Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-03-11

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

All times shown according to UTC.

Time Nick Message
00:00 Vynce i see a lot of grains['foo'] and some salt['grains.get']('foo', 'default') — i understand the latter has to do with graceful behavior when the grain is not defined, but i' unclear what the behavior of grains['nothere'] actually is.
00:01 Vynce will the whole state fail?  or will a minion missing a grain jsut never match a test for that grain?  or … ?
00:04 forrest Vynce: Regarding your earlier comment, for sure. Just didn't want you to think we were ignoring you on purpose
00:04 iggy in a state, that'll fail
00:04 iggy you'll get a render error
00:05 Vynce cool.  hard crash on misconfig can be a feature.
00:06 Vynce also, i don't see a good generalized point in teh docs about salt['foo']() vs foo() — just some throwaway side comments. is there a good reference for that?
00:06 iggy well, these days there are salt.states.test.*
00:06 iggy but yeah, before that was your best bet
00:07 jaloren iggy: it does in fact look like some kind race/deadlock condition. The sleep seems to make the problem go away.
00:08 jaloren @iggy: seems like a bug. Why wouldn't the message bus simple queue up each state.sls and execute them in order.
00:08 iggy jaloren: maybe a bug? Worth opening an issue either way
00:08 jaloren iggy: agreed....if nothing else they c an tell me what i am doing wrong and then i can fix my script!
00:08 jaloren :)
00:08 iggy and maybe document the behavior if it's expected/unfixable
00:12 jerematic joined #salt
00:13 opennode_ joined #salt
00:14 TyrfingMjolnir joined #salt
00:19 bfoxwell joined #salt
00:29 Pixionus Hey so I was messing with my servers and it looks like those that are running 2014.1.11 respond much faster to test module commands than 7.x
00:30 Pixionus It's very consistent. I always see all the minions running the older version report back before those running 7.x
00:31 timoguin joined #salt
00:32 Vynce hm.  well, my excludepat didn't work, but i'm not sure why.
00:35 Pixionus iggy, basepi do you guys or the person currently checking into this issue know about this?  If not maybe I should create a ticket?
00:35 Vynce salt docs seem o jsut say "E@" means regex, without getting into details of which regex. anybody happen to know?  POSIX?  PCRE?
00:35 iggy python re
00:36 iggy see the python docs for more info
00:36 iggy Pixionus: I skipped 2014.7... nfc
00:37 iggy 2015.2 seems faster than 2014.1
00:37 Pixionus The servers most likely to be forgotten or time out (both it looks like) are all of the 2014.7.x flavor
00:37 Pixionus was that already released or used internally?
00:38 iggy not released yet
00:38 iggy but wait
00:38 iggy are you using 2 different minions versions with the same master?
00:38 Pixionus yes.
00:38 iggy make sure you put that in the issue you open so they can close it immediately
00:38 Pixionus It was working so we didn't fix it.
00:39 Pixionus So if I open an issue like this that alone is reason to dismiss it?
00:39 moapa joined #salt
00:39 Pixionus Seems to be diagnostic information that may or may not have been missed so figured I'd offer it up.
00:40 iggy it's not suggested to run old minions with new masters (and it's completely unsupported to use new minions with old masters)
00:40 Pixionus I thought Salt Masters were backwards compatible.  I could have sworn I read that somewhere.
00:41 jerematic joined #salt
00:41 Pixionus Nah.  Masters are updated
00:41 Pixionus Do you work at Salt iggy?
00:42 Pixionus or are you just a supper helpful user?
00:43 iggy mmmm, supper
00:43 iggy No, I couldn't handle the Utah weather.
00:46 Pixionus Mind if I ask where you work?  You sound to me to be an early adopter so I assume you guys use Salt at work
00:47 iggy Small Software company in Houston (OnCenter.com for the morbidly curious). We do use it at work. I wouldn't say we were early adopters (we just launched our product 3 months ago and I've only been here for 6 months)
00:48 iggy that was a brutal first 3 months
00:48 bluenemo_ joined #salt
00:49 Pixionus ouch.
00:49 Pixionus Both the brutal months and the Houston
00:49 Pixionus I lived in Houston most of my life and I am glad to be out of that humidity.
00:50 Pixionus The mosquitoes alone...  Ended up moving to Az and loving it so far though I miss the rain.
00:50 iggy ehh, every place is going to have +/-'s
00:50 Pixionus yeah.  I just found the negatives out here less negative.
00:51 Pixionus Culture out here seems more startup friendly too, and I love having mountains, desert, and forest in reasonable distance to go camping in.
00:51 anotherZero joined #salt
00:52 Pixionus But to each his own.  My parents will NOT be leaving Houston any time soon and neither will many of my friends.
00:53 Pixionus Though strangely enough a good number of them seem to have been moving to Huntsville of all places.
00:53 Pixionus Didn't even know there was a job market there.
00:53 iggy I've had offers other places... Not really a fan of Cali. I almost did NYC, but it was a pay cut (factoring CoL). Not a lot of places I'd really like to move
00:53 iggy New Orleans, but that's got the same climate for the most part
00:53 Pixionus Yeah.
00:54 gar1 joined #salt
00:54 Pixionus Don't think I could do it myself.
00:55 Pixionus Thanks for often being willing to chime in on problems we have.  All this time I just assumed it was part of where you worked :P
00:55 iggy #1 mistake
00:55 Pixionus mine or yours? :P
00:55 Pixionus helping or assuming
00:55 iggy lol... both
00:56 Pixionus lol
00:56 iggy but it is 8pm, so... *poof*
00:56 Pixionus (and I did find a ticket that talks about what I was finding so that's resolved of sorts)
00:56 Pixionus ah yes.  forget about the time gap.  night
00:57 iggy I'm just leaving the office :/
00:57 desposo joined #salt
01:01 schristensen joined #salt
01:02 Pixionus :/
01:06 yomilk joined #salt
01:08 aqua^mac joined #salt
01:14 p0rkbelly joined #salt
01:21 bhosmer__ joined #salt
01:25 ocdmw joined #salt
01:30 MatthewsFace joined #salt
01:34 mafro joined #salt
01:48 viq joined #salt
01:48 viq joined #salt
01:48 Singularo joined #salt
01:51 shaggy_surfer joined #salt
01:53 otter768 joined #salt
02:09 josephleon joined #salt
02:13 yawniek_ joined #salt
02:19 donmichelangelo joined #salt
02:20 seblu joined #salt
02:24 Singularo joined #salt
02:29 catpiggest joined #salt
02:34 forrest joined #salt
02:36 jaloren iggy: salt.modules.state.running(concurrent=False) Return a dict of state return data if a state function is already running. This function is used to prevent multiple state calls from being run at the same time.
02:38 aparsons joined #salt
02:40 smcquay joined #salt
02:41 jaloren That would be my problems
02:45 evle joined #salt
02:47 ilbot3 joined #salt
02:47 Topic for #salt is now Welcome to #salt | SaltConf 2015 is Mar 3-5! http://saltconf.com | 2014.7.2 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/
02:51 ocdmw joined #salt
02:52 ocdmw joined #salt
02:55 pravka joined #salt
03:06 favadi joined #salt
03:09 ajw0100 joined #salt
03:19 XenophonF joined #salt
03:19 XenophonF left #salt
03:20 neogenix joined #salt
03:30 mosen joined #salt
03:30 jxFive joined #salt
03:35 TyrfingMjolnir joined #salt
03:38 ocdmw joined #salt
03:42 bash1245_ joined #salt
03:43 zzach joined #salt
03:45 XenophonF joined #salt
03:54 otter768 joined #salt
03:58 JlRd joined #salt
04:00 mgw joined #salt
04:01 hasues joined #salt
04:01 hasues left #salt
04:08 tmwsiy joined #salt
04:09 XenophonF left #salt
04:19 robot_hands joined #salt
04:28 jalbretsen joined #salt
04:32 mafro joined #salt
04:43 garthk joined #salt
04:45 Furao joined #salt
04:50 kermit joined #salt
05:02 thehaven joined #salt
05:05 balltongu_ joined #salt
05:08 mafro /quit/quit
05:20 pravka joined #salt
05:24 felskrone joined #salt
05:31 Furao look like SuSE is switching to salt
05:36 cberndt joined #salt
05:38 linjan joined #salt
05:40 aparsons joined #salt
05:43 jalbretsen joined #salt
05:43 mosen Furao: nice
05:44 ramteid joined #salt
05:44 aw110f joined #salt
05:45 aparsons_ joined #salt
05:46 zatricky joined #salt
05:50 zatricky Hey all. Is there a set of guidelines for handling "side effect" package changes when uninstalling a package from a minion? For example: If, via salt, you install a package that causes a conflicting package to be disabled/uninstalled - but later on decide to remove the installed package: Can you reliably return that other package to the original state? Is there a guideline for this scenario?
05:54 malinoff joined #salt
05:55 otter768 joined #salt
06:01 iggy nope, not really
06:01 iggy redeploy and call it a day
06:06 pravka joined #salt
06:12 zatricky Thanks, iggy
06:14 zatricky Sometimes "redeploy" isn't a good solution though. For VMs ... easy. For hundreds of off-site Raspberry Pis ... ;)
06:14 zatricky Its easy to sort this one out - but was hoping there was a guideline for doing it in the formula definitions. I guess not. :)
06:20 seshan joined #salt
06:27 andygrunwald joined #salt
06:39 Vynce joined #salt
06:45 Vynce joined #salt
07:04 stoogenmeyer joined #salt
07:05 AndreasLutro joined #salt
07:06 pravka joined #salt
07:07 colttt joined #salt
07:16 che-arne joined #salt
07:21 mattiasr joined #salt
07:26 favadi joined #salt
07:26 joshin joined #salt
07:26 krzysztof joined #salt
07:27 krelo joined #salt
07:30 seshan joined #salt
07:34 Furao joined #salt
07:34 flyboy joined #salt
07:38 clintberry joined #salt
07:40 colttt_ joined #salt
07:40 krelo joined #salt
07:46 Linuturk joined #salt
07:47 toanju joined #salt
07:48 jtang joined #salt
07:49 kawa2014 joined #salt
07:49 che-arne joined #salt
07:52 linjan joined #salt
07:53 Auroch joined #salt
07:56 cberndt joined #salt
07:56 otter768 joined #salt
08:01 hebz0rl joined #salt
08:02 KermitTheFragger joined #salt
08:04 cberndt joined #salt
08:06 bash1245_ joined #salt
08:06 jri joined #salt
08:08 cberndt joined #salt
08:09 eseyman joined #salt
08:23 lietu joined #salt
08:23 markm joined #salt
08:31 Andre-B joined #salt
08:36 kawa2014 joined #salt
08:39 trikke joined #salt
08:47 lb1a joined #salt
08:47 Furao_ joined #salt
08:48 Vynce joined #salt
08:49 intellix joined #salt
09:07 Xevian joined #salt
09:09 \ask joined #salt
09:13 devweasel joined #salt
09:16 losh joined #salt
09:16 lietu joined #salt
09:21 MatthewsFace joined #salt
09:22 illern joined #salt
09:22 PI-Lloyd joined #salt
09:22 N-Mi joined #salt
09:24 fredvd joined #salt
09:32 huddy joined #salt
09:34 Vynce joined #salt
09:36 illern joined #salt
09:43 stoogenmeyer_ joined #salt
09:43 agend joined #salt
09:44 devweasel left #salt
09:46 JlRd joined #salt
09:50 iromli joined #salt
09:57 otter768 joined #salt
09:57 I3olle joined #salt
10:03 giantlock joined #salt
10:06 andreyst joined #salt
10:13 andygrunwald joined #salt
10:14 aquassaut joined #salt
10:20 pkruithof joined #salt
10:28 Furao joined #salt
10:30 devweasel joined #salt
10:31 che-arne joined #salt
10:37 midihenry joined #salt
10:38 midihenry Hello
10:38 PI-Lloyd morning
10:38 damon_ joined #salt
10:39 midihenry I've ran into a problem with the salt virt module
10:40 midihenry I've been trying to spin up VMs but it always uses the br0 default network configuration although I changed the default configuration using pillars
10:41 midihenry I found a similar issue on github but no solution
10:42 ladistwo joined #salt
10:43 ladistwo left #salt
10:46 amcorreia_ joined #salt
10:46 andygrunwald joined #salt
10:49 midihenry somehow it is not aware of the configuration
10:49 midihenry I hope someone has the anwser
10:50 midihenry Thanks in advance
10:53 babilen Which issue do you refer to?
10:54 midihenry issue #11345
10:56 bluenemo joined #salt
10:56 bluenemo joined #salt
10:56 evle1 joined #salt
10:57 gar joined #salt
10:57 gar joined #salt
11:01 jespada joined #salt
11:02 flebel joined #salt
11:12 wnkz joined #salt
11:12 che-arne joined #salt
11:13 gar left #salt
11:14 I3olle_ joined #salt
11:17 che-arne joined #salt
11:20 toanju joined #salt
11:21 toastedpenguin joined #salt
11:22 che-arne joined #salt
11:26 hobakill joined #salt
11:27 che-arne joined #salt
11:28 johtso joined #salt
11:28 kiorky joined #salt
11:29 \ask joined #salt
11:30 jeddi joined #salt
11:36 illern joined #salt
11:37 che-arne joined #salt
11:44 che-arne joined #salt
11:45 jhauser joined #salt
11:58 otter768 joined #salt
12:04 mikkn joined #salt
12:10 illern_ joined #salt
12:15 opennode joined #salt
12:20 jri joined #salt
12:26 illern joined #salt
12:27 BtbN joined #salt
12:28 BtbN Hello, is it possible to allow a minion to send commands to other minions? I want to use salt for icinga checks, but i wonder if icinga has to be the salt master, or if icinga can be another minion.
12:31 I3olle joined #salt
12:32 yomilk joined #salt
12:32 Ahrotahntee BtbN: salt commands, or shell commands?
12:33 BtbN salt commands.
12:33 BtbN At least i think so, just started setting this up.
12:33 Ahrotahntee publish.publish would work
12:33 Ahrotahntee it lets one minion target and invoke salt commands on another
12:33 BtbN Currently our icinga is using ssh to monitor everything, which is getting more and more inefficient
12:39 aphoriser joined #salt
12:42 dendazen joined #salt
12:43 aphorise joined #salt
12:44 andygrunwald joined #salt
12:46 aphoriser joined #salt
12:48 cpowell joined #salt
12:51 jeffspeff joined #salt
12:53 illern joined #salt
12:53 aphorise joined #salt
12:53 peters-tx2 joined #salt
12:54 TyrfingMjolnir joined #salt
12:55 overyander joined #salt
12:56 denys joined #salt
12:58 subsignal joined #salt
12:58 MatthewsFace joined #salt
12:58 mariusv joined #salt
12:58 mariusv joined #salt
13:02 mariusv joined #salt
13:02 mariusv joined #salt
13:02 mariusv joined #salt
13:02 mariusv joined #salt
13:03 nexsja joined #salt
13:03 subsignal joined #salt
13:06 bhosmer joined #salt
13:06 illern joined #salt
13:06 subsignal joined #salt
13:06 pravka joined #salt
13:09 I3olle joined #salt
13:09 Roee joined #salt
13:10 bhosmer joined #salt
13:10 Roee hi, does someone knows if there is an option to create an AMI for instance using salt-cloud
13:10 Roee what is the command ?
13:15 jdesilet joined #salt
13:16 edrocks joined #salt
13:16 I3olle_ joined #salt
13:19 rayha joined #salt
13:20 perfectsine joined #salt
13:25 jespada joined #salt
13:27 mpanetta joined #salt
13:28 scarcry joined #salt
13:30 dunz0r Hmm... if I want to match on ipcidr in a jinja if-case... how would I do that?
13:31 dunz0r I've considered grains['ipv4'] == "172.*" but that seems dirty
13:31 racooper joined #salt
13:31 amcorreia_ joined #salt
13:33 ja-s joined #salt
13:35 jerematic joined #salt
13:37 MWheelz joined #salt
13:40 Deevolution dunz0r: Have you considered a custom grain to wrap your match criteria?
13:41 amcorreia_ joined #salt
13:42 yuhl_work_ left #salt
13:42 yuhl_work_ joined #salt
13:42 yuhl_work_ how do I do to have a watch on a file that is not managed by salt. I'd like to have the watch triggering when the content of the file (not managed by salt) change.
13:42 yuhl_work_ Is there some. file.change ??
13:43 hybridpollo joined #salt
13:48 yuhl_work_ In my case, I try to make a script  working on /etc/fstab that execute only if the /etc/fstab has changes or if the script itself changed.
13:49 Stucaz joined #salt
13:49 racooper sounds like you're looking more for a file monitor than file management. something like zabbix can do what you're asking for...
13:49 cmcmacken joined #salt
13:50 Tauro joined #salt
13:51 edrocks is the latest salt-minion version 2014.7.1? My maaster is 2014.7.2 but I cant seem to get my minions updated
13:52 dunz0r Deevolution: You mean setting some grain based on the subnet?
13:52 hasues joined #salt
13:54 hasues left #salt
13:56 dunz0r I'm contemplating using the salt.utils.network.in_subnet function, but I'm not sure it's avaliable in a jinja-template
13:59 otter768 joined #salt
13:59 Furao joined #salt
13:59 kaptk2 joined #salt
14:00 josephleon joined #salt
14:01 hybridpollo joined #salt
14:01 perfectsine joined #salt
14:01 wooks joined #salt
14:02 timoguin joined #salt
14:03 lietu- joined #salt
14:04 aphoriser joined #salt
14:05 dunz0r I /could/ of course just add a new state to my top.sls, but I don't want to crowd it with exceptions
14:08 lothiraldan joined #salt
14:09 steffo joined #salt
14:10 che-arne joined #salt
14:11 froztbyte hmm
14:11 Deevolution dunz0r: That is what I mean.  I'm not currently doing that for CIDR, but I do something similar to indicate which interface is the 'primary' one on multi-interface machines.
14:11 froztbyte so I've been trying to make time for this for nearly 2~3 months now, finally getting to it
14:11 egil joined #salt
14:11 ltsampros Hello
14:11 froztbyte question: how sane is it to potentally have multiple pillars/states per module type?
14:12 ltsampros are there any plants for ssh_async to get implemented? Is this on the roadmap?
14:12 froztbyte like, I want to make a bunch of reusable modules ('users', 'webserver', etc, for example), and then import these as git submodules into salt repos (that have to be separate by necessity)
14:13 froztbyte but it seems like I'd have to put n-many loading paths into the various top files to do that
14:13 froztbyte guess I just need to play around for a while and see
14:15 edrocks iggy: It seems my minions didnt update the ip after 30min. But i found out you can restart the minion with `sudo service salt-minion restart` so I didn't need to restart each instance
14:16 \ask joined #salt
14:16 wincyj joined #salt
14:20 Tauro left #salt
14:21 rayha left #salt
14:22 clintberry joined #salt
14:23 andrew_v joined #salt
14:24 amcorreia_ joined #salt
14:27 triplecorn joined #salt
14:27 TheoSLC joined #salt
14:30 bluenemo joined #salt
14:30 bluenemo joined #salt
14:31 pdayton joined #salt
14:33 murrdoc joined #salt
14:37 triplecorn joined #salt
14:37 neogenix joined #salt
14:38 dyasny joined #salt
14:40 jdesilet joined #salt
14:42 steffo joined #salt
14:43 ltsampro` joined #salt
14:45 rayha joined #salt
14:47 v0rtex does anyone know if the talks/slides from saltconf are up in a centralized location somewhere?
14:47 MatthewsFace joined #salt
14:48 v0rtex and especially links to code and examples (for those speakers who saved that)
14:53 funzo joined #salt
14:54 conan_the_destro joined #salt
14:59 esogas_ joined #salt
15:01 opennode joined #salt
15:02 rojem joined #salt
15:05 ghanima joined #salt
15:06 ghanima hello has anyone successfully configured raet with saltstack
15:08 babilen v0rtex: They have not been published in a coherent manner. Some speakers made their slides available, but I think that saltstack's policy is to not publish slides and videos during/shortly after the conference to get more people to attent the conference itself (monies monies monies)
15:08 zwi joined #salt
15:08 babilen ghanima: Be the first! (but I would think that, at least, the developer managed to do it)
15:08 brick_ joined #salt
15:09 _JZ_ joined #salt
15:09 ghanima babilen: I wouldn't mind going through the paces but from what I can tell it doesn't look to be included in 2014-7-1-1 from epel
15:09 ghanima I am getting python import error on my salt master
15:09 babilen Wasn't it a feature preview in 2017.7 ?
15:10 v0rtex babilen: yeah, I wondered. I'm sure it's also not easy to try and get it all together. It's just unfortunate when my notes weren't good enough on certain examples I wanted to refer back to (especially due to my laptop dying with the lack of power at the tables)
15:11 babilen v0rtex: I haven't even been able to attend the conference (flying over from Europe was not an option this time) and would (have) love(d) to watch the videos, but I guess that we'll have to wait a bit longer (6 months) before Saltstack decide to publish them
15:11 babilen ghanima: What exactly did you do, what did you expect to happen and what is the exact error you see? Could you use a pastebin such as http://refheap.com to provide that information?
15:12 murrdoc u guys want links
15:12 murrdoc to slides
15:12 babilen Do you have some?
15:12 babilen (or videos)
15:12 murrdoc videos base pi said that should be on youtube asap
15:12 murrdoc they contracted it out
15:12 babilen Do we have to join #salt-backalley ?
15:12 murrdoc but i can get u some slide links
15:13 babilen oh, ASAP sounds nice
15:13 murrdoc http://www.slideshare.net/stevendgonzales/saltconf2015-saltstack-at-scale-automating-your-automation
15:13 v0rtex murrdoc: I'm definitely interested
15:13 murrdoc thats the emc talk about how they formularize their stuff
15:13 murrdoc good read
15:13 ghanima https://www.refheap.com/98310
15:13 ghanima babilen: https://www.refheap.com/98310
15:14 murrdoc https://github.com/saltstack/salt/blob/develop/salt/engines/sqs_events.py https://github.com/saltstack/salt/blob/develop/salt/engines/__init__.py
15:14 murrdoc this is some cool shit that thatch talked about
15:14 babilen ghanima: http://docs.saltstack.com/en/latest/topics/transports/raet/index.html#using-raet-in-salt do you have those packages in the repos ?
15:15 v0rtex murrdoc: I was just digging through that code yesterday - pretty cool
15:15 murrdoc babilen:  http://www.slideshare.net/SaltStack/spot-trading-salt-stack-devops-summit from the main man ali
15:15 babilen ta!
15:15 murrdoc gareths kickass scheduler talk https://github.com/garethgreenaway/presentations/blob/master/SaltConf2015-Dr-SatlStack.pdf
15:16 murrdoc cloudflares kickass talk http://www.slideshare.net/lovenyberg/using-saltstack-to-orchestrate-microservices-in-application-containers-at-saltconf15
15:16 murrdoc thats all i have
15:16 murrdoc the linkedin talk is out there too
15:16 murrdoc i dont have that link handy
15:17 murrdoc yw
15:17 v0rtex nice, thanks murrdoc
15:18 murrdoc thatch is crazy smart
15:18 v0rtex I actually missed those talks so it's nice to be able to review them
15:18 babilen ghanima: You might have to install some of this stuff from pypi
15:18 murrdoc like srsly i have sat through dev talks before
15:18 murrdoc but the raet talk from thatch was interesting, i didnt nod off at all
15:18 ghanima babilen: I don't see a reference to the RPM exactly but there are two version references on the page 2014.7.0 and 2015.2.0rc1 and 2014.7.2 I have 2014.7.1 installed
15:19 esogas_ joined #salt
15:19 v0rtex murrdoc: did you go to the training before or just his conference talk?
15:20 ghanima babilen: I am little confused because according to the article it was introduced in 2014.7.0
15:20 murrdoc conference man, got no time for training :|
15:20 ghanima babilen: I would assume that would be the python module is installed
15:20 v0rtex haha
15:20 v0rtex I went to the training and it was something like 2.5 hours of in-depth raet/ioflo
15:20 v0rtex and digging into the code
15:21 v0rtex crazy stuff
15:21 v0rtex I don't think I would have ever had a true grasp on flow-based programming if not for his explanations there
15:24 jimklo joined #salt
15:28 analogbyte joined #salt
15:29 debian112 joined #salt
15:29 clintberry joined #salt
15:30 lietu joined #salt
15:31 crashmag joined #salt
15:33 iggy babilen: it shouldn't take as long to get the videos out this year... they actually hired a company to do all the editing, etc.
15:34 babilen \o/
15:35 iggy yeah, I think last year, rhett did it all himself
15:37 ghanima anyone else have any experience raet
15:38 iggy it next on my list
15:38 murrdoc raet about now
15:38 iggy I wouldn't bother trying it with 2014.7
15:38 Ahlee anybody got a slick trick for getting output on one line?  I want minion name, file name, md5sum on one line after running cmd.run md5sum /foo
15:39 rocket joined #salt
15:39 iggy in 0.17.ahlee? no clue
15:39 ghanima iggy: I am trying resolve an issue with with minion not reporting back within a certain amount of time. I have hosts spread across multiple datacenters with a total of 1100 nodes. and trying to see if I can optimize the transport
15:39 ghanima sorry that meant to say minion... was trying to how performance raet would be
15:39 murrdoc —out =json | .jq
15:39 iggy ghanima: okay... I still wouldn't try raet with 2014.7
15:40 rocket so for a gitfs backend do you have to do anything else special for the environments to work?  I am not able to get my develop environ to work properly
15:40 rocket keep getting a message that there is no top file
15:40 iggy rocket: does "base" work?
15:41 Ahlee murrdoc: thank you!
15:41 murrdoc :)
15:41 murrdoc whats this geekon thing
15:42 murrdoc how do i attend it ahlee
15:42 ghanima iggy: sorry is it ready for use looking at the documentation there seem to be some client additions that are pending but from what I read its enabled
15:42 analogbyte joined #salt
15:42 rocket iggy: I switched my node to not be part of a default environment.
15:43 rocket iggy: I see this error now.  No matching salt environment for environment 'develop' found
15:43 Ahlee murrdoc: i'll fill you in when i have details
15:43 analogbyte joined #salt
15:45 Ahlee ah, i was mistaken.  It's geekfest, not geekon
15:45 analogbyte joined #salt
15:48 analogbyte joined #salt
15:49 murrdoc http://www.meetup.com/Geekfest/ ?
15:49 Ahlee aye
15:50 aparsons joined #salt
15:51 analogbyte joined #salt
15:51 evle joined #salt
15:52 smcquay joined #salt
15:52 evle1 joined #salt
15:53 ghanima iggy: just curious the 2015 release is available do you think its more mature in that release
15:53 CouperAlex joined #salt
15:53 iggy yes
15:54 iggy rocket: I don't use environments... and now I'm starting to remember why
15:54 CouperAlex Hey. SaltStack Noob here - just starting to look into saltstack for the first time in a couple of years. *SO* many awesome features.
15:55 murrdoc iggy:  get a ssce stamp
15:55 iggy because anybody where I work would know what that meant
15:55 rocket I think I figured out my issue .. my git push didnt send --all initially
15:56 rocket thanks anyway
15:56 rojem joined #salt
15:56 CouperAlex Getting an issue where upon applying a very simple state file I get an error on the master saying "Data failed to compile", though looking at the minion it seems to have run fine.
15:58 CouperAlex https://gist.github.com/alexcouper/ebd4ececd3e34668670b
15:59 iggy CouperAlex: do you have 2 minions running?
15:59 lietu- joined #salt
15:59 CouperAlex only one
15:59 iggy it looks like you have a minion running in the foreground with -l debug, make sure there's not another copy running in the background
16:00 otter768 joined #salt
16:00 jespada joined #salt
16:01 CouperAlex iggy: awesome debugging. works
16:01 CouperAlex well spotted.
16:02 murrdoc lazyloader ?
16:03 iggy it's new
16:03 iggy and hopefully I hit _every_ bug with it for you guys
16:04 thedodd joined #salt
16:05 murrdoc yeah i saw the linkedin talk, i didnt realize it was in 2014.7
16:06 giantlock joined #salt
16:09 sjwoodr joined #salt
16:11 StDiluted joined #salt
16:14 tligda joined #salt
16:14 balltongu joined #salt
16:15 esharpmajor joined #salt
16:15 kormoc joined #salt
16:15 techdragon joined #salt
16:15 djinni` joined #salt
16:15 Vye joined #salt
16:15 thehaven_ joined #salt
16:15 wnkz joined #salt
16:15 adrianhannah joined #salt
16:16 packeteer joined #salt
16:16 mt joined #salt
16:16 GrueMaster joined #salt
16:16 ThomasJ joined #salt
16:16 chutzpah joined #salt
16:16 analogbyte joined #salt
16:16 kivihtin joined #salt
16:17 danielcb joined #salt
16:17 xenoxaos joined #salt
16:18 manytrees joined #salt
16:18 davedash joined #salt
16:19 iggy was it? I didn't realize that either
16:19 JonGretar joined #salt
16:19 natewalck joined #salt
16:19 akoumjian joined #salt
16:20 codekobe joined #salt
16:20 elfixit joined #salt
16:20 collinanderson joined #salt
16:20 zipkid joined #salt
16:20 cmcmacken joined #salt
16:23 ja-s joined #salt
16:23 pravka joined #salt
16:25 mpanetta_ joined #salt
16:29 igorwidl joined #salt
16:29 pravka joined #salt
16:29 KyleG joined #salt
16:29 KyleG joined #salt
16:30 aparsons joined #salt
16:32 bluenemo joined #salt
16:32 shaggy_surfer joined #salt
16:34 analogbyte joined #salt
16:34 jay_d joined #salt
16:34 shaggy_surfer joined #salt
16:37 MatthewsFace joined #salt
16:37 analogbyte joined #salt
16:38 MatthewsFace joined #salt
16:38 sarlalian joined #salt
16:38 analogbyte joined #salt
16:39 analogbyte joined #salt
16:39 sarlalian joined #salt
16:42 signull iggy, murrdoc: either of you have experience with the dockerio state?
16:42 hybridpollo joined #salt
16:43 signull im trying to utilize the new docker.latest_running however I am getting:  Comment: State 'docker.latest_running' was not found in SLS 'some_sls'
16:43 iggy paste code
16:44 signull the state.sls ?
16:44 signull sure let me sanitize first
16:44 iggy the more the merrier
16:46 Auroch joined #salt
16:48 signull iggy: http://pastebin.com/AhP8rt2F
16:50 murrdoc iggy u done any reboot based on onchages flag
16:50 murrdoc or anything like
16:50 iggy reboot? no
16:50 iggy I don't reboot... redeploy
16:51 murrdoc ok cool boy
16:51 Ahlee murrdoc: we atd reboot with -order: last
16:52 Ahlee although that isn't working since some HPs won't actually disable hyperthreading on a warm boot, gotta power it down
16:52 iggy isn't that what listen gets you (not that you'd know about listen ;)
16:53 Ahlee we started toying with orchestrate to power down and power up via iLO over a proxy minion
16:53 Ahlee but then we realized we weren't making it easier
16:53 murrdoc Ahlee:  i have the case for listen_in + order: last
16:53 murrdoc and a case for reboot now
16:54 murrdoc trying to figure out both
16:54 murrdoc if u use listen_in + order last, it event.fires the reactor
16:54 murrdoc which reboots the server
16:54 Ahlee yeah i don't get listen_in here
16:54 Ahlee oh
16:54 murrdoc listen_in fires the state at the end of run
16:55 Ahlee yeah, This example will cause apache2 to be restarted when the apache2.conf file is changed, but the apache2 restart will happen at the end of the state run.
16:55 Ahlee interesting
16:55 Ahlee meh
16:55 I3olle joined #salt
16:56 dendazen if i create my salt state on mac, is there a way to check syntax against it before i push the change to the git repo?
16:56 murrdoc meh ?
16:56 robawt dendazen: run it on a VM :)
16:56 robawt vagrant + virtualbox are awesome for checking states
16:56 iggy at doesn't sound like a terrible idea, cmd.run "shutdown +2 'salt initiated' &"
16:57 dendazen yeah, that’s what i ma currently doing.
16:57 Ryan_Lane joined #salt
16:57 dendazen just thought to eliminate that step
16:57 dendazen and check natively
16:57 dendazen on mace
16:59 Gareth morning morning
16:59 signull iggy: any idea?
16:59 iggy signull: what version of salt?
17:00 signull iggy: this mornings build
17:00 signull for 2015.2
17:00 iggy do me a favor
17:00 iggy go file a bug that they didn't version annotate that function
17:01 iggy it's only in devel
17:01 signull ahh ok
17:01 murrdoc what funtion
17:01 murrdoc morning gareth
17:02 murrdoc this whole reboot servers during state.hightstate is fucking fantastic problem
17:02 murrdoc imho
17:02 pravka joined #salt
17:03 murrdoc anyone got thoughts ?
17:03 * murrdoc holds out pennies
17:03 jespada joined #salt
17:04 iggy I like at
17:04 signull iggy: https://github.com/saltstack/salt/issues/21547
17:04 murrdoc signull:  u can easily fix the documentation
17:04 murrdoc step up to the plate!
17:04 iggy NO
17:04 iggy TEACH THEM A LESSON
17:05 iggy public shaming is the best teacher
17:05 murrdoc hahah
17:05 solvik left #salt
17:05 signull murrdoc: there are so many docs that need updating. ive gone through so many modules and states and found hidden optons and functions that arent documented
17:05 * iggy can't wait to have iggy brood
17:05 TheoSLC joined #salt
17:05 murrdoc signull:  fix it for the love of salt man
17:05 murrdoc u ll make salt better
17:06 murrdoc or iggy it and complain and do it
17:06 murrdoc dont murrdoc and do nothing and ask other people to do it
17:06 murrdoc gawd
17:06 heyya joined #salt
17:06 * murrdoc paints himself in to a corner
17:07 Gareth murrdoc: hey :) hows it going?
17:07 * signull thinks saltstack dev team, or someone who is a better python programmer then himself. That there needs to be a LINT to look for non hidden functions (* not prepended with an underscore _privFunction ) must be in the documentation or the build fails
17:07 aphorise joined #salt
17:08 iggy any new function should have version annotations too (trying to make build green)
17:08 mgw joined #salt
17:09 lietu joined #salt
17:10 signull ehh well: https://github.com/saltstack/salt/issues/21550
17:10 murrdoc Gareth:  good man, trying to figure out how to implement reboot now and reboot-at-end-of-state-run for dells
17:12 Gareth murrdoc: so like a do stuff....reboot...then continue where the run left after the reboot?
17:12 murrdoc except some states need to reboot 'onchanges'
17:12 murrdoc and some can wait till end of state run
17:13 murrdoc similar to how failhard runs
17:14 perfectsine joined #salt
17:14 murrdoc and then use this man page, http://docs.saltstack.com/en/latest/ref/runners/all/salt.runners.state.html, as a how to work with reactor to reboot minion
17:14 murrdoc thoughts Gareth
17:15 pravka joined #salt
17:17 Gareth Interesting idea.
17:19 linjan joined #salt
17:19 andrew_v joined #salt
17:21 ejk joined #salt
17:22 mpanetta joined #salt
17:23 shaggy_surfer joined #salt
17:28 elfixit joined #salt
17:28 jtang joined #salt
17:29 murrdoc cool https://github.com/felskrone/salt-eventsd
17:30 spookah joined #salt
17:31 germs_ joined #salt
17:31 pravka joined #salt
17:32 neogenix joined #salt
17:34 jespada joined #salt
17:37 Vynce joined #salt
17:39 jtang joined #salt
17:41 ksalman hm.. is salt not supported on ubuntu raring anymore? i have added the repo but i don't see new versions of salt available
17:42 forrest joined #salt
17:44 iggy raring is EoL'ed
17:44 murrdoc what u talking about
17:44 ksalman right
17:44 murrdoc its raring to go
17:45 ksalman that does not mean that salt is unsupported on it
17:47 evilrob I'm writing a state for a mysql_grant, and having issues with the host parameter.  The docs say the host param indicates which network/host the grant should apply to, but when trying to actually apply the grant, the state fails with a "TypeError: 'bool' object is not iterable".  It looks (from the code) like it's trying to connect to the host listed in the host param instead of using that value in the grant.
17:48 iggy ksalman: that means that people aren't going to bother making packages for it (and indeed I don't think launchpad supports building for it anymore)
17:48 iggy feel free to install from source/pip/whatever
17:49 forrest joehh: Are you still building the ubuntu packages for raring?
17:49 jdesilet joined #salt
17:49 forrest I'm hoping the answer is no.
17:49 ksalman i should still be able to install the last version that was built on it, no?
17:50 amcorreia_ joined #salt
17:50 evilrob this is with 0.17.5+ds-1
17:50 evilrob on ubuntu 14.04.1
17:51 iggy lol
17:51 evilrob https://paste.ee/p/qM0tO
17:51 iggy a_h_l_e_e over here with the 0.17
17:52 jdesilet joined #salt
17:53 johanek joined #salt
17:53 Heartsbane evilrob: you have me for 9 minutes, but happens if you remove your require
17:53 Heartsbane because that looks the same as my setups
17:54 Heartsbane I have a meeting in 7
17:55 nebuchadnezzar joined #salt
17:55 evilrob Heartsbane: I have a meeting at the tope of the hour too.  I'm working on updating my minion.  I just noticed it's a year old.
17:55 Ahlee This is awesome.  In a _state/ I'm calling __salt__['cmd.run_all']('mkdir -p /foo/bar/baz').  Retcode returns 0, but directory isn't being created
17:55 iggy the code looks correct (in the mysql state and module)
17:55 perfectsine_ joined #salt
17:55 thayne joined #salt
17:56 evilrob It's not the require.  I have a similar grant right above it without the host param and it works just fine
17:56 Heartsbane hmmmmm
17:56 evilrob and if I remove that host param, it also works.
17:56 Aidan_L joined #salt
17:57 aphoriser joined #salt
17:57 evilrob if I ran it from the command-line with "mysql.grant_add 'ALL' 'keystone.*' 'openstack' '%'" it works jsut fine, then that state (in paste.ee) runs and says the grant exists
17:58 evilrob and jsut to be complete, I removed the require:  same result
18:00 iggy check it with debug on, it'll output the full query generated, see if it looks right
18:01 otter768 joined #salt
18:03 wincyj joined #salt
18:03 Ahlee so __grant_generate, '{3!r}'.format(user) and '{4!r}'.format(host)
18:04 MatthewsFace joined #salt
18:04 Ahlee eh, i have nothing to add to this conversation, and need to get to the bottom of my cmd.run_all woes
18:04 Ahlee good luck evilrob
18:05 elfixit joined #salt
18:06 aphorise joined #salt
18:12 dendazen i get this error:
18:12 dendazen Detected conflicting IDs, SLS IDs need to be globally unique.
18:12 dendazen The conflicting ID is "sudo" and is found in SLS "base:packages.sudo.foreman" and SLS "base:packages.sudo
18:13 iggy what's the question?
18:13 dendazen so if i include ‘-packages’ for base *, and then do 'ops1.sec1.*':
18:13 dendazen - packages.sudo.foreman
18:13 dendazen it gives me an error
18:14 iggy so change the state id
18:14 wincyj joined #salt
18:14 dendazen what do you mean?
18:14 gmoro joined #salt
18:14 dendazen sorry new to salt
18:14 kbyrne joined #salt
18:14 rocket joined #salt
18:15 cmcmacken joined #salt
18:15 iggy the top level part (unindented) should be unique across every file you have
18:16 iggy so somewhere you have something like "sudo:\n  pkg.installed" and then somewhere else you have "sudo:\n  module.function"... you need to change the id of one of those
18:16 iggy try pasting your sls files somewhere
18:16 iggy protip: gist.github.com allows multiple files per paste
18:17 Andre-B_ joined #salt
18:17 dendazen you want top.sls?
18:17 dendazen or that state sls
18:17 dendazen for packages.sudo.foreman
18:17 iggy packages.sudo and packages.sudo.foreman
18:17 dendazen one sec
18:19 dalexander joined #salt
18:19 dendazen https://gist.github.com/anonymous/eba7170c188abc2b9be7
18:19 dendazen Thank you
18:20 dendazen oh i got you
18:20 iggy you have "sudo:\n  pkg.installed" in both files... do you really need it installed that many times?
18:20 dendazen i need to remove from my foreman
18:20 dendazen pckg installed
18:20 dendazen Thanks, sorry for stupid question
18:21 iggy no problem, it's a little confusing when getting started
18:21 flamin_scotsman joined #salt
18:21 dooshtuRabbit1 joined #salt
18:21 toanju joined #salt
18:21 dooshtuRabbit joined #salt
18:23 dendazen i guess it’s like in python when you try to import module which includes in upper module
18:23 dendazen behavior like
18:23 v0rtex using file.managed with an externally hosted file I have a tar.gz.sig file available but I'm guessing it's not simply a hash and instead is a pgp key since it's a bunch of binary data - is there any way to use that for my source_hash parameter?
18:25 ajw0100 joined #salt
18:26 MWheelz joined #salt
18:26 dendazen there is also sudoers_inc_path
18:26 dendazen which is duplicated
18:27 dendazen needs to be removed as well it seems
18:27 MTecknology We better get more of these really sexy books at saltconf next year!
18:28 murrdoc http://www.amazon.com/Learning-Saltstack-Colton-Myers/dp/1784394602
18:28 iggy that guy doesn't know anything
18:28 MTecknology damnit!
18:28 murrdoc true
18:28 * MTecknology was ninja'd
18:29 drawsmcgraw If I wanted to schedule an event (say, a service restart) across a cluster. But I wanted the event 'rolling' so that not all the nodes restart the service at the same time.
18:29 drawsmcgraw What's the best way to do that?
18:29 MTecknology orchestrate
18:29 Ryan_Lane drawsmcgraw: batch mode
18:29 murrdoc salt '*' cmd.run 'minion.reboot'
18:29 MTecknology or that
18:29 murrdoc or what ryan said
18:29 * basepi kicks murrdoc and iggy
18:29 drawsmcgraw I had thought orchestrate, but I think I was looking for batch_mode.
18:29 drawsmcgraw thanks Ryan_Lane!
18:30 Ryan_Lane yw
18:30 Ryan_Lane I think you can use batch in orchestrate
18:30 murrdoc hi basepi
18:30 murrdoc dont even need orchestrate, just str8 reactor is fine
18:30 * iggy whistles innocently
18:30 drawsmcgraw okay. It's just restarting a single service on a few nodes. I'm sure I can find a simple approach with those options
18:31 MTecknology basepi: I have my badge with your pi on it hanging up in my cube :)
18:31 iggy --batch=1 should probably work
18:31 basepi MTecknology: haha!
18:31 iggy the problem arises when you need time between the restarts (for clusters to regain quorum, etc.)
18:32 MTecknology I've now figured out how to migrate ldapd->sssd across 190 servers.   http://dpaste.com/0CDWPDS
18:32 stoogenmeyer_ joined #salt
18:32 MTecknology I'm sure I can coordinate that with orchestrate, but I might just do it by hand
18:32 drawsmcgraw iggy: good point. I'll keep that in mind. This is pretty small and fairly resilient
18:32 murrdoc it would be nice to have  a 'pillar' that did canaries
18:33 murrdoc like something that maintained state of where a salt state was applied and has a rollout mode
18:33 murrdoc basepi:  can u write something like that in this version
18:33 murrdoc and break it in the next one
18:33 murrdoc all <3
18:33 iggy salt --batch=1 '*cluster*' cmd.run "service cluster restart ; sleep 1m"
18:33 murrdoc nah
18:33 murrdoc i am thinking more
18:34 murrdoc salt batch some size
18:34 murrdoc if 90% of states are successful
18:34 murrdoc then do a bigger batch size
18:34 basepi hehe
18:34 murrdoc and a bigger batch size
18:34 v0rtex anyone see my message? I'm wondering if I have to bypass this whole process and cmd.run a wget instead to get around the source_hash requirement
18:34 iggy there definitely needs to be some better "rolling" functionality
18:34 iggy it seems like the last thing salt doesn't really handle well
18:35 murrdoc also basepi u should probably share http://smile.amazon.com/dp/1784394602/ as the url so that each purchase also goes to a charity
18:35 murrdoc imho
18:35 iggy v0rtex: you can't just download it and generate the source_hash once?
18:35 murrdoc v0rtex:  look at cp.cache_file
18:36 Ahlee iggy: Verio gave an interesting talk on their map reduce runner/returner that was effectively batch scheduling
18:36 perfectsine joined #salt
18:36 drawsmcgraw v0rtex: In my experience, I've just wget'd or curl'd the file down
18:36 murrdoc Ahlee:  yeah ?
18:36 v0rtex I suppose I could but I don't really want to have to do that for every version change
18:36 murrdoc got slides
18:36 Ahlee murrdoc: no, it was at saltconf
18:36 drawsmcgraw v0rtex: Any reason just cmd.run->wget won't work for you?
18:38 murrdoc damnit
18:38 v0rtex drawsmcgraw: it will work - I was just trying to make it properly stateful
18:38 evilrob well updating the salt master and salt minion to 2014.7.2 fixed my mysql problems, but broke all my requires
18:38 drawsmcgraw I can understand that :)
18:38 Ahlee generating the hash either remotely or downloading once and getting it is the salt-ish way
18:38 iggy evilrob: you really shouldn't need many requires
18:38 subsigna_ joined #salt
18:39 Ahlee i also wish you could pass the no_trust_me_i_know_waht_im_doing flag
18:39 Ahlee and have it ignore
18:39 iggy I have 4 in my entire state tree
18:39 v0rtex Ahlee: yes, I just thought I would like to take advantage of the already existing .sig file on the remote host
18:40 evilrob I put them in because of some things I'd read about the mysql modules.  I've pulled them.  I'm going to blow away the vagrant box and start from scratch and see how it goes.
18:40 iggy that would require uncommon tools to check
18:40 Ahlee v0rtex: pull requests welcome to extend file.managed to use it :)
18:40 murrdoc thats what cp.cache_file is for , trust me fetch it
18:40 v0rtex murrdoc: looking..
18:41 Ahlee murrdoc: looks like requires the file to be in the file_roots though?
18:42 v0rtex Ahlee, iggy: I do realize it requires gpg to be installed but if support only on systems that have gpg would be useful I could look at implementing and making a PR
18:42 drawsmcgraw To be clear, the 'splay' parameter means that the job will run at the time you specify, give-or-take the splay value, right?
18:42 drawsmcgraw http://docs.saltstack.com/en/latest/ref/states/all/salt.states.schedule.html
18:43 Ahlee drawsmcgraw: only verified is + splay
18:43 murrdoc ahale:  http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.cp.html#salt.modules.cp.get_url
18:43 murrdoc <— was wrong
18:44 drawsmcgraw Ahlee: That's fine.
18:44 iggy v0rtex: the problem is, you have to put that gpg check in the virtual function... so you end up losing the entire file module if you don't have gpg installed
18:44 drawsmcgraw So basically, it adds <number> seconds to your scheduled time, where <number> is what you give 'splay' ?
18:45 Ahlee i think it adds 0..random(splay)
18:45 v0rtex iggy: I see... well not a big deal
18:45 iggy drawsmcgraw: it should be time +/- splay/2 (iiuc)
18:45 drawsmcgraw iggy: Okay. Thanks
18:46 Rockj joined #salt
18:46 perfectsine joined #salt
18:47 debian112 hello all: I have a question about my config file. I have a DNS file that works on new CentOS 7 servers, but fails on Debian servers. Any idea on what I should be looking for: http://paste.debian.net/160785/
18:48 debian112 error: http://paste.debian.net/160786/
18:48 debian112 Basically it is failing at the if statement
18:49 debian112 or not reading the variables below the if statement
18:50 iggy I told you what was wrong with it yesterday
18:50 debian112 salt-minions are the same versions
18:50 debian112 what was that?
18:50 Ahlee heh
18:50 debian112 let me check salt archive
18:50 iggy missing salt['grains.get'] before all those ('nodename') bits
18:50 MTecknology Howto reload user/group cache after changes nsswitch ??
18:51 yidhra joined #salt
18:51 Ahlee MTecknology: nscd -i passwd, nsswitch itself doesn't cache
18:51 iggy reboot
18:51 debian112 iggy: yeah that works
18:51 debian112 but
18:52 debian112 what I sent works on new CentOS 7 servers
18:52 debian112 but fails on debian servers
18:52 Kelsar joined #salt
18:53 debian112 I was wondering if I should be looking at anything else
18:53 debian112 Weird to me!
18:53 iggy well, fix that first
18:54 MTecknology Ahlee: I mean within a running salt session. The system itself works perfect and great. The issue is that salt changes nsswitch after salt has already cached user/group data (using python-sss) and I'm changing nsswitch during the execution. That's causing some issues for me.
18:54 iggy I suspect your centos boxes are matching the first test (the actual grains lookup)
18:54 Ahlee MTecknology: interesting
18:54 MTecknology Running 'salt-call state.higstate; salt-call state.highstate' works, I'd just prefer do it once only
18:55 debian112 iggy: yeah that's the question I am asking. Why it might fail on debian servers
18:55 Ahlee MTecknology: double call just the part you need through a 'helper' state
18:55 iggy debian112: I had to look at it a few times to figure out why you weren't getting syntax errors
18:55 dharper_ii joined #salt
18:55 josephleon joined #salt
18:55 iggy debian112: I couldn't say without knowing more about your setup
18:56 Vynce joined #salt
18:56 MTecknology Ahlee: how would that fix the issue? My states depend on nsswitch.conf and such, the states failing /would/ be just fine if it weren't for the caching.
18:56 iggy debian112: but basically because none of your debian servers match salt['grains.get']('nodename').endswith('pc.greentoads.net') OR salt['grains.get']('nodename').endswith('conyers.simon.com')
18:56 Ahlee MTecknology: so you're saying for the length of the run, the incorrect values are cached?
18:56 MTecknology I'm pretty sure salt used to not cache user/group stuff, but it does because of some massive performance issues that resulted.
18:56 MTecknology yup
18:56 Ahlee ah
18:56 yidhra joined #salt
18:56 Ahlee i remember that thread, too
18:57 iggy MTecknology: turn off caching while your states are running?
18:57 Ahlee i'm unfamiliar with python-sss
18:57 Ahlee but, you're likely going ot need to make the caching configurable and disable it
18:57 jalbretsen joined #salt
18:57 MTecknology eh...
18:57 Ahlee which i'm sure thatch'll love.
18:58 Kelsar joined #salt
18:58 * MTecknology is responsible for python-sss being looked at in the first place :P
18:58 debian112 iggy: let me try trimming it down and see if it is because of that
18:58 MTecknology I didn't write it, but I did make my company pay the guy that did
18:58 _JZ_ joined #salt
18:59 evilrob yeah, only issue I'm having now is my ntp service is failing to restart the first time I run the state.  It's probably not running yet, so the restart is moot.
18:59 murrdoc simple scrum system reads like an oxymoroon
19:00 MatthewsFace joined #salt
19:00 jdesilet Is anyone on who is successfully using raet? I'm trying to get a lab setup and running but when I add: transport: raet to my server config and restart I get an error.
19:00 jdesilet http://paste.ofcode.org/x9EMVhFbsMnjArYfpgz4Sb
19:01 jdesilet I'm not sure if I'm missing a library or what
19:01 baweaver joined #salt
19:01 mimianddaniel joined #salt
19:03 MTecknology jdesilet: you have raet and (what were the salt crypto libs) libraries installed?
19:03 jdesilet I installed raet and libsodium
19:04 jdesilet that has been it for a clean install on CentOS 7 / 2014.7.2
19:04 mimianddaniel is it normal for syndic to reauth with its master every second  ???
19:04 jdesilet ioflo was already installed when I checked
19:04 MTecknology I'm pretty sure that should be all you need. I haven't done any testing at all with it.
19:04 iggy debian112: without the salt['grains.get'] , ('nodename').endswith('conyers') means "does the python tuple with an entry of 'nodename' endwith 'conyers'" which it never will
19:05 iggy jdesilet: libnacl
19:05 iggy or pynacl
19:05 iggy something like that
19:06 dyasny joined #salt
19:06 debian112 thanks iggy, let me investigate further, and report back!
19:06 iggy hmm, there's both libnacl and PyNaCl in pip
19:06 Ahlee oh crazy.
19:06 murrdoc libnacl ftw
19:06 jdesilet iggy: checking right now
19:06 Ahlee __salt__['cmd.run'] and __salt__['cmd.run_all'] behave very, very different.
19:07 forrest Ryan_Lane: are you around?
19:07 jdesilet libnacl is already installed and up to date
19:07 Ryan_Lane forrest: yeah. what's up?
19:07 apergos joined #salt
19:08 Ryan_Lane actually. back in 20 or so
19:08 Ryan_Lane lunch
19:08 forrest Ryan_Lane: How are you handling provisioning of the masterless minions for getting your actual unique salt state code onto the systems? Are you doing a package?
19:08 forrest Ryan_Lane: Sure
19:08 iggy jdesilet: the python module libnacl (not the c lib)?
19:09 jdesilet iggy: correct [root@salt-master salt]# pip install libnacl --upgrade
19:09 jdesilet Requirement already up-to-date: libnacl in /usr/lib/python2.7/site-packages
19:09 iggy okay, then I got nothing
19:10 dendazen what would be the best way to disable selinux? I do not want to run selinux module as it is only sets mode to permissive
19:10 dendazen should i just sed /etc/selinux/config
19:11 dendazen and put ‘disabled’
19:17 jcsp joined #salt
19:19 debian112 iggy, thanks my mistake. I can be slow at times
19:19 debian112 LOL
19:19 debian112 I got it working now
19:20 debian112 I need to keep using: salt['grains.get'] for each match. Instead of just one time
19:20 debian112 thanks
19:20 debian112 for your help
19:21 allmightyspiff dendazen: the command "setenforce 0" should work too
19:22 allmightyspiff although i think that might just put selinux in permissive mode....
19:22 jaloren joined #salt
19:22 jaloren murrdoc: you there?
19:23 dendazen Thanks
19:23 diegows joined #salt
19:23 camel_ joined #salt
19:24 camel_ having a small problem with saltstack, the question is listed here. http://superuser.com/questions/888427/why-is-oracle-java8-installer-pkgrepo-managed-failing-in-saltstack
19:24 camel_ related to oracle-java8-installer pkgrepo.managed failing
19:24 camel_ any help is appreciated
19:25 hybridpollo joined #salt
19:26 murrdoc jaloren:  i am
19:27 numkem Are files in the _pillar directory executed before or after the files put in the top.sls file?
19:29 iggy camel_: can you run that command manually on the server? can that server connect to keyserver.ubuntu.com?
19:30 camel_ trying that now.  Also, another weird thing about it, is that half the times I run state.highstate, it gives me those errors, and the other half, it returns nothing.
19:31 ckao joined #salt
19:32 XenophonF joined #salt
19:32 johanek joined #salt
19:33 iggy sounds like intermittent connection problems to the keyserver
19:33 iggy shitty firewall?
19:33 camel_ nah, super expensive firewall, and no other problems with it.
19:33 camel_ doesn't mean there isn't now
19:33 camel_ but i doubt it's that.
19:34 iggy could be an issue on the keyserver.ubuntu.com end (or anywhere in-between)
19:34 iggy but it definitely looks like a connectivity problem
19:34 camel_ on the server, I was able to run the commands and everything work fine.
19:35 desposo joined #salt
19:36 v0rtex tee hee: https://gist.github.com/ev0rtex/d70e1c454dfd47cd7cee#file-raet-sls
19:36 v0rtex who needs docs?
19:37 iggy camel_: so does the highstate run now?
19:37 tomh- joined #salt
19:37 ghanima joined #salt
19:38 dendazen in cmd.run
19:38 Ryan_Lane forrest: nope
19:38 dendazen what stateful: True means?
19:38 Ryan_Lane we generate tarball artifacts and stick them into S3
19:39 forrest Ryan_Lane: bah okay thanks
19:39 Ryan_Lane forrest: we use IAM policy to restrict which clusters can access which artifacts
19:39 forrest Yeah that's what I was looking at
19:39 MTecknology Hm.... how can I write this into a pillar? ..   *: git-rev: prod  ||  *-dev: git-rev: dev  ||  *-test: git-rev: test    hrm....
19:39 Ryan_Lane I'm not a huge fan of packages
19:39 forrest Ryan_Lane: Neither am I, especially for salt code since it's so dynamic
19:39 denys joined #salt
19:39 forrest thus why I was asking, I was hoping you weren't using IAM policies.
19:39 camel_ iggy, highstate will run now, but I have to manually add the java8 ppa
19:40 Ryan_Lane well, we manage the IAM policies per-service and run the orchestration code for every single deploy
19:40 camel_ the rest of it runs after I do that, but for some reason it's not wanting to do that.
19:40 Ryan_Lane we're using roles, so when the instances boot, they immediately have access, without having to have keys
19:40 camel_ I know it's not syntax, because it was working before.  I'm checking into the intermittent connection problems from my hypervisor.  I agree, i think most likely it's something like that.
19:40 forrest Ryan_Lane: Yeah that makes sense.
19:40 josephleon joined #salt
19:41 v0rtex for those too lazy to click my link it's an SLS file that will install and configure a master or minion to use RAET: just do "salt-call --local state.sls raet"
19:41 v0rtex just for fun :)
19:41 camel_ I'll hit you up again after I've reset all the network stuff if it still doesn't work.  Thanks for the input
19:41 Ryan_Lane forrest: I think S3 + IAM policy is a reasonable approach. I can't think of another simpler way :)
19:41 forrest Ryan_Lane: thanks
19:41 Ryan_Lane yw
19:42 Ryan_Lane we use an external pillar for secrets, too
19:42 Ryan_Lane so those aren't in the artifacts
19:42 rojem joined #salt
19:42 MTecknology Is there a .matches() in jinja2 instead of .startswith()?
19:43 iggy MTecknology: {% if 'foo' in 'foobar' %}
19:43 aboe joined #salt
19:43 MTecknology woah..
19:43 MTecknology iggy: :D Thanks!
19:44 iggy #ssce #mindblown #Idon'teventwatter
19:44 Ryan_Lane MTecknology: basically all of the string functions work
19:44 Ryan_Lane .format() works, for instance
19:44 Vynce is there a way to see the rendered (post grain substitution, post templating, etc.) state file for a minion without executing it?
19:44 MTecknology Ryan_Lane: nice .. I need to learn what's limited and what's not
19:45 v0rtex Vynce: state.show_sls
19:45 Ryan_Lane jinja sucks, but occasionally it has nice features
19:45 ipmb joined #salt
19:45 MTecknology I /usually/ like it
19:45 Vynce bertiful, v0rtex, thanks.
19:45 murrdoc also salt has fun stuff like { import_yaml and so on
19:45 MTecknology it is rather limited .. but apparently by design
19:46 MTecknology We need syntax highlighting for vim
19:46 aboe ping s0undt3ch
19:47 murrdoc MTecknology:  https://github.com/saltstack/salt-vim/pulls
19:47 jaloren murrdoc: remember last night i was having a problem with data failing to compile with a state.sls?
19:47 MTecknology :D
19:47 numkem where could I find more information regarding the _pillar and _grains behaviors?
19:48 murrdoc jaloren:  from the command line
19:49 forrest timoguin: what da fuuuuck man
19:49 timoguin who did wut?
19:49 iggy forrest: ikr...
19:50 murrdoc u dont like all the ng formulas going to saltstack-formulas-dev ?
19:50 forrest timoguin: what is with this man: https://groups.google.com/forum/#!topic/salt-formulas/2B_7EwgCT7s
19:50 forrest I just had to comment on it
19:51 forrest because you guys are over here making shit complicated for no reason
19:51 forrest murrdoc: no, don't split the code base further, just make it's own repo in saltstack-formulas
19:51 forrest then they show up right next to each other alphabetically
19:51 murrdoc sure
19:51 murrdoc new org or new repo is fine
19:51 forrest more orgs = more confusion
19:52 forrest then we get some garbage like puppet forge where it's the wild west of 16 different apaches.
19:52 murrdoc the argument being if all the ng is in a diff org
19:52 iggy the way they are going, I like the calling them something other than formulas idea
19:52 murrdoc then it will need to be promoted into the main org
19:52 timoguin forrest: *shrug*
19:52 murrdoc repos vs org is not that big a difference
19:53 jaloren forrest: a hardy amen to that
19:53 williamthekid joined #salt
19:53 iggy bechtoldt already has his own copies of half the shit in saltstack-formulas (that make me want to claw my eyes out)
19:53 schlueter joined #salt
19:53 murrdoc also saltstarters.org
19:53 timoguin I wasn't advocating making anything but a single org what new users see. But a few people doing a bigger refactor seemed good for an org.
19:53 forrest murrdoc: For you and me it isn't, but what happens when I google 'saltstack apache formula'
19:53 forrest and the ng one shows up first in a different org?
19:54 murrdoc forrest:  tim and i asked for the same thing (see puneet reply)
19:54 murrdoc they ran with it and want new or
19:54 murrdoc org
19:54 murrdoc aint tim's bad
19:54 dendazen the only thing is
19:54 forrest Yeah I know
19:54 murrdoc oh wait
19:54 murrdoc idk
19:54 * murrdoc gives up
19:54 dendazen when i run setenforce 0
19:54 dendazen my state shows failed
19:55 dendazen since command complains about SELiux is disabled
19:55 forrest timoguin: I just don't see why we need a 'refactor' just move all the ng terribleness to its own repo. The idea of more orgs is just brutal, what happens when ng isn't configurable, ng+ ?
19:55 dendazen and stderror is not 0
19:55 dendazen i need that to be universal
19:55 dendazen some servers have selinux disabled manually already
19:55 forrest *configurable enough that is
19:57 timoguin forrest: I guess I haven't kept up with all the "ng" stuff. I was just thinking another org would be a good place to work together on a refactor before sending clean-up PRs.
19:57 timoguin But I can see how that org would probably end up taking on a life of its own.
19:57 rayha1 joined #salt
19:57 forrest timoguin: The ng stuff is a total warping of the existing formulas to make every single little thing configurable.
19:57 forrest timoguin: I'm all for doing less work and merging PRs, just seems like there is no reason for another org you know?
19:58 forrest also not sure what the problem is with working off of forks/forks of branches, then squashing commits once stuff works.
19:58 forrest I'm just worried it will get left by the side, and then there won't be any road patrol to come scoop up the carcass.
19:59 timoguin Personal forks makes more sense, I guess.
19:59 iggy "Templates and Formulas as Domain-Specific Language Serializers" I can't even...
20:00 murrdoc https://gist.github.com/puneetk/664eed2dd0eee2b6232c
20:00 jaloren iggy: geez what's with the curse words :)
20:00 badon joined #salt
20:00 murrdoc this is what i am doing for formulas now iggy forrest
20:00 dendazen iggy is there a way  to do cmd.run unless another cmd.run returned particualr stdout word
20:01 forrest seems fine to me murrdoc
20:01 dendazen like run selinux enforce unless getenforce.sdout=‘disabled'
20:01 iggy dendazen: not easily
20:01 dendazen oh ok
20:01 forrest dendazen: I'd do an unless
20:01 otter768 joined #salt
20:01 forrest dendazen: and just have it do a return code value on a 1/0 of selinux enforcing or something
20:01 dendazen i will have to create temp file i guess
20:02 iggy murrdoc: it dissappeared
20:02 murrdoc oi
20:02 murrdoc uh
20:02 murrdoc let me recreate
20:02 murrdoc basically the defaults.yml thing i was talking about yesterday
20:03 murrdoc default = import_yaml defaults.yml + os map;  formula_SEttings = salt.pillar.get(formula:lookup, default, merge=True);
20:03 murrdoc thats the map.jinja
20:08 garthk joined #salt
20:10 murrdoc start a saltstack-formula/cookiecutter repo
20:10 murrdoc so formula strucutres are set in cookie cutter stone
20:11 timoguin definitely agree there
20:12 drawsmcgraw If I have a pillar file like this: http://dpaste.com/3X2E8VZ
20:12 drawsmcgraw What's the best way to target, say, just the compute nodes?
20:12 timoguin drawsmcgraw: -I "openstack:compute"
20:12 drawsmcgraw Specifically in a pillar topfile. Or do I run into an inception-type situation there?
20:13 timoguin that's pillar-ception territory
20:13 drawsmcgraw timoguin: I tried that one (with -I) and no one matched it...
20:13 drawsmcgraw Yeah I was afraid of that
20:13 drawsmcgraw My actual problem is I'm trying to wire up the compute nodes to a schedule
20:13 camel_ I'm still having the same problem after resetting the network.  Can't get pkgrepo.managed for oracle-java8-installer through salt, but I can do it manually.  No connection problems on the server, just salt returning negative with the errors posted http://superuser.com/questions/888427/why-is-oracle-java8-installer-pkgrepo-managed-failing-in-saltstack
20:14 drawsmcgraw I want the schedule in Pillar but want to avoid having to remember to assign grains to machines as I add them to the cluster.
20:14 timoguin drawsmcgraw: I misunderstood the pillar. (fast scan)
20:14 drawsmcgraw ah :/
20:16 rocket_ joined #salt
20:16 drawsmcgraw Maybe I should have a file.managed that iterates over Pillar and fills out /etc/salt/grains accordingly....
20:17 drawsmcgraw I think I'll try that approach
20:17 murrdoc also consider have a state, not in highstate, that cleans the /etc/salt/grains file
20:17 timoguin Doable. I set grains when I launch machines via salt-cloud
20:18 murrdoc other option is to put your managed grains in /etc/salt/minion.d/grains file
20:18 murrdoc so that u dont have collisions when u run grains.set from a state
20:18 dendazen can i do somehting like cmd.run unless getenforce | grep "Disabled"?
20:18 iggy if you are using some cloud provider that has a metadata service, there are a number of custom grains in salt-contrib that will pull that data and put it in grains
20:18 dendazen that would work wouldn’t it?
20:19 iggy onlyif
20:19 iggy grep -q
20:19 dendazen oh
20:20 dendazen thanks
20:20 murrdoc read up on requisites
20:20 drawsmcgraw iggy: murrdoc: Thanks. I've actually been spoiled by salt-cloud. I'm so used to just setting grains in there and forgetting about it.
20:21 drawsmcgraw No salt-cloud here. It's all metal
20:21 drawsmcgraw But I may change up to use /etc/salt/minion.d/grains instead.....
20:22 triplecorn joined #salt
20:23 iggy eww.. bare metal
20:23 johanek joined #salt
20:24 camel_ iggy, it was my crazy firewall.  Just got a beast palo alto box, and it was denying some stuff.  Thanks
20:24 iggy — SSCE SAID SO
20:24 jodv joined #salt
20:30 timoguin drawsmcgraw: no VMs?
20:30 drawsmcgraw not here, no :(
20:30 drawsmcgraw I mean, it's an openstack cluster. But I'm (trying to) manage the actual bare metal
20:30 drawsmcgraw timoguin: ping
20:31 viq joined #salt
20:31 triplecorn joined #salt
20:32 drawsmcgraw What the hell am I missing here? http://dpaste.com/2VZMGYP
20:33 evilrob yeah, my eventual need is going to be using salt to manage an openstack farm
20:33 drawsmcgraw I keep getting TypeError: 'NoneType' object is unsubscriptable when Salt goes out to render the file
20:33 evilrob so metal (for most of it)
20:33 eliasp drawsmcgraw: tried salt['pillar.get']('openstack:controller') instead?
20:34 drawsmcgraw evilrob: My first cut is wiring up a schedule to restart services. I'm ashamed of it, but it's often enough the case that we have a problem that's just solved by restarting some damned service.
20:34 iggy drawsmcgraw: paste the sls?
20:34 drawsmcgraw iggy: the SLS is the text under 'STATE FILE' in that paste
20:34 drawsmcgraw ugly formatting ,sorry
20:35 baweaver joined #salt
20:35 iggy drawsmcgraw: the "JINJA FILE" is where?
20:35 drawsmcgraw oh sorry. It's in /srv/salt/files/.
20:35 drawsmcgraw BUT
20:36 iggy ahh, I see the relationship there
20:36 drawsmcgraw I think eliasp called me on it. There's no such thing as -> pillar.get('openstack:compute_nodes')
20:36 drawsmcgraw I mean that's not how you use Pillar
20:36 iggy salt['pillar.get']('openstack:compute_nodes')
20:36 drawsmcgraw Have to either do what eliasp said. Or use -> pillar['openstack']['compute_nodes']
20:36 drawsmcgraw yep
20:36 drawsmcgraw works
20:36 drawsmcgraw Thanks!
20:37 eliasp yw
20:37 bash1245_ joined #salt
20:37 iggy that's what you should be using
20:38 murrdoc joined #salt
20:38 iggy (kind of wish we could remove pillar/grains/etc from jinja variables)
20:38 ClausA joined #salt
20:39 TheoSLC joined #salt
20:39 bluenemo joined #salt
20:40 murrdoc just dont use it
20:40 TheoSLC greetings.  I'm having trouble with s3fs.  I've aded s3 to my list of fileserver_backend on the master and added my bucket in the s3 list.  But I don't see my bucket in the file lists
20:41 TheoSLC when I run the runner fileserver.update with debug output I can see it updating my gitfs backends, but nothing for s3.
20:42 TheoSLC I'm currently running version 2014.7.1
20:45 evilrob ooohhh... strangeness...  run highstate, get 14 out of 37 failed.  come back 5min later with fresh coffee, re-run it, and they all pass green (no changes made).  So reset vagrant, try again, and after the failure happens (couldn't get lock on /var/lib/apt/lists) apt processes are still running even after my highstate failed and exited
20:50 iggy I just don't see us getting to a common ground between people that want formulas to be cookbooks and those of us that think that's a terribad idea
20:50 baweaver joined #salt
20:50 hal58th evilrob sounds like a connection issue to apt or something like that. just a thought.
20:50 baweaver joined #salt
20:51 iggy the moment someone tries to add code to any formulas I maintain that turns it into plague on humanity just to further their goals is going to be sorely upset
20:51 baweaver joined #salt
20:52 forrest I wish Chad was in here: We've not been able to use a single first-gen formula off-the-shelf without having to make changes to it.
20:52 forrest THAT IS THE POINT
20:52 murrdoc that is not the point as they see it
20:52 timoguin Then it seems we have two separate points.
20:52 murrdoc you should be able to write code to solve every possible use cases
20:52 iggy then they aren't talking about formulas
20:52 murrdoc even the ones u havent run into yet
20:53 murrdoc invite the dude to join irc
20:53 timoguin One side wants examples and baselines to help get started, other side wants to just do some pillar config.
20:53 murrdoc its more one side wants a clean, simple way to do it
20:53 iggy they want more than that (from the sounds of it)
20:53 murrdoc so that include formula is sufficient
20:53 forrest I mean writing all the code as python
20:53 murrdoc no ?
20:53 timoguin I, personally, want them be be more like cookbooks.
20:54 forrest we've now just eliminated every user who can't write python
20:54 murrdoc yeah cookbooks is what forrest / iggy want
20:54 murrdoc but with a functional default mode
20:54 iggy if you want every formula to consider every possible use case, you might as well throw out 99% of the formulas that are there now and start over
20:54 murrdoc invite chad and bechtold to the ircs
20:54 murrdoc shit have a google hangout
20:55 murrdoc shouldnt be too tough
20:55 forrest I mean look at this shit: https://github.com/miketheman/nginx
20:55 timoguin So many option...
20:55 TheoSLC Solved: documentation is incorrect.  you need to use fileserver_backen: - s3fs.  NOT s3
20:55 murrdoc not bad
20:55 TheoSLC how do I fix http://docs.saltstack.com/en/latest/ref/file_server/all/salt.fileserver.s3fs.html ?
20:57 forrest TheoSLC: I believe that one is pulled straight from the file server code
20:57 timoguin TheoSLC: It's in salt/fileserver/s3fs.py
20:57 XenophonF hey speaking of s3fs
20:57 forrest hmm or maybe not...
20:57 XenophonF i can't seem to get it to work
20:58 forrest oh yeah it is, good
20:58 iggy TheoSLC: https://github.com/saltstack/salt/blob/develop/salt/fileserver/s3fs.py#L12
20:58 iggy PR that ho
20:58 TheoSLC iggy: thanks :)
20:59 forrest murrdoc: I really wouldn't know what to say. I've made this point many times directly to Chad
20:59 forrest so it would be me rehashing the point again
21:00 iggy I think the best thing to do right now is call them something else
21:00 forrest I'm literally even fine with formula-ng
21:00 forrest nginx-formula-ng, whatever
21:00 iggy but they want more than that
21:00 Ryan_Lane or do what I do and tell people to write their own code from scratch and make it as specific as possible :)
21:00 XenophonF ah so that's a documentation error
21:01 forrest iggy: What more do they want?
21:01 iggy they want mgmt of "formulas" etc...
21:01 Ryan_Lane only abstract the stuff you need to abstract for just your own use-cases :)
21:01 forrest iggy: In regards to what sort of management?
21:01 forrest damn Chad was here, I think he's cheus
21:01 forrest wish I had know, could have pinged him
21:03 steffo joined #salt
21:03 timoguin iggy: So it's not any special kind of "management of formulas"
21:03 * murrdoc rights ship
21:03 iggy i.e. not just putting them in gitfs_remotes
21:03 timoguin It's a formula for the salt master to put the formulas in place on the master.
21:04 timoguin EMC specifically packages their as .deb and puts them under /srv/formulas
21:04 timoguin So restarting the master isn't necessary
21:04 perfectsine joined #salt
21:04 murrdoc even if they were under /srv/salt restart would have been unnecessary
21:05 murrdoc only thing necessary is sync_all to get the _modules/_states dirs out
21:05 timoguin Right.
21:05 murrdoc i was at the emc talk
21:05 timoguin But modifying gitfs_remotes means a restart.
21:05 timoguin I was telling iggy
21:05 iggy noted... I have no problem throwing a git url in the config file (using the existing salt-formula) and restarting the master
21:05 murrdoc yeah, the gitfs_remotes stuff doesnt work
21:05 murrdoc my bad tim
21:05 forrest is the EMC talk the one where they put everything in map.jinja?
21:05 murrdoc no
21:06 timoguin forrest: yea
21:06 timoguin no?
21:06 murrdoc well yes and no
21:06 forrest lol
21:06 murrdoc sorry
21:06 mdye joined #salt
21:06 murrdoc he didnt do it like saltstack-formula/hadoop/settings.sls
21:06 iggy you're not going to get nearly the amount of contributions if you require every formula to look like those horendous -ng formulas (myself included)
21:07 forrest murrdoc: Right, he put everything in the map.jinja itself from what I was told
21:07 murrdoc he did it like, include defaults, apply os/lsb_logic, and merge pillar
21:07 murrdoc similar to what i showed u earlier in the day
21:07 forrest ahh
21:07 murrdoc but the ng stuff is this weird fuckton of login in settings.sls
21:07 murrdoc and so on
21:07 murrdoc like its not like this at all
21:07 murrdoc https://github.com/saltstack-formulas/php-formula/blob/master/php/ng/map.jinja#
21:07 murrdoc right tim ?
21:08 murrdoc that is confusing for me to read, and i know what its doing
21:08 timoguin Yea EMC's wasn't quite like that.
21:09 iggy how is that better than just having config files and file.recurse'ing... fml
21:09 iggy nvm, I'm out
21:09 murrdoc its like 3pm
21:09 murrdoc where u headed
21:09 murrdoc :)
21:09 iggy booze
21:09 timoguin Good ideas...
21:10 timoguin I don't like reading complex Jinja dictionaries, that's for sure.
21:10 XenophonF so with s3fs, do i also have to configure it using s3fs.keyid, etc.?
21:10 timoguin But I see the logic behind that.
21:11 iggy this is all so very frustrating because I'm a stubborn asshole and refuse to be flexible
21:12 murrdoc that vs https://gist.github.com/anonymous/6edc341b7810d886444d
21:12 murrdoc thats a badly obfuscated map.jinja from my code
21:12 murrdoc but u ll get the idea
21:12 iggy ceph huh?
21:13 murrdoc s/ceph/formulaname
21:13 murrdoc so the pillar.example became defaults.yaml
21:13 JonGretar joined #salt
21:13 murrdoc the os specific stuff was done in os_map
21:13 murrdoc and user overrides come from pillar.get
21:14 iggy the point of pillar.example is also to show some alternate uses... it's not just defaults
21:15 murrdoc yeah
21:15 murrdoc but u either do defaults.yml
21:15 murrdoc or use pillar.get('key, default) everywhere
21:16 iggy I don't generally have a problem with that
21:16 murrdoc its not fun when u are trying to trace down a config
21:16 murrdoc looking in the defaults.yml file and the pillar for a minion is easier
21:17 iggy it does get tedious sometimes, but it's at least clear when you're looking at your sls file
21:17 iggy instead of having to dig back through some godawful settings.sls or whatever
21:17 murrdoc settings.sls != defaults.yml
21:17 murrdoc that settings.sls is fucking hideous
21:19 XenophonF left #salt
21:19 BretFisher joined #salt
21:19 iggy I still think your method is going to end up looking just as bad with a sufficently complex task
21:20 murrdoc nah, all snowflakes go in a pillar
21:20 murrdoc not in the formula
21:20 murrdoc <— is maintaining
21:20 timoguin The point is to have a single place to access config data without having to dig through templates.
21:20 SmileyChris joined #salt
21:20 CheKoLyN joined #salt
21:20 timoguin And if you don't set anything via pillar, you get something like the distro's package maintainer sets as the default
21:21 cheus joined #salt
21:21 iggy this all sounds great in theory
21:21 murrdoc your face
21:21 timoguin I dont' think every single config option needs to be listed though.
21:21 murrdoc ever overrideable config should be
21:21 murrdoc every*
21:22 SmileyChris say I had a bunch of public keys as files in a pillar/state directory -- is there a way of referencing these without manually having to enter them in a state as a jinja array?
21:22 iggy but it's not going to stay nice and clean, you're going to have to change a shit ton of existing formulas, a lot of those are going to get less maintainable, and in the long run you're going to get less contributions because of the complexity
21:22 murrdoc look at tim linked from ceph
21:23 murrdoc they had a 'recipes' directory
21:23 iggy SmileyChris: there's a ticket open for pillar:// uri's
21:23 murrdoc which is what salt-formulas needs
21:23 SmileyChris iggy: they're just public keys, they could live in salt://
21:23 murrdoc a 'solutions' or examples dir, where the snowflake solutions are put
21:23 peters-tx joined #salt
21:23 SmileyChris iggy: if I needed to do more than just copy them though, can I somehow access a salt:// glob as a jinja array?
21:24 iggy SmileyChris: oh, you want a {% for file in dir.list('/foo') %} or something
21:24 iggy nope
21:24 SmileyChris that answers that then :)
21:24 iggy SmileyChris: there's file.recurse
21:28 timoguin murrdoc: That defaults list is going to have to be very closely maintained, especially when config options change between app versions.
21:28 timoguin But I can't think of a better way.
21:28 giantlock joined #salt
21:29 timoguin For a lot of apps, my config file only needs to contain the options I set that aren't default. Everything else will be default.
21:30 timoguin Then any new option that's introduced will automatically get default values.
21:30 timoguin Without it having to be added the a defaults list and without having to modify any templates.
21:30 mgw joined #salt
21:31 smcquay joined #salt
21:32 shaggy_surfer joined #salt
21:33 murrdoc timoguin:  yeah you right, but i have to write these for use across mroe than one role
21:33 murrdoc so i put generic for those in defaults.yml
21:33 murrdoc and then i have the ubuntu flavor specific stuff
21:34 iggy make it next week, I'm f'ing done talking about formulas this week
21:34 murrdoc ?
21:35 iggy any hangout or whatever
21:35 jimklo_ joined #salt
21:36 iggy can you put up a good (useful) example of defaults.yaml before then (not just some highly redacted/hacked C&P of something you guys use in house)
21:36 murrdoc i will after work hours
21:39 rocket joined #salt
21:41 pdx6 joined #salt
21:48 camel_ joined #salt
21:49 Singularo joined #salt
21:50 mosen joined #salt
21:50 camel_ Alright, here's my problem, I am adding pkgrepo.managed ppa: webupd8team/java.  It is failing because it's trying to use port 11371 to connect to keyserver.ubuntu.com.  If I just add-apt the repository, it uses port 80, and works.  Is there a way to make salt use port 80 when getting any ppa?
21:51 camel_ or anything other than that crazy port?
21:51 camel_ I'm behind a corporate firewall, so it's not a big fan of port 11371
21:51 camel_ link to my problem and error log http://superuser.com/questions/888427/why-is-oracle-java8-installer-pkgrepo-managed-failing-in-saltstack
21:53 iggy don't use pkgrepo... just cmd.run add-apt-repo
21:54 camel_ alright, well that solves that problem :)
21:55 MTecknology wrong channel..
21:55 MTecknology I just sent "wrong channel" to the wrong channel
21:55 _JZ_ joined #salt
21:55 camel_ lol
21:57 theologian joined #salt
22:02 otter768 joined #salt
22:03 dalexander joined #salt
22:03 murrdoc joined #salt
22:07 nexsja^ joined #salt
22:07 nexsja^ hello
22:07 nexsja^ i'm very new to salt, and I'm currently having difficulties with testing my configuration
22:08 nexsja^ is there any way to simply run an .sls file (under vagrant) and see if anything happens?
22:10 arif-ali joined #salt
22:10 forrest nexsja^: You can run a single state (if running a masterless minion) with salt-call --local state.sls <state_name>
22:11 nexsja^ "Function init.sls is not available" :/
22:11 nexsja^ I've tried something like that once
22:12 murrdoc state_name is the filename without sls
22:12 nexsja^ oh, wait. 'state.sls'...
22:12 murrdoc in this case
22:12 murrdoc whatever dir init is in
22:12 murrdoc what u will put in place of <state_name>
22:13 nexsja^ oh, cool! it worked! I had to be in a lower dir
22:13 nexsja^ thanks!
22:13 eliasp nexsja^: https://github.com/simonmcc/kitchen-salt
22:14 aquassaut joined #salt
22:14 nexsja^ I'm actually starting to like SaltStack even though I haven't heard of it until yesterday. I've heard only of Chef, Puppet and Docker, but I've been asked to help out with a project
22:14 lietu joined #salt
22:14 camel_ Sooo, on my salt-master, when I run a state.highstate against a server, it will return output about 50% of the time.  It works 100% of the time, and there are no errors (that it's reporting), but there is 0 output about half the time.
22:14 camel_ do i have something configured wrong, or is that just weird?
22:14 ghanima heygys anyone on that is familar with raet
22:14 nexsja^ and it has saltstack already and I'm just trying to wrap my head around Vagrant and SaltStack together... Never worked with any of those before :/
22:14 ghanima does anyone now if it is support in 2015.7.1
22:15 nexsja^ thanks eliasp!
22:16 forrest camel_: Probably need to increase your timeout: http://docs.saltstack.com/en/latest/ref/configuration/master.html#timeout
22:18 camel_ thanks, I'll try that.
22:18 forrest np
22:18 iggy ghanima: you're going to have to compile the dependencies from scratch, so you might as well also go ahead and use a bleeding edge version of salt
22:19 kermit joined #salt
22:20 arif-ali joined #salt
22:26 jerematic joined #salt
22:29 loz-- joined #salt
22:30 MTecknology *grumble*
22:30 dalexander joined #salt
22:30 iromli joined #salt
22:31 MTecknology I added another gitfs_remote and it seems to be pulling without issues. However, the addition to the top file doesn't seem to be looking at that repo... :S
22:32 MTecknology typo!
22:34 shaggy_surfer joined #salt
22:34 loz-- joined #salt
22:36 ecdhe MTecknology, do you have your firewall sls handy?
22:37 MTecknology YES!
22:38 MTecknology I was 10sec from walking out the door
22:38 theo__ joined #salt
22:38 MTecknology https://gist.github.com/MTecknology/ebd746dc9d7cc7d351f0
22:38 ecdhe Thanks so much!
22:38 _JZ_ joined #salt
22:39 wincyj joined #salt
22:39 cberndt joined #salt
22:40 catpig joined #salt
22:40 MTecknology hope you enjoy that much after trying to use it
22:40 ecdhe MTecknology, is it a bear?
22:42 ghanima joined #salt
22:42 MTecknology nope, I think it's incredibly easy and sexy
22:43 MTecknology to be honest... I barely do anything with it anymore. At most, I add a single package to the list. Even that's rare at this point.
22:44 relidy joined #salt
22:44 ecdhe MTecknology, I run salt on between 10 and 100 nodes .  What order of magnitude do you manage?
22:45 aphorise joined #salt
22:46 MTecknology 500
22:46 murrdoc pssh
22:46 MTecknology ~15 at home
22:46 murrdoc * 30 easy
22:46 MTecknology murrdoc: it's more exciting if you listened to my presentation
22:46 iggy I've got 2
22:46 hal58th iggy, mTecknology, I was able to successfully register for ssc.saltstack.com.
22:47 ecdhe MTecknology, link?
22:47 murrdoc MTecknology:  i would have if i could have but i didnt :(
22:47 MTecknology ecdhe: it's not online yet
22:47 MTecknology hal58th: you don't need anything to register
22:47 iggy I told you all not to come to my talk, so that's all on you guys
22:47 MTecknology :P
22:48 ecdhe I was blown away by the bookings site having 15000 nodes...  I didn't know you would need that many for a site of that scope.
22:48 MTecknology I was planning to visit a different one until I listened to the keynote and realized that I'd prefer to not attend that one. So I went to yours instead
22:48 nexsja^ Guys, I'm getting the hang of saltstack - If I understand correctly, then there's a salt-master that has all the configuration and there are salt-minions that communicate with the master via zmq, receive instructions and install necessary software, right?
22:48 nexsja^ So, if I add formulas to master they will be populated to minions
22:48 iggy correct
22:49 nexsja^ Since I don't have a server and won't need one for month or so I've setup a masterless minion
22:49 ajw0100 joined #salt
22:49 nexsja^ but I can't quite get what a minion actually is. From the top.sls config I get that there's... '*' - which refers to all minions.
22:50 nexsja^ and you can specify instructions per each minion.
22:50 MTecknology ecdhe: could partly be because of poor coding... maybe they used ruby
22:50 nexsja^ My project has that - there are instructions for the web minion, db minion, '*web*', '*db*', etc.
22:51 iggy you can actually setup a master and a minion on a single server
22:51 nexsja^ Could somebody explain - what a minion actually is and how to force a masterless minion to read configuration for all minions and install everything else? Is that even possible? Or I'm not getting something
22:51 iggy (I prefer it for testing over masterless)
22:51 MTecknology apparently I have nginx starting, but it's not binding to any ports. Weird. :S   Time to take a break, though
22:51 timoguin joined #salt
22:51 forrest MTecknology: You sure it's starting :P
22:51 ecdhe MTecknology, kinda seems like scaling gone overboard... look at whatsapp, solving the C10M problem with their erlang server.
22:52 nexsja^ iggy: I figures as much and I will do that at some point, but for what I'm doing now I actually need this setup
22:52 nexsja^ figure*
22:53 MaliutaLap joined #salt
22:53 jespada joined #salt
22:53 MTecknology forrest: ps says it's running. I think I /might/ see the issue
22:53 forrest cool
22:54 tligda joined #salt
22:54 hal58th Right, I was able to use my e-mail I provided for the conference and create an ID. https://ssc.saltstack.com/accounts/signup/
22:54 hal58th Mtecknology
22:55 MTecknology I was working on a prod branch for a dev box so it couldn't find the other end of the symlink
22:55 MTecknology hal58th: that's what I did. I figure we'll get the update when there's something exciting to see
22:56 hal58th Mtecknology, were you able to see your test results?
22:56 rocket joined #salt
22:57 MTecknology nope, not until they upload info
22:57 MTecknology patience is the name of this game
22:58 hal58th Yeah, I was trying to say that mine is uploaded.
22:58 MTecknology oooh
22:59 MTecknology gotta run
22:59 neogenix_ joined #salt
23:00 petong POP
23:00 petong mischan!
23:00 cberndt joined #salt
23:01 jespada joined #salt
23:08 MaliutaLap left #salt
23:08 linjan joined #salt
23:10 markm joined #salt
23:21 smcquay joined #salt
23:25 johngrasty joined #salt
23:25 MTecknology hal58th: heh... I used the same address as well, but nothing listed
23:26 hal58th hmmm, guess they didn't get to you yet Mtecknology
23:26 nexsja^ is there a way to undo what a salt-minion installed?
23:26 smcquay joined #salt
23:26 johngrasty Ok. I'm on 2014.7.2. I just upgraded. Now I get errors saying the the source files are not found. Any ideas?
23:26 iggy nexsja^: is there a way to undo what a human installed?
23:27 nexsja^ iggy: there's no automatic way... :))
23:27 smcquay joined #salt
23:27 MTecknology hal58th: so that means people can verify your cert
23:28 iggy in all seriousness though, not really
23:28 hal58th yes, I got a lookup code. 0x9A6D1E8A
23:29 MTecknology number 194
23:29 MTecknology It's kinda odd that they list that number..
23:29 hal58th Yeah, I would also like a total number of SSCE
23:30 MTecknology I'd prefer neither be listed
23:30 MTecknology My SSC ID is ECB8599F .. but no results currently
23:31 baweaver joined #salt
23:31 nexsja^ iggy: gotcha. Is my understanding correct that one server should have one minion?
23:32 nexsja^ seems kind of... stupid. So I must be wrong
23:32 MTecknology It looks like they've limited it to a maximum of 4,294,967,295 people that can register for the site (and some without the cert will register)
23:33 jespada joined #salt
23:35 debian112 one last question for tonight: do jinja supports append, like if I wanted to create a list
23:35 debian112 example: http://paste.debian.net/160837/
23:35 debian112 of what I am talking about
23:36 hal58th yes debian112 http://jinja.pocoo.org/docs/dev/templates/#expression-statement
23:37 iggy Mine keeps saying I don't have any certifications
23:37 hal58th must have not loaded your data yet. I didn't get an e-mail or anything. just decided to try it
23:37 debian112 hal58th thanks, I will check it out
23:39 debian112 hal58th: do i have to load some extensions for do?
23:39 debian112 append
23:39 hal58th not sure debian112
23:43 iggy debian112: http://paste.debian.net/160840/
23:43 patrek joined #salt
23:45 MatthewsFace joined #salt
23:45 aquinas joined #salt
23:45 MatthewsFace joined #salt
23:46 JPaul joined #salt
23:47 johngrasty Ok. I'm on 2014.7.2. I just upgraded. Now I get errors saying the the source files are not found. Any ideas?
23:48 johngrasty It worked yesterday. I upgraded the master, then the minion, and now it doesn't.
23:50 VaneStoniest joined #salt
23:50 iggy that's not an error message I'm familiar with
23:51 dstokes is there a cache on the salt master for custom module responses from minion? running the module on the minion returns the correct data, but running from the master returns stale data
23:52 theologian joined #salt
23:52 iggy yes
23:52 dstokes how do you clear it? clear_cache from both locations doesn't fix it
23:52 iggy wait, what kind of module?
23:52 dstokes a custom python module in _modules
23:52 iggy oh, no
23:53 iggy for some reason I was thinking you were asking about grains
23:54 amcorreia_ joined #salt
23:54 dstokes when you run a module, the minion executes it and returns data right? if that's the case, it appears the master is caching the response
23:55 iggy it doesn't cache it, the return value is stored in the job cache
23:56 iggy can you paste the module code?
23:56 dstokes iggy: https://gist.github.com/dstokes/ac4e7c42947cf17cb1b5
23:57 dstokes ah, what's probably happening is the master has cached the minion grains
23:57 dstokes how do i clear those?
23:58 iggy sync_grains should do it
23:59 dstokes that did it :D
23:59 dstokes iggy++

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