Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-12-03

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

All times shown according to UTC.

Time Nick Message
00:04 genediazjr joined #salt
00:11 smcquay joined #salt
00:14 glyf joined #salt
00:17 TheThing joined #salt
00:18 Damon joined #salt
00:18 unpaidbi1l am i missing something here? it looks like yamlex is acting different when run from within salt than from the command line http://pastebin.com/WnUyTsrA
00:18 conan_the_destro joined #salt
00:20 murrdoc joined #salt
00:21 otter768 joined #salt
00:23 elfixit joined #salt
00:23 xDamox joined #salt
00:24 ajolo joined #salt
00:25 murrdoc joined #salt
00:26 murrdoc http://bencane.com/2014/07/17/integrating-saltstack-with-other-services-via-salt-api/
00:26 murrdoc excellent article
00:30 shaggy_surfer joined #salt
00:33 bhosmer_ joined #salt
00:34 ajolo joined #salt
00:40 ajolo joined #salt
00:42 patrek joined #salt
00:44 glyf joined #salt
00:45 rypeck joined #salt
00:47 kickerdog joined #salt
00:54 ajolo joined #salt
00:54 druonysuse joined #salt
00:54 druonysuse joined #salt
00:57 bhosmer_ joined #salt
01:00 ajolo joined #salt
01:01 tristianc joined #salt
01:01 TK_ joined #salt
01:02 aparsons joined #salt
01:04 davet joined #salt
01:05 MugginsM joined #salt
01:06 torment joined #salt
01:07 snuffeluffegus joined #salt
01:08 dunz0r joined #salt
01:09 druonysuse joined #salt
01:09 druonysuse joined #salt
01:10 shaggy_surfer joined #salt
01:10 genediazjr joined #salt
01:10 joehh1 joined #salt
01:10 khalieb joined #salt
01:15 paha joined #salt
01:19 Ryan_Lane joined #salt
01:24 ingwaem joined #salt
01:25 ingwaem left #salt
01:28 ajolo joined #salt
01:35 bhosmer_ joined #salt
01:42 aqua^mac joined #salt
01:58 ice_ joined #salt
01:59 ice_ anyone else?
02:00 malinoff joined #salt
02:01 TK_ joined #salt
02:04 druonysus joined #salt
02:16 bhosmer_ joined #salt
02:17 Ryan_Lane joined #salt
02:20 dork ice_: missed what you asked
02:20 dork think the network was split
02:21 otter768 joined #salt
02:25 jimklo_ joined #salt
02:40 desposo joined #salt
02:41 shaggy_surfer joined #salt
02:49 TheThing joined #salt
02:53 nitti joined #salt
02:54 aurynn is the minion ID stored in the grains or the pillar, at all?
02:55 malinoff aurynn, grains
02:57 kermit joined #salt
02:59 aurynn malinoff, great, thanks
03:01 TK_ joined #salt
03:03 Ryan_Lane joined #salt
03:08 Mso150_p joined #salt
03:09 shaggy_surfer joined #salt
03:12 TK_ joined #salt
03:28 monkey66 joined #salt
03:31 aqua^mac joined #salt
03:33 CeBe1 joined #salt
03:37 bdols joined #salt
03:42 MugginsM so. I'm writing a big complex fabric script to deploy ... salt master. Is there a better way?
03:43 MugginsM very chicken and eggy
03:43 MugginsM needs to set up repos, install things like git, salt, pull config from git, etc.
03:43 malinoff MugginsM, use already installed salt master to deploy salt-master
03:43 malinoff an other salt-master*
03:44 MugginsM yeah
03:44 aurynn I use salt-ssh for this particular yak
04:01 sophocles joined #salt
04:01 __number5__ MugginsM: we use AWS AMI so less eggy :) and you can call your states in master-less mode to deploy the salt-master
04:04 MugginsM we've got about 6 sites each will have their own master so automated deploy for that is good
04:05 MugginsM yeah, except we need quite a lot to get to the point where salt can run.  new apt repository, git, creds for access to our git repo, user, etc
04:05 MugginsM tried it with a standalone minion, but is easier with fabric (which we already use)
04:05 MugginsM seems to be working now so at least there's that
04:06 MugginsM also have salt-master in a docker container, which seems to be working nicely
04:08 __number5__ if it's working without any big issue then don't bother to change. Salt is very flexible
04:08 MugginsM I guess salt-ssh is the best way to kick start new minions?
04:09 MugginsM we'll have an ip address and some ssh keys for each one
04:13 __number5__ MugginsM: if you already have many hosts need to be "saltified" yes. For new minions, better to install salt-minion directly (e.g. via salt-bootstrap.sh)
04:14 MugginsM salt-bootstrap.sh doesn't really work for us, can't have production sites calling out to random internet sites
04:14 MugginsM we've got a local apt-mirror that we use and the packages are fine
04:17 ajolo joined #salt
04:17 __number5__ MugginsM: then I don't see why you'll need salt-ssh. just apt-get install and set the master ip you'll be fine
04:17 MugginsM something has to do the apt-get install :)
04:19 Steve7314 left #salt
04:20 genediazjr joined #salt
04:22 otter768 joined #salt
04:23 shoemonkey joined #salt
04:26 thedodd joined #salt
04:26 jimklo joined #salt
04:30 meylor joined #salt
04:33 TheThing_ joined #salt
04:43 voxxit joined #salt
04:52 genediazjr joined #salt
04:59 dvestal joined #salt
05:05 jalbretsen joined #salt
05:13 TK_ joined #salt
05:20 aqua^mac joined #salt
05:22 monkey66 left #salt
05:29 TK_ joined #salt
05:30 jhoffa joined #salt
05:35 linjan joined #salt
05:37 sophocles_ joined #salt
05:37 felskrone joined #salt
05:39 TyrfingMjolnir joined #salt
05:40 felskrone1 joined #salt
05:41 genediazjr joined #salt
05:46 ramteid joined #salt
05:54 TK_ joined #salt
05:58 Furao joined #salt
06:01 beneggett joined #salt
06:04 catpiggest joined #salt
06:16 jhoffa joined #salt
06:17 T-12 joined #salt
06:18 T-12 left #salt
06:23 otter768 joined #salt
06:24 snuffeluffegus joined #salt
06:25 shoemonkey joined #salt
06:29 oyvjel joined #salt
06:32 jp_ joined #salt
06:32 jp_ is there a way to shorten the time it takes to run salt commands i.e. shorteingin the cycle when stuff is pickuped up from the queue?
06:39 gildegoma joined #salt
06:41 Valdo joined #salt
06:41 SpX joined #salt
06:41 Diaoul joined #salt
06:42 Deevolution joined #salt
06:42 lahwran joined #salt
06:46 Blacklite joined #salt
06:46 Guest6778 joined #salt
06:47 simonmcc_ joined #salt
06:47 simonmcc_ joined #salt
06:50 genediazjr joined #salt
06:50 linjan joined #salt
07:00 jmux joined #salt
07:01 smcquay joined #salt
07:05 colttt joined #salt
07:09 aqua^mac joined #salt
07:11 smcquay left #salt
07:28 lcavassa joined #salt
07:35 snuffeluffegus joined #salt
07:36 flyboy joined #salt
07:39 jhoffa joined #salt
07:42 yomilk joined #salt
07:48 jhoffa_ joined #salt
07:52 lb1a joined #salt
07:54 hojgaard joined #salt
07:58 lcavassa joined #salt
07:58 Outlander joined #salt
08:01 genediazjr joined #salt
08:04 linjan joined #salt
08:11 stoogenmeyer joined #salt
08:12 stoogenmeyer hey is there anybody here using salt to orchestrate a docker deployment to multiple hosts and can give some recommendation on how to get started?
08:13 tomspur joined #salt
08:15 aurynn I'm doing something similar
08:16 aurynn basically I stole http://bitjudo.com/blog/2014/05/13/partial-continuous-deployment-with-docker-and-saltstack/ as the base and went from there
08:18 bhosmer_ joined #salt
08:20 malinoff Is there any article describes that managing multiple docker containers with a single service within is *better* than managing a single VM with multiple services?
08:20 malinoff s/that/why
08:21 malinoff i really can't understand why people use docker, i can't get its purpose
08:21 aurynn docker makes my developers happier; they can write stuff that Just Works
08:21 malinoff i can write stuff that just works without docker, so it's not an answer
08:21 JlRd joined #salt
08:22 aurynn ... I'm pretty sure "you" didn't enter my sentence at all
08:23 aurynn In fact, I said it makes my developers happy, because they can easily provision containers that work without jumping through VM hoops
08:23 babilen docker is a VM
08:23 aurynn no
08:23 babilen (of sorts)
08:23 babilen well, container
08:23 malinoff I actually did. Here is the example: python makes me happy, because I don't need to write tons of boilerplate code to work with simple data structures. It saves my time. What docker saves?
08:23 aurynn it's a security threat
08:23 BigBear joined #salt
08:24 aurynn malinoff, perhaps you should venture into #docker and ask?
08:24 otter768 joined #salt
08:25 malinoff I don't like to ask such questions in interested community, it's better to hear some thoughts from people who just use it
08:26 malinoff I'm asking that, because i don't really see the difference between launching a VM or running a container if you have a provisioner (e.g. salt), and i'm curious why people involve an other tool to solve the same problem
08:26 shoemonkey joined #salt
08:26 aurynn yes, but you aren't listening to the answer. you're asserting that your world is better regardless of the answer, and while that's fun for the first five minutes, maybe you could do that somewhere else?
08:27 Mso150 joined #salt
08:28 malinoff With such assumptions, the discussion will not be good
08:28 aurynn indeed, that's what I just said
08:29 malinoff well, thanks for your input anyway
08:37 slafs joined #salt
08:37 slafs left #salt
08:39 genediazjr joined #salt
08:42 karimb joined #salt
08:52 superted666_ Morning all, got a strange problem where an update to a custom module is causing the salt minion to fail after it's started.
08:52 superted666_ There are no errors about the actual code and it works through salt call, the code is quite simple. :
08:53 superted666_ if 'string' in open('/etc/salt/grains').read():
08:53 superted666_ print "String detected, not running"
08:53 superted666_ exit(0)
08:53 superted666_ But salt just silently dies when i add this to the module
08:53 superted666_ even in debug
08:53 superted666_ This is on 2014.1.13
08:56 aurynn wrap it in a try/except and see what logs?
08:57 aqua^mac joined #salt
08:58 superted666_ Would that output the error to the salt logs or the console?
08:59 aurynn if you use logging.whichever in the except block, it should log to salt's logs
08:59 summers joined #salt
09:04 spo0nman joined #salt
09:07 CeBe joined #salt
09:11 dRiN joined #salt
09:13 genediazjr joined #salt
09:15 Pixionus joined #salt
09:16 ninkotech_ joined #salt
09:17 lb joined #salt
09:23 potens joined #salt
09:25 stooj joined #salt
09:26 wnkz joined #salt
09:26 intellix joined #salt
09:27 wnkz__ joined #salt
09:28 jhauser joined #salt
09:28 Auroch joined #salt
09:29 abo1234 joined #salt
09:29 abo1234 hi
09:30 abo1234 left #salt
09:39 mick3y blaffoy: i looked at the update_url option, and I already have 'salt-minion' in update_restart_services. the problem is when i run highstate, which will update salt-minion on the host, the execution stops and i get 'minion did not return' (or something similiar). the minion will come back up after restarting and i can re-run required state agains the host. it's just tad annoying
09:39 fredvd joined #salt
09:40 Auroch joined #salt
09:42 Pixionus joined #salt
09:46 CeBe joined #salt
09:48 Outlander joined #salt
09:50 BigBear joined #salt
09:52 felskrone joined #salt
09:55 blaffoy mick3y: right. UtahDave was on here yesterday evening saying that the esky update option isn't ready for primetime yet. So I'm trying my update a different way. That's led to other problems though.
09:57 mick3y blaffoy: the easiest way IMVHO would be to run salt-minion update/upgrade as the very last state
09:58 sieve joined #salt
10:00 ntropy i have state_verbose: False and state_output: changes in my minion conf, but why do i still get verbose output from highstate when called with salt-call?
10:00 blaffoy mick3y: what OS(es) are you working on? I've my master on Ubuntu, and my minions are Windows 7, Windows Server 2008, and Linux. Updating Linux minions should be a simple pkg call (except I'll need to make sure that they're pulling from the right pkg), but upgrading the windows minions will be more tricky.
10:01 flyboy82 joined #salt
10:02 eject_ck joined #salt
10:06 Mso150 joined #salt
10:10 Auroch joined #salt
10:16 danburbridge joined #salt
10:17 Auroch_ joined #salt
10:20 wnkz joined #salt
10:25 otter768 joined #salt
10:27 shoemonkey joined #salt
10:28 eliasp blaffoy: this is how I'm doing my windows updates: https://gist.github.com/eliasp/cd962142e14c34569504
10:29 blaffoy eliasp: that looks solid.
10:34 Auroch joined #salt
10:36 Auroch joined #salt
10:37 wnkz joined #salt
10:39 N-Mi_ joined #salt
10:39 N-Mi_ joined #salt
10:40 zerthimon joined #salt
10:42 Auroch_ joined #salt
10:43 Auroch_ joined #salt
10:43 ckao joined #salt
10:46 mick3y blaffoy: centos all over - master on 6.5 and 5.x minions (+few 6.5). will get to this sometime later.
10:46 aqua^mac joined #salt
10:48 giantlock joined #salt
10:50 eject_ck joined #salt
10:54 wnkz__ joined #salt
10:55 Auroch joined #salt
10:55 Auroch joined #salt
10:57 X86BSD joined #salt
11:01 Auroch joined #salt
11:11 jrluis joined #salt
11:14 danburbridge pillar layout question - I understand the pillar top file targets hosts, I have database servers which I want specific databases to exist on (and only on) do I need to create a pillar sls file for each database?
11:14 diegows joined #salt
11:19 Furao you can match multiple host such as ‘db-*’ and include the same .sls on all hosts that start with ‘db-‘
11:20 Furao and mix that with macro and other jinja if/else
11:21 Furao or http://reclass.pantsfullofunix.net/
11:22 bhosmer_ joined #salt
11:32 TK_ joined #salt
11:41 felskrone joined #salt
11:41 Auroch joined #salt
11:42 Auroch joined #salt
11:43 TK_ joined #salt
11:52 Auroch joined #salt
11:52 faust joined #salt
11:53 Auroch joined #salt
11:54 Auroch joined #salt
11:54 Auroch joined #salt
12:01 glyf joined #salt
12:10 jonatas_oliveira joined #salt
12:16 tafa2 joined #salt
12:16 bhosmer_ joined #salt
12:16 qx joined #salt
12:19 qx_ joined #salt
12:22 CeBe joined #salt
12:24 CeBe1 joined #salt
12:26 otter768 joined #salt
12:28 shoemonkey joined #salt
12:34 johtso__ joined #salt
12:35 aqua^mac joined #salt
12:35 CeBe joined #salt
12:39 funzo joined #salt
12:40 wvds-nl joined #salt
12:45 JlRd joined #salt
12:47 TK_ joined #salt
12:47 gildegoma joined #salt
12:50 ericof joined #salt
12:52 TK_ joined #salt
12:54 Svake joined #salt
12:55 brayn joined #salt
12:56 wm-bot4 joined #salt
12:58 wm-bot4 joined #salt
12:59 ericof_ joined #salt
13:00 wm-bot4 joined #salt
13:02 wm-bot4 joined #salt
13:03 wm-bot4 joined #salt
13:04 sieve joined #salt
13:04 sieve joined #salt
13:05 sieve joined #salt
13:05 sieve joined #salt
13:06 wm-bot4 joined #salt
13:06 sieve joined #salt
13:06 sieve joined #salt
13:07 wm-bot4 joined #salt
13:09 wm-bot4 joined #salt
13:11 wm-bot4 joined #salt
13:14 wm-bot4 joined #salt
13:17 wm-bot4 joined #salt
13:17 hcl2 joined #salt
13:18 wm-bot4 joined #salt
13:20 wm-bot4 joined #salt
13:22 jonatas_oliveira joined #salt
13:22 wm-bot4 joined #salt
13:25 wm-bot4 joined #salt
13:26 wm-bot4 joined #salt
13:28 wm-bot4 joined #salt
13:29 oyvjel1 joined #salt
13:30 wm-bot4 joined #salt
13:30 oyvjel joined #salt
13:32 wm-bot4 joined #salt
13:32 maze I'm getting random "No matching sls found" errors on my minion, this is my fileserver config: https://maze.io/paste/p/oje555w2asies1z3
13:32 maze am I missing something?
13:32 maze current version is 2014.7.0
13:33 maze also, it seems that my custom grains dissapear all the time .. I must be doing something wrong?
13:33 wm-bot4 joined #salt
13:33 oyvjel1 joined #salt
13:36 wm-bot4 joined #salt
13:37 CeBe joined #salt
13:37 Gareth joined #salt
13:38 wm-bot4 joined #salt
13:40 thawes joined #salt
13:40 wm-bot4 joined #salt
13:42 wm-bot4 joined #salt
13:43 tafa2 joined #salt
13:44 wm-bot4 joined #salt
13:46 wm-bot4 joined #salt
13:47 eofs is there a way to make sure local.saltutil.sync_grains is run before highstate using Reactor System?
13:47 eject_ck joined #salt
13:48 wm-bot4 joined #salt
13:50 thawes joined #salt
13:52 TK_ joined #salt
13:54 maze ls
13:57 thawes joined #salt
14:00 Al_ joined #salt
14:02 Guest56489 What's the difference between salt.state vs salt.module in documentation? The documentation seems to be mixing both, it's not clear to distinguish
14:05 shoemonkey joined #salt
14:06 goal in file.replace, I don't suppose there's a way to say "don't do the replacement if the replacement already exists"
14:06 goal ?
14:06 eliasp Guest56489: "state" → handler for states as in your SLS files, "module" → execution module doing the hard work in the background
14:07 lcavassa joined #salt
14:07 eliasp goal: there is… and there's also a bugfix you should use in this case… https://github.com/saltstack/salt/issues/18612
14:07 eliasp goal: sorry, this is what you need: https://github.com/saltstack/salt/pull/18615
14:08 eliasp goal: the param you're looking for is search_only=True
14:09 eliasp goal: hmm, sorry… I think I got your initial question wrong
14:09 goal heh yes I was just about to say that ;)
14:09 goal but thanks for the link, as that would have affected me
14:10 eliasp ok, great :)
14:10 patrek joined #salt
14:11 goal but the original question remains
14:11 blaffoy joined #salt
14:12 thehaven joined #salt
14:12 eliasp goal: which params do you currently use and what does your original line look like, in which the replacement should be done?
14:12 eliasp goal: please nopaste…
14:14 cpowell joined #salt
14:14 blaffoy Hi, I tried to upgrade to my salt-master (ubuntu 14.04) and salt-minion (Windows 7) to the latest version of salt: 2014.7, but this seems to have broken a lot of commands. I can no longer use state configurations like pkg.installed and archive.extracted.
14:15 blaffoy I get an error like: State 'archive.extracted' found in SLS 'dev-win' is unavailable
14:15 quantumriff joined #salt
14:15 briffle_ joined #salt
14:15 thawes joined #salt
14:16 blaffoy What could have happened to these?
14:16 goal say there's a two lines, one '#key = foo' and the other 'key = foo'. I want to replace the value with 'bar', but also ensure it exists without the comment. If I replace using a regex that optionally includes the comment, I will result in two identical lines. If I could say 'only replace if the replacement doesn't exist' then it would prevent that.
14:17 Ahlee goal: you'd be better of taking the file under full management of salt and doing the logic there, rather tahn trying to play with regexes
14:17 goal In writing that I've just thought file.uncomment might be a good idea first
14:17 eliasp goal: depreceated in favor of file.replace… ;)
14:17 goal In my situation, as much as I'd like to do that, it's not really an option.
14:18 goal (was @ Ahlee )
14:18 goal argghh
14:18 Ahlee I hear ya.  Good luck man
14:18 goal so I'll have to replace for the commented version, then do the replace for the value
14:18 goal ..I'm guessing
14:19 eliasp goal: just trying something here… one moment
14:19 patrek joined #salt
14:20 briffle_ left #salt
14:20 quantumriff I am needing to move to a new master.. I have to build a new VM (salt was installed on a physical machine, that was doing lots of other stuff)
14:20 quantumriff any idea how to copy the data over to the new master, in a way that the minions all accept it?
14:21 quantumriff obviously, DNS needs updated, firewalls (this new vm is in a new datacenter), etc
14:22 Ahlee quantumriff: /etc/salt/pki should hold all the keys you need to move
14:23 Ahlee so, move that directory, move your file_roots and the like
14:23 Ahlee and you should be fine
14:23 quantumriff so really, its just a matter of installing salt-master, copying over /etc/salt (well, the master parts and pki) and then move /srv/ (which has salt, pillar, and reactor)?
14:23 sieve joined #salt
14:23 Ahlee Pretty sure that's it, yeah
14:23 quantumriff thanks very much!
14:23 Ahlee I don't believe the master key cares about the master's hostname
14:23 eliasp goal: meh… it empties my file trying to get done what you want to do… looks like I need to fix another bug in file.py :)
14:24 sieve joined #salt
14:24 mapu joined #salt
14:24 sieve joined #salt
14:24 mpanetta joined #salt
14:24 aqua^mac joined #salt
14:24 blaffoy I've lost access to pkgrepo.managed, pip.installed, archive.extracted after upgrading to 2014.7
14:24 eliasp Ahlee: subscribe to #18615 if you want… I'll use your example to describe the 2nd bugfix to it, so you'll get updated with a solution (which hopefully works for you) once I fixed the bug
14:24 sieve joined #salt
14:25 blaffoy I notice that these are all package management related.
14:25 sieve joined #salt
14:25 blaffoy Could there be some configuration of these tools that got nuked during the upgrade?
14:25 sieve joined #salt
14:26 Ahlee eliasp: I think you meant goal?
14:26 sieve joined #salt
14:26 sieve joined #salt
14:26 eliasp Ahlee: eh, sure… sorry ;)
14:26 eliasp goal: ^^
14:26 sieve joined #salt
14:27 otter768 joined #salt
14:27 sieve joined #salt
14:27 sieve joined #salt
14:28 sieve joined #salt
14:28 sieve joined #salt
14:28 sieve joined #salt
14:29 nitti joined #salt
14:29 oyvjel joined #salt
14:29 goal eliasp: watching now, thanks. I've applied those fixes anyway
14:31 giantlock joined #salt
14:32 micah_chatt joined #salt
14:33 dude051 joined #salt
14:34 micah_chatt_ joined #salt
14:39 CeBe joined #salt
14:39 sieve joined #salt
14:41 BigBear joined #salt
14:41 Auroch joined #salt
14:41 Roee joined #salt
14:42 Roee Hi All, question please
14:42 Roee i'm trying to configure a basic top.sls that will run a single state
14:42 superted666_ Does salt master need to be restarted to read in config from /etc/salt/master.d ?
14:43 Ahlee yes
14:43 superted666_ Okay doke, thanks
14:44 blaffoy Anybody? Please.
14:44 blaffoy I've tried completely removing the master and minion installations and reinstalling them
14:45 blaffoy And I've reverted the minion to 2014.1.11
14:45 Roee when runing state.highstate i'm getting an error
14:45 blaffoy I can't figure out how to revert the master.
14:45 Roee here you can see the whole pic : https://gist.github.com/roeera/3235f6be932db05e1d0c
14:46 quickdry21 joined #salt
14:46 Roee can someone tell why im getting the error please >
14:46 Roee ?
14:46 thedodd joined #salt
14:48 thawes joined #salt
14:49 dvestal joined #salt
14:49 kaptk2 joined #salt
14:51 blaffoy Roee: could you try hard coding the string "CollectServerParams.sh" into the line: /scripts/{{info_script}}:
14:52 elfixit joined #salt
14:52 lcavassa joined #salt
14:53 canci joined #salt
14:53 Roee i dont think the problem is inside the state, because it run perfect with state.sls comnand
14:53 Roee with no erorr
14:54 housl joined #salt
14:54 rbstewart joined #salt
14:56 Roee any other idea ?
14:57 glyf joined #salt
14:58 cjduan joined #salt
14:59 cjduan say hi
14:59 Kelsar hi
14:59 patrek joined #salt
14:59 jeddi joined #salt
15:02 goal Roee: increase log level or run with trace?
15:02 Roee seems that i'm getting this error even if i'm deleting my top file
15:03 Roee maybe the problem related to the minion ?
15:04 goal I would suggest not, if its a yaml error, but I'm no expert
15:05 goal as i say, increase verbosity/log level and see what you get
15:05 Roee i will check this
15:06 tafa2 joined #salt
15:09 asmaps left #salt
15:12 goal eliasp: Perhaps this is the problem you already know of, but it seems that if append_if_not_found=True, then it will append even if the replacement value already exists.
15:16 goal I don't see a way around that without somewhow having an option something like append_if_replacement_not_found (or otherwise you would first need to do a search/replace with identica path+pattern, in order to append the replacement. The do another search for the pattern you want to replace, in order to replace it with the repl you need)
15:16 goal convoluted!
15:19 mikkn joined #salt
15:23 _prime_ joined #salt
15:24 glyf joined #salt
15:26 sijis left #salt
15:28 markm_ joined #salt
15:29 TheThing joined #salt
15:30 sroegner joined #salt
15:31 UtahDave joined #salt
15:32 workingcats say, is there a way to change the tab width used by salt's output?
15:33 jaimed joined #salt
15:37 intellix joined #salt
15:39 tedski joined #salt
15:40 eliasp workingcats: the output is defined by so called outputters: http://docs.saltstack.com/en/latest/ref/output/all/index.html
15:40 workingcats eliasp, cheers, having a look
15:40 eliasp workingcats: so it depends on what the outputter you're using is capable of
15:40 workingcats ah ok
15:41 eliasp workingcats: you could also write a custom output module or just use plain/json/yaml output and process it outside of Salt
15:41 workingcats hum well the issue is that i want salt to print tabs as size 4 when printing the diff of files
15:42 workingcats i need to look at the txt outputter for htis i think?
15:42 murrdoc joined #salt
15:42 workingcats (as in, the diffs in the output of "salt-call state.highstate test=True")
15:43 eliasp AFAIK these diffs are generated by the respective state modules…
15:43 _JZ_ joined #salt
15:43 eliasp workingcats: you mean the diff of e.g. file.managed?
15:43 workingcats eliasp, yeah
15:43 CVVS joined #salt
15:43 workingcats exactly that one
15:44 eliasp ok, before we dive deeper into this: what's the reason you want to change this?
15:45 rickh563 joined #salt
15:45 eofs Should this Reactor state work? http://pastebin.com/yNuPet4J
15:46 eofs it seems that "require" is ignored
15:47 eliasp workingcats: btw: the diff is generated in modules.file.check_file_meta() around line 2891
15:48 Carlos_2 joined #salt
15:48 Carlos_2 Good Morning.
15:49 workingcats eliasp, well, i do my tabs as 4 width in my text editor, so the alignment (think crontab or fstab) is according to that
15:49 Carlos_2 I am trying to install salt
15:49 Carlos_2 Halite,  I download the Zip files form github and I execute the setup.py, but I cannot find more information in how to configure the gui
15:49 workingcats now when salt uses 8 width it's harder to read due to the "broken" visual alignment
15:50 Carlos_2 sorry I am trying to install halite*
15:53 eliasp workingcats: hmm, why are you importing the resulting diffs into your editor? I'm interested in your workflow here…
15:53 workingcats eliasp, i'm not importing the diffs
15:53 intellix joined #salt
15:53 workingcats but they're displayed in the shell with the wrong tab width
15:54 eliasp workingcats: looks like this is caused by 'difflib' which defaults to tabsize=8: https://docs.python.org/2/library/difflib.html#difflib.HtmlDiff.__init__
15:55 workingcats ah great
15:56 workingcats so i'll just hack it for now (will probably also fix that i have the same "bug" in git-cola) and when i have time (tm) clean it up and send a patch
15:56 StDiluted joined #salt
15:56 KennethWilke joined #salt
15:57 jab416171 joined #salt
16:02 pr_wilson joined #salt
16:05 conan_the_destro joined #salt
16:08 blaffoy Okay, after a bit more work, I've determined that upgrading to 2014.7 didn't break anything. There were a bunch of borked sls files in my config that I apparently hadn't run in a while until I did a state.highstate after my upgrade.
16:08 aquinas joined #salt
16:09 Ozack1 joined #salt
16:11 stevednd joined #salt
16:13 ajolo joined #salt
16:13 aqua^mac joined #salt
16:13 slav0nic joined #salt
16:19 TheThing joined #salt
16:23 patrek joined #salt
16:25 iggy I'm considering a cron'ed state.highstate test=True just for that reason
16:26 TyrfingMjolnir joined #salt
16:27 mreider joined #salt
16:27 mreider hello wondering if anyone can help w issue starting master
16:28 otter768 joined #salt
16:28 jimklo joined #salt
16:28 iggy without knowing what the problem is, I'd say start it in the foreground with -l debug
16:29 patrek joined #salt
16:30 ntropy why would some grains only be present when grains.items is called on a minion (with salt-call)?  when grains.items is called from the master, i dont see them
16:31 mreider getting ErrNo 99
16:31 mreider [root@centos2 salt]# cd /etc/init.d [root@centos2 init.d]# salt-master start -l debug [DEBUG   ] Reading configuration from /etc/salt/master [DEBUG   ] Guessing ID. The id can be explicitly in set /etc/salt/minion [INFO    ] Found minion id from generate_minion_id(): centos2 [DEBUG   ] Configuration file path: /etc/salt/master [INFO    ] Setting up the Salt Master [WARNING ] Unable to bind socket, error: [Errno 99] Cannot assign reques
16:31 RedundancyD joined #salt
16:31 Nazzy joined #salt
16:31 mreider checking  netstat ports seem to be open
16:33 UtahDave mreider: try this     service salt-master stop
16:33 UtahDave salt-master -l debug
16:37 Nazzy_ joined #salt
16:37 Nazzy_ joined #salt
16:38 StDiluted joined #salt
16:39 iggy ntropy: sounds like a bug, what version?
16:39 ntropy iggy: 2014.1.13, im just going through the logs, have some errors there, will check if thats the reason things arent working as expected
16:40 mreider [root@centos2 init.d]# salt-master start -l debug
16:40 mreider [DEBUG   ] Reading configuration from /etc/salt/master
16:40 mreider [DEBUG   ] Guessing ID. The id can be explicitly in set /etc/salt/minion
16:40 mreider [INFO    ] Found minion id from generate_minion_id(): centos2
16:40 mreider [DEBUG   ] Configuration file path: /etc/salt/master
16:40 mreider [INFO    ] Setting up the Salt Master
16:41 mreider [WARNING ] Unable to bind socket, error: [Errno 99] Cannot assign requested address
16:41 mreider The ports are not available to bind
16:41 mreider firewall is disabled
16:41 ntropy mreider: is the port already open by another process?  what does your netstat say?
16:42 iggy mreider: you probably have some phantom master processes running... you'll have to kill them by hand
16:42 iggy I've seen that happen a time or two
16:42 hybridpollo joined #salt
16:43 iggy ntropy: there's a bug that was fixed in 2014.7 that caused weird grain/mine return values
16:45 wnkz joined #salt
16:45 jalbretsen joined #salt
16:47 mreider did a ps -ef | grep salt
16:47 mreider not processes running
16:47 mreider what port should I grep for for netstat?
16:48 mreider 4505?
16:48 mreider also restarted the entire VM hosting master, that shouldve cleared any fantom procs
16:49 eliasp mreider: 4505, 4506 and in case you're using Halite 8080
16:50 thawes joined #salt
16:52 wendall911 joined #salt
16:56 djaime joined #salt
16:56 ericof joined #salt
17:01 jespada joined #salt
17:02 manytrees hi, what's the recommended way to manage the firewall on centos7? disable firewalld and use iptables?
17:02 spookah joined #salt
17:04 eject_ck joined #salt
17:05 babilen I am using the mysql-formula and would like to grant a user the ability to connect remotely .. What would I have to set to set in its pillar? It looks as if I'd use "host: '%'" or something along those lines. (cf. http://docs.saltstack.com/en/latest/ref/states/all/salt.states.mysql_user.html )
17:05 mreider manytree:  http://docs.saltstack.com/en/latest/topics/tutorials/firewall.html
17:06 babilen manytrees: Last time I checked there wasn't a very nice iptables formula around, but that might have changed in the interim. firewalld would be a frontend to iptables anyway ... you might want to take a look at ferm if you are looking for another.
17:07 gildegoma joined #salt
17:08 davet joined #salt
17:09 babilen What would I set salt.states.mysql_user.present's host value to if I want to allow connections from any host?
17:09 manfred manytrees: firewalld is terrible, i disable and then mask it, then use the iptables state to manage them
17:13 manytrees mreider: thanks, but i think that's initial salt setup not managing the rules with salt; babilen: there is an iptables state which looks ok; manfred: that's what i've done pre-salt, i'll just continue that way with salt --- thanks
17:14 CeBe1 joined #salt
17:14 iggy mreider: are you setting a listen address in the config?
17:16 monkey661 joined #salt
17:16 KyleG joined #salt
17:16 KyleG joined #salt
17:16 zlhgo_ joined #salt
17:17 conan_the_destro joined #salt
17:17 eliasp could anyone try to reproduce this? https://github.com/saltstack/salt/issues/18680
17:19 KennethWilke joined #salt
17:20 monkey66 joined #salt
17:22 kermit joined #salt
17:23 eject_ck joined #salt
17:24 kickerdog joined #salt
17:25 troyready joined #salt
17:26 sieve Hi, I am trying to find a very basic example of salt-cloud for making some AWS instances
17:26 UtahDave joined #salt
17:27 eject_ck joined #salt
17:31 rickh563 joined #salt
17:32 Svake joined #salt
17:34 budman joined #salt
17:34 hal58th joined #salt
17:37 budman Getting ready to roll salt out ... any tips/tricks :) or gotchas?
17:37 aparsons joined #salt
17:38 Gareth budman: prepare for awesomeness!
17:38 budman lol i cant wait. I see basic master/minion setup is cake. I assume you can then assign or group nodes into (web, db etc..)
17:39 desposo joined #salt
17:39 budman Any good guis? everyone using CLI?
17:39 Gareth No GUIs.
17:39 nkuttler there is a web ui in development
17:39 Gareth I've heard rumors of people using Foreman with Salt though.
17:39 nkuttler but yeah, what would you need a gui for anyway, most of the work is writing configs
17:40 blaffoy I am seeing a problem when installing a win repo package for mongodb. The package installs just fine on the minion, but on the master, after "salt '*' state.sls mongodb" it still reports an error
17:41 smcquay joined #salt
17:41 eliasp blaffoy: "reports an error" makes it impossible for us to help you…
17:41 eliasp blaffoy: what error?
17:41 Svake_ joined #salt
17:41 blaffoy eliasp: yep, just in the process of copying it out.
17:42 mpanetta joined #salt
17:42 eliasp ok ;)
17:42 budman I dont need a gui, but 2 other friends were mentioning about letting people push tasks via the web (not what I would want).
17:42 budman I dont see a huge benefit to the gui.
17:42 mpanetta joined #salt
17:43 blaffoy eliasp: actually, I don't get any error without turning on logging. with logging I see something like: [DEBUG] get_returns for jid 2014120317 sent to set(['myminion.myorg.local']) will timeout at 17:38:23
17:43 blaffoy Without logging, it just says "result: failed"
17:44 eliasp blaffoy: how does your state look like?
17:44 blaffoy But it will work just fine if I `salt-call state.sls mongodb`
17:44 blaffoy The state is literally just "mongodb:\n  pkg:\n    - installed"
17:45 eliasp hmm
17:46 eliasp blaffoy: what happens if you do "salt your-minion pkg.install mongodb"?
17:46 blaffoy Running `salt "myminion" pkg.install mongodb` works.
17:46 eliasp now that is actually weird
17:47 blaffoy The package installs. And I get a zero return code for the command
17:47 blaffoy I thought it might have been a bug in 2014.7, but I downgraded to 2014.1.11 and saw the same behaviour. Though I never saw this happen before upgrading yesterday.
17:52 glyf joined #salt
17:53 Andre-B joined #salt
17:54 budman where is a good GROUPing break down
17:54 rickh563 joined #salt
17:55 kickerdog joined #salt
17:56 kickerdog1 joined #salt
17:57 rickh563 joined #salt
18:01 Ryan_Lane joined #salt
18:02 aqua^mac joined #salt
18:04 transmutated joined #salt
18:04 rickh563 joined #salt
18:04 transmutated How to apply my sls file to a specific minion? I had it in a wiki at my previous employer, and I need it again:)
18:04 blaffoy Actually, I might have something. I tried downgrading the minion to 2014.1.11, and it worked again.
18:05 blaffoy This might be a bug with the 2014.7 win32 minion
18:05 jswanson_ joined #salt
18:06 workingcats budman, i noticed halite earlier, looks nice
18:07 ajolo UtahDave: hey again :)
18:07 felskrone joined #salt
18:07 budman workingcats: thanks, I believe I read a blog post/review about it being a bit janky (I could be completely wrong).
18:07 murrdoc joined #salt
18:07 workingcats budman, to clarify, i literally meant *looks* nice, havent tried it yet
18:09 aparsons joined #salt
18:09 esharpmajor joined #salt
18:09 eliasp pah, nowadays it's all about the looks… what about inner values? ;)
18:13 lz-dylan transmutated: $ salt 'nodename' state.sls statefile
18:14 diegows joined #salt
18:16 genediazjr joined #salt
18:17 lz-dylan transmutated: alternately you can select for your minion in your topfile. http://docs.saltstack.com/en/latest/ref/states/top.html
18:18 linjan joined #salt
18:18 Andre-B joined #salt
18:18 cpowell joined #salt
18:18 bhosmer_ joined #salt
18:20 blaffoy Okay. Trying a few combinations. master-2014.1.11/minion-2014.1.11 works. master-2014.7/minion-2014.1.11 works. master-2014.7/minion-2014.7 FAILS. master-2014.1.11/minion-2014.7 FAILS.
18:20 blaffoy So, it looks like something wrong with the minion, for sure.
18:20 blaffoy I wonder if anybody else could reproduce this?
18:21 blaffoy I'm not seeing any output in var/log/salt/minion
18:22 eliasp blaffoy: http://docs.saltstack.com/en/latest/ref/configuration/logging/index.html#log-level-logfile
18:22 eliasp blaffoy: you might have to set the loglevel for the logfile explicitely/seperate from the CLI
18:22 blaffoy Cheers, turning that on now
18:23 thawes joined #salt
18:23 geekatcmu joined #salt
18:23 blaffoy eliasp: would setting "log_level: all" in the salt\conf\minion not work?
18:24 dstokes hey guys. i've got a custom execution module that returns a dictionary. every time i run it i get the expected yaml output, but the exit code is 11. what am i missing?
18:24 cnelsonsic joined #salt
18:25 dstokes figure it might be this: https://github.com/saltstack/salt/blob/59398dc4e10369fd64abe8741922b03016a250c8/salt/cli/__init__.py#L226, but other modules in the salt repo return strings etc with a 0 exit code
18:26 forrest joined #salt
18:29 otter768 joined #salt
18:29 mreider figured out my Master startup issue
18:29 jonbrefe1 joined #salt
18:29 mreider no ports available ErrNo 99
18:29 hal58th joined #salt
18:29 mreider my IPv4 nic was set to DHCP so it switched an IP on me once I rebooted VM, the master config was set to a specific IP
18:29 hal58th joined #salt
18:30 rypeck joined #salt
18:30 eliasp mreider: ouch… ok, that's something which should result in a better error message… could you file an issue for this?
18:30 blaffoy 2014-12-03 18:27:44,036 [salt.state       ][ERROR   ] The following packages failed to install/update: mongodb=2.6.1
18:30 eliasp something like "IP 1.2.3.4 configured for master not found on any interface"
18:30 mreider how do i file issue, on github? sorry im 2 days old with salt
18:31 eliasp mreider: I'll do it for you… :) which version was your master on which OS?
18:31 mreider centos 6
18:31 mreider salt 2014.7.0 helium
18:31 eliasp mreider: alright … will reproduce in a container + then file an issue… will ping you when it's filed so you can subscribe if you want
18:32 mreider great thanks
18:33 jhauser joined #salt
18:33 pwf joined #salt
18:34 lz-dylan hey folks! anyone have idea for restarting salt-minion in mid-run? it's okay if I have to restart my highstate or sls run (and also okay if it picks up from where it left off).
18:34 genediazjr joined #salt
18:35 forrest lz-dylan: You can't really do that, you can use 'at' to restart it once the run finishes, but the minion needs to finish the run before restarting.
18:36 eliasp lz-dylan: see also: https://github.com/saltstack/salt/issues/6792
18:36 forrest Yep, I created that a while back ;P
18:37 blaffoy And... that seems to be all of the relevant information
18:37 eliasp yeah… and I'd really love to have a solution for this which is programmatically accessible
18:37 eliasp but it looks like we're currently stuck with an "overstate"-ish solution
18:38 eliasp have something which aggregates all reboot-requests and then (when not inhibited) reboots the minion… once it's up again, all states which had the reboot as requisite are continued
18:38 eject_ck joined #salt
18:38 dstokes anybody know how to control the exit code of a custom execution module?
18:40 lz-dylan Shoot. I can try to work with that, but it would really help to have a way to Stop Running Minion Right Now. The goal is to make sure that minion reloads itself after docker-py is installed before it hits a state dependent upon that. I got some advice from UtahDave yesterday to try using module.service.restart, and so have a module.run right in my statefile. It returns True and the execution time of the salt-minion process updates, which is *almost*
18:41 lz-dylan Thanks for the super-fast response! Makes me wish I had more of a Python background so I could hack on this. :P
18:42 lz-dylan I'll know if this is actually as not-working as it looks in like 5m; waiting on a new instance to spin up
18:45 cpowell_ joined #salt
18:45 shaggy_surfer joined #salt
18:45 druonysus joined #salt
18:45 druonysus joined #salt
18:48 patarr joined #salt
18:48 patarr joined #salt
18:50 hal58th1 joined #salt
18:51 smcquay joined #salt
18:51 blaffoy It looks like pkg.installed makes a call to pkg.list_pkgs to verify that a package installation completed correctly
18:52 eject_ck1 joined #salt
18:52 blaffoy But if I run `salt myminion pkg.list_pkgs` against a minion running 2014.7, I get a stack trace
18:52 eliasp blaffoy: might be because you have packages with non-ASCII chars in their fullnames…
18:53 eliasp blaffoy: see these commits: https://bpaste.net/show/e8a208455713
18:53 eliasp blaffoy: they fix this issue
18:54 blaffoy eliasp: I think you nailed it. "UnicodeDecodeError: 'utf8' codec can't decode byte 0x82 in position 12: invalid start byte
18:55 eliasp blaffoy: I hope 2014.7.1 is release soon which will include these fixes
18:55 blaffoy eliasp:
18:55 eliasp blaffoy: until then, distribute the fixed modules via http://docs.saltstack.com/en/latest/ref/file_server/dynamic-modules.html
18:55 blaffoy eliasp: okay, cool. I'll give that a go.
18:55 blaffoy Thanks for your help.
18:55 eliasp yw
18:56 arif-ali joined #salt
18:57 lz-dylan .....yeah, so re: prior discussion, salt-minion may report itself as reloaded & update its process execution time, but states dependent on docker-py (which gets installed in highstate) fail in that run.
18:58 eliasp mreider: I was just trying to reproduce your master IP issue… setting a non-available IP results in "[WARNING ] Unable to bind socket, error: [Errno 99] Cannot assign requested address" here, so I'd say this message is sufficient
18:59 eliasp "Cannot assign requested address" tells IMHO enough
18:59 eliasp 2014.7.0 Master on CentOS
18:59 dstokes looks like the retcode from __salt__[cmd.run]() sets the exit code of parent modules?
19:00 Ahlee That sounds right, dstokes
19:01 dstokes what if i don't care if a cmd.run fails in relation to the success of my custom module?
19:01 dstokes Ahlee: ignore_retcode?
19:01 eliasp http://docs.saltstack.com/en/latest/ref/states/all/salt.states.test.html might also be useful
19:01 Ahlee I'm not sure, I've never tried to ignore the return code
19:02 lz-dylan forrest: if I just cmd.run /etc/init.d/salt-minion restart am I likely to break something in the process?
19:02 dstokes seems weird that my module could make a call to a subshell that controls the success state of my module..
19:02 Ahlee dstokes: well, you could always make your cmd.run "my_command || true"
19:02 lz-dylan (can always just drop startup highstate in the minion config, which takes care of half the issue)
19:03 forrest lz-dylan: Usually the salt-minion process will just hang and get weird on the run, I've only done it one time, and started using 'at' to restart after runs finished.
19:03 dstokes Ahlee: yeah, that's how i confirmed it was the command that was causing the mod to fail (even though i'm not throwing) but that's an epic hack. there's got to be a way to handle this properly..
19:03 dstokes ignore_retcode doesn't seem to do anything
19:03 Ahlee dstokes: Well.  To be fair, you're asking how you can ignore the return code, which in unix world is an epic-er hack :)
19:04 Ahlee i'm checking cmd.py
19:04 lz-dylan forrest: huh. so is there no way to programmatically terminate the current state run?
19:04 Ahlee what version you on dstokes?
19:04 dstokes i'm asking how i can gracefully handle the failure of a command in a script. if [[ $? -ne "0" ]] is not a hack ;)
19:04 dstokes Ahlee: salt 2014.7.0rc2 (Helium)
19:05 forrest lz-dylan: I mean you could kill -9 or something... but not that I can think of gracefully.
19:05 dstokes my python module is funning the equiv of `set -e`..
19:06 Ahlee and you want to not fail a state run when one of the cmd.run definitions returns a non-zero exit code
19:06 RichardTheRed joined #salt
19:07 dstokes yeah. utilizing the command line to check system state, in a way that can fail, shouldn't mean my entire module fails. or maybe it should and i'm just making the wrong assumptions about module authoring
19:07 Ahlee all my _modules/ and _states assume a failure is critical and bail, but I hear ya.  I remember doing this back when I used puppet so I know there's use cases
19:08 Ryan_Lane joined #salt
19:08 dstokes if i __salt__['cmd.retcode']('grep foo bar') to check for a line in a file, do i rly want that to be considered a module failure?
19:08 dstokes (if the line doesn't exist)
19:10 lz-dylan forrest: d'ya think there's a way to shove a module.run saltutil.kill_job into a state? I actually don't know if saltutil stuff is considered modules at all.
19:11 forrest lz-dylan: Hmm, I don't THINK so, but I've never tried it. You could always right a module that calls that somehow maybe if it's not in there?
19:11 dstokes Ahlee: thx for your input. guess i'll just `;exit 0` for now :(
19:13 Ahlee dstokes: yeah< i need more time to crawl over cmdmod.py
19:13 Ahlee as, well, it takes the option, and it's checking it for log levels at least
19:13 dstokes in there right now, trying to figure what "ignore_retcode" is _supposed_ to do
19:13 dstokes yea
19:13 Ahlee my guess is it's just checking it too deep, and that should move above the logging checks
19:14 Ahlee lz-dylan: I use a state with -order: last to schedule a restart of the salt-minion (in my case a kill -9 and I depend on supervisord to restart the minion)
19:14 wendall911 joined #salt
19:15 dstokes Ahlee: i can see the retcode being returned from the private _run method, but def run then returns stdout. not sure how it's getting the retcode back to the master..
19:16 lz-dylan I'm thinking something like: stop this run if we just installed docker-py: \ module.mod_watch: \ - name: saltutil.term_job {{ some smart jinja to get our jobid }} - watch: pip: docker-py
19:17 Ahlee dstokes: the _run in this case is in your _module/ ?
19:17 dstokes Ahlee: https://github.com/saltstack/salt/blob/develop/salt/modules/cmdmod.py#L156
19:18 dstokes Ahlee: called by "run" function
19:18 Ahlee i used to use _modules extensively, but recently moved everything to _states/ to control the ret dictionary
19:18 Ahlee ah, i see
19:18 dstokes yeah. i'll have to open an issue / question
19:18 dstokes Ahlee: thx again
19:18 Ahlee good luck
19:19 lz-dylan Ahlee: that's not a bad idea! I'll keep it in mind for restarting minion after a full run. I'd like to break out of the current cun if run if I can, but we'll see if that's even a pissbossibilitiy without breaking everything...
19:19 Ahlee i have notes somewhere around here for the latest suggested method
19:19 Ahlee I don't recall what it is now adays
19:20 lz-dylan s/crazy/possibility (wifi crapped out)
19:21 lz-dylan y'know, I'm not even sure I need to restart the minion.
19:21 lz-dylan I think I _just_ need to terminate the current job.
19:21 Ababab joined #salt
19:22 Ababab Hi. How do I specify "--salt-transport=raet" when installing with pip? I'm not too familiar with Python
19:22 lz-dylan yeah. without a minion restart, the run subsequent to docker-py installation has access to newly-enabled states.
19:22 lz-dylan I just need to kill the current job.
19:24 lz-dylan Ababab: http://docs.saltstack.com/en/latest/topics/transports/raet/index.html tl;dr install its dependencies and enable in /etc/salt/master
19:25 lz-dylan Ababab: deps are libsodium, libnacl, ioflo and raet
19:25 manfred Ababab:  pip install --install-option="--salt-transport=raet" salt
19:26 Ababab lz-dylan: well, I can't compile zeromq on Cygwin, so I can't install salt without that option
19:26 lz-dylan or ask the guy who knows what he's doing :)
19:26 manfred :)
19:26 manfred it is on the install documetation page http://docs.saltstack.com/en/latest/topics/installation/
19:26 lz-dylan brilliant!
19:27 Ababab manfred: tried that, but it passes --salt-transport=raet to zeromq setup.py
19:27 manfred there is only one setup.py
19:27 Ababab http://pastebin.com/U0g3KuVa
19:27 thawes joined #salt
19:28 BigBear joined #salt
19:28 Ababab So, the dependencies are calculated before parsing the option - pip problem or Salt one?
19:31 manfred i would open an issue on the github tracker just to have it looked into to see if it is something we can fix
19:31 Ababab manfred: okay! Thanks
19:31 Ababab lz-dylan: thank you too
19:32 manfred Ababab: it could be the version of pip you have, if it is old enough?
19:32 Ababab pip 1.5.6 from /usr/lib/python2.7/site-packages (python 2.7)
19:32 Mso150 joined #salt
19:32 lz-dylan Ababab: good luck!
19:33 lz-dylan is the current job ID available in jinja?
19:34 manfred yeah, that is the newest or close to it, so it should work
19:35 TheThing joined #salt
19:35 Ababab Yup, I also upgraded setuptools just to be sure
19:36 thawes joined #salt
19:36 toscano joined #salt
19:37 lz-dylan maybe something like {{ salt['saltutil.running'] }} ... ?
19:41 Aut0Exec joined #salt
19:42 Aut0Exec hmmm
19:42 Aut0Exec what is this channel about?
19:44 robawt Aut0Exec: mostly salt the configuration management tool
19:44 robawt but sometimes we get on wild tangents about other things
19:45 slk_ joined #salt
19:46 slk_ do salt minions need to be able to resolve the hostname salt? i've got a minion configured to connect to the master via IP address, but the logs are complaining about not being able to resolve the salt hostname, and the minion isn't connecting to the master
19:46 Aut0Exec wtf is this salt configuration ?
19:47 Aut0Exec i'm lost here
19:47 robawt slk_: 'salt' is the default in the config from the package, you can change that value to anything you want or set your DNS to resolve 'salt' to something
19:47 jonbrefe1 any idea how to match a salt call depending on the file root?
19:47 jonbrefe1 something like
19:48 TheThing joined #salt
19:48 mschiff when including pillar data with "include:", how can I define subkeys in "key:"? using "global.foo" will use a key named "global.foo", and not foo as a child of global …
19:48 jonbrefe1 salt file_root_base_ test.ping
19:48 Aut0Exec left #salt
19:51 thawes joined #salt
19:51 aqua^mac joined #salt
19:57 fishdust joined #salt
19:57 kaictl joined #salt
19:58 Ababab manfred: in case you're interested, I think I found the problem.. The setup gets confused if pyzmq doesn't compile and you retry with --salt-transport=raet. Here are more details: https://github.com/saltstack/salt/issues/18685#issuecomment-65477539
19:59 thedodd joined #salt
19:59 manfred maybe
19:59 manfred thanks!
19:59 manfred thanks for making the issue :)
20:00 stoogenmeyer joined #salt
20:01 stoogenmeyer hey i want to try out salt, is there a way to have a master and a slave be the same machine?
20:01 glyf joined #salt
20:01 repl1cant master/minion can be on the same box, yes
20:02 Ababab manfred: np, happy to help! Though you already had quite many tickets and I'm piling up more :) Now off to bed and the quest continues tomorrow
20:03 eject_ck joined #salt
20:03 dude051 joined #salt
20:04 ingwaem joined #salt
20:04 iggy stoogenmeyer: in fact, I think it's fairly common for your master to also be a minion (the salt-formula kind of assumes as much)
20:05 stoogenmeyer cool, thanks
20:06 stoogenmeyer im hoping to be able to orchestrate some docker containers, the thing is i'll need some kind of a dynamic discovery system (like etcd for coreos). can pillar do this?
20:06 bVector joined #salt
20:06 iggy it might handle some parts of it (especially if you start getting into ext_pillar setups)
20:07 iggy there's also the mine
20:07 bVector whats the way to do yum groupinstall "Development Tools" in a state?
20:07 KaaK_ joined #salt
20:07 iggy which is what we use to figure out what hosts do what (along with grains for services, etc.)
20:08 ingwaem joined #salt
20:09 boredatwork joined #salt
20:15 span_ joined #salt
20:16 bfoxwell joined #salt
20:18 diegows joined #salt
20:21 geekatcmu joined #salt
20:28 shaggy_surfer joined #salt
20:28 nitti_ joined #salt
20:29 wm-bot4 joined #salt
20:30 otter768 joined #salt
20:30 lz-dylan okay. so, my mission to get this thing to install docker-py and then terminate current statefile run is not going well.
20:30 lz-dylan is there a way to have salt-cloud do a pip install on instantiation?
20:30 lz-dylan that would be just as good.
20:30 lz-dylan I've already got it installing a few packages from apt with -p
20:31 wm-bot4 joined #salt
20:31 lz-dylan (I'm using the bootstrap script that salt-cloud fires up by default)
20:33 wm-bot4 joined #salt
20:33 g3cko joined #salt
20:33 Mso150 joined #salt
20:34 chitown is pillar.get() broken inside pillars? (2014.1.13)
20:34 wm-bot4 joined #salt
20:34 iggy chitown: you can't use pillar.get inside pillars... because... you have no pillars to pillar.get
20:35 chitown https://groups.google.com/forum/#!topic/salt-users/gH7DHC0Ck88
20:35 debian112 joined #salt
20:35 chitown it looks others do it...
20:35 chitown and if the pillar i need is defined earlier in the top file, it SHOULD work
20:35 eliasp I see a traceback ending in 'psutil/_pswindows.py' on Win7, but I can't find where psutil is located in the win build… any ideas?
20:36 eliasp chitown: it's not that easy to implement this properly in a non-broken way… the devs are on it, it's a known (and often requested) issue, but not solved yet
20:36 eliasp chitown: there are workarounds
20:36 wm-bot4 joined #salt
20:36 eliasp chitown: you could use reclass or plain python or other external pillar methods to achieve what you want
20:36 chitown ugh...
20:36 chitown that is a ton of work
20:36 iggy JRY
20:38 iggy odd
20:38 iggy reading that thread, it appears a lot of people expect that to actually work
20:38 chitown ya, i would think that that pillar dict could be built progessively
20:38 chitown so, as long as you need a var AFTER its defined....
20:38 iggy whereas I wouldn't expect it to work at all (and I'm surprised that it kind of worked at some point for some people)
20:39 chitown im in the middle of moving the data layer into yaml files and then accessing with jinja that poulates pillar
20:40 mpanetta_ joined #salt
20:40 chitown but, i have a short term issue.... ofc :/
20:40 chitown either way, it would be cool if it either "worked" or if jinja complained about an undefined function/macro/...
20:41 wm-bot4 joined #salt
20:41 chitown but, thanks. i will move away from using pillar.get inside pillars :)
20:42 wm-bot4 joined #salt
20:43 budman what is the best/proper way to add a user/sshkey/sudo? any good links for me?
20:44 wm-bot4 joined #salt
20:45 eliasp http://docs.saltstack.com/en/latest/ref/states/all/salt.states.ssh_auth.html
20:45 eliasp http://docs.saltstack.com/en/latest/ref/states/all/salt.states.user.html
20:45 jonatas_oliveira joined #salt
20:45 eliasp https://github.com/saltstack-formulas/sudoers-formula
20:46 budman thanks :) wondering if there was a formula that did them all in once :)
20:46 thawes joined #salt
20:46 wm-bot4 joined #salt
20:46 eliasp well, there's no relation between ssh keys and sudo at all… so it's unlikely to have such a formula
20:46 kickerdog joined #salt
20:47 iggy some people also like the way the reverse-users-formula works, but I don't know that it supports sudo as well
20:47 karimb joined #salt
20:48 wm-bot4 joined #salt
20:49 nitti joined #salt
20:50 wm-bot4 joined #salt
20:51 desposo1 joined #salt
20:52 wm-bot4 joined #salt
20:54 wm-bot4 joined #salt
20:55 murrdoc joined #salt
20:55 wm-bot4 joined #salt
20:56 UtahDave1 joined #salt
20:58 mapu joined #salt
20:58 wm-bot4 joined #salt
20:59 cleme1mp joined #salt
21:01 kickerdog left #salt
21:01 wm-bot4 joined #salt
21:01 lz-dylan Hiya UtahDave! :)
21:02 lz-dylan Or hello to your automatically-reconnected nick
21:02 dude051 joined #salt
21:03 druonysuse joined #salt
21:03 wm-bot4 joined #salt
21:04 UtahDave :)  I'm here lz-dylan.  How's it going?
21:04 murrdoc joined #salt
21:04 rickh563 joined #salt
21:05 wm-bot4 joined #salt
21:05 Mso150 joined #salt
21:05 X86BSD joined #salt
21:05 eject_ck I want to use salt to automate updates check on Win
21:05 eject_ck salt "chi-iis-s10" cmd.run  'Import-Module PSWindowsUpdate ; Get-WUInstall -MicrosoftUpdate -IgnoreUserInput -WhatIf -Verbose' shell='powershell'
21:05 eject_ck why this not works?
21:06 eliasp eject_ck: there's a builtin windows update module in 2014.7 … tried this one yet?
21:06 lz-dylan Not bad! Just finally figuring out a way to have my instances come up with docker-py ready to rock before highstate hits. I'm settling for putting it in reactor and scheduling highstate to run a few minutes later. How're you?
21:07 UtahDave cool!
21:07 wm-bot4 joined #salt
21:08 budman joined #salt
21:09 wm-bot4 joined #salt
21:11 linjan joined #salt
21:11 wm-bot4 joined #salt
21:13 wm-bot4 joined #salt
21:13 arif-ali joined #salt
21:17 Mso150_o joined #salt
21:19 lz-dylan If I do a salt-cloud run on AWS, then snapshot the instance into an AMI and use that image in my cloud.profiles, will I get a speed boost out of not having to install all those dependencies on instantiation? And will my minions all have the same key, or will salt-cloud overwrite that?
21:20 aurynn I think it still has to test for the dependencies to be present on instantation.
21:20 aurynn but, I don't know
21:21 theologian joined #salt
21:21 lz-dylan It'd be nice to shave a couple minutes off initial boot.
21:24 asyncsrc joined #salt
21:26 Ryan_Lane joined #salt
21:29 giantlock joined #salt
21:30 iggy lz-dylan: you should test and let us know ;)
21:31 lz-dylan iggy: I may just! wanna get a current project to MVP first though.
21:31 cpowell joined #salt
21:31 asyncsrc Is there a way to pass in a string into a module like state.sls when calling a state file in order to have that string referenced as a variable within the state file? For instance, I have a server that has multiple sites on it, and I'd like to run a set of steps against a particular site on a server and would love to be able to pass in which site i'd like the steps in the state
21:31 asyncsrc file to be run against.  The only thing i can think of at the moment is to write a custom module that sort of does that instead of relying on state.sls
21:32 iggy that question was asked here as well, we came to the conclusion that shaving a little time off boot wasn't important enough to end up maintaining our own images
21:34 UtahDave asyncsrc: yeah, in your state you can put   - context:\rn  mykey: myvalue
21:34 iggy asyncsrc: state.sls takes pillar data as a command line variable (there's even an example in the docs)
21:34 UtahDave asyncsrc: then you can access that variable in jinja     {{ mykey }}
21:34 __gotcha joined #salt
21:34 asyncsrc okay thank you, UtahDave and iggy
21:34 UtahDave asyncsrc: ah, yeah.  iggy's example is what you want
21:35 iggy http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.state.html#salt.modules.state.sls
21:36 iggy asyncsrc: look at the cli examples in that section
21:36 asyncsrc iggy: thank you.  I think you were referencing this one? -- salt '*' state.highstate pillar="{foo: 'Foo!', bar: 'Bar!'}" -- it looks pretty good.  I'm not sure how I missed it
21:36 iggy yeah
21:37 iggy you could also do an env per site I guess and then just change the saltenv
21:38 wolfpackmars2 joined #salt
21:40 aqua^mac joined #salt
21:41 lz-dylan so, 2014.7 has this great new htpasswd state that manages .htpasswd files. It looks excellent. I'd love to use it!
21:42 lz-dylan But on http://docs.saltstack.com/en/latest/ref/states/all/salt.states.htpasswd.html there's examples of using both htpasswd.user_exists and webutil.user_exists
21:43 lz-dylan ...and _neither_ of them appears to work. :/
21:43 iggy make sure you have the prerequisites installed
21:43 lz-dylan I get "State 'whatever.user_exists' found in SLS whatever is not available"
21:44 iggy most modules are gated in their init function on some module/util/etc being installed
21:44 bVector whats the way to do yum groupinstall "Development Tools" in a state?
21:44 lz-dylan iggy: the docs don't list any prerequisites...
21:45 lz-dylan iggy: ...for either states.htpasswd or modules.htpasswd, at that.
21:45 iggy that module in particular requires the "htpasswd" util to be installed
21:46 hal58th1 Their definitely needs to be better documentation with that. Ran into that problem while using mysql module.
21:46 lz-dylan Shoot, and here I was hoping I could get rid of my apache states. I was using them _exclusively_ for htpasswd.
21:47 aurynn there needs to be better docs all around
21:48 iggy hal58th1: lz-dylan: aurynn: PRs welcome
21:49 aurynn iggy, working on it :)
21:49 lz-dylan duly noted. won't be my first.
21:50 shaggy_surfer joined #salt
21:50 hal58th1 I really need to figure out how to do my own documentation updates. I couldn't figure out how to link other webpages in my webpage.
21:50 iggy oddly, I have no PRs against core (but tons against -formulas)
21:50 aurynn I intend to have a bunch against docs
21:51 aurynn and one against logrotate
21:51 Syn1 joined #salt
21:51 Ryan_Lane joined #salt
21:52 iggy well, don't wait on the docs ones... I believe they hired someone to work on docs recently, so the sooner you get stuff in there, the better
21:52 hal58th1 that's good to hear iggy, the one complain i see consistently about Salt is the lack of coherent documentation
21:53 iggy which is odd, because I think the docs are pretty damn good for a project this size
21:54 aurynn Not ... really.
21:54 lz-dylan I'd agree that the docs are pretty good for reference, and the intro is better than most projects. Intermediate users are kind of in a tough place, though -- making the leap from My Little Pony Highstate to more sophisticated cloud-deploying setups with Reactor and Mine is a tough one.
21:54 aurynn there's a lot of docs
21:54 aurynn but they're not accurate, and the examples are often incomplete or not helpful
21:55 iggy that's not been my experience
21:56 iggy I will agree that they do kind of jump around a little and yeah, there's a missing niche between the very basic and... API docs
21:56 iggy but I haven't seen inaccurate docs
21:58 lz-dylan I'd say I hit inaccuracies once in a great while. I PR'd a few.
22:00 yomilk joined #salt
22:01 micah_chatt joined #salt
22:01 martoss joined #salt
22:02 martoss left #salt
22:02 aparsons_ joined #salt
22:03 jonatas_oliveira joined #salt
22:03 mpanetta_ Yo, how goes?
22:03 iggy who knows, my brain might just gloss over them... I have been reading shoddy open source docs for years now
22:03 mpanetta_ Anyone here know when the keyword arg 'merge' for grains.filter_by was added?  What version of salt that is.
22:03 micah_chatt_ joined #salt
22:03 aurynn there's been a few places I've found inaccuracies. Incomplete docs where I have to go to the code are annoying.
22:04 davet joined #salt
22:05 Ahlee damnit UtahDave.  C:\salt\salt-0.17.5-52-g2d4772c.win32 vs C:\salt\salt-0.17.5-52-g2d4772c.win-amd64
22:05 UtahDave ???
22:05 Ahlee . vs - seperating CPUARCH
22:06 Ahlee :)
22:07 budman joined #salt
22:07 UtahDave like  x86 instead of win32?
22:07 Ahlee .win32 vs -amd64
22:07 Ahlee so now I need to get clever-er in my jinja on the . vs -
22:07 Ahlee so of course I blame you (in jest, only)
22:07 iggy mpanetta: 0.17 maybe?
22:08 UtahDave :)  ah, I see.  sorry about that!  :)
22:08 mpanetta iggy: Yeah, I see that grains.filter_by was supposedly added in that version, but apparently it is in 0.16.4 as well...
22:09 mpanetta I'm trying to revive a very old salt setup...
22:09 iggy mpanetta: merge was not in 0.16.x (filter_by without merge was)
22:09 mpanetta Ah ok cool
22:09 mpanetta Thanks iggy
22:10 iggy I just clicked some stuff on github
22:11 mpanetta Well thanks for the clicks :P
22:13 kballou joined #salt
22:21 dannyn left #salt
22:21 hal58th1 I just ran into an issue where I don't think my hourly highstate was pulling in new files (according to my logs). How can I tell if the minion is solely running of caches and unable to talk to master?
22:22 Steve7314 joined #salt
22:23 hal58th1 To go along with that incident, I ran a test.ping from master and none of my minions responded in 5 seconds, including minion located on the box. I ran test.ping again and a few minions replied. Finally with a third try, all my minions responded.
22:24 StDiluted joined #salt
22:24 wolfpackmars2 i've increased the time my master waits to 60 seconds
22:24 Ahlee how would one test if a file is present on windows minions
22:25 UtahDave Ahlee: file.exists
22:25 smcquay joined #salt
22:26 Ahlee so i'm trying to remove a bogus file, if i set up a file.exists on that file, then cmd.run del /file with a -watch on the file.exists?
22:26 hal58th1 wolfpackmars2: That may help, but odd that multiple runs made the minion respond almost instantly
22:26 UtahDave or if you want to make sure it's never there, then you can use     file.absent
22:26 Ahlee guess hte logic is backwards
22:26 Ahlee file.absent just returns wether it is or isn't there, right? it will not remove the file and requires cmd.run rm to actually purge/
22:27 wolfpackmars2 I've noticed this behavior as well.  Before I increased the timeout, it seemed like I would have to run test.ping 2 or 3 times before all the minions would start responding
22:27 Ahlee nope, nevermind
22:27 UtahDave I think the file.absent state will delete the file if it is there.
22:27 Ahlee reading docs hard
22:27 Ahlee i'll go sit in the corner
22:27 hal58th1 wolfpackmars2: Ah ok, I didn't realize you had the exact same issue
22:27 wolfpackmars2 not sure why, not sure if it's a bug...  Debian minions shouldn't be "sleeping" but it may have something to do with the vps hosts idling the net connection...  really not sure
22:28 dude^2 joined #salt
22:28 Ahlee oh, i was confusing missing with absent
22:28 Ahlee thanks UtahDave
22:28 UtahDave anytime!
22:28 hal58th1 I'm all local but it is running on vmware. Not sure if that makes a difference though. I am on Ubuntu 14.04
22:28 wolfpackmars2 however, the main reason I increased the timeout was because when doing a highstate where packages need to be installed, the default of (5?) seconds didn't give enough time for the action to complete and the minion to return results
22:29 wolfpackmars2 the actions would still complete, but I would need to do a second highstate command to see that the packages were installed
22:30 hal58th1 Sure, I knew that. Now I am curious if anyone has a response to my first question.
22:30 hal58th1 "I just ran into an issue where I don't think my hourly highstate was pulling in new files (according to my logs). How can I tell if the minion is solely running of caches and unable to talk to master?"
22:30 wolfpackmars2 it would seem that the intuitive thing to do would be for the minion to acknowledge that it received the command, then acknowledge that it is processing the command, and the master should then wait until the minion responds with the results of the command.
22:30 otter768 joined #salt
22:31 iggy hal58th1: I've seen the same fwiw
22:32 iggy debian wheezy on GCE
22:32 wolfpackmars2 well, I was thinking that if you can test.ping then the highstate command should be pulling active data
22:32 wolfpackmars2 active=current
22:33 hal58th1 I have a separate environment that I just ran an extended test.ping on and it took up to 45 seconds for a minion to respond and that minion was located on the master.
22:33 wolfpackmars2 if you manually run state.highstate, do you observe the expected behavior ? (that is, updates are pushed as you would expect)
22:33 cjohn joined #salt
22:33 wolfpackmars2 is cpu usage high on the minion?
22:33 hal58th1 wolfpackmars2: that's how I noticed the issue. I ran a manual highstate and noticed it just got new files that it should have gotten days ago.
22:34 rickh563 joined #salt
22:34 hal58th1 No it isn't
22:34 dude051 joined #salt
22:35 wolfpackmars2 you're sure the hourly state.highstate is running?
22:36 hal58th1 yeah, I use a cron and pipe the logs to /var/log/salt/minion and see it running hourly.
22:37 hal58th1 Pretty nifty command actually. I do a test to see if I should run highstate.
22:37 hal58th1 if [ ! -e /etc/nosalt ]; then salt-call state.highstate -l info >> /var/log/salt/minion 2>&1; fi
22:38 wolfpackmars2 so you're running salt-call using a cronjob from the minions?
22:40 hal58th1 woops, take that back. I use a cron to make sure the salt-minion service is running. I use the salt scheduler with cmd.run and that command above. But like I said, I see the output in the logs. I want to know if there is a line in there somewhere, that basically says that it can't grab newer files.
22:42 hal58th1 or is that even possible. will highstate error if it can't talk to master?
22:43 wolfpackmars2 i've noticed that if salt times out, you may not get any response.
22:43 iggy why are you running the highstate from cmd.run? Just to be able to check /etc/nosalt?
22:43 wolfpackmars2 so if I understand correct, you have a scheduled command on the master that runs "cmd.run salt-call state.highstate" hourly on your minions
22:44 hal58th1 iggy, correct. wolfpackmars2, also correct.
22:44 UtahDave hal58th1: what logs are you looking at. Right now jobs run from the salt scheduler don't show up in the salt job cache
22:44 wolfpackmars2 so... why not just run salt '*' state.highstate using an hourly cronjob on the master?
22:45 hal58th1 UtahDave, I pipe the output. cmd.run "if [ ! -e /etc/nosalt ]; then salt-call state.highstate -l info >> /var/log/salt/minion 2>&1; fi"
22:45 wolfpackmars2 if you run salt '*' -v state.highstate then it should report any minions that timed out
22:46 hal58th1 wolfpackmars2, because I want to not run highstate on some minions. Say a developer is doing work and doesn't want his file overwritten.
22:47 shaggy_surfer joined #salt
22:48 iggy to get back to your original problem, there's not a setting that I know of that says "run from cache"
22:49 hal58th1 Ok, so there is no possibility of the minion not grabbing the latest data from master and continuing the highstate run from cache?
22:49 iggy I don't know if the default is to run from cache if the master doesn't respond fast enough
22:49 mattjb joined #salt
22:54 cjohn joined #salt
22:55 Gareth UtahDave: that feature exists in the develop branch if you use the 'job_return' option.
22:55 hal58th1 I can definitely see the master was updated three days ago, and yet the minion did not get the file until I ran a highstate manually. It may be because of timeouts somewhere
22:58 debia joined #salt
23:03 genediazjr joined #salt
23:06 wolfpackmars2 iggy what cache are you talking about?  does a minion cache master files?
23:08 wolfpackmars2 hal I setup a test minion and manually ran the "cmd.run" command you specified.  if the file /etc/nosalt exists, then basically nothing happens and nothing is logged anywhere that the command was attempted.  if I removed the /etc/nosalt file, then the master didn't show any output from the cmd.run but the minion logged the highstate results
23:09 hal58th1 wolfpackmars2: Yeap, that's what I designed it for. but that totally misses my problem
23:09 iggy yes /var/cache/salt/minion/
23:09 wolfpackmars2 i'm not familiar with salt caching so can't say whether there is some other timeout somewhere.  One thing I might recommend is adding an "else" statement to your if block where something gets put into the minion log if the command is not run
23:09 hal58th1 Is there a possibility of the minion not grabbing the latest data from Master witthout any errors being written to the logs. How can I figure out if this is happening.
23:10 iggy I would switch to -l debug and see if it isn't pulling new files for some reason
23:10 hal58th1 That's my only question
23:10 ajolo joined #salt
23:10 wolfpackmars2 e.g. "echo >> /var/log/salt/minion "cmd.run not run due to presence of /etc/nosalt file""
23:10 hal58th1 yeah iggy, I will have to do that and put some new files in salt
23:11 nethershaw joined #salt
23:14 iggy I know I've seen messages before in debug output about files being skipped because the latest was already in cache
23:15 hal58th1 Well that's what it says when successful. But I don't know what it says when it's unsuccessful.
23:15 iggy the question still remains though, what is different between your cron run jobs and when you run it by hand
23:15 TheThing_ joined #salt
23:16 hal58th1 iggy, my current theory is that I run my manual highstate a few minutes after the scheduled job. Because of that, my connection was fresh and I was able to grab the new files.
23:16 TheThin__ joined #salt
23:16 hal58th1 *I ran my
23:17 iggy so maybe throw a salt-call test.ping ; sleep 30s in your job?
23:17 iggy or 3 test.ping's
23:17 hal58th1 heh, that would be a nice hack fix if that's all it needed
23:17 iggy that seems to be the magic number for everyone
23:19 wolfpackmars2 also evaluate the output with -l debug for a period of time to see if anything stands out as to why the files may not be refreshing
23:20 eykd joined #salt
23:20 wolfpackmars2 I would expect the output to be something along the lines of "Master unreachable; using cached version instead" if there were a communication problem
23:20 __number5__ it might be the minion/master keys is out of sync with the salt master
23:23 ekristen joined #salt
23:24 eykd Is there a way to get realtime feedback during a highstate run? We’re still pretty early in our salt adoption, and waiting a few minutes to find out if a production deploy succeeded is a real nailbiter.
23:25 forrest eykd: In terms of what sort of feedback? Just seeing the status of the run?
23:25 hal58th1 eykd, currently no. forrest, I think he wants output of states in realtime.
23:26 eykd forrest: Yeah. especially if there’s a state failure.
23:26 forrest Yeah, the only way I do that is by reviewing the debug stuff, so you can do salt state.highstate -l debug
23:26 forrest and it will pump out a bunch of data.
23:26 Al_ joined #salt
23:26 eykd Sounds like fun. :)
23:26 forrest It's a lot of info, but it's usually pretty useful stuff.
23:26 eykd I’m used to a tool like Fabric, where I see the commands and their output in realtime.
23:26 iggy you shouldn't be biting nails when you push to dev/staging
23:26 forrest eykd: yeah try -l debug
23:27 eykd forrest: Cool, thanks.
23:27 forrest eykd: Yeah np, it's not perfect, but better than nothing
23:27 hal58th1 eykd, running a highstate or any function is treated as a job. job's don't give real time feedback.
23:28 eykd hal58th1: I understand if you’re pushing to 1000 servers, you wouldn’t want it. :) Right now, I’m just pushing to one. Is there any plan for introspection into a running job?
23:28 aqua^mac joined #salt
23:29 hal58th1 eykd, you can check the github saltstack issues page to see if anyone has written a request in the past.
23:29 eykd hal58th1: Good idea. Thanks.
23:30 iggy presumably, for 1000 servers you're going to be running some sort of returner that dumps stuff to a database of some sort
23:30 bhosmer joined #salt
23:34 MugginsM joined #salt
23:35 diegows joined #salt
23:35 aparsons joined #salt
23:38 snuffeluffegus joined #salt
23:49 glyf joined #salt
23:53 Singularo joined #salt
23:57 micah_chatt_ joined #salt

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