Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-06-30

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

All times shown according to UTC.

Time Nick Message
00:03 Nexus_x1 joined #salt
00:11 zz_cro joined #salt
00:30 tkharju1 joined #salt
00:34 otter768 joined #salt
00:35 tkharju2 joined #salt
00:38 tkharju3 joined #salt
00:38 Luke_ joined #salt
00:45 oz_akan_ joined #salt
00:47 krow joined #salt
00:50 krow1 joined #salt
00:52 krow joined #salt
00:57 Luke_ joined #salt
01:07 malinoff joined #salt
01:08 bhosmer joined #salt
01:24 ifur joined #salt
01:30 jhauser_ joined #salt
01:31 mgw joined #salt
01:46 oz_akan_ joined #salt
01:52 Luke_ joined #salt
01:52 gzcwnk joined #salt
01:58 patrek joined #salt
02:00 vejdmn joined #salt
02:04 vejdmn joined #salt
02:04 gzcwnk anyone around pls?
02:05 ml_1 joined #salt
02:15 malinoff gzcwnk, I am
02:18 kermit joined #salt
02:19 aquinas_ joined #salt
02:20 tkharju4 joined #salt
02:20 kermit joined #salt
02:24 CeBe joined #salt
02:24 CeBe1 joined #salt
02:30 gzcwnk do you know about writing top.sls files and running them?
02:30 gzcwnk I am trying to remove packages to ahrden servers
02:31 malinoff gzcwnk, top.sls is required only for highstate
02:31 gzcwnk http://pastebin.com/465u238S
02:31 gzcwnk well im trying to figure out how to do things
02:32 gzcwnk i created a top.sls poinitng at security.sls to remove 2 packages
02:32 gzcwnk at least I think that is how it works
02:35 malinoff gzcwnk, can you run a single security.sls? salt vuwunicopatch02 state.sls security
02:35 malinoff Will that run without issues?
02:36 malinoff It seems to me that you have a malformed security.sls file
02:37 malinoff It should look like that: http://pastie.org/9339045
02:37 gzcwnk trying
02:37 malinoff http://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkg.html?highlight=pkg.installed#salt.states.pkg.removed
02:37 gzcwnk ah
02:38 gzcwnk http://pastebin.com/RFDM17UU
02:38 justlooks joined #salt
02:39 chadhs joined #salt
02:39 malinoff gzcwnk, illegal indentation is not the only problem - you must specify multiple package names as 'pkgs' argument
02:40 gzcwnk doesnt like tabs
02:40 malinoff http://www.yaml.org/faq.html
02:40 gzcwnk so how is the files spaced? with a spacebar?
02:40 malinoff 2 spacebars
02:41 malinoff Check my paste: http://pastie.org/9339045
02:41 justlooks hi, i have a problem ,  in salt init.sls i use include to include 2 two sls file , like tomcat.package ,tomcat.newinstance ,  but when i execute command like "salt 'alex' state.sls tomcat " just first sls file executed ,the second can not execute, why?
02:42 gzcwnk it still isnt happy
02:42 gzcwnk http://pastebin.com/F4V3xh9y
02:42 malinoff justlooks, I'm not sure, but afaik including an sls does not mean it will be executed. You must refer to it in requisites
02:43 malinoff gzcwnk, missing space between dash and pkgs
02:43 gzcwnk wow picky.
02:44 gzcwnk I dont recall the docs saying its pedantic  lol
02:44 gzcwnk still failing
02:44 gzcwnk http://pastebin.com/PFid4z2y
02:44 malinoff Yes, but it is still the easiest thing for writing in comparison with json/xml
02:45 malinoff missing colon after pkg.removed
02:45 gzcwnk thanks
02:46 gzcwnk ok that worked
02:46 gzcwnk :D
02:46 malinoff You're welcome
02:47 oz_akan_ joined #salt
02:47 gzcwnk thanks, now Im beginning to get the picture
02:47 gzcwnk so I should be able to call security.sls from teh top.sls?
02:48 malinoff yes, definitely
02:48 malinoff just try
02:48 gzcwnk i'll just re-install https
02:48 vbabiy joined #salt
02:48 malinoff you can specify test=True argument (not sure if it works with highstate), which means no actual actions will be executed, but salt will check if those actions CAN be executed
02:49 malinoff specify in the cli
02:49 malinoff like salt \* test.ping test=True
02:50 gzcwnk hmmm, failed,  http://pastebin.com/kKgFfyp5
02:51 gzcwnk think i missed a colon
02:52 gzcwnk yep it was teh colon
02:53 funzo joined #salt
02:56 bhosmer joined #salt
02:56 acabrera joined #salt
02:57 justlooks malinoff: but if i have no relation between first file op and second file op , all op of the two file will execute?
02:58 malinoff justlooks, sorry, can't understand you
02:58 ajolo joined #salt
03:01 justlooks malinoff: i means the operation of second file have no requisites with first file ,like i do transform a config file in first sls , and copy dir in second sls ,both can do first
03:02 malinoff justlooks, I guess, it' just a salt limitation. You can try if that works by putting a simple require statement
03:09 ajolo_ joined #salt
03:21 tkharju joined #salt
03:24 justlooks i tested it https://gist.github.com/justlooks/4e8949bffeb41f7c4edd
03:25 malinoff so it works
03:25 malinoff maybe you have syntax errors, or something like that in your state files
03:25 malinoff can you try to execute them separately?
03:26 justlooks malinoff: here is my real work https://gist.github.com/justlooks/e35e1251b73c992fdd9f
03:27 justlooks malinoff: i tested it seperate use state.sls
03:28 malinoff sooo many repetitions...
03:29 malinoff And if you separately call state.sls it works, right?
03:29 justlooks malinoff: yes
03:30 kermit joined #salt
03:31 malinoff justlooks, I can't imagine how that may work, since you're referring to pkg: tomcat in newinstance.sls, but that state is in another file. In my mind, to use something in a requisite, you must include a state file contains that 'something'
03:31 malinoff Or, in your case, include: tomcat.newpackage
03:31 justlooks malinoff: before i add this require statement it's works
03:32 malinoff justlooks, if you want to use pkg: tomcat in require, you must include package.sls in newinstance.sls
03:34 justlooks malinoff: if i include package.sls in newinstance.sls ,in init.sls i just need include newinstance.sls ,right?
03:34 malinoff yep
03:34 ramishra joined #salt
03:35 mgw joined #salt
03:36 bhosmer joined #salt
03:43 sag47 left #salt
03:47 oz_akan_ joined #salt
03:47 happytux_ joined #salt
03:54 funzo joined #salt
03:59 jacksontj joined #salt
04:07 mgw joined #salt
04:07 thayne joined #salt
04:08 yomilk joined #salt
04:10 nyov joehh: I'd like that. much appreciated, thanks
04:11 acabrera joined #salt
04:26 TyrfingMjolnir joined #salt
04:45 bhosmer joined #salt
04:48 joehh nyov: which release do you use?
04:48 oz_akan_ joined #salt
04:53 ramteid joined #salt
04:55 funzo joined #salt
05:04 nyov joehh: the one which works? sorry, debian or salt?
05:08 nyov ah well, I mix things. so, stable release + the sid salt pkg, dependencies lowest possible
05:21 schimmy joined #salt
05:21 jeddi joined #salt
05:29 gildegoma joined #salt
05:33 acabrera joined #salt
05:38 justlooks joined #salt
05:39 justlooks why  it is still old sls file on minion cache ,when i update it on master ?
05:41 linjan joined #salt
05:42 felskrone joined #salt
05:48 joehh nyov: I meant debian - I take it you are running sid (or atleast using the sid package)
05:49 oz_akan_ joined #salt
05:54 ndrei joined #salt
05:56 funzo joined #salt
05:56 justlooks f
05:58 * xmj looks at topic
05:59 * xmj checks salt announce
06:05 joehh xmj: no news yet - I think it is still sunday in the us
06:06 xmj ye
06:06 joehh even then, my guess is late monday or early tuesday us time...
06:07 aqua^^ joined #salt
06:07 xmj heh.
06:07 xmj i just wanna make sure we're the first. ;)
06:08 malinoff You're the first who will face a bunch of new bugs? :)
06:08 marco_en_voyage joined #salt
06:08 joehh :)
06:15 ipalreadytaken joined #salt
06:15 nyov joehh: yes, only using the sid package. (though I'l likely build me a backported package after the next release)
06:17 picker joined #salt
06:17 xmj malinoff: nah, just wanna be able to tweet that our port was first up to the new version :p
06:17 xmj (packages aren't built until wednesday, so... we lose anyway :p)
06:19 malinoff xmj, you can build packages by yourself :)
06:20 xmj yea but i can't kick our big box to put them on pkg.freebsd.org :(
06:20 malinoff Don't you use a custom repo?
06:21 xmj that's beside the point
06:33 bhosmer joined #salt
06:33 joehh xmj: do you do the packaging for freebsd?
06:35 seventy3 joined #salt
06:38 xmj joehh: nah, cedwards does
06:39 xmj i just make sure his work gets into our portstree asap :)
06:43 joehh fair enough
06:50 ipalreadytaken joined #salt
06:50 oz_akan_ joined #salt
06:57 funzo joined #salt
07:17 justlooks left #salt
07:21 ml_1 joined #salt
07:27 N-Mi joined #salt
07:41 steffo joined #salt
07:41 chiui joined #salt
07:43 slav0nic joined #salt
07:50 oz_akan_ joined #salt
07:51 xsteadfastx i have a problem with installing a pkg on a windows box over salt. for the others it works... with thunderbird it just stucks on the master. no output. anywhere i can check for logs or a more verbose mode?
07:52 darkelda joined #salt
07:52 jeddi xsteadfastx:  you can change to 'trace' or 'debug' in salt master config (have to restart the master)
07:52 jeddi i usually do that first - i guess you can bump up the salt minion on the windows box too.
07:53 xsteadfastx try the debug log level now :)
07:54 xsteadfastx mh no output at all
07:54 jeddi i have the utmost respect for people brave enough to try managing windows with anything other than a short piece of thick timber.
07:54 xsteadfastx its hell
07:55 xsteadfastx but there is no single good way to do this
07:55 xsteadfastx i tried wpkg... but on windows 7 it just dont want to work for everything
07:55 malinoff There is - to drop windows :)
07:55 xsteadfastx i wish we could
07:55 jeddi periodically i ponder setting up a win2k8 box to play with salt on.
07:56 xsteadfastx our company needs windows boxes. it sucks big time
07:56 xsteadfastx i was already happy i could roll out thunderbird insted of outlook
07:56 xsteadfastx was a big step for me ;-)
07:57 funzo joined #salt
08:22 workingcats joined #salt
08:25 Lomithrani joined #salt
08:48 londo joined #salt
08:49 _mel_ joined #salt
08:51 oz_akan_ joined #salt
08:57 TheThing joined #salt
08:59 funzo joined #salt
09:00 bhosmer joined #salt
09:06 ndrei joined #salt
09:16 vu joined #salt
09:17 ramishra_ joined #salt
09:26 thayne joined #salt
09:31 ndrei joined #salt
09:31 Lomithrani when lookup_jid on a state.highstate doesn't return anything would that mean that the state hasn't finished / is stuck ?
09:32 Lomithrani it's been a while I launched the highstate (it is normal there is some compiling involved) but I don't see salt launching any saltutil.fin_job which it usually do .
09:32 giantlock joined #salt
09:37 yomilk joined #salt
09:42 BillChung joined #salt
09:52 oz_akan_ joined #salt
09:56 xsteadfastx joined #salt
09:59 jdmf joined #salt
09:59 funzo joined #salt
09:59 TheThing joined #salt
10:02 pressureman joined #salt
10:09 bhosmer joined #salt
10:11 jeddi Lomithrani: i don't know whta the official answer is .. but if you run the same again with test=True you can quickly find out.  i thought salt-run jobs.active is meant to tell you (only recently looked that one up)
10:12 jeddi Lomithrani: or hop onto the box (or cmd.run to it) and see if there's some salty activity going on.
10:13 Lomithrani jeddi: thanks , it was indeed doing its job I just wasn't sure of it :) , now I'm struggling with my smf :( my service is supposed to be up and running and its not
10:14 Lomithrani in fact if anyone have an idea , I'll pastebin the smf
10:16 Lomithrani http://pastebin.com/TtJRvZt3   obviously typing "/opt/forge/target/universal/stage/bin/forge -Dpidfil.path=run.pid -Dconfig.file=/opt/forge/conf/prod.conf" in the CLI works perfectly
10:16 jeddi almost goes without saying.
10:18 jeddi http://pastebin.com/NLqEsy5d for people who like to use the syntax highlighting feature :)
10:22 ndrei joined #salt
10:23 jpcw joined #salt
10:24 Lomithrani yes sorry ^^
10:24 Lomithrani jeddi: anything seems wrong to you with my smf ?
10:32 beebeeep joined #salt
10:37 beebeeep hello folks, can anyone explain, what i'm doing wrong - there is my sls http://pastebin.com/4gw7yxUT, when i'm applying it to some host, i get "The state "disk_mpfs_mongos" in sls disk_mpfs_mongos is not formed as a list"
10:38 beebeeep if i remove top-level id 'disk_mpfs_mongos', everything is ok
10:39 mrchrisadams left #salt
10:44 ggoZ joined #salt
10:53 oz_akan_ joined #salt
10:54 ndrei joined #salt
10:58 ramishra_ joined #salt
11:00 funzo joined #salt
11:07 non7top_ joined #salt
11:18 bhosmer joined #salt
11:42 Lomithrani joined #salt
11:51 dyu joined #salt
11:52 dyu where should i place the sls files?
11:53 viq dyu: file_roots in master config, by default /srv/salt
11:53 oz_akan_ joined #salt
11:54 dyu got it. thanks!
11:57 bhosmer_ joined #salt
11:58 ghaering joined #salt
11:59 ghaering I'm trying to use the s3 fileserven in masterless mode, but AFAIC the minion simply ignores any s3 settings. Is there a way around that?
12:01 funzo joined #salt
12:01 ghaering minion config file: https://gist.github.com/ghaering/3f9e81de8c0d77f4e6d9
12:07 ndrei joined #salt
12:07 taterbase joined #salt
12:08 diegows joined #salt
12:10 dyu salt still can't find my sls. i've got it in /var/salt/postgres/server.sls
12:11 viq dyu: did you adjust file_roots ?
12:12 dyu yup. i'm running it like this... sudo salt pegasus postgres.server test=True
12:12 dyu file_roots is set to /srv/salt
12:13 viq well then no wonder it can't find your file in /var/salt if it's looking for it in /srv/salt
12:14 dyu sorry. meant /srv/salt
12:14 dyu /srv/salt/postgres/server.sls
12:14 viq and what error are you getting?
12:14 viq aaah, I see
12:14 viq you're doing the same thing I was doing :P
12:15 viq you're missing 'state.sls' in there
12:15 viq sudo salt pegasus state.sls postgres.server test=True
12:16 dyu nice. i'm getting new errors now which means progress. thanks!
12:16 viq :P
12:18 logix812 joined #salt
12:18 babilen heh
12:26 workingcats joined #salt
12:27 DammitJim joined #salt
12:38 mapu joined #salt
12:38 vejdmn joined #salt
12:39 vu joined #salt
12:44 dyu ack. i'm back. anyone tried making a salt state for postgres using apt.postgresql.org? salt wants me to pass in a keyserver
12:45 funzo joined #salt
12:46 babilen dyu: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkgrepo.html#salt.states.pkgrepo.managed accepts keyserver, keyid and key_url
12:50 dyu nice. finally. got it installed. thanks
12:50 dyu XD
12:52 babilen Have fun :)
12:52 oz_akan_ joined #salt
12:57 oz_akan_ joined #salt
13:01 bhosmer joined #salt
13:06 racooper joined #salt
13:07 aquinas_ joined #salt
13:11 kermit joined #salt
13:16 jaimed joined #salt
13:18 mgw joined #salt
13:23 Lomithrani hello again , http://pastebin.com/NLqEsy5d does anyone have any idea what could be wrong with that ? the command does work but salt tells me the service is running when it's not  :(
13:27 babilen Lomithrani: "the command" ? What did you paste there?
13:27 Lomithrani /opt/forge/target/universal/stage/bin/forge -Dpidfil.path=run.pid -Dconfig.file=/opt/forge/conf/prod.conf
13:27 Lomithrani I paster a smf manifest
13:28 Lomithrani and I'm in the wrong irc :D
13:28 Lomithrani wrong tab sorry
13:28 babilen I am utterly confused and have no idea what you are talking about, so that seems to be likely
13:29 Lomithrani yeah its a smartos problem nvm sorry for the confusion :)
13:30 babilen yeah, take your ugly XML elsewhere :-þ
13:31 Kalinakov joined #salt
13:33 abele joined #salt
13:34 or1gb1u3 joined #salt
13:35 mgw has anybody ever heard of a highstate removing files from /var/cache/salt/minion/extmods/renderers/?
13:36 fxdgear joined #salt
13:36 mgw I run sync_renderers, it's there, state.sls works fine… but when I run state.highstate, it fails (renderer not available) and the renderer is deleted
13:39 dyu1 joined #salt
13:40 kaptk2 joined #salt
13:44 elfixit joined #salt
13:45 jcsp joined #salt
13:45 ekristen joined #salt
13:49 beebeeep well, after reading http://docs.saltstack.com/en/latest/ref/states/highstate.html twice, i've realized, where i was wrong :)
13:50 beebeeep i'm sure that saltstack guys should place this doc on their main documantation page printed with HUGE letters :)
13:54 E1NS but what was the problem?
13:55 beebeeep 10:37 < beebeeep> hello folks, can anyone explain, what i'm doing wrong - there is my sls http://pastebin.com/4gw7yxUT, when i'm applying it to some host, i get "The state "disk_mpfs_mongos" in sls disk_mpfs_mongos is not formed  as a list"
13:55 beebeeep 10:38 < beebeeep> if i remove top-level id 'disk_mpfs_mongos', everything is ok
13:57 beebeeep the problem was in misunderstanding of highdata layout
13:59 ramishra joined #salt
13:59 E1NS xD
13:59 ndrei joined #salt
13:59 E1NS beebeeep: i see. is true ..  page printed with HUGE letters =P
14:01 dude051 joined #salt
14:02 ajprog_laptop1 joined #salt
14:04 ccase joined #salt
14:05 ponie joined #salt
14:05 ponie howdy,
14:08 ponie having problems running state.highstate for the first time
14:08 ponie I have configured my master with the correct file_roots
14:08 ponie confirmed that there is a top.sls that matches my minion
14:09 ponie in the master when I look at the mtime_map file I can see that the correct path for my top.sls is in place
14:09 ponie however when I run it
14:09 ponie I get No Top file or external nodes data matches found
14:09 bhosmer_ joined #salt
14:10 ponie this is the path under my mtime_map file in the master "/srv/salt/top.sls:1404137249.57"
14:10 ponie any ideas?
14:11 ponie I had been fighting this all weekend, and I thought of just doing fresh install
14:11 ponie but I get the same result
14:16 ajolo joined #salt
14:16 xinkeT joined #salt
14:22 FL1SK joined #salt
14:23 dude^2 joined #salt
14:24 dude^2 joined #salt
14:25 CheKoLyN joined #salt
14:25 dude^2 joined #salt
14:26 dude^2 joined #salt
14:27 acabrera joined #salt
14:27 dude051 joined #salt
14:29 dude^2 joined #salt
14:29 dude051 joined #salt
14:30 dude^2 joined #salt
14:32 vejdmn joined #salt
14:34 toastedpenguin joined #salt
14:36 conan_the_destro joined #salt
14:36 acabrera joined #salt
14:38 ksalman how can i sync grains in a state file?
14:41 babilen ksalman: That is being done on every highstate ... Do you really have to call it from a state?
14:43 ksalman babilen: we have a test enviromment and folks deploy various state files without doing a state.highstate after a minion vm starts
14:44 babilen ksalman: So you want to sync grains before you run the first highstate?
14:44 babilen Or rather when the minion starts?
14:44 jalbretsen joined #salt
14:44 ksalman highstate might never happen, I want to sync the grains before soemone tries to deploy a given state file
14:45 ksalman eg, salt-call state.sls test
14:45 ksalman i was wondering if i could sync the grains in the test.sls file
14:45 babilen Okay, I am simply thinking of http://docs.saltstack.com/en/latest/topics/reactor/index.html#syncing-custom-types-on-minion-start
14:45 diegows joined #salt
14:46 babilen But I am not sure if "when minion starts" is applicable in your setup
14:47 tyler-baker joined #salt
14:47 ksalman hm i think that may work
14:47 babilen You can react to other events too, but syncing grains on start might be all you need (and is a common problem that is typically handled in this manner)
14:48 ksalman thanks
14:48 babilen np
14:50 zach is there a way to apply the ACL to salt-cloud?
14:50 zach I want to use salt-cloud via Jenkins but I do not want to give Jenkins root access
14:51 ksalman babilen: on that reactor page, do you know what the difference is between 'salt/minion/*/start' and 'minion_start'?
14:51 penguin_dan joined #salt
14:51 GnuLxUsr joined #salt
14:54 babilen ksalman: minion_start is syntactic sugar for 'salt/minion/*/start'. I'd use the namespaced form if you are likely to have other events that will have to use that form
14:54 maxskew joined #salt
14:54 ksalman oh isee
14:55 babilen I have no idea why they introduced minion_start
14:56 babilen The only effect it seems to have is to confuse people who learned earlier in the documentation that "salt/minion/*/start" events are being fired when a minion starts :)
14:56 ksalman yup
14:56 thayne joined #salt
14:59 mgw how can extension modules (e.g., runners, output) be loaded form gitfs?
14:59 mgw s/form/from/
15:00 thedodd joined #salt
15:04 babilen External grains can be placed in _grains, you cannot yet place reactor files in _reactor (cf. https://github.com/saltstack/salt/issues/13671), runners and output do require a master restart don't they and can AFAIR not yet be read automatically from any _FOO in file_roots
15:06 kballou joined #salt
15:07 ipalreadytaken joined #salt
15:09 mgw babilen: that's the conclusion I'm coming to
15:09 mgw _grains, _states, _modules etc work form gitfs/file_roots
15:09 mgw but master-side modules (e.g, pillar, runner, output) don't seem to work from the tree
15:10 mgw which seems to mean they can't be in gitfs
15:10 mgw if that's true, it's a shame
15:11 slav0nic joined #salt
15:17 racooper Can anyone help get a module.run quota.set working? I seem to be missing something in syntax. https://gist.github.com/racooper/970be82b500a5c044afb
15:18 babilen mgw: You can, naturally, put them in your git repo and even check them out "manually" git, say, states.git.latest but the master cannot access them in GitFS
15:18 babilen mgw: I mean the master would have to restart itself to use them wouldn't it?
15:19 mgw actually, not for _runners
15:19 mgw salt-run loads them
15:19 mgw it just doesn't look in the gitfs tree
15:19 mgw babilen: ^
15:19 babilen Ah, in that case it would make perfect sense to load them dynamically from _runners wouldn't it?
15:19 mgw yes, I would think so
15:20 babilen I'd file a bug .. I had the same thing with _reactor and upstream was happy about the suggestion. I mean salt is a project that is heavily developed .. Just because something isn't there doesn't mean that it was necessarily exluded by design.
15:20 kivihtin joined #salt
15:21 mgw to clarify… i can add runners to the directory that runner_dirs points to and they're immediately availabe
15:21 mgw babilen: absolutely agree
15:21 mgw I'm looking at how it can be added
15:21 babilen I would guess that it was simply not yet implemented... GitFS isn't *that* super old either and the "I want to do everything in GitFS" fraction seems to grow steadily ..
15:21 scarcry joined #salt
15:22 tkharju joined #salt
15:23 oz_akan_ joined #salt
15:24 ksalman So I am using the reactor to sync grains before doing a highstate, and sometimes it does not work. https://gist.github.com/ksalman/df58936f572ff4785665
15:24 ksalman And smoetimes it does work
15:25 ksalman it errored out even thouugh grains got synced?
15:25 troyready joined #salt
15:26 ndrei joined #salt
15:27 Ahlee does salt support a signal to rotate logs without restarting?
15:28 pholbrook joined #salt
15:29 pholbrook In one of our pillars, we have these lists of minion names that are getting excessively long:
15:30 pholbrook foo: L@m1.ourdomain.com,m2.ourdomain.com, ....... etc
15:30 pholbrook can we use the yaml convention of '>' to split those over multiple lines?
15:30 pholbrook foo: >
15:30 agend joined #salt
15:30 pholbrook L@m1.ourdomain.com
15:31 pholbrook m2.ourdomain.com,
15:31 pholbrook etc
15:31 pholbrook (assume comma after m1)
15:31 diegows joined #salt
15:31 pholbrook in yaml the '>' is supposed to join those multiple lines into a single string with no CRs
15:35 laubosslink joined #salt
15:36 slav0nic joined #salt
15:36 slav0nic joined #salt
15:40 yomilk joined #salt
15:40 zach salt & jenkins sure don't play nicely
15:40 vu joined #salt
15:41 zach salt <server> state.sls <state> -- even if it is successful returns an exit code of 1 which causes Jenkins to think it failed
15:41 zach at least on my machine..
15:44 tligda joined #salt
15:46 rallytime joined #salt
15:46 timoguin zach: looks like there's been a lot of work done with return codes in the develop branch (Helium)
15:46 timoguin https://github.com/saltstack/salt/issues/4176
15:46 timoguin https://github.com/saltstack/salt/pull/11337
15:48 ksalman any idea why this reactor configuration works only sometimes? Othertimes it complains that custom grains wern't found even though the sync happend before highstate https://gist.github.com/ksalman/6e37c0d31a535f1ae7fe
15:50 ponie can I rename an aws minion using salt-cloud?
15:52 pdayton joined #salt
15:53 ponie BTW, regarding my earlier post on minion unable to get to top.sls file. I was using the install that came from ubuntu repos...Once reinstall using the one provided on salt docs it worked..
15:53 oz_akan_ joined #salt
15:53 Eivind joined #salt
15:55 oz_akan__ joined #salt
15:57 jnials joined #salt
16:00 Lomithrani joined #salt
16:00 pdayton joined #salt
16:01 Lomithrani left #salt
16:02 mateoconfeugo joined #salt
16:04 Eivind left #salt
16:04 chiui joined #salt
16:06 UtahDave joined #salt
16:09 slav0nic joined #salt
16:09 thedodd joined #salt
16:14 KyleG joined #salt
16:14 KyleG joined #salt
16:15 doddstack joined #salt
16:15 vejdmn joined #salt
16:23 bhosmer joined #salt
16:24 thedodd joined #salt
16:24 meteorfox joined #salt
16:24 pdayton joined #salt
16:29 jnials joined #salt
16:29 jnials joined #salt
16:36 bhosmer joined #salt
16:39 ramteid joined #salt
16:39 chitown is there a way to delete say, 10 minion keys?
16:40 chitown out of a couple of thousand
16:40 timoguin random ones?
16:40 chitown random = it doesnt match a nice glob, then yes
16:40 joehillen joined #salt
16:40 chitown eg a comma sep list
16:41 dvogt joined #salt
16:41 timoguin don't think there's a built-in way, no
16:41 timoguin but you could just use xargs or something
16:41 viq for i in 1 2 3 4 ; do salt-key -y -d $i ; done
16:41 chitown that doesnt work either
16:42 chitown every time you do that, the master key is regenerated
16:42 chitown when you have 10k minions, thats kind of a problem :)
16:42 viq whaaaa?
16:43 bmatt the master key is wha?
16:43 chitown looking for code ref... need a few mions
16:43 chitown mins*
16:43 malinoff joined #salt
16:44 bmatt chitown: if that's the case, that might explain a really odd bug we've been chasing :)
16:44 chitown but, yes; master key is changed when you delete (but, i have this info 3rd hand... trying to verify)
16:45 slav0nic joined #salt
16:45 ramishra joined #salt
16:46 Ahlee The master key is not changed
16:46 Ahlee salt-key is a wrapper around file system management of hte keys directory
16:47 chitown hmm... we have a weird situation that the master key is regenerated on every minion delete
16:47 chitown one sec
16:47 Ahlee Are you positive you're using individual minion keys?
16:47 Ahlee like, is it possible every one of your minion keys is the same/
16:47 acabrera joined #salt
16:48 chitown lol... no :)
16:48 chitown but, we do have auto accept on
16:48 Ahlee Same.
16:48 Ahlee /etc/salt/pki/master/minions
16:49 chitown ok... but, the keys are not all the same
16:49 oz_akan_ joined #salt
16:49 Ahlee Ok.  in /etc/salt/pki/master/, your master.pem file is getting removed?
16:49 Ahlee or at least, recreated?
16:50 Ahlee and to reprodue this, you're running salt-key -y -d <fqdn>
16:50 Ahlee well
16:50 Ahlee sorry, minion_id, not fqdn
16:50 bhosmer_ joined #salt
16:50 dstokes is there an option in salt-bootstrap to install the stable release w/o using a package manager, similar to the way installing latest runs?
16:51 viq dstokes: you can give it git tag to install, yes
16:51 chitown yes... that command (in our case, fqdn == minion id, but the precision is appreciated :> )
16:51 dstokes viq: duh... thx ;p
16:51 manfred dstokes: -P should install it with pip
16:51 vexati0n joined #salt
16:51 manfred (i can't remember if it does pip for salt, or just for the depends)
16:52 vexati0n so... i have a minion that is fully connected to the master but not receiving any commands from the Pub port
16:52 Ahlee chitown: Which version of salt?
16:52 vexati0n i can use tcpdump to watch the ack packets go back and forth, and pass nonsense on a telnet connection and everything shows up
16:52 Ahlee salt-call --versions
16:52 vexati0n but salt never gets a command
16:52 pdayton joined #salt
16:52 Ahlee vexati0n: Positive the key is accepted?
16:52 vexati0n Ahlee: yes the key is accepted
16:53 Ahlee what version of zmq?
16:53 vexati0n it's a brand new key, i even went to open auth mode
16:53 vexati0n 3.2.2
16:53 manfred salt-call --version-report
16:53 manfred sorry, --versions-report
16:53 UtahDave vexati0n: are both ports 4505 and 4506 open on your master's firewall?
16:54 Ahlee vexati0n: has it ever worked?'
16:54 chitown Ahlee: 2014.01.4
16:54 vexati0n yes . well it isn't 4505 and 4506 because we changed them, but yes they're open and i have 600 other minions that work just fine.
16:54 chitown Ahlee: salt.crypt.dropfile(self.opts['cachedir'], self.opts['user'])
16:54 vexati0n Ahlee: this minion is new so no it has never worked
16:54 chitown from salt/key.py
16:55 shaggy_surfer joined #salt
16:55 chitown the comment on that method says it updates the session key
16:56 higgs001 joined #salt
16:56 Ahlee yeah, https://github.com/saltstack/salt/blob/develop/salt/key.py#L580
16:56 Ahlee interesting.  er, not develop
16:56 Ahlee stupid github
16:56 vexati0n UtahDave: the ports are open, otherwise a telnet connection would be established, but it is, and i can watch tcpdump on the master to see the telnet session connect and write data back and forth
16:56 Ahlee anyway
16:57 chitown thats fine, i follow :)
16:57 Ahlee vexati0n: have you turned the logging up to debug to verify nothing strange is going on?
16:58 druonysus joined #salt
16:58 vexati0n yeah i have the master on debug and the minion on trace... they both just act like they don't see anything.
16:58 vexati0n and when i run tcpdump on the minion, then send a command from the master, nothing happens incoming from the pub port.
16:58 Ahlee chitown: crazy
16:58 Ahlee er, vexati0n
16:58 vexati0n it's like the master is just not giving any craps about that minion when it pushes commands.
16:58 Ahlee i've seen similar with duplicate named minions
16:59 Ahlee but, that was a while ago.
16:59 chrisjones joined #salt
17:00 bmatt I just deleted a key on our 2014.1.4 test setup and couldn't reproduce
17:00 viq vexati0n: when you restart the minion, does tcpdump show it sending packets towards master?
17:00 vexati0n the only data sent from the salt-master is ACK packets in response to ZMQ from the minion
17:00 viq ah, it does
17:01 chitown Ahlee: ok, i think i understand... the dropfile() method drop the entire cachedir (correct?) and that is called for every delete request
17:01 chrisjones joined #salt
17:01 chitown is that right?
17:02 Ahlee chitown: that's how i'm reading it, but i'm trying to remember how to pull in __opts__ to test
17:02 chitown :)
17:02 chitown so, the problem with us is that we have 10k-ish minions on a master
17:02 chitown so, a single delete has a pretty substantial effect :/
17:03 Ahlee who needs a cache, right? ;)
17:03 chitown DELETE ALL THE THINGS!
17:03 chitown clea slate wfm
17:03 chitown clean*
17:03 viq nuke it from orbit, it's the only way to be sure!
17:04 chitown :)
17:04 chitown in the case of a delete; does it make more sense to just delete the affected minion cache?
17:05 malinoff I guess you can use /dev/null http://devnull-as-a-service.com/ to store your cache. It's fast, and it scales!
17:05 chitown that would remove some of our problems with 10k minions
17:05 malinoff Agreed
17:05 chitown 1 minion, 1 MILLLION.. all the same
17:06 vexati0n i cache nothing and dump it all to mongo. i call it the "see what sticks" method.
17:06 Gareth morning
17:06 chitown ya, we are moving some stuff to couchbase
17:06 chitown mostly job return for the moment
17:07 vexati0n we're actually using salt as the platform for managing field appliances. works pretty well when TCP doesn't fail.
17:07 bmatt chitown: I'm still not totally clear on what you're seeing
17:08 bmatt I deleted a minion key from the master without the master reissuing
17:08 chitown ya, i think there was some confusion on my end...
17:08 chitown but, it does look like on a delete, the entire cache dir is deleted?
17:09 Ahlee i'm in salt.utils.__init__.py rightnow looking at their fopen overload
17:09 Ahlee as https://github.com/saltstack/salt/blob/develop/salt/crypt.py#L39-L67
17:09 Ahlee er, not far enough
17:10 Ahlee shutil.move, looks like it's just moving the affected file out of the way, from dfnt to dfn
17:10 forrest joined #salt
17:10 Ahlee but why is it in cachedir at all for keys?
17:11 jaimed http://blog.pagerduty.com/2014/06/10-common-ops-mistakes/
17:11 jaimed check point 3
17:11 Ahlee jaimed: neat!
17:11 logix812 joined #salt
17:12 bmatt "up to four or five server types, going bigger will require a more robust tool like Chef or Puppet"
17:12 bmatt what.
17:12 chitown lol
17:12 chitown i think i know what pagerduty guys use internally
17:12 Ahlee heh
17:12 Ahlee i didn't read that far
17:12 Ahlee not neat
17:13 picker joined #salt
17:13 timoguin lol that's bananas
17:15 timoguin that article is... on the weak side
17:15 malinoff I may partially agree with that point. Complex deployments with salt or ansible is a pain
17:15 jaimed Funny thing is how they just say that without offering any explanation about their reasoning
17:15 timoguin don't use personal accounts when setting up infrastructure for a business.... hmmm.... okay
17:15 shaggy_surfer joined #salt
17:15 timoguin 2. use a test environment.... hmmmm.... ookay
17:16 tristianc|Alt joined #salt
17:16 timoguin have a process. use backups. don't ignore alerts.
17:16 Ahlee chitown: i'm setting up a test environment for 2014.1.4 with some additional logging, going to write out contents of cachedir before and after
17:16 timoguin pagerduty you have taught me nothing!
17:17 malinoff The most suitable tool for me is AWS OpsWorks in conjunction with AWS CloudFormation. Complex deployments are easy as pie, but you can't really manage infrastructure with it
17:18 forrest timoguin, I feel like this article is for managers
17:18 timoguin probably. i stopped reading.
17:18 forrest yea I skimmed
17:18 matrix3000 joined #salt
17:19 Eugene Getting stuff off my personal Amazon account is one I'm putting off
17:20 Eugene I don't have anything important running, but it looks like its gonna be a pain to migrate S3 bukkits & EC2 AMIs :-/
17:20 schimmy joined #salt
17:20 malinoff Eugene, use s3cmd, Luke
17:20 Eugene Also I still have a few hundred bucks worth of "free credit" from a Trial offer to use up
17:20 Eugene Yeah, that's what I'm doing for the upload right now. But it takes a while to move several hundred thousand objects.
17:20 forrest Eugene, I blew through 100 bucks of free credit from pycon 2013 in a month, and ended up owing them money
17:20 forrest was like 'yea no more amazon'
17:21 Eugene I have $700 >_>
17:21 forrest so... like two weeks with 4 servers?
17:21 Eugene Actually I've had 4 c1.xlarge up for two months now
17:21 Eugene Spot Instances ftw
17:21 forrest he
17:21 forrest *h
17:21 AviMarcus joined #salt
17:21 Eugene Look what it does for my stats! http://stats.distributed.net/team/tmsummary.php?project_id=28&amp;team=20673
17:22 chitown Ahlee: so, it looks like it just changes {cachdir}.dfn to .dfnt
17:22 chitown and that is some aes key string
17:22 ndrei joined #salt
17:23 forrest this website blows Eugene
17:23 Eugene Welcome to 1998
17:23 forrest yea, I love their usage of the PHP logo
17:23 chitown sorry, other way around; it creates a new one (.dfnt) and moves that on top of .dfn
17:23 forrest PROUD PHP USER!
17:23 Eugene It's a hell of a drug
17:23 Eugene Also: if it ain't broke, don't fix it, etc.
17:24 schimmy1 joined #salt
17:25 Ahlee chitown: yeah
17:25 matrix3000 PHP FTW!
17:25 matrix3000 just ask facebook
17:25 chitown why does changing that file force all of the minions to reauth?
17:27 chitown Ahlee: https://github.com/saltstack/salt/blob/develop/salt/master.py#L669-L693
17:27 to_json joined #salt
17:28 masterkorp matrix3000: lolwat
17:29 miqui joined #salt
17:31 matrix3000 i was just adding to forrest's comment, facebook used PHP and then even HipHop for PHP.
17:32 dvogt joined #salt
17:33 tristianc|Alt I'm trying to use a salt state to append multiple entries to iptables
17:33 ldlework tristianc|Alt: cool story
17:33 tristianc|Alt But when I apply the state, I'm told there are conflicting IDs
17:33 tristianc|Alt https://gist.github.com/tristianc/e8493d21f25d0e4a4eed
17:33 ldlework tristianc|Alt: because you have the same state multiple times
17:34 schimmy joined #salt
17:34 tristianc|Alt How do I use iptables.append to apply multiple entries?
17:34 tristianc|Alt Example documentation also uses identical state names:
17:34 tristianc|Alt http://docs.saltstack.com/en/latest/ref/states/all/salt.states.iptables.html
17:35 ldlework tristianc|Alt: right and you can't see the difference between the documentation's version and your version?
17:36 marco_en_voyage joined #salt
17:38 tristianc|Alt As far as I can tell, the difference is that each iptables.append rule is nested in it
17:38 tristianc|Alt 's own identically named state
17:38 tristianc|Alt https://gist.github.com/tristianc/2b42b4270ebd89cb5fae
17:39 tristianc|Alt when I apply the above, I get a similar error: Conflicting ID "nfs"
17:41 malinoff Just found some YAML's cool stuff, it may save your time. Not sure if this works in salt, but you can try: http://yaml.org/spec/1.1/current.html#id857905 example 2.10
17:42 malinoff More complex example here: http://yaml.org/spec/1.1/current.html#id859060 example 2.27
17:43 malinoff In a two words - you can make an anchor to some structure (more or less complex) and then refer to it elsewhere
17:43 felskrone joined #salt
17:43 oz_akan_ joined #salt
17:48 luminous when using require_in in one state, to another, do you need to include the formula that has the 'other' state?
17:50 ml_1 joined #salt
17:51 luminous tristianc|Alt: you might also want to take a look at reclass, as a means for (sensibly) defining dynamic pillar config, per host. I save my sanity by writing formula that uses reclass for stuff like you are talking about - eg, you define the 'details' of the ip tables rules in your reclass class definitions, and the the formula uses that to simplify your life
17:52 kermit joined #salt
17:57 AviMarcus left #salt
17:58 rglen joined #salt
18:01 kiorky joined #salt
18:01 Ryan_Lane joined #salt
18:02 oz_akan_ joined #salt
18:03 pmcg joined #salt
18:03 oz_akan_ joined #salt
18:03 forrest here's cachedout talking about stuff at velocity: https://www.youtube.com/watch?v=Y4JTq2tUhM4&amp;feature=share&amp;list=PL055Epbe6d5aC6OQSRlkYCv7bu-5bc_Is&amp;index=6
18:03 forrest his beard is getting crazy
18:06 Flusher joined #salt
18:08 dvogt joined #salt
18:09 \ask joined #salt
18:10 monokrome joined #salt
18:10 monokrome joined #salt
18:11 cheus joined #salt
18:12 smcquay joined #salt
18:12 malinoff forrest, thanks for the link, nice explanation about why people shouldn't use docker
18:12 forrest lol
18:12 cheus Hi, is there a way to send some type of warning message to state output (from an .sls)?
18:13 malinoff Well, that sounds like that ;)
18:13 dvogt joined #salt
18:13 bmatt cheus: nope
18:13 bmatt cheus: such a thing isn't really possible due to the way state modules work
18:14 bmatt unless you wrote a "warning" state which always returns failure with a message
18:15 mateoconfeugo joined #salt
18:16 austin987 joined #salt
18:18 cheus bmatt, That's good enough, I think.
18:19 ndrei joined #salt
18:19 mateocon_ joined #salt
18:21 ckao joined #salt
18:25 thedodd joined #salt
18:27 bhosmer joined #salt
18:28 pjs joined #salt
18:29 vejdmn joined #salt
18:31 oz_akan_ joined #salt
18:33 ndrei joined #salt
18:34 oz_akan__ joined #salt
18:39 dvogt joined #salt
18:46 luminous when using the salt client (salt.client.Caller()/etc) , what is the best way to capture log ouput?
18:49 mapu joined #salt
18:52 ecdhe joined #salt
18:53 oz_akan_ joined #salt
18:53 kermit joined #salt
18:54 toastedpenguin1 joined #salt
18:55 napper joined #salt
18:56 Damoun joined #salt
18:59 davet joined #salt
19:08 pdayton joined #salt
19:10 oz_akan_ joined #salt
19:10 forrest Does anyone remember if there's a good way using a jinja if statement to force a state to bomb out?
19:10 forrest basically if pillar_val set, do thing, else, exit
19:14 manfred forrest: use an unless statement, wrap it in jinja, and then just make it return 1, and do hardfail
19:15 forrest manfred, ugh
19:15 manfred yeah it is ugly
19:16 bmatt {% if pillar('foo') %} state: ... {% else return 1 %} ?
19:16 manfred bmatt: don't know that that causes it to hard fail out of the state run
19:16 obimod joined #salt
19:16 bmatt the only think that jinja could make fail is the parser, right?
19:16 bmatt because once it gets to the salt module level, it's about Real Daya
19:16 bmatt er, Data
19:17 manfred right, but he wants it to hard fail during the state run, not the parser
19:17 bmatt er, state module. man, I need coffee
19:17 forrest *checks time* you sure you don't mean lunch bmatt ?
19:17 bmatt forrest: coffee is an any-time drink
19:18 forrest bleh
19:19 bmatt so there are a few steps, right? template the sls file, parse the sls file into a dict, call the state modules in the dict, execute the states via execution modules
19:19 bmatt at which point should things fail?
19:20 forrest preferably as soon as it sees that pillar variable isn't set.
19:20 forrest because it should always be set
19:20 forrest I MIGHT just use a default which doesn't exist (using git.latest and the rev)
19:20 forrest I was just hoping there was some quick way I was missing
19:21 bmatt yeah, if the pillar value doesn't exist it'll bomb out at the templating step
19:21 bmatt but it's not elegant
19:21 forrest bmatt, elegance be damned!
19:23 Hell_Fire joined #salt
19:24 Thiggy joined #salt
19:24 mgw joined #salt
19:25 scarcry joined #salt
19:27 alekibango joined #salt
19:30 obimod bah... SaltReqTimeoutError
19:31 xzarth_ joined #salt
19:34 VictorLin joined #salt
19:34 pdayton joined #salt
19:39 oz_akan_ joined #salt
19:43 oddsock2 joined #salt
20:04 oz_akan_ joined #salt
20:05 vejdmn joined #salt
20:05 or1gb1u3_ joined #salt
20:11 smcquay joined #salt
20:12 mateoconfeugo joined #salt
20:13 nEcko__ joined #salt
20:13 colinbits joined #salt
20:16 AdamSewell joined #salt
20:17 active8 joined #salt
20:20 scarcry joined #salt
20:21 mateoconfeugo joined #salt
20:29 Ryan_Lane when I use s3:// in file.managed, why does it try to connect to 169.254.169.254 ?
20:30 Ryan_Lane I'm providing s3.key and s3.keyid in my pillars
20:34 UtahDave Ryan_Lane: Just a guess, is that the ip of the internal aws metadata server?
20:34 druonysuse joined #salt
20:34 Ryan_Lane yep. it's probably trying to get the role credentials from the metadata server
20:34 Ryan_Lane but I'm providing keys and I'm in vagrant
20:34 Ryan_Lane so there's no reason it should be trying to access metadata
20:35 chrisjones joined #salt
20:36 UtahDave yep. I haven't been in that code at all, but I wonder if that use case was an oversight
20:38 chrisjones joined #salt
20:39 Ryan_Lane when I use s3.get things work properly
20:40 ipalread_ joined #salt
20:40 Ryan_Lane I wonder if the code actually needs keys in the minion config, rather than in grains
20:40 Ryan_Lane err
20:40 Ryan_Lane pillars
20:40 Ryan_Lane sigh
20:40 Ryan_Lane it does
20:41 UtahDave Hm. It should allow for either.
20:42 ksalman can onfail be used for pkg.install?
20:42 schimmy1 joined #salt
20:42 Ryan_Lane UtahDave: https://github.com/saltstack/salt/issues/13850
20:43 UtahDave thanks!
20:43 kermit joined #salt
20:44 schimmy2 joined #salt
20:45 felskrone joined #salt
20:52 druonysuse joined #salt
20:55 ksalman any idea why syncing grains via reactor, before deploying a state file via reactor, does not always work? https://github.com/saltstack/salt/issues/13843
20:55 mortis just bought a new tshirt http://www.cafepress.com/mf/85190173/lowsodium-freaks_tshirt?productId=1206835773
20:56 UtahDave nice, mortis
20:56 mortis :)
20:56 UtahDave ksalman: are the grains finished syncing before the state file is executed?
20:57 ksalman UtahDave: I would assume so? it works sometimes so maybe they don't finish syncing when it does not work?
20:58 smcquay_ joined #salt
20:58 ksalman shouldn't it wait for it to sync?
21:00 tligda1 joined #salt
21:02 MTecknology How would I write a template that's  {% foreach boothost in salt.key('boothost*') %} {{ boothost['id'] }}, {{ boothost['eth0']['ip_addr'] }} {% endfor %}    I'm sure this is easy for someone smarter than me...
21:02 tligda2 joined #salt
21:03 MTecknology Can I even do that? Templates are rendered client side, so salt keys and node info isn't available...
21:04 UtahDave correct
21:05 MTecknology Can I pass that info using context?
21:05 vejdmn joined #salt
21:06 MTecknology UtahDave: is a pillar of that data the right choice here?
21:06 druonysuse joined #salt
21:06 druonysuse joined #salt
21:06 kiorky joined #salt
21:07 UtahDave MTecknology: Yeah, that's one way. You might also look at using the mine, perhaps
21:07 MTecknology yup, mine, mine was created for this, huh?
21:08 UtahDave Yeah, pretty much.  :)
21:09 ajprog_laptop1 joined #salt
21:09 MTecknology thanks!
21:09 kermit joined #salt
21:10 MTecknology oh, bummer... I don't want to need to change all of the minion configs...
21:10 MTecknology makes sense, though. I'll probably play with this more tomorrow.
21:11 ipalreadytaken joined #salt
21:11 ksalman MTecknology: you could manage the minion's config via salt =)
21:12 MTecknology I do, but pushing an update to 180 servers over a T1 isn't a small task.
21:13 ksalman ah
21:14 Ahlee and restarting minion is always a crap shoot
21:15 ksalman so far i've only had issue restarting minion on windows
21:16 MTecknology Ahlee: we'll find out if this is gonna actually work or not... I now set an at task to restart it 1min after salt finishes
21:17 aw110f joined #salt
21:21 Ahlee That's our method, as well.  Start atd, and then if atd starts if usually works
21:21 Ahlee unfortunately, atd likes to return 0 but not start for us
21:22 MTecknology I just leave it always running
21:22 MTecknology I have one box that I just can't get responding to salt
21:23 MTecknology salt-minion -l debug  acts like it connects with issues, salt 'id' says it's not returning and nothing from stdout... *grumble8
21:23 MTecknology s/8/*/
21:24 doddstack joined #salt
21:24 Ahlee I haven't got a reliable enough way of monitoring jobs run through at to leave it running always
21:26 MTecknology weird...
21:26 MTecknology I had to reboot the box
21:27 rallytim_ joined #salt
21:27 higgs001_ joined #salt
21:27 schimmy joined #salt
21:27 devx_ joined #salt
21:27 phx_ joined #salt
21:28 EntropyWorks_ joined #salt
21:28 tligda joined #salt
21:28 rallytim_ joined #salt
21:28 horus_plex joined #salt
21:29 Xiao joined #salt
21:29 rallytim_ joined #salt
21:29 rallytim_ joined #salt
21:29 thayne_ joined #salt
21:29 esogas` joined #salt
21:30 ipalreadytaken joined #salt
21:30 yomilk joined #salt
21:30 tempspace_ joined #salt
21:31 peno_ joined #salt
21:31 E1NS_ joined #salt
21:32 Spark__ joined #salt
21:34 rallytim_ joined #salt
21:34 rallytim_ joined #salt
21:35 seb`- joined #salt
21:35 tligda1 joined #salt
21:36 TaiSHi joined #salt
21:36 rallytime joined #salt
21:36 TaiSHi Hi all
21:36 TaiSHi Is there a way to start a minion in debug mode?
21:37 vlcn joined #salt
21:37 babilen Sure, "salt-minion -ldebug"
21:37 Daviey joined #salt
21:37 juice joined #salt
21:37 TaiSHi I'm provisioning my minions with salt-cloud and the VM goes up, salt is bootstrapped, then it fails to auth with the master and dies
21:37 Yoda-BZH joined #salt
21:37 TaiSHi If I restart the minion, all is good in the world
21:37 Yoda-BZH joined #salt
21:37 Phibs joined #salt
21:38 Phibs can anyone tell me how to show the SWAP usage via a salt module? I'm not finding anything quite useful yet.... (2014.1.5)
21:40 twobitsprite joined #salt
21:40 DenkBrettl joined #salt
21:40 hardwire joined #salt
21:42 active8 joined #salt
21:42 penguin_dan joined #salt
21:43 kballou joined #salt
21:44 racooper joined #salt
21:45 TaiSHi babilen: I guess I could, not sure if it'll work
21:46 hardwire joined #salt
21:46 robawt joined #salt
21:46 TaiSHi hmm
21:48 TaiSHi babilen: Wont do :(
21:50 Phibs can anyone tell me how to show the SWAP usage via a salt module? I'm not finding anything quite useful yet.... (2014.1.5)
21:51 felskrone joined #salt
21:52 forrest Phibs, http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.mount.html#salt.modules.mount.swaps ?
21:53 Phibs hmm
21:54 Phibs forrest: danke sir
21:54 forrest np
21:54 Phibs I'd not expect it under 'mounts' more of a 'system info' cmd ;)
21:56 dude051 joined #salt
21:58 andredieb joined #salt
21:59 ksalman any idea why syncing grains via reactor, before deploying a state file via reactor, does not always work? https://github.com/saltstack/salt/issues/13843
22:05 oz_akan__ joined #salt
22:09 shaggy_surfer joined #salt
22:11 dvogt joined #salt
22:18 mgw joined #salt
22:23 TaiSHi Any configuration file option that might help me enable debug on minion?
22:24 babilen TaiSHi: Yes, you stop the minion and then run "salt-minion -ldebug"
22:24 dstokes TaiSHi: http://docs.saltstack.com/en/latest/ref/configuration/minion.html#log-level-logfile "debug"
22:24 TaiSHi babilen: The thing is, if I run the minion manually
22:25 TaiSHi It will work :P
22:25 TaiSHi It isn't working if I provision with salt-cloud
22:25 babilen How is that a problem?
22:25 babilen Well, now you are just making it complicated ;)
22:25 babilen But see the documentation on log-files and how to set log levels that dstokes just pointed out. :)
22:27 smcquay joined #salt
22:28 TaiSHi babilen: that seems to be exactly it
22:28 * TaiSHi hugs babilen and dstokes
22:30 Luke_ joined #salt
22:32 danielbachhuber joined #salt
22:37 * TaiSHi ¬¬
22:37 TaiSHi It seems now
22:37 TaiSHi [salt.minion      ][INFO    ] Authentication with master successful!
22:38 kiorky joined #salt
22:40 TaiSHi So, if minion authed successfully
22:40 TaiSHi startup_states: highstate
22:40 TaiSHi Why the f isn't it being provisioned?
22:42 garthk joined #salt
22:43 bhosmer joined #salt
22:43 TaiSHi bootstrapping again
22:52 rlarkin|2 joined #salt
22:52 mgw joined #salt
22:52 shaggy_surfer joined #salt
22:59 robawt joined #salt
22:59 TaiSHi When installing from source, are there default init scripts somewhere?
23:00 mapu joined #salt
23:00 TaiSHi There we go
23:00 Eureka_ joined #salt
23:01 manfred should be in pkg
23:02 TaiSHi Yeah, had to delete old ones from deb packages
23:02 Eureka_ Hi there. If anyone has experience using salt reactor on 214.1.5 I would really appreciate it. It seems like its not firing my events after the auth completes.
23:02 manfred Eureka_: did you check using the eventlisten.py test script?
23:03 manfred https://github.com/saltstack/salt/blob/develop/tests/eventlisten.py
23:04 Eureka_ Yes. It shows that I am succeeding with the Salt/Auth but no even is fired after that. If i manually run a command from the salt server such as salt '*' test.echo 'foo: bar' I see the full event pass between the server and the minions. If i run a salt-call from the minions I am not seeing the event on the server.
23:04 Eureka_ no event*
23:04 * TaiSHi is firing a deployment again
23:04 TaiSHi Wish there were apt for dev branch
23:06 TaiSHi Was using old master with new minions
23:06 TaiSHi And over the weekend it stopped failing
23:06 function07 joined #salt
23:06 TaiSHi Failed sign in <-- that
23:06 TaiSHi Google gave no answers
23:07 manfred Eureka_: i haven't experienced that at all.
23:07 Eureka_ I was worried about that o.0
23:08 Eureka_ running the server in debug im not really seeing anything usful either. I do see "Gathering reactors for tag salt/auth" I dont get anything after that thought.
23:08 Eureka_ However... I am not running any reactors for salt/auth
23:08 Eureka_ (other than what is already built in i suppose)
23:08 manfred match something on it, and see if it shows up.
23:09 manfred Eureka_: so
23:09 * TaiSHi is using raet
23:09 manfred Eureka_: the events you see in eventlisten.py should be the event.
23:09 TaiSHi I know I shouldn't
23:09 manfred there is nothing else using the events except the reactor
23:10 Eureka_ @manfred That seems to make sense. However When I am using eventlisten.py I am not getting anything past the "end public key" section of the authentication. Unlike what I have seen posted by other peoples examples.
23:11 manfred ahh
23:11 Eureka_ @manfred An example graphic from the web: http://vbyron.com/blog/wp-content/uploads/2014/03/salt307.png
23:11 manfred yeah, then i have no idea
23:11 Eureka_ I only get to the end of that first *set*
23:12 TaiSHi Well it was master's issue
23:12 TaiSHi Can't wait for Helium
23:12 Eureka_ @manfred Darn. I did attempt to copy the example exact. Guess ill spin up a couple VM's and see if its the version I am using from the repo.
23:13 Eureka_ @manfred Thanks for trying to help =) Ill let you know if I figure it out.
23:13 TaiSHi Eureka_: Didn't really read it all (fever again) but check if master/minion are using same version or master > minion
23:13 TaiSHi Just ran into that
23:13 TaiSHi Hoping THIS fixes my issue (first try ran out well)
23:13 chrisjones joined #salt
23:14 Eureka_ @TaiSHi I just did that ;) Saw your issue.
23:15 manfred TaiSHi: how are you dealing with raet endlessly createing yards and filling up your filesystem?
23:16 mgw joined #salt
23:16 TaiSHi manfred: I just ...
23:16 TaiSHi just...
23:16 TaiSHi installed salt-master with raet
23:17 TaiSHi Ok.
23:17 manfred raet from git or from a packaged version?
23:17 TaiSHi dev
23:17 manfred so from git?
23:17 manfred it is all dev
23:17 TaiSHi I guess, I just python setup.py install --salt-transport=raet
23:17 manfred so it installed raet from pip?
23:19 TaiSHi Well, it should...
23:19 TaiSHi I guess
23:19 manfred ok
23:19 * TaiSHi loves throwing things into production
23:19 TaiSHi It installed them from the source from git
23:19 TaiSHi I think
23:20 TaiSHi Yep
23:20 TaiSHi Or not, it didn't install it at all
23:20 manfred how did you install raet?
23:20 manfred yeah
23:20 TaiSHi Great, problem solved.
23:20 manfred yeah, raet doesn't work really yet
23:21 TaiSHi Someone suggested it to me
23:21 TaiSHi Very recently
23:21 TaiSHi On this channel
23:21 manfred the only thing you need for raet is that you put transport: raet in /etc/salt/{master,minion}.d/transport.conf or some minion/master thingy
23:21 TaiSHi Damn fever I can't think
23:21 manfred they shouldn't, it hasn't stableized from what I have experienced
23:21 TaiSHi Sorry, mentioned, not suggested
23:21 manfred yar
23:21 TaiSHi There is no reference to raet in /etc/salt/master
23:21 rallytime joined #salt
23:22 Eureka_ @manfred Ouch. Looks like it might be a problem with the master/minion available in epel-testing.
23:22 TaiSHi YAY try #2 worked as well
23:22 Eureka_ Just installed on 2 systems with the epel stable and its working fine.
23:22 manfred terminalmage: ^^
23:23 TaiSHi Firing up 3 instances
23:23 TaiSHi Any way to improve deploy speed with salt-cloud ?
23:24 manfred just deploy them all at the same time?
23:24 manfred i mean, it really depends on your provider...
23:24 manfred most of the time is spinning up the device
23:25 TaiSHi It's a lot of time doing apt-get update :P
23:25 manfred sure
23:25 manfred but that is a consequence of your provider not updating their images :P
23:25 Eureka_ Yeah ;)
23:25 TaiSHi Even from DO's own services
23:25 whytewolf joined #salt
23:25 TaiSHi Sometimes it runs at 1.5kB/s
23:25 manfred put in a support ticket and ask them why their stuff is slow
23:26 TaiSHi I will
23:26 TaiSHi Oh, my old provider is on a 21 hour downtime for my 4 KVM machines
23:26 manfred but yeah, beyond that, there isn't much you can do
23:27 sxar_ joined #salt
23:28 Eureka_ bah
23:28 bhosmer joined #salt
23:28 terminalmage manfred: ?
23:28 Eureka_ Try applying to my actual test env and its still busted for some reason.
23:28 manfred terminalmage: check eureka's messages, says something is broken with events in 2014.1.5 in epel-testing?
23:29 terminalmage I'm kinda busy right now
23:29 manfred kk
23:29 Eureka_ Its possible. It might have something to do with my modules as well. Im working on it o.0
23:29 terminalmage can someone give me the CliffNotes version?
23:29 manfred terminalmage: no events in the reactor past auth in the testing version, they work in the none testing version?
23:29 manfred should look like this http://vbyron.com/blog/wp-content/uploads/2014/03/salt307.png
23:30 manfred but doesn't
23:30 manfred that is all i got
23:30 manfred i can't replicate the issue
23:30 forrest TaiSHi, one of the quickest ways you could speed up deployment is by mirroring the repo that is needed to install a lot of packages
23:30 terminalmage ok, well, file an issue with a test case
23:30 terminalmage I am not going to be able to look at it tonight
23:30 forrest this guy and his east coast time
23:30 terminalmage and I haven't done anything with the event system anyway
23:30 forrest it's like he has a life to live
23:31 TaiSHi forrest: but how do I give them to salt-cloud ?
23:31 manfred TaiSHi: you would have to spin  up using your own image
23:31 forrest TaiSHi, you'd have to ensure they were there for the initial image that spins up.
23:31 terminalmage forrest: why does everyone think that Chicago is Eastern
23:31 terminalmage ?
23:31 terminalmage :P
23:31 terminalmage you west coast people
23:31 TaiSHi I guess I could spin my own version
23:31 forrest it's 6:30 there
23:31 forrest might as well be east coast
23:31 TaiSHi 20:30 here, diaf
23:31 forrest TaiSHi, also add that repo for any other packages you install, so that you use them specifically
23:32 forrest you're working at 8:30 at night?
23:32 forrest do you work for a sweat shop?
23:32 Eureka_ FYI I am on centos 6.5 final fully updated and with epel enabled. If it helps anyone.
23:32 forrest even 9 hours is too much, WHERE IS MY STAR TREK FUTURE
23:32 forrest I want to work like 4 hours a day
23:32 terminalmage also, we have a really bad storm that
23:32 terminalmage that just hit
23:32 terminalmage so I may lose my internet at any moment
23:33 forrest I'd share some of this sunshine, but I need it all for the winter
23:33 TaiSHi I work from 16 to 00
23:33 forrest sorry
23:33 TaiSHi Sadly
23:33 forrest that's not too bad
23:33 terminalmage though ZNC will keep me logged in and it'll look like I'm just ignoring you
23:33 forrest 2nd shift > 3rd shift
23:33 KyleG my wife works those hours 16-00
23:33 TaiSHi It is because I wake up 7:30 am
23:33 KyleG Always so tired at midnight afterwards. :(
23:33 TaiSHi And I get home ~ 1:30
23:33 forrest TaiSHi, well, whose fault is that?
23:33 forrest you're taking 1.5 hours to get home?
23:33 forrest sounds like it's time to move
23:33 TaiSHi Mine, and noone but mine
23:34 TaiSHi Well I live halfway between both universities I attend
23:34 TaiSHi Moving would require me to wake up earlier
23:34 forrest Ahh yes, the debt machines
23:34 TaiSHi They're free here
23:34 TaiSHi Well, at least the one I go to
23:35 TaiSHi (it's 1 U with 2, hmm, "places" or whatever the english word for it is)
23:35 Eureka_ Ah, Looks like I might have had issues upgrading the base "salt" package on centos. Force removing it now to reinstall the 214.1.5-1 from epel.
23:36 forrest TaiSHi, university?
23:36 TaiSHi Yeah
23:36 TaiSHi 1 institution with 2 universities
23:36 forrest gotcha
23:36 TaiSHi Or the other way around
23:36 forrest *shrug* it's all the same
23:36 TaiSHi Either way I go to class on both of them
23:36 schimmy joined #salt
23:36 TaiSHi I still don't get why I can't work from home
23:37 forrest We all ask this question every day
23:37 KyleG so much
23:37 forrest internet here, internet there, why does it matter?
23:37 forrest same reason, why do I have to work 8 hours? Who cares, the work is done in 4, I should go home
23:37 TaiSHi Well, when I'm at home
23:37 TaiSHi I can work naked
23:37 TaiSHi And do stuff whenever I'm not working
23:37 forrest another plus
23:37 TaiSHi Probably stuff you don't want to hear about
23:38 forrest yea no thank you
23:38 forrest just do everyone a favor and put down a towel
23:38 schimmy1 joined #salt
23:38 TaiSHi I wonder why it took salt-cloud to start 'highstate' on the first VM, right after 2nd VM finished up creating
23:39 kiorky joined #salt
23:39 forrest probably the package install?
23:39 TaiSHi It was finished
23:39 forrest or some configuration?
23:39 forrest the highstate was finished, but returned a second finished?
23:40 TaiSHi No no, I was logged in to the first one
23:42 TaiSHi Know what? I'm not even asking, I guess I'll see how long it takes whenever I deploy a single machin
23:42 TaiSHi e*
23:43 TaiSHi Now it's time to test this gluster thingie
23:43 TaiSHi I haven't found a good solution yet, current version of gluster doesn't work well when adding/removing bricks
23:45 TaiSHi xt: ping
23:49 mosen joined #salt
23:51 KyleG1 joined #salt
23:52 KyleG joined #salt
23:52 KyleG joined #salt
23:53 KyleG1 joined #salt
23:53 KyleG1 joined #salt
23:56 KyleG1 joined #salt
23:57 TaiSHi KyleG seems to be having network issues
23:57 KyleG1 yes
23:57 KyleG1 Power surge
23:57 KyleG1 :(
23:57 TaiSHi My UPS is burned
23:57 TaiSHi Now that you mention
23:57 TaiSHi I need to fix it ASAP
23:57 KyleG1 I can't seem to get Adium to accept the /nick command either
23:57 KyleG1 -_
23:57 KyleG1 usually I open and close it and it works
23:57 KyleG1 which is shitty
23:59 KyleG1 brb 1 more time :(
23:59 KyleG joined #salt
23:59 TaiSHi w
23:59 TaiSHi lol...
23:59 KyleG joined #salt
23:59 KyleG joined #salt

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