murrdoc paste your state there
23:29 murrdoc does it look like a nested dict
23:30 murrdoc salt totally needs a 'state.render_yaml' mode
23:30 iggy I generally use 4 either way... doesn't seem to complain
23:31 tligda joined #salt
23:31 markmarine murrdoc: I pasted it there and it just complained about the jinja
23:31 murrdoc oh
23:31 murrdoc sorry
23:31 markmarine '{{'
23:31 murrdoc without the jinja
23:31 murrdoc put temp text in for the jinja
23:31 murrdoc if possible
23:32 markmarine looks good w/o the jinja
23:33 markmarine so that means the jinja isn't being parsed in this state.sls test=True during my test run?
23:33 PovAddict well there is a show_sls command which can help
23:33 murrdoc true
23:34 markmarine I'm just trying to use this: salt -G 'node_type:events_server' state.sls uwsgi_test test=True
23:34 markmarine and hoping I get the jinja/yaml output in the diff of the conf file
23:36 aw110f joined #salt
23:37 markmarine how is this not a dictionary? http://pastebin.com/aZXqe4de
23:37 markmarine here is the error:      Comment: Context must be formed as a dict
23:38 iggy no dash
23:38 iggy I typed that wrong earlier
23:38 iggy it's just - context:\n  user: {{ user }}
23:38 PovAddict dashes are for lists
23:39 markmarine right, sorry. that was there from trying to figure out why it did this: http://pastebin.com/aNP8r9sG which happens with no dash
23:40 markmarine does that mean I can't use the pillar 'user' ? which would make a little bit of sense and I can change it
23:41 iggy that probably means you had some other error
23:41 murrdoc TypeError: must be string, not OrderedDict
23:42 murrdoc whats your state look like
23:42 neogenix joined #salt
23:46 nitti joined #salt
23:53 markmarine ok, I don't know why for sure this was happening, but when I ran that with state.sls test=True it failed... but putting it in as a highstate and then actually running it got me a workable file
23:54 markmarine so really, the only way I can test these jinja templating files is running them? I must be missing something, it brings me back to my first question... how do you guys debug these issues?
23:54 TheoSLC joined #salt
23:55 murrdoc build on vagrant
23:57 badon joined #salt

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