Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-04-13

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

All times shown according to UTC.

Time Nick Message
00:02 ITChap joined #salt
00:07 autonomy joined #salt
00:07 dude051 joined #salt
00:12 scbunn joined #salt
00:12 dude051 joined #salt
00:14 fusionx86 joined #salt
00:14 fusionx8_ joined #salt
00:17 glyf joined #salt
00:18 fusionx8_ joined #salt
00:22 fusionx86 joined #salt
00:24 amkread joined #salt
00:25 amkread Anyone change a minion ID for a Windows machine via a salt command before? Any help apprecaited.
00:26 amkread *appreciated
00:27 MatthewsFace joined #salt
00:30 pravka joined #salt
00:31 SeeDickCode joined #salt
00:34 semarj joined #salt
00:37 dude051 joined #salt
00:46 fusionx86 joined #salt
01:12 fusionx86 joined #salt
01:27 kaictl joined #salt
01:28 SheetiS1 joined #salt
01:29 fusionx86 joined #salt
01:31 kusams joined #salt
01:31 ckao joined #salt
01:32 kusams_ joined #salt
01:35 otter768 joined #salt
01:36 ITChap joined #salt
01:39 kusams_ joined #salt
01:49 tdjb_ joined #salt
01:59 jalaziz joined #salt
02:12 ITChap joined #salt
02:12 MatthewsFace joined #salt
02:22 armguy joined #salt
02:32 micko1 joined #salt
02:36 evle joined #salt
02:37 michelangelo joined #salt
02:43 kusams joined #salt
02:44 favadi joined #salt
02:47 dude051 joined #salt
03:07 autonomy joined #salt
03:16 tdjb_ joined #salt
03:28 Furao joined #salt
03:29 armguy joined #salt
03:33 lictor36 left #salt
03:52 MatthewsFace joined #salt
03:57 ITChap joined #salt
04:04 Furao joined #salt
04:05 dude051 joined #salt
04:10 lictor36 joined #salt
04:10 p66kumar joined #salt
04:10 ITChap joined #salt
04:16 lahwran joined #salt
04:19 bhosmer joined #salt
04:27 ITChap joined #salt
04:29 Furao joined #salt
04:35 ITChap joined #salt
04:48 lahwran joined #salt
04:48 keimlink joined #salt
04:54 echo joined #salt
05:08 ITChap joined #salt
05:09 baloo_ joined #salt
05:09 baloo_ hey guys
05:09 baloo_ can you tell me pls what's wrong with this syntax: https://gist.github.com/ciandro/5ab2e72a814e29d1dfe8
05:10 baloo_ \Rendering SLS 'base:jdk-post-config' failed: Jinja variable 'cacerts' is undefined
05:12 BretFisher joined #salt
05:16 BretFisher anyone ever use the deb_apache module in state files? http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.deb_apache.html
05:16 BretFisher I must be missing something to getting it working
05:17 eliasp baloo_: your default points to "non-prod-cacerts' instead of 'non-production'
05:18 baloo_ oh crap
05:18 baloo_ :)
05:18 baloo_ you the man
05:18 baloo_ but wait
05:18 baloo_ then the whole thing craps?
05:19 BretFisher hmm, maybe I'm not understanding how to call modules from state files :/
05:20 baloo_ because it's not really detecting the grain
05:20 baloo_ it's just assigning the default value
05:24 glyf joined #salt
05:38 iggy BretFisher: you generally don't... you use state modules in state files
05:39 BretFisher iggy: thanks, I think I'm figuring that out
05:39 catpigger joined #salt
05:44 loz-- joined #salt
05:55 otter768 joined #salt
05:56 aparsons joined #salt
05:57 aparsons_ joined #salt
06:01 rdas joined #salt
06:02 colttt joined #salt
06:03 huleboer joined #salt
06:12 ITChap joined #salt
06:14 stephanbuys joined #salt
06:16 Singularo joined #salt
06:18 AndreasLutro joined #salt
06:19 huleboer joined #salt
06:34 oxae joined #salt
06:38 JayFK joined #salt
06:40 __alex joined #salt
06:40 KermitTheFragger joined #salt
06:40 nene1 joined #salt
06:42 hojgaard joined #salt
06:42 stoogenmeyer joined #salt
06:45 I3olle joined #salt
06:45 favadi joined #salt
06:49 iwishiwerearobot joined #salt
06:50 dRiN joined #salt
06:55 Furao joined #salt
06:59 ml623 joined #salt
07:00 ggoZ joined #salt
07:01 ml623 Hello... Trying to create a custom outpuuter... just a function named output(ret) which returns pprint.pformat(ret)... when I try: salt 'log*' test.ping --out gu I get [ERROR   ] Invalid outputter gu specified, fall back to nested
07:02 ml623 Any idea whats I am missing? my file is /srv/salt/_outputters/gu.py
07:04 hebz0rl joined #salt
07:05 lb1a joined #salt
07:05 kawa2014 joined #salt
07:08 gibmachine joined #salt
07:09 slav0nic joined #salt
07:09 slav0nic joined #salt
07:09 markm joined #salt
07:15 Auroch joined #salt
07:15 eseyman joined #salt
07:21 spookah joined #salt
07:36 Grokzen joined #salt
07:40 gdm85 joined #salt
07:45 autonomy joined #salt
07:47 Romlok joined #salt
07:50 CeBe joined #salt
07:51 harkx joined #salt
07:55 BbT0n joined #salt
07:55 otter768 joined #salt
07:59 pf_moore joined #salt
08:01 TyrfingMjolnir joined #salt
08:17 Xevian joined #salt
08:18 Furao joined #salt
08:20 bhosmer joined #salt
08:21 cberndt joined #salt
08:28 ujjain joined #salt
08:29 bdf joined #salt
08:30 dopesong joined #salt
08:32 glyf joined #salt
08:43 favadi joined #salt
08:44 ktosiek joined #salt
08:45 crashmag joined #salt
08:48 _mel_ joined #salt
08:48 kusams joined #salt
08:51 N-Mi_ joined #salt
08:58 dopesong_ joined #salt
08:59 Romlok joined #salt
09:00 madduck joined #salt
09:03 quist joined #salt
09:15 catpig joined #salt
09:27 gdm85 is it possible to print a pillar variable via CLI on a masterless minion?
09:27 kawa2014 joined #salt
09:28 Furao salt-call pillar.get my:pillar:key
09:29 babilen You might have to set --pillar-root=PILLAR_ROOT if that differs from defaults
09:29 echo joined #salt
09:31 ekle joined #salt
09:33 autonomy joined #salt
09:34 dorvan joined #salt
09:35 derelm joined #salt
09:36 Twiglet is salt-minion meant to trigger a highstate as soon as it starts?
09:38 plindgren joined #salt
09:38 plindgren Hi guys, im having an issue when trying to set purge to true in user.absent
09:39 babilen Twiglet: No, would you like that?
09:39 Twiglet Nope, the issue is that it keeps doing it
09:39 plindgren http://pastebin.com/FEpXdEq5
09:40 plindgren can someone check that paste and maybe help me realize what im doing wrong here?
09:40 Twiglet I didn't think it  should do it by default but it does constly, at least in 2014.7.1
09:41 ekle @plindgren missing :
09:41 ekle in line 11
09:42 plindgren Doh!
09:42 plindgren thanks :)
09:42 Twiglet babilen: is there a config setting? Trying to track down why it's doing it without much luck
09:43 ekle @Twiglet: yes in the minion config
09:43 ekle startup_states: 'state.highstate'
09:43 babilen Twiglet: There is, yeah. http://docs.saltstack.com/en/latest/ref/states/startup.html
09:44 keimlink joined #salt
09:45 Twiglet hmm, see I don't have that enabled :/
09:45 Twiglet http://hastebin.com/xosoxevoso.avrasm
09:45 Twiglet (yes I know, it's insecure, this is my vagrant setup)
09:45 babilen Twiglet: it might also be triggered by a reactor
09:45 babilen Twiglet: Oh, you can set that in vagrant too.
09:46 Twiglet Yeah, I'm triggering it with vagrant which then comes back saying a highstate is already running
09:46 Twiglet which is making me think it's the minion process kicking it off
09:46 babilen run_highstate in http://docs.vagrantup.com/v2/provisioning/salt.html
09:47 Twiglet this is my vagrant/salt setup: https://github.com/ryanwalder/saltstack-lxc-vagrant
09:47 malinoff joined #salt
09:48 Twiglet the only difference being that I have changed the vagrantfile to "config.run_highstate = true"
09:49 Twiglet And this is the end of the vagrant up on a minion: http://hastebin.com/fumuqaweho.pas
09:50 Twiglet Which is leading me to think it's the minion treiggering the highstate
09:51 echo joined #salt
09:56 babilen don't necessarily see something obvious
09:56 otter768 joined #salt
09:57 Twiglet grepped my states/pillar and can't see any startup_states calls either
09:57 Twiglet Oh well, cheers for the help. I'll carry on digging
09:58 trinken joined #salt
10:00 mfournier joined #salt
10:01 denys joined #salt
10:06 yggdrasil joined #salt
10:11 plindgren is there a way to only trigger running one specified state to a minion?
10:11 plindgren something like salt '*' state.highstate onlyrunthis.state
10:12 malinoff plindgren, state.sls
10:12 echo joined #salt
10:12 plindgren what?
10:13 Romlok plindgren: salt \* state.sls onlyrunthis.state
10:13 plindgren ah
10:13 plindgren thanks
10:14 Romlok apparently there will be state.apply in the next version
10:14 Romlok which makes a lot more sense than "highstate" or "state.sls"
10:14 Romlok http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.state.html
10:14 plindgren anybody seen this before? http://pastebin.com/gGggDaYU
10:15 plindgren everywhere tailspintoys_dev is state is from the same pillar item and for some reason, the last one gets sent in with an escape sign before the _
10:16 jeddi joined #salt
10:24 refnode joined #salt
10:25 plindgren was a mysql privilege issue
10:25 plindgren needed WITH GRANT OPTION
10:25 plindgren on that user
10:25 colttt joined #salt
10:26 gdm85 Furao, babilen thanks!
10:31 _unicoletti joined #salt
10:34 echo joined #salt
10:43 lietu joined #salt
10:45 kusams joined #salt
10:48 aquassaut joined #salt
10:56 echo joined #salt
10:59 Grok joined #salt
11:01 PI-Lloyd joined #salt
11:02 bluenemo joined #salt
11:02 bluenemo joined #salt
11:02 Twiglet babilen: found the problem with the minion triggering a highstate on start, seems if you have a schedule in the pillar it'll trigger on start
11:03 Twiglet babilen: eg: http://hastebin.com/harenegize.sm
11:03 babilen Makes sense
11:03 bhosmer joined #salt
11:03 Twiglet kinda
11:04 Twiglet not exactly expected behaviour but yeah, kinda makes sense
11:04 babilen Well, you have to establish a baseline and "all schedules have been run" is something that sort of makes sense to ensure when you start
11:04 Twiglet Yeah
11:05 giantlock joined #salt
11:09 cberndt joined #salt
11:12 uwjesq joined #salt
11:13 uwjesq What does salt do? The sucky website only says it walks your dog and regrows your hair but nothing concrete.
11:15 uwjesq And everybody in the irc channel just idles. I go read the source to find out. Thanks for nothing.
11:15 uwjesq left #salt
11:17 echo joined #salt
11:24 AndreasLutro haha
11:24 AndreasLutro okay then
11:27 babilen wtf?
11:30 APLU joined #salt
11:30 malinoff joined #salt
11:34 elfixit joined #salt
11:34 ujjain joined #salt
11:39 echo joined #salt
11:39 viderbit joined #salt
11:39 denys joined #salt
11:48 evle joined #salt
11:48 ThomasJ O.o
11:49 amcorreia joined #salt
11:50 glyf joined #salt
11:52 APLU joined #salt
11:53 monkey66 joined #salt
11:53 monkey66 left #salt
11:56 bhosmer_ joined #salt
11:57 otter768 joined #salt
12:01 echo joined #salt
12:03 eseyman left #salt
12:06 MK_FG joined #salt
12:07 iMil joined #salt
12:09 bhosmer joined #salt
12:10 moos3 joined #salt
12:11 autonomy joined #salt
12:11 Tyrm joined #salt
12:16 markm joined #salt
12:19 TooLmaN joined #salt
12:22 echo joined #salt
12:24 glyf joined #salt
12:27 gdm85 is there a way to output only the value of a pillar variable? I am using --out json but what I's like to see is really only the value on stdout
12:28 Twiglet salt blah pillag.get key ?
12:28 Twiglet pillar.get*
12:28 gdm85 Twiglet: that gives local: and then on a new line the value
12:29 gdm85 right now I am using jq this way: salt-call --out json pillar.get my_var | jq -r .local
12:29 gdm85 I was wondering if I can skip the call to jq at all
12:29 timoguin joined #salt
12:29 dexter91x joined #salt
12:30 dexter91x Hello,
12:30 dexter91x How can i add custom function to an existing saltstack module ?
12:30 cmcmacken joined #salt
12:30 Twiglet There must be a way of doing it sensibly, however I do every things from states and use file.managed with pillar_contents for something like that
12:31 londo joined #salt
12:34 primechuck joined #salt
12:34 kusams joined #salt
12:35 jeddi joined #salt
12:44 dexter91x ?
12:44 echo joined #salt
12:49 dendazen joined #salt
12:53 dyasny joined #salt
12:59 chiui joined #salt
12:59 jdesilet joined #salt
13:00 JDiPierro joined #salt
13:01 timoguin dexter91x: you could put the whole module into _modules in your file roots and that version of the module will override the one that's built in
13:05 CeBe1 joined #salt
13:05 dyasny joined #salt
13:06 echo joined #salt
13:07 echo_ joined #salt
13:08 babilen Any news on the .4 release?
13:10 dexter91x timoguin: i don't wont to override a module , all i want is to add a new function to it
13:10 cpowell joined #salt
13:11 murrdoc joined #salt
13:11 timoguin dexter91x: fork the repo and send in a pull request
13:12 murrdoc fork pull ftw
13:12 _JZ_ joined #salt
13:13 dexter91x so i will need to follow hack the source section which inside the documentation ?
13:13 timoguin Yep
13:14 timoguin And you'll have to wait until that version is released to use that new function.
13:15 AndreasLutro dexter91x: just put it in a custom module
13:17 dexter91x AndreasLutro: to be more specific , i need to create a custom function related to service module so i can use service.custom_function, thats why i was asking about modifying module without overriding it
13:17 mpanetta joined #salt
13:18 chiui joined #salt
13:18 fusionx86 joined #salt
13:19 kusams joined #salt
13:24 CeBe1 joined #salt
13:25 andrew_v joined #salt
13:25 racooper joined #salt
13:25 pdayton joined #salt
13:26 Deevolution joined #salt
13:27 echo joined #salt
13:29 AndreasLutro dexter91x: I'd still recommend putting it in a custom module so it's explicitly clear that it's not a standard salt function
13:30 AndreasLutro (besides it sounds like it's not possible to add your own functions to an existing module anyway)
13:31 gdm85 what would be the idiomatic way to recursively copy a directory with source on same filesystem? (not salt://)
13:33 Tecnico1931 joined #salt
13:34 tdjb_ joined #salt
13:37 mpanetta_ joined #salt
13:39 primechuck joined #salt
13:39 iggy don't?
13:40 murrdoc file.recurse ?
13:40 AirOnSkin Am I right in assuming that states/httpd in top.sls would match both states/httpd.sls and states/httpd/init.sls ?
13:40 murrdoc yes
13:40 AirOnSkin murrdoc: Thank you
13:40 iggy but you could probaby use salt.modules.file.copy
13:40 murrdoc hey man if u have more easy questions
13:40 murrdoc lob em over
13:41 murrdoc :D
13:42 murrdoc iggy:  i have the same requirement as gdm85
13:42 gdm85 iggy: why not?
13:42 gdm85 murrdoc: I tried file.recurse
13:42 murrdoc have to source a file from the code placed by devs on a server, and push it to config dirs like for monit and ssh keys
13:42 murrdoc gonna have to write a state to do the source.hash stuff
13:42 murrdoc and then file.copy if needed
13:43 gdm85 murrdoc: file.recurse requires a salt:// URI
13:43 murrdoc i know
13:43 murrdoc i re read your question
13:44 murrdoc basically u ll have to write a state function to do it gdm85
13:44 gdm85 :|
13:44 kawa2014 joined #salt
13:45 murrdoc they are easy to write ,and it makes sense that salt doesnt support it
13:45 stoogenmeyer_ joined #salt
13:46 dopesong joined #salt
13:46 murrdoc https://github.com/saltstack/salt/blob/v2014.7.2/salt/modules/file.py#L1982
13:46 gdm85 murrdoc: I'll rather just use an rsync cmd.run
13:47 dopeson__ joined #salt
13:47 murrdoc cmd.run is evil
13:48 kaptk2 joined #salt
13:49 echo joined #salt
13:49 denys joined #salt
13:50 echo_ joined #salt
13:51 dude051 joined #salt
13:52 murrdoc iggy:  salt environments merge right ?
13:52 murrdoc like if there is a state in prod and in base
13:52 murrdoc the stuff in base is available to all
13:53 scoates joined #salt
13:53 Brew joined #salt
13:54 chiui joined #salt
13:55 quist joined #salt
13:56 gdm85 murrdoc: thinking that way, basically all scripts need to become salt templates sooner or later :P
13:58 otter768 joined #salt
13:58 kusams joined #salt
14:00 murrdoc gdm85:  as they should be
14:00 dexter91x joined #salt
14:00 riftman joined #salt
14:00 murrdoc only because with how environments and stuff work ( in *sh) the cmd.run stuff is flakey
14:00 murrdoc also say in the future you get a mandate from security that salt cant run as root
14:00 devnull_ joined #salt
14:01 murrdoc keeping it all in python lets you write the sudo support into your python easier
14:01 murrdoc its extra work no doubt, but the benefits are there
14:01 murrdoc way better error handling
14:01 murrdoc money power women
14:01 murrdoc you know
14:01 murrdoc all the good stuff
14:02 rhodgin joined #salt
14:02 I3olle joined #salt
14:02 gdm85 murrdoc: what if I don't want to convert my scripts to python?
14:03 gdm85 I mean..I want to use Salt for deployment automation, ideally it shouldn't be bound to a specific language :)
14:03 murrdoc cmd.script is safer obviously
14:03 murrdoc keep the scripts where ever that serves as a salt file root
14:03 murrdoc and then source the script and run it on the minion
14:04 murrdoc cmd.script gives you the same things i recommended
14:04 murrdoc logging, sudo, error handling
14:04 gdm85 I guess I can use cmd.script, yes
14:04 murrdoc personally i dont want to run anything on a minion from salt that doesnt give me those things
14:05 murrdoc logging, sudo-ability, error handling
14:05 murrdoc thats why i say cmd.run is evil
14:05 murrdoc (personal preference)
14:05 murrdoc basepi:  https://github.com/saltstack/salt/issues/12483#issuecomment-64181598 :|
14:06 debian112 joined #salt
14:06 devnull_ Hi #salt! I'm running into some difficulty setting up a custom returner for a MasterJobCache... Anyone got some bandwidth to help me troubleshoot?
14:11 ndrei joined #salt
14:11 devnull_ I have written a basic custom returner (https://gist.github.com/anonymous/c07a1fd81c8e0c1e146c) intended to simply let me know if its working.  I've put that returner in /srv/salt/_returners, and set master_job_cache: cust_mssql in /etc/salt/master.  However, when I run a salt command with this setup, I get this error: https://gist.github.com/anonymous/d6651f8abe25f05d7d0a.
14:11 iggy murrdoc: true... I want to say they merge all the way up
14:12 murrdoc i am not liking it iggy
14:12 iggy murrdoc: so if in your config you have base, prod, qa, dev... dev will have dev/qa/prod/base... but it could just be dev/base
14:12 murrdoc i was hoping dev wont have access to prod
14:13 JDiPierro joined #salt
14:16 iggy oh, so it's not merged up, it's merged alphabetically?
14:16 iggy or is that just top merging (vs actual env merging)
14:17 iggy I don't use environments like that (for these very reasons)
14:17 murrdoc The code seems to simply merge and overwrite, in alphabetical order. Which contradicts entirely the documentation.
14:17 murrdoc
14:17 iggy we have a master per env and it's on a branch
14:17 murrdoc one year now
14:17 murrdoc i have to work with two 'purists'
14:17 murrdoc expect much head desking
14:18 iggy I think I'd rather roll with no merging whatsoever than some complex set of merging rules that nobody understands and isn't documented
14:18 iggy maybe a merge_envs: False setting?
14:19 * iggy -> work
14:19 murrdoc o/
14:19 murrdoc iggy:  merge my shit
14:19 murrdoc i got more pulls to rewrite formulas
14:19 murrdoc :)
14:19 murrdoc late
14:19 murrdoc left #salt
14:20 StDiluted joined #salt
14:23 tdjb_ left #salt
14:25 gdm85 I'd also prefer that it fails, like when you have 2 conflicting ids
14:26 timoguin joined #salt
14:26 Chris_____ joined #salt
14:28 Workflow Hi guys, wondering if someone could help me with passing pillar data through on the command line: https://gist.github.com/anonymous/cbca6f6174e894d39831
14:29 armguy joined #salt
14:29 Workflow Data is still 2.4.3 ^
14:30 Workflow On 2014.07.0 (Helium)
14:32 _unicoletti joined #salt
14:33 devnull_ Workflow: Have you tried changing 'zabbix={"version": "2.4.4"}' to 'pillar={"zabbix": { " version" : "2.4.4" } } '   ...I believe you have to encapsulate pillar data in the 'pillar' named structure, rather than simply passing the dictionary contents.
14:34 devnull_ See section 3.2.6.4 here: http://docs.saltstack.com/en/latest/topics/tutorials/pillar.html
14:37 bmac2 joined #salt
14:37 Romlok is there some gotcha I'm missing with multi-line strings in YAML?
14:37 Romlok http://pastebin.com/Ts0tkcLt
14:38 Romlok it doesn't matter what I put in the string, it always complains about the second line
14:39 yuhl_work_ joined #salt
14:39 timoguin Romlok: I think you need to indent that block 4 more spaces
14:39 babilen No, it should be two
14:39 bmac2 word to the wise, if a salesman names Steven Travaglini from Datto, selling Backupify calls you, he is a total jerk.  Once he was told we didn't need his software, he emailed the owner of our company calling the guy who talked to him an idiot.  And lied that he had been called an F**king asshole by the tech
14:39 timoguin something like that. ;-)
14:40 Romlok if it has more than two spaces, the string appears to be ignored completely :/
14:40 bmac2 so beware of backupify and their sales people
14:40 babilen Ah, not sure about - content: ..., but the yaml parser in Python insists on two spaces (as in the paste) after a |
14:40 Workflow I get "ERROR executing 'state.sls': Pillar data must be formatted as a dictionary" when I try that devnull_
14:40 devnull_ Workflow: Can you gist the command you're typing for me?
14:41 Romlok in fact, it complains about the line after the first "content" line, even if that line isn't part of the string
14:41 Workflow salt 'vm-testying' state.sls zabbix.upgrade pillar={"zabbix": { " version" : "2.4.4" } }
14:41 ITChap joined #salt
14:41 Workflow Sorry, lemme gist it!
14:41 devnull_ Workflow: Wrap the data for pillar in single quotes.      ...pillar='{"zabbix": ...... }'
14:42 Workflow Ok, got a little further there! Thank you devnull_  :)
14:43 devnull_ You're welcome!
14:43 babilen Romlok: That is not actually related, but I'd recommend to keep your keys in the pillar and to references them with pillar_source
14:43 riftman joined #salt
14:43 Workflow I got it to work with what you suggested I believe - pillar='{"zabbix": { "version" : "2.4.4" } }'
14:43 Romlok babilen: yeah, there's a reason this state is called "test.sls" :)
14:43 babilen Romlok: Ah, you are testing the gpg renderer, aren't you?
14:44 Romlok yuperoonie
14:44 babilen I'd try adding another two spaces. The python yaml parser is quite peculiar
14:44 seblu joined #salt
14:44 devnull_ Anyone available to give me a hand with a KeyError on a custom Master Job Cache returner?
14:45 devnull_ Workflow: Glad to hear that worked.  Command line can be finicky at times.
14:47 dexter91x joined #salt
14:49 Romlok the file.managed state is apparently ignoring the "content" altogether
14:50 Romlok if I set the value to a plain one-line string, the file is still empty
14:50 rm_jorge joined #salt
14:51 p66kumar joined #salt
14:52 SheetiS joined #salt
14:52 hasues joined #salt
14:54 SheetiS joined #salt
14:55 iggy Romlok: it should be 4 spaces
14:55 colttt joined #salt
14:55 Romlok okay, so I guess that was working with four spaces
14:55 devnull_ Romlok: According to salt docs... your key should be 'contents', not 'content'.   http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.managed
14:55 harkx joined #salt
14:55 Romlok ahahahahahaah
14:56 Romlok I'm a dufus
14:56 Romlok well spotted devnull_
14:56 andrew_v joined #salt
14:56 devnull_ hahah... Only reason I spotted it was that I've made the same mistake countless times before.
14:57 babilen I'd argue that "contents" is a badly chosen name for that parameter
14:57 devnull_ I would agree with you babilen.
14:57 babilen It's "the file content" not "the file contents"
14:57 Romlok right now, so would I
14:57 dexter91x joined #salt
14:58 Romlok I'm not sure either one is incorrect
14:58 Romlok my instinct is to use "contents" in spoken language
14:58 devnull_ I believe, grammatically, they are technically both valid.
14:59 iggy content refers to the amount of something in something else, contents refers to the something
14:59 devnull_ iggy: Huh. Learned something new today.  And in light of that, the 'contents' key seems to make more sense.
14:59 iggy this drink has high alcohol content, the contents of this drink are vodka... and ice
15:00 clintberry joined #salt
15:00 Romlok I'm not sure that's always true iggy
15:00 faust joined #salt
15:00 Romlok "content of their character", for example
15:00 Romlok though that might just've been oration style
15:01 Romlok nevertheless, I now have a sekrit file with sekrit contents \o/
15:01 devnull_ romlok: Glad it worked :)
15:02 Romlok I'm not sure I like that the sekrit contents were printed out in the salt changelog's diff, though... :/
15:04 chiui joined #salt
15:05 Romlok hah
15:05 Romlok "diff: <show_diff=False>"
15:05 Romlok nice that the feature is already there, but the UI could do with some work :)
15:05 dopesong joined #salt
15:06 jalbretsen joined #salt
15:06 Romlok sweet, thanks all
15:06 Romlok always like to end the workday on a high
15:09 scbunn joined #salt
15:13 ndrei joined #salt
15:16 toastedpenguin joined #salt
15:18 favadi left #salt
15:21 ALLmightySPIFF joined #salt
15:23 gladiatr joined #salt
15:23 ALLmightySPIFF joined #salt
15:25 dexter91x joined #salt
15:26 raygunsix joined #salt
15:26 jonlangemak joined #salt
15:29 jonlangemak joined #salt
15:30 jonlangemak I recently installed satl on a set of new centos7 boxes and the minions appear to be having issues registering to the master.  On the mater, I get this error -> salt-master.service: Got notification message from PID 2969, but reception only permitted for PID 2957 when I check the systemd journal logs.  A little googling makes this seem related to systemd but Im not sure what the fix is.  Anyone seen this?
15:34 p66kumar joined #salt
15:36 peters-tx joined #salt
15:37 dexter91x joined #salt
15:37 PI-Lloyd joined #salt
15:39 PI-Lloyd joined #salt
15:42 riftman joined #salt
15:42 manfred jonlangemak:  that is expected
15:42 manfred it sends that messages a couple times
15:42 dexter91x joined #salt
15:42 manfred and should be fixed in the next release
15:43 keimlink joined #salt
15:43 manfred it was put in a place where a child process could send the notification for Type=notify
15:44 manfred https://github.com/saltstack/salt/issues/15721
15:44 manfred https://github.com/saltstack/salt/issues/20531
15:48 cheus joined #salt
15:50 StDiluted_ joined #salt
15:50 jonlangemak manfred: Thanks!  I assumed that was my issue with minions not registering
15:50 jonlangemak But I guess I need to do more debugging
15:50 manfred make sure that you open up the firewall on the master
15:51 manfred by default centos 7 comes with firewalld enabled iirc
15:51 manfred systemctl mask firewalld
15:51 manfred systemctl stop firewalld
15:51 manfred and you can then just use regular iptables
15:52 jonlangemak Oh my gosh.  I had totally disabled that when I built this box.  Guess I must have not disabled it and it picked it up on reboot
15:52 jonlangemak thanks!
15:53 SeeDickCode joined #salt
15:53 smcquay joined #salt
15:53 ndrei joined #salt
15:55 StDiluted_ joined #salt
15:56 synestine joined #salt
15:56 dopesong joined #salt
15:57 SheetiS joined #salt
15:58 Eureka_ joined #salt
15:58 manfred stupid firewalld
15:58 manfred needs to be masked
15:59 otter768 joined #salt
16:01 BretFisher joined #salt
16:02 autonomy joined #salt
16:04 iggy or... shouldn't salt be installing rules for itself?
16:04 iggy (disclaimer: I know nothing about firewalld)
16:05 manfred firewalld
16:05 manfred firewalld is just RedHats version of ufw from ubuntu
16:05 manfred the only thing it is good for is just a wrapper around iptables to implement zones
16:06 manfred on a server, it is useless
16:06 manfred https://fedoraproject.org/wiki/FirewallD#Runtime_zone_handling
16:06 ndrei joined #salt
16:07 desposo joined #salt
16:10 timoguin iggy: I know salt installs rules for UFW at least.
16:10 * iggy prefers tools that give him plenty of rope to hang himself with
16:11 bhosmer joined #salt
16:11 manfred there are some firewalld state modules that were in a pull request the other day.
16:11 iggy I don't need no stinking firewalls (on by default)
16:13 dexter91x joined #salt
16:13 mgw joined #salt
16:13 timoguin joined #salt
16:13 babilen exactly
16:14 quintinadam joined #salt
16:14 manfred I wonder if openstack + neutron private networks can be configured with a global nat ip… because then I could configure my whole setup without allowing anything to connect to the servers outside of the datacenter… which would be nice
16:15 MatthewsFace joined #salt
16:15 wendall911 joined #salt
16:15 iggy we only allow access to the master on ssh from 3 IPs (main office and 2 other locations)
16:16 iggy aside from that, everything goes through a LB
16:16 manfred i think I could do it, but would need a vyatta appliance node
16:16 enarciso joined #salt
16:17 aparsons joined #salt
16:18 glyf joined #salt
16:18 kusams joined #salt
16:24 coval3nce joined #salt
16:26 faust joined #salt
16:27 slav0nic joined #salt
16:27 slav0nic joined #salt
16:27 jeremyalons what other options are there if saltutil.refresh_pillar doesn't actaully refresh the pillar?
16:27 jeremyalons looks like the minion is just stuck.
16:28 jeremyalons bounced the minion
16:28 jeremyalons would prefer not to have to do that.
16:28 iggy is this on an upgrade?
16:30 iggy or wait... is this on 0.17.5.Ahlee?
16:30 iggy because if so... you get to keep all those pieces
16:31 jeremyalons This minion is 2014.7.1
16:31 jeremyalons but i've seen this same issue since 0.13
16:31 iggy I've never had to restart a minion to get pillar data updated
16:31 jeremyalons in healthy environments, I don't either :(
16:31 jeremyalons I can't isolate what leads to this
16:32 iggy make your environment more healthy?
16:32 jeremyalons oh ok
16:32 iggy ;)
16:32 jeremyalons good thinking, not sure why I didn't think of that
16:32 bhosmer joined #salt
16:32 babilen The solution to your problem is: Fix it!
16:32 jeremyalons i am still on 0.17.ahlee, just a smattering of test instances on differnet environment
16:32 iggy <-- Mr Helpful
16:32 JDiPierro joined #salt
16:32 babilen jeremyalons: Which version is your master running?
16:33 iggy but you said you've seen this on multiple versions
16:33 jeremyalons babilen: this specific environment was 2014.7.1
16:33 babilen Okay, what happened in between "working" and "not working" ?
16:33 jeremyalons 2014.7.1 master with a dozen minions, 6 2008, 6 linux
16:33 babilen And what are common attributes amond the minions that don't refresh?
16:34 babilen +speling
16:34 jeremyalons babilen: my buest guess is ext_pillar is the culprit, I saw the master reach out and query zookeeper
16:34 jeremyalons running pillar.items against the minion showed the correct ext_pillar value
16:34 iggy if you can find a reproduceable test case... A/B testing?
16:34 jeremyalons but salt -I ext_pillar:value would not return
16:34 jeremyalons I gotta reproduce it again since i bounced the minion to get t back up and working
16:35 babilen jeremyalons: Could you be more specific on "but salt -I ext_pillar:value would not return" ? You can't even call it like that
16:35 iggy can you put a default in file pillar and see if that shows up when (you think) the ext_pillar is acting up?
16:35 jeremyalons babilen: i target on my external pillars, I can look them up, but targeting would not return
16:35 jeremyalons iggy: of course, but as i said I see the master reach out
16:36 iggy doesn't mean it's getting back the data it wants/expects
16:36 babilen So a "salt -I 'something:foo' test.ping is what we are talking about? And "something" is coming from an external pillar. Can you target on non-external pillar values and which external pillar(s) are involved here?
16:36 quintinadam joined #salt
16:37 cheus joined #salt
16:37 iggy so you could have a temporary zookeeper issue... it says "try again later" and the ext_pillar is seeing that as "no pillar data"
16:37 jeremyalons iggy: correct, however if the value wasn't correct the master would log an exception
16:38 jeremyalons babilen: that's an excellent test case. I'll give that a go next time it occurs
16:38 jeremyalons i've started the test suite agianst hte test master again, i'll see if i can get it back into the state
16:39 babilen +1
16:39 jeremyalons has anybody seen a max pillar size? I have, how do you say, a large amount of pillar data
16:39 babilen Always easier to resurrect a dead horse than a living one
16:39 iggy there's a max zmq msg size
16:39 iggy don't know if that limits pillar data though
16:39 jngd joined #salt
16:40 jeremyalons hrm. good thinking iggy
16:41 jeremyalons the potential-pillar-size-issues is back on 0.17.ahlee, so i'll let that subject die until i can test against the 2014 instance
16:41 jeremyalons and i'll upgrade my test instance to 2014.7.2 since i see from /topic that's actually 'latest'
16:41 jeremyalons despite 2014.7.4 lurking somewhere
16:41 iggy .4 is tagged
16:41 iggy just not in most repos
16:42 jeremyalons word
16:42 iggy .3... who tf knows
16:42 jeremyalons i'll rebuild and throw it in mine
16:42 timoguin .3 got tossed because issues found in testing
16:44 theologian joined #salt
16:45 writtenoff joined #salt
16:46 aparsons joined #salt
16:47 babilen Wasn't .4 to scheduled for release today?
16:47 stanchan joined #salt
16:47 * iggy avoids taking shots at release dates
16:48 babilen Well, so should salt ... 2015.2 is, well, unfortunate ;)
16:48 dexter91x joined #salt
16:49 hal58th joined #salt
16:50 babilen The new versioning scheme is simply FUBAR. There is absolutely no semantic to it and the only thing it does ensure is that version numbers are monotonically increasing
16:50 spookah joined #salt
16:51 mpanetta babilen: And don't forget confusing and incorrect :P
16:51 timoguin version 2016-to-the-MOON.rc1
16:51 babilen Well, hence the "no semantic"
16:52 jeremyalons semver or bust
16:52 babilen I wasn't strictly thinking in terms of SemVer, but rather in "no meaning whatsoever".
16:52 mapu joined #salt
16:52 jamilner joined #salt
16:53 jamilner how come when you preseed a minion with a key and you run "service salt-minion restart" another set of keys are generated with the names of "minion.pem" and "minion.pub"?
16:54 iggy I'd rather see 2015.1, 2015.2, 2016.1, 2016.2, 2016.3 or something (if they really want to stick with year versions)
16:54 BtbN joined #salt
16:54 iggy jamilner: what were the preseed'ed file names called?
16:55 jamilner this causes the master to subsequentally deny the preseeded key believing the auto generated key to be the legit minion key to the master
16:55 jamilner iggy: the preseeded file names are "minionTest.pem" and "minionTest.pub"
16:55 jamilner iggy: do the filenames have to be under a specific format?
16:55 babilen iggy: Yeah, but those numbers should have *some* relation to *some* action that happened in that month
16:56 jeremyalons ...why not preseed them minion.pem and minion.pub?
16:56 jeremyalons shouldn't be required of course
16:57 jamilner well what happens if you have multiple servers and they all have minion.pem as the key?
16:57 jamilner then i'm guessing in that scenario authentication with master wouldn't work
16:57 bhosmer joined #salt
16:57 timoguin jamilner: the master doesn't care about the filename of the minion key
16:58 timoguin it just ties the key contents to a minion_id
16:58 babilen Wait, aren't we talking about the key name on the minion side?
16:58 babilen s/key name/file name/
16:58 timoguin yea babilen
16:58 iggy I thought so
16:58 babilen anyway, /me is off
16:58 iggy which explains my perplexed look
16:59 babilen But in that case it shouldn't matter if you have a minion.pem (even a different one) on each minion
17:01 jamilner ok so are you saying the auto-generated minion.pem and minion.pub keys are the keys that should be accepted on master?
17:04 ndrei joined #salt
17:06 notnotpeter joined #salt
17:06 subsignal joined #salt
17:07 scoates joined #salt
17:08 iggy we're saying the keys you tell the master to accept should be installed on the minion as /etc/salt/pki/minion/minion.p{em,ub}
17:08 jamilner ok
17:08 iggy what you call them on the master side is irrelevant, that's where the minion looks for keys
17:09 andabata left #salt
17:09 monkey66 joined #salt
17:10 jamilner so does the master key have to match up with the minion_id?
17:10 KyleG joined #salt
17:10 KyleG joined #salt
17:10 iggy yes
17:10 jamilner or does master automatically recognize the salt-minion key as long as it's titled minion.pem or pub
17:11 iggy the minion doesn't pass a file to the master as it's key... it's all over the wire directly
17:11 antani joined #salt
17:12 jamilner ok i think i got it, so the preseeded minion key has the name of minion.pem and pub and then the master identifies the minion through the FQDN
17:12 kaptk2 joined #salt
17:13 zer0def ok, i'm yet to play around with syndic, but i'm curious: how does a syndic present grains of it's minions, if at all?
17:13 iggy through the minion_id I believe (which doesn't necessarily have to match the fqdn, but does by default)
17:14 denys joined #salt
17:15 j-saturne joined #salt
17:16 randomuser joined #salt
17:17 prwilson joined #salt
17:22 Ixan joined #salt
17:24 coval3nce in a top.sls, is it possible to use the “- env: formula_name” syntax under a target stanza like you can in the “include:” stanza of a normal state file?
17:24 hal58th What is
17:24 hal58th - env " from?
17:25 Ixan joined #salt
17:25 coval3nce http://docs.saltstack.com/en/latest/ref/states/include.html shows you can use a syntax in an state include statment to signify which enviroinment a formula should be pulled from - wondering if the same can be dont in a top.sls
17:26 iggy no, the formula would match the environment it's listed in (and subsequent merging)
17:27 coval3nce iggy: so if i am in environment “A”, and i call for formula “ntp”, would it be found in the “base” environment if not explicitly called out in envrionment “A”’s gitfs backend?
17:28 UtahDave joined #salt
17:28 nickdew joined #salt
17:28 coval3nce still tooling with idea of how to set upstream formula sources without having to fork every formula and re-create my env/branch structure
17:28 stoogenmeyer_ joined #salt
17:29 nickdew joined #salt
17:29 iggy yes
17:29 iggy well... should
17:29 timoguin joined #salt
17:29 iggy if not, I'd consider it a bug
17:29 coval3nce hmm, that is not working for me…will have to look into it more
17:30 stoogenmeyer_ joined #salt
17:30 BretFisher I'd like to help fix these SmartOS bugs if possible: https://github.com/saltstack/salt-bootstrap/issues/473
17:30 perfectsine joined #salt
17:30 BretFisher Not new to github/PR's, but new to salt
17:31 BretFisher I've tested that each is still an issue and the fixes work
17:31 coval3nce i’ve got one github “site” for my states/pillars as a gitfs remote, then added a bunch of upstream formulas as gitfs remotes as well, but seem to be getting errors of the formula not existing in my environment, which should exist in the base/master upstream of one of my gitfs remotes
17:31 BretFisher but wondering if "stranger PR" is the proper next step?
17:31 stoogenmeyer_ joined #salt
17:32 iggy BretFisher: the salt devs love people contributing... even if it's only a couple lines
17:32 iggy (I mean so long as it's actually correct)
17:33 stoogenmeyer_ joined #salt
17:33 iggy coval3nce: what version of salt?
17:34 coval3nce 2014.7.0
17:34 kusams joined #salt
17:34 keimlink joined #salt
17:34 iggy might try a newer version to see if it's fixed (a _lot_ of stuff was fixed in 2014.7.4... although most distros don't have packages for it yet)
17:35 coval3nce will do
17:35 coval3nce ill go puruse the changelogs
17:36 BretFisher iggy: thanks, I keep hearing this is a good community and number of commiters is impressive, but didn't want to do the work of putting together a PR just to have it sit. Most/all of these fixes are just a few lines I think
17:37 iggy well, salt-bootstrap is maintained by a smaller subset of people than salt itself, but you shouldn't wait too long
17:37 iggy but do keep in mind they are working on 2 different releases right now
17:37 iggy so... a couple days isn't out of the question
17:37 stoogenmeyer_ joined #salt
17:38 bhosmer joined #salt
17:41 UtahDave coval3nce: We try to fix bugs as far back as possible, so if you can fix it in 2014.7.0, that would be really awesome
17:41 Gareth morning morning
17:43 coval3nce UtahDave: as soon as i can confirm this is a bug i’ll look at it ;)  I was expecting a hierarchy merge behavior for formulas in gitfs, but was seeing otherwise.  Very much new to salt so still trying to confirm if what i expected is indeed the intended behavior or not.
17:43 iggy there's a bug open about env merging
17:45 preachermanx iggy: Due to all the fixing going on, is best practice right now to build salt-master + salt-minion from scratch and distribute it to your clients?
17:46 coval3nce iggy: can you post me that issue link?
17:47 hal58th_ joined #salt
17:48 iggy preachermanx: it's python, so doesn't exactly require building... I install from source, but that's for historical reasons... I'd eventually like to get back to using packages
17:49 murrdoc joined #salt
17:49 murrdoc whats a good way to manage files under /sys/modules
17:49 kusams joined #salt
17:49 iggy coval3nce: https://github.com/saltstack/salt/issues/12483
17:50 speedlight joined #salt
17:50 speedlight joined #salt
17:52 kusams Can somehelp me
17:52 kusams in setting virtualenv to have certain version of python
17:52 manfred with salt?
17:53 kusams 2014.7.0
17:53 kusams yes
17:53 manfred http://docs.saltstack.com/en/latest/ref/states/all/salt.states.virtualenv_mod.html#salt.states.virtualenv_mod.managed
17:53 manfred the python= variable
17:53 manfred python: /usr/bin/python3
17:53 manfred or whatever
17:54 kusams thx i will try tht
17:55 beneggett joined #salt
17:59 cheus joined #salt
18:00 otter768 joined #salt
18:01 aparsons joined #salt
18:01 rockinwaffle joined #salt
18:05 rockinwaffle joined #salt
18:05 rockinwaffle left #salt
18:05 baweaver joined #salt
18:14 nickdew joined #salt
18:18 sandah joined #salt
18:18 moos3 anyone run into a issue where the salt master runs on a lxc and and everything command has a major delay in response ?
18:18 dopesong_ joined #salt
18:18 oxae joined #salt
18:21 bhosmer joined #salt
18:23 monkey66 joined #salt
18:25 druonysus joined #salt
18:25 druonysus joined #salt
18:25 UtahDave Ryan_Lane: you around?
18:28 quist` joined #salt
18:28 Gareth UtahDave: hey.
18:28 UtahDave how's it going, Gareth!
18:29 Gareth UtahDave: Going well :) How's it going with you?
18:29 UtahDave pretty good!  Finishing up a talk I'm giving in a couple days.
18:30 Gareth Nice.
18:30 UtahDave I'm finishing up collecting some historical stats and events about Salt
18:31 Gareth Doing a talk on the history of Salt?
18:31 baweaver joined #salt
18:32 UtahDave a small part of the talk will be.
18:33 aparsons joined #salt
18:33 Gareth Cool.  I hope you included the LoTR bit around the reason for the name :)
18:34 UtahDave lol, yeah, that's popular.  :)
18:36 bluenemo_ joined #salt
18:38 yggdrasil joined #salt
18:38 AJ___ joined #salt
18:39 baweaver joined #salt
18:39 AJ___ Hey, I am wondering what the difference is between SaltStack Community and Enterprise versions. Is there a restriction on number of nodes in the Community version? Any help would be much appreciated
18:40 manfred there are not any restrictions, you just get more support, and now they have a pretty awesome web gui for controlling your masters
18:40 bhosmer joined #salt
18:41 AJ___ thanks so much for that. The enterprise version has a web gui u mean?
18:41 iggy AJ___: very much _not_ like puppet/chef enterprise
18:41 iggy AJ___: yes, SSE has a web UI
18:42 StDiluted_ joined #salt
18:42 AJ___ Oh okay, great. Thank you so much for that !!
18:46 kusams joined #salt
18:48 ben_NN joined #salt
18:50 timoguin joined #salt
18:56 cro joined #salt
18:56 druonysuse joined #salt
18:56 druonysuse joined #salt
18:57 jeremyalons I don't suppose there's an 'uptime' option in salt for the minion itself is there
19:01 coval3nce iggy: thx for that issue link, read the whole thing, and i did notice one guy claiming the behavior i am seeing in terms of a deep merge when using gitfs environments
19:02 coval3nce iggy: e.g - if including a forumla in an environment other than base/master, that formula does ineeed need the corresponding environment branch
19:02 cpowell joined #salt
19:02 timoguin coval3nce: yes
19:03 timoguin I had to create environment branches in every formula I forked
19:03 coval3nce i’d love to see a way to override that so that i can include upstream formulas
19:03 coval3nce and even specify a tagged release for example
19:03 timoguin Let me know if you figure it out. :)
19:03 coval3nce it does make sense to have the strict pivots of the environments, but the hierarchy makes sense in some workflows too
19:03 resting joined #salt
19:04 baweaver joined #salt
19:04 coval3nce i wonder if this behavior is just a gitfs thing or if the same holds true if using file_roots
19:04 timoguin I pulled all mine down into one folder and ran a bash loop to create and push all the branches at once
19:04 timoguin So if I work on any formula I'm using, I start out in a local testing branch
19:04 \ask joined #salt
19:04 murrdoc i rewrite it
19:05 murrdoc so its consistent with what my style
19:05 murrdoc :D
19:05 bhosmer joined #salt
19:06 timoguin I keep mine to where I can easily merge in upstream changes. And easily PR back.
19:06 timoguin Try to, at least
19:06 coval3nce I was hoping to follow the puppet r10k setup for upstream modules, but doh ;)
19:06 murrdoc :)
19:06 andrew_v joined #salt
19:06 murrdoc i send it over anyways
19:06 murrdoc then it waits for someone to merge
19:07 murrdoc that would be nice
19:10 intellix joined #salt
19:10 rap424 joined #salt
19:11 intellix left #salt
19:17 iggy I won't merge your PRs (now that I have tha powa)
19:19 JlRd joined #salt
19:19 murrdoc ass
19:21 iggy I thought you had one earlier, but I never saw it
19:21 viq joined #salt
19:22 niseak joined #salt
19:23 murrdoc https://github.com/saltstack-formulas/sysstat-formula/pull/9
19:23 murrdoc 8 days now
19:23 iggy oh that's why... not a new one
19:24 iggy murrdoc: do you mind moving that sources/ to files/ to match _every_ other formula that exists?
19:24 murrdoc boooooo
19:24 iggy and the upstream docs
19:25 iggy I actually kind of like sources/ but it's confusing with what exists
19:25 iggy (and what the docs say)
19:26 murrdoc pushed
19:33 clintberry joined #salt
19:39 jalaziz joined #salt
19:44 lictor36 joined #salt
19:45 jalaziz joined #salt
19:48 kusams joined #salt
19:50 yggdrasil joined #salt
19:50 giantlock joined #salt
19:52 p66kumar joined #salt
19:52 kusams joined #salt
19:55 hal58th joined #salt
19:58 aparsons joined #salt
19:59 j-saturne joined #salt
20:01 otter768 joined #salt
20:02 j-saturne joined #salt
20:04 cberndt joined #salt
20:04 kusams joined #salt
20:04 notnotpeter Is there a timeline when 2014.7.4 is likely to be released?
20:04 benegget_ joined #salt
20:06 manfred now?
20:06 manfred it should be in most repositories
20:06 manfred they were going to announce it today
20:06 manfred it all depends on when the rest of the packagers finish
20:08 * iggy hates that packaging lag crap
20:08 notnotpeter Thanks Manfred. Guess I'll just be patient and keep this tab open: https://launchpad.net/~saltstack/+archive/ubuntu/salt?field.series_filter=trusty
20:09 manfred notnotpeter:  joehh1 i think said he started working on them this last weekend.
20:09 iggy god, we'd just now be getting linux-2.x if we waited for distros to package crap
20:09 manfred arch has 4.0 in the testing repository
20:10 iggy that's not even one of the distros they wait for
20:10 notnotpeter Well 2014.7.4 is not in pip either...so it's not just distros ;) https://pypi.python.org/pypi/salt
20:10 baweaver joined #salt
20:10 murrdoc dont forget to hit f5
20:10 manfred iggy:  i know, cause i package the arch one :P
20:10 manfred notnotpeter: https://github.com/saltstack/salt/blob/develop/salt/utils/openstack/nova.py#L134-138
20:10 manfred bah
20:11 manfred notnotpeter:  https://pypi.python.org/pypi/salt/2014.7.4
20:11 manfred yeah it is
20:11 sjwoodr joined #salt
20:11 manfred you have to turn on 'unreleased' or something
20:11 manfred no idea how to get a release past the current stable with pip
20:12 manfred curl -sL https://bootstrap.saltstack.com | sh -s — git v2014.7.4
20:12 aparsons joined #salt
20:12 beneggett joined #salt
20:12 sjwoodr hey guys - i can do:  salt 'foo' s3.get .....    and it works fine, but if i put this in a state file, i get the error: State 's3.get' found in SLS 'blah' is unavailable
20:13 manfred s3.get is an execution module, not a state module
20:13 manfred if you did
20:13 notnotpeter manfred this works: pip install salt==2014.7.4
20:13 manfred something:
20:13 manfred module.run:
20:13 manfred - name: s3.get
20:13 manfred - m_name: thing
20:13 manfred it would work
20:13 manfred but you have to use the module state module
20:13 iggy but don't
20:13 sjwoodr whoops.  @manfred - thanks.  that's clear
20:13 sjwoodr but dont ?
20:13 manfred yeah
20:13 manfred don't do that
20:14 manfred go find what the state module is
20:14 manfred or use the s3 file backend
20:14 iggy file.managed (and other state modules) have built-in support for s3
20:14 manfred sjwoodr:  http://docs.saltstack.com/en/latest/ref/file_server/all/salt.fileserver.s3fs.html
20:14 iggy what are you trying to do exactly?
20:14 sjwoodr just pull files from s3
20:14 sjwoodr nothing fancy
20:15 manfred use the s3 fileserver
20:15 iggy what kind of files?
20:15 iggy tarballs?
20:15 manfred then you can reference them as s3://bucket/…
20:16 sjwoodr ok.  one thing i noticed is that i had to use verify_ssl: False with the execution module.... becuase our bucket names have periods in them.  Is that going to be a problem with the s3fs file backend ?
20:16 manfred maybe?
20:16 manfred never used it
20:16 sjwoodr :D  one way to find out
20:17 manfred yup!
20:18 ndrei joined #salt
20:20 hal58th_ joined #salt
20:20 Vye joined #salt
20:20 tkharju joined #salt
20:22 cptsupermrkt joined #salt
20:26 keimlink joined #salt
20:26 awerner joined #salt
20:27 aw110f joined #salt
20:28 aw110f I’m getting “State pip.installed found in foo is unavailable” I’m on 2014.1.13 running centos 6.4
20:29 jalaziz joined #salt
20:29 aw110f i have multiple version of python installed
20:30 iggy do you have pip installed?
20:30 aw110f yes i do
20:30 manfred can you run salt \* pip.list ?
20:30 aw110f I’m running python 3.4.3
20:31 iggy salt doesn't run under python3
20:31 manfred yeah
20:31 manfred you can't use pip like that from salt
20:31 manfred because it has to import the module
20:31 iggy not sure if the pip state looks for the module or the cmdline
20:31 aw110f but I’d like salt to manage python 3.4.3 package
20:31 manfred it does not
20:31 iggy cmd.run
20:31 manfred you are gonna be sol until salt fully runs on 3.4.3
20:31 manfred you could do cmd.run
20:31 manfred or you could use the virtualenv state
20:31 manfred and make a virtualenvironment
20:32 manfred that is what I do
20:32 aw110f there’s a bin_env option in the pip state which specifies the pip bin path
20:32 manfred i don't think that matters
20:32 manfred i have never heard of it working
20:32 manfred like the idea is there
20:32 manfred but the problem is that it actually imports the pip module to check if it exists iirc
20:32 iggy import pip\nHAS_PIP = True
20:33 iggy so it's definitely trying to import it
20:33 manfred yup
20:33 manfred which breaks it for the state
20:33 manfred the module should still work, cause it doesn't have an import
20:34 manfred so it is purely a problem with how they used the execution module to implement the state
20:34 cptsupermrkt What are salt's plans to support Python3?  Is it actively in the works? https://github.com/saltstack/salt/issues/11995 looks kind of dated.
20:34 iggy define actively
20:34 manfred because it does stuff like this https://github.com/saltstack/salt/blob/develop/salt/states/pip_state.py#L122
20:34 hal58th joined #salt
20:35 manfred cptsupermrkt:  you can install it with python3
20:35 aw110f Looks like it’s going to be a problem for users who have multiple version of python ?
20:35 manfred but a large number of moduels don't work yet
20:35 manfred aw110f:  maybe? not if you use virtualenvs
20:35 iggy aw110f: no, just for people who want to manage system modules on a version different than that running salt
20:35 manfred the virtual env stuff is solid, and you can specify a requirements.txt for pip packages to install inside it
20:36 manfred cptsupermrkt:  http://docs.saltstack.com/en/latest/topics/releases/2015.2.0.html#python-3-updates
20:37 cptsupermrkt Cool, thanks.
20:37 aw110f iggy: right
20:39 manfred aw110f:  don't do that, just have the main package, and then python<ver>-devel and setuptools, and you are all set to go to setup a million different virtualenvs, and not mess with the system python, which to be honest
20:39 manfred is really fragile and prone to breaking
20:40 awerner joined #salt
20:41 sunkist joined #salt
20:42 diegows joined #salt
20:42 diegows hi
20:43 diegows this command is working salt -t 300 -v minion utils.get_appenvs
20:43 diegows but
20:43 diegows it's failing in a sls file
20:43 diegows Rendering SLS 'base:composer' failed: Jinja variable 'salt.loader.LazyLoader object' has no attribute 'dwsutils'
20:43 diegows but not on all platform, only in centos
20:43 diegows python 2.6.x
20:43 manfred can you put the sls file on something like gist?
20:44 manfred dnsutils?
20:44 iggy diegows: do you call a module directly anywhere?
20:44 diegows well, this is the important line {%- for app_env in salt.dwsutils.get_appenvs() %}
20:44 diegows it's a custom module
20:44 manfred oh
20:44 diegows iggy, I'm calling it from cmdline and works
20:44 manfred salt['dwsutils.get_appenvs']()
20:44 manfred it is an object in the salt dictionary
20:45 iggy https://github.com/saltstack/salt/issues/22444
20:45 diegows manfred, that syntax works
20:45 hasues left #salt
20:45 diegows but it's failing for some reason there :)
20:45 iggy salt.module.dwsutils
20:45 iggy ...
20:45 manfred oh yeah
20:45 manfred that bug
20:45 diegows ooops
20:45 iggy modules
20:45 diegows good to know that I'm not alone...
20:45 jeremyalons whiteinge: Is there any way to get the equivalent of -v on the netapi modules?  It doesn't look like passing verbose as a kwarg=True is sufficient, and there's no verbose parameter
20:45 iggy but that issue I linked might also be related
20:46 awerner1 joined #salt
20:46 diegows let me try :)
20:46 iggy without seeing the actual state, it's hard to say
20:47 diegows well, it works now
20:47 manfred diegows:  did you try the salt.modules.dwsutils.get_appenvs() ?
20:47 diegows it's running, so it was rendered
20:47 manfred i had no idea we could do it that way, i always just use the salt[''] method
20:47 iggy as well you should
20:47 manfred it is better <3
20:48 diegows I like the salt.module.func() syntax
20:48 diegows :)
20:48 manfred salt.modules.module.func()
20:48 iggy except it doesn't work
20:48 iggy ^
20:48 diegows but it used to work
20:48 iggy salt['module.func']() seems shorter to me
20:48 manfred probably wasn't the intended use case
20:48 manfred and something else was broken
20:49 diegows I don't like to [] '' :)
20:49 diegows manfred, salt.modules.dwsutils.get_appenvs doesn't work either
20:49 jhauser joined #salt
20:49 iggy because of that bug
20:50 diegows yes, I supposed
20:50 iggy what version are you on?
20:50 diegows 2015.2.0rc2
20:51 iggy okay, good
20:51 manfred i put money on that bug
20:51 iggy so you're at least expecting a little breakage
20:51 manfred which is a blocker for the stable release of 2015.2
20:51 iggy ^
20:51 iggy along with mine ;)
20:51 perfectsine joined #salt
20:52 diegows subscribed to the bug :)
20:53 iggy there's a patch for it, just not merged
20:57 perfectsine joined #salt
21:00 yml joined #salt
21:00 yml hello
21:00 murrdoc does not parse
21:00 yml is there anyone sprinting @pycon ?
21:03 iggy I think base of pi is there
21:03 iggy dunno about anyone else
21:03 murrdoc based pi is giving his talk there
21:04 soren joined #salt
21:05 yml I see
21:05 yml because I am sitting in the room where the saltstack sprint was suppose to happen
21:05 benegget_ joined #salt
21:05 yml and there is only ansible people
21:05 yml so I was wondering :-)
21:06 soren I'm trying to work out how best to convert a defined type in Puppet to something in Salt. I have a consul_service type which takes a few arguments: a port number and a check script.
21:06 iggy basepi: ping ^
21:06 soren Each service I have also instantiate one of these consul_service types with the port number it runs on and a check script.
21:06 soren ...and then that defined type does the right thing (create config file, SIGHUP consul, etc.).
21:07 iggy realistically if you are trying to match functionality 1:1 between puppet and salt, you are doing it wrong
21:07 soren I'm sure.
21:07 soren So help me think about this in a more Salty way.
21:08 iggy (also... I know next-to-nothing about puppet, so I couldn't help convert stuff like that anyway)
21:08 iggy I imagine there are others that didn't spend much time with other tools before they ended up on salt either
21:09 soren Basically, I have a resource type which encapsulates a bunch of other resources.
21:09 soren Suppose I use Apache to terminate SSL for a number of web services.
21:10 soren In Puppet, I might define a resource type that creates a virtualhost, adds it to my monitoring system, etc. That resource type would simply take a few arguments: port number, backend url, friendly name
21:12 soren So, for each of these services, I'd just reference this custom resource type and it would take care of all those things for me.
21:12 iggy you could write a custom module to do it
21:12 soren Instead of each of them having to create a virtualhost resource, a monitoring resource, etc.
21:13 soren Ok. I took a quick look at the docs and it seemed custom modules was mostly for things that do something special.
21:14 soren I'm not sure how I'd go about serializing the current state of a collection of contained resources.
21:14 soren I mean... Sure, I can do it, but I wonder if there's some generic method for it.
21:14 iggy no
21:15 iggy and in general, most salt states are so short (or so overwhelmingly complex) that it wouldn't make sense to try to roll them all up together
21:16 iggy I mean you could lay out your pillars in a certain way and have a couple different states all just feed off that same pilar
21:17 soren Hm.
21:17 soren Not only do these custom types in Puppet let me encapsulate several resources that always need to be created together, but it also makes it easy to make changes to all of them in one go.
21:17 soren I don't know... This is so incredibly useful, I can't imagine not having it. I must just be thinking about this all wrong.
21:18 iggy it's not a pattern I've seen in salt
21:19 soren The web service example is a real one. The various web services are very, very different, but this one part is common for all: (Port number, backend base URL, friendly name) -> Apache virtualhost, consul monitoring script, logrotate config + a couple more things.
21:19 iggy but then, outside of formulas, I haven't looked at many people's states
21:19 aw110f Hi, about peer communication, I cannot seem to get a minion to be able to run salt commands to other minions
21:20 aw110f [INFO     ] publishing ‘cmd.run’ to tcp://hidden_ip:4506
21:20 aw110f local:
21:20 soren iggy: So this would generally involve having a state per service and then just also adding this service to a list in a pillar which then results in all these other resources being created?
21:20 aw110f ———---
21:21 mward1028 joined #salt
21:21 iggy aw110f: make sure you actually have the perms setup for publish
21:22 mward1028 left #salt
21:22 aw110f yea i set them in the master config and restarted the master
21:22 aw110f I actually tried to allow all minions to do peer communication
21:22 aw110f peer:
21:22 iggy soren: I mean you could throw it all in one file and have that assigned in the top file, but yeah... I've never seen a module that does multiple things like that
21:22 aw110f .*:
21:23 aw110f -*
21:24 bhosmer joined #salt
21:24 iggy I am pretty sure publish doesn't work on 2014.1.x
21:24 iggy I had problems with it and abandoned the idea
21:25 aw110f does it work on 2014.7 ?
21:26 soren iggy: So if you had 3 different, unrelated services they would of course be in 3 different states...  and if you then wanted to do something for all three. Like add an apache virtualhost, logrotate config and a config file for a monitoring system.. Would you generally directly create those three things in each of the 3 states?
21:28 iggy aw110f: I skipped 2014.7, but it looks like it's working fine in 2015.2 (with the same config that _didn't_ work in 2014.1)
21:30 aw110f iggy: what are the work around to get minions to talk to other minions?
21:30 iggy aw110f: I don't know... I never got it working
21:31 iggy soren: most of my services are grouped by role
21:31 murrdoc sometimes i get pissed
21:31 murrdoc salt 'minionname' sysctl.persist net.bonding.bond1 1 passes
21:31 murrdoc but in a state it fails
21:31 soren The alternative is to have a fourth state that creates these resources and make it pull its data from a pillar that has the config for the 3 services.
21:31 soren s/The alternative/An alternative/
21:31 soren ...but that also doesn't feel very elegant.
21:31 aw110f ok thanks iggy:
21:36 cansis_ joined #salt
21:36 Ligthert_ joined #salt
21:36 tmmt_ joined #salt
21:36 sophomer1c joined #salt
21:36 sk_0_ joined #salt
21:36 wintamut1 joined #salt
21:37 soren iggy: I don't think I understand what that means.
21:38 murrdoc thoughts on what should be managing the /proc/net/*
21:38 bVector joined #salt
21:38 iggy file.managed
21:38 ecdhe joined #salt
21:39 lictor36 joined #salt
21:40 iggy soren: use a roles grain and have everything applied to hosts based on that... multiple states, yes, but they are mostly pillar data driven, so my web servers get their ssl certs, config, etc from pillar and then have a nginx, supervisord, virtualenv, etc states applied
21:40 clintberry joined #salt
21:41 soren murrdoc: You mean /proc/sys/net, right?
21:41 murrdoc nah
21:41 soren murrdoc: Not much to do in /proc/net
21:42 murrdoc /prob/net/bond/bonding
21:42 coval3nce I’ve verified gitfs origin branches with correct commit values under /var/cache/salt/master/gitfs/<hash>/.git/refs/remotes/origin, but my environment which maps to a branch isn’t being pulled to populate the cache files on the minions, any ideas on how to troubleshoot that outside of “salt-master -l debug” ?
21:42 soren murrdoc: I'd advise against doing anything in /proc/net.
21:42 murrdoc i just need to verify it
21:42 murrdoc like does it exist and if it does, is it equal to 1
21:43 soren murrdoc: I would use the respective commands. ifenslave or whatever the cool kids use nowadays.
21:43 iggy coval3nce: salt-run -l debug fileserver.update may give some insight
21:43 murrdoc soren:  yeah that makes sense
21:43 iggy systemd... duh
21:44 murrdoc no ifenslave module
21:46 murrdoc https://github.com/saltstack/salt-contrib/blob/master/modules/linux_netconfig.py
21:46 coval3nce iggy: thx, shows no errors and return code of 0
21:46 murrdoc is there from like 2011
21:47 LtLefse murrdoc: I just peeked at the monitoring plugin I use to check bonding and it reads /sys/class/net/$bondname/operstate
21:47 murrdoc yeah
21:47 murrdoc is that a systemd implementation ?
21:47 jhauser joined #salt
21:47 LtLefse no, rhel6
21:48 LtLefse iirc /sys and systemd are unrelated
21:48 soren iggy: Thanks, btw.
21:49 soren LtLefse: Correct.
21:58 baweaver joined #salt
21:58 Guest70 joined #salt
22:02 otter768 joined #salt
22:11 beneggett joined #salt
22:12 murrdoc joined #salt
22:12 kunit joined #salt
22:12 kunit hey saltines, what do I do if I need to add user that has same name as service?
22:14 iggy svc_name:\n  user.present: []\n  pkg.installed: []\n  service.running: []
22:14 iggy or whatever
22:14 JDiPierro joined #salt
22:15 kunit so, nest user in service?
22:16 nzero kunit, user.present is a child of svc_name, not service.running, but yes, nest all things with the same name under one tag
22:17 amcorreia joined #salt
22:17 kunit excellent, gracias the both of you.
22:17 nzero or, you can do something like svc_user:\n user.present:\n - name: user_name
22:17 dalexander joined #salt
22:18 basepi yml: We didn't sign up for a sprint this year....we're all home, actually
22:18 murrdoc joined #salt
22:18 basepi Did we still get put on the list?
22:18 murrdoc ?
22:18 basepi (pycon)
22:18 fusionx8_ joined #salt
22:20 basepi yml: if we're on the list at pycon, could you please remove us from said list?
22:22 subsignal joined #salt
22:22 timoguin joined #salt
22:23 primechuck joined #salt
22:28 _JZ_ joined #salt
22:28 dalexand_ joined #salt
22:32 florinandrei joined #salt
22:33 echo joined #salt
22:36 murrdoc joined #salt
22:37 raygunsix joined #salt
22:40 Gareth bah.  devel is busted.
22:41 * Gareth throws kraft singles at basepi
22:42 basepi :(
22:42 ndrei joined #salt
22:43 * basepi is actually out sick today, so he gets to pass the buck.
22:44 nzero *pass the cheese
22:45 Gareth Throwing cheese isn't a sign of blame.  In some cultures (see what I did there?) it's the highest compliment.
22:45 * iggy boos
22:47 aquassaut joined #salt
22:50 dude051 joined #salt
22:52 dude051 joined #salt
22:54 dude^2 joined #salt
22:55 rhodgin joined #salt
22:55 _JZ_ joined #salt
22:58 autonomy joined #salt
23:00 vietor joined #salt
23:01 irctc773 joined #salt
23:02 coval3nce iggy: “gitfs_root: states”…..yeah that bit me pretty good ;)
23:03 iggy you are trying to put states and pillars in the same repo?
23:03 coval3nce yeah i was, thinking i should break that out
23:03 coval3nce make it clear
23:03 iggy yeah, don't do that
23:04 coval3nce ;)
23:06 jalaziz joined #salt
23:06 iggy the default backend doesn't even support doing that
23:07 hal58th joined #salt
23:10 buMPnet joined #salt
23:10 baweaver joined #salt
23:14 keimlink joined #salt
23:14 murrdoc joined #salt
23:17 lahwran joined #salt
23:17 I3olle joined #salt
23:20 keimlink_ joined #salt
23:27 rhodgin_ joined #salt
23:35 baweaver joined #salt
23:36 theologian joined #salt
23:38 jalaziz joined #salt
23:46 baweaver joined #salt
23:49 bfoxwell joined #salt
23:52 raygunsix joined #salt

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