Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-12-20

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

All times shown according to UTC.

Time Nick Message
00:00 sroegner_ joined #salt
00:00 KyleG JordanRinke: You must practice thine google fu
00:00 KyleG wow so results, such search.
00:02 scoates wow.
00:05 KyleG scoates: I have that affect sometimes.
00:06 fxhp joined #salt
00:09 honestly eah
00:09 honestly augh
00:10 kermit joined #salt
00:26 bhosmer joined #salt
00:26 cachedout joined #salt
00:34 Sheco joined #salt
00:35 quickdry21 joined #salt
00:55 goncalo_ joined #salt
00:55 davet joined #salt
00:58 alunduil joined #salt
01:00 cnelsonsic joined #salt
01:02 anitak joined #salt
01:09 dstanek joined #salt
01:13 BenCoinanke joined #salt
01:21 fllr joined #salt
01:23 Chrisje joined #salt
01:45 fllr joined #salt
01:47 stefanmonkey joined #salt
01:47 swa_work joined #salt
01:57 favadi joined #salt
01:57 fxdgear any advice on manually building/installing packages from a tar.gz file?
01:58 fxdgear aside from just doing cmd.run "make" , etc...
01:58 honestly fxdgear: build your own salt runner
01:58 honestly or salt state
01:59 honestly a 'built' state that you can give a file and configure options and make options would be pretty snazzy
01:59 fxdgear would be! ;)
02:02 higgs001 joined #salt
02:03 anitak joined #salt
02:09 stantont joined #salt
02:12 jalbretsen joined #salt
02:16 whiteinge fxdgear: here's an example of how to do something very simple:
02:16 whiteinge https://github.com/saltstack-formulas/redis-formula/blob/master/redis/common.sls
02:17 fxdgear whiteinge +1 thanks!
02:18 cachedout joined #salt
02:22 xl1 joined #salt
02:24 RoboTeddy joined #salt
02:25 RoboTeddy is there a way to pin the version of salt-minion that's bootstrapped on machines brought up by salt-cloud?
02:25 RoboTeddy right now it seems like every time salt upgrades, and I want to bring up a new box, I _have_ to upgrade salt-masters+salt-minions if I want to use salt-cloud
02:31 jergerber joined #salt
02:33 fllr joined #salt
02:33 fllr joined #salt
02:34 mgw joined #salt
02:35 jcsp joined #salt
02:45 elithrar Anyone familar with the iptables module on Ubuntu 12.04? (iptables v1.4.12 w/ salt 0.17.2)
02:47 elithrar Looks similar to this but not the same Salt version (not sure if 0.17.2 fixed the issue) https://github.com/saltstack/salt/issues/8419
03:02 BenCoina- joined #salt
03:03 anitak joined #salt
03:05 rojem joined #salt
03:14 dwfreed elithrar: it's fixed in 0.17.3 and 0.17.4
03:14 dwfreed elithrar: 0.17.4 was released last Friday
03:15 austin987 joined #salt
03:25 smccarthy joined #salt
03:28 malinoff joined #salt
03:31 jdenning joined #salt
03:34 dstanek joined #salt
03:41 elithrar dwfreed: Ah, great. For some reason I'm still on 0.17.2 (via Vagrant & Packer) but can't see if/why they are pinning the version
03:45 favadi joined #salt
03:50 elithrar hm, for some reason it looks like bootstrap.saltstack.com is still installing 0.17.2
03:52 fxdgear whiteinge you still around?
03:55 higgs001 joined #salt
03:56 quickdry21 joined #salt
03:58 diegows joined #salt
03:58 kossy joined #salt
03:59 fxdgear so I have this state file: https://gist.github.com/fxdgear/8d8f83dbceefe940093a and i"m trying to configure/make/install openssh...
03:59 fxdgear but the configure commnd will run
03:59 fxdgear but then the make and make install commands faile...
03:59 fxdgear fail
03:59 fxdgear I'm wondering if the reasoning is cause the configure is taking too long...
04:00 jcsp joined #salt
04:02 kossy fxdgear: why aren't you using a package? pulling down a pkg {.deb,.rpm,etc} might work better
04:03 fxdgear @kossy cause due to client contractual obligations I need a version that's not avail for ubuntu 12.04 via apt... :/
04:03 shadowsun fxdgear: -t 99999
04:03 shadowsun fxdgear: But, what's the end of configure / what's make's error
04:03 kossy you can make your own package
04:04 kossy fxdgear: http://askubuntu.com/questions/146343/how-to-create-a-deb-package-that-installs-a-series-of-files
04:04 fxdgear shadowsun https://gist.github.com/anonymous/633376ce7a7c6d84d68c
04:04 fxdgear but the configure runs just fine...
04:04 shadowsun fxdgear: configure did not complete.
04:04 anitak joined #salt
04:04 shadowsun fxdgear: also, does your cmd all set the same cwd?
04:05 shadowsun fxdgear: Honestly it's easier to write a script that does it and run that on the remote server
04:05 shadowsun fxdgear: That being said, kossy's solution is actually correct
04:05 fxdgear @shadowsun the configure is completing: https://gist.github.com/anonymous/049754a60b1437d93908
04:05 fxdgear this is is in the output of running my state
04:05 MK_FG joined #salt
04:05 MK_FG joined #salt
04:05 shadowsun fxdgear: Then subsequent commands are run in the directory where Makefile does not exist
04:06 shadowsun no target means the Makefile is either 1) missing or 2) incomplete
04:06 Gareth w 44
04:06 Gareth erm
04:06 shadowsun wd40!
04:06 shadowsun that'll fix it!
04:06 shadowsun Thanks Gareth
04:06 fxdgear shadowsun yeah the configure/make/make install are all run in the same dir
04:06 fxdgear https://gist.github.com/fxdgear/8d8f83dbceefe940093a#file-openssh-sls-L38-L41
04:06 shadowsun fxdgear: Then why is your Makefile missing or incomplete :)
04:06 fxdgear well
04:07 fxdgear I can can run make
04:07 fxdgear AFTER the state fales
04:07 fxdgear and it makes just fine
04:07 fxdgear which makes me think there's a timing issue...
04:07 shadowsun yes
04:07 shadowsun there could be
04:07 shadowsun You want to put that as three blocks which watch the previous block
04:08 fxdgear ok... I'll give that a shot... then look into kossy's suggestion
04:08 shadowsun or - again - a script that does it
04:08 shadowsun which will (frankly) be more reliable
04:08 shadowsun but Kossy's solution is correct because you can easily deploy it to more than one box ;)
04:09 fxdgear __which__ I will have to do :)
04:10 kossy fxdgear: its more logical to build it and then deploy a binary. saves time not having to compile it over and over.
04:10 fxdgear yeah no that does make sense
04:11 fxdgear we have to have openssh >= 6.2 and when i was looking at ubuntu's repos... that was only avail with ubuntu 13.10 and up
04:11 fxdgear and we're on 12.04 (lts reasons)
04:11 kossy yeah i get that. i roll my own packages for almost everything because of that reason
04:12 jacksontj joined #salt
04:16 fllr joined #salt
04:16 sroegner_ joined #salt
04:19 Dinde joined #salt
04:20 jfzhu_us joined #salt
04:32 nlb joined #salt
04:33 nlb hey folks, question about syntax for grains inside a state
04:33 nlb I'm trying to do something like this:
04:33 nlb http://pastebin.com/4dA36qAJ
04:33 nlb but getting a compile error
04:38 bemehow_ joined #salt
04:40 pcurry_nomi joined #salt
04:49 ravibhure joined #salt
04:50 anuvrat joined #salt
04:51 ravibhure1 joined #salt
04:52 Dinde joined #salt
04:57 sroegner_ joined #salt
05:03 davet joined #salt
05:05 anitak joined #salt
05:13 ajw0100 joined #salt
05:15 elfixit joined #salt
05:15 forresta joined #salt
05:20 pcurry_nomi joined #salt
05:43 ajw0100 joined #salt
05:47 dstanek joined #salt
06:05 anitak joined #salt
06:39 sashka_ua joined #salt
06:42 sunand joined #salt
06:56 jalbretsen joined #salt
07:08 junedm joined #salt
07:10 fllr joined #salt
07:17 jfzhu_us Can you call salt modules from within a states file?
07:20 rojem joined #salt
07:21 wasd_ jfzhu_us, you can require a state to have being run prior to another.
07:23 jfzhu_us wasd_: I am confused by your answer. Perhaps I should've made clearer my intention. I am trying to call a salt module (specifically user.list_users) in order to dynamically generate a states file based on that information. I am not sure how your answer relates to that.
07:24 malinoff jfzhu_us, {% set users = salt['user.list_users']() %}
07:25 jfzhu_us malinoff: Thanks. I couldn't seem to find the documentation for this. Got any keywords I can use in google?
07:25 malinoff jfzhu_us, http://docs.saltstack.com/ref/states/vars.html
07:26 malinoff jfzhu_us, I'm not sure you can google it. So diving in docs is the best idea
07:26 jfzhu_us malinoff: Will do. thanks for the advice.
07:26 malinoff jfzhu_us, My pleasure
07:26 gammalget joined #salt
07:27 wasd_ jfzhu_us, perhaps I am still fuzzy about the difference between modules, states and functions.
07:28 jfzhu_us wasd_: You wouldn't be the only one ;-). I have shot myself quite a few times in the foot before i even realized that modules were what i was looking for.
07:29 malinoff Module is an execution concept. When you run a command via cli, you execute a module, each time. State is a configuration management concept. You can think about states like it is a sequence of actions that should be applied on a minion
07:30 malinoff So let's say you have 'pkg' module and 'pkg' state
07:30 malinoff 'pkg' module should be used this way: "install a package"
07:30 malinoff 'pkg' state should be used this way: "ensure a package is installed"
07:31 malinoff Function is python concept
07:31 malinoff Each function describes an action in a module or in a state
07:32 malinoff In the example above, you can run 'salt \* pkg.install nginx'
07:32 malinoff pkg is execution module
07:32 malinoff and install is python's function in this module
07:32 malinoff If you have such state:
07:32 malinoff Nginx installation:
07:33 malinoff - pkg.installed:
07:33 malinoff - name: nginx
07:33 malinoff pkg is state module
07:33 malinoff and installed is python's function in this module
07:34 malinoff So basically, when you run something via the cli - you run an execution module
07:34 malinoff When you write a state - you run a state module
07:34 wasd_ Thanks, malinoff! You made it clear, now I need to practice, practice, practice :)
07:34 malinoff salt \* state.sls yourstate - here state is execution module, sls is python's function
07:35 malinoff wasd_, I know by myself salt's concepts are difficult to understand :)
07:35 malinoff For a newbie, of course
07:36 wasd_ Guess I should start using salt to remote execute stuff more. I started a couple of day ago and went straight into using it as a configuration manager for my cluster
07:37 techdragon joined #salt
07:38 malinoff wasd_, I like salt right because of it's configuration management abilities. I can run any command I need via ssh, so remote execution stuff is not very important for me :)
07:44 wasd_ I use pdsh to push commands to all machines, but I guess that when I get every machine expressed in salt using salt to bulk remote execute will be better. Specially, being able to filter by grain will be invaluable.
07:44 jfzhu_us malinoff: Based on your explanation it would appear that a state module has significantly more logic than an execution model since one of the requirements is that a state must be ensured
07:47 malinoff jfzhu_us, yeah. 'pkg.install' execution will just try to install a package - and silently fail if any error will occur. And 'pkg.installed' state module will get already installed packages at first, check that the package is not installed already, try to install and check that this package installs
07:52 higgs001 joined #salt
07:56 eliasp joined #salt
07:56 helminthe joined #salt
07:58 netzmonster joined #salt
07:59 jfzhu_us malinoff: The remote execution stuffs is what actually drew me into puppet, chef, and salt but I didn't realize at the time that it was an entirely separate concept. Being able to ping all my vms on a unstable vmware infrastructure was very appealing for me at the time I learned about this software but now that I actually got my hands on it. I sitll don't see the distinction between the execution concept vs the state concept since they both either succeed or fail
07:59 balboah joined #salt
08:03 jfzhu_us Or maybe I am thinking about it wrong and state is just a subset of remote execution
08:04 malinoff jfzhu_us, For example, difference between exec module and state module is that exec module operates with imperative actions, and state module operates with declarative actions
08:05 malinoff This is the most significant difference, btw
08:05 jfzhu_us malinoff: Ah that clarifies it perfectly. Ty
08:06 anitak joined #salt
08:08 Nexpro joined #salt
08:11 gammalget joined #salt
08:12 harobed joined #salt
08:15 MrTango joined #salt
08:17 mr_chris joined #salt
08:33 bhosmer joined #salt
08:37 giantlock_ joined #salt
08:40 gammalget joined #salt
08:40 fllr joined #salt
08:52 ConceitedCode joined #salt
08:56 gammalget joined #salt
08:57 middleman_ joined #salt
08:58 anuvrat joined #salt
09:03 helminthe joined #salt
09:05 carlos joined #salt
09:07 anitak joined #salt
09:17 bemehow joined #salt
09:18 tonthon joined #salt
09:38 junedm left #salt
09:40 fllr joined #salt
09:51 anitak joined #salt
10:01 mr_chris joined #salt
10:02 pengunix joined #salt
10:05 joehh joined #salt
10:05 sroegner_ joined #salt
10:09 psyl0n joined #salt
10:10 fllr joined #salt
10:24 bhosmer joined #salt
11:00 scott_w joined #salt
11:10 fllr joined #salt
11:10 diegows joined #salt
11:12 Whissi joined #salt
11:14 joehh packages for 0.17.4 on ubuntu now in ppa
11:20 mortis_ joehh: \o/ sweet
11:23 jdenning joined #salt
11:27 Furao joined #salt
11:54 malinoff joehh, How about debian repo? :)
12:01 cowyn_ joined #salt
12:04 Psi-Jack joined #salt
12:05 ahammond joined #salt
12:05 djinni` joined #salt
12:10 fllr joined #salt
12:11 Hefeweizen joined #salt
12:12 fllr joined #salt
12:14 joehh malinoff: uploading now
12:15 joehh not that you are here :)
12:20 joehh packages for 0.17.4 on debian now at debian.saltstack.com (and in debian sid)
12:21 Sheco joined #salt
12:21 middleman_ joined #salt
12:39 viq joehh: sweet :)
12:39 viq Anyone managing snmp (users, communities and such) with salt, especially snmp-v3 users? How are you doing it?
13:02 harobed joined #salt
13:05 elithrar joined #salt
13:10 fllr joined #salt
13:13 ajw0100 joined #salt
13:20 jslatts joined #salt
13:21 blee joined #salt
13:23 rojem joined #salt
13:25 junedm joined #salt
13:25 bvd_ joined #salt
13:26 fllr joined #salt
13:27 _fllr_ joined #salt
13:29 yota joined #salt
13:31 xinkeT joined #salt
13:32 brianhicks joined #salt
13:32 JasonG_TA joined #salt
13:34 patrek joined #salt
13:35 cachedout joined #salt
13:43 des__ joined #salt
13:43 des__ hi
13:44 des__ here something strange
13:44 des__ cat top.sls base:   'ubnt-grid-(11|12)-*':     - selenium-node     - last.browsers
13:44 des__ oops
13:44 des__ top.sls
13:44 des__ base:
13:44 des__ 'ubnt-grid-(11|12)-*':
13:44 des__ - last.browsers
13:45 des__ minion id is - ubnt-grid-11-1387559684
13:45 th3reverend joined #salt
13:45 des__ but if i type salt-call state.highstate
13:45 des__ No Top file or external nodes data matches found
13:47 des__ and no logs in master log
13:48 scuwolf joined #salt
13:50 bvd_ Hi des.  Is that top.sls in the right path for file_roots?
13:51 des__ yes
13:51 des__ if i change ubnt-grid-(11|12)-* to ubnt-grid-11-*
13:52 des__ all works fine
13:52 Gifflen joined #salt
13:52 fxhp des__: http://docs.saltstack.com/topics/targeting/compound.html
13:52 viq des__: oh, glob vs pcre matching
13:52 fxhp basically you are doing it wrong
13:52 y0j joined #salt
13:53 fxhp """The default matcher is a glob match"""
13:53 fxhp if you want to use a different matcher
13:54 fxhp You need to use the compound matcher statement
13:55 des__ 'E@ubnt-grid-(11|12)-*':
13:55 des__ don`t work too
13:56 pass_by_value joined #salt
13:57 oz_akan_ joined #salt
13:57 ange joined #salt
13:57 jrb28 joined #salt
13:57 viq des__: - match: pcre
13:58 des__ hm
13:58 viq des__: http://docs.saltstack.com/topics/targeting/globbing.html end of page
13:59 viq The end of that page has exactly what you're trying to do.
13:59 oz_akan_ joined #salt
13:59 mgw joined #salt
14:00 des__ viq thank you!
14:00 des__ viq: thank you!
14:01 Gifflen joined #salt
14:04 vejdmn joined #salt
14:06 junedm joined #salt
14:07 junedm left #salt
14:10 tdilly joined #salt
14:10 fllr joined #salt
14:15 shiin joined #salt
14:16 juicer2 joined #salt
14:16 shiin is there a way to pkg.install without debian recommends?
14:16 che-arne joined #salt
14:18 GothAck joined #salt
14:19 des__ you can use APT::Install-Recommends "false";
14:20 GothAck Afternoon saltees
14:21 nocturn joined #salt
14:21 * pass_by_value waves
14:21 shiin des__: thanks
14:21 * GothAck waves back
14:22 GothAck Seeing a couple of exceptions in 0.17.2, not sure if anyone else has seen the, google / issue search didn't come up with much :/ basically salt-call doesn't return, seemingly because a socket doesn't get closed:
14:23 GothAck Exception TypeError: 'an integer is required' in <bound method SREQ.__del__ of <salt.payload.SREQ object at 0x3511c10>> ignored and Exception TypeError: "'NoneType' object is not callable" in <zmq.core.socket.Socket object at 0x36b4d50> ignored
14:23 shiin des__: that turns it off globally?
14:23 des__ yes
14:23 des__ shiin: yes
14:24 GothAck Anyone seen similar before?
14:24 shiin des__: do you think thats generally better than having it on for convenience?
14:25 shiin GothAck: what state is that in?
14:25 pcurry_nomi joined #salt
14:26 des__ shiin: yes
14:27 JasonSwindle joined #salt
14:27 shiin I shall contemplate that.
14:27 GothAck shiin: state.sls performing some yum/pip installs & adding configs for salt itself
14:28 stantont joined #salt
14:28 smccarthy joined #salt
14:28 pass_by_value I just checked out 0.17.2 (I got a warning) but my salt-call to test.ping did return.
14:29 GothAck test.ping doesn't get a warning at all for me :/
14:29 JasonSwindle left #salt
14:30 GothAck seems to leave a socket open to 4506 during some state.sls runs :/
14:30 GothAck s/during/after/
14:33 racooper joined #salt
14:33 rojem joined #salt
14:34 ggoZ joined #salt
14:34 pass_by_value afaik, and please correct me if I am wrong, the salt master will bind to both 4505 and 4506
14:34 rojem joined #salt
14:35 pass_by_value GothAck what os are u running?
14:35 scoates joined #salt
14:35 sroegner_ joined #salt
14:35 GothAck running centos 6.4
14:36 GothAck yeah, salt-master is bound to the two, but salt-call creates a connection to the master and leaves the one to 4506 open
14:37 GothAck tis a little bizarre, going to try packaging 0.17.4 to our internal repo and see if I hit the same issue
14:41 blee joined #salt
14:42 toastedpenguin joined #salt
14:45 pengunix joined #salt
14:48 anuvrat joined #salt
14:49 mpanetta joined #salt
14:52 sroegner_ joined #salt
14:58 viq What do you use to manage snmp, especially the v3 users, using configuration management? Push out ready files, or add using commands?
14:59 tyler-baker joined #salt
14:59 tyler-baker joined #salt
15:01 rgbkrk joined #salt
15:01 rgbkrk Does anyone in here have a good example of how they're using webhooks or GitHub hooks with salt?
15:01 MTecknology so... 0.17.4 for rhel boxes is... where?
15:02 cachedout joined #salt
15:04 viq Apparently not on epel-testing
15:05 rlarkin joined #salt
15:09 colinbits joined #salt
15:09 kaptk2 joined #salt
15:10 davidfischer joined #salt
15:14 renoirb Just upgraded the version, that was quick!
15:17 elfixit joined #salt
15:19 scoates anyone have experience with this? does this solve the minions-dont-connect problem? it looks like I'm running zmq3, not 2. https://github.com/saltstack/salt/issues/6231#issuecomment-29878818
15:22 aleszoulek joined #salt
15:23 timoguin joined #salt
15:24 th3reverend left #salt
15:24 MTecknology viq: I guess I'll just have rhel boxes on 0.17.2...
15:26 viq MTecknology: I want to play with the frozen deployments, this is one of the reasons why
15:27 MTecknology frozen deployments?
15:27 shiin left #salt
15:27 viq http://docs.saltstack.com/topics/tutorials/esky.html
15:28 nahamu I had trouble getting esky builds to be nice and portable under Linux, but it works great on SmartOS.
15:29 MTecknology ooh
15:30 nahamu viq: did you get it to work under Linux?
15:30 viq I only read about it, didn't look at it yet
15:32 nahamu ah
15:33 viq What do you mean "nice and portable"?
15:33 austin987 joined #salt
15:34 jasg75 joined #salt
15:35 austin987 joined #salt
15:36 nlb Need help with using grain var in a state file
15:36 nlb trying to do something like this:
15:36 nlb http://pastebin.com/4dA36qAJ
15:36 nlb but getting compile error
15:37 nlb what is correct syntax?
15:37 nahamu viq: I had trouble building on one version of fedora and running it on another, though I probably made the mistake of building on a newer rather than an older one.
15:37 nahamu I should try again at some point.
15:39 viq nahamu: ah. Well, I did somewhat assume from the description that it's "per OS version", though I'm looking at (I think) more diverging systems between versions: (Free|Open}BSD, Debian
15:40 Vye nlb: Could you include the error you're getting?
15:42 nahamu viq: nice thing about illumos is how carefully it tries to remain backwards compatible.
15:43 nahamu build on a nice old system and anything newer should still be able to run it.
15:43 viq hm, nice
15:43 viq I haven't done much with solaris
15:43 viq What are you using yours for, if I may ask?
15:43 fllr joined #salt
15:44 nahamu viq: I'm not doing much with it yet.
15:44 nahamu The developers here are deploying django apps on fedora
15:44 nahamu they're automating / CM'ing with salt
15:44 pcurry_nomi joined #salt
15:44 nahamu So since I want to move them to SmartOS I figured I should get salt working for them first.
15:45 jergerber joined #salt
15:46 quickdry21 joined #salt
15:47 danielbachhuber joined #salt
15:49 davet joined #salt
15:49 netzmonster joined #salt
15:50 nlb Vye: error pasted here: http://pastebin.com/xiWuHXDG
15:50 nlb Vye: sorry for delay, have to jump on guest network to paste to pastebin
15:51 anuvrat joined #salt
15:51 scoates nlb: salt://salt/minion/configs/managed_settings.{{grains['id']}}.conf
15:52 scoates the `+` is only to concatenate once within the replacement context (you don't need it; there's nothing to concatenate)
15:53 [diecast] joined #salt
15:53 nlb woot! awesome
15:54 nlb first couple of days with salt. coming from puppet
15:55 scoates nlb: I just moved one of our projects off of that steaming pile, myself. (-:
15:55 chromak__ joined #salt
15:56 nlb At new job and working on project that depends on remote code execution for testing purposes, I *have* to have cm of course, and this requires nodes all over the world behind nat
15:56 nlb mcollective didn't seem the right fit
15:56 nlb this, so far, is great
15:56 mgw joined #salt
15:57 nlb not much of a learning curve
15:57 scoates yeah; it's very Pythonic. Stuff makes sense.
15:57 nlb I do most of my development in ruby, python is pretty new to me, but I can read it for the most part
15:58 nlb Think I'm going to fork the salt repo and take a stab at adding a feature.
15:58 nlb something I always wanted in puppet, basically a header at the top of managed filesa
15:59 nlb that says the file is managed, what template or source it came from.
16:01 forresta joined #salt
16:01 seanz joined #salt
16:03 chromak__ left #salt
16:07 colinbits joined #salt
16:07 tdilly joined #salt
16:08 tdilly Are the values of grains always strings?
16:08 cowyn_ joined #salt
16:08 yota never understood why functions are in state module and not in exec modules cand onversely
16:08 cachedout joined #salt
16:09 tdilly For example this doesn't work as I would expect: {% if grains['osmajorrelease'][0] > 6 %}
16:09 scoates tdilly: not always strings
16:10 scoates `salt nodename grains.get pythonpath --output=json` <-- array
16:10 scoates (array is clearer in json than yaml)
16:10 colinbits joined #salt
16:10 ovorp joined #salt
16:11 tdilly ahh osmajorrelease IS a string, thanks scoates
16:11 Dale joined #salt
16:13 ovorp just getting started and jumped in a little too quickly with salt-cloud without understanding basics. Now I'm working through the tutorials and I have a questions about the pillar tutorial and walkthrough
16:13 ovorp when running the command with a target of something like 'webserver_role:dev'
16:14 ovorp is that role set as the minion hostname or is some other meta added to a minion at creation?
16:15 frantou joined #salt
16:16 jslatts joined #salt
16:17 blee_ joined #salt
16:19 ovorp agh found it in grains
16:21 utahcon I run a salt command and all my servers respond back that 0 succeeded and 0 failed. So it seems they aren't getting the sls properly. I created a new whoami for testing and sure enough it isn't working right. Ideas?
16:21 [diecast] joined #salt
16:22 scoates utahcon: on the minion, `salt-call -l debug state.highstate` might give you a bit more info
16:30 utahcon scoates: No Top file or external nodes data matches found
16:30 toastedpenguin joined #salt
16:31 utahcon Other than that, I didn't see anything that looked bad.
16:31 Marion joined #salt
16:31 utahcon Are the minions not able to pull down the states?
16:32 favadi joined #salt
16:32 StDiluted joined #salt
16:32 favadi joined #salt
16:33 scoates utahcon: 1) are you sure your master's file_roots: base: {} is pointing at the right place? and if so 2) can you pastebin your top.sls?
16:33 scoates heh. Now I'm supporting Salt. Oh, OSS. Got me again.
16:33 utahcon hmm... good call on teh file_roots, will check
16:33 utahcon thanks
16:34 oz_akan_ joined #salt
16:34 scoates if you stop your salt-master and then start it (interactively) with: `salt-master -l debug` ; you'll get output. Check also in /var/log/salt/master (or wherever your distro puts it)
16:35 utahcon scoates: that was it
16:35 utahcon thanks!
16:35 scoates oh, nice
16:35 utahcon facepalm
16:35 scoates that's easy (-:
16:35 lineman60 joined #salt
16:35 gadams999 joined #salt
16:35 utahcon Hey, I didn't think of it :D
16:36 [diecast] is there an integration method between fabric and salt
16:36 forresta [diecast], Not that I've seen.
16:36 jakubek [diecast]: what do you mean?
16:37 [diecast] i would like to have developers keep maintaining their fab files and have salt integrate the changes
16:37 [diecast] either as a wrapper or parser
16:37 forresta I mean you could use salt to do a cmd.run with fabric
16:37 forresta but there's nothing in salt to convert fabric to salt
16:38 scoates we're planning on replacing our (limited) fabric scripts with salt calls.
16:38 scoates as long as I can keep these minions connected. /-:
16:38 [diecast] currently fabric is heavily used here and im implementing salt
16:38 forresta scoates, you just rewrote your fabric scripts though right?
16:39 scoates forresta: yes, that's the plan. total replacement.
16:39 [diecast] to have salt use a cmd.run would maybe be a problem for dependency failures
16:39 forresta scoates, yea that's how I'd do it too.
16:39 forresta [diecast], yea it's a dirty work around, and I wouldn't do it.
16:39 [diecast] i am currently writing the complete replacement to salt
16:39 Marion joined #salt
16:40 [diecast] was wishing for a parser at a minimum =)
16:40 forresta [diecast], you should write one!
16:40 [diecast] oh yes, i hae the time!
16:40 [diecast] hah
16:40 forresta that would be cool, I keep saying that I should write one for puppet
16:40 forresta but there are only so many hours in the day
16:40 [diecast] ya, idealistically if i spent the hours writing it it could theoretically be the same time as writing it out
16:41 forresta yea
16:41 [diecast] but we know how that ends u
16:41 [diecast] *up
16:41 forresta :P
16:42 forresta [diecast], you could always create a feature request on that
16:43 jfzhu_us joined #salt
16:44 bhosmer joined #salt
16:44 forresta Does anyone know enough about sphinx to see wtf is causing this issue: https://github.com/saltstack/salt/issues/9366
16:45 oeuftete joined #salt
16:47 gadams999 Should the salt-bootstrap develop version install startup files for CentOS? I see one issue opened/closed earlier this year: https://github.com/saltstack/salt-bootstrap/issues/58
16:49 forresta s0undt3ch already fixed that I thought
16:50 forresta gadams, are the files not getting dropped into /etc/init.d?
16:50 forresta err gadams999
16:50 gadams999 nope, nothing changed in /etc/init.d -  no salt files there
16:51 gadams999 mlocate only shows /usr/bin/salt-master on my system. no startup files
16:51 forresta weird
16:52 oeuftete joined #salt
16:53 gadams999 indeed. not a problem as this is dev. I can always pull the file from the rpm
16:54 BRYANT__ joined #salt
16:54 forresta gadams999, you aren't getting that issue if you just use the standard bootstrap script right?
16:54 forresta from http://bootstrap.saltstack.org
16:54 gadams999 let me check
16:55 forresta ok
16:56 kermit joined #salt
16:58 shadowsun forresta: Yeah, now I got a Q for you
16:59 forresta shadowsun, ok, I didn't read the scrollback from a couple nights ago as a heads up
16:59 shadowsun forresta: that's fine, the question is newer.
16:59 shadowsun I'm using gitfs for environments.
16:59 forresta what's up
16:59 shadowsun so in git I have a branch named dev.
16:59 forresta ok
17:00 shadowsun the top.sls is in the master branch.
17:00 shadowsun It matches based on a pcre, which is working
17:00 shadowsun However, state.highstate blows up
17:01 Marion left #salt
17:01 shadowsun I'm still running salt 0.17.2 because EPEL doesn't have the new versions
17:02 shadowsun Now, highstate is blowing up because
17:02 [diecast] joined #salt
17:02 shadowsun ...
17:02 * shadowsun rummages
17:02 rojem joined #salt
17:03 gadams9991 joined #salt
17:03 rojem joined #salt
17:04 shadowsun self.avail[env] doesn't contain a key named 'dev'
17:05 forresta even though you have dev defined in your top.sls?
17:05 shadowsun Actually, it's worse than that.
17:05 shadowsun I've never seen python do this before.
17:05 shadowsun I added a couple lines to write the contents to a file
17:05 shadowsun File "/usr/lib/python2.6/site-packages/salt/state.py", line 2224, in render_highstate lfd.writelines(self.avail[env])
17:06 shadowsun KeyError: 'dev'
17:07 forresta I've never seen that error before :\
17:07 shadowsun No
17:07 gadams9991 forresta - straight bootstrap pulls the 0.17.2 RPMs in, which do have the startup files….
17:07 shadowsun I've never seen writing a list throw a keyerror.
17:07 gadams9991 curl -L http://bootstrap.saltstack.org | sudo sh -s -- -M
17:08 forresta gadams9991, ahh good point. I forgot it used the RPM for that. Are you specifically calling the git source when you did the develop bootstrap?
17:08 forresta shadowsun, yea
17:09 scoates I think my multiple-masters configuration might be a problem. /-:
17:09 forresta gadams9991, https://github.com/saltstack/salt-bootstrap/blob/v1.5.9/bootstrap-salt.sh#L1823
17:09 forresta scoates, you using syndic?
17:09 scoates I don't know what that is, so I guess not? (-:
17:09 forresta gadams9991, you can see there that the centos_stable function still installs the rpm
17:09 ckao joined #salt
17:09 forresta scoates, heh ok
17:10 gadams9991 sure - back in a bit though, pizza is here
17:10 forresta ok
17:10 shadowsun forresta: I'm going to pickle the data instead.
17:10 scoates oh. I see why.
17:10 forresta shadowsun, interesting, that should provide some more data.
17:11 forresta scoates, are you actually running two master values set in each minion conf file?
17:11 shadowsun forresta: If pickle blows up too, then I suspect I'm actually looking at a python bug
17:11 scoates forresta: yep
17:11 forresta Someone asked me about that during a presentation I did, and I had no answer
17:11 forresta scoates, cool, so it does work then
17:11 scoates no (-:
17:11 forresta shadowsun, that would suuuuuuck
17:11 scoates it only LOOKS like it works.
17:11 scoates http://paste.roguecoders.com/p/2e640373840354642605b24b0a6288e2.txt
17:11 scoates (the second master tries to auth, and then the minion fails to validate the key)
17:11 shadowsun scoates: are you having a problem with a dual active master config?
17:12 forresta ahh, so it only has the ability to auth with a single master, interesting
17:12 shadowsun Both amsters have to have the same key
17:12 shadowsun O.o
17:12 shadowsun both masters have to have the same key
17:12 forresta yea just make sure both hamsters have the same key
17:12 forresta problem solved
17:12 shadowsun LOL
17:12 mgw joined #salt
17:12 shadowsun bot hamsters!
17:12 forresta heh
17:12 shadowsun scoates: but yeah, I ran into the same problem when I was first setting up dual masters
17:13 * scoates nods
17:13 [diecast] joined #salt
17:13 scoates I can copy it over I guess.
17:13 forresta shadowsun, did any of the devs give the reason for why you can't have multiple masters set with different keys?
17:13 scoates seems less than ideal?
17:13 shadowsun forresta: I never asked, it's documented that you need to do it this way
17:13 forresta scoates, I agree it's less than idea, but at least it's an easy solution
17:13 shadowsun forresta: At a guess, it's because the functions were originally only written for one master
17:13 forresta scoates, if you want the docs: http://docs.saltstack.com/topics/tutorials/multimaster.html
17:14 forresta shadowsun, yea that makes sense.
17:14 [diecast] joined #salt
17:14 scoates thanks, forresta
17:15 scoates it's not a big deal; the secondary master is for my VM for testing only.
17:15 forresta yea np, I'm just glad shadowsun had done it before.
17:15 scoates (and also thanks shadowsun )
17:15 forresta gotcha
17:15 shadowsun (welcome :D)
17:21 pcurry_nomi joined #salt
17:21 jalbretsen joined #salt
17:23 scoates success on the key copy
17:26 scoates despite the manual key copy, this is SO much more sane than puppet.
17:27 forresta HOW DARE YOU SIR
17:28 forresta Look buddy, how dare you insult the fact that if your times aren't identical you're screwed
17:28 ovorp joined #salt
17:28 scoates forresta: ever set up puppetdb?
17:28 scoates I want that week of my life back.
17:28 forresta I'm going to have to at my job soon probably
17:28 forresta and I will want to kill myself
17:29 forresta we had it at my old job, thankfully I dodged that bullet
17:29 or1gb1u3 joined #salt
17:29 scoates the key management is a disaster
17:29 forresta yea we had issues with that
17:29 scoates I had to alias stuff in my /etc/hosts file just so I didn't have to expose it on the public network
17:29 forresta lol
17:30 nebuchadnezzar hello
17:30 forresta hi
17:30 scoates the icing on the cake is that it was taking around 2 minutes from launch to sockets opening. so if something was wrong, it took FOREVER to debug.
17:30 forresta that's dumb
17:31 forresta that reminds me, has anyone ever figured out what the puppet user does in here?
17:31 forresta puppet, is this a bot?
17:31 timoguin they're just watching
17:31 timoguin with competive, beady eyes
17:31 forresta nope, weird, I could have sworn I had gotten output back at some point
17:31 forresta The puppet dudes are nice guys
17:32 timoguin that's what they want us to think
17:32 forresta lol
17:32 ajw0100 joined #salt
17:33 nebuchadnezzar I'm thinking about the password problem, where to store it etc. I'm wondering if some king of Diffie-Hellman between minions could be investigated, with a mecanism to determine if password are in sync and generate a new one when it's not
17:34 forresta nebuchadnezzar, what do you mean 'password problem'?
17:34 nebuchadnezzar forresta: storing passwords in files
17:34 forresta you should put that data in pillar
17:34 forresta then reference it in states that need it.
17:34 timoguin i just responded to someone on the list about the same thing
17:35 nebuchadnezzar pillar is not files?
17:35 forresta oh you mean files on the master?
17:35 nebuchadnezzar yes, sorry
17:35 timoguin it is, but only the master has access to it if that's how you set it up.
17:35 timoguin i'm not sure how to deal with unhashed passwords like the mysql
17:35 timoguin but for user passwords i store just the hash in pillar
17:36 jacksontj joined #salt
17:37 frantou left #salt
17:38 scoates idea: when not globbing on a host, maybe the salt command should make `--show-timeout` implicit?
17:39 DanGarthwaite joined #salt
17:39 * DanGarthwaite is going to SaltConf
17:39 forresta nice DanGarthwaite
17:40 nebuchadnezzar forresta: saw the thread on the ML
17:40 forresta ahh ok
17:45 forresta scoates, we've been discussing the timeout flags and data returned here: https://github.com/saltstack/salt/issues/8761
17:46 whiteinge fxdgear: sorry i missed your follow up ping. you get that state working?
17:47 forresta hey whiteinge, did you get steamOS installed? I didn't want to blow away a system, and couldn't get it working inside of virtualbox for anything,.
17:48 pcurry_nomi1 joined #salt
17:48 whiteinge i never tried. i'm super excited for it but i'll wait for a commercial steambox release
17:48 forresta gotcha
17:48 fxdgear whiteinge I started looking into building my own package
17:48 fxdgear but it wasn't working when I triedto install the package on a new server and it failed... :/
17:49 scoates forresta: ty
17:49 whiteinge sorry to hear that. i wouldn't wish having to figure out building .debs on my worst enemy...  ;-)
17:49 forresta scoates, np
17:49 shadowsun scoates: that's why I use salt and not puppet. It's more saner.
17:49 scoates more saner indeed (-:
17:50 Ahlee gitfs, master is supposed to map to base, right?
17:50 Ahlee No matching sls found for 'graphite.get_login_stats' in env 'base'
17:50 Ahlee environment's Staging (git branch staging) and UAT (git branch UAT) find fine
17:51 forresta the master branch maps to the base env
17:51 Ahlee no environment set on minions in prod, expecting base to resolv to master
17:51 Ahlee no love.
17:51 forresta what does your gitfs_remotes look like?
17:51 timoguin Ahlee, so you have that state in the staging and UAT branches?
17:51 Ahlee sorry, misspoke, Staging (git branch Staging) - cap is important
17:51 timoguin and the master branch?
17:52 Ahlee timoguin: yes. All merged in and happy
17:52 timoguin hmm
17:52 Ahlee independent salt masters, all salt masters are capable of cloning gitfs remotes as root (user salt-master is running as, 0.17.2 across the board)
17:52 vkurup left #salt
17:52 shadowsun Oh, wow
17:53 shadowsun Ahlee: line 2222?
17:53 shadowsun Ahlee: in state.py?
17:53 AdamSewell joined #salt
17:53 AdamSewell joined #salt
17:53 Ahlee statefiles = fnmatch.filter(self.avail[env], sls_match) ?
17:53 shadowsun Yes
17:53 shadowsun I've got the same problem
17:53 shadowsun KeyError, but mine is 'dev' of course
17:53 Ahlee so let me back this up
17:54 Ahlee sorry, need to understand what render_highstate is doing
17:54 shadowsun Well
17:54 shadowsun Gather the state files and render them into a single unified salt
17:54 shadowsun high data structure.
17:54 shadowsun Are you on CentOS 6.5?
17:54 Ahlee 6.4
17:54 shadowsun Okay
17:54 Ahlee well, 6.4 and 6.2, i'm excluding my 5.6 minions from this test
17:54 shadowsun hrm
17:55 Ahlee python 2.6
17:55 shadowsun yeah
17:55 shadowsun 2.6.6
17:55 Ahlee i did rebuild for zmq 3
17:55 shadowsun So did I
17:55 shadowsun Custom repo
17:55 Ahlee yeah, pulled srpm's from koji
17:55 shadowsun zmq3.2.4?
17:55 Ahlee um, 3.2.2
17:55 shadowsun Okay
17:55 shadowsun SO we can rule out zmq
17:55 shadowsun probably
17:55 Ahlee yeah
17:55 Ahlee probably
17:55 shadowsun everything else we're damn near identical on
17:56 shadowsun :|
17:56 Ahlee that has it's pluses, at least.
17:56 shadowsun I tried to writelines() self.avail[env] and got a magical keyerror
17:56 shadowsun I'm going to pickle it and see what happens
17:56 Ahlee do you have any local states?
17:56 shadowsun I'm afraid it might be a python bug
17:56 shadowsun local states?
17:57 Ahlee i'm only having issues with gitfs backed items
17:57 shadowsun yes
17:57 ovorp joined #salt
17:57 shadowsun I only have gitfs backed items
17:57 Ahlee ok
17:57 [diecast] joined #salt
17:57 shadowsun Because multiple masters
17:57 shadowsun You know
17:57 shadowsun I can do a pip-install of salt-minion
17:57 shadowsun sec
17:57 shadowsun pickle.dump(self.avail[env], lfd)
17:57 shadowsun KeyError: 'dev'
17:58 shadowsun wait a sec.
17:58 shadowsun I bet you env is set to 'dev'.
17:58 shadowsun and I'm just being dumb, it's probably set to dev.
17:59 Ahlee $ grep ^environment /etc//salt/minion
17:59 Ahlee $
17:59 Ahlee so, no env set for me
17:59 shadowsun no
17:59 * Gareth waves
17:59 shadowsun I don't have mine set anywhere
17:59 shadowsun interesting
17:59 shadowsun The data structure has the production environment
17:59 shadowsun so
18:01 * shadowsun waves at Gareth
18:02 * Gareth waves back at shadowsun
18:04 Ahlee shadowsun: interesting.
18:04 Ahlee well, need to run.  Will continue to dig
18:04 shadowsun Ahlee: g'luck
18:04 forresta let us know Ahlee
18:06 shadowsun >>> print a.keys()
18:06 shadowsun ['production', 'base']
18:07 higgs001 joined #salt
18:07 shadowsun Okay, that's effed up.
18:07 forresta why are you failing on your envs shadowsun?
18:07 forresta jeez!
18:07 * shadowsun nukes the cache on the minion and tries again
18:07 shadowsun forresta: My top.sls doesn't even HAVE  production anymore
18:07 shadowsun so
18:07 shadowsun it "can't" be there.
18:07 forresta ahh nice
18:07 Ahlee shadowsun: yeah, i nuked my gitfs cache as well
18:08 Ahlee i even ran this on a new minion
18:08 shadowsun I'm nuking the entire minion cache... yeah, you're right, it still goes boom, but
18:08 Ahlee doh, reactor, so i guess "fresh" is relative
18:08 shadowsun 'production', 'base'
18:08 shadowsun interesting!
18:08 Ahlee oh well, guess i'll load this up onto a laptop and tinker over the holidays
18:09 shadowsun The minion magically has the production enviroment which doesn't exist.
18:09 * shadowsun double checks
18:09 Ahlee just a for statefiles.iterkeys()?
18:09 shadowsun and here's the messed up part
18:09 Ahlee or what's a
18:09 Ahlee i'll do same on thi shost
18:09 shadowsun the rendered top.sls doesn't have production in it either
18:09 shadowsun Oh
18:10 shadowsun I've modified state.py
18:10 shadowsun Lines 221-223 are now:
18:10 shadowsun with open("/log.txt", "w") as lfd:
18:10 shadowsun 2222                 import pickle
18:10 shadowsun 2223                 pickle.dump(self.avail, lfd)
18:10 shadowsun So it dumps a pickle of the data format to a file
18:10 shadowsun I'm then just loading the pickle in python cli to look at it
18:14 thrashr888 joined #salt
18:14 Ahlee thanks for that debug idea too, btw
18:14 Ahlee that'll come in handy.
18:14 Ahlee this time for real, gotta jet
18:14 shadowsun Ahlee: welcome - tc :)
18:18 cdcalef joined #salt
18:21 mpanetta Quick question... is salt-cloud the only part of salt that uses libcloud or even has access to cloud provider info?
18:22 forresta mpanetta, a quick search through the source says yes.
18:22 shadowsun forresta: hrm. Do I have to list individual branches somehow for gitfs?
18:23 shadowsun forresta: to tie them in with the enviroments? The documentation makes it sound like git branches should be magically mapped
18:23 mpanetta forresta: Would a salt state module for cloud load balancers be a welcome addition to salt?
18:23 timoguin yea they're supposed to be directly mapped to the envs that are defined in top.sls
18:23 forresta uhh I don't think so, you just set the fileserver_backend, set up the gitfs_remotes, and then it maps the branch names to environment minus master.
18:23 forresta which you have to use as master.
18:23 shadowsun forresta: Okay. Sec
18:23 bemehow_ joined #salt
18:24 forresta mpanetta, sounds kick ass to me man.
18:24 mpanetta Because I think I may be writing one in the near futre, if such a thing is possible.
18:24 forresta mpanetta, I'd say open an issue on it, as that will get some eyes on it from the dev team
18:24 mpanetta Ok
18:24 forresta worst case, you write it, then you put it in https://github.com/saltstack/salt-contrib
18:25 forresta so it would be worth doing regardless.
18:25 mpanetta Well, regardless I will preobably be writing one, we need something to help automate LB rotations on servers as we upgrade them.
18:25 shadowsun forresta: _get_envs says otherwise
18:25 forresta yea that would be slick
18:25 forresta shaodwsun, hmm
18:25 shadowsun forresta: file_roots is there
18:25 shadowsun forresta: but it only contains production and base(master)
18:25 shadowsun forresta: It's missing qa and dev branches completely.
18:26 shadowsun So this is def a bug minion side
18:26 forresta shadowsun, I thought you didn't use file_roots with gitfs...
18:26 shadowsun forresta: You don't, but the internal code does
18:26 forresta ahh ok
18:26 forresta can you trash /var/cache/salt on the minion?
18:26 forresta then restart the minion service?
18:26 shadowsun forresta: I've already done that
18:26 forresta lame
18:26 shadowsun yeah
18:26 shadowsun I tried that like an hour ago?
18:26 forresta so can you see that the environments exist on the master?
18:26 shadowsun :|
18:27 shadowsun ?
18:27 shadowsun How would I even do that?
18:27 forresta I'm trying to think of a way to confirm that it's being seen by the master
18:27 forresta I don't know :P
18:27 shadowsun Oh
18:27 shadowsun If the minion is actually using the production environment, everything works
18:27 shadowsun but
18:27 shadowsun the other two environments are missing
18:27 forresta yea but it's not seeing the dev environment even right? And it's part of the dev matching in your top.sls
18:27 shadowsun bingo
18:27 shadowsun and it's a branch on the remote
18:28 forresta yea that's the problem, it would make sense if the minion wasn't part of the dev env, so it only saw prod or whatever, I could accepte that
18:28 shadowsun yeah
18:28 shadowsun but this minion isn't part of prod
18:28 forresta right
18:28 shadowsun so it's all kinds of wtf
18:28 shadowsun you know
18:28 shadowsun let me use pip on my other test box to get salt 0.17.4
18:28 forresta ok
18:28 timoguin wonder if it's something to do with the top.sls file? doesn't gitfs have some issues where it combines the top.sls from all envs into one?
18:28 shadowsun timoguin: the only top.sls is in base
18:28 timoguin well then that's out
18:29 shadowsun timoguin: I moved everything else into an orphan branch and then branched that for my environments
18:29 forresta shadowsun, as dumb as it sounds, can you try this: http://docs.saltstack.com/topics/tutorials/gitfs.html#why-aren-t-my-custom-modules-states-etc-syncing-to-my-minions
18:29 [diecast] hey forresta i lost connection earlier. could you explain advantages of using the bootstrap.sh
18:30 shadowsun forresta: I'm not having this problem
18:30 shadowsun forresta: I checked debug output, the rendered files contain the updated information
18:30 forresta [diecast], give me a few, troubleshooting some work stuff
18:30 shadowsun also
18:30 shadowsun If I disable gitfs, *I have nothing*. everything is in git. :/
18:31 [diecast] shadowsun gitfs_remotes
18:32 shadowsun [diecast]: what about it
18:32 [diecast] maybe supply a local option
18:32 shadowsun [diecast]: What do you mean, supply a local option?
18:32 shadowsun I wasn't aware that gitfs_remotes took flags?
18:33 [diecast] http://docs.saltstack.com/topics/tutorials/gitfs.html#multiple-remotes
18:33 shadowsun Oh, using file://
18:33 shadowsun umm
18:34 [diecast] yes, if you start with a repo (in any state, likely stale) then you have a fallback
18:34 shadowsun that might be a bit challenging, I'd have to actually get the raw file structure off the git server
18:34 [diecast] correct
18:34 shadowsun yes
18:34 shadowsun and then sync it between three servers.
18:34 shadowsun And keep it in sync.
18:34 [diecast] not really
18:34 shadowsun With each push.
18:34 [diecast] i mean, that would be nice right
18:34 shadowsun yes, I kinda would.
18:34 shadowsun well
18:34 shadowsun Everything is in the git server
18:34 [diecast] but if you are talking about strictly provisioning it shouldn't be manditory
18:34 shadowsun so if I dupe it locally, I have to keep the damn thing in sync to make a change.
18:35 shadowsun I'm not talking about provisioning
18:35 shadowsun I'm talking about the minions can't see whole environments.
18:35 shadowsun And I'm trying to troubleshoot why environments are missing.
18:35 [diecast] and you found that it does not have access to a remote gitfs?
18:35 shadowsun what the hell?
18:35 [diecast] hmm
18:36 timoguin [diecast], it has access. the branches are supposed to map to environments.
18:36 timoguin anything that's not his production env isn't working
18:36 [diecast] yes, ok
18:36 forresta [diecast], the benefit of the bootstrap.sh is that it's OS agnostic, so you can get the salt minion or master installed on a variety of systems (in the event you aren't using salt-cloud, or need a local option), the other benefit is you can specify a specific release if you want to.
18:36 [diecast] so the user doesnt have access maybe
18:36 forresta shadowsun, deep breaths :P
18:36 shadowsun forresta: yess.
18:36 [diecast] shadowsun did you manually try to pull/clone/etc from those instances.
18:36 shadowsun [diecast]: I do it all the time.
18:37 timoguin [diecast], best part of the bootstrap to me is being able to specify specific git revisions, branches, or tags
18:37 shadowsun [diecast]: the master has full access. There's no problem with the masters access.
18:37 timoguin faster testing with newer releases
18:37 shadowsun [diecast]: Who do you think writes the stuff that's in those branches? :P
18:37 [diecast] shadowsun i dont know, you i hope? ;)
18:38 shadowsun [diecast]: yeah
18:38 [diecast] but what does that have to do with remote access via a certain account to a specific branch of a remote repo?
18:38 [diecast] i author all day from my desktop
18:38 shadowsun [diecast]: did you know that minions contact the master, and the master pulls from the git server? The minions never have direct contact with it.
18:38 [diecast] not logged into a remote master
18:38 shadowsun [diecast]: git branch maps to salt environment
18:38 [diecast] shadowsun yes, i assumed that
18:38 nlb I'm having trouble ordering a state
18:39 nlb I have a state that include the creation of a directory
18:39 [diecast] right, i have mine setup as branches as well.
18:39 nlb then creates a file in that direcotry
18:39 shadowsun [diecast]: So if I have a git branch named "dev" and an environment named "dev", the minion that's in the dev environment gets the file version from the dev branch.
18:39 shadowsun Right. Okay
18:39 nlb how do I require the directory from the file.managed: require:
18:39 shadowsun So
18:39 [diecast] ya
18:39 shadowsun How come the minion is missing the enviroment?
18:40 shadowsun like, it goes to check the state files it has for environment dev
18:40 shadowsun and it has no environment dev and KeyErrors
18:40 [diecast] log onto a minion, git clone with the salt user or whatever you run it as
18:40 shadowsun I can't
18:40 shadowsun O.o
18:40 [diecast] what happens when you git -b branch
18:40 shadowsun ...
18:41 shadowsun I can't do that from a minion
18:41 shadowsun the minion never contacts git directly
18:41 shadowsun there's a firewall allow for the masters
18:41 timoguin the master has the access, and it serves up the files to minions
18:41 shadowsun yes
18:41 shadowsun what timoguin said
18:41 timoguin or in this case, it doesn't
18:41 [diecast] heh
18:41 shadowsun well
18:41 shadowsun it' ought to anyway ;P
18:41 [diecast] hmm, why not allow them
18:42 [diecast] they're not internal in a VPC?
18:42 shadowsun [diecast]: because they don't contact the git server directly
18:42 xmltok joined #salt
18:42 [diecast] why not
18:42 shadowsun ...
18:42 timoguin because that defeats the purpose of using the master as a file server
18:42 shadowsun ^^^
18:43 shadowsun :q
18:44 [diecast] but, but
18:44 [diecast] it is an interesting design
18:44 shadowsun [diecast]: If the minion is trying to to contact git directly, then there's a *huge* bug in salt
18:44 [diecast] im not sure that level of granularity is needed
18:45 shadowsun since it's explicitly written to not do that.
18:45 shadowsun actually, it is
18:45 shadowsun Since I need to have different minions in different environments so I can write new code without breaking production boxes.
18:45 [diecast] right
18:45 [diecast] i have setup multiple vpc, one for each dev/prod/staging
18:46 [diecast] so permissions within can be a bit loose, of course not wide open
18:46 [diecast] using gitlab internally instead of github so a lot of "remote" is actually private
18:46 shadowsun Okay.. I don't see where there's a permissions problem anywhere
18:46 shadowsun The masters have access to the git server
18:46 shadowsun there's no problem there
18:47 [diecast] the user running master and the directories maybe
18:47 shadowsun ...
18:47 shadowsun okay, stop.
18:47 [diecast] is there any selinux stuff going on?
18:47 shadowsun I don't even know what the fuck you're trying to troubleshoot
18:47 shadowsun but it's not the problem I'm trying to troubleshoot
18:47 [diecast] ok fine
18:47 shadowsun So lets start from the top
18:47 shadowsun File "/usr/lib/python2.6/site-packages/salt/state.py", line 2255, in render_highstate statefiles = fnmatch.filter(self.avail[env], sls_match)
18:47 shadowsun KeyError: 'dev'
18:48 shadowsun The KeyError is caused because self.avail has an incomplete list of environments
18:48 shadowsun It has base and it has production, but it's missing qa and dev
18:48 timoguin shadowsun, i'm not using any branches in my gitfs yet since i'm currently only doing a dev env, so would you mind letting me know when/if you figure this issue out?
18:48 timoguin 'cause i'll be using them very soon
18:49 shadowsun timoguin: I will try to yeah, but this is rapidly getting into "I'm just going to file a bug" territory
18:49 shadowsun since I just replicated on 0.17.4
18:49 [diecast] and everything is good in the top file?
18:49 shadowsun [diecast]: yes.
18:49 timoguin yea, please link me to the bug when you file it then. :)
18:49 shadowsun [diecast]: and if I put the minion on the production environment, everything works as expected.
18:49 [diecast] im just playing devil's advocate here a bit so don't get too mad ;)
18:50 shadowsun I'm not
18:50 shadowsun lol
18:50 [diecast] what if you rename default branch to production
18:50 shadowsun I'm frusterated with the machine
18:51 shadowsun um
18:51 shadowsun production is an orphan branch
18:51 shadowsun I can't turn it into master
18:51 [diecast] i wonder if there is some odd association with "master"
18:51 [diecast] as a name
18:51 shadowsun base IS master
18:52 shadowsun your base enviroment is the master in git
18:52 [diecast] ok
18:52 timoguin and that's working
18:52 shadowsun timoguin: yes.
18:52 shadowsun What I need to know is
18:53 shadowsun what populates self.opts['file_roots']
18:53 [diecast] where is this stored? fnmatch.filter(self.avail[env], sls_match)
18:53 shadowsun [diecast]: It's part of the class
18:53 shadowsun it's populated by _get_envs(), which pulls info out of self.opts['file_roots']\
18:53 [diecast] i mean the object's value of self.avail[env] and sls_match
18:53 [diecast] ah, ok
18:53 shadowsun What?
18:53 [diecast] so, yes. that is important to know
18:53 shadowsun It's in state.py
18:54 [diecast] self.opts['file_roots']
18:54 shadowsun yes
18:54 [diecast] that value would prob give a serious clue
18:54 [diecast] =)
18:54 shadowsun It has
18:54 shadowsun it's missing half of the enviroments
18:54 jacksontj whiteinge: yt?
18:54 shadowsun I'm trying to figure out why
18:54 shadowsun ;P
18:54 shadowsun thats why I went back to the top when you started talking about FS permissions
18:55 [diecast] ok, and now that i have the full picture
18:55 shadowsun yeah
18:55 [diecast] hahah
18:55 shadowsun Sorry, I thought you'd read the scroll of the last two hours of forresta and I looking at this
18:55 [diecast] no, i apologize
18:55 [diecast] i lost connection
18:55 shadowsun ahhh
18:55 whiteinge jacksontj: que?
18:55 shadowsun Now it all makes sense :)
18:55 [diecast] hehe
18:55 bhosmer joined #salt
18:55 [diecast] well, now i have something to think about
18:55 [diecast] thanks for re-sharing
18:55 shadowsun Welcome :)
18:56 jacksontj whiteinge: we got a nice pull req for you :D https://github.com/saltstack/salt/pull/9374
18:56 shadowsun I'm going to grab some food and then root around in source code more
18:56 jacksontj drops our CPU usage on the master by ~75%
18:56 jacksontj and thats the "small" master
18:56 whiteinge O_o
18:56 MTecknology forresta: as far as the documentation goes... yes, I finally got around to it.
18:56 jacksontj should be even bigger savings on the "big" one :D
18:56 MTecknology whiteinge: special thanks!
18:56 whiteinge whoa
18:57 bemehow joined #salt
18:57 rojem joined #salt
18:57 MTecknology hm?
18:57 jacksontj whiteinge: so basically we are stating all the minion key files on every auth request just so we can log
18:57 jacksontj lol
18:58 jacksontj when you have 20 minions, thats probably ok
18:58 jacksontj with 10k, not so much
19:00 whiteinge jacksontj: that is incredible
19:01 jacksontj wanted to get it in, since we are about to cut a new release. We like to commit upstream before backporting ;)
19:01 whiteinge nice work tracking that down. and your pull req is freakin' awesome :-D
19:02 Sheco joined #salt
19:02 jacksontj not mine, its rbtz's ;)
19:02 MTecknology jacksontj: mine's bigger
19:03 vejdmn joined #salt
19:03 shadowsun when is EPEL going to get caught up with current salt releases?
19:04 bemehow_ joined #salt
19:06 whiteinge MTecknology: lol
19:06 shadowsun nevermind
19:06 shadowsun 0.17.4 hit epel today
19:10 harobed joined #salt
19:11 shadowsun well
19:11 shadowsun is hitting.
19:11 MTecknology shadowsun: awesome!!
19:12 shadowsun MTecknology: very. I can get it off epel-testing on two of my boxes.
19:12 MTecknology I'll be on 0.17.4 for at least a year
19:12 shadowsun I doubt that
19:13 shadowsun the salt guys submit updates to it rather more frequently than that
19:13 MTecknology ya- but I won't be updating to it
19:13 shadowsun Oh I bet you will
19:13 MTecknology I actually downloaded the deb and will be making a state to hold it at that version
19:13 shadowsun why?
19:14 shadowsun i.e. why not allow bugfixes and stuff?
19:14 MTecknology because of how often a newer version becomes incompatible with old versions, I can't do it
19:14 shadowsun ah
19:14 jcockhren I partly agree with MTecknology
19:14 shadowsun I haven't had that problem
19:15 dstanek joined #salt
19:15 MTecknology I'm having a lot of issues now because I updated the master and the image we deploy has a version that can't even connect.
19:15 jcockhren though. I'd freeze major versions. not bugfixes
19:15 MTecknology I have to recreate the image to handle this
19:15 shadowsun Ah
19:15 shadowsun I deploy and install from repo
19:16 MTecknology I'm just going to stick with 0.17.4 (no others) until we fully deploy these new servers
19:16 UtahDave joined #salt
19:17 MTecknology UtahDave!
19:17 MTecknology hi
19:17 UtahDave Hey, MTecknology!
19:17 shadowsun UtahDave: sup!
19:18 UtahDave sup
19:18 forresta MTecknology, Are you submitting a pull request for the docs? Sorry been dealing with some prod issues
19:18 UtahDave !
19:18 Gareth UtahDave: howdy
19:18 Thiggy joined #salt
19:18 gadams999 joined #salt
19:18 UtahDave hey, Gareth!
19:18 MTecknology forresta: it's already been submitted
19:18 UtahDave Gareth: lots of joy expressed for your debian/ubuntu network states!
19:18 forresta oh cool
19:18 MTecknology where's kuffs been? :(
19:18 Gareth UtahDave: yay :)
19:19 forresta MTecknology #?
19:19 whiteinge MTecknology: yeah, that is a real problem. your plan makes sense to me. fwiw
19:19 MTecknology forresta: https://github.com/saltstack/salt/pull/9373
19:20 jacksontj whiteinge: https://github.com/saltstack/salt/pull/9376
19:20 MTecknology whiteinge: ya... it's a "catch 22" issue... You can either require things remain backward compatible with minions, or you can keep this insane level of development.
19:20 jacksontj moving it to the loop job instead of on each auth
19:21 forresta MTecknology, are your minions running a 0.16.x release?
19:21 pcurry_nomi joined #salt
19:22 MTecknology yup
19:22 Gareth UtahDave: glad it was a beneficial addition :)
19:22 forresta psssh, those have been incompatible since 0.17.0! Get with the times! :P
19:22 shadowsun UtahDave: BaseHighState gets created with opts - do minions get their opts from the master?
19:23 MTecknology forresta: I know, but I built the image with 0.16 and changing the image is extremely painful
19:23 forresta Yea I'm just joking with you
19:23 forresta I've gotta go grab some food before something else catches on fire, I'll be back in a bit
19:24 MTecknology *grumble*
19:24 MTecknology I really need to make these salt states work with centos boxes
19:25 UtahDave shadowsun: hm. not sure if some of the opts comes from the master, but opts mostly the minion's config
19:25 shadowsun UtahDave: Hrm. It seems to contain file_roots on the minion.
19:25 JasonSwindle joined #salt
19:25 JasonSwindle Howdy, all
19:25 MTecknology any of you guys have rhel, debian, and windows on a master?
19:26 MTecknology I'm curious what a top file would look like in that situation. I'm guessing you don't use '*'
19:26 vejdmn joined #salt
19:27 UtahDave MTecknology: It depends on how your states are built.  I have several states that work across all three of those OSes.
19:27 MTecknology what does that state do?
19:28 jacksontj UtahDave: whiteinge got another awesome pull req :) https://github.com/saltstack/salt/pull/9376
19:28 UtahDave :)  just install vim.  lol
19:29 UtahDave jacksontj: I saw that!  very cool.
19:29 jacksontj working on getting you a nice CPU graph ;)
19:29 UtahDave jacksontj: I love that graph on the first one
19:30 MTecknology I'm in the top 50 comitters! :D
19:30 timoguin MTecknology, check out how the apache-formula works with both debian- and redhat-bases systems
19:30 timoguin https://github.com/saltstack-formulas/apache-formula/blob/master/apache/map.jinja
19:30 cdn joined #salt
19:31 cdn Hi Guys, what is the ideal way to duplicate a folder (and all it's sub folders) from a directory on the minion to another folder on the same minion
19:31 MTecknology timoguin: neat
19:31 shadowsun timoguin: Oh, oh no.
19:31 UtahDave MTecknology: woot!  ^5
19:31 jcockhren timoguin: yo!
19:32 timoguin sup jcockhren?
19:32 cdn file.recurse or file.direcory.recurse
19:32 Gareth woo on the board :)
19:33 MTecknology I want to get to top 15... but I want to do it with good quality commits (I'd like to think most of them have been good so far
19:33 MTecknology )
19:33 cdn I seem to be able to create the new dir, but nothing is in there :)
19:34 jcockhren timoguin: the thursday after next, come out to bbh if you can
19:34 timoguin jcockhren, what the jazz is bbh?
19:34 jcockhren timoguin: Broadway Brew house.
19:35 timoguin jcockhren, pynash evening meetup?
19:35 jcockhren Beerz 'n tech
19:35 jcockhren umm... mostly any tech
19:37 timoguin jcockhren, ah the 2nd. pynash is starting evening meetups too. first one is the 9th at Stratasan
19:37 timoguin is there a meetup page or anything for what you're tlaking about?
19:38 jcockhren timoguin: it's not really a meetup 'thing'. more like a tradition of developers
19:38 jcockhren just coming together
19:38 jcockhren so no sponsorship or web page
19:39 jcockhren just friends in tech getting together and inviting people. There's tends to be a special guest every once in a while
19:39 timoguin sounds cools. i'll get my coworker to come too
19:39 jcockhren word
19:39 timoguin i've been learning him some salt
19:40 jcockhren so yeah. Gene Kim dropped by once
19:40 shadowsun timoguin: Soooo
19:40 jcockhren back when there was a Agile Conf in Nash or something. arafat brought him
19:40 MTecknology *grumble* Where the crap is /var/lib/snmp/snmpd.conf (debian) on an rhel box?...
19:41 jcockhren not to drop names.... just saying... come nerd out over beerz
19:41 seanz left #salt
19:41 gadams999 joined #salt
19:41 jcockhren that invitation goes for ANY one else in here if they're in Nashville at some point
19:41 timoguin jcockhren, i do know arafat actually. :)
19:41 * jcockhren waves at everyone
19:41 MTecknology jcockhren: I owe some people in here some drinks
19:41 jcockhren timoguin: I know you know him.
19:41 timoguin pssshhhh cause i just told you
19:42 MTecknology I think we should have an IRC get together...
19:42 bemehow joined #salt
19:42 MTecknology jcockhren: is that what you meant?
19:42 timoguin shadowsun, sorry. i gotta roll out for now....
19:42 shadowsun timoguin: I has an answer for you
19:42 shadowsun Oh, nvm then
19:42 higgs001 joined #salt
19:43 timoguin i'll come back. :)
19:43 jcockhren timoguin: he told me like "I have a friend tim that doing a talk on salt at pyTN"
19:43 shadowsun timoguin: short version: It'll work fine if you don't break your git server in subtle ways.
19:43 shadowsun :)
19:43 jcockhren was like, "yeah, I know"
19:43 shadowsun [diecast]: You put me on the right track, although it was completely tangential to what you though
19:43 shadowsun t
19:43 [diecast] sometimes talking is helpful to get those juices going
19:44 shadowsun Indeeeeed
19:44 jcockhren MTecknology: not really an "IRC" get together. We have a develop group that already gets together. Always looking for great devs to come by and hang
19:44 jcockhren like I said, not really a "meetup" in the traditional sense. more like a hangout that has grown organically
19:45 jcockhren over 7+ years
19:45 bemehow_ joined #salt
19:45 MTecknology jcockhren: oh.. :(
19:45 MTecknology well... I hope you guys have fun. :(
19:46 MTecknology *sniffle*
19:46 jcockhren MTecknology: why sad?
19:46 jcockhren you don't like beer?
19:46 jcockhren haha
19:46 MTecknology I was guilting you into inviting me.
19:46 MTecknology trying to be funny
19:47 jcockhren oh. I already invited you ;)
19:47 MTecknology oh
19:47 MTecknology I'd enjoy that
19:47 * jcockhren officially invites MTecknology
19:47 MTecknology :D
19:49 bemehow joined #salt
19:50 MTecknology jcockhren: so.. when/where?
19:50 MTecknology I still need to get my room reserved and transportation to the place set up
19:55 jcockhren MTecknology: Nashville at the Broadway Brew House every Thursday
19:56 jcockhren MTecknology: let me know whenever you're in town. we can hang
19:56 rgarcia_ joined #salt
19:56 jcockhren MTecknology: there's word some of the devs will be doing similar something in SF
19:56 jcockhren b/c many of them work at github
19:57 jcockhren "at" being sometimes in SF
19:57 harobed_ joined #salt
19:57 rgarcia_ is it possible to run salt modules to fill in pillar data? I'd like to avoid writing an external pillar to expose functionality already in a salt module
19:57 rgarcia_ e.g. fill a pillar key with the result of __salt__["some.cmd"](arg1, arg2, ...)
19:58 harobed_ joined #salt
19:58 harobed_ joined #salt
19:59 UtahDave rgarcia_: yeah, you can do that.   Just remember that any commands executed run in perspective of the Salt Master.
19:59 UtahDave rgarcia_: only __grains__ is specific to the minion
20:00 dstanek joined #salt
20:01 rgarcia_ UtahDave: cool, thanks! I'm using a module that makes an API call that is the same from any perspective
20:01 UtahDave cool
20:01 thrashr888 joined #salt
20:01 rojem joined #salt
20:05 ovorp joined #salt
20:11 diegows joined #salt
20:12 william_20111 joined #salt
20:16 carmony good afternoon fellow salters :)
20:17 thrashr888 joined #salt
20:18 sroegner joined #salt
20:21 Sheco joined #salt
20:23 jfzhu_us Pillar templates with modules get executed on the master and not the minion? Can anyone confirm this?
20:25 dstanek anyone every deploy apps (like a Django app) using salt?  i know you can, but i'm debating in my head if i should use salt, fabric, ansible, etc.
20:31 pcurry_nomi joined #salt
20:32 fxdgear whiteinge hey I missed if you responded to me...
20:32 fxdgear once I got to work i gotcrazy busy and pulled into lots of meetings
20:32 fxdgear and my scroll back has lost our conversation
20:35 Ahlee oh neat. i wonder if those stat issues are what kills our masters @100 minions
20:36 thrashr888 joined #salt
20:37 shadowsun Ahlee: Wow, I haven't seen any problems like that
20:37 shadowsun Ahlee: btw.. Your git repo may be subtly broken.
20:38 paul joined #salt
20:39 jcsp joined #salt
20:40 UtahDave hey, carmony!
20:41 UtahDave dstanek: yeah, lots of people deploy Django using Salt.  I think whiteinge has an example online somewhere
20:41 UtahDave jfzhu_us: Yes, Pillars are compiled on the master.
20:41 UtahDave back in a minute
20:41 jnials joined #salt
20:42 jnials Where do I find the changelog file for the various version of salt?
20:42 UtahDave joined #salt
20:43 Gareth jnials: eg http://docs.saltstack.com/topics/releases/0.17.0.html
20:43 Gareth latest release.
20:43 Gareth http://docs.saltstack.com/topics/releases/0.17.4.html
20:45 jnials Gareth:  Thx.
20:45 Gareth np
20:45 dstanek UtahDave: would they just run state.highstate to deploy?
20:46 mgw joined #salt
20:49 thrashr888 joined #salt
20:57 thrashr888 joined #salt
20:57 UtahDave dstanek: you could set it up that way
20:59 dstanek UtahDave: is there another way? i was thinking about making my own command like app.deply
20:59 mesmer joined #salt
20:59 dstanek app.deploy that is
20:59 UtahDave dstanek: Yeah, you could create a module that does the deploy for you.
21:00 dstanek UtahDave: thanks for the insight
21:00 zfouts joined #salt
21:00 pengunix joined #salt
21:01 UtahDave any time!
21:04 dstanek one more question...is there a way to combine pillar data? for instance, say i have a list of users in someapp:users and anotherapp:uses
21:04 dstanek would have i have to have two different for loops in my sls to create all of the user states?
21:06 scoates I am also interested in this.
21:06 blee joined #salt
21:07 dstanek the case i have right now is in prod i have to different nodes one for each app and each has a user list; in my staging environment both apps run on the same node and both sets of users need to be on the node
21:07 dstanek in my current solution i maintain lots of duplicate lists of users
21:07 blee does anyone have any ideas as to what this means?  I havent changed my states at all and this started happening: http://pastie.org/private/x3assfoocil8eu0tzg3hea
21:08 rgbkrk joined #salt
21:08 dstanek blee: have you tried to see if there is a detailed error on the minion?
21:09 bhosmer joined #salt
21:11 erezool joined #salt
21:12 erezool left #salt
21:19 vejdmn joined #salt
21:19 Veticus joined #salt
21:20 UtahDave dstanek: You can't really combine them right now in pillar
21:21 UtahDave dstanek: One solution is to use an external pillar.  The external pillar environment lets you do whatever you want with the pillar dict
21:21 blee dstanek, not much going on in the minion logs either
21:23 dstanek UtahDave: i'll take a look at that; was considering writing my state in Python like they show at the end of http://dev.mlsdigital.net/posts/SaltStackBeyondJinjaStates/
21:23 shadowsun Hrm
21:24 shadowsun Can a minion get stuff from multiple nevironments?
21:27 UtahDave shadowsun: Yeah, it can.  It just depends on what th minion matches in the top file
21:27 shadowsun Okay
21:27 shadowsun That actually makes my life a lot easier.
21:27 shadowsun Now.
21:27 jslatts joined #salt
21:27 shadowsun which enviroment will the minion get cmd.script files from?
21:32 mpanetta Is it possible to use salt-cloud modules in salt?
21:32 alexandrel joined #salt
21:36 bhosmer joined #salt
21:40 armonge joined #salt
21:43 JasonSwindle joined #salt
21:53 quickdry21 joined #salt
21:56 thrashr888 joined #salt
22:00 shadowsun UtahDave: Okay, maybe I'm not understanding something here. I've got a dev environment - it has a script in it. I've even statically set the environment in the minion config. Why is it still trying to pull the version of the script in base instead of from the correct branch?
22:01 DanGarthwaite joined #salt
22:01 giantlock_ joined #salt
22:01 higgs001 joined #salt
22:03 Chrisje joined #salt
22:06 UtahDave shadowsun: can you pastebin your top.sls and your file_roots config option from your /etc/salt/master     ?
22:07 shadowsun UtahDave: I can't pastebin from here, but file_roots is gitfs
22:07 shadowsun I can dcc you, or I can just run home, grab grub and pastebin from there
22:08 forresta lol
22:08 forresta you can't gist at work?
22:09 * whiteinge thinks we need a dcc-based IRC pastebin bot
22:10 forresta whiteinge, that would be cool
22:10 dstanek do people typically control the uids and gids in salt?
22:11 forresta dstanek, a lot of the time yes
22:11 forresta dstanek, take a look at https://gist.github.com/UtahDave/3785738
22:12 shadowsun forresta: rules at work
22:12 forresta shadowsun, fair enough
22:12 shadowsun forresta: Yeah :|
22:12 forresta *shrug*
22:12 shadowsun UtahDave: I'll run home and pastebin 'em from there. It's hot here anyway.
22:12 forresta it happens, I've had places where IRC was banned
22:12 shadowsun I don't know what IRC is.
22:12 shadowsun >.>'
22:13 shadowsun ;)
22:13 shadowsun You usually don't see me on here in the day for a reason
22:14 whiteinge heh
22:16 ovorp joined #salt
22:16 JasonSwindle forresta:  It looks like I am going to SaltConf
22:16 dstanek forresta: thanks that was extremely helpful
22:20 mordred joined #salt
22:27 thrashr888 joined #salt
22:27 foxx joined #salt
22:29 forresta dstanek, np
22:29 forresta JasonSwindle, nice!
22:30 JasonSwindle I now have to get ready for that….
22:37 forresta heh, I'll be back online in 20.
22:45 jeffro joined #salt
22:50 forresta joined #salt
22:50 forresta whiteinge are you around?
22:51 ovorp joined #salt
22:51 shadowsun UtahDave: http://pastebin.com/hT07EbUM
22:51 nineteeneightd joined #salt
22:52 shadowsun UtahDave: Additionally, I have environment set to dev in the /etc/salt/minion but it still pulls scripts from base for cmd.script
22:52 _fllr_ joined #salt
22:54 kolbe left #salt
22:55 UtahDave shadowsun: if you set the environment to dev in the minion config, then it should pull ALL its files from the dev environment
22:55 UtahDave shadowsun: so you have a dev branch in your git repo, right?
22:55 higgs001 joined #salt
22:55 shadowsun yes
22:56 shadowsun it's where the state definitions are for devbox
22:56 shadowsun and for highstate it does pull from dev
22:56 shadowsun but when I cmd.script it pulls from base, and if the script isn't there it can't find it even if it is in dev
22:57 shadowsun I mean, it contacts the master for the script, but the salt-master can't find it if it's in the dev branch but not in base
22:58 william_20111 joined #salt
22:58 jeffrom joined #salt
22:59 whiteinge forresta: hi
22:59 forresta hey, so regarding https://github.com/saltstack/salt/issues/9366
23:00 forresta what was the original logic behind calling the actual state pip_state, instead of pip?
23:00 forresta Does that create a conflict with actual pip?
23:00 Marion joined #salt
23:01 kermit joined #salt
23:02 whiteinge i don't know the history behind that one
23:02 forresta ok
23:03 whiteinge (it did use to be called pip.py...)
23:03 forresta so now that you've modified the reference, when the docs are rebuilt it will destroy the existint state.pip.html page right?
23:03 whiteinge no
23:03 forresta I know, that's why I was confused, because I remember it being called that :\
23:03 forresta *existing
23:03 forresta oh it won't?
23:03 whiteinge no. there shouldn't be *links* to the old page anywhere but the html file still exists on the webserver
23:04 forresta hmm
23:04 whiteinge i'm working on a fix
23:04 forresta It was treating those as though they were overlayed almost, the navigation was the same
23:04 forresta whiteinge, ok cool
23:04 forresta if you figure out why the state is pip_state, let me know, because it bothers me :(
23:05 whiteinge gah. my OCD! ...the goggles do nothing!
23:05 whiteinge (it bothers me too ;)
23:06 forresta lol, I'm not even ocd. It just annoys me, because it messes with the 'standard', especially that setup in the docs.
23:06 * whiteinge nods
23:11 pcurry_nomi joined #salt
23:11 defunctzombie anyone have tips for making salt-call STFU
23:12 defunctzombie it is way way noisy
23:12 defunctzombie would like to run it as a cron job periodically
23:12 defunctzombie but only want to see error output
23:12 rgarcia_ joined #salt
23:12 defunctzombie ideally by just running the salt-call command
23:13 defunctzombie but if it is easier to just connect to the minion service to log errors, then maybe that is the route I gotta take
23:13 defunctzombie I see some notes about sentry logging
23:14 elfixit joined #salt
23:16 jer_ joined #salt
23:16 jcsp joined #salt
23:17 jer_ good afternoon
23:17 jer_ anyone using gitfs with salt?
23:17 shadowsun yes
23:17 jer_ I am having trouble adding a .deb file and having it show up on our master
23:17 jer_ and am wondering if this is ignored somewhere
23:18 shadowsun No, it's not
23:18 shadowsun How have you added it?
23:18 shadowsun Also, define "show up" :)
23:18 jer_ heh, sorry :)
23:19 jer_ yah so I have a salt state that I have added elastisearch
23:19 jer_ in this directory I added the most current deb file from there downloads
23:19 jer_ the added and commited this to our salt rep
23:19 jer_ repo
23:19 shadowsun did you push it?
23:19 jer_ for sure
23:19 shadowsun Cool
23:19 shadowsun pastebin your state file?
23:20 jer_ I just had a coworker pull the latest and it is there
23:20 jer_ actually
23:20 shadowsun yeah, I do that all the time too
23:20 jer_ ya one sec
23:23 jer_ oh shoot one minute I changed it in a deperate attempt to download it
23:26 defunctzombie where do I specify the external logger configuration?
23:26 defunctzombie I tried putting it in the minion conf but that didn't seem to work :/
23:27 StDiluted joined #salt
23:28 jer_ http://pastebin.com/EZQtgHnG
23:29 jer_ i have verified that deb file is in the repo
23:29 sgflt joined #salt
23:37 Thiggy joined #salt
23:40 ovorp joined #salt
23:43 sgflt if i all want to run is local salt scripts (salt-call --local state.highstate), which ubuntu package do i need? minion, master or syndic?
23:45 forresta just minion
23:46 jer_ so I just verified it is not a gitignore
23:47 jer_ I added a txt file and that is not updating either
23:47 jer_ I am misssing something
23:52 pcurry_nomi joined #salt

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