Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-10-08

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

All times shown according to UTC.

Time Nick Message
00:02 nickermire joined #salt
00:04 rim-k joined #salt
00:07 dh__ joined #salt
00:14 cilkay Hello. I'm pulling/updating a Node.js application from a Mercurial repo. That application is being started/stopped by supervisord. On my standalone minion, I know that I can run: salt-call --local supervisord.status api-server and restart the Node app (api-server).
00:15 cilkay How can I have that happen when there is a pull/update from the hg repo?
00:15 cilkay er, that should be salt-call --local supervisord.restart api-server
00:16 cilkay In other words, when the code base changes, I want the Node.js app, api-server, to be restarted.
00:16 nikogonzo maybe a post commit hook to a salt api instance?
00:17 nikogonzo oh standalone. maybe an inotify script?
00:17 cilkay It's easy enough with daemons to use "watch" and restart the daemon accordingly. Maybe something like that?
00:17 kevinquinnyo1 joined #salt
00:19 * cilkay has an idea
00:19 bfoxwell joined #salt
00:21 pm90_ joined #salt
00:23 janne_ joined #salt
00:25 mortis_ joined #salt
00:25 is_null joined #salt
00:27 cilkay I attempted to treat api-server as a named service but that didn't work. If I had a systemd script for it, it probably would.
00:28 kitplummer joined #salt
00:28 Cottser joined #salt
00:28 ekkelett joined #salt
00:28 ekkelett joined #salt
00:32 iggy salt has supervisord states that support normal watch statements
00:38 zwi joined #salt
00:39 cilkay iggy: https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.supervisord.html#salt.modules.supervisord.restart
00:39 cilkay I don't see a watch statement there.
00:40 iggy "states"
00:41 JDiPierro joined #salt
00:41 falenn joined #salt
00:42 jodv joined #salt
00:43 cberndt joined #salt
00:52 zsoftich3 joined #salt
00:57 JDiPierro joined #salt
01:04 kevinquinnyo1 left #salt
01:08 ^C whats the general opinion on using GitFS? is it slow? or add a point of failure?
01:09 nickermire joined #salt
01:09 ^C just trying to work out if its worthwhile using over just downloading the states locally
01:12 nickermire joined #salt
01:13 iggy I wouldn't suggest it if you are planning on using saltenv's
01:15 ^C dang, really?
01:16 kermit joined #salt
01:18 subsignal joined #salt
01:23 sunkist joined #salt
01:28 favadi joined #salt
01:29 TheoSLC joined #salt
01:31 falenn joined #salt
01:37 favadi joined #salt
01:37 aparsons joined #salt
01:41 kitplummer joined #salt
01:45 larsfronius joined #salt
01:48 catpigger joined #salt
01:53 pm90_ joined #salt
02:09 Bryson joined #salt
02:12 Jimlad joined #salt
02:13 ageorgop joined #salt
02:13 Cyis Gareth: I figured out how to get the v2 notifications to work and change who they're from. It's got nothing to do with having hipchat.from_name set... with v2 that's not used. It gets the from based on the v2 API token itself... If it is a personal access token for your account it uses your name/profile... if you set it up as a per room token it takes the Label from when you create the token. The v2 still requi
02:13 Cyis res the Content-Type though to work
02:13 stanchan joined #salt
02:23 nocturn joined #salt
02:24 _JZ_ joined #salt
02:24 krymzon joined #salt
02:30 iggy Cyis: you mind opening a docs bug about that?
02:31 Cyis iggy: yeah... I'll even do a little better and see if I can't get a PR created if I have time. I'm in the packing up tonight to travel to see my family after I finish with work tomorrow
02:32 iggy I usually file a ticket, then I get to close my own ticket with my PR
02:32 iggy boosts my GH stats
02:35 Cyis iggy: true...
02:44 dthom91 joined #salt
02:53 evle joined #salt
02:55 Cyis iggy: submitted #27777 for the Header update for v2 API and #27778 is the PR to fix it... #27779 submitted for the doc update
02:59 writtenoff joined #salt
03:00 TheoSLC joined #salt
03:12 rhodgin joined #salt
03:14 TheoSLC joined #salt
03:15 otter768 joined #salt
03:26 auzty joined #salt
03:45 ^C hmmmm. can you do pillar.get from inside a pillar?
03:49 aurynn ^C, not last I checked
03:50 ^C hmmm
03:50 ^C bummer
03:50 iggy ^C: nope
03:50 ^C that would be nice to be able to group common config into one easily editable place
03:51 otter768 joined #salt
03:51 iggy pillarstack, reclass, etc might help
03:51 ^C i see tickets about it on github
03:51 ^C iggy: will go read
03:54 Val_ joined #salt
04:03 ^C wow. cool.
04:03 ^C cant seem to do simple things without introducing extra layers of abstraction
04:04 ^C i'm probably doing something wrong :(
04:04 Nazzy joined #salt
04:04 gmoro joined #salt
04:04 ^C still, it'll probably do the job... i'll have a play
04:05 s0undt3ch joined #salt
04:05 ^C dev branch only? dammit
04:06 danielcb joined #salt
04:06 Cyis can a cmd.wait be set to run on success of another cmd.run ?
04:07 armyriad joined #salt
04:07 SteamWells joined #salt
04:07 Cyis I have 2 commands that need to be ran... the first runs and creates a file which I can test for, it then needs another command run that doesn't actually make a change I can test for but only needs to be ran if the first was successful
04:07 arapaho joined #salt
04:08 stooj joined #salt
04:08 sinonick joined #salt
04:08 moogyver Cyis: Not that I've seen, besides writing a runner to do that.
04:10 Cyis moogyver: hmm... may end up having to move to using cmd.script then and do all the commands within it
04:10 moogyver The orchestrate runner should be the place for it, but it generally considers most things 'successful'
04:10 moogyver I have a Q up on the google groups about it
04:11 munhitsu joined #salt
04:12 Cyis well the nice thing is I have an example script already to use... it is just currently interactive but I can gut that for use via salt
04:12 moogyver Cyis: cool, that helps then.
04:12 Cyis It's a one-time setup that should have to run twice unless there were an issue
04:13 moogyver Cyis: but yeah, orchestrations of execution modules seems to be a shortcoming i've found.  there isn't any really set way to do it.  you can do lots of things to patch it all together though.
04:13 moogyver and runners can do it, since they can execute command modules
04:14 Cyis moogyver: https://github.com/UGNS/sks-formula/blob/master/sks/init.sls this is what I'm working on
04:14 Cyis it's likely not really too usable for someone else at this point but I'd like to make it so
04:17 moogyver Cyis: if you're using states, you may be able to use orchestrate then
04:18 moogyver https://docs.saltstack.com/en/latest/topics/tutorials/states_pt5.html#orchestrate-runner
04:18 moogyver It doesn't work for me because I'm not using states
04:19 sunkist joined #salt
04:20 fedepalla joined #salt
04:20 Cyis moogyver: not sure if using an orchestration runner with the state would be any less convoluted... the state would have to run to install the package and configure but it wouldn't be able to run because of the initial build steps not having taken place
04:21 moogyver hrm, you may not even need orchestrate ( I'm not super familiar with states )
04:21 Cyis unless I'm not following completely
04:22 moogyver But couldn't you do the cmd.run, then just do a state.file.exists
04:22 moogyver and then specify another cmd.run and have it require the file.exists?
04:24 Cyis when I do the cmd.run for the 'build' step it creates a file I can test for so I know I don't need to run it another time... the 'cleanup' cmd.run can't be called until 'build' has been completed and before 'pbuild' is called. The 'pbuild' also creates a file I can test for to ensure it only runs once, but the 'cleanup' doesn't create any files/directories it just merges and dedups data inside the files
04:32 moogyver Cyis: could you do something like this? https://gist.github.com/sjmh/e17282f097b0cb6dd8fc
04:33 Cyis moogyver: possibly... I might try that if what I just put together doesn't work
04:33 moogyver that was me assuming that DB/key was the file you are checking on
04:33 Cyis I'm about to nuke one of my servers and run this against it from scratch. Thankfully it's a cluster so I can take them down like that
04:35 Cyis the problem was that without the cleanup step executing it wouldn't join the cluster and peer
04:36 Cyis state runs nicely against an existing server and behaves as expected
04:37 Cyis gawd I love salt-cloud and mappings :)
04:40 Cyis while that runs I'll go take care of things around here that have to be done by morning... I've got to interview a candidate in just over 13 hours and then have to start an 18 hour drive... gonna be a long day and it's going on 01:00
04:41 moogyver Cyis: blech
04:43 Cyis and of course I have nothing packed yet so have to throw that together
04:43 prog_ joined #salt
04:47 lowfive joined #salt
04:49 evle2 joined #salt
04:50 subsignal joined #salt
04:54 anmolb joined #salt
04:59 markm_ joined #salt
05:03 larsfronius joined #salt
05:03 malinoff joined #salt
05:12 dh joined #salt
05:15 mehakkahlon joined #salt
05:36 anmolb joined #salt
05:38 Cyis moogyver: I think I got it... I switched to the cmd.script for the build and then I added another with an onfail on the build step to perform clean-up
05:39 Joeskyyy joined #salt
05:41 ^C iggy: think i've got  hacky work around - i've imported config file into a pillar variable using import_yaml
05:42 ^C dont know how well it'll work globally, but its working within this pillar file at least
05:48 zot joined #salt
06:00 rdas joined #salt
06:04 katyucha joined #salt
06:09 favadi joined #salt
06:09 colttt joined #salt
06:13 und1sk0 joined #salt
06:19 iggy ^C: yeah, that should work, it's just usually not an option for most
06:28 rim-k joined #salt
06:29 slav0nic joined #salt
06:35 KermitTheFragger joined #salt
06:39 felskrone joined #salt
06:41 jbrnds joined #salt
06:42 edulix joined #salt
06:42 JonGretar joined #salt
06:42 lb joined #salt
06:43 AndreasLutro joined #salt
06:49 zot joined #salt
06:54 Ztyx joined #salt
06:56 msx joined #salt
07:04 xcv joined #salt
07:04 colttt joined #salt
07:11 _aboe joined #salt
07:11 eseyman joined #salt
07:13 deus_ex joined #salt
07:15 felskrone1 joined #salt
07:21 favadi joined #salt
07:22 aparsons joined #salt
07:27 AndreasLutro does anyone want to try and reproduce this bug? https://github.com/saltstack/salt/issues/27783
07:28 Ztyx joined #salt
07:32 rim-k joined #salt
07:33 jhauser joined #salt
07:35 Grokzen joined #salt
07:39 Cyis does this https://github.com/UGNS/sks-formula/blob/master/sks/init.sls#L44-L63 make sense?
07:39 denys joined #salt
07:43 SMillerNL joined #salt
07:43 trapha joined #salt
07:48 SMillerNL Hello everyone, I have the following question on the module "file.rename". When I run a high state test it tells me my file is already moved. What does it check for before giving that message?
07:50 CeBe joined #salt
07:50 evle joined #salt
07:52 subsignal joined #salt
07:55 GreatSnoopy joined #salt
07:56 Cyis SMillerNL: I'm not positive as I've not used it personally... but from reading the docs on file.rename I'd hazard a guess that it's checking to see if the 'source' file exists and finding it doesn't and that the 'name' value does
07:58 SMillerNL That would explain my issue. My assumption would have been that it checks for the resulting file
07:58 SMillerNL Thanks
08:00 elsmo joined #salt
08:00 Cyis SMillerNL: Well according to the doc it's only going to rename the file if it finds the 'source' and the 'name' doesn't exist unless you 'force' it
08:02 thalleralexander joined #salt
08:09 linjan joined #salt
08:10 fredvd joined #salt
08:12 s_kunk joined #salt
08:13 PI-Lloyd joined #salt
08:16 Xevian joined #salt
08:16 chiui joined #salt
08:18 jhauser joined #salt
08:20 trph joined #salt
08:32 dthom91 joined #salt
08:38 babilen eliasp: I really like your pillar+$RENDERER://.... idea
08:38 keimlink joined #salt
08:41 babilen Can I mark minions as temporarily down somehow? (as in: "Yes, dear salt, it is expected to be that way")
08:42 huddy joined #salt
08:46 losh joined #salt
08:51 kukacz joined #salt
08:55 markm_ joined #salt
08:55 mpanetta joined #salt
08:58 Guest62_ joined #salt
09:00 TheoSLC joined #salt
09:01 TheoSLC joined #salt
09:01 TheoSLC joined #salt
09:04 cberndt joined #salt
09:05 Guest62_ anyone able to help with winrepo errors?
09:05 Guest62_ when i run winrepo-genrepo I get a parse error
09:06 Guest62_ File "/usr/lib/python2.7/dist-packages/yaml/scanner.py", line 257, in fetch_more_tokens
09:06 Guest62_ % ch.encode('utf-8'), self.get_mark())
09:06 Guest62_ ScannerError: while scanning for the next token
09:06 Guest62_ found character '%' that cannot start any token
09:06 Guest62_ in "<string>", line 7, column 6:
09:06 Guest62_ {% if grains['cpuarch'] == 'AMD64' %}
09:10 jbrnds joined #salt
09:11 dkrae joined #salt
09:13 markm__ joined #salt
09:15 ericof joined #salt
09:17 zot joined #salt
09:20 larsfronius joined #salt
09:21 zot joined #salt
09:23 impi joined #salt
09:24 larsfron_ joined #salt
09:25 froztbyte is https://github.com/saltstack/pepper considered ready for use?
09:28 N-Mi joined #salt
09:28 marekb joined #salt
09:34 lb joined #salt
09:40 jespada joined #salt
09:45 linjan joined #salt
09:48 Gardinero joined #salt
09:50 markm__ joined #salt
09:51 Gardinero Hello, can anybody help me? After upgrading from 2015.5.5 to 2015.8.0 using "sudo salt 'minion' pkg.refresh_db" does not work any more...
09:52 permalac joined #salt
09:54 babilen Is it drinking Gin & Tonic again or what does it do? How did you upgrade?
09:56 is_null left #salt
09:56 Gardinero I upgraded the master using "salt apt-get upgrade" and the (Windows)-minions by installing the new Salt Minion Package
09:57 Gardinero Before I received "Minionname: True" after running pkg.refresh_db. No I only get "Minionname:2
09:57 Gardinero "Mionionname:"
09:58 babilen So the return value changed, but you can't be sure if something changed. Could you run "salt-call -ldebug pkg.refresh_db" and "salt --versions-report" and the minion and also include the versions report on the master and paste all that to one of http://refheap.com, http://paste.debian.net, https://gist.github.com, http://sprunge.us, http://dpaste.de, … ?
09:58 babilen s/and the minion/on the minion/
09:59 harkx joined #salt
10:00 the_drow joined #salt
10:01 agend joined #salt
10:01 zot left #salt
10:01 zot joined #salt
10:07 the_drow Can anyone tell me what do I do if I get: Received bad data when setting the match from the top file from the minion?
10:09 anmolb joined #salt
10:10 giantlock joined #salt
10:11 babilen the_drow: Could you paste your config, states, top file (and so on, whatever you deem relevant) to one of http://refheap.com, http://paste.debian.net, https://gist.github.com, http://sprunge.us, http://dpaste.de, … ? I don't quite understand what you mean by "when setting the match from the top file from the minion"
10:12 Gardinero Here are the logs: https://www.refheap.com/110414 (pkg.refresh_db), https://www.refheap.com/110413 (minion version), https://www.refheap.com/110412 (master-version)
10:13 babilen Gardinero: Thank you! (it's typically fine to put them all in the same paste as that cuts down on the number of open tabs)
10:13 elsmo joined #salt
10:13 the_drow babilen: That's what I see in the minion logs
10:13 the_drow Sure I'll paste what I have.
10:14 Gardinero Ok, sorry. I could paste them to a new one...
10:14 babilen Gardinero: It's perfectly fine the way it is ... The salt-call output looks alright too and it might just be that the return value of pkg.refresh_db changed
10:14 kbyrne joined #salt
10:16 babilen https://www.refheap.com/110414#L-21 it seems to do something, but then it doesn't return the 'database_name': Bool dict :-/
10:19 Gardinero @babilen But when trying to install a package I get the following message: Unable to locate package <packagename>"
10:19 Gardinero www.refheap.com/110415
10:20 DanyC joined #salt
10:22 babilen Gardinero: That looks like a bug. The refresh_db in 2015.8.0 now returns the cached files (cf. https://github.com/saltstack/salt/blob/2015.8/salt/modules/win_pkg.py#L391 vs https://github.com/saltstack/salt/blob/2015.5/salt/modules/win_pkg.py#L384 )
10:22 the_drow babilen: This is my kitchen.yml file which defines the pillars top and states top https://dpaste.de/jUV4
10:23 babilen I have no experience with Windows and salt minions on it, so I can't really help much further. The win_pkg execution module saw quite some refactoring between .5 and .8 though
10:23 the_drow do you need the rest as well?
10:23 Gardinero Would it be easy to roll back my master to .5?
10:25 babilen the_drow: It literally says "when setting the match from the top file from the minion" ?
10:25 the_drow It says "Received bad data when setting the match from the top file"
10:26 babilen Yeah, that I would have expected
10:26 the_drow but it's from the minion's log
10:28 babilen yeah, that was the missing information I needed for proper tokenisation and phrase detection
10:28 AndreasLutro the_drow, can you show your actual top.sls file? the kitchen.yml may be messing things up
10:29 the_drow those are the only tops I have
10:29 AndreasLutro okay?
10:30 DanyC all, anyone knows a workaround such that when you run salt-bootstrap.sh -M stable (which install the master/ minion) to force overwriting the minion/ master cfg files should an old version exist ?
10:30 Rebus joined #salt
10:39 amcorreia joined #salt
10:40 nc1024 joined #salt
10:40 mr-op5 joined #salt
10:42 the_drow AndreasLutro: So what did I do wrong?
10:45 AndreasLutro no idea the_drow, I don't know how your kitchen.yml things works and unless you can find where your actual top.sls is saved I can't help you further
10:46 the_drow can it be the pillars tops?
10:48 breakingmatter joined #salt
10:54 the_drow because the highstate runs correctly
10:54 the_drow but the pillar changes are not affected
10:55 zot joined #salt
10:56 giantlock joined #salt
10:59 pfallenop joined #salt
11:11 marekb1 joined #salt
11:12 bluenemo joined #salt
11:15 fredvd joined #salt
11:17 zot joined #salt
11:20 DanyC i found the answer - is not possible, we need to remove the old v before installing the new one.. once i get some time i'll push a PR on bootstrap
11:21 babilen the_drow: Could very well be that case. Check for non-ascii characters and other funny stuff in your top file
11:25 Xevian_ joined #salt
11:26 zot joined #salt
11:29 Seanie joined #salt
11:30 DanyC For whoever is intrested/ will browse the channel logs, to install salt-master/ minion/ cloud using salt-bootstrap on CentOS distro you can run ./install_salt.sh -D -M -P -L -Z -p salt-cloud . The option -D is for debug hence you can take it out
11:30 Seanie Hello guys.:)
11:31 DanyC to be on the safe side you can specify stable 2015.5.5 :) Happy salting :D
11:32 zot joined #salt
11:33 Seanie I have a question as always. Is there anyway dynamically to see what is happening during high state, That is not -l debug
11:35 AndreasLutro not really Seanie
11:39 Seanie :(
11:39 amcorreia joined #salt
11:39 Seanie Thats one problem I have guess I can run in debug via bamboo.
11:40 Seanie Thanks
11:40 Seanie Until next time peace
11:41 pfallenop joined #salt
11:45 seanie joined #salt
11:45 seanie Anyone at IP EXPO london?
11:46 JDiPierro joined #salt
11:48 zot joined #salt
11:53 viq joined #salt
12:00 zot joined #salt
12:08 quasiben joined #salt
12:11 krymzon joined #salt
12:11 evle joined #salt
12:11 quasiben joined #salt
12:15 TooLmaN joined #salt
12:16 rhodgin joined #salt
12:16 falenn joined #salt
12:16 teryx510 joined #salt
12:17 dthom91 joined #salt
12:17 zot joined #salt
12:18 mackripeum_ joined #salt
12:19 evle1 joined #salt
12:19 aqua^c joined #salt
12:19 rim-k hello
12:19 subsignal joined #salt
12:20 mackripeum_ hi
12:20 evle joined #salt
12:20 anmolb joined #salt
12:23 mackripeum_ any idea when salt 2105.5.6 will be available on epel (https://docs.saltstack.com/en/latest/topics/releases/2015.5.6.html)
12:24 AndreasLutro mackripeum_: it's not released at all yet, I think
12:25 AndreasLutro no git tag for it
12:32 elsmo joined #salt
12:36 pm90_ joined #salt
12:36 dendazen joined #salt
12:37 lahwran joined #salt
12:40 KennethWilke joined #salt
12:41 Dev0n joined #salt
12:42 pm90__ joined #salt
12:42 is_null joined #salt
12:42 is_null hi all, how to list all failed jobs on an overlord ?
12:44 sbrefort joined #salt
12:49 sbrefort Hello, I just pushed a Hugo formula on github, and I'd like to submit it to saltstack-formulas, here it is : https://github.com/sbrefort/hugo-formula
12:49 sbrefort For now this formula just installs Hugo's package for Debian/Ubuntu
12:51 shiriru joined #salt
12:55 oravirt joined #salt
12:57 subsignal joined #salt
12:58 ferbla joined #salt
12:59 JDiPierro joined #salt
13:00 Dev0n joined #salt
13:00 Dev0n joined #salt
13:01 favadi joined #salt
13:03 DammitJim joined #salt
13:05 toastedpenguin joined #salt
13:10 mapu joined #salt
13:13 aqua^c joined #salt
13:18 racooper joined #salt
13:20 jdesilet joined #salt
13:22 honestly joined #salt
13:23 honestly whelp, just ran into a brick wall trying to use ansible.
13:23 honestly I have sinned, father
13:23 * DammitJim gets the stick and asks honestly to put hands on the table
13:24 Tecnico1931 joined #salt
13:27 Cyis lol
13:27 I joined #salt
13:27 numkem joined #salt
13:28 jeddi joined #salt
13:31 pprkut can you have an "onlyif" and an "unless" statement combined for the same state?
13:33 X67r joined #salt
13:35 DaveQB joined #salt
13:35 honestly so now I have to remember how salt works
13:35 honestly it's been two years :(
13:36 honestly (no, I didn't abandon salt for ansible two years ago)
13:37 impi joined #salt
13:38 DammitJim honestly, it'll be easy... don't worry, you can do it!
13:42 pm90_ joined #salt
13:43 toastedpenguin joined #salt
13:43 TOoSmOotH joined #salt
13:44 JDiPierro joined #salt
13:44 Ztyx1 joined #salt
13:45 mpanetta joined #salt
13:45 breakingmatter joined #salt
13:46 kaptk2 joined #salt
13:46 cpowell joined #salt
13:50 jeffpatton1971 joined #salt
13:52 honestly hmm... how do I use pillar data with salt-ssh?
13:53 babilen honestly: You define pillar_roots and put it somewhere suitable
13:53 jeffpatton1971 joined #salt
13:53 babilen (in the master config)
13:53 * honestly scratches head
13:53 dalibro http://pastebin.com/6EAqaJy6 - documentation says relative include works, but it doesn't like this ? what am I doing wrong? Is the \. relative to parent folder of sls having include ?
13:54 honestly I want to be master and minion less
13:54 andrew_v joined #salt
13:56 impi joined #salt
13:56 honestly ugh
13:56 babilen honestly: You might want to read https://docs.saltstack.com/en/develop/topics/ssh/index.html#configuring-salt-ssh and pay close attention to "Salt SSH takes its configuration from a master configuration file."
13:56 honestly I want ansible's "just use ssh for everything" combined with salt-ssh's "not being terrible"
13:56 honestly :P
13:57 honestly I'm reading that right now
13:57 * AndreasLutro uses salt-ssh
13:57 babilen How does salt-ssh not fit the bill?
13:57 AndreasLutro what seems to be the problem
13:57 babilen AndreasLutro: How do you deal with reactors/mine ? (or don't you)
13:57 babilen Well, you didn't say that you use it exclusively
13:57 AndreasLutro I don't, I know reactors don't work and mines are limited
13:57 timoguin joined #salt
13:58 babilen ack
13:58 AndreasLutro honestly: just like you would with a salt master, put it in /srv/pillar (or another dir if you so configure)
13:58 babilen God, I hope they release a new version soon as I really like to use 2015.8
13:59 sbrefort Hello, I just pushed a Hugo formula on github, and I'd like to submit it to saltstack-formulas, here it is : https://github.com/sbrefort/hugo-formula
13:59 sbrefort For now this formula just installs Hugo's package for Debian/Ubuntu
13:59 AndreasLutro best thing about salt-ssh is how easy it is to use git HEAD :)
13:59 babilen honestly: I typically define a "config dir" that I pass via "-c /path/to/that/dir" to salt-ssh in which I have a master configuration file appropriate for that setup. Just like a normal salt master salt-ssh would use /srv/pillar as pillar_root if you don't define anything else
14:00 honestly babilen: yeah that sounds good
14:00 AndreasLutro Saltfile will let you drop the -c
14:00 favadi joined #salt
14:00 DammitJim am I doing something wrong? I'm not using salt-ssh
14:00 DammitJim :(
14:00 AndreasLutro DammitJim: no
14:00 babilen AndreasLutro: But requires to adjust CWD
14:01 AndreasLutro babilen: true, hopefully not for long re: your XDG issue on github
14:01 babilen AndreasLutro: Yeah, can't wait for that
14:01 honestly babilen: why is a pki_dir required?
14:01 AndreasLutro honestly: it's where ssh keys are stored
14:01 babilen Right now I've defined a couple of aliases (salt-ssh-foo) to handle different -c as I can't be arsed to cd back and forth
14:01 honestly AndreasLutro: well... I don't need that
14:02 AndreasLutro honestly: you definitely do
14:02 honestly all ssh keys I need are stored in ~/.ssh
14:02 kitplummer joined #salt
14:02 honestly AndreasLutro: do you mean "salt needs this or it will refuse to work"?
14:02 AndreasLutro salt-ssh keys need to be passwordless, and the user on the server needs to be root OR have passwordless sudo
14:02 babilen indeed
14:02 AndreasLutro you should not set up either of those with your ~/.ssh key
14:02 honestly I have passwordless sudo.
14:03 AndreasLutro you can make salt-ssh use your personal keys, but it'll still generate some keys at the start of each run
14:03 AndreasLutro afaik
14:03 honestly yeah.
14:03 honestly that's lame.
14:04 honestly this is the part that just works without thinking about it, much less configuring anything, when using ansible.
14:04 honestly I just ssh -A to the "salt master" server
14:04 honestly well, or ansible master
14:05 honestly and then I can configure everything where I have a sudo account.
14:07 scoates joined #salt
14:07 honestly what does "salt-ssh 0.17.5" mean in terms of which doc version I need to look at?
14:08 AndreasLutro means ancient
14:08 dalibro honestly: if you still didn't figure it out: http://pastebin.com/UP8rVkVu (ssh password is optional)
14:08 jeffpatton1971 @sbrefort does the hugo install correctly? I was having problems trying to install .deb's via salt states
14:08 honestly AndreasLutro k
14:08 honestly brb, finding ppa
14:10 honestly dalibro: why do I need a running salt-master to use salt-ssh?
14:10 AndreasLutro you don't, you can skip that step
14:10 jeffpatton1971 I thought you needed that for the roster file?
14:10 AndreasLutro but salt-ssh shares a lot of config with salt master
14:10 Tecnico1931_ joined #salt
14:11 babilen It does, however, not require the master to run on your workstation
14:11 babilen honestly: https://docs.saltstack.com/en/latest/topics/installation/ubuntu.html (assuming "ppa" means that you are running Ubuntu)
14:12 dalibro anyone familiar with relative includes ? :)
14:12 honestly babilen: already got it, I'm on 2015.5.3 (this Joe Healy guy needs to upgrade the package :P )
14:12 babilen honestly: You didn't install from repo.saltstack.com in that case
14:13 babilen (which are automagically generated packages of, well, automatic quality)
14:13 honestly I didn't
14:13 zot joined #salt
14:13 babilen There are multiple repos right now
14:13 honestly yeah
14:13 honestly I can see that
14:14 honestly am I going to start tearing my hair out using 2015.5 because there are really useful missing features?
14:14 markm joined #salt
14:15 davisj Any salt-ssh users out there having success using custom grains? Mine don't seem to be making it to the minion. This is on 2015.8. After several invocations of saltutil.sync_all. Works fine in normal master/minion mode.
14:17 numkem joined #salt
14:17 rhodgin joined #salt
14:18 mpanetta joined #salt
14:19 Xevian joined #salt
14:19 AndreasLutro davisj: just tested here, works fine - I'm using 2015.8 git HEAD though
14:20 chiui_ joined #salt
14:21 davisj AndreasLutro: hmmm... I must be hitting some kind of chicken and egg problem here. Thanks for verifying.
14:24 Gardinero Hello. Is there anybody using version 2015.8.0 with a local Windows Repository without any problems?
14:25 teebes joined #salt
14:25 tmclaugh[work] joined #salt
14:26 zot joined #salt
14:29 zwi joined #salt
14:31 honestly ugh, apparently proper documentation for salt-ssh is a "nice to have" feature :|
14:31 jbrnds joined #salt
14:31 jeffspeff joined #salt
14:32 honestly do i have to look in the source to find out what --wipe *actually* does and how to generally translate salt-ssh commandline options to config file lines? >_>
14:32 Akhter joined #salt
14:32 pm90__ joined #salt
14:33 jeffspeff i'm not sure if it's just a coincidence or not but after i updated my salt master from 2015.5.5.0 to 2015.8.0.3 the master keeps losing network connection. i have to restart the network service to regain connection. anybody else experiencing this with salt 2015.8.0.3 and centos 7?
14:33 Akhter_ joined #salt
14:34 TyrfingMjolnir joined #salt
14:34 cberndt joined #salt
14:36 Grokzen joined #salt
14:37 Brew joined #salt
14:40 honestly oh. I need to work on my reading comprehension.
14:41 honestly the docs actually tell me that yes, I need to go read the source.
14:44 pravka joined #salt
14:44 subsignal joined #salt
14:46 pm90_ joined #salt
14:46 dthom91 joined #salt
14:47 sbrefort @jeffpatton1971: it's running on my ubuntu server 14.10
14:47 armyriad joined #salt
14:48 sbrefort @jeffpatton1971: and it's working ;)
14:48 aurynn joined #salt
14:49 ALLmightySPIFF joined #salt
14:49 zmalone joined #salt
14:51 ALLmightySPIFF joined #salt
14:52 zot for remote execution, does salt have any notion of metadata (tags, other things that might want to be stored with the results of the remote execution), separate from args? the salt command has a metadata flag, but that doesn't sound like it matches up with my meaning.
14:54 aurynn joined #salt
14:54 clintberry2 joined #salt
14:55 zmalone joined #salt
14:57 Nazca joined #salt
15:01 hasues joined #salt
15:02 debian112 joined #salt
15:02 _JZ_ joined #salt
15:02 tracphil joined #salt
15:02 hasues left #salt
15:02 rhodgin joined #salt
15:04 davisj zot: can you decribe what you're trying to accomplish?
15:04 MadHatter42 joined #salt
15:05 MadHatter42 is this room about salt configuration management ?
15:05 Twiglet Salt's ordering makes me want to kill things sometimes
15:05 Twiglet MadHatter42: yes
15:05 MadHatter42 Twiglet, thnx
15:06 Nazca__ joined #salt
15:06 MadHatter42 i was thinking to start using a configuration management and was unsure about using salt or chef
15:06 Twiglet Well, I'm pretty sure we'll vote for salt here ;)
15:06 MadHatter42 yeah but could you elaborate
15:07 MadHatter42 as I need to answer this question as well if I tell my manager that we should move to salt
15:07 ktenney joined #salt
15:07 MadHatter42 start using*
15:07 MadHatter42 any good guides about getting started with salt Twiglet
15:07 MadHatter42 ?
15:08 illern_ joined #salt
15:08 ktenney joined #salt
15:08 Twiglet My arguments for salt over other CM systems are: It's really easy to get going, it's really easy to write CM states for, the orchestration is a first class citizen (not tacked on as an after thought)
15:08 Slimmons joined #salt
15:09 Twiglet MadHatter42: The official docs are quite good these days https://docs.saltstack.com/en/latest/topics/tutorials/walkthrough.html
15:09 Twiglet Think I also saw a seried on the digital ocean tutorials pop up a couple of days ago
15:10 Twiglet ah yes, https://www.digitalocean.com/community/tutorial_series/managing-development-environments-with-saltstack
15:10 davisj This new doc is shiny https://docs.saltstack.com/en/getstarted/fundamentals/index.html
15:10 Twiglet no idea if that is any good though
15:10 Twiglet davisj: that is indeed new and shiny
15:11 Dev0n Hey, is there a ufw state or am I just expected to either use iptables or cmd to control ufw?
15:12 Twiglet I've not seen one
15:12 zmalone MadHatter42: the argument for is listed above, the argument against is bugs bugs bugs
15:12 Dev0n Twiglet: ahh
15:12 JDiPierro Dev0n: Private networking is the way to go if that's available. My company has our own state/module for managing UFW and it's just a nightmare... go with private networks if you can.
15:13 MadHatter42 zmalone, what bugs ?
15:13 Twiglet zmalone: this is true, core features do break a lot more often than i'd like, but then if you don't want that don't run bleeding edge ;)
15:13 MadHatter42 i cant use something that breaks too often
15:14 Twiglet MadHatter42: it's not quite that bad, as I say just don't run bleeding edge
15:14 MadHatter42 ah thats ok
15:14 Dev0n JDiPierro: I've gone private for most of the internal services, it's just for the remote staging server that needs to be accessed publicly but restricted to a few ips.
15:14 zmalone newly announced features won't work from the moment they are release (Tornado), core states will be broken for years running, some platforms that seem to be supported are not really.
15:14 Twiglet I usually run a couple of (minor) releases behind and have lots of automated testing for new releases
15:14 Dev0n So private networking wouldn't work for that unfortunately
15:15 Twiglet (testing my states, not salt's code that is)
15:15 MadHatter42 Twiglet, what does an automated test consist of ?
15:15 Twiglet firing up a butt-ton of vagrant machines with jenkins and running the CM states
15:16 zmalone Here's an example of a common one: https://github.com/saltstack/salt/issues/22356
15:16 JDiPierro Twiglet: What are CM states?
15:16 Twiglet which reminds me, another downside that infuriates me is the lack of testing frameworks for salt
15:16 zmalone for some reason pkg can't handle dependency resolution for some packages, so you just shell out for that single package
15:16 zmalone you'll run into lots of stuff like that and just have to shell out to work around it
15:16 Twiglet JDiPierro: CM = Config Management, so in this case Salt states
15:17 JDiPierro I just discovered ServerSpec not long ago and we're about to get started using it for testing our states: http://serverspec.org/
15:17 kitplummer joined #salt
15:17 JDiPierro Twiglet: Ahh, gotcha :P
15:17 Twiglet zmalone: at least shelling out is nice and not a dirty thing in salt (mostly)
15:17 favadi joined #salt
15:17 zmalone except when it is, like the security fix in early 2015
15:18 zmalone It's a workable system, just prepare for some frustration vs. more mature configuration management dsls
15:18 peppper joined #salt
15:19 peppper is there a way to assicate a existing elastic IP with an ec2 instance (not vpc) via salt-cloud?
15:20 perfectsine joined #salt
15:21 MadHatter42 Twiglet, can salt help me with standardization ?
15:21 MadHatter42 and reporting
15:21 MadHatter42 as in I need the same sets of users / software / configurations on all my servers
15:22 zer0def joined #salt
15:22 iamnota joined #salt
15:23 iamnota 'salt -v 172.26.2.2 cp.get_dir salt://7-zipPortable C:\\distr\\7-zipPortable' where is the mistake? i have no folder on minion
15:26 jalbretsen joined #salt
15:26 fredvd joined #salt
15:26 elsmo joined #salt
15:28 _JZ__ joined #salt
15:28 murrdoc joined #salt
15:29 perfectsine_ joined #salt
15:29 davisj MadHatter42: the enterprise version of salt ($) has a reporting interface. There's also  https://github.com/tinyclues/saltpad Apart from that, you have many options to return salt command results to 3rd party databases via returners https://docs.saltstack.com/en/latest/ref/returners/ Or, you can write a custom event listener if none of those works for you.
15:30 jeffspeff iamnota, have you tried without the double back-slashes in the target dir?
15:30 MadHatter42 i'll start using it soon
15:30 MadHatter42 i'm a bit unsure about tests but i'm sure I do need them
15:31 MadHatter42 as for example ldap logins after an upgrade and service status after an update and stuff like that
15:31 Twiglet MadHatter42: Yes it can, assuming you don't have remot changes on the systems salt is managing then all changes would be made by salt. So you'd have a state which sets up your users, one that install postgres and sets it's databases up, one that sets up nginx & config etc... then each run salt would makes sure the configs/service are running etc...
15:31 Twiglet each time it run you can have salt return it's output to something if you want with returners, so you could have it output to a database for example
15:32 Twiglet however if a user logs in an makes a lod of changes to files/services/etc... not managed by salt then salt won't revert those changes
15:33 iamnota <jeffspeff> ty, but not help
15:34 kawa2014 joined #salt
15:35 edulix joined #salt
15:35 Brew joined #salt
15:36 thalleralexander joined #salt
15:39 thalleralexander joined #salt
15:45 peppper is there a way to assicate a existing elastic IP with an ec2 instance (not vpc) via salt-cloud?
15:51 felskrone joined #salt
15:58 aboe joined #salt
16:00 tomspur joined #salt
16:00 debian112 joined #salt
16:00 jodv joined #salt
16:02 RedundancyD joined #salt
16:02 Bryson joined #salt
16:06 Akhter joined #salt
16:07 elsmo joined #salt
16:10 JDiPierro joined #salt
16:10 kitplumm_ joined #salt
16:10 Akhter_ joined #salt
16:11 larsfronius joined #salt
16:14 debian112 left #salt
16:17 CheKoLyN joined #salt
16:17 trj joined #salt
16:17 aparsons joined #salt
16:24 ipmb joined #salt
16:25 writtenoff joined #salt
16:25 patchedmonkey joined #salt
16:32 Ztyx joined #salt
16:33 aparsons joined #salt
16:41 stanchan joined #salt
16:41 winsalt joined #salt
16:42 Lionel_Debroux_ joined #salt
16:43 ashirogl joined #salt
16:45 JDiPierro joined #salt
16:51 EvaSDK joined #salt
16:51 EvaSDK joined #salt
16:54 murrdoc left #salt
16:54 murrdoc joined #salt
16:56 perfectsine joined #salt
17:06 winsalt joined #salt
17:08 malinoff joined #salt
17:09 winsalt i have a pillar that contains a list, but when I do pillar.get it returns nothing.  Is there something I am missing?
17:09 murrdoc the data ?
17:09 murrdoc AMIRITE!
17:10 murrdoc salt <minion> pillar.items
17:10 murrdoc then pillar.get pillarname
17:10 murrdoc see if its blank in both
17:10 winsalt its there in pillar.items
17:10 murrdoc and it has data ?
17:10 murrdoc as u expect it to
17:11 winsalt yeah
17:11 murrdoc then u are using the key wrong
17:12 winsalt ill put it in a gist
17:12 winsalt https://gist.github.com/rmarcinik/7e6420e25c6fc59f56da
17:13 ashirogl joined #salt
17:16 jodv joined #salt
17:17 kitplummer joined #salt
17:19 moogyver winsalt: works for me? i added a comment to your gist
17:20 winsalt did you mock it up as a pillar file?  maybe it is because  this is a custom pillar
17:21 kitplumm_ joined #salt
17:21 _JZ_ joined #salt
17:21 moogyver you mean it's an external pillar?  yeah, mine is just using the normal pillar file stuff
17:22 krymzon joined #salt
17:22 winsalt yeah, ill mess around with how im returning the data from ext_pillar
17:22 engblom left #salt
17:28 stanchan joined #salt
17:30 cliluw joined #salt
17:32 cyborglone joined #salt
17:32 Akhter joined #salt
17:36 test joined #salt
17:41 larsfronius joined #salt
17:44 aqua^c joined #salt
17:48 malinoff joined #salt
17:48 edeefelt joined #salt
17:48 Akhter joined #salt
17:49 denys joined #salt
17:50 ageorgop joined #salt
17:53 ashirogl1 joined #salt
17:55 giantlock joined #salt
17:56 edeefelt left #salt
17:57 timoguin joined #salt
17:57 big_area anyone have a practical example of using engines?  I'm trying to wrap my head around them
17:57 kukacz joined #salt
17:58 perfectsine joined #salt
17:59 debian112 joined #salt
18:00 iggy they are pretty new, probably not a whole lot of docs/walkthru's on them
18:00 kukacz joined #salt
18:02 oznah joined #salt
18:02 felskrone joined #salt
18:02 rim-k joined #salt
18:04 bhosmer_ joined #salt
18:05 patchedmonkey joined #salt
18:07 Akhter joined #salt
18:07 winsalt my pillar problem solved itself, but I dont think i changed anything..
18:09 katyucha joined #salt
18:10 mpanetta joined #salt
18:11 mpanetta joined #salt
18:12 oznah I need to write a state to create a bunch of directories. The problem isn't creating /dir1/dir2 etc. The problem is, inside of dir2 a dir for each letter of the alphabet
18:12 writtenoff joined #salt
18:13 Akhter joined #salt
18:13 oznah last message should read... dir2 I need a dir for each letter of the alphabet
18:13 oznah seems like jinja should be able to help, but I am not sure how exactly
18:14 baweaver joined #salt
18:14 oznah call me lazy, but I am trying to avoid writing salt state for each /dir1/dir2/a; /dir1/dir2/b... etc
18:16 felskrone joined #salt
18:17 winsalt would you be able to have the directory tree you want on the master, and just copy it over?
18:20 oznah I guess I could. I briefly considered that.
18:20 pravka joined #salt
18:20 kitplummer joined #salt
18:21 oznah to a wrinkle directory a-z actually need to be symlinks to /home
18:21 oznah does that make it easier or harder
18:21 oznah sorry... to add a wrinkle...
18:24 oznah The actual dir structure will look like this /dir1/dir2/a-z/a-z --> /home
18:25 Adam____ joined #salt
18:25 Guest9366 left #salt
18:26 Adam____ hi. where can i find information on what OS salt supports?
18:26 tanta um use brace expansion
18:26 tanta bash has a built-in feature that can accomplish that without looping or weird steps oznah
18:27 tanta such as: mkdir {a..z}
18:30 oznah ok, so bash is the best tool, how would I saltify that.
18:31 genomer joined #salt
18:33 whytewolf oznah: cmd.run with - creates
18:35 oznah whytewolf: i was just looking at that
18:35 GreatSnoopy joined #salt
18:38 moogyver big_area: there's a few examples in the git repo of different engines
18:38 big_area thanks moogyver
18:39 moogyver big_area: i've been looking at them as well.  it's hard to pinpoint exactly where they're useful.  i don't like the test engine example because i think it's sorta backward as to how they're useful ( ie, listening to events and doing something with them ), as that's really the reactor's purview
18:40 moogyver i think you get more use out of them if you're using them to talk to some other service and based on results from those services, do something with salt ( start a exec mod, a state, exec a runner, etc )
18:40 big_area ya it seems liek theyre an evolution of returners and the api. useful in either direction
18:41 moogyver there's a slack engine in there that's a good example - allows you to send commands to salt from the slack client
18:42 zmalone that seems terrifying
18:42 Fiber^ joined #salt
18:42 tanta I hate slack
18:43 Ryan_Lane tanta: it's the future. it's going to replace email and irc and phones and video chat and probably breathing at some point.
18:43 moogyver zmalone: there's restraints around it, not like it's purely open
18:43 moogyver and tanta: why?
18:44 tanta because it's just one more channel for clueless idiots to ask me dumb questions and it confers an air of immediacy that even email lacks
18:44 murrdoc damn
18:44 murrdoc for angel
18:44 murrdoc u have angre
18:44 murrdoc anger*
18:45 moogyver so, you hate slack not because it's client or it's tech, but because it's just another chat client.
18:45 moogyver got it
18:45 timoguin_ joined #salt
18:45 tanta aye, that's the main complaint, I prefer the old days when 'IM at work' was considered a bad idea
18:46 tanta plus it's in the middle of our new open-concept office where I can already smell everyone on my team
18:46 murrdoc we have a slack chat for salt ?
18:46 * murrdoc hates it already
18:47 iggy the only thing slack is good for is immediate UD lookups, bear facts, and kitten gifs
18:49 murrdoc yes
18:49 murrdoc sucks
18:50 Norrland Simple, don't use it.
18:50 iggy "THANK YOU FUR SIGNING UP FOR BEAR FACTS! BEARS CAN INSTALL UBUNTU IN ONLY 1856 TRIES!"
18:51 iggy http://i.giphy.com/ECwTCTrHPVqKI.gif
18:52 iggy there, now you can skip Slack for the day
18:53 tanta it creates more cross-team synergies by breaking down silos and enabling real-time dynamic collaboration
18:55 big_area and it does it all in the cloud
18:55 big_area because it is agile
18:57 tanta you guys are awesome
18:59 mike_s joined #salt
19:00 zmalone joined #salt
19:01 baweaver joined #salt
19:01 mike_s Hello.  I'm trying to use Salt to manage an Apache Tomcat web application.  When I change my server /srv/salt/webserver/tomcat.sls  entry for the .war file so that the next time I do "salt 'webfoo*' state.highstate" the .war is updated, I want to
19:02 mike_s trigger the automatic deletion of compiled Java jsps and other cache files and a restart of the service.   I'm sure this kind of thing is in the docs, but I'm having problems finding it.
19:02 iggy prereq maybe?
19:02 mike_s I can put something in the .sls file that always deletes the compiled jsps and such when I run state.highstate, but I'd prefer to only have it trigger off changes to the app.war
19:03 mike_s If I put the prereq at the top level of the .sls file, in its own stanza, I think it runs every time, right?
19:03 pm90__ joined #salt
19:04 falenn joined #salt
19:04 pm90_ joined #salt
19:05 clintberry2 joined #salt
19:05 Akhter joined #salt
19:05 dendazen_ joined #salt
19:06 mike_s I guess that's not the end of the world.  I'll try to get it running that way.  Deleting the precompiled .jsps too often is better than not often enough.  Thanks anyway.
19:07 blast_hardcheese joined #salt
19:08 iggy prereq is a built-in requirement specifier (like creates, require, onchanges, etc.)
19:08 iggy https://docs.saltstack.com/en/latest/ref/states/requisites.html#prereq
19:14 murrdoc1 joined #salt
19:16 patchedmonkey joined #salt
19:17 Akhter joined #salt
19:18 mike_s Thanks!  I think I might be able to use the 'onchanges' hook on the same page to do what I need.
19:23 jeffpatton1971 joined #salt
19:25 onovy hi, is 2015.5.6 already released? can't found it on pypi, no tag, but release note (https://docs.saltstack.com/en/2015.5/topics/releases/2015.5.6.html) exists :)
19:26 jeffpatton1971 what's the easiest way to push salt-minion to a bare server?
19:27 moogyver jeffpatton1971: there's the bootstrap stuff
19:27 jeffpatton1971 is that in the docs?
19:27 moogyver https://github.com/saltstack/salt-bootstrap
19:27 jeffpatton1971 lol thank you :)"
19:27 moogyver https://docs.saltstack.com/en/latest/topics/tutorials/salt_bootstrap.html
19:27 moogyver it works ok - it's not fantastic
19:28 moogyver it still requires you to have certain repos setup
19:29 DanyC joined #salt
19:30 jeffpatton1971 so you still need to setup the ppa's in advance?
19:30 jeffpatton1971 hrm
19:31 moogyver PPA's?
19:31 lexter joined #salt
19:32 moogyver oh, debian/ubuntu thing i guess. :)
19:33 kitplumm_ joined #salt
19:34 jeffpatton1971 ya sorry
19:35 moogyver it'll setup the salt one's - at least it sets it up for the rhel yum repos
19:35 moogyver but if you're missing certain packages that are only available in the os repos, it doesn't set those up
19:36 baweaver joined #salt
19:37 iggy jeffpatton1971: saltify salt-cloud provider
19:38 jeffpatton1971 I've not had any luck with that on 2015.5.5
19:38 jeffpatton1971 we're all up in the azure
19:38 iggy there's your problem...
19:38 * iggy runs
19:38 jeffpatton1971 lol
19:38 hal58th joined #salt
19:39 jeffpatton1971 it's not too big of a problem at the moment, we've got about 520 cores in azure all Linux which is super entertaining to me
19:40 patchedmonkey joined #salt
19:43 khaije1 are there any best practices around managing minion config from a master?
19:45 pm90_ joined #salt
19:47 babilen khaije1: Use the salt-formula
19:49 chiui joined #salt
19:50 moogyver is there a way when using __runners__ to fire another runner to do it async?
19:52 thebwt joined #salt
19:53 baweaver joined #salt
19:54 larsfronius joined #salt
19:55 patchedmonkey joined #salt
19:56 bVector what is a good way to tell (on the minion) that the minion is failing auth or not connecting?
19:57 bVector I want to have an out-of-band system to remediate disconnected minions
19:57 blast_hardcheese bVector: Whether the minion can talk to the master?
19:57 bVector yeah
19:57 blast_hardcheese salt-call test.ping immediately comes to mind
19:58 blast_hardcheese there might be a more graceful way to do it though
20:04 Nebraskka if anybody had such experience, is it possible to install salt-minion on windows as one exe, without rolling out the whole system-wide dependency packages? having debates with my colleague
20:04 denys_ joined #salt
20:05 Nebraskka his idea - there are tons of devops tools, but none simple enough for infrastructure orchestration (he can't just easily rollout pip on every workstation, internet here is restricted, and other limitations)
20:06 Nebraskka and he don't want to mess around with internetless multiple *.msi with dependencies
20:07 Nebraskka if there is no such solution currently, it's nothing :) maybe just someone saw such compiled 1-exe minion versions
20:08 aparsons joined #salt
20:08 kitplummer joined #salt
20:08 Nebraskka he searching for something that don't require to rollout 10+ dependencies, and if possible, serverless
20:08 timoguin joined #salt
20:10 felskrone joined #salt
20:11 dthom91 joined #salt
20:14 Akhter joined #salt
20:18 baweaver joined #salt
20:18 andrew_v joined #salt
20:27 kawa2014 joined #salt
20:31 felskrone joined #salt
20:31 dendazen_ is there a salt state to update root password on 500 hosts using salt?
20:31 dendazen_ and is it ok way to do it?
20:32 bVector user.present would work
20:32 bVector but you would need to test and make sure you have a backup plan in case you lock yourself out :P
20:32 qybl joined #salt
20:34 dendazen_ well we use keys to access
20:34 dendazen_ anyway
20:34 dendazen_ but in general do people use salt to do this procedure?
20:34 dendazen_ currently i have a perl script i run
20:35 dendazen_ but i am tired of it.
20:35 babilen Sure, why not?
20:36 dendazen_ Ok, thanks, guys i will create the state.
20:36 murrdoc1 whats the —out for seeing only changes
20:37 bVector depending on how well the perl script is written (being amiable) salt would probably give you a much better experience
20:37 pm90_ joined #salt
20:41 stanchan joined #salt
20:43 traph joined #salt
20:43 MadHatter42 joined #salt
20:43 conan_the_destro joined #salt
20:45 murrdoc1 Gareth:  u around
20:45 murrdoc1 question about mounts, seeing as u fixed a bunch of them
20:46 pm90__ joined #salt
20:46 sunkist joined #salt
20:47 Gareth murrdoc1: I am.  mounts....there be monsters.
20:48 bluenemo joined #salt
20:48 murrdoc1 :D
20:48 murrdoc1 Forced unmount and mount because options (nfsvers=3) changed
20:48 murrdoc1 thats wrong right
20:48 murrdoc1 like it shouldnt remount
20:49 murrdoc1 originb:/ifs/LOL/LOL              /LOL/LOL      nfs     _netdev,nfsvers=3,proto=tcp,rsize=32768,wsize=32768,intr,soft,timeo=60,nosuid,sync        0 0
20:49 murrdoc1 its in the fstab
20:49 murrdoc1 anything stand out ?
20:49 subsignal joined #salt
20:52 dthom91 left #salt
20:56 cpowell joined #salt
20:56 alemeno22 joined #salt
20:57 jcockhren fixes for the salt provisioner -> https://github.com/mitchellh/vagrant/pull/6382
20:57 jcockhren vagrant ^
20:58 twork i think i've askd this before, but... is it common/advised practice to extend a formula by adding states that refer to stuff inside another formula? i've been doing so (putting vsftpd stuff inside my deployment of the 'users' formula) but i wonder if i'm setting myself up for grief.
20:59 tanta make the users pillar available globally and do whatever work you need to in the vsftpd formula
21:00 twork tanta: ok, if i follow you, that's what i've been doing. tx.
21:01 murrdoc1 Gareth:  figured it out
21:01 murrdoc1 nfsvers=3 bad
21:01 murrdoc1 vers=3
21:01 murrdoc1 good
21:01 murrdoc1 no remount if you do the latter
21:01 murrdoc1 Gareth:  eom
21:02 Gareth murrdoc1: like I said...there be monsters.
21:04 conan_the_destro joined #salt
21:05 Akhter joined #salt
21:05 perfectsine joined #salt
21:09 murrdoc1 you sir
21:09 murrdoc1 need to moveto the beach
21:09 murrdoc1 so i can thank you in person
21:09 murrdoc1 o/
21:09 DanyC joined #salt
21:11 Gareth murrdoc1: saltconf '16 :)
21:11 murrdoc1 you will be hugged
21:11 murrdoc1 #hugops
21:11 Rebus joined #salt
21:12 aurynn :)
21:12 aurynn consensual hugging
21:13 jcockhren murrdoc1++
21:14 clintberry2 joined #salt
21:16 Akhter joined #salt
21:18 chiui joined #salt
21:18 PI-Lloyd joined #salt
21:19 teebes joined #salt
21:19 rim-k joined #salt
21:20 cberndt joined #salt
21:24 hal58th 6 months from now you are going to forget but murrdoc1 won't and it will be awkward.
21:26 Akhter joined #salt
21:26 bVector hugging at a conference is always awkward
21:26 bVector you never know if someone has changed distros behind your back
21:27 moogyver bVector: :O
21:27 murrdoc1 consensual?
21:27 murrdoc1 no way
21:28 murrdoc1 its consensual if i yell surprise hug
21:28 murrdoc1 then i ll make gareth solve a problem on a whiteboard
21:28 murrdoc1 #throwbackthursday
21:29 bVector guys, this is how you get con flu
21:29 mpanetta or the zombie plague
21:30 murrdoc1 con FLU
21:30 murrdoc1 also not coming to saltconf
21:31 murrdoc1 unless base of pi gives me a usable 2015.5.5 deb
21:31 murrdoc1 then maybe
21:33 jalbretsen joined #salt
21:34 Akhter joined #salt
21:37 kitplummer joined #salt
21:38 mvensky joined #salt
21:39 quique joined #salt
21:44 baweaver joined #salt
21:47 Bryson joined #salt
21:48 druonysus joined #salt
21:48 druonysus joined #salt
21:53 breakingmatter joined #salt
21:54 MadHatter42 can i use salt for a certain report
21:54 bVector yes
21:54 MadHatter42 like openssh version installed on my minions
21:54 MadHatter42 openssl and other stuff
21:54 MadHatter42 ?
21:55 MadHatter42 hmm, how ?
21:55 bVector depends on how you want to do it
21:55 bVector you can use states to assert that a certain version is installed, and then report discrepancies
21:55 zmalone cmd.run "ssh —version"
21:55 zmalone etc.
21:55 MadHatter42 i decided to go with salt and now i'm reading documentation
21:56 MadHatter42 interesting
21:56 MadHatter42 beside the cmd.run
21:56 gekitsuu joined #salt
21:56 MadHatter42 is there a package.list | grep ...
21:56 MadHatter42 ?
21:56 geomyidae_ joined #salt
21:57 bVector which OS
21:57 bVector https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.pkg.html
21:57 MadHatter42 debian
21:57 MadHatter42 well i know i can cmd.run "dpkg --list | grep openssh "
21:57 MadHatter42 but thats quite not the point
21:57 bVector aptpkg.info_installed
21:58 bVector err
21:58 bVector pkg.info_installed
21:58 bVector https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.aptpkg.html#salt.modules.aptpkg.info_installed
21:58 MadHatter42 nice
21:59 bVector to get the most out of salt, you might look at turning your question into a statement
21:59 iggy MadHatter42: test=True is fun to see what might change too
22:00 bVector think about it like 'I want these packages installed, at this version', make a state, and use test=True and just report on packages that dont match what you think should be there
22:00 bVector or, let salt fix it for you automatically
22:00 MadHatter42 uuuuu
22:00 MadHatter42 that sound awesome
22:00 whytewolf you can query the pkg database with get_selections also.
22:01 MadHatter42 i've been using perl mostly
22:01 whytewolf https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.aptpkg.html#salt.modules.aptpkg.get_selections
22:01 MadHatter42 for years
22:01 bVector you would use pkg.installed in states
22:01 MadHatter42 for these stuff
22:01 bVector https://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkg.html
22:02 whytewolf welcome to sanity then
22:02 bVector ^
22:07 bVector ahh #$!@ yes! aws lambda supports python now!
22:11 _JZ__ joined #salt
22:12 druonysuse joined #salt
22:13 moloney joined #salt
22:14 aqua^c joined #salt
22:14 moloney It seems like the service.running state will call "update-rc.d" when "enable=True".  But shouldn't it be calling "insserv" instead so that LSB headers are used?
22:15 moloney On ubuntu/debian systems at least
22:15 kitplumm_ joined #salt
22:25 kitplummer joined #salt
22:30 aidalgol joined #salt
22:31 kitplummer joined #salt
22:31 aidalgol Why do the version numbers here all end in '150'? https://github.com/saltstack/salt-winrepo/blob/master/python2_x86.sls
22:33 whytewolf aidalgol: so that if the true package version  doesn't change but they need to increase the package number they can? I honestly have no idea but that is my guess
22:33 hal58th_ joined #salt
22:39 iggy moloney: ubuntu doesn't have anything called insserv
22:39 iggy (at least not the ubuntu box I have access to)
22:40 whytewolf aidalgol: unforchantly the commiter for the change had such lovely code commit as "Update python2_x86.sls"
22:41 moloney iggy: its not on the path, its at /usr/lib/insserv/insserv
22:42 aidalgol whytewolf: fun. that's as good a guess as any, I suppose. :P
22:42 moloney iggy: using update-rc.d doesn't seem to respect the LSB headers in my init script, but using insserv directly does
22:44 bfoxwell joined #salt
22:48 iggy I'm guessing they didn't opt for it for some reason
22:48 iggy you could always open a ticket to ask why (it should probably be documented somewhere why...)
22:49 stanchan joined #salt
22:49 bVector does a disconnected minion send events?
22:49 bVector trying to listen chatter on the local minion if auth fails or the like
22:50 cav3at joined #salt
22:52 blue0ctober joined #salt
22:53 baweaver joined #salt
22:54 markm joined #salt
23:00 iggy I think it depends on the version
23:00 iggy I know reactors don't work (yet)
23:00 iggy but I think there's an event bus and some chatter on it
23:01 murrdoc1 salt needs a runonce
23:01 murrdoc1 cmd.runonce
23:02 Ryan_Lane murrdoc1: isn't that cmd.run + unless/onlyif? :)
23:02 murrdoc1 no
23:02 murrdoc1 it isnt ryan
23:02 murrdoc1 it isnt
23:03 murrdoc1 :)
23:05 Ryan_Lane :)
23:06 murrdoc1 you know why not ?
23:06 murrdoc1 cos not all scripts actually touch a file
23:06 Ryan_Lane unless/onlyif can run commands
23:06 timoguin joined #salt
23:06 Ryan_Lane not just check to see if a file was touched
23:06 murrdoc1 u sir
23:06 murrdoc1 have that green field optimism
23:07 * murrdoc1 is jelly
23:07 * Ryan_Lane shrugs
23:07 murrdoc1 its a 4 year old python script
23:07 murrdoc1 cant tocuh it , cos well cant
23:07 murrdoc1 but it needs to be run once at provision time with flags
23:07 murrdoc1 and then its cron'd
23:07 Ryan_Lane you /could/ also do: cmd.run: -name: do_action && touch /tmp/did_action - unless: test -f /tmp/did-action
23:07 murrdoc1 it doesnt touch a file, but cant touch it
23:07 murrdoc1 yeah
23:08 Ryan_Lane you can also trigger a grain set
23:08 quique joined #salt
23:08 Ryan_Lane and only run the state if the grain isn't set
23:08 Ryan_Lane I guess it would be more convenient if salt kept track of that for you, though
23:08 murrdoc1 what i was hoping from runonce was that it put the file either in cache or in /etc
23:09 murrdoc1 well yeah what u said
23:09 murrdoc1 :)
23:09 * murrdoc1 goes back to work, thanks for listening to rant
23:11 cav3at left #salt
23:16 mosen joined #salt
23:22 mvensky_ joined #salt
23:24 mvensky_ joined #salt
23:24 dendazen joined #salt
23:27 khaije1 is there a state that will allow me to do the equivalent of passwd -l some_user ?
23:35 kermit joined #salt
23:39 iggy khaije1: do you really want to just lock it (they can still login with keys, etc.)?
23:39 iggy if so, I think you can just manually put ! at the beginning of the hash
23:42 und1sk0 left #salt
23:46 quasiben joined #salt
23:49 conan_the_destro joined #salt
23:50 conan_the_destro joined #salt
23:55 larsfronius joined #salt

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