Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-04-07

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

All times shown according to UTC.

Time Nick Message
00:06 Luke__ joined #salt
00:10 wedgie joined #salt
00:16 snuffeluffegus joined #salt
00:23 Linz joined #salt
00:23 mafrosis joined #salt
00:26 elfixit joined #salt
00:28 travisfischer joined #salt
00:34 ndrei joined #salt
00:44 pydanny joined #salt
00:46 Luke__ joined #salt
00:53 pydanny joined #salt
00:57 malinoff joined #salt
00:58 hotsnow joined #salt
00:59 Gordonz joined #salt
01:05 kuser joined #salt
01:08 Luke joined #salt
01:16 mgw joined #salt
01:17 Linz joined #salt
01:23 joehoyle joined #salt
01:24 xzarth_ joined #salt
01:25 joehoyle_ joined #salt
01:40 pfallenop joined #salt
01:47 ilbot3 joined #salt
01:47 Topic for #salt is now Welcome to #salt | 2014.1.1 is the latest | Please be patient when asking questions as we are volunteers and may not have immediate answers | Channel logs are available at http://irclog.perlgeek.de/salt/
01:59 ajw0100 joined #salt
02:12 Linz joined #salt
02:21 joehoyle joined #salt
02:40 TyrfingMjolnir joined #salt
02:51 mgw joined #salt
02:55 cnelsonsic joined #salt
03:05 Ryan_Lane joined #salt
03:06 Linz joined #salt
03:13 borgstrom joined #salt
03:14 scarcry joined #salt
03:39 swa_work joined #salt
03:47 pydanny joined #salt
03:48 jeremyBass joined #salt
03:49 jeremyBass left #salt
03:53 AdamSewell joined #salt
03:57 travisfischer joined #salt
04:00 Linz joined #salt
04:04 anuvrat joined #salt
04:08 srage joined #salt
04:12 faldridge joined #salt
04:19 Networkn3rd joined #salt
04:33 gildegoma joined #salt
04:37 smcquay joined #salt
04:53 mikemar10 joined #salt
04:54 Linz joined #salt
04:58 ravibhure joined #salt
04:59 Networkn3rd joined #salt
05:00 sroegner__ joined #salt
05:12 anuvrat joined #salt
05:19 cyrusdavid joined #salt
05:26 yomilk joined #salt
05:28 kedo39 joined #salt
05:29 lyddonb joined #salt
05:33 aarontc joined #salt
05:34 Eugene joined #salt
05:34 zz_Cidan joined #salt
05:37 hardwire joined #salt
05:38 bhosmer joined #salt
05:44 ipalreadytaken joined #salt
05:44 z3uS| joined #salt
05:47 anuvrat joined #salt
05:48 Linz joined #salt
05:56 ninnypan_ joined #salt
05:58 Ryan_Lane joined #salt
06:21 ndrei joined #salt
06:35 balboah joined #salt
06:42 Linz joined #salt
06:49 sroegner__ joined #salt
06:50 kadel joined #salt
06:52 linjan joined #salt
06:53 ewong_ joined #salt
06:55 mtford joined #salt
06:56 joehoyle joined #salt
07:04 obimod joined #salt
07:04 swa_work joined #salt
07:08 harobed_ joined #salt
07:12 ravibhure joined #salt
07:12 Kenzor joined #salt
07:14 faldridge joined #salt
07:21 viod1 joined #salt
07:22 harobed joined #salt
07:22 ndrei joined #salt
07:35 TyrfingMjolnir joined #salt
07:36 Linz joined #salt
07:44 ndrei joined #salt
07:47 johtso joined #salt
07:50 epcim joined #salt
07:50 viod2 joined #salt
08:08 che-arne joined #salt
08:09 maveas joined #salt
08:12 mortis if i'm having a top.sls matching a prod-env for instance, and pointing to a custom.sls from there ... how can i go ahead and match on hostname inside that custom.sls? i wanna say something like myserver: - subfolder.item
08:13 mortis but i end up getting Too many functions declared in state <something>
08:15 ndrei joined #salt
08:16 nkuttler mortis: grains
08:16 mortis cant just match on hostname?
08:16 mortis well ofc i can match on hostname in the grain :)
08:19 hotsnow joined #salt
08:21 MK_FG joined #salt
08:31 Linz joined #salt
08:37 ndrei joined #salt
08:37 easylancer joined #salt
08:38 sroegner__ joined #salt
08:39 giantlock joined #salt
08:40 joehoyle joined #salt
08:46 Ryan_Lane joined #salt
08:50 mike25ro joined #salt
08:53 CeBe joined #salt
08:54 ndrei joined #salt
09:07 crazysim joined #salt
09:15 Luijk joined #salt
09:16 viq https://groups.google.com/forum/#!topic/salt-users/IwtntfbxoJA - maybe someone here will have ideas how to deal with this?
09:21 che-arne joined #salt
09:25 Linz joined #salt
09:26 easylancer joined #salt
09:28 johtso joined #salt
09:31 bhosmer joined #salt
09:38 yomilk joined #salt
09:40 ndrei joined #salt
09:44 mike25ro hi guys ... i have a weird question... on a vm minion i have these grains: lsb_distrib_id: CentOS  lsb_distrib_release: 6.4  BUT on a physical machine i have no such grains. my salt version is older.... 0.17.1 .... any idea why? the lsb_* are missing on the physical machine?
09:45 xmj do you run the same version both in vm and on bare metal?
09:49 yomilk joined #salt
10:05 carlos joined #salt
10:19 Linz joined #salt
10:27 sroegner__ joined #salt
10:29 mike25ro xmj: yep... same version
10:41 che-arne|2 joined #salt
10:44 engblom joined #salt
10:49 engblom Could anyone help me to get started? I have made top.sls with this content: http://pastebin.com/zjm0fUC9 . Then I also have created files.sls and packages.sls. "salt-call state.sls top" will complain with "Too many functions declared in state "*" in sls top"
10:53 mortis im having sort of the same question as engblom really :) ...well atleast the same error message. Trying to include a file in top.sls and match on hosts in the included file: pseudo-code here: http://pastie.org/9000429 should explain what im trying to do under the dev-part (custom)
10:53 mortis i understand that the custom.sls is trying to run functions, but i really want another layer of matching first, and dont know how
10:54 mortis the reason for this ...is to stop the editing of the top.sls
10:54 mortis and for it to be a bit more failsafe
10:57 ggoZ joined #salt
10:57 yomilk joined #salt
11:00 bitmand engblom: instead of a files.sls and packages.sls, try files/init.sls and packages/init.sls ( directories )
11:02 bitmand mortis: not sure how to include sls files in the same directory as top :)
11:02 mortis bitmand: ah, its not same dir
11:02 mortis its dev/custom.sls
11:03 mortis but id like some way to continue matching inside dev :)
11:03 bitmand mortis: you wan't to load dev/custom.sls instead of dev/init.sls ?
11:03 mortis not sure its possible
11:03 engblom bitmand: It apparently still behaves in the same way
11:03 mortis well, would it work differently if i called it init.sls? in the way of matching hosts?
11:05 bitmand mortis: I think I misunderstood the question :)
11:05 mortis ah :)
11:05 bitmand you kind of want to group inside groups .. ?
11:06 mortis yeah, sort of
11:06 mortis the dev filter in top, matches only the dev-pillar
11:06 mortis and so i want to move the rest of the logic out of the topfile
11:08 bitmand you can match in host in the state files, but not sure you can do what you want :/
11:09 bitmand please remove first 'in' from that sentence :)
11:09 mortis hmm yeah
11:09 bitmand ( and although I have done it myself, I wont recommend it :) )
11:10 rawtaz does salt master and minions run as root by default? if yes, why?
11:12 babilen rawtaz: Minions run as root as they have to perform many tasks for which superuser privileges are required
11:12 Shish yes; minions because they need to do all sorts of system-level things like installing packages and creating users; master because it needs root to integrate with PAM (though not everybody uses that feature, I don't...)
11:12 rawtaz babilen: there's this thing called sudo :)
11:13 Linz joined #salt
11:13 babilen rawtaz: Sure, you can configure that and run minions as a different user
11:13 viq rawtaz: and you can define every single thing minion will do?
11:14 rawtaz babilen: it just strikes me as odd that its using root by default, if it *can* work without permanent root privs
11:14 rawtaz viq: i dont understand your question
11:14 babilen rawtaz: http://docs.saltstack.com/en/latest/ref/configuration/nonroot.html
11:14 viq rawtaz: master can and does run without root privileges. minion, as with chef and puppet and everything else, needs to do a lot of things that require root, so by default runs as root
11:15 viq rawtaz: and it would be a bitch to try to define in sudoers every single thing that minion may try to do. And if you just let it execute every command as root, then what's the difference?
11:16 joehoyle Hey, I am trying to cross-call a costom state to git.latest, but having problems with __salt__ __grains__ etc not being defined for the git.latest, anyone know of any documentation on how to do this?
11:16 joehoyle I found https://github.com/saltstack/salt/issues/3513 but the examples don't seem to work
11:16 babilen viq: The major difference is that auth.log would show salt gaining privileges, but apart from that I agree with you
11:16 rawtaz viq: right, yeah i see what you mean by that
11:17 viq babilen: by the same account you can have in minion log a trace of everything it does
11:18 rawtaz but if there's a vulnerability in the minion it has a greater chance of causing harm if it runs as root than if it doesnt, right?
11:19 viq ish. If you give it a right to do everything, then it has a right to do everything anyway
11:19 viq Even if it's running as a different user
11:19 babilen viq: sure :)
11:19 anuvrat joined #salt
11:20 rawtaz to use the sudo feature it has to perform the action in a controller manner, while with a vulnerability it might be able to do it in a different way
11:20 rawtaz but ok. i see your points and it is what it is. i was just curious about the rationale. i think the article babilen linked to clarifies it a lot, too
11:20 babilen yay :)
11:21 rawtaz thanks both :)
11:24 TyrfingMjolnir joined #salt
11:29 ggoZ joined #salt
11:30 easylancer joined #salt
11:32 yomilk joined #salt
11:34 msil joined #salt
11:41 engblom I have a few sls-files (files.sls and packages.sls), how do I run both of them for all minions from top.sls?
11:43 ekristen joined #salt
11:46 babilen engblom: base: '*': - files - packages -- or something along those lines
11:46 viq engblom: if they're in the same dir as top.sls then your top looks fine, you just run "salt \* state.highstate"
11:49 engblom I can not get it too work. I will post all files on pastebin.
11:50 babilen good idea, please don't use pastebin.com
11:50 xmj dpaste.de << is good
11:51 babilen There are plenty (e.g. http://paste.debian.net http://refheap.com http://paste.ubuntu.com http://gist.github.com ...)
11:54 topochan joined #salt
11:54 yomilk joined #salt
11:55 engblom top.sls: http://pastie.org/9000532  files.sls: http://pastie.org/9000533  packages.sls: http://pastie.org/9000536
11:57 engblom so: "salt '*' state.sls files" and "salt '*' state.sls packages" works, but not from top.sls
12:00 babilen Here's a single paste of all of them: https://www.refheap.com/72975
12:01 babilen engblom: What happens if you try the highstate run?
12:04 workingcats joined #salt
12:05 harobed joined #salt
12:06 engblom babilen: This happens when I run "salt-call state.highstate" on the minion: http://pastie.org/9000555
12:07 Linz joined #salt
12:07 babilen engblom: Which version of salt is running on the master and minion and what happens if you try the command viq just mentioned?
12:08 engblom If I run the command viq just mentioned, I just get the id of the minion and :
12:09 engblom computername:
12:09 engblom Nothing more
12:10 engblom Master: 0.17.4, Minion: 0.16.4
12:11 babilen That is really quite old and I have no experience with those versions. I would also strongly recommend to run the same salt version on the master an minions. Please upgrade both to 2014.1.1 (or at least to the same version!) and try again.
12:13 engblom :/ This is were I hate debian from all of my hearth for. Even though I used wheezy-backport the version was 0.16.4 and unless someone has made a 3:rd party repository for debian it will be a nightmare to keep track of the security updates for this package.
12:16 mortis why not use salts own repo?
12:16 Shish --> deb http://debian.saltstack.com/debian wheezy-saltstack main
12:18 engblom Thanks, that solves the problem with maintaining updates myself!
12:21 gammalget joined #salt
12:25 joehh babilen: managed to get some time - doing ubuntu packages for 2014.1.1 now
12:25 engblom How old is 0.17.4?
12:27 joehh engblom: released (packaged for debian) in jan 2014
12:30 babilen engblom: I was rather referring to 0.16.4 which is also not that ancient, but the most important part is, I guess, to keep minions and master at the same version
12:30 babilen engblom: And please don't hate Debian
12:30 abe_music joined #salt
12:31 babilen joehh: Ah, wonderful! Did you run into any problems apart from ENOTIME?
12:33 engblom babilen: You are right in that I should not hate Debian. Sometimes it just feels frustrating when the packages from even backport is old.
12:34 babilen engblom: "stable" simply means "not changing" and has *nothing* to do with "working" or "bug free"
12:35 engblom So, now I have upgraded the minion (debian) to 2014.1.0 from the reporitory provided by salt. The master is OpenBSD running salt 0.17.4
12:35 babilen hmm, did I mention "same version" ?
12:36 engblom Yes, you did. I just do not know how I will achieve this easily... that is why I am asking how old 0.17.4 is
12:36 babilen Well, give it a try and see how it behaves now. Can't you upgrade the master to 2014.1.1 too?
12:36 babilen Oh, you have 2014.1.0 ?!
12:37 engblom I wrote wrong... 2014.1.4 I meant
12:37 babilen ITYM 2014.1.1+ds-1~bpo70+1
12:41 joehh engblom: you can get 0.17.x packages from
12:41 workingcats hello,
12:41 joehh deb http://debian.saltstack.com/debian wheezy-saltstack-017 main
12:41 workingcats i just installed the 2014.01 versions from the debian repo, and i was wondering where salt takes the host keys from?
12:41 workingcats does it use the SSH host key?
12:42 joehh otherwise, all the packages are still in http://debian.saltstack.com/debian/pool/main/s/salt/
12:42 joehh workingcats: keys are generated when the minion first starts
12:42 wkf joined #salt
12:42 workingcats ah i see, thanks
12:43 elfixit joined #salt
12:44 workingcats also, i think i found an error in either the 2014.1 repo or the docs. when i had "deb http://debian.saltstack.com/debian wheezy-saltstack-2014-01 main" in my sources.list it couldnt install due to an outdated python-zmq
12:44 workingcats however, using just wheezy-saltstack without the version it works fine
12:44 joehh babilen: pure ENOTIME - recentish addition to family makes working at nights much harder
12:44 TyrfingMjolnir joined #salt
12:44 joehh workingcats: thanks for letting me know, I'll check it out
12:45 workingcats sure, thanks
12:45 babilen joehh: Ack, is there a salt-team?
12:45 babilen pkg-salt or somesuch?
12:45 joehh sort of
12:46 joehh not terribly active beyond me
12:46 joehh debian is easy enough for me, but ubuntu is more of a stretch
12:46 xmj cedwards: does freebsd
12:47 joehh not as aware of release differences... tend to get to know the LTS releases, but the in betweens are a different storey
12:47 babilen joehh: I'm not active in Ubuntu at all, but I am sure that sbuild can be beaten into submission for that too. Wouldn't mind spending some time in the coming months on this, so ping for team details
12:48 ipmb joined #salt
12:48 engblom Success!! Thank you babilen and joehh!
12:48 babilen \o/
12:48 jeffrubic joined #salt
12:48 joehh I've started down that approach - have a cowbuilder environment for precise, but trade off between time spent on directly
12:48 joehh doing package update and improving workflow
12:48 jeffrubic joined #salt
12:49 scott_w joined #salt
12:50 babilen joehh: I am primarily working on backports for $JOB these days and used to maintain some patches against salt locally. Probably the wrong person for Ubuntu work, but just thought that I express that I'd hopefully be able to spend some of my time on that
12:51 engblom Is there a good way for syncing a whole directory? Doing each file manually with file.managed is ok for a few fils, but not for a whole directory
12:52 babilen lsyncd
12:53 joehh babilen: good to hear
12:55 babilen engblom: I mean you can also use file.recurse or drdb or unison or .... it really depends on what you want/need
12:55 jslatts joined #salt
12:56 engblom babilen: I want /etc/skel to be synced from the salt-server so it is always the newest version. Old files should automatically be removed.
12:56 babilen engblom: And just as a side note: You might want to use the proper salt modules (e.g. pkgrepo) to define your sources.list rather than simply syncing those files
12:56 engblom babilen: Thanks, I will have to look into that
12:57 babilen http://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkgrepo.html for example
13:01 Linz joined #salt
13:01 babilen engblom: Hmm, not sure if "file.recurse" deletes files :-/
13:03 mike25ro left #salt
13:04 babilen I'd probably use lsyncd or unison, but that takes it out of salt if your tests show that file.recurse does *not* delete files. If it does I'd use that.
13:05 ndrei joined #salt
13:06 jaimed joined #salt
13:07 viq engblom: 0.17.1 was a flag day, stuff older than this can't talk to stuff newer than this
13:07 engblom viq: Thanks!
13:08 workingcats oh, but (modulo bugs) 0.17.1 can talk to everthing newer and vice versa?
13:09 * babilen likes the %bugs
13:09 viq workingcats: it's recommended to keep it as close as possible, and master newer than minions. But that's a barrier where stuff definately will NOT work
13:09 workingcats viq, understood, cheers
13:09 babilen +1
13:10 viq That being said, I have 2014.1.1 minions happily talking to a 0.17.4 master for what little I do with them
13:10 viq And 0.17.3 minions talking to 2014.1.1 and 2014.1.0 masters
13:11 workingcats well i'm starting off wit 2014.1.1 on debian7.4 and current gentoo
13:11 engblom viq: Excactly with 2014.1.4 talking with 0.17.4 master did not work with those files I had earlier. I had to downgrade the minion to 0.17.x...
13:11 djaime joined #salt
13:12 viq engblom: I do _very_ little with my 0.17.4 master, pretty much just remote execution
13:12 workingcats but there's some.. more mature... servers where probably nobody made 2014.1 packages for ;)
13:12 viq engblom: if possible it's recommended to have master newer than minions, can you upgrade your master to 2014.1.1 ?
13:12 engblom No... not easily, but it is working as I got it now.
13:12 viq workingcats: the only one that comes to mind is OpenBSD and I'm trying to work on that ;)
13:13 workingcats viq, well i was more talking out-of-support debians
13:14 workingcats but before i dare "reinstalling" them i want to unify them as much as possible with their designated successor. plus then i'll have the successor fully salted straight away
13:15 joehh workingcats: lenny? or earlier?
13:15 joehh key requirement is python2.6
13:15 babilen workingcats: Do you really expect recent salt backports for lenny and older or did I misunderstand?
13:15 viq workingcats: you're running master on an old debian?
13:16 vbabiy joined #salt
13:16 workingcats viq, no, master was setup today on deb7.4
13:16 viq workingcats: then you do have master newer than minions :P
13:16 workingcats joehh, mostly lenny. i'll see how i deal with it system by system, but i'll start with the new deb6/7 ones first to get used to salt
13:17 workingcats babilen, i am hoping to, but i dont expect to ;)
13:17 joehh I've got a fair bit of incentive to get it going on lenny (number of old systems not able to upgrade), but have never
13:17 joehh gotten over the python2.6 hurdle
13:17 babilen Well, lenny really isn't supported anymore and I would *strongly* recommend upgrading all boxes to squeeze (or better wheezy unless we actually get a squeeze LTS team)
13:18 workingcats well salt on those servers would only be a temporary measure until i can replace them with current setups (which will then be kept current)
13:18 babilen Ah yeah ... that wonderful day in the future when everything will work fine.
13:19 workingcats exactly ;)
13:19 workingcats when i move to should-land
13:22 engblom file.recurse seem to not take care of old deleted files
13:22 workingcats well, considering what the lower bound on the length of "temporary" is.. babilen i'll try to remember to ping you if i find a way to get it to work on <=deb5
13:22 workingcats lenny is 5 right?
13:22 babilen yeah
13:23 racooper joined #salt
13:25 viq engblom: have you seen 'clean' parameter to it?
13:25 faldridge joined #salt
13:25 viq http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.recurse
13:27 babilen viq: Wonderful! I should have double checked the documentation I linked earlier
13:28 viq babilen: :P
13:28 engblom viq: Too many functions declared in state "file" in sls files
13:28 engblom That I get when I add - clean
13:29 babilen - clean: True?
13:29 vejdmn joined #salt
13:29 babilen engblom: Could you paste what you have?
13:29 babilen brb
13:30 engblom babilen: Thanks! that worked!
13:30 babilen engblom: It would otherwise be interpreted as "file.clean"
13:30 engblom I did not know I had to add True
13:31 MTecknology viq: did you ever look at gitlab-ci? I haven't yet.
13:31 viq MTecknology: no, I haven't even had time/energy to finish properly the gitlab formula...
13:31 babilen engblom: I typically use "file: \n - recurse \n - name: .... \n - ..." to make this more obvious. The "file.recurse" is just syntactic sugar isn't it?
13:33 MTecknology I've been deep into writing python applications for windows and packaging them for distribution. :(
13:37 viq MTecknology: eh, I'm not even sure where my time goes...
13:39 sroegner__ joined #salt
13:42 viq since there seems to be more people awake now, maybe someone will know how to approach this? https://groups.google.com/forum/#!topic/salt-users/IwtntfbxoJA
13:44 kermit joined #salt
13:44 anuvrat joined #salt
13:44 mpanetta joined #salt
13:47 joehoyle joined #salt
13:49 davet joined #salt
13:50 joehoyle_ joined #salt
13:53 joehh packages for ubuntu precise and lucid copying over to salt ppa now
13:54 kzx joined #salt
13:55 Linz joined #salt
14:01 babilen dput(-ng) them all!
14:02 ndrei joined #salt
14:03 Darnoth joined #salt
14:03 GradysGhost joined #salt
14:06 arrose joined #salt
14:08 joehh babilen: the others need some care - something changed in python-support so .pyc files are not handled correctly in newer releases
14:08 ggoZ1 joined #salt
14:10 joehh tossing up between hand coding pre/post inst/rm scripts and upgrading to dh_python2
14:10 joehh think I will upgrade
14:12 kaptk2 joined #salt
14:12 jeremyBass1 joined #salt
14:13 toastedpenguin joined #salt
14:14 Linz joined #salt
14:17 Linz_ joined #salt
14:18 srage joined #salt
14:19 Linz_ joined #salt
14:21 pydanny joined #salt
14:23 Kenzor joined #salt
14:27 ajprog_laptop joined #salt
14:30 opapo joined #salt
14:31 anuvrat joined #salt
14:31 ajprog_laptop1 joined #salt
14:32 pass_by_value joined #salt
14:32 wendall911 joined #salt
14:32 seapasulli joined #salt
14:39 abe_music joined #salt
14:39 joehoyle joined #salt
14:45 seapasulli joined #salt
14:45 derelm joined #salt
14:50 srage joined #salt
14:52 scoates Turns out I'll be at PyCon after all. Looking forward to meeting some of you there.
14:54 chitown_ joined #salt
14:54 timoguin joined #salt
14:54 dwiden joined #salt
14:58 AviMarcus joined #salt
14:58 alunduil joined #salt
14:58 dwiden Hi all.  Does anyone happen to know if salty vagrant plays nicely with windows minions?
14:59 timoguin dwiden, it works if you use the vagrant-windows plugin
14:59 timoguin well, kind of.
14:59 timoguin what i did was use a script for provisioning. it's a powershell script that silently installs the salt-minion installer
15:00 dwiden Hmm, that is actually what I need it to do
15:01 dwiden unfortunately I have to make my own vagrant windows box first
15:01 timoguin well
15:01 timoguin http://vagrantbox.msopentech.com/
15:01 timoguin that's an already-made evaluation box
15:01 conan_the_destro joined #salt
15:02 dwiden thank you, that will be useful to test with before I need to make my own
15:04 dwiden how did you handle giving your minions unique ids?  Or did you only have one minion?
15:04 easylancer joined #salt
15:05 thedodd joined #salt
15:05 timoguin Each minion gets its own folder and Vagrantfile, with a minion id assigned in a minion config file
15:06 joehoyle joined #salt
15:06 dwiden that makes a lot of sense.  I was trying to figure out how to do it with a single folder/Vagrantfile
15:06 dwiden But that might be really messy
15:10 TheRealBill joined #salt
15:10 rawzone joined #salt
15:11 GradysGhost joined #salt
15:13 dwiden timoguin: I'm fairly new to using vagrant, and all of the guides focus on really simple deployments.  Is it typical to have a separate directory/Vagrantfile for each machine that you spin up?  I know that this isn't strictly a salt question, but it would really help out or deployment
15:15 che-arne joined #salt
15:16 viq dwiden: no, I'd say the power of vagrant is being able to spin up multi-machine environments
15:16 viq dwiden: two examples: mine https://github.com/viq/cm-lab-salt and someone's much smarter than me ;) https://github.com/ripienaar/mcollective-vagrant
15:17 scott_walton joined #salt
15:17 timoguin dwiden, I use single-instance setups to test platform-specific formulas, but I'm working on some multi-machine setups for integration tests
15:17 timoguin but, yea, multi-machine is pretty damn useful and powerful
15:18 dwiden thank you both very much, it looks like I should be able to do what I need to do while minimizing how hackish it gets
15:18 smurfy__ joined #salt
15:19 ipmb joined #salt
15:20 baniir joined #salt
15:21 davidone so, to bootstrap a new server cobbler is a more suitable tool?
15:21 viq davidone: depends what you mean by "bootstrap"
15:21 viq davidone: if "take from bare metal to running OS" then yeah
15:22 sandbender1512 joined #salt
15:22 baniir is it possible to include environment vars to the pkg.installed state
15:22 davidone yes, I mean: I have a new server, I boot it with a livecd/pxe iso and salt (or another tool) provides disk formatting and so on
15:22 viq davidone: also I think it was me who suggested cobbler on the list ;)
15:22 davidone viq: yes, I was looking at that thread :)
15:22 timoguin there's an external pillar module for cobbler as well
15:23 xmj gnihihi
15:23 viq davidone: and yeah, probably salt can do that, but I'd say either look at cobbler, or if you want higher level and more fancy tools look up foreman or puppet razor
15:23 davidone well, if I can do the complete task with salt is ok, if I need salt+cobbler for a task where only cobbler is necessary I think I'll take a look at cobbler
15:24 davidone cobbler seems like a 'redhat' tool
15:24 davidone anyway, I'll ask in cobbler chan
15:25 timoguin i think the developer of cobbler worked for redhat
15:26 davidone that's why I don't see any cobbler for freebsd :)
15:26 jalbretsen joined #salt
15:27 viq davidone: one of first hits ;) http://blog.hostileadmin.com/2013/04/11/installing-freebsd-via-cobbler/
15:27 kballou joined #salt
15:28 viq davidone: also I see here some WIP port for OpenBSD https://github.com/jasperla/openbsd-wip/tree/master/www/py-cobbler
15:29 davidone I know I'm completely off-topic but I'm reading https://fedorahosted.org/cobbler/wiki/SupportForOtherDistros and their support seems far from being "production ready" :)
15:29 conan_the_destro joined #salt
15:30 xmj viq: when i wrote 'gnihihi' I *knew* you'd post something about obsd within the next 5min.
15:30 xmj viq: Thank you. :)
15:31 viq xmj: you were off by 30 seconds :P
15:31 xmj viq: ever seen the movie Primer ?
15:32 viq Doesn't ring a bell, no
15:32 chrisjones joined #salt
15:34 HeadAIX joined #salt
15:34 viq davidone: I think generally FreeBSD doesn't quite have native support for unattended installation
15:35 viq davidone: though I see things like http://lists.freebsd.org/pipermail/freebsd-hackers/2012-November/041162.html
15:35 davidone viq: probably a custom livecd is what I need
15:35 davidone that's why I am interested in this thread
15:35 xmj viq: the post is deprecated
15:36 xmj viq: with NanoBSD and bsdinstall unattended installations are entirely possible
15:36 viq And I know OpenBSD is working on getting there as well, but I don't really know where FreeBSD stands in this
15:36 davidone xmj: yes but is not something you can control from a master node
15:36 viq xmj: yeah, trying now to find something more recent than 2012
15:36 ndrei joined #salt
15:37 xmj davidone: of course you can (though it'll be a PITA to set up the first time)
15:39 xmj build a custom image that comes with base, kernel, ports-mgmt/pkg and sysutils/py-salt  (+ configuration for IP/DNS/... and the salt-minion) along with a pre-seeded key for the minion on some storage that can be network-mounted
15:40 viq Or yeah, what xmj says
15:40 xmj then install everything else via salt stats
15:40 xmj states even
15:41 viq When I think about it it actually makes some sense - boot from PXE an image, and highstate/whatever will install the system, then you reboot and take from there
15:41 xmj viq: reboot? c'mon, it's not ubuntu ;)
15:41 viq xmj: what, you're going to run off of netbooted image forever?
15:42 davidone I was thinking a custom image (with freesbie, nanobsd or similar)
15:43 xmj viq: you're the openbsd guy
15:43 davidone that acts like a livecd where I install py27-salt
15:43 xmj viq: standard obsd advice: "just pxeboot bsd.rd and be done with it."
15:43 davidone so I can have a salt-master to use with *-boostrap agent
15:43 KyleG joined #salt
15:43 KyleG joined #salt
15:43 davidone and then retrieving information from agent to decide what to do
15:44 viq xmj: well, I'm not necessarily a stereotypical obsd guy - I'm right now preparing a yaifo image to upgrade a box because I can't be bothered to walk to the next building :P
15:44 xmj viq: tongue in cheek ;)
15:45 viq xmj: just make sure you don't bite before you remove it from there ;)
15:45 davidone :)
15:46 davidone xmj: I'm noticing now that you are in #freebsd, too :)
15:47 xmj davidone: i use 11.0 on my laptop, single boot
15:47 davidone well, production ready :)
15:48 xmj not so sure about that, haha
15:48 viq xmj: btw, you played with bhyve?
15:49 xmj viq: yes
15:49 xmj i'm actually saltifying a centos/openvz based system on it
15:50 viq nice
15:50 viq What are your experiences with it, and on what hardware?
15:51 toastedpenguin joined #salt
15:51 viq mainly pondering if it would run on http://pastie.org/9001008
15:52 srijan4 joined #salt
15:52 workingcats having a little, probably very silly, problem.. just set up my salt, switched to using the init scripts to start master&minions, and have set all minions to default to test. it worked before when i had started masters&minions manually, but now "salt '*' state.highstate test=False" no longer applies the changes
15:52 workingcats any pointers?
15:53 viq workingcats: "salt \* test.ping" do they respond?
15:53 workingcats viq, yep
15:54 workingcats and the command i pasted does show all the differences (and successes)
15:54 workingcats one sec
15:55 workingcats back
15:55 Gordonz joined #salt
15:56 workingcats come to think of it, i may not have restarted the minions after changing their setting to default to test=True
15:56 xmj viq: Asus UX32VD 13.3" ZenBook
15:57 xmj viq: AMD isn't supported even on 11.0-CURRENT yet (there's one bug needing to be fixed before it can be merged)
15:57 viq xmj: ah, thanks
15:57 viq xmj: I guess I'll try running logstash in jails then ;)
15:58 xmj viq: that being said, you could run it if you use the project branch  http://svnweb.freebsd.org/base/projects/bhyve_svm/  i think
15:58 workingcats oh i think i see my problem
15:58 xmj viq: just takes a little... src/ hackery.
15:58 chitown_ is there a way to disable a grain module from running?
15:58 chitown_ specifically, external_ip.py
15:59 viq xmj: eh, not sure if I am that desperate ;)
16:00 workingcats hrm no i dont see my problem
16:00 workingcats my whole state config (not long): http://pastebin.com/JSDnKskc
16:01 workingcats i thought adding lines 57+58 would fix it
16:03 viq fix what?
16:03 workingcats it's not changing the system to match the defined state
16:04 [diecast] joined #salt
16:04 viq What is it (not) doing?
16:04 johtso joined #salt
16:05 workingcats e.g. i changed /srv/salt/bashrc on the master, and it is showing me how the minions differ, but it is not changing the minions to match the new file contents
16:05 workingcats tried "salt '*' state.highstate test=False" on master and "salt-call state.highstate test=False" on minion
16:06 workingcats same for ntp.conf
16:07 workingcats by changing the minions i mean editing/creating those files
16:09 Xe joined #salt
16:10 glenbot joined #salt
16:13 gildegoma joined #salt
16:13 workingcats i just can't see what I'm doing wrong here, http://docs.saltstack.com/en/latest/ref/states/testing.html#default-test is quite clear. only changes i made in the minion conf is to type in master's IP and set test to True
16:14 hunter joined #salt
16:16 jeremyfelt joined #salt
16:17 vilelm joined #salt
16:18 meteorfox joined #salt
16:20 davidone see you
16:25 Linz joined #salt
16:27 Gordonz joined #salt
16:30 _mel_ joined #salt
16:31 _mel_ Hi. i like to update a config and only restart a service, if the service is running already. if its stopped it should stay stopped. can i do this?
16:35 _mel_ if i don't set "- running" i got the error: "No function declared in state "service" in sls nginx"
16:39 jmreicha joined #salt
16:44 Gareth_ morning.
16:44 kermit joined #salt
16:44 ThomasJ|d joined #salt
16:46 elfixit joined #salt
16:47 zeroC joined #salt
16:49 Linz joined #salt
16:49 mgw joined #salt
16:54 ndrei joined #salt
16:56 redondos joined #salt
16:57 manfred _mel_: you need to specify - running, or you need to use service.running:
16:58 manfred that is what declares which function to use from https://github.com/saltstack/salt/blob/develop/salt/states/service.py#L256
16:59 manfred _mel_: you should be able to specify - reload: True, https://github.com/saltstack/salt/blob/develop/salt/states/service.py#L30
16:59 manfred which will reload the service
17:02 manfred _mel_: i prefer to use service.mod_watch for some of those, because it allows me to do a full restart
17:06 Employee312 joined #salt
17:07 Employee312 joined #salt
17:07 Ryan_Lane joined #salt
17:11 travisfischer joined #salt
17:15 bhosmer joined #salt
17:18 hunter joined #salt
17:22 joehoyle joined #salt
17:23 faldridge joined #salt
17:23 timoguin joined #salt
17:24 Kenzor joined #salt
17:24 workingcats well i'm off home but if anyone has a pointer for my presumably incredibly idiotic problem highlight me pls ;)
17:26 hunter joined #salt
17:27 ajw0100 joined #salt
17:29 joehoyle joined #salt
17:30 bhosmer joined #salt
17:36 hunter joined #salt
17:37 ggoZ joined #salt
17:39 wkf joined #salt
17:40 glenbot joined #salt
17:41 druonysus joined #salt
17:42 yomilk joined #salt
17:42 sgviking joined #salt
17:43 ndrei joined #salt
17:46 Kenzor joined #salt
17:47 themadcanudist joined #salt
17:47 themadcanudist hey guys. are there any good docs on how the salt pki system works in detail?
17:51 jcockhren I'm open to suggestions: https://github.com/sophicware/benderbot/issues/3
17:51 pydanny joined #salt
17:53 ndrei joined #salt
17:58 ipalreadytaken joined #salt
17:59 GradysGhost joined #salt
17:59 arrose joined #salt
17:59 crazysim joined #salt
18:01 arthabaska joined #salt
18:03 hunter joined #salt
18:08 ipmb joined #salt
18:12 hunter joined #salt
18:15 redondos joined #salt
18:17 yomilk joined #salt
18:17 catpig joined #salt
18:19 smcquay joined #salt
18:21 GradysGhost joined #salt
18:23 easylancer joined #salt
18:23 Luijk left #salt
18:24 jY salt-call event.fire_master '{"data": "message for the master"}' 'tag'
18:24 jY what is the tag for
18:24 jY ?
18:27 timoguin the tag is what the reactor matches events on. salt has internal events that are named space like salt/foo
18:27 timoguin you could call yours customtag/foo/whatever
18:27 timoguin and match based on that
18:28 jY ok thanks
18:28 jY clears it up
18:33 Mr_N So two problems.
18:33 Mr_N It seems the salt bootstrap script is sporadically REALLY not ok with SL machines
18:33 Mr_N as is documented in many of the related issues to https://github.com/saltstack/salt-bootstrap/issues/343
18:34 Mr_N and is claimed to be "closed" but I can still reproduce it.
18:34 Mr_N Two, is it silly of me to request that 17.5 remain in yum?  (or am I blind, and it is, because I can't seem to find it.)
18:35 Corey Mr_N: That's an EPEL thing.
18:35 Mr_N They only allow for the latest version or something?
18:35 themadcanudist left #salt
18:39 Employee_ joined #salt
18:43 Mr_N Yeah; even the develop version fails to install on my SL62
18:43 sandbender1512 basepi: here?
18:43 Mr_N Yay, time to add more to my pile of unresolved issues...
18:45 toastedpenguin joined #salt
18:47 smcquay joined #salt
18:49 big_area joined #salt
18:51 big_area is there a way to uninstall a package its version is >x?
18:53 kermit joined #salt
18:55 Gordonz joined #salt
18:56 Gordonz joined #salt
18:58 mgw joined #salt
18:59 ipalreadytaken joined #salt
19:00 bhosmer joined #salt
19:02 TheRealBill_here joined #salt
19:03 thedodd joined #salt
19:06 pydanny joined #salt
19:06 basepi sandbender1512: am now
19:07 sandbender1512 nm sorta figured out my question already :)
19:08 basepi sandbender1512: love it when that's the answer!  xD
19:08 sandbender1512 side note: hopefully I'll see you at pycon
19:08 sandbender1512 haha
19:08 basepi sandbender1512: nice!  come see us at the booth!
19:08 sandbender1512 yeah will do
19:08 sandbender1512 :)
19:09 kermit joined #salt
19:13 Gareth basepi enters and all of a sudden all the questions are answered!
19:13 basepi Mwahahaha.  The servers fear my wrath.
19:16 joehoyle joined #salt
19:19 rager joined #salt
19:20 vbabiy joined #salt
19:24 armonge joined #salt
19:31 thedodd joined #salt
19:32 joehoyle_ joined #salt
19:33 thedodd joined #salt
19:36 viod joined #salt
19:37 rager joined #salt
19:38 doddstack joined #salt
19:38 UtahDave joined #salt
19:39 catpig oh you answer all questions? ;)
19:42 catpig i'm still confused by why my salt minions dont apply the state when i tell them to. brand new install, master is 2004.1.1 on debian 7.4, minions debian 7.x and a gentoo, all 2004.1.1. i've set test: True on all minions (which includes the master), but override it with: salt '*' state.highstate test=False
19:42 catpig obviously i'm making some mistake in telling it, i just can't find what
19:42 krow joined #salt
19:43 Guest50566 joined #salt
19:47 toastedpenguin joined #salt
19:48 viod joined #salt
19:50 Kenzor joined #salt
19:56 pydanny joined #salt
20:02 nebuchad` joined #salt
20:03 jamesog_ joined #salt
20:03 gwmngilfen|afk joined #salt
20:03 aurigus_ joined #salt
20:04 pydanny joined #salt
20:04 ninkotech_ joined #salt
20:04 funzo joined #salt
20:04 smurfy joined #salt
20:04 steveoliver_ joined #salt
20:04 fedgoat joined #salt
20:05 ldleworker joined #salt
20:05 Mr_N Hmmm.
20:05 Mr_N Is it intentional that in 14.1
20:05 tmmt_ joined #salt
20:05 Mr_N For a SL 6.2 minion;  os: Scientific
20:05 jillr_ joined #salt
20:05 hotbox_ joined #salt
20:05 rogst_ joined #salt
20:05 bdf_ joined #salt
20:05 Mr_Magic_ joined #salt
20:05 Mr_N (it used to be ScientificLinux)
20:06 sulky_ joined #salt
20:06 mikkn_ joined #salt
20:06 Ymage_ joined #salt
20:07 harobed joined #salt
20:09 bwq- joined #salt
20:09 ThomasJ|o joined #salt
20:09 aarontc_ joined #salt
20:09 basepi_ joined #salt
20:09 kaiyou_ joined #salt
20:09 EugeneKay joined #salt
20:12 alunduil joined #salt
20:16 n0c joined #salt
20:16 Jarus joined #salt
20:16 Ixan joined #salt
20:17 Jarus joined #salt
20:19 yomilk joined #salt
20:19 diegows joined #salt
20:21 funzo joined #salt
20:26 Gordonz_ joined #salt
20:27 jslatts hey salt folks, what is the right terminology for discussing a salt master and its collection of minions?
20:28 jslatts like: a "pool" of minions? a "cluster" of minions? a "gaggle" of minions?
20:28 xmj i always wondered. stalactite versus stalagmites?
20:28 jslatts I'm writing a blog post and I don't really have a satisfactory answer
20:29 dave_den joined #salt
20:29 smurfy_ joined #salt
20:30 jcockhren jslatts: I think you had a it right. "coillection of minions"
20:30 jcockhren collection*
20:30 jslatts "when the new VM comes up, it will automatically join the master's collection of minions"
20:30 jslatts i guess that works
20:31 jcockhren yeah. b/c there can be sub-collections of minions (see syndic setups)
20:31 jslatts i'm actually writing about syndics
20:32 jcockhren well... there it is
20:32 jslatts we use syndics to facilitate autoscaling
20:32 jcockhren \o/
20:32 Gordonz joined #salt
20:32 jcockhren ohhh! nice
20:32 Ryan_Lane jslatts: oh? how are you doing that?
20:32 jslatts just really hard to phrase it right
20:32 jcockhren jslatts: link that blog post when done. :)
20:32 Ryan_Lane yeah, for sure. I'm interested in that as well
20:33 jslatts Ryan_Lane: we have syndics that are set to promiscuous mode and we bake in the syndic master into our AWS ami
20:33 Gordonz joined #salt
20:33 jslatts when a new AMI comes up, it joins the local syndic and does an auto-highstate
20:33 jslatts which then pulls the latest build from S3 and self-deploys our app
20:33 jslatts as well as pick up and config updates
20:33 pydanny joined #salt
20:33 Ryan_Lane how is that different than putting the master into autosign mode?
20:34 jslatts well, the syndics are only available inside the VPCs
20:34 sulky joined #salt
20:34 Ryan_Lane ah. I see
20:34 jslatts not directly accessible, so we can then manage multiple AWS regions with one master
20:34 jcockhren jslatts: that's slick
20:34 jslatts but the master of masters is not set to autosign
20:34 jslatts so its for security
20:35 jslatts yeah, it actually works well. we can then redeploy the running VMs any time. as well as change or run config queries
20:35 jslatts use grains for all targetting
20:35 [diecast] joined #salt
20:36 jslatts the only real issue I have is that the master of masters -> syndic responses are somewhat broken. We will often have no results comes back to the MoM even though the states ran
20:42 oleg__ joined #salt
20:43 Luke___ joined #salt
20:43 oleg__ Hi guys,
20:43 oleg__ I am writing formula for djbdns (dnscache right now) and have some problem
20:43 oleg__ I have pillar file like this: http://pastebin.com/FtwL1HMY
20:43 oleg__ Now I should read clients_files files, merge names from these files together with clients... and....
20:43 oleg__ 1) remove from some directory all files which not preset in any clients (in list 'clients' or files)
20:43 oleg__ 2) create files (touch) which present in any clients (in list 'clients' or files)
20:43 oleg__ How I can do this? Easy way :)
20:44 jslatts oleg__: you want to iterate through each IP?
20:44 jslatts or concat the list with your pillar?
20:45 oleg__ Step-by-step, ok?
20:46 oleg__ 1. Iterate over list and create file for each value. I know how to do it ({% for c in clients %}.... {%endfor %})
20:46 oleg__ 2. Scan directory and remove all files which not present in some list 'clients'. How?
20:47 oleg__ jslatts: ^
20:47 jslatts can you clear the entire directory then re-create the valid files?
20:47 higgs001 joined #salt
20:47 oleg__ jslatts: on every state applying?
20:47 jslatts yeah, thats not good
20:47 oleg__ jslatts: ugly solution, indeed. Daemon would be affected :(
20:48 oleg__ jslatts: Do you know some clear solution?
20:48 jslatts thinking, but nothing off the top of my head. I would probably try to right a python state definition and do it in code
20:48 jslatts s/right/write
20:49 jslatts oleg__: http://docs.saltstack.com/en/latest/ref/renderers/index.html
20:50 jslatts you could enumerate the directory and then do the compare yourself. not simple, but certainly doable
20:50 krow joined #salt
20:50 oleg__ jslatts: It's ok. Only single concern: do we have way to distribute custom state over formulas?
20:50 jslatts generate removes/creates as needed
20:50 oleg__ >>> you could enumerate the directory
20:50 oleg__ how?
20:51 jslatts ugh, you are right. you need that to run on the minion itself
20:51 higgs001_ joined #salt
20:52 oleg__ jslatts: yeap :)
20:52 oleg__ jslatts: so, any way to distribute custom state over formula?
20:53 jslatts custom SLS?
20:54 yomilk joined #salt
20:55 pydanny joined #salt
20:55 oleg__ jslatts: python state :)
20:55 oleg__ jslatts: custom... Mmmm... Renderer? I do not know exactly ;)
20:55 jslatts oleg__: yes, you just include it in your state tree
20:56 oleg__ jslatts: cool! can you show me some example for github? :)
20:56 jslatts oleg__: i think it could work actually, because the minion runs the state, but someone smarter than me would have to chime in. or you could try it :)
20:57 jslatts sure, illl show you a custom one i wrote
20:57 ckao joined #salt
20:58 * oleg__ is waiting
20:58 jslatts oleg__: https://gist.github.com/jslatts/10053524
20:59 jslatts looks for any pillar with user_add_foo and adds those users to the server
20:59 oleg__ jslatts: >>> #!py
20:59 oleg__ jslatts: very interesting shobang
20:59 oleg__ jslatts: can you refer to some documentation how to write scripts like this?
21:00 catpig jslatts, i like gaggle of minions
21:01 jslatts oleg__: I had to figure it out with trial and error. not many docs. http://docs.saltstack.com/en/latest/ref/states/highstate.html
21:01 joehoyle joined #salt
21:02 oleg__ jslatts: thank you very much
21:02 jslatts oleg__: anyway, if I was feeling frisky, I would try to write code that looks at the directory structure and creates the correct high state output to make it conform to your list
21:03 jslatts oleg__: i would think it would work because the minion itself is executing the state compilation, but thats me guessing
21:03 oleg__ jslatts: https://github.com/z-saltstack/djbdns-formula/tree/master/djbdns
21:03 jslatts catpig: I'll send a PR for the official docs to add Gaggle in :)
21:04 oleg__ jslatts: I am trying to write dnscache_config.py which would read pillar dict to input and fix directory layout according to pillar settings
21:04 jslatts oleg__: yes, keeping a list of "users to remove" is how I currently handle this same issue you have
21:04 Ryan_Lane1 joined #salt
21:05 oleg__ jslatts: ya-ya. Thanks :)
21:06 colinbits joined #salt
21:08 bhosmer joined #salt
21:11 skogen joined #salt
21:11 krichardson joined #salt
21:15 pydanny joined #salt
21:15 oleg__ jslatts: can you show me example how did you use you script?
21:16 oleg__ I mean, in anothe sls file
21:17 hunter joined #salt
21:19 ianawilson joined #salt
21:22 hunter joined #salt
21:22 wendall911 joined #salt
21:26 ianawilson Hi all. How do I require another state declaration? I see that I need a single key dictionary, but I'm not sure how to say that state "foo" is required for state "bar".
21:27 toastedpenguin joined #salt
21:29 oleg__ foo:
21:29 oleg__ - require: bar
21:29 jslatts oleg__: that script is just included in the top file
21:29 oleg__ http://docs.saltstack.com/en/latest/ref/states/ordering.html
21:30 oleg__ use 'include' directive
21:30 dvogt joined #salt
21:30 oleg__ and 'require' after that
21:30 oleg__ or 'require_in' inside bar
21:30 oleg__ bar:
21:30 oleg__ - require_in: foo
21:30 ianawilson oleg__: so, it's not an entire other state file, but one of the state declarations (sorry if these are the wrong terms -- very new) inside an .sls file
21:31 jslatts ianawilson: you can use the include function to include states outside of that file if you need to
21:31 ianawilson jslatts: this is all in one file
21:31 ianawilson let me make an example
21:32 ianawilson http://dpaste.com/1773320/
21:32 majoh joined #salt
21:33 ianawilson these are in one sls file, and i'd like bar to require that foo was successful
21:33 ianawilson (in my specific example, foo is a cmd.run)
21:33 ianawilson *specific case
21:33 pydanny joined #salt
21:34 jslatts you have to be more specific. if foo is a file.present
21:34 dvogt Does anyone know, If I update the /etc/salt/master.d/reactor.conf do I need restart the Salt master ?
21:34 jslatts then require needs to indicate file: foo
21:34 ianawilson jslatts: so, foo is a cmd.run, and i don't know how to require that it was run
21:34 ianawilson can i say "cmd.run: foo" or something in my require list?
21:35 dfinn joined #salt
21:36 dfinn our company is huge into graphs and "metrics" lately.  I'm being asked to try and quantify and graph some of the salt stuff too.  an example might be how long it takes to run a highstate on a minion.  I'm wondering if anyone is doing anything like this currently?
21:36 jslatts ianawilson: http://pastebin.com/6J1GrY8q
21:36 jslatts something like that
21:37 ianawilson jslatts: mm, i'll try it
21:38 jslatts ianawilson: try and avoid doing too much with cmd.run. the only way salt knows if it worked/failed is based on shell exist codes
21:38 jslatts ianawilson: which can lie. It also may or may not be idempotent
21:38 ianawilson jslatts: yeah
21:39 ianawilson jslatts: i'm using it because i need to compile a specific version of something that is not available in an apt repository. is there a better way to deal with compiling code like that?
21:40 jslatts ianawilson: not really. thats about the only placed I have used it as well. only better option would be to precompile it and distribute the binary instead
21:40 jslatts ianawilson: which doesn't help much if you have a variety of environments
21:40 ianawilson jslatts: mmmm
21:40 ianawilson gotcha
21:40 ianawilson that's a good option to keep in mind, though
21:40 krow joined #salt
21:40 pydanny joined #salt
21:41 ianawilson jslatts: thanks a lot for the help! using require: - cmd: foo worked!
21:41 jslatts ianawilson: its just so slow to compile. you may find salt timing out the returners
21:41 ianawilson mmm, that's good to know too
21:41 jslatts np, good lucky
21:41 jslatts luck. sheesh. can't type today
21:46 krichardson been trying to get reactor to fire only if there are changes in highstate (https://gist.github.com/khrisrichardson/10063575). for some reason the for loop causes a failure to render the reactor formula. without it, everything is peachy. any ideas? checked output from halite, eventlisten.py, and `salt-master -l debug'. no dice.
21:53 smurfy_ joined #salt
21:53 cewood joined #salt
21:57 pydanny joined #salt
21:59 thehaven joined #salt
21:59 sinh joined #salt
22:01 yomilk joined #salt
22:03 joehoyle joined #salt
22:06 kermit joined #salt
22:07 snave joined #salt
22:14 eliasp joined #salt
22:16 UtahDave joined #salt
22:16 chrisjones joined #salt
22:16 Gareth uh oh.
22:20 catpig okay, tried again, i commented the "test: True " in /etc/salt/minion again, and now "salt '*' state.highstate test=False" (and True) behave as expected. but when i uncomment the line in the config then it does test=True whether i say False on the CLI or not
22:20 catpig any ideas? i'd prefer to have default with test on
22:25 krow joined #salt
22:25 ipalreadytaken joined #salt
22:26 nextdoorwarren joined #salt
22:27 hunter joined #salt
22:28 hunter joined #salt
22:32 smurfy_ joined #salt
22:44 thayne joined #salt
22:47 djinni` joined #salt
22:49 ivan joined #salt
22:57 meteorfox joined #salt
23:05 UtahDave1 joined #salt
23:05 Gareth two UtahDave's.
23:06 greghaynes joined #salt
23:06 UtahDave Gareth: muhahahaahahahahaha
23:10 yomilk joined #salt
23:12 elfixit joined #salt
23:21 jslatts joined #salt
23:31 meteorfox joined #salt
23:31 ajprog_laptop joined #salt
23:33 sroegner__ joined #salt
23:34 druonysuse joined #salt
23:34 druonysuse joined #salt
23:42 krichardson joined #salt
23:53 hunter joined #salt
23:55 garthk joined #salt
23:55 krow joined #salt
23:57 Ryan_Lane1 joined #salt

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