Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-05-13

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

All times shown according to UTC.

Time Nick Message
00:02 RandalSchwartz I think so
00:09 cewood joined #salt
00:11 uestra joined #salt
00:16 pydanny joined #salt
00:22 Nazca joined #salt
00:22 Nazca joined #salt
00:29 ashw7n joined #salt
00:29 RandalSchwartz not /sys/boot?
00:29 RandalSchwartz oops
00:31 aw110f joined #salt
00:32 georgemarshall joined #salt
00:32 faldridge joined #salt
00:32 faldridge joined #salt
00:35 alunduil joined #salt
00:38 schimmy1 joined #salt
00:43 vbabiy joined #salt
00:43 schimmy joined #salt
00:52 troyready joined #salt
00:58 DaveQB joined #salt
01:00 mgw joined #salt
01:00 zekoZeko joined #salt
01:02 akka_ joined #salt
01:07 jnials joined #salt
01:08 sheldon joined #salt
01:09 sheldon can anyone tell me how to invoke Salt-Key directly ? i'm new to salt
01:09 RandalSchwartz salt-key ... :)
01:10 sheldon not cli invoke,i write a third-party pthon program, i dont how to pass the argument to salt-key function
01:11 RandalSchwartz ahh, not a python guy, dunno
01:11 RandalSchwartz and trying very hard never to learn python
01:11 z3uS wow really?
01:11 z3uS and you're using salt why?
01:11 sheldon ahh, i 'm also new to python
01:11 z3uS and why would you not want to learn python
01:11 RandalSchwartz because salt is the best of breed
01:12 RandalSchwartz and because python just isn't for me
01:12 z3uS python is like the best of bash n' perl with good syntaxing
01:12 z3uS :D
01:12 RandalSchwartz no - python is missing a lot of good things from Perl
01:12 * z3uS smh.
01:12 uestra joined #salt
01:12 RandalSchwartz and was designed by someone who doesn't understand object oriented programming very well
01:12 RandalSchwartz I know... I was in a classroom with him
01:13 RandalSchwartz in fact, I was on a weeklong cruise with him
01:13 RandalSchwartz got plenty of time to talk to Guido
01:13 * z3uS yawns.
01:13 RandalSchwartz I don't mind if other people choose python, but it won't be for me.
01:13 z3uS so sorry you have to suffer the wrath of python in order to use Salt then
01:14 z3uS it must make you angry
01:14 z3uS unless you just have a kink for yaml
01:14 z3uS then I could see the appeal
01:14 RandalSchwartz I also kinda hate YAML too
01:14 z3uS or is it cause you don't like reading right to left with ruby?
01:14 RandalSchwartz you really don't know who I am, do you? :)
01:15 z3uS I just got around to googling about the same time you were typing that
01:15 z3uS :D
01:16 z3uS I find it intersting that you call Salt best of class yet don't particularly like python or yaml, but interesting in a good way
01:16 z3uS :)
01:17 * z3uS has a thing against ruby and DSL's like chef/puppet
01:17 z3uS and primarily write bash and perl for the last 4yrs for autmation but recently migrating to python
01:17 RandalSchwartz I am but a user of Salt
01:18 RandalSchwartz puppet has a weird DSL
01:18 z3uS agreed
01:18 RandalSchwartz chef requires full ruby knowledge
01:18 RandalSchwartz I'm hoping as a user of Salt, I never have to write lines of Python :)
01:18 z3uS lol
01:18 z3uS use Getopt::Long;
01:18 z3uS use Config::Simple;
01:18 z3uS my fav's
01:18 z3uS :)
01:19 RandalSchwartz I interviewed the cfengine, puppet, chef, and salt leaders for FLOSS Weekly
01:19 RandalSchwartz in fact, hatch sometimes claims I helped put salt on the map
01:20 z3uS I found it randomly early last year when I had to chance to choose the platform for my last org's automation
01:20 z3uS it was already a python shop so why introduce ruby or a dsl just to control automation
01:20 z3uS tried salt, stuck with it
01:20 z3uS the zeromq backend is amazing
01:20 RandalSchwartz yeah, I interviewed the zeromq guy too
01:20 z3uS enjoy the parallel shell stuf
01:20 RandalSchwartz he was a remarkable interview
01:20 z3uS I'm doing a fuckload of interviews right now
01:21 z3uS trying to decide what to do next
01:21 z3uS so we'll see
01:21 z3uS twitter, salesforce, facebook, or brigade
01:21 z3uS *shrug*
01:21 z3uS Brigade is kinda my top pick, but I still need to get on the call with Sean
01:21 RandalSchwartz use ziprecruiter.com to search dozens of job boards at once. :)
01:21 RandalSchwartz if you're in the US
01:21 z3uS I'm in SF and no need to search, have enough recruiters chasing me constantly
01:22 RandalSchwartz zipalerts are pretty amazing too
01:22 z3uS no use for much of anything but linkedin really
01:22 RandalSchwartz email every morning for new jobs from your search criteria
01:22 z3uS :)
01:22 RandalSchwartz yeah - they link to linked in to
01:22 RandalSchwartz too
01:23 faldridge joined #salt
01:24 RandalSchwartz geez... I have 163 new connection requestions in linkedin
01:24 z3uS lol
01:24 RandalSchwartz sucks to be famous sometimes. :)
01:24 z3uS º¿º
01:24 z3uS ¯\_(?)_/¯
01:25 Networkn3rd joined #salt
01:27 RandalSchwartz not sure what those chars were supposed to be
01:27 RandalSchwartz are you sending UTF-8?  or latin-1?
01:27 z3uS utf-8
01:27 RandalSchwartz odd... normally those come through
01:27 z3uS first one is random
01:27 z3uS 2nd one was a little guy with his hands raised
01:27 shaggy_surfer joined #salt
01:27 RandalSchwartz instead I got \272\277\272
01:28 z3uS ahh
01:28 z3uS its the same as the thing that shows on my twitter
01:28 RandalSchwartz I don't think that's utf8
01:28 z3uS https://twitter.com/z3usy
01:28 RandalSchwartz I think that's latin 1
01:28 z3uS its utf8
01:28 z3uS pretty sure
01:28 z3uS but could easily be wrong
01:28 RandalSchwartz Oh - I probably don't have that font
01:28 RandalSchwartz not inside emacs :)
01:28 z3uS ahh
01:28 z3uS emacs for irc?
01:29 shaggy_surfer anyone now how to call salt within a bash script…. I am trying this:   sudo salt "'"${HOSTNAME}"'" state.highstate, but it doesn't work…. however echo'ing that on the commandline works and evaluated and replace $HOSTNAME
01:29 RandalSchwartz Terminal.app -> ssh -> tmux -> emacs -> irc -> (you)
01:29 RandalSchwartz yes
01:29 RandalSchwartz and getting that all to be utf8 was... fun
01:29 z3uS heh interesting :)
01:29 RandalSchwartz
01:30 Eugene why run tmux if you've got emacs
01:30 * Eugene ducks
01:30 RandalSchwartz because my connection to my server breaks from time to time
01:30 RandalSchwartz and emacs cannot survive on its own
01:30 RandalSchwartz it needs a daddy :)
01:30 xzarth joined #salt
01:30 RandalSchwartz Oh... no wait
01:30 RandalSchwartz not tmux.
01:30 RandalSchwartz mosh
01:30 Eugene I'm sure there's a daemonization patch
01:31 RandalSchwartz Terminal.app -> mosh -> emacs -> irc
01:31 RandalSchwartz but sometimes I run emacs inside tmux
01:31 RandalSchwartz just not here
01:31 RandalSchwartz no wait - I am running it here
01:31 RandalSchwartz ... [main] 2:emacs*
01:32 RandalSchwartz so Terminal.app -> mosh -> tmux -> emacs -> irc :)
01:32 jesusaurus ... is emacs your irc client?
01:33 z3uS well I'm off, need to grab some food
01:33 z3uS good meeting you RandalSchwartz
01:33 z3uS peace all!
01:34 RandalSchwartz yes.. my irc is in emacs.  I'm typing this in emac.
01:34 RandalSchwartz emacs too.
01:40 jnials joined #salt
01:41 meteorfox joined #salt
01:47 ckao joined #salt
01:47 shaggy_surfer anyone have an answer to the question above?  thanks
01:49 XenophonF joined #salt
01:50 acabrera joined #salt
01:50 wt shaggy_surfer, what question?
01:51 shaggy_surfer anyone now how to call salt within a bash script…. I am trying this:   sudo salt "'"${HOSTNAME}"'" state.highstate, but it doesn't work…. however echo'ing that on the commandline works and evaluated and replace $HOSTNAME
01:52 shaggy_surfer I am trying to execute salt with sudo inside a bash script…. but since the <target> requires single quotes around it, it's making it difficult to do.
01:52 shaggy_surfer I can get it to work with an echo, but not running the command with sudo
01:53 XenophonF the single quotes aren't necessary
01:53 shaggy_surfer really?
01:53 XenophonF they exist solely to prevent globbing by the shell
01:53 shaggy_surfer ouch… I thought they were
01:54 XenophonF in UNIX, the shell handles glob expansion (turning * into a list of files passed to the command)
01:54 XenophonF whereas in DOS/Windows, that's the program's job, not the shell
01:54 XenophonF so if you do:
01:54 XenophonF salt '*' state.highstate
01:55 XenophonF the '*' bit tells the shell (sh, bash, csh, etc.) not to expand * into a list of whatever files are in the current working directory
01:55 XenophonF so you can do:
01:55 XenophonF salt host.example.com state.highstate
01:55 XenophonF you can do:
01:55 shaggy_surfer perfect
01:55 shaggy_surfer thank you very much
01:55 shaggy_surfer that worked.
01:55 XenophonF oh good
01:56 XenophonF you can do this too:
01:56 XenophonF salt -G os:CentOS state.highstate
01:56 XenophonF or salt -G cpuarch:x86_64 grains.item num_cpus
01:57 XenophonF even though the examples in section 4.2 of the salt documentation has the grains target escaped
01:57 shaggy_surfer you guys are awesome, thx for this :)
01:59 XenophonF i'm trying to find a comprehensive, clearly written doc for you but coming up empty :/
02:00 XenophonF this would be the kind of thing covered in a Unix shell scripting book
02:01 XenophonF hey all i'm getting a KeyError when I try to use pkgng.update_packaging_site
02:01 XenophonF this is salt 2014.1.4
02:01 XenophonF http://pastebin.com/b75VE3Vx
02:02 XenophonF i could update PACKAGESITE in /usr/local/etc/pkg.conf manually as a workaround
02:03 XenophonF although for the current version of pkgng, I don't think you're supposed to use PACKAGESITE any more
02:03 XenophonF does anyone use this functionality on FreeBSD?
02:05 l0x3py joined #salt
02:08 Darnoth joined #salt
02:13 elfixit joined #salt
02:15 uestra joined #salt
02:17 ravibhure joined #salt
02:17 ravibhure left #salt
02:21 XenophonF left #salt
02:25 faldridge joined #salt
02:27 anuvrat joined #salt
02:28 ashw7n_ joined #salt
02:29 mateoconfeugo joined #salt
02:33 [diecast] joined #salt
02:34 ajw0100 joined #salt
02:42 Comradephate joined #salt
02:43 ccase joined #salt
03:06 ajw0100 joined #salt
03:10 frasergraham joined #salt
03:15 uestra joined #salt
03:16 kermit joined #salt
03:31 Ryan_Lane joined #salt
03:34 elfixit joined #salt
03:38 user joined #salt
03:45 ashw7n joined #salt
03:48 sheldon joined #salt
03:48 ipalreadytaken joined #salt
03:49 sheldon when i commit a job to install software, the result job return like this ,"the function 'state.highstate' is running as PIO5467 "?
03:50 sheldon dose anybody can tell me why?
03:50 manfred because a highstate is alraedy running on that server?
03:50 manfred sheldon: salt-run jobs.active
03:50 sheldon if i want to execute job concurrent, how i should do ?
03:51 manfred you can only be executing one job at a time on each minion iirc
03:51 manfred well, states related
03:51 sheldon states related?
03:52 manfred yeah, state.highstate or state.sls
03:52 manfred i can run state.highstate, and get test.ping back while that is still running
03:52 sheldon i want to execute two or more jobs at a time , it seems state.highstate can not work
03:52 manfred it is saying that state.highstate is alraedy running
03:54 sheldon i commit a job to install software, and this job is not done yet, then i want to commit a job to install another software, is there a method to make it ?
03:54 manfred you can't do that
03:55 Kumar joined #salt
03:55 Kumar Hi
03:55 manfred you could use states to have them run in sequence, but you cannot do what you are saying
03:55 Kumar need a help on salt mine
03:55 Kumar root@myserver1:/srv/salt# salt '*' 'mine.get' 'roles:hadoopmaster' fqdn grains.item myserver1:     ---------- rhserver1:     ----------
03:55 Kumar this doesn't return any value
03:55 Kumar but
03:55 Kumar root@myserver1:/srv/salt# salt '*' 'mine.get' '*' grains.item myserver1:     ----------     rhserver1:         ----------         fqdn:             rhserver1         roles:             - hadoopmaster rhserver1:     ----------     rhserver1:         ----------         fqdn:             rhserver1         roles:             - hadoopmaster root@myserver1:/srv/salt#
03:55 Kumar returns full mine list
03:56 Kumar how can i filter the output based on mine grains ?
03:56 Kumar anything wrong with salt '*' 'mine.get' 'roles:hadoopmaster' fqdn grains.item
03:56 Kumar ?
03:58 sheldon hi manfred, i use state.highstate to execute job, but it doesnt run in sequence. if there is a job still running ,then state.highstate immediatly return "is running as ****" result
03:58 manfred yes
03:59 sheldon in fact , the job i commit later didnt execute ,right ?
03:59 manfred correct
03:59 manfred because a highstate is still executing
03:59 manfred it is alraedy in highstate, so you cannot execute another highstate until that one finishes
04:01 sheldon that's not good, i write a html to end user, they complain software deploy not work
04:02 manfred then you are doing something wrong, you should only need to go into highstate once, to set the environment and do the whole deploy
04:02 manfred you shouldn't need to run it again
04:02 sheldon because they concurrent commit the job
04:02 manfred immediately after unless something is broken
04:03 sheldon thanks manfred
04:03 zain_ joined #salt
04:03 sheldon i have lunch now.
04:04 Kevin joined #salt
04:04 frasergraham joined #salt
04:04 Furao joined #salt
04:05 obimod joined #salt
04:08 frasergr_ joined #salt
04:10 schimmy joined #salt
04:13 otter768_ joined #salt
04:13 schimmy1 joined #salt
04:15 otter768 joined #salt
04:16 uestra joined #salt
04:23 jnials joined #salt
04:24 ifnull joined #salt
04:25 jgilik joined #salt
04:26 jgilik Hi all, I'm seeing an issue where mine.get is returning an empty dict when trying to use it with network.ip_addrs (as well as network.interfaces -- anyone know any other things to try to triage the problem?)
04:26 jgilik command line I'm trying is
04:26 jgilik salt-call -l all mine.get '*' network.ip_addrs 'glob'
04:26 jgilik though I am ultimately going to be matching by grain
04:27 jgilik a regular call to network.ip_addrs /w salt-call network.ip_addrs returns the IP normally, which makes me suspect this is not an issue in the network module
04:28 jgilik Any advice on how to triage this?  In the source, it seems to ultimately reach an RPC step, which ultimately returns {} ("ret = sreq.send(load)" in salt.modules.mine.get())
04:28 manfred do you see the ips with salt-call mine.get \* network.ip_addrs ?
04:29 jgilik no
04:29 manfred did you enabled network.ip_addrs: [] in your mine_functions on each minion?
04:29 jgilik interesting
04:29 jgilik let me look up how to do that :)
04:30 manfred jgilik: http://docs.saltstack.com/en/latest/topics/mine/
04:30 jgilik thanks -- I've been refering to http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.mine.html
04:30 jalbretsen joined #salt
04:30 jgilik I'll see if I can contribute a link to the docs of that module
04:31 manfred yeah, you still need to configure it in the /etc/salt/minion, or you need to do that mine.send first before it can do a mine.get
04:31 jgilik I'm trying to do a master-only configuration
04:31 jgilik I assume /etc/salt/minion is the config on the minion?
04:31 jgilik if so, I'll look into mine.send
04:32 manfred put the mine.send as part of the scheduler
04:32 manfred to have the minions send it even X minutes
04:32 manfred http://docs.saltstack.com/en/latest/topics/jobs/schedule.html
04:33 manfred useing the module.run state to run teh mine.send module
04:34 jgilik thanks, will look into it
04:34 jgilik I appreciate the guided tour!
04:34 manfred np
04:36 schimmy joined #salt
04:41 schimmy joined #salt
04:41 jgilik adding misc to my /srv/pillar/top.sls and populating it thus fixed the issue:
04:41 jgilik # cat /srv/pillar/misc.sls
04:41 jgilik mine_functions:
04:41 jgilik network.ip_addrs: []
04:42 jgilik (using '*' as the matching clause for misc.sls)
04:42 jgilik I did need to restart the minion for it to take effect, but I'll see if there are issues when standing up new minions; hopefully not
04:43 manfred should be able salt \* saltutil.refresh_pullar for that maybe? not really sure if it needs to be read before/around/with the minion config to do stuff
04:43 manfred ¯\(°_o)/¯
04:44 jgilik hahahahaha -- I got the impression from the docs
04:44 jgilik that there is a certain refresh interval for salt mine information
04:44 jgilik and that the only case it gets updated outside of that interval is on minion start
04:45 manfred there is that
04:45 jgilik this is why I restarted the minion: not to refresh the pillar data, but to force a mine update
04:45 manfred possibly
04:45 jgilik I imagine there's a more clever and suave way of doing this, but... it's 9:45pm
04:45 jgilik and it had the desired effect:
04:45 jgilik Comment: File /etc/keystone/keystone.conf updated
04:46 jgilik Succeeded: 5
04:46 manfred nice
04:46 jgilik Hooray!  You've unblocked me :)  Thanks again!
04:47 manfred np!
04:48 tr_h joined #salt
04:50 Furao joined #salt
05:00 frasergraham joined #salt
05:06 pydanny joined #salt
05:07 malinoff joined #salt
05:12 lionel joined #salt
05:17 uestra joined #salt
05:24 CeBe joined #salt
05:26 lionel joined #salt
05:36 anuvrat joined #salt
05:46 tr_h joined #salt
05:53 picker joined #salt
05:57 ajw0100 joined #salt
06:13 skyion joined #salt
06:14 n8n joined #salt
06:17 uestra joined #salt
06:18 saravanans joined #salt
06:24 mateoconfeugo joined #salt
06:32 briner joined #salt
06:47 briner_ joined #salt
06:55 uestra joined #salt
06:56 BigBear joined #salt
06:58 bottleneck joined #salt
07:01 kdorf joined #salt
07:03 slav0nic joined #salt
07:03 slav0nic joined #salt
07:07 rupsky________ joined #salt
07:11 ndrei joined #salt
07:13 harobed_ joined #salt
07:14 harobed_ joined #salt
07:14 harobed_ joined #salt
07:15 londo_ joined #salt
07:15 harobed_ joined #salt
07:17 jnials joined #salt
07:22 ifnull joined #salt
07:24 TyrfingMjolnir joined #salt
07:25 siciliansun joined #salt
07:28 anuvrat joined #salt
07:30 ghartz joined #salt
07:30 martoss joined #salt
07:31 TyrfingMjolnir joined #salt
07:33 superted joined #salt
07:34 superted Anybody around that could help me with a potentially stupid question? :)
07:34 malinoff superted, ask it
07:34 superted Cool, thanks!
07:35 superted Well, i've been keeping an eye on a few fixes and bugs, i notice them get merged into develop... then i lose track. How can i track if a fix has made it into the next release?
07:35 superted Is it just from release notes or is it taken from a different branch?
07:36 cast joined #salt
07:37 malinoff superted, this may be interesting for you: https://groups.google.com/forum/?pli=1#!topic/salt-users/usIleqkiHQg
07:38 superted Will have a read now thanks!
07:39 superted Thanks for the link, i'm still not 100% clear though on the process. Is the release version taken from the develop branch? Is there anyway through github i can track the ticket to see if it's in X release?
07:40 TyrfingMjolnir joined #salt
07:41 malinoff superted, as far as I know, PRs are taken from different branches, but internally I'm not sure if they have a separate branches... Looks like they don't. So I think yes, develop is the mainline
07:42 malinoff About releases, they are quite spontaneous
07:42 malinoff As i see ofc
07:42 ghartz_ joined #salt
07:42 superted Okay so if i look at release notes for release Salt 2014.1.4 then it means that as of the date that was released it was the develop branch at that time?
07:43 superted Thanks for the info by the way
07:43 Kenzor joined #salt
07:43 malinoff superted, they use tags though
07:43 malinoff https://github.com/saltstack/salt/tree/v2014.1.4
07:43 superted Ahh, that's the missing link
07:44 superted Yep, that makes sense. thanks for the heads up
07:44 topochan joined #salt
07:44 malinoff One more thing to complain about this new versioning :)
07:46 siciliansun1 joined #salt
07:47 ghartz_ joined #salt
07:50 darkelda joined #salt
07:50 darkelda joined #salt
07:55 siciliansun1 left #salt
08:00 ghartz_ joined #salt
08:07 bottleneck help, is there any way of obtaining the list of ip-addresses of all the minions of a specific role - lets say web
08:07 ndrei joined #salt
08:09 concord joined #salt
08:13 superted ^ You can as the IP is stored as a grain. 2 secs, i'll dig out an example
08:13 saravana_ joined #salt
08:14 cast grains.item ipv4, though never used a role [yet :] so not sure how to target based on them. are they merely a grain?
08:14 lee__ joined #salt
08:15 ghartz joined #salt
08:16 superted beat me too it
08:17 cast about my favorite module/function ever is the sys.doc :)
08:26 lee__ hello
08:31 picker bottlneck: i guess you are looking for something like http://docs.saltstack.com/en/latest/topics/mine/#example
08:32 TyrfingMjolnir joined #salt
08:35 saravanans joined #salt
08:37 orbit_darren joined #salt
08:37 BigBear joined #salt
08:42 bottleneck will try that in a while
08:48 pydanny joined #salt
08:54 Comradephate_ joined #salt
09:03 masterkorp hello
09:03 masterkorp guys how do i subscribe to the mailing list without needing a gmail account ?
09:05 ghartz joined #salt
09:06 babilen masterkorp: I think you need one for google groups (it's *not* a mailing list)
09:06 Guest7273 joined #salt
09:06 masterkorp fuck
09:06 babilen masterkorp: https://support.google.com/groups/answer/46438?hl=en (1 second google)
09:06 masterkorp all i need to get emails and send emails back
09:08 MrTango joined #salt
09:18 zain_ joined #salt
09:19 tdk2fe joined #salt
09:23 elithrar joined #salt
09:30 skyion joined #salt
09:31 uestra joined #salt
09:41 ghartz joined #salt
09:43 anitak joined #salt
09:45 giantlock joined #salt
09:48 agend joined #salt
09:49 pdayton joined #salt
09:51 ndrei joined #salt
09:55 bram_ joined #salt
09:56 bram_ hi everyone… simple question: what would be the best site(s) to post job-offerings that are looking for some salt-enabled people :-)
09:56 bram_ of course stack overflow, but… can you guys recommend me some more?
10:06 nkuttler bram_: where are you located?
10:07 londo__ joined #salt
10:07 nkuttler fwiw, you might want to try local usergroups
10:14 TyrfingMjolnir joined #salt
10:19 pdayton joined #salt
10:20 jnials joined #salt
10:26 sulky joined #salt
10:28 ajprog_laptop joined #salt
10:33 bram_ joined #salt
10:33 CeBe joined #salt
10:33 bram_ nkuttler: germany/berlin!
10:33 bram_ sorry I was called into a meeting
10:33 bram_ fucking meetings...
10:34 bram_ is there a user group berlin?
10:36 nkuttler not according to http://www.saltstack.com/community/
10:36 nkuttler but berlin should have plenty of sysadmin groups and such
10:36 nkuttler actually, i'll be at the aws summit in berlin this thursday. was wondering if any salt people would go
10:37 nkuttler http://www.meetup.com/blndevops/ scheint relevant zu sein
10:38 bram_ nkuttler: also going to the AZS summit :)
10:40 nkuttler nice :)
10:40 bram_ aws, djeez… :)
10:41 bram_ thx for the devops list
10:41 bram_ I’ll drop them an email to see if it’s acceptable to throw something in there
10:44 nkuttler good luck :)
10:48 AviMarcus joined #salt
10:51 viq bram_: out of curiousity, you have a link to the job offer? ;)
10:51 bhosmer joined #salt
10:51 bram_ http://goo.gl/KbXXYy & http://goo.gl/f3EyCJ
10:52 bram_ read between the lines… I don’t like HR messing with my initial list of requirements :/
10:52 bhosmer joined #salt
10:53 bhosmer joined #salt
10:54 saravanans joined #salt
10:54 viq Hm, I'm about 2/3 - 3/4 of that, I wonder if I'd feel like moving to Berlin ;)
10:55 quist joined #salt
10:59 elfixit joined #salt
11:03 faldridge joined #salt
11:04 Outlander joined #salt
11:06 harobed_ joined #salt
11:07 harobed_ joined #salt
11:09 skyion joined #salt
11:12 ml_1 joined #salt
11:13 diegows joined #salt
11:14 ekristen joined #salt
11:26 bottleneck I deleted one of my minion's key, accepted it again. But now I'm not able to get that minion via test.ping. What may be the cause of that
11:29 CeBe joined #salt
11:30 bottleneck restarted both master and minion, still not able to get
11:31 BigBear joined #salt
11:32 bottleneck can anyone please point out what might be missing
11:32 jrdx joined #salt
11:35 viq bottleneck: check logs
11:38 younqcass joined #salt
11:38 jaycedars joined #salt
11:39 saravana_ joined #salt
11:44 ggoZ joined #salt
11:46 nkuttler bottleneck: if necessary run minion + master with -l debug
11:52 bottleneck In master log: Authentication request from minion
11:52 bottleneck and Authentication accepted from minion
11:52 bottleneck displays continuously
11:52 bottleneck but I'm able to see the minion via salt-key -L
11:53 viq and in minion logs?
11:53 bottleneck under accepted keys
11:54 bottleneck there is nothing in minion log
11:54 sulky joined #salt
11:58 viq is the minion running?
11:58 viq I have seen minions kill themselves when server refuses the key, so you may need to restart the service there
11:58 Outlander joined #salt
12:01 uestra joined #salt
12:05 NewGeek joined #salt
12:06 sgviking joined #salt
12:08 tessellare joined #salt
12:13 uestra joined #salt
12:14 bastion1704 joined #salt
12:15 sulky joined #salt
12:21 ravibhure joined #salt
12:21 jnials joined #salt
12:23 skyion bottleneck: might also be worth clearing the local cache directory on the salt minion
12:23 albertid_ joined #salt
12:23 bottleneck trying things out
12:24 albertid_ Can I have cmd.wait run only if a specific file has changed?
12:27 bhosmer joined #salt
12:31 ndrei joined #salt
12:32 FarrisG joined #salt
12:32 __alex_ i installed salt master from git(branch 2014.1), now i want to switch to ppa pakets(2014.1.3), plan is to stop salt-master, delete /usr/local/bin/salt*, any hint? remove /var/cache/salt/master?
12:34 __alex_ git install is versioned 2014.1.0-20-g0005c75
12:34 cast deleting /usr/local/bin/salt* would only be a partial uninstall
12:35 __alex_ what else?
12:35 hernantz joined #salt
12:37 __alex_ oh, /usr/local/lib/python2.7/dist-packages/salt for example
12:37 cast check out the contents of the PPA .deb's to get an idea,
12:39 cast i've only ever installed from a repo so can't help you much with cleaning up your install
12:40 hernantz hi! i'm getting Failed to create database ligonier with this state: http://pastebin.com/0UUGPYkR, how can i get more insight on whats going on?
12:41 bottleneck File "checkrc.pxd", line 21, in zmq.backend.cython.checkrc._check_rc (zmq\back end\cython\socket.c:6174) zmq.error.ZMQError: Address in use - when starting the minion
12:43 jslatts joined #salt
12:44 bottleneck and "no", I'm not trying to start another instance of the minion while one is already started
12:48 obimod joined #salt
12:54 uestra joined #salt
12:57 pviktori joined #salt
12:59 hernantz joined #salt
12:59 hernantz salt-call -l debug state.highstate was what i was looking for ^
13:00 Comradephate joined #salt
13:03 topochan joined #salt
13:04 cpowell joined #salt
13:06 faldridge joined #salt
13:06 mgw joined #salt
13:06 novastar2 joined #salt
13:09 jaycedars joined #salt
13:10 rojem joined #salt
13:15 dude051 joined #salt
13:15 victorpoluceno joined #salt
13:16 resmike joined #salt
13:16 BigBear joined #salt
13:18 Networkn3rd joined #salt
13:22 vbabiy joined #salt
13:22 racooper joined #salt
13:28 jaycedars joined #salt
13:28 Dudz_ joined #salt
13:29 GradysGhost joined #salt
13:30 CeBe joined #salt
13:32 ashw7n joined #salt
13:33 jaycedars joined #salt
13:34 NewGeek Hi there ! I'm new to salt and I'm loving it ! I want to deploy salt over one hundred servers but I still have some doubts on how to make the best of it. One of my biggest problem is how to regroup servers (to use these group names in both pillars and states). The problem with nodegroups (in /etc/salt/master) is that one needs to restart the master once it is modified. Do you know a better way ?
13:35 viq NewGeek: you can assing custom grains, or you can assign pillars to machines. Though you can't use pillars in pillars, so probably custom grains would be better for that
13:36 viq NewGeek: although master comes up pretty much immediately after restart, so is this really a problem?
13:39 NewGeek viq: Thanks viq. I've thought of grains. The problem is that I'm storing sensitive data in pillars. And it seems that's a bad ideia to match sensitive data in pillar with grains...
13:40 NewGeek <NewGeek> viq: i'ts not a very big problem to restart it each time. But I would rather not if there is another solution :)
13:40 jaycedars joined #salt
13:43 NewGeek <NewGeek> viq : and i've tried to use pillars to assign roles... but has you said, I can not use pillar data in pillars...
13:43 faldridge joined #salt
13:43 ipmb joined #salt
13:44 Cuy joined #salt
13:45 viq NewGeek: I haven't used it, but there is stuff that could make it easier for you, like reclass http://reclass.pantsfullofunix.net/ or range (which I have to find)
13:45 kaptk2 joined #salt
13:45 frasergraham joined #salt
13:46 happytux joined #salt
13:46 hernantz left #salt
13:47 BigBear joined #salt
13:49 quickdry21_ joined #salt
13:50 NewGeek <viq>: Seems interesting. I'll give it a try ! Thank's a lot.
13:53 Networkn3rd joined #salt
13:55 saguilar joined #salt
13:55 bhosmer joined #salt
13:58 pdayton joined #salt
14:00 xmj does anyone here have salt-states around for jenkins / swarm?
14:01 viq swarm ?
14:01 viq http://savannah.nongnu.org/projects/swarm ?
14:02 timoguin there's a jenkins formula: https://github.com/saltstack-formulas/jenkins-formula
14:02 xmj https://wiki.jenkins-ci.org/display/JENKINS/Swarm+Plugin
14:03 xmj I've been trying the last few hours to get this $#9& running as non-root. Fails, because it tries to write maven jars to / ... #jenkins is no Help at all, so I'm hoping to find people elsewhere :|
14:04 toastedpenguin joined #salt
14:05 to_json joined #salt
14:06 ghartz joined #salt
14:08 uestra joined #salt
14:09 lilmatt joined #salt
14:10 ajprog_laptop joined #salt
14:11 rupsky________ joined #salt
14:14 Eugene WHen I encounter an app that "has to" run as root I just spin up another VM and call it a day
14:15 to_json joined #salt
14:15 nkuttler hm, jenkins formula sounds nice
14:15 sroegner joined #salt
14:15 nkuttler i like the ui of travis, but i hear running it is a bitch
14:16 viq How can target a couple machines with a single state in top.sls ?
14:16 viq AKA "what's the easiest way to..."
14:16 viq Can I just 'machine1,machine2': and call it a day?
14:16 nkuttler viq: http://docs.saltstack.com/en/latest/ref/states/top.html
14:17 nkuttler yeah, that should work
14:17 AdamSewell joined #salt
14:17 AdamSewell joined #salt
14:17 viq Although I need - match: list
14:17 viq OK then
14:17 timoguin or -L on the cli
14:17 siciliansun joined #salt
14:18 * viq nods
14:18 jnials joined #salt
14:21 tyler-baker joined #salt
14:25 Cuy Hi! Can someone tell me, roughly how well Salt scales? I'm supposed to find a configuration management system to deploy and manage OpenStack, so the number of servers that can be managed from one Salt master is of much interest to my boss.
14:26 viq Cuy: LinkedIn supposedly has ~10k servers per master
14:26 siciliansun left #salt
14:27 viq But it obviously depends how you use salt
14:27 Cuy Wow, that sounds a lot like what Chef boasts
14:27 viq I also heard claims/recommendations of up to 20k nodes per master with chef
14:28 RandalSchwartz you're aware of salt syndic?
14:28 Cuy not yet, no
14:28 RandalSchwartz it allows you to increase fan-out
14:28 danielbachhuber joined #salt
14:29 RandalSchwartz master talks to ten syndics then talks to 1000 nodes each
14:29 viq Cuy: it's like a proxy. you connect minions to a syndic, and then several syndics to a central master. So you can have the ~10k per syndic, but from central master be able to issue commands to all your mininos
14:29 viq minions
14:30 Cuy alright - that sounds nice! I'll surely put that into my documentation
14:30 viq It requires considerations, checking, etc, bla bla bla. But it's there in the flyer ;)
14:30 Cuy =)
14:31 Cuy For now, it's about figuring out how to build an OpenStack environment
14:31 viq Also you certainly can do things to salt that will case it to have problems coping at a much lower scale, but I guess you can with everything
14:31 Cuy So far, I'm trending towards Juju with an additional CMS for the power (like Python in SLS or Ruby in a Chef recipe)
14:32 viq I know some people use puppet for that. I want to play with standing up an OpenStack environment too, but haven't even looked at the issue, not to mention touching it
14:32 viq Cuy: SLS files are by default in YAML, not python
14:32 Cuy Yes, I saw that, but the documentation claimed, you could alternatively also render them in Python
14:32 tedski that would be cumbersome
14:32 viq yeah
14:33 Cuy And I imagine, that you should then have the full power of the language at your command?!?
14:33 tedski Cuy: even with yaml, you have the full power of the language
14:33 jaimed joined #salt
14:33 Cuy Or am I completely missing the point here?
14:33 bhosmer joined #salt
14:33 tedski Cuy: since you can embed python in the yaml using jinja
14:33 Cuy Ah, OK - I was not aware of that
14:33 tedski Cuy: but yaml keeps it simple and fast
14:34 nkuttler tedski: jinja can do that?
14:34 nkuttler i thought it had no code blocks
14:34 sroegner joined #salt
14:35 tedski sorry, it does not... i'm going wacky
14:35 timoguin There are a handfull of openstack formula implementations out in the interwebs
14:35 Cuy =D
14:35 tedski you can make calls to salt to run python modules on the state's behalf though
14:35 nkuttler ah, ok :)
14:35 nkuttler mako does though
14:35 nkuttler (afaik, not used it)
14:36 uestra joined #salt
14:36 tedski i.e. key: {{ salt['module.method'](args) }}
14:36 tedski that kind of use
14:36 nkuttler can confirm, modules are super useful
14:36 Cuy Well, so far I've been thinking about using Chef, bcfg2 or Salt, simply because of the extensibility by a "real" programming language (and I know a bit of Python, so right now I'd like to go with bcfg2 or Salt)
14:37 tedski Cuy: out of curiosity, why isn't puppet on your evaluation list?
14:37 Cuy I've stumbled upon some bad remarks about scalability of Puppet
14:38 Cuy Mainly, that a single master server dies as soon as you try to manage more than 1500 nodes
14:38 viq Cuy: also look up reactor system in salt, I find the concept sexy
14:38 tedski how big is your environment?  another team at my company didn't run into scaling issues until 100k nodes or so
14:38 tedski oh, single master... yeah
14:38 Cuy So far, there is no environment
14:38 tedski i'm not saying you should use puppet... don't get me wrong... i think salt wins out on all fronts
14:38 Cuy I'm in a little data center and we would like to offer OpenStack as a service
14:39 Cuy And I'm the intern who looks up the possible ways of implementation ;)
14:39 viq :D
14:39 tedski i'm doing something similar
14:39 tedski our build & release system will use salt-cloud to call our private openstack buildout
14:41 Cuy How far would you go with the open source version of Salt? As in - rough estimate of # of nodes?
14:42 tedski tens of thousands
14:42 RandalSchwartz "the open source version"?  You mean there's another version? :)
14:42 tedski many tens of thousands
14:42 tedski RandalSchwartz: yes
14:42 timoguin considering you can scale horizontally with syndics and masters of masters, you can keep going
14:42 ndrei joined #salt
14:42 RandalSchwartz when did salt become freemium?
14:42 nkuttler tedski: where is that? thomas has always said he's against open core and such
14:42 tedski and the only reason i limit it to that is because i use gitfs with a github enterprise backend
14:42 Cuy Nice, that will be EXACTLY what my boss wants to hear!
14:42 timoguin there's not a non-open source version. they offer enterprise support for the same code base as the main project
14:42 tedski oh, it's just a support contract from what i understand
14:43 nkuttler yeah, support..
14:43 RandalSchwartz yeah... like what I do for Perl
14:43 Cuy Ah, sorry, then I got the concept wrontg
14:43 RandalSchwartz Perl is free... I cost money
14:43 tedski this is what happens when you don't sleep... your brain half fires
14:43 nkuttler off to bed! :)
14:43 Cuy why would I sleep, it's the middle of the day ;)
14:44 timoguin perfect nap time!
14:44 Cuy hehe
14:44 nkuttler i sleep whenever i feel sleepy. freelancing has it's benefits..
14:44 Cuy yeah... I'm sitting here right in front of my boss' office door...
14:44 tedski i woke up at 0130 with the brain racing about test cases :/
14:44 nkuttler hehe
14:44 viq Cuy: sleep on his doorstep ;()
14:44 tedski couldn't get back to bed until after 3
14:44 viq ;)
14:45 Cuy If I sleep in the doorstep, people will trip over me
14:45 Cuy but really, I'm in Germany - it's not even 5 p.m. here
14:46 lz-dylan joined #salt
14:46 nkuttler sounds like beer time
14:46 tedski offices... i remember those days... open floor plans are so noisy :(
14:46 Cuy No reason to feel sleepy, with the coffee machine only 10 steps away
14:46 KennethWilke joined #salt
14:46 alunduil joined #salt
14:46 Cuy sorry to disappoint you, nkuttler, but I despise beer with a passion...
14:47 KennethWilke moar for the rest of us!
14:47 nkuttler Cuy: funny, i stopped drinking it.. too carby
14:47 nkuttler now i don't even like the taste any more
14:47 Cuy hehe I'm on intermittent fasting right now - but low carb could go on top of that.
14:48 Cuy Well, I'll have to fill out the rest of my excel sheet now - Thanks, everyone! You've been very helpful and I'm almost sure, I'll be here again!
14:49 tedski have fun!
14:49 tedski good luck with your internship
14:49 tedski and welcome to the club :)
14:49 Cuy I think it's going quite well
14:49 Cuy =)
14:49 Cuy Thanks
14:51 resmike joined #salt
14:54 jeffasinger joined #salt
14:54 EntropyWorks joined #salt
14:54 anuvrat joined #salt
14:55 che-arne joined #salt
15:01 nocturn joined #salt
15:02 rupsky________ joined #salt
15:03 brianhicks joined #salt
15:03 ggoZ joined #salt
15:04 resmike joined #salt
15:08 patrek joined #salt
15:09 frasergraham joined #salt
15:12 nw0428 joined #salt
15:12 nw0428 Hello all!
15:13 nw0428 Does anyone know if it is ok to put backticks into your states?
15:13 [diecast] joined #salt
15:14 nkuttler nw0428: try it?
15:14 nkuttler if not you can probably just quote them
15:16 nw0428 I guess what I would like to know if there is a way to use backticks in such a way that they are run on the minion machine
15:16 glennyb joined #salt
15:16 nw0428 I am trying to create a salt-state that installs the salt master
15:16 glennyb hi folks, anyone got time for a salt/google compute engine question?
15:17 pydanny joined #salt
15:17 viq nw0428: how about salt-bootstrap script?
15:17 nw0428 That works for installing the minion but not, i think the master
15:17 viq nw0428: it can install master as well depending on flags you pass to it
15:18 abe_music joined #salt
15:18 timoguin nw0428: are you just wanting your states to shell out onn the minion?
15:19 nw0428 Timoguin: yeah
15:19 viq nw0428: have you seen cmd.run state?
15:19 nw0428 It may be easier to just use the salt bootstrap script instead though
15:19 glennyb running into issues re-creating a gce virtual server after its been deleted (i'm setting up a demo environment I can blow away and recreate as needed)
15:19 gamingrobot joined #salt
15:19 resmike joined #salt
15:19 conan_the_destro joined #salt
15:19 nw0428 viq: yeah but I was trying to work through pkgrepo.managed to add the salt ppa
15:19 rupsky________ joined #salt
15:20 johanek joined #salt
15:21 EntropyWorks joined #salt
15:21 [diecast] joined #salt
15:23 viq And why do you need to shell out for that?
15:23 schimmy joined #salt
15:24 mateoconfeugo joined #salt
15:24 rojem joined #salt
15:25 EntropyWorks joined #salt
15:26 schimmy1 joined #salt
15:27 EntropyWorks joined #salt
15:27 nw0428 to get the ubuntu version (i.e. precise)
15:27 nw0428 `lsb_release -sc`
15:28 nw0428 I think I will just pass over the salt boostrap script
15:28 jbub joined #salt
15:29 acu joined #salt
15:33 viq nw0428: that should be available in grains
15:34 resmike joined #salt
15:35 abe_music does anyone know if the newer state.orchestrate is any faster than the state.overstate? we're trying to speed up a distributed install of hadoop, hbase, and a few other things on a 200+ node cluster....any thoughts are welcome
15:36 ajprog_laptop1 joined #salt
15:38 babilen abe_music: You could cut down on states by enumerating only those you want with state.sls_id or state.sls (if you know that a number of states are "done" and you don't want to wait for their tests to finish)
15:38 babilen No idea about orchestrate
15:38 babilen Ah, state.sls is state.orchestrate :)
15:39 timoguin abe_music: I'm not sure if it's any faster, but I do know that overstate is being deprecated.
15:40 abe_music babilen: yeah, i was confused when i read that yesterday :) i do like the new name though
15:40 abe_music timoguin: right, i read that too....we'll move over to that eventually, but the documentation in that area has been pretty lacking
15:40 Gareth ahoy hoyu.
15:41 abe_music noticed the docs were updated though, so i might start jumping into it soon
15:41 abe_music (especially if it can speed things up :) )
15:41 timoguin yea there have been several name changes that have gotten me...
15:41 babilen abe_music: I find it a bit less obvious as to what it does ... We use the state.sls_id and state.orchestrate (well, state.sls until now) functionality quite often if we "just" want to roll out specific changes ...
15:41 babilen It is a lot faster if you have plenty of states that are in the correct state already
15:42 conan_the_destro joined #salt
15:42 abe_music babilen: right, a second orchestrate is a lot faster...installing all of the packages the first time does take a while, but i'm wondering if there are any ways to speed up the execution and returning of the individual states
15:43 abe_music i'm currently messing with the "worker_threads" parameter to see if that helps...also curious if using an external returner like redis or something will not bog down the master
15:43 babilen abe_music: I would argue that if there would be a general way that it would have been used already. If you have to run *all* states anyway then you can (and probably should) use state.highstate
15:43 ajw0100 joined #salt
15:44 smcquay joined #salt
15:44 abe_music babilen: but they have to come in a particular order, for example, in hadoop your namenode needs to be up and running before datanodes, same sort of thing for hbase
15:44 abe_music btw, babilen here's the overstate file we're generating to use during our orchestration: https://gist.github.com/abemusic/4e46b20d3ebd8fe4ae46
15:44 babilen abe_music: Ah, you are also orchestrating multiple nodes (sorry, didn't quite grasp that immediately)
15:45 abe_music it's pretty straight forward and it's meant to be generic so that users can build new orchestration routines in our platform fairly easily using the UI
15:45 abe_music babilen: yeah, 204 nodes to be exact :)
15:45 hunter_ joined #salt
15:45 abe_music 200 data nodes, 1 namenode, 3 zookeepers
15:45 abe_music they're all provisioned in the order of that gist
15:45 timoguin nice
15:47 abe_music we've been saying it for a while, but we're really meaning to open source what we're building....my company/boss just needs to agree and let us pull the trigger
15:47 rlarkin_ joined #salt
15:49 jcockhren abe_music++
15:49 resmike joined #salt
15:50 uestra joined #salt
15:51 slav0nic left #salt
15:51 timoguin abe_music: your salt infrastructure?
15:52 darkelda joined #salt
15:52 tligda joined #salt
15:53 jcockhren timoguin: abe_music is with digital reasoning
15:53 abe_music timoguin: our platform we built on top of salt to make it easier for non-systems folks to manage their infrastructure. think a bunch of engineers and data science people needing hadoop or cassandra or whatever, but they don't know how or need to be bothered with how to make that happen
15:53 abe_music jcockhren: this is true :)
15:54 timoguin jcockhren: yea i knew that. didn't know if he meant the salt stuff or their larger learning platform
15:54 abe_music heard you guys and my guys spoke a while back....*makes note to see where we're at with that*
15:54 rlarkinhome joined #salt
15:54 timoguin abe_music: sounds nice
15:54 jeffasinger joined #salt
15:54 abe_music timoguin: oh, yeah just the salt stuff....the learning will remain closed :)
15:54 jcockhren heh
15:54 timoguin i assumed so. :)
15:55 resmike joined #salt
15:55 abe_music of course i would prefer it to be open, but not everyone has the same viewpoint haha
15:57 rojem joined #salt
15:58 NV joined #salt
16:03 jalbretsen joined #salt
16:07 ashw7n joined #salt
16:11 schimmy joined #salt
16:12 KyleG joined #salt
16:12 KyleG joined #salt
16:12 schimmy1 joined #salt
16:15 Ryan_Lane joined #salt
16:16 redondos joined #salt
16:16 redondos joined #salt
16:17 TheRealBill_here joined #salt
16:18 uestra joined #salt
16:19 resmike joined #salt
16:20 vbabiy joined #salt
16:20 forrest joined #salt
16:23 meteorfox joined #salt
16:25 troyready joined #salt
16:27 bemehow joined #salt
16:27 Teknix joined #salt
16:27 stephas joined #salt
16:28 uestra joined #salt
16:29 bemehow joined #salt
16:30 ashw7n_ joined #salt
16:34 vbabiy joined #salt
16:34 resmike joined #salt
16:34 dave_den joined #salt
16:34 eriko joined #salt
16:37 kballou joined #salt
16:40 viq kiorky: what is the proper syntax for the tls module to use subjectAltName? subjectAltName='DNS:text.example.com' ?
16:41 jeffasinger joined #salt
16:41 uestra joined #salt
16:41 resmike joined #salt
16:42 ndrei joined #salt
16:48 CeBe1 joined #salt
16:48 uestra joined #salt
16:49 kballou joined #salt
16:49 londo_ joined #salt
16:52 jimklo joined #salt
16:52 Joseph joined #salt
16:53 Joseph I am trying to get the stateful option to work for the cmd state module. However, no matter what my script returns on the last line, salt always marks it as successful. http://salt.readthedocs.org/en/v2014.1.1/ref/states/all/salt.states.cmd.html?highlight=cmd#module-salt.states.cmd
16:53 analogbyte is there a way to disable the automatic merging of top files from multiple environments?
16:54 viq Joseph: that is based on return code
16:55 Joseph oh okay...yes i can get that working. i misunderstood the purpose of stateful then. Its only supposed to tell if you something has changed from the scirpts point of view. It doesn't control whether salt deems it a failure or success.
16:55 Joseph got it
16:55 * viq reads about the stateful
16:55 [diecast] joined #salt
16:55 viq Yeah, I guess so
16:56 nkuttler oh.. hey.. i just ran virt.hyper_info and see that salt already knows about my libvirtds...
16:57 zirpubolci joined #salt
16:57 joehillen joined #salt
16:58 nkuttler that seems too easy :) anyway, just wanted to spin up a few vms locally..
17:01 mateoconfeugo joined #salt
17:01 KyleG1 joined #salt
17:02 * viq nudges kiorky
17:02 vbabiy joined #salt
17:04 forrest viq, it's 7 PM in france, he might not be around
17:04 opapo joined #salt
17:05 ml_1 joined #salt
17:05 viq meh
17:06 viq Well, it's exactly the same time here :P
17:06 viq Trying to figure out how to use subjectAltName in tls.create_csr and since he wrote it... ;)
17:06 ccase joined #salt
17:06 [diecast] joined #salt
17:06 ashw7n joined #salt
17:10 uestra joined #salt
17:12 nkuttler so i'm just running a virt.init and the process just seems to hang? no cpu activity, no log output
17:15 aw110f joined #salt
17:17 eculver joined #salt
17:17 eculver joined #salt
17:17 aw110f hi, is it possible to store a command output to a variable in the sls state file?
17:20 n8n joined #salt
17:20 travisfischer joined #salt
17:22 nkuttler gah.. can't really tunnel the master over dsl while the backup runs..
17:23 Ryan_Lane joined #salt
17:23 ndrei joined #salt
17:24 anuvrat joined #salt
17:26 patrek_ joined #salt
17:27 bhosmer joined #salt
17:28 frosty996 joined #salt
17:28 frosty996 hey guys, anyone have a suggestion on where to start looking to debug this error:
17:29 frosty996 https://gist.github.com/jfrost/b45324547390dc205828
17:29 frosty996 I was guessing it’s in my top file, but it’s pretty simple.
17:29 frosty996 salt-master is 2014.1.3-2precise1
17:29 will joined #salt
17:29 frosty996 salt-minion is 2014.1.3+ds-2trusty2
17:33 jhulten joined #salt
17:33 jhulten joined #salt
17:34 xt joined #salt
17:34 KyleG joined #salt
17:34 KyleG joined #salt
17:35 jeffrubic joined #salt
17:36 jnials joined #salt
17:36 frosty996 oh, perhaps this is a clue from the minion log: https://gist.github.com/jfrost/b743801097662f475aa0
17:36 abe_music I've already increased my worker_threads to 10, should I go higher? What's the downside? [ERROR   ] Salt request timed out. If this error persists, worker_threads may need to be increased.
17:37 bhosmer joined #salt
17:41 Ryan_Lane I feel dirty making AWS code commits while I'm sitting at the openstack summit: https://github.com/saltstack/salt/pull/12741
17:42 shaggy_surfer joined #salt
17:43 abe_music lol
17:43 shaggy_surfer HI all, can someone tell me how you can do ordering in salt.  I want to make sure a mount point exists before trying to remove a lock file on a database
17:44 aw110f_ joined #salt
17:44 abe_music shaggy_surfer: check out using require statements in your sls files: http://docs.saltstack.com/en/latest/topics/tutorials/starting_states.html
17:45 abe_music the state you write to remove the lock would require the state for verifying the required mount point
17:45 kermit joined #salt
17:46 shaggy_surfer right, is there an example of a require mount point for a file?  I see example of require group, pkg, etc
17:46 ghartz joined #salt
17:49 druonysuse joined #salt
17:49 druonysuse joined #salt
17:51 anuvrat joined #salt
17:51 nkuttler shaggy_surfer: - file: /mnt/point
17:53 googolhash joined #salt
17:54 schmutz joined #salt
17:54 dangra joined #salt
17:55 andrein Hi guys, i would like to use salt to replace /root/.my.cnf on a group of servers. I don't feel confortable putting that in a state, i'd like to do it one-off with a salt call. how can I do this?
17:58 giannello joined #salt
17:59 aw110f_ joined #salt
18:00 giannello I'm trying to use a gitfs remote which contains /formulas and /pillar folders, and I've configured the master with the right gitfs_remote (from the salt master I can clone the repository) and "gitfs_root: formulas"
18:01 vbabiy joined #salt
18:01 giannello every time I try to execute a salt 'target' state.highstate, the job is not executed, and it hangs forever
18:01 giannello if I move the files from the git repo to /srv/salt, it works. If I use a file:// URL as the git repo (manually cloned), it still hangs
18:02 giannello is there anything I'm missing or it could be a bug?
18:02 giannello I've used other gitfs repos, without gitfs_root without any problem
18:03 ipalreadytaken joined #salt
18:04 brianhicks joined #salt
18:04 pydanny joined #salt
18:04 whg joined #salt
18:08 Furao joined #salt
18:14 kmshultz joined #salt
18:14 Ahlee Thoughts: [salt.minion      ][ERROR   ] The return failed for job 20140513131241082661 could not find master/primary
18:17 bhosmer joined #salt
18:20 kmshultz does anyone know if there was ever traction on "definitions" in salt as per https://github.com/saltstack/salt/issues/3460 ?
18:21 kmshultz i.e. other than the pydsl example posted in that issue
18:25 Networkn3rd joined #salt
18:26 timoguin doesn't look like it
18:27 kmshultz of course no new comments in that issue, but didn't know if this was ever added and the issue was never closed
18:30 ashw7n joined #salt
18:30 timoguin yea i don't think so
18:32 kmshultz ok just figured i'd check! sometimes things go undocumented :)
18:39 ifnull joined #salt
18:43 mateoconfeugo joined #salt
18:48 dccc joined #salt
18:49 schimmy joined #salt
18:52 schimmy1 joined #salt
18:53 fragamus joined #salt
19:01 kermit joined #salt
19:01 magenbrot joined #salt
19:01 Whissi joined #salt
19:02 djaime joined #salt
19:03 Ryan_Lane joined #salt
19:04 jimklo joined #salt
19:04 whiteinge Ahlee: multi-master? did one of the masters go down during the job?
19:07 ipmb joined #salt
19:08 Ahlee no, so no
19:09 Ahlee i believe it's actually it's unable to comuincate with the returner
19:09 Ahlee so that's an interesting timeout exception being recorded
19:09 londo_ joined #salt
19:10 meteorfox joined #salt
19:10 mprymek joined #salt
19:12 schmutz joined #salt
19:13 kryss joined #salt
19:15 kryss hello, i've got a problem, i'm making a state with jijnja like this: - text: "ListenIP= {{ salt['network.ip_addrs'](None,False,'192.168.0.0/16') }}" and i have a result like this: ListenIP= ['10.9.33.10']  how to remove brackets?
19:20 KennethWilke left #salt
19:20 KennethWilke joined #salt
19:24 catpig joined #salt
19:25 linjan joined #salt
19:26 MindDrive joined #salt
19:27 ecdhe joined #salt
19:28 jnials joined #salt
19:28 tligda kryss: I'm not sure I understand, but I took a shot at it. Look at http://pastebin.com/8m60hWhi
19:28 krichardson joined #salt
19:29 whiteinge Ahlee: odd error message. new to me anyway
19:29 tligda The brackets indicate you are getting a python list object. You can use the [0] in the example I gave you to return the first item in the list, which should be the IP address you are after without the brackets.
19:29 nkuttler kryss: loop over your list and output the elements
19:29 whiteinge Ahlee: i don't see that message (via a very quick search) in the salt source. any idea where it's coming from?
19:32 tligda kryss: It should work as long as you are confident you will not get a list with more than one item.
19:32 nkuttler kryss: though jinja probably has some join filter or something
19:33 kryss thanks for advices
19:33 kryss tligda: it returns an error: failed: Jinja variable list object has no element 0;
19:33 bhosmer joined #salt
19:35 kryss oh, sorry, my mistake, it work like a charm!
19:35 kryss thanks :)
19:35 tligda kryss: Awesome. If you need more help, just pastebin the whole file so we can see the whole thing.
19:35 repl1cant joined #salt
19:35 kryss ok, i will remember
19:37 ndrei joined #salt
19:39 Ahlee whiteinge: yeah, guessing it's trickling up from the mongo
19:39 allanparsons joined #salt
19:40 ipmb joined #salt
19:40 allanparsons qq - how do i negate in a top file for grain data?
19:41 allanparsons i.e.  'G@ec2_apps:*myapp* and not  'webserver' in grains.get('ec2_roles', [])|default(False)':
19:48 jimklo joined #salt
19:49 Ryan_Lane joined #salt
19:51 whiteinge ah
19:55 gildegoma joined #salt
19:59 gothix how would i include states from one envirnment into another
20:00 gothix i am trying to get roles working some of my states that get should get include in a role are in my base tree
20:00 whiteinge gothix: there's an example mid-way down in this doc:
20:00 whiteinge http://docs.saltstack.com/en/latest/ref/states/include.html
20:00 rjc joined #salt
20:02 jnials joined #salt
20:05 meteorfox joined #salt
20:05 gothix whiteinge, thanks :)
20:07 frozenice joined #salt
20:08 meteorfo_ joined #salt
20:09 adizam joined #salt
20:09 XenophonF joined #salt
20:10 nrgaway joined #salt
20:12 schimmy joined #salt
20:15 schimmy1 joined #salt
20:18 nineteeneightd Hmmmm
20:19 XenophonF Indeed
20:19 nineteeneightd I'm getting a bunch of "failed to read virtual function for module" around a number of my custom modules after upgrading to 2014.1.4
20:20 nineteeneightd Any ideas?
20:20 nineteeneightd Now, I'll be honest, my "upgrade" process was to just utilize the salt-bootstrap.sh to install the newer version...I'm starting to think that might not have been the best idea
20:20 Recon_ joined #salt
20:24 ekristen is there any way to include mulitple pillar files that have the same root element for the same minion?
20:25 ekristen IE, globally for a development environemtn I want to define an array of users in pillar format, but for a single host I want to add an additional user
20:25 XenophonF nineteeneightd: how difficult would it be to re-install salt?
20:25 XenophonF just the python module, not your configs
20:27 nineteeneightd I mean, what do you think that entails? A little more background: I've been using the salt-bootstrap.sh to install from git.
20:28 XenophonF i mean switching to the native package (deb/rpm/etc.)
20:28 ReconD joined #salt
20:28 XenophonF i don't know how salt-bootstrap works, am assuming it uses pip or easy_install?
20:28 nineteeneightd I was actually just about to give that a shot
20:29 XenophonF i think you can do something like "python setup.py install --record files.txt" to get a list of what salt installs
20:29 XenophonF then cat that to xargs rm -f and everything should get deleted
20:30 XenophonF might want to circle back and clean up the module directories too
20:30 schimmy joined #salt
20:30 ashw7n joined #salt
20:30 Ryan_Lane joined #salt
20:30 XenophonF on my computer (FreeBSD), python packages get installed to /usr/local/lib/python2.7/site-packages/
20:30 XenophonF and i see both salt/ and salt-2014.1.4-py2.7.egg-info/ as sub directories
20:30 XenophonF if i had to do it manually, i'd delete those first
20:31 XenophonF then re-install the native package
20:31 XenophonF using apt-get/dpkg/yum/rpm/pkg/etc.
20:31 supplicant joined #salt
20:31 XenophonF i dunno
20:32 schimmy1 joined #salt
20:32 XenophonF depends on what's causing the error, but i'm guessing a messed up install
20:32 supplicant I'm trying to figure out how to write part of a config file in yaml by pretty must just taking a nested dictionary from pillar. is there a "right" way to serialize the dictionary from pillar into yaml?
20:32 XenophonF iirc there's a "pip uninstall" command, too
20:33 XenophonF might have to do "pip freeze" first to get a list of installed packages
20:35 sroegner joined #salt
20:37 supplicant ah, http://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.jinja.html#salt.utils.jinja.SerializerExtension
20:41 mproper joined #salt
20:47 mgw1 joined #salt
20:49 rupsky________ joined #salt
20:52 anuvrat joined #salt
20:53 XenophonF left #salt
20:54 martoss joined #salt
21:04 [diecast] joined #salt
21:05 schmutz joined #salt
21:07 trevorj joined #salt
21:08 mgarfias joined #salt
21:09 Gareth joined #salt
21:11 nliadm joined #salt
21:11 berto- joined #salt
21:12 Ryan_Lane joined #salt
21:19 scooby2_ joined #salt
21:22 jgarr I'm trying to use pillar to grab data from nested states. what's the best way to test the output without running the states?
21:24 jgarr currently I have pillar.get('foo:bar:baz') but I think that only applies to values inside the foo state right? What if it's not in the init.sls file?
21:34 ashw7n joined #salt
21:36 uestra joined #salt
21:37 jgarr joined #salt
21:40 Ryan_Lane salt is the 3rd most used non-openstack specific tool for deployment, but ranks one result lower than ansible for application developers: https://twitter.com/SquidDLane/status/466314650402172929
21:40 Eugene joined #salt
21:40 Ryan_Lane thought I'd share that with you guys ;)
21:42 Ryan_Lane it's higher in proof of concepts than in others, which is probably a good sign.
21:43 Ryan_Lane whoops. by 14 results for the app developer survey
21:43 Ryan_Lane which likely means openstack needs better salt-cloud support
21:43 jnials That's because developers don't have to maintain their own crap. ;)
21:44 jnials They throw it over the wall to the sysadmins.
21:44 Ryan_Lane most developers are also operations now
21:44 Ryan_Lane lots of orgs are embracing the "if you build it you run it" model
21:44 Ryan_Lane so we want to be #1 in the dev results
21:45 jnials Yup, which is why I get paid so much money to fix their problems.
21:45 yurei joined #salt
21:45 jnials which is also why I dumped ansible for salt in my last job.
21:47 jnials Salt is more maintainable, ansible is quicker for first deployment.  But once you have to update something complex Salt is much much more robust.
21:48 kermit joined #salt
21:49 Ryan_Lane it's important to figure out why salt is harder initially than ansible and to address that. initial impression means a lot.
21:49 jnials That last idea should be part of Salt's evanbelization effort.
21:49 Ryan_Lane evanbelization?
21:49 jcockhren haha
21:49 jnials Heh.  Bad spelling, new keyboard.
21:50 * jcockhren waits for definition
21:50 jnials Evangelization.
21:50 Ryan_Lane evangelism?
21:50 timoguin sounds bad.
21:50 Ryan_Lane you can evangelise all you want, but if people's first impression when they try it is bad, then they'll leave
21:50 vbabiy joined #salt
21:50 timoguin i don't want to evilangelize, but i would like to help make it easier for people getting started with Salt
21:51 meteorfox joined #salt
21:51 DaveQB joined #salt
21:52 steveoliver joined #salt
21:53 jnials Salt defines the state you want to get to, Ansible lets you get there like were doing it manually.  They both end up at the same end state.
21:53 steveoliver joined #salt
21:54 Ryan_Lane if you use ansible's playbooks and modules there's no major difference
21:54 Eugene joined #salt
21:55 Ryan_Lane and the initial use of the playbooks is easier. we should make sure the initial states impression is stellar :)
21:55 pdayton joined #salt
21:56 jnials Not with the end result, but there is in how you get there.  If you have a complicated playbook, it can be difficult to decompose so as to change one thing at the end.  Whereas with salt you may just change a definition here or there.
21:56 timoguin most people don't look at complicated states/playbooks right off the bat
21:57 timoguin you'd tend to start with the most basic examples
21:57 Ryan_Lane yep
21:57 jnials Exactly.  I didn't either, which is why I initially discarded salt for ansible.
21:57 Ryan_Lane most people start with: I want to write this state, then I want to run this state
21:57 frosty996 left #salt
21:58 jnials I think having a good cookbook of examples would go a long way.  Not complaining about what's out there, but more is better. :)
21:59 Ryan_Lane and then the docs are like "so, install your master, then your minion, then put your configs in /etc/salt/<blah>, now restart your master and minion, now sign your minion key, now put your state in /srv/salt/<blah>, also make sure to add a top file, which includes the * host, now run this weirdly named command called highstate."
21:59 Ryan_Lane and the user says "but all I want to do is see what'll happen when I run this specific file I just wrote"
21:59 uestra joined #salt
21:59 eliasp gnah, deploying software on windows is like constantly getting hit in the balls… Saltstack makes it at least bearable
22:00 jnials And clear documentation.  Salts docs are not great, ansible's docs suck rocks.
22:00 cewood joined #salt
22:00 Ryan_Lane jnials: I'm often told the opposite by new users
22:01 Ryan_Lane I'm not downing on salt, btw. I'm one of the most active proponents ;).
22:02 jnials I got burned by the constant doc searching I had to do in ansible, and the constant changes (not improvements) they made as they went corporate.  Maybe they have improved since then.
22:03 jnials I gave up after trying to upgrade my ansible deployment to use roles and discovered nothing made sense.
22:03 pdayton joined #salt
22:05 DaveQB joined #salt
22:05 andrej joined #salt
22:11 diegows joined #salt
22:14 mr_chris joined #salt
22:15 l0x3py joined #salt
22:16 XenophonF joined #salt
22:18 stevednd Salt's docs would be helped a lot by defaulting docs.saltstack.com to the latest *release* version, and not just latest in develop. You glance to the side, and see "Latest Salt release: 2014.1.4", and miss the comment about it being for the dev version, and struggle in massive frustration trying to figure out why features that it claims are present, aren't there
22:19 stevednd make getting the latest develop docs a deliberate act, as I would think most users would be using official releases in their package managers and such
22:19 * jnials agrees
22:20 elfixit joined #salt
22:20 pydanny joined #salt
22:20 stevednd or if not that, then have a big ass banner across the top stating that docs.saltstack.com are for the latest development, and present a link to the latest release docs
22:20 jcockhren stevednd: yep. like django's docs
22:21 ashw7n joined #salt
22:22 pydanny joined #salt
22:23 stevednd jcockhren: not familiar with django's docs, but assuming they're like most other sane project docs, then yes.
22:24 stevednd I think linking to latest was probably a good strategy when salt was younger, but as it has matured more it probably needs to shift
22:24 jcockhren https://docs.djangoproject.com/en/dev/
22:25 jcockhren look at the top
22:25 stevednd yeah, that's pretty obvious
22:25 jcockhren on the bottom right, select any older release, it goes away
22:25 stevednd and even just going to django's site and clicking documentation takes you to the docs for the latest release
22:25 anuvrat joined #salt
22:26 [diecast] joined #salt
22:26 linjan joined #salt
22:27 ccase joined #salt
22:27 linjan joined #salt
22:36 sroegner joined #salt
22:36 Ztyx joined #salt
22:36 alunduil joined #salt
22:37 ajprog_laptop joined #salt
22:37 stevednd https://github.com/saltstack/salt/issues/12754
22:37 stevednd I just created an issue for changing the documentation presentation if anyone wants to +1 that, or chime in on it
22:39 jerrcs joined #salt
22:41 linjan joined #salt
22:42 eliasp is there a way to have a "no-op" pkg? http://pastebin.kde.org/pmhf4lzk0
22:42 eliasp see lines 19..22
22:43 eliasp I need to require "endnote-upgrade-state" which only does something in case an upgrade is available…
22:43 eliasp in case one is available, it is rendered as "pkg", but in case not… nothing should happen… how to deal with this?
22:44 attrition joined #salt
22:44 jgarr does highstate every run automatically if a minion is connected to a master? or is it something that is always scheduled or triggered?
22:45 attrition left #salt
22:45 jgarr I'm pretty new to salt. more used to puppet scheduling
22:45 eliasp jgarr: http://docs.saltstack.com/en/latest/topics/jobs/schedule.html
22:46 eliasp jgarr: by default, Minions won't do anything…
22:46 eliasp unless you configure a schedule, you'd have to run something like "salt some-minion state.highstate" manually
22:46 jgarr eliasp: k, that's what I was trying to figure out (which is great for me right now testing things in prod)
22:50 sm1ly joined #salt
22:51 jgarr another beginner question I didn't see in the docs...my state tree is /srv/salt/{states,pillar}/{prod,dev}. Should my top.sls go in the states folder or in /prod and /dev?
22:51 sm1ly re2all. can anyone tell me plz, when I using salt on 20 or more hosts for deploy some soft+configs, how I can see only errors?
22:51 jgarr everything I seen references says put it in /srv/salt/states/ but I'm not sure how to specify which environment to pull the right state file from
22:59 shaggy_surfer joined #salt
23:00 shaggy_surfer wow… I have recure on a file.directory and it's really slow….. is this normal?
23:00 shaggy_surfer recurse*
23:00 shaggy_surfer doing a user and group mod
23:07 TyrfingMjolnir joined #salt
23:10 jnials joined #salt
23:15 nkuttler any guesses why a minion would return 'virt.query' is not available.? other virt commands succeed
23:20 TyrfingMjolnir joined #salt
23:20 manfred sounds like it is erroring on the minion
23:20 manfred try using salt-call -l debug on the minion and see if you get a python error
23:22 hunter joined #salt
23:22 nkuttler manfred: yeah, i had a look and there are several error, but i guess i'll have to look at the salt source
23:23 hunter joined #salt
23:24 to_json joined #salt
23:33 hunter joined #salt
23:38 younqcass joined #salt
23:38 whiteinge !salt.states.test
23:38 whiteinge :(
23:40 whiteinge oh
23:40 whiteinge !states.test.succeed_without_changes
23:40 wm-bot4 http://docs.saltstack.com/en/latest/ref/states/all/salt.states.test.html#salt.states.test.succeed_without_changes
23:40 forrest awwwwwwwwww shiiiiiiiiiiiiiiiiiiiiiiit
23:40 forrest dat documentation
23:40 whiteinge ^^ eliasp that module can be used for no-op stuffs
23:40 forrest how long has that bot been in here?
23:40 robawt you two are funny
23:41 whiteinge forrest: it's been in for a week or so, but not fully seeded until yesterday-ish
23:41 forrest !states.pkg.installed
23:41 wm-bot4 http://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkg.html#salt.states.pkg.installed
23:41 forrest whiteinge, nice
23:41 forrest that's very cool
23:41 whiteinge see also: http://bots.wmflabs.org/~wm-bot/dump/%23salt.htm
23:41 whiteinge stats!
23:41 whiteinge (at the bottom)
23:42 forrest very cool
23:43 fragamus joined #salt
23:43 whiteinge hm. we need links to the module-level
23:45 forrest whiteinge, does that bot only record real time since it was activated? I was thinking about parsing the logs to see what the values were there.
23:51 whiteinge forrest: yeah, it keeps it's own logs afaik
23:52 forrest gotcha
23:57 ipalreadytaken joined #salt
23:58 jab416171 I made it on the top 100? whoa that's crazy
23:58 jab416171 oh is the bot only since like this month?
23:58 forrest yea
23:59 jab416171 still neat
23:59 nkuttler not much traffic in here anyway

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