Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-07-13

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

All times shown according to UTC.

Time Nick Message
00:16 ndrei joined #salt
00:17 quasiben joined #salt
00:29 petyanca joined #salt
00:52 desposo joined #salt
00:52 mrbiggle_ joined #salt
00:54 desposo1 joined #salt
00:58 baweaver joined #salt
00:58 aw110f joined #salt
00:59 baweaver_ joined #salt
01:01 amcorreia_ joined #salt
01:02 mvensky joined #salt
01:21 baweaver joined #salt
01:24 beauby joined #salt
01:29 otter768 joined #salt
01:38 favadi joined #salt
01:44 beauby joined #salt
01:49 sgargan joined #salt
01:54 sgargan joined #salt
01:55 timoguin joined #salt
02:00 beauby joined #salt
02:09 pirogoeth wrote up a blog post about fixing the problems with salt on freebsd
02:10 beauby joined #salt
02:11 yomilk joined #salt
02:11 scoates joined #salt
02:12 pirogoeth https://blog.maio.me/problems-wi/
02:16 tmclaugh[work] joined #salt
02:21 malinoff joined #salt
02:26 saltstackbot joined #salt
02:30 beauby joined #salt
02:31 saltstackbot [reddit-saltstack] remove files that are no longer managed? https://www.reddit.com/r/saltstack/comments/3d1wq0/remove_files_that_are_no_longer_managed/ - 2015-07-12 - 22:22:04
02:31 jmreicha joined #salt
02:31 desposo joined #salt
02:32 quasiben joined #salt
02:34 evle joined #salt
02:37 quasiben joined #salt
02:38 desposo1 joined #salt
02:49 beauby joined #salt
02:52 favadi joined #salt
02:57 kusams joined #salt
03:04 beauby joined #salt
03:30 otter768 joined #salt
03:48 ITChap joined #salt
03:49 packeteer joined #salt
03:49 gladiatr joined #salt
04:06 ssbarnea joined #salt
04:09 ITChap Hi
04:10 gladiatr joined #salt
04:10 ITChap when using git as an external pillar source. Is it possible to define the ssh user ?
04:10 ITChap like for gitfs
04:29 stoogenmeyer__ joined #salt
04:29 favadi joined #salt
04:42 timoguin joined #salt
04:48 iggy ITChap: I set all that via ssh config aliases
04:50 iggy ext_pillar:  - git: master git+ssh://git@salt-pillars-github.com/iggy/salt_pillars.git
04:52 iggy Host salt-pillars-github.com\n  Hostname github.com\n  IdentityFile /etc/salt/pki/master/salt-pillars-ssh-key\n  StrictHostKeyChecking no
04:52 iggy that goes in /etc/ssh/ssh_config
04:53 napsterX joined #salt
05:00 ITChap iggy: thanks
05:04 al joined #salt
05:04 kaiyou joined #salt
05:06 keimlink joined #salt
05:09 rdas joined #salt
05:11 vandemar joined #salt
05:20 campee_ joined #salt
05:21 ITChap Is there an equivalent to fileserver.update for the pillar ? I found saltutil.refresh_pillar but the output is not verbose enough
05:21 ITChap even with -l all
05:21 ITChap I cannot see the result of the connection too my git repo
05:21 ITChap I would like to check if it work
05:23 sgargan joined #salt
05:25 karlthane With Halite being deprecated, what is the prefered dashboard / ui now? Foreman? or something else?
05:26 aparsons joined #salt
05:27 ramteid joined #salt
05:31 otter768 joined #salt
05:33 mosen karlthane: I saw someone recommend another one recently...
05:33 mosen ahh it was SaltPad, but it does say early alpha/beta
05:37 ValF Hi, got a weird behavior with salt 2015.5.0+ds-1precise1 (same version on master and minion). Here is what happens
05:39 stoogenmeyer__ joined #salt
05:40 ValF 1 - [master]  : salt 'minion' cmd.run 'command'   # master sends the command to the target minion
05:40 ValF 2 - [minion] :  executes command under a 'jid'
05:40 campee joined #salt
05:41 ValF 3 - [minion] : Returns information for job
05:42 ValF 4 - [minion] : executes AN OTHER  job doing the same thing with previous jid as argument
05:43 ValF that second jid returns nothing  : {}
05:43 ValF 5 - [master] :  returns " Minion did not return. [No response]"
05:44 ValF after executing test.ping, output is finally normal and following jobs execute correctly
05:44 ValF Has anybody experienced that ?
05:46 jhauser joined #salt
05:47 ValF Note that the command output at step 3 is what I expect to see on the salt master, but doesn't actually get printed
05:50 campee_ joined #salt
05:54 catpigger joined #salt
05:58 ageorgop joined #salt
06:00 colttt joined #salt
06:04 Guest89 joined #salt
06:09 dopesong joined #salt
06:10 TyrfingMjolnir joined #salt
06:12 sgargan joined #salt
06:17 sirex joined #salt
06:22 ndrei joined #salt
06:24 sirex joined #salt
06:32 AndreasLutro joined #salt
06:33 soren joined #salt
06:36 loz-- joined #salt
06:46 julienlavergne joined #salt
06:53 khebbie joined #salt
06:57 baweaver_ joined #salt
07:07 Ztyx joined #salt
07:09 kawa2014 joined #salt
07:11 TyrfingMjolnir joined #salt
07:12 thalleralexander joined #salt
07:17 intel joined #salt
07:18 khebbie_ joined #salt
07:22 slav0nic joined #salt
07:29 is_null joined #salt
07:32 otter768 joined #salt
07:36 Sniper^^ joined #salt
07:41 linjan_ joined #salt
07:47 zer0def joined #salt
07:49 Aidin joined #salt
07:54 bfoxwell joined #salt
08:02 ndrei joined #salt
08:03 bones050 joined #salt
08:05 is_null hi all, would PRs removing usage of superlatives and marketing talk like "The Salt system is amazingly simple and easy to configure" out of the documentation (seen here http://docs.saltstack.com/en/latest/ref/configuration/master.html ) be interresting ?
08:05 is_null i'm not sure how this got there initially maybe there's a reason I'm missing :D
08:06 is_null i don't think that's good even from a marketing point of view
08:09 is_null still reading about the salt master "Turning on the Salt Master is easy -- just turn it on!" because I'm trying to find a definition of all the things the salt-master service does for our internal "onboarding" documentation and I didn't find it ( http://docs.saltstack.com/en/latest/topics/tutorials/walkthrough.html#setting-up-the-salt-master )
08:11 muep indeed, imo it would seem more convincing if it just said something like "The Salt system is designed to be simple and easy to configure"
08:11 is_null "simple and easy to configure" compared to what ?
08:12 is_null from a marketing / NLP point of view that's not even the kind of things you want in your landing page ^^
08:12 muep no idea really
08:12 is_null or in an article, or anywhere ...
08:13 ajw0100 joined #salt
08:13 is_null here's the definition of saltstack-master I came up with for our internal docs, is it any good ? how would you suggest to improve it ? thanks ! http://dpaste.com/198QAN1
08:14 impi joined #salt
08:15 julienlavergne hello
08:16 bfoxwell joined #salt
08:16 julienlavergne I try to test the salt cloud vsphere driver; it well connects to the esx but I cannot have any information any error
08:17 stephanbuys joined #salt
08:21 Xevian joined #salt
08:23 khebbie joined #salt
08:30 eseyman joined #salt
08:31 ingslovak joined #salt
08:37 s_kunk joined #salt
08:48 is_null hi all, could anyone clarify a confusion i have please ? I thought that `file_roots` was used /only/ by the `roots` fileserver backend, but it is mentioned that only SLS files in the `file_roots` configuration would be part of the salt state tree wheras i thought other fileserver_backend could add files to the salt state tree. Src: the highstate docs ( http://docs.saltstack.com/en/latest/ref/states/highstate
08:48 is_null .html )
08:50 keimlink joined #salt
08:54 lempa joined #salt
08:55 jim_ joined #salt
08:56 jim_ Hi, salt-run manage.status just hangs and salt-key -L shows no keys in any fields
08:56 jim_ salt services are running
08:56 jim_ should I just resend the keys
08:56 jim_ ?
08:57 is_null jim_: did you check logs ?
08:57 jim_ hm no
08:57 jim_ where are they located?
08:57 jim_ hey
08:58 jim_ actually, running salt-run manage.status --log-level=debug worked
08:58 jim_ as in it didn't hang and returned an answer
08:58 jim_ not really the one I wanted though tbf
08:59 baweaver joined #salt
08:59 joeto joined #salt
09:01 is_null maybe there has been a network connectivity issue
09:01 is_null ( in such case, always blame the network xD )
09:03 stephanbuys joined #salt
09:04 ahale joined #salt
09:08 Norrland verify with tcpdump
09:08 Norrland on both master and minion.
09:08 KermitTheFragger joined #salt
09:08 jim_ found the issue, hostname -i wasn't giving me the master's true IP address
09:10 VSpike joined #salt
09:14 joshin joined #salt
09:14 joshin joined #salt
09:14 aqua^c joined #salt
09:15 kbyrne joined #salt
09:17 is_null that would definitely not help indeed :D
09:17 ajw0100 joined #salt
09:20 jim_ hey do I curl or telnet the ports to see if minions can access them?
09:25 is_null i'm not sure so please anyone correct me if i'm wrong but i'd say if 0mq communication is fine then it should work
09:26 is_null i'm just trying here to confront what i think as a newb to experts lurking this channel ;)
09:26 jim_ seems it reverted to raet somehow
09:26 jim_ what's the best way to set the master and minions back to 0mq?
09:29 shiin joined #salt
09:32 shiin Is there a way to resolve special variables like __salt__ when editing python code like for execution modules?
09:33 otter768 joined #salt
09:33 shiin or better yet: can I use a less magical way?
09:34 jim_ ignore me, all working again : )
09:39 Aidin joined #salt
09:39 dopesong joined #salt
09:41 markm joined #salt
09:42 impi joined #salt
09:50 The_Loeki joined #salt
09:50 permalac joined #salt
09:56 ITChap hi
09:56 ITChap get some weird issues with the mine
09:56 morsik it's possible somehow to do  {{ salt['cmd.run']('hostname') }}  in pillar in a way it'll return hostname from minion point of view, not master?
09:56 ITChap I am trying to get the internal ip of my servers like using the doc as a base
09:57 ITChap https://bpaste.net/show/d35cefc53cf3
09:57 ITChap but doesn't seem to work
09:57 morsik I started writing salt module that needs this… and ugh… it looks like I can't do this :<
09:57 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 81: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
09:58 ajw0100 joined #salt
09:59 The_Loeki @ITchap: you can use something like {{ salt.dnsutil.A('my.example.com') }} in your pillar stuff, or {{ opts.id }} for minion id; beware that using it like you do now would invalidate pillars function of being a reliable master-generated store
10:00 ITChap The_Loeki: I think it was for morsik
10:00 The_Loeki ah sorry yes
10:01 morsik The_Loeki: actually cmd.run was just example, but i'm trying to write submodule for… dnsutil :P
10:01 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 119: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
10:01 morsik The_Loeki: but I need it to be executed on minion when compiling pillar, not on master
10:02 AndreasLutro haha
10:02 AndreasLutro stupid saltstackbot
10:02 The_Loeki youi should *NEVER* ask the minion to do *ANYTHING* as part of your pillar composition; like I said, it defeats the purpose of the pillar; you should just go with a module instead in that case.
10:03 The_Loeki (or grain, or data, of course)
10:03 ITChap AndreasLutro: What is happening to the bot
10:03 morsik ITChap: hmm… it was mine ellipsis character :D
10:03 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 47: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
10:03 morsik damn! again.
10:04 The_Loeki LOL
10:04 morsik I use elipsis all the time normally :<
10:04 AndreasLutro can't handle unicode? …
10:04 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 64: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
10:04 The_Loeki I'm actually here for a unicode issue as well lol
10:04 AndreasLutro
10:04 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 42: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
10:04 morsik The_Loeki: but how this module can work…
10:04 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 75: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
10:04 malinoff а если по-русски?
10:04 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xd0 in position 38: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
10:04 malinoff :(
10:04 ITChap lol
10:04 morsik The_Loeki: ok, full case:   i need to resolve host using gethostbyname.
10:04 The_Loeki @malinoff: Speaking Russian apparently trips over poor ol' saltstack bot :)
10:05 morsik The_Loeki: just do {{ something('redis') }}  and it'll be automatically resolved from the minion point of view
10:05 morsik cause I have minions in different subdomains
10:05 malinoff The_Loeki, definitely :)
10:05 malinoff > Willie is a simple, lightweight, open source, easy-to-use IRC Utility bot, written in Python. It's designed to be easy to use, run and extend.
10:06 malinoff yeah, yeah
10:06 morsik and this would automatically resolve IP for 'redis.any.domain' when minion is in '.any.domain'. and other IP for 'redis.other.domain' when minion is in '.other.domain'
10:06 morsik we have prod and itg in different domains, and this would speed up work with salt without thinking about including IPs inside pillars
10:07 morsik and dev, etc…
10:07 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 48: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
10:07 The_Loeki So you want to know how a certain minion resolved a certain address because it might be different than a master would? If so, why would you want to put that in a pillar? The 'proper' place should be either grains, mine, or data module
10:07 ndrei joined #salt
10:07 ajw0100 joined #salt
10:08 ITChap Anyone has any idea regarding my mine function issue ? Or some additional doc
10:08 ITChap I start to lose hope
10:09 The_Loeki @ITchap; are you positive the data actually gets pushed to the mine? I suspect (but not for sure) that the docs state that all mine function conf stanza's get accumulated, but that's not true; *one (config/pillar)* will win. Check your running config on your minion to see if it is actually (trying to) publish the command to the mine
10:11 The_Loeki also, try salt-call -l debug mine.update see what happens
10:11 The_Loeki you should definitely see the minion attempting to publish it's functions then
10:13 morsik The_Loeki: so… how can I put it inside grains automatically?
10:13 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 49: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
10:13 morsik HELL NO, someone should kick this bot -.-
10:14 The_Loeki LOL
10:14 zer0def joined #salt
10:14 The_Loeki @morsik; you could do a few things actually:
10:16 ITChap The_Loeki: the command timeout
10:16 illern joined #salt
10:16 Norrland so…
10:16 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 40: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
10:16 Norrland heh.
10:17 ITChap https://bpaste.net/show/eae765e5f92f
10:17 morsik The_Loeki: ?
10:20 ITChap on the minion it looks like things are happening https://bpaste.net/show/30346be2f13e
10:20 The_Loeki If you just want resolution to work, put shortname in your pillar and configure 'search' in your resolv.conf or whatever; that way, 'redis' in 'search sub1.example.com' would try to resolve 'redis.sub1.example.com' which should work across the board. Another idea would be to force the config to use the minion's fqdn, e.g. 'my-server = redis.{{ grains.domain }}'  Yet another idea would be to include this in the hosts file using the
10:21 The_Loeki @ITchap; your mine.update times out??? The debug output should tell you why then
10:22 The_Loeki @morsik; sorry about the linecrap; the state is called host.present; it would circumvent regular name resolution (not recommended) and put the ip (semi-)statically in the hosts file
10:23 The_Loeki @ITChap; you're timing out on the bus; are you positive this minion is actively connected to a/it's master?
10:23 The_Loeki brb
10:25 ITChap The_Loeki: yes the minion is connected and I run it with -l all. This is what I get on the minion side https://bpaste.net/show/6cc5ea2059c9
10:27 ITChap Is this https://bpaste.net/show/90a2f6d1f6f2 related ?
10:30 jim_ hey, salt should let work with something like salt 'minion[1-5].*' right?
10:30 jim_ *just not let
10:30 ITChap That's funny when I run my mine.get I see it in the output of the minion without any error
10:32 ITChap Is there a way to list what is in the mine ?
10:33 aidin joined #salt
10:35 ITChap Have some progress https://bpaste.net/show/b257d15232b3
10:37 jim_ think i may have found a bug
10:37 jim_ minion[1-8].* finds targets but minion[9-12] doesn't
10:38 jim_ I don't think salt does double digits well
10:38 Mate jim_: never seen a variant of regex/wildcard expansian that does
10:38 Mate o
10:39 jim_ ok, so I have saltstack running on 14 Raspberry Pis
10:39 jim_ they're named pi0 through to pi13
10:40 Mate but [] matches single characters
10:40 jim_ running salt 'pi[10-13].*' cmd.run returns results from pi1 and pi3, not pi10, 11, 12 or 13
10:40 Mate and - specifies char code ranges,
10:41 Mate [10-13] matches a single character '1', '0' to '1' or '3'
10:41 jim_ so I need to type something like ['10'-'13']?
10:41 Mate pattern matching doesn't know anything about numbers
10:42 jim_ bollocks
10:46 jim_ figured it out now, thanks Mate
10:54 The_Loeki @ITchap hi; the last output is a-OK, right? :)
10:55 ITChap The_Loeki: yeah but it still doesn't work when using my function alias
10:55 ITChap The_Loeki: and I cant confirm that it works when the function is defined in the pillar instead of CLI
10:56 The_Loeki @ITchap; that's OK, at least it is publishing, which means that the minion has picked up the config either way. The alias part I don't know; we don't currently use it. It might be a YAML config issue or simply a bug
10:57 pfallenop joined #salt
10:59 ITChap The_Loeki: so basically it would be smarter for me to use mine function that send for example the whole network config of all the interface an later filter with Jinja in my sls/files ?
11:03 ITChap The_Loeki: ok it seems to work now when defining functions in the pillar. But I need to use module names not aliases
11:04 fsteinel joined #salt
11:04 The_Loeki Well, you could play around with the YAML config for a while to see if you can get it to work; most notably the interface: net0 part seems suspicious
11:04 The_Loeki you could try -arg: - net0 or something like that
11:05 The_Loeki -args, -arg not exactly sure
11:06 ITChap The_Loeki: I will give it a try
11:06 ITChap The_Loeki: anyway thank you very much for your help
11:06 pcdummy Someone may help me to replace this line: https://github.com/pcdummy/saltstack-salt_patches-formula/blob/master/salt_patches/init.sls#L25 ?
11:07 giantlock joined #salt
11:13 pfallenop joined #salt
11:16 ravdeep joined #salt
11:18 kushniro joined #salt
11:20 kushniro hello, i'm new to salt and have been trying to use salt-cluster to set up a cluster on AWS. I've managed to create VMs run salt bootstrap on them etc. but my minion file isn't being copied over. Is there some configuration or manual step needed for this or is this supposed to happen as part of the salt-cluster setup?
11:20 matthew-1arlette joined #salt
11:22 AndreasLutro kushniro: salt-cluster is a thing?
11:22 denys joined #salt
11:23 kushniro sorry :)
11:23 kushniro salt-cloud
11:24 AndreasLutro kushniro: I don't think there's any functionality for copying over the minion config file - there is a minion dict in the cloud config file you can use http://docs.saltstack.com/en/latest/topics/cloud/config.html#minion-configuration
11:26 ndrei joined #salt
11:26 gnudon joined #salt
11:27 kushniro yeah, i've tried that but whatever I seem to try from the docs I still end up with a masterless minion configuration on all of the minions salt-cloud creates.
11:29 aqua^c joined #salt
11:34 otter768 joined #salt
11:35 evle joined #salt
11:39 ndrei joined #salt
11:40 illern joined #salt
11:41 _JZ_ joined #salt
11:46 _JZ__ joined #salt
11:49 illern joined #salt
11:50 baweaver joined #salt
12:01 stephanbuys1 joined #salt
12:02 stephanbuys joined #salt
12:06 Whissi joined #salt
12:07 Whissi left #salt
12:19 N-Mi joined #salt
12:19 N-Mi joined #salt
12:21 tmclaugh[work] joined #salt
12:22 soren joined #salt
12:24 XenophonF joined #salt
12:25 sk_0 joined #salt
12:25 SheetiS joined #salt
12:28 AndreasLutro anyone know where salt gets its ssl info from? getting a certificate error even though curl and wget work fine
12:28 kawa2014 joined #salt
12:29 yomilk joined #salt
12:30 amcorreia_ joined #salt
12:32 redLED joined #salt
12:33 XenophonF AndreasLutro: can you post the error?
12:33 XenophonF use gist or something, natch
12:34 AndreasLutro it's a file.managed state, the error is  Unable to manage file: Error: The certificate was invalid. Error returned was: CertificateError("hostname 'files.phpmyadmin.net' doesn't match either of ...
12:34 redLED hi! do i understand correctly that as of latest version, salt-ssh has the ability to set grains, but not to target specific hosts by grains?
12:35 XenophonF iirc it eventually uses openssl, so perhaps the CA list isn't set up properly
12:35 XenophonF i've seen this on windows minions
12:35 AndreasLutro it's a debian machine
12:35 AndreasLutro and like I said, both curl and wget work, so it must be something specific to salt/python
12:36 XenophonF a copy of the state and the complete error message might help
12:37 Sniper^^ joined #salt
12:38 markm joined #salt
12:40 XenophonF maybe try writing a python script that does the download - urllib, i think. could be your python is broken
12:42 DammitJim joined #salt
12:47 XenophonF left #salt
12:50 XenophonF joined #salt
12:52 gnudon joined #salt
12:52 XenophonF joined #salt
12:57 subsignal joined #salt
13:01 AndreasLutro XenophonF: https://bpaste.net/show/55997d970e5c
13:02 numkem joined #salt
13:05 DammitJim joined #salt
13:06 dyasny joined #salt
13:07 PI-Lloyd when using reactors, is there a way I can pull info from the event path.. for example pull hostname from 'syndic/<hostname>/event/<minion_id>'
13:08 PI-Lloyd i want to trigger a reactor that only sends a command back to the originating syndic, not all of them
13:09 __monty__ joined #salt
13:12 AndreasLutro turns out it's an upstream/wontfix bug fixed in 2015.8 https://github.com/saltstack/salt/issues/23103
13:12 saltstackbot [#23103]title: Unable to manage file: 'handle' in 2015.2.0 | I want to download the couchdb sources and extract them, but always fail with: ...
13:12 AndreasLutro time to upgrade!
13:14 elfixit joined #salt
13:15 ekristen joined #salt
13:16 Tecnico1931 joined #salt
13:19 aidin joined #salt
13:20 AndreasLutro or not, that broke more things than it fixed
13:20 pviktori joined #salt
13:21 thalleralexander hi i want to add a line (or block) of text before a certain line. in ansible i had something like lineinfile with the insertbefore option but when i checked the file.blockreplace module i coudnt find something like that. should i just use a template and manage the file or is there a better way?
13:22 primechuck joined #salt
13:22 speriyasamy joined #salt
13:23 markm joined #salt
13:24 kawa2014 joined #salt
13:25 furrowedbrow joined #salt
13:27 eliasp is there a proper Salt-ish way to use the results of an execution module's function for "onlyif" statements? e.g. I'd like to replace "onlyif: test -e {{ configfile }}" with "onlyif: file.file_exists {{ configfile }}"
13:27 eliasp using Jinja to determine this at the render-time of the SLS is no option, as the conditions to be checked /might/ change during runtime…
13:27 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 171: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
13:29 hasues joined #salt
13:29 eliasp I could create separate 'module.run' states to depend on, but this makes all this unecessarily complex
13:29 hasues left #salt
13:31 FRANK_T joined #salt
13:31 speriyasamy joined #salt
13:31 timoguin joined #salt
13:31 peters-tx joined #salt
13:31 AndreasLutro eliasp: I forget if module function calls are done on the master or the minion, but if they're done on the minion you can just wrap your state in an {% if salt.file.file_exists(configfile) %}
13:31 bluenemo joined #salt
13:33 eliasp AndreasLutro: that's exactly the problem… by wrapping it in {% … %} it will be determined at render-time, not execution-time…
13:33 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 76: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
13:34 bhosmer joined #salt
13:34 AndreasLutro ah yeah nevermind
13:35 otter768 joined #salt
13:35 ljt joined #salt
13:39 napsterX joined #salt
13:41 markm joined #salt
13:42 impi joined #salt
13:44 jdesilet joined #salt
13:45 cpowell joined #salt
13:49 aparsons joined #salt
13:49 SheetiS eliasp: Is it acceptable to have a failure in your return if the file does not exist?  If so I can outline a way with a file exists state and a requires line.
13:49 perfectsine joined #salt
13:50 eliasp SheetiS: no, there should be no failure… just a different "path" of applied states…
13:50 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 75: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
13:51 eliasp SheetiS: I'm aware, that I could build such a requirement using other "module.run" states and building appropriate requisites, but that's way more than what should be necessary…
13:51 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 212: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
13:51 eliasp so it looks like I'll have to come up with a PR to do this…
13:51 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 94: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
13:51 eliasp here saltstackbot - take this: … … … … … … … :)
13:51 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 67: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
13:52 SheetiS well I was thinking of using the salt state of file.exists and then have your things that work if that exists require that state
13:52 SheetiS and then have an onfail for the file.exists that executes the not exist path
13:52 SheetiS but it would leave failure messages in the return even though it did what you wanted most likely
13:52 eliasp SheetiS: yes, something like this… but it't still way more "verbose" in terms of LoC than a simple "onlyif" line
13:52 saltstackbot UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 69: ordinal not in range(128) (file "/usr/local/lib/python2.7/dist-packages/willie/modules/chanlogs.py", line 135, in log_message)
13:53 babilen ...
13:54 babilen saltstackbot: You really shouldn't use ascii anymore ;)
13:54 wpot joined #salt
13:56 stoogenmeyer__ joined #salt
13:57 favadi joined #salt
13:57 andrew_v joined #salt
13:58 AndreasLutro does anyone here know if pkg.version_cmp works on debian? I can't seem to get it to work
13:58 icflournoy joined #salt
14:00 iggy saltstackbot: quit
14:00 weylin_ joined #salt
14:00 tmclaugh[work] joined #salt
14:01 sroegner joined #salt
14:04 impi joined #salt
14:11 speriyasamy joined #salt
14:11 tmclaugh[work] joined #salt
14:12 clintberry joined #salt
14:14 giantlock joined #salt
14:14 kusams joined #salt
14:18 bhosmer joined #salt
14:19 aqua^c joined #salt
14:21 subsignal joined #salt
14:21 joeto1 joined #salt
14:24 __alex_ hi. i have a set of minions which don't talk to the master. i see that they can authorize but test.ping doesn't work in either direction. release is 2014.1.7.
14:24 __alex_ any idea what else to debug?
14:24 tmclaugh[work] joined #salt
14:26 impi joined #salt
14:29 speriyasamy joined #salt
14:29 SheetiS __alex_: have you verified that you can communicat on tcp 4505 and 4506 to the master from those minions (via nc or some other tcp port test?)
14:29 DammitJim joined #salt
14:29 racooper joined #salt
14:30 __alex_ SheetiS: nc on both ports good
14:30 SheetiS hmm
14:30 SheetiS both master and minion running 2014.1.7?
14:30 __alex_ yep
14:32 __alex_ when i do "salt-call -l trace some.other.machine  test.ping" i notice that i echos "Loaded minion key" twice and hangs after the second one
14:32 __alex_ between the 2 echos there is " Decrypting the current master AES key"
14:33 __alex_ it's on ubuntu 12, old zeromq that needs an upgrade?
14:34 SheetiS what version is it? (pastebin a salt-call --versions-report for all the goodies)
14:36 gadget-girl joined #salt
14:36 gadget-girl can relocate where salt look for it
14:36 gadget-girl it's config files
14:37 DammitJim joined #salt
14:37 __alex_ SheetiS: http://pastebin.com/GBVtRFW2
14:37 __alex_ i just tried multiprocessing=False, didn't help
14:38 SheetiS ZMQ of 4.0.4 is newer than what a lot of my infra runs, so I think you're ok there.
14:39 __alex_ SheetiS: got other minions with zmq 3.2.2, they work
14:42 SheetiS In case it was a key issue for any reason, you could wipe the keys on the minion and try again (stop service on minion, use salt-key to remove key on master, rm -rf /etc/salt/pki/minion on minion, restart service on minion, accept key on master)
14:46 bhosmer joined #salt
14:47 __alex_ SheetiS: done, no luck
14:47 SheetiS key showed up on master to accept though?
14:49 SheetiS I'm not sure what to try next.
14:51 sunkist joined #salt
14:51 __alex_ SheetiS: yes, master accepts auth request, minion tells about it, too.
14:54 __alex_ i'll blame the intern and take a month of vacation ;)
14:54 SheetiS sounds like a plan.
14:59 programmerq joined #salt
15:00 iggy __alex_: you could also try upgrading to something a bit newer (than 18 months ago)
15:00 dewdrop joined #salt
15:01 Gareth morning morning
15:02 __alex_ iggy: i'd like to, but 50 minions are fine and 3 are not working. got scruples.
15:06 icflournoy joined #salt
15:06 jondonas joined #salt
15:08 sdm24 joined #salt
15:08 IanV0rn2341 joined #salt
15:09 TyrfingMjolnir joined #salt
15:10 stephanbuys joined #salt
15:24 jespada joined #salt
15:29 iggy __alex_: well, you'll have to do it at some point
15:29 baweaver joined #salt
15:31 ljt joined #salt
15:32 quasiben joined #salt
15:35 otter768 joined #salt
15:36 mephx joined #salt
15:36 muep joined #salt
15:36 unusedPhD joined #salt
15:36 Arendtsen joined #salt
15:36 Ahlee joined #salt
15:36 btorch joined #salt
15:36 amatas joined #salt
15:36 shadowsun joined #salt
15:36 terratoma joined #salt
15:36 smkelly joined #salt
15:36 runnyspot joined #salt
15:36 pacopablo joined #salt
15:36 runnyspot joined #salt
15:37 aqua^lsn joined #salt
15:37 sophomeric joined #salt
15:37 LtLefse joined #salt
15:37 CSa_ joined #salt
15:37 JoeHazzers joined #salt
15:37 cansis_ joined #salt
15:37 nodens joined #salt
15:37 superseb joined #salt
15:37 binocvlar joined #salt
15:37 lz-dylan joined #salt
15:37 is_null joined #salt
15:37 ]V[ joined #salt
15:37 nkuttler joined #salt
15:37 preachermanx joined #salt
15:37 hacfi joined #salt
15:37 mattiasr joined #salt
15:37 lazybear joined #salt
15:37 SaveTheRbtz joined #salt
15:37 erjohnso joined #salt
15:37 chitown joined #salt
15:37 zsoftich joined #salt
15:37 erjohnso joined #salt
15:37 ValF joined #salt
15:37 wintamute joined #salt
15:37 Gilou joined #salt
15:37 cwyse joined #salt
15:37 g3cko joined #salt
15:37 Sypher joined #salt
15:37 jcockhren joined #salt
15:37 agj joined #salt
15:37 johtso_ joined #salt
15:37 Norrland joined #salt
15:37 ska-fan joined #salt
15:37 analogbyte joined #salt
15:37 UForgotten joined #salt
15:37 coreping joined #salt
15:37 dstokes joined #salt
15:37 spootly joined #salt
15:37 moapa joined #salt
15:37 babilen joined #salt
15:37 mrtrosen joined #salt
15:37 toddnni joined #salt
15:37 Ssquidly joined #salt
15:37 eXistenZNL joined #salt
15:37 jalbretsen joined #salt
15:37 cb joined #salt
15:37 TaiSHi joined #salt
15:38 danemacmillan joined #salt
15:38 ujjain joined #salt
15:38 ujjain joined #salt
15:39 gadget-girl this is the error "No Top file or external nodes data matches found" where should i be looking
15:43 gadget-girl can i run master and minion on the same system
15:43 iggy yes
15:43 Bateau_ joined #salt
15:43 iggy I'd guess your top file is wrong
15:43 gadget-girl ok help
15:43 gadget-girl my salt ping failed
15:44 lexter joined #salt
15:44 gadget-girl gadgetgirl-HP-15-Notebook-PC:
15:44 gadget-girl 'ping' __virtual__ returned False
15:44 sdm24 what is the exact call you are using?
15:44 gadget-girl sudo salt '*' ping
15:45 sdm24 you want test.ping
15:45 sdm24 salt '*' test.ping
15:45 DammitJim joined #salt
15:45 gadget-girl so you think something wrong with my top.sls
15:45 CheKoLyN joined #salt
15:46 pguinardco Where is your top file located and how are you targeting minions?
15:46 grepory joined #salt
15:46 pguinardco Could also be connectivity if you can't get
15:46 sdm24 can you paste your top.sls to gist?
15:46 gadget-girl '*'
15:46 pguinardco test.ping working
15:46 linjan__ joined #salt
15:47 gadget-girl http://pastebin.com/HGNnMpRC
15:47 frankS2 joined #salt
15:48 gadget-girl and yes test.ping is working
15:48 CeBe joined #salt
15:49 mihait joined #salt
15:49 abele joined #salt
15:49 lkannan joined #salt
15:50 gadget-girl https://gist.github.com/sarasfox/b28bd0b1da1923da6580
15:50 sdm24 do you have a minion named 'saltminion*'?
15:50 sdm24 if you do not, that can ause that error
15:50 mattl joined #salt
15:50 iggy what's teh output from salt '*' test.ping?
15:50 basepi joined #salt
15:50 sdm24 because salt is trying to target something that doesn't exist
15:50 bhosmer joined #salt
15:50 ramishra joined #salt
15:51 gadget-girl what sudo be their
15:53 gladiatr joined #salt
15:53 kermit joined #salt
15:55 XenophonF joined #salt
15:55 SteamWells joined #salt
15:56 twodayslate joined #salt
15:57 sarerto joined #salt
15:57 goki joined #salt
15:59 gadget-girl https://gist.github.com/sarasfox/b28bd0b1da1923da6580
15:59 gadget-girl still not working add the error to the gist
16:00 mikepea joined #salt
16:02 murrdoc joined #salt
16:03 XenophonF gadget-girl: that init.sls file is in a subfolder named something like /usr/local/etc/salt/states/nginx, right?
16:03 quantumsummers joined #salt
16:04 gadget-girl yes
16:04 XenophonF gadget-girl: on your minion try running "salt-call cp.list_master"
16:05 IanV0rn2341 joined #salt
16:05 XenophonF that will show you what the master's telling the minion
16:06 XenophonF you should see top.sls and nginx/init.sls
16:06 gadget-girl yes
16:06 gadget-girl look at the gist
16:06 gadget-girl https://gist.github.com/sarasfox/b28bd0b1da1923da6580
16:07 whytewolf states/base/nginx would be states.base.nginx not nginx
16:07 subleq joined #salt
16:07 XenophonF there you go
16:07 aqua^c joined #salt
16:07 subleq Is there any way to populate a pillar value from a file?
16:08 whytewolf subleq: you could look into ext_pillars
16:08 XenophonF subleq: or the different renderer's ability to load data from files
16:08 zsoftich1 joined #salt
16:09 anotherZero joined #salt
16:09 XenophonF iirc there's specifically a file pillar or something that reads a given pathname into a pillar with the same name, so foo/bar/baz.txt matches a pillar called foo:bar:baz
16:10 XenophonF subleq: http://docs.saltstack.com/en/latest/ref/pillar/all/salt.pillar.file_tree.html#module-salt.pillar.file_tree
16:11 subleq that looks interesting
16:11 XenophonF oh users-formula, why you no set the right permissions on ssh keys :(
16:12 whytewolf gadget-girl: i see another error in your gist. - source: salt://home/gadget-girl/tiretask/salt/master/nginx/nginx.conf should be salt://nginx/nginx.conf
16:13 whytewolf gadget-girl: and top.sls should be one directory down. so that it isn't in states
16:14 gadget-girl top is in /base
16:14 TyrfingMjolnir joined #salt
16:14 whytewolf gadget-girl: not according to that gist. unless you moved it. it is showign as states/top.sls
16:15 gadget-girl states/ has base/ and top.sls
16:17 gadget-girl hum
16:17 gadget-girl that worked
16:18 whytewolf i forgot top.sls that is in the file_root dirctly doesn't show on a cp.list_master [although the ones that are not in file_root do]
16:22 murrdoc joined #salt
16:23 subleq Is SaltPad the canonical web frontend for salt?
16:24 tmclaugh[work] joined #salt
16:24 LtLefse no, that would be the UI that comes with Saltstack Enterprise :)
16:25 LtLefse saltpad is the most popular free one though
16:25 LtLefse as far as I can tell
16:26 spookah joined #salt
16:26 iggy true ^
16:28 amcorreia_ joined #salt
16:30 gadget-girl whytewolf i am lose how to point the at right nginx.conf
16:31 klocek joined #salt
16:32 DammitJim joined #salt
16:33 whytewolf gadget-girl: salt://home/gadget-girl/tiretask/salt/master/nginx/nginx.conf would point to [file_root]/home/gadget-girl/tiretask/salt/master/nginx/nginx.conf and looking at the output of the gist. you have a nginx/nginx.conf
16:33 theologian joined #salt
16:34 gadget-girl so it should be salt:/nginx/nginx.conf
16:34 whytewolf gadget-girl: yes
16:35 whytewolf well salt://nginx/nginx.conf
16:36 XenophonF there wouldn't be a formula for deploying saltpad, would there?
16:36 iggy I think there's an "example" state in the saltpad reop
16:36 iggy repo
16:39 gadget-girl ooh boy look forward all the fun written salt support for arangodb
16:42 murrdoc joined #salt
16:43 jmreicha Is there a way to get stdout output of what a cmd.run is actually doing?  I have a really long script I'm trying to run and don't know if it is hanging or doing stuff
16:43 ageorgop joined #salt
16:45 pguinardco Try running it in screen? (screen -d -m command)
16:45 pguinardco then screen -r to attach on the client
16:45 iggy salt isn't built to send data until it's completely done with what you told it to do
16:46 iggy so you'll have to do something else to see it's output (log to a file and check the file outside salt, screen as pguinardco said, or whatever else works best for you)
16:47 pguinardco Either that or append the output to another file with > or >>
16:47 jmreicha I see
16:47 LtLefse or command | tee /var/log/file
16:48 KyleG joined #salt
16:48 jmreicha Got it, thanks
16:48 KyleG joined #salt
16:48 jmreicha I am new to Salt and still trying to get it all figured out
16:50 XenophonF jmreicha: iirc there's a vt mode or something that might do what you want - check the salt man page
16:51 XenophonF jmreicha: depending on what your script does, there might be a way to express it via salt states
16:51 XenophonF like on windows, i don't need to run a bunch of powershell or dism commands to install windows features - i can use the win_servermanager state module
16:52 XenophonF (although under the covers, it calls out to powershell, so maybe that's not the best example)
16:52 msheiny joined #salt
16:52 jmreicha XenophonF: Yeah it's probably possible but I just haven't got that far yet :)
16:54 murrdoc joined #salt
16:56 jmreicha I have another (unrelated) question.  Is it best practice to clone formulas in to their own directory e.g. /srv/formulas or just put them in /srv/salt
16:56 murrdoc joined #salt
16:57 whytewolf jmreicha: personally if i use a formula i will setup a seperate section of my filesystem for them. to seperate it from my own work.
16:57 XenophonF salt-formula puts them into their own folder, e.g., /usr/local/etc/salt/formulas or /srv/formulas
16:58 XenophonF jmreicha: if you want to see a semi-complete and mostly working example config, take a gander at mine - https://github.com/irtnog/salt-states and https://github.com/irtnog/salt-pillar-example
16:58 XenophonF i use salt-formula to configure the master
16:58 * iggy barfs at salt-formulas "formula support"
16:59 XenophonF it is a little ugly
16:59 iggy horrendously ugly and completely against what the salt docs say (whether those should be revised or not is another story)
16:59 iggy hooray for spm?
17:00 perfectsine_ joined #salt
17:00 XenophonF the problem i have with just listing the formulas in my master config is that i'm not sure how to deal with the different environments i've set up
17:00 XenophonF i dunno
17:00 XenophonF for now running the salt.formulas SLS at the end of a bunch of git pulls isn't a big deal
17:01 XenophonF (i've forked all the formulas i use)
17:01 XenophonF in any case, jmreicha, you can see my example salt master config in https://github.com/irtnog/salt-pillar-example/blob/master/salt/example/com/init.sls, which uses salt-formula
17:02 jmreicha XenophonF: Thanks
17:04 jmreicha XenophonF: how many servers do you manage, if you don't mind me asking?
17:05 frosty996 joined #salt
17:05 XenophonF worldwide, less than 100, but everything's so spread out, and my team is so small/diverse, that change management is a really big deal
17:05 impi joined #salt
17:06 jmreicha I see
17:06 frosty996 What’s the easiest way to specify availabilty zone for instances?
17:06 ajw0100 joined #salt
17:06 jmreicha Thanks for the examples, they are really helpful
17:06 XenophonF we have 500-1000 direct users worldwide, for whom we provide email/file/print/etc.
17:08 XenophonF i pointed you at my home setup - don't have approval to share our actual configs with the world yet, although ultimately my goal is to release at least a few of our states as reusable formulas on github
17:08 XenophonF e.g., shibboleth, comanage, active directory, sharepoint
17:08 XenophonF brb
17:08 XenophonF left #salt
17:11 stanchan joined #salt
17:12 Guest89 joined #salt
17:13 frosty996 huh, the docs on docs.saltstack.com indicate: To allow for a more extensible configuration, --providers-config, which defaults to /etc/salt/cloud.providers, was added to the cli parser. It allows for the providers' configuration to be added on a per-file basis.
17:14 frosty996 but my 2015.5.2 indicates no such option: --providers-config
17:15 pirogoeth left #salt
17:16 conan_the_destro joined #salt
17:17 frosty996 oh I’m looking at Develop. :-/
17:19 Bryson joined #salt
17:20 bhosmer joined #salt
17:20 stoogenmeyer__ joined #salt
17:20 frosty996 maybe that came out in 2015.5.3 and that version hasn’t landed in the apt repos yet.
17:20 murrdoc joined #salt
17:24 iggy develop is what will be 2016.2 (or thereabouts)
17:24 iggy 2015.8 has already been forked and I doubt there will be another release this year
17:27 soren joined #salt
17:27 twork OT but relevant: i'm having trouble getting vim to use the vim-ansible-yaml plugin, and google hasn't helped. i've folowed (i think) the instructions, but every time i run vim, it complains: "Not an editor command: Bundle '.vim/bundles/vim-ansible-yaml'"
17:28 wryfi joined #salt
17:28 wryfi can someone explain: what is the macine_id grain?
17:29 wryfi i have several virtual machines that i cloned from a single master, and they all have the same machine id. is that to be expected?
17:29 wryfi is that a problem for salt?
17:29 perfectsine joined #salt
17:29 iggy I don't see any machine_id grain (only id)
17:30 ajw0100 joined #salt
17:30 wryfi interesting. it's there on my ubuntu vm's, but not on centos.
17:31 murrdoc joined #salt
17:31 wryfi chaumes@clamps-admin:~$ salt '*' grains.get machine_id
17:31 wryfi clamps-ihat.sfo.plos.org:
17:31 wryfi 0bc77ef2c27249e3cdab86c4559ee0b7
17:31 wryfi scruffy-centos01:
17:31 wryfi clamps-slanger.sfo.plos.org:
17:31 wryfi 0bc77ef2c27249e3cdab86c4559ee0b7
17:31 wryfi clamps-atlas.sfo.plos.org:
17:31 wryfi 0bc77ef2c27249e3cdab86c4559ee0b7
17:31 wryfi clamps-tahi.sfo.plos.org:
17:31 iggy if it's the same as id, then it's bad to have duplicates
17:31 wryfi 0bc77ef2c27249e3cdab86c4559ee0b7
17:32 wryfi the 'id' is equal to the hostname, and those are all unique
17:32 iggy try to paste that much output to gist next time
17:33 sdm24 hmm, my ubuntu minions seem to have a machine_id, and it too is the same
17:33 theologian joined #salt
17:33 iggy ditto (all ubuntu images in DO probably all cloned from the same image)
17:34 zmalone joined #salt
17:34 keimlink joined #salt
17:35 wryfi interesting, thanks for the data points iggy and sdm24
17:35 andrew_v_ joined #salt
17:36 otter768 joined #salt
17:36 aron_kexp joined #salt
17:37 iggy fdde6ec2231ff9e1ace717e4534ff7c5 - Ubuntu-14.04
17:37 iggy 51b84e16174dc44b46f1a68e4b99f34f - Ubuntu-15.04
17:37 ajw0100 joined #salt
17:38 wryfi mine are all ubuntu 14.04.2 - 0bc77ef2c27249e3cdab86c4559ee0b7
17:38 druonysus joined #salt
17:38 * wryfi is increasingly curious where these things come from
17:40 twork re. my bit a few lines up, about vim plugin not working, any hints on where to look for help? i know it's not exactly a salt thing but hoped somebody here might point me in a good direction.
17:41 iggy /var/lib/dbus/machine-id or /etc/machine-id
17:41 wryfi ah-ha!
17:41 redLED left #salt
17:42 iggy so it looks like a systemd thing
17:42 UForgotten joined #salt
17:43 wryfi that would explain why it doesn't exist in centos 6
17:44 iggy and the Debian-7 machines I checked
17:45 wryfi http://www.freedesktop.org/software/systemd/man/machine-id.html
17:45 wryfi it sounds like it's intended to be unique for each system.
17:47 * wryfi is irritated that this man page doesn't exist anywhere on his system
17:49 frosty996 Any idea when 2015.5.3 might land in the PPA?
17:51 iggy when the volunteer that makes the packages gets some free time to work on it
17:51 wryfi and systemd-machine-id-setup doesn't exist in trusty :-/
17:52 gimpy973 joined #salt
17:52 aw110f joined #salt
17:54 denys joined #salt
17:54 * iggy smacks murrdoc
17:54 murrdoc SUP
17:54 murrdoc getting saltrained by Raine
17:54 clintber_ joined #salt
17:55 jm-sn joined #salt
17:56 aqua^c joined #salt
17:57 murrdoc i see it
17:57 jm-sn Hey guys - I'm trying to execute cmd.run on the salt master itself (not a minion) via the reactor netapi. Is this possible? It seems to require a minion target.
17:57 wendall911 joined #salt
17:57 murrdoc setup a minion on the salt amster
17:58 donmichelangelo joined #salt
17:58 twork n00b question, jm-sn: why is your master not a... yeah, waht mrrdoc said
17:58 jm-sn So I just need salt-minion running on master?
17:58 gillesMo joined #salt
18:00 XenophonF joined #salt
18:00 XenophonF back
18:01 iggy jm-sn: so many things become possible when your master is itself a minion
18:01 jm-sn Thanks all. I found this which will get me started down the golden path. http://stackoverflow.com/questions/23791968/saltstack-is-it-possible-to-apply-states-on-the-master-and-if-so-how
18:03 TOoSmOotH joined #salt
18:03 bhosmer joined #salt
18:04 mks1945 joined #salt
18:08 primechuck joined #salt
18:09 viq joined #salt
18:10 gimpy973 Trying out Salt on an HPC cluster (with Warewulf), my nodes boot from a single PXE image, what is a good way to have them get a unique key after they boot?  I can preseed keys on the master, but how to securely give them to nodes?
18:10 quinox left #salt
18:12 XenophonF oh i get it, key pairs managed via ssh_keys_pillar don't get the right permissions
18:12 XenophonF time to fork/patch/pull request
18:17 Fiber^ joined #salt
18:18 __monty__ Is anyone using salt for workstations? Kind of like boxen?
18:21 Guest89 joined #salt
18:23 iggy gimpy973: there are a number of examples out there for how to do automatic key acceptance (with various levels of security)... wouldn't really be any different (conceptually) for pxe images
18:24 iggy __monty__: I know there are some people in here using it with Windows workstations
18:25 Ztyx joined #salt
18:26 __monty__ iggy: So it's possible but there isn't a specific project targeted to it? Not particularly interested in windows but good to know that it's not unheard of.
18:31 iggy nah, as far as Salt is concerned, a computer is a computer
18:31 iggy whether that be a raspberry pi, a user's desktop, or a server with 512 CPUs and 2T of ram
18:32 __monty__ iggy: Sure, the same holds for puppet but people still thought boxen was a good idea. I was wondering if anyone did the same for salt or if salt is so good out of the box that that's not even necessary :)
18:33 iggy I have to admit to now knowing what boxen is, but no... there's nothing specific
18:33 iggy at least not that I know of
18:33 iggy maybe you've found your calling...
18:33 XenophonF __monty__: i'm not using salt for workstations, yet, but it's in my roadmap
18:34 amcorreia_ joined #salt
18:34 XenophonF somebody, maybe eliasp? was using salt to push out things like microsoft office
18:35 baweaver joined #salt
18:36 XenophonF __monty__: what are you trying to do with your workstations?
18:36 __monty__ Mmm, didn't even think of all the software that relies on installers. Been a pretty long time since I've seen one of those but I guess a *real* solution couldn't avoid them.
18:37 __monty__ XenophonF: Nothing special, I just like the idea of being able to totally fubar my system and then have it in a pristine state at the push of a button.
18:37 pguinardco Most likely a cmd.run to do it and some wmi queries to tell if it's installed. It's a lot heavier than a package manager but hey it's windows so you are expected to throw hardware at the problem right?
18:38 perfectsine joined #salt
18:38 soren joined #salt
18:39 khris joined #salt
18:40 ageorgop joined #salt
18:40 XenophonF __monty__: if you use salt's package manager, installing stuff that supports silent installs is pretty easy
18:41 murrdoc joined #salt
18:42 martoss joined #salt
18:42 martoss left #salt
18:44 bhosmer joined #salt
18:47 druonysus joined #salt
18:52 zleslie left #salt
18:53 __monty__ XenophonF: I'd expect no less but of course that's only part of the problem.
18:55 XenophonF __monty__: dunno what kind of workstations you have, but my approach with windows so far is to let GPO do as much as it can, then supplement with Salt
18:55 __monty__ I was originally planning to use plain puppet because boxen is osx only. But I think I'm convinced that salt is superior.
18:55 julienlavergne joined #salt
18:55 XenophonF Mac OS X is a little harder because of the even worse package management situation over there.
18:56 markm joined #salt
18:56 XenophonF i'm also not sure if salt supports making changes to plist files natively
18:57 ahammond I want to use salt to find out the package version of a single package. I'm using pkg.list_pkgs but that gives me the info for every package on the system. Is there a more targeted approach?
18:57 dopesong joined #salt
18:57 ahammond it's a centos system, so yum, in case that's relevant
18:58 rhodgin joined #salt
18:58 XenophonF at least there's support for brew and maybe pkgsrc
18:59 speriyasamy joined #salt
18:59 sdm24 ahammond: did you try pkg.latest_version?
19:00 __monty__ XenophonF: salt.states.pkg docs don't mention homebrew nor macports, is it elsewhere in the docs? Initially I'm using it on linux boxes anyway so it's not a dealbreaker.
19:00 ahammond sdm24 that tells me what's available, but not necessarily installed.
19:00 sdm24 ah gotcha
19:01 ahammond and... also doesn't appear to work? Strange.
19:01 tvinson how can i set a grain value with the output of a command without using jinja? i want to make sure i capture the value during the execution time of the module.
19:03 __monty__ ahammond: The docs here mention a command "pkg.version", http://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkg.html
19:03 SheetiS tvinson: anything in a state file (regardless of the renderer) would be pre-rendered prior to execution.  Your might need a custom execution module to do what you want.  Able to go into further detail regarding your needs at all?
19:04 murrdoc this cpan module cant possibly work
19:04 murrdoc it doesnt configure cpan
19:05 whytewolf ahammond: if you are just looking for the module.function yumpkg.version [pkg.version]
19:05 X67r joined #salt
19:05 whytewolf ahammond: http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.yumpkg.html#salt.modules.yumpkg.version
19:07 pullphinger joined #salt
19:07 bhosmer joined #salt
19:08 pullphinger joined #salt
19:10 quasiben joined #salt
19:10 Ztyx joined #salt
19:11 tvinson SheetiS: http://pastebin.com/anRzPgMf is what i'm working with. in the last part of the state where the grain gets set i'm pretty sure the jinja variable is getting set before anything else in the state is run, so i don't wind up with the grain set to the new version on the first run of the state.
19:11 XenophonF http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.pkg.html
19:11 XenophonF there's an entry for brew
19:11 twork (still) not a salt problem per se, but (still) can't track this down: all the docs i find for syntax highlighers refer to the vim comand 'Bundle', but my vim says that's not a vim command. anybody got a hint?
19:11 XenophonF and one for openbsdpkg, which i thought was based on pkgsrc
19:11 twork vim docs don't do me any good.
19:12 twork example: https://github.com/chase/vim-ansible-yaml
19:12 XenophonF holy crap salt supports solaris pkg and pkgin and pkgutil?
19:13 whytewolf twork: the Bundle command is part of the Vundle vim plugin.
19:13 ahammond whytewolf that's exactly what I was looking for. Thanks!
19:14 whytewolf twork: https://github.com/VundleVim/Vundle.vim#quick-start
19:15 soren joined #salt
19:16 twork whytewolf: just go there, thanks
19:16 twork *got
19:16 whytewolf also it isn't a command. but a config line item
19:18 whytewolf one of these days i should put my dotfiles out on github
19:19 RandyT joined #salt
19:20 twork others have... and i appreciate it muchly... but i've been cargo-culting cut-and-pasting for days...
19:21 SheetiS tvinson: The goal in what you have is just to install the package if you do not have the latest version and then clean up after yourself and set a grain for the version installed?
19:21 whytewolf twork: lol. don't worry about it. most of my dotfiles have by lifted from a buddy of mine anyway.
19:22 tvinson SheetiS: that's the idea. the package is maintained as a shell script full of a bunch of rpm's (which is why there's all the weirdness v. a simple pkg.installed)
19:23 XenophonF left #salt
19:23 ajw0100 joined #salt
19:24 SheetiS Were you using the grain just to track the install here, or was is it something you want to keep for later?
19:26 bhosmer joined #salt
19:28 Vynce joined #salt
19:28 tpaul joined #salt
19:28 tvinson SheetiS: it may be used for other states
19:29 Vynce anybody have an idea about why my salt-master might be consuming 50% of system memory while (afaik) idle?
19:29 twork whytewolf: along my way earlier today i happened across: http://yanpritzker.com/2012/01/20/the-cleanest-vimrc-youve-ever-seen/
19:31 twork ...which is nice 'n' all, but still doesn't do me much good at sorting out the "stuff that matters" from the "stuff he happened to like" from the examples i find
19:31 jonlangemak joined #salt
19:31 perfectsine joined #salt
19:31 twork i love vim. but who has time to spend a day learning its mojo.
19:32 SheetiS Well obviously you've found that all of it cannot be put together into one state if you want to do it in a single run.  (It can be done to work on 2 highstates).  If you wanted it all to work in one pass, you could do it with an orchestration.  I'd probably skips some of the intermediate grain stuff and only set the grain at the end if you needed.  I'd also probably use pkg.list_pkgs rather than 'rpm -qa | grep foo' as my test to see if a package is instal
19:35 whytewolf twork: it just kind of comes from use.
19:35 whytewolf twork: it isn't something to learn in a day. but to grow experence with through a lifetime
19:35 __monty__ SheetiS: Isn't pkg.version more appropriate to check if a single package is installed?
19:36 Sniper^ joined #salt
19:36 Vynce … and a full CPU?
19:37 twork whytewolf: i learned the gospel of vi/vim in about 1997 or so. but i got it to "good enough for me" and stopped. only last week or so did i hit the point where (significant whitespace) was no longer something i could muddle along without a crutch.
19:37 otter768 joined #salt
19:37 Vynce google('salt master memory leak') returns references to nothing that looks relevant
19:38 dopesong_ joined #salt
19:38 SheetiS __monty__: indeed, but it sounded like there could be multiple names for the package because of the way it was handled so it might have taken some regex from a list
19:39 SheetiS if it is always the pkg name kitten-open-123 for any non-upgraded package
19:39 SheetiS rather than kitten-open-[0-9] as the regex pattern that was passed to grep
19:40 __monty__ SheetiS: Valid point, although personally I wouldn't want to open that can of worms, managing a package that changes names that is...
19:41 RandyT joined #salt
19:43 SheetiS I'll give an example using pkg.version for tvinson here in a moment
19:43 quantumsummers left #salt
19:44 tomasq SheetiS: here was my terrible pass on it: http://pastebin.com/zZ9VAMrp
19:45 Vynce is it just a thing that i need to restart salt-master every couple weeks?  looking at my graphs, it looks like system load and memory and CPU all just grow constantly while salt-master is running.
19:47 SheetiS tvinson: I'd probably do something like this if i wanted it to become 'correct' in 2 highstates: https://bpaste.net/show/5e00a9f3e8fd
19:48 SheetiS otherwise i'd take the 'if' and the 'else' sections and put them into an orchestration and run them in order.
19:49 Vynce1 joined #salt
19:50 iggy murrdoc: if we don't hear anything back on that salt-formula PR by tomorrow morning do you mind reverting it?
19:51 tvinson thanks SheetiS tomasq
19:51 saltstackbot joined #salt
19:52 tvinson it even looks like tomasq's works in one pass
19:53 iggy someone ping me if that bot starts going crazy again
19:55 SheetiS tvinson: technically it's 2 passes, but he's calling the 2nd pass from the first via salt-call :).   You might run into a concurrency/race condition depending on which one completes first.
19:55 SheetiS so if you plan to use that, make sure to test it fairly thoroughly.
19:59 baweaver joined #salt
19:59 RandyT joined #salt
20:03 evilrob joined #salt
20:05 johngrasty joined #salt
20:07 Vynce joined #salt
20:07 RandyT greetings
20:07 jonlangemak joined #salt
20:07 RandyT anyone around to give a greenhorn some help mapping an ext_pillar to github repo?
20:08 iggy shoot
20:08 baweaver joined #salt
20:09 RandyT iggy: standby and let me throw some things in gist
20:09 RandyT https://gist.github.com/rterbush/c1212a8ceddc26e6727e
20:10 RandyT probly best help would be some hints as to how to debug...
20:10 RandyT private repo
20:10 iggy git+ssh://git@salt-pillars-github.com/iggy/salt_pillars.git
20:11 murrdoc joined #salt
20:11 ajw0100 joined #salt
20:13 pcdummy https://github.com/saltstack-formulas/salt-formula/pull/151 @ iggy :)
20:13 saltstackbot [#151]title: Install cherrypy/tornado from pip only if use_pip is True for salt-api. |
20:13 ndrei joined #salt
20:13 iggy bad
20:13 iggy all kinds of bad
20:15 perfectsine joined #salt
20:15 breakingmatter joined #salt
20:16 breakingmatter left #salt
20:16 zmalone joined #salt
20:17 twork hrm. well, i think i've now fumbled my way through getting vim-ansible-yaml installed. at least, when i ":Bundle 'chase/vim-asible-yaml'" i get an error about a name collision, so yay! but a salt file with "# vim:ft=ansible:" at the bottom shows up just as plaintext as anything else.
20:18 RandyT iggy: any hint as to how to debug this to see if it is visible from the master?
20:18 iggy RandyT: salt-run -l debug fileserver.update ?
20:18 Vynce1 joined #salt
20:21 SheetiS twork: tried https://github.com/saltstack/salt-vim?
20:21 SaveTheRbtz joined #salt
20:23 twork SheetiS: not yet. my quest continues...
20:23 twork thanks
20:23 SheetiS It's what I use locally, and I've been happy enough.
20:25 stoogenmeyer__ joined #salt
20:25 RandyT iggy: no mention of the ext_pillar and no errors, so assuming it is good. Off to figure out why a software dependency is not installing
20:25 RandyT thanks
20:26 murrdoc iggy:  murrdoc: if we don't hear anything back on that salt-formula PR by tomorrow morning do you mind reverting it?
20:26 murrdoc yeah i ll revert
20:26 jespada joined #salt
20:26 twork hm. the instrux for vundle say, in .vimrc, "filetype off" is required. so i did that, thinking the plugin might be doing that magic itself? but when i ":syntax on", i do get colors, if not quite accurate ones. fwiw.
20:26 murrdoc twerK!
20:26 twork haven't tried that
20:27 twork mixed company...
20:27 iggy RandyT: oh, I'm not sure root= works (it didn't in 2014.1 when we started rolling out this deployment... we ended up having to use separate repos)
20:28 SheetiS I set mine up a year ago, so I am not sure what I had to do in vundle to get it how I wanted.
20:28 RandyT iggy: interesting, will try that as well.
20:28 RandyT running dev branch here as dependent on much of the aws cloud stuff
20:29 iggy yeah, I never went back and tested the root= stuff for pillar
20:29 iggy tbh, if I was going to do it all over again, I wouldn't use gitfs or ext_pillar:git
20:30 iggy I'd do salt-api + webhooks + reactor + git.latest
20:30 RandyT iggy: tks, that is over my head at the moment so will do some research
20:31 iggy we have so many formulas/git repos that we notice the minutely fileserver.update running in monitoring
20:31 * SheetiS uses git.latest and a jenkins job that polls the repo because lazy.
20:31 RandyT tks folks
20:32 iggy someone should do a blog post on that kind of setup (vs gitfs+ext_pillar:git
20:32 RandyT yes, somebody should. :-)
20:32 pcdummy iggy: https://github.com/saltstack-formulas/salt-formula/pull/152
20:32 saltstackbot [#152]title: Fix stuff suggested by @iggy. |
20:33 pcdummy iggy: hopefully puneetk isn't merging now.
20:33 pcdummy so you can review first :)
20:34 DammitJim joined #salt
20:36 perfectsine joined #salt
20:38 giantlock joined #salt
20:39 bhosmer joined #salt
20:41 iggy pcdummy: there's still more stuff (in the first PR) that wasn't fixed
20:42 BretFisher joined #salt
20:44 pcdummy iggy: what?
20:44 Trevor joined #salt
20:44 Vynce1 left #salt
20:45 ajw0100 joined #salt
20:46 Guest19508 left #salt
20:46 conan_the_destro joined #salt
20:46 tkarjani joined #salt
20:47 baweaver joined #salt
20:47 iggy You only fixed a few of the things I mentioned in the very first PR (from yesterday)
20:49 twork SheetiS: vim-yaml (a) works, huzzah! and (b) looks pretty good so far. thanks.
20:49 SheetiS nice
20:50 twork also i've had quite the remedial education in making vim do tricks. thanks everyone. i'll forget it all in a matter of hours because documentation is for weenies, but it's been a good ride.
20:50 sunkist joined #salt
20:52 trevorj joined #salt
20:53 amcorreia_ joined #salt
20:55 ndrei joined #salt
20:59 Gareth Booo. Someone broke develop.
20:59 perfectsine joined #salt
21:02 jmreicha joined #salt
21:02 tkarjani I have a question about grains like but dynamic data representing the current state of a minion.
21:02 tkarjani Does anyone currently store this data, in addition to grains, on a server to be horizontally scaled and tracked over time and does salt stack have an outlook for providing a solution to handling this kind of data in the future?
21:03 tkarjani Like a minion side reactor that could be used to return data about a minion on connection or when a particular module is executed ...
21:03 iggy salt isn't a time-series database
21:03 tkarjani Well I didn't expect you to be the database.
21:04 tkarjani But you are the mechanism that changes the data.
21:04 iggy salt gives you the tools to do that kind of stuff (by connecting to other tools), but it doesn't (probably never will, and probably shouldn't) do that itself
21:05 iggy we have modules that send events to graphite and riemann for tracking
21:06 soren joined #salt
21:07 tkarjani So you have to poll the minion order acquire the data over time? I my point is that salt knows when particular modules execute and those modules are usually doing the manipulation.
21:07 tkarjani So from that perspective it seems cumbersome to use an external service to handle sending events with that kind of data back to a server.
21:07 tkarjani Or am I approaching the issue wrong?
21:09 tkarjani I guess an example would be software installation. I don't want to send the list of installed software every time the cache file changes. I want to send it from the pkg module executes a function that changes it.
21:09 mbrgm joined #salt
21:09 cberndt joined #salt
21:09 tkarjani Excuse me, bad typing. I want to send it when the pkg module executes a function that changes it. *****
21:09 mbrgm hey! could someone point out when I should publish facts via grains and when via the salt mine?
21:11 jaybocc2 joined #salt
21:12 tkarjani Mine is executed over an interval
21:12 ljt joined #salt
21:13 tkarjani So if that data changes after a minion starts the change be picked at the interval.
21:13 tkarjani But not necessarily when the data actually changes?
21:14 tkarjani This may have a negative affect on use experience for client-application to salt.
21:15 tkarjani You also need to be careful when working with hardware that is CPU sensitive.
21:17 tkarjani Some data is best polled.
21:19 mbrgm ok, so what's the use case for mine at all?
21:19 sunkist1 joined #salt
21:21 tkarjani "The Salt Mine is used to collect arbitrary data from minions and store it on the master. " Its implementation seems to be polling.
21:23 tkarjani Sometimes polling is a symptom of not solving a problem with an event style solution.
21:23 iggy the mine is queryable across all minions from a single minion
21:23 iggy whereas grains are only available on the minion currently processing
21:23 tkarjani Ew, security.
21:23 RandyT looking for more help in debugging my salt foo
21:23 RandyT https://gist.github.com/rterbush/7d0ee0dfd93a697bb5db
21:24 RandyT trying to apply a state against a remote install
21:24 RandyT running: salt proxy.00.west.lhn state.highstate
21:24 ljt joined #salt
21:24 Guest89 joined #salt
21:24 RandyT applied to a minion named as such.
21:24 RandyT debug on the minion in that gist.
21:25 RandyT seems the only issue potentially
21:25 iggy tkarjani: it's not enabled by default
21:26 tkarjani Ok, the mine isn't or the sharing?
21:26 clintberry joined #salt
21:26 iggy RandyT: those saltenv's match git branches... I'm pretty sure those errors are safe to ignore for the most part
21:26 iggy RandyT: except for the base one obviously
21:27 iggy well, that's also safe to ignore if this is for a formula (as the top.sls is going to be in another location)
21:28 zmalone joined #salt
21:28 RandyT have moved top.sls locally to master for now
21:31 tkharju joined #salt
21:31 RandyT iggy: are the sls files refreshed for the minions when the highstate runs, or when master restarts?
21:34 aqua^c joined #salt
21:34 a7p joined #salt
21:36 baweaver joined #salt
21:36 danielcb joined #salt
21:36 hax404 joined #salt
21:36 tkarjani joined #salt
21:38 iggy every highstate
21:38 otter768 joined #salt
21:38 RandyT I think I have gotten confused about pillar_roots vs file_roots.
21:38 RandyT I have top.sls in pillar_roots..
21:41 iggy you need a top file in both
21:44 zmalone1 joined #salt
21:46 lpmulligan joined #salt
21:50 RandyT iggy: lightbulb starting to brighten... tks
21:51 murrdoc joined #salt
21:51 jaybocc2 joined #salt
21:52 mrpull joined #salt
21:53 soren joined #salt
21:55 baweaver joined #salt
22:03 zer0def joined #salt
22:06 jaybocc2 joined #salt
22:08 Ryan_Lane it's nearly impossible to find documentation for modules now :(
22:08 gnudon joined #salt
22:09 Ryan_Lane I used to click "all modules" then search for the module I needed, but now there's that damn sidebar that's massive and I can't even find the list of state modules
22:09 Ryan_Lane especially when a section is expanded
22:09 Ryan_Lane you can't filter the sidebar at all either
22:10 Ryan_Lane ah. the top bar still has the link to all modules
22:14 ajw0100 joined #salt
22:16 jespada joined #salt
22:17 iggy Ryan_Lane: be sure to file bugs, jacobhammons seems to be pretty good about getting things fixed up
22:17 quasiben joined #salt
22:20 jespada_ joined #salt
22:23 gnudon joined #salt
22:23 zmalone left #salt
22:25 gnudon joined #salt
22:26 froztbyte yeah that sidebar thing is a bit tricky
22:26 froztbyte but I don't know what to suggest as an improvement
22:28 Ryan_Lane well, I don't really know what to file as a bug
22:28 Ryan_Lane it's just difficult to navigate using the sidebar
22:30 iggy seems like an okay subject to me
22:37 ndrei joined #salt
22:38 spookah joined #salt
22:38 timoguin joined #salt
22:38 hacfi joined #salt
22:41 quasiben joined #salt
22:42 juanito gday guys, im trying to find information about states concurrent executions.
22:42 frosty996 left #salt
22:43 juanito sometimes salt throws an error saying that an other job is already running
22:43 juanito eg : 'The function "state.sls" is running
22:44 BretFish_ joined #salt
22:44 juanito is that possible to run different states at the same time ? i can see there is a concurrent option to state.sls module
22:45 juanito i think i got my answer use the queue option
22:50 murrdoc joined #salt
22:51 baweaver joined #salt
22:53 iggy juanito: that error usually pops up because you've got a rogue minion process running on the host (i.e. you restarted once and the old process lingers)
22:53 iggy juanito: stop the minion on that host, double check to make sure nothing is still running, then start the minion
23:03 twork i'm back to having something fundamental upside down again. excerpt of my top.sls: https://gist.github.com/mjinks/f814999ccc8e38313b6d
23:04 iggy the second part isn't indented
23:04 iggy 'app... should be at the same indentation as '*':
23:04 twork what if i don't want it to get the stuff under '*' though
23:05 twork i was about to elucidate
23:05 iggy you can't exclude
23:05 twork here's my situation:
23:05 twork i want all my minons to get base:
23:05 iggy (there's a bug open about it, but I don't expect it'll ever be solved)
23:06 twork i want minions that aren't under 'app-stage*' to not ge that stuff
23:06 iggy so don't use *
23:06 iggy just use 'app-stage*'
23:07 iggy * isn't some magical required entry (you just see it in a ton of examples because people are lazy)
23:07 twork ok, yeah, but why is a minion that isn't under app-stage* getting that stuff?
23:07 twork that's my real problem
23:07 iggy getting what stuff?
23:07 twork tomcat-logs, logstash-forwarder
23:08 iggy because it's formatted wrong
23:08 ajw0100 joined #salt
23:08 iggy I commented
23:10 twork ok, we're closing in on the fundamental that i don't understand.
23:10 twork [ponder ponder]
23:10 stanchan_ joined #salt
23:11 Guest89 joined #salt
23:12 iggy also, you really don't need the match: pcre part
23:12 twork hm. before i put that in (iirc) i was having trouble.
23:12 mosen joined #salt
23:12 twork but whatevs, that's a minor issue for me right now
23:13 mosen hi salties
23:13 twork so... in my version, with 'app-stage.*' at the first column, what's the meaning of that, if not to specify a particular set of minions?
23:14 twork oh! "because of the bug, dumbass"?
23:16 ndrei joined #salt
23:16 iggy I'm honestly surprised it didn't throw some odd errors like that
23:16 twork well that does fix it. thanks, even if i remain a little confused.
23:17 iggy but if not it was reading 'app-stage.*' as a saltenv and anything below it as minion matches
23:18 twork ah, that makes sense.
23:21 iggy next time you might try playing around with the data in a yaml parser
23:21 markm joined #salt
23:21 twork ha ha, i've spent the past few days trying to get a yaml plugin to work in vim.
23:22 twork i thought i was... 85% there
23:22 whytewolf valid yalm does not always mean it will do what you are expecting
23:22 aqua^c joined #salt
23:22 twork (my biggest remaining complaint was/is that it colored my closing #} tags funny)
23:23 Xiao joined #salt
23:23 twork dammit, no, my sls is still wrong.
23:24 twork with that indent fixed, "evil" is still getting everything that i only want to go to a specific subset of minions.
23:25 twork ...but i thought it was correct for a bit. so i've probably goofed up my format wrong again.
23:25 twork any tips on a good yaml parser (that i haven't tried already)?
23:25 twork which i guess means "one that isn't a vim plugin"
23:27 coreping left #salt
23:27 lexter joined #salt
23:27 furrowedbrow joined #salt
23:30 twork oh balls. there's one in my browser cache.
23:30 iggy http://yaml-online-parser.appspot.com/
23:30 iggy but as whytewolf said, valid yaml != correct yaml
23:31 twork yeah. my yaml is fine.
23:33 kusams joined #salt
23:35 ^_^ joined #salt
23:36 twork current state: https://gist.github.com/mjinks/bc9fff61044a807afedf
23:36 twork (now with some of the "i don't think this matters" left in)
23:37 Guest89 joined #salt
23:39 otter768 joined #salt
23:39 jmreicha What are folks doing to bootstrap nodes in to their salt environments?
23:40 baweaver joined #salt
23:40 iggy twork: so what isn't working?
23:41 iggy jmreicha: we have a couple line script that gets shoved into our instances by GCE
23:42 jmreicha iggy: Do you do cloud it via cloud-config?
23:42 iggy it's... similar
23:42 jmreicha *do it via
23:42 iggy closer to AWS' original "boot script" functionality
23:43 jmreicha iggy: is it running that salt bootstrap shell script deal?
23:44 twork iggy: the states i have specified under 'app.*' are going out to all my minions, not just them
23:45 whytewolf twork: example of a minion id that it is showing up on
23:45 iggy jmreicha: yes
23:45 iggy I also use salt-cloud (which also runs salt-bootstrap via ssh)
23:46 jmreicha iggy: Nice, thanks
23:47 twork whytewolf: you mean the contents of /etc/salt/minion_id?
23:47 twork if so, one is: evil.prlss.int
23:47 whytewolf twork: i mean the id that salt knows the minions as. as that is what you are trying to target
23:47 whytewolf ahh ok
23:47 bfoxwell joined #salt
23:48 Ryan_Lane jmreicha: cloud-init
23:48 Ryan_Lane jmreicha: are you doing master/minion or masterless?
23:48 iggy i.e. the output of "salt -E 'app.*' test.ping"
23:48 iggy twork: ^
23:49 jmreicha Ryan_Lane: That's what I'm thinking, I'll probably want to do both
23:49 Ryan_Lane salt-cloud works, but I'm pretty sure it doesn't do autoscale groups, which isn't terribly useful
23:49 Ryan_Lane I don't see why anyone would use anything other than autoscale groups
23:49 aqua^c joined #salt
23:49 Ryan_Lane (even for singletons)
23:50 Ryan_Lane jmreicha: have you seen the boto_* state and execution modules?
23:50 jmreicha Ryan_Lane: no, I'm still very new
23:50 Ryan_Lane jmreicha: http://ryandlane.com/blog/2014/08/26/saltstack-masterless-bootstrapping/
23:50 jmreicha Ryan_Lane: Thanks
23:51 twork iggy: that comes back just as i want
23:51 Ryan_Lane they're really good. best tooling around imo. I did write most of them, so take that with a grain of salt ;)
23:51 zsoftich1 joined #salt
23:51 Ryan_Lane I'm not sure there's support for creating ec2 instances. :D
23:51 whytewolf twork: do you have more than 1 top file?
23:51 iggy twork: then you've got something wrong
23:51 Ryan_Lane there /is/ support for autoscale groups, though
23:51 twork iggy: surely
23:52 jmreicha Ryan_Lane: did you do a talk about this?
23:52 Ryan_Lane which you should use. they are superior to individual instances in basically every way
23:52 Ryan_Lane jmreicha: yeah. http://ryandlane.com/blog/2015/04/02/saltconf15-masterless-saltstack-at-scale-talk-and-slides/
23:52 twork i think i'm back to starting from a blank top.sls and pasting things in one hunk at a time
23:53 jmreicha ha, I have it queued up on my YouTube list
23:53 Ryan_Lane heh
23:54 iggy twork: salt '*' state.show_top
23:54 ajw0100_ joined #salt
23:54 Ryan_Lane jmreicha: using ASGs makes things a bit harder (like naming your nodes), but they're totally worth it
23:54 iggy twork: make sure you only have one top (including in every branch of every git repo
23:54 * iggy gotta run
23:55 jmreicha iggy: Thanks for the input
23:55 twork thanks iggy
23:55 Ryan_Lane jmreicha: if you have any questions about the boto states, let me know
23:55 jmreicha Ryan_Lane: I'm going to poke around :)
23:56 jaybocc2 joined #salt
23:57 twork for anyone still folowing my soap opera: 'salt.show_top' returns "'salt' __virtual__ returned False" for each of my minions.
23:59 twork but 'find . -name top.sls' returns just the one i want
23:59 twork (the one i've been pasting out of)

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