Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-08-24

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

All times shown according to UTC.

Time Nick Message
00:04 overyander joined #salt
00:29 jeadre joined #salt
00:30 dopesong joined #salt
00:36 moss joined #salt
00:55 pravka joined #salt
01:03 jeadre joined #salt
01:05 bbbryson joined #salt
01:25 dthom91 joined #salt
01:31 sk_0 im getting Pillar failed to render with the following messages: Specified SLS 'rd' in environment 'stg' is not available on the salt master
01:32 sk_0 i've moved everything out of /srv/pillar/stg and just made a simple list. then pulled everything out of /srv/salt/stg and just made a simple rule to get an item from this list and i still get this error
01:32 sk_0 this is driving me mad. no where in /srv/pillar/stg is anything call 'rd'
01:54 jaybocc2 joined #salt
02:06 favadi joined #salt
02:32 evle joined #salt
02:32 UForgotten joined #salt
02:38 zwi joined #salt
03:00 timoguin joined #salt
03:02 jeadre joined #salt
03:08 thayne joined #salt
03:09 cachedout joined #salt
03:10 XenophonF joined #salt
03:14 XenophonF figured out why i can't get the wsgi version of rest_cherrypy to talk to the salt master, which seems obvious in hindsight: two of the four unix domain sockets the master sets up are only readable/writable by root,
03:14 XenophonF namely, master_event_pull.ipc and publish_pull.ipc
03:15 XenophonF if i chgrp those to www and chmod those g+rw, i can do things like fire webhooks very easily
03:16 XenophonF i guess that short term, that's an acceptable workaround
03:16 XenophonF but i wonder if it would be better to (a) run the salt-master as non-root and (b) run the wsgi process in the same account
03:17 XenophonF i'm uncomfortable with the idea of running rest_cherrypy as the httpd user
03:18 XenophonF maybe it doesn't matter - i'm not intimately familiar with how well httpd and mod_wsgi are insulated from one another
03:20 Furao joined #salt
03:23 foo_ joined #salt
03:32 dopesong joined #salt
03:38 evle left #salt
03:51 XenophonF left #salt
04:02 ajw0100 joined #salt
04:02 quan joined #salt
04:02 quan hello
04:08 jeadre joined #salt
04:09 ramteid joined #salt
04:13 jeadre joined #salt
04:18 sunkist joined #salt
04:27 womble What would be the least-worst way to emulate something akin to Puppet's concept of "defined types"?  I'm looking to tell Salt to create a bunch of containers on a machine, but having trouble with how to do it.  The only way I've come up with so far is a Jiant Jinja template which loops through a nested data structure and defines all the necessary states for each container I list, but that feels icky.  Are there other options?
04:27 hvn joined #salt
04:28 foob joined #salt
04:32 * foob slaps ahammond around a bit with a large fishbot
04:35 jeadre joined #salt
04:38 zblum joined #salt
04:50 thayne joined #salt
04:55 Fiber^ joined #salt
05:01 UForgotten joined #salt
05:02 catpigger joined #salt
05:04 darsh_ joined #salt
05:10 dthom91 joined #salt
05:17 dthom91 joined #salt
05:18 dopesong joined #salt
05:31 vignesh_ joined #salt
05:31 vignesh_ hi
05:31 vignesh_ How to install saltstack
05:31 vignesh_ any begineer guie
05:32 favadi joined #salt
05:35 rdas joined #salt
05:41 AmVu joined #salt
05:53 Norrland finding the irc channel before the guide is quite unusual
05:56 jhauser joined #salt
05:56 womble Some people are just talented.
05:56 CeBe joined #salt
06:03 colttt joined #salt
06:04 harkx joined #salt
06:05 katyucha joined #salt
06:09 orion Anyone run in to a KeyError: 'git.present' error when using the git state?
06:09 orion http://lpaste.net/8844867070253334528
06:14 malinoff joined #salt
06:16 lb1a joined #salt
06:16 hojgaard joined #salt
06:20 shiriru joined #salt
06:20 hvn joined #salt
06:23 AndreasLutro joined #salt
06:35 hvn joined #salt
06:44 KermitTheFragger joined #salt
06:46 shiriru joined #salt
06:47 shiriru joined #salt
06:52 jeadre joined #salt
06:53 viq joined #salt
07:00 ITChap joined #salt
07:03 Ztyx joined #salt
07:12 Ztyx joined #salt
07:16 linjan joined #salt
07:16 Ztyx joined #salt
07:17 eseyman joined #salt
07:29 ninkotech_ joined #salt
07:30 mattiasr joined #salt
07:30 wendall911 joined #salt
07:39 favadi joined #salt
07:59 supersheep joined #salt
08:00 Ztyx joined #salt
08:23 kbaikov joined #salt
08:26 Grokzen joined #salt
08:26 aqua^c joined #salt
08:27 stephanbuys joined #salt
08:34 aqua^c_ joined #salt
08:36 impi joined #salt
08:39 cberndt joined #salt
08:46 micko joined #salt
08:54 user371803 joined #salt
08:55 obimod hello all
08:57 obimod anyone know if require.service.dead.name: xorg     # would work?
08:57 obimod or require.service.dead: \n\t- name: xorg
09:01 mike25de joined #salt
09:02 monkey66 joined #salt
09:20 Antiarc joined #salt
09:25 katyucha Hi all
09:31 tkharju joined #salt
09:39 zerthimon joined #salt
09:42 Hydrosine joined #salt
09:43 keimlink joined #salt
09:45 aristedes joined #salt
09:50 hvn joined #salt
09:50 hvn joined #salt
09:58 cberndt joined #salt
09:59 hvn joined #salt
09:59 hvn joined #salt
10:07 aristedes left #salt
10:15 ITChap joined #salt
10:19 DanyC joined #salt
10:20 zer0def joined #salt
10:32 AndreasL1tro joined #salt
10:32 babilen obimod: No, that wouldn't work. What are you trying to achieve?
10:33 obimod babilen: just wondering :)
10:37 babilen A state consists of at least three things: 1. An ID (require.service.dead.name in your example) 2. The state function being called (missing in your example and 3. The 'name' argument (which is the same as ID by default)
10:48 mortis_ I've asked this maybe a year ago, but never found a solution to the problem and finally figured i'd revisit it. Does anyone know if its possible in anyway to match on pillars in the pillar top.sls?
10:48 mortis_ like this:
10:48 mortis_ prod:
10:49 mortis_ - match: pillar
10:49 mortis_ paste missed a line ...brb pastebin :P
10:49 AndreasL1tro mortis_: pillar top.sls needs to be loaded before salt knows what the pillar data is-  how could it?
10:49 mortis_ http://pastebin.com/2GJMtd56
10:49 mortis_ AndreasL1tro: exactly
10:49 mortis_ thats the problem
10:50 jakubm joined #salt
10:50 mortis_ we use external pillars
10:51 mortis_ so it makes sense to match on them
10:51 mortis_ but cant find a nice way to do it
10:51 giantlock joined #salt
10:51 mortis_ ofc i could set grains based on the pillars, and match on them in the pillar top.sls
10:52 yml joined #salt
10:52 stefan_ joined #salt
10:53 lynxman joined #salt
10:55 AndreasL1tro I don't agree that it make sense - external pillars could be loaded before or after top.sls, for example, making the outcomes unreliable
10:55 AndreasL1tro in any case I'm pretty sure it's not possible in the present state
10:56 techdragon joined #salt
10:56 wintamute joined #salt
10:56 AndreasL1tro https://github.com/saltstack/salt/issues/23910 may be relevant
10:56 saltstackbot [#23910]title: Please implement static pillars | Hi,...
10:57 jeadre joined #salt
10:57 LyndsySimon joined #salt
10:58 bytemask joined #salt
10:58 SteamWells joined #salt
10:58 simonmcc joined #salt
10:58 jchadwick joined #salt
10:58 hillna joined #salt
10:59 OliverMT joined #salt
10:59 mortis_ yeah, i can see my colleague next to me have commented on that too
11:00 Laogeodritt joined #salt
11:00 munhitsu joined #salt
11:00 linjan joined #salt
11:00 davedash joined #salt
11:00 PrincessZoey_ joined #salt
11:01 al joined #salt
11:01 impi joined #salt
11:01 tr_h joined #salt
11:01 ThomasJ joined #salt
11:01 kaiyou joined #salt
11:01 berto- joined #salt
11:01 synical joined #salt
11:01 mortis_ maybe this could help us, i'll have to try https://github.com/bbinet/pillarstack
11:02 [vaelen] joined #salt
11:02 packeteer joined #salt
11:04 imanc joined #salt
11:06 skarn joined #salt
11:08 zz_Cidan joined #salt
11:12 babilen I still think that no-fuss static pillars would be nice
11:12 evle1 joined #salt
11:13 babilen (I'd very much like to use mine lookups in my pillars too)
11:13 AndreasLutro I'd like some form of compile passes for pillar data
11:13 Nazca joined #salt
11:15 stephanbuys joined #salt
11:15 AndreasLutro wishful thinking: if my "web_sites" pillar has information about a user a specific website should be ran/owned by, I should be able to just shuffle that into the "users" pillar which is then used by my "users" state
11:17 babilen Yeah, pillar renaming is important. In particular if you want to use formulas and semantically structured pillars that bundle entire "setups"
11:17 babilen Maybe s/renaming/symlinking/ ;)
11:18 AndreasLutro I guess a python rendered pillar could achieve what I want.. but I like yaml :(
11:19 TyrfingMjolnir joined #salt
11:22 aristedes joined #salt
11:24 aristedes If I want to call a module function from a custom runner, how can I construct the __context__ dictionary to allow the module to execute?
11:24 Hazelesque joined #salt
11:25 Mate love the regressions on every single salt point-release
11:26 Mate 2015.5.0 -> 5.3 update prevented half of my minions failing to restart \o/
11:26 giantlock joined #salt
11:27 babilen Mate: Which OS ?
11:27 Mate trusty and centos7
11:28 Mate master_type: str
11:28 Mate in minion config
11:28 Mate generated by salt-cloud (bootstrap)
11:30 aristedes1 joined #salt
11:31 babilen "prevented ... failing to restart" doesn't sound too bad though ;)
11:31 stephanbuys1 joined #salt
11:31 Mate babilen: ...
11:31 babilen We had no issues whatsoever with that upgrade, how did you upgrade your minions and what was needed to get them running again?
11:31 Ztyx joined #salt
11:32 Mate sudo sed -i /etc/salt/minion -e '/master_type/ s/str$/standard/'; sudo service salt-minion start
11:32 Mate this seems to be working
11:32 Mate and I did a  salt \* pkg.install salt-minion refresh=True
11:32 babilen It should be str (i.e. the default) -- See http://docs.saltstack.com/en/latest/ref/configuration/minion.html#master-type -- "standard" is not a valid value AFAICT
11:33 stephanbuys joined #salt
11:33 Mate interesting, "str" throws 2015-08-24 12:56:26,615 [salt.minion      ][ERROR   ] Invalid keyword 'str' for variable 'master_type'
11:34 stephanbuys joined #salt
11:34 Mate while standard is correct
11:35 babilen Mate: Which version do you have installed on that minion?
11:35 Mate https://github.com/saltstack/salt/blob/v2015.5.3/salt/minion.py#L776 this one
11:36 Mate https://github.com/saltstack/salt/commit/b9a756ff5fab76d448e104f1e106e042750d5753
11:37 Mate this changed the values of a configuration
11:37 Mate and no change in user docs
11:37 babilen https://github.com/saltstack/salt/blob/2015.5/salt/minion.py#L781
11:37 babilen (current HEAD of 2015.5)
11:37 Mate and commits like this are always getting merged
11:38 Ztyx1 joined #salt
11:38 Mate https://github.com/saltstack/salt/commit/0dc28ad3e498b5d08bc4070bb11dbaae4e7054d2 it was reverted here
11:38 Mate but not in the release
11:40 babilen So the real problem is that, while 'str' is correct for subsqeuent releases, the configuration file generated for your minion should have included 'standard' for 2015.0.3
11:40 babilen Why do you set it at all?
11:40 Mate 1340 -!- Topic for #salt: Welcome to #salt | 2015.5.3 is the latest
11:40 babilen Ah, it has been generated by salt-cloud
11:41 Mate but why was a patch like this merged?
11:41 babilen Looks as if salt-cloud generates broken configurations for 2015.5.3 then
11:41 Mate without a deprecation warning or anything
11:42 babilen Mate: Because "str" communicates the "type" better, but it was, obviously, incomplete and not planned well.
11:42 * babilen shakes fist at cachedout
11:43 Mate i had to add       minion:
11:43 Mate master_type: failover  # workaround unreleased https://github.com/saltstack/salt/commit/f732be36
11:43 Mate to my salt-cloud earlier
11:43 Mate but my old minions broke today
11:43 babilen I would guess that this change will be well documented in 2015.8.0 release notes, but the problem is that salt-cloud should *still* generate configs with "standard" in there
11:43 babilen (for/in 2015.5.3 that is)
11:43 Mate failover because i expected standard will not work later
11:43 Mate as it was reverted
11:44 Mate so the fix was also incompatible with the already released version
11:45 AndreasLutro salt has relatively poor test coverage - and their team seems to be all about move fast and break things
11:45 babilen So, lets get this straight: The current stable version wants 'standard' in there, which did not change from the version before (2015.5.0). Correct?
11:45 AndreasLutro not the best combination maybe!
11:46 Mate 5.0 still had 'str'
11:46 ashashyne joined #salt
11:46 babilen Mate: Okay, that is a problem then. Did you report that?
11:46 kbaikov joined #salt
11:47 AndreasLutro it's already been fixed in 2015.5.4 and .5 - but those releases haven't been packaged yet
11:47 Mate there are many different reports about this
11:48 Mate but it seems like this issue-flow would need some discussion about the processes of development
11:48 babilen Yeah, that is an issue that should have been caught if saltstack tested with minions that use configurations created for them by salt-cloud
11:49 babilen (or configurations that test various setting combinations, but that is much harder to implement)
11:49 Mate but i don't think the resolution (the revert patch) was correct
11:50 babilen In the end configuration value changes "just for the sake of it" shouldn't be merged. We are getting *much* better at this in the context of formulas, but salt-proper seems to still be a bit too lenient.
11:50 babilen Mate: What do you consider to be correct?
11:51 Mate keep the change and add compatibility with "str" (possibly with a deprecationwarning)
11:51 babilen ack
11:51 babilen +1
11:51 mattiasr joined #salt
11:51 Mate the revert patch broke correctet configs again
11:51 Mate d
11:51 babilen "now for twice the fun!"
11:51 Mate (:
11:52 babilen Could you link some of the issues you referred to earlier?
11:52 mike25de hi guys... i am redoing my salt setup... and i want to have a base-pillar which stores some data and minion-pillar for each minion. Is it possible to merge those 2 pillars so the minion can overwrite values from the base-pillar? Thanks
11:52 babilen (just curious)
11:52 Mate https://github.com/saltstack/salt/issues/26291 this links to two others
11:52 stephanbuys1 joined #salt
11:52 saltstackbot [#26291]title: Invalid keyword 'str' for variable 'master_type' | Hello,...
11:52 babilen mike25de: Pillars are merged already -- http://docs.saltstack.com/en/latest/ref/configuration/master.html#pillar-source-merging-strategy
11:53 babilen Mate: ta!
11:53 XenophonF joined #salt
11:54 thayne joined #salt
11:54 XenophonF what's the right way to get configuration data into a reactor script?
11:54 mike25de thanks babilen
11:55 XenophonF i have a secret key that i want to set---presumably in the master config file
11:56 babilen mike25de: fwiw, I wouldn't make your pillar setup too complicated as it *might* become tricky to keep track of where which k/v pair actually comes from. There are other approaches to this if you want to implement hierarchies of pillars that are supported by external pillars in salt. Notable mentions include reclass, varstack, ... http://docs.saltstack.com/en/stage/ref/pillar/all/index.html
11:56 jhauser joined #salt
11:57 mike25de babilen: thanks a lot for your input mate
11:58 AYM_ joined #salt
11:59 AYM_ Hi guys
11:59 babilen mike25de: It is fine to do some "merging" , but I found that it could become hard to manage if not done properly. This does, however, only hold true if pillars further down the hierarchy replace values that are higher. I very often split monolithic pillars up into smaller ones, but take care that there is no overlap in the "leafs" that actually hold the values.
12:00 AYM_ Am looking into integrating SALT (which we use for all our servers) and HipChat (which is our default groupchat application)
12:00 AYM_ Any advise, tuts, links to docs would be very much welcome.
12:00 babilen AYM_: "integrate" how ?
12:01 AYM_ Through notifications for instance. But basically if what is available from http://www.rtop-monitor.org/rtop-bot/ could be achieved directly from Salt would be great
12:01 Ztyx joined #salt
12:01 mike25de babilen: i am thinking how to do it best. My pillar key-var is actually in mysql... so i am thinking if i should merge before generating the pillar or generate 2 sets of pillars and let salt do the merge... but babilen your ideas are good
12:04 babilen mike25de: I'd say that either approach is fine, but take care that the order in which values are being merged is absolutely clear and that everyone who might come in contact with it is aware of it.
12:05 AYM_ Saw this http://docs.saltstack.com/en/latest/ref/states/all/salt.states.hipchat.html
12:05 babilen AYM_: Does such a bot exist for hipchat already?
12:05 mike25de sure babilen ... i should keep it simple, but thanks a lot for your help - much appreciated
12:05 AYM_ But not sure how to implement it
12:06 AYM_ Yes, the bot above (rtop-monitor) does already exist, and it was while trying to implement the agent on our Salt master that the idea came to me to use Salt directly
12:07 babilen AYM_: I am not aware of any bot that talks to salt (be it for IRC or HipChat). The states above essentially provide the means to send messages from an ongoing salt run to your chat (e.g. to report "I'm done"), but that doesn't allow you to trigger salt actions from within HipChat
12:08 AYM_ Okay, thank you Babilen.
12:08 babilen If you have such a bot for HipChat already it might be possible to hook it up to salt's API (cf. http://docs.saltstack.com/en/latest/ref/netapi/all/index.html#all-netapi-modules )
12:08 babilen Also ping the mailing list, maybe somebody is working on that already
12:09 AYM_ Am glad I found this place through my search.
12:09 AYM_ Thanks for the links and advice.
12:09 AYM_ Cheers
12:14 fredvd joined #salt
12:20 diegows joined #salt
12:22 zwi joined #salt
12:22 amcorreia joined #salt
12:25 teryx510 joined #salt
12:27 dyasny joined #salt
12:27 AndreasLutro joined #salt
12:28 tmclaugh[work] joined #salt
12:29 xf10e joined #salt
12:29 xf10e tzag *
12:30 xf10e does anyone know how to access modules synced via saltutil.sync_utils in a (py-renderer) state?
12:31 xf10e got them synced, but how to access them? utils is undefined and __utils__ is undefined, too
12:38 GreatSnoopy joined #salt
12:39 lrojas joined #salt
12:41 pm90__ joined #salt
12:54 cpowell joined #salt
12:55 cpowell joined #salt
12:58 numkem joined #salt
12:59 rideh joined #salt
13:00 favadi joined #salt
13:02 rideh joined #salt
13:05 numkem joined #salt
13:12 babilen xf10e: What are you doing exactly? Could you paste modules and state to any of http://refheap.com, http://paste.debian.net, https://gist.github.com, http://sprunge.us, http://dpaste.de, … please?
13:14 pravka joined #salt
13:14 bhosmer joined #salt
13:14 Twiglet anyone know if clean on file.directory can only require files?
13:15 babilen clean doesn't require anything, it is an option you can set ?-)
13:15 Twiglet I mean so it doesn't blast the files I want salt to manage
13:15 Twiglet http://hastebin.com/sekalosoni.sm
13:15 Twiglet I want it to clean the dir, unless salt is managing the file with pkgrepo
13:16 Twiglet but it doesn't seem to honor it
13:16 babilen That's a circular dependency
13:16 Twiglet clean
13:16 Twiglet Make sure that only files that are set up by salt and required by this function are kept.
13:17 Twiglet that suggest it'll clean anything unless salt has managed it
13:17 Twiglet no it's not, core-apt-sources-default != core-apt-sources-default-dir-clean
13:18 Twiglet Whichever way I order it it deletes the ubuntu.list file
13:18 babilen Ah, nvm... I saw file: core-apt-sources-default and thought it to be "file: core-apt-sources-dir-clean"
13:18 Twiglet ;)
13:18 babilen So, yeah, that looks fine
13:18 Twiglet my bad for missing that one out of the file
13:19 Twiglet it'll either clean the dir then add it, or add it then clean it
13:20 babilen The only problem is that pkgrepo.managed is not really a managed file, but another state. AFAIUI you can use clean with other file.* states
13:20 Twiglet that was my assumption, didn't really want to have to move it out to a file when pkgrepo is there
13:20 babilen Maybe a file.exist on the file created by pkgrepo works
13:20 nixfloyd joined #salt
13:20 Twiglet ah, will try that
13:21 xf10e babilen: I want to replace the helper functions in https://github.com/fraunhoferfokus/openvswitch-formula/blob/master/networking/config.sls
13:22 nixfloyd left #salt
13:22 xf10e I'm trying to figure out which functions defined there I can replace with equivalents from salt.utils.network and which I have to get into salt.utils.network ;)
13:23 thayne joined #salt
13:23 jdesilet joined #salt
13:24 Edgan joined #salt
13:25 jdesilet joined #salt
13:26 Twiglet babilen: nope, doesn't like file.exists :/
13:26 Twiglet will have to use file.managed I guess
13:27 kaiyou joined #salt
13:27 babilen Twiglet: Yeah, file.exist is a bit "speshul" :(
13:27 fe92 joined #salt
13:27 babilen xf10e: So, assuming you replaced some in your execution module. Why can't you call them?
13:29 xf10e it's common network functions, stuff that might be useful else so I don't think it belongs (directly) in a execution module
13:29 timoguin joined #salt
13:29 xf10e s/else/elsewhere/
13:30 xf10e just like the other stuff in salt.utils.network
13:30 bhosmer joined #salt
13:31 babilen Sure, totally agree. I might have used "execution module" a bit loosely earlier. I essentially meant: "You've replaced your functions now. What is the actual problem?"
13:31 perfectsine joined #salt
13:32 xf10e I can't access them in the py-rendered state. _utils get's synced but I have no idea how to access the code. there's no `utils` or `__utils__` I can use like `salt` or `grains`
13:32 al_ joined #salt
13:34 xf10e thus I can't update salt.utils.network w/o replacing ... what's its name... /usr/lib/python2.7/dist-packages/salt/utils/network.py on my ubuntu-minions
13:36 mapu joined #salt
13:36 xf10e b/c I can import salt.utils.network just fine. but then I won't get my _utils/network.py...
13:36 bhosmer joined #salt
13:37 denys joined #salt
13:38 babilen Why don't you get your _utils/network.py after syncing?
13:38 seb` joined #salt
13:38 babilen Did you restart the minion?
13:39 racooper joined #salt
13:39 babilen http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.saltutil.html#salt.modules.saltutil.refresh_modules might be what you need, but not entirely sure
13:40 babilen sync_all should™ call that already btw
13:40 xf10e nope, ran saltutil.sync_all, restarted the minion several times. the file is on the minion but importing salt.utils.network gives me /usr/lib/python2.7/dist-packages/salt/utils/network.py (which I see in my stacktraces)
13:42 babilen Ah, sync_utils takes refresh=True
13:43 orion Is there an easy way to determine if there are any unreferenced pillar values?
13:43 babilen My assumption would be that you get the network.py from _utils after syncing with "refresh=True" or restarting the minion afterwards.
13:43 mpanetta joined #salt
13:44 babilen orion: Can't think of one
13:44 subsignal joined #salt
13:45 babilen xf10e: If that is not the case it is either a bug or we misunderstand how _utils and sync_utils are supposed to behave.
13:45 mattrobenolt joined #salt
13:45 babilen (don't use it, so I only go by documentation)
13:45 racooper joined #salt
13:48 zmalone joined #salt
13:48 xf10e yeah... I didn't find any documentation on _utils
13:48 XenophonF FYI when running rest_cherrypy via WSGI, make sure the user account it runs as has access to the master_event_pull.ipc and publish_pull.ipc Unix domain sockets
13:50 xf10e babilen: I guess I have to open a documentation bug about this ;)
13:51 schristensen joined #salt
13:53 xf10e babilen: btw, it seems __utils__ is the way to go in a exec or state module
13:54 babilen xf10e: Well, it might be a genuine bug in that modules in _utils should override those in .../dist-packages/salt/utils
13:54 anmol joined #salt
13:54 anmol hi
13:54 ekristen joined #salt
13:55 DammitJim joined #salt
13:55 timoguin joined #salt
13:55 anmol I am having a three node salt setup
13:55 anmol 1-master and 2 minions
13:56 anmol I am trying to install and configure collectd using salt
13:56 anmol While I could get it configured on one of my nodes on the other I am hitting an issue
13:56 ajw0100 joined #salt
13:56 anmol Data failed to compile:
13:56 anmol ----------
13:56 anmol Rendering SLS 'base:collectd' failed: Jinja error: collectd/map.jinja
13:56 anmol Traceback (most recent call last):
13:56 anmol File "/usr/lib/python2.7/site-packages/salt/utils/templates.py", line 306, in render_jinja_tmpl
13:56 dthom91 joined #salt
13:56 anmol output = template.render(**decoded_context)
13:56 anmol File "/usr/lib/python2.7/site-packages/jinja2/environment.py", line 969, in render
13:57 anmol return self.environment.handle_exception(exc_info, True)
13:57 anmol File "/usr/lib/python2.7/site-packages/jinja2/environment.py", line 742, in handle_exception
13:57 anmol reraise(exc_type, exc_value, tb)
13:57 anmol File "<template>", line 1, in top-level template code
13:57 babilen anmol: Please use one of http://refheap.com, http://paste.debian.net, https://gist.github.com, http://sprunge.us, http://dpaste.de, … and do *not* paste into the channel
13:57 anmol File "/usr/lib/python2.7/site-packages/salt/utils/jinja.py", line 143, in get_source
13:57 anmol raise TemplateNotFound(template)
13:57 anmol TemplateNotFound: collectd/map.jinja
13:57 anmol ; line 1
13:57 babilen Oh dear ...
13:57 anmol ---
13:57 anmol {% from "collectd/map.jinja" import collectd_settings with context %}    <======================
13:57 anmol include:
13:57 anmol - collectd.service
13:57 anmol collectd:
13:57 anmol [...]
13:57 anmol babilen, sure my apologies
13:58 babilen So, you are using https://github.com/saltstack-formulas/collectd-formula /
13:58 babilen ?
13:58 anmol yep
13:59 anmol babilen, I am seeing the error as in http://paste.debian.net/304485/
13:59 dthom91 joined #salt
14:00 babilen How do you include the formula? How do your two minions differ?
14:02 rawzone^ joined #salt
14:02 Akhter joined #salt
14:02 DenkBrettl joined #salt
14:03 Norrland Running onto some trouble updating debian packages with my state-file. the package version is "3.7.7-10-abcd" and the update is "3.7.7-15-foo", It just says "Package: foo-bar is already up-to-date"
14:03 babilen anmol: Please keep it in the channel. Could you paste the output of "salt 'someminion' cp.list_master" ?
14:04 rdutch joined #salt
14:04 Twiglet it'll either clean the dir then add it, or add it then clean ithttp://hastebin.com/dewowawamu.sm
14:05 XenophonF when targeting minons, is there a shortcut for self (i.e., this master)?
14:05 Twiglet i've tried with '(blah|ubuntu)' '(blah|ubuntu).list' '(blah|ubuntu)\.list' all with no joy
14:05 babilen Norrland: Debian version numbers should end in a number and -abcd strikes me as rather uncommon. Which package is this exactly and what are the exact version numbers involved here?
14:05 XenophonF ah never mind i need to target a different way, anyway
14:06 ipmb joined #salt
14:06 anmol babilen, http://paste.debian.net/304489/
14:07 Akhter_ joined #salt
14:08 babilen anmol: Could you verify that map.jinja is indeed in /srv/formulas/collectd-formula/collectd ?
14:08 jeddi joined #salt
14:08 Norrland babilen: internal packages built for our environment.
14:08 Norrland so it's <version>-<buildhash>
14:08 bhosmer joined #salt
14:09 babilen Norrland: -abcd is a rather uncommon debian_revision. Could you be more specific about the exact versions involved here? Why did you chose such a non-standard one?
14:09 Twiglet Norrland: We had to deal with something similar, we ended up going to data based versioning
14:09 anmol babilen, its fact in a level higher i.e, srv/formulas/collectd-formula/
14:09 Twiglet s/data/date
14:09 Norrland babilen: -abcd is an example.
14:10 Twiglet the issue is that apt doesn't see the hash as a numerical version unless it ends in a number
14:10 Norrland Version table: 3.7.4-18-geaa9fad 0
14:10 Norrland for example
14:10 babilen Norrland: Well, it implies that you use only [a-zA-Z] there which is quite uncommon and might be the reason for the behaviour you are seeing. Take a look at https://www.debian.org/doc/debian-policy/ch-controlfields.html#s-f-Version which makes your software version "3.7.7-10" and the debian revision "-abcd"
14:11 Norrland babilen: I'll look at that.
14:11 babilen Norrland: What are you trying to achieve exactly? Are you building git HEAD ?
14:12 XenophonF you know what, i wish thach45 would have just written the damn thing in lisp
14:12 babilen +1
14:12 * Twiglet runs away screaming
14:12 babilen s/lisp/clojure ?
14:12 XenophonF i mean actual Common Lisp, the real deal
14:12 babilen Well, you didn't specify which one, but I assumed you mean common
14:12 Norrland babilen: I have a setup where jenkins builds the git HEAD, and produces an .deb package, which it in turn installs by calling "salt 'myminion' state.sls mystate"
14:13 Norrland babilen: the thing is that it stopped installing the new packages.
14:13 dthom91 joined #salt
14:13 XenophonF then, i'd only have to keep track of one kind of paren, and quoting would be sane
14:13 andrew_v joined #salt
14:14 anmol babilen, I moved map.jinja into the location you told me and it worked thanks but I wonder how it worked on the other node with the same setup
14:14 babilen Norrland: Okay, the common naming scheme for these packages would be something along the lines of "X.Y.Z+git$DATE-$SHA-1" with X.Y.Z being the latest stable version (+ sorts after, hence the +)
14:14 anmol babilen, thanks
14:15 Norrland babilen: okey
14:15 babilen anmol: I have no fucking clue, but could you paste the entire contents of your /srv/formulas (run "tree" therein) directory and the "file_roots" bit of your master's configuration
14:15 Norrland babilen: will look into how we name the packages.
14:16 babilen Norrland: If you require more than one build per day you would use YYYYMMDDHHMM for DATE which ensures that packages build later on are "newer" in terms of their version.
14:17 zwi joined #salt
14:19 lothiraldan joined #salt
14:20 Norrland babilen: yep
14:20 babilen Norrland: "grep -h Version /var/lib/apt/lists/*_binary*Packages|grep git|sort|uniq" for Versions found in Debian already. I get this on a current unstable system: http://paste.debian.net/304496/
14:21 Norrland babilen: mkay, thanks. Looking at our jenkins-setup atm.
14:21 XenophonF i just typed "] } } } } ] } ] }" into a #!py reactor script
14:21 babilen Norrland: The choice between + and ~ is essentially down to what you want to express. + is more commonly used as it expressed "build from SHA *after* last stable version X.Y.Z", but ~ (which sorts lower) might make sense if you want to express "working up to next stable version X.Y.Z"
14:21 XenophonF dammit, left out one "}" too
14:21 babilen XenophonF: You need paredit and rainbow-parantheses!
14:22 LotR joined #salt
14:22 anmol babilen, /srv/formulas/    ==>  http://paste.debian.net/304497/    and "file_roots" bit of master's configuration ==> http://paste.debian.net/304498/
14:22 XenophonF oooh i knew i forgot to install something when i re-did my emacs config a couple of weeks ago
14:23 XenophonF thanks for the reminder babilen! :)
14:23 Deevolution joined #salt
14:23 Norrland babilen: mhm, trying to figure out the magic oneliner found in jenkins that my devs put there a long time ago.
14:25 babilen XenophonF: https://github.com/clojure-emacs/cider#integration-with-other-modes has a nice overview of "lisp-y" modes/packages available for emacs. I'm quite partial to RainbowDelimiters and https://github.com/Fuco1/smartparens (alternative to paredit)
14:25 XenophonF babilen: i will definitely check those out
14:25 dthom91 joined #salt
14:25 zmalone joined #salt
14:25 Twiglet and file.directory clean doesn't even respect require: file now
14:26 babilen That little bastard!
14:26 Twiglet yup
14:26 Twiglet doesn't seem to work with exclude_pat either
14:26 XenophonF in a multi-master setup with reactors, do reactor scripts run on all of the masters?
14:27 babilen anmol: So, where was map.jinja before? That looks just fine.
14:27 DanyC joined #salt
14:27 DanyC left #salt
14:27 XenophonF like if they're all configured identically
14:27 gnudon joined #salt
14:28 Andrei111 joined #salt
14:28 anmol babilen, I moved it after you told me. It was in /srv/formulas/collectd-formula before
14:28 malinoff joined #salt
14:28 dthom91 joined #salt
14:29 babilen anmol: It shouldn't have been there! I mean https://github.com/saltstack-formulas/collectd-formula/tree/master/collectd is what you should have cloned and that has map.jinja in collectd as you can see.
14:29 babilen anmol: Not sure how it moved in your setup
14:30 lothiraldan joined #salt
14:30 hasues joined #salt
14:30 hasues left #salt
14:30 anmol babilen, thanks a lot may be I missed it up unintentionally
14:30 mike25de hi guys again... me and my ideas: is it possible to run a state on a minion automatically ... right after the minion was accepted by master?  No highstate was run yet.
14:31 XenophonF mike25de: yes
14:31 XenophonF salt reactor
14:31 mike25de XenophonF: thanks man
14:32 XenophonF look for a salt/key event with act=accept, id will be the minion ID whose key you just accepted
14:33 fxhp joined #salt
14:34 XenophonF in jinja it'd be something like {% if data['act'] == 'accept' %} followed by a highstate run with tgt set to {{ data['id'] }}
14:34 babilen mike25de: https://docs.saltstack.com/en/develop/topics/reactor/index.html is the documentation you want to read
14:34 XenophonF mike25de: http://docs.saltstack.com/en/latest/topics/reactor/#mapping-events-to-reactor-sls-files
14:35 XenophonF mike25de: http://docs.saltstack.com/en/latest/topics/reactor/#knowing-what-event-is-being-fired
14:35 babilen eventlisten.py ftw!
14:36 babilen Ah, these days it is probably easier to use "salt-run state.event pretty=True" I guess :)
14:36 XenophonF LOL
14:36 XenophonF IIUC the salt/key event only fires when you do something with the salt-key command
14:36 babilen It does, yeah
14:37 babilen mike25de: fwiw (not saying that reactors aren't awesome), but http://docs.saltstack.com/en/latest/ref/states/startup.html might be of interest too
14:37 mike25de babilen:  XenophonF thanks guys
14:38 Twiglet some days I really wonder why we use salt, other days I love it. it's like an abusive relationship
14:38 mike25de babilen:  you are the SHIT!!! that's what i needed
14:39 XenophonF heh well i've been digging into reactors for the last week, so, you know, when all you have is a hammer...
14:39 lothiraldan joined #salt
14:40 bhosmer joined #salt
14:40 _JZ_ joined #salt
14:41 mike25de XenophonF:  thanks man... reactors are cool as well. I had for 2 years salt 0.17 .. now i am moving on to the new version and redo some of the states / logic
14:42 mike25de just descovering now the new features of salt (new to me )
14:42 Twiglet That's a big old jump
14:42 kaptk2 joined #salt
14:43 XenophonF mike25de, dude, babilen showed me part of his crazy awesome reactor setup, where minions can use the reactor system to send X.509 certificate signing requests on a certificate authority running on another minion
14:44 mike25de XenophonF: babilen is awesome .. i saw he did some crazy stuff there ... too much for me :)
14:44 dthom91 joined #salt
14:45 DanyC joined #salt
14:45 mike25de i think salt... has made a huuuge leap in the last year... although ... there are still a lot of people stuck in the puppet chef world
14:45 Sketch that sounds rather similar to what i want to set up today...
14:45 favadi joined #salt
14:45 babilen Sketch: What does?
14:45 mike25de regarding to STARTUP STATES ... are the states run everytime the minion is restarted or just only once after the minion is accepted by the master?
14:46 babilen mike25de: The former
14:46 Sketch sending signing requests to a CA on another machine
14:46 mike25de cool babilen
14:46 babilen Sketch: I plan to open-source my solution soon™, but have to clean it up for general consumption as it includes some idiosyncrasies still
14:46 mike25de thanks man ... i really owe you some beers babilen :)
14:46 babilen If you are ever in Berlin ping me
14:47 mike25de babilen: if you are ever in Frankfurt :) ping me :P
14:47 impi joined #salt
14:47 lothiraldan joined #salt
14:47 mike25de i might visit Berlin soon.. so i might call you babilen
14:48 babilen Sure, I'm in .uk almost half of the time, but let me know
14:48 mike25de will do that!
14:48 scoates joined #salt
14:49 Sketch babilen: well, it's not _exactly_ what i want to do, so any general guidelines would be helpful.  i am new to reactors (and salt in general)
14:53 babilen Sketch: So, the general idea I have is that I have an execution module that contains all functions for generating, reading and writing of certificates. I use event.fire_master to fire events from the minion to the master and include all data that might be needed in its "data" argument
14:54 armguy joined #salt
14:55 babilen Sketch: So I basically pack "data" dictionaries with *everything* that is needed. This typically includes information that uniquely identifies a specific CA and the payload (certificates). On the master side I then have reactor files that listen to *custom* events. One of them is "saffron/cert/request" (saffron is the name of the module, and 'request' says "I want one"
14:55 mike25de i am being stupid now... so i see that the states startup can be configured on the minion conf... but that would mean to already have access to it. Can i setup the startup sls without modifying the minion conf?
14:56 mapu joined #salt
14:57 babilen Sketch: The data dictionary also includes data['ca_name'] that uniquely identifies a CA that I configured on a specific minion (say pki-minion) and data['ca_host'] = pki-minion ... So a minion requests data from another minion and uniquely identifies that data with a given key.
14:59 babilen Sketch: The key is important as I use something along the lines of saffron:$CA_KEY:{setting1, setting2, ...} for CA settings. These can be set in the configuration files, grains or pillars and are exclusively referenced by http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.config.html#salt.modules.config.get in the execution module.
14:59 babilen That way I can easily configure CAs in a way that allows me to identify them easily later on.
15:00 stomith joined #salt
15:02 perfectsine joined #salt
15:03 pbrooko joined #salt
15:04 cachedout joined #salt
15:05 babilen Sketch: So, a minion then requests some data from another one. The master receives that event and calls saffron.get_certificate(....) on the CA minion. The execution module checks if it has generated a cert before and sends it if it has or generates a new one. The first is done via a saffroncert/cached event and the latter with a saffron/cert/new one. The master then receives the cert and returns it to the m
15:05 babilen inion that requested it. To be able to do ...
15:05 babilen ... so it is important to continue including details of the minion that initially requested the cert.
15:10 Akhter joined #salt
15:10 rawzone joined #salt
15:15 sdm24 joined #salt
15:16 mike25de again about startup state: the state has to be defined in the minion conf file. But i need to be able to run a script/state right after the minion is accepted byu the master (the minion conf file is the default). Is that possible using startup state? or i should try reactor?
15:17 mike25de it seems that rreactor is the way to go
15:17 mike25de thanks babilen
15:19 XenophonF sie sind alles in Deutschland? wunderbar!!!
15:20 mike25de XenophonF: yeah :) i am in Frankfurt but i don't speak Deutsch
15:21 XenophonF that's awesome. i'm 2nd generation German American, have family in both cities (and elsewhere).
15:21 XenophonF my German's pretty poor, though
15:21 mike25de XenophonF: i moved to DE 2 years ago... but being in a multinat company ... i didn't learn any DE
15:22 zblum joined #salt
15:22 mike25de it's on my todo list :D
15:22 XenophonF :)
15:23 XenophonF well, the funny thing is, i learned german so i could converse with my grandma, who's from southern German
15:23 mike25de haha
15:23 mike25de :)
15:23 XenophonF so my accent is this weird amalgam of both English and Suddeutsch, like if somebody explicitly wanted to learn English with a southern American drawl :)
15:23 mike25de :))))
15:24 XenophonF OK so I think I have my #!py reactor script ready to test
15:24 * XenophonF tries to get back on topic
15:25 XenophonF i'm seeing events on the bus from github
15:25 DaveQB joined #salt
15:25 XenophonF a simple reactor script sees them and fires off accordingly
15:25 XenophonF time to see how buggy my python code is :)
15:28 sdm24 intentionally add bugs so you have something to do later haha
15:28 duendecat joined #salt
15:28 bhosmer joined #salt
15:29 XenophonF LOL
15:29 pravka joined #salt
15:30 _JZ_ joined #salt
15:34 Ztyx joined #salt
15:35 Ztyx1 joined #salt
15:37 TranquilityBase joined #salt
15:37 coval3nce joined #salt
15:38 zblum joined #salt
15:40 dthom91 joined #salt
15:41 numkem joined #salt
15:42 ThomasJ|d joined #salt
15:43 Akhter joined #salt
15:46 dyasny joined #salt
15:47 pravka joined #salt
15:48 johnkeates joined #salt
15:51 ngks joined #salt
15:54 NV joined #salt
15:55 dyasny joined #salt
15:55 ALLmightySPIFF joined #salt
15:58 jeadre joined #salt
15:58 zsoftich3 joined #salt
16:00 zblum joined #salt
16:05 scurry joined #salt
16:05 zmalone joined #salt
16:08 Bryson joined #salt
16:08 orion grr...
16:09 orion When you use "clean" with file.directory, should the clean be ordered before or after the files you want to keep?
16:09 yanivtr joined #salt
16:09 orion No matter which way I order it, the "clean: true" directive *always* wipes the whole directory.
16:10 mapu joined #salt
16:12 zmalone joined #salt
16:13 yanivtr hello
16:14 jalbretsen joined #salt
16:14 DammitJim joined #salt
16:15 Fiber^ joined #salt
16:17 shaggy_surfer joined #salt
16:21 dthom91 joined #salt
16:22 yanivtr how can we change the salt-minion environment variables that it passes to commands that it runs?
16:22 yanivtr for example:
16:22 yanivtr we have an sls file that checks if a package is installed via yum
16:22 johnkeates left #salt
16:23 yanivtr since it runs on an internal network, access to external yum repositories require to have http_proxy environment variable configured
16:23 yanivtr we have http_proxy configured in /etc/profile.d/proxy.sh
16:24 yanivtr when we manually log in via ssh, this variable is defined  in the environment
16:24 shaggy_surfer joined #salt
16:25 yanivtr but when commands are being run (such as repoquery) by salt-minion as part of the pkg.installed state, this environment variable does not exist
16:28 nahamu how are you launching the salt minion?
16:28 nahamu could you modify the environment variable before the minion gets launched?
16:29 nahamu for example if you're using systemd you could modify the unit file...
16:30 Grokzen joined #salt
16:32 indispeq Is there a smart way to store binary files in the pillar, or some sort of state that can take care of this?
16:33 indispeq At the moment we end up base64 encoding files
16:33 indispeq and encrypting them using PGP and putting them in pillar
16:33 indispeq but the base64 decoding with cmd.run is a bit clunky and repeats itself at every run
16:33 indispeq did anybody run into this problem?
16:33 cberndt joined #salt
16:34 indispeq for example in storing binary certificate stores
16:34 rideh joined #salt
16:34 XenophonF indispeq: i have
16:34 XenophonF i use the !!binary yaml type
16:34 XenophonF i don't encrypt that with pgp
16:35 capricorn_1 joined #salt
16:35 XenophonF i haven't had time to look at pgp-encrypted git-backed external pillars
16:35 indispeq where could I find more info on the !!binary yaml type? Could I still use the GPG renderer with that?
16:35 XenophonF yaml.org
16:36 indispeq ah ok, I see it. And the yaml renderer in salt just automagically knows how to interpret it then I assume
16:37 aparsons joined #salt
16:37 XenophonF it's a standardized part of the YAML spec, so, yes
16:38 XenophonF i haven't had a chance to use the gpg renderer yet
16:38 writtenoff joined #salt
16:38 indispeq awesome, thank you, I did not know about this ability. This should make things easier if I can make it work with the GPG renderer.
16:38 XenophonF i assume it's the same as saying yaml|jinja|gpg
16:38 XenophonF if so, just put your keymat in there and encrypt it like you would any other pillar
16:39 XenophonF hang on i have an example for you
16:39 khaije1 is there a way to get salt-ssh to return info in something closer to a stream, (for when jobs with lots of return info) ?
16:40 XenophonF well, i thought i had an example but it looks like all my example certs are in PEM format, so i don't use !!binary
16:41 indispeq XenophonF, no problem. Thank you, this is a good starting point for me. I appreciate the help
16:41 yanivtr i am launching the salt-miion using the init.d script
16:43 manfred khaije1: you can't do that with salt, it returns all the information at the end of the run, whether it is salt-ssh or with the master
16:45 yanivtr how can we change the salt-minion environment variables that it passes to commands that it runs?
16:45 yanivtr i am launching the salt-miion using the init.d script
16:46 XenophonF yanivtr: by altering the environment in which salt-minion gets launched
16:46 manfred https://github.com/saltstack/salt/blob/develop/pkg/salt-minion.upstart.rhel6#L15 it reads /etc/default/salt-minion
16:46 yanivtr however, commands (like repoquery) will not see the environment variable http_proxy that is defined in /etc/profile.d/proxy.sh
16:46 manfred or /etc/sysconfig/salt-minion
16:46 manfred like every other init.d script?
16:46 tanta joined #salt
16:47 bhosmer joined #salt
16:47 manfred https://github.com/saltstack/salt/blob/develop/pkg/rpm/salt-minion#L45
16:47 yanivtr got it
16:47 XenophonF on freebsd you'd need to alter the default path in /etc/rc.conf
16:47 XenophonF there's probably similar ways to set system-wide environment variables on linux
16:48 yanivtr thanks manfred
16:48 markm joined #salt
16:49 pm90_ joined #salt
16:50 troyready joined #salt
16:51 yanivtr i think i will use the /etc/default/salt-minion
16:52 yanivtr and define the variables there
16:52 khaije1 manfred: gotcha thanks for the info
16:54 manfred yanivtr: you will want to check which initscript you have, and then use whichever default file is appropriate
16:54 manfred it looks like sometimes it is /etc/default/salt and sometimes salt-minion
16:58 indispeq XenophonF, I just tested the binary encoding, works great if you use the canonical !!binary "\ and then close quotes. I'll just need to make it work with the GPG renderer now.
16:58 aarontc joined #salt
16:58 XenophonF indispeq: good, glad to hear it
16:59 XenophonF that reminds me to rtfm about the gpg renderer
16:59 zsoftich3 speaking of http_proxy...do modules like pkgrepo use that env var? I seem to be having a bit of difficulty getting things to work behind a proxy
17:00 diegows joined #salt
17:04 forrest joined #salt
17:06 jaybocc2 joined #salt
17:12 stomith is there a way to run salt-cloud 'create_snapshot' non-interactively?
17:12 indispeq XenophonF, it works, but there is a bit of trickery in the rendering order. The doc says to do jinja | yaml | gpg
17:12 manfred stomith: anything you can run from the commandline, you should be able to run from python code using the CloudClient class
17:13 indispeq thus I have to gpg encrypt the binary by catting it | gpg --encrypt --armor | base64 >> output.yaml
17:13 indispeq but the end-result works
17:13 manfred if you just want it to take the snapshot, you might try adding -y
17:13 indispeq hacky but it seems to work on randomly selected binary executables and the md5sums match up
17:13 stomith manfred, I tried that, didn't seem to work. I'll look into cloudclient class, thanks
17:16 stomith I mean, it DOES if I put the -y in the right place. Duh.
17:18 ajw0100 joined #salt
17:18 stomith although it doesn't seem to take wildcards
17:22 impi joined #salt
17:27 baweaver joined #salt
17:27 coval3nce joined #salt
17:28 coval3nce Anyone have suggestions on how to use a network state to add multiple loopback ip’s to lo?
17:28 iggy step 1. Don't use the network state
17:29 iggy step 2. Use anything other than the network state
17:29 iggy step 3. Profit
17:29 coval3nce Ha…yeah the network state didn’t seem like it was going to do it for me ;)
17:30 theologian joined #salt
17:32 coval3nce oh, maybe i can use “- ipaddr0” instead of “ - ipaddr” for the network state?
17:32 coval3nce Whats the reason for that statement there iggy ?
17:34 iggy it isn't really up to par with the rest of salt code
17:35 sirex joined #salt
17:36 CeBe joined #salt
17:36 shiriru joined #salt
17:40 coval3nce sounds good, thx man
17:41 drawsmcgraw For running Powershell scripts via cmd.script(), does the 'env' parameter work? Anyone have any experience there?
17:42 drawsmcgraw I'm deciding between templating my script with Jinja or just having the script pull from env variables.
17:43 _JZ_ joined #salt
17:43 zer0def joined #salt
17:44 Akhter joined #salt
17:46 ajw0100 joined #salt
17:50 numkem drawsmcgraw: I'd go with jinja simply because you could use pillar also or even grains to add more variables.
17:51 drawsmcgraw numkem: Good point. I'm curious now but yeah, every instance I think of using env variables involves at least one more abstraction layer
17:54 drawsmcgraw First pass at using env={'donkey': 'vampires'} looks like a no-go.
17:55 drawsmcgraw That's with the module anyway, not sure of the state. Either way, I'm going with a template.
17:58 Akhter joined #salt
18:00 iggy it's more "Salthonic" too
18:00 rdutch left #salt
18:02 aparsons joined #salt
18:03 JoeHazzers when i call  mine.get '*' 'alias', i get an empty response from all minions, even ensuring that mine_functions:alias is a dict in the pillar on the minions, and running mine.update
18:03 JoeHazzers what's up with that?
18:06 perfectsine joined #salt
18:06 pravka joined #salt
18:06 forrest iggy: salthonic??
18:06 * forrest facepalms
18:07 forrest I like this IRC urban salt dictionary, salthonic, saltines, etc.
18:08 tanta joined #salt
18:08 iggy JoeHazzers: what's alias?
18:08 sdm24 JoeHazzers: whats the output of salt '*' pillar.get mine_functions? did you refresh the pillar on the minions?
18:13 yanivtr joined #salt
18:15 zero_shane joined #salt
18:20 ahammond I'm changing the ID on some salt minions. I've updated the id: foo, wiped /var/cache/salt and restarted the minion. Generally, the minion restarts just fine. On those boxes, I see two salt-minion -d processes. On some boxes, I restart and see only a single process and the minion doesn't respond.
18:21 JoeHazzers iggy: alias in this case is vpn_ip_addrs
18:21 JoeHazzers sdm24: hrm, it seems i've had some misconceptions about the data structure. i'll come back when i've got my head around it. thanks.
18:22 rodr1c joined #salt
18:22 linjan joined #salt
18:23 iggy JoeHazzers: what's vpn_ip_addrs?
18:23 sdm24 JoeHazzers: it can definitely be confusing at times. 'pillar.get mine_functions' will just show what data (for mine_functions) is on the minions. It will just show if the minions are correctly getting the data from the master
18:23 ahammond I've run strace salt-minion and see that it's doing a wait4() on a pid. When I check the pid, it's a 2nd salt-minion pid. This is different from when I just run service salt-minion start in that there's a 2nd process
18:24 iggy sdm24: mine_functions can also be set in the minion config
18:24 rodr1c joined #salt
18:25 iggy so probably use config.get
18:25 sdm24 iggy: definitely, he just said that he was doing it through pillar. Always good to cover all the bases though
18:26 ahammond I'm going to try rebooting the box to see if that unblocks things... but this is really strange to me.
18:26 ahammond Anyone else had problems restarting a minion after changing it's id?
18:26 JoeHazzers iggy: yeah, vpn_ip_addrs is a dict for a mine_function in the manner shown in the docs, in the pillar
18:27 sdm24 ahammond: nope. When I do it, I just edit 'id: newname' in /etc/salt/minion, restart the minion, delete the old minion's key on the master (salt-key -d oldname), then salt-key -a newname
18:28 unholycrab joined #salt
18:28 ahammond sdm24 yeah, I did that, plus also wiping the cache. Sigh and grumble.
18:28 iggy JoeHazzers: maybe try gist'ing some of your code/config
18:28 iggy ahammond: does /etc/salt/minion_id exist?
18:29 sdm24 I though /etc/salt/minion id: takes priority over minion_id?
18:29 sdm24 thought*
18:29 JoeHazzers iggy: i think i figured it out. the pillar didn't seem to be refreshing on a highstate (or i was doing something wrong), but it seems to be working now.
18:29 JoeHazzers as usual, PEBKAC
18:29 spark_ joined #salt
18:30 sdm24 JoeHazzers: salt '*' saltutil.refresh_pillar will automatically refresh the pillar and update any changes you made
18:30 pm90_ joined #salt
18:30 JoeHazzers sdm24: yes, i did that, and mine_functions is now identical to what i expect
18:30 JoeHazzers cheers for helping me through my stupidity :3
18:31 sdm24 JoeHazzers: stupidity isn't an issue. Ignorance is ;)
18:31 unholycrab stupidity is more fun
18:31 JoeHazzers i swear i ran a highstate after i changed pillar files and changed some state, but obviously not
18:38 jeffspeff if i want to run a command on multiple systems, would i just comma separate them like 'salt foo1, foo2, foo3 test.ping' ?
18:38 wendall911 joined #salt
18:38 shaggy_surfer joined #salt
18:39 iggy no space
18:39 jeffspeff ty
18:40 pcn Is it possible to use archive.extracted with a salt path as the source?
18:40 zero_shane @jeffspeff - great salt document on targeting minions at:  http://docs.saltstack.com/en/latest/topics/targeting/index.html
18:41 zero_shane and specifically answering your question: http://docs.saltstack.com/en/latest/topics/targeting/globbing.html#lists
18:41 pcn Ah, never mind, I think I found my oops
18:42 jeffspeff ty zero_shane
18:44 baweaver joined #salt
18:47 bhosmer joined #salt
18:48 unholycrab i want to spin up an entire pipeline on AWS into a VPC, including EC2 instances, RDS, ELBs, subnets, etc. is Salt Cloud a good candidate for this?
18:48 unholycrab autoscaling groups?
18:49 unholycrab im researching this as an alternative to a pile of python scripts which create a cloudformation template
18:49 DanyC joined #salt
18:49 unholycrab s/template/stack
18:50 bluenemo joined #salt
18:52 DanyC joined #salt
18:52 DammitJim joined #salt
18:54 londo joined #salt
18:54 DammitJim joined #salt
18:54 pcn That wasn't my problem after all.  Can I use archive.extracted with a salt:// source?
18:55 iggy unholycrab: should be, maybe with some of the boto_* modules/states thrown in for whatever the salt-cloud maps don't support
18:55 iggy pcn: yes
18:56 sirex joined #salt
18:56 unholycrab iggy: i just skimmed your saltconf15 presentation on youtube :P
18:56 ahammond iggy turns out we have resolver issues on the box. go figure.
18:56 apinoworky joined #salt
18:58 apinoworky I was wondering if anyone could help me out with this error regarding the salt-api? "Not starting 'salt.loaded.int.netapi.rest_cherrypy'. Options 'ssl_crt' and 'ssl_key' are required if SSL is not disabled."
18:59 unholycrab iggy: can i use salt-cloud maps to describe a VPC ?
18:59 timoguin unholycrab: it's pretty good for spinning up instances, but there are some shortcomings.
18:59 timoguin and no, salt-cloud won't manage VPC things.
18:59 revjim joined #salt
19:00 timoguin I'm looking at switching over to just using auto-scaling groups with a reactor that will bootstrap the instances upon creation
19:00 DammitJim I would like to set up all my minions to use the ubuntu ppa for salt-minion
19:01 unholycrab has anyone compared it to troposphere?
19:01 revjim I've installed salt-formula. When I do "salt-ssh mynewhost state.sls salt.minion" however, I get an error about template files and such. TemplateNotFound: salt/formulas.jinja
19:01 revjim What did I miss?
19:01 iggy unholycrab: don't do that... (in my defense it was the conference I've ever been to, and the first one I've presented at... I'll be better next year)
19:01 iggy *was the first conference
19:02 unholycrab it was helpful to me, iggy
19:02 iggy revjim: usually means you didn't install the formula correctly... how did you "install" it?
19:03 timoguin unholycrab: the boto modules/states have good support for VPC, autoscaling groups, launch configs, RDS, etc.
19:03 timoguin An issue I ran into a few days ago is that it doesn't look like salt-cloud has any support for creating encrypted EBS volumes
19:03 revjim iggy: mkdir /srv/formulas; git clone salt-formula; vi /etc/salt/master add /srv/formulas/salt-formula to file_roots. restart salt-master
19:04 iggy revjim: ls /srv/formulas/salt-formula/salt/formulas.jinja
19:05 iggy revjim: also `salt-call cp.list_master | grep formulas.jinja`
19:05 revjim iggy: file is there.
19:05 revjim iggy: my salt master is not also a minion. so salt-call doesn't work.
19:06 pcn iggy: I think my problem is with archive.extracted.  If I point it to /opt, per the docs, it'll say "/opt/ already exists" and don't extract anything.
19:06 iggy that's odd, but okay... check it on a minion?
19:07 revjim iggy: chicken and egg problem. I have no minions. Haha. Let me make one the "manual" way and check it.
19:07 timoguin You can also try the fileserver runner on the master. salt-run fileserver.dir_list and fileserver.file_list
19:07 * iggy takes note
19:07 revjim timoguin: awesome...
19:08 revjim # salt-run fileserver.file_list | grep formulas.jinja
19:08 revjim - salt/formulas.jinja
19:08 revjim iggy: i did what timoguin said:
19:08 revjim iggy: so I think that means it sees the file.
19:08 * iggy really wants to rip out that formulas.jinja crap anyway
19:09 revjim let me paste the full error... maybe we just need more details iggy
19:09 iggy revjim: how are you running the state if you have no minions?
19:09 revjim iggy: salt-ssh... https://gist.github.com/dlashua/63cac5038fa998f027c3
19:10 * iggy runs
19:10 revjim iggy: hahaha! noooo.
19:10 revjim iggy: I have to minionize 30 hosts. I'm hoping to do it with a script instead of 30 times by hand. this was my "solution". it HAS to work. :)
19:11 pcn Ah, if_missing.
19:11 iggy revjim: saltify?
19:11 hamburglar joined #salt
19:11 mark_____ joined #salt
19:11 revjim iggy: yes. saltify. minionize is the wrong word. ha.
19:11 iggy I mean are you using that to do the minion?
19:12 revjim iggy: wha? that's a thing?
19:12 mark_____ hey guys, i'm using terraform to spin up aws instances and i'm spinning up N of each type. The thing i'm stuck on is hostnames, i'm using duplicate hostnames on the boxes. Can I get salt to identify a minion on something other than hostname?
19:12 revjim iggy: http://docs.saltstack.com/en/latest/ref/clouds/all/salt.cloud.clouds.saltify.html looking at this now.
19:12 iggy mark_____: you can set minion_id to whatever you want
19:13 revjim iggy: this looks interesting... but I'm having trouble finding info on how to use it. have any docs?
19:14 mark_____ cool thanks iggy i'll just write id: to minion using a python script as the box comes up to web-uuid or something
19:14 iggy revjim: not really in salt docs... there's some stuff out there though
19:15 iggy mark_____: or /etc/salt/minion_id
19:15 revjim iggy: I'm googling for it. thanks.
19:15 mark_____ even easier
19:15 mark_____ thank you
19:19 aron_kexp joined #salt
19:20 godlike what would be the correct way to apply a state or not, based on the timestamp of a file? (i.e. apply the state if the file is older than, say, 1 day) I am thinking way too many twisted things to accomplish this and wished to know if there was an 'acceptable' way
19:25 g3cko joined #salt
19:29 edrocks joined #salt
19:30 zsoftich3 so I figured out the http_proxy issue
19:30 zsoftich3 https://github.com/saltstack/salt/blob/develop/salt/utils/http.py#L328-L329
19:30 DanyC joined #salt
19:31 zsoftich3 I had an https key_url in pkgrepo.managed
19:32 zsoftich3 changing the key_url to an http source works just fine since it looks like urllib will use env *_proxy vars
19:33 spark_ joined #salt
19:35 forrest Has anyone seen this? https://gist.github.com/gravyboat/b3e528dde9b8bd629da5 on 2014.7.6
19:36 forrest I understand the error, but I don't understand why the error is occurring, it's just a pillar call to a value
19:39 dexter91x joined #salt
19:40 DammitJim how do I get salt to update a keyid for a ppa?
19:40 DammitJim it is still using the old one
19:40 baweaver joined #salt
19:44 iggy forrest: I think there's an issue open about that (it might not actually be "open" anymore
19:44 forrest noooo
19:47 forrest Hmm, I don't see an identical issue, the last closed on was for 2014.1.3
19:47 Akhter joined #salt
19:48 pdayton joined #salt
19:48 forrest iggy: This is a total blocker, and absolute shit. It just started randomly happening too
19:48 forrest Was fine up until two provisions ago
19:48 forrest same release of salt :|
19:49 KyleG joined #salt
19:49 KyleG joined #salt
19:49 iggy that line is shit... it should be checking password in user_attr first
19:49 zmalone joined #salt
19:49 iggy file a bug
19:51 forrest They're going to tell me to upgrade salt
19:51 forrest which I can't do because of other bugs in salt
19:51 forrest lol
19:52 iggy well, the line looks the same in develop
19:52 iggy so it's still a bug
19:53 iggy https://github.com/saltstack/salt/blob/develop/salt/states/postgres_user.py#L187
19:53 DammitJim what is the keyid I need to use to add the saltstack ppa for ubuntu trusty?
19:54 intel joined #salt
19:55 forrest DammitJim: Did you already look here? http://docs.saltstack.com/en/latest/topics/installation/ubuntu.html
19:55 DammitJim I did... but I am confused when creating the state to do this
19:56 DammitJim https://gist.github.com/anonymous/c2abcc33bfbe76a4011c
19:56 whytewolf DammitJim: the keyid "should" import with the ppa so no need for the keyid. but if you really want the info for the ppa you can find it here https://launchpad.net/~saltstack/+archive/ubuntu/salt
19:56 manfred DammitJim: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkgrepo.html look at the second example
19:57 DammitJim manfred, what about it? I used that as a template
19:57 DammitJim is the keyid the same for saltstack's ppa?
19:57 manfred no
19:57 manfred the keyid is the signing key id number
19:57 DammitJim then what keyid do I use?
19:57 iggy ppa's don't require a key to be set
19:57 manfred the one in the gpg keyserver
19:58 DammitJim iggy, so, I don't need a keyid nor a keyserver?
19:58 theologian joined #salt
19:58 manfred 0E27C0A6
19:58 manfred so yo should already have the right one
19:59 iggy DammitJim: no
19:59 manfred you could also just use the ppa example and not spell it out like you are
19:59 manfred base:
19:59 manfred pkgrepo.managed:
20:00 whytewolf DammitJim: you have added ppa's to ubuntu before right? you need about as much info as you do when adding a ppa
20:00 manfred - ppa: saltstack/salt
20:00 manfred or whatever the ppa name is
20:00 DammitJim I have, I don't know why I am getting confused with these examples
20:00 manfred DammitJim: look at the third example on the page where they directly just add the ppa, instead of spelling out the full repo information
20:01 DammitJim ok
20:02 mnaser hi everyone, i have a few nodes managed with saltstack... i want to have a few node specific pillars, whats the best way going about this?
20:02 DammitJim oh, I see
20:02 DammitJim is this one of those things where the protocol being used uses a different port other than 80?
20:03 manfred it shouldn't be anything different
20:03 iggy DammitJim: no, ppa: is a special identifier that tells salt to lookup all the data at launchpad
20:03 manfred apt-add-repository just knows where ppa's are located
20:04 ahammond what is the best way to test a salt-minion's connectivity to the master from the minion? I'd like something for zabbix or even monit to check.
20:07 forrest Check the outbound netstat for an established connection to the master.
20:08 iggy cp.cache_file ?
20:10 whytewolf forgetting about useing the minion. and use salt-run manage.up from the master to get a list of minions that are currently up and known to it?
20:11 zsoftich3 manage.down be more appropriate?
20:11 sdm24 manage.status to see both?
20:11 bhosmer joined #salt
20:12 iggy or cp.get_file{,_str}
20:13 iggy but yeah, it's far easier for the master to tell you about everything at once
20:14 forrest iggy: https://github.com/saltstack/salt/issues/26616
20:14 saltstackbot [#26616]title: postgresql_user state generates a bool error | I get the following error:...
20:16 spark_ joined #salt
20:18 iggy forrest: now, copy that module into your _modules, add a check for password in user_attr and move on with life
20:19 forrest lol
20:19 forrest iggy: That's not really feasible with this setup
20:19 forrest good idea though
20:20 shaggy_surfer joined #salt
20:22 DammitJim https://gist.github.com/anonymous/5a9813bc365df4b15481
20:23 DammitJim that's what I get when I try to do the addition of the ppa
20:23 DammitJim that's why I was asking if it retrieves the key through a different tcp port
20:23 square1 joined #salt
20:23 DammitJim I ran into this before (I think) and someone on this channel said something like that
20:24 opensource_ninja joined #salt
20:24 whytewolf port 443?
20:24 manfred should be just 80 or 443
20:26 DanyC joined #salt
20:26 manfred though it appears that some of them also have 11371 open
20:27 manfred http://keyserver.ubuntu.com:11371/
20:27 square1 Good evening all, I've never used Salt before but having come across puppet in my new job - when doing some research i found salt which I seem to prefer over puppet - it just feels right ;)  - now having never used either, I bought two new servers and wondered whether I could use it to set up my new servers or whether I should set up my new servers first then configure salt afterwards?
20:27 manfred but i don't know if it redirects or not, cause that appears the same as http://keyserver.ubuntu.com/
20:27 DammitJim this is more of a Ubuntu question probably, but is there a way I can just download the key file and source it through salt?
20:28 zwi joined #salt
20:31 whytewolf DammitJim: of coarse. you don't have to use the ppa. ppa is just auto config stuff for manually adding it. all the info you need to add the repo is here https://launchpad.net/~saltstack/+archive/ubuntu/salt
20:32 whytewolf although it does look like port 11371 is the gpg port used by launchpad
20:33 whytewolf keyfile can be found here http://keyserver.ubuntu.com/pks/lookup?fingerprint=on&amp;op=index&amp;search=0x7A82B743B9B8E46F12C733FA4759FA960E27C0A6
20:33 iggy square1: depends what you mean by set up your new servers... I work all in the cloud currently, but our process is basically: deploy base instance, deploy salt-minion, make salt do the rest
20:34 square1 that'
20:34 square1 woops.
20:34 square1 that's exactly what I want to do
20:34 square1 but I don't really have a 'base' yet as i just have two new servers
20:35 bhosmer joined #salt
20:35 square1 seems I may need a salt master server first though
20:35 square1 I'm just playing about, none of this is critical
20:35 whytewolf square1: salt can be done masterless.
20:35 iggy you could use masterless
20:35 whytewolf square1: http://docs.saltstack.com/en/latest/topics/tutorials/quickstart.html
20:35 iggy salt-ssh is another option
20:36 iggy DammitJim: can you paste the state stanza that you have in that gist?
20:37 keimlink joined #salt
20:37 square1 I'll have a read, thanks whytewolf and iggy
20:42 jeadre joined #salt
20:43 pdayton1 joined #salt
20:44 DammitJim iggy, https://gist.github.com/anonymous/b317ce652f78a8530b86
20:44 iggy DammitJim: you have a locked down network?
20:45 DammitJim yes
20:45 DammitJim that's why I'll probably should just get the key file and pass it to the minions
20:45 iggy proxy server?
20:45 DammitJim no, just a firewall thing
20:46 DammitJim these servers can only do http out
20:47 iggy then yeah, key_url with a salt:// url is probably your best bet
20:51 spark_ Hi, can someone help me with redis job cache with salt runner??? I'm trying to set up only specific jobs to store job result to redis, but figured runner doesn't accept --return option
20:52 Arbusto joined #salt
20:52 SheetiS joined #salt
20:52 jeadre joined #salt
20:52 DammitJim how do I download the key file, though?
20:55 Arbusto In a multi master setup, can I have a shared external job cache so I don't lose a job when a master goes down?
20:57 zsoftich3 DammitJim: you can do apt-key list; apt-key export <keyid> if you already have the repo key added
20:58 tsultrim joined #salt
21:00 jeadre joined #salt
21:00 DammitJim zsoftich, so I just export it to a file
21:01 DammitJim put it in the master, and do a file.manage for it on the minion?
21:02 zsoftich3 DammitJim: key_url: salt://some_file.gpg
21:02 DammitJim oh cool
21:02 zsoftich3 ^^ in pkgrepo
21:02 troyready joined #salt
21:04 baweaver joined #salt
21:05 ajw0100 joined #salt
21:11 arthoo joined #salt
21:11 jeadre joined #salt
21:11 tsultrim Has anyone else had an issue where a state file fails (regardless of content) but you can still execute commands via cli?
21:12 tsultrim so `salt-ssh` commands work fine, but state files are returning "Failed to return clean data." Any ideas?
21:13 arthoo hi, I'm trying to a simple replace with salt but can't get the syntax right.  let's say I'm trying to replace "foo" with "bar" in /tmp/foobar, I was trying to do a 'salt \* /tmp/foobar "(foo, bar)"' but I'm obviously not doing something right.
21:14 jeffspeff joined #salt
21:16 arthoo er, 'salt \* file.replace /tmp/foobar "(foo,bar)"' that is
21:17 DammitJim why would I get this on my master? Minion did not return. [No response]
21:17 iggy arthoo: salt '*' file.replace /path/to/file pattern='foo' repl='bar'
21:17 DammitJim yet, the minion was still doing stuff
21:17 aurynn DammitJim, Because the minion didn't return before the timeout
21:18 iggy arthoo: like the docs say
21:18 aurynn you can keep checking on it via the JID
21:18 DammitJim hhmmmm... never happened before
21:18 sdm24 DammitJim: I get that too sometimes, I'm not sure exactly why, but it will still accept commands
21:18 sdm24 well now I know why, ty aurynn
21:18 arthoo thanks, iggy
21:18 iggy it could be timeout related, or if you have something that restarts the minion it can happen as well
21:18 johnkeates joined #salt
21:19 iggy (i.e. a state that restarts the minion)
21:19 johnkeates left #salt
21:21 subsignal joined #salt
21:22 teryx510 coval3nce: thanks for the jinja tip yesterday. That was exactly what I needed.
21:26 rideh joined #salt
21:27 DammitJim iggy, you nailed
21:27 DammitJim it
21:27 DammitJim it's restarting the minion since it is updating the package LOL
21:28 Arbusto joined #salt
21:31 baweaver joined #salt
21:33 diegows joined #salt
21:44 rideh joined #salt
21:46 hamburglar joined #salt
21:48 shiriru joined #salt
21:49 zero_shane joined #salt
21:49 gazarsgo how do i use --pillar-root with `salt-call --local` ? i don't see much docs around salt local usage :(
21:50 gazarsgo `--log-level debug` also gives me no clue as to what pillar-root is doing :(
21:51 hamburglar I am new to salt and am working on learning what I need to begin porting over from puppet.  Puppet has templates that can be used to generate dynamic content.  I am looking for the salt way to do something similar with grain and pillar data. When looking at templates in salt, it seems like its more for conditionals within a state.  What is the terminology that I should be looking for instead of template?
21:53 iggy hamburglar: it can be used either way... one common way you see things done is a state with file.managed with the context arg set to a dict, then the jinja template fills in it's options based on that dict
21:53 iggy or templated files directly pulling in pillars
21:53 SheetiS1 joined #salt
21:53 iggy looping over pillar lists
21:53 iggy etc...
21:54 cliluw joined #salt
21:54 hamburglar iggy: ah, interesting.  I was looking at http://docs.saltstack.com/en/latest/topics/tutorials/states_pt3.html and it led me to believe that it was only for loops/conditionals in a state.  I'll have to go mess with it a bit.  Thanks!
21:55 iggy that's templating sls modules
21:55 theologian joined #salt
21:56 pdayton joined #salt
21:57 hamburglar iggy: ah, perfect! I found an example. I guess I was on the right track after all. http://www.pythian.com/blog/managing-files-with-saltstack/ thanks again!@
22:00 gazarsgo is top.sls supposed to be the entry point for pillar data ? are the variables there supposed to come from /etc/salt/grains ?
22:00 GreatSnoopy joined #salt
22:00 pdayton1 joined #salt
22:02 iggy gazarsgo: it targets the data
22:03 gazarsgo i don't know what 'targets the data' means ... would you say top.sls is a mapping between grains and pillar data ?
22:03 iggy not necessarily grains, but yes
22:04 hamburglar gazarsgo: pillars can contain the data too.  its just a yaml list of values and those values could be from grains.  at least thats my understanding of it so far.
22:04 iggy you can map hostnames, regexes, compound matches, etc
22:05 chamunks joined #salt
22:07 hamburglar gazarsgo: this example helped me understand how to use pillars  https://gist.github.com/UtahDave/3785738
22:07 ujjain joined #salt
22:07 ujjain joined #salt
22:09 gazarsgo iggy: what do you call any of the things you can map on besides grains? grains are both automatically supplied by salt (hostname, cpu_flags, etc) or user supplied (/etc/salt/grains) ... or are the automatic things something els e?
22:09 jgelens joined #salt
22:09 iggy hostname is probably the most popular
22:09 gazarsgo no i mean hostname is still a grain ?
22:10 iggy http://docs.saltstack.com/en/stage/topics/targeting/index.html
22:10 iggy it technically is, but that's not the simplest way to match hostnames
22:12 saltstackbot [reddit-saltstack] Reading a plain text file into a template? https://www.reddit.com/r/saltstack/comments/3i99cr/reading_a_plain_text_file_into_a_template/ - 2015-08-24 - 23:10:14
22:13 pm90__ joined #salt
22:15 pm90__ joined #salt
22:21 coval3nce teryx510: no worries
22:22 baweaver joined #salt
22:24 Zachary_DuBois joined #salt
22:28 forrest iggy: So that error from earlier isn't always present
22:28 forrest iggy: I'm not sure what exactly causes it, but I built a brand new VM in a clean repo, and it does not show up
22:30 jgelens joined #salt
22:30 iggy kill with fire?
22:31 forrest I guess, it's still a bug either wy
22:31 forrest *way
22:40 pdayton joined #salt
22:42 pdayton1 joined #salt
22:44 adelcast joined #salt
22:48 coval3nce how do you add/modify a grain from the master again?
22:50 sdm24 http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.grains.html?
22:51 coval3nce ah damn the module, thanks
22:51 sdm24 np
22:53 rabbai joined #salt
22:53 rabbai hello all, hope you all having a good day
22:55 forrest Can't complain, what's up?
22:57 jgelens joined #salt
23:00 forrest rabbai: If you have a question you can just ask, a lot of people have gone home for the day
23:01 rabbai oooh my bad sorry
23:01 rabbai I am trying to setup a paas to use with soft layer cloud and someone pointed me to saltstack
23:01 rabbai any success with that?
23:03 forrest I know that salt has support for softlayer: http://docs.saltstack.com/en/latest/topics/cloud/softlayer.html
23:03 sbogg joined #salt
23:04 forrest rabbai: And no worries, past that stuff for the cloud, as far as I'm aware they're just a standard VPS host so using salt there shouldn't be a problem.
23:05 forrest And they did a talk at saltconf
23:05 rabbai ahhh ok
23:05 forrest That's all I really know, I don't have any examples using them, sorry.
23:05 rabbai now is salt opensource or enterprice .... I see salt community ... just heard of it for the first time today
23:06 forrest rabbai: Salt is open source. The enterprise aspect is support, and you can get an enterprise dashboard, though there is a team writing an open source dashboard as well.
23:06 rabbai awesome
23:07 forrest Yup\
23:07 rabbai Thanks for the pointers ... i appreciate it
23:08 forrest Yeah np, you might want to try out the new getting started guide: http://docs.saltstack.com/en/getstarted/ , and if you come across anything that is confusing or not clearly explained as a new user, please create an issue on GitHub.
23:09 rabbai ok cool I will do just that
23:09 rabbai already looking at it now
23:09 forrest Sweet!
23:09 jgelens joined #salt
23:12 schristensen joined #salt
23:17 sgargan joined #salt
23:17 forrest rabbai: Are you good to go for a while with regards to general questions? I am planning on heading out.
23:18 rabbai yeah I am good to go
23:18 ronrib joined #salt
23:18 rabbai I will do my research and ping you if I have issues
23:18 rabbai Thanks alot
23:18 XenophonF left #salt
23:18 forrest NP best of luck, feel free to just ask in the channel, usually people have time to help.
23:20 twork i have an ftp server, where all the home directories need to have the same gid. i've been back and forth through the 'file' and 'user' states and can't figure out how to make that work.
23:20 twork i know the 'file' state has a 'directory' option but don't know how to make it match a wild card, that can come and go with other states...
23:21 iggy do it the old fashioned way
23:21 twork shell?
23:21 robawt why would you match a wildcard?
23:22 twork robawt: because i don't care who owns the directory, i want the gid to be the same
23:22 robawt either manage with file that requires the user to exist or set the user's primary group
23:22 twork yeah, can't set the primary gid
23:22 twork then the user will access other directories
23:23 kevinquinnyo joined #salt
23:23 robawt so you want the users to have same GIDs but not allowed to access the other directories?
23:23 twork no
23:24 twork i want all the user accounts to have their own individual uid/gid. i want the home dir.s to have the user's uid, and a gid that no user is a member of
23:24 twork did that make sense?
23:25 twork then, an automatic account will have write perm's in those directories, which have perms 0770
23:26 twork ...there has to be a better way to accomplish what i want but i'm in a hurry. :\
23:28 twork y'know, drop this. initially i just had my daemon's account be a member of all those usrs' gid, then my boss did a chgrp on all those directories and i think he may have been up the wrong tree
23:29 jeffspeff joined #salt
23:30 chequers left #salt
23:30 twork and my clock is running off anyhow. thanks guys, more babble tomorrow no doubt.
23:30 robawt twork: your boss was up the wrong tree
23:31 robawt your initial idea of the daemon being a member, or a member of a common group is more right
23:33 theologian joined #salt
23:33 shiriru joined #salt
23:33 timoguin joined #salt
23:35 peno joined #salt
23:36 edulix joined #salt
23:36 ws2k3_ joined #salt
23:36 fhh joined #salt
23:37 mosen joined #salt
23:37 Rockj joined #salt
23:38 wiqd joined #salt
23:38 SteamWells joined #salt
23:40 ajw0100 joined #salt
23:40 godlike joined #salt
23:40 godlike joined #salt
23:42 peno joined #salt
23:44 ws2k3_ joined #salt
23:44 aristedes joined #salt
23:44 aristedes left #salt
23:52 c00p3r joined #salt
23:53 aqua^c joined #salt
23:54 t730 joined #salt
23:56 bfoxwell joined #salt

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