Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-07-14

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

All times shown according to UTC.

Time Nick Message
00:00 mateoconfeugo joined #salt
00:01 mateoconfeugo joined #salt
00:02 oz_akan_ joined #salt
00:02 mateoconfeugo joined #salt
00:03 mateoconfeugo joined #salt
00:03 mateoconfeugo joined #salt
00:04 mateoconfeugo joined #salt
00:05 mateoconfeugo joined #salt
00:06 mateoconfeugo joined #salt
00:07 sag47 left #salt
00:07 mateoconfeugo joined #salt
00:08 mateoconfeugo joined #salt
00:08 mateoconfeugo joined #salt
00:09 mateoconfeugo joined #salt
00:10 mateoconfeugo joined #salt
00:11 mateoconfeugo joined #salt
00:11 dccc joined #salt
00:12 mateoconfeugo joined #salt
00:12 mateoconfeugo joined #salt
00:14 mateoconfeugo joined #salt
00:15 mateoconfeugo joined #salt
00:16 mateoconfeugo joined #salt
00:17 mateoconfeugo joined #salt
00:17 mateoconfeugo joined #salt
00:19 mateoconfeugo joined #salt
00:20 mateoconfeugo joined #salt
00:20 Shenril joined #salt
00:21 mateoconfeugo joined #salt
00:22 mateoconfeugo joined #salt
00:22 jbirdman joined #salt
00:22 mateoconfeugo joined #salt
00:22 jbirdman left #salt
00:24 mateoconfeugo joined #salt
00:25 mateoconfeugo joined #salt
00:26 mateoconfeugo joined #salt
00:28 cruatta joined #salt
00:45 acabrera joined #salt
00:49 mechanicalduck joined #salt
01:02 oz_akan_ joined #salt
01:08 rojem joined #salt
01:10 jslatts joined #salt
01:11 otter768 joined #salt
01:15 bhosmer joined #salt
01:35 arthabaska joined #salt
02:03 oz_akan_ joined #salt
02:17 mateoconfeugo joined #salt
02:23 mateoconfeugo joined #salt
02:25 TheThing joined #salt
02:26 ajolo joined #salt
02:31 otter768 joined #salt
02:37 imon joined #salt
02:38 imon left #salt
02:54 TyrfingMjolnir joined #salt
02:55 Luke joined #salt
03:01 XenophonF left #salt
03:01 XenophonF joined #salt
03:02 XenophonF any one else run into pkg_resources.DistributionNotFound errors after upgrading to Salt 2014.1.7 on FreeBSD?
03:03 otter768 joined #salt
03:03 bhosmer joined #salt
03:04 oz_akan_ joined #salt
03:05 XenophonF hm, maybe my build process is borked
03:05 XenophonF the package built using poudriere didn't work
03:06 XenophonF but i just rebuilt sysutils/py-salt using portmaster and had no problems
03:07 ramishra joined #salt
03:13 Hipikat joined #salt
03:16 RandalSchwartz haven't up graded salt yet
03:16 RandalSchwartz is there a new one?
03:21 manfred 2014.1.7 will be announced tomorrow morning, but it should be in all distributions now
03:22 XenophonF RandalSchwartz: I'm zeroing out poudriere and rebuilding everything, just in case I messed something up somewhere along the way.
03:22 RandalSchwartz ok
03:22 XenophonF I caught it in testing so no harm done, I guess.  Just surprising.
03:25 anuvrat joined #salt
03:42 yomilk joined #salt
03:44 quickdry21__ joined #salt
03:48 XenophonF this is odd - i get the same pkg_resources.DistributionNotFound error when installing Salt from the FreeBSD pkgng repo
03:48 XenophonF I just build a new 10.0-RELEASE VM
03:49 XenophonF installed Salt post-install using "env ASSUME_ALWAYS_YES=yes pkg install py27-salt"
03:49 XenophonF that bootstrapped pkgng at the same time
03:50 XenophonF but when I run service salt_minion start I get
03:50 XenophonF pkg_resources.DistributionNotFound: salt==2014.1.5
03:51 acabrera joined #salt
04:04 XenophonF left #salt
04:05 oz_akan_ joined #salt
04:18 mosen joined #salt
04:37 terminalmage joined #salt
04:38 arthabaska joined #salt
04:39 Topic for #salt is now Topic for #salt is "Welcome to #salt | 2014.1.5 is the latest | Please be patient when asking questions as we are volunteers and may not have immediate answers | Channel logs are available at http://irclog.perlgeek.de/salt/"
04:39 bhosmer joined #salt
04:39 Topic for #salt is now Welcome to #salt | 2014.1.5 is the latest | Please be patient when asking questions as we are volunteers and may not have immediate answers | Channel logs are available at http://irclog.perlgeek.de/salt/
04:39 m1crofarmer joined #salt
04:40 terminalmage left #salt
04:40 ipalreadytaken joined #salt
04:53 bhosmer joined #salt
04:54 anuvrat joined #salt
04:56 acabrera joined #salt
04:57 ramteid joined #salt
05:05 oz_akan_ joined #salt
05:24 fragamus joined #salt
05:37 yomilk joined #salt
05:38 [HACKING-TWITTER joined #salt
05:41 [HACKING-TWITTER joined #salt
05:47 [HACKING-TWITTER joined #salt
06:03 CeBe joined #salt
06:07 greyhatpython joined #salt
06:22 ipalreadytaken joined #salt
06:26 ramishra joined #salt
06:34 mikkn joined #salt
06:41 bhosmer joined #salt
06:43 yomilk joined #salt
06:45 ipalreadytaken joined #salt
06:46 aquinas joined #salt
06:49 anuvrat joined #salt
06:59 Hell_Fire_ joined #salt
07:07 oz_akan_ joined #salt
07:13 anuvrat joined #salt
07:16 Kenzor joined #salt
07:19 linjan joined #salt
07:20 alanpearce joined #salt
07:25 guojun04 joined #salt
07:26 cheftestn joined #salt
07:28 jhauser joined #salt
07:30 cheftestn joined #salt
07:33 felskrone joined #salt
07:34 martoss joined #salt
07:36 slav0nic joined #salt
07:45 ixokai joined #salt
07:47 martoss1 joined #salt
07:51 chiui joined #salt
08:08 oz_akan_ joined #salt
08:12 Nexpro joined #salt
08:13 TyrfingMjolnir I have a self signed cert that I need to make pre approved for my terminal server users, how to do this in W2K3 Outlook Express, W2K12 Live Mail?
08:22 xsteadfastx anyone got cp.get_file working with a windows minion?
08:22 xsteadfastx no matter what destination i get a "permission denied"
08:23 [HACKING-TWITTER joined #salt
08:26 [HACKING-TWITTER joined #salt
08:27 matthias_ joined #salt
08:30 [HACKING-TWITTER joined #salt
08:30 bhosmer joined #salt
08:41 ml_1 joined #salt
08:41 aberdine joined #salt
08:46 giantlock joined #salt
08:46 ggoZ joined #salt
08:50 TheThing joined #salt
08:51 Shenril joined #salt
08:52 LordOfLA|Broken joined #salt
08:56 anuvrat joined #salt
09:03 gmoro joined #salt
09:05 mike25de joined #salt
09:06 ramishra joined #salt
09:06 intellix joined #salt
09:08 oz_akan_ joined #salt
09:12 TyrfingMjolnir I have a self signed cert that I need to make pre approved for my terminal server users, how to do this in W2K3 Outlook Express, W2K12 Live Mail? Does this have to be done pr user account? Or can this self signed cert be pre approved for all users on this machine?
09:18 damianball joined #salt
09:21 Dattas joined #salt
09:27 yomilk joined #salt
09:29 BbT0n_ joined #salt
09:29 nebuchad` joined #salt
09:30 jhauser_ joined #salt
09:31 sirtaj_ joined #salt
09:32 analogbyte joined #salt
09:32 mosen joined #salt
09:33 canci joined #salt
09:33 sectionme joined #salt
09:34 majoh joined #salt
09:34 sindreij joined #salt
09:34 MK_FG joined #salt
09:34 alanpearce joined #salt
09:35 sectionme Has anyone experianced file.recurse working intermittently? Seems I get random failues with "Comment: The directory 'salt://<REDACTED>' does not exist on the salt fileserver in saltenv 'base'".
09:37 crane joined #salt
09:45 ramishra joined #salt
09:49 mariusv joined #salt
09:56 TheThing joined #salt
10:06 ze- joined #salt
10:17 babilen When I write an execution module which keys do have to present in the dictionary that I return from the function that I call?
10:18 babilen Or what can I return there? I've seen modules simply return a string or something, but what are the expectations there?
10:19 bhosmer joined #salt
10:20 ramishra joined #salt
10:33 giannello joined #salt
10:34 TheThing joined #salt
10:49 picker joined #salt
10:51 intellix joined #salt
11:03 ilako joined #salt
11:10 oz_akan_ joined #salt
11:14 netzmonster joined #salt
11:17 Dutchsmoker_ joined #salt
11:17 Dutchsmoker_ anyone around?
11:24 the_drow joined #salt
11:26 q4brk joined #salt
11:26 q4brk joined #salt
11:31 the_drow It's been a couple of days and I still got no reply for https://github.com/saltstack/salt/issues/14175 I'd really like to know what do you guys think
11:31 picker joined #salt
11:34 XenophonF joined #salt
11:34 dzen mh
11:35 dzen using asyncio would require a huge amount of job
11:35 intellix joined #salt
11:35 dzen first, salt would be python 3.3 only
11:35 dzen or > 3.3
11:36 the_drow dzen: Which is bad because?
11:37 mechanicalduck_ joined #salt
11:37 the_drow Most (if not all) modern operating systems have Python 3 installed
11:39 Dutchsmoker_ left #salt
11:40 dzen oh, I would like to have python 3.3
11:40 dzen redhat does not provide it
11:40 tkharju3 joined #salt
11:41 dzen debian stable is sticke to 3.2
11:41 dzen and there is no backport available
11:42 dzen *sticked to*
11:42 the_drow Really?
11:42 jrdx joined #salt
11:42 dzen I'm using asyncio in production, and this part is a nightmare ;)
11:43 the_drow We can still use something like Trollius although the aio0mq library does not support it
11:43 the_drow dzen: use pyenv :)
11:43 the_drow I contributed the pyenv module and state
11:43 dzen we do not use venv in production
11:43 the_drow dzen: No need for venv
11:43 dzen we package them
11:43 dzen the_drow: or similar ;)
11:43 the_drow pyenv is what downloads and compiles python for you
11:43 dzen whoah
11:44 the_drow dzen: you mean with pants?
11:44 dzen compiling python in the whole platform
11:44 dzen on 400 machines, no seriously :)
11:44 the_drow dzen: You can compile once and package it into a binary with pants
11:45 dzen we're deploying applications using debian packages
11:45 the_drow dzen: Yes, and that's how you should do it
11:45 the_drow But the application itself can be a self contained binary
11:45 dzen I'm not a big fan of having multiple installation procedures
11:46 the_drow http://pantsbuild.github.io/python-readme.html
11:46 the_drow The interpeter is embedded inside the executable with the code
11:47 the_drow dzen: if you are deploying web applications then uwsgi can do the same for you
11:47 dzen it's interesting
11:47 dzen but it's the same problem in having two different installation procedures
11:48 dzen And imho it's hard to maintain
11:48 the_drow dzen: How is it two different installation procedures?
11:48 dzen mh, not really two différent installation procedures but
11:48 the_drow dpkg/apt-get installs the binary to where ever it should and also installs the dependencies
11:49 dzen when the project have to be released, there should be a pex, and a debian|whateverdistro package
11:49 dzen and, pants have to be available in the distro
11:49 dzen and probably at the right version
11:50 the_drow dzen: No. pants should be available on the build server
11:50 dzen ah yes
11:50 mechanicalduck joined #salt
11:50 the_drow dzen: http://uwsgi-docs.readthedocs.org/en/latest/Embed.html - the uwsgi way to do it
11:50 dzen I misread
11:51 the_drow dzen: It's the same process as compiling an application
11:51 dzen so, for the moment, all the distro are not helping the python community by not making python 3 the default :(
11:51 bhosmer joined #salt
11:52 hobakill joined #salt
11:53 the_drow But it's pretty simple to create a binary of your python app
11:53 vbabiy joined #salt
11:54 jaimed joined #salt
11:56 the_drow dzen: Isn't it possible to vendor your own Python like Vagrant does with ruby?
11:56 bhosmer joined #salt
12:04 the_drow dzen: take a closer look about embedding your applications with pants or uwsgi. It's really useful
12:04 jeddi joined #salt
12:06 mechanicalduck_ joined #salt
12:08 bhosmer_ joined #salt
12:09 sectionme joined #salt
12:11 mechanicalduck joined #salt
12:11 dzen I note it ;)
12:14 diegows joined #salt
12:20 mechanicalduck joined #salt
12:24 elfixit joined #salt
12:30 matthiaswahl joined #salt
12:33 oz_akan_ joined #salt
12:36 marnom joined #salt
12:40 tkharju4 joined #salt
12:40 bastion1704 joined #salt
12:51 bhosmer joined #salt
12:52 matthew-parlette joined #salt
12:56 bhosmer joined #salt
12:57 ccase joined #salt
13:00 elfixit joined #salt
13:01 miqui joined #salt
13:03 ggoZ joined #salt
13:04 bhosmer_ joined #salt
13:04 elfixit1 joined #salt
13:08 scott_w joined #salt
13:08 ramishra joined #salt
13:11 elfixit joined #salt
13:12 mpanetta joined #salt
13:16 racooper joined #salt
13:19 logix812 joined #salt
13:20 robertkeizer joined #salt
13:27 rojem joined #salt
13:28 vejdmn joined #salt
13:31 vejdmn joined #salt
13:31 rnts joined #salt
13:32 ajprog_laptop1 joined #salt
13:36 kaptk2 joined #salt
13:38 elfixit joined #salt
13:41 Deevolution joined #salt
13:41 aquinas joined #salt
13:44 intellix joined #salt
13:44 chose joined #salt
13:49 mechanicalduck_ joined #salt
13:50 acabrera joined #salt
13:55 aquinas joined #salt
14:02 mapu joined #salt
14:04 kermit joined #salt
14:05 chose Installation of printer TestsAuto   cmd.script:     - name: 'salt://scripts/run-printer-install.bat'
14:05 chose can anyone tell me why the state can't run cmd.script ?
14:08 mateoconfeugo joined #salt
14:08 koyd chose: I believe it's supposed to be cmd.script: -name "run-printer-install.bat' -source: "salt://scripts///scripts/run-printer-install.bat"
14:08 koyd as in: name = name + args, source = where to download
14:09 chose Thanks I will try it right away
14:09 koyd according to http://docs.saltstack.com/en/latest/ref/states/all/salt.states.cmd.html#salt.states.cmd.script
14:13 jalbretsen joined #salt
14:14 XenophonF oh i didn't know about cmd.script
14:14 XenophonF i've been using file.managed + cmd.run states
14:15 dude051 joined #salt
14:22 funzo joined #salt
14:22 chose Thank koyd it's seem to work I was getting a render error it seem that the arg -source:.. is needed to avoid the render error since it's the only diff
14:22 koyd np chose, glad it worked :)
14:28 ipmb joined #salt
14:30 masterkorp hello everyone
14:32 rallytime joined #salt
14:32 Twiglet joined #salt
14:34 perfectsine joined #salt
14:36 al joined #salt
14:41 pdayton joined #salt
14:44 elfixit joined #salt
14:45 Lloyd_ joined #salt
14:46 jeremyBass joined #salt
14:47 Lloyd_ Hi, probably me being stupid and missing something but, having a slight issue when updating users managed by salt. For example, salt creates user, i realise they are in incorrect group, i update pillar/state, run a highstate, salt says user already up to date and no change occurs. Any ideas?
14:49 Eureka_ joined #salt
14:52 housl joined #salt
14:52 babilen Lloyd_: Works for me -- are you using the user formula? Are you sure that your pillar changes have been picked up by the master? Could you share states/pillars + highstate run output on http://paste.debian.net ?
14:52 Lloyd_ that on 2014.1.7?
14:54 dimeshake hm, anyone know when 2014.1.7 will hit epel?
14:54 thedodd joined #salt
14:55 babilen Lloyd_: I am using .7 yes, but did not have to change user attributes since I've upgraded
14:56 Lloyd_ babilen: running highstate with "-l debug" shows the updated pillar data, but it definately isn't applying it
14:56 racooper dimeshake,  2014.1.7 is in EPEL testing since 7/11, I think it's something like 14 days unless it gets tested and positive feedback to release earlier.
14:57 dimeshake gotcha. perhaps I should look at using epel-testing for salt
14:57 dimeshake thank racooper
14:58 Reiner030 joined #salt
14:58 speed145a joined #salt
14:59 hobakill Good morning all. Anyone have much experience using the windows software repo in salt? i cannot get it to, well, install packages  - even from the git repo
15:02 ndrei joined #salt
15:02 eliasp hobakill: I might be able to help… what's your specific issue?
15:03 hobakill well eliasp ran through the documentation and i could't get it to install the firefox example... or my webdeploy
15:03 ipalreadytaken joined #salt
15:04 hobakill [root@salt salt-winrepo.git]# salt "*" pkg.available_version Firefox
15:04 hobakill SALTWINTEST2.NA1.ad.group:
15:05 hobakill (and then nothing)
15:05 eliasp hobakill: did you do "salt your-minion pkg.refresh_db"?
15:06 eliasp hobakill: this will make sure the list of installed and available packages is up-to-date on the minion
15:06 hobakill i did
15:06 hobakill [root@salt salt-winrepo.git]# salt "SALTWIN*" pkg.refresh_db
15:06 hobakill SALTWINTEST2.NA1.ad.group:
15:06 hobakill True
15:06 eliasp hobakill: did you run "salt-run winrepo.genrepo" before that to generate the winrepo.p file?
15:07 hobakill eliasp, i have yes.
15:07 eliasp ok, could you nopaste the output of "salt-run winrepo.genrepo"?
15:07 hobakill sure one sec.
15:10 hobakill http://pastebin.com/Xswe2cmx
15:12 ajolo joined #salt
15:15 Reiner030 Hello,
15:15 logix812 joined #salt
15:15 ajw0100 joined #salt
15:16 eliasp hobakill: I can confirm, that at least "pkg.available_version" doesn't return something in this case… what happens when you simply try to install firefox (salt your-test-minion pkg.install Firefox)?
15:17 eliasp hobakill: hmm, ok… probably relevant for pkg.available_version: "If the latest version of a given package is already installed, an empty string will be returned for that package."
15:18 eliasp hobakill: does "salt your-test-minion pkg.list_available Firefox" anything?
15:18 hobakill eliasp, i get nothing returned again
15:18 hobakill eliasp, it does not
15:18 eliasp hobakill: what windows version does your Minion run?
15:19 hobakill 2012 i think
15:19 hobakill confirmed elfixit
15:19 hobakill whoops eliasp
15:20 hobakill eliasp, [root@salt ~]# salt -G "os:Windows" pkg.list_available Firefox
15:20 hobakill SALTWINTEST2.NA1.ad.group:
15:20 hobakill [root@salt ~]#
15:22 teskew joined #salt
15:23 hobakill eliasp, FWIW - master is CentOS 6.5 running the latest stable salt
15:23 conan_the_destro joined #salt
15:28 eliasp hobakill: could you try to deploy the following files using dynamic module distribution (http://docs.saltstack.com/en/latest/ref/file_server/dynamic-modules.html), as I fixed some windows related issues in 'develop':
15:28 eliasp https://raw.githubusercontent.com/saltstack/salt/develop/salt/modules/win_pkg.py
15:28 eliasp https://raw.githubusercontent.com/saltstack/salt/develop/salt/modules/pkg_resource.py
15:28 eliasp https://raw.githubusercontent.com/saltstack/salt/develop/salt/states/pkg.py
15:28 helderco joined #salt
15:28 eliasp these fixes might not be part of 2014.1.x …
15:29 mateoconfeugo joined #salt
15:29 davet joined #salt
15:29 mateocon_ joined #salt
15:30 danielbachhuber joined #salt
15:31 helderco Hey guys… having some trouble, I guess with cache. “Pillar failed to render with the following messages: Specified SLS 'users' in environment 'base' is not available on the salt master”. That pillar is in another envirnoment which is correctly defined in the top.sls file. I tried the following to no avail: service salt-minion stop && rm -r /var/cache/salt && service salt-minion start. Any ideas how to proceed? saltutil.refresh_pillar re
15:31 helderco None.
15:32 hobakill eliasp, i can... is it worth snapshotting my master before doing this? how will it effect future master updates from EPEL?
15:34 eliasp hobakill: if it doesn't work, all you have to do is to remove these files from your directory again, run 'salt '*' saltutil.sync_all' and they're gone from the minions
15:34 eliasp it won't have any effect on the master
15:34 Lloyd_ babilen: after doing a bit more poking around with the users problem, it only seems to affect adding or removing groups in the 'groups:' part of the user state. Updating uid or gid works fine without any issues, just adding more groups seems to crap out and do nothing
15:35 hobakill ok - eliasp thanks. stand by
15:35 eliasp if you're using gitfs, you can also create the corresponding subdirs (_{modules,grains,renderers,returners,states}) in your git root
15:37 hobakill sorry - so i'm not totally clear where to put these files eliasp
15:37 hobakill modules, grains, renderers....etc... should go in my root win file structure?
15:38 hobakill and then where do the py files go?
15:38 eliasp hobakill: no, in the root of your states directory… see also: http://docs.saltstack.com/en/latest/ref/file_server/index.html
15:39 eliasp so if you store your states in /srv/salt/, create /srv/salt/_states and /srv/salt/_modules and place them there
15:40 hobakill eliasp, i understand. thanks.
15:41 Lloyd_ babilen: http://paste.debian.net/hidden/a361b4c7/
15:42 hobakill and those py files should go into each of the directories i just created eliasp  ? (sorry for the n00bing)
15:42 bastion1704 joined #salt
15:43 eliasp hobakill: exactly… depending on their original location… files from salt/states/ go to _states, from salt/modules/ to _modules …
15:43 aquinas joined #salt
15:44 hobakill ok pushing the git.. stick by eliasp
15:45 ramteid joined #salt
15:46 bastion1704 Hello, is it possible to change de default returner for multiple ones ?
15:47 hobakill ok eliasp - files are on the master. now just salt "win-min" saltutil.sync_all ?
15:47 erjohnso joined #salt
15:47 eliasp exactly
15:48 eliasp then again: "salt win-min pkg.refresh_db" and "salt win-min pkg.install Firefox"
15:48 cDR_ joined #salt
15:49 cDR_ Hi
15:49 hobakill eliasp, is the sync_all supposed to return a true or anything? i just got blank things like grains: modules: outputters: etc
15:50 eliasp it is supposed to list at least the files you put in _states and _modules
15:50 hobakill :/
15:50 cDR_ question; how would you remove (blank) lines form a text file using a state file? For now I'm using cmd.run with sed, but it feels wrong
15:51 eliasp hobakill: otherwise there might a problem with your salt-fileserver config
15:51 hobakill [root@salt _modules]# ls
15:51 hobakill pkg_resource.py  win_pkg.py
15:51 hobakill [root@salt _modules]# pwd
15:51 eliasp cDR_: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.replace
15:51 ipalreadytaken joined #salt
15:51 tligda joined #salt
15:52 eliasp hobakill: try "salt win-min cp.list_master"
15:52 eliasp hobakill: it should show you all files stored on the master known to the minion
15:52 eliasp hobakill: if nothing shows up, check your fileserver configuration: http://docs.saltstack.com/en/latest/ref/file_server/file_roots.html
15:52 catpig joined #salt
15:53 kaiserpathos joined #salt
15:54 hobakill eliasp, interesting. it only sees what appears to be my base dir
15:54 hobakill nothing under /srv/salt/win /srv/salt/_modules etc
15:56 babilen Lloyd_: Sorry, I left work and will be busy with something else for the next couple of hours (hence my failure to reply)
15:57 hobakill eliasp, it would appear i have some lines to add in that configuration. :/ stick tight.
15:58 hobakill eliasp, argh i know why this isn't working now.....
15:59 eliasp messed up fileserver configuration, so even the generated winrepo.p won't be distributed? ;)
15:59 Lloyd_ babilen: no worries. I will be heading home myself in about 30 mins, I don't use IRC at home so can either catch up with you tomorrow or PM you my email address for if/when you get time.
16:00 hotbox joined #salt
16:01 hobakill eliasp, the file structure was definitely wrong but the minion still doesn't see /srv/salt/win  (i had it set to /srv/salt/windows barrrrrrrrrgh...) do i need to restart the minion service? (already restarted master)
16:01 cDR_ eliasp, thanks but I don't know how to delete a line, i can only replace it with a blank line using file.replace, is this possible at all?
16:01 giannello is there any limit for "extends" in a cloud profile file?
16:04 bhosmer joined #salt
16:06 wendall911 joined #salt
16:07 xzarth joined #salt
16:08 vejdmn joined #salt
16:08 druonysus joined #salt
16:08 druonysus joined #salt
16:11 hobakill eliasp, yeah it WAS messed up but it's fixed now but the minion still doesn't see it
16:11 jcsp joined #salt
16:11 UtahDave joined #salt
16:12 helderco Hey guys… I’m not being able to fix a cache issue. “Pillar failed to render with the following messages: Specified SLS 'users' in environment 'base' is not available on the salt master”. That pillar is in another envirnoment which is correctly defined in the top.sls file. I tried the following to no avail: service salt-minion stop && rm -r /var/cache/salt && service salt-minion start. Any ideas how to proceed? saltutil.refresh_pillar
16:12 helderco returns None.
16:12 Eugene Wheee.... $DAYJOB is having me write up a proposal for replacing a (poorly-maintained) bcfg2 install with Salt.
16:13 UtahDave Eugene: woot!
16:13 UtahDave helderco: can you pastebin your sanitized top.sls?
16:15 helderco UtahDave: sure, give me a moment…
16:15 patrek_ joined #salt
16:18 notpeter_ joined #salt
16:19 nebuchadnezzar joined #salt
16:20 helderco UtahDave: http://pastebin.com/ZasByuY5
16:21 ajolo joined #salt
16:22 anuvrat joined #salt
16:22 KyleG joined #salt
16:22 KyleG joined #salt
16:22 UtahDave helderco: is there anything in    /srv/base/pillar   ?
16:23 helderco Yes, everything’s there
16:23 KyleG joined #salt
16:23 KyleG joined #salt
16:24 UtahDave helderco: your paste has stuff in  /srv/prod/pillar/users.sls    but I see no mentionof    /srv/base/pillar/users.sls
16:25 helderco users is in prod, it has to be in base?
16:25 helderco — /srv/prod/pillar/users.sls
16:26 UtahDave hm. just a sec
16:26 rglen joined #salt
16:27 aquinas joined #salt
16:32 UtahDave ok, let's try 2 things.  First, make sure all your indents are 2 spaces in your pillar top.sls.  The mismatching indents might be just your pastebin, but make sure.
16:32 hobakill well i'm completely lost with this idiotic windows repo that my minion can't seem to talk to
16:33 UtahDave helderco: Second, stop the salt-master service,  then run it in another terminal with     sudo salt-master -l debug
16:33 UtahDave helderco: then run refresh_pillar on that minion and see if you get any stacktraces or other related errors
16:33 UtahDave hobakill: what's going on?
16:34 helderco UtahDave: on it
16:34 hobakill hey UtahDave - well i've seemingly set it up properly but i cannot , well, install anything. eliasp was helping earlier but they've disappeared i think
16:34 babilen Lloyd_: I'd suggest to just come back here and ask everyone .. feel free to highlight me afterwards and I'll comment if I have something to say. Have a good evening :)
16:35 UtahDave hobakill: Did you send that email regarding TortoiseSVN?
16:36 hobakill i did not no
16:36 UtahDave Hm. Ok, seems like you're not the only person having issues with that.
16:36 UtahDave hobakill: can you pastebin the steps you've taken so far along with the errors you've gotten?
16:37 RandalSchwartz hmm.  got an answer to my mailing list question overnight.  Hope it works.
16:37 hobakill that's the thing. there are no errors - just blank returns. UtahDave
16:37 helderco UtahDave: no errors on refresh_pillar… will attempt an highstate
16:37 RandalSchwartz is there any way to see a highstate while it is in progress?
16:38 RandalSchwartz like a verbose flag?
16:38 RandalSchwartz maybe somethign with a salt-call on the affected host?
16:38 hobakill RandalSchwartz, i think you have to change the debug level or do a salt-call from the minion and that has some verbosity
16:38 bhosmer joined #salt
16:38 UtahDave hobakill: ok, pastebin what you've done so far, then
16:39 bhosmer_ joined #salt
16:39 UtahDave RandalSchwartz: adding   -l  debug    will give you more info on what the Master is doing, but not what the minion is doing
16:39 helderco UtahDave: err… actually refresh_pillar did return the error: Specified SLS 'users' in environment 'base' is not available on the salt master
16:40 helderco i think i have an idea
16:40 hobakill UtahDave, initial salt-run http://pastebin.com/s9hUQtVP
16:40 t0rrant joined #salt
16:41 UtahDave RandalSchwartz: One thing I've been doing with lately is writting a runner that listens on the event bus and does things with events coming from a highstate. You can even send your own events from your sls files.
16:42 hobakill UtahDave, then attempts to refresh db on minion and get firefox version: http://pastebin.com/YjmBwr7q
16:42 helderco UtahDave, have a /srv/dev/pillar/top.sls with users and hosts on base. Because my file_roots has a different order on dev to have it working locally while testing. On the salt-master -l debug terminal, the error is comming after that top.sls file.
16:42 RandalSchwartz ok thanks
16:43 XenophonF did something change in 2014.1.7 to make the salt command less verbose?  if i run a command like "salt uxedevlnxsvr02 grains.items" or "... state.highstate", I get nothing back
16:43 XenophonF i have to set "-l debug" to get the customary output
16:44 helderco trying to fix
16:44 UtahDave helderco: having multiple top.sls file gets complicated fast. I usually recommend having only one.
16:44 XenophonF with debugging enabled, I get the error that jid whatever did not return in time
16:44 kballou joined #salt
16:44 UtahDave hobakill: so is the wmi_provider the one you're having issues with?
16:45 UtahDave XenophonF: try adding more to your timeout     -t 20    for example
16:45 hobakill no. firefox... or a custom one i wrote for Web Deploy.... but i'm just trying to get it to do the basics as this point UtahDave
16:46 helderco UtahDave: yeah, but I not sure how I would do otherwise
16:47 hobakill UtahDave, the fact that i get this tells me that the minion cannot see something: http://pastebin.com/22DwJX1r
16:47 UtahDave hobakill: can you pastebin your application sls file?
16:47 XenophonF UtahDave: there's a good chance the minion in question has crashed - nothing to do with salt, sorry
16:47 UtahDave XenophonF: no problem!
16:47 XenophonF yeah, it panicked
16:48 XenophonF i guess  the hv_balloon driver in centos 7 is still buggy
16:48 UtahDave ah
16:48 XenophonF i was kind of hoping that got fixed
16:48 hobakill UtahDave, sure......http://pastebin.com/0Jdi66TQ
16:49 helderco UtahDave: now I’m getting another error. It seams he’s not getting the gitfs_remotes on prod…
16:49 hobakill UtahDave, that's for my "custom" one - but i'd think it would work for the 'built-in' firefox install from the salt git repo
16:49 helderco failed to get users and mysql.server on prod, which come from users-formula.git and mysql-formula.git
16:50 schimmy joined #salt
16:52 helderco debug in master shows fetching from git, but only after the minion has output failure
16:54 XenophonF hey all where's the grains code that says whether a minion is virtual or not?
16:55 XenophonF on CentOS 7 grains.items['virtual'] returns 'VirtualPC'
16:55 saru11 joined #salt
16:55 XenophonF on FreeBSD 10 it returns 'physical'
16:55 XenophonF in both cases they're running under Hyper-V 2012
16:55 bhosmer_ joined #salt
16:56 bhosmer_ joined #salt
16:56 saru11 hey guys, I have noticed some difference between salt and salt-call in the output they produce
16:56 saru11 when I run salt MINION my_module --output=yaml, it produces standard YAML format as I know
16:57 saru11 but when I switch to salt-call my_module --output=yaml, string attributes are prefixed with '~~python/unicode' string
16:58 UtahDave XenophonF: in the salt repo    salt/grains/core.py
16:58 saru11 sorry, with '!!python/unicode'
16:58 UtahDave XenophonF: I'm not surprised we're not catching cent7 yet
16:59 linjan joined #salt
17:00 UtahDave hobakill: and then did you run these 2 commands on the master?  http://docs.saltstack.com/en/latest/topics/windows/windows-package-manager.html#generate-repo-cache-file
17:00 ipmb_ joined #salt
17:01 Eugene What would be the best way to talk to somebody about SaltStack Enterprise? Phone call?
17:01 vejdmn joined #salt
17:01 hobakill UtahDave, i have yes.
17:01 UtahDave Eugene: Yeah, probably.  Want me to pm you?
17:01 XenophonF UtahDave: in both cases it picks up that the hw manufacturer is Microsoft and the productname is virtual machine
17:01 XenophonF which is probably good enough for me now
17:02 UtahDave XenophonF: if you come up with a patch that catches that correctly, that would be awesome.
17:02 schimmy1 joined #salt
17:03 forrest joined #salt
17:04 cwyse joined #salt
17:04 XenophonF you bet
17:04 XenophonF say i wrote some modules for user/group management on freebsd
17:04 XenophonF what's the best way to get them to you?
17:04 XenophonF they probably need more testing
17:08 UtahDave XenophonF: A pull request on https://github.com/saltstack/salt   is the best way
17:08 XenophonF OK
17:09 ndrei joined #salt
17:09 cbaesema joined #salt
17:09 kballou joined #salt
17:10 UtahDave hobakill: could you open an issue on github for this issue?  There may be a bug there
17:10 hobakill UtahDave, this is interesting though. running: salt "win min" cp.list_minion env=win doesn't show any of my win\repo files
17:11 ml_1 joined #salt
17:11 UtahDave hobakill: are you setting up a different environmnet for your windows machines?
17:11 hobakill UtahDave, http://pastebin.com/u51U7gMV
17:11 hobakill it's under /srv/salt/win
17:13 hobakill UtahDave, my fileroots are pretty basic right now: http://pastebin.com/ZYY12hgP
17:15 UtahDave hobakill: in your master config what's the config for the win repo path set to?
17:16 hobakill UtahDave, http://pastebin.com/eUWhajkK
17:16 MTecknology Could someone with minion and master at 2014.1.5 try out   salt-call event.fire_master "{'host': 'moses', 'action': 'highstate'}" 'minion_request'   and tell me if eventlisten.py shows the event making its way through? I'm not seeing it happen and it was working fine in a previous version. I want to make sure I'm not insane.
17:18 manfred MTecknology: i believe that was a bug in 2014.1.5, besically determined that was just a super broken release, 2014.1.7 should be announced today, (it is already packaged)
17:18 MTecknology oh
17:19 MTecknology welp- thanks!
17:19 MTecknology Guess I'll wait.
17:19 manfred what are you running on?
17:19 MTecknology debian and windows mostly, a few centos
17:20 UtahDave hobakill: yeah, that looks right.    Could you open an issue for this?
17:20 hobakill sure. got a URL?
17:20 MTecknology debian boxes are the ones that need event.fire_master
17:20 UtahDave hobakill: https://github.com/saltstack/salt/issues/new
17:20 hobakill thansk
17:20 UtahDave thank you, hobakill!
17:21 hobakill odd that this hasn't come up before tho yeah? i'm using the EPEL 2014.1.4 release.
17:23 UtahDave yeah. for sure.
17:25 FarrisG joined #salt
17:28 Reiner030 Hello,
17:28 Reiner030 I have a little question where I cannot find any google matches for when I searches longer with creative keyword combinations.
17:28 Reiner030 Is it perhaps possible to use also one/multiple saltmasters on same server which different company/projects so
17:28 Reiner030 that it's initially possible to minimize server maintenance/costs and later separate the individual saltmaster
17:28 Reiner030 to an external source?
17:28 Reiner030 Or must there each an own (v)server running?
17:29 forrest Reiner030, while you could do that, it sounds like a bad (and overly cheap) plan.
17:29 forrest it's a security concern as well
17:29 dstokes hey guys. getting a '_Serving' has no attr 'session' when trying to use the run endpt in salt-api. anybody else run into this?
17:30 dstokes passing auth and eauth method in request
17:30 aw110f joined #salt
17:30 travisfischer joined #salt
17:30 Reiner030 ok, was only the idea like webserver vhosts to maintain first a single production deploy and "export" later the config to the company
17:31 yomilk joined #salt
17:32 MTecknology manfred: thanks again for that info. I was beating my head pretty hard and am extremely excited to know that I can relax and be patient. :)
17:32 Reiner030 then thanks for the hint - I would check howto setup a small vserver environment for each project then (for startup projects)
17:32 d_d_d joined #salt
17:32 manfred np
17:33 n8n joined #salt
17:34 hobakill UtahDave, issue #14203 submitted.
17:34 chrisjones joined #salt
17:34 anuvrat joined #salt
17:36 d_d_d hello all. I am trying to extend some of the salt python code and push up a change. That being said, I would obviously like to test my change. When I use setup.py to rebuild and deploy the my new code, then restart the salt master, I get an error saying that the function on the module that I am trying to call is not available
17:36 d_d_d does anybody have some ideas as to what I am missing?
17:39 rojem joined #salt
17:39 Eureka_ joined #salt
17:40 UtahDave d_d_d: Have you tried adding   --force   to your setup.py install command?
17:41 d_d_d no i did not... i'll give that a shot, thanks
17:45 scristian joined #salt
17:45 damianball joined #salt
17:45 jaimed joined #salt
17:46 tkharju joined #salt
17:47 badon_ joined #salt
17:47 scristian hi, I'm looking for a diagram with Salt architecture, in docs I can't find one, does anyone have it and is willing to share it ? thank you
17:47 smcquay joined #salt
17:48 d_d_d Utah_Dave: that worked,,, thanks
17:49 forrest scristian, what do you mean? The architecture of salt itself? Or a system using salt?
17:49 scristian for the salt itself
17:50 Reiner030 left #salt
17:51 Ryan_Lane joined #salt
17:51 gothix joined #salt
17:51 gothix joined #salt
17:51 forrest scristian, I don't think such a thing exists, or if it does I've never seen it
17:52 scristian ok, thank you
17:52 Ahlee_ so it looks like custom returners stay in cache even after syncing a new version with saltutil.sync_returners
17:52 manfred scristian: there aren't any really
17:52 manfred scristian: http://docs.saltstack.com/en/latest/topics/topology/proxyminion/index.html
17:52 manfred that one is how proxy minions work
17:52 manfred http://raet.docs.saltstack.com/en/latest/topics/introduction.html
17:52 manfred that is the raet diagram
17:53 gothix can this be used in a salt state  acl.addfacl ? I get an error saying  its unavailable
17:53 manfred scristian: and that is all i found
17:54 scristian manfred: thank you
17:55 ajolo joined #salt
17:56 forrest gothix, it's only a module, call it using http://docs.saltstack.com/en/latest/ref/states/all/salt.states.module.html
17:57 gothix forrest, Great thanks!
17:57 amcnea joined #salt
17:57 forrest np
17:57 kivihtin joined #salt
18:01 mico joined #salt
18:01 mico howdy
18:04 conan_the_destro joined #salt
18:05 silviudicu joined #salt
18:05 silviudicu left #salt
18:05 Eureka_ joined #salt
18:06 silviud joined #salt
18:07 matthew-parlette joined #salt
18:07 silviud hi - i'm looking to copy a file from a windows minion to master but the exe that installs the minion doesn't contain the salt.exe - what can be done ?
18:08 anuvrat joined #salt
18:08 jslatts joined #salt
18:09 kalessin left #salt
18:12 Theo-SLC joined #salt
18:12 Theo-SLC Hale ye salts
18:12 q4brk_ joined #salt
18:13 UtahDave hola, Theo-SLC!
18:13 UtahDave silviud: use salt-call
18:15 talwai joined #salt
18:15 Theo-SLC I'll be using salt to deploy OpenShift Origin today.  Before I start, has anybody here done this before?
18:16 ipalreadytaken joined #salt
18:16 talwai The docs don't state it explicitly, but I assume that when defining an ext_pillar I can use a git+ssh://somerepo.git style remote just like i would define gitfs_remotes?
18:16 talwai http://docs.saltstack.com/en/latest/ref/pillar/all/salt.pillar.git_pillar.html
18:17 silviud @UtahDave - do you have an example with salt-call ?
18:18 UtahDave talwai: I think that's true. I haven't tried it myself, but give it a try
18:18 UtahDave silviud: first you have to set an option in your config to allow the minion to upload to the master.
18:19 UtahDave silviud: file_recv: True
18:19 silviud file_recv: True - it';s there
18:19 silviud on the master
18:20 zime joined #salt
18:20 UtahDave silviud: then use cp.push     http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.cp.html#salt.modules.cp.push
18:20 zime using salt-cloud, how to by pass the accept conditions of AWS when launching a ami?
18:20 UtahDave so from the minion that would be      salt-call cp.push '/tmp/myfile.txt'
18:21 UtahDave zime: usually salt-cloud "Just Works" (tm)
18:21 UtahDave are you using a specific ami with an eula or something?
18:21 zime yea, when requesting a centos
18:22 zime ubuntu works as expected..
18:23 zime when I run salt-cloud -p profile_info centos_test
18:23 zime I get
18:23 zime http://paste.debian.net/109718/
18:24 UtahDave Hm. what ami are you using?
18:24 ckao joined #salt
18:24 UtahDave zime: I don't know the answer to this one.  Could you open an issue for this?
18:24 zime k
18:26 quantumriff joined #salt
18:33 talwai What's the recommended way to restart the salt-master from the machine that's running the salt-master?
18:33 talwai "service salt-master restart" doesn't seem to work
18:34 KyleG talwai: Might be salt_master instead
18:35 kermit joined #salt
18:35 quantumriff On many of my minions, I need to automate the deployment of the Oracle database client software.  I'm trying to figure out the best way to do it with saltstack.  Normally, I scp the large zip file, unzip it, and run the installer (I can pass an asnwer file to it if needed) but this seems like something ugly to do in salt.  Especially because of how many files are in the zip.. (thousands of JAR files)
18:35 vxitch left #salt
18:35 quantumriff would this be something like a single file for the zip, then a bunch of cmd.wait commands in the right order, that watch that file?
18:36 anuvrat joined #salt
18:36 jslatts joined #salt
18:36 talwai KyleG: nope it's salt-master, 'service salt-master status' runs fine
18:38 quantumriff talwai: I have seen mine hang before.. what happens if you do 'service salt-master stop' ? does it stop, but leave some processes hanging?
18:38 aboe joined #salt
18:40 talwai quantumriff: it says "stop: Unknown instance: "
18:41 toastedpenguin joined #salt
18:46 mgarfias is there a way to pre-seed a set of files using salt-cloud?
18:47 miqui joined #salt
18:48 miqui joined #salt
18:48 jaimed joined #salt
18:49 miqui joined #salt
18:49 Theo-SLC joined #salt
18:49 dmendez joined #salt
18:49 netzmonster joined #salt
18:52 dRiN joined #salt
18:52 silviud @UtahDave - i got it working - thanks
18:52 anuvrat joined #salt
18:54 troyready joined #salt
18:54 UtahDave silviud: great!
18:56 aboe joseph are you online?
18:58 sebnil joined #salt
18:58 badon joined #salt
18:59 sebnil Hello, i'm having some problems with the iptables module. I'm using append and every time salt is running a new iptables rule with the same values. How can i tell salt to check if it's already exists and skips adding a new one every time?
19:00 penguin_dan joined #salt
19:01 bmatt sebnil: my guess is that the module doesn't check, since duplicated rule-setting is idempotent
19:01 krow joined #salt
19:01 bmatt so the short answer is "re-write a portion of the module"
19:01 n3wbie joined #salt
19:02 sebnil hmmm.
19:02 manfred bmatt: it does
19:02 manfred it checks for it
19:02 manfred sebnil: what OS?
19:02 sebnil CentOS
19:02 manfred 6 or 5?
19:02 sebnil 6
19:03 manfred you are not going to reliably get iptables module to work
19:03 manfred because it just does regex against your iptables-save data
19:03 manfred with shortened flags and stuff
19:03 manfred the best thing to do on any os that doesn't have the iptables --check flag is to just generate your /etc/sysconfig/iptables configuration and reload iptables
19:04 sebnil So manually push the entire table? That could work.
19:04 quantumriff talwai: it sounds like it didn't start via the service command, so it isn't tracking it.  I guess worst case, you can do a 'chkconfig --list' (on a RHEL server) and see if salt-minion is there, and set to start automatically, and then kill the processes manually, and start them
19:04 sebnil or, manually write.
19:04 rojem joined #salt
19:05 n3wbie Hello! I'm looking for a good way to do sequentially remote execution across server farms where processes ping the status of the previous step for successful completion before execution for both local and remote hosts.  Is there a good module for this?  (ex. rolling server restarts with service comands to shutdown running tomcat instances in a specif order)
19:05 juancri joined #salt
19:06 n3wbie Best I have so far is a FOR loop script with commands listed in order on master and waiting for exitcode before running next step.
19:06 RandalSchwartz alright - solved my nested down/up problem - https://groups.google.com/d/msg/salt-users/fk5l1Hluq1w/f4oDit6tv2sJ
19:07 RandalSchwartz but now pgpool is being a bear.
19:09 bmatt n3wbie: the service module has specific hooks for pre- and post-restart steps, but I've done that in a more hacky way using chained cmd.run states
19:10 bmatt er, I thought service did.
19:10 bmatt hm.
19:10 n3wbie My question is very similar to what you are trying to accomplish Randal.  I was looking at both those options, just didn't know if there was a better way to flag the command to run sequentially across hosts found via Glob or if its ALWAYS parallel
19:11 alanpearce joined #salt
19:11 linjan joined #salt
19:11 sebnil ls
19:11 bmatt A voice says, "PLOUGH."
19:12 bmatt s/O//
19:13 AdamSewell joined #salt
19:13 n3wbie bmatt: any chance you could provide an example for me?  I don't see the any hooks in service module.
19:13 gamedna joined #salt
19:13 nmistry joined #salt
19:13 ipalreadytaken joined #salt
19:14 bmatt http://docs.saltstack.com/en/latest/ref/states/requisites.html
19:14 bmatt hahaha
19:15 bmatt n3wbie: I'm thinking of the "prereq" requisite, but it looks like it's been lost to the mists of time or something?
19:15 n3wbie awesome... haha
19:21 sectionme joined #salt
19:23 jslatts joined #salt
19:23 gothix So im getting this response "Function acl.addfacl is not available
19:23 gothix " what do i need to do make it availabe
19:23 gothix ?
19:25 martoss joined #salt
19:25 silviud from all the netapi deployment options which one will be the best - tornado ?
19:25 RandalSchwartz so pgpool is being annoying.  I can't use the built-in service restart.  I need to send a stop, then wait a bit, then a start
19:26 RandalSchwartz can I do a cmd.run watch on service: pgpool?
19:26 RandalSchwartz and let it manage the restart instead of service?
19:26 martoss left #salt
19:32 simmel_ joined #salt
19:33 bmatt RandalSchwartz: sounds like you're solving the wrong problem :)
19:36 Theo-SLC joined #salt
19:37 tedski whiteinge: any plans to update readthedocs with current documentation? 2014.1.7 doesn't seem available
19:38 tedski whiteinge: also, salt.rtfd.org says latest release is 2014.1.5
19:43 ajolo joined #salt
19:46 RandalSchwartz I know
19:46 RandalSchwartz where's the docs for requisites again.  the googles fail
19:46 RandalSchwartz they point at the wrong page, and the wrong page doesn't have a link to the right page.
19:47 RandalSchwartz broken page - http://docs.saltstack.com/en/latest/ref/states/requisites.html
19:48 RandalSchwartz until I can get this sorted out, I can't let pgpool watch its package or conf file.  that's silly
19:48 RandalSchwartz and that also means I can't restart postgresql either. :(
19:49 n8n joined #salt
19:50 tedski RandalSchwartz: http://salt.readthedocs.org/en/v2014.1.4/ref/states/requisites.html
19:50 tedski RandalSchwartz: you might ant cmd.watch instead of cmd.run
19:51 jrdx joined #salt
19:52 RandalSchwartz oooh.  cmd.watch might be in
19:52 RandalSchwartz it
19:55 jrdx joined #salt
19:56 FeatherKing joined #salt
19:57 FeatherKing manfred: you around?
19:58 FeatherKing anyone used saltify before? i have a profile and map setup but when i run it i get "The profile 'make_salty' is defining 'saltify' as the provider. Since there's no valid configuration for that provider, the profile will be removed from the available listing"
19:59 FeatherKing but in the docs it says saltify doesnt need a provider setup
20:05 jslatts joined #salt
20:07 anuvrat joined #salt
20:09 manfred i am, but i am busy at work sorry
20:12 schimmy joined #salt
20:14 RandalSchwartz ugh... pgpool is still giving me headaches
20:14 schimmy1 joined #salt
20:15 RandalSchwartz I have the right rules to stop/start apache properly bracketing a cmd.watch, but the actual stop/start of pgpool doesn't run the same from salt as it does from the command line
20:18 CeBe joined #salt
20:20 RandalSchwartz hopefully someone answers this while I move on
20:20 Eureka_ joined #salt
20:22 miqui joined #salt
20:24 whiteinge Correct requisites doc: http://docs.saltstack.com/en/latest/ref/states/global_state_arguments.html
20:24 whiteinge (Sent a pull req to fix the link.)
20:24 jslatts joined #salt
20:25 martoss joined #salt
20:25 whiteinge tedski: thanks for the RTD heads-up
20:25 tedski no prob
20:28 whiteinge RandalSchwartz: can you tell what is the difference between how pgpool start/stop runs manually vs runs via salt? salt should just be shelling out to whichever service tool is default for that platform. (is this on a bsd system?)
20:29 whiteinge if you run via ``salt-call -l debug service.stop pgpool`` you can see the actual CLI command that salt is running
20:29 d3vz3r0 joined #salt
20:29 vejdmn joined #salt
20:35 RandalSchwartz Yeah let me look
20:35 RandalSchwartz ... /usr/sbin/service pgpool onestop
20:35 RandalSchwartz Yeah, that's correct
20:36 RandalSchwartz I think it's a timing issue.
20:36 druonysuse joined #salt
20:36 druonysuse joined #salt
20:36 RandalSchwartz apache has to be fully stopped for a few seconds before pgpool can stop
20:36 kermit joined #salt
20:36 RandalSchwartz oooh - you didn't tell me that would actually stop pgpool
20:36 * RandalSchwartz quickly resets the live services )
20:37 jdmf joined #salt
20:38 Eureka_ RandalSchwartz: If you want to "test" a command you should look here. http://docs.saltstack.com/en/latest/ref/states/testing.html
20:38 Eureka_ RandalSchwartz: That will explain how to run a command without it actually running (get to see the output)
20:38 RandalSchwartz yes - been using that a lot
20:39 RandalSchwartz But still something is different about how it runs under salt
20:39 RandalSchwartz apache gets stopped and started just fine. :)
20:39 RandalSchwartz for some reason, pgpool is being persnickity
20:41 RandalSchwartz ahh... pgpool doesn't handle *restart* correctly
20:41 Eureka_ RandalSchwartz: Do you have environment var's getting set on your user for the server when you log in?
20:41 RandalSchwartz stop and start are fine
20:41 Eureka_ RandalSchwartz: Ah, in that case try modifying the start/stop section of the initd file for pgpool to add a sleep 5s (sleep 5 seconds) between the stop and start.
20:41 Eureka_ RandalSchwartz: it may just be triggering too fast.
20:42 RandalSchwartz no wait... deeper problem
20:43 RandalSchwartz I used salt to stop apache22
20:43 RandalSchwartz then used salt to stop pgpool
20:43 RandalSchwartz and it hung
20:43 RandalSchwartz but "service stop pgpool" worked just fine
20:44 Eureka_ RandalSchwartz: Stop the minion on your host and run it with salt-minion -l debug it will keep it in the foreground so you can see error output more easily.
20:45 RandalSchwartz I never get this message when using salt: stop request sent to pgpool. waiting for termination
20:45 krow joined #salt
20:46 doctorsound joined #salt
20:46 doctorsound left #salt
20:47 RandalSchwartz only when I say "service pgpool stop"
20:47 Eugene UtahDave - ping
20:48 RandalSchwartz No - I mixed up my tests.
20:48 matthiaswahl joined #salt
20:49 RandalSchwartz I was trying restart, not stop
20:49 RandalSchwartz so yeah, the problem is that I need to wait about 10 seconds between stop and start
20:50 RandalSchwartz I think maybe I wasn't waiting long enough
20:50 jaimed joined #salt
20:50 Eureka_ RandalSchwartz: That seems like it makes sense.
20:50 ajolo joined #salt
20:53 ipalreadytaken joined #salt
20:53 RandalSchwartz ok - trying a 15 second delay now
20:53 RandalSchwartz I won't see the command output until the command fully runs?
20:54 RandalSchwartz ugh - still not bringing pgpool down :(
20:54 Eureka_ RandalSchwartz: I believe that is correct.
20:54 Eureka_ RandalSchwartz: Trigger a stop, sleep, start manually rather than using the build in restart/reload?
20:55 RandalSchwartz yes
20:55 RandalSchwartz and for some reason, even that fails
20:55 RandalSchwartz I'm using a cmd.wait now
20:56 RandalSchwartz In another window, I verify that it is hung trying to bring pgpool down
20:57 Eureka_ RandalSchwartz: That is strange. I feel like it has something to do with not having a tty or something.
20:57 RandalSchwartz yeah maybe
20:58 RandalSchwartz maybe I can redirect to/from /dev/null
20:58 ndrei joined #salt
21:00 jslatts joined #salt
21:01 RandalSchwartz yeay... /dev/null worked!
21:01 Eureka_ o.0
21:01 eliasp RandalSchwartz: regarding your postgresql/pgpool/apache service dependencies issue… any chance you're running this on top of an initsystem which is service-dependency aware? so you could simply let the init-system handle this?
21:02 RandalSchwartz freebsd doesn't have service stop awareness
21:02 eliasp oh, looks like you're discussing the whole topic right now… should've read the backlog… just read the mail on the ML
21:02 aquinas joined #salt
21:02 RandalSchwartz it has only startup ordering
21:02 eliasp RandalSchwartz: ah, k… good old SysV-Init?
21:02 RandalSchwartz ...         service pgpool restart </dev/null >/dev/null 2>&1
21:02 RandalSchwartz that worked
21:02 ecdhe joined #salt
21:04 yomilk joined #salt
21:04 rojem joined #salt
21:07 avn_ Hi all. Are possible use docker to develop/test/debug states before push them to production?
21:07 RandalSchwartz ... https://groups.google.com/d/msg/salt-users/fk5l1Hluq1w/ZpZJX-dwzPwJ
21:07 faulkner Has anyone made any progress on https://github.com/saltstack/salt/issues/5004?
21:08 forrest faulkner, not that I'm aware of
21:12 RandalSchwartz now I just need to figure out how to add the third level
21:12 RandalSchwartz triggering restart_pgpool around any update of postgresql
21:12 RandalSchwartz which then hopefully triggers the apache down/up
21:14 eliasp avn_: sure… Salt's Docker support is really good
21:14 RandalSchwartz would adding prereq_in: cmd: restart_pgpool do that?
21:14 eliasp avn_: ah, you meant the other way round… test states inside Docker containers
21:14 tedski avn_: using vagrant is handy for that
21:14 eliasp avn_: well, this should work as well
21:14 schimmy joined #salt
21:15 tedski avn_: i have a vagrant file that spins up a multi-machine cluster and preseeds the master... through targeting and synced folders, you could easily test states
21:15 bhosmer_ joined #salt
21:15 eliasp avn_: but you'll have to live with minor differences running something inside an unprivileged Docker container, as you don't have full privileges (eg. mounting is not possible)
21:18 schimmy1 joined #salt
21:18 avn_ eliasp: no, haven't mounts at all. And prefer docker over vagrant when it possible
21:18 ssj joined #salt
21:18 eliasp avn_: sure… Docker is way faster and lightweight spinning up new instances…
21:19 gzcwnk hi, anyone in?   for teh salt state on selinux, is there an option just to report the status and not change it?
21:19 eliasp avn_: so I'd say Docker is fine (for testing) depending on the kind of states you want to test… if it's just stuff like pkg, file, … fine… if you want to do low-level stuff it might get tricky or even impossible
21:20 avn_ eliasp: any recipes about getting both master and minions in linked containers?
21:20 eliasp gzcwnk: are you looking for "salt your-selinux-minion state.highstate test=True"?
21:20 avn_ eliasp: yea, mostly pkg stuff and little bit of compilation
21:21 avn_ for stuff which touch mount/sysctl/... I just spin few extra EC2 instances
21:21 eliasp avn_: I don't have any recipes regarding linking for you, but http://docs.saltstack.com/en/latest/ref/states/all/salt.states.dockerio.html should give you all you need… so you could go full-circle… running your test-environment based on itself ;)
21:21 gzcwnk I can try that, thanks
21:22 gzcwnk I need it in a sls file though
21:22 avn_ eliasp: cool, so I can just drop docker to salt master, and do all testing there ;)
21:22 eliasp avn_: you should also be aware of https://github.com/simonmcc/kitchen-salt which you probably could use/adapt to run it inside your containers
21:22 bhosmer_ joined #salt
21:25 RandalSchwartz Hmm.  Still having trouble figuring out how to add the third level to this
21:25 davet joined #salt
21:25 kermit joined #salt
21:25 RandalSchwartz if I need to update postgresql, how to trigger apache down and pgpool down
21:25 RandalSchwartz especially when pgpool has to be done with a cmd wait
21:26 Eureka_ RandalSchwartz: Are you using watch?
21:26 RandalSchwartz yes
21:27 Eureka_ RandalSchwartz: Hum... That should handle it I would think? Otherwise you can use a reactor?
21:27 RandalSchwartz this was easy with two steps... but now I have three. :)
21:27 RandalSchwartz I can't use service.dead on pgpool, because we now know it's tty-picky
21:28 RandalSchwartz I guess I could just fully script it.
21:28 trevalen joined #salt
21:29 RandalSchwartz restart_postgresql as a cmd.wait
21:29 druonysuse joined #salt
21:29 RandalSchwartz and have it shut down apache, then pgpool
21:29 RandalSchwartz and then restart postgres
21:30 trevalen hello all, i am looking at salt for use in a very large environment.  for the multi-master setup, does this have to be explicity stated on all minions, or can i use a vip to front a fleet of masters?  not been able to find anything about using multi-master in this type of config.
21:30 RandalSchwartz I kinda just wanted the dependencies to figure that out
21:31 scoates joined #salt
21:31 UtahDave trevalen: what do you mean by a very large environment?
21:32 ajolo joined #salt
21:33 jbirdman joined #salt
21:34 savvy-lizard joined #salt
21:34 trevalen_ joined #salt
21:35 trevalen_ potentially growing to 100k servers
21:36 UtahDave ok.   So we have users that are using the Salt master behind a vip.  You just need to make sure that the masters all have the same pub/priv keys.
21:36 Hollinski joined #salt
21:36 UtahDave also ensure that their file_roots are identical, either by copying them all there yourself, or use gitfs
21:37 eliasp gitfs FTW! :)
21:37 UtahDave also you might look at using syndics if there are logical separations in those 100k, like if they are in different DCs and things like that.
21:38 trevalen_ the other question there then is when do you hit a ceiling on number of servers being managed via the vip?  is there a point at which snat from the masters via the vip is exhausted?
21:38 trevalen_ or are all conns client initiated?
21:39 eliasp trevalen_: as long as you don't use 'salt-ssh', all connections are client initiated
21:39 UtahDave trevalen_: Hm. I haven't explored that at that type of scale myself. I'm not sure.
21:40 UtahDave right now, we have users managing about 14K to 15K minions per master. with the master on pretty decent hardware.
21:40 trevalen joined #salt
21:40 trevalen thanks for the answers, it definitely gives me a better starting point to begin testing on
21:41 UtahDave The next version of Salt will have a new transport called "raet" that's supposed to handle quite an increase in the number of minions, but it's still in heavy development. beta still, really
21:41 trevalen i have to test using raet
21:41 martoss joined #salt
21:41 UtahDave trevalen: you're very welcome.  Let me know if you have any questions.
21:41 eliasp trevalen_: you might wanna watch this to learn a bit about RAET: https://www.youtube.com/watch?v=SI5J43UkarM
21:42 trevalen watched that, it is the inclusion of raet that makes salt a possibility
21:42 trevalen the use of the lgplv3 on zeromq makes management forbid it
21:42 troyready joined #salt
21:42 eliasp uh… wow…
21:43 forrest lol
21:43 trevalen got to love lawyers....
21:43 forrest what org forbids gpl3?
21:43 forrest crazy people
21:43 UtahDave trevalen: Yeah, we've had several enterprise customers have the same problem.
21:44 RandalSchwartz Hmm.  I borked my recipe by refactoring
21:44 RandalSchwartz a prereq: cmd: restart_pgpool doesn't inherit any of its watch/watch_in
21:44 RandalSchwartz that seems odd.
21:45 avn_ /etc/salt/master.d/*  have priority over /etc/salt/master or not?
21:46 trevalen thanks again for the answers, i am off to go start some testing now
21:46 RandalSchwartz I've worked plenty of places that are allergic to restrictive licences like GPL
21:47 avn_ Hm, or I can just run my test states masterless, via salt-call, and one container will be enought
21:47 RandalSchwartz that's why I encourage developers to use permissive licenses, like BSD or Apache
21:48 RandalSchwartz their code will be more widely used
21:48 UtahDave RandalSchwartz: yep. exactly.
21:49 forrest Yea my problem is when companies require that, then whine and moan when you want to contribute back what you wrte
21:49 forrest *wrote
21:49 matthiaswahl joined #salt
21:49 forrest how dare you try to contribute back to this thing we depend heavily upon!
21:50 codekobe_ Question about pillar data:
21:50 codekobe_ I am working on a remediation action for a system, that would change the contents of a config file during a period of high load.  I have an existing state for this system, that pulls config values form pillar data. I don't want to permanently modify the pillar data I just want to modify the values for this single run.  I am trying to figure out the best way
21:50 codekobe_ to do this.  So i had 2 ideas, but don't know if either are possible.  Can i pass values to override pillar data when I manually run the state?  Or Can i modify the pillar data being used by wrapping the use of the state in an execution module, and modify the pillar data used from within the execution module?
21:52 UtahDave codekobe_: yeah, you can add  pillar="{'pillar_thingy': 'my value'}"
21:53 UtahDave whiteinge: is that the correct pillar cli syntax?  ^^
21:53 forrest UtahDave, Pretty sure it is: https://github.com/saltstack/salt/issues/4467
21:54 forrest UtahDave, we should probably add that to the docs if it doesn't already exist.. If it does, I don't know where.
21:55 anotherZero joined #salt
21:58 ajprog_laptop1 joined #salt
22:00 codekobe_ i looked through the docs
22:00 codekobe_ and using the salt cli
22:01 RandalSchwartz Hmm.. had to add both "watch_in" and "prereq_in" to a given state.  weird.
22:01 codekobe_ i did salt -h and looked through the help menu
22:04 avn_ salt-call --id=$minion --local   (for each minion's id)  coould be a good testsuite for development?
22:05 savvy-lizard joined #salt
22:05 troyready joined #salt
22:05 codekobe_ thanks for the help guys, I love this channel!
22:09 rallytim_ joined #salt
22:10 rallytim_ joined #salt
22:16 oz_akan_ joined #salt
22:17 avn_ Guys, why some operations with apt works, and others say "/var/lib/dpkg/lock taken by other process"  during one salt state.highstate run?
22:17 forrest avn_, If it was an rpm system I'd say that's just because of the lock file ensuring only one thing can happen in the repo at a time
22:18 forrest I don't know if its' the same for apt.
22:19 avn_ it is ubuntu trusty running on EC2
22:19 avn_ no others users except me, and no other operations with apt except salt's one
22:20 rallytime joined #salt
22:30 forrest avn_, I'd suggest to start going through the states then to determine which one is causing the problem, or if the issue only occurs on a highstate.
22:36 RandalSchwartz yeah - solved my two-level down/up problem with brute force
22:36 yomilk joined #salt
22:41 troyready joined #salt
22:42 vejdmn joined #salt
22:47 lz-dylan Is there any way to manage mounted disks/mkfs using states (not modules)? I'd love for salt to automatically mkfs.ext4 and mount a volume created by salt-cloud and attached to the instance. If not, would a better way be to just create a blank container, format it, and build off that as a snapshot?
22:48 eliasp lz-dylan: I don't know of an mkfs-capable state yet, but I'd say writing a custom one on top of udisks2 (using it's D-Bus API) should be rather straightforward
22:50 matthiaswahl joined #salt
22:51 n8n joined #salt
22:51 lz-dylan eliasp: never heard of it! I'll read up.
22:52 eliasp lz-dylan: http://www.freedesktop.org/wiki/Software/udisks
22:57 avn_ Another randomly appeared issue: -- "Comment: The directory 'salt://reprepro/keys' does not exist on the salt fileserver in saltenv 'base'"
22:58 forrest does it exist?
22:58 avn_ exists of course (it error from file.recurse state)
23:01 lz-dylan is there an easy way to contribute to/comment on documentation?
23:01 Hell_Fire joined #salt
23:02 jayne joined #salt
23:02 lz-dylan (context: http://docs.saltstack.com/en/latest/topics/cloud/aws.html#configuration describes the example as 'any file in /etc/salt/cloud.providers.d/ directory', but salt-cloud appears only to read .conf files from that location, ignoring all else)
23:03 forrest lz-dylan, http://docs.saltstack.com/en/latest/topics/development/conventions/documentation.html talks about some of the stuff, and you can always create an issue if you don't want to contribute.
23:03 forrest well, contribute directly, the issues still help
23:03 jayne Anybody know why salt-jenkins keeps sending me (j) notices about builds failing?
23:03 forrest jayne, to your email?
23:03 jayne no, to me on IRC, as a notice
23:04 forrest jayne, didn't you have this issue before? Or was that another freenode staff member?
23:04 jayne salt-jenkins (~salt-jenk@166.78.186.222): Build Salt Rackspace Arch is STILL FAILING: http://jenkins.saltstack.com/job/salt-rs-arch/3771/
23:04 jayne forrest: um... I don't remember, to be honest
23:05 forrest jayne, ok no worries. I'll let s0undt3ch know, he runs the jenkins instance, sorry about that
23:05 jayne k, thanks
23:06 forrest jayne, yea np.
23:07 RandalSchwartz "capturing configurations... one file at a time..."
23:07 forrest he lives in Portugal so I don't know when he'll be around.
23:07 notpeter_ Can anyone provide me a copy of a working cloud profile for AWS with VPC that's working? I got it going on the "classic" instance types, but the newer instance types require VPC.
23:08 cheus joined #salt
23:08 notpeter_ When I follow the example here: http://docs.saltstack.com/en/latest/topics/cloud/aws.html I always get errors like "Network interfaces and an instance-level security groups may not be specified on the same request"
23:09 notpeter_ there's a note at the bottom at that page which I think obscurely acknowledges this, but I'm still not quite getting it.  "Note that it is an error to assign a 'subnetid' or 'securitygroupid' to a profile where the interfaces are manually configured like this. These are both really properties of each network interface, not of the machine itself."
23:09 Ryan_Lane notpeter_: I'm not totally sure any of them are working for VPC
23:09 Ryan_Lane notpeter_: we have some changes incoming for the boto states to fix this
23:10 notpeter_ Hmmm...ok.
23:10 avn_ Hmm. Another problem -- salt '*' state.highstate report me, than another highstate job is running
23:10 Ryan_Lane newer instance types should work fine in classic
23:10 Ryan_Lane I'm using c3s in classic
23:10 Ryan_Lane you just can't use certain availability zones. you need to use 1a 1d 1e
23:10 Ryan_Lane 1b and 1c are generally unavailable
23:11 bhosmer joined #salt
23:11 Ryan_Lane heh. the boto stuff I wrote also doesn't have support for single nodes (only autoscale groups)
23:12 Ryan_Lane notpeter_: oh, wow. the aws stuff has support for VPC already?
23:13 * Ryan_Lane hasn't used that yet
23:13 notpeter_ Documention suggests it does, it apparently does not have /working/ support.
23:13 Ryan_Lane heh
23:13 lz-dylan forrest: thank you!
23:13 forrest lz-dylan, np
23:14 Ryan_Lane well, that sucks.
23:14 Ryan_Lane all the boto stuff I wrote will eventually replace the aws stuff
23:14 Ryan_Lane but you'd need to use autoscaling groups
23:15 arthabaska joined #salt
23:16 Luke joined #salt
23:23 notpeter_ I figured out what that documentation was hinting at.  I had declared the security group at the provider level and at the network interface level.
23:24 Outlander joined #salt
23:24 joehillen joined #salt
23:24 Ryan_Lane ah
23:24 notpeter_ Second, apparently the t2 instances (t2.micro, small, medium) are apparently VPC only.
23:24 smcquay joined #salt
23:24 Ryan_Lane ah
23:25 smcquay joined #salt
23:25 avn_ Looks I do something wrong -- first pass of salt state.highstate put salt://reprepro/keys to  /srv/reprepro/keys as it should be, second pass complains that salt://reprepro/keys not exists
23:26 bmatt hm, very odd issue that's been reported but no resolution identified
23:26 bmatt I'm seeing "AttributeError: 'module' object has no attribute 'get_group_list'"
23:27 bmatt from the call "return salt.utils.get_group_list(name)" in useradd.py line 486
23:27 bmatt how is the salt.utils object populated?
23:28 RandalSchwartz looking at installing crons now
23:29 RandalSchwartz Hmm... doesn't look like @reboot or @daily are supported
23:30 RandalSchwartz unless minute: '@reboot' is valid
23:30 RandalSchwartz but then it would need to know not to put hour etc in
23:30 mosen joined #salt
23:30 ajprog_laptop1 joined #salt
23:32 RandalSchwartz oh... maybe I want cron.file
23:33 RandalSchwartz no - that replaces the whole file
23:36 ajolo joined #salt
23:39 RandalSchwartz cron.set_special supports @reboot... can I call that from a state?
23:40 RandalSchwartz maybe module.run?
23:40 lz-dylan So, in /etc/salt/master the default timeout is 5 seconds. I pretty frequently ran into minions not returning, so boosted it to 10, which usually works...but not on a fresh deploy. Is there any harm in a really long timeout, say 60s?
23:43 forrest lz-dylan, not really
23:43 lz-dylan cool. I like it when minions return :)
23:43 forrest lz-dylan, unless you screw something up and it hangs for a long time
23:43 forrest but that's usually user error
23:44 ipalreadytaken joined #salt
23:44 RandalSchwartz or you're trying to restart pgpool. :)
23:51 matthiaswahl joined #salt
23:52 lz-dylan I can spare a minute for my own problems :)
23:52 aquinas joined #salt
23:57 DaveQB joined #salt
23:59 ipalreadytaken joined #salt

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