Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-11-29

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

All times shown according to UTC.

Time Nick Message
00:15 scott_w joined #salt
00:18 scott_w_ joined #salt
00:19 oz_akan_ joined #salt
00:20 ccase joined #salt
00:20 scott_w joined #salt
00:27 elfixit joined #salt
00:27 nmistry joined #salt
00:34 anitak joined #salt
00:38 zandy joined #salt
00:40 rojem joined #salt
00:41 nmistry joined #salt
00:44 smccarthy joined #salt
00:52 backjlack joined #salt
00:54 nmistry joined #salt
00:56 [diecast] joined #salt
01:01 sebest joined #salt
01:06 sroegner__ joined #salt
01:06 sroegner__ joined #salt
01:11 zandy joined #salt
01:12 pass_by_value joined #salt
01:22 [diecast] joined #salt
01:34 forrest joined #salt
01:34 anitak joined #salt
01:37 Sheco joined #salt
01:37 nmistry joined #salt
01:37 oz_akan_ joined #salt
01:52 jeter___ joined #salt
01:54 jeter___ question about top.sls, i have 4 environments (base, dev, prod and other). dev inherits base, prod inherits base and other inherits base
01:55 jeter___ when i run a highstate w/ debug on, i see all 4 top.sls files requested. is that right? i thought that shouldn't happen
01:57 jeter___ The conflicting ID is "nginx-repo" and is found in SLS "dev:nginx.repo" and SLS "other:nginx.repo"
01:57 jeter___ im getting that message when im updating a server part of dev environment
02:00 forrest jeter___, are you on 0.17.2?
02:00 jeter___ 17.1
02:00 jeter___ centos epel version
02:01 forrest so when your top runs, it applies stuff in the base to your dev branch as well since you're trying to match that. I believe there is an issue open right now where having something in both base and dev for example, causes an issue. I don't know if it affects 0.17.2
02:05 jeter___ file_roots:
02:05 jeter___ base:
02:06 jeter___ - /srv/states/base
02:06 jeter___ dev:      - /srv/states/dev      - /srv/states/base    other:      - /srv/states/ocean      - /srv/states/base    prod:      - /srv/states/prod      - /srv/states/base
02:06 jeter___ argh, that came out poorly
02:06 forrest can you put that on pastebin or something?
02:06 jeter___ sure
02:10 jeter___ http://pastebin.com/dAXtJ8U4
02:10 jeter___ shows the file_root & then the logs the top.sls being requested
02:11 forrest uhh I don't think you need to include the base again for each environment for the file root
02:11 forrest just include it once, then make sure you're deciding what gets applied where in the top.sls
02:12 jeter___ ah, k
02:13 jeter___ just 3 environments w/ /srv/states/base as kinda a failover
02:13 forrest well, the base states will get applied to everything right? And then each other environment just has what is unique to that environment
02:13 jeter___ thats the goal
02:14 forrest yea, so in your top if you apply the base to everything, then only certain things get applied from prod, you should be good
02:14 jeter___ it seems top file is requested for each environment which seems strange but maybe thats supposed to happen
02:17 techdragon joined #salt
02:19 pass_by_value left #salt
02:22 jeter___ i seem to still get this error: The conflicting ID is "nginx-repo" and is found in SLS "dev:nginx.repo" and SLS "ocean:nginx.repo"
02:22 jeter___ http://pastebin.com/dAXtJ8U4
02:22 jeter___ of my file_root and top files
02:23 forrest ok, so can you fix the conflict?
02:23 forrest seems like you're including nginx.repo twice
02:23 forrest Like I said, I'm PRETTY sure that is a bug
02:23 forrest it shouldn't be doing that
02:24 jeter___ ok, just want to be sure im not doing something stupid
02:24 jeter___ thx for the help
02:24 forrest nope, let me see if I can find the issue.
02:25 forrest pretty sure it's related to this jeter___, https://github.com/saltstack/salt/issues/7526
02:25 forrest yea, the bottom example is a dup of what you're seeing :\
02:25 forrest but other than your duplicating in the file_roots, you would be good to go if this wasn't an issue :D
02:27 stefanmonkey joined #salt
02:29 jeter___ my issue seems to be slightly different
02:30 jeter___ the issue in the ticket seems to be duplicates in the same environment
02:30 littleidea joined #salt
02:31 jeter___ where mine is an environment that doesn't apply is causing a duplicate issue
02:31 forrest I'm pretty sure it's related, because when your states are all compiled together (after it determines which states to apply due to the top.sls) is where the issue occurs
02:31 forrest If you have a github account, I'd add your example that you pasted to me above as a comment on that issue
02:31 [diecast] joined #salt
02:32 jeter___ yes, i do (jeteokeeffe)
02:32 forrest Ok cool, yea if you add your example that has the error and how your states are ordered that you provided earlier, it would probably help the guys figure it out
02:32 forrest The overall issue is basically 'same state gets applied, an error is generated'
02:32 forrest which shouldn't be an issue.
02:35 jeter___ ok
02:35 anitak joined #salt
02:36 forrest your case is good because it shows that even in different environments, the issue is still a problem, which sucks
02:36 nmistry joined #salt
02:48 Linz joined #salt
02:57 jeter___ k, i posted it
02:58 jeter___ github gist sadly doesn't allow directory structures in the file names .... its not as straight forward as it should be
02:59 forrest yea you can just link to your pastebin
02:59 forrest jeter___, can you add the error you were getting to the gist?
02:59 forrest looks fine to me though
03:00 jeter___ its in the title
03:00 forrest oh I see what you are seeing
03:00 forrest my bad
03:04 mgw joined #salt
03:07 favadi joined #salt
03:08 littleidea joined #salt
03:13 khaije|io joined #salt
03:21 minaguib Hey folks. I'm debugging one hell of an issue. Clean installation of salt on a server. I start `salt-master -l all` then a `salt-minion -l all` and they refuse to talk to each other.  I suspect it's zmq related - here are my observations:  1. TCP handshake happens ok 2. minion calls socket.send() with the first packet including its minion ID and public key  2. I do NOT see that packet 's data with tcpdump  3. The master's socket.recv() returns an empty s
03:22 minaguib What I think is happening is that zeromq is "negotiating" and not actually reaching a successful state where the message is actually sent
03:23 minaguib Any ideas how to debug this further ?
03:25 Corey minaguib: You cut off at the start of step 3.
03:26 Corey minaguib: How did you do the install, and what does salt --versions spit out? (to a pastebin or gist please)
03:26 minaguib Corey: 3. The master's socket.recv() returns an empty string '' - which then causes an exception in msgpack._unpacker.unpackb .  In terms of the whole thing on the wire he's what I see: http://pastebin.com/bmzV6TBZ
03:26 forrest minagiub, Also, have you confirmed the keys were properly accepted?
03:27 minaguib Corey: installed via "emerge salt" on gentoo.  salt --versions output here: http://pastebin.com/eDiuWmm4
03:28 forrest minion version is the same minaguib?
03:28 matanya joined #salt
03:28 minaguib forrest: Nowhere near that part. I wiped out /etc/salt and /var/cache/salt before rerunning this.  This is literally the first contact between the minion and master - the minion's first payload has the key, but the master never receives anything meaningul to store it in the minions_pre pki dir
03:29 forrest so what happens when you run salt-key -L from the master?
03:29 Corey minaguib: Are these both on the same LAN?
03:29 forrest it isn't even showing the key there?
03:29 minaguib forrest: salt-minion --versions reports the exact same thing
03:29 minaguib Corey: yes - not only that - I eliminated the lan completely and am running both on the same server
03:29 ravibhure joined #salt
03:29 minaguib so they're talking over 127.0.0.1 as you can see in the tcpdump
03:30 xl1 joined #salt
03:31 minaguib forrest: salt-key -L lists nothing under Accepted/Unaccepted/Rejected
03:31 forrest awesome
03:31 minaguib Here's the last bit the minion logs: http://pastebin.com/cxABK7VP
03:32 minaguib The last 2 lines are my own print() statements to make sure I'm not crazy
03:32 forrest It's ok, I think you're crazy
03:32 minaguib So there is data being put into socket.send(), and it returns None
03:32 minaguib but it doesn't show up in tcpdump, and so the master has no chance in hell seeing it
03:33 minaguib but even weirder, the master's loop receives an empty string ''
03:33 minaguib so the last bit the master logs is: http://pastebin.com/8YAftnQW
03:36 anitak joined #salt
03:36 nmistry joined #salt
03:36 littleidea joined #salt
03:37 minaguib I'm almost certain I'm looking into a pyzmq/zmq/kernel issue, but am just wondering if someone's seen salt behaving this way before
03:37 forrest haven't seen that before, granted we don't get many gentoo users in here.
03:38 kermit joined #salt
03:44 forrest minaguib, just as a test have you tried spinning up VMs at any providers with gentoo to see what happens?
03:46 minaguib forrest: Honestly no, though I do have a central master with ~60 gentoo minions connecting to it successfully without issues
03:47 minaguib forrest: This popped up on the last 3 minions I was upgrading to 0.17.2
03:47 forrest oh weird, are these machines somehow different?
03:47 minaguib forrest: Not in any significant way
03:48 forrest same kernel and everything?
03:48 Corey I'd posit that all gentoo machines are inherently unique snowflakes.
03:48 minaguib Yup
03:48 forrest lol
03:48 forrest and you have enough file handles and everything on the salt master?
03:49 minaguib forrest: as far as I can tell, yes - though in this particular exercise I'm not using my main production master - just getting one of the problematic minions to talk to a local salt-master on the same server
03:49 forrest gotcha
03:49 minaguib I'll try to put together my observations into an email to the mailing list
03:50 forrest sounds good, I was just gonna say I have to leave for a bit but was hoping to see the answer.
03:54 sebest joined #salt
04:00 anuvrat joined #salt
04:06 anitak joined #salt
04:16 abing joined #salt
04:20 ckao joined #salt
04:20 ajw0100 joined #salt
04:24 Furao joined #salt
04:29 druonysus joined #salt
04:29 druonysus joined #salt
04:29 malinoff joined #salt
04:38 techdragon Is there any good way to get more information about just what Jinja is choking on?
04:40 techdragon I found the 'wrong line number' bug, and fortunately all I'm after is what the missing variable is, yet all i can get is 'undefined ninja variable'.
04:44 sandGorgon joined #salt
04:45 dvogt joined #salt
04:46 malinoff techdragon, What version of salt do you use?
04:48 techdragon 0.17.2, and i just found this ticket. https://github.com/saltstack/salt/issues/7625 - the matching Line number is suspicious to me. I'm beginning to wonder if its something to do with using grains
05:17 Ram joined #salt
05:19 techdragon Ive changed the lines around a bit, line 353 is still causing an issue, I just wish I could work out which template its in, I suspect its something being brought in, or generated, not sure yet.
05:21 Guest19281 we are deploying many packages through saltstack.how do we track whihc version installed latest across environmnet for all packages using saltstack?
05:22 favadi joined #salt
05:28 techdragon Guest19281 Do you want to know the versions after a high state call? As part of the results each time you allow salt to install things? Or are you looking to get salt to report back their version numbers whenever you need to check?
05:29 Ryan_Lane joined #salt
05:34 favadi joined #salt
05:43 sebest joined #salt
05:49 NV techdragon: {% if 'postfix' in pillar['servers'][hostname]['roles'] %}
05:50 NV use pillar.get() instead
05:51 NV as it wont crap out if pillar['servers'][hostname] isn't defined for example
05:51 NV well salt['pillar.get']('servers:myhost:roles', []) rather
05:52 NV also, why dont you just have a pillar['roles'] dict?
05:53 lemao joined #salt
05:53 NV servers shouldn't need to access other servers role data no?
05:53 NV ah wait hang on, that isn't your ticket
05:54 NV serves me right for not reading properly
05:56 techdragon It is however similar.
05:56 techdragon But i worked out my current issue.
05:56 techdragon I had that one last night. the pillar.get vs salt['pillar.get']
05:57 techdragon For that I feel a documentation overhaul is due.
05:57 techdragon But my current one may be a bug.
05:58 techdragon using "{% if grains['nodename'] == 'server1' %}" inside a sly file seems to cause some kind of rendering issue. Like salt can't deal with grains.
05:59 techdragon I wonder if its like the pillar.get thing. I couldn't find any alternate grain syntax however.
05:59 Guest19281 reading it ... thx for yiur reply
06:00 malinoff techdragon, grains is a salt's concept lol :)
06:00 malinoff techdragon, I use grains everywhere
06:01 techdragon malinoff thats why i spent 2 hours rewriting and editing a fairly simple state that worked under 0.16.x before trying the big things and swapping out grains for an alternative method of per server config. now it workes.
06:02 Destro joined #salt
06:02 techdragon so far, 0.17 has been rough to me lol
06:03 malinoff techdragon, Well, I use a custom execution module and separate config files for each group of servers I have
06:03 techdragon Ive been mostly using stock salt tools. Keeping things simple, state definitions and pillar.
06:03 rmt joined #salt
06:03 NV techdragon: salt['grains.get']() works exactly like salt['pillar.get']() ftr ;)
06:04 NV compare https://github.com/saltstack/salt/blob/develop/salt/modules/pillar.py#L37 vs https://github.com/saltstack/salt/blob/develop/salt/modules/grains.py#L79
06:04 NV :)
06:05 malinoff tallpaul, stock tools can't handle all my needs :)
06:06 techdragon NV thanks for that, could have saved 2 hours, but i suppose its a case of minimum time before I feel comfortable complaining in IRC
06:06 techdragon and it can be hard to google through github revisions.
06:06 NV mhmm, i find looking at the source helpful at times :D
06:06 NV python is pretty easy to read
06:08 anitak joined #salt
06:08 techdragon One of the reasons its on my preferred language list.
06:09 malinoff basepi, are you here?
06:14 FL1SK joined #salt
06:25 dwyerj joined #salt
06:33 ravibhure joined #salt
06:44 forrest malinoff, he's logged in on his phone, but probably asleep by now
06:44 forrest it 'blinks' at him even if he's not at work
06:44 forrest so I try not to message him outside the standard 9-5 :P
06:46 ravibhure1 joined #salt
06:47 malinoff forrest, my 9-5 standard is quite different from US 9-5 :) And it's difficult to catch US standard time...
06:47 NV it's quarter to 5 here :D
06:47 forrest malinoff, yea I'm not saying don't message him, just saying don't expect a response outside those hours
06:48 forrest plus stupid US holidays
06:49 fivethreeo joined #salt
06:50 forrest I don't know if any of the guys are going to be around tomorrow, I know they have the day off
06:50 malinoff It's OK :) I'll ask the next week
06:51 malinoff Just wanted to ensure my pull request was seen
06:51 forrest which one?
06:51 forrest sorry I was goofing off out in the living room
06:51 malinoff https://github.com/saltstack/salt/pull/8850
06:52 forrest ahh cool
06:53 forrest hey xt are you around?
06:54 forrest xt, when you are around would you be able to comment on https://github.com/saltstack/salt/issues/8682 ?
06:55 dwyerj joined #salt
07:00 forrest oh man dwyerj? What broke :P
07:01 Ryan_Lane joined #salt
07:03 Nazzy forrest, tis a very interesting feature that one
07:03 forrest which one?
07:03 Nazzy 8682
07:03 Nazzy <-- Kaithar
07:03 forrest ahh ok
07:04 forrest Yea I think it could be cool, reporting at least would be nice
07:04 forrest just one more thing to make Salt look badass
07:06 malinoff I think at first salt should get rid of stupid regressions and typo bugs. All these chat bots are interesting feature, but if you can't properly use the basics you will never get to chat bots...
07:06 NV sorting instability in minion-master comms above all though really
07:06 Nazzy I've been playing with the client api quite a bit for web interface needs ... and I have the start of an IM bot framework ... and they're both centered around python tornado loop
07:07 forrest malinoff, I've already spoken to UtahDave regarding the issues with regression testing.
07:07 NV minions dropping off randomly and not reconnecting until the minion is restarted is a pita
07:07 Nazzy NV, I've noticed that ... any idea wtf causes it?
07:07 NV not really
07:07 Nazzy bugger
07:07 NV zmq2 has issues, but im experiencing it on zmq3 as well anyway
07:08 forrest malinoff, He also spoke with Tom about it, and Tom has been putting http://docs.saltstack.com/topics/tutorials/writing_tests.html together, and there's been good effort by guys like cachedout to do stuff like this: https://github.com/saltstack/salt/commit/3ee05e605afe43871bbeb4d91cc42cf6f7650324
07:08 NV i do have some zmq2 minions though, but my master is zmq3 and most minions are zmq3 as well
07:08 malinoff I mean, there are a lot of things to do. Things like fixing retcode to automatically understand what's going on
07:08 NV (rhel5 == no zmq3 packages)
07:08 forrest rhel5, boooooooooo
07:08 forrest we're using some rhel 5 boxes at work too
07:08 forrest crap drives me nuts
07:08 NV going to manually build some zmq3 packages for rhel5 though, and update to v13 of pyzmq and see if that helps
07:08 Nazzy zmq3 is much nicer than zmq2 ... naively I'd say it's improper retry handling, but I doubt it could be that simple
07:09 NV Nazzy: its funny, netstat shows the minion as being connected - and if you restart the master the minion does reconnect, but test.ping still fails
07:09 forrest malinoff, the regression issues are a huge problem, can't have releases breaking shit so often for some of the most dedicated guys
07:09 malinoff forrest, Community will never write tests. People want to write some code instead. And I'm sure salt internal guys should cover the tests by themselves
07:09 forrest there were a couple guys in the IRC that wanted to write tests
07:09 NV so the socket is good, just zmq having an aneurysm
07:09 Nazzy NV, interesting ... and curiously I've noticed it tends to be the same boxes that drop
07:10 forrest shit I'll work on them myself, because I hate when shit breaks.
07:10 malinoff Just a couple vs 550 contributors
07:10 forrest a few is better than none
07:10 NV Nazzy: eh, all of my boxes drop, although its rather intermittent and relatively rare - days pass without issues
07:10 forrest malinoff, besides, people should be adding tests when they add a new feature
07:11 NV although ive noticed the windows boxes crap out more often than the rhel ones
07:11 forrest NV, not gonna comment on that ;P
07:12 malinoff forrest, yes, of course, but I don't have an idea why people should write tests for something that was added in... Wait, in v0.9 or something?
07:12 forrest because they don't exist and that feature keeps breaking on updates?
07:12 Nazzy NV, might be worth slapping trace on the minions and seeing what the last message was
07:12 forrest lol
07:12 forrest it pisses me off when stuff like pkgrepo doesn't work
07:12 forrest or right now, the issue with including the same state twice breaking shit
07:12 malinoff forrest, that's what I'm talking about
07:13 forrest there's no disagreement from the salt team that tests are missing and should have been written from what I've seen.
07:14 Nazzy malinoff, it's worth noting that in any project people tend to have areas of expertise ... I don't know the system well enough to go write coverage (and my boss wouldn't be impressed at me for doing so) but I know chatter bot mechanics like the back of my hand
07:14 malinoff forrest, yes but why the hell I see docker, Puppet integration, chat bots, etc. when I can't properly set up a repo?
07:14 forrest just like what Nazzy said, the Salt guys aren't working on a chat bot
07:14 forrest they're just talking to us about it
07:14 cym3try joined #salt
07:15 forrest most of the stuff you see like docker wasn't written by the main dev team
07:15 Nazzy malinoff, it's nice to say we need a certain bit doing, but to do it, you need to find someone who knows that bit well enough to correctly automate it ... there's a bunch of things I'd like to have but don't trust myself to write modules for
07:15 malinoff forrest, they spend a time on discussing about chat bots instead of something that should be working
07:16 forrest malinoff, I'll take that over no discussion on any issues ever, like the puppet guys do
07:16 forrest move the issue to 'being worked on', 3 years later they haven't touched it with even a comment
07:16 forrest it's quick to talk about something and provide input, it's not quite as fast to write code.
07:17 Nazzy also, discussing is a very good way to get other people to write code for you ;)
07:17 forrest yep
07:18 malinoff I'm not gonna say that discussing is something bad :) Just want to have a more core-stable system. And when the core becomes stable, you can write everything else around it
07:19 malinoff And I think, stuff like pkgrepo is core-related that MUST be stable
07:19 forrest Yea you won't here any disagreement from anyone on that fact
07:19 forrest We all want a more core stable system
07:20 forrest I don't like hanging around having to tell people shit is broken any more than the next guy.
07:20 Nazzy stable or new features, it's a trade off sadly
07:20 forrest yep
07:20 malinoff Nazzy, core-stable. Not just 'stable'
07:20 forrest and they are being pushed hard for new features, plus all the official support they have to provide
07:20 Nazzy salt has a very big core :p
07:21 Nazzy also, main dev team likely knows about the pkgrepo stuff quite well, given all the travis tests are failing because of it
07:22 forrest as much as it annoys us when regression breaks, imagine how the team writing it feels? You just have a new release, and something is broken as shit that you thought worked
07:22 Nazzy (I think it's pkgrepo that's killing them, I've slept since I read the ci output)
07:22 forrest I've been on those projects before, it can be really frustrating
07:22 forrest lol
07:23 malinoff forrest, that's why I think internal guys should write tests for the stuff like pkgrepo by themselves. To be 100% sure that a breaking in a core feature will never pass silently
07:24 forrest You won't here any argument man
07:24 forrest *hear
07:24 matanya joined #salt
07:24 Nazzy https://travis-ci.org/kaithar/salt/jobs/14663285#L2149 <--- yup, pkgrepo is indeed the source of the ci failures
07:25 forrest oh yea Nazzy
07:25 forrest I almost had to move all the damn 'still failing' errors to a subfolder in my email
07:25 forrest was getting spammed by them when I went to bed last night
07:28 Nazzy don't suppose those emails have the first failing commit?
07:28 forrest nope
07:28 forrest This was last night when I majorly screwed my repo
07:28 forrest so it was all trying to clean up
07:29 forrest did you know when you roll back history on your personal repo and push that (to a previously committed item) travis CI runs again? WELL IT DOES
07:29 Nazzy yea, but rewriting history is a Bad Thing (tm)
07:29 forrest yea that's why I didn't do it
07:30 forrest I somehow made a commit to my own repo that had the same commit details as the last one, so it was like a copy with the same changes plus my changes
07:30 forrest so I just deleted all those commits back to the same stage the master repo was on, and confirmed everything was looking good on a fetch/merge.
07:31 forrest the git documentation on how to fix that friggin blows.
07:31 sebest joined #salt
07:32 forrest anyways time for me to go to bed, have a good one guys.
07:33 Nazzy sounds weird ... for things like that you fix your local refs and force push ... then apologize heh
07:33 Nazzy bugger, too slow
07:38 xinkeT joined #salt
07:39 hotbox joined #salt
07:40 apergos joined #salt
07:41 juasiepo joined #salt
07:41 sandGorgon joined #salt
07:46 carlos joined #salt
07:54 lemao joined #salt
07:55 harobed joined #salt
07:59 carlos joined #salt
08:00 juasiepo joined #salt
08:00 slav0nic joined #salt
08:01 ollins joined #salt
08:02 balboah joined #salt
08:07 ajw0100 joined #salt
08:09 MrTango joined #salt
08:11 Ixan I'm getting a weird issue with a state. I can apply it to one minion and it renders fine, but when applying to another minion, it says the sls could not be found. Any ideas on where i can start to look for debugging?
08:12 Destro joined #salt
08:19 giantlock joined #salt
08:20 malinoff Ixan, salt minion-id state.sls yourstate -l debug
08:21 Ixan checked it earlier, doesnt state much at all
08:23 malinoff Ixan, check permissions, ACL's, top files and so on
08:31 dangerousbeans joined #salt
08:33 Destro joined #salt
08:45 balboah if you do extend a state from another, will it replace the thing you extend with your options or append to it?
08:46 malinoff balboah, replace, as I know
08:46 balboah ok
08:51 balboah trying to have a cmd.wait for a service watch. But even if I stop the service and salt executed start again it says "is already enabled, and is running" and cmd.wait is not triggered :(
08:52 balboah aha it's the enable: true that is causing the bug
08:54 ravibhure joined #salt
08:58 Furao joined #salt
08:59 scott_w joined #salt
09:00 wiggy joined #salt
09:01 mike251 joined #salt
09:01 dangerousbeans joined #salt
09:01 * mike251 morning from Frankfurt :)
09:10 tiriel joined #salt
09:10 tiriel Hi everyone
09:11 lemao joined #salt
09:11 tiriel I'm having some trouble writing my sls file. I want to restart apache at an arbitrary point in the file but I can't get it to work
09:12 Destro joined #salt
09:12 aleszoulek joined #salt
09:13 _ikke_ tiriel: statefile don't work like that, they're not procedural
09:13 scott_w joined #salt
09:13 tiriel alright, let's phrase it another way
09:14 scott_w_ joined #salt
09:15 tiriel at some point in my sls I run "a2dissite" via a cmd statement, for this change to be recognised by apache I need to call apache reload afterwards, how would I accomplish this?
09:18 malinoff tiriel, http://pastebin.com/KeNwqt4W
09:18 malinoff http://docs.saltstack.com/ref/states/all/salt.states.service.html?highlight=service#salt.states.service
09:18 jpaetzel_ joined #salt
09:19 dh__ joined #salt
09:19 prooty joined #salt
09:19 sebest joined #salt
09:19 jcsp joined #salt
09:19 lemao_ joined #salt
09:20 eclectic_ joined #salt
09:20 fatbox_ joined #salt
09:20 pmcg_ joined #salt
09:20 jefferai_gone joined #salt
09:21 bdf_ joined #salt
09:22 yidhra_ joined #salt
09:22 scott_w_ joined #salt
09:22 zach___ joined #salt
09:22 Damoun_ joined #salt
09:22 abele_ joined #salt
09:22 mike251 tiriel: you can have the watch statement.
09:22 supplica1t joined #salt
09:22 eskp_ joined #salt
09:24 linjan_ joined #salt
09:24 tiriel Thank you very much, that worked!
09:24 prooty hello. how do you securely use the mysql_* states? i'd rather not have the root password written out in a pillar or in the connection_user argument.
09:24 prooty *connection_pass
09:24 scott_w__ joined #salt
09:25 crop joined #salt
09:25 mike251 prooty: good one... i want to install mysql these days.. so i will look into it also :)
09:25 malinoff prooty, http://docs.saltstack.com/ref/configuration/master.html?highlight=ext_pillar#ext-pillar - you should use external encrypted storage
09:26 Psi-Jack- joined #salt
09:26 litheum joined #salt
09:26 mike251 malinoff: ext_pillar... is not cached on the minion ? just asking silly questions...
09:27 malinoff mike251, Yes, pillars are cached
09:27 prooty malinoff, thanks. i'll try looking at that.
09:28 mike251 but i still don't get how .. is that securing mysql
09:28 defunctzombie_zz joined #salt
09:29 prooty mike251, from what i can tell it only secures the mysql root password so that it isn't written out in plaintext.
09:30 prooty which is what i'm looking for. securing mysql is a different matter.
09:30 krak3n` joined #salt
09:31 scalability-junk joined #salt
09:31 mike251 ok..
09:32 alunduil_ joined #salt
09:32 gasbakid joined #salt
09:33 sandGorgon joined #salt
09:33 dccc joined #salt
09:33 rmt joined #salt
09:34 ggoZ joined #salt
09:35 krak3n` hey guys, how do external pillars map to pillar routes and / or environments in a top.sls, it's not clear form the docs how they relate, like the states gitfs where branches map to environments in the top.sls, is it the same for external pillars?
09:35 JasonG_TA joined #salt
09:35 mirko joined #salt
09:35 16WABMFVL joined #salt
09:36 mike251 can i have a pillar named like:  minion01.host.com.sls ? - it seems i get an error if i use this name... is that normal?
09:38 malinoff mike251, No, you can't. "one.two.three.sls" converts to "one/two/three.sls"
09:38 malinoff mike251, And sls file != pillar, it's just a file that contains pillars
09:41 jeddi joined #salt
09:44 zooz joined #salt
09:45 krak3n` ah found the doc i was looking for: http://docs.saltstack.com/ref/pillar/all/salt.pillar.git_pillar.html
09:47 scott_w joined #salt
09:48 sebest joined #salt
09:51 prooty i'm a bit confused. for example i have an ext_pillar entry "  - cmd_yaml: cat /srv/mysql.sls"
09:52 prooty then in /srv/mysql.sls i have "root_pw: password"
09:52 JasonG_TA joined #salt
09:52 prooty how is this attached to the top.sls pillar?
09:58 balboah top.sls tells which minion that will get that pillar data
09:58 mike251 malinoff: thanks buddy
09:59 balboah prooty: you could for example have a grain with roles: mysql, which you match in the pillar top so that every minion that is of this role also get the pillar data required for it
10:01 rojem joined #salt
10:02 prooty weird, when i try pillar.items it returns "Specified SLS 'mysql' in environment 'base' is not available on the salt master"
10:02 rojem joined #salt
10:02 prooty but pillar.raw shows the contents of the external pillar.
10:03 prooty scratch that, sorry. pillar.raw seems to be cached.
10:06 prooty balboah, i tried "base:\n  '*':\n    - mysql" for top.sls.
10:07 balboah prooty: so make sure mysql.sls is in the pillar root defined in your master config
10:08 scott_w joined #salt
10:08 prooty it's supposed to be an external pillar.
10:31 prooty i'm still confused on how external pillars work exactly.
10:32 sin_ joined #salt
10:32 sin_ hello eveyone!
10:33 mike251 prooty:  i am reading through external pillars and i am confused too
10:33 malinoff sin_, Hi
10:34 sin_ I have dynamic IP adress on my notebook and I can not set  master: 10.0.0.1
10:35 sin_ http://docs.saltstack.com/topics/configuration.html
10:35 mike251 sin_:  what is your laptop? master? or minion?
10:35 sin_ master
10:35 sin_ minoin is vps
10:35 sin_ *minion
10:35 mike251 and why dont' you use a static ip for master?
10:36 mike251 master in my opinion is a static server
10:36 sin_ My provider do not support this
10:37 mike251 sin_:  so you have your minion as a vps in a datacentre and your laptop at home? with a dynamic IP?
10:37 sin_ Can I use my laptop as master ?
10:37 sin_ mike251: yes
10:38 malinoff sin_, You can, but on each IP address change you will have to change config files on all your minions
10:39 sin_ it mean I can not)
10:39 malinoff sin_, I'd rather commend to get AWS micro instance with elastic IP - it costs almost nothing for 1st year
10:39 sin_ mike251: malinoff : thanks
10:43 william_20111 joined #salt
10:45 spuzirev joined #salt
10:47 prooty malinoff, could you explain how to integrate an external pillar with the top.sls pillar?
10:47 spuzirev Hi. Can you help me? I'm trying to use develop branch and salt perfomance is very poor. Using 0.17.2 I have execution time 5.7 seconds for 105 states. In develop build on same machines and same states it takes 48 seconds. Is it by design or I make any mistakes in building develop branche to deb-package?
10:48 malinoff prooty, Nope, sorry. Haven't used that
10:49 joehh spuzirev: I think the packaging is unlikely to make a difference - which os, ubuntu or debian?
10:49 spuzirev ubuntu.
10:49 spuzirev it is precise. I do: git clone && git checkout develop && dpkg-buildpackage
10:50 joehh then almost certain not to be packaging related
10:50 prooty thanks anyway malinoff.
10:51 spuzirev can you guess any other reason?
10:52 joehh regression? either semi deliberate (for new feature and known about) or not?
10:53 joehh I'd be tempted to get something simple and reproduceable and file an issue
10:55 malinoff spuzirev, try to use a profiler - and check differences between 0.17.2 and develop
10:55 pengunix joined #salt
10:56 joehh good idea - also try using top/watch ps/pstree to see if there are many apt-get update or other slow calls
10:58 sroegner__ joined #salt
10:58 spuzirev As I watch in INFO log, all operations are slow. apt-get, file managing, checking services if they are running. In develop log shows very slow, about 2-4 messages per second, and in 0.17.2 all operations except 'apt-get update' are extremely fast.
11:00 malinoff spuzirev, Try to use a profiler. I recommend statprof
11:00 mephx joined #salt
11:01 aleszoul3k joined #salt
11:02 JasonG_TA joined #salt
11:15 malinoff joined #salt
11:15 ertac joined #salt
11:17 scott_w joined #salt
11:17 pengunix joined #salt
11:24 cyberbob joined #salt
11:28 _ikke_ left #salt
11:28 scott_w_ joined #salt
11:30 solarradiation joined #salt
11:32 Ertac joined #salt
11:34 xet7 joined #salt
11:36 JasonG_TA joined #salt
11:36 xl1 left #salt
11:41 scott_w joined #salt
11:41 LordOfLA joined #salt
11:49 LordOfLA joined #salt
11:57 LordOfLA joined #salt
11:58 Sheco joined #salt
12:03 lemao joined #salt
12:26 harobed_ joined #salt
12:31 khaije|io whats the over-arching difference between pillars and grains?
12:32 scott_w joined #salt
12:32 sunand joined #salt
12:35 zandy joined #salt
12:38 scott_w joined #salt
12:40 scott_w_ joined #salt
12:48 puppet joined #salt
12:49 scott_w joined #salt
12:51 balboah khaije|io: grains is used to match on stuff
12:51 balboah or it can
12:53 wiggy joined #salt
12:53 malinoff Pillars can be used for matching too. The main difference is that grains is something that born and live on a minion, but pillars is something that born and live on the master, but CACHED on a minion
12:53 balboah malinoff: can you make an example of targeting minions with a specific pillar?
12:54 malinoff balboah, salt -I 'yourpillar:pillarvalue' test.ping
12:54 malinoff man salt
12:56 balboah confusing indeed
12:56 balboah one feature I often feel is missing for pillars is to override a value instead of having to copy the whole pillar file contents if you want to change just one value of a minion
12:57 scott_w joined #salt
13:01 mattmtl joined #salt
13:03 diegows joined #salt
13:06 scott_w joined #salt
13:07 pniederw joined #salt
13:13 sandGorgon joined #salt
13:14 JasonG_TA joined #salt
13:14 tsing joined #salt
13:15 suzie_11 joined #salt
13:15 suzie_11 left #salt
13:18 zandy joined #salt
13:22 harobed_ joined #salt
13:22 survietamine joined #salt
13:23 survietamine hello
13:23 tsing left #salt
13:24 tiriel Hi, I am writing a simple formula that includes apache-formula
13:24 tiriel If I have my state tree locally on the master everything goes fine, but if I retrieve it from git then the apache.vhosts.default state is skipped
13:25 tiriel what could be causing this and how would I go about fixing it=
13:25 tiriel ?
13:27 tiriel more info, there is a template file mising on the minion, I can't figure out why wouldn't it fetch it
13:29 malinoff left #salt
13:37 harobed joined #salt
13:38 ipmb joined #salt
13:42 survietamine left #salt
13:43 scott_w joined #salt
13:43 Chrisje joined #salt
13:47 Chrisje joined #salt
13:50 shiin joined #salt
13:50 shiin is the mysql default_user stuff implemented yet? it was said to be in 17.0
13:53 wiggy left #salt
14:04 tiriel damn! it works if I run salt-master in the foreground but not if I use the ubuntu upstart script!
14:08 Gifflen joined #salt
14:11 scott_w joined #salt
14:13 sroegner__ joined #salt
14:15 ipmb joined #salt
14:17 jergerber joined #salt
14:17 mgw joined #salt
14:25 littleidea joined #salt
14:31 bezaban how is it handled if a minion matches under multiple sections of top state with conflicting values eg. a file?
14:32 JasonG_TA joined #salt
14:32 bezaban will salt try to determine the closet match or is this best avoided?
14:32 bezaban closest*
14:35 pniederw I think it just executes all matching states
14:36 pniederw and it complains if it comes across two states with the same id
14:38 scott_w joined #salt
14:39 Nazzy as I understand it, it builds all matching states sequentially and then executes them, but I've not confirmed if I understand that correctly
14:40 Nazzy thus a conflicting state would result in the last one encountered being the winner
14:40 bezaban ah right. that's an option too of course
14:41 bezaban excluded it from the nodegroup for now and set up a specific section for the one machine that requires password auth (for now)
14:42 bezaban still not quite decided on what is good grouping of hosts for my use
14:42 bezaban or if I should just have an entry for every host
14:43 defunctzombie joined #salt
14:43 bezaban latter is probably a good idea and then just group common options
14:44 quickdry21 joined #salt
14:45 pniederw I assign roles to hosts in pillar, and match on the roles in state top.sls. one host can have multiple roles.
14:49 gammalget joined #salt
14:51 scott_w joined #salt
14:51 pengunix_ joined #salt
14:52 pniederw how do I call an execution function that I've dropped into _modules/hosts.py from jinja? this doesn't appear to work: {% for host in salt['hosts.all']() %}
14:57 pniederw or maybe the module simply isn't synced to minions. from what I understand, auto-loading of dynamic modules should take care of this though, w/o me having to configure anything.
15:01 th3reverend joined #salt
15:02 JasonG_TA joined #salt
15:02 pniederw anyone around who has used custom modules before?
15:05 th3reverend left #salt
15:06 william_20111 joined #salt
15:08 JasonG_TA joined #salt
15:10 kiorky pniederw: did you salt '*' saltutil.sync_modules ?
15:11 mike251 silly Q: in a conf file i want to import a jinja map... but the name is picked from a pillar... how do i concatenate strings with pillar['environment']  ? ;     {%- from 'files/conf/all/'      pillar['environment']    '.jinja' import jinjamap with context -%}
15:11 elfixit joined #salt
15:12 pniederw kiorky: no, but if I understand http://docs.saltstack.com/ref/file_server/dynamic-modules.html correctly, it shouldn't be necessary
15:17 pniederw kiorky: thanks, that solved it. not sure if that's a bug. I really don't want another step beyond checking module code into my gitfs repository.
15:17 kiorky that's not a bug i think.
15:18 pniederw have you read the docs I link too, in particular auto-loading of dynamic modules? why would it not apply here?
15:18 kiorky at (re)start time
15:20 pniederw you mean I have to restart something? docs don't say anything about this. they say: "This option makes the state system refresh all dynamic modules when states are run." besides, restarting the minion service (which I tried) didn't help.
15:21 shiin left #salt
15:21 pniederw either something is broken here, or the docs are wrong. anyway, thanks for the hint.
15:34 wilywonka joined #salt
15:35 [diecast] joined #salt
15:39 harobed joined #salt
15:40 Psi-Jack I'm trying to determine how best to do this approach. I'
15:41 harobed joined #salt
15:42 gasbakid joined #salt
15:44 mikkn Is there a way to limit key auto acceptance in salt to a specific subnet?
15:44 smccarthy joined #salt
15:45 blee joined #salt
15:46 Psi-Jack Working with this snippet of code, state and jinja: http://paste.linux-help.org/view/54b31528 -- on line 36, I want to only handle Protocol if the pillar defines protocol, so if defined {{server['protocol']}}, put in "Protocol=\"value\""
15:46 donatello joined #salt
15:53 blee joined #salt
15:55 wiggy joined #salt
15:57 Chrisje joined #salt
15:57 yidhra joined #salt
15:57 sroegner__ joined #salt
16:00 Linz joined #salt
16:04 Psi-Jack I'm guessing I'd just have to do something like: {% if server['protocol'] is defined %} Protocol="{{ server['protocol'] }}"{% end if %}   ?
16:04 jkleckner joined #salt
16:06 jkleckner left #salt
16:09 littleidea joined #salt
16:14 Psi-Jack Actually, heck, it would be better to just loop through key/value pairs and use the key and value accordingly.
16:19 littleidea joined #salt
16:22 Linz joined #salt
16:34 [diecast] joined #salt
16:36 mkoistinen joined #salt
16:37 mkoistinen If anyone in here uses Digital Ocean, be sure to read their Twitter feed real soon now.
16:40 sroegner__ joined #salt
16:41 sroegner___ joined #salt
16:42 zandy joined #salt
16:42 pgreen_ccic joined #salt
16:44 pgreen_ccic I just learned about salt and I'm trying it out on my windows box.  I installed 0.17.2 and as far as I can tell, the install went correctly.  However, when I try to run salt-master, I'm met with the following error: [ERROR   ] An un-handled exception was caught by salt's global exception handler : ImportError: No module named resource
16:44 pgreen_ccic Is this I known issue, I haven't found much about it on google
16:45 pengunix joined #salt
16:53 littleidea joined #salt
16:56 sroegner___ joined #salt
17:05 lemoi joined #salt
17:05 lemoi joined #salt
17:10 harobed joined #salt
17:11 harobed joined #salt
17:16 zandy joined #salt
17:21 renoirb Question about best practice with 0.17 and the states.file.patch  when a file has changes coming from  a package
17:27 ajw0100 joined #salt
17:31 pgreen_ccic I'm looking at the source; master.py method __set_max_open_files(), it uses resource to configure the maximum open files, which is not something i'm familiar with in windows; also the resource module is unix only; based on this, is salt-master expected to work on windows?
17:33 dwyerj joined #salt
17:42 Nazca um, basepi, it might be an idea to unset the rest of those modes :p
17:42 Nazca unless you want the channel set +ip ;)
17:44 basepi Whoops. O.o
17:44 basepi Stupid typo
17:45 Nazca yus :p
17:46 mike251 left #salt
17:48 basepi any idea which flags were there before?
17:49 Nazca you did +b -o!*@* and +spi ... since chanserv unset the s, /mode #salt -pib -o!*@*
17:50 bastion1704_ joined #salt
17:51 ajw0100 joined #salt
17:51 basepi alright, i think the rest of those mode flags are fine
17:51 Nazca yup
17:51 pengunix_ joined #salt
17:51 basepi that's what i get for messing with irc on my phone
17:51 jcockhren that's what you get basepi!
17:51 jcockhren nah.
17:52 jcockhren lol
17:53 Nazca ah, phone irc ... the tool of an addict suffering withdrawl ... *hides his own phone irc client*
17:54 jcockhren irc is a great tool. my dev group uses it for everyday stuff as well. chat-ops-ish
17:54 jcockhren I'm the only one that doesn't have irc on my phone, so I be missing stuff
17:54 jcockhren :(
17:55 Nazca oh I love irc, couldn't live without it ... it's just a little painful on a small screen and yucky keyboard
17:56 jcockhren that'sreally my only gripe with it being on a phone.
17:56 basepi yep.
18:07 sroegner___ joined #salt
18:13 zandy joined #salt
18:17 bemehow joined #salt
18:19 xmltok joined #salt
18:22 mkoistinen joined #salt
18:27 bemehow joined #salt
18:29 vejdmn joined #salt
18:41 hhatfield joined #salt
18:46 zandy joined #salt
18:51 Linz joined #salt
18:54 wilywonka joined #salt
18:56 rojem joined #salt
18:57 pengunix joined #salt
19:02 bastion2202 joined #salt
19:04 bastion2202 Anyone have seen this zeromq error when starting the master :
19:04 bastion2202 salt-master
19:04 bastion2202 Process Publisher-2:
19:04 bastion2202 Traceback (most recent call last):
19:04 bastion2202 File "/usr/lib64/python2.7/multiprocessing/process.py", line 258, in _bootstrap
19:04 bastion2202 self.run()
19:04 bastion2202 File "/usr/lib64/python2.7/site-packages/salt/master.py", line 409, in run
19:04 bastion2202 pub_sock.setsockopt(zmq.HWM, self.opts.get('pub_hwm', 100))
19:04 bastion2202 File "socket.pyx", line 330, in zmq.core.socket.Socket.set (zmq/core/socket.c:3301)
19:04 bastion2202 File "checkrc.pxd", line 21, in zmq.core.checkrc._check_rc (zmq/core/socket.c:5838)
19:04 bastion2202 ZMQError: Invalid argument
19:08 Psi-Jack Okay.. This.. Is crazy..
19:09 Psi-Jack I have a jinja situation where a dict is being treated as a string!
19:16 pengunix joined #salt
19:21 linjan_ joined #salt
19:37 wilywonka joined #salt
19:40 jergerber joined #salt
19:46 jeter___ joined #salt
19:46 zandy joined #salt
19:48 zandy joined #salt
19:57 quickdry21 joined #salt
19:58 Linz joined #salt
20:02 Ryan_Lane joined #salt
20:02 quickdry21 joined #salt
20:02 Psi-Jack Okay. http://paste.linux-help.org/view/202a5023    If I comment out line 6 and uncomment 4 or 5, it stops working, because neither of those are objects anymore. Somehow, somehwere, jinja or salt is destroying objects and making them into strings.
20:04 Psi-Jack I do not understand why. However, uncommenting line 4 results in an "undefined jinja variable" while uncommenting line 5 results in a completely off-the-wall error with a line number that doesn't even exist.
20:08 zooz joined #salt
20:10 foxx joined #salt
20:11 ajw0100 joined #salt
20:18 futurisk joined #salt
20:25 Tekni joined #salt
20:26 scott_w joined #salt
20:26 dvogt joined #salt
20:26 rojem joined #salt
20:26 oz_akan_ joined #salt
20:27 ajw0100 joined #salt
20:28 Kizano joined #salt
20:30 Kizano hey all
20:31 Kizano So, I'm curious about the architecture around the "standards" if any about how the sal states should be built.
20:31 Kizano I see the state contrib at https://github.com/saltstack/salt-contrib, but it looks like some of it was out of date since there are references to v0.9...
20:31 Kizano Right now, I'm building my states like this:
20:32 Kizano We have 1 pillar that is "the golden pillar" since >1 pillar made salt-call's take 2x [and more] execution time. This contains all the data for our servers.
20:33 Kizano The states are built using the branching mechanism for the different environments, we have 1 top.sls in the master branch to orchestrate the whole of single-machine instances.
20:33 Kizano In each state.sls or state/init.sls, we are using {% set %} at the top of the files to assort our data, the data comes from salt['pillar.get']() and salt['grains.get']()
20:34 Kizano ^ Parametarizing the data similar to how puppet did in the classes
20:34 Kizano (also coming from a puppet background) :x
20:34 ajw0100 joined #salt
20:35 Kizano What are the accepted standards on how people are building their states + pillar data?
20:35 Kizano or, better yet - how are people handling the use of different data among environments in the states.
20:40 oz_akan_ joined #salt
20:41 lemoi modules/portage_config.py does not support/handle comments in /etc/portage/package.{use,mask,unmask,accept_keywords}
20:42 lemoi I see there's a bug already opened
20:44 oz_akan__ joined #salt
20:46 zandy joined #salt
20:51 pengunix_ joined #salt
20:54 harobed_ joined #salt
21:00 sroegner___ joined #salt
21:02 Ryan_Lane joined #salt
21:06 lamper joined #salt
21:07 ktenney joined #salt
21:18 oz_akan__ joined #salt
21:23 xmltok joined #salt
21:24 pengunix joined #salt
21:24 mwmnj joined #salt
21:37 Linz joined #salt
21:38 anitak joined #salt
21:41 sroegner___ joined #salt
21:45 Ulli joined #salt
21:46 zandy joined #salt
21:48 forrest joined #salt
21:53 ccase joined #salt
21:57 scott_w joined #salt
22:03 dvogt joined #salt
22:03 scott_w joined #salt
22:13 ccase_ joined #salt
22:24 mannyt joined #salt
22:27 pengunix joined #salt
22:41 Linz_ joined #salt
22:46 zandy joined #salt
22:50 mapu joined #salt
22:51 lemao joined #salt
22:52 sroegner___ joined #salt
22:58 dwyerj joined #salt
22:59 pengunix_ joined #salt
22:59 cachedout joined #salt
23:03 swa_work joined #salt
23:03 opapo joined #salt
23:10 Nazca joined #salt
23:14 copelco joined #salt
23:30 pniederw Kizano: From what I've read, salt-contrib has been superseded by https://github.com/saltstack-formulas
23:31 forrest pniederw, that's not quite true
23:31 davet joined #salt
23:31 forrest everything in salt-contrib is add-ons to salt, so new states that didn't previously exist, modules, and so on
23:32 forrest saltstack-formulas is specifically for formulas.
23:32 pniederw maybe not, I'm pretty sure thomas hatch said it in an interview :-)
23:32 pniederw ...but I'm ...
23:33 forrest really? Salt-contribs and salt-formulas serve two completely different purposes. Are you sure you're not thinking of https://github.com/saltstack/formulae ?
23:35 pniederw not sure. I remember him saying that the plan was to get the usable bits of salt-contrib (or formulas?) into salt proper and then move on.
23:36 forrest weird, there's even a pull request from 15 days ago saying he thinks something should go into salt-contrib https://github.com/saltstack/salt/pull/8483
23:46 zandy joined #salt
23:46 mannyt joined #salt
23:47 pniederw forrest: can't find that statement anymore (whatever it was). you are probably right then, was more likely about formulae vs. saltstack-formulas.
23:47 forrest ok let me know if you find it, I'd really be curious to see what he was talking about
23:51 pniederw have been working crazy hours recently to get the cm to a working state. probably doesn't help my brain. I know someone said something, though. :-)
23:51 forrest heh
23:51 forrest yea I'm glad to have the day off today
23:54 jams_ joined #salt
23:59 davet joined #salt

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