Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-12-12

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

All times shown according to UTC.

Time Nick Message
00:00 APLU joined #salt
00:03 austin987 joined #salt
00:05 akoumjian Anyone have experience setting up halite? I'm just trying to create a user who can login, following the quickstart: https://github.com/saltstack/halite#installation-quickstart
00:05 quickdry21 joined #salt
00:06 pipps_ joined #salt
00:06 akoumjian sonofagun. Nevermind
00:06 NotreDev joined #salt
00:06 akoumjian Kids, remember to restart your master when making config changes.
00:07 jacksontj joined #salt
00:08 Steve__ anyone from saltstack on here?
00:10 davet joined #salt
00:14 shinylasers joined #salt
00:17 meteorfo_ joined #salt
00:22 dvogt I am having an issue invoking a custom outputter from local command line API. I am using like minions = client.cmd_cli(tgt=mytgt, fun=deployer_module, arg=[application], timeout=1, verbose=True, expr_form='list')
00:22 dvogt for minion in minions:
00:23 dvogt salt.output.display_output(minion, 'customouptter', client.opts)
00:23 dvogt This just defaults to nested.
00:23 dvogt It works from the command line.
00:23 dvogt Anybody have experience with this ?
00:30 jcsp joined #salt
00:30 dstanek joined #salt
00:34 ajw0100 joined #salt
00:36 kermit joined #salt
00:53 markizano joined #salt
00:54 dstanek anyone control their supervisor jobs using salt?
00:55 dstanek when my config in /etc/supervisord/conf.d/ changes i want supervisor to reload so that the config reloads, but it's not currently doing that
00:57 Steve__ Does this actually work ? https://gist.github.com/SEJeff/5257789
00:57 fllr joined #salt
00:59 NotreDev joined #salt
01:00 zandy joined #salt
01:03 Sheco joined #salt
01:03 redondos joined #salt
01:09 markfletcher joined #salt
01:12 zandy joined #salt
01:12 mesmer joined #salt
01:14 jacksontj joined #salt
01:17 markfletcher Hi - anyone here using Salt to manage windows servers?
01:17 OrbHacker joined #salt
01:20 OrbHacker hello
01:21 markfletcher hi
01:24 OrbHacker first time in salt, having major issues since 17.2
01:24 Gifflen joined #salt
01:25 teepark quit
01:25 teepark wups :()
01:26 OrbHacker hehe
01:26 sroegner joined #salt
01:26 OrbHacker its quiet
01:26 psyl0n joined #salt
01:27 mgw joined #salt
01:29 OrbHacker anyone else having problems with salt 17.2 and ZMQ version?
01:29 OrbHacker Salt: 0.17.2
01:29 OrbHacker Python: 2.6.8 (unknown, Nov  7 2012, 14:47:45)
01:29 OrbHacker Jinja2: unknown
01:29 OrbHacker M2Crypto: 0.21.1
01:29 OrbHacker msgpack-python: 0.1.12
01:29 OrbHacker msgpack-pure: Not Installed
01:29 OrbHacker pycrypto: 2.3
01:29 OrbHacker PyYAML: 3.08
01:29 OrbHacker PyZMQ: 2.2.0.1
01:29 OrbHacker ZMQ: 3.2.2
01:31 cachedout joined #salt
01:34 halfss joined #salt
01:35 fandikurnia01 joined #salt
01:35 OrbHacker anyone else having problems with salt 17.2 and ZMQ version?
01:37 Psi-Jack OrbHacker: Yes, ZeroMQ 2.x versions on our most hated CentOS 5.x servers, but other than that. No.
01:37 DanGarthwaite joined #salt
01:38 OrbHacker Psi-Jack: I had to hack in the ZMQ to version 3.2.2
01:39 OrbHacker found someone that made custom python-zmq and zeromq packages
01:39 OrbHacker it works from the minion side, ex: state.highstate
01:39 OrbHacker but fails every time from the master side, ex: salt '*' state.highstate
01:40 OrbHacker Psi-Jack: what other distros and versions do you have it working on?
01:40 OrbHacker RHEL6? CentOS 6?
01:40 Psi-Jack CentOS 6.x it works.
01:40 Psi-Jack Debian 7, Ubuntu 12.04 as well.
01:40 OrbHacker can i ask a dum question?
01:40 Psi-Jack You just did.
01:41 OrbHacker are you installing via epel or git?
01:41 OrbHacker haha
01:41 OrbHacker ;P
01:41 fandikurnia01 joined #salt
01:41 Psi-Jack EPEL, of course.
01:41 OrbHacker did ZMQ install via epel in version 3.2.2?
01:42 Psi-Jack Though, because EPEL's running short a bit, I plan to take the src.rpm and update to 0.17.4 manually from it.
01:42 Psi-Jack zeromq3-3.2.4-1.el6.x86_64
01:42 OrbHacker hmmm
01:42 OrbHacker guess ill try that
01:42 Psi-Jack so, that's a no. :)
01:43 OrbHacker its higher though, so should be good?
01:43 Psi-Jack has bug fixes I'm sure yes.
01:43 frosty996 I wish epel would release 0.17.4
01:43 OrbHacker ok, ill just stop the server, and build one in rhel6 and see what happens I guess
01:43 OrbHacker me too
01:44 OrbHacker ive spent 48 hours chasing this down...
01:44 OrbHacker ive got it to work frmo minion side, butnot master
01:44 oz_akan_ joined #salt
01:45 srage_ joined #salt
01:45 oz_akan_ joined #salt
01:45 OrbHacker did you use the bootstrap to call the epel installs or justamke yum calls yourself?
01:45 OrbHacker sorry about my space key... it sucks!
01:46 Psi-Jack Uhh, I just use yum to manually install epel, then I install salt, and let salt run from there.
01:46 OrbHacker ok
01:47 OrbHacker i found git during the process and it seemed to help me
01:47 OrbHacker
01:52 djbclark joined #salt
01:52 djbclark joined #salt
01:53 xl1 joined #salt
01:57 OrbHacker <Psi-Jack>: thanks for trying to help!
01:58 NotreDev joined #salt
02:06 dstanek joined #salt
02:06 frosty996 left #salt
02:10 troyready joined #salt
02:11 Chrisje joined #salt
02:11 OrbHacker well thats a drag
02:11 OrbHacker just booted up a template of RHEL6
02:12 Nexpro joined #salt
02:12 jbub joined #salt
02:14 higgs001 joined #salt
02:14 OrbHacker anyone know if zeromq is still required for 17.2?
02:14 Psi-Jack Of course it is.
02:14 OrbHacker i have zeromq3, but is zeromq needed?
02:14 OrbHacker it didnot isntall as a dependancy..?
02:15 Psi-Jack I wouldn't recommend using the zeromq3 repo, it broke so much, while EPEL's has a solid package.
02:15 OrbHacker roger: I used the epel
02:15 OrbHacker but when I did it just now, I see that zeromq3 was installed
02:15 Psi-Jack zeromq3 is not in epel, zeromq is
02:15 Psi-Jack Oh yeah, nevermind.
02:15 OrbHacker however, the older zeromq was not...
02:16 OrbHacker Available Packages
02:16 OrbHacker zeromq.i686                                                         2.2.0-4.el6                                                epel
02:16 OrbHacker zeromq.x86_64                                                       2.2.0-4.el6                                                epel
02:16 Psi-Jack zeromq3 is correct.
02:16 OrbHacker ok
02:16 OrbHacker cool
02:16 OrbHacker just like you said, zeromq3 is epel: 3.2.4-1.el6
02:16 OrbHacker awesome sauce
02:16 OrbHacker hope this works better!
02:17 OrbHacker Salt: 0.17.2
02:17 OrbHacker Python: 2.6.6 (r266:84292, May 27 2013, 05:35:12)
02:17 OrbHacker Jinja2: unknown
02:17 OrbHacker M2Crypto: 0.20.2
02:17 OrbHacker msgpack-python: 0.1.13
02:17 OrbHacker msgpack-pure: Not Installed
02:17 OrbHacker pycrypto: 2.0.1
02:17 OrbHacker PyYAML: 3.10
02:17 OrbHacker PyZMQ: 2.2.0.1
02:17 Psi-Jack Pleae stop flooding already.
02:17 OrbHacker ZMQ: 3.2.4
02:17 OrbHacker jesh
02:17 OrbHacker :p
02:18 prooty hello. i have a question about the includes statement.
02:18 prooty in file main.sls, i include sub1 and sub2.
02:19 prooty can sub1 access a state in sub2 even if i don't include sub2 in sub1?
02:20 Psi-Jack prooty: No.
02:20 prooty i see, thanks Psi-Jack.
02:27 dave_den prooty: what do you mean 'access'?
02:27 prooty dave_den, specifically there's a file state in sub2 that i'd like to require in a state in sub1.
02:27 dave_den like 'require' a state from sub2 in sub1 when you run your main.sls state?
02:28 dave_den prooty: yes, you can do that if you run the state from main.sls so that it includes sub1 and sub2.
02:29 prooty oh, that's good news then.
02:29 dave_den salt will compile the entire highstate data before it runs the state.
02:29 dave_den so as long as something incliudes sub1 and sub2 in the state run, then sub1 can require states from sub2.
02:31 dave_den prooty: for example:  https://gist.github.com/dlanderson/bc055629f58ada98c24b
02:34 prooty thanks.
02:36 dave_den no problem
02:37 APLU joined #salt
02:38 OrbHacker schwing
02:38 OrbHacker RHEL6 is working well...
02:38 OrbHacker thanks again <Psi-Jack>
02:40 Katafalkas joined #salt
02:42 OrbHacker is it expected behavior for a state.highstate call to return nothing back on theconsole?
02:45 prooty can i test if a state exists before requiring it?
02:47 _ilbot joined #salt
02:47 Topic for #salt is now Welcome to #salt - SaltConf in January!! http://saltconf.com | 0.17.2 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/
02:49 dave_den prooty: you can, but if you are having to do that, then maybe you may want to rethink the structure of your states
02:50 dave_den prooty: you can use cp.list_states to see the states that exist on the master for your envirnment
02:50 dave_den prooty: http://docs.saltstack.com/ref/modules/all/salt.modules.cp.html#salt.modules.cp.list_states
02:51 prooty hmm...
02:55 ajw0100 joined #salt
02:56 OrbHacker night, thanks again!
02:56 OrbHacker left #salt
02:57 higgs001 joined #salt
02:57 OrbHacker joined #salt
02:58 OrbHacker joined #salt
03:02 sephoreph joined #salt
03:04 favadi joined #salt
03:04 sephoreph Hi, I'm using a simple init.sls script for an nginx installation with a custom built RPM.  It works fine, but I'm trying to upgrade it from 1.2.x to 1.4.x, but because nginx itself is already installed it's not removing and then upgrading to the new RPM.  I can't use pkg.latest because I'm using Sources: (salt://nginx/nginx-{{ nginx_ver }}.{{ grains['cpuarch'] }}.rpm)
03:04 sephoreph Do I have to first purge the old package and then install the new one or is there a way to automate this?  rpm -qa | grep nginx returns nginx-1.2.8-1.ngx so it should be detecting the version change I'd have thought?
03:09 cnelsonsic joined #salt
03:11 fllr joined #salt
03:15 bhosmer joined #salt
03:19 cowyn__ joined #salt
03:20 jpcw_ joined #salt
03:20 mattmtl joined #salt
03:22 mgw joined #salt
03:27 sroegner joined #salt
03:28 Nub682 joined #salt
03:33 Gifflen joined #salt
03:34 StDiluted joined #salt
03:35 jslatts joined #salt
03:40 shinylasers joined #salt
03:42 nmistry joined #salt
03:50 mgw joined #salt
03:50 forresta joined #salt
03:58 fllr joined #salt
04:16 xmltok joined #salt
04:24 jergerber joined #salt
04:33 jcsp joined #salt
04:40 fllr joined #salt
04:40 apergos joined #salt
05:04 zach joined #salt
05:14 forresta whiteinge, you around?
05:18 bemehow joined #salt
05:27 StDiluted joined #salt
05:27 sroegner joined #salt
05:35 fllr joined #salt
05:37 amckinley joined #salt
05:39 giantlock joined #salt
05:40 nmistry joined #salt
05:44 ckao joined #salt
05:45 Marion joined #salt
06:02 Destro joined #salt
06:05 fandikurnia01 joined #salt
06:22 vipul joined #salt
06:34 fllr joined #salt
06:57 amckinley joined #salt
07:00 redondos joined #salt
07:04 prooty when there are multiple include statements in an sls file, is the last the only one used?
07:05 prooty not two items in a single include statement. multiple include statements.
07:17 prooty left #salt
07:21 troyready joined #salt
07:28 sroegner joined #salt
07:37 matanya joined #salt
07:39 fllr joined #salt
07:41 juasiepo joined #salt
07:42 fllr joined #salt
07:46 ninkotech joined #salt
07:46 svx joined #salt
07:54 tempspace joined #salt
07:55 robinsmidsrod joined #salt
07:57 jasonrm joined #salt
07:58 junedm joined #salt
08:01 balboah joined #salt
08:05 ntt_ joined #salt
08:07 ConceitedCode joined #salt
08:07 jpcw joined #salt
08:10 harobed joined #salt
08:12 agh hello to all
08:13 malinoff Hi
08:13 junedm joined #salt
08:13 junedm left #salt
08:13 agh i've a file managed with jinja, and i want to do this in this file :
08:13 agh {% if grains['host'] == '*-priv-*' %}
08:13 agh bla bla bla
08:13 agh {% endif %}
08:13 agh but... it does not work :/
08:13 agh there is no error
08:13 agh but it seems that '*-priv-*' does not work
08:13 malinoff agh, You should specify template: jinja argument in file.managed state
08:14 agh malinoff: it's done
08:14 malinoff Ah
08:14 malinoff You want to use globbing
08:14 malinoff Right?
08:14 agh heu, what is globbing ?
08:14 malinoff http://en.wikipedia.org/wiki/Glob_%28programming%29
08:15 agh malinoff: ah yes, that's it
08:15 agh malinoff: doing this is working in pillars, but does not seem to work in jinja files
08:15 malinoff agh, Jinja is more like python, not bash. You should use something like {% if '-priv-' in grains['host'] %} blah {% endif %}
08:16 agh malinoff: i'll test that
08:17 agh malinoff: does not match
08:18 ntt_ Hi. Someone can help me with this -> http://pastebin.com/BrN6ekF5   I cannot print value from a dictionary with jinja in an sls file
08:18 malinoff agh, Write {{ grains['host'] }} and show the result
08:19 malinoff Or just run salt your-minion-id grains.get host
08:19 agh malinoff: dc1-web-gw-priv-1
08:19 malinoff ntt_, {{ n['hostname'] }} ?
08:19 malinoff ntt_, Python is not JavaScript
08:20 malinoff agh, strange thing...
08:20 ntt_ malinoff: i have the same error...
08:20 agh malinoff: agreed... it's working in pillars, so, i'm going to got by pillars instead. But yeah, it's strange
08:21 malinoff agh, Write {{ '-priv-' in grains['host'] }} and watch the result
08:21 agh malinoff: ok
08:21 malinoff ntt_, What error?
08:21 malinoff Ah
08:21 malinoff I see
08:21 ntt_ ok
08:22 malinoff So you can't use dictionaries, that's what the error about
08:22 agh malinoff: True
08:23 malinoff agh, so the check works fine
08:23 malinoff Interesting
08:23 agh malinoff: yes
08:23 ntt_ but i have a list of hostname and i should assign some key/value to each of them. How can i solve?
08:23 malinoff Can you pastebin the whole state?
08:24 ntt_ malinoff: are you talking with me?
08:24 malinoff ntt_, No, with agh
08:24 agh malinoff: oh
08:24 malinoff ntt_, But I'm trying to understand your error message
08:24 agh malinoff: it's working now :)
08:24 malinoff agh, Lol
08:24 agh malinoff: i really do not understand :/
08:24 malinoff agh, Programming is funny
08:24 agh malinoff: i've just removed the '-' ('priv' and not '-priv-'
08:25 pengunix joined #salt
08:25 agh malinoff: ok. Thanks a lot for your help
08:25 srage joined #salt
08:27 slav0nic joined #salt
08:29 giantlock_ joined #salt
08:32 ntt_ malinoff: have you another solution for my problem? basically i have an hostname and i want to relate to this hostname a list of key/value items. This items should be readed from sls files using the hostname (or grains['id']). Is this possible?
08:35 agh is it possible to do {% elif %} alone, without condition ?
08:35 agh {% if  'priv' in grains['host'] %}
08:35 agh bla bla
08:35 agh {% elif %}
08:35 agh toto
08:35 agh {%endif%}
08:36 redondos joined #salt
08:39 srage_ joined #salt
08:39 fllr joined #salt
08:40 bhosmer joined #salt
08:47 mike251 joined #salt
08:48 mike251 hey guys .... i am doing a custom mysql installation ... and i want to run also some queries ... what is in your opinion the best option to run some queries after the mysql is setup? should i add them to a *.sh and use cmd.wait ? ... any other ideas?
08:49 mike251 what about setting up the root password for mysql on centos? how should i do that?! ... on centos one can use mysql_secure_installation
08:49 mike251 how can i proceed with salt to set up a mysql root password?
08:53 malinoff agh, of course
08:54 agh malinoff: ok. Thanks again a lot for your help. Everything is fine now
08:55 psyl0n joined #salt
08:56 Veticus joined #salt
09:03 Veticus joined #salt
09:07 ConceitedCode joined #salt
09:07 zooz joined #salt
09:15 nebuchadnezzar hello
09:17 prooty joined #salt
09:18 Destro Hey hey
09:18 nebuchadnezzar any idea why pillar flattening does not work as describe in the documentation, my problem is related in http://paste.debian.net/70473/
09:29 achileuss joined #salt
09:29 sroegner joined #salt
09:30 ninkotech_ joined #salt
09:34 Iwirada joined #salt
09:36 jinnko joined #salt
09:37 carlos joined #salt
09:39 fllr joined #salt
09:41 davidone joined #salt
09:42 jakubek {% if value['ip'] == salt['network.ip_addrs'](value['if'])[0] %} /// how to propely insert value['if'] into this bracket?
09:43 jakubek is there any good way to do this?
09:43 malinoff jakubek, Looks good
09:43 malinoff What's the problem?
09:45 jakubek malinoff: http://pastebin.com/QxNivPS3, value is not inserted to template
09:45 nebuchadnezzar relative to http://paste.debian.net/70473: the pillar flattening order is first one in alphabetical order
09:46 Shenril joined #salt
09:46 jakubek malinoff: i have updated paste with error
09:48 malinoff jakubek, Right, because pillar['ha_nodes'].items() returns {'jakubest1': {'if': 'eth1', 'ip': '192.168.212.221', 'ip_class': 22, 'master': 1}}
09:48 malinoff jakubek, So you need one more loop
09:49 malinoff Oh wait
09:49 malinoff This should work...
09:50 slav0nic joined #salt
09:50 malinoff jakubek, Can you just write {{ value['if'] }} ?
09:50 malinoff Will that work?
09:51 jakubek malinoff: like this? salt['network.ip_addrs']({{ value['if'] }})[0]
09:51 jakubek because this value needs to be in bracket
09:51 malinoff jakubek, No, just {{ value['if'] }} in the for loop
09:53 elfixit joined #salt
09:54 jakubek malinoff: no i can't :/
09:54 prooty hello. the first time the highstate is called for a supervisord.running state it always returns ERROR (already started). subsequent calls are fine: pastebin.com/aXVP1EdK
09:55 prooty http://pastebin.com/aXVP1EdK
09:55 malinoff jakubek, can you change 'if' to 'if_'?
09:55 pengunix joined #salt
09:56 malinoff Both in pillars and in the template
09:58 jakubek malinoff: and how should bracket look like?
09:58 malinoff jakubek, {{ value['if_'] }}
09:59 che-arne joined #salt
09:59 jakubek malinoff: if value['ip'] == salt['network.ip_addrs']({{ value['ifx'] }})[0], like this? i changed that for ifx
10:00 malinoff jakubek, yeah
10:00 jakubek not working :/
10:00 jakubek same situation
10:01 jakubek malinoff: http://pastebin.com/QxNivPS3
10:01 malinoff Damn. Like that: if value['ip'] == salt['network.ip_addrs'](value['ifx'])[0]
10:01 jakubek different error
10:02 malinoff You should not use quotes around value['ifx']
10:02 malinoff Just copy the string above and try
10:03 jakubek malinoff: http://pastebin.com/QxNivPS3, nope ;/
10:03 jakubek malinoff: when i insert just salt['network.ip_addrs']('eth1')[0] its working fine
10:04 jakubek but hardcording is not good idea in this situation :-)
10:04 malinoff jakubek, what will {{ value.items() }} return in the for loop?
10:06 jakubek malinoff: +[('ip', '192.168.212.222'), ('ifx', 'eth2'), ('ip_class', 22), ('master', 0)]
10:07 pengunix joined #salt
10:08 malinoff jakubek, So there IS 'ifx' key in value dict
10:08 malinoff Idk what else to say
10:08 malinoff Have no idea what's going on :(
10:14 jakubek malinoff: :/
10:22 pengunix_ joined #salt
10:28 sylvinus joined #salt
10:33 nmistry joined #salt
10:34 pengunix joined #salt
10:37 sylvinus joined #salt
10:39 fllr joined #salt
10:40 sylvinus_ joined #salt
10:43 faust joined #salt
10:44 ntt_ can i declare a jinja variable global? i'd like to access this variable from 2 sls files.
10:45 scott_w joined #salt
10:46 backjlack joined #salt
10:47 malinoff ntt_, you can write global.template with content {% set globalvar = 'abc' %} and use {% from global.template import globalvar %}
10:48 malinoff http://jinja.pocoo.org/docs/templates/#import
10:54 ntt_ malinoff: where i have to put global.template file? i'm putting in /srv/salt/ but doesn't work..... i have an "undefined jinja variable " error
10:55 malinoff ntt_, this file should be present on minions
10:56 malinoff ntt_, But I think, pillars/grains is way better to do that
10:56 ntt_ yes
11:13 davidone joined #salt
11:14 jotterbot1234 joined #salt
11:23 Spidfire joined #salt
11:30 sroegner joined #salt
11:30 Spidfire I don't seem to understand the "require" attribute naming, http://pastebin.com/Wx8eiB8D  does someone have a hint?
11:34 ntt_ http://pastebin.com/TQ0nLtm2  --> How can i access to the nested elements in the grains item?
11:34 malinoff ntt_, salt['grains.get']('gpus:model')
11:34 malinoff ":" is the trick
11:36 anuvrat joined #salt
11:36 ntt_ i'm doing  sudo salt '*' grains.item gpus:model    but doesn't work
11:37 ntt_ (not working with .get also)
11:38 malinoff What about salt '*' grains.item "gpus:model" ?
11:38 matanya joined #salt
11:38 ntt_ malinoff: nothing
11:38 ntt_ (uff)
11:39 fllr joined #salt
11:39 malinoff ntt_, get should work, item should not
11:39 malinoff Try salt '*' grains.get "gpus:model"
11:40 malinoff http://docs.saltstack.com/ref/modules/all/salt.modules.grains.html?highlight=grains.item#salt.modules.grains.get
11:40 ntt_ malinoff: i'm trying.... but nothing
11:41 malinoff ntt_, So you're doing something wrong, or you're using a buggy version
11:42 ntt_ i'm using salt 0.17.2
11:42 ntt_ is this a bugged version?
11:42 malinoff A bit :)
11:42 malinoff Like any other salt version :D
11:43 malinoff But v0.17.2 should work :)
11:43 malinoff Uhm
11:43 malinoff MUST work
11:44 ntt_ this is frustrating....
11:51 Spidfire Does someone have any idea how you can add a mysql user and mysql database as a requirement to an other state?
11:53 bhosmer joined #salt
11:54 pengunix_ joined #salt
11:58 bezaban is the 'spawning of lots of minion states when master is unavailable' bug in 0.17.2?
11:58 zandy joined #salt
12:05 aleszoulek joined #salt
12:06 mikkn Spawning sounds ominous
12:14 zooz joined #salt
12:21 mikkn Does anyone know if you could programmatically activate an ext_pillar instead of doing it through the conf file?
12:24 sambaiah joined #salt
12:25 svx joined #salt
12:27 dstanek joined #salt
12:37 jcockhren bezaban: that was one of the bugs supposedly sqauashed in 0.17.2
12:39 junedm1 joined #salt
12:39 fllr joined #salt
12:41 Sheco joined #salt
12:42 viq Question: I have a pkg.installed state with sources link to a specific version of a package. I changed the version of the package in the sources line, but running the state does not update the package. How would I make it update?
12:43 viq 0.17.2
12:43 MrTango joined #salt
12:48 sylvinus joined #salt
12:50 slav0nic joined #salt
12:50 viq ah, https://github.com/saltstack/salt/issues/7772
13:01 viq https://gist.github.com/viq/7927588 - any idea why this fails?
13:02 junedm joined #salt
13:02 junedm left #salt
13:02 _ikke_ I've seen that earlier
13:02 whiskybar joined #salt
13:06 jfzhu_us joined #salt
13:06 _ikke_ viq: https://github.com/saltstack/salt/issues/8945
13:07 blee joined #salt
13:11 cron0 joined #salt
13:16 ConceitedCode joined #salt
13:17 oz_akan_ joined #salt
13:17 viq _ikke_: ah, thanks
13:18 viq I guess off to patch yet another file waiting for newer release to hit repos
13:18 oz_akan_ joined #salt
13:22 sylvinus joined #salt
13:25 che-arne joined #salt
13:29 btorch older versions of salt is able to talk to a new 0.17.2  master without any issues right ?
13:29 btorch this is just while an upgrading is done
13:29 jeddi joined #salt
13:30 viq btorch: should be, I believe. General advice is master needs to be at least as new as the minions
13:30 viq AFAIK
13:30 btorch ok so upgrade master first and then minions
13:30 viq Hm, I have file.managed with source http, I updated the path and checksum, yet file wasn't updated, is that expected?
13:30 viq yeah
13:31 sroegner joined #salt
13:39 fllr joined #salt
13:42 robinsmidsrod joined #salt
13:44 jslatts joined #salt
13:44 dstanek joined #salt
13:46 mike251 joined #salt
13:47 brianhicks joined #salt
13:55 sunand joined #salt
13:56 bhosmer joined #salt
13:56 pengunix joined #salt
13:56 sunand hi, is there an option like - y(yes for all it asks )  for salt-ssh command ?
13:57 ipmb joined #salt
13:57 tyler-baker joined #salt
13:57 _ikke_ "yes salt-ssh .."?
13:57 sunand yes
13:57 _ikke_ "yes | head"
13:57 _ikke_ there is a command called yes
13:59 sunand i should call like   "yes salt-ssh * salte.sls <statename>"
13:59 sunand is it?
14:00 _ikke_ yup
14:01 sunand okay. thank you
14:03 pengunix_ joined #salt
14:04 sunand didnt work :(   neither   "yes | salt-ssh * salte.sls <statename>"
14:07 fllr joined #salt
14:07 pengunix joined #salt
14:08 fllr joined #salt
14:11 viq Hm, there is a freebsdservice module, which takes a 'config' argument. Is there a way to specify a default value for that so I don't have to enter it every time as I don't like the hardcoded one?
14:11 juicer2 joined #salt
14:11 Gifflen joined #salt
14:12 vejdmn joined #salt
14:13 smccarthy joined #salt
14:13 racooper joined #salt
14:14 pengunix_ joined #salt
14:17 mpanetta joined #salt
14:19 sunand _ikke_: any other suggestions?
14:19 rgbkrk joined #salt
14:20 _ikke_ sunand: nope
14:22 viq Any idea how to provide a 'config' option to all service states for all FreeBSD machines?
14:24 penguin_dan joined #salt
14:24 yannj joined #salt
14:25 yannj Hello, is there a page to explain the good practices ? between pillar and grains as an example, is there a drawback in using pillar targetting in states?
14:29 pdayton joined #salt
14:32 juasiepo joined #salt
14:33 viq Ah, you can set a pillar with it
14:34 pengunix hi
14:34 pengunix Can salt execute all unrelated operations in parallel?
14:34 nmistry joined #salt
14:35 yannj viq: is it an answer to my question?
14:36 Destro left #salt
14:38 quickdry21 joined #salt
14:39 th3reverend joined #salt
14:39 sroegner joined #salt
14:39 th3reverend left #salt
14:39 fllr joined #salt
14:41 favadi joined #salt
14:41 Sheco joined #salt
14:42 amahon joined #salt
14:45 NotreDev joined #salt
14:45 favadi joined #salt
14:47 gammalget joined #salt
14:49 pengunix ?
14:51 jergerber joined #salt
14:53 JulianGindi joined #salt
14:55 wherebot joined #salt
14:59 wherebot joined #salt
15:01 a1j joined #salt
15:05 JulianGindi left #salt
15:10 rlarkin joined #salt
15:11 wkf joined #salt
15:14 zgre joined #salt
15:17 dccc joined #salt
15:18 fllr joined #salt
15:19 bhosmer joined #salt
15:20 btorch is there a way to overwrite the state_output that is configured in the master config using the cli ?
15:20 btorch I've tried using --state-output and --out but still returning output as terse
15:22 db joined #salt
15:26 bhosmer joined #salt
15:30 AdamSewell joined #salt
15:30 AdamSewell joined #salt
15:33 forresta joined #salt
15:33 kermit joined #salt
15:34 MZAWeb joined #salt
15:40 lineman60 joined #salt
15:44 viq yannj: no, it was an answer to mine ;)
15:44 xmltok joined #salt
15:44 viq yannj: as for your question, personally I am trying to avoid setting custom grains, mainly for one reason: grains are local, I want to have as much centralised as possible
15:44 FL1SK joined #salt
15:44 kiorky joined #salt
15:48 dstanek joined #salt
15:48 robinsmidsrod joined #salt
15:53 robinsmidsrod joined #salt
15:54 NotreDev joined #salt
15:54 FL1SK joined #salt
15:57 robinsmidsrod joined #salt
15:58 pdayton joined #salt
15:59 robinsmidsrod joined #salt
16:00 rics joined #salt
16:02 vejdmn joined #salt
16:02 robinsmidsrod joined #salt
16:02 amahon joined #salt
16:06 smccarthy joined #salt
16:08 davet joined #salt
16:08 Iwirada left #salt
16:11 scoates Psi-Jack: any chance you might have had epiphanies about what I'm doing wrong? (-:
16:12 meteorfox joined #salt
16:13 robinsmidsrod joined #salt
16:13 meteorfo_ joined #salt
16:14 cachedout joined #salt
16:18 FL1SK joined #salt
16:21 mgw joined #salt
16:22 NotreDev joined #salt
16:23 robinsmidsrod joined #salt
16:25 hunternet93 joined #salt
16:26 hunternet93 How can I allow all users in a certain group on the Salt master to execute commands? I know I can enable individual users with client_acl, how can I use a group?
16:28 mike251 left #salt
16:28 hunternet93 After more searching, external_auth seems like what I need, but I still don't see a way to specify a group.
16:29 forresta hunternet93, there was this issue opened: https://github.com/saltstack/salt/issues/3238 , which leads to https://groups.google.com/forum/#!topic/salt-users/IPWYAyzxDF8
16:31 dfinn joined #salt
16:31 hunternet93 forresta: I don't see a resolution in that topic, did I miss something or has the feature not been implemented yet?
16:31 dfinn is this the wrong way to have a file require a package be installed?  it seems right according to the docs however I'm getting an error.  code and error here:
16:31 dfinn http://pastebin.com/BtdXnESR
16:31 forresta hunternet93, sorry got distracted and didn't finish my last thought. From what I see it isn't implemented yet
16:31 forresta hey basepi, did you ever talk to Tom about https://github.com/saltstack/salt/issues/3238 ?
16:32 forresta dfinn, does the pkg exist in the same state?
16:32 dfinn yes
16:32 forresta ok, can you paste your state?
16:32 dfinn and on the machine that is giving the error it's actually already installed
16:32 dfinn yes, one sec
16:33 dfinn http://pastebin.com/P15ZCALB
16:33 forresta yea it won't matter if it's installed because when you do that check, it's looking for the declaration of that item, NOT against the server itself. The pkg check actually looks for that :P
16:33 MrTango joined #salt
16:33 forresta dfinn, yea here is why
16:33 dfinn hmm…not sure I follow
16:33 forresta dfinn, ok let's look at it like this
16:33 forresta on your paste line 32
16:33 hunternet93 dfinn: Try requiring nagios_pkg instead of net-snmpd
16:34 forresta exactly
16:34 dfinn ah, ok
16:34 forresta you have an item which you have defined, nagios_pkgs
16:34 forresta inside of that are multiple items which are associated with what you have declared
16:34 Veticus joined #salt
16:34 forresta Salt sees those items as all part of nagios_pkgs.
16:34 forresta so when you try to require something from inside of it, Salt doesn't understand, because it doesn't see ONLY net-snmp
16:35 dfinn i think i get it, i thought it was referencing the literal package but I guess not
16:35 forresta correct
16:35 meteorfox joined #salt
16:35 dfinn that worked, thanks guys!
16:35 forresta yup
16:37 dottedmag left #salt
16:37 jalbretsen joined #salt
16:38 nmistry joined #salt
16:38 sylvinus_ joined #salt
16:39 forresta hunternet93, I'm curious as to how much of a change that would take to the existing client_acl code to resolve
16:41 Ryan_Lane joined #salt
16:41 hunternet93 for pam, probably not a lot
16:41 shinylasers joined #salt
16:41 hunternet93 not sure about ldap
16:44 FL1SK joined #salt
16:45 Furao joined #salt
16:46 alunduil joined #salt
16:46 mr_chris joined #salt
16:49 shinylasers joined #salt
16:51 jcsp joined #salt
16:57 scoates is there a way for me to minimize network traffic when doing a highstate? e.g. I don't want it to implicitly do `apt-get update` or all of the `git ls` that seems to be going on…
16:58 * Gareth waves
17:00 jeddi scoates: I think (!) there's a way of setting a refresh time / expiry time for running an update.   Alternatively set up something like apt-cacher-ng ... so that these things aren't impacting your external network bandwidth / quota.
17:00 scoates jeddi: I'll check out the cacher. thanks.
17:00 hunternet93 scoates: you could set up an apt-cache server, not exactly a salt fix but will save a ton of bandwidth
17:00 jeddi I think, by the way, that apt-get update doesn't pull down release or patch files that it already has - so you'll probably find that the network usage is actually quite low, despite the frequency.
17:00 scoates hunternet93: we already have our apt repos on s3. This is for VM testing, and I'm on a bad connection today, so I'm really noticing the lag
17:01 jeddi local network for apt-cacher-ng makes a huge difference.
17:01 FL1SK joined #salt
17:01 jeddi especially once you have 2 or more machines. :)
17:04 jinnko joined #salt
17:06 Tekni joined #salt
17:06 rgarcia_ joined #salt
17:07 sylvinus_ joined #salt
17:10 cdcalef joined #salt
17:11 troyready joined #salt
17:13 sine_nitore joined #salt
17:13 druonysus joined #salt
17:14 junedm joined #salt
17:16 jdenning joined #salt
17:21 xinkeT joined #salt
17:22 zandy joined #salt
17:23 StDiluted joined #salt
17:25 vejdmn joined #salt
17:28 KyleG joined #salt
17:28 KyleG joined #salt
17:29 nmistry joined #salt
17:35 redondos joined #salt
17:36 NotreDev joined #salt
17:41 JordanRinke Morning
17:41 forresta hey
17:49 dave_den joined #salt
17:49 bhosmer joined #salt
17:50 mVla joined #salt
17:50 mVla hello
17:51 mVla when I add a module in _modules directory, di I have ti restart teh stlt master ?
17:53 zandy joined #salt
17:59 dave_den mVla: no, you should not have to restart the master
17:59 mVla ok thanks
18:00 dvogt joined #salt
18:00 bemehow joined #salt
18:00 mVla anyway , all my minions answered "gridengine.set_admin_host" is not available.
18:01 nebuchadnezzar mVla: it looks like not, if I understand correctly http://docs.saltstack.com/ref/configuration/minion.html#autoload-dynamic-modules
18:01 mVla gridengine.py is present on all my minions
18:01 jinnko joined #salt
18:01 mgw joined #salt
18:01 forresta http://docs.saltstack.com/ref/#minion-execution-modules
18:01 forresta that states you need to add the modules to the modules directory then restart the minion
18:02 jacksontj joined #salt
18:02 mVla oh !
18:02 forresta can you try to restart one minion's salt-minion service, then query the grains for that machine?
18:02 forresta if these are minion execution modules that is :P
18:02 mVla ok
18:02 mVla :)
18:02 forresta litle bit of ambiguity there
18:02 forresta so maybe it will work
18:03 mVla I restarted the salt minion, still don't works
18:04 nebuchadnezzar should I open a bug for http://paste.debian.net/70473: the pillar flattening order is first one in alphabetical order and not last declared one like suggested in the documentation
18:05 mVla thank you all
18:05 mVla I'll see all that tomorow
18:05 mVla bye
18:05 mVla && thanks
18:06 forresta nebuchadnezzar, I believe an issue already exists for that, might be worth checking
18:06 nebuchadnezzar forresta: thanks
18:06 forresta np
18:07 shinylasers joined #salt
18:07 mVla left #salt
18:10 thrashr888_ joined #salt
18:12 vbabiy is there a way to load a salt console that to help debug modules?
18:13 vbabiy using ipython or something
18:16 vbabiy Also with gitfs, is there a way to force a sync
18:16 blee_ joined #salt
18:16 vbabiy There are times when I have to run highstate multiple times to get it to sync?
18:18 Ryan_Lane joined #salt
18:19 KyleG joined #salt
18:19 KyleG joined #salt
18:20 [diecast] joined #salt
18:24 amckinley joined #salt
18:32 backjlack joined #salt
18:34 * Gareth ponders adding a module for incrond
18:36 forresta vbabiy, you should be able to import and mess with modules using some form of http://docs.saltstack.com/ref/modules/#cross-calling-modules I believe.
18:36 vbabiy I was hoping to get access to __salt__
18:36 vbabiy so I can call mine.get
18:36 vbabiy or anything along those lines
18:37 forresta what about using the pydsl?
18:37 forresta http://docs.saltstack.com/ref/renderers/all/salt.renderers.pydsl.html
18:38 tyler-baker joined #salt
18:38 tyler-baker joined #salt
18:42 redondos joined #salt
18:43 vbabiy forresta have you used gitfis?
18:43 vbabiy s/gitfis/gitfs
18:46 meteorfox Hi all, has anybody tried GCE with salt-cloud? I can deploy the instance but the master fails to ssh and bootstrap the minion. I set both ssh_username, and ssh_keyfile in the providers files, but it still doesn't work. any ideas?
18:48 bitz joined #salt
18:48 higgs001 joined #salt
18:53 zandy joined #salt
18:53 forresta vbabiy, I've read through the docs, but I haven't set up an instance that uses it. Regarding your question above though, the gitfs stuff should be sync'd whenever you run a highstate. You could try deleting the cache dir if you wanted, but you shouldn't need to. Are you using an old version of Salt?
18:54 vbabiy nope: salt-call 0.17.0-4672-g276fcdd
18:54 fllr joined #salt
18:54 vbabiy yeah I am having a hard figuring out what is causing it. There is nothing in the master log
18:55 vbabiy *time
18:57 forresta Hmm
18:57 forresta yea I'm not sure, haven't seen that issue too much since it was fixed in 0.16.4
18:59 forresta meteorfox, you went through the directions here I assume? http://docs.saltstack.com/topics/cloud/gce.html
18:59 meteorfox yes
19:00 NotreDev joined #salt
19:01 forresta I don't see the guy who wrote it in here, s0undt3ch, have you been in communication with the dude who wrote the GCE support at all? He hasn't been on IRC in 9 weeks.
19:02 meteorfox forresta: no
19:04 meteorfox forresta: looking at the debug trace, it tries to ssh with root, and there it doesn't use the key file.
19:05 forresta hmm, is there anything weird going on with the format for your gce config lines?
19:05 forresta seems pretty straightforward in the docs :\
19:05 uomobonga joined #salt
19:06 forresta meteorfox, I'm wondering if the ssh_keyfile: section has certain character limits or something
19:06 Ryan_Lane joined #salt
19:06 pentabular joined #salt
19:06 meteorfox forresta: I know. I think everything is fine.
19:06 meteorfox forresta: mines is just ssh_keyfile: /root/.ssh/google_compute_engine
19:06 forresta with whatever your user is?
19:06 forresta for the line above
19:08 forresta I mean it's using libcloud, so it should be ok for that aspect...
19:08 meteorfox forresta: right
19:08 forresta what perms do you have on the key
19:09 meteorfox forresta: I don't think it evens get to try to use the key or the user I set, this is the ssh command is trying according to the debug trace  ssh -oStrictHostKeyChecking=no -oUserKnownHostsFile=/dev/null -oControlPath=none root@108.59.84.147 date
19:09 forresta uhh root?
19:09 forresta you can't use the root user
19:10 meteorfox forresta: I know, but I set ssh_username to salt
19:10 jacksontj anyone here know how to pass in kwargs to the salt-api?
19:10 forresta odd
19:10 meteorfox forresta: and it doesn't even try to use a key
19:11 forresta jacksontj, did you look at this page? http://docs.saltstack.com/ref/clients/index.html#client-apis
19:11 forresta there's an example there where they use a kwarg
19:11 forresta meteorfox, yea that is odd.
19:11 forresta meteorfox, if everything else looks good on the google compute engine side, can you create a post on the message board?
19:11 jacksontj forresta: apologies, i meant the rest api (salt-api project) that doc page only includes local clients
19:11 forresta maybe we can get someone who knows more or has actually used it since it just came out :P
19:12 forresta jacksontj, ahh gotcha, yea I'm not sure then, maybe a similar format though?
19:12 jacksontj not sure how to encode a dict in a dict in query params :/
19:12 meteorfox forresta: ok, I'll try to debug it, and if I can't get it I'll post something to the message board. Thanks for your help
19:12 forresta meteorfox, yea sorry I couldn't be more help, if you get it figured out let me know
19:12 forresta jacksontj, yea.
19:13 nebuchadnezzar does it mean something to someone http://paste.debian.net/70554/ ? I found nothing related on github issue tracker
19:14 jacksontj forresta: looks like you can post JSON over, trying that out
19:16 redondos joined #salt
19:16 redondos joined #salt
19:17 bitz joined #salt
19:19 austin987 joined #salt
19:24 [diecast] joined #salt
19:29 KyleG joined #salt
19:29 KyleG joined #salt
19:32 bemehow_ joined #salt
19:34 bemehow joined #salt
19:35 cdk does salt allow you to query what packages have been installed using salt?
19:35 pengunix joined #salt
19:36 vmdsch joined #salt
19:37 allanparsons joined #salt
19:37 allanparsons hello everyone.  I'm using Ubuntu 12.04 + Supervisor (via apt repo) + guinicorn
19:37 scoates Is there a smarter way to do this? (perhaps without the iteration?) http://paste.roguecoders.com/p/28d772c1cd9fbe788a0cba74b3e3074e.txt
19:38 allanparsons and, i'm using the service module rather than the supervisord module.  but, if i create a watch on a file (/etc/supervisor/conf.d/*), it doesnt seem to monitor that directory for changes
19:39 allanparsons so, i deploy via state.highstate, my sls removes the init.d script and replaces with upstart, and then writes files to the conf.d directory, but guinicorn doesnt cycle.  if i run state.highstate a second time, it'll cycle...
19:40 zooz joined #salt
19:40 cdk say i had salt install redis, postgresql and vim using the pkg module on a minion. Can I query the minion and get a list of those packages with their version numbers etc?
19:40 allanparsons http://paste.roguecoders.com/p/9c4be871916a79c854e7e0dcb5d056f8.txt
19:44 pentabular left #salt
19:45 lineman60 when i run salt '*' state.highstat it list the servers but no log of operations is returned nothing in logs any ideas?
19:46 lineman60 also I see a bunch of  Caught signal 15, stopping the Salt Master
19:50 nebuchadnezzar arf, I badly search https://github.com/saltstack/salt/issues/9128
19:51 Gareth lineman60: do you see anything if you turn on more verbose logging?  like -l debug?  so salt -l debug '*' state.highstate
19:52 NotreDev joined #salt
19:53 zandy joined #salt
19:55 ajw0100 joined #salt
19:55 lineman60 Gareth a  missing configueration file .salt.  one server shows nothing, (just server:) the other
19:55 lineman60 [DEBUG   ] loading output in ['/var/cache/salt/master/extmods/output', '/usr/lib/pymodules/python2.7/salt/output']
19:55 lineman60 [DEBUG   ] Skipping /var/cache/salt/master/extmods/output, it is not a directory
19:55 lineman60 [DEBUG   ] Loaded no_out as virtual quiet
19:55 lineman60 [DEBUG   ] Loaded json_out as virtual json
19:55 lineman60 [DEBUG   ] Loaded yaml_out as virtual yaml
19:55 lineman60 [DEBUG   ] Loaded pprint_out as virtual pprint
19:56 Gareth lineman60: You have them configured with some states?
19:56 lineman60 yes, I'm debuging my states
19:57 Gareth Can you put them into a pastebin?
19:57 lineman60 do you think it's a missconfiguered state?  yes I can
19:57 ajw0100 joined #salt
19:58 sylvinus joined #salt
19:59 lineman60 http://pastebin.com/qVmtjghX
20:00 lineman60 That's the most recetn one i've worked on
20:00 lineman60 still need to add piliars
20:01 Gareth lineman60: what does your top.sls look like?
20:05 srage joined #salt
20:06 lineman60 Garth I think i found it I was useing file_roots with no base
20:06 vbabiy Has anyone else had issues with gitfs changes not being applied right away?
20:07 Gareth lineman60: that would do it :)
20:08 kermit joined #salt
20:08 lineman60 Garth forgot I was working on that yesterday
20:08 lineman60 thansk
20:08 Gareth no worries.
20:09 meteorfox forresta: I think I found the bug with GCE. If you look at https://github.com/saltstack/salt/blob/develop/salt/cloud/clouds/gce.py#L545-L546  The dictionary is being created with the keys ssh_username, ssh_keyfile and with pdb I inspected that the values been passed were actually the ones I set. Now, at line 616 it passes that dict to salt.utils.clouds.deploy_script
20:09 meteorfox forresta: which is this file https://github.com/saltstack/salt/blob/develop/salt/utils/cloud.py#L499-L511
20:10 meteorfox forresta: if you look at the kwargs being unpacked, it doesn't match the keys, they are username, and key_filename instead
20:11 [diecast] joined #salt
20:13 hunternet93 left #salt
20:14 allanparsons vbabiy - they usually get applied for me in ~ 10s after checkin
20:16 sroegner vbabiy: i had to go to version 0.17.2 to get gitfs to materialize at all after startup
20:17 vbabiy sroegner was there something changed in 0.17.2?
20:18 sroegner vbabiy: :-) evidently
20:18 allanparsons anyone know how to correctly monitor a directory for changes?
20:18 allanparsons watch:  file:  /path/to/dir/* doesnt seem to work
20:18 allanparsons for recycling a service :(
20:19 allanparsons maybe all conditions in a watch statement need to be met for a service to cycle?
20:20 forresta meteorfox, hmm, seems odd that it wouldn't convert over in some way
20:21 forresta jacksontj, any luck>
20:21 jacksontj forresta: yea, so you can pass it in as JSON
20:21 jacksontj you have to also send it as a list of lowstate objects
20:21 forresta ahh ok
20:21 jacksontj otherwise it stacktraces
20:21 jacksontj https://github.com/saltstack/salt-api/pull/131
20:21 jacksontj i added a check ;)
20:22 forresta nice
20:22 forresta more meaningful errors go
20:22 mmlb joined #salt
20:24 toastedpenguin joined #salt
20:26 bemehow joined #salt
20:30 Corey basepi, terminalmage: Salt have a presence at FOSDEM?
20:31 smccarthy2 joined #salt
20:31 ninkotech_ joined #salt
20:31 basepi Corey: I doubt it.  That'll be *right* after SaltConf, so I think we're all going to be sleeping.
20:32 basepi But I don't know for sure.
20:33 UtahDave joined #salt
20:33 Corey basepi: That was my plan until FOSDEM accepted my salt talk.
20:33 Corey UtahDave: Nice timing.
20:33 UtahDave you're speaking at fosdem?  cool!
20:33 Corey UtahDave: Apparently so!
20:34 lineman60 Does anyone have a 2 or 3  server saltstack enviroment  example posted onilne?
20:34 basepi forresta: sorry i forgot to reply to your mention this morning.  I have not reminded Tom about that issue, but I will.
20:35 Corey UtahDave: Any idea if Salt is sending a delegation?
20:37 jcsp joined #salt
20:38 UtahDave Corey: Yeah, Tom's going to make sure we have a couple people there.  (I'm hoping to be one of them.)
20:40 ajw0100 joined #salt
20:42 nebuchadnezzar is there a way to tell salt to undo a state if it does not apply anymore?
20:42 UtahDave nebuchadnezzar: No. You have to define your own state that puts the minion into the old state you wanted.
20:43 nebuchadnezzar thanks
20:44 Heartsbane joined #salt
20:44 Corey UtahDave: I'll be flying direct from SLC to Brussels. I can apparently drag up to 8 people with me into Economy Plus on United if that helps anything. Cold comfort really, but it costs me nothing to do. :-)
20:44 forresta basepi, no worries.
20:45 Heartsbane joined #salt
20:45 Heartsbane joined #salt
20:46 Heartsbane joined #salt
20:46 UtahDave Corey: dude. drag me with you!
20:46 UtahDave United?
20:48 Corey UtahDave: Always.
20:48 forresta who is this UtahDave imposter?
20:48 Corey UtahDave: FYI, FOSDEM is Feb 1 and 2, then ConfigCamp is 3 and 4. I figured I'd stay the extra two days.
20:48 UtahDave lol.  I'm alive!
20:48 davidfischer joined #salt
20:48 UtahDave Corey: Yeah, I think we've been asked to stay for ConfigCamp, too
20:49 Corey UtahDave: Superb. I'll be booking after Jan 1.
20:49 Corey (I have to spend $10K on United in 2014 to retain my status)
20:50 forresta Corey, oh man I wish I had some vacation time left after saltconf I'd totally be down.
20:50 Corey forresta: Call it work development.
20:50 Corey "I'm not asking you to pay for my conference attendance-- just don't ding me for vacation time" sells REALLY well in most places.
20:51 forresta lol, I'm using my own vacation time and I'm presenting at saltconf, it doesn't work like that.
20:51 Corey forresta: Do you work for the devil?
20:51 forresta No
20:51 forresta back in a few
20:52 Corey k
20:52 dfinn1 joined #salt
20:53 zandy joined #salt
20:54 whiteinge jacksontj: thanks for the pull req! to answer your original question: for urlencoded input, kwargs are passed as strings in the "key=value" format (just like the CLI)
20:54 whiteinge sending JSON is a lot nicer though :)
20:55 JordanRinke process question, if I template a machine with salt already installed - what is the best way to make it generate new pks when it starts its new life after being cloned?
20:55 JordanRinke firs time working with an image that is already bootstrapped
20:59 whiteinge JordanRinke: if you can write the image with an empty /etc/salt/pki dir the minion will generate one the first time it starts, then send it to the master as usual. if you're working with an existing image, you could replace the existing key with salt-cloud's saltify driver (via ssh)
21:00 jacksontj whiteinge: when i did that it showed up as key, val pairs in the lowstate-- not in the kwargs :/
21:00 hellome joined #salt
21:00 forresta Corey, we don't go to conferences that aren't associated with a vendor or product we use
21:00 jacksontj whiteinge: i was thinking about it actually, do we want to do some better validation of that lowstate object?
21:01 jacksontj make sure it has a tgt a fun etc.
21:01 whiteinge the translation from string to dict happens in salt itself, not salt-api
21:02 whiteinge uses the same code that salt uses to process CLI args
21:02 JordanRinke whiteinge: Cool, there is a touch of manual to it right now (vmware clones) so I will just add removing that dir to the runbook for provisioning/naming a new clone
21:02 whiteinge jacksontj: i was thinking the same thing! more validation would be very helpful
21:03 jacksontj so, does salt have a util library for validating lowstates?
21:03 jacksontj if not then we'll have to make one ;)
21:04 jacksontj whiteinge: ah, so i should send (in urlencoded) kwarg={'a': 1}
21:04 whiteinge no. :-/   and each of salt's *Client interfaces are sufficiently different that there are some consistency annoyances when you have to use them directly (e.g., from salt-api or salt.client.api)
21:04 Zuru joined #salt
21:05 JordanRinke whiteinge: ah, I also have to delete minion_id
21:05 jacksontj :/
21:05 pentabular joined #salt
21:05 jacksontj well, we could do some basics-- like is it a dict
21:05 jacksontj :D
21:06 rgbkrk joined #salt
21:07 pentabular left #salt
21:08 whiteinge heh :)
21:09 whiteinge sec...i'll paste an example of kwargs
21:10 forresta alright jacksontj, steal the example and add it to the docs :P
21:10 pengunix joined #salt
21:12 whiteinge curl -sS localhost:8000 -d client='local' -d tgt='*' -d fun='test.kwarg' -d arg='key=value'
21:13 whiteinge but note, that format is different for runner and wheel modules which require the name of the actual kwarg the function is expecting:
21:13 whiteinge curl -sS localhost:8000 -d client='runner' -d fun='jobs.lookup_jid' -d jid=20131206075053468906
21:14 jacksontj i see, so its pretty "raw" with what the underlying salt client arg/kwargs
21:14 whiteinge very. I'm not a fan of the usability implications of that, but that's what salt's python APIs are expecting
21:15 cewood joined #salt
21:16 rgarcia_ joined #salt
21:16 bemehow_ joined #salt
21:17 prooty does the configuration here for salt-cloud still work (https://salt-cloud.readthedocs.org/en/latest/topics/map.html).
21:17 prooty i tried setting minion configuration options but i received KeyError: 'minion'
21:18 whiteinge prooty: what version of salt-cloud?
21:18 prooty salt-cloud 0.17.0-5755-gceab83f
21:18 prooty i installed the latest from git on the salt repo.
21:19 Gifflen_ joined #salt
21:19 whiteinge prooty: oh! no, that doc is wrong
21:19 whiteinge prooty: http://docs.saltstack.com/topics/cloud/map.html
21:20 whiteinge (indent grains up by one level)
21:22 prooty oh i see. thanks.
21:22 prooty that needs to be fixed though.
21:24 bemehow joined #salt
21:25 prooty another question about salt-cloud. i currently have a VM with salt as the hostname.
21:25 prooty if i define a mapping with salt as the hostname, will salt-cloud create a new VM or update the current one?
21:26 prooty the key for the salt VM is already added to the salt-master (itself).
21:28 sylvinus joined #salt
21:31 whiteinge those salt-cloud project docs have been EOL'ed, we should add a note mentioning that...
21:32 whiteinge prooty: salt-cloud won't touch any VMs that already exist when you're using a map file
21:33 whiteinge if you do want to update an exsiting VM the saltify driver in salt-cloud can do that
21:35 Steve____ joined #salt
21:36 SteveL_ joined #salt
21:37 prooty whiteinge: okay, i'll take a look at that.
21:37 [diecast] joined #salt
21:38 prooty how about if i change the grains in the mapping then call salt-cloud -m mapping again, would salt-cloud update the grains in the hosts?
21:38 hellome joined #salt
21:40 hellome joined #salt
21:41 fivethreeo joined #salt
21:41 SteveL_ anyone familiar with file_roots and environments?
21:42 SteveL_ Have not been able to get basic search paths working with environments (with gitfs)
21:43 UtahDave SteveL_: what question do you have?
21:43 hellome joined #salt
21:43 SteveL_ I have a prod env defined with a branch in git and setup file_roots with base and prod
21:44 SteveL_ the minion gets targeted properly but can't pick up the state from a base directory (its a prod target)
21:44 SteveL_ so file_roots:
21:45 SteveL_ base:
21:45 SteveL_ - /srv/salt/base
21:45 SteveL_ prod:
21:45 SteveL_ - /srv/salt/base/prod
21:45 SteveL_ - /srv/salt/base
21:46 SteveL_ have a branch in git called prod and master...
21:46 UtahDave does the state exist in both  / srv/salt/base/prod   and /srv/salt/base    ?
21:46 SteveL_ only in base
21:47 SteveL_ was hoping the search path could just pick up the inherited one in base
21:47 psyl0n joined #salt
21:47 whiteinge prooty: no
21:47 SteveL_ ie: just define the differences in the prod branch
21:47 prooty i see.
21:48 SteveL_ but it can't find the state in prod
21:48 zooz joined #salt
21:48 UtahDave SteveL_: If   foo.sls   doesn't exist in   /srv/salt/base/prod, it should then look in   /srv/salt/base
21:48 SteveL_ works ok if target is base env
21:48 SteveL_ Dave, thats what I thought but its not working
21:49 SteveL_ does using gifts effect that?
21:49 SteveL_ gitfs
21:49 SteveL_ base:
21:49 psyl0n joined #salt
21:49 SteveL_ '*':
21:50 zandy joined #salt
21:50 SteveL_ - core.tools
21:50 SteveL_ prod:
21:50 whiteinge SteveL_: will you please use a pastebin for that?
21:50 SteveL_ ok
21:51 SteveL_ http://pastebin.com/8xx3BJjM
21:51 blee joined #salt
21:51 whiteinge ty
21:52 SteveL_ Any thoughts UtahDave?
21:52 whiteinge prooty: you can update existing grains via salt. salt-cloud only handles that initial deployment. if you use the saltify driver on an existing VM though, salt-cloud will re-bootstrap salt there
21:53 UtahDave SteveL_: what's the path the    smartapps.py?
21:54 SteveL_ its in a gifts repo in the master branch ... path is webapps/smartapps/init.sls
21:54 SteveL_ if I target from base it works ok... just does not see it in prod env
21:55 SteveL_ i'm using salt-call from a minion.... not sure if that effects anything (I do have a master)
21:55 SteveL_ ie:     sudo salt-call state.highstate
21:55 hellome joined #salt
21:55 SteveL_ (on vagrant instance)
21:55 UtahDave what's the name of your minion?
21:56 SteveL_ precise64
21:57 SteveL_ root@ip-10-80-195-125:~# salt-key -L Accepted Keys: precise64 Unaccepted Keys: Rejected Keys:
21:57 ninkotech_ joined #salt
21:58 SteveL_ local:     Data failed to compile: ----------     No matching sls found for 'webapps.smartapps' in env 'prod'
21:58 SteveL_ error above
21:58 UtahDave oh, ok
21:59 UtahDave smartapps.sls needs to be in a "prod" branch in your git repo
21:59 UtahDave The "master" branch maps automatically to Salt's "base" environment
22:00 SteveL_ I guess based on the file_roots I assumed that it would still search base
22:00 jacksontj joined #salt
22:01 SteveL_ as there are 2 paths in the prod: env
22:03 SteveL_ do paths in file_roots map to gitfs branches ?
22:03 sylvinus joined #salt
22:04 UtahDave no
22:04 UtahDave did you set your  "fileserver_backend"  option in your Master config?
22:05 SteveL_ yes
22:05 SteveL_ with roots and git
22:06 UtahDave ok, did you restart your master after modifying the config
22:06 SteveL_ yes
22:06 toastedpenguin1 joined #salt
22:08 UtahDave and is webapps/smartapps/init.sls in a  "prod" branch of your git repo?
22:08 Brew_away joined #salt
22:08 SteveL_ no ... only in base
22:08 UtahDave No. Gitfs doesn't look at the file_roots at all
22:09 SteveL_ ok is there anyway to support inheritance with git?
22:09 UtahDave that's why it's not working.   "prod" environment in Salt looks for a "prod" branch or tag in git
22:09 [diecast] joined #salt
22:10 SteveL_ like with file_roots
22:11 UtahDave Yeah, just like you're trying right now
22:11 UtahDave have some things in your master branch, which maps to the "base" environment. Then have other things in your "prod" branch, which maps to your "prod" environment.
22:11 UtahDave like you have in your top file right now
22:12 Psi-Jack UtahDave: You're back!
22:12 UtahDave yep!
22:12 Psi-Jack UtahDave: Good trip?
22:13 SteveL_ UtahDave: If thats the case then should I not be able to pick up the webapps/smartapps/init.sls from base in prod?
22:14 UtahDave Yeah, it was!  thanks
22:14 SteveL_ I'm missing something here :)
22:14 Psi-Jack heh
22:14 UtahDave SteveL_: no
22:15 Psi-Jack UtahDave: I'd been hunting for you regardnig some pretty interesting environment stuff, but then basically found out my approach would never work without some core changes. heh, fun fun.
22:15 SteveL_ how does the inheritance work then? :)
22:15 UtahDave ah, joy!
22:16 Psi-Jack UtahDave: So, of course, there's a issue in for it. 9186. Stacking States. :)
22:16 scoates environment whatnow? (-:
22:16 forresta scoates, let's not discuss it again
22:16 forresta 3 days in a row is too much for me.
22:17 forresta :P
22:17 scoates I worked around my problem (-:
22:17 scoates couldn't get overwrites/merges to work properly, so I just designed around it. Seems good enough for now.
22:18 cdcalef- joined #salt
22:18 ninkotech_ joined #salt
22:19 Psi-Jack Hmmm, really strange, scoates.
22:19 Psi-Jack I'm actually in the process of converting back FROM gitfs based states and pillars, because that was just a nightmare.
22:21 SteveL_ Psi-jack: kind of wondering the same thing
22:22 Psi-Jack Wondering?
22:22 SteveL_ going back to file based vs git based states
22:22 * Psi-Jack nods.
22:23 SteveL_ Still not clear how the inheritance stuff works ...
22:23 Psi-Jack yeah. gitfs states can get seriously out of whack quickly, with just 3 environments. pillars, a little more ... managable, but still painful to maintain.
22:23 Psi-Jack There is little to no actual inheritence.
22:24 SteveL_ It seems that way
22:24 SteveL_ The file / directory approach seems straight forward... still use git for maintenance but use a directory structure instead
22:25 Psi-Jack heh
22:26 Psi-Jack I'm switching back to git filesystem based, and using make to deploy it. Basically a simple Makefile in the base directory of the git repo that has two functions. push, and deploy. I could add more, but all I need for now is those two, so when I'm in the main directory, all I need to do to update the main server is run "make", and it pushes, and ssh'd into the salt master and pulls.
22:30 KyleG1 joined #salt
22:30 ninkotech_ joined #salt
22:32 jdenning_ joined #salt
22:32 joehoyle joined #salt
22:33 KyleG joined #salt
22:34 joehoyle Hey, anyone know if it is possible to apply a change to apply a change to a file.managed in a file.sed from another sls? Currenly I get "Detected conflicting IDs, SLS IDs need to be globally unique." when trying to do this as the filename is the ID
22:34 Psi-Jack SteveL_: One problem I had for example was gitfs, pushing highstate doesn't work on the first time after an update to git. it seems to only refresh the local gitfs cache /after/ compiling the states. Not before. heh
22:36 NotreDev joined #salt
22:37 joehoyle to answer my own question: Yes, just give the file.sed a unique ID, and use the `- name: ` for the filename
22:37 forresta joehoyle, I was just gonna suggest the same thing :P
22:37 forresta why are you stealing my answers?
22:37 joehoyle haha
22:43 srage joined #salt
22:43 jm-ck joined #salt
22:47 sylvinus joined #salt
22:48 Gifflen joined #salt
22:48 jm-ck hey, does anyone currently have a way to test .sls/jinja rendering, like ... "salt-lint" or something? doesn't look like there is currently a tool
22:49 jm-ck looking for a test that can be run on our master branch to try and prevent syntax-erroring things from getting pushed
22:49 [diecast] joined #salt
22:52 Psi-Jack heh
22:52 srage joined #salt
22:52 Psi-Jack I'm trying to do better, have the ability to test salt states in development environments before pushing them to production or base. ;)
22:53 ninkotech__ joined #salt
22:53 Gareth Psi-Jack: I bet you could do something with jenkins+vagrant for that.  on a commit, spin up a test box, run a highstate, record the output.
22:53 Psi-Jack Ewwww
22:54 Gareth Ew?
22:54 Psi-Jack Yes, eww.
22:54 Psi-Jack For one, key management would be a PITA.
22:54 Psi-Jack Salt uses keys, afteral.
22:55 Gareth should be pretty easy to automate that though.  Keep the machine names the same.  Delete the key after the run.
22:55 Psi-Jack I would sooner, rather, use docker anyway than vagrant.
22:56 Psi-Jack but, even then. No, just stacked states would do the trick rather quite well. ;)
22:57 Gareth *shrug* :)
22:58 Psi-Jack I hate both VirtualBox and VMWare. As hypervisors, they are substandard crap.
22:59 forresta jm-ck, currently something like that doesn't exist. This is the closest thing https://github.com/jesusaurus/salt-shaker
22:59 jcockhren Psi-Jack: what would be standard/up to par?
23:00 Psi-Jack jcockhren: kvm, moslty, but lxc is definitely moving forwards rather quite well.,
23:01 Psi-Jack Both will have their places. :)
23:01 jcockhren lxc+user sandboxing will be awesome
23:01 jm-ck thanks guys, that answers my question
23:01 Psi-Jack Will be? No, IS awesome. :0
23:02 jdenning joined #salt
23:02 jm-ck that would be sweet to have something automatically test it in a container (currently just doing that with masterless salt + vagrant on my workstation)
23:02 jcockhren Was merged in the kernel right? is it already in general distribution? I haven't checked
23:02 jm-ck but uh, i want it to happen automagically and not have to be so careful about syntaxy things
23:02 jm-ck or rather, worry about other team members's syntax
23:03 Psi-Jack heh
23:03 mgw joined #salt
23:03 forresta jm-ck, yea I gotcha, right now there isn't a great way to do it because Salt needs to compile all the data (such as pillar info) together, that script that jesusaurus wrote is basically just passing the test option through when it calls the state. We've discussed getting some sort of syntax checker going, it's just tough to do.
23:03 Psi-Jack One of these days, I might actually check out openstack.. maybe in a virtual machine of my proxmox to see how it works. :)
23:03 forresta Psi-Jack, I've got about 4000 emails from the mailing list right now :\
23:04 Psi-Jack Eh?
23:04 forresta openstack
23:04 forresta just preparing you
23:04 Psi-Jack Oh. heh. Again, I hate mailing lists. :)
23:04 forresta I know
23:04 forresta lol
23:04 jesusaurus forresta: that script can take a --no option to only compile the pillar and not even try to compile the state. i use it as a yaml linter
23:04 Psi-Jack heh
23:05 jesusaurus because currently the only yaml linters out there are rails apps :(
23:05 sylvinus joined #salt
23:05 jesusaurus Psi-Jack: one does not simply check out openstack
23:05 Psi-Jack Sure they do.
23:05 jesusaurus its a huge ecosystem of projects and getting a simple base-case system up and running is not trivial
23:06 Psi-Jack Right now, I'm quite moslty happy with Proxmox VE in my 4-hypervisor node + 3 ceph storage node home cluster setup of virtual machines. OpenStack... I'm just not sure can meet up to my standards. :0
23:06 jfzhu_us joined #salt
23:06 forresta jesusaurus, ahh gotcha, I didn't know you had that option in there.
23:06 jesusaurus forresta: only 4000?
23:06 forresta that's just from the 'help me' mailing list :(
23:06 forresta and that's less than a years worth
23:06 pdayton joined #salt
23:06 Psi-Jack I hate RHCS, which Proxmox VE uses, but.. I see Pacemaker's dev is now at Red Hat Software these days, and integrating pacemaker to work with RHCS now.
23:06 forresta I was gonna try to contribute a while back, but they just drop you in there without any details
23:06 Psi-Jack To replace rgmanager. :0
23:07 jesusaurus yeah, openstack doesnt have a shallow end, you just need to jump in somewhere
23:07 jesusaurus and wherever you pick, its going to be daunting
23:07 forresta yep, it's a tough one to jump into if you don't need to use it eveyr day
23:07 forresta *every
23:07 * jesusaurus recommends starting with python-*client projects
23:08 forresta for the time being I'm gonna stick with Salt, that takes up enough of my off hours work :P
23:12 SteveL_ I switched to using file based inheritance and am now having an inheritance problem :(
23:12 SteveL_ The conflicting ID is "core.tools" and is found in SLS "base:core.tools" and SLS "prod:core.tools"
23:12 Psi-Jack SteveL_: yeah, that's common. Salt states don't currently stack, they conflict.
23:12 SteveL_ have a core/tools.sls in prod and base environments
23:14 dfinn1 is it too redundant to have a salt state that make's sure salt is installed?  obviously none of this will work without it.  we did that in puppet but I'm not sure it's worth moving over
23:14 SteveL_ oops... I meant I only have that in base
23:14 forresta dfinn1, a lot of people run the minion service on there master to salt the Salt master :P
23:15 SteveL_ cleared out all caches also
23:15 dfinn1 ah, that's interesting.
23:15 Psi-Jack forresta: I do! I do!
23:15 Psi-Jack I even use salt to grain my minions.
23:15 forresta There you go dfinn1, so start harrassing Psi-Jack :P
23:15 Psi-Jack dfinn1: using salt to insure salt is installed though.. Not useful.
23:15 Psi-Jack Configuring salt, however.. That's useful
23:16 dfinn1 i guess in the case that we run a custom salt config on the minions it make make sense but as of now we are using the default one
23:17 Psi-Jack I, for example, use my base to handle ALL global things, including the basic salt setup needed for the minions. If grains aren't defined yet, the first highstate sets up the basic initial stuff it needs and applies the base groundwork. The second call to highstate finishes the job of any changes it needs to do by installing all the states based on the whole grains setup.
23:18 dfinn1 good to know
23:18 Psi-Jack I'm planning to, if possible, attempt to inherit the primary base state to help stack it up automatically so only one highstate is needed, but that takes time to work out. :)
23:19 ninkotech_ joined #salt
23:19 canci joined #salt
23:21 SteveL_ Psi-jack: any thoughts on the id conflict? turns out I only have the core.tools in base not prod
23:23 pdayton joined #salt
23:23 ConceitedCode joined #salt
23:24 forresta SteveL_, that should be getting fixed if there is no issue. I'd say start by trashing the /var/cache/salt dir (after backing it up of course), and then restarting the service if it's not getting picked up. If it still occurs after that then something else is probably going on
23:25 NotreDev joined #salt
23:25 srage joined #salt
23:26 Psi-Jack SteveL_: Whats your file_roots look like?
23:27 ninkotech__ joined #salt
23:27 SteveL_ http://pastebin.com/FU63c5we
23:28 SteveL_ i have git repo in the SaltAWS dir
23:32 Psi-Jack Ahh there you go.
23:33 Psi-Jack You're doube-defining roots per environment.
23:33 Psi-Jack This doesn't work as you might want, depending on your top.sls, which is best held in your base only.
23:33 SteveL_ i have top.sls only in base
23:34 ninkotech__ joined #salt
23:34 Psi-Jack You don't need base in every environment, base always is run first.
23:34 Psi-Jack Only include your new environment. :0
23:35 Psi-Jack When this https://github.com/saltstack/salt/issues/9186 gets done, stacked states will be much more possible and overridable by environements. But right now, duplicate states just causes conflict, rather than overrides. :)
23:36 jschairb left #salt
23:36 Psi-Jack Anyway, I'm off to get dinner. Good luck. :0
23:36 kermit joined #salt
23:36 SteveL_ Thanks for the help jack
23:38 Psi-Jack No problem. :)
23:38 Psi-Jack lemme know if that does the trick. Should.. ;)
23:42 dstanek joined #salt
23:46 db joined #salt
23:47 ninkotech__ joined #salt
23:49 KyleG1 joined #salt
23:49 sylvinus joined #salt
23:51 jdenning joined #salt
23:53 NotreDev joined #salt
23:57 sky__ joined #salt
23:58 ninkotech__ joined #salt
23:59 ConceitedCode joined #salt

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