Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2016-09-02

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

All times shown according to UTC.

Time Nick Message
00:04 ninjada joined #salt
00:06 edrocks joined #salt
00:07 mosen joined #salt
00:07 netcho joined #salt
00:10 brotatochip joined #salt
00:11 subsignal joined #salt
00:11 kus joined #salt
00:17 woodtablet left #salt
00:28 subsignal joined #salt
00:29 netcho joined #salt
00:31 sp0097 joined #salt
00:39 subsignal joined #salt
00:40 mattp_ hi guys. is there any way of a per run metadata?
00:55 flowstate joined #salt
00:56 John_Kang joined #salt
01:04 subsignal joined #salt
01:17 justanotheruser joined #salt
01:18 jab416171 joined #salt
01:18 sp0097 joined #salt
01:19 brotatochip joined #salt
01:30 netcho joined #salt
01:42 mosen_ joined #salt
01:45 catpigger joined #salt
01:55 flowstate joined #salt
01:55 aphor joined #salt
02:13 ageorgop joined #salt
02:14 bastiandg joined #salt
02:22 aphor https://trac.macports.org/ticket/52154 <-- yay salt
02:28 fannet joined #salt
02:31 netcho joined #salt
02:41 XenophonF macports is the best
02:43 XenophonF git.latest keeps saying "Repository would be updated to 5edb73a, but this is not a fast-forward merge."
02:44 XenophonF i'm the only one committing to this repo, so i don't understand why it can't do a fast-forward merge
02:48 ageorgop joined #salt
02:49 cyrus_mc joined #salt
02:52 brotatochip joined #salt
02:54 flowstate joined #salt
03:02 DEger joined #salt
03:04 justan0theruser joined #salt
03:05 subsignal joined #salt
03:09 edrocks joined #salt
03:10 DEger joined #salt
03:12 sp0097 joined #salt
03:16 ninjada joined #salt
03:18 sp0097 joined #salt
03:18 debian112 joined #salt
03:28 whitenoise joined #salt
03:32 netcho joined #salt
03:41 alvinstarr joined #salt
03:44 evle joined #salt
03:46 intr1nsic joined #salt
03:53 jaybocc2 joined #salt
03:53 flowstate joined #salt
03:59 jerredbell joined #salt
04:01 debian112 joined #salt
04:05 raspado is it possible to do a salt-call on a function within a state file?
04:05 iggy yes
04:05 mosen_ iirc you can use the salt dict to get a func
04:06 mosen_ ah iggy is here, take it away
04:06 raspado im trying to do something like this but its not working "salt-call state.apply backup.run_backup" where backup is a python script in _modules
04:06 hasues joined #salt
04:06 hasues left #salt
04:06 iggy state.sls_id
04:07 raspado salt-call state.sls_id backup.run_backup ?
04:08 iggy !salt modules.state.sls_id
04:08 saltstackbot https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.state.html#salt.modules.state.sls_id
04:08 raspado ahhhh ok thx iggy!
04:15 penguinpowernz joined #salt
04:16 jimklo joined #salt
04:24 DEger joined #salt
04:33 netcho joined #salt
04:34 debian112 joined #salt
04:41 jaybocc2 joined #salt
04:42 kshlm joined #salt
04:44 Topic for #salt is now Welcome to #salt! | Latest Versions: 2015.8.12, 2016.3.3 | Support: https://www.saltstack.com/support/ | Logs: http://irclog.perlgeek.de/salt/ | Paste: https://gist.github.com/ (please don't multiline paste into channel) | See also: #salt-devel, #salt-offtopic | Ask with patience as we are volunteers and may not have immediate answers
04:54 flowstate joined #salt
05:06 ageorgop joined #salt
05:07 cyborg-one joined #salt
05:10 edrocks joined #salt
05:19 yuhlw_ joined #salt
05:21 pipps joined #salt
05:22 bltmiller joined #salt
05:27 netcho joined #salt
05:33 smcquay joined #salt
05:36 nidr0x joined #salt
05:42 ajv joined #salt
05:50 ronnix joined #salt
05:51 akhter joined #salt
05:51 bltmiller joined #salt
05:54 flowstate joined #salt
05:55 onlyanegg joined #salt
06:02 ttrumm joined #salt
06:05 ub joined #salt
06:05 rdas joined #salt
06:07 DEger joined #salt
06:21 raspado joined #salt
06:21 jimklo joined #salt
06:22 jxm_ joined #salt
06:23 rem5 joined #salt
06:26 jimklo_ joined #salt
06:30 fannet joined #salt
06:33 akunin joined #salt
06:33 bltmiller joined #salt
06:42 akunin joined #salt
06:43 ravenx joined #salt
06:44 raspado joined #salt
06:45 akunin joined #salt
06:50 akunin joined #salt
06:53 jaybocc2 joined #salt
06:55 flowstate joined #salt
06:55 toanju joined #salt
06:55 hasues joined #salt
06:56 hasues left #salt
07:05 akunin joined #salt
07:09 ninjada joined #salt
07:09 jaybocc2 joined #salt
07:12 akunin joined #salt
07:15 remyd1 joined #salt
07:16 akunin need some jinja help, how can I get the IP of the NIC (multihomed host) which has the default gateway set?
07:17 dariusjs joined #salt
07:21 remyd1 Hi folks. I have some issues on this state: http://paste.debian.net/803278/
07:21 remyd1 the "onfail" does not work
07:21 remyd1 I am not able to insert a correct path to the sls file
07:21 remyd1 indeed, I am working in a subdirectory
07:21 remyd1 I tried with just dot, slash, and like here with backslash
07:21 ravenx i also need some help, as of salt-master/minion 2016.3.3 (Boron), I am getting these:  /usr/lib/python2.7/dist-packages/salt/grains/core.py:1493: DeprecationWarning: The "osmajorrelease" will be a type of an integer.
07:21 ravenx is there a way i could suppress it?
07:22 fracklen joined #salt
07:23 ravenx remyd1: i think your onfail needs to read:  cmd: sge_service_exec
07:23 ravenx so:   -onfail:
07:23 ravenx - cmd: sge_service_exec:
07:23 remyd1 thx ravenx. I will try that immediately*
07:24 Elsmorian joined #salt
07:26 remyd1 It seems to work even I do not understand why you have to use "cmd" in this case
07:27 bltmiller joined #salt
07:27 ravenx because the docs says so:  https://docs.saltstack.com/en/latest/ref/states/highstate.html#requisite-reference
07:27 ravenx lol
07:29 remyd1 Hehe. I usually read the doc, but, too fast...
07:30 remyd1 ravenx, on Boron 2016.3.2 I have "osrelease" in my grains but no "osmajorrelease"
07:31 ravenx remyd1: dang..so you are getting those deprecation warnings too?
07:31 remyd1 No, I have not, but I am not using "osmajorrelease" in my recipes
07:32 remyd1 Do you ?
07:32 rsys joined #salt
07:32 ravenx nope
07:32 remyd1 ok
07:32 ravenx i never use that grain, which is why i'm puzzled
07:33 remyd1 Did you look at the source code on github to check the differences on this file ?
07:33 ravenx yup, i found the place:  https://github.com/saltstack/salt/commit/45a12ccaf092ca16e2b196a2440b7cb117622809
07:33 ravenx which looks liek it has to do with grains['osmajorrelease'] that i dont' use.
07:33 ravenx i mean it is removed in the development branch, i ahve to wait until it is released.
07:35 akunin joined #salt
07:35 sfxandy joined #salt
07:36 remyd1 Yes, or you could edit the file on your own :) Just comment it if you want. You will have differences with the main repo if you use git, but you can undo it before fetching/merging
07:37 ravenx ah, so like, just change the source code?
07:38 remyd1 I think it would work
07:38 remyd1 make a copy
07:38 remyd1 before
07:38 ravenx sounds good
07:38 remyd1 changes
07:38 ravenx ah, i installed my salt-master on a venv
07:38 remyd1 ok then
07:38 ravenx so if shit hits the fan i could always just pip install salt again
07:39 ravenx lol
07:42 binocvlar joined #salt
07:43 binocvlar Perhaps a simple one (though I haven't found my answer in the docs so far) - I'm trying to include MANY formulas beneath the 'base' environment in my top.sls, however, I don't want to list them all manually. Is there a simple way to include all sls files beneath a particular directory?
07:44 ravenx ls dir/
07:44 ravenx or are you looking for something that lists it in salt states?
07:44 ronnix joined #salt
07:45 ravenx binocvlar: i dont have that many, but i have 3 i 'import' in
07:45 ravenx {% import 'states/scripts.sls' as scripts %}
07:45 ravenx that's how i do it
07:45 ravenx i have common states that get re-used like git pull, in states/
07:45 ravenx then i import it like so in an init.sls and i can use it.
07:46 binocvlar ravenx: Sounds promising. I was thinking I might be able to just glob all files beneath a directory, but that doesn't seem to work.
07:46 binocvlar Your approach seems simple enough though - I'll give it a try :)
07:47 ravenx my approach maty not work for you cuz you may be importing like 50 .sls files lol
07:47 ravenx you could always try wildcards :P
07:47 binocvlar Yeah, I tried a glob (minions/* in my case), but no go :(
07:48 binocvlar I was thinking of doing something programatic in a statefile to detect other files in the directory and then template out the topfile, but that just seemed... icky
07:49 amy_ joined #salt
07:51 mohae_ joined #salt
07:53 flowstate joined #salt
07:54 remyd1 In the "require" you can put a sls file
07:54 remyd1 You would have to make a loop to retrieve all of your files
07:55 remyd1 but do not know how exactly
07:55 ttrumm joined #salt
07:55 netcho joined #salt
07:55 remyd1 binocvlar, perhaps the best thing to do is either to create your own python module or make a pillar with all your files
07:56 remyd1 you can imagine generate that pillar with a simple python script
07:56 remyd1 that you could run with cron/salt scheduler
07:57 remyd1 and then you include each of it within the loop
08:04 ttrumm joined #salt
08:09 akunin left #salt
08:09 Rumbles joined #salt
08:10 JohnnyRun hi all. On production I feel state.highstate is slow. I test sls states with docker on my machine and the full state.highstate takes 20/25 seconds  for 150 states run. On production it takes abort 90 seconds.
08:11 JohnnyRun How to debug? I feel yum may be slow..
08:11 binocvlar thanks for your advice remyd1 - I'll see what I can work out
08:11 binocvlar Detatching for a sec...
08:11 edrocks joined #salt
08:12 jaybocc2 joined #salt
08:12 binocvlar joined #salt
08:26 jaybocc2 joined #salt
08:28 m0ntez joined #salt
08:28 lero joined #salt
08:31 fannet joined #salt
08:33 Taters_ joined #salt
08:33 m0ntez Hello, if someone has time to lend a hand. On 3 minions out of 80 service salt-minion stops after 5 min or after 30 min. error in /var/log/salt/minion is "[salt.scripts][ERROR   ][10616] argument of type 'int' is not iterable"and "[salt.scripts][WARNING ][10616] ** Restarting minion **". CPU is arm. OS is debian v7, salt-minion version is 2015.5.3. Any ideas please
08:36 idontbyt_ joined #salt
08:37 artemz joined #salt
08:38 Sokel joined #salt
08:39 Rumbles_ joined #salt
08:48 ravenx how do you guys store your returne salt states?
08:48 ravenx influxdb?  postgres?  i'm at a loss to choose one.
08:51 Rumbles joined #salt
08:55 flowstate joined #salt
08:56 Rumbles joined #salt
08:57 Rumbles joined #salt
08:58 fannet joined #salt
09:05 keimlink joined #salt
09:18 Rumbles hello salt people, I was hoping to build a dictionary for use in a config file for a python script used in a nagios check, something like this: https://paste.fedoraproject.org/419722/
09:21 babilen Sounds as if you want to use pillars
09:21 babilen (which are dictionaries)
09:21 Rumbles yep
09:21 Rumbles I was getting to that but I'm struggling to copy it in the right format :)
09:21 remyd1 ravenx, I am using mongodb
09:21 Rumbles trailing spaces being a pain and making it unreadable
09:21 Rumbles 1 sec
09:22 Rumbles I thought pillars were just yaml files babilen
09:22 remyd1 it is JSON document oriented, easily "parsable", light, nosql...
09:23 babilen Rumbles: The yaml is being parsed into Python dictionaries
09:25 Rumbles so this is the pillar I created: https://paste.fedoraproject.org/419728/28082941/
09:25 Rumbles the idea being I should have 3 db-nodes and each one should have the mongodb.log file there
09:25 Rumbles and 2 api-srv with both of those log files
09:25 Rumbles the second file there is just to demonstrate that I would have a number of files
09:26 babilen Taht won't compile (you want a list .. start with "-")
09:26 babilen (line 5, 8 and 9)
09:27 Rumbles ahhhh
09:27 Rumbles that would explain the error complaining about -
09:27 Rumbles thanks
09:27 babilen np
09:28 babilen Why do you use 3 and 2 in there?
09:28 babilen Or even db-node and api-srv?
09:29 Rumbles the script is to be run on my rsyslog server, I want to check for missing files
09:29 mage_ how can I reload a service (mountd) but only if it's running ? is there a better way than a - onlyif: pgrep -u root mountd ?
09:30 Rumbles I wrote this script babilen: https://paste.fedoraproject.org/419732/
09:33 Rumbles the script works if I get the dictionary in the right format
09:33 Rumbles I wrote this jinja template to take the pillar data and turn it in to the dictionary: https://paste.fedoraproject.org/419734/
09:34 Rumbles but I get: SaltRenderError: Jinja syntax error: expected token 'in', got '-'; line 10
09:34 Rumbles line 10 being: {% for host-type, count, filetypes in salt['pillar.get']('rsyslogmonitor', {}).iteritems() %}
09:34 Rumbles I tried removing the - in rsysog-monitor in my test because I was trying to rle that out as being the cause for the issue.... it didn't help :/
09:36 babilen Pillars are dictionaries .. you don't have to write code for that to happen
09:36 babilen I'd also make the "count" explicit
09:39 babilen https://www.refheap.com/122709 something along those lines. That way you can do "for file_check in salt['pillar.get']('rsyslog:file_check', []) .... "
09:39 babilen file_check.name
09:39 babilen (or file_check.get('name'))
09:40 babilen The - on line 3, 6 and 9 essentially ensure the following datastructure: {'rsyslog: {'file_check': [{'name': ...}, {'name': ...}, ...]}}
09:46 Rumbles thanks babilen I'll look that over
09:46 Rumbles people keep asking me questions though so it's hard to concentrate on one thing :)
09:48 babilen I would just find that a lot easier to read/use
09:48 Rumbles hmmm, yeah but I would have to re-write my script to make that structure work :/
09:48 Rumbles I don't mind about readability
09:48 Rumbles I love cryptic dictionaries :D
09:48 * Rumbles is evil
09:49 babilen I would strongly recommend not to be ;)
09:49 Rumbles hehe
09:49 pjs joined #salt
09:50 Rumbles thanks I'll try it out
09:50 Rumbles you're right, I should try to make it readable :)
09:52 Rumbles if I wanted multiple files for a certain type of host
09:52 Rumbles could I just have name: 'file1', 'file2'
09:52 Rumbles or would I have multiple names ?
09:53 tooth joined #salt
09:54 flowstate joined #salt
09:56 babilen Well, my recommendation would be a little different. I would suggest to ensure that the list of files that are being checked is being sent to each host in toto (in the pillar)
09:57 babilen You would then ensure that the host type is being send the correct list of files by suitable targeting or logic in your pillar
09:57 babilen (and you'd use a list of those files over which you iterate)
09:57 Rumbles I'm not checking each host
09:57 Rumbles I'm checking whether the files are recieved on the rsyslog server
09:58 Rumbles we've had an issue where logrotate broke rsyslog collecting mongodb files
09:58 Rumbles I fixed the issue, but I want a nagios check that makes sure those files are in place
09:58 Rumbles and that each host has a folder that is created, each day
09:59 babilen Right
09:59 haam3r joined #salt
09:59 babilen In that case I'd make "name" a list and use something like "file_names: - foo - bar - baz ..."
10:00 babilen (pick a better variable name if you like)
10:05 netcho joined #salt
10:08 keimlink joined #salt
10:09 DEger joined #salt
10:15 Rumbles thanks, I'm going to have to coe back to this later
10:15 Rumbles developers have mre priority work for me :D
10:21 richardl Hi everyone I have a quick question for you
10:22 richardl I am trying write a loop and have an issue with it
10:22 richardl the code is located on http://pastebin.com/XmeW6MFi
10:23 richardl I am trying have the variable that is in the for loop as part of the name inside a pillar call however I am getting a failure when I try run this
10:24 rem5 joined #salt
10:24 jhauser joined #salt
10:28 west575 joined #salt
10:43 Brijesh1 joined #salt
10:47 fracklen joined #salt
10:50 dariusjs joined #salt
10:50 keimlink joined #salt
10:54 haam3r joined #salt
10:55 keimlink joined #salt
10:55 flowstate joined #salt
10:56 flowstat_ joined #salt
10:59 netcho joined #salt
11:00 djinni`_ joined #salt
11:05 VSpike richardl: how is your pillar data structured?
11:07 VSpike richardl: it's more likely to be something like https://bpaste.net/show/3a342e6dcfbf
11:07 keimlink joined #salt
11:08 bluenemo joined #salt
11:08 VSpike richardl: better highlighting https://bpaste.net/show/f0b3292cffb6
11:09 fleaz joined #salt
11:09 fracklen joined #salt
11:09 tooth joined #salt
11:09 M-MadsRC joined #salt
11:09 LeProvokateur joined #salt
11:09 cmek_ joined #salt
11:09 pprkut joined #salt
11:09 akw joined #salt
11:09 Satyajit joined #salt
11:09 goal_ joined #salt
11:09 kevc joined #salt
11:09 dandelo joined #salt
11:09 Edur joined #salt
11:09 inire joined #salt
11:09 haam3r joined #salt
11:14 roock joined #salt
11:14 bastiandg joined #salt
11:15 yidhra joined #salt
11:20 haam3r joined #salt
11:20 flowstate joined #salt
11:30 amcorreia joined #salt
11:38 haam3r joined #salt
11:39 pcn joined #salt
12:03 oida joined #salt
12:06 richardl VSpike: Thanks I ran that code and it worked
12:07 VSpike Excellent!
12:15 liskl joined #salt
12:24 barmaley joined #salt
12:25 rem5 joined #salt
12:26 akhter joined #salt
12:30 edrocks joined #salt
12:33 SubOracle joined #salt
12:37 numkem joined #salt
12:39 RodrigoEduardo__ joined #salt
12:42 gh34 joined #salt
12:42 dendazen joined #salt
12:42 idontbyt_ joined #salt
12:43 akhter joined #salt
12:44 dariusjs joined #salt
12:46 akhter_1 joined #salt
12:46 raspado joined #salt
12:49 SirMikkalot joined #salt
12:57 OliverMT uhm, how do I do a - watch on a whole folder of .conf files for a service?
12:59 eichiro joined #salt
13:00 JohnnyRun joined #salt
13:04 keimlink joined #salt
13:06 idontbyte joined #salt
13:07 mapu joined #salt
13:08 Cadmus OliverMT: Are they all put in by one statement, or does each one have its own? I've had success putting "listen_in" on config files instead of using watch.
13:09 OliverMT I just want a catchall for the whole folder
13:09 OliverMT I think watch on a service restarts the service if a .conf changes
13:09 akhter joined #salt
13:09 OliverMT which is the behaviour I want
13:11 ttrumm joined #salt
13:12 akhter joined #salt
13:13 flowstate joined #salt
13:15 keimlink joined #salt
13:23 akhter joined #salt
13:29 dyasny joined #salt
13:29 akhter joined #salt
13:30 tapoxi joined #salt
13:30 lazybear joined #salt
13:33 skrobul joined #salt
13:34 gh34 joined #salt
13:38 akhter joined #salt
13:42 akhter joined #salt
13:45 oida joined #salt
13:47 akhter_1 joined #salt
13:50 marie1972 joined #salt
13:50 cmarzullo is each conf file managed by salt?
13:51 marie1972 left #salt
13:53 oida joined #salt
13:56 jerredbell joined #salt
13:58 fracklen joined #salt
13:59 Taters_ joined #salt
14:03 edrocks_ joined #salt
14:04 rem5 joined #salt
14:04 sjmh joined #salt
14:04 cyborg-one joined #salt
14:06 _JZ_ joined #salt
14:10 akhter joined #salt
14:11 DEger joined #salt
14:13 richardl Got a quick question for you all, does any one have an idea on how to make a salt-state wait for about 30 seconds before running? I am deploying a docker box and it loads up after about 30 seconds and the state makes calls to the database
14:13 richardl I would like to wait for the loading to be complete
14:15 m0ntez Hello, if someone has time to lend a hand. On 3 minions out of 80 service salt-minion stops after 5 min or after 30 min. error in /var/log/salt/minion is "[salt.scripts][ERROR   ][10616] argument of type 'int' is not iterable"and "[salt.scripts][WARNING ][10616] ** Restarting minion **". CPU is arm. OS is debian v7, salt-minion version is 2015.5.3. Any ideas please
14:19 sp0097 joined #salt
14:20 dariusjs joined #salt
14:22 hasues joined #salt
14:22 hasues left #salt
14:33 mpanetta joined #salt
14:35 _aeris_ joined #salt
14:35 _aeris_ hello #salt :)
14:36 _aeris_ i have some trouble with external pillar, unable to have a working env :'(
14:38 _aeris_ https://paste.imirhil.fr/?eb8baa5a9bc3cc8c#qYhnUhJJ31BtnPXsGNeNiOb7+s+AX9wvBr5xYLCu+V4=
14:38 _aeris_ am i wrong somewhere ? :(
14:38 sjorge joined #salt
14:39 rem5 joined #salt
14:40 akhter joined #salt
14:42 flowstate joined #salt
14:43 CampusD joined #salt
14:46 CampusD Hi All! I had a few questions about some the core grains. Are either machine_id or server_id unique identifiers of some sort? Are the randomly generated by the salt-minion? or they are part of the machine/vm ?
14:47 mapu joined #salt
14:48 mpanetta joined #salt
14:50 flowstate joined #salt
14:51 jav joined #salt
14:52 flowstate joined #salt
14:54 jxm_ joined #salt
14:54 kus joined #salt
14:58 win_salt _aeris_, is your modules folder names _modules? I think thats what salt expects
14:58 armonge joined #salt
14:59 fracklen joined #salt
15:01 _aeris_ win_salt > i find one of my error : modules are search in {{root_dir}}/modules, not directly on modules
15:01 _aeris_ but so, i have another question : i don’t find a way to have subfolder on modules
15:01 _aeris_ if i create modules/ssh/authorized_keys.py, no way to use it
15:02 _aeris_ i expect it will be ssh.authorized_keys, but it’s not :(
15:04 win_salt yeah its flat be default, but you can configure it https://docs.saltstack.com/en/latest/ref/configuration/master.html#module-dirs.  You can put modules/ssh in that list and it will find it i think
15:05 _aeris_ dot on __virtual__ seems not work too
15:05 _aeris_ https://paste.imirhil.fr/?d7bce82aa08c533d#C6dIRH43Fvh5rvR+G38HBK/4Kp1zivaTcTfHtbYqr94=
15:06 _aeris_ hum, not important here in fact, can directly set the pillar name in module \o/
15:07 jimklo joined #salt
15:07 jaybocc2 joined #salt
15:08 wangofett joined #salt
15:09 manji joined #salt
15:11 oida joined #salt
15:14 raspado joined #salt
15:17 ferbla joined #salt
15:19 cro joined #salt
15:20 BattleChicken joined #salt
15:21 haam3r joined #salt
15:22 Brijesh1 joined #salt
15:23 tercenya joined #salt
15:25 sp0097 joined #salt
15:28 mikecmpbll joined #salt
15:32 XenophonF OliverMT: how are you deploying the directory full of conf files?
15:33 XenophonF OliverMT: if you're using a file.recurse state, you can just watch that for changes
15:34 philhuk joined #salt
15:34 haam3r joined #salt
15:40 philhuk joined #salt
15:41 cscf file.recurse is really handy
15:42 flowstate joined #salt
15:42 Reverend yhou're telling me :P
15:42 akhter joined #salt
15:44 edrocks joined #salt
15:44 onlyanegg joined #salt
15:48 akhter joined #salt
15:49 Brijesh1 joined #salt
15:52 flowstate joined #salt
15:52 kojiro joined #salt
15:57 haam3r joined #salt
16:01 ageorgop joined #salt
16:03 sp0097 joined #salt
16:03 akhter joined #salt
16:04 CampusD so, ok I tried clonings on virtual box and vmware and the UUID seems to be the unique identifier when cloning vms
16:05 gh34 joined #salt
16:06 jaybocc2 joined #salt
16:07 ALLmightySPIFF joined #salt
16:09 marie1973 joined #salt
16:10 titilambert joined #salt
16:14 marie1972 joined #salt
16:15 marie1972 left #salt
16:16 kojiro joined #salt
16:18 DEger joined #salt
16:18 g3cko is there a bug in latest 2016.3.3 where a minion when matched on a grain may execute things twice? when I do a `alt -v -G apt:true pkg.list_upgrades` I get 1 minion responding twice (once with an apt locking error due to running two at once)
16:19 g3cko I've removed the minion's key and re-accepted it, its only in the list of keys once
16:19 g3cko if I do the same salt -v -G apt:true test.ping it shows up twice in there as well
16:21 DammitJim joined #salt
16:22 jaybocc2 joined #salt
16:22 dps joined #salt
16:23 dps I have a dump question.  It is possible to keep the top.sls for the base environment in git and have no top.sls on the filesystem for that environment at all?  I basically want to make git authorotative for base (i.e. no fs_roots for base at all, other than git).
16:24 Brijesh1 joined #salt
16:24 pipps joined #salt
16:32 writtenoff joined #salt
16:34 alwaysatthenoc joined #salt
16:40 akhter joined #salt
16:43 Brijesh1 joined #salt
16:45 subsignal joined #salt
16:51 netcho joined #salt
16:54 amcorreia joined #salt
16:56 jholtom joined #salt
16:58 XenophonF dps: that's how i do it
16:58 jgarr so is kubernetes working fine with docker 1.12.1 in 1153.0.0?
16:58 XenophonF dps: https://github.com/irtnog/salt-states
16:58 jgarr woops, wrong channel
16:59 XenophonF dps: i have setup my git repo to have branches separate from the master branch, as well
16:59 GreatSnoopy joined #salt
16:59 XenophonF dps: so top.sls is in master (a/k/a base) and then dev/test/prod states are in the corresponding branches
17:00 ageorgop joined #salt
17:00 edrocks joined #salt
17:02 sjmh so if you have a minion reporting to two hot masters - does it send event data to both when you do local minion side things, like salt-call or mine functions?
17:02 gtmanfred I believe in 2016.3, it only sends to one of them, and there isn't a way to know which one it is going to send to
17:03 gtmanfred but i believe that there was work to fix that... not sure
17:03 sjmh ok - but if one of the masters actually does the publish, it should only send back to that master, correct?
17:04 gtmanfred it will still report back through the salt commandline, but any events fired, may appear on the other master
17:04 sjmh hm
17:05 gtmanfred like, if you do an event.send state, it could go to the other master
17:05 sjmh ah, ok.
17:05 sjmh but if i was just doing something like cmd.script, that output would at least always come back to the master who published it
17:05 gtmanfred yes
17:05 gtmanfred it should just work ™
17:06 sjmh :)
17:06 sjmh we're getting to the point where we have some teams wanting to use some automation that's going to be putting heavier loads on the master - we are considering splitting our master into two masters, one for users to do tasks, and one for automation jobs
17:06 sjmh wasn't sure if it'd actually help or not
17:07 gtmanfred We have some big customers using it, and it definitely got more stable in 2016.3
17:07 akhter joined #salt
17:07 gtmanfred and i know they wanted that feature, it would be good to test in small first, cause i can't remember what was successfully fixed
17:08 gtmanfred cause it was almost 6 months ago
17:08 sjmh using 'it' - meaning the multiple hot masters?
17:08 gtmanfred yeah, there were improvements to that in 2016.3
17:08 sjmh and yeah, we're one of those 'bigger' customers :)
17:09 gtmanfred nice
17:10 gtmanfred the bigger one it was made nicer for is already using it and giving good reviews :)
17:10 sjmh I think our use case would mostly revolve around publishes from a master, just to run scripts/get results, so I don't think we'd be hit by the events going elsewhere.
17:10 gtmanfred yeah, if you have reactors on both setup the same, it should be all good
17:11 sjmh now the question is - is this actually going to help from a load perspective?
17:12 gtmanfred i would acutally suggest setting up a syndic
17:12 sjmh yeah, we went that route once upon a time.
17:12 sjmh it fell over
17:12 gtmanfred in 2015.8?
17:12 sjmh yeah
17:12 sjmh at around the ~16k node mark
17:12 gtmanfred you might consider trying again, cause there were significant improvements made in 2016.3
17:13 gtmanfred one customer is running 27k nodes
17:13 sjmh yeah, I think that one was us
17:13 sjmh that's us ( the 27k nodes )
17:13 gtmanfred ahh ok :)
17:13 gtmanfred nice
17:13 sjmh or at least, that's around what we are, so probably us? :P
17:13 sjmh We had tried to split our hosting zones via syndics for load purposes and we ran into some stampeding herd issues
17:14 sjmh I know you guys made some big improvements based off that, we just haven't had a chance to get to 2016.3
17:14 gtmanfred ahh yes, it was yall :P
17:14 sjmh we were waiting for 2016.3.3 to drop to fix the multiple restart issue on RHEL
17:14 gtmanfred love your products
17:14 sjmh lol, thanks.
17:15 ttrumm joined #salt
17:15 gtmanfred started using the budgeting product last week
17:16 sjmh My friends laugh when I tell them I use an accountant :/
17:16 gtmanfred My dad is a CPA, so i seek advice from him when i need it
17:17 sjmh so we'll revisit syndics when we get 2016.3.3 going I think then, although the single master simplicity is nice
17:19 gtmanfred +1
17:26 ageorgop joined #salt
17:29 btorch joined #salt
17:31 brotatochip joined #salt
17:32 onlyanegg joined #salt
17:33 dariusjs joined #salt
17:33 Brijesh1 joined #salt
17:39 jholtom to specify an environment to use on the commandline when running a state its saltenv=dev right?
17:44 sp0097 joined #salt
17:48 whitenoise joined #salt
17:50 lungaro joined #salt
17:50 pipps joined #salt
17:52 brotatochip joined #salt
17:52 gtmanfred yes
17:54 btorch left #salt
17:56 hasues joined #salt
17:56 hasues left #salt
17:56 jholtom I should not need to specify specific branches that I want to also be environments, correct?
17:59 tapoxi joined #salt
18:00 tapoxi there an easy way to eliminate all package repos aside from the ones I've specified in my sls?
18:00 tapoxi joined #salt
18:05 tapoxi joined #salt
18:06 brotatochip joined #salt
18:08 jholtom I just configured gitfs and have it ticking over happily for the base environment.  However, when I go to apply a state with saltenv=dev (dev is a branch) it does not find a top file
18:08 jholtom salt-run fileserver.envs backend=git howerver, does succesfully return all the branches that are allowed as environments
18:08 jholtom It does find a top file for the base environement
18:08 gtmanfred you do need to specify the other branches
18:09 gtmanfred iirc
18:09 gtmanfred actually
18:09 gtmanfred no you shouldn't, i think that was only for the old gitpillar setup
18:09 jholtom I specified my branch for ext_pillar which has no issues
18:09 jholtom as it doesn't have a dev branch as of now
18:09 gtmanfred yeah, you only have to do it for git_pillar, you shouldn't need to do it for the states :/
18:10 jholtom so for a one off run in a different environment, the appropriate way is with saltenv=dev (or another branches name)
18:11 jholtom I have the top_file_merging_strategry set to same, so that each of the branches will read its own top file
18:12 flowstate joined #salt
18:12 jholtom even if I create a fresh branch off of master (which works fine for calling states) and push it up to the git server and then speicify it with saltenv= it cannot find the file
18:16 philhuk joined #salt
18:17 brotatochip joined #salt
18:17 jholtom oh I really am retarded this time
18:18 jholtom never mind me guys
18:19 gtmanfred :)
18:24 armin joined #salt
18:30 mikecmpbll joined #salt
18:32 catpig joined #salt
18:33 jaybocc2 joined #salt
18:34 brotatochip joined #salt
18:38 debian112 joined #salt
18:39 pipps joined #salt
18:46 tapoxi smooth sailing on 2016.3.3 for everyone?
18:49 jaybocc2 joined #salt
18:53 flowstate joined #salt
18:56 mapu joined #salt
18:56 workthrick joined #salt
18:57 jimklo joined #salt
19:08 jimklo joined #salt
19:11 fredrick joined #salt
19:11 Taters_ joined #salt
19:11 zer0def erm, quick question: `Passed invalid arguments to state.highstate: unhashable type: 'list'` is known? getting this by calling `salt '*' state.highstate`
19:11 fredrick Anyone using hubblestat with 2016.3.3
19:12 _aeris_ left #salt
19:12 fredrick err hubblestack
19:15 Fiber^ joined #salt
19:15 fredrick Actually I see the module in the /srv/spm/salt/_modules/hubble.py but if I do a sync.modules it does not pick it up.
19:16 akhter joined #salt
19:21 jaybocc2 joined #salt
19:28 west575 joined #salt
19:33 toanju joined #salt
19:40 raeven joined #salt
19:42 raeven Hi saltminers, I am new to saltstack and i was wondering if there is a way to get the minions ip and add it to a configuration file for some servic that is running on the minion?
19:43 zer0def yes, through Salt Mine
19:43 raeven zer0def, Do you have any examples that is can use?
19:43 jnials I'm seeing a problem using a grain in an orchestration.  I have a snippet of the orchestration along with the debug output from salt-run.  I could really use some help. https://gist.github.com/anonymous/5587d703a4762508226c975f04e7c997
19:44 raeven Think I have found it. https://docs.saltstack.com/en/latest/topics/mine/
19:45 dendazen joined #salt
19:45 zer0def the tl;dr is: set a execution module to be called in mine_functions and call it wherever you need to
19:45 zer0def just can't remember whether mine_functions is set on the minion you want to pull the data from or into
19:47 Llmiseyhaa zer0def: From.
19:48 zer0def thank you, Llmiseyhaa
19:48 Llmiseyhaa Notta problem
19:50 west575 joined #salt
19:51 jnials This is the error message (truncated) that I see that.  [ERROR   ] Data passed to highstate outputter is not a valid highstate return:
19:53 flowstate joined #salt
19:56 ajw0100 joined #salt
19:58 zer0def as for the issue i was nearly about to report - i had two of the same requisite in a state
20:01 pipps joined #salt
20:04 ageorgop joined #salt
20:05 pipps joined #salt
20:08 notnotpeter joined #salt
20:08 ivanjaros joined #salt
20:08 ajw0100 joined #salt
20:13 pipps joined #salt
20:31 cmarzullo Does the salt-mine work across syndics?
20:34 fredrick jnials: stop_my_proc2: salt.function <--- missing :
20:36 sp0097 joined #salt
20:41 Bryson joined #salt
20:41 brotatochip joined #salt
20:43 voytek joined #salt
20:44 jnials OMFG.  I had 2 other people look at that and not one of us spotted. that.  I think I've been working too much this week.
20:44 nidr0x joined #salt
20:44 voytek i'm trying to setup salt+foreman... i've got salt running with one minion accepted, but anytime i try to run any salt command, i get 'load' returned in green, nothing else
20:44 edrocks joined #salt
20:45 jnials Thanks for spotting that fredrick.  I'm going to go fix my error, and start the weekend drinking a little early.
20:48 ajw0100 joined #salt
20:49 ageorgop joined #salt
20:51 dendazen joined #salt
20:52 flowstate joined #salt
20:55 sjorge joined #salt
20:57 fredrick jnails: enjoy the drinks.
20:58 flowstate joined #salt
21:06 lungaro joined #salt
21:14 debian112 joined #salt
21:18 lungaro joined #salt
21:20 lungaro joined #salt
21:23 brotatochip joined #salt
21:24 lungaro joined #salt
21:27 lungaro joined #salt
21:37 intr1nsic joined #salt
21:38 akhter joined #salt
21:38 ronp_usa joined #salt
21:43 iggy raeven: if you just need the local minion IP, you don't need mine for that... (i.e. you need a listen addr in a local conf file)
21:44 iggy mine is for if you want info from other minions on the minion (i.e. cluster IPs, openstack controller address on the compute nodes, etc)
21:44 DEger joined #salt
21:51 debian112 joined #salt
21:51 raeven iggy: What do i add in the conf file to get the ip?
21:51 flowstate joined #salt
21:53 iggy {{ grains.ipv4|first }}
21:53 iggy or something similar
21:54 raeven iggy: Oh, so easy huh
21:54 iggy well, mostly, yeah... there's a little more to it (like making sure your file.managed has template: jinja, etc)
22:00 ekristen joined #salt
22:02 DEger joined #salt
22:06 cebreidian joined #salt
22:10 flowstate joined #salt
22:14 DEger joined #salt
22:14 aphor https://trac.macports.org/ticket/52154
22:15 barmaley_ joined #salt
22:18 aalmenar joined #salt
22:18 aalmenar joined #salt
22:20 justanotheruser joined #salt
22:20 lero joined #salt
22:21 fredrick joined #salt
22:22 raspado joined #salt
22:22 armyriad joined #salt
22:22 DEger joined #salt
22:23 cyborg-one joined #salt
22:25 debian112 joined #salt
22:26 lero joined #salt
22:33 DEger joined #salt
22:34 lero joined #salt
22:41 ajw0100 joined #salt
22:53 debian112 joined #salt
23:02 DEger joined #salt
23:08 sp0097 joined #salt
23:10 marie1973 joined #salt
23:15 Reverend joined #salt
23:21 ageorgop joined #salt
23:29 jeddi joined #salt
23:40 DEger joined #salt
23:52 flowstate joined #salt
23:56 mattp_ can you parameterize state sls files?
23:56 mattp_ and/or pass arguments?
23:57 mattp_ or is it a necessity to make said thing a python state module if you need that functionality
23:57 iggy usually you do that with pillars
23:57 mattp_ for example, say you have a common REST request to update something, but you dont want to copy and paste this everywhere
23:58 mattp_ iggy: say youve got a sls like what ive described for POST'ing a log note, where a custom message is what you want to pass
23:58 mattp_ how would you do that with a pillar?
23:59 iggy well, states describe the system, so if you are posting log messages from an sls file, you're probably already doing something wrong
23:59 mattp_ hmm. true ;)
23:59 mattp_ iggy: its a blurry line though
23:59 iggy if I was trying to log things, I'd use an outputter

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