Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-04-16

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

All times shown according to UTC.

Time Nick Message
00:02 napper joined #salt
00:03 UtahDave kickerdog: Yeah, we had a bunch of people at pycon.
00:03 UtahDave unfortunately I wasn't able to go. customer onsite
00:03 Eugene Cursed customers and their needs
00:04 UtahDave :)  luckily they're really awesome.  so not bad.
00:06 jcockhren Katafalk_: Hey. did you ever figure it out?
00:07 kickerdog My CTO said he saw you guys. I miss all the old Beyonders.
00:11 UtahDave kickerdog: ah, did you work with Tom and Joseph?
00:19 kickerdog yeah, back in the day
00:19 UtahDave cool
00:20 jcockhren Katafalk_: tag me in a github issue and I'll look at it in a few
00:22 rome joined #salt
00:28 rome joined #salt
00:31 Leech joined #salt
00:33 ipalreadytaken joined #salt
00:35 fragamus joined #salt
00:36 manfred Eugene: is it bad that I gave myself a papercut trying to open the box?
00:39 zain_ joined #salt
00:41 redondos joined #salt
00:52 bhosmer joined #salt
00:54 TyrfingMjolnir joined #salt
00:56 toastedpenguin joined #salt
01:01 Thart joined #salt
01:04 rome joined #salt
01:05 napper joined #salt
01:16 quickdry21 joined #salt
01:17 rome joined #salt
01:18 TyrfingMjolnir joined #salt
01:24 jvasallo joined #salt
01:45 rome 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/
01:50 jvasallo1 joined #salt
01:53 mgw joined #salt
01:56 snuffeluffegus joined #salt
02:03 kickerdog joined #salt
02:04 TyrfingMjolnir joined #salt
02:09 ajw0100 joined #salt
02:10 ckao joined #salt
02:16 srage joined #salt
02:19 srage_ joined #salt
02:20 taion809 joined #salt
02:28 ekristen joined #salt
02:28 rushm0r3 joined #salt
02:31 TyrfingMjolnir joined #salt
02:35 connie joined #salt
02:39 possibilities joined #salt
02:51 travisfischer joined #salt
02:52 ZombieFeynman joined #salt
02:55 Ryan_Lane joined #salt
02:56 ldlework joined #salt
02:57 mgw joined #salt
02:59 MasterNayru joined #salt
03:08 ixokai- joined #salt
03:16 possibilities joined #salt
03:17 travisfischer joined #salt
03:18 Katafalkas joined #salt
03:18 jeremyfelt joined #salt
03:18 nmistry joined #salt
03:18 icarus joined #salt
03:23 possibilities joined #salt
03:26 ravibhure joined #salt
03:33 bemehow joined #salt
03:36 anuvrat joined #salt
03:39 xl1 joined #salt
03:41 quickdry21 joined #salt
03:48 catpiggest joined #salt
03:51 napper joined #salt
03:51 Nexpro1 joined #salt
03:51 googolhash joined #salt
03:51 nmistry joined #salt
03:52 Leech what's the best way to reuse states in a state tree for different configurations?
03:53 ZombieFeynman joined #salt
03:54 meteorfox joined #salt
03:56 Hipikat joined #salt
04:03 philipsd6 joined #salt
04:10 martin__ joined #salt
04:10 kermit joined #salt
04:12 sgviking joined #salt
04:12 UtahDave joined #salt
04:19 thayne joined #salt
04:19 quickdry21 joined #salt
04:23 ajw0100 joined #salt
04:27 mikemar10 joined #salt
04:36 mafro joined #salt
04:36 mafro lo salters
04:36 mafro 2014.1.2 is failing for me on install
04:36 mafro File "/home/vagrant/.virtualenvs/boomerang/build/salt/setup.py", line 516, in <module>
04:36 mafro NameError: name 'zmq' is not defined
04:36 mafro others must have hit this one too
04:39 nmistry joined #salt
04:41 schimmy joined #salt
04:43 rushm0r3 joined #salt
04:44 mgw joined #salt
04:46 schimmy joined #salt
04:47 l0x3py joined #salt
04:51 CeBe joined #salt
04:52 Gordonz joined #salt
04:53 Gordonz joined #salt
05:04 srage joined #salt
05:07 UtahDave mafro: yeah, we've paused packaging while we iron out a couple things
05:07 mafro UtahDave: ah cool, I was just writing to the mailing list, but if you guys are on it...
05:07 UtahDave mafro: could you create an issue with the info you pasted here?  We want to get everything cleaned up quickly
05:08 mafro it’s a pretty easy fix here
05:08 mafro yeah sure
05:08 UtahDave 2014.1.2 hasn't been officially released. We just released it the packagers today when we found a couple bugs.
05:08 UtahDave thanks, mafro!
05:11 arthabaska joined #salt
05:11 mafro https://github.com/saltstack/salt/issues/12031
05:15 acabrera joined #salt
05:15 rushm0r3 joined #salt
05:25 xmj UtahDave: cedwards: cool! do you have an ETA when we can get a PR to FreeBSD ?
05:25 * xmj would really much like to be among the first OS that were to ship 2014.1.2
05:29 notimpossible joined #salt
05:31 it_dude joined #salt
05:32 engblom file.recurse seem to not copy any hidden file (dot-file). Is there anything I could do about this?
05:37 it_dude joined #salt
05:38 smcquay joined #salt
05:41 srage joined #salt
05:42 budrose joined #salt
05:43 anuvrat joined #salt
05:49 UtahDave xmj: cedwards is usually pretty quick to package
05:49 xmj I know
05:50 UtahDave engblom: Hm. I've never noticed that dot files don't get copied down.  I think that would be a bug
05:50 UtahDave what version of Salt are you on, engblom?
05:50 engblom UtahDave: 0.17.4
05:51 xmj UtahDave: that was more of a poking Him and make sure You see it too :-)
05:51 UtahDave xmj: :)
05:51 UtahDave engblom: can you test against a later version to see if it's still occuring?
05:51 engblom UtahDave: Not easily in this case...
05:52 UtahDave engblom: ok, let me see if I can test it here.
05:53 sroegner joined #salt
05:55 kickerdog joined #salt
05:58 UtahDave engblom: Hm.  It just worked for me.
06:00 ndrei joined #salt
06:01 xmj cedwards: had you seen ports/186397 ?
06:16 xmj UtahDave: do you know who does the EPEL pkging
06:16 xmj ?
06:16 UtahDave I think it's been terminalmage
06:16 flebel joined #salt
06:17 bhosmer joined #salt
06:18 xmj derp
06:19 fllr joined #salt
06:19 xmj UtahDave: is there a CHANGELOG or release announcement with #bugfixes and #newfeatures?
06:19 xmj I want to sell the upgrading as #relevent :)
06:20 _fllr_ joined #salt
06:21 UtahDave xmj: yeah, there's release notes for each release
06:23 mafro joined #salt
06:28 xmj UtahDave: mind sending those to salt-announce?
06:28 xmj (can't see them there yet)
06:28 engblom UtahDave: Did you also use the "clean" option?
06:29 UtahDave xmj: the release hasn't been announced yet.
06:29 UtahDave engblom: no.  can you pasetbin what you're using?
06:29 xmj UtahDave: ah!
06:31 engblom UtahDave: http://pastie.org/9084096
06:32 engblom UtahDave: .mozilla, .locale and .config are not copied by that paste
06:35 rushm0r3 joined #salt
06:36 UtahDave engblom: Yeah, it's working for me on the latest develop.  Must be a bug that has been fixed.
06:36 engblom UtahDave: Thanks for testing. I will have to see what I could do for upgrading
06:42 octarine joined #salt
06:43 UtahDave you're welcome
06:44 rewerlin joined #salt
06:44 nebuchadnezzar joined #salt
06:46 cele joined #salt
06:46 rewerlin Hi all, maybe somebody can help me. Ive got Oracle Linux ANd Salt. I see that the host is  on Accept keys. Im running "salt salt.intech-global test.ping" but its not publishing smth, in log:[INFO    ] Clear payload received with command publish [INFO    ] User root Published command test.ping with jid 20140416031401919097 [DEBUG   ] Published command details {'tgt_type': 'glob', 'jid': '20140416031401919097', 'tgt': 'salt.intech-glo
06:51 ilako joined #salt
06:52 Kenzor joined #salt
06:54 stubee joined #salt
06:59 it_dude joined #salt
07:01 UtahDave joined #salt
07:08 kedo39 joined #salt
07:10 harobed_ joined #salt
07:16 stubee joined #salt
07:17 engblom viq: Are you the one maintaining the OpenBSD port for salt? In what state is https://github.com/jasperla/openbsd-wip/tree/salt/sysutils/salt ? Is it working, but not enough tested?
07:18 chiui joined #salt
07:19 fllr joined #salt
07:28 malinoff joined #salt
07:32 nadley joined #salt
07:32 nadley hi all
07:36 southquist joined #salt
07:37 it_dude joined #salt
07:39 giantlock_ joined #salt
07:40 q1x joined #salt
07:40 malinoff Hi
07:42 sroegner joined #salt
07:44 stubee joined #salt
07:49 esogas_` joined #salt
07:53 Shenril joined #salt
07:56 anuvrat joined #salt
08:01 stubee joined #salt
08:05 kadel joined #salt
08:07 jalaziz joined #salt
08:09 giantlock joined #salt
08:10 smcquay joined #salt
08:10 copasa joined #salt
08:12 kuffs joined #salt
08:13 rome joined #salt
08:19 fllr joined #salt
08:22 chiui_ joined #salt
08:33 che-arne joined #salt
08:40 z3uS joined #salt
08:40 twobitsprite joined #salt
08:41 jacksontj joined #salt
08:42 ewong joined #salt
08:48 TyrfingMjolnir joined #salt
09:02 ajw0100 joined #salt
09:15 Hipikat joined #salt
09:19 fllr joined #salt
09:21 tinuva joined #salt
09:28 googolhash joined #salt
09:31 sroegner joined #salt
09:31 chiui joined #salt
09:38 bhosmer joined #salt
09:42 redondos joined #salt
09:45 babilen Hi all -- Running state.highstate results in quite a lot of output and makes it hard to answer the question: "Did it work the same on all minions and what failed?" -- Is there a way to cut down on the output and have it display only a global summary at the end?
09:46 babilen Or only information about failed states for each minion?
09:49 yomilk joined #salt
09:49 mike25ro babilen: i think that ... on the minion ... /var/log/salt/ you can see only the errors
09:50 babilen mike25ro: Okay - It's just that we finished deploying 20 new boxes and I kept on scrolling through the massive output and thought: "Hmm, all I really care about right now are minions on which a state failed"
09:51 mike25ro i was actually prepared to ask a similar question : i am deploying with salt servers that are on another continent... and the response from the minion is timedout. How can i still see what has happened on the minion ... like the full log of that job? (packages installed, states failed)
09:52 stubee joined #salt
09:52 mike25ro babilen: i know... what you mean ... i am ... building a web app on top of salt... and one of the goals is to easily spot errors. I take the out=json and parse it in php ...  i still have to work on the app a bit though.. not yet as  i want it
09:52 mike25ro i have no idea if just in pure salt is possible to filter only the failed states
09:53 babilen It will be in a year ;)
10:01 giannello joined #salt
10:14 ndrei joined #salt
10:18 lareth joined #salt
10:19 fllr joined #salt
10:22 elfixit joined #salt
10:35 ksk joined #salt
10:35 ksk hey.
10:35 southquist babilen: If you want to cut down on the state-output, take a look at the state_verbose and state_output settings in /etc/salt/master.
10:35 ksk where is a salt-module stored on a minions filesystem?
10:41 ksk find shows me /var/cache/salt/minion/extmods/modules/ and /var/cache/salt/minion/files/base/_modules - whats the difference?
10:54 yomilk joined #salt
10:54 southquist The copy at /var/cache/salt/minion/extmods/modules/ is the one that will be imported into salt and executed. One difference is that /var/cache/salt/minion/files is only readable by root, since other sensetive information might be cached there.
10:54 viq engblom: I'm not one maintaining it, but I'm the one trying to upgrade it. On openbsd-wip it's somewhat still work in progress, it's running on my box and seems to work for remote execution, I haven't tried any configuration management with it, and I want to enable GitFS support before I submit it
10:55 viq engblom: and it certainly could use testing, yes ;)
10:57 viq engblom: though I believe I already replied to your email ;)
10:58 stubee how do most people run salt on a schedule? say every 30mins - using cron? or is there something built-in?
10:59 babilen southquist: Ah, state_output set to mixed is great
10:59 viq stubee: http://docs.saltstack.com/en/latest/topics/jobs/schedule.html
11:01 stubee viq: cheers mate
11:03 viq stubee: but there's also http://docs.saltstack.com/en/latest/ref/states/all/salt.states.cron.html#module-salt.states.cron  ;)
11:04 stubee the problem with cron is salt may be still running when the next cron kicks of another run…
11:07 viq I think I saw a setting somewhere for that
11:07 viq And I don't think the scheduler solves that either
11:08 viq Oh, no, there is a provision for that in the scheduler, my bad.
11:12 yomilk joined #salt
11:13 xmj viq: i'm curious: do many obsd users consider salt even barely secure enough to use?
11:13 Calum joined #salt
11:13 viq xmj: how should I know? ;)
11:13 xmj viq: you're using it ... :p
11:13 viq xmj: I thought you wanted an answer larger than "2 that I know of" ;)
11:14 xmj ah :D
11:14 xmj viq: i'll take that as an answer. heh
11:14 viq xmj: also, is puppet any better? And there are certainly openbsd users that use that.
11:19 fllr joined #salt
11:19 sroegner joined #salt
11:34 giannello joined #salt
11:35 ekristen joined #salt
11:38 rome joined #salt
11:40 elfixit joined #salt
11:42 rome joined #salt
11:44 MZAWeb joined #salt
11:49 jrdx joined #salt
11:52 harobed joined #salt
11:53 ravibhure joined #salt
11:53 mgw joined #salt
11:55 napper joined #salt
11:59 chiui joined #salt
12:04 mgw joined #salt
12:15 AdamSewell joined #salt
12:15 epcim joined #salt
12:19 fllr joined #salt
12:20 Leech joined #salt
12:21 mgw joined #salt
12:28 jslatts joined #salt
12:28 Shenril joined #salt
12:29 bhosmer joined #salt
12:31 mike25ro guys..... i have salt-minion installed on several older machines.  My problem is that i can not run there the highstate .. because there are differences from the old machines and the newer ones... What is a best-practice ... on getting a report of the differences between highstate and what will be changed on the minions ? ... i do not want to lose settings/configs on the minions.
12:33 viq mike25ro: test=True
12:40 mike25ro viq:  i know that ... but there are times when is not enough ... for example... sometimes when managing files i just get.. file is about to be changed... not also the diff.... i will look more into it.
12:40 mike25ro thanks viq
12:40 thayne joined #salt
12:41 sroegner joined #salt
12:42 viq mike25ro: well, I guess depends how the file is put there. But in most situations I did get the diff
12:43 faldridge joined #salt
12:43 mike25ro viq:  i think if i use file.append... or smth like that
12:43 mike25ro anyway thanks
12:50 diegows joined #salt
12:51 rome joined #salt
12:53 jraby joined #salt
12:59 ndrei joined #salt
13:02 ndrei_ joined #salt
13:02 GradysGhost joined #salt
13:07 Shenril joined #salt
13:10 Teknix joined #salt
13:10 harobed_ joined #salt
13:13 Teknix joined #salt
13:15 ZombieFeynman joined #salt
13:17 mpanetta joined #salt
13:17 Teknix left #salt
13:19 Teknix joined #salt
13:19 fllr joined #salt
13:22 mpoole joined #salt
13:22 faldridge joined #salt
13:23 xintron joined #salt
13:23 xintron Playing around with docker and thought I should try a provision solution (instead of plain shell commands) and saw the support for salt (which I've heard about before).
13:25 xintron Should I consider my vagrant VM to be a minion and my host to be the master (even if I'll only have one minion in total) and is it possible to have multiple master "profiles/groups-of-minions"?
13:25 zimski joined #salt
13:25 zimski hello !
13:25 zimski i have problem when i run file.copy on windows minion
13:26 zimski it say : 'file.copy' in not available !
13:27 zimski my salt minion version is 2014.1.1
13:40 fllr joined #salt
13:42 _fllr_ joined #salt
13:42 sroegner joined #salt
13:47 Kraln- joined #salt
13:49 quickdry21 joined #salt
13:51 vejdmn joined #salt
13:55 MTecknology So, this works, but... I'm no python guru, so please pick this sucker apart for me. http://dpaste.com/1781773/
13:56 jimk joined #salt
13:57 jimk hello everyone. I'm new to salt, trying to saltify our current ec2 environment
13:57 MTecknology using anything now?
13:58 jimk lots and lots of php and bash. we have chef deploying nagios on a few machines, but that's it
13:58 yusuket joined #salt
13:58 rnts you'll be so happy when you've figured out how to use salt properly ;-)
13:58 rnts good luck
13:59 undefinednull joined #salt
14:00 jimk yeah, thanks. What I can't quite get my head around right now is how to deal with the ec2 dynamics. Basically, how do i deal with the fact when I launch something, I have no control over hostname and ip address
14:02 * MTecknology needs to start playing with ec2 and learn all of that
14:03 MTecknology when you create the instance, ip and host are returned and you should be able to use that to register whatever you need wherever it's needed; I know nothing beyond that
14:03 rnts jimk: you can enter post-create-hooks into ec2
14:04 rnts which will report back the hostname/ip-addresses to the salt-mster or some specific system that coordinate stuff
14:04 rnts sorry i can't help you more with the ec2-stuff, we're moving away from them as we speak
14:05 timoguin MTecknology: glad you found out the issue. :)
14:06 AdamSewell joined #salt
14:06 AdamSewell joined #salt
14:07 acabrera joined #salt
14:08 jimk ok, I'll start looking at post-create-hooks. There must be a way to define node types based on amazon tags.
14:09 Networkn3rd joined #salt
14:10 jeremyBass joined #salt
14:11 mikemar10 joined #salt
14:14 mattmtl joined #salt
14:15 jimk when I terminate an instance, how does the salt-master purge that key?
14:18 timoguin jimk: if you're not using salt-cloud it doesn't.
14:18 timoguin salt-cloud, however, will pre-seed the keys when you create a new instance, and remove the key when you destroy one
14:20 jimk so can I remove the key by hand then?
14:20 timoguin yes. salt-key -d minionid
14:21 jimk ok. so is there a way to find minions that are disconnected?
14:22 timoguin salt-run manage.down
14:22 timoguin that basically runs a test.ping and gives you all the minions that didn't respond
14:22 eriko joined #salt
14:23 jimk cool. that worked.
14:25 jimk now salt '*' commands are running faster. they would blow through everything then sit there for a few seconds. I guess they were waiting on minions that I had terminated to timeout.
14:26 jimk went away once i deleted old keys
14:29 stevednd joined #salt
14:29 kaptk2 joined #salt
14:30 stevednd is there a way to use a template with file.blockreplace? It doesn't look like it accepts a template, but is there maybe a way to render a template inline?
14:30 Whissi When I run "salt '*' service.restart foo" against servers where foo service is available but not running, will service.restart start them?
14:31 Whissi (I am looking for a way to restart a service only on minions where the service is actual running... on servers where the service is available but not yet running, salt shouldn't do something)
14:34 fllr joined #salt
14:35 danielbachhuber joined #salt
14:36 timoguin Whissi: try it. :)
14:36 timoguin and report back plz
14:37 helderco joined #salt
14:37 conan_the_destro joined #salt
14:37 helderco Hey
14:37 Whissi I tried it. And service.restart started the service on servers where it wasn't running before :/
14:37 helderco fileserver_backend and gitfs_remotes aren’t in /etc/salt/minion… can they be used in a masterless setup?
14:38 timoguin helderco: I think there is an open issue for getting gitfs working for masterless
14:39 helderco timoguin: ok, thanks
14:39 mgw joined #salt
14:39 epcim_ joined #salt
14:40 timoguin helderco: https://github.com/saltstack/salt/issues/6660
14:44 it_dude joined #salt
14:47 timoguin Whissi: fyi, for upstart, service.restart calls 'service restart servicename'. systemd calls 'systemctl restart servicename'
14:47 timoguin and the fall-back service module calls /etc/init.d/servicename restart
14:49 Whissi timoguin: Yup... it is questionable what should happen at the end. One could argue that restart should fail, when the service which should restart isn't yet running (so it cannot be stopped...) ... but other can argue that the result (the service should be up and running at the end) is important... so if it wasn't yet running... well, at the end, it is. :>
14:50 thedodd joined #salt
14:50 Whissi But in my case I was looking for something I can use on the CLI... but I can write a state which will check the status first.
14:50 ZombieFeynman joined #salt
14:53 napper joined #salt
14:56 giannello joined #salt
14:56 sandbender1512 joined #salt
14:56 napper joined #salt
14:57 swissknife joined #salt
14:58 Leech joined #salt
14:59 calvinhp_mac joined #salt
15:00 calvinhp_mac got my environments from gitfs being picked up yesterday
15:00 calvinhp_mac but I'm having a strange issue highstating them
15:00 calvinhp_mac the base env specifies a new repo to install and some deps that get installed
15:00 calvinhp_mac and my acceptance env requires one of those packages for a conf file
15:01 calvinhp_mac but the conf file errors saying it can't find that dep
15:01 calvinhp_mac show_highstate shows that the two required states are in order before the managed.file
15:01 calvinhp_mac but they don't show when I acutally do the highstate
15:01 swissknife Hi, please can I specify more than 1 source for a File.managed ? I'm using #!pyobjects
15:01 alunduil joined #salt
15:01 vbabiy joined #salt
15:01 jalbretsen joined #salt
15:02 calvinhp_mac seems to only be using the states from the more specific environemnt and not base
15:02 calvinhp_mac is this an expected behavior?
15:04 jvasallo joined #salt
15:05 conan_the_destro joined #salt
15:05 ajprog_laptop joined #salt
15:09 jeremyfelt joined #salt
15:09 faldridg_ joined #salt
15:17 anuvrat joined #salt
15:17 ZombieFeynman joined #salt
15:18 timoguin swissknife: that's not currently possible
15:18 [diecast] joined #salt
15:20 swissknife timogui: thks ; basically I want to make a state that distribute a file.conf if there are not custom file.conf__hostname available
15:21 swissknife so I have a default case, and several exceptions ; how to manage it ? thks
15:22 Gordonz joined #salt
15:26 Gordonz joined #salt
15:27 Leech joined #salt
15:27 lareth left #salt
15:31 viq swissknife, timoguin - I thought you could, I thought you could specify several, and first one found would be used
15:32 gothix can i use a map.jinja for testing the oscodename and assigning values based on that
15:32 viq swissknife, timoguin - http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html at the top has an example of doing just that
15:33 viq No idea about pyobjects though
15:34 ndrei joined #salt
15:34 swissknife thks viq, I notice now that double "- salt://foo.conf " in that link , I'll try
15:35 viq gothix: I think so - http://docs.saltstack.com/en/latest/topics/development/conventions/formulas.html and http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.grains.html#salt.modules.grains.filter_by
15:37 swissknife viq, trying http://pastebin.com/BuK6PRRL ; closing for today.. bye
15:40 tligda joined #salt
15:41 snowghost joined #salt
15:42 mgw joined #salt
15:44 crazysim joined #salt
15:44 giantlock joined #salt
15:44 rawzone joined #salt
15:49 helderco joined #salt
15:50 napper joined #salt
15:50 thayne joined #salt
15:56 jvasallo joined #salt
16:01 ndrei joined #salt
16:06 bhosmer joined #salt
16:06 seme_work joined #salt
16:07 RazaR joined #salt
16:07 seme_work I'm getting an error when I try to run salt-run winrepo.genrepo says Failed to compile wmi_provider/init.sls.... how can I troubleshoot this?
16:07 seme_work j
16:08 ajprog_laptop joined #salt
16:11 ndrei joined #salt
16:14 jvasallo joined #salt
16:18 schimmy joined #salt
16:18 masterkorp hello
16:18 masterkorp can i send notifcations from a state to another
16:18 masterkorp ?
16:19 masterkorp eg I have file and a service state that runs that same file
16:19 zain_ joined #salt
16:19 masterkorp now i want to restart the service if that file changes
16:19 faldridge joined #salt
16:20 Eugene watch:
16:20 cowyn_ joined #salt
16:20 Eugene http://docs.saltstack.com/en/latest/ref/states/all/salt.states.service.html
16:20 schimmy1 joined #salt
16:21 ZombieFeynman joined #salt
16:22 masterkorp hmm
16:22 masterkorp can you drop me an example ?
16:22 Eugene There's a great one right there in the docs. ;-)
16:22 masterkorp where ?
16:26 Eugene Where I linked?
16:27 viq masterkorp: and more on that http://docs.saltstack.com/en/latest/ref/states/requisites.html
16:27 masterkorp can it watch 2 things ?
16:27 viq yes
16:27 Eugene Yes, just add them in under watch:
16:28 abe_music joined #salt
16:28 jalaziz joined #salt
16:30 calvinhp_mac why would you ever use require if watch is require with more (and if the module doesn't provide mod_watch it does exactly what require does)
16:30 possibilities joined #salt
16:30 smcquay joined #salt
16:31 viq because sometimes you don't care whether some file changed or not
16:32 abe_music anyone know why minions would constantly be authenticating with the master? I'm seeing this in my master logs: https://gist.github.com/abemusic/10903470
16:34 meteorfox joined #salt
16:34 calvinhp_mac viq:  wouldn't that not matter, if you changed the file, salt should update it, if you didn't it won't
16:34 jaimed joined #salt
16:34 calvinhp_mac it doesn't look at the state of the file on the target and make a decision on that does it?
16:39 TheRealBill joined #salt
16:41 chrisjones joined #salt
16:43 linuxlewis joined #salt
16:45 viq calvinhp_mac: I mean eg a service that only cares that the file is present before it starts, but I don't want to restart the service just because some file (or package) changed
16:46 techhat joined #salt
16:46 techhat left #salt
16:47 ghanima joined #salt
16:47 viq Say I'm using logstash, with kibana and nginx. There's no point starting nginx before kibana was checked out, but there's no point restarting nginx every time the kibana checkout changed.
16:47 ghanima Hey guys
16:48 ghanima I have a really strange issue and apologize if it sounds confusing, but here goes
16:48 calvinhp_mac viq:  ok, that is a good example
16:48 calvinhp_mac thanks
16:49 ghanima I have a saltstack environment running the latest version of salt with 220 minions and 1 salt-master. For some reason when starting everything for the first time(please note I have done multiple installations) my worker_thread usage maxes out(have it set to 30) and my load average never comes down.
16:49 ghanima When I do a job list I get a empty dictionary returned {}
16:49 ghanima when looking at the saltmaster log
16:49 ghanima the only error that comes up is
16:49 ghanima An inconsistency occurred, a job was received with a job id that is not present on the master
16:50 ghanima any thoughts
16:51 astrofog joined #salt
16:51 ghanima thinking about doing a scortched earth, but  I am trying to understand the problem before doing that
16:55 aw110f joined #salt
16:55 linuxlew_ joined #salt
16:56 ghanima Just curious is there a way to remove a queued up transaction on the minion
16:56 travisfischer joined #salt
16:57 TheRealBill joined #salt
16:57 ghanima I presume there in this directory
16:57 ghanima var/cache/salt/minion/proc
16:59 svs_ joined #salt
17:00 juaqim joined #salt
17:00 linuxlewis joined #salt
17:00 jvasallo joined #salt
17:01 thayne joined #salt
17:03 A||SySt3msG0 Can you elaborate on that scortched earth procedure?
17:05 calvinhp_mac joined #salt
17:05 mgw joined #salt
17:06 ghanima A|Syst3msG0: reinstall minion agent and master agent
17:06 ghanima remove rpm. delete all data dirs
17:06 ghanima then reinstall
17:06 ghanima assuming I got all datadirs that would be scorched earth
17:11 jvasallo joined #salt
17:11 __alex i read docs over and over again, i'm still confused: can i set the environment in pillar so the states are from dev/qa/whatever?
17:13 ghanima _alex yes you can
17:14 __alex ghanima: what's the variable name then in pillar? "env" or "environment"?
17:15 timoguin environment
17:15 mgw joined #salt
17:15 saguilar joined #salt
17:16 __alex ghanima: i why did i type it wrong in my pillar sls and didn't see it?
17:16 __alex ghanima: thanks
17:17 ghanima ____alex: assumption you ave defined your pillar_roots: correct
17:18 ldlework left #salt
17:19 andredieb joined #salt
17:20 rushm0r3 joined #salt
17:20 rushm0r3 left #salt
17:21 gadams999 joined #salt
17:21 travisfischer joined #salt
17:22 arthabaska joined #salt
17:22 ghanima anyone see this error Salt request timed out. If this error persists, worker_threads may need to be increased
17:22 faldridge joined #salt
17:22 ghanima odd thing is I get this consistenly as root and rooot is in the wheel group I believe
17:26 snowghost joined #salt
17:27 seme_work joined #salt
17:29 ajw0100 joined #salt
17:30 jslatts joined #salt
17:30 arthabaska joined #salt
17:32 kermit joined #salt
17:33 MTecknology Is there any way to make specific servers run state.highstate when a specific git repo on a gitlab server is updated?
17:34 Ryan_Lane joined #salt
17:34 MTecknology I'm guessing I'll have to write an api on the salt server to react to events from the gitlab server
17:35 calvinhp_mac joined #salt
17:36 druonysus joined #salt
17:36 druonysus joined #salt
17:37 timoguin MTecknology: yea that's what i'd do. repo post hooks to an api that fires salt events
17:38 MTecknology I see github/saltstack/salt-api. I doubt that can be a drop-in solution since I want to control what can actually be done, but maybe a good reference
17:39 MTecknology oh.....
17:39 MTecknology All authentication is done through Salt's external auth system.
17:40 whiteinge MTecknology: it sounds like you're looking for this: http://salt-api.readthedocs.org/en/latest/ref/netapis/all/saltapi.netapi.rest_cherrypy.html#saltapi.netapi.rest_cherrypy.app.Webhook
17:42 jimklo joined #salt
17:43 MTecknology whiteinge: I could set up salt-api and just have the external auth system restrict what can be executed and where and then have gitlab post to that on commits?
17:43 possibilities joined #salt
17:45 whiteinge MTecknology: if you can add authentication credentials to the HTTP request that gitlab's hook sends out then, yes that should do it.
17:46 mateoconfeugo joined #salt
17:46 MTecknology All I'm trying to do is have state.highstate triggered on specific servers when a git repo is updated.
17:46 MTecknology This seems like an overly complicated solution, but also something that's incredibly flexible and future-proof.
17:48 whiteinge you have to install salt-api but everything else is quite simple
17:48 whiteinge you don't need to use eauth for this if you don't want to
17:48 MTecknology I don't want to use anything that I don't need to.
17:50 whiteinge is your gitlab server running a minion?
17:50 MTecknology there's no init script in the repo :(
17:50 timoguin you could also write a little daemon that accepts the post hooks from gitlab and then fires a custom event that the salt-master reacts to
17:51 whiteinge MTecknology: look in the "pkg" directory
17:51 MTecknology timoguin: that's what I meant
17:51 MTecknology there it is!
17:52 timoguin following what whiteinge said, if the gitlab server is on a minion, you could probably just fire an event directly from there to the master
17:52 MTecknology I installed it from pip, I'm wondering if I should just use the ppa, but this is a debian box and I don't like intermixing them.
17:53 whiteinge the PPA hasn't been updated for the current release
17:53 MTecknology timoguin: it is a minion and the salt mine would probably suffice, but if I want to extend this in the future, the bigger option might be better. (i'm thinking anyway..)
17:53 whiteinge i'm working on that thsi week
17:53 arthabaska joined #salt
17:54 MTecknology whiteinge: the mine?
17:54 whiteinge MTecknology: if it's a minion, you can just use ``salt-call event.fire_master``
17:54 ghanima whiteinge: hey just to confirm that fix you gave me on Monday worked
17:55 whiteinge ghanima: nice! (-ish)
17:55 whiteinge still need to get that working without hacks, but at least you can get some work done :-P
17:56 ghanima whiteinge: is there a way to refresh state in saltstac I am constanntly getting timeouts eerrors in the log and I have played aroung increasing my worker threads but still no avail
17:56 ghanima the error I am seeing is salt.client      ][ERROR   ] Salt request timed out. If this error persists, worker_threads may need to be increased.
17:57 ajw0100 joined #salt
17:58 whiteinge ghanima: refresh my memory, what version of salt master and minion are you running?
17:58 ghanima 2014.1.0-1
17:59 faldridg_ joined #salt
18:00 whiteinge ghanima: unless you have a ton of minions (600+), that error may be due to a bug in the current release that is causing too many auth requests from the minions
18:01 ghanima whiteinge: oh wow really do u recommend downgrading
18:01 whiteinge it's been fixed and we're cutting a new release today, though it'll take a few days for the packages get built
18:01 MTecknology whiteinge: so make the gitlab minion somehow get that hook on the system and have it run salt-call '{"data":"update_moses"}' '??' and then use the reactor to trigger that update?
18:02 timoguin MTecknology: you should be able to define a post-commit hook on the gitlab server that shells out to salt-call to fire an event
18:02 timoguin and then the reactor would trigger the appropriate deploy state
18:02 nmistry joined #salt
18:02 calvinhp_mac joined #salt
18:04 whiteinge ghanima: no, there's a config option you can set now. sec...
18:04 calvinhp_mac left #salt
18:06 ajw0100 joined #salt
18:06 CeBe joined #salt
18:06 whiteinge ghanima: look at the config options in this comment:
18:06 whiteinge https://github.com/saltstack/salt/issues/11781#issuecomment-39869201
18:07 MTecknology timoguin: that sounds like the bestest best solution... now to figure out how to do that
18:08 whiteinge ghanima: only trick is those options need to go in the minion config. you could use a single file.managed state to drop a file in /etc/salt/minion.d/auth_flood_fix.conf though
18:08 ghanima whiteinge: ahhh that's on the minion got it
18:08 ghanima just curious than if that is the temporary fix what is the permanant one that will be released shortly
18:10 whiteinge ghanima: http://paste.fedoraproject.org/94761/39767179
18:10 whiteinge ghanima: you can remove that file after 2014.1.3. basically what happened is the default in the code was erroneously changed to a short timeout, but the config option is still respected.
18:12 ajolo joined #salt
18:13 ghanima whiteinge: ahhh gotcha... man that would have took me foorever to troubleshoot  that looking at the github history that took 12 days to troubleshoot huh
18:14 MK_FG joined #salt
18:15 thedodd joined #salt
18:16 MTecknology timoguin: salt-call can only be run from root, and I don't want any salt-call command to be executed by git hooks, so... I'll need to create a wrapper script that commits that, won't I?
18:17 Ryan_Lane if I need to have a state generate something, then have it passed into another state, is that possible?
18:18 ghanima whiteinge: so is this tempoorary fix will this delay the execution of commands
18:18 Ryan_Lane specifically, I'm looking at launch configurations for EC2. they can't be modified, only deleted and created, and they can't be deleted if they are connected to an autoscaling group
18:18 ghanima I am concerned about the recon_randomize setting
18:18 Ryan_Lane so, I need to generate a unique name, and have an autoscaling group use the unique name
18:19 whiteinge ghanima: no, i don't think so
18:19 UtahDave joined #salt
18:20 ZombieFeynman joined #salt
18:21 timoguin MTecknology: I guess so: https://groups.google.com/forum/#!topic/salt-users/mXK4464Jlk8
18:21 timoguin ^ wrapper script with sudo access
18:23 ohmygoshjosh joined #salt
18:23 ghanima whiteinge: this is working perfectly now... GOD I have been scratching my head on this for hours. I hate when things come together like that... whiteinge thanks for the assist on that don't think I would have been able to do that on my own.... so you said a new release in the next couple of days... are release notes or changelog notes available to review want to see
18:23 ghanima what is coming down the pipe other than that change
18:24 whiteinge ghanima: http://docs.saltstack.com/en/latest/topics/releases/2014.1.2.html
18:24 whiteinge those release notes aren't quite complete yet. check them again once we send out the official announcement
18:26 ghanima whiteinge: as you can tell I am new to the project so I really apologize on the newbie questions... but is there an announce Distribution List or just the IRC channel
18:27 Ryan_Lane could I use the watch interface to do what I want?
18:27 whiteinge we have to mailing lists, salt-users and salt-announce
18:27 whiteinge https://groups.google.com/forum/#!forum/salt-users
18:27 whiteinge https://groups.google.com/forum/#!forum/salt-announce
18:27 whiteinge ghanima: ^^
18:27 Ryan_Lane I want to generate a unique name from a launch config state, and have the autoscale state watch the launch config state, getting the name and changing it in the autoscale group if it changed
18:28 ghanima whiteinge: thanks man
18:29 whiteinge np! and we're plenty happy to field newbie questsions in here :)
18:29 __alex what's the syntax for require: file: for multiple files?
18:30 ghanima whiteinge: yeah everyone has been super friendly and polite which is my history is not always the case on IRC so thank you for that
18:31 whiteinge :)
18:31 ndrei joined #salt
18:32 whiteinge Ryan_Lane: could you lean on Salt Mine for that?
18:32 baniir joined #salt
18:32 Ryan_Lane nope. I want this to work 100% masterless
18:32 whiteinge __alex: require statements take a list, so: ``require: - file: foo\n- file: bar``
18:32 Ryan_Lane and I also want it to be able to work in a single state run
18:33 arthabaska joined #salt
18:33 Ryan_Lane if I push something into the mine it would require at least two runs, right?
18:35 whiteinge Ryan_Lane: unsure. let me puzzle over taht for a sec. (gotta run afk)
18:35 * Ryan_Lane nods
18:36 Ryan_Lane watchers definitely have the ability to send data back that's change
18:36 Ryan_Lane *changed
18:36 catpiggest joined #salt
18:36 Ryan_Lane and the interface that's watching it should be able to consume that data
18:36 __alex whiteinge: hu, thought that would be duplicate dict keys, but it's not. works great, thx.
18:38 vejdmn joined #salt
18:39 vejdmn joined #salt
18:41 toastedpenguin joined #salt
18:42 tligda joined #salt
18:44 KyleG joined #salt
18:44 KyleG joined #salt
18:44 ghanima in al the tutorials I am reading there seems to be a generic reference of if you want to test the function of your saltstack installation run salt '*' test.ping however I am seeing problems with this when minions are offline and or if there salt master is being overhwlemed is there another check that is faster in the two scenerios I decribed
18:45 kermit joined #salt
18:45 timoguin ghanima: that's pretty much the shortest module there is
18:45 timoguin the CLI is just waiting on all the minions to respond before returning
18:48 kickerdog joined #salt
18:48 kickerdog left #salt
18:49 napper joined #salt
18:50 ghanima timoguin: got it thank you sir
18:50 ipalreadytaken joined #salt
18:52 possibilities joined #salt
18:53 stubee joined #salt
19:02 smcquay joined #salt
19:02 undefinednull joined #salt
19:05 schimmy joined #salt
19:14 jvasallo joined #salt
19:16 seme_work Hey guys... I'm getting an error on windows workstation minions where python can't write the winrepo.p cache file.  Is there a workaround?  I'm using 2014.1.0-1 on both master and minion
19:18 seme_work I enabled debugging and it looks like the minion can't open the file...google seems to indicate that there is a bug in how it is written but I'm not sure if that is exactly the same issue I'm having
19:18 ghanima question: is there any performance gains with cython enabled
19:19 whiteinge Ryan_Lane: could you solve this at the jinja level? generate a unique name in jinja that is then used in both states?
19:20 whiteinge Ryan_Lane: perhaps by calling out to an execution module, i.e.: {% set unique_name = salt['ec2.get_or_set_unique_name'](foo, bar) %}
19:20 whiteinge ?
19:21 Ryan_Lane whiteinge: well, I'm writing state modules
19:21 Ryan_Lane which I'd like to upstream
19:22 Ryan_Lane so I'd like to have a generic approach to this, where it can use a set name, or have an option to auto-generate one
19:22 UtahDave seme_work: I think that was fixed.  Can you test with a minion using 2014.1.1?
19:23 Ryan_Lane I was going to base the name on an md5 sum of the ordered key json (or yaml) output of the data being passed into the launch config state
19:23 seme_work UtahDave: I don't think I can use 2014.1.1 on my minion with 2014.1.0 on my master.
19:23 seme_work I was getting errors with 2014.1.1 on my minion but it failed silently
19:23 Ryan_Lane so that if the options change, the resource name stays the same, but a new launch configuration is generated
19:23 UtahDave seme_work: Hm. Ok, I can do some testing later this evening.
19:24 Ryan_Lane then the generated launch config's name would be passed from launch config to autoscale, maybe via watchers
19:24 Ryan_Lane back in a bit
19:24 seme_work I'll try it again... now to see if I can get it to work
19:24 seme_work is there a file:/// type format for win_repo_cachefile?
19:24 seme_work instead of using the salt://win/repo/winrepo.p arg
19:27 bhosmer joined #salt
19:28 Ryan_Lane whiteinge: back.
19:29 Ryan_Lane whiteinge: the biggest issue here is that if I make a launch config state module that it won't work as people expect unless it can dynamically generate names
19:29 Ryan_Lane because a user could name a launch config x, then change its config and there's no way for the state module to actually keep that name
19:29 Ryan_Lane since it can't delete launch configs in use and it can't modify launch configs
19:29 ghanima I am researching this site: http://docs.saltstack.com/en/latest/topics/tutorials/states_pt5.html that has two mechs .... orchestration overstate & orchestrate.....  Outside of this doc is there any more documentation or video that describes implementation looking for comparisons between the two
19:30 ndrei joined #salt
19:35 jslatts joined #salt
19:39 seme_work UtahDave: ok so I upgraded the minion and I get the same result
19:39 juaqim joined #salt
19:39 seme_work it doesn't print an error though just says [salt.loaded.int.module.cp][ERROR   ] Unable to cache file 'salt://win/repo/winrepo.p' from saltenv 'base'.
19:40 seme_work I also changed the perms so that anyone could do anything to the entire salt directory with the same result.
19:40 seme_work I also tried running the minion as a user that has full local admin rights, same result
19:41 Cidan is there a way to pass arguments to pip in pip state?
19:41 Cidan ie: --user
19:41 Cidan pip install --user pyes
19:42 aranhoide joined #salt
19:42 toastedpenguin joined #salt
19:43 aranhoide I have a salt master that won't communicate with any of its minions anymore.  not even with (the minion service in) itself.  this has been running alright for months.  currently it's at version 2014.1.1 (which, again, has been working alright for a while).  I can ssh into both the cloudmaster and any minions.  any pointers on how to troubleshoot this?
19:44 whiteinge Ryan_Lane: i'm with you on the general-use state modules, sorry for the mental detour there.
19:45 whiteinge Ryan_Lane: the thing i'm puzzling over is how (or even whether) this workflow matches the intent of States. i'm trying to think of an analog with existing states where one state directly depends on data from a previous state
19:46 MTecknology timoguin: sorry, got side tracked with a meeting about legal stuff ... always fun
19:46 MTecknology timoguin: that looks like exactly what I want. :D
19:48 possibilities joined #salt
19:49 schimmy joined #salt
19:49 stevednd Is there any method available or accepted way to send an email, or some other kind of notification as a result of applying states?
19:53 whiteinge stevednd: there are many ways of doing that. :-)   you could use the SMTP returner, you could use Salt's event bus to call a function that sends an email (or does anything else)
19:54 whiteinge stevednd: what specific info are you interested in having in the notification?
19:55 stevednd I need to include the host name, and a newly generated pubkey
19:58 stevednd whiteinge: can a returner be used from inside a statefile? The examples I see show it being used just through the command line
19:59 slawek- joined #salt
19:59 danielbachhuber joined #salt
20:00 baniir joined #salt
20:00 schimmy joined #salt
20:00 stubee joined #salt
20:03 schimmy1 joined #salt
20:05 thedodd joined #salt
20:07 srage_ joined #salt
20:09 monokrome joined #salt
20:11 mgw joined #salt
20:12 slawek- Hi guys. Newbie's security concern. It's about roles and keeping them in grains. Let's say my load balancer gets compromised. As it's a minion, the attacker could edit grains to change the role to something more critical like db machine and by running highstate get some sensitive data from the master. What would you recommend to prevent this?
20:12 Eugene Don't store sensitive data in salt://
20:12 tligda joined #salt
20:12 Eugene Don't trust minions about anything; I do all of my role defining in a pillar based upon a system's ID
20:13 Eugene Don't run salt:// over an open network
20:13 Eugene And especially don't do Open Masters
20:13 ghanima the command salt '*' sys.doc
20:14 ghanima is that output available in a pdf
20:15 Eugene If you're worried about compromised systems(as you always should be), dont' enable any of the peering features.
20:15 calvinhp_mac joined #salt
20:15 ghanima Eugene: everything you are describing is this the default config
20:16 Eugene Security-sensitive items(ssl keys, for example) should be moved via a different system. Deployment using SSH works.
20:16 whiteinge stevednd: you can specify a default returner in the master or minion config
20:16 Eugene If it's something that you'll need to "roll into" a minion image(eg, a clone-able web server) consider keeping it in a git repo, accessed via SSH keys.
20:17 linuxlewis joined #salt
20:17 slawek- Hi Eugene, thanks. I'm not planning to use salt over an open network, nor running open master. I'm also considering storing roles in pillars by minion ID. What are those peering features you're mentioning?
20:17 Eugene You can reference this in your state configs, but only certain minions will be able to get to it
20:17 Eugene http://docs.saltstack.com/en/latest/ref/peer.html
20:17 stevednd whiteinge: does that just email the log of an application of state on a minion?
20:17 slawek- oh ok
20:17 whiteinge obviously that will get triggered for all salt executions though
20:18 lukeg s
20:18 whiteinge stevednd: if you need more granularity than that, there's an awesome third-party project that watches salt's event bus and lets you take more granular action. sec i'll find the link...
20:18 timoguin that project's getting a lot of attention. :)
20:18 whiteinge stevednd: https://github.com/felskrone/salt-eventsd
20:19 whiteinge stevednd: you can make a custom returner easily enough to format/process the data. the default SMTP returner will email the full result of the state run (what you see when you run state.highstate --out=json)
20:19 manfred whiteinge: should build in a returner: for use in all states like require: or watch:
20:21 napper joined #salt
20:21 stevednd whiteinge: thanks, I'll look at that. I think I'm thinking along the lines of what manfred mentioned
20:22 whiteinge manfred: i could see some utility in that. depending on the returner it could be a little heavy. a possible alternative is to fire custom events from a state using the module.wait state
20:22 harobed_ joined #salt
20:22 manfred yeah, I am going to look into this :)
20:22 whiteinge then take some action on the custom event. like with salt-eventsd or with a custom script
20:23 harobed_ joined #salt
20:23 belekas joined #salt
20:23 manfred i am going to try adding the returner to salt/state.py internal_keywords, and see if I can make it work :)
20:23 slawek- It's funny, because even if I don't store any highly sensitive data in salt, using this roles in grains approach will still in most cases allow the attacker to learn a great deal about my internal network whenever my public-facing machine is a minion
20:23 seme_work hey guys on my minion where should file_roots: base: be pointing to... right now it points to c:\salt\srv\salt which is the default but I don't think thats right... should it be C:\salt\var\cache\salt\minion\files\base
20:24 harobed_ joined #salt
20:24 slawek- Eugene: I'll start with taking your advice about only storing roles in pillar by minon id. thanks
20:24 whiteinge manfred: if you intend to send that upstream, keep in mind we're *extremely* careful about changes to the state compiler. emiting custom events is probably the way to go -- there's a ticket for more easily emitting events from states
20:25 whiteinge manfred: https://github.com/saltstack/salt/issues/11762
20:25 manfred sounds good, I am just going to use it as an excuse to dig into how the states stuff actually works :)
20:25 whiteinge cool :)
20:25 manfred cool thanks!
20:25 stevednd whiteinge: is interacting/handling events the reactor system described in the docs?
20:26 stevednd or is that different from what you're talking about here?
20:26 whiteinge stevednd: http://docs.saltstack.com/en/latest/topics/reactor/index.html
20:26 whiteinge the salt-eventsd project doesn't use the reactor and listens to salt's event bus natively
20:26 manfred whiteinge: oohh i like ryan_lane's idea better
20:28 fxhp basepi:  NameError: name 'zmq' is not defined
20:28 fxhp on 2014.1.2
20:28 zain_ joined #salt
20:29 napper joined #salt
20:29 monokrome joined #salt
20:32 eriko joined #salt
20:33 ipalreadytaken joined #salt
20:34 smcquay joined #salt
20:37 napper joined #salt
20:38 seanz Is it really wise for saltstack to be creating their own messaging service right now? There are so many issues left with the provisioning aspect of the code.
20:40 ekristen anyone know how I can do a {{ variable }} replace in a set with quotes?
20:40 linuxlew_ joined #salt
20:40 seanz ekristen: Is this in bash?
20:40 seanz For a salt module call?
20:40 ekristen seanz: no in a state
20:40 seanz ekristen: Why do quotes matter?
20:41 seanz jinja doesn't care if variables are surrounded by quotes or not.
20:41 seanz Have you tried it?
20:41 ekristen {%- set vns_mongodb_viaurl = "mongodb://{{ mongodb_host }}:27017/{{ mongodb_db }}" -%}
20:41 seanz ekristen: Oh, as part of an assignment.
20:41 ekristen yes
20:42 seanz I don't think python does string interpolation without the .format() function. Maybe jinja has support for that.
20:42 ekristen the {{ }} inside the quotes don’t get replaced
20:42 ekristen hrm I might be able to do it without the set
20:42 ekristen now that I look at my state
20:44 seanz ekristen: Some people are saying {{ "%s"|format(myvar) }} works.
20:44 manfred it should
20:44 manfred {{ "%s - %s"|format("Hello?", "Foo!") }}
20:44 manfred -> Hello? - Foo
20:44 manfred just
20:44 ekristen kk good to know
20:45 ekristen thanks!
20:45 manfred "mongodb://%s:27017/%s"|format(mongodb_host, mongodb_db)
20:45 manfred ekristen: http://jinja.pocoo.org/docs/templates/#format
20:45 ekristen thanks
20:46 ndrei joined #salt
20:47 Ryan_Lane joined #salt
20:49 manfred whiteinge: any idea if the default returner is split out at all? or is it just builtin?
20:49 whiteinge manfred: how do you mean?
20:50 whiteinge specifying a default returner is the exact same thing as always adding one at the CLI, e.g.: salt '*' test.ping --ret=somereturner
20:50 whiteinge new/custom returner modules are easy to add
20:51 manfred i mean like the just regular, return to the master and print to the command line.  isthat split into it's own file in salt/returners, or is it just the return from the State().call
20:51 whiteinge oh, gotcha
20:51 whiteinge no, the return to the master happens regardless. it's not a returner module
20:52 whiteinge it's in salt/cli.py (i think)
20:52 Gareth 'lo
20:52 manfred ok, so even if you use the smptd returner, it still returns to the master. awesome
20:52 manfred thanks
20:52 * whiteinge nods
20:52 whiteinge np
20:53 MTecknology I have a server where users are being cached. I don't know why. I restarted nscd and nslcd isn't on here. Incorrect data is cached and I can't flush it. *grumble*  legacy systems. :(
20:54 kermit joined #salt
20:54 whiteinge Gareth!
20:56 Gareth whiteinge: howdy :)  hows it going?
20:58 whiteinge other than spotty internets (which i'm pretty sure is basepi's fault), not too shabby :)
20:58 Gareth basepi is torrenting again?
20:58 Ryan_Lane whiteinge: what I was aiming for was for the states to not have to depend on each other
20:58 Ryan_Lane where you can explicitly name a launch config, and use that explicit name in the autoscaling group
20:59 manfred Ryan_Lane: this one ? https://github.com/saltstack/salt/issues/11762
20:59 Ryan_Lane but also to have an option where the launch configuration can dynamically generate a resource and pass it to the autoscaling group via a watcher
20:59 Ryan_Lane manfred: heh, nah, this is a new one
20:59 manfred ahh
20:59 Ryan_Lane not event based
20:59 cnelsonsic joined #salt
20:59 manfred well i am looking at that one, and trying to figure out how states work and might be doing that one over the next couple days :)
20:59 manfred i like the idea
20:59 manfred +1
20:59 Ryan_Lane but where one state would generate a resource and another state would get the generated resource via watch
21:00 manfred Ryan_Lane: you have the best ideas
21:00 manfred
21:00 Ryan_Lane heh
21:00 Ryan_Lane thanks
21:00 manfred now to go get my new laptop bag and then dig back in to states.py :)
21:00 manfred peace o/
21:00 bhosmer joined #salt
21:04 garthk joined #salt
21:04 whiteinge Ryan_Lane: i just realized i've been wanting similar functionality to what you're describing for the "module.wait" state
21:05 whiteinge "grr. gone again!"
21:05 whiteinge (nightmare before christmas quote)
21:05 Ryan_Lane joined #salt
21:07 mgw joined #salt
21:08 pssblts joined #salt
21:09 ndrei joined #salt
21:10 whiteinge Ryan_Lane: i just realized i've been wanting similar functionality to what you're describing for the "module.wait" state
21:10 whiteinge Ryan_Lane: it may even be possible with the current version of salt. i'm not familiar with that code though
21:10 robawt wuddup whiteinge, missed you at Pycon
21:10 whiteinge states can indeed pass info between them with a __context__ dictionary. the pkg states do just that
21:11 gothix joined #salt
21:11 whiteinge robawt: likewise, man!
21:11 whiteinge would have been nice to catch up over beers
21:11 robawt there's always next time
21:11 whiteinge i haven't missed a pycon in five or six years :(
21:11 robawt too busy living the good life to stop and visit Pycon eh?
21:12 robawt we'll have to make up for it at Scale then
21:12 stubee joined #salt
21:13 whiteinge living the good life...or traveling so much for Salt that if i can sit one out one trip i might be able to stave off my ultimate divorce by another month or two  ;-)
21:13 robawt heh
21:14 alunduil joined #salt
21:18 napper joined #salt
21:19 schimmy joined #salt
21:21 schimmy1 joined #salt
21:21 whiteinge jacksontj: ping me if you want help getting docs for the saltnado module up. i stopped just short of adding those because I didn't want to pre-announce anything for you
21:21 nmistry joined #salt
21:22 whiteinge jacksontj: if you add the following file it will Just Work (TM): doc/ref/netapis/all/saltapi.netapi.rest_tornado.rst
21:22 zain_ joined #salt
21:22 jacksontj whiteinge: will do :) i actually think i might re-work the url structure a bit ;)
21:23 whiteinge awesomes
21:26 gothix can anyone tell me if this is valid in my map.jinja {% set tomcat = salt['grains.item '](
21:27 gothix I want to bas my tomcat version off my version of RHEL if oscodename is Santago then my versionwill be 6
21:27 Ryan_Lane whiteinge: heh
21:28 Ryan_Lane whiteinge: I'm pretty sure it's possible with watch
21:29 Ryan_Lane since watch has a return data structure and the watcher gets that data structure
21:33 ZombieFeynman joined #salt
21:33 jeremyBass joined #salt
21:38 linuxlewis joined #salt
21:39 cewood joined #salt
21:40 ZombieFeynman joined #salt
21:41 stubee joined #salt
21:41 whiteinge Ryan_Lane: does it? do you have an example state handy?
21:45 kermit joined #salt
21:47 jimklo greetings... trying to install saltstack on my mac as a development master to connect to vmware minions... installed via homebrew, however when I go to launch salt-master, the yaml module isn't found... any advice?
21:47 Ryan_Lane whiteinge: I don't have one handy
21:47 Ryan_Lane whiteinge: I was looking at this: http://docs.saltstack.com/en/latest/ref/states/ordering.html#watch-and-the-mod-watch-function
21:48 Ryan_Lane ah. it looks like it doesn't actually take any usable input
21:48 Ryan_Lane :(
21:51 whiteinge Ryan_Lane: ok, we're on the same page. i think that could be added though. i'm looking into how...
21:51 tligda joined #salt
21:52 Ryan_Lane whiteinge: take a look at states/docketio.py
21:52 Ryan_Lane and its mod_watch
21:52 Katafalkas joined #salt
21:53 whiteinge hm. that does look like what we need...
21:55 Ryan_Lane though it's still just providing the calling function
21:55 Ryan_Lane and not any data from the function
22:00 napper joined #salt
22:02 stubee joined #salt
22:05 snuffeluffegus joined #salt
22:06 UtahDave joined #salt
22:06 jeffrubic joined #salt
22:09 MindDrive joined #salt
22:10 helderco joined #salt
22:15 Ryan_Lane whiteinge: I see the kwargs are passed in from the state, do you know what the *args passed in are?
22:15 Ryan_Lane hm. I wish thatch was in here
22:18 UtahDave joined #salt
22:22 faldridge joined #salt
22:24 ldlework joined #salt
22:26 gothix joined #salt
22:27 gothix joined #salt
22:27 xinkeT joined #salt
22:28 whiteinge Ryan_Lane: i think so. we just had a convo about it here in the office. sec...experimenting
22:28 [diecast] joined #salt
22:28 Ryan_Lane it looks like it may be the return data?
22:29 philipsd6 I don't know why anyone would want to make a salt-ssh roster identical to your current minions, but here's a one-liner to do just that:
22:29 philipsd6 salt '*' --out=txt cmd.run template=jinja "printf '{{ grains.id }}:\n  host: {{ grains.fqdn }}\n  user: yourname\n  priv: /path/to/id_rsa\n  sudo: false'" | awk -F':' '{ st = index($0,":")+1;print substr($0,st+1)}'
22:30 philipsd6 fwiw, I needed it for salt-ssh testing. :)
22:31 fragamus joined #salt
22:34 bhosmer joined #salt
22:37 UtahDave awesome, philipsd6!
22:38 rome joined #salt
22:41 Ryan_Lane ugh. I hate having to install homebrew. is swig only needed because of m2crypto?
22:41 Ryan_Lane and that's also the only reason I need to use homebrew on osx?
22:42 bemehow joined #salt
22:43 bhosmer joined #salt
22:46 travisfischer joined #salt
22:48 jimklo Ryan_Lane: have you successfully gotten saltstack installed via homebrew?
22:48 Ryan_Lane well, I'm really trying to get a dev environment up
22:48 jimklo I'm trying the same thing...
22:48 Ryan_Lane and I need brew to install gcc and swig so that I can use pip to install m2crypto
22:48 Ryan_Lane it's kind of a nightmare
22:48 Ryan_Lane m2crypto needs to die in a fire
22:49 jimklo well you need XCode for those things
22:49 Ryan_Lane gcc is linked to clang, which won't work
22:49 jimklo at least for gcc
22:49 Ryan_Lane you need actual gcc, which you need to get from brew :'(
22:50 jimklo ahh... I did the whole `brew install saltstack` but none of the module dependencies are getting discovered
22:50 Ryan_Lane yeah, I have the same issue there
22:51 Ryan_Lane so, I'm just going to go via pip after getting all the dependencies fixed
22:52 jimklo yep... thats what I'm in the process of trying to do
22:54 APLU joined #salt
23:08 googolhash joined #salt
23:08 bhosmer joined #salt
23:09 ewhitake joined #salt
23:12 Whissi Mh. With salt-2014.1.2 something has changed. Previously, with 2014.1.1, when I run a long running command with "cmd.run" the salt command keeps running until the last minion returned or I pressed CTRL+C to send the command to the background.
23:13 Whissi Now it seems like that after a timeout, the command will be send to background and the prompt will return.
23:13 manfred Whissi: there is a known regression, 2014.1.3 is already in the works to fix it
23:13 UtahDave Whissi: yeah, we just fixed that.  don't use 2014.1.2   It hasn't been officially released
23:15 tligda joined #salt
23:15 Whissi OK. Thanks for the information.
23:25 MTecknology this is fun... the first minion I'm deploying from a template since moving salt states to gitfs is breaking... http://dpaste.com/1782442/
23:26 MTecknology nope... apparently this is broken for all minions. those things shouldn't be keys :S
23:26 meteorfo_ joined #salt
23:26 MTecknology those particular files are pillars, not states
23:28 MTecknology oooooooh.......
23:34 napper joined #salt
23:43 ajprog_laptop joined #salt
23:48 ipalreadytaken joined #salt
23:50 nmistry joined #salt
23:51 Eugene Out of curiousity, what does it take to become a "SaltStack Certified Trainer"
23:51 Eugene Be a Salt employee?
23:52 manfred take the certification test, not sure how you get to take the test anymore.  The first batch was if you went to saltconf
23:52 * Eugene may or may not be padding his resume in preparation for a new job
23:53 aranhoide joined #salt
23:53 whiteinge Ryan_Lane: so, good news is the thing we need to make this happen will be in Helium
23:54 Ryan_Lane heh. bad news is that won't be for a few months? :)
23:54 whiteinge yeah. the RC will probably be cut in May
23:55 whiteinge state modules will have access to the __running__ dictionary which means they can pull the results from previous states right from there
23:55 * Ryan_Lane nods
23:55 Ryan_Lane do they not currently have access to that? it looks like it's injected right now
23:55 Ryan_Lane oh. I'm of course looking at the develop branch
23:56 whiteinge i'm looking at the same...
23:56 whiteinge not sure
23:56 Ryan_Lane current release may not have that
23:58 aranhoide I have a master with ~100 minions.  it gets spammed with authentication requests, to the extent that salt-master processes cumulatively take ~100% CPU.  the master logs that the authentication requests are being accepted, but the minions get SaltReqTimeoutErrors (60 seconds) trying to authenticate so they keep retrying
23:58 Ryan_Lane whiteinge: looks like that global is always injected in call
23:59 Ryan_Lane even in the current release
23:59 Ryan_Lane if I can ever get a developer environment up on my laptop I'll do some tests :)
23:59 aranhoide has anyone experienced anything like this?

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