Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-04-04

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

All times shown according to UTC.

Time Nick Message
00:02 andrej timoguin - it wasn't (isn't) erroring in the state I posted 2nd to last
00:02 andrej Thanks for showing me the diff, though
00:05 garthk andrej: hey there again!
00:06 fllr joined #salt
00:06 garthk andrej: I got my miniature CA working, with stunnels everywhere
00:06 garthk meanwhile
00:06 garthk Is there some way I can have the salt command on my laptop fire commands to my salt hierarchy without setting up a syndic?
00:13 mgw joined #salt
00:14 andrej garthk : nice :)
00:17 bhosmer joined #salt
00:17 sroegner_ joined #salt
00:18 ajprog_laptop joined #salt
00:20 yomilk joined #salt
00:23 rgbkrk joined #salt
00:25 diegows joined #salt
00:31 thayne joined #salt
00:34 ekristen joined #salt
00:35 ndrei joined #salt
00:36 taion809 joined #salt
00:39 fllr joined #salt
00:46 xunil joined #salt
00:56 andrej Oooooph
00:56 andrej working
00:56 andrej I don't understand why
00:56 andrej all I changed is the FORMAT of my add_minion.sls
00:57 andrej The initial form follwed the command on leftmost end: form, with cmd.run indented
00:58 andrej I changed that from that to raandom_name:, indent cmd,run, next indent - name: command
00:58 andrej and the sls invoked from reactor now does what I want
00:58 andrej I'm stumped
01:05 smurfy__ joined #salt
01:05 BogdanR joined #salt
01:12 andrej How do I "translate" salt '*' event.fire_master '{"data":"my event data"}' 'tag' to something I can put into a state file?
01:12 smcquay joined #salt
01:13 Ryan_Lane andrej: so, you'll use http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.event.html#salt.modules.event.fire_master in your state
01:13 halfss joined #salt
01:13 Ryan_Lane andrej: you'll use the module state to run it: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.module.html#module-salt.states.module
01:13 Ryan_Lane the data is passed in as parameters
01:14 andrej Ryan_Lane : thanks - starting to read now :)
01:14 halfss joined #salt
01:14 Ryan_Lane andrej: what you'll really want is likely: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.module.html#salt.states.module.run
01:15 Ryan_Lane though depending on what you're doing wait and watch could also be really useful ;)
01:15 andrej heh
01:16 andrej All I need to do is to fire an event with a yet to be defined (albeit unique) tag from a minion to the master when a package gets installed, so I can open a port on another minion via the reactor
01:16 andrej Whiteinge suggested that listen to ALL jobs and filtering out relevant ones might be a tad inefficient :}
01:17 andrej Hence the desire to fire a specific tag out
01:17 Ryan_Lane ah, then you may actually want to watch the package
01:17 Ryan_Lane and fire the custom event on that
01:17 Ryan_Lane it would be kind of awesome if every state had an event param
01:17 andrej :}
01:18 Ryan_Lane then you could fire an event on any normal state action
01:18 andrej Neat thought ...
01:18 * Ryan_Lane adds a github issue for that
01:18 andrej heh
01:20 andrej Ryan_Lane - how would I stand up that watch, and, more importantly, fire the event on success?
01:20 Ryan_Lane I'm not 100% sure how watch works
01:20 ckao joined #salt
01:21 Ryan_Lane https://github.com/saltstack/salt/issues/11762 :)
01:25 Ryan_Lane andrej: would that actually meet your use case?
01:25 Ryan_Lane notice that I added an example use case
01:25 Ryan_Lane if it would, maybe add your own use case?
01:25 Ryan_Lane and maybe an alternative way of specifying the event, if my way isn't very good :)
01:26 fllr joined #salt
01:28 andrej I guess that would kind of do. I'm really just after a certain tag, so I can keep the post-processing in jinja/yaml/whatever to a minimum :)
01:28 yomilk joined #salt
01:28 rgbkrk joined #salt
01:30 middleman_ joined #salt
01:31 * Ryan_Lane nods
01:31 Ryan_Lane andrej: my 2nd example may be more useful, since you'd be able to specify a tag name and/or data
01:32 Ryan_Lane anyway, good luck with firing the events
01:32 andrej Heh
01:32 andrej ta
01:53 ipalreadytaken joined #salt
02:18 sroegner_ joined #salt
02:21 Linz joined #salt
02:34 abe_music joined #salt
02:37 pydanny joined #salt
02:44 possibilities joined #salt
02:50 rgbkrk joined #salt
02:51 napper joined #salt
02:52 schimmy joined #salt
02:57 schimmy1 joined #salt
03:00 jeremyfelt joined #salt
03:05 mgw joined #salt
03:07 dccc joined #salt
03:11 gorran joined #salt
03:18 zz_cro joined #salt
03:19 eykd joined #salt
03:22 eykd Hey, I've got a question on Stack Overflow regarding Salt that I'm trying to get an answer to: http://stackoverflow.com/questions/22841230/introspection-into-forever-running-salt-highstate
03:22 eykd I've got a highstate that runs forever, and I don't know why.
03:23 forrest eykd, did you try to just do state.sls on nginx?
03:24 eykd Not sure what you mean?
03:24 forrest salt -v 'web*' state.sls nginx_state
03:25 forrest to run that single state, to see if that is where the problem is
03:25 eykd Hey, I didn't know I could do that. :)
03:25 forrest also, can you pastebin the entire state?
03:25 eykd Can do, just a minute.
03:25 forrest cool, also try that state.sls, you might want to add on -l debug.
03:26 eykd Actually, better yet, I've got the full state here: https://github.com/worlds-enough-studios/nginx-formula
03:26 forrest eykd, if the state.sls hangs, I'd also suggest running it against a single machine to eliminate the possibility of a minion related issue.
03:26 eykd This is a single machine.
03:26 forrest gotcha
03:27 eykd It's a fresh digital ocean droplet. I've destroyed/re-provisioned several times.
03:27 forrest gotcha, ok yea, then run state.sls and see what happens there.
03:27 schimmy joined #salt
03:28 eykd OK, just a moment, I have to destroy and reprovision.
03:28 forrest no worries.
03:28 forrest ugh I really don't like the nginx formula
03:28 forrest I hate the usage of init to include other states, bleh
03:29 eykd Oh yeah?
03:29 eykd I'm just getting into salt, so I don't really know what the best practices are.
03:29 forrest eykd, yea we don't have all the formulas standardized in that sense yet.
03:29 eykd I didn't like how the nginx installation was set up though, so I tweaked it considerably. Hence my current trouble. :P
03:30 forrest heh
03:30 forrest yea I'm curious as to whether it's getting caught in a loop or something.
03:30 eykd I wonder.
03:32 eykd What would cause a loop?
03:32 Ryan_Lane joined #salt
03:32 forrest eykd, nothing that I can see looking at the states.
03:33 forrest default.conf doesn't require or interact with nginx.conf in any way. I wonder if the service is having a freak out..
03:33 eykd OK, provisioned, so here we go.
03:33 ninnypan_ joined #salt
03:34 eykd Or wait, default.conf?
03:34 forrest ?
03:34 forrest just run the state.sls as state.sls nginx
03:34 eykd OK. :)
03:34 ninnypan_ if I'm storing users in pillar and want to push their public keys to the server where is the best place to put them?
03:34 forrest and it will apply the nginx init.
03:35 forrest which includes those other ones
03:35 ninnypan_ should I put them in the file root and get them with salt://?
03:35 forrest ninnypan_, Unless you have privacy concerns over public keys, I'd just make them managed files.
03:35 forrest or you can do like utahdave did here: https://gist.github.com/UtahDave/3785738
03:35 forrest and you could drop them in pillar, then do some 'fun' variable substition to drop them into files.
03:36 eykd Oh shoot, I forgot to add -l debug. :\
03:36 forrest no worries, is it just running?
03:37 fxhp http://pad.yohdah.com/241/refactor-and-test-python-salt-state-internal-keyword-cleaning
03:37 eykd Just running.
03:37 eykd And my guess is, it will continue to run.
03:38 forrest eykd, yea, so I'd start by modifying that watch that you pasted on stackoverflow, to remove the nginx.conf check
03:38 forrest just as a test
03:38 forrest so you can confirm that it's getting hung up on the default.conf
03:38 forrest or if it's a combo of the two
03:38 forrest and just as a question, both master and minion are running the same release of salt right?
03:39 eykd Oh, you know, I already tried removing the default.conf watch. That's in github now.
03:39 eykd Yes, both the same.
03:39 eykd Fresh installations.
03:39 eykd Ubuntu 12.04
03:39 forrest so when you're just watching the default.conf, it still hangs?
03:39 forrest but as soon as you remove that, and just watch nginx.conf, no problems?
03:39 eykd Er, no, actually, it still hangs after I remove the watch nginx.conf
03:39 ninnypan_ forrest: ok that looks similar to how I was going to do it thanks
03:39 eykd So it might be one of the other changes I made.
03:39 forrest ninnypan_, yea np
03:40 forrest eykd, ok, at this point then, I'd remove all of the watches, thus forcing a restart ANY time the state is run
03:40 forrest if it still borks up, then we know something else you did is causing it
03:40 eykd OK, I'll try that.
03:40 eykd I'm also wondering if there's anything fishy-looking in common.sls: https://github.com/worlds-enough-studios/nginx-formula/blob/master/nginx/common.sls
03:41 forrest eykd, yea I need to diff your changes versus the original state since I know that one works. I've gotta head to the gym, but I'll take a look when I get back if you don't figure it out.
03:41 eykd OK, thanks for your help, I appreciate it.
03:41 forrest yea np, you'll probably figure it out by the time I'm back, just start removing your changes till it works :P
03:41 eykd I'll just start ripping things out one at a time, yeah. :)
03:46 anuvrat joined #salt
03:53 possibilities joined #salt
03:54 ipalreadytaken joined #salt
04:07 malinoff joined #salt
04:19 sroegner_ joined #salt
04:29 jalbretsen joined #salt
04:30 rompop joined #salt
04:31 l0x3py joined #salt
04:40 kiorky joined #salt
04:41 MedicalJaneParis joined #salt
04:43 fllr joined #salt
04:46 Sypher joined #salt
04:51 hotsnow joined #salt
04:53 smcquay joined #salt
04:58 borgstrom joined #salt
05:00 garthk Comment: Unable to manage file: Jinja variable 'redis_tls_host' is undefined; line 13
05:00 garthk connect = {{ opentsdb_tls_host }}:{{ opentsdb_tls_port }}    <======================
05:00 garthk !?!?
05:01 garthk aah, there’s a reference in line 2
05:09 thayne joined #salt
05:13 fllr joined #salt
05:15 _fllr_ joined #salt
05:16 anuvrat joined #salt
05:21 ipalreadytaken joined #salt
05:24 forrest eykd, did you figure that out?
05:25 eykd Close to it, I think. I commented everything out, then went through and restored it piece by piece.
05:25 eykd service.running seems to be the problem, actually.
05:25 eykd And I think it's a problem w/ nginx, not salt.
05:25 forrest oh?
05:25 eykd As in, nginx isn't starting up properly.
05:25 forrest does the service just hang if you're on the box?
05:25 eykd Yeah.
05:26 forrest lame
05:26 eykd So I probably screwed up the conf file. :(
05:26 forrest nginx should throw some errors I thought when it has a problem with the conf, just like apache.
05:27 eykd Usually it does. Not sure what I've done yet.
05:28 garthk salt-call state.sls to the rescue!
05:30 forrest eykd, if you confirm it's nginx, can you answer your own question and mark it as answered so the salt guys don't see it and spend time troubleshooting it?
05:30 eykd Oh! I bet it's the daemon setting. The formula was originally using upstart.
05:30 eykd That could cause the service start to "hang"
05:33 eykd Success! That was it.
05:33 forrest nice
05:33 ipalreadytaken joined #salt
05:33 forrest now you just need to put all your stuff back in :P
05:34 forrest or pull it back from the repo, lol
05:34 eykd Yeah.
05:34 eykd GitFS is nice, but it's really annoying for debugging.
05:34 forrest yea
05:34 forrest it's awful to debug
05:34 forrest one of the joys of git I think :P
05:34 eykd Ha. :)
05:35 eykd Thanks a lot for your help--I feel like I have a better grasp on things now.
05:35 forrest yea np
05:35 forrest you saw my comment earlier about the stackoverflow question right?
05:35 eykd Yeah, I'll clear it.
05:35 forrest great thanks
05:35 forrest otherwise I know utahdave will get in there and mess with it for a solution :P
05:37 eykd I've posted my answer, but I can't accept it until tomorrow.
05:37 CeBe joined #salt
05:37 forrest eykd, awesome
05:37 forrest as long as there is an answer up there, should be good to go.
05:37 eykd Cool.
05:37 forrest so you're not using upstart?
05:38 forrest or did you just have to modify a conf setting
05:41 fllr joined #salt
05:53 ndrei joined #salt
05:54 smurfy__ joined #salt
05:54 Valdo joined #salt
05:55 hotsnow windows minion log: ZMQerror: no buffer space available
05:56 zach joined #salt
05:58 viod1 joined #salt
05:58 eykd forrest, I prefer to keep the default nginx init.d service setup that I'm used to, though I love upstart for everything else.
05:58 forrest fair enough, I just wanted to double check that the formula itself wasn't causing you a problem.
06:02 eykd No, it was my screwing around w/ the formula that caused the problem. I never did well at chemistry. ;)
06:02 mtford joined #salt
06:02 forrest heh
06:05 xmj moin
06:06 xmj are there salt virt states (much like salt virt modules, 20.16.182 salt.modules.virt) ?
06:06 hhenkel joined #salt
06:08 MedicalJaneParis joined #salt
06:08 forrest xmj, nope, but you could use: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.module.html
06:10 xmj hm, that sounds a bit fugly
06:11 forrest you coudl always write the virt state :P
06:11 xmj like, for instance, salt.modules.virt.create() is totally not salt.states.virt.created() :-)
06:11 xmj forrest: I knew that answer was coming.
06:11 forrest hah
06:11 xmj forrest: better than "patches accepted" (what i usually reply ... ;-)
06:12 forrest Yea
06:12 xmj forrest: we'll see
06:13 xmj forrest: probably need to work on libvirt first before going ahead and integrangvirt states..
06:13 xmj integrating..*
06:13 forrest yea
06:19 sroegner_ joined #salt
06:28 travisfischer joined #salt
06:29 gammalget joined #salt
06:36 xDam0x joined #salt
06:49 eykd joined #salt
07:03 Katafalkas joined #salt
07:06 balboah joined #salt
07:09 Katafalkas joined #salt
07:12 fllr joined #salt
07:12 vvlad_ joined #salt
07:19 ekristen joined #salt
07:20 obimod joined #salt
07:20 obimod whaaaaaatuppp
07:27 Kenzor joined #salt
07:28 ravibhure joined #salt
07:33 ndrei joined #salt
07:34 videre joined #salt
07:37 harobed joined #salt
07:45 videre joined #salt
07:46 ndrei joined #salt
07:47 joehoyle joined #salt
07:50 johtso joined #salt
07:51 johtso joined #salt
07:57 topochan joined #salt
07:59 ndrei joined #salt
08:10 carlos_ joined #salt
08:10 babilen whiteinge: Thank you, I appreciate your comments a lot and both approaches are suitable. I opted for the first, non-macro, one in this particular case, but will have to explore Jinja Macros some more (they happify my Clojure side)
08:12 fllr joined #salt
08:13 think-free_ joined #salt
08:14 _fllr_ joined #salt
08:16 joehoyle joined #salt
08:20 sroegner_ joined #salt
08:35 micko joined #salt
08:38 giantlock joined #salt
08:40 ggoZ joined #salt
08:42 fllr joined #salt
08:43 redondos joined #salt
08:46 ajw0100 joined #salt
08:50 MrTango joined #salt
08:56 joehoyle joined #salt
08:57 ajw0100 joined #salt
08:58 Ryan_Lane joined #salt
09:00 grosjean joined #salt
09:00 grosjean yo all
09:02 grosjean i'm trying to use cmd_async with get_cli_returns to have a blockant job
09:02 grosjean but i'm a bit lost with get_cli_returns
09:08 kermit joined #salt
09:24 gildegoma joined #salt
09:28 bhosmer joined #salt
09:31 bhosmer joined #salt
09:31 vvlad joined #salt
09:33 Ryan_Lane joined #salt
09:39 joehoyle_ joined #salt
09:42 fllr joined #salt
09:48 joehoyle joined #salt
09:56 srijan4 joined #salt
09:57 xunil__ joined #salt
10:00 rtucker joined #salt
10:10 millz0r joined #salt
10:10 TamCore saltstack.com says latest release is 2014.1.1 (since 18. march), but why isn't there a .deb, yet? latest deb is only 2014.1.0 :/
10:16 babilen TamCore: There is: http://paste.debian.net/91620/
10:17 bezaban yeah, there is in the salt deb repo and it has made it into sid
10:18 babilen Quite shortly after the release even, so not sure what TamCore is referring to. Maybe some derivative?
10:19 londo_ joined #salt
10:20 TamCore babilen: was looking in saltstack ppa
10:20 babilen Ubuntu?
10:20 babilen You might want to consider running Debian if you want up-to-date software
10:20 * babilen couldn't resist
10:21 sroegner_ joined #salt
10:21 bezaban hehe
10:22 TamCore I'd like to use Debian, but my employer prefers Ubuntu :x
10:23 babilen TamCore: Sure, no worries. I am not involved in Ubuntu nor in the salt packaging for it, so I am not sure what the current status of it is. joehh might have run into prolems (again) while working on the Ubuntu packages or maybe has not yet found the time to finalise them.
10:34 Ryan_Lane joined #salt
10:42 fllr joined #salt
10:45 obimod [ERROR   ] An un-handled exception was caught by salt's global exception handler: RSAError: invalid header
10:45 obimod on the minion -- anyone have any idea?
10:46 obimod M2Crypto.RSA.RSAError: invalid header
10:50 Shish just one minion out of several giving that message, or all of them, or is there only one to test with?
10:52 obimod only one to test with
10:52 obimod it was with ubuntu... trying debian now, but debian has failed for me when deploying with salt-cloud -m cloud.map
10:57 TyrfingMjolnir joined #salt
11:00 paramah exit
11:05 obimod yyyup http://pastebin.com/xtwUbW57
11:07 babilen obimod: You have to accept the repo key -- http://docs.saltstack.com/en/latest/topics/installation/debian.html
11:07 ndrei joined #salt
11:08 obimod on the remote minion? should i add a pull request to salt-bootstrap?
11:11 babilen obimod: Not familiar with salt-bootstrap, but that is what is missing I think
11:11 obimod kk - thank you
11:12 MZAWeb joined #salt
11:16 obimod nope : (
11:17 obimod after accepting the repo key (responded OK) ... http://pastebin.com/2KDhjSyV
11:20 babilen obimod: Could you upgrade debian-archive-keyring ?
11:21 obimod oh yea, lol
11:21 obimod thx
11:22 harobed_ joined #salt
11:23 harobed_ joined #salt
11:24 taion809 joined #salt
11:27 obimod yes the stable http://bootstrap.saltstack.org needs to be updated : (
11:27 halfss_ joined #salt
11:28 halfss_ joined #salt
11:28 ndrei joined #salt
11:33 faldridge joined #salt
11:34 faldridge joined #salt
11:34 Ryan_Lane joined #salt
11:42 diegows joined #salt
11:42 fllr joined #salt
11:43 obimod babilen: this is what is weird... http://pastebin.com/qzLjjBdJ
11:44 obimod even though it's in the bootstrap scripts it doesn't seem to work
11:46 joehoyle joined #salt
11:50 abe_music joined #salt
11:50 babilen obimod: yes, backports have been moved to the official mirrors. A suitable line for your sources.list would be: "deb http://http.debian.net/debian-backports/ squeeze-backports main [contrib] [non-free]" (adjust as needed, uses a metamirror that picks a "good" mirror automagically)
11:52 halfss joined #salt
11:54 joehoyle_ joined #salt
12:04 bastion1704 joined #salt
12:05 TyrfingMjolnir joined #salt
12:07 dimension joined #salt
12:11 obimod babilen: still didn't work : ( http://pastebin.com/uzrpBau0
12:11 srage joined #salt
12:11 obimod babilen: it works when i apt-install debian-archive-keyring after the salt-cloud -m cloud.map fails and i login and run the install and re-run the /tmp/.saltcloud/deploy.sh
12:13 obimod babilen: fyi https://github.com/saltstack/salt-bootstrap/pull/361 - bed time!
12:13 obimod thx
12:14 joehoyle joined #salt
12:16 TyrfingMjolnir joined #salt
12:22 sroegner_ joined #salt
12:27 higgs001 joined #salt
12:29 bhosmer joined #salt
12:29 napper joined #salt
12:31 untamo13 joined #salt
12:35 Ryan_Lane joined #salt
12:42 wkf joined #salt
12:42 fllr joined #salt
12:44 anuvrat joined #salt
12:46 fllr joined #salt
12:48 srage joined #salt
12:48 jslatts joined #salt
12:48 _fllr_ joined #salt
12:49 Fin1te joined #salt
12:51 Fin1te left #salt
12:59 vbabiy joined #salt
13:01 thayne joined #salt
13:03 joehoyle joined #salt
13:06 ndrei joined #salt
13:06 joehoyle_ joined #salt
13:13 oz_akan_ joined #salt
13:19 jaimed joined #salt
13:22 elfixit joined #salt
13:22 untamo13 joined #salt
13:25 ipmb joined #salt
13:26 mpanetta_ joined #salt
13:29 Linz joined #salt
13:31 pydanny joined #salt
13:34 oz_akan_ joined #salt
13:35 jgelens I have multiple pkgrepo.managed statements and for every statement it executes an apt-get update. Is it possible to do this only once ?
13:38 ndrei joined #salt
13:40 Psi-Jack I have a wierd issue in Jinja. In the top of my jinja, I {%- set logdir = "/var/log/supervisor", and while in a for...loop, if a piece of pillar data has a logdir=value, it runs another set to change logdir, but in another loop outside of the former loop, logdir isn't changed, it's using only the originally set.
13:41 davidone joined #salt
13:42 fllr joined #salt
13:44 davidone I need to create a server from scratch, booting with a livecd where I have salt-minion
13:44 davidone so I need to do a lot of action (fdisking disks, installing packages, users and so on)
13:44 davidone what is the 'best' way to do this kind of task?
13:45 davidone I'm creating a bootstrap/top.sls file using jinja instructions
13:45 jaimed joined #salt
13:45 davidone but probably I should use a complete different approach, creating a library or something else
13:47 Gordonz joined #salt
13:49 londo__ joined #salt
13:51 JasonSwindle joined #salt
13:53 pydanny joined #salt
13:54 GradysGhost joined #salt
13:55 fllr joined #salt
13:58 seapasulli joined #salt
13:59 Damoun joined #salt
14:04 vvlad joined #salt
14:04 sroegner_ joined #salt
14:13 londo__ joined #salt
14:18 arrose joined #salt
14:20 _mel_ joined #salt
14:22 _mel_ Hi. I'm currently trying salt on wheezy with tow computers. master and one minion. if i change something eg. a sls file and try to run a state.highstate i often get this error: "The function "state.highstate" is running as PID 6583 and was started at 2014, Apr 04 16:18:26.962633 with jid 20140404161826962633" after restating the minion it works again. What am i doing wrong?
14:24 ndrei joined #salt
14:25 ajprog_laptop joined #salt
14:25 tyler-baker joined #salt
14:25 toastedpenguin joined #salt
14:26 _mel_ salt-call from the clients works
14:26 thedodd joined #salt
14:27 ajprog_laptop1 joined #salt
14:27 bhosmer joined #salt
14:30 joehoyle joined #salt
14:33 Psi-Jack {%- if defined pillar["supervisor-progs"][group]["logdir"] %}   -- Why am I getting this error: inja syntax error: expected token 'block_end', got 'pillar'
14:35 jalbretsen joined #salt
14:35 fllr joined #salt
14:37 alunduil joined #salt
14:39 devx joined #salt
14:41 bhosmer_ joined #salt
14:41 ashb mmmm okay can anyone explain to me how stateconf renderere is not just crazy madness?
14:41 ndrei joined #salt
14:41 ashb It seems to have implemented an extra layer of templating on top of yaml/jinja
14:42 ashb amy I missing something? What is the use case its deisgned for?
14:43 bhosmer joined #salt
14:49 Psi-Jack {%- if pillar["supervisor-progs"][group]["logdir"] is defined %}
14:49 Psi-Jack {%- set logdir = pillar["supervisor-progs"][group]["logdir"] %}
14:50 Psi-Jack This, is resulting in {{ logdir }} being the value: "missing"
14:50 Psi-Jack Yet, it's defined, the defined value is not missing, it's actually "/var/log/supervisor"
14:52 jaimed joined #salt
14:52 kaptk2 joined #salt
14:55 bhosmer joined #salt
14:58 UForgotten Psi-Jack: did you
14:58 UForgotten salt ’*’ saltutil.refresh_pillar
14:59 Psi-Jack I figured out a better solution. ;)
14:59 Psi-Jack {{ pillar["supervisor-progs"][group]["logdir"]|default("/var/log/supervisor") }}  this way it pulls from pillar but if undefined, provides a sane default.
14:59 Psi-Jack Heh, yay wierd jinja pipes!
15:00 UForgotten Psi-Jack: I read in the manual the way to do that is (salt[’pillar.get’](’pillarname’, ’default'))
15:01 Psi-Jack Yeah, there's that too, but that's also been flattened a bit, to salt['pillar.get']('pillar:name:subname', 'default')   so I wasn't sure the other would work.
15:01 Psi-Jack Since it's multiple levels deeps.
15:01 UForgotten right,and you're using variable substitution
15:01 Psi-Jack Yep.
15:02 Psi-Jack group is from an earlier for..loop in the jinja.
15:03 ashb salt['pillar.get']('supervisor-progs:' + group + ':logdir', '/var/log/supervisor')
15:03 fxhp Psi-Jack - the issue with the pillar+jinja-pipe with default is a keyerror can occur if pillar is not presented to the minion
15:03 Psi-Jack Hmmm..
15:04 Psi-Jack YEah, this is kinda wierd.. if my pillar value isn't defined, default is not making any bit of difference, which is wierd by itself.
15:04 Psi-Jack Ahh, nevermind, yes, it worked..
15:05 Psi-Jack fxhp: Ahhh, hmmm..
15:09 Psi-Jack fxhp: While it's unlikely, it could happen, so..
15:09 Psi-Jack ashb: Thanks, I'll use that method if it solves the issue fxhp mentioned? ;)
15:10 higgs001 joined #salt
15:11 hotsnow joined #salt
15:12 napper joined #salt
15:12 jeremyfelt joined #salt
15:14 wendall911 joined #salt
15:14 masterkorp hello
15:14 masterkorp http://pastie.org/private/tznadgnrnubtvunxbavmg
15:14 masterkorp i just bootstrapped a server
15:14 masterkorp test.ping works
15:15 masterkorp now i want to add the state.sls users to it so i can ssh to it
15:21 fxhp Psi-Jack: yes the pillar.get (Salt Execution Module Function) supports dictionary traversal and allows for a default if the traversal fails.
15:21 Psi-Jack Cool
15:21 Psi-Jack Well, thank you both then. :)
15:21 fxhp No problem
15:23 kermit joined #salt
15:27 conan_the_destro joined #salt
15:28 rgbkrk joined #salt
15:30 vejdmn joined #salt
15:30 Katafalkas joined #salt
15:39 ajw0100 joined #salt
15:44 bhosmer joined #salt
15:45 eliasp any idea why I get this when running "salt MN1221-C0008 state.show_lowstate"?
15:45 eliasp TypeError encountered executing state.show_lowstate: 'NoneType' object is not iterable. See debug log for more info.  Possibly a missing arguments issue:  ArgSpec(args=['queue'], varargs=None, keywords='kwargs', defaults=(False,))
15:45 gildegoma joined #salt
15:46 eliasp ah, please ignore me… outdated minion version
15:50 smcquay joined #salt
15:50 seapasulli_ joined #salt
15:51 ipalreadytaken joined #salt
15:52 Damoun hello
15:52 Damoun i'm doing some states to use on my personal server
15:53 Damoun i'm working on a state to configure a MX Backup
15:54 Damoun Do I have to write a module to select a mx hostname from DNS record ?
15:56 TheRealBill_here joined #salt
15:56 Linz joined #salt
15:56 ndrei joined #salt
15:57 fedgoat use the built in dnsutil module?
15:57 moos3 by default how often does the minion's refresh /srv/salt for changes ?
15:57 fedgoat salt ns1 dig.MX domain.com
15:57 smcquay joined #salt
15:57 masterkorp anyone ? i am cluekless
15:57 fedgoat http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.dnsutil.html
15:59 Damoun fedgoat: yes, but I have 3 MX, How can I allocate MX hostname per host ?
16:02 fedgoat not sure what you mean?  when you dig the MX records for a hostname it will put the MX records in a list
16:02 vejdmn joined #salt
16:03 Kenzor joined #salt
16:06 topochan joined #salt
16:07 vejdmn joined #salt
16:08 Damoun in fact, I don't know how to do :/
16:09 ajw0100 joined #salt
16:10 yomilk joined #salt
16:10 KyleG joined #salt
16:10 KyleG joined #salt
16:14 ndrei joined #salt
16:21 eliasp shit… just deployed a salt-minion update to 15 Win clients and left a ' too much in the /master=… silent installation switch… manual mode engaged! :)
16:22 eriko joined #salt
16:22 Damoun ok, I think I'm doing it wrong, I should use interface instead of hostname
16:24 joehillen joined #salt
16:24 Damoun https://gist.github.com/Damoun/9977917 my states now (for opensmtpd ;))
16:26 meteorfox joined #salt
16:28 eykd joined #salt
16:28 schimmy joined #salt
16:28 ajw0100 joined #salt
16:29 druonysus joined #salt
16:29 druonysus joined #salt
16:29 UtahDave joined #salt
16:30 sandbender1512 @basepi or @thatch45 sitting in here by any chance?
16:30 sandbender1512 (github usernames)
16:30 basepi definitely note.
16:30 UtahDave basepi is here.
16:30 basepi not*
16:31 basepi man alive, I broke my joke.
16:31 basepi stupid fingers.
16:31 UtahDave go take a nap, basepi
16:31 * basepi zzzzzzzzzzzzzzzzz
16:31 sandbender1512 lol
16:31 sandbender1512 basepi: you guys're gonna hate me by the time I get this sls working
16:31 sandbender1512 (if you don't already ;)
16:31 basepi hahaha
16:31 basepi another bug?
16:32 sandbender1512 indeed
16:32 sandbender1512 good news is yesterday's fix worked for that bug/case
16:32 basepi file away!  i'll forward it to thatch45, who doesn't come on IRC because people mob him when he does.
16:32 sandbender1512 but I hit another setup that results in inf. recursion
16:32 basepi yay recursion!
16:32 * sandbender1512 nods
16:33 sandbender1512 I'm about to lunch but I'll type it up after lunch
16:33 basepi cool
16:33 sandbender1512 still gotta finish simplifying the test sls as much as possible for reproduction purposes
16:34 eriko Has anyone been successful getting Windows minions to stay connected across a VLAN/firewall? I’ve been experimenting with all sorts of things including pinging them from a cron job every minute, but no dice. Minions on the same VLAN have no issues.  Just wondering if I should keep trying or give up :)
16:35 basepi sandbender1512: we appreciate the simplifications.  not everyone takes that step, but it sure makes testing and narrowing down the problem easier.
16:35 basepi sandbender1512: does the 2 on the end mean you're away or something?  o.O
16:35 basepi or did my irc client truncate your name?
16:36 oz_akan_ joined #salt
16:36 basepi this is what i saw, but then tab-completion gave me 1512:  http://cloud.basepi.net/image/2L092O3N430F/Image%202014-04-04%20at%2010.35.58%20AM.png
16:36 basepi if it's truncating names i need to fix that
16:36 basepi eriko: do you have the latest version of the windows minion?
16:37 basepi eriko: ZMQ keepalive was broken on Windows until ZMQ 4.0.4, which is only included in the latest set of installers
16:38 eriko basepi: good to know, I was running 2014.1.0 until about an hour ago, I’ll keep trying with 2014.1.1. Thanks!
16:38 basepi eriko: kk
16:38 basepi eriko: I'm sure there are still probably some network setups that will struggle, but we've seen a lot of the problems go away
16:38 sandbender1512 the 2 on the end means I'm too lazy to auth against NickServe and get my 'real' nick of sandbender most of the time ;)
16:39 basepi sandbender1512: it does appear that my client is truncating, though.  curses.
16:39 sandbender1512 1512 is an old school nick-suffix I've been using for ages
16:40 masterkorp ok i am clueless
16:40 masterkorp http://pastie.org/private/tznadgnrnubtvunxbavmg
16:40 masterkorp any idea ?
16:40 oz_akan_ joined #salt
16:41 viod1 left #salt
16:44 sandbender there we go :)
16:44 chrisjones joined #salt
16:47 basepi masterkorp: it appears you're trying to template using mako without mako being properly installed
16:47 basepi without seeing the states, I can't tell for sure.
16:48 masterkorp how do i show you the states ?
16:48 basepi masterkorp: pastebin the users.sls file.  you should remove any information you don't want public, of course.
16:50 smurfy__ joined #salt
16:51 kballou joined #salt
16:51 masterkorp it just creates 2 users and gets ssh keys in there
16:51 masterkorp noting special
16:53 basepi does it have a shebang at the top or any sort of rendering pipe?
16:53 basepi somewhere in there you're telling it to render with mako, but it can't import mako
16:54 basepi either in your config, or in the state itself, do you have any references to mako?
17:00 masterkorp nope
17:00 basepi ...hrm....
17:00 masterkorp i search for mako on the file and nothing
17:01 masterkorp hint; i have no idea what mako is
17:01 basepi mako is a templating engine like jinja
17:02 basepi can you give me the output of `salt '*' test.versions_report` please?
17:06 JasonSwindle joined #salt
17:09 jmerripen joined #salt
17:10 vbabiy joined #salt
17:14 Ryan_Lane joined #salt
17:15 Dattas joined #salt
17:16 travisfischer joined #salt
17:17 rgbkrk joined #salt
17:18 JasonSwindle joined #salt
17:18 mortis is there a way to invert a run? like "oh shit i ran this crap on all our servers" and want to run salt "prod*" invertstate.sls only-on-dev-plz.sls
17:18 mortis :D
17:19 vbabiy joined #salt
17:19 Dattas If you use the backup function you can restore the files it changed
17:20 mortis was thinking more like pkg.install = pkg.purge and file.managed = file.remove or whatever it is
17:20 mortis without doing it manually bit by bit
17:21 Guest40962 win 3
17:23 Employee312 joined #salt
17:23 Gareth morning.
17:24 Employee312 joined #salt
17:29 stanchan joined #salt
17:31 meteorfox joined #salt
17:32 diegows joined #salt
17:32 druonysuse joined #salt
17:35 vbabiy joined #salt
17:35 ejames123 joined #salt
17:36 ejames123 example in => http://docs.saltstack.com/en/latest/topics/pillar/  is giving me this error => The state "packages" in sls packages is not formed as a list
17:36 raizyr joined #salt
17:37 ejames123 anyone around that would be willing to help me?
17:37 manfred which example?
17:37 ejames123 5.1 and 5.2
17:38 manfred you have the - ?
17:38 manfred it looks correct to me
17:38 ejames123 I copy and pasted from the web page.
17:38 ejames123 top.sls has the - packages
17:39 manfred can you pastebin it?
17:39 ejames123 I’m new to salt.  I’m running this on the minion: salt-call state.highstate
17:39 vbabiy joined #salt
17:39 manfred curl -F 'f:1=<-' ix.io < /srv/salt/top.sls
17:40 ejames123 on a vagrant host, no curl.  one sec.
17:40 ejames123 k done
17:40 ejames123 http://ix.io/btX
17:40 manfred and then the datadog/init.sls ?
17:41 ejames123 again, I copy and pasted.  I’m working with pillars
17:41 manfred oh sorry
17:41 ejames123 so it should have been /srv/pillar/top.sls
17:41 manfred i meant /srv/pillars/top.sls
17:41 ejames123 one sec
17:41 manfred yeah
17:41 manfred pillar*
17:41 ejames123 http://ix.io/btY
17:42 ejames123 packages.sls - http://ix.io/btZ
17:42 manfred hrm
17:42 manfred can I also see your datadog sls?
17:42 manfred just so i can see the whole environment?
17:42 ejames123 sure, one sec.
17:43 ejames123 http://ix.io/bu1
17:44 manfred hrm...
17:44 manfred i don't see anything. And I have to go.
17:44 ejames123 the error is around the pillar.
17:44 ejames123 thanks for your help manfred
17:44 ejames123 ;)
17:45 manfred well, it sys the 'state' packages in sls packages
17:45 manfred good luck!
17:45 ejames123 thanks
17:45 ejames123 packages is a pillar, so...
17:46 ejames123 I’m assuming it’s talking about the state of the pillar
17:46 ejames123 again, I’m new to salt
17:47 ajprog_laptop1 I am setting up gitfs. how do I test and see what is available at salt://?
17:48 Ahlee in a ext_pillar, is it best to just try: except block around checking if a __grain__ is set?
17:49 mgw joined #salt
17:49 joehoyle joined #salt
17:51 Psi-Jack Okay, so I have a state that has service:\n  - running  --- But, if I just remove that it won't try to make sure it's actually running, correct?
17:51 Psi-Jack I'm phasing out a service and running it under supervisord instead of lsb-init. :)
17:52 sdsk joined #salt
17:53 vbabiy joined #salt
17:53 Ahlee I believe so
17:54 ipmb joined #salt
17:55 bhosmer joined #salt
17:55 moos3 if I change a file in on my /srv/salt on the master, how do I get the minions to pull that file
17:55 rgbkrk joined #salt
17:58 srage joined #salt
18:01 eykd I know that for pillars, there's `salt '*' saltutil.refresh_pillar`
18:01 eykd But I don't think you have to do anything special for /srv/salt
18:02 eykd My changes take when I run `salt '*' state.sls mystate`
18:02 moos3 is there away to make the minions refresh that automatically
18:03 eykd cron job? :) I'm not sure how the pros do that, or if they even do. Sounds kind of dangerous, actually.
18:04 moos3 yeah true
18:06 nkuttler moos3: what kind of file do you mean? why can highstate and/or the file module not handle it?
18:07 raizyr joined #salt
18:08 moos3 i want to be able to update files in the my pillars and just have it push out
18:08 moos3 I can just make a wrapper to force a refresh
18:09 gamedna joined #salt
18:09 nmistry joined #salt
18:10 arthabaska joined #salt
18:11 meteorfox joined #salt
18:13 eliasp simply use the scheduler: http://docs.saltstack.com/en/latest/topics/jobs/schedule.html
18:15 moos3 eliasp do I up that on the master or the minions is my question
18:18 rgbkrk joined #salt
18:19 joehoyle joined #salt
18:21 juanlittledevil joined #salt
18:23 ajw0100 joined #salt
18:26 ashtonian joined #salt
18:30 ggoZ joined #salt
18:30 eykd I'm trying to use either a pip.installed or a virtualenv.managed state to set up a virtualenv using a requirements file that lives on the minion (installed by an earlier git.latest state). With either, when I try and use the requirements argument, I get this error: "Could not open requirements file: [Errno 2] No such file or directory: '/tmp/requirements.txt'", though that's not the path I'm passing to requirements.
18:31 eykd This looks kind of like a bug, like maybe /tmp/requirements.txt is hard-coded, because it expects me to use a salt:// URL instead of a local file path.
18:33 oz_akan_ joined #salt
18:33 sandbender basepi: #11773
18:34 * basepi looks
18:35 tedski joined #salt
18:35 nkuttler eykd: did you specify an absolute path to the requirements?
18:35 eykd nkuttler: Yes, I have.
18:35 JordanRinke joined #salt
18:38 eykd Here's the state in question: http://pastebin.com/LF4FKyj8
18:40 sijis joined #salt
18:40 sijis joined #salt
18:42 basepi sandbender: replied
18:44 ndrei joined #salt
18:45 srage_ joined #salt
18:45 mgw1 joined #salt
18:47 sandbender thx, I just replied to your reply :)
18:48 eykd I've posted my question on Stack Overflow, with additional details: http://stackoverflow.com/questions/22870751/setting-up-a-python-virtual-environment-with-salts-virtualenv-managed-or-pip-in
18:50 nkuttler eykd: /opt/django/mysite/requirements/production.txt exists on the minion, right?
18:50 eykd It does, I've confirmed that much. ;)
18:51 nkuttler eykd: and the debug log? the mysite user also has +rw?
18:52 oraqol joined #salt
18:52 nkuttler to me it looks like the tmp file is a fallback attempt
18:53 eykd the pip.log just has the traceback: InstallationError: Could not open requirements file: [Errno 2] No such file or directory: '/tmp/requirements.txt'
18:53 nkuttler i find it odd that requirements isn't documented for the pip module.. i've only used it with virtualenv.manage
18:53 JasonSwindle left #salt
18:54 nkuttler eykd: ah.. odd
18:54 eykd pip.installed looks to have all the same arguments as the pip 1.5.x CLI.
18:54 eykd nkuttler: I wondered about the fallback, but I never see an error about my requirements file, and it does exist.
18:55 nkuttler eykd: yeah, no idea, i was really just curious
18:55 srage joined #salt
18:56 eykd nkuttler: Ah well, thanks anyway.
18:56 [diecast] joined #salt
18:59 eliasp hmm, master + minion both on 2014.1, minion is Win7 amd64: state.show_lowstate returns: TypeError encountered executing state.show_lowstate:
18:59 eliasp 'NoneType' object is not iterable. See debug log for more info.  Possibly a missing arguments issue:  ArgSpec(args=['queue'], varargs=None, keywords='kwargs', defaults=(False,))
19:00 nkuttler eykd: fwiw, Could not open requirements file happens after a try: f = open(url)
19:01 nkuttler eykd: i'd try to run the command by hand on the minion. sounds like the http request is failing
19:01 nkuttler (and that error message doesn't seem to make much sense)
19:01 nkuttler ah, url seems to mean the local file..
19:02 ndrei joined #salt
19:03 eliasp can I pass the log-level to the minion at runtime when executing any modules from the master's CLI? I don't want to change the minion config each time + restart it do get some quick debug/log messages…
19:03 Psi-Jack Can you do multi-line comments in state files with {#\n\n\n\n#} ?
19:04 nkuttler Psi-Jack: jinja does that, yeah
19:04 Psi-Jack I've only ever done it by {# content ... #} for each line .;)
19:05 ksalman what is wrong with this state? the file exists in the filesystem https://gist.github.com/anonymous/9981142
19:06 Psi-Jack Cool that saves me a bit of work. All I needed due to this migration is that memcached state have pkg:installed and nothing else anymore. ;)
19:06 ksalman i don't know why the error has the filename with random characters in it
19:06 eliasp something like "{% if False %} whatever you want {% endif %}" _might_ work as well
19:07 eliasp or maybe it needs to be "{% if false is true %}…" … give it a try
19:10 oraqol I'm hoping you can help me with a very annoying issue I've noticed since upgrading my master and minions to 2014.1.0
19:10 oraqol [DEBUG   ] Attempting to authenticate with the Salt Master at xxx.xxx.xxx.xxx
19:10 oraqol [DEBUG   ] Loaded minion key: /etc/salt/pki/minion/minion.pem
19:10 oraqol [DEBUG   ] Loaded minion key: /etc/salt/pki/minion/minion.pem
19:10 oraqol some of the minions just hang at that
19:11 eliasp ksalman: had something like this in the past (not in Salt) and it turned out to be a weird non-rendered unicode char which somehow slipped in… maybe try to delete the line containing the path completely and rewrite it manually (no copy paste) again
19:11 jcsp1 joined #salt
19:11 oraqol no trace of them when I run the master in debug mode, but I can see the minion attempting to communicate when I run tcpdump on the master.  I already raised the ulimit to much more than necessary.  Any ideas?
19:11 eliasp oraqol: had this in 2014.1.0 as well and I believe I haven't seen it in 2014.1.1, so maybe try an upgrade to 2014.1.1 first
19:13 oraqol well i just ran apt-get install, which should upgrade the minion, but it says up to date at 2014.1.0
19:13 oraqol Has the new version been pushed out to the repos?
19:14 oraqol I'm running 10.04 btw
19:18 combusean joined #salt
19:18 UtahDave oraqol: Ubuntu 10.04?
19:18 oraqol yes
19:18 oraqol the ppa I'm using is ppa:saltstack/salt
19:18 oraqol is this correct?
19:19 babilen oraqol: Even though a new version has been released by salt and can be found in the Debian repositories I was under the impression that the Ubuntu PPA has not yet been updated. (not sure why)
19:19 oraqol dangit
19:19 ipalreadytaken joined #salt
19:19 oraqol that's up to canonical to update, right?
19:19 babilen oraqol: So, you might want to consider running Debian if you want the new version.
19:20 babilen oraqol: No, I think that is also handled by the Debian maintainer, but let me check that.
19:20 combusean hey, is there any particular reason I would get SaltReqTimeoutError: Waited 60 seconds in the minion log from the rpm.  /etc/salt/master and minion have minimal configuration in them and i can ping the salt master fine and telnet into it on 4505/4506 from the minion
19:20 combusean erm, in salt installed from the rpm
19:21 oraqol and THIS is exactly why I've been harping on my higher-ups to implement an internal god$$%#ed repository :/
19:21 babilen oraqol: Yeah, joehh is definitely involved there too. See https://launchpad.net/~saltstack/+archive/salt for an overview
19:21 UtahDave combusean: set     recon_randomize: True   in your minions' configs
19:22 combusean the minion's just have master and user configured.
19:22 UtahDave combusean: how many minions?
19:22 combusean 1 so far
19:22 combusean the master hasn't been able to accept its key yet
19:23 combusean is recon_randomize documented?
19:25 tyler-baker joined #salt
19:25 oraqol babilen: I'm not sure what I'm looking at here.  Is 2014.1.1 up on the ppa, is that what the link means?
19:26 UtahDave combusean: oh, that's odd.  What version of Salt are you on?
19:26 oraqol oop, nevermind, I see the latest version of salt is .0
19:26 combusean nm, found it
19:26 oraqol Do you guys have an estimated time when the new version will be uploaded?
19:26 oraqol I hate be a pain but my bosses are freaking out :/
19:27 combusean # salt --version
19:27 combusean salt 2014.1.0
19:27 UtahDave combusean: that shouldn't be a problem with only one minion.
19:27 combusean heh
19:27 oraqol yeah combusean, I think we're are experiencing the same problem here
19:29 combusean does salt-key -L list anything on your master oraqol?
19:29 oraqol yes
19:30 combusean i don't have anything yet
19:30 oraqol I have many minions already enrolled
19:30 oraqol and the minions experiencing this issue have already been enrolled
19:30 combusean just to be sure the minions don't need any open ports, do they?
19:30 oraqol they return when I salt-key -L | grep xxx
19:30 raizyr joined #salt
19:30 eliasp combusean: only on the master… see also: http://docs.saltstack.com/en/latest/topics/tutorials/firewall.html
19:31 combusean yeah i was just checking
19:31 combusean doesn't appear to be a fw connection issue
19:31 oraqol I can second that, tcpdump shows the minion hitting the master
19:34 combusean [root@ip-10-0-1-251 salt]# grep master /etc/salt/minion
19:34 combusean master: deployer.staging.[redacted]-labs.com
19:34 combusean [root@ip-10-0-1-251 salt]# telnet deployer.staging.[redacted]-labs.com 4505
19:34 combusean Trying 10.0.1.249...
19:34 combusean Connected to deployer.staging.[redacted]-labs.com.
19:34 combusean
19:34 combusean [root@ip-10-0-1-251 salt]# telnet deployer.staging.[redacted]-labs.com 4506
19:34 combusean Trying 10.0.1.249...
19:34 combusean Connected to deployer.staging.[redacted]-labs.com.
19:34 combusean
19:35 combusean sorry about the redacteds, i'm under an nda
19:36 Psi-Jack Wait a sec.. You;'re from puppet-labs, aren't you? :D
19:36 UForgotten Psi-Jack: I had the same thought, lol
19:36 Psi-Jack hehe
19:36 eliasp I thought about bell labs ;)
19:36 UForgotten that dog food got stale I suppose
19:36 combusean no
19:36 oraqol LOL
19:36 combusean it's a mobile security startup
19:37 Psi-Jack Heh. Okay. Was kidding anyway. ;)
19:37 combusean i just don't want people googling them and finding stuff in here
19:37 chrisjones joined #salt
19:37 combusean as i'm not sure we release the details of our infrastructure
19:37 UForgotten not if its security related... :)
19:37 combusean :)
19:38 oraqol So guys...what do I do?  Just wait for the update to be rolled out to the ppa?
19:38 combusean should I try installing this from the bootstrapper instead of the epel rpm?
19:38 jcockhren oraqol: or create a salt package yourself. see fpm
19:38 jcockhren salt *rpm*
19:40 UForgotten can also override whatever's broken in your /srv/salt/_modules _states etc on the master
19:40 UForgotten but I guess you'd have to determine what is broken, haven't been tuned in long enough to see
19:40 UForgotten and I'm a total salt noob myself :)
19:40 kzx any chance of updating salt rpms in EPEL? I use those for now but 2014.1 has some problems so waiting for an update (2014.1.1)
19:41 ajw0100 joined #salt
19:41 babilen joehh: We are getting quite a number of questions regarding the upload of 2014.1.1 to the Ubuntu PPA - What is preventing it from going ahead and would it be an option for users to try, say, to update the packages with uupdate? Is the .dsc available somewhere?
19:42 babilen oraqol: Which Ubuntu version are you on again?
19:43 eliasp babilen: I don't know much about Debian/Ubuntu packaging, but it seems like PPA is causing quite some pain for some projects… Docker migrated away from it as well ~6 months ago
19:44 babilen eliasp: I am not involved in Ubuntu nor am I particular fond of PPAs, but recompiling a package and updating it to a new minor version shouldn't be too hard.
19:45 masterkorp @basepi | can you give me the output of `salt '*' test.versions_report` please?
19:45 babilen oraqol: You are also running quite an old Ubuntu version, it might be that some of the build dependencies are, well, tricky to work with salt.
19:45 oraqol I'm on 10.04
19:46 oraqol babilen: I would agree with you, but only a handful of the minions are exhibiting this behavior, and most of them run fine, all of them are on 10.04
19:46 juanlittledevil joined #salt
19:46 oraqol which means they install fine, and may even start out working fine, but break somewhere along the line
19:46 masterkorp basepi: http://pastie.org/private/j7iv5qsyknszfd4npxuq
19:47 masterkorp sorry for the delay
19:47 babilen oraqol: Either way, you could grab the source package with, say, "dget https://launchpad.net/~saltstack/+archive/salt/+files/salt_2014.1.0-1lucid3.dsc" and the tarball for salt 2014.1.1. Use uupdate to update the source package (e.g. "cd PACKAGE_DIR ; uupdate ../salt_2014.1.1.tar.gz" and then recompile the package.
19:48 babilen oraqol: If you don't have a package builder setup you can use something like "dpkg-buildpackage -uc -us" -- See also https://www.debian.org/doc/manuals/maint-guide/update.en.html#newupstream
19:49 babilen oraqol: Sorry, I can't really do more right now and would hope that an "official" version will be uploaded soon, but that should give you something to work on in the interim.
19:50 oraqol Ok, I'll send it up the foodchain, we may just end up waiting for the ppa to be updated, as adding anything to our internal repo is a pain in the @$$
19:50 oraqol we have one, btw, i just asked :P
19:52 faldridge joined #salt
19:53 babilen oraqol: It shouldn't be a pain to add something to your internet repo, but just a "dput YOURNAME foo_0.1-1.changes", but then you might have all sorts of bureaucracy between you and the repo. Another easy way to get a repo fast is to "dpkg-scanpackages . /dev/null | gzip > Packages.gz" in the directory with your packages. Then just make that available to the clients and you are done.
19:54 babilen oraqol: That really is the most basic way to build and host a package, but it might be all you need to deal with the current fallout. I can suggest increasingly sophisticated schemes, but then .. in the end .. you really want the maintainer to just upload a new version to the PPA.
19:54 babilen oraqol: Might also want to mention that Debian has the new version already so a switch might be an option too.
19:55 babilen Alternatively forget the entire package approach and deploy using the boostrap script and use the git tag ... http://docs.saltstack.com/en/latest/topics/tutorials/salt_bootstrap.html
19:56 oraqol Yes, looks like I'll be making the switch soon.  And yes its technically easy to add things to repos, but I meant a red-tape pain in the butt
19:56 ksalman how miight i run an arbitrary command based on some grain?
19:56 ksalman like 'salt '*' test.ping'
19:56 thayne joined #salt
19:56 manfred salt -G 'os: Ubuntu' test.ping
19:56 oraqol salt -G "key:pair" cmd.run "sudo rm -rf /"
19:56 eliasp ksalman: see http://docs.saltstack.com/en/latest/topics/targeting/grains.html
19:57 oraqol DON'T RUN THAT
19:57 oraqol lol
19:57 manfred oraqol: needs more --no-preserve-root
19:57 ksalman thanks =)
19:57 babilen oraqol: Err, please do not suggest commands like that. Even in jest.
19:57 oraqol sorry sorry
19:57 babilen oraqol: I mean it wouldn't do anything, but then you never know what people might actually run
19:57 oraqol mea culpa [ashamed]
19:57 babilen oraqol: That is something that would own you a +q or +b in #debian and a couple of other channels. Either way, lets move on.
19:58 oraqol Thank you!
19:59 eliasp #debian … where I got banned once ~15 years or so ago because I stated I'd be AFK for 15 minutes to pick someone up… "we're not idling here…" yeehaw ;)
20:00 oraqol please don't ban me, I promise not to do it again :/
20:01 Psi-Jack Anyone that ever suggests an rm command like that, especially via salt, SHOULD be banned.
20:01 Psi-Jack It's /never/ funny.
20:02 nkuttler manfred: also check the targeting docs
20:02 conan_the_destro joined #salt
20:02 babilen eliasp: I would be open to discuss lifting that ban ;)
20:03 manfred wat?
20:03 eliasp babilen: well, I think it was 14.95 years ago or so ;)
20:03 babilen hehe
20:04 oraqol left #salt
20:05 oraqol joined #salt
20:05 oraqol left #salt
20:06 basepi masterkorp: weird one.  can you file an issue on github?  include as much information as possible.
20:10 danielbachhuber joined #salt
20:11 TheRealBill joined #salt
20:12 oraqol joined #salt
20:13 oraqol left #salt
20:14 rglen joined #salt
20:15 ajw0100 joined #salt
20:18 higgs001 joined #salt
20:22 yurei joined #salt
20:23 masterkorp will do
20:28 harobed joined #salt
20:32 naiyte joined #salt
20:33 jaimed joined #salt
20:40 [diecast] joined #salt
20:41 ashtonian joined #salt
20:45 druonysus joined #salt
20:47 obimod joined #salt
20:48 druonysus joined #salt
20:50 eliasp had all my SLS as foo/init.sls until now and restructed everything a bit… e.g. I have now win/common.sls: http://pastie.org/8995008 running "state.show_low_sls win.common" results in 9 times "The state "common" in sls win.common is not formed as a list"
20:51 eliasp any idea what should be wrong with this YAML/SLS?
20:51 juanlittledevil left #salt
20:55 eliasp I already tried to make a list out of all the states listed in this file, but the result is then "Name common in sls win.common is not a dictionary"
20:56 FarrisG Is anyone using the nginx formula to install nginx from source (or any similar formula) who can help me modify the state/formula to handle patches as well as extra modules? I want to include the syslog patch from https://github.com/yaoweibin/nginx_syslog_patch
20:59 smcquay joined #salt
21:01 ksalman at what point do i need to worry about more then 1 salt master? i have 300+ minions and when i do things like 'salt \* test.ping' it just times and says the worker threads may need to be increased. I changed threads to 20 but it's still the same
21:01 Kenzor joined #salt
21:02 ksalman more over the salt-master processes are using all the available cpu
21:02 gladiatr joined #salt
21:02 ksalman is my saltmaster just getting DOS'd by the minions?
21:04 gladiatr greetings. Does it work to use wildcards with pillar includes?  My initial instinct is to build pillar parameter files for non-default configs per host.  is it possible to do something like: include:^J^I- service.hosts.*
21:18 ajw0100 joined #salt
21:23 eliasp converting a SLS from a foo/bar/init.sls to foo/bar.sls should be just indenting all existing content by 1 + adding "bar:" on top, right?
21:25 nkuttler eliasp: hrm, just moving the file doesn't do?
21:25 nkuttler not really sure why you'd need to adjust anything
21:26 eliasp nkuttler: uuh, ok… embarassing… where did I pick this up? it works now
21:26 eliasp I think I got it here from: http://docs.saltstack.com/en/latest/topics/tutorials/states_pt3.html#name-declaration
21:27 eliasp so simply moving the file is actually all that's needed without repeating the SLS name top-level in the file
21:27 ashtonian joined #salt
21:27 rglen joined #salt
21:28 eliasp but I see now… the repeated "mywebsite" in the example above doesn't refer to the SLS name at all, so it's just a coincidence they're having matching names
21:29 combusean should I worry about not having dmidecode or smbios?
21:30 combusean pops up in chef-minion -l trace
21:30 eliasp combusean: not really… I might provide some helpful additional grains, but it won't do any harm as long as you don't have to rely on the grains provided by them
21:30 eliasp s/I might/it might/g
21:30 * combusean sighs
21:31 combusean i can telnet to the salt master on 4505 and 4506 but salt-minion times out trying to talk to it
21:32 jeremyfelt joined #salt
21:32 harobed_ joined #salt
21:33 rglen joined #salt
21:35 harobed joined #salt
21:36 eliasp nkuttler: ok, all problems solved… my highstate applies now properly… also made the mistake to use relative instead of canonical names for my "include" statements
21:40 napper joined #salt
21:41 druonysuse joined #salt
21:47 hunter joined #salt
21:49 ksalman i am getting a lot of tcp connections on the master. I started the master with debug log level and i am seeign a lot of these messages https://gist.github.com/anonymous/9983755
21:49 ksalman any idea why?
21:50 ksalman lots of auth and the "clear payload received" messages
21:50 ksalman clients auth and gets accepted over and over
21:50 eliasp ksalman: that's nothing to worry about… the minions are just asking your master for new jobs
21:51 ksalman well, should they be connecting to the master over in a matter of seconds?
21:52 eliasp ksalman: at least your log-excerpt doesn't list any minion twice, so it looks good to me…
21:52 ksalman the master appears to be getting DOS'd because things time out when I try to run a highstate
21:52 ksalman eliasp: there are same minions over and over, i just pasted an excerpt
21:53 ksalman all of the minions are in the list, over and over
21:53 ksalman all 200+ of them
21:53 ndrei joined #salt
21:54 harobed_ joined #salt
21:57 joehoyle joined #salt
22:00 ipalreadytaken joined #salt
22:02 joehoyle_ joined #salt
22:05 JordanRinke joined #salt
22:09 harobed joined #salt
22:25 napper joined #salt
22:26 thayne joined #salt
22:26 meydvr joined #salt
22:27 Katafalkas joined #salt
22:28 louie joined #salt
22:29 chrisjones joined #salt
22:42 harobed joined #salt
22:43 Ryan_Lane joined #salt
22:46 jeremyfelt joined #salt
22:52 ajprog_laptop joined #salt
22:53 Katafalkas joined #salt
22:55 Fin1te joined #salt
22:55 Fin1te left #salt
23:03 dmick joined #salt
23:04 Damoun joined #salt
23:04 dmick Hi guys; I have what may be a very silly question.   Our project is using salt-call --local state.template to execute a .sls for some local setup (as part of a larger salt scheme)
23:05 dmick I need to make some of it conditional on os type, which would lead me to think of grains, but since minion selection isn't involved, I'm not sure there's a way to express that
23:05 dmick it feels like I'm barking up the wrong tree, but, can it be done somehow I'm just not seeing?
23:07 pydanny joined #salt
23:10 Katafalkas joined #salt
23:10 UtahDave dmick: you can check for the os grain inside of jinja in your sls file
23:10 harobed joined #salt
23:11 dmick ooh, that's a thought
23:11 dmick (the fact that salt-call has a -g that works leads me to believe it actually has the info...)
23:11 dmick that might well work.  thanks.
23:12 kermit joined #salt
23:18 bhosmer joined #salt
23:22 alunduil joined #salt
23:24 Nazca__ joined #salt
23:26 ndrei joined #salt
23:28 UtahDave druonysuse: you around?
23:31 possibilities joined #salt
23:31 Katafalkas joined #salt
23:42 Ryan_Lane joined #salt
23:51 Katafalkas joined #salt
23:52 Katafalkas joined #salt

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