Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-04-25

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

All times shown according to UTC.

Time Nick Message
00:03 drags joined #salt
00:03 Teknix joined #salt
00:05 ajw0100 joined #salt
00:07 joehoyle joined #salt
00:12 shawnjgoff joined #salt
00:13 nebuchad` joined #salt
00:17 ajw0100 joined #salt
00:18 dramagods joined #salt
00:19 Teknix joined #salt
00:24 xinkeT joined #salt
00:24 fllr joined #salt
00:31 joehoyle joined #salt
00:32 Ryan_Lane1 joined #salt
00:32 jbub joined #salt
00:32 Teknix joined #salt
00:33 mway joined #salt
00:34 Teknix joined #salt
00:39 ianawilson_ joined #salt
00:40 abele_ joined #salt
00:41 Sypher joined #salt
00:42 herzi__ joined #salt
00:43 KennethWilke joined #salt
00:44 BrendanGilmore joined #salt
00:44 wiqd_ joined #salt
00:45 gldnspud__ joined #salt
00:45 mgw joined #salt
00:46 _jslatts joined #salt
00:47 balltongu_ joined #salt
00:48 jgelens joined #salt
00:48 Jarus joined #salt
00:48 rocket joined #salt
00:51 fllr joined #salt
00:52 drags joined #salt
00:52 blueblade joined #salt
00:54 RandalSchwartz joined #salt
00:54 RandalSchwartz I'll be neck-deep in Salt this weekend
00:54 joehoyle_ joined #salt
00:55 ajw0100 joined #salt
00:57 jimklo_ joined #salt
01:00 it_dude joined #salt
01:01 UForgotten RandalSchwartz bring extra water so you don’t get dehydrated :)
01:02 eliasp if anyone's bored: I could use a short review: https://github.com/saltstack/salt/pull/12279 … will take care of feedback in ~6h
01:02 eliasp thanks a lot
01:05 stanchan joined #salt
01:08 ldlework left #salt
01:09 pydanny joined #salt
01:11 vbabiy joined #salt
01:13 Steve_ joined #salt
01:13 Steve_ Trying to get cp.push_dir to work but I'm getting 'cp.push_dir' is not available.
01:13 Steve_ I do have file_recv: True
01:13 cb joined #salt
01:14 Steve_ This is on version salt 2014.1.1
01:16 gw joined #salt
01:16 Steve_ Looking at the source I see its not actually in that version
01:16 Steve_ Has this been released?
01:18 Steve_ maybe its in 2014.1.3 ?
01:19 whiteinge [MT]: nice! glad you got it working. you're in uncharted waters, so i'd be interested to hear how you like that setup
01:25 xzarth_ joined #salt
01:32 Ryan_Lane1 joined #salt
01:35 doanerock joined #salt
01:36 cast joined #salt
01:36 Sypher joined #salt
01:36 jeddi I'm a bit stuck with adding users to a group - or rather, forcing users (using members: - i was using addusers: before but that wasn't working, either).   Paste http://rn0.me/show/EZOtzdN31dRnC8k6E4EZ/ - showing debug-level logs on minion  ... any clues?
01:40 ajw0100 joined #salt
01:40 eliasp joined #salt
01:45 rome joined #salt
01:45 bhosmer_ joined #salt
01:45 joehoyle joined #salt
01:46 halfss joined #salt
01:49 ajolo joined #salt
01:51 j4son joined #salt
01:53 ajolo_ joined #salt
02:03 ajw0100 joined #salt
02:08 fllr joined #salt
02:09 mateoconfeugo joined #salt
02:12 feiming joined #salt
02:18 joehoyle_ joined #salt
02:18 CeBe joined #salt
02:21 [MT] whiteinge: I was'nt going to ping you again, but... it's actually not working 100%, the authentication happens. I can execute things with salt-call cmd, but what I try to push something to the minion with salt 'minion' cmd, there's an auth failure in the logs
02:22 ninkotech joined #salt
02:23 dramagods joined #salt
02:23 mateoconfeugo joined #salt
02:26 Networkn3rd joined #salt
02:29 djinni` joined #salt
02:30 jperras joined #salt
02:39 Luke_ joined #salt
02:45 bhosmer joined #salt
02:47 ninkotech joined #salt
02:48 rome joined #salt
02:50 alekibango joined #salt
02:52 rome joined #salt
02:57 anuvrat joined #salt
02:59 CeBe1 joined #salt
03:08 catpigger joined #salt
03:09 rome joined #salt
03:14 rome joined #salt
03:15 joehoyle joined #salt
03:17 ajolo__ joined #salt
03:22 ajolo joined #salt
03:23 dramagods joined #salt
03:27 feiming joined #salt
03:28 meteorfox joined #salt
03:33 jalaziz joined #salt
03:34 Ryan_Lane joined #salt
03:37 possibilities joined #salt
03:38 joehoyle_ joined #salt
03:38 rome joined #salt
03:39 ajolo_ joined #salt
03:39 rome joined #salt
03:39 joehoyl__ joined #salt
03:43 ajolo__ joined #salt
03:47 halfss joined #salt
03:53 robinsmidsrod joined #salt
03:54 MZAWeb joined #salt
04:01 mapu joined #salt
04:03 mateoconfeugo joined #salt
04:07 l0x3py joined #salt
04:10 Ryan_Lane joined #salt
04:13 mway joined #salt
04:14 kermit joined #salt
04:15 UtahDave joined #salt
04:15 jeremyBass joined #salt
04:15 jeremyBass left #salt
04:18 mway has there been any recent discussion about including a cloud provider for docker (in addition to the existing dockerio module) so containers can be provisioned the same way other cloud VMs are?  all I can find is the original discussion before dockerio was initialled added, so not sure if anyone else sees value in having that ability
04:22 mway seems like it might get a little messy and not necessarily fit into the same structure as other providers (just starting to do some digging), any thoughts?
04:22 mateoconfeugo joined #salt
04:24 dramagods joined #salt
04:28 UtahDave mway: I don't recall exactly how those discussions went.  Writing a salt-cloud driver isn't very difficult.
04:28 UtahDave It might be interesting to explore how that would work with Docker
04:28 srage joined #salt
04:29 MZAWeb that would be really really cool
04:29 mway yeah, I'll try to ask during the day as well to maybe get a few more eyes on it for overall discussion
04:30 srage joined #salt
04:30 mway my concern so far is, is there an issue with it effectively only implementing a subset of other providers' functionality, e.g. regions/predefined images and sizes/etc, since it's sort of arbitrary based on whatever you have/need built
04:33 mway might get a little meta… but it's a major use case for me at least, so unless there's a dealbreaker or general opposition, I'll definitely poke around and see what can be done
04:35 srage joined #salt
04:36 UtahDave Yeah, I see no problem with exploring and seeing how that works out.
04:51 anuvrat joined #salt
05:00 meteorfox joined #salt
05:02 malinoff joined #salt
05:03 joehoyle joined #salt
05:03 yusuket joined #salt
05:05 jchen i'd like to be able to manage my users' ssh pubkeys via pillar, but I don't want to have to maintain the pillar file every time a key gets published, but would instead like to use my keys that are published at https://github.com/<user>.keys. What's the recommended way to do this?
05:06 layer3switch joined #salt
05:06 jchen in addition, I'm using the salt-formulas/users-formula
05:08 jchen maintaining a list of pubkeys in a yaml file seems pretty silly :(
05:09 bmonty joined #salt
05:10 layer3switch joined #salt
05:10 ajolo__ joined #salt
05:15 jeddi jchen: i do something a bit messier using ssh public keys that i loop through in jinja on the /srv/salt side ... I'm not convinced there's a beautiful way of doing this yet.
05:16 jeddi I'm a bit stuck with adding users to a group - or rather, forcing users to be present using members: (I was using addusers: before, but that didn't work).   Paste http://rn0.me/show/EZOtzdN31dRnC8k6E4EZ/ - showing debug-level logs on minion  and before-and-after (without change)... any clues?
05:17 jeddi jchen: are all your users' keys published on github btw?
05:17 jchen jeddi: yeah, I was thinknig I would just jank it up and just create the dir and do a file.managed to download the file from github?
05:17 jchen jeddi: my keys are all published on github
05:17 jchen rn i don't care about multiple users, but the issue is still kinda the same.
05:18 jeddi jchen: yup - was thinking about that as an option here, it's a useful pattern for a few other things - maintaining a git repo of keys, f.e, and extracting that as needed.  means some crufty wrapper around maintaining a separate repo .. if you go that path.
05:18 jchen you dont' need a git repo of keys
05:18 jeddi what mechanism would you pull the github.com/xyz.keys out with?
05:18 jchen if you add your keys to the "authorized keys" on github, it publishes the list at https://github.com/<user>.keys
05:18 jchen oh
05:18 jchen file.managed?
05:19 jchen cant you do a source of https://github.com<user>.keys?
05:19 jeddi ooh - not sure.   only used sources of salt://fs/... (ie. local)
05:19 jeddi oh, no, there you are - you can use http or https ..
05:20 jeddi http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html - half way down in salt.states.file.managed // source
05:20 jeddi jchen: that might actually be a quite lovely approach.
05:21 jchen yeah, that woudl be great
05:21 jchen might consider submitting a patch to make users-formula smarter.
05:21 jchen vim apparently doesn't like it when one line in your file is a long-ass pubkey
05:21 Sypher joined #salt
05:23 yomilk joined #salt
05:24 flebel joined #salt
05:24 jeddi i'm sure vim doesn't mind it too much :)
05:25 Jarus joined #salt
05:25 dramagods joined #salt
05:27 jchen not according to the 100% cpu each time i cursor over teh one line :)
05:28 jchen also, ugh, salt.states.ssh_auth seems a bit clunky
05:29 * jchen throws up his hands and just uses file.managed
05:36 UtahDave jchen: this was just merged into develop:  https://github.com/saltstack/salt/pull/12236
05:37 jchen <3 UtahDave
05:42 UtahDave it's not in a release yet
05:46 Corey Wow, their API exposes it?
05:48 halfss joined #salt
05:52 ajw0100 joined #salt
05:56 jchen Corey: yes
05:56 jchen it's really useful.
06:04 ravibhure joined #salt
06:14 ndrei joined #salt
06:15 gildegoma joined #salt
06:18 xl1 joined #salt
06:20 slav0nic joined #salt
06:20 slav0nic joined #salt
06:21 sgviking joined #salt
06:26 dramagods joined #salt
06:27 nebuchadnezzar joined #salt
06:27 nebuchadnezzar .
06:29 nebuchadnezzar joined #salt
06:34 TyrfingMjolnir joined #salt
06:35 ajw0100 joined #salt
06:46 bram_ joined #salt
06:50 bram_ hi everyone, what's the easiest way to figure out when a fix is released in a stable release?
06:57 Kenzor joined #salt
07:00 Sypher This sls: http://paste2.org/ZyDAJVP2 results in annoying traceback: http://paste2.org/dv4GmVhZ
07:01 Sypher for some reason the for falls under the user, even though they are separate
07:01 otsarev_home joined #salt
07:03 jgelens joined #salt
07:07 DaveQB joined #salt
07:12 harobed joined #salt
07:13 ml_1 joined #salt
07:13 it_dude joined #salt
07:19 anuvrat joined #salt
07:26 dramagods joined #salt
07:44 gammalget joined #salt
07:49 halfss joined #salt
07:56 Ryan_Lane1 joined #salt
08:04 Konrad_ joined #salt
08:06 topochan joined #salt
08:27 dramagods joined #salt
08:38 slav0nic joined #salt
08:38 slav0nic joined #salt
08:56 Daemonik joined #salt
08:57 giantlock joined #salt
08:57 Daemonik Can the bootstrap.sh script point to an alternative (for me, an internal) git repo?
08:57 Daemonik github.com[0: 192.30.252.131]: errno=Connection timed out
08:57 Daemonik fatal: unable to connect a socket (Connection timed out)
08:57 Daemonik Initialized empty Git repository in /tmp/git/salt/.git/
08:57 Daemonik * ERROR: Failed to run install_centos_git_deps()!!!
08:57 Daemonik You have new mail in /var/spool/mail/root
08:57 Daemonik =(
09:00 millz0r http://boostrap.salt.org/
09:00 millz0r Daemonik: tried that?
09:02 londo__ joined #salt
09:02 millz0r might be a stupid question
09:03 Whissi joined #salt
09:05 Daemonik millz0r, I got the bootstrap.sh script from bootstrap.saltstack.org ; bootstrap.sh attempts to connect to github.com and raw.github.com, github.com is timing out, so I'm wondering if I can git pull from git hub, then specify an internal git repo, at least until I get around to adding a Jenkins task to git pull Salt and create RPMs and DEBs for internal use.
09:08 dramagods joined #salt
09:10 millz0r oh
09:11 millz0r you can actually just download the deb pacakge
09:11 millz0r and use dpkg to install it
09:11 millz0r the bootstrap script doesnt do much magic
09:11 millz0r http://docs.saltstack.com/en/latest/topics/installation/debian.html
09:20 ggoZ joined #salt
09:22 Daemonik There are debs available for v2014.1.3 ?
09:23 che-arne joined #salt
09:26 kvbik joined #salt
09:36 ravibhure joined #salt
09:37 faldridge joined #salt
09:49 dramagods joined #salt
09:50 halfss joined #salt
10:09 millz0r there is 1.3?
10:12 che-arne joined #salt
10:15 malinoff millz0r, https://groups.google.com/forum/?pli=1#!topic/salt-users/g-JFfSi4ei0
10:16 millz0r nice
10:16 millz0r good to see them ironing bugs out
10:23 f0086 joined #salt
10:24 f0086 I have some problem understanding the pacman module (https://github.com/saltstack/salt/blob/develop/salt/modules/pacman.py)
10:25 f0086 Installing a package can take some time (sometimes minutes)
10:25 f0086 How does the server knows the progress of the installation?
10:25 f0086 I can't find any hint in the code
10:28 malinoff f0086, the server (aka salt master) does not know anything about remote execution progress
10:28 N-Mi joined #salt
10:28 N-Mi joined #salt
10:29 f0086 I saw some progress bar in the web UI in the webinar yesterday
10:30 f0086 malinoff: Is that just some estimate fake-progress-bar?
10:30 N-Mi joined #salt
10:30 N-Mi joined #salt
10:31 malinoff f0086, can you point me to the video? In v17.x salt could not do that
10:32 malinoff Probably, something changed
10:32 f0086 malinoff, http://www.saltstack.com/saltstack-events/saltstack-demo-webinar
10:32 f0086 This was a webinar
10:33 malinoff Unfortunately, I can't find it's video so I don't really know
10:34 mekstrem joined #salt
10:34 malinoff It's better to ask your question on https://groups.google.com/forum/?pli=1#!forum/salt-users mailing list
10:36 f0086 malinoff, http://static.squarespace.com/static/524cf70fe4b05018590c3fb3/t/52a0043fe4b0f564cbc010ec/1386218563577/prog_events_png.png (the blue progress bars)
10:36 malinoff Ah, this
10:36 blub_ joined #salt
10:38 malinoff As I think, this bar is a fake - it shows you how many single states were applied versus the number of all states that must be applied
10:38 malinoff But not a progress of a single state
10:39 f0086 malinoff, ah thanks. That makes sense. In the module itself there is no sign for "status update" logic
10:39 malinoff Yep
10:45 f0086 If I execute multible commands at once to a minion which will take ... say 10 minutes each ... will the commands run in parallel?
10:46 malinoff f0086, yes, they should run in parallel
10:46 malinoff How many commands do you run at once?
10:46 f0086 And if so, can I define limits for certain commands/modules (only execute this command 2 in parallel but this can be run 50 in parallel)?
10:47 f0086 malinoff, hm, it depends ... 2-100
10:50 malinoff I was sure salt is running about 5 workers to handle commands...
10:50 malinoff But I can't find this right now
10:50 malinoff Give me a sec
10:53 malinoff Ops
10:53 malinoff Nope
10:53 malinoff It does not
10:53 malinoff Lol
10:54 malinoff Or I can't find this
10:54 malinoff Ah
10:54 malinoff Alright
10:54 malinoff https://github.com/saltstack/salt/blob/develop/salt/minion.py#L788-L800
10:54 malinoff It will launch a separate process/thread to run a command
10:54 f0086 the reason i want this is we have some tasks which are strain cpu or disk. Such tasks shouldn't run to often/much in parallel (to strain the complete server) but other tasks like "show me the uptime" or "change this setting in some config file" are blazing fast and can run 100x in parallel to fasten up reaction time
10:55 f0086 ah, thanks!
10:57 f0086 I think I need to do this logic outside of salt :/
10:57 malinoff f0086, yeah, seems so
11:01 bhosmer joined #salt
11:09 joehh1 Daemonik: for both ubuntu(ppa) and debian (debian.saltstack.com)
11:16 logix812 joined #salt
11:26 Matt_ joined #salt
11:26 ajw0100 joined #salt
11:26 ml_1 joined #salt
11:26 Guest49953 \nick MattCoffey
11:26 Guest49953 forgot how to use irc haha
11:28 ajw0100 joined #salt
11:28 MattCoffey I want to parameterise my sls file to differentiate between test and production so that it can do things slightly differently, I think I have to write a custom grain
11:29 logix812 Aside from doing calls on the system itself, can you get the grains of the salt master? (specifically looking for it's IP)
11:29 ajw0100 joined #salt
11:30 MattCoffey I am a developer trying to develop deployment automation and I am starting off by testing my sls file with standalone masterless minion
11:30 MattCoffey maybe that is a bad idea
11:31 MZAWeb joined #salt
11:31 ajw0100 joined #salt
11:32 ajw0100 joined #salt
11:34 ajw0100 joined #salt
11:35 ajw0100 joined #salt
11:37 ajw0100 joined #salt
11:38 mikkn MattCoffey: That's how I do it too. It does limit you a little in that from what I've gathered, you can't use ext_pillars for instance
11:39 ajw0100 joined #salt
11:40 MattCoffey thanks, do you have a really simple example of a custom grain that just returns a flag stating whether it's production or not?
11:40 ajw0100 joined #salt
11:40 MattCoffey trying to bootstrap myself on this stuff
11:41 jY MattCoffey: we have a grain that just grabs part of a hostname for prod or dev
11:41 jY our hostnames are like web1.datacenterlocation.domain.net
11:41 MattCoffey good call!
11:41 MattCoffey that can work for me
11:41 mikkn MattCoffey: Yeah, we do that too
11:42 ajw0100 joined #salt
11:42 jY so we have a company_env grain that says production/staging/qa
11:43 mikkn Well, pretty much. We have hostname.prod.datacenter.domain.com, so the type of env is straight in the host name
11:43 ajw0100 joined #salt
11:43 ndrei joined #salt
11:45 ajw0100 joined #salt
11:46 ajw0100 joined #salt
11:48 ajw0100 joined #salt
11:49 ajw0100 joined #salt
11:50 halfss joined #salt
11:51 ajw0100 joined #salt
11:51 elfixit joined #salt
11:53 ajw0100 joined #salt
11:54 ajw0100 joined #salt
11:55 taion809 joined #salt
11:56 ajw0100 joined #salt
11:57 ajw0100 joined #salt
11:59 ajw0100 joined #salt
12:00 topochan joined #salt
12:00 ajw0100 joined #salt
12:01 workingcats joined #salt
12:01 jpcw joined #salt
12:01 A||SySt3msG0 joined #salt
12:01 halfss joined #salt
12:02 ajw0100 joined #salt
12:04 ajw0100 joined #salt
12:17 kiran2 joined #salt
12:19 Gordonz joined #salt
12:20 rome joined #salt
12:22 mgw joined #salt
12:23 kiran2 hello guys, I have written my first custom execution module. Synchronization of the module from the master to registered minions with the command 'salt my_minion sysutil.sync_modules' works and I can see it then on the minion under /var/cache/salt directory . What I don't understand is how to call it from the master. I can run it from the minion with 'salt-call my_module param1 param2 ...' , I can see it in the list of modules with'sa
12:24 kiran2 But when I run 'salt my_minion sys.list_modules' it's not there and so I can't eve run it like 'salt my_minion my_module param1 param2 ...'
12:25 kiran2 So my question is whether such behaviour is normal one or have I forgotten something?!?
12:25 kiran2 I meant when I run it from the master then, it doesn't work.
12:26 malinoff kiran2, try to restart minion
12:28 kiran2 OMG!
12:28 kiran2 thanks malinoff, any idea why it's required to restart it?!?
12:28 malinoff kiran2, does that help? :)
12:29 ckao joined #salt
12:29 kiran2 yep!
12:29 malinoff I think, salt does not properly load your module, but when you restart your minion, minion can properly load your module
12:30 malinoff There were some issues about that
12:33 rome joined #salt
12:38 harobed joined #salt
12:38 micko joined #salt
12:45 toastedpenguin joined #salt
12:46 jslatts joined #salt
12:48 ipmb joined #salt
12:49 doanerock joined #salt
12:52 rome joined #salt
12:52 feiming joined #salt
12:54 tru_tru joined #salt
13:02 gadams999 joined #salt
13:06 wincus joined #salt
13:07 rome joined #salt
13:08 rome_ joined #salt
13:09 giannello joined #salt
13:10 diegows joined #salt
13:12 shawnjgoff joined #salt
13:15 racooper joined #salt
13:15 Rory joined #salt
13:17 Rory How do I use the Python client API (LocalClient) to run a command on hosts matching a grain? Ie emulate "salt -G 'foo:bar'"
13:18 Rory All the docs just use '*' as an example
13:18 eliasp Rory: are you looking for this?: http://docs.saltstack.com/en/latest/topics/targeting/compound.html
13:18 nix2170 joined #salt
13:19 darrend joined #salt
13:19 Rory eliasp: Probably
13:19 Rory eliasp: Yes I know how to do it on the CLI
13:20 eliasp Rory: ah, so you're talking about using it programatically…
13:20 Rory But if I use '-G foo:bar' as a tgt it says "No matching Targets..."
13:20 mway joined #salt
13:21 Networkn3rd joined #salt
13:21 logix812 Rory: how are you calling it programmatically: myclient.???
13:22 Rory Well that's what I'm asking how to do. myclient.cmd()
13:22 Rory Docs say "by default a target is passed in as a bash shell glob"
13:22 logix812 myclient.cmd('*', 'test.fib', [10])
13:22 Rory but client.cmd('-G foo:bar', 'test.ping') doesn't work with the error: "No Matching Targets for glob : -G foo:bar"
13:23 logix812 hmm
13:23 logix812 and it matches if you run that command from the cli?
13:23 Rory And all the docs just use '*' as an example, but I need to do something to all hosts matching a grain
13:23 Rory Which I assume is a common pattern
13:24 Rory It seems weird to actually put a '-G' in there, so I was asking if there's a proper way
13:24 logix812 Rory you see this: https://github.com/saltstack/salt/blob/develop/salt/client/__init__.py#L478-L488
13:25 logix812 Rory: I don't have an answer I am just looking myself for the first time
13:26 Rory OK so I want something like: client.cmd('expr_form=grain...
13:27 faldridge joined #salt
13:27 logix812 I think you want: client.cmd('foo:bar', 'test.ping', expr_form='grain')
13:27 Rory Yeah I'm thinking so
13:28 logix812 That work?
13:28 mapu joined #salt
13:29 bastion1704 joined #salt
13:31 bram_ noob question: I have an already large state+pillar system set up with a highstate... however, for development I would rather just run the state I'm working on right now instead of writing sudo salt-call --local state.highstate all the time. How can I apply one particular state and still get my pillar data?
13:31 Rory bram_: salt-call state.sls foo
13:32 Networkn_ joined #salt
13:32 bram_ Rory, YAY! :)
13:32 faldridge joined #salt
13:32 bram_ I should have asked this two weeks ago :D :D
13:33 Rory highstate time is checking Reddit time
13:33 vbabiy joined #salt
13:33 mpanetta joined #salt
13:33 bram_ :)
13:33 bram_ just like vagrant up
13:34 vejdmn joined #salt
13:36 giannello what's the best way to pass parameters to the orchestrate runner?
13:36 sroegner joined #salt
13:37 tkharju3 joined #salt
13:37 giannello let's say I want to define a few steps to be executed to startup a well-defined infrastructure for my project, and I want to spin up a production and a staging instance
13:37 giannello is there a way to pass this parameters at runtime?
13:37 giannello or when invoking the runner?
13:38 jalbretsen joined #salt
13:39 seb____ joined #salt
13:39 pydanny joined #salt
13:42 Rory And yes, thanks bram_ that works
13:42 Rory The syntax I ended up using is: client.cmd('grain_name:value_to_match', 'test.ping', expr_form='grain')
13:43 tkharju4 joined #salt
13:44 donatell0 joined #salt
13:48 mateoconfeugo joined #salt
13:48 pdayton joined #salt
13:49 herzi_ is there a way to set --force-yes as a cmdline argument for salt?
13:50 herzi_ (for apt) I was trying to replace sysvinit w/ systemd-sysv and I get this error:
13:50 herzi_ E: There are problems and -y was used without --force-yes
13:50 logix812 if I want to issue a wheel command over the salt-api (cherrypy), is the workflow: /login -> POST to / {'client': 'wheel', ...data...} -> /logout?
13:53 Rory herzi_: --force-yes is an argument for apt not salt
13:53 joehoyle joined #salt
13:58 techdragon joined #salt
13:59 jraby joined #salt
14:01 stritz joined #salt
14:01 herzi_ Rory: yes, i added that in my second line after realizing that i skipped „[for] apt with [salt]“ in the first line
14:02 dramagods joined #salt
14:02 Rory herzi_: Can you try adding "force_yes: True" to your state?
14:03 sijis joined #salt
14:03 sijis joined #salt
14:08 ajprog_laptop joined #salt
14:08 GradysGhost joined #salt
14:11 saltn joined #salt
14:12 thedodd joined #salt
14:13 saltn Does cloud.providers handle pillar data ?
14:13 vejdmn joined #salt
14:13 rome joined #salt
14:17 mgarfias joined #salt
14:19 N-Mi hi, I have a question I already asked like 1 year ago, but things may have been changed with all the features that have been added since. I try to define things on a minion, that will be used by another minion.
14:19 N-Mi for example, serverA defines how it needs to be monitored and backed up
14:20 N-Mi monitoring server gather all these information for each servers to generate its nagios conf
14:20 N-Mi backup server  do the same for backup configuration
14:21 N-Mi the idea is not to have to modify state/pillar in 3 places (server1, monitoring server, backup server)
14:21 Rory herzi_: Did it work?
14:23 N-Mi it could have been done easily with an extend mechanism  in pillar data, but this feature request (#3991) is not implemented
14:25 shawnjgoff Oy... everything I try to Google for salt results in garbage results. Is there a list of salt version names?
14:26 N-Mi I had a try with the new minion fileserver feature, but I see not way to do something like "salt monitoring-server cp.get_file salt://*/etc/config/monitoring/* /etc/monitoring/hosts/"
14:26 mrj joined #salt
14:28 mrj joined #salt
14:28 shawnjgoff There are references to version names all over the documentation, like this: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.test.html . "New in version Helium." I have no idea what that means because the package repositories for RedHat and Debian don't use those names.
14:29 jeremyBass1 joined #salt
14:33 nix2170 shawnjgoff: I usually have more luck googling for "salt stack" instead of just "salt".
14:33 shawnjgoff nix2170, I try salt first, then saltstack.
14:34 shawnjgoff That link above is the first result for "salt release version names helium", and it is a completely inappropriate response.
14:35 Ahlee 2014.1.0 is Hydrogen, the version to follow will be Helium
14:35 Ahlee I do not like the naming scheme either
14:35 Ahlee nor do I like the variation from semver
14:35 shawnjgoff Ah. Thanks.
14:35 Ahlee i believe it will be 2014.2.0 == Helkum
14:35 Ahlee er, helium
14:36 shawnjgoff At least they're going in order there.
14:37 andrein__ Hi guys, I'm setting up an Nginx server using saltstack. my sls file is at http://paste.opensuse.org/57516779 I want salt to manage all of the conf.d directory, but it seems it only makes sure that whatever is on my salt master in that directory is applied to the minions. Any ideas how to tell it to make sure only the files I want are present on the minion?
14:37 Ahlee true, but cute/clever shouldn't replace obvious
14:38 mpanetta Quick question... Is there any way to access minion grains from reactor?
14:38 mpanetta The reactor documentation does not make this obvious...
14:39 mpanetta I need to get the minions environment_id when it is created via salt-cloud...
14:39 mateoconfeugo joined #salt
14:41 alunduil joined #salt
14:43 Ahlee I do not believe you can pull values via reactor, I believe you'll need to use the mine
14:44 Ahlee and I do not know how that would be accomplished
14:44 mpanetta Hmm ok
14:44 mpanetta Yeah I am looking at mine now.
14:53 stritz joined #salt
14:53 topochan joined #salt
14:55 ml_1 joined #salt
14:55 timoguin i don't think Helium will be 2014.2. as far as i know the release numbers go by release date, i.e., 2014.4 if it's released this month
14:55 timoguin if not i don't think the numbering makes a lot of sense
14:59 taion809 joined #salt
14:59 giannello joined #salt
15:00 perfectsine joined #salt
15:02 ndrei joined #salt
15:03 ipmb joined #salt
15:03 conan_the_destro joined #salt
15:04 faldridg_ joined #salt
15:09 rome joined #salt
15:10 Rory How do I use client.cmd('*', cmd.run, [bash_command]) in cases where the output of bash_command is multiple lines, and I want to process them all?
15:10 MrTango joined #salt
15:12 Rory calling values() on the results of the above gives me one line with a \x01 character at the end
15:13 Rory I get the output when running on the CLI: salt '*' cmd.run "bash_command"
15:13 Rory but not when using the API
15:15 Rory I've discovered the problem is not that the output of bash_command is multi-line, but that it uses pipes
15:15 mway joined #salt
15:16 Mr_N Anyone here eloquent at Jinja?  Handwavy explanation, I have a list of pillar variables, and I want to do something for only the first "valid" one I find, then stop.
15:16 Mr_N But jinja doesn't allow for breaking out of loops
15:16 Mr_N and perhaps I'm uncreative, but I'm running in circles looking for other patterns to represent this as.
15:17 Mr_N I could go all the way to raw python in the state I guess, but I'd really rather not if I can avoid it.
15:22 faldridge joined #salt
15:23 happytux joined #salt
15:23 gothix_ joined #salt
15:31 paul75 joined #salt
15:32 paul75 sorry for the stupid question: my top.sls is located  in salt but salt/base how do I get salt-call to find it?
15:33 paul75 Not located in salt but salt/base  argggh
15:34 timoguin paul75: for masterless you either set 'file_client: local' in the minion config or pass --local at the CLI
15:35 conan_the_destro joined #salt
15:37 paul75 timoguin:  Sorry I'm using --local   The problem is the top.sls file is not in the salt directory, but salt/base
15:38 bram_ hey everyone, can state require something more highlevel than a single state? I.e. is it possible to require a complete SLS to be correctly installed - for example?
15:38 qybl joined #salt
15:39 kickerdog joined #salt
15:41 timoguin bram_: you can call state.sls to run a single formula, and there's state.highstate to run all the SLS formulas assigned to a minion
15:41 paul75 Sorry for the repeated question but my typo probably confused the issue so here it is again:  my top.sls is NOT located  in salt but salt/base how do I specify the directory in salt-call so that it finds it?
15:41 mpanetta Are there any full examples around that show how to pull a particular grain from a minion using mine?
15:41 bram_ timoguin, mmm, not really what I'm asking for
15:42 schimmy joined #salt
15:42 possibilities joined #salt
15:42 bram_ timoguin, what I mean is, when you list your requirements in a state after "require:", can you use requirements which are more "high level" than other single states
15:44 taterbase joined #salt
15:44 timoguin bram_: ah, i see what you mean. other than including an SLS in another, i'm not sure if that feature is available
15:44 timoguin paul75: edit your file_roots in the minion config
15:44 timoguin not sure if you can pass it on the CLI
15:45 paul75 k thanks
15:45 bram_ meh the problem is just that the project I'm trying to saltify has many spaghetti-dependencies :/
15:45 bram_ no clear separation of parts
15:46 jgelens joined #salt
15:46 Jarus joined #salt
15:46 cb joined #salt
15:46 wendall911 joined #salt
15:47 ajw0100 joined #salt
15:53 nix2170 I'm trying to match in top.sls but I can't get the syntax right because of the space. <problem>'G@osfinger:Scientific Linux-6':\n  - match: compound</problem>
15:53 [MT] perhaps osfinger:Scientific\ Linux-6
15:54 vejdmn joined #salt
15:54 andrein__ Hi guys, how can I run a command when the contents of a directory change?
15:55 nix2170 <[MT]> That didn't error but also didn't match.
15:55 nix2170 I have also tried Tried 'G@osfinger:"Scientific Linux-6"': w/o success.
15:55 Ahlee andrein__: Salt isn't really set up that way.  You'd likely need to set up an inode watcher on the minion, and then salt-call from the minion when it changes
15:56 tligda joined #salt
15:58 andrein__ Ahlee: I'm trying to use this to regenerate the repo data in an rpm repository (just in case someone forgets to regenerate it when they drop an rpm there). I'm currently managing only the repository root, not it contents using salt.
15:59 andrein__ can you elaborate on the inode watcher?
15:59 nix2170 FYI: this is not what I wanted but it works: 'G@osfinger:"Scientific*Linux-6":
15:59 Ahlee andrein__: Salt has no knowledge of the rpm changing, so you'd have to set up some process to watch the directory and react accordingly
15:59 nix2170 Sorry I meant  'G@osfinger:Scientific*Linux-6:
16:00 giannello joined #salt
16:00 jalbretsen joined #salt
16:01 andrein__ Ahlee: so, salt has no knowledge that the contents of the directory changes since the last scheduled run?
16:01 bastion1704 joined #salt
16:01 Ahlee andrein__: it will pick up on the next run, but it has no native ability that i'm aware of to react to the run between scheduled runs
16:02 Ahlee so like, you could set up a process using inotify watchers (pynotify for example), and on change trigger salt to run via salt-call, but that's ugly and failure prone
16:05 jslatts does anyone know if an s3 external pillar is in the works?
16:05 andrein__ Ahlee: I guess I'll just have to depend on humans for regenerating the metadata when they change stuff :)
16:05 Ahlee andrein__: just cron createrepo
16:05 Ahlee :)
16:11 [MT] If I have a master set up and a syndic and the syndic is pointing at the master, how do I use the master to talk through the syndics to the minion?
16:13 tkharju joined #salt
16:13 gildegoma joined #salt
16:14 meteorfox joined #salt
16:14 viq [MT]: syndic is irrelevant
16:14 timoguin [MT]: as far as i know, the syndic is a pass-through, so you should be able to send commands from the master normally
16:14 viq [MT]: just like you would normally do
16:14 timoguin i'm answering this one, viq!
16:14 timoguin ;-)
16:14 KyleG joined #salt
16:14 KyleG joined #salt
16:14 viq timoguin: like hell you are, I was first ;P
16:15 mpanetta hahaha
16:15 TyrfingMjolnir joined #salt
16:15 [MT] timoguin: I'm not able to. On the master of masters, test.ping never shows success from anything, but from the syndic servers they're connected to, test.ping returns right away
16:16 [diecast] joined #salt
16:17 [MT] the syndic config is part of the master loading, right? So with salt-master and salt-syndic installed, I can put the config in /etc/salt/master.d/syndic.conf, right?
16:18 viq [MT]: yeah, that should work. Though did you configure both the syndic to tell it where the master is, and the master to tell it it is to support syndics?
16:19 [MT] "order_masters: True" <-- on the master ;; "syndic_master: salt.corp.domain.tld" <-- on the syndic
16:19 viq yeah
16:20 [MT] yup, got that
16:20 viq [MT]: also I don't know with 2014.1 but with 0.17 I had the minion on syndic be connected to syndic itself instead of higher level master for the communication to other minions to work
16:21 [MT] these boxes are all connected to the syndic and not the higher master
16:21 viq I mean if you're running a minion on the syndic itself
16:22 Ansem joined #salt
16:22 Ansem Hey, I'm having redis problems on Precise64.
16:22 Ansem The service is up and running but unreachable. I've added service.restart to my state, but it doesn't seem to be doing anything?
16:22 Ansem When I ssh in and restart the service, it's fine, but salt can't seem to get it working
16:22 Ansem any advice?
16:23 viq timoguin: for the record, if it wasn't clear, I was joking ;)
16:24 timoguin viq: nope, we are un-friends now.
16:24 viq Now I have a question. When you install certificates on system, you often need to symlink them for stuff to see them, you can get the name of what the symlink should be called with something like "openssl x509 -noout -hash -in syslog-ng.cert" - how would I automate that with salt?
16:25 viq timoguin: aww, and I don't even have facebook where you could unfriend me to reflect that, double offence!
16:25 jeddi viq: do you have meow-meow-beans?
16:26 timoguin Ansem: i think i've experienced that before too. for me it was because there was no upstart script installed for redis, which is what the service module was trying to use.
16:26 viq jeddi: my mother has cats, does that count?
16:26 jeddi viq: to answer your question - i'd (but i'm a brute force kinda guy) write a small bash script, push it out, run it, and then set a flag somewhere to say it's done (or just rely up an 'unless: ls -l ...' (the symlink file))
16:26 jeddi viq: it does not count.
16:27 * jeddi is very much looking forward to the (very much bigtime impending) death of upstart.
16:27 timoguin Ansem: https://gist.github.com/bdotdub/714533
16:27 viq jeddi: I was thinking of that as a possible option, but looking if there's a cleaner way to do it
16:27 Ansem Thanks timoguin
16:28 jeddi viq: step 1 - get it running, no matter how cruftily.  step 2 - tidy it up.
16:28 viq jeddi: yeah
16:28 jeddi viq: remember, you get extra points in step 1 if it's spectacularly crufty.
16:28 [MT] viq: ok, that makes sense
16:31 [MT] viq: all minions are on the syndic, I have them configured how I mentioned, I can test.ping minions from the syndic the minions are connected to, I can't test.ping them from the master; using 2014.1.3 on minions, syndic, and master
16:31 viq jeddi: I wonder if something like {{ salt['cmd.run']('name of script to get that hash') }} would work as - name in file.symlink
16:32 viq [MT]: machine that is running salt-syndic process - does it have a salt-minion process as well?
16:32 [MT] ya
16:32 mpanetta ARgh... Ok I am trying to use recator, but I am getting the following error: [ERROR   ] Failed to render "/srv/salt/reactor/update_hosts.sls"
16:32 [MT] and the minion is pointed at the syndic address
16:33 mpanetta I have enabled debugging in the master and that is still all it says
16:33 mpanetta How can I see what the failure was?
16:33 viq [MT]: and can you test.ping the syndic from the master?
16:34 mpanetta I'm trying to use mine to get an env so I can set it in the reactor formula, but it is apparently failing :(
16:34 [MT] viq: nope
16:34 mpanetta Anyone have any ideas?
16:34 viq [MT]: does the master show in logs connections from the syndic?
16:35 mpanetta timoguin: Any ideas?
16:36 [MT] viq: it didn't, I'm running the master now with -l debug
16:36 [MT] viq: nope- nothing showing up
16:37 [MT] gotta run for a sec- brb
16:37 tkharju joined #salt
16:38 mpanetta hrm, this is frustrating...
16:39 viq mpanetta: can you show your /srv/salt/reactor/update_hosts.sls ?
16:39 timoguin mpanetta: so nothing in the logs except that line?
16:39 timoguin like... why it failed to render?
16:39 mpanetta https://gist.github.com/bentwire/91a38ff9d7614503b0f7
16:39 adizam joined #salt
16:39 mpanetta Yep, that is all it says.
16:40 mpanetta I just enabled trace level messages to see if that helps.
16:40 mpanetta viq: see gist above
16:40 UtahDave joined #salt
16:40 cruatta joined #salt
16:41 cruatta joined #salt
16:41 UtahDave joined #salt
16:42 [diecast] joined #salt
16:43 sarlalian joined #salt
16:43 viq mpanetta: shouldn't line 6 have == and not = ?
16:43 mpanetta OOPS
16:43 mpanetta Yes
16:43 mrj joined #salt
16:44 [MT] aaand back
16:44 mpanetta Hopefully that is the issue... Gonna take a minute to test it...
16:45 joehillen joined #salt
16:45 kballou joined #salt
16:46 kivihtin joined #salt
16:46 mpanetta viq: Yeah, that was an error, but apparently not what is causing the rendering failure :(
16:47 mpanetta And using trace level messages still does not say why it errored timoguin
16:47 mpanetta Any suggestions?
16:47 mpanetta brb
16:49 viq mpanetta: sorry, not really from me
16:49 viq Time to start the weekends ;)
16:51 mpanetta haha
16:51 mpanetta yeah, no kidding
16:52 mpanetta maybe UtahDave knows, he knows everything right? :P
16:52 UtahDave mpanetta: I missed the back log.  What's the error you're getting?
16:53 mpanetta [ERROR   ] Failed to render "/srv/salt/reactor/update_hosts.sls"
16:53 mpanetta But no messages saying why it failed
16:53 mpanetta so I don't know what to fix heh
16:53 mpanetta I'm trying to run a reactor that updates hosts in my env when a new device is added via salt-cloud
16:54 ldlework joined #salt
16:55 yomilk joined #salt
16:56 mpanetta UtahDave: Basically I am trying to figure out how I can see the error output from the render, I tried enabling debug and then trace level messages, but neither give me what I need.
16:57 UtahDave mpanetta: debug mode on the minion?
16:57 tkharju1 joined #salt
16:57 mpanetta On the master.
16:57 UtahDave Oh, the reactor
16:57 mpanetta Yeah
16:57 mpanetta It is the reactor formula that is failing to render, I assume that is done on the master correct UtahDave?
16:58 chrisjones joined #salt
16:59 UtahDave yeah, reactor is done on the master
16:59 stanchan joined #salt
16:59 UtahDave can you pastebin a sanitized version of your update_hosts.sls?
16:59 mpanetta sure
17:00 mpanetta https://gist.github.com/bentwire/f9824f90dc8f9cd53c62
17:01 schimmy joined #salt
17:02 blueblade joined #salt
17:04 mpanetta UtahDave: Only thing I can figure is that maybe mine.get isn't returning anything, but I can't tell for usre.
17:04 mpanetta *sure
17:05 [MT] viq: if I restart the master and syndic on the syndic server and have the master running in debug, I should see connections, right?
17:05 timoguin mpanetta: does the mine.get return properly on the CLI?
17:05 UtahDave mpanetta: Yeah, I would try putting in some dummy data first to see if the rest of it is working
17:05 mpanetta timoguin: It did at one time.
17:06 ajolo joined #salt
17:07 mpanetta UtahDave: I don't quite get what you mean.  Do you mean remove the jinja?
17:07 mpanetta Hmmm
17:09 jclarke joined #salt
17:09 mpanetta I think that formula may be running after the minion was deleted, that may be my issue.
17:09 UtahDave Yeah, put some good data in there instead of the mine.get call to see if it still works
17:09 mpanetta Trying to see if it gives the same issue when creating the minion...
17:09 mpanetta UtahDave: Oh, yeah I did that before trying to use mine, it worked fine.
17:10 UtahDave ok, so you're thinking the minion may already be deleted?
17:10 Guest29771 i
17:11 Guest29771 left #salt
17:11 mpanetta Yeah, here is my reactor.conf: https://gist.github.com/bentwire/250a690fa1fbf7c718e1
17:11 mpanetta I am trying to see if it works on minion start now.
17:12 bastion1704 joined #salt
17:12 ajolo joined #salt
17:14 defmikekoh joined #salt
17:14 dstanek trying is use salt-cloud with ec2 and i'm running into a KeyError - http://dpaste.com/1797135/
17:14 defmikekoh joined #salt
17:14 Guest29771 joined #salt
17:15 dstanek i have grains in my profile config, but not in my provider config because I assumed the profile grains would be inherited
17:15 dstanek am i wrong?
17:15 ghanima joined #salt
17:15 ghanima hello all......
17:16 Guest29771 Can someone tell how to update linux packages using salt ?
17:16 ghanima Guest29771: if you just want to use the standard runner
17:16 ghanima you could do something like salt '*' cmd.run "yum -y install package name"
17:17 ghanima Guest29771: or are you trying to build a state file
17:17 Ryan_Lane1 joined #salt
17:17 Guest29771 I'm building a statefile
17:17 timoguin use pkg.upgrade or pkg.install
17:17 ghanima Guest29771: oh my bad
17:19 Guest29771 Ok thanks
17:20 Guest29771 Uh wait won't "upgrade", upgrade the packages and not just update them ?
17:20 ghanima Guest29771: without understanding how you have salt deployed I found the State's tutorial pages really helpful. If your already familar with those I would say the simplest way would be create your top file top.sls in /srv/salt and then add your pkg in that file if you need for it to apply to all servers managed by your master
17:20 UtahDave dstanek: yeah, you can put grains in the provider config
17:20 schimmy1 joined #salt
17:21 ghanima But I recommend you look at the tutorials because it shows examples of how to do grouping based off of role
17:21 ghanima Guest29771: I hope that helps
17:21 dstanek UtahDave: the problem seems to be that if i don't have a minion section in my provider config i get the key error
17:21 [MT] timoguin: I don't suppose you're still hanging around?
17:22 UtahDave dstanek: Hm can you pastebin a sanitized version of your provider config?
17:24 dstanek UtahDave: http://dpaste.com/1797147/
17:25 dstanek UtahDave: i said that wrong; if i don't have grains in my profile config i get the keyerror
17:26 ghanima Question: does anyone have any experience with the reactor system..... Do I understand correctly that its a way to allow for Salt Master to check the state of a system based off of what's defined on the master and then send an alert that the state has changed and perform an action on it immediately
17:26 dstanek i added a minion entry here to make it work: http://dpaste.com/1797150/
17:26 ghanima It's like a preimptive monitoring solution for like process state I presume
17:26 UtahDave dstanek: hm, you should be able to put it in the provider as well
17:27 UtahDave ghanima: A reactor listens for events on the event bus and then executes the sls file you specify when it sees that event
17:27 UtahDave ghanima: the reactor sls files have access to the full power of Salt to do whatever you want
17:27 gadams999 joined #salt
17:29 Heartsbane is there a way in jinja to pull the hostname of the machine it is running on?
17:29 Heartsbane for a if statement
17:30 Heartsbane {% if hostname %}
17:30 Heartsbane do this thing
17:31 Heartsbane {% endif %}
17:31 Heartsbane Anyone have any good examples I could look at
17:31 Sypher Heartsbane: you should be able to grab that from the grains
17:32 nix2170 Heartsbane: salt['grains.get']('fqdn', '')
17:32 nix2170 oops that between {% and %} of course :)
17:32 arthabaska joined #salt
17:33 Gareth ahoy hoy.
17:33 Heartsbane thanks.
17:33 * Heartsbane needs stronger coffee.
17:33 nix2170 Heartsbane: you can also get a lot of other types of hostname. Run on master cli # salt 'minion' grains.ls for a list, then # salt 'minion' grains.get 'os' for example.
17:35 Heartsbane ya I have done that ... I don't know why I didn't think to look in the grain info
17:35 Heartsbane I typically use grain.item
17:36 jclarke joined #salt
17:36 jaimed joined #salt
17:37 jclarke @ghanima: I was logged in as guuest a minute ago i just thought that i'd tell you i got what i needed
17:37 ghanima UtahDave: What is actually firing the events? I am having hard time connecting what in salt is responsible for sending the event
17:37 ajolo joined #salt
17:37 jclarke @ghanima: I just needed to add this function to my state "uptodate".
17:38 UtahDave ghanima: Salt itself sends events on the event bus for various things.  You can also send your own custom events using   the event execution m odule
17:38 timoguin ghanima: take a look at tests/eventlisten.py in the source. you can run that on your master and it will dump the event bus
17:39 zz_Cidan hm
17:39 mpanetta eventlisten.py is fun :)
17:39 zz_Cidan does anyone know how I can get raw responses from salt cloud when I get a 400 response code from amazon?
17:39 zz_Cidan ie
17:39 zz_Cidan [ERROR   ] EC2 Response Status Code: 400 Bad Request
17:39 zz_Cidan but I don't know why
17:39 zz_Cidan -lall doesn't tell me
17:40 Sypher zz_Cidan: try running in debug mode (-l debug), or does all already include that too?
17:40 abe_music joined #salt
17:40 Heartsbane Thanks Sypher and nix2170 that was what I needed
17:40 zz_Cidan it does include that
17:43 jimklo joined #salt
17:43 joehoyle joined #salt
17:46 MZAWeb joined #salt
17:46 abe_music i'm attempting to launch/delete VMs via the salt.cloud modules directly instead of using the command line app....i can get things starting, but i'm getting logging issues with the VT module
17:46 abe_music Failed to spawn the VT. Error: 'LoggingProxy' object has no attribute 'fileno'
17:47 abe_music any ideas what i might be doing wrong here? i've messed around with salt.log.setup, but haven't had any luck and the google isn't helping me out either
17:49 ajolo joined #salt
17:50 druonysus joined #salt
17:50 druonysus joined #salt
17:51 UtahDave abe_music: Hm. I haven't tried using salt cloud like that.
17:52 UtahDave If nobody in here knows, send an email to the mailing list and I'll make sure the salt cloud maintainer sees it
17:52 abe_music UtahDave: will do...working on quick test case now
17:53 mpanetta Damn, it is still failing to render update_hosts :(
17:53 catpigger joined #salt
17:53 mpanetta UtahDave, timoguin: do you guys have any idea how to get more info from aslt-master as to why the reactor sls could not render?
17:55 beardo joined #salt
17:56 rojem joined #salt
17:56 UtahDave mpanetta: whiteinge has a doc somewhere explaining that.  I'll have to search for it. whiteinge is on a plane right now, I believe
17:57 gildegoma joined #salt
17:57 mpanetta UtahDave: Ah ok.  Yeah the reactor docs don't tell you much other then use eventlisten.py and debug on the master :(
17:59 UtahDave mpanetta: Yeah, I think this is it: http://docs.saltstack.com/en/latest/topics/reactor/#debugging-the-reactor
17:59 UtahDave I'll check with whiteinge when he lands
17:59 UtahDave back in a bit
17:59 mpanetta UtahDave: That is exactly what I am doing already :(
17:59 mpanetta UtahDave: Ok
18:01 travisfischer joined #salt
18:03 [MT] what files does salt-syndic read when it starts? It's starting to seem like it doesn't read anything I created for it.
18:04 [MT] ah- debug, duhr
18:13 roddergreg joined #salt
18:14 schimmy joined #salt
18:14 roddergreg Hello group. i'm having issues setting the set-uid bit using salt. I'm using the file managed module with '4755' but the setuid 's' is not there. should i be doing this a different way?
18:16 [diecast] joined #salt
18:19 MZAWeb joined #salt
18:21 [vaelen] joined #salt
18:21 swa_work joined #salt
18:24 debian112 joined #salt
18:24 davet joined #salt
18:25 debian112 Question for group: What is the best way to override a managed file from base? Should I use extend or something else?
18:26 danielbachhuber joined #salt
18:28 debian112 or should I copy a template from base and allow each enviroments to override with there settings?
18:29 roddergreg guess i can do it with cmd.run. Thanks!
18:29 meteorfox joined #salt
18:30 redondos joined #salt
18:30 redondos joined #salt
18:32 joehoyle joined #salt
18:34 schimmy joined #salt
18:37 schimmy1 joined #salt
18:37 Whissi Can I tell salt when running "salt '*' cmd.run 'foo'" to only show output of failed clients?
18:38 Philip joined #salt
18:39 timoguin Whissi: not built-in, but there's this: http://russell.ballestrini.net/filter-salt-stack-return-data-output/
18:39 JasonSwindle joined #salt
18:41 Whissi Ah ok, parsing the output... thanks.
18:41 rome joined #salt
18:42 meteorfox joined #salt
18:43 brianabbet634613 joined #salt
18:44 halfss joined #salt
18:44 dstokes joined #salt
18:44 brianabbet634613 i have a bit of a strange issue with a couple providers... rackspace, --list-locations returns an empty set, joyent, --list-images returns an empty set. other providers and other functions for those providers work fine. is this a known issue or am i doing something wrong?
18:45 anuvrat joined #salt
18:48 KennethWilke joined #salt
18:54 Voziv Is there a way I can get the ip address of eth1 on a machine using pillar?
18:55 KennethWilke Voziv: they throw that stuff in grains for you
18:55 Philip hostname
18:55 KennethWilke there's a dict at grains['ip_interfaces'] and you can get list of ip's each interface is configured for
18:57 KennethWilke Voziv: for example you could do "salt \* grains.get 'ip_interfaces:eth1'"
18:58 danielbachhuber joined #salt
18:59 Voziv KennethWilke: Ah nice, would I be correct in using this to get it with jinja templating? {{ grains.get('ip_interfaces:eth1') }} ?
18:59 jalaziz joined #salt
19:01 KennethWilke yeah I believe that would work
19:02 nix2170 Some days ago someone pointed me to https://koji.fedoraproject.org/koji/packageinfo?packageID=13129 to get old salt packages. At the time I didn't notice that the salt-minion packages are not there. Does anybody knows where I can find them? (I specifically need salt-minion-0.17.5-1.el5)
19:03 vkurup joined #salt
19:03 Philip I've got another weird one for you guys today
19:03 Philip in a jinja templated file, `{{ frains.get('sample:foo') }}` returns None, but calling salt-call grains.get 'sample:foo' on that same minion returns the right value
19:04 joehoyle joined #salt
19:05 nix2170 oops actually it is there in a link. Sorry for the noise. And thanks to terminalmage for the rpms :)
19:05 KennethWilke Voziv: it actually didn't work for me in that way, but it does work when i do '{{ grains.get('ip_interfaces')['eth1'][0] }}'
19:06 KennethWilke i think that works the same for what Philip is talking about, inside of jinja templates you'll probably need to do the same
19:07 KennethWilke like {{ grains['sample']['foo'] }} or {{ grains.sample.foo }}
19:08 terminalmage nix2170: so you did find them, then?
19:09 Philip KennethWilke changing the notation to `[]` does work, but I do wonder why colon based grains.get doesnt work in jinja templates
19:09 Voziv KennethWilke: thanks, that did work
19:09 KennethWilke Philip: The : interpretation might be something the salt-cli is handling instead of the execution module
19:10 KennethWilke since there's not a good built in way to handle dictionaries within a shell
19:10 KennethWilke but in jinja there is
19:10 brianabbet634613 anybody here working with rackspace?
19:10 KennethWilke hello!
19:11 Katafalkas joined #salt
19:11 Philip ah, see it was nice to do the colon notation because then you wouldnt get keyerrors on dicts without one of the keys you were digging down into
19:11 snarfy^ joined #salt
19:12 Katafalkas Hey, is there a convenient way of executing states from a salt runner ?
19:12 snarfy^ i love salt but i'm watching my sodium intake as per my doctor's orders
19:12 Katafalkas i asked it in googlegroups too: " https://groups.google.com/forum/#!topic/salt-users/v0SufDG-PUw "
19:12 Philip but then again... you are just passing a string to the grains.get function, whether it's cli or in a jinja template
19:12 KennethWilke yeah, but i think the cli is expanding out what ':' means while the jinja template is taking it as a literal
19:12 Philip so it's confusing that it'd act differently depending on where you're calling it from
19:12 KennethWilke yeah it is
19:14 gq45uaethdj26jw6 joined #salt
19:14 brianabbet634613 .
19:14 KennethWilke Philip: if you do a "salt '*' grains.setval 'test:with:colon' 'value'" it'll make the grain key contain the semicolon
19:15 KennethWilke and then you can't get that value with "salt '*' grains.get 'test:with:colon'"
19:15 KennethWilke but in jinja it would be returned from {{ grains.get('test:with:colon') }}
19:16 Philip KennethWilke: right, so grains.get in the context of jinja acts like a regular python dict.get while grains.get from salt-call or salt does some magic
19:16 KennethWilke yes! indeed
19:17 Philip and furthermore, as you've pointed out, grains.set_val from salt and salt-call dont behave in the way .get behaves on the same interface
19:17 KennethWilke yeah, and that's interesting
19:17 KennethWilke something to be aware of
19:17 Philip I think something to be discussed in an issue ;)
19:17 KennethWilke perhaps, i'd say for all cases but on the CLI it works as I expect
19:18 KennethWilke and for interactions in the CLI I'm not sure what a good recommended alternative would be
19:18 Philip yeah... I feel like if you are going to override the way grains.get works on the cli, it should either carry over or be another method name or at the very list explained in the docs that this is cli-only behavior
19:19 KennethWilke brianabbet634613: did you mean using salt with rackspace cloud or were you looking or a rackspace employee
19:19 Philip and that set should mirror get in it's semantics on the cli
19:21 brianabbet634613 KennethWilke: i'm using salt-cloud to spin up vm's in rackspace, and then manage them with salt. I'm having a very strange issue properly setting the location, they all spin up in dallas and the --list-locations command returns an empty result. Not sure if it's a libcloud issue or if I did something wrong or what
19:21 KennethWilke yeah, but it's handy to be able to get a nested value through the cli, where otherwise you'd probably get a YAML body back that you'd need to dig further into
19:21 brianabbet634613 I got it working with some other providers no problem
19:21 [MT] viq: got it working! now for the load balancer thing
19:22 Ryan_Lane1 [MT]: blog post ;)
19:22 KennethWilke brianabbet634613: are you using the rackspace provider or openstack provider?
19:22 Philip oh I totally agree with the usefulness, but I think it's not intuitive that set doesn't behave by the colon->nesting semantics, but then again: how would you set a key that has a colon in it's name
19:23 Philip for that matter, how do you retrieve a key with a colon in its name using that notation?
19:23 nix2170 terminalmage: yep. I had read very very quickly the page and mistook the src download link for "the" download link. Only on a second look noticed there were other rpms to download. Unfortunately I think the machine I'm trying to install into is borken, 'cause the python2.6 prereqs can't be satisfied. Not a big issue - I just wanted to get a minion in a legacy server. I'll migrate the service to a new host in a couple of weeks.
19:23 sarlalian joined #salt
19:23 terminalmage nix2170: sounds good
19:23 brianabbet634613 KennethWilke: rackspace
19:23 viq [MT]: cool, what did you have to do to make it work?
19:24 gildegoma joined #salt
19:24 KennethWilke i think the rackspace connector will only work for the first-gen rackspace cloud servers, and i don't know if they provision on that setup anymore
19:24 brianabbet634613 KennethWilke: i have a similar issue with joyent, but it is with setting the image. joyent wont return a list of images for some reason, but everything else works
19:24 KennethWilke you
19:24 KennethWilke you'll want to use the openstack connector for rackspace nowadays
19:24 terminalmage nix2170: just FYI, all those "salt" links refer to the name of the RPM spec. We generate 3 packages out of that one spec. So, the package names and the names of the specs in the build system aren't all 1-to-1
19:25 brianabbet634613 KennethWilke: i can get vm's provisioned, just not in the right region. i'll try it with openstack, but i got an exception when i tried but maybe i can run that down
19:25 terminalmage it can be confusing if you've never seen it before, I should have probably described it better
19:25 KennethWilke Philip, yeah thats an odd case where you can set grains with : in the get but not get them, likewise you can get nested values with get but not set them
19:25 terminalmage when I sent you the link last week
19:25 Philip I just tried it: grains.get on a grain with a key that contains colons will fail because of the expansion
19:25 KennethWilke with : in the key rather*
19:25 Philip right
19:25 KennethWilke yeah
19:26 Philip that seems... very strange
19:26 Katafalk_ joined #salt
19:26 KennethWilke there might be a set that will follow expansion and a get that doesn't
19:27 nix2170 terminalmage: I tried once to be the "packager guy" for a software project I work in my uni. OMG. I never ever again will try that.
19:27 KennethWilke well no, i don't see one lol
19:27 Philip but even then, I think keying off ":" for expansion doesn't make sense since ":" is perfectly valid in a key name
19:27 KennethWilke I think it would be nice if there were both versions, both are handy
19:27 Philip fortunately, but unfortunately in this case, _any_ string is valid in a key name
19:28 [MT] Ryan_Lane1: not yet... :(
19:28 Philip perhaps the : should be escaped for it to work, and the cli parser would then take care of expansion if it is
19:28 [MT] viq: type things correctly...
19:28 otsarev_home joined #salt
19:28 Ryan_Lane1 [MT]: heh. you just told me to remind you ;)
19:28 Philip or different semantics altogether should be employed (preferably a different method name to keep this from happening really)
19:29 terminalmage nix2170: haha
19:29 otsarev_home joined #salt
19:29 KennethWilke brianabbet634613: i think the first gen servers can only be spun up in ORD
19:29 terminalmage nix2170: yeah it's not the rock star lifestyle it appears
19:29 KennethWilke but the openstack servers you can spin up in 5 DCs
19:30 brianabbet634613 KennethWilke: oh really? well that would certainly explain it....
19:30 KennethWilke brianabbet634613, yeah, you want to avoid those boxes anywho, the new stuff is much more performant
19:30 KennethWilke and i think a little cheaper too
19:31 ghanima Ok this is going to be a little winded question so please bear with me
19:31 * KennethWilke inhales
19:31 viq [MT]: lol, yeah, that often helps ;)
19:31 ghanima So the Reactor System vs the Overstate System
19:31 brianabbet634613 KennethWilke: alright, i mostly was using it because openstack was throwing an exception that i was having a hard time nailing down. i'll see if i can fix my config
19:32 KennethWilke brianabbet634613, I don't have my salt cloud setup for rax right now, but maybe this has something helpful: http://salt-cloud.readthedocs.org/en/latest/topics/rackspace.html
19:33 diegows I havea  weird issue, may be someone sees it.... if I setup a repo in a state fiel and then I install a package from the repo doesn't work
19:33 diegows I have to call highstate twice
19:33 brianabbet634613 KennethWilke: thanks, thats what i'm going off of as we speak
19:33 KennethWilke i know sometimes i find another part of the salt cloud docs that have the old stuff, but this is representative of the new stuff
19:33 diegows looks like a yum issue (centos 6)
19:33 ndrei joined #salt
19:33 KennethWilke brianabbet634613, bah! i was hoping it would be more helpful then :(
19:33 KennethWilke but i know i've had to finick with it before
19:36 KennethWilke brianabbet634613, I'm gonna spin up an ubuntu box and try to get salt-cloud working with it again
19:36 brianabbet634613 KennethWilke: alright, openstack works if i define my compute_region in the provider (some let me do it in the profile). it works now! thank you for your help!
19:36 KennethWilke ah nice!
19:36 KennethWilke you're welcome :)
19:37 brianabbet634613 KennethWilke: two part issue, i just hadn't combined the two in troubleshooting
19:37 joehoyle joined #salt
19:38 KennethWilke brianabbet634613, i feel fairly sure i've done the same lol
19:38 ajw0100 joined #salt
19:38 oleg joined #salt
19:39 brianabbet634613 KennethWilke: one more quick question, when making a provider that extends another provider, is there a way to identify it as extendable only? i dont want it to actually run queries against a provider that i have partly configured... throws errors
19:40 KennethWilke hmm, i dunno on that. nothing that I am aware of
19:40 brianabbet634613 cool, thanks again
19:40 KennethWilke you could manage your salt.cloud provider with salt and jinja template it! but that's my crazy idea
19:40 ghanima Overstate system looks like it has the ability to introduce an hierarchy and dependency state... The reactor is looking at any even that is coming across the event bus.... Let's say that I have 100 hosts with multiple processes per hosts... I was thinking that I should
19:41 [diecast] joined #salt
19:41 ghanima setup a reactor to monitor for processes and respond if one were down.. but I do have some app dependencies and I trying to decide is it better to do this all within overstate
19:41 ghanima or reactor
19:41 ghanima Curious anyone's thoughts if you have them
19:41 brianabbet634613 hm, that might do it... i'm configuring a ton of providers and they have a lot of overlapping fields. trying to build a dependency tree, but it tries to run queries against every item i define
19:41 KennethWilke diegows, did you install that package through a state as well or though the CLI?
19:42 KennethWilke ghanima, that's a tricky one, I don't have enough experience with that stuff to lend guidance
19:42 mpanetta ghanima: I'm not sure reactor monitors the things you think it does.
19:43 mpanetta I have been playng with it today.
19:43 diegows KennethWilke, state
19:43 ghanima KennethWilke: no worries appreciate the comment though
19:43 KennethWilke mpanetta, salt mine maybe?
19:43 diegows if I run highstate twice, works :P
19:43 mpanetta ghanima: Go get eventlisten.py and watch what happens.  It is very informative.
19:43 KennethWilke diegows, do you have a require statement that says the pkg requires the pkgrepo?
19:43 diegows yes
19:44 mpanetta KennethWilke: Yeah I am having issues getting mine tow ork in my reactor state :(
19:44 ghanima mpanetta: I was under the impressesion that salt can monitor for any event injected onto the event bus. That event state can have a sls rule applied to it to take action... Is that not accurate?
19:44 mpanetta ghanima: That is accurate
19:44 diegows http://paste.ubuntu.com/7332175/
19:44 KennethWilke ghanima, that sounds about right, but i think it's just salt events
19:44 mpanetta ghanima: But to use it to monitor processes you would have to write your own event generator.
19:44 KennethWilke sounds like yours might be an event of your own definition
19:45 mpanetta yeah exactly
19:45 KennethWilke with salt mine you're supposed to be able to run arbitrary salt functions at a specified interval, but it's another aspect i've not tinkered with very  much
19:46 ghanima So my main question is dependency structure if I have app a, app b and app c that are being monitored... and I need app c to check if app a is before trying to start... Is it easier to just do that within Overstate but checking the return code status of a command or doing that as an event....
19:46 ghanima From what I can tell reactor doesn't have a way to tie events to current state
19:46 mpanetta I don't believe it does
19:46 mpanetta In the bits of playing I have been doing today
19:47 KennethWilke diegows, hmmmm maybe the 'fromrepo' parameter from pkg.installed could help there
19:47 diegows I think that I've tried that
19:48 diegows but I'll try again
19:48 [MT] Ryan_Lane1: I really really don't wanna have to write some trippy system to pin specific minions to specific masters...
19:48 KennethWilke yeah i'm not too sure beyond that, your state file looks solid
19:49 KennethWilke i'd expect it to work in 1 run as well :\
19:49 Ryan_Lane1 [MT]: yeah, it would be ideal if 0mq had cluster support
19:49 ghanima Another question... I have been deploying custom grains to multiple nodes.... and am curious if there is a performance gain from defining this as pillar var than a grain
19:49 ghanima I have a deployment of 600 servers and the checking time is so quick I never saw a real point
19:50 mpanetta ghanima: Using grains vs pillar is more of a usage thing, not performance really.
19:50 [MT] Ryan_Lane1: I can't even picture what that would look like... I'd probably hash the minion id and based on the range the hash falls into, pick a syndic and the next one above it. I really really don't wanna do that...
19:51 Guest49182 joined #salt
19:51 [MT] I have our load balancer balancing requests to a specific master and using a hash of the source IP to keep that minion talking to that syndic and unless the syndic goes down, the minion will stay connected to it.
19:55 [MT] When the minions are connected to the syndics via the LB, I see this in the minion logs - http://dpaste.com/1797304/
19:55 diegows KennethWilke, fromrepo is the name of the .repo file with without the suffix, right?
19:56 [MT] UtahDave: any chance you could make sense of this? Why would the fact that this is being sent through a load balancer upset this section of code?
19:57 Eugene Has there been any discussion/an Issue opened for the possibility of recognizing 'all' as a magic hotkey in place of '*', for the purposes of `salt` ?
19:57 KennethWilke diegows, i think in debian-family context it's stuff like 'dist' or '10gen' from your line 15
19:57 Eugene Or even -A, etc
19:57 diegows in centos
19:57 diegows in debian/ubutnu works fine this
19:57 KennethWilke oooooho derp, sorry lol
19:57 * diegows hates yum
19:58 * [MT] agrees
19:58 [MT] Eugene: what if someone named a minion all?
19:58 KennethWilke i'm crazy enough to be a portage guy myself :x
19:58 KyleG joined #salt
19:58 KyleG joined #salt
19:59 KennethWilke i think in yum context it'd be 'mongodb'
19:59 Eugene [MT] - hence, a discussion. :-p. IMO it would be acceptable to break that one edge case, but I ain't in charge of the project.
19:59 KennethWilke the humanname for your repo
19:59 diegows KennethWilke, ok, thanks!
19:59 KennethWilke diegows, don't thank me yet! who knows if i'm right :)
19:59 KennethWilke but you're welcome either way :p
19:59 diegows right! :)
19:59 Eugene AFAIK "-A" and "--all" would be nonconflicting, and match the existing options -R -L etc
20:00 MBroadhead joined #salt
20:00 kermit joined #salt
20:03 [diecast] joined #salt
20:03 jalaziz joined #salt
20:04 yusuket joined #salt
20:06 jcockhren Do modules, runners and returners occupy the same namespace?
20:06 jcockhren For example, can you have a returner and a runner that have a __virtualname__ = 'foo'?
20:07 jcockhren s/__virtualname__/__virtual__/
20:07 herzi_ Rory: Thanks for the hint about force_yes, now I'm getting this (trying to install systemd-sysv and uninstalling sysvinit w/ wheezy-backports):
20:07 herzi_ You are about to do something potentially harmful. To continue type in the phrase 'Yes, do as I say!' ?] Abort.
20:07 jcockhren basepi: ^^
20:07 kermit joined #salt
20:08 Eugene https://github.com/saltstack/salt/issues/12301
20:09 ajolo Hey guys, little typo in: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html it says - file_mode: 744; 644 would be more suitable
20:09 jcockhren ajolo: PR that thing bro. :)
20:09 KennethWilke don't take away my exec bro
20:09 KennethWilke :)
20:10 jcockhren ha
20:10 KennethWilke its my file! i'll run it if i want
20:10 pydanny joined #salt
20:12 tr_h joined #salt
20:12 ajolo lol
20:13 timoguin ajolo: line 174 of salt/states.file.py. :)
20:15 Ansem left #salt
20:16 basepi jcockhren: as far as i know, different namespaces.
20:17 jcockhren basepi: hm. ok. I know that there's a cmd state but the module had to be called 'cmdmod'
20:18 jcockhren I'll try it out and see what happens or doesn't regarding the other salt elements
20:18 basepi jcockhren: right, but we have both a file state and a file module
20:18 basepi the cmdmod thing was for some other reason, trying to remember.....
20:20 Ansem joined #salt
20:20 Ansem How to tell a service to restart?
20:21 Ahlee hrm, -creates doesn't seem to be stopping this cmd.run from running every time
20:21 Ansem Service get's installed, then my config is put in, and then I need Salt to restart it to pickup the new config
20:21 rome joined #salt
20:21 jcockhren Ansem: there's a service state that can do that
20:21 jcockhren in fact, you can tell it to watch a file for changes
20:22 Ansem Hm.. any links to an example?
20:22 Ansem The docs are a bit vague beyond getting a service running
20:26 jcockhren Ansem, for example: https://gist.github.com/jcockhren/f6bde4a6fef332001703
20:27 Ahlee oh well, unless works
20:33 herzi jcockhren: doesn't watch include require (so the require isn't needed)?
20:33 ajw0100 joined #salt
20:36 jcockhren herzi: :shrug: That's pulled from a state file from work.
20:36 joehillen joined #salt
20:37 kermit joined #salt
20:37 pdayton joined #salt
20:37 jcockhren There require is used for testing presence
20:37 jcockhren the*
20:38 jcockhren the watch is for changes in the package itself. For example, internal patched versions
20:43 Ansem jcockhren Got it, thanks
20:43 Ansem I had to re-deploy the VM I was using for it to work (of course), but it's exactly what I needed
20:43 Ansem (Vagrant + Salt)
20:43 diegows github has changed something that breaks curl when you want to download salt-bootstrap. :P
20:45 manfred using curl -L ?
20:45 manfred how exactly is it breaking?
20:46 diegows es
20:46 diegows yes
20:46 arthabaska Can someone help me with a package installation problem? I'm installing a custom yumrepo using pkgrepo and then trying to use it just after, ending with the error " Package 'mongodb-org' not found (possible matches: mongodb-org)" log and state: https://gist.github.com/donjohnson/208411bb01c62803f559
20:46 diegows the webserver return 400
20:46 diegows only happens in CentOS 6, the ubuntu versio works fine
20:46 diegows weird
20:47 arthabaska ** I should note that a second run of state.highstate succeeds in installing the package
20:48 joehoyle joined #salt
20:48 diegows arthabaska, same problem here :)
20:48 manfred diegows: did you update openssl in centos 6 to make sure it wasn't a weird certificate thing?
20:48 alunduil joined #salt
20:48 diegows trying a solution suggested right now
20:49 diegows :)
20:49 diegows hmm
20:49 diegows good point
20:49 manfred or ca-certificates, whatever that thingy is
20:49 diegows but it doesn't looks like a ssl error
20:50 arthabaska diegows haha I see that! I'm on the amazon linux ami btw
20:51 diegows manfred, http://paste.ubuntu.com/7332655/
20:53 manfred weird
20:53 manfred i got nothing
20:53 diegows where?
20:53 diegows centos?
20:53 diegows I works for me on Ubuntu 14.04
20:54 manfred no idea, it was working for me ealier today on centos 6.5
20:54 diegows yes, I was working for me too
20:54 diegows :)
20:54 timoguin arthabaska: you can call pkg.refresh_db using module.run after the state that adds the repo. like this: https://github.com/timoguin/flexnet-salt-states/blob/master/redis/init.sls
20:54 diegows arthabaska, the suggested solution didn't fix the issue :P
20:55 timoguin that will work the first state run, but it's a bit annoying because it will run every time
20:55 timoguin and highstate already does a pkg.refresh_db
20:55 diegows is there a way to refresh, purge caches or something in yum?
20:56 chuffpdx_ joined #salt
20:57 diegows timoguin, I didn't see that...
20:57 diegows trying
20:57 arthabaska yum clean ; yum makecache might be right diegows
20:57 arthabaska timoguin trying that
21:01 AdamSewell joined #salt
21:01 AdamSewell joined #salt
21:04 Eugene `yum clean all`
21:04 arthabaska timoguin: refresh_db ran, but still no dice: https://gist.github.com/donjohnson/dc8d045238998005afc1
21:04 jaimed joined #salt
21:05 rome joined #salt
21:05 timoguin arthabaska: and it works on the second run?
21:05 arthabaska this is on a completely fresh launch of an ami
21:05 arthabaska yes, 2nd run works
21:06 timoguin this line is funny to me: Package 'mongodb-org' not found (possible matches: mongodb-org)
21:06 timoguin yes... that's the one....
21:06 Katafalkas joined #salt
21:07 arthabaska if it matters, that's a "meta package" pointing to four mongo packages from the vendor
21:10 diegows timoguin, yes, really funny :)
21:13 arthabaska here's a new gist with debug logs--repoquery is definitely running correctly and returning correctly. its exit value is 0 as well (from testing): https://gist.github.com/donjohnson/6cb6ec5d902ebfcc6d63
21:14 arthabaska I'm going to dive into the code and see if I can find this
21:15 Steve_ joined #salt
21:15 Steve_ What is the general release schedule for salt? I need function cp.push_dir which is in master branch but not yet released?
21:16 diegows arthabaska, it's weird, I have a similar recipes to add a repo to install  php 5.4 and works
21:16 diegows but there is something with mongo repo and yum
21:17 gq45uaethdj26jw6 left #salt
21:17 diegows now, trying a workaround
21:17 diegows yum install -y mongodb-org-server mongodb-org-shell mongodb-org-tools:
21:17 diegows cmd.run
21:17 Steve_ joined #salt
21:17 diegows :(
21:18 lukeg1 joined #salt
21:18 arthabaska yeah that's dirty...does it work? :)
21:18 thedodd joined #salt
21:19 jcockhren https://github.com/saltstack/salt/issues/11147
21:19 jcockhren :)
21:20 Steve_ Anybody know when the next release of salt is due? I need function cp.push_dir which is in master branch
21:21 herzi Steve_: I usually copy the fixed file to the machines and let the next upgrade overwrite it
21:22 joehoyle_ joined #salt
21:22 diegows arthabaska, I'll let you know :)
21:23 timoguin Steve_: they're hoping to have the first RC for the current develop branch out in the next few weeks
21:24 Steve_ ok, thanks ... if you copy a single file to salt master is that automatically distributed to minions?
21:24 Steve_ Want to use cp.push_dir to distribute virtualenvs across minions
21:25 Corey Hey guys.
21:25 Corey Gareth is stuck in an elevator and livetweeting it.
21:25 Corey @garethgreenaway
21:25 UtahDave Steve_: drop the cp.py you want in  /srv/salt/_modules
21:25 UtahDave and then run   salt \* saltutil.sync_all   or salt \* saltutil.sync_modules
21:26 rogst joined #salt
21:26 timoguin Corey: very exciting to watch the drama unfold!
21:26 UtahDave lol
21:26 Steve_ ok, thanks Dave. I have developed some module/states and distributed them that way...I guess if names overlap it takes the one from _modules?
21:27 Corey I'm mocking CNN as we go.
21:28 UtahDave Steve_: yep, that is correct
21:28 mgw joined #salt
21:29 wincus left #salt
21:30 nlb joined #salt
21:30 DaveQB joined #salt
21:31 Steve_ Thanks
21:32 diegows arthabaska, the dirty thing works!
21:32 diegows :)
21:33 timoguin pkg and pkgrepo y u no werk?
21:34 mgw1 joined #salt
21:34 diegows no
21:34 diegows pkgrepo and cmd.run 'yum install' works !
21:34 diegows !?
21:34 jaimed joined #salt
21:36 arthabaska the code in pkg.py is accidentally in _find_install_targets....just not sure where yet. will know soon.
21:39 lukeg1 hi all. i just upgraded to 2014.1.3 and when I run salt '*' salt.highstate I get "Failed to authenticate, is this user permitted to execute commands?"
21:39 arthabaska lukeg1 running that as root?
21:40 lukeg1 yes logged in as reg_user then su - to root the run salt command
21:40 Steve_ Did not like it when I tried to sync cp.py
21:40 Steve_ Passed invalid arguments: cache_dir() takes at most 5 arguments (6 given)
21:41 jeff joined #salt
21:41 jeff joined #salt
21:41 mgw joined #salt
21:42 lukeg1 arthabaska:: I also stopped and started master and minion services
21:42 Steve_ i'll try that, thanks
21:43 Steve_ Same problem
21:46 Steve_ I guess where are some dependencies in the general code base on cp.py
21:48 ndrei joined #salt
21:49 arthabaska lukeg1 sorry, thought that might be a simple non-root issue. might be something with external auth?
21:51 sverrest joined #salt
21:55 Steve_ Yikes... my install seems to be hosed now
21:57 lukeg1 arthabaska: is that something that needs configured? I didn't touch that before - when everything was working.
21:57 Steve_ Anyone have any ideas on how to remove a module which seems to be cached somewhere and it causing problems?
21:57 Steve_ [INFO    ] Loading cache from salt://_modules, for base) Passed invalid arguments: cache_dir() takes at most 5 arguments (6 given)
21:58 Steve_ keep getting this
21:58 arthabaska lukeg1 I'm not sure, as I haven't worked with that area either. still pretty new to salt myself
22:00 kermit joined #salt
22:01 it_dude joined #salt
22:01 A||SySt3msG0 joined #salt
22:02 lukeg1 arthabaska: ok -thanks. i'll keep troubleshooting and what else I can break along the way. :)
22:05 Katafalk_ joined #salt
22:05 Steve_ Anyone know how to clear out a broken module from cache?
22:05 Katafalk_ joined #salt
22:06 Katafalk_ Hey, anyone could help me with this question ? "https://groups.google.com/forum/#!topic/salt-users/v0SufDG-PUw"
22:08 ajolo Hey guys, is there anyway of passing using and password to the git state ? I need to clone a private repo from GitHub
22:08 ajolo s/using/username/
22:09 Gareth well that was interesting..
22:15 rome joined #salt
22:17 UtahDave Steve_: try making sure all *.pyc files are deleted
22:18 lukeg1 left #salt
22:19 UtahDave Katafalk_: try using cmd.async, that will return the jid and you can check the job status
22:19 UtahDave Gareth: you survived!
22:19 Steve_ Just deleted my install and rebuilding
22:19 Katafalk_ UtahDave: ok. i ll give it a go ! danke
22:19 rome joined #salt
22:20 Steve_ I'll keep that in mind for next time
22:20 halfss joined #salt
22:21 Gareth UtahDave: I did!  It was tense in there.
22:21 UtahDave he he
22:22 pdayton joined #salt
22:23 ajolo hey UtahDave :)
22:25 UtahDave hey, ajolo!
22:27 Katafalk_ UtahDave: how do I check if the job was complete ? and if so the result of it ?
22:29 avienu joined #salt
22:31 cruatta joined #salt
22:32 UtahDave salt \* saltutil.lookup_jid <job id>
22:32 mgarfias joined #salt
22:33 rome joined #salt
22:37 Katafalk_ it prints out that every minion does not have that JID.
22:37 Katafalk_ how do I find the result of it ?
22:38 avienu joined #salt
22:39 Katafalk_ nvm
22:39 ahammond are there current known issues with salt / vagrant? I'm seeing this: https://gist.github.com/ahammond/11305504   with a Vagrantfile that used to work. I've upgraded vagrant to latest. I'm guessing it's salt-bootstrap again?
22:39 Katafalk_ runner - jobs.lookup_jid
22:39 Katafalk_ git me result
22:39 ajolo UtahDave: how are you doing ? I'm returning to BA tomorrow :/
22:44 guest814 joined #salt
22:44 ajprog_laptop joined #salt
22:45 UtahDave joined #salt
22:45 guest814 hey all quick question or maybe not....
22:45 guest814 i looking at away to aggregate the out put of multiple minions
22:47 guest814 example salt '*' cmd.retcode 'some command'... something that would look at the output and aggregate the result of retcode and display the host that are different
22:48 allanparsons joined #salt
22:48 allanparsons do you guys know if salt installs rvm?
22:48 allanparsons i use "rvm.gemset_present" and get an error: Comment: RVM is not installed.
22:52 JasonSwindle joined #salt
22:54 manfred allanparsons: it doesn't have rvm in the $PATH for cmd.has_exec
22:54 manfred allanparsons: https://github.com/saltstack/salt/blob/develop/salt/states/rvm.py#L112
22:54 manfred https://github.com/saltstack/salt/blob/develop/salt/states/rvm.py#L112
22:54 manfred bah
22:54 manfred this is the second link https://github.com/saltstack/salt/blob/develop/salt/modules/rvm.py#L53
22:55 allanparsons manfred
22:55 allanparsons but, i'm running rvm as a user named "rvm"
22:55 allanparsons if i su - rvm
22:55 allanparsons and do which rvm...
22:56 allanparsons $ rvm -v
22:56 allanparsons rvm 1.25.24 (stable) by Wayne E. Seguin <wayneeseguin@gmail.com>, Michal Papis <mpapis@gmail.com> [https://rvm.io/]
22:56 manfred but does salt's cmd.has_exec see it
22:56 allanparsons which rvm == /usr/local/rvm/bin/rvm
22:57 manfred what is the result of cmd.has_exec rvm ?
22:57 gothix_ Anyone an admin sharing access with developers to the salt master?
22:57 happytux joined #salt
22:57 allanparsons salt-call cmd.has_exec rvm local:     True
22:57 allanparsons that''s as root
22:59 manfred allanparsons: are you using runas: ?
22:59 allanparsons i am
22:59 manfred https://github.com/saltstack/salt/blob/develop/salt/modules/rvm.py#L25
22:59 allanparsons runas is deprecated :)  I'm using user:
22:59 manfred so
23:00 manfred if you choose user:
23:00 manfred it only checks if rvm exists in the users home
23:00 manfred https://github.com/saltstack/salt/blob/develop/salt/modules/rvm.py#L25
23:00 allanparsons yeah :( ls: cannot access /home/rvm/.rvm/bin/rvm: No such file or directory
23:00 manfred so
23:00 manfred you will want to put it there for now, and I will look at fixing this later tonight
23:01 allanparsons well, so here's the deal... i brought a new box up and attempted to install ruby via rvm.installed
23:01 allanparsons it bitched that it couldnt find rvm, so i just did a cmd.run on the curl -sSL https://get.rvm.io | bash -s stable
23:02 allanparsons but, it became a huge mess because i dont want to give rvm user sudo access for the install
23:02 allanparsons and, so i had to install as the root user which installs it systemwide
23:03 allanparsons (not ideal)
23:03 manfred well, if you install it system wide, it doesn't let you do user: at least how it is written now
23:04 allanparsons so how would i get around that?
23:04 manfred you should install it as that user in ~/.rvm
23:04 manfred and then do user:
23:04 manfred or get rid of user:
23:04 manfred until I can get a minute to fix it :)
23:04 allanparsons i want it in ~/.rvm
23:04 manfred su - rvm
23:04 Katafalkas joined #salt
23:05 manfred then run the rvm bootstrap to install it there
23:05 manfred manuall
23:05 manfred manually
23:05 manfred if it is broken in salt right now
23:05 manfred I will look at it in a minute if it is broken
23:05 allanparsons okay
23:05 allanparsons 1 sec
23:05 allanparsons lemme su - to it
23:06 allanparsons manfred: http://pastebin.com/0aE8f6sA
23:09 googolhash joined #salt
23:09 ghanima_ joined #salt
23:10 manfred looks like the curl command for rvm.install needs to have a -L
23:10 mgw joined #salt
23:10 logix812 joined #salt
23:10 arthabaska diegows, timoguin : found the bug, and just as I was going to fix it, it seems that terminalmage fixed it 2 months ago haha: https://github.com/saltstack/salt/commit/2461e94d0da555a7d74ad4882fef14a66fe08e8a (thanks terminalmage !)
23:11 allanparsons http://rvm.io/rvm/install
23:11 allanparsons ah... in Salt?  not my .sls?
23:11 terminalmage arthabaska: no prob!
23:11 logix812 master_tops: so what does that conf option expect right now.. I see examples for cobbler and reclass and they are different. I want to mess around with this, do I just pass it something that is executable on the cli?
23:11 manfred allanparsons: yeah, in salt
23:11 arthabaska turns out an internal cache of available yum packages (an expensive call) is stale after adding a pkgrepo, and is believed to be true
23:11 allanparsons woohoo.  found a bug!
23:11 manfred allanparsons: https://github.com/saltstack/salt/blob/develop/salt/modules/rvm.py#L82
23:11 manfred i will push something in a minute
23:12 allanparsons thanks
23:13 manfred https://github.com/saltstack/salt/pull/12313
23:15 ghanima_ joined #salt
23:15 logix812 hmmm looks like I need a top() function.. and some other drama (__virtual__, etc) and I stick it in /srv/salt/states as a python file
23:16 allanparsons manfred.. +1
23:16 doanerock joined #salt
23:20 mpoole Hey, so if I fire an event to master from the command line it works fine. If I fire an event from python nothing ever hits the master bus
23:20 mpoole https://gist.github.com/reedox/b81b3d054359b5992dfd
23:20 mpoole any ideas?
23:21 mpoole salt-call event.fire_master  works fine
23:23 mpoole using 2014.1
23:24 ghanima_ is it true you can install a salt-minion with just python pip
23:24 ghanima_ mpoole: I think you are going to have to enter the fqdn of the master no?
23:27 mpoole ghanima_ doesn't work
23:28 manfred ghanima_: no, if you use pip, you get all of salt, salt-minion is only split out by distribution packagers
23:29 tr_h joined #salt
23:29 ghanima_ manfred: gotcha :) thanks
23:36 allanparsons argh rvm is so annoying to use via supervisor
23:42 allanparsons if supervisor starts as root, how do i use rvm as a diff user?
23:43 manfred allanparsons: http://ix.io/bVW that works for me now
23:43 allanparsons @manfred... thanks
23:43 allanparsons do you use supervisor with rvm?
23:44 manfred i do not
23:44 allanparsons dag
23:44 manfred i haven't done any of that stuff, was just testing that the rvm.installed and that stuff works now
23:44 allanparsons okay
23:44 allanparsons thanks :)
23:44 manfred np
23:44 anuvrat joined #salt
23:44 UtahDave ajolo: I hope you had a great time while you were here!
23:47 meteorfo_ joined #salt
23:48 mrj joined #salt
23:49 arthabaska does anyone know of a good rpm of 2014.1.3 for el6/Amazon yet?
23:50 manfred 2014.1.3 should be in epel-testing
23:52 manfred salt-2014.1.3-1.el6.noarch.rpm                                                19-Apr-2014 05:38  2.8M  RPM Package
23:52 manfred salt-api-0.8.3-1.el6.noarch.rpm                                               05-Dec-2013 21:45   53K  RPM Package
23:52 manfred salt-master-2014.1.3-1.el6.noarch.rpm                                         19-Apr-2014 05:38   36K  RPM Package
23:52 manfred salt-minion-2014.1.3-1.el6.noarch.rpm                                         19-Apr-2014 05:38   21K  RPM Package
23:52 manfred https://dl.fedoraproject.org/pub/epel/testing/6/x86_64/
23:52 arthabaska manfred many thanks! trying it now
23:53 TyrfingMjolnir joined #salt
23:54 ghanima_ manfred: awesome news
23:54 jchen hm, looks like https://github.com/saltstack/salt/commit/0080e01a35cba87c4fdaea16dab2256fad94bffe didn't make it into 2014.1.3, will it be in the next release? how do things get added to releases?
23:55 manfred jchen: have to make a request against the 2014.1 branch with that commit cherry picked
23:55 joehoyle joined #salt
23:56 jchen manfred: kkty

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