Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-09-19

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

All times shown according to UTC.

Time Nick Message
00:00 baconbeckons oh, you’re not talking to me :)
00:00 dstokes sry, *can't*
00:00 dstokes i've installed redis w/ pip on the master. the module loads, but it's not the redis module..
00:00 __number5__ dstokes: paste your code and error traceback somewhere
00:01 __number5__ dstokes: my guess will be you have a redis.py file in your $PYTHONPATH
00:01 dstokes https://gist.github.com/dstokes/f3c7f57bc2c26759dfe4
00:04 __number5__ dstokes: can you also paste /srv/salt-states/modules/pillar/envs.py?
00:04 dstokes i did. it's pillar.py
00:05 dstokes renamed in the gist
00:06 dstokes did a locate and found a redis.pyc. removed it and now i'm getting something else (specified ext_pillar interface envs is unavailable). guessing there's an error in my code somewhere
00:06 dstokes __number5__: thx for the tip (on redis.py file)
00:11 __number5__ dstokes: np
00:13 teebes joined #salt
00:17 mapu joined #salt
00:21 dstokes __number5__: you know how i can get stack error output from an ext_pillar module? salt seems to be gobbling the error. raising saltinvocationexception and just letting it die don't work
00:21 dstokes all i get is [CRITICAL] Specified ext_pillar interface envs_pillar is unavailable
00:22 oz_akan joined #salt
00:22 __number5__ I'm not sure, do you already have debug flag on salt-master?
00:22 dstokes yea
00:24 dstokes looks like this is killing me https://github.com/saltstack/salt/blob/795ab7161718cf0c24879a3d27bfd0b38ccf6de7/salt/pillar/__init__.py#L506
00:24 ph8 joined #salt
00:24 dstokes problem matching keys..
00:32 kermit joined #salt
00:33 obimod joined #salt
00:34 aquinas joined #salt
00:35 mechanicalduck joined #salt
00:35 dstokes reprovisioned the master server (vagrant) and everythings working. must have been a bad cache somewhere
00:36 dstokes these ext_pillars are tempermental ;)
00:38 __number5__ yep, cache are sticky
00:38 __number5__ XD
00:39 dstokes OMG IT'S WORKING
00:39 dstokes thx guys
00:39 fragamus joined #salt
00:42 catpigger joined #salt
00:48 kingel joined #salt
01:03 dalexander joined #salt
01:09 gadams joined #salt
01:12 tligda joined #salt
01:13 malinoff joined #salt
01:15 wt joined #salt
01:16 wt is it possible to use two states from the same state module in one named stanza?
01:16 wt e.g.
01:16 wt /etc/blah.cfg:
01:16 wt file.managed:
01:16 wt ...
01:16 wt file.recurse:
01:16 wt ...
01:16 wt ?
01:16 Ryan_Lane wt: no
01:16 wt annoying
01:16 manfred it is not
01:16 wt ok
01:17 manfred it is possible to have two file.manageds, kind of, in 2014.7
01:17 manfred by exploiting the changes to -names:
01:17 Ryan_Lane I'm weird, but I don't like combining more than one module/function under a single ID
01:17 Ryan_Lane I actually wish I could completely disallow it :)
01:17 wt well, I was going to create a directory and populate it with recurse
01:18 Ryan_Lane right, so do it with two separate states
01:18 wt I've never actually wanted to use it before now.
01:18 wt Ryan_Lane, I am.
01:18 manfred why not file.managed with recurse: and makedirs: True
01:18 manfred http://docs.saltstack.com/en/latest/ref/states/highstate.html#names-declaration last example shows the thing I added
01:19 manfred actually
01:19 manfred that is what file.recurse does
01:19 manfred http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.recurse
01:19 Ryan_Lane wt: well, not really, here's how you'd use two states: https://gist.github.com/ryan-lane/69607524dca7b152651e
01:20 manfred why do you want it seperated into two states?
01:20 Ryan_Lane that said, you'd want to use makedir: True
01:20 Ryan_Lane as manfred is suggesting
01:20 wt managed doesn't have recurse according to the docs
01:20 tligda joined #salt
01:21 manfred I meant file.recurse
01:21 manfred it can do makedirs: True , and do your place files in place
01:21 Ryan_Lane heh. I did too :)
01:21 malinoff I was sure I could do something like that:http://pastie.org/9574074
01:21 manfred you can't
01:22 manfred because the - managed, gets grabbed out first, to make itback into a dictionary
01:22 wt recurse doesn't seem to have makedirs by the docs
01:22 wt is that it's default behavior?
01:22 Ryan_Lane I think it's the default
01:22 manfred it should be able to go in kwargs
01:22 manfred because it just uses file.manage
01:22 Ryan_Lane if not you can use two separate states
01:22 manfred file.manage_file module
01:22 wt magic
01:23 Ryan_Lane anyway, it's usually better/less confusing to always split states apart into their own states
01:24 Ryan_Lane then you'll never run into edge cases like this
01:24 Ryan_Lane (meaning don't do two function calls under the same ID)
01:24 manfred https://github.com/saltstack/salt/blob/develop/salt/states/file.py#L3935
01:24 manfred yeah, manage_file
01:24 wt then why is it even supported to have multiple states?
01:25 Ryan_Lane because some people like it
01:25 manfred actually that one is under serialize
01:25 * Ryan_Lane is a normal user, who has strong opinions
01:25 wt It's the paradox of choice.
01:25 Ryan_Lane well, some people find it really convenient
01:25 manfred i can do it whichever way I want, even if it is wrong
01:28 gadams joined #salt
01:30 otter768 joined #salt
01:30 n8n joined #salt
01:32 manfred yall have a good one, I am going to the bar.
01:32 manfred o/
01:39 NotreDev joined #salt
01:40 wt is archive.managed smart enough to reextract an archive file if the source file is updated?
01:40 wt It's unclear from the docs.
01:41 wt seems like a bad idea in any case not to change the filename of the tarball, but I was just wondering
01:44 wt archive.extracted, I mean
01:45 jhauser_ joined #salt
01:55 bhosmer joined #salt
02:08 tk75 joined #salt
02:10 wendall911 joined #salt
02:12 Gnouc joined #salt
02:12 kt76 joined #salt
02:25 bhosmer joined #salt
02:31 dalexander
02:32 tmh1999 joined #salt
02:36 ramishra joined #salt
02:36 baconbeckons joined #salt
02:36 kingel joined #salt
02:45 NotreDev joined #salt
02:45 andrej How can I accomplish something like the "onchange" that 2014.7 will bring in 2014.1.10? I'm about to deploy a bunch of replacement system config files, and after they're deployed I would like to run "sysctl -p" and "ufw reload"
02:56 otter768 joined #salt
02:59 dude051 joined #salt
03:06 kelseelynn joined #salt
03:07 malinoff joined #salt
03:08 kelseelynn left #salt
03:09 beneggett joined #salt
03:20 Hell_Fire joined #salt
03:23 Zachary_DuBois joined #salt
03:24 Hell_Fire joined #salt
03:26 iggy baconbeckons: github limitation
03:26 baconbeckons iggy: how so?
03:27 higgs001 joined #salt
03:28 iggy they don't allow you to use the same deploy keys for multiple private repos
03:28 baconbeckons oh, ok
03:30 spookah joined #salt
03:30 spookah joined #salt
03:38 ajolo joined #salt
03:40 ramishra_ joined #salt
03:48 aarontc joined #salt
03:50 smcquay joined #salt
03:50 Gnouc joined #salt
03:52 vbabiy joined #salt
04:01 englishm joined #salt
04:05 joehoyle1 joined #salt
04:06 ajolo_ joined #salt
04:10 smcquay joined #salt
04:15 smcquay joined #salt
04:22 yomilk joined #salt
04:25 kingel joined #salt
04:26 ramishra joined #salt
04:30 smcquay joined #salt
04:31 smcquay joined #salt
04:33 martinp joined #salt
04:39 jalbretsen joined #salt
04:41 smcquay joined #salt
04:43 englishm joined #salt
04:48 smcquay joined #salt
04:49 TheRealBill joined #salt
04:51 epcim joined #salt
04:54 yomilk joined #salt
04:54 oz_akan joined #salt
04:56 davet joined #salt
04:59 Gnouc joined #salt
05:06 ramishra_ joined #salt
05:11 felskrone joined #salt
05:18 wavis joined #salt
05:19 __number5__ joined #salt
05:24 Eugene joined #salt
05:28 aquinas joined #salt
05:29 aquinas_ joined #salt
05:34 yomilk joined #salt
05:35 troyready joined #salt
05:41 Hell_Fire joined #salt
05:41 kingel joined #salt
05:42 oyvjel joined #salt
05:46 Eugene joined #salt
05:51 tmh1999 joined #salt
05:54 g3cko joined #salt
05:55 catpig joined #salt
05:55 oz_akan joined #salt
05:59 ramishra joined #salt
05:59 Jahkeup joined #salt
06:19 bhosmer joined #salt
06:22 mechanicalduck_ joined #salt
06:23 Hell_Fire joined #salt
06:27 felskrone joined #salt
06:27 wnkz_ joined #salt
06:28 lopez joined #salt
06:30 asmaps joined #salt
06:34 sectionme joined #salt
06:38 lcavassa joined #salt
06:41 asmaps joined #salt
06:51 slav0nic joined #salt
06:51 slav0nic joined #salt
06:52 Sweetsha1k joined #salt
06:56 oz_akan joined #salt
07:00 otter768 joined #salt
07:02 viq joined #salt
07:06 kingel joined #salt
07:07 oyvjel1 joined #salt
07:14 sectionme joined #salt
07:22 yomilk joined #salt
07:25 oz_akan joined #salt
07:39 yomilk joined #salt
07:40 alanpear_ joined #salt
07:43 obimod_ joined #salt
07:51 PI-Lloyd joined #salt
07:53 baconbeckons joined #salt
07:56 chiui joined #salt
07:57 Daviey joined #salt
07:58 felskrone joined #salt
07:58 KaaK_ joined #salt
07:58 Hell_Fire joined #salt
07:59 intellix joined #salt
08:08 bhosmer joined #salt
08:14 che-arne joined #salt
08:26 oz_akan joined #salt
08:26 ndrei joined #salt
08:26 ndrei_ joined #salt
08:28 oz_akan_ joined #salt
08:28 saltstacker joined #salt
08:29 saltstacker hello
08:29 saltstacker anybody home?
08:30 * sectionme is at work ;)
08:30 yomilk joined #salt
08:37 VSpike saltstacker: it's best just to ask your question and then wait. IRC is a high latency medium :)
08:38 peters-tx joined #salt
08:41 pviktori joined #salt
08:42 ndrei joined #salt
08:42 ndrei_ joined #salt
08:43 Nazzy joined #salt
08:50 fe92 joined #salt
08:51 saltstacker left #salt
08:52 thehaven joined #salt
08:58 viq Also, as many people put it, "don't ask to ask"
08:59 fe92 is it possible to use salt remotely on a host without a salt minion or anything else installed on the target host? salt ssh will be able to do this?
08:59 fredvd joined #salt
09:04 viq fe92: well, not "without anything else installed", you need python a couple other libs, but in theory it should work, yes
09:05 viq fe92: though depends a bit what exactly you want to do
09:07 lionel joined #salt
09:09 darkelda joined #salt
09:10 Outlander joined #salt
09:10 NotreDev joined #salt
09:11 fe92 viq: bootstrap clean installs of centos/freebsd without having to run a manual ssh command first (requiring python2.7 is fine)
09:12 AxelFooley joined #salt
09:12 AxelFooley hi all
09:12 AxelFooley need someone who can help me to set up git remote for states and pillars
09:12 viq If you don't expect to depend on stuff like mine or publish, then I expect it should work
09:13 AxelFooley salt version 2014.1.10
09:13 viq AxelFooley: AxelFooley http://docs.saltstack.com/en/latest/topics/tutorials/gitfs.html
09:14 fe92 viq: ok, thanks I'll experiment a bit with it
09:14 viq fe92: which freebsd?
09:14 AxelFooley viq, followed that tutorial, but i'm getting "Error when fetching: fatal: unable to connect to git@gitlab.rete.farm:' returned exit status 2: None"
09:14 viq fe92: <10 requires forcing pkg provider to pkgng, won't detect by itself, AFAIK
09:14 AxelFooley but from the server running salt-master, i can connect to the git repository
09:15 viq AxelFooley: wrong URI format, you need to use / as repo separator, not :
09:15 viq AxelFooley: gitfs_remotes: - ssh://git@github.com/example/salt-states.git
09:15 AxelFooley ok, i'll try
09:16 viq at least that was my experience so far, the docs seem to say you can do it...
09:16 AxelFooley same error :(
09:16 viq AxelFooley: can you paste somewhere that bit of config?
09:17 AxelFooley sure i can, give me 1 minute
09:17 fe92 viq: think its 10.0 so should be fine, otherwise I might just run a minion on the bsd machines since we dont have that many
09:17 viq fe92: I much prefer the setup with minions myself
09:18 AxelFooley viq, here we go, https://www.refheap.com/90430
09:19 viq AxelFooley: the root=something doesn't work in 2014.1, it is a 2014.7 feature. Also, I'd try ssh:// instead of git://
09:20 AxelFooley k then, i'm going to try
09:21 AxelFooley getting no errors, maybe it's working
09:22 viq try asking one of your machines for a pillar it's supposed to have from git
09:22 sectionme joined #salt
09:22 AxelFooley viq, my repo has this structure: repo.git/states and repo.git/pillars
09:22 AxelFooley how i can instruct salt-master to get pillars from that subdir?
09:22 viq I don't think you can currently
09:23 AxelFooley so i must put pillars into the repo root dir, isn't it?
09:23 viq yeah
09:23 viq Or use 2014.7 RC1
09:23 AxelFooley meh :/
09:23 AxelFooley ok, i'll do some tests
09:23 AxelFooley thank you so much, for now
09:24 viq Sure, I had my share of beating that problem into submission, so now I have some answers ;)
09:25 N-Mi joined #salt
09:25 N-Mi joined #salt
09:28 oz_akan joined #salt
09:29 ghaering joined #salt
09:30 AxelFooley viq, seems pext_pillars doesn't works
09:30 AxelFooley 2014-09-19 11:30:06,848 [salt.pillar.git_pillar                      ][ERROR   ] Unable to fetch the latest changes from remote ssh://git@gitlab.rete.farm/sistemi/saltstack.git: 'git fetch' returned exit status 128: fatal: could not read Username for 'http://gitlab.rete.farm': No such device or address
09:30 PI-Lloyd AxelFooley: it's better to have states and pillar in seperate repos... also bear in mind with gitfs, if you have branches on your repo, salt will pick those up as different environments, for example dev stage prod
09:31 viq AxelFooley: rm -rf /var/cache/salt/* restart salt master and try again
09:31 AxelFooley i don't know where it picks up that http thing, i haven't defined it anywhere
09:31 AxelFooley k
09:31 AxelFooley PI-Lloyd, thanks, may be i'll go for that solution
09:31 viq AxelFooley: it may be an old checkout, and it remained in the git config there, that's why it's sometimes a god idea to nuke cache when you're playing with it
09:32 viq yeah, that's what I curently do as well, separate repos for states and pillars, mainly because so far you couldn't easily have them in a single repo
09:32 PI-Lloyd AxelFooley: also take a look at http://docs.saltstack.com/en/latest/topics/tutorials/gitfs.html#gitfs-authentication for the ssh authentication
09:33 babilen viq: I'm really looking forward to Helium as it will finally allow you to keep *everything* in a single repository.
09:33 PI-Lloyd if you have an ssh key to use, i would recommend using GitPython for your gitfs remotes as well
09:36 viq babilen: I'm more looking forward to pillar merging, though that formula for reverse-users (or however it was called) is interesting and may solve my immediate need
09:37 * viq bought himself the Salt Essentials (pre-release) book ;)
09:37 babilen Oh, there is a book now?
09:38 PI-Lloyd babilen: http://www.amazon.co.uk/Salt-Essentials-Craig-Sebenik/dp/1491900636
09:38 viq http://shop.oreilly.com/product/0636920033240.do
09:38 viq technically will be out in december
09:38 PI-Lloyd xmas day i believe is release date
09:38 PI-Lloyd lol
09:39 viq I prefer oreilly. you get epub and pdf in addition to mobi, drm free, and I actually got access to it already, what is ready of it
09:39 viq Also it was available with 50% discount ;)
09:40 PI-Lloyd nice
09:40 babilen yeah, O'Reilly's shop is wonderful. I guess I'll buy that book too!
09:41 PI-Lloyd might as well jump on the bandwagon and make the purchase also :)
09:41 babilen Do we know which version it is written for?
09:42 viq babilen: do you know any thing that changed that I could look for?
09:43 Outlander joined #salt
09:44 babilen viq: Does chapter 4 cover requisites? If so: Do they mention the new requisites that will be introduced in Helium such as onfail, listen, ...
09:45 PI-Lloyd i'm guessing it's for hydrogen, i can't see the onfail or listen stuff in the requisites in the preview version im looking at
09:45 viq So far there's ~1 page about requisites, and it lists only the simple require
09:45 babilen Pity ... It would really make sense to update it for Helium. That way it will be relevant for a lot longer.
09:46 viq I think the person writing it frequents here, but I can't for the life of me remember the nick
09:49 viq Ah, looks like we can blame chitown
09:57 bhosmer joined #salt
09:58 CeBe joined #salt
10:03 felskrone joined #salt
10:07 VSpike I really feel the lack of a state for windows environment variables (preferably one that can manage arrays masquerading as env vars, like PSMODULEPATH). I know there's one for PATH but I haven't seen a general one. Am I missing it or does it not exist (yet)?
10:08 lcavassa joined #salt
10:09 lcavassa joined #salt
10:10 aurynn joined #salt
10:12 lcavassa joined #salt
10:16 rodo /join #odm
10:21 TheThing joined #salt
10:23 intellix joined #salt
10:24 bhosmer joined #salt
10:27 giantlock joined #salt
10:27 diegows joined #salt
10:29 oz_akan joined #salt
10:33 yomilk joined #salt
10:35 maggo joined #salt
10:37 maggo Hi all, i searching for hours now and i don't get my mind pimped to understand how i can order included states in a state file. Can someone point me in the right direction please?
10:41 viq maggo: http://docs.saltstack.com/en/latest/ref/states/requisites.html
10:44 apergos joined #salt
10:46 peters-tx joined #salt
10:47 kingel joined #salt
10:50 apergos cna someone give me a pointer to the projected release schedule for 2014.7.0, if there is one?
10:50 apergos or else discussion about rc testing, either will do
10:51 maggo viq: Yes i know that site and i uses this in the states it self.
10:52 viq maggo: if you do includes you can also require whole sls, or particular states from it
10:52 vbabiy joined #salt
10:52 maggo viq: puh how should i explain it. I don't want to do the ordering in the states it self because on other systems the ordering are different so i have to order the packages in my host state
10:52 viq maggo: or I'm not understanding what you're asking
10:54 maggo viq: let me give you some more details. I don't use the highstate because its to risky for me
10:55 maggo viq: i use a state for every host so i can call a state like that: salt server.domain.tld state.sls path.to.my.servers.statefile
10:56 maggo viq: in that init.sls i have in the first line an include an after that all my states i need listed
10:56 * viq nods
10:57 viq (though I would suggest getting your hosts to a state where you can run highstate, makes life easier and more predictable. But I know, life and reasons and time)
10:58 viq (not that long ago got most of the machines and states here so that highstate works - though doesn't do yet everything wanted. But that's my infrastructure, not yours)
11:01 viq maggo: is there more coming?
11:02 maggo viq: hm maybe i don't get the problem i have in right words.... But highstates wouldn't help here i think. ... let me mediate about that.. i think i will write it on the mailinglist with more details
11:03 stephanbuys joined #salt
11:04 viq maggo: wouldn't that be solved by eg. 'my awesome app': file.archive: something - require: - sls: packages.deps.awesomeapp  ?
11:04 viq and at the top include the packages.deps.awesomeapp
11:04 AxelFooley viq, and PI-Lloyd thank you for your support, i've successfully integrated gitfs infto salt, now everything works, finally i've opted for a two repos architecture, one for states and one for pillars.
11:05 VSpike I'm having probs with the right syntax for this https://bpaste.net/show/cbe864bc25c0 .. i want to create a directory and fill it from the salt file server, cleaning anything that shouldn't be there
11:05 viq AxelFooley: cool it's working for you now. Yeah, such separation is needed for the time being
11:05 VSpike What's the proper way to do it?
11:06 viq VSpike: - requires: - file: {{ pillar['PSModuleDir'] }}\FMPSLib
11:06 viq in requires you need the module the state you require uses (here: file), and the name of the state (your case: {{ pillar['PSModuleDir'] }}\FMPSLib)
11:07 viq VSpike: also, you can't have two file calls under single id
11:07 VSpike viq: yeah, that was the bit I was having trouble with... so split them into two and use -name ?
11:08 viq VSpike: also, http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.recurse  look for "clean"
11:08 viq so you don't need the file.directory state
11:08 VSpike Oh :)
11:08 VSpike Thanks .. heh
11:09 VSpike If I did need the two states, I'd have to use -name, right?
11:09 viq hm?
11:09 NotreDev joined #salt
11:09 VSpike Because you can't have two states under and an ID, and you can't have two IDs the same
11:10 viq you can have multiple states under same ID, just not using same module (here: file)
11:10 kingel joined #salt
11:10 VSpike So I'd have to have one called makethedir: and one called makethefile: and pass -name: {{ pillar['PSModuleDir'] }}\FMPSLib
11:10 VSpike ah right so
11:11 viq Also, IDs can be arbitrary and include spaces. And then yes, in -name you pass the name of the file you want
11:11 viq Also with how you did that every time that state would run it would first remove all the files from the directory, and then repopulate it anew (I think)
11:15 VSpike thanks viq ... much appreciated
11:21 intellix joined #salt
11:41 N-Mi joined #salt
11:41 N-Mi joined #salt
11:46 bhosmer_ joined #salt
11:54 ndrei joined #salt
11:54 ndrei_ joined #salt
11:59 wnkz_ joined #salt
12:01 Outlander joined #salt
12:14 ndrei joined #salt
12:14 ndrei_ joined #salt
12:14 kiorky joined #salt
12:14 CookQ joined #salt
12:19 fredvd joined #salt
12:21 hobakill joined #salt
12:22 vbabiy joined #salt
12:23 micah_chatt joined #salt
12:24 eunuchsocket joined #salt
12:25 maggo viq: sorry for the delay. you wrote something about file.archive .. that funktion i can't find in the sysdoc what is it?
12:26 viq It was just an example, and I got the wrong name, I meant http://docs.saltstack.com/en/latest/ref/states/all/salt.states.archive.html
12:31 vbabiy joined #salt
12:32 maggo Hm ok i don't understand why you choose states.archive for that but if i use it like you have written i get requisites not found
12:36 dark_helmet joined #salt
12:38 vbabiy joined #salt
12:38 scbunn joined #salt
12:39 maggo viq: let me explain my problem again: I have the php-formula and there are states for php-apc php-gd php-imagic. now i include the apache formula and the 6 of the php states. Now i have to order the states to ensure that apache is installed at last with restart or that afert the last php state the apache is restarted. How can i arange that without changing the forumula?
12:39 ndrei joined #salt
12:40 ndrei_ joined #salt
12:41 vbabiy joined #salt
12:42 geekatcmu joined #salt
12:49 vbabiy joined #salt
12:52 otter768 joined #salt
12:54 obimod joined #salt
12:57 ndrei joined #salt
12:57 ndrei_ joined #salt
12:57 bhosmer joined #salt
12:59 dude051 joined #salt
13:00 robinsmidsrod joined #salt
13:01 mapu joined #salt
13:03 dude051 joined #salt
13:04 bhosmer joined #salt
13:04 kermit joined #salt
13:05 englishm joined #salt
13:05 cpowell joined #salt
13:07 racooper joined #salt
13:07 Twiglet_ joined #salt
13:11 fe92 joined #salt
13:11 ndrei joined #salt
13:11 pass_by_value joined #salt
13:11 ndrei_ joined #salt
13:15 pass_by_value left #salt
13:15 giantlock joined #salt
13:16 Twiglet_ joined #salt
13:18 viq maggo: have you looked at extending formulas?
13:19 viq maggo: http://docs.saltstack.com/en/latest/ref/states/extend.html http://docs.saltstack.com/en/latest/topics/tutorials/states_pt3.html
13:20 maggo viq: yes but thats the same just turned around. But i think i got it now. From time to time you have to get your head out of the ass ;) The failure was sitting on the chair ;)
13:21 vbabiy joined #salt
13:21 viq PEBKAC, eh? ;)
13:21 maggo viq: i just removed the require ithe states where i thought i need a require and salt sorted the states all all is fine now ;)
13:21 maggo jep
13:21 viq cool
13:23 mpanetta joined #salt
13:24 maggo viq: But thank you a lot for talking about that ;)
13:25 vbabiy joined #salt
13:25 sysadmin75 joined #salt
13:25 viq :)
13:26 lulzsp00n joined #salt
13:27 spo0nman joined #salt
13:28 teebes joined #salt
13:28 nitti joined #salt
13:30 mechanicalduck joined #salt
13:31 spo0nman hello can a salt-syndic not have multiple masters?
13:31 spo0nman just as a minion subscribes to multiple masters, a syndic should subscribe to multiple master-of-masters
13:33 spo0nman from the syndic configuration it appears as if there can be only one master of masters
13:34 Denny_ joined #salt
13:34 Denny_ Hello everyone
13:34 Sypher_ joined #salt
13:35 dvriesman I am looking for a guy that could help me with SALT in Windows Environment !
13:36 viq Don't know about help, but certainly VSpike can commiserate ;)
13:37 jslatts joined #salt
13:38 Sypher__ joined #salt
13:39 jemejones joined #salt
13:41 cpowell joined #salt
13:41 cnelsonsic joined #salt
13:42 cpowell joined #salt
13:44 cpowell_ joined #salt
13:47 dvriesman Well, my problem is that everything works on Linux but when I've tried to apply a simple directory copy directive on a windows host.. the salt command is very very very slow and sometimes returns nothing else nothing.
13:47 blarghmatey joined #salt
13:49 balltongu joined #salt
13:50 ndrei joined #salt
13:50 ndrei_ joined #salt
13:51 viq joined #salt
13:51 viq joined #salt
13:52 FeatherKing joined #salt
13:52 mgw joined #salt
13:52 apergos I have  packaging issue to report for trusty, change in behavior that causes salt-minion to play poorly with puppet (plese no complaints about why we use puppet :-P)
13:53 mpanetta Is it possible to use the saltenv variable in an orchestration runner?
13:54 apergos I'm looking at salt-minion_0.17.5+ds-1_all.deb and the debian/postinst file used to just check if the standard named config file wa in place and start the minion up if so
13:54 balltongu joined #salt
13:54 mpanetta joined #salt
13:54 mpanetta Is it possible to use the saltenv variable in an orchestration runner? ... (got disconnected...)
13:54 apergos this worked well for us, since we use a different location for the minion conf file, which is deployed by puppet after the package is installed
13:54 mpanetta Is it possible to use the saltenv variable in an orchestration runner? ... (got disconnected...)
13:55 ndrei joined #salt
13:55 ndrei_ joined #salt
13:55 mpanetta I'm trying to force the runner to only act on minions in th environment listed on the command line...
13:55 apergos with trusty, it will generate a default conf file if you don't have one, and then start salt.  that's annoying and appear causes some issues with accepting the salt key on the actual master later too
13:57 apergos if the packager has a workaround, that would be awesome
13:58 jaimed joined #salt
14:22 ilbot3 joined #salt
14:22 Topic for #salt is now Welcome to #salt | 2014.1.10 is the latest | Help us test the 2014.7 RC! http://bit.ly/salt-rc | SaltStack is hiring! http://www.saltstack.com/careers | 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/
14:23 FeatherKing joined #salt
14:23 bhosmer joined #salt
14:24 QuinnyPi1 joined #salt
14:25 ndrei joined #salt
14:25 ndrei_ joined #salt
14:25 chitown PI-Lloyd: viq: babilen: yes, the book is mine :)
14:26 chitown it was started back in jan/feb. updating it with for the new helium features is something i want to do
14:26 chitown (there is some pressure to get it done soon)
14:26 chitown so, i can update with helium, but then it would likely not be out until feb
14:26 chitown so, pros and cons :)
14:27 chitown something to be aware of; the current state is VERY rough
14:27 kedo39 joined #salt
14:27 viq chitown: I assume you are aware of horrible amount of spelling mistakes in "Some Quick Examples" part? ;)
14:27 chitown i.e. there will likely be a lot more on requisite states
14:27 chitown lol
14:27 chitown yes... unfortunately, i cant type worth crap
14:27 chitown and this is all "pre editorial"
14:28 viq yeah, I figured
14:28 chitown also, i finally got spellcheck into vim (i know; duh...)
14:28 chitown anyway, yes, those have been fixed
14:28 viq I just got it and started looking through, so not many comments yet.
14:28 babilen chitown: to be honest: Helium introduces such a large number of changes that it would IMHO definitely make sense to target that release with your book. I'd hope that upstream would start focussing on getting every feature in shape after the release rather than continuing with the "moar features"
14:28 Furao joined #salt
14:28 chitown i need to grab the train... back in about an hour...
14:28 babilen cya!
14:29 djstorm joined #salt
14:29 ajolo joined #salt
14:30 H3lraiseR joined #salt
14:31 bhosmer_ joined #salt
14:33 oyvjel joined #salt
14:33 babilen Hello - I am running into infinite recursions when I use mount.mounted to mount NFS filesystems. There seem to be two bugs at play here (cf. #15696): The first problem is that I can't seem to be able to run the state with "test=True" if there is a dependency on a not yet installed service. This is annoying but understandable somehow.
14:35 oz_akan joined #salt
14:35 babilen The more severe problem is that one can't just change the IP of the server referenced in device as this causes *two* entries in the fstab (with the old one first) which then causes problems during the mount and salt can also never satisfy the dependenc{y,ies} of the mount.mounted state
14:35 bhosmer joined #salt
14:35 babilen Does this sound familiar to anybody? I'm currently trying to gather as much data on this problem as I can.
14:36 dccc joined #salt
14:40 smcquay joined #salt
14:41 bhosmer__ joined #salt
14:41 colonD joined #salt
14:41 cztanu__ joined #salt
14:42 spiksius joined #salt
14:43 ntropy chitown: about time for a salt book :)  im esp. interested in the helium release, whats the best way to find out if the book will cover it or not?
14:43 ntropy chitown: if yes, ill just go ahead and buy the book now
14:44 bhosmer joined #salt
14:44 t0rrant_ joined #salt
14:44 elfixit1 joined #salt
14:45 jalbretsen joined #salt
14:45 pviktori_ joined #salt
14:46 [vaelen]_ joined #salt
14:46 Hell_Fire_ joined #salt
14:46 viq < chitown> it was started back in jan/feb. updating it with for the new helium features is something i want to do
14:46 rogst_ joined #salt
14:46 viq < chitown> (there is some pressure to get it done soon)
14:46 viq < chitown> so, i can update with helium, but then it would likely not be out until feb
14:46 viq < chitown> so, pros and cons :)
14:46 viq ntropy: ^
14:46 ze-_ joined #salt
14:47 codekobe____ joined #salt
14:47 scarcry_ joined #salt
14:47 linjan joined #salt
14:47 acabrera joined #salt
14:47 fe92_ joined #salt
14:47 _jslatts joined #salt
14:47 SheetiS joined #salt
14:47 ntropy viq: thanks, i did see that, the answer seems inconclusive though :)
14:47 conan_the_destro joined #salt
14:47 acabrera joined #salt
14:47 kingel_ joined #salt
14:47 analogby1e joined #salt
14:47 tru_tru_ joined #salt
14:48 mikepea_ joined #salt
14:48 EWDurbin joined #salt
14:48 mgw joined #salt
14:48 mephx_ joined #salt
14:48 mikepea_ joined #salt
14:48 EWDurbin joined #salt
14:48 arapaho joined #salt
14:48 geekatcmu_ joined #salt
14:48 viq ntropy: how I understand it is "it wasn't written with helium in mind, I would have and like to update it, but it would take time and make the book come out in february instead of planned december"
14:48 mfournier1 joined #salt
14:49 CaptTofu___ joined #salt
14:49 heresjohnnee joined #salt
14:49 wiqd joined #salt
14:52 Sypher joined #salt
14:52 mr_chris joined #salt
14:52 esogas joined #salt
14:53 Gareth chitown: You should ping forrest, not here now but should be in a bit, he had started writing a Salt book too.  Probably something you guys could collaborate on, if you aren't already.
14:53 ndrei joined #salt
14:53 ndrei_ joined #salt
14:53 garphy joined #salt
14:56 johngrasty joined #salt
14:56 AvengerPenguin joined #salt
14:58 A||SySt3msG0 joined #salt
14:58 bhosmer_ joined #salt
14:59 alanpear_ joined #salt
14:59 blarghmatey joined #salt
15:00 perfectsine_ joined #salt
15:00 rallytime joined #salt
15:00 geekatcmu joined #salt
15:02 claytron joined #salt
15:03 lulzsp00n joined #salt
15:03 pviktori joined #salt
15:04 robins joined #salt
15:04 cpowell joined #salt
15:04 higgs001 joined #salt
15:04 UtahDave joined #salt
15:04 kingel joined #salt
15:05 mgw1 joined #salt
15:05 penguin_dan joined #salt
15:05 heresjohnnee left #salt
15:06 whitepaws joined #salt
15:06 vejdmn joined #salt
15:06 kingel_ joined #salt
15:07 workingcats joined #salt
15:08 maggo Can someon tell me if it is posible to reboot a system within a state and continuing at that statepoint after reboot?
15:09 viq maggo: I believe currently it isn't
15:09 arapaho joined #salt
15:10 maggo viq: ok, read something about startup_states: on the minion configuration, but i don't know if i have courage to do that all the time ;)
15:10 pdayton joined #salt
15:11 Sypher joined #salt
15:12 VSpike What's the difference between cmd.run and cmd.script in states?
15:13 xDamox joined #salt
15:14 VSpike maggo: yeah, I saw that too. http://stackoverflow.com/a/25731608
15:14 ekristen joined #salt
15:15 viq VSpike: "here's a command I want you to run" vs "download this script and run it"
15:15 viq Or was this a tricky question?
15:15 ekristen hey all, question for you — is there any way to reload modules/states to recognize a third party python library is now availble on the server without having to restart the salt minion
15:15 eunuchsocket joined #salt
15:16 ekristen my use case is — I have a state that installs docker and the docker-py pip module, the docker module/state require that docker-py be present on the server for those states to be available but it seems that I have to restart the salt minion for it to see the docker-py module
15:17 ekristen I’d like to dynamically reload if possible, otherwise my current work around is to have the salt-minion restart itself if it indeed installs the docker-py python library
15:19 VSpike viq: that's what I thought ... it could just be the docs confusing me. http://docs.saltstack.com/en/latest/ref/states/all/salt.states.cmd.html#salt.states.cmd.run has an example which shows a script file, but with cmd.script so probably a mistake ...
15:19 geekatcmu joined #salt
15:19 VSpike .. and http://docs.saltstack.com/en/latest/ref/states/all/salt.states.cmd.html#salt.states.cmd.script suggests that "name" can be either URI of a script to download or a command and parameters
15:19 _mel_ joined #salt
15:20 ekristen UtahDave: you around?
15:21 VSpike Oh wait.. << Either "cmd arg1 arg2 arg3..." (cmd is not used) or a source "salt://..." >> does that mean you can supply a source of salt://foo/bar/baz.sh and a name of "whocares param1 param2" to pass param1 and param2 to the script?
15:23 _mel_ Hi. on debian xen kernel right after the boot process, grains.items (executed via /etc/rc.local) shows no entries for virtual or virtual_subtype
15:29 knot joined #salt
15:29 geekatcmu joined #salt
15:29 wendall911 joined #salt
15:30 chitown_ joined #salt
15:30 oeuftete_ joined #salt
15:30 mik3_ joined #salt
15:30 Ymage_ joined #salt
15:30 Sacro_ joined #salt
15:30 goal_ joined #salt
15:31 heise_ joined #salt
15:31 kwmiebach______ joined #salt
15:31 wedgie_ joined #salt
15:31 funzo_ joined #salt
15:31 ashb_ joined #salt
15:31 rblackwe_ joined #salt
15:31 thick_mcrunfast joined #salt
15:31 smkelly_ joined #salt
15:31 grep_awez joined #salt
15:31 Heartsbane_ joined #salt
15:31 twobitsp1ite joined #salt
15:31 seb_ joined #salt
15:31 rofl____ joined #salt
15:31 kiorky joined #salt
15:31 moos3 joined #salt
15:31 whyzgeek joined #salt
15:31 ropes_ joined #salt
15:31 redondos- joined #salt
15:31 Kakwa joined #salt
15:31 ampex joined #salt
15:31 erjohnso_ joined #salt
15:31 rcsheets_ joined #salt
15:31 uzomg joined #salt
15:32 jeblair joined #salt
15:32 gwmngilfen joined #salt
15:32 ubuntu1 joined #salt
15:32 Alan_S_ joined #salt
15:32 LordOfLA|Weechat joined #salt
15:32 __alex joined #salt
15:32 mage_ joined #salt
15:32 gmoro joined #salt
15:32 mattikus joined #salt
15:32 lude1 joined #salt
15:32 rnts joined #salt
15:32 minaguib joined #salt
15:33 irated joined #salt
15:33 andredieb joined #salt
15:33 hareem joined #salt
15:34 hareem hello everyone
15:34 Whissi joined #salt
15:34 xintron joined #salt
15:34 iamtew joined #salt
15:34 masterkorp joined #salt
15:34 apergos joined #salt
15:34 kingel joined #salt
15:35 tyler-baker joined #salt
15:36 UtahDave _mel_: could you open a bug on that?  Thanks!
15:36 nkuttler joined #salt
15:36 UtahDave ekristen: hey!
15:36 ekristen how goes?
15:36 vejdmn joined #salt
15:36 ekristen UtahDave: did you see my question a litle earlier up in the chat?
15:36 UtahDave good!
15:36 UtahDave Yeah, I did
15:36 ekristen UtahDave: is the only way to do it to restart the minion?
15:37 UtahDave so from what I've heard 2014.7 will have a fix so you don't have to restart the minion after installing docker
15:37 UtahDave but in the demos I've done, I had to restart the salt-minion process after installing docker
15:37 UtahDave salt 'docker*' service.restart salt-minion
15:38 ekristen yeah the module/state seems to not see the docker-py module until after reboot
15:38 UtahDave yeah, a salt-minion restart works though, no need to reboot the whole vm
15:38 ekristen yeah bad terminology on my part, I meant salt-minion restart
15:39 ekristen I basically have a service.restart watching pip.installed: docker-py and if it installs then it triggers a restart, and since my startup state is a high state it seems to work, but its pretty dirty IMO
15:39 UtahDave Yeah, I agree.
15:39 UtahDave what version of salt are you on?
15:39 conan_the_destro joined #salt
15:39 JPaul joined #salt
15:39 ekristen um, for my minions I’m on v2014.1.10
15:40 ekristen my master is much further behind
15:40 _mel_ UtahDave: ok
15:40 ekristen I still haven’t figured out a good way to test upgrade on my master so I make sure nothing breaks
15:40 ekristen I tried upgrading to 2014.1.7 once and all my pillar data from git stopped working whenever states ran on my minions :/
15:40 ekristen so I had to revert my master
15:41 UtahDave that sucks
15:43 ekristen yeah, so now I’m a little gun-shy to get my master upgraded
15:44 ndrei joined #salt
15:44 ekristen I really want the new eventing subsystem that is in 1.7 or 1.8 I forget, but its really cool from what I can tell
15:44 geekatcmu joined #salt
15:44 smcquay joined #salt
15:44 ndrei_ joined #salt
15:44 lynxman_ joined #salt
15:44 _mel_ UtahDave: how would i file a bug?
15:44 Hipikat_ joined #salt
15:44 Micromus1 joined #salt
15:45 UtahDave _mel_: click here:   https://github.com/saltstack/salt/issues/new
15:45 maggo VSpike: Jep i will try something like that with changingt the startup_states after the last needed reboot... maybe i can provide a solution for that...
15:45 UtahDave _mel_: thanks for doing that, by the way!
15:45 _Dinde joined #salt
15:45 atoponce joined #salt
15:45 _mel_ UtahDave: sure, no problem. I#d like to have this feature :)
15:46 Kakwa joined #salt
15:46 UtahDave ekristen: was there an issue created for the git pillar breakage?
15:46 moos3 joined #salt
15:46 etw_ joined #salt
15:46 apergos joined #salt
15:46 whyzgeek joined #salt
15:46 ekristen yes, but the guy who looked into it couldn’t seem to re-produce it
15:46 UtahDave _mel_: Yeah, occasionally there are a few missed cases in there.
15:46 goki_____ joined #salt
15:46 gldnspud__ joined #salt
15:47 UtahDave ekristen: do you have a ticket number for that?
15:47 iamtew joined #salt
15:47 ekristen one sec
15:47 ekristen https://github.com/saltstack/salt/issues/14483
15:48 ekristen UtahDave: I’m going to try and get a “testing” environment up of salt in the next week or so, maybe I’ll be able to reproduce it or at the very least be able to test upgrade again
15:49 UtahDave ekristen: that would be great.  If this is still a problem, I'd like to see it fixed for the 2014.7 release
15:49 ekristen I wonder if I could use a single VM to test this
15:50 kiorky joined #salt
15:50 UtahDave ekristen: yeah, probably
15:50 nkuttler_ joined #salt
15:50 rofl____ joined #salt
15:50 higgs001 joined #salt
15:50 ajolo joined #salt
15:50 UtahDave ekristen: also, could you add the related configs to that ticket.  Basically anything to help us recreate the environment where you're seeing this issue
15:51 chitown joined #salt
15:51 nitti_ joined #salt
15:51 UtahDave including the output of    salt-master --versions-report    as well as    salt '<affected minion' test.versions_report
15:53 Whissi joined #salt
15:53 codysoyland joined #salt
15:54 CF839989 joined #salt
15:55 _mel_ UtahDave: i think there are more grain items missing right after the boot process.
15:56 rostam joined #salt
15:56 _mel_ i utahcon_ i think anything below lsb_distrib_release
15:56 pdayton joined #salt
15:57 troyready joined #salt
15:57 UtahDave ok. Yeah put all the info in the issue.  We'll probably get someone on that very soon so it can be fixed for the next release.
15:59 intellix joined #salt
16:00 ekristen UtahDave: will do
16:00 ekristen I’m duping everything into VMware right now
16:00 UtahDave cool
16:01 crane i stored my private key in a pillar. this key is a multiline key (http://dpaste.com/3X9NR8P). i see my key correctly parsed in content_pillar but for some reasons salt has problems parsing it?
16:01 crane anything wrong with the syntax i stored my key?
16:01 crane and don't worry, this is a dummy setup. the key will not run in production :3
16:01 viq crane: try having two less spaces in front of the pasted key
16:02 englishm joined #salt
16:02 crane viq: you mean wrong indention?
16:02 viq yeah
16:02 yano joined #salt
16:03 viq I guess generally try converting to a unit of indentation being 2 spaces, not 4, see if that changes things
16:03 viq Mind you, this could be a cargo cult on my part
16:03 crane viq: hmm... this does not solves my problem. i tried with two spaces and zero spaces
16:04 crane so the indention is lined up with private_key
16:04 viq No, you need an extra indent there
16:04 geekatcmu joined #salt
16:04 tligda joined #salt
16:04 viq You're certain you don't have any tabs in that file?
16:04 spookah joined #salt
16:04 crane viq: yep, i hate tabs and killed them in vim ^^
16:05 viq Good, because so does yaml ;)
16:05 viq But the standard yaml indentation seems to be 2 space
16:06 crane viq: hmm... i still don't get it yet. seeing private_key as indention level 0. what does my key need?
16:06 Cyphus joined #salt
16:07 viq crane: I'd try http://dpaste.com/39FWYMB
16:08 viq Also, what error are you getting?
16:08 crane viq: Rendering SLS "base:tor.key" failed: Unknown yaml render error; line 21
16:09 crane viq: maybe is here an error? - content_pillar: {{ salt['pillar.get']('tor:lookup:private_key') }}
16:09 Ozack1 joined #salt
16:10 babilen crane: Any particular reason why you reference :lookup: directly?
16:10 babilen (sorry for sidetracking)
16:10 crane babilen: every tipp is nice ;)
16:10 crane no not really
16:10 dusel joined #salt
16:10 crane i tried to be align with the salt best practises
16:11 viq I guess it could be even http://dpaste.com/2DK7Z13 and then you'd be calling tor:private_key
16:11 tligda joined #salt
16:11 crane viq: hmm... okay, that solves my problem
16:12 crane err.. wait
16:12 babilen crane: Okay, the "lookup" is used wrongly in a number of places. It stems from formulas that make use of map.jinja which typically contains a map of OS specific settings (e.g. package/service names). This map is a "lookup" table and you would typically use ":lookup:" in your pillars if you want to override values in map.jinja
16:12 fllr joined #salt
16:12 scottpgallagher joined #salt
16:12 crane viq: sorry, it does not solve my problem
16:12 crane babilen: aah, thx :)
16:13 viq crane: could you paste your whole sls file?
16:13 babilen crane: So if you had a "tor" formula it would define a number of settings in map.jinja and that map would merge the "tor:lookup" pillar to allow users to override the settings in the map. Outside of this I consider it an anti-idiom.
16:14 babilen And the pyyaml parser is quite specific about the number of indents that you can use in this situation.
16:14 babilen (which is an idiosyncrasy of the implementation and not part of the standard)
16:15 crane viq: https://github.com/craneworks/tor-salt
16:15 * babilen should finally finish the PR for the "best practices" document regarding the ":lookup:" that is used for the template.
16:16 babilen crane: What is the exact error you get with that?
16:16 * crane would like to be the laboratory rat for the updated document
16:17 crane babilen: Rendering SLS "base:tor.key" failed: Unknown yaml render error; line 21
16:17 babilen Ah, and I see it :)
16:17 geekatcmu joined #salt
16:17 crane hm? :o
16:18 viq crane: also, line 14 - that should be file, not pkg
16:19 tru_tru joined #salt
16:19 crane viq: fixed :)
16:21 englishm joined #salt
16:22 viq crane: does that fix it complaning as well?
16:22 babilen You'd use "- contents_pillar: 'tor:private_key'" (note the 's' in contents_pillar too)
16:22 crane viq: unfortunately not
16:22 babilen https://github.com/craneworks/tor-salt/blob/master/salt/salt/tor/key.sls#L19
16:23 babilen I guess that I am to blame for that
16:23 babilen (due to https://www.refheap.com/89746/)
16:23 patarr joined #salt
16:23 patarr joined #salt
16:24 crane babilen: i remember that paste ^^
16:24 * babilen blushes
16:25 crane but with the additional s the error is still there
16:25 babilen I blame 'content' vs 'contents_pillar'
16:26 babilen crane: Did you also change the value of 'contents_pillar' from {{ salt['pillar.get']('tor:private_key') }} to 'tor:private_key' ?
16:27 babilen https://www.refheap.com/90452
16:27 crane aaand... fixed :)
16:28 nitti joined #salt
16:28 crane babilen: thx :)
16:29 KyleG joined #salt
16:29 KyleG joined #salt
16:30 englishm joined #salt
16:30 chrisjones joined #salt
16:30 crane babilen: is there a difference between contents_pillar and content_pillar or was it just a typo?
16:31 xDamox joined #salt
16:31 viq crane: contents_pillar is a valid keyword, content_pillar isn't
16:32 crane so just a typo ^^
16:32 nkuttler1 joined #salt
16:33 baconbeckons joined #salt
16:33 VSpike I have some software to put in the windows repo, but it has a 32 bit and 64 bit version. Am I right that the directory name is the package name for installation, and that it's otherwise arbitrary?
16:34 VSpike So I could create foo32 and foo64 directories in /srv/win/repo and put each in each
16:34 FeatherKing joined #salt
16:36 Eugene joined #salt
16:37 CF395544 joined #salt
16:41 nkuttler joined #salt
16:42 geekatcmu joined #salt
16:49 c0d3x joined #salt
16:49 c0d3x Hey guys... this game will challenge your problem solving skills: www.codex.sifionsolution.com
16:49 geekatcmu joined #salt
16:52 ajprog_laptop joined #salt
16:56 lulzsp00n joined #salt
16:56 fragamus joined #salt
16:57 linjan joined #salt
17:00 forrest joined #salt
17:05 martinp joined #salt
17:11 geekatcmu joined #salt
17:13 babilen What's the current policy for backports of bug fixes to 2014.1 ?
17:17 kermit joined #salt
17:17 forrest babilen, You can ask for it, I know ryan_lane has had some success doing so
17:18 babilen Wonderful. It's nothing too onerous and would be nice to have. Up until then I simply add my own cherry-picked version to _modules/
17:20 englishm joined #salt
17:20 bhosmer joined #salt
17:21 forrest babilen, yeah I think you can ask basepi via a PR against that specific branch
17:21 forrest but I haven't done it
17:21 gothix joined #salt
17:22 fragamus joined #salt
17:24 SheetiS forrest: Thanks for PR merge on openssh-formula just now
17:24 forrest SheetiS, yeah it looks good, and it already required the pillar anyways, so it wasn't really adding any complexity
17:29 SheetiS It should also be backwards compatible if you don't have 'matches: <matches>' in the pillar, it should render just as before.
17:29 forrest SheetiS, yeah I saw that
17:30 forrest thanks for ensuring that it worked.
17:30 TheThing joined #salt
17:35 SheetiS Is there a cleaner way of doing this in jina when the dict only has 1 entry? -> Match {{ match['type'].keys()[0] }} {{ match['type'].values()[0] }} <-
17:35 SheetiS I really think that looks ugly, but a for loop is wasted there.
17:37 babilen One thing that I dislike about salt's contribution workflow is that people are not encouraged to clean up their commits in the PR by aggressively rebasing them. This always leads to so many "Fixed error in previous commit that is also part of the PR" commits.
17:37 mordax joined #salt
17:37 SheetiS babilen: I'm guilty of that :(
17:38 Ryan_Lane joined #salt
17:38 SheetiS If it's a problem to people, I can behave better ;-)
17:38 babilen SheetiS: It's just that I just had a look at the changes and those could have easily been rebased into a single clean commit.
17:38 babilen You are definitely not the only one and I see it a lot. Sorry, I really didn't mean to pick on you personally.
17:38 forrest babilen, ehh it happens, I don't worry about it much
17:39 SheetiS Understood.
17:39 forrest babilen, some people aren't that familiar with git that are doing the commits
17:39 SheetiS Won't happen again from me.
17:40 babilen But I have worked too much today. Before I open an issue for project A in project B again I'll better head off :)
17:40 englishm joined #salt
17:43 SheetiS babilen: take it easy
17:44 aparsons joined #salt
17:45 egalano joined #salt
17:45 aparsons_ joined #salt
17:48 conan_the_destro joined #salt
17:48 FeatherKing joined #salt
17:49 iggy that's one of the things holding me back from pushing some stuff back upstream
17:49 SheetiS iggy: what's that?
17:49 iggy I want to squash a number of those "fixup X" additional commits into one clean one
17:50 iggy it's not just the committers that don't encourage that though
17:50 iggy it's kind of a byproduct of the way gitfs works (at least for me)
17:50 englishm joined #salt
17:51 n8n joined #salt
17:51 iggy I fork a fomula, add it to our master config, commit changes, notice a bug, another commit, repeat last 2 until things work the way I want
17:51 iggy then I'm left with all those little tail commits
17:52 iggy I'm thinking I need to start developing differently, but we are trying to keep dev/qa/prod as close to each other as possible
17:53 SheetiS I do something similar, but I could still squash things for upstream push.  I just didn't realize people cared really.
17:53 iggy yeah, I can (and intend to), that's why I've got 3+ formulas with changes in my fork that I need to push upstream
17:54 SheetiS Well the longer and further out of sync they get, the more work you create for yourself :D
17:54 iggy I think the only one that's changed is salt-formula
17:55 iggy and it's going to be a pita to cleanup anyway
17:56 kermit joined #salt
17:56 crane babilen: now that you now my state... could you give me a tipp for transforming this state which supports multiple hosts so that i can ship each host with different configs and private keys?
17:58 cpowell joined #salt
17:59 chitown follow up from this morning's convo about the book:
17:59 TheThing joined #salt
18:00 chitown editor and i have (tentaively) agreed to delay the publishing until i can update it for helium
18:01 sctsang joined #salt
18:02 cpowell joined #salt
18:03 dstokes does anyone know how to bust the salt master module cache?
18:07 jeffspeff If i issue a pkg.remove to a windows minion that is offline, will it remove the package when the system starts up?
18:07 SheetiS jeffspeff: it won't ever see the message
18:08 jeffspeff oh, i thought salt-master queued the package and the minion checked in when it was online
18:09 jeffspeff is there anyway to accomplish that?
18:09 perfectsine joined #salt
18:10 kermit joined #salt
18:12 iggy highstate on boot?
18:12 iggy oh wait, you mean manually issuing that command
18:12 iggy ... not that I've seen
18:14 17SAAASJI joined #salt
18:14 1JTAADDOD joined #salt
18:16 A||SySt3msG0 joined #salt
18:18 jeffspeff iggy, i'm completely new to salt, just installed a few hours ago. what is the recommended way to manage off-line minions?
18:19 aparsons joined #salt
18:19 iggy I have no idea... we are using salt in a cloud environment... so no off-line minions
18:20 iggy maybe instead of manually running commands targeted at minions, set everything in the state and just have them do a highstate whenever they can reach the master?
18:20 nkuttler jeffspeff: just a guess, but i think salt-ssh delivers a payload.. maybe you don't need to be networked for that
18:20 nkuttler jeffspeff: but it might try to call back
18:20 aparsons joined #salt
18:20 SheetiS salt-ssh is probably not ideal for windows minions :P
18:21 forrest SheetiS, I think someone actually got it sort of working
18:21 forrest SheetiS, but they were just playing around
18:21 SheetiS using cygwin that was elevated or something?
18:21 forrest SheetiS, I have no idea
18:21 forrest SheetiS, I try to keep my windows knowledge to a minimum, my brain doesn't have that much space
18:21 SheetiS I know I _could_ make it work that way, but .......
18:22 oncallsucks joined #salt
18:23 SheetiS understood.  I often _try_ to disavow windows knowledge, but I also hate waiting on broken things, and our infrastructure does have a little bit of windows to it.
18:23 forrest yeah that's fair
18:27 xDamox joined #salt
18:30 irctc646 joined #salt
18:31 irctc646 \NICK talwai
18:31 irctc646 anyway to access runners (e.g. salt-cp) in custom execution modules? They don't seem to be under the __salt__ dictionary
18:36 ajolo joined #salt
18:37 jeffspeff when making the "init.sls" file, how can i make more than one option for the uninstaller location? for example, the installer could be in %programfiles% or %programfiles(x86)% depending on the system.
18:38 forrest jeffspeff, you'd probably want to check the architecture of the system and do use some jinja if/else statements
18:39 agend joined #salt
18:39 skyler_ I have been reading the documentation on salt reactor, and my mind is a little bit boggled. Is there a way to fire an event after state.highstate completes?
18:39 jeffspeff forrest, can you point me to a doc or something about that?  I'm referencing http://docs.saltstack.com/en/latest/topics/windows/windows-package-manager.html  right now
18:39 skyler_ I want state.highstate to get the minion into a basic state, then use the reactor to fire an event which will cause the minion to then be fed data. So state.highstate makes a generic server, then the reactor should cause it to fetch files and a mysqldump.
18:40 forrest jeffspeff, http://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.jinja.html#jinja-in-states those would be the docs, but I don't know if it works in the repo files...
18:40 forrest I've never tried it
18:40 forrest jeffspeff, shouldn't there be two versions of the package? One x86, one not?
18:41 jeffspeff not always, a lot of msi's package the 32bit and 64bit versions together
18:41 forrest jeffspeff, oh, hmm, well I would try the jinja stuff to see if it works, if not I am not sure how that is handled.
18:41 jeffspeff ok, thanks
18:41 jeffspeff i'll check out that page
18:41 forrest np, sorry I can't provide more than that
18:44 englishm joined #salt
18:51 blarghmatey joined #salt
18:51 SheetiS skyler_: I currently run a reactor item on highstate failures.  I could share part of to help maybe give a basis of what to do.
18:52 SheetiS It's a little bit different than what you are wanting of course, but might give you a better picture of what you can do in a reactor.
18:55 InAnimaTe joined #salt
18:57 blackhelmet joined #salt
18:57 blackhelmet forrest: thanks for accepting that PR
18:57 englishm joined #salt
18:58 SheetiS skyler_: This might be a little confusing but this is how I am matching for returns.  I'll have another paste that shows a little more pertinent stuff soon: https://bpaste.net/show/8a6c72ed3391
18:59 SheetiS show you what things I'm matching on and why
18:59 egalano joined #salt
18:59 fragamus joined #salt
19:01 mapu joined #salt
19:01 talwai joined #salt
19:01 talwai Any way to use 'salt-cp' within an execution module? Should I even be doing this?
19:03 SheetiS skyler_: to add to the paste there, here's what one of the events that would be processed by the reactor would look like: https://bpaste.net/show/3a5ae1ee932a
19:04 SheetiS talwai: what is your end goal with the execution module that is currently not supported?
19:06 ndrei_ joined #salt
19:06 ndrei joined #salt
19:06 skyler_ SheetiS: Thanks! I am looking over what you pasted now.
19:08 aparsons joined #salt
19:11 thistle joined #salt
19:13 root____1 joined #salt
19:19 cpowell joined #salt
19:30 aparsons joined #salt
19:31 thayne joined #salt
19:31 sysadmin75 Hello, does a highstate have to be scheduled on minion, or can it be scheduled on master?
19:32 rjc joined #salt
19:33 forrest sysadmin75, it can be set on the master: http://docs.saltstack.com/en/latest/topics/jobs/schedule.html
19:34 sysadmin75 forrest: I'm actually looking at that page but the Highstates section says to set it on the minion.
19:34 forrest That might just be a lack of documentation, I'd say try it on the master
19:34 forrest if it doesn't work we'll need to update the docs to be more clear.
19:35 sysadmin75 ok I'll give it a try. Thanks.
19:35 forrest np
19:39 InAnimaTe joined #salt
19:40 sysadmin75 forrest: http://pastebin.com/RKriHPyf
19:40 anitak joined #salt
19:40 sysadmin75 Maybe I just have the syntax wrong.
19:41 forrest sysadmin75, try setting it in the pillar maybe?
19:41 forrest sysadmin75, as a second question, why not just push this change to the minion confs?
19:42 claytron joined #salt
19:42 sysadmin75 I was just trying to change one config file instead of many. But I could. I'll try it on the minion and see if that works.
19:43 forrest sysadmin75, gotcha
19:43 forrest sysadmin75, I mean it says it works in the pillar, so you could put it in the pillar data, then sync that out
19:43 forrest I haven't set that up before though
19:43 sysadmin75 I haven't either. I'll mess with the minion config first since the docs do state it can be set there.
19:44 forrest sysadmin75, alright
19:45 sysadmin75 Ha, I'm getting the same error on the minion.
19:45 forrest after a minion restart?
19:45 sysadmin75 yes but the error is in the log
19:46 anitakrueger hey guys…I am trying to run state.orchestrate with a salt.state declaration in it. Based on the result of that state execution, I'd like to deliver a "failed" or "success" message. Unfortunately for salt.state within orchestrate, the "onfail" option doesn't seem to work. Any ideas on how to enhance orchestration further so that actions can be taken based on the result of individual steps in the orchestration sls?
19:46 sysadmin75 I can't be the only person that has this problem. How does everyone else handle pushing out their configs automaticallY?
19:46 sysadmin75 cron?
19:46 forrest usually the schedule system *shrug*
19:47 forrest you're the first person I've seen have an issue with it
19:47 Ahlee sysadmin75: Through scheduling software, yes
19:47 Ahlee forrest: We've had a long lasting request for more granular scheduling
19:47 Ahlee i think they just implemented that, actually
19:47 sysadmin75 Ahlee, as in cron or something similar?
19:47 Ahlee sysadmin75: similar
19:48 stuck_ joined #salt
19:48 sysadmin75 ok thanks
19:48 Ahlee We use active batch for scheduling
19:48 Ahlee It's expensive, bulky, and mostly junk
19:48 Ahlee but it's what the L1s watch, so it's what we used
19:48 forrest Ahlee, schedule is saying version 0.12.0
19:48 sysadmin75 cron will handle what I need.
19:48 forrest Ahlee, unless some functionality was added this release and no one put a note...
19:48 perfectsine joined #salt
19:48 Ahlee forrest: yes, and scheduling "run this ever X" is useless
19:48 forrest heh
19:49 Ahlee i thought i remember seeing cron-like scheduling options were added
19:49 Ahlee *shrug*
19:49 forrest yeah you're thinking of the schedule system
19:49 forrest Ahlee, http://docs.saltstack.com/en/latest/topics/jobs/schedule.html
19:49 Ahlee is that not what we're talking about?
19:50 forrest Ahlee, yeah you're right, there's a small bullet for the fact it was added in 2014.7.0
19:50 forrest sigh
19:50 * Ahlee wins
19:50 Ahlee :)
19:50 forrest but that appears to be for number of jobs, and time... not schedule itself
19:50 sysadmin75 hahe, I see the note now too.
19:50 forrest that's only for the specific format though
19:50 forrest oh well whatever.
19:50 Ahlee ok, so back to the subject
19:51 Ahlee yeah, when we played with schedule we did it on the minions to check in, not on the master
19:51 Ahlee so we just used our scheduling soft to salt -G kernel:Linux
19:52 fragamus joined #salt
19:52 Ahlee we hope to migrate to http://nerds.airbnb.com/introducing-chronos/ sometime in 2015
19:53 scbunn joined #salt
19:53 bmonty joined #salt
19:54 Ahlee anitakrueger: I gave up on orchestrate, and I make multiple calls to the API, parsing the ret values myself
19:55 anitakrueger oh my...ok
19:55 Ahlee for k,v in ret.get('return')[0].iteritmes(): \ if not v.get('result'): handle fail
19:56 anitakrueger I kinda liked it…up until now, where sometimes one of the states in the orchestration sls fails and I can't do anything about it
19:56 anitakrueger apart from requiring it to be successful, for the next state
19:56 Ahlee I do'nt remember why exactly I gave up on orchestrate now, I should revisit
19:57 ndrei joined #salt
19:57 Ahlee oh, i recall now, orchestrate doesn't like non-base roots on 0.17.x
19:57 anitakrueger that's good to know
19:58 ndrei_ joined #salt
20:00 mapu joined #salt
20:09 pfallenop joined #salt
20:12 ekristen joined #salt
20:18 kingel joined #salt
20:19 xDamox joined #salt
20:20 vejdmn joined #salt
20:20 mapu joined #salt
20:22 bmonty is there a way to force the pip state to use python 3?
20:24 forrest bmonty, Did you try the bin_env variable already to specify the python 3 pip?
20:24 bmonty forrest: not sure what that is
20:25 forrest it's an option in the pip state: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.pip_state.html
20:25 bmonty checking, thanks!
20:25 forrest np
20:29 eunuchsocket joined #salt
20:31 mapu joined #salt
20:31 perfectsine joined #salt
20:33 chrisjones joined #salt
20:33 n8n joined #salt
20:34 chrisjones joined #salt
20:38 sysadmin75 left #salt
20:44 bhosmer joined #salt
20:45 bmonty isn't pip_state part of the default states?  I'm getting an error "State pip.installed found in sls python_env is unavailable".
20:46 forrest bmonty, is pip installed?
20:46 forrest on the target machine
20:46 bmonty yes, I have a require on installing the python3-pip package
20:46 bmonty which looks like it is installing correctly from the logs
20:46 forrest bmonty, might have to restart the minion
20:47 bmonty between the install of pip and installing a package with it?
20:47 forrest correct
20:49 jalaziz joined #salt
20:50 TheThing joined #salt
20:53 pdayton joined #salt
20:54 pdayton joined #salt
20:56 pdayton joined #salt
20:57 bmonty so to make it work, you have to install python-pip (i.e. the 2.7 package) and then add "reload_modules: true" to force salt to figure out that pip is installed
20:57 forrest ahh right the reload modules optionf
20:57 forrest *forgot about that
20:57 bmonty I guess the pip state only recognized the 2.7 binary?
20:57 forrest bmonty, it might by default only look for the 2.7 binary to confirm that it's available
20:58 forrest but when you modify the bin_env, it should allow you to use the other one
20:58 forrest bmonty, if you can confirm that, opening a ticket on it would be good so that can be fixed.
20:58 bmonty that's what I'm trying right now...and I'll open a ticket if it's true
20:58 forrest bmonty, cool
20:59 bmonty I'm using salt to provision a vagrant vm, so starting over is easy :)
20:59 forrest heh
20:59 stuck_ joined #salt
21:01 forrest hiding in the shadows timoguin ?
21:01 babilen Is salt doing any regression testing at the moment? I think I found one and would like to implement a test for a bug that is easy to reproduce, but seems to have snuck in at a later time.
21:02 babilen I can't seem to find any and mostly see unit/integration tests.
21:02 timoguin forrest: don't be peekin' me.
21:03 forrest babilen, I don't know, that would probably be best to post on the mailing list
21:03 forrest timoguin, heh
21:03 cpowell joined #salt
21:03 perfectsine joined #salt
21:04 babilen ack
21:05 roolo joined #salt
21:07 bmonty is there a way to force salt to run using python3?
21:08 skyler_ There appears to be a lot of overlap between state.over and state.orch. Which should I use?
21:08 SheetiS skyler_: I think the plan is that eventually overstate will be deprecated, so I use orchestrate
21:09 SheetiS http://docs.saltstack.com/en/latest/topics/tutorials/states_pt5.html#the-orchestrate-runner mentions something to that effect.
21:11 SheetiS bmonty: what OS for your question?
21:11 bmonty SheetiS: ubuntu trusty
21:12 aquinas joined #salt
21:13 SheetiS if it uses the same init script as the osfamily redhat, you can set PYTHON=<path/to/python3> in /etc/default/salt and it sources it
21:13 SheetiS I'll double check for ubuntu
21:13 anitak joined #salt
21:14 bmonty also, I'm working with using salt to provision vagrant
21:14 bmonty I don't think vagrant installs the ubuntu package
21:15 bmonty nvm, it does install packages
21:16 nitti joined #salt
21:16 rallytime joined #salt
21:18 skyler_ SheetiS: Good to know, thanks!
21:18 UtahDave joined #salt
21:19 kballou joined #salt
21:21 SheetiS bmonty: it looks like the debian init script (which is what I expect ubuntu uses) does source /etc/default/$name (where name would be salt-master or salt-minion), but I don't see where that init script references the python version.
21:22 bmonty :q
21:22 SheetiS it might work how I said still, but I don't have something ubuntu handy to test.
21:22 bmonty I'm going to test it now
21:22 SheetiS ok
21:22 pdayton joined #salt
21:22 bmonty I think if PYTHON is in the environment it should work
21:22 goodwill joined #salt
21:23 SheetiS let me know how it works for you so I can say for sure in the future, and good luck!
21:23 bmonty will do, and thanks for the help
21:24 bhosmer joined #salt
21:26 bmonty SheetiS: it won't work without changing the "salt-minion" script.  It has "#! /usr/bin/python" at the top.
21:26 SheetiS hmm :(
21:26 babilen Salt can run with python3 ?
21:27 pdayton joined #salt
21:27 bmonty babilen: actually, I think the answer is no
21:28 bmonty since I just tried starting it with python3 and it immediately throws an exception because the salt package isn't install for python3
21:28 bmonty by the ubuntu package
21:28 bmonty but my goal is to build a python 3 virutalenv in a vm
21:28 UtahDave Salt can't run under python3 quite yet
21:29 bmonty so the work around is to install pip for 2.7 and 3 (2.7 is necessary for salt to load the pip state)
21:29 bmonty then you can add reload_modules to a state
21:30 bmonty and then use "bin_env" on any pip state that you want to use python 3
21:31 bmonty here is what I'm doing: http://dpaste.com/2A7ZXDF
21:31 Ryan_Lane joined #salt
21:31 nitti joined #salt
21:33 Ryan_Lane joined #salt
21:41 englishm joined #salt
21:44 laxity joined #salt
21:44 tharkun joined #salt
21:45 Hell_Fire_ joined #salt
21:49 englishm joined #salt
21:50 ryu joined #salt
21:50 ryu haroo
21:51 Cyphus joined #salt
21:52 vbabiy joined #salt
21:52 ajolo joined #salt
21:53 cromark joined #salt
21:54 bhosmer joined #salt
21:58 ryuhei joined #salt
21:59 pdayton joined #salt
22:02 ryuhei could anyone help beginner questions please?
22:02 forrest ryuhei, Feel free to ask, if someone knows the answer they'll usually respond with information or some documentation links
22:02 iggy you may have just missed everybody, but you should probably ask anyway
22:03 ryuhei thanks.!
22:04 ryuhei when i download/install the salt release.. from github.. it always say Salt: 2014.1.0 regardless of if it was 1.5 or .10. or .11
22:04 forrest how are you doing the download/install?
22:04 ryuhei i downloaded the tar.gz file of those version and python setup.py install
22:05 forrest what OS are you on?
22:05 ryuhei ubuntu 14.04 64bit
22:06 forrest why don't you just install the package
22:06 ryuhei 3.13.0-24-generic #47-Ubuntu
22:06 forrest http://docs.saltstack.com/en/latest/topics/installation/ubuntu.html
22:06 ryuhei because cherrypy didnt pass the argument of cloud.profile properly..
22:07 iggy my guess would be that you're ending up with 2 versions (possibly one from package, one from tarball, etc)
22:07 ryuhei and reading salt-cloud is merged with salt.. i tried to use newer version.
22:07 ryuhei i d think so if it's not fresh installed OS.
22:07 forrest the version from the ppa has salt-cloud in it
22:07 ajprog_laptop joined #salt
22:08 ryuhei right.. i have another os running with package version installed.
22:09 ryuhei i couldn't get cloud.profile working
22:09 ryuhei it passes the argument as a list of characters... not strings
22:10 ryuhei I'm trying to create droplet on digital ocean, but it successfully run the command.
22:10 ryuhei but say if i want to creat a droplet named "test", it will end up creating droplets of t,e,s,t
22:10 pdayton joined #salt
22:12 englishm joined #salt
22:12 kermit joined #salt
22:12 ryuhei i couldnt figure out why so i wanted to use newer version and hoping it's fixed.
22:15 Ryan_Lane joined #salt
22:15 ryuhei curl -sikv https://ipaddress:port/run -H "Accept: application/x-yaml" -d client='runner' -d username='someuser' -d password='somepass' -d eauth='pam' -d fun='cloud.profile' -d prof='ubuntu_512MB_nyc3' -d names='ab,cd'
22:15 ryuhei this would get me a, b, c, d droplets..
22:15 ryuhei supposedly it should create ab and cd
22:18 ryuhei if i can resolve this, i can stick with the package version on ubuntu.
22:18 UtahDave ryuhei: try   names='["ab", "cd"]'
22:18 ryuhei ok will do that now
22:21 ryuhei no luck.. it just created a
22:21 perfectsine joined #salt
22:32 mordonez joined #salt
22:33 bhosmer joined #salt
22:33 StDiluted joined #salt
22:34 StDiluted Question regarding directory structure for state definitions: can i nest the directories like /srv/salt/states/core/thing/init.sls
22:34 StDiluted where /srv/salt/states is my root
22:35 StDiluted and then call that as core.thing
22:35 UtahDave yep
22:35 StDiluted I thought so. it’s not working so I’m trying to figure out why
22:36 chrisjones joined #salt
22:36 perfectsine_ joined #salt
22:37 StDiluted so I have /srv/salt/states/core/check_mk_agent/init.sls
22:38 StDiluted and in my top.sls I have base: ‘*’: core.check_mk_agent
22:38 StDiluted but I get No matching sls found for core.check_mk_agent
22:41 StDiluted ah wait
22:41 StDiluted it looks like it was a missing ‘
22:41 StDiluted lol
22:41 StDiluted thanks!
22:41 babilen StDiluted: Use make that "base: '*': core.check_mk_agent" rather than "base: ‘*’: core.check_mk_agent"
22:42 babilen s/Use//
22:43 StDiluted ah yeah I have single quotes, my client changed it though
22:43 StDiluted irc client that is
22:44 babilen That's not good. (and where were you missing a single ' ?)
22:45 StDiluted {% if pillar.get('libsodium', {}).get(install_from_source') %}
22:46 * babilen hasn;t seen that before. Another example of "Just paste the entire SLS files to a pastebin and errors can be found much faster" :)
22:46 babilen Glad you spotted i though
22:46 StDiluted yeah, I like pastebin for that as well
22:46 StDiluted it’s just a bunch of separate files
22:47 StDiluted so it’s more pain that it’s worth
22:47 StDiluted still getting a no matching sls
22:47 StDiluted but less render errors now :)
22:48 StDiluted so weird
22:48 StDiluted hang on let me pastebin some stuff and maybe you can see where my issue is
22:49 babilen use http://refheap.com or http://paste.debian.net please
22:49 StDiluted no gist?
22:49 babilen Well, anything but pastebin.com :)
22:50 StDiluted hehe
22:50 StDiluted I think i figured it out actually
22:50 babilen even better
22:50 StDiluted i had some include statements
22:50 StDiluted that didnt have core.thing.whatever
22:50 StDiluted just thing.whatever
22:50 StDiluted fixed those and the highstate is running
22:50 StDiluted ;)
22:50 StDiluted confused myself with the nesting
22:56 SheetiS joined #salt
23:06 StDiluted is there a way to make file.managed not require a source_hash?
23:09 JKT_ joined #salt
23:10 babilen I've never used that. Be more specific please.
23:11 StDiluted https://gist.github.com/dginther/d820cba132259997f379
23:11 babilen StDiluted: I mean: What happens if you don't provide it?
23:11 StDiluted the file.managed section is failing because
23:11 StDiluted i posted the fail message as well
23:12 StDiluted Comment: Unable to determine upstream hash of source file http://download.dnscrypt.org/libsodium/releases/libsodium-0.7.0.tar.gz
23:12 babilen "If the file is hosted on a HTTP or FTP server then the source_hash argument is also required"
23:12 StDiluted ugh
23:13 StDiluted damn
23:14 UtahDave StDiluted: Yeah, the source_hash is required. I think I'd like to add an option to make it optional.
23:14 StDiluted that would be nice
23:17 babilen You could use git.latest and pull it by tag from GH :-/ -- although that wouldn't be their released tarball.
23:17 StDiluted for instance the SHA-256 for libsodium is a DNS call
23:17 StDiluted dig +dnssec +short txt <file>.download.libsodium.org
23:17 babilen (assuming tarball == "git archive $tag")
23:17 StDiluted kind of a pian to get
23:17 StDiluted pain*
23:18 babilen No https?
23:18 Schmidt joined #salt
23:19 catpig joined #salt
23:19 babilen Ah, broken certificate
23:20 bezaban ~hm
23:21 babilen https://download.libsodium.org/libsodium/ looks better
23:21 StDiluted so it returns a hash but the has is in double quotes. am I going to have to chop the quotes as well?
23:23 blarghmatey joined #salt
23:23 babilen https is not an option?
23:23 StDiluted i suppose it is, though this is failing on a few different files and not sure it’s an option for all of them
23:24 babilen I'm also not sure if "HTTP server" does not apply to https too. That might very well be the case.
23:24 StDiluted still wants a source hash
23:24 babilen ack
23:27 englishm joined #salt
23:28 englishm joined #salt
23:29 StDiluted what the hell, this is ridiculously hard
23:29 babilen StDiluted: Alternatively: Use cmd.run and a suitable curl call and then archive.extracted
23:29 StDiluted is there a way in jinja/sls file to get the result of a dig command inside the sls?
23:29 jbub joined #salt
23:30 babilen cmd.wait rather (with unless that checks if the file is there)
23:30 babilen salt['cmd.run'](....)
23:30 StDiluted yeah, I can do that, but that’s on the minion, I think the hash gets checked on the salt master
23:31 StDiluted anyway, time for me to go home
23:31 babilen You would need the has in that case anymore.
23:31 babilen *hash
23:31 StDiluted yeah, I guess i could just grab the file instad of using the file,managed
23:31 StDiluted ugh
23:31 StDiluted that is a pain in the butt
23:32 StDiluted thanks for your input, babilen
23:32 babilen You have two solutions now that you can try. I agree that it would be best if you could simply use https without having to specify the source_hash, but that is how it is right now.
23:33 babilen Three solutions if you count using git.latest with a tag of your choice.
23:37 babilen The dig call in salt['cmd.run']('dig ...') is a bit problematic as you will have to get the actual hash and jinja might make that unnecessarily hard. You can, however, probably get pretty far with the replace builtin filter.
23:37 smcquay joined #salt
23:38 anitak joined #salt
23:40 englishm joined #salt
23:41 babilen I would probably either implement that state in pure Python/PyObjects or write an external module that provides the "get hash from dns" functionality. (cf. http://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.jinja.html#calling-salt-functions and http://docs.saltstack.com/en/latest/ref/modules/ )
23:42 babilen I don't think that blindly downloading tarballs without specifying at least *something* to verify that what you get is what you want is bad.
23:42 babilen s/bad/good :)
23:43 teebes joined #salt
23:45 jbub joined #salt
23:49 jalaziz joined #salt
23:56 fllr joined #salt
23:59 ndrei joined #salt

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