Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-11-18

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

All times shown according to UTC.

Time Nick Message
00:00 stooch1122 joined #salt
00:01 DanGarthwaite joined #salt
00:43 sroegner_ joined #salt
00:55 zandy joined #salt
00:58 DanGarthwaite left #salt
01:05 sroegner_ joined #salt
01:09 sroegner_ joined #salt
01:12 halfss__ joined #salt
01:17 bemehow joined #salt
01:21 alekibango joined #salt
01:23 ajw0100 joined #salt
01:32 pdayton joined #salt
01:37 jjasinski_ joined #salt
01:52 malinoff joined #salt
01:55 malinoff Hi
01:55 zandy joined #salt
01:56 rgarcia_ joined #salt
01:57 winway joined #salt
02:01 jjasinski_ joined #salt
02:03 winway left #salt
02:12 snewell joined #salt
02:13 hvn joined #salt
02:13 jalbretsen joined #salt
02:14 hvn left #salt
02:15 xl1 joined #salt
02:23 franky joined #salt
02:23 sroegner_ joined #salt
02:40 thelorax123 joined #salt
02:55 zandy joined #salt
03:00 tulu joined #salt
03:33 Baribal_ joined #salt
03:34 thelorax123 joined #salt
03:35 ravibhure joined #salt
03:36 nicksloan yay!
03:36 nicksloan it's working!!!!
03:37 thelorax123 joined #salt
03:37 nicksloan well, okay. Maybe it works.
03:39 malinoff nicksloan, it is :)
03:40 nicksloan I'm beginning to think it's not
03:40 nicksloan oh, nice. You must have turned the apt mirror back on for me. Thanks. :-)
03:41 zakm joined #salt
03:42 nicksloan balling out of control
03:42 nicksloan why did I spend so much time on puppet
03:43 malinoff nicksloan, dunno what are you talking about, but it works :) Especially, salt works much better than puppet
03:46 nicksloan after getting salt working for the first time:
03:46 nicksloan [22:46:14] <nicksloan> alright, I have to go to bed now so I can make it in to work early tomorrow. Can't wait!!!!
03:49 malinoff Hah, I bet you will complain about salt for some time :) Docs are huge and uninformative, there are bugs and regressions. But of course, salt is awesome
03:49 malinoff Don't be afraid to ask here or in the mailing list
04:06 jankowiak joined #salt
04:07 ckao joined #salt
04:09 __number5__ BenBen: /topic
04:09 __number5__ oops
04:10 __number5__ I thought 0.17.2 was released...
04:21 bemehow joined #salt
04:21 junedm joined #salt
04:27 Katafalkas joined #salt
04:31 malinoff __number5__, it is, on PyPi
04:31 malinoff It should be in repositories today/tomorrow, as I heard
04:32 __number5__ cool!
04:35 __number5__ topic needs update :)
04:36 crazysim joined #salt
04:41 anuvrat joined #salt
04:55 zandy joined #salt
05:08 druonysus joined #salt
05:08 druonysus joined #salt
05:12 junedm left #salt
05:31 redondos joined #salt
05:34 thelorax123 joined #salt
05:42 forrest joined #salt
05:55 zandy joined #salt
06:19 rmt joined #salt
06:25 junedm joined #salt
06:30 davet joined #salt
06:32 matanya joined #salt
06:32 micko joined #salt
06:42 Furao joined #salt
06:55 zandy joined #salt
06:58 redondos joined #salt
07:06 slav0nic joined #salt
07:07 ollins joined #salt
07:07 hhenkel Good morning everybody. I'm currently looking into gitfs/hgfs backend for salt and I'd like to know if it is possible to use this with authentication.
07:08 hhenkel Or is it only possible to use this features with public web repositories?
07:11 _ikke_ hhenkel: You could use ssh keys
07:12 junedm what if my ssh key have the passphrase ?
07:13 _ikke_ github for example has specialized deploy keys
07:13 _ikke_ Those are passwordless ssh keys, but limited to only one project
07:13 EWDurbin downside to GH deploy keys is they still have write access :/
07:13 EWDurbin and they don't necessarily have to be passwordless
07:15 hhenkel _ikke_: okay, I saw that. So web authentication (we currently use rhodecode) is not possible?
07:15 _ikke_ hhenkel: You mean over http?
07:15 hhenkel _ikke_: yes
07:16 _ikke_ hhenkel: one possibility would be to provide the credentials in the url
07:16 _ikke_ Another would be to use curlrc
07:16 bhosmer joined #salt
07:16 _ikke_ another option (but more work) is to use credential helpers in git
07:17 hhenkel _ikke_: ah, good point, totally forget about providing them in the url, regarding the curlrc I'll check.
07:18 hhenkel _ikke_: ++
07:18 hhenkel Hmm, no karma bot around it seems... :(
07:22 Nexpro1 joined #salt
07:41 andy___ joined #salt
07:42 andy___ hello
07:42 andy___ is any one here?
07:43 _ikke_ nope
07:53 tom_ joined #salt
07:54 tom_ Hi, I would look for some assistance for executing a git.latest state using a pwd-protected ssh key. see the question at SOF: http://stackoverflow.com/questions/20000993/how-can-i-execute-the-salt-state-git-latest-with-password-protected-private-key . Thanks for your help!
07:56 IJNX joined #salt
07:56 magicbill joined #salt
07:58 _ikke_ tom_: Either install an ssh agent that has the key, or use a passwordless key
07:59 redondos joined #salt
07:59 joehh hhenkel: we use http://hgtip.com/tips/advanced/2009-10-01-configuring-user-auth-https/ with the password
07:59 joehh in the hgrc file
08:00 joehh and the hgrc file managed by file.managed
08:00 joehh havent done so with hgfs though, only the hg state
08:01 harobed joined #salt
08:03 tom_ ok thanks to you both. I think using an ssh agent which has the key is the way to go. which would mean I have to login to my minions first and set the password in the ssh agent.
08:03 balboah joined #salt
08:10 networkpadawan joined #salt
08:25 carlos_ joined #salt
08:30 hhenkel joehh: Thanks for the link!
08:32 redondos joined #salt
08:37 linjan joined #salt
08:52 bemehow joined #salt
08:55 zandy joined #salt
08:57 the_drow joined #salt
08:58 the_drow Hi guys I'm having trouble with installing pyenv using salt
08:58 the_drow pyenv is a bash tool that depends on editing .bashrc
09:00 the_drow https://dpaste.de/qxOu - First of all for some reason I have to source .bashrc every time. Second of all it seems that even though I'm using bash and sourcing it doesn't catch the appended text correctly.
09:01 the_drow Does anyone have any leads?
09:04 giantlock_ joined #salt
09:06 aib_233 joined #salt
09:15 sroegner_ joined #salt
09:21 zandy joined #salt
09:25 mekstrem joined #salt
09:25 Iwirada joined #salt
09:25 Koma left #salt
09:27 Psy-Q_ joined #salt
09:32 redondos joined #salt
09:34 tom_ I have another thought regarding pwd protected keys on minions.. Do you think it is somehow possible in saltstack to share the master keychain with the minion. this would allow me to setup the passphrases in my keychain at the master. in ssh this could be done with the -A switch: http://superuser.com/a/249712/62892
09:38 ravibhure1 joined #salt
09:40 balboah I see 0.17.2 was tagged in github 4 days ago. Any plans on updating the ubuntu ppa?
09:41 Ryan_Lane joined #salt
09:46 aleszoulek joined #salt
09:46 joehh balboah: it is waiting for the official announcement
09:46 balboah ok
09:47 joehh if you wish to get a sneak preview, you could use the salt-depends ppa (https://launchpad.net/~saltstack/+archive/salt-depends)
09:47 joehh at or around the time of the announcement, they will be copied from there to the main repo
09:48 balboah cool, I got it from pip now just for a test
10:02 Furao joined #salt
10:04 baoboa joined #salt
10:04 hhenkel Hi all, what is the best option to have a semi random number (for cronjobs) with salt? It would be nice to generate a number from a string.
10:09 hhenkel Okay, just saw the "random" option. Is there another option to use something like this within a regular file
10:09 Furao joined #salt
10:19 joehh hhenkel: you could do something similar to https://groups.google.com/forum/#!topic/salt-users/AyhLudjvqvg with a custom module that
10:19 joehh returned random numbers/strings that were appropriate for your use
10:19 Furao joined #salt
10:20 matanya joined #salt
10:22 Valdo joined #salt
10:22 jpcw joined #salt
10:25 mastrolinux joined #salt
10:33 redondos joined #salt
10:40 Damoun joined #salt
10:46 djn joined #salt
10:47 Ryan_Lane joined #salt
11:13 timbobsteve joined #salt
11:13 timbobsteve Hi all
11:13 timbobsteve has anyone got any working examples of SaltMine ?
11:13 timbobsteve The docs don't do it much justice
11:14 Furao timbobsteve: check module doc it's more useful
11:14 timbobsteve Furao: http://docs.saltstack.com/ref/modules/all/salt.modules.mine.html#module-salt.modules.mine ?
11:15 sroegner_ left #salt
11:16 Furao yeah
11:16 sroegner_ joined #salt
11:16 Furao I don't know if this was fixed in 0.17.x but in 0.16.x mine.update do nothing
11:17 timbobsteve am I right in thinking that Mine just periodically calls a normal module function (e.g. cmd.run, or network.interfaces) and sends the values to be stored on the master for safe keeping?
11:32 backjlack joined #salt
11:33 malinoff timbobsteve, No. Mine system periodically gets a data from a minion and transfer it to all minions via the master
11:35 timbobsteve malinoff: But the minion doing "collecting"... it just calls module functions right to find the values to send around?
11:40 malinoff timbobsteve, yes, something like that
11:43 beardo joined #salt
11:47 Ryan_Lane joined #salt
11:49 bhosmer joined #salt
11:50 the_drow Hi, how do I run a command using a login shell?
11:51 timbobsteve Furao: You mention that mine.update does nothing on some versions of Salt... I am running it and it does seem to talk back to the master, but none of the minions mine data is updated... if I do mine.send from the same minion the values update fine. I'm on 0.17.1, so it looks like it might still be an issue.
11:52 IJNX joined #salt
11:55 diegows_ joined #salt
11:57 the_drow Furao: Hi man. Got any idea?
11:57 hhenkel joehh: okay, thanks. That looks good.
11:59 malinoff the_drow, you mean, you need to ask a username/password before running a command?
12:00 hhenkel Is it possible to have states only enforced during defined times and if possible only once (a day)?
12:01 malinoff hhenkel, scheduler
12:01 hhenkel Puppet uses a scheduler resource for such things, what is the salt equivalent to that?
12:01 rjc joined #salt
12:01 malinoff http://docs.saltstack.com/topics/jobs/schedule.html
12:01 hhenkel malinoff: okay, will look into that, thought it's only possible to apply it every n hours or so
12:03 malinoff hhenkel, but of course, you can always use cronjob
12:08 hhenkel malinoff: hmm, I only want to schedule a single resource. At the moment we're rolling out a jar file at night and this creates the need to restart the weblogic servers afterwards.
12:08 hhenkel I want to be able to apply the states during the day, but I don't want anything to happen during daytime..
12:08 malinoff hhenkel, ▲ that link should help :)
12:09 jcsp joined #salt
12:09 malinoff I'm afraid, salt can't stop you to run a command if you have access to do that
12:10 malinoff Use your system mechanisms to prevent any critical runs
12:13 elfixit joined #salt
12:14 backjlack joined #salt
12:15 nicksloan hmm
12:15 nicksloan I'm kind of bummed out by the format of salt formulas.
12:16 hhenkel malinoff: I don't want salt to stop me. I'm looking for a mechanism that executes a resource only between special hours, something like: http://docs.puppetlabs.com/references/latest/metaparameter.html#schedule
12:17 nicksloan seems they could have easily been set up such that they could easily be used as git submodules.
12:17 hhenkel As far as I understand the link this kind of schedule provides the ability to rerun something every n hours or x minutes.
12:18 malinoff hhenkel, I don't think salt has such ability
12:19 hhenkel malinoff: okay, I think I'll file a feature request then.
12:19 malinoff hhenkel, try to search at first, I'm 90% sure I heard about that
12:20 hhenkel malinoff: I searched the salt page or do you mean within the issues?
12:21 malinoff hhenkel, the issues
12:21 malinoff And mailing list
12:24 hhenkel malinoff: https://github.com/saltstack/salt/issues/8122
12:26 malinoff hhenkel, Yeah! That's it
12:26 timbobsteve malinoff: If I use "__salt__['mine.get'](minion_id,'network.interfaces')" will the value be in a dict with the minion-name as the key?
12:27 malinoff timbobsteve, you can test it by yourself: # salt-call grains.get os --out=raw
12:27 malinoff {'local': 'CentOS'}
12:28 timbobsteve legend!... had not realised that was available
12:28 malinoff :)
12:30 malinoff Legend-wait-for-it-dary :)
12:30 timbobsteve hahahaha
12:32 juasiepo joined #salt
12:35 redondos joined #salt
12:35 nicksloan can anyone provide some context for these commits: https://github.com/saltstack-formulas/salt-formula/commit/46652ebfda2603bf355c508a7e97d134579bcc85 https://github.com/saltstack-formulas/salt-formula/commit/fa29decc39bf10ee5f6b2eb7a00205c36818b948
12:36 nicksloan basepi: ^
12:36 nicksloan I liked your way better.
12:36 harobed_ joined #salt
12:40 sroegner_ joined #salt
12:44 timbobsteve Anyone hit an integer conversion issue in minion_pub()? I'm running "publish.publish" in a state file via Jinja and I'm getting TypeError because of the ret['jid'] value not being converted :_/
12:48 Ryan_Lane joined #salt
12:50 dangerousbeans joined #salt
12:50 dangerousbeans OH HAI
12:51 timbobsteve dangerousbeans: Hai :P
12:51 dangerousbeans timbobsteve: whats cooking?
12:52 dangerousbeans I'm here because inevitably I'll need to ask lots of noob questions soon, but not yet...
12:52 timbobsteve dangerousbeans: that's fine. I'll answer what I can, but I'm also partaking in the noob-spam approach to IRC ;-)
12:58 the_drow malinoff: nope I mean : #/usr/bash -l
12:59 blee joined #salt
13:04 anitak joined #salt
13:07 micko_ joined #salt
13:09 cron0 joined #salt
13:12 timbobsteve left #salt
13:33 dox joined #salt
13:34 Furao joined #salt
13:34 jankowiak joined #salt
13:35 redondos joined #salt
13:36 dangerousbeans joined #salt
13:37 brianhicks joined #salt
13:40 aserdp joined #salt
13:40 jslatts joined #salt
13:41 sroegner_ joined #salt
13:43 NotreDev joined #salt
13:44 stooch1122 joined #salt
13:49 Ryan_Lane joined #salt
13:49 lemao joined #salt
13:53 mpanetta joined #salt
13:55 bhosmer_ joined #salt
13:57 juicer2 joined #salt
13:59 racooper joined #salt
14:02 Brew joined #salt
14:02 balboah so if you want to have a custom returner to only log things not succeeding, you should iterate ret["return"] looking for entry["result"] != True?
14:03 balboah or can you configure the default to not populate returners with succeeding stuff?
14:03 druonysus joined #salt
14:03 druonysus joined #salt
14:06 mpanetta_ joined #salt
14:09 Gifflen joined #salt
14:11 dangerousbeans Hey, I'd like to run a command, but only if a file does not exist
14:11 dangerousbeans HOW IS IT DONE?
14:11 dangerousbeans - unless?
14:12 ipmb joined #salt
14:14 vejdmn joined #salt
14:14 _ikke_ - unless test ! -f path/to/file?
14:14 _ikke_ - unless test -f path/to/file?
14:15 dangerousbeans _ikke_: awesome thnx
14:18 _ikke_ yw
14:22 blafountain joined #salt
14:23 jeddi joined #salt
14:25 timoguin joined #salt
14:30 harobed joined #salt
14:35 ravibhure joined #salt
14:36 redondos joined #salt
14:39 JulianGindi joined #salt
14:44 tty-1 joined #salt
14:49 mattmtl joined #salt
14:54 Bram joined #salt
14:55 JasonSwindle joined #salt
14:55 Bram hey, I have a strange problem, salt-minion just stop to work for a very strange reason (and I can't find a solution on the internet), apparently it didn't managed to load the grains, here is the traceback http://paste.debian.net/hidden/62408e3b/
14:56 Bram anyone have an idea of how I could solve this?
14:57 donatello joined #salt
14:57 jslatts joined #salt
14:57 donatello hi all
14:58 Bram (oh, and I'm on ubuntu LTS using the ppa running 0.17.1)
14:58 NotreDev joined #salt
15:01 donatello is it possible at this point (i.e. 0.17.1)  to create throw away salt-ssh roster files and use them to just setup the salt-minion?
15:02 donatello it would be great if it were possible to use a custom roster file. its not very clear what the "--roster-file" option works. any pointers would be great
15:05 smccarthy joined #salt
15:07 mkoistinen joined #salt
15:08 nicksloan where does salt get files that I refer to with salt:// from?
15:08 mkoistinen 'SaltRenderError: Undefined jinja variable' in salt-ssh. Offending line is simply: {% for p in grains['projects'] %} what gives?
15:09 anuvrat joined #salt
15:10 _ikke_ mkoistinen: projects is a valid key$
15:10 mkoistinen yes
15:11 mkoistinen salt-ssh 'xxx' grains.get projects => ['aaaa', 'bbbb']
15:13 Valda joined #salt
15:13 mkoistinen Also just tried removing one of them 'bbbb' and same error.
15:14 mkoistinen And, now I've just commented out the lines, and it still bitches about it!
15:14 mkoistinen Clearly a Jinja issue.
15:14 AdamSewell joined #salt
15:16 alunduil joined #salt
15:16 NotreDev joined #salt
15:20 quantum-x joined #salt
15:21 quantum-x Hey all. I'm deploying a state across a nodegroup. One of the files deployed in the state needs the IP of the machine it's being deployed too. I'm gone through the documentation, but can't seem to find an obvious solution to this challenge ?
15:31 mkoistinen Also: salt-ssh 'host' state.show_state mysql =? no matching sls found for 'm'; no matching sls found for 'y'...
15:32 NotreDev joined #salt
15:33 eclectic quantum-x: have you tried using the ipv4 grain?
15:33 timoguin quantum-x, you'll want to use a template in the file and accessing the minions grains like so: {{ grains['ipv4' }}
15:33 amahon joined #salt
15:33 timoguin or something similar
15:37 quantum-x OK, thank you very much :)
15:37 opapo joined #salt
15:37 redondos joined #salt
15:38 eclectic Does this look legit: {{ grains['host'] }} to refer to the host itself, and {{ salt['grains.get'](‘ip_interfaces:eth0′)[0] }} to get the host’s IP
15:40 cnelsonsic joined #salt
15:45 bhosmer_ joined #salt
15:47 pentabular joined #salt
15:51 zandy joined #salt
15:52 quantum-x timoguin, that looks like the perfect solution - one question - the box i am on has multiple IPs, how would I go about grabbing just one IP (the first in the list?)
15:52 d0ugal joined #salt
15:53 [diecast] joined #salt
15:54 timoguin quantum-x, I think grains['ipv4'][0] might do the trick for grabbing the first one
15:54 quantum-x how logical :) thanks kindly sir
15:55 bhosmer joined #salt
15:55 zandy joined #salt
15:57 tyler-baker joined #salt
15:57 kaptk2 joined #salt
16:01 pentabular joined #salt
16:03 nicksloan can I call multiple functions on a given state?
16:04 jcockhren nicksloan: yes. per stateID
16:04 jankowiak joined #salt
16:05 quickdry21 joined #salt
16:05 jergerber joined #salt
16:06 nicksloan jcockhren: https://gist.github.com/nicksloan/5cc981acd587ea4c7f29
16:06 nicksloan so that's wrong?
16:06 nicksloan I need a separate name for each one?
16:07 jcockhren I think I misunderstood. But it does look correct
16:07 foxx joined #salt
16:08 viq nicksloan: from my playing around, yes, you need separate block for each file
16:08 jcockhren check with basepi on if you have to list the file state name
16:08 jcockhren viq++
16:08 forrest joined #salt
16:10 jeffasinger joined #salt
16:11 nicksloan hmm. Must have misunderstood something
16:12 nicksloan I have salt, pillar, and files in the directory in my file_roots directive
16:12 nicksloan thought I could get to files by doing salt://files/whatever/fun.conf
16:12 nicksloan where whatever/fun.conf is in the files directory
16:12 nicksloan is that wrong?
16:13 viq nicksloan: https://gist.github.com/7530516
16:14 viq Near the end, I have file managed in sentry and sentry_service, beacuse I couldn't have two in one top-level block
16:14 nicksloan viq: yeah. I split them up
16:14 basepi nicksloan: yes, each file declaration needs its own ID, so you can't put them all under the `supervisor` id
16:14 nicksloan now my issue is that my masterless minion can't find the files
16:15 viq nicksloan: also, if you want salt://files/whatever/fun.conf then the file needs to be in /srv/salt/files/whatever/fun.conf
16:15 basepi nicksloan: for masterless, have you set `fileclient: local` and are you using `--local` on your calls?
16:15 basepi http://docs.saltstack.com/topics/tutorials/standalone_minion.html
16:16 nicksloan basepi: I am using --local
16:16 nicksloan I have not set fileclient: local. I assume that goes in my minion config?
16:16 nicksloan oh, but I have set file_client: local, per the page you linked
16:16 ipmb joined #salt
16:17 viq nicksloan: and where are your files located?
16:17 basepi oh, sorry, forgot the underscore
16:17 nicksloan \/srv/salt/files (where /srv/salt contains salt, pillar and files directories)
16:17 nicksloan blurgh
16:17 nicksloan omit the first slash
16:18 nicksloan /srv/salt/files (where /srv/salt contains salt, pillar and files directories)
16:18 nicksloan better
16:18 basepi nicksloan: i assume you've defined file_roots in your minion config?
16:18 nicksloan yep
16:20 basepi nicksloan: hrm, that's odd.  can you submit a pull req?  include all your version info, gists of your config and states, etc.
16:20 basepi sorry, meant issue
16:20 basepi not pull req
16:20 basepi (still waking up my fingers, apparently)
16:20 nicksloan haha
16:21 teskew joined #salt
16:24 aleszoulek hey guys..
16:24 aleszoulek having a problem with saltutil.refresh_pillar
16:25 aleszoulek When I call this, the minions get unresponsive for a few seconds (~15?), until they get up again...
16:25 aleszoulek Any thougts?
16:25 JasonSwindle left #salt
16:25 viq aleszoulek: anything in logs/
16:26 aleszoulek Nope...
16:26 aleszoulek What I've found out is that it will just fire pillar_refresh pillar and return (the module)
16:26 aleszoulek And when catching the signal, the minion is refreshing all modules/functions
16:26 aleszoulek That's basicaly all I've got
16:27 viq How much of the pillar data do you have?
16:28 viq And how many minions?
16:28 aleszoulek A lot of pillar data (and ext_pillar as well)
16:28 aleszoulek Minions about 60
16:29 Iwirada left #salt
16:30 aleszoulek When I refresh just on one node, it get's online again very fast
16:30 aleszoulek 1sec ?
16:30 viq aleszoulek: hm, IO on the master?
16:31 aleszoulek viq: master is standard AWS large instance, it should handle much more
16:32 NotreDev joined #salt
16:33 pentabular joined #salt
16:33 aleszoulek trying calling it on bigger and bigger groups of boxes, to see, what may be wrong..
16:34 viq aleszoulek: well, kind of a workaround, maybe, but maybe batching the refresh would be an idea?
16:34 forrest can you call that module with the -l debug option aleszoulek?
16:34 viq Ah, here comes someone who actually knows what they're talking about ;)
16:34 jacksontj joined #salt
16:34 forrest Also, how does your ulimit look?
16:34 forrest viq, I wish :\
16:34 viq forrest: better than me :P
16:35 forrest lol
16:38 jrdx joined #salt
16:39 aleszoulek ulimit on master?
16:39 BenBen joined #salt
16:40 aleszoulek Turned on the debug on one of the minions to see where is it hanging...
16:40 nicksloan basepi: https://github.com/saltstack/salt/issues/8609
16:40 BenBen hi there, i've got a question regarding pillar : could it be possible to inject pillar values into non-sls files, such as .cfg file, so that my password are stored only in the pillar's sls file ?
16:41 JasonG_TA joined #salt
16:41 basepi nicksloan: ah, i see the problem
16:41 basepi salt:// references from your environment
16:41 basepi so your `files` directory needs to be under /srv/salt/salt
16:42 basepi that's why it can't find the files, because they are not in any of your file_roots
16:42 nicksloan aha
16:43 quantum-x timoguin, I created the template / grains info etc - itcompiles, but the template doesn't appear to be replaced: http://pastebin.com/MRf0Dp25
16:44 timoguin quantum-x, "templte" typo in line 6 might be why
16:44 MTecknology I wanna blame salt for this, but deep down, I know my issue is running salt on an OS that has been mutilated... http://dpaste.com/1471216/
16:44 bhosmer joined #salt
16:44 quantum-x timoguin, thanks, I appreciate the extra pair of eyes :)
16:44 UtahDave joined #salt
16:45 alunduil joined #salt
16:45 aleszoulek forrest: ulimit -n gives 32768, which should be just enought..
16:46 MTecknology i lied... looks like a bug exists
16:46 aleszoulek when running the minion in debug, I can see that it is reloading modules, and loading pem keys..
16:46 aleszoulek Nothing especialy different
16:47 BenBen I read pillar's documentation but it seems to process sls files only. So, is there any way i could inject my db connection string from pillar settings into configuration files ? I thought maybe using something like a sed module that would take a regexp as a search parameter and a new value.
16:48 forrest aleszoulek hmm.
16:48 aleszoulek Seems like it takes some time in between Loading minion keys...
16:49 whiskybar joined #salt
16:49 forrest BenBen, you can use pillar data inside of configuration files.
16:49 forrest here's an example that I'm using: https://github.com/gravyboat/hungryadmin-sls/blob/master/salt/hungryadmin/files/hungryadmin.conf
16:50 forrest aleszoulek, unless someone else has another suggestion, I think it's time to start reviewing strace :\
16:50 BenBen forrest: i used pillar["dbuser"] is some cfg file that is managed by salt, but nothing got replaced
16:50 pdayton joined #salt
16:50 viq BenBen: and did you say that the file is templated?
16:50 tempspace Morning!
16:50 BenBen viq: probably not
16:51 aleszoulek forrest: https://gist.github.com/aleszoulek/f18f1bc5d4d6a8dbefa1
16:51 BenBen viq: didn't know how you would do that
16:51 viq BenBen: sec, will find an example
16:51 aleszoulek And the gap seems to be larnger when more minions involved...
16:51 BenBen viq: thx a lot
16:51 forrest aleszoulek, how many machines were associated with that run?
16:52 forrest it only took 7 seconds
16:52 forrest hey tempspace
16:52 MTecknology I think we should move to enforcing tests... :)
16:52 viq BenBen: with that example, https://github.com/gravyboat/hungryadmin-sls/blob/master/salt/hungryadmin/app.sls - last block
16:52 viq erm, next to last
16:52 aleszoulek forrest: just 11
16:52 viq BenBen: - template: jinja
16:52 MTecknology man... that would be a bitch to write all of those at this point
16:52 forrest MTecknology, they've been working on it
16:52 forrest trying to get to 100% coverage
16:53 aleszoulek But this is just a fragment of the log. This was an obvious place it waited in when looking on tail -f
16:53 ccase joined #salt
16:53 forrest aleszoulek, gotcha
16:53 MTecknology oooh.. I see them now
16:53 aleszoulek MTecknology: Ohh... Tests... Yes. That would be nice :)
16:53 tempspace forrest: chapter one off to a good start
16:53 forrest tempsapce ?
16:54 forrest *tempspace
16:54 mr_chris joined #salt
16:55 MTecknology It seems like a lot of regression happens because of all the beginners like me. Unfortunately, that's part of the reason that salt has had the insane growth. (and because it makes it easy for the new guys to jump into)
16:55 * jcockhren blames MTecknology
16:55 MTecknology I'd be open to helping with writing tests if someone wanted to teach me the basics.
16:55 forrest aleszoulek, unless someone else has another suggestion, I'm not sure where to go from here, the debug data isn't providing enough info, my next step would be to strace the command
16:55 BenBen viq: thx, will try that right now
16:56 forrest MTecknology, might be worth talking to terminalmage about, he's been writing quite a few. Maybe he'd be willing to write up some beginner's guide in the contributing to salt docs :P
16:56 aleszoulek Hmm... That is what I was afraid of..
16:57 troyready joined #salt
16:58 forrest or you could dive into the code and see what happens during the loading of the minion key
16:58 MTecknology terminalmage: Wanna teach me to write tests and I'll document the process and then give you the docs to proof (so the clueless guy notes the confusing stuff that's simple to others), and I'll start trying to make some useful tests?
16:58 BenBen viq : seemed to work. Thx
17:01 nicksloan hmm... can pkg.installed's name argument be a list of names?
17:01 nicksloan or am I just obsessed with jamming things together that can't go together
17:01 MTecknology -pkgs:
17:01 viq nicksloan: https://github.com/viq/cm-lab-salt/blob/master/salt/roots/salt/packages.sls
17:02 blafountain joined #salt
17:02 elfixit joined #salt
17:03 blafountain is anyone around who can comment on this commit, https://github.com/saltstack/salt/commit/1bb87186a9dd1fedb5dcc6089eb3db22e5937a38#diff-a2285d74287dff5e21803f14e2323cd5R314
17:03 kermit joined #salt
17:03 nicksloan viq: of course.
17:04 nicksloan I read about that too. Just not hard enough.
17:04 blafountain this breaks the pem/pub pushing to the newly deployed cloud servers, the new minion never gets the pem generated by the master, starts up generates a new one.. then we get a conflict of pub files.... the master has the one that it generated and the minion has the one it does
17:04 blafountain so i need to delete the key on the master, then have it re-register
17:05 forrest blafountain, techhat usually does't join the IRC (unless he is hiding somewhere).
17:06 pipps_ joined #salt
17:06 moos3 I have a series of bash scripts that I wrote, is it hard to convert them to salt ?
17:06 ktenney joined #salt
17:06 terminalmage MTecknology: I'm by no means a test master, I've only written about a half dozen
17:06 forrest blafountain, I'd make an issue, then reference the pull request and mark @techhat so he sees.
17:06 forrest moos3, what do your scripts do?
17:07 blafountain forrest: okay, will due
17:07 blafountain thanks
17:07 forrest np
17:07 pentabular joined #salt
17:07 forrest he's usually fast to respond on that stuff if he isn't out of the office on business.
17:07 terminalmage MTecknology: I'm also with a client all this week
17:08 forrest oh my bad then terminalmage, I guess I was just looking at pull request where you had put tests
17:08 forrest who is considered the 'test master'?
17:08 pentabular joined #salt
17:08 terminalmage forrest: no biggie
17:08 terminalmage fishpen0: s0undt3ch is the test master
17:08 moos3 forrest stuff like this https://gist.github.com/moos3/2bed4f7e615a7b5d6402
17:10 MTecknology terminalmage: doesn't have to be this week... I just want to increase my contribution and I'm not finding anything for low hanging fruit or anything that I've feel very comfortable handling. Tests would be beneficial for me to learn and would be beneficial
17:10 terminalmage yeah
17:11 forrest moos3, most of that wouldn't be bad, some of it you'd still probably have to run as an actual command line call because I don't believe that make is available as a salt state.
17:11 moos3 forrest ok
17:12 forrest oterh than that though, you're looking at some pkgrepo.managed, file.managed, pkg.install, and git.latests
17:12 forrest *latest
17:13 MTecknology terminalmage: ping me when you'd be up for it; it'd be nice to keep regression issues from ever happening. :)
17:13 terminalmage yeah I agree :)
17:17 KyleG joined #salt
17:17 KyleG joined #salt
17:17 redondos joined #salt
17:18 pentabular left #salt
17:19 bemehow joined #salt
17:20 matanya joined #salt
17:20 elfixit joined #salt
17:22 amckinley joined #salt
17:22 dlloyd is it possible to see what sls included a specific module?
17:23 MTecknology How do you set a different pull location with git?
17:25 Linz joined #salt
17:27 pentabular joined #salt
17:32 davet joined #salt
17:34 bhosmer_ joined #salt
17:34 aleszoulek forrest: Where, I've got so far, that it is calling module_refresh which takes insane amount of time...
17:34 aleszoulek Had the same problem before with calling sync_all
17:35 aleszoulek It takes the nodes down for a minute or so
17:35 aleszoulek (In larger deployments)
17:35 BenBen Anyone knows of a way to run a script after a postgres_database.present only when the db is created ? That's for launching a sql script that would create  initial data
17:36 nicksloan do you tend to put the configuration in sites-available/ and then link it into sites_enabled/ like ubuntu wants, or do you just put the file into sites_enabled and get on with life because when is it really going to not be enabled?
17:36 dave_den joined #salt
17:37 EWDurbin nicksloan: i do conf.d/*.conf via an include in nginx.conf
17:37 forrest aleszoulek, weird
17:38 forrest BenBen, consider using cmd.wait http://docs.saltstack.com/ref/states/all/salt.states.cmd.html#should-i-use-cmd-run-or-cmd-wait
17:39 amckinley joined #salt
17:39 BenBen forrest: thx again
17:39 forrest np
17:39 BenBen forrest: told you i would come back with a lot of question :)))
17:39 forrest hah
17:39 forrest did you finish your deployment this weekend that you were working on?
17:39 BenBen nope
17:39 forrest :\
17:39 BenBen i forgot that i changed my flask configuration, i spend sunday fixing my nginx + uwsgi install
17:39 BenBen today's db day
17:40 forrest ahh ok
17:40 BenBen then tonight's going to be celery worker night. That means maybe tomorrow i'll start deploying on ec2..
17:40 forrest nice
17:40 BenBen yeah, well...we'll see :) i seriously start to lack a bit of sleeping..
17:41 forrest I haven't messed with celery before
17:41 forrest It looks cool from what I read about it though
17:42 BenBen celery in itself is cool, but you need to be really carefull when mixing it with flask. If you want to reuse the configuration files (which you surely would like to), then there a some tricks to do that and not mess around. I actually had to look at a patch in django source code to fix some issues.
17:42 aleszoulek forrest: You want to tell me, that it's normal for sync_all to finish quickly without having iresponsive nodes? I mean my friend in completely different infrastructure has similar behaviour...
17:42 BenBen then there are some tricks
17:42 brianhicks joined #salt
17:43 forrest aleszoulek, I'm not saying that
17:43 forrest aleszoulek, I don't know enough about how the backend of that works to provide you with a valid answer.
17:44 forrest BenBen, ahh ok, I haven't seen it mixed with flask, some job I applied for a while back was using it to convert documents
17:46 matanya joined #salt
17:46 BenBen forrest: actually it's an extremely good exercise. Because you need to really understand the way http requests - threads and flask context work together in order to make it work properly. Actually, now i remember what made it a bit funky. I'm also using sqlalchemy in my celery worker. So i had to create fake request context in my celery task , in order to isolate the sqlachemy context from different worker.
17:48 BenBen forrest: and the issue was in the way postgresql handled connections and transaction context (iirc). Hopefully i won't have to dig in that thing this week, but that's why i'm a bit anxious about "celery worker night" :))
17:48 forrest heh
17:51 anitak joined #salt
17:51 utahcon what is the best course of action when a salt command seems to hang?
17:51 utahcon 2 out of 3 servers completed, but #3 seems to be dead or just taking forever.
17:54 matanya joined #salt
17:54 abe_music joined #salt
17:54 forrest utahcon, I'd find the process on the minion and see what's going on with it (might need to strace that process), then if it's just hanging, kill it off with jobs.kill_job <job id>, and then from there re-run the job with debug enabled on the minion/master.
17:55 forrest also see what's happening with the logs on the minion
17:55 diegows_ joined #salt
17:55 zandy joined #salt
17:57 pipps joined #salt
17:57 zandy joined #salt
17:57 abe_music what would cause a pkg.installed in my SLS to fail with an error about not finding the package to install even though a yum list or search finds the package? i can also manually run pkg.install on the minion and it installs fine
17:58 forrest abe_music, what does your state look like?
17:58 forrest and what OS?
18:01 matanya joined #salt
18:01 _jslatts joined #salt
18:02 BraveSirRobin_ joined #salt
18:02 BraveSirRobin_ Hello!  Simple question (I hope) - how do I call a runner from a minion state file?
18:02 davet joined #salt
18:02 BraveSirRobin_ there's always cmd.run....
18:02 BraveSirRobin_ but is there a "native" way?
18:03 matanya_ joined #salt
18:04 xmltok joined #salt
18:04 forrest BraveSirRobin_, http://docs.saltstack.com/ref/runners/all/salt.runners.state.html
18:04 forrest maybe
18:05 BraveSirRobin_ other way round, I think
18:05 BraveSirRobin_ so, in init.sls I want to have something which calls the runner
18:05 aleszoulek forrest: OK. Sry. Thanks a lot anyways..
18:06 forrest aleszoulek, Yea np. I wish I had a better answer for you
18:06 danielbachhuber joined #salt
18:07 aleszoulek forrest: :) Thx, rly. I need to dig in what's going on there..
18:09 forrest aleszoulek, let me know if you figure out what it is
18:09 racooper joined #salt
18:13 austin987 joined #salt
18:13 Teknix joined #salt
18:16 BraveSirRobin_ gotacha :)
18:16 BraveSirRobin_ module.run / publish.runner
18:19 BenBen weird. I'm trying to stop an ubuntu service using the service.dead state , but it doesn't stop it..
18:21 forrest BenBen, what does your state look like for that
18:22 CheKoLyN joined #salt
18:22 pipps_ joined #salt
18:24 DerekRBN joined #salt
18:24 alunduil joined #salt
18:25 DerekRBN Hello, I'm still having a problem with the ext_job_cache feature, anyone around to help?
18:25 tty-1 joined #salt
18:28 test___ joined #salt
18:31 backjlack joined #salt
18:31 jalbretsen joined #salt
18:34 DerekRBN It seems as if the only way to fix this is to set ext_job_cache in the minon config. Is this the only way or is there somewhere i can over ride with the master?
18:34 austin987 joined #salt
18:36 matanya joined #salt
18:37 Ryan_Lane joined #salt
18:37 cjh973 joined #salt
18:38 pipps joined #salt
18:40 utahcon so I set my minion to debug
18:40 utahcon and I watch the debugging, it says Returning info for job: blah...
18:41 utahcon however the master never seems to get that.
18:42 forrest utahcon, what happens if you just test.ping from the master to the problematic minion
18:50 thelorax123 joined #salt
18:52 matanya joined #salt
18:53 cnelsonsic joined #salt
18:55 zandy joined #salt
18:57 Vivek joined #salt
18:59 bemehow joined #salt
18:59 JasonSwindle joined #salt
19:03 DerekRBN basepi i see you touched the issue im refering to where ext_job_cache is undefined in the minion config. Any idea when that will be cut into a version?
19:04 ipmb joined #salt
19:04 basepi DerekRBN: link?  i participate in a lot of issues.  =P
19:05 DerekRBN https://github.com/saltstack/salt/pull/7953
19:06 basepi DerekRBN: I think those fixes should be in 0.17.2.
19:06 basepi Though I'm not 100% sure
19:07 jacksontj joined #salt
19:08 dangerousbeans joined #salt
19:08 DerekRBN Thats what I heard at re:invent, updated my server, updated my minion still undefined. I can work around it by defining it but ive got some unknown amount of autoscaling groups i have to update. Is there anything I can do on the master to get around this?
19:11 basepi DerekRBN: ya, I just verified that it was cherry-picked.  So maybe that fix wasn't actually the fix?
19:11 basepi `ext_job_cache` should always be defined now, because we put it in the DEFAULTS dict in config.py
19:12 DerekRBN Let me see if my minion is 100% up to date because i use the bootstrap
19:12 DerekRBN I just used pip to install 17.2 but im not sure if that over wrote my minion
19:13 basepi It should.....
19:13 utahcon forrest: test.ping works fine
19:13 basepi Unless you have some weirdness in your install
19:14 DerekRBN Thats what im looking into. Does the bootstrap get updated with every version cut?
19:14 basepi with no additional args, bootstrap will install using system packages by default
19:14 basepi or you can use args to override to a specific version or branch from git
19:15 basepi so it's possible it installed from system packages which haven't been updated with 0.17.2 yet (depending on your system)
19:16 DerekRBN interesting looks like im using yum and not the bootstrap. sorry about the confusion. Let me check yum, im on amazon linux
19:17 nahamu DerekRBN: pip installation can screw up and pull in the wrong version of pyzmq https://github.com/saltstack/salt/issues/8390
19:17 DerekRBN Nahamu, i just found that out.
19:18 ajw0100 joined #salt
19:18 DerekRBN pkg_resources.DistributionNotFound: pyzmq>=2.1.9
19:18 matanya joined #salt
19:20 sfz joined #salt
19:21 patrek does file.replace in a state support pcre expressions like \s* (to capture whitespaces) ? And if so, what would be the exact syntax?
19:22 bhosmer_ joined #salt
19:25 bemehow joined #salt
19:28 wramthun joined #salt
19:29 basepi DerekRBN: 0.17.2 isn't on epel-testing yet, unfortunately.  but you can use the bootstrap script to install from git with the tag `v0.17.2`
19:30 DerekRBN Ok, just started a new node and epel didnt have it still on 17.1. do we have a timeline when that might go though? If not I've got a lot of work to do to lock down version. I should probably do it anyways to prevent issues like these
19:31 japage joined #salt
19:31 basepi DerekRBN: should make testing in the next day or two (terminalmage will hopefully have time to get the packages going later today, though they can still take 24 hours to hit the mirrors)
19:31 basepi then two weeks to EPEL proper
19:33 DerekRBN Alright thanks for all the info. I'm going to go back through and lock down all my auto scaling groups to 0.17.2 and update them as needed after having tested in my own envrionment
19:34 bemehow joined #salt
19:35 terminalmage DerekRBN: you can install 0.17.2 using salt-bootstrap if you really need it. but depending on your platform, it may not pull in the right init scripts
19:35 terminalmage I think it installs the upstart conf files but not sysvinit scripts
19:35 terminalmage another downside is that it installs with setup.py and no system packages
19:36 terminalmage so it's not remove-friendly
19:36 basepi DerekRBN: the other possibility would be to drop the 0.17.2 version of config.py into your salt install location
19:37 basepi DerekRBN: that would workaround that specific issue.
19:38 DerekRBN That might work. I would prefer to use yum to manage my packages. the config.py work around might get me through the next 2 weeks and then i can lock down my yum install to 0.17.2
19:39 basepi right.  that's the advantage of just monkey-patching that specific file is that yum is none the wiser, so it doesn't mess anything up in the future.
19:39 basepi but it is a more manual process, which is a negative
19:40 DerekRBN Being that these are autoscaling i have to come up with a weird solution to do any of the above :p
19:40 xet7 joined #salt
19:40 BenBen joined #salt
19:40 pipps joined #salt
19:41 basepi ya.  =\
19:42 basepi i mean, by later this week it will be on epel-testing and that will be a much easier workaround
19:44 * Gareth waves
19:48 tty-1_ joined #salt
19:48 nicksloan here's the problem: supervisor comes with an (awful) init script on Ubuntu 12.04. I have an upstart job that works much better, but alas, apt starts supervisor on installation (which is an absolutely maddeningly stupid thing to do, IMO)
19:49 matanya joined #salt
19:50 nicksloan so the question is, how do I install supervisor, stop it with the init script, remove the init script, add my upstart job, and start the upstart job with salt, without my formulas becoming stupid?
19:51 teskew who is in charge of ubuntu builds?  any idea when .17.2 will be built and in the repo?
19:53 DerekRBN nicksloan: have you tried installing it with pip? I believe most of my salt configs install it with pip and then later call it's service to ensure it's running tho it might again run automatically upon install
19:54 * Gareth goes to registere "all-salt-packages-available-yet.com"
19:54 terminalmage teskew: Corey does ubuntu, I believe
19:55 zandy joined #salt
19:56 guest101 joined #salt
19:57 DerekRBN @basepi thanks for you help with this, the config.py did the trick. I'm going to upload the 0.17.2 to s3 or something alike and pull it down in my userdata. When 0.17.2 is available on epel i'll lock down all my versions. Moving forward I cant risk running different versions across nodes
19:57 [diecast] joined #salt
19:57 bt joined #salt
19:57 bemehow joined #salt
19:58 Corey terminalmage: I do!
19:58 Corey teskew: It's built now. I'm just waiting on the release announcement to take it live. :-)
19:58 terminalmage Corey: yeah I didn't know if you shared time or were the sole maintainer
19:59 grep_awesome joined #salt
19:59 nicksloan DerekRBN: I suppose I could install it from pip. I'd guess that it won't run on startup from pip, unless the supervisor guys are just spiteful jerks.
19:59 Corey Credit where due, joehh helped a lot this time.
19:59 guest101 Hello Salters. I've had some modest initial success playing with pydsl, creating a state object ('mystate') and setting Salt states  with syntax such as 'mystate.sysctl.present(name=KEY)'.
19:59 guest101 However, for a general purpose script the Salt module/function name needs to be specified as a variable, something like 'mystate.SALTMODULE(name=KEY)'.
19:59 guest101 Can someone please tell me how I can implement that? The syntax or the special sauce.
20:00 bt anyone have recommendations on creating a "shadow" state, where you might have a "widget_prod" state and would like a "widget_dev" state that only has minimimal differences from widget_prod?
20:02 carlos joined #salt
20:02 matanya joined #salt
20:02 teskew thanks, Corey :)
20:03 Corey teskew: No worries. Blame utahdave for the lack of release announcement.
20:03 teskew will do.  what a slacker
20:04 dwyerj joined #salt
20:04 moos3 anyone have a good way to deploy ssh keys ?
20:04 Gareth Corey: you should do an original performance of "Blame UtahDave" to the tune of "Blame Canada".
20:05 Corey Gareth: That's actually my SaltConf presentation.
20:05 tempspace lol
20:05 Corey moos3: There's a provider for that.
20:06 aserdp_ joined #salt
20:06 IJNX joined #salt
20:08 DerekRBN Corey: i cant wait to see it
20:08 DerekRBN Also I hold you to it
20:08 Corey I've also shoved it into Homebrew for mac.
20:10 moos3 Corey you have a link on it
20:10 Corey moos3: Just so http://docs.saltstack.com/ref/states/all/salt.states.ssh_auth.html
20:11 moos3 Corey that looks like its only public keys and not private keys
20:11 Corey If you want to deploy private keys, first off "don't," and secondly if you must, store them in pillar.
20:11 moos3 yeah I have a pillar for making a user, but I can't figure out how to make pillar run
20:12 moos3 on a node
20:13 forrest utahcon, oh sorry I went to lunch, did you figure that out?
20:14 forrest Corey, did you guys get the raring issues figured out?
20:14 Corey SOrt of.
20:14 Corey "It compiles, ship it!
20:14 Corey "
20:14 forrest lol
20:15 pkimber joined #salt
20:17 linjan joined #salt
20:17 Gareth "I'm Corey and I stand beside these packages....not behind....beside."
20:18 Corey "Shoot them, not me!"
20:19 nahamu is there a way to watch events happening from the CLI (kind of like how halite can show them)?
20:21 forrest can you not run the log in debug mode nahamu? I don't have a salt instance at work or else I'd just check myself
20:23 mwmnj joined #salt
20:24 Teknix joined #salt
20:24 nahamu I found this: http://docs.saltstack.com/topics/event/index.html
20:25 nahamu 4 lines of python isn't too bad.
20:25 forrest oh you're going off this? I thought you were talking about http://docs.saltstack.com/ref/cli/
20:25 nahamu I was hoping for a CLI tool
20:25 gkze joined #salt
20:26 nahamu but I can write one in 4 lines of python so I don't need a prepackaged one.
20:26 gkze is there a way to add custom global dynamic grains?
20:26 gkze like load average/disk stats/etc
20:27 apergos joined #salt
20:27 Ryan_Lane when using multiple masters, how does the master_finger config work?
20:28 Ryan_Lane utahcon: ^^
20:28 Ryan_Lane heh
20:28 Ryan_Lane well, at a conference, that should make getting an answer harder :)
20:30 Pikachu320 joined #salt
20:31 dvogt joined #salt
20:31 dave_den Ryan_Lane: the same key is used by all masters
20:31 troyready joined #salt
20:31 dave_den so the fingerprint will be the same
20:31 Ryan_Lane the same master key is used??
20:32 dave_den Ryan_Lane: http://docs.saltstack.com/topics/tutorials/multimaster.html
20:32 Ryan_Lane huh. interesting. I keep forgetting this isn't x509
20:32 apergos nope it's not
20:33 dave_den gkze: grains are meant to contain static data that doesn't change often
20:33 gkze how would I collect that kind of stuff with salt then?
20:33 ajw0100 joined #salt
20:34 dave_den well, within salt itself you could report that data to the mine, but the mine only keeps the most recent data
20:34 dave_den are you looking to record it and graph it for trends?
20:35 ajw0100 joined #salt
20:37 blee joined #salt
20:39 jefferai joined #salt
20:39 dave_den plop plop
20:39 dave_den fizz fizz
20:39 forrest ?
20:39 dave_den oh what a relief it is?
20:39 forrest I'm just curious about the context
20:39 nahamu indigestion?
20:40 nahamu (the context, not the alka-seltzer)
20:40 dave_den - context: taco bell
20:40 forrest is that in reference to graphing trends with salt?
20:40 dave_den :P
20:40 forrest oh you had taco bell?
20:40 nahamu haha
20:41 forrest Spicy food doesn't bother me, probably helps that I eat it 4-5 days a week (granted taco bell isn't really spicy)...
20:41 dave_den gkze: when you come back, you're probably looking for this: http://docs.saltstack.com/topics/jobs/schedule.html#scheduler-with-returner
20:41 gkze sorry
20:41 dave_den when i occasionally do hit the bell, i put about 5 fire sauces on a single taco
20:41 dave_den that's what gets me
20:41 forrest lol
20:41 forrest that's why you just buy their hot sauce in the bottle
20:42 gkze yes I was looking for a time series thing for graphing that data
20:42 forrest only thing from there that I like is their hot sauce (and it does't have any HFCS)
20:42 gkze kinds of an alternative to collectd/statsd
20:43 jacob_s joined #salt
20:43 dave_den chipotle + cholula = nom
20:43 KyleG1 joined #salt
20:43 mpanetta Quick question: When you do an include: in a state formula, does it run the included state immediately?
20:43 forrest mpanetta, according to what UtahDave said, the included state is supposed to run first
20:44 jacob_s I'm having an issue with the npm state. When salt checks if an npm package is installed it runs
20:44 jacob_s npm list --json --global "grunt"
20:44 mpanetta forrest: If I do some stuff and then include some states after, it will run first then?
20:44 mpanetta Hrm...
20:44 forrest dave_den, yea I use chipotle for lunch when it's gonna be a heavy workout night, you can get a bowl and the tortilla on the side, then double rice, double beans, and half and half meat (which is usually more than you'd get) for the same price as a normal bowl
20:44 jacob_s If this is not installed npm returns {} with an exit code of 1. Which causes the state to fail immediately.
20:44 dave_den mpanetta:  the included state files get included after the initial top files are walked.
20:45 forrest mpanetta, I don't think I've ever seen someone use include part of the way through a state.
20:45 mpanetta I have a start.sls and a stop.sls and I want to include them in an upgrade.sls to be run after the upgrade...
20:45 mpanetta forrest: I guess I am being spechul then? ;)
20:45 JasonSwindle left #salt
20:45 mpanetta I'm still quite the noob with salt.
20:45 mpanetta And our guru is out today.
20:45 forrest why would you not have the start.sls include the upgrade.sls, then the stop.sls include the start.sls?
20:46 forrest so it goes upgrade -> start -> stop
20:46 dave_den mpanetta:  you can either use overstate or configure 'requires/watches' to execute them when needed
20:46 forrest yea I'd consider using requires for tihs as well
20:46 mpanetta er I want upgrade, stop, start tho.
20:46 forrest to force the order
20:46 mpanetta Ok so I can include *and* use watch?
20:46 forrest sure
20:47 mpanetta er rather in this case, requires I suppose.
20:47 forrest mpanetta, https://github.com/gravyboat/hungryadmin-sls/blob/master/salt/hungryadmin/app.sls
20:47 forrest there I include a bunch of stuff, then require things like the pkg of virtualenv from python.virtualenv no problem
20:47 mpanetta The stop.sls and the start.sls were never really meant to be run inside another state the way they were written I think.
20:47 forrest yea it doesn't sound like they were
20:48 forrest why don't you just reload the service based on when the upgrade stuff goes?
20:48 forrest the file gets modified or whatever.
20:49 forrest Like this: https://github.com/gravyboat/hungryadmin-sls/blob/master/salt/ssh/init.sls
20:50 forrest or maybe you guys are already doing that, I can't remember, who is the other guy that you work with mpanetta?
20:50 mpanetta oz_akan
20:50 forrest ahh right right
20:51 mpanetta Oh, this piece of software is not managed as a service (it should be IMO, but isn't, yet.)
20:51 forrest ahh ok
20:51 KyleG joined #salt
20:51 KyleG joined #salt
20:53 mpanetta I'm just going to dup code until I get a chance to refactor all these sls files.
20:54 forrest yea as long as you keep them apart that's an ok workaround.
20:54 mpanetta Cool, thanks forrest
20:54 forrest :P I didn't do much
20:54 mpanetta You guys are always a good help :)
20:55 forrest except say 'stop doing things that way!'
20:55 mpanetta forrest: Sometimes that is all one needs to do to get someone going the right direction :)
20:55 forrest hah
20:55 forrest Yea that's why I have a rubber duck on my desk at home
20:55 zandy joined #salt
20:55 Ryan_Lane joined #salt
20:55 mpanetta You ask it questions? heh
20:55 forrest Yea
20:55 mpanetta I know someone that does something similar
20:56 mpanetta Maybe I should try it
20:56 mpanetta heh
20:56 forrest It works well, talking it through outloud is really helpful
20:56 forrest if that doesn't do it, I take a pad of paper and a pen then go for a walk
20:56 forrest I mean I look like a crazy person walking around the block talking to myself scribbling down garbage
20:56 forrest but it helps
20:57 mpanetta makes sense
20:58 forrest Yea I learned about the walking thing from a coworker at my old job, he'd walk around outside our whole building.
20:59 NotreDev joined #salt
21:00 forrest Uh Oh, NotreDev is back :P
21:00 NotreDev hah
21:00 forrest no rage on monday
21:01 pipps_ joined #salt
21:02 cewood joined #salt
21:03 jacob_s Okay my npm issue was fixed a few days ago https://github.com/saltstack/salt/issues/8517.
21:05 cewood joined #salt
21:06 oz_akan_ joined #salt
21:09 mrevans joined #salt
21:09 pdayton joined #salt
21:10 bhosmer_ joined #salt
21:11 sroegner_ joined #salt
21:12 pipps joined #salt
21:19 bhosmer_ joined #salt
21:21 xmltok joined #salt
21:23 funzo joined #salt
21:23 Pikachu320 hello
21:24 Pikachu320 I'm looking for a clean way to do something like this http://dev.mlsdigital.net/posts/SaltStackBeyondJinjaStates/
21:24 Pikachu320 (end of the page), but with users in different groups depending on the hosts
21:25 Pikachu320 to be more clear: I'll have a pillar with my "global" users, another with my "provider1" users
21:27 Pikachu320 and some hosts will have "global" users in a group called "sst", while some other hosts will have "global" + "provider1" users in a group called "rpu", any idea of a clean way to do it?
21:30 Pikachu320 I'm kinda noob in python and salt for now, so I don't know the possible if 'host group' syntax
21:30 Pikachu320 (if it's a solution)
21:33 forrest Pikachu320, you could always use jinja, so you have your pillar that gets compiled, with your global set of users, and your providers1 set of users, then depending on the value of certain items, you loop through a group
21:35 jslatts joined #salt
21:35 forrest jesusaurus is doing something that MIGHT work as a starting point for you Pikachu320, where he has this pillar: https://github.com/jesusaurus/hpcs-salt-state/blob/master/apt/sample.pillar
21:35 forrest and then he runs through the distros here: https://github.com/jesusaurus/hpcs-salt-state/blob/master/apt/repo/distro
21:35 forrest granted depending on how you configure your users it might be different
21:36 Guest42076 w
21:36 Pikachu320 I see
21:38 lineman60 joined #salt
21:38 andrej joined #salt
21:41 lyska joined #salt
21:46 godber joined #salt
21:47 joshuamckenty joined #salt
21:47 godber Hi, is it legit to provide a list of files to watch like this ? https://github.com/godber/publicsalt/blob/master/salt/roots/rabbitmq/init.sls#L36
21:48 godber Using salt 0.17.1 ... I swear it worked in my test vagrant vm
21:51 ajw0100 joined #salt
21:51 DerekRBN godber: those look just fine. are you getting an error or is it just not doing the watch?
21:52 forrest godber, it looks fine to me
21:52 forrest what error are you getting?
21:53 forrest godber, maybe try a watch_in: within the actual file.managed sections?
21:53 dangerousbeans joined #salt
21:54 godber http://pastebin.com/d79EjVg1
21:54 godber those are the errors
21:54 godber oh the server is 17.0 BTW
21:55 zandy joined #salt
21:55 godber at least I think it was complaining about those file declarations (I'm not too good at grokking salt errors yet)
21:56 godber oh I think I am misreading the errors
21:57 forrest 0.17.0? Damn godber, playing with fire
21:58 forrest is this in develop where you can safely upgrade to 0.17.1?
21:58 forrest hmm, so it's actually complaining about your managed files having unhashable types?
21:58 forrest why is that...
21:59 godber maybe I should sort out my bad version, let me read the changelog for 0.17.1
21:59 forrest I'm curious as to whether it has to do with the grain reference
21:59 forrest can you try to remove the grains['fqdn'] from the source godber?
21:59 forrest one one of them
22:00 forrest and then run it on a test machine
22:02 godber forrest: I changed one to this:
22:02 godber http://pastebin.com/ePExiWsh
22:02 godber still no go
22:02 godber same unhashable
22:03 forrest Hmm, I'm not quite sure then godber, maybe take a look at the lines it's complaining about in that error within the saltstack code to see if there is a hint?
22:04 jimallman joined #salt
22:04 godber well, I will upgrade my master, since thats the one difference
22:04 pipps_ joined #salt
22:05 DerekRBN I'm not 100% on this but it appears as you're giving it a list of source files.
22:05 forrest godber, what version are your minions before you do that?
22:06 godber forrest: :) About 20% of them are now dead
22:06 DerekRBN it's failing on the source files not the watch.
22:06 amckinley joined #salt
22:06 godber yeah, its failing on the source
22:06 DerekRBN what are you trying to do with those source files?
22:07 DerekRBN if it has a fqdn then give it this file if not give it the non fqdn?
22:07 godber I had been using that as a brute force way of overriding configurations
22:07 godber if the file with the fqdn is not there ... it should fall through and use the default
22:08 DerekRBN hmm didnt know we could do that.
22:10 godber DerekRBN: even with those removed, I get the same unhashable error
22:11 Brew joined #salt
22:12 godber DerekRBN: See this URL, http://docs.saltstack.com/ref/states/all/salt.states.file.html#module-salt.states.file
22:12 godber "The source parameter can be scpeified as a list ..."
22:13 godber and theres my problem
22:13 godber there should be no "file: "
22:13 DerekRBN ahh i see
22:13 forrest for the watch?
22:13 godber for the source:
22:13 sroegner_ joined #salt
22:14 godber oh wait no, I was confused
22:14 forrest Are you talking about in the file? https://github.com/godber/publicsalt/blob/master/salt/roots/rabbitmq/init.sls#L41
22:14 godber I don't have it in the file
22:14 forrest I don't see file: there
22:14 godber forrest: you're right, I was looking at the wrong line
22:14 forrest oh ok
22:14 forrest thought I was going crazy
22:14 shinylasers joined #salt
22:15 godber forrest: are you in PHX?
22:16 forrest yea
22:16 forrest I was curious as to whether you were Austin or not, lol
22:17 godber yeah :)
22:17 moos3 when i do salt 'minionname' saltutil.refresh_pillar it returns none
22:17 moos3 ideas on this
22:18 forrest moos3, can you try with -l debug?
22:18 forrest or check the log to see if anything is happening, as well as the active jobs?
22:18 moos3 https://gist.github.com/moos3/3bdd5d43f765e5d54e4e
22:20 forrest moos3, you confirmed it doesn't have the latest pillar data already?
22:20 moos3 not sure how to check that
22:20 ipmb_ joined #salt
22:21 forrest moos3, you have a state that is relying on this to work right
22:21 forrest *?
22:21 forrest on your pillar data being updated
22:21 moos3 nothing, just setup user data
22:22 godber Ok, I now at least understand the unhashable error .... thats what you get when NONE of the items listed in the source list are there
22:22 godber aha!
22:23 forrest ahh ok
22:23 forrest so even your default files weren't there?
22:23 godber correct
22:23 godber which was me being dumb
22:23 forrest lol
22:23 godber sorry for bothering you all :-//
22:23 moos3 forrest i'm trying to get this to work http://serverfault.com/questions/538218/deploy-ssh-key-from-master-to-minion-via-salt-pillars
22:23 forrest all good
22:24 godber now, if anyone can elaborate on the warning at the top of this page:
22:24 godber http://docs.saltstack.com/topics/releases/0.17.1.html
22:25 godber what happens if I don't update my minions
22:25 forrest if your minions are running something other than 0.17+ they will not connect
22:25 godber I have some CentOS boxes running older (pre 17) minions
22:25 forrest ok then don't update.
22:26 godber well, I have, but I can still contact them
22:26 forrest That seems... strange
22:26 forrest some stuff might not work, I don't have the full list
22:26 forrest but there was a protocol change between 0.16 and 0.17 for security purposes.
22:27 godber yeah
22:29 forrest That's why I was asking if this was a dev environment you could safely mess with, umm best bet is probably to go back if you can.
22:30 godber No, its fine, I was aware that there were consequences
22:30 forrest oh ok cool
22:30 forrest Was worried there for a second
22:31 jbirdman joined #salt
22:31 Teknix joined #salt
22:32 BenBen joined #salt
22:37 godber so, should salt say something to the effect of "No files found" rather than "Unhashable" error?
22:40 forrest It probably should, but right now it does not
22:41 forrest There's been one guy who has been working on making the error messages more friendly, and I believe he's just chugging along
22:41 forrest I don't remember who though
22:42 Cidan Does anyone know if there is a way to make salt-cloud assign a random public ip address to a VPC instance on create?
22:42 Cidan (EC2)
22:43 forrest godber, if you want to file that as an issue it might be considered low hanging fruit and get fixed
22:43 godber ok will do
22:43 forrest cool
22:45 pipps_ joined #salt
22:45 jslatts joined #salt
22:46 snuffeluffegus joined #salt
22:47 Cidan I'm going to take that as a no, tsch..
22:48 forrest Cidan, yea sorry man, I saw the question, but I don't have a solution. Haven't messed with salt-cloud much
22:48 Cidan hm hm hm
22:48 dwyerj joined #salt
22:49 dwyerj hey everyone, im about to start deployment on ~50 servers.  is salt a good choice for this? previously i had some bash scripts for deployment.  but only had around 5 servers
22:49 Cidan dwyerj: I'll be 100% honest with you, I love salt, but it breaks, a lot.
22:50 forrest dwyerj, it could be, can you provide a bit more information regarding your needs, is your code mostly ruby or python, etc.
22:50 dwyerj thanks for the honesty!
22:50 forrest and Cidan is correct, Salt is awesome, but still heavily developed some sometimes stuff does break :D
22:50 forrest *so
22:50 dwyerj i code in python, so i am biased. but i was asked by someone else to do this deployment and they have code mostly in php (and its pretty rubbish)
22:51 dwyerj I saw that 0.17 now has the API released, is that up for much change?
22:51 dwyerj they want to be able to communicate with all 50 servers after deployment also, i was originally going to use salt and extend halite
22:52 dwyerj but they just want to do deployment with salt, then i think do like salt "*" cmd.run 'ping localhost' (for example)
22:52 forrest what do they want communication wise?
22:52 dave_den to be fair to salt, most of the stuff that breaks is when upgrading to a new release of salt, and that is mostly due to the rapid and large code cycle
22:52 forrest oh yea, for initial deployment/continued deployment salt, as well as most other config management stuff is nice.
22:52 forrest have you used a configuration management tool before?
22:53 forrest dave_den, yea I think that's a given, it's still very young (but has a lot of cool stuff)
22:53 forrest I'll take the breaking stuff over puppet any day
22:53 dwyerj not really, ive played with ansible, breifly with salt and kickstart/anaconda
22:53 forrest ahh ok
22:53 dwyerj mostly just used bash scripts for deployment, which i find get far too tedious
22:53 forrest yea
22:54 NotreDev joined #salt
22:54 forrest I mean I'd use Salt, but clearly, bias.
22:54 dwyerj yeah but thats fine, as long as you would recommend it
22:54 forrest It's nice to be able to spin up a master, and get it hooked to your minions in about 5 minutes if you know what to do
22:54 forrest well, what OS are you running ?
22:54 dwyerj centos 6
22:54 forrest ok sweet, so you have python2.6 already
22:54 forrest that's good
22:54 forrest salt only runs on 2.6+
22:55 heewa dwyerj: I've used manual bash before, then moved to puppet, then moved to salt. Massive improvement each time. But warning: puppet/chef/salt/etc are more work in the short term, much much less in the longterm.
22:55 heewa So stick with it, through the innevitable frustrations. There's a light at the end of the tunnel.
22:55 dwyerj thats ok, im getting paid by the hour :)
22:55 forrest lol
22:55 heewa lol
22:55 moos3 hrm anyone have a good way of deploying ssh priv keys ?
22:55 zandy joined #salt
22:55 forrest honestly dwyerj, I'd say spin up a master/minion combo on digitalocean or something, and play with it
22:55 forrest see if you like it
22:56 dwyerj ive spun it up befre, i think when it was around >0.15 and i liked it at the time
22:56 forrest moos3, I don't, maybe you could adapt UtahDave's user stuff? https://gist.github.com/UtahDave/3785738
22:56 forrest ahh yea, it's only gotten better since then
22:56 dwyerj was super fast getting it going, so ive always been tempted to give it a crack. but never needed to until now
22:56 heewa moos3: I know this isn't your question (I hate when people do this), but is it possible not to? Can you generate them using salt on the minions, and get the public key back? Instead of generating them centralling and distributing?
22:56 forrest yea that's my biggest thing, Salt is awesome to get up and running
22:56 dwyerj what about communbication with 50+ servers
22:57 dwyerj after ive deployed them, is it viable ?
22:57 forrest dwyerj, no problem
22:57 dwyerj sweet
22:57 moos3 I have about 500 boxes that need to be able to slide into with ssh keys for root, because of how gpfs works
22:57 heewa dwyerj: You might need a simple script to initially install salt on the boxes, but that could be a simple ssh one-liner.
22:57 forrest heewa, he could also use the bootstrap
22:58 dwyerj i was going to make my own bootstrap and just have it on like gist.github (kinda like theres)
22:58 heewa forrest: Right, ssh to each box to curl & run the bootstrapper.
22:58 forrest dwyerj, check out this vid: http://www.youtube.com/watch?v=uWGDC1PdySQ&amp;feature=c4-overview&amp;list=UUpveTIucFx9ljGelW63-BWg That's UtahDave and Tom using salt-ssh (the slower version you don't install) to build a 100 node riak cluster
22:58 Cidan moos3: hey moose, :D
22:58 moos3 hey man
22:58 dwyerj thanks, will watch it now
22:59 heewa moos3: For getting into your boxes, you'd want to distribute your public key to them. You don't want their private key. In fact, they don't really even need any private keys of their own.
22:59 forrest cool
22:59 forrest keep in mind that's salt-ssh, so it's a lot more resource intensive, and slower
22:59 forrest Are any of the LinkedIn guys here?
22:59 forrest or someone from the rackspace team using Salt
22:59 dwyerj yeah alright, is regular salt secure? i remember there was a CVE released.  and last time i used salt there was only talk of salt-ssh
23:00 forrest so I mean there are obvious issues if your master gets compromised
23:00 Corey dwyerj: That's a loaded question.
23:00 forrest lol
23:00 moos3 heewa true, but when the boxes have to be able to connect back to the masters on a GPFS cluster they need to have a private thats on all the nodes, so its just easy to distrubute the same ssh private key to all the nodes
23:00 Corey But yeah, I'd not hesitate to use salt.
23:00 Corey I wouldn't have it speak across the internet, but neither would I with any other config management solution.
23:00 Corey That includes via SSH.
23:00 forrest dwyerj, the only people I've seen complain were guys working with government agencies, and they ended up using salt-ssh, still in a controlled environment of course.
23:01 dwyerj ah alright, well it would be live on the net.
23:01 heewa moos3: I see. One method I've used in the past is to reverse-tunnel to the remote box, but that requires initiating the action from the central box, not the edge nodes.
23:01 dvogt joined #salt
23:01 forrest dwyerj, are you talking within Amazon or something?
23:01 dwyerj leaseweb
23:01 moos3 my stuff isn't in the cloud :D they are in a private dc :D
23:02 heewa moos3: Would it be bad to keep the private key on salt master & just put it on the minions with a file.managed directive?
23:02 forrest Yea I don't know if anyone is doing that with a production environment where it isn't part of a 'private network' within AWS or somethign along those lines. I mean I have some boxes on digitalocean, but that's just stuff I spin up.
23:02 kermit joined #salt
23:02 dwyerj though, i believe that the added requirments of salt-ssh wouldnt be a problem.   can i develop it as plain salt, then easily migrate to salt-ssh if i want ?
23:03 forrest the code you write for salt-ssh works identically to that of normal salt
23:03 moos3 heewa i could
23:03 forrest so you create your states like normal
23:03 dwyerj yeah too easy then, ill just dev with salt then switch to salt-ssh
23:04 dwyerj it doesnt need to be super high avaliability. currently boxes go down all the time. and its like 10min cronjobs to get responses (also they are just plain POST requests anyone can see(
23:04 forrest yea, unless someone can show me evidence to the contrary, I don't think there is anything unique you'll have to configure for salt-ssh to get everything working (I didn't)
23:04 forrest lol
23:04 dwyerj excellent, im glad to hear all this
23:05 blee_ joined #salt
23:05 pentabular joined #salt
23:05 forrest yea it's pretty great
23:07 Gareth for vim users.  this is awesome: https://github.com/scrooloose/syntastic
23:08 jcockhren Gareth: yep
23:08 jcockhren Gareth: if doing python, and writing tests check out sniffer
23:09 Gareth will do.
23:09 jcockhren autoruns any code on file changes
23:09 Gareth very cool.
23:13 Gifflen joined #salt
23:14 mwillhite joined #salt
23:15 mwillhite is there any sort of try/catch for salt? I'd like it to not stop if something goes wrong…
23:15 forrest not that I've ever seen or tried to jerry rig mwillhite
23:25 mwillhite hm ok…I'm using salt/vagrant to setup a dev environment with a dozen or so apps…I don't really want it to blow up if an app has problems
23:25 forrest well, each item should be self contained
23:25 forrest I mean the run will continue onwards for other states, unless dependencies you have defined fail
23:26 forrest but if they actually spit back an error, then yea the run can bomb out
23:26 mwillhite hm, yeah maybe I just need to organize better
23:26 forrest https://github.com/terminalmage/djangocon2013-sls
23:26 mwillhite but for example…if I don't have permission for a particular git repo and I do a pull…it stops
23:26 forrest that's a pretty good example
23:26 forrest when you use the git.latest state?
23:27 mwillhite yes
23:27 mwillhite or…need to double check that
23:27 mwillhite I set this up a couple months ago, just now coming back to it
23:28 forrest gotcha
23:29 mwillhite actually…it's from a bundle command
23:29 mwillhite I'm requesting a gem from a private repo
23:29 mwillhite and I hadn't setup my keys correctly, so it blew up on that
23:29 mwillhite in that case I supposed its with the cmd.run state
23:30 vespakoen joined #salt
23:30 Cidan woohoo
23:31 Cidan obscure undocumented options for the win
23:31 Cidan jesus.
23:31 forrest for what Cidan?
23:32 forrest and you should update the docs and make a pull request to fix that :P
23:32 mwillhite forrest or for example a gem that fails to install brings the whole thing to a stop
23:32 Cidan setting up an instance in a VPC and having it obtain a public IP
23:32 mwillhite whats the whole for cmd.run…if it exists with something other than 0 we get stop the whole thing?
23:33 forrest mwillhite, I'm not sure, when there is an actual run breaking error, it will fail, so if the cmd.run returns an error, Salt probably sees that and stops the run?
23:33 forrest because I know I've had states where certain items have failed (git.latest for example), and it continues onwards
23:33 forrest Cidan, oh nice, what was the solution to that anyways?
23:34 Cidan in your profile configuration for a machine type
23:34 Cidan add
23:34 Cidan network_interfaces:
23:34 Cidan - {DeviceIndex: 0, SubnetId: subnet-071a1165, SecurityGroupId: [sg-03cadd61, sg-1a061178], AssociatePublicIpAddress: True}
23:34 forrest uhh forgive my ignorance, what are the values?
23:34 forrest the sg-03cadd61?
23:34 alunduil joined #salt
23:35 Cidan EC2 VPC security group id's
23:35 forrest ahh ok
23:35 vespakoen Hey guys, I want to use the mount state, and in the "opts" arguments, want to pass in a uid and gid, but to figure those out, I would like to run a command, so I am guessing I need some jinja2 - run a command and use the output of it thingy, does anyone know if there is something like this?
23:35 forrest vespakoen, take a look at the jinja2 docs: http://jinja.pocoo.org/docs/templates/#assignments
23:36 forrest if you don't want to make the pull request Cidan, can you at least open an issue and paste your solution in there? Then I'll try to get to it.
23:36 vespakoen great! Thanks a lot
23:36 forrest vespakoen, yea np, might take a bit of modification, but you should be able to do it
23:36 forrest vespakoen, it might make more sense however to try and base uid/gid either off of pillar data, or grain data.
23:37 dvogt joined #salt
23:38 forrest vespakoen, what command are you trying to run to base this info off of?
23:39 vespakoen id -u "user"   / id -g "group"
23:40 forrest ok, so why not create your users where they have a default uid across your environment? Then what you can do is have that data available, and just set it the same for everything?
23:40 forrest *.
23:40 dangerousbeans_ joined #salt
23:40 vespakoen this is the command I would normally run: mount -t vboxsf -o uid=`id -u azion`,gid=`id -g www-data` /home/azion/ivizi-app /home/azion/ivizi-app   but I am trying to use mount.mounted since it has support for persisting the rule in fstab
23:41 forrest so I'd probably user a bunch of user.present lines (like this example https://gist.github.com/UtahDave/3785738), then you can just pull that data right out, no issue.
23:41 vespakoen I could probably just execute the command and add it to fstab myself with another command if going this route is to hard / stupid
23:41 forrest It seems overly complicated to me
23:41 forrest when you can set the uid when the user is created, and simply reference that data.
23:41 vespakoen so basically, predefine the UID / GID ?
23:41 forrest yea
23:42 forrest like in that example I linked
23:42 vespakoen I see, well that seems like a better way since salt seems to be built around the id's instead of the names
23:42 forrest so you know user 'vespakoen' will always have uid 5021 or whatever.
23:42 vespakoen I see, will do it that way, thanks for the help!
23:42 forrest yea np!
23:43 vespakoen oh, one more thing, do you perhaps know what "- device:"  should be in the mount.mounted properties for my command (mount -t vboxsf -o uid=`id -u azion`,gid=`id -g www-data` /home/azion/ivizi-app /home/azion/ivizi-app) it's what vagrant uses
23:47 garron joined #salt
23:54 ajw0100 joined #salt
23:55 zandy joined #salt

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