Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-03-31

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

All times shown according to UTC.

Time Nick Message
00:01 gladiatr joined #salt
00:01 ipalreadytaken joined #salt
00:02 joehillen joined #salt
00:03 jeremyfelt joined #salt
00:06 ajw0100 joined #salt
00:20 bemehow joined #salt
00:27 gladiatr left #salt
00:32 n8n joined #salt
00:33 napper joined #salt
00:33 younqcass joined #salt
01:24 ilbot3 joined #salt
01:24 Topic for #salt is now Welcome to #salt | 2014.1.1 is the latest | Please be patient when asking questions as we are volunteers and may not have immediate answers | Channel logs are available at http://irclog.perlgeek.de/salt/
01:25 rushmore joined #salt
01:26 garthk Reactor, maybe.
01:31 shenril1 joined #salt
01:34 jmreicha joined #salt
01:37 cachedout_home joined #salt
01:38 bemehow joined #salt
01:38 jeremyfelt joined #salt
01:39 oz_akan_ joined #salt
01:44 xunil joined #salt
01:49 andrej Ok ... so Thursday & Friday last week I was barking up a few wrong trees (I didn't know, and am still not certain, which are the right trees :}) ... but on the WE I read up on the reactor system, which looks promising for my wish to to something to MinionB when a state executes on MinionA.
01:51 garthk andrej: yeah, reckon I might give that a shot for this logging keys thing
01:52 andrej My problemm now is that can't see an easy way to match the response to the state execution on the minion on the master ...
01:53 andrej garthk _ just did a scroll-back; looks like you have a similar task at hand as me :)
01:54 andrej I need to poke a hole in the host-based firewall of a minion when a package pertaining to port 9997 gets deployed on other minions
02:00 andrej ok, I see a pattern in the events - now I need to whip up some matching-magic and a suitable response-sls :)
02:00 andrej wish me luck
02:02 ipalreadytaken joined #salt
02:02 jmreicha joined #salt
02:11 toastedpenguin joined #salt
02:13 garthk Nuts.
02:14 garthk The way to watch a directory for changed files is to use file.recurse, which means entirely taking over how you fill it in.
02:14 garthk Bah.
02:14 garthk Aah, clean: false might help
02:14 andrej :)
02:29 xl1 joined #salt
02:29 favadi joined #salt
02:35 garthk So good to see a TCP port on loopback forwarding via a pair of stunnels to the same TCP port on loopback on another server.
02:35 codekoala joined #salt
02:36 n8n joined #salt
02:36 garthk Finding it hard to resist urge to add WARNING: SALT ADDED to every config file
02:38 garthk Huyh
02:38 garthk Why resist?
02:38 garthk # WARNING: SALT ADDED (master: {{ grains['master'] }})
02:46 Ryan_Lane joined #salt
03:00 Ryan_Lane joined #salt
03:02 ajw0100 joined #salt
03:06 vejdmn joined #salt
03:06 fllr joined #salt
03:10 pydanny joined #salt
03:20 Gifflen joined #salt
03:35 googolhash_u joined #salt
03:35 redondos joined #salt
03:37 kermit joined #salt
03:43 ipalreadytaken joined #salt
03:44 ipalreadytaken joined #salt
03:47 Ryan_Lane joined #salt
03:52 fllr joined #salt
03:56 TyrfingMjolnir joined #salt
03:58 yomilk joined #salt
04:01 yomilk joined #salt
04:08 TyrfingMjolnir joined #salt
04:15 fllr joined #salt
04:26 MZAWeb joined #salt
04:28 schimmy joined #salt
04:31 pydanny joined #salt
04:31 ravibhure joined #salt
04:33 googolhash_d joined #salt
04:35 mekstrem joined #salt
04:38 fllr joined #salt
04:46 toastedpenguin joined #salt
04:46 shenril1 joined #salt
04:48 jmreicha joined #salt
04:48 torrancew joined #salt
04:51 pydanny joined #salt
04:53 tmmt joined #salt
04:54 hazzadous joined #salt
04:57 smcquay joined #salt
05:00 tmmt joined #salt
05:04 xl1 joined #salt
05:08 tmmt joined #salt
05:10 dh joined #salt
05:10 fllr joined #salt
05:12 fllr joined #salt
05:24 slav0nic joined #salt
05:24 slav0nic joined #salt
05:32 NV joined #salt
05:33 dave_den joined #salt
05:35 arthabaska joined #salt
05:49 Nexpro1 joined #salt
05:49 gildegoma joined #salt
05:50 googolhash joined #salt
05:55 j4son joined #salt
06:02 schimmy joined #salt
06:06 zain_ joined #salt
06:10 fllr joined #salt
06:16 ipalreadytaken joined #salt
06:17 fragamus joined #salt
06:17 Hipikat left #salt
06:28 gammalget joined #salt
06:38 dccc joined #salt
06:38 srage joined #salt
06:45 ChaosPsyke joined #salt
06:57 Kenzor joined #salt
07:01 balboah joined #salt
07:05 srage joined #salt
07:08 harobed_ joined #salt
07:10 fllr joined #salt
07:10 mekstrem joined #salt
07:20 ChaosPsyke joined #salt
07:37 eligro91 joined #salt
07:37 thehaven joined #salt
07:38 balboah joined #salt
07:39 eligro91 left #salt
07:42 ipalreadytaken joined #salt
07:42 blitz`` joined #salt
07:52 swissknife joined #salt
07:57 Ryan_Lane joined #salt
07:57 swissknife left #salt
07:57 johtso joined #salt
07:59 johtso joined #salt
08:00 vortec joined #salt
08:01 stephanbuys joined #salt
08:01 shoma joined #salt
08:07 topochan joined #salt
08:10 fllr joined #salt
08:11 yomilk joined #salt
08:21 ggoZ joined #salt
08:21 yomilk_ joined #salt
08:24 dRiN joined #salt
08:29 mekstrem_ joined #salt
08:31 obimod joined #salt
08:31 obimod hey hey
08:40 ipalreadytaken joined #salt
08:40 Daviey joined #salt
08:49 zach_ joined #salt
08:49 Shish joined #salt
08:50 jeremyBass1 joined #salt
08:51 20WAAB3JM joined #salt
08:53 nebuchadnezzar joined #salt
08:56 giantlock joined #salt
08:57 ivan joined #salt
08:57 icarus joined #salt
08:58 faldridge joined #salt
09:01 djinni` joined #salt
09:02 nocturn joined #salt
09:03 johtso joined #salt
09:04 johtso joined #salt
09:04 johtso joined #salt
09:05 Username_ joined #salt
09:07 johtso joined #salt
09:07 topochan joined #salt
09:09 mekstrem joined #salt
09:10 ninkotech joined #salt
09:10 * obimod yawns
09:10 mekstrem joined #salt
09:10 fllr joined #salt
09:15 helderco joined #salt
09:19 twobitsprite joined #salt
09:24 MrTango joined #salt
09:28 hazzadous joined #salt
09:37 yomilk joined #salt
09:38 blitz`` hi guys, trying to follow the centos installation guide
09:38 blitz`` installed via epel repo, centos 6.5 x64, i get the following when trying to start the salt-master service
09:38 blitz`` http://pastebin.com/HdQX5UcU
09:40 joehoyle joined #salt
09:45 helderco joined #salt
09:58 wunki left #salt
10:10 fllr joined #salt
10:11 Whissi joined #salt
10:16 gildegoma joined #salt
10:43 zeroC joined #salt
10:44 zeroC hi there... :)
10:44 yomilk joined #salt
10:46 zeroC just a short question, i want to exclude some packages from yum groupinstall base, i tried salt '*' pkg.group_install 'Base' skip='pkg1,pkg2' but it started the installation with pkg1 etc
10:46 zeroC is there a way to exclude packages from a yum group?
11:01 mekstrem joined #salt
11:10 fllr joined #salt
11:11 blitz`` joined #salt
11:17 younqcass joined #salt
11:17 prooty joined #salt
11:18 joehoyle joined #salt
11:19 prooty hello. i have a salt master version 0.17.2-2precise1. i launched a new instance whose salt minion version is 2014.1.0-1precise1. when i try to call state.highstate on it the error "No Top file or external nodes data matches found" is shown. trying state.highstate on my older minions works.
11:20 prooty my older minions have the same version as the master.
11:22 johtso Can anyone see any issues with this state in regards to the supervisord process group being properly restarted when the pip package is updated? http://hastebin.com/xayamawitu.yaml
11:23 johtso I'm finding that after the package is updated, the process is still running the old version. Possibly because the restart is happening before the package is updated?
11:23 DaveQB joined #salt
11:23 ndrei joined #salt
11:24 johtso Is it possible to pass arguments to mod_watch?
11:25 johtso I'm thinking I need to tell mod_watch to update+restart, and remove update=True from the supervisord.running state
11:25 diegows joined #salt
11:26 ekristen joined #salt
11:27 johtso oh, the parameters are taken from the parent state.. I'm stumped then
11:33 oc- salt has become insanely slow since upgrading to 2014.1.1. Any changes I should be aware of (haven't updated any of my states yet)
11:33 oc- any <particular>
11:36 oc- takes about 7 minutes to highstate a server, used to take 20sec
11:36 johtso ah, I've found the issue.. https://github.com/saltstack/salt/issues/6692
11:36 johtso does anyone know if there would be a workaround for the above?
11:45 topochan joined #salt
11:46 prooty i downgraded my new minion to 0.17.2-2precise1 and it worked.
11:46 prooty good thing i have copies of the old deb files.
11:47 gadams999 joined #salt
11:49 toastedpenguin joined #salt
11:54 jslatts joined #salt
11:54 johtso can I call a salt module function from a state somehow?
11:56 Calum joined #salt
12:04 yomilk joined #salt
12:06 Gifflen joined #salt
12:09 dangra joined #salt
12:10 fllr joined #salt
12:24 gildegoma joined #salt
12:28 oz_akan_ joined #salt
12:35 jeremyBass joined #salt
12:36 elfixit joined #salt
12:37 spiette joined #salt
12:37 isomorphic joined #salt
12:40 isomorphic left #salt
12:43 DammitJim joined #salt
12:43 DammitJim hi guys
12:52 wkf joined #salt
12:55 Gordonz joined #salt
13:00 babilen DammitJim: o/
13:01 zeroC Hey DammitJim :)
13:07 jcsp joined #salt
13:08 jeremyBass left #salt
13:09 Gifflen joined #salt
13:10 fllr joined #salt
13:13 R_Shackleford joined #salt
13:13 joehoyle joined #salt
13:16 oz_akan_ joined #salt
13:20 higgs001 joined #salt
13:27 babilen I was wondering if there are established workflows to manage security sensitive data in pillars. We are using git pillars so far, but I am not really happy with the fact that a bunch of developers are running around with a collection of private keys and whatnot on their laptops.
13:29 yomilk joined #salt
13:30 mortis babilen: how about storing the sensitive pillars in a different repo that the devs have no access to?
13:30 fllr joined #salt
13:30 babilen mortis: I am considering something like that - or documenting that this particular pillar is only to be edited on the master ...
13:31 faldridge joined #salt
13:31 babilen mortis: It's more that we have a bunch of developers (whom I all trust, that's really not the point) and I am trying to mitigate the "OMG, everything is horrible" moment if one of them loses a laptop
13:31 _fllr_ joined #salt
13:32 mortis hehe
13:34 racooper joined #salt
13:34 babilen I think the only way to really prevent that is to either prevent external access to that repository or to mix git pillars and "normal"/"local" ones :-/
13:34 nahamu like using a git submodule, perhaps
13:35 nahamu and that submodule can be hosted somewhere that requires credentials only you (and whoever else ins ops...) and the salt masters have
13:36 nahamu so that when setting up a salt master it's as easy as "git submodule init; git submodule update"
13:36 nahamu but when developers check out the main pillar git repo, they don't get they keys
13:37 nahamu babilen: ^
13:37 babilen yeah ... my colleagues are already complaining about the number of git repos they have to handle, but that sounds like a good alternative
13:37 babilen (mr to the rescue btw!)
13:37 mpanetta joined #salt
13:38 JasonSwindle joined #salt
13:38 nahamu right, the goal is to make it easy for those with the right credentials to check out the pillar repo, run a couple of extra commands to have it slurp in the sensitive submodule, and for those without to be unable to do so.
13:39 jaimed joined #salt
13:40 salt80 joined #salt
13:41 timoguin joined #salt
13:41 babilen Can't I just define multiple ext_pillars?
13:41 nahamu it does mean that when keys get updated you and your colleagues will have to correctly commit to the submodule *and* the main pillar repo, and whatever mechanism currently updates the master's copy of the repo will also need to do a git submodule update.
13:41 nahamu babilen: that would probably work too
13:41 nahamu I also don't know that git-fs would be able to handle submodules.
13:41 salt80 hello, I try to test windows reg state module
13:42 faldridge joined #salt
13:42 nahamu I haven't worked with multiple pillars yet.
13:42 babilen I'd rather do that as it is more explicit - I am using multople GitFS repositories already and I don't see a benefit in throwing in the whole "submodule" part
13:42 nahamu good point.
13:42 ipmb joined #salt
13:42 nahamu I was coming at it from a git workflow perspective
13:42 salt80 the documentation is very terse in this case so I checked the module code and prepared this basic test
13:43 nahamu but multiple pillars is probably cleaner from the salt master's perspective.
13:43 babilen nahamu: Neither have I, but I try to keep the whole setup as homogeneous as possible (this seems to define everything I do at $WORK these days) - My colleagues are not too happy with the "Oh, why another repo for that too?!" approach so I think I'll spare them the "Yes, and this is a submodule" one :)
13:44 babilen nahamu: From the git workflow perspective I am using http://myrepos.branchable.com/ to tie it all together
13:44 Networkn3rd joined #salt
13:45 babilen nahamu: Okay, thank you for the discussion so far. I'll try that approach :)
13:45 younqcass joined #salt
13:46 nahamu babilen: let us know how it goes!
13:46 babilen will do!
13:48 salt80 http://justpaste.it/exba
13:49 nahamu salt80: is it not working?
13:49 salt80 the state seems to be ran successfully  but the key in registry is missing
13:52 salt80 btw, why the example within the reg module itself uses double backslash to separate the keys?!?
13:52 seapasulli joined #salt
13:52 nahamu because python treats single backslashes as means to escape characters. "\n" means newline character. Did you try using double backslashes?
13:55 salt80 wit double backslashes, the state ends with an error that the path is invalid so single backslash seems to be correct
13:55 MZAWeb joined #salt
13:55 salt80 does anybody have any working example of working with windows registry?!?
13:56 krak3n` joined #salt
13:58 Gifflen_ joined #salt
14:00 joehoyle_ joined #salt
14:04 joehoyle joined #salt
14:06 sashka_ua joined #salt
14:07 UForgotten joined #salt
14:08 joehoyle joined #salt
14:08 sandbender1512 joined #salt
14:12 seapasulli left #salt
14:12 ajprog_laptop joined #salt
14:12 helderco joined #salt
14:14 joehoyle_ joined #salt
14:15 xunil joined #salt
14:20 joehoyle joined #salt
14:24 yomilk joined #salt
14:25 DammitJim man, stalking?
14:25 AdamSewell joined #salt
14:27 Darnoth joined #salt
14:27 Gifflen joined #salt
14:28 rtucker joined #salt
14:28 jalbretsen joined #salt
14:31 opapo joined #salt
14:31 wendall911 joined #salt
14:32 felixhummel joined #salt
14:33 felixhummel hi! will salt 2014.1.1 be available in the ubuntu PPA soon or should we migrate to the pypi install to get updates?
14:33 Networkn3rd joined #salt
14:33 napper joined #salt
14:34 tyler-baker joined #salt
14:35 AirOnSkin joined #salt
14:35 ndrei joined #salt
14:35 mortis felixhummel: isnt it allready there?
14:36 babilen felixhummel: Debian packages have been uploaded ages ago, but IIRC there were some Ubuntu specific issues
14:36 rtucker I just added ~1700 sls files to my salt master, and now my minions won't finish a run
14:36 AirOnSkin Hey guys. Quick question. I'm trying to figure out the differences of SaltStack Community and the enterprise edition. Are these two products with the same features, limited support for certain things or is the enterprise edition necessary anyway if you're a company trying to manage a few hundred servers?
14:36 rtucker am i missing some kind of setting?
14:37 babilen rtucker: How long did you wait? 1700 SLS are quite a lot.
14:37 rtucker babilen: i've waited ~5 minutes
14:37 sroegner joined #salt
14:38 rtucker babilen: also, nothing is using the sls files, they're just on the file system
14:38 jcockhren hey. Is it known that the settings/directions for using Google Compute Engine for salt-cloud or the cloud module are out of date?
14:38 babilen rtucker: Wait longer, refer to http://docs.saltstack.com/en/latest/topics/jobs/index.html for commands to see active jobs and on how to query their status. Did you also set a higher timeout?
14:38 babilen rtucker: Ah, okay - I misunderstood that.
14:38 jcockhren Can anyone verify that GCE is working with salt cloud with the newer dashboard?
14:39 rtucker babilen: i kind of feel like it's a ulimit issue
14:39 felixhummel mortis: No, 2014.1.0 is latest on the PPA: https://launchpad.net/~saltstack/+archive/salt
14:39 rtucker babilen: but, i've got my ulimit cranked way up
14:39 rtucker ulimit -n
14:39 rtucker 65535
14:39 conan_the_destro joined #salt
14:39 babilen rtucker: Are the jobs still running/active ?
14:39 alunduil joined #salt
14:40 rtucker babilen: doesn't appear as such
14:40 djaime joined #salt
14:41 babilen rtucker: So, why do you think that they are still running?
14:41 rtucker babilen: i don't believe that i said they were still running
14:42 babilen "won't finish a run"
14:42 rtucker running salt-call state.highstate -l debug on the client just hangs
14:42 rtucker no timeout occurs
14:43 babilen felixhummel: joehh would know the exact status (I couldn't find anything in my (complete) irc logs regarding this though)
14:43 rtucker [DEBUG   ] Loaded minion key: /etc/salt/pki/minion/minion.pem
14:43 rtucker is where the minion just sits forever
14:43 rtucker [INFO    ] AES payload received with command _file_list
14:44 rtucker is on the master also just hanging forever
14:44 felixhummel thanks, babilen. I'll wait. :)
14:44 eliasp rtucker: similar things happen here on one of the Win7 minions ;-/(
14:44 babilen felixhummel: Ah, just switch to Debian ;)
14:44 Gordonz joined #salt
14:45 trevorj joined #salt
14:45 eliasp I have a 'state.highstate' running since 25 minutes on one of the Win7 clients…
14:45 rtucker eliasp: boo :(
14:45 Gordonz joined #salt
14:46 jaimed joined #salt
14:46 felixhummel babilen: definitely will do. ever head of "whoopsie"? canonical's spyware on all ubuntu boxes. that was the last straw
14:46 joehoyle joined #salt
14:47 felixhummel http://crunchbang.org/ has some appeal
14:47 felixhummel ok, enough OT ;)
14:53 mgw joined #salt
14:55 seapasulli joined #salt
14:56 rallytime joined #salt
14:56 schimmy joined #salt
14:59 kaptk2 joined #salt
15:01 frasergraham joined #salt
15:01 rdorgueil joined #salt
15:01 rdorgueil joined #salt
15:01 chrisjones joined #salt
15:03 rdorgueil Hello. I'm looking for companies providing services arount salt and configuration management using salt. Do you have pointers where I should look for that ?
15:04 ndrei joined #salt
15:05 ipalreadytaken joined #salt
15:06 joehoyle_ joined #salt
15:08 fxdgear joined #salt
15:08 timoguin rdorgueil, do you mean companies that provide support?
15:08 timoguin Saltstack, Inc does
15:10 rdorgueil timoguin, I mean companies that based on some need definition can do the actual coding work so that I can run my infrastructure based on it
15:10 rdorgueil don't know if it make sens
15:10 rdorgueil e
15:10 ipalreadytaken joined #salt
15:11 timoguin rdorgueil, well, Saltstack will do consultations and help with implementation if that's what you mean
15:12 joehoyle joined #salt
15:14 danielbachhuber joined #salt
15:14 rdorgueil timoguin, I'm looking for companies that can actually realize the work, not just help us doing it.
15:15 timoguin well I'd bet they can do that. you just need to contact sales
15:15 AirOnSkin Can anyone tell me the differences of SaltStack Community and the enterprise edition? Are these two products with the same features, limited support for certain things or is the enterprise edition necessary anyway if you're a company trying to manage a few hundred servers?
15:15 R_Shackleford joined #salt
15:15 timoguin AirOnSkin, it's the same code-base. enterprise is just supported, including any necessary business-specific things
15:16 AirOnSkin timoguin: great, thanks!
15:19 terminalmage AirOnSkin: if you'd like additional info on the enterprise edition, send an email to info@saltstack.com
15:21 Gareth morning
15:22 ndrei joined #salt
15:23 rrives joined #salt
15:26 Ahlee anybody wanna save me some code diving and give me the 1000' overview of how manage.down determines status?  is it intersection("salt \* test.ping", ls(path/to/keys)) ?
15:29 machmalabala joined #salt
15:31 tdawg joined #salt
15:34 jcockhren Ahlee: https://github.com/saltstack/salt/blob/develop/salt/runners/manage.py#L98
15:34 ndrei joined #salt
15:35 jcockhren Basically, uses the client library to call test.ping
15:35 fllr joined #salt
15:35 jcockhren https://github.com/saltstack/salt/blob/develop/salt/runners/manage.py#L26
15:36 jcockhren See also lines 43-44
15:36 Ahlee jcockhren: word, guess I need to look into salt.key.Key(__opts__)
15:36 Ahlee guessing that's what's building the list of accepted keys
15:38 UtahDave joined #salt
15:39 davet joined #salt
15:40 Ahlee is there anything in the workflow for renegotiation of keys?  I see removekeys option, but that seems dangerous since a minion just failing to return at that moment gettings its key yanked out from under it seems bad
15:41 pass_by_value joined #salt
15:41 seapasulli_ joined #salt
15:42 srage joined #salt
15:42 tdawg G'morning! I'm having some trouble using an sls file in a subdirectory. Will that even work? Here's what I'm trying to do in my top.sls: http://hastebin.com/mexuyojoti.coffee
15:42 jcockhren Ahlee: if a key is removed, then if the minion is alive it'll just appear back in the list of keys waiting to be accepted
15:42 hunter joined #salt
15:42 ekristen joined #salt
15:43 timoguin tdawg, use a period to denote directories instead of a slash
15:43 AirOnSkin terminalmage: I'm still evaluating, but thanks for the info :)
15:43 terminalmage AirOnSkin: no prob
15:44 _fllr_ joined #salt
15:44 tdawg timoguin: Really? That's all it is? Man I struggled with this forever last week lol, let me give it a try Thanks!
15:44 srage_ joined #salt
15:45 Gifflen joined #salt
15:45 tdawg timoguin: Some of my minions have periods in their minion id, do I need to escape them for this to work?
15:45 pydanny joined #salt
15:47 timoguin tdawg, errm, i'm not sure, actually.
15:47 timoguin my minions have periods in their ids too, but i don't have any paths with periods in them.
15:48 rawzone joined #salt
15:48 timoguin I would avoid using periods for paths in the salt fileserver
15:49 tdawg I don't have any paths with periods either, are you saying that switching: "subdir/my.minoin" to "subdir.my.minion" it should work?
15:49 tdawg only my minions have periods, no dirs
15:50 ipalreadytaken joined #salt
15:50 timoguin tdawg, looking at your paste: /salt-git-repo/minions/particular-minion/init.sls
15:50 timoguin so "particular-minion" might be "my.minion"?
15:50 tdawg correct
15:50 srage joined #salt
15:51 timoguin tdawg, so there'd be a period in the path to the SLS file.
15:51 timoguin Which I don't think is supported
15:51 tdawg hmmm. I see
15:53 xerxas joined #salt
15:53 tdawg timoguin: Do you used something besides "init.sls" for your states?
15:54 timoguin I use init.sls in addition to other named sls files, so applying 'foostate' will give you the base states, and foostate.otherstuff will install other optional things
15:56 tdawg timoguin: Thanks, I think I misunderstood what you meant about files in paths earlier, it all makes sense now.
15:56 timoguin glad i could help
15:58 mgw joined #salt
16:00 icarus joined #salt
16:01 Gordonz joined #salt
16:02 iondrip joined #salt
16:02 Gordonz joined #salt
16:06 Ahlee jcockhren: /win 10
16:06 Ahlee sorry
16:07 jcockhren LOL
16:08 Gifflen_ joined #salt
16:09 schimmy joined #salt
16:09 viq Is there a way to force a package provider on a host other than setting it in minion config?
16:12 viq Setting it via pillar would be extra nice
16:13 arthabaska joined #salt
16:14 dfinn joined #salt
16:14 jeremyfelt joined #salt
16:14 toastedpenguin joined #salt
16:16 oz_akan_ joined #salt
16:17 KyleG joined #salt
16:17 KyleG joined #salt
16:19 jslatts joined #salt
16:20 oz_akan_ joined #salt
16:20 meteorfox joined #salt
16:20 balboah joined #salt
16:21 tdawg Just saw this in the docs: "Information transferred via pillar is guaranteed to only be presented to the minions" does this mean that everything in the salt root dir is transferred to every minoin?
16:23 schimmy joined #salt
16:23 timoguin tdawg, yes
16:23 timoguin all states and modules are made available to every minion.
16:24 srage joined #salt
16:25 tdawg timoguin: Thanks
16:27 joehoyle joined #salt
16:28 NV tdawg: its not sent unless the minion requests it so dont worry if you have a big tarball in it
16:28 Gordonz joined #salt
16:28 NV but any minion can enumerate and fetch all files
16:29 UtahDave viq: I think you might be able to set it in pillar
16:29 viq UtahDave: any idea how exactly?
16:30 UtahDave just match on the correct minions and set the provider option just like you would in the minion config
16:30 tdawg NV: Ahh ok, I have a private key for one minion that I should now move to a pillar, but the fact that the minion doesn't fetch files until requested is good to know (for tarballs, etc).
16:31 NV yup anything sensitive to pillar
16:31 viq UtahDave: so just http://pbot.rmdir.de/ZgljnqaI65z1OFbqv6E2fw?
16:31 UtahDave viq: yep!
16:32 UtahDave Test it first, but the minion should look for config items in it's own minion config, then pillar, then the master config
16:33 viq pillar.get providers shows me pkg: pkgng, yet salt-call ran on that machine tries to use the other provider for pkg :(
16:34 NV viq: what does test.providers show?
16:35 NV also what fbsd and salt version?
16:35 arrose joined #salt
16:35 viq http://pbot.rmdir.de/6R6yi9lDxQmrAhB1KOsUdw
16:36 UtahDave viq: Hm. I thought that should work.
16:36 viq http://pbot.rmdir.de/sjl5gVomJFykggE7hj0obg
16:36 viq UtahDave: I was hoping it would
16:37 NV ah does 9.2 default to pkg still?
16:37 viq Yes, only 10 defaults to pkgng
16:37 NV just upgrade to 10 and hayo
16:37 viq NV: "just"
16:37 [diecast] joined #salt
16:38 viq NV: not quite an easy option for production machines
16:38 NV arguably salt should parse make.conf on pre-10 to work out if to use pkgng or not...
16:39 srage joined #salt
16:41 bemehow joined #salt
16:42 druonysus joined #salt
16:42 druonysus joined #salt
16:44 tdawg Is there anywhere in the docs that describe how to convert the salt cli format into the sls YAML format? I'm struggling to get the pip module working, I'm not structuring it correctly.
16:44 viq NV: well, all those ways require going into the machines and doing something - even if it's just editing some file, it's still an overhead I'd rather avoid. That's why I'm cetrally managing my servers via salt, because I want to centrally manage my servers, not to go into each and every one of them to do something.
16:44 jcockhren tdawg: need an example for using the pip state module?
16:45 tdawg jcockhren: sure, that will help me out a great deal :)
16:45 viq tdawg: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.pip_state.html is not enough ?
16:46 jcockhren viq++
16:46 NV viq: but to use pkgng on 9.x you need to have WITH_PKGNG="YES" in make.conf already no?
16:46 tdawg viq: *facepalm* I've always looked at the modules docs, wasn't aware of the separate 'states' doc, that is exactly what I needed
16:46 viq NV: only if you use ports to build stuff
16:46 NV true
16:47 viq tdawg: ;)
16:47 Gordonz joined #salt
16:47 NV anyeay you could patch your pkgng state module and push it out to minions using the _states foldrr
16:47 NV folder
16:47 giantlock joined #salt
16:48 viq I guess it's an option, although admittedly somewhat ugly
16:48 NV or use salt to manage the minion config
16:49 Gordonz joined #salt
16:49 viq Yeah, and I'm trying to use salt-formula for that, but one of the first things it does is to make sure that the package is installed... See where this is going? ;)
16:49 NV true...
16:50 hunter joined #salt
16:51 NV actually why not just override the provider for pkg.installed in your formula?
16:51 NV http://docs.saltstack.com/en/latest/ref/states/providers.html
16:52 NV first example
16:52 viq Because I'm also managing centos and debian boxes with it. And I'd like to push the changes upstream
16:52 fedgoat joined #salt
16:53 viq Yeah, I know, {% if grains['os'] == 'FreeBSD' %} but still I think there should be a better way to do it
16:54 NV mhmm
16:54 HeadAIX joined #salt
16:54 NV manage make.conf and autodetect off that imho seems cleanest
16:55 NV or perhaps on the existance of pkgng being installed?
16:55 viq and how will you detect that?
16:55 NV check files?
16:56 Ryan_Lane joined #salt
16:56 viq Actually the current pkgng provider needs a change in that respect. It looks for presence of /usr/local/etc/pkg.conf - which is optional
16:56 [diecast] joined #salt
16:56 viq (and doesn't exist on FreeBSD 10, but I haven't looked there yet)
16:57 alexandrel joined #salt
16:58 alexandrel the mysql state/execution module is doing some funky stuff to pass/grab any mysql exception: return sys.modules[__salt__['test.ping'].__module__].__context__.pop('mysql.error', None)
16:59 viq Eh, I guess I'll file an issue for that
16:59 alexandrel can anyone tell me why it's using the test.ping module's context?
16:59 Gordonz_ joined #salt
16:59 alexandrel well basically, I was thinking of doing something similar in my own state module.
16:59 alexandrel it's ugly, but if it works, it's good.
16:59 Gordonz_ joined #salt
17:00 joehillen joined #salt
17:04 Employee312 joined #salt
17:04 mgw joined #salt
17:05 yomilk joined #salt
17:08 viq hah, https://github.com/saltstack/salt/issues/11666
17:08 viq 11666 :D
17:08 Shenril joined #salt
17:10 alekibango joined #salt
17:11 kballou joined #salt
17:12 ajw0100 joined #salt
17:14 bhosmer joined #salt
17:20 krak3n` joined #salt
17:30 travisfischer joined #salt
17:37 n8n joined #salt
17:38 MedicalJaneParis joined #salt
17:38 toastedpenguin joined #salt
17:40 JasonSwindle joined #salt
17:44 kermit joined #salt
17:45 ggoZ joined #salt
17:45 schimmy joined #salt
17:46 JasonSwindle joined #salt
17:52 tyler-baker left #salt
17:52 abe_music joined #salt
17:55 tyler_ joined #salt
17:56 tyler-baker joined #salt
17:56 tyler-baker joined #salt
17:57 higgs001 joined #salt
17:57 arthabaska joined #salt
18:00 jaimed joined #salt
18:00 conan_the_destro joined #salt
18:01 quickdry21 joined #salt
18:03 tdawg How can I track down the cause of a "Data failed to compile" error? I've got logging turned up to 11...
18:05 RandalSchwartz Oh - I've been getting so many of those
18:06 RandalSchwartz there's usually some sort of hint along with it
18:06 RandalSchwartz although cryptic
18:06 ndrei joined #salt
18:06 RandalSchwartz like "expecting a list but you gave it a hash", etc
18:06 chrisjones joined #salt
18:06 tdawg RandalSchwartz: THE Randal Schwartz? PERL guru and Floss Weekly host?
18:07 RandalSchwartz Not PERL, but Perl, yes
18:07 RandalSchwartz PERL = Protecting Escarpment Rural Lands (perlofburlington.org)
18:07 RandalSchwartz I have nothing to do with that. :)
18:07 Employee312 joined #salt
18:07 tdawg RandalSchwartz: lol
18:08 RandalSchwartz and yeah, I just saltified my first project
18:08 RandalSchwartz builds a vagrant system from a basebox exactly to mimic $client's live site
18:08 RandalSchwartz I learned a lot about salt during that week of head-bang-on-desk :)
18:08 UtahDave RandalSchwartz: woot!
18:08 tdawg strangely I restarted the minion (turned off 'garbage' loggin setting) and the highstate worked...
18:09 oz_akan_ joined #salt
18:09 UtahDave tdawg: usually it means either your pillar or your sls file faile to compile.  something stacktraced.
18:09 forrest joined #salt
18:09 UtahDave off to lunch. back in a few
18:09 stephanbuys joined #salt
18:09 n8n joined #salt
18:09 RandalSchwartz lunch?  Oh yeah, I forget, I'm Turkey
18:09 RandalSchwartz timeline very different here
18:10 RandalSchwartz I'm *in* Turkey
18:10 * Eugene puts away the baster
18:10 RandalSchwartz baster child :)
18:11 Employee312 joined #salt
18:12 tdawg RandalSchwartz: How are you liking salt so far? Was there ever a floss weekly about salt? I can't recall?
18:13 shoma joined #salt
18:14 stephanbuys hi all - is there a module to check the presence of a service? At the moment I’m using file.managed with a url and checking the content of the banner using a pre-computed hash
18:17 Employee312 joined #salt
18:19 Eugene Service as in /etc/init.d/ service? Or something else
18:20 stephanbuys Eugene: in this case its to check if a web service is running on a specific host and port
18:21 stephanbuys its a bit of a monitoring usecase
18:22 Eugene I think you're hunting for the Reactor system, but I've not used it enough to know if it supports monitoring like that.
18:23 Eugene I would use Nagios / Munin myself, possibly having it invoke a reactor call
18:25 jaimed joined #salt
18:25 jcockhren is it me, or does the salt.cloud module seems a bit out of place. To me it's would fall under a runner
18:29 timoguin jcockhren, there's a runner too.
18:30 timoguin the idea of it being a module is so it can be executed on any configured minion.
18:31 ChrisC1 joined #salt
18:31 ChrisC1 hey guys! noob here
18:31 Ryan_Lane ChrisC1: howdy
18:32 ChrisC1 I'm trying to auto-deploy software to Windows minions
18:32 ChrisC1 i can do this manually through salt 'WINDOWS-MINION'' pkg.install OSSEC
18:32 ChrisC1 and it works just fine
18:32 ChrisC1 how can i include this in the highstate?
18:33 ChrisC1 i can't seem to reference the init.sls file directly in top.sls from the windows repo
18:33 timoguin You'd create an SLS that calls the pkg.installed function
18:33 ChrisC1 yes
18:33 ChrisC1 i did that
18:33 ChrisC1 and it worked
18:33 ChrisC1 however
18:34 ChrisC1 Comment: The following packages failed to install/update: OSSEC=2.7.1.
18:34 ChrisC1 it came back that it failed
18:34 ChrisC1 even though it installed properly
18:34 ChrisC1 which i guess I'm ok with, just seems not all the way polished ;)
18:35 timoguin it's not. that could be a bug.
18:35 Networkn3rd joined #salt
18:35 ChrisC1 when it checks for pkg.installed
18:35 ChrisC1 what query is it actually running?
18:35 ChrisC1 pkg.list_pkgs ?
18:36 kzx ChrisC1: I found I had to hand tweak versions of windows packages in the repo file, otherwise after installing the package, and then matching the version during the pkg.installed state it failed
18:36 MindDrive joined #salt
18:36 kermit joined #salt
18:37 kzx Like installing Python 2.7.6 was failing as it turns out windows was reporting its version as 2.7.6150 for some reason
18:39 ChrisC1 is pkg.installed supported in Windows?
18:39 timoguin it is
18:41 sroegner forrest: you there?
18:41 forrest sroegner, yea what's up?
18:42 sroegner i am thinking about making a supervisor formula - do you use supervisor anywhere?
18:43 Masin joined #salt
18:43 kzx ChrisC1: it is, but it needs version matching to work well, so after you install OSSEC=2.7.1 and pkg.installed checks it, it needs OSSEC version reported by windows to be 2.7.1 if it is 2.7.1000 it might not work, so it might try to install it again
18:44 sroegner forrest: it would be very basic - just setting up foundation and helping with platforms
18:45 sroegner forrest: but don't want to re-invent the wheel either and there seems to be quite some support in modules and even a state
18:45 johtso joined #salt
18:46 johtso joined #salt
18:46 ChrisC1 kzx: when I run pkg.list_pkgs on the minion, it reports back as:
18:46 ChrisC1 OSSEC HIDS 2.7.1:
18:46 ChrisC1 2.7.1
18:46 ChrisC1 so how do i check for this in the SLS file for pkg.installed?
18:46 forrest Yea I don't use supervisor at all
18:46 hunter joined #salt
18:46 sroegner forrest: oh, nvmd then
18:47 forrest sorry :\
18:47 ChrisC1 kzx: appears i need to do some more reading here http://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkg.html
18:47 ChrisC1 ;)
18:47 Masin left #salt
18:49 srage joined #salt
18:54 tyler-baker joined #salt
18:54 tyler-baker joined #salt
18:55 tdawg left #salt
19:01 srage_ joined #salt
19:04 seapasulli joined #salt
19:06 gladiatr joined #salt
19:07 colinbits joined #salt
19:08 sroegner forrest: did you notice that most of the links to the salt formula documentation in the formula READMEs are broken now?
19:08 yomilk joined #salt
19:08 MZAWeb joined #salt
19:08 forrest sroegner, again??
19:08 sroegner when was the last time?
19:08 forrest last week
19:08 sroegner same thing then
19:08 forrest they are working for me right now, which links specifically?
19:09 sroegner i'm just behind :/
19:09 forrest ok
19:09 forrest yea I fixed those last week, I stupidly made them all https the first time around
19:09 sroegner :-)
19:09 renoirb Hey guys. Is there a good page in the docs about pillars, Jinja and YAML syntax. Problem is not that I do not understand, is that in our world, we have to jungle with many implementation of same concepts and the syntax twaks are confusing.
19:09 timoguin I just got a 404 coming from the users formula
19:10 renoirb Mostly to loop through a Pillar, keys, content. how to access them through a state.
19:10 johtso timoguin, I fixed the link in the README, but the repo info still has a broken link
19:11 gladiatr greetings.  what mechanism is recommended for setting grains?  I've attempted to execute grains.append (this is for a list of active nagios plugins) from a file template and I've been playing with doing it from a state definition with (so far) no luck.  Any suggestions?
19:11 gladiatr setting grains, specifically from within some part of the state definition
19:11 renoirb … ok, that one is complex enough for me, think it has what I needed: https://github.com/saltstack-formulas/apache-formula/blob/master/apache/register_site.sls
19:12 mattmtl joined #salt
19:13 vbabiy joined #salt
19:20 twobitsprite joined #salt
19:20 gladiatr bah.  nm.  trying to use an executor module rather than the grains state module.
19:22 uzomg joined #salt
19:25 vbabiy_ joined #salt
19:25 vbabiy joined #salt
19:33 hazzadous joined #salt
19:38 helderco joined #salt
19:41 n8n joined #salt
19:52 eliasp trying to develop my first own module from scratch… the final purpose: manage systemwide printers on Windows-Minions… so I have this rough skeleton: http://pastie.org/8983530 it synced successfully to the minion but executing now 'salt my-minion printer.list' just results in "'printer.list' is not available." … did I forget something basic?
19:53 yomilk joined #salt
19:53 eliasp the output when syncing it to the minion was simply "- modules.win_printer" which looks IMHO ok
19:54 bstr joined #salt
19:58 UtahDave eliasp: try running a minion in debug mode in a terminal and then try executing it from the master.
19:58 UtahDave eliasp: that will often tell you if there is a stacktrace or something that is keeping it fromloading
19:58 UtahDave eliasp: I'm excited to see a printer module for windows!!
20:00 UtahDave if that paste is your actual module you're pushing down, the  incomplete  "add" function might be causing a stacktrace
20:02 eliasp UtahDave: ok… well… I'm not used enough to python… wrote Perl for too long… my syntax is screwed up (missing : after method names)
20:02 eliasp UtahDave: I'll better enable pylint or something else in my vim ;)
20:04 hunter joined #salt
20:04 UtahDave eliasp: that's very helpful!  :)
20:04 UtahDave there's a good salt-vim repo
20:05 eliasp just found that as well… looks nice, thanks! ;)
20:05 UtahDave np.
20:06 UtahDave let me know if you need any help with that module. looks cool.
20:06 eliasp UtahDave: if I get this working, quite some more win stuff might follow
20:06 UtahDave cool
20:06 eliasp UtahDave: it would be nice to get this in a state at some point, where it becomes merge-worthy…
20:07 eliasp for now I'll ship this locally here and put it at least on github once the basics are there
20:07 UtahDave cool.   We also have the salt-contrib repo where you can send it if you want to have it looked at or tested before contributing to salt proper
20:08 eliasp ok, great
20:08 whiteinge timoguin, forrest: gah! the docs refactor from a couple weeks ago changed the URL to the formula docs. thanks for pointing that out
20:09 forrest whiteinge, ?
20:09 forrest did someone make an issue?
20:09 whiteinge not afaik
20:10 whiteinge that's a whole lot of links to change...
20:10 * whiteinge ponders alternatives
20:10 forrest sed all the things?
20:11 whiteinge yeah. and i could update the repo URLs via the API easily enough
20:11 timoguin amazon turk
20:11 forrest lol
20:12 timoguin reminds me, they still owe me $8...
20:12 forrest WOAH
20:12 forrest big bucks
20:12 timoguin gotta make my house payment, amazon...
20:14 eliasp timoguin: nah, better use kickstarter… this will get you a 3.5 Mio Mansion
20:14 ChrisC1 question for anyone willing to help: I'm deploying a package via SLS pkg.installed to windows servers.  what would be the best way to blast out an email when the SLS file runs to alert staff to configure the new package?
20:16 rdorgueil joined #salt
20:17 bja joined #salt
20:18 n8n joined #salt
20:19 timoguin ChrisC1, if your windows servers can send email you might try Powershell's Send-MailMessage
20:19 whiteinge timoguin, forrest: updated the URL in the repo title. now for the READMEs...
20:19 UtahDave ChrisC1: there's an smtp returner, as well
20:19 UtahDave http://docs.saltstack.com/en/latest/ref/returners/all/salt.returners.smtp_return.html#module-salt.returners.smtp_return
20:19 ChrisC1 i thought about that, and it's possible
20:20 ChrisC1 i'd like to include the hostname of the minion in question in the email message
20:22 Gareth ChrisC1: there is an smtp module & state...but its only in develop at the moment.
20:23 shoma joined #salt
20:24 ChrisC1 interesting
20:24 ChrisC1 OK
20:24 MZAWeb joined #salt
20:24 n8n joined #salt
20:24 arrose left #salt
20:25 harobed joined #salt
20:26 harobed joined #salt
20:27 harobed joined #salt
20:30 ndrei joined #salt
20:32 teskew joined #salt
20:32 kzx1 joined #salt
20:33 spiette joined #salt
20:37 schimmy joined #salt
20:37 jnylen0 joined #salt
20:37 Supermathie joined #salt
20:38 jnylen0 hi all... I get an error "msgpack.exceptions.UnpackValueError: Data is not enough." on a fresh install of salt-minion on Windows... can anyone help?
20:39 Supermathie Hey everyone. I'm trying to get up and running with salt. I've got a Win and Linux salt client connect and I'm trying to define some custom grains.
20:39 [diecast] joined #salt
20:40 Supermathie I've configured the fileserver to look at /srv/salt as per the suggested default and have a file: /srv/salt/_grains/windows_virtual.py that I'm expecting to get run on the nodes, but it doesn't even seem to be visible.
20:42 mortis Supermathie: maybe you need to take a look at the extension_modules setting in the masterconfig
20:43 UtahDave Supermathie: run   salt \* saltutil.sync_grains   or  salt \* saltutil.sync_all
20:43 Networkn3rd joined #salt
20:43 mortis or highstate i guess
20:43 UtahDave jnylen0: can you pastebin the error?
20:43 yomilk joined #salt
20:44 MZAWeb_ joined #salt
20:44 Supermathie UtahDave: have synced grains before but not sync_all. Should there be files/etc listed in the output of sync_all (there's only grains:\nmodules:\noutputters\n... with nothing listed after)
20:45 Supermathie mortis: it's not set, looking into it
20:45 mortis Supermathie: i had to use that for pillars, dunno if its needed for grains
20:45 jnylen0 UtahDave: http://pastebin.com/9ahHU2FL
20:46 UtahDave jnylen0: is that traceback from the windows minion or from the linux master?
20:47 UtahDave Supermathie: have you tried restarting the minion service?
20:47 jnylen0 UtahDave: from the windows minion
20:47 hunter joined #salt
20:48 Supermathie https://gist.github.com/Supermathie/9901883 <- my salt master config
20:48 Supermathie UtahDave: just did, no change
20:48 UtahDave jnylen0: what was the command you executed that triggered that error?
20:48 Supermathie I would expect salt '*' cp.list_master to list any files on the master?
20:49 jnylen0 UtahDave: just run salt-minion.exe at the command line (I don't have admin rights so the service install didn't work.)
20:49 UtahDave Supermathie: can you stop the minion service, and then from an administrator terminal with elevated permissions could you start the minion in debug mode?    c:\salt\salt-minion.exe -l debug -c c:\salt\conf
20:50 jnylen0 UtahDave: also i'm forwarding ports 4505 and 4506 to the minion machine using ssh
20:50 UtahDave jnylen0: It's the master that needs port 4505 and 4506 open
20:50 UtahDave Supermathie: Yes, in your file_roots
20:51 jnylen0 UtahDave: let me clarify... on the client machine, ports 4505 and 4506 are listening for connections and are forwarded back to the master
20:51 jnylen0 the minion is pointed at localhost
20:52 jnylen0 it does at least attempt to connect... that error only happened after i accepted the minion's key
20:52 Supermathie UtahDave: running as debug, yes. Don't see anything obvious.
20:52 jnylen0 I'm not sure if that's relevant though, the only program i've ever had issues with forwarding ports like that is M$ remote desktop
20:53 DammitJim ls
20:55 UtahDave jnylen0: I've never tried that myself. I'm not sure.
20:55 UtahDave Supermathie: can you pastebin your grains module?
20:55 Supermathie UtahDave: just updated the gist I posted above: https://gist.github.com/Supermathie/9901883
20:56 jnylen0 UtahDave: I can try from a VM with a direct connection if that helps...
20:56 UtahDave jnylen0: yeah, let's try that real quick, to isolate that.
20:57 UtahDave Supermathie: Hm. that should work.   Just a minute. Let me test here
20:59 shoma joined #salt
21:00 Supermathie UtahDave: Yeah, I don't think I'm doing anything weird :) updated the gist with the client config (again, standard install, nothing weird)
21:02 shoma_ joined #salt
21:02 cewood joined #salt
21:03 Supermathie UtahDave: hah apparently i installed this before the 2014.1.0 release... running 0.17.5. Should I update my master?
21:04 robertlai joined #salt
21:06 oz_akan_ joined #salt
21:07 Supermathie UtahDave: Oh! I updated my salt master and all of a sudden everything shows up. Guess it was a version incompatibility?
21:10 Supermathie salt 'win7-salt' grains.item virtual -> virtual: kvm  (excellent :)
21:10 jnylen0 UtahDave: works fine from vm =/
21:11 Supermathie UtahDave: Yeah it's all good now. Figured adding some more virtualization detection would be a lightweight place to start and I have a few good test cases here.
21:12 jnylen0 are there any other ports that salt needs besides 4505 and 4506?
21:17 Ryan_Lane jnylen0: 22 on each minion if using salt-ssh
21:17 Ryan_Lane (or whatever your ssh port is)
21:17 Ryan_Lane on the master I think that's the only two ports needed
21:23 oz_akan_ joined #salt
21:23 Supermathie Is there an "easy" way to push an updated grains/core.py to a client for testing?
21:24 UtahDave Supermathie: awesome!  Yeah, you should always have your master be the same or newer version
21:25 UtahDave jnylen0: Hm. Not sure what's going on there with the other networking setup
21:25 UtahDave Supermathie: just drop a core.py in   /srv/salt/_grains/      and run a grains sync
21:25 UtahDave Supermathie: that should take precedence over the default core.py
21:28 MZAWeb_ joined #salt
21:30 jnylen0 UtahDave: I noticed that the vm was connecting on port 5905
21:31 UtahDave something misconfigured?
21:31 jnylen0 not sure... I just used the default settings
21:32 elfixit joined #salt
21:32 jnylen0 I forwarded that port as well, and now i am getting hard crashes and ZMQ errors
21:35 johtso joined #salt
21:36 johtso joined #salt
21:36 faldridge joined #salt
21:37 diegows looking for some ideas
21:37 diegows I have an application the uses a backend, an API
21:37 diegows every app requires a key to access the api, they key is obtained from the api servers
21:37 diegows everything is running salt
21:38 diegows what can I use to get the app key every time a new app server is launched ?
21:38 diegows I have some ideas (peer communication, reactor) but I would like to hear another opinions
21:38 * Gareth slips UtahDave a fiver to merge his pull request
21:38 alunduil joined #salt
21:39 UtahDave diegows: I would just have the master push down the appropriate key
21:40 diegows there are two parts, 1- get the key, 2- set the key
21:40 diegows the things is how to create something statul of this
21:40 snuffeluffegus joined #salt
21:40 diegows stateful
21:42 * UtahDave accepts Gareth's fiver if he replaces the double quotes with single quotes.    :)
21:42 snuffeluffegus joined #salt
21:42 Gareth double quotes? hmm
21:42 * Gareth looks
21:42 UtahDave Tom hates 'em.  :)
21:42 Ahlee that's cool, nobody can be right all the time
21:43 UtahDave diegows: So you have a variable number of salted applications?
21:43 UtahDave Ahlee: :)
21:44 Supermathie Getting the "KeyError: 'cmd.run'
21:44 rps_ joined #salt
21:44 Supermathie ... problem referencing __salt__ in core.py.
21:44 Gareth UtahDave: so all double quotes?
21:45 higgs001_ joined #salt
21:45 UtahDave no.  Use single quotes everywhere, unless you have to use double quotes
21:45 rps_ I would like to configure two systems where one has to know about the other. I know how to do this via pillars but is there a way to do this dynamically using salt-cloud? Just looking for a pointer
21:45 diegows UtahDave, yes
21:45 UtahDave Supermathie: __salt__ isn't available in core.py
21:46 dvogt joined #salt
21:46 Gareth UtahDave: sorry. I meant...get rid of all the double quotes.  I think I see where they are.
21:46 UtahDave Gareth: there's just a few
21:46 diegows UtahDave, I have N frontends and M backends, front end needs a key to access the backends, BTW, the same key works for all the backends, so the number of backends doesn't matter
21:47 Supermathie UtahDave: Isn't that what https://github.com/saltstack/salt/blob/develop/salt/grains/core.py#L38 does as a workaround?
21:47 Supermathie ... hence, I can't just push a new core.py into /srv/salt/_grains to test changes
21:48 UtahDave Supermathie: Hm. I see. Shouldn't it still work after syncing the new core.py, though?
21:48 joehoyle joined #salt
21:48 Supermathie %s/__salt__/__coresalt__/g makes it work :D
21:49 dvogt Hi all. I have an interesting problem. using 2014.1.0
21:49 dvogt trying to use the reactor system
21:49 dvogt I have it working but the reactor only kicks off once
21:49 UtahDave diegows: How are you creating the new key for each new application?
21:50 diegows running a command that return it from stdout
21:50 xunil___ joined #salt
21:50 dvogt Subsequent events don't fire off the reactor
21:51 kermit joined #salt
21:51 UtahDave dvogt: try running your salt-master in a terminal in debug mode and see if you're getting a stacktrace in the reactor.  That can cause it to stop
21:51 dvogt utahdave: thanks, i will try that.
21:51 UtahDave diegows: how are you authenticating the new applications?
21:52 diegows UtahDave, they need the key in a specific file
21:53 UtahDave diegows: I mean, do you trust them when they're requesting an api key?
21:54 xunil joined #salt
21:55 diegows UtahDave, well, I want to trust in Salt not in the app :)
21:55 rps_ I will keep searching google, but a way to use the IP of an instance created by salt-cloud to configure another instance would be really useful. Any pointers very welcome
21:55 diegows rps_, there is a way
21:56 diegows rps_, I write an answer in stackoverflow about that
21:56 diegows wait
21:56 rps_ diegows: yay!
21:57 smurfy_ joined #salt
21:57 diegows http://stackoverflow.com/questions/20897796/using-the-output-of-salt-to-be-used-as-input-for-an-sls-state-or-pillar/20940053#20940053
21:57 n0c joined #salt
21:58 smurfy_ are module functions intended for use directly in administration, or is it best practice to reference these from inside states/templates?
21:59 rps_ diegows: neat! Terrifying regarding the security model but neat!
21:59 rps_ diegows: thanks for the help!
21:59 diegows rps_, you are welcome
22:00 UtahDave OK, well, assuming that if they've been accepted by the Salt master, then I would create a state that pushes the key down to the minion, and I would have a little bit of jinja that looks for the key in the minion's pillar and which will run the command to create the key if it doesn't exist
22:00 danielbachhuber joined #salt
22:01 UtahDave smurfy_: it's fine to use them in both ways.  Using them in a state is a little safer because it's easier to have the command double-checked by someone.
22:01 cyrusdavid joined #salt
22:02 UtahDave Gareth: ready for me to merge that yet?
22:02 smurfy_ what about using it directly, i.e. useradd, how do you get that into pillar?
22:02 ndrei joined #salt
22:02 smurfy_ or is it better to edit the pillar data and run it from inside state
22:02 UtahDave smurfy_: I'm a little confused. Why would you put useradd into pillar?
22:03 smurfy_ in this example i'd be stroring users in pillar
22:03 bhosmer joined #salt
22:03 dvogt joined #salt
22:04 smurfy_ it seems like running a module function directly could cause a conflict with a file managed by the state
22:04 fedgoat say i wanted to add an entry into resolv.conf    so i create a module to salt '*' dns.add nameserver blahblah.com  this updates the resolv.conf on the minions but the data source / pillar dosnt get updated.  Wouldn't it make sense i fyou're going to push changes to the minions files that you also update the pillar and do a commit to git also in 1 step rather than 3 steps
22:04 Gareth UtahDave: Yeah.  couple minutes.  I had started adding a state as well so I had made some changes.
22:04 ajw0100 joined #salt
22:05 yomilk joined #salt
22:06 diegows UtahDave, that's fine, but I should run the command in the other minion, how can I run the command in the other minion? using peering?
22:07 UtahDave Gareth: cool
22:08 UtahDave diegows: You could do that.  I would probably send a message through the event system and have the reactor tell the other minion to run the command.  Just to have a bit more control over who does what.
22:08 diegows ok
22:13 * Heartsbane blames whiteinge.
22:13 joehoyle_ joined #salt
22:14 Gareth UtahDave: changes pushed, just waiting on jenkins to finish.
22:15 UtahDave cool
22:16 dvogt @UtahDave: regarding the reactor, At the end of the first rector run I got:
22:16 dvogt Exception AttributeError: "'ReqServer' object has no attribute 'work_procs'" in <bound method ReqServer.__del__ of <salt.master.ReqServer object at 0x16a7310>> ignored
22:16 dvogt Exception TypeError: 'an integer is required' in <bound method SaltEvent.__del__ of <salt.utils.event.SaltEvent object at 0x1008d10>> ignored
22:16 dvogt No other stack traces.
22:17 UtahDave dvogt: what version of Salt are you running?
22:17 dvogt Maybe I need an exit code
22:17 dvogt 2014.1.0
22:17 dvogt The same problem happen 0.17.5
22:18 Supermathie It's against the version I'm running (v2014.1.0) not head, but do people like https://github.com/NetDirect/salt/commit/b31cec606d3be92b0b5a01fd1430997d57e5318b ?
22:18 Supermathie TTFN
22:18 yomilk joined #salt
22:20 UtahDave dvogt: can you pastebin the code and configurations related to this?
22:21 bstr joined #salt
22:22 UtahDave Gareth: it passed the lint test, so I'm going to merge it.
22:22 Gareth UtahDave: Okay cool.  Thanks :)
22:24 JasonSwindle joined #salt
22:24 higgs001 joined #salt
22:24 zain joined #salt
22:25 dvogt UtahDave: I am in a meeting right now, I will post them later. Thanks
22:25 UtahDave dvogt: ok. I might not be online at that time. If not, just send it to the salt-users email list
22:26 garthk joined #salt
22:30 ipalreadytaken joined #salt
22:44 [diecast] joined #salt
22:45 BogdanR joined #salt
22:51 quickdry21 joined #salt
22:56 conan_the_destro joined #salt
22:58 oz_akan_ joined #salt
23:00 joehoyle joined #salt
23:00 teskew joined #salt
23:02 robertlai joined #salt
23:03 honestly joined #salt
23:05 robertlai joined #salt
23:08 ipalreadytaken joined #salt
23:09 andrej How do I pass the a grain from a state to a reactor?
23:09 andrej I think I have understood the selection mechanism on the the reactor side.
23:10 andrej but how do I get e.g. the IP of the minion that the state was run against into a corresponding reactor sls?
23:13 ipalreadytaken joined #salt
23:13 dvogt joined #salt
23:14 coderanger joined #salt
23:14 mateocon` joined #salt
23:14 coderanger left #salt
23:15 n8n joined #salt
23:17 ekristen joined #salt
23:18 KyleG1 joined #salt
23:19 KyleG joined #salt
23:21 baniir joined #salt
23:29 robertla_ joined #salt
23:31 MZAWeb_ joined #salt
23:33 untamo13 joined #salt
23:33 meteorfo_ joined #salt
23:37 arthabaska hi everyone, say I'm setting custom grains in /etc/salt/grains on minions at boot with cloudinit on ec2, but they can't be seen from/by the master for targeting until after a * highstate. what is the correct way to get these to sync, assuming the grains file will be in place before the minion starts?
23:40 Ryan_Lane arthabaska: you could have your cloudinit do a highstate run locally
23:42 shoma_ joined #salt
23:45 arthabaska Ryan_Lane: ah, that makes perfect sense. thanks
23:49 sroegner joined #salt
23:50 joehoyle_ joined #salt
23:53 Ryan_Lane yw
23:53 n8n joined #salt
23:58 oz_akan_ joined #salt

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