Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-12-08

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

All times shown according to UTC.

Time Nick Message
00:07 aqua^c joined #salt
00:20 AdamSewell joined #salt
00:22 N-Mi_ joined #salt
00:23 N-Mi joined #salt
00:23 zmalone joined #salt
00:25 snarfy^ joined #salt
00:27 pbx_ joined #salt
00:31 brianfeister joined #salt
00:32 tbaker57 joined #salt
00:32 shaggy_surfer joined #salt
00:32 bhosmer_ joined #salt
00:33 abednarik joined #salt
00:35 baweaver joined #salt
00:38 zmalone joined #salt
00:39 hacks joined #salt
00:39 Striki joined #salt
00:41 otter768 joined #salt
00:43 Guest44611 joined #salt
00:44 yidhra joined #salt
00:45 JDiPierro joined #salt
00:45 iggy forrest: LA, not SF
00:45 forrest iggy, huh?
00:46 forrest Man I was just about to go to the gym, I don't think I've said anything in here all day, lol
00:46 forrest what about LA?
00:46 iggy outdated conversation... I've been on the road since last week
00:46 iggy carry on
00:46 forrest iggy, ahh okay
00:46 forrest Alright I'm bouncing for the evening, have a good one
00:46 ahammond why not SF? :)
00:46 forrest $$$$$
00:46 iggy ^
00:47 forrest Lemme know when your company wants to pay me to work from my house since I do the same thing as in the office ;)
00:47 iggy LA is bad enough... I can't hang with the FB-illionaires
00:47 mikepea joined #salt
00:47 forrest 'ahh yes this thing broke and I fixed it, here are the steps' 'Ahh yes *takes off headphones* this thing broke and I fixed it, here are the steps'
00:47 forrest iggy, You just need your daily kool-aid injections
00:47 ahammond forrest I wish they'd pry their heads out of the sand on this one... we'll see.
00:47 forrest ahammond, As do I, it's making it a real pain in the ass to find a job for a cool company
00:48 ahammond we've had 3 hiring reqs open for over a year now.
00:48 ahammond devops: apparently impossible to hire for.
00:48 forrest lol
00:48 iggy invent teleportation
00:48 forrest Are you in SF?
00:48 iggy then we'll all be set
00:48 forrest ahammond, ^
00:48 ahammond we're in Menlo Park
00:49 forrest I've been scouring for devops jobs, no luck on remote gigs.
00:49 forrest ahh
00:49 moderation joined #salt
00:49 forrest Well, I should rephrase that. I've had luck, but not the kind I'm interested in.
00:49 forrest It is kind of hard to do remote if you aren't already working with the tools
00:50 forrest but how many companies aren't at this point? I don't know. I had an interesting discussion with the CEO of sentry on Twitter earlier today
00:50 ahammond forrest yeah, that's not an easy sell.
00:50 terratoma we're hiring in SF and in Irvine
00:50 forrest ahammond, If I was willing to move to SF, I could have had a dozen or two dozen jobs by now
00:50 ahammond forrest yeah, I hear you.
00:50 forrest So the people you want are out there, they just aren't going to pay 5k for some shithole apartment without a dishwasher/washer and dryer and step over human feces ever day
00:51 forrest *every
00:51 ahammond forrest I just moved to Santa Clara. I've been with the company for 2 years now, so I can WFH a day or two every couple of weeks, but... yeah. :(
00:51 forrest I can imagine, are you commuting by train?
00:51 forrest or BART?
00:51 forrest I can't remember if it goes out there
00:52 ahammond forrest nope, no bart. I just drive. it's 25 min without traffic. about 40 min most days.
00:52 ahammond fortunately there are books on tape.
00:53 forrest Yeah that makes it a little more bearable, still losing 1-1.5 hours per day
00:54 forrest I told a place in downtown seattle (I don't have a car, so it's ~1 hour each way by bus) that if they wanted to include bussing in my hours I'd consider it
00:54 forrest they weren't happy about that
00:54 terratoma i would defend SF and say i dont step over human feces every day but... i just did, nevermind
00:54 forrest terratoma, lol, yeah my buddy at Twitter considers it a good day when he doesn't have to step over crap on the way to work
00:54 terratoma they are part of the problem!
00:54 terratoma that is a rough block
00:55 forrest Yeah I've been around there, not a fan.
00:55 geekatcmu "Maybe you should consider hiring some competent managers who can, you know, 'manage' work-from-home people appropriately."
00:56 forrest Alright I've gotta go to the gym. ahammond if you find any remote devops jobs for companies that aren't banks/investment firms, and aren't building dog photo filter apps, let me know.
00:56 forrest geekatcmu, It's pretty easy if the people you hire are responsible adults about their job
00:57 moloney joined #salt
00:57 geekatcmu "responsible adult" is a great target to aim for, but as a hiring criteria ... HR generally hates it.
00:57 baweaver joined #salt
00:57 * forrest shrugs
00:57 Nazzy joined #salt
00:57 forrest Anyways, gym time, see you guys tomorrow
00:57 whytewolf I have found HR typically has issues with competent also.
00:57 jjasinski joined #salt
00:58 geekatcmu "Sounds like discrimination to me!"
00:58 geekatcmu I wish I were kidding
00:59 whytewolf I know you are not. when i worked at the paper. we had issues with trying to fire people that ... cost the company a ton of money in mistakes that could have been avoided.
01:01 moloney How do you guys manage package updates? I don't like enabling automatic updates, but managing every single package update through salt seems a bit tedious. I wish apt could easily export/import a list of package versions.
01:02 _JZ_ joined #salt
01:03 whytewolf moloney: if you don't want to manage every single pkg. by it's self. then pkg.uptodate would be your option. personally I perfer each pkg on it's own so i keep a tight grip.
01:06 iggy constant redeploy so updates aren't necessary
01:06 Sucks joined #salt
01:07 moloney whytewolf: Yeah I was hoping for some middle ground. Ideally I would upgrade all package to latest before testing for deployment and after deploying I would carefully install updates in some limited fashion to contain breakage. New deployments would then use the tested set of package versions.
01:08 geekatcmu moloney: pkg.uptodate and only ever talk to your own repositories.  Best of both worlds.
01:08 whytewolf ^^
01:09 whytewolf keep your own repos. and do you testing on another setup. like a enterprise shop
01:10 geekatcmu Basically, treat your OS exactly the same way you treat your product: release management, testing, etc.
01:10 moloney geekatcmu:  That is a good idea, will definately look into that approach
01:12 Diaoul joined #salt
01:14 RobertChen117 joined #salt
01:18 brianfeister joined #salt
01:18 jjasinski_ joined #salt
01:19 abednarik joined #salt
01:23 dendazen joined #salt
01:24 onovy joined #salt
01:24 moloney geekatcmu: This looks like a useful tool for your approach to package management (at least for Debian derivatives) -> http://www.aptly.info/
01:25 geekatcmu yeah.
01:25 geekatcmu We've got an Agile Storage plugin for it.
01:25 geekatcmu We'd be using it for our internal repo management already but ... the team that was working on setting that all up basically just got fired, so ...
01:31 RobertChen117 anyone know what skip_suggestions does ? (pkg.installed)
01:32 Eugene No idea. I know apt has a "suggested packages" feature, probably related
01:35 cyborg-one joined #salt
01:38 falenn joined #salt
01:39 aqua^c joined #salt
01:42 ahammond geekatcmu the problem isn't the managers. heck, the only reason my manager cares if I'm at my desk rather than at my kitchen table is the politics of us being seen around the office and working. The problem is the co-owners. But... we're workign on that.
01:42 digitalhero joined #salt
01:45 shaggy_surfer joined #salt
01:48 digitalhero joined #salt
01:49 favadi joined #salt
01:50 Sucks joined #salt
01:53 digitalhero joined #salt
01:54 spuder joined #salt
01:56 hightekvagabond joined #salt
01:59 RobertChen117 joined #salt
02:02 Lionel_Debroux joined #salt
02:04 solidsnack joined #salt
02:05 bfoxwell joined #salt
02:06 bfoxwell joined #salt
02:07 ajw0100 joined #salt
02:12 digitalhero joined #salt
02:15 RobertChen117 joined #salt
02:16 burp_ joined #salt
02:22 ITChap joined #salt
02:24 catpigger joined #salt
02:27 yomilk joined #salt
02:31 nszceta joined #salt
02:32 otter768 joined #salt
02:39 boargod hi folks.  Trying to move into a salt-master from some prototype salt-ssh stuff, but hitting an error: "get() got an unexpected keyword argument 'merge'". This line for example https://github.com/saltstack-formulas/consul-formula/blob/master/consul/map.jinja#L3
02:47 racooper joined #salt
02:50 anmolb joined #salt
02:56 boargod Ugh, disregard, my minion here was installed without the proper repo and got the ancient Ubuntu version
02:59 clintberry joined #salt
03:00 nszceta left #salt
03:00 dyasny joined #salt
03:12 oida joined #salt
03:17 jaybocc2 joined #salt
03:22 bhosmer joined #salt
03:26 Sucks joined #salt
03:31 subu joined #salt
03:33 subu hey - anyone know how AES key is initialized so that it is unique and random each time
03:38 Sucks joined #salt
03:46 evle joined #salt
03:47 larsfronius joined #salt
03:55 oida joined #salt
04:06 yomilk joined #salt
04:10 colegatron joined #salt
04:14 anmolb joined #salt
04:24 yomilk_ joined #salt
04:35 Sucks joined #salt
04:36 yomilk joined #salt
04:37 n8n joined #salt
04:41 armguy joined #salt
04:48 nethershaw joined #salt
04:54 quasiben joined #salt
05:01 LondonAppDev__ joined #salt
05:03 oida joined #salt
05:08 spuder joined #salt
05:10 rdas joined #salt
05:11 hasues joined #salt
05:11 hasues left #salt
05:14 alvinstarr joined #salt
05:14 ekristen joined #salt
05:15 solidsnack joined #salt
05:17 malinoff joined #salt
05:20 drawsmcgraw joined #salt
05:20 burp_ joined #salt
05:25 jaybocc2 joined #salt
05:25 PeterO joined #salt
05:27 oida joined #salt
05:30 bhosmer_ joined #salt
05:36 rawzone joined #salt
05:43 ageorgop joined #salt
05:49 anmolb joined #salt
05:51 av_ joined #salt
05:57 favadi joined #salt
05:59 Guest63 joined #salt
06:00 calvinh joined #salt
06:03 solidsnack joined #salt
06:13 cyteen joined #salt
06:20 Joren__ joined #salt
06:20 ageorgop joined #salt
06:21 oida joined #salt
06:28 otter768 joined #salt
06:29 solidsnack joined #salt
06:31 stanchan joined #salt
06:35 zer0def joined #salt
06:46 Guest63 joined #salt
06:47 felskrone joined #salt
06:50 solidsnack joined #salt
06:51 yomilk joined #salt
06:54 aqua^c joined #salt
06:56 ageorgop joined #salt
07:04 brianfeister joined #salt
07:06 sfxandy joined #salt
07:18 jaybocc2 joined #salt
07:18 burp_ joined #salt
07:25 oida joined #salt
07:31 bhosmer_ joined #salt
07:33 slav0nic joined #salt
07:37 KermitTheFragger joined #salt
07:38 mattiasr joined #salt
07:40 tbaker57 joined #salt
07:42 Grokzen joined #salt
07:42 jhauser joined #salt
07:49 larsfronius joined #salt
07:52 AndreasLutro joined #salt
07:56 dkrae joined #salt
07:58 oida joined #salt
07:58 elsmo joined #salt
08:02 stanchan joined #salt
08:11 solidsnack joined #salt
08:14 Granjow joined #salt
08:15 Granjow Good morning! I'm trying to find out, in a formula, if node --version is 'v5.1.0', but did not find a good way yet. Atm. I have an unless requisite with: - test $(node --version) = 'v5.1.0'
08:16 Granjow This does not work though, probably because the shell started by python is not the one I expect.
08:16 mortis Granjow: cant you use pkg.version?
08:18 Granjow mortis: This returns 5.1.0-1 when I call it with salt '*' pkg.version node
08:18 kshlm joined #salt
08:19 Granjow I guess the -1 comes from checkinstall
08:19 Granjow Jinja or salt don't directly have support for storing stdout of a call in a variable?
08:20 mortis no, but you could call it via salt
08:20 Granjow Like how?
08:20 felskrone joined #salt
08:21 mortis {% set something = salt['pkg.version']('node') %} <- something like this
08:22 mortis they discuss something similar here https://groups.google.com/forum/#!topic/salt-users/6r3gCPyzOyg
08:26 Kusen_ joined #salt
08:27 Granjow Ahh, thanks mortis! This works. Just needs a bit of fine-tuning now to get rid of the -1. http://codepad.org/4WMAFY9c
08:27 mortis Granjow: awesome :)
08:28 solidsnack joined #salt
08:29 otter768 joined #salt
08:30 thalleralexander joined #salt
08:33 mortis guess its dirty to use |replace("-1","")
08:33 mortis :D
08:34 jamesp9 joined #salt
08:35 dgutu joined #salt
08:36 thalleralexander joined #salt
08:38 dgutu joined #salt
08:52 hightekvagabond joined #salt
08:53 brianfeister joined #salt
08:53 Grokzen joined #salt
08:53 Granjow mortis: Haha, that's what I've come up with as well. Dirty or not, it works! And there are no regular expressions, so ...
08:55 mortis oh you can use match also
08:56 linjan_ joined #salt
08:56 ingslovak joined #salt
08:56 Granjow match is a salt function?
08:57 mortis no, jinja
08:57 mortis google jinja regex
08:58 Granjow Is there a reason some Jinja expressions use -%} at the end, instead of just %}?
09:00 babilen There is (whitespace control)
09:01 Granjow Ah. Thanks babilen. Did not find this in the template overview where all possible {% syntaxes are listed. Except for this one :D
09:02 eseyman joined #salt
09:03 keimlink joined #salt
09:04 fredvd joined #salt
09:07 Granjow mortis: Found it. A bit too much magic, will stick with replace for now :)
09:09 zerthimon joined #salt
09:15 acsir joined #salt
09:16 Granjow Is there a way to add something to the information that is shown when a state is applied?
09:16 brianfeister joined #salt
09:20 pppingme joined #salt
09:20 oida joined #salt
09:21 ashmckenzie joined #salt
09:23 Xevian joined #salt
09:23 hemebond Granjow: Like what?
09:24 hemebond I can't think of a way but I've never tried.
09:24 N-Mi_ joined #salt
09:26 elsmo joined #salt
09:27 Granjow hemebond: Debug information mainly, e.g. 'node already at version 5.1.0, not installing'
09:27 hemebond You can change the log level, yes. By default if no changes are made, nothing is reported.
09:27 hemebond It's a change you can make in the master config.
09:27 yomilk joined #salt
09:28 Granjow But now I have another problem anyway that my minion returns with non-zero exit state, and I have no idea why. So probably anyway a good time to take a look at the master config :)
09:31 s_kunk joined #salt
09:31 s_kunk joined #salt
09:32 rotbeard joined #salt
09:33 GreatSnoopy joined #salt
09:36 chiui joined #salt
09:36 Granjow Did not help :(
09:38 berto- joined #salt
09:40 Granjow Ah. If I have node.sls containing include: node, then it includes itself quite recursively. Any way to tell it to include from node-formula?
09:40 hemebond I don't think so. Like Python includes, your modules (states, etc) need to be unique.
09:41 Granjow Aye. Thanks.
09:47 fxhp joined #salt
09:47 mohae joined #salt
09:47 oida joined #salt
09:50 ziro` joined #salt
09:51 N-Mi_ joined #salt
09:56 basepi joined #salt
09:57 esharpmajor joined #salt
09:57 mullein joined #salt
09:57 chutzpah joined #salt
10:10 oida joined #salt
10:11 yomilk joined #salt
10:12 colegatron joined #salt
10:18 shorty_mu joined #salt
10:20 RobertChen117 joined #salt
10:20 shorty_mu Hi, since a few days salt-bootstrap fails under Centos 7 with the error: :https://repo.saltstack.com/yum/rhel7/python-zmq-14.7.0-1.el7.x86_64.rpm: [Errno -1] Package does not match intended download."  Looks like something with the RPM/Repo-Build went wrong.
10:20 RobertChen117 how to detect resource dependancies dead loop ?
10:20 anmolb joined #salt
10:21 LondonAppDev joined #salt
10:23 terinjokes joined #salt
10:30 otter768 joined #salt
10:32 giantlock joined #salt
10:34 om2 joined #salt
10:34 mohae joined #salt
10:35 denys joined #salt
10:35 larsfronius joined #salt
10:36 om2 joined #salt
10:36 yomilk joined #salt
10:40 justanot1eruser joined #salt
10:41 om2 joined #salt
10:42 linjan_ joined #salt
10:49 |Trend| joined #salt
10:51 amcorreia joined #salt
10:52 malinoff_ joined #salt
10:57 inad922 joined #salt
10:59 rim-k joined #salt
10:59 rim-k hello everyone
11:01 rim-k in a autoscaling context, what will be the best and safe way to handle/accept new minions keys
11:01 rim-k ?
11:03 mohae joined #salt
11:10 Grokzen joined #salt
11:19 favadi joined #salt
11:20 GothAck joined #salt
11:21 justanotheruser joined #salt
11:21 aurynn joined #salt
11:24 oida joined #salt
11:27 N-Mi_ when a repository is unreachable, salt doesn't want to install packages (because apt update failed). Is there a way to bypass this error with salt ?
11:28 apergos joined #salt
11:29 apergos I notice that git clone of the salt repo is hellaciously slow these days, has anyone considered a repack or some other cleanup?
11:32 bhosmer joined #salt
11:37 shorty_mu joined #salt
11:38 bhosmer joined #salt
11:43 oida joined #salt
11:45 dendazen joined #salt
11:46 sixninetynine joined #salt
11:47 shanemhansen joined #salt
11:47 jaybocc2 joined #salt
11:47 FTfz joined #salt
11:51 blue0ctober joined #salt
11:56 burp_ joined #salt
11:57 irctc856 joined #salt
11:58 irctc856 hi there. I am trying to get started with salt. I am wondering what the best method is to distribute private python code that resides in a git repository. any pointers/ideas would be greatly appreciated
12:00 armguy joined #salt
12:01 aqua^c joined #salt
12:01 Jimlad joined #salt
12:02 om joined #salt
12:03 burp_ joined #salt
12:05 burp_ joined #salt
12:09 burp_ joined #salt
12:09 blue0ctober joined #salt
12:10 dgutu joined #salt
12:10 sixninetynine joined #salt
12:10 favadi joined #salt
12:10 yomilk joined #salt
12:12 Ch3LL joined #salt
12:12 dkrae joined #salt
12:13 mortis irctc856: sound like you should start here https://docs.saltstack.com/en/latest/topics/tutorials/starting_states.html
12:14 Corey joined #salt
12:15 burp_ joined #salt
12:20 burp_ joined #salt
12:23 burp_ joined #salt
12:23 mdupont joined #salt
12:28 evle joined #salt
12:31 otter768 joined #salt
12:33 om joined #salt
12:33 shorty_mu Hi folks, I'm struggling with casting from "int" to "string" and unicode (i guess?). More here: https://gist.github.com/bemeyert/f55c6ff6b28fd62c1d71 If anyone could help how to get rid of the "u" I would be _very_ happy. Cheers
12:33 bhosmer joined #salt
12:34 bhosmer_ joined #salt
12:36 yomilk joined #salt
12:36 jaybocc2 joined #salt
12:38 nidr0x joined #salt
12:42 mortis shorty_mu: maybe {{ somethingsomething | safe }} works?
12:42 bluenemo joined #salt
12:43 shorty_mu Already tried that to no effect. I always get the bloddy "u". This happens when using "~" to cast from int to string.
12:43 mortis tried join?
12:44 shorty_mu Hmm, I need to look this up. Gimme a moment...
12:47 JDiPierro joined #salt
12:49 daniel joined #salt
12:50 shorty_mu mortis: Nope. "Jinja error: 'int' object is not iterable" when I use it like that: "'storage_size': storage_size | join + 'M'"
12:53 Guest13659 hi @all, i'm new to salt and we actual evaluate the migration from ansible to salt-ssh...as we have host files for ansible but i can not get it to work with salt so i do an simple test ans generate an 2-line file with an [test] header an an server entry...the ansible roster does not find anything: [DEBUG   ] Matched minions: {'[test]': {}}
12:53 Guest13659 i've called it as following: salt-ssh --verbose -l all --roster-file /home/daniel/test * test.ping --roster ansible
12:54 shorty_mu mortis: Stupid me. Use as a list and it doesn't fail: 'storage_size': [ storage_size ] |join + 'M'  But I still get the bloody "u":  u'498M'
12:54 mortis shorty_mu: nice :)
12:54 mortis but not so nice
12:54 mortis hehe
12:55 shorty_mu What? ahh, yeah...
12:55 mortis can you use [0] in jinja? i duno
12:55 mortis like in python
12:56 shorty_mu Why should I? There is only one element in the list. Wouldn't make sense, would it?
13:02 job joined #salt
13:02 job good afternoon
13:02 job I'm trying to iterate over a yaml list in a pillar: https://github.com/NLNOG/ring-salt/blob/master/pillars/users/former_users.sls
13:03 job but according to the master log, 'former_users' is not defined, and if i replace 'former_users' with pillar.get('former_users', []) it gives other errors
13:04 job i am terribly new to salt
13:04 mrbobbytables shorty_mu: could you truncate the last character then cast it as an int?
13:06 shorty_mu mrbobbytables: Without the "M" and left as "int" (you can see in the Gist that it is an int) it works. I just get the 498.
13:06 shorty_mu ... but concatenating the "M" (meaning stringifying the int) breaks things.
13:07 mrbobbytables oooh derp, sorry missed that above
13:07 shorty_mu np ;)
13:08 mrbobbytables oh wow yeah, read that totally backwards somehow and going from string to int
13:09 job is this illegal approach? https://github.com/NLNOG/ring-salt/blob/master/pillars/users/former_users.sls
13:09 job error I get is:      Comment: Unable to render top file: Jinja variable 'None' has no attribute 'items'
13:09 teryx510 joined #salt
13:09 akhter joined #salt
13:10 rim-k any idea on my concern? please :)
13:11 ggoZ joined #salt
13:13 teryx510 joined #salt
13:14 shorty_mu I threw away the concatenation (and stringification) and use it in the template like this: {{ varnish.storage_size }}{{ varnish.storage_size_unit }} So Problem not solved, but circumvented. Cheers all
13:14 akhter joined #salt
13:16 toastedpenguin joined #salt
13:20 mrbobbytables job: I'm still pretty new to salt as well, but it looks like you have a state in a pillar
13:20 TooLmaN joined #salt
13:21 mortis afaik you cant use pillar.get in a pillar.sls
13:21 job i've updated my pillar file
13:21 job https://github.com/NLNOG/ring-salt/blob/master/pillars/users/former_users.sls
13:22 job this seems rather unwieldy
13:22 job any hints on how i can do away with the duplication?
13:22 irctc856 left #salt
13:22 mortis move them to the state
13:22 mortis only have the usernames in the pillar
13:22 mortis if they're gonna be the same
13:22 toastedpenguin joined #salt
13:23 job but seems i then cannot use the 'users' formula
13:23 job as the 'users' formula expects this type of data in the pillar
13:23 babilen job: You could iterate over a list and generate that structure
13:24 babilen for user in $LIST_OF_USERS ...
13:24 AdamSewell joined #salt
13:24 b0r joined #salt
13:25 favadi joined #salt
13:25 job where do i define that list
13:25 job in the pillar?
13:27 tpaul joined #salt
13:30 Jimlad joined #salt
13:33 giantlock joined #salt
13:33 abednarik joined #salt
13:36 lompik joined #salt
13:37 subsignal joined #salt
13:37 denys joined #salt
13:42 mapu joined #salt
13:43 mkjgore joined #salt
13:45 dendazen joined #salt
13:45 JDiPierro joined #salt
13:46 babilen job: In the file (it simply is an easier way to generate that data .. I would probably generate it in Python (render the pillar in Python and return a suitable dictionary), but it should be absolutely doable in jinja too
13:47 babilen {% for user in ['user1', 'user2', ....] %} .... {% endfor %}
13:50 subsignal joined #salt
13:51 favadi joined #salt
13:52 thalleralexander joined #salt
13:54 TyrfingMjolnir joined #salt
13:54 job that makes for super long lines and thus is fugly
13:54 job messes with git blame :)
13:56 quasiben joined #salt
14:01 melrakki joined #salt
14:03 racooper joined #salt
14:03 toastedpenguin joined #salt
14:05 DammitJim joined #salt
14:07 yomilk joined #salt
14:09 thalleralexander joined #salt
14:10 favadi joined #salt
14:11 JDiPierro joined #salt
14:11 bhosmer joined #salt
14:13 acsir joined #salt
14:14 viq joined #salt
14:16 protoz joined #salt
14:18 mpanetta_ joined #salt
14:18 mrbobbytables actually, as I'm still pretty new to salt -- if anyone that is more familiar with salt best practices / styling wouldn't mind looking over a formula of mine and letting me know if I'm on track it'd be  greatly appreciated :) - https://github.com/mrbobbytables/salty-docker
14:19 mrbobbytables I intend on getting everything going in salt, but want to make sure I'm not doing anything really dumb from the get-go
14:19 numkem joined #salt
14:20 marsdominion joined #salt
14:20 marsdominion joined #salt
14:21 toastedpenguin joined #salt
14:21 JDiPierr_ joined #salt
14:22 mage_ I have a question regarding organization: is it better for example to have single pillar users.sls and do the filtering in the state, or pillars users/webapps.sls, users/admin.sls, and a single state file ?
14:23 AndreasLutro mage_: single state, different pillars is usually the way to go
14:24 mage_ ok.. thanksç
14:31 Ztyx joined #salt
14:31 Ztyx left #salt
14:31 toastedpenguin joined #salt
14:32 otter768 joined #salt
14:32 shiriru joined #salt
14:34 oida joined #salt
14:35 emaninpa joined #salt
14:37 lorengordon joined #salt
14:41 giannello joined #salt
14:44 giannello hi folks - is it normal for pillars to be refreshed every time even if just a simple module is called? I have a pretty big encrypted pillars list, and just running a "grains.setval" will make the salt-master process them, slowing down everything
14:44 protoz joined #salt
14:44 protoz joined #salt
14:45 shorty_mu left #salt
14:47 jaybocc2 joined #salt
14:48 cpowell joined #salt
14:49 cpowell joined #salt
14:50 mitsuhiko joined #salt
14:58 favadi joined #salt
14:58 nethershaw joined #salt
14:59 toastedpenguin joined #salt
15:01 zmalone joined #salt
15:03 jaybocc2 joined #salt
15:03 andrew_v joined #salt
15:06 favadi joined #salt
15:06 timoguin joined #salt
15:10 akhter_1 joined #salt
15:10 blw0rm joined #salt
15:10 spuder joined #salt
15:12 timoguin_ joined #salt
15:15 kaptk2 joined #salt
15:17 _JZ_ joined #salt
15:18 ocdmw joined #salt
15:19 dijit yep, they'll be processed every time
15:19 dijit same with the templating system.
15:31 gtd joined #salt
15:41 bhosmer joined #salt
15:43 zerthimon joined #salt
15:45 ericof joined #salt
15:45 spuder joined #salt
15:46 clintberry joined #salt
15:48 chiui joined #salt
15:50 Tanta joined #salt
15:50 abednarik joined #salt
15:52 fredvd joined #salt
15:53 cyteen joined #salt
15:55 hasues joined #salt
16:00 Sucks joined #salt
16:02 clintberry joined #salt
16:05 AlberTUX joined #salt
16:05 AlberTUX hi ppl
16:06 AlberTUX has anyone used status.process?
16:06 rim-k heya
16:06 rim-k "curl -Ssl" is missing in this page : https://docs.saltstack.com/en/latest/topics/installation/debian.html
16:06 rim-k at Run the following command to import the SaltStack repository key:
16:06 AlberTUX i'm trying to match a process by its name but it fails
16:07 mariusv joined #salt
16:07 mariusv joined #salt
16:08 djstorm joined #salt
16:10 AlberTUX like this:
16:10 AlberTUX https://gist.github.com/bert-o-git/19bcef84354dfb4faad9
16:11 AlberTUX is there a specific way to pass the name of the process to match?
16:12 colegatron joined #salt
16:13 jaybocc2 joined #salt
16:13 Joker75 joined #salt
16:16 drawsmcgraw Uh.... I just got this when trying `dockerng.running()` -> Failed to create new container: 'Client' object has no attribute 'create_host_config'
16:16 drawsmcgraw Anyone seen that before?
16:18 oida joined #salt
16:19 PeterO_ joined #salt
16:21 ziro` joined #salt
16:23 Grokzen joined #salt
16:23 zerthimon joined #salt
16:24 abednarik joined #salt
16:29 mariusv_ joined #salt
16:30 favadi joined #salt
16:30 cberndt joined #salt
16:33 drawsmcgraw Apparently you need to install python-docker
16:33 otter768 joined #salt
16:33 shiin joined #salt
16:33 drawsmcgraw Except not any through the normal package distributions. You need the latest from Pip
16:33 drawsmcgraw http://stackoverflow.com/questions/33270253/salt-dockerng-virtual-returned-false
16:33 brianfeister joined #salt
16:38 denys joined #salt
16:38 drawsmcgraw Do I misunderstand, or is the lazy_loading of modules not seem to work the way it should?
16:39 drawsmcgraw For example, I have states to install python-pip, followed by states that pip.installed() things
16:39 drawsmcgraw But pip.installed() *always* fails on a new machine because it says pip doesn't exist
16:39 drawsmcgraw It does, it's just that it didn't exist before the Salt run
16:40 drawsmcgraw The result is I have to run some state file two or more times because Salt's not loading the modules I'm asking for.
16:40 acsir joined #salt
16:41 favadi joined #salt
16:43 ivanjaros joined #salt
16:44 linjan__ joined #salt
16:45 dabb joined #salt
16:45 ivanjaros hi. i am trying to use git.clone and pass --depth=1 --branch=master arguments to it via opts parameter but i am getting some python errors.
16:46 mr_const joined #salt
16:47 winsalt joined #salt
16:48 cpowell_ joined #salt
16:49 whytewolf drawsmcgraw: https://docs.saltstack.com/en/latest/ref/states/#reloading-modules reload_modules isn't called everytime a state is run. would slow down the run way to much.
16:49 cpowell_ joined #salt
16:50 drawsmcgraw whytewolf: Ah! Thank you so much! I was not aware of that capability.
16:54 indispeq hi! I was wondering if it was possible to use ext_pillar to read pillar-data from a file that is also used to source bash Environment variables (e.g.: export FOO=bar, export ABC=def, etc...), and to then look for values using salt['pillar.get']('FOO') ?
17:01 akhter joined #salt
17:03 Bryson joined #salt
17:04 favadi joined #salt
17:06 ziro` joined #salt
17:09 crd joined #salt
17:12 lexter joined #salt
17:12 emaninpa joined #salt
17:13 Rumbles joined #salt
17:14 Rumbles hi, can anyone suggest how to set a users umask using salt?
17:16 justanotheruser joined #salt
17:16 writtenoff joined #salt
17:16 zmalone edit shell profile files for that user using file.managed or set it system wide in /etc/profile, /etc/bashrc, etc.
17:17 blw0rm joined #salt
17:18 emaninpa joined #salt
17:18 Rumbles thanks zmalone
17:19 emaninpa joined #salt
17:27 thalleralexander hi is it possible to execute states into a folder or a chroot?
17:28 shaggy_surfer joined #salt
17:28 oida joined #salt
17:28 aw110f joined #salt
17:28 zmalone I'm not sure that you can chroot an already running salt, but you should be able to run salt within a chroot
17:28 zmalone or you could treat the folder like any other from outside of the chroot
17:29 zmalone (I have not done this)
17:30 tpaul Does anyone know of any known issues using zeromq 4.1.3 or py-zmq 14.4.1?
17:31 ericof joined #salt
17:34 JDiPierro joined #salt
17:40 netcho joined #salt
17:41 ageorgop joined #salt
17:42 hightekvagabond joined #salt
17:48 AlberTUX any idea what do i need to match a process with status.process? https://gist.github.com/bert-o-git/19bcef84354dfb4faad9
17:49 emaninpa joined #salt
17:49 emaninpa joined #salt
17:51 Fiber^ joined #salt
17:55 toastedpenguin joined #salt
17:58 pppingme joined #salt
18:00 nafg joined #salt
18:02 programmerq joined #salt
18:02 oida joined #salt
18:05 thalleralexander joined #salt
18:07 nethershaw joined #salt
18:08 akhter joined #salt
18:18 pjs joined #salt
18:19 cpowell joined #salt
18:24 toastedpenguin joined #salt
18:26 bhosmer_ joined #salt
18:26 abednarik joined #salt
18:28 jgee joined #salt
18:31 toastedpenguin joined #salt
18:31 VSpike joined #salt
18:34 otter768 joined #salt
18:34 oida joined #salt
18:37 toastedpenguin joined #salt
18:40 Ashald_ joined #salt
18:40 Ashald_ Hi guys
18:41 Ashald_ I hope you don't mind a question. I need to create an AWS EC2 Cloud Formation stack from the Salt. I'm not allowed to put any credentials onto the Minions, so I'm thinking whether it's possible to create a runner that will execute boto_cfn modules/states from within the master?
18:42 Ashald_ The goal is to create a stack in response to a particular event
18:42 Ashald_ That will be tied to runner via the reactor
18:42 Ashald_ And triggered through salt-api
18:42 Ashald_ presumably with pepper :)
18:43 Ashald_ Is it something reasonable or I'm trying to do something insane? :)
18:44 whytewolf Ashald_: instead of that. why not salt the master and turn it into a "minion" since it sounds like you can put the creds on the master
18:44 Ashald_ you mean like install minion on the same machine?
18:44 whytewolf yes
18:44 Ashald_ That's might be an option
18:45 Ashald_ I even considered creating a small pool of 'trusted minions' that can have these credentials
18:45 Ashald_ Though in this case IDK how can I trigger only 1 instance of a job
18:46 Ashald_ So only 1 stack will be created instead of <number-of-trusted-minions>
18:46 Ashald_ Is it possible at all? I mean, from within the reactor run a single instance of some job on a minion from the pool (obviously filtered by grains/etc)?
18:47 ageorgop1 joined #salt
18:48 aboe joined #salt
18:49 whytewolf not sure what you are asking. each event is seperate for each minion that sends it.
18:49 elsmo joined #salt
18:49 Ashald_ So I have a reactor that listens for a particular event
18:50 Ashald_ and then it should trigger creation of a CF stack
18:50 Ashald_ Let's say I have pool of 5 'trusted minions' that have AWS credentials
18:50 Ashald_ so I need to runt the job to launch the stack once
18:50 Ashald_ but not on all of the minions
18:50 Ashald_ is is possible?
18:52 akhter joined #salt
18:52 whytewolf Ashald_: drop the thought of "many trusted"
18:53 whytewolf it isn't going to help you
18:53 Ashald_ So just keep a 1 'trusted minion'?
18:53 whytewolf the best thought i would think with a many trusted approch would be to randomly select one
18:54 Ashald_ Hm... I see, thanks
18:54 hightekvagabond joined #salt
18:54 Ashald_ now I guess how salt-cloud works...
18:54 Ashald_ does it create instances from its own process or it delegates instance creation to minions?..
18:55 whytewolf salt cloud has it's own api that runs on the master
18:55 marsdominion joined #salt
18:55 marsdominion joined #salt
18:56 shaggy_surfer joined #salt
18:56 Ashald_ I wasn't able to find that out on the net, but maybe there is a way to create extensions for the salt-cloud? So I can put my CF stuff into salt-cloud and then use it as though it was built into it?
18:58 marsdomi_ joined #salt
18:58 Ashald_ hm.., I found a doc how to extend Salt-Cloud by adding custom providers
18:58 whytewolf I don't think salt-cloud is extendable at all currently
18:58 Ashald_ Maybe I can make my CF stuff into a custom provider...
18:58 jaybocc2 joined #salt
18:58 Ashald_ I assume that I can trigger salt-cloud commands from reactor :)
18:58 marsdomi_ joined #salt
19:00 bhosmer__ joined #salt
19:01 forrest joined #salt
19:01 whytewolf Ashald_: if you are reading anything on readthedocs it most likely is useless.
19:01 whytewolf the docs for salt-cloud have been moved to docs.saltstack.com
19:02 whytewolf and salt-cloud has gone through a ton of reworking since it was on readthedocs
19:02 Ashald_ Actually I'm reading https://docs.saltstack.com/en/develop/topics/cloud/cloud.html
19:02 marsdominion joined #salt
19:02 marsdominion joined #salt
19:02 Ashald_ And the only thing that's unclear for me so fat is how to register my cloud driver module once it's done
19:02 Ashald_ Except putting it into Salt-Stack's dirs within virtualenv/wherever it's installed
19:04 whytewolf Ashald_: just having it in the python module directory should do it. cloud doesn't have a nice _cloud directory like modules states and grains do yet.
19:04 whytewolf salt/cloud.clouds
19:04 whytewolf err
19:04 whytewolf salt/cloud/clouds
19:05 drawsmcgraw joined #salt
19:06 saltstackbot [reddit-saltstack] Backend storage (or strategy) for binary files https://www.reddit.com/r/saltstack/comments/3vyrn6/backend_storage_or_strategy_for_binary_files/ - 2015-12-08 - 19:04:55
19:07 cpowell joined #salt
19:08 Ashald_ thanks!
19:09 cwyse joined #salt
19:12 ajw0100 joined #salt
19:14 numkem joined #salt
19:14 huddy joined #salt
19:15 hightekvagabond joined #salt
19:17 tkharju joined #salt
19:19 job what am i doing wrong in this command:
19:19 job root@master01:~# salt '*.ring.nlnog.net and not *.infra.ring.nlnog.net' self.ping
19:19 job No minions matched the target. No command was sent, no jid was assigned.
19:19 job i have hosts that are XXX.ring.nlnog.net and YYY.infra.ring.nlnog.net
19:20 whytewolf job: salt -C '*.ring.nlnog.net and not *.infra.ring.nlnog.net' self.ping
19:20 job thanks
19:23 racooper joined #salt
19:27 linjan joined #salt
19:28 perfectsine joined #salt
19:28 quasiben joined #salt
19:28 lorengordon joined #salt
19:28 FreeSpencer If using module.run if something returns false is there a way to not have it come up as an error?
19:30 abednarik joined #salt
19:31 winsalt joined #salt
19:31 ziro` joined #salt
19:35 mapu joined #salt
19:39 brianfeister joined #salt
19:43 Bryson joined #salt
19:45 solidsnack joined #salt
19:48 akhter joined #salt
19:51 JDiPierro joined #salt
19:52 JDiPierro joined #salt
19:54 giantlock joined #salt
19:54 giantlock_ joined #salt
19:55 giantlock joined #salt
19:55 cyborg-one joined #salt
19:57 denys joined #salt
19:58 hightekvagabond joined #salt
19:59 hightekvagabond joined #salt
20:00 Destreyf joined #salt
20:02 keimlink joined #salt
20:02 abednarik joined #salt
20:04 cwyse joined #salt
20:04 jaybocc2 joined #salt
20:07 s_kunk joined #salt
20:07 s_kunk joined #salt
20:07 aidalgol joined #salt
20:09 akhter joined #salt
20:10 akhter joined #salt
20:15 racooper is there a way to make pkgrepo.managed create the .repo file (for yum on CentOS) if it doesn't exist?
20:15 mariusv joined #salt
20:15 mariusv joined #salt
20:17 emaninpa joined #salt
20:20 Bryson joined #salt
20:23 abednarik joined #salt
20:24 job who here sets up munin & munin-nodes with salt?
20:24 beardedeagle joined #salt
20:24 job if i look at this pillar example and the code in munin-formula, i don't have the feeling that all 'hosts' entries are created whenever you add in a pillar that a bunch of nodes is a munin-noe
20:25 job munin-node*
20:25 lunarlamp joined #salt
20:25 akhter joined #salt
20:26 mariusv joined #salt
20:31 cberndt joined #salt
20:31 portaluser joined #salt
20:35 otter768 joined #salt
20:35 mariusv joined #salt
20:37 penguinpowernz joined #salt
20:38 emaninpa joined #salt
20:39 mariusv_ joined #salt
20:40 penguinpowernz joined #salt
20:42 akhter racooper: Why not use file.managed?
20:45 racooper I didn't want to have to have another file to manage? the yumpkg.modrepo execution module can create the the repo file if it doesn't exist,  not sure why the state module wouldn't do that as well.
20:45 lemur joined #salt
20:47 FreeSpencer Anybody using the iptables.append and having -table always set to filter even if its set to mangle?
20:51 Sucks joined #salt
20:51 ziro` joined #salt
20:55 whatapain racooper: if you're managing the repo and it doesn't exist it should automatically create it.  is it not doing that?
20:55 conan_the_destro joined #salt
20:56 racooper not using states it doesn't. but nothing in documentation for the state says it should, either. the execution module docs specifically say that if the file doesn't exist it will be created.
20:56 shaggy_surfer joined #salt
20:57 marsdominion joined #salt
21:00 jerryc joined #salt
21:01 marsdominion joined #salt
21:01 marsdominion joined #salt
21:02 whatapain racooper: not sure i understand, the whole point of repo.managed is that it will create the repo files if they don't exist... this is how i'm doing it for rsyslog https://gist.github.com/mshirley/02387082084a93a82388
21:02 spuder_ joined #salt
21:02 whatapain that creates /etc/yum.repo.d/rsyslog.repo in centos
21:02 oida joined #salt
21:03 racooper I agree that it should work, but I haven't been able to make that happen yet. if the file doesn't exist, the state fails.
21:03 rim-k joined #salt
21:03 racooper for me.
21:03 whatapain racooper: checkout my gist and see if it's a syntax issue
21:03 lemur joined #salt
21:03 bhosmer_ joined #salt
21:04 GreatSnoopy joined #salt
21:05 racooper it looks very close, I have some additional fields: https://gist.github.com/racooper/465a5f5012f1d4d69578
21:07 AdamSewell joined #salt
21:07 whatapain racooper: grains['os'][0] doesn't looks right
21:08 racooper that was just my latest attempt. I have local repos where I've tried the same thing with no jinja variables, and it still fails.
21:08 whatapain racooper: i'd drop your minion into debug mode and see what's happening
21:09 hightekvagabond joined #salt
21:09 racooper I'll test it out again. I've moved on to using a repo file and that seems to work for now.
21:10 whatapain *nod*, pkgrepo.managed should work though
21:10 onovy hi. is it possible to specify for "watches" is reload: true applied? I want to do service restart for example if package is upgrade (service.running: watch: pkg: nginx) but when config changes i want just reload (service.running: watch: file: config)
21:10 onovy for which "watches"
21:13 perfectsine joined #salt
21:13 onovy https://gist.github.com/onovy/61670414405144e84549
21:13 oida joined #salt
21:17 aron_kexp joined #salt
21:23 aw110f joined #salt
21:24 * whatapain crosses his fingers that he can use a file glob in an unless: statement
21:26 solidsnack joined #salt
21:27 lorengordon can anyone help me understand the deprecation warning on cmd.run, re: python_shell? i'm trying to figure out how to rewrite the state to avoid the warning and avoid future breakage.
21:28 whytewolf whatapain: unless is just a cmd.retcode with python_shell=True. so a file glob by it's self is nothing. need to have something return false.
21:28 lorengordon the warning says to use python_shell=True, or to switch to cmd.shell, but cmd.shell doesn't exist as a state module and cmd.run doesn't have a python_shell parameter (and it ignores kwargs)
21:29 catpig joined #salt
21:29 pppingme joined #salt
21:29 lorengordon if there's an issue or user group post that i missed explaining it all, i would appreciate a link
21:30 Grokzen joined #salt
21:32 whatapain whytewolf: i briefly misunderstood unless, i was thinking of creates
21:33 perfectsine joined #salt
21:34 jaybocc2_ joined #salt
21:35 whytewolf lorengordon: cmd.shell is a cmdmod exacution module function. what exactly are you passing to the cmd.run state that is causing deprecation?
21:35 perfectsine joined #salt
21:35 akhter Anyone ever do a for loop inside an if condition before?
21:35 jgee \join cordova
21:36 lorengordon whytewolf: basically any cmd.run state appears to generate the deprecation warning, i didn't think it was anything exotic
21:37 whytewolf lorengordon: I'm not getting it and use cmd.run states pretty extensivly. and am running 2015.8.3
21:38 akhter Seems that order is off.  (if condition) (for loop) some stuff (end for loop) (end if condition)  But for some reason salt complains that it's expecting an endfor at the end instead of ending the if condition.
21:38 akhter Makes no sense....
21:38 lorengordon yeah, i just created a more basic state and it didn't happen. let me try some variations...
21:38 lorengordon it looks like it may be coming from the map.jinja import
21:39 lorengordon ahh yes, i had a cmd.run in there to set a jinja variable to the programfiles path
21:39 cberndt joined #salt
21:39 whytewolf ahh
21:42 lorengordon yep, switching that to cmd.shell works. thanks for the gut check. probably headed off some crazy pointless troubleshooting path! :D
21:44 subsignal joined #salt
21:44 lorengordon but cmd.run with python_shell=True still displays the dep warning. kind of annoying.
21:44 whytewolf thats odd
21:46 lorengordon cmd.run with python_shell=False also prints the warning. meh.
21:47 lorengordon now i'm not sure what the 'secure' invocation is supposed to be
21:55 nafg joined #salt
22:01 cberndt joined #salt
22:02 foundatron is there an equivalent in saltstack land to the Puppet Forge, or is it just the official formulas hosted at https://github.com/saltstack-formulas/
22:03 foundatron Coming from puppet, I always check there bfore writing my own module. Just wondering if there is something similiar in Salt that i've missed
22:03 blw0rm joined #salt
22:04 zmalone my experience with saltstack-formulas was more broken stuff and one off formulas then you might expect in the chef supermarket, or presumably puppet forge
22:04 zmalone I think the lack of dependency declarations and formula versioning doesn't help
22:05 aw110f_ joined #salt
22:05 oida joined #salt
22:06 Sucks joined #salt
22:06 foundatron huh
22:08 akhter joined #salt
22:08 hightekvagabond joined #salt
22:09 solidsnack joined #salt
22:10 colegatron joined #salt
22:10 Ryan_Lane yeah, that's the spot
22:10 Ryan_Lane I've always found forge/berkshelf/galaxy/etc to be mostly useless
22:10 Ryan_Lane reusable config management code is a falacy :)
22:12 Ryan_Lane what you end up with is a poorly implemented DSL using pillars (or hiera, variables, or whatever the analog is in whichever system) that doesn't do exactly what you need
22:14 zmalone berkshelf isn't a public thing, it's just a tool for doing dependency management and stuff, whether internal or external
22:14 zmalone while there is a salt-kitchen, I don't think the berkshelf tool can be salted because you don't have version or dependency metadata at this time
22:15 zmalone supermarket is what they call their public reusable community code site, and it's honestly pretty good these days,
22:16 whytewolf honestly my experence with all of them [including salt-formulas] is that they get to be bloated, not handle all the edge cases I might need, or are just someones attempt that only handles there own situation.
22:16 whytewolf s/there/their
22:17 ecdhe Ryan_Lane, I'd never heard that so succinctly.  I had considered taking to the blogosphere to complain about the status quo in reusable config.
22:17 ecdhe It is oddly comforting to mentally just stop trying.
22:18 ecdhe I mean, none of my efforts were globally reusable anyway, so I said to myself, "I will refactor this formula one day!"
22:18 ecdhe Now, I can take all that refactoring out of my bug tracker as WONTFIX.
22:19 ecdhe The dude lives!
22:20 whytewolf the dude abides
22:22 sfxandy joined #salt
22:23 Ryan_Lane ecdhe: http://ryandlane.com/blog/2014/10/08/config-management-code-reuse-isnt-always-the-right-approach/
22:25 pcn If I am changing pillars in a way that would break my states until my states are also updated, is there a flag I can set that will tell minions to chill and spin until the change is made?
22:26 abednarik joined #salt
22:29 ecdhe Ryan_Lane, well said.  Various upstream formulas like users-formula are quite handy to spin up a quick little one-off.
22:29 ecdhe One problem, even with a perfect upstream formula, is that in order to stay DRY, you often need one formula to actually get data from the pillar of the other.
22:30 ecdhe Which, once again, works only for users who need both formulas.
22:30 clintberry joined #salt
22:31 ecdhe I was thinking that it could be good for salt to standardize on a few different pillar schemas, so that things like user names, ssh keys, etc could be stored in one place.
22:32 jeffspeff for some reason i get this error https://gist.github.com/jeffclay/e3542daa9e4d2728fc78 on some of our minions. i have tried uninstalling and re-installing the salt-minion but it continues to give this error. most of the systems return the package list properly, only a few random systems are doing this. any ideas?
22:33 whytewolf ecdhe: I'm not sure if it should be salt, or the formula comunity that decideds that. as I don't use formulas so would be hard pressed to find a reason to use salts "defined standrd" where as with formulas that does make sense
22:34 titilambert How often the develop branch doc is compiled and updated on web site ?
22:34 burp_ joined #salt
22:35 jfindlay titilambert: all three branches of the docs are rebuilt every day
22:35 zmalone jeffspeff: did you try restarting salt after installing?
22:35 otter768 joined #salt
22:35 titilambert jfindlay: thanks !
22:36 jeffspeff zmalone, yes, restarting the minion
22:36 jeffspeff i get the same error on the minions having this issue if i try to use pkg.install
22:36 zmalone well, the first step is running pkg.list_pkgs to check for matching dependency versions
22:36 zmalone (I'll get my coat)
22:37 zmalone that's really weird, and I can't replicate in my environment, but I'm strongly guessing it has to do with having a mismatched package somewhere.
22:37 jeffspeff zmalone, that's the command that returns the error i showed in the link
22:37 zmalone yeah, that was meant as a joke
22:37 jeffspeff oh
22:37 jeffspeff :P
22:37 jeffspeff it's been a long day
22:38 whytewolf could use a salt-call -l trace on the minion see if something funny turns up
22:39 jeffspeff so, if it's a mismatched package, then i should be able to determine the the culprit by removing packages from my winrepo and using pkg.refresh_db then pkg.list_pkgs until it works right?
22:40 whytewolf windows? are you using a odd distrobution of pythong?
22:40 whytewolf s/pythong/python
22:40 jeffspeff whytewolf, no
22:40 zmalone Sorry, I assumed *nix, I have no idea if pkg will work there.
22:41 jeffspeff the only python app that i can think  of is salt
22:41 jeffspeff zmalone, pkg works with windows
22:42 jeffspeff i've just noticed this issue here lately on a few minions i've needed to work on... we have about 300 windows workstations running salt like a charm
22:42 whytewolf pkg uses the win_repo pkgs iirc [I'm not a windows guy so take everything said about it with an ocean of salt]
22:42 jeffspeff lol
22:43 whytewolf jeffspeff: maybe force a db refresh
22:44 whytewolf [who knows maybe winrepo was broke for a bit during a db update]
22:45 Ryan_Lane ecdhe: yeah, doing quick PoC stuff with the formulas works well
22:45 Rumbles joined #salt
22:46 ecdhe It was great for PoCs on the raspberry pi until salt stopped working on raspbian.
22:46 cberndt joined #salt
22:46 portaluser Hey folks, I'm running into an issue with an orchestrate runner.  I've turned a salt command into a runner (salt.function) but when I use the runner it throws results in my face with a "Data failed to compile:"
22:47 burp_ joined #salt
22:47 abednarik joined #salt
22:48 pppingme joined #salt
22:48 whytewolf portaluser: some things don't jiva with that sentance. salt command and runner don't make sense. salt.function doens't turn something into a runner. a runner is a master side code. salt.function just calls exacution modules on a minion that matches the tgt
22:49 whytewolf portaluser: salt.runner is for running a runner
22:50 whytewolf either way that is neither here nor there. please post a bit of code to examine. if you can
22:51 portaluser sure: here's the runner definition: https://gist.github.com/anonymous/b666a57fd959f12e515a
22:51 burp_ joined #salt
22:53 portaluser which I think should be equivalent to this: https://gist.github.com/anonymous/3bd7fd957d690320de33
22:55 whytewolf portaluser: you are correct. that should be the same. and does look correct. however to be safe. try surrounding sec_osqueries in single quotes
22:56 shaggy_surfer Has anyone seen this one, I just upgraded to 2015.8.3 last week and am seeing this in the logs on ubuntu 14.04.3 LTS, I am using official salt repo packages,  "[salt.loaded.int.module.tls][ERROR   ][20300] You should upgrade pyOpenSSL to at least 0.14.1 to enable the use of X509 extensions"
22:56 whytewolf never mind that shouldn't be needed
22:56 whytewolf portaluser: what version of salt?
22:56 shaggy_surfer I looked on salt issues and found this one https://github.com/saltstack/salt/issues/28872
22:56 saltstackbot [#28872]title: bad error/warning message, 'You should upgrade pyOpenSSL' | I have found a few salt.module calls that do not work for Debian based OS's, complaining ``You should upgrade pyOpenSSL`` to a version that is one version higher than is installed on the box.  For example, I have tested with the following distro versions and version of python-openssl/pyOpenSSL:...
22:57 shaggy_surfer bit ot
22:57 shaggy_surfer but it's not the same issue exactly
22:58 whytewolf shaggy_surfer: sorry not having that issue. but i also am not using the systems openssl. I pip installed 0.15.1
22:58 portaluser whytewolf: Looks like I'm seeing the same issue.  Changes: <minion-id>: Data failed to compile
22:59 portaluser whytewolf: Still showing as succeeded, but I believe it's causing issues with a returner I'm trying to use
22:59 jgelens joined #salt
23:00 burp_ left #salt
23:00 burp_ joined #salt
23:00 whytewolf portaluser: I don't see a function for pillar_queries in osquery. is this a custom copy of osquery?
23:00 burp_ left #salt
23:01 burp_ joined #salt
23:01 portaluser whytewolf: it is
23:02 whytewolf humm. I'm pretty sure the _modules work with orch.
23:02 portaluser whytewolf, I'm on Lithium, 2015.5.5
23:03 whytewolf I'm also wery of the fact you kept calling it a runner. how is that module setup?
23:03 portaluser whytewolf: It's an orchestration runner, trying to run it using `salt-run state.orchestrate orch.osquery`
23:04 whytewolf orchestrate is a runner. but the items in orchestrate are still states. also I was asking about osquery
23:05 iggy whytewolf: they don't
23:05 portaluser gotcha, the pillar_queries() is the custom chunk
23:05 iggy you have to set extension_modules
23:06 iggy or module_dirs
23:06 iggy can't remember off the top of my head
23:06 portaluser whytewolf: the extra defined function is here: https://gist.github.com/anonymous/279350ea9a0accf2a530
23:06 bfoxwell joined #salt
23:07 portaluser iggy: is that a master config option?
23:07 iggy si
23:07 whytewolf https://docs.saltstack.com/en/latest/ref/configuration/minion.html#module-dirs
23:08 portaluser Right now I'm just distributing by putting them in /srv/salt/_modules and doing a saltutil.sync_modules
23:08 iggy apparently it's both, but you want it in the master conf file
23:08 portaluser does that not work for modules used by orhcestration runners?
23:08 iggy no
23:08 portaluser OK, lemme give it a shot
23:09 cwyse joined #salt
23:09 iggy that's for minions... runners/orch run in the master context
23:10 portaluser I can use the same /srv/salt/_modules/ as a module dirs though?
23:11 iggy if that's where you actually keep thme, yes
23:11 marekb joined #salt
23:11 iggy we used gitfs, so I had to lay those files down on the master somewhere
23:12 mimiandi joined #salt
23:12 iggy but same difference either way
23:12 portaluser hmm, same issue.  Data failed to compile
23:12 whytewolf huh. orchestrate worked fine for me with a custom module without any added config
23:13 portaluser So it's running the commands and returning it in "Changes" (red text) with Data failed to compile
23:13 whytewolf portaluser: does it run fine from the command line without issue?
23:13 mimiandi hi,  how would you pass dict as a value in sls?
23:14 mimiandi i am trying to use reactor and one of the value i fire to saltmaster has a variable that is dict
23:14 mimiandi without having to do 'key': 'value' explicitly
23:14 portaluser whytewolf: Yeah, salt-call osquery.pillar_queries sec_osqueries works fine
23:17 cberndt joined #salt
23:17 eliasp portaluser: oh, there's now an osquery execution module? totally missed that… or is that your own/internal one?
23:17 hacfi joined #salt
23:17 wintamute joined #salt
23:18 whytewolf eliasp: both, there is a new osquery module in 2015.8.* however he is on 2015.5.5 and has their own
23:18 eliasp ok
23:18 portaluser yeah, there's the osquery execution module, I pulled the one from github and added in another function to just run a list of queries as defined in pillar
23:22 marekb hi everyone, is there a way to access metadata (or the whole job info) while in event_return() "salt/job/<JID>/new" event? salt.returners.local_cache.get_jid(jid) doesn't work inside there because of "NameError: global name '__opts__' is not defined"
23:22 whytewolf try running the salt-run with -l debug
23:22 * whytewolf is running out of ideas
23:23 eliasp portaluser: great, thx
23:25 portaluser Generally what does the "Data failed to compile:" mean?
23:26 portaluser it looks like the job is being distributed correctly, the data is coming back correctly, it's saying that the job succeeded in the Summary, but it's this error which I'm guessing is related to why the returner isn't launching
23:26 whytewolf thats the odd part. Data failed to compile generally means there was an error compileing an sls.
23:27 whytewolf things such as an already existing state.
23:27 whytewolf there is normally more to the error. like an error under it
23:28 portaluser yeah, it's the entirety of the execution module's return data
23:29 portaluser wait a sec, can a function in an execution module return an array, or does it be a dictionary?
23:30 lemur joined #salt
23:30 cwyse joined #salt
23:30 whytewolf I thought it needed to be a list.
23:31 whytewolf maybe it is a dict
23:31 portaluser hmm, let me try tweaking that
23:32 portaluser yeah, that was it.
23:33 portaluser Wrapped the array in a single-element dictionary and it's not throwing the error when launching the orcestrate runner
23:33 whytewolf ahhh. nice
23:36 protoz joined #salt
23:38 cpowell joined #salt
23:38 crazyphil joined #salt
23:38 M-liberdiko joined #salt
23:42 drawsmcgraw left #salt
23:42 portaluser Ha, now the returner still isn't working right.
23:42 seblu joined #salt
23:42 portaluser Oh well, that's a Wednesday problem
23:43 portaluser whytewolf: Thanks for your help
23:43 whytewolf no problem, wish i could have helped more
23:43 whytewolf but I am still on sundays brain
23:45 seblu joined #salt
23:46 pppingme joined #salt
23:46 pppingme joined #salt
23:53 akhter joined #salt
23:54 lemur joined #salt
23:59 cwyse joined #salt

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