Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-01-16

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

All times shown according to UTC.

Time Nick Message
00:02 Edgan joined #salt
00:02 Edgan Anyone using foreman-salt?
00:03 stevednd Edgan: no, but I was considering it. Please report back if you end up getting it working
00:04 Edgan stevednd: I have it kinda of working. But trying to figure out how to make it usable.
00:05 davebarnes left #salt
00:05 1JTAABLY5 .
00:06 bhosmer_ joined #salt
00:07 Ryan_Lane joined #salt
00:08 davebarnes joined #salt
00:08 franz_ joined #salt
00:09 franz_ can someone confirm that the file state doc has an invalid example
00:09 franz_ http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html
00:10 franz_ - file_mode: 744
00:10 iggy franz_: what's wrong with it?
00:10 franz_ but the file.managed definition only includes "mode" and "dir_mode"
00:10 elco_ joined #salt
00:11 portablejim joined #salt
00:11 iggy franz_: yes, file_mode only makes sense for some states (directory with recurse, recurse, etc.)
00:11 cads joined #salt
00:11 davebarnes left #salt
00:12 iggy it doesn't make sense for file.managed because that is implied by mode
00:12 miqui_ joined #salt
00:12 franz_ ok, thanks iggy
00:13 phpdave11 joined #salt
00:13 paha joined #salt
00:14 iggy I'd personally think they should just add them in for consistency, but I get it
00:18 jeremati_ joined #salt
00:20 FarrisG joined #salt
00:21 atoponce i think the 'extend' state is actually the right solution
00:21 ryuhei joined #salt
00:21 bhosmer__ joined #salt
00:21 atoponce for my case
00:21 atoponce and indeed, it works as it should
00:21 ryuhei hello.. anyone using 2014.7 with digitalocean?
00:24 abrahamrhoffman hi i did a few months ago
00:24 abrahamrhoffman salt cloud for provisioning
00:24 ryuhei private_networking: True is not working anymore with apiv1 for some reason :/
00:24 abrahamrhoffman hm
00:24 abrahamrhoffman havent tried
00:25 abrahamrhoffman probably raise a ticket on the DO side
00:25 abrahamrhoffman ensure that providers linking is proper in new API
00:25 ryuhei right.. will do that as well
00:25 ryuhei thanks for the advise!
00:26 abrahamrhoffman same applies for private networking i imagine
00:26 abrahamrhoffman sorry not much help :P
00:26 ryuhei heh
00:27 phpdave11 joined #salt
00:28 phpdave11 joined #salt
00:28 ryuhei the problem is that it works on 2014.1.11 but not on 2014.7
00:31 Singularo joined #salt
00:36 druonysus joined #salt
00:36 druonysus joined #salt
00:37 otter768 joined #salt
00:37 ccarney_ROCC joined #salt
00:41 g3cko joined #salt
00:44 bhosmer_ joined #salt
00:50 brianfeister joined #salt
00:51 dave_den joined #salt
00:52 theologian joined #salt
00:53 nitti joined #salt
00:54 nitti joined #salt
00:55 Bryanstein joined #salt
00:57 favadi joined #salt
01:04 ryuhei found out why..
01:05 Edgan yumpkg is crazy slow. Anyone have any tips?
01:08 robawt joined #salt
01:11 zzzirk joined #salt
01:13 MugginsM joined #salt
01:18 oeuftete joined #salt
01:18 nickdew Edgan: yumpkg is just runs yum from the CLI, so it is as fast as yum.
01:20 Edgan nickdew: It should not be running repoquery per package for already installed packages. This is the same crazyiness that ansible does. All it needs to do is a rpm -q stuff to know a package is installed.
01:20 nickdew Edgan: So, local mirrors help.
01:21 Edgan nickdew: repoquery, with caching of metadata, is like 1.3s per package listed. That adds up way too fast.
01:22 nickdew Edgan: Agreed. I'm actually thinking about taking package management out of Salt and moving it to Pulp.
01:22 Edgan nickdew: ?
01:23 nickdew Edgan: http://www.pulpproject.org/
01:24 Edgan nickdew: Sounds like katello, but just sounds like repo management, not package management
01:24 nickdew Edgan: Pulp is part of Katello. I thought it could do package management?
01:25 nickdew Edgan: Haven't actually got around to deploying it. Still using Spacewalk.
01:25 Edgan nickdew: Sounds like you can push stuff, but that is more orchestration and not configuration management.
01:27 nickdew Edgan: True. I wonder if consolidating package installs into one state (with lots of Jinja) would make a difference.
01:28 Edgan nickdew: looks like I can replace repoquery with rpm, and use almost the same syntax. The only thing that repoquery adds is the repo. I am going to see if I can find a faster way to get that information, or find a way to not need it.
01:29 nickdew Edgan: That could be promising. I was thinking about doing the initial setup in Salt, but then having Pulp to manage package versions and updates.
01:29 nickdew Edgan: Like, using the content management in Pulp to control what versions prod vs dev are on.
01:30 Edgan nickdew: I prefer things to be unified in CM. I also see a nice improvement if I disabled plugins. So I need to find which plugin is slowing it down.
01:30 yomilk joined #salt
01:32 druonysuse joined #salt
01:32 raghu joined #salt
01:32 Edgan nickdew: Looks like fastest mirror is doing quite a bit of it
01:33 nickdew Edgan: Makes sense. Disable that and using local mirrors probably would help.
01:33 Edgan nickdew: trying highstate again with it disabled. I am currently setting up a local mirror, so fastest won't be needed
01:34 beneggett joined #salt
01:34 Edgan nickdew: Twice as fast without :)
01:35 nickdew Edgan: Oh nice! Very good to know. This is the first shop I don't have local mirrors.
01:35 Edgan nickdew: Sadly the package is now a dependency of yum. So I have to rewrite the config to disable it.
01:36 nickdew Edgan: Yeah, which you would do if you are creating local mirrors anyway.
01:36 Nazzy joined #salt
01:38 nickdew Edgan: Kind of odd that yum and yum-plugin-fastestmirror both depend on each other.
01:39 Edgan nickdew: didn't used to, probably did it to stop people from just uninstalling it :\
01:43 murrdoc local mirrors for the win
01:56 manytrees hal58th1: thanks for the tip, turned out i can't immediately do salt-call state.highstate after bootstrapping salt, things get confused
01:58 mrjk joined #salt
02:04 yomilk joined #salt
02:05 portablejim joined #salt
02:05 portablejim joined #salt
02:05 portablejim joined #salt
02:07 forrest joined #salt
02:10 sbx joined #salt
02:10 dude051 joined #salt
02:16 druonysuse joined #salt
02:16 druonysuse joined #salt
02:19 hvn joined #salt
02:21 TyrfingMjolnir joined #salt
02:29 tracphil joined #salt
02:29 nmadhok joined #salt
02:34 budman joined #salt
02:37 otter768 joined #salt
02:38 Nazzy joined #salt
02:40 nmadhok joined #salt
02:40 nmadhok joined #salt
02:41 portablejim joined #salt
02:41 portablejim joined #salt
02:41 portablejim joined #salt
02:42 yomilk joined #salt
02:45 saltymoli left #salt
02:52 otter768 joined #salt
02:53 nmadhok joined #salt
02:59 favadi joined #salt
03:07 bhosmer joined #salt
03:07 jaylark joined #salt
03:08 chrischris joined #salt
03:08 jcockhren joined #salt
03:11 stotch joined #salt
03:13 vukcrni joined #salt
03:14 hobakill joined #salt
03:16 hobakill hey guys - trying to run my first masterless salt setup. running into an odd error on a VERY VERY basic configuration: https://www.refheap.com/96178
03:18 otter768 joined #salt
03:18 monkey66 joined #salt
03:19 hobakill what's odd is that i'm not doing anything at all with pillars so i don't know why it even brings that up
03:21 JlRd joined #salt
03:24 hobakill ok nevermind. i figured it out. i had to create /srv/pillar.... not totally sure why but after i did that i was able to run the highstate as expected.
03:28 dude051 joined #salt
03:29 MugginsM it was probably looking to see if you had any
03:32 Brew joined #salt
03:49 aqua^mac joined #salt
03:51 monkey661 joined #salt
03:52 hvn joined #salt
03:54 hvn joined #salt
03:55 hvn joined #salt
03:55 hvn joined #salt
03:57 bhosmer_ joined #salt
04:03 abrahamrhoffman joined #salt
04:03 lamasnik joined #salt
04:03 yomilk_ joined #salt
04:04 favadi joined #salt
04:05 elfixit joined #salt
04:11 eagles0513875 joined #salt
04:11 eagles0513875 joined #salt
04:26 favadi joined #salt
04:41 Furao joined #salt
05:00 bhosmer_ joined #salt
05:06 jramnani` joined #salt
05:08 Andre-B joined #salt
05:08 bhosmer_ joined #salt
05:20 neogenix joined #salt
05:23 beneggett joined #salt
05:24 saurabhs left #salt
05:33 portablejim left #salt
05:43 maxskew joined #salt
05:49 druonysus joined #salt
05:49 druonysus joined #salt
05:53 _ether_ joined #salt
05:54 TomWork joined #salt
05:54 jerematic joined #salt
06:00 Furao joined #salt
06:13 catpiggest joined #salt
06:15 ramteid joined #salt
06:16 kermit joined #salt
06:17 svx joined #salt
06:21 druonysus joined #salt
06:21 druonysus joined #salt
06:21 hvn joined #salt
06:21 hvn joined #salt
06:32 Furao joined #salt
06:35 Nazca__ joined #salt
06:50 ponpanderer joined #salt
06:50 ponpanderer hello
06:52 ponpanderer i seem to be noticing some odd new behavior rendering a script template with jinja + salt mine. it always worked previously and I just noticed since moving to 2014.7 I get an error rendering jinja for: {% for server, addrs in salt['mine.get']('I@tags:user:key:value', 'network.ip_addrs', expr_form='compound').items() %}
06:53 ponpanderer the error being that whatever the mine function is doing it's not iterable. error is: SaltRenderError: Jinja variable 'str object' has no attribute 'items'
06:54 ponpanderer what's odd is doing a compound match works fine for running scripts, only with salt-mine does it appear to be a problem.
06:54 qybl joined #salt
06:54 qybl joined #salt
06:57 ponpanderer just to add using a glob works just fine for pulling data from salt mine. it's only the compound match in a script rendered in jinja for salt mine
06:58 nebuchadnezzar joined #salt
07:03 calvinh joined #salt
07:03 felskrone joined #salt
07:03 hvn joined #salt
07:03 hvn joined #salt
07:05 jerematic joined #salt
07:06 Andre-B joined #salt
07:08 iMil joined #salt
07:14 Haskron joined #salt
07:14 Haskron .f Someone is here ?
07:16 otter768 joined #salt
07:19 Raging_fenrir joined #salt
07:25 arno joined #salt
07:27 AndreasLutro joined #salt
07:46 vukcrni joined #salt
07:48 Furao joined #salt
07:52 Auroch joined #salt
07:54 trikke joined #salt
07:56 elco__ joined #salt
07:57 hebz0rl joined #salt
07:58 JlRd joined #salt
07:58 CycloHex joined #salt
07:59 jtang joined #salt
08:07 dh joined #salt
08:13 felskrone joined #salt
08:18 Kakwa joined #salt
08:22 Vye joined #salt
08:22 jramnani` joined #salt
08:23 smkelly joined #salt
08:23 intellix joined #salt
08:26 vukcrni joined #salt
08:26 kawa2014 joined #salt
08:28 jhauser joined #salt
08:29 slafs joined #salt
08:29 slafs left #salt
08:31 flyboy joined #salt
08:32 tux_ joined #salt
08:42 KermitTheFragger joined #salt
08:43 lb1a joined #salt
08:46 babilen joined #salt
08:49 favadi joined #salt
08:49 jtang joined #salt
08:50 chiui joined #salt
08:50 CeBe joined #salt
08:51 tomh- joined #salt
08:52 ndrei joined #salt
08:54 jerematic joined #salt
09:11 eliasp joined #salt
09:13 go joined #salt
09:17 otter768 joined #salt
09:18 xsteadfastx joined #salt
09:18 Grokzen joined #salt
09:21 bhosmer_ joined #salt
09:21 Mindfart joined #salt
09:22 Kakwa joined #salt
09:23 xsteadfastx joined #salt
09:26 active8 joined #salt
09:27 xsteadfastx joined #salt
09:28 dkrae joined #salt
09:30 masterkorp joined #salt
09:31 Rawkode joined #salt
09:36 zadock joined #salt
09:36 eliasp joined #salt
09:44 jrluis joined #salt
09:48 N-Mi joined #salt
09:48 N-Mi joined #salt
09:48 paulm- joined #salt
09:51 paulm-- joined #salt
09:56 d0xb joined #salt
09:58 thayne joined #salt
10:00 CeBe joined #salt
10:01 davromaniak joined #salt
10:02 favadi joined #salt
10:04 linjan joined #salt
10:08 Raging_fenrir joined #salt
10:10 zadock joined #salt
10:11 aquinas joined #salt
10:12 Raging_fenrir joined #salt
10:12 tux_ hello, how do i enable modules like sysbench? i got "'sysbench.cpu' is not available"...
10:14 tux_ ok nvm just realized i have to install sysbench ;)
10:35 karimb joined #salt
10:36 CeBe joined #salt
10:43 jerematic joined #salt
10:46 xsteadfastx joined #salt
11:01 calvinh joined #salt
11:09 che-arne joined #salt
11:12 CeBe joined #salt
11:18 otter768 joined #salt
11:18 Furao joined #salt
11:19 giantlock joined #salt
11:23 CeBe1 joined #salt
11:24 karimb joined #salt
11:24 zadock joined #salt
11:29 linjan joined #salt
11:33 calvinh_ joined #salt
11:34 markm_ joined #salt
11:40 istram joined #salt
11:42 robinsmidsrod joined #salt
11:43 gmoo joined #salt
11:44 gmoo left #salt
11:45 gmoo joined #salt
11:45 calvinh joined #salt
11:48 bhosmer joined #salt
11:52 robinsmidsrod joined #salt
11:55 vbabiy joined #salt
11:56 ecdhe joined #salt
11:57 gmoo left #salt
11:59 gmoo joined #salt
12:04 calvinh_ joined #salt
12:04 favadi joined #salt
12:06 xsteadfastx joined #salt
12:08 kbyrne joined #salt
12:12 saltymoli joined #salt
12:15 mikkn joined #salt
12:21 giantlock joined #salt
12:22 jtang joined #salt
12:26 bhosmer joined #salt
12:32 akafred joined #salt
12:34 jtang joined #salt
12:36 booly-yam joined #salt
12:39 paulm-- How can I modify files in a user's home directory immediately after they're created with states? I can't template their home directory because it won't exist at the time the template is compiled!
12:39 paulm-- That is, {{ salt['user.info'](username).home }} won't exist yet
12:42 saltymoli paulm--, are you using salt to create the home directories? if so, you could just tack on another state. if you're using a module, you could have the module send an event back to the salt master, and setup the reactor system
12:45 paulm-- saltymoli: I am using states. I cannot just "tack on another state" because that other state doesn't know where the home directory is without querying the server (since home directories are configurable)
12:45 paulm-- That is, home directories can be configured outside the purview of salt. Home directory locations are not being managed by salt
12:46 paulm-- I cannot make assumptions about the location of a home directory therefore I must query the system to discover the location. I cannot do this using templating, though, since the user may not yet exist!
12:47 saltymoli ah, i see
12:47 saltymoli are the home directories created (you just said configured) outside of the purview of salt as well?
12:48 paulm-- Salt creates users and home directories are also created at this time but the system determines the location. Besides this, other users with home directories configured in a different manner may also exist
12:49 saltymoli you could write a custom execution module that is able to query the system and then perform a follow-on task (and you could tack the execution module to the end of the salt state)
12:52 mrjk joined #salt
12:53 calvinh joined #salt
12:55 JWK joined #salt
12:56 JWK hi
12:58 calvinh_ joined #salt
12:58 JWK I’m trying to use saltstack with gentoo and I have a lil trouble with setting use flags in sls. for instance this sls: http://pastebin.com/BxJtCcgz returns error  InvalidAtom: Invalid use dep: '"fpm"'
12:59 JWK anyone has experienced similar problems?
13:00 toddnni joined #salt
13:02 wnkz joined #salt
13:02 tkharju joined #salt
13:03 redzaku joined #salt
13:04 calvinh joined #salt
13:07 hvn joined #salt
13:07 hvn joined #salt
13:07 jcsp1 joined #salt
13:08 JDog Need a bit of a hand with mysql setup. Been at it all morning, so any help would be appreciated. Need set up mysql, but keep getting that it is not accessible with or without password. This is the pillar https://gist.github.com/jontyneedham/5b0dbf087df28fa8f92e
13:09 zzzirk joined #salt
13:10 wnkz_ joined #salt
13:11 shrihari joined #salt
13:11 d0xb joined #salt
13:13 calvinh_ joined #salt
13:13 jonatas_oliveira joined #salt
13:17 calvinh joined #salt
13:18 otter768 joined #salt
13:20 wnkz joined #salt
13:20 intellix joined #salt
13:21 jerematic joined #salt
13:24 phx am i seeing it right that with pygit2 it is possible to use the same git repo for both states and pillars, by setting a different rootdir?
13:24 CeBe joined #salt
13:26 robinsmidsrod joined #salt
13:27 sgate1 joined #salt
13:27 thehaven joined #salt
13:33 zadock joined #salt
13:34 zadock joined #salt
13:35 babilen phx: You don't have to use pygit2 for that
13:36 babilen (it works with gitpython too)
13:37 paulm- joined #salt
13:40 robinsmidsrod joined #salt
13:45 favadi joined #salt
13:45 hvn joined #salt
13:45 jtang joined #salt
13:48 elfixit joined #salt
13:50 yomilk joined #salt
13:52 Nazzy joined #salt
13:52 Nazzy joined #salt
13:58 toastedpenguin joined #salt
13:59 hobakill joined #salt
14:00 bhosmer joined #salt
14:05 JDiPierro joined #salt
14:08 nullptr joined #salt
14:14 forrest_ joined #salt
14:14 oldmantaiter joined #salt
14:14 Kelsar Quick Question: http://dpaste.com/31V6S3Z Is this the way to state an exact directory state?
14:16 toastedpenguin joined #salt
14:17 JWK joined #salt
14:18 bhosmer joined #salt
14:22 perfectsine joined #salt
14:24 racooper joined #salt
14:24 racooper joined #salt
14:30 mpanetta joined #salt
14:30 grok-johan joined #salt
14:31 perfectsine_ joined #salt
14:32 ksalman i have a custom grain which gets data from a database. It hits the database evertime a minion needs that grain. Is there any way to cache that data?
14:33 racooper joined #salt
14:33 paulm-- joined #salt
14:34 cpowell joined #salt
14:34 viq joined #salt
14:34 viq joined #salt
14:38 dude051 joined #salt
14:39 aqua^mac joined #salt
14:43 zzzirk joined #salt
14:44 favadi joined #salt
14:50 yomilk joined #salt
14:51 JDiPierro joined #salt
14:51 yomilk joined #salt
14:54 Kelsar who to call pkg.uptodate in a state? Either it does nothing for me or i get syntax error.
14:55 zzzirk_ joined #salt
14:56 JDiPierro Kelsar: I've never used it before but looking at the Doc it should be something like this: http://pastebin.com/Ms5sqTj3
14:58 JDiPierro Kelsar: Except "pkg.uptodate" not "up_to_date". It's early :P
14:58 Kelsar JDiPierro: no, error, but seems not to do anything... well will debug it later
14:58 JDiPierro Perhaps your system is already up to date? :P
14:59 Kelsar JDiPierro: it is too fast... portage is pretty slow nowadays ;)
14:59 ksalman Kelsar: it is new in 2014.7.0. Perhaps your salt is older?
14:59 Kelsar ksalman: no, 2014.7.0-r1
15:00 andrew_v joined #salt
15:00 zzzirk joined #salt
15:00 Kelsar atm i got a more urgent problem, who to manage a directory with managed files and clean
15:01 Kelsar right now it will delete the files (clean) and then reload them into it
15:03 zzzirk_ joined #salt
15:03 JDiPierro Do your file.managed have a require for the file.directory?
15:04 jeremyr joined #salt
15:06 kaptk2 joined #salt
15:06 perfectsine joined #salt
15:07 Kelsar JDiPierro: no, just looking it up
15:08 housl joined #salt
15:08 JDog Anyone for a MySQL question? Go on :-)
15:08 fredvd joined #salt
15:10 jerematic joined #salt
15:10 Sacro Is there a way to get the results of a previously run salt job without it running again?
15:10 Sacro saltutil.find_job(jid) seems to refresh the results which isn't what I require
15:13 Kelsar JDiPierro: k, this worked
15:13 Kelsar JDiPierro: or not... it will still do the same thing
15:17 pdayton joined #salt
15:18 nitti joined #salt
15:19 otter768 joined #salt
15:21 schlueter joined #salt
15:21 Mso150 joined #salt
15:21 favadi joined #salt
15:23 twellspring joined #salt
15:25 Kelsar Ok, now it won't remove the files, but will update them every run. why??
15:26 twellspring joined #salt
15:27 amustafa joined #salt
15:28 amustafa Hey guys, I'm having some trouble with the pip module and I was hoping to get some help.
15:28 amustafa I'm trying to install a python package from my own hosted server
15:29 jerematic joined #salt
15:29 amustafa I can pip install -f http://example.com/build_dis/ -u package
15:29 amustafa And I made this state to mimic that:
15:29 amustafa package:
15:29 amustafa pip:
15:29 grok-johan joined #salt
15:29 amustafa - installed
15:30 amustafa - find-links: http://example.com/build_distributions/release/package-core/
15:30 amustafa - upgrade: True
15:30 amustafa I know the name don't line up with my original example. They do in my code.
15:31 amustafa When I run the state I get " Failed to install package PACKAGE. Error: Downloading/unpacking PACKAGE   Could not find any downloads that satisfy the requirement PACKAGE
15:31 Brew joined #salt
15:32 amustafa the .pip/pip.log  error basically says that it's search in pypi for the package instead of my custom site
15:33 amustafa Not sure what I'm doing wrong. Any help is greatly appreciated!
15:33 Mso150 joined #salt
15:36 booly-yam joined #salt
15:37 Auroch joined #salt
15:42 hasues joined #salt
15:42 hasues left #salt
15:45 markm__ joined #salt
15:45 iggy amustafa: try index_url or extra_index_url
15:47 CeBe joined #salt
15:47 asyncsrc joined #salt
15:48 Ozack joined #salt
15:51 ksk joined #salt
15:51 ksk hey guys
15:51 ksk I would like to use pkgrepo - is it someway possible to provide the signing key as a file (in contrast to having it on a keyserver)? thanks!
15:53 iggy ksk: I'm pretty sure salt:// uri's are supported for key_url
15:53 franz_ joined #salt
15:53 iggy but we just put our in the root of our repo server
15:53 meylor joined #salt
15:54 franz_ need some env assistance
15:54 franz_ cmd.run:     - name: ant deploy-war     - env:       - ANT_HOME: '/opt/ant'       - JAVA_HOME: '/opt/java'       - PATH: '$ANT_HOME/bin:$JAVA_HOME/bin:$PATH'
15:54 franz_ /bin/bash: ant: command not found
15:54 amustafa Hey iggy, thanks for the suggestion. Things still are working exactly right. It does find the package now, but there are other erros. I'm going to keep digging.  Is this a bug in the salt code that "find-links" didn't work?
15:54 franz_ am i setting the PATH incorrectly?
15:55 karimb joined #salt
15:55 ksk iggy: ah okay, ill try that. thanks!
15:55 franz_ i.e. the command "ant" exists in /opt/ant/bin
15:56 iggy amustafa: never used that functionality so I don't know... did you look for existing bugs (possibly even closed if it's in newer version than you have)
15:57 iggy franz_: what if you put the full path to the binary?
15:57 amustafa There is a closed bug involving this tag, but it wasn't the same thing that was happening to me.
16:00 JWK joined #salt
16:02 franz_ /opt/ant/bin/ant: 1: /opt/ant/bin/ant: uname: not found                   Build failed
16:02 redzaku joined #salt
16:02 franz_ found the "ant" binary but in there, couldn't find "uname" - i.e. my path is horked
16:02 dude051 joined #salt
16:03 franz_ presumably, - env:
16:03 franz_ - PATH: '$ANT_HOME/bin:$M2_HOME/bin:$JAVA_HOME/bin:$PATH'
16:04 franz_ isn't working as I had hoped
16:04 dude051 joined #salt
16:05 markm__ joined #salt
16:07 booly-yam joined #salt
16:09 denstark joined #salt
16:09 paulm- joined #salt
16:12 phx how should i generate my state top.sls using python? if i do it over the py renderrer then it might not run on all minions since it's heavier than what the minion have access to
16:12 robinsmidsrod joined #salt
16:13 Ryan_Lane joined #salt
16:16 jab416171 joined #salt
16:16 wendall911 joined #salt
16:21 JDiPierro joined #salt
16:22 oeuftete joined #salt
16:22 abrahamrhoffman joined #salt
16:23 twellspring joined #salt
16:24 lz-dylan_ Hey, folks! I'm trying to store TLS certificates in pillar, and after some digging think that the file_tree external pillar is the best approach. Can someone tl;dr on how to use it? Per https://github.com/saltstack/salt/issues/19262 it's not documented.
16:25 jramnani` joined #salt
16:25 bhosmer_ joined #salt
16:27 vgasrs joined #salt
16:28 favadi left #salt
16:28 aqua^mac joined #salt
16:30 gladiatr joined #salt
16:30 gladiatr joined #salt
16:33 schristensen joined #salt
16:33 iggy you'll probably have to figure it out by reading the source
16:33 heaumer hi; what could to suddenly losts connexion with ~70% of minions?
16:34 heaumer last thing done were two shells commands (grep; grep)
16:34 heaumer salt is 2014.7.0; zmq 3.2.3
16:34 iggy heaumer: any defunct master processes? are the minions actually running (did they crash somehow, are they busy, etc.)
16:34 lz-dylan I'll give it a shot. I'm not entirely sure how this is organized, though -- is it considered an "external" pillar if it's in salt/pillar/? ie. will I have to fiddle with /etc/salt/master?
16:34 heaumer (and the shell commands were successfully executed everywhere; connexion was apparently lost a few minutes ago)
16:34 heaumer minions are running
16:35 heaumer and apparently, connection to master is still ESTABLISHED (netstat -a) from minions
16:35 iggy heaumer: so how are you determining they lost connection?
16:35 heaumer test.ping fails
16:35 heaumer (but for some hosts, eh)
16:35 iggy repeated attempts?
16:35 nitti_ joined #salt
16:35 heaumer yes
16:35 iggy sme hosts each time?
16:35 heaumer same host still showing up
16:35 heaumer let me double check that
16:36 lz-dylan iggy: okay so there's documentation in the comments :P I'll manage.
16:36 gladiatr joined #salt
16:36 gladiatr joined #salt
16:37 heaumer iggy: yes i think it was (most of the minions were restarted in the meantime, but i'm 90% sure)
16:39 lz-dylan generally speaking, I wish configuration data pertaining to how pillar works were readable out of /srv/pillar or the like; it feels like the necessary data to drop on a fresh new salt-master and have it operate as expected for a given pack of states is spread across a bunch of locations that don't logically get stored together.
16:40 jalbretsen joined #salt
16:40 iggy lz-dylan: ENOPARSE (aside from a general "the docs aren't organized the way I expect")
16:41 zzzirk joined #salt
16:41 karimb joined #salt
16:42 bhosmer_ joined #salt
16:43 lz-dylan iggy: Sorry, haven't had my coffee :) I wish I could grab a single directory from my running salt-master (say /srv) and drop it on a fresh instance with salt-master installed and have it either Just Work or be *really close* without having to munge around in /etc/salt/master
16:44 lz-dylan becase /etc/salt/master's defaults don't pull in things like external pillars or file roots. it'd be nice to be able to write those to a dotfile inside /srv/ or the like.
16:45 gladiatr lz-dylan, just make /etc/salt/master.d a symlink into /srv/
16:46 iggy yeah, I have a "bootstrap" script for my master's that does all that, then they take over from there with the salt-formula
16:48 lz-dylan Symlink might be the way to go for now. That's a good thought.
16:49 booly-yam joined #salt
16:50 thayne joined #salt
16:51 ksalman i have a custom grain which gets data from a database. It hits the database everytime a minion needs that grain. Is there any way to cache that data?
16:52 iggy I was under the impression that grains were only read at minion start
16:53 ksalman iggy: doesn't appear to be. The minion is hitting the database evertyime i do something like 'salt-call grains.items'
16:54 ksalman it's a #py file
16:55 ksalman maybe grains in /etc/salt/grains are only loaded at start
16:55 iggy I have a custom grain in /srv/salt/_grains and it only gets run at startup
16:55 gladiatr custom grains are run whenever a query is sent.  I think there is a master-side option to cache grains, though.  You'll need to give your custom grain some kind of cacheing capability.  It can be as simple as storing the data in a file (not /etc/salt/grains) and checking its age on query.  If its new enough, return the contents of the file, otherwise, requery.
16:56 ksalman iggy: that is where i have my custom grain
16:56 gladiatr iggy, interesting.  Dunno.  Its actually a useful thing when I was debugging (watch minion debug log)
16:56 iggy I stand corrected
16:56 ksalman gladiatr: that is a good idea
16:56 iggy I guess I just wasn't waiting long enough for them to update before
16:57 iggy but yeah, I think gladiatr's idea is probably your best bet... build that caching into your module
16:57 ksalman I didn't care earlier but now i have 1000+ minions and that is not helping the database =)
16:58 gladiatr ksalman, how much data is being returned?
16:58 ksalman gladiatr: actually not much but the database isn't beefy at all and we didn't anticipate hitting it this much
16:59 ksalman the database machine isn't beefy that is
16:59 nullptr joined #salt
17:00 twellspring joined #salt
17:02 gladiatr ksalman, cool.  Yup.  I'd just make a little nest under /var/cache/salt and put such data bits in there--just check the mtime when the grain is called then only refresh at the interval that makes sense.
17:02 ksalman gladiatr: thanks, neat idea =)
17:02 gladiatr no problem!
17:03 dbray joined #salt
17:08 Ch3LL_ joined #salt
17:08 MooreGuy joined #salt
17:09 intellix joined #salt
17:09 abrahamrhoffman1 joined #salt
17:10 Ch3LL_ joined #salt
17:13 abrahamrhoffman joined #salt
17:17 dude051 joined #salt
17:17 lz-dylan anyone know of another way to store entire files as data in pillar? I basically want to be able to dump my .crt and .key files for a particular class of server somewhere in /srv/pillar and then use states to populate those files on matching minions.
17:17 Ryan_Lane joined #salt
17:17 lz-dylan file_tree explicitly targets only by minion name.
17:17 lz-dylan am I just best off including the base64 encoded string right in a pillar SLS file? it'd work but it just doesn't feel right.
17:19 markm__ joined #salt
17:19 booly-yam joined #salt
17:19 ksk using states i would like to install a package and several config-files. however it seems i must not use more then one "file.managed" statement in one sls? how would i do that?
17:20 lz-dylan ksk: you can use as many file.managed as you like in a state, but they need to have different name properties
17:20 felskrone hey all, is rallytime around?
17:20 nullptr you can have as many as you want, they just can't have the same id
17:20 spookah joined #salt
17:20 lz-dylan s/name properties/id/ :)
17:20 otter768 joined #salt
17:21 ksk so, what would a name property or an id be in that case? http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.managed shows no such field
17:22 ksk right now my sls looks like this: http://paste.debian.net/140932/
17:22 KyleG joined #salt
17:22 KyleG joined #salt
17:23 lz-dylan ksk: it's the enclosing statement in the yaml. so you could have:        /etc/somefile:\n  file.managed\n    - property\n\n /etc/someotherfile:\n  file.managed\n    - property
17:25 lz-dylan ksk: https://gist.github.com/arubis/9a0d299114a551d46c2a
17:26 lz-dylan in fact, you may want also to drop the '- name: update-motd' depending on the name of the actual package you're installing (I don't know if your package name is 'update-motd' or just 'motd
17:26 lz-dylan ')
17:26 ksk ah okay, because yaml is key-value based and i have that "file.manage" key many times? guess I got it. many thanks guys!
17:26 Guest63579 joined #salt
17:26 ksk update-modt is some ubuntu package in that case
17:27 lz-dylan yeah. in that case you can either leave it as is, or also 'update-motd:\n  pkg.installed' is its functional equivalent
17:27 lz-dylan because pkg.installed will inherit name from the id of the enclosing statement
17:29 speed145a joined #salt
17:31 btorch joined #salt
17:31 btorch are there any major issues going from 0.17.5 to latest ? like anything that could happen that would break the states ?
17:31 hebz0rl joined #salt
17:32 iggy lz-dylan: we just keep ours in pillar files, yes
17:33 redzaku joined #salt
17:34 iggy felskrone: he doesn't do irc afaik
17:35 iggy btorch: probably... they try to maintain backwards compat, but there's no telling really without trying
17:35 forrest joined #salt
17:36 felskrone iggy: she(!) does on request, we've been fixing things a few times :-)
17:36 ndrei joined #salt
17:37 ulgoro joined #salt
17:39 tligda joined #salt
17:40 iggy I was thinking of someone else
17:40 btorch thanks
17:40 desposo joined #salt
17:41 lz-dylan iggy: gotcha, thanks.
17:41 iggy I was thinking of cachedout for some reason... weird mixup, but there you have it
17:43 iggy you said she, and I was picturing that goofy tom green looking avatar pic.... mind boggled
17:49 ulgoro Hello, does anybody know if there's been any change to the behaviour of pip.installed on Windows? I see nothing recently in this ticket: https://github.com/saltstack/salt/issues/8405
17:51 nitti joined #salt
17:53 shaggy_surfer joined #salt
17:54 JoeJulian If I'm reading this correctly, I could use reactor to trigger a highstate run after completing some state change on a minion. So, for instance, I could install and configure a ceph in a single-monitor configuration (on server1), then install another monitor (server1) and trigger a highstate on server1 to change its configuration file to include the newly added server2. Am I reading that right?
17:54 JoeJulian Gah, used server1 twice.
17:55 iggy JoeJulian: a better fit might be orchestration
17:55 iggy but it depends how you are expecting all this to work really
17:56 iggy with the reactor route, you'd have to fire events on your own to signal things being complete, etc.
17:57 JoeJulian thanks
17:57 iggy with the orchestration route you just make sure everything in your orchestration sls file is in the proper order
17:57 jeremyr joined #salt
17:57 zzzirk joined #salt
17:57 iggy but if you are planning on OSDs/MONs/etc to spin up automatically (via autoscale groups, etc.) then the reactor route might be better
17:58 shaggy_surfer joined #salt
17:58 spookah joined #salt
18:03 JoeJulian Yeah, reading about orchestration, that wouldn't work for us. Thanks for the confirmation that I'm not going the wrong way with reactor. I didn't want to throw a bunch of time at it if I was just way off.
18:03 _prime_ joined #salt
18:04 higgs001 joined #salt
18:04 _prime_ hi.  Any word on when salt 2014.7.1 will be released.  we're going into a seveal week freeze in about a week or so, and I'd like to get 2014.7.1 tested and if possible upgraded before then.
18:05 ekristen joined #salt
18:07 jcsp1 joined #salt
18:08 iggy _prime_: it's possible, I've heard QA has the final cut for it, but they had 2014.7.0 for 4 months... so...
18:11 booly-yam joined #salt
18:11 Mso150_y joined #salt
18:14 mohae joined #salt
18:14 rap424 joined #salt
18:16 shaggy_surfer joined #salt
18:17 mohae joined #salt
18:17 schlueter1 joined #salt
18:17 aqua^mac joined #salt
18:18 monkey66 joined #salt
18:20 hvn joined #salt
18:20 hvn joined #salt
18:25 schlueter joined #salt
18:29 cotton joined #salt
18:29 aquinas joined #salt
18:31 JDiPierro joined #salt
18:31 cotton hello - trying to deploy a dev/qa/prod docker container, using this as an example: http://docs.saltstack.com/en/latest/topics/tutorials/states_pt4.html
18:31 lz-dylan iggy: am I under the wrong impression that orchestration was being deprecated? or was that just overstate?
18:31 murrdoc joined #salt
18:31 cotton once I try to start it with "docker_role:dev", I get this message on the minion, and doesn't seem to match: [DEBUG   ] Attempting to match 'dev' in 'docker_role' using delimiter ':'
18:31 desposo joined #salt
18:32 cotton I'm following it verbatim, besides "s/webserver/docker/" and "s/foobarcom/registry/"
18:32 lz-dylan cotton: try to make it respond with `salt -G 'docker_role:dev' test.ping`
18:32 cotton No minions matched the target. No command was sent, no jid was assigned.
18:33 budman joined #salt
18:33 cotton nothing on minion debug even
18:33 lz-dylan gotcha. can you see what grains the minion has assigned? `salt 'minionid' grains.ls`
18:33 SneakyPhil did that grain get applied to the minion in the first place?
18:33 lz-dylan (replace minionid with the minion id of course :)
18:34 cotton http://pastebin.com/raw.php?i=ajUZt5db
18:35 cotton yes mition list works fine
18:35 cotton minion*
18:35 cotton root@salt:/srv/pillar# salt dev1.supersalon grains.ls|grep docker
18:35 cotton root@salt:/srv/pillar#
18:36 lz-dylan yeah, looks like the grain didn't get assigned
18:36 cotton how does my pillar look..?
18:36 lz-dylan I'm not sure just putting it in pillar would do it like that but I'm also still learning pillar myself, so don't take that as canonical
18:37 lz-dylan I've mostly used salt-cloud to assign grains on instantiation :/
18:37 SneakyPhil I am not very well versed in pillars or top files or that sort of thing, but does your pillar apply the openssh top file to each machine?
18:38 lz-dylan to the best of my reading it does look okay, but I am also having similar issues that I'm debugging in a different window
18:38 thayne joined #salt
18:39 chiui joined #salt
18:39 lz-dylan if I figure out what's going on on my end before you get an answer, I'll double back with more insight :)
18:40 lz-dylan you could always try `salt 'minion' saltutil.refresh_pillar`
18:42 cotton that didn't seem to help
18:42 cotton shouldn't the first state run add those grains?
18:43 JWK joined #salt
18:43 lz-dylan yeah, it should
18:43 lz-dylan I've gotta run for a bit, should be back in an hour. i'd love to know how this turns out!
18:44 MTecknology joined #salt
18:46 twellspring joined #salt
18:48 xliiv joined #salt
18:50 Ryan_Lane joined #salt
18:50 ipmb joined #salt
18:51 Ryan_Lane anyone have any idea why this may be failing for me and how I can solve it? https://github.com/saltstack/salt/pull/19791
18:52 stevednd Ryan_Lane: I went down a rabbit hole of stats stuff yesterday after your grafana message. what are your thoughts on sensu?
18:53 Ryan_Lane I haven't tried it :)
18:55 stevednd maybe someone else here will have some thoughts, even though it's not strictly the right channel
18:56 murrdoc i prefer ichinga 2 to sensu stevednd
18:56 TheThing joined #salt
18:56 meylor joined #salt
18:57 meylor I am running a state that is saying that uwsgi is running, though it is not running on the server. Thoughts on debugging?
18:58 ficek joined #salt
18:58 stevednd murrdoc: is icinga2 easy to extend to monitor different information? Like if I wanted to monitor things inside my applications?
18:59 murrdoc 2 reasons, transition from nagios is easy and you can restart the master
19:00 murrdoc and if the restart fails, then the slaves just keep working
19:09 aparsons joined #salt
19:09 booly-yam joined #salt
19:10 hobakill joined #salt
19:13 aparsons joined #salt
19:14 ajw0100 joined #salt
19:17 druonysus joined #salt
19:17 druonysus joined #salt
19:21 otter768 joined #salt
19:23 Edgan Can anyone point me to sls that treats user management as functions?
19:25 murrdoc stevednd:  and yes you can monitor anything that you can script, and it links up with graphite really well
19:27 babilen murrdoc: Is ichinga 2 just a typo of icinga2 or some hidden super awesome software that mere mortals are unable to find?
19:27 signull joined #salt
19:28 murrdoc oh man
19:28 murrdoc thanks for pointing that out babilen
19:28 murrdoc https://www.icinga.org/icinga/icinga-2/distributed-monitoring/
19:28 babilen (it's just that you seem to use that consistently)
19:28 babilen yeah, I know icinga :)
19:28 murrdoc i was spelling it phonetically for some reason
19:28 shel3over joined #salt
19:28 redzaku joined #salt
19:29 babilen eyechingar!
19:29 murrdoc :|
19:29 babilen ack, rock on
19:29 murrdoc hahah
19:29 murrdoc #fridayftw
19:29 pdayton joined #salt
19:29 murrdoc flying back home today, sleeping in a hotel is not all its cracked up to be
19:29 signull Hey can anyone point me to a good resource to go RTFM if I want to use docker in aws have it managed by saltstack?
19:31 robawt murrdoc++
19:33 redzaku joined #salt
19:34 cpowell joined #salt
19:35 nickdew joined #salt
19:35 Sypher Hi folks! I was hoping someone could help me (brainstorm) about an salt-minion issue I have.
19:35 Sypher We have two parties who both want to use the minion but differently. Party 1 wants to issue commands while party 2 wants to do commands + (scheduled) highstates.. But they cannot share a server nor files/keys
19:36 babilen signull: I take it that you are familiar with: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.dockerio.html
19:36 signull I've seen that but i feel like there is a disconnect between the dockerio sls and salt-air
19:37 Sypher what's practical here? Running two minions with different roots (pki/cache/config)? Running one in virtualenv and one normal?
19:37 Sypher the latter doesn't feel very practical to me maintenance wise
19:38 signull babilen: ^my previous message^
19:39 pdayton joined #salt
19:39 nickdew Sypher: we have several different snowflakes in our environment too. I make heavy use of grains and pillars to deal with them.
19:40 babilen signull: I'm not familiar with salt-air
19:40 Sypher nickdew: the trouble of this is that both parties need individual access to salt. Something most likely can only be achieved with multiple minion daemons
19:41 babilen Sypher: Maybe a multi syndic approach? (not sure if that would be feasible though)
19:42 Sypher babilen: both masters should have 0 relation to eachother :) There are basically two standalone systems. At one point its likely that both do highstates of their own, the catch then is to coordinate which party does what
19:43 Sypher its probably not a widely used approach :-)
19:43 schlueter1 joined #salt
19:43 dude051 joined #salt
19:44 iggy lz-dylan: just overstate (being deprecated in favor of orch)
19:44 troyready joined #salt
19:45 robinsmidsrod joined #salt
19:49 abrahamrhoffman anyone have any experience with RAET vs ZeroMQ?
19:49 abrahamrhoffman Large cluster running ZeroMQ 4 currently - looking for pros/cons
19:49 markm__ joined #salt
19:50 iggy abrahamrhoffman: con for raet... it's not ready
19:51 robinsmidsrod joined #salt
19:51 abrahamrhoffman what kinds of errors are you seeing?
19:51 Brew joined #salt
19:52 abrahamrhoffman we currently get +- .5% loss
19:52 abrahamrhoffman hoping RAET would push that lower
19:54 DolourousEdd1 joined #salt
19:54 robinsmidsrod joined #salt
19:54 arno__ joined #salt
19:57 zadock joined #salt
19:57 iggy it should be much improved in 2015.2, I'd just wait until then personally
19:58 dude051 joined #salt
19:58 robinsmidsrod joined #salt
19:59 robinsmidsrod joined #salt
20:02 moos3 joined #salt
20:06 aqua^mac joined #salt
20:09 hvn joined #salt
20:09 hvn joined #salt
20:10 nmadhok joined #salt
20:13 shaggy_surfer joined #salt
20:18 buxy joined #salt
20:18 zzzirk joined #salt
20:19 * buxy is looking for a saltstack employee to include https://github.com/rhertzog/debootstrap-formula in the official set of formulas (following the recommendation of the documentation)
20:20 desposo joined #salt
20:21 KyleG left #salt
20:22 badon joined #salt
20:23 bfoxwell joined #salt
20:23 signull_ joined #salt
20:23 diegows joined #salt
20:24 iggy buxy: it doesn't need to be an employee, there are people outside the company that can add them
20:25 iggy babilen seems to be the only one around right now
20:25 buxy iggy: fine. s/saltstack employee/someone/ :)
20:25 iggy oh, whiteinge too
20:25 buxy I guees over the WE a non-employee is more likely even :-)
20:25 ksalman speaking of custom grains hitting the database every time. Shouldn't grains be static? I'd think that the minion would run the custom grain module one time and then have the grains in memory. If it hits the databse every time, doesn't that make grains dynamic?
20:26 Slimmons joined #salt
20:27 iggy ksalman: yeah, that's how I thought they worked as well, but I guess they assume people won't be doing database lookups for grains (as that tends to imply something a little more dynamic by definition)
20:28 ksalman maybe it's a bug with custom grains? =)
20:29 ksalman iggy: even if it's not ddoing a database looking. Say it's just reading from a file, it seems like it'd read form the file every time then
20:29 iggy I dunno, but if you open a ticket, let me know the number. I'd be interested to know more about it
20:29 ksalman okay
20:29 ndrei joined #salt
20:39 williamthekid joined #salt
20:44 StDiluted joined #salt
20:44 Slimmons setting up gitfs for the first time.  I've looked through the gitfs backend walkthrough, and one thing I don't understand is on their example, where it says "Assume that each repository contains some files:", it looks like the file structure they are writing out is in a top.sls, but it's actually setting the top.sls.  Where would that actually go, if not in the top.sls. https://salt.readthedocs.org/en/v0.17.4/topics/tutorials/gitfs.ht
20:44 StDiluted can someone tell me what the syntax for a compound grain match in a pillar top.sls should look like?
20:46 bhosmer_ joined #salt
20:46 murrdoc joined #salt
20:46 babilen iggy: I can't unfortunately. I would, but GH doesn't allow people to grant privileges on such a fine grained level
20:47 babilen buxy: hey! Your best bet would be to write a message to the ML (salt-users on google groups). It'll be added soon. Thank you for writing a debootstrap formula, that looks quite useful :)
20:49 StDiluted should this work: https://gist.github.com/dginther/8065559209e6fc0f5a2f
20:49 iggy babilen: huh, I thought you could add formulas... weird so how does that normally work?
20:50 iggy StDiluted: G@key:value
20:50 babilen iggy: Only owners can add formulas but not contributors
20:50 StDiluted iggy: that dodnt seem to work
20:50 iggy and - match: compound
20:50 StDiluted didnt
20:50 StDiluted yeah, i updated the gist with that method which didnt seem to work
20:50 StDiluted so the real way to do it is the second way I have listed in the gist, yes?
20:50 iggy babilen: StDiluted yes
20:51 iggy err -babilen
20:51 StDiluted that’s weird because that actually doesn’t work
20:51 StDiluted and the wrong way, what I have listed first, is working
20:51 iggy babilen: is there  a way to see the owners? all I ever found was a contributors list (and thought they all could)
20:52 rogst joined #salt
20:52 iggy StDiluted: unless it's just stripping everything after the first space (so only matching ec2_environment), that shouldn't work
20:53 ksalman iggy: https://github.com/saltstack/salt/issues/19801
20:55 iggy danke
20:55 babilen iggy: https://github.com/orgs/saltstack-formulas/teams/owners
20:56 babilen (biite)
20:56 babilen -i+t
20:57 cotton_ joined #salt
20:58 cotton joined #salt
20:58 StDiluted iggy: hm, at first it wasn’t working but now it seems to be
20:58 StDiluted weird
21:00 perfectsine joined #salt
21:00 nitti_ joined #salt
21:01 bhosmer_ joined #salt
21:01 nitti_ joined #salt
21:02 gattie joined #salt
21:07 nitti joined #salt
21:08 nitti joined #salt
21:20 lz-dylan cotton: any luck?
21:22 otter768 joined #salt
21:23 perfectsine joined #salt
21:23 cotton not yet, had to move to another project for the day unfortunately
21:24 cotton both -I and -G are saying no matches lz-dylan
21:24 lz-dylan so, here's what I managed to gather
21:24 cotton I may touch base again monday, thanks for your help so far
21:24 lz-dylan pillar != grains :P
21:24 cotton right
21:25 lz-dylan try out `salt 'yourminion' pillar.get docker_role`
21:25 lz-dylan it's probably there. if you want it in grains you'll have to set it another way.
21:26 lz-dylan if you're all closed up, I should be around Tuesday :)
21:27 shaggy_surfer joined #salt
21:29 druonysuse joined #salt
21:29 cpowell joined #salt
21:29 druonysuse joined #salt
21:29 gladiatr joined #salt
21:33 KyleG joined #salt
21:33 KyleG joined #salt
21:35 druonysuse joined #salt
21:35 Singularo joined #salt
21:36 twellspring joined #salt
21:37 whiteinge buxy: i can grab that formula
21:38 buxy whiteinge: great, let me know if you have any question
21:39 jeremati_ joined #salt
21:40 gladiatr left #salt
21:40 gladiatr joined #salt
21:41 desposo joined #salt
21:42 whiteinge buxy: you should get an invite to join the COntributros tea
21:44 moos3 joined #salt
21:44 buxy whiteinge: got it, thanks!
21:47 iggy ksalman: did you see grains_cache: False in the minion config?
21:47 ksalman iggy: no
21:48 cpowell joined #salt
21:49 iggy not sure if that does what we're thinking, but it does defualt to False
21:49 JDiPierro Jinja doesn't seem to like this, Is there a way to do something like this or am I barkin' up the wrong tree? {%- set service_name = {{ service.get('environment','production') }}_{{ service.get('service') }} -%}
21:49 ksalman huh
21:49 ksalman i don't see grains_cache in here http://docs.saltstack.com/en/latest/ref/configuration/minion.html
21:51 iggy ksalman: it's in the example in the source (conf/minion)
21:51 ksalman i see that it was added here http://docs.saltstack.com/en/latest/topics/releases/2014.1.0.html
21:51 ksalman okay
21:51 bhosmer joined #salt
21:51 ksalman interesting
21:51 iggy so yeah, looks like at the very least, there's a docs problem
21:51 ksalman i'll pay with it
21:51 ksalman play
21:53 iggy JDiPierro: what is service?
21:53 JDiPierro A dictionary
21:53 FRANK_T joined #salt
21:54 iggy {%- set service_name = service.get ~ "_" ~ service.get -%}
21:54 housl joined #salt
21:55 aqua^mac joined #salt
21:55 JDiPierro Sweet deal!
21:58 hvn joined #salt
21:58 ksalman huh!
21:58 Edgan Does salt have the equivalent of puppet defines? So I can reuse the same code over and over.
21:59 ksalman iggy: grains_cache appear to do the trick. Even for my custom grain that hits an external API
21:59 iggy ksalman: excellent... leave the bug open, but amend it to say the option should be documented maybe
21:59 ksalman okay
22:00 dude051 joined #salt
22:00 iggy Edgan: assume we are stupid about puppet... in face don't assume
22:00 iggy *in fact
22:01 murrdoc edgan, you would make a state in python
22:02 murrdoc and then call it in your jinja/yaml states
22:02 murrdoc put it in the _states dir
22:02 Edgan murrdoc: But I just want to combine two existing states. A state module is the closest I have gotten to the idea, but going out into python when I just want to combine two things seems excessive.
22:02 ksalman iggy: you'd think grains_cache: True would be the default
22:03 ksalman i'll just comment on bug
22:03 murrdoc Edgan:  you can look at 'extending states'
22:03 murrdoc that might solve your use case
22:03 hebz0rl joined #salt
22:04 Edgan murrdoc: thanks
22:06 dude051 joined #salt
22:09 Edgan murrdoc: That just seems to append some stuff to one sls from another. I want to combine user and ssh_auth into want reusable thing. Otherwise I will be repeating that code every time I am saying create a list of users.
22:09 murrdoc ah
22:09 murrdoc then u use formulas
22:10 murrdoc i like this one for that use case
22:10 murrdoc https://github.com/saltstack-formulas/users-formula
22:10 iggy I think the user-formula has ssh support anyway
22:11 Edgan I saw that too, but can I reuse it anywhere?
22:12 GP joined #salt
22:12 GP Hi
22:12 murrdoc you use it with a pillar, and in the pillar you will specfy the various users
22:12 murrdoc which is probably not what you want
22:13 murrdoc the way i did the equivalent of define from puppet is to make my own state which calls out to the salt modules as a group
22:13 murrdoc its not ideal yes, but it works for my use case
22:15 murrdoc but the salt way is to define pillars and iterate on those
22:16 Edgan murrdoc: Which makes sense. I think the users formula might be what I want, but it has almost no docs and the example is super basic.
22:17 babilen https://github.com/saltstack-formulas/users-formula/blob/master/pillar.example is basic?
22:17 babilen What are you missing?
22:18 murrdoc Edgan:  pillar.example is the usage docs
22:18 murrdoc sorry usages is just include:\n\t- users
22:19 babilen Ah, yeah. That should be noted if you aren't familiar with formulas. The README might be a little bit sparse, but the pillar.example should have an example of everything you can do.
22:19 murrdoc and then what all the formula does you have to read the sls files to understand what all you are getting with it
22:19 Edgan babilen: That is just a list of three users. I want groups. Then I want to be able to call the groups from where ever. I think it might be better if it was agroup: auser buser cuser  bgroup: xuser yuser zuer
22:19 murrdoc that example file is not used
22:20 murrdoc you haev to change it to fit your needs
22:20 Edgan murrdoc: I know, but the example is super basic, and I am just wrapping my brain around formulas.
22:20 babilen Edgan: The user formula is entirely data driven and you can organise the data in whatever way you want. If you want to have a users/foo_group.sls in your pillar that contains three users then you can do that.
22:20 murrdoc ah
22:21 murrdoc formulas are 'grouped states configured by pillars'
22:21 murrdoc imho
22:21 Guest67932 Hi , I have to two states defined in my sls file. I want one state to execute before another . But it does not seems to follow. [ Intent is to read a file on disk and then install packages based on file contents.]  In order to achieve that , I have to run <b>salt * state.highstate</b>    two times.
22:22 Guest67932 I am trying to do that in single run
22:22 murrdoc so like in puppet you would have a 'define' for nginx where you could pass in the paths to the config files
22:22 murrdoc and it would set up the stuff as needed
22:22 babilen Yeah, all "moving bits" are configured via pillars (or populated with sensible defaults). The users formula doesn't do anything by default (that is if you don't define any pillar values)
22:22 murrdoc in salt, you use the formula, and then the paths are setup in the pillar
22:24 signull joined #salt
22:25 giantlock joined #salt
22:29 echtish joined #salt
22:31 rawzone joined #salt
22:31 conan_the_destro joined #salt
22:37 nitti_ joined #salt
22:39 schristensen joined #salt
22:39 redzaku joined #salt
22:40 rogst joined #salt
22:40 druonysuse joined #salt
22:40 ldlework Did anything recently change in salt that would screw up like input/output buffering or anything like that?
22:41 Ryan_Lane joined #salt
22:41 ldlework I have a command that I run during provisioning that is supposed to pipe yes | cmd, and this just doesn't work anymore, without change.
22:41 ldlework We install salt from source each deploy so I'm just wondering if salt updated recently
22:43 viq joined #salt
22:43 iggy ldlework: I don't know of anything specific, but we've had some stuff break recently too (a tool that called gpg started failing saying no tty available)
22:44 ldlework no tty would perfectly explain it
22:44 ldlework now the question is whyy
22:51 utahcon_ Data failed to compile: No matching sls found -- that is genereated on the master, right? That isn't being reported from the minion
22:52 redzaku joined #salt
22:54 iggy that's from the minion (after it syncs everything from the master)
22:55 iggy data failed to compile though... so there's got to be more to the error message (or you need to run it with salt-call -l debug)
22:55 kermit joined #salt
23:05 twellspring joined #salt
23:05 ckao joined #salt
23:07 twellspring joined #salt
23:11 twellspring joined #salt
23:17 wincyj hello
23:17 wincyj i recive that error
23:17 wincyj salt-ssh: error: no such option: --askpass
23:17 wincyj when
23:17 wincyj salt-ssh --askpass -v ${HOST_NAME} ${DEBUG} state.highstate ${TEST}
23:17 wincyj how to fix that?
23:18 wincyj on the other machine it worked well
23:21 jonatas_oliveira joined #salt
23:22 signull_ joined #salt
23:23 otter768 joined #salt
23:25 druonysuse joined #salt
23:26 brycekahle joined #salt
23:26 brycekahle is it possible to get salt-cloud to use a git filesystem?
23:27 stevednd brycekahle: you can manage your cloud configs on the master the same way you manage other files
23:28 brycekahle ok. I wasn't sure if it would use the fileserver or not
23:28 brycekahle thanks stevednd
23:30 wincyj anyone?
23:30 wincyj :>
23:30 wincyj seems like my version of salt-ssh doesnt have --askpass
23:30 wincyj ;/
23:30 wincyj salt-ssh 0.17.5
23:31 iggy wincyj: yeah, it was added in 2014.1
23:32 wincyj is there package for mint/ubuntu?
23:32 iggy there's a ppa
23:33 wincyj ok will check
23:39 dstokes joined #salt
23:40 aparsons joined #salt
23:40 JDiPierro joined #salt
23:41 stevednd anyone have any issues with archive.extracted? Particularly on an ubuntu 10.04 machine?
23:42 dave_den left #salt
23:43 aqua^mac joined #salt
23:46 iggy you deserve problems using 10.04
23:46 hvn joined #salt
23:46 hvn joined #salt
23:47 stevednd iggy: yes, I am aware
23:47 stevednd in the process of getting off of it, but still have to maintain some of them until that time
23:49 druonysuse joined #salt
23:51 iggy what problems are you actually having?
23:52 stevednd https://gist.github.com/dnd/7fe6e9450e685d19b3f0
23:52 stevednd it's not specific to this file either
23:52 stevednd it happens any time I use archive.extracted
23:52 stevednd the file is in the right location
23:53 stevednd I thought maybe it wasn't waiting for the file to arrive, but after the first time the file is pulled from the cache
23:53 stevednd so obviously it's there

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