Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-11-19

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

All times shown according to UTC.

Time Nick Message
00:00 cruatta my coworker told me he tried it like that but I'll have to verify for myself
00:00 UtahDave OK.  I'll be back in a little bit.
00:02 linjan joined #salt
00:02 ekristen joined #salt
00:08 cholcombe973 looks like my rest_cherrypy server is running properly but it's not sending events to my reactor sls files
00:14 cholcombe973 nvm i figured it out
00:16 cholcombe973 salt: how can i test my reactor sls files?
00:16 cholcombe973 my master says it failed to render but doesn't tell me what failed
00:16 aurynn development environments!
00:17 cholcombe973 lol
00:17 cholcombe973 aurynn: can i call the sls file directory?
00:17 aurynn ?
00:17 cholcombe973 like salt-call something.sls
00:17 cholcombe973 to just run that one file
00:17 iggy reactor debugging is definitely a weak point in salt
00:18 iggy best I can suggest is turning up debug and watch the logs
00:18 aurynn dunno; I set up a master/minion for testing my reactor stuff
00:18 skyler Anyone using the bind formula on ubuntu?
00:18 hal58th salt-call '*' state.sls somestate
00:18 iggy or run the master in the foreground with -l debug
00:18 cholcombe973 oh cool ok
00:19 cholcombe973 iggy: yeah i'm running the master with debug
00:19 cholcombe973 it doesn't seem to be giving enough info so i'll comment out my sls file and do it piece by piece
00:21 jonatas_oliveira joined #salt
00:21 jaimed joined #salt
00:21 jonatas_oliveira joined #salt
00:28 spielberg joined #salt
00:28 UtahDave joined #salt
00:31 holler joined #salt
00:34 cholcombe973 so in reactor can you not call cmd.run?  it seems to want to use cmd.async
00:40 yomilk joined #salt
00:47 manfred it can call runners, i do not believe it can call modules
00:47 manfred runners or states*
00:47 manfred it technically can only call .sls files
00:48 manfred actually
00:48 TTimo joined #salt
00:48 manfred cholcombe973: it can run cmd.run
00:48 manfred local.cmd.run
00:48 manfred http://docs.saltstack.com/en/latest/topics/reactor/#understanding-the-structure-of-reactor-formulas
00:48 lowkeyliesmith joined #salt
00:48 cholcombe973 manfred: thanks.  I missed that part :)
00:49 manfred local. is the localclient, which is anything that is run from a module. runner. is a salt-run thing, and wheel. is some wheel client thing, usually with salt-key i believe
00:49 manfred looks like local. was only added in 2014.7
00:49 manfred ahh, renamed cmd. to local.
00:50 druonysus joined #salt
00:53 aqua^mac joined #salt
00:55 cruatta so this works
00:55 cruatta https://gist.github.com/anonymous/257072d0a9de916485ce
00:55 delinquentme is master > minion communicvation default encrypted?
00:56 cruatta but this fails miserably https://gist.github.com/anonymous/dff84f2317660dcede23
00:56 * ingwaem is back
00:56 cruatta kwargs is definitely not making it from the pkg state to the pkg module
00:57 manfred https://github.com/saltstack/salt/blob/develop/salt/states/pkg.py#L930 it is right there
00:58 manfred and it isn't being overwritten anywhere
00:58 cruatta i'm not sure why it's failing
00:58 nitti joined #salt
00:58 cruatta i saw the same thing
00:59 manfred and i double checked, it isn't getting overwritten anywhere
00:59 cruatta This is what i'm getting when i use the failing code
00:59 manfred what version of salt is on the minion?
00:59 cruatta https://gist.github.com/anonymous/a0bfdc7735df7358aacd
00:59 cruatta 2014.7.0
01:00 manfred oh,
01:00 manfred try
01:00 manfred - kwargs:
01:00 manfred - disablexcludes: main
01:00 manfred it might need to be an ordered dictionary?
01:00 manfred nah, that shouldn't matter
01:01 zions joined #salt
01:02 manfred cruatta: did you mean disablerepo: main?
01:02 cruatta nope. disableexcludes
01:02 cruatta we use yum excludes on some packages
01:03 manfred but that just disables a package named main doesn't it?
01:03 mapu joined #salt
01:03 manfred disable excludes*
01:03 manfred or does it disable all the excludes on packages in the main repository?
01:06 cruatta it refers to the section in yum.conf where you specify the exclude statement
01:06 manfred oh
01:06 manfred i know
01:06 cruatta it disables all in main
01:06 manfred kwargs in that is just refering to everything else you have in the state
01:06 manfred so use
01:07 manfred php54-bcmath:
01:07 manfred pkg.installed:
01:07 manfred - disableexcludes: main
01:07 manfred that should work
01:07 manfred since disableexcludes isn't one of the defined thing in the def installed(), it then gets put into **kwargs,
01:08 manfred **kwargs gets all the key=value assignments passed to a function that aren't actually defined in the function assignment
01:08 manfred http://stackoverflow.com/questions/1769403/understanding-kwargs-in-python
01:09 manfred i think it might still fail, because the kwargs don't appear to get passed to the repoquery
01:09 manfred that checks if the package exists
01:10 manfred [ERROR   ] Command 'repoquery --plugins --queryformat="%{NAME}_|-%{VERSION}_|-%{RELEASE}_|-%{ARCH}_|-%{REPOID}"  --disableexcludes=\'main\' --all --quiet --whatprovides php' failed with return code: 2
01:10 manfred [ERROR   ] stderr: Usage: repoquery [options]
01:10 manfred repoquery: error: no such option: --disableexcludes
01:10 manfred it is going to fail on that
01:11 anotherZero joined #salt
01:11 manfred alternatively, you could do a module.run like you have, and then add an - unless: rpm -q php54-bcmath
01:11 manfred then the state will only run if php54-bcmath isn't already installed
01:12 cruatta yeah. i got the same failure you mentioned
01:14 manfred the problem is that repoquery doesn't support disableing excludes, and it would need to, even if it didn't fail there, at that point, the whole state would stop, because it only goes onto the pkg.install if it finds the package in the repositories
01:14 manfred alternatively, you could modify it, and drop a custom one without the repoquery in /srv/_states
01:19 UtahDave mr_chris: Hey, sorry. I missed you again.  catch you tomorrow!
01:19 UtahDave left #salt
01:23 delinquentme joined #salt
01:23 pssblts joined #salt
01:37 lpmulligan joined #salt
01:42 hunmonk_ joined #salt
01:44 hunmonk_ i need to do something once on initial server config, and never have it run again during a highstate, is there an accepted way to accomplish this?
01:46 manfred you can touch a file in a later state that only happens when your state gets run, like with onchanges
01:46 TTimo joined #salt
01:46 manfred and set unless: test -f /path/to/file
01:47 manfred or you could wrap it with a grain, so that the grain gets created on the first run, but the state doesn't get added in later runs
01:47 manfred wrap it with a jinja if ! something.defined or whatever
01:50 skyler Anyone here know bind? I am using the bind formula, and I have found section that is causing it not to work, but I would like to see if there was a reason it is there before changing it.
01:51 ingwaem you could create an sh file that runs your script, but at the end replaces the contents of the file with either blank, or a "date" call, that way if the highstate does call it again it will be harmless to the system
02:03 TTimo joined #salt
02:03 aw110f joined #salt
02:04 aquinas joined #salt
02:05 smcquay joined #salt
02:05 BrendanGilmore joined #salt
02:05 aw110f Hi, when targetting a set of minions, when one of the minion is down, is it possible for salt to continue running on the rest of the machines ?
02:07 skyler aw110f: If I understand you correctly, this is the default behavior. You will see changes on all of the minion that are up. The  ones that are not, will not report anything.
02:10 Ryan_Lane left #salt
02:10 aw110f skyler: So how can you apply states or run remote commands on a cluster of many machines where down nodes will be expected?
02:10 Ryan_Lane joined #salt
02:11 Ryan_Lane joined #salt
02:11 skyler aw110f: Say you want to run a command on all machines on a large cluster, you could do this: salt '*' test.ping
02:12 skyler you would only see results for the nodes that are up. The down nodes would just be missing.
02:12 skyler this works for any other commands as well: salt '*' state.highstate
02:12 skyler If you would like a list of nodes that are down, you can run: salt-run manage.down
02:14 skyler I currently have 28 minions for my salt master. 1 of them is down. The one being down does not reduce salt's usability at all.
02:16 malinoff joined #salt
02:16 aw110f yep i did a test shutting down few minions at a time and the count of minions that are up is correct
02:17 aw110f thanks skyler:
02:17 skyler aw110f: np
02:22 DaveQB joined #salt
02:23 laderhiton joined #salt
02:23 jeddi joined #salt
02:25 yomilk_ joined #salt
02:31 SheetiS joined #salt
02:34 Guest7935 joined #salt
02:40 bhosmer joined #salt
02:45 hasues joined #salt
02:45 hasues joined #salt
02:45 hasues left #salt
02:53 jimklo_ joined #salt
02:54 possibilities joined #salt
02:59 perfectsine joined #salt
03:01 glyf joined #salt
03:02 Outlander joined #salt
03:02 mr_chris joined #salt
03:03 cads joined #salt
03:06 catpig joined #salt
03:07 miqui joined #salt
03:09 glyf joined #salt
03:11 bhosmer joined #salt
03:26 feythinl joined #salt
03:26 Outlander joined #salt
03:27 possibilities joined #salt
03:31 pssblts joined #salt
03:36 Mso150 joined #salt
03:41 possibilities joined #salt
03:44 hunmonk_ joined #salt
03:49 delinquentme joined #salt
03:49 delinquentme calling a salt-cloud instancing event ... won't place the minions in a highstate right?
03:50 delinquentme like i can get replies back from the minions without giving them .sls files ... as the events for .sls files come after a highstate event
03:51 funzo joined #salt
04:00 ecdhe joined #salt
04:10 Gnouc joined #salt
04:21 yomilk joined #salt
04:27 smcquay joined #salt
04:40 oldmantaiter joined #salt
04:43 TyrfingMjolnir joined #salt
04:46 styly joined #salt
04:46 tligda joined #salt
04:50 jp_ joined #salt
04:56 mapu joined #salt
05:00 jonatas_oliveira joined #salt
05:02 jonatas_oliveira joined #salt
05:02 thayne joined #salt
05:04 jonatas_oliveira joined #salt
05:06 jonatas__ joined #salt
05:08 jonatas_oliveira joined #salt
05:10 jonatas_oliveira joined #salt
05:12 jonatas__ joined #salt
05:14 jonatas_oliveira joined #salt
05:15 clinta joined #salt
05:15 jonatas_oliveira joined #salt
05:17 jonatas_oliveira joined #salt
05:19 jonatas_oliveira joined #salt
05:21 jonatas__ joined #salt
05:22 dooshtuRabbit joined #salt
05:23 jonatas_oliveira joined #salt
05:25 jonatas_oliveira joined #salt
05:27 yomilk_ joined #salt
05:27 jonatas_oliveira joined #salt
05:29 jonatas_oliveira joined #salt
05:31 jonatas_oliveira joined #salt
05:32 jonatas_oliveira joined #salt
05:34 jonatas__ joined #salt
05:36 jonatas_oliveira joined #salt
05:38 felskrone joined #salt
05:38 jonata___ joined #salt
05:40 jonatas_oliveira joined #salt
05:42 jonatas_oliveira joined #salt
05:44 jonatas_oliveira joined #salt
05:46 jonatas_oliveira joined #salt
05:48 jonatas_oliveira joined #salt
05:49 jonatas__ joined #salt
05:50 ramteid joined #salt
05:50 superted666_ joined #salt
05:51 jonatas_oliveira joined #salt
05:53 jonatas_oliveira joined #salt
05:54 jimklo joined #salt
05:54 techdragon joined #salt
05:55 jonatas__ joined #salt
05:57 jonatas_oliveira joined #salt
05:59 jonatas__ joined #salt
06:01 jonatas_oliveira joined #salt
06:03 jonatas__ joined #salt
06:04 micko joined #salt
06:05 jonatas_oliveira joined #salt
06:07 jonatas__ joined #salt
06:09 jonatas_oliveira joined #salt
06:10 jonatas_oliveira joined #salt
06:12 jonatas__ joined #salt
06:14 jonatas__ joined #salt
06:16 jonatas_oliveira joined #salt
06:18 jonatas__ joined #salt
06:19 micko joined #salt
06:20 jonatas_oliveira joined #salt
06:22 kermit joined #salt
06:22 jonatas_oliveira joined #salt
06:24 jonatas__ joined #salt
06:26 jonatas_oliveira joined #salt
06:27 jonatas__ joined #salt
06:29 jonatas_oliveira joined #salt
06:30 swa_work joined #salt
06:31 jonatas_oliveira joined #salt
06:33 jonatas_oliveira joined #salt
06:34 Furao joined #salt
06:35 jonatas_oliveira joined #salt
06:37 jonatas_oliveira joined #salt
06:39 jonatas_oliveira joined #salt
06:41 jonatas__ joined #salt
06:43 jonatas__ joined #salt
06:45 jonatas_oliveira joined #salt
06:46 jonatas__ joined #salt
06:47 catpigger joined #salt
06:49 jonatas_oliveira joined #salt
06:50 linjan joined #salt
06:50 jonatas_oliveira joined #salt
06:52 jonatas_oliveira joined #salt
06:54 jonatas_oliveira joined #salt
06:56 jonatas_oliveira joined #salt
06:57 CeBe joined #salt
06:58 jonatas_oliveira joined #salt
06:59 Hell_Fire joined #salt
07:00 yomilk joined #salt
07:00 jonatas_oliveira joined #salt
07:01 TheThing joined #salt
07:02 jonatas_oliveira joined #salt
07:04 jonatas_oliveira joined #salt
07:05 colttt joined #salt
07:05 jonatas__ joined #salt
07:07 jonatas_oliveira joined #salt
07:08 herzi joined #salt
07:08 DaveQB joined #salt
07:09 jonatas__ joined #salt
07:10 Hell_Fire_ joined #salt
07:11 jonatas_oliveira joined #salt
07:12 bhosmer joined #salt
07:13 jonatas__ joined #salt
07:15 jonatas__ joined #salt
07:15 oyvjel joined #salt
07:17 jonatas_oliveira joined #salt
07:19 jonatas__ joined #salt
07:21 jonatas_oliveira joined #salt
07:23 jonatas_oliveira joined #salt
07:23 Hell_Fire_ joined #salt
07:23 __gotcha joined #salt
07:24 jonatas_oliveira joined #salt
07:26 jonatas_oliveira joined #salt
07:28 jonatas_oliveira joined #salt
07:30 jonatas__ joined #salt
07:30 linjan joined #salt
07:32 Hell_Fire_ joined #salt
07:35 JlRd joined #salt
07:37 flyboy joined #salt
07:38 CeBe joined #salt
07:41 Hell_Fire joined #salt
07:43 styly_ joined #salt
07:47 Ryan_Lane joined #salt
07:51 Hell_Fire joined #salt
07:51 tomspur joined #salt
07:56 Hell_Fire joined #salt
07:59 yomilk_ joined #salt
08:04 chiui joined #salt
08:06 lcavassa joined #salt
08:08 slafs joined #salt
08:08 slafs left #salt
08:09 slafs joined #salt
08:16 agrundy joined #salt
08:21 __gotcha1 joined #salt
08:22 Hell_Fire joined #salt
08:23 Mso150 joined #salt
08:24 picker joined #salt
08:29 Hell_Fire_ joined #salt
08:31 bhi joined #salt
08:38 eliasp just restarted my master (2014.1.13), and now all my minions (2014.1.10 on Win7 and 2014.1.13 on Ubunu 14.04) stopped responding… restarted one of the minions manually and it responds now again, but what about all the other minions? I don't feel like they're recovering on their own - this is going on now for ~15 minutes
08:38 Hell_Fire_ joined #salt
08:39 eliasp any ideas what's going on here? seen this before?
08:48 Flusher joined #salt
08:49 eliasp hmm, getting "The minion failed to return the job information for job 20141119093940465845. This is often due to the master being shut down or overloaded. If the master is running consider incresing the worker_threads value." on the minions…
08:49 shorty_mu joined #salt
08:53 delinquentme joined #salt
09:02 PI-Lloyd joined #salt
09:05 eliasp the minions are constantly (every ~70 seconds) reporting: "[salt.crypt][WARNING ] SaltReqTimeoutError: Waited 60 seconds" … this is weird
09:09 linjan joined #salt
09:10 ckao joined #salt
09:14 karimb joined #salt
09:17 alanpearce joined #salt
09:18 slav0nic joined #salt
09:19 jonatas_oliveira joined #salt
09:25 fredvd joined #salt
09:28 DolourousEdd joined #salt
09:28 CeBe joined #salt
09:30 Hell_Fire_ joined #salt
09:35 oyvjel joined #salt
09:35 ingwaem joined #salt
09:37 DaveQB joined #salt
09:37 Couch joined #salt
09:44 otter768 joined #salt
09:46 __gotcha joined #salt
09:48 linjan joined #salt
09:55 feythin_ joined #salt
10:01 pdayton joined #salt
10:02 alexr__ joined #salt
10:05 wnkz joined #salt
10:06 viq Bah, any idea when the stable debian packages will be out?
10:10 DaveQB joined #salt
10:16 Hell_Fire joined #salt
10:16 eliasp damn… no communication between master + minions at all since more than 2h … all configurations look ok, moved the master + minion cache directories away… no changes so far
10:16 eliasp looks like it's time for tcpdump ;(
10:16 P0bailey joined #salt
10:17 P0bailey joined #salt
10:18 ingwaem joined #salt
10:22 pdayton joined #salt
10:24 flyboy82 viq: I'm expecting that too much too eagerly
10:25 petan joined #salt
10:26 viq Annoying thing is that part of our infrastructure is centos, and it's already in epel
10:28 flyboy82 my orchestrator depends on it... I'm using mine.get in the orch sls and it's not running cuz there's this bug that's been fixed in 2014.7
10:28 wnkz joined #salt
10:29 ckao joined #salt
10:29 CeBe joined #salt
10:36 __gotcha joined #salt
10:38 rawkode joined #salt
10:49 prognostikos joined #salt
10:51 ingwaem joined #salt
10:53 Outlander joined #salt
10:57 Fbraad joined #salt
10:58 felskrone joined #salt
11:06 joehh viq, flyboy82: I'll be cpoying from testing to the main repos in the next few minutes
11:07 luminous eliasp: that SaltReqTimeoutError might need a service restart for master/minion
11:08 flyboy82 cheers joehh!!!
11:08 luminous I'm pretty sure I've seen that, though I'm not sure if the restart resolved it.
11:08 jonatas_oliveira joined #salt
11:08 eliasp luminous: yeah… walking now through the whole building and restarting the service on each Win minion manually… ;-(
11:08 luminous eliasp: try restarting the master first?
11:08 eliasp luminous: reminds me not to disable SSH on Linux minions yet…
11:08 luminous hah!
11:08 luminous yes
11:08 eliasp luminous: I did, several times, cleared the cache… waited… etc.
11:08 luminous eliasp: do not depend on salt remote exec..
11:09 luminous enjoy it
11:09 luminous but ensure you have other methods
11:09 luminous to prevent you from walking around
11:09 luminous :)
11:09 eliasp luminous: thanks! might be a good reason to deploy now psexec to Win minions
11:09 luminous yea, I don't do windows anymore, so I don't have good recommendations there
11:10 eliasp luminous: you can't really "do windows", it's more like "windows does you" … :)
11:10 joehh can you try now - I think I have copied all the deps in too
11:11 flyboy82 u're awesome!
11:12 * flyboy82 buys joehh a beer if he ever comes to Greece
11:12 eliasp joehh: does this mean we'll see 2014.7.x on Ubuntu too in the near future or are there issues to be solved first?
11:12 Furao joined #salt
11:14 brayn joined #salt
11:15 joehh eliasp: the same issue remains on both (salt-api not yet packaged)
11:15 eliasp ok
11:15 joehh not expecting a hold up there other than getting time over the next few days
11:16 eliasp joehh: great! thanks a lot for your work!
11:16 joehh one other ubuntu issue - hasn't been packaged for lucid yet
11:17 agend joined #salt
11:17 joehh request has been made to launchpad for copying of packages
11:17 joehh could take 5 minutes, could take an hour
11:18 eliasp \o/
11:19 ze- mmm... anyone knows when is master/minions/xxx/data.p saved ?
11:20 flyboy82 if that's the mine_functions returned data I think it's every $mine_interval  seconds
11:21 ze- there is a mine.d is that directory (and I don't use mines)
11:21 diegows joined #salt
11:21 ze- the data.p is used by CkMinions.connected_ids
11:23 peters-tx joined #salt
11:23 luminous thanks joehh!
11:23 luminous where are you that this is a reasonable time to be working?
11:23 luminous :P
11:24 NightMonkey joined #salt
11:25 joehh depends on your definition of reasonable - AU, melbourne currently 10:24pm
11:26 joehh copying the packages isn't real work anyway - the real work was building them and backporting the deps a little while ago
11:26 istram joined #salt
11:28 joehh looks like the launchpad packages have copied for ubuntu
11:29 ssteinerX joined #salt
11:29 ssteinerX left #salt
11:35 Furao can we just have mobile hardware we can run xen or whatever and have both android and a linux desktop?
11:35 Furao they have enough ram cores
11:37 agend joined #salt
11:41 danburbridg joined #salt
11:43 linjan joined #salt
11:43 Furao oups wrong window :)
11:45 otter768 joined #salt
11:47 mick3y joined #salt
11:47 mick3y g'morning
11:49 mick3y could someone please explain this message to me:
11:49 mick3y warnings: - '__reqs__' is an invalid keyword argument for 'service.mod_watch'. If you were trying to pass additional data to be used in a template context, please populate 'context' with 'key: value' pairs. Your approach will work until Salt Lithium is out. Please update your state files.
11:49 Ancient joined #salt
11:49 mick3y i get this warning when service state fails
11:50 mick3y or when service is being restarted
11:50 jonatas_oliveira joined #salt
11:52 goodwill joined #salt
11:52 bhosmer joined #salt
11:57 Ancient joined #salt
11:58 linjan_ joined #salt
11:59 N-Mi_ joined #salt
11:59 N-Mi_ joined #salt
12:07 jonatas_oliveira joined #salt
12:15 Couch Hi, I had a question about Salt. -> Does Salt provide anything to perform hierarchical lookups of data ? If so, does it support merging hashes and arrays from different levels of the hierarchy. ?
12:15 Couch In short, does it have replacement functionality for Puppet's Hiera ? Specifically the merged lookup functionality
12:17 CeBe joined #salt
12:22 dooshtuRabbit joined #salt
12:26 glyf joined #salt
12:27 luminous Couch: in short, yes
12:27 luminous Couch: Salts 'hiera' is an interface known as 'pillar'
12:29 Couch luminous, thank you. and does 'pillar' also support those features ?
12:29 luminous there is a set of internal pillar modules that compile templates (rendered by one or more renderers) as .sls files into a yaml dictionary to associate with a minion. There are then ext_pillar modules that let you connect in other data sources, be it a command, hiera, postgresl, etc
12:29 Couch luminous, ok so it's possible. thanks very much
12:29 luminous the most recent version of salt includes the type of more intelligent merging you are talking about
12:29 Couch awesome
12:29 luminous earlier versions will step on hash keys
12:30 luminous Couch: it is worth looking at reclass (https://github.com/madduck/reclass) if the pillar merging does not work to your liking. I would also recommend looking at reclass for other reasons (simplify inventory management, create parametized classes as formula+pillar, etc)
12:31 luminous goodluck
12:31 viq joehh: sweet, thank you! :)
12:31 Couch luminous, thanks a lot
12:37 londo_ joined #salt
12:39 JlRd joined #salt
12:41 rjc joined #salt
12:46 scottpgallagher joined #salt
12:46 multani joined #salt
12:47 __gotcha joined #salt
12:48 johtso_ joined #salt
12:50 joehh viq: no worries
12:50 viq Time for lunch, and then to roll this out ;)
12:54 glyf joined #salt
12:56 z0rc left #salt
12:57 joehh viq: I would recommend testing on a limited set first...
12:58 joehh although it has taken approximately 5 months for the rc process and saltstack have drastically increased their QA
12:58 joehh it is still "new"...
12:59 thawes joined #salt
13:04 picker joined #salt
13:05 ajolo joined #salt
13:05 ecdhe joined #salt
13:05 multani Is there any way to speed up the integration tests? I'd like to run exactly *one* test (that I'm working on), and it takes about 2 minutes to launch everything and just a blink to execute my test :(
13:06 multani (I'm using -n of course)
13:08 ze- multani: what command are you launching/where ?
13:08 clinta joined #salt
13:09 multani ze-: I'm launching './tests/runtests.py --no-clean -n integration.states.debconfmod' from the root of the repository
13:10 ze- mmm... not familiar with such. but as it seems to be "loading" alot, it might be related to a problem i found on my servers.
13:11 ze- here, at times, we have salt-run -d foo that already takes a long time, instead of being instant.
13:11 ze- it is linked to having too many files in cache dir (old jobs), and i have to clean them by hand.
13:11 ze- maybe strace your test, and check where it lose times. (maybe there also are history files that are loaded, and that you could purge)
13:12 multani ze-: hum, I've started to run the tests from a brand new clone yesterday, and it was slow all the time
13:12 multani ze-: but there are 'integration' tests, I think they are expected to run slowly, but less slow would be great :)
13:13 spielberg joined #salt
13:14 clinta joined #salt
13:16 styly joined #salt
13:18 scottpgallagher joined #salt
13:19 tafa2 joined #salt
13:23 thayne joined #salt
13:25 aristedes joined #salt
13:26 jeddi joined #salt
13:27 __gotcha joined #salt
13:32 AirOnSkin joined #salt
13:32 claytron joined #salt
13:33 AirOnSkin Where do I start debugging if my Master won't show the new minions with salt-key? There are already minions accepted by the master... this has worked before.
13:33 NV joined #salt
13:36 Reiner0316 joined #salt
13:36 Reiner0316 Hello,
13:37 AirOnSkin Nevermind, I got it...
13:37 AirOnSkin networking issue
13:37 Hell_Fire joined #salt
13:38 Reiner0316 short question - is it possible to set pillar.get fallback values with another pillar.get variable? As I could tested it it seems that only strings/integegers and dictionary/lists can be setup but no another pillar.get function ?
13:39 Reiner0316 what I like to have is somethin like this: {%- set newrelic_app_name    = salt['pillar.get']('tomcat:newrelic:app_name', pillar.get('newrelic:app_name', 'My Application')) %}   (checking server role based variablen, then project based).    If server role variable exists it works but not with 2nd project (environ) based variable - then he used automatically last fallback definition.
13:41 aristedes joined #salt
13:41 __gotcha1 joined #salt
13:41 spielberg joined #salt
13:43 DaveQB joined #salt
13:44 jeddi joined #salt
13:46 pduersteler joined #salt
13:46 otter768 joined #salt
13:46 GnuLxUsr joined #salt
13:51 miqui joined #salt
13:54 pduersteler hi all. quick question, I don't understand why I get an error here saying that it can't watch the package: https://gist.github.com/pduersteler/537084fb6de49cd79df0
13:54 pduersteler the package gets installed properly, why does salt not "find" it afterwards?
13:55 claytron_ joined #salt
13:58 Reiner0316 package id is zabbix.agent... so he uses this name if you don't use -name: zabbix-agent in -pkgs:
13:58 vbabiy joined #salt
13:59 Reiner0316 ... instead of -pkgs: and -zabbix-agent you should write -name: zabbix-agent
14:00 pduersteler Reiner0316: thx, let me try...
14:01 Reiner0316 (was sometimes also my problem ;))
14:01 __gotcha joined #salt
14:02 pduersteler Reiner0316: but when I specify - name: zabbix-agent in -pkgs then I get "The following packages failed to install/update: name=zabbix-agent." Do you have an example about how it should look like?
14:06 Reiner0316 ah, do you need to watch packages ? Debian e.g. automatically stops/starts service when updating... searching for example
14:07 pduersteler Reiner0316: ah, you're right. Still wondering though, as it does work for other packages. not sure if it has something to do that I am getting the package from a different repo
14:07 CeBe1 joined #salt
14:07 Reiner0316 i think i knew only -requre: -pkg: ... not -watch ...
14:08 Reiner0316 like e.g. https://github.com/saltstack-formulas/openssh-formula/blob/master/openssh/init.sls
14:08 mapu joined #salt
14:08 Fbraad joined #salt
14:09 pduersteler yes, but watch has a different purpose AFAIK. I'm leaving the pkg line away for now
14:09 racooper joined #salt
14:09 Reiner0316 perhaps you want a "monitor" like apache uses
14:09 Reiner0316 https://github.com/saltstack-formulas/apache-formula/blob/master/apache/init.sls
14:10 pduersteler actually not, no. I got that watch from 0815 basic examples from the saltstack docs
14:10 Reiner0316 https://github.com/saltstack-formulas/apache-formula/blob/master/apache/mod_proxy.sls  could be triggered with -watch_in
14:10 Reiner0316 ok ;)
14:10 pduersteler yes, but I'm not in a subpackage wanting to "modify" a parent package ;)
14:17 nitti joined #salt
14:18 perfectsine joined #salt
14:19 ITChap joined #salt
14:20 summers joined #salt
14:21 Guest71585 joined #salt
14:24 __gotcha joined #salt
14:24 jerematic_work joined #salt
14:25 bhosmer joined #salt
14:27 alxbse joined #salt
14:28 flyboy82 I'm trying to use mine.get in one of my orchestration sls'es... but mine doesn't return anything. I'm rendering the sls in debug and the supposed dictionary's length is 0
14:29 flyboy82 any ideas how to use m ine.get in orchestrate runner? btw, mine.get with salt-run from master's cli works fine
14:29 funzo joined #salt
14:33 jaimed joined #salt
14:34 che-arne joined #salt
14:37 masterkorp Hello
14:37 SkyRocknRoll__ joined #salt
14:38 NV joined #salt
14:46 cpowell joined #salt
14:47 flyboy82 guys? anyone? I was waiting almost a month for mine to work on the master and still no luck in orchestrate runner...
14:49 aparsons joined #salt
14:49 bhosmer joined #salt
14:50 aparsons_ joined #salt
14:50 thawes_ joined #salt
14:53 glyf joined #salt
14:53 ndrei_ joined #salt
14:54 abe_music joined #salt
14:54 CSa_ :w
14:55 catpig joined #salt
14:58 ndrei joined #salt
14:59 ndrei_ joined #salt
14:59 ndrei__ joined #salt
14:59 perfectsine joined #salt
14:59 TheThing joined #salt
15:00 viq joined #salt
15:03 gngsk joined #salt
15:03 thedodd joined #salt
15:04 vbabiy joined #salt
15:05 FarrisG joined #salt
15:05 housl joined #salt
15:08 flyboy82 ok, other question. how do I call a runner from an orchestrate sls?
15:08 _JZ_ joined #salt
15:09 spielberg joined #salt
15:10 perfectsine joined #salt
15:11 StDiluted joined #salt
15:15 danburbridge joined #salt
15:17 felskrone joined #salt
15:18 thawes joined #salt
15:20 Guest71585 joined #salt
15:27 mohae joined #salt
15:29 smcquay joined #salt
15:31 Hell_Fire_ joined #salt
15:31 ekristen joined #salt
15:32 ekristen I’ve got a problem, for some reason one of my states on a minion cannot successfully get publish data from the master, but the search works from the master
15:32 ekristen so I can’t seem to figure out what the problem is or where the problem lies
15:33 bhosmer joined #salt
15:33 pr_wilson joined #salt
15:33 martino joined #salt
15:33 martino howdy
15:33 styly_ joined #salt
15:33 jngd joined #salt
15:34 martino question, can a salt master implement the states that it mandates to the minion to itself?
15:34 perfectsine joined #salt
15:34 flyboy82 left #salt
15:35 fredvd joined #salt
15:35 martino I have a redis state that configures redis on my minions, well realized that I need to install redis on my master now- so I was wondering if the master can call the state itself and install it
15:36 ekristen martino: just install the minion on your master
15:36 slav0nic is here any solution for https://github.com/saltstack/salt/issues/17313 ?
15:36 iggy Reiner0316: you could try 2 set lines: {% set nrapp = pillar.get %}\n{% set nrsapp = pillar.get('foo', nrapp) %}
15:37 iggy that's terrible shorthand btw, but should tell you what I'm talking about
15:37 Reiner0316 3 {%- set newrelic_app_name    = salt['pillar.get']('tomcat:newrelic:app_name', 'My Application') %}
15:37 Reiner0316 4 {%- if newrelic_app_name == 'My Application' %}
15:37 Reiner0316 5 {%-   set newrelic_app_name = salt['pillar.get']('newrelic:app_name', 'My Application') %}
15:37 Reiner0316 6 {%- endif %}
15:37 Reiner0316 my workaround created some minutes before ^^
15:38 iggy mine is shorter
15:38 iggy but yeah, I don't know exactly what you're trying to do, so it might not be exactly correct
15:38 catpig joined #salt
15:38 Reiner0316 yes, thx i take a look ... i had first also 2 lines but i like it clearly written
15:39 iggy martino: there's nothing that says a master can't also have a minion running on it... I think it's quite common in fact (and at least the salt-formula expects it)
15:42 _mel_ joined #salt
15:43 aparsons joined #salt
15:45 Reiner0316 iggy: question is what is the result of 1st line ... pillar.get without parameters "drops" the whole pillar tree and not only 1 variable ?
15:47 glyf joined #salt
15:47 otter768 joined #salt
15:49 iggy you mean in mine? yeah... I said that was terrible shorthand
15:49 iggy it was just to put you on the right track
15:52 bigl0af joined #salt
15:53 Reiner0316 ok... thx I think there is somehow a nice "howto"/background tutorial missing which helps better understanding jinja/saltstack rendering. I often have the problem that it didn't work as expected with knowlege of perl, bash and other non-python/ruby based stack... and some examples in saltstack docu weren't working because of typos/changes ^^
15:55 iggy if you see bugs, report them
15:55 iggy and I felt like the docs were pretty good, but then I come from a django background
15:59 thayne joined #salt
15:59 Reiner0316 yes, already notified (support@ ... or is there also a github issue tracker for docu ? - not info found for it)
16:00 conan_the_destro joined #salt
16:05 iggy the docs are in the main repo (usually inline with the code)... so use the main repo issue tracker
16:05 smcquay_ joined #salt
16:05 Reiner0316 ok
16:06 spielberg joined #salt
16:06 smcquay_ joined #salt
16:07 ekristen basepi: I’ve got a problem with mine.get and publish.publish, I’m still on rc5
16:07 ekristen doing the search from the master against the minion with the compound search works
16:07 Reiner0316 you have peers allowed in master config?
16:07 ekristen but when my state runs it on the minion, it can’t get a result
16:08 Reiner0316 and pushed the data to master on each minion? that was my problem last time
16:08 Reiner0316 # http://stackoverflow.com/questions/17158665/how-to-get-a-list-of-all-salt-minions-in-a-template
16:08 Reiner0316 ## running scripts on other nodes:
16:08 Reiner0316 ## http://stackoverflow.com/questions/20678176/how-do-you-use-the-publish-module-in-your-own-module
16:08 iggy I also had problems with getting the peer stuff enabled for publish to work
16:09 ekristen peer: .*: - network.* is set on my master
16:09 ekristen and running it from the master, like sudo salt ‘minion’ mine.get ‘compoun_search’ network.interfaces compound works
16:09 ekristen but doing it in a state in the minion fails
16:09 ekristen its been working for a year without issue
16:09 ekristen not it fails
16:09 ekristen now*
16:10 ekristen keep getting element 0 not available in jinja template
16:10 Reiner0316 try:   salt '*' mine.send network
16:10 iggy do you have missing .p files in the cache?
16:10 Reiner0316 and then it should work
16:11 Reiner0316 thats e.g. one of the not clear docu parts of saltstack ... the mine is not automatically filled...
16:12 ekristen like I said it works from the master with the salt command
16:12 ekristen just not from the minion in a state
16:12 ekristen so I’m perplexed
16:12 ekristen but right now a deployment is on hold cause it isn’t working
16:12 Reiner0316 yes, was in my setup also the case... after the mine.send i could access the variables on minions, too
16:13 ekristen kk
16:13 wnkz joined #salt
16:13 ekristen working on that now Reiner0316, we’ll see how it goes
16:14 iggy if that doesn't help, check out https://github.com/saltstack/salt/issues/15673
16:14 Reiner0316 hope it helps... for "generic"... the google matches shows also always the "public" docu which isn't very helpful if i implement features of 2014.7.x which is development version but package is still 2014.1.x ^^... (and there are several options which forgotten info "implemented in 2014.7.x")
16:15 smcquay joined #salt
16:16 ekristen still fails
16:16 ekristen wtf!!!!!!!!!!!!!!!
16:18 Reiner0316 the docu itself is always a problem child ... so its already very nice but has some bigger footsteps which creates blog entries like "one-week-of-salt-frustrations-and-reflections" (several time found these when searching for help) ^^
16:18 hasues joined #salt
16:20 vejdmn joined #salt
16:22 __gotcha joined #salt
16:22 igorwidl joined #salt
16:22 Reiner0316 ekristen: sorry, then I have no idea ... I used later publish.publish for my small setup
16:22 ekristen Reiner0316: publish publish keeps returning weird results
16:22 ekristen when i test it from the master
16:22 RedundancyD joined #salt
16:23 ekristen sometimes I’ll get 2 hosts sometimes I’ll get 5 or 7
16:23 ekristen when I have 10 that match the compound search
16:24 SheetiS joined #salt
16:25 Reiner0316 yes, that was my case, too... that sounds like the bug url iggy posted before... I first thought that my micro AWS instance was too small / limited so that this problem occurred.
16:26 racooper joined #salt
16:26 Reiner0316 in http://irclog.perlgeek.de/salt/2014-03-13/text I found nice publish.publish examples... again a docu problem... in which cases I need which formatters   (in this function/parameter: expr_form=comound")
16:30 iggy it would be nice if you would actually give examples of how the docs are bad instead of just continually saying they are bad
16:31 kermit joined #salt
16:31 geekatcmu Next you'll start asking for people to change from whining to constructive critism, to active participation.  WHERE WILL IT END?!?
16:32 StDiluted joined #salt
16:32 linjan joined #salt
16:32 tomspur istram, did you report the bug for zeromq4?
16:33 btorch joined #salt
16:33 jimklo joined #salt
16:33 tmh1999 joined #salt
16:33 tligda joined #salt
16:33 Reiner0316 was that not a practicable example ? howto use the formatter ? I had made try&error with on other pages found snippets with formatters because its not clear which variable needs which formatter to get a (printable) result back
16:34 rm_jorge joined #salt
16:34 Reiner0316 if Its clear I can also open an issue... but where to start ? main github repo with documentation url as reference ?
16:35 btorch hi, I have a state where I have common/packaging/init.sls common/packaging/files.sls common/packaging/repos.sls . the init.sls is basically just including those two others . When I use a require: sls: common.packaging on another state it seems to always fail
16:35 Reiner0316 btorch: do you have also included pillars with same name ? That was often my problem when getting such errors
16:36 btorch pillars with the same name as what ?
16:36 eunuchsocket joined #salt
16:37 btorch all I get when running it with test=true is Comment:   The following requisites were not found require: sls: common.packaging
16:41 jonatas_oliveira joined #salt
16:42 Ozack joined #salt
16:46 Reiner0316 btorch: I think in your case pillar common.packaging (minimum common)
16:47 btorch Reiner0316: I have no pillar of such sort
16:48 vejdmn joined #salt
16:48 mr_chris Still struggling with the MySQL grant state being broken. Filed a bug report for it. https://github.com/saltstack/salt/issues/18260
16:48 mr_chris Attempting to fix it.
16:48 iggy btorch: gist the state's causing the problem
16:49 Reiner0316 yes, thats the problem... e.g. my failure message was: "Specified SLS 'newrelic' in environment 'dev' is not available on the salt master"
16:49 aparsons joined #salt
16:49 Reiner0316 after cp -p pillar/test/newrelic pillar/dev/   it was running in my setup test
16:50 btorch Reiner0316: that's a different issue though
16:50 btorch iggy: ok
16:50 Reiner0316 ok
16:51 possibilities joined #salt
16:54 Guest71585 joined #salt
16:57 btorch iggy: http://pastebin.com/6FYCJsi3
16:58 slafs joined #salt
16:59 perfectsine joined #salt
17:01 troyready joined #salt
17:01 iggy I wonder if it's not treating common/packaging/init.sls the way we expect because there aren't actually any states in it
17:01 jalbretsen joined #salt
17:01 iggy can you put a dummy state in there like: /mnt:\n  file.directory
17:02 Reiner0316 require:sls ... is supported at 2014.7.x only - you have this version running ?
17:03 iggy fwiw, that all looks correct
17:03 Ahlee I don't think that's correct Reiner0316.  Pretty sure I use require: sls on 0.17.5
17:04 ajolo joined #salt
17:04 Reiner0316 ah sorry. http://docs.saltstack.com/en/latest/ref/states/requisites.html#require-an-entire-sls-file ... that was another require function i was thinking about
17:05 jimklo joined #salt
17:05 iggy it was added in 0.16
17:05 Reiner0316 yaml... perhaps you must indent the require like the pkgs line (2 more spaces)
17:07 berserk joined #salt
17:09 btorch iggy: sure
17:10 btorch iggy: ha! that was it. I added a file.absent state in there of a non-existing file and it works now
17:11 TTimo joined #salt
17:11 TheThing joined #salt
17:11 iggy btorch: I would actually (first look for, and then) file a bug on that... it should either be documented somewhere or work as (both of us obviously) expected
17:11 btorch iggy: so I guess I need to add something dummy on init.sls files that have just includes
17:12 iggy but for now, yeah, you have a workaround
17:12 btorch iggy: thanks
17:12 btorch iggy: I'm using 0.17.5 though so not sure if that happens on the latest salt ... will test it out later
17:13 RedundancyD iggy: one example where the docs are not ‘Bad” but more confusing is the quick start and walk-thoughs that not really linar but more parrallel and jump around alot.
17:13 CeBe joined #salt
17:14 iggy RedundancyD: fwiw, those are about to undergo some signficant work (they just hired a doc writer from what I heard)
17:14 iggy but that's good stuff to hear
17:15 luminous does salt-vim support other renderers? or only jinja?
17:15 jonatas_oliveira joined #salt
17:16 zlhgo joined #salt
17:17 luminous whiteinge: does salt-vim support pyrenderer/etc or only jinja?
17:18 SheetiS luminous: it looks to be jinja-yaml only from what I can tell.
17:18 SheetiS I try and name anything that I use the py renderer on .py so that my python highlighting automatically takes over (and I know it was python without having to look inside first).
17:19 SheetiS of course I don't typically py render states but mostly underlying files.
17:19 SheetiS so I can get away with that.
17:20 luminous SheetiS: oh, will salt find the .py in file roots?
17:20 luminous I didn't even test that..
17:21 Reiner0316 I use in debian vim-nox: vim compiled with support for scripting with Lua, Perl, Python, Ruby, and Tcl but no GUI.
17:21 desposo joined #salt
17:22 Deevolution When using a for loop in Jinja over pillar data, how do I maintain the ordering of the pillar data (it appears to be reordering it alphabetically)?
17:23 Reiner0316 and if you need other extensions renderes as ruby, too: http://blog.arangamani.net/blog/2013/11/27/syntax-highlighting-in-vim-for-vagrantfile-gemfile-and-berksfile/
17:23 slafs left #salt
17:24 luminous Reiner0316: vim-nox has syntax highlighting for salt .sls and its various renderers?
17:24 Reiner0316 vim-nox: not salt but the written above
17:25 iggy Deevolution: python doesn't order dicts, lists, etc. consistently
17:25 SheetiS luminous: it won't.  That's why I said I do most of my py rendered stuff in managed files and not states (and therefore can get away with it).
17:25 Reiner0316 also vim-gnome e.g. (
17:26 Deevolution iggy: any way I can get it to order them as written?  Otherwise I'm in for some pain to get them to execute in an explicit order...
17:26 SheetiS luminous: you could change your vimrc to parse the shebang at the top of a py-rendered state though http://stackoverflow.com/questions/8890668/how-to-make-vim-detect-filetype-from-shebang-line
17:26 hal58th joined #salt
17:26 KyleG joined #salt
17:26 KyleG joined #salt
17:26 iggy Deevolution: afaik, you'd have to use a pillar_ext if you need strict ordering
17:27 zirpu left #salt
17:27 iggy yaml just gets read in as standard python types (which are unordered)
17:27 Deevolution Iggy: Okay.  Have to sort that one out via a different route.  Thanks.
17:28 luminous SheetiS: thanks for that last link, I was wondering if that was going to be my best route here.
17:28 iggy Deevolution: you might want to wait around or ask on the list to see if there's a convenient way of doing it
17:28 luminous Deevolution: I think reclass might keep ordering
17:30 smcquay joined #salt
17:30 alexr joined #salt
17:31 luminous SheetiS: I <3 vim, that was a 4 second fix
17:31 SheetiS Deevolution: if you use a list (or a list of dicts if you need dictionary style data) rather than a straight dict, it will maintain order
17:32 luminous whiteinge: for when you are around, I would appreciate a quick chat to see how to update salt-vim to include this fix more properly
17:34 SheetiS Deevolution: do you have an example of what you are trying to accomplish, and I can show how I would use a list to maintain order?
17:34 jimklo joined #salt
17:34 Deevolution luminous: I've played with reclass, but we can't use it currently (doesn't quite match how we're using pillar; we may patch it down the road to allow it's use).
17:35 Deevolution SheetiS: I'd considered that.
17:36 thedodd joined #salt
17:36 Deevolution SheetiS: I'm managing software installs from files that don't automatically handle dependencies.  This is why I'm trying to get the ordering more explicit.
17:37 istram tomspur: you still there? yes, I put down the BZ for 0mq
17:39 Reiner0316 could be dependent of the type of module... pkg.* in debian e.g. sorts automatically all packages... I can't install e.g. postfix before bsd-mailx (which installs exim4 as dependency automatically)
17:40 tomspur istram, so you are interested in an zeromq4 package, or do you just want to have the zeromq updated to 4 there too?
17:40 alanpearce joined #salt
17:41 carmony What is the command line option you can pass to salt for when you're running a highstate it will report which servers still haven't reported back yet?
17:41 Reiner0316 for files/software installs... perhaps you can setup file-requirements for each package so that the order must be the one you want ?
17:41 viq carmony: -v / --verbose
17:41 thayne joined #salt
17:41 iggy Deevolution: is there a reason you are using pillars for that? If you just used states, they are processed top-down already
17:42 Reiner0316 ekristen: because of cmdline option ... it helps to increase timeout -t 600 to get all instances reporting back
17:44 istram tomspur: I'd love to see the zeromq4 package in epel rather then in copr because the pkgs in copr aren't (and can't) be signed
17:44 Deevolution iggy: I have states reading the pillar data (and currently iterating over a list in the pillar data of packages and versions).
17:45 Deevolution iggy: It's cleaner and more readable (in my opinion) to have a single pillar that reads something like 'packagea: 1.2.3' and iterate those then to have loads of pkg statements in the salt code.
17:45 iggy I see where you're coming from
17:45 Deevolution It also means if we have people testing things, it's much easier to let them update their pillar data.
17:46 Deevolution iggy:  Yeah.  I'm going to see if I can get the dependencies down to a few 'stages' and setup separate pillars for that.
17:46 teebes joined #salt
17:47 iggy you could maybe use a number based system and use the |dictsort filter in the jinja
17:48 iggy 1: packagea\n- 1.2.3
17:48 CeBe1 joined #salt
17:48 otter768 joined #salt
17:48 iggy but I imagine if that was workable a normal list would work too
17:49 tomspur istram, I'd prefer not to maintain yet another package just for epel...
17:49 Deevolution iggy: I considered that also.  My concern is if you need to add something in the middle, you have to renumber all of them.
17:49 Deevolution iggy:  Workable, but kind of a PITA.
17:50 SheetiS can you use pillar['list'].index('item') in jinja?
17:50 SheetiS That's how I'd avoid knowing numbers in python of a list
17:50 SheetiS I can test if unknown
17:51 istram tomspur: well, not forcing you to :) just asked if that's possible. If not then I'll probably have to put up with 0mq v3 or push the v4 packages from copr to own repo and sign them.
17:52 tomspur istram, hmm isn't zeromq3 enough, why don't you use that?
17:52 tomspur python-zmq is build against it on epel
17:53 SheetiS Deevolution: this works for getting the index: https://bpaste.net/show/0801a5523191
17:54 SheetiS so you could extend that to whtaever you needed for the list
17:54 SheetiS or even macro it
17:55 istram tomspur: there have been reports that salt is (can be?) more stable on 0mq v4.
17:55 gngsk joined #salt
17:56 SheetiS So no renumbering required ;-)
17:56 tomspur istram, uh... :/
17:56 istram tomspur: not sure about the sources as my colleague responsible for salt came with it...
17:56 vejdmn joined #salt
17:56 istram tomspur: I'm just running the pkg show cause I know more about it.
17:57 capricorn_1 joined #salt
17:57 jdesilet joined #salt
17:59 hasues joined #salt
17:59 hasues left #salt
18:00 jdesilet I'm just getting into salt and looking to see if I can replace our puppet manifests with salt. One of the things we've done in puppet is have a manifest that drops a bunch of gpg keys onto our servers so they can be imported. Is it possible to build a salt stack where you can specify multiple files or at least contain it to one ID declaration?
18:00 gngsk joined #salt
18:00 carmony viq: : thanks... for some reason I thought --verbose was -log-level :P
18:02 forrest joined #salt
18:03 Nacmac joined #salt
18:03 Nacmac go away jdesliet. :)
18:04 alexr joined #salt
18:05 Deevolution SheetiS:  I'll take a look at that.
18:05 Deevolution Thanks!
18:06 ekristen did mine.get change
18:06 ekristen like the output format?
18:06 ekristen nm
18:06 murrdoc joined #salt
18:09 ekristen dang, this whole minion cache/salt key not matching and whatnot is really a big PITA basepi
18:09 ekristen the issue we thought we had fixed a week or so ago was causing issues again today
18:11 iggy jdesilet: you'll have to expound on what you're trying to do a little more I think, but as far as gpg keys go, you're going to want to keep those in pillars
18:12 dnai23 joined #salt
18:13 jdesilet iggy the file type isn't too important here as I'll have a need to do this for other stacks as well. What I'm hoping to be able to do is just reference one ID and then have a list of files be deployed.
18:14 jdesilet I haven't made it to pillars yet in the documentation. If you think that will solve this I'll start digging into more of the documentation there.
18:15 cheus joined #salt
18:17 jdesilet a better example might be when we're trying to push out multiple custom nagios checks. We'd like all the servers to have a group of additional checks and if it's possible I would like to only have to call the ID nagios ( in this example ) top.sls.
18:17 iggy pillars are just where you put data that you don't want _every_ minion to see
18:17 iggy there's nothing special about pillars that will help you solve... whatever it is you're trying to solve
18:19 Reiner0316 jdesilet: there is also an icinga2 formula which offers your requirements: https://github.com/saltstack-formulas/icinga2-formula
18:20 __TheDodd__ joined #salt
18:20 Reiner0316 when I checked it out it has a small bug that it installs icinga on each instance and not only the icinga server... should be checked before setup
18:21 iggy okay, so what it sounds like, you could target the pillars containing the gpg keys (various different ones at different hosts), then have a single state that says {% for key in salt['pillar.get']('gpg:keys', []) %}  file.managed:  - contents: {{ key }} {% endif%}
18:21 perfectsine joined #salt
18:21 gngsk joined #salt
18:23 JohanR joined #salt
18:25 styly joined #salt
18:25 Reiner0316 what I am also searching for "weeks" (several days now from time to time again"): How can I reference in jinja template the actual id of for loop of sls file ? {{ id }} gave back a nice SaltRenderError: Jinja variable 'id' is undefined; line 7
18:25 otter768 joined #salt
18:25 ajolo joined #salt
18:27 perfectsine joined #salt
18:27 Reiner0316 SLS content: {% for id, data in salt['pillar.get']('tomcat:instances', {}).items() %}
18:27 harkx joined #salt
18:27 druonysus joined #salt
18:28 Reiner0316 I tried it already with a context: line id: {{ id }}   but that didn't help either
18:33 shaggy_surfer joined #salt
18:38 _prime_ joined #salt
18:39 wincus joined #salt
18:40 Mso150 joined #salt
18:42 eliasp yeehaw… 2014.7.0 on Ubuntu! thanks joehh
18:43 jdesilet iggy: I'll give you an example of what I'm trying to do and maybe you can tell me if it's even possible or what's the better way to do it.
18:43 jdesilet derp:
18:43 jdesilet /var/tmp/10:
18:43 jdesilet file:
18:43 jdesilet - managed
18:43 jdesilet - name: /var/tmp/10
18:43 jdesilet - source: salt://derp/10
18:43 jdesilet /var/tmp/11:
18:43 jdesilet file:
18:43 jdesilet - managed
18:43 eliasp jdesilet: please use a nopaste service!
18:43 jdesilet - name: /var/tmp/11
18:43 jdesilet - source: salt://derp/11
18:43 luminous Deevolution: how is reclass not able to flow with your use of pillar?
18:43 spielberg joined #salt
18:43 rlarkin I'm building wheezy debs for python-botocore and some other python deps...I hope to get it running on wheezy but I have found a debianized source package for orderddict yet.
18:45 MugginsM joined #salt
18:46 jdesilet Better example http://pastebin.com/XGLgF5PT
18:46 basepi ekristen: dangit
18:46 iggy jdesilet: for that (hopefully) very contrived situation, you'd do {% for i in [10, 11] %}\n  /var/tmp/{{ i }}:\n    file.managed:\n      source: salt://derp/{{ i }}\n{% endfor %}
18:47 Topic for #salt is now Welcome to #salt | SaltConf 2015 Call for Speakers is open! http://saltconf.com/call-for-speakers/ | 2014.7.0 is the latest | Help us test the 2014.7 RC! http://bit.ly/salt-rc | 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/
18:47 Topic for #salt is now Welcome to #salt | SaltConf 2015 Call for Speakers is open! http://saltconf.com/call-for-speakers/ | 2014.7.0 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/
18:47 iggy the "- name: ..." is redundant if you are using IDs like that
18:48 jdesilet ok, I'm only 2 days into playing with this
18:48 teebes joined #salt
18:48 jdesilet still a lot to learn about syntax and whatnot :)
18:49 iggy jdesilet: yeah, it's cool... I'd read the jinja docs (separately from what's in the salt docs)
18:49 iggy jinja.pocoo.org or something
18:49 jdesilet sounds good, thank you
18:50 schristensen joined #salt
18:50 iggy the more you actually understand about jinja strengths and weaknessess the better your states/etc will be
18:50 aw110f joined #salt
18:51 whiteinge luminous: it tries to detect the correct salt renderer: https://github.com/saltstack/salt-vim/blob/master/ftdetect/sls.vim
18:51 bfoxwell joined #salt
18:51 luminous whiteinge: interesting, my #!py was not seen as python
18:52 luminous whiteinge: I had to add a modified snippet from SO
18:52 iggy fwiw, my first salt deployment used almost no jinja... I went back and rewrote pretty much the entire thing
18:52 jdesilet hehe
18:52 iggy formulas are good to look at for examples too
18:52 iggy well, some of them are
18:52 whiteinge luminous: what was it?
18:52 luminous jinja is awesome, though I'm finding pyrenderer so much better for some stuff
18:52 perfectsine joined #salt
18:53 luminous whiteinge: if getline(1) =~# '^#!py\>'   then used setfiletype. a bit more than this, but here is the magic line. I can look up the link too if you want
18:56 whiteinge luminous: hm. i have used this with a #!py shebang before. what version of Vim? what fileformat is that file?
18:57 Ryan_Lane joined #salt
18:57 shaggy_surfer joined #salt
18:58 toddnni joined #salt
18:58 Reiner030 joined #salt
18:59 luminous whiteinge: the file format is .sls, checking version..
18:59 cwyse joined #salt
18:59 luminous whiteinge: VIM - Vi IMproved 7.3 (2010 Aug 15, compiled May  4 2012 04:25:35)
19:03 meylor joined #salt
19:04 meylor left #salt
19:05 diegows joined #salt
19:05 cpowell_ joined #salt
19:05 thedodd joined #salt
19:06 druonysuse joined #salt
19:06 _prime_ joined #salt
19:08 glyf joined #salt
19:08 mohae joined #salt
19:09 perfectsine joined #salt
19:09 Reiner030 joined #salt
19:11 spielberg joined #salt
19:12 skyler Is there a good way to manage apparmor.d files with salt? I mean is there a better way than file.managed?
19:13 Cidan are you guys aware that ppa management is broken with this version of salt?
19:13 Cidan what a mess
19:13 perfectsine joined #salt
19:14 perfectsine joined #salt
19:15 Ryan_Lane Cidan: in what way is it broken?
19:15 Cidan new ppa's don't get added
19:15 Ryan_Lane Cidan: I'm using it successfully (and have been since before the RCs)
19:17 Cidan and adding ppa's works for you?
19:17 Reiner030 could it be a caching / proxy problem ? one of our devs had it with some others repos locally...
19:17 Cidan they show up on the server?
19:17 Cidan because I get a success in salt, but it's not actually added
19:18 aparsons joined #salt
19:18 Cidan the repo isn't added at all
19:18 Cidan this is new with 2014.7
19:19 Cidan starting to lose my religion with salt, I need to learn to not upgrade
19:20 Ryan_Lane hm. yeah, repo gets added for me
19:20 Ryan_Lane what's your state look like?
19:20 Reiner030 that was my problem with 2014.1.11 ... it was buggy when i installed fresh servers... so I need to install them by git tag
19:21 Cidan nodejs-repo:
19:21 Cidan pkgrepo.managed:
19:21 Cidan - ppa: chris-lea/node.js
19:21 Deevolution luminous: (First sorry for delay, missed your message) Reclass interacts with Salt environments (and how we're storing pillar data in source control) in such a way as we'd have to have a single master list of nodes.  Since I want my development branch(s) to allow developers to merge into them freely while still preventing changes to prod branch(s), having a single set of nodes is not going to work.
19:21 Ryan_Lane I'm doing a code review on something right now that's using pkgrepo.managed in 2014.7
19:21 Reiner030 btw... if there is someone from saltstack... is it possible to hold not only latest but also previous release of a version please ?
19:21 Ryan_Lane Cidan: I think that syntax is wrong
19:22 Cidan it's not
19:22 Cidan it hasn't changed in months
19:22 Ryan_Lane - name: ppa:chris-lea/node.js
19:22 Cidan unless salt changed in 2014.7
19:22 CeBe1 joined #salt
19:22 Reiner030 because of the ppa: ... must it perhaps included in '' / "" ?
19:22 cholcombe973 salt: can i call saltutil.refresh_pillar in my state file?
19:22 Ryan_Lane ah. seems that's true
19:23 aqua^mac joined #salt
19:23 Cidan downgrading. sigh.
19:25 span joined #salt
19:25 teebes joined #salt
19:25 Ryan_Lane Cidan: make sure to open an issue in github
19:25 Cidan am
19:25 eunuchsocket joined #salt
19:27 Mso150 joined #salt
19:29 istram joined #salt
19:33 ldlework Does anyone have any recommendations for installing the requirements of custom salt modules?
19:36 spielberg joined #salt
19:37 rheaume joined #salt
19:40 cholcombe973 salt: can my reactor call multiple state files or just one?
19:40 cholcombe973 it seems like it calls the first one and then stops
19:41 iggy ldlework: be more specific
19:41 ldlework iggy: custom module depends on PackageX
19:41 ldlework how do I ensure that new nodes have PackageX so my custom module works?
19:41 iggy write a state?
19:42 skyler I have a pull request that I need to update for a formula. If I make the necessary changes and then squash them into a single commit, will that cause any problems?
19:43 cholcombe973 ok i'm stuck.  i can't seem to get salt to force a ext_pillar refresh in the middle of a state file change
19:43 cholcombe973 anyone know how to do that?
19:43 iggy skyler: as long as nobody has the intermediate commits, it should be fine
19:44 skyler iggy: Even if I have already issued the pull request (but nobody has used it)?
19:45 gngsk joined #salt
19:45 aparsons joined #salt
19:48 iggy just close the existing one, force push, open a new PR (assuming that I'm understanding you correctly)
19:48 mapu joined #salt
19:49 chiui joined #salt
19:50 possibilities joined #salt
19:50 whiteinge luminous: not sure about that one. My regex-fu is not up to the task of identifying that closing bracket (end-of-word?)
19:51 whiteinge The way it's written upstream should match py, pydsl, pyobjects so I'm not sure we should change it.
19:53 whiteinge Thoughts?
19:56 sroegner joined #salt
19:57 alexr joined #salt
20:00 cholcombe973 is there a way to force salt to rebuild a file on every run?
20:00 thayne joined #salt
20:01 alanpearce joined #salt
20:03 alanpearce_ joined #salt
20:05 tempspace Has anybody seen this going on with 2014.7? from urllib3.util import Timeout as TimeoutSauce
20:07 luminous whiteinge: do you have a code reference/link I can look at?
20:08 nitti joined #salt
20:09 druonysus joined #salt
20:09 druonysus joined #salt
20:09 thedodd joined #salt
20:13 aparsons joined #salt
20:13 iggy tempspace: search the issue tracker
20:13 tempspace I saw an old issue that was closed, just checking to see if anybody here has seen anything that may not have created an issue yet...
20:15 iggy well... I could swear I saw something today or yesterday
20:15 iggy I wasn't just being an asshole telling you to search the bug tracker
20:15 LotR joined #salt
20:15 iggy maybe it was somewhere else... one sec
20:15 izibi joined #salt
20:16 izibi hi, how can i override the default implementation of the pkg module without patching salt?
20:17 eliasp izibi: place your custom module in _modules/pkg.py
20:17 eliasp izibi: if your improvements could serve all other Salt users too, please create a PR to get your changes upstream into 'develop'
20:17 jonatas_oliveira joined #salt
20:18 iggy tempspace: what distro?
20:19 TheoSLC joined #salt
20:19 tempspace Debian Wheezy
20:20 iggy yeah, I swear I saw something about this
20:20 shaggy_surfer1 joined #salt
20:21 iggy tempspace: what's the actual command you are running to see this error?
20:21 tempspace salt-minion -l debug
20:21 tempspace it also wouldn't start my master
20:21 tempspace some of my machines are fine, others aren't
20:22 iggy okay... not the same thing I was thinking then (it was specific to salt-ssh)
20:22 tempspace not seeing any differences in pip or dpkg yet
20:22 iggy Open a new issue I guess
20:22 izibi eliasp: tried that. i'm only able to add modules this way, not override existing ones
20:23 eliasp izibi: this should work… I've used this approach for the last ~14 months or so… check also: http://docs.saltstack.com/en/latest/ref/file_server/dynamic-modules.html
20:23 madduck Deevolution: it would be great if you could take the time to write a message to the reclass list with your exact use case and how reclass does not cater for it. Maybe we can make it work!
20:24 bhosmer joined #salt
20:24 eliasp izibi: if you can't get this working, that's IMHO a bug
20:24 eliasp izibi: modules in _{modules,grains,renderers,returners,states} should always override the builtin ones
20:25 gamingrobot_ joined #salt
20:25 MTecknology dang... salt gets chatty with 590 minions
20:26 LotR with gitfs, is there a way to say "each of the subdirs of this dir is a repo" instead of having to list all of them seperately?
20:26 CeBe1 joined #salt
20:26 eliasp LotR: what's your idea of a "repo"?
20:27 eliasp LotR: just a regular git repository?
20:29 eliasp LotR: the GitFS configuration just defines a list of remotes… per remote you can specify some specific config params: http://docs.saltstack.com/en/latest/topics/tutorials/gitfs.html#per-remote-configuration-parameters
20:29 LotR eliasp: yes. I clone all the formula repos I want to use locally and currently use file: urls in gitfs_remotes. but since I clone them all in the same parent dir, I thought it would make sense to be able to specify that
20:29 linjan joined #salt
20:30 rap424 joined #salt
20:30 eliasp LotR: well, it's possible to define 1 root per repo (http://docs.saltstack.com/en/latest/topics/tutorials/gitfs.html#serving-from-a-subdirectory), but I don't think it's possible to define more than 1
20:30 murrdoc joined #salt
20:30 iggy I just symlink all my formulas into my main state dir on my dev salt server
20:30 rojem joined #salt
20:31 LotR eliasp: I don't want to specify a directory inside the repos, I want to be able to say "this place contains a bunch of repos; use them all"
20:31 eliasp the filesystem layout of formulas is IMHO a bit unfortunate… I can clearly see what's your problem LotR
20:31 eliasp LotR: I just don't have a "proper" solution to it…
20:32 LotR the only 'problem' I have is that I need to change the master config each time I add a formula. it works just fine, it's just not quite optimal
20:33 eliasp LotR: yeah… that's why I have a repo containing manual copies of formulas here where I also drop the root directory of each formula repo to get a flatter hierarchy
20:34 eliasp LotR: this also gets rid of the issue to trust "random" formula repo updates
20:34 eliasp but it's still a bit ugly
20:35 alexr joined #salt
20:35 LotR eliasp: that's why I clone them, which is recommended in the docs. I would have to do a git pull before it updates
20:35 pppingme joined #salt
20:35 eliasp yeah, right
20:36 vejdmn joined #salt
20:39 jimklo joined #salt
20:41 LotR eliasp: I'm a bit confused why you want a flatter hierarchy than you already get. all the ones I use have init.sls, which you can leave out completely, so you just put e.g. '- users' in your top.sls to use the users-formula. I don't see how it can be more flat than that
20:43 eliasp LotR: cloning e.g. users-formula into a formula collection repository would result in having users-formula/users/init.sls, so I'd have to use 'users-formula.users' in my top.sls … instead I prefer to simply assign 'users'
20:44 eliasp LotR: so I get rid of the files in the toplevel directory (mostly LICENSE, README.{md,rst}, pillar.example) and just merge all this into my repository's root directory…
20:45 eliasp s/all this/the formula's directory/g
20:46 LotR eliasp: but then you lose the benefit of it being a repo, don't you? or how do you pull in updates you do want?
20:46 eliasp LotR: yeah, that's right… I don't use formulas much here (yet), so this worked fine for me so far… it probably won't in the future as this approach doesn't really scale
20:48 TheoSLC Any SaltStack employees on?
20:49 eliasp TheoSLC: they're all busy being awesome… :)
20:53 giantlock joined #salt
20:53 TheoSLC eliasp: Do you think SaltStack would reject a speaker submission that was about NOT using salt-cloud? (chuckle)
20:53 alexr joined #salt
20:54 murrdoc i ll watch that
20:54 aparsons joined #salt
20:54 TheoSLC I'm integrating SaltStack with ManageIQ due to the self-service features.  I think it's useful.
20:54 murrdoc yeah man thats a good talk
20:55 murrdoc send it over
20:55 murrdoc <— not an employee obviously
20:55 TheoSLC okay.  you've convinced me.  I'll submit it. :).. probably will be rejected . lol
20:56 murrdoc you can always practice at the local saltstack user group
20:57 hal58th My dev guy made a script that will install minions and configure them to talk to a master that you specify and then run highstate. You run the script from master. All you need is a user that has sudo access
20:57 eliasp hal58th: he was re-inventing the wheel… that's already done ;)
20:57 iggy I fully expect my talk to be declined, so we can commiserate together
20:58 hal58th eliasp: Well send it on over.
20:58 TOoSmOotH joined #salt
21:00 eliasp hal58th: you can either use salt-bootstrap (https://github.com/saltstack/salt-bootstrap) or salt-cloud's provisoning…
21:00 TheoSLC hal58th: I have a similar script that I use to pass into cloudinit.
21:00 delinquentme joined #salt
21:01 TheoSLC eliasp: here is the one I use https://gist.github.com/deuscapturus/8f18d28d1a1ccef6327c
21:01 Outlander joined #salt
21:01 perfectsine joined #salt
21:02 Ironhand joined #salt
21:03 TheoSLC eliasp: wow, the official bootstrap script is almost 5,000 lines long.
21:03 hal58th So salt-bootstrap installs a minion but it does not configure the master, nor run highstate. Which are the two things we really needed.
21:03 hal58th oo, take that back, you can seed master
21:03 Ahlee then if you really want to get fancy, have highstate tied to a reactor for minion connect
21:04 Ahlee (I don't actually recommend that, as minions can reconnect at inopportune times)
21:04 eliasp also have a look at: http://docs.saltstack.com/en/latest/ref/states/startup.html
21:04 badon joined #salt
21:04 murrdoc or configure a startup state
21:04 Ahlee eliasp: huh, didn't know about that.  Neat
21:05 Ahlee minion calls that via salt-call I'm presuming?
21:05 TheoSLC I like my short little script.  it seeds master config and grains, but also fits in nice small base64 package.
21:05 Mso150 joined #salt
21:06 druonysuse joined #salt
21:06 druonysuse joined #salt
21:06 UtahDave joined #salt
21:07 Ahlee during kickstart we install an RPM that contains our own configs to point to our master(s) and go from there
21:07 Ahlee no scripts installing stuff all willy-nilly
21:08 Deevolution madduck: I'll try to find time (and my notes from my eval of it) to do so.  I did take some time looking at the code and modifying it to allow it to work (really interested in the much better hierarchical key merging over the default pillar provider), and it didn't look too bad to get it working for us, but it got pushed back behind getting some other pieces of Salt implemented.
21:08 TheoSLC Ahlee: that's works.  but your stuck rebuilding an rpm to change those settings.
21:09 Ahlee TheoSLC: as it should be. :)
21:10 jimklo joined #salt
21:12 alexr joined #salt
21:12 TheoSLC Ahlee: True.  For us, with ManageIQ, the user needs the ability to pass all minion settings
21:13 Ahlee word
21:21 LotR hmm. how do you create a feature request issue in github?
21:23 hal58th https://github.com/saltstack/salt/issues Add an issue and specify it as a feature request
21:24 aquinas joined #salt
21:25 aquinas_ joined #salt
21:26 saravanans joined #salt
21:28 LotR hal58th: oh, you get to specify that after the first dialog? ok
21:33 Ryan_Lane joined #salt
21:35 aparsons joined #salt
21:35 kballou joined #salt
21:37 hal58th LotR: No you don't, I would just specify that in your description.
21:37 skyler LotR: The mods get to do the tagging.
21:38 LotR I see
21:39 iggy grr... select (or map) is not a valid filter
21:40 bhosmer joined #salt
21:40 unpaidbi1l someone who has used yamlex and reclass - what are your thoughts on each of them and which did you ultimately end up using?
21:41 madduck url for yamlex?
21:41 unpaidbi1l http://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.yamlex.html
21:42 glyf joined #salt
21:44 uber joined #salt
21:44 madduck unpaidbi1l: for a start, reclass (which I initially wrote) allows you to build hierarchies of classes and put the information where it belongs, and then it hides the implementation.
21:44 madduck and reclass can be used for other tools, not just salt.
21:44 uber joined #salt
21:45 uber joined #salt
21:45 unpaidbi1l those two points are why i am leaning towards reclass.  if it already supported chef it'd be a no brainer - that's what i really need it to work with
21:45 uber joined #salt
21:46 uber joined #salt
21:46 unpaidbi1l on the other hand, the only thing i was lacking before yamlex was that aggregation function.  however i think the node/class hierarchy makes more sense to me and is easier to read.  i also like that it keeps the pillar and state information closer together if that makes sense
21:46 wt joined #salt
21:46 uber joined #salt
21:47 madduck unpaidbi1l: right.
21:47 madduck unpaidbi1l: writing an adapter for chef should be fairly straight forward…
21:47 uber joined #salt
21:48 uber joined #salt
21:48 uber joined #salt
21:50 thayne joined #salt
21:50 unpaidbi1l just makes a little more work is all - not a big deal.
21:51 thawes joined #salt
21:52 CeBe1 joined #salt
21:54 __gotcha joined #salt
21:57 Marga28 joined #salt
21:57 Marga28 left #salt
21:58 spookah joined #salt
21:59 nitti joined #salt
22:02 Ryan_Lane joined #salt
22:02 alanpearce joined #salt
22:03 perfectsine joined #salt
22:05 adrianhannah joined #salt
22:09 glyf joined #salt
22:13 aparsons joined #salt
22:14 cads joined #salt
22:15 aparsons_ joined #salt
22:18 skyler_ joined #salt
22:21 aparsons joined #salt
22:29 cholcombe973 salt: anyone want to take a peek at: https://github.com/saltstack/salt/issues/18279
22:30 renoirb Hi all!
22:30 renoirb Glad to see Salt 2.14.7 out
22:30 renoirb I mean 2014.7
22:31 spookley joined #salt
22:31 Hell_Fire_ joined #salt
22:32 iggy cholcombe973: you can't run salt from within salt
22:32 cholcombe973 lol
22:32 cholcombe973 aww come on :D
22:32 whatapain hehe
22:32 whatapain quit trying to hack the matrix
22:32 cholcombe973 you can run docker in docker ;)
22:32 mick3y joined #salt
22:32 whatapain yo dawg
22:33 cholcombe973 iggy: so is there any way to force a pillar refresh?
22:33 SheetiS cholcombe973: but the docker inside the inside docker doesn't modify the outside docker to modify the inside docker.
22:33 Ryan_Lane cholcombe973: you should use refresh_modules: True
22:33 Ryan_Lane does that actually refresh pillars?
22:33 Ryan_Lane it may not
22:33 cholcombe973 yeah i was looking at the refresh_modules.  i'm not sure it it'll refresh pillar
22:33 Ryan_Lane it doesn't refresh grains either, which is annoying
22:33 cholcombe973 i could try it
22:33 cholcombe973 oh :(
22:33 alanpearce joined #salt
22:34 Ryan_Lane it refreshes the grain /modules/ which is worthless
22:34 SheetiS might have to do it as part of an orchestration to get it to refresh the pillar in that fashion
22:35 cholcombe973 SheetiS: i'll look into the orchestration.  my other thought was I could break my state up into several states and run them with different api endpoints
22:35 iggy cholcombe973: module.run saltutil.refresh_pillar
22:35 iggy but (afaik
22:35 iggy ) it'll only run the refresh on the previously targeted minions
22:36 cholcombe973 iggy: i'm not sure i'm following about the previously targeted minions bit
22:36 hal58th joined #salt
22:37 iggy well, you wouldn't run "salt '*' states.updatecephconf"
22:38 iggy so whatever you were actually targeting would be the only thing that would run the refresh
22:38 cholcombe973 iggy: i see
22:38 iggy you probably want to use orchestrate as you can then target each step individually that way
22:38 cholcombe973 iggy: that's fine.  That's all i really wanted.  The rest of the minons can refresh later
22:38 cholcombe973 ok. i'll read up on that
22:41 Ryan_Lane anyone know how I can do a git based pip install that uses raet instead of 0mq?
22:42 aparsons joined #salt
22:43 Ryan_Lane bleh. it looks like the egg in the git repo always specifies zmq and m2crypto
22:43 cholcombe973 iggy: can i call orchestrate from the reactor?
22:44 iggy unknown
22:44 cholcombe973 ok
22:45 aqua^mac joined #salt
22:46 ekristen joined #salt
22:46 renoirb Hi all, did somebody tested etcd with Ubuntu 14.04 w/ salt?
22:47 renoirb The documentation isn’t clear yet :(
22:47 linjan joined #salt
22:48 hal58th1 joined #salt
22:48 N-Mi_ joined #salt
22:48 N-Mi_ joined #salt
22:49 Outlander joined #salt
22:56 thayne joined #salt
22:59 Outlander joined #salt
23:02 murrdoc joined #salt
23:06 murrdoc joined #salt
23:08 stevednd joined #salt
23:11 hal58th joined #salt
23:14 pdayton joined #salt
23:21 skyler_ What is the best way to map some values for debian, redhat, with some of the debian values changed for ubuntu. In other words, I want to map some values for debian, some for redhat, then set change some values if the debian based system is ubuntu in particular.
23:22 iggy what's the colon lookup syntax called i.e. pillar.get('foo:bar')
23:23 iggy skyler_: a double merge with one merge being based on os instead of os_family
23:23 MugginsM joined #salt
23:24 skyler_ iggy: Thanks. Do you know of any formulas with an example of this?
23:24 iggy postgres formula
23:24 iggy sort of
23:24 iggy it actually keys off the release version (to have different settings for ubuntu 14.04)
23:24 iggy but same concept
23:24 rojem joined #salt
23:25 thayne joined #salt
23:25 skyler_ iggy: Yep, that is what I was looking for.
23:26 kermit joined #salt
23:27 shaggy_surfer joined #salt
23:28 Ryan_Lane there's so many freaking reasons it's impossible to use raet right now
23:28 iggy you know at this point it's probably slower than 0mq
23:29 Ryan_Lane I actually just want to avoid m2crypto
23:29 Ryan_Lane I don't even use a transport
23:32 aparsons_ joined #salt
23:35 OnTheRock joined #salt
23:38 chikawaztla joined #salt
23:44 linjan joined #salt
23:46 aparsons joined #salt
23:48 pentabular joined #salt
23:48 wnkz_ Hi, since I updated to 2014.7 ; on every highstate where mount.mounted is present I get "Remount would be forced because options changed" but mount options are already good ; do you know where can this come from ?
23:49 mpanetta joined #salt
23:50 mpanetta joined #salt
23:50 vbabiy joined #salt
23:51 TTimo joined #salt
23:51 murrdoc Ryan_Lane:  how do make all the servers run salt-call —local
23:52 murrdoc like you mentioned all your servers are masterless
23:52 murrdoc do you do an ssh loop to get to all the servers to run the call or just leave it in cron
23:53 Ryan_Lane cron
23:53 Ryan_Lane you could also run the minion and a schedler
23:53 Ryan_Lane scheduler*
23:53 skyler_ What do you do when you need a value from map.jinja in a configuration file? All of the examples I have seen use map.jinja only for sls files.
23:54 iggy skyler_: usually pass it as context to file.managed
23:56 skyler_ iggy: Thanks, the only thing that worries me about that is that it becomes unclear where the variable came from inside the actual file.
23:56 iggy yeah
23:56 iggy and it's one of the things we're working on standardizing in the -formulas
23:57 iggy so it still won't be perfectly obvious where exactly it came from, but at least there will be a standard path to check
23:58 skyler_ iggy: Perhaps we could pass the values in the form of a dict called 'context' or some sort, and then in the file you would say {{ context['variable'] }}. Then you would at least know where it came from.
23:58 iggy I want to say I've seen a config file import a map before
23:59 iggy but I can't think off the top of my head where I saw it
23:59 styly joined #salt

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