Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-01-29

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

All times shown according to UTC.

Time Nick Message
00:00 jalaziz_ joined #salt
00:00 stanchan joined #salt
00:03 shaggy_surfer joined #salt
00:05 Guest3774 joined #salt
00:08 robm Thanks for the response number5, I am using it already, but was curious if there was something for limits. I will use the method that iggy mentioned.
00:08 aron_kexp joined #salt
00:13 ajw0100 joined #salt
00:15 vu joined #salt
00:16 theologian joined #salt
00:16 ekristen joined #salt
00:17 otter768 joined #salt
00:18 dude051 joined #salt
00:19 dude051 joined #salt
00:23 viq joined #salt
00:24 peters-tx joined #salt
00:25 scott2b joined #salt
00:27 andrej What could be causing an intermittent behaviour where the master will sometimes echo out results for a salt command, and sometimes it won't?
00:29 conan_the_destro joined #salt
00:29 lesel left #salt
00:29 dusel joined #salt
00:31 hal58th minions aren't responding fast enough?
00:31 bluenemo_ joined #salt
00:32 aqua^mac joined #salt
00:33 scott2b joined #salt
00:34 iggy yeah, throw a couple test.pings in there first
00:36 andrej they come back within 0.995 - 1.005 seconds
00:37 andrej I keep seeing message authentication failures on both the master and the minion(s)
00:38 iggy check issues on github and see if any of the common fixes help that
00:38 andrej heh
00:40 GabLeRoux joined #salt
00:45 scott2b joined #salt
00:45 bhosmer_ joined #salt
00:59 Tritlo Would you recommend ansible or salt for personal use?
01:03 KyleG joined #salt
01:03 KyleG joined #salt
01:03 markmari_ I would like to use gitfs to transfer a dir tree to some machines, so my salt code is /srv/www: file.recurse: - source: salt://tos/www
01:03 markmari_ is that possible? when I do a salt call salt is reporting that none of the sources are there
01:05 jimklo_ joined #salt
01:05 markmari_ there aren't any sls files in that machine, but I'm trying to keep from installing git and ssh keys, etc. on these webservers
01:06 markmari_ and I want my files inside the VPC when the box comes up so they come online faster
01:12 jer_ joined #salt
01:17 tkharju joined #salt
01:20 aqua^mac joined #salt
01:26 __number5__ file.recurse can only copy stuff *inside* your salt states tree
01:28 scott2b joined #salt
01:28 mohae joined #salt
01:29 GabLeRoux joined #salt
01:33 fxhp cmd.append('{0}'.format(repository)) <- that sort of looks funky
01:34 fxhp could we just cast to str?
01:34 fxhp is this a style thing?
01:36 aparsons joined #salt
01:36 ZafarHussaini joined #salt
01:43 nitti joined #salt
01:46 nethershaw joined #salt
01:49 keeth joined #salt
01:49 zanhsieh joined #salt
01:50 eliasp joined #salt
01:52 shaggy_surfer joined #salt
01:55 druonysus joined #salt
01:58 yetAnotherZero joined #salt
01:58 yetAnotherZero joined #salt
02:00 msheiny joined #salt
02:02 cberndt joined #salt
02:03 scott2b left #salt
02:04 jasonrm joined #salt
02:07 yomilk joined #salt
02:08 brendanashworth joined #salt
02:08 brendanashworth left #salt
02:13 fllr joined #salt
02:16 ckao joined #salt
02:17 otter768 joined #salt
02:20 schlueter joined #salt
02:22 schlueter1 joined #salt
02:26 speed145a joined #salt
02:40 eligos joined #salt
02:40 fxhp https://github.com/russellballestrini/salt/commit/bc517c0e1c73a8f6f19fae8da2aa6255b7b5090c
02:41 fxhp some reason cmd.run doesn't let this work
02:45 s51itxsyc joined #salt
02:45 dude051 joined #salt
02:45 dude051 joined #salt
02:47 ilbot3 joined #salt
02:47 Topic for #salt is now Welcome to #salt | SaltConf 2015 is Mar 3-5! http://saltconf.com | 2014.7.1 is the latest | Please be patient when asking questions as we are volunteers and may not have immediate answers | Channel logs are available at http://irclog.perlgeek.de/salt/
02:50 eightyeight joined #salt
02:52 echtish joined #salt
02:57 yomilk_ joined #salt
03:03 mohae joined #salt
03:04 yetAnotherZero joined #salt
03:04 markmarine joined #salt
03:07 Furao joined #salt
03:09 markmarine joined #salt
03:12 TyrfingMjolnir joined #salt
03:13 elfixit1 joined #salt
03:15 yetAnotherZero joined #salt
03:15 fllr joined #salt
03:16 jeddi I'm trying to do a file.managed for a file with utf-8 characters in it - a '.toprc' file as it happens.  I appear to be hitting this bug:  https://github.com/saltstack/salt/issues/16651  ... I can't really bump up to the most recent version on master and minions.  Any clues on a workaround for managing that file in the short-term?
03:17 jeddi The traceback I'm seeing (with file.managed) is:  http://rn0.me/show/DZen0JLjSis8GUzFDZck/
03:19 Mso150 joined #salt
03:23 CeBe1 joined #salt
03:32 tligda joined #salt
03:34 battaglia joined #salt
03:37 Mso150_e joined #salt
03:38 markmarine joined #salt
03:39 Mso150 joined #salt
03:41 forrest joined #salt
03:42 markmari_ joined #salt
03:44 tligda joined #salt
03:44 schristensen joined #salt
03:45 agend joined #salt
03:49 Furao joined #salt
04:00 jer_ joined #salt
04:04 bhosmer joined #salt
04:05 cotton joined #salt
04:17 jerematic joined #salt
04:19 otter768 joined #salt
04:20 druonysus joined #salt
04:21 monkey66 joined #salt
04:25 vbabiy joined #salt
04:36 monkey66 left #salt
04:38 jimklo_ joined #salt
04:46 Ryan_Lane joined #salt
04:48 aurynn joined #salt
04:48 Micromus joined #salt
04:51 cberndt joined #salt
05:00 tomspur joined #salt
05:02 smcquay joined #salt
05:03 smcquay joined #salt
05:13 gmoro joined #salt
05:13 timbyr_ joined #salt
05:19 ajw0100 joined #salt
05:22 ajw0100 joined #salt
05:31 yomilk joined #salt
05:34 rome_390 joined #salt
05:36 felskrone joined #salt
05:48 joehh2 debian packages for 2014.7.1 going up to debian.saltstack.com now
05:53 joehh2 ubuntu going up to ppa shortly
05:54 __number5__ cool, thanks joehh2
06:01 Saltn00b joined #salt
06:06 Saltn00b Anyone hanging out here tonight? I'm running into a problem with a new installation of Salt that I was hoping someone could help with.
06:07 Saltn00b I set up salt-ssh and it was working great. After playing around with it, I used the bootstrapper to install the full version onto my master, so that I could try out installing the agent on my minions.
06:07 Saltn00b Ever since I did the full install on the master, though, salt-ssh doesn't work. I get:
06:07 Saltn00b NameError: global name 'msgpack' is not defined
06:08 Saltn00b I know this is a known bug that's fixed in the latest version, but I can't get it to disappear. I've tried reinstalling, but no luck.
06:08 Saltn00b Is there a way to completely uninstall salt, then reinstall?
06:11 jeddi Saltn00b: what distro are you on?  if you're on debian or derivative - dpkg --purge salt-minion salt-common should do it.
06:11 jeddi i'm assuming salt state and pillar files are stored in a dvcs, somewhere safe.
06:11 jcockhren "safe"
06:11 jcockhren sorry. couldn't resist. ;)
06:12 jeddi safer than on the system you're about to purge stuff from.
06:13 jcockhren yep.
06:13 jeddi safety is always relative.  :)
06:13 Saltn00b I'm on Amazon Linux. Same packages with yum?
06:13 Saltn00b And yes, my config files are safe. :)
06:13 jeddi dpkg purge *usually* leaves data files around, and from memory usually snots config files unless they've been changed ... but it's still something i'm wary of.
06:13 jeddi I'm not sure what Amazon Linux means.
06:14 jeddi If you mean EC2, then that doesn't define the distro.
06:15 mikeywaites joined #salt
06:15 Saltn00b It's the Amazon Linux distro on EC2 (as opposed to the other distros they offer). It's their own CentOS-based distro.
06:16 jeddi aha ... okay.   been a while since i used ec2.
06:16 jeddi Saltn00b: not sure what the yum equiv for dpkg --purge would be - but it can't be hard to determine?  have you already tried flushing the entire contents away?  sounds like you have.
06:17 Saltn00b No worries. I appreciate the help!
06:17 Saltn00b That's the next step. I just wanted to make sure that I wasn't missing something obvious, like "install_salt.sh --uninstall" or something....
06:18 Furao joined #salt
06:19 otter768 joined #salt
06:19 timbyr_ joined #salt
06:20 favadi joined #salt
06:25 calvinh joined #salt
06:27 mikkn joined #salt
06:28 Furao joined #salt
06:28 Pixionus joined #salt
06:29 jeddi yeah - stick with the package management tools first ...
06:31 Saltn00b Will do, thanks. Another quick question -- the bug report on this issue describes a workaround where they cleared the thin client code on the minions. Is there a command to force a thin client update, or do I have to ssh into each of the devices and delete the cached version?
06:32 jeddi ooh - no idea, sorry.   i've not used the salt-ssh stuff at all yet.
06:32 jeddi i'm a strictly master/minion kind of chap.
06:32 jeddi simple life, and all that.
06:33 Saltn00b I'm starting to think that may be the smart way to go.... :)
06:34 markmarine joined #salt
06:34 jeddi i think on the provisioning front, there's some advantages to salt-ssh ... but ... i really haven't looked into it, as i say, or automating provisioning with salt at all, in fact.
06:36 Saltn00b The nice thing about salt-ssh is that I was able to set it up on my master and immediately talk to minions, without having to install anything on them.
06:37 Saltn00b They're behind firewalls and connecting to the master via ssh tunnels, so I wrote a roster plugin to get the list of connected devices and dynamically generate the roster.
06:37 markmari_ joined #salt
06:37 Saltn00b I figured I'd get up and going like that, then set up the minions using salt-ssh.
06:38 jeddi the master is behind a firewall, or the minions are?  the minions only need to be able to talk to the master , remember.
06:38 Saltn00b Only the minions are behind the fw. They can contact the master fine.
06:39 jeddi ah, then they'll be fine to 'phone home' and use vanilla master / minion comms (no need for salt-ssh)
06:39 calvinh joined #salt
06:39 Saltn00b Yes, that's the eventual goal. I just wanted to kick the tires without having to install agents on all the minions.
06:39 jeddi though there are other reasons for using salt-ssh apparently.  as I say, have'nt used it, haven't needed to.
06:39 jeddi ahh - gotcha.
06:40 Saltn00b I was actually trying out the minion install when the problem happened. Unfortunately, the "upgrade" from just salt-ssh to the full salt install seemed to torpedo the thin client.
06:40 Saltn00b I'm still hopeful it's pilot error, though, and not something more insidious with running both at once.
06:40 jeddi i wouldn't think so ... you said there's a known bug, is that relating to conflicts with salt-ssh and salt-minion?
06:40 jeddi concurrent, same client machine, i mean?
06:41 Ryan_Lane joined #salt
06:42 Saltn00b https://github.com/saltstack/salt/issues/7913
06:42 jeddi I'm stuck trying to work around this utf-8 problem where file.managed tries to show a diff of the file, and chokes ... weirdly file.managed works just fine on straight up binary files without choking.  wonder if there's a way of telling file sls to not try to be smart about it (just do a binary diff, or whatever python equiv it's doing under the hood).  nothing doc'd.
06:42 Saltn00b The thin client was expecting the msgpack module, which won't necessarily be there.
06:43 jeddi aha... msgpack, yeah, rings a bell.
06:43 jeddi not as much fun as it used to be to get zmq running on a raspberry pi ... but fun nonetheless, I'm sure.
06:43 Saltn00b Funny, that's exactly what my minions are....
06:44 jeddi ah.
06:44 Saltn00b The Pi side worked swimmingly, though. I installed Salt on one of them at the same time I upgraded the master, and afterwards that was the only one that would respond.
06:45 Saltn00b No doubt because Salt installed msgpack.
06:45 jeddi i used to have four of them ... but kept on giving them away to relations to run xbmc on (etc) ... i need to order another handful.    the way i've done my state files means it takes ages to compile and then implement initial runs on those poor little devices.
06:45 jeddi you're using raspbian wheezy?
06:45 jimklo_ joined #salt
06:46 Saltn00b Yes, for now. We may move to a custom Arch build eventually, but for now Wheezy is getting it done.
06:47 calvinh_ joined #salt
06:47 jeddi ack.
06:47 Saltn00b We've got about 2 dozen deployed in our customers' houses. That's why I'm looking at Salt -- it's quickly becoming a pain to ssh into each of them every time we need to do an upgrade.
06:47 jeddi i could imagine!
06:49 ajw0100 joined #salt
06:49 stoogenmeyer_ joined #salt
06:49 calvinh_ joined #salt
06:49 jeddi i'm starting ona  project to run a half dozen of them around the perimeter of the property with onboard cameras and 'motion' package , to keep an eye on things, as it were.   been pondering (non-salt) problem of distributing everything each unit sees as soon as it sees it, in case that one device is compromised.
06:50 jeddi i'm pre-configuring with salt - though it's probably as easy to just make a gold 8gb sd card build and replicate then change hostname / ip - but this way i can periodically upgrade them.
06:50 jeddi but, yeah, as i say, watch out for the massive performance hit you take when your recipes get lengthy.  i've bumped the timeouts up a bit for minions to respond.
06:50 krelo joined #salt
06:51 felskrone Saltn00b: depending on what you need to to do, dsh might help you already and its easier to setup than salt http://www.netfort.gr.jp/~dancer/software/dsh.html.en
06:52 felskrone it nothing compared to salt, but you can manage groups of hosts
06:55 Saltn00b We're doing the golden image SD card thing right now. It's fine for bootstrapping the business, but exponentially grows into a huge pain.
06:56 calvinh joined #salt
06:56 joehh2 packages for ubuntu precise and trusty have gone to launchpad ppa
06:56 Saltn00b And thanks for the advice -- I'll keep an eye on the timeouts. As it is, even the testing I was doing with salt-ssh was slow. Ping responses were taking 20-30s. Another reason why I wanted to try using minion agents.
07:00 jeddi Saltn00b: you may find the slightly hacky way of using salt to push out major stuff primarily, and perhaps ship out sh / py / etc scripts that then do the actual work ... may be slightly more performant.  but, well, i've found once the boxes are configured, performance isn't so bad.
07:01 jeddi felskrone, Saltn00b: that might be a good way of doing the initial provisioning
07:02 colttt joined #salt
07:03 smkelly_ joined #salt
07:04 felskrone yeah, maybe :-)
07:08 AndreasLutro joined #salt
07:10 Saltn00b felskrone: Thanks for the tip. I'm looking at that site now, and I'll definitely download it and try it out.
07:10 TheThing joined #salt
07:11 stoogenmeyer_ joined #salt
07:12 Saltn00b jeddi: That makes sense. We don't do too many changes to device configuration. It's primarily code updates, and those are not too big; the overhead of doing all of the connections is the biggest concern. Salt is showing promise for that, and I've already been able to pull some pretty useful reporting information using salt-ssh.
07:13 krelo joined #salt
07:14 jeddi i do some code distributions, but usually pulling (at the minion) from a read-only gitolite repo ... rather than distributing files / packages directly from the salt master.  still, single location, i guess.
07:15 markmarine joined #salt
07:15 Saltn00b I'm outta here gentlemen. Thanks again for the help!
07:22 bfoxwell joined #salt
07:26 KermitTheFragger joined #salt
07:26 Furao joined #salt
07:31 flyboy joined #salt
07:31 slafs joined #salt
07:33 Morbus joined #salt
07:35 mikeywaites joined #salt
07:35 krelo joined #salt
07:37 toanju joined #salt
07:38 slafs left #salt
07:38 Auroch joined #salt
07:39 laax joined #salt
07:47 jerrcs joined #salt
07:48 glyf joined #salt
07:50 ralala joined #salt
08:03 hebz0rl joined #salt
08:04 smkelly I'm running 2014.7.0 on FreeBSD 10 and have a state {'zabbix_agentd': {'__env__': 'base', '__sls__': 'zabbix', 'service': [{'enable': True}, {'restart': True}, {'require': [{'pkg': 'zabbix24-agent'}]}, 'running', {'order': 10001}]}. If the service is enabled in rc.conf but not running, teh state doesn't start the service but just errors with: Service zabbix_agentd is already enabled, and is dead
08:04 smkelly Any idea what I'm doing wrong?
08:05 bhosmer_ joined #salt
08:06 smkelly hm, hold that thought. the service may be failing to start
08:06 smkelly yup
08:06 smkelly question retracted; I'm dumb.
08:11 fredvd joined #salt
08:12 lb1a joined #salt
08:15 timbyr_ joined #salt
08:15 gmoro joined #salt
08:17 malinoff joined #salt
08:18 krelo joined #salt
08:20 otter768 joined #salt
08:23 kawa2014 joined #salt
08:24 TheThing joined #salt
08:28 eseyman joined #salt
08:32 _blackjid joined #salt
08:34 egil I have a situation where I installed develop on my master. Now I wan't to back to stable but I'm having some issues ..
08:35 egil 1) I tried bootstrapping stable version. salt --versions-report still showed develop
08:35 egil 2) I unistalled the master and tried to bootstrap it again
08:35 egil now salt-master refuses to startup
08:35 egil any tips?
08:36 zadock joined #salt
08:43 bluenemo joined #salt
08:43 bluenemo joined #salt
08:48 intellix joined #salt
08:52 trikke joined #salt
08:54 krelo joined #salt
08:59 jeddi 'yellow denotes a future expected change in configuration' -- does this mean syntax, or that my state files are going to change something?   because the text next to each of my yellow paragraphs is 'is in the correct state', which indicates that entity won't be adjusted.
09:00 jeddi egil: tail -f /var/log/salt/master &   .... and    /usr/bin/salt-master
09:00 karimb joined #salt
09:00 jeddi oh, maybe 'which salt-master' - to confirm you are running the one you think you are.  and/or salt-master --version
09:03 catpig joined #salt
09:07 jtang joined #salt
09:07 krelo joined #salt
09:07 dkrae joined #salt
09:09 grrrrr I'm testing salt-ssh and just can't get it working... followed this tutorial http://www.giantflyingsaucer.com/blog/?p=5061 and getting "No hosts found with target * of type glob" after running "sudo salt-ssh '*' test.ping"
09:09 trikke joined #salt
09:10 grrrrr any ideas?
09:10 grrrrr sudo salt-ssh --roster
09:10 grrrrr heheesa
09:13 egil jeddi: the problem is, now salt-master wont start and there is No log entry
09:14 I3olle joined #salt
09:16 joehh2 egil: is a master still running?
09:16 joehh2 ps or netstat to see if something is listening on the port
09:17 wnkz joined #salt
09:21 egil joehh2: no ports are taken, and no salt services running
09:23 egil running /usr/bin/salt-master i get a stacktrace:
09:23 egil ImportError: No module named cli.caller
09:24 N-Mi joined #salt
09:24 N-Mi joined #salt
09:30 jeddi egil: so you're sure master's not running?  then what happens if you bump up logging level to debug (/etc/salt/master) and then launch salt-master directly in the foreground (/usr/bin/salt-master ... as opposed to init.d / service / systemctl)
09:30 ramteid joined #salt
09:30 jeddi aha - got it - stacktrace.
09:30 egil jeddi: it wont start
09:30 jeddi sorry - catching up slowly :)
09:31 jeddi did you remove the packages via your package manager?   and then check if there's any salt remnants on the box?
09:31 egil I tried removing using apt, some files remained though
09:31 egil I can try and delete everything again
09:35 davidone joined #salt
09:35 davidone hi all
09:35 davidone is there a way to force a propagation to minions of files managed by salt?
09:37 jhauser joined #salt
09:42 joehh2 egil: that is probably best
09:42 joehh2 davidone: file.managed?
09:42 jeddi egil:  dpkg --purge is your friend.
09:43 egil joehh2: Think I've managed to remove everything now (wow, does salt leave traces everywhere), but do you think it's best to use bootstrap or just use apt?
09:43 egil jeddi: thanks!
09:43 egil never thought of that one
09:50 davidone joehh2: salt minion state.sls file.managed?
09:51 jeddi egil: apt ... always package management tools if you have the option.
09:51 felskrone joined #salt
09:52 jeddi davidone:  -->  http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.managed
09:53 egil jeddi: yeah, but doesn't bootstrap try to use package manager if available?
09:54 mikkn joined #salt
09:55 egil jeddi,joehh2: thanks guys, its working again now
09:56 egil Salt: 2014.7.1
09:56 egil now to try and update minions :)
09:56 jeddi egil: glad to hear it. :)
09:57 jeddi egil:  ... not sure about bootstrap's inner workings.  haven't really had to defer to it.
09:57 egil jeddi: I'll remember dpkg --purge though :)
09:58 jeddi egil: dpkg --purge is a tremendously handy thing to have in your toolkit.   knowing to look at /var/lib/dpkg/info/*list  (where * is your package name) is also handy ...
09:58 davidone jeddi: so if I have something like this: http://pastebin.com/zbkheN5g
09:58 davidone which command shoud I run on the master to update /usr/local/etc/clamd.conf on minions?
09:58 jeddi won't necessarily show you generated configuration and data files, but will tell you if there's remnants sticking around.  similar to 'dpkg -l | grep -v ^ii'
09:58 davidone salt '*' state.sls file.managed /usr/local/etc/clamd.conf ?
09:59 egil jeddi: yeah, I saw that now when looking for traces of salt, so thanks for that. I learned something new :)
09:59 jeddi davidone: no, you'd refer to the state (file) name.   so, if that file you pasted was called 'foo.sls' you'd run 'salt '*' state.sls foo test=True'
09:59 jeddi and once you looked at the output, and see that it isn't doing something Surprising (!), you'd then run that again without the 'test=True' bit
10:00 jeddi davidone: does that make sense?
10:00 davidone so if I have that stanza in a bigger file with many instructions, I cannot 'isolate' the file.managed block
10:00 jeddi egil: apt-get / apt-cache do 95% of the things you need.   occasionally ... dpkg needs to be summoned.
10:01 jeddi davidone: no.   the trick typically there, as I understand it, is to break things up into separate files.  you can have sub-directories that refer to groupings of activities - for instance i have /srv/salt/base/ - and within that an 'init.sls' which includes a bunch of other files in there.  so I can, if I want, run state.sls base, or I can run state.sls base.root-config-files (for example).
10:02 jeddi of course, the counter to this approach is that dependencies need to be referenceable from within each file (or at least you need to be able to deal with on-screen complaints / errors if you've got a configuration type state, that refers bcak to a separate state file that may do the package installation, as one example)
10:03 jeddi davidone: but .. in general terms, break things into logical groupings, small groupings go into files in a sub-directory under /srv/salt, and larger groupings are those sub-directories.    of course, this is merely how I do it ... and may not be Best Practice <tm>.
10:03 jeddi dinner beckons.
10:11 Furao joined #salt
10:12 trikke joined #salt
10:13 akafred joined #salt
10:20 TyrfingMjolnir joined #salt
10:21 otter768 joined #salt
10:23 Cidan joined #salt
10:25 davidone jeddi: yeah, same feelings here
10:25 davidone ty
10:28 aquinas joined #salt
10:30 fxhp joined #salt
10:30 karimb joined #salt
10:31 karimb joined #salt
10:41 ekle joined #salt
10:41 ekle hi, i have an endless loop of "Running scheduled job: __master_alive" and "/usr/lib/python2.7/dist-packages/salt/modules/config.py:136: DeprecationWarning: pillar_opts will default to False in the Lithium release" after salt '*' status.master
10:41 ekle any suggestions ?
10:53 harkx joined #salt
10:54 teogop joined #salt
11:03 calvinh joined #salt
11:05 giantlock joined #salt
11:23 Grokzen joined #salt
11:27 che-arne joined #salt
11:27 AndreasLutro joined #salt
11:27 intellix joined #salt
11:33 hojgaard joined #salt
11:34 krelo joined #salt
11:38 zadock joined #salt
11:44 paulm- joined #salt
11:44 paulm- Can you say "require x OR y" instead of the usual "require x AND y"?
11:45 Morbus joined #salt
11:46 Zachary_DuBois joined #salt
11:46 bytemask joined #salt
11:48 APLU joined #salt
11:49 jeddi When I run a state.sls or highstate, there's a stack of yellow-paragraphs with 'this is in the correct state', and they obviously don't change anything.  Any idea why they're coming up yellow?  These are exclusively file.managed and file.recurse stanzas.
11:49 aqua^mac joined #salt
11:49 shel3over joined #salt
11:53 joehh2 jeddi: guessing you are running 2014.7.1 with test=true?
11:54 joehh2 probably also debian/ubuntu.... recently upgraded...
11:54 joehh2 I believe it is a regression in 2014.7.1
11:54 karimb joined #salt
11:55 joehh2 https://github.com/saltstack/salt/issues/18312
11:56 calvinh_ joined #salt
11:59 trikke joined #salt
11:59 _ether_ paulm-: depending of the structure of your sls files and how they are included, you may achieve this with require_in.
12:03 giantlock joined #salt
12:04 agend joined #salt
12:05 vbabiy joined #salt
12:07 bhosmer joined #salt
12:12 intellix joined #salt
12:13 felskrone joined #salt
12:15 colttt joined #salt
12:17 paulm- Is it possible to use mysql_database states with only salt-ssh?
12:18 JDog Hi. When I use the virtualenv salt state, how do I install the virtual environment as not the root user?
12:22 _ether_ JDog: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.virtualenv_mod.html#salt.states.virtualenv_mod.managed did you try the user kwarg?
12:22 otter768 joined #salt
12:23 JDog No - - have got that page up and didn't see that listed -- silly me. Thanks!
12:23 bhosmer joined #salt
12:26 ]V[ joined #salt
12:28 monkey661 joined #salt
12:39 Furao joined #salt
12:40 xsteadfastx joined #salt
12:44 wnkz joined #salt
12:45 favadi joined #salt
12:49 jeddi joehh2: that very much looks like it ... and I see you're on the task, and in AU also.  nicely done. :)
12:49 I3olle Hey there. Is it possible to use the return output from a .py file in a .sls formula? I would like to have the output of the following simple script and then use a for loop to generate some files on the master
12:49 I3olle import salt.config
12:49 I3olle import salt.runner
12:49 I3olle opts = salt.config.master_config('/etc/salt/master')
12:49 I3olle runner = salt.runner.RunnerClient(opts)
12:49 I3olle grains = runner.cmd('cache.grains', [])
12:49 I3olle pillar = runner.cmd('cache.pillar', [])
12:50 jeddi joehh2: it's not just that the stanzas are yellow, and saying 'result: none', but that they appear at all with test=True ... they clog up the output.  usually i run test=True before any production run, just to make sure a small sls / pillar change won't bork things savagely ... and expect only a few paragraphs of output, of course.   anyway - will wait for an updated package - thank you!
12:52 jakubek joined #salt
12:52 jakubek any idea how to pass http_proxy to all state.sls?
12:52 tomspur joined #salt
12:54 jakubek i want to run pkg.install in all states over http_proxy because my server don't have access to internet and i need to tunnel proxy
13:01 felskrone joined #salt
13:06 bhosmer joined #salt
13:11 intellix joined #salt
13:12 I3olle joined #salt
13:17 felskrone1 joined #salt
13:25 jerematic joined #salt
13:33 ]V[_ joined #salt
13:34 favadi left #salt
13:38 aqua^mac joined #salt
13:43 adilroot joined #salt
13:44 rypeck joined #salt
13:49 jeddi all 'handling users and groups via pillar' tutorials seem to assume you want the same set of users on every host.
13:49 cotton joined #salt
13:50 xsteadfastx joined #salt
13:53 babilen jeddi: Take a look at https://github.com/saltstack-formulas/users-formula
13:55 egil Running a local master and some minions i Azure: I loose contact with minions all the time
13:55 egil if I run test.ping maybe 10 times, they respond again
13:55 dude051 joined #salt
13:55 egil has anyone seen this behaviour?
13:58 elfixit joined #salt
13:59 nitti joined #salt
14:00 jeremyr joined #salt
14:00 Ymage joined #salt
14:01 jerematic joined #salt
14:04 racooper joined #salt
14:04 pestouille joined #salt
14:04 pestouille Hi
14:05 GabLeRoux joined #salt
14:05 ptinkler joined #salt
14:06 pestouille I don’t know if it’s possible with Salt to do this : state A depend on B and C, I would the state C to reload the service in B but only if B is installed...
14:06 ptinkler can salt states/directory paths not have dots in them?
14:07 ptinkler i.e. is /home/vagrant/.virtualenvs/myapp:  an illegal name for a state?
14:08 pestouille you can have ID with that name (I think)
14:08 pestouille but state file must always have .sls extension
14:08 Ahlee It's a valid name.  What error are you getting?
14:08 ze- well, the ':' is probably not part of the name :)
14:09 ze- pestouille: I don't get what you are trying to do exactly (from your example with A B C)
14:09 ptinkler "Rendering SLS python.python-pip failed, render error: expected '<document start>', but found '<block mapping start>'" is the error
14:09 ptinkler and points to the start of my next state in the file, which I don't think there's anything wrong with
14:09 pestouille ze-: I have A (php5) with B (apache) and C (newrelic) dependencies
14:09 pestouille If I install newrelic (or change its configuration file) I need to reload Apache
14:10 Ahlee ptinkler: can you gist or similar the state? It's likely a syntax issue
14:10 pestouille but… If I install php-fpm instead of php for php, I must tell newrelic to reload php-fpm instead of Apache
14:10 ptinkler sure I'll write one, one mo
14:10 mdasilva joined #salt
14:10 ze- A B and C are pkg.intall ?
14:10 pestouille yep
14:11 pestouille I have included service.running also in B
14:11 ze- that's a B', other state :)
14:11 pestouille It worked well with watch_in in newrelic state but… it doesn’t know if it must reload apache2 or php-fpm
14:11 ze- so, have B' watch C. service.running: reload: True watch: C
14:11 pestouille exactly !
14:12 ptinkler Ahlee: http://pastebin.com/10DXKCwL
14:12 ze- how do you install php-fpm or apache2 ?
14:12 pestouille so I might have to declaration for newrelic
14:12 pestouille ?
14:12 Ahlee ptinkler: testing
14:12 ze- pestouille: I guess you have a state installing php-fpm or mod-php, depending on something ?
14:12 pestouille nop
14:13 pestouille i have in my top.sls => - php5-fpm (or mod-php)
14:13 ze- well, your "something" is in top.sls
14:13 ze- you can never have php5-fpm and mod-php loaded at once
14:14 ze- get a "php-service" state (same name) in both files, reloading either apache or php-fpm.
14:14 pestouille ze-: that s it
14:14 ze- and your newrelic can watch_in that state name.
14:14 _prime_ joined #salt
14:14 pestouille ze-: let me show you a pastebin of all of this :)
14:17 pestouille http://pastebin.com/MH4xDhv2
14:18 pestouille at the end, when I change newrelic.ini I should reload either apache2 or php5-fpm depending on what have been previously installed...
14:18 Ahlee ptinkler: so what is python-pkgs?
14:18 JDiPierro joined #salt
14:19 jeddi babilen: woo - thanks for that link - it's certainly got much more complex than the last time i trawled through the saltstack-formulas repo.   mind, at 230 lines, that init.sls is pretty frightening :)
14:19 ptinkler Ahlee: just installing python libs, so it installs python-pip, python-dev, python-virtualenv and build-essential
14:20 ptinkler the whole thing works when I remove the dots from the python-pip sls, which is why I assumed it was that
14:20 Ahlee strange
14:20 Ahlee what version?
14:20 pestouille ptinkler: If you have a doubt on /home/vagrant/.virtualenvs/myapp
14:20 pestouille you can change this to virtualenv-myapp:
14:20 pestouille - name: /home/vagrant/.virtualenvs/myapp
14:20 subsignal joined #salt
14:21 babilen jeddi: We have been using in production for quite some time and it is working great though
14:21 pestouille but it should work the same in both situations
14:21 Ahlee also the witespace before /home/vagrant
14:21 ptinkler pestouille: I'll try that now
14:21 Ahlee you're offset one space on the first stanza, not offset on second
14:21 Ahlee it's invalid yaml
14:22 xmj oh noes :(
14:22 ze- pestouille: you have include troubles.
14:22 jeddi babilen: nice.  will have a deeper ponder on the morrow - it's a bit much for my after-midnight elderly brain to take in now.  i do recall previously trying to work out how to get ssh keys handled sanely between hosts, given they're a function of both where you are, and where you're configuring (if that makes sense).
14:22 pestouille ze-: probably :)
14:22 ze- php5/web includes newrelic/php which includes php5/fpm
14:22 cpowell joined #salt
14:22 ze- so, you install both, php5-fpm and mod php
14:22 ptinkler Ahlee: ah, stupid me :(
14:23 pestouille ze-: indeed… the include was a requirement for the « watch_in »
14:23 otter768 joined #salt
14:23 Ahlee ptinkler: don't worry about it, a linter is sorely needed
14:24 ptinkler yeah it's working now :) thanks
14:24 ze- pestouille: i don't see any easy way where you can state.sls newrelic.php directly.
14:25 pestouille I was thinking of some kind of global var using pillar. But It sounds ugly
14:25 ze- pestouille: but get 2 different files, that both provide php (apache & mod php - or - apache & php-fpm), and both should provide a php: service, but restarting the adequate service
14:25 pestouille it should be a common case that salt might address, but I can’t think of an elegant way to do it
14:25 xmj left #salt
14:26 pestouille the problem is not with php & apache
14:26 pestouille it s the third dependencies on newrelic
14:26 vbabiy joined #salt
14:26 pestouille I already have to different php sls files
14:26 pestouille one for php-fpm and another for apache
14:27 ze- pestouille: get the same state name with different action.
14:27 pestouille I thought ID state must be unique across all states
14:28 ze- pestouille: yeah, but it's not a problem if you can only include one of those.
14:28 ze- globaly unique, and fails if you try to install both (mod-php & fpm)
14:28 redzaku joined #salt
14:33 karimb joined #salt
14:33 mpanetta joined #salt
14:35 ze- pestouille: http://pastebin.com/UNbP5P2H
14:35 ze- and in top, you can either include php.mod or php.fpm to install the version you want.
14:37 MTecknology viq: I don't imagine you ever found time to poke at gitlab-ci perhaps?
14:38 gngsk joined #salt
14:39 keeth joined #salt
14:43 _prime_ joined #salt
14:44 pestouille joined #salt
14:45 pestouille ze-: thank you very much
14:45 pestouille ze-: I’m going to test that
14:46 glyf joined #salt
14:48 faliarin joined #salt
14:52 andrew_v joined #salt
14:52 trikke joined #salt
14:52 viq MTecknology: no, I didn't yet, but there's a much better chance now than last year
14:53 MTecknology :D
14:54 MTecknology I haven't touched it either. :(
14:54 FRANK_T joined #salt
14:54 pestouille ze-: it works… you rock :)
14:54 giantlock joined #salt
14:55 FRANK_T Do you guys know how salt manage yum package. for example a create a yum package that install a lot of thinking and the manual command is yum groupinstall @package name
14:55 MTecknology viq: will I see you at saltconf?
14:55 FRANK_T Do you know hot salt handle those packages?
14:55 viq MTecknology: wrong side of the pond
14:55 FRANK_T how salt handle those packages*
14:55 MTecknology bummer
14:56 Ouzo_12 joined #salt
14:58 Brick joined #salt
14:59 hax404 joined #salt
15:00 kaptk2 joined #salt
15:02 housl joined #salt
15:02 krelo_ joined #salt
15:08 twellspring joined #salt
15:11 dude^2 joined #salt
15:12 dude051 joined #salt
15:13 favadi joined #salt
15:14 Brew joined #salt
15:17 mdasilva joined #salt
15:20 bostonq joined #salt
15:22 Saltn00b joined #salt
15:24 iggy FRANK_T: for the most part, salt calls the yum command line util and passes through a lot of what you tell it
15:27 aqua^mac joined #salt
15:28 FRANK_T iggy how do I add that to my .sls list
15:29 pdayton joined #salt
15:30 FRANK_T I have a lot of packages like 50 and 10 group I am able to install the packages
15:30 tomh- joined #salt
15:30 FRANK_T but my problem is the groups that I have
15:30 FRANK_T I do not know if they did implement this https://github.com/saltstack/salt/issues/5504
15:31 aquinas joined #salt
15:31 iggy it's in the docs
15:32 FRANK_T Give me the link if you can please. I cant find that.
15:34 monkey66 joined #salt
15:35 smcquay joined #salt
15:36 iggy you aren't trying very hard
15:36 iggy http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.yumpkg.html#salt.modules.yumpkg.group_install
15:37 tkharju joined #salt
15:38 FRANK_T iggy I thought that this is for a group of nodes
15:39 FRANK_T I read this document
15:39 iggy presumably, everything in salt is for a group of nodes
15:39 iggy or at least has the possibility of being such
15:39 ccarney_ROCC joined #salt
15:40 FRANK_T I do not know if I am not asking the correct question.
15:40 ccarney_ROCC left #salt
15:40 twellspring looking for the best way to install minion on existing servers.   salt-cloud + saltify, bootstrap-salt.sh, something else?
15:41 FRANK_T iggy https://www.refheap.com/fee14917c2c5664562c129b45
15:41 lytchi joined #salt
15:42 lytchi win 25
15:42 lytchi oups sorry :>
15:42 FRANK_T If I install the one with the add manually I just have to do  yum groupinstall "package name"
15:43 FRANK_T I do not have a problem with the one without @
15:43 iggy there's no such option as names to pkg.installed
15:43 iggy oh, so  there is
15:43 iggy but what you really want is pkgs
15:47 FRANK_T and iggy this works for me
15:48 FRANK_T https://www.refheap.com/a8dd2b9afd5e0a5c4e83dfe2b
15:48 chrism_ joined #salt
15:49 chrism_ left #salt
15:50 chris_m_ joined #salt
15:50 jtang joined #salt
15:52 chris_m_ morning
15:52 chris_m_ ran this command: ./salt.sh '*' test.ping -b 50   reports back on all 102 minions, and then just hangs. any ideas?
15:55 glyf joined #salt
15:56 paulm- joined #salt
15:56 tkharju joined #salt
15:56 FRANK_T chris_m_ try salt '*' test.ping
15:57 FRANK_T hy -b 50
15:57 FRANK_T why -b 50?
15:58 chris_m_ only want to report back so many at a time.     like the scaling documentation suggest.  could be an overload to have them all report bck at once.
15:58 iggy chris_m_: salt-run manage.down
15:59 FRANK_T got it.
15:59 Brew1 joined #salt
16:00 ajw0100 joined #salt
16:00 SheetiS joined #salt
16:01 manfred just booked my travel to saltconf
16:03 kermit joined #salt
16:04 trikke joined #salt
16:04 glyf joined #salt
16:04 thedodd joined #salt
16:05 pestouille joined #salt
16:06 stqism_ joined #salt
16:07 chris_m_ thank-you iggy. will research that one
16:07 SheetiS1 joined #salt
16:10 FRANK_T iggy You were right I was able to install the group using pkg.group_install
16:14 pestouille_ joined #salt
16:14 FRANK_T I was able to do it like this salt '*' pkg.group_install SC_Lustre
16:14 jimklo joined #salt
16:15 brian__ joined #salt
16:15 bluenemo joined #salt
16:15 bluenemo joined #salt
16:16 rojem joined #salt
16:16 dmorrow joined #salt
16:17 iggy see you there manfred
16:17 jer_ joined #salt
16:17 jimklo joined #salt
16:18 manfred I am pumped.
16:18 dmorrow hello everyone, I was wondering does anoyone know is there anywhere official that historical RPM's are stored?. I found some on http://rpmfind.net but was wondering is there a saltstack official one?
16:22 iggy saltstack has never published rpms
16:22 iggy but the spec file is in the source. They are easy to make
16:22 clintberry joined #salt
16:24 otter768 joined #salt
16:24 chris_m_ left #salt
16:25 jimklo joined #salt
16:26 quickdry21 joined #salt
16:29 forrest joined #salt
16:30 conan_the_destro joined #salt
16:30 glyf joined #salt
16:33 lb1a joined #salt
16:33 josephleon joined #salt
16:38 Furao joined #salt
16:38 Ozack-work joined #salt
16:39 dmorrow ok cool thanks
16:41 mdasilva joined #salt
16:42 JDiPierro joined #salt
16:42 hasues joined #salt
16:42 hasues left #salt
16:44 tkharju joined #salt
16:45 anotherZero joined #salt
16:45 jalbretsen joined #salt
16:46 felskrone joined #salt
16:46 tomspur joined #salt
16:48 tligda joined #salt
16:48 _JZ_ joined #salt
16:51 Andre-B joined #salt
16:54 paulm- joined #salt
16:56 pestouille left #salt
17:01 StDiluted joined #salt
17:02 hebz0rl joined #salt
17:04 desposo joined #salt
17:04 hobakill anyone doing any sssd/winbind/whatever ad joins via salt?
17:04 keeth joined #salt
17:05 tligda1 joined #salt
17:06 lothiraldan joined #salt
17:06 AlexStraunoff joined #salt
17:06 timoguin joined #salt
17:06 tligda1 joined #salt
17:07 cpowell joined #salt
17:09 theologian joined #salt
17:10 spookah joined #salt
17:12 lothiraldan_ joined #salt
17:15 TyrfingMjolnir joined #salt
17:15 aqua^mac joined #salt
17:20 bluenemo joined #salt
17:20 bluenemo joined #salt
17:21 wendall911 joined #salt
17:22 KyleG joined #salt
17:22 KyleG joined #salt
17:26 lothiraldan joined #salt
17:27 pdayton joined #salt
17:27 Deevolution joined #salt
17:28 agend joined #salt
17:28 jla joined #salt
17:29 hobakill any chance we can get 2014.7.1 out of EPEL testing and into stable? i realize no one here has direct access but i think upvotes help.
17:30 Guest89335 joined #salt
17:30 forrest hobakill: do you have the link?
17:30 forrest hobakill: I can log in and upvote
17:32 hobakill forrest, i don't. i always lose it. let me find and post it. in the meantime i just upgraded from epel testing. :)
17:32 forrest heh
17:32 TyrfingMjolnir joined #salt
17:32 iggy push saltstack to move the packaging in house sooner than later
17:32 hobakill iggy, amen
17:32 vectra joined #salt
17:33 aparsons joined #salt
17:33 iggy (and preferably with separate repos for the major versions... which I think they do for ubuntu or debian already)
17:33 Ryan_Lane joined #salt
17:34 mdasilva joined #salt
17:34 hobakill forrest, https://apps.fedoraproject.org/packages/salt-master
17:35 kitp joined #salt
17:35 hobakill forrest, https://admin.fedoraproject.org/updates/salt-2014.7.1-1.el6
17:35 hobakill not sure which one you want
17:36 forrest I'm looking at all of them, ugh I always forget how to upvote these...
17:36 paulm-- joined #salt
17:36 rodder joined #salt
17:37 forrest 2015-01-28 05:53:59
17:37 forrest This update has reached 7 days in testing and can be pushed to stable now if the maintainer wishes
17:37 forrest hobakill: ^
17:37 timoguin joined #salt
17:37 hobakill i thought it was 14 for some reason
17:37 rodder Hello everyone. Does anyone know of way to use an entire subnet in the /etc/salt/autosign.com
17:37 rodder autosign.conf
17:38 forrest hobakill: Nah, that was just the non cent versions anyways
17:38 rodder id like to auto accept any keys coming from a kickstart network.
17:38 iggy rodder: I'm pretty sure autosign accepts globs
17:38 iggy but I've only ever seen it used with names, so...
17:38 forrest hobakill: should be there the day after tomorrow, it has zero upvotes, so mine won't help much :P
17:38 rodder ive tried the usualy CIDR notation
17:38 forrest just going to let it hang out.
17:38 rodder no love
17:39 hobakill forrest, sounds good.
17:39 monkey661 joined #salt
17:40 hobakill forrest, one of these days i'm going to learn how to package an RPM from the dev branch ... haven't had much luck so far
17:41 forrest hobakill: did you grab the spec and try rpmbuild and it didn't work?
17:41 hobakill forrest, yeah it kept looking for old-ass versions of salt that, frankly, i wasn't sure if i needed to tar up and put in the directory
17:42 forrest hobakill: ahh yeah I'm not sure either, if terminalmage shows up today, you could ask him if you can't figure it out.
17:42 hobakill forrest, sounds good. i should also spend more than 45 minutes trying to get it to work!
17:43 forrest The joys of RPMs, where 45 minutes with a pre-made (working) file doesn't result in a working install
17:43 hobakill no joke.
17:43 hobakill another quick question. can we upgrade windows salt minions via salt yet? i've been manually doing them all and it's a real pain
17:44 neogenix joined #salt
17:45 josephleon joined #salt
17:45 shaggy_surfer joined #salt
17:51 aw110f joined #salt
17:54 neogenix_ joined #salt
17:57 Ouzo_12 joined #salt
18:01 forrest hobakill: I honestly don't know, I rarely touch the windows stuff.
18:01 saltine joined #salt
18:01 desposo joined #salt
18:03 Saltn00b Good morning everyone. Can anyone help me with a salt-ssh problem? If I install from bootstrap, then when I try to run salt-ssh, it fails on the minions with "NameError: global name 'msgpack' is not defined".
18:03 Saltn00b The versions report returns Salt: 2015.2.0-307-g1fc9a52
18:03 Saltn00b If I install from pip, it works fine.
18:04 Guest78488 Saltn00b: sounds like the git install doesn't pull in all the dependencies for you
18:04 hobakill forrest, you are lucky
18:04 RedundancyD joined #salt
18:04 Saltn00b In that case, versions report returns Salt: 2014.7.1
18:04 Saltn00b Guest78488: That's my thought, but the minions shouldn't need msgpack (or anything) for salt-ssh.
18:05 Saltn00b There's a similar issue open about this (https://github.com/saltstack/salt/issues/7913), but it was fixed in 2014.7.1.
18:05 Saltn00b Could the later code have introduced a regression?
18:06 holler joined #salt
18:07 pdayton joined #salt
18:08 holler hello, I just started getting an error when doing vagrant up (w/salt provision) to rackspace.. ive tried reverting to past commits and same thing.. nothing seems to have changed that I know of so not sure what's causing the error? UnicodeEncodeError: 'ascii' codec can't encode character
18:08 gmoro joined #salt
18:08 timbyr_ joined #salt
18:09 bhosmer_ joined #salt
18:09 dgiagio joined #salt
18:10 iggy holler: A. salt is python2 only B. python2 has shit unicode/charset handling
18:11 holler iggy: would it be anything relating to my actual application code? or something in the provisioner? I havent changed any of the salt files
18:11 Guest78488 Saltn00b: good question!
18:11 Guest78488 Saltn00b: ping one of the salt devs/ops
18:11 Ryan_Lane iggy: the next version of salt supports python3 ;)
18:12 iggy "supports"
18:12 glyf joined #salt
18:12 iggy I know they've done a lot of work, but it's a big project to claim "full support"
18:13 iggy holler: if python2 is the default on the systems you are using, then my guess would be that you had some unicode characters sneak in somewhere (often from c&p'ing from a webpage and your ' turns into the unicode version that look like `)
18:16 monkey66 joined #salt
18:17 holler iggy: crap... this seems like it could be a needle in a haystack@
18:17 holler !
18:17 holler the weird thing is that my local vagrant/salt works
18:18 holler the main difference between the local version and rackspace version is rackspace pulls the actual code from github and sets it up wheras the local just symlinks to the host machine
18:19 holler how could I find unicode character in a file tree?
18:19 Mso150 joined #salt
18:20 cpowell joined #salt
18:23 Auroch joined #salt
18:23 iggy holler: throw all your yaml in http://yaml-online-parser.appspot.com/ (with jinja variables filled in appropriately)
18:24 shaggy_surfer joined #salt
18:25 otter768 joined #salt
18:30 mdasilva joined #salt
18:31 rap424 joined #salt
18:33 schlueter joined #salt
18:37 salten joined #salt
18:38 salten hello, when I try to use dockerio states/modules on a set of  hosts, I get 'docker.foo' is unavaible. I have docker installed, docker-py installed (with pip, even install --upgrade), and the salt-minions refreshed, what am I missing?
18:40 forrest salten: which version of salt are you using?
18:42 salten forrest: salt-minion 2014.7.0 (Helium)
18:43 salten forrest: ok, so I'm probably referencing states/modules from develop
18:43 salten or more recent version
18:43 forrest salten: That's what I was wondering as well
18:43 forrest but I see this salten: https://github.com/saltstack/salt/blob/v2014.7.0/salt/states/dockerio.py
18:44 forrest so maybe you are using something that doesn't exist in that release?
18:44 salten I will find out
18:44 salten what is the latest stable?
18:46 timoguin joined #salt
18:48 hax404 joined #salt
18:49 timoguin joined #salt
18:49 hax404 joined #salt
18:50 conan_the_destro joined #salt
18:50 adnauseaum left #salt
18:50 toanju joined #salt
18:54 Guest99681 joined #salt
18:55 GabLeRoux Hey there, How would I install php module php_curl (that's for legacy code support). php-mysql worked (pkg: - installed), but I can't seem to find how I could install php_curl with salt
18:57 GabLeRoux Oh well, that's not a yum package, that's why
18:58 iggy salten: try with docker-py-0.5 (I think latest is more recent and people have had issues with it)
18:58 BigBear joined #salt
19:02 Grokzen Is raising a exception from a module method the only way to say that something within the method failed? Can i somehow set some retcode or say success=False?
19:02 iggy I think most modules just return false
19:02 iggy exception's are frowned upon
19:03 jimklo joined #salt
19:03 jimklo joined #salt
19:04 josephleon joined #salt
19:04 ajw0100 joined #salt
19:04 aqua^mac joined #salt
19:06 Grokzen iggy, mkay
19:06 salten iggy: thanks for the pointer
19:10 Mso150 joined #salt
19:11 Grokzen iggy, Just returning False do not change the success or retcode in the event that is transmitted on the eventbus, it just sets the 'return' field to False
19:11 iggy well, that's not what you said to begin with...
19:12 iggy I'm 98% sure there isn't a way
19:12 Grokzen it was not O.o
19:13 iggy reactors/schedulers, etc are notoriously bad about error handling
19:15 hax404 joined #salt
19:17 druonysus joined #salt
19:24 josephleon joined #salt
19:24 davet joined #salt
19:25 jla my salt-master for just a handfull, several, minions is using 1.3g virt and 1.2g res and every minute or so pegs at 99% cpu for several seconds. Does that sound normal?
19:26 spootly joined #salt
19:27 kitp joined #salt
19:28 iggy cpu peggin maybe (depends on cpu, etc.)
19:28 iggy memory usage, no
19:30 iggy my master with 40+ minions is using less than 512M of mem
19:30 iggy res
19:30 iggy virt is much higher, but I don't ever pay attention to virt
19:31 jla I'll have to re-check what I'm doing. The salt/master log is empty, nothing in messages, daemon, or dmesg related to salt or looking out of the ordinary. I didn't even think I was doing that much with salt yet.
19:31 jla thanks
19:32 iggy jla: oh, you know one time I had a problem with something like that... We have a git tree that holds our "dist" files (packages, code archives, etc.), and when I used to have it as a gitfs backend mem usage was through the roof
19:32 ALLmightySPIFF joined #salt
19:35 davet joined #salt
19:36 druonysuse joined #salt
19:39 glyf joined #salt
19:40 berserk joined #salt
19:42 jla I had to go check the configs because I had started into salt a couple months ago then got busy and am trying to dive in again. My master doesn't have any git stuff defined. The fileserver_backend is all commented out so defaults
19:42 iggy well, there goes that idea :/
19:44 jla I remember thinking about trying multi-master but according to my minion configs I'm not doing that either. What's the best way to see what the master is doing?
19:44 druonysus joined #salt
19:44 druonysus joined #salt
19:45 johanek joined #salt
19:45 kitp joined #salt
19:45 jeremyr joined #salt
19:48 Guest99681 joined #salt
19:48 shaggy_surfer joined #salt
19:48 bhosmer__ joined #salt
19:49 supersheep joined #salt
19:49 arif-ali_ joined #salt
19:49 lionel_ joined #salt
19:49 mik3 joined #salt
19:50 iggy__ joined #salt
19:51 toanju joined #salt
19:51 philipsd6_ joined #salt
19:52 GnuLxUsr_ joined #salt
19:52 scoates_ joined #salt
19:52 malinoff joined #salt
19:53 rbjorkli1 joined #salt
19:55 mtanski_ joined #salt
19:56 smkelly_ joined #salt
19:56 MK_FG joined #salt
19:56 monkey661 joined #salt
19:56 MK_FG joined #salt
19:57 Tahm joined #salt
19:57 iamtew joined #salt
19:58 rhand joined #salt
19:58 catpig joined #salt
19:58 al joined #salt
19:58 whytewolf joined #salt
19:58 lamasnik joined #salt
19:58 xist joined #salt
19:59 Alan_S joined #salt
19:59 Mso150_t joined #salt
20:00 thedodd joined #salt
20:00 shaggy_surfer joined #salt
20:00 malinoff_ joined #salt
20:02 Grokzen iggy__, Aparently you can set the retcode with __context__["retcode"] = 1337 inside the module function
20:02 Grokzen that should be enough to handle errors
20:10 giantlock joined #salt
20:14 BigBear joined #salt
20:18 aparsons joined #salt
20:18 chris_m_ joined #salt
20:19 quickdry21 joined #salt
20:19 chris_m_ afternoon all (those in EST)
20:19 johtso joined #salt
20:20 chris_m_ qq instead of using the --config-dir everytime I execute a salt command (since i am running as a non-root user).  Is there any master profile where I can auto-define this config path
20:20 urtokk joined #salt
20:20 toastedpenguin joined #salt
20:20 felskrone joined #salt
20:21 ThomasJ chris_m_: Not sure, but could you not solve it with an alias?
20:22 gattie joined #salt
20:23 toastedpenguin joined #salt
20:23 chris_m_ yes, that would work too.    but, is there any environment properties file (like .bashrc or .profile) that salt reads before executing any commands - would be better :)
20:24 ThomasJ Absolutely, can't help you there though :\
20:24 chris_m_ thx Thomas for your help
20:26 otter768 joined #salt
20:26 supersheep_ joined #salt
20:27 serenecloud left #salt
20:30 Morbus joined #salt
20:31 andrew_v joined #salt
20:32 redzaku joined #salt
20:33 chris_m_ Thomas - there is two variables that store the environment properties location: SALT_MASTER_CONFIG & SALT_MINION_CONFIG
20:34 mr_chris joined #salt
20:34 chris_m_ is it as simple as setting these in my user profile to what I need?   and then, I don't have to worry about adding the --config-dir everytime I run salt?
20:34 jimklo joined #salt
20:38 ThomasJ easiest way to test: export SALT_MASTER_CONFIG=''   and then just try :)
20:38 Mso150 joined #salt
20:38 ThomasJ With the desired path of course
20:39 aparsons joined #salt
20:40 zadock joined #salt
20:41 big_area joined #salt
20:42 ALLmightySPIFF joined #salt
20:42 durana joined #salt
20:44 Mso150_f joined #salt
20:45 cberndt joined #salt
20:46 josephleon joined #salt
20:48 bluenemo joined #salt
20:48 bluenemo joined #salt
20:52 jimklo joined #salt
20:53 aqua^mac joined #salt
20:53 ALLmightySPIFF joined #salt
20:54 Mso150 joined #salt
21:01 Ozack-work joined #salt
21:03 kermit joined #salt
21:03 kermit joined #salt
21:04 conan_the_destro joined #salt
21:05 laax joined #salt
21:06 murrdoc joined #salt
21:07 snuffychi joined #salt
21:08 andrej Is there a way to only capture data pertaining to a certain minion in the masters log?
21:10 iggy minions don't log to the masters log
21:11 iggy you can use a returner to get the ouput somewhere else useful
21:12 twellspring joined #salt
21:13 hal58th1 joined #salt
21:19 jtang joined #salt
21:20 badon joined #salt
21:21 CeBe joined #salt
21:23 CeBe joined #salt
21:25 snuffychi joined #salt
21:26 druonysus joined #salt
21:26 druonysus joined #salt
21:32 josephleon joined #salt
21:35 _2_Maggie joined #salt
21:35 _2_Maggie hey guys :D
21:40 mdasilva joined #salt
21:42 josephleon joined #salt
21:43 * iggy ponders when 2015.2 will actually be released
21:44 jtang i dont suppose anyone here uses salt-virt as a cloud controller?
21:44 jtang i've seem to have run into a bunch of regressions that I'm not sure if others have come across
21:44 iggy lots of people do
21:44 Ryan_Lane salt-virt?
21:45 Ryan_Lane lots of people use salt-cloud, I haven't seen many who use salt-virt
21:45 jtang specifically runners.virt
21:45 Brick joined #salt
21:45 iggy oh, that
21:45 iggy yeah, you're on your own
21:45 jtang the virt.modules are a bit broken as well in 2014.7.1
21:45 jtang at least for me on ubuntu they are a bit borked
21:46 jtang iggy, its convenient for spinning up vm's on my laptop for testing
21:47 Ryan_Lane I'm sure there's others using it
21:47 Ryan_Lane I've seen people talk about it once or twice in here
21:47 jtang its just a little frustrating with these regressions
21:48 Ryan_Lane jtang: you should open an issue on github
21:48 Ryan_Lane in the saltstack/salt repo
21:48 mosen joined #salt
21:48 Ryan_Lane I know the saltstackinc folks use it
21:49 jtang Ryan_Lane, i've already logged it
21:49 Ryan_Lane ah. gotcha
21:49 jtang I'm trying to track down where the problem is
21:53 bfoxwell joined #salt
21:54 rbjorkli1 Hi, does anyone know the .sls syntax for the docker.running argument links?
21:55 rbjorkli1 I've been trying:
21:55 rbjorkli1 - links:
21:56 rbjorkli1 /linked_container:/container_with_link/alias
22:00 josephleon joined #salt
22:00 iggy rbjorkli1: space is important in yaml (I don't kno wanything about docker.running, but that doesn't look right in a strictly yaml sense)
22:02 jimklo_ joined #salt
22:03 andrew_v joined #salt
22:05 yomilk joined #salt
22:06 rbjorkli1 iggy: You are correct, with 4 more spaces it's now almost doing what I wanted
22:07 shaggy_surfer joined #salt
22:08 iggy rbjorkli1: it looks like whatever you specify there get's passed all the way down to docker.utils.create_host_config which wants a dict or list it looks like
22:09 smcquay joined #salt
22:11 jimklo joined #salt
22:14 clintberry joined #salt
22:17 hellerbarde joined #salt
22:26 glyf joined #salt
22:26 otter768 joined #salt
22:34 jimklo joined #salt
22:37 druonysus joined #salt
22:42 aqua^mac joined #salt
22:44 dude051 joined #salt
22:48 lothiraldan joined #salt
22:48 dude051 joined #salt
22:54 josephleon joined #salt
22:56 smcquay joined #salt
22:59 conan_the_destro joined #salt
23:00 rogst joined #salt
23:07 GabLeRou_ joined #salt
23:10 glyf joined #salt
23:21 the_lalelu the gpg renderer is nice. thx again salt upstreams. ;)
23:27 crane joined #salt
23:32 Edgan I am trying to get a redis.conf's contents to dependent on a pillar existing or not. The pillar depends on the value of a grain. It all looks right, except it doesn't behave as expected. http://pastebin.com/FWJHUdaG
23:33 kermit joined #salt
23:33 kermit joined #salt
23:34 iggy you have include redis.config, but the only file you have there is /srv/pillar/config.sls
23:35 Edgan iggy: that is a typo on my part, it is correct in reality
23:35 iggy and you have top targing redis and no redis file
23:35 Edgan http://pastebin.com/rmaMxTf7
23:36 Edgan If it didn't find slave, then it wouldn't write anything in redis.conf
23:36 josephleon joined #salt
23:36 Edgan I expect to write slaveof on 02 and 03, but it does it on 01 too
23:37 stevednd grains['id'] is a system grain, and I'm pretty sure you can't overwrite it
23:37 iggy don't use defined
23:37 stevednd nor should you try to
23:37 stevednd that's the actual minion id grain
23:37 Edgan stevednd: I am not trying to overwrite it. I am just trying to if != it
23:37 stevednd call your instance_id or something
23:38 iggy foo=False <-- that's still defined
23:38 Edgan iggy: yes, but on top it is slave: True on 01
23:38 stevednd Edgan: yes, but given what you posted, if 'id' cannot be overriden then grains['id'] will not be whatever you have '01'
23:38 Edgan iggy: I expect it not to be defined on 01
23:39 iggy if '01' in grains['id']
23:39 stevednd grains['id'] will equal redis-01.foo.com or whatever your actual minion id is
23:39 iggy or not in
23:39 Edgan iggy: != is only for integers?
23:39 stevednd so either use another gain, or do as iggy suggested
23:40 stevednd if not '01' in grains['id']
23:40 iggy no, but as stevednd said, your id isn't going to be 01, it's redis-01.foo.com
23:41 Edgan iggy: I don't know where you get that. I think your suggestion will work.
23:45 Edgan iggy: ok, so that worked, but http://docs.saltstack.com/en/latest/topics/tutorials/pillar.html suggests what I did should have worked,  {% if grains['os_family'] == 'RedHat' %}
23:46 tomh- joined #salt
23:46 Edgan iggy: unless is there some inconsistent difference between == and !=
23:48 stevednd Edgan: there is no inconsistent difference between the two. In python, you place `not` before the condition. `if not 'something' == yourvar`
23:49 rudi_s Hi. I'm new to salt and have a general question about packages and services. At the moment my "template" to create and start a service looks like this: https://pbot.rmdir.de/EZNKpCXk-wo3nm5KI7IkiA - if the service depends on e.g. config files, I add watch_in: service: nslcd to the config file. - Is this the recommended way to install/start/enable/restart a service or should I use a different way? Thanks.
23:50 Edgan stevednd: That is just silliness. != should equal not ==
23:50 rudi_s Also, is the require: pkg: nslcd necessary?
23:50 rudi_s (And while I'm at it, this creates quite some boilerplate, is there a recommended way to shorten this?)
23:51 anotherZero joined #salt
23:55 neilf______ joined #salt
23:56 scalability-junk joined #salt
23:57 meylor joined #salt
23:59 hal58th1 rudi_s I believe you want to break that into two separate states. Then the require on the service state would make more sesnse

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