Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-06-12

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

All times shown according to UTC.

Time Nick Message
00:00 bemehow joined #salt
00:01 jschadlick left #salt
00:07 godber juanlittledevil: don't you need to run ...
00:08 godber salt -G '*db*' state.highstate
00:08 godber to get a grain match rather than id match?
00:09 juanlittledevil hmm… that's a good point.. do'h trying ...
00:09 z3uS joined #salt
00:11 juanlittledevil actually… it was: salt -G 'roles:prod' state.show_highstate
00:11 godber oh that cant be it, since db is in the name
00:11 godber hmm ok
00:12 juanlittledevil actually. I'm not sure this is correct either
00:12 juanlittledevil show_highstate shows me no tree
00:12 godber ah
00:13 Nazca joined #salt
00:13 fragamus joined #salt
00:14 godber geez, I didnt even know about state.show_highstate
00:14 juanlittledevil heh
00:16 juanlittledevil Hmmm well I guess I better re-read the docs. I'm sure it's something silly I'm doing in how I'm configuring the minion (or the master)
00:16 godber prod shows up an aweful lot in your top.sls
00:16 godber don't you have to have a base?
00:16 juanlittledevil I do.
00:16 godber (or override it in master)
00:16 godber ok
00:17 juanlittledevil I'm just playing with this at this point.
00:17 juanlittledevil enough to make me very dangerous… prod right now == sandbox ;)
00:18 godber same here, though my prod is prod
00:18 bemehow joined #salt
00:18 godber good luck
00:18 juanlittledevil thanks.
00:18 KyleG So this is you guys then: http://rationalmale.files.wordpress.com/2011/09/doitlive.jpeg
00:19 juanlittledevil hahaha… no. this is in a very controlled environment. and isn't going to go live until I've fully tested it and fully understand it.
00:19 godber KyleG: no, no, I have a test environment, I am only taking well tested baby steps in prod
00:20 juanlittledevil hehehe… this is a good one: http://imgur.com/y7Hm9
00:20 KyleG lol yeah that's a good one
00:20 godber except it makes me want a beer
00:20 KyleG 10 minutes until beer thirty here in CA. <3 having a keg in the office
00:21 juanlittledevil nice.
00:21 juanlittledevil no keg… only boxes of beer in the fridge. same diff
00:22 luminous juanlittledevil: please use a pastebin service when you need to share code/etc
00:22 juanlittledevil sounds good
00:22 juanlittledevil will do
00:22 KyleG juanlittledevil: gist.github.com is really nice as well, especially isnce you can track it an do private/public ones
00:22 KyleG since*
00:23 juanlittledevil nice!
00:24 KyleG there's also a chrome extension for managing your gists
00:24 juanlittledevil yeah I don't think I'm goign to be much further in 10 minutes. I'll read up some more tonight on environments and try again tomorrow.
00:24 juanlittledevil thanks for you help guys.
00:24 mannyt joined #salt
00:24 KyleG http://www.gistboxapp.com/
00:30 bemehow joined #salt
00:34 terminalmage joined #salt
00:39 jerkbot1 joined #salt
00:48 bemehow_ joined #salt
00:52 UtahDave joined #salt
00:55 admineric left #salt
00:56 mgw joined #salt
01:14 austin987 joined #salt
01:32 auser joined #salt
01:32 auser hey all
01:40 m_george|away joined #salt
01:44 jeddi joined #salt
02:02 milind joined #salt
02:16 oz_akan joined #salt
02:19 fxhp woot, I wrote a "central-log-server" syslog-ng salt state
02:20 Furao joined #salt
02:34 mannyt joined #salt
02:42 austin987 joined #salt
02:52 longdays_ joined #salt
02:54 raydeo joined #salt
03:08 clintberry joined #salt
03:09 cxz joined #salt
03:20 sgviking joined #salt
03:20 kmwhite joined #salt
03:27 Ryan_Lane joined #salt
03:31 longdays joined #salt
03:36 Furao let's say that to deploy an infra you need to download 150 files from various open source repository/sources
03:36 Furao and I don't mean APT repositories, these are relatively well mirrored.
03:37 Furao I mean, pypi, github, bitbucket, ppa, etc
03:38 Furao and my conclusion is that 95% of errors are caused by these source unavailable or that cut the connection in the middle
03:41 Furao and I wonder how to handle this. should a big salt deployment mirror all the third party software?
03:41 Furao but some installer don't support salt:// protocol
03:41 Furao so, it need to be under http://
03:44 Furao keeping a local copy of everything makes installation faster
03:46 ragnar0k joined #salt
03:48 joehh Furao: we've struggled with this too
03:48 joehh mainly with windows installers - some are in the hundreds of mb and a pain to keep in repos or move around
03:49 Furao I got few big archives too, such as graylog2 server requires around 120 mb of files
03:49 joehh we've ended up mirroring them and rsyncing around as required
03:49 joehh for us it is matlab compiler run times...
03:50 Furao in my case, is that I wrote a testing tool for salt states, a single test run can execute the same states 20 times with various settings/pillars. and I'm hit with those errors more often than usual
03:50 Furao that's my current idea to handle it
03:50 Furao a centralized big file repo available trough rsync and/or http
03:50 Furao it hold copy of everything
03:51 Furao master can rsync, or it run a http cache proxy (only requested files are kept on disks)
03:51 joehh yes
03:52 Furao this way, files are to minion even if they don't have internet access
03:52 basepi joined #salt
03:52 Furao are AVAILABLE
03:53 Furao save on bandwidth
03:53 hvn joined #salt
03:58 joehh yes - generally, our environments are heavily restricted in terms of internet access
04:03 ragnar0k joined #salt
04:13 sarkis joined #salt
04:17 LarsN joined #salt
04:17 LarsN I appear to be running into this bug: https://github.com/saltstack/salt/commit/75c6172df8062d0d6aa9a999756575db9bee6c09
04:17 LarsN or more specifically https://github.com/saltstack/salt/issues/4629
04:18 Corey LarsN: Have a gist?
04:18 LarsN when trying to push my ubuntu laptop to highstate I see these in the minion logs
04:18 LarsN 2013-06-11 23:16:02,712 [salt.loaded.int.module.apt][WARNING ] Unable to import "softwareproperties.ppa": No module named softwareproperties.ppa
04:19 LarsN it's only configured at the moment to install vim and sockstat
04:20 bemehow joined #salt
04:21 LarsN master is 0.15.1, and my minion is 0.15.3 if that matter
04:21 bemehow_ joined #salt
04:23 xet7 joined #salt
04:24 Corey LarsN: And uh... don't do that.
04:24 Corey LarsN: *always* upgrade the master *first*
04:24 Corey LarsN: I'd be curious to see if this problem magically goes away once that's done.
04:25 xet7 how do I set timezone in salt state?
04:25 Corey xet7: Good question! I cheat!
04:25 Corey xet7: http://pastebin.com/fUwxDsbH
04:27 LarsN master upgraded,
04:27 LarsN testing highstate again.
04:28 xet7 Corey: thanks, I was missing force: true :)
04:28 Corey LarsN: Not a huge deal with 0.15.1 to 0.15.3, but some releases contain breaking changes.
04:28 Corey xet7: A common problem. Glad you got it sorted!
04:28 LarsN is there a preferred pastebin?
04:29 LarsN perhaps that warning is a red herring, and I'm doing something else stupid.
04:29 LarsN (highly likely)
04:29 Corey LarsN: One that works. :-)
04:29 Corey LarsN: You'll note I just used pastebin.com
04:30 Corey LarsN: But basically "don't flood the channel."
04:30 Corey LarsN: I'd be idly curious to see the state file that triggers this, if you can isolate it.
04:31 LarsN http://pastebin.com/xKKgTjhX
04:31 Corey Let's see.
04:31 LarsN and the associated minion log
04:31 LarsN http://pastebin.com/FigwK1jw
04:32 Corey LarsN: Hmm. Humor me a second here, switch  salt -G "os_family:debian" state.highstate -v to salt '$MINION_YOU_ARE_HAVING_TROUBLE_WITH' state.highstate -v
04:32 Corey Grain matching is needless complexity for this, and I only want to see it on the minion we're diagnosing.
04:33 Corey LarsN: Also, which Ubuntu version are you on?
04:34 LarsN surly is 13.04, salt01 (the master) is 12.04
04:34 Corey LarsN: And you see this on the master when you target it as a minion?
04:34 Corey I lack a 13.04 box.
04:34 LarsN the logs are on the minion
04:35 Corey Understood.
04:35 LarsN I could switch the top.sls to upgrade the master.
04:35 efixit joined #salt
04:35 LarsN which would be updating itself if you like.
04:35 Corey LarsN: That "no top.sls found" is disturbing.
04:35 Corey :-)
04:35 LarsN it's only installing two packages.
04:36 xet7 how do I then force ntp to restart after timezone change?
04:36 Corey LarsN: What does top.sls look like?
04:36 LarsN it's in the first paste.
04:36 Corey xet7: Why would you?
04:36 LarsN that was the output, cat top.sls, and cat ubuntu.sls
04:36 Corey xet7: With regard to timezones, ntp is like a honey badger, in that it just doesn't care.
04:37 xet7 Corey: clock was not correct on vagrant virtualbox
04:37 xet7 Corey: restarting fixed it
04:37 Corey Ah, there it is.
04:37 Corey xet7: How off was it?
04:37 xet7 about 3 minutes
04:37 Corey xet7: ntpdate to get it synced, then start ntpd.
04:38 xet7 Corey: can it be done in salt state?
04:38 Corey LarsN: I think your /srv/salt heirarchy is... a little off. :-)
04:38 LarsN entirely possible.
04:38 LarsN <--- really really new to salt.
04:39 LarsN and not very smart either :(
04:39 LarsN :)
04:39 Corey LarsN: Ah, okay. For some reason I thought you were an old hand at this.
04:39 Corey Hang on, let me dig out mine.
04:39 LarsN ubuntu.sls needs to be under /srv/salt/dev/ubuntu.sls
04:39 LarsN because I defined dev: in the top.sls sound about right?
04:40 Corey LarsN: http://pastebin.com/rh96MPNk
04:40 Corey That's how I've set mine up.
04:41 Corey LarsN: (I'm also the Ubuntu packager these days, so if there's a bug I'd like to know about it!)
04:42 LarsN It seems to have installed fine.
04:42 LarsN the upgrade on the server was missed because it was held back and I overlooked it.
04:42 LarsN dist-upgrade rather and upgrade.
04:44 Corey xet7: Yes, you *can* do it in Salt states, just not sure why you would.
04:44 xet7 Corey: is your salt tree for one or more servers?
04:46 LarsN fixed
04:46 Ryan_Lane joined #salt
04:47 hvn left #salt
04:48 LarsN one small change, and it's now working.
04:48 Corey xet7: It's a number of systems, I use wildcard matching in the minion name to target.
04:48 Corey LarsN: Sweet, what was your change?
04:48 LarsN moved "surly" into "base:" rather than "dev:"
04:48 LarsN which tells me I need to look at my salt hierarchy in the master config.
04:49 LarsN I now have another, likely noobish, question.
04:49 Corey LarsN: Fire away.
04:49 LarsN with Chef, the "chef client" executes on the client ever X minutes checking the chef-server for updates,changes
04:50 Corey LarsN: Correct.
04:50 LarsN if I do set something up for all "os_family:debian" machines
04:50 LarsN and two of them are powered off at the time I push highstate, what kind of magic is required for those nodes to get caught up when they do come online?
04:50 Corey LarsN: That seems to be a somewhat odd and arbitrary targeting distinction, why'd you pick it?
04:50 LarsN is it as silly as putting a cron entry on the master to push highstate every x minutes.
04:51 Corey salt \* state.highstate is what I use. :-)
04:51 LarsN what if it was more like -G "vm_image:CLI-Testing"
04:52 Corey LarsN: Why use -G at all?
04:52 LarsN I suppose I could just push higstate to everything.
04:52 Corey That's what I do. If highstate doesn't work everywhere, fix it. That's my personal position on it.
04:52 Corey Now, on to your question
04:52 LarsN regardless, is the proper method to catch nodes that were offline, by just putting highstate into cron?
04:52 Corey LarsN: Okay, now we're moving away from "Salt" specific questions into more general guidance around best practices. That's fine, you just need to be aware that I'm no longer advocating what YOU should do, but what I do. I'm opinionated around this.  :-)
04:53 LarsN no problem.
04:53 Corey Originally I put a cron job on each minion to call salt-call state.highstate.
04:53 LarsN thankfully my use case is very narrow at the moment.
04:53 Corey This screwed me.
04:53 Corey "Okay, I'm testing some changes and OH GOD THEY WENT OUT TO PRODUCTION"
04:53 Corey On the master I cronned salt \* state.highstate every 15 minutes or so, which was fine.
04:54 Corey Then pushed out a file, /etc/cron.d/salt
04:54 Corey It was one line: @reboot /usr/bin/salt-call state.highstate
04:54 LarsN that might actually work pretty well for my specific needs.
04:54 Corey This neatly solved the "one touchpoint to start making changes without breaking prod" and "keep systems from coming up with old code" problems.
04:54 LarsN "team" shares a common VM image.
04:55 LarsN option a) upload / download new .img files every time we make changes
04:55 LarsN option b) keep all the important bits, and expect co-workers to update via git at the beginning of each shift
04:55 LarsN or option c) automate that @#$%
04:55 LarsN options a and b suck.  ;)
04:55 Corey Option C doesn't preclude A or B. :-)
04:56 Corey Ensure latest git checkout with the git module, is one way to do it.
04:56 Corey What do these VM images do?
04:56 LarsN they contain all the CLI clients for a large public cloud.
04:57 Corey LarsN: Sounds like AWS. ;-)
04:57 Corey I've written a number of tools around that.
04:57 LarsN it's like AWS, but more opensource.
04:57 Corey OpenStack / Rackspace sounding then.
04:57 LarsN I'm at HP Cloud Services.
04:57 Corey "Not that open." "Joyent."
04:57 Corey Ooh.
04:57 Corey LarsN: I've heard good things about your group.
04:58 LarsN other parts of the group are using salt.  I'm in the NOC currently
04:58 JordanRinke <- Rackspace
04:58 LarsN so these are "test" vms, since there's a lot to test, and things change rapidly.
04:58 Corey LarsN: Richard Buckingham was there for a bit heading the group IIRC.
04:58 JordanRinke I think we are supposed to have some kind of battle of snarky remarks or something now.
04:58 LarsN JordanRinke: I think you and I have been in the same room as one another on at least three occations. :)
04:58 LarsN it's always been civil :)
04:59 Corey LarsN: Pushing out VM images does indeed suck mightiliy.
04:59 Corey "We changed six files, have another 3GB image."
04:59 LarsN especially when you're the ONE guy not running vmware, and insist on converting the vm disk to qcow2 for using in kvm
05:00 Corey "We've checked the files into git, because we're a pack of morons. git clone now takes six weeks."
05:00 Corey LarsN: Oooh.
05:00 JordanRinke Well in that case, I am going to sleep.
05:00 LarsN I'd be more likely to put the 6 changes into git, and ask everyone to clone.
05:00 JordanRinke LarsN: Best of luck to you.
05:00 Corey LarsN: Right, the problem with that is you need the VM images to self-update to latest on boot, which is what the @reboot cron job does.
05:01 LarsN JordanRinke: '73
05:01 LarsN Corey: I don't know why I didn't think of that.
05:01 LarsN that's a brilliant solution to a simple problem.
05:01 Corey LarsN: Because you're too close to the problem.
05:01 LarsN the rest of version control we get :)
05:02 Corey LarsN: I presume you're not booting these things in an offline state?
05:02 dthom91 joined #salt
05:02 LarsN honestly, these test vms are stupid ubuntu installations, with some magic related to environment variables, and a metric ton of python virtualenvs
05:03 LarsN nothing fancy, just a lot of headaches I'd rather never have to manually recreate
05:03 LarsN :)
05:03 Corey This is kinda the point of configuration management. :-)
05:04 LarsN I at $oldjob spent 6 months getting cfengine to a point we could shoe horn it into our existing server infrastructure.
05:04 Corey Are you me? :-)
05:04 Corey What I like about Salt is that you can stick it in strictly as remote execution to start.
05:04 Corey Then start moving things into it piece by piece.
05:04 LarsN two weeks before d-day, the CTO came in and said...  We need crowbar & openstack, which means learn Chef.
05:05 Corey *shudder*
05:05 LarsN I left before we finished implimenting chef.  They've since given up on openstack and crowbar.  Are now using Cloudstack
05:05 LarsN and I doubt they've ever done anything about configuration management.
05:06 Corey LarsN: Are you guys strictly Ubuntu at this point?
05:06 LarsN I don't think so.
05:06 LarsN but for my testing purposes, I'm lazy, and ubuntu was easy.
05:07 Corey LarsN: Well, the reason I ask is for grain targeting.
05:07 Corey LarsN: Also, one thing that will definitely save your bacon, both with salt as well as many other things, is a strictly enforced naming convention.
05:07 LarsN yeah, for the work we're doing in the NOC, that's already in good shape.
05:07 Corey hosttype001.vlan.dc.internaldomain.com is one I like, but it'll depend on your area.
05:08 LarsN my private lab is all sorts of chaos.
05:08 LarsN which works because if I can make it work in chaos, then in sanity it'll be easy.
05:08 LarsN at least that's the theory.
05:08 Corey Then in your top.sls you can target, say, /etc/resolv.conf to differentiate based upon the vlan string.
05:16 timl0101 joined #salt
05:18 xet7 How do you include substates in init.sls, like /srv/salt/projectname/init.sls includes /srv/salt/projectname/nginx/init.sls ?
05:18 oz_akan joined #salt
05:19 Corey xet7: I don't believe salt supports that at present.
05:19 Corey (Saltstack folks, feel free to disprove that!)
05:21 KFDM xet7: Corey: can you not use the include directive ?  https://salt.readthedocs.org/en/latest/ref/states/include.html
05:22 Corey KFDM: I've... somehow never seen that before. Amazing!
05:22 KFDM :)
05:23 Furao saltstack folks usually sleep at this hour
05:23 Furao https://github.com/bclermont/states plenty of example of include usage
05:23 LarsN Sleep is a totally inappropriate substitute for caffeine.
05:24 xet7 KFDM: Thanks, seems to work! :)
05:27 LarsN can you do matching in a .sls based on grains values?
05:28 LarsN looks like the compound matchers document is what I'm looking for.
05:28 Corey LarsN: yes.
05:29 Corey Digging out a quick example.
05:29 Corey <     {% if grains['os'] == 'CentOS' or grains['os'] == 'Fedora' %}
05:29 Corey <       - yum-plugin-priorities
05:29 Corey (For package installation)
05:33 LarsN perfect.
05:34 cxz LarsN: I have a similar problem
05:34 cxz our existing infrastructure is non-cf managed
05:35 cxz and it all exists within bash scripts :<
05:35 cxz in a web interface called rightscale
05:35 LarsN One of the guys that just left my team went to some training with Rightscale last year.
05:37 cxz hah
05:37 cxz how did he find that
05:37 cxz that sounds...infuriating
05:38 LarsN I "think" they might be a partner.  Looks like it according to our site.
05:39 cxz hehe
05:44 MFen joined #salt
05:51 maspwr joined #salt
06:05 koolhead17 joined #salt
06:05 koolhead17 joined #salt
06:06 krissaxton joined #salt
06:09 Ryan_Lane joined #salt
06:17 Ryan_Lane joined #salt
06:22 f4cl3y joined #salt
06:22 f4cl3y joined #salt
06:24 pcarrier_ joined #salt
06:30 dthom91 joined #salt
06:34 `3rdEden joined #salt
06:42 pcarrier_ joined #salt
06:44 ragnar0k joined #salt
06:48 sebgoa joined #salt
06:54 dhgbrg joined #salt
06:55 pcarrier_ joined #salt
07:02 dhgbrg joined #salt
07:02 stevetodd joined #salt
07:05 Teknix joined #salt
07:09 hhtp joined #salt
07:14 nielsbusch joined #salt
07:17 vaxholm joined #salt
07:18 bbeermann joined #salt
07:19 scott_w joined #salt
07:20 stephen__ joined #salt
07:20 bbeermann hi guys. is there anyone having experience with using salt and buildout?
07:28 linjan joined #salt
07:30 dthom91 joined #salt
07:38 Teknix joined #salt
07:39 austin987 joined #salt
07:44 krissaxton joined #salt
07:54 ageron joined #salt
07:55 adotbrown joined #salt
07:58 stevetodd joined #salt
08:08 fredvd joined #salt
08:11 canci_ is there anyone maintaining a pacemaker configuration with salt?
08:12 felixhummel joined #salt
08:13 carlos joined #salt
08:15 sebgoa joined #salt
08:19 Valdo joined #salt
08:27 Teknix joined #salt
08:31 dthom91 joined #salt
08:33 krak3n` joined #salt
08:42 dhgbrg joined #salt
08:44 p3rror joined #salt
08:48 xerxas joined #salt
08:57 __gotcha joined #salt
08:57 __gotcha joined #salt
09:00 Teknix joined #salt
09:01 zooz joined #salt
09:03 Teknix joined #salt
09:03 efixit joined #salt
09:04 agend joined #salt
09:08 jesusaurus joined #salt
09:10 canci_ hm, it seems I managed to break the salt master. It is not reacting to any command it seems and there are zombie processes hanging around
09:18 entil joined #salt
09:31 dthom91 joined #salt
09:34 waverider joined #salt
09:52 bhosmer_ joined #salt
09:52 jeddi joined #salt
09:58 Teknix joined #salt
10:01 vaxholm joined #salt
10:06 lyddonb joined #salt
10:07 linjan joined #salt
10:07 jasonrm joined #salt
10:08 Kyle joined #salt
10:13 Teknix joined #salt
10:15 entil what's the regexp like for file.uncomment? I have this toggle whether or not an nginx installation is public or not, and I'd like to uncomment the acl. that's allow|deny but as a regexp that didn't work
10:19 unicoletti_ joined #salt
10:20 Furao file.sed, file.comment and file.uncomment aren't good idea. it's better to use file.managed with template
10:20 entil myeah, now that you say it I'm going "why didn't I think of that" :D maybe I can use a jinja if statement to render it in there
10:23 Furao file.sed, file.uncomment and file.comment is still kind of using bash
10:23 Furao you can't be sure 100% what will be the file at the end
10:24 entil just kicked up a new instance with an if block in the template, we'll know in about 10 minutes or so
10:25 Furao anyway it's easier to write/understand code in jinja than reverse engineer some file.comment/uncomment/sed
10:25 entil sure, I guess I just had my brain in a broken mode thinking about this ;) "hey I could comment these lines out.. actually I want the default to be commented out.. file.uncomment" instead of "how do I solve this the best" :P
10:30 Furao I almost got 30 000 lines of states and templates, and I couldn't survive without file.managed :)
10:31 Teknix joined #salt
10:31 pjs joined #salt
10:32 entil my biggest gotcha this week was that a lot of salt's slowness is actually extra work in my states. that I had to break out something into a completely other state and such
10:37 monokrome joined #salt
10:37 echos joined #salt
10:37 up_the_irons joined #salt
10:38 MFen joined #salt
10:38 N-Mi joined #salt
10:39 EugeneKay joined #salt
10:45 entil looks good with the if thing
10:46 efixit joined #salt
10:53 Teknix joined #salt
10:54 faust joined #salt
10:55 aleszoulek joined #salt
10:58 Furao slowness?
10:58 Furao it's probably because your state execute stuff when not necessary
10:59 Furao such as extract a file on each state.highstate
10:59 Furao extract it only if it had been changed
11:00 Furao salt '*' nrpe.run_all_checks
11:00 Furao it run all nagios nrpe check on all hosts
11:00 Furao so useful during deployments :)
11:01 Furao I should add a param to only return failed check
11:07 Katafalkas joined #salt
11:08 Katafalkas hey. has anyone tried doing some autoscaling module for salt ?
11:13 bhosmer_ joined #salt
11:13 danielbachhuber joined #salt
11:13 Furao why paraguay domain are so hard to get, I'd love to get a .py :(
11:14 Teknix joined #salt
11:19 adotbrown joined #salt
11:36 Teknix joined #salt
11:38 logix812 joined #salt
11:42 Vivek joined #salt
11:44 aleszoulek Katafalkas: We're using salt and autoscaling.. But I would hardly call that a module :)
11:45 Katafalkas aleszoulek: u mean u are using aws autoscaling ?
11:46 aleszoulek Katafalkas: No. We're using statsd to monitor req/s on loadbalancer and trigger salt runner, which creates/destroys a web node.
11:47 Teknix joined #salt
11:47 MrCeeJ joined #salt
11:48 MrCeeJ Hello! I am new to salt and am trying to get some information from our Cobbler build server into some Pillars to use for configuring our servers, the only real documentation I have been able to find is http://docs.saltstack.com/ref/pillar/all/salt.pillar.cobbler.html but that doesn't really help at all
11:48 entil Furao: yeah, I peppered a bunch of test -f statements, and broke off some stuff into separate states and now I have a partial idea on what to do next wrt checking files (there's a loop and whatnot) but no time to implement just yet
11:49 MrCeeJ is there anywhere I can go for some more detailed information / instructions / how-too ?
11:49 entil the powers that be had to be convinced this salt stuff wasn't just yak shaving
11:49 Katafalkas aleszoulek: statsd is that from etsy ?
11:50 backjlack joined #salt
11:51 aleszoulek Katafalkas: yep
11:52 aleszoulek Katafalkas: But that doesn't mind. You just need any service monitoring, which is able to trigger events when some predefined conditions (for example req/s) are met
11:53 dhgbrg joined #salt
11:56 Teknix joined #salt
11:56 Katafalkas is there a specific way you connect those two ? db ?
11:57 Katafalkas aleszoulek:
11:58 jgelens_ joined #salt
12:04 canci_ usually there is something a 1:n mapping from state instances to files: a state may create/modify some files
12:04 Teknix joined #salt
12:04 canci_ is there a way (or an example) how to implement an n:1 mapping for states? so to generate one file containing all instances of a specific state?
12:08 Teknix joined #salt
12:14 Teknix joined #salt
12:20 jeddi canci_: can you explain a use case?
12:22 jslatts joined #salt
12:25 jay joined #salt
12:26 jay hi Guys, loving getting into salt, finding a complete lack of documentation as to how I might get pillar data from my cobbler server
12:27 jay I've found the salt.pillar.cobbler page, but its rather lacking in actual how one might use it
12:27 jay can someone give me some pointers please ?
12:37 Vivek joined #salt
12:42 canci_ jeddi: I have a service that runs and monitors programs. I have to configure it with one single file. However I want to populate that file from the states of the individual services
12:43 Kamal_ How does salt figure out the machine's hostname by default?
12:44 jeddi canci_: oh, okay.   i don't have any insight unfortunately, sorry.
12:45 jeddi canci_: this program, btw, does it have the ability to take its configuration from a bunch of {daemonname}.d/*.conf files?
12:45 Kamal_ socket.getfqdn()
12:45 canci_ jeddi: no, unfortunately not
12:46 canci_ jeddi: I could of course create many those .d files and then collect them into one single file, but that feels quite hacky
12:46 jeddi canci_: dang.  there are some structures for doing the kind of stuff you're taling about  .. *perhaps* .. such as the way that ssh keys are handled.  or the augeas (sp?) thing.   you may need to write your own for a custom thing like this.
12:46 tempspace Kamal_: if socket.fqdn is something other than localhost, it uses that, then it checks /etc/hosts for something that isn't localhost and maps to 127.0.0.1
12:46 jeddi canci_: i wish more apps would implement the xyz.d/* approach to configuration files.
12:47 Kamal_ It returns localhost.localdomain
12:47 Kamal_ but hostname -f returns the correct fqdn
12:47 canci_ tempspace, Kamal_: had that issue too. The first entry in the first line matching 127.0.0.1 in /etc/hosts should be the fqdn
12:47 Kamal_ but why
12:48 Kamal_ I used to fix it that way
12:48 Kamal_ but it seemed hackish
12:48 jeddi Kamal_: what canci_ said.  check /etc/hosts as well as /etc/hostname   in the hosts file you'll often find the fqdn and the hostname on the same line.  reverse the order.
12:48 jeddi Kamal_: or just confgure the hostname in the minion configuration.
12:48 Kamal_ that's what I've been doing lately
12:48 tempspace I add it to /etc/hosts at the top or I hardcode the minion
12:48 tempspace hardcode the id rather
12:49 canci_ basically, socket.fqdn() does something different than the usual fqdn determination
12:49 jeddi Kamal_: i found some discrepancies even between debian and ubuntu on the fqdn being returned .. i lost interest trying to work it out, and i maintain a custom grains conf for each minion from the master now.
12:49 canci_ the first just reads the first entry for 127.0.0.1 in /etc/hosts (or the first entry at all? not sure) while the second does a reverse lookup on the own ip
12:50 jeddi Kamal_: it does feel a bit arse-about, but it solves the problem, as well as a few other things where i can't intuit [env] and [roles] just from the hostname (legacy hostnames that aren't embedded with information)
12:50 Kamal_ I see
12:50 jeddi Kamal_: keep it in mind anyway.  we all have different use cases.  and different levels at which arms are thrown in the air and we go 'that's not *too* hacky, now that I think about it'
12:51 tempspace I had to submit some code for the hosts module because it used to break salt if I put other entries for 127.0.0.1 in there because it would group them together and lose its place at the top of /etc/hosts
12:51 canci_ Kamal_: I stand corrected, I just read socket.getfqdn again and what I said is wrong. seems I confused it with another fqdn implementation then :/
12:52 Kamal_ canci_, putting the actual thing at the top fixes it
12:52 juicer2 joined #salt
12:52 Kamal_ weird
12:52 Kamal_ if I uncomment id: in /etc/salt/minion - it keeps crashing
12:52 Kamal_ no errors in /var/log/salt/minion
12:52 canci_ Ah no, found it
12:52 tempspace It's in https://github.com/saltstack/salt/blob/4a5a0c1c4372edf448415c0a60f0df30a91fe49a/salt/config.py get_id
12:52 tempspace if you're curious
12:53 oz_akan joined #salt
12:53 canci_ salt/config.py:542. It uses the first entry in /etc/hosts for 127.*. Which happend to be the unqualified hostname for me
12:53 Kamal_ I do have 127.0.0.1 localhost.localdomain localhost in /etc/hosts though
12:54 canci_ I had 127.0.0.1 oberon oberon.top.level.doma.in in /etc/hosts which never was an issue, until that code came along
12:54 Kamal_ it should be ip    fqdn alias1 alias2 etc
12:55 canci_ hm yes, hosts(5) agrees
12:58 Teknix joined #salt
12:59 Kamal_ ffs
12:59 Kamal_ 2013-06-12 12:59:12,428 [salt.minion    ][WARNING ] Starting the Salt Minion
12:59 Kamal_ and it just crashes
13:00 Kamal_ it works if I comment out id: in /etc/salt/minion
13:00 linjan joined #salt
13:01 Kamal_ alright what
13:01 DredTiger joined #salt
13:01 Kamal_ screw you salt :/
13:01 Kamal_ left #salt
13:02 Kamal_ joined #salt
13:02 Khollowa_ joined #salt
13:03 Gifflen joined #salt
13:06 Teknix joined #salt
13:06 bhosmer_ joined #salt
13:07 bejer Hi, how would you run multiple 'file.sed' on the same file? Write multiple state declarations, as it doesn't execute multiple "file:\n<indent>-sed" statements in the same state declaration ?
13:08 Furao bejer: under multiple state name
13:08 ronc joined #salt
13:08 Furao but use file.sed, file.comment and file.uncomment are not a good idea. it's like still using bash to manage configs. use file.managed - template: jinja instead
13:10 entil bejer: as someone who just today tried solving something with file.uncomment and had a duh moment, I testify to what furao is saying ;)
13:10 bejer Okay hoped I wouldn't have to make one for each sed - yeah could aswell just manage the whole file then :)
13:10 entil it's a cognitive glitch of old habits to not do jinja
13:11 Furao find . -type f -name '*.sls' | xargs cat | wc -l
13:11 Furao 17495
13:11 Furao not a single file.sed, comment or uncomment in theses
13:12 entil I got 840 lines :o
13:12 Furao if I would rely on see I would still troubleshoot my first states
13:12 Furao see -> sed
13:12 Furao my states can cook my breakfast :P
13:12 bejer nice
13:13 entil git push && saltutil.sync_all && state.sls went wtf, sync_all && state.sls worked
13:13 entil some kind of lag, that's happened to me a couple of times
13:14 entil dunno if the second sync_all was unnecessary or if just waiting would have helped, but that's annoyin
13:14 bejer Furao: Do you support many different distributions/systems with your states?
13:14 Furao bejer: no just ubuntu 12.04 so far
13:14 entil dunno if it's true but christophe pettus mentioned ubuntu 12.04 in a talk as being the go-to distro on aws
13:14 Furao it's a product/service by itself that i sell
13:15 entil I have no idea what he bases that on, I've never had a hitch with debian I wouldn't have caused myself :P
13:15 entil .. maybe it's just too old for most people
13:16 Furao for few things ubuntu 12.04 is already too old
13:16 Furao I replace packages as they contains few bugs, mostly pip/virtualenv
13:17 entil I offload a lot, maybe even too much, to pip
13:17 Furao the problem with pip is that it don't retry connection if pypi fail
13:18 dcrouch joined #salt
13:18 entil or github..
13:18 Furao and yes github/bitbucket
13:18 Furao it's the major source of failure in my testing tool
13:18 entil maybe you could do a wrapper with return code evaluation
13:18 Furao I just asked one of my guy to build a mirror of everything we need
13:18 entil but there's a lot of crap out there that exits 0 on clear errors
13:19 Furao in fact, there is a bug in salt that cause salt.modules.pip to return success: True if pip fail
13:19 Furao in those cases when pip fail with a stacktrace
13:19 entil we had our own pypi and I half considered mirroring all our requirements there, but it's a bother to uphold the packages so I decided on the opposite route and do everything in github
13:19 lorimar left #salt
13:19 lorimar joined #salt
13:20 Furao one of my client manage 300 servers in cell phone stores that don't have internet access
13:20 Furao by setting up our own mirror we solve this problem as well
13:20 Furao damn you vodafone with those IIS proxy
13:21 entil I heard a lot about .my censorship back in the day
13:21 lorimar hi all. I'm just getting started with salt and I'm using it to install some RPMs that are stored locally on the salt server (don't exist in any repos). Is it possible to have salt install a package with a wildcard in the name? Tried this without any luck: http://pastebin.com/EkxNwm8P
13:21 Furao ah I had been in .my for 1 week :P
13:22 Furao but yeah look like Najib want to control internet, he's learning from obama
13:22 entil lorimar: didn't try it but as it's python it probably fails
13:22 entil lorimar: and never did that empty sources: either
13:22 entil furao:
13:22 entil asdf
13:23 Furao ?
13:23 lorimar entil: was going based off the sources section of this: http://docs.saltstack.com/ref/states/all/salt.states.pkg.html
13:23 entil furao: there used to be a guy online who's mom did university courses, she couldn't get papers that were published in the universities of essex and sussex
13:23 entil maybe it's obama who's doing the learning...
13:24 entil lorimar: sure, but you need to expand the sources, now it's like empty
13:24 Furao I did not seen any trace of censorship so far
13:25 lorimar entil: not sure I follow you. Isn't that what I'm doing on the next line down (though I should have indented another 2 spaces)
13:25 Furao I watched the news few days ago and some intelectual was arrested because he said something against "the prophet" on Facebook.
13:26 entil lorimar: makes a lot more sense when you point out the indentation, then it's better. but I'm quite sure that evaluates to a literal * and you're screwed
13:26 lorimar entil: ok, is there any way to treat that as a wildcard then?
13:26 entil lorimar: I guess one is expected to know the files and versions. remember that if you repeat them you can set them as variables
13:26 entil lorimar: I dunno, this is my second week or so doing salt ;)
13:26 lorimar lol, gotcha
13:27 entil furao: oh whoah. and also, I hear muslims aren't allowed to go into bars to drink soda and juice, because it's a bar. regardless of that, scantily-veiled women serve haraam food during the ramadan to foreigners
13:27 Furao it seem to be very relaxed on this here
13:28 Furao yes I hear the call to prayers 8 times a day starting at 5h30 am
13:28 Furao there is a mosque on each side of my condo tower
13:28 entil like an alarm clock you can't snooze
13:29 entil do they pray during the day? we had this maroccoan girl with nokia, and I asked ("if it's not weird or personal") and she just laughed me off and told me she saves the prayers up and does them all at home after work
13:30 entil that was so cool
13:31 Furao the island i'm on is almost 50% chinese and they don't go to mosque
13:31 racooper joined #salt
13:31 Furao alcohol is very expensive
13:32 Furao in supermarket there is a separate section for "non-hallal" products
13:32 Furao which are mostly chinese sausage, alcohol and bacon
13:32 Teknix joined #salt
13:32 entil where are you from originally if you're not from .my?
13:32 Furao i'm a french-canadian
13:32 entil cool
13:32 Furao I left canada 2 years ago with my wife traveling and working around
13:32 entil sounds like an interesting thing to do
13:33 jeddi joined #salt
13:33 Furao yes :) did not felt winter for the past 30 months :P
13:34 entil the summer in finland's been insanely hot, though now it started monsooning, rains like crazy
13:34 teskew joined #salt
13:34 entil very atypical, even the winters; we've never had this many winters this cold in a row before
13:36 Furao I spent the last winter in Hong Kong, we've been able to swim in the sea starting march. and the year before we were in Santiago, Chile and it was a blue sky no cloud and 30 celcius in the middle of the day for 4-5 months
13:36 Furao it rained few minutes in the 6 months we've been there
13:36 teskew joined #salt
13:37 entil sick
13:37 Furao it's a big difference with -40 celcius in january in Quebec :)
13:38 entil that makes us appreciate the gulf stream
13:38 Furao there was snow this year at the end of May
13:38 entil .fi was somewhere around -30 iirc, but that's real bad
13:38 jescalan joined #salt
13:38 Furao when I was a kid I remember deliver newspapers to my neighbour at -50 celcius with wind
13:38 entil yeah, our snow lingered for quite a while too, then the temps got hto
13:39 entil *hot
13:39 entil dunno why they say .fi is cold, it's nowhere as bad as what you're describing. or siberia.
13:40 Furao yes you're surrounded by the sea
13:40 Furao .ca and .su are vast and winter is hard in the middle of the land
13:43 entil yeah
13:43 Katafalkas joined #salt
13:44 Furao http://en.wikipedia.org/wiki/Penang#Climate
13:44 Furao except rainfall, it's the same weather all year long
13:44 entil if I'm not entirely mistaken the online .my dude was from penang as well
13:44 lorimar ok, new question. Not going to bother with wildcards, instead I'm just removing the version number from the rpm filename. Here is my new  state: http://pastebin.com/aecAzZLY
13:45 entil lorimar: is that broken somehow?
13:45 lorimar entil: getting the following error on my minion when i run it: http://pastebin.com/NXcFigwt
13:46 entil lorimar: haven't seen that myself :(
13:47 lorimar nevermind. apparently the "name' needs to match the package filename itself
13:47 lorimar got it working
13:48 entil ah, cool
13:49 entil so it's check_mk-agent: salt://... right?
13:53 sarkis joined #salt
13:55 lorimar entil: yep. check_mk-agent: salt://check_mk/check_mk-agent.rpm
13:57 entil got to remember that
13:57 entil I think I've accidentally used the same names
13:57 lorimar is there not currently a state for controlling iptables?
13:57 lorimar https://github.com/saltstack/salt/issues/3597
14:01 entil apparently not
14:01 Furao lorimar: I use something like https://github.com/bclermont/states/tree/master/states/firewall
14:02 Furao but this is old, my new version is a little better. it's very hard to handle iptables in a flexible way to all use cases.
14:02 Furao i think it's better to use iptables-save|restore and jinja template
14:02 lorimar Furao: so this adds an iptables state to salt?
14:03 clintberry joined #salt
14:03 Furao it configure iptables on a host, it's not a state by itself (in the sense of _states/ or salt.states.$something)
14:06 lorimar Furao: gotcha. I think for now I'll probably just have a file with the desired state of the iptables for each of my minion roles (web, ldap, mail, etc) on my salt master and just serve them out from there
14:07 maspwr joined #salt
14:08 alexl joined #salt
14:09 aberant joined #salt
14:14 oz_akan joined #salt
14:15 alexl left #salt
14:15 krissaxton joined #salt
14:17 aat joined #salt
14:20 karlp is it expected that I can do require: -user: {{variable}}_user  (which has a -name bit) but I can't do - runas: {{variable}}_user ?
14:20 karlp runas seems to want the real username, but requires works with the logical name in the sls file?
14:21 DredTiger joined #salt
14:22 mannyt joined #salt
14:24 Furao yes it need a real username
14:26 LyndsySimon joined #salt
14:27 kho joined #salt
14:27 kho joined #salt
14:28 Teknix joined #salt
14:29 aranhoide joined #salt
14:33 Teknix joined #salt
14:33 karlp Furao: well, yes, I see that it does, but why does it require a username, when in all the other places I can refer to the user state blob.
14:33 karlp what's the point of the -name field if I can't use it?
14:34 Furao there is a difference as state name and user name
14:34 Furao as -> between
14:35 abe_music joined #salt
14:35 Furao require specify a state name
14:35 karlp I don't like it, but ok.  it feels really inconsistent :|
14:36 Furao it's consistent
14:37 Furao what if you want to create a user called abc, then install package abc, and create virtualenv abc
14:37 Furao you can't always have the state name as the value of the state
14:39 jescalan joined #salt
14:39 aberant joined #salt
14:41 karlp I see what you mean, but the fact that some things take states, and some things don't is where I feel the inconsistency is.
14:41 multani joined #salt
14:42 multani is there a way to configure the default output of state.highstate so that states which aren't changed are displayed as terse whereas the modified ones are shown as verbose?
14:43 andrewclegg joined #salt
14:46 milind joined #salt
14:52 Newt[cz] joined #salt
14:55 Gifflen_ joined #salt
14:56 carmony good morning. I'm getting a pkg.latest failing for apache2, but I Can't figure out why it is failing
14:56 carmony any idea on debugging it?
14:57 Gifflen_ joined #salt
14:57 multani carmony: did you check your minion's logs?
14:58 carmony well, this is using vagrant-salt, so I called "salt-call -l debug state.highstate" directly
14:58 carmony and I don't see any obvious error
14:58 nliadm will a batch group hang forever if one minion doesn't respond? that seems to be the behaviour I'm seeing
15:00 carmony infact in the logs I see "apache2 is already the newest version."
15:00 carmony so I don't know why that state is failing
15:04 mikedawson joined #salt
15:06 krissaxton joined #salt
15:13 billyw joined #salt
15:19 faeroe joined #salt
15:22 jalbretsen joined #salt
15:22 Bastion2202 joined #salt
15:22 maspwr joined #salt
15:24 clintberry joined #salt
15:24 Bastion2202 hey guys, with becoming part of our core infra, we need to monitor it using nagios. because nagios is not our salt-master I cannot use the nagios server to run test.ping for example. Is there a way on the minion to test if salt is running correctly and perhaps is connected to the master ?
15:24 brettlangdon left #salt
15:24 Furao_ joined #salt
15:29 jschadlick joined #salt
15:29 jacksontj joined #salt
15:30 dthom91 joined #salt
15:31 multani Bastion2202: you can use salt-call test.ping from the minion to ping the master
15:35 Bastion2202 multani: it seems that it tests if the master is up. I am trying to test if the minion is up and running.
15:35 Oct Am I mistaken in believing that state.highstate requires the master to have direct TCP access to all minions ?
15:38 dthom91 left #salt
15:38 Oct bejer: what about: salt '*' test.ping ?
15:39 Oct oups
15:39 Oct Bastion2202: ^
15:39 bemehow joined #salt
15:40 Bastion2202 Oct: the prob with that you will only get answer from the minion that are up. if salt-minion is down on a host you would not notice
15:41 EugeneKay No ports are required to be open on minions; they leave a connection open to the master.
15:41 omame- joined #salt
15:41 EugeneKay Oct - http://docs.saltstack.com/ref/topology.html
15:41 bemehow_ joined #salt
15:43 Furao Bastion2202: t'as juste a faire un check du proc du master, des ports et oui tu peux faire un check nrpe pour qu'il liste tout les minions via l'api de salt
15:43 racooper could you use a nagios service check on the minion to see if the process is running?
15:43 milind joined #salt
15:44 Furao https://github.com/bclermont/states/tree/master/states/salt/master
15:44 Furao https://github.com/bclermont/states/blob/master/states/salt/master/nrpe.jinja2
15:44 felixhummel joined #salt
15:44 Furao these are old, today I got more checks
15:45 Furao and I use salt mine to auto detect infra topology and which check to perform on which host
15:45 afx_hero joined #salt
15:45 afx_hero left #salt
15:45 Newt[cz] joined #salt
15:46 Bastion2202 I already created a service check but it does not mean that it salt-minion works. I've seen situation where minion is up but not accessible. that is why I am looking for something like a test.ping on the minion to himself for exampe
15:46 Bastion2202 Furao:  tx for the link and the comment I will look at them
15:47 Furao you can run the check on the minion instead of the master
15:48 LarsN if you run salt-call test.ping on the minion, it's doing virtually exactly the same thing that would happen if the salt master ran:  salt "$thathost" test.ping
15:48 LarsN in both cases the minion executes python, and talks to the master.
15:48 KFDM Bastion2202: if you use the peer interface, you could run something like "salt-call publish.publish '*' test.ping"
15:49 KFDM So your salt minion, would tell the master to run "salt '*' test.ping" against all minions
15:49 Bastion2202 LarsN: Tx :) in fact salt-call will return true even if salt-minion service is down :)
15:49 zwevans joined #salt
15:50 Bastion2202 Tx for all the feedback guys
15:50 KFDM Ah true. I guess salt-call doesn't always return a "sane" status code :/
15:51 KFDM If you don't mind writing a bit of Python glue code, you could use the Python api to do a similar thing as above
15:51 KFDM https://salt.readthedocs.org/en/v0.12.1/ref/python-api.html
15:51 KFDM Then you would have more control over the return code
15:51 KFDM https://salt.readthedocs.org/en/latest/ref/python-api.html
15:51 KFDM Guess my search  shortcut picked up the older docs. Woops
15:52 LarsN It's an ugly hack, but you could always use "monit" on the minions to keep salt-minion in a running state.
15:53 UtahDave joined #salt
15:54 LarsN I blame UtahDave for my presence here.  :)
15:54 LarsN Morning Dave :)
15:54 nliadm I'm getting an error running a frozen v0.15.3: an ImportError for esky
15:54 nliadm anyone else seen/dealt with this?
15:54 LarsN vim:
15:54 LarsN pkg:
15:54 LarsN - installed
15:54 LarsN sorry....
15:54 ageron joined #salt
15:56 ageron Hi everyone.  Apparently salt does not run on python 3 yet.  Is it in the roadmap somewhere?  (this is probably a frequent question but I couldn't find the answer on google).
15:57 LarsN Bastion2202, KFDM: doesn't the salt-master log any minions that don't respond to a job?  Couldn't you parse that log to get a list of minions that aren't in a sane state?
16:00 KFDM It may. I have not studied the job log very much :)
16:01 LarsN I just remember UtahDave saying something similar to that in Training a few weeks ago.
16:01 LarsN personally I like my ugly hack.  Monit is a stupid solution to dumb problems :)
16:01 aranhoide joined #salt
16:02 Bastion2202 LarsN: I just test with on minion down, nothing in the log.
16:03 Bastion2202 LarsN: and then I will need a nagios check for monit on servers ;)
16:03 jacksontj joined #salt
16:03 LarsN Bastion2202: it's a vicious cycle isn't it. :)
16:05 Oct EugeneKay: ok, thank you, it seems to be working, in fact, my bad :
16:05 Oct :)
16:06 Bastion2202 LarsN: indeed.
16:17 baniir joined #salt
16:18 nliadm UtahDave: ever seen something like this? it doesn't say where's it's actually encountering this problem...
16:18 KyleG joined #salt
16:18 nliadm http://pastebin.com/A2Hkq008
16:29 Teknix joined #salt
16:30 moondawg joined #salt
16:31 UtahDave nliadm: how did you build this system you got this error on?
16:31 moondawg from the top file, how can I create a list from pillar data and match the minion based on that list. My pillar data is a 2 deep array.
16:32 jacksontj joined #salt
16:33 UtahDave moondawg: - match: pillar
16:34 nliadm UtahDave: using this: https://gist.github.com/hdonnay/5700506  I can try stripping out bits I'm including to make modules work and see if that helps
16:34 dlam joined #salt
16:35 UtahDave nliadm: oh,   pip install bbfreeze and pip install esky
16:35 nliadm yeah, doing that alread
16:35 nliadm y
16:35 nliadm line 37
16:37 UtahDave can you import esky from a python prompt?
16:37 moondawg @UtahDave: this is what I have. pillar['uat0']['web']: - match: pillar   (YAML format of course).  No errors, the minion just falls through.
16:37 Oct is there an idiomatic way to cache data inside a module: for example, to run a command once on the minion and be able to reuse its output later ?
16:38 UtahDave moondawg: try 'uat0:web':\n  - match: pillar
16:38 nliadm UtahDave: yes
16:39 UtahDave Oct you can store arbitrary data using the data module
16:39 UtahDave nliadm: on your minion?
16:40 nliadm inside the virtualenv the freeze is built with
16:40 Oct UtahDave: ok thanks. I'd prefer something more transient, but this can do the trick
16:40 bhosmer_ joined #salt
16:41 cbloss_ joined #salt
16:41 UtahDave nliadm: I'm not sure, you'll have to ask the guy who wrote that script.
16:41 nliadm I did
16:41 nliadm I wrote that script
16:41 UtahDave I use esky all the time for the windows stuff and haven't seen that.
16:41 UtahDave :)
16:41 UtahDave ha!
16:41 nliadm it's the steps I was taking before with the 0.14 series
16:41 nliadm and took with some modifications for 0.15.1
16:43 UtahDave I was thinking that maybe you're running into a virtualenv module bug
16:44 UtahDave but you're obviously not using Salt's virtualenv module here
16:44 nliadm hm
16:45 UtahDave so are you sure you're running the salt built from within that virtualenv?
16:45 oz_akan joined #salt
16:46 nliadm UtahDave: hm, okay I'll make everything as verbose as possible and report back
16:47 lex joined #salt
16:48 faeroe joined #salt
16:48 UtahDave cool
16:49 aranhoide joined #salt
16:49 moondawg @UtahDave: yeah that did not work. Here is what I have.     http://pastebin.com/aNwXjsGH
16:51 UtahDave ok, here's your problem, moondawg.  salt-minion1 doesn't have a pillar entry
16:51 jarnet313 joined #salt
16:52 AviMarcus joined #salt
16:52 UtahDave the way you have your pillar set up, only a minion with the id of  'uat0'  will have any pillar data
16:52 moondawg @UtahDave: ah ha...got it
16:52 ZenoTasedro joined #salt
16:53 jacksontj joined #salt
16:54 ageron It took me ages to get this to work.  Can you tell what I did wrong?  Hint: yaml…
16:54 ageron salt 'config' cmd.run 'echo "I love $SW"' 'env={SW: salt}'
16:55 moondawg @UtahDave: I forgot to include the top.sls file in the paste bin for the pillar. http://pastebin.com/EXV6xdmz
16:57 UtahDave moondawg: so in your /srv/salt/top.sls  do you intend for  'uat0' to be an environment?
16:57 moondawg @UtahDave: yes
17:00 UtahDave moondawg: ok, in your /srv/salt/top.sls try this to test the matching
17:00 UtahDave change the pillar match to just be a regular glob match and match on 'salt-minion1'
17:00 UtahDave so remove the -match: pillar   line
17:01 UtahDave and just have  'salt-minion1':
17:01 UtahDave I want to see if your minion will match that way in that environment
17:01 bemehow joined #salt
17:01 moondawg @UtahDave: matched and returned okay
17:02 Teknix joined #salt
17:02 UtahDave ok, and pastebin the output of  salt 'salt-minion1' pillar.data
17:03 moondawg http://pastebin.com/aBn4WTvj
17:04 ZenoTasedro is there a way to get salt-cloud to skip cert verification for openstack?
17:04 UtahDave trying pillar matching on   'uat0:web:salt-minion1':
17:05 UtahDave ZenoTasedro: I'm being told that no, there isn't a way to do that right now.
17:06 UtahDave ZenoTasedro: redbeard2 would love it if you'd open an issue on that.
17:06 ZenoTasedro hacks shall be made! this is unacceptable! :)
17:06 ZenoTasedro i blame libcloud
17:06 moondawg @UtahDave: that worked, but I need to dynamically match
17:06 ZenoTasedro an issue will be made
17:07 moondawg @UtahDave: I did this and it still worked. 'uat0:web:*' could this be it. I am only testing with one minion right now.
17:10 ZenoTasedro @UtahDave, issue made, thank you!
17:10 faeroe_ joined #salt
17:11 jkleckner joined #salt
17:13 UtahDave thanks, ZenoTasedro!
17:13 UtahDave ok
17:17 Kamal_ oooh wow
17:17 Kamal_ I'm so derp
17:18 mikedawson joined #salt
17:20 mephx regarding multiple environments, having a setup of /var/lib/salt/states/ and /var/lib/salt/pillar/ can modules also have a different root? I wanted to split the git repos for states and modules/statemodules
17:20 ZenoTasedro UtahDave: if i recall correctly you said salt-cloud is moving away from libcloud right?
17:20 ZenoTasedro for speed and such
17:20 mephx like a module_roots for master conf or something
17:21 faeroe1 joined #salt
17:21 UtahDave ZenoTasedro: We're generally moving away from it. It's still in use, though
17:21 krissaxton joined #salt
17:22 UtahDave mephx: I'm not quite understanding what you're trying to do.
17:22 ZenoTasedro yeah i see that it's shelling out to curl to perform it's rest operations
17:22 ZenoTasedro so probably a good move in the long run
17:22 ZenoTasedro :p
17:23 ZenoTasedro but i think i found a config for libcloud that'd allow invalid certs
17:23 mephx UtahDave: _modules and _states /(for modules and state modules) need to be in the same folder as state files top etc?
17:23 ZenoTasedro will add my info to the issue
17:23 dhgbrg joined #salt
17:23 mephx UtahDave: it would make managing environments through sepperate git repos
17:24 mephx ppa*
17:24 oz_akan joined #salt
17:24 billyw With remote execution, is there a way to define groups of servers under a single name? i.e. "voip1", "voip2", and "telco1" servers are all referenced under "phone"
17:25 Teknix joined #salt
17:26 UtahDave billyw: you could do that with nodegroups
17:27 billyw UtahDave: Ah, haven't learned about that yet... I'll look for it in the docs. Thanks!
17:27 faeroe_ joined #salt
17:27 UtahDave billyw: http://docs.saltstack.com/topics/targeting/nodegroups.html
17:29 UtahDave mephx: have you looked at gitfs?  gitfs maps each branch and tag to environments
17:29 ageron salt 'config' cmd.run 'echo "I love $SW"' '"env={SW: salt}"'
17:29 ageron Without the extra quotes, the last argument gets parsed into a dict { "env={SW"  :  "salt}" }
17:29 UtahDave Also, you might look at your file_roots option: http://docs.saltstack.com/ref/configuration/master.html#file-roots
17:29 UtahDave ageron: Yeah, passing in arguments like that to cmd.run can be tricky
17:30 mephx UtahDave: yes, I moved away from it sicne from my understading I will have to use hooks to hae the same behaviour with pillars (we have different pillar envs too)
17:30 Ryan_Lane joined #salt
17:30 UtahDave yeah, gitfs doesn't deal will pillar
17:30 mephx so i decided to do it all by hooking and have total control of it (whereas I feel gitfs can get out of control easily)
17:31 kvbik joined #salt
17:31 ioni joined #salt
17:31 dthom91 joined #salt
17:31 Ixan joined #salt
17:31 loz_hurst joined #salt
17:31 logix812 joined #salt
17:32 UtahDave mephx: ok, so then I'd set up your file_roots like in that example and manage those git repos however you'd like
17:32 mephx UtahDave: still, _modules and _states can't really have a different path from the state files right?
17:33 mephx I just wanted to confirm this
17:33 dcolish joined #salt
17:33 mephx the docs are really clear regarding this, and it is not possible
17:33 UtahDave _modules and _states have to be in the root of your file_roots
17:34 mephx ok, thanks
17:38 aberant joined #salt
17:42 kermit joined #salt
17:46 nliadm UtahDave: installing things as eggs fixed it
17:46 UtahDave ok, interesting, nliadm
17:47 nliadm apparently the pip that got installed in the venv didn't have the '--egg' option, either
17:48 Sypher joined #salt
17:49 aranhoide joined #salt
17:57 dhgbrg joined #salt
18:00 Kamal_ Is it possible to have a state that makes sure a file has a specific line in it and if not, insert it?
18:00 xt yes!
18:00 Kamal_ :-o
18:00 Kamal_ educate me
18:02 Kamal_ hmm I see it's possible using cmd.run but I think I'm overengineering this
18:04 krissaxton joined #salt
18:04 Shell Kamal_: the file.append state should do it
18:04 Kamal_ hah how did i miss that
18:04 Kamal_ thanks Shell
18:12 jeddi This might be more of a #vim question, but do you guys have any configs ready to go for ensuring the various yaml files open with the right syntax?  obviously *.sls is easy, but i see also (as examples) files called 'master', 'minion', and some *.conf files
18:14 akoumjian What is everyone's preferred way of setting environment variables? Would I throw it in .profile?
18:14 jeddi akoumjian: for a given user only?
18:14 UtahDave jeddi: have you looked at salt-vim?
18:14 jeddi UtahDave: ooh, i have not.
18:15 akoumjian jeddi: Let's say for root, but global is fine too
18:15 jeddi UtahDave: actually, i have, but had forgotten it :)   yeah - i've got that configured already .. but it doesn't pick up files that aren't called .sls, right?
18:15 ZenoTasedro UtahDave: I put in a pull request for that openstack ssl thing
18:16 oz_akan joined #salt
18:16 UtahDave jeddi: I'm not sure, actually
18:16 UtahDave ZenoTasedro: awesome, thanks!
18:16 ZenoTasedro UtahDave: no problem, felt better than waiting :p
18:16 ZenoTasedro and i am already up and running as a result
18:16 jeddi UtahDave: yeah - just checked that out and confirmed . it's based on .sls extension only.
18:17 jeddi akoumjian: for global, /etc/profile.d/{file} fragments for me.
18:17 UtahDave :)  We do our best to get things done, but help from our awesome community is REALLY appreciated.
18:19 Shell is there not a special comment you can put in a file to set the syntax? there's ones for tabstop, etc...
18:19 jeddi Shell: yeah, this rings a bell ..
18:19 jeddi Shell: it'll break for deploys where i'm using salt-cloud and relying on the default /etc/salt/minion file (as one example).  but i can sort this out i'm sure from master later.
18:19 Shell a google doesn't turn up anything for me, but maybe if you asked #vim they might be able to help.
18:20 jeddi plus, i really shouldn't be editing these files on minions.  just during debugging it's nice to hop onto boxes and see what's going on .. and not be visually offended by bad syntax highlighting. :)
18:20 bemehow joined #salt
18:20 jschadlick joined #salt
18:22 mahimahi joined #salt
18:23 Teknix joined #salt
18:27 auser joined #salt
18:28 auser hey all
18:28 bemehow joined #salt
18:28 krissaxton joined #salt
18:28 dthom91 joined #salt
18:30 bemehow joined #salt
18:32 bemehow_ joined #salt
18:33 bemehow joined #salt
18:36 UtahDave hey, auser!
18:36 auser Hey UtahDave
18:36 auser how goes
18:37 UtahDave pretty good!
18:39 auser from my talk last night: https://speakerdeck.com/auser/reactive-saltstack
18:39 racooper Howdy.  Quick query, is there an example of syntax to use pkg.group_install in a .sls state file?
18:40 UtahDave ooh, awesome, auser.  I'm going to read it right now while eating lunch!
18:40 auser nice
18:42 baniir joined #salt
18:42 whiteinge auser: holy cow. neat stuff
18:42 auser thanks whiteinge
18:42 auser I build my states to be as dynamic/shareable as possible
18:42 jalbretsen UtahDave: in ssh_auth building, does setting the user option imply the user must be present?  Thus using a require statement would be redundant?
18:44 jalbretsen building?  I'm on crack.  Referring to the "BUILT IN" state
18:49 Katafalkas joined #salt
18:49 ronc joined #salt
18:59 UtahDave jalbretsen: no, normally Salt doesn't assume anything.  You most likely will need a require there.
19:00 UtahDave racooper: you probably just want to use pkg.installed with the - pkgs argument
19:00 racooper the idea is to do the equivalent of "yum groupinstall base"
19:01 UtahDave Hm. terminalmage might know.
19:02 terminalmage my ears are burning
19:02 terminalmage what's up?
19:02 terminalmage I think there is a group install feature there already
19:02 terminalmage checking
19:02 racooper Hi, the question was looking for examples of the syntax of using pkg.group_install in a sls statefile
19:03 KyleG joined #salt
19:03 terminalmage yeah, I know there is a pkg.group_install function... but state support might not be implemented
19:04 racooper that might explain why I can't find anything through googlefu....
19:04 terminalmage yeah I don't think there is state support for this
19:04 terminalmage are you on github?
19:05 racooper ah, apparently I do have an acct there.
19:05 terminalmage if so, please file an issue and we will try to get that feature added for 0.17.0. I'm not sure we'll get it in time for 0.16.0
19:05 terminalmage we're in bugfix mode right now
19:06 racooper no problem. Thanks for the answer!
19:08 dave_den joined #salt
19:09 ageron joined #salt
19:09 danielbachhuber joined #salt
19:11 Ahlee anybody get grain_pcre working?
19:14 ageron What's the best practice for minion roles?  Setup roles in minion's config and access that info through grains ?  Or setup static groups in the master's config ?
19:16 auser ageron: I set it with grains
19:16 auser sec
19:17 auser https://speakerdeck.com/auser/reactive-saltstack look for grains.template (not sure what slide it is)
19:17 ageron Yes, I just when through your slides, and that's why I'm asking ! :-)
19:18 auser ah
19:18 ageron I see roles:hadoop_slave, match: grain
19:18 auser I set those roles with grains.setval roles hadoop_slave,zookeeper etc
19:18 melinath joined #salt
19:19 ageron auser: I haven't tried roles with grains, but I was fearing 1) tough to update (is it?)  and 2) security hole if a hacker gains access to just one minion, he can set the group to admin and perhaps gain access to sensitive data?
19:19 ageron s/set the group/set the roles/
19:21 jschadlick I have been trying to implement roles through pillar. Do you recommend grains instead?
19:23 ageron I suppose, grains make it easy to spawn an instance with the appropriate role(s) configured inside it.  Whereas with static groups (and pillars as well), you would have to 1) update the master config (static groups or pillars) 2) spawn the instance
19:24 auser joined #salt
19:24 auser gr, darn internet
19:26 waverider joined #salt
19:26 atn34 joined #salt
19:27 jescalan joined #salt
19:29 FreeSpencer joined #salt
19:29 bemehow joined #salt
19:31 dcrouch joined #salt
19:32 melinath joined #salt
19:32 aranhoide I get the following when trying to apply a salt-cloud map containing three instances:
19:33 aranhoide http://pastebin.com/6GrJJG1A
19:33 aranhoide one of the instances already exists when I run salt-cloud -m map
19:34 aranhoide but not the one salt-cloud is complaining about
19:34 aranhoide (this is 0.8.8)
19:38 m_george left #salt
19:39 aranhoide OK, doesn't happen in github HEAD..
19:43 ZenoTasedro is there a way to target boxes with ext_pillar?
19:44 atn34 I am running salt in a virtual environment, with the root of the virtual environment set as root dir in the master config. I am running salt-master and salt-minion daemonized and on the same machine. I tried salt '*' test.ping, and it says it cannot access /var/cache/salt/master. It should be trying to access $VIRT_ENV/var/cache/salt/master. Any ideas?
19:45 p3rror joined #salt
19:45 UtahDave have you set the correct paths in your master config?
19:46 mgw joined #salt
19:46 atn34 root_dir is set to $VIRT_ENV
19:48 aranhoide does anyone have an ETA for next salt-cloud release?
19:49 aranhoide atn34 are you running that as root?
19:49 aranhoide or under sudo
19:49 chrisgilmerproj1 joined #salt
19:49 aranhoide I mean, maybe it's a permissions problem, not a path problem
19:50 alekibango joined #salt
19:51 atn34 not running as root. salt is not installed for root
19:55 atn34 ahh I need to run salt -c $VIRT_ENV/etc/salt '*' test.ping
19:55 melinath joined #salt
19:55 atn34 need to pass config_dir to that command as well
19:55 atn34 thanks guys
20:00 mgw joined #salt
20:02 nliadm did the scheduler bug get fixed in 0.15.3?
20:03 wunki joined #salt
20:03 wunki I just released a project which (I think) truly shows the power of Saltstack.
20:03 wunki One-click django development environment with Vagrant: https://github.com/wunki/django-salted#omg-you-also-put-the-private-key-on-github
20:03 wunki let me know if you find any improvements :)
20:08 UtahDave nliadm: yeah, I think so. Have you looked at the issue tracker?
20:08 Teknix joined #salt
20:08 UtahDave wow, that looks awesome, wunki!
20:09 wunki UtahDave: thanks! It truly is one-click and your setup. At least at my PC :)
20:09 nliadm I did, but I did a 'config.get schedule' and didn't see what my schedule items
20:09 jschadlick left #salt
20:10 jschadlick joined #salt
20:11 mnemonikk joined #salt
20:13 gordonm joined #salt
20:14 jschadlick So i have a system where the fqdn and roles for hosts is managed in a database. What would be the best way to target hosts by role in the case where roles can change quickly?
20:15 nliadm jschadlick: I ended up writing my own grain for that
20:15 jschadlick I see, that is something we are considering as well.
20:16 jschadlick how do you quickly update your grains?
20:20 gordonm joined #salt
20:21 kcb joined #salt
20:22 gordonm left #salt
20:30 atn34 joined #salt
20:35 DerekRBN joined #salt
20:35 gordonm joined #salt
20:35 bhosmer__ joined #salt
20:38 mgw1 joined #salt
20:43 mikedawson joined #salt
20:44 UtahDave jschadlick: That would probably be best done with an external pillar
20:46 bhosmer_ joined #salt
20:47 ZenoTasedro wunki: nice work there!
20:48 jescalan joined #salt
20:49 jschadlick does the host with the db have to be running salt in order for an external pillar implementation to work?
20:50 ZenoTasedro jschadlick: i'm in the depths of working with external pillar as well, and no
20:51 jschadlick I see
20:51 ZenoTasedro like if you use mongodb for pillar data, that can be on a box that's just running mongo
20:52 ZenoTasedro though i'd have a salt-minion on the mongo box for administration, that's my preference
20:52 nliadm can you use pillar items elsewhere in the pillar?
20:52 ZenoTasedro i'm not sure, i don't think so
20:53 nliadm hm, I think that's the reason I went with a grain
20:53 ZenoTasedro in the end of all things pillar data gets flattened into a data structure like the states do
20:53 ZenoTasedro i like grains for the concept of a role as well
20:54 ZenoTasedro i think both are about equally valid though
20:54 UtahDave jschadlick: nope
20:54 ZenoTasedro one thing to keep in mind though, i believe grains are intended to be static bits of information about a server
20:54 ZenoTasedro though you can set and update and etc of course
20:55 nliadm mine are ldap lookups
20:56 nliadm which I keep meaning to re-write to use the salt ldap module
20:58 ZenoTasedro nliadm: the auth ldap module?
20:58 dthom91 joined #salt
20:58 nliadm ZenoTasedro: there's an ldapcmd module
20:59 ZenoTasedro ah okay
21:00 nliadm but I wrote my own py module thing
21:00 jacksontj joined #salt
21:01 robinw joined #salt
21:06 jescalan joined #salt
21:12 krissaxton joined #salt
21:24 jay for the love of god, I've spent about 5 hours on getting pillar data from cobbler
21:25 jay its starting to make me angry now, I'm no closer than when I started and I can find bugger all meaningful documentation
21:25 jay does ANYONE have a useful link please !
21:25 dthom911 joined #salt
21:26 dthom911 joined #salt
21:30 auser joined #salt
21:33 ZenoTasedro jay: thats a sad story :( were you looking for docs on ext_pillar or that module specifically?
21:34 ZenoTasedro i dunno jack about cobbler
21:34 jay I'll take anything to be honest
21:34 krissaxton joined #salt
21:34 ronc joined #salt
21:34 ZenoTasedro i was trying to use a premade ext_pillar module, but i'm about to start writing my own instead
21:34 jay a decent example of how to pull pillar data from cobbler and or puppet or anything else I can work from
21:35 jay and maybe how to test it .. at this point I'm assuming that sasalt '*' pillar.data will show my classes
21:35 dave_den joined #salt
21:35 ZenoTasedro at my level of experience with external pillar i don't have as much to offer as i'd like to
21:35 ZenoTasedro do you have a means of programatically getting your won data from cobbler/puppet?
21:36 ZenoTasedro your own* data
21:37 ZenoTasedro i don't have any good docs or examples for what you're looking for, but https://github.com/saltstack/salt/blob/develop/salt/pillar/cmd_yaml.py shows an super simple and nearly useless pillar module
21:38 jacksontj joined #salt
21:38 ZenoTasedro https://salt.readthedocs.org/en/latest/topics/development/external_pillars.html#configuration and this page shows the master config section for ext_pillars
21:39 ZenoTasedro and in that example, if 'example_a' was 'cmd_yaml' and the argument was 'cat /tmp/file.yaml', that argument would go into the 1st argument of cmd_yaml's ext_pillar function
21:39 ZenoTasedro in that example it'd apply the yaml as pillar data for any node
21:40 ZenoTasedro but the module can access grain information and the rest of salt to do much more useful things
21:41 ZenoTasedro oh! also, try running your master manually with
21:41 ZenoTasedro '-l debug'
21:41 ZenoTasedro maybe the logs will show you something useful
21:42 ZenoTasedro jay: is that helpful to you?
21:43 ZenoTasedro i suppose not!
21:49 jay joined #salt
21:55 billyw I'm having difficulty getting package management to work properly on a FreeBSD minion
21:55 billyw Running:  salt 'freebsd.box' pkg.install wget   will have the server return an empty response
21:56 billyw And running:  salt 'freebsd.box' freebsdpkg.install wget  will return that no freebsdpkg.install module is available
21:56 UtahDave billyw: yeah, you can't call the freebsd pkg state directly
21:57 UtahDave ZenoTasedro: there's an external pillar for cobbler already
21:59 UtahDave billyw: I'd run the minion in the foreground in debug mode to see if there are any indications as to what's going on
21:59 billyw UtahDave: Thanks, I'll do that, appreciate all your help
22:01 billyw UtahDave: Ahhh, it's because the pkgng repos are still threadbare since the security incident last November
22:02 jay ZenoTasedro: many thanks .. my daft internet just blipped between me and my remote server .. just catching up on what you posted .. looks loads better than where I was at tho
22:02 mannyt joined #salt
22:03 jay my problem is while I maybe a 15year linux sysadmin I am not a python programmer so most of the stuff is nearly useless to me, and I'm not really wanting to learn python for a deployment engine
22:03 jay like .. how the bloody hell would I actually use https://github.com/saltstack/salt/blob/develop/salt/pillar/cmd_yaml.py ... wheres the useful info !
22:04 jay sorry - that sounds a bit ranty .. its not aimed at you .. its just highly frustrating
22:05 mgw joined #salt
22:07 UtahDave jay, any reason you haven't tried the cobbler external pillar?
22:07 UtahDave https://github.com/saltstack/salt/blob/develop/salt/pillar/cobbler.py
22:07 jay I've been trying that _all_ day
22:08 sarkis joined #salt
22:09 UtahDave jay: can you pastebin the relevant part of your master config?
22:09 jay there is loads of cyptic comments everywhere, salt doesnt give a crap if your config is valid just if its syntactically correct yaml
22:09 jay just doing that
22:10 jay http://pastebin.com/dtAa4Y4D
22:11 jay is my last attempts
22:11 jay and I've put that in /etc/salt/master at the bottom
22:12 logix812 joined #salt
22:12 UtahDave ok, just a moment. reading
22:12 jslatts joined #salt
22:13 Newt[cz] joined #salt
22:13 jay thank you for any input, I really appreciate it
22:14 UtahDave OK, so first of all, let's split this all out and do one thing at a time.
22:14 UtahDave Let's get just the cobbler external pillar working
22:15 druonysus joined #salt
22:15 druonysus joined #salt
22:15 jay I *think* thats all I need anyhow
22:16 UtahDave OK, so first of all, I'm assuming you have cobbler running on the same machine as your master
22:16 jay correct
22:16 UtahDave and that it's api is available at http://localhost/cobbler_api
22:16 jay correct
22:16 UtahDave ok, good
22:17 UtahDave So try using this in your master config instead of what you had before
22:17 UtahDave then restart the master
22:17 UtahDave then on the master run:    salt \* pillar.data
22:18 austin987 joined #salt
22:18 ronc joined #salt
22:19 jslatts joined #salt
22:19 aranhoide joined #salt
22:20 aranhoide does anyone know from the top of their head how to get the public IP of an EC2 instance?
22:20 DanGarthwaite joined #salt
22:20 aranhoide it seems not to be in grains by default?
22:20 jay UtahDave: I didnt miss a line there did I .. I assume you're gonna pastebin me something
22:20 UtahDave aranhoide: you have to use the public dns name of your minion.  That's in the grains
22:21 UtahDave jay: Bah, sorry. I did forget to paste. http://pastebin.com/DHb4arwc
22:21 jay LOLOL
22:22 bhosmer_ joined #salt
22:23 aranhoide UtahDave: I can't find that in the grains either
22:23 UtahDave aranhoide: Yeah, I'm looking right now, too, and am not seeing it.  I was pretty sure I had seen it just recently
22:24 UtahDave There's an ec2 custom grains module in the salt-contrib module that might output what you need
22:24 jay okay, that all runs
22:24 aranhoide I know it gets returned to the salt-cloud that built it
22:24 dcrouch joined #salt
22:24 aranhoide or at least the salt-cloud prints it out on completion
22:25 aranhoide UtahDave: OK, I'll look into that
22:25 kmwhite joined #salt
22:26 UtahDave jay: is that all you need?  I can try to help you get the master_tops working, too
22:26 baniir joined #salt
22:26 chrisgilmerproj1 left #salt
22:26 jay erm, but I'm not seeing what I might say is useful ... bear with me 1 sec
22:30 jay http://pastebin.com/Rsxme5q7
22:30 jay am I missing the point here .. surely I should see Class1 somewhere in my pillar data ?
22:30 dthom91 joined #salt
22:31 evxd joined #salt
22:36 UtahDave yeah, I think it should
22:36 UtahDave try running   salt \* saltutil.refresh_pillar
22:37 jay returns BuildSrv02.build.domain.net: \n None
22:38 jay still nothing using the grep command
22:39 UtahDave what's the output of salt \* pillar.data   after running the saltutil.refresh_pillar
22:39 jay you want the whole lot pastebin'd
22:39 aranhoide custom states go in /srv/salt/_states/ by default, right?
22:40 UtahDave aranhoide: yes
22:40 UtahDave jay: yeah, you can pm it to me if you don't want it public
22:41 jay http://pastebin.com/KK28qg1Z
22:41 aranhoide UtahDave: thanks!
22:43 jb___ joined #salt
22:43 UtahDave jay, can you confirm that cobbler and cobbler are valid usernames and passwords?
22:43 jay they are
22:44 horses joined #salt
22:44 UtahDave try adding - only: classes    to your config
22:47 jay no change ... http://pastebin.com/yaBdP8VF
22:48 UtahDave ok, open up a second terminal into your master
22:49 UtahDave stop the salt-master service
22:49 oz_akan joined #salt
22:49 UtahDave and start it in the foreground in debug mode
22:49 UtahDave sudo salt-master -l debug
22:50 UtahDave the run salt \* pillar.data in the original terminal window
22:50 UtahDave I'm looking for any errors or stacktraces to be reported there.
22:50 jay [WARNING ] Authentication failure of type "user" occurred.
22:51 jay on each run
22:51 jay I know categorically cobbler:cobbler is the username and password for cobbler tho, I have logged out and logged back in
22:52 adotbrown joined #salt
22:52 DredTiger joined #salt
22:53 mannyt joined #salt
22:55 UtahDave does that user have rights to use he cobbler api?
22:59 jay I'm thinking so .. I've got [zmlrpc_service_users] admin = cobbler, which is correct as I understand it, I can however go off and poke cobbler until I dont get the authentication failed as I now have something to go on :)
22:59 UtahDave ok, good
22:59 jay thanks very much
22:59 UtahDave I"ll be here about another 20 minutes, then I'll be back on in the morning
23:00 jay I'll report back and let you know either way
23:00 UtahDave cool
23:01 mgw joined #salt
23:02 aranhoide any estimate on when a new salt-cloud pip release comes out?
23:02 adotbrown joined #salt
23:03 aranhoide 0.8.8 doesn't seem to match the map file format in the 0.8.8 docs, while git HEAD does
23:03 aranhoide in 0.8.8 I had to put the grains under the 'minion:' dict for them to be picked up
23:07 UtahDave I don't know exactly when, but I think they were talking about cutting a release in the next week or so
23:07 adotbrow1 joined #salt
23:08 bemehow joined #salt
23:09 bemehow_ joined #salt
23:12 mgw joined #salt
23:16 melinath joined #salt
23:17 godber_ joined #salt
23:19 godber_ hello! with jinja templates of sls modules, is it possible to test for the presence of a config file under salt:// ?
23:20 mannyt joined #salt
23:22 UtahDave godber_: you could probably use {% salt['cp.get_file'](salt://mypath/myfile) %}   and test for a positive response
23:23 dthom91 joined #salt
23:24 godber_ hmmm, would I benefit from reading about other renderers perhaps
23:24 evxd aranhoide: apparently 0.8.9 should be ready in a week or two, just from reading the issues on github
23:24 godber_ that is a good suggestion and I will try it, thanks UtahDave
23:25 UtahDave you're welcome.  Hope it helps you find a good solution
23:25 kermit joined #salt
23:25 evxd aranhoide: https://github.com/saltstack/salt-cloud/issues/628 also related to your grains issue
23:37 jeddi So with salt-cloud - I have to define the provider within the profile - so if I have two or three providers (different accounts on rackspace, in my case), and I want to launch an Ubuntu LTS image in any of the three, then I need to name my profiles something like 'ubuntu_lts_companya', 'ubuntu_lts_companyb' and so on.  is this really the way it's done?
23:38 adotbrown joined #salt
23:40 cxz joined #salt
23:43 mahimahi1 joined #salt
23:45 krissaxton joined #salt
23:46 Gifflen joined #salt
23:47 aat joined #salt
23:54 jschadlick left #salt

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