Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-11-26

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

All times shown according to UTC.

Time Nick Message
00:01 torment joined #salt
00:02 ggoZ joined #salt
00:02 bhosmer joined #salt
00:15 TheThing joined #salt
00:15 Gues480631 joined #salt
00:16 ajolo joined #salt
00:24 TyrfingMjolnir joined #salt
00:28 catpig joined #salt
00:32 HACKING-FACEBOOK joined #salt
00:35 shaggy_surfer joined #salt
00:39 ysiad joined #salt
00:42 Pixionus joined #salt
00:42 bfraser joined #salt
00:49 Ryan_Lane joined #salt
00:49 flebel joined #salt
00:52 beneggett joined #salt
00:57 aqua^mac joined #salt
01:00 TyrfingMjolnir joined #salt
01:01 juice joined #salt
01:01 span joined #salt
01:17 pdayton joined #salt
01:23 ysiad joined #salt
01:23 jimklo joined #salt
01:27 TK_ joined #salt
01:30 thayne joined #salt
01:31 asdf___ joined #salt
01:32 jswanson_ joined #salt
01:33 TheThing joined #salt
01:34 otter768 joined #salt
01:36 atbell joined #salt
01:38 atbell joined #salt
01:43 catpig joined #salt
01:48 jimklo joined #salt
01:50 Zachary_DuBois Anyone know why salt returns a blank response when running a state.sls?
01:50 Zachary_DuBois I am about to run it with -v
01:51 pppingme joined #salt
01:52 evidence so this is odd.. under gitfs_env_whitelist unless i list base and master, it doesn't load anything
01:58 Zachary_DuBois Yeah
01:58 Zachary_DuBois I get "Minion did not return. [No response]" when running `salt -v 'tmp' state.sls first-setup`
02:00 TheThing_ joined #salt
02:03 Zachary_DuBois That's weird
02:03 Zachary_DuBois So, I looked up the job ID and all the output is there
02:03 Zachary_DuBois It didn't print the output when I ran the command though
02:04 jimklo_ joined #salt
02:07 babilen joined #salt
02:07 StDiluted joined #salt
02:12 Zachary_DuBois Seems to be the timeout
02:12 Zachary_DuBois I added -t 300 and it worked
02:12 Zachary_DuBois Just need to edit the master config now
02:21 alexr joined #salt
02:21 hasues joined #salt
02:21 StDiluted joined #salt
02:21 hasues left #salt
02:27 ssteinerX joined #salt
02:40 nitti joined #salt
02:47 ilbot3 joined #salt
02:47 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/
02:51 ysiad joined #salt
02:57 evidence soo.. if you have GPG in any of your state files referenced by top.sls, _ALL_ minions must have the python-gnupg tools, even if the top.sls doesn't actually include that state file for them
02:57 malinoff joined #salt
02:58 evidence ie i have some state files only being loaded by masters, but until i install the gnupg package on the minions, they refuse to run highstate.  debug confirms they are not loading that sls
02:59 ramishra joined #salt
03:00 Gues480631_ joined #salt
03:09 thayne joined #salt
03:13 glyf joined #salt
03:16 glyf joined #salt
03:16 CeBe joined #salt
03:30 TyrfingMjolnir joined #salt
03:35 otter768 joined #salt
03:38 feythin joined #salt
03:41 CeBe1 joined #salt
03:42 ctrlrsf interesting.
03:44 Mso150 joined #salt
03:44 seshan joined #salt
04:02 seshan Hi
04:04 bhosmer joined #salt
04:13 StDiluted joined #salt
04:23 smcquay joined #salt
04:30 bfoxwell_ joined #salt
04:34 atbell joined #salt
04:46 ecdhe joined #salt
04:49 TK_ joined #salt
05:21 jonatas_oliveira joined #salt
05:22 atbell joined #salt
05:34 TK_ joined #salt
05:35 TK___ joined #salt
05:35 otter768 joined #salt
05:37 TK_ joined #salt
05:48 kormoc joined #salt
05:49 ramteid joined #salt
05:51 Samos123 just filed an issue about mine data not getting updated on github: https://github.com/saltstack/salt/issues/18505 but it may be my stupidity any help would be great :)
05:53 iggy evidence: can you gist the sls files you are referring to?
05:57 iggy Samos123: did you try a mine.update?
05:57 Samos123 i did
05:58 iggy oh yeah, I see it now
05:58 iggy well, it's not what I thought it was, that should be fixed in 20147
05:58 iggy 2014.7
05:59 Samos123 salt '*' config.option mine_functions weird thing is that it shows that i didnt configure mine, although it was working before
06:00 iggy you notice the mine_functions are different right?
06:01 Samos123 i did i already removed the pillar one and only left the minion configuration of mine
06:02 Samos123 as that was my previously working config
06:07 Samos123 same result though
06:08 Samos123 possibly related: https://github.com/saltstack/salt/issues/11501
06:10 TyrfingMjolnir joined #salt
06:11 druonysuse joined #salt
06:11 druonysuse joined #salt
06:12 catpigger joined #salt
06:23 favadi joined #salt
06:25 druonysuse joined #salt
06:32 span joined #salt
06:40 thayne joined #salt
06:44 slafs joined #salt
06:45 linjan joined #salt
06:46 jhauser joined #salt
06:47 slafs left #salt
06:47 kermit joined #salt
06:49 TyrfingMjolnir joined #salt
06:51 oyvjel joined #salt
06:55 TyrfingMjolnir joined #salt
07:05 shorty_mu joined #salt
07:10 jonatas_oliveira joined #salt
07:12 douardda joined #salt
07:17 Leonw joined #salt
07:28 seshan Was trying to install django in a virtualevn, and ended up with this error
07:28 seshan raise CommandNotFoundError('Could not find a `pip` binary') CommandNotFoundError: Could not find a `pip` binary
07:28 seshan any clue ??
07:30 Gues480631 joined #salt
07:36 otter768 joined #salt
07:39 nkuttler seshan: pip is not installed?
07:45 flyboy joined #salt
07:52 xsteadfastx joined #salt
07:53 stanchan joined #salt
07:56 chiui joined #salt
08:00 CeBe joined #salt
08:02 xsteadfastx joined #salt
08:03 zamuz joined #salt
08:07 seshan joined #salt
08:07 seshan hi
08:08 seshan Just tracing back this error :  CommandNotFoundError: Could not find a `pip` binary
08:08 seshan and someone had asked is pip installed, and yes pip in installed
08:09 druonysuse joined #salt
08:09 druonysuse joined #salt
08:09 malinoff seshan, well, salt is not so sure
08:09 seshan so how di I tell salt to be sure pip exists ??
08:10 nkuttler seshan: put it on a PATH where salt can find it?
08:10 nkuttler seshan: (on the minion obviously)
08:11 seshan nkuttler: I quite didnt follow you ??
08:11 nkuttler seshan: er, which part?
08:11 malinoff seshan, what is the exact traceback? Can you paste it to http://pastie.org
08:11 malinoff ?
08:14 seshan malinoff: here you go
08:14 seshan https://www.refheap.com/94006
08:14 seshan nkuttler: posted the traceback
08:15 __gotcha joined #salt
08:15 malinoff seshan, check these conditions: https://github.com/saltstack/salt/blob/develop/salt/modules/pip.py#L43-L51
08:16 malinoff where bin_env is path to your virtualenv
08:16 seshan malinoff: this is failing
08:17 seshan i dont see a pip her
08:17 malinoff seshan, what is failing exactly?
08:18 seshan I dont see pip in the bin_path/bin
08:18 malinoff os.path.isdir or os.path.isfile ?
08:18 malinoff so it wasn't installed into the virtual environment
08:18 sieve joined #salt
08:19 seshan malinoff: isfile according to the traceback
08:20 seshan the env itself is not created properly, I dont see activte or deactivate and everything else, which should exist
08:20 malinoff bingo
08:21 seshan was refering this article : http://www.barrymorrison.com/2013/Mar/11/deploying-django-with-salt-stack/
08:23 malinoff not sure what's wrong, try to use state.sls -l debug to find out the issue
08:27 wvds-nl joined #salt
08:29 Mso150 joined #salt
08:30 Roee joined #salt
08:30 Roee Hi all, i've a very basic question
08:31 Roee Is there a way to run a state on a minion and see the output in live...
08:31 Ixan on the minion: salt-call -l debug state.sls yourstate
08:32 Roee this is very basic... and sometimes running of a state didn't return back any output event at the end of the process
08:32 Roee thanks Ixan, but i'm working only from the salt-master
08:33 Roee and let's say that i'm running a state on a few minions...how can I know at the end of the process that everything passed ok in all of them
08:35 Roee another problem that i'm not feel comfortable with, is that i can't monitor the status of the state in live
08:35 Roee I mean that I need to run the state, and just wait... this is not make sense
08:36 Roee I'm sure that there is option to collect info in live...
08:36 Roee maybe some logs in the master, i don't know...
08:36 malinoff Roee, nope, you can't do that right now
08:37 Roee very bad...
08:37 malinoff Roee, just because modules are written without any possibility to fire events from within
08:37 malinoff so you can just see this: (pre-run) -> (result)
08:38 Roee what do you mean pre-run ?
08:39 malinoff like "before executing state X" and "after executing state X"
08:39 __gotcha joined #salt
08:40 Roee Do  you have an idea why the results are not coming all the time - I mean that i'm running a state and from some reason i don't get the output from server. just getting the prompt
08:40 CeBe joined #salt
08:40 malinoff timeouts
08:40 malinoff or flood
08:41 Roee and how can this be solved ?
08:42 malinoff Have you tried to google it? https://www.google.ru/search?q=salt+minion+did+not+return
08:42 malinoff 367k results
08:43 druonysuse joined #salt
08:43 Roee :)
08:43 Roee I'm looking only for one solution, don't need 367k :)
08:44 mike25de joined #salt
08:44 malinoff I can't give you just one solution that will fix your particular case, you know
08:45 slav0nic joined #salt
08:45 slav0nic joined #salt
08:45 mike25de hi all ... is there a way to check within a state if a file on the salt master exists?  and deploy something on the minion if on the master that file exists... Thanks guys
08:48 lcavassa joined #salt
08:49 __gotcha joined #salt
08:50 flyboy82 mike25de: states are being rendered and executed minion-side, so what you want to do would imply that the minion has filesystem access to  the master... kinda insecure. If you want though you can use "- unless:" in your state and use a script that communicates with the master via some secure manner and returns 0 if the file isn't there and 1 if it is
08:51 mike25de flyboy82: i know .. that's why i was asking...
08:52 mike25de i thought maybe there is a way of pre- logic before rendering :) thanks anyway
08:52 sieve joined #salt
08:52 malinoff mike25de, i'd suggest to rethink your states layout
08:52 mike25de i will then use GIT ... to map to each minion some ... files
08:53 malinoff because configuration management is not scripting, you shouldn't do something depending on something else, you should just do things
08:53 mike25de malinoff:  is not a real state.... i just want to know if i might come to that :)
08:53 mike25de good point malinoff
08:53 karimb joined #salt
08:53 malinoff mike25de, well, please avoid such situations :)
08:53 mike25de thanks for your input guys
08:59 jonatas_oliveira joined #salt
09:00 gildegoma joined #salt
09:05 __gotcha joined #salt
09:08 iwishiwerearobot joined #salt
09:10 jonatas_oliveira joined #salt
09:15 intellix joined #salt
09:19 jhauser joined #salt
09:29 valentinbud joined #salt
09:30 valentinbud Hello Salted people :).
09:30 valentinbud Can I pass pillar from within on state to another?
09:30 gildegoma joined #salt
09:31 valentinbud *one
09:31 felskrone joined #salt
09:32 valentinbud I have the nginx formula with the nginx pillar. Now I have a new formula that needs a vhost and I would like to store the config file for nginx under this formula's pillar not nginx's.
09:35 Samos123 just fyi about issue: https://github.com/saltstack/salt/issues/18505 it just started magically working again after waiting for several hours.... the way salt caches and refreshes is magic which fails....
09:37 otter768 joined #salt
09:43 darvon joined #salt
09:46 __gotcha joined #salt
09:52 intellix joined #salt
09:58 abele joined #salt
10:00 CeBe joined #salt
10:02 lothiraldan joined #salt
10:02 aqua^mac joined #salt
10:02 glyf joined #salt
10:02 ggoZ joined #salt
10:04 viq valentinbud: pillars are pillars, you just access them
10:05 viq valentinbud: or, should I say, you access them _appropriately_ - ie. you have to tell your states which pillars to grab, where they are. So possibly you'd have to change the formula. But nothing is stopping you from doing so
10:06 bhosmer joined #salt
10:06 mikepea Is there a standard work-around to prevent Salt from triggering updates on directories marked as Clean: True?
10:07 mikepea in particular, when you are also managing files that are underneath the cleaned dir
10:07 valentinbud Thanks viq, so what I described above is not possible, right?
10:14 mikepea hmm, have found https://github.com/saltstack/salt/issues/12965 which might solve my problem
10:17 TK_ joined #salt
10:18 CeBe1 joined #salt
10:18 viq valentinbud: sure it's possible, you just need to change some things
10:19 CeBe joined #salt
10:19 valentinbud viq thinking about it what I would like to achieve is something along the lines of specifying pillar via the command line but from within states. What should I change?
10:20 chiui joined #salt
10:21 viq valentinbud: first, make up your mind ;) From command line, or from within states?
10:21 viq Erm, sorry, misread
10:22 valentinbud viq I have made up my mind, from within states, I was just giving an example to be clear what I want to achieve.
10:22 viq valentinbud: you can't define pillars from states, you define pillars from pillars. But, you can assign variables from pillars, if that helps. Or, you can simply assign variables, and use them elsewhere
10:22 viq valentinbud: yeah, sorry, I misread and then realized my mistake
10:23 valentinbud I understand what you are saying viq. I know I can't define pillar from within states.
10:23 N-Mi_ joined #salt
10:23 valentinbud Real world scenario: I have an nginx formula that iterates over nginx:vhosts and configures them with file.managed.
10:25 valentinbud I am about to write a kibana formula and I would like to run from within this formula/state the nginx state/formula with a specific nginx:vhosts. Does it make sense?
10:25 sieve joined #salt
10:25 babilen valentinbud: Why don't you simply include the SLS files in question from Kibana and define an additional pillar that will then get merged?
10:26 sieve joined #salt
10:26 sieve joined #salt
10:26 valentinbud babilen: How would I do that, can you give me an example please?
10:27 babilen Or just include the nginx SLS files in their default and add a file.managed that adds yet another vhost (this time hard-coded defaults for kibana/populated from kibana pillar)
10:27 felskrone joined #salt
10:27 sieve left #salt
10:28 babilen valentinbud: "include: - nginx" in your Kibana formula and you are done. The pillar merge assumes that you use a suitable merge strategy (cf. http://docs.saltstack.com/en/latest/ref/configuration/master.html#pillar-source-merging-strategy)
10:28 valentinbud Thanks babilen I think I'll do just that, it's way easier.
10:28 babilen For the latter I'd also "include: - nginx" (or whatever is appropriate) and then just add an additional file.managed for the kibana vhost in question and tie that into the nginx formula with watch/listen/...
10:29 babilen In the latter case you could also allow for configuration of the "kibana vhost" from the kibana pillar rather than the nginx one
10:30 babilen I think that I prefer the second approach as it is more explicit and doesn't require the user to setup a pillar in every case. I'd just make that "kibana.nginx" and include the nginx formula SLS files that you require from there
10:32 valentinbud babilen: I understand. I think I'll choose the same approach too. Thanks.
10:34 CeBe1 joined #salt
10:35 badon joined #salt
10:36 ckao joined #salt
10:41 fredvd joined #salt
10:44 giantlock joined #salt
10:52 TK_ joined #salt
10:58 valentinbud joined #salt
11:03 alexr_ joined #salt
11:05 chiui joined #salt
11:07 CeBe joined #salt
11:08 linjan joined #salt
11:08 istram joined #salt
11:08 rjc joined #salt
11:10 diegows joined #salt
11:10 CeBe1 joined #salt
11:13 blackshadow joined #salt
11:17 TK_ joined #salt
11:18 johtso_ joined #salt
11:35 BigBear joined #salt
11:38 otter768 joined #salt
11:41 Katafalkas joined #salt
11:42 CeBe joined #salt
11:44 seshan testing
11:45 ggoZ joined #salt
11:45 AirOnSkin Hmm, when I run the following command: salt myhost cmd.run 'yum -y install git' it returns timeout errors. When I run the same command on the local host it works. I'm pretty sure it has to do with our proxy settings. So I'm guessing that Salt doesn't know about environment variables?
11:46 viq AirOnSkin: salt times out or yum times out?
11:46 viq AirOnSkin: also, why not pkg.installed?
11:47 viq AirOnSkin: and I believe you need to configure proxy in yum.conf, I remember talking with someone about it here.
11:47 diegows joined #salt
11:48 jonatas_oliveira joined #salt
11:48 brayn joined #salt
11:49 AirOnSkin viq: yum show's the timeout error. I use pkg.installed in states, but they fail. So for testing I wanted to try to just install one package from the master. I could define a proxy in YUM, but we have our own repo server internally. YUM honors the http_proxy environment variables but also the no_proxy variable. See here (works on the client):
11:50 AirOnSkin http://hastebin.com/ihefozijuv.cpp
11:51 aqua^mac joined #salt
11:51 AirOnSkin The proxy is configured globally but an exception is made for our repo server... What I can't explain is, that IF Salt would read these settings, then the no_proxy variable would take effect, if it doesn't it wouldn't matter ...
11:51 AirOnSkin It should work either way, but it doesn't if the action comes from the master
11:51 viq meh, another one of those that won't show you anything without javascript...
11:52 AirOnSkin Sorry about that. I like it because it's clean and fast
11:52 viq AirOnSkin: well, read from where?
11:53 AirOnSkin viq: That would be /etc/profile.d/proxy.sh
11:53 viq That's read by shell, salt doesn't use shell
11:53 AirOnSkin Hmm...
11:54 AirOnSkin It uses python?
11:54 viq I guess so
11:54 AirOnSkin To call YUM I mean...
11:54 AirOnSkin I see
11:55 AirOnSkin Ok, that helps. I'll see what more I can find out.
11:55 viq AirOnSkin: I know you can put a bunch of those settings in yum.conf, maybe the no_proxy too?
11:55 viq (also you can pass env to cmd.run)
11:56 alanpearce joined #salt
11:57 viq https://docs.oracle.com/cd/E37670_01/E37355/html/ol_yum_config.html#ol_proxy_config
11:58 AirOnSkin oO, a doc on YUM by Oracle ^^
11:58 AirOnSkin I'll look into it :) Thanks
11:59 felskrone joined #salt
12:00 jonatas_oliveira joined #salt
12:00 ssteinerX joined #salt
12:02 AirOnSkin viq: You can only define a proxy in yum.conf. You can't specify no_proxy or something like it in yum.conf. You can however add 'proxy=_none_' to each repository file. This will override the global settings in yum.conf
12:02 AirOnSkin Tested it and it works :)
12:03 viq there ya go ;)
12:03 ecdhe joined #salt
12:03 viq And since you already have salt in place it will be a breeze to put it on machines ;)
12:06 CycloHex joined #salt
12:11 wincus joined #salt
12:11 Cyclohex joined #salt
12:13 Cyclohex When trying to change or add grains through CLi (salt "targets" grains.setval foo bar), my minion sometimes crashes.. It is then down. I have t manually go into the system and restart the service. Is this a known problem?
12:17 TK_ joined #salt
12:17 AirOnSkin viq: I'm still in testing. The repo files are managed by Salt as well, so the only thing I need to do, is add the option to the repo files on the Salt master... :) (I like that a lot!)
12:20 badon joined #salt
12:20 blaffoy joined #salt
12:21 blaffoy Hi, I'm having trouble understanding how roles work. Can anybody help?
12:22 blaffoy I've a windows minion configured with two roles, salt-test and developer
12:22 blaffoy I can confirm on the master with grains.items that these roles are set
12:22 Cyclohex blaffoy, roles is a grain. You can then add values to that grain, like developer, webserver, dataserver etc. They are moslty use din top-files.
12:22 blaffoy roles:
12:22 blaffoy salt-test
12:22 blaffoy developer
12:23 Cyclohex then in your top file: base:  'G@roles:developer':  - match: grain - SLS
12:23 blaffoy But then I can only get test.ping to work against the first role, not the second.
12:24 blaffoy `sudo salt -C "G@roles:salt-test" test.ping` responds
12:24 blaffoy but `sudo salt -C "G@roles:developer" test.ping` does not
12:25 blaffoy Hi Cyclochex. Thanks for responding.
12:25 Cyclohex try sudo salt -G 'roles:developer' test.ping
12:26 Cyclohex -G matches on grains, -C is mostly used to match on multiple things, grains, ids or w/e
12:26 Cyclohex -C stands for compound matching
12:26 blaffoy Right, I had paired it down from a more complicated match condition
12:27 blaffoy Same outcome: `sudo salt -G 'roles:developer' test.ping` does not respond, but `sudo salt -G 'roles:salt-test' test.ping` does
12:29 blaffoy Can you confirm the normal way to assign a role to a minion?
12:29 blaffoy I have a file on the minion at C:\salt\conf\grains
12:29 blaffoy with contents: roles:\n  - salt-test\n  - developer
12:30 Cyclohex not sure about the location.. I don't use windows with salt
12:31 Cyclohex But you confirmed that you hav eminions with the grain roles set to developer?
12:32 blaffoy This is a test configuration, just one master (on ubuntu) and one minion (windows 7)
12:33 linjan joined #salt
12:33 blaffoy I'm planning to change some infrastructure in my org using salt roles. Right now there's a lot of ad hoc machine config in the saltstate
12:33 blaffoy So I can't confirm that the 'developer' role works on any minion
12:34 blaffoy But I can't see why it wouldn't work on this one.
12:35 Cyclohex What's the output for the command 'sudo salt '*' grains.item roles
12:36 blaffoy Another question: I've changed the order of the roles in my salt/conf/grains file, to see if that makes any difference. If I salt-call grains.items on the minion, it reflects this change, but if I salt "*" grains.items on the master it does not reflect the change. How can I force the master to read the update minion config?
12:36 blaffoy Output to that command is:
12:36 blaffoy minion-machine.my-org.local:
12:36 blaffoy roles:
12:37 blaffoy salt-test
12:37 Cyclohex whenever you change something to pillars or grains it's best practice to sync them using the saltutil module
12:37 blaffoy developer
12:37 Cyclohex salt '*' saltutil.sync_grains
12:37 Cyclohex sync_all can be used as well
12:37 blaffoy Cheers
12:38 blaffoy Huh, it looks like sync_grains fixed the problem.
12:39 blaffoy Now both roles respond to ping
12:39 blaffoy I've switched them back around in the conf/grains and synced again, and it still works
12:40 __gotcha joined #salt
12:40 Cyclohex ok
12:40 Cyclohex Glad it works ;)
12:40 blaffoy I'm not sure I understand this; the caching scheme must work differently to how I had assumed, but it works.
12:40 nosleep77 joined #salt
12:40 felskrone joined #salt
12:44 CeBe joined #salt
12:44 drawsmcgraw joined #salt
12:45 JlRd joined #salt
12:45 drawsmcgraw I don't suppose there's any closure to the "minions fall asleep" issue? https://github.com/saltstack/salt/issues/16518
12:46 bhosmer joined #salt
12:47 intellix joined #salt
12:51 VSpike Would you folks tend to have multiple salt-masters based on the environment? e.g. one for internal systems, one for development/qa systems, and one for production?
12:51 VSpike Or do you tend to stick to one for all purposes?
12:52 __gotcha joined #salt
12:58 genediazjr joined #salt
13:02 LittUp joined #salt
13:08 linjan joined #salt
13:08 lothiraldan joined #salt
13:08 hcl2 joined #salt
13:10 Gues480631 joined #salt
13:13 Roee Hi, does someone can point to good example of how to use templates in salt states...
13:14 Roee I wrote few states, but every running i have few arguments that I want it to take automatically from a template file (instead of update the state itself...)
13:17 TK_ joined #salt
13:17 mike25de joined #salt
13:20 viq Roee: I think you may mean a map file?
13:22 viq Roee: template would be for example "here's nginx config, put vhosts here". Map would be "here's where to put nginx config file on this system" for example
13:23 drawsmcgraw VSpike: I indulge in my own personal salt master for dev stuff
13:23 drawsmcgraw That way, when I'm ready to promote to production, I just merge my states/files into the master branch on my Git repo
13:24 drawsmcgraw And it prevents any careless `salt '*' cmd.run something-awful` situations :)
13:24 Rory joined #salt
13:25 Roee viq : I tell what i need, I've created state,sls files that using few arguments that need to be changes from time to time
13:26 viq cmd.run nuke_from_orbit the-only-way-to-be-sure=True
13:26 Roee I would like to define a file (maybe template) that I will need to change the argument within and the state will read the arguments from this file
13:26 _JZ_ joined #salt
13:27 viq Roee: plenty of ways to do it. You could set some variables in jinja in your sls file. You could do the same via map and include it. Or you could pass those variables via pillars.
13:27 Rory Roee: You're describing pillars
13:28 thawes joined #salt
13:28 Rory But I will just idle now or I'll steal R[tab] from people who are actually able to help you
13:28 linjan joined #salt
13:28 Roee Ok great ! so how can I find the way to build a pillar that my state will collect the argument from it ?
13:29 Rory You refer to the pillar as (for example) {{ pillar["base"]["route53_domain"] }}
13:29 Rory So in that case, the "route53_domain" pillar value would be pulled from the "base" pillars
13:30 Rory which looks like
13:30 Rory base:
13:30 Rory route53_domain: foo
13:30 Roee OK thank you very Much !
13:30 Rory Roee: http://docs.saltstack.com/en/latest/topics/tutorials/pillar.html as well
13:30 Roee I will check this...
13:34 TheThing joined #salt
13:39 otter768 joined #salt
13:39 aqua^mac joined #salt
13:40 aquinas joined #salt
13:44 TyrfingMjolnir joined #salt
13:45 CeBe joined #salt
13:47 oqewfhuge joined #salt
13:47 thawes joined #salt
13:51 viq A Ro{ee,ry} conversation. You're even marked the same colour by my IRC client ;)
13:52 nitti joined #salt
13:52 Guest14357 joined #salt
13:53 imanc joined #salt
13:54 Roee :)
13:54 che-arn8 joined #salt
13:54 che-arne joined #salt
13:55 Roee Yes I actually little bit confused:)
13:55 wiqd joined #salt
13:55 Guest33099 joined #salt
13:56 VSpike drawsmcgraw: I was thinking the same kind of thing. Also allows you test master/minion upgrades more safely
13:56 VSpike I certainly think a dev/prod split makes sense
13:57 VSpike If I end up using salt for internal stuff at work like domain controllers, file servers, even workstations, then it might make sense to have that separate too
13:57 elfixit joined #salt
13:57 thawes joined #salt
13:59 Cidan joined #salt
14:00 pipeep joined #salt
14:00 Ouzo_12 hmm if i do "salt '*' test.ping -l debug" on a newly installed salt-master server with about 20 minions connected to i get a lot of "True" from the servers but i does not finish it just repeats the same 3 linies at the end
14:01 Ouzo_12 [DEBUG ] MasterEvent PUB socket URI: ipc:///var/run/salt/master/master_event_pub.ipc
14:01 Ouzo_12 [DEBUG ] MasterEvent PULL socket URI: ipc:///var/run/salt/master/master_event_pull.ipc
14:01 Ouzo_12 [DEBUG ] Checking whether jid 20141126144917304555 is still running
14:01 m0nky joined #salt
14:07 bhosmer joined #salt
14:08 BigBear joined #salt
14:08 racooper joined #salt
14:10 genediazjr joined #salt
14:10 nliadm joined #salt
14:11 fredvd joined #salt
14:11 g3cko joined #salt
14:16 Cidan joined #salt
14:17 toddnni joined #salt
14:18 mschiff joined #salt
14:18 mschiff joined #salt
14:19 mordonez joined #salt
14:21 douardda joined #salt
14:23 manytrees joined #salt
14:24 mpanetta joined #salt
14:24 Cidan joined #salt
14:24 Cidan joined #salt
14:29 lacrymology joined #salt
14:30 Shiv joined #salt
14:30 Shiv Hi
14:31 lacrymology I have a beginner's question. I have a web server with a couple of deployments, they're all pretty much identical except for content details that can be defined by pillars, such as, project name, repository address, the database name should be the same as the project name, password should be generated on creation, etc. What would be the best way to maintain this with salt and, specially, to deploy a new projecy?
14:31 Shiv If the minion disk is full and i issue "salt "*" test.ping" on Master does it not display any output ?
14:32 lacrymology I understand enough to create the states that describe the server and a generic deployment that can be looped with the different names in jinja, but I wonder if I just add a new deployment to the pillars and highstate, won't my existing deployments be reset somehow?
14:32 Shiv I m able to see the key and increased the timeout also .... but not output after execution of any salt command
14:34 __gotcha joined #salt
14:34 lacrymology Shiv: if the disk is really full it's very possible that either the minion process or even the OS is not being able to do things like create temporary files that it might need for the execution or communication.
14:34 viq lacrymology: if they wouldn't on "normal" highstate, then no, they shouldn't
14:34 lacrymology viq: what is "normal" highstate, and again, I'm not sure exactly how to know if they will or won't
14:34 viq lacrymology: "oh, hey, I have another position/place to do something now for"
14:35 viq lacrymology: all depends how you define your states
14:35 lacrymology of course
14:35 lacrymology but let's assume what one thinks of as standard behavior. I guess something like db.managed would noop if it already existed
14:36 viq lacrymology: you can always run with test=True and see what would be done.
14:36 Rucknar___ joined #salt
14:37 ajolo joined #salt
14:37 lacrymology yes, of course, but I'm kind of thinking about how to write the states exactly. Actually just asking the question I think dislodged a couple of doubts I had
14:37 viq lacrymology: for i in "get list of projects" ; do stuff ; done
14:37 lacrymology viq: is that how you'd do it? to add a new project, you just add it to the deployments pillar or whatever, and highstate?
14:37 viq ;)
14:38 viq If they're so similiar then yeah, probably
14:38 nliadm joined #salt
14:38 lacrymology ok, good
14:38 bfoxwell_ joined #salt
14:38 lacrymology and.. had one more
14:38 lacrymology and
14:38 lacrymology ah
14:39 lacrymology I wanted to make myself a bootstrapping function for creating new projects, but I think it's best if I just wrote that in bash or python
14:39 viq bootstrapping being what in this case?
14:39 Rory So, zombie salt-master processes being left around and causing various instability issues... salt devs keep saying it'll be fixed "in the next release", but in the meantime is anyone working around this?
14:40 Hazelesque joined #salt
14:41 mikepea joined #salt
14:42 lacrymology viq: cloning a base git repo (same for all), creating a couple specialized packages and settings files (templateable), adding and committing them, activating a python virtualenv, running a couple of commands (syncdb), creating a new repo in bitbucket and pushing this there
14:42 felskrone joined #salt
14:43 lacrymology viq: all of this in my workbox, of course
14:43 viq Depends... You could manage such dev environments on people's machines via salt. Or you could have a script that does that.
14:44 lacrymology viq: that for dev, for the server side it would just be cloning this newly created repo in a specific directory and doing similar things, but in the server's case it'd be done with salt easily enough
14:44 lacrymology yeah, but this would be just for me
14:44 lacrymology something like 'startproject <foo>' should be enough
14:44 arapaho joined #salt
14:45 viq right
14:45 lacrymology well
14:46 lacrymology I gotta go to bed
14:46 lacrymology thanks for your time
14:47 Hazelesque joined #salt
14:47 mikepea joined #salt
14:48 Cidan joined #salt
14:48 xenoxaos joined #salt
14:51 basepi joined #salt
14:52 nliadm joined #salt
14:53 spielberg joined #salt
14:54 balltongu joined #salt
14:54 wincus joined #salt
14:56 che-arne joined #salt
14:57 jaimed joined #salt
14:57 housl joined #salt
14:58 jakubek joined #salt
14:59 nliadm joined #salt
15:00 viq joined #salt
15:01 jakubek im looking for better solution, at this moment i have 3 state files which are installing different versions of ruby. state files have name ruby212.sls ruby193.sls ruby215.sls. the difference in these files is rpm file. maybe there is better way to do it?
15:01 zz_Cidan joined #salt
15:02 jakubek i can create pillar for ruby_version and then check it in sls but maybe you have other ideas :-)
15:05 InAnimaTe joined #salt
15:07 viq jakubek: there was a blog post, "pillar driven infrastructure" I think, you could look at that
15:07 viq jakubek: do you foresee installing multiple at once?
15:08 pdayton joined #salt
15:09 linjan joined #salt
15:10 bhosmer joined #salt
15:13 jakubek viq: nope, always only one version
15:14 viq then putting that eg in pillar and/or doing it like in that blog post should work
15:16 nliadm joined #salt
15:16 xenoxaos joined #salt
15:17 TK_ joined #salt
15:21 troyready joined #salt
15:22 Katafalkas joined #salt
15:22 jakubek viq: k, ill try to find that post, thanks
15:22 OnTheRock joined #salt
15:24 viq jakubek: http://www.willdurness.com/post/101277984950/salt-pillar-driven-design-pattern
15:25 oyvjel joined #salt
15:25 davet1 joined #salt
15:26 jakubek viq: thank you :)
15:28 agend joined #salt
15:28 aqua^mac joined #salt
15:29 VSpike How would you handle failover of salt masters? We have a DR site, but everything currently points to the salt master at the primary site. Obviously if everything on the master is in git, then it's easy to keep a spare server up-to-date. Just manage the failover by DNS?
15:29 VSpike Or is there a more cleverer way?
15:30 iggy you'll need the keys synced from the master as well
15:31 ale joined #salt
15:37 pr_wilson joined #salt
15:37 tafa2 joined #salt
15:38 evidence soo.. if you have GPG renderer enabled on the master, _ALL_ minions must have the python-gnupg installed to function, even if you don't have any references to gpg stuff in state files the minions parse
15:38 evidence seems like that might be a bug, is the master somehow requesting the minion have it?
15:39 evidence 2014.7.0 ^
15:39 spielberg joined #salt
15:40 otter768 joined #salt
15:45 InAnimaTe joined #salt
15:47 conan_the_destro joined #salt
15:48 babilen evidence: That came up on the ML recently I think. Let me check ... (someone definitely mentioned that before)
15:48 TheThing joined #salt
15:48 babilen No, must have been in here.
15:48 iggy it was in here
15:49 bhosmer joined #salt
15:50 babilen Ah, did they figure out a solution or was an issue filed?
15:50 iggy not that they mentioned
15:51 iggy I think it was actually the same person
15:51 babilen Ah, in that case:
15:52 babilen evidence: Did you file a bug report about this? I have to confess that I have not yet played with the GPG renderer and can therefore not comment on your issue.
15:53 YiP joined #salt
15:59 jrluis1 joined #salt
15:59 Gues480631 joined #salt
16:02 InAnimaTe joined #salt
16:02 evidence left #salt
16:02 evidence joined #salt
16:03 evidence babilen/iggy: yeah it was me prior, but figured everyone was idle :p  i wanted to check and see if anyone had run into it first before i dropped an issue in github.  might try to poke around and see what causes the client to receive that message
16:04 iggy I don't see that feature mentioned much in here (i.e. I doubt many people use it)
16:04 monkey66 left #salt
16:05 evidence when the master has 'renderer: jinja | yaml | gpg' it seems to trigger it, so guessing somehow the requirement gets passed to the minions
16:05 evidence iggy: thx, yeah it's new in helium
16:05 evidence real neat though, let's you store stuff encrypted in pillar, for revision control or whatnot
16:05 StDiluted joined #salt
16:09 iggy I figure if someone gets into our revision control, the last thing we need to worry about is an extra layer of gpg encryption of some pillar data
16:09 hasues joined #salt
16:09 jalbretsen joined #salt
16:10 Ahrotahntee paranoia is the new black
16:10 Ahrotahntee you can't trust anyone, let alone your employees/staff members
16:10 seanz joined #salt
16:10 evidence sometimes 'someone' is people who do have access
16:10 evidence heh yea
16:10 glyf joined #salt
16:10 hasues left #salt
16:11 Ahrotahntee the only safe place left in the world is the inside of your head; maybe that's enough.
16:11 iggy yeah, if I didn't trust someone to have access to our source control, I wouldn't give them access, but I realize that's not an acceptable answer for everyone
16:15 Ozack joined #salt
16:15 SheetiS joined #salt
16:15 rvankleeck joined #salt
16:15 Bryanstein joined #salt
16:17 viq Ahrotahntee: not necessarily, they are finding ways to figure out what you're thinking about ;)
16:17 TK_ joined #salt
16:19 beneggett joined #salt
16:24 rvankleeck is there a way to get non-static data from minions? E.G. cpu load, ram usage, etc.
16:25 viq rvankleeck: there are modules for that, I believe
16:25 SheetiS http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.status.html
16:25 tmh1999 joined #salt
16:26 iggy for some stuff... but I've said it before and I'll say it again... salt is not a monitoring system
16:27 Wagahai left #salt
16:27 so-user joined #salt
16:27 rvankleeck Thanks viq , SheetiS , iggy . that's very helpful!
16:29 VSpike iggy: good point
16:30 babilen iggy: I think the point is rather that you can easily host repos publicly *or* don't have to worry if a dev loses a laptop (that was, for some reason that we will never understand, not encrypted)
16:31 babilen rvankleeck: Look into riemann, grafana and statsd/...
16:31 rvankleeck SheetiS, that pointed me in the direction I needed to look. http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.yumpkg.html#module-salt.modules.yumpkg
16:32 quickdry21 joined #salt
16:33 BigBear hi folks
16:33 Ahrotahntee hi
16:34 BigBear I have finally gotten 100% proof that the latest salt-minion installer for 2014.7.0 for 32-bit is broken. What is the best and fastest way to report this? I already have found what is wrong, who can I tell so they can fix it?
16:34 babilen BigBear: On which operating system?
16:34 iggy the github issue tracker
16:35 * babilen guesses Windoof
16:35 iggy or yeah... that is assuming it's not platform specific
16:35 BigBear habilea: on windows 7 pro 32-bit - basically the 32-bit installer has the 64-bit binary of nssm.exe in it . ;-)
16:36 BigBear iggy: so where do I go and report this on github usng the issue tracker?
16:37 iggy I'm really not sure about the Windows installers
16:37 ajolo joined #salt
16:37 vlcn joined #salt
16:38 BigBear iggy: ok - thanks.
16:39 babilen BigBear: Most IRC clients support tab completion of nicknames. Try "bab<TAB>".
16:40 babilen BigBear: Just report it against saltstack, they will let the packagers know and that way other people who run into this have a chance to stumble over it.
16:41 BigBear babilen: thanks ;-) learned something new that will save me loads of time. ;-) the second I spotted my misspelling - i thought - oh they'll know ... ;-)
16:41 BigBear babilen: thanks
16:41 intellix joined #salt
16:41 babilen (you can also press tab multiple times to cycle through alternatives)
16:46 Frank_ joined #salt
16:48 __gotcha joined #salt
16:51 eagles joined #salt
16:52 eagles Hey guys, I am trying to make it but it returns me an error.. Is there any idea of what is going on?
16:52 eagles {% set 'value' = {{ salt['pillar.get']('mysql:lookup:user') }} %}
16:52 eagles The error is: "........ failed: Jinja syntax error: can't assign to 'const'; line 1"
16:52 mohae joined #salt
16:52 iggy you can't put {{ }} inside {% %}
16:53 iggy just leave that out and it should work
16:53 eagles I tried to remove {{ }} as well but it returned the same error, let me try it again
16:53 tmh1999 joined #salt
16:53 iggy oh
16:53 iggy also, remove the '' around value
16:53 eagles ohhhhh
16:54 iggy did you see that in an example somewhere?
16:54 babilen eagles: And are you sure that :lookup: is correct there? It should typically only be used for overriding values in that lookup table (that is: map.jinja)
16:55 eagles No, I tried to id by myself according to this page http://docs.saltstack.com/en/latest/topics/best_practices.html
16:55 eagles it is a pillar with all those values filled out
16:55 * babilen should finally get around to fix the :lookup: idiocy in there
16:55 babilen eagles: fwiw, are you aware of https://github.com/saltstack-formulas/mysql-formula/ ? It is a rather nice formula
16:56 eagles hmm let me check it out
16:56 jonatas_oliveira joined #salt
16:57 eagles Guys, actually the code is
16:57 eagles {% set servico = pillar.get['services:appserver_apo:tmp_path'] %}
16:57 eagles And now with that changes it is returning
16:58 eagles Rendering SLS 'protheus:nameofthestate' failed: Jinja variable 'builtin_function_or_method object' has no attribute 'services:appserver_apo:tmp_path'
16:58 babilen [ vs (
16:58 eagles hmmm
16:59 troyready joined #salt
16:59 babilen You can use: pillar.get('foo', default_value), salt['pillar.get']('foo', default_value) or pillar['foo']
16:59 eagles Not it didn't show me any error, good!
16:59 eagles hmmm
16:59 eagles good
17:00 eagles Guys, now there is no error but the state I am doind is trying to echo the var "servico" that is filled out with the return of the pillar
17:00 eagles Despite there was no error, the echo returned me "None"
17:00 eagles When I run as module like it works
17:01 eagles "salt -v minion pillar.get services:appserver_apo:tmp_path" it returns the value but in the state it returns the var "servico" empty
17:02 Frank_ joined #salt
17:02 eagles test:   cmd.run:     - name: echo {{ servico }}
17:03 babilen Paste your state and commands in their entirety to http://refheap.com please. It is hard to follow you in here.
17:03 eagles ok
17:05 Frank123 joined #salt
17:06 Gues480631 joined #salt
17:06 meylor joined #salt
17:07 eagles https://www.refheap.com/94041
17:08 alanpearce joined #salt
17:08 KyleG joined #salt
17:08 KyleG joined #salt
17:08 iggy appserverapo vs appserver_apo
17:11 eagles Yeah! Just fixed but still "      Command "echo None" run            "
17:11 eagles :(
17:11 meylor left #salt
17:11 iggy just fixed the state or pillar? did you do a fileserver.update (if necessary)?
17:11 shaggy_surfer joined #salt
17:13 eagles I did saltutil.refresh_pillar already
17:13 eagles I fixed the pillar
17:14 Frank123 joined #salt
17:14 Frank123 test
17:14 babilen FAILED
17:14 Frank123 first time on IRC sorry
17:15 babilen Frank123: There are typically channels such as #test for testing. Do you have a saltstack related question we can help you with?
17:15 Frank123 New to Salt . . . want to install Dev Studio, but Googling did not find a standard reciept
17:15 KyleG babilen: don't be a dick
17:15 KyleG to the newbie
17:15 babilen KyleG: Excuse me?
17:15 KyleG you're exscused.
17:15 KyleG excused*
17:16 babilen wtf, calm down please.
17:16 jrluis joined #salt
17:16 KyleG I'm so angry, rawrrrr i'm a dinosaur? lol
17:16 babilen ...
17:16 babilen Frank123: So, you want to install that on Windows I guess?
17:16 nethershaw joined #salt
17:17 Frank123 sorry . . . yes . . . I assumed it would just be obvious, but it isn't (at least to me)
17:17 eagles iggy, any idea of that is happing? :)
17:17 Frank123 windows
17:17 aqua^mac joined #salt
17:17 TK_ joined #salt
17:17 eagles of what ***
17:17 babilen I haven't been working with Windows at all, but I believe that you would want to make use of http://docs.saltstack.com/en/latest/topics/windows/windows-package-manager.html for that.
17:18 iggy eagles: no, but then I've never tried echo'ing things to test... I'm not even sure if that's expected to work
17:18 jp__ joined #salt
17:18 babilen eagles: Could you provide a similar pastebin to the one above with the, now fixed, pillar reference?
17:18 Frank123 looks like a great place to start . . . let me read and I'll be back
17:18 jdesilet joined #salt
17:19 eagles I will make a new paste...
17:19 babilen I've also never done that and am not sure if what you are trying there is actually supposed to work. In fact I believe that it wouldn't. Do you want to actually do something with that value, eagles ? Is the pillar setting correct if you run "salt 'yourminion' pillar.item services ?
17:20 babilen Frank123: Good luck :)
17:20 eagles Actually, it is not just a text, I will post the idea of what I will do better in then share with you... that was the first test to build all the states I need to do
17:20 mike25de joined #salt
17:21 eagles I will paste everything there, only few minutes to change the sensitive data please
17:23 jonatas_oliveira joined #salt
17:23 Frank123 good great . . . leave lots for me to learn, but enough I should be able to make some progress
17:23 Frank123 THANK YOU
17:24 RedundancyD joined #salt
17:25 aparsons joined #salt
17:28 torment joined #salt
17:32 TK_ joined #salt
17:34 aparsons joined #salt
17:41 otter768 joined #salt
17:42 spookah joined #salt
17:43 jrluis joined #salt
17:43 eagles Hello guys, I am back
17:43 eagles Please, check it out
17:43 eagles https://www.refheap.com/94043
17:44 eagles The idea is for each service, run the 3 functions on the state file
17:45 bhosmer joined #salt
17:46 iggy you can use archive.extracted (it supports s3 urls just like file.managed)
17:48 eagles I tried but it complained about the source when I tried to run it... but i think I did some mistake... no problems, I will try to use that after... the issue is that I need to use the one "for" of the each service (aaa, bbb)
17:48 jonatas_oliveira joined #salt
17:50 iggy I would convert the pillar to just be a dict instead of a list of dicts
17:50 ph8 joined #salt
17:50 Moss joined #salt
17:51 SeanM joined #salt
17:51 eagles You mean to use archive instead of cmd.run ?
17:51 Moss Hi, simple question: How do I match a string to a regex in jinja? i.e. {{ if "string" == "^s.*" }} ??
17:52 iggy just in general (too me it makes looping easier
17:52 iggy Moss: you don't
17:53 eagles let me try it
17:54 SeanM I was wondering if someone could help me with this. Im running Debian Jessie and I installed Salt however the installer is not creating the minion or master config files in /etc/salt
17:55 shaggy_surfer joined #salt
17:55 jswanson joined #salt
17:56 Moss I want to match a hostname using a regex inside an sls. How can I do it then? Something like this to be more accurate: {% if grains['host'] == 'dev[^1]' %}
17:56 iggy Moss: if that is your actual goal (and not a contrived example), you could use {% if string[0] == "s" %}
17:56 eagles In the archive.extracted, the " - source " shall be the full path of my tar file ?
17:56 iggy Moss: you would normally do that in your top.sls file
17:57 Moss I mostly do, but in this particular case I prefer to do it in another sls file
17:57 Moss can't I do it with an if then?
17:57 iggy jinja isn't really meant to be used like that
17:59 hal58th1 Moss: This is the only thing you can use. startswith and endswith.. https://groups.google.com/forum/#!topic/salt-users/DQ3RouJF_9Y
17:59 SgtMalicious joined #salt
17:59 forrest joined #salt
18:00 Moss ok i'll have to do it from the top sls then... thanks
18:00 __gotcha joined #salt
18:01 iggy yeah, you can use all the standard python stuff... that does not include regex's though
18:02 Moss something like =~ would be nice :)
18:02 * iggy shudders
18:04 eagles-salt joined #salt
18:05 eagles-salt iggy: I updated it
18:05 eagles-salt https://www.refheap.com/94045
18:06 ecdhe Anyone here configured a ubiquiti router with salt?
18:07 iggy A. I don't see any difference... B. I'm not sure what problem you're actually trying to solve at this point
18:07 iggy those definitely aren't complete
18:08 eagles-salt the function to extract is using archive.extracted instead of cmd.run with tar
18:09 iggy you can't use local paths for archive.extracted (or file.managed)
18:10 iggy it's got to be a salt://, http(s)://, or s3 url
18:10 iggy or ftp I guess, but I've never tried it
18:10 eagles-salt the source will be {{ s3_path }} , right ?
18:11 eagles-salt and there will be not {{ tpm_path }}
18:11 iggy afaik
18:11 eagles-salt and there will be no {{ tpm_path }}
18:12 eagles-salt updated again
18:16 shaggy_surfer joined #salt
18:17 TK_ joined #salt
18:28 eagles-salt iggy, well, to make the state to know the arguments from the services (aaa , bbb), how whould I buildthe state file?
18:28 eagles-salt I mean, to use for with that one
18:28 eagles-salt I was not able to find a good example
18:29 iggy there are plenty of examples of how to loop over pillar data in jinja templates
18:30 eagles-salt I could find coulpe of them but either I was not able to understand or they do not do what I was expecting, like the example I pasted here
18:30 glyf joined #salt
18:31 shanemhansen joined #salt
18:32 snuffeluffegus joined #salt
18:32 shanemhansen I just installed salt-master in a docker container and there are 6 salt-master processes using up all my cpu. The processes recycle every couple seonds. I don't see any errors in the log other than a warning about dmidecode not being installed.
18:32 shanemhansen Is this a known issue? I couldn't find it on the google.
18:32 douardda joined #salt
18:33 shanemhansen (centos6 using salt-master from epel repos)
18:34 hal58th joined #salt
18:36 iggy salt does certain house keeping tasks on a schedule
18:37 drawsmcgraw I hadn't thought about putting a Master inside a Docker container....
18:37 shanemhansen It does seem to have settled down, but it was burning 100% cpu for several minutes. strace showed a whole bunch of : close(107796)                           = -1 EBADF (Bad file descriptor)
18:38 shanemhansen drawsmcgraw, I'm just playing around with a local test cluster.
18:38 drawsmcgraw shanemhansen: interesting
18:38 drawsmcgraw The master wants ports 4505 & 4506. I don't suppose it could be failing to get those ports because it's in a container?
18:38 drawsmcgraw or are you exposing those?
18:39 shanemhansen It has those ports. I can access my master from my minion on those ports.
18:40 gildegoma joined #salt
18:40 drawsmcgraw okay. I got nothin', then. My Salt internals knowledge isn't quite that good :/
18:40 shanemhansen I think the default docker networking setup is basically like virtualbox's internal networking setup. Each instance has it's own ip.
18:41 Katafalkas joined #salt
18:42 bhosmer joined #salt
18:42 shanemhansen It's not a showstopper for me, but this seems bad: https://gist.github.com/shanemhansen/cb03cc78ad3a65401d26
18:42 shanemhansen I'll try and get a stack trace.
18:44 waygeekierthanu joined #salt
18:44 faust joined #salt
18:44 eagles-salt iggy, sorry to insist on that :) but I only need now to set the var with the value of that pillar
18:44 eagles-salt like it
18:44 eagles-salt {% set service = pillar.get('services:appserver_apo:tmp_path') %}
18:44 eagles-salt do see you anything wrong here?
18:44 iggy don't
18:44 roblarkin joined #salt
18:45 waygeekierthanu Hi all. I'm trying to use salt-ssh, but I'm unsure how to use a gateway with it? Anybody know how that works? I have it working fine in my ~/.ssh/config file
18:45 jonatas_oliveira joined #salt
18:45 iggy if you loop over the pillar data correctly, you'll have access to everything without having to use pillar.get
18:47 roblarkin I am struggling to create a lxc container with salt-cloud for some days now.  no matter what I do the name of the container comes back in error 'Configuration (name) needs to be set'.  I just need to create one container with salt-cloud.  Anyone have a link to something bullet proof?  if lxc-list returns a name I can go home.
18:49 riftman joined #salt
18:50 pdayton joined #salt
18:50 InAnimaTe joined #salt
18:50 Ryan_Lane joined #salt
18:52 eagles-salt thank you
18:53 simmel What is the easiest way to get *one* file from a git repo into a specific path?
18:53 glyf joined #salt
18:54 simmel With salt.states.git I must check out the repo and then move the file (and possibly remove the checked out repo). And the initial checkout should depend on if the file exists or not.
18:55 forrest roblarkin: Did you already look at this: http://docs.saltstack.com/en/latest/topics/cloud/lxc.html#provider-configuration ?
18:56 cedwards joined #salt
18:56 Kurt___ joined #salt
18:57 roblarkin yeah, that's where I started.
18:57 roblarkin my provider is just like that
18:58 simmel I decided it was much easier to use file.managed and source: https://.. and source_hash
18:58 roblarkin but I've gone through a ton of iterations wrt profile , and how I call
19:00 rager joined #salt
19:00 roblarkin my intuition is that it's related to bootstrapping or creating the key....every path I go down results in the container name needing to be set. i.e Failed to create VM mycontainer. Configuration value 'mycontainer' needs to be set
19:01 roblarkin the circular nature of that error message has me pulling on my last hair
19:02 rager joined #salt
19:03 UtahDave joined #salt
19:03 nliadm joined #salt
19:03 jonatas_oliveira joined #salt
19:04 pr_wilson joined #salt
19:05 thayne joined #salt
19:05 waygeekierthanu Hi all. I'm trying to use salt-ssh, but I'm unsure how to use a gateway with it? Anybody know how that works? I have it working fine in my ~/.ssh/config file... anybody have any thoughts?
19:06 InAnimaTe joined #salt
19:06 aqua^mac joined #salt
19:07 nosleep77 left #salt
19:13 wincus joined #salt
19:15 rager joined #salt
19:16 linjan joined #salt
19:20 nliadm joined #salt
19:31 xenoxaos joined #salt
19:31 nliadm joined #salt
19:36 hoster joined #salt
19:39 UtahDave waygeekierthanu: I'm not sure if salt-ssh has an option to use a gateway.  Could you open a bug report on Github requesting that feature? I think that's something we should support
19:42 otter768 joined #salt
19:44 baoboa joined #salt
19:45 Ryan_Lane joined #salt
19:45 msheiny joined #salt
19:45 pr_wilson joined #salt
19:47 paha joined #salt
19:55 yorjo joined #salt
19:55 nliadm joined #salt
19:57 sebest joined #salt
19:58 bhosmer joined #salt
19:59 sebest Hi All, i'm testing the docker support and i think there is some issue with the support of docker "tag"
20:00 rager hi - how do I apply a state IFF a directory exists?
20:00 sebest it seems that most of the module/states expect the tag to be the default "latest", and things start to break if we use a custom tag
20:00 rager what I'm trying to do is remove ruby if it's been installed as root
20:00 rager (through rvm)
20:00 jdesilet joined #salt
20:00 rager there doesn't seem to be an rvm state for "rvm remove [ruby name]"
20:00 xenoxaos joined #salt
20:00 rager nor is it good to just uninstall ruby if it's not installed as root
20:01 UtahDave rager: you can use jinja to check if the directory exists to gate a section in your yaml
20:01 UtahDave sebest: which modules are failing?
20:02 rager I'll go with bash, then
20:03 rager would it be bad to just do a bash command like
20:03 zz_Cidan joined #salt
20:04 rager if [ -d "/usr/local/rvm/rubies/ruby-1.9.3-p551" ]; then rvm remove ruby-1.9.3; fi
20:04 Cidan joined #salt
20:04 UtahDave rager: it's really easy to use jinja for that.  Let me put together an example.
20:05 kballou joined #salt
20:07 UtahDave rager: https://gist.github.com/UtahDave/b106dcc802523381854a
20:07 rager ty
20:07 rager can't find much on using jinja
20:09 nliadm joined #salt
20:10 rager what about the onlyif conditional?
20:11 UtahDave yeah, you can try that, too
20:11 __gotcha joined #salt
20:12 xenoxaos joined #salt
20:13 rager onlyif: test -d [path]
20:14 sebest UtahDave: docker.pulled
20:14 sebest if i use a tag, it keeps on downloading the image
20:15 sebest the reason is that docker.pull use inspect_image()
20:15 sebest and inspect_image should be called like that inspect_image(name + ':' + tag if tag != None else 'latest')
20:16 sebest instead of inspect_image(name)
20:17 UtahDave sebest: could you open an issue on that?
20:18 Cidan joined #salt
20:18 sebest UtahDave: yes, i'll attach a possibe solution in a pull request
20:19 UtahDave sebest: that would be fantastic! thanks!
20:20 nliadm joined #salt
20:20 denys joined #salt
20:22 BigBear joined #salt
20:24 jhauser joined #salt
20:24 denys is there a rule of thumb to properly dimension the number of worker_threads?  I have about 300 minions.
20:24 ggoZ joined #salt
20:26 xenoxaos joined #salt
20:28 iggy denys: # cores?
20:28 hal58th joined #salt
20:28 iggy not sure it's worthwhile going much higher than that
20:28 thawes joined #salt
20:29 denys iggy: at present 4, but I can raise it.
20:29 darrend joined #salt
20:32 iggy I was saying that's probably the upper limit
20:34 denys iggy: right. I wish I knew for sure :-(
20:34 thawes joined #salt
20:35 iggy There's not a good rule of thumb I've heard of
20:38 sebest joined #salt
20:42 VSpike UtahDave: I was thinking of reworking salt.modules.win_service a bit more
20:43 VSpike UtahDave: there's a lot of repetition in it, and it spawns a lot of shells (which is expensive in Windows). It could be improved by refactoring, but I was actually wondering about switching to use the Win32 API directly as it will be much faster
20:44 VSpike Unfortunately, there's no way for unmanaged code to access the powershell API directly, so spawning powershell.exe and parsing the output is always going to be slow and soemwhat error-prone
20:45 VSpike Unless someone fancies getting the minion working in IronPython, but that could be a challenge. And I don't know what the long-term future of it is.
20:46 thawes joined #salt
20:47 VSpike Someone suggested using the WS-Man protocols to do powershell remoting via HTTP calls since you'd be interacting with a single instance that way. But a few minutes reading the specs plus this diagram http://msdn.microsoft.com/en-us/library/dd305175.aspx was enough to put me off even experimenting with that idea
20:47 VSpike Protocols layered two deep on top of SOAP? The horror!
20:49 UtahDave denys: with 300 minions 4 or 5 worker threads should be plenty
20:50 UtahDave VSpike: yeah, I'd love to move more of the windows modules to using the win32 apis.  we're already including most of the python windows modules
20:51 thawes joined #salt
20:52 eliasp meh… I'm getting tons of 'Failed to authenticate message' log entries on my 2014.7.0 master… around 1MB/Minute
20:52 eliasp trying now to figure out how to enhance this log message to also include the minion involved in this…
20:53 jhauser joined #salt
20:53 UtahDave eliasp: I'm seeing that, too
20:53 iggy if you do, open an issue
20:53 eliasp there are two existing issues regarding this… will either comment for re-open or open a new one… need to dig a bit further first
20:53 iggy I've seen that before and it annoys me that it's such a useless error message
20:54 Mso150 joined #salt
20:55 aqua^mac joined #salt
20:55 VSpike UtahDave: If you're happy with that as a direction, I'm happy to have a stab at it. I was just wondering how it fitted into long-term plans for Windows support in Salt, and that led me to trying to guess what they might be
20:56 VSpike The other option, which is similar but not quite the same, is to use WMI through COM
20:56 VSpike I expect like the current mix of PS and legacy cmd tools, both could have their place
20:56 eliasp the win32 API is such a PITA compared to all the nice things we have in the Python world… did some Salt/Win stuff in the past… it was painful ;(
20:57 thawes joined #salt
20:57 UtahDave VSpike: Yeah, if you want to give that a try, that would be awesome.
20:57 UtahDave we started using powershell because it has a restart method, so we could actually restart the salt-minion itself
20:57 UtahDave win32serviceutil apparently has that, too
20:57 UtahDave http://code.activestate.com/recipes/59872-manipulating-windows-services/
20:58 eliasp awesome
20:58 rager is there a way to see what states failed after doing a highstate?
20:58 rager I can't see much for all the output
20:58 VSpike Yeah the Win32 API is horrid, mostly. I don't even like Windows, but I have Windows boxes to support and it's a pain point, and I think I can ways to improve it. And I like Python, and I have a background in Windows coding that I'm mostly trying to forget these days :)
20:58 eliasp rager: you could use --state-output=terse
20:59 UtahDave rager: search in your master config for the "terse" option
20:59 VSpike So seems like something useful to do
20:59 UtahDave VSpike: well, any help would be greatly appreciated.
20:59 Katafalkas joined #salt
20:59 VSpike Cool... thanks
20:59 iggy rager: I generally change state_output to mixed
20:59 eliasp VSpike: great! I'd also love to have some win improvements in some areas, as I'm also managing all Win7 clients via Salt here
20:59 rager but there's no way to review an already-run highstate?
20:59 eliasp VSpike: did some things in the past, but there's still a lot room for improvement
21:00 BigBear joined #salt
21:00 VSpike Plenty I think, yep :)
21:00 denys UtahDave: since the upgrade to helium, we are experiencing utterly ridiculous slow downs after a while.  restarting the master helps for while. but then we fall back into "slow as molasses" mode, when practically nothing gets done.
21:00 eliasp rager: http://docs.saltstack.com/en/latest/topics/jobs/#lookup-jid
21:01 iggy ^
21:01 rager tyty
21:01 UtahDave denys: Which OSes?  could you open an issue on that?
21:01 rager crud: No such file or directory: '/var/cache/salt/master/jobs'
21:02 thawes joined #salt
21:04 denys UtahDave: ubuntu 14.(04,10).  as for opening an issue: I am reluctant; I am really not sure whether it's salt or it's us doing something stupid.
21:04 spielberg joined #salt
21:05 Mso150 joined #salt
21:07 shaggy_surfer joined #salt
21:08 saravanans joined #salt
21:09 Heartsbane joined #salt
21:09 Heartsbane joined #salt
21:12 skinner_ joined #salt
21:14 thawes joined #salt
21:15 giantlock joined #salt
21:16 sebest what is the best way to escape {{ in a yaml file? using {% raw %} ?
21:17 Cidan joined #salt
21:17 TK_ joined #salt
21:17 iggy {{ '{{' }} is another way
21:18 aurynn can I connect to the Salt event bus over tcp/ip instead of the local socket?
21:20 orion__ joined #salt
21:23 thawes joined #salt
21:24 UtahDave denys: go ahead and open an issue. Please include the output of versions_report from your master and minions and any details you think might be helpful
21:29 denys UtahDave: ok, but I need to instrument things a bit more before I do.  I don't feel right complaining without evidence (there are many issues with the network service provided by the university, and they have gotten worse)
21:30 UtahDave ok, that's cool, denys.
21:30 spielberg joined #salt
21:31 hal58th joined #salt
21:31 __gotcha joined #salt
21:34 eliasp gnaah… now I believed to have figured out where to add the required exceptions in crypt.py and master.py to get some more out of the 'Failed to authenticate message" log entries… and after restarting my master to apply the code changes, they seem to be gone :-(
21:34 eliasp let's see if I can somehow trigger this again
21:35 eliasp yeeehaw, a backtrace! :)
21:35 atbell joined #salt
21:38 thawes joined #salt
21:38 eliasp but still no way to properly reproduce it…
21:40 PWR_ joined #salt
21:41 eliasp ok… every now and then I still get the backtrace I'm able now to cause when message authentication fails… there' still hope ;)
21:43 otter768 joined #salt
21:46 alanpearce joined #salt
21:47 thawes joined #salt
21:47 JordanTesting joined #salt
21:52 SgtMalicious left #salt
21:54 patrek joined #salt
21:57 nliadm joined #salt
21:57 jaimed joined #salt
22:00 zz_Cidan joined #salt
22:00 Eliz_ joined #salt
22:01 Cidan joined #salt
22:01 shaggy_surfer joined #salt
22:03 alexr joined #salt
22:03 Katafalkas joined #salt
22:04 spookah joined #salt
22:04 kermit joined #salt
22:05 Kyle joined #salt
22:05 Eliz_ joined #salt
22:06 balltongu joined #salt
22:06 thawes joined #salt
22:07 wt joined #salt
22:07 pr_wilson joined #salt
22:07 wt hi, I am suffering with s3fs again
22:07 wt seems that the cache for the metadata reading sometimes files, and the thing that reads that data doesn't act well when that happens
22:08 wt I have a hacky patch for a fix
22:13 bhosmer joined #salt
22:16 atbell joined #salt
22:17 zz_Cidan joined #salt
22:17 Cidan joined #salt
22:17 TK_ joined #salt
22:17 nliadm joined #salt
22:22 eliasp giving up for now… I just can't get the logging right in case it breaks… simply seems to ignore my log statement… will continue tomorrow :-(
22:22 eliasp good n8
22:22 alexr joined #salt
22:24 aurynn hm
22:24 shanemhansen left #salt
22:25 aurynn so I've got an event listener program running on a minion, and it looks like I can't send a message to myself from the minion on the minion without crossing to the master?
22:30 bhosmer joined #salt
22:32 UtahDave aurynn: what command are you using to send the message?
22:32 aurynn UtahDave, on the minion, salt-call event.fire "{'a': 'woo'}" 'tag'
22:32 aurynn and on the minion, just listening for all events in a Python shell
22:33 UtahDave Hm. I was under the impression that event.fire only sent the message locally on the minion
22:33 aurynn yes
22:33 aurynn that's what the docs say, but (on the same minion) I'm never receiving the message
22:33 UtahDave so you're seeing it on the master as well?
22:33 aurynn no
22:33 aurynn I'm justn ot getting a message
22:34 UtahDave the thing listening has to be running as the same user as the salt-minion
22:34 aurynn ah
22:34 aurynn right
22:35 aurynn okay, salt-minion is running as root, and so is the pyshell
22:36 BigBear joined #salt
22:36 Ramses_ joined #salt
22:38 aurynn this is on 2014.1.13
22:39 UtahDave what are you using to listen to the events?
22:39 aurynn salt.util.events
22:39 aurynn literally, the python examples on http://docs.saltstack.com/en/latest/topics/event/index.html
22:40 UtahDave can you try using the eventlisten.py that's in the test directory of the salt git repo?
22:40 evidence joined #salt
22:40 aurynn sure
22:41 UtahDave back in a minute
22:42 patarr joined #salt
22:42 patarr joined #salt
22:42 aurynn eventlisten shows nothing
22:44 aqua^mac joined #salt
22:46 aurynn interestingly
22:46 aurynn eventlisten is reporting a different ipc handle than what exists in the socket dir
22:46 Ryan_Lane joined #salt
22:46 atbell joined #salt
22:48 spielberg joined #salt
22:51 avn joined #salt
22:51 alanpearce joined #salt
22:52 UtahDave aurynn: can you pastebin the exact python  you were using?
22:52 jonatas_oliveira joined #salt
22:53 aurynn sure
22:54 aurynn UtahDave, http://pastebin.com/VVagLLzp
22:54 aurynn it hangs endlessly
22:54 UtahDave oh, I think you have to have a loop.
22:55 aurynn iter_events() responds identically
22:55 aurynn but I think I've found it
22:55 aurynn eventlisten.py isn't listening on the same IPC socket as is created by the minion
22:57 aurynn which makes me wonder if my code is
22:58 aurynn or.. ?
22:58 aurynn I dunno
23:01 glyf joined #salt
23:01 aurynn yeah
23:02 alexr joined #salt
23:03 cads joined #salt
23:04 alexr joined #salt
23:09 TyrfingMjolnir joined #salt
23:11 aurynn so the id being wrong was the problem with eventlistener not getting anything
23:12 istram joined #salt
23:15 Jimlad joined #salt
23:15 aurynn and the same with salt.utils.event
23:16 aurynn by default it's not getting the minion ID
23:22 waygeekierthanu okay UtahDave I'll file a report for salt-ssh using a gateway. Thanks!
23:23 UtahDave aurynn: can you pastebin all that?
23:24 UtahDave waygeekierthanu: thanks!
23:24 waygeekierthanu https://github.com/saltstack/salt/issues/18540
23:24 Gareth hrm.
23:29 atbell joined #salt
23:29 atbell joined #salt
23:43 glyf joined #salt
23:44 otter768 joined #salt
23:51 joehh SeanM: if you see this, ping me - assuming you are installing via deb, I'm a little puzzled, which packages are you installing?
23:51 TheThing joined #salt
23:52 ajolo joined #salt
23:56 bhosmer joined #salt

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