Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-11-07

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

All times shown according to UTC.

Time Nick Message
00:00 jankowiak joined #salt
00:00 MTecknology sentx: I didn't see your name on the roster yet, so you have nothing to worry about it.
00:00 MTecknology yet*
00:03 sentx Possibly stupid question, do I need salt on the host and exposed via PATH or would vagrant deal with pre-configuration on its own?
00:03 MTecknology sentx: DUH!!!
00:03 MTecknology ... I know nothing about vagrant
00:04 MTecknology sentx: if it tries to launch anything with just the command then you need $PATH to be set. If you're passing the full path with the command, then it's not.
00:04 jcockhren sentx: you mean boostrapping a vagrant with salt?
00:04 sentx jcockhren, correct
00:04 sentx MTecknology, vagrant is literally violence written in software as it started off as a utility library to wrap Virtualbox's "api"
00:05 Cidan I use vagrant + salt
00:05 jcockhren sentx: https://gist.github.com/jcockhren/37469ab8475d35a923d3
00:05 Cidan but I just jabbed something into /etc/rc.local that a) downloads state files from a remote source b) runs salt-call state.highstate
00:06 MTecknology sentx: ... that sounds terrible.
00:06 MTecknology I didn't even know vbox had an api or anything of the like
00:06 jcockhren sentx: just a sample.
00:07 jcockhren sentx: I have a multi-host setup as well
00:07 jcockhren sentx: with 1 master in 2 minions
00:07 jcockhren s/in/and
00:07 sentx MTecknology, Yeah it doesn't really.  There's a xcom like thing but most of vagrant they're parsing VboxManage's stdout
00:08 pentabular joined #salt
00:08 sentx jcockhren, thanks I appreciate the help
00:08 anitak for the fact that is is "just" parsing, vagrant is doing a fine job
00:08 MTecknology I wonder how many minions one master can manage at a time
00:08 jcockhren MTecknology: 1 MILLION
00:08 __number5__ million
00:08 anitak MTecknology, ten's of thousand's, no?
00:08 jcockhren MTecknology: nah. I don't know
00:08 anitak :)
00:09 MTecknology I'm referring to a single master
00:09 MTecknology without needing a syndic server in the mix
00:09 sentx anitak, indeed.  I have a side project to make a deskbar widget for quick ssh connection to virtualbox instances and VboxManage isn't very friendly
00:09 __number5__ OK, I'm joking. The right number is same as the minions following Gru
00:09 MTecknology HOLY HELL!! What happened to amanda bynes?
00:10 __number5__ sentx: use vagrant!
00:11 jcockhren sentx: or salt to manage your vms ;)
00:11 sentx jcockhren, I am leaning towards using salt in a devops role and hoping my experiments of vagrant + salt will be cheaper than using AWS for testing/experimenting
00:12 jcockhren sentx: that's exactly how I'm using it
00:12 __number5__ sentx: I'm doing exactly like that
00:12 sentx salt + aws?
00:12 jcockhren sentx: salt + aws, linode, digitalocean
00:12 sentx I chose ansible at the start of the year and while its a nice CM system, it just didn't seem to fit right with a cloud environment
00:12 __number5__ yep, for aws use but developing with vagrant
00:13 MTecknology I'm going to get to talk to the lead salt admin for linkedin next thursday
00:13 MTecknology this could be fun
00:14 jcockhren sentx: my goal is to deploy sites through git-backed pillar
00:14 MTecknology anything I should make sure to ask him?
00:14 forrest MTecknology, let us know if they're doing anything cool we haven't seen.
00:15 forrest Ask if they're using the Salt Mine
00:15 jcockhren sentx: right now, I'm doing mostly maintenence and provisioning through salt. setting up needed formulas+sates as I go
00:16 jacksontj joined #salt
00:16 anitak left #salt
00:17 MTecknology forrest: good question.. next one?
00:17 sentx jcockhren, Still learning saltstack so maybe missing something.  pillar is just shared/managed state?
00:17 forrest How much do they pay their salt team :P, no I'm joking don't ask that.
00:17 jcockhren sentx: more like a datastore
00:18 sentx forrest, What is it like working for LinkedIn and whats the general opinion of python?
00:18 forrest I don't work there
00:18 sentx yeah I meant that for MTecknology
00:18 MTecknology forrest: well... if they decide to hire me, I would know that. That's the position they asked me to discuss.
00:18 forrest oh ok
00:18 forrest Ahh, well good luck
00:19 sentx I know LinkedIn has sent their development managers ( and some of their code monkies ) to the last two US PyCon's
00:19 forrest In reality though my primary question would be how they handle state contributions, and the environments that they push to.
00:19 forrest Yea they hit me up on LinkedIn a few days ago, but it wasn't for one of the Salt spots.
00:20 anitak joined #salt
00:20 MTecknology sentx: python, php, perl, and something else... that's the four main ones they use?
00:20 forrest You should be asking them how they handle dev/test/prod pushes, is it all done from GIT? Do you they use a branching scheme? Do developers write Salt code? How do they handle code approvals and pushes from one environment to the next in terms of Salt? To what extent do developers influence the way salt is configured, etc.
00:20 mwillhite joined #salt
00:21 forrest The workflow they use is pretty important
00:21 forrest because if that sucks, your life sucks
00:21 forrest especially if their developers are almighty gods who don't lift a finger to do the work of the peasant.
00:22 MTecknology sentx: ruby.. that's the 4th
00:22 sentx forrest, I worked for one company where every line of code had to pass jslint/pep8 + 100% test coverage.  It was not very funny
00:22 forrest Yea, sounds like they had that right though :\
00:22 forrest you should have full coverage and pep8
00:23 sentx MTecknology, everything but perl doesn't sound too bad
00:23 MTecknology sentx: I hate php and perl and ruby; but love python
00:23 forrest yea, another good question to ask
00:23 forrest to what extent are they using PHP especially
00:23 sentx forrest, Ruby isn't so bad... Rails on otherhand is not too fun
00:23 forrest perl I can tolerate, and ruby I can also tolerate
00:24 forrest but php, no conventions, no one cares.
00:24 forrest yes, rials is meh, ruby is ok
00:24 forrest *rails
00:25 patrek joined #salt
00:25 pdayton joined #salt
00:27 mgw joined #salt
00:27 forrest MTecknology, you might also want to ask what their long term company goals are for the next 5 years, obviously shareholder stock is up at this point, but what does that mean for the company? Are they just squeezing every penny out to please stockholders?
00:28 forrest Plus how will that affect your day to day activities, because public companies can sometimes be kinda meh, might be better there because they have to compete with other large guys to retain talent.
00:29 Boohbah joined #salt
00:32 MTecknology How much does your group deal with perl, php and ruby?
00:32 MTecknology there's another. :)
00:32 forrest Yea just basic stuff.
00:33 MTecknology I'm interviewing them as much as they're interviewing me. Honestly, I don't really want the job. Unless I'm offered something like $300k, I probably won't take it.
00:33 MTecknology So... I get to ask whatevery the heck I wan.t
00:33 forrest yea you have to interview them when you go
00:33 forrest that's the point
00:34 forrest I don't know if any of their team hangs out in here, never seen anyone
00:34 MTecknology I need good questions that make them sweat and squirm in their seats! The salt lead is just the guy I get to talk to for fun.
00:34 forrest Oh
00:34 forrest well, if you want to be a real bastard, you can ask about what their retirement process is looking like for their 'ancient php technology stack'
00:35 MTecknology HAH! Nice!
00:35 forrest as well as other terrible questions relating to mid-term goals
00:36 MTecknology holy shit... http://www.collegehumor.com/pictures/gallery/6934672/you-could-probably-watch-these-forever-23-oddly-satisfying-gifs
00:36 forrest why you spamming the IRC with gifs :(
00:36 MTecknology that's just fucking amazing
00:38 forrest I wish the linkedin guys were in here from their saltstack team, I'd be really interested to learn about their setup and how they're handling so many boxes, just to see what that looks like.
00:40 MTecknology forrest: exactly! tell me what to ask!
00:40 goodwill MTecknology: hey big chief :)
00:40 MTecknology HI!
00:42 forrest I don't really have anything more to ask other than for them to actually share some of what they are doing
00:43 forrest Does anyone know someone who works at Basis?
00:43 jacksontj joined #salt
00:43 pipps__ joined #salt
00:43 MTecknology forrest: If you have questions that you would ask them, then you should ask me and in a week (just over), I can answer them. :)
00:44 forrest the linkedin guys?
00:45 MTecknology ya
00:46 sfz joined #salt
00:47 forrest I don't have anything more, I just want examples of large scale Salt implementations.
00:48 jcockhren forrest: define large. 100+?
00:49 forrest yea
00:49 forrest I wanna see a setup using the mine and salt reactor and such
00:50 MTecknology By the end of next year, I'll have 300 servers on salt. I need to make sure one master can tolerate that
00:51 forrest should be able to easily
00:51 forrest UtahDave spun up a 100 node riak cluster from his laptop with salt-ssh, lol. zeromq salt should destroy that.
00:54 pipps_ joined #salt
00:57 eculver joined #salt
01:06 DallaRosa joined #salt
01:12 teebes joined #salt
01:13 redondos_ joined #salt
01:15 packeteer i've never had a successful salt-ssh run  :/
01:16 Gifflen joined #salt
01:28 teebes joined #salt
01:28 Gifflen joined #salt
01:34 pipps joined #salt
01:37 MTecknology I've never used salt-ssh
01:38 packeteer i cant wait for it to be usable. gonna start using it to do my work
01:39 packeteer tho, the fact that my boss has mandated chef realy makes life difficult
01:39 fllr joined #salt
01:40 fllr How does the running state work? I'm asking cause I wanna install a uwsgi server on my machine, but I don't know how to keep it running, restarting if a file change, etc...
01:40 jslatts joined #salt
01:44 sentx fllr, for ubuntu there's upstart for managing process/services - http://uwsgi-docs.readthedocs.org/en/latest/Upstart.html
01:46 fllr sentx: ah! cool. thanks!
01:50 Ryan_Lane joined #salt
01:56 jankowiak joined #salt
02:00 ddv joined #salt
02:03 BrendanGilmore joined #salt
02:03 pipps joined #salt
02:03 redondos joined #salt
02:04 pipps_ joined #salt
02:11 teebes joined #salt
02:13 ddv joined #salt
02:15 jhulten joined #salt
02:22 MTecknology Is there any dummies guide to salt 0.17.1 yet?
02:24 forrest MTecknology, in what sense?
02:24 forrest there's the walkthrough which works fine with 0.17.1
02:26 MTecknology forrest: I want to learn everything about what exists at this moment so I can attempt to keep up on changes and know what's out there before saltconf
02:26 pdayton joined #salt
02:26 oz_akan_ joined #salt
02:28 forrest MTecknology
02:28 forrest that is a lot of stuff
02:28 forrest lol
02:28 forrest Just start hammering on the docs is all I can suggest.
02:28 oz_akan_ joined #salt
02:29 MTecknology forrest: I'd prefer the book... I'd like it in audio format.
02:30 goodwill yeah, there is no such thing?
02:30 forrest there's no pro salt if that's what you're asking
02:31 gkze joined #salt
02:31 MTecknology fuck man... I know. I was trying to be funny. :(
02:32 tempspace MTecknology: I was playing around with the idea of making a Salt Course
02:32 gkze guys, question: have you experienced difficulty cmd.run-ing commands with sudo?
02:32 forrest tempspace, I was thinking of just writing pro salt :P
02:32 gkze I tried to cmd.run 'sudo -u anotheruser command' but that didn't pick up that users' environment variables properly
02:33 gkze so I fixed it by force exporting the var beforehand: cmd.run 'export VAR=something; sudo -u anotheruser command'
02:33 gkze and that ^ worked just fine, but I'm not entirely sure how
02:34 tempspace gkze: you're trying to do this from the command line or from a state?
02:34 gkze cmd.run
02:34 gkze not a state
02:34 gkze command like yes
02:34 gkze *line
02:35 MTecknology tempspace: I'm hoping to be at saltconf and would like to be well read up on it before going so I can have the opportunity to pick up and learn more advanced and finer points.
02:36 tempspace MTecknology: Yeah, salt's getting started is a little lacking
02:36 forrest tempspace, what do you feel would make it better?
02:37 tempspace forrest: I asked that to salt-users, let me find the thread
02:37 forrest cool
02:37 tempspace forrest: https://groups.google.com/forum/#!searchin/salt-users/course/salt-users/DAs32w3CTC0/sJjO_Dijv-UJ
02:38 xl1 joined #salt
02:38 forrest you're Will?
02:39 tempspace yes
02:39 forrest how is this video tutorial going
02:39 tempspace gkze: try salt target cmd.run 'command' runas=useryouwanttorunas
02:40 tempspace forrest: I made an introduction video just to see how big of a pain in the butt it was
02:40 tempspace forrest: the answer was a lot :)
02:40 jimallman joined #salt
02:40 MTecknology salt has so far been a tool that expands and grows with what you learn. The more you learn, the more you can make it do. When I started, I used modules (remote execution) only. Then I started learning states and realized that's where the real power of salt is. Then I ran into an issue where I didn't like the states and learned about pillars. I then got to learn how pillars made things even
02:41 MTecknology better. Then I ran into an issue that salt couldn't solve... until I learned about the reactor. Same thing happened that caused me to learn about ext_pillars.
02:41 MTecknology Once I had to write a state because it didn't exist yet. It was dead simple.
02:42 forrest tempspace, yea making videos sucks
02:44 tempspace forrest: yep, so I just stalled out unable to make a decision
02:44 forrest gotcha
02:44 tempspace Was kind of hoping someone else was going to step up and do a dummies guide to Salt before I did
02:44 tempspace since I'm a big a dummy as you can find :)
02:44 MTecknology I love that you can still just pick up and start using salt without knowing much and then the more you want to expand with it the more you find out what it's capable of... before I need something, someone else usually came up with it
02:44 forrest I've been planning to do one.
02:44 forrest just working on lots of stuff, you know how it is
02:44 forrest MTecknology, yeah I gotcha
02:44 tempspace MTecknology: I hope you learn what you need! And I'll see you at SaltConf if you're there
02:44 MTecknology forrest: sorry for the rambling; it's just something that makes me love it so much.
02:45 forrest no I understand
02:45 forrest that's why I like it as well
02:45 tempspace me too
02:46 forrest I've gotta get some food, I'll be back in a bit;
02:46 tempspace Once salt gets some better unit testing and qa procedures, it'll be even more awesome. I have a regression issue with just about every single release
02:46 forrest tempspace, ugh let's not start the pitchforks there, shhhhhh
02:47 forrest had a big discussion about that two or three nights ago
02:47 forrest anyways, food
02:47 MTecknology I can agree with that one... unit testing is a bitch and relies on people writing good tests, but golly gosh it's useful
02:47 tempspace Testing salt is very difficult
02:48 tempspace as we can see
02:48 tempspace It'll get there, I have faith in the team
02:48 * MTecknology is hungry too...
02:48 gkze tempspace: thanks, I think that should work
02:49 tempspace gkze: awesome! That used all the enviornment variables you needed?
02:49 MTecknology tempspace: especially when soooo much is added or changed in every release. It's crazy. However, that's rather typical of a younger project. Salt really isn't very old at all yet.
02:49 tempspace MTecknology: yep
02:50 tempspace looking forward to seeing how other people use Salt at SaltConf
02:50 MTecknology I'm excited to learn how linkedin uses it
02:51 tempspace I do something weird with my pillars and states that I don't see anybody else doing, so I'm excited to see other models
02:51 MTecknology what's that?
02:52 MTecknology how's that? *
02:52 MTecknology I actually don't make much use of pillars
02:53 tempspace MTecknology: I found that when I wanted to add something like postgresql, I was assigning pillar values about postgresql, like the version I want to install
02:53 tempspace MTecknology: Then I want the state to run, so I'd have to assign something the postgresql pillar and the postgresql state
02:53 tempspace MTecknology: So for each new server, I'm doing double duty
02:54 tempspace MTecknology: So all my states are dependent on pillars, so I only assign the pillar, and the states get assigned based on the pillars I assign for a particular host
02:55 tempspace MTecknology: If you don't use pillars, what I said may not make a lot of sense
02:55 tempspace (and maybe even if you do...)
02:55 Jahkeup tempspace: that's not a bad idea actually. top.sls taking care of both is a great idea i think
02:55 MTecknology it makes sense, but sounds like there has to be an easier way...
02:56 MTecknology it sounds like laundry changing time
02:56 Jahkeup or at least I can think of use for it, but I do like my globbing right now ;) may experiment in the future
02:57 tempspace I'm not a pillar I just glob a lot
02:57 tempspace Jahkeup: I glob a lot too, but I keep on bringing up servers outside my globs
02:57 packeteer so, using salt to run up an ec2. is that salt-cloud?
02:58 Jahkeup tempspace: ahhh gotcha. yeah then that usecase makes sense ;)
02:58 tempspace packeteer: salt-cloud does the talking to EC2 API to provision the VM and automates the deployment of salt
02:59 noob2 joined #salt
02:59 tempspace packeteer: or SoftLayer, Rackspace or HP Cloud or DigitalOcean, Linode etc...
03:00 gkze tempspace: nope, that still didn't work
03:00 tempspace so salt-cloud brings up the VM and installs salt, which gets the VM in the state you want it to be in
03:00 gkze not picking up LC_ALL
03:00 gkze that's the env var I need
03:00 packeteer tempspace: ok, thanks
03:00 gkze cause root's is C and the user's is en_US.UTF-
03:00 gkze -8*
03:01 gkze cmd.run 'command' user=otheruser didn't work
03:02 tempspace gkze: try cmd.run with 'su -l useryouwanttorunas -c 'commandyouwannarun'
03:02 redondos joined #salt
03:02 redondos joined #salt
03:03 tempspace packeteer: np
03:03 BrendanGilmore joined #salt
03:04 gkze tempspace: bingo!
03:04 gkze worked like a charm
03:05 tempspace awesome
03:05 gkze I wonder what's different between 'su -l user -c command' and 'sudo -u user command'
03:05 gkze But I guess that's off topic, I will go google around
03:05 tempspace -l simulates a full login which gets the environment variables
03:05 gkze sudo -u doesn't?
03:06 gkze just changes the uid?
03:06 tempspace I believe it does not
03:06 gkze hm ok
03:06 gkze well thanks anyway, I've been at this for over a day now
03:06 packeteer -l just listes allowed commands
03:06 packeteer you may be thinkning of -i
03:07 packeteer -l (list) option will list the allowed (and forbidden) commands for the invoking user
03:07 tempspace packeteer: http://unixhelp.ed.ac.uk/CGI/man-cgi?su
03:08 packeteer oh, sorry. i was man sudo
03:15 jalbretsen joined #salt
03:23 jhulten joined #salt
03:24 ipmb joined #salt
03:27 pipps_ joined #salt
03:30 bhosmer joined #salt
03:31 genkimind left #salt
03:41 cachedout joined #salt
03:41 Ryan_Lane joined #salt
03:51 redondos joined #salt
03:56 sentx joined #salt
03:57 sentx With the 1.3 salt/vagrant, how do I work around the invalid option:  stdin is not a tty error?
04:02 forrest I'm not sure sentx :\
04:02 sentx forrest, brute forcing it by using the :shell provisioner to hack around some issues
04:02 forrest hh
04:02 forrest *ahh
04:03 sentx https://github.com/saltstack/salty-vagrant/issues/66#issuecomment-15654580
04:03 sentx I am assuming the shell is throwing up when it hits `#!/bin/sh -`
04:06 _ikke_ joined #salt
04:10 __number5__ sentx: what's in your Vagrantfile, I'm using the salt provisioning without any issue
04:12 sentx I just updated vagrant to 1.3.5 so perhaps its a regression, the spaghetti I've got so far is this - https://gist.github.com/devdave/e8c02d8db77012e025f8
04:13 sentx __number5__, what does `vagrant -v` say?  Curious if I just downloaded too fresh/new of an install
04:18 mannyt joined #salt
04:19 jhulten_ joined #salt
04:19 pipps joined #salt
04:22 mgw joined #salt
04:24 forrest terminalmage, did anyone build python-halite for any distro other than the EPEL ones?
04:28 forrest or I should say distros that can handle rpms
04:28 __number5__ sentx: I'm using vagrant 1.3.5 as well
04:29 mnemonikk joined #salt
04:29 pipps_ joined #salt
04:30 sentx __number5__, interesting, perhaps its a conflict from the update ( pre-1.3 ), will restart the workstation and wipe out vagrant ( Have you tried turning it off and back on again )
04:33 sentx joined #salt
04:33 __number5__ sentx: I guess what you need is just changing the :salt to "salt"
04:33 __number5__ in your Vagrant file
04:34 sentx would laugh if that's it
04:42 anuvrat joined #salt
04:43 lemao joined #salt
04:47 xl1 joined #salt
05:02 xl1 left #salt
05:02 redondos joined #salt
05:09 Teknix joined #salt
05:13 xl1 joined #salt
05:14 xmltok joined #salt
05:15 pipps joined #salt
05:20 cachedout joined #salt
05:22 jacksontj joined #salt
05:24 jankowiak joined #salt
05:25 jacksontj_ joined #salt
05:29 clone1018_ joined #salt
05:30 joonas joined #salt
05:31 gkze joined #salt
05:35 defunctzombie_zz joined #salt
05:46 sfz joined #salt
05:49 nmws joined #salt
05:50 nmws Hello I have a question about the salt libvirt module: Does it support vmware and/or aws?
06:02 ctdawe joined #salt
06:05 gkze joined #salt
06:08 UtahDave joined #salt
06:13 kung joined #salt
06:14 Guest86331 hello
06:19 UtahDave hey, Guest863311
06:27 wy-fund joined #salt
06:29 wy-fund joined #salt
06:38 NV on RHEL5/6, what is the preferred way to get zmq 3 instead of zmq2 (which apparently has connectivity issues?)
06:41 matanya joined #salt
06:44 lemao joined #salt
06:44 matanya joined #salt
06:44 Ryan_Lane joined #salt
06:46 defunctzombie joined #salt
06:49 NV RHEL5 sorry, RHEL6 already has zmq 3
06:50 __number5__ NV: I think EPEL has libzmq3 http://docs.saltstack.com/topics/installation/rhel.html
06:51 NV on RHEL6 it does, but not on RHEL5 afaict
06:51 __number5__ :/
06:51 NV RHEL5 just has zmq2, which apparently has communication issues etc
06:55 UtahDave NV: have you tried recompiling the rhel6 zmq rpms for rhel5?
07:06 sebgoa joined #salt
07:11 fink_ployd joined #salt
07:14 ravibhure joined #salt
07:23 NV UtahDave: nah, was kinda hoping that there would be something that just works without having to setup a yum repo, etc...
07:32 creich joined #salt
07:38 PoLuX joined #salt
07:44 carlos joined #salt
07:49 linjan joined #salt
07:51 creich hi i got a strange behavior.. i tried to copy some big files from my master to my minion using the file.managed command within a state-file.
07:51 creich the strange thing is that after 6minutes i got an timeout-error which sais that 60seconds are gone
07:52 creich while that hapens i permanently got new entries within my master-log saing 2013-11-07 08:49:32,118 [salt.master      ][INFO    ] AES payload received with command _serve_file
07:52 honestly creich: timeout is just for the cli interface
07:52 creich i checked the minion also and there was somthing going on
07:52 creich after waiting a while the files were all present
07:52 honestly creich: either do -t 3600 or --async
07:52 creich hmm async
07:52 creich ok
07:52 honestly you can check job status with the jobs module
07:53 alunduil joined #salt
07:53 creich ok sounds great
07:53 creich that was what i wanted to ask for
07:53 honestly --async will give you the job id right away
07:53 creich if there is some way to avoid this message cause it seems all to be ok
07:53 creich cool
07:53 creich :D
07:53 creich thx
07:53 honestly and then get the output with job.lookup_jid
07:53 honestly jobs*
07:53 creich i'll try that
07:54 creich time salt 'jenkins*' state.highstate --async
07:54 honestly and running jobs with jobs.active
07:54 creich but got no job-id
07:54 creich looks still like before
07:54 honestly hm
07:54 honestly put the --async right after the salt
07:55 creich ok
07:55 creich same behavior
07:55 honestly huh
07:55 gildegoma joined #salt
07:55 honestly maybe that changed in 0.17
07:56 honestly hit ctrl+c while it's running
07:56 creich but it looks like the minoin does nothing
07:56 honestly does that give you the jid?
07:56 honestly maybe that's because the job is already running
07:56 creich no in both cases
07:56 creich i think the first job was finisehd
07:57 creich w8
07:57 creich i'll stop everything
07:57 honestly what does salt-run jobs.active return?
07:57 creich reset my VM and test it again
07:57 honestly ok
07:57 slav0nic joined #salt
07:57 creich salt-run jobs.active
07:57 creich {}
07:57 honestly hmm
07:57 creich ok so i'll restart my VM
07:57 honestly ok
07:58 honestly UtahDave: you around?
08:00 creich ah sry, i have to leave for a meeting...
08:00 creich will be back in around 2h
08:00 creich thx for your help
08:01 honestly k
08:01 honestly np
08:03 balboah joined #salt
08:05 Iwirada joined #salt
08:11 harobed_ joined #salt
08:12 harobed_ joined #salt
08:31 malinoff Guys, can someone give an advice on logging in custom modules? I can't figure out how to properly set up the logger...
08:32 scott_w joined #salt
08:33 honestly log = logging.getLogger()
08:33 honestly log.info("hey, i'm the root logger!)
08:34 honestly if saltstack uses the logging system
08:34 honestly i hope it does
08:36 Ryan_Lane joined #salt
08:38 Furao honestly: that's the way to do it
08:38 honestly ok (:
08:38 scott_w it does
08:39 honestly no logger substring?
08:39 honestly like 'salt.mymodule'?
08:39 scott_w anyone mind if i make a pull request with a simple SQS integration? (#6444)
08:40 honestly prolly not
08:40 honestly but they might not merge it
08:40 honestly get in touch with utahdave
08:41 scott_w are they all still in bed? :P
08:41 honestly seems so
08:42 malinoff honestly, just getLogger, without any arguments?
08:42 malinoff I see in the source code that getLogger is called with __name__ arg
08:43 honestly do that
08:43 malinoff honestly, It's not working.
08:44 honestly ask Furao then
08:44 * malinoff asking Furao
08:45 honestly or show some code and say what you'rre e expecting to see
08:45 honestly [C
08:46 malinoff honestly, The code is very simple - I'm trying to log.info something in a function in custom module - and there's nothing in the log file/output
08:47 Furao malinoff: привет
08:47 Furao maybe you need to run -l debug or -l info
08:47 Furao to see?
08:48 malinoff Привет, Furao
08:48 malinoff Tried that, nothing helped
08:48 scott_w are you using salt-call
08:48 scott_w ?
08:49 malinoff looks like getLogger returns nothing (or something inconsistent)
08:49 malinoff scott_w, Nope, I call a function from the master
08:49 scott_w salt-call --local -l info module.command
08:49 scott_w maybe that's why? i don't know if it returns logging output to the master
08:49 malinoff scott_w, I can't do that, my module relies on target string
08:49 scott_w ah
08:50 Furao I know this: https://github.com/bclermont/states/blob/master/states/_modules/npm.py
08:50 Furao works in term of logging
08:53 malinoff Furao, I have the same logging stuff... Don't know what's the problem
08:54 Furao UnicodeEncodeError: 'ascii' codec can't encode character u'\u0441' in position 13: ordinal not in range(128)
08:54 Furao I need to remove my russian keyboard :)
08:55 sfz joined #salt
08:55 Furao I couldn't see the difference with latin "c"
08:55 bhosmer joined #salt
08:58 hazzadous joined #salt
08:59 malinoff Furao, only in numerical view :)
09:01 malinoff Furao, Lazy Russians developed 'Punto Switcher' - this program analyzes current input and switches layout if needed (with correction of previous typed words)
09:01 Furao oh really, useful, but not for me :)
09:02 Furao is "c" the only character mapped to the same key in US than russian?
09:03 malinoff Furao, yes
09:03 Furao oh that's why
09:05 networkpadawan joined #salt
09:06 UtahDave joined #salt
09:07 giantlock joined #salt
09:08 Furao_ joined #salt
09:10 ninkotech joined #salt
09:10 UtahDave honestly: hey, what's up?
09:10 honestly urgh, unicode in python
09:10 honestly just don't even try
09:11 honestly UtahDave: wanted to ping you on creich's problem
09:11 malinoff Interesting. Looks like that logging actually works, but I can't get any logging messages on the master
09:11 honestly seems he's experiencing weirdness
09:11 malinoff except debug msgs
09:11 UtahDave what's he seeing?
09:12 honestly do you see the backlog?
09:12 honestly I think you can ignore it for now, he said he'll be back soon
09:12 ninkotech joined #salt
09:13 UtahDave ok, cool
09:19 Furao UtahDave: you're lucky, tomorrow is supposed to be day of the week in term of weather
09:20 Furao to be the BEST day
09:20 Furao sunday it's back to usual rain
09:21 Furao yay first salt runner that works
09:24 malinoff Can I somehow know whether my module runs locally or from the master?
09:27 Furao malinoff: module run in minion
09:27 malinoff Furao, I know. I can run with 'salt minion-id mymodule.func' from the master or with 'salt-call mymodule.func' locally on a minion
09:28 UtahDave nice!
09:28 Furao if you run salt $minion mod.func from the master, the module will still run on the minion
09:28 malinoff I need to see the difference between these runs in the module
09:28 Furao to make it simple, /usr/bin/salt will ask the master to ask the minion to run mod.func
09:29 malinoff Furao, let me clarify
09:29 malinoff The first way always has target string
09:29 malinoff The second - has not
09:30 Furao salt-call on minion is almost the same as running salt minionid-of-where-salt-call-is-executed
09:30 malinoff Furao, I pointed the difference above
09:30 Furao the difference is that salt-call results don't ends in job cache
09:31 fr0de joined #salt
09:33 fr0de hi, quick question from a total n00b: If I am to store my salt config in a git repo, is there a way to ensure that sensitive data like db passwords are not in plaintext in the repo? In puppet's hiera, there's a plugin that lets you store the values in encrypted format. Anything similar for salt?
09:34 Furao fr0de: pillar
09:34 fr0de it's encrypted?
09:35 Furao I keep my pillar git repo on the master itself
09:35 Furao no
09:35 Furao there is a hiera ext_pillar module
09:36 fr0de so then I'd need hiera installed too?
09:36 Furao I never used it
09:37 fr0de hmm... I might have to think about developing something for pillar to do this stuff.. :)
09:48 qba73 joined #salt
09:51 sebgoa joined #salt
09:54 NV I'm not sure what you're trying to accomplish that pillar (especially when combined with ext_pillar, or even just a python rendered pillar sls file) can't achieve
09:55 Minioncito joined #salt
09:56 Minioncito Hi all
09:56 fr0de good question - I (as mentioned) am a total newbie to salt - I've done a lot of puppet, but not yet put salt into any live environment. My concern is that we store sensitive data in the git repo and would need it to be encrypted. if pillar can do that easily, great :) I just didn't see how from the doco..
09:57 whiskybar joined #salt
09:59 fr0de with the hiera plugin, we store the key-value like "mysql_root_password: [ENC:<some encrypted string>]"
09:59 fr0de so that if the git repo is compromised, the passwords are not
09:59 fr0de (using a private bitbucket repo)
10:01 networkpadawan joined #salt
10:01 Minioncito I'm going to explain some weird behaviour of syndic, and if anyone know I appreciate any help :(. I'm using a masterofmasters with some regular masters for each environment and lot of minions depending of each regular master. When I try to match any minion from each regular master with regexp, everything OK; but when try from the masterofmasters it works... sometimes! This is a kind of random behaviour, e.g: my minions are called minion[
10:11 Ryan_Lane joined #salt
10:14 creich honestly, UtahDave thx for regarding my issue. i am back now ;)
10:15 creich so my initial problem was that i somehow often get some timeouts even if the minion is still working and finally will get the job done
10:15 creich but i also got some connection issues all the time
10:16 honestly the timeout is normal
10:16 bhosmer joined #salt
10:16 creich i think that after a few hours  (maybe minutes) my master can't reach the minions
10:23 creich honestly, yeah you said that but the really problem with that was taht the --sync does not work, but that seems to be a conection problem too
10:23 creich cause after restarting the minions the async options works as expected
10:23 creich so the question now is, is there anybody else experiencing some sonnection issues
10:24 creich s/sonnection/connection/
10:24 honestly hmm
10:24 honestly is there nat bullshit between your master and the minions?
10:25 NV fr0de_afk: that would require you keep your encryption key secure - so why not just keep your pillar data secure?
10:25 NV you can write a python script that can provide pillar data
10:25 NV you could keep your non-sensitive pillar data in git, and source your sensitive data from elsewhere for example
10:26 creich honestly, somehow... so my setup is the following
10:26 creich i'm using 2 VMs (virtual box) one is the master the other one as minion
10:27 creich host-OS is Win7
10:27 creich using NAT for inet connection but host-only-adapter for connection between the master and minion machines
10:27 bhosmer_ joined #salt
10:27 creich so i am not sure if the NAT is the problem here
10:28 honestly hmmm
10:28 fink_ployd left #salt
10:28 honestly make sure the host-only connection is working properly
10:28 creich but i guess you are only answering cause you don't have this problems in a real environment
10:29 honestly and then set the master address on the minions to the host-only IP of the master
10:29 creich i am using the host-only for my ssh-conections without any problems
10:29 NV what OS?
10:29 creich so i think everything is working properly
10:29 creich master-address is the host-only ip
10:30 honestly well, ssh would work find through nat
10:30 honestly fine*
10:30 creich host-os is Win7 guest os is ubuntu 12 lts
10:30 honestly or maybe not
10:30 creich master is desktop ubuntu minion is server
10:31 creich but that should not matter i thiink
10:31 creich s/thiink/think/
10:31 NV using zeromq3?
10:31 creich NV, where to check that? i am using salt 0.17.1
10:33 NV creich: check your package manager - it will be a dependancy of salt
10:33 viq creich: are you familiar with vagrant ? ;)
10:34 joehh salt-minion --versions-report
10:34 joehh creich: ^^
10:35 NV ah yeah that's probably a better idea :P
10:35 joehh I would expect to see a versoin of python-zmq > 13
10:35 joehh dpkg -l | grep zmq
10:36 creich http://pastebin.com/11mxVg4e
10:37 creich so yes zmq 3.2.2
10:37 creich pyzmq = 13.0.0
10:38 creich i used the bootstrap script to install the packages
10:38 creich not the packages
10:38 creich is there a difference?
10:39 joehh not a significant difference - just need to watch if you switch between the two
10:39 joehh have you looked at https://github.com/saltstack/salt/issues/8295
10:39 joehh could it be that?
10:39 creich sofar the --async option works if my connections are up and running
10:39 creich joehh, i'll have a look
10:41 creich i assume the conclusion of that ticket is different architechtures 32- VS 64-bit
10:41 creich that is not the case here
10:42 creich i have both 64bit
10:42 creich and i think the problems only occur if i let the minions and master idle around some time
10:42 creich but it's not that long
10:43 creich btw, can i set the --async as a default option within the master config?
10:43 joehh not to sure
10:43 bezaban going to be selling salt at todays operations meeting
10:44 joehh that was regarding async
10:44 bezaban so probably running a test deployment this week.
10:45 creich didn't found such an option yet
10:45 creich but maybe i can use an alias for that
10:45 lemao joined #salt
10:46 lemao joined #salt
10:46 joehh also not too sure about conclusion either
10:47 joehh edomaur saw it on debian with 64bit on both master/minion
10:47 creich anyway, there is no solution within that ticket yet as far as i can see
10:48 joehh no agree on tha
10:48 joehh t
10:49 creich but i assume that these connection problems are VM related somehow
10:49 joehh I would add either a me too or a new ticket - there haven't been minion connectivity issues
10:49 creich i am not sure about that but if noone else has this problems in real world
10:49 joehh with zmq3 for a long time
10:49 creich ok
10:51 creich ah besides i didn't notice that this ticket was related to the timeout... sry... the timeout problem seems to be solved using the "--async"
10:51 scott_w joined #salt
10:51 creich the main problem i encounter everytime was related to the connection
10:51 creich for now i am trying the "keep-alive" settings for the minions
10:51 creich maybe that will help
10:53 joehh maybe
10:58 creich seems not :(
10:58 creich while browsing and reading tickets and mailing list ... after calling a test.ping i don't reach my minion again
11:04 creich ok so it seems to be a problem with the VM set-up
11:04 shar_HS joined #salt
11:04 creich one of my coworkers allready installed some minions in an real set-up and they encounter no connection problems...
11:05 shar_HS Hey guys, I'm looking for a way to use salt without deploying minions - and i figured salt-ssh is the way to go. but i haven't found solid documentation on how to use salt-ssh with states. can anyone help me out?
11:06 IJNX joined #salt
11:08 diegows joined #salt
11:11 N-Mi joined #salt
11:11 N-Mi joined #salt
11:13 ggoZ joined #salt
11:15 viq shar_HS: I think just the same way you would with normal salt
11:15 viq shar_HS: also, have you seen ansible?
11:27 networkpadawan joined #salt
11:28 JasonG_TA joined #salt
11:35 UtahDave joined #salt
11:46 Minioncito Hi, I've compiled info about my problem if anyone can help: http://pastebin.com/ev7ES0Pg
11:54 UtahDave great pastebin, Minioncito.  still reading
11:55 UtahDave Minioncito: That looks like it's a bug to me.  Would you mind opening an issue on this? including that great description?
11:56 Minioncito UtahDave: of course! Thank you very much
11:56 jcockhren Minioncito: I've experienced that as well
11:57 tomspur joined #salt
11:57 UtahDave thanks for letting us know about that!
11:58 Minioncito UtahDave: jcockhren: I have been debugging for a couple of hours and didn't find nothing, so I was also starting to think that it could be a bug. I'm opening the issue
11:59 jcockhren UtahDave: yeah. my bad. I've been slow on logging syndic related issues. Totally saw that Tuesday
12:00 jcockhren Minioncito: To be clear, you've installed salt-master, salt-syndic and salt-minion on your syndics?
12:01 pears joined #salt
12:04 Minioncito jcockhren: yeah, did it yesterday when I set up the whole environment. Btw, my syndics are regular masters and only have salt-master and salt-syndic services running. I installed the salt-minion package because the /etc/salt/minion file is necessary to run salt-syndic service
12:05 Minioncito (bad packaging I guess? Because salt-syndic service is part of salt-master package, but /etc/salt/minion file wich is needed by salt-syndic, is actually included only in salt-minion package... EPEL repositories)
12:05 jcockhren Minioncito: really? I wasn't able to get syndics to be directly targeted without have a minion service running
12:06 jcockhren I think it's possible that I assumed the minion service had to be running and not just have the minion config file present
12:06 Minioncito jcockhren: really. Only the file is necessary (at least in my case), not the service running
12:07 Minioncito And the file is necessary because of 'id:' parameter, because if you don't set it, salt-syndic starts with the 'salt syndic minion' set to 'None' and never authenticates through the masterofmasters
12:07 Minioncito (I found the problem also yesterday :P)
12:07 logix812 joined #salt
12:08 jcockhren Minioncito: yeah
12:09 scott_w joined #salt
12:09 jcockhren Minioncito: can you "salt 'salt0[1-2]*' test.ping" your syndics?
12:09 Minioncito despite all of this, I'm really happy with salt. It's very powerful imho
12:09 jcockhren I see in your pastebin you can call the minions from the masterofmaster
12:10 jcockhren I don't see your syndic showing up in anywhere other than your salt-key command
12:10 Minioncito no, I ca't call dmy regular masters directly from my masterofmasters
12:10 Minioncito I guess it's because actually they are not real minions?
12:10 viq https://gist.github.com/viq/7353629 - any idea why it creates groups, but does not assign users to them ?
12:11 UtahDave Minioncito: that's correct
12:11 jcockhren Yeah. so running the salt-minion makes them directly targetable
12:12 Tekni joined #salt
12:12 jcockhren In my test case, my syndics had lxc containers as minions
12:12 jcockhren so I'd want direct targeting of my syndic to manage those containers
12:13 Minioncito (thank you guys, I'm going to have lunch now, will let you know my issue when it's open)
12:13 hotbox_ joined #salt
12:13 elfixit joined #salt
12:13 hazzadous_ joined #salt
12:14 jcockhren fwiw UtahDave, lxc minions contained in a syndic worked nicely excluding Minioncito's issue
12:14 elm-1 joined #salt
12:16 BogdanR_ joined #salt
12:16 pkruithof joined #salt
12:16 bigmston1 joined #salt
12:16 btorch_ joined #salt
12:17 drogoh_ joined #salt
12:17 Bosse_ joined #salt
12:17 bui_ joined #salt
12:20 johnsocp_ joined #salt
12:20 dcolish_ joined #salt
12:20 mr_chris_ joined #salt
12:20 bcc- joined #salt
12:20 [ilin]- joined #salt
12:20 marcel- joined #salt
12:21 [ilin] joined #salt
12:21 totte joined #salt
12:23 networkpadawan joined #salt
12:26 qba73 joined #salt
12:26 UtahDave jcockhren: oh, yeah? that's cool.
12:26 VSpike joined #salt
12:27 SEJeff_work joined #salt
12:27 xerxas joined #salt
12:27 baoboa joined #salt
12:27 gmoro joined #salt
12:27 yota joined #salt
12:28 modafinil joined #salt
12:29 srijan4 joined #salt
12:30 Corey joined #salt
12:30 malinoff UtahDave, could you help me? :)
12:30 elm- joined #salt
12:31 viq malinoff: how about asking the question first ? ;)
12:31 malinoff just want to ensure that the person has time and desire to answer
12:32 viq Well, someone else might now something too ;)
12:32 tempspace Good morning!
12:32 UtahDave :)  malinoff.  shoot!
12:34 malinoff UtahDave, Why the hell in https://github.com/saltstack/salt/blob/develop/salt/modules/saltutil.py#L36 env argument is not 'base'? And why is it trying to parse top.sls file instead of just taking 'base' env by default?
12:34 mwillhite joined #salt
12:36 UtahDave malinoff: because that way the top file can determine what environment to use
12:36 Corey joined #salt
12:38 xl1 joined #salt
12:38 fr0de NV: thanks - I'll keep that in mind, if my workplace decides to go with salt
12:38 fr0de :)
12:38 malinoff UtahDave, strange thing, all other modules use 'base' as default value, and _sync isn't
12:41 malinoff UtahDave, a slight difference with the rest of salt
12:42 bezaban I got the AOK to start testing a salt deployment :)
12:43 bezaban a little concerned it won't do anything with log centralization (off a random internet article)
12:44 joehh malinoff: did you get an answer/figure out yesterdays question?
12:44 viq bezaban: it has a syslog returner
12:45 viq And I think I even saw somewhere something about logstash
12:45 joehh re the two functions from a single module
12:45 bhosmer_ joined #salt
12:46 malinoff joehh, nope :) looks like that this restriction  is taking place deeply in salt.
12:46 bhosmer_ joined #salt
12:46 viq https://salt.readthedocs.org/en/latest/ref/configuration/logging/handlers/salt.log.handlers.logstash_mod.html
12:47 viq bezaban: ^
12:49 UtahDave malinoff: It's possible that is an oversight.  My brain is too fried right now to know.   :)
12:49 malinoff UtahDave, I'm going to create a pull request to change this behavior. Tomorrow :)
12:49 joehh UtahDave: I like the look of the new docs!
12:50 joehh or is that just a read the docs thing
12:50 joehh looks like it is...
12:52 UtahDave joehh: that is nice!
12:52 UtahDave much better than what they used to have
12:53 bezaban viq: neat, thanks :)
12:54 bezaban need some setup for log signing too.  Not sure if I'm going to sign the entire log or set up something that signs each entry
12:54 bezaban got to check requirements
12:55 bezaban sorry, thinking out loud
12:55 tempspace man Dave, what are you doing in here so early?
12:55 tempspace (or late...)
12:56 snikkers joined #salt
12:57 blee joined #salt
13:00 UtahDave joined #salt
13:05 NV fr0de: you won't regret it - I'm in the process of going from Puppet, it's glorious :D
13:08 fr0de NV: if you don't mind me asking, how many systems are you migrating (I'll have about 1,000, in various roles (web server, app server, db server) running a mix of stuff (tomcat, unicorn, custom apps..)
13:08 fr0de ?
13:10 Pate_ joined #salt
13:13 Pate_ Database Migration question when using TeamCity and Salt: mySQL is deployed on RDS. Using Alembic for database migrations. Connection string is templated by Pillar. How to tell one minion to upgrade the database?
13:13 Pate_ Preferably TeamCity should call 'alembic upgrade head', but TeamCity doesn't know the connection string and possibly does not have python/alembic.
13:15 gareth_ joined #salt
13:16 NV fr0de: not that many, we have around that many systems but only a fraction (100 or so) managed by puppet for various reasons (the primary of which is my section doesn't control them all)
13:17 fr0de NV: sounds like you should have a pretty easy time of it, then. especially if the systems are fairly similar.. :)
13:18 NV quite the opposite really, we're relatively small in the grand scheme of things, most things usually only had 1 or 2 nodes at most for a particular service
13:19 fr0de ahh.. dozens of "handcrafted works of art" servers?
13:19 NV so quite a lot of different configuration, although a fair amount of shared config between the hosts too for general stuff like user accounts, mail, etc
13:19 NV I work at a University, go figure
13:19 fr0de I guess, migrate the common stuff first.. :)
13:20 NV I wouldn't call them handcrafted works of art, just either the software used either doesn't scale out, or we just haven't had enough load to warrant more than 2 for the task (I mean we have 4 frontend webservers for example, but thats the exception not the norm)
13:20 NV but we do run a _LOT_ of different systems
13:20 elm-1 joined #salt
13:20 fr0de lots of 'legacy' systems?
13:21 NV because every second faculty decides they want to use a different bit of software to do pretty much the same thing, etc
13:21 NV and of course thanks to politics, they usually get their way
13:21 fr0de lol, I feel your pain :)
13:21 NV fair amount of legacy stuff, but that also predates our puppet installation
13:22 NV most of the legacy stuff runs on old sloaris sparc (or as I like to call them, craps) machines
13:22 NV with oracle shafing us on maintenance for the hardware cause its so far EOL'd
13:22 fr0de :D - gotta love them!
13:23 NV but yeah, we do middleware, web and applications
13:24 fr0de if my workplace decides to migrate to salt, we have three legacy puppet environments to manage, including some really old servers... *sigh*
13:25 NV networking systems (including stuff like dns), and backend systems like storage (fileshares, san, etc) are all managed by different sections - none of which have adopted any form of configuration management
13:25 NV heck, i don't think they even use revision control for stuff...
13:25 fr0de scary..
13:25 NV yeah...
13:26 NV what's worse is the guys that manage backups and the virtual infrastructure invested in a rather terrible bit of software
13:26 NV about once every 2-4 weeks one VM will shit itself
13:26 NV at random
13:26 fr0de ouch!
13:26 fr0de what software? vmware?
13:26 NV yeah
13:27 NV the backup software does a quiesced snapshot of the linux vm
13:27 NV which fails (according to the vmware.log)
13:27 NV and the linux kernel seems to lose its ability to talk to the disk at all as a result
13:27 NV until its rebooted
13:27 fr0de so the backup kills the vm? beautiful...
13:27 NV yup..
13:28 NV combine with a very dangerous cowboy who practically invented linux if you believe what he says who refuses to admit responsibility for anything, fun times
13:30 Brew joined #salt
13:30 NV speaing of which, should probably hit the sack so im not dead in the morning :P
13:31 * NV gallops off into the sunset
13:32 fr0de cheers
13:32 fr0de and good luck!!
13:40 juicer2 joined #salt
13:42 amahon joined #salt
13:44 ipmb joined #salt
13:48 gkze joined #salt
13:49 Minioncito hi all again, just want to confirm... I'm opening an issue for reporting a bug, is here the right place? https://github.com/saltstack/salt/issues/
13:49 creich is there a convenient way to add one new user into the sudoers file?
13:53 viq bezaban: if you want signing I believe your options are either samhain or ossec
13:53 viq creich: either add him to group that has sudo rights, or work with /etc/sudoers.d/
13:54 cron0 joined #salt
13:57 viq https://gist.github.com/viq/7353629 - any idea why it creates groups, but does not assign users to them ?
14:00 gareth_ Hello gentlemen. First off, I must apologise for joining the channel as a noob and then asking questions straight off the bat. I've been trying to create a Mongo ext_pillar with Salt 0.16.4 on RHEL6 and have come up against the following error:
14:00 gareth_ [CRITICAL] Specified ext_pillar interface mongo is unavailable
14:00 tyler-baker joined #salt
14:00 tyler-baker joined #salt
14:00 gareth_ A quick hack in /usr/lib/python2.6/site-packages/salt/pillar/__init__.py leads me to wonder if the mongo ext_pillar module is being loaded at all.
14:01 gareth_ As the contents of 'self.ext_pillars' appears to only contain the following:
14:02 gareth_ cmd_yaml, cmd_json, libvirt, hiera, puppet, cobbler
14:02 gareth_ I was wondering whether anyone else has come up against this problem or if anyone is using a MongoDB ext_pillar, please?
14:03 networkpadawan joined #salt
14:04 honestly [A
14:04 honestly augh
14:04 timoguin joined #salt
14:05 snikkers gareth_: not sure, but have you installed the mongodb python module ?
14:06 bhosmer joined #salt
14:06 oz_akan_ joined #salt
14:07 oz_akan_ joined #salt
14:09 baffle Hi, I have a grain on a few nodes, to set wich roles they have. Basically they are a dict entry with a list. Example: ('role', ['zookeeper', 'cassandra'])
14:10 bhosmer joined #salt
14:10 baffle How would I in a jinja template get a list of all hosts (or, preferably, ipv4 addresses) that has the role "zookeeper"?
14:11 baffle I'm trying to use the "publish.publish" module, and I seem to be able to get a list of all minions that have a role.. But not targeting one specific role. :)
14:13 baffle And should I be using the saltmine instead? If so, how? Doesn't seem to be much documentation. :)
14:13 racooper joined #salt
14:15 gareth_ joined #salt
14:16 jslatts joined #salt
14:16 toastedpenguin joined #salt
14:22 Khollowa_ joined #salt
14:23 mwillhite joined #salt
14:26 aleszoulek joined #salt
14:27 ashtonian joined #salt
14:28 Guest18138 joined #salt
14:29 networkpadawan joined #salt
14:29 jumperswitch joined #salt
14:30 santagada joined #salt
14:30 Furao joined #salt
14:32 viq Any of the salt gurus would know maybe why it doesn't put the users in the groups (I think) I tell it to?
14:32 juicer2 joined #salt
14:32 scott_w joined #salt
14:33 baffle viq: You probably should show your examples of your state/pillar configuration.
14:33 quickdry21 joined #salt
14:34 tty-1_ joined #salt
14:34 Furao_ joined #salt
14:35 viq https://gist.github.com/viq/7353629 - any idea why it creates groups, but does not assign users to them ?
14:35 viq baffle: ^
14:36 brianhicks joined #salt
14:36 UtahDave joined #salt
14:36 baffle Should this give me IPv4 addresses of all hosts that have role: ['zookeeper'] set in grains? salt '*' publish.publish 'role:zookeeper' 'grains.item' 'ipv4' expr_form=grain
14:37 mannyt joined #salt
14:39 baffle viq: It just confused me, don't think I can help. :)
14:40 viq :P
14:43 aleszoulek joined #salt
14:46 fishpen0 joined #salt
14:46 baffle {% for host,ips in salt['publish.publish']('role:zookeeper', 'grains.item', 'ipv4', expr_form='grain').items() %}{{ ips.get('ipv4')|first }},{%- endfor %}
14:46 baffle It works, but I somehow think this is bad..
14:51 fishpen01 joined #salt
14:51 carmony Hey, is there a interactive CLI for salt?
14:51 micah_chatt joined #salt
14:52 EricHogue joined #salt
14:52 baffle Haven't tried. Maybe you can just do "import salt" in a python shell? :)
14:54 tty-1 joined #salt
14:54 fishpen01 joined #salt
14:54 xmltok joined #salt
14:55 sroegner joined #salt
14:55 timoguin carlos, there was totally a useful interactive shell for cmd.run that someone shared recently
14:55 timoguin errr carmony
14:55 timoguin i'm trying to find it
14:56 timoguin carmony, https://gist.github.com/stevearc/6027355
14:56 timoguin might be useful for you
14:57 viq Hm, functools, so requires func installed
14:57 mgw joined #salt
14:59 fishpen0 joined #salt
15:00 Minioncito UtahDave: hi, I opened that issue: https://github.com/saltstack/salt/issues/8322
15:01 ipmb joined #salt
15:01 carmony timoguin: awesome. I'm doing a live tutorial, and you just proved to 20+ people how awesome the IRC channel is :)
15:01 mr_chris joined #salt
15:01 timoguin awesome! :)
15:05 slav0nic salt-ssh does't work from non-roor user =/?
15:06 ashtonian joined #salt
15:08 mm_mannen joined #salt
15:08 Kareem joined #salt
15:11 amckinley joined #salt
15:13 petersabaini joined #salt
15:17 kermit joined #salt
15:21 slav0nic joined #salt
15:22 elm- joined #salt
15:23 jankowiak joined #salt
15:28 forrest joined #salt
15:28 baffle Hmm, I want to call some of the functions in a object inside a jinja tempalte. Right now I'm using a throwaway object to actually get to my object, is there a better way? I.e. I'm doing this: {% set tmp = zookeepers.append(ips.get('ipv4')|first) %} where I really would want to just do {% zookeepers.append(ips.get('ipv4')|first) %}
15:29 fwiles joined #salt
15:29 forrest baffle, are you unable to use pillar/grains for that data?
15:30 logix812 joined #salt
15:30 UtahDave joined #salt
15:31 JulianGindi joined #salt
15:32 petersabaini hi — i'd like to enable a service (zabbix) on boot but that doesn't work out. instead i get a message "Enable is not available on this minion, service zabbix-agent is in the desired state". whats going on?
15:33 forrest petersabaini, which OS?
15:33 EricHogue left #salt
15:34 baffle forrest: Well. I can statically define that these hosts/IPs are zookeepers in pillar. That works. Or I can assign "roles" on the hosts by using grains, and doing magic to get the IPs of the hosts that has that role.. Wich I'm doing.. :-/
15:34 petersabaini forrest: SL 6.3, salt 0.17.1
15:34 baffle I kinda feel I'm doing it all wrong. :)
15:34 jcsp1 joined #salt
15:34 forrest baffle, ahh I see.
15:35 forrest baffle, are you setting multiple IPs in there as a list or what?
15:35 networkpadawan joined #salt
15:36 forrest petersabaini, can you log onto the actual box and check the status of the service please?
15:36 forrest seems like it's zabbix agent?
15:36 petersabaini forrest: exactly, and its running
15:36 baffle forrest: Yeahh.. Well, here is what I do..
15:36 baffle {% set zookeepers = [] %}
15:36 baffle {% for host,ips in salt['publish.publish']('role:zookeeper', 'grains.item', 'ipv4', expr_form='grain').items() %} {% set tmp = zookeepers.append(ips.get('ipv4')|first) %}
15:36 baffle {% endfor %}
15:36 baffle # zookeepers: {{ zookeepers|join(',') }}
15:37 forrest petersabaini, ok can you show me what your state file looks like then where the service is enabled?
15:37 forrest scientific linux is a RHEL offshoot right?
15:37 forrest I can never remember.
15:37 baffle So now "zookeepers" contain (in my case) "10.101.0.88,10.101.0.81,10.101.0.86" :-)
15:37 petersabaini forrest: yes, its basically RHEL 6. state file excerpt: zabbix-agent:
15:37 petersabaini service:
15:37 petersabaini - enabled
15:38 forrest ok, can we start with a chkconfig --list | grep zabbix to ensure that the service was enabled for the associated runlevels? And if you want the service itself to be turned on, you needed '- running' under service
15:39 forrest baffle, yea that is not so great, hmm
15:39 forrest I'd probably create that as a list inside of pillar baffle
15:39 baffle ORLY?
15:39 forrest well, it's just a lot of work
15:39 forrest is what I mean by not so great
15:39 forrest it's a good solution :P
15:40 petersabaini # chkconfig | grep zabbix
15:40 petersabaini zabbix-agent   0:off1:off2:off3:off4:off5:off6:off
15:40 forrest so ptersabaini, what happens when you do chkconfig zabbix on
15:40 forrest *petersabaini
15:40 petersabaini forrest: sure, that works but i thought salt would do that?
15:40 forrest it should, that's why I wanted to make sure it worked on the system
15:41 petersabaini -bash-4.1# chkconfig zabbix-agent on
15:41 forrest ok chkconfig it back off, go back to the master, re-run your salt command from the master, with -l debug
15:41 petersabaini -bash-4.1# chkconfig | grep zabbix
15:41 petersabaini zabbix-agent   0:off1:off2:on3:on4:on5:on6:off
15:41 forrest maybe we can get some more output
15:41 khodgson joined #salt
15:44 viq https://gist.github.com/viq/7353629 - any idea why it creates groups, but does not assign users to them ?
15:44 petersabaini forrest: i've put it into a gist: https://gist.github.com/sabaini/7356722#file-gistfile1-txt (some keys elided)
15:46 forrest petersabaini, looks like the enable error isn't unique to zabbix-agent.
15:46 petersabaini forrest: correct
15:46 forrest Is all of this in the same state file? Or is there a single state file for zabbix?
15:46 scott_w joined #salt
15:46 petersabaini no, thats for several state files
15:47 scott_w joined #salt
15:48 khodgson joined #salt
15:49 faust joined #salt
15:50 jY if i have this "- source: salt://tripwire/Makefile"  is there anyway for the master to know when the file is changed so it can push it out to the minions?
15:50 scott_w yo, is it possible to set a salt module to return a failure result for the purposes of testing?
15:50 scott_w jY: not to the best of my knowledge, it will check it everytime you run state.highstate
15:50 amckinley joined #salt
15:51 jY ok
15:51 forrest ptersabaini, you didn't by chance just upgrade to 0.17.1? Is this a brand new build?
15:52 petersabaini forrest: yes, the minion is on a pretty new box
15:52 scott_w jY: try a cronjob running every x hours?
15:52 forrest what about the master though
15:53 petersabaini that one has been upgraded some time ago, about 2 wks
15:54 forrest hmm, did you restart the mater after doing the update? And did the enable just now stop working?
15:55 petersabaini forrest the master was restarted after the upgrade.
15:55 petersabaini i believe the enable stuff did work before, but not sure about that
15:55 khodgson joined #salt
15:55 viq OK, WTF? http://pbot.rmdir.de/8c2S5jpATiEDKC6_yik4ww
15:56 PoLuX joined #salt
15:56 petersabaini forrest: do you think there
15:56 cnelsonsic joined #salt
15:56 petersabaini * theres some cached state around?
15:57 giantlock joined #salt
15:57 forrest petersabaini, well, can we confirm that these states work on another machine?
15:58 forrest viq, I haven't had a chance to look at your thing yet, so many states :P
15:58 forrest and this whole 'work' thing
15:58 forrest getting in theway
15:58 forrest *the way
15:58 sciyoshi any way to get feedback during a long-running highstate?
15:58 sciyoshi like each state's status as it completes
15:58 forrest sciyoshi, you can check the state of the job
15:58 viq forrest: aye, that it sometimes does ;)
15:58 sciyoshi forrest: hmm?
15:59 forrest sciyoshi, try salt-run jobs.active
15:59 petersabaini forrest: got the same problem on another minion
15:59 forrest then salt-run jobs.lookup_jid <id of running job>
15:59 forrest petersabaini, also SL, also new?
16:00 petersabaini forrest: yes
16:00 forrest do you have any non SL machines?
16:00 forrest other rhel based distros that is
16:00 sciyoshi forrest: doesn't return anything until the whole state run is finished though :/
16:01 forrest sciyoshi, ahh ok bummer, I couldn't remember off the top of my head, that would be a cool feature
16:01 petersabaini forrest: sorry, only SL and ubuntu
16:01 forrest does it work on ubuntu?
16:01 sciyoshi do all the jobs use the returner system for their results?
16:01 sciyoshi maybe some kind of partial results would be doable
16:02 forrest I can't remember sciyoshi :\
16:02 JulianGindi joined #salt
16:02 sciyoshi k guess i'd have to dig into code :P
16:02 utahcon On minion "crwmi-test" I am expecting revision to be "release_4.15.0" but I am getting "release_4.16.0" which means this isn't working how I would expect it to, is it possible to do what I am trying, but I am doing it wrong? : http://pastebin.com/gQ5YjScP
16:03 forrest yea utahcon
16:03 forrest 'cr*': matches crwmi-test
16:03 khodgson joined #salt
16:03 forrest so all the base stuff gets applied
16:03 elm- left #salt
16:03 utahcon forrest: so it is one and done?
16:03 petersabaini forrest: interestingly, 'enable' did work on another, not freshly built box. i'
16:03 petersabaini * i'll investigate with ubuntu some more
16:04 forrest petersabaini, which version is on that SL machine that worked? Of Salt
16:04 forrest utahcon, honestly I don't know, previously it seemed like it overwrote, but this is the second instance since the 0.17 release where I've seen an issue like yours where it doesn't get overwritten.
16:04 petersabaini 0.17.1 as well
16:05 utahcon forrest: I expected it to work like the salt top.sls, which seems to work the way I am expecting... hmm
16:05 forrest petersabaini, weird, can you try and restart the minion service on the newer machines, and then on the master trash /var/cache/salt, then restart the salt master and try again?
16:05 QauntumRiff joined #salt
16:05 forrest utahcon, yea I don't get why it pulls it in that way.
16:05 tty-1_ joined #salt
16:06 forrest but I'm pretty sure it just says 'ok this matches this first item, it gets base, bye'
16:06 forrest as a test, try changing cr* under base to crwmi
16:06 forrest then the test state should get applied
16:06 QauntumRiff I'm getting a strange error on most of my minions after updating to 17 (i'm running centos 5.9, and using EPEL for updates)
16:06 QauntumRiff [salt.minion][WARNING ] TypeError encountered executing saltutil.find_job: __init__() got an unexpected keyword argument 'with_communicate'. See debug log for more info.  Possibly a missing arguments issue:  ArgSpec(args=['jid'], varargs=None, keywords=None, defaults=None)
16:07 pdayton joined #salt
16:07 QauntumRiff rpm -qa | grep salt: salt-minion-0.17.1-1.el5, salt-0.17.1-1.el5
16:08 forrest utahcon, or maybe try moving base to the bottom?
16:08 gasbakid joined #salt
16:08 rockey QauntumRiff: that's a python error mostly thrown when you use python syntax wrong
16:08 forrest QauntumRiff, I only see this issue from a monht ago that's similar to that: https://github.com/saltstack/salt/issues/7717
16:09 rockey surely in a function
16:09 ctdawe joined #salt
16:09 rockey (ie "" instead of '')
16:09 fishpen01 joined #salt
16:10 khodgson joined #salt
16:10 terminalmage forrest: I believe halite is available for OpenSUSE
16:10 forrest terminalmage, well screw you buddy, I already put in the doc update pull request! :P
16:10 terminalmage :)
16:13 santagada left #salt
16:13 mwillhite joined #salt
16:15 aurigus Does anyone know if there is a way to see just which states failed? I get a nice summary with the latest update but with lots of states it is hard to look at the log
16:15 aurigus (that is, when running state.highstate)
16:15 forrest aurigus, failed states should print in a different color in that output
16:15 renoirb left #salt
16:16 aurigus Hmm, not im my current client at least. But I have colors and such turned off
16:17 aurigus I'd also have to increase my scrollback buffer in that case
16:17 QauntumRiff forrest: rockey: thanks, is there any good way to 'narrow it down' other than uncomment everything, and start adding one at a time?
16:18 timoguin joined #salt
16:18 QauntumRiff I guess I could manually pass the state files one at a time as well, instead of state.highstate
16:18 forrest yea I'd say pass each state
16:18 forrest and find which one is an issue, then trace what gets pulled in, etc.
16:18 cachedout joined #salt
16:19 khodgson joined #salt
16:20 slav0nic joined #salt
16:22 s0undt3ch joined #salt
16:23 forrest hey s0undt3ch, can you either reject or add a comment on issue 248 for the bootstrap?
16:24 noob2 left #salt
16:24 forrest diegows also has a pretty cool addition to merge in
16:26 utahcon on the cli how do you pass cwd to cmd.run? "salt '*' cmd.run name="command
16:26 utahcon " cwd="/some/dir"
16:26 utahcon ?
16:26 forrest I don't understand
16:27 forrest c,d.run 'ls -l /some/dir'
16:27 forrest *md
16:27 utahcon trying to run git branch in /var/www/common
16:27 s0undt3ch forrest: yeah, I've been holding that one back
16:27 mwillhite joined #salt
16:27 forrest utahcon, cmd.run 'cd /var/ww/common; your command'?
16:27 khodgson joined #salt
16:28 utahcon thanks
16:28 forrest s0undt3ch, Well, you addressed it previously in that question I asked, which I notated in there, I just wasn't sure if you were gonna provide an update or response back as well.
16:28 forrest I spoke to ifnull in here when he was around as well and he wanted to wait for your opinion *shrug*
16:28 viq utahcon: how about using http://docs.saltstack.com/ref/modules/all/salt.modules.git.html#module-salt.modules.git  ;)
16:29 forrest viq, that doesn't support branch
16:29 viq ah
16:29 viq uh, it says it does
16:29 viq CLI Examples:
16:29 viq salt '*' git.checkout /path/to/repo somebranch user=jeff
16:30 forrest yea that's on a checkout though
16:30 forrest not sure if utahcon is doing it that way, or has local content he has to branch to :P
16:30 forrest but that would be a better way to do it.
16:30 QauntumRiff this might be a really dumb question.. but I have only used the top.sls before.. what is the syntax to do a single state file? ie, "salt 'testServer1' ?????? test.sls" ?
16:31 forrest state.sls
16:31 utahcon viq: that'd be to obvious!
16:31 forrest that's actually a good point, should probably update the states tutorial on the first page with that data.
16:32 utahcon oh, yeah I am not checking out
16:32 utahcon checking which branch it is on
16:32 s0undt3ch forrest: I'm updating the ticket
16:32 jdenning joined #salt
16:32 druonysus joined #salt
16:32 forrest s0undt3ch, awesome.
16:32 forrest sorry to bug you about it, it's just been in there a while
16:34 ctdawe joined #salt
16:34 pipps_ joined #salt
16:35 khodgson joined #salt
16:35 s0undt3ch forrest: yeah, I've been getting less free time these past days, which it actually a good thing
16:35 forrest heh
16:35 s0undt3ch s/it/is
16:36 teskew joined #salt
16:36 forrest I can never remember, do you work on the Salt team now?
16:37 viq http://pbot.rmdir.de/8c2S5jpATiEDKC6_yik4ww - any idea what I am doing wrong?
16:37 s0undt3ch forrest: Nope :)
16:41 kaptk2 joined #salt
16:42 ravibhure joined #salt
16:44 QauntumRiff so.. I have no states for my minion.  I run 'salt-minion -l debug' and then do a 'test.ping' from the server
16:45 QauntumRiff I see my process list going over and over and over (I have thousands of processes, very large server)
16:45 QauntumRiff and this error intermixed: Traceback (most recent call last):
16:45 QauntumRiff File "/usr/lib64/python2.6/logging/handlers.py", line 796, in emit
16:45 QauntumRiff self.socket.send(msg)
16:45 QauntumRiff error: [Errno 105] No buffer space available
16:45 QauntumRiff [INFO    ] Returning information for job: 20131107104304434802
16:45 QauntumRiff is that related to one of the new 'monitoring' features in 17? that would also cause the error I pasted above ?
16:46 Corey QauntumRiff: What OS is this on?
16:46 QauntumRiff centos 5.9
16:47 QauntumRiff salt-minion --versions
16:47 QauntumRiff Salt: 0.17.1
16:47 QauntumRiff Python: 2.6.8 (unknown, Nov  7 2012, 14:47:45)
16:47 QauntumRiff Jinja2: unknown
16:47 QauntumRiff M2Crypto: 0.21.1
16:47 QauntumRiff msgpack-python: 0.1.12
16:47 QauntumRiff msgpack-pure: Not Installed
16:47 QauntumRiff pycrypto: 2.3
16:47 QauntumRiff PyYAML: 3.08
16:47 QauntumRiff PyZMQ: 2.1.9
16:47 QauntumRiff ZMQ: 2.1.9
16:49 Corey QauntumRiff: For god's sake dude, pastebin.
16:49 QauntumRiff sorry, I'll use that
16:51 racooper also, I'd recommend doing a yum update to 5.10. there are several dozen packages upgraded since 5.9, including some kernel security fixes.
16:53 Corey *nod* 5.9 is deprecated.
16:54 viq Speaking of pastebin
16:54 linjan_ joined #salt
16:54 viq http://pbot.rmdir.de/8c2S5jpATiEDKC6_yik4ww - any idea what I am doing wrong?
16:54 QauntumRiff also reproduceable on 5.8..  Sorry, were actually running OEL 5.9.. not centos.  There is no release yet of 5.10 for oracle.
16:54 QauntumRiff and unfortunately, not all our DB software is 'certified' on the 6.x series yet
16:55 xmltok joined #salt
16:55 forrest ahh Oracle
16:55 forrest my favorite word every day when I come to work
16:55 QauntumRiff forrest: yeah.. I know
16:55 nmistry joined #salt
16:55 forrest I don't know how they stay in business
16:55 forrest oh wait I do
16:55 forrest because they have a flashy conference
16:56 forrest and executives eat it up
16:59 mgw joined #salt
17:00 grep_awesome Oracle stays in business because Larry Ellison is a fantastic lier and a better salesman
17:01 grep_awesome *liar
17:01 danielbachhuber joined #salt
17:01 whiskybar joined #salt
17:03 mwmnj joined #salt
17:03 QauntumRiff but anyways, I had our master (running Centos 6.4)  somehow update itself from v.16 to v.17 of salt.  and things got really weird.  people suggested updating the minions.. so I did.. (via epel) and now i'm seeing these types of errors
17:03 KyleG joined #salt
17:03 KyleG joined #salt
17:03 QauntumRiff is there any way to downgrade the minions?
17:03 cnelsonsic joined #salt
17:05 racooper you can get the 0.16.4 RPMs from koji. I had the link yesterday...just a sec
17:06 racooper http://koji.fedoraproject.org/koji/buildinfo?buildID=462945
17:06 racooper if you do that, don't forget to add an exclude to your epel config.
17:07 redondos joined #salt
17:07 QauntumRiff racooper: thanks.. then do something like salt 'server1' run.cmd 'rpm -i --force <location to rpm>' ?
17:07 forrest grep_awesome, yea pretty much
17:08 racooper I think you'll have to do something different, becuase you'll need to remove the newer minion first.
17:08 grep_awesome forrest: if I remember correctly, they never had a 1.0 release. They went straight into 2.0 so that it looked like an established product. Even then that wasn't finished. He sold it and the told his team to finish it with "bug fixes"
17:08 viq Anyone knows WTF is going on here? http://pbot.rmdir.de/8c2S5jpATiEDKC6_yik4ww
17:08 forrest heh
17:08 Minioncito QauntumRiff: why don't you have a frozen repo?
17:09 QauntumRiff minioncit: yeah.. I know.. one of those "on the list" kind of things
17:09 forrest viq you're still working on that?
17:10 viq forrest: kinda at a loss, so shooting it out every once in a while, maybe someone will know, or jarr me onto a new track
17:10 forrest viq, gotcha
17:11 Minioncito QauntumRiff: it's very dangerous not having that if you have software that requires specific versions of another software... Setting up a frozen repository for a kind of software and another for the system itself is pretty quick
17:11 anitak joined #salt
17:12 anitak1 joined #salt
17:15 troyready joined #salt
17:18 khodgson joined #salt
17:21 VertigoRay joined #salt
17:22 jalbretsen joined #salt
17:22 pipps joined #salt
17:24 VertigoRay joined #salt
17:24 pipps_ joined #salt
17:25 VertigoRay joined #salt
17:26 utahcon viq: So do pillars not cascade like salt(s)?
17:27 viq utahcon: not sure what you mean
17:27 utahcon viq: http://pastebin.com/gQ5YjScP
17:28 utahcon on minion crwmi-test I would expect revision to be "release_4.15.0", instead I am getting the value set in cr "release_4.16.0"
17:28 utahcon it would seem matching stops at the first match in top.sls
17:28 viq utahcon: 'salt <minion> pillar.items' - see for yourself ;)
17:29 dave_den joined #salt
17:29 utahcon I guess I am more asking, should it cascade... becuase I am not seeing that to be the case
17:29 viq utahcon: what does that command give you?
17:31 viq utahcon: it could be that you do get all the data, but don't use it, or something
17:31 nmws Hello everyone. Does anyone know if Salt's virt module can provision vmware or aws vms?
17:32 utahcon hmm, interesting, that command had shows an error, thanks! That is something I can work with.
17:32 Chocobo joined #salt
17:32 Chocobo joined #salt
17:32 viq nmws: not quite an answer, but also have a look at salt-cloud
17:33 viq utahcon: when changing pillars, run saltutil.refresh_pillar; when running saltutil.refresh_pillar look at master log for errors
17:33 jeffstoner joined #salt
17:34 nmws thanks viq will take a look :)
17:34 claudep joined #salt
17:35 claudep it would be soooo nice if the pattern replace param was a bit more documented
17:36 forrest which doc claudep?
17:36 viq How to modify this to make this work? Right now it sees only the last group listed. https://gist.github.com/viq/7358604
17:36 claudep forrest: this one http://docs.saltstack.com/ref/modules/all/salt.modules.file.html#salt.modules.file.rename
17:36 jeffstoner is it possible to mount and umount file systems using the file system label (not the mount point or device name)?
17:36 claudep e.g.: "bind-address\s*=
17:37 claudep found unknown escape character 's'
17:38 Iwirada left #salt
17:38 viq claudep: uhm, http://docs.python.org/3/library/re.html#re.sub ?
17:39 claudep viq: excepted that \s is valid in re.sub()
17:39 viq claudep: try ' instead of " ?
17:40 bhosmer joined #salt
17:40 viq or double-escape
17:40 * viq nudges https://gist.github.com/viq/7358604 towards the gurus ;)
17:42 utahcon viq isnt the guru?!
17:42 claudep viq: thanks, but grmblll... yaml crap!
17:42 viq utahcon: no, I'm just starting :P
17:42 forrest claudep
17:42 forrest https://github.com/saltstack/salt/issues/8326
17:43 utahcon viq: you have been a blessing, give your self more credit ;)
17:44 viq utahcon: my only claim to fame is that I read a bunch of docs, my extent of experience with salt can be seen here https://github.com/viq/cm-lab-salt  :P
17:45 forrest ok viq let's look at this thing
17:45 forrest so the groups get created, but the users aren't getting added?
17:45 s0undt3ch forrest: commented
17:45 viq forrest: well, with that the show_sls actually shows only the last of each of the pair of the groups as "make sure they are there"
17:46 forrest s0undt3ch, sweet
17:46 viq While with that list of groups, users state actually finally sees that users are to be part of some groups... But only the first group on the list.
17:47 viq Actually I guess I'll add third group to those lists, to see if it sees only one, or cuts off one
17:47 forrest ok
17:48 whitepaws guys, i'm trying to use zeromq 3.2.2 on my aging centos 5 boxes.
17:48 teebes joined #salt
17:48 whitepaws i was able to rebuild the rpm for python26-zmq based on the python26-zmq-2.1.9 SRPM but linked against 3.2.2 headers etc.
17:48 whitepaws it seems to have worked...
17:48 teskew bugs like this are a real pain in the ass when it affects hundreds of machines after upgrading.  https://github.com/saltstack/salt/pull/7953   getting a config file to the minions would be easy, ya know, if the minion wouldn't crash on startup.
17:48 rgarcia_ joined #salt
17:48 whitepaws is there any thing i should try to see if anything else breaks?
17:49 whitepaws salt '*' test.ping   ...comes back fine
17:49 forrest whitepaws, try running a state
17:49 forrest just something basic
17:49 whitepaws forrest that sounds like a great idea. i haven't graduated to states yet ;) so, no time like the present. thanks
17:49 ajw0100 joined #salt
17:50 forrest whitepaws, sounds good!
17:50 viq forrest: hah. groups to be created: only last on each of those lists. Groups for user to be in: all but last in relevant list.
17:50 forrest viq, :(
17:50 forrest why you gotta break it and make a weird error state
17:51 forrest I forgot, what was your reason for having so many cascading pillars?
17:51 viq forrest: ooh, I do have a talent for exposing weird bugs / making weird errors, don't I ? ;P
17:51 forrest only in your own code viq :P
17:51 viq forrest: different groups of users, belonging to different groups based on machine they end up on
17:52 forrest So why do you not have a top file that based on the system includes specific states that contain all the users, with an overall map.jinja file that determines group names based on OS?
17:53 viq But the group names are later to be not only based on OS but also specific machine. And, well, I thought pillar would be a cleaner way to do it...
17:54 forrest Hmm
17:55 viq apply clue-by-four if I'm thinking wrong ;)
17:56 forrest To me it doesn't seem cleaner, especially since you aren't storing user hashes
17:56 forrest which would be the main reason to use pillar
17:56 utahcon turns out my matching was off <facepalm>
17:56 forrest but I'm also not coming at it from your perspective.
17:56 forrest utahcon, for the ordering?
17:56 forrest in your top?
17:57 viq forrest: I would use the hashes... Except it seems the only cross-platform mechanism is MD5, ugh
17:57 forrest yea that sucks
17:57 bhosmer joined #salt
17:57 utahcon forrest: yeah I will post my fix
17:57 forrest cool
17:58 bemehow joined #salt
17:59 utahcon http://pastebin.com/diff.php?i=MLrYaSHR
17:59 utahcon I shuffled a few things, but basically my crwmi-test is actually crwmi-test.rest.of.my.domain
17:59 utahcon so I had to add *
17:59 forrest ahhh ok
17:59 forrest awesome
17:59 utahcon purely boneheaded
18:01 avienu joined #salt
18:01 khodgson joined #salt
18:01 viq forrest: and if I were to store password hashes, what would you recommend?
18:03 khodgson joined #salt
18:03 viq Or would it still be "most in states, just sensitive data in pillars" ?
18:03 viq And just general advice to move that data to states?
18:06 forrest viq, password hashes should always go in pillar for security
18:06 forrest It depends, I've seen some people store all their users in pillar
18:06 forrest but they weren't doing it the same way you are where different OS, and different servers were getting unique users
18:06 viq That's what I was trying to do
18:06 forrest it was just 'these users go on all my boxes'
18:06 viq mhmm
18:07 forrest You might scrape the mailing list, I know I've seen a similar setup before :\
18:07 forrest just don't remember where
18:07 khodgson joined #salt
18:07 viq well, there's for example https://gist.github.com/UtahDave/3785738
18:08 viq or http://russell.ballestrini.net/understanding-salt-stack-user-and-group-management/
18:08 forrest ahh yea UtahDave's
18:09 viq yeah, that's what I'm basing mine on
18:10 viq Or, well, trying to ;)
18:11 modafinil is this a bug? [INFO    ] Executing state file.recurse for /home/nick [INFO    ] {'/home/nick/../nickc': {'group': 'nick', 'user': 'nick'}, '/home/nick/../nickc/.ssh': {'group': 'nick', 'user': 'nick'}}
18:12 amckinley joined #salt
18:12 modafinil it's sure buggin' me ba-dum-bum-pshhhh
18:12 rockey you'll operate in the folder /home/nickc/, not /home/nick as intended
18:12 modafinil why is that happening though?
18:13 rockey depends on what you're tryint todo :)
18:13 modafinil i am trying to have 3 users 'nick', 'nickb', 'nickc' -- each with a home dir
18:13 gildegoma joined #salt
18:13 modafinil and not have salt chown all of it to 'nick' and then back to the right owner afterwards D:
18:13 modafinil every run
18:14 rockey then you should do an array of users and do /home/<user-input>/ instead of /home/nick/../nickc/ (which is the same as /home/nickc/)
18:14 srage joined #salt
18:15 renoirb joined #salt
18:15 modafinil right, thats what im doing
18:15 modafinil https://gist.github.com/jzellner/e4fa3a56d13a0e9b6adf
18:15 modafinil that wacky log is whats HAPPENING, not what i want to happen :D
18:17 sfz joined #salt
18:18 viq forrest: OK, that's enough for today, thanks for help and a kind word :)
18:19 forrest viq, sorry I couldn't be of more help, let me know if you figure it out :\
18:19 danielbachhuber joined #salt
18:19 viq forrest: will do
18:20 josephholsten joined #salt
18:22 Kareem joined #salt
18:22 rockey modafinil: try using pillars
18:22 rockey modafinil: http://dev.mlsdigital.net/posts/SaltStackBeyondJinjaStates/
18:22 rockey there is an example
18:22 rockey modafinil: your code and iteration doesnt look that bad tbh, dont know why fucks with you
18:23 modafinil yeah, i will check that out -- i cannot figure out wtf is up with these 'nicks' -- if i add more nickX users, it happens to all of them as well
18:23 modafinil there is nothing funky like a symlink going on
18:25 rofl____ joined #salt
18:25 sroegner joined #salt
18:25 auser joined #salt
18:26 bhosmer joined #salt
18:26 tty-1 joined #salt
18:28 z3uS joined #salt
18:29 rofl____ should i worry if not all my minions reply on test.ping on every run?
18:29 rofl____ https://gist.github.com/anonymous/5069d07c2cbbf8f0b2f5
18:30 forrest seems kind of odd that they wouldn't
18:30 rockey rofl____: network glitches, unresponsive minion and a lot more can accure
18:30 rofl____ they are all locally
18:30 rofl____ on a kvm box
18:30 rofl____ with 0.17.1
18:30 forrest oh, hmm
18:31 rockey hm, thats wierd
18:31 pears you should dig into why that is
18:31 forrest can you check on your master in /etc/salt/master what the value of max_open_files is?
18:31 jcsp1 joined #salt
18:31 rofl____ ex cmd.run always replies
18:31 rofl____ forrest: its the default 100000
18:32 rofl____ why would test.ping be more glitchy than cmd.run?
18:32 rofl____ that sounds just strange to me
18:32 forrest hmm, shouldn't be an issue then.
18:32 jhulten joined #salt
18:32 forrest yea it is strange
18:33 rofl____ yeah my guess is that its ok
18:33 rofl____ just really strange
18:33 rofl____ i tried to debug a minion
18:33 rofl____ seems it got the ping request
18:33 rofl____ it just doesnt get the reply or send a reply
18:34 forrest is salt.intra also a minion of itself?
18:34 rofl____ yes
18:34 rockey you should time the reply and match with the masters configuration
18:34 rockey if saltmaster have an open connection for the respons
18:35 rofl____ how do i time it?
18:35 rockey logs?
18:35 rockey strace process?
18:35 rofl____ yeah
18:35 rofl____ theres no builtins or verbose triggers for salt trigger
18:35 rofl____ ?
18:35 rockey tcpdump network
18:36 rockey there is alot of different options
18:36 rofl____ yeah
18:36 rockey oh i dont know actually :P
18:37 rofl____ also a question for highstate's
18:37 rofl____ is there any good way to force a push
18:37 rofl____ for all minions
18:37 rofl____ other than salt '*' state.highstate ?
18:38 forrest that's the standard, is there something that isn't doing you want to accomplish?
18:38 forrest you can also do salt '*' state.sls state_name
18:38 rofl____ can i ex setup the minions to do pulls instead?
18:38 rofl____ instead of hackish cron solutions
18:39 rofl____ on some interval and ex some splay time
18:40 forrest rofl____ http://docs.saltstack.com/topics/jobs/schedule.html
18:41 rofl____ forrest: excellent
18:41 rofl____ thanks
18:41 forrest np
18:41 rofl____ just alot of docs to read up on
18:42 rofl____ didnt find this one
18:42 forrest yes I know
18:42 forrest sometimes I forget where stuff is too
18:43 MTecknology Isn't it possible to only send something to a number of minions at a time? Like file.cp only sending to ten minions at a shot
18:43 forrest Yea batch size http://docs.saltstack.com/topics/targeting/batch.html
18:44 rofl____ wow
18:44 rofl____ nice
18:44 MTecknology ya.. that one! thanks
18:44 forrest np
18:45 MTecknology I shall commence ro-Friday (last day of the week for me)
18:45 forrest I wish it was Friday
18:45 MTecknology http://docs.saltstack.com/genindex.html <-- HOLY HELL this got big!
18:45 forrest yea
18:46 MTecknology http://docs.saltstack.com/contents.html .... WOW!
18:46 btorch hello . is this somewhat possible on a pillar .sls file ? param = {% salt[.. ip on eth1 of another node that has role X ..] %}
18:47 MTecknology btorch: I don't quite get what you're after; but what you want is definitely possible with ext_pillar
18:49 NotreDev joined #salt
18:50 btorch MTecknology: well I'm trying to have a config opton on this pillar file be updated with the IP on eth1 of another minion node. So node X is already deployed and has role X ... now I'm deploying node Y with role Y and I want a config on node Y to be updated with the IP found of node X
18:50 NotreDev where can i find some states that have already been built (recipes)
18:50 ctdawe joined #salt
18:50 forrest NotreDev, https://github.com/saltstack-formulas
18:50 NotreDev thanks!
18:50 dave_den btorch: you may want to use the mine
18:51 Teknix joined #salt
18:51 forrest np
18:52 whitepaws so now that i have written some states, how do i make the minions "do" the state?
18:52 forrest whitepaws, I'd say review http://docs.saltstack.com/topics/tutorials/starting_states.html
18:53 forrest if you just want the quick and dirty answer, if you have things set up properly, you can just do salt 'name of minion' state.sls state_name
18:53 MTecknology btorch: are you looking for http://docs.saltstack.com/topics/mine/index.html ?
18:54 whitepaws forrest,thanks again! i checked it and this worked flawlessly: salt '*' state.highstate
18:54 btorch ah cool thanks
18:56 forrest whitepaws, yea keep in mind state.highstate applies ALL states
18:56 forrest so it's not great for testing single states.
18:56 mapu joined #salt
18:57 jcsp1 joined #salt
18:57 fragamus joined #salt
18:59 mwillhite joined #salt
18:59 ausmarton joined #salt
19:00 Kareem joined #salt
19:01 alunduil joined #salt
19:01 khodgson joined #salt
19:02 smccarthy joined #salt
19:05 whitepaws forrest, thanks. kinda looks like http://docs.saltstack.com/ref/modules/all/salt.modules.state.html would be a good source of things to try to learn more about how states "happen" to the minions
19:07 jeffstoner when using salt-key, the options "-l all"  "--list=all"   "--list all" return nothing. Only "--list-all" actually lists keys.
19:07 MTecknology Hah! I'm making a web ui for some basic stat stuff; I ended up calling it saltstat. I thought the name was awesome.
19:08 jalbretsen So this kind of error is making me scratch my head
19:08 PoLuX joined #salt
19:08 jalbretsen The conflicting ID is "yum-pkgs" and is found in SLS "base:core.yum" and SLS "base:core.yum"
19:12 jalbretsen looks like for this server, I have yum-pkgs define, and I include the file with it in top.sls, but then I also do in a formula file so I can reference it in a "require"
19:13 btorch not sure I'm doing this right to get this mine thing working ... http://goo.gl/cvEnrQ
19:13 btorch I've added mine_functions to the end of the minion file and restarted the minion service
19:13 UtahDave joined #salt
19:14 jalbretsen I bet UtahDave can save me
19:14 hazzadous joined #salt
19:15 UtahDave lol, what's up , jalbretsen?
19:16 jalbretsen So on a server, I am getting this on a state.highstate:  The conflicting ID is "yum-pkgs" and is found in SLS "base:core.yum" and SLS "base:core.yum"
19:17 crazysim joined #salt
19:17 btorch dave_den: is that wrong ?
19:17 jalbretsen Looking closer, I have the formula file that defines it in base.core.yum.  That is included in top.sls for all servers to use
19:18 NotreDev i'm trying to build a pillar for my users. i only want some users to be included for certain grains. how can i include 'users.notredev', and export it under `users`?
19:19 jalbretsen I have specific server however that has a specific formula written where I include base: core.yum so I can reference "yum-pkgs" in a require statement
19:19 jalbretsen worked dandy in 0.16, now I get that error in 0.17
19:20 jalbretsen I'm wondering if with your changes, since I have that in my top.sls file for the server, that in automagically knows about it now
19:20 jalbretsen so now I'm including it twice and it's getting the error
19:22 Corey ROTHGAR!
19:23 dave_den btorch: yes, that's correct. have you restarted your minion?
19:23 rgbkrk joined #salt
19:23 btorch dave_den: yep
19:24 ipmb joined #salt
19:25 dave_den btorch: can you manually send data to the mine?
19:25 dave_den e.g.: salt 'yourminion' mine.send network.ip_addrs '[eth0]'
19:25 jalbretsen UtahDave.... and my theory appears to be right.  I just took out the include, and the error went away
19:25 btorch dave_den: I did salt '*' mine.send network.interfaces eth0 and came back true
19:26 dave_den btorch: and mine.get shows the proper data?
19:26 btorch opps let me change to ip_addrs
19:26 [1]VertigoRay joined #salt
19:27 btorch dave_den: yes now it does after the push
19:27 UtahDave jalbretsen: you should be allowed to include it more than once. I think this may be fixed in develop.  can you test on the develop branch?
19:27 btorch dave_den: but it returns the ips of both interfaces and not only eth0
19:28 dave_den btorch: ok, clear the mine data and restart the minion again. it should send mine functions at startup and at mine_interval. check your logs for any errors.
19:28 btorch dave_den: ok doing a flush and restarting it .. the logs haven't shown any errors when I was testing before
19:28 jalbretsen UtahDave:  I kinda like the idea actually...... if I include something in top and need to reference later in a specific sls file.... it's already there
19:29 bhosmer joined #salt
19:30 sroegner joined #salt
19:30 jalbretsen UtahDave, I don't have a setup like that unfortunately.  Just EPEL rpms
19:30 bhosmer joined #salt
19:30 jalbretsen I know I know, I need to get with it
19:31 UtahDave lol, ok. well, if you think this behavior is wrong, please open an issue.  I need to go back to bed. I'm in Hong Kong for the openstack conference and it is 3:30 am.
19:32 prooty joined #salt
19:33 btorch dave_den: there must be something wrong on my config ... cause it doesn't work . I even set the interval to 15
19:33 btorch also have it running in debug
19:36 prooty hello. how can i be sure that the opts parameter to the hg.latest state is being passed? they don't seem to be included in the hg command when looking at the minion in debug mode. using salt 0.17.0.
19:36 khodgson joined #salt
19:37 GradysGhost joined #salt
19:37 GradysGhost hola
19:38 GradysGhost So, I've been at this for a few days here, and am still hitting a really strange problem. state.show_highstate notes that a file is managed and must be created. state.highstate results in no action taken or reported.
19:38 foxx joined #salt
19:38 GradysGhost master/minion are both 0.16.4
19:39 * GradysGhost runs updates
19:40 Thiggy joined #salt
19:42 viraptor_ joined #salt
19:43 jcsp2 joined #salt
19:45 NotreDev i've got a pillar file that compiles fine, but i can't access the pillar information in my state file -> http://pastebin.com/eMJq5nQD
19:45 NotreDev well, the attributes of `user` aren't accessible
19:45 giantlock joined #salt
19:45 GradysGhost Ok, I updated, I'm running master/minion 0.17.1, I fixed the code bug so the master will start, and I'm experiencing the same problem. Has anyone here seen this happen before?
19:46 btorch ok not quite sure what else to do to get this mine working. only works if I first do the mine.send .. seems to ignore the section on the config I guess
19:47 dave_den btorch: what version are you running?
19:47 btorch 0.17.1
19:47 dave_den from an rpm or deb install?
19:48 btorch from a deb using the saltstack repo
19:48 ckao joined #salt
19:48 gildegoma joined #salt
19:49 khodgson joined #salt
19:49 khodgson joined #salt
19:50 dave_den btorch: does it work if you put the mine_functions config in the minion's pillar?
19:50 Kareem joined #salt
19:50 dave_den i don't use the mine myself, so i haven't tested it working from the minion config
19:51 pears what's mine for?  I can't figure it out by googling
19:51 pears like what is a potential use case
19:51 dave_den pears: http://docs.saltstack.com/topics/mine/index.html
19:52 dave_den it just stores the output from a module and function on the master and all other minions can see that data
19:52 mgw Are there any isc dhcp/bind experts here that could share some thoughts on integration with salt?
19:53 Kareem joined #salt
19:53 mgw I've developed the omapi and ddns modules, and they're useful to me, but I think I'm missing some best practices in the overall setup of my network.
19:53 pears I see, so if you wanted to do something on a minion depending on something on another minion that is not know a priori
19:53 pears s/know/known/
19:53 btorch dave_den: nope
19:56 dave_den btorch: does salt 'yourminion' mine.update work?
19:56 pdayton joined #salt
19:58 btorch dave_den: ok this is weird .. now it works after I did that update
19:58 btorch even if I restart the minion it still keeps on working
19:59 dave_den have you flushed the mine data and restarted the minion to see if it still works after a flush?
19:59 btorch no
19:59 dave_den salt 'yourminion' mine.flush
19:59 btorch hehe ok it's gone now
20:00 dave_den it seems like your scheduler isn't firing the mine.update
20:00 btorch but it is
20:00 btorch 2013-11-07 19:59:47,643 [salt.utils.schedule][DEBUG   ] Running scheduled job: __mine_interval
20:00 GradysGhost I made a bit of progress re: the state non-application. The targeted minion throws this error when I run a highstate:
20:00 GradysGhost [WARNING ] TypeError encountered executing state.highstate: 'NoneType' object is not iterable. See debug log for more info.  Possibly a missing arguments issue:  ArgSpec(args=['test', 'queue'], varargs=None, keywords='kwargs', defaults=(None, False))
20:00 GradysGhost Is this a config problem?
20:00 btorch dave_den: I took it down to 1 even
20:02 dave_den wtf
20:02 dave_den btorch: do you see anything in the master log?
20:03 btorch need to enable debug ... but right now with warning enabled the last line was from 35min ago when I kept running mine.get calls
20:07 gkze joined #salt
20:10 claudep left #salt
20:10 cedwards joined #salt
20:11 JesseC joined #salt
20:13 jkleckner joined #salt
20:13 cjh973 #scala
20:13 cjh973 oops
20:15 forrest joined #salt
20:20 tyler-baker joined #salt
20:20 tyler-baker joined #salt
20:22 alekibango joined #salt
20:23 Thiggy joined #salt
20:24 Thiggy joined #salt
20:27 jeffstoner left #salt
20:33 jankowiak joined #salt
20:34 dave_den GradysGhost: that warning is probably preventing the highstate run. you'll have to find out where the error is being caused
20:35 gildegoma joined #salt
20:36 dave_den have you tried to run a highstate with debug logging?
20:37 modafinil is there a general function i can use to get a host-specific random number? ala http://docs.puppetlabs.com/references/3.stable/function.html#fqdnrand
20:43 pipps joined #salt
20:45 Cidan is there a timeframe for 0.17.2?
20:46 GradysGhost dave_den: I get stack traces, but nothing that points me in the right direction
20:46 dave_den GradysGhost: what do the stacktraces show?
20:47 GradysGhost http://pastebin.com/B7TNUUmq
20:47 petersabaini joined #salt
20:48 dave_den GradysGhost: do you have a state function that has both args 'test' and 'queue'?
20:48 srage joined #salt
20:48 GradysGhost I don't believe I have any state functions with either argument.
20:50 GradysGhost This is another piece that looks relevant. It's the "Results of YAML rendering" for the issue in question:
20:50 GradysGhost OrderedDict([('include', ['roles.rad_server.python', 'roles.mongodb_server', 'roles.nginx_server']), ('/etc/init/rad.conf', OrderedDict([('file.managed', [OrderedDict([('source', 'salt://roles/rad_server/files/upstart.rad.conf')]), OrderedDict([('mode', 644)]), OrderedDict([('user', 'root')]), OrderedDict([('group', 'root')]), OrderedDict([('template', 'jinja')])])])), ('/etc/nginx/conf.d/rad.conf', OrderedDict([('file.managed', [OrderedDict([('source'
20:50 GradysGhost , 'salt://roles/rad_server/files/nginx.conf')]), OrderedDict([('mode', 644)]), OrderedDict([('user', 'root')]), OrderedDict([('group', 'root')]), OrderedDict([('template', 'jinja')]), OrderedDict([('require', [OrderedDict([('pkg', 'nginx')])])])])]))])
20:51 jkleckner joined #salt
20:52 dave_den GradysGhost: I don't think your problem is in that state.
20:52 GradysGhost I just grepped through all my state files and didn't find any instances of the term 'test' or 'queue'.
20:52 GradysGhost That error comes up right after some stuff from my mongo role. Lemme look into the mongo thing.
20:52 dave_den GradysGhost: the error indicates you're missing a required argument for a state function somewhere
20:52 zach last zach
20:52 zach er
20:53 NotreDev joined #salt
20:55 dave_den GradysGhost: pay special attention to your require/watch args in your states. the traceback ends at evaluating a requisite
20:55 GradysGhost For the pkgrepo.managed function on an RPM-based distro, is the mirrorlist option required?
20:55 GradysGhost I have a baseurl
20:56 GradysGhost I may have just found it.
20:56 GradysGhost Lemme dig around a bit. It's not fixed yet, but I affected a change.
20:58 gkze joined #salt
20:58 dave_den it says either baseurl or mirrorlist is required in the docs, but i haven't oooked at the code
20:58 cewood joined #salt
20:59 jschairb joined #salt
21:00 mwillhite joined #salt
21:03 matanya joined #salt
21:04 KyleG1 joined #salt
21:04 pipps_ joined #salt
21:05 KyleG joined #salt
21:05 KyleG joined #salt
21:05 KyleG left #salt
21:06 GradysGhost dave_den: I think I got it. It was a combination of things.
21:06 GradysGhost I had some {% %} tags when I needed {{ }} tags
21:06 GradysGhost I had some dumb broken requires
21:07 dave_den GradysGhost: cool, glad you found it
21:07 scott_w joined #salt
21:08 matanya joined #salt
21:09 NotreDev how can i iterate over pillar data, and access items in sub dictionaries?
21:09 xmltok is anyone using pydsl? it seems like the python looping and variables would be easier to debug than jinja+yaml. i havent tried pydsl much yet though.
21:10 scott_w_ joined #salt
21:16 _ikke_ xmltok: I have used it
21:17 DanGarthwaite joined #salt
21:17 NotreDev http://blog.mattcrampton.com/post/31254835293/iterating-over-a-dict-in-a-jinja-template
21:17 NotreDev that's the answer to iterating over dictionaries in jinja2
21:17 DanGarthwaite How to access / test for an environment variable on the minion?
21:17 GradysGhost left #salt
21:18 khodgson joined #salt
21:18 teskew the release schedule for salt-cloud really needs to be synced with the release of salt-common...it breaks constantly unless using bleeding edge for both salt-master and salt-cloud
21:18 Cidan salt-cloud is going to be included into salt it self eventually iirc
21:19 teskew for instance, if you install the latest git version of salt-cloud to fix this problem: https://github.com/saltstack/salt-cloud/issues/852   it creates this problem: https://github.com/saltstack/salt-cloud/issues/890
21:20 teskew unless you have the git version of salt-common, which nobody should really be running in a real prod environment with hundreds of machines
21:23 NotreDev how do i render a state without executing it?
21:23 teskew so in reality, the "stable" version salt-cloud, really isn't stable, because it was released before salt .17.
21:23 _ikke_ NotreDev: I only know appending test=True to the command, which lets so see what salt would do
21:23 jgarr joined #salt
21:29 mwillhite joined #salt
21:31 jacksontj joined #salt
21:31 pentabular joined #salt
21:34 icarus joined #salt
21:39 jgarr I have a general directory structure question about salt (I new to salt but have used puppet & ansible)
21:39 josephholsten joined #salt
21:40 jgarr puppet has modules -> manifests -> resources What is Salts structure? State tree -> ??
21:43 NotreDev why is this yaml not compiling properly? http://pastebin.com/prrzsHt9
21:43 NotreDev the list of groups should be a list (at least that's how pyyaml is parsing it), not a string
21:45 pears what did you expect to happen by enclosing it in quotes?
21:46 pears remove the double quotes around the square brackets and your problem should go away
21:47 pears oh sorry that's from the output, ignore me
21:47 NotreDev pears: yeah
21:48 viraptor_ joined #salt
21:49 pears I would guess something is str()-ing the list somewhere
21:52 pears NotreDev: are you doing that from the master?
21:52 NotreDev ?
21:52 NotreDev salt-call --local?
21:52 bemehow joined #salt
21:53 pears yeah
21:53 pears I only have salt-call on my minions so that probably answers that
21:53 NotreDev it's a masterless minion. no, i don't think that answers anything.
21:53 pears oh
21:53 NotreDev pears: http://docs.saltstack.com/topics/tutorials/quickstart.html
21:53 matanya joined #salt
21:55 canci joined #salt
21:57 ange hi
22:00 gkze joined #salt
22:02 KyleG joined #salt
22:02 KyleG joined #salt
22:08 diegows joined #salt
22:08 icarus joined #salt
22:09 scott_w joined #salt
22:20 icarus joined #salt
22:23 druonysus joined #salt
22:23 oz_akan_ joined #salt
22:25 Kholloway joined #salt
22:27 jkleckner joined #salt
22:30 Gifflen joined #salt
22:31 NotreDev wow. i'm exploring utils/jinja.py, and i'm amazed by the parse_yaml method on SerializerExtension class.
22:32 NotreDev parse_load is even cooler. https://github.com/saltstack/salt/blob/develop/salt/utils/jinja.py#L331
22:33 NotreDev sorry for the sarcasm, I've just never seen anyone return something like that. the review process for accepting code should probably be improved.
22:33 forrest lol
22:36 Ryan_Lane joined #salt
22:37 xmltok ive had some annoying problems with yaml parsing too, when using pillar.get it would transform my default value into True, when i really wanted 'yes'
22:38 xmltok im inclined to go with pydsl for our internal states but i need to vet it out further
22:38 gkze joined #salt
22:38 scott_w joined #salt
22:42 pears I ended up using the py renderer for almost everything
22:42 pears which hopefully won't bite me in the ass later
22:49 cedwards terminalmage: submitted a new FreeBSD specific salt bug (8336)
22:49 troyready joined #salt
22:51 xmltok i'm unable to do __pillar__.get("p1:p2:p3", []) in pydsl, does the __pillar__ here not have the : nesting?
22:53 kermit joined #salt
22:53 ajw0100_ joined #salt
22:55 blee_ joined #salt
22:55 xmltok yeah, i have to use __salt__['pillar.get']. its kind of a bummer how in some places its a regular dictionary and in other places its enhanced
22:57 jcsp1 joined #salt
22:59 terminalmage whoa, a cedwards sighting!
22:59 terminalmage :)
23:00 terminalmage cedwards: thanks, I'll look at it
23:00 jacksontj joined #salt
23:00 pdayton joined #salt
23:00 cedwards terminalmage: lol. it has been a while. school is in session again..
23:01 terminalmage have you had a chance to test any of my recent FBSD goodness in develop?
23:01 terminalmage new package management, and ports support
23:01 cedwards I've used pkg.upgrade which seems to work fine. Haven't had a need to build any ports.
23:03 pipps_ joined #salt
23:04 terminalmage yeah I'm pretty happy with the ports support: http://docs.saltstack.com/ref/states/all/salt.states.ports.html
23:04 forrest terminalmage did you grow a beard for winter?
23:05 terminalmage forrest: I'm swedish/british, so no
23:05 terminalmage best I can do is a scraggly neckbeard
23:05 forrest Ok, was gonna say if you were doing so your freebsd support would instantly be complete
23:05 forrest yea I can't grow one either
23:05 terminalmage which, I realize, makes me the quintessential programmer stereotype
23:05 forrest once a week shave go.
23:07 terminalmage yeah, it's a victory for me that one can tell I haven't shaved after 5 days or so
23:07 forrest Yea
23:07 NotreDev joined #salt
23:08 cedwards time to head for class.
23:08 cedwards terminalmage: thanks for taking a look at that bug
23:08 forrest later cedwards
23:08 forrest don't be a stranger, I can't answer the freebsd questions :(
23:09 cedwards I'm always lurking here. highlight me if you need.
23:09 forrest alright
23:10 jhulten joined #salt
23:13 rgarcia_ joined #salt
23:15 akitada joined #salt
23:15 modafinil joined #salt
23:16 xerxas joined #salt
23:16 UtahDave joined #salt
23:16 jhulten joined #salt
23:17 forrest UtahDave get outta here, you were just here 3 hours ago!
23:25 * Gareth waves
23:26 cdcalef joined #salt
23:26 mwillhite joined #salt
23:28 cdcalef hi
23:38 jhulten joined #salt
23:40 eightyeight how do i update the salt-minion to the latest version? using salt.modules.yumpkg5 remote execution, that is
23:40 eightyeight pkg.upgrade_available returns True. now, to update it
23:41 eightyeight use pkg.install?
23:47 cdcalef part of me wishes this project had a more unique name.. trying to google something with 'salt' sometimes doesn't provide the right answers. :)
23:48 Jahkeup cdcalef: yeaaaah morton is a common result
23:49 jacksontj joined #salt
23:50 drags cdcalef: I wind up using 'saltstack' instead of salt everywhere.. google seems to understand :)
23:51 cdcalef i've started doing that as well.
23:51 jslatts joined #salt
23:52 trickyher joined #salt
23:53 cdcalef still trying to fully grok the mapping between sls files, the data structure, and how those map to functions, modules, states..
23:53 cdcalef still liking it better than puppet
23:56 KyleG joined #salt
23:56 KyleG joined #salt

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