Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-05-06

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

All times shown according to UTC.

Time Nick Message
00:01 murrdoc yes u care
00:01 murrdoc cos u want that new sweet shit
00:01 murrdoc but honestly u care cos its too much work to ship around a gpg private / public key
00:03 aurynn less obnoxious security is always good
00:03 murrdoc my current plan is environment vars for user/pass for the vault auth
00:04 murrdoc and setup a pillar to use http api
00:04 murrdoc to fetch stuff
00:04 murrdoc also i hope to have some time to do this
00:05 rhodgin joined #salt
00:05 guesswhat_ joined #salt
00:05 guesswhat_ Everytime I install a package i have to call service.systemctl_reload so that I can start it (set it to running). is it correct? It had no problems before. I think its a bug with newly installed salts thru saltbootstrap
00:05 VR-Jack2 looks cool
00:05 guesswhat_ It was working, but now with newly bootstrapped minions, it isn't for me in CentOS7. the systemd cache isnt updated after a pkg.installed, so a service.running doesnt find the <name>.service used by systemd
00:10 bemehow_ joined #salt
00:12 smcquay joined #salt
00:13 rhodgin joined #salt
00:13 druonysuse joined #salt
00:13 druonysuse joined #salt
00:14 MatthewsFace[SEA joined #salt
00:16 dendazen joined #salt
00:18 bhosmer joined #salt
00:18 otter768 joined #salt
00:19 aquassaut joined #salt
00:20 baweaver joined #salt
00:22 hemebond joined #salt
00:23 KyleG joined #salt
00:23 KyleG joined #salt
00:27 guesswhat_ salt is so slow installing packages
00:34 smcquay joined #salt
00:34 smcquay left #salt
00:35 druonysus joined #salt
00:43 bemehow joined #salt
00:49 thayne joined #salt
01:00 penguinpowernz joined #salt
01:03 msheiny joined #salt
01:03 UtahDave guesswhat_: It should only be as slow as the package manager on your system.  what kind of times are you seeing?
01:07 cztanu joined #salt
01:08 cromark joined #salt
01:14 ITChap joined #salt
01:16 mapu joined #salt
01:20 primechuck joined #salt
01:26 joren joined #salt
01:28 dh joined #salt
01:28 zpotoloom joined #salt
01:34 sunkist joined #salt
01:35 amcorreia joined #salt
01:45 dh joined #salt
01:45 scbunn joined #salt
01:50 ageorgop joined #salt
01:52 pipeep joined #salt
01:54 SeeDickCode joined #salt
02:01 dh joined #salt
02:02 DammitJim joined #salt
02:09 cromark joined #salt
02:16 ChesFTC I'm trying to get onchanges_in to work with an entire sls, like the example for require in http://docs.saltstack.com/en/latest/ref/states/requisites.html#require-an-entire-sls-file
02:17 JDiPierro joined #salt
02:17 ChesFTC Essentially, if any state is changed in an included sls, I want a notification triggered (via a custom module)
02:17 ChesFTC I get the error "Cannot extend ID 'secgroup' in 'aws-master:main'. It is not part of the high state."
02:17 ChesFTC Any suggestions? 'secgroup' is included about 3 lines previously
02:19 bhosmer joined #salt
02:20 smcquay joined #salt
02:21 smcquay joined #salt
02:22 mattrobenolt joined #salt
02:22 mattrobenolt Is it possible to query a grain from within another grain?
02:23 JDiPierro joined #salt
02:23 mattrobenolt Something like: https://gist.github.com/mattrobenolt/b9c95ad69c2a8880f43f
02:23 mattrobenolt If that makes sense.
02:23 mattrobenolt A contrived example, but a formula is expecting a grain with some configs, but I want the values of that config to be derived from other grain values.
02:24 mattrobenolt So I want to provide a grain effectively just for that service, which pulls them together.
02:25 Furao joined #salt
02:25 otter768 joined #salt
02:29 mattrobenolt Looking in the code, this is definitely a chicken and egg problem. :(
02:34 smcquay joined #salt
02:35 evle joined #salt
02:36 poi__ joined #salt
02:38 speedlight joined #salt
02:38 speedlight joined #salt
02:38 aqua^mac joined #salt
02:41 pdayton joined #salt
02:47 otter768 joined #salt
02:49 __number5__ postgres_user is broken in 2014.7.5, anyone have a workaround?
02:49 jdputsch joined #salt
02:50 michelangelo joined #salt
02:51 jdputsch_ joined #salt
02:52 matthew-parlette joined #salt
02:54 __number5__ https://github.com/saltstack/salt/issues/22462
02:54 favadi joined #salt
02:56 jdputsch joined #salt
03:03 jerematic joined #salt
03:12 jdputsch left #salt
03:12 jdputsch joined #salt
03:13 writtenoff joined #salt
03:23 hasues joined #salt
03:23 guesswhat_ __number5__ it seems to work if the service is running
03:30 StDiluted joined #salt
03:32 ponpanderer any way to know the process of a script executed through cmd.script on a windows minion? Looks like the salt-minion process is initially the output with jobs.active and the later when the task is complete jobs.print_job displays the pid used by cmd.exe
03:33 hasues left #salt
03:36 thayne joined #salt
03:38 loz-- joined #salt
03:38 DammitJim joined #salt
03:39 mosen joined #salt
03:58 pdayton joined #salt
04:06 pdayton joined #salt
04:08 TyrfingMjolnir joined #salt
04:10 cromark joined #salt
04:12 pdayton joined #salt
04:13 TyrfingMjolnir_ joined #salt
04:14 vstoniest joined #salt
04:16 Gareth joined #salt
04:19 bhosmer_ joined #salt
04:20 vstoniest joined #salt
04:22 pdayton joined #salt
04:23 schristensen joined #salt
04:25 rdas joined #salt
04:36 jimklo joined #salt
04:39 pdayton joined #salt
04:42 vschum1 joined #salt
04:46 clintberry joined #salt
04:47 ramteid joined #salt
05:04 jerematic joined #salt
05:17 otter768 joined #salt
05:22 c10 joined #salt
05:40 yexingok joined #salt
05:40 yggdrasi1 joined #salt
05:51 rdas joined #salt
05:53 stoogenmeyer_ joined #salt
05:58 Furao joined #salt
05:59 djinni` joined #salt
05:59 joeto joined #salt
06:01 viq joined #salt
06:03 izibi joined #salt
06:06 colttt joined #salt
06:12 AndreasLutro joined #salt
06:16 steffo joined #salt
06:20 bhosmer_ joined #salt
06:22 flyboy joined #salt
06:28 cozby_ how does one execute a sls file
06:28 cozby_ without it being in top.sls
06:29 ThomasJ state.sls <file>
06:29 ThomasJ state.sls ntp   for example
06:30 ThomasJ Keep in mind that also bypasses any checks you might have in top.sls
06:30 cozby_ ThomasJ: hmm ok thanks
06:33 eightyeight joined #salt
06:33 cozby_ ThomasJ: brilliant, that worked
06:34 ThomasJ \o/
06:35 kevc joined #salt
06:37 yexingok I am curious if anyone know what 'sls' suffix stands for in a sls file? does this mean something like 'service level state'?
06:37 KermitTheFragger joined #salt
06:38 cozby_ argh, why is the documentation so difficult to grok
06:38 ThomasJ SaLt State file
06:38 cozby_ I just had a page where it showed how to set the env setting in an sls file
06:38 cozby_ I can no longer find it
06:38 preachermanx joined #salt
06:39 cozby_ can someone point me in the right direction?
06:39 cozby_ I would like to set some environment variables in my sls file
06:39 Aikar joined #salt
06:39 JayFK joined #salt
06:39 ThomasJ cozby_: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.environ.html ?
06:40 yexingok Thanks ThomasJ
06:40 cozby_ you know in your sls file when you do - cwd: / - env: ?
06:40 cozby_ there
06:41 cozby_ I guess that link is correct...
06:41 cozby_ but I saw much simpler examples
06:42 ThomasJ cozby: In general, the state you are using will be fully documented. So if you are wanting to cmd.*, google: saltstack state cmd
06:42 ThomasJ cmd.run allows you to do: - env:   for example
06:43 ChesFTC I've got a question about requisites: I've got a state that sends a notification to slack, and I want it to be run should any state in an sls file (that includes other child sls files) have changes. Is that possible?
06:43 cozby_ ThomasJ: right, thx
06:47 agj joined #salt
06:47 trikke joined #salt
06:49 paolo joined #salt
06:49 joren joined #salt
06:51 soren joined #salt
06:53 jerematic joined #salt
06:55 jab416171 joined #salt
06:58 ecdhe joined #salt
07:00 Emantor joined #salt
07:02 steffo joined #salt
07:02 kawa2014 joined #salt
07:02 eseyman joined #salt
07:03 linjan joined #salt
07:03 malinoff joined #salt
07:12 cromark joined #salt
07:13 lb1a joined #salt
07:15 SaltLearner joined #salt
07:17 SaltLearner Hello Group. I wanted to now more about Returners that can be used with SALT. I am trying to run this command "salt '*' test.ping --return syslog" which ideally should write the output in /var/log/syslog file on ubuntu based master. But this is not happening. Is there anyone who has worked on this and can guide me further?
07:21 steffo joined #salt
07:24 hger joined #salt
07:25 hger hello, I need some help with a particular setup and if it is possible to solve using salt
07:26 SaltLearner Hello Group. I wanted to know more about Returners that can be used with SALT. I am trying to run this command "salt '*' test.ping --return syslog" which ideally should write the output in /var/log/syslog file on ubuntu based master. But this is not happening. Is there anyone who has worked on this and can guide me further?
07:26 hger so we have a loadbalancer infront of three nodes. All servers involved are managed by salt. this is what I would like to accomplish:
07:29 hger change one (1) value in the load balancer to point so other node. Then that node should load a different config that is well defined. But I would like to only change one value on the loadbalancer and that change should trigger a change on the other nodes. I was considering grains but I guess that would require two runs of highstate. Any ideas?
07:30 hger and ideally only using salt :)
07:34 Laogeodritt joined #salt
07:38 jimklo joined #salt
07:39 fbergroth joined #salt
07:41 akitada joined #salt
07:41 Laogeodritt joined #salt
07:44 jimklo joined #salt
07:46 yggdrasi1 joined #salt
07:52 pannon joined #salt
07:55 CeBe joined #salt
07:57 favadi joined #salt
08:01 ede joined #salt
08:01 trikke joined #salt
08:04 agend joined #salt
08:04 ede joined #salt
08:08 joeto1 joined #salt
08:08 rdas joined #salt
08:09 flebel joined #salt
08:09 favadi left #salt
08:10 saltie joined #salt
08:10 saltie question: how can I execute a cmd.run after nginx reload?
08:11 ITChap joined #salt
08:14 cromark joined #salt
08:17 Xevian joined #salt
08:18 AndreasLutro saltie: cmd.wait and watch/watch_in
08:20 VR-Jack3 joined #salt
08:21 toastedpenguin joined #salt
08:21 bhosmer_ joined #salt
08:22 saltie AndreasLutro: do you have snippet, not sure exactly how to do it
08:26 toastedpenguin joined #salt
08:27 Grokzen joined #salt
08:29 Grokzen joined #salt
08:29 chiui joined #salt
08:31 cberndt joined #salt
08:42 jerematic joined #salt
08:42 JayFK joined #salt
08:45 st_iron joined #salt
08:45 st_iron hello
08:46 flebel joined #salt
08:46 eliasp joined #salt
08:47 st_iron how can I create dependency between two states? I mean, state1 deletes a file onlyif the file exists, but I want to run state2 in case state1 must have deleted a file?
08:47 st_iron am I clear?
08:48 _JZ_ joined #salt
08:48 nocturn joined #salt
08:48 st_iron so if the file was already deleted and state1 did not run, then state2 is useless
08:49 keimlink joined #salt
08:55 jhauser joined #salt
08:55 eliasp st_iron: for the 1st state, use "file.absent"… then use a requisite like this in your 2nd: "- onchanges: \n - file: your-1st-state"
08:55 eliasp st_iron: see also: http://docs.saltstack.com/en/latest/ref/states/requisites.html#onchanges
08:56 djaykay joined #salt
08:57 st_iron thank you eliasp!
08:59 favadi joined #salt
08:59 nocturn joined #salt
09:00 ITChap joined #salt
09:00 yggdrasi1 joined #salt
09:00 malinoff joined #salt
09:00 ecdhe joined #salt
09:00 babilen joined #salt
09:00 denys joined #salt
09:00 desposo joined #salt
09:00 Jimlad joined #salt
09:00 ckao joined #salt
09:00 ALLmightySPIFF joined #salt
09:00 tkharju joined #salt
09:00 alexhayes joined #salt
09:00 jY joined #salt
09:00 TOoSmOotH joined #salt
09:00 lempa joined #salt
09:00 zer0def joined #salt
09:00 bVector joined #salt
09:00 paha joined #salt
09:00 emid_ joined #salt
09:00 7GHAAC66H joined #salt
09:00 msciciel joined #salt
09:00 dec joined #salt
09:00 ropes joined #salt
09:00 drags joined #salt
09:00 tmmt_ joined #salt
09:06 flebel joined #salt
09:09 st_iron eliasp: onchanges is very good, thank you again! it works now
09:19 otter768 joined #salt
09:21 illern joined #salt
09:23 favadi joined #salt
09:23 huddy joined #salt
09:25 ndrei joined #salt
09:25 wincyj joined #salt
09:26 wincyj hello
09:26 wincyj despite of fact that variable mentioned belowe is present in pillar file i have no idea whats wrong
09:26 wincyj - Rendering SLS 'base:db-mysql.server' failed: Jinja variable 'dict object' has no attribute 'mysql_root_pass'
09:26 wincyj have you got any ideas?
09:26 wincyj i use: salt-ssh 2014.7.5 (Helium)
09:29 eliasp st_iron: you're welcome
09:30 eliasp wincyj: please nopaste your failing SLS on a platform like http://refheap.com or http://gist.github.com
09:30 eliasp wincyj: have you verified whether the minion actually get's the required pillars assigned (e.g. "salt your-minion pillar.get mysql")?
09:31 wincyj i use salt-ssh
09:32 wincyj eliasp: http://pastebin.com/0X5by6E8
09:32 wincyj here's state i use with this variable
09:37 eliasp what you're doing there was most likely already solved in https://github.com/saltstack-formulas/mysql-formula/
09:37 ntropy wincyj: can you share the pillar that has the mysql_root_pass attribute as well
09:42 peters-tx joined #salt
09:44 wincyj ntropy: eliasp http://pastebin.com/7VUXvUHu
09:44 wincyj its not a big deal, just variable and assigned value :/
09:49 Berty_ joined #salt
09:53 teogop joined #salt
10:02 ntropy ah, when you say that this works "salt your-minion pillar.get mysql"
10:02 ntropy then your mysql_root_pass attribute is mysql:mysql_root_pass
10:02 ntropy pillar['mysql:mysql_root_pass'] instead of pillar['mysql_root_pass']
10:04 wincyj well
10:04 wincyj ive changed that and now i have: Rendering SLS 'base:db-mysql.server' failed: Jinja variable 'dict object' has no attribute 'mysql:mysql_root_pass'
10:04 wincyj seems like it cannot find any of vars in pillar
10:05 wincyj ntropy: or maybe i did it wrong?
10:07 ntropy difficult to tell at this point
10:07 ntropy please paste your complete pillar and state, not just a snippet
10:08 ntropy does "salt minion pillar.get mysql:mysql_root_pass" work?
10:09 nocturn_ joined #salt
10:09 wincyj ntropy: pillar: http://pastebin.com/xJgq4rXs
10:09 wincyj state: http://pastebin.com/vCeLzmYc
10:15 johndeo joined #salt
10:18 johndeo if I have an install file with predefined configurations, how to put it on the filesystem of the saltmaster?
10:20 bhosmer joined #salt
10:22 st_iron can I trigger a state3 run if my state1 fails?
10:22 st_iron I want state3 to run only if state1 failed to run
10:22 SaltLearner joined #salt
10:22 SaltLearner has anyone worked on syslog returner of SALT. I want to get the output in /var/log/syslog file of ubuntu based SALT Master
10:23 seb` joined #salt
10:24 st_iron answer to me: onfail
10:24 st_iron thanks :)
10:26 illern joined #salt
10:30 jerematic joined #salt
10:31 c10 joined #salt
10:36 linjan joined #salt
10:43 dendazen joined #salt
10:43 wincyj ntropy: any ideas?
10:44 ntropy no, you didn't share your complete pillar that has the mysql_root_pass attribute :)
10:46 wincyj i DID:>
10:46 jimklo joined #salt
10:46 wincyj http://pastebin.com/xJgq4rXs
10:46 wincyj ntropy: 12:10 < wincyj> ntropy: pillar: http://pastebin.com/xJgq4rXs
10:46 wincyj 12:10 < wincyj> state: http://pastebin.com/vCeLzmYc
10:49 ndrei joined #salt
10:52 giannello joined #salt
10:55 ntropy ok my bad :)
10:55 ntropy so, what do you get when you run salt your-minion pillar.get mysql_root_pass?
10:57 SaltLearner Got it working.
11:02 evle1 joined #salt
11:11 cromark joined #salt
11:12 lb1a joined #salt
11:18 johndeo saltlearner: Care to share?
11:20 otter768 joined #salt
11:22 primechuck joined #salt
11:24 CeBe joined #salt
11:25 wincyj ntropy: i use salt-ssh
11:25 SaltLearner Yeah. So Returner requires working rsyslog already between master and minion
11:26 SaltLearner Which I wasn't using. I found out binary to be corrupt and rsyslog port was not opened on minion to forward the logs to master syslog file
11:26 SaltLearner Fixed it and it got working !! Now i will try using redis as returner for better data management :D
11:27 N-Mi joined #salt
11:28 giantlock joined #salt
11:28 VSpike joined #salt
11:31 VSpike manfred: thanks for the answer yesterday on default minion config for salt-cloud bootstrap. The problem with that mechanism is that it possibly prevents you from overriding options in /etc/salt/minion.d doesn't it?
11:31 VSpike Since it seems to specify a default value for most options explicitly
11:32 jerematic joined #salt
11:35 Auroch joined #salt
11:36 jonatas_oliveira joined #salt
11:37 pdayton joined #salt
11:37 CeBe joined #salt
11:42 bhosmer joined #salt
11:46 jimklo joined #salt
11:50 bluenemo joined #salt
11:52 mapu joined #salt
11:54 I3olle joined #salt
11:57 c10 joined #salt
11:58 johndeo saltlearner: Thanks!
12:09 numkem_ joined #salt
12:11 bluenemo can I use ssh_auth's source option to load complete ssh keys from pillars as well?
12:14 babilen bluenemo: I don't think so
12:14 babilen (it would be great if one could simply use pillar:// in lieu of salt:// everywhere)
12:15 babilen Or any URI really
12:15 pdayton joined #salt
12:19 aquassaut joined #salt
12:28 favadi joined #salt
12:31 cmcmacken joined #salt
12:31 iggy there's a bug open for pillar:// uris
12:32 ponpanderer joined #salt
12:32 c10 joined #salt
12:33 jerematic joined #salt
12:34 tkharju joined #salt
12:37 pdayton joined #salt
12:37 primechuck joined #salt
12:37 primechu_ joined #salt
12:44 elfixit joined #salt
12:44 eliasp iggy: opened by babilen :)
12:45 eliasp iggy: https://github.com/saltstack/salt/issues/18406
12:46 amcorreia joined #salt
12:47 jimklo joined #salt
12:49 iggy yeah, well... at least I knew _something_ about it
12:50 pdayton1 joined #salt
12:51 ltsampros joined #salt
12:51 ltsampros hellp
12:51 ltsampros I've been useind 2014.7.5 in a test setup with 4-5 VMs as minions
12:51 ltsampros I've been trying to test out presence_events and although enabled, they are always empty when I check with salt-run state.event pretty=True
12:52 ltsampros found some similar bugs, but they were referring to 2015. version
12:54 babilen hehe
12:54 jerematic joined #salt
12:54 babilen ltsampros: empty?
12:54 cpowell joined #salt
12:54 oblio joined #salt
12:55 denys joined #salt
12:55 ltsampros babilen: salt/presence/present { "_stamp": "2015-05-06T12:51:33.848954","present": [] }
12:55 favadi left #salt
12:56 DaveQB joined #salt
12:56 babilen ltsampros: And if you "cycle" a minion?
12:57 canci joined #salt
12:57 ltsampros babilen: no change and no events fired :/
12:57 seb` joined #salt
12:57 babilen ltsampros: Give me a second and I check my infrastructure
12:58 babilen (or 10 minutes rather, if you have that time)
12:58 ltsampros yeah no rush.
12:58 edulix joined #salt
12:59 ltsampros the most interesting thing I see is a different ZMQ version between minions/master but that shouldn't matter because presence events according to the docs are generated by reading /proc/net/tcp
13:00 dendazen joined #salt
13:00 ltsampros never mind. most probably it doesn't work because I'm behind NAT and all machines have the same IP to the master :/
13:00 ltsampros damn
13:01 st_iron what is the most elegant way to write a state if I know that the first run will fail surely (download .deb, try to install it, then install dependencies)
13:01 JDiPierro joined #salt
13:01 ltsampros st_iron: install the dependencies first :P
13:02 losh joined #salt
13:03 st_iron ltsampros: unfortunately those depend on the downloaded .deb, but of course, it would be the wisest way :)
13:03 stoogenmeyer_ hey can anyone tell me how i can output something from pillar without the word "local:"
13:03 FeatherKing joined #salt
13:04 manfred pipe to grep -v local ? :P
13:04 stoogenmeyer_ i'm pulling this: salt-call pillar.get my_service:tag --output=txt and I get in response local: 829
13:04 stoogenmeyer_ I can use grep or awk but I was wondering if there's some option to salt that will do it without
13:05 manfred i am not aware of a way to make the outputter do that
13:06 stoogenmeyer_ ok thats fine. I'll just use awk then.
13:06 stoogenmeyer_ thanks!
13:07 manfred np, you could write your own outputter and send it upstream :P
13:07 manfred maybe*
13:07 manfred i don't know if it just shoves local in there after the fact :/
13:08 stoogenmeyer_ don't know how to do that really.. but did find that you can do it with newline_values_only
13:08 stoogenmeyer_ salt-call pillar.get my_service:tag --output=newline_values_only returns the wanted response
13:13 manfred nice!
13:13 jdesilet joined #salt
13:15 samnmax joined #salt
13:16 dyasny joined #salt
13:18 bhosmer joined #salt
13:20 otter768 joined #salt
13:21 subsignal joined #salt
13:22 thayne joined #salt
13:23 bluenemo joined #salt
13:25 ekle joined #salt
13:26 racooper joined #salt
13:31 ekle_ joined #salt
13:31 ekle hi, how can i use the queue flag for highstate schedules ? here is what i have: http://pastebin.com/hehZF86h
13:32 hasues joined #salt
13:32 Tyrm joined #salt
13:33 hasues left #salt
13:36 Furao joined #salt
13:38 pdayton joined #salt
13:38 ebayer joined #salt
13:39 ebayer joined #salt
13:40 perfectsine joined #salt
13:43 erdembayer joined #salt
13:43 st_iron the idea was cool, I install the dependencies from the .deb file now, thanks for the help again :)
13:43 bhosmer_ joined #salt
13:48 Ahlee JID is YYYYMMDDepoch right?
13:48 bluenemo babilen, thx
13:48 Ahlee nope, that's not right
13:48 jimklo joined #salt
13:51 timoguin joined #salt
13:55 Ahlee oh lovely. salt -b 10% \* state.highstate decided to target 92 out of the 614 minions attached to this master
13:55 Slimmons joined #salt
13:56 colonD joined #salt
13:56 Ahlee day before only 485/614 were targeted
13:57 Ahlee So. If you can't trust salt \* to target everything, what do you do?
13:57 Slimmons I was looking through the salt docs here (http://docs.saltstack.com/en/latest/ref/states/top.html), and in the "1.26.18.2.Other Ways of Targeting Minions" section, I noticed that they have a top file that has base: '*': -salt-minion.  Does that mean that it will try and set up salt-minion on the machine?  because that makes 0 sense to me, and can't be right.  I've got to be thinking about that incorrectly.
13:57 Ahlee salt -L $(salt-key -l accepted | awk {magic here})?
13:58 Ahlee Slimmons: That means it will run salt-minion state against all servers
13:58 AndreasLutro Slimmons: it does, yes, but the salt-minion state in that example most likely just does config and updating, not actually installing (that would be impressive)
13:58 djaykay joined #salt
13:58 Slimmons awesome thanks.  Yeah, salt was going to go way up on my list of magic software if it were installing itself before it existed
13:58 Slimmons /waves hand, "magic!"
13:59 Ahlee Well, it has been said to do things before people know they want them ;)
13:59 AndreasLutro think you could do it in theory, if you set up salt-ssh
13:59 AndreasLutro for the initial run
13:59 Ahlee You can.
14:00 StDiluted joined #salt
14:00 Ahlee my kickstarts just have a very basic "get connected to hte master" config that just includes the hostname, and then the salt state reconfigures everyting for hte config
14:01 mpanetta joined #salt
14:02 andrew_v joined #salt
14:12 toastedpenguin joined #salt
14:13 StDiluted joined #salt
14:14 kaptk2 joined #salt
14:14 erdembayer left #salt
14:15 Slimmons I'm reading through the best practices page (http://docs.saltstack.com/en/latest/topics/best_practices.html), and was wondering if I'm understanding the structure they are using correctly, so here's a few questions....             /srv/salt/tomcat, apache, mysql, otherPackagesIMayWantOnAServer, Project1, Project2, Project3, top.sls  ................Then in my top.sls I would have something like ```Project1: 'project1*':  - mysql -apac
14:15 Furao joined #salt
14:15 rojem joined #salt
14:15 Slimmons Project2: 'project2*': -  tomcat - mysql
14:15 Slimmons I really couldn't think of a better way to format that question in an irc lol
14:15 toastedpenguin joined #salt
14:15 ntropy Slimmons: thats unreadable, please see topic for suggestions
14:16 Slimmons yeah
14:16 Slimmons my bad
14:16 Slimmons I can't even tell what I wanted out of that re-reading it
14:17 Slimmons I'll post the question on a forum, then link it here
14:17 Brew joined #salt
14:18 ekle_ how do i prevent that schedules to started concurrent and therefor do not run ?
14:18 ekle_ i often get the error: The function "state.single" is running as PID 17322 and was started at ...
14:20 iggy scheduler by default will only allow one running instance of the scheduled job
14:21 ekle_ which is fine, but i want to get rid of the error
14:22 ekle_ if i have to schedules which run ever minute, one will always fail, with the above error message
14:22 cromark joined #salt
14:23 ekle_ i tried it with queue: True, but it doesn't seem to work
14:25 iggy if you actually want more than one running, set maxrunning >1
14:26 TooLmaN joined #salt
14:27 iggy don't know where you got the queue option from, but it isn't valid for the scheduler
14:27 yaryarrr joined #salt
14:27 Heartsbane joined #salt
14:29 pickledpig joined #salt
14:30 rhodgin joined #salt
14:32 malinoff joined #salt
14:33 MatthewsFace[SEA joined #salt
14:33 CedNantes joined #salt
14:34 bluenemo what criteria are being used in ssh_auth to determine if two public keys are equal and the second should overwrite the first? I'm having two keys which overwrite each other all the time. formula snippet I'm using: http://paste.debian.net/hidden/d9f30a2b/
14:34 bluenemo they are not equal in options, comments and name
14:36 jimklo joined #salt
14:36 clintberry joined #salt
14:36 xintron_ joined #salt
14:36 CedNantes Hello, another quick question : salt-cloud is using ssh to connect to it's remote minion and deploy salt-minion package, how could i change the default port (22) to another one ? I mean : i'm cloning minion from a template with a different ssh port (like 1234) so salt-cloud is stuck before even starting to do anything
14:36 I3olle joined #salt
14:36 timoguin joined #salt
14:36 jschroeder joined #salt
14:37 keimlink_ joined #salt
14:37 Furao joined #salt
14:38 timoguin joined #salt
14:46 ndrei joined #salt
14:47 lumtnman joined #salt
14:47 mage_ is there a way to tell salt to "execute" requisites when I'm executing a single sls file ?
14:48 tkharju joined #salt
14:49 VR-Jack3 I'd think it'd fail if the requisites weren't loaded
14:50 mage_ yes, that's whatI would like to avoid
14:52 keltim joined #salt
14:53 mage_ in short I have this https://dpaste.de/e4BM
14:53 bluenemo cant get ssh_auth not to overwrite my keys.. wtf
14:53 mage_ it fails on the require: stuff
14:53 mage_ but the info is in other sls file, is there a way to "run" them without running a highstate ?
14:54 VR-Jack3 have an include in the state file, or have a separate sls that includes both
14:55 VR-Jack3 here's my auth/init.sls for example.
14:55 VR-Jack3 include:
14:55 VR-Jack3 - auth.users
14:55 VR-Jack3 - auth.config
14:55 mage_ ok, so having a lot of includes is "ok" ?
14:56 mage_ I'm trying to split everything in separate files, to be able to regenerate all py virtualenvs for ex
14:56 VR-Jack3 a lot of people recommend here having things in seperate files and just including those files into the activate state file
14:56 bluenemo argh my bad
14:56 mage_ ok :)
14:57 VR-Jack3 common layout is a services file and a packages file. include both in the init file
14:59 mage_ any examples ?
15:00 scbunn joined #salt
15:01 VR-Jack3 here's my personal ssh init.sls, though I didn't do packages on it yet.
15:01 VR-Jack3 include:
15:01 VR-Jack3 - ssh.config
15:01 VR-Jack3 - ssh.service
15:01 nocturn Hi, can I use salt in pull mode like puppet?  This is for a project managing laptops, so not always online systems
15:01 VR-Jack3 ssh.config has a listen_in to the service.
15:02 VR-Jack3 nocturn: you can reactor based on the minion starting, issuing a highstate
15:03 jalbretsen joined #salt
15:03 debian112 joined #salt
15:04 sandah joined #salt
15:04 nocturn VR-Jack3, I just googled reactor, just not sure how this does a pull
15:06 VR-Jack3 It's not exactly a pull. In a master/minion setup, it's all about when the minion connects to the master. In a masterless setup, it'd probably be about when you refresh the local git
15:07 VR-Jack3 reactor just allows the master to do things when it receives an event; such as minion start.
15:08 nocturn Ok
15:08 VR-Jack3 ie, you are online and connected, so let's get to business
15:10 tmh1999 joined #salt
15:12 SeeDickCode joined #salt
15:16 bhosmer_ joined #salt
15:16 fyb3r joined #salt
15:21 otter768 joined #salt
15:22 jimklo joined #salt
15:22 sdm24 joined #salt
15:24 conan_the_destro joined #salt
15:24 st_iron nocturn: salt-call on the minions can help you
15:25 st_iron I do the same topology, my minions run salt-call from crom
15:25 st_iron *cron
15:25 nocturn st_iron, that is exactly what I am looking for
15:25 st_iron anyway, I shared one of my states on github: https://github.com/stiron/salt-icaclient
15:26 st_iron everybody is welcome to comment my mistakes in it :)
15:26 nocturn Cool
15:27 tmh1999 joined #salt
15:28 tmh1999 joined #salt
15:30 tmh1999 joined #salt
15:31 Furao joined #salt
15:31 _prime_ joined #salt
15:33 tmh1999 joined #salt
15:34 StDiluted joined #salt
15:35 tmh1999 joined #salt
15:39 nzero joined #salt
15:39 Auroch joined #salt
15:41 andrew_v_ joined #salt
15:44 murrdoc joined #salt
15:44 bhosmer joined #salt
15:45 nesv joined #salt
15:48 VSpike Does anyone have any tips on how I'd create and manage RDS instances with salt? Can I create them with salt-cloud for example? Or do I need to use the boto_rds module?
15:49 iggy I imagine boto_rds
15:49 iggy salt-cloud is kind of a "common functionality across cloud providers" kind of thing
15:51 murrdoc joehh:  2014.7.5+ds-1ubuntu1 vs 2014.7.1+ds-1trusty1
15:51 murrdoc was there a reason for the package name switch up
15:51 VSpike I also wonder if you're committing to AWS whether it's better to use cloud formation, and then hook salt into that to do the configuration. Which I think is possible, although I haven't seen much information on it
15:53 VR-Jack3 ahhh. good to know. dicts merge, but lists do not
15:55 pedromaltez joined #salt
15:55 ageorgop joined #salt
16:03 thayne joined #salt
16:03 supersheep joined #salt
16:04 supersheep can you remotely execute salt commands without being root on the master?
16:04 murrdoc yes
16:04 murrdoc it requires setup
16:04 murrdoc but yes
16:04 supersheep salt-api and REST interaction?
16:05 murrdoc yes
16:05 supersheep is there a command-line client that will interact with the REST API?
16:05 VR-Jack3 you can run salt cli as a normal user
16:06 murrdoc there is pepper
16:06 murrdoc its in github/saltstack
16:06 supersheep thanks; my inexpert googling was not turning up anything useful
16:10 sk_0_ how to clean out salt-keys for retired minions?
16:10 greniesa joined #salt
16:11 murrdoc try the salt manage runner
16:13 kawa2014 joined #salt
16:17 MatthewsFace[SEA joined #salt
16:19 iggy sudo salt-run manage.down removekeys=True
16:21 murrdoc joined #salt
16:24 spookah joined #salt
16:24 sk_0_ iggy: very cool. thank you
16:25 iggy I'd probably run it without the removekeys first to make sure it looks right
16:26 Matthews_ joined #salt
16:27 VSpike murrdoc / manfred : you kindly answered my question about where the salt minion config comes from when using salt-cloud. Looks like it explicitly sets all the defaults in the /etc/salt/minion file. Doesn't this create issues if you want to set config using /etc/salt/minion.d, e.g. using the salt.minion formula?
16:28 VSpike Or do settings in the equivalent .d directory have higher precedence over ones in the regular file?
16:28 sdm24 Salt will first read config files in minion.d first
16:28 beneggett joined #salt
16:28 murrdoc i d say test it
16:28 murrdoc put a config in the minion.d directory
16:29 murrdoc then on the minion run salt-call config.get configname
16:31 smcquay joined #salt
16:31 sk_0_ iggy: certainly
16:32 VSpike Cool. It was the "config.get" knowledge that was preventing me finding out if it worked, but it seems like it does. Grand
16:35 N-Mi joined #salt
16:38 wt joined #salt
16:39 bbradley hello
16:40 aparsons joined #salt
16:40 ckao joined #salt
16:40 bbradley i need a feature from v2015.2. currently using the saltstack debian repository. do i have to package my own or is there another way?
16:40 Grokzen joined #salt
16:41 TyrfingMjolnir joined #salt
16:43 writtenoff joined #salt
16:47 linjan joined #salt
16:48 ndrei joined #salt
16:51 Gareth morning morning
16:52 rojem joined #salt
16:53 kunersdorf joined #salt
16:53 gladiatr joined #salt
16:57 iggy bbradley: what feature?
16:57 bbradley https://github.com/saltstack/salt/pull/15079
16:58 iggy you can try to put the updated modules/states in /srv/salt/{_modules,_states}
16:58 MatthewsFace[SEA joined #salt
16:59 iggy or you can just disable recommends the normal way
16:59 bbradley what do you mean?
16:59 bbradley (by the second statement)
16:59 bbradley yea i thought about just patching it.
17:00 VR-Jack3 is there a state equiv of unless? I'd like to do a cmd.run unless a state is true
17:00 iggy /etc/apt/apt.conf.d/01norecommend:\n  file.managed:\n    - contents: |\n        APT::Install-Recommends "0";\n        APT::Install-Suggests "0";
17:01 theologian joined #salt
17:02 bbradley i suppose i could do that for only one package also.
17:02 bbradley thanks that should be good.
17:03 MatthewsFace[SEA joined #salt
17:04 aparsons joined #salt
17:05 bemehow joined #salt
17:05 tracphil joined #salt
17:06 baweaver joined #salt
17:07 tracphil Hi all. I have been using Ansible and now I am moving to Salt. In Ansible there is a concept of items/with_items like so https://gist.github.com/tracphil/7b0eeeed3580d15a0704 Is there anything like that in Salt?
17:08 baweaver joined #salt
17:10 iggy requisites
17:10 iggy if I'm understanding that correctly... not super familiar with ansible
17:10 catpig joined #salt
17:11 Ahlee fresh set of eyes so I'll re-ask, seeing salt -b $NUm \* not match all hosts (sometiems as few as 1/10th, sometimes 9/10ths
17:11 tracphil thanks iggy, basically what I want to do is instead of doing two service.running, just do one stanza that references two services.
17:11 iggy - names: may be what you want then
17:12 iggy it's a list of what would normally be in the - name:
17:12 tracphil ahhh
17:12 tracphil let me dig into that. Thanks!
17:12 philipsd6_ tracphil: iggy is right about this case, but in general if you want to loop using with_items like you do in Ansible, you need to do it explicitly in Jinja. Ansible abstracts the loop away from you in with_items.
17:13 tracphil very nice! Thanks for the info guys!
17:15 CaptainMagnus joined #salt
17:16 murrdoc yw
17:21 VR-Jack3 hmm. guess I can just do cmd.run with unless and do another shell command. Annoying, though. Prefer state results.
17:22 VR-Jack3 I like my script for creating VMs. sadly, it's not a "state".
17:22 otter768 joined #salt
17:24 VR-Jack3 script generates and passes 15+ options into virt-install. the module.virt didn't seem to have near the flexibility I needed.
17:25 Ryan_Lane basepi: so, I had an idea about handling salt keys in AWS really well, but... the key system still isn't pluggable, right?
17:25 ek6 Do you still need to specify raet at install time or setting transport config option all it takes now?  (assuming you have installed all prereqs)
17:25 Ryan_Lane basically, it would allow us to accept keys in an incredibly simple way
17:26 Ryan_Lane including generating and rotating keys in a dead-simple way
17:26 basepi Ryan_Lane: Nope. >.<
17:27 fbergroth joined #salt
17:27 basepi You should run your idea past Tom, if it's cool enough maybe he'll make the key system pluggable tonight. xD
17:27 Ryan_Lane it's *really* good :)
17:27 yggdrasi1 joined #salt
17:27 basepi btw, just a heads up -- 2015.2.0 will never exist. It is 2015.5.0.
17:27 * basepi stands back and waits for the confusion to begin
17:27 Ryan_Lane -_-
17:27 Ryan_Lane why?
17:27 Ryan_Lane that's horrible
17:28 basepi Because Tom didn't take my adive.
17:28 basepi advice*
17:28 VR-Jack3 rofl, and it will release........ *soon!*
17:28 Ryan_Lane you realize that every blog post that I've referenced 2015.2 in is now inaccurate?
17:28 ek6 @basepi your not allowed to do that unless you created new element to go with it
17:28 basepi And named it 2015.2.0 instead of 2014.12.0, and then the date slipped
17:28 basepi 2015.2.0 tells you *nothing* useful
17:28 Ryan_Lane it's supposed to be based on the branch point
17:28 basepi Not when it was released, not when it was branched
17:28 Ryan_Lane right?
17:28 Ryan_Lane -_-
17:28 basepi That was the original idea. Tom changed that for 2015.2.0
17:28 Ryan_Lane you guys really need to fix your release process
17:28 ek6 oh yeah because the periodic naming is real intuitive
17:29 basepi Because he thought we *definitely* could hit February
17:29 Ryan_Lane s/guys/folks/
17:29 basepi I know...
17:29 murrdoc call the next version 2 weeks
17:29 Ryan_Lane I announced a bunch of shit via blog posts
17:29 murrdoc so people stop holding you to a picked version date
17:29 basepi Anyway, I don't think it will actually be a big deal for most people. Just wanted to give a heads up to you IRC folks
17:29 subsigna_ joined #salt
17:30 Ryan_Lane I guess I can go back and edit them... that sucks, though
17:30 basepi Yeah, I'm sorry. I don't think anyone thought about community blog posts.
17:30 basepi =\
17:30 Ryan_Lane can we just pick a 6 month process, do a feature freeze at some point before it, and then actually stop adding features to it?
17:30 basepi In related news: https://github.com/saltstack/salt/tree/v2015.5.0
17:31 Ryan_Lane every release has been "ok, we're freezing now". then 3 months of adding features to the branch
17:31 Ryan_Lane so, it's actually tagged now? :)
17:31 basepi Yep.
17:31 Ryan_Lane \o/
17:32 ek6 they are freezing....they just throw one ice cube in at a time over a long period
17:32 Ryan_Lane -_-
17:32 murrdoc \o/
17:32 iggy basepi: it should be 2015.1 = first release of 2015, 2015.2 = seconde release of 2015, 2015.3 = third release of 2015
17:32 Ryan_Lane releases that come out in a consistent time are more exciting then a giant release with a ton of new features
17:32 murrdoc let it go iggy
17:32 basepi Will be on PyPI soon, and packaged sometime
17:32 murrdoc oh lord no
17:32 murrdoc not pyppi
17:32 murrdoc pypi has gone the way of ruby in some ways
17:32 basepi Luckily you don't have to use pypi
17:33 basepi You can wait for packages. Or use bootstrap. =P
17:33 murrdoc yes
17:33 Ryan_Lane when's the packaging going to be integrating in the ci process? :)
17:33 basepi iggy: I think I probably agree with you
17:33 Ryan_Lane *integrated
17:33 murrdoc <3 the packaging strategy
17:33 basepi iggy: but let's face it, we would try to have a 2015.4 in october and it would slip to 2016
17:33 basepi >.<
17:33 iggy your words...
17:33 basepi It really is ridiculous, and we know it's ridiculous, Ryan_Lane
17:33 murrdoc also basepi do u know 'nicole thomas'
17:33 basepi We're taking measures to fix it going forward.
17:34 aasirc joined #salt
17:34 Ryan_Lane it's pretty easy to avoid slippage, if you actually freeze when you say you're going to freeze :)
17:34 basepi Yes, rallytime (Nicole) works 20 feet from me. She's awesome.
17:34 murrdoc we been trolling her in our saltstack meetings
17:34 murrdoc :)
17:34 basepi hahahaha
17:34 ek6 not sure about ridiculous...   it just wanders between funny haha and funny sad
17:34 Ryan_Lane btw, testing and QA has gotten much better :)
17:34 basepi murrdoc: which meetings? who do you work for?
17:34 Ryan_Lane I don't want to be all negative :D
17:34 basepi Yes, it has.
17:34 murrdoc basepi:  edgecast
17:34 murrdoc mi amo puneet
17:34 murrdoc but uh on a srs note
17:35 murrdoc salt db
17:35 murrdoc is it going to be available in the community edition ?
17:35 murrdoc like a pluggable returner ?
17:35 pedromaltez so Lithium is gonna be v2015.5.0?
17:35 pedromaltez or rather, v2015.5.x
17:36 murrdoc basepi:  salt db ?
17:36 murrdoc please ?
17:36 basepi yes, Lithium is 2015.5.0
17:36 Ryan_Lane basepi: anyone else I should include in this email to thomas?
17:36 basepi murrdoc: I'm unsure of what you mean by salt db
17:36 basepi Ryan_Lane: me please. e-mailing about the release process?
17:37 Ryan_Lane can we change the naming convention, btw? :) I didn't take enough chemistry to know what's going on
17:37 murrdoc basepi:  the maria db backed salt db, is that going to be something that the community edition can use
17:37 Ryan_Lane basepi: about pluggable key
17:37 basepi Ryan_Lane: oh, right. Yes, include me, I'm interested to hear the conversation. And I'll bug him if he doesn't respond.
17:37 Ryan_Lane can you pm me your email address?
17:38 basepi murrdoc: you speak of the GUI backed by maria db?
17:38 murrdoc i dont want the gui
17:38 murrdoc just the db part
17:38 Ryan_Lane basepi: he means sdb
17:38 murrdoc thanks ryan
17:38 Ryan_Lane murrdoc: if you look in the repo, there's sdb modules
17:39 Ryan_Lane for writing/reading
17:39 murrdoc k i ll read the code
17:39 Ryan_Lane via the sdb:// url
17:39 basepi murrdoc: the DB part of the GUI is just the mysql returner. It's in open source already.
17:39 basepi But it sounds like we may be talking about different things. =P
17:39 murrdoc any instructions on how to setup the maria db instance
17:39 basepi I think there's a scheme in the docs for the mysql returner
17:40 basepi schema*
17:40 ph8 joined #salt
17:40 murrdoc kk
17:41 murrdoc i ll dig in
17:41 jab416171 joined #salt
17:41 bemehow_ joined #salt
17:42 keltim joined #salt
17:45 bhosmer joined #salt
17:45 VR-Jack3 bah. packages are so slow. They should be up by now.
17:45 VR-Jack3 :)
17:48 murrdoc debs > pip, gem, whatever the fuck node calls it
17:50 pedromaltez Tried asking in #vagrant to no avail so far - but obviously there might be people here who know. Anyone have any knowledge of how to tell vagrant to use a specific version of salt-bootstrap? You can set the version of salt to install as an argument to the bootstrap script from within the vagrant provisioner block, but there doesn't seem to be a way
17:50 pedromaltez to specify the version of salt-bootstrap to use. The only real way I have found is described in this blog post: http://blog.publysher.nl/2013/08/infra-as-repo-using-git-submodules-for.html , where this person includes a copy of salt-bootstrap within the project and checks out the relevant version of the script
17:53 pedromaltez and then specifies within the provisioner block where the bootstrap script is located. But ideally there would be a way to set the version through a tag instead... like you can do for the version of salt to install.
17:53 desposo joined #salt
17:54 pedromaltez of course that presumes git is already installed in the target system, which is partly the bootstrap script's responsibility...
17:54 pmcg joined #salt
17:56 OneAngry_ joined #salt
17:56 ek6 gotta try again..just dont trust the docs at all here......Do you still need to specify raet at install time or setting transport config option all it takes now?  (assuming you have installed all prereqs)
17:56 tmclaugh[work] joined #salt
17:58 UtahDave joined #salt
17:59 st_iron joined #salt
18:01 martoss joined #salt
18:04 stoogenmeyer_ joined #salt
18:06 jonatas_oliveira joined #salt
18:09 scbunn joined #salt
18:09 N-Mi joined #salt
18:11 murrdoc http://docs.saltstack.com/en/latest/topics/releases/2014.5.0.html :(
18:11 DSGamer joined #salt
18:13 DSGamer What's the typical way most people install modules for apache?
18:14 Gareth murrdoc: what did you do.
18:14 murrdoc got my hopes up Gareth
18:15 murrdoc got my hope up
18:15 Gareth tsk tsk.
18:15 Gareth Wouldn't it be 2014.5.2 anyway? :)
18:15 Gareth actually 2015.2.
18:15 Gareth oh. nm.
18:15 Gareth ignore me.
18:16 * Gareth goes back to the corner
18:16 murrdoc uh hmm
18:16 * murrdoc nods from his corner
18:18 murrdoc well shit http://docs.saltstack.com/en/latest/topics/releases/2015.2.0.html works now
18:23 Gareth You're being punked.  Look for cameras.
18:24 * murrdoc waves
18:25 VR-Jack3 Now I'm confused. Thought it was renamed
18:26 rm_jorge joined #salt
18:26 babilen It has been tagged ages ago
18:26 babilen (prematurely IMHO)
18:28 furball365 joined #salt
18:28 babilen s/tagged/branch created/
18:29 keimlink joined #salt
18:30 rojem joined #salt
18:30 ndrei joined #salt
18:32 stoogenmeyer_ joined #salt
18:32 babilen It was indeed renamed
18:32 murrdoc ja
18:34 babilen "d6b62f7 - (HEAD, tag: v2015.5.0, origin/2015.5, 2015.5) Update manpages to reflect 2015.2 rename to 2015.5 (80 minutes ago)"
18:34 ProT-0-TypE joined #salt
18:34 babilen It seems to be happening
18:35 babilen And; Hooray, they listened to babilen and actually created a branch *during* the release. Now, if the next release branch wouldn't use any crystal ball naming I'd be super happy
18:37 twork joined #salt
18:38 twork left #salt
18:39 MatthewsFace[SEA joined #salt
18:39 twork joined #salt
18:40 supersheep joined #salt
18:40 clintber_ joined #salt
18:41 MatthewsFace[SEA joined #salt
18:41 twork hi all. unix oldster, salt n00b here. rtfm has made my head swim and i come here seeking guidance.
18:44 twork common issue; the docs are fine for 'what does [...] mean' but not as good for 'how best does one...'
18:44 twork and i haven't found 'salt for dummies'. but i hate to ask bonehead questions in a venue like this.
18:45 ageorgop joined #salt
18:45 VR-Jack3 irc is home of the bonehead. don't worry. :)
18:46 twork :)
18:47 twork well then!
18:48 DSGamer I am in the same boat, twork
18:48 DSGamer The "how do I do x" boat
18:48 iggy twork: I struggled that when I started too
18:48 MatthewsFace[SEA joined #salt
18:48 cromark joined #salt
18:49 iggy short answer: you do whatever works best for you
18:49 kossy joined #salt
18:49 StDiluted joined #salt
18:49 numkem_ joined #salt
18:49 teogop joined #salt
18:49 peters-tx joined #salt
18:49 vstoniest joined #salt
18:49 Gareth joined #salt
18:49 loz-- joined #salt
18:49 matthew-parlette joined #salt
18:49 pipeep joined #salt
18:49 awerner_ joined #salt
18:49 renoirb joined #salt
18:49 dingo joined #salt
18:49 Deevolution joined #salt
18:49 marcinkuzminski joined #salt
18:49 kalessin joined #salt
18:49 robsavino joined #salt
18:49 stevednd joined #salt
18:49 jesusaurus joined #salt
18:49 admgre joined #salt
18:49 philipsd6 joined #salt
18:49 morsik joined #salt
18:49 hax404 joined #salt
18:49 ksalman joined #salt
18:49 mens joined #salt
18:49 _ether_ joined #salt
18:49 numkem joined #salt
18:49 gwmngilfen joined #salt
18:49 gthank joined #salt
18:49 InAnimaTe joined #salt
18:49 kzrl joined #salt
18:49 twork i figure, best way to learn (for me at least) is by tackling problems one at a time.
18:49 djinni` joined #salt
18:49 pmcg joined #salt
18:49 tracphil joined #salt
18:49 aparsons joined #salt
18:49 murrdoc joined #salt
18:49 greniesa joined #salt
18:49 jdesilet joined #salt
18:49 seb` joined #salt
18:49 primechu_ joined #salt
18:49 ecdhe joined #salt
18:49 babilen joined #salt
18:49 Jimlad joined #salt
18:49 ALLmightySPIFF joined #salt
18:49 alexhayes joined #salt
18:49 jY joined #salt
18:49 TOoSmOotH joined #salt
18:49 lempa joined #salt
18:49 zer0def joined #salt
18:49 bVector joined #salt
18:49 paha joined #salt
18:49 emid_ joined #salt
18:49 7GHAAC66H joined #salt
18:49 msciciel joined #salt
18:49 dec joined #salt
18:49 ropes joined #salt
18:49 drags joined #salt
18:49 tmmt_ joined #salt
18:49 twork it's getting started that's daunting.
18:49 dingo oh it does? shucks i just assumed it doesn't, great!
18:50 kossy joined #salt
18:50 timoguin joined #salt
18:50 numkem_ split server issue? Whole bunch of people came back
18:50 iggy we could definitely do better with the docs on things like complete walkthroughs for certain features, etc.
18:50 dingo i tried to find evidence of string splicing in jinja ala string[:-2] on jinja2.readthedocs.org, never found it... i'm suprised it works
18:51 dingo its called 'slicing' in python, but there is no mention of it in jinja's docs
18:51 twork iggy: glad to hear you say that, because the example cases i've found, while nice as far as they go, are proving hard for me to generalize from
18:51 numkem_ me too, glad to see it that does!
18:51 babilen dingo: Typically methods of standard types just work .. it is some weird "in-between" Python support in jinja
18:52 numkem_ babilen: some things don't work tho, its sort of a trial and error kind of thing
18:52 twork ...and i feel like an idiot
18:53 twork present case: i found and installed the "users" Formula. i haven't figured out to populated it/tweak it to our environment.
18:54 twork one issue, we use samba and have very few local users.
18:54 denys joined #salt
18:54 tmclaugh[work] joined #salt
18:54 twork simple enough, but everything i try lands me with a stack that won't compile. so my basic question (right now) is, how to debug a fouled up salt?
18:55 dingo -l debug ? :)
18:55 dingo if its just compiling, well, it tells you the offending file
18:55 twork ...context? "salt '*' -debug"?
18:55 dingo know your yaml, i guess
18:56 sdm24 its salt '*' ... -l debug
18:56 babilen numkem_: yeah, hence the "weird"
18:56 dingo when compiling the states, its trying to coerce it all into nice OrderedDict and lists and things, match dependencies between 'requires', build a graph, etc.
18:56 dingo so if its failing there, you have the equivalent of a syntax error of some kind
18:56 seev if you have directory users, you should integrate it with PAM like you normally would without salt
18:57 twork seev: that's done already
18:58 twork rather... i know how to do the integration. what i don't now how to do is The Salt Way.
18:58 seev I don't think Salt would interface directly with the LDAP service except as a client, in an ideal setup
18:58 seev http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.ldapmod.html
18:59 seev but there is a custom module for configuring the ldap parameters and performing basic LDAP functions
18:59 seev although this looks pretty limited, actually
19:00 twork the compile error i'm getting is just "Data failed to compile: " ...followed by "----------" and then a message telling me the PID of salt.sls
19:00 twork "know your YAML" indeed, i guess.
19:00 twork (notice, i don't yet, really)
19:01 twork seev: thanks much. reading.
19:01 soren joined #salt
19:02 seev https://github.com/saltstack-formulas/pam-ldap-formula/
19:02 seev there is also a formula for managing the LDAP client configuration
19:02 tomh- joined #salt
19:03 twork i now live in this channel. beware.
19:03 VR-Jack3 twork: when testing a single sls, I usually prefer to run it with salt-call -l debug on the minion itself.
19:04 seev you can do salt-call state.sls <state> -l debug
19:04 twork just being reminded that "-l" exists has helped. reading the docs the first time... i think i retained about 5%
19:05 VR-Jack3 don't feel bad. I've over 100 hours clocked on docs
19:05 seev I have also used a YAML syntax checker online to catch errors like missing colons
19:05 sdm24 also, state.show_sls will show you what salt is doing in the state
19:05 twork "i know i saw a way to do this..." [skim skim skim... skim skim skim skim skim skim skim...]
19:06 twork ooh, online YAML syntax checker, now there's a clue.
19:07 VR-Jack3 Make sure you include Ryan Lane's blogs on salt, too. Few other good blog posts out there, as well.
19:09 smcquay joined #salt
19:10 twork sdm24: this is a fine example of n00b stuff that make me feel dumb. "state.show_sls" isn't turning up in docs i've checked so far and i think that's because i misunderstand what you mean by it.
19:10 sdm24 http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.state.html#salt.modules.state.show_sls
19:11 sdm24 twork: no worries, I'm a salt noob too
19:12 twork i've got a pretty good hunch that this is one of those places where the first few steps are the hardest. i have the "how to eat an elephant" problem. but i also have deliverables to deliver, so that elephant is sort of dancing around in front of me...
19:14 jab416171 joined #salt
19:14 VR-Jack3 start simple, expand
19:14 flebel joined #salt
19:15 twork yeah. ...i thought "define user accounts" would be that.
19:16 twork the super-simple stuff is simple.
19:16 VR-Jack3 can be, but depends on how you try to do it. without pillar is easier than with pillar, of course.
19:17 twork it is? ah, good. i was trying with.
19:17 VR-Jack3 well, with pillar you have jinja
19:17 twork yeah, and i suspect that jinja is part of my trouble.
19:17 giantlock joined #salt
19:18 VR-Jack3 here's my personal with pillar. https://gist.github.com/anonymous/adde3ff46ce9052c56b5
19:18 VR-Jack3 it's not even complex. Just does what I need
19:18 babilen We have been using the users-formula for a while now in production
19:19 smcquay joined #salt
19:19 StDiluted joined #salt
19:19 numkem_ joined #salt
19:19 teogop joined #salt
19:19 peters-tx joined #salt
19:19 loz-- joined #salt
19:19 matthew-parlette joined #salt
19:19 awerner_ joined #salt
19:19 dingo joined #salt
19:19 marcinkuzminski joined #salt
19:19 kalessin joined #salt
19:19 robsavino joined #salt
19:19 stevednd joined #salt
19:19 jesusaurus joined #salt
19:19 admgre joined #salt
19:19 philipsd6 joined #salt
19:19 morsik joined #salt
19:19 hax404 joined #salt
19:19 ksalman joined #salt
19:19 mens joined #salt
19:19 _ether_ joined #salt
19:19 numkem joined #salt
19:19 gwmngilfen joined #salt
19:19 kzrl joined #salt
19:20 twork so... here's a straightforward question. the users pillar makes sense when i read through it but, where does it get the data to fill in all those elements?
19:21 rojem joined #salt
19:21 baweaver joined #salt
19:21 VR-Jack3 well, in my case, I'm using the pillar root of /srv/pillar/ where I have a top.sls and a users.sls file
19:22 twork oh, sorry, should have been more specific
19:22 sdm24 I believe the actual "data" you want is in the example.pillar
19:22 Deevolution left #salt
19:22 JDiPierro joined #salt
19:22 Deevolution joined #salt
19:22 sdm24 and the formula wants you to create your own pillar in /srv/pillar that is contextually similiar to the example
19:23 twork i meant the data to fill in the users pillar from saltstack-formulas on github
19:23 twork but, i'm hardly married to that
19:23 otter768 joined #salt
19:23 VR-Jack3 there's an example.pillar in the users formula. it shows an example of what the pillar should look like
19:23 twork sdm24: looking at yours now
19:24 VR-Jack3 in my case, I just designed my own.
19:25 Dev0n joined #salt
19:26 VR-Jack3 if done right, pillars are the foundation of salt. learning when it merges/overwrites and establishing good top.sls is fundamental, imo
19:29 froztbyte it can overwrite?
19:30 froztbyte I guess on conflicting namespace for a key
19:30 VR-Jack3 yeah. dicts merge, but lists do not from what I've noticed.
19:31 keltim joined #salt
19:32 babilen twork: I don't understand which data you mean. *You* provide that data. The users-formula can't (and shouldn't) know which users you want to create and what kind of settings you'll use for that. All settings are exemoplified in the pillar.example files,but you don't have to set everything.
19:32 VR-Jack3 and I'm probably not even saying that right, as a dict that has a value and not another dict also doesn't merge
19:33 babilen pillars are merged smartly these days -- http://docs.saltstack.com/en/latest/ref/configuration/master.html#pillar-source-merging-strategy
19:33 ckao joined #salt
19:34 twork babilen: i know i provide it, but i was having a hard time figuring out where it goes -- where the templates looks
19:34 twork i know it should be obvious
19:34 babilen twork: http://docs.saltstack.com/en/latest/ref/configuration/master.html#pillar-source-merging-strateg -- You would, naturally, replace "auser" and "buser" with the usernames of the users you want to create.
19:35 babilen twork: It doesn't look anywhere. But I seem to understand what you are having problems with. Let me elaborate.
19:35 VR-Jack3 it's actually slightly above the merging strategy. pillar_roots is what you're looking for.
19:36 babilen A pillar is simply a Python dictionary that is being specifically targeted to minions. This targeting happens in the pillar's top.sls. So if you have something like "'foomininon': - users.someuser" in that top it would sent whatever data is in users/someusers.sls to that minion
19:37 twork babilen: okay, yeah. this is an example of "i read this somewhere..." but when i reached the context where i needed to apply the item, my mental dictionary failed me.
19:37 babilen The content of that file would follow the pillar.example (I tend to have one file for each user) and you'd place that in /srv/pillar/users/foouser.sls with /srv/pillar/top.sls containing a suitable top.,sls
19:38 twork thanks. this is helping.
19:39 babilen The main point about pillars is that the data they contain is *only* made available to the minions for which you have a suitable entry in the top.sls. That makes them suitable for sensitive data that shouldn't be seen by every minion in your infrastructure
19:39 manfred iggy:  name changed
19:39 babilen Byt they are just dictionaries .. nothing fancier than that
19:39 manfred iggy:  https://github.com/saltstack/salt/releases/tag/v2015.5.0
19:40 babilen yeah, thank you
19:41 babilen It really doesn't make sense to use YYYY.MM with a "try to guess the release date" crystal ball approach. Create the branch when you tag the .0 release
19:41 manfred doesn't matter, 2015 released finally tagged
19:41 twork babilen: yeah, the "just dictionaries" part, i got that. the examples make that clear.
19:41 manfred packging for arch now
19:42 babilen manfred: I haven't seen anything on salt-pkg ... did I miss it?
19:42 manfred it literally just got sent
19:43 babilen and there it is
19:43 babilen :)
19:44 murrdoc joined #salt
19:44 babilen No idea what Daniel means though
19:44 desposo joined #salt
19:45 manfred that is me
19:45 hybridpollo joined #salt
19:45 manfred he said in the end of his first email, that they will announce the release on april 6th
19:45 manfred which was a month ago
19:45 babilen Ah!
19:45 babilen Daniel is manfred! :)
19:45 manfred ==> Creating clean working copy [daniel]...
19:45 manfred yes
19:46 bhosmer__ joined #salt
19:49 manfred and it is in the arch testing repos
19:51 murrdoc manfred is daniel
19:52 iggy hah... I'm going to stay on 2015.2 just to mess with their heads
19:53 babilen For some reason I assumed that people who use "real" names as IRC nick actually use their actual first/second/family name. That is, apparently, not the case
19:53 babilen iggy: heh, make sure to cherry pick everything apart from the version bits and file bug reports with --versions-report ;)
19:53 timoguin There's a salt-pkg mailing list?
19:53 iggy yeah :)
19:53 iggy that's the plan
19:54 babilen timoguin: see pm
19:55 murrdoc why not just use salt-release
19:56 murrdoc cant we all just along
20:00 Tyrm joined #salt
20:01 baweaver joined #salt
20:02 MatthewsFace[SEA joined #salt
20:03 baweaver joined #salt
20:04 baweaver joined #salt
20:06 monkey66 joined #salt
20:06 Not_ joined #salt
20:06 murrdoc @SaltStackInc: The SaltStack Los Angeles meetup w/ @OpenX is tomorrow. Also, don't miss us at @openwestconf this week. http://t.co/f8Tm239YoB
20:07 _JZ_ joined #salt
20:12 jschroeder joined #salt
20:16 bhosmer__ joined #salt
20:16 rojem joined #salt
20:22 Gareth murrdoc: If the FOSS software has taught us nothing...it is that we can not and will not get along.
20:23 ageorgop joined #salt
20:23 MatthewsFace[SEA joined #salt
20:26 murrdoc Gareth:  :)
20:26 murrdoc true
20:27 murrdoc lets start a mailing list to discuss that
20:28 Gareth Better call a meeting to discuss the formation of a mailing list first.
20:28 Chadk I just wanna make sure I got this right: If I'm developing stuff with salt on windows, do I have any way of executing the salt recipes from my windows machine to the minions?
20:29 spookah joined #salt
20:29 murrdoc Gareth:  first lets take a poll about where to have the meeting, and since all pollig websites are game able, i am going to go write one, so we can all use it
20:29 murrdoc #2weeks
20:30 Gareth murrdoc: gather some focus groups too.
20:30 ponpanderer Anyone know a way on a windows minion to determine which job is associated to which child process while the job is running? Seems the only pid I see with salt-run jobs.list_job is the actual salt-minion pid. when the job is completed salt-run jobs.print_job then reveals the actual pid of cmd.exe, etc
20:31 keltim joined #salt
20:31 hemebond ponpanderer: You can't see the cmd.exe PID if you use print_job before it's finished?
20:31 UtahDave Chadk: you can use salt-call to run salt states
20:31 UtahDave ponpanderer: Salt doesn't have a way to get the thread id of the actual job that's running.  We're working on that right now, actually
20:31 sandah joined #salt
20:33 thehaven joined #salt
20:33 Chadk UtahDave, that only allows me to run local salt states, right?
20:34 UtahDave that will pull the states from the master
20:34 Chadk Ah, I see!
20:34 UtahDave if you want to pull local states, then you need to add the --local flag and set up your file_roots configs in your minion config to point to a directory on your minion itself
20:35 UtahDave the nice thing about using salt-call is it helps to get a better idea of what's actually going on when you run the state
20:35 UtahDave salt-call state.sls mynewstate -l debug
20:35 Chadk Ugh, these terms are still a little alien to me. I'm developing on windows, deploying stuff to linux VMs that may or may not exist yet(I need to work out dynamic provisioning still).
20:36 ponpanderer UtahDave: ah, this is what i thought. thanks for confirming! i guess the only way to kill a job using cmd.script (or anything else) is to just kill all child processes on the minion?
20:36 Chadk That is to say, when I say "deploying to", I mean. I've got absolutely nothing, I've yet to execute any salt commands ever.
20:37 UtahDave Chadk: try    salt-call disk.usage
20:37 ponpanderer right now i've resorted to this abomination: http://pastebin.com/mF40nMEK which uses cmd.script to kill all salt-minion children while leaving the minion up
20:38 ponpanderer UtahDave: any rough ETA on either getting the thread id or kill_job in a stable release?
20:39 andrej The latest version of salt for ubuntu 12.04 breaks python-mako ... how come it depends on 0.7 when 0.5.0-1 is the latest version available?
20:44 desposo joined #salt
20:48 cromark joined #salt
20:48 amcorreia joined #salt
20:50 adelcast joined #salt
20:55 OneAngry_ joined #salt
20:56 iggy because nobody tests >3 year old versions of ubuntu... you should open an issue
21:00 arif-ali joined #salt
21:01 martoss joined #salt
21:03 baweaver joined #salt
21:05 SeeDickCode joined #salt
21:17 keltim joined #salt
21:21 UtahDave yes, please open an issue, andrej.  thanks!
21:22 UtahDave ponpanderer: We're still in the very early stages of working on switching the windows minion from using threads to processes. I'm still not sure how soon it will be ready.
21:24 otter768 joined #salt
21:24 CaptainMagnus joined #salt
21:25 Slimmons joined #salt
21:26 trueCamelType joined #salt
21:27 trueCamelType Hey guys, salty noob here, and I have a pretty elaborate question about how to get closer to best practices using salt.  Any help is appreciated.  thanks, http://superuser.com/questions/911666/saltstack-best-practices-documentation-clarification
21:29 SeeDickCode joined #salt
21:29 iggy base is like a fallback
21:30 iggy so you could put apache in /srv/salt/base (don't use /srv/salt for base if you have other env's in there)
21:31 iggy then you could have an apache.config state in the other environments if necessary (although I tend to make my nginx config part of the actual application)
21:32 iggy you could ditch nodegroups and just put 'L@dev-project1,qa-project1,prod-project1' in top.sls directly
21:32 Ahlee Hrm. When did presence_events get added
21:33 iggy would obviate the need to restart the master to update nodegroups
21:34 VR-Jack3 I usually just stick with a base and label minion id's descriptively
21:34 trueCamelType when you say "don't use /srv/salt for base if you have other env's in there", do you mean do something like what I have (ie set a specific file root for each project)?
21:35 iggy yeah, we actually have separate masters for our environments, so we don't really use salt environments at all
21:35 trueCamelType that's actually not a bad idea at all
21:35 iggy so as usual... tons of ways to do the same thing
21:35 trueCamelType that would simplify things pretty drastically
21:36 VR-Jack3 just to be safe, I still reset base to /srv/salt/base/
21:36 VR-Jack3 I hate changing things after the fact when I change my mind
21:36 iggy we still have one git repo that has everything (well, one states and one pillars)
21:36 cromark joined #salt
21:38 VR-Jack3 I'm still working on my separation issues.
21:38 VR-Jack3 ie, one master with base + company1 + company2, and then separate masters that are base + localcompany
21:42 VR-Jack3 probably gonna just git the main master, then do an orchestrate to tar/transfer to new masters.
21:45 martoss joined #salt
21:46 nzero joined #salt
21:46 trueCamelType so, if I ditch the nodegroups, I can actually have multiple minions specified in the top.sls file like this,   project1:   'L@dev-project1,qa-project1,prod-project1'.   I didn't see that in the documentation for top files that I could list multiples (I'm sure its' somewhere, I just haven't come close to reading it all)
21:46 bhosmer_ joined #salt
21:46 trueCamelType because I hate having to use those nodegroups, that would be great
21:47 lz-dylan heya folks
21:47 lz-dylan I'm trying to figure out grain matches
21:47 trueCamelType hi!
21:47 lz-dylan in particular I'm using the ec2_tags custom grain, which seem to work for the most part
21:47 baweaver joined #salt
21:48 lz-dylan doing a `salt \* grains.get ec2_tags` yields a dict, which is lovely
21:48 lz-dylan but I can't successfully select nodes based on its contents
21:48 SeeDickCode joined #salt
21:48 VR-Jack3 trueCamelType: be sure to read up on match types, especially Compound.
21:48 baweaver joined #salt
21:48 lz-dylan ie. doing a `salt -G 'ec2_tags:ENVIRONMENT:*resque*` just says no matches.
21:48 sdm24 joined #salt
21:50 lz-dylan heck, even `salt -G 'ec2_tags:ENVIRONMENT:*` fails to target anything.
21:51 lz-dylan So, how do I get at least _something_ to match?
21:52 VR-Jack3 Did you sync the grains first?
21:52 trueCamelType lol thanks vr-jack3
21:52 lz-dylan Let me forcibly resync...
21:52 sdm24 lz-dylan, does "salt '*' grains.item ec2_tags" return anything?
21:53 lz-dylan doing a saltutil.sync_all fixed it...but when I ran the sync_all it didn't appear to do anything.
21:53 lz-dylan hm. okay, that's a step. thanks!
21:53 lz-dylan sdm24: yes, even prior to doing a resync.
21:54 lz-dylan sdm24: or to be more clear, I was doing a `grains.get ec2_tags`; don't know if that's a major difference from `grains.item`
21:54 sdm24 nope
21:54 VR-Jack3 by default, grains don't sync until after first highstate or if you do a sync
21:55 VR-Jack3 in reactor docs, there's an example to sync on minion startup
21:55 c10 joined #salt
21:55 lz-dylan VR-Jack3: and I've actually got a reactor in there for exactly that
21:56 sdm24 random question, does pillar and/or mine_functions sync on startup?
21:57 murrdoc yes
21:57 murrdoc no
21:57 murrdoc not startup
21:57 murrdoc yes on highstate
21:57 sdm24 ok thanks
21:58 lz-dylan I am confused by this. /var/log/salt/master reflects that the reactor is running and target is `- 'minion_start':` then points to a sync-all directive file that exists.
21:58 lz-dylan le sigh
21:58 lz-dylan probably not worth it; I can just do this based on minion ID
22:02 iggy trueCamelType: - match: compound in the top.sls
22:03 Eric-RB joined #salt
22:03 iggy trueCamelType: from the looks of your hostnames, you could just globs and do '*project1*':
22:07 Eric-RB Hey all, I'm using salt-cloud to provision RHEL 6.5 VMs on AWS.  The minion gets started up, but it doesn't install salt.  I used the --keep-tmp option and it looks like the deploy.sh is missing.  I encountered this before and updating the bootstrap script fixed it, but it's not working this time.  Do you know what might cause this?
22:08 wt when doing pip.installed with a requirements parameter and on success, the state seems to report a result True and changes of {'requirements': True}. However, watching it from another state does activate the watching state while onchanges does not. Is that expected?
22:10 andrej thanks iggy & UtahDave, I'll do that.  And 12.04 is LTS, so age *shouldn't* be a problem :)
22:10 wt Here's an example the states in question: https://gist.github.com/wt/41f96fe1d1f9f8720758
22:11 wt I would have expected both of the cmd.wait_run states to execute
22:11 wt Also, I would have expected the troposphere state to report that tropophere was installed.
22:12 UtahDave wt: can you pastebin the output you're seeing?
22:13 forrest joined #salt
22:14 iggy andrej: the packages for it are what I would call "community maintained", the devs and volunteers do their best but we definitely rely on users to report issues they find
22:14 * forrest waves at iggy
22:15 iggy I didn't do it
22:19 fyb3r left #salt
22:23 andrej cheers iggy, done
22:23 andrej https://github.com/saltstack/salt/issues/23432
22:23 dynamicudpate joined #salt
22:24 andrej Is that sufficient info?
22:24 nzero joined #salt
22:26 rhodgin joined #salt
22:26 StDiluted joined #salt
22:28 murrdoc andrej:  put a ``` before and after your bash output
22:28 Singularo joined #salt
22:31 andrej thanks murrdoc - will do
22:32 Guest29910 joined #salt
22:34 VR-Jack3 hmm. irc returner sounds good
22:34 MatthewsFace[SEA joined #salt
22:35 UtahDave left #salt
22:36 ek6 vr-jack3: what do you think iggy is?  one big saltstack irc returner
22:36 VR-Jack3 ikr
22:36 rojem joined #salt
22:40 mpanetta joined #salt
22:42 murrdoc anyone done 'stages' with salt
22:42 murrdoc like i want a state to execute before any/all states that get executed
22:42 murrdoc not just in highstate
22:43 VR-Jack3 orchestrate does best, I think
22:43 murrdoc right
22:43 aurynn yeah, orchestrate is for that
22:43 VR-Jack3 otherwise, I was looking at using the local minion to modify a state pillar if I need it.
22:44 iggy murrdoc: what do you want to know about orchestration
22:45 VR-Jack3 reactor can sorta handle an automated stage even without orchestration. it's messy, though.
22:46 mosen joined #salt
22:46 clintber_ joined #salt
22:47 desposo joined #salt
22:50 murrdoc iggy:  i dont want to know about orchestrate
22:50 murrdoc i mean i do
22:50 murrdoc but right now i have a different problem
22:51 murrdoc i have an ext pillar that can go away (cos fuck if i know)
22:51 murrdoc so i am trying to put a 'validate pillars before going further' state
22:51 murrdoc before all states that use values from the ext_pillar
22:52 murrdoc not sure if that makes sense
22:52 iggy put it at the top of top.sls {% if salt['pillar.get']('something:specific:to:ext_pillar') test.fail:\n failhard: True {% endif %}
22:52 joehh andrej: re python-mako, packaging oversight/incompatibility with some later versions
22:52 murrdoc in every single state ?
22:52 murrdoc thats painful
22:52 joehh I'll be doing some packaging work tonight and will try and sort it out
22:52 murrdoc hey joehh quick question was testing 2014.7.5 on trusty
22:53 VR-Jack3 just the top.sls. it'll fail the whole processing
22:53 murrdoc wanted to know why you changed the package name from trusty1 to ubuntu1
22:53 murrdoc VR-Jack3:  yes but i dont need it in highstate
22:53 murrdoc but in any state that may use that pillar
22:54 joehh ubuntu1 is the default produced by dch and I missed it when uploading
22:54 VR-Jack3 then you'd need to have it active for any sls that might use the pillar
22:54 joehh even missed the lintian warning referring to it :(
22:54 rhodgin joined #salt
22:55 murrdoc k, cos u used to have trusty1
22:56 murrdoc 2014.7.5 had a nasty code change
22:56 murrdoc added exception in a util file, but no catch for it in the module
22:56 murrdoc so stuff started bombing out
22:56 murrdoc not a fan
22:57 ek6 ouchie
22:57 iggy nobody seems real fond of 2014.7.5
22:57 iggy but it's okay, because 2015.2 will be out soon
22:57 VR-Jack3 2015.5 :P
22:58 * iggy smacks VR-Jack3
22:58 iggy there's more of that to go around
22:58 mosen heh
22:58 VR-Jack3 lol. renaming online docs is easy. renaming iggy's habbits, take forever
22:59 iggy Salt: 2015.2.0rc2-190-g003ac71
22:59 iggy everybody can suck it
22:59 __number5__ iggy: do you have a work-around for postgres_user bug in 2014.7.5?
22:59 mosen good version but i dont like the release codename
22:59 ek6 this web page just asked me to try the shiny new release candidate of salt, v2015.2.0.rc2
22:59 iggy __number5__: I use 2015.2
23:00 rap424 joined #salt
23:00 murrdoc u are missing a .2
23:00 murrdoc :P
23:00 murrdoc __number5__:  whats the bug
23:01 iggy my workaround would be to downgrade to 2014.7.4 (in general)
23:01 murrdoc or put the working files in _modules,_states
23:01 murrdoc is what iwould recommend
23:04 ndrei joined #salt
23:09 __number5__ murrdoc: this one https://github.com/saltstack/salt/issues/22462
23:13 primechuck joined #salt
23:14 murrdoc https://github.com/saltstack/salt/blob/v2014.7.5/salt/states/postgres_user.py hasnt changed since 2014
23:14 desposo1 joined #salt
23:15 scbunn joined #salt
23:15 murrdoc https://github.com/saltstack/salt/commits/v2014.7.5/salt/modules/postgres.py is a 2015.2 prep fix
23:16 codehotter joined #salt
23:17 murrdoc https://github.com/saltstack/salt/blob/v2014.7.5/salt/modules/postgres.py#L568-L572
23:17 murrdoc for sure doesnt return a dict
23:17 murrdoc on error or failure
23:18 murrdoc if user_attr is not None:
23:18 murrdoc not sure if that will check for 'false'
23:19 __number5__ murrdoc: thanks, will look at it
23:19 murrdoc so basically in your salt code
23:19 murrdoc change https://github.com/saltstack/salt/blob/v2014.7.5/salt/states/postgres_user.py#L185 to if user_attr
23:20 murrdoc also check your logs to see if u have an error like "Could not retrieve Postgres role. Is Postgres running?"
23:20 murrdoc on the minion
23:20 murrdoc if thats the case then the fix is what i recommended
23:20 murrdoc send a pull
23:20 murrdoc get a package
23:20 murrdoc fuckitshipit#
23:23 murrdoc good talk murrdoc
23:23 murrdoc why thanks murrdoc
23:23 murrdoc k gtfbtw then
23:23 murrdoc ok murrdoc
23:23 murrdoc o/
23:23 mosen you're losing it murrdoc
23:24 iggy *lost
23:24 __number5__ lol
23:24 VR-Jack3 a-team lost it
23:24 Tyrm joined #salt
23:25 otter768 joined #salt
23:28 cromark joined #salt
23:33 KyleG joined #salt
23:33 KyleG joined #salt
23:35 ajw0100 joined #salt
23:38 Tyrm joined #salt
23:44 c10 joined #salt
23:44 dendazen joined #salt
23:48 bhosmer_ joined #salt
23:49 MindDrive *sigh* Two of my three environments suddenly have most of the minions no longer responding.  I log into several of them to see traceback errors with 'AuthenticationError: message authentication failed'.  A restart fixes it, but now I've got to use something _other_ than Salt to restart all the minions, since the master can't talk to them.
23:50 primechuck joined #salt
23:50 iggy salt-ssh with salt.roster.cache?
23:51 MindDrive I don't have salt-ssh.
23:51 MindDrive (At least, I'm not seeing it on the Salt masters)
23:52 nzero joined #salt
23:52 iggy if you installed from source, it's part of salt, if you installed from packages, it's probably separate
23:52 murrdoc its not installed by default
23:52 murrdoc install it
23:52 murrdoc then do salt-key | grep <whatever makes sense> | xargs -P 6 -I {} salt-ssh {} 'command'
23:52 murrdoc or something
23:53 iggy salt.roster.cache pulls data from minion cache on the master
23:54 MindDrive We manage everything on our systems (nearly all CentOS, some RHEL) via RPMs, so I'll need to track it down.
23:54 iggy there's also the scan roster if your minions are accessible
23:58 murrdoc there is a rpm for salt-ssh

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