Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-11-04

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

All times shown according to UTC.

Time Nick Message
00:00 conan_the_destro joined #salt
00:01 ipmb joined #salt
00:01 diehard_ joined #salt
00:04 glyf joined #salt
00:05 tafa2 joined #salt
00:07 tafa2 joined #salt
00:07 ice799 joined #salt
00:08 kermit joined #salt
00:09 ice799 Hi - I've been trying to track down some docs or some tutorial or something about how to write salt modules and best practices around that. Anyone know where I might be able to find more info? I have been searching around docs.saltstack.com a bunch but can't seem to find anything that shows a good example of how to build a module and what is expected
00:09 tafa2 joined #salt
00:09 aparsons joined #salt
00:13 ipmb_ joined #salt
00:13 alex-mesos joined #salt
00:15 baconbeckons joined #salt
00:17 ipmb joined #salt
00:24 mdasilva joined #salt
00:27 murrdoc http://docs.saltstack.com/en/latest/ref/modules/index.html
00:32 halfss joined #salt
00:32 jforest joined #salt
00:37 Ryan_Lane joined #salt
00:39 forrest_ joined #salt
00:45 diehard_ joined #salt
00:45 tligda joined #salt
00:50 diehard_ joined #salt
00:53 cads joined #salt
00:55 smcquay joined #salt
00:55 aqua^mac joined #salt
00:56 snuffeluffegus joined #salt
00:59 StDiluted joined #salt
01:00 meylor joined #salt
01:04 ipmb joined #salt
01:09 baniir joined #salt
01:12 aparsons joined #salt
01:12 crocket joined #salt
01:13 crocket Vagrant makes it a lot easier to debug salt states.
01:13 aurynn vagrant makes everything easier
01:14 racooper joined #salt
01:16 crocket up, destroy, up, destroy
01:16 crocket Clean, easy, fast
01:18 halfss joined #salt
01:23 teepark the salt provisioner has been pretty buggy though, I end up resorting to shell provisioning a lot
01:24 aurynn I use shell provisioning to get to bring up a salt master, then provision my secondaries from that
01:25 rojem joined #salt
01:27 baconbeckons joined #salt
01:27 notpeter_ joined #salt
01:34 notpeter_ joined #salt
01:48 notpeter_ joined #salt
01:53 NaPs joined #salt
01:55 crocket left #salt
01:58 diehard_ joined #salt
02:04 MindDrive joined #salt
02:07 _sifusam_ joined #salt
02:08 TheThing joined #salt
02:08 nitti joined #salt
02:12 eightyeight joined #salt
02:23 esogas_ joined #salt
02:27 diehard_ joined #salt
02:28 otter768 joined #salt
02:36 tristianc joined #salt
02:39 mdasilva joined #salt
02:47 jalaziz_ joined #salt
02:49 jalaziz joined #salt
03:03 dalexander joined #salt
03:07 b1nar1 joined #salt
03:07 mosen joined #salt
03:08 bezeee joined #salt
03:11 thayne joined #salt
03:12 dmccc joined #salt
03:13 rgrtom joined #salt
03:13 rgrtom yho
03:16 meylor joined #salt
03:24 meylor joined #salt
03:25 rgrtom ext
03:25 rgrtom \q
03:29 psidox joined #salt
03:30 zlhgo joined #salt
03:33 thayne joined #salt
03:34 bhosmer joined #salt
03:35 n8n joined #salt
03:36 rojem joined #salt
03:45 psidox joined #salt
03:48 ndrei_ joined #salt
03:48 ndrei joined #salt
03:54 Ryan_Lane joined #salt
04:05 ipmb joined #salt
04:05 srage_ joined #salt
04:08 jalaziz joined #salt
04:10 dalibro joined #salt
04:16 I2PNinja left #salt
04:19 aparsons joined #salt
04:20 TyrfingMjolnir joined #salt
04:22 meylor joined #salt
04:23 mage_ joined #salt
04:28 mdasilva joined #salt
05:01 ajolo joined #salt
05:03 ajolo joined #salt
05:06 smcquay joined #salt
05:06 cpowell joined #salt
05:12 kermit joined #salt
05:23 bhosmer joined #salt
05:25 tligda joined #salt
05:33 TheThing joined #salt
05:48 ndrei joined #salt
05:48 ramteid joined #salt
05:49 mosen joined #salt
05:52 gmcwhistler joined #salt
05:53 ndrei_ joined #salt
05:55 ghanima joined #salt
05:57 bhosmer joined #salt
05:57 diehard_ joined #salt
05:58 skullone anyone run salt without agent certs? is it possible?
06:02 jhauser joined #salt
06:02 psidox joined #salt
06:03 ldlework Does Salt still run its own handrolled crypto?
06:10 Setsuna666_ joined #salt
06:11 zartoosh joined #salt
06:12 shookees joined #salt
06:14 Hammerheart joined #salt
06:14 CSa_ joined #salt
06:16 schristensen_ joined #salt
06:17 mdasilva joined #salt
06:17 jacksontj_ joined #salt
06:17 Blacklite joined #salt
06:27 patarr joined #salt
06:27 patarr joined #salt
07:04 meylor joined #salt
07:12 bhosmer joined #salt
07:13 flyboy joined #salt
07:19 cberndt joined #salt
07:24 ramishra joined #salt
07:24 duncanmv joined #salt
07:25 SpX joined #salt
07:30 meylor joined #salt
07:42 tinuva joined #salt
07:42 lothiraldan joined #salt
07:53 lothiraldan joined #salt
07:55 Mso150 joined #salt
07:56 TyrfingMjolnir joined #salt
07:57 ramishra joined #salt
07:58 trikke joined #salt
07:59 b1nar1_ joined #salt
08:03 h8 joined #salt
08:04 slav0nic joined #salt
08:06 mdasilva joined #salt
08:07 tomspur joined #salt
08:10 P0bailey joined #salt
08:10 P0bailey joined #salt
08:11 eseyman joined #salt
08:14 zz_Cidan joined #salt
08:14 Cidan joined #salt
08:19 lcavassa joined #salt
08:25 CeBe joined #salt
08:34 TyrfingMjolnir joined #salt
08:36 hotbox joined #salt
08:37 Mso150_x joined #salt
08:45 agend joined #salt
08:45 cberndt joined #salt
08:49 felskrone joined #salt
08:54 cofeineSunshine joined #salt
08:54 cofeineSunshine joined #salt
08:55 intellix joined #salt
08:56 iwishiwerearobot joined #salt
08:59 karimb joined #salt
09:01 bhosmer joined #salt
09:04 ajolo joined #salt
09:05 agend joined #salt
09:08 ajolo joined #salt
09:09 TyrfingMjolnir joined #salt
09:11 rjc joined #salt
09:12 brayn joined #salt
09:12 TyrfingMjolnir joined #salt
09:13 CeBe1 joined #salt
09:20 CycloHex joined #salt
09:21 heww joined #salt
09:21 N-Mi joined #salt
09:30 viq skullone: there's salt-ssh, but otherwise I don't think so
09:35 zartoosh joined #salt
09:36 duncanmv joined #salt
09:41 wnkz joined #salt
09:45 fbettag joined #salt
09:47 intellix joined #salt
09:48 swa_work joined #salt
09:49 wnkz joined #salt
09:55 mdasilva joined #salt
09:56 alex-mesos joined #salt
09:58 Schmidt left #salt
10:33 PI-Lloyd joined #salt
10:34 workingcats skullone, why would you want to do that, and what do you plan to use instead..?
10:34 yomilk joined #salt
10:35 brayn Mormning. I'm running salt in a AWS VPC which is in development so not traffic spikes present. One of the issues I have is that hosts (there are about 8 present) sometimes respond to commands like test.ping very quickly but other times even with -t 60 some won't appear until I run the command the second time
10:47 intellix joined #salt
10:50 bhosmer joined #salt
10:50 stooj joined #salt
10:51 stooj Hi. Trying out salt for the first time - I'm trying to get flavours working, but I'm getting a no matching sls found error.
10:51 irctc326 joined #salt
10:51 babilen flavours?
10:52 stooj Sorry, not flavours. Formulas.
10:52 stooj Specifically users-formula
10:52 samed joined #salt
10:52 stooj I'm trying to use gitfs, have installed python2-pygit2 (on my arch master)
10:52 viq stooj: what exactly did you do so far?
10:53 irctc326 How can i run this through salt state  docker inspect --format "{{ .NetworkSettings.IPAddress }}"
10:53 stooj Added git as a fileserver_backend
10:53 babilen Okay, could you paste whatever parts of your setup should make it work (e.g. master config, local state trees, ...), your pillar and the output to, say, http://refheap.com ?
10:53 stooj Forked the repo and added my version in gitfs_remotes
10:54 irctc326 i am getting failed: Jinja syntax error: unexpected '.';
10:54 viq irctc326: would you run the command exactly like that from a terminal?
10:55 viq including the {{}} ?
10:55 irctc326 yes
10:55 stooj Here is my master file, with all commented lines removed: https://www.refheap.com/92681
10:55 irctc326 i included { { } }
10:56 irctc326 i can run the same command in termina;
10:56 irctc326 terminal*
10:56 viq irctc326: {{}} is jinja syntax, and it tries to evaluate that, failing. You'd have to escape it somehow
10:56 stooj Here's my /srv/salt/top.sls file: https://www.refheap.com/92682
10:57 viq stooj: and how are you trying to use the users-formula?
10:57 irctc326 http://bitjudo.com/blog/2014/05/13/partial-continuous-deployment-with-docker-and-saltstack/
10:57 irctc326 he says its working
10:57 stooj viq - I understood that if you added it to the gitfs_remotes and then referenced the directory from top.sls, it should just "find" it
10:58 viq irctc326: it's his blog, he can say whatever he likes ;)
10:58 viq stooj: and what errors are you getting?
10:58 viq stooj: also, did you restart the master?
10:58 stooj viq - heh, yep. Restarted.
10:59 stooj viq: Error message: https://www.refheap.com/92683
11:00 babilen stooj: Please also run the master in debug mode with "salt-master -ldebug" and check for errors that might occur during a git clone. Am not using pygit2, but you are aware of the dependencies as noted in http://docs.saltstack.com/en/latest/ref/file_server/all/salt.fileserver.gitfs.html, aren't you?
11:01 babilen You can run "salt-run fileserver.update" to trigger an upgrade.
11:02 viq irctc326: https://github.com/audreyr/cookiecutter/issues/11
11:02 stooj babilen: OK. I will. I can confirm that I have libgit2 and pygit2 installed
11:02 viq irctc326: or http://grokbase.com/t/gg/ansible-project/1488csaxdy/how-to-escape-a-variable
11:03 babilen stooj: Can you clone that repo manually as the user salt runs as? Without *any* interaction?
11:04 stooj Hmm - debug output: Git fileserver backend is enabled in configuration but could not be loaded, is GitPython installed?
11:05 stooj I thought it prefered pygit2
11:05 aquinas joined #salt
11:06 babilen That might very well be a thing introduced in 2014.7 - not entirely sure, but you could try setting http://docs.saltstack.com/en/latest/ref/configuration/master.html#std:conf_master-gitfs_provider explicitly
11:06 babilen Ah, that is new in 2014.7
11:06 babilen "Beginning with version 2014.7.0, both pygit2 and Dulwich are supported as alternatives to GitPython. "
11:07 babilen So, either run that version or install/use GitPython
11:07 stooj Ah, great! Thanks for your help
11:08 zekoZeko_ joined #salt
11:09 irctc326 thanks @viq and @irctc326
11:09 irctc326 salt-call  cmd.run "docker inspect --format  \{\{..NetworkSettings.IPAddress\}\}  container" works fore me
11:16 stooj Hmm - manually specifying pygit2 with gitfs_provider still didn't work, but installing GitPython seems to have fixed it. Thanks viq, babilen
11:17 ramishra joined #salt
11:17 viq :)
11:28 babilen stooj: pygit2 is *not* supported in the version you are, presumably, running.
11:30 stooj So I gather. I couldn't find any mention of support being dropped in the repo, but I must have missed it or something
11:32 viq stooj: rather "wasn't added yet"
11:33 viq stooj: I assume you're running 2014.1, a lot of docs reflect 2014.7 for which we're still waiting
11:33 viq ooh ooh ooh! https://pypi.python.org/packages/source/s/salt/salt-2014.7.0.tar.gz  !!!
11:34 viq 16h ago
11:35 stooj Haha - You mean if I'd waited a couple of days, everything would have worked perfectly? :D
11:35 viq probably ;)
11:39 babilen % NEWMAJORVERSIONPROBLEMS, yes
11:41 stooj Are formulas the way to go? This users one isn't working all that well, since the rackspace image I am using includes ubuntu as uid 1000, and I'd like to remove that user and replace it with me
11:41 stooj (with the same uid)
11:42 stooj Do you find that there are usually edge-cases that mean you always roll your own states, even for basic stuff?
11:43 glyf joined #salt
11:43 mdasilva joined #salt
11:46 ShibMcNe stooj: you can create pull requests if something is missing in the formulas, rencently I made one because the apache formula was acting strange with ubuntu 14.10 and/or Apache 2.4.
11:46 brayn I'm running salt in a AWS VPC which is in development so not traffic spikes present. One of the issues I have is that hosts (there are about 8 present) sometimes respond to  commands like test.ping very quickly but other times even with -t 60 some won't appear until I run the command the second time. Thanks
11:46 ShibMcNe I try to stick to formulas, and enhance them to cope with my use case
11:47 stooj ShibMcNe: Give me a couple of days to find my way around first :D
11:48 sgate1 joined #salt
11:49 ShibMcNe brayn: maybe https://github.com/saltstack/salt/issues/14343 ?
11:50 brayn ShibMcNe: Looking at it now, thanks
11:50 ShibMcNe I think I encountered the same issue twice, but did not have time to look further into it.
11:51 intellix joined #salt
11:54 Reiner030 joined #salt
11:55 brayn ShibMcNe: It might be. Sometimes it works instantly even after a long pause. What whould be the best way of tracking it down, I see the issues has no real solution there
11:56 ShibMcNe using debug logs on the master and minions, and then try to reproduce the problem
11:57 brayn ShibMcNe: OK, I'll have to take an evening and figure this out
11:58 ShibMcNe brayn: I'll also try on my side and comment on the issue if I find something
11:59 bhosmer joined #salt
11:59 Reiner030 Hello, i can't find a solution in docu/blogs if there is the possibility by salt-cloud/salt to write the AWS public ips and other infos that are written after setup and which are not requestable by clients - by salt - to push into pillars or into the mine... (btw. the "unsalty" way could be a request to http://169.254.169.254/latest/meta-data/public-ipv4 I found yet out). Is there perhaps informations I missed? Thx.
11:59 brayn Thanks for the help
11:59 nitti joined #salt
12:00 Reiner030 (the "unsalty" way I found in this blog entry: http://www.ducea.com/2009/06/01/howto-update-dns-hostnames-automatically-for-your-amazon-ec2-instances/ )
12:05 scottpgallagher joined #salt
12:12 cetex joined #salt
12:12 cetex hm. i have an issue.
12:13 cetex when a package has been installed as a dependency to something, then included in salt to be installed, then we uninstall the main package that pulled the dependency the package isn't set as manually installed, so it's removed when doing uninstall of the main package
12:14 shookees joined #salt
12:14 Reiner030 yes, thats Debian "cleanup" behavior... (i guess you have Debian as I)
12:14 shookees joined #salt
12:15 cetex the actual issue: we install salt into a chroot (that's going to be a pxe image) which pulls in python-yaml, run salt --local to get lots of small tweaks which also installs python-yaml and then remove salt again which removes python-yaml.
12:16 Reiner030 http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.pkgng.html#salt.modules.pkgng.autoremove
12:16 cetex but if i would do "apt-get install python-yaml" instead of salt's "pkg.installed" it would mark the package as manually installed and it wouldn't be autoremoved.
12:16 cetex :)
12:17 diehard_ joined #salt
12:17 Reiner030 ah, that was the manager for FreeBSD ^^
12:18 Reiner030 ah, that must then an expert answer ^^... I just beginner for salt
12:18 cetex I guess salt is checking dpkg to see if the package is installed at all, and not checking if it's installed as a dependency or manual, and therefore not doing anything (marking it as manually installed) when i tell it to set it to manually installed? :)
12:18 cetex yeah. most likely
12:18 cetex it's some voodoo somewhere.
12:19 Reiner030 http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.aptpkg.html#salt.modules.aptpkg.install
12:19 Reiner030 > Dependencies are automatically resolved and marked as auto-installed.
12:20 Reiner030 seems there is no option to override the auto-install tagging.
12:20 cetex yeah. as they should be, but if i then set one package that is installed as a dependency to "pkg.installed" it should mark it as manually installed and not autoremove it.. :>
12:21 viq cetex: how are you installing salt? Maybe you could install python-yaml beforehand?
12:22 viq cetex: also, why do you remove salt? ;)
12:22 cetex i guess i could, but then i need to do stuff outside of salt which is unclean.
12:22 cetex it's a pxe image, we reboot when we upgrade image
12:22 cetex it should be static.
12:22 cetex :)
12:22 viq mhmm
12:23 viq though you'd get remote execution if you kept it ;)
12:23 cetex and we try to strip it down.
12:23 * viq nods
12:23 cetex and i would need master.
12:23 cetex :>
12:23 cetex which i try to avoid.
12:23 viq true, that
12:23 viq I still don't get why people do that :P
12:23 Reiner030 but there is also a problem i have ...when i install bsd-mailx and postfix... he first installs bsd-mails + exim ... and later postfix by removing exim again ^^
12:24 viq Reiner030: what if you reversed the order?
12:24 Reiner030 It seems that salt is sorting them alphabetically and installing then the packages
12:24 viq ah
12:25 cetex we're aiming for a single-image datacenter, no machine will be more important than anyone else by default, some may launch mesos-master stuff, others may launch cassandra, all will run hdfs-datanodes, if we change anything we'll build a new image and reboot machines to get them up on new software.
12:25 Reiner030 and I am a little wondering because salt seems to install packages step by step and not in one command... (2014.1.13)
12:27 cetex and everything should be stateless except for the stuff stored in cassandra or hdfs. :)
12:28 viq I guess that is a reason ;)
12:30 viq cetex: apt-mark(8)
12:30 cetex :)
12:30 cetex yeah. i'll do something half-ugly like that :)
12:30 cetex lunch! :)
12:30 Reiner030 when I was searching for differnt things I found an interesting blog http://humankeyboard.com/saltstack/2013/execution-order-and-overstate.html which shows howto order server setup sequence ;)
12:34 ITChap joined #salt
12:37 bhosmer joined #salt
12:39 bhosmer_ joined #salt
12:40 wm-bot4 joined #salt
12:40 X86BSD joined #salt
12:40 felskrone joined #salt
12:41 felskrone joined #salt
12:47 babilen I am using a state such as: https://www.refheap.com/92684 and am a bit surprised that the apache2 service is not restarted *after* either /etc/apache2/sites-available/001-foo-bar have been created or a2ensite 001-foo-bar has been run. Those states execute without problems, but are the last states that execute (i.e. no service restart afterwards)
12:47 babilen Any idea what I am doing wrong there?
12:48 claytonk joined #salt
12:50 viq babilen: would it make a difference if you changed watch_in from apache2 to apache2-service ?
12:51 babilen I can try that, but it shouldn't.
12:51 babilen (you can refer to state ID and name)
12:51 maoroo joined #salt
12:51 viq Yeah, I know
12:51 ipmb joined #salt
12:51 Reiner030 I have the reload problem in 2014.1.13 ... there is a python error in this service
12:52 viq I know this is a simplification, but you don't have service require packages
12:52 babilen viq: I'm trying it anyway. It would probably make sense to implement a service reload as in the apache formula anyway, but I was just a bit surprised by that.
12:52 viq yeah, I would expect it to work
12:52 shookees joined #salt
12:52 viq babilen: also, shouldn't the a2ensite be a cmd.wait ?
12:53 Reiner030 I used a part of https://github.com/saltstack-formulas/apache-formula/tree/master/apache ... perhaps it helps you to find your problem
12:53 babilen viq: I pondered that question and thought that cmd.wait is for states that watch other states (i.e. they shouldn't execute every single time) -- It is definitely worth a try too, but then the documentation is a bit out of line with reality ;)
12:54 Reiner030 yes, there is in init.sls a module.wait ...
12:54 babilen indeed
12:54 viq babilen: yeah, and watch the 001-foo-bar state. Because otherwise a2ensite would run every time you run the state, and thus every time your apache would be restarted
12:55 viq (if it worked as it's supposed to that is :P)
12:55 babilen viq: You mean use cmd.wait for a2ensite with a watch on 001-foo-bar and a watch_in: - service: apache2 in the 001-foo-bar one?
12:55 Reiner030 and... normally an apache.reload is sufficient ... or is there a good reason to use restart instead (debian system) ?
12:56 ndrei__ joined #salt
12:56 babilen Just the fact that I haven't implemented apache-reload yet :)
12:56 viq babilen: yeah, though watch_in in both, I think
12:56 Reiner030 because reload creates a "graceful" restart which didn't interrupt the service...
12:56 babilen I know
12:56 Reiner030 ok... question was because the apache-formula uses also only restart
12:58 ramishra joined #salt
12:58 babilen viq: Wait, why would I use cmd.wait for the a2ensite and then "watch_in: - file: /etc/apache...." ?
12:59 Reiner030 i think it helps to not reload/restart apache 10 times when you changes multiple configs/services for it?
12:59 viq babilen: no, watch_in service apache
13:00 viq and watch the file
13:01 babilen ack
13:07 ladistwo joined #salt
13:10 felskrone1 joined #salt
13:12 ladistwo hello, i have often problems, when i running `salt 'minion' state.sls  name_state`, salt return this:
13:12 ladistwo minion:
13:12 ladistwo Data failed to compile:
13:12 ladistwo -----------
13:12 ladistwo The function "state.sls" is running as PID 20805 and was started at 2014, Nov 04 14:07:06.255691 with jid 20141104140706255691
13:12 ladistwo State is was running in minion, but no output on the server
13:13 intellix joined #salt
13:13 claytonk joined #salt
13:14 Reiner030 yes I have it also when fresh highstate is run... setting your timout higher should solve the output problem if you need it
13:14 Reiner030 (i hope there is a timeout configurable... not searched for it yet)
13:16 ramishra joined #salt
13:16 _ikke_ Is it me, or is the setup.py in develop corrupt? around line 592 there is a random return statement not belonging there, and there are closing brackets missing for the large datastructure
13:16 babilen ladistwo: You can also run it as "salt -v 'minion' state.highstate" and then check the JID with "salt-run jobs.lookup_jid $JID" or check "salt-run jobs.active" to see running jobs
13:16 ndrei joined #salt
13:16 lothiraldan joined #salt
13:17 babilen You can also raise the timeout with "-t 360" (or in the master config)
13:17 yomilk joined #salt
13:18 ladistwo So I try to set a larger timeout
13:19 ramishra joined #salt
13:21 ladistwo babilen: i set timeout on 360, problem is still
13:23 babilen ladistwo: There is no problem. Run it with "-v" and then wait for the job to complete.
13:26 ladistwo babilen: and why print this message? I use parameter "-v"
13:26 tafa2 joined #salt
13:26 Reiner030 ladistwo... perhaps you need even higher timeout ? for me it works (just tested ^^)
13:27 babilen ladistwo: Which message?
13:27 shookees joined #salt
13:28 Reiner030 he means the "-v" ?
13:28 ladistwo babilen: Data fialed to compile: etc.
13:28 Reiner030 mmh, but "Data failed to compile" is a "good" explanation why you don't get any output ;)
13:29 babilen ladistwo: That is an error. Please paste it to http://refheap.com along with the state it references
13:29 Reiner030 in this case I had also no output on my master but when run direct salt-call on minion.
13:29 the_drow joined #salt
13:30 ipmb joined #salt
13:32 mgw joined #salt
13:33 babilen ladistwo: You will also have to give us the resulting URL :)
13:36 shookees joined #salt
13:36 shookees joined #salt
13:37 ladistwo babilen: https://www.refheap.com/92688
13:37 Morbus g'day. i've got a "apache2-reload" module defined in apache2/init.sls.
13:37 Morbus i'd like to USE that inside php/init.sls
13:37 Morbus but i'm doing "module.run name: apache2-reload", and the salt logs are saying it can't be found.
13:38 babilen ladistwo: You did not paste the error -- could you add it at the bottom?
13:38 babilen (use "fork")
13:38 Morbus (is fork to me?)
13:38 babilen no
13:39 babilen Morbus: Did you include "- apache2" in the SLS file in which you reference it?
13:39 Morbus i did.
13:39 Morbus lemme pastebin the whole thing.
13:39 babilen (please don't use pastebin.com)
13:40 Morbus http://privatepaste.com/3d065a2b2c
13:40 ladistwo babilen: I add error
13:41 babilen ladistwo: What's the new URL?
13:43 ladistwo babilen: https://www.refheap.com/92689
13:48 ale_ joined #salt
13:48 thayne joined #salt
13:49 lcavassa joined #salt
13:49 Morbus ladistwo: i switched to watch_in, and now i'm getting a Python error: http://privatepaste.com/70a007e131
13:50 babilen Morbus: I just ran into the same error -- looks like a bug
13:50 ndrei joined #salt
13:51 Morbus hrm...
13:51 Morbus babilen: this was based off https://github.com/saltstack-formulas/apache-formula/blob/master/apache/mod_proxy.sls
13:51 babilen I know, it is a bug in salt
13:51 ipmb joined #salt
13:51 cpowell joined #salt
13:53 Morbus babilen: alright. is there a differnet way of doing this? in the past, i was just doing a cmd.run "service apache2 restart", which is dirty and non-portable.
13:53 yomilk joined #salt
13:54 pressureman joined #salt
13:54 miqui joined #salt
13:54 aqua^mac joined #salt
13:55 __gotcha joined #salt
13:56 __gotcha Has anyone succeded to find pygit2 package on EPEL 6 ?
13:59 eseyman joined #salt
14:00 ladistwo left #salt
14:00 TaiSHi joined #salt
14:04 mike_perdide left #salt
14:07 gngsk joined #salt
14:08 ladistwo joined #salt
14:09 mdasilva joined #salt
14:10 peters-tx joined #salt
14:15 racooper joined #salt
14:17 fredvd joined #salt
14:18 TyrfingMjolnir joined #salt
14:19 nitti joined #salt
14:19 mdasilva morning
14:20 nitti joined #salt
14:20 * Morbus is having a tough time getting a service to restart after a cmd.run...
14:21 * Morbus has tried listen, watch, module.run, etc.
14:21 gmcwhistler joined #salt
14:24 jaimed joined #salt
14:25 Reiner030 cmd.wait was discussed before for service restart ;)
14:25 flyboy82 there's also listen_in you may want to give it a lookup
14:26 wr3nch joined #salt
14:26 Morbus yeah, i've tried all the _in variations too.
14:26 Morbus i must be doing something massively wrong.
14:28 flyboy82 I assume you are including the sls in which the service state resides
14:28 bhosmer_ joined #salt
14:28 Morbus i am including the sls, but not sure what you mean by "service state resides".
14:28 flyboy82 if you want to restart apache you have to include the sls containing the apache service.running state
14:28 nebuchadnezzar joined #salt
14:29 Morbus here is one such example that i tried: http://privatepaste.com/70a007e131
14:31 rostam joined #salt
14:31 flyboy82 aah, the error it throws is certainly not due to any of your syntax
14:32 Morbus sure. but i've tried other approaches and DON'T get that particulr error
14:32 Morbus but the reload/restart never seems to happen.
14:32 Morbus driving me a little bonkers. getting close to just adding a "cmd.run: service. apache2 restart" somewhere.
14:32 Ahlee When the master goes under distress (say, applying highstate to ~500 minions), [ERROR   ] An un-handled exception was caught by salt's global exception handler: NameError: global name 'log' is not defined
14:36 glyf joined #salt
14:36 karimb joined #salt
14:37 karimb joined #salt
14:37 __gotcha lemme try a second time : Has anyone succeded to find pygit2 package on EPEL 6 ?
14:38 bhosmer joined #salt
14:38 perfectsine joined #salt
14:38 MTecknology Any of you have super magical salty states for mysql db/user/pass/grant combos? I'd like to have my pillar file look like this... http://dpaste.com/3SE0ZMR
14:40 iggy __gotcha: it's a relatively new project, it's probably not widely packaged
14:40 dude051 joined #salt
14:40 iggy f.ex. debian sid doesn't even have it packaged
14:42 iggy MTecknology: the mysql-formula doesn't work for you? The pillar isn't exactly the same, but it's not far off
14:44 MTecknology I didn't look at the formula. Will do
14:44 xintron joined #salt
14:46 iggy the difference is it ties the db grants to the user instead of it being separate
14:46 thematthewgreen left #salt
14:47 ladistwo left #salt
14:48 ramishra joined #salt
14:49 ramishra_ joined #salt
14:50 xintron I have problems with supervisord state. I want to have watchers on a package and a file that (if updated) should trigger supervisor to restart the process, but only if the file/pkg has been modified. Currently it's always restarted so the supervisord.mod_watch/running seems to always trigger a restart+update
14:51 saravanans joined #salt
14:51 iggy you sure the file/pkg isn't constantly updating?
14:52 ramishra_ joined #salt
14:52 iggy even if doesn't seem to be, sometimes what humans see as a change and what computers do is wildly different
14:52 xintron Not according to the debug output, no.
14:53 xintron But do you have any way for me to verify this (to SEE if salt watched a change or not) I would be happy to learn
14:53 viq Huh, the 2014.7 release notes don't mention anything about merging pillars...
14:54 iggy I think there's a setting for it now
14:55 racooper joined #salt
14:55 iggy or I'm totally wrong... either way
14:55 xintron iggy, "Package xxx is already up-to-date." + "File /etc/supervisor/conf.d/xxx.conf is in the correct state". Both the file and the pkg watchers reported correct state, yet the mod_watch block triggered
14:56 __gotcha1 joined #salt
14:56 xintron Correct me if I'm wrong but http://docs.saltstack.com/en/latest/ref/states/all/salt.states.supervisord.html#salt.states.supervisord.mod_watch should only trigger if the watchers triggers?
14:56 iggy in general yes
14:57 iggy try pasting your states, etc. somewhere
14:57 xintron sure
14:59 xintron iggy, https://gist.github.com/xintron/8bb2c26a5d297cb6732c
15:04 bhosmer joined #salt
15:05 rspectre joined #salt
15:06 iggy I've never used mod_watch in that way... I've always just used it as watch or watch_in
15:06 xintron iggy, Do you have an example?
15:07 iggy I would (personally) put watch_in's on the pkg-jebediah and the config.json file
15:08 iggy but that just usually makes more sense to me
15:09 hobakill joined #salt
15:11 ekristen joined #salt
15:11 intellix joined #salt
15:13 ekristen joined #salt
15:14 ekristen so I see that 2014.7 was tagged 20 hours ago, that mean its officially out?
15:15 xintron iggy, No difference. Seems like mod_watch always triggers and is not dependent on the watchers at all.
15:15 timoguin ekristen: yes. just waiting for packagers before the announcement
15:15 ekristen nice
15:15 iggy xintron: what do you have now? something like https://gist.github.com/iggy/bc19c411fd446a730171
15:16 iggy 2 weeks
15:16 xintron iggy, And the restart task as well. I want the restart to trigger if there is a new package installed.
15:17 viq timoguin: and I'm already asking questions about it ;)
15:17 iggy xintron: try my version
15:18 xintron iggy, Running it as we speak :)
15:18 iggy there's a watch_in on the package, it should do what you want
15:18 xintron iggy, You found that by reading the source or?
15:18 timoguin viq: you should be ashamed!
15:18 iggy xintron: found what?
15:19 xintron sorry, nvm
15:20 kaptk2 joined #salt
15:20 srage joined #salt
15:21 xintron iggy, oh, that worked!
15:21 * iggy flexes
15:21 xintron Running one last test though (with update: True)
15:21 viq timoguin: I probably should... naaah, who am I kidding :P
15:21 iggy and you saved 8 lines of cruft
15:22 rojem joined #salt
15:22 xintron iggy, ok, update: true will each and every time update the config files (even if there are any changes or not).
15:22 xintron update: true; also seems to break the watchers
15:23 viq Anyone knows for sure if pillar merging made it into 2014.7? https://groups.google.com/forum/#!topic/salt-users/Qe5z3hZ8IB0
15:23 iggy xintron: so don't use it?
15:23 iggy what are trying to do?
15:23 xintron iggy, I have the supervisor config file handled by salt and if that one updates I need supervisor to run an update
15:24 xintron So: 1. If pkg updates; restart process. 2. If supervisor config updates; run supervisorctl update.
15:24 xintron Maybe I need two different blocks for this then
15:24 iggy right, but why the use of update=true? and where are you even putting that?
15:24 xintron iggy, under supervisord.running: - update: True
15:25 iggy so if you leave that out and stuff changes, what happens?
15:26 xintron iggy, the process is restarted. But not updated
15:26 dalexander joined #salt
15:26 ndrei joined #salt
15:26 iggy I'm not a supervisord expert, so you might want to open a ticket or wait around for someone who knows it better
15:29 housl joined #salt
15:29 timoguin xintron: I use a managed file for the supervisor config and watch that with supervisord.running state.
15:29 timoguin I'm pretty sure it's been reloading the configs appropriately.
15:29 nicksloan joined #salt
15:30 iggy that's what he's got now (only with a watch_in on the file instead of a watch on the .running)
15:30 nicksloan I saw that there is a 2014.7 release on Github. Is 2014.7 done, or is did I misunderstand?
15:30 iggy 2 weeks
15:30 MTecknology iggy: that was fun! http://dpaste.com/39ST0VF
15:31 iggy nicksloan: I kid, I kid... think they are just waiting on packagers
15:31 timoguin xintron: iggy: I'm actually using a watch_in AND a watch. for whatever reason
15:31 nicksloan iggy: oh! I hadn't thought of that. Makes sense
15:31 iggy RAIW
15:33 mpanetta joined #salt
15:34 jngd joined #salt
15:34 xintron timoguin, I just tested it and it didn't do an update. It did an restart and I guess I'll have to live with that for now.
15:36 timoguin xintron: well... there is are some more execution modules that may be able to help: supervisord.reread and supervisord.update
15:36 timoguin I think I used those initially but switched to the state
15:37 mpanetta_ joined #salt
15:37 xintron timoguin, Can you use modules in a state file?
15:38 timoguin xintron: you can with module.run/module.wait
15:38 xintron ah, a module state
15:38 glyf joined #salt
15:38 timoguin ideally the state should have all the functionality you need, but if it doesn't you still gotta get stuff done. ;-)
15:38 xintron Indeed :)
15:40 mpanetta joined #salt
15:40 conan_the_destro joined #salt
15:42 mpanetta_ joined #salt
15:45 ajolo joined #salt
15:47 rspectre Question on salt-cloud: what sync_after_install isn't super clear within the nomenclature of salt configuration.  Does this mean the host performs a highstate after installation?
15:47 mpanetta joined #salt
15:48 SheetiS joined #salt
15:48 pdayton joined #salt
15:51 mdasilva joined #salt
15:51 elfixit1 joined #salt
15:53 mpanetta joined #salt
15:53 mpanetta joined #salt
15:55 johtso joined #salt
15:57 johtso How do I stop git.latest for pulling after doing a force reset and a checkout?
15:57 johtso *from
15:58 glyf joined #salt
15:58 babilen johtso: git.latest will always fetch HEAD
15:58 iggy rspectre: it does a sync_all (iirc)
15:58 xintron timoguin, supervisord.reread triggered a change event, supervisord.update did not. (when using module.wait-state).
15:59 rspectre iggy: how is that distinct from a highstate?
15:59 thedodd joined #salt
15:59 iggy rspectre: it doesn't run any states at all
15:59 johtso babilen: it ends up doing this: http://hastebin.com/qufajolune.vbs
15:59 hasues joined #salt
15:59 rspectre iggy: Guess I'm a little confused on that distinction.  sync just pulls does the states and pillars for the host?
15:59 rspectre iggy: highstate runs them?
16:00 johtso babilen: I just want to force reset to origin/ref and make sure it's checked out into the working tree
16:00 hasues left #salt
16:00 babilen johtso: You might want to remove the repo beforehand. You also wouldn't be able to do that manually
16:00 babilen I mean "force reset" isn't exactly something one should do in git.
16:01 vuba joined #salt
16:01 vuba howdy
16:01 babilen bbl ($HOME)
16:01 vuba has anyone played with syndic?
16:01 iggy rspectre: correct (sync can also sync custom modules, grains, etc. if you have any of those)
16:01 babilen vuba: Lots of people
16:01 vuba I have a very modular infrastructure
16:01 vuba and have 3 submasters
16:01 rspectre iggy: gotcha.  Does sync pull down all the states and grains?  Or just the ones specified for that host?
16:02 Rojematic joined #salt
16:02 iggy rspectre: we use it to sync our custom grains before we run a mine update
16:02 iggy minions always have all states/grains
16:02 vuba but what it seems to me is that salt-syndic only proxies the cmd from the master.master
16:02 vuba how can I send cmd directly to the submaster
16:02 vuba ?
16:02 iggy pillars are the only thing that are limited to the minion
16:02 rspectre iggy: Gotcha.  What would be the canonical method of initiating a highstate post-install then?
16:03 johtso babilen: hmm, well if I'm using a git repository for code deployment, I don't want to be doing any merging or anything if branches diverge, I just want make sure the checked out code matches the remote code at that ref
16:03 iggy rspectre: either with a reactor or startup_states in the minion conifg
16:04 conan_the_destro joined #salt
16:06 rspectre iggy: I see.  Haven't used reactors yet.  Will startup_states work with the order of operation using salt-cloud?  My sense is salt-cloud spins up the host, runs bootstrap, adds the key... not sure if startup_states will fire then.
16:06 mdasilva question for those using gitfs, how do you manage multiple environments? do you cherry-pick commits? rebase? keep them separate as long running branches?   With file_roots I kept base with common fomulaes each envronment hads states only specific to that environment.  Adapting this to gitfs i feel i'm littering git with a ton of commits and constantly switching between master and branches
16:07 iggy rspectre: it will run when the minion process starts
16:07 iggy afaik, _every_ time it starts up
16:08 rspectre iggy: Alrighty.  Will give it a go then - appreciate the pointers.
16:09 karimb joined #salt
16:09 conan_the_destro joined #salt
16:09 bezeee joined #salt
16:10 vuba anyone?
16:10 viq vuba: run minion on syndics as well
16:11 Setsuna666_ joined #salt
16:13 dalexander joined #salt
16:14 KennethWilke joined #salt
16:16 wnkz Hi, I want to start coding a custom returner ; ofc I read http://docs.saltstack.com/en/latest/ref/returners/ _but_ I still don't know *where* I should place my file so I can use it with salt
16:17 viq wnkz: http://docs.saltstack.com/en/latest/ref/file_server/dynamic-modules.html
16:17 ipmb joined #salt
16:18 wnkz viq: beleive or not I had "Place custom returners in a _returners directory within the file_roots specified by the master config file." in front of me for 10 good minutes .. -_-'
16:18 CeBe1 joined #salt
16:18 viq :P
16:18 wnkz viq: thanks anyway :)
16:18 viq What's the saying, "it's darkest under the lantern" ? ;)
16:20 iggy I think I got approval to go to SaltConf15
16:21 StDiluted joined #salt
16:21 wnkz Please make a SaltConf in Europe :3
16:21 klj joined #salt
16:22 klj joined #salt
16:22 p2 joined #salt
16:23 iggy if it keeps gaining popularity, I wouldn't be surprised
16:23 klj joined #salt
16:26 rawzone joined #salt
16:33 karimb joined #salt
16:34 berserk joined #salt
16:37 dave_den joined #salt
16:46 gwb joined #salt
16:47 Ozack1 joined #salt
16:49 tligda joined #salt
16:50 ghanima joined #salt
16:50 aqua^mac joined #salt
16:52 bhosmer_ joined #salt
16:54 tristianc joined #salt
16:55 thayne joined #salt
16:56 tligda1 joined #salt
16:57 KyleG joined #salt
16:57 KyleG joined #salt
16:58 LGSilva joined #salt
16:58 meylor joined #salt
16:59 felskrone joined #salt
16:59 alex-mesos joined #salt
17:00 LGSilva hello, I'm trying to recurse directory permissions using file.directory (http://pastebin.com/JnrCqWnD) but I keep getting this error: Types for "recurse" limited to "user", "group" and "mode"
17:00 LGSilva salt version is salt-minion 2014.1.13 (Hydrogen)
17:00 LGSilva same for master
17:01 alex-mesos joined #salt
17:02 iggy that's a very odd use
17:02 TheThing joined #salt
17:03 wnkz do you know where I can find all existing fields in a 'ret' object inside a returner ?
17:03 iggy and also... I'm starting to see why everybody hates on pastebin.com so much
17:03 iggy f'ing flash ads
17:03 LGSilva adblocker :)
17:04 LGSilva so the file.directory is not the best way to change those permissions?
17:06 iggy that error message is pretty clear though
17:07 iggy take out ignore_files from the recurse
17:07 lopezio hi fellow salters.. anyone here that can tell me how to correctly call a built in state module within a custom state module?
17:07 LGSilva so the documentation on the website is wrong?
17:07 troyready joined #salt
17:08 iggy it also looks like your indentation is messed up
17:08 LGSilva that's pastebin, I checked my file once I saw it too
17:08 babilen joined #salt
17:09 LGSilva this is what I'm getting the ignore_files from: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.directory
17:09 viq LGSilva: "New in version Lithium."
17:09 LGSilva facepalm... I'm blind.
17:09 LGSilva sorry about that
17:10 forze joined #salt
17:12 iggy I really hope they get the versioned docs situation sorted out soon
17:12 iggy I have a feeling it's going to be even worse with this release than it was the last one
17:13 druonysus joined #salt
17:14 dalibro joined #salt
17:15 babilen iggy: You should be able to, at least, browse docs in a specific branch
17:18 iggy you mean the unrendered versions in git?
17:18 iggy I'd rather castrate myself with rocks
17:18 smcquay joined #salt
17:18 thedodd joined #salt
17:19 viq that could be arranged.... ;)
17:20 rojem joined #salt
17:21 bhosmer_ joined #salt
17:21 avienu joined #salt
17:22 aparsons joined #salt
17:23 mike_perdide joined #salt
17:23 Gareth morning morning
17:24 babilen iggy: No, you should be able to browse the rendered version of all available branches individually.
17:24 eriko joined #salt
17:24 iggy where?
17:24 babilen *should*
17:24 babilen (as in: that would be nice)
17:24 iggy oh, yeah
17:24 mike_perdide left #salt
17:24 iggy that's why I said I hope they get it sorted
17:25 iggy last I heard, it was all moving from readthedocs to docs.saltstack.com
17:25 iggy but I haven't heard anything since then
17:27 babilen I'm sure this is a "patches welcome" situation
17:27 iggy uhh... how am I going to patch their infrastructure
17:27 iggy it's not like there's some link broken or some weirdly worded text
17:28 babilen But maybe we (as in: the community) will have some time to focus on "unimportant" things like this for a while now after the Helium release. I would definitely like to see a focus on getting *all* formulas working, documentation in order and bugs fixed.
17:28 babilen true
17:29 iggy the docs for 2014.1 haven't been updated since 2014.1.4 on readthedocs
17:30 iggy and there is no way (at least that I can see) to switch versions on docs.saltstack.com
17:32 babilen there isn't
17:33 SheetiS babilen: I definitely agree on formulas.  It seems so many are ubuntu specific.  I've been trying to add PR when possible for os_family RedHat since I work with that mainly, but I really feel bad for the FreeBSD guys on it.
17:34 SheetiS Not so bad for the Windows ones since usually the entire formula is not applicable, but would like to see more windows-based formulas.
17:35 babilen SheetiS: It is also quite unpleasant to find out that they have been developed against Ubuntu and break on Debian.
17:35 babilen (or use random PPAs if there are perfectly fine packages available in the repos already)
17:36 glyf joined #salt
17:36 SheetiS babilen: I hadn't thought of that as well
17:36 babilen Another huge thing is proper support for multiple versions of configuration files
17:36 iggy I'm not going to insist someone making a formula make it work on distros they aren't familiar with (that's worse than no support at all)
17:37 iggy case in point... the graphite formula
17:37 babilen Surely not, but the distributions that *are* supported should be supported right.
17:37 SheetiS iggy: I agree with that.
17:37 viq joined #salt
17:38 babilen What's the problem with the graphite formula?
17:38 iggy someone wrote it for redhat, and put not only incomplete, but mostly incorrect debian/ubuntu support in it
17:38 babilen (I see no issue)
17:38 babilen Could you report an issue?
17:38 SheetiS I'm kinda in a weird spot since my most recent job has been an Amazon Linux shop that had no config mgmt (I am working through to salt all the things in my first 3 months so far)
17:38 iggy it would have taken me less time to add debian/ubuntu support than to fix what was there
17:38 SheetiS So even os_family redhat isn't always right for me
17:39 iggy it is using a redhat init script
17:39 iggy on debian/ubuntu
17:39 Eugene What exactly is Amazon Linux derived from, anyway? Mix of RHEL & Fedora?
17:39 babilen That is just ... wrong
17:40 SheetiS iggy: I had the opposite on the salt nginx formula.  The one provided in the formula is the ubuntu one.
17:40 beneggett joined #salt
17:40 SheetiS at least you can install the lsb packages to make it mostly work.
17:40 SheetiS but... yuck.
17:40 iggy ugh... the nginx formula
17:41 SheetiS Eugene: kinda.  It looks like it started as CentOS6, and EPEL6 still works with it
17:41 iggy I looked at it... wrote my own in 15 minutes
17:41 SheetiS But it uses a modern kernel and newer packages for a lot of things
17:41 Eugene So it's more like an Ubuntuization
17:41 cberndt joined #salt
17:41 SheetiS kinda
17:42 SheetiS some things are way out of date
17:42 iggy oel is similar (rhel6 + new kernel and other bits)
17:42 SheetiS like it still uses libgmp 4. whatever
17:42 Eugene But nobody cares about those :-p
17:42 SheetiS You know this thing: /usr/lib64/python2.6/site-packages/Crypto/Util/number.py:57: PowmInsecureWarning: Not using mpz_powm_sec.  You should rebuild using libgmp >= 5 to avoid timing attack vulnerability.
17:42 SheetiS _warn("Not using mpz_powm_sec.  You should rebuild using libgmp >= 5 to avoid timing attack vulnerability.", PowmInsecureWarning)
17:43 thayne joined #salt
17:44 SheetiS I'm still not sure the 'right' way to fix that on Amazon linux.  I've chosen _not_ to suppress the message programmatically, which is what I saw some.
17:44 SheetiS s/some/some do/
17:44 steverweber joined #salt
17:44 carmony is 2014.7.0 live?
17:46 ShibMcNe iggy: the Apache formula is not perfect either, but I think a lot of people are losing by reimplementing their formulas instead of fixing the existing ones.
17:46 iggy for the most part I fix things
17:46 seydu joined #salt
17:46 iggy but the nginx formula is beyond that imo
17:47 SheetiS I got it working os_family RedHat recently.  I'm still not happy with it
17:47 babilen The apache formula is horrible
17:47 SheetiS but it at least works
17:48 iggy I have sent PRs for collectd-formula, graphite-formula, salt-formula, salt-contrib, and rsyslog-formula at least
17:48 wnkz does someone know how can I debug a returner on a minion ? I did run "saltutil.sync_returners" and got True as return but then nothing happens when I specify my returner. Note: It does work when calling salt on my master (where the returner is)
17:48 ShibMcNe I mostly agree with you, I hesitated between patching it and creating cutom one…
17:48 SheetiS I don't use the rsyslog formula.  I just rolled up my own on that one.  I was in a hurry that day.
17:48 ekristen_ joined #salt
17:48 babilen redis is also a bit weird in that it doesn't use the configuration files that come with the packages, but the default ones from upstream. That makes for a weird "packaged-but-unpackaged" mix that caused me to abandon it pretty damn fast
17:48 babilen iggy: Have they been merged?
17:48 mapu joined #salt
17:48 iggy some
17:49 mgw joined #salt
17:49 iggy some I need to separate and/or squash the commits
17:49 ShibMcNe a lot of formulas are quite odd, it seems that a lot of people prefer to use custom ones.
17:50 iggy oddly, my salt-formulas PRs sit for a while usually... my salt-contrib ones have been merged within minutes usually
17:50 SheetiS https://github.com/saltstack-formulas/nginx-formula/pull/56/files look at how nasty that is just to make os_family RedHat work for nginx.  There's still a problem with the way the sysvinit script works for nginx that causes the state to kinda freeze, but I wanted to use upstart anyhow, so all good-ish.
17:50 babilen iggy: If you want to we can take an hour and work through them
17:50 babilen Ah, https://github.com/saltstack-formulas/collectd-formula/pull/4 ?
17:51 wendall911 joined #salt
17:51 iggy that's one
17:51 Ahlee Anybody seen this issue before? https://gist.github.com/jalons/c05016aafc6dd8d8569f
17:51 babilen iggy: Please rebase/fixup those commits into a single one. Nobody wants to merge "hopefully fix df config template" -- just rebase and force push into your branch
17:51 iggy but I noticed after I started doing the PR that it needed to be split up
17:52 SheetiS Ahlee: it looks like you are trying to do a |startswith on a list instead of a string.
17:52 SheetiS I assume inc_sls is a list?
17:52 unpaidbi1l anyone have an example of setting default values with grain filters in a mako template?
17:52 Ahlee SheetiS: no idea.
17:53 SheetiS so you may need to iterate over it (for loop or whatever), and then do your test on each item in the list.
17:53 unpaidbi1l i'm having some trouble replicating the jinja way of doing it in mako
17:53 babilen iggy: Well, ping me. I'd happily merge what I understand and deem correct
17:53 flyboy82 joined #salt
17:53 Ahlee SheetiS: I am include: \n - my__state\n - .foo
17:53 jalbretsen joined #salt
17:53 Ahlee but a recursive grep of my file_root just returned no matches for |startswith
17:54 SheetiS oh ok so that looks like it's not in a state then but maybe a bug.  What version are you running?
17:54 Ahlee 0.17.5
17:54 iggy babilen: the problem is (at least for that one) that I want feedback on some things... and a lot of the formulas seem abandoned
17:55 SheetiS Did 0.17.5 support multiple includes like that?  It appears the test is not accounting for the list.
17:55 iggy I can throw features at formulas all day long, but if I'm the only one using them, they need to be generalized
17:55 CeBe1 joined #salt
17:55 Ahlee I thought it did.  I know it includes . support, betting the issue is it ddoesn't support the mix
17:55 Ahlee will try breaking them out
17:55 Ahlee thanks SheetiS
17:56 babilen iggy: I totally agree, but the only thing I can offer right now is to discuss and merge changes that you submitted. If *that* is the bottleneck in getting them in shape.
17:57 murrdoc joined #salt
17:57 vlcn what is the expected value of the virtual grain on a minion that is a vmware vm which is a openvz (or smilar) node?
17:57 forrest joined #salt
17:58 SheetiS Ahlee: that _should_ work in 0.17.5 according to this http://salt.readthedocs.org/en/v0.17.5/ref/states/include.html, but I don't know about the mix.  You cannot have 2 separate 'include' lines in a single file though because of yaml limitation if you try and break them apart.  I do know that.
17:58 vlcn It seems like the 'VMware' value simply gets overidden with 'openvzhn'
17:58 Ahlee SheetiS: hrm.  Good point.
17:58 Ahlee SheetiS: include will name collide
17:58 SheetiS Exactly.
18:02 SheetiS iggy: you were planning to go to saltconf15 right?  I'm considering it myself, but not 100% sure yet.
18:02 iggy I am
18:03 aqua^mac joined #salt
18:03 iggy my boss likes the idea, just need to convince his boss and HR
18:03 iggy (I'm trying to get it to count as work days instead of using PTO)
18:03 murrdoc shiet
18:03 iggy and they are trying to find training money, but I'm okay with paying for it myself
18:03 murrdoc i be there too
18:04 rypeck joined #salt
18:04 murrdoc you know
18:04 murrdoc if you guys want to say wassup
18:04 SheetiS Similar here.  I'm not worred about the time as my place of employment offers 'unlimited' vacation days, but they are also usually open to paying for these kinds of things.
18:05 SheetiS murrdoc: Would be kinda cool to put faces to everyone on IRC ;-).
18:05 iggy I could never work there... I'd be gone every Monday
18:05 murrdoc sweet
18:05 murrdoc there is a dope vodka brewery that was a sponsor last time
18:06 murrdoc i know in utah
18:06 murrdoc :)
18:06 SheetiS iggy: The only metric my boss cares about is whether work gets done or not.  If I am meeting expectations, it doesn't matter when/how I am working too much.
18:08 berto- joined #salt
18:09 notbmatt joined #salt
18:10 pdayton joined #salt
18:13 Ahlee if you include: something, can you later require: something that would have executed in that include?
18:14 SheetiS Ahlee: yes, I do that all the time.
18:14 SheetiS At least in 2014.1 and 2014.7 I do that.  I don't want to speak for 0.17.5 as I don't recall.
18:14 Ahlee no problem
18:14 KyleG joined #salt
18:14 KyleG joined #salt
18:15 Ahlee Not a lot of the core functionality has really changed.
18:16 ingwaem joined #salt
18:17 gwb Should a brew install of saltstack automatically create /etc/salt directory? My install did not.
18:18 forrest joined #salt
18:19 Ahlee SheetiS: It looks like the actual issue was unrelated, and was from requiring a cmd.run output that's a bit...flakey.
18:19 murrdoc gwb:  nope it doesnt
18:19 SheetiS Ahlee: weird.
18:20 SheetiS Locked into 0.17.5 or looking to push forward sometime soon?  I'm hating it enough that I'm not using 2014.7 yet ;-)
18:20 ndrei joined #salt
18:20 gwb murrdoc: thanks. What salt files should initially reside in /etc/salt or the equivalent salt home dir?
18:20 Ahlee holding strong to 0.17
18:20 murrdoc top.sls
18:21 murrdoc and the state files you care about
18:21 Ahlee The devil you know is better than the devil you don't
18:21 murrdoc thats why perl is still a thing ahale
18:21 murrdoc ahlee *
18:21 murrdoc sorry ahale
18:21 SheetiS Ahlee++ so true.
18:21 gwb thx
18:21 Ahlee plus at this point, the man hours required to convert all my env= to saltenv=
18:22 Ahlee i'll wait until whatever-replaced-halite comes out before I move
18:22 SheetiS Right now my salt install looks so Frankenstein with so much _modules and _states from 2014.7 that I should upgrade myself ;-)
18:22 murrdoc i d also mkdir pillars _modules _renderes _states
18:22 Ahlee as all i want is the stupid reactor
18:22 murrdoc put a salt/top.sls
18:22 Ahlee SheetiS: i hear ya there.
18:22 diegows joined #salt
18:22 murrdoc uh pillars/top.sls
18:22 murrdoc then start cloning stuff from salt-formulas repos
18:23 murrdoc and have fun gwb
18:23 gwb murrdoc: where do I find the initial master file template?
18:23 SheetiS http://docs.saltstack.com/en/latest/topics/tutorials/walkthrough_macosx.html has an ok macosx master with vagrant/virtualbox walkthrough for local stuff.  Assuming that is what is going on when i see 'brew install'.
18:23 KyleG damn applecare is WAY cheaper ($100 cheaper) from other vendors rather than from Apple themselves
18:23 murrdoc there you go
18:24 gwb perfect
18:26 n8n joined #salt
18:29 gngsk Anyone using the iptables formula? I can't seem to get a service whitelistd that isn't part of the original declarationi of firewall.service.
18:30 spookah joined #salt
18:31 pdayton joined #salt
18:35 rojem joined #salt
18:35 MugginsM joined #salt
18:36 jeffspeff can someone point out what i'm doing incorrectly? here's part of my state file and the error it generates. http://pastebin.com/acs4DVLr
18:40 Ryan_Lane joined #salt
18:42 superted666 joined #salt
18:43 jalaziz joined #salt
18:44 seydu I would like to manage some files without linking them to any particular package. How can I do that ?
18:44 jeffspeff seydu, will file.manage work?
18:44 murrdoc git state
18:45 murrdoc file.recurse
18:45 n8n joined #salt
18:46 seydu No. If I defined them in a file without any package they are not shown when I run 'highstate'. I had to to add them in the slsfile that defines 'nano'
18:46 MugginsO joined #salt
18:48 seydu murrdoc: can you show me some examples please (docreference would be ok too)
18:48 jaimed joined #salt
18:49 murrdoc http://docs.saltstack.com/en/latest/ref/states/all/salt.states.git.html
18:49 murrdoc http://docs.saltstack.com/en/latest/topics/tutorials/gitfs.html
18:49 iggy babilen: should be good to go
18:49 murrdoc two options
18:51 Reiner030 if you haven't defined e.g. package nano in top.sls then it's state wouldn't run with highstate but only with salt-call sls.state nano
18:51 saurabhs joined #salt
18:52 mdasilva joined #salt
18:52 Reiner030 other way ^^... state.sls ... (and it must not included in other sls files)
18:54 seydu Reiner030: I get that. But when I put the files management in a separate state file state.highstate wont see them
18:55 gngsk Here's some debug output for the issue i'm dealing with now: http://pastebin.com/bUa1e7t7  For whatever reason salt seems to think my sls file doesn't exist but you can clearly see that it does. Any idea what might be going on here? I'm following the directions for augmenting the firewall.service pillar at the bottom of https://github.com/saltstack-formulas/iptables-formula
18:55 murrdoc highstate can see whatever you list in top.sls
18:55 seydu I ended up adding them in nano.sls
18:55 SheetiS jeffspeff: I don't know if you had your question answered, but you are trying to use salt['file.file_missing'] which does not exist.  I'd try and do an if not salt['file.file_exists'] instead
18:55 SheetiS jeffspeff: http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.file.html#salt.modules.file.file_exists
18:55 Kyle joined #salt
18:56 seydu But I dont think I can add the files in top.sls !
18:56 pdayton joined #salt
18:56 jeffspeff SheetiS, thanks. i figured since file.file_exists and file.directory_exists actually exist that it was just the naming scheme for all file modules
18:56 Reiner030 but you can define a package "filesystem" which handle only this kind of files and include it in top.sls ?
18:57 rojem joined #salt
18:58 Reiner030 and i just used try out for my problem a file.managed: with "source: None" which also creates only empty file and do no more management on the file.   or you can also use a file.managed with option replace: False
18:59 rojem joined #salt
19:00 aparsons_ joined #salt
19:03 b0fh joined #salt
19:04 Phibs I have a salt master doing 150Mbps of traffic, and nobody is running anything.  Seems some clients are just spewing at it
19:04 Phibs any ideas?
19:05 jchen what kind of traffic
19:05 Phibs salt master.
19:06 jchen no, you'll want to use `tcpdump` to actually see what's up
19:06 Phibs I know
19:06 jchen ok
19:07 SheetiS Phibs: How many minions?  You could also look for events via this http://docs.saltstack.com/en/latest/topics/reactor/#knowing-what-event-is-being-fired.  I know it's part of the reactor docs, but that script can give some decent insight into what is going on.
19:07 babilen Any active jobs on the master?
19:07 Phibs babilen: nope
19:07 Phibs hmm
19:07 mpanetta SheetiS: That script is a very nice tool
19:08 mpanetta For all sorts of things
19:08 Phibs yum updated salt master and it restarted heh my bad
19:08 SheetiS I have it running for all sorts of things
19:08 Phibs I'll have to figure out next time it gets hosed
19:08 Phibs SheetiS: probably ag ood 800
19:08 mpanetta It helped me figure out that I was passing arguments to a module incorrectly :)
19:08 SheetiS :)
19:08 babilen Did that happen after you restarted the master? It might trigger authentication events that can look like a ddos
19:09 Phibs no, before.
19:09 Phibs something gets wedged and minions just spew
19:09 babilen (or did you you in some way trigger it by, say, pinging/communicating with minions after a long time of being idle)
19:09 Phibs its a few minions taking up all the bw
19:09 Phibs doing jack.
19:09 babilen I killed one of my masters today :(
19:10 mpanetta babilen: Killed?  In what way?
19:10 babilen I am using the salt-formula and, for some stupid reason, the minion failed to render. That caused the master to essentially nuke its configuration (pillar config included) and I had to resurrect it manually. (etckeeper to the rescue)
19:10 iggy babilen: I missed one :/
19:10 babilen Still a very chilling experience ... not sure how to guard against that
19:11 hybridpollo joined #salt
19:11 mpanetta babilen: OUCH!  I didn't even know that was possible..
19:11 iggy babilen: you mean bits of pillar data were rendered in the config file?
19:11 mpanetta So the files got deleted from master.d or something?
19:12 babilen iggy: No, I use the salt-formula to manage my master. The configuration is defined in the pillar and, well, the pillar wasn't there and the master therefore nuked the old configuration with the default one.
19:12 iggy oh...
19:12 babilen No, just overwritten with the "new" default one
19:12 mpanetta Oh
19:13 mpanetta babilen: You scared me for a minute :P
19:13 babilen Which also included the git pillar setup (and therefore an easy way to undo that after fixing the pillar)
19:13 iggy yeah, we use the salt-formula as well... every once in a while we see bits of serialized pillar data in the config files
19:13 babilen wut?
19:14 babilen It was complicated by the fact that I run my master as a different user and that, therefore, suddenly all the directories (/var/cache/salt, ...) had the wrong owner and so on.
19:14 iggy yeah... it's happened quite a few times since we started using the salt-formula (and afaik, we only see it with that formula)
19:14 babilen I mean it wasn't that hard to get around as we use etckeeper on *every* box so we have commits of every change, but still ...
19:14 babilen iggy: I have not seen that before and would be very interested in a way to reproduce that
19:14 iggy p.s. thanks for looking at those so fast
19:15 iggy yeah, I've been trying to find a way
19:15 iggy if I ever do find a way...
19:15 mpanetta babilen: What are the advantages to using salt-formula to manage your master?  I've not used it yet.  We just manually manage them here, not that we change master confs otfen...
19:16 iggy reproduceability?
19:16 babilen mpanetta: I like to add formulas that I start using and also define different master configurations for, say, production, vagrant boxes, ...
19:17 mpanetta Ah ok
19:17 iggy I can push out an entirely new copy of dev/qa/prod in ~20 minutes with 1 button
19:17 babilen exactly
19:17 mpanetta Ah!
19:18 babilen Which reminds me that I should write some states for the initial configuration of salt-masters and minions that I can use via salt-ssh to get new boxes "saltified"
19:18 mpanetta So you use it to manage your pillar and ext configs and so forth?
19:18 iggy we keep everything in git and use the state and pillar git backends for everything
19:19 babilen mpanetta: Well, I define the entire master configuration in my pillar. Whenever I require a new formula I also add it and run orchestration that will "update" the master (so that it knows about that new formula) before running a highstate on the minion that will then, in turn, use that formula
19:19 babilen yeah, everything is in git
19:19 mpanetta Ah ok I see
19:20 babilen The problem being: If salt messes up your master it is a bit of work to get it back on track :)
19:20 iggy that's on my TODO (orchestration to do it in order)... right now we just run highstate a few times in a row
19:20 * iggy whistles innocently
19:21 mpanetta Yeah we run orchestrations but our salt formula are layed out a bit different at the moment, so we don;t use any of the ones un git yet.
19:21 n8n joined #salt
19:21 * babilen hopes that he will be "done" with salt soon and can focus on finally work on monitoring (riemann, graphene, ...)
19:21 babilen But given that 2014.7 has just been released I will have a lot of work ahead of me ... also I should finally restructure my salt states to be more generic
19:22 babilen *even more
19:23 possibilities joined #salt
19:23 ckao joined #salt
19:23 babilen And release "saffron" to the salt community ... (easy-rsa salt module/reactor for creating/delivering certificates from a central CA to minions)
19:23 MTecknology Is it possible to have a state that recurses over files in a directory?... {% for filename in '/usr/local/etc/relay-files/*' %} {% set otherfile = os.path.basename(filename) %} state_stuff {% endfor %}
19:24 alex-mesos joined #salt
19:24 babilen MTecknology: What are you really trying to do? I wouldn't think that you can do what you exemplified ...
19:25 MTecknology I 'think'... I want to symlink every *.cf file pulled from a git repo
19:25 MTecknology maybe not...
19:25 babilen It helps if you know what you want to do ;)
19:26 babilen Anyway, bbl
19:27 iwishiwerearobot joined #salt
19:28 iggy babilen: when you get around to that stuff, I've got some stuff that I'm using to tie together salt and graphite/riemann (to graph salt events on my monitoring graphs, etc.)
19:29 babilen yay
19:31 babilen Do you deploy graphene?
19:31 MrFuzz joined #salt
19:32 babilen s/graphene/grafana
19:32 seydu I created I file phpconfig.sls. I put all the definitions of the managed files in it. In the top file a liste - phpconfig
19:32 gngsk *sigh* finally found out what was causing my issue
19:32 gngsk root cause, outdated documentation :(
19:33 forrest gngsk: can you file an issue for that? Or submit a PR to fix the docs?
19:33 gngsk well, it's a formula
19:33 gngsk so i suppose i should should branch the README and submit a PR?
19:34 babilen exactly
19:34 forrest gngsk: Yea, just fork the formula, make the changes, then make a PR and we'll get it merged in.
19:34 gngsk I will do that and feel good about saving someone else the headache.
19:34 forrest gngsk: Thanks!
19:34 gngsk Although, it was a good exercise in actually understanding how states, etc are compiled.
19:35 babilen gngsk: Let us know once you have the PR ready (a link or so)
19:35 dalibro joined #salt
19:36 aparsons joined #salt
19:36 gngsk will do
19:38 iggy graphite with the stock dashboard (for now.... muahahahaha)
19:38 felskrone joined #salt
19:38 babilen better than munin ;)
19:39 jerematic joined #salt
19:40 jerematic joined #salt
19:40 jcockhren iggy: upgrade. ;) http://grafana.org/
19:41 babilen yeah, just mentioned that earlier
19:41 jcockhren ah
19:41 jeffspeff SheetiS, that worked like a charm thank you.
19:41 jcockhren I'm late. heh
19:41 SheetiS jeffspeff: glad to hear it.  Thought it might based upon what you were doing :D
19:42 jeffspeff SheetiS, is there a better way to do if statements, other than nesting them?
19:42 SheetiS You could combine them with an 'and'
19:43 SheetiS {% if <condition1> and <condition2> %}
19:43 SheetiS you can span multiple lines with the condition as well if it is too long
19:43 jeffspeff ok, i'm not familiar with yaml at all. the c# in me would have added '&&'
19:43 SheetiS {% if <condition1> and \n     <condition2> %}
19:43 SheetiS something like that
19:44 SheetiS well it's jinja2 which is a templating system built upon python
19:44 patrek joined #salt
19:44 iggy I've just been running the stock dash for now since it's django which I'm pretty familiar with
19:44 SheetiS so much of the logic is very pythonic
19:44 n8n joined #salt
19:46 djstorm joined #salt
19:46 MTecknology spamassassin is an apache thing? ... now I get why it's so painful to work with! :D
19:46 iggy but it has been on my list to split up the graphite formula (or at least make what gets installed configurable)
19:46 iggy apache.org ... where dead projects go to die
19:47 MTecknology I'm trying to figure out how to include cf files from another directory. Having a tough time. The alternative is to make a symlink for every file
19:47 alex-mesos joined #salt
19:49 iggy the software doesn't have the ability to include from a directory?
19:50 MTecknology "include //usr/local/etc/relay-files/*.cf" is giving me the result ''warn: util: refusing to untaint suspicious path: "/usr/local/etc/relay-files/*.cf"''
19:51 MTecknology If I specify a specific file, it'll be included. I can't figure out how to include a directory. I want to include all files in there. It feels more I'm being left with the loop option which is icky. The channel is dead. I'm whining.
19:51 smcquay joined #salt
19:52 iggy did you try just including the dir?
19:52 iggy i.e. w/o the *.cf
19:52 MTecknology hm....
19:52 MTecknology it's not whining now...
19:53 MTecknology iggy: THANKS!!!!!!
19:53 peters-tx Does anyone know the best way to enforce minion responses?
19:53 peters-tx I run into a lot of "oh, these minions never completed" here and there
19:53 iggy MTecknology: it's actually including the files in there (as well as not bitching)?
19:54 peters-tx I would rather wait as long as it takes than be missing data
19:54 iggy you can change the timeout
19:54 babilen peters-tx: You are never missing data (cf. "salt-run jobs.lookup_jid $JID_OF_THE_JOB"), but you can adjust the timeout with "-t TIMEINSECONDS"
19:54 MTecknology iggy: yessir! You're a brilliant man! :D
19:55 babilen peters-tx: I'd just run the command in verbose mode and then lookup the JID. "salt -v '*' some.function"
19:55 MTecknology babilen: thanks for making me not try to do a silly thing
19:56 peters-tx iggy, babilen Ok I will give that a shot
19:56 bluenemo joined #salt
19:56 babilen MTecknology: Always ask about the thing you are trying to achieve not something that you think will be the solution (while it might not be) :)
19:57 bluenemo reading http://en.wikipedia.org/wiki/Salt_%28software%29, I do not get the difference between state and execution modules. can sbd explain that to me?
19:57 mpanetta execution module do all the work, state modules generally just call them.
19:57 babilen bluenemo: Execution modules are the things that actually "do" things and are typically used either interactively or by states to perform changes
19:58 Mso150 joined #salt
19:58 babilen bluenemo: States are typically a combination of tests and calls to execution modules that are being done when the tests fail.
19:58 forrest bluenemo: https://github.com/gravyboat/salt-book/blob/46b787a40ce3511e1136407b9a5fece366d08a19/en/chapter02-masterlessminion/chapter02.rst#the-difference-between-salt-states-and-salt-modules
19:58 bluenemo ok, so I understood this right. have to explain it to ppl tomorrow, so I thought I'd ask again. thank you guys :)
19:59 babilen bluenemo: So, say you have a "pkg.installed: - name: foo" state. The state will check if the package is installed and then, if it isn't, call, for example, the pkg.install function in the pkg execution module to actually install the package.
20:00 bluenemo ah ok. perfect example, thank you.
20:00 babilen (there are more layers as pkg would actually be aptpkg, but just to give an example)
20:02 kballou joined #salt
20:04 babilen bluenemo: https://github.com/saltstack/salt/blob/develop/salt/states/pkg.py#L918 would be the call to pkg.install execution module function in the pkg.installed state. The stuff around it essentially figures out "has the desired state already been achieved?".
20:04 smcquay joined #salt
20:05 bluenemo ah ok, cool. thank you babilen :)
20:05 gngsk hrmm, so it turnsn out i got the sls files to compile and run but that the desired affect has not been achieved
20:05 gngsk the hunt continues
20:07 possibilities joined #salt
20:07 smcquay joined #salt
20:08 babilen gngsk: You could paste them to http://refheap.com and include the output of your highstate/state run ... we might be able to spot something
20:09 gngsk is there a debug print function i can use in an sls file?
20:09 gngsk right now i just need to see if a jinja variable has a certain variable
20:10 babilen You can pass "-ldebug" to the master, minion and salt call. You can also use "salt-call ..." locally on the minion. For your case http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.state.html#salt.modules.state.show_sls might come in handy
20:11 bluenemo -l debug (with a whitespace)
20:12 bluenemo i prefer to set the debug levels on both master and minion to debug (or garbage) and then just tail -f the logfiles. sadly, especially with renderer errors (which are sometimes evil to track), this doesnt help much either (in my experience)
20:13 bluenemo gngsk, I sometimes do just a cmd.run: - name: echo {{ foobar }}, which fails but prints the variable / string / whatever in the debug output
20:13 babilen bluenemo: I write it without whitespace (I'm lazy, I know)
20:13 bluenemo havent found a cooler solution so far.. there IS  a way to just print strings without states, but I cant seem to find it anymore :(
20:13 bluenemo ah ok, didnt know it works without the whitespace.
20:14 gngsk bluenemo: thanks, i also found there's {{ show_full_context() }}
20:14 gngsk which may be a bit heavy handed, but hey!
20:16 b1nar1 joined #salt
20:17 bluenemo gngsk, the show_full_context is cool, I dont get exactly how to use that. how do you use taht? cmd.run: - name: echo {{ show_full_context() }} ?
20:17 Mso150_l joined #salt
20:21 Mso150_l_m joined #salt
20:24 gngsk yeah, but apparently i don't have a recent enough version of salt to use it
20:27 possibilities joined #salt
20:29 bluenemo that is really how one would use it? babilen isnt there a cooler way?
20:29 bluenemo also can it be that http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.state.html should specify which versions of salt are required? state.apply looks cool and like a nice shortcut to seperate between highstate and state.sls, but its not available for me
20:30 babilen A cooler way to do what? (and as compared to what?)
20:31 bluenemo compared to cmd.run - name: echo {{ show_full_context }}
20:31 Sylphid|home joined #salt
20:31 babilen I never mentioned that
20:32 bluenemo no, I was just asking you about it :)
20:32 bluenemo thought maybe you have an idea
20:35 nicksloan left #salt
20:35 Mso150 joined #salt
20:35 anotherZero joined #salt
20:36 bluenemo how do I show the salt version again, as in 2014.7.0. or helium or sth. my stupid ubuntu package is only outputting the ubuntu version string for --version and --versions-report
20:37 babilen bluenemo: You would use that in a jinja context ... say a manged file or a state.
20:38 babilen bluenemo: If you combine that with state.show_sls you would get the entire context (and note that it is show_full_context() not show_full_context)
20:38 bluenemo ok, so I was rigth with my echo. just a bit ugly, but good for debugging
20:38 babilen And you are running 2014.7 ?
20:38 agend joined #salt
20:38 bluenemo i'm trying to find out, but ubuntu is against me ;)
20:38 bluenemo stupid ubuntu..
20:38 babilen "salt --version-report"
20:38 iggy salt --version-report
20:39 * iggy shakes his fist
20:39 bluenemo yeah, it gives me Salt: 0.17.5 m(
20:39 babilen well, that is quite ancient
20:39 babilen You should, at least, upgrade to 2014.1.13
20:40 babilen (in which you also cannot use show_full_context as that was introduced in Helium (2014.7.0+))
20:40 babilen http://docs.saltstack.com/en/latest/topics/installation/ubuntu.html
20:40 gngsk I'm on 2014.1.10 (Hydrogen)
20:41 babilen I don't like recommending third-party repositories, but in this case it is an unfortunate necessity
20:41 bluenemo uh damn I joust found my error
20:41 bluenemo used the original ubuntu repos
20:41 bluenemo wtf
20:42 babilen Well, that would be quite an okay thing to do if it weren't for the fact that salt is a little special when it comes to packages. Does Ubuntu have backports too?
20:42 bluenemo wtf i installed it from the repos  of ubuntu. how did that happen? :D sorry for the stupid questions then.
20:42 babilen Well, but then 0.17.5 is still the version in wheezy-backports too so those wouldn't help you on Debian either
20:43 * babilen wonders what prevents an upload of .13 to backports
20:43 b1nar1_ joined #salt
20:43 bluenemo no i usually always take the salt provided repos.. just forgot about that I guess
20:44 bluenemo there we go: salt-master 2014.1.13 (Hydrogen)
20:44 babilen It shouldn't be necessary though. It is, IMHO, not a good thing to rely on third party repositories
20:44 bluenemo hm. in case of salt and omd / check_mk I dont really care. why not
20:46 bluenemo hm. still gives me http://paste.debian.net/130317/
20:46 babilen Because distributions have working repositories with an established trust path already. They also have the necessary infrastructure to ensure that broken upstream releases don't enter production environments (e.g. uploads to wheezy-backports wouldn't be allowed unless the package made it into testing)
20:47 babilen It also allows maintainers to tailor packages for the specific environment and backport patches whenever they see fit. Furthermore it wouldn't cause delays for all platforms after a release ...
20:47 bluenemo thats true, yes
20:48 bluenemo uah i'm stupid. my version is to low, sorry (pastebin)
20:48 bluenemo ok its getting late now :P
20:48 babilen It is a *very* peculiar setup that causes problems for maintainers and the salt platform in general. "Normally" software is released and *then* packaged by maintainers whenever they want to.
20:50 patrek joined #salt
20:51 babilen I don't even want to get started on security/certification issues that arise once you start using third-party packages. I'm happy that nobody started complaining, but I can *not* use salt for some projects based simply on the fact that I cannot use third-party (uncertified) repositories.
20:52 meylor joined #salt
20:52 babilen Which is ludicrous as the maintainer for saltstack's Debian packages is actually the Debian maintainer, but given salt's peculiar demands the in-house support is ... well ... overlooked
20:53 Mso150 joined #salt
20:58 MugginsM joined #salt
20:59 jstrom joined #salt
20:59 druonysuse joined #salt
21:00 loz-- joined #salt
21:01 jstrom hi, new salt user here. Running on two FreeBSD 10.0 machines with Salt 2014.1.13 from ports. I'm having some slowness issues when testing. Basic test.ping responds quick (0.3s), pkg.installed is quite fast too. but a simple state which should create a few small dotfiles, or install some packges seems to be a lot slower. Seems config is slow to load: http://pastebin.com/TnaYd3bi
21:01 jstrom 21:57:45,770 [DEBUG   ] Reading configuration from /usr/local/etc/salt/minion
21:01 jstrom 21:57:48,362 [DEBUG   ] Loaded minion key: /usr/local/etc/salt/pki/minion/minion.pem
21:01 jstrom should it really take a few seconds to read the config?
21:02 jstrom I've got the same issue on two slaves, one small Atom box, but the other oe is a beefy Xeon
21:03 jalaziz joined #salt
21:03 babilen jstrom: Make the web a better place and stop using poopy pastebins such as pastebin.com. There are many nice alternatives such as http://refheap.com, http://paste.debian.net, http://bpaste.net, http://gist.github.com, ...
21:04 babilen Oh, a captcha .. well, I'm not going to open that
21:04 walkingDude joined #salt
21:04 Mso150 joined #salt
21:05 jstrom babilen: hum, never seen a captcha on pastebin, but here is a gist: https://gist.github.com/stromnet/486b1c7e701c6991b541
21:05 ndrei joined #salt
21:05 babilen ta!
21:05 walkingDude hi, my salt formula uses something like: mongodb_configuration:  66   file.managed:  67     - name: {{ mongodb.conf_path }}  68     - user: root  69     - group: root  70     - mode: 644  71     - source: salt://mongodb/files/mongodb.conf.jinja  72     - template: jinja  73     - context:  74         dbpath: {{ db_path }}  75         logpath: {{ log_path }}  76         port: {{ settings.get('port', 27017) }}  77         bind_ip: {
21:06 tcolgate joined #salt
21:06 tcolgate hi
21:06 tcolgate I've got a quick question about custom grains and gitfs
21:06 walkingDude where the values for settings like this are being set? ->  settings.get('port', 27017) s
21:06 babilen jstrom: It happens, they also mangle input, a horrible layout, bad syntax highlighting support, useless "social media integration", no way to easily add to pastes, .... and simply make money of the fact that the great unwashed googles "pastebin" and uses their site.
21:07 babilen jstrom: Anything special in your minion config?
21:07 walkingDude is this taken from the pillar..?
21:07 tcolgate if I've got files checked in _grains in the git repo, are they distributed when I do a sync?
21:07 jstrom babilen: nope, straight copy from the saple config execpt the console output logging format (added timestamps)
21:07 jstrom babilen: and nothing configured on master either
21:08 jstrom everytime i run the state, it seems to do all those things (reading config and so on)
21:08 jstrom not sure if that is expected or not though.. :)
21:09 rojem joined #salt
21:09 ekristen joined #salt
21:09 tcolgate The config mentions file roots, but it seems a shame to have to maintain a manual checkout in /srv/salt just for grains
21:10 mosen joined #salt
21:11 ipmb joined #salt
21:11 babilen tcolgate: Why would that be necessary?
21:11 active8 joined #salt
21:11 babilen jstrom: I can take some measurements tomorrow .. my highstates run a little longer so that I don't really care about 3s delay at the beginning :)
21:12 babilen tcolgate: yes, they will be distributed
21:12 babilen (same for _modules, _states, ...)
21:12 jstrom babilen: well true, but I watched some demo video where they continously bragged about sub-second state execution and speed, so I figure something is wrong here.. ;)
21:13 tcolgate babilen: do I need to do anything more than have a fileserver_backend setup and do the sync?
21:13 tcolgate I guess I probably have something going wrong somewhere else
21:14 babilen jstrom: It is much faster for me
21:15 babilen jstrom: I can get you decent measurements tomorrow (but then I'm also on Debian which might not be a good comparison)
21:15 jstrom babilen: sure, I'll hang around
21:15 babilen tcolgate: What is your actual problem?
21:15 zekoZeko_ joined #salt
21:15 tcolgate the grains don't seem to be syncing
21:16 tcolgate babilen, it worked fine when I had a roots backend in /srv
21:16 tcolgate states do seem to be being pulled, so I think gitfs is working ok.
21:16 babilen tcolgate: Which grain is not syncing and why do you think that it should? Is it executable? Did you restart the minion and are all modules available on the minion that are used by the grain?
21:17 dooshtuRabbit joined #salt
21:17 babilen Trust me, it works fine with GitFS
21:17 tcolgate it's an custom .py script I've had in /srv/salt/_grains. and has been working ok
21:17 pdayton joined #salt
21:17 jstrom babilen: i'll upload some more verbose logging, might help..
21:17 tcolgate babilen, not to worry, I've only just really started fiddling with this stuff, I'll poke at it for a while longer.
21:18 babilen So all you did was to move the file from a local /srv/salt/_grains/foo.py to one of your git repositories? What is the mode of the file in the repo? Do you find it in the gitfs cache on the master?
21:18 babilen Or in extmods on the minion?
21:18 babilen (/var/cache/salt/....)
21:19 babilen How do you sync it? Could you paste your grain and the rest of the information to, say, http://refheap.com ?
21:19 babilen But then .. I think I'll go to bed now. :D
21:19 tcolgate I coldn't find it under /var/cache, I check the permission, this is all new to me, as I say, will play on for a bit
21:19 mosen joined #salt
21:20 ingwaem but the sun is shining babilen, can't sleep now :D
21:20 tcolgate the fact it's supposed to work is enough for me, I;ll work it out.
21:22 b1nar1 joined #salt
21:23 karimb joined #salt
21:24 jstrom babilen: as verbose as it gets.. ;) https://gist.github.com/stromnet/fb30f8523b6a11d6dd22, my execution starts at line 913
21:24 Ryan_Lane I'm upgrading salt during a highstate, and I tell it to reload modules, which seems like it does, but then later states fail
21:25 Ryan_Lane is it not possible to have salt manage itself via virtualenv.managed?
21:27 jstrom babilen: looks like the search path lookups are slow?
21:28 SheetiS Ryan_Lane: I've not found a way to do it myself.  I've always had to isolate my 'salt upgrades itself' state and run it individually, then go back and run the highstate for 'everyting else' after.  Of course I could be doing it all wrong as well.
21:28 Ryan_Lane well, it's looking like that's what I'll need to do
21:30 mephx joined #salt
21:31 iggy tcolgate: "salt-call cp.list_master" is helpful at times
21:33 aparsons_ joined #salt
21:36 glyf joined #salt
21:36 ipmb joined #salt
21:38 possibilities joined #salt
21:39 Eugene Well, this is weird. I just received an email from a "Travis Sabin" concerning a "BYU-SaltStack Student Survey"
21:39 Eugene According tot he email, I've chosen not to use Salt.
21:40 Eugene I wonder if this is because I had an inquiry into the Enteprise product pricing, heh
21:40 jstrom babilen: nope, added some extra trace statements between every 'could not be found' error; the delay is somewhere before which() is called
21:46 aparsons joined #salt
21:54 monkey66 left #salt
21:56 dalexander joined #salt
22:00 glyf joined #salt
22:02 aparsons joined #salt
22:08 hasues joined #salt
22:08 iggy forrest: did you ever get anywhere with your aptly formula?
22:08 aparsons joined #salt
22:09 forrest iggy: oh yeah it's mostly done, I just didn't finish up the keys portion: https://github.com/gravyboat/aptly-formula/blob/master/aptly/aptly_config.sls#L69
22:09 forrest iggy: so if you just comment out those lines, it should work.
22:09 rap424 joined #salt
22:10 hasues left #salt
22:11 iggy you are putting your secret key in your state dir?
22:11 meylor joined #salt
22:11 forrest iggy: it would live in pillar
22:12 iggy salt:// paths resolve to pillar files?
22:12 riceandbeans joined #salt
22:13 forrest iggy: oh sorry I was looking at line 69+, I forgot I left that other section in there as the 'ghetto' workaround
22:13 forrest while I was doing the pillar stuff.
22:13 forrest just comment that out if you want, and ensure that aptly:secure isn't in pillar, and it just won't look for the key.
22:15 zekoZeko_ joined #salt
22:15 iggy hmm, but I do actually want to sign stuff... guess I'll use that as a starting point... save myself a little time
22:15 iggy thanks forrest
22:16 forrest iggy: let me make a PR against the actual formula, then you can fork and continue working on it if you want, hang on
22:17 forrest iggy: unless you want to fork mine *shrug*
22:17 iggy yeah, I can do that
22:17 forrest I just don't like to make PRs for formulas that don't work.
22:17 forrest okay cool
22:17 iggy I can always fixup the origin later
22:17 * timoguin glances at aptly
22:19 forrest timoguin: it's pretty awesome.
22:19 forrest iggy: yeah, just make a PR against mine if you get the key stuff working, it's mostly implemented now. I just got sidetracked :P
22:21 LGSilva joined #salt
22:22 LGSilva hello guys. when I include a state so I can use "require: - sls: foo" will that state be executed twice as it is also included in the top file?
22:23 aurynn require: just sets up precedence
22:23 aurynn as far as being run twice, I don't know. I don't think it would.
22:24 LGSilva ok, thank you aurynn
22:24 iggy it doesn't
22:24 timoguin No, the order will be evaluated at run-time, and states should only run once
22:24 iggy if it did, I'd be in trouble because I have lots of redundant includes to get at stuff for watch/require/etc
22:31 forrest Does anyone remember if centos has something weird with pkg.installed? I'm getting The following packages failed to install/update: ruby=2.0.0p481 when I specify the version, but the package is already installed, and it did install the package, but requires are failing now.
22:34 SheetiS forrest: I've not tried that specific thing on CentOS6, but I constantly have problems with older versions of all kinds of packages making my life difficult with salt on CentOS6.  (Old Jinja2 version comes to mind at first.  My bootstrap actually updates it from the supplied OS package).
22:35 forrest SheetiS: hmm, yeah I need a very specific release of ruby, it's weird it's installing it but thinking it isn't, I'll try doing the 'full' version to see if that helps at all.
22:36 SheetiS Forrest:  I always use the rvm stuff for specific ruby versions (though my states might apply more quickly if I packaged it instead).
22:37 forrest SheetiS: yeah RVM isn't an option in this situation.
22:37 SheetiS Ahh.  I could see that, especially if you don't want gcc and friends on the system.
22:39 nyx_ joined #salt
22:40 nitti joined #salt
22:40 iggy too bad nmadhok already went night night :/
22:43 nyx_ joined #salt
22:43 forrest iggy: why? Do you need a formula merged?
22:43 active8 joined #salt
22:43 iggy I mean if you're bored... collectd-formula
22:44 CeBe1 joined #salt
22:44 forrest iggy: he already merged it
22:45 iggy lol
22:45 iggy I was even hitting refresh to see if someone caught it
22:48 Ryan_Lane 2014.7 released and I'm already having to patch it :D
22:49 Ryan_Lane (with our own changes being upstreamed in dev)
22:51 kermit joined #salt
22:51 oeuftete I'm using salt-cloud to deploy to ec2 with the minion configured to run highstate when it starts.  How can I reliably check to see when state.highstate is done?  saltutil.running looked like the answer, but it starts returning an empty list long before the job is complete
22:53 forrest SheetiS: I got it figured it, the issue was that it needed the full release of 2.0.0p481-1.el6 not just 2.0.0p481
22:55 SheetiS forrest: that makes sense.  It'd just have to match whatever you'd have to do for the equivalent in yum.  I was actually kinda peeking through https://github.com/saltstack/salt/blob/2014.1/salt/modules/yumpkg.py ti see what was specifically called under the hood.
22:55 forrest SheetiS: yeah I don't know what was up with it, it was catching that straight off the command line, so I'm not sure
22:56 saurabhs joined #salt
22:56 iggy there's a setting in 2014.7 that might make that easier
22:57 iggy maybe not
22:58 iggy normalize was what I was thinking of, but it looks like it just chops off the arch if it's there
22:59 SheetiS forrest: the package name is split on '.' in the source it looks like
22:59 forrest SheetiS: yeah there was an issue with how things were split a while back that was resolved.
23:00 SheetiS Anyhow, I guess it doesn't matter too much as long as you got it working
23:01 GrueMaster joined #salt
23:02 CeBe1 joined #salt
23:04 forrest SheetiS: Heh, yeah
23:05 rojem joined #salt
23:07 gngsk joined #salt
23:07 baconbeckons joined #salt
23:08 Gareth Ryan_Lane: it's officially out?
23:09 Ryan_Lane it was tweeted, so it looks like it :)
23:09 dalexander joined #salt
23:09 Gareth Twitter makes everything official :)
23:09 baconbeckons i have file_roots defined for my salt master. i am using salt formulas and i thought that i could include the formula roots in base, but salt complains because the formula aren’t defined in the other environments. do i need to explicitly list the paths for the formula in every single file_root environment?
23:10 StDiluted Has anyone written a salt/packagecloud.io integration?
23:11 iggy I just link the formula subdir's into my main salt states dir (when I'm using that)
23:11 baconbeckons iggy: that might be a better solution :)
23:12 gngsk babilen: https://github.com/saltstack-formulas/iptables-formula/pull/5
23:14 Corey_ joined #salt
23:14 pssblts joined #salt
23:15 aparsons_ joined #salt
23:22 alex-mesos joined #salt
23:22 nitti joined #salt
23:23 Corey {%- set _vip = grains.internal_vip %} <-- if I want to convert this to using pillar, would it become {%- set _vip = pillar[internal_vip] %} ?
23:23 bhosmer_ joined #salt
23:23 manfred yes
23:23 manfred or
23:23 manfred pillar['internal_vip']
23:24 gngsk how about salt['pillar.get']("internal_vip") ?
23:25 manfred you could do that
23:25 Corey Is there a meaningful differece?
23:25 manfred and you can set a deafult with that
23:25 iggy or pillar.get('internal_vip', 'default')
23:25 manfred the biggest difference is you can do stuff like salt['pillar.get']('vips:internal', '')
23:25 manfred and chain together dictionary keys
23:25 iggy ^
23:32 marshalls joined #salt
23:32 marshalls Is there anything out there to monitor SaltStack installations besides halite? I have it installed but it's not really what we are looking for. It would be nice if we could have a web gui or even a daily email that showed the status of all the minions. I'm pretty handy with scripting but I don't know what to script.
23:33 dalexander joined #salt
23:33 dalexander joined #salt
23:37 iggy I don't know of anything
23:38 iggy I know some people have talked about rigging up something with returners/outputters to a database of some sort and then writing some views into that in some sort of way
23:38 Ryan_Lane marshalls: you could always send your highstates to another source with a returner
23:39 Ryan_Lane heh
23:39 Ryan_Lane if you use the returner, then you would have all the data necessary in a database, which you could scrape for reports
23:39 psidox joined #salt
23:39 LonEagle joined #salt
23:39 Ryan_Lane you could likely get the same info from the job data on the master
23:40 Ryan_Lane I find working with the jobs to be a pain, though
23:40 iggy yeah, honestly, most people seem to be more interested in seeing what failed and when rather than what halite offers (which is basically just 'salt "*" state.highstate test=True'
23:41 iggy saltstack enterprise apparently has a fancy dashboard/gui
23:41 hopthrisC joined #salt
23:41 marshalls So if I used a mysql returner does each minion write back to the db or does the master write to it?
23:42 nitti joined #salt
23:43 Ryan_Lane iggy: eh? I haven't heard that there's a dashboard for enterprise
23:44 Ryan_Lane the salt folks had said they were aiming at being fully open-source. that would be a bit annoying :)
23:44 LonEagle there is something I haven't been able to figure out, looking at salt's documentation. is there a dynamic way to select which configuration is highstate?
23:44 tafa2 joined #salt
23:44 iggy the salt meetup that was televised the other day was supposed to have a demo of it I read
23:44 Ryan_Lane iggy: ah. I watched that
23:44 Ryan_Lane they have mockups for what they're building
23:44 iggy maybe it was just halite and they were making it sound better than it really is
23:45 Ryan_Lane apparently also have a live-demo, but weren't sure it would work correctly, so showed the mockups
23:45 Ryan_Lane it's not halite
23:45 Ryan_Lane it's something new
23:46 iggy and after I just got done making a big stink about halite being abandoned the other day
23:46 meylor joined #salt
23:48 murrdoc its a maria db backed thing
23:48 murrdoc not sure why maria db when they could POSTGRES!
23:48 murrdoc the mockups look pretty cool
23:51 druonysus joined #salt
23:51 druonysus joined #salt
23:55 LonEagle my situation is that i have a farm of computers that are either under automation control or user control.
23:55 aquinas joined #salt
23:55 LonEagle if a user has a system checked out, I can't do as invasive of configuration management as I otherwise would.
23:58 bhosmer_ joined #salt

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