Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-03-20

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

All times shown according to UTC.

Time Nick Message
00:00 rocket joined #salt
00:02 baweaver joined #salt
00:07 catpiggest joined #salt
00:08 mikaelhm joined #salt
00:10 Kelvin joined #salt
00:12 iggy [x] add gitfs backend installs to salt-formula
00:12 Kelvin problem here: salt master shows, randomly, this message ager issuing a command to a minion:
00:12 Kelvin Failed to authenticate!  This is most likely because this user is not permitted to execute commands, but there is a small possibility that a disk error occurred (check disk/inode usage).
00:13 Kelvin I have not found much on this in the docs, am I the only one with this problem?
00:13 hal58th What's the command Kelvin? Are you sudoed?
00:13 iggy I suppose you checked inode usage?
00:13 Kelvin I am sudoed
00:13 Kelvin inodes not an issue (nagios says)
00:14 Kelvin also, this happens whether or nit I am sudo or just root
00:14 Kelvin (not)
00:15 hal58th What's the command, or is it all commands?
00:15 dude051 joined #salt
00:15 iggy shutdown salt-master, check/kill any rogue salt processes, start salt-master
00:15 forrest in addition to hal58th and iggy, can you confirm it occurs for random machines? All machines? Only some machines?
00:15 Kelvin the command can be as simple as cmd.run ls or a state.sls with lots of clauses
00:16 hal58th forrest it's a salt-master command
00:16 forrest hal58th: maybe I am missing something, he says issuing a command to the minion
00:16 forrest from the master
00:16 Kelvin sorry, I am on the master
00:16 hal58th from master, you get the error on the master
00:16 Kelvin yes
00:17 forrest hal58th: Yeah that's fair, I was thinking it was during a highstate and that was returning after X successful runs
00:17 mikaelhm joined #salt
00:17 forrest iggy has the right idea, stop salt-master, kill rogue process, start it back up
00:17 forrest good place to start at least
00:18 Kelvin OK, I will kill off the salt-master now and try again. I have not done this yet
00:18 hal58th make sure the processes are all gone.
00:18 iggy the only 2 times I've seen that error was A. ran my shit out of inodes B. rogue salt-master processes
00:18 forrest iggy: Same here, minus the one time someone said they were running as sudo, when they weren't
00:19 forrest also, doesn't nagios only do a check every X minutes or something? I'd still check inodes directly from the system
00:19 iggy ^
00:19 iggy I mean presumably, you're right there already
00:19 forrest I know our nagios is always a few minutes late.
00:19 forrest right
00:19 Kelvin these are reallt stable minions, guys, no inode issues, I promise
00:19 overyander do i have to specify a condition in an elif statement? can i just do if condition do this, elif do this  ?
00:20 forrest Kelvin: We're just confirming, I mean, trusting nagios? PSSSshhhh :D
00:20 iggy the inode issue is on the master (specifically the cache dir runs out of inodes)
00:20 iggy overyander: else
00:20 iggy just else
00:20 overyander thanks iggy
00:20 * iggy throws python books at overyander
00:21 Kelvin OK, I am actually testing with check_mk in omdistro, only thing Nagios does is sit there looking pretty
00:21 cmaldonado joined #salt
00:21 overyander iggy, i need them. also need the time to read them. lol
00:21 iggy I was thinking a simple "df -i" but that works too
00:21 overyander thanks for all the help though.
00:21 forrest df -ih
00:21 forrest iggy: Screw your unreadable shennanigans
00:22 Kelvin salt-master has lots and lots of free inodes
00:22 iggy overyander: I can't help with time (if I could, I'd probably not be here, lol)
00:23 robawt hey team
00:23 robawt anyone have the salt air episode on runners?
00:23 overyander tv show? i'm lost. (as usual)
00:24 iggy I thought they were all on the youtubez
00:24 robawt youtube thing Salt does
00:24 mordonez joined #salt
00:24 Kelvin restarted the salt-master - there were zombie salt-masters and old salt-master processes. thank you all
00:25 iggy ^5
00:25 Kelvin right back
00:26 forrest the episodes don't have any proper tags/description, so you can't find the specific on one salt runners
00:26 iggy SSCEs: 1 Bugs: 50
00:26 robawt forrest: yeah :\
00:26 forrest let's see, when was the runner system introduced...
00:26 robawt forrest: i'm thinking 14-16?
00:27 hal58th Awesome Kelvin. Glad we figured it out.
00:27 nitti joined #salt
00:27 dalexander joined #salt
00:27 forrest robawt: I see it as far back as .12
00:28 forrest .13 seems to be where it got an update.
00:28 jaloren joined #salt
00:28 jaloren hi all...i am having problem with salt-cp
00:28 jaloren it just seems to have stopped woorking
00:28 Kelvin add to check_mk: look for zombie salt-masters
00:28 jaloren logs say nothing
00:28 jaloren but nothing gets colpied over
00:28 jaloren it just returns an empty dict
00:29 jaloren any points of where i could look to find out why its failing
00:29 jaloren minion logs are quiet
00:29 robawt forrest: i've given up and resorted to trial and error
00:29 robawt thanks for helping anyway
00:29 forrest jaloren: Run the command with -l debug
00:30 itru joined #salt
00:30 forrest robawt: Yeah np, why are you trying to find the video anyways?
00:30 forrest it's pretty out of date
00:30 jaloren forrest: i did that and trace
00:30 robawt forrest: a "this is how I " intro
00:31 forrest robawt: Uhh did the salt air talk about that?
00:31 robawt forrest: it's been a long day i could be wrong
00:31 robawt :)
00:31 jaloren forrest: https://gist.github.com/jaloren/aa3f59311c7243f090d5
00:31 AnxiousGarlic joined #salt
00:31 iggy are you trying to write a new runner?
00:31 forrest robawt: The docs are a bit lacking :\
00:31 forrest sad panda
00:31 forrest http://docs.saltstack.com/en/latest/ref/runners/ has something, but not much
00:32 forrest jaloren: can you try it again, but increase your timeout value?
00:32 forrest jaloren: like -t 30 or something depending on how big that zip is
00:33 jaloren sure
00:33 forrest jaloren: Also, what is the status of this jid: 20150320003021251647
00:33 forrest salt-run jobs.lookup_jid 20150320003021251647
00:33 forrest jaloren: ^
00:34 forrest check that too, see if there is anything
00:34 itru if I usually set jinja variable like {% set something = { ‘name’ : ‘blabla’ %} and then override it by merge=salt['pillar.get’](‘something:lookup'))) from pillar….
00:34 itru can I have a YAML file in salt (not in pillar!) and load this jinja variable from that YAML file? because this jinja data scructures are ugly comparing to YAML :(
00:34 jaloren forrest: https://gist.github.com/jaloren/75ea54725dced3156985
00:34 spookah joined #salt
00:35 jaloren forrest: did the jid  lookup and it returns zilch
00:35 jaloren also this is like 3 mb....it should take like a second to copy lol
00:35 iggy itru: yes
00:35 forrest itru: Yeah you can, take a look at the redis example for a (somewhat complicated) example: https://github.com/saltstack-formulas/redis-formula/tree/master/redis
00:35 iggy check out the salt-formula's map.jinja
00:35 * iggy flexes
00:36 itru wow guys, thank you for quick answers! O_o
00:36 forrest pssh, my copypasta won iggy
00:36 forrest salt-formula is probably a better example though :P
00:36 iggy mine actually uses yaml
00:36 forrest HEY
00:36 forrest shhhhh
00:36 forrest shhhhhhhh
00:36 forrest jaloren: That sucks
00:37 jaloren forrest: lol i agree. I am stumped....plus my salt-cp was working not more than 5 minutes ago
00:37 baweaver joined #salt
00:37 forrest jaloren: Did you double check to make sure there weren't any salt-cp bugs on whatever release you are running?
00:37 forrest jaloren: oh well shit then
00:37 forrest jaloren: Let's pull an iggy here, maybe stop the master, kill any bad jobs, restart
00:37 jaloren i don't think anything radically changed to the system in that time frame
00:37 forrest probably not
00:37 jaloren i did try the service restart
00:37 jaloren i whacked the master and the slaves
00:37 forrest jaloren: Yeah but try actaually stopping the service
00:37 jaloren should i try clearing the cache too?
00:37 forrest and confirm there aren't any salt processes still running
00:38 hal58th map.jinja is from pillar though right? which itru did not want
00:38 jaloren ohhhh
00:38 iggy master's got plenty of space (AND INODES!)
00:38 iggy ?
00:38 jaloren okay i see
00:38 jaloren just a sec
00:38 iggy hal58th: go look... it's smexy
00:38 hal58th kk
00:38 forrest hal58th: The map is merged into the pillar
00:38 iggy I can't do it justice with words
00:39 forrest I can
00:39 forrest *blows raspberry*
00:39 itru forrest: in redis we have {% set redis_settings = {
00:39 itru 'appendfilename': 'appendonly.aof',
00:39 itru ….
00:39 itru what I want is to have
00:39 itru redis_settings:
00:39 itru appendfilename: appendonly.aof
00:39 itru ….
00:39 itru in some file in salt… not to deal with jinja {{{}}}}}
00:39 forrest itru: Then look at iggy's example
00:39 forrest https://github.com/saltstack-formulas/salt-formula/blob/master/salt/defaults.yaml
00:39 forrest you can see how it's handled there, then imported into the map.jinja
00:40 forrest It's really good, I just like poking fun at iggy
00:40 forrest yeesh 5:40 already, jaloren once we figure that out (or if it's 6 before we do) I am outta here.
00:41 itru forrest: oh yeah, that’s what I wanted! thanks again!
00:41 JoshuaX joined #salt
00:41 forrest itru: Yep it's all iggy, so all thanks should be directed to him!
00:41 jaloren forrest: lol forest this might be one you want to declare surrender on. So get this...
00:41 itru iggy: thank you!
00:41 jaloren 1. stopped all daemons....2. verified nothing was running with ps 3. cleared the caches 4. restarted everything
00:41 mikaelhm joined #salt
00:41 jaloren and it still doesn't work
00:42 forrest jaloren: As stupid as this sounds, the file you're copying does exist on the master right?
00:42 jaloren i am on centos
00:42 forrest you didn't trash it or change dirs?
00:42 jaloren ls says its there
00:42 jaloren :)
00:42 forrest jaloren: Damn you and your use of salt-cp! Write a state to handle this, YOU HEATHEN :P
00:42 forrest Okay let's do this then
00:42 forrest stop the salt-master again
00:43 jaloren i even tried absolute paths
00:43 jaloren still nada
00:43 forrest then run it in debug with salt-master -l debug
00:43 jaloren okay sure
00:43 forrest let's run the whooooole master with debug
00:43 jaloren ah sure
00:43 joshfu joined #salt
00:45 forrest Dear github
00:45 forrest why is your search such shit
00:45 forrest sincerely, forrest
00:45 jaloren forrest: ....so not sure what to make of this. But in the foreground the master is spewing binary data out...i've lost complete control of the terminal
00:45 jaloren maybe its doing something?
00:45 forrest yeah you'll lose control of the terminal, need a new one
00:45 forrest it's piping everything to stdout
00:45 forrest so technically when you run real commands it should dump to that window
00:46 jaloren it shouldn't have any problem with zip file should it?
00:46 forrest jaloren: it shouldn't.
00:46 forrest You know, can you try it without the compound matcher?
00:47 forrest and instead, target one system directly (after running the original command once more with debug set )
00:47 jaloren you mean with asterisk?
00:47 jaloren oh just target a single minion
00:47 jaloren sure
00:47 jaloren though i already did that before :)
00:47 jaloren this time with debug
00:47 forrest booo
00:47 jaloren everything is better with debug...its like beer!!
00:47 forrest you shouldn't need to run the command with debug since the master is already running in debug mode
00:48 iggy DEBUG THE PLANET
00:48 iggy or something
00:49 jaloren hmmm
00:49 jaloren this is something
00:49 jaloren https://gist.github.com/jaloren/9c4464a20fd1cd08cdad
00:49 jaloren let me check my disk utilization ......
00:49 jaloren all my disks are only 17
00:49 jaloren %
00:49 jaloren utilized
00:49 iggy space or inodes?
00:50 jaloren space should be fine according to df
00:50 jaloren its possible inodes are exhausted but these are brand new vms
00:50 jaloren and like nothing is on them
00:50 forrest jaloren: What's the ulimit?
00:50 iggy that job_cache tho
00:50 forrest what does it look like that is
00:50 forrest that's some progress at least.
00:51 forrest Was that when you tried to run across all of them? Or when only a single VM was targeted?
00:51 jaloren unlimited
00:51 forrest okay cool
00:51 jaloren forrest: single vm
00:51 forrest okay that's good at least.
00:51 forrest and it's unlimited on both the master, and the minion?
00:51 forrest shouldn't really matter for the minion, but hey, better safe than sorry
00:52 forrest jaloren: Actually...
00:52 forrest jaloren: What version of salt/zmq is on the master, versus the version on the minion?
00:52 jaloren forrest: all unlimited
00:52 jaloren forrest: sure i can check that....should be the same but yea i'll check
00:52 forrest jaloren: Better safe than sorry, especially since they are new.
00:52 jaloren forrest: minion and master are 2014.7.1 (Helium)
00:53 jaloren this is centos 6.4
00:53 jaloren i wished that was the problem!!
00:53 forrest jaloren: What about zmq?
00:53 forrest same release on both?
00:54 jaloren yes sir
00:54 forrest hmm
00:54 forrest Can you run other commands?
00:54 jaloren i checked python and mq based
00:54 forrest like a cmd.run
00:54 forrest okay cool
00:54 forrest or do we get the same error for all salt commands?
00:54 jaloren forrest: cmd.run works
00:54 forrest can you touch a file with cmd.run on the same minion we've been working on?
00:54 forrest in /root
00:55 forrest Also, no duplicate keys/funkiness going on with the auth dirs on the master?
00:55 jaloren forrest: touch worked fine
00:56 jaloren forrest: what do you mean about funkiness?
00:56 forrest Just anything in general, dup keys or anything along those lines that would cause salt to blow up
00:56 zwi joined #salt
00:56 jaloren not that i know of
00:57 jaloren at one point i had a bad dns problem with hostnames but i fixed that
00:57 dooshtuRabbit1 joined #salt
00:57 jaloren or so i thought
00:57 ajw0100 joined #salt
00:57 jaloren minion id picked up an old hostname from /etc/hostname
00:57 forrest well, if that was still the problem cmd.run wouldn't work
00:57 jaloren but i caught that when i added the salt-key and cleaned that up
00:57 forrest jaloren: Just now? Or previously?
00:57 forrest gotcha
00:57 forrest As much as I friggin hate to say it, can you upgrade these to a more recent release?
00:58 jaloren forrest: right...and i've been doing a bunch of stuff with salt like running custom python clients and fairly complex states
00:58 jaloren those worked fine
00:58 forrest Yeah, thus why it's lame
00:58 jaloren forrest: ....ha!!
00:58 forrest I'm not super familiar with the internal workings of salt-cp
00:58 jaloren such as?
00:58 forrest jaloren: 2014.7.2?
00:58 jaloren forrest: i can but not at the moment i am working on some stuff for my boss that i need to complete tongith
00:58 jaloren i'll have to come back to it tomorrow
00:58 jaloren but i do find it bizaree
00:59 forrest jaloren: I also find it weird, I'm curious if you can dup the issue on a more recent release.
00:59 forrest for the time being though
00:59 jaloren i'll keep you updated tomorrow....if nothing else i am sure you are curious
00:59 iggy but for realz though... use file.managed
01:00 jaloren iggy: i am just doing some throw away work...and needed to copy some files around
01:00 jaloren salt-cp seemed like the guy for the job
01:00 forrest yeah I was going to suggest the same.
01:00 iggy you're the first person I've heard of using salt-cp (I only vaguely knew it existed)
01:00 forrest jaloren: I updated your gist: https://gist.github.com/jaloren/9c4464a20fd1cd08cdad
01:01 forrest I've seen people use it before. Just not super familiar with it.
01:01 dooshtuRabbit1 joined #salt
01:01 iggy so that should tell you how much testing it probably gets
01:01 forrest jaloren: On that note, how dare you build systems that aren't fully salted!
01:01 forrest copying files like some kind of... madman!
01:01 forrest HOW WILL YOU KEEP IT DRY!
01:02 jaloren forrest: wanna hear something even nuttier
01:02 jaloren ' cp.get_file works fine
01:02 jaloren only salt-cp ...isn't copying
01:02 forrest jaloren: awesome
01:02 forrest well, screw it then
01:02 forrest just write some ghetto states, then when you get a chance update to a newer release to see if it can be duped. If so, then create an issue.
01:02 jaloren forrest:  and lol to keeping it dry...i am sorry i didn't build a docker container and run it unchanged
01:02 forrest If not, I'm going to chock it up to 2014.7.1 bugs.
01:03 forrest jaloren: :P
01:03 forrest jaloren: You should be ashamed
01:03 jaloren forrest: sure...but its going to drive baseapi nuts cause i have no logs
01:03 jaloren and i am not even sure what i am doing to cause the problem
01:03 jaloren lol
01:03 jaloren developers love those kinds of bugs
01:03 forrest jaloren: Well, we have the logs you just ran, but I would bet money on it that it goes away with an upgrade.
01:03 jaloren its broke and i have no information about what or why
01:03 forrest Yeh
01:03 forrest *yeah
01:03 forrest Anyways, it's 6:03, so I'm leaving
01:04 forrest good luck with your copy'ing
01:04 iggy you should have seen my first salt bug report... didn't even have "salt --versions" output in it
01:04 jaloren forrest: yea upgrades fix everything unless they break it
01:04 jaloren :)
01:04 forrest iggy: lol
01:04 iggy you'd think I just tripped and fell into a devops job without having ever used any open source software before
01:04 jaloren speaking of upgrades....has anyone used raet?
01:04 forrest jaloren: True, but I'd still suggest going to at least 2014.7.2. I'm not willing to go back and scour the 2014.7.1 bugs
01:04 forrest Nope, I run all masterless now
01:05 jaloren reaaaally?
01:05 jaloren how do you distribute jobs?
01:05 forrest What do you mean?
01:06 scbunn joined #salt
01:07 iggy ryan (lane) has some write ups on how they do masterless (uses tarballs snagged from s3 iirc)
01:08 forrest yeah he was storing most of his salt stuff in S3, then using IAM roles to determine what had access where.
01:08 forrest It's a good system, we aren't quite there yet
01:08 aqua^mac joined #salt
01:09 subsignal joined #salt
01:12 forrest Alright I am out of here guys, have a good evening.
01:21 jerematic joined #salt
01:23 otter768 joined #salt
01:24 baweaver joined #salt
01:25 dooshtuRabbit joined #salt
01:27 iromli joined #salt
01:27 dooshtuRabbit joined #salt
01:35 favadi joined #salt
01:36 favadi left #salt
01:38 aparsons joined #salt
01:38 treborTech joined #salt
01:41 subsignal joined #salt
01:41 robawt can runners accept arguments?
01:45 robawt answer: you can screw around and feed it sys.argv starting @ 2 since 0 = runner 1 == runner name
01:46 otter768 joined #salt
01:50 zwi joined #salt
01:52 hemebond joined #salt
01:53 hemebond Is there a term for an operation that does nothing if something is already configured as it should be?
01:53 hemebond e.g., when you run highstate, nothing is done if the system is already configured the correct way.
01:53 hemebond So you can run highstate as much as you like.
01:53 hemebond Without changing anything.
01:53 hemebond Is that a thing?
01:57 subsignal joined #salt
01:59 robawt hemebond: that's the idea of a system designed around idempotence
01:59 hemebond robawt: Thanks! I just this very second found the wiki page for Idempotence.
01:59 hemebond Thank you.
01:59 active8 joined #salt
02:00 everbird joined #salt
02:00 everbird joined #salt
02:01 everbird joined #salt
02:01 enarciso joined #salt
02:03 bhosmer joined #salt
02:03 catpigger joined #salt
02:03 everbird joined #salt
02:05 everbird joined #salt
02:06 I3olle joined #salt
02:07 everbird joined #salt
02:09 Negher joined #salt
02:12 TyrfingMjolnir joined #salt
02:13 dude051 joined #salt
02:16 jerematic joined #salt
02:17 fllr joined #salt
02:21 fusionx86 joined #salt
02:24 michelangelo joined #salt
02:32 JDiPierro joined #salt
02:41 evle joined #salt
02:48 ajw0100 joined #salt
03:02 beerguy2015 joined #salt
03:05 fllr joined #salt
03:07 Karlthane joined #salt
03:13 mwpher joined #salt
03:16 clintber_ joined #salt
03:20 favadi joined #salt
03:21 mwpher Hey, I'm trying to go through a pillar dict in Jinja, but Jinja seems to think that the dict is a string. The error message is very short but it basically says 'Jinja variable <str object> has no attribute "name"' (an attribute that only shows up at the top level of the variable)
03:22 mwpher I've tried getting the dict through both pillar[] and salt['pillar.get], it makes no difference
03:23 kermit joined #salt
03:24 scbunn joined #salt
03:26 Furao mwpher: the pillar value is a string and not a dict
03:26 Furao pillar[“name”] = “a string”
03:26 Furao look at your pillar structure
03:26 Furao run pillar.data to see the entire pillar values
03:29 otter768 joined #salt
03:29 mwpher I'm trying pillar.data right now, but I know that pillar.items shows the increasing indentation of a dict with the row of dashes
03:30 mwpher Yes, pillar.data shows it too
03:31 mwpher Is pillar data just always returned as strings?
03:32 mwpher I know that the example code I was looking at used {% for item in pillar[] %} so I'm not sure why it wouldn't work, unless something changes between versions
03:32 Furao pillar data should return a dict
03:35 mwpher Ok, thank you. Is there a way to enable Jinja debugging? The error messages are very terse.
03:36 mwpher And when you say 'pillar data' I assume you don't mean the pillar.data method
03:41 Furao jinja is render on the master
03:42 Furao run master @ debug level (what i do is stop it and run “salt-master -l all” in a term)
03:42 Furao pillar is rendered on the master
03:45 mwpher Giving that a shot now
03:46 mwpher Amazing how many modules it has
03:46 johnkeates joined #salt
03:47 Furao and it took forever to load :)
03:47 Furao I have like 12 additional custom modules :|
03:47 Linuturk joined #salt
03:48 overyander I just saw that 2015.2 was in the downloads folder. is that ready to use now?
03:48 germs_ joined #salt
03:49 overyander http://docs.saltstack.com/downloads/
03:50 Furao pick the version that is less buggy for you
03:50 Furao i’m still on 2014.1.x
03:52 johnkeates is there a method to see how far along salt releases are in the dev-testing-production path?
04:00 johnkeates left #salt
04:04 mwpher joined #salt
04:04 _JZ_ joined #salt
04:07 Furao joined #salt
04:24 hintss joined #salt
04:27 mapu left #salt
04:31 Nazca__ joined #salt
04:32 hintss so, uh
04:32 hintss we lost the box running our salt master, and didn't have a backup of the keys
04:32 hintss how do I get the minions to accept a new master?
04:43 Vynce http://docs.saltstack.com/en/latest/ref/states/testing.html is not explicit — is state.highstate test=True a dry-run (aka no-op)?
04:43 Vynce hintss: it hink you just change the master in /etc/salt/minion
04:45 Furao hintss: rm -f /etc/salt/pki/ some files there that are the old master pub key
04:45 Furao and restart minion
04:49 Hipikat yeah kill all /etc/salt/pki/minion/minion_master.pub :)
04:50 hintss I just rm'd /etc/salt/pki, and I'm still getting
04:50 hintss 2015-03-20 00:49:50,302 [salt                                     ][ERROR   ] Attempt to authenticate with the salt master failed
04:50 Furao hintss: + salt-key -a on the master
04:50 hintss salt-key doesn't show any unaccepted keys
04:51 dude051 joined #salt
04:52 Furao stop minion and rm -rf /var/cache/salt/minion/*
04:54 hintss nope :S
04:56 Furao who; check any logged IP from NSA range; kill those session and restart minion
04:56 hintss lol
04:57 hintss oh
04:57 hintss file permissions
04:57 hintss >_<
04:58 Furao probably NSA
04:58 hintss wait, no
04:59 hintss :S
05:01 hintss ima reinstall the mster and minion and see how that goes
05:05 hintss ok, reinstalling seems to have worked
05:08 mgw joined #salt
06:01 malinoff joined #salt
06:01 Karlthane Looking at testing salt and I have a question that my google-fu has failed. What are the minimum system requirements for a salt master?
06:15 Vynce joined #salt
06:15 Vynce left #salt
06:18 Furao joined #salt
06:21 gibmachine joined #salt
06:23 gibmachine Hey all, I'm having some trouble and I can't find anything using google fu. I'm trying to use states.win_servermanager.installed to make sure a windows server has a role installed, but it comes back to me saying that there is no ServerManager module, which it tries to import with Import-Module... Anyone know why this might be?
06:24 gibmachine I'm guessing that the module tries to perform some powershell commands via the minion agent, which seems to work, but the importing business is broken for me
06:24 gibmachine The same powershell commands work fine locally
06:24 gibmachine In fact, it's not even required to import the module
06:27 sieve joined #salt
06:27 krelo joined #salt
06:28 dopesong joined #salt
06:31 oxae joined #salt
06:38 esogas_ joined #salt
06:40 n8n joined #salt
06:47 dopesong_ joined #salt
07:08 AndreasLutro joined #salt
07:20 ajw0100 joined #salt
07:22 felskrone joined #salt
07:23 felskrone joined #salt
07:23 pipps joined #salt
07:27 dopesong joined #salt
07:31 pipps joined #salt
07:34 jhauser joined #salt
07:39 Plotek joined #salt
07:42 KermitTheFragger joined #salt
07:45 TinuvaMac joined #salt
07:46 pravka joined #salt
07:52 apergos joined #salt
07:52 apergos left #salt
07:54 sieve joined #salt
08:00 andygrunwald joined #salt
08:01 hebz0rl joined #salt
08:04 krelo joined #salt
08:04 intellix joined #salt
08:05 bhosmer_ joined #salt
08:07 Auroch joined #salt
08:07 eseyman joined #salt
08:08 __gotcha joined #salt
08:24 oxae joined #salt
08:25 jri joined #salt
08:26 trikke joined #salt
08:28 __gotcha1 joined #salt
08:31 wincyj joined #salt
08:36 mcescher joined #salt
08:44 dopesong joined #salt
08:47 dopesong_ joined #salt
08:47 kawa2014 joined #salt
08:54 alfborge joined #salt
08:57 Xevian joined #salt
09:01 pipps joined #salt
09:03 I3olle joined #salt
09:11 N-Mi_ joined #salt
09:22 a7p joined #salt
09:22 netzmonster joined #salt
09:24 thayne joined #salt
09:26 everbird joined #salt
09:35 dopesong joined #salt
09:36 dopesong_ joined #salt
09:37 sieve joined #salt
09:39 dkrae joined #salt
09:40 jri joined #salt
09:43 paulm- joined #salt
09:56 bluenemo joined #salt
10:02 pipps joined #salt
10:02 wnkz joined #salt
10:04 fllr joined #salt
10:04 sieve joined #salt
10:05 peters-tx joined #salt
10:06 wnkz_ joined #salt
10:12 aquassaut joined #salt
10:20 Furao joined #salt
10:24 ndrei joined #salt
10:26 favadi joined #salt
10:27 jri joined #salt
10:29 mcescher joined #salt
10:34 wnkz joined #salt
10:34 paulm-- joined #salt
10:36 sieve joined #salt
10:37 ndrei joined #salt
10:45 alexbst joined #salt
10:50 amcorreia joined #salt
10:53 diegows joined #salt
10:54 tzero joined #salt
10:59 CeBe joined #salt
10:59 Grokzen joined #salt
10:59 dooshtuRabbit joined #salt
11:02 bhosmer joined #salt
11:03 pipps joined #salt
11:05 fredvd joined #salt
11:06 bluenemo joined #salt
11:12 __gotcha joined #salt
11:14 Morbus joined #salt
11:16 johtso joined #salt
11:23 giantlock joined #salt
11:25 fusionx86 joined #salt
11:35 nitti joined #salt
11:36 lb1a joined #salt
11:42 otter768 joined #salt
11:42 nitti joined #salt
11:45 vstoniest joined #salt
11:45 paulm- joined #salt
11:48 __gotcha joined #salt
12:01 dRiN joined #salt
12:04 pipps joined #salt
12:04 bhosmer_ joined #salt
12:09 mpanetta joined #salt
12:10 zwi joined #salt
12:14 mpanetta_ joined #salt
12:17 ndrei joined #salt
12:18 malinoff_ joined #salt
12:18 malinoff joined #salt
12:19 sieve joined #salt
12:21 armguy joined #salt
12:22 mpanetta_ joined #salt
12:23 huddy joined #salt
12:29 cmcmacken joined #salt
12:29 giantlock joined #salt
12:30 bfoxwell joined #salt
12:35 nitti joined #salt
12:35 refnode joined #salt
12:39 jerematic joined #salt
12:41 numkem joined #salt
12:43 edrocks joined #salt
12:53 denys joined #salt
12:54 subsignal joined #salt
12:57 jri joined #salt
12:58 everbird joined #salt
12:59 nitti joined #salt
13:02 BtbN joined #salt
13:03 everbird joined #salt
13:04 intellix joined #salt
13:04 pipps joined #salt
13:07 murrdoc joined #salt
13:08 JDiPierro joined #salt
13:09 nebo joined #salt
13:09 nebo Hi :)
13:11 evle1 joined #salt
13:13 nexsja joined #salt
13:14 fusionx86 joined #salt
13:14 murrdoc :)
13:14 lb1a au revoir
13:15 I3olle_ joined #salt
13:17 everbird joined #salt
13:17 everbird joined #salt
13:18 everbird joined #salt
13:20 jdesilet joined #salt
13:22 jri left #salt
13:23 Nazca joined #salt
13:23 nitti joined #salt
13:25 everbird joined #salt
13:25 renoirb joined #salt
13:25 everbird joined #salt
13:26 everbird joined #salt
13:26 nitti_ joined #salt
13:28 dude051 joined #salt
13:30 Tecnico1931 joined #salt
13:31 yomilk joined #salt
13:34 cpowell joined #salt
13:35 lomeroe joined #salt
13:38 andabata joined #salt
13:39 dyasny joined #salt
13:39 dyasny_ joined #salt
13:40 nebuchadnezzar joined #salt
13:43 ekristen joined #salt
13:43 bhosmer joined #salt
13:43 otter768 joined #salt
13:46 paulm- joined #salt
13:50 timoguin joined #salt
13:53 mariusv joined #salt
13:53 mariusv joined #salt
13:56 debian112 joined #salt
14:00 timoguin joined #salt
14:01 andrew_v joined #salt
14:05 pipps joined #salt
14:06 gladiatr joined #salt
14:07 wincyj joined #salt
14:11 SheetiS joined #salt
14:13 StDiluted joined #salt
14:15 bluenemo joined #salt
14:16 _JZ_ joined #salt
14:19 Brew joined #salt
14:23 teskew joined #salt
14:23 __gotcha joined #salt
14:25 che-arne joined #salt
14:27 ndrei joined #salt
14:27 mbrgm joined #salt
14:28 mbrgm hi! is there a way to find out the number of (physical) nics of a minion, and match on that number? i.e. I want to set a state for all minions that have 2 physical nics?
14:29 murrdoc Easiest way i can think of is making a network-nic-count custom grain, match on that
14:40 jeddi joined #salt
14:45 TyrfingMjolnir joined #salt
14:45 Eureka_ joined #salt
14:48 iggy {% if grains['hwaddr_interfaces']|length > 1 %}
14:49 dendazen joined #salt
14:49 dendazen Hi guys i am having an issue with archive extract policy
14:49 dendazen this is my policy
14:49 dendazen http://www.paste.org/77748
14:49 dendazen and this is an error i am getting
14:49 dendazen http://www.paste.org/77749
14:49 iggy protip: gist supports multiple files per paste
14:50 dendazen but when i ‘ls -la’ that file in the error it exist in that path
14:50 XenophonF joined #salt
14:50 iggy take z and x out of tar_options
14:50 dendazen Thanks.
14:51 iggy and you might need keep: True in there as well (someone reported needing that the other day, and all of my archive.extracted's have it)
14:51 murrdoc yeah
14:52 dendazen “keep: True”?
14:52 dendazen but why z and x options are not suitable?
14:52 dendazen can you please elaborate?
14:52 iggy the docs (pretty clearly I thought) say they aren't necessary
14:53 dendazen “Main operators like -x, --extract, --get, -c and -f/--file should not be used here”
14:53 dendazen Is that it?
14:54 iggy yeah, and any tar in the last 10 years is smart enough to know x, j, J, etc on it's own
14:55 XenophonF anyone here use apache-formula and php-formula together? i'm looking for an example config
14:55 iggy I print them out and use them as TP... very rough TP
14:55 nesv joined #salt
14:55 sastorsl joined #salt
14:56 murrdoc iggy is into sadism
14:56 iggy that would be masochism
14:56 murrdoc my bad
14:56 murrdoc engrish
14:57 iggy if I made somebody else do it, it would be sadism (which I'm not personally against in the case of those formulas, so... partially correct)
14:57 murrdoc XenophonF:  no i havent used them together
14:57 murrdoc or used them at all
14:57 murrdoc should be (best case) a pillar file and two includes
14:58 sastorsl I'm writing up a state to configure the salt-master, where any settings beside the default should be set in a pillar named "salt_master_custom" (or something similar). I'm then using file.managed and "contents" to add this pillar to /etc/salt/master.d/master_config.conf
14:58 iggy sastorsl: salt-formula?
14:59 sastorsl But I'm running into issues with the yaml-filter and indent
15:00 paulm-- joined #salt
15:01 sastorsl iggy, not getting you there..
15:01 murrdoc iggy is saying, are u using the salt-formula
15:01 iggy https://github.com/saltstack-formulas/salt-formula
15:01 murrdoc if not can u paste in your pillar, and state in to a gist
15:01 murrdoc so we can look at the code
15:01 murrdoc or we can try to guess what u wrote
15:01 murrdoc :D
15:01 XenophonF sastorsl: i use salt-formula to configure my master and minions
15:01 XenophonF it works really well
15:01 iggy I'm saying you should probably be using the salt-formula (unless it just really doesn't suit your needs)
15:02 XenophonF the only challenge is that it complicates your master bootstrapping process
15:02 XenophonF so if you need to rebuild your master from scratch, you need to check out salt-formula on the new master manually
15:03 mikaelhm joined #salt
15:03 scbunn joined #salt
15:03 speedlight joined #salt
15:04 sastorsl We use gitfs
15:04 XenophonF sastorsl: as do I
15:04 speedlight hi ppl
15:04 hobakill gitfs ftw
15:04 XenophonF hobakill: gitfs+s3 ftw
15:05 sastorsl So during bootstrap the minions are installed, but the master is of course set up as pr needed.
15:05 hobakill XenophonF, oooh!
15:05 speedlight trying to install a minion in a whezzy box but i get deps error for python-requests
15:05 XenophonF hobakill: if only i could get s3fs working
15:05 Plotek joined #salt
15:05 XenophonF but the s3:// URLs seem to work OK, and setting my minions up with a read-only API key was easy enough
15:06 sastorsl But master config aside. What I'm running into is that the yaml filter does not seem to indent properly when encoutering lists.
15:06 XenophonF sastorsl: my bootstrap script installs salt-master, salt-minion, and GitPython
15:06 pipps joined #salt
15:06 XenophonF sastorsl: it clones salt-formula (plus a few other commonly used formulas, e.g., git-formula)
15:06 speedlight anyone have a wheezy minion install ?
15:07 XenophonF sastorsl: then it lays down a minimal master/minion config, automatically accepts its own key, and runs highstate on itself
15:07 XenophonF sastorsl: which applies the official configuration
15:07 hobakill speedlight, sorry - cent here but i think that issue came up a few days ago.  have you searched the channel logs?
15:07 StDiluted joined #salt
15:09 hobakill speedlight, http://irclog.perlgeek.de/salt/ has saved my tail a few times.
15:09 XenophonF speedlight: what problems are you encountering?
15:09 speedlight no yet hobakill, searching the issues, for example 497, marked as solved but donoo how i can install then..
15:09 speedlight thnkz hobakill
15:09 sastorsl i.e. XenophonF, and how do you customize?
15:10 XenophonF sastorsl: pillar
15:10 XenophonF sastorsl: the master and minion configs can be customized in pillar
15:10 timoguin joined #salt
15:11 XenophonF sastorsl: let me sanitize my salt-master pillar and paste it for you
15:11 speedlight XenophonF: http://pastebin.com/rMQejBsi
15:12 speedlight and if a try to install salt-common: http://pastebin.com/Z7K1bu55
15:12 StDiluted joined #salt
15:13 sastorsl XenophonF: that would be great
15:13 speedlight at least the majority of the servers i need as minions are centos :) i want to scape from the chef nightmare XD
15:13 StDiluted joined #salt
15:14 murrdoc how are u installing this
15:14 speedlight me?
15:14 XenophonF sastorsl: https://bpaste.net/show/8b5cc1dec4a1
15:15 XenophonF speedlight: are you following the debian install instructions on docs.saltstack.com?
15:16 murrdoc XenophonF:  your paste is missing this data
15:16 hobakill XenophonF, i'm ashamed to admit i only found out about salt-formulas last night. would have been super handy back in the day!
15:16 murrdoc -----BEGIN RSA PRIVATE KEY-----
15:16 murrdoc -----END RSA PRIVATE KEY-----
15:16 murrdoc :P
15:16 hobakill murrdoc, hahaahhaa
15:16 XenophonF murrdoc: LOL
15:17 XenophonF hobakill: you might have missed me posting my GitHub SSH keys for all to see last week
15:17 overyander i noticed 2015.2 was released yesterday, any ideas when that will make it into the epel repo or is there a more current salt-stack repo i should be using?
15:18 hobakill XenophonF, i did! oh boy. :/
15:18 murrdoc yeah, thats why it was funny
15:18 XenophonF speedlight: i followed http://docs.saltstack.com/en/latest/topics/installation/debian.html to the letter on my Raspberry Pi (Raspbian-Wheezy), and I encountered no issues
15:18 XenophonF yeah, and I have cookies disabled by default so i couldn't delete the paste
15:18 XenophonF had to rekey toute suite
15:19 sastorsl XenophonF, thx. I'm new to formulas too, so I'll need to read up.
15:19 sastorsl But.
15:19 speedlight XenophonF: the same steps i followed, to centos, opensuse, debian jessie boxes
15:19 sastorsl Given a pillar like:
15:19 sastorsl pillar_roots:
15:19 sastorsl base:
15:19 sastorsl - /srv/pillar/sandbox
15:19 XenophonF sastorsl: the only thing i have to do manually on the master is configure the aforementioned GitHub SSH key
15:20 speedlight but only in this wheezy i have this problem.. the repos are the minimal of debian stable
15:20 sastorsl and using file.managed with
15:20 XenophonF sastorsl: i'm using git for pillar data via ext_pillar
15:20 sastorsl - contents: | {{ config | yaml(False) | indent(8) }}
15:21 sastorsl the list is missing the extra 2 indends
15:22 XenophonF sastorsl: not sure i'm following you
15:23 XenophonF speedlight: maybe something in the apt sources config got messed up?
15:24 XenophonF speedlight: like i said, i followed the instructions verbatim on my raspbian-wheezy install
15:25 wnkz joined #salt
15:25 sastorsl XenophonF, https://bpaste.net/show/ab9598f79f23
15:27 sastorsl I'll still look into the formula, but the use case can still come in handy.
15:28 sieve joined #salt
15:29 speedlight strange, is just wheezy main, updates and nothing more
15:29 speedlight :(
15:30 kawa2014 joined #salt
15:31 clintberry joined #salt
15:32 pipps joined #salt
15:32 XenophonF sastorsl: it isn't shown in salt-formula/pillar.example, but it's possible to define additional pillar_roots via salt-formula - https://github.com/saltstack-formulas/salt-formula/blob/master/salt/files/master.d/_defaults.conf#L645
15:33 wnkz_ joined #salt
15:33 XenophonF this reminds me - i need to configure the master to run as non-root
15:33 iggy overyander: if it's anything like the past... 2 weeks
15:33 sastorsl Just did a "[3~git clone https://github.com/saltstack-formulas/salt-formula.git
15:33 sastorsl :-)
15:34 meylor joined #salt
15:35 Aikar hmm, does pillar top.sls not suffer the same (imo flaw)/issue of roots/top.sls that if it matches in base, it wont match in an environment? I noticed I have a '*': global in base, wouldnt that bone me if it matches behavior of roots?
15:36 Aikar but i know it is loading environment stuff still
15:36 Aikar just, is there a disconnect in behavior between pillars and roots top.sls and environment matching?
15:37 rvankleeck joined #salt
15:37 overyander thanks iggy. i've been working on a new update method for our windows minions and was about to start pushing out 2014.7.2 last night when i saw that 2015.2 was released. i think i'll just wait.
15:37 thedodd joined #salt
15:38 iggy I wouldn't consider moving to 2015.2 for a point release or two
15:38 speedlight hmmm the apt policy show the 2.0.0 of python deps in the saltstack repo, but debian repo have 0.12.0
15:38 iggy we had to make a lot of changes to move to 2015.2
15:39 speedlight i dont have preferences configured, how i tell apt to use that repo deps?
15:39 rvankleeck Is there a setting on the minion for the publish port? i am going to need to have salt-masters behind a port forwarder, and don't want to set each one on the master
15:40 overyander iggy, i had seen on here the other day that 2014.7.2 had a bug of minions disappearing. do you know if that's a widespread issue or is it generally safe to use?
15:40 iggy dunno, we skipped 2014.7 completely (2014.1 -> 2015.2)
15:40 speedlight he works with apt-get install -t wheezy-saltstack
15:40 dendazen saltstack-formulas
15:40 dendazen what are these?
15:41 dendazen never mind
15:41 iggy handy starting points
15:42 N-Mi_ joined #salt
15:42 StDiluted joined #salt
15:42 JoshuaX joined #salt
15:43 relidy_ joined #salt
15:43 dendazen Thanks.
15:43 pipps99 joined #salt
15:44 zwi joined #salt
15:44 otter768 joined #salt
15:44 dyasny joined #salt
15:44 sastorsl overyander: We've had issues in 2014.7 with minions not responding.
15:45 relidy_ left #salt
15:45 sastorsl Might be that we are at zmq3
15:45 overyander sastorsl, i have issues with 2014.7.0 with the master not responding and having to be restarted every day, that's why i was looking to update to a newer version
15:46 overyander i updated to zmq4 and didn't resolve it
15:46 sastorsl XenophonF, any input on the issue in https://bpaste.net/show/ab9598f79f23
15:47 dyasny joined #salt
15:47 speedlight https://github.com/saltstack/salt-bootstrap/issues/497
15:47 speedlight the only place I found to comment :)
15:48 zer0def joined #salt
15:48 ajw0100 joined #salt
15:49 XenophonF sastorsl: yeah, create a pillar key called salt:master:pillar_roots with a list of directories on the master as its value
15:51 Eureka_ Hey All. Any idea why my salt-minions are using their IP address as their ID? They used to use hostname I thought.
15:52 neogenix_ joined #salt
15:52 iggy they use the output socket.getfqdn()
15:53 iggy which would probably be the IP if /etc/hosts was fubar and there was no name resolution
15:53 Eureka_ ah, I get you. That makes sense for this case I suppose.
15:53 Eureka_ Ill try a few things. Thank you =)
15:56 murrdoc fun fact you can disable states globally
15:56 meylor i'm trying to get grain data to return in a mine. trying salt '*' mine.get '*' 'fqdn_ip4'
15:56 meylor but i'm getting ————— returned
15:56 fusionx86 joined #salt
15:57 meylor can anyone point me in the right direction
15:57 Ahlee Eureka_: you can specify the id in the minion config, or in /etc/salt/minion_id
15:57 iggy fun fact something in shell environment handling changed in 2015.2 and messed up a bunch of stuff... tread carefully
15:57 murrdoc based pi ?
15:57 Eureka_ Ahlee: Thanks =) I know I can specify the ID in the minion file. I was just hoping to be able to avoid doing that.
15:57 SheetiS meylor: can you sharea a pastebin of what your mine_functions looks like?
15:57 pf_moore joined #salt
15:58 meylor maybe that's the issue SheetiS
15:58 iggy meylor: fqdn_ip4 is probably not what you want
15:58 meylor I haven't defined anything else
15:58 iggy salt['mine.get']('tags:mon', 'network.get_hostname', 'grain')
15:58 iggy that's one of our mine lines
15:59 iggy so notice the lookup is a module.function
15:59 iggy in newer versions you can have aliases
15:59 neogenix joined #salt
15:59 murrdoc put mines in pillars
16:00 SheetiS iggy is correct that you probably want to use the network module in your mine functions, but you need to have the mine function defined as well
16:00 teskew joined #salt
16:00 andreyst1 joined #salt
16:00 meylor do those go in master conf
16:00 SheetiS http://docs.saltstack.com/en/latest/topics/mine/#mine-functions
16:00 SheetiS pillar or minion conf.  pillar is better.
16:00 andreyst1 SheetiS:  hahahaha
16:01 murrdoc mine.conf ftw
16:01 andreyst1 SheetiS: you won't believe it but you've answered my yet unasked question
16:01 andreyst1 thanks, man
16:02 * SheetiS has become better at precognition! (1)
16:03 jtang joined #salt
16:03 JoshuaX joined #salt
16:03 murrdoc is salt cmd.run 'stop salt-minion' the only way to stop minions
16:04 SheetiS salt '*' service.stop salt-minion <== does this not work?
16:05 murrdoc same thing
16:05 SheetiS it should be sysvinit/upstart/etc agnostic that way
16:05 murrdoc i was hoping manage had something
16:05 ek6 joined #salt
16:06 SheetiS ahh like a runner you hoped
16:06 murrdoc salt-run manage.pause
16:06 _ale1_ joined #salt
16:06 murrdoc basically saying dont run any jobs for a bit
16:06 TyrfingMjolnir joined #salt
16:06 murrdoc while i clean house
16:07 Ahlee murrdoc: build a roster with current running hosts, use salt-ssh to kill; maintenance; salt-ssh to start?
16:08 murrdoc Ahlee:  yeah, i can do that too
16:08 iggy in new versions there's a roster module that will automatically talk to everything that has a key on the master
16:08 Ahlee neat
16:08 nickdew joined #salt
16:08 Tecnico1931 joined #salt
16:08 Ahlee we're upgrading to $next_version
16:08 murrdoc i guess what i want is, ignore any scheduled jobs and/or jobs created on master
16:08 jalbretsen joined #salt
16:08 Ahlee murrdoc: stop the master?
16:09 murrdoc yeah, but then when i bring it back up
16:09 murrdoc WHEEE
16:09 iggy the scheduler runs on the master, so not a lot of choices there
16:09 Ahlee but if the master's down
16:09 Ahlee there's no jobs going out
16:09 pravka joined #salt
16:09 iggy Ahlee: what? no more 0.17.Ahlee?
16:09 Ahlee iggy: i know, right!?
16:10 murrdoc no way
16:10 murrdoc u upgradeded ?
16:10 Ahlee we poc'd it, our returner changes didn't make it into 2015.2 so we're wiating for next
16:10 Ahlee windows upgrades are a pain in the ass.
16:10 murrdoc basically what i want is the ability to put a minion into no op mode
16:10 meylor joined #salt
16:10 iggy who are we going to harass when someone comes in here with distro packaged 0.17.5 and we want someone to make fun of after we tell them to upgrade?
16:10 murrdoc an ignore all ops mode
16:10 hobakill Ahlee, +1 re: windows updates
16:11 nickdew I have an open pull request for salt-vim, could someone accept it?
16:11 Ahlee i mean, don't get me wrong, linux upgrades are way harder than they should be, too
16:11 iggy mine was easy
16:11 iggy 2014.1 -> 2015.2 anyway
16:12 andreyst joined #salt
16:12 murrdoc http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.schedule.html#salt.modules.schedule.disable does no op for schedule
16:12 murrdoc so jobs dont run even if schedules
16:12 murrdoc Ahlee:  i smell a saltstack chicago meetup
16:12 sastorsl Seems I'm running into issues with the salt-formula because I'm on Centos 6 (and RHEL 6 in production) with regards to python-jinja at level 2.2
16:12 sastorsl Unable to manage file: Jinja syntax error: no test named 'mapping'
16:12 murrdoc 'how i  upgraded to latest and greatest after fixing everything in 0.17'
16:13 sastorsl In CentOS 6 the jinja version is python-2.6.6-52.el6.x86_64
16:13 Ahlee i also still need to figure out why apache is changing application/json to application/x-www-form-urlencoded
16:13 sastorsl python-jinja2-2.2.1-2.el6_5.x86_64 even
16:13 hobakill murrdoc, Ahlee i'm in!
16:14 Ahlee murrdoc: top 10 things I learned at saltconf 15 is currently planned intro presentation
16:14 murrdoc hobakill:  did u sign up on meetup.com yet
16:14 Ahlee i've got engines and beacons for 2, need 8 more
16:14 murrdoc Ahlee:  i can do a … what should your formula look like thing
16:14 murrdoc if u want
16:14 murrdoc engines, beacons, raet
16:14 hobakill murrdoc, yeah i'm on that site..... but didn't know there was a meetup on there.
16:15 bbhoss when building top files, I have some things that need to be done in all my environments. I would put that in my base env dir/top.sls, correct?
16:15 Ahlee but if i talk about raet i need to learn about raet
16:15 Ahlee hobakill: i created it two days ago, should go public today or tomorrow
16:15 hobakill gotcha.
16:15 murrdoc iggy:  u are invited
16:15 murrdoc we can facetime u in
16:15 murrdoc and mute your ass at the same time
16:15 murrdoc for max fun
16:15 JDiPierro joined #salt
16:16 murrdoc Ahlee:  waht about, salt scheduler now does AT jobs
16:17 Ahlee scheduler in general, good one
16:18 Ahlee 20k* nodes, no problem! *if you're just running _modules/ against them
16:18 hasues joined #salt
16:18 iggy sastorsl: pip install jinja ?
16:19 Ahlee huh, cisco CML uses salt for deployment/entitlement
16:19 danlsgiga joined #salt
16:19 hasues left #salt
16:22 dalexander joined #salt
16:23 Ahlee up to 8
16:23 Ahlee need two more
16:23 johanek joined #salt
16:23 danlsgiga Hi all, I'm moving from Puppet to Salt and we use puppet-doc a lot to auto-generate documentation for our puppet modules. Is there anything compared to that for Salt?
16:23 KyleG joined #salt
16:23 KyleG joined #salt
16:23 meylor SheetiS :https://gist.github.com/meylor/8ea4aa38d84dd7bfcc06
16:25 mgw joined #salt
16:26 bhosmer joined #salt
16:26 twobitsprite is there a way to make salt-ssh non-interactive? Every now and then I have minions that don't accept my password (either LDAP timed out, or isn't reachable or whatever) and salt-ssh prompts if I want to push the ssh key and waits for a Y/n response
16:26 mikaelhm joined #salt
16:28 kawa2014 joined #salt
16:29 drawsmcgraw danlsgiga: Wow, good question. If there is, I'm unaware. I'm not terribly familiar with puppet-doc but the best I've done with my Salt work is comments in the state files
16:29 prwilson joined #salt
16:29 khris danlsgiga: once your modules are synced to the minions, you can run sys.doc to output the docstrings. see http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.sysmod.html#salt.modules.sysmod.doc
16:30 clintberry joined #salt
16:31 drawsmcgraw I seem to have misunderstood what puppet-doc does :)
16:31 speedlight the file_roots config si an absolute path?
16:31 speedlight is*
16:32 speedlight so if i want the files in my home dir I need to put /home/user/srv/foo/bar ?
16:32 robawt speedlight: file_roots is an absolute path
16:32 speedlight thnkz
16:33 robawt you may have more than one
16:33 danlsgiga khris: To be really honest I don't know if I actually need it, since everything in Salt is Yaml and Jinja2, there's no DML and so no need to create "modules"
16:33 speedlight and acepts symlinks?
16:33 danlsgiga drawsmcgraw: To be really honest I don't know if I actually need it, since everything in Salt is Yaml and Jinja2, there's no DML and so no need to create "modules"
16:34 drawsmcgraw danlsgiga: Ah, I think I see what you mean now.
16:35 denys joined #salt
16:35 lnr joined #salt
16:35 speedlight or whats the good practice to have the files not in root path ? :)
16:35 danlsgiga drawsmcgraw: Does it make sense?
16:35 lnr joined #salt
16:35 robawt speedlight: keep them backed up ;)
16:35 lnr joined #salt
16:35 speedlight haha
16:36 robawt the best part about a good configuration management tool is it gets out of your way, if you want them somewhere special that's totally fine
16:36 lnr joined #salt
16:36 robawt a
16:36 aparsons joined #salt
16:36 lnr joined #salt
16:36 robawt a lot of folks will create a salt state for their salt master to bootstrap from and set configs + file paths
16:36 drawsmcgraw speedlight: I use Git repos
16:36 lnr joined #salt
16:37 lnr joined #salt
16:37 lnr joined #salt
16:37 drawsmcgraw danlsgiga: Having never written a Puppet module before, I have to use my imagination a bit. But it sounds like because Salt uses more 'standard' vehicles (YAML, Jinja2...), it obviates the need to generate documentation for states.
16:37 lnr joined #salt
16:38 lnr joined #salt
16:38 Fiber^ joined #salt
16:38 lnr joined #salt
16:38 lnr joined #salt
16:38 XenophonF danlsgiga: salt doesn't really supply a documentation generator
16:39 speedlight drawsmcgraw: so u push from the user and then pull from root?
16:39 lnr joined #salt
16:39 lnr joined #salt
16:39 drawsmcgraw speedlight: I use the GitFS fileserver backend
16:39 lnr joined #salt
16:39 drawsmcgraw speedlight: http://docs.saltstack.com/en/latest/topics/tutorials/gitfs.html
16:39 speedlight was reading about, dont fully understand yet
16:40 drawsmcgraw It's the way to go, in my opinion.
16:40 lnr joined #salt
16:40 danlsgiga drawsmcgraw: You are correct! The salt states and pillars are self explanatory (obviously if you comment the code), while puppet modules needs to be documented because you have the need to understand what the module accepts has parameters and what each module does.
16:40 XenophonF danlsgiga: maybe you can use doxygen or something similar alongside salt?
16:40 lnr joined #salt
16:40 drawsmcgraw danlsgiga: Ah, I see. Puppet modules are no small matter
16:40 lnr joined #salt
16:41 __gotcha joined #salt
16:41 lnr joined #salt
16:41 drawsmcgraw speedlight: I treat my Salt states like a software project. I branch off of 'master', develop states, test them, and merge them back in.
16:41 danlsgiga drawsmcgraw: Yep... I'm just realizing that! =P
16:41 lnr joined #salt
16:41 drawsmcgraw Once they're merged, I push them up to the Git server (say, Github), and the master automatically pulls down the new states
16:41 lnr joined #salt
16:41 XenophonF speedlight, drawsmcgraw: i'm using a similar workflow here
16:42 XenophonF only the master branch is empty except for top.sls
16:42 speedlight oh so the git backend have a like a push event hook
16:42 XenophonF speedlight: the salt master does a git pull automatically
16:43 danlsgiga drawsmcgraw: I'm reading the gitfs conversation... one more question... does salt accept s3 as a backend fs? If it does, salt is a great candidate to be a container =)
16:43 drawsmcgraw default is every 60 seconds but it can be adjusted
16:43 drawsmcgraw yes! Yes it does danlsgiga
16:43 speedlight yes, from the begining want to have that king of workflow, many things here work with git flow :D
16:43 XenophonF drawsmcgraw: have you gotten s3fs to work?
16:43 drawsmcgraw It's a game-changer for me, especially for serving up huge tarballs, etc
16:43 XenophonF because i haven't
16:43 drawsmcgraw I did a while back. I remember it being a little fiesty
16:43 danlsgiga drawsmcgraw: Awesome!
16:43 XenophonF i've had to use s3:// URLs in the places where I need to deploy large files
16:44 XenophonF as opposed to using salt:// URLs that resolve to an S3 bucket
16:44 speedlight i use gitlab, i guess theres no difference for salt
16:44 danlsgiga We use s3 a lot at the company. This is a huge win for the project! =D
16:44 drawsmcgraw XenophonF: Yeah, pay special attention to how the docs say to set up your buckets.
16:44 prwilson joined #salt
16:44 drawsmcgraw It's possible you're getting hung up on that. There are two different way to describe them.
16:44 spookah joined #salt
16:44 drawsmcgraw I don't remember the details but I *do* remember getting hung up on that part
16:44 XenophonF drawsmcgraw: would you mind posting a complete (but sanitized) example?
16:44 hal58th danlsgiga this is the file server backends that Salt supports. http://docs.saltstack.com/en/latest/ref/file_server/all/
16:45 drawsmcgraw Here we are: http://docs.saltstack.com/en/latest/ref/file_server/all/salt.fileserver.s3fs.html
16:45 drawsmcgraw XenophonF: I wish I could. I've since left the environment where I successfully wired it up :/
16:45 Anna_ joined #salt
16:45 Guest23224 joined #salt
16:46 XenophonF drawsmcgraw: ah too bad.  i did rtfm but couldn't get it to work
16:46 drawsmcgraw It's not the easiest..
16:46 XenophonF drawsmcgraw: i had the right s3.key etc. settings in the master config, and i had the right bucket config, or at least it matched what was in the docs
16:46 drawsmcgraw If you want to give it another try and tell us the errors you're seeing, I'm sure the room can get you set up
16:46 XenophonF drawsmcgraw: that's the thing - i didn't see _any_ indication that s3fs was even active
16:47 danlsgiga drawsmcgraw: Great!! Thanks!!
16:47 drawsmcgraw XenophonF: I would start salt-master in the foreground with '-l debug' and watch
16:47 drawsmcgraw I remember seeing it reach out to S3 and building the cache
16:47 XenophonF drawsmcgraw: roger that - i'll have to cons up a dev instance of my master to mess with this
16:48 drawsmcgraw There's also the 's3' modules to interactively (attempt to) fetch objects from S3
16:48 drawsmcgraw module, not modules
16:48 khris joined #salt
16:48 XenophonF drawsmcgraw: those work fine when i call them via salt-call (i.e., the minion executes it)
16:48 drawsmcgraw XenophonF: heh. Yep. I remember all this pain now. I'm sorry I don't have a better answer for now
16:48 XenophonF like s3.get and stuff work fine
16:49 drawsmcgraw right....
16:49 XenophonF as do s3:// URLs in the state
16:49 XenophonF er, states
16:49 drawsmcgraw right. But salt://location URNs don't seem to be searching the S3 bucket(s) ?
16:49 XenophonF exactly
16:49 XenophonF git or file system only
16:50 drawsmcgraw Brutal. But yeah, try it again, start the salt-master with -l debug and watch to see if it's trying to sync with S3
16:50 XenophonF it's highly likely i was doing it rong
16:50 drawsmcgraw It's easy to mess up, for sure.
16:51 sieve joined #salt
16:51 XenophonF if i can get it to work, i'm going to set up syndic and multiple masters in different AWS regions
16:52 itru joined #salt
16:52 XenophonF it will be glorious :)
16:52 drawsmcgraw XenophonF: fancy :D
16:54 XenophonF my boss will likely complain that it's over-engineered =)
16:55 amcorreia_ joined #salt
16:58 speedlight from a newbie point of view XD is better pygit2, gitpython...etc?
16:59 numkem whats the order that the pillar data is compiled? one line after the other in the order that they are specified in the top file?
16:59 n8n joined #salt
16:59 speedlight or the "advantages" is more for api devs, advances customization and so on?
17:00 pipps joined #salt
17:00 ek6 speedlight: pygit2 still a pita to install on ubuntu,  if you get past that its the most full featured...if its easy to install on your os id go with that
17:00 hobakill speedlight, i've had the most luck with gitpython but i have no real hard evidence to show that.
17:01 speedlight i have gitpython in the repo but no pygit.. so i guess i'll go with gitpython
17:01 johanek pygit2 supports more config options around git auth: http://docs.saltstack.com/en/latest/topics/tutorials/gitfs.html#per-remote-configuration-parameters
17:03 Ryan_Lane joined #salt
17:04 nesv joined #salt
17:04 iggy I really need to push my gitfs modules to the salt-formula
17:12 jespada joined #salt
17:12 _ale1_ joined #salt
17:15 ndrei joined #salt
17:17 murrdoc dew it
17:17 nesv Does the grains.append state not create a grain if it doesn't exist?
17:19 hal58th "Append a value to a list in the grains config file. If the grain doesn't exist, the grain key is added and the value is appended to the new grain as a list item." http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.grains.html
17:19 iggy I would think list_present would take care of that
17:19 hal58th nesv
17:19 joshfu joined #salt
17:20 nesv hal58th: Thanks, but that's for a module.
17:20 nesv iggy: Thanks, I'll give that a try.
17:20 nesv Please forgive my dumb questions; I'm still learning Salt, and it's ways.
17:21 iggy fwiw, salt.states.grains.append calls salt.modules.grains.append internally
17:22 iggy so the docs for the grains.append module apply here
17:22 nesv iggy: Ah, okay. Thank you. :)
17:25 isbm joined #salt
17:26 isbm hi
17:26 isbm anyone dev around?
17:27 germs_ joined #salt
17:27 iggy we don't have devs... only a QA dept
17:27 germs_1 joined #salt
17:27 chandy joined #salt
17:27 evle3 joined #salt
17:27 iggy (I kid... I'm not associated with SaltStack in any official capacity)
17:28 iggy if I was, I'd be too busy to spend my days in irc
17:28 nebo joined #salt
17:29 murrdoc is he right!
17:30 isbm OK :/
17:30 hal58th why do you ask isbm?
17:30 murrdoc if u have questions hal58th can solve em
17:30 isbm I am looking how I can get data from minion to master
17:30 murrdoc hes really good
17:31 hal58th Don't need a dev for that. Look at "salt mine".
17:32 evle joined #salt
17:32 isbm hal58th: no, I mean how it is implemented :)
17:32 hal58th also depends on what kind of data you need to get from minion to master
17:32 hal58th ah
17:32 isbm file
17:32 isbm I want to take file from client to the master, for example
17:32 iggy fileserver.py and fileclient.py
17:32 nebo hey basepi you ve got a spam bot called Anna_ on your chat
17:33 kawa2014 joined #salt
17:33 basepi "on your chat"... wher?
17:33 basepi where?
17:34 basepi Was Anna_ spamming or something?
17:34 rgarcia_ joined #salt
17:35 iggy yes
17:35 isbm hal58th: Also, can Salt ensure only one line in the config?
17:35 iggy isbm: yes
17:36 isbm iggy: where to RTFM that? :)
17:36 basepi Can anyone give me a timestamp on the spamming so I can verify?
17:36 isbm basepi: Anna_ bot is spamming privately :(
17:36 basepi Ah, well.....nothing I can do about that.
17:36 gladiatr joined #salt
17:37 basepi Does Freenode have a reporting mechanism?
17:37 iggy you can boot them out of here
17:37 fredvd joined #salt
17:37 Anna_ was kicked by basepi: Anna_
17:37 iggy so they don't see when people come and go from this channel
17:37 Gareth Corey: ping
17:37 basepi valid.
17:37 iggy <3 base of pi
17:37 basepi =)
17:37 isbm iggy: do you know an example how exactly to ensure config line?
17:37 cpowell joined #salt
17:38 iggy file.managed contents
17:38 murrdoc based pi!
17:38 gladiatr left #salt
17:39 gladiatr joined #salt
17:39 Guest70 joined #salt
17:39 _ikke_ joined #salt
17:39 isbm iggy: ??
17:40 isbm iggy: It will use string as the whole content of the file?
17:40 iggy the file.managed state module has an option called contents that will make the contents of said file be the entirety of that string
17:40 isbm iggy: Nooooo...
17:41 isbm iggy: I am looking for this: I want to change e.g. in httpd.conf only port. And let Salt takes care ONLY that, while client can manually change something else.
17:41 isbm iggy: e.g. cfengine can manage only one line in the configuration file without touching the rest. It even will leave my comments there untouched.
17:42 isbm iggy: I cannot so far find something similar in Salt. Any help?
17:42 iggy file.replace? yes... it's possible... there are likely quite a few different ways to handle it
17:42 iggy pick whichever one is best for you
17:42 isbm it is *replace*
17:42 isbm If string is not there, it will fail.
17:42 isbm What I am looking is that if no string, it is added, if different value, it is fixed.
17:43 isbm How it is typically done in Salt?
17:43 isbm (before I opened a new PR :D _
17:43 isbm )
17:43 iggy typically, you lay down a custom file with file.managed and template the values you need... you control the custom file, so you know it has the Port line in it
17:44 iggy you're not going to get much traction on a PR for such a feature
17:44 isbm iggy: and if I am allowed to manage ONLY that line?
17:44 iggy there's no generic way to do that
17:44 iggy there are hundreds (if not thousands) of different config file formats that salt would have to take into account for something like that
17:45 Ahlee augeas or similar is better spent for something like that
17:45 otter768 joined #salt
17:45 iggy or quit half-assing your config mgmt
17:45 iggy you either manage your config files or you don't...
17:46 Ahlee there's always file.sed or file.blockreplace, etc
17:46 petong what about file.append
17:46 murrdoc file.accumulated
17:46 Ahlee append wouldn't work with apache config
17:46 isbm iggy: https://docs.cfengine.com/docs/3.5/reference-promise-types-files-edit_line.html
17:46 isbm I want this in Salt.
17:46 mbrandeis or http include file statement
17:46 iggy I'm not reading cfengine docs
17:46 hal58th the key here is he wants to edit a line if it exists, otherwise append the line. Trying to figure out how to do that with requisites
17:46 iggy I've had a bad enough week without that shit
17:46 Ahlee Well, you shouldn't, but http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.blockreplace
17:47 isbm hal58th: yep. Exactly.
17:47 Corey Gareth: pong
17:47 isbm hal58th: Folks hates our YaST (SUSE) in past, because it was killing all manual changes :)
17:48 iggy my original point stands... there's no generic way to do this... salt would have to understand config file formats... it's not a simple matter of trying replace and falling back to append
17:48 isbm hal58th: And we fixed that. Now, if we offer the same again, I kind of foresee the reaction.
17:48 andreyst1 joined #salt
17:49 isbm iggy: We doing it in YaST and cfengine can. There is no problem how to do it, question is: can Salt do it at the moment or not? :)
17:49 hal58th unfortunately that is the one and ONLY thing I missed about cfengine. It is currently not possible elegantly in Salt.
17:49 isbm hal58th: Ah, then OK. I am preparing PR for that. :)
17:50 JDiPierro joined #salt
17:50 isbm hopefully that may come to 2015.2 if will work right.
17:50 hal58th Makes me think there should be a requisite for "nochanges". That would solve this.
17:50 iggy you missed the boat for 2015.2... by like 2 months
17:50 johanek managing part of your config with CM and part manually seems like a bad place to be :(
17:51 isbm iggy: why?
17:51 hal58th 2015.2 is feature complete.
17:51 iggy it was branched from dev in Decemeber... i.e. no new features
17:51 isbm hal58th: Do you have any thoughts how it could work better?
17:51 isbm iggy: I've just pushed yesterday few more features there. :)
17:52 isbm iggy: maybe it was a time machine...
17:52 evle joined #salt
17:52 iggy like what?
17:52 iggy because I will raise hell
17:53 iggy if they are adding features instead of getting that month late release out the door, they all need to go back to CS 101
17:53 _ikke_ joined #salt
17:53 hal58th If you had a state that did a file.sed and had a requisite for "nochange" (opposite of "onchange"), then you could do a file.append or whatever.
17:54 hal58th you would just have to make sure not to get in an infinite loop. But that requisite doesn't exist
17:54 isbm iggy: Sorry, my part was only open the PR, test it and back port it.
17:54 danlsgiga iggy: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.augeas.html
17:54 amcorreia joined #salt
17:55 druonysus joined #salt
17:55 mpanetta You know, it might be easier if you just left the configuration to salt and just file.managed the whole thing...
17:55 hal58th back to work for me. IRC crack is to stronk
17:55 iggy yeah, I'm out this bitch too... see you guys Monday
17:57 LtLefse isbm: have you looked at the append_if_not_found and not_found_content parameters on file.replace?
17:58 edrocks joined #salt
17:58 isbm LtLefse: and if I need to insert after specific line? (e.g. config has { ... } ..... { ... } or is an ini file etc)
17:59 auser joined #salt
17:59 auser Hola
18:00 mbrandeis isbm, the augeas state package looks like it handles line changes and insertion after elements
18:01 isbm mbrandeis: I did not yet look at that. Does it need the entire augeas installed on the client?
18:03 mwpher joined #salt
18:03 mbrandeis i haven't used it. it would need some kind of augeas module installed on the client. not sure if augeas just has an api/dev install. they have an api. just not sure what salt requires, and how augeas breaks up its api vs client
18:05 johanek joined #salt
18:09 mwpher I'm trying to see the Jinja environment by putting show_full_context() in a template, but when I try to execute it, I get the error 'unable to determine the upstream hash of source file'
18:10 TheoSLC joined #salt
18:11 mwpher Any ideas what that could mean? I've never seen that error with any other file.
18:12 mwpher I also see that error when using 'contents' instead of 'source'
18:13 baweaver joined #salt
18:16 Auroch joined #salt
18:16 sinh joined #salt
18:18 mbrgm left #salt
18:19 nitti joined #salt
18:20 netzmonster joined #salt
18:21 joshfu joined #salt
18:21 mwpher Any ideas what an error like that means? I can't see any hints in the logs either
18:21 mwpher This is 2014.7.2 on FreeBSD 10.1, by the way
18:22 timoguin joined #salt
18:23 mwpher ... Well, serverfault it is I guess
18:24 iggy I would give more info there
18:24 mwpher What other details do you recommend?
18:24 iggy the state at the very least
18:25 iggy but traditionally, "salt --versions" is also a good go to
18:25 mwpher The state? It's just a file.managed line that makes a text file with show_full_context() in it
18:26 iggy that's not any clearer tbh
18:26 mwpher Ok
18:26 iggy but I've never messed with show_full_context
18:26 aron_kexp joined #salt
18:29 mwpher Are there any salt forums?
18:29 mwpher Or someplace equivalent
18:29 iggy mailing list
18:29 mwpher Ok
18:29 mwpher I'll try that
18:29 mwpher Thank you
18:30 amcorreia joined #salt
18:32 baweaver joined #salt
18:32 timoguin joined #salt
18:33 ndrei joined #salt
18:40 meylor I'm trying to mine a grain that is set and i'm not getting anything out? Can someone assist
18:40 aron_kexp joined #salt
18:41 iggy paste stuff
18:41 meylor Does the ability to mine need to be set in the master?
18:42 meylor iggy https://gist.github.com/meylor/8ea4aa38d84dd7bfcc06
18:42 denys joined #salt
18:42 mgw joined #salt
18:43 iggy that's not how mine_functions work
18:44 iggy it's a _function_
18:46 chandy joined #salt
18:46 aron_kexp joined #salt
18:46 iggy so do grains.items and access the data through the entire set of grains or use one of the network modules instead
18:47 meylor iggy: re-reading the mine topic. I basically want to query all of the nodes that have a grain set and make that available to salt. Wouldnt' that be in a mine
18:47 wincyj joined #salt
18:47 bhosmer__ joined #salt
18:48 iggy you're almost there, you've just got a few small issues
18:49 thayne joined #salt
18:51 eclectic joined #salt
18:54 iggy https://gist.github.com/iggy/c785f73e055bb8ead30e
18:54 iggy meylor: ^
18:55 Guest70 joined #salt
18:56 meylor okay. dumb question. say i were to set that, just like you have. are you expected to bounce salt-master to be able to use them (when do they get applied)
18:56 johanek joined #salt
18:56 neogenix_ joined #salt
18:56 TheoSLC joined #salt
18:56 meylor iggy: ^^^
18:56 iggy that's in a pillar, so you don't have to bounce the master
18:56 amcorreia joined #salt
18:57 iggy but the default mine update time is pretty long (an hour maybe?)
18:57 meylor there's no way to force that?
18:57 meylor as in, force an update
18:57 spookah joined #salt
18:59 mgw1 joined #salt
19:00 __gotcha joined #salt
19:00 iggy mine.update
19:01 iggy but it's never really worked for me
19:02 teskew joined #salt
19:05 jtang joined #salt
19:06 murrdoc restart all servers
19:06 murrdoc works everytime
19:06 murrdoc (please dont)
19:06 meylor iggy thanks!
19:07 meylor set my pillar just like yours.. so mine.update also didn't work for me. but i did kill the node and brought it back up and i see the mine now
19:07 ajw0100 joined #salt
19:11 pipps joined #salt
19:12 mwpher joined #salt
19:13 TheoSLC joined #salt
19:16 dendazen in this case
19:16 dendazen <% syslog_config = pillar.get('syslog_ng', {}) %>\
19:17 krelo joined #salt
19:17 dendazen where pillar syslog_ng should be
19:17 dendazen in the file structure
19:18 _JZ_ joined #salt
19:18 numkem for those who are like me and are using Foreman, I've create an external_pillar that looks through folders depending on organization and roles
19:18 numkem https://gist.github.com/numkem/75706423e8f0063278f0
19:22 iggy dendazen: I never really got what you were trying to figure out
19:23 dendazen oh
19:24 dendazen i am looking at this guy https://github.com/saltstack-formulas/syslog-ng-formula/tree/master/syslog_ng
19:24 dendazen and there is ‘pillar.example’ file
19:24 dendazen so is this file would be as an input here? <% syslog_config = pillar.get('syslog_ng', {}) %>\
19:25 dendazen that line is from files/syslog-ng.conf mako template
19:26 baweaver joined #salt
19:26 iggy mako template?
19:26 giantlock joined #salt
19:27 dendazen well, what is this syntax?
19:27 dendazen https://github.com/saltstack-formulas/syslog-ng-formula/blob/master/syslog_ng/files/syslog-ng.conf
19:29 joshfu joined #salt
19:29 dingo yes mako https://github.com/saltstack-formulas/syslog-ng-formula/issues/2
19:30 fusionx86 joined #salt
19:31 fusionx8_ joined #salt
19:31 mikaelhm joined #salt
19:32 iggy use something else (unless you just really want to install make on all your minions for no apparent reason)
19:32 dendazen i do not.
19:32 dendazen i can just i guess use jinja
19:32 dendazen that is used by default
19:33 dendazen oh looked at that bug
19:33 dendazen that is fine, i will try to convert to jinja
19:35 iggy saltconf videos out?
19:35 khris dendazen: does syslog-ng finally allow includes ala conf.d/*
19:36 theologian joined #salt
19:37 iggy EMCs talk is up
19:37 iggy but not all of them yet
19:38 khris https://www.youtube.com/results?search_query=saltconf15
19:39 iggy Ryan Lane's talk about sequentially ordered execution was good, and I heard lots of stuff about EMCs talk
19:39 ndrei joined #salt
19:39 ckao joined #salt
19:39 iggy *good stuff
19:40 amcorreia joined #salt
19:41 dendazen yeah it does
19:41 dendazen you can create /etc/syslog-ng/conf.d
19:42 dendazen you will have to add ‘include’ line before log defintions in syslog-ng.conf
19:42 khris that's fortunate. the lack thereof fueled my transition to rsyslog once upon a time. glad to hear that's an option now
19:44 baweaver joined #salt
19:46 otter768 joined #salt
19:47 chandy joined #salt
19:50 nbari joined #salt
19:50 nbari joined #salt
19:50 baweaver joined #salt
19:51 mikepea_ joined #salt
19:52 baweaver If one executes a lowstate through the API, is there a reliable way to get the exit status back from it?
19:53 murrdoc left #salt
19:53 goki joined #salt
19:57 supermike_ joined #salt
19:58 nbari one question relatd to mines, if I want to get the IPs from only certain minions, on the pillars/top.sls I need to create a 'node_type:web' ?
19:58 nbari I mean is the way salt undestand to apply only the mine to that group?
19:58 iggy you could do it that way
19:59 iggy but you can also just have all minions report back the same mine data and then use the mine tgt_type to narrow the search results (simpler imo)
19:59 jtang1 joined #salt
20:00 nbari so basically i can later just filter in jinja using something like node_type, 'glob' etc right ?
20:00 prwilson joined #salt
20:00 iggy {% set dbhost = salt['mine.get']('G@tags:db and G@tags:primary', 'network.interfaces', 'compound') %}
20:01 iggy you could use just grain matches (or globbing or any other matches that salt supports)
20:01 nbari ok
20:02 nbari for testin I change the IP of a minion, but some one I think I have a cache since my conf did not change, (following the haproxy example)
20:03 akitada joined #salt
20:03 nbari I did an mine.delete/mine.flush but no working
20:05 iggy the mine only updates by default every hour (iirc)
20:06 abele joined #salt
20:06 ek6 ok..so its friday and so ive gone off the deep end...but even using arch I cant figure out an easy way  to install 2015.2.rc1 with raet
20:07 krelo joined #salt
20:07 ek6 i can install 2014.7 with raet easily enough...but i dont want that..   no bootstrap support and hand makepkg salt-git gives me zmq
20:08 relidy Question about requirements: If I have multiple states that require a particular group to exist on the system, do I have to "include" my groups state in each of them in order to require the groups to be set up first, or can I reference the dependency more directly?
20:10 hal58th no, you just have to require the state id. Salt will build a list of all the states when it runs
20:10 iggy bootstrap supports installing packages and installing salt from git... should be good enough
20:10 hal58th i think if i read you right...
20:10 nbari iggy: anyd idea of what could I be doing wrong: http://pastebin.com/mdLYmsmK
20:10 pf_moore joined #salt
20:10 iggy using pastebin.com
20:11 nbari I just tried to pug globally my mine but can't get any output
20:11 nbari :) besides patebin
20:11 relidy hal58th: So, my require statement would have a key that's just the state ID?
20:11 hal58th relidy, you need to have a state that makes sure group exists, then any state can do a require on that state id.
20:11 hal58th yes.
20:11 nbari salt '*' mine.get '*' network.ip_addrs
20:11 nbari I am using that just for testing but I get no output
20:11 relidy hal58th: Thanks.
20:12 iggy what is mine_interval there? I don't think you can do that
20:12 CaptTofu_ joined #salt
20:12 hal58th relidy:   require:  /n   - group: groupname
20:12 hal58th *sorry not groupname, stateid is what i meant
20:12 relidy hal58th: Ah, that's the magic I was missing. Thank you.
20:13 iggy basepi: did you ever hear anything about /etc/salt/minion.d/_schedule.conf ?
20:13 nbari every time I alter something on the pillars I should salt '*' saltutil.refresh_pillar
20:13 nbari ?
20:14 basepi iggy: I totally spaced it. It's been a busy week trying to get 2014.7.3 and 2015.2.0rc2 out the door
20:14 ek6 iggy: no bootstrap support for raet whatsoever that i can see
20:14 cmcmacken joined #salt
20:14 doriftoshoes joined #salt
20:14 basepi iggy: I'm not in office today, but I'll find out more on Monday, I just set myself a reminder.
20:14 iggy basepi: no worries, just curious if you'd heard anything (in case I get bored over the weekend and feel like working on formulas... unlikely)
20:15 basepi hehe
20:15 iggy ek6: yeah... I don't think they really want people using raet without having to put in a little effort
20:16 MaZ- joined #salt
20:16 ek6 iggy: im ok with a little effort...im even putting it up on arch (which i dont typically run)  just ASSuMEd that if i was going to try it id do it on something close to trunk and not 2014.7
20:17 imanc_ joined #salt
20:17 Jouke left #salt
20:17 oxae joined #salt
20:18 SheetiS 2014.7.3 going to be out before 2014.7.2 gets out of epel-testing, heh. <3 epel...
20:18 iggy yeah, so install raet by hand and install salt via bootstrap... BAM!
20:19 SheetiS I'm going to just start packing my own rpms for salt.  I'm already managing a local yumrepo, so might as well.
20:20 iggy DO IT!!!!!
20:20 thayne joined #salt
20:20 iggy that seems like what salt people want to tell me every time I piss and moan about it taking 2 weeks from release tag to tarball available
20:21 baweaver joined #salt
20:24 SheetiS hehe
20:26 chandy joined #salt
20:29 baweaver joined #salt
20:30 ek6 iggy: these the same salt people that are explaining to you thats its feb 48th?
20:31 iggy no, everybody seems to feel genuinely bad about that particular hold-up
20:32 iggy you could see it in their dead little eyes every time it was brought up at saltconf... and that was when it was just 4 days late
20:33 amatas joined #salt
20:33 Guest70 joined #salt
20:34 hal58th joined #salt
20:36 ek6 iggy: aye..
20:38 iggy hooray, collectd-formula groundwork laid
20:38 iggy layed?
20:39 iggy nfc
20:43 nbari I am trying a simple mine with test.ping: [] but from while running salt '*' mine.get test.ping got no reply
20:43 nbari how to force/clean previous mines ?
20:44 relidy In regards to my previous group requirement question, that solution (require:\n-group:<name>) only seems to work if the state that declared the group(s) has been included first. Is there a way to make salt aware that it is managing the group without including the entire group state first? Or, is there harm in including the entire group state all over the place?
20:45 itru what if I want to “watch: file” but I don’t want to manage that watched file? is it possilble?
20:46 matthew-parlette joined #salt
20:47 SheetiS relidy: Includes to allow for requirements to work is actually the correct way to handle that situation.
20:47 nexsja^ joined #salt
20:47 itru can I have just
20:47 itru file
20:47 relidy SheetiS: Okay, that sounds reasonable. I just wanted to make sure I wasn't making more work for the system or for myself by including that stuff.
20:48 __gotcha joined #salt
20:48 CheKoLyN joined #salt
20:49 ndrei joined #salt
20:50 JDiPierro joined #salt
20:51 Ixan will any module called by ext_pillar be called only on the master or on every individual minion, when requesting pillar data?
20:51 SheetiS Ixan: anything run from an ext_pillar will be executed on the master.
20:52 Ixan okay, good. The docs are a bit hazy on that particular execution
20:52 Ixan thanks, SheetiS
20:54 SheetiS itru: there'd be no way to know if things are changed. you can do file.exists if you want to make suer it is in place for a dependency, but watching for changes on an unmanaged resource isn't possible becase the way salt determines changes are changes that it wants to apply.
20:55 itru yep, used file.exists, it works, thanks!
20:55 speedlight haha im so confused with roles and pillars XD
20:55 iggy relidy: salt is smart enough not to rerun the included states repeatedly
20:55 speedlight i guess is a hard learning curve for a noob like me :D
20:56 iggy nbari: it's salt '*' mine.get '*' test.ping
20:56 relidy Thanks, iggy. I assumed as much, but you know that they say about that.
20:57 iggy itru: not yet... 2015.2 has inotify beacons that will probably do everything you need it to do in that situation (although it uses reactor and not watch'es on states)
20:57 SheetiS speedlight: 'roles' are concept that would be more human logic, where a pillar is a dictionary of data to be used by states.  What are you trying to accomplish?  It's usually easiest to point someone in a direction when you share an idea of where you want to go. :)
20:58 SheetiS iggy: so many things I'm looking forward to in 2015.2.
20:59 Guest92235 joined #salt
21:01 baweaver joined #salt
21:01 nbari iggy: thanks, test.ping working but I can't get the IP's, how to test from the command line specifying a CIDR ?
21:01 itru my head blows up with current set of features, please have mercy!
21:03 iggy nbari: you passed the cidr as part of the mine function... so it's past when the minion loads that data in the mine (not when you query it)
21:04 nesv I'm trying to define some mine_functions in a pillar, but whenever I try to call them, I don't get anything.
21:04 nesv ...any results.
21:04 iggy so in your previous paste it'd just be "salt '*' mine.get '*' network.ip_addrs"
21:04 nbari salt '*' mine.get '*' network.ip_addrs(10.1.0.0/15) ?
21:04 nesv If I manually put the mine_functions configuration in place, they work.
21:04 nbari some times work others not, I have no idea what I may have wrong
21:05 speedlight SheetiS: letme clear my mind first cuz reading and trying all these cause some cerebral paralisis XDD
21:06 speedlight SheetiS: im starting from scrath, but now if have a problem :/
21:06 speedlight No Top file or external nodes data matches found
21:06 speedlight running state.highstate
21:06 auser joined #salt
21:07 ndrei joined #salt
21:08 SheetiS speedlight: do you have a top.sls defined (in both your file_root and your pillar_root)?
21:09 speedlight i change the file_root base to /srv/states and inside of these i have the top.sls
21:09 kermit joined #salt
21:09 ajw0100 joined #salt
21:09 speedlight now dont care about pillars, the top.sls have:
21:09 speedlight base:
21:09 speedlight '*':
21:09 speedlight - app.pacdeps
21:09 speedlight i hv the app folder and a pacdeps.sls file
21:10 iggy that's not proper yaml
21:10 speedlight with the simplest thing of httpd pkg installed
21:10 iggy the 3rd line should be indented more
21:10 speedlight oh
21:11 speedlight the - needs to be under the : level of the 2nd line?
21:11 speedlight i putit like this
21:11 speedlight http://docs.saltstack.com/en/latest/topics/tutorials/states_pt1.html#preparing-the-top-file
21:12 nbari now is working, I just rename /pillars/mine to /pillars/haproxy, the stupid hame was causing conflicts
21:14 nbari simm that the dirnames have to follow the python way when chossing hames
21:16 sijis left #salt
21:17 amcorreia_ joined #salt
21:17 bash124512 joined #salt
21:20 JDiPierro joined #salt
21:20 edrocks joined #salt
21:21 edrocks how do you clear the salt masters minion public key cache? I got this error after updating my minions "The Salt Master has cached the public key for this node, this salt minion will wait for 10 seconds be fore attempting to re-authenticate"
21:22 rsimpkins joined #salt
21:23 drawsmcgraw left #salt
21:26 dendazen joined #salt
21:27 hal58th edrocks not sure what happened. What did you do?
21:27 hal58th I just want to give you best solution edrocks
21:27 edrocks I ran "salt '*' cmd.run 'apt-get upgrade -y' to update everything then my minions lost connection
21:28 edrocks I don't think it updated them but it took them offline
21:28 edrocks so I connected to run an ran it manually and it asked to use the provided salt-minion config or keep mine and I updated it but I think that broke it
21:29 hal58th you updated it to use the apt config?
21:29 edrocks yes which broke my config I'm going to replace it with the old one
21:30 jonher937 You should have your own config in /etc/salt/minion.d/myminion.conf and only include master + minionid
21:30 jonher937 edrocks: this way you will be able to upgrade without breaking the config
21:30 iggy python one liner to read mine structure (nbari nesv etc... this might help): python -c "import msgpack; f = open('/var/cache/salt/master/minions/<minion_id>/mine.p', 'rb'); print msgpack.unpackb(f.read())"
21:31 fogrito joined #salt
21:31 hal58th edrocks, if it generated a new key for the salt minion, you will have to re-add it
21:31 edrocks i just fixed it
21:32 edrocks its because the new config had the default id of my hostname but I created my minions with salt-cloud on gce so the key id was wrong
21:32 hal58th but yes, jonher937 idea is a really, really good suggestion
21:32 hal58th ah I see
21:32 edrocks I'll have to get it setup how jonher937 said so I can do updates safly
21:33 hal58th yeah, and do one minion config option per file. I have one for master, schedule, etc
21:33 edrocks I think I just need to manually restart the other minions now since their config didn't change they just dont show up in test.ping
21:33 edrocks do you use file.managed or something similar to sync the config to each minion?
21:33 nesv iggy: That just gives me "{}"
21:34 edrocks hal58th: ^
21:34 [7hunderbird] joined #salt
21:36 refnode_ joined #salt
21:36 edrocks well I have to go the want to lock up. Thanks for all the help
21:36 iggy nesv: that means you don't have any data in the mine
21:36 iggy (for that host)
21:37 iggy I should have mentioned, replace <minion_id> with an actual minion_id from salt-key -L
21:37 nesv iggy: Okay, I figured as much. I'm just lost as to why I can get mine data for a host when I put the mine_functions into a config file, but not when I put it into a pillar.
21:38 iggy because when you put it in the config file, you have to restart the minion for it to take effect (and mine update is something that happens on minion start)
21:38 iggy otherwise you have to wait for mine_interval
21:39 teskew joined #salt
21:40 nesv Yes, that I understand. Even after setting the mine_interval to 5, and waiting a half-hour, the pillar-based mine_functions never return any data
21:41 iggy *shrug* works for me
21:41 iggy make sure mine_functions shows up in pillar.items output for the minions
21:43 nesv iggy: Which it does
21:43 nesv Je suis is feeling defeated by this stack of salt
21:47 otter768 joined #salt
21:49 iggy mine isn't usually the first thing people jump into when new to salt
21:51 baweaver joined #salt
21:56 danlsgiga left #salt
22:02 thayne joined #salt
22:03 Guest70 joined #salt
22:12 beneggett joined #salt
22:14 aquassaut joined #salt
22:15 JDiPierro joined #salt
22:22 pravka joined #salt
22:25 yomilk joined #salt
22:28 baweaver joined #salt
22:29 YogeeBear joined #salt
22:31 YogeeBear left #salt
22:37 iromli joined #salt
22:45 thayne_ joined #salt
22:45 ajw0100_ joined #salt
22:49 fogrito The minion config file at the top states that values that are commented out but have no space after the comment are default that need not be set in the config. However, the "test: True" comment has no space after the comment, yet my minion doesn't operate in test=True mode by default. Why is this?
22:50 scbunn_ joined #salt
22:54 iggy file a bug report
22:54 fogrito Will do
22:55 MindDrive *sigh* So since I never got a definitive answer last time, want this: <non-root user on a system running salt-call/salt.client.Caller()> -> <Salt master running as root> -> <Salt minions running as root>
22:55 MindDrive Need to find a secure way to do the first part.
22:55 MindDrive My attempts up to this point have failed.
22:56 cmcmacken joined #salt
22:57 mikaelhm joined #salt
23:00 HoloIRCUser2 joined #salt
23:04 SheetiS MindDrive: With salt-call you basically open the whole root 'can-o-worms' to anyone who has access.  What are you trying to secure (and in what way) with that?
23:04 Vynce joined #salt
23:05 MindDrive The ability to run a very specific module on the minions which does package management (yum in this case).
23:05 MindDrive And yes, I know this can be dangerous, which is why I asked if it could be done securely.
23:06 Vynce heh.  i was just wndering over as i begin research on getting salt to apt-get upgrade all my minions. ( :
23:07 SheetiS if it were a static command (didn't have to have different parameters for different packages etc), you could put it in a script and then do a custom sudoers to allow execution of just that script.
23:07 JDiPierro joined #salt
23:07 SheetiS I don't think there is a concept of client_acl (http://docs.saltstack.com/en/latest/ref/clientacl.html) when doing salt-call from the minion
23:07 dendazen joined #salt
23:07 MindDrive sudo will not allow me to handle this programmatically, which was the main point/hope.
23:08 ecdhe Mindrive, sudo doesn't have to ask for a password.
23:08 ecdhe You can configure it to not require a password for certain commands.
23:08 ecdhe And for certain users running certain commands... from certain terminals.
23:08 SheetiS nopassword and even !Requiretty if needed
23:08 MindDrive ecdhe: I can't call salt.client.Caller() if I require sudo.  I have to call out to the command line.
23:09 iggy it can't be done
23:09 MindDrive That's sad, because MCollective can do it.
23:09 MindDrive And I really wanted to move away from MCollective.
23:09 * iggy slow claps
23:10 ecdhe MindDrive, I think sudo will be your key to victory.
23:10 iggy salt can't solve every problem for everybody (yet)
23:10 ecdhe salt.client.Caller() calls cmd.run calls sudo calls the script you need to run.
23:16 bhosmer joined #salt
23:18 timoguin joined #salt
23:21 timoguin joined #salt
23:22 fxhp joined #salt
23:27 rvankleeck joined #salt
23:27 rvankleeck joined #salt
23:30 davet joined #salt
23:31 davet1 joined #salt
23:33 yomilk joined #salt
23:36 davet joined #salt
23:36 viq joined #salt
23:37 fbettag joined #salt
23:39 baweaver joined #salt
23:39 TaiSHi Can you quiet out an output ?
23:39 TaiSHi Just for one state
23:39 TaiSHi file.directory / clean is rather spammy when deleting cache folders
23:39 bhosmer_ joined #salt
23:44 pipps joined #salt
23:45 pipps99 joined #salt
23:50 iggy state_output in newer versions
23:50 iggy or something like that
23:51 iggy output_lovlevel: quiet
23:57 scarcry joined #salt

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