Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-10-22

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

All times shown according to UTC.

Time Nick Message
00:00 packeteer oh ffs, my work has "standardized" on chef
00:01 amckinley anyone around to help debug a problem invoking salt-call? http://pastebin.com/UbaYzhuw
00:03 Furao joined #salt
00:03 amckinley my state module: http://pastebin.com/UHnS8bVe
00:04 cjh_ i don't see a get call anywhere in your state module
00:04 amckinley cjh_: ?
00:05 cjh_ your stack trace says list object has no attr get on it
00:05 cjh_ i was guessing you were calling get on a list
00:05 amckinley yeah, coming from the sale internals
00:05 amckinley *salt
00:05 amckinley im assuming this is getting thrown right before or right after my code gets invoked
00:05 cjh_ yeah i don't know the internals at all
00:06 cjh_ it looks like line 18 is a list of dict objects?
00:06 amckinley i thought i might be returning a list somewhere salt expects a dict, but that didnt change anything
00:06 cjh_ right
00:06 cjh_ i think it expects a dict back
00:06 amckinley yeah one of these:
00:06 amckinley ret = {'name': path, 'changes': {}, 'result': True, 'comment': ''}
00:06 cjh_ that seems legit
00:07 amckinley you can see on line 46 where i hacked it to just return right away
00:07 cjh_ yup
00:07 cjh_ i have a python module that works.  lemme dig it up
00:07 cjh_ it might help you
00:08 cjh_ amckinley: http://fpaste.org/48472/13824005/
00:08 cjh_ that works for me
00:09 amckinley cjh_: thanks
00:09 cjh_ i'm returning a list of dicts i believe
00:09 cjh_ sorry a list of lists of dicts, it's confusing
00:10 cjh_ in any event it runs haha
00:10 amckinley cjh_: cool. ive got some more debugging info; looks like my module has a missing dependency, which is causing salt to choke later on
00:11 cjh_ oh..
00:11 cjh_ that'd do it haha
00:11 amckinley still pretty heinous to blow up the way it does… might file a bug on salt if i get this working
00:12 juicer2 joined #salt
00:13 ronc joined #salt
00:16 racooper joined #salt
00:21 jslatts joined #salt
00:23 Furao joined #salt
00:26 amckinley im having a hard time getting my code/test/debug cycle working because salt keeps caching stale versions of the modules im working on. is there an easy way to tell the minion to always pull straight from the filesystem instead of checking cache first?
00:27 taylorgumgum joined #salt
00:29 Brew joined #salt
00:29 redondos joined #salt
00:34 __number5__ amckinley: IIRC, if you do salt-call --local on the minion it won't use the cache (or actively refresh the cache)
00:35 amckinley __number5__: yeah, it looks like that flag just tells the minion to stop attempting to fetch from the server. definitely still uses the cache
00:39 jslatts joined #salt
00:39 __number5__ hmm, I never have caching problem when develop modules using masterless minion, but might be my modules all just a single python files
00:40 amckinley __number5__: yeah i have to call salt-call saltutil.sync_all to get salt to pick up my changes
00:43 sixninetynine joined #salt
00:43 Furao joined #salt
00:45 AdamSewell joined #salt
00:47 cshuman joined #salt
00:51 Drawsmcgraw joined #salt
00:52 yota joined #salt
00:53 __number5__ gvio
00:53 __number5__ sorry
00:54 Gifflen joined #salt
01:00 TJ1980 joined #salt
01:01 pviktori_ joined #salt
01:01 __number5__ amckinley: I know why I'm not having the syncing problem, because I'm working directly on /srv/salt/_modules on minion directly
01:03 Furao joined #salt
01:19 rgbkrk joined #salt
01:22 avienu joined #salt
01:22 kevin joined #salt
01:22 phab3k joined #salt
01:24 chuffpdx joined #salt
01:26 ajw0100 If I change fileserver_backend to only pull from git and then restart the master service I am getting the following result after running salt '*' state.highstate test=true: https://gist.github.com/ajw0100/7093274
01:27 ajw0100 If I continue to run salt '*' state.highstate test=true a bunch of times eventually it picks up the remote top.sls
01:27 ajw0100 but there is no consistency in how long it takes
01:27 ajw0100 is this expected?
01:31 redondos joined #salt
01:34 Furao joined #salt
01:34 mannyt joined #salt
01:37 jslatts joined #salt
01:38 deepakmd_oc joined #salt
01:40 mannyt joined #salt
01:41 oz_akan_ joined #salt
01:42 Lue_4911 joined #salt
01:46 v0id_ joined #salt
01:49 m0hit joined #salt
01:52 mannyt_ joined #salt
01:53 taylorgumgum joined #salt
01:53 berto- joined #salt
01:54 Furao joined #salt
01:55 smccarthy joined #salt
01:58 jtrucks joined #salt
01:59 UtahDave joined #salt
01:59 ajw0100 I reverted back to 0.17.0 and I don't have this issue
02:04 Furao joined #salt
02:06 UtahDave ajw0100: what issue are you seeing?
02:07 m0hit joined #salt
02:08 ajw0100 with 0.17.1-1precise if I change fileserver_backend to only pull from git and then restart the master service I am getting the following result after running salt '*' state.highstate test=true: https://gist.github.com/ajw0100/7093274
02:08 ajw0100 If I continue to run salt '*' state.highstate test=true a bunch of times eventually it picks up the remote top.sls
02:08 Lue_4911 joined #salt
02:08 ajw0100 but there is no consistency in how long it takes
02:09 Gareth hm. is it possible to look up something in pillar using a value from another pillar item?
02:09 ajw0100 with 0.17.0 it picks up the remote top.sls after a few seconds, which I'm guessing is just the time it takes to clone the repos into the local cache.
02:10 malinoff joined #salt
02:11 ajw0100 with 0.17.1 it can take over a minute and it's not consistent so somethings is clearly not right
02:12 malinoff Hi all
02:13 ajw0100 the consistency might be my connection or something. but in comparison it is still taking way longer on 0.17.1 than on 0.17.0
02:26 carmony joined #salt
02:37 joehh anyone know how I'd get in touch with redbeard?
02:37 joehh a name that I can search for an email would be handy
02:37 packeteer build a time machine?  :P
02:39 xl1 joined #salt
02:41 UtahDave ajw0100: That's odd.  Are you getting any errors in the logs?
02:41 renoirb joined #salt
02:41 UtahDave joehh: Yeah, I'll pm you his info
02:41 redondos joined #salt
02:42 joehh thanks UtahDave
02:42 UtahDave you're welcome
02:44 ajw0100 UtahDave I ran it with -l debug and I don't remember seeing anything unusual but let me switch back to 17.1 and check
02:44 UtahDave ok
02:47 AdamSewell joined #salt
02:51 ajw0100 UtahDave added the output of -l debug to the gist: https://gist.github.com/ajw0100/7093274
02:53 jalbretsen joined #salt
02:54 carmony UtahDave: So I migrated my personal servers to Digital Ocean
02:55 malinoff joehh, Do you have a news about salt v0.17.1 in the debian repository? :)
02:55 UtahDave ajw0100: Dang, nothing helpful there.   would you mind opening an issue about this?
02:55 UtahDave carmony: ah, cool.  How many vms?
02:55 carmony UtahDave: I just consolidated down to 1 vm
02:55 carmony I did have 2
02:56 ajw0100 UtahDave sure
02:56 UtahDave thanks, ajw0100
02:57 UtahDave carmony: cool.  They have a nice api, too.  So far they've been pretty great to work with, especially for the price
02:57 foxx[cleeming] joined #salt
02:57 foxx joined #salt
03:01 UtahDave carmony: did you get your vagrant / salt-master thing going?
03:03 carmony UtahDave: yeah, made some progress on it
03:04 carmony I fought vagrant-salt until I realized that it installs an older version of salt
03:04 Furao joined #salt
03:04 Gareth Should this be possible:
03:05 Gareth {% set pvalue = "available_zones:" + key + ":file" %}
03:05 Gareth {% set file = salt['pillar.get'](pvalue) %}
03:05 Gareth inside a SLS.
03:05 carmony I also setup salt to manage my personal stuff
03:06 taylorgumgum joined #salt
03:06 UtahDave Gareth: Yeah, I think that should work
03:06 UtahDave carmony: cool
03:08 lesnail joined #salt
03:10 Gareth UtahDave: looks like it does :) was looking at the wrong bit of output. thanks :)
03:11 UtahDave Gareth: ah, cool.   :)
03:12 redondos joined #salt
03:13 lesnail joined #salt
03:17 carmony UtahDave: alright, I'm going to try to see if my salt-cloud stuff now works :P
03:17 taylorgumgum joined #salt
03:18 ajw0100 joined #salt
03:18 UtahDave carmony: :)  I'd better go to bed now, then. lol
03:19 carmony lol, hey now...
03:21 ajw0100 joined #salt
03:24 Furao joined #salt
03:24 carmonyWeb joined #salt
03:24 carmonyWeb lol, and as soon as I start singing their praises, I loose connectivity to their servers
03:24 nmistry joined #salt
03:24 carmonyWeb lose*
03:25 carmonyWeb UtahDave: Can you get to www.digitalocean.com ?
03:28 rgarcia_ joined #salt
03:28 Gareth carmonyWeb: doesn't work for me.
03:28 pentabular joined #salt
03:28 Gareth actually now it does.
03:28 pentabular joined #salt
03:28 avienu joined #salt
03:29 redondos joined #salt
03:29 redondos joined #salt
03:30 phab3k joined #salt
03:30 cshuman joined #salt
03:31 carmonyWeb yeah, looks like they are experiencing some problems
03:33 MTecknology joined #salt
03:34 carmony joined #salt
03:34 mua joined #salt
03:34 [ilin] joined #salt
03:35 joehh malinoff: still working on them - mainly tracking down a few debian policy items
03:36 joehh I'll definitely have something up tonight, even if it is not absolutely perfect and ready to submit to the debian archive...
03:37 cachedout joined #salt
03:38 malinoff joehh, Cool :)
03:38 pmrowla joined #salt
03:38 throwanexception joined #salt
03:39 pmrowla joined #salt
03:40 lesnail Hey everyone, I want to push messages to my IRC when certain events in salt happen. What I'm currently doing is using a reactor sls file with the pure python renderer pushing the message and returning an empty dictionary so no other action is triggered. I wonder if there is a more idiomatic way of doing this.
03:42 lesnail Or maybe a best practice for doing these kind of things
03:48 UtahDave joined #salt
03:49 carmony UtahDave: lol, I really did think you ran from me ;D
03:52 druonysus joined #salt
03:53 malinoff lesnail, are you talking about firing events? What kinds of events do you have?
04:04 Furao joined #salt
04:09 druonysus joined #salt
04:09 cshuman joined #salt
04:11 m0hit joined #salt
04:15 mwillhite joined #salt
04:19 mafrosis joined #salt
04:19 josephholsten joined #salt
04:20 UtahDave joined #salt
04:21 mafrosis yo Salt peeps
04:22 mafrosis is anyone aware of a bug concerning apt-get sometimes failing since 0.17?
04:22 [ilin] joined #salt
04:23 mafrosis I have a pkg.latest state which has recently been failing "with return code: 100"
04:23 mafrosis some other installs do the same thing, but only intermittently
04:23 josephholsten joined #salt
04:23 mafrosis some other packages*
04:26 jefimenko joined #salt
04:27 josephholsten joined #salt
04:37 UtahDave left #salt
04:37 throwanexception joined #salt
04:39 carmony joined #salt
04:41 [ilin] joined #salt
04:41 rgarcia- joined #salt
04:44 pmrowla joined #salt
04:49 [ilin] joined #salt
04:50 pmrowla joined #salt
04:54 Furao joined #salt
04:56 pmrowla joined #salt
04:59 [ilin] joined #salt
05:02 redondos joined #salt
05:09 josephho_ joined #salt
05:13 pmrowla joined #salt
05:14 carmony joined #salt
05:14 redondos joined #salt
05:14 redondos joined #salt
05:15 pentabular joined #salt
05:16 throwanexception joined #salt
05:17 [ilin] joined #salt
05:23 jasonrm joined #salt
05:24 josephholsten joined #salt
05:24 Furao joined #salt
05:25 pentabular joined #salt
05:25 vipul joined #salt
05:30 josephho_ joined #salt
05:36 packeteer hmm, i'm getting thintar errors using salt-ssh (17.1 package on ubuntu) ie. NameError: global name 'thintar' is not defined
05:36 TheCodeAssassin joined #salt
05:42 josephholsten joined #salt
05:48 josephholsten joined #salt
05:59 redondos joined #salt
05:59 redondos joined #salt
06:02 carmony joined #salt
06:03 pmrowla joined #salt
06:03 anuvrat joined #salt
06:04 [ilin] joined #salt
06:05 Destro joined #salt
06:06 middleman_ joined #salt
06:07 throwanexception joined #salt
06:10 MK_FG joined #salt
06:11 josephholsten joined #salt
06:12 gildegoma joined #salt
06:14 josephholsten joined #salt
06:16 matanya joined #salt
06:17 balboah joined #salt
06:23 MohShami joined #salt
06:25 josephholsten joined #salt
06:27 bud joined #salt
06:32 pmrowla joined #salt
06:33 phab3k joined #salt
06:33 [ilin] joined #salt
06:36 jhujhiti joined #salt
06:38 josephholsten joined #salt
06:40 josephholsten joined #salt
06:41 josephholsten joined #salt
06:45 spicyWith joined #salt
06:48 phab3k joined #salt
06:53 anuvrat joined #salt
06:53 josephho_ joined #salt
06:53 taylorgumgum joined #salt
06:58 phab3k joined #salt
07:06 puffrfish joined #salt
07:17 Lue_4911 joined #salt
07:22 linjan_ joined #salt
07:33 m0hit joined #salt
07:36 networkpadawan joined #salt
07:42 anuvrat joined #salt
07:43 josephholsten joined #salt
07:44 Destro joined #salt
07:44 Destro Any other web gui than Halite, which might be nice to have a look into ?
07:55 Destro joined #salt
07:56 backjlack joined #salt
07:57 abishekk92 joined #salt
07:57 felixhummel joined #salt
07:58 [ilin] joined #salt
08:07 Furao destro: halite is just web shell over /usr/bin/salt
08:07 Furao salt-ui too
08:07 Furao if you find one, let me know
08:07 Furao I'll stop develop my own
08:07 Destro :-( and here me think that I was going for gold :p
08:08 Destro Thanks, I find the command line easier however thought of giving a web gui a go
08:09 dmwuw joined #salt
08:11 carmony joined #salt
08:12 [ilin] joined #salt
08:16 s0undt3ch joined #salt
08:26 dmwuw joined #salt
08:51 jcsp joined #salt
08:57 _mzito___ joined #salt
09:00 josephholsten joined #salt
09:01 Destro left #salt
09:05 krissaxton joined #salt
09:05 jcsp joined #salt
09:12 pentabular joined #salt
09:13 austin_laptop joined #salt
09:16 ronc joined #salt
09:17 ronc joined #salt
09:20 austin_laptop joined #salt
09:28 bhosmer joined #salt
09:39 lemao joined #salt
09:45 mikko__ joined #salt
09:47 mikko__ has anyone encountered "ZMQError: Operation cannot be accomplished in current state" on a minion in release 0.15.3? I have trouble executing commands on the master and i only get timeouts and all 5 salt-master processes ramp up to 20% cpu each engulfing all cpu.
09:47 mikko__ The master does that for a while and you can't execute any commands without getting errors. When it's not using up 100% cpu i can execute a couple of commands without problems then i get the same problem again
09:52 giantlock joined #salt
09:53 ronc joined #salt
09:55 ronc_ joined #salt
09:55 jcsp joined #salt
10:06 packeteer what version of zmq?
10:07 packeteer if pre v3, you really need to upgrade
10:08 mikko__ packeteer: ZMQ: 3.2.3
10:08 krissaxton joined #salt
10:08 jcsp joined #salt
10:09 mikko__ packeteer: i'm not sure if this is a firewall issue tho.
10:09 packeteer ok, sorry I don't know then
10:09 mikko__ i can still see that the minion is returning the job
10:09 mikko__ really odd
10:15 ronc joined #salt
10:24 Nexpro joined #salt
10:28 packeteer hmm, salt-minion is failing on centos6.4
10:29 packeteer salt is v17.1 from epel
10:33 packeteer hmm, not failing, just throws python erros on minion restart
10:34 mikko__ packeteer: what errors
10:34 Furao salt.client.LocalClient fail in 0.17.1
10:34 Furao yay no fun
10:37 alexheneveld joined #salt
10:40 packeteer with salt-ssh, do we have to put all hosts we're going to talk to into the roster file?
10:42 packeteer ie. must targets exist in roster file?
10:53 jcsp joined #salt
11:00 phab3k joined #salt
11:09 diegows_ joined #salt
11:21 blee joined #salt
11:30 jcsp joined #salt
11:37 srage joined #salt
11:38 austin987 joined #salt
11:38 srage joined #salt
11:42 zooz joined #salt
11:43 rgbkrk joined #salt
11:45 srage_ joined #salt
11:47 jslatts joined #salt
11:52 xl1 left #salt
12:05 ipmb joined #salt
12:12 matanya joined #salt
12:16 LucasCozy joined #salt
12:21 alexheneveld joined #salt
12:23 ronc joined #salt
12:24 austin987 joined #salt
12:26 xinkeT joined #salt
12:38 rgbkrk joined #salt
12:40 krissaxton joined #salt
12:41 anteaya joined #salt
12:42 whiskybar joined #salt
12:51 danielbachhuber joined #salt
12:55 linjan_ joined #salt
12:56 lemao joined #salt
12:58 brianhicks joined #salt
12:59 krissaxton joined #salt
13:01 jcsp joined #salt
13:03 fishpen0 joined #salt
13:06 Gifflen joined #salt
13:08 oz_akan_ joined #salt
13:08 groovehunter hi there. How to set an environment var for cmdrun , version is 0.15.1 ??
13:12 mwillhite joined #salt
13:13 racooper joined #salt
13:13 jslatts joined #salt
13:15 Khollowa_ joined #salt
13:17 blee joined #salt
13:20 mpanetta joined #salt
13:23 LucasCozy joined #salt
13:24 6JTAAJI8V joined #salt
13:26 alexheneveld joined #salt
13:26 BrendanGilmore joined #salt
13:30 matanya joined #salt
13:30 tumm joined #salt
13:30 juicer2 joined #salt
13:32 colinbits joined #salt
13:38 Brew joined #salt
13:40 Brew joined #salt
13:43 rgbkrk joined #salt
13:44 groovehunter test
13:46 tyler-baker joined #salt
13:46 tyler-baker joined #salt
13:50 avienu joined #salt
13:52 samsalt joined #salt
13:52 m_george left #salt
13:53 mannyt joined #salt
13:54 ipmb joined #salt
13:54 DredTiger joined #salt
13:56 micah_chatt joined #salt
14:03 lemao joined #salt
14:09 derelm joined #salt
14:10 aberant joined #salt
14:10 alexheneveld joined #salt
14:12 derelm joined #salt
14:13 mgw joined #salt
14:13 jcsp joined #salt
14:13 penguin_dan joined #salt
14:14 c0bra joined #salt
14:26 mwillhite joined #salt
14:28 cachedout joined #salt
14:28 opapo joined #salt
14:29 kermit joined #salt
14:29 tyler-baker joined #salt
14:30 Linz joined #salt
14:31 abe_music joined #salt
14:35 cammm joined #salt
14:35 alunduil joined #salt
14:37 roddergreg joined #salt
14:37 amahon joined #salt
14:37 roddergreg has anyone else upgraded to salt 0.17.1 using the epel testing packages. Having issues with syndic now?
14:37 Furao other stuff are broken
14:37 Furao reverted to 0.16.4, again
14:38 roddergreg ok thanks! I'll do the same. its a bummer.
14:40 cnelsonsic joined #salt
14:43 mgw joined #salt
14:47 mua joined #salt
14:47 alexheneveld joined #salt
14:52 eliasp what's the best way to debug pkg.install on Win minions? I have 5 packages of which only 1 (7zip) works (because I copied it straight from the salt-winrepo from GH :) )…
14:53 cshuman joined #salt
14:53 jcsp joined #salt
14:53 eliasp my minion log file on Win keeps being zero-sized… I might better investigate this first ;)
15:01 jslatts joined #salt
15:04 mapu joined #salt
15:06 UtahDave joined #salt
15:06 scott_w joined #salt
15:07 matanya_ joined #salt
15:09 lineman60 joined #salt
15:09 utahcon Just convinced the boss to start using Salt, immediately getting red tape from the executives about SOX compliance. Anyone have any experience in this field?
15:11 utahcon as far as I can tell, Salt is as far removed from the topic of SOX as possbile.
15:11 matanya_ joined #salt
15:12 roddergreg I know two types. Red abd White :-)
15:12 jcockhren utahcon: if anything salt can be leveraged to help achieve SOX
15:12 jalbretsen joined #salt
15:13 kula every time i've had to deal with compliance issues (FERPA, HIPPA, etc.) the only important people are your council, because its their interpretation of the compliance rules that count (because they're going to be the ones arguing in court if there's an issue....)
15:13 utahcon kula: thanks, that makes sense. I just happen to work for a bunch of lawyers, this should be fun.
15:14 kula "It's lawyers all the way down, man."
15:14 rgbkrk joined #salt
15:14 * racooper looks around for the bottle of holy water...
15:14 kula It's amusing you think that holy water will repel laywers.
15:14 kula :-D
15:14 UtahDave utahcon: So there's a few things here with SOX
15:14 honestly_ joined #salt
15:15 utahcon UtahDave: I am listening
15:15 UtahDave first, kula is right "it's lawysers all the way down."  :)
15:15 racooper well, it's the first line of defense for the unholy :P (my sister is a lawyer, and we make jokes about it all the time)
15:15 jbunting joined #salt
15:16 UtahDave utahcon: So I don't know every detail about SOX, but our clients who have discussed their SOX compliance with us usually focus on tracking and auditing who did what to the infrastructure
15:16 UtahDave So they disallow people logging in as root to use Salt
15:16 cachedout joined #salt
15:17 UtahDave They'll usually  use external authentication such as PAM, ldap or AD authentication
15:17 utahcon ok
15:17 matanya joined #salt
15:17 UtahDave Then they'll use an external job cache to keep a long term history of who did what.
15:17 utahcon ok
15:18 UtahDave That seems to satisfy most of their worry. I'm by no means a SOX compliance expert or a legal advisor, but that's what I've heard from customers.
15:18 nmistry joined #salt
15:18 amahon joined #salt
15:18 utahcon That is a help.
15:18 utahcon Thanks UtahDave
15:18 forrest joined #salt
15:19 UtahDave And we have customers in quite a few fields that have lots of compliance regulations.  Government, Financial trading, etc
15:20 fxhp joined #salt
15:20 eliasp does anyone have an idea, why the logfile of my minion (0.17.1, amd64, Win7 Prof) remains empty? I've tried /var/log/salt/minion and C:\salt\var\log\salt\minion (log_level: debug) but both didn't work… starting the minion on an admin CMD with --log-level=debug shows the output on stdout just fine…
15:24 all_the_fisch joined #salt
15:24 mnemonikk joined #salt
15:26 brimpa joined #salt
15:26 redondos joined #salt
15:26 mgw joined #salt
15:28 UtahDave You're welcome, utahcon!  Please let me know if you have any other questions about that.  I can put you in contact with some other people that might be able to help if there's other issues with that
15:29 UtahDave eliasp: does c:\salt\var\log\salt\minion exist already?
15:29 eliasp UtahDave: yes, but it keeps being zero-sized…
15:30 brimpa joined #salt
15:30 eliasp UtahDave: ah, nvm…
15:30 eliasp UtahDave: seems to work now… no idea what happened ;)
15:30 eliasp UtahDave: sorry for the noise
15:31 UtahDave no problem, eliasp. Let me know if this resurfaces again.
15:31 eliasp now I just need to find a decent 'tail' for windows… "… is in use by another application and cannot be accessed" … oh Windows, you're so funny!
15:32 eliasp UtahDave: well, I think the default config for windows minions should use C:\salt\var\log\salt\minion instead of /var/log/salt/minion
15:32 eliasp UtahDave: I had to change this manually
15:33 ctdawe joined #salt
15:33 eliasp also interesting: starting salt-minion manually on an admin CMD… it doesn't seem to catch Ctrl+C at all
15:36 berto- joined #salt
15:36 UtahDave eliasp: really?  It should default to c:\salt\var\log\salt\minion
15:36 UtahDave eliasp: yeah, catching Ctrl+C is iffy.   Ctrl+Break works better on windows.
15:37 UtahDave eliasp: is your root_dir  config option set to   c:\salt   ?
15:38 eliasp UtahDave: I'll verify this, but I'm pretty sure by default there was no value set, so the default /var/log/salt/minion should be used… have to check the code how the defaults are determined, whether they handle OS specific things like this…
15:38 eliasp UtahDave: but after all, I had to change the path to make it work…
15:39 eliasp UtahDave: but unless I have verified this, feel free to ignore me … gonna do that now
15:39 eliasp UtahDave: I used the default config and only specified the master host + minion name in the installer…
15:40 cjohnston_work joined #salt
15:41 bhosmer_ joined #salt
15:41 smccarthy joined #salt
15:41 derelm_ joined #salt
15:42 UtahDave eliasp: You did that correctly. You should only have to set the master ip/hostname and the minion's ID if you want.
15:42 jcsp joined #salt
15:46 dottedmag joined #salt
15:47 teskew joined #salt
15:48 eliasp UtahDave: the salt-minion service isn't started after the installation… starting it now manually…
15:48 CheKoLyN joined #salt
15:48 fusionx86 joined #salt
15:48 Khollowa_ joined #salt
15:48 eliasp UtahDave: log file remains empty… changing now the config…
15:49 druonysus joined #salt
15:49 druonysus joined #salt
15:50 cshuman joined #salt
15:50 cshuman joined #salt
15:51 balltongu joined #salt
15:53 backjlack joined #salt
15:53 avienu joined #salt
15:53 druonysus joined #salt
15:55 eliasp weird… even with manually setting "log_file: c:\salt\var\log\salt\minion" and "log_level: debug" I get nothing in the logfile now…
15:55 KennethWilke joined #salt
15:55 eliasp probably another Windows weirdness I don't want to know about… ;)
15:56 cshuman joined #salt
15:56 eliasp I should focus on getting my win_repo working and then concentrate on my Linux-boxes… debugging on Windows is just a nightmare and really clumsy
15:59 eliasp weird… the logfile is only written to when starting the minion manually on an admin CMD, but not when starting it as a service…
16:01 mgw joined #salt
16:01 hazzadous joined #salt
16:02 eliasp ok, at least found now the reason why my win_repo packages don't work… only the 'installer' file is transfered to the minion, but not the directory containing it which provides additional files required by the MSI…
16:03 mwillhite joined #salt
16:03 troyready joined #salt
16:04 hazzadous I'm having difficulty with setting limit nofile for a service what is salt managed.  It appears that despite assigning all users a non-default file limit, cmd.run 'ulimit -Hn' still registers the default.  When suing to root I get the correct limit, I assumed as salt ran as root that the result of cmd.run would also give roots limits.  Anyone know what I could be doing wrong?
16:04 UtahDave ah, yeah.  that's a limitation of the current implementation
16:05 hazzadous which*
16:05 hazzadous that*
16:05 UtahDave eliasp: the workaround for that right now is to use a file.recurse to drop the directory needed by the msi into the correct location first
16:06 micah_chatt I'm trying to write a custom grain, can I access the other existing grains?
16:06 eliasp UtahDave: ok, are there any plans for a "proper" solution (e.g. data_dir: salt://win/repo/microsoft/office/2013)? couldn't find a related issue…
16:06 nearmiss joined #salt
16:07 UtahDave eliasp: plans in my head.  :)  Would you mind opening an issue on that to make sure I remember?
16:07 eliasp UtahDave: sure, I'll do
16:07 UtahDave thanks!
16:07 UtahDave hazzadous: which OS are you on?
16:07 micah_chatt when I try accessing __grains__ in my custom grain, all I get is an empty dict
16:07 hazzadous ubuntu 12.10 @UtahDave
16:08 hazzadous salt 0.16.0
16:08 eliasp UtahDave: there's anyways a pile of other win_repo improvements I had in my mind… expect another 2-3 issues in the next minutes ;)
16:09 UtahDave micah_chatt: Yeah, because __grains__ doesn't exist yet at that point
16:09 UtahDave eliasp: VERY much appreciated!
16:10 micah_chatt UtahDave: The reason i'm asking is I want 2 different custom grains to be different keys in a dict under the same grain name
16:10 micah_chatt but instead of merging grains, it overwrites
16:10 mr_chris Having a difficulty getting the mysql states and module to work on a couple of centos 5 boxes. Salt 16.4, the MySQL-python module is installed, the user is properly configured in /etc/salt/minion, but it I keep running into "mysql_user.present found in sls mysql.users is unavailable". Same goes for every other mysql related state or module operation.
16:10 VSpike If you use "make_master: True" with salt-cloud, does it automatically configure the other machines to use that as the master? How do you point your minions at the new master as part of the deployment?
16:10 micah_chatt UtahDave: worthy of an issue?
16:11 bitz joined #salt
16:11 mr_chris Debug gives nothing useful. It just gives ERROR no changes made
16:11 redondos joined #salt
16:11 mr_chris Could it be an issue with having a slightly older version of the python mysql module?
16:12 KyleG joined #salt
16:12 KyleG joined #salt
16:13 taylorgumgum joined #salt
16:14 UtahDave micah_chatt: I think it's probably working as expected right now.  A custom grain returns a dict that gets added to __grains__    The second custom grain to run will probably overwrite the first
16:15 micah_chatt right, but I would like them to be merged. Kind of a namespace rather than just having a bunch of grain names
16:16 jdenning joined #salt
16:16 hazzadous sorry, 12.04 @UtahDave
16:16 micah_chatt UtahDave: the specific case being that I have an grain['ec2']
16:16 jcsp1 joined #salt
16:16 micah_chatt with different instance attributes
16:16 jcsp joined #salt
16:16 UtahDave micah_chatt: can you have one custom grain file that returns both values in the same return?  Otherwise it's going to be a fair amount of work to rework the grains creation to allow for merging and determing when merging is the correct behavior
16:16 micah_chatt grain['ec2']['instance-id']
16:16 micah_chatt yea, thats what I've resorted to
16:17 micah_chatt I think it will be ok that way
16:17 forrest mr_chris, have you installed the mysql-client, and mysql-python packages on the associated minion?
16:18 carmony UtahDave: Hrm, so my salt-cloud deployments are failing to install the minion and connect to the master
16:18 forrest carmony, why do you always have to break it? :D
16:18 mr_chris forrest, The client is installed. The MySQL-Python package is installed.
16:18 forrest on the minion you're installing onto?
16:18 carmony I help make it a better product! :D
16:18 forrest or is this a masterless minion config?
16:18 mr_chris forrest, Figured it out.
16:19 forrest mr_chris, what was it?
16:19 mr_chris It also needed python26-mysqldb
16:19 forrest ahh ok
16:19 mr_chris Which I don't remember my centos 6 boxes needing.
16:20 UtahDave VSpike: This isn't documented very well, but look here:  http://salt-cloud.readthedocs.org/en/latest/topics/releases/0.8.9.html#mapping-independent-hierarchies
16:20 forrest mr_chris, it looks like someone posted on SO about this a while back
16:20 forrest http://stackoverflow.com/questions/16987594/how-to-install-python26-mysqldb-on-a-rhel5
16:21 forrest I don't know what is in the python26-mysqldb RPM though :\
16:21 mr_chris I didn't have any trouble installing it.
16:21 mr_chris It's in the repos.
16:21 UtahDave micah_chatt: feel free to open an issue on that, but you may have to make a good case to change the behavior for Tom to want to change it
16:21 mr_chris forrest, Thanks for the help.
16:21 forrest yea that guy is using RHEL, so it probably wasn't in his default repo, just saying it looks like it's been that way for a while as a point of reference
16:21 forrest yea np
16:22 UtahDave hazzadous: I'm not sure what's going on there. If your minion is indeed running as root, then I'm not sure why it's not getting the correct values.  Would you mind opening an issue on that?
16:22 hazzadous UtahDave: sure
16:23 UtahDave mr_chris: looks like your minion is not loading the mysql_user module for some reason.
16:23 forrest he already fixed it UtahDave
16:24 mr_chris UtahDave, Got it worked out. For some reason python26-mysqldb AND MySQL-Python is required. On centos6, only MySQL-Python is required.
16:24 forrest mr_chris, what is the rpm -qi data on python26-mysqldb?
16:24 mr_chris UtahDave, Thanks anyways.
16:24 backjlack joined #salt
16:24 UtahDave ah, interesting.  Glad you figured that out, mr_chris!  I'll make a mental note on that.  In fact, we should probably add that to the docs.
16:24 UtahDave mr_chris: Would you mind adding a quick note about that to the docs?
16:25 UtahDave carmony: have you looked at this? http://salt-cloud.readthedocs.org/en/latest/topics/releases/0.8.9.html#mapping-independent-hierarchies
16:25 mr_chris UtahDave, Would love to. How?
16:25 UtahDave all the docs are either in the docstrings of the modules themselves, or in the docs directory of the repo.
16:26 alexheneveld joined #salt
16:26 mr_chris forrest, http://pastie.org/8421748
16:26 forrest mr_chris, thanks!
16:26 mr_chris UtahDave, OK. I'll see if I can figure it out.
16:26 forrest wait, so what does MySQL-Python provide then?
16:27 cbloss with 0.17.1, one of my minion hostnames is now "li659-43.members.linode.com" instead of what is being set in /etc/hostname. According to issue #7558, it looks like socket.getfqdn() is now prioritized over /etc/hostname. anyone know the best way to manage minion's names across several different cloud providers? Am I setting something up incorrectly so my fqdn isn't being assessed correctly by salt?
16:27 Lue_4911 joined #salt
16:27 UtahDave mr_chris: here: https://github.com/saltstack/salt/blob/develop/salt/modules/mysql.py
16:28 UtahDave basepi: what was the final decision on the minion hostname priority order that cbloss is seeing?
16:28 mr_chris UtahDave, Ah! Easy enough. I've had a few things of mine added there already!
16:28 UtahDave cool, thanks, mr_chris!
16:29 pipps joined #salt
16:29 jefimenko joined #salt
16:31 xmltok joined #salt
16:31 mr_chris UtahDave, Any formatting conventions that I should be aware of?
16:31 hazzadous UtahDave: just had a thought, I'm guessing that it's due to using the included upstart scripts in the deb
16:31 carmony UtahDave: not yet, but I was planning on it :) I wanted to get just 1 provider worker
16:32 carmony and I think I figured out my problem
16:32 hazzadous Although I wouldn't have thought spawned subprocesses should inherit the limits
16:32 mr_chris Would something like :note: work? Is there a reference for recognized tags that the doc system uses?
16:33 VSpike UtahDave: many thanks for that. I'm getting this when I try to test http://sprunge.us/AVGi (using latest git)
16:33 forrest if you wann do note do .. note::
16:33 VSpike UtahDave: actually, i'm using salt-cloud-git with the release version of salt. Is that perhaps a bad idea?
16:34 CheKoLyN joined #salt
16:36 jcockhren as a note. I just updated my minions to 0.17.1 without issue it seems
16:36 jcockhren Ubuntu 12.04
16:38 roddergreg joined #salt
16:39 anuvrat joined #salt
16:41 berto- joined #salt
16:43 mr_chris UtahDave, https://github.com/saltstack/salt/pull/8018
16:44 mr_chris Hopefully that's done right. If not, I'll continue the conversation in the pull request.
16:45 matanya joined #salt
16:46 jcsp joined #salt
16:46 dave_den hazzadous:  how are you setting the limits? if you're trying to do it trhough /etc/secutiry/limits, that is parsed for login shells from pam_limits.so. Upstart scripts will not use /etc/security/limits. you must specify your limits in the upstart service file.
16:48 hazzadous dave_den: have added a ticket to update the docs on this.  The reason for salt not picking up limits I can understand.  However, that the managed services are inheriting the settings is annoying me at the moment.
16:48 dave_den the only limit set from salt's upstart files is the nofile limit.
16:50 hazzadous dave_den: ah yes, I was using 0.16 before, which didn't set the limits in the upstart file
16:50 hazzadous looks like that's changed.
16:50 basepi UtahDave: in 0.17.1 it's back to socket.getfqdn() as the default
16:50 dave_den oh, i read your statement wrong. sorry.
16:50 UtahDave thanks, mr_chris!
16:50 basepi if that returns localhost, then it will fallback on /etc/hostname
16:51 basepi and then ip addresses after that, iirc
16:51 basepi also in 0.17.1 it caches minion IDs rather than guessing every time.
16:52 dave_den hazzadous: i bet if you try cmd.run with runas: root, you will see the limits you set in security/limits.
16:53 UtahDave VSpike: that should be fine.  Let me pass on your pastebin error to one of the salt-cloud devs
16:57 carmony UtahDave: ok, so it can create the server, but it is failing to provision it with salt
16:57 cbloss basepi: my linode minons are now showing up as "li659-43.members.linode.com" etc. Is there a proper way to set minion ids across several different cloud providers?
16:57 carmony I get this error:
16:58 carmony https://gist.github.com/JustinCarmony/7104215
16:59 basepi cbloss: i mean, the most consistent way will always be to manually set your ID in the minion config
16:59 basepi then it will never change unexpectedly
16:59 basepi cbloss: that's the minion ID i would expect to see on all versions except 0.17.0 when the regression happened
16:59 dave_den cbloss: if you want to be sure your minion id is always set to what you want, I would recommend setting it in the minion config.d directory explicitly. E.g. echo "my_minion_id" > /etc/salt/minion.d/id.conf
17:00 toofer joined #salt
17:00 carmony is there a way to tell it to re-provision my cloud server? like try to re-install salt on it?
17:00 anteaya joined #salt
17:00 cbloss i'd have to do that prior to joining a minion to the master? I control salt's minion/master config using the salt-formula
17:00 UtahDave carmony: are you setting the "script" arg in your profile?
17:01 carmony ummmm, I don't think so?
17:01 UtahDave carmony: can you pastebin your /etc/cloud.profiles?  (sanitized, if needed)
17:01 jkleckner joined #salt
17:02 dave_den cbloss: yes, you would want to set your minion id before trying to connect to the master for the first time, since the minion will send the RSA key as your currently set minion ID (which is soc.getfqdn() as default in 0.17.1 as basepi said)
17:03 carmony UtahDave: https://github.com/JustinCarmony/vagrant-cloud/tree/master/saltstack/salt/master/files/etc
17:03 cbloss thanks
17:03 clintberry joined #salt
17:03 UtahDave carmony: yeah, yank out "script: Ubuntu"  that's deprecated and doesn't work very well.
17:04 carmony k, afk for a bit, doing an interview
17:05 clintberry Hey UtahDave. Every time I google salt stuff, it returns the docs I want in the results but adds this to each title " - What is Salt Stack?". Kind of makes the results look weird. Just thought I would let you know :-)
17:05 UtahDave clintberry: url?
17:06 clintberry http://imgur.com/UKTDKCb
17:06 clintberry https://www.google.com/search?q=saltstack+kvm&oq=saltstack+kvm&aqs=chrome..69i57j69i60l3.3356j0j4&sourceid=chrome&espv=210&es_sm=91&ie=UTF-8
17:06 eliasp only applies to docs.saltstack.com: https://www.google.com/search?q=saltstack+kvm&oq=saltstack+kvm&aqs=chrome..69i57j69i60l3.3356j0j4&sourceid=chrome&espv=210&es_sm=91&ie=UTF-8#es_sm=91&espv=210&q=saltstack+kvm+inurl:docs.saltstack.com&safe=off
17:07 eliasp that could be probably solved by a custom sitemap which provides the site titles
17:07 clintberry kvm built in? awesome
17:07 UtahDave Ah, I think that's because "What is Salt Stack" is the first <h1> on the main page.
17:08 hazzadous dave_den: didn't try runas but you very well could be right.  I guess it just sus' to root for cmd.run.
17:08 clintberry UtahDave: weird that it would put it on every page, though
17:09 hazzadous dave_den: I guess the question then is why the spawned service inherits
17:09 hazzadous anyway, will investigate
17:10 dave_den hazzadous:
17:10 joehh joined #salt
17:11 dave_den hazzadous: that's just the way limits, login shells, init services work
17:11 eliasp UtahDave: regarding your suggested workaround using file.recurse to sync a win_repo package directory to the minion… can this be done inline the package state file or do I have to do this outside of it (sync files, then depend the pkg installation on it…)?
17:11 hazzadous dave_den: ok
17:11 UtahDave yes, the second scenario.
17:11 eliasp ok
17:11 UtahDave eliasp:
17:12 UtahDave eliasp: Yeah, it's not optimal, for sure.
17:12 eliasp UtahDave: sure, it's just a workaround
17:12 xmltok anyone out there using salt in a large enterprise environment and care to chat with me about it?
17:12 eliasp UtahDave: just wanted to make sure I don't miss the easiest way for the workaround
17:12 forrest tons of people are xmltok! Just ask your questions in here.
17:13 nmistry joined #salt
17:13 ctdawe joined #salt
17:13 xmltok i got a lot! i want to know how they monitor the automated runs for errors, how are you managing versioning/releasing/workflow for formulas, how large is the deployment, how many people working on formulas, any best practices, if you are pci compliant -- how do you keep your secure information separate so that engineering can also work on salt formulas
17:14 nmistry joined #salt
17:15 pentabular joined #salt
17:15 eliasp xmltok: you can use different environments for versioning/releasing… use the GitFS fileserver backend in which each git branch represents an environment… http://docs.saltstack.com/topics/tutorials/gitfs.html
17:16 fusionx86 hey guys, is it possible to have two s3.keyid and s3.key in the same /etc/salt/master file? I need to have my master access s3 buckets from two different aws accounts.
17:16 cshuman joined #salt
17:16 eliasp secure information should be kept in a "pillar": http://docs.saltstack.com/topics/pillar/
17:16 forrest and the pillar itself could be in a different git repo that only certain people have access to.
17:16 eliasp xmltok: ^^
17:16 xmltok i was looking at that -- my understanding was that you would manage releases through merges, which is OK but potentially rough if the environments get out of sync a little, you dont really have a good way of testing the merge
17:17 forrest xmltok, sure you do
17:17 forrest use the master branch as your 'good' copy, then in your test file, just pull from a 'dev' branch to run in your test env.
17:17 forrest or you could use multiple repos, and do a 'pull, check-in, push to new repo' style
17:17 xmltok oh right right, i am thinking of a single repo for a few different formulas, you would have a single repo for each formula
17:18 xmltok it requires a master restart everytime a new formula repo is added right? thats kind of a bummer but not a big deal i guess
17:18 eliasp xmltok: for repos: yes, for branches: no
17:18 xmltok a branch name == environment right
17:18 clintberry dude, I should just hang out here always. gitfs is awesome!
17:18 eliasp xmltok: exactly
17:19 eliasp clintberry: yes, it's Puppet's librarian/r10k just builtin… one of the many reasons why I'm making the switch from Puppet for Salt ;)
17:19 xmltok can it be made to abort immediately if a part of a formula fails? i admit i have not tried a very complicated formula yet, but it appeared to keep on truckin when i asked it to install an invalid package
17:20 dave_den xmltok: overstate does, highstates don't
17:20 xmltok does salty vagrant pull in formula dependencies, or can you define a dependency in a formula? berkshelf was awesome for that, i could work on a cookbook that depended on other cookbooks very easily
17:20 xmltok dave_den: aha, i need to read more on the difference between overstate nad highstate then, i was only looking at highstate to replicate how i am used to chef working
17:21 amckinley joined #salt
17:21 xmltok basically the continuous exectution to ensure my node is up to date on a system level, i have not dug into application releases with salt yet
17:22 dave_den that's a recurring question on the salt-users mailing list, app release cycles and approach
17:22 UtahDave xmltok: if you set    failhard: True in your master config, then your highstate will completely stop if there's an error
17:22 xmltok excellent
17:22 dave_den but only on one minion, right UtahDave?
17:23 dave_den if he's deploying an app across many minions and different server roles that depend on successful completion ins sequence, overstate is needed in that case
17:24 xmltok if salt can orchestrate states across multiple minions then i totally missed that and its a huge benefit over using chef/puppet
17:24 dave_den xmltok: yes, i gather overstate is what you should look at.
17:24 foxx joined #salt
17:25 forrest agreed dave_den
17:25 dave_den http://docs.saltstack.com/ref/states/overstate.html
17:25 xmltok already there.. :)
17:25 dave_den good, get reading! ;)
17:25 dave_den you'll find lots more you like, i am sure
17:25 xmltok for some reason i assumed this was a general configuration that was applied individually, not something that could be applied to all nodes at once, if you know what i mean
17:25 networkpadawan joined #salt
17:25 xmltok however, the first sentence on that page makes it totally clear
17:26 forrest yea, overstate is awesome
17:26 forrest except when you link someone highstate instead of overstate like I do sometimes..
17:26 xmltok ok, so overstate is a state across minions, and highstate is a state on a single minion
17:26 dave_den salt is very flexible and able to operate as a single, masterless setup, to entrprise with sequence orchestration and rollout, event-driven real-time infrastructure, etc.
17:27 cshuman joined #salt
17:27 dave_den that's what i like about it is the real-time infrastructure capability
17:28 forrest and don't forget the reactor system dave_den, awww yea
17:29 cjh_ i haven't used reactor yet
17:29 xmltok i saw someone complaining about minions getting disconnected from the master and it being hard to know when it happens ,or something like that. does that ring any bells? is it still an issue?
17:29 dave_den xmltok: how may minions are you estimating to have?
17:29 cjh_ looks cool though
17:30 xmltok dave_den: 30k across 3 colos
17:30 cjh_ xmltok: awesome :D
17:30 dave_den xmltok: the minion reconnects were usualy a zmq 2.x issue. it is not a problem when you are using zmq 3.x which all linux based minions will, afaik.
17:30 xmltok excellent
17:30 forrest damn you dave_den, typing faster than I can
17:31 cjh_ how do i get salt to timeout on starting a service if the init script hangs?  is that something salt should handle or should i modify the init script to be smarter?
17:31 xmltok ok, i think i have everything i need right now, you guys have been a huge help :)
17:31 dave_den xmltok: you will need to do some perf tuning for that many minions. not sure a single master can handle 30k minions. you can do multi-master setup or synic, tho.
17:31 xmltok thanks a lot!
17:31 dave_den forrest: ;)
17:31 forrest :D
17:31 xmltok yeah i would do multi-master, probably 2-3 in each colo for redundancy and load distribution
17:32 xmltok the key syncing seemed like a pretty easy problem and i could probably use salt to reconfigure masters for additional gitfs repos
17:32 dave_den xmltok: absolutely. please post whatever you can about your large deployment if you are able. i like to see large deployments and how they are handled.
17:32 dave_den and the salt-user mailing list is something you might want to join
17:32 xmltok oh i saw something about salt mine (i think) where the minions can send data up to the master, so i can make minions confiugre themselves based on other minions, does that work in multimaster mode?
17:32 forrest It should
17:33 forrest Because you can set up multiple masters in the config file for the minion, so that info should be pushed to all of them.
17:33 xmltok yeah, we will be starting a blog and the CMS selection will be posted
17:33 xmltok forrest: got it
17:33 dave_den i havent' tested mine across multi master myself, but from what i know about mine, you may have inconsistencies or stale data depending on how you use it.
17:33 forrest I haven't tested it xmltok
17:33 forrest dave_den, I think you could work around that though
17:33 forrest would take a bit of doing though
17:34 xmltok it may not be needed, i have an inventory system that i can pull all of the information i need from, i would probably have the minions publish to the inventory system and pull from there if it comes down to it
17:34 dave_den forrest: the mine calls a single master calls, so until a minion makes a few mine.sends, the data may only be on one master
17:34 forrest That would be a question for th salt guys though, UtahDave, have you configured the salt mine to work with multiple masters? How does it handle that? Can you have issues with stale data?
17:34 forrest dave_den, ahh right right, I forgot about that.
17:34 forrest my apologies xmltok
17:35 xmltok that is fine, that design makes me feel like the multimaster mode is more reliable for my scale
17:35 dave_den xmltok: also, you can do custom grains for minions to store that non-sensitive info, then you can do use the peer system (or peer_run) to pull that info into states on the minion that needs it.
17:36 dave_den which is actually more real-time that the mine
17:36 dave_den by up to 59 seconds (by default), if that matters
17:36 dave_den because you know you want it super fresh  :)
17:37 UtahDave forrest: I think there's an open bug that the mine doesn't work with multi-master.
17:37 forrest UtahDave, yea https://github.com/saltstack/salt/issues/7697
17:38 pentabular joined #salt
17:38 ccase joined #salt
17:39 xmltok the overstate system works on multimaster, even if minions are connected to other masters?
17:39 UtahDave xmltok: yeah, that should work just fine
17:40 pentabular1 joined #salt
17:43 ajw0100 joined #salt
17:45 woebtz joined #salt
17:49 m_george joined #salt
17:49 berto- joined #salt
17:50 druonysuse joined #salt
17:50 druonysuse joined #salt
17:51 clintberry Salt Virt is ready for prime time, right? even though these docs don't exist yet? http://docs.saltstack.com/topics/tutorials/cloud_controller.html#using-salt-virt
17:51 bhosmer joined #salt
17:52 amckinley joined #salt
17:52 UtahDave clintberry: I know there are people using it in production.  I haven't had a chance to use it yet, myself
17:53 MatheusOl joined #salt
17:53 clintberry okay cool. You guys iterate so dang fast. how many devs do you have at SaltStack now?
17:54 clintberry what the, awesome! https://github.com/saltstack/salt-api
17:54 clintberry you guys are ridiculous! so much good stuff!
17:55 luminous anyone familiar with the postgres module? and postgres.owner_to in particular?i
17:55 Ryan_Lane joined #salt
17:56 ajw0100 joined #salt
17:56 luminous I've got a state using postgres.owner_to with module.wait, and the output is not so reassuring: https://dpaste.de/Hg1t
17:56 luminous when I review table/db permissions with psql, I see no changes
17:59 ninkotech joined #salt
18:01 anteaya joined #salt
18:03 UtahDave luminous: sorry, I'd like to help, but have almost no experience using postgres.  You might ask on the mailing list if nobody here right now can help.
18:03 luminous UtahDave: understandable. I think the module is broken
18:06 krissaxton joined #salt
18:08 Ryan_Lane joined #salt
18:08 giantlock joined #salt
18:08 ckao joined #salt
18:09 abe_music joined #salt
18:10 anteaya joined #salt
18:10 AdamSewell joined #salt
18:12 cjh_ salt: what returncode does my init script need to return to indicate to salt that the service needs to be restarted?
18:12 cjh_ salt seems to think my service is running but it's actually dead
18:12 pentabular joined #salt
18:13 cjh_ i'm guessing just some non zero return will do it
18:16 josephholsten joined #salt
18:16 anteaya_ joined #salt
18:17 UtahDave luminous: Would you mind opening an issue on that?
18:18 UtahDave cjh_: Yeah,I think you are correct
18:18 mgw joined #salt
18:18 luminous sure, anything special to note other than "I don't think this works"?
18:19 tyler-baker joined #salt
18:20 UtahDave luminous: :)  Well, include anything that will be helpful in tracking down the issue.  Any stacktraces you're seeing, the configs you're using, anything you feel is out of the ordinary.
18:21 * luminous nods
18:22 cjh_ UtahDave: i added a return code in the right place and it looks like salt figured out what to do.  i'm good
18:22 cjh_ it was returning zero even on failures
18:26 rgbkrk joined #salt
18:27 juicer2 joined #salt
18:30 luminous UtahDave: https://github.com/saltstack/salt/issues/8023
18:30 luminous hope that helps
18:31 UtahDave thanks, luminous!
18:31 luminous likewise :)
18:31 tumm joined #salt
18:33 Ryan_Lane joined #salt
18:36 jslatts joined #salt
18:38 mapu joined #salt
18:38 mapu left #salt
18:40 Kingdutch joined #salt
18:43 ajw0100 joined #salt
18:45 mapu joined #salt
18:46 colinbits joined #salt
18:51 amckinley joined #salt
18:51 abe_music joined #salt
18:52 eliasp UtahDave: I'm just playing with a proper directory structure for my states to make use of the win_repo workaround we talked about… but it seems the minion MN1221-C0001 doesn't pick up anything… is there anything fishy in this config: http://pastie.org/8422057
18:53 UtahDave eliasp: I think that looks fine.   You know, a few days ago I ran into a problem getting the winrepo runner to correctly compile the packages.  It had to do with the saltstack repo having some jinja, which hasn't been implemented in the runner yet
18:54 UtahDave I took out the jinja that I could find, but I'm not sure it's entirely working correctly.
18:54 UtahDave have you seen any other stacktraces?
18:55 tumm joined #salt
18:55 colinbits joined #salt
18:56 pentabular joined #salt
18:56 eliasp UtahDave: well, I'm wondering: do I have do to anything to make the minion pick up this states? having the minion MN1221-C001 in dev/top.sls should be enough, right? or do I have to trigger it somehow to update?
18:57 eliasp UtahDave: so until now there's no output in the minion's log at all… it reacts just fine to manual actions like test.ping
18:57 VSpike UtahDave: many thanks :)
18:57 UtahDave Oh, yeah. Salt doesn't do anything until you tel it to.   once you have the top.sls set up correctly you have to initiate it
18:57 UtahDave salt 'MN1221-C001' state.highstate
18:57 UtahDave VSpike: any time!
18:58 eliasp UtahDave: ah… still have to get some basic Salt concepts… thanks! ;)
18:58 UtahDave you're very welcome, eliasp!
19:00 eliasp UtahDave: hmm, only https://github.com/saltstack/salt/issues/7954 on the minion… is there some kind of state validation to validate syntax, structure etc of state files?
19:01 UtahDave Ah, yeah. I have to finish fixing that one.
19:02 VSpike UtahDave: looks like you have yourself a full-time job here :)
19:02 eliasp or to put it differently… is there a way to introspect/output the states to see whether they're picked up correctly by the master?
19:03 eliasp VSpike: working hard on keeping him busy ;)
19:03 UtahDave VSpike: he he. You have no idea.  :)  Plus, we have a bunch of enterprise customers that I'm supporting, too.
19:04 UtahDave eliasp: are you referring to the Windows states you're trying to put together?  Or your states in general
19:04 alunduil joined #salt
19:04 eliasp UtahDave: the states in general…
19:04 UtahDave try this:   salt '<minion name>' state.show_highstate
19:04 m_george left #salt
19:04 UtahDave salt '<minion name>' state.show_sls
19:05 eliasp UtahDave: awesome!
19:05 jslatts joined #salt
19:05 UtahDave there's several related function in the state module like that.  They're often very helpful
19:05 eliasp UtahDave: exactly what I was looking for…
19:05 m0hit joined #salt
19:06 derelm joined #salt
19:12 rgarcia_ joined #salt
19:12 elfixit joined #salt
19:14 eliasp UtahDave: well,… always runs into the UnpackError, so it looks like my win-clients have to wait for #7954 …
19:15 eliasp UtahDave: was already trying to have a look at #7954 on my own, but this involves unpacker.pyx/unpacker.cpp and I don't feel like touching that ;)
19:15 UtahDave eliasp: Yeah, I'm afraid that's going to be a blocker. I'm hoping to get a new Windows installer out tonight or tomorrow morning
19:15 aleszoulek joined #salt
19:15 eliasp UtahDave: oh, that fast… awesome!
19:15 UtahDave eliasp: I think it's actually easier than that.
19:15 UtahDave eliasp: I think the bbfreeze isn't catching all the deps it is supposed to.
19:16 eliasp UtahDave: ah, so it's not a code but a packaging issue…
19:17 UtahDave eliasp: I *believe* so.
19:17 eliasp believing is better than having no clue at all (at least when it comes to bugs, I tend to think differently when it comes to religion :) )
19:17 alexheneveld joined #salt
19:18 roddergreg has anyone had any luck with Solaris Minions?  The package available in the unstable repo of Open CSW is Salt 0.14.0.
19:19 roddergreg I'm running this version on some Solaris servers and I've installed via git on one of my development boxes and its running 0.17.0.  Are there any plans in the works to create a 0.17.1 version on Open CSW?
19:21 rgarcia_ joined #salt
19:22 rgarcia_ is there an easy way to debug hostname matching in top.sls? running a highstate w/ debug logs but not seeing any useful info
19:22 toofer left #salt
19:23 hjubal joined #salt
19:23 eliasp rgarcia_: you mean like: "show all minions which match the hostname compound"?
19:24 juicer2 joined #salt
19:24 alexheneveld joined #salt
19:24 UtahDave rgarcia_: try    salt '<minion name>' state.show_top
19:24 rgarcia_ eliasp: sort of--I'm just running salt-call --local
19:25 rgarcia_ ah yes
19:25 rgarcia_ thanks!
19:25 UtahDave roddergreg: The person who initially packaged Salt for Solaris hasn't been around for a while.
19:26 UtahDave roddergreg: do you use a lot of Solaris machines?
19:26 eliasp rgarcia_: well, a bit of a poor-man's approach, but what about "salt -C 'your-hostname-regex' test.ping"?
19:27 eliasp rgarcia_: I'm pretty sure there's a better way to do this, but that's what I can come up with using my (still) limited Salt knowledge ;)
19:30 abishekk92 joined #salt
19:30 nahamu I'd love to get the freezing process working on Unix.
19:30 rgarcia_ eliasp: I'm just doing salt-call --local right now w/ a local state tree
19:30 nahamu (I'm on SmartOS, but could also be useful for Solaris)
19:31 eliasp rgarcia_: ah, sorry… no idea then
19:31 UtahDave nahamu: It should be working. I do it for ubuntu and centos quite often
19:31 nahamu the SmartOS pkgsrc repos lag just far enough behind that it would be nice to be able to ship salt by a different mechanism.
19:31 rgarcia_ UtahDave: is there a way to see the hostname that salt-call --local uses to match?
19:31 nahamu UtahDave: in that case I'll try it again and perhaps bug you with questions?
19:31 UtahDave rgarcia_: salt-call --local grains.item id
19:31 UtahDave nahamu: sure!
19:31 nahamu thanks!
19:32 rgarcia_ UtahDave: amazing, thanks. It all makes sense now
19:33 roddergreg UtahDave - Hundreds
19:33 UtahDave roddergreg: Are you familiar with packaging for Open csw?
19:33 roddergreg it would be great to have a precompiled package available in OpenCSW to manage this
19:34 jslatts joined #salt
19:34 carmony UtahDave: I think the script issue was definitely the problem
19:34 roddergreg I've been looking at the packaging process using mgar
19:34 UtahDave carmony: also, digitalocean's api has been flaky and down all morning.
19:34 UtahDave completely unusable at the moment
19:35 roddergreg but if the groundwork has been started shouldnt the maintainer keep it current?
19:35 carmony UtahDave: ahh, yeah, well salt-cloud still failed to connect to the server
19:35 phab3k joined #salt
19:35 UtahDave roddergreg: I could probably get the original guy to help you get started.
19:35 carmony I think this issue was the server hadn't completely come up
19:35 carmony before salt-cloud tried to ssh into it to provision it
19:35 roddergreg ok that sounds great :-)
19:36 UtahDave roddergreg: Yeah, but it's kind of hard to force a volunteer to keep maintaining it when their job pulls them elsewhere.
19:36 UtahDave carmony: yeah, you're exactly right.  As soon as their api is back up I'll continue troubleshooting that
19:37 carmony UtahDave: sounds good. :) Maybe I'll try the AWS one right now :P
19:37 UtahDave cool.
19:38 JPmx joined #salt
19:39 jslatts joined #salt
19:40 berto- joined #salt
19:40 phab3k joined #salt
19:41 ipmb joined #salt
19:43 juicer2 joined #salt
19:44 eliasp UtahDave: I got rid of the unpack issue by clearing the minion cache (https://github.com/saltstack/salt/issues/7707#issuecomment-26009314)
19:44 eliasp UtahDave: still can't get my pkg installation working, but that's now due to some other issue… still have to figure this out
19:44 UtahDave eliasp: Ah, cool!
19:45 derelm joined #salt
19:46 UtahDave carmony: that worked on Rackspace just now
19:47 rodvdka joined #salt
19:48 juicer2 joined #salt
19:49 josephholsten joined #salt
19:50 throwanexception joined #salt
19:50 MTecknology joined #salt
19:51 [ilin] joined #salt
19:51 carmony joined #salt
19:51 [ilin] joined #salt
19:51 pmrowla joined #salt
19:55 srage joined #salt
19:57 rodvdka left #salt
19:57 colinbits joined #salt
20:01 kaptk2 joined #salt
20:03 krissaxton joined #salt
20:05 ldlework joined #salt
20:07 nahamu UtahDave: is this a problem when doing a bdist_esky: "Warning: don't know how to handle binary dependencies on this platform (sunos5)"
20:08 pentabular joined #salt
20:08 UtahDave nahamu: Ah, probably.
20:10 redondos joined #salt
20:10 redondos joined #salt
20:11 nahamu is that probably esky or bbfreeze's fault?
20:11 UtahDave nahamu: I'm leaning toward bbfreeze, but i'm not sure. Looking at bbfreeze's docs and it looks like it support Linux and Windows.
20:12 mapu joined #salt
20:12 gmcwhistler left #salt
20:12 UtahDave although it does say that it works on Windows and UNIX-like operating systems.
20:13 UtahDave nahamu: what's the output of      python setup.py bdist_bbfreeze
20:13 nahamu I'll pastebin, one moment
20:15 nahamu http://paste.ec/?e7654a310db60a33#MtqRTgv/OW5vy78p87LsDCTEB3kbCCM6INih7gp7WqI=
20:15 zooz joined #salt
20:16 UtahDave nahamu:  try this:       python setup.py bdist_bbfreeze
20:21 nliadm iirc, I had a problem with it last time I tried
20:21 nliadm for 0.17.0
20:21 alexheneveld joined #salt
20:24 nahamu I may have broken my build environment... http://paste.ec/?30be60ba3226a0c0#tUaTBrOyAngcB1byJSVhN9WHf5A+RDXKc1SGKYrh5ao=
20:24 nahamu nliadm: should I try building e.g. 0.16 just to help isolate the problem?
20:24 pentabular1 joined #salt
20:24 UtahDave you might try uninstalling esky
20:25 UtahDave esky messes around with bbfreeze
20:25 nahamu I can make a fresh virtualenv.
20:26 nliadm I'm actually about to mess with this, if I can remember where I had my build environment...
20:27 Snow-Man joined #salt
20:28 nahamu uhh... error: invalid command 'bdist_bbfreeze'
20:28 Snow-Man I've got a 'traditional' setup w/ salt/pillar where I have definitions for all my users in pillar/users.sls, which is included in pillar/top.sls through a 'base: "*" - users' construct.
20:28 UtahDave really, I don't think esky is necesarry unless you want the remote update capability of esky
20:29 Snow-Man And then in my salt states, I specify which users go with which systems.
20:29 Snow-Man I'd like to move that information into pillar instead though, but I don't want to break up my user dictionary into multiple files.
20:29 Snow-Man Is it possible to get pillar to essentially render subsets of a given set?
20:29 nliadm I need esky because it's the only way to get a new enough zmq on the machines
20:29 nahamu I was kind of hoping for the remote update capability... :)
20:30 UtahDave ah, ok.
20:30 nahamu do the esky builds need a system python or do they bundle that too?
20:30 nmistry joined #salt
20:31 UtahDave nahamu: no, everything it needs is bundled
20:31 * Snow-Man has a feeling he's stuck breaking up his "users.sls" file into multiple files.
20:31 UtahDave Snow-Man: you could use some jinja in your users.sls that makes it so it only returns the proper uses depending on the your criteria
20:31 Snow-Man And then trying to work out the perfect "groupings", which will be next to impossible. :/
20:32 Snow-Man UtahDave: yeah, that's rather ugly also. :/  Lots of 'if this, if that' madness
20:32 UtahDave yeah.
20:32 Snow-Man I really just want to have a complete dictionary in users.sls and then a simple mapping, similar to what I can do with the current pillar/salt split
20:32 UtahDave Snow-Man: you could create an external pillar. an external pillar allows you to do almost anything you want with the pillar data
20:33 Snow-Man UtahDave: I had heard/read online..  Are there any examples you could point me to?
20:33 Snow-Man UtahDave: I'd love to use pillar to simply connect to a PG database...
20:35 Snow-Man Where I could then have a normalized schema, with the user data in one table, the mapping in another, and then a join which brnigs them together for the pillar config of a given box..
20:36 UtahDave Snow-Man: yeah, you can totally do something like that.  just a sec
20:36 nliadm UtahDave: hit it again, I get 'ImportError: No module named site' with an esky build
20:36 UtahDave nliadm: add  'site' to the setup.py in the freezer includes list
20:36 nliadm will try
20:37 diegows_ joined #salt
20:38 UtahDave Snow-Man: Read through this doc: http://docs.saltstack.com/topics/development/external_pillars.html  especially the "useful globals" almost at the very bottom.
20:38 Snow-Man yea, found that..
20:38 UtahDave Snow-Man: some built in pillars are here: http://docs.saltstack.com/py-modindex.html#cap-p
20:39 Snow-Man was hoping for a more.. fully fledged example, heh.
20:39 Gareth Snow-Man: I'm doing something almost similar to what you're trying to do, but its for bind.  I can post it, might give you some ideas.
20:39 Snow-Man hmmm.
20:39 UtahDave doesn't look like we have a PG external pillar
20:39 Snow-Man Gareth: Would be great, thanks!
20:39 hazzadous joined #salt
20:40 taylorgumgum joined #salt
20:41 nahamu well that's more interesting: IOError: [Errno 20] Not a directory: '/root/salt-test/salt-0.17.1.solaris-2_11-i86pc_64bit/library.zip/orig-prefix.txt'
20:41 dave_den nliadm: i had issues with bdist_esky in a virtualenv on ubuntu 12.04 the other day.  unfortunately i didn't spend much time looking into the cause. I got the same 'no module named site'. i did as utahdave is suggesting now, which gave me another error
20:41 forrest Is the frequency at which the salt user group emails are sent out set by something, or is it just based on number of messages?
20:41 dave_den heh, which was that orig-prefix error.
20:42 Gareth Snow-Man: no worries.  Might be a bit...$dayjob gets in the way of playing with Salt :)
20:42 nahamu UtahDave: you were able to build 0.17.1 into an esky dist on Ubuntu, right?
20:43 nahamu (i just tried hacking bbfreeze to tell it to treat 'sunos5' the same as 'freebsd')
20:43 Snow-Man sure, np.
20:43 UtahDave nahamu: I haven't tried with 0.17.1 yet.
20:43 nahamu UtahDave: ah, which one did you last build?
20:43 UtahDave 0.16.4
20:44 Brew joined #salt
20:44 forrest Gareth, yea same here.
20:44 UtahDave I'll be doing 0.17.1 sometime this week.
20:45 nliadm UtahDave: no joy, still says no module named site
20:45 nahamu ImportError: cannot import name Queue
20:46 UtahDave nliadm: you added site to this list here?  https://github.com/saltstack/salt/blob/develop/setup.py#L349
20:47 nliadm I added an append after the linux check
20:47 ausmarton joined #salt
20:48 nliadm after line 376
20:48 wkf joined #salt
20:48 UtahDave ok, that should probably work.
20:49 dave_den Queue should also be added to the freeze list
20:49 nliadm it doesn't work, though
20:49 eliasp is there a better way  to access files in win_repo through salt:// than symlinking the win_repo into the corresponding environment?
20:49 UtahDave nliadm: what OS are you on?
20:50 nliadm linux, gentoo
20:50 nliadm running all this in a fresh virtualenv, though
20:50 UtahDave nliadm: hm. ok. I've never attempted those builds on gentoo nor in a virtualenv.
20:51 puffrfish joined #salt
20:51 nliadm it's more likely the virtualenv is the issue
20:51 UtahDave nliadm: very likely
20:51 dave_den nliadm: try uninstalling easy_install in the virtualenv
20:51 nliadm will give it a shot
20:52 dave_den easy_install monkeypatches itself to disrupt the loading of the real site module
20:52 dave_den then it does crap and then load the real site module
20:52 dave_den it's caused me problems in the past
20:52 nahamu Oh, you guys don't do builds in virtualenvs?
20:53 nahamu "You should probably set up a virtualenv in order to limit the scope of Q/A." made me think otherwise.
20:53 honestly can I put files into pillar?
20:54 honestly I have some private keys that need to be distributed to minions
20:54 nliadm Cannot uninstall requirement easy-install, not installed
20:54 honestly or is there another way to do this?
20:55 honestly I don't *really* care about minions not seeing each others' private keys but it would be nice to have
20:55 nliadm is it distribute or setuptools, I can never remember which is which
20:55 nahamu honestly: could you just put the private keys themselves into the pillar and then have the file be templated to pull the pillar data?
20:55 toastedpenguin joined #salt
20:55 kula i really want to write an addon to the file module that allows you to go "here's the file, and this pillar contains the decryption key to decrypt it"
20:56 nahamu (I'm assuming you're talking about .ssh/authorized_keys or /etc/ssh/<host key files>)
20:57 honestly they're rsa keys for certificates actually
20:57 nahamu honestly: ah
20:57 honestly I'm doing an openvpn setup
20:57 nahamu I wish I knew offhand how the guy setting up our webservers ships the SSL private keys.
20:58 nahamu but I do not.
20:58 honestly meh, I'll just put them into salt
20:59 nahamu should I make a fresh environment for the esky build but not use virtualenv?
21:00 jumperswitch joined #salt
21:00 cewood joined #salt
21:01 mgw joined #salt
21:05 woebtz_ joined #salt
21:06 eliasp I'm running into https://github.com/saltstack/salt/issues/7967 so I need to replace cp.py on my Windows minion, but there exists only the compiled module cp.pyc, so how would I go about replacing cp.pyc with https://raw.github.com/saltstack/salt/develop/salt/modules/cp.py
21:06 JasonSwindle joined #salt
21:07 eliasp ah, it might work automagically by simply placing cp.py next to cp.pyc
21:07 eliasp let's see
21:07 nahamu I think I did it!
21:07 nliadm nahamu: do tell
21:07 UtahDave carmony: I found the error that was cutting the wait timeout too short.  There's an open pull req for it on the salt-cloud github repo
21:08 carmony UtahDave: nice
21:08 UtahDave eliasp: just delete the cp.pyc and drop your cp.py in there
21:09 rlarkin joined #salt
21:09 eliasp UtahDave: worked! awesome!
21:09 nahamu nliadm: I did it in a zone. I installed enough to get pip to be able to install and upgrade all deps
21:09 eliasp UtahDave: now let's see how my win_repo workaround works ;)
21:09 nliadm in a zone?
21:09 nahamu I hacked bbfreeze to treat sunos like freebsd
21:09 nliadm oh
21:09 nahamu I'm on SmartOS, not gentoo
21:09 Gareth Snow-Man: still a work in progress but might give you some ideas: https://gist.github.com/anonymous/7108221
21:10 nahamu but I didn't use a virtualenv.
21:10 nliadm did you hit the 'site' problem?
21:10 nliadm ah
21:10 nliadm ok
21:10 nahamu I don't think so.
21:10 colinbits joined #salt
21:10 nahamu but if I remove the system python it can't find libpython2.7.so.1.0 so it's not perfect yet.
21:11 pipps joined #salt
21:11 thlig joined #salt
21:12 nahamu seems like patchelf isn't being invoked by bbfreeze...
21:14 nahamu it also didn't include the "py" binary that https://pypi.python.org/pypi/bbfreeze/ says it should...
21:15 soukihei joined #salt
21:16 nahamu though maybe that's esky's fault
21:16 nliadm do you have patchelf in your path?
21:17 nliadm it just keeps chugging even if it can't find it
21:17 yidhra joined #salt
21:19 nahamu I do.
21:19 honestly anybody want to proofread my init.sls?
21:19 nahamu I could test to see if it's actually getting called...
21:19 honestly http://ix.io/8Hr
21:21 nahamu nliadm: yeah, it's never even calling patchelf
21:21 UtahDave FYI  Rackspace has a Developer Trial that gives you up to $50 of free Rackspace service usage per month for the first 6 months.  http://developer.rackspace.com/devtrial/
21:21 forrest UtahDave is shilling for Rackspace? :O
21:21 amckinley joined #salt
21:22 dave_den honestly: youre requires are not correct
21:22 dave_den your
21:22 UtahDave forrest: ha ha!
21:22 honestly I thought so
21:22 forrest honestly, are you on 0.17?
21:22 honestly All I want to achieve is to get openvpn restarted after it's all fixed up
21:22 honestly should be
21:23 dave_den http://docs.saltstack.com/ref/states/ordering.html#the-require-requisite
21:23 honestly lemme check
21:23 honestly root@bb1:~# salt-minion --version
21:23 JasonSwindle left #salt
21:23 honestly salt-minion 0.17.0
21:23 forrest just trash the requires then
21:23 honestly y?
21:23 forrest it should process in order, so it will do the client.conf, the crts, and the key, then restart the service if the update was found
21:24 honestly ok
21:24 forrest you don't have to do requires any longer in 0.17.
21:24 forrest it just processes in order now (the option should be enabled by default)
21:24 forrest your master is running 0.17 right?
21:24 honestly 0.16.4 ):
21:24 honestly let's fix that
21:24 forrest UtahDave, do you guys choose the frequency at which the user group sends out emails?
21:24 forrest honestly, ok
21:25 honestly does salt still resolve dependencies though?
21:26 nahamu ah, my patch to bbfreeze is wrong.
21:26 forrest what do you mean honestly?
21:26 dave_den you still have to declare requires in 0.17 if you want functions to depend on eachother for execution.
21:27 honestly right
21:27 honestly so requires still *work*
21:27 forrest yea of course
21:27 dave_den right
21:27 forrest but if you do it in order
21:27 forrest it will do A, then B, then C, and so on
21:27 honestly I guess I should watch all of those files
21:27 dave_den but forrest is probably implying that you can leverage state_auto_order to have your states run in order
21:27 forrest correct
21:27 dave_den which makes things a bit more natural
21:27 forrest all you have to do is reorganize your state that you have there
21:28 forrest so do ca.rt, then the other crts/keys.then the client.conf, and do the openvpn service with that watch command.
21:29 dave_den honestly: yeah, a change to any of those files means you probably want to restart openvpn, so i would just watch them in the openvpn service state
21:29 honestly http://ix.io/8Hx
21:29 dave_den looks good
21:30 forrest *thumbs up*
21:30 carmony dave_den: so if I wanted to test your pull request, what should I do?
21:30 honestly should I set up an openvpn-client salt-formula for this?
21:30 forrest dave_den, gotta leverage the auto_order, helps keep the confusion to a minimum when explaining salt.
21:31 honestly I don't need a jinja shebang to have the grains[] there evaluated by jinja, or do I?
21:31 forrest that would be cool honestly, there isn't one in the repo yet
21:31 forrest You do
21:31 dave_den carmony: ?   did you mean UtahDave?
21:31 forrest do #!jinja|yaml
21:31 honestly ok
21:31 carmony dave_den: yes, sorry :P
21:32 * dave_den sings "These are the Daves I know"...
21:32 honestly is there a template repo for a saltstack formula?
21:33 jcsp joined #salt
21:34 forrest https://github.com/saltstack-formulas/
21:34 amckinley joined #salt
21:34 forrest honestly, just build it in your own github repo, then I think you have to email whiteinge or something to get it implemented? I keep forgetting to do that with my fail2ban formula
21:34 honestly root@bb1:~# salt-call -l debug state.sls openvpn test=True
21:35 honestly is something wrong with that call?
21:35 honestly it throws AttributeError: 'list' object has no attribute 'get'
21:35 honestly oh
21:35 honestly do I need to include the state in my top.sls for state.sls to work?
21:35 forrest no
21:35 nahamu if I can figure out why bbfreeze isn't calling patchelf, I might be done...
21:35 forrest the top file is only run through when you use highstate
21:36 honestly the openvpn dir is next to my top.sls
21:36 honestly and the init.sls is in there
21:36 forrest and then you have the init.sls?
21:36 honestly so "state.sls openvpn" should be the right invocation
21:36 forrest you don't have a target
21:36 honestly I'm callign that on the minion
21:36 honestly it's salt-call
21:37 micah_chatt Sooo does anyone from saltstack know why 0.17.1 is not packaged for ubuntu 13.04?
21:37 forrest oh duh, do --local
21:37 honestly why --local?
21:37 honestly is that a new 0.17 thing?
21:37 forrest salt-call --local state.sls openvpn -l debug test=True
21:37 forrest no it's not new, that's just how I've always run my masterless minion *shrug*
21:37 honestly lol
21:37 forrest so I figured it might be worth a shot
21:37 honestly ...wat
21:38 honestly I get sort-of empty output
21:38 forrest micah_chatt, there is a bug
21:38 mgw joined #salt
21:38 forrest at least there was with 0.17 for raring
21:38 carmony I need like a week of being off work
21:39 carmony to just play around with salt-cloud
21:40 forrest joehh, do you know if the issues got fixed for raring?
21:41 honestly http://ix.io/8HC
21:41 honestly oh
21:41 honestly no matching sls found
21:41 honestly took me way too long to see that
21:42 honestly do I need to refresh something to get the minion to find the state?
21:42 honestly (it's not a masterless setup)
21:43 forrest the state lives on the master right?
21:44 forrest can you use salt-call to apply a state locally as opposed to checking in with the master? hmmm
21:44 forrest I haven't tried that...
21:44 honestly nvm
21:45 lineman60 joined #salt
21:45 forrest honestly, did you figure it out?
21:45 honestly it works if I do it from the master
21:47 honestly is this going to make it ensure the package is installed? http://ix.io/8HF
21:48 honestly hm
21:48 honestly that doesn't work
21:48 APLU joined #salt
21:48 honestly because yaml
21:48 honestly how do I make that work?
21:50 honestly there we go, http://ix.io/8HG is the right one
21:52 forrest huzzah yaml!
21:53 honestly if only json had comments
21:53 c0bra joined #salt
21:53 forrest ehhhhhh
21:53 elfixit joined #salt
21:56 c0bra joined #salt
21:57 m0hit joined #salt
21:59 Kingdutch left #salt
22:00 blee_ joined #salt
22:00 Gareth hm. can you have a watch on one file do a restart but a watch on another do a reload?
22:00 joehh forrest: not that I am aware of
22:01 eliasp honestly: why don't you re-use the Salt CA, so you don't have to manage 2 CAs (one for Salt, one for OpenVPN)?
22:02 m0hit joined #salt
22:02 rgarcia_ joined #salt
22:03 Linz joined #salt
22:05 UtahDave carmony: I just merged it.  If you clone https://github.com/saltstack/salt-cloud  and then cd into salt-cloud   and then run  sudo python setup.py install --force  you'll have the latest and greates
22:05 UtahDave t
22:06 dave_den Gareth: if you have two ID declarations for your service, it will probably work fine.
22:07 dave_den but you can't do it in one
22:07 UtahDave forrest: The email list should send you the emails as they come in unless you choose the daily digest option
22:08 jcockhren hey. in order to get halite running, does it require salt to be install via pip or the setup.py script even if salt-master is already present via apt?
22:09 UtahDave jcockhren: No, I don't think it matters how you've installed salt
22:09 UtahDave jcockhren: just that it's running and configured correctly
22:09 phab3k just make sure it's kosher :D
22:11 jcockhren UtahDave: I'm getting a "ImportError: No module named salt.exceptions"
22:11 jcockhren this is using halite installed from pip
22:11 UtahDave what version of Salt, jcockhren?
22:12 jcockhren 0.17.1
22:12 jcockhren ubuntu 12.04
22:13 UtahDave samsalt: do you know why jcockhren would be getting the error above?
22:14 ncjohnsto joined #salt
22:15 jcockhren I see
22:15 drogoh joined #salt
22:15 jcockhren UtahDave: I needed to rn halite as sudo
22:16 eliasp is it somehow possible to provide multiple full_names for a win_repo pkg? e.g. MS Office installs multiple products through a single installer…
22:16 eliasp otherwise the other installed components can't be matched…
22:17 rgarcia_ joined #salt
22:18 forrest UtahDave, I'll have to double check when I get back, was noticing I was getting 1 a day, then 2 a day, and now it's something like 4 a day
22:19 pipps joined #salt
22:20 UtahDave jcockhren: Ah, yeah. Halite has to be running as the same user as salt-master is running
22:21 jcockhren ok. So it starts, however, in the browser I get a "File Not Found" message
22:21 rgarcia_ joined #salt
22:21 jcockhren I'm guessing it needs to app/main.html file generated?
22:21 jcockhren samsalt: ^
22:24 alunduil joined #salt
22:25 UtahDave are you running it in debug mode?
22:25 jcockhren let me try that
22:26 pipps1 joined #salt
22:26 UtahDave I think you might have to connect to port 8080 or 8000 or something like that.
22:26 jcockhren yeah.
22:26 UtahDave jcockhren: I have to head home now.  I can help you in the morning if you haven't gotten it working correctly
22:27 jcockhren I'm pulling up the browser on the correct point
22:27 jcockhren UtahDave: word.
22:27 jcockhren port*
22:30 rgarcia_ joined #salt
22:31 m0hit_ joined #salt
22:32 pipps joined #salt
22:33 druonysuse joined #salt
22:33 druonysuse joined #salt
22:34 rgarcia_ joined #salt
22:43 ccase joined #salt
22:44 oz_akan_ joined #salt
22:45 mgw ddns.delete (which I wrote) takes a zone, a host, and an optional record tpe. In the case of reverse zones, the host is an int-like string (e.g.: ddns.delete 4.20.172.in-addr.arpa 60 PTR)
22:46 pentabular joined #salt
22:46 mgw It seems to me that something is converting that 60 to an int
22:46 mgw before it gets to the ddns module
22:46 dthom91 joined #salt
22:46 joehh joined #salt
22:49 mgw I can bypass the issue by casting it back to a string in the ddns module… but is there a better way?
22:51 pentabular joined #salt
23:05 kermit joined #salt
23:05 lineman60 joined #salt
23:09 pabelanger joined #salt
23:17 josephholsten joined #salt
23:20 pipps joined #salt
23:20 Gareth Is there an issue with having dots in the name of a pillar file?
23:21 Gareth damn.  there would be.
23:24 amahon joined #salt
23:38 wkf joined #salt
23:41 amckinley joined #salt
23:46 cachedout joined #salt
23:50 AdamSewell joined #salt
23:59 Thiggy joined #salt

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