Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2018-04-05

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

All times shown according to UTC.

Time Nick Message
00:00 onslack joined #salt
00:04 tiwula joined #salt
00:35 shiranaihito joined #salt
00:36 KevinAn275773 joined #salt
00:42 __xor joined #salt
00:43 nomad_fr joined #salt
00:45 cholcombe_ joined #salt
00:45 infinity1 joined #salt
00:46 sommerfe1t joined #salt
00:46 redkrieg_ joined #salt
00:46 cmbrnt_ joined #salt
00:46 esteban__ joined #salt
00:46 pfalleno1 joined #salt
00:46 Corey_ joined #salt
00:49 exarkun_ joined #salt
00:49 weylin- joined #salt
00:49 Nazzy joined #salt
00:49 stewgoin joined #salt
00:50 k1412_ joined #salt
00:55 LeProvokateur joined #salt
00:56 nledez joined #salt
00:56 nledez joined #salt
00:56 vali joined #salt
00:58 Puckel_ joined #salt
00:58 egilh joined #salt
00:58 mritchie joined #salt
00:58 tzero joined #salt
00:59 EvaSDK joined #salt
01:06 ddg joined #salt
01:18 mechleg joined #salt
01:22 gmoro joined #salt
01:40 zerocoolback joined #salt
01:57 ilbot3 joined #salt
01:57 Topic for #salt is now Welcome to #salt! <+> Latest Versions: 2017.7.5, 2018.3.0 <+> Support: https://www.saltstack.com/support/ <+> Logs: http://irclog.perlgeek.de/salt/ <+> Paste: https://gist.github.com/ <+> See also: #salt-devel, #salt-offtopic, and https://saltstackcommunity.herokuapp.com (for slack) <+> We are volunteers and may not have immediate answers
02:04 carmony joined #salt
02:33 Nahual left #salt
02:52 hoonetorg joined #salt
03:05 _xor joined #salt
03:13 sybix joined #salt
04:03 evle1 joined #salt
04:23 tiwula joined #salt
04:39 Church- So question when trying to set a file using file.managed why would I get such an error?
04:41 Church- http://paste.ubuntu.com/p/YngNYmfZQR/
04:41 Church- Only get it on some hosts however, the sls file is this.
04:42 Church- http://paste.ubuntu.com/p/CHy5PVtWjK/
04:44 indistylo joined #salt
04:44 hemebond Iiiiiis there a root account on the machine?
04:45 hemebond Oh.
04:46 hemebond That's a bit odd. Is the sh file under the minion salt cache?
04:48 hemebond Is there any error in the master log Church-?
04:49 Church- One sec.
04:50 Church- hemebond: The cache on the minion?
04:51 hemebond /var/cache/salt/
04:51 Church- Sec
04:52 Church- hemebond: It's under some of the minions cache's.
04:53 Church- And nothing in /var/log/salt/master
04:54 hemebond Does it work the second time?
04:54 hemebond Or does it fail every time?
04:54 Church- Fail every time.
04:54 Church- Worked for some minions the first but not others, and continued as such each successive time.
04:57 cyborg-one joined #salt
04:57 hemebond Church-: https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.saltutil.html#salt.modules.saltutil.regen_keys :-D
04:57 hemebond For your earlier question/issue.
04:57 cyborg-one left #salt
04:57 hemebond What if you do saltutil.sync_all on the failing minion?
04:58 Hybrid joined #salt
04:59 Church- Sec.
04:59 Church- hemebond: Should regen_keys return None after executing on each minion?
04:59 hemebond Probably.
04:59 hemebond I've never had to regen minion keys.
05:00 hemebond Have your keys been compromised or something?
05:01 Church- Possibly, two servers got ssh bruteforced when I disabled fail2ban and keyfiles while troubleshooting an ssh issue.
05:02 Church- So regening everything, and regening new vpn tokens and reploying.
05:02 hemebond What if you do `salt minion file.copy salt://key/fix_salt_keys.sh /root/test.sh` ?
05:02 Church- That's what I'm doing next.
05:02 hemebond Ouch.
05:05 Church- Yep, now to see if this works however.
05:12 cmbrnt_ left #salt
05:13 cmbrnt joined #salt
05:18 Church- Welp seems my plan didn't work, still have to ssh in.
05:18 Church- On the plus side I can just run the script locally.
05:30 hemebond What didn't work?
05:38 DanyC joined #salt
05:40 colegatron joined #salt
05:40 colegatron left #salt
05:48 Church- hemebond: Running the script through salt '*' cmd.run 'bash ...'
05:48 Church- Got it all fixed finally however.
06:08 mauli joined #salt
06:17 eekrano joined #salt
06:19 chesty joined #salt
06:26 chowmeined joined #salt
06:45 aruns joined #salt
06:54 cyteen joined #salt
07:00 auha joined #salt
07:00 auha Hey all, I'm running into an issue when I try to use slack engine
07:01 auha It throws and error that says "argument must be an int"
07:01 auha And it seems like it is dealing with the tornado ioloop.py
07:06 aldevar joined #salt
07:18 Pjusur joined #salt
07:19 aviau joined #salt
07:19 Hybrid joined #salt
07:22 cewood joined #salt
07:29 jesusaur joined #salt
07:41 DanyC joined #salt
07:45 shoogz joined #salt
07:52 jrenner joined #salt
07:58 darioleidi joined #salt
08:08 onslack <msmith> rivyn: if you see this please @tag me
08:10 mikecmpbll joined #salt
08:11 canci joined #salt
08:11 turambar joined #salt
08:12 _KaszpiR__ joined #salt
08:12 deadpoet joined #salt
08:13 infinity_ joined #salt
08:13 lionel joined #salt
08:13 rcvu joined #salt
08:13 hop joined #salt
08:14 asoc joined #salt
08:14 hoonetorg joined #salt
08:14 hemebond joined #salt
08:15 icedev joined #salt
08:16 g3cko joined #salt
08:17 c4rc4s joined #salt
08:18 Tyrant joined #salt
08:18 emid joined #salt
08:20 Church- joined #salt
08:22 benjiale[m] joined #salt
08:22 benasse joined #salt
08:22 xet7 joined #salt
08:24 gmoro joined #salt
08:25 major joined #salt
08:28 fl3sh joined #salt
08:28 onslack joined #salt
08:29 chutzpah joined #salt
08:29 mbologna joined #salt
08:31 jerrykan[m] joined #salt
08:31 g3cko joined #salt
08:31 fujexo[m] joined #salt
08:31 v12aml joined #salt
08:32 Church- joined #salt
08:32 c4rc4s joined #salt
08:32 Tyrant joined #salt
08:32 nledez joined #salt
08:32 nledez joined #salt
08:33 systeem[m]1 joined #salt
08:34 nixjdm joined #salt
08:35 Tenyun[m] joined #salt
08:35 hemebond joined #salt
08:35 ThomasJ|m joined #salt
08:35 benjiale[m] joined #salt
08:35 rtr63gdh[m] joined #salt
08:36 Processus42 joined #salt
08:41 Valfor joined #salt
08:41 Valfor joined #salt
08:53 frew joined #salt
09:00 cmbrnt I'm using pkgrepo.managed to maintain a couplo of remote (public) repos on a bunch of Ubuntu servers, but I keep getting errors whenever the remote repo is unreachable for whatever reason.
09:00 hemebond cmbrnt: Yip. That's normal.
09:01 hemebond If Salt can't refresh the DB for any reason then the state fails.
09:04 pkkz joined #salt
09:12 cmbrnt hemebond: Allright, so to get rid of the daily error messages, I'd have to maintain the repo files "manually"? There's no way to stop requiring successful reachability?
09:13 cmbrnt I've looked over the source and can probably answer that myself. But I'm curious about what other people do.
09:13 onslack <msmith> you could convert to a cmd.run that ignores the failure, but is that any better?
09:13 onslack <msmith> you could convert to a cmd.run that ignores the failure, but is that any better?
09:13 cmbrnt I'm tempted to mirror the external repos locally. :)
09:14 cmbrnt msmith: That's an idea!
09:14 hemebond Uh, if you're using those repos to install packages there's possibly no way around it.
09:14 onslack <msmith> that would definitely work, and also give you control over whether new packages get held or not in case of an upstream issue
09:14 onslack <msmith> that would definitely work, and also give you control over whether new packages get held or not in case of an upstream issue
09:14 hemebond Am I the only one seeing duplicate messages from the Slack bridge?
09:15 cmbrnt I get duplicate messages as well.
09:15 hemebond Okay, good :-)
09:15 msmith joined #salt
09:15 * onslack <msmith> checks
09:15 * onslack <msmith> checks
09:15 msmith yes, it does indeed. heh
09:16 msmith @gtmanfred isn't happy already :D
09:16 cmbrnt hemebond: So if I run states that depend on packages from those repos then I can't assume that the repo is available?
09:16 msmith those will certainly fail if the repo isn't available, yes. the mirror idea would get around that
09:17 hemebond Well, when Salt tries to install packages it refreshes the database and if that refresh fails (because the repo is unavailable) then the pkg install states will likely fail their requisites.
09:17 hemebond What msmith said.
09:17 msmith bit more work i'd imagine, but with clear benefits
09:18 cmbrnt Yeah, absolutely. I have no issue with an install failing if the repo is unavailable - that would be a good thing. My issue is that I get errors on several state.apply each day probably because the repo is overloaded or failing over.
09:19 cmbrnt I'd like to push the repo info and only retry a connection to the repo if the repo info has changed in any way.
09:19 Cadmus joined #salt
09:20 msmith so `pkgrepo.managed` with a clever `unless`?
09:22 cmbrnt Maybe :)
09:23 cmbrnt I'm running a home environment on CentOS servers where I simply have salt push .repo files to the servers.
09:24 cmbrnt So I'll have to manually figure out if anything changes and update the .repo files centrally, but on the other hand it never fails...
09:25 lkthomas joined #salt
09:31 LevitusCommander joined #salt
09:35 LevitusCommander joined #salt
09:52 peter___ joined #salt
09:53 pkkkz joined #salt
09:55 pkkkz hi
10:07 exarkun joined #salt
10:16 FL1SK joined #salt
10:34 ProT-0-TypE joined #salt
11:01 edrocks joined #salt
11:01 mikecmpbll joined #salt
11:08 aanriot joined #salt
11:18 aanriot joined #salt
12:04 sjorge joined #salt
12:05 aanriot joined #salt
12:06 aanriot joined #salt
12:12 pf_moore joined #salt
12:12 dkehn_ joined #salt
12:15 adriano joined #salt
12:16 n45760 joined #salt
12:31 ProT-0-TypE joined #salt
12:34 indistylo joined #salt
12:36 aruns joined #salt
12:36 aruns joined #salt
12:40 deuscapturus joined #salt
12:50 onslack joined #salt
12:56 onmeac joined #salt
13:05 aviau joined #salt
13:13 Cadmus Is there a way of asking salt "Please run this command on some minions, and return a list of those with a non-zero exit status"?
13:14 cgiroua joined #salt
13:16 XenophonF Cadmus: You could always write your own returner.
13:16 XenophonF Or you could filter the job status through something like jq.
13:16 XenophonF Or you could return job status to a database and query that (I think).
13:17 XenophonF Or, quick and dirty, pipe to a file and search for "Result: False", which is what I do all the time.
13:17 Cadmus Ugh, I wish I understood salt's internals better.
13:18 babilen Take a look at the way "--summary" is implemented
13:19 Cadmus Basically I've broken something, and the existence of a file matching a glob shows if it's fixed or not. So I'm just doing a "salt '*' cmd.run 'ls /foo/bar/working*'" and want it to spit out a list of those machines that either don't respond or return non-zero.
13:19 darix joined #salt
13:19 Cadmus I'll take a look into your suggestions
13:20 babilen You could probably also use json output and filter with jq
13:20 Cadmus "--out=json --static", right?
13:20 shoogz joined #salt
13:20 babilen Start with "jq -s '.[]'" and take it from there
13:21 babilen Ah, yeah
13:21 babilen --static should take care of that
13:21 babilen noice
13:22 Cadmus cmd.run doesn't seem to return the exit code, guess I could wrap the command somehow
13:22 hammer065 joined #salt
13:22 babilen Do you need this often?
13:22 babilen But yeah .. Salt really needs better "summarisers" for various bits of information
13:25 Cadmus No, not often, but it's come up once or twice. For example I'm trying to keep all my patching up to date and the incredibly useful 'needs-restarting' is useful afterwards, I don't car *what* needs restarting, just whether a machine needs looking at. So again it's exit-code I'm concerned with
13:26 babilen Well, if you need it more than once you could define a custom grain that is "True" when that file is present on the minion
13:26 XenophonF Cadmus: if you're just looking for a file, you could use a file.file_exists or file.find
13:26 XenophonF https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.file.html#salt.modules.file.find
13:27 babilen XenophonF: You'd still have to filter the list of minion by "value"
13:27 babilen Salt is good at "map", it needs to be improve its "reduce" :)
13:27 babilen -be
13:27 hax404 joined #salt
13:28 Cadmus XenophonF: Looks like it takes globs so that will do for my immediate problem.
13:29 XenophonF you're right of course, but file.find might give him/her better results to filter on
13:34 edrocks joined #salt
13:35 Nahual joined #salt
13:37 shoogz joined #salt
13:38 Cadmus Now to figure out how to select the things I want with JQ, I must admit to hating JSON ever since first setting up a transmission server
13:40 babilen fwiw, it will probably be fast to write a custom grain that tests for the file to be present, than grokking jq
13:42 Cadmus enh, this is a one off because I did something stupid, tomorrow I will do something different (and also stupid), so learning how to parse arbitary salt output is worth doing
13:43 babilen That's why I asked if you require this more than once
13:43 babilen You sounded as if you don't require it frequently, but as if it will come up every now and then
13:49 tyx joined #salt
14:05 shoogz joined #salt
14:23 babilen Cadmus: Did you figure it out?
14:26 Cadmus Not yet, but this is more for doing cleanup towards the end of the mess I have, dropped a query into #jq and I'll do some reading when I'm not knee-dep in the doot.
14:27 babilen Hopefully there is a better way, but something like the following should work: salt '*' file.file_exists /path/to/file|jq 'to_entries[]|select(.value == true).key'
14:29 mchlumsky joined #salt
14:29 Cadmus Find returns a list, so it was 'to_entries[] | select(.value != []).key'
14:30 Cadmus In this particular example the emptly list shows healthy
14:30 TomJepp joined #salt
14:30 babilen find?
14:31 babilen Are you looking for more than one file?
14:31 Cadmus I wasn't sure when I started, and hey, knowing how to use file.find might be useful down the line.
14:31 exarkun how do I make a service restart when a package is installed or uninstalled?
14:31 Cadmus In this case I only care about '0' vs 'some'
14:32 Cadmus exarkun: listen_in might be what you want
14:34 babilen exarkun: Could you provide more details? Do you want that to happen when somebody is making local changes to the minion? If a specific state (pkg.installed for example) has changes, ...
14:35 Cadmus As always babilen has the more pertinent questions :)
14:35 mikecmpb_ joined #salt
14:35 exarkun Only if the package is installed or uninstalled as a result of a salt state.  If someone logs on and starts running random commands, they can clean up their own damn mess.
14:36 racooper joined #salt
14:36 exarkun something like https://gist.github.com/exarkun/53da46fe3b7213129a1fbcb65b4c9e00 ?
14:36 tiwula joined #salt
14:36 exarkun maybe `name: ` in the middle there
14:37 babilen Well, requisites are specific to a particular state. In your case you are referring to a "pkg.* state with the id/name of 'megacli'"
14:37 babilen There is no way to express "if any unspecified state installs megacli"
14:38 babilen For that you'd need a beacon that triggers whenever that package is installed and then triggers a reaction on the master
14:38 babilen (that would capture the 'user makes local changes' bit also)
14:39 babilen Generally speaking you want to be explicit with salt: You decide if megacli is installed and *iff* you install it you trigger the service restart via listen/listen_in requisite
14:39 babilen Keep in mind that "- pkg: 'megacli'" refers to a state and *not* to the package itself
14:40 bigjazzsound joined #salt
14:42 exarkun ok, thanks.
14:43 babilen So: Decide if you want to install megacli, if so .. trigger the service restart
14:45 exarkun And "Decide if you want to install megacli" means "define a pkg.installed state with the right arguments" and "trigger the service restart" means "define a service.running state with a listen argument to react to changes in that pkg.installed state"?
14:46 babilen Exactly
14:49 mikecmpbll joined #salt
14:50 babilen You define *exactly* what you want the state a particular system is in and then define relationships between those states
14:51 babilen megacli won't be installed *unless* you install it explicitly
14:52 babilen You can, naturally, also build a system that reacts to dynamic changes on the minion. The tools to use in that context are beacons and reactors. The former provides a stream of events to which you can then react with the latter
14:52 edrocks is `file.managed` working with binary data for anyone on 2018.3.0? It was working in 2017.x but all the workarounds i found aren't working (setting encoding to utf_8, disabling show_changes)
14:53 babilen inotify, log, something akin to the pkg beacon might come in handy to watch for "megacli" related changes
14:53 edrocks I'm getting errors like this https://github.com/saltstack/salt/issues/46648#issuecomment-378961367
14:53 babilen Hmm, something like that was reported recently
14:53 edrocks babilen my issue?
14:54 babilen https://github.com/saltstack/salt/issues/46858
14:55 babilen Workaround: show_changes: False
14:55 babilen Ah, you've done that already
14:56 babilen You might want to try the py3 packages and provide additional information in that bug report
15:02 edrocks I'm going to try downgrading to 2017.7.6 and see if that fixes it. I'm on ubuntu 14 and don't really want to get py3 installed
15:02 babilen Sure
15:02 babilen .6 ?
15:03 edrocks apparently it was released? https://docs.saltstack.com/en/2017.7/topics/releases/2017.7.6.html
15:03 edrocks I know the irc says .5
15:03 babilen Hasn't landed in my repos yet
15:03 edrocks i see
15:04 babilen .5 was released 2 days ago .. I guess some people were not impressed with the pillar behaviour changes
15:05 babilen I guess I'll also wait for people to report all the .0 bugs, before upgrading to .1 :)
15:06 babilen But then .. I *really* need the @jinja_filter functionality
15:11 edrocks lol different bugs in .5
15:15 peters-tx MTecknology, I believe I've found my problem
15:15 nixjdm joined #salt
15:16 peters-tx MTecknology, are you there?
15:16 peters-tx MTecknology, you're the only one that's responded thus far :)
15:20 babilen Assume MTecknology is here and share your findings!
15:21 peters-tx Well then
15:21 peters-tx I'm still trying to determine the cause of it, but it looks like my Salt Minions are getting a messed up PATH on my systems
15:22 peters-tx And on the Minion startup, it tries to run RPM commands but fails, saying it can't find the command (which is due to the PATH being bungled for still unknown reasons)
15:22 babilen edrocks: *sigh*
15:22 peters-tx The issue is this happens after I upgrade from 2017* to 2018*.  The Minion no longer starts
15:23 peters-tx I'm still digging into it
15:25 edrocks babilen multiple docker related issues now
15:27 babilen Yeah .. massive changes for Docker in 2018.*, so if you rely on those states you probably do want to use that version
15:27 babilen It's rather annoying that you run into encoding issues.
15:28 babilen Given that one major motivation for py3 was to unfuck encodings/strings/binary/unicode in py2, you really might want to try those packages
15:29 edrocks I'll probably upgrade to ubuntu 16 at some point. I really don't want to put too much work into it before i do that
15:32 Rumbles joined #salt
15:32 KyleG joined #salt
15:32 KyleG joined #salt
15:49 XenophonF "The following git_pillar remotes have conflicting cachedirs: b, b." LOLWUT?
15:49 * XenophonF STFWs.
15:51 XenophonF This is what I get for seeing the new release and going, "oooh shiny!!!"
15:53 XenophonF I'm going to nuke /var/cache/salt/master/gitfs and restart.  Hopefully, that fixes it.
15:54 XenophonF well, that and pillar_gitfs and git_pillar
16:00 babilen heh
16:01 dezertol joined #salt
16:01 aldevar left #salt
16:08 zerocoolback joined #salt
16:14 DanyC joined #salt
16:15 DanyC_ joined #salt
16:17 sybix joined #salt
16:17 FL1SK joined #salt
16:18 Edgan XenophonF: 2018.3 pillar gitfs bugs?
16:41 gh34 joined #salt
16:43 edrocks joined #salt
16:52 izrail joined #salt
16:58 scivola joined #salt
17:18 mchlumsky joined #salt
17:19 pcgod joined #salt
17:24 Nahual left #salt
17:30 mikecmpbll joined #salt
17:31 ymasson joined #salt
17:32 pcgod joined #salt
17:39 Gabemo joined #salt
17:44 cewood joined #salt
17:48 tzero is there a trick to pass environment variables to pkg.installed, or would it just be easier to use cmd.run + creates ?
17:49 tzero this is for a package that wants ACCEPT_EULA to be set (or via interactive prompt)
17:52 babilen tzero: environ.setenv could be used
17:55 sjorge joined #salt
17:55 MTecknology peters-tx: Nope, I'm not here, but I am now.
17:56 MTecknology peters-tx: Nope, I'm not here, but I was now. **  (I like that better)
17:57 tzero babilen: oh, there's a debconf module too; I'll try environ.setenv if that doesn't work
17:58 babilen debconf is not the same as environment variables
17:58 babilen Which package is this about?
17:58 babilen I mean .. if you can set it via debconf in lieu of using an environment variable, I'd use that
17:59 babilen You just hadn't mentioned debconf until now :)
18:04 MTecknology debconfmod... my first salt contribution. It looks almost nothing like what I originally wrote. (it doesn't suck now)
18:09 tzero ah, sorry... everyone uses ubuntu, right?! this is for mssql-tools and other MSSQL related packages :(
18:11 edrocks joined #salt
18:11 MTecknology I strongly dislike ubuntu as a server os
18:12 Edgan MTecknology: It has a few advantages, but sadly in my experience the majority of startups are already using Ubuntu.
18:12 ecdhe joined #salt
18:13 MTecknology I don't mind ubuntu as a consolation prize, but debian gets me hard and makes me feel warm inside.
18:13 MTecknology (mostly because of the alcohol consumed while developing for it)
18:14 exarkun do you realize you're in a public forum
18:14 fl3sh joined #salt
18:14 sarlalian ^
18:15 * MTecknology blinks
18:15 MTecknology if that bothers you, I'm gonna laugh and move on
18:17 sarlalian No one said it bothered them, but it certainly could bother others.  All you have to do is look at any recent code of conduct to realize that.
18:18 rivyn joined #salt
18:18 rivyn Hey guys.  Is there a way to process an SLS such that the Jinja is processed and I can see the resulting SLS that comes out that is then processed by salt?  Would be a useful debugging step.
18:19 MTecknology rivyn: you've been pointed at debug logging a few times, ya?
18:19 pcgod joined #salt
18:19 MTecknology -l debug
18:19 rivyn MTecknology: not that I recall...  Where does that flag go?
18:19 rivyn on the salt command for state.apply?
18:20 onslack <msmith> rivyn: you probably want to look at cp.get_template
18:20 sjorge joined #salt
18:21 onslack <msmith> and btw you really REALLY need to sanitise your pastes. and now, change that password
18:22 rivyn change what password?
18:23 rivyn The hashed http authorization that only works on an internal network in a development environment?  I'm not worried about it.
18:23 onslack <msmith> fair enough. i thought i'd mention it anyway as a good habit to get into
18:23 MTecknology that statement makes me weep for security :(
18:24 MTecknology "You can't use it anyway... until we oops and you can."
18:25 rivyn Sure, I'll give you that it's ideal to sanitize out, but we have much bigger security problems if you are able to hack in to the internal network anyways, frankly.
18:26 rivyn it should probably be moved to a pillar I suppose
18:40 whytewolf rivyn: lots of ways to get rendered data. cp.get_template is a good one. sometimes slsutils.renderer can be useful sometimes. my personal fav well mainly because i wrote it. https://github.com/whytewolf/salt-debug
18:42 eekrano joined #salt
18:42 whytewolf actually in 2018.3.0 slsutil.renderer looks like it got a lot more useful now that it supports cp.get_url URI's [such as salt://, http://, etc]
18:48 J0hnSteel joined #salt
18:48 rivyn I have a weird issue where a state that normally works fine isn't working on one particular server - instead of creating the directory that's not there, it's saying "Name: /var/log/postgresql/10/main - Function: file.directory - Result: Clean" and leaving the directory uncreated...
18:48 rivyn this is the state:  https://ghostbin.com/paste/zz5hp
18:49 rivyn hmm, is it because it's re-evaluating the unless condition for each one of the "names"?
18:49 rivyn doesn't explain why I didn't see this issue on other servers...maybe the directory was left there inadvertently.
18:50 whytewolf well, it would attach the unless to each of the created states
18:53 schemanic joined #salt
18:53 schemanic Hello. I need pillar to not try to load a file if it can't find one
18:54 schemanic I'm rendering the names of pillars to include in my topfile
18:54 whytewolf ignore_missing?
18:55 schemanic that's a setting I can place in my pillar top?
18:56 Laogeodritt joined #salt
18:56 whytewolf yes
18:57 schemanic whytewolf can you you please explain just a bit more? I can't find anything that explains just what that does. Is that similar to a match statement in the pillar?
18:57 schemanic I'm trying to find a usage example but I can't find one
19:00 whytewolf sorry there is documentation coming for it finally. but it was merged not to long ago so hasn't made it to the documentation
19:00 whytewolf https://github.com/saltstack/salt/pull/46833/files
19:00 deuscapturus joined #salt
19:00 whytewolf [sad for a setting that has been there since ... i want to say 2015.something
19:00 deuscapturus joined #salt
19:01 schemanic whytewolf, do I have to put it at the end?
19:01 whytewolf no
19:02 schemanic Thank you. You just saved my whole setup
19:09 cewood joined #salt
19:16 racooper Well yay, it looks like the dependency clashes with RHEL/CentOS base packages have finally been resolved. Are there any caveats upgrading from 2015.5.11 to 2018.3.0?
19:17 edrocks babilen I found a workaround. Use 2018.3, convert your binary file as base64 under a different name and then chain it with a cmd.run to decode it back to binary on the target machine
19:17 racooper or better question, is there any documentation on doing such a large version jump?
19:17 edrocks bypasses the docker issues I had on 2017.7 and gets me running on 2018.3 without any issues
19:17 deuscapturus joined #salt
19:18 deuscapturus joined #salt
19:19 deuscapturus joined #salt
19:19 Hybrid joined #salt
19:21 viq I still have no idea why thorium refuses to see that there indeed is top.sls in proper location x_x
19:24 ecdhe joined #salt
19:24 ecdhe joined #salt
19:26 deuscapturus joined #salt
19:31 MTecknology What was the page that describes how things in salt are named?
19:35 pcgod_ joined #salt
19:37 deuscapturus joined #salt
19:39 viq https://www.digitalocean.com/community/tutorials/an-introduction-to-saltstack-terminology-and-concepts was linked from main docs page
19:40 exarkun joined #salt
19:40 AvengerMoJo joined #salt
19:41 jesusaur joined #salt
19:47 MTecknology Thanks, I think that's the one I was thinking of.
20:00 dwfreed joined #salt
20:37 deuscapturus joined #salt
20:40 XenophonF Edgan: looks like the config syntax changed or something
20:40 XenophonF I need to read through the release notes to see what else broke.
20:40 XenophonF once I updated my git ext_pillar config to match the current docs, it started working
20:41 whytewolf XenophonF: were you still using the old style?
20:42 whytewolf the pre 2015.8.0 style
20:44 whytewolf cause that looks like it was what changed. dropping support for a style that was decomm 2 and a half years ago.
20:50 Edgan XenophonF: Now you have me even more curious. I use pillar gitfs, and was considering upgrading
20:51 Edgan XenophonF: Hmm, I am probably good if the docs have the right format
20:51 Trauma joined #salt
20:51 Edgan XenophonF: Even more fun when I converted to gitfs again recently was making sure all the branches were right for pillars
20:53 deuscapturus joined #salt
20:57 jeffspeff joined #salt
20:57 deuscapturus joined #salt
21:02 Edgan XenophonF: One of the things salt could be much better about is throwing errors on bad syntax, or options instead failing silently.
21:09 eekrano joined #salt
21:15 tcolvin joined #salt
21:22 tzero welp, I dug into the mssql-tools preinst script, and they clear the debconf entry before setting it to true, so cmd.run with env: is the only way
21:23 tzero love the debconf module though, works like a charm
21:46 whytewolf leave it to microsoft to mess up how to do packages right
21:47 tzero they get so close though! distributing actual native binaries, package repo (that broke/changed without warning a couple weeks ago), gpg key, but not debconf
21:48 whytewolf close only counts in horseshoes and hand gernades :P
21:49 whytewolf i will give them prompts for at least coming to the table though. it shows groth
21:49 whytewolf growth even
22:16 sjorge joined #salt
22:23 Younder joined #salt
22:24 Younder followed https://docs.saltstack.com/en/getstarted/event/beacons.html, but the example didn't work.. Is this common?
22:35 hemebond Younder: It's not unheard of. Any errors in the master log?
22:35 Church- So just curious do we have any infra CI tools like test kitchen/Kitchen-CI for Chef?
22:36 hemebond Church-: I've seen some formulas that use Kitchen for tests.
22:39 Younder hemebond, nothing in /var/log/salt/minion of interest
22:40 hemebond So there's no event sent to the master?
22:41 Younder no nothing
22:41 hemebond Have you run the minion with debug logging?
22:43 Younder hemebond, no, guess I should try that
22:49 aphor joined #salt
22:59 hemebond Younder: Anything?
22:59 deuscapturus joined #salt
23:03 pcgod joined #salt
23:04 Church- hemebond: Good to know, that's something I guess.
23:06 aphor How is Oxygen release so far?
23:28 aphor DeprecationWarning: zmq.eventloop.ioloop is deprecated in pyzmq 17. pyzmq now works with default tornado and asyncio eventloops.
23:28 aphor async.py:56
23:32 aphor this on python 3.
23:32 hemebond ?
23:32 hemebond It's just a deprecation warning.
23:32 hemebond There's probably an issue open for it.
23:33 aphor there's an error too now on 2017.7.4-5
23:33 hemebond Is 2017.7 compatible with Python 3?
23:33 aphor Experimental.
23:33 hemebond I thought Python 3 support was mostly for 2018
23:36 aphor 2017.7.4 has been working in my sandbox on Python3
23:36 mateothegreat joined #salt
23:38 aphor I think it's mainly a pyzmq update issue.
23:40 aphor [ERROR   ] Future exception was never retrieved
23:40 aphor future: <Future finished exception=TypeError("__init__() got an unexpected keyword argument 'io_loop'",)>
23:41 aphor Now I need to figure out where this appeared in the pyzmq history, and how to test for it.
23:44 aphor https://pyzmq.readthedocs.io/en/latest/changelog.html <-- pyzmq ioloop changes appeared in pyzmq 17.0.0.0
23:48 aphor I'm going to try with 2018.3.0
23:56 justanotheruser joined #salt
23:57 aphor [ERROR   ] Future exception was never retrieved
23:57 aphor future: <Future finished exception=TypeError("__init__() got an unexpected keyword argument 'io_loop'",)>
23:57 aphor ...
23:57 aphor TypeError: __init__() got an unexpected keyword argument 'io_loop'
23:58 aphor File "../tornado/iostream.py", line 1169, in __init__
23:58 aphor super(IOStream, self).__init__(*args, **kwargs)
23:59 aphor from
23:59 aphor File "../salt/transport/ipc.py", line 334, in _connect
23:59 aphor io_loop=self.io_loop

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