Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2016-01-29

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

All times shown according to UTC.

Time Nick Message
00:00 sjmh just actually the cli.
00:00 jack__ joined #salt
00:00 sjmh https://gist.github.com/sjmh/f97129d4f49eb4a49741
00:01 invsblduck joined #salt
00:01 iggy sure it's not just your shell "fixing" things for you?
00:01 totzky yeah, although most users don't really care about what's happening underneath, i've got to the point where i debugged the code and saw it was yamlifying the args which is counterintuitive
00:02 totzky @iggy, i've stepped in the salt cli code and its the yamlifying part which is stripping off the quotes
00:03 totzky AFAIK here's the input parsing flow:  salt optparse => yamlify => pass to module func
00:03 sjmh does escaping the quotes work for the boto module?
00:03 sjmh ie - salt \*dev00\* test.echo '\"help\"'
00:04 sjmh boto module may not like it
00:04 totzky @sjmh the problem is that the yamlify part unexpectedly transforms the arg
00:04 sjmh well, with escapes, it returns \"help\"
00:04 totzky @sjmh ok your example will result to -> help
00:04 sjmh but if that got passed to boto, not sure boto is goin to escept that
00:04 sjmh totzky - no it doesn't
00:04 sjmh it results to \"help\"
00:05 sjmh https://gist.github.com/sjmh/2debe07b4e114c4e4c40
00:05 totzky @sjmh yamlify_arg('\"hep\"') will result to 'help'
00:06 woodtablet joined #salt
00:06 totzky i see let me try that echo thing is handy!
00:06 woodtablet hey guys, how do i ask the salt-minion to send its key to the master ?
00:07 jtylers joined #salt
00:09 L4nce joined #salt
00:09 evidence woodtablet: it just does that on initial connect
00:09 L4nce Hi all
00:09 djgerm it should when you start the service
00:09 om2 anyone seen this on windows minions?
00:09 om2 Module function system.join_domain threw an exception. Exception: argument of type 'NoneType' is not iterable
00:09 djgerm and if it doesn't, check the log. it should say why
00:09 brianfeister joined #salt
00:10 om2 woodtablet: make sure your /etc/salt/minion has the master: set
00:10 om2 then restart salt-minion
00:10 om2 on the minion of course
00:10 om2 then on master do:  salt-key to list it
00:11 om2 minion by default doesn't know what the master is, you have to tell it where it is by IP or resolving DNS
00:11 om2 gosh, windows is a mare
00:12 woodtablet evidence & om - must be a HW firewall issue then.
00:12 om2 keep getting this:
00:12 woodtablet thanks guys =D
00:12 om2 Module function system.join_domain threw an exception. Exception: argument of type 'NoneType' is not iterable
00:12 L4nce I would like to know if I can run an exec module from a state then use the data returned in my state
00:12 om2 try to telnet to master  TCP salt ports
00:12 om2 that will tell ya
00:13 woodtablet om2 - i can from some, and i cant from others.. must be the hw restrictions. i ll go ask them to open that baby up
00:13 woodtablet thanks guys -D
00:14 totzky @sjmh - thanks, yes you are correct but boto didn't like the result
00:14 sjmh yeah i figured it probably wouldn't like it
00:15 L4nce I can't find a doc about this
00:16 L4nce I would like to call network.get_route
00:16 L4nce then use the source IP returned in a template file
00:17 jack__ joined #salt
00:19 sjmh L4nce - you have access to the exec modules via jinja.
00:19 sjmh in the __salt__ dunder dir.
00:19 gimpy927 left #salt
00:19 L4nce ho, I'm going to check that
00:20 djgerm when I have a thing like "    - source: salt://binddns/files/named.conf.options" does that mean /binddns/ under the file_root in the master config?
00:20 robertsonai joined #salt
00:21 sjmh totzky - looks like you don't have much of a choice.  maybe throw up an issue on github about it and ask for some flag to turn off the yamlfying via a cmd switch
00:21 L4nce Many thanks sjmh
00:21 sjmh L4nce - sure
00:22 totzky @sjmh thanks a lot!
00:22 sjmh djgerm - it should, depending on how you have your file_root setup.
00:22 sjmh djgerm - depends on if you are using environments and such
00:23 djgerm oh just base environment
00:23 sjmh if you just have base and it's set to /srv/salt, then that file should be in /srv/salt/binddns/files/named.conf.options
00:24 djgerm ok thanks!
00:26 rem5 joined #salt
00:26 rem5 joined #salt
00:33 [7hunderbird] joined #salt
00:35 yomilk joined #salt
00:35 jack__ joined #salt
00:39 om2 joined #salt
00:40 cberndt joined #salt
00:42 zmalone joined #salt
00:44 lompik joined #salt
00:47 kevinquinnyo1 joined #salt
00:52 bhosmer_ joined #salt
00:54 abednarik joined #salt
00:58 bhosmer_ joined #salt
01:13 notnotpeter joined #salt
01:13 yomilk joined #salt
01:18 woodtablet left #salt
01:26 baweaver joined #salt
01:30 _JZ__ joined #salt
01:31 stooj joined #salt
01:34 keekz joined #salt
01:37 berserk joined #salt
01:43 mrtrosen joined #salt
01:47 iceyao joined #salt
01:52 bhosmer_ joined #salt
01:57 zmalone joined #salt
02:01 catpiggest joined #salt
02:08 donmichelangelo joined #salt
02:12 rem5 joined #salt
02:23 yomilk joined #salt
02:27 rem5 joined #salt
02:31 iceyao_ joined #salt
02:32 jack__ joined #salt
02:37 racooper joined #salt
02:38 totzky joined #salt
02:45 akhter joined #salt
02:48 onlyanegg joined #salt
02:51 JDiPierro joined #salt
02:58 onlyanegg joined #salt
02:59 yomilk joined #salt
02:59 catpigger joined #salt
03:07 jack__ joined #salt
03:14 jack__ joined #salt
03:14 anmolb joined #salt
03:15 brianfeister joined #salt
03:23 mcfallen left #salt
03:26 ajw0100 joined #salt
03:26 JDiPierro joined #salt
03:41 evle joined #salt
03:41 invsblduck joined #salt
03:42 totzky joined #salt
03:45 writtenoff joined #salt
03:46 invsblduck joined #salt
04:03 pcn in salt-cloud what's the difference in an ec2 profile between block_device_mappings and volumes?  The documentation shows them both, but it's very ahrd to suss out the intended configuration or behavior for them
04:09 jack__ joined #salt
04:12 treaki joined #salt
04:17 voidspacexyz joined #salt
04:18 Sen_ joined #salt
04:19 jack__ joined #salt
04:34 jack__ joined #salt
04:35 anmolb joined #salt
04:49 HoloIRCUser joined #salt
04:57 brianfeister joined #salt
05:11 jack__ joined #salt
05:29 ruxu joined #salt
05:32 jirwin left #salt
05:38 brianfeister joined #salt
05:47 impi joined #salt
05:56 jack__ joined #salt
05:58 calvinh joined #salt
06:02 writtenoff joined #salt
06:02 DanyC joined #salt
06:04 DanyC joined #salt
06:05 favadi joined #salt
06:08 DanyC joined #salt
06:09 lemur joined #salt
06:10 baweaver_ joined #salt
06:13 DanyC Hi, i am trying to secure Salt Master and so is any way i can set some permissions such that OS/ Linux users can highstate/ run cmd on a pool of minions?
06:14 DanyC i'm pretty sure is not possible but maybe someone found a different solution
06:23 cdog joined #salt
06:24 cdog anyone here
06:24 cdog left #salt
06:24 om2 yes.  but it varies by time
06:25 om2 DanyC: you might need enterprise UI for something like that... not sure
06:26 DanyC om2: never thought about that, i might look into what it offer then. I was more looking from/ while i'm ssh'ed into salt master ;)
06:26 rotbeard joined #salt
06:27 beardedeagle joined #salt
06:39 jack__ joined #salt
06:45 malinoff joined #salt
06:46 malinoff joined #salt
06:46 linjan joined #salt
06:47 yomilk joined #salt
06:55 jhauser joined #salt
07:00 akhter joined #salt
07:01 yomilk joined #salt
07:02 giantlock joined #salt
07:09 colttt joined #salt
07:10 dlam joined #salt
07:11 dlam hmm anyone know how to debug why a computer isnt matching a pattern in top.sls ?
07:12 dlam i have 'app*':  in it, and the box's `hostname` is app    ...but it aint matching or something
07:12 voidspacexyz Ok, I am facing this weird error, I am trying to install some python requirements from the requirements.txt file.  If my package name contains a * , like "uwsgi==2.0.*". I get an error saying  "ValueError: ("Expected ',' or end-of-list in", 'uwsgi==2.0.*', 'at', '*')"
07:14 AndreasLutro where did you read that 2.0.* works with pip?
07:16 voidspacexyz Ah, how did I miss that
07:17 voidspacexyz Its a pip issue
07:17 voidspacexyz sorry, have been straight up for 24 hrs, and just didnt think
07:17 bhosmer joined #salt
07:18 peters-tx joined #salt
07:19 sjmh dlam - can you paste the whole top.sls somewhere?
07:19 sjmh or just the issue part
07:20 sjmh AndreasLutro - i submitted a PR for that yield issue in the log4mongo stuff.  After reading it, it was just a bad use of yield.
07:21 AndreasLutro sjmh: must be a reason it's used though, yield is also used in all the other log handlers
07:21 sjmh yeah, not sure why they yield instead of returning - i haven't looked at the handler code stuff.
07:21 sjmh but obviously they want set_handlers as a generator.
07:22 sjmh and they call it multiple times.
07:26 jack__ joined #salt
07:35 frewtc4ke joined #salt
07:36 jack__ joined #salt
07:38 dgutu joined #salt
07:38 totzky joined #salt
07:39 patrek joined #salt
07:40 Rumbles joined #salt
07:44 brianfeister joined #salt
07:45 frewtc4ke left #salt
07:45 jack__ joined #salt
07:54 impi joined #salt
07:57 Knuta dlam: check what it is configured to use as the node name. It might be different from the hostname.
08:00 Knuta dlam: check the content of /etc/salt/minion_id. That's what it's identifyint itself as.
08:03 tmclaugh[work] joined #salt
08:04 blckbit10 joined #salt
08:04 babilen dlam: Also take a look at the output of "salt-key -L"
08:06 jack__ joined #salt
08:11 bhosmer joined #salt
08:13 rubendv joined #salt
08:13 rubendv joined #salt
08:20 viq joined #salt
08:30 impi joined #salt
08:35 ruxu joined #salt
08:36 voidspacexyz joined #salt
08:37 dgutu joined #salt
08:42 guanophobic joined #salt
08:42 LondonAppDev joined #salt
08:44 jack__ joined #salt
08:49 iceyao joined #salt
08:51 ianmac1966 joined #salt
08:52 yomilk joined #salt
09:00 fredvd joined #salt
09:01 jack__ joined #salt
09:03 s_kunk joined #salt
09:03 slav0nic joined #salt
09:04 s_kunk joined #salt
09:05 GreatSnoopy joined #salt
09:06 bhosmer_ joined #salt
09:06 amcorreia joined #salt
09:08 keimlink joined #salt
09:11 Rumbles joined #salt
09:13 jack__ joined #salt
09:21 losh joined #salt
09:21 geomacy joined #salt
09:23 mariusv joined #salt
09:25 LondonAppDev joined #salt
09:26 anmol joined #salt
09:32 brianfeister joined #salt
09:33 lothiraldan joined #salt
09:36 fredvd joined #salt
09:41 sesa joined #salt
09:52 dlam ahh thanks Knuta   ...it says  "ip-172-31-23-64.us-west-2.compute.internal"  :O
09:52 dlam me go write that tip down
09:53 dlam i like made a new EC2 server, then manually changed it a la `sudo hostname app`
09:53 dlam (guess maybe that not the way to do it)
09:55 brianfeister joined #salt
10:00 bhosmer joined #salt
10:01 blckbit10 joined #salt
10:05 aarontc joined #salt
10:06 kawa2014 joined #salt
10:16 jaybocc2 joined #salt
10:18 jack__ joined #salt
10:21 yomilk joined #salt
10:24 voidspacexyz joined #salt
10:26 amcorreia joined #salt
10:27 baweaver joined #salt
10:28 yomilk joined #salt
10:29 giantlock joined #salt
10:33 yomilk joined #salt
10:33 totzky joined #salt
10:43 krymzon joined #salt
10:56 jack__ joined #salt
10:56 N-Mi joined #salt
10:56 N-Mi joined #salt
10:56 akhter joined #salt
10:57 interpeix joined #salt
11:07 bluenemo joined #salt
11:25 treaki joined #salt
11:37 totzky joined #salt
11:37 voidspacexyz I was looking at this, https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.djangomod.html.  Just not clear how to mention the settings_module parameter in this ?
11:38 XmodeEmonX joined #salt
11:39 XmodeEmonX Hi there. Can someone tell me if I should be able to use pillar information within a cmd: state definition? As in cmd: \n echo {{ pillar['somevalue'] }} as it doesn't seem to pick it up.
11:39 XmodeEmonX errors with dict object has no attribute 'somevalue'
11:39 AndreasLutro XmodeEmonX: yes, it's possible, the pillar just isn't defined for the minion
11:40 XmodeEmonX AndreasLutro: hmmm. I wish I knew exactly what you meant.
11:41 colegatron_origi Hi
11:41 colegatron_origi I am using saltstack to deploy the servers software but also deploy the websites/apps of my company. is that a typical use case or people uses salt/CM only to deploy server's software?
11:41 PeterO_ joined #salt
11:41 XmodeEmonX AndreasLutro: Do you mean I need to conduct a salt '*' saltutil.refresh_pillar first?
11:42 AndreasLutro XmodeEmonX: maybe, or maybe you just haven't added the pillar data "somevalue" to the minion
11:44 jack__ joined #salt
11:44 CeBe joined #salt
11:45 guanophobic joined #salt
11:45 PeterO_ Is it possible to use Jinja templating in a top.sls? Something like {% if grains['os'] == 'RedHat' %}
11:47 XmodeEmonX AndreasLutro: Thanks, I'll have a play :-)
11:48 bhosmer joined #salt
11:50 XmodeEmonX AndreasLutro: Aha! Yes, that works. *shakes head at self embarrasingly*
11:54 favadi joined #salt
11:54 evle joined #salt
11:56 LondonAppDev Has anyone experience an issue with apache not restarting even if I have - restart: True in my salt state?
12:05 tmclaugh[work] joined #salt
12:05 denys joined #salt
12:06 PeterO_ no doesn't appear Jinja templating will work in top.sls
12:07 ruxu joined #salt
12:07 _JZ_ joined #salt
12:13 jack__ joined #salt
12:17 kbaikov joined #salt
12:18 eliasp PeterO_: what about targeting instead ('G@os:RedHat') in top.sls?
12:19 jaybocc2 joined #salt
12:19 PeterO_ I need to do some not equals, is that possible? Basically I need " 'G@os:Amazon' and not 'G@region:us-west-2' "
12:20 totzky joined #salt
12:23 jack__ joined #salt
12:23 ITChap joined #salt
12:25 PeterO_ oh brother.. I just found the docs..
12:25 PeterO_ It literally is "and not"
12:31 PeterO_ https://docs.saltstack.com/en/latest/topics/targeting/compound.html
12:31 PeterO_ Not sure why I couldn't find that doc earlier..
12:33 linovia Is salt python3 compatible ?
12:34 avozza joined #salt
12:34 jack__ joined #salt
12:38 malinoff joined #salt
12:42 bhosmer_ joined #salt
12:49 quasiben joined #salt
12:49 ianmac1966 Help with an error -  I'm new to salt and was hoping someone can point me in the right direction
12:49 ianmac1966 salt-call --local state.highstate -c . test=true
12:50 ianmac1966 [INFO    ] Loading fresh modules for state activity
12:50 ianmac1966 [INFO    ] Fetching file from saltenv 'base', ** done ** u'top.sls'
12:50 ianmac1966 [ERROR   ] Unable to render top file: Illegal tab character
12:50 ianmac1966 local:
12:50 ianmac1966 ----------
12:50 ianmac1966 ID: states
12:50 ianmac1966 Function: no.None
12:50 ianmac1966 Result: False
12:50 ianmac1966 Comment: No Top file or external nodes data matches found.
12:50 ianmac1966 Started:
12:50 ianmac1966 Duration:
12:50 ianmac1966 Changes:
12:53 jack__ joined #salt
13:00 jack__ joined #salt
13:01 Steven- joined #salt
13:03 akhter joined #salt
13:09 morissette joined #salt
13:09 jack__ joined #salt
13:27 jack__ joined #salt
13:28 av_ joined #salt
13:29 lothiraldan joined #salt
13:34 kawa2014 joined #salt
13:34 jaybocc2 joined #salt
13:37 bhosmer_ joined #salt
13:40 gh34 joined #salt
13:43 jack__ joined #salt
13:44 CrummyGummy Hi, I have my salt-master configured to get it's top file from git. It's stopped complaining so I assume it's connecting ok. How do I make sure that it's checking the git repo for updates?
13:44 CrummyGummy I've committed a top file. It doesn't seem to see it.
13:44 geomacy joined #salt
13:44 CrummyGummy 'I've committed a top file. It doesn't seem to see it.'
13:44 CrummyGummy ^ sorry
13:45 CrummyGummy "Comment: No Top file or external nodes data matches found."
13:45 babilen CrummyGummy: You could run "salt-run fileserver.update"
13:45 babilen It doesn't check *all the time*, but every 60s IIRC
13:45 CrummyGummy I pushed yesterday though
13:46 babilen Then we need more information (such as the command that you ran, its output, the top.sls, ...) -- paste it to one of http://refheap.com, http://paste.debian.net, https://gist.github.com, http://sprunge.us, … please
13:47 ianmac1966 joined #salt
13:47 CrummyGummy ok
13:49 CrummyGummy https://gist.github.com/anonymous/1c555a5ffa6d43706392
13:49 jack__ joined #salt
13:51 CrummyGummy babilen: I hope that's enough data
13:53 JDiPierro joined #salt
13:54 babilen CrummyGummy: I'll take a look shortly
13:54 illern joined #salt
13:55 CrummyGummy thanks
13:56 subsignal joined #salt
13:57 disbound joined #salt
13:58 edrocks joined #salt
13:58 subsigna_ joined #salt
14:01 tmclaugh[work] joined #salt
14:04 babilen CrummyGummy: Okay, that should have worked. What does "salt '*' state.show_top" give you?
14:04 babilen And "salt '*' cp.list_master"
14:06 CrummyGummy https://gist.github.com/anonymous/7edf504434b463a0a132
14:08 babilen Where are "nginx", "vim" (and so on) ?
14:08 babilen It would probably also not be a bad idea to start the salt-master manually in debug mode (salt-master -ldebug) and to check if that shows any problem. Anything in /var/log/salt/master ?
14:09 CrummyGummy I haven't made the vim/nginx etc. Just trying to confirm that it's working to this point first.
14:10 CrummyGummy Am I getting what I should be getting?
14:11 CrummyGummy I expected an error like, nginx not found, not 'no top file'
14:16 Crypticus joined #salt
14:17 JDiPierro joined #salt
14:17 jack__ joined #salt
14:21 thejrose1984 joined #salt
14:21 cpowell joined #salt
14:22 babilen CrummyGummy: It's essentially telling you that no states have been targeted to that box
14:23 babilen s/box/minion
14:23 babilen Why don't you start with a single, existing, SLS first and target that?
14:23 dyasny joined #salt
14:25 CrummyGummy Thanks, I'm doing that atm. cleaning out and implementing a simple vim pkg install.
14:27 blckbit10 joined #salt
14:30 babilen Sounds like a good idea
14:30 CrummyGummy same outcome
14:30 babilen Okay, check the master log and output
14:31 bhosmer_ joined #salt
14:31 babilen Is the additional file now listed in "salt '*' cp.list_master"
14:31 CrummyGummy nothing of interest. I'll make another pastebin now.
14:32 CrummyGummy no
14:32 babilen Did you run "salt-run fileserver.update" before trying?
14:32 giantlock joined #salt
14:32 PeterO joined #salt
14:33 CrummyGummy I did now, no change. I can see in the debug log that it checks git every minute or so aswell.
14:34 perfectsine joined #salt
14:35 babilen And the file is still not listed in salt '*' cp.list_master ?
14:36 CrummyGummy no :(
14:36 CrummyGummy https://gist.github.com/anonymous/5e16de930626d057d10e
14:36 CrummyGummy ugg, that may have been my messup. Commit fail
14:39 CrummyGummy no, it definitely is committed.
14:39 tracphil joined #salt
14:41 antpa joined #salt
14:41 antpa joined #salt
14:42 irctc524 joined #salt
14:42 spiette joined #salt
14:43 irctc524 Looking for some documentation on how to implement `watch_in` and `require_in` on a custom state
14:43 tracphil joined #salt
14:44 Tyrm joined #salt
14:45 racooper joined #salt
14:45 irctc524 closest thing I've found is the source for `/salt/states/file.py` `def accumulated(name, filename, text, **kwargs)`
14:45 tracphil joined #salt
14:46 CrummyGummy babilen: I've figured it out. I need to run update-server-info on the git repo for the http downloads to work. Thanks for your time.
14:46 geomacy joined #salt
14:47 babilen Great
14:47 babilen Which "git repo" is that?
14:47 anthpa joined #salt
14:48 babilen Ah .. you might want to use git://
14:48 CrummyGummy the one that I have my top file in. My own installation. Not sure why I have to run that command at all though.
14:49 CrummyGummy I don't think I have the git protocol setup, only ssh over git and a bad http interface. I definitely need do work on that.
14:51 babilen ssh is not a bad option
14:51 avozza joined #salt
14:52 zmalone joined #salt
14:52 CrummyGummy ja, I tried that first. It needs py2git or something, that's dependency hell to install on my LTS ubuntu installation.
14:52 anthpa Would anyone be able to help me create a state file to deployment multiple vhost nginx configuration files for 3 domains to one minion?  My current saltstack config is set up to manage single domain names on individual minions but a client's requirement is forcing me to create something custom.
14:53 mapu joined #salt
14:53 CrummyGummy anthpa: As a local n00b I'll be making notes.
14:53 CrummyGummy that's also in my todo list.
14:54 anthpa CrummyGummy: nice.  i've spent the better part of the last 3 months setting up saltstack to manage our client's servers.  I have almost all of them converted... Just working on the weird one's now...
14:56 CrummyGummy uggg, I hope it doesn't take me that long (no offence)
14:56 CrummyGummy not that I'm doing that well atm. Most of my problems have been git related though.
14:56 DammitJim joined #salt
14:57 anthpa CrummyGummy: I missed your initial question.  Maybe I can help.
14:58 CrummyGummy anthpa: I'm good thanks, I had to update my git repo to run update-server-info after commit. Winning finally.
14:59 CrummyGummy (for now)
14:59 rm_jorge joined #salt
15:01 jack__ joined #salt
15:01 CrummyGummy fixed, yay
15:01 andrew_v joined #salt
15:01 CrummyGummy cheers anthpa babilen I'm off to Friday
15:01 babilen Enjoy!
15:02 babilen anthpa: https://github.com/saltstack-formulas/nginx-formula might come in handy
15:02 hal58th joined #salt
15:02 giantlock joined #salt
15:04 irctc524 Looking for some documentation on how to implement `watch_in` and `require_in` on a custom state. Closest thing I've found is the source for `/salt/states/file.py` `def accumulated(name, filename, text, **kwargs)` but, something a little more clear would be great.
15:04 amcorreia joined #salt
15:09 sroegner joined #salt
15:14 KennethWilke joined #salt
15:18 PeterO joined #salt
15:19 XmodeMunZ joined #salt
15:19 PeterO joined #salt
15:20 XmodeMunZ Hi all. Anyone suggest the best way to extract an array of information held in a pillar? Currently the use of pillar['addresses'] results in square brackets and comma delimits. I just want to use the comma delimits.
15:20 jack__ joined #salt
15:21 antpa joined #salt
15:21 tmclaugh[work] joined #salt
15:22 XmodeMunZ I used a for loop that works, but thought there would be a nicer method direct form the pillar method.
15:23 asco1aro joined #salt
15:25 antpa babilen: nginx is being installed perfectly.  If you'd like to see what I have so far, https://gist.github.com/anthonyscolaro/46c1633e88f3322a93f5.  The problem is that my current nginx config is using the domain to create the server name... I need to find a way to pull the server name from pillar in a different way.  Like some kind of lookup.
15:25 mrwboilers left #salt
15:26 babilen antpa: And https://github.com/saltstack-formulas/nginx-formula/blob/master/pillar.example#L88 is not to your liking in that regard?
15:27 Tyrm joined #salt
15:28 antpa babilen: It's working perfectly if I am hosting a single website on a single webserver.  Now that I'm being asked to host 3 websites on a single server, I am hoping to use some kind of lookup.  For instance, I could specify 3 domain names within pillar somewhere and then have the nginx state deploy 3 different vhost files, each with a different server name. Currently it's static to the server's domain name.
15:29 babilen Yes, isn't that what the formula does?
15:29 jack__ joined #salt
15:31 antpa babilen: I don't see it but I'll look again.
15:32 babilen https://github.com/saltstack-formulas/nginx-formula/blob/master/nginx/ng/files/vhost.conf + https://github.com/saltstack-formulas/nginx-formula/blob/master/nginx/ng/vhosts_config.sls
15:33 kawa2014 joined #salt
15:35 jaybocc2 joined #salt
15:37 Tyrm joined #salt
15:37 dgutu joined #salt
15:37 Tyrm joined #salt
15:39 berserk joined #salt
15:40 numkem joined #salt
15:42 antpa babilen: Thank you!
15:42 AdamSewell joined #salt
15:44 iceyao joined #salt
15:54 andrew_v_ joined #salt
15:57 bowhunter joined #salt
15:59 ruxu joined #salt
16:00 jack__ joined #salt
16:00 spiette joined #salt
16:04 antpa_ joined #salt
16:04 anthpa joined #salt
16:06 onlyanegg joined #salt
16:11 beardedeagle joined #salt
16:11 PeterO joined #salt
16:12 lloesche joined #salt
16:14 StolenToast I think this ntp formula is all sorts of whacky
16:15 StolenToast it should just be plug-n-play but it doesn't work out of the box and I think the pillar references don't reflect pillar.example
16:17 troubled joined #salt
16:18 avozza joined #salt
16:18 troubled Heya. I was just curious if anyone had any advice on a salt state that could enumerate all files in a specified directory tree and look for a missing .md5 (or similar) file, generating one if missing
16:19 troubled I was thinking of just doing a shell/py/perl/etc script for this, or possibly a py module, but thought I should ask first in case there is any easy way already in salt that I am not considering
16:19 bhosmer_ joined #salt
16:19 jack__ joined #salt
16:21 onlyanegg joined #salt
16:21 nidr0x joined #salt
16:23 mohae joined #salt
16:24 soma joined #salt
16:27 Brew joined #salt
16:29 hasues joined #salt
16:29 hasues left #salt
16:30 jack__ joined #salt
16:31 zsoftich1 joined #salt
16:33 winsalt joined #salt
16:34 ruxu joined #salt
16:36 gladiatr joined #salt
16:36 berserk joined #salt
16:40 antpa joined #salt
16:41 antpa__ joined #salt
16:44 gladiatr joined #salt
16:44 jack__ joined #salt
16:46 snicers-work joined #salt
16:47 snicers-work I have updated my cloud map file and would like to ship those changes to my EC2 instances, in this case I have added 2 minions to my production environment.
16:48 voidspacexyz joined #salt
16:52 gladiatr greetings all.  I’m working on providing tests with my pull requests (last batch was my first time meeting this goal!).  I’ve got a patch set for the jobs runner (jobs.list_jobs search_target=thing is currently broken) but to write proper initegration tests I need to be able to generate some properly formed job data to read/process/return for the test. What is the established way to do this?  The idea I have is to spin up a couple
16:52 gladiatr additional minion processes, pass them some test functions and, run the tests, assert that things are (or are not) as they should be and return a result.  What I’m not finding (I have *not* exhaustively reviewed all the available test functions) is how to do this.  Is this even the best way to go for this sort of test?  There are very few runner tests to compare against—I would like to help change that :)
16:53 StolenToast I can't seem to query pillar data using pillar.item(s)
16:54 StolenToast the response is simply empty, though pillar data is definitely used within states
16:54 iggy gladiatr: you've now stumbled upon the problem with testing something like salt :/ It sucks, it's not an easy problem to solve, I don't think there is a consistent way to solve it either
16:54 sjmh joined #salt
16:55 iggy StolenToast: then your pillar data probably isn't targeted correctly
16:55 StolenToast but shouldn't pillar.items show every bit that's exposed to my selected minion?
16:56 KennethWilke joined #salt
16:56 gladiatr StolenToast, also, check your salt-master logs for any sort of stacktrace connected to pillar processing. In certain cases, if processing pillar fails you can end up with an empty dictionary being returned to your minions
16:56 snicers-work How do I bring up a minion w/ a modified version of the /etc/salt/minion
16:57 StolenToast that's what I thought, maybe an error in the salt file, but states the rely on salt data seemed to continue to run
16:57 StolenToast *in the pillar file
16:57 StolenToast and... on *pillar data
16:58 StolenToast hmm, no pillar errors in the log
16:59 gladiatr iggy, I hear you. I guess it’s a matter of enough testing-infrastructure things getting built that can be re-used without too much pain.  I’ve got some decent ipython (run-these things on startup) profiles for local testing salt-master/minion code, but I’m still trying to get my head around how to translate these things into pyunit
17:00 akhter joined #salt
17:02 mpanetta joined #salt
17:04 jack__ joined #salt
17:06 geomacy joined #salt
17:06 gladiatr Is raet still being worked on?
17:06 jaybocc2 joined #salt
17:07 sighten joined #salt
17:07 akhter joined #salt
17:10 sighten Hi - Anyone around to answer a quick question about gpg/pillar best practices?
17:12 racooper joined #salt
17:12 gladiatr sighten, what’s your plan?
17:13 zmalone gladiatr: I don't think it's seeing much development any more
17:13 sighten Cool, thanks - basically, I've got it working as documented on a master/minion setup where master has the private key
17:14 sighten Just wondering how rendering would work on a masterless configuration
17:14 zmalone sighten: I started adopting it, but the limits on why you can encrypt (no special characters), the pain of encrypting every pillar value separately, random bugs, and the "all decryption happens on the master" stuff made me dump it
17:14 akhter_1 joined #salt
17:15 zmalone it was a lot of work and it only protected the secrets at rest in my salt pillar repo
17:16 sighten I can certainly see that. Did you wind up pre-encrypting the data somehow?
17:16 PeterO joined #salt
17:16 sjmh gladiatr - from what i heard, it wasn't solving all they thought it would, so it wasn't a priority.
17:16 sjmh (anymore)
17:16 zmalone I didn't, I'm waiting on external pillars in an environment with a salt master to adopt Vault, or another similar tool.
17:17 zmalone sjmh: And they decided to go with Tornado, and promptly ran into more problems
17:17 gladiatr sjmh, zmalone: ah, that’s too bad.  It sounded like a neat design idea.
17:17 sjmh zmalone - sure, async is hard.
17:18 sjmh gladiatr - i don't think it's dead, just think they are concentrating on other things first.
17:18 zmalone I don't think that transporting data is why Salt has problems, but I can't fault anyone for working on what they are interested in.
17:18 sjmh like, bug fixes.
17:19 gladiatr zmalone, I thought its primary motivation was to dodge the limits the per-IP tcp connection limits—get away from the handshake overhead, etc…
17:20 jfindlay gladiatr: the current test runner is very limited in the way that it uses salt daemons, and to make the changes you need would involve some effort
17:20 Hetman Hello :-) Hello, anyone can advise on that problem: http://pastebin.com/y2gVjW2J
17:20 jfindlay gladiatr: yes, the plan with RAET is to port it to tornado
17:21 zmalone Hetman: looks like an unneeded space after "rhscl-"
17:21 gladiatr jfindlay, I’m having a look at the TestDaemon code.  I don’t mind putting in the work—I just wanted to be sure that I wasn’t looking into solving an already-solved thing (nod)
17:21 berserk joined #salt
17:22 jfindlay gladiatr: also, I and some of the other QA/automation engineers are working on a new testing infrastructure that uses salt rather than python to create tests
17:22 jfindlay gladiatr: not that I know of
17:22 gladiatr jfindlay: nice!
17:23 jfindlay so that, in a simple sense, a test is sls data, which should hopefully be much more accessible to a new user than salt-quirked python, or worse salt-testing-quirked python :-)
17:23 jack__ joined #salt
17:24 zmalone I run into enough cases where my bugs end up being with salt, that I do all my testing from a non-Salt testing framework :/
17:24 onlyanegg joined #salt
17:24 dlam joined #salt
17:24 zmalone right now I'm using inspec for integration tests
17:25 gladiatr jfindlay: testing seems to be its own form of Gordian knot—to test things you have to write code to provide the environment that can prove the code’s functional validity but then you have to make sure that you can (provably) provide those services with that code.
17:25 tristianc joined #salt
17:28 jfindlay zmalone: the basic idea is to run the tests in a cmd.run state over salt-ssh so that the resulting data can be returned to py-test or something similar.  The salt-ssh connection is meant to be kept as simple as possible, whereas salt breaking while running the test, is the idea :-)
17:29 writtenoff joined #salt
17:29 berserk joined #salt
17:30 terratoma joined #salt
17:30 snicers-work Is there a way for me to configure the minion port on salt-cloud creation?
17:30 jim__ joined #salt
17:31 jfindlay snicers-work: if you use a cloud map, you can specify any minion config
17:31 jim__ I'm looking for some help with the official node-formula
17:31 jim__ it's asking about the npm pkg
17:32 jack__ joined #salt
17:33 snicers-work jfindlay, like this? https://gist.github.com/anonymous/a281c73806db9882968c
17:34 giantlock joined #salt
17:34 Tanta joined #salt
17:34 jfindlay snicers-work: yes
17:34 snicers-work Nice. Thanks.
17:35 snicers-work Do you know how to update a cloud provider with your new environment map?
17:35 snicers-work jfindlay, ^
17:35 jfindlay snicers-work: `salt-cloud -m /path/to/mapfile -P`
17:36 snicers-work It will just handle it eh, that's nice.
17:36 snicers-work I will have to test that for hardness.
17:42 rhys joined #salt
17:43 berserk joined #salt
17:45 soma joined #salt
17:45 jack__ joined #salt
17:48 armyriad joined #salt
17:49 evle joined #salt
17:58 snicers-work jfindlay, can I put the minion config in my profile as well instead of the map?
17:59 snicers-work I have the config key and value I want to change but it doesn't seem to reflect when I do my salt-cloud -m
18:01 jfindlay snicers-work: I think so
18:01 jfindlay I forgot about that one, sorry
18:01 snicers-work It is strange, according to that my salt minion should be configured correctly but is not.
18:02 ericof joined #salt
18:02 snicers-work does this look right to you jfindlay https://gist.github.com/anonymous/400f670aafa9a4c70164
18:07 JDiPierro joined #salt
18:07 jfindlay snicers-work: I've never used extends before, but it looks correct to me.  Is there a reason you have quoted the master_port?  Also, regarding `start_action`: https://github.com/saltstack/salt/issues/23873
18:07 saltstackbot [#23873]title: deprecate/remove start_action from salt-cloud, document using reactor instead | `start_action` is an experimental feature that has never worked and is superseded by better functionality in reactor.
18:07 bhosmer_ joined #salt
18:09 snicers-work Gotcha
18:09 akhter joined #salt
18:12 JDiPierro joined #salt
18:14 jrgochan Hey guys. Anyone have any luck getting the ext_pillar configuration for the master to work using pygit2 over ssh?
18:14 Pixionus so this may be a stupid question but, when highstating from the master what's the best way to record the results?  I mean it time out's invariably since our highstates on production tend to be somewhat large, so use a huge time out, an async and out-file?  I don't know.  I'm still pretty ignorant here.
18:14 jrgochan gitfs_remotes is working just fine, but ext_pillar keeps crashing out
18:15 ktosiek joined #salt
18:18 shoemonkey joined #salt
18:18 HardWall joined #salt
18:19 jrgochan pleaaaase ignore my previous request. I'm always forgetting about them permissions
18:19 jrgochan it's working now
18:20 jack__ joined #salt
18:20 jfindlay Pixionus: use a returner: https://docs.saltstack.com/en/latest/ref/returners/all/index.html#all-salt-returners
18:22 blckbit10 joined #salt
18:24 berserk joined #salt
18:25 Pixionus jfindlay: thanks this looks way more useful than an out-file
18:25 Pixionus do I need to still assign a crazy long timeout
18:25 Pixionus ?
18:25 Rockj joined #salt
18:26 jfindlay Pixionus: I don't think so.  There are details that I am unfamiliar with, but the minions should return to the master when they're done, no matter how long it takes
18:26 Pixionus Normally when I check my job by job id it has no return from the minions, but the minions finished highstating
18:28 jfindlay I think the timeout you specify on the command line alters the time the `salt` command waits for a response to a query to the minion as to whether the job is still going
18:28 jfindlay but I could be wrong
18:28 snicers-work jfindlay, I just found out that changing the minion name in the environment looses scope on the original minion.
18:29 snicers-work This seems like a problem, how would you handle something like that.
18:29 jfindlay snicers-work: I'm unsure what you mean
18:30 blckbit10 joined #salt
18:30 snicers-work I changed 'Salt-Minion' to 'Salt-Minion1' and when I did `salt-cloud -m` it spun up a new instance and left the old one running.
18:30 snicers-work I was hoping for a rename, or removal process.
18:32 snicers-work jfindlay, also it doesn't seem that the cloud.profiles.d file is giving the minion the right /etc/salt/minion file with my specified port and master address...
18:32 rhys joined #salt
18:32 ageorgop joined #salt
18:33 jack__ joined #salt
18:37 aboe basepi, can you take a look at the saltstack-formulas firewalld and see if you can accept the pull request?
18:37 jaybocc2 joined #salt
18:37 jfindlay snicers-work: by default salt-cloud will not autodelete VMs.  If you want to do that, you have to use the cloud map and explicitly enable `enable_hard_maps` in your cloud config
18:38 snicers-work jfindlay, nice, thanks for that. I just did `salt-cloud -d Salt-Minion` I will keep that in mind though.
18:39 ktosiek joined #salt
18:39 Crypticus joined #salt
18:39 gladiatr snicers-work, yeah, a good point to remember is that the minion_id (combined with the minion’s RSA key) is Saltstack’s source-of-truth
18:40 gladiatr snicers-work: if either the key or the minion_id should change, that minion will no longer have validity with regards to the master
18:40 gimpy927 joined #salt
18:41 Pixionus that sounds like what I remember reading, but I just don't get why my jobs are returning nothing then
18:41 gimpy927 Is anyone here using anything to manage and monitor their Salt instance?  I realize I can make a returner to shove Salt output into a DB and use that but I'm curious if better tools already exist.
18:42 jfindlay gimpy927: you can try salt beacons
18:42 snicers-work gimpy927, nothing yet, but I am moving in that direction over time.
18:42 snicers-work https://docs.saltstack.com/en/latest/topics/beacons/
18:43 invsblduck joined #salt
18:43 aw110f joined #salt
18:44 gimpy927 jfindlay: That doesn't seem to be what I'm looking for.  What I need is a way to see what the previous salt runs were, what their status was, and what the putput was.
18:45 teryx510 joined #salt
18:45 jfindlay gimpy927: you probably want to use a returner, then: https://docs.saltstack.com/en/latest/ref/returners/all/index.html#all-salt-returners
18:45 snicers-work jfindlay, is it possible for me to propegate profile changes to running instances or do I need to destroy and recreate them?
18:46 malinoff joined #salt
18:46 gimpy927 jfindlay: Yea...I said that I know I can do that...however, I'm looking for something more than "use a returner to shove into a DB"...I was trying to see if higher-level monitoring software has been made for Salt or if I need ot invent my own
18:46 jfindlay snicers-work: good question.  Once the VM is created, you'll have to manage it with salt.  salt-cloud is not really idempotent, but there are currently development efforts to make it more so
18:47 jfindlay snicers-work: you could check out the salt formula for that or do something custom that is simpler: https://github.com/saltstack-formulas/salt-formula
18:47 NightMonkey joined #salt
18:47 jfindlay gimpy927: that makes sense, although I don't really have any answers
18:50 snicers-work Interesting, so I really have to nail the profile or build into my process to destroy and recreate boxes.
18:50 snicers-work Not a real problem.
18:50 jfindlay gimpy927: the new salt enterprise GUI should do that really well though :-)
18:51 edrocks joined #salt
18:51 jfindlay gimpy927: the monitoring aspect is looking really good at least
18:52 hasues joined #salt
18:53 gimpy927 jfindlay: Have a link?  Can't find much on the site, not even a screenshot.
18:54 jack__ joined #salt
18:55 lompik joined #salt
18:56 snicers-work jfindlay, moving the minion configuration off of the extended base on directly onto the profile fixed this issue. Think that is worthy of a bug report?
18:57 ruxu joined #salt
18:58 jfindlay snicers-work: yes, thanks for reminding me.  It seems that the minion data is not merged, but overwritten.  I think it make sense that the profile data should override the base data, even down into nested items.
18:59 snicers-work Agreed.
18:59 snicers-work Unless explicitly set to override I suppose, you still might want that behaviour.
18:59 bowhunter joined #salt
18:59 jfindlay salt already has tools to do this, they just need to be used in this case
18:59 drawsmcgraw joined #salt
18:59 DanGarthwaite joined #salt
19:00 BhavyaM joined #salt
19:02 bhosmer__ joined #salt
19:02 DanGarthwaite left #salt
19:03 tracphil joined #salt
19:05 * snicers-work is a salt rookie, I need to get some actual training I think.
19:06 jfindlay snicers-work: it looks like you've made a good start :-)
19:06 jack__ joined #salt
19:07 jfindlay although we have remote training available if that's what you're asking, https://saltstack.com/training/
19:08 nma joined #salt
19:11 fleaz Hi, my master is 2015.8 and one of my minions is 2015.3 and they can't talk to each other. Is it because of the version mismatch ?
19:12 zmalone fleaz: Probably not, I've had a 2015.8 master talk to 2015.5 minions.
19:12 jfindlay fleaz: not likely
19:12 fleaz okay, thanks
19:12 fleaz then i will search for some others issues ;)
19:12 nma can you be a bit more specific than can't talk to each other? :)
19:13 fleaz i got the key on the master and accepted it. but "test.ping" says "Minion did not return. [No response]"
19:14 baweaver joined #salt
19:16 jfindlay do you have a firewall on the master blocking the ZMQ ports?
19:16 jfindlay or anywhere inbetween?
19:17 toanctruong joined #salt
19:17 jack__ joined #salt
19:18 jfindlay i.e., can you `telnet master 4506`?
19:21 fleaz yes
19:21 fleaz nmap shows 4505 and 4506 as open
19:21 spiette joined #salt
19:25 zsoftich1 joined #salt
19:25 denys joined #salt
19:25 beardedeagle joined #salt
19:28 jaybocc2 joined #salt
19:29 GreatSnoopy joined #salt
19:33 blckbit10 joined #salt
19:34 scarcry joined #salt
19:36 jack__ joined #salt
19:37 racooper joined #salt
19:40 akhter joined #salt
19:41 dlam hmm is there an easy way to update `salt-call`?    (i think im running into https://github.com/saltstack/salt/issues/30234)
19:41 saltstackbot [#30234]title: Pillar Environment not working with the parameter pillarenv=qa u others environments | Version SaltStack 2015.8.3 (Beryllium)...
19:41 giantlock joined #salt
19:42 dlam me on ubuntu
19:43 jfindlay dlam: not without updating salt itself
19:44 dlam ohh ok that's fine i can update both
19:44 dlam maybe i use that bootstrap_salt.sh thingy
19:44 CeBe joined #salt
19:46 jfindlay dlam: 2015.8.4 is at https://repo.saltstack.com/staging/
19:48 zmalone staging/ doesn't have links to staging, so if you want the packages, the browsable open directories are at https://repo.saltstack.com/staging/apt/
19:48 zmalone although I'd be wary about running pre-release packages
19:49 jfindlay zmalone: we're about to announce 2015.8.4
19:50 jfindlay should be ready to go, just some bureaucratic things left
19:50 zmalone Thank you for classifying that issue, it's a little bit of an ugly situation now
19:50 _beardedeagle joined #salt
19:51 zmalone in six months of releases, there might end up being three different labels on a cronjob, and three identical jobs as a result
19:52 jfindlay zmalone: yeah, unfortunately I don't have the time to investigate it now.  Since AndreasLutro made that change, I think he should be able to figure out what to do
19:54 Blode- joined #salt
19:54 beardedeagle joined #salt
19:56 snicers-work Is there no RDS salt state for 2015.5.9?
19:56 jack__ joined #salt
19:56 bhosmer joined #salt
19:58 snicers-work Yea, it is now in 2015.8.0. Does anyone know if it is possible to do AWS RDS pre 2015.8.0?
19:59 zmalone Not without back-porting it yourself, boto_rds is 2015.8.0 and on.
19:59 zmalone https://docs.saltstack.com/en/latest/ref/states/all/salt.states.boto_rds.html the docs are pretty good about marking when features were added
20:03 beardedeagle winrm pisses me off
20:11 _JZ_ joined #salt
20:11 perfectsine joined #salt
20:14 jack__ joined #salt
20:20 winsalt win* pisses me off
20:20 beardedeagle error: [Errno 104] Connection reset by peer all day everyday
20:20 Edgan jfindlay: 2015.8.4 fixing the package dependency issues?
20:22 mapu joined #salt
20:23 jfindlay Edgan: what odo you mean?
20:23 jfindlay s/odo/do/
20:23 beardedeagle gotta be some firewall crap getting in the way. winexe is not really an option anymore since we have been forced to centos 7.2/samba 4.2
20:24 beardedeagle and I am too lazy to maintain a working winexe-1.1 for the foreseeable future.
20:25 jfindlay I don't understand why the epel maintainers don't just upgrade winexe
20:26 jfindlay beardedeagle: we've been forced to add winexe to our deps on repo.saltstack.com because of this :(
20:26 Edgan jfindlay: the python-tornado-salt like problems
20:26 beardedeagle oh really?
20:27 jfindlay Edgan: I'm not sure what problems you're referring to, but current releases should not have problems from what I remember
20:27 beardedeagle I have a working winexe 1.1 for cent7.2 and samba 4.2 but I just don't wan't to maintain it for the entire company since we are moving everything to salt-cloud for provisioning.
20:29 Edgan jfindlay: Things like conflicts between python packages in CentOS and the repo.saltstack.com repository. I have heard it said it is a work in progress, and was hoping it was ready for 2015.8.4.
20:29 jfindlay beardedeagle: it doesn't look like it's there yet, but it should be soon, I think
20:29 jfindlay maybe Boron?
20:29 beardedeagle now for everyone's favorite question: bout when abouts is that dropping?
20:30 jfindlay Edgan: I've tested the packages multiple times against multiple variations of CentOS/RHEL and haven't had any problems
20:30 Edgan jfindlay: Which is why EPEL hasn't upgrade salt, and may not. I don't think they are cool with how you guys are packaging salt now.
20:31 winsalt beardedeagle, do you have ssl with winrm
20:31 Edgan jfindlay: https://github.com/saltstack/salt/issues/29132
20:31 saltstackbot [#29132]title: Saltstack Yum Repo overwriting packages from CentOS Base | It appears that Salt 2015.8.1 from repo.saltstack org is pulling two dependencies from that repo that are overwriting the Base CentOS repo.  PyYAML 3.10 and python-crypto 2.0.1 are being replaced....
20:31 beardedeagle @winsalt: sure do
20:32 toanju joined #salt
20:34 baweaver joined #salt
20:35 DammitJim if my minions are running tomcat6 and tomcat7
20:35 beardedeagle @winsalt: https://gist.github.com/beardedeagle/eb6f0c8b93e8a4367744
20:35 DammitJim how can I use salt to do a service.restart tomcat ?
20:35 DammitJim do I need to do something on the minion where I 'ln -s tomcat tomcat7' ?
20:35 DammitJim or is there something I can do on the master?
20:36 snicers-work ok, I updated to 2015.8.3, do I need to use the salt boto_rds module to create rds instances?
20:36 jack__ joined #salt
20:36 tracphil joined #salt
20:36 zmalone Edgan: the packages in https://repo.saltstack.com/staging/yum/rhel6/ have the same naming scheme as the packages in that issue, so I'd assume 2015.8.4's packaging on repo.saltstack.com will continue to try and overwrite OS packages.
20:37 JPT joined #salt
20:38 baweaver joined #salt
20:39 jfindlay Edgan: that issue does not seem to have been fixed yet
20:39 Edgan zmalone: They were going to start doing things like python-name-salt to avoid the conflicts
20:39 Edgan zmalone: So probably still WIP
20:39 bronz joined #salt
20:39 bronz joined #salt
20:40 jfindlay Edgan: the problem is that we have to provide some EPEL deps for users who cannot access EPEL due to policy, etc.
20:41 snicers-work does boto_rds.present actually create an rds instance?
20:41 jfindlay also, yes, that is a bad bug
20:42 Edgan jfindlay: yeah, I am sure people are in a silly places where they will trust SaltStack but not EPEL.
20:43 ajw0100 joined #salt
20:46 om2 win_system.join_domain is always failing with False on windows server 2012 r2
20:46 om2 any ideas?
20:46 jack__ joined #salt
20:47 om2 DammitJim: are you minions running tomcat7 and tomcat6 on the same minion?
20:47 om2 or what?
20:48 berserk joined #salt
20:49 berserk joined #salt
20:49 snicers-work I am very confused on how rds instances are supposed to be created. Can I not have it as part of my cloud.maps.d file?
20:50 baweaver joined #salt
20:50 totte joined #salt
20:51 bhosmer joined #salt
20:53 DammitJim no, separate minions
20:53 DammitJim but I want to be able to say restart tomcat on the different minions
20:54 DammitJim or is that just not possible/smart?
20:54 ruxu joined #salt
20:55 berserk_ joined #salt
20:55 JDiPierro joined #salt
20:55 winsalt om2 i had errors with the win_system too, so i just used a ps script
20:56 om2 winsalt: thanks!
20:56 om2 I was thinking I was the only one!
20:56 Tyrm joined #salt
20:56 om2 winsalt: any other gotcha's you noticed?
20:56 Tyrm joined #salt
20:57 jack__ joined #salt
20:57 om2 with windows modules and w2k12 ?
20:58 dlam hey if a salt/pillar/database   folder ...do i gotta put a top.sls in the 'database' sub folder to make it work?
20:59 om2 DammitJim: https://github.com/saltstack-formulas/tomcat-formula
20:59 om2 but not sure it supports restart... :p
20:59 winsalt nothing specific, but i usually find it easier to do windows stuff in powershell anyway
20:59 om2 I'll keep that in mind
20:59 om2 so you add ps script with salt and have salt run it?
20:59 DammitJim om2, I've done most of that
21:00 DammitJim the restart piece is what is ddriving me nuts
21:00 om2 DammitJim: just use cmd.run
21:00 om2 and unless
21:00 om2 or only
21:00 om2 I bet you were trying to avoid that
21:00 DammitJim om2, no, my problem is
21:00 DammitJim how do I call from the master to reset tomcat6 and tomcat7 services?
21:01 DammitJim can I do something like: 'salt '*' service.restart tomcat* ?
21:01 DammitJim I don't think it's working for me
21:02 amcorreia joined #salt
21:02 forrest joined #salt
21:03 snicers-work This is strange, I am getting the following error after updating to 2015.8.3
21:03 snicers-work AttributeError: 'list' object has no attribute 'setdefault'
21:04 om2 salt 'minion' cmd.run 'whatever-restarts-tomcat-cmd'
21:04 om2 DammitJim: ^^
21:05 DammitJim om2, but the tomcat services on the minions are tomcat6 or tomcat7
21:05 DammitJim is there a way I can just say: tomcat?
21:06 om2 make a symlink like you said
21:07 om2 as long as it's in PATH you should be fine
21:07 om2 right?
21:07 onlyanegg joined #salt
21:07 adelcast so I have a minion running on my embedded system, which runs another application that could leverage from calling salt modules
21:08 ablemann joined #salt
21:08 DammitJim symlink on the master?
21:08 om2 no!
21:08 om2 on the minions
21:08 adelcast I want to be able to call salt, but if I do salt-call, another process launches...is there a way to call into the already running minion?
21:09 om2 DammitJim: the symlink will depend on how you restart Tomcat
21:09 antpa joined #salt
21:09 om2 if you have an init script... called tomcat7, you symlink tomcat to that inside the same dir
21:09 om2 I believe
21:09 om2 try that
21:09 asco1aro joined #salt
21:10 om2 if you do not have an init script, you need to symlink the binary file that you use to demonize
21:10 om2 but why bother with this?
21:10 om2 tomcat6 tomcat7 tomcat8124
21:11 Tanta use systemd or upstart
21:11 om2 it's a couple diff chars
21:11 DammitJim I can restart all my tomcat services on all my minions at the same time
21:11 DammitJim I don't have to figure out what server is running what version of tomcat
21:11 om2 oh ok
21:11 DammitJim I'm sure there is a different approach
21:11 Tanta I moved to using systemd on Centos for service management, integrates seamlessly with salt
21:11 DammitJim maybe I need to define a state that does this?
21:11 om2 what command do you use to restart manually on the minions?
21:11 nma joined #salt
21:11 om2 yea, that's an idea!
21:11 DammitJim sudo service tomcat6(7) restart
21:11 om2 create conditionals
21:12 DammitJim I guess I can have a service.restart in a state?
21:12 om2 making a custom state would be ideal for this
21:12 om2 and use conditional to check whether it's tomcat6 or tomcat7
21:13 om2 or fork the tomcat link I sent you and augment and send pull request :)
21:13 DammitJim no, I already know what version of tomcat is running on each minion
21:13 beardedeagle it's odd. my userdata is being ran and completes. but pywinrm fails to connect after to get salt installed.
21:14 om2 you do, but salt doesn't inside a state
21:14 DammitJim right
21:15 DammitJim I have pillar data for this, I think
21:15 DammitJim oh no, actually, I don't
21:15 DammitJim so, I need to query the minion first?
21:15 rmnuvg joined #salt
21:15 om2 yea.
21:16 ruxu joined #salt
21:17 jack__ joined #salt
21:18 jfindlay beardedeagle: in response to your everyman's question: https://github.com/saltstack/salt/issues/30740
21:18 saltstackbot [#30740]title: mine sometimes does not return correct result when using compound matching | We found an issue with mine matching using grains when moving from 2015.5.9 to 2015.8.3. The contents of the mine are fine but when using compound to narrow extraction, results are not correct until a saltutil.sync_grains is done. This is a change in behavior from 2015.5.9 where an explicit sync is not needed. Presumably something changed so that minion
21:18 jfindlay sorry, this one: https://github.com/saltstack/salt/issues/30740
21:18 saltstackbot [#30740]title: mine sometimes does not return correct result when using compound matching | We found an issue with mine matching using grains when moving from 2015.5.9 to 2015.8.3. The contents of the mine are fine but when using compound to narrow extraction, results are not correct until a saltutil.sync_grains is done. This is a change in behavior from 2015.5.9 where an explicit sync is not needed. Presumably something changed so that minion
21:19 jfindlay https://github.com/saltstack/salt/issues/30658
21:19 saltstackbot [#30658]title: win.exe package for RH 7 | see win.exe package for RH 6 #21256...
21:19 jfindlay dude
21:20 Tyrm joined #salt
21:21 beardedeagle lol thanks. looking.
21:21 heaje joined #salt
21:23 rhys joined #salt
21:23 beardedeagle ah yeah
21:23 beardedeagle that git repo that is linked will not work with centos 7.2
21:25 heaje When I use jinja to use a do...update on a dictionary, it is passing in a string as unicode.  Is there a way to prevent that?  I've tried using the safe and string functions in Jinja on the string being passed in.
21:25 beardedeagle @jfindlay: this will get you a compiled version on cent7.2 : http://rand0m.org/2015/08/05/winexe-1-1-centos-6-and-windows-2012-r2/
21:26 heaje for instance, this is what is being shoved into the dictionary on the update() => 'dfs.datanode.data.dir': u'file:///var/lib/hadoop-hdfs/dfs/data'
21:26 jfindlay beardedeagle: would you mind commenting on that issue? :)
21:26 heaje I have verified that the string before being put in is NOT unicode (at least, just printing the string doesn't show it as unicode)
21:26 beardedeagle sure
21:26 jfindlay thanks
21:26 jfindlay heaje: is the unicode string causing you problems?
21:27 beardedeagle commented
21:27 heaje jfindlay: Yes, later on when I try to pass it into a jinja template for a file using the "- defaults" option for file.managed, it breaks
21:27 jfindlay heaje: can you paste the sls file and the error?
21:28 heaje Sure, it's pretty obfuscated with jinja all over the place though
21:28 heaje just warning :)
21:28 beardedeagle @jfindlay: eff it, I am in week 3 of vacation, I am just going to rewrite the linked github repo to work with the needed updates and roll my own rpm.
21:29 beardedeagle bored anyway
21:29 heaje jfindlay: Here's the SLS.  I'm working on producing the error...
21:29 heaje https://gist.github.com/heaje/6e47f59c3d45643cd7bd
21:30 jack__ joined #salt
21:31 heaje jfindlay: Here's the error => https://gist.github.com/heaje/6ada914f3c52fe632452
21:31 heaje If I get rid of the unicode strings, it works great
21:32 perfectsine joined #salt
21:34 om2 does 'unless' work with powershell commands?
21:34 om2 winsalt: ^
21:35 Tyrm joined #salt
21:35 winsalt yeah i think so
21:35 Tyrm joined #salt
21:36 winsalt if you are making a script, remember to use stateful=True and format what the script outputs according to :  https://docs.saltstack.com/en/latest/ref/states/all/salt.states.cmd.html#module-salt.states.cmd
21:37 perfectsine joined #salt
21:42 nma joined #salt
21:43 andrew_v_ joined #salt
21:45 bhosmer joined #salt
21:49 jack__ joined #salt
21:51 nma joined #salt
21:52 akhter joined #salt
21:52 beardedeagle @winsalt: you spoke as if salt-cloud can do winrm without ssl support with ease?
21:53 jfindlay beardedeagle: cool.  Let us know if there's anything you can pass on to https://github.com/saltstack/salt-pack
21:53 sfz- joined #salt
21:53 beardedeagle oh. thats a thing.
22:00 Eugene Is the osfinger grain the one I should be using for reliable detection?
22:01 jfindlay Eugene: it should be
22:01 jfindlay it shouldn't tell you anything different from the other os grains.  If it does, it's a bug :)
22:01 Eugene I guess it's more of a style question.
22:01 jfindlay yeah, I prefer the split os grains
22:02 Rumbles joined #salt
22:06 jfindlay heaje: that may be a jinja problem
22:07 heaje @jfindlay: That's what I was afraid you'd say
22:07 aj_ joined #salt
22:07 heaje Is anyone aware of a way to update the value for a key in a dictionary without using the update() method?
22:08 jfindlay heaje: you should be able to do `d['key'] = new_value`
22:08 aj_ hi all
22:09 aj_ happy friday to everyone
22:09 heaje jfindlay: I'll give it a shot.  I thought I had tried that before.  I hope you're right :)
22:09 aj_ has anyone ever tried this before:
22:09 jfindlay heaje: update is a convenience for changing/adding multiple values: `d.update({'k1: new_1, 'k2': new_2})`
22:10 aj_ i'd want to populate a value in a pillar, by running getting a salt mine from a node\
22:10 dlam anyone got some tips/tricks on how to debug pillar variables that aint there?     like im getting this:  "Rendering SLS 'base:webserver' failed: Jinja variable 'dict object' has no attribute 'webserver"
22:11 Eugene You need to use 'if foo is defined' before referencing things that don't exist
22:11 dlam for probably this in my template: {{ pillar.webserver.static_root }};
22:11 dlam
22:11 Eugene Yeah. Look at your pillar, how is webserver being defined?
22:13 dlam like i changed up my pillar_roots  so it has like... /srv/salt/pillar/production
22:13 dlam vagrant@app:~$ ls /srv/salt/pillar/production
22:13 dlam apache.sls  app.sls  cache.sls  celery.sls  top.sls  webserver.sls
22:13 tracphil joined #salt
22:13 dlam (^ thats the ls, i like moved stuff out of a big top.sls into separate files :D)
22:14 Eugene Are you including those other files now?
22:14 aj_ ipinfo = sudo salt 'master' mine.get 'beta-s5*' network.ip_addrs  and then ipinfo key would have the ip address
22:15 dlam Eugene:  maybe not,  is it like... - include: somewhere?
22:15 jack__ joined #salt
22:15 Eugene Indeed it s.
22:15 Eugene https://docs.saltstack.com/en/latest/topics/pillar/#including-other-pillars
22:15 subsignal joined #salt
22:16 om2 thanks winsalt
22:16 om2 whats the best way to debug whether jinja vars are passed correctly to jinja powershell script template?
22:17 oraqol joined #salt
22:17 om2 the ps script works well manually... but fails with salt jinja template
22:17 heaje @jfindlay: Using the `d['key'] = new_value` stuff doesn't work
22:18 jfindlay heaje: how doesn't it work?
22:18 heaje I get an error about an unknown tag.  If I try to do it using "set my_dict['path']='whatever'" that doesn't work either.  I get => Jinja syntax error: expected token '=', got '.'
22:19 oraqol Hi guys, is there a way to have all the salt minions' grains reside on the master, instead of on the master?
22:19 heaje @jfindlay: Here's the error using "set" https://gist.github.com/heaje/6309f92bf1770679319b
22:20 hasues left #salt
22:20 heaje @jfindlay: and here's the error without using set => https://gist.github.com/heaje/55ec6dad8920979a3d74
22:23 Tyrm joined #salt
22:29 jack__ joined #salt
22:39 blckbit10 joined #salt
22:40 oraqol anyone?
22:41 sjmh oraqol - the master has a cache of the grains.
22:41 sjmh but no, the grains live on the minion.
22:41 sjmh is there a reason why you want grains to reside on the master?
22:41 oraqol bummer, ok guys, thanks
22:41 sjmh oraqol - maybe explain what you're trying to do?
22:42 jfindlay heaje: it looks like you're right and you have to use update due to jinja's limitations
22:42 oraqol instead of having to seed each minion, just have an associative array on the master that links to grain profiles
22:43 jfindlay heaje: `{% do d.update({'key': new_val}) %}`
22:43 oraqol minionID => grain
22:43 zenlot1 joined #salt
22:44 heaje jfindlay: Yup, that's what I'm doing that introduces the unicode.  It might be because I'm appending txt in the value.  I'm testing out variations on that now.
22:45 totzky joined #salt
22:45 sjmh oraqol - is there a reason you can't use pillar?
22:46 jfindlay heaje: how are you appending text?
22:46 heaje 'test' ~ datanode_data_dir_formatted
22:46 heaje jfindlay: and I did just confirm that when I concatenate text together, I get the unicode
22:46 jfindlay I am unfamiliar with that syntax, what is ~?
22:46 heaje it's not the "do" statement that does it
22:47 heaje jfindlay: according to the jinja docs, that's the "correct" way to concatenate strings
22:48 jfindlay heaje: in one of your pastes, you had a salt:// URL in u'', and that's likely the problem.  SLS file names can be unicode in general (actually can be encoded to the system locale but are decoded to unicode for salt handling)
22:49 jfindlay and there are some unfortunate things in how python2 handles unicode, whereas all strings are unicode by default in py3
22:49 heaje jfindlay: It's not the issue.  If I force those strings I've been working on to be non-unicode, it works fine
22:49 heaje jfindlay: Anyway, I just found that if I use "+" to concatenate strings, I don't get the same issue
22:49 jfindlay that's interesting
22:50 jfindlay I still suspect a '{!r}' formatter somewhere in jinja's source code
22:52 ajw0100 joined #salt
22:54 sjmh https://docs.saltstack.com/en/latest/topics/development/logging.html
22:54 sjmh that made me sad
22:54 sjmh :(
22:55 jfindlay oh, that :)
22:56 jfindlay sjmh: what do you want to know about logging?  I've done some things with it
22:56 Crazy67 joined #salt
22:56 sjmh jfindlay - I wanted to setup a logger in a runner that's getting executed ( quite a bit, and concurrently ) by the reactor - I wanted to log stuff to a separate file.
22:56 sjmh Just wasn't sure how safe it'd be to setup my logger w/in the runner.
22:57 sjmh Since, from what I've read, you've got to do some special stuff if you have multiple procs logging to the same file
22:57 sjmh Just losing my messages in the swarm of debug logs in the default master log file
22:57 jack__ joined #salt
22:58 jfindlay you should be able to create a separate logging handler with standard python logging tools
22:58 om2 winsalt: unless is returning True even though it should be returning False and actually run.
22:58 om2 have you seen that too?
22:59 atmosx joined #salt
23:00 sjmh jfindlay - hrm.  figured it'd barf if the runner was running concurrently.
23:00 sjmh i'll give it a shot tho
23:02 jfindlay it might, I've never experimented with multiproc logging in python
23:04 sjmh from what I've read, it doesn't like it, which is why I was wondering how you guys accomplished it.
23:04 sjmh and thus, sadness.
23:05 baweaver joined #salt
23:12 jack__ joined #salt
23:16 RandyT Howdy all
23:16 RandyT Is there a state to decompress a file to a minion?
23:17 RandyT File is not "archived", so not joy with archive.extracted.
23:17 RandyT I see module to gunzip, but was looking for possible way to do this in a state.
23:18 oraqol haven't thought of using pillars
23:19 oraqol I'll look into that
23:19 oraqol thank you
23:23 sjmh RandyT - can't you just execute the module from the state?
23:24 RandyT sjmh: I can... was looking for something cleaner and wanting to make sure I have not missed anything. Which often happens
23:26 RandyT not clear if archive.gunzip will work in this case.
23:26 RandyT pulling source from s3://
23:26 Crazy67 joined #salt
23:31 jack__ joined #salt
23:32 perfectsine joined #salt
23:32 RandyT https://gist.github.com/rterbush/26939b79b0f7c96d6d06
23:33 RandyT And that state runs with success, but no file...
23:33 bhosmer_ joined #salt
23:37 zmalone joined #salt
23:40 iceyao joined #salt
23:41 oraqol left #salt
23:41 Crazy67 joined #salt
23:43 PeterO joined #salt
23:52 Crazy67 joined #salt
23:56 notnotpe_ joined #salt
23:56 catpig joined #salt

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