Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-02-19

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

All times shown according to UTC.

Time Nick Message
00:00 toastedpenguin joined #salt
00:04 zzzirk_ joined #salt
00:07 diegows joined #salt
00:08 UtahDave hunter: All pull requests should be against the develop branch
00:09 hunter I'll post it somewhere as a patch then - my git-fu is weak
00:10 Corey hunter: Ooh, ooh!
00:10 Corey I wrote a blog about how to do this!
00:10 Corey http://www.taos.com/2013/12/23/pull-requests-a-semi-sane-workflow/
00:10 Corey It uses Saltstack as an example!
00:11 mgw So, I'm working on a multi-minion orchestration that involves removing minions from a load balancer, upgrading, and re-adding. At first I thought I would use a chain reaction to do that… but I think it would be easier with a runner. How are others doing this?
00:13 bhosmer joined #salt
00:14 lude joined #salt
00:15 hunter Corey: ugh - let me fight the solaris boxen to a standstill and then I'll do that - I need to hack on the zfs/zpool module anyway.
00:15 hunter I wonder what the odds on a 0.17.6 are
00:16 whiteinge mgw: have you thought about using events and the reactor for that?
00:17 lude Does anyone know how to tell salt to use systemd on distro it doesn't normally use? eg, Gentoo?
00:17 mgw yes
00:17 fllr joined #salt
00:17 mgw whiteinge: that's what I meant by a chain reaction
00:17 whiteinge ah
00:17 mgw But I'm thinking about using events and a runner
00:17 mgw a runner can wait on events
00:18 mgw After working with borgstrom's pyobjects I'm loath to have to put too much more logic in yaml
00:19 whiteinge +1 to that
00:19 whiteinge pyobjects has a really slick and simple syntax
00:19 whiteinge i'm looking forward to playing with it
00:19 tyler-baker joined #salt
00:19 mgw Now, if there were a way to work with the overstate and reactor system with pyobjects...
00:20 akl joined #salt
00:21 whiteinge mgw: ooc, have you tried changing the renderer with a shebang line in a reactor sls?
00:21 mgw no, but i know it wouldn't work
00:21 whiteinge oh?
00:22 mgw I've dug pretty deep into the pyobjects code
00:22 fllr joined #salt
00:22 mgw thinking how that could be made to work
00:23 UtahDave lude: you need to set the "provider" option to do that.  Let me find a doc for you
00:24 UtahDave http://docs.saltstack.com/ref/configuration/minion.html?highlight=provider#providers
00:24 mgw whiteinge: do you know where the expected output of run() from a renderer rendering a reactor sls would be?
00:25 Gifflen joined #salt
00:25 cachedout joined #salt
00:26 whiteinge mgw: indirectly. you can see the rendered yaml if you run the master in the foreground with debug logging. the jinja has been processed at that point so the result of unserializing that yaml is the datastructure you'd need to build
00:27 mgw ok, thanks
00:27 whiteinge i have one nearby. sec...
00:28 jtang1 joined #salt
00:33 whiteinge not as nearby as i thought
00:33 whiteinge {'touch_a_file': {'cmd.cmd.run': [{'tgt': '*'}, {'arg': ['touch /tmp/sothisworked']}]}}
00:33 whiteinge mgw: ^^
00:35 tempspace joined #salt
00:35 zzzirk joined #salt
00:36 mgw whiteinge: thanks
00:36 frasergr_ joined #salt
00:37 ccase joined #salt
00:37 bemehow joined #salt
00:38 whiteinge if you make any headway on that runner idea, i'd be interested in the result
00:38 whiteinge i don't think a multi-minion chain-reaction with custom events would be too complicated but a runner would be easier to reason about since it would be central and proceedural
00:40 mgw whiteinge: yeah, i think i'll go that route
00:40 mgw unfortunately it'll be pretty single purpose
00:40 mgw but if I can think of how to abstract some of it out, i will
00:41 whiteinge even still it'd serve as a useful example
00:41 mgw at the same time, i've piqued my own interest in #!pyobject rendering for reactor etc
00:41 whiteinge i'm always surprised that custom runner modules aren't more common
00:41 ndrei joined #salt
00:41 fragamus joined #salt
00:41 mgw whiteinge: same here
00:43 sroegner joined #salt
00:48 hunter not even sure what they are - any intro docs around?
00:50 hunter EPEL now has 0.17.5 -- if anyone else was waiting for it.
00:54 mua joined #salt
00:56 thayne joined #salt
01:01 hunter joined #salt
01:02 sroegner joined #salt
01:07 glen_ joined #salt
01:08 ndrei joined #salt
01:10 krow joined #salt
01:15 joehoyle joined #salt
01:20 joehoyle joined #salt
01:21 jergerber joined #salt
01:22 jtang1 joined #salt
01:24 liwen joined #salt
01:30 forrest joined #salt
01:32 Hipikat joined #salt
01:34 joehoyle joined #salt
01:36 jankowiak joined #salt
01:41 BenCoinanke joined #salt
01:41 schimmy joined #salt
01:42 FL1SK joined #salt
01:42 jtang1 joined #salt
01:43 bhosmer joined #salt
01:44 oz_akan_ joined #salt
01:47 joehoyle joined #salt
01:50 oz_akan_ joined #salt
01:50 diegows joined #salt
01:59 alienresidents joined #salt
02:03 schimmy joined #salt
02:05 dcolish joined #salt
02:05 hardwire in multi master mode do all masters get event.fire_master?
02:12 joehoyle joined #salt
02:18 ravibhure joined #salt
02:20 HeadAIX joined #salt
02:20 ravibhure left #salt
02:27 yidclare joined #salt
02:27 jeremyBass joined #salt
02:27 jeremyBass left #salt
02:34 calvinhp_mac joined #salt
02:36 scoates ooh. found a legitimate bug in 0.17.5's mysql module.
02:36 jtang1 joined #salt
02:44 y0j joined #salt
02:50 vejdmn joined #salt
02:55 xl1 joined #salt
02:57 glen_ joined #salt
02:59 rojem joined #salt
03:04 higgs001 joined #salt
03:04 scoates already fixed in develop, heh
03:07 bhosmer joined #salt
03:09 Nazca joined #salt
03:10 simonmcc joined #salt
03:11 sroegner joined #salt
03:22 zzzirk joined #salt
03:22 rostam joined #salt
03:25 TamCore joined #salt
03:26 LUO joined #salt
03:27 Nazca__ joined #salt
03:31 jtang1 joined #salt
03:36 LUO hi ,how to set mysql connection args in command,such as ip and port , in command "salt '*' mysql.db_exists 'test'"
03:38 jcockhren http://docs.saltstack.com/ref/modules/all/salt.modules.mysql.html#module-salt.modules.mysql
03:38 jcockhren LUO: ^
03:38 krow joined #salt
03:39 jcockhren you can also pass the connection args at the CLI
03:39 LUO salt.modules.mysql.db_exists(name, **connection_args)
03:39 LUO I try,but fail
03:40 jcockhren for example,
03:40 dave_den LUO: the docs don't tell you, but you can do:   salt 'yourminion' mysql.db_exists 'test' connection_user=root connection_pass=yourpass
03:40 jcockhren salt '*' mysql.db_exists 'test' user=root
03:40 jcockhren dave_den: beat me too it
03:40 jcockhren :)
03:41 LUO I want to set port ,ip ,user
03:42 LUO would you give a full format
03:42 LUO not find in doc
03:42 dave_den LUO: the list of connection options is here: https://github.com/saltstack/salt/blob/develop/salt/modules/mysql.py#L233
03:45 scooby2 joined #salt
03:45 psyl0n joined #salt
03:47 LUO not find, dbc = _connect(**connection_args)?
03:48 dave_den LUO: dbc is the database connection handle initiated with the connection_ars
03:49 dave_den are you getting an error?
03:49 LUO _connarg('connection_port', 'port') ?
03:51 dave_den conncetion_port=3306
03:51 dave_den connection_port=3306
03:52 LUO salt '*' mysql.db_exists 'test2' connection_host=127.0.0.1 connection_user=db_monitor connection_pass=XXX  connection_port=6301
03:52 LUO it seems ok
03:53 LUO thank you :)
03:53 dave_den LUO: you're welcome. also, it might just be easier to use a mysql.default_file or set the options in the minion config as described in http://docs.saltstack.com/ref/modules/all/salt.modules.mysql.html
03:55 LUO but if I run several mysql instances in one machine,one config option is not enought
03:55 dave_den luo: you're correct.
03:55 LUO ^_^
03:55 LUO lunch time ,see you
04:04 hunter joined #salt
04:07 favadi joined #salt
04:13 xmj joined #salt
04:13 xmj hi
04:13 glen_ joined #salt
04:13 xmj terminalmage: when's the plan to release the salt version your FreeBSD ports/packages fix will be in?
04:13 terminalmage xmj: any day
04:14 terminalmage you're welcome to checkout the 2014.1 branch and build it
04:14 xmj ...which is? :)
04:14 terminalmage which is... any day
04:14 terminalmage if I had an exact date, I would have said so
04:14 terminalmage right now we have some issues with our test system
04:15 xmj terminalmage: I'm waiting for it to be in freebsd ports/sysutils/py-salt/Makefile ${PORTVERSION} :)
04:15 terminalmage and we need to iron those out so we can get the test suite to reliably run and make sure the tests all pass
04:15 xmj I see
04:15 terminalmage xmj: well, you'll be waiting for at least a few more day
04:15 terminalmage days
04:15 terminalmage we have SCaLE this weekend
04:15 xmj the id-like "we'll release when it's done" :)
04:15 terminalmage we're hoping to get it released by then
04:15 xmj whats that?
04:16 terminalmage https://www.socallinuxexpo.org/scale12x
04:17 terminalmage the release is likely ready to go, we just want to make sure our test suite runs without issue
04:17 terminalmage and that all tests pass
04:17 xmj :)
04:19 Hipikat joined #salt
04:20 xmj terminalmage: how well does salt integrate with freenas ?
04:20 terminalmage xmj: excellent question
04:20 xmj Thank you!
04:20 terminalmage I've never used it on FreeNAS
04:20 terminalmage so I do not know
04:20 terminalmage :)
04:20 xmj ok
04:20 terminalmage should be able to install it
04:20 terminalmage since freenas can install from ports
04:25 jtang1 joined #salt
04:26 JulianGindi joined #salt
04:34 oz_akan__ joined #salt
04:36 oz_akan_ joined #salt
04:38 luanvn joined #salt
04:40 zwevans joined #salt
04:53 schimmy joined #salt
04:56 mgw joined #salt
05:01 erya1007 joined #salt
05:02 rojem joined #salt
05:07 zzzirk joined #salt
05:08 sroegner joined #salt
05:09 yomilk joined #salt
05:12 BenCoinanke_q joined #salt
05:12 cachedout joined #salt
05:16 snuffeluffegus joined #salt
05:18 zzzirk joined #salt
05:19 jtang1 joined #salt
05:23 jnials joined #salt
05:29 schimmy joined #salt
05:30 liwen joined #salt
05:33 whiteinge joined #salt
05:38 thayne joined #salt
05:41 kermit joined #salt
05:49 anuvrat joined #salt
05:58 joehoyle joined #salt
06:08 zzzirk joined #salt
06:13 aleszoulek joined #salt
06:13 jtang1 joined #salt
06:21 luo joined #salt
06:22 luo curl -L http://bootstrap.saltstack.org | sudo sh -s -- -M -N , this install command ,can we set which version to install?
06:22 bhosmer joined #salt
06:28 mr_chris joined #salt
06:33 luo I  try curl -L http://bootstrap.saltstack.org | sudo sh -s -- -M -N  git v.0.17.0
06:34 luo error: pathspec 'v.0.17.0' did not match any file(s) known to git.  * ERROR: Failed to run install_ubuntu_git_deps()!!!
06:37 gammalget joined #salt
06:38 krow joined #salt
06:38 kloplop321 joined #salt
06:57 sroegner joined #salt
07:00 joehoyle joined #salt
07:03 schimmy joined #salt
07:06 __number5__ luo, you have too many dots in your version tag, try v0.17.0 (i.e. no dot between 'v' and '0')
07:07 jtang1 joined #salt
07:13 oz_akan_ joined #salt
07:17 Shenril joined #salt
07:26 krow joined #salt
07:29 [diecast] joined #salt
07:41 luo pass!git v0.17.0 ok
07:41 luo thank _number5_
07:42 jtang1 joined #salt
07:43 xt joined #salt
07:49 ml_1 joined #salt
07:54 clearclaw joined #salt
07:55 n8n joined #salt
07:59 NV joined #salt
07:59 Katafalkas joined #salt
07:59 oz_akan_ joined #salt
08:07 slav0nic joined #salt
08:07 rvm joined #salt
08:08 Nazca joined #salt
08:08 Nazca joined #salt
08:09 cast joined #salt
08:14 LinuxBill joined #salt
08:14 LinuxBill left #salt
08:15 Kenzor joined #salt
08:15 LinuxBill joined #salt
08:16 LinuxBill left #salt
08:19 higgs001 joined #salt
08:21 LinuxBill joined #salt
08:21 LinuxBill left #salt
08:21 harobed joined #salt
08:23 LinuxBill joined #salt
08:23 LinuxBill left #salt
08:24 schimmy joined #salt
08:26 unicoletti joined #salt
08:26 linjan joined #salt
08:27 zooz joined #salt
08:28 giantlock_ joined #salt
08:30 HeadAIX joined #salt
08:35 aleszoulek joined #salt
08:40 bigmstone joined #salt
08:40 btorch joined #salt
08:41 luminous joined #salt
08:41 miguitas joined #salt
08:41 Brew_away joined #salt
08:41 keekz joined #salt
08:41 ahale joined #salt
08:41 emilisto joined #salt
08:42 jrdx joined #salt
08:44 txmoose joined #salt
08:46 sroegner joined #salt
08:47 yomilk joined #salt
08:56 mike25ro joined #salt
09:00 oz_akan_ joined #salt
09:03 BogdanR Hello
09:03 BogdanR I configured the minions to be pooling every hour through pillar
09:03 BogdanR How do I see a log of what actually worked or failed?
09:04 magnusj_ Hi there. I'm about to do some deployment with linux/windows base images (with cloudstack and salt). Trying to figure out how to handle the hostname/computername setting. Anyone here who has experience with this and can share their method? Thinking about eventually populating grains with hostname from cloudstack useradata and let saltstack set it. Don't know if there's a better method.
09:14 carlos joined #salt
09:20 MrTango joined #salt
09:24 jtang1 joined #salt
09:27 raj_ joined #salt
09:27 raj_ salt
09:28 assd joined #salt
09:29 assd how can be salt configured to install virtual box
09:31 middleman_ joined #salt
09:32 rvm left #salt
09:42 Hipikat left #salt
09:43 millz0r joined #salt
09:44 giantlock_ joined #salt
09:50 mike25ro hey guys ... i need to use a feature that is available since New in version 2014.1.0. ... i have salt 0.17.1 ... current version is 0.17.5 ... what is that 2014 version all about?
09:59 ml_1 joined #salt
10:01 oz_akan_ joined #salt
10:03 Damoun joined #salt
10:09 APLU joined #salt
10:10 hardwire joined #salt
10:18 xmj mike25ro: I'm sure you'll find the difference between 0.17.5 and 2014.1.0 the same way you've found the new stuff in 2014.1.0
10:18 jtang1 joined #salt
10:20 mike25ro xmj i just saw that there is a new feature.. i was browsing docs and i saw there is :  salt.states.file.blockreplace .. anyway i will dig some more
10:24 higgs001 joined #salt
10:34 sroegner joined #salt
10:45 xmj mike25ro: (point being, youll find the rest of the differences in the docs too.)
10:53 n8n joined #salt
10:53 sroegner joined #salt
10:55 che-arne joined #salt
10:57 aleszoulek joined #salt
11:01 oz_akan_ joined #salt
11:05 bemehow joined #salt
11:06 bemehow_ joined #salt
11:11 gammalget joined #salt
11:12 jtang1 joined #salt
11:17 Katafalkas Hey, is there a place where you can watch videos from SaltConf 2014 yet ?
11:18 jrdx Gooooood question, I'd also be interested in those... especially the Best Practices presentation....
11:41 krak3n` joined #salt
11:44 elfixit joined #salt
11:58 higgs001 joined #salt
12:02 oz_akan_ joined #salt
12:05 sroegner joined #salt
12:05 zzzirk joined #salt
12:06 jtang1 joined #salt
12:24 bhosmer joined #salt
12:27 Psyke joined #salt
12:33 mike25ro hi guys... need some help with pillar / jinja .... http://pastebin.com/944SGAuM
12:34 mike25ro it works :) i am just stupid :D
12:35 ravibhure1 joined #salt
12:35 gmtech Hi all.. Don't know how to publish this, but I've written an external pillar which supports files, and encrypted data. Although its a gem, I will rewrite it in python at some point - you can find it here https://github.com/gtmtechltd/salthiera
12:36 ravibhure1 joined #salt
12:37 gmtech In the meantime, does anyone know if you can watch every file in a directory - e.g. watch: /some/log/directory/*.log
12:38 viq gmtech: watch a recursive state maybe? I think you can watch only stuff managed by salt
12:38 mike25ro gmtech: good question ... i am curious myself..
12:38 Shenril joined #salt
12:39 gmtech So in my case - I am managing a bind server in salt, I have hundreds of bind files and hundreds of zones, and i want to run an rndc reload every time one of these files changes (because of salt)
12:40 gmtech I figured it would be easier to watch /var/named/db.* , rather than put a watch_in: service: bind  on every single zone file i am writing
12:41 ravibhure joined #salt
12:41 psyl0n joined #salt
12:41 viq gmtech: maybe another idea: have the zones in one sls file, and I think you could do watch: - sls: zones
12:41 gmtech cool - thanks for the tip
12:42 viq I'm not sure it will work, but it's worth a try
12:44 mikkn I asked this question a couple of days ago, but I forgot to check for answers which was really silly of me... but here goes: Are pillars read in at the beginning of a state run or could you have a state that triggers a new pillar to be generated and then read it in the next state?
12:45 faust joined #salt
12:46 diegows joined #salt
12:46 viq mikkn: both, I think ;)
12:46 krak3n` joined #salt
12:46 gmtech The iptables module REALLLLLLLLLLLLLLY sucks
12:46 mikkn viq: Hah
12:47 gmtech Anyone know how to solve this? I'm doing iptables.append:, proto: udp, sport: 53 -- but when it runs the command line its putting sport BEFORE the proto which is invalid iptables command
12:47 gmtech How do I force the order?
12:47 viq mikkn: I think you could trigger a pillar refresh with http://docs.saltstack.com/ref/states/all/salt.states.saltmod.html
12:49 che-arne joined #salt
12:52 v101 joined #salt
12:52 v101 left #salt
12:53 luminous joined #salt
12:56 scuwolf joined #salt
12:59 pdayton joined #salt
13:00 jtang1 joined #salt
13:01 ravibhure left #salt
13:10 joehoyle joined #salt
13:15 elfixit joined #salt
13:15 balboah joined #salt
13:20 mgw joined #salt
13:22 mpanetta joined #salt
13:27 dan_ joined #salt
13:29 jaimed joined #salt
13:30 Guest9454 Hi everyone, I've come to ask a quick question. I am trying to deploy minion servers from a template. I think as I have already installed salt-minion they all have the same public key (as per generated in the initial install). Is there a way to regenerate the  key?
13:33 jrdx joined #salt
13:35 Ahlee Guest9454: rm /etc/salt/pki/minion/minion*; service salt-minion restart
13:35 jslatts joined #salt
13:35 ipmb joined #salt
13:35 joehoyle joined #salt
13:39 sroegner joined #salt
13:43 Guest9454 that works thanks!
13:44 Guest9454 however it still uses the hostname of the template and not the new hostname
13:44 andersb Anybody here know anything about Salt Cloud not installing on Ubuntu 12.04?
13:44 Ahlee Guest9454: rm /etc/salt/minion_id
13:45 zwevans joined #salt
13:48 Guest9454 fantastic Ahlee, thanks so much!!!!
13:51 Ahlee Guest9454: you're very welcome
13:54 brianhicks joined #salt
13:54 jtang1 joined #salt
13:57 calvinhp_mac joined #salt
14:00 cachedout joined #salt
14:04 zzzirk joined #salt
14:05 oz_akan_ joined #salt
14:06 calvinhp_mac joined #salt
14:10 juicer2 joined #salt
14:10 Gifflen joined #salt
14:13 BenCoinanke joined #salt
14:15 racooper joined #salt
14:16 juicer2 joined #salt
14:17 calvinhp_mac left #salt
14:22 mgw joined #salt
14:25 justBob__ joined #salt
14:25 vejdmn joined #salt
14:25 bemehow joined #salt
14:26 faldridge joined #salt
14:26 bemehow_ joined #salt
14:26 Gordonz joined #salt
14:34 krak3n` joined #salt
14:34 psyl0n joined #salt
14:36 fragamus joined #salt
14:39 oz_akan_ joined #salt
14:39 bryano joined #salt
14:41 bhosmer joined #salt
14:44 krow joined #salt
14:46 HeadAIX joined #salt
14:49 jtang1 joined #salt
15:06 psyl0n joined #salt
15:06 higgs001 joined #salt
15:08 vejdmn joined #salt
15:09 colinbits joined #salt
15:10 krow1 joined #salt
15:12 tyler-baker joined #salt
15:16 jergerber joined #salt
15:19 mgw joined #salt
15:21 rgbkrk joined #salt
15:21 quickdry21 joined #salt
15:22 Katafalkas joined #salt
15:22 kaptk2 joined #salt
15:23 rostam joined #salt
15:23 alunduil joined #salt
15:27 elsmorian1 joined #salt
15:29 elsmorian1 I have a state which calls the djangomod.collectstatic module- is there a way I can run this command as a certain user?
15:30 xmltok joined #salt
15:32 scoates elsmorian1: looks like there's no provision for user switching for that, if I'm reading it correctly: https://github.com/saltstack/salt/blob/develop/salt/modules/djangomod.py#L173
15:32 mgw joined #salt
15:32 elsmorian1 scoates:  :( Thats a shame
15:33 abe_music joined #salt
15:34 kolaman joined #salt
15:36 MTecknology I don't know who changed it so that salt actually sits and waits forever and ever for responses from minions. Whoever did it: I love you.
15:37 scoates MTecknology: did you set your timeout to 0? (-:
15:38 yidclare joined #salt
15:39 MTecknology scoates: actually... I just noticed that my timeout is 30
15:40 scoates MTecknology: hmm. one minion or a glob of many?
15:40 MTecknology glob of many; each minion takes 1-4hr to run state.highstate
15:41 linjan joined #salt
15:42 MTecknology I used to have a long hang and then nothing ever returned. It seems like there's a heartbeat in place now.
15:43 jtang1 joined #salt
15:49 scoates MTecknology: interesting. Heartbeat will be helpful to me, too.
15:49 scoates 4h highstate sounds like a nightmare (-:
15:49 krow joined #salt
15:50 jaimed joined #salt
15:50 linjan joined #salt
15:52 gadams999 joined #salt
15:54 yidclare joined #salt
15:59 MZAWeb joined #salt
15:59 vejdmn joined #salt
16:00 d1rk joined #salt
16:02 GradysGhost joined #salt
16:03 joehoyle joined #salt
16:04 forrest joined #salt
16:04 sifusam_ joined #salt
16:06 mua joined #salt
16:06 pass_by_value joined #salt
16:08 sifusam_ left #salt
16:08 sifusam joined #salt
16:10 cachedout joined #salt
16:11 davet joined #salt
16:11 anuvrat joined #salt
16:11 rojem joined #salt
16:14 troyready joined #salt
16:14 psyl0n joined #salt
16:16 Gordonz joined #salt
16:16 jalbretsen joined #salt
16:16 danielbachhuber joined #salt
16:18 Teknix joined #salt
16:24 bhosmer_ joined #salt
16:24 frasergraham joined #salt
16:25 Gifflen_ joined #salt
16:26 Gifflen_ joined #salt
16:27 Gifflen_ joined #salt
16:30 dancat joined #salt
16:31 joehoyle joined #salt
16:32 vejdmn joined #salt
16:33 mgw joined #salt
16:37 jtang1 joined #salt
16:40 jeremyfelt joined #salt
16:49 krow joined #salt
16:50 linjan_ joined #salt
16:52 thayne joined #salt
16:57 psyl0n joined #salt
16:59 hunter joined #salt
17:01 vejdmn1 joined #salt
17:01 joehoyle joined #salt
17:05 res0nat0r joined #salt
17:05 bemehow joined #salt
17:06 JordanRinke Morning
17:06 fragamus joined #salt
17:07 bemehow_ joined #salt
17:10 dave_den hoy hoy
17:10 KyleG joined #salt
17:10 KyleG joined #salt
17:12 mua_ joined #salt
17:14 cpenner461 joined #salt
17:15 matt__p joined #salt
17:16 meteorfox joined #salt
17:18 bchristianv joined #salt
17:19 JulianGindi joined #salt
17:19 bchristianv Question: Can I use '#' to insert comments into my state tree files?
17:20 R_Shackleford joined #salt
17:20 schimmy joined #salt
17:20 jdenning joined #salt
17:21 yidclare joined #salt
17:21 johtso joined #salt
17:22 JulianGindi joined #salt
17:22 johtso What's the correct way to specify a salt state dictating that a supervisord process group should be running, and should be restarted when something happens?
17:22 dancat joined #salt
17:22 dave_den bchristianv: yes
17:22 johtso "supervisord.restart process:" works fine from the command line
17:23 dave_den johtso: you need to use watch or watch_in
17:23 dave_den johtso: http://docs.saltstack.com/ref/states/requisites.html#watch
17:23 johtso but if I set the "name" of the supervisord.running state to be "process:"  I get  Function:  running   Result:    False   Comment:   Starting service: processor:
17:23 budrose joined #salt
17:24 bchristianv Thanks. Had not seen an example in all my searching and wasn't sure.
17:24 bchristianv left #salt
17:24 johtso dave_den, I have a watch, and I think it's triggering correctly
17:24 johtso but the process isn't being restarted
17:25 dave_den johtso: is 'process' the actual name of your service's configuration file in supervisord?
17:25 rgbkrk Has anyone bootstrapped salt on Windows?
17:26 rgbkrk Possibly with winrm or the like?
17:26 rgbkrk I see that the salt cloud stuff doesn't have any direct options
17:26 R_Shackleford anyone here have any experience with using the iptables.insert state in an sls, or just the iptables insert module using the command line in general?
17:27 johtso dave_den, the state looks like this: https://gist.github.com/johtso/57979127d57e22d8d81f
17:27 rgbkrk *salt minion on Windows
17:27 MZAWeb joined #salt
17:28 johtso dave_den, the .conf looks like this: https://gist.github.com/johtso/57979127d57e22d8d81f#file-gistfile2-txt
17:28 manfred R_Shackleford: nerd
17:28 rojem joined #salt
17:28 R_Shackleford whats interesting is that the example given in the documentation doesn't even work
17:29 R_Shackleford I keep getting a "Error: Position needs to be specified or use append (-A)" error
17:29 dave_den johtso: 'process:' should probably be 'process'
17:29 jeremyBass joined #salt
17:30 dave_den R_Shackleford: what distro?
17:30 R_Shackleford CentOS 6.4
17:30 R_Shackleford the problem with append is that its going to put my rule after the REJECT all rule
17:31 R_Shackleford otherwise, i wouldn't mind just settling on append
17:31 jtang1 joined #salt
17:31 pdayton joined #salt
17:32 dave_den R_Shackleford: the iptables module used to have issues on centos. i haven't used it myself. what does your state look like?
17:32 schimmy joined #salt
17:32 manfred i have had issues with the same thing on ubuntu 12.04 and 13.10
17:32 johtso dave_den, that gives me "process: ERROR (no such process)"
17:33 R_Shackleford https://gist.github.com/mparker1001/9097020 <--- the state in my sls file
17:33 dave_den johtso: where is your supervisor config file located?
17:35 johtso dave_den, https://gist.github.com/johtso/57979127d57e22d8d81f#file-gistfile3-yml
17:35 dave_den R_Shackleford: are you using salt from git?
17:35 joehillen joined #salt
17:36 Doormatt joined #salt
17:36 R_Shackleford R_Shackleford: no from epel. v 0.17.4
17:36 manfred gross, 0.17.5 is where it is at
17:37 R_Shackleford manfred: that upgrade from .4 to .5 doesnt show anything in the changelog related to iptables specifically
17:37 mua joined #salt
17:37 dave_den R_Shackleford: strange, iptables.insert shouldn't even be working in 0.17.4 - it's marked as new in 2014.1.0 http://docs.saltstack.com/ref/states/all/salt.states.iptables.html#salt.states.iptables.insert
17:38 bhosmer joined #salt
17:38 manfred lol
17:38 manfred R_Shackleford: ♥
17:38 R_Shackleford well then that may explain it
17:38 manfred it certainly would
17:39 R_Shackleford i think ill just disable my iptables in the mean time and set my access controls on the application level
17:40 gnugnu5 joined #salt
17:41 dave_den johtso: what does 'salt-call supervisord.status' show?
17:43 bt joined #salt
17:45 johtso dave_den, no such process
17:45 johtso dave_den, I don't think status works with group syntax
17:45 amckinley joined #salt
17:46 johtso dave_den, I've actually just taken a look at the state.py and there seem to be group related changes that aren't in the current release
17:46 dancat joined #salt
17:46 bemehow joined #salt
17:47 dave_den johtso: supervisord.status should show all of your services that supervisord is managing. you may need to specify the path to your main supervisord.conf file
17:47 dave_den e.g.  salt-call supervisord.status conf_file=/etc/supervisord/supervisorod.conf
17:48 dave_den http://docs.saltstack.com/ref/modules/all/salt.modules.supervisord.html#salt.modules.supervisord.status
17:48 johtso dave_den, oh sorry, I meant when explicitly specifying the process. supervisord.status with no arguments returns http://hastebin.com/mivigebelu.vhdl
17:49 johtso *process name
17:49 johtso as expected
17:49 krow joined #salt
17:50 dave_den hrm, not sure what's up, then.
17:50 johtso dave_den, https://github.com/saltstack/salt/commit/40cfa293d0f9d8a5b1abd4b79bf774703d380b00
17:50 johtso looks like I want that
17:51 dave_den ah, yeah
17:52 dave_den dang, authored 6 months ago
17:52 dave_den but only made it into 2014.1
17:53 jankowiak joined #salt
17:53 johtso feeling like I might need to start using more bleeding edge salt..
17:56 rojem joined #salt
17:59 teskew joined #salt
18:00 teskew does anyone have a formula/set of states for mesos?
18:00 xmltok joined #salt
18:00 micah_chatt joined #salt
18:02 scoates johtso: fwiw, you can provide your own modules in roots/_modules to override problems in salt-proper. I do this for a couple things (most recently: mysql_grants fix, yesterday)
18:07 johtso scoates, yeah, with thinking I could pull in the latest supervisord module and state.. assuming it doesn't depend on other stuff though
18:08 * scoates nods
18:14 yidclare joined #salt
18:15 dwiden joined #salt
18:15 fragamus joined #salt
18:19 timoguin joined #salt
18:24 ml_1 joined #salt
18:25 jtang1 joined #salt
18:26 sibyfi joined #salt
18:26 sibyfi is anyone using Salt with NetApp storage?
18:29 bdf we have both salt and netapp in our enviornment, so I suppose that probably qualifies
18:30 kula i have salt. someone else runs a netapp. i have nfs space. but salt and netapp don't really do anything with each other.
18:30 bdf have I attempted to do any kind of integration of the two besides maybe mounting nfs? not really
18:30 DanGarthwaite joined #salt
18:31 Ryan_Lane joined #salt
18:31 schimmy joined #salt
18:31 cachedout bdf: Could be a good integration point for the new Salt proxy minion.
18:32 bdf but next time im looking for a way to stew my storage admins, maybe some proxy-minion will be in our future
18:32 rojem joined #salt
18:32 krow joined #salt
18:33 sibyfi am researching wether salt would be a good fit for provisioning netapp controllers for a lab environment.  ~2k-3k Netapp controllers reset every other day or so
18:34 bdf real ones or the demo filer VM? (if that still exists)
18:35 sibyfi both, the vsims would reside within vmware so salt would need to talk to vmware server but we most have physical controllers.
18:36 ajw0100 joined #salt
18:37 rmnuvg joined #salt
18:41 guest_towel joined #salt
18:42 DanGarthwaite joined #salt
18:42 higgs001_ joined #salt
18:43 DanGarthwaite joined #salt
18:44 hunter joined #salt
18:45 oz_akan_ joined #salt
18:47 bt joined #salt
18:51 mgw Any ideas why salt would be reporting this (https://gist.github.com/mgwilliams/bb635018e6ede8849308) when the package is both installed and running?
18:53 sibyfi left #salt
18:54 KyleG mgw: postgresql-8.4=8.4.11-*.
18:54 KyleG Is probably an invalid package name
18:54 KyleG maybe try just postgresql-8.4.11
18:54 KyleG or even just postgresql
18:54 mgw It's a valid package+version
18:54 KyleG Right I understand that
18:54 mgw And actually installed it
18:54 KyleG But it's probably doing some weird regex
18:54 dave_den the =8.4.11 is just an artifact of salt checking the version
18:54 KyleG against the list of packages
18:55 KyleG looking for "postgresql-8.4=8.4.11-*"
18:55 KyleG rather than "postgresql-8.4.11"
18:55 dave_den mgw: there was a bug with installing packages that contained versions in their names not too long ago, iirc
18:55 mgw There's no package such as postgrsql-8.4.11
18:55 mgw the package is postgresql-8.4
18:55 mgw and the version is 8.4.11-1
18:55 KyleG well if you want to know exactly why its' doing it
18:55 KyleG Set your log level to debug
18:55 KyleG It should show you the exact package command it's running to look @ the values
18:56 KyleG and the output, and why it errored or not.
18:56 meteorfox joined #salt
18:57 dave_den mgw: https://github.com/saltstack/salt/pull/8921
18:57 mgw ah, it's the comparison that's failing
18:57 mgw even though it is valid for the install
18:58 mgw dpkg --compare-versions '8.4.11-*' lt '8.4.11-1
18:58 mgw errors out
18:58 meteorfox joined #salt
18:58 KyleG that's what I said!
18:58 KyleG lol
18:58 KyleG :)
18:58 mgw KyleG: Sorry, I thought you meant it was invalid for the apt-get install
18:59 mgw for apt-get install it's fine
18:59 mgw so I wonder how I can make this work
19:00 mgw I need exactly 8.4.11, but don't care if -1 changes to -2 in the future
19:00 KyleG Does apt-get care if you add-on the -*
19:00 mgw apt-get won't install the package without it
19:00 zain joined #salt
19:01 mgw I guess I'll just spell out the -1
19:02 mgw and change it in the future if needed
19:02 dave_den if the package name is postgresql-8.4 and pkg.installed can't handle that, it's a bug and needs to be fixed...
19:07 danielbachhuber joined #salt
19:09 mgw_ joined #salt
19:09 Ryan_Lane is there any way for external pillars to handle environments?
19:10 CaptTofu joined #salt
19:11 Ryan_Lane also, if I'd like environments to map to a location, is that doable? something like this? http://pastebin.com/626TZkRs
19:12 Ryan_Lane err. let me amend that
19:12 mgw dave_den: it's the version that's the issue
19:12 mgw the globbing
19:12 Ryan_Lane would I be able to do this with an external top?
19:12 mgw which is valid for apt-get, but not dpkg —compare-versions
19:12 bhosmer joined #salt
19:19 zain joined #salt
19:19 jtang1 joined #salt
19:22 hunter joined #salt
19:23 GradysGhost joined #salt
19:24 jtang2 joined #salt
19:31 hunter_ joined #salt
19:32 toastedpenguin joined #salt
19:33 dwiden_ joined #salt
19:34 MZAWeb joined #salt
19:38 ml_1 joined #salt
19:43 joehoyle joined #salt
19:43 MedicalJaneParis joined #salt
19:44 dancat joined #salt
19:44 joehoyle joined #salt
19:46 hunter_ joined #salt
19:46 Ryan_Lane joined #salt
19:59 MZAWeb joined #salt
20:00 timoguin_ joined #salt
20:01 Mr_N joined #salt
20:02 Mr_N Hey guys; rather stupid question since I'm sure I read something on a builtin function to do this, but I'm being absolutely fail at re-finding any reference to it;
20:02 Mr_N Is there any way to check for existence of a given state/module
20:03 Mr_N I'm trying to do some stuff where new minions automatically try to apply a state identified by their host grain if such a state exists
20:03 Mr_N (I may also be approaching solving this problem in the wrong sort of mindset, in which case I'd appreciate a pointer as to the proper way to approach it ><)
20:04 Mr_N Any advice would be appreciated :)
20:06 n8n joined #salt
20:06 zooz joined #salt
20:06 jcockhren Mr_N: depends on how you're writing your state
20:06 jcockhren (or rather which tool)
20:06 druonysus joined #salt
20:06 Mr_N jcockhren: The relevent yaml is inside my top.sls
20:06 jcockhren but first outline what you're trying to achieve, the bif pictuer
20:06 jcockhren s/bif/bit
20:07 Mr_N For any new minion; when I highstate, I want salt to first check if a state exists with the matching hostname to the minion, and if so, apply it.
20:07 Mr_N (is the big picture)
20:07 Mr_N This is my attempt to in system deployment, if needed, just drop an appropriately named state file in an appropriate dir, and essentially automate per host CMS membership.
20:08 DanGarthwaite left #salt
20:08 jcockhren ok.
20:08 thayne joined #salt
20:09 kieran_ joined #salt
20:09 Mr_N_ Hm
20:09 hunter_ joined #salt
20:09 dimeshake joined #salt
20:09 Mr_N_ Sorry about that; just got my ssh session killed for no good reason
20:09 jcockhren another way to achieve that is to have the  top file to match base on a grain
20:09 jcockhren for the a specific grain have it spply certain states
20:10 jcockhren for a new minion, a grain can be applied dynamically or through it's config file
20:10 psyl0n joined #salt
20:10 jcockhren Mr_N_: get all that?
20:10 Mr_N_ I do.
20:11 Mr_N_ I'm not sure I'm completely content with that sort of model
20:11 Mr_N_ but that's probably my not seeing all the constraints.
20:11 jcockhren yeah.
20:11 jcockhren give me a larger view
20:11 jcockhren I
20:11 jcockhren I'll try to help . ;)
20:12 Mr_N_ I liked the idea of tweaking a per-node state on the salt master, rather than tweaking the grains on a client to change what states get pushed to that particular host
20:12 jcockhren ok.
20:12 jcockhren a couple quick notes:
20:13 jcockhren 1. grains are designed as 'grains of truth'
20:13 jcockhren 2. the top file has a matching system then says which states are to be sent to the minions
20:14 Mr_N_ 1. Indeed, 2. I'm aware, that's the functionality I've been using thus far, I had been trying to do it based on hostname, and that's exactly where the "detect if a state exists" question came in.
20:14 jcockhren are we talking about custom states?
20:14 jcockhren (and custom modules?)
20:14 Mr_N_ Since rather than just cluttering my top file with "if hostname == x, do y, if hstname == z, do z2", I figured it would juts say "if a state exists with the hostname specified, use that one"
20:14 Mr_N_ Yes.
20:14 Mr_N_ Sorry if I was not clear
20:14 jcockhren even better
20:15 jcockhren if you write your states in python or pydsl
20:15 Mr_N_ I can have them not bomb out
20:15 Mr_N_ I guess I could add an "abstraction state"
20:15 Mr_N_ where in my top file I call a "apply host state" state on all hosts
20:15 Mr_N_ which is a pure python state which checks for existence of the "true" host state
20:16 Mr_N_ That doesn't sound too horrible, but I guess it surprises me that there's not a builtin for checking which states are available.
20:16 Mr_N_ (thanks by the way, even if I don't realize any better solutions, this is certainly better than what I had been working with.)
20:17 Mr_N_ Or rather; like, an -optional directive.
20:18 jcockhren Mr_N_: there's requisite system as well
20:18 Mr_N_ Indeed, but it's an all or nothing.
20:18 Mr_N_ when you do an -include or a -require it gets really sad if that state doesn't exist
20:18 Mr_N_ at least AFAIK
20:19 hunter_ Mr_N_ another way is to use pillar so that the master specifies truth rather than asking the client (grains)
20:19 zooz joined #salt
20:19 jcockhren hunter++
20:19 Mr_N_ hunter: Indeed; I'm using it for some components.  The "big big" picture for all of this, is that I'm trying to make my kickstart system tie into Salt in a really unobtrusive fasion
20:20 Mr_N_ it would be _great_ if a kickstart could just drop a <machinename>.sls file onto the salt-master and have "everything work out"
20:20 Mr_N_ without having to do line modifications on a piller file
20:20 krow joined #salt
20:20 Mr_N_ But I guess that's another opportunity to use the pure python templating
20:20 hunter Mr_N_: using grains (pulling info from client) seems to be more common in salt than pillar (for this purpose -pillar/grains have many purposes/uses) but its definately a case that both WORK and pick based on site preference.
20:20 Mr_N_ Do it in the pillar to automatically look for host-information pillars, and apply them accordingly...
20:21 hunter joined #salt
20:21 Mr_N_ Yeah; this has been one of the problems I've had picking up salt.  There are _tons_ of ways to solve similar problems.
20:21 jcockhren it's a "feature"
20:21 jcockhren lol
20:21 Mr_N_ And I'm nervous about "building bad habits", because I watched that all go to hell in my use of nagios :P
20:21 hunter Mr_N_: I'm in a similar boat and was thinking I'd have the master create the kickstart templates based on pillar/grain data.
20:21 hunter Mr_N_: its interesting to hear about someone thinking of having kickstart add to salt instead.
20:22 Mr_N_ Indeed; I find your approach the other direction to what's intuitive to me
20:22 Mr_N_ I think it's probably due to the mindset of the systems each of us are coming from
20:22 Mr_N_ whereas in mine, kickstarts are the "source of truth"
20:23 Mr_N_ Your idea is certainly interesting, I'd wonder what you'd actually do the kickstart creation/deployment with; I assume a python-coded state?
20:23 chbaker joined #salt
20:23 faldridge joined #salt
20:23 chbaker Anyone awak in here?
20:24 chbaker s/awak/awake/
20:25 Mr_N_ Finally.
20:25 vbabiy joined #salt
20:27 hunter Mr_N: haven't gotten that far. For me kickstart is just another mechanism to implement truth.
20:28 Mr_N Right now we use kickstart to deploy all our machines, so I'm somewhat inclined to bootstrap off that
20:28 chbaker What's the trick to executing a single state?
20:28 sulky joined #salt
20:28 chbaker Like "salt 'someserver' state.sls MyState"
20:29 Mr_N I'm really really tempted to make a feature request for -include_if_present:
20:29 chbaker when you have separate environments like dev, qa, prod
20:29 darless salt 'someserver' state.sls MyState env='dev'
20:29 darless chbaker: -^
20:30 chbaker darless: thanks. I had tried saltenv='dev' and it still used base
20:32 tyler-baker joined #salt
20:33 chbaker so does that env='dev' tell salt that the baseurl is /srv/salt/dev ?
20:33 jxqz joined #salt
20:34 jxqz ~locate redbeard
20:34 chbaker I got back that it cat find for example MyStatedev/MyState/MyState.conf
20:34 * jxqz searches in vain
20:34 chbaker Even though I have /srv/salt/dev/MyState/MyState.conf
20:35 chbaker oops, s/cat/can't/
20:36 UtahDave joined #salt
20:37 UtahDave left #salt
20:37 krow joined #salt
20:38 Mr_N Hmm, even in a pure python renderer I don't see any immediately visible way to check existence of a given state
20:40 jcockhren Mr_N: one sec. let ,e show you an example
20:40 Mr_N sure.
20:42 jcockhren https://github.com/saltstack/salt/blob/develop/salt/states/mysql_database.py#L21
20:42 dave_den chbaker: you need to have a file_roots path defined in your master config for env 'dev'
20:42 jcockhren the basically, every state has what is called a __virtualname__ so that's its registered for executioni
20:42 jcockhren check if it's there
20:42 Mr_N See, I had been trying to do something similar
20:43 Mr_N if <statename> in salt:
20:43 Mr_N (but that in the default template)
20:43 Mr_N But that doesn't seem to work for my custom states.
20:44 Mr_N Could be that I'm misunderstanding what makes up the virtualname
20:46 jcockhren a virtualname is simply a name that used interally by the salt system to find the module to execute code to do stuff
20:46 jcockhren it's a function
20:46 dave_den Mr_N and jcockhren: I think there is a misunderstanding of nomenclature
20:46 jcockhren totally possible I'm missing what he's trying to do
20:46 diegows joined #salt
20:46 jcockhren (switching context and ish)
20:46 Mr_N Hm.  Let me try to explain better
20:47 dave_den when Mr_N is referring to states, he is referring to the states that comprise the highstate data
20:47 dave_den not state modules
20:47 Mr_N No, I was referring to state modules
20:47 dave_den oh, my mistake then :)
20:47 Mr_N I am using the highstate
20:47 Mr_N So:
20:47 Mr_N All I'm doing is coming up with a braindead way to associate arbitrary hosts and state files, by hostname
20:48 Mr_N I want a system such that when I bring a minion into salt; it applies any state that matches its hostname.
20:48 dave_den that's easy
20:48 Mr_N (if such a state exists)
20:48 Mr_N Well, I have it working
20:48 Mr_N but it bombs out if it doesn't exist (naturally)
20:48 Mr_N and I'm not sure how to check.
20:49 Mr_N It's a grand total of 5 lines; if there's no rules against small code dumps I can throw it here or in pastebin, it might be more informative as to where I have a gap in my knowledge ><
20:49 dave_den ah, so in your top file you are able to apply a state file to a minion, but it's possible that state file is not present in the master's file_roots
20:49 Mr_N Yes.
20:49 Mr_N Precisely
20:50 dave_den what do you want it to do if there is no .sls file that matches the minion hostname?
20:50 Mr_N For the purpose of being completely clear; here is my currently (broken_ stanza
20:50 Mr_N http://pastebin.com/E9SESaSF
20:50 Mr_N I want it to go on with life.
20:51 Mr_N I broke down my top.sls file into a few "Sections."  The '*' rule that pushes common to all hosts, a few rules that push states based on subnet, rules that push states based on role grains, and rules that push based on hostname.
20:51 Mr_N (since those are really the categories we "select" on when making decisions deploying systems)
20:51 dave_den ok
20:52 dave_den 1 - you can't do the my_hostname in salt
20:52 Mr_N I mean; that component works, FWIW.
20:53 Mr_N if I remove the if, it behaves exactly as expected.
20:53 Mr_N (succeeds by calling the proper state for hosts that have the state existing under file_roots, bombs out on hosts that don't have it.)
20:53 dave_den well, the 'salt' dict in templates is a special dict where each key is a callable to the state_mdule.function_name
20:54 Mr_N Oh; sorry
20:54 dave_den e.g.  salt['cmd.run']('echo foo')
20:54 Mr_N I completely misparsed
20:54 Mr_N what you said.
20:54 dave_den so when you do if my_hostname in salt you are saying
20:54 Mr_N When you said my_hostname in salt I brainfarted and thought you meant I can't use the set my_hostname as I did anywhere in salt :P
20:54 dave_den if 'my.minion.fqdn.net' in salt, which will never be true
20:55 Mr_N IIRC grains[host] isn't FQDN?
20:55 Mr_N e.g. if I have "sensor6.pha.jhu.edu" it returns "sensor6"
20:55 dave_den ah, yes my bad
20:56 Mr_N And I have a state, sensor6.sls in my file_root
20:56 Mr_N and that all works yadda yadda
20:56 dave_den ok, but the same reasoning applies. the salt dict holds 'state_module.function'
20:56 dave_den so file.managed, cmd.run, etc.
20:56 Mr_N So I guess the question is; is there any way to referr to the sensor6.sls state in that fasion?
20:56 dave_den you're referring to state files
20:56 Mr_N Since I think I assumed it would just be in "the top level" or under like "states.<name>" or something
20:57 dave_den the salt dict refers to state modules and functions.
20:57 dave_den they are different
20:57 dave_den so
20:57 dave_den to do what you want
20:57 dave_den you can use the 'cp.list_master' function to get a list of files in the master's file_roots.
20:58 Mr_N Oh; that's useful.  Thank you very much for all your time, by the way.
20:58 Mr_N Saving my ass from making a terribly hacky workaround :)
20:58 dave_den and you can use logic in your template to see if grains['host'] in salt['cp.list_master'] …
20:58 n8n joined #salt
20:58 Mr_N I'd probably have to call salt['cp.list_master']() ?
20:59 dave_den and i have actually been saying the salt dict hold 'state_module.function', but it really holds 'execution_module.function'
20:59 dave_den Mr_N: yes, correct
20:59 jstadler joined #salt
21:01 Gifflen joined #salt
21:01 Mr_N Woo!  it will work once I beat it with a hammer a little bit.
21:01 cewood joined #salt
21:01 dave_den cool
21:01 Mr_N since it's returning the right stuff when I run it as a standalone, but isn't applying the state in the template
21:01 Mr_N I can bang that out though; thanks again for all the help :)
21:02 dave_den you're welcome
21:02 Gifflen joined #salt
21:02 xmj question re 2014.1.0
21:02 xmj is that a change in that salt's version numbering will change to a new format?
21:03 xmj or... will salt get bumped to 0.18.xx?
21:04 toastedpenguin joined #salt
21:05 meteorfox hi, executing salt-call event.fire_master '...omitted' '...omitted' is printing DeprecationWarning: This minion received a job where kwargs were passed as string'd args, what is the new way to this?
21:05 akoumjian joined #salt
21:06 dave_den xmj: 2014.1.0 is the new format.
21:06 xmj cedwards: don't forget to bump PORTEPOCH :)
21:07 Saint-N so does salt have a specific order that it runs commands? cuz it doesnt seem to be top to bottom
21:07 Saint-N ie install package then run a command then modify files then restart
21:07 xmj you can make it do that
21:07 Saint-N how so?
21:07 dave_den Saint-N: http://docs.saltstack.com/ref/states/ordering.html#state-auto-ordering
21:07 Saint-N ooo
21:07 Saint-N thanks
21:08 xmj and require's
21:08 Saint-N ahh thats what i was forgetting
21:08 Saint-N thanks
21:10 meteorfox I'm following the same format as in http://docs.saltstack.com/topics/reactor/index.html?highlight=events#fire-an-event, but I'm getting the deprecation warning.
21:10 krow joined #salt
21:11 synfin joined #salt
21:12 meteorfox I'm running salt 2014.1.0-1784-ga3d4d72
21:12 Saint-N xmj and dave_den  yeah fergot the "pkg" part of the requrie
21:13 Saint-N klnew it was something dumb heh
21:13 * xmj installs mod_speling on Saint-N
21:13 krow1 joined #salt
21:13 Saint-N tahnkz! :)
21:13 ffggh joined #salt
21:13 MrTango joined #salt
21:13 Ryan_Lane joined #salt
21:15 diegows joined #salt
21:15 schimmy joined #salt
21:18 cjbarnes18 joined #salt
21:19 ahwitz joined #salt
21:20 cjbarnes18 eveing all,  been playing about with halite, got one machine working with the doc search feature, but cant seem to replicate it on a second.  Am I missing something?
21:20 ahwitz quick question (hopefully) - a variable is set on my minion in the bash_profile and can be accessed when I type in the command line there. echo $VAR_NAME and env do not include it even when run as a user in a cmd.run salt state call. any suggestions on how to get it to show up?
21:21 rgbkrk joined #salt
21:21 dave_den meteorfox: here's the commit comment that is causing that message: https://github.com/saltstack/salt/commit/01f2d119b1ac42fa780fbbd243fddf8923a7fa6b
21:21 GradysGhost ahwitz: Probably has something to do with the way the shell gets set up. Likely salt is running sh instead of bash or something like that keeping your shell from being set up the way you expect.
21:22 meteorfox dave_den: thanks, looking...
21:22 dave_den meteorfox: it doesn't look like salt-call has been patched to be consistent with that change.
21:22 GradysGhost I wouldn't count on env vars, though. Maybe a grain or pillar is more appropriate? What's the goal?
21:22 ahwitz GradysGhost: I had a similar problem using "source" instead of "." that might be related. I'll check into it and get back in a sec.
21:23 dave_den ahwitz: when you start salt-minion in init script or upstart, it is not started in a login shell, so bash_profile is not sourced.
21:24 pass_by_value Hi cjbarnes18
21:24 dave_den ahwitz: i am actually rewriting the new environ modules that will help with this:  https://github.com/saltstack/salt/blob/develop/salt/modules/environ.py
21:24 cjbarnes18 Hi pass_by_value
21:24 Gifflen_ joined #salt
21:24 pass_by_value does the checkbox not activate on another machine?
21:25 yidclare joined #salt
21:25 pass_by_value for search docs?
21:25 ahwitz dave_den and GradysGhost: I got it working by simply sourcing the bash profile manually for the single cmd.run - I'm trying to automate installation of Cappuccino, which requires a bash variable set to point to its build directory for a particular step.
21:26 cjbarnes18 pass_by_value: no I just get the no enrty mouse pointer when I mouse over.
21:26 dave_den ahwitz: that is a common issue, and the environ module and state modules will be able to do that.
21:26 ahwitz dave_den: to value stability over functionality, I'm gonna stay off the bleeding edge releases (still in 0.17 something I think?) but good to hear it'll be out soon
21:27 dave_den ahwitz: but they will not likely be in stable for a while.
21:27 ahwitz thanks a ton for your help, guys!
21:27 ahwitz roger!
21:27 dave_den heh, yes exactly
21:27 pass_by_value If you click "Job" do you see sys.doc in green?
21:28 cwright joined #salt
21:28 cjbarnes18 pass_by_value: it's not listed, but runner.manage.present is listed with red light.
21:29 Gifflen joined #salt
21:29 pass_by_value hmm If that failed then Halite won't "boot up". Does a page refresh help?
21:30 Saint-N so here's another im stupid question
21:30 Saint-N if i have primary init.sls in a state.. and i include blerg.sls itll run through init first and when thats done itll run blerg yeah?
21:31 Saint-N sort of a 2 staged thing?
21:31 Ryan_Lane joined #salt
21:32 cpenner461 joined #salt
21:33 Gifflen joined #salt
21:36 cjbarnes18 pass_by_value: the machine that has doc search working aso has failed runner.manage.present
21:36 dancat joined #salt
21:37 pass_by_value I would try "salt-run manage.present" on the command line
21:37 justBob_ joined #salt
21:37 pass_by_value That way we will know if there is some issue with that runner function actually working cjbarnes18
21:37 cjbarnes18 Function 'manage.present' is unavailable
21:38 forrest Saint-N, when you include a state in another state, the included state should run first.
21:38 pass_by_value what version of salt master and minions
21:38 pass_by_value salt —versions-report
21:38 pass_by_value and salt \* grains.get saltversion
21:39 cjbarnes18 0.17.5
21:39 timoguin Saint-N, foo/init.sls can be applied independently of foo/bar.sls, and vice versa, if you're not actually including them in each other
21:40 timoguin otherwise what forrest said applies
21:41 cjbarnes18 pass_by_value: the one that works is  0.17.4
21:41 pass_by_value I am gonna check to see if there is some issue after checking out 0.17.5 cjbarnes18
21:43 robawt joined #salt
21:43 ckao joined #salt
21:44 pass_by_value cjbarnes18: I get that error on 0.17.5 as well
21:44 kermit joined #salt
21:44 Saint-N thanks
21:45 Saint-N im coing from ansible so im trying to get my head around this whole seemingly orderless thing ;)
21:45 Saint-N coming
21:45 cjbarnes18 pass_by_value: is manage manage.present runner in the core?
21:46 pass_by_value afaik it should be part of the standard salt core, yes
21:46 ffggh Hi guys, stumbled upon this channel. What is salt?
21:47 cjbarnes18 pass_by_value: accoring to readthedocs its not in 0.17.4 but is in latest
21:48 cjbarnes18 pass_by_value: I susspect this is not the couse of the docs issue however
21:48 pass_by_value I am on develop so I always get to see it
21:49 pass_by_value It is not a direct cause, you are right. But Halite needs to know the minions that are present before it can make the sys.doc call
21:49 Saint-N ffggh,  a infrstructure managment tool  or an essential chemical compound your choice :)
21:50 cheus joined #salt
21:50 pass_by_value so if manage.present fails there are no calls made for sys.doc and any subsequent functions
21:51 cjbarnes18 pass_by_value: minions are listed
21:51 ffggh Saint-N: Thanks. Thought it might be for the Salt cryptocurrency which is currently being developed.
21:51 pass_by_value Listed via test.ping or something?
21:51 jason75 joined #salt
21:52 cheus Howdy. Does anyone have a best-practice recommendation for setting up recurring events in a multi-master (ha) scenario? Cron on one node isn't HA and cron on two could cause the same events to fire. I'd usually use a rabbitmq work queue for this but hoping something's baked in.
21:52 ffggh left #salt
21:53 _the3rdman_ joined #salt
21:54 QuantumRiff joined #salt
21:54 jtang1 joined #salt
21:55 bashcoder joined #salt
21:55 krow joined #salt
21:55 troyready joined #salt
21:56 QuantumRiff I hope this is a basic question, but is it possible to 'nest' jinja "IF" statements in my state files?
21:57 harobed joined #salt
21:57 QuantumRiff if OS = Centos 5 AND if subnet = B then
21:57 QuantumRiff etc
21:57 cjbarnes18 pass_by_value: clicking on briefcase icon.
21:58 harobed joined #salt
21:58 QuantumRiff trying to check if its centos 5 or 6, then, depending on domain name (or IP subnet) send to a different syslog server
21:58 cjbarnes18 pass_by_value: then minions appear in the minons tab and i can drill into grains.
21:58 MTecknology scoates: yup, there's a heartbeat type system in here now. That's absolutely effing amazing
21:58 cheus QuantumRiff, You can use a compound match in your grains
21:59 cheus QuantumRiff, I wouldn't nest something like that (or use conditional logic in my grains, loops are sometimes OK but conditionals are better handled through grains matches)
21:59 forrest joined #salt
22:00 pass_by_value yes that will make the grains call, lemme check and see if we can somehow use that info to our advantage cjbarnes18
22:00 QuantumRiff something similar to "{% if '5.' in grains['osrelease'] AND if 'domainA' in grains ['domain'] %} ?
22:01 yomilk joined #salt
22:03 cheus QuantumRiff, I'd just use it in my topfile to match, 'G@osrelease:5 and G@domain:domainA'
22:03 cheus QuantumRiff, see http://docs.saltstack.com/topics/targeting/compound.html
22:05 alexthegraham joined #salt
22:05 pass_by_value cjbarnes18 I am curious to know what target it used to get the grains when you clicked the briefcase icon. Could you please check under Event? There should be a grains.items there
22:05 pass_by_value I tried the same thing you did but since I had no target my grains job did not go through
22:07 pass_by_value Also for 0.17.5 there is no function present in runners/manage.py file
22:08 pass_by_value function named present, I should say
22:08 QuantumRiff cheus: I have several syslog related things to do, and with 3 domains, and 2 OS releases, thats 6 state files to maintain. (i guess I could use some includes) but that seems a bit "icky", doesn't it? I guess I haven't used compound matchers in my top.sls before
22:09 elfixit joined #salt
22:10 alexthegraham Anyone have a few to help me suss out whether what I'm seeing is a bug or normal operations?
22:10 cjbarnes18 pass_by_value: where to I find the target?  I see "fun: grains.get"
22:11 dave_den alexthegraham: don't ask to ask - just ask :)
22:11 cheus QuantumRiff, It all depends on your perspective. A good best practice is avoiding as much jinja logic as possible, it's very difficult to read and to get the spacing right and leads to deleterious effects. I don't know that I'd use that exact compound match because I'm not sure what the whole context of your question is, but it's cleaner than a jinja approach.
22:11 pass_by_value Under Event if you click the littlt arrow next to grains.items it should show tgt in the expanded view
22:11 pass_by_value little*
22:12 QuantumRiff cheus, good point
22:12 budrose joined #salt
22:13 alexthegraham Hah. OK. Here goes: I have a state file that throws the 'argument of type 'StrictUndefined' is not iterable' error until I install the package dmidecode. Once that's installed, no problem. I can't see any problem with my YAML and I can't figure out why having dmidecode installed would make a difference.
22:13 cheus QuantumRiff, as a rule of thumb, it's better to put logic into processing and setting intelligent grains and pillar matches than using otherwise. Remember that it's all just data, you could (the best approach), set up a dictionary with the keys of osversion and domain to do the same thing
22:14 dave_den alexthegraham: can you gist your state file?
22:14 cheus QuantumRiff, That's probably how I'd do it, using those combinations to give me a specific role or state I need to implement
22:15 QuantumRiff cheus: interesting.. any thing to lookup for seeing how to do the dictionary with keys of osversion and domain?
22:15 QuantumRiff cheus: never dealt with that before
22:16 cjbarnes18 pass_by_value: tgt: router.mc.local
22:16 alexthegraham dave_den : https://gist.github.com/alexthegraham/87be1315cafb83592d66
22:17 thayne joined #salt
22:17 pass_by_value Assuming that is your minion it seems like there is a target present for it
22:18 pass_by_value Kind of surprising that it got a target when manage.present failed
22:18 alexthegraham dave_den - also, this only fails on CentOS 6 if that helps.
22:19 QuantumRiff i'm looking at alexthegraham's gist.. looks like he's doing something similar to what I was trying
22:19 cjbarnes18 pass_by_value: am getting the feeling that I'm gonna have to wait for the next release.
22:19 cjbarnes18 pass_by_value: an indications when that might be?
22:19 pass_by_value But it seems like manage.present does not appear in the source tree anywhere before 2014.1
22:20 dave_den alexthegraham: is that the only state file you run when you get the traceback?
22:20 pass_by_value Any chance you could try out anything above (and including) release 2014.1?
22:20 pass_by_value cjbarnes18 ^^
22:21 CaptTofu joined #salt
22:21 cjbarnes18 pass_by_value: i'll spin up a new vm and give is a go, I dont suppose theres debs available?
22:22 alexthegraham dave_den : Yeah. Fresh install w/ all updates of CentOS6 + salt-minion 0.17.5.
22:22 pass_by_value Just to be clear when I said 2014.1 I really meant v2014.1. I will have to get back to you on the debs
22:23 alexthegraham pkgmgr just has an include for - .yum, which is that init.sls
22:24 dave_den alexthegraham: can you gist the debug log output when running that state file?
22:24 alexthegraham Sure. One sec.
22:27 dave_den dmidecode is used to populate grains, but you should be getting that traceback any time your minion starts or you run a salt-call. not just with one state file
22:28 yidclare joined #salt
22:29 alexthegraham dave_den : https://gist.github.com/alexthegraham/2a711b183e4ccd0c06f1
22:29 dave_den alexthegraham: also, without dmidecode installed, can you run salt-call grains.items
22:29 linjan joined #salt
22:30 alexthegraham dave_den : Yeah, that works just fine. Want the output?
22:30 dave_den yes pls
22:31 andersb Anybody here care to help me with some salt-cloud stuff? Pretty basic stuff I would believe :)
22:32 schimmy so I have a custom module that I'm including with '-m /path/to/module/', but it does not get passed to the target with salt-ssh
22:32 schimmy am I wildly overexpecting functionality, or is this a bug?
22:32 Ryan_Lane joined #salt
22:32 pass_by_value cjbarnes18: What os are you using?
22:33 schimmy I get this message before it quits:
22:33 schimmy [DEBUG   ] STDERR localhost
22:33 schimmy Function crypt_helper.get_encrypted is not available
22:33 schimmy where crypt_helper is the module in my /path/to/module directory
22:33 schimmy this works for salt-call —local
22:33 salt-new-user joined #salt
22:34 alexthegraham dave_den : https://gist.github.com/alexthegraham/43aed8e10b5ff4d8f163
22:36 dave_den alexthegraham: there is no grains['productname'] in there. do you have a custom grains module that sets 'productname'] grain?
22:36 cjbarnes18 pass_by_value: debian wheezy
22:37 alexthegraham That value does exist after installing dmidecode.
22:37 pass_by_value Installing salt using apt-get? After adding the ppa?
22:37 kermit joined #salt
22:37 cjbarnes18 pass_by_value: apt-get, I should be file with the bootstrap though.
22:38 cjbarnes18 assuming that gets v2014.1
22:38 pass_by_value bootstrap might be better, lemme get the right flag to get v2014.1
22:38 dave_den alexthegraham: [WARNING ] Although 'dmidecode' was found in path, the current user cannot execute it. Grains output might not be accurate.
22:39 dave_den productname is actually one of the grains that needs dmidecode installed
22:39 alexthegraham So that's the problem, huh? bummer.
22:39 linjan joined #salt
22:39 blee joined #salt
22:39 alexthegraham I mean, not a big deal to work around, but I've been battling that damn error for *way* too long now.
22:39 schimmy I guess, how do I send along the module to the client I'm salt-sshing into?
22:41 alexthegraham dave_den: Where are you seeing that Warning?
22:41 dave_den alexthegraham: https://github.com/saltstack/salt/blob/develop/salt/grains/core.py#L1264
22:41 pass_by_value cjbarnes18: sudo sh salt_install.sh git v2014.1.0rc3 is an example
22:41 dave_den alexthegraham: you will get that warning in your minion log if you do not have dmidecode installed
22:43 andersb Anybody with some salt-cloud skill that can help me a couple of minutes?
22:43 pass_by_value cjbarnes18: you can use that to get the version of your choice. Versions are listed here (https://github.com/saltstack/salt/tags). More info on the bootstrap installation can be found (https://github.com/saltstack/salt-bootstrap#bootstrapping-salt) - if you were not already aware of that
22:44 chrisjones joined #salt
22:45 sifusam @andersb I'm not expert, but I use salt-cloud
22:45 faldridge joined #salt
22:46 meteorfox joined #salt
22:47 andersb sifusam: Great. Well, I am having some problems with UnauthorizedOperation.
22:47 bretep Does anyone have time to look at this issue? The batching is not working. https://github.com/saltstack/salt/issues/10470
22:47 andersb sifusam: I believe it's related to my iAM/role.
22:47 bretep There wasn't enough log output for me to debug
22:48 andersb sifusam: What rights do you give the role?
22:48 bretep If not, then I'll code dive. basepi did comment on it
22:48 cjbarnes18 pass_by_value: thanks, running now
22:48 krow joined #salt
22:49 andersb sifusam: The doc only says: Configure an IAM Role from the IAM Management Console.
22:51 dave_den bretep: your log output in that issue shows that the master is batching and executing the jobs properly. have you tried just increasing your timeout value? -t 300
22:52 joehoyle joined #salt
22:52 dave_den bretep: also, have you checked your minion logs to see if they are actually executing the highstate?
22:52 bretep dave_den: yes on the second
22:52 bretep I'll try the -t 300
22:52 bretep one sec
22:53 bretep the highstate did run on them
22:53 btorch after renaming a server hostname do I need to delete anything on the minion or master so that it picks up the new name ? I laready deleted the old key
22:53 jtang1 joined #salt
22:53 FL1SK joined #salt
22:53 joehoyle joined #salt
22:54 bretep dave_den: increasing the timeout to 300 seconds worked.
22:54 sifusam andersb: can't help you there. I only use rackspace cloud
22:55 meteorfox joined #salt
22:55 bretep dave_den: What is your github username?
22:55 bretep I'm making a comment on the issue
22:55 dave_den bretep: dlanderson
22:55 dave_den bretep: your issue was different than the OP on that issue
22:55 bretep oh
22:56 dave_den his master is not able to determine which minions match the glob. your master was able to match the glob, but the highstates of the minions took longer than the default timeout value on the master, so the master did not display the output
22:56 andersb sifusam: Ahh, thx anyway.
22:57 bretep Got it
22:57 bretep dave_den: thanks!
22:58 dave_den bretep: no prob
22:59 bretep Is the /etc/salt/master timeout: 15 the only change I need to make in the master config?
23:00 bretep Looks like it
23:03 bhosmer joined #salt
23:04 Nazca__ joined #salt
23:05 rojem joined #salt
23:05 mgw1 joined #salt
23:05 ajw0100 joined #salt
23:07 Gifflen_ joined #salt
23:07 rgbkrk joined #salt
23:10 linjan joined #salt
23:10 pdayton1 joined #salt
23:11 dancat joined #salt
23:12 bhosmer joined #salt
23:14 cpenner461 joined #salt
23:14 salt-new-user how do I pipe multiple commands in powershell....using salt  cmd.run
23:15 salt-new-user is the following syntax correct:  salt '*' cmd.run 'powershell.exe get-service m* | format-table Name,Displayname,Status'
23:15 Corey salt-new-user: No idea, sorry; I don't have a Windows environment handy.
23:15 pdayton joined #salt
23:16 pdayton1 joined #salt
23:16 cjbarnes18 pass_by_value: wow, that is soo much more responsive :D
23:18 blee_ joined #salt
23:18 krow joined #salt
23:22 yidclare joined #salt
23:24 bemehow joined #salt
23:28 diegows joined #salt
23:28 pdayton joined #salt
23:29 bhosmer joined #salt
23:30 pdayton1 joined #salt
23:36 zain joined #salt
23:36 MindDrive joined #salt
23:36 n8n joined #salt
23:37 sashka_ua joined #salt
23:40 jaimed joined #salt
23:42 faef joined #salt
23:43 faef joined #salt
23:43 faef anyone know why local file client doesn't obey fileserver_backends?
23:43 joehillen joined #salt
23:43 faef is it only because pillar uses it too, and pillar could only handle normal paths?
23:44 BenCoinanke_g joined #salt
23:44 tyler-baker left #salt
23:48 zain joined #salt
23:50 n8n joined #salt
23:53 rojem joined #salt
23:56 meteorfox joined #salt
23:57 BenCoinanke joined #salt
23:57 kedo39 joined #salt

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