Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-02-27

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

All times shown according to UTC.

Time Nick Message
00:00 allanparsons @leonardinius @corey - http://narrabilis.com/mybook/puppet/autosign.conf
00:01 Corey I'm aware of what it does.
00:01 leonardinius allanparsons, is it somewhere similar  http://docs.saltstack.com/ref/configuration/master.html#autosign-file?
00:01 skynet joined #salt
00:02 Xe joined #salt
00:02 allanparsons @leonardinius - perfect
00:02 allanparsons but not sure why that gets flagged as INSECURE but auto-accept: True doesnt get flagged as INSECURE in that documentation
00:02 leonardinius allanparsons, plus as I can see - you might try to use reactor events for that, not sure whether key requests events are emitted
00:03 AdamSewell joined #salt
00:04 leonardinius allanparsons, don't pay attention to me anyway, I'm complete salt rookie
00:05 baniir joined #salt
00:06 leonardinius Guys anyone - could please someone point me out to documentation / examples - how I might provision Amazon AWS instances in VPC? I see subnetid parameter, I have seen someone mentioned `network_interfaces`, but could not find any decent piece of doc / decent example.
00:07 baniir is config a supported option in service.enabled? i'm not seeing it in the docs
00:07 allanparsons 1 second leonardinus
00:07 allanparsons i have done that
00:07 allanparsons is there a paste link avail?
00:07 jpaetzel joined #salt
00:08 leonardinius allanparsons, ok. what do you mean by paste link?
00:08 leonardinius baniir, yes I believe it is
00:09 leonardinius baniir, I evene suceeded to install *.sh script as service that way on ubuntu
00:10 DaveQB joined #salt
00:10 allanparsons @leonardinius : http://pastie.org/private/ghihxdy6mohalsvuanw
00:10 pydanny joined #salt
00:10 baniir leonardinius: nice. have you ever seen an error like: "Service foo is already enabled, and is dead"? the result of service.enabled is false, and I'm not exactly sure what this means
00:11 david_a joined #salt
00:11 fllr joined #salt
00:11 leonardinius baniir, https://gist.github.com/leonardinius/9241614 is snippet I used
00:12 aleszoul3k joined #salt
00:13 ndrei joined #salt
00:13 bhosmer joined #salt
00:13 baniir leonardinius: i'm not seeing config
00:14 clearclaw joined #salt
00:14 allanparsons @leonardius can you start it manually?
00:14 allanparsons may be respawning
00:15 allanparsons oops, disregard. sorry.
00:15 leonardinius allanparsons, thanks a lot. will try it first thing tomorrow morning (it's 02 am my timezone now). Anyway, you saved my day. Thanks
00:16 leonardinius baniir, what do you mean by config ?
00:16 leonardinius baniir, what does -l all debug statement say?
00:17 allanparsons @leonardinius you have to run that with salt-cloud (i assume thats what you're trying to do)
00:18 allanparsons but to be honest, I liked the cloudformation templates a little more.  i'd recommend using those instead
00:18 ashb I'm a little bit confused how to structure my environments for pillars. If I *dont* specify an environment it deep merges how I'd expect. If i do set an environment then It doesn't
00:18 baniir leonardinius: as an example, see: https://github.com/saltstack/salt/issues/6875 (line 7 of the first code block)
00:18 leonardinius allanparsons, yes, exactly.
00:19 leonardinius allanparsons, cloudformation templates, hmm. I need to do a little research
00:20 allanparsons just dont write everything in one huge template
00:20 allanparsons stack them.
00:20 allanparsons we use a BaseVPC / Networking / Security Group template
00:20 allanparsons then we add a salt template
00:20 allanparsons then an elastic search template
00:20 allanparsons etc.
00:48 ilbot3 joined #salt
00:48 Topic for #salt is now Welcome to #salt | Thanks to everyone who made SaltConf 2014 amazing! | 2014.1.0 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/
00:49 saurabhs joined #salt
00:52 schimmy1 joined #salt
00:52 alunduil joined #salt
00:53 saurabhs left #salt
00:55 fllr joined #salt
00:57 whiteinge xDamox: if you're talking about a top.sls file for salt states, that gets rendered on the minions so you can call out to the local system via jinja
00:57 ajw0100 joined #salt
00:59 xDamox hello whiteinge, I have a directory on the master that has pillar data e.g. /srv/pillar/important/file1, file2, file3 etc
00:59 krow joined #salt
00:59 xDamox I wanted to ls /srv/pillar/important and the results I wanted to loop over and render the top.sls
01:01 whiteinge xDamox: ah, gotcha. yes. you can put jinja in the pillar top file
01:02 seapasulli joined #salt
01:02 Ryan_Lane1 joined #salt
01:03 xDamox whiteinge, do you know how I could list the /srv/pillar/important directory and capture all the filenames that I can then loop over?
01:05 TheRealBill left #salt
01:07 whiteinge xDamox: something like {% for file in salt['cmd.run']('ls /srv/pillar/important').splitlines() %}
01:09 xl1 joined #salt
01:10 ipalreadytaken joined #salt
01:11 xDamox whiteinge, cmd module is disabled for security reasons
01:11 snuffeluffegus joined #salt
01:11 clearclaw joined #salt
01:13 kvbik what is the first thing you'd check, if a command on one minion runs in 3 seconds and on all of them (just 65 minions) you have to wait about 2min? salt==0.17.4
01:13 jpcw_ joined #salt
01:14 whiteinge xDamox: http://docs.saltstack.com/ref/modules/all/salt.modules.file.html#salt.modules.file.readdir
01:14 whiteinge you might have to manually trim up some of the output though
01:14 sgviking joined #salt
01:15 whiteinge xDamox: or you could write a custom two-line execution module, i suppose. :)
01:15 justlooks joined #salt
01:16 liwen joined #salt
01:16 zaz\ joined #salt
01:16 whiteinge kvbik: during that time are you waiting for the ``salt`` CLI command to return? it might be just one wonky minion causing things to hang
01:17 justlooks hi, i find salt master will cache the init.sls content , if i modify my sls file then test it ,it will cause error ,how to handle this satuation?
01:17 whiteinge kvbik: you could watch salt's event bus to see which minions return at what time
01:17 kvbik whiteinge: sure. and how would fing that one?
01:17 kvbik whiteinge: ah.. can you point me to that one?
01:17 xDamox Thanks whiteinge, I think I'll write the custom module
01:18 xDamox :)
01:18 whiteinge kvbik: https://github.com/saltstack/salt/blob/develop/tests/eventlisten.py
01:18 whiteinge kvbik: this also may be helpful: http://docs.saltstack.com/topics/reactor/index.html#knowing-what-event-is-being-fired
01:19 fllr joined #salt
01:20 kvbik whiteinge: thx. going to check this!
01:22 Math` joined #salt
01:25 mgw joined #salt
01:25 david_a joined #salt
01:27 taion809 joined #salt
01:28 thayne joined #salt
01:38 justlooks hi,anyone can solve this ? https://gist.github.com/justlooks/9242568
01:39 srage joined #salt
01:41 justlooks why the commented state still get work?
01:42 Ryan_Lane joined #salt
01:45 zdong justlooks, i did not really go through what you did and i am new to salt too. but i notice you mispell init.sls in /opt/pillar/mytest/iniit.sls .
01:46 srage joined #salt
01:47 zdong timoguin, did you get a chance to test module on git repo ?
01:48 kingloger joined #salt
01:54 justlooks zdong: sorry for the mistake , i correct it ,still same problem
01:54 hillna joined #salt
01:55 justlooks how can i upload a file which has user execution permission
01:57 justlooks i find there is a mode option in file.managed
01:57 krow joined #salt
01:58 mgw joined #salt
01:58 quickdry21 joined #salt
02:01 vejdmn joined #salt
02:02 Ryan_Lane joined #salt
02:03 quickdry21_ joined #salt
02:13 kvbik whiteinge: not sure if i've found out anything in the event bus. just the thing that there is almost 0.5s between all "requests" and also "responses" which can lead to the time i am getting to...
02:14 kvbik i've tried to run the command manually sequitially - on all minions after each other and no one took much more than the 3secs...
02:15 Math` joined #salt
02:15 justlooks how to call a module in sls file?
02:16 krow joined #salt
02:20 rojem joined #salt
02:20 Furao joined #salt
02:20 zdong justlooks, see this on how to run execution module from sls. http://docs.saltstack.com/ref/states/all/salt.states.module.html#module-salt.states.module
02:28 ajw0100 joined #salt
02:30 sinh joined #salt
02:36 yomilk joined #salt
02:40 pydanny joined #salt
02:44 n8n joined #salt
02:49 krow joined #salt
03:07 Networkn3rd joined #salt
03:15 pydanny joined #salt
03:20 AdamSewell joined #salt
03:22 l0x3py joined #salt
03:22 pydanny joined #salt
03:23 rostam joined #salt
03:27 HeadAIX joined #salt
03:28 Katafalkas joined #salt
03:36 xmltok joined #salt
03:38 Daemonik joined #salt
03:39 Daemonik In the page http://docs.saltstack.com/ref/states/all/salt.states.gem.html, what is "addressable:", I don't understand that line.
03:53 n8n joined #salt
03:53 rojem joined #salt
03:54 rojem joined #salt
03:55 rojem joined #salt
03:59 forrest joined #salt
04:01 favadi joined #salt
04:01 luminous I don't quite understand why the minion gets shafted so many times
04:02 luminous why doesn't the minion (or salt-call --local) get to use master_tops??
04:03 tvle joined #salt
04:06 mgw joined #salt
04:08 Linz joined #salt
04:10 UtahDave joined #salt
04:26 kedo39 joined #salt
04:31 clearclaw joined #salt
04:34 liwen joined #salt
04:35 saurabhs joined #salt
04:37 AdamSewell joined #salt
04:39 liwen joined #salt
04:45 zain joined #salt
04:50 jeffro joined #salt
04:50 n8n joined #salt
04:51 dwfreed joined #salt
04:51 krow joined #salt
04:53 yomilk joined #salt
04:55 Ryan_Lane joined #salt
05:00 ajw0100 joined #salt
05:01 Daemonik joined #salt
05:01 pydanny joined #salt
05:05 fllr joined #salt
05:06 smcquay joined #salt
05:09 pydanny joined #salt
05:11 seapasulli joined #salt
05:13 rgbkrk joined #salt
05:25 DaveQB joined #salt
05:28 smcquay joined #salt
05:29 Math` joined #salt
05:35 zzzirk joined #salt
05:44 anuvrat joined #salt
05:49 damianh` joined #salt
05:51 krow joined #salt
05:51 damianh` left #salt
05:57 thayne joined #salt
05:57 pydanny joined #salt
05:58 schimmy joined #salt
05:59 krow joined #salt
06:02 ChaosPsyke joined #salt
06:07 schimmy joined #salt
06:20 UtahDave joined #salt
06:34 pydanny joined #salt
06:38 cedwards1 joined #salt
06:38 Katafalkas joined #salt
06:40 dcolish joined #salt
06:41 schimmy joined #salt
06:49 krow joined #salt
06:54 fllr joined #salt
06:54 smcquay joined #salt
06:54 joehillen joined #salt
06:56 Ryan_Lane2 I have a really weird issue
06:56 Ryan_Lane2 I have a runner that runs service.restart batched against a grain target
06:56 Ryan_Lane2 it runs against all targets
06:57 Ryan_Lane2 but if I run publish.runner from another node, it runs against all targets minus one
06:57 Ryan_Lane2 consistently minus one, but never the same one is missing
06:59 kolaman joined #salt
07:04 lessismore joined #salt
07:06 astol joined #salt
07:07 fllr joined #salt
07:13 gammalget joined #salt
07:13 Ryan_Lane2 ah hah: https://github.com/saltstack/salt/issues/8678
07:13 Ryan_Lane2 https://github.com/saltstack/salt/issues/6456
07:17 Rojematic joined #salt
07:18 slav0nic joined #salt
07:25 liwen joined #salt
07:25 ndrei joined #salt
07:27 rojem joined #salt
07:30 leonardinius joined #salt
07:31 nn0101 joined #salt
07:32 Iwirada joined #salt
07:36 leonardinius_ joined #salt
07:37 fllr joined #salt
07:39 fllr joined #salt
07:40 johtso joined #salt
07:44 leonardinius joined #salt
07:51 ndrei joined #salt
07:55 kedo39 joined #salt
07:58 MedicalJaneParis joined #salt
07:59 schimmy joined #salt
08:00 Kenzor joined #salt
08:07 fllr joined #salt
08:08 naiyte joined #salt
08:09 leonardinius joined #salt
08:13 linjan_ joined #salt
08:14 elfixit joined #salt
08:14 Ryan_Lane2 it's not even that
08:15 Ryan_Lane2 it's simply that publish.runner doesn't have a timeout and calls to it aren't turned into jobs on the master
08:15 Ryan_Lane2 which really sucks.
08:15 Ryan_Lane2 it means calling publish.runner is completely unreliable
08:15 Hipikat left #salt
08:15 martoss joined #salt
08:16 Ryan_Lane2 unless you do at most a single salt call
08:16 Ryan_Lane2 the individual calls get turned into jobs
08:17 harobed joined #salt
08:23 Katafalkas joined #salt
08:31 kadel joined #salt
08:32 yomilk joined #salt
08:34 jeffro joined #salt
08:37 miguitas joined #salt
08:42 zooz joined #salt
08:49 scott_w joined #salt
08:52 alienresidents joined #salt
08:52 fishpen0 joined #salt
08:52 rojem joined #salt
08:55 justlooks hi,how can i run awk '$4~/19\/Nov\/2013/&&$6~/"GET/&&$7~/\/aywin\?/{print substr($4,2)" "$7}' /tmp/mytest > /tmp/myout in cmd.run ' xxx '?
09:01 jeddi holy dooly.   justlooks I'd push the command out into a .sh file, and run it that way.  :)
09:01 justlooks jeddi: OK
09:02 jeddi justlooks: i'd just be worried about escaping that safely otherwise.
09:02 jeddi (I'm sure it could be done ... just not by me :)
09:05 Daemonik joined #salt
09:05 giantlock joined #salt
09:05 ndrei joined #salt
09:05 jeffro joined #salt
09:05 jeffro joined #salt
09:39 n8n joined #salt
09:39 17SAAKRNV joined #salt
09:45 N-Mi joined #salt
09:45 N-Mi joined #salt
09:49 martoss1 joined #salt
09:49 martoss joined #salt
09:50 carlos joined #salt
09:51 mortis joined #salt
09:52 sfello joined #salt
09:55 Valdo joined #salt
09:56 aleszoul3k joined #salt
09:57 martoss1 joined #salt
10:02 che-arne joined #salt
10:08 astol joined #salt
10:08 fllr joined #salt
10:15 calvinx joined #salt
10:16 alexj joined #salt
10:17 mike25ro joined #salt
10:19 TheRealBill_here joined #salt
10:20 lessismore joined #salt
10:26 yomilk joined #salt
10:26 jinnko_ joined #salt
10:32 babilen Hi all - I am quite partial to environments and corresponding git branches, but am unsure how to deal with my top file. I am, in particular, unsure how to come up with a sensible merging strategy as I don't really want to merge from master (base) into dev but only the other way round. Using different top files in the branches is also not really an option as that would mean that changes to those will also be merged.
10:32 babilen (all using GitFS, yes)
10:33 yomilk joined #salt
10:35 ndrei joined #salt
10:40 xmj joined #salt
10:40 xmj Hi, how do I reference a pillar from within a salt state?
10:40 xmj can I simply do pillar['name'].entry
10:40 xmj ?
10:42 tmmt xmj: salt['pillar.get']('name:entry')
10:43 babilen xmj: There are various ways (e.g. pillar['foo']['bar'], pillar.get(...), ...) -- I'd recommend to read http://docs.saltstack.com/topics/pillar/
10:48 mike25ro xmj: i also use pillar[x][y] ..depending on how many layers you have .. so for a pillar:   myPillar: myValue   i use   {{ pillar['myPillar'] }}
10:49 xmj well, FN_derp: - hostname derp \n - ip: 192.168.0.1 should be pillar['FN_derp']['ip'] and ..['hostname'] respectively
10:50 xmj the one thing i was wondering is if I could save myself the last few brackets and go with .ip directly.
10:54 mike25ro i think is safer with the [][] ... i know i used the .foo.bar ... but at some point i needed to switch to [][]... maybe because of the naming conventions.. or smth like that xmj
10:55 martoss joined #salt
10:59 TamCore Are there ways to use cmd.run's output to create grains/pillars? For example: simply reading a file on minion and passing it's content into a grain?
11:01 xmj mike25ro: the question is, does it >work< within salt states?
11:02 mike25ro xmj i remember i used that in states of course
11:04 xmj k
11:10 Math` joined #salt
11:10 david_a joined #salt
11:12 topochan hi everyone, I am developing the nagios module and one of the issues are to configure plugin path
11:14 topochan I am trying to solve this issue with a pillar data but I ca not use __salt__ before the module is loaded, is there any other way to get the pillar data before the module is loaded?
11:27 zzzirk joined #salt
11:27 beardo_ joined #salt
11:29 Math` joined #salt
11:30 quickdry21 joined #salt
11:31 keith4 joined #salt
11:33 MrTango joined #salt
11:41 psyl0n joined #salt
11:42 bhosmer joined #salt
11:43 diegows joined #salt
11:46 sgviking joined #salt
11:49 lessismore joined #salt
11:51 oznt joined #salt
11:52 jdenning joined #salt
12:01 Math` joined #salt
12:07 ndrei joined #salt
12:08 londo joined #salt
12:09 fllr joined #salt
12:19 liwen joined #salt
12:24 baoboa joined #salt
12:34 yomilk joined #salt
12:38 clearclaw joined #salt
12:51 ggoZ joined #salt
12:51 akoumjian joined #salt
12:54 zzzirk joined #salt
13:02 viq Is there any ETA on debian and centos packages?
13:03 johtso joined #salt
13:05 xmj joehh is the right person to ping on that one, i think
13:06 joehh viq: just trying to solve some build issues now
13:07 joehh debian not centos
13:07 viq OK, thanks
13:10 mike25ro hey guys .. i am using salt 0.17.1 and sometimes when i run highstate for a minion on master i get no returned data... if i run it one more time, it works. I get this on more minions .... any idea what might be the cause?
13:15 rizumu joined #salt
13:22 fishpen0 joined #salt
13:22 anuvrat joined #salt
13:28 Nazca joined #salt
13:28 babilen mike25ro: Anything special about your setup or otherwise remarkable? GitFS for example?
13:33 elfixit joined #salt
13:33 gnugnu6 joined #salt
13:37 jeddi mike25ro: and what's the network transit times like?
13:38 jeddi mike25ro:  you can try the show timeout option ... may provide some insight:   salt --show-timeout {host} state.highstate
13:38 ipmb joined #salt
13:39 jeddi i see it semi-regularly, usually because either the master takes too long to compile the state, and/or the minion takes too long to complete.  the minion actually is still heading for highstate in the background (you'll get told that if you try to run the highstate again).   you can monitor the activity on the minion at the time (top, nmon, etc) to see if that's it.  i think there's some timeout values in /etc/salt/master that you
13:39 jeddi could play with.
13:42 leonardinius_ joined #salt
13:42 ndrei joined #salt
13:42 leonardinius__ joined #salt
13:42 mike25ro jeddi thanks .. will look into that
13:44 harobed joined #salt
13:49 jeddi mike25ro: just checked mine - i bumped my timeout up to 20s a while back, as i run my salt master from digital ocean, but lots of my minions are down residential grade (ie. cruddy) adsl links, and they'd frequently do as you have obesrved.
13:49 faldridge joined #salt
13:51 TamCore I see the same behaviour every day in my setup. Mostly when I try to run highstate on a newly accepted Minion. Network shouldn't be a problem in my case because all Minions and my Master are sitting in the same rack (behind the same switch and partly on the same hardware node)
13:52 quickdry21 joined #salt
13:56 jslatts joined #salt
13:56 vejdmn joined #salt
13:57 ahammond joined #salt
13:58 gildegoma joined #salt
14:02 juicer2 joined #salt
14:05 lessismore joined #salt
14:05 che-arne joined #salt
14:07 zzzirk joined #salt
14:09 harobed_ joined #salt
14:10 harobed_ joined #salt
14:10 fllr joined #salt
14:11 ipmb joined #salt
14:13 musonic joined #salt
14:13 musonic Hi all.
14:13 musonic I'm just getting started with salt and could really do with some help
14:13 alexj joined #salt
14:13 musonic i'm wanting to use salt to provision my vm in partnership with vagrant
14:14 faldridge joined #salt
14:14 musonic one thing i'm not sure about at the moment is how i know the name of the packages that can be installed. for example, if i want to install mysql how do i know the package is called mysql and not something else like mysql-client or whatever?
14:14 Saint-N_ joined #salt
14:14 musonic is there a list of available packages somewhere?
14:15 saint-n joined #salt
14:15 Mr_N musonic: You can do that in a few ways. (the proper package naming thing)
14:16 Mr_N There are some good examples in the official salt docs, I'm skimming briefly to see if I can find that for you
14:16 Mr_N But basically, normally the pkg: installed statement is sufficient, and the package manager handles it
14:17 Mr_N but in the case that there's still some naming mismatch, you can explicitely specify the package name to use; and condition this on e.g. os_family
14:17 musonic thanks Mr_N. I've been reading through the docs and all I could find was that you need to use the correct package name for your os
14:17 bryano_ joined #salt
14:17 musonic but which package manager does salt use by default? apt-get?
14:18 Mr_N Different ones for different OS's.
14:18 Mr_N So, I stole this example from the docs for my own use, having trouble finding the original doc it's from, but it looks something like:
14:18 Mr_N vim: pkg: -installed {%if grains['os_family'] == 'RedHat' %} -name: vim-enhanced
14:19 musonic ok. so my vagrant setup is installing Ubuntu 12.04.01 LTS
14:19 Mr_N so it'd be using apt.
14:20 musonic Mr_N so if I find any Ubuntu package I can install it via salt using the apt package name?
14:20 Mr_N Yup.
14:20 musonic Mr_N. cool. that makes sense!
14:20 musonic thanks
14:20 rgbkrk joined #salt
14:20 musonic i haven't got as far as grains yet...
14:20 Mr_N Keep in mind, naturally, the apt package name wouldn't work on a RHEL system (as the snipped I gave earlier specified)
14:21 Mr_N that's why you need to utilize grains and such to select properly
14:23 racooper joined #salt
14:23 Ztyx joined #salt
14:24 faldridge joined #salt
14:25 Ztyx Looking to fetch external Pillars from a subdirectory in a Git repository. Is that possible? I'm looking for equivalent of "gitfs_root", that is.
14:25 HeadAIX joined #salt
14:25 Kenzor_ joined #salt
14:25 kermit joined #salt
14:26 MTecknology I have a *lot* of these happening... http://dpaste.com/1666173/   I can't figure out exactly why. It only happens in fil.{managed,recurse} states.
14:26 fllr joined #salt
14:26 Mr_N I'm still trying to figure out why refresh_pillar doesn't cause pillar.raw and pillar.items to converge...
14:27 fllr joined #salt
14:28 gcbirzan what version of salt should I use, as a rule?
14:29 mpanetta joined #salt
14:29 MTecknology eh... looks like thatch and cro are working on it..
14:29 Mr_N 0.17.5, unless I'm talking out of my ass?
14:30 gcbirzan salt-minion is incredilby broken on hosts whose hostname is localhost
14:30 gcbirzan in 0.17.5
14:30 gcbirzan I'm gonna go for the git version :P
14:30 * Shish is finding 2014.1.0 to be nice and have lots of very useful features, but not packaged for redhat yet :(
14:30 Mr_N That's fun.  why is that a thing?
14:30 gcbirzan because, in master, shit's fixed
14:31 gcbirzan if you're talking about locaqlhost, it's because I want the IPs to be the minion_id
14:31 clearclaw joined #salt
14:31 gcbirzan wee, works :)
14:31 Shish Speaking of which, anyone know when 2014.1.0 should hit EPEL / is there anything I can do to make it happen? (erjohnso?)
14:31 Mr_N 2014.1 hitting EPEL would make my life wonderful, if it does indeed "fix shit"
14:31 xmj Shish: "wait"
14:32 xmj Shish: http://dl.fedoraproject.org/pub/epel/testing/6/x86_64/ if i
14:32 xmj 'm not mistaken it's there somewhere
14:32 Networkn3rd joined #salt
14:33 Shish aha, in epel-testing already \o/  will wait then :)
14:34 Mr_N I'm almost tempted to push to that to see if it addresses my pillar syncing issue...
14:35 liwen joined #salt
14:35 Kenzor joined #salt
14:35 JasonSwindle joined #salt
14:36 jaimed joined #salt
14:38 faldridge joined #salt
14:40 seapasulli joined #salt
14:41 nmap911 joined #salt
14:42 nmap911 hi. anyone here?
14:45 rawtaz are you going to sue me if i am?
14:45 manfred yes
14:49 timoguin joined #salt
14:50 gnugnu6 joined #salt
14:52 mgw joined #salt
14:52 AdamSewell joined #salt
14:52 AdamSewell joined #salt
14:53 Mua joined #salt
14:54 oznt is there a quicker alternative to 'high.state' ? what if I want to run the stuff related only to webserver?
14:55 Shish state.sls <statename>
14:55 JasonSwindle oznt:  You can target just the webserver
14:55 JasonSwindle Does the web server have something in it's name or a grain?
14:55 oznt but it will do all the stuff that is already done ?
14:56 ggoZ left #salt
14:56 ggoZ joined #salt
14:56 oznt JasonSwindle, to be clear: i followed the tutorial, and now I have srv/salt/webserver and below I have vts
14:57 Shish "salt mywebserver.foo.com state.sls apache" will only run the stuff in apache.sls and ignore everything else
14:57 backjlack joined #salt
14:57 GradysGhost joined #salt
14:57 jaimed joined #salt
14:57 JasonSwindle or salt '*webserver*' state.highstate
14:57 Mr_N Anyone here very familiar with pillar?
14:57 JasonSwindle There are many ways to target just one system, but I am not sure when one you picked. :)
14:58 GradysGhost Mr_N: There's probably a few of us.
14:58 JasonSwindle GradysGhost:  Yo
14:58 GradysGhost Hey JasonSwindle, how goes it?
14:58 Mr_N GradysGhost: Feel free to shout out that I'm just being a stupid; but I have a deployment for which pillar.items and pillar.raw return two different things
14:58 Mr_N and refresh_pillar doesn't cause them to converge.
14:58 backjlack joined #salt
14:59 Mr_N Neither does sync_all, but AFAIK this is known.
14:59 oznt Shish, thank you
14:59 GradysGhost Mr_N: There's a note in the documentation that suggests that may sometimes be the case. http://docs.saltstack.com/ref/modules/all/salt.modules.pillar.html#salt.modules.pillar.raw
14:59 GradysGhost "Contrast with items() which calls the master to fetch the most up-to-date Pillar."
15:00 Mr_N Indeed.  I'm just looking on how to get them to converge.
15:00 GradysGhost So it sounds like maybe you need to highstate the minion.
15:00 Mr_N since I need to run get() and I can't do that on the outdated stuff.
15:00 GradysGhost You have a master/minion setup?
15:00 Mr_N Indeed
15:00 GradysGhost Try running something like "salt 'target*' state.highstate"
15:00 Mr_N My understanding from the documentation was that refresh_pillar was exclusively what was needed to sync the pillar data
15:00 GradysGhost on the master
15:00 GradysGhost hrm, but that doesn't work for you?
15:01 Mr_N I believe I've highstated since this problem occured, is the thing.  I'm trying again now, I have 0 memory ><
15:01 che-arne joined #salt
15:01 GradysGhost If that fails, I'd say restart the minion into debug mode: salt-minion -ldebug
15:01 Mr_N Yeah; no luck.
15:01 GradysGhost And then try again and see if there's any interesting logging.
15:02 Mr_N Fair nuff.  Trying that now.  Thanks for the pointer :)
15:02 toastedpenguin joined #salt
15:02 GradysGhost I'm going out for a bagel, but I'll follow up in a few minutes if nobody else here does.
15:02 toastedpenguin left #salt
15:02 Mr_N Thanks a lot :) Enjoy your foods.
15:03 Mr_N ... That's some interesting outcome.
15:03 Mr_N starting it in -ldebug fixed the problem.
15:03 R_Shackleford joined #salt
15:03 Mr_N I may be hallucinating, but I could have sworn I restarted the service multiple times trying to fix this
15:04 babilen Hi all - I am using a, very simply, recipe http://paste.debian.net/84375/ and now started to use different environments. How would I deal with this most idiomatically?
15:04 Mr_N This does beg the question, however, if there's a better way to refresh the pillars than knocking over the minion.
15:04 Mr_N And why refresh_pillar (which would seem to claim to do that) wasn't doing so.
15:07 pdayton joined #salt
15:08 mortis_ joined #salt
15:09 Mr_N (I've managed to reproduce the situation, but running salt-minion in -ldebug, in case there's anything I could poke that might be more informative)
15:10 Hefeweizen joined #salt
15:10 Mua_ joined #salt
15:11 diegows joined #salt
15:12 oznt Shish, i tried: sudo salt 'backoffice*' state.sls vts, but I am getting an error I dont understand:   No matching sls found for 'vts.sls' in env 'base'
15:13 jpcw joined #salt
15:13 che-arne joined #salt
15:13 JasonSwindle1 joined #salt
15:14 kedo39 joined #salt
15:15 luminous oznt: salt isn't finding your vts.sls
15:15 _Flusher joined #salt
15:15 oznt luminous, so I have /srv/salt/webserer/vts , and salt is not looking into webserver?
15:16 Mr_N If it is not specified in file_roots, no.
15:16 wendall911 joined #salt
15:16 kvbik_ joined #salt
15:16 arnoldB_ joined #salt
15:16 mephx joined #salt
15:16 fllr joined #salt
15:16 Heartsbane_ joined #salt
15:16 Mr_N Actually; this is something I had some confusion on, so maybe I shouldn't speak up
15:16 babilen_ joined #salt
15:17 luminous oznt: try state.sls webserver.vts
15:17 luminous oznt: /srv/salt is your file_roots
15:17 mikkn_ joined #salt
15:17 luminous so every .sls is relative to that
15:17 harobed__ joined #salt
15:17 luminous if you have foo/bar.sls, that translates to foo.bar
15:18 luminous and foo implies foo/init.sls
15:18 oznt luminous, thanks. it's not that clear from the docs. Although I saw that when configuring the windows repository
15:18 jeddi note that foo.bar is how you refer to those state (files) within your other state files, too, as ithappens (eg:  for 'include')
15:18 luminous right, that's how we reference .sls
15:19 babilen Sorry, Freenode hickup (hence potential double-post): I am using a, very simple, recipe http://paste.debian.net/84375/ and now started to use different environments. How would I deal with this most idiomatically?
15:19 Daemonik__ joined #salt
15:19 vbabiy joined #salt
15:21 vipuls joined #salt
15:22 cedwards joined #salt
15:22 funzo_ joined #salt
15:23 rostam joined #salt
15:26 oznt luminous, I have cmd.wait:  test -L /etc/apache2/sites-enabled/mywebsite.com, in my vts/init.sls salt show that it passed, but on the machine I don't see that link. What am I missing?
15:27 baoboa joined #salt
15:27 luminous oznt: use file.symlink to create the link instead of cmd.wait
15:28 luminous but then I'm guessing about what you're doing
15:28 JasonSwindle1 left #salt
15:28 lahwran joined #salt
15:28 luminous I rarely have luck with cmd.run/wait respecting/reflecting my intentions
15:28 luminous with respect to that test
15:28 oznt luminous, I want to issue the command a2ensite, if i don't find the link already
15:29 racooper joined #salt
15:29 luminous oznt: does a2ensite do anything more than create a symlink?
15:29 oznt so my expectation is not to find /etc/apache2/sites-enabled/mywebsite.com, hence the command should run
15:29 oznt luminous, no
15:29 luminous oznt: then just create the symlink with file.symlink state
15:29 luminous that is part of the point to salt
15:29 oznt but's it's the way to do it in debian
15:29 luminous embrace states
15:30 luminous oznt: fine, then do that and wrestle with cmd.run's test
15:30 luminous I am no good there, sorry
15:30 oznt I tried folloiwng this: http://docs.saltstack.com/ref/states/highstate.html, look for a2ensite on this page, you will see what I copied to my stated
15:30 oznt *state
15:30 R_Shackleford joined #salt
15:30 homelinen joined #salt
15:31 cetex joined #salt
15:31 luminous oznt: what was the output of the command that was successful;?
15:31 luminous can you set --log-level=debug and review the output?
15:32 babilen Hmm, I am still not sure why I get "Detected conflicting IDs, SLS IDs need to be globally unique" or rather what I could do about it. How do you deal with identical states in different environments?
15:32 luminous babilen: when do you get that error?
15:33 luminous you should not be including those identical sls in one run
15:33 oznt http://paste.debian.net/84387/
15:33 babilen luminous: http://paste.debian.net/84375/ -- Just started playing with environments (I want a "dev" branch for testing in vagrant)
15:33 bt joined #salt
15:33 cofeineSunshine joined #salt
15:33 oznt luminous, i passted also...
15:33 tyler-baker joined #salt
15:33 luminous babilen: I've never used env, my advice is not worth too much there
15:34 babilen luminous: I simply lack all intuition and am not sure how to start dealing with this :)
15:34 luminous oznt: oh, you're command is in a good state, but it isn't executing
15:34 luminous salt doesn't think it needs to run it
15:35 lessismore joined #salt
15:35 seapasulli joined #salt
15:35 oznt that is why i am wondering: test -L etc/apache2/sites-enabled/vts.cfg  on backoffice returns 1
15:36 luminous oznt: no, you are using cmd.wait, and your watch isn't telling salt it needs to apply the state
15:36 luminous your state isn't even evaluating the test
15:36 luminous is my guess
15:37 [diecast] joined #salt
15:37 platforms joined #salt
15:38 R_Shackleford joined #salt
15:39 schimmy joined #salt
15:39 psyl0n joined #salt
15:40 mgw joined #salt
15:41 Mr_N Actually; this is something I had some confusion on, so maybe I shouldn't speak up
15:41 Mr_N Ak
15:42 Mr_N sorry, scrolled up while reading scrollback.
15:42 Mr_N Ignore me.
15:44 Ztyx1 joined #salt
15:44 gfa how to set defaults for a file.managed on pyobjects?
15:45 oznt luminous, I know what is the problem, if the file is not changes, the command won't run. So I moved to cmd.run instead of cmd.wait
15:45 luminous oznt: exactly
15:46 oznt now my question is can I run cmd.run with a condition that it is only run when the previous cmd.run was successful ?
15:46 fllr joined #salt
15:46 tmwsiy joined #salt
15:46 luminous oznt: that's cmd.wait with watch set
15:46 forrest joined #salt
15:47 tmwsiy I am looking to replace my monitoring system and was looking at Thomas's salt-monitor. Is this a good bet at this point or will this likely get dumped/changed drastically?
15:48 psyl0n joined #salt
15:48 tessellare joined #salt
15:48 forrest tmwsiy, salt-monitor hasn't been updated in nearly 2 years.
15:48 luminous tmwsiy: what are your goals with respect to monitoring? use that to determine which options are the better choices
15:49 tmwsiy well I am just looking for something more modern than nagios, was looking at shinken
15:49 mattmtl joined #salt
15:49 forrest heh
15:49 forrest oh nagios
15:50 tmwsiy like the idea of using zeromq so was looking for something in the salt space
15:50 gfa tmwsiy: shinken is cool, i use it
15:51 fllr joined #salt
15:52 Mr_N forrest: The sad part is, I haven't found many better options :(
15:53 Mr_N icinga is looking promising, though.
15:53 forrest Mr_N, to nagios?
15:53 forrest isn't icinga just a fork of nagios though?
15:53 Mr_N Indeed.
15:53 Mr_N (to both)
15:53 Mr_N It's also far more polished.
15:53 forrest does adding rules with it still suck like nagios?
15:53 fllr joined #salt
15:54 oznt luminous, I have some syntax error, i don't know how to tell one command to watch another, can you check this: http://paste.debian.net/plain/84393?
15:54 Mr_N No idea, actually.  I've been focused mostly on learning Salt, my coworker has been the monitoring guru
15:54 Mr_N I know it has some sort of a RESTful API
15:54 Mr_N no idea whether it's in any way useful for programmatic modification. (which would make the rule adding thing way simpler)
15:55 forrest oznt - watch in your cmd.wait should be - watch: \n  - cmd: a2ensite vts.cfg
15:55 rjhintz joined #salt
15:55 ipmb joined #salt
15:55 forrest Yea I don't mess with monitoring much, I'd lik to see how shinken is since it's written in python
15:56 tmwsiy seems like salt would be perfect to have a co-existent monitoring framework built on top I saw that salmon project that started out on salt but then moved away
15:56 Mr_N shinken, eh.  I have no idea if that was on our list of "technologies to vette" but I've made a note of it.
15:56 tmwsiy forrest: shinken is billed as a nagios fork rewritten in python
15:56 kula a co-worker of mine looked at shinken, didn't think it was ready for prime-time yet. we've been focusing on icinga here.
15:56 oznt forrest, I changed it, now I am seeing: " The require or watch statement in state "/etc/init.d/apache2 reload" in sls "webserver.vts" needs to be formed as a list..."
15:56 kula which is sad, i liked the concepts behind shinken. and the fact it is in python
15:57 forrest oznt, it's probably the space
15:57 oznt sorry, yes ...
15:57 forrest oznt, try this https://gist.github.com/gravyboat/9252904
15:58 oznt I love salt :-)
15:58 forrest yea it's pretty awesome
15:58 oznt and with this fuzzy nice feeling, i am going home! a they say in germany: Feierabend!
15:58 forrest have a good evening!
15:58 oznt thank you all!
16:00 Mr_N Enjoy
16:00 Mr_N kula: That's somewhat similar to other opinions I've been hearing, FWIW.
16:00 Iwirada left #salt
16:02 frasergraham joined #salt
16:03 mgw joined #salt
16:05 tmwsiy ok guess I will stick with nagios for now.. its at least a VERY WELL KNOWN quantity lol
16:05 RobSpectre joined #salt
16:06 frasergr_ joined #salt
16:06 RobSpectre qq - have files specific to  a host that I need managed.  Have them in a pillar, but that appears to not be installing them...
16:07 RobSpectre What is the canonical method?  a host specific salt state?
16:07 fraserg__ joined #salt
16:11 dave_den joined #salt
16:12 musonic joined #salt
16:13 tvle joined #salt
16:13 Linz joined #salt
16:14 tl-ml joined #salt
16:15 forrest RobSpectre, a host specific salt state would probably be the best. If you could instead create a 'general' state which then uses a templatized managed file (for reuse if needed) that would be even better.
16:15 jeremyBass joined #salt
16:15 RobSpectre forrest: check.  Much obliged.
16:16 forrest np
16:17 clearclaw joined #salt
16:20 bhosmer joined #salt
16:21 liwen joined #salt
16:21 clemenko joined #salt
16:21 bhosmer joined #salt
16:22 Ztyx1 left #salt
16:24 ndrei joined #salt
16:26 jalbretsen joined #salt
16:27 Mr_N Anyone have good examples of using pillar match inside a .sls to deploy a state?
16:28 Nate_ joined #salt
16:28 clearclaw joined #salt
16:30 GradysGhost joined #salt
16:33 Gifflen joined #salt
16:36 rojem joined #salt
16:36 yomilk joined #salt
16:37 mgw joined #salt
16:38 clemenko left #salt
16:38 hunter joined #salt
16:39 forrest ok so who wants to copy this: https://github.com/jenkinsci/puppet-jenkins
16:42 Gifflen joined #salt
16:43 Kenzor joined #salt
16:47 danielbachhuber joined #salt
16:50 Flusher joined #salt
16:52 pydanny joined #salt
16:53 jslatts joined #salt
16:54 TheRealBill_here joined #salt
17:01 btorch hmm, when we assign roles to a minion... salt applies them from bottom to top ?
17:02 lessismore joined #salt
17:02 foxx joined #salt
17:03 aleszoul3k joined #salt
17:04 smcquay joined #salt
17:07 Linz_ joined #salt
17:08 MedicalJaneParis joined #salt
17:08 btorch the call to "state.show_lowstate" just shows me things in a weird order even though I'm uring requires to organize things
17:09 astol joined #salt
17:09 jeremyfelt joined #salt
17:11 rgbkrk joined #salt
17:11 KyleG joined #salt
17:11 KyleG joined #salt
17:13 scott_walton joined #salt
17:14 mephx joined #salt
17:16 srage joined #salt
17:17 kaptk2 joined #salt
17:25 joehillen joined #salt
17:27 ipalreadytaken joined #salt
17:30 NotreDev joined #salt
17:30 tharkun joined #salt
17:32 seanz Greetings. Is there an easy way to render out a salt state file with jinja variables?
17:32 NotreDev is there anyway to see the logs from a state.highstate?
17:32 NotreDev 1 state failed, but it's beyond my buffer
17:33 tvle joined #salt
17:33 NotreDev my stdout/stderr buffer
17:33 seanz NotreDev: Can you increase the buffer and/or pipe stdout and stderr to less?
17:33 NotreDev hmm
17:33 NotreDev i have 151 states. i feel like i'm doing logging of salt wrong
17:34 dave_den seanz: if you mean you just want to see what the file looks like when rendered, use http://docs.saltstack.com/ref/modules/all/salt.modules.cp.html#salt.modules.cp.get_file
17:34 xmltok joined #salt
17:34 saint-n it would be awesome if salt did a summary at the end of pass fail like ansible does for each host
17:36 seanz dave_den: Trying...
17:36 btorch is state.show_lowstate not a good way to find out the order of the states ?
17:38 dave_den btorch: that, or show_highstate
17:40 btorch I don't know it just doesn't seem to show a proper order even though requires are in place ... I'm going over state.highstate test=True now to see there
17:41 seanz dave_den: Thanks. I cp.get_template did what I wanted it to, but you pointed me in the right direction. :)
17:41 hurtz joined #salt
17:42 dave_den seanz: cool. fyi, both methods will do it, but you have to pass template= to get_file
17:42 schimmy joined #salt
17:43 seanz dave_den: Oh, interesting. I guess I misread the docs. It looked like passing template= to get_file would allow me to use variables in the salt call, but I didn't know that extended all the way to the salt master.
17:43 zirpu joined #salt
17:45 seanz Is there a case where I would want to see the pre-rendered file? I see that in my editor.
17:46 dave_den if you're getting the wrong version from a gitfs backend or something maybe
17:46 alexthegraham joined #salt
17:47 teskew joined #salt
17:47 seanz dave_den: Got it.
17:47 seanz Just not a use case I've run into yet, but nice to know it's available.
17:50 btorch dave_den: not quite sure what's throwing off the order though when checking with show_lowstate ... "state.highstate test=True" seems more in order ... http://goo.gl/bJJGuI
17:51 btorch dave_den: no matter what I do, show_lowstate always seem to go object -> swift -> account -> container
17:51 quickdry21 joined #salt
17:51 bhosmer joined #salt
17:53 schimmy joined #salt
17:53 dave_den btorch: http://docs.saltstack.com/ref/states/layers.html?highlight=compiled
17:55 tvle joined #salt
17:55 Mr_N dave_den: You're good at "this sort of stuff", is there any reason why a refresh_pillar wouldn't refresh the pillar, and would necessitate a salt-minion restart? (which performs the reset successfully)
17:56 dave_den saltutil.refresh_pillar should do it.
17:56 dave_den not sure why it wouldnt
17:56 Mr_N Yeah; that doesn't, I have it in a reproducable state, too
17:56 Mr_N (running salt-minion -ldebug, too)
17:56 btorch dave_den: thanks
17:56 Mr_N (version 0.17.5)
17:57 dave_den check your master in debug mode to see if there are errors when compiling the pillar for any reason
17:57 toastedpenguin joined #salt
17:57 Mr_N Sure.  (running to a meeting, will continue debugging when I'm sitting there.  Also, I'd expect if there were errors the "restart minion" wouldn't work successfully (it does), but shall try explicitely)
17:58 dave_den yeah, i would agree. you just never know, though
17:59 mgw joined #salt
18:00 tolkil joined #salt
18:00 CeBe joined #salt
18:01 diegows joined #salt
18:04 ipalreadytaken joined #salt
18:04 Mr_N dave_den: Nada :(
18:04 aleszoul3k joined #salt
18:06 tvle joined #salt
18:06 ajw0100 joined #salt
18:08 themadcanudist joined #salt
18:09 themadcanudist hey guys, does anyone have a working example of a json_jinja .sls file?
18:09 hunter joined #salt
18:11 Gareth morning
18:12 forrest hey Gareth
18:12 Ryan_Lane joined #salt
18:12 Gareth forrest: hey. hows it going?
18:12 forrest Gareth, almost Friday!
18:12 Gareth almost :)
18:14 Ryan_Lane joined #salt
18:17 res0nat0r joined #salt
18:17 alexthegraham What's the best way to keep a state from being rendered until a package has been installed?
18:18 alexthegraham i.e. I'm trying to test if grains['productname'] has a value before installing a package, but that grain doesn't exist until dmidecode is installed.
18:18 alexthegraham And if the grain doesn't exist, rendering the state fails.
18:19 ndrei joined #salt
18:19 ipalreadytaken joined #salt
18:19 NotreDev joined #salt
18:24 benza joined #salt
18:25 benza Hi guys
18:27 jeremyfelt joined #salt
18:28 benza I'm busy playing around with saltstack. Trying to deploy a Web hosting enviroment with varnish, nginx and mysql. I'd like my varnish servers to automatically add nginx back end s when they get deployed. I've got the sls files and everything in place to do this but my states won't get executed by the reactor. Is there something I need to enable on the master to get reactors working?
18:30 saurabhs joined #salt
18:31 forrest benza, did you try these steps already? http://docs.saltstack.com/topics/reactor/#debugging-the-reactor
18:32 benza I haven't,  that looks like a good start point. I'm gonna try it when I get home. Thanks a bunch!
18:33 forrest benza, yea no problem.
18:33 naiyte joined #salt
18:36 alunduil joined #salt
18:37 yomilk joined #salt
18:38 themadcanudist hey guys, in a jinja templated .sls file, how can i throw a properly formatted error (if certain pilllar data is missing, etc.?)
18:40 dave_den themadcanudist: https://groups.google.com/d/msg/salt-users/o4aBUzZVGpM/gK6NpnNLIqsJ
18:42 themadcanudist dave_den: seriously, you can't raise an exception in jinja.. *sigh*
18:42 forrest themadcanudist, you could always tweet at Armin :P
18:42 themadcanudist oiy
18:42 dave_den not unless you write a custom filter, i suppose
18:43 gfa SALT IS AWESOME!!!
18:43 gfa sorry for the spam, but i just finish to automate a huge pile of crap i use to do manually
18:43 gfa :D
18:44 cnelsonsic joined #salt
18:44 forrest gfa, nice!
18:44 themadcanudist Sorry, who's armin?
18:44 forrest themadcanudist, the dude who wrote jinja
18:48 whiteinge themadcanudist: you could write a small execution module that you call from jinja which could do the checking and raise a proper python exception otherwise
18:48 * themadcanudist nods
18:48 whiteinge it would probably just take 3 or 4 lines
18:48 themadcanudist i think i will investigate the custom jinja filter
18:48 themadcanudist that sounds super-simple
18:49 whiteinge not a jinja filter but an actual salt execution module
18:50 themadcanudist whiteinge: I want something more generic so I can raise specific error messages based on template conditionals
18:50 themadcanudist something simple, so if pillar data isn't structured in a particular way, then just raise an exception with the error specified in the template filter?
18:51 whiteinge fair enough. a jinja filter may be the best way to go then
18:51 whiteinge since i typed it out though, alternate (less generic) idea: http://paste.fedoraproject.org/81023/35269921
18:51 Linz joined #salt
18:52 themadcanudist ah, thank you sir!
18:54 astol is it possible to somehow include another salt:// file from file.managed: which has template: jinja option?
18:54 astol i.e. include template from template?
18:55 forrest whiteinge, you should put that in contrib!
18:56 [diecast] joined #salt
18:56 forrest astol, what do you mean?
18:56 ausmarton joined #salt
18:56 forrest a template inside a template?
18:58 astol forrest: I mean I have my basic haproxy state, which file.managed's some base haproxy.conf, and I want to extend it for different servers
18:58 forrest astol, so why not just handle that in the templated file>?
18:58 astol forrest: how exactly? can you give me pointers?
18:59 whiteinge forrest: actually we should put that in the best practices docs, now that you mention it. we talk about making execution modules specifically for use in jinja during the training but it's not in the docs afaik
18:59 forrest whiteinge, yea I don't think I've ever seen that in the docs
18:59 forrest whiteinge, that would be great to have though
18:59 forrest astol, yea hang on
19:00 druonysus joined #salt
19:02 themadcanudist hah, my second question now is how do i register my custom jinja filter within salt? =D there is exactly one post i can find via google, where the user ended up using a completely different methodology (rather than acustom jinja filter)
19:02 themadcanudist is it possible without having to hack the salt installation?
19:04 forrest astol, https://gist.github.com/gravyboat/9256809
19:04 fragamus joined #salt
19:05 astol forrest: yeeeeeah, thanks, that's based on grains, and I wanted to write another file/part of file
19:05 astol to not put it grains  but to have just a separate file
19:06 forrest astol, so you want to create a second file if it's a specific machine?
19:06 astol forrest: https://gist.github.com/andreyst/0d191586d798e90b4d05
19:06 astol something like this
19:07 astol the trick is that I should put everything in 1 file
19:07 astol haproxy doesn't allow to include multiple files (actually it does but that's too difficult for me atm)
19:07 astol it doesn't allow it easily via conf.d or include statement
19:07 astol so I wanted to extend basic haproxy config with different additions on each load balancer
19:08 ajw0100 joined #salt
19:08 hunter joined #salt
19:10 forrest refresh your gist astol
19:10 forrest still wrong here?
19:10 forrest in what you want
19:11 forrest astol, I guess I'm confused as it seems like a templated file based on LB details would be the best bet.
19:11 forrest since you only want one file
19:12 astol forrest: nope, that's not it. I think I got something in templates
19:12 astol forrest: yeah! will check out templtated file
19:12 astol thanks
19:12 forrest astol, yea np
19:13 ajw0100 joined #salt
19:14 cjbarnes18 joined #salt
19:14 dN0t joined #salt
19:14 dN0t iptables module - what is the keyword argument for setting a DROP policy on INPUT, OUTPUT and FORWARD?
19:15 johtso Any news on the 2014.1.0 ubuntu package?
19:15 dN0t Understand I set - name for the table, but not clear what additional value I set to DROP?
19:15 Gareth dN0t: jump.
19:15 ajw0100 joined #salt
19:16 Gareth dN0t: assume you're using build_rule?
19:16 dN0t Gareth: affirm.  Just want to set default policy to DROP, add my own chain for the inbound/outbound permissions...
19:17 dN0t Gareth: Issue is the -P switch for iptables doesn't use -j.
19:17 dN0t Gareth: the next argument is just the policy you wish to set for the table.  Seems from this interface with iptable.set_policy that wouldn't work?
19:17 Gareth dN0t: ahh. hm. good point.
19:18 kermit joined #salt
19:18 dave_den joined #salt
19:18 Gareth dN0t: ahh. set_policy is what you want.
19:18 Gareth not build_rule.
19:19 dN0t Gareth: yeah that's what I think, but I'm not sure what keyword argument I need to set for the policy.
19:19 dN0t Gareth: Documentation is a little sparse here.
19:20 Gareth dN0t: doesn't look like there is a keyword.  just salt '*' iptables.set_policy filter INPUT ACCEPT
19:20 dN0t gotcha.
19:20 dN0t Much obliged sir.
19:20 Gareth glad to help :)
19:20 chrisjones joined #salt
19:22 hunter joined #salt
19:26 AdamSewell joined #salt
19:26 AdamSewell joined #salt
19:26 dN0t well.  crap.  salt install is borked after pip upgrade...
19:26 dN0t Needs apache-libcloud?
19:28 hunter joined #salt
19:29 dN0t is that the only new requirement?
19:30 toastedpenguin joined #salt
19:32 * Gareth fires off a rather large pull request
19:32 GradysGhost I don't suppose Salt has a state module that manages trusted SSL certs and CAs, does it?
19:33 GradysGhost Aside from managing the files themselves, that is.
19:33 jab416171 joined #salt
19:34 giantlock joined #salt
19:36 NotreDev joined #salt
19:37 themadcanudist does anyone know how to pull a custom jinja filter in to the scope of salt?
19:39 Gifflen joined #salt
19:42 ashb joined #salt
19:42 zain joined #salt
19:43 ckao joined #salt
19:43 srage joined #salt
19:44 Linuturk so, what happens if I manage the salt mininon configuration with salt, and I issue a salt-minion restart? will I get the response or will it restart before the minion can respond?
19:46 dave_den joined #salt
19:46 srage Linuturk: my experience, at least with 0.17.5 is that the minion will stop and not restart.  Working around it with salt-ssh at the moment.
19:46 Ryan_Lane joined #salt
19:49 martoss1 joined #salt
19:51 madduck :((
19:51 madduck [ERROR   ] Received incomplete call from blue.virt.local for '_ext_nodes', missing 'tok'
19:51 madduck so 0.16.4 minions cannot connect to 0.16.4 anymore
19:51 madduck uh to 2014.1
19:52 forrest madduck, ?
19:52 forrest madduck, you technically aren't even supposed to use 0.16.4 minions with a 0.17 master
19:53 saint-n bah technically
19:53 saint-n when has that ever stopped an admin
19:54 hunter joined #salt
19:55 dave_den madduck: yes, there is a backwards compatibility change
19:56 madduck i think that's just bad.
19:56 forrest madduck, why? There was a protocol change for security purposes that wasn't backwards compatible
19:57 madduck protocols have version numbers for a treason
19:57 madduck reason ;)
19:57 forrest *shrug*
19:58 madduck well, it's just a shame that debian backporting has to take so long due to the licencing issues and binary files in the source tree
19:59 seanz Can someone tell me if file.recurse copies a directory to another location and then creates a symlink to the old location?
19:59 seanz I'm not sure what sym_mode is supposed to represent.
19:59 dave_den sym_mode is new and is meant for determining how to handle symlinks on the master
19:59 seanz Oh, ok.
19:59 forrest dave_den, it just sets the symlink permissions
19:59 seanz So file.recurse will copy a file to another location.
20:00 forrest according to the docs
20:00 dave_den oh, sorry, i was thinking of keep_symlinks
20:01 forrest yea I figured :D
20:01 seanz forrest: The docs were a bit tough to understand.
20:01 seanz "The directory to set the recursion in" - that's odd wording.
20:01 dave_den seanz: file.recuse just copies down a directory form the master, recursively
20:01 ndrei joined #salt
20:01 n8n joined #salt
20:01 madduck am i the only one disturbed by the amount of duplication in log output and the suspicion of much deeper problems being the cause of that?
20:02 seanz dave_den: Is there a simple command to move a directory?
20:02 seanz From one place to another on the minion?
20:02 forrest seanz, as a one time thing?
20:02 forrest might be better to use a file.recurse, then delete the old directory
20:02 dave_den madduck: it depends on what is being logged. you may have seemingly 'duplicate' lines for some things which is an artifact of the multiprocess architecture of salt
20:02 seanz forrest: Yes. Kind of an odd thing I have to do.
20:02 Ryan_Lane joined #salt
20:03 dave_den seanz: cmd.run
20:03 seanz This application creates temp dirs in one spot. I need to move them, but if I set them up ahead of time then the application won't start up.
20:03 seanz dave_den: Thanks.
20:03 forrest seanz, that's... an odd application
20:03 seanz forrest: Yes, well, I chose it.
20:03 madduck dave_den: but then I wonder why the config isn't read and virtual services set up, etc., and then forked…
20:03 madduck rather than forking and doing the setup work n times
20:04 dave_den madduck: which mesages are you referring?
20:04 madduck for instance http://slexy.org/raw/s2qB0XMyK9
20:06 dave_den madduck: yes, that does not look normal
20:06 snuffeluffegus joined #salt
20:07 madduck i have a lot more if you want ;)
20:07 madduck (and this is a plain develop checkout in a venv just run for the first time)
20:10 dave_den oh, i thought that was a minion log
20:10 madduck nah, master
20:10 dave_den yes, for the master that is normal if you have more than one worker thread. default worker threads is 5
20:11 madduck again, I don't think it's normal for all threads to do the same setup work
20:11 madduck but whatever, too late to do anything about that now
20:11 dave_den it's 'normal', meaning it's expected.
20:12 jesusaurus are there any known issues about "Rendering SLS "base:rabbitmq" failed: could not determine a constructor for the tag 'tag:yaml.org,2002:python/object/apply:collections.OrderedDict'"? It looks like using the yaml jinja-filter is barfing on a list of lists pulled from pillar
20:12 alexthegraham Anyone got a sec to take a look at a state and tell me what I'm doing wrong?
20:13 forrest alexthegraham, can you pastebin it, as well as the error?
20:13 alexthegraham forrest: how 'bout a gist? https://gist.github.com/alexthegraham/03256a3f05c5581691b2
20:13 forrest yea that's fine
20:14 alexthegraham forrest: that state is not creating my /etc/mail/aliases file and is not restarting postfix.
20:14 martoss joined #salt
20:15 dave_den jesusaurus: funky. what does your state look like?
20:16 forrest alexthegraham, ok, you seem to be misusing exists: http://docs.saltstack.com/ref/states/all/salt.states.file.html#salt.states.file.exists
20:16 jesusaurus dave_den: thats coming from https://github.com/jesusaurus/salt-formula-rabbitmq/blob/master/rabbitmq/init.sls#L247
20:16 jesusaurus the full error is https://gist.github.com/jesusaurus/b4e8a2b52376ab69166c
20:17 alexthegraham since "create Default is True, if create is set to False then the file will only be managed if the file already exists on the system." , shouldn't it create the file if it doesn't exist?
20:19 dave_den jesusaurus: might want to make sure your user args don't contain any characters that will be interpreted by yaml
20:19 jesusaurus alexthegraham: file.exists and file.append are two different states
20:20 jesusaurus dave_den: I don't think anything in https://github.com/jesusaurus/salt-formula-rabbitmq/blob/master/pillar.example#L37 is being interpreted
20:20 alexthegraham jesusaurus that would explain why it's not appending any data there, but doesn't explain why the file isn't being created.
20:20 forrest dave_den, ++
20:20 forrest oh, jesusaurus ++, not dave_den
20:20 forrest no credit for you dave!
20:21 dave_den sad face
20:21 jesusaurus /yoinks
20:21 alexthegraham So if I want to append data to a file (if it exists), and create it if it doesn't, how do I set that up?
20:23 alexthegraham I can do this with a few cmd.run's, but would rather manage the file.
20:23 jesusaurus can you just use a file.managed?
20:23 forrest alexthegraham, yea I was gonna ask the same thing, file.managed with a templated jinja file
20:24 whiteinge themadcanudist: there may not be a hook for custom jinja filters.
20:24 forrest then just drop the pillar data in
20:24 zooz joined #salt
20:25 whiteinge themadcanudist: it should be easy to add something to the minion config. the code is here: https://github.com/saltstack/salt/blob/develop/salt/utils/jinja.py#L305
20:25 alexthegraham But some systems will have aliases that others don't. Or are you saying use pillar data for that and keep different pillar contents for each system?
20:25 [diecast] joined #salt
20:26 jesusaurus alexthegraham: yeah, i would do the latter
20:26 jesusaurus template the aliases using pillar data
20:26 alexthegraham I realize that's what pillars are for, but that seems like a PIA when all I want to do is make sure that a single line of text is at the end of a file. And create it if it doesn't exist.
20:27 jaimed joined #salt
20:28 jesusaurus it would be a little bit more work now... but it would save you a lot of effort every time you want to add an alias to a host
20:29 dave_den alexthegraham: use jinja logic to test if the file exists or not. if it exists, use file.append, otherwise use file.managed with the 'content' option.
20:30 dave_den or use a cmd.run to do 'touch' on the filename, then use file.append
20:31 faldridge joined #salt
20:32 Ahlee anybody got a sample or know where in the code __opts__ is populated?
20:33 kula get very familiar with salt/loader.py
20:34 kballou joined #salt
20:34 Ahlee ty
20:34 Ryan_Lane joined #salt
20:34 dave_den Ahlee: the opts are created in config.py, but if you are using __opts__ in a module then it is built by loader.py
20:35 kula nearly every time i want to figure out something interesting in salt, i end up reading loader.py
20:35 Ahlee troubleshooting a returner, want to suck in __opts__ from my test script
20:35 Ahlee suppose i could just create a saltclient and abuse that
20:38 yomilk joined #salt
20:39 Ahlee in fact, that sounds much smarter.
20:39 johtso joined #salt
20:40 sroegner joined #salt
20:40 n8n joined #salt
20:47 alexthegraham BAM. Thanks dave_den. Here's the working version: https://gist.github.com/alexthegraham/5fd297ba5ca228def0b7
20:47 alexthegraham And thanks forrest and jesusaurus
20:48 ipalreadytaken joined #salt
20:50 madduck is there a recommended way to run salt from source that does not involve virtualenv?
20:50 blee joined #salt
20:50 madduck I don't want a virtualenv, I want to run salt *on the host* from source
20:51 dave_den alexthegraham: looks good. you might want to make the file.append require the cmd.run mkdir/touch, but it's not totally necessary
20:51 schimmy joined #salt
20:51 dave_den madduck: i would recommend running it in a virtualenv, but if you just want to install it system-wide, just run ./setup.py install
20:52 alexthegraham Yeah. I don't want it to be dependent on that because if the file already exists with some other values in it, I just want to append that data.
20:52 dave_den alexthegraham: it will still work
20:53 dave_den the 'unless' will evaluate to true, and the state will be marked as successful, so file.append will still run
20:53 alexthegraham dave_den : 'cause the "unless" will return true, the cmd.run will return True?
20:53 naiyte joined #salt
20:53 alexthegraham Oh, cool.
20:53 forresta joined #salt
20:53 madduck dave_den: the problem is upgrades when actually copy to /usr/local
20:53 madduck unfortunately, setup.py does not come with install --develop
20:54 toastedpenguin joined #salt
20:54 dave_den —develop?
20:54 madduck dave_den: https://groups.google.com/forum/#!topic/salt-users/p5NjaSETLgs
20:54 madduck dave_den: that installs a reference into /usr/local that points to the source checkout
20:55 madduck so the source is available system-wide except it actually runs from source. ;)
20:55 AdamSewell joined #salt
20:55 madduck i don't know, one of the many setuptools variants has it
20:55 cewood joined #salt
20:55 dave_den i'd still recommend virtualenv. that's what i do and it works well
20:55 madduck is there a way to do without pip?
20:56 dave_den without pip meaning what
20:56 madduck I don't want to install pip, so can I just create a virtualenv and install salt into it from source?
20:57 dave_den i just do:  virtualenv —system-site-packages /somepath then activate teh virtualenv, cd /git/salt; git checkout tag; ./setup.py install
20:57 alexj_ joined #salt
20:57 Gordonz joined #salt
20:58 hunter joined #salt
20:59 madduck and how do you upgrade, i.e. after switching to another tag?
20:59 dave_den if you want to do it in the same virtualenv, just check out the new tag and run ./setup.py install again
20:59 madduck will that remove obsolete files?
20:59 dave_den i would recommend making a clean virtualenv
21:00 madduck which is hard to do using salt itself. ;)
21:00 dave_den that's the nice thing about virtualenvs
21:00 dave_den it is?
21:00 madduck how would you do it?
21:00 dave_den http://docs.saltstack.com/ref/states/all/salt.states.virtualenv_mod.html
21:00 madduck I want to run salt from source on my minions and be able to upgrade to newer tags by changing pillar data
21:01 madduck dave_den: sure, I can create one with salt. I can also maintain a git clone with salt
21:01 dave_den yes, you have all the tools you need
21:01 madduck and I can probably run setup.py install when the git clone changes
21:01 Ryan_Lane1 joined #salt
21:01 madduck but how do I maintain throwaway virtualenvs?
21:02 Nazca__ joined #salt
21:02 dave_den write a state that deletes anything older than x, or is not current virtualenv
21:02 JordanRinke joined #salt
21:03 JordanRinke Hello
21:03 kballou joined #salt
21:05 Katafalkas joined #salt
21:05 dave_den howdy
21:05 GradysGhost Got a weird one, pros. Using file.append to add the contents of an SSL cert to a file which contains trusted certs. I'm doing this with this state: http://www.hastebin.com/jexeleviqo.mel
21:06 GradysGhost And that references a pillar as defined here: http://www.hastebin.com/lekoruvoce.vbs
21:07 GradysGhost When I apply the highstate, the -----BEGIN CERTIFICATE----- line is omitted
21:07 GradysGhost When I run a show_highstate, it is clearly included.
21:07 CeBe joined #salt
21:07 GradysGhost When I do a pillar.item on that variable, it's included.
21:07 dave_den GradysGhost: try escaping the lines in yaml
21:08 GradysGhost What do you mean? What needs escaping?
21:08 jeremyBass joined #salt
21:08 dave_den - "{{ line }}"
21:08 GradysGhost I see
21:08 GradysGhost Otherwise, the ----- get seen as objects.
21:08 GradysGhost Er, syntax
21:08 GradysGhost I can't word today.
21:10 GradysGhost No dice.
21:10 GradysGhost Also no BEGIN CERTIFICATE
21:11 GradysGhost Trying single quotes
21:11 GradysGhost Still no
21:12 ipalreadytaken joined #salt
21:12 GradysGhost Odder still is that this exact setup works in other salt configs I have.
21:12 zach magnets
21:13 GradysGhost Well, I guess not exact. Those files are created as empty files and then appended to, so it's not really the same I guess.
21:13 * GradysGhost turns on debug logging
21:14 NotreDev joined #salt
21:14 GradysGhost But it is something to do with those dashes. I'm not getting the END CERTIFICATE line either
21:14 dave_den GradysGhost: you could just do {{ pillar['my-cert'] | yaml }} instead of looping, since you have the lines formatted as a list in your pillar anyway. which is kind of odd
21:15 GradysGhost I had no idea that was a thing
21:16 skynet joined #salt
21:16 DaveQB joined #salt
21:19 forrest joined #salt
21:19 hunter none of my centos-5 machines are responding to a test.ping from the master
21:19 hunter unless I restart the minion
21:19 dave_den GradysGhost: i would store the cert as a string value in your pillar and just use - text: |\n        {{ pillar['my-cert']|indent(8) }}
21:19 hunter Are there known issues with the 0.17.5 minion on centos/rhel-5 ?
21:20 FarrisG total n00b question: Is there a good way to handle setting up states for packages that require ruby plus a bunch of gems? I have ruby and rubygems installed correctly via sls, but I have a package that includes, as an installation step, the need to run "bundle install" based on Gemfile. Is there a formula for that, or is it kosher to just dig out the list of gems from the Gemfile and use gem.installed?
21:21 baniir joined #salt
21:21 ipmb joined #salt
21:22 dave_den FarrisG: assuming you're not using rvm/rbenv?
21:23 dave_den bundle install is meant to do the gem installation for you
21:23 hunter yeah -all these nodes are dead until a service restart
21:23 ninkotech joined #salt
21:23 dave_den hunter: other than the old zmq stuff, i'm not aware of particular rhel5 issues
21:24 dave_den but it's been a long known issue that zmq < 3 have persistent connectivity issues
21:24 hunter dave_den: theres a google hit about them not reconnecting after a master restart, but the last master restart predates those minions starting up.
21:24 budrose joined #salt
21:25 budrose joined #salt
21:25 hunter Hmmm. That might be it.
21:25 dave_den hunter: i don't recall exaclty, but there was also an issue of simply losing connection after a time i think
21:25 chrisjones joined #salt
21:25 hunter dave_den: just on zmq 2, rhel 5 or minions in general?
21:25 dave_den all zmq < 3
21:26 zirpu left #salt
21:26 hunter well I just hit that then
21:26 hunter poo
21:26 hunter I wonder how much of a faff rebuilding zmq is
21:28 dave_den the zmq 2.2 in epel5 testing might be good to try
21:28 mgw joined #salt
21:28 dave_den it's always hard to tell what patches redhat has backported
21:29 forrest you don't need to rebuild zmq
21:29 harobed joined #salt
21:29 xDamox Hello dave_den, do you know a good way I can list a directory on the mast and the files inside that directory get used in a state?
21:29 dave_den for unofficial packages, there is http://zeromq.org/distro:centos
21:29 hunter dave_den: just found that
21:29 forrest dave_den, exactly, and technically they consider that their 'official' one almost...
21:29 forrest which is dumb
21:30 harobed joined #salt
21:30 hunter forrest: python links against the zeromq library doesn't it -
21:30 dave_den xDamox: cp.list_master will list all files on the master, state.show_states will show the states applied to a minion for highstate runs
21:30 david_a joined #salt
21:31 xDamox dave_den, I am trying to list files under /srv/pillar/import and then in /srv/pillar/top.sls I an trying to use the filenames
21:31 psyl0n joined #salt
21:31 rojem joined #salt
21:32 dave_den xDamox: you cannot get a list of files under your pillar roots path from the minions
21:33 dave_den oh
21:33 dave_den i see what you mean
21:33 xDamox :)
21:33 xDamox I can hard code the files into the top.sls and do a for loop
21:33 dave_den xDamox: there is not a way to do that in simple jinja
21:34 dave_den yes, you could
21:34 xDamox Ok
21:35 xDamox But I would like to be more dynamic and perform an ls of the directory
21:36 xDamox would you recommend writing a salt runner and executing salt['name of runner'] in the Jinja for loop?
21:37 mgw joined #salt
21:38 madduck dave_den: one way would be to watch the git.latest function from cmd.run, which removes the virtualenv, and to have a cmd.run to run setup.py requiring both git and the virtualenv_mod function…
21:39 madduck but how do I ensure that the watching cmd.run gets run before the virtualenv_mod that is required by the cmd.run running setup.py?
21:40 dave_den xDamox: you could use mako to do it
21:40 ipalreadytaken joined #salt
21:40 forrest dave_den, dun dun dunnnnnnn
21:40 forrest dave_den, when what you want to do is complex enough to write mako, you might as well go full python! :P
21:41 dave_den nah, not for a mostly templated file
21:41 forrest oh sorry I thought this was a state, the joys of being in meetings
21:41 njpatel joined #salt
21:41 dave_den it would just require one line of mako, and you could keeo the rest of it jinja, then just send it through the render pipe with #! mako | jinja | yaml
21:42 forrest does the | with a space work dave_den?
21:42 forrest I've never tried it
21:42 forrest always just done #!mako|jinja|yaml
21:42 dave_den yeah, you can have a space
21:43 forrest nice
21:44 dave_den madduck: use cmd.wait for the cleanup, have it require the cmd.wait of the setup, which requires the virtualnv and git states
21:46 ninkotech joined #salt
21:46 dave_den you probably don't want to delete your current virtualenv on the running salt-minion though
21:47 jalbretsen joined #salt
21:48 jinnko_ joined #salt
21:49 madduck probably not, but this does raise an interesting question: how do I restart the minion? ;)
21:50 madduck then again, I am using salt-call for regular upgrades
21:50 dave_den madduck:  some people on the mailing list use 'at'. i have written a POC module that will allow salt to restart itself.
21:51 madduck at sounds horrible ;)
21:51 dave_den but it may be a while before it gets merged, if ever
21:51 madduck right
21:51 dave_den there's always the esky way
21:51 n8n joined #salt
21:51 madduck yikes
21:51 bemehow joined #salt
21:52 madduck i might also just run windows
21:52 bemehow hello everyone, just searching for another pair of eyes to look into the problem. Could you spare a second and share what you think?
21:52 bemehow https://github.com/saltstack/salt/issues/10810
21:54 Katafalkas joined #salt
21:55 abe_music joined #salt
21:55 dave_den bemehow: are you sure state_auto_order is True on the minion?
21:55 bemehow yes
21:56 bemehow dave_den: without state_auto_order show.lowstate  doesn't display 'order' fields
21:57 dave_den indeed. what version?
21:58 rojem joined #salt
21:58 jacksontj_ joined #salt
21:58 jacksontj_ basepi: yt?
22:00 dave_den bemehow: when you run 'salt-call -l debug state.show_lowstate', do you see the rendered lowstate in the debug info contained in an OrderedDict?
22:00 bemehow dave_den: let me verify
22:01 dave_den or rather the rendered yaml state info
22:02 ninkotech joined #salt
22:02 dave_den bemehow: it will show  [DEBUG   ] Results of YAML rendering:\n OrderedDict(…..
22:03 Katafalkas joined #salt
22:05 basepi jacksontj_: hey
22:05 jacksontj_ wanted to pick your brain on https://github.com/saltstack/salt/pull/10374
22:06 bemehow dave_den: https://gist.github.com/anonymous/927ffe841a4f1f097dae
22:06 jacksontj_ so i believe my concept of compatibility is correct, but it sounds like the jenkins machinery doesn't enforce that compatability modell-- right/
22:09 ausmarton joined #salt
22:09 glen_ joined #salt
22:09 dave_den bemehow: what version?  salt-call —versions-report
22:09 baniir joined #salt
22:09 hunter Ok, I have the zeromq pkgs from the OBS. I just need to get an rpm for python26-zmq that links against it and that should fix the issue.
22:10 ninkotech joined #salt
22:10 hunter Sadly, I need i386 and x86_64 - <sigh>
22:10 kermit joined #salt
22:11 dave_den hunter: oh hey, the x86_64 is all here: http://docs.saltstack.com/downloads/cent5/
22:12 bemehow dave_den: https://gist.github.com/anonymous/fb3728d08f4247a84b67
22:12 bemehow dave_den:  0.17.5 salt codebase
22:12 ninkotech joined #salt
22:13 hunter dave_den: awesome - do you know where, if anyplace the src rpms can be had from?
22:13 dave_den bemehow: hrm it seems like a bug, since the top.sls is rendering as an ordereddict. i'd have to dig into the state compiler though
22:13 hunter dave_den: I really do need i386 and for x86_64 even from past pain I always try to have the src rpms available.
22:14 dave_den hunter: i don't know, i just happened upon that directory in google
22:14 pydanny joined #salt
22:15 basepi jacksontj_: sorry, got distracted
22:15 lzhang joined #salt
22:15 basepi jacksontj_: so there's a master that runs on the jenkins server which uses salt cloud to spin up the minions, bootstraps them, installs things required for the test run
22:16 basepi jacksontj_: that master doesn't upgrade itself automatically, as far as i know
22:16 lzhang How do I do a cmd.run if a directory is missing?
22:16 hunter dave_den: cool
22:16 basepi jacksontj_: so i think that was the problem (though i haven't asked anyone if that was actually the problem, that's just what i understood from reading the thread)
22:17 jacksontj_ k, so if thats the case it sounds like i need to commit the master and config wrap the minion wait a while then change the minion default
22:17 jacksontj_ who would i talk to to verify that?
22:17 dave_den hunter: dunno if you want to trust it, but: http://www.comingchina.com:8080/testing/ownsource/5/
22:18 forrest lol
22:19 mgw is this a bug or a feature: https://gist.github.com/mgwilliams/487aaa77fe3b6a6f870e
22:19 forrest mgw, what release is that?
22:19 hunter dave_den: interesting url
22:19 mgw forrest: 0.17.5
22:19 mgw maybe it's been fixed?
22:19 forrest mgw, hmm, nope then the bug I am thinking of isn't related to that.
22:20 forrest mgw, the bug I was thinking of was fixed prior to 0.17.5
22:20 dave_den hunter: linked from: http://www.jsxubar.info/centos-5-zeromq2-update-zeromq3.html
22:20 mgw this is caused by the conflict between oracle-java7-installer and openjdk-6-jre-headless
22:20 mgw the former "provides" the latter
22:20 mgw but isn't the same as
22:20 mgw so Salt is trying to remove it when it is not installed
22:21 forrest Hmm, yea seems like it is almost looking at the wrong value
22:22 dave_den mgw: looks like a bug. does purged also break?
22:22 mgw not sure, it's not my system
22:22 RobSpectre joined #salt
22:23 valgrind joined #salt
22:24 bemehow dave_den: so I looked and it seems OrderedDict is new in 2.7, im running at 2.6 wonder how this is implemented;
22:25 ninkotech joined #salt
22:25 dave_den bemehow: salt implements its own ordereddict for salt on py2.6
22:26 dave_den so it should work properly..
22:26 lzhang left #salt
22:28 themadcanudist so, when trying to use mako (as well as jinja+yaml) in my .sls files, i'm getting an exception: ImportError: No module named mako.exceptions…. i have maki installed and can import both mako and mako.exceptions in my python REPL
22:29 themadcanudist and as a side question, is jinja+yaml the recommended combo for .sls files?
22:29 dave_den themadcanudist: you have mako installed on your minions?
22:29 themadcanudist i know this is all religious preference...
22:29 themadcanudist dave_den… ah, i do not.
22:29 themadcanudist right, i keep thinking this is like puppet
22:29 dave_den themadcanudist: jinja + yaml is just the default
22:29 themadcanudist compiles on the master, applies on the minion, but it's not
22:30 themadcanudist it fires everythign over to the minion for execution
22:30 dave_den yeah, the minions do the template rendering themselves
22:30 themadcanudist gotcha
22:31 themadcanudist i suppose there is no way to protect the formulas that are sent to the minions and encrypt them or something?
22:31 themadcanudist ie. once they're there, they can be read by the superuser, correct?
22:31 forrest themadcanudist, that's why you put secure data inside of pillar.
22:32 themadcanudist yep and set pillar_opts: false
22:32 dave_den themadcanudist: correct, the salt minion will cache some things. the salt-users mailing list has archives of discusssions about that sort of thing
22:32 themadcanudist forrest: understood, just wondering.
22:32 forrest cool
22:32 themadcanudist dave_den: interesting, is there any push to add protection for that?
22:32 forrest dave_den, If you printed them out I think you could fill a few hefty trash bags :P
22:32 themadcanudist forrest: lol!
22:33 forrest themadcanudist, not really as far as I know, you don't put secure data in states.
22:33 jalbretsen joined #salt
22:33 * themadcanudist nods
22:33 aleszoul3k joined #salt
22:33 dave_den themadcanudist: i don't know how much saltstack (the company) is getting for it, but it could be done from what i know of the source code
22:34 themadcanudist dave_den: everything can always be done (software wise), it's always a question of will and resources ;)
22:34 Ryan_Lane1 joined #salt
22:35 dave_den indeed
22:36 dave_den bemehow: fyi if you are invetigating the odict thing, the code is in salt/utils/odict.py
22:37 bemehow dave_den: looking at the State objects now, there is a function that order chunks that seems interesting
22:37 basepi jacksontj_: i would ping s0undt3ch, in #salt-dev
22:37 basepi (sorry, got distracted again.  i'm all over the place today)
22:38 ninkotech joined #salt
22:38 dave_den bemehow: state.py order_chunks?
22:39 dN0t joined #salt
22:41 jacksontj_ basepi: no one is on that channel… is that internal or something?
22:42 basepi jacksontj_: sorry, i meant #salt-devel
22:42 NotreDev joined #salt
22:43 jacksontj_ basepi: no worries ;) thanks for the help!
22:43 basepi jacksontj_: yep
22:44 bemehow dave_den: class State method order_chunks in state file
22:45 bemehow dave_den: line 891
22:46 NotreDev is there a way to log the colors of a salt's output to a file. i.e. i want to scroll through a file and see the green and red
22:46 ajw0100 joined #salt
22:48 baniir joined #salt
22:48 rojem joined #salt
22:49 dave_den NotreDev:  check —help
22:49 ninkotech joined #salt
22:52 NotreDev k
22:53 chrisjones joined #salt
22:54 ipmb when do you think 2014.01 will hit the ubuntu ppa?
22:55 ipmb 2014.1.0 :)
22:58 themadcanudist hey, about the default renderers… yaml_jinja or jinja_yaml is default…  i want to use jijna|mako|yaml… which i can set in a shebang, but how do I make this default in the master conf?
22:59 ipalreadytaken joined #salt
22:59 alexj_ joined #salt
22:59 kintel joined #salt
23:00 Ryan_Lane joined #salt
23:00 baniir joined #salt
23:01 ninkotech_ joined #salt
23:02 dave_den themadcanudist: it's a minion setting, 'renderer'
23:02 themadcanudist yeah, i see it
23:02 themadcanudist i'm wondering what the option is… it's not formatted like the way you string together renderers in a sehbang .sls
23:02 themadcanudist would i just use underscore instead?
23:03 themadcanudist jinja_mako_yaml
23:03 themadcanudist ?
23:03 jinnko_ joined #salt
23:03 themadcanudist or yaml_mako_jinja? or what? =D
23:04 dave_den ah, good question
23:04 NotreDev joined #salt
23:04 kintel Anyone have a smart way of getting the pkg state to trigger a checkinstall process (for building packages from source) ?
23:05 urban joined #salt
23:06 dave_den themadcanudist: try just using pipes, but enclose them in doulbe quotes
23:06 dave_den "mako | jinja | yaml"
23:06 forrest dave_den, the conf should support the old method
23:07 forrest it has an example right here: http://docs.saltstack.com/ref/configuration/minion.html#renderer
23:07 themadcanudist yeah, that example works, but i'm talking about adding a third to the chain. i'm not sure if that's a static enumerated value, or if it's a _ delimited value
23:07 themadcanudist and what order should they be in, etc...
23:07 themadcanudist it's unclear
23:07 dave_den forrest: that only supports two
23:08 vejdmn joined #salt
23:08 dave_den themadcanudist: the code suggests using "mako | jinja | yaml" will work fine
23:08 urban hello, unable to get salt-cloud 2014.1.0 to work with hpcloud, got error and configuration like here - http://dpaste.com/1667731/
23:09 urban is it problem with hpcloud's openstack? or there's something wrong with my configuration?
23:09 urban have anyone tried salt-cloud with hpcloud?
23:10 themadcanudist dave_den: confirmed, works like a charm
23:10 dave_den cool
23:10 schimmy1 joined #salt
23:11 forrest dave_den, gotcha
23:11 armonge joined #salt
23:11 armonge Hi guys, is there a way to purge or the gitfs cache?? or make it pull latest changes?
23:13 armonge i added a new git remote:  https://github.com/saltstack-formulas/tomcat-formula then restarted the salt-master, however i don't see tomcat when i issue salt-call cp.list_master
23:13 dave_den armonge: salt-run fileserver.refresh should trigger a refresh
23:13 dave_den bbiab
23:18 yomilk joined #salt
23:18 synfin is there a way to override the minion config's root_dir variable for one setting?  Specifically I'm trying to create a sock_dir less than 107 characters in length
23:19 mgw forrest, dave_den: I was able to reproduce that apt/pkg issue with java. If there's a partially installed package that cannot complete its installation, all future pkg operations will fail.
23:21 armonge Hi, i used ~$ sudo salt-run fileserver.refresh
23:21 armonge Function 'fileserver.refresh' is unavailable
23:21 armonge i'm using vagrant@salt:~$ salt --version
23:21 armonge salt 0.17.5
23:21 ninkotech_ joined #salt
23:24 thayne joined #salt
23:25 armonge @dave_den thanks, it was
23:25 armonge salt-run fileserver.update
23:28 synfin never mind, i figured out a good workaround.  Just use sock_dir=../../../../your/directory, just go up enough parent directories as needed.
23:30 hunter Ok, i've downloaded or rebuilt as needed zmq and related bits for centos-5 and added them to the local repo.
23:30 GoKage joined #salt
23:31 hunter I'll try and remember to check those minions regularly to see how they do.
23:31 faldridge joined #salt
23:31 ahammond is there something similar to python's __name__ in salt's .sls states? I'd like to say file.managed: source: salt://{{ basedir }}/myfilename
23:31 GoKage hi i'm having some problems with my sls files, or maybe it is my server. i'm trying to get salt-ssh to work
23:31 hunter dave_den: I don't have a place to put them but these MIGHT be desireable by the rest of the community (your urls helped but nothing was a complete set of both binary and source rpms for both i386 and x86_64)
23:32 ahammond where basedir would be the path of the current salt file.
23:32 hunter dave_den: do you think the salt people would want them?
23:32 ahammond GoKage, oh, I've seen that problem before. :)
23:33 GoKage when i run test.ping, and cmd.run, both works
23:33 baniir joined #salt
23:33 GoKage but state.highstate doesn't seem to work
23:33 ninkotech_ joined #salt
23:33 GoKage i got the same files to work on a different set of VMs but when i copied those same files, unchanged to another set of VMs, it appears to run but no configurations were changed
23:34 ahammond GoKage top.sls matching issue?
23:34 fishpen0 joined #salt
23:34 GoKage i have a roster file
23:35 GoKage it is mapping the same thing with correct IPs and everything...
23:35 fishpen0 left #salt
23:36 ahammond run it in debug mode where it works and again where it doesn't and compare the output. where's the first change?
23:38 NotreDev joined #salt
23:38 GoKage let me give that a try. i'll try it now
23:38 AdamSewell joined #salt
23:38 AdamSewell joined #salt
23:38 aleszoul3k joined #salt
23:39 sinh joined #salt
23:40 elfixit joined #salt
23:40 kainswor joined #salt
23:41 kainswor is anyone here using multi-master with 2014.1?
23:41 skynet joined #salt
23:41 kainswor I built a package off the 2014.1 tag, and when I upgrade with it I get a StopIteration exception in the tune_in
23:42 kainswor Not sure if it's my package-making that broke it or if this is an issue
23:42 philipsd6 joined #salt
23:43 mephx_ joined #salt
23:43 darrend joined #salt
23:43 akl_ joined #salt
23:43 akl_ joined #salt
23:43 forrest joined #salt
23:43 seapasulli_ joined #salt
23:43 yml_ joined #salt
23:44 smcquay_ joined #salt
23:44 GoKage ok now i get "TypeError: Undefined is not JSON serializable"
23:44 Troicc joined #salt
23:45 20WABFIH9 joined #salt
23:45 svx joined #salt
23:45 xDamox joined #salt
23:45 tmwsiy joined #salt
23:46 mikkn joined #salt
23:46 kainswor joined #salt
23:47 chrisjones joined #salt
23:48 codysoyland joined #salt
23:48 jpcw joined #salt
23:49 zirpu joined #salt
23:49 themadcanudist do most of you use jinja or mako?
23:49 forrest jinja
23:50 themadcanudist cool, any specific reason you wouldn't use mako?
23:50 forrest the only time most people want to use mako is when they want to do something that is too complex, and at that point they just want to use python stuff.
23:50 themadcanudist yeah
23:50 forrest I like simplicity, and I try to stay simple even when using jinja if I can, or try to redesign the solution to be simpler
23:50 forrest the more jinja/mako the harder it is for other users to understand, and myself when I look at it later.
23:52 schimmy joined #salt
23:53 Networkn3rd joined #salt

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