Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2016-01-28

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

All times shown according to UTC.

Time Nick Message
00:00 Xmonk8989 So rather than modular, this becomes more of a snapshot kind of scenario with using git?
00:01 Xmonk8989 Rather than update an ntp repo I'd have to download the whole lot of formulae, and then edit just that subfolder, and then push back with the changes.
00:01 Xmonk8989 It seems to be counter-prodctive.
00:01 Xmonk8989 (I'm new, so please tell me if i'm wrong)
00:01 MTecknology ehm... you'd have to change how you think about writing/managing states to be able to adopt what I do.
00:02 bhosmer_ joined #salt
00:03 MTecknology I built a single monolithic structure that keeps everything in an expected location. A lot of times, I'll see people struggling to figure out which sls is causing the thing they don't expect. I don't have that problem
00:04 Xmonk8989 Hmm. So, if I for example wanted to use repo1:/salt-states.git - but I wanted to use the old /ntp formulae, and the newly created /nginx formulae, even though someone has recently updated the /ntp to the new version, I couldn't mix the versions?
00:04 Xmonk8989 (it's late, that may not make sense)
00:04 MTecknology If it's a state, it's in one repo. If it's a managed file, that data is in a data repo. Almost all packages installed on systems are always unique to that system or, at minimum, the configs that pillar controls are unique
00:04 MTecknology formula*
00:04 Xmonk8989 I see. I noticed the "source" arguments to configuration files.
00:05 kevinquinnyo1 joined #salt
00:05 Xmonk8989 MTecknology: doesn't that lead to a lot of duplication of your states? For example in mixed dev, qa, live environments, and different ranges of OS's?
00:06 qu9 joined #salt
00:06 MTecknology In my states repo, I have ./sys/{packages,files,cron,users>/<name>.sls and ./top.sls. If you're fighting a state that's modifying rsyslog, you know instantly to go look at sys/files/rsyslog.sls
00:07 MTecknology In the data repo, the structure is what you'd find on a file system, ./etc/salt/minion.d/main.conf
00:08 MTecknology In my old role, I had top.sls in its own repo with an empty.sls that did nothing. Then the states repo had three branches (environments) for prod, dev, test
00:08 jack__ joined #salt
00:09 MTecknology In the incredibly rare case where systems need a different version of something, I'd just tag the file.  ./etc/samba/smb.conf%%syslog
00:09 MTecknology I think that happened three times across >500 servers
00:09 Xmonk8989 MTecknology: It begs the question "Why have the top.sls in a separate repo at all?"
00:10 MTecknology because of the environments thing I mentioned...
00:10 MTecknology otherwise salt has three versions of top.sls that it needs to merge and that's gonna get ugly even if they're all the same contents.
00:12 MTecknology At home, there's no need because I don't separate environments like that.
00:13 Xmonk8989 I see. I need to learn more about salt of course. It seems that it's a shame it doesn't support the ability to "call in" other specific repositories on the fly.
00:13 MTecknology just add them to the config and restart salt-master...
00:13 onlyanegg joined #salt
00:14 dlam hey in a pillar file,  can I reference a pillar variable above it? (so i dont have duplicate stuff)
00:14 MTecknology vim ./etc/salt/master.d/gitfs.conf; git commit -am 'Adding neat repos'; git push; <done>
00:14 djgerm Howdy! I asked last night about this, but was hoping to perhaps get more info: I am evaluating DNS formula/modules for saltstack. I'd like to ultimately spin up a couple masters, update forward and reverse… you know all the things you'd need DNS internally for. Any recommendations on where to start? Or just… install bind, manage a couple config files, use ddns?
00:15 Xmonk8989 It kills the ability of community sharing of the salt formulas though. You cannot simply "call-in" an nginx from a random github, or another ntp from a gitlab.
00:15 MTecknology why can't you?
00:15 MTecknology I mean... if you don't copy my system.
00:16 keimlink_ joined #salt
00:16 Xmonk8989 MTecknology: because from what I have understood, you need to define these in the master config daemon. I would like to simply reference it within a top.sls dynamically.
00:17 Xmonk8989 Having other repos defined this way, would lead to very messy gitfs_remote lists.
00:17 MTecknology you can't pull statets from a repo without them being defined in the master config
00:17 Xmonk8989 MTecknology: examtly.
00:17 Xmonk8989 MTecknology: Waw. There's the tiredness kicking in.
00:18 MTecknology what you're trying to do is likely going to get messy no matter how you make it happen.
00:18 MTecknology That's another nice thing about how I do things... three repos, never more.
00:18 Xmonk8989 MTecknology: It would be nice to be able to define such as: base: '*' \ - nginx \ git_repo:github.com/somerepo - branch:master
00:19 GreatSnoopy joined #salt
00:19 MTecknology within your top file?!
00:19 Xmonk8989 MTecknology:  as an example.
00:19 MTecknology my brain is bleeding...
00:22 MTecknology Xmonk8989: at a minimum, look through formulas and make sure you understand what's going on before sucking it in
00:24 Xmonk8989 MTecknology: I'm trying. :-)
00:26 ageorgop joined #salt
00:29 jack__ joined #salt
00:31 zsoftich1 joined #salt
00:32 berserk joined #salt
00:32 om <om> Hi all.  Is there something wrong with this syntax?
00:32 om <om>   - unless: {% win_system.get_domain_workgroup %} = {{ salt['pillar.get']('windows_join_domain:correct_windows_domain:domain', '') }}
00:32 om <om> I get this error
00:32 om <om>   Rendering SLS 'development:windows_join_domain' failed: Jinja syntax error: expected token 'end of print
00:32 om refreshed pillars even, doesn't work
00:34 berb joined #salt
00:36 * MTecknology sighs, shaked his head, thinks about a beer
00:37 MTecknology om: https://docs.saltstack.com/en/latest/ref/states/requisites.html#unless
00:37 ageorgop joined #salt
00:37 [dee] github is down .. weee
00:37 om wee
00:37 MTecknology [dee]: but at least the magical unicorn is fighting for us!
00:38 [dee] i like their status page though
00:38 [dee] and the unicorn haha
00:40 MTecknology indeed, it makes me not care
00:41 jack__ joined #salt
00:44 Xmonk8989 MTecknology: Thanks for your time. Gonna go sleep and read more tomorrow.
00:44 * MTecknology should go nappy too
00:45 Xmonk8989 MTecknology: nn
00:45 * MTecknology sighs .. was about to provide links :(
00:46 hal58th joined #salt
00:52 zhashuyu joined #salt
00:53 jack__ joined #salt
00:56 zhashuyu hello I'm wondering could salt-ssh work with salt-api ?
00:57 bhosmer_ joined #salt
00:59 GnuLxUsr joined #salt
01:00 gladiatr joined #salt
01:00 onlyanegg joined #salt
01:01 nethershaw joined #salt
01:07 subsignal joined #salt
01:13 tligda joined #salt
01:13 jack__ joined #salt
01:15 iggy zhashuyu: nope
01:16 RobertChen117 joined #salt
01:29 jack__ joined #salt
01:30 ageorgop joined #salt
01:36 amcorreia joined #salt
01:37 baweaver joined #salt
01:37 subsignal joined #salt
01:40 invsblduck joined #salt
01:42 invsblduck joined #salt
01:51 bhosmer_ joined #salt
01:54 rem5 joined #salt
01:56 brianfeister joined #salt
01:59 jack__ joined #salt
02:01 devtea joined #salt
02:02 catpigger joined #salt
02:05 om hi
02:05 om getting this error on salt run
02:05 om 'win_system.get_domain_workgroup' is not available.
02:05 om left #salt
02:07 om joined #salt
02:07 michelangelo joined #salt
02:07 om any idea why if that's a windows server?
02:13 om seems something is missing...
02:13 om the wins/ modules directory
02:13 om ??
02:23 jack__ joined #salt
02:25 JDiPierro joined #salt
02:30 hasues left #salt
02:33 jack__ joined #salt
02:35 brianfeister joined #salt
02:45 colegatron_origi joined #salt
02:45 bhosmer_ joined #salt
02:46 jack__ joined #salt
02:53 Bryson joined #salt
02:59 om Module for managing windows systems.
02:59 om depends:
02:59 om win32net
02:59 om how/where do I get that and install it?
03:00 zhashuyu left #salt
03:03 alexlist joined #salt
03:04 jack__ joined #salt
03:13 kk_ joined #salt
03:17 evle joined #salt
03:18 jack__ joined #salt
03:19 kawa2014 joined #salt
03:25 jaybocc2 joined #salt
03:26 om joined #salt
03:29 om Anyone have ideas why no windows modules work?
03:29 om something is missing
03:29 om 'win_system.get_computer_name' is not available.
03:30 ageorgop joined #salt
03:33 jack__ joined #salt
03:33 colegatron_origi joined #salt
03:36 JDiPierro joined #salt
03:39 bhosmer joined #salt
03:40 dimeshake joined #salt
03:47 quasiben joined #salt
03:50 beardedeagle joined #salt
03:57 Bryson joined #salt
04:04 racooper joined #salt
04:09 treaki__ joined #salt
04:25 om joined #salt
04:31 justanotheruser joined #salt
04:33 bhosmer joined #salt
04:35 jack__ joined #salt
04:39 impi joined #salt
04:44 jack__ joined #salt
04:46 jaybocc2 joined #salt
04:47 TyrfingM1olnir joined #salt
04:48 TyrfingMjolnir joined #salt
04:51 ageorgop joined #salt
05:09 kshlm joined #salt
05:14 toanctruong joined #salt
05:15 ralish joined #salt
05:16 ralish Hi there, I have a horrid feeling this is an obvious question, but I've yet to find the answer to it: is there a way to implement the reverse of the "onchanges" requisite?
05:16 ralish I.e. I want to run a state only if a different state *didn't* report any changes
05:17 jaybocc2 joined #salt
05:26 anmol joined #salt
05:28 bhosmer_ joined #salt
05:37 hasues joined #salt
05:48 ageorgop joined #salt
05:51 invsblduck joined #salt
05:53 colegatron_origi joined #salt
05:55 jack__ joined #salt
05:56 hasues left #salt
05:57 ageorgop joined #salt
05:59 calvinh joined #salt
06:07 jalbretsen joined #salt
06:07 jack__ joined #salt
06:10 brianfeister joined #salt
06:22 bhosmer_ joined #salt
06:29 dMoult joined #salt
06:31 blckbit10 joined #salt
06:38 jack__ joined #salt
06:38 invsblduck Thanks for the merge basepi.
06:51 om what's the most common issue when top.sls cannot be found/read?
06:52 om I keep getting “No Top file or external nodes data matches found.”
06:52 om even though file_roots:
06:52 om base:
06:52 om - /srv/salt/states/base
06:52 om top.sls is there
06:52 avozza joined #salt
06:52 om this used to work until I rebuilt the master server
06:56 jack__ joined #salt
07:03 ralish om: The top file belongs in the pillar? Which is defined by pillar_roots?
07:03 om no
07:03 om it's in file_roots
07:04 om for some crazy reason, it doesn't find the match for the node
07:05 ralish check the master/minion logs in /var/log/salt?
07:06 ralish Also, if you rebuilt the master, you'll want to make sure your minion is up-to-date. Version mismatches can cause such weirdness.
07:07 saffe joined #salt
07:15 om yes I checked all that
07:15 om it's a windows minion
07:15 om I'll trying reinstalling the minion
07:16 bhosmer_ joined #salt
07:17 jaybocc2 joined #salt
07:19 jack__ joined #salt
07:26 om :p
07:27 om did that, but that's obviously not the issue... Master cannot find top.sls
07:28 voidspacexyz joined #salt
07:29 av_ joined #salt
07:30 om the master connects to the minion fine
07:31 om just no top.sls found even though its there.
07:31 Rumbles joined #salt
07:33 jack__ joined #salt
07:34 favadi joined #salt
07:37 giantlock joined #salt
07:40 jhujhiti_ joined #salt
07:41 KermitTheFragger joined #salt
07:45 jack__ joined #salt
07:48 impi joined #salt
07:48 yomilk joined #salt
07:52 federicob joined #salt
07:57 voidspacexyz I am trying to understand the underlying architecture of the salt-setup, like what are the underlying components, how does the calls happed, and things like that
07:57 voidspacexyz Could you suggest some good articles or videos on the same
07:58 colttt joined #salt
07:58 brianfeister joined #salt
08:02 colttt joined #salt
08:10 bhosmer joined #salt
08:11 elsmo joined #salt
08:14 rubendv joined #salt
08:14 rubendv joined #salt
08:14 jack__ joined #salt
08:19 dimeshake_ joined #salt
08:19 permalac joined #salt
08:21 KermitTheFragger joined #salt
08:21 brianfeister joined #salt
08:21 evle joined #salt
08:23 izrail joined #salt
08:29 AndreasLutro I don't think there are any
08:36 om any windows salt people out here?  I have an issue where linux minions work fine, but windows come back with this:
08:37 om State: - no
08:37 om Name:      states
08:37 om Function:  None
08:37 om Result:    False
08:37 om Comment:   No Top file or external nodes data matches found.
08:37 om Changes:
08:37 om whoops
08:37 om sorry for the multi-pings
08:37 om the targeting is correct...  so I have no idea why the master reports this.  Possible bug??
08:38 om test.ping comes back tru
08:40 jhauser joined #salt
08:44 fredvd joined #salt
08:48 fredvd joined #salt
08:49 jack__ joined #salt
08:54 slav0nic joined #salt
08:56 keimlink joined #salt
09:02 LondonAppDev joined #salt
09:04 ruxu joined #salt
09:06 Rumbles joined #salt
09:08 jack__ joined #salt
09:18 jaybocc2 joined #salt
09:18 GreatSnoopy joined #salt
09:19 s_kunk joined #salt
09:19 s_kunk joined #salt
09:21 voidspacexyz joined #salt
09:21 justanotheruser joined #salt
09:23 jespada joined #salt
09:24 amcorreia joined #salt
09:24 CeBe joined #salt
09:27 om yep
09:27 om it's a  bug!
09:27 om version 2015-08-1 works while 2015-08-3 doe not
09:33 denys joined #salt
09:34 blckbit10 joined #salt
09:34 rotbeard joined #salt
09:35 jack__ joined #salt
09:36 avozza joined #salt
09:38 chiui joined #salt
09:42 malinoff joined #salt
09:48 avozza joined #salt
09:48 impi joined #salt
09:48 rotbeard joined #salt
09:57 blckbit10 joined #salt
09:59 bhosmer joined #salt
09:59 jack__ joined #salt
09:59 An_T_oine joined #salt
10:01 yomilk joined #salt
10:03 RandyT joined #salt
10:13 favadi joined #salt
10:13 jayprakash joined #salt
10:14 jack__ joined #salt
10:15 favadi joined #salt
10:25 impi joined #salt
10:29 kevinquinnyo1 joined #salt
10:32 jack__ joined #salt
10:47 jack__ joined #salt
10:51 giantlock joined #salt
10:53 viq joined #salt
11:04 ruxu joined #salt
11:08 An__T__oine joined #salt
11:10 avozza joined #salt
11:11 msciciel joined #salt
11:11 impi joined #salt
11:17 voidspacexyz joined #salt
11:18 Gareth joined #salt
11:18 xf10e joined #salt
11:19 jaybocc2 joined #salt
11:19 jack__ joined #salt
11:21 amcorreia joined #salt
11:23 xf10e tzag everyone
11:26 * xf10e is starting to think using pillar.get in states is broken in 2015.8 on Ubuntu...
11:31 xf10e srsly... {% if salt['pillar.get']('pillar:key', False) %} -> never gets run. {% if not salt['pillar.key']('pillar.key', False) %} -> always gets run.
11:31 blckbit10 joined #salt
11:32 xf10e copy & pasted the freakin key to `salt $MINION pillar.get 'pillar:key'` and get the value I set...
11:34 AndreasLutro xf10e: typo? pillar:key and pillar.key are not the same
11:35 felskrone joined #salt
11:39 voidspacexyz joined #salt
11:39 babilen Maybe some actual commands and output would help
11:39 xf10e oh, yeah, just a typo here. the key is 'openstack:skip_pillar_check' in both cases
11:39 xf10e babilen: If you want to... *g*
11:40 babilen Well, we can guess as long as we want ;)
11:40 babilen And in which context do you run that?
11:41 xf10e babilen: it's probably still the same problem I had yesterday. see https://gist.github.com/0xf10e/4fe9e5eec65bb5f992dd, which seems to work on FreeBSD but not on Ubuntu
11:41 * xf10e creates a new gist...
11:41 colegatron_origi joined #salt
11:42 babilen I wouldn't work on Ubuntu either ;)
11:43 jeddi joined #salt
11:43 babilen But maybe include the state that contains {% if salt['pillar.get']('pillar:key', False) %}, applicable information about your pillar and output that might shed light
11:45 kevinquinnyo1 joined #salt
11:47 bhosmer joined #salt
11:47 babilen xf10e: And are you sure that your pillar is available in all environments? Also: The whole notion of an "openstack" environment is a little, well, weird (at least to me)
11:49 xf10e babilen: since when is pillar environment specific? I think there's still an issue on this topic... maybe I missed an update, let's see
11:49 voidspacexyz I am looking to understand the internals of salt. Any leads ? Like what are the underlying components, how to they communicate and things like that
11:49 xf10e babilen: gist, btw: https://gist.github.com/0xf10e/d19f5c2a758d5669e731
11:49 voidspacexyz any help ?
11:50 jack__ joined #salt
11:50 AirOnSkin Is it possible to display IDs rather then the content of 'name' in highstate output?
11:50 babilen xf10e: Since when are they not?
11:51 babilen voidspacexyz: That is a *very* generic question and, just like the last time you asked it, nobody will have an answer where to begin I guess
11:51 agend joined #salt
11:52 harkx joined #salt
11:53 SVQTQ joined #salt
11:54 voidspacexyz Ok, trying to be specific, I am trying to understand, how is the communication between the salt master to minion actually happens.
11:54 babilen xf10e: There's little to say about that without knowing pillar_roots and *all* your pillar top.sls files (i.e. from all environments). And an openstack environment is a little weird, but then it is just semantics I guess
11:54 xf10e babilen, last year it was not or I wouldn't have sort about this: https://github.com/0xf10e/blog/blob/master/SaltStack_useful_envs_for_pillar.rst
11:55 babilen voidspacexyz: That depends on how you configured it, but the default would be 0mq
11:55 xf10e ..."sort"?.... ah, s/sort/thought/
11:56 babilen the "all thrown together" part might be due to top.sls file merging
11:56 xf10e my problem is probably related to this: https://github.com/saltstack/salt/issues/28237
11:56 saltstackbot [#28237]title: Pillar no longer obeys client set environment in 2015.8 | Here is a sample scenario:...
11:58 xf10e babilen: well, I'll test this but AFAIR I had different envs with differen topfiles and ended up with one big chunk of pillar data no matter in which saltenv I've used them
11:58 anmolb joined #salt
11:58 babilen xf10e: Well, I'm using a completely different setup that uses separate masters to model environments such as "dev" and "prod". You seem to see environments as akin to, well, likened to sets of services, which I find baffling.
11:58 xf10e and I'm using saltenv=openstack pretty much since I started this stupid "openstack-formula" thing ;)
11:59 jack__ joined #salt
11:59 babilen Why do you need an environment for that at all?
12:00 hoonetorg joined #salt
12:01 xf10e I want the openstack crap seperated from other states so no co-worker could accidently break a host but telling salt to apply a custom network configuration state or s/t like that
12:01 xf10e s/but/by/
12:02 rem5 joined #salt
12:02 ruxu joined #salt
12:02 * xf10e may need more caffein... or more sleep...
12:03 xf10e but pillar now being env sensitive might explain this.
12:04 xf10e btw in 2015.5 on Ubuntu 14.04 this worked fine for me. Had some issue yesterday, saw a fix in 2015.8, made my test-setup use the new saltstack repos and boom
12:07 gmoro joined #salt
12:08 xf10e babilen: and #23907 is also about pillars from different envs being merged together
12:10 babilen I guess https://github.com/saltstack/salt/issues/19992 is what you actually want
12:10 saltstackbot [#19992]title: Prevent untargeted states from being executed  | Related to #19991 it would make sense if there was an option to restrict which states a minion may execute. The topfile already links states to the minion, but it isn't stopping them from doing a `salt-call state.sls other.state` and running that....
12:11 babilen But then, what would prevent users from passing saltenv and a suitable pillar and thereby breaking the system again?
12:12 babilen In a way .. my recommendation would be to not perform manual state.sls runs at all, but simply highstates and to review what is being merged into your master branches
12:12 xf10e yeah, #19992 is the overall issue
12:12 evle1 joined #salt
12:13 xf10e only highstate is difficult with all the distributed parts openstack consists of
12:14 babilen orchestrate then
12:14 babilen Point being: Do not allow people you don't trust to execute random states manually.
12:16 babilen But then .. much like manual state.sls runs, you would target multiple minions with your highstate runs. The difference being that state.sls runs allow you to run states that haven't been targeted and that highstate runs would execute *all* targeted states.
12:16 babilen Don't quite understand why the distributed manner of openstack changes that
12:17 OliverMT https://gist.github.com/olivermt/991aa80d193c8330e94f ID pkgrepo.managed in SLS psql is not a dictionary
12:17 OliverMT anyone see what I am doing wrong?
12:17 xf10e babilen: restricting access to salt is a little hard when you're the only one on the team of admins who's deep enough into saltstack to write new modules/fix existing code where needed ;)
12:18 babilen So, you don't trust yourself?
12:19 babilen OliverMT: Where are your state IDs?
12:19 babilen OliverMT: Or, in other words: Is that everything in your SLS file?
12:19 OliverMT yes
12:19 xf10e if I would be the one who had to take care of the openstack crap alone I wouldn't have time to do _anything_ else.
12:19 OliverMT its an init.sls I use in top.sls
12:19 OliverMT just a folder called psql
12:19 xf10e openstack is crap, the specs get changed to match the implementation
12:20 xf10e the API doesn't work as documented, the documentation is out of date and in places just wrong
12:20 xf10e different components using the same common code behave differently
12:20 babilen OliverMT: https://www.refheap.com/114176 (or something along those lines) is what you want. Every state requires an ID
12:21 OliverMT in other packages the package name itself has been good enough
12:21 OliverMT for example this works:
12:21 OliverMT git:
12:21 OliverMT pkg.installed
12:21 xf10e when nova (controlling the hypervisors) couldn't talk to neutron (managing virtual networks) the webinterface stacktraced
12:21 OliverMT anyhow, thanks :)
12:22 jack__ joined #salt
12:22 babilen OliverMT: Yeah, in that state "git" is the idea. If you don't provide a "name" argument the ID is being used
12:22 xf10e when neutron can't talk to it's database it runs just fine - until the service get's restarted and then everything is gone b/c not written to the DB
12:22 OliverMT ah, gotcha
12:22 babilen s/idea/ID :)
12:22 babilen damn
12:25 colttt joined #salt
12:25 xf10e best to not ask me about openstack ;)
12:27 giantlock joined #salt
12:27 morissette joined #salt
12:33 jack__ joined #salt
12:33 akhter joined #salt
12:36 impi joined #salt
12:42 bhosmer joined #salt
12:44 mohae joined #salt
12:58 Hetman joined #salt
13:05 agend joined #salt
13:11 spuzirev joined #salt
13:16 futuredale joined #salt
13:20 jaybocc2 joined #salt
13:21 iceyao joined #salt
13:25 fleaz joined #salt
13:26 fleaz Hi
13:27 fleaz I have a nodegroup in my salt master config which works just fine. Now I want another nodegroup with every minion except the ones which are in the first group. How do I realize something like this?
13:30 KennethWilke joined #salt
13:32 xf10e fleaz: `salt -C 'N@groupA and not N@groupB' mod.function` I think
13:32 xf10e -C for the compound matcher and N@ for matching a nodegroup (if I remember correctly)
13:36 bhosmer_ joined #salt
13:41 fleaz xf10e: not exactly what i want. i have one group "foo" with a few hosts in my master conf. And I want a group "bar" with every host which is not in "foo".
13:41 JDiPierro joined #salt
13:41 babilen fleaz: '* and not N@foo' then
13:42 quasiben joined #salt
13:42 babilen And I think you can drop the "* and"
13:42 babilen (in certain versions of salt)
13:43 fleaz i'm running 2015.8.3 on debian 8 and the '* and not N@foo' isn't working :/
13:44 fleaz #
13:44 fleaz https://paste.linuxlounge.net/#/VqQ_syrZjypNKMRCvcH05c5n34o!-9G7sX0H1Z5m8H2Zo-YksJBQOffktSM6CjXRqpt1CaA my master conf
13:45 fleaz root@salt ~ # salt -N noncritical test.ping
13:45 fleaz No minions matched the target. No command was sent, no jid was assigned.
13:45 fleaz ERROR: No return received
13:45 babilen Define "not working" please. Is it hanging out on the couch drinking g&t? What did you try and what was the result?
13:47 babilen What does "salt -C "N@critical" test.ping" give you?
13:48 lothiraldan joined #salt
13:48 babilen And did you restart your master after making changes?
13:48 fleaz same error as above.
13:48 av_ joined #salt
13:49 subsignal joined #salt
13:49 fleaz yes, did a restart
13:49 babilen Well, you have to get to the point were you can define a single nodegroup. The definition (i.e. the list of nodes) looks alright on first glance.
13:49 treaki__ joined #salt
13:49 babilen Anything interesting in the logs if you run the master in debug mode?
13:50 fleaz i can run "salt -N "critical" test.ping" which works just fine
13:51 DammitJim joined #salt
13:51 subsigna_ joined #salt
13:52 jack__ joined #salt
13:52 AndreasLutro there isn't actually a compound matcher for node groups
13:52 babilen yeah
13:52 babilen Which is weird as they show its usage in the nodegroup definition
13:52 AndreasLutro that's odd
13:53 babilen But that is obviously an "enhanced" compound matcher :)
13:53 babilen https://docs.saltstack.com/en/develop/topics/targeting/nodegroups.html has 'G@os:Debian and N@group1'
13:54 babilen But you are right, N@ is not mentioned anywhere in https://docs.saltstack.com/en/latest/topics/targeting/compound.html
13:56 gh34 joined #salt
13:56 babilen Still don't quite see why the noncritical nodegroup definition wouldn't work.
14:01 linovia Hi there
14:02 linovia I'm currently reading Salt essentials. It mentions that I can get pillar informations from external sources
14:03 linovia in the book, it's a shell script but I can't make it work and looking at the documentation I've only seen mentions of external pillar being in python
14:03 linovia does anyone have a pointer about getting pillar informations from external sources (bash for example) ?
14:04 VSpike linovia: you can create pillar data with Python, but I'm not sure that counts as an external pillar.. i thought that was usually mysql or something like that?
14:04 xf10e linovia: "json_cmd" external pillar or s/t like that
14:05 VSpike https://docs.saltstack.com/en/latest/ref/pillar/all/salt.pillar.cmd_json.html#module-salt.pillar.cmd_json
14:05 VSpike Just what I was looking at too
14:06 linovia VSpike: thanks, looks like that's it
14:06 ianmac1966 joined #salt
14:06 linovia xf10e: thanks too :)
14:09 racooper joined #salt
14:11 xf10e linovia: you're welcome. I've played with this some time ago and added the %s parameter to pass a minion id ;)
14:14 cpowell joined #salt
14:15 rem5 joined #salt
14:17 iceyao joined #salt
14:18 numkem joined #salt
14:20 akhter joined #salt
14:21 rem5 joined #salt
14:21 Pie_Mage hf2: you around?
14:22 Pie_Mage whoops, wrong channel :P
14:23 hasues joined #salt
14:24 rm_jorge joined #salt
14:25 hasues left #salt
14:26 voidspacexyz joined #salt
14:29 permalac joined #salt
14:30 aphor joined #salt
14:30 dHod joined #salt
14:30 bhosmer_ joined #salt
14:31 dHod left #salt
14:32 aphor orion_: I think the existence/absence of a VM in the salt-cloud map is idempotent, but salt-cloud ONLY looks at the VM configuration details at creation time, so if you change stuff like setting minion grains in the map file and the VM exists, salt-cloud will do nothing.
14:37 XenophonF joined #salt
14:40 stooj_ joined #salt
14:42 Tyrm joined #salt
14:43 Tyrm joined #salt
14:44 ageorgop joined #salt
14:47 tpaul joined #salt
14:48 giantlock joined #salt
14:49 will___ joined #salt
14:50 zmalone joined #salt
14:50 colegatron_origi joined #salt
14:54 Pie_Mage joined #salt
14:58 ruxu joined #salt
14:58 ferbla joined #salt
15:01 amcorreia joined #salt
15:02 deus_ex joined #salt
15:02 Vaelatern Hi, I'm writing a module for a package manager. Is it okay for me to add optional parameters to the function declaration?
15:03 penguin_dan joined #salt
15:03 xf10e babilen: I still think there's something buggy with pillar and saltenvs but moving everything into base works.
15:04 xf10e Vaelatern: there are some minor differences between the various flavors of modules loaded as 'pkg' so I guess as long as the basic functionality is the same it should be fine
15:10 berserk joined #salt
15:11 btorch hi stupid question here .. I want to transition my current salt setup from running as root to a regular user
15:11 babilen xf10e: great
15:12 btorch I would setup a regular user for salt, change any perms on salt related files/paths and restart salt
15:12 jhujhiti joined #salt
15:12 btorch how would salt know that my own regular user is able to install something using pkg.install for example ?
15:13 buhman btorch: https://docs.saltstack.com/en/latest/topics/eauth/access_control.html#acl
15:13 andrew_v joined #salt
15:13 buhman btorch: you can add an acl for pkg.install specifically, for example
15:14 btorch cheers
15:16 jack__ joined #salt
15:17 mapu joined #salt
15:17 Vaelatern What about service managers? Can I add optional custom arguments?
15:19 JDiPierro joined #salt
15:20 jaybocc2 joined #salt
15:21 subsignal joined #salt
15:23 aphor Vaelatern: I think optional custom arguments are fine. Convention over configuration is the principle.
15:24 JDiPierr_ joined #salt
15:24 bhosmer_ joined #salt
15:28 Brew joined #salt
15:31 JDiPierro joined #salt
15:34 blckbit10 joined #salt
15:37 rem5 joined #salt
15:37 _JZ_ joined #salt
15:44 zwobot joined #salt
15:46 pdayton joined #salt
15:50 spuder joined #salt
15:51 bowhunter joined #salt
15:54 zsoftich1 joined #salt
15:55 eriko_ joined #salt
16:00 jack__ joined #salt
16:00 rem5 joined #salt
16:02 berserk_ joined #salt
16:04 JDiPierro joined #salt
16:08 jack__ joined #salt
16:10 illern joined #salt
16:10 bbradley joined #salt
16:10 t0nyhays joined #salt
16:15 DammitJim joined #salt
16:19 bhosmer_ joined #salt
16:22 fyb3r joined #salt
16:27 RandyT_ joined #salt
16:32 xf10e left #salt
16:33 dfinn joined #salt
16:35 tligda joined #salt
16:38 rem5 joined #salt
16:39 gladiatr joined #salt
16:40 asd joined #salt
16:42 dfinn joined #salt
16:43 djgerm Is there a Gary Larizza of the salt stack world that I can follow and learn all the best secrets from?
16:44 jack__ joined #salt
16:44 * jfindlay knows all the secrets
16:45 jfindlay maybe
16:45 jfindlay actually not
16:45 djgerm heh where can I sign up?!
16:46 AlberTUX1 joined #salt
16:46 jfindlay djgerm: you're welcome to ask me here :-)
16:46 aphor Salt has secrets?
16:46 zmalone joined #salt
16:46 djgerm I am not sure how the pro's use salt-stack to best effect regarding making DNS servers and updating DNS "dynamically" with saltstack.
16:47 djgerm I found a formula that seems more difficult than I need, I found some ddns modules that look promising…
16:47 djgerm so many options
16:47 aphor djgerm: I could be incentivized to making a BIND state that uses RDNC or even a proxy-minion.
16:48 jfindlay I haven't done that much with DNS and salt, although I know there are people here who have such experience
16:48 teryx510 joined #salt
16:48 sjmh joined #salt
16:48 aphor A domain with SOA is autonomous when the root servers start delegating to it.
16:49 aphor Everything in a domain would, IMHO be grains, mines.
16:50 mapu joined #salt
16:51 djgerm The problem's not particularly complex, when approached via old-fashioned ways… (install bind, config the zones, keys, start populating records)… but what's the saltstack way I guess is my question
16:51 aphor Each zone could be read in, and then dumped out in a dictionary format.
16:51 iggy figure out how you'd do it without salt, and then use salt to make it easier
16:51 aphor djgerm: how much control do you want/need?
16:51 iggy there's not magical saltstack ways of doing everything
16:52 djgerm it'll be internal DNS that I'll want salt to update everytime a new minion is created
16:52 aphor DNS is one of the seminal mysteries of the Internet. It's actually pretty complicated and awesomely powerful.
16:52 djgerm A and PTR
16:52 aphor djgerm: no AAAA ?
16:52 djgerm well sure DNS globally is. but running a couple rinky dink name servers internally isn't
16:52 aphor djgerm: no CNAME?
16:53 iggy You can use the salt mine to help pull some of the data for the configs
16:53 jack__ joined #salt
16:53 iggy I doubt there's any "formulas" out there that do it
16:53 iggy I only know of maybe 2 formulas that use mine
16:53 aphor +1 what Iggy said: use mine when you can't tolerate a DNS zone TTL.
16:54 aphor Do you want minions to be able to assert their own reverse (PTR)?
16:54 djgerm yes.
16:54 djgerm and yeah, I want saltstack to control all RRs, but specifically I want minions to be added to forward and reverse on creation
16:55 fyb3r >_> Just found that my MoM was running with the default open files limit.... no wonder I was getting so many timeouts when 7000 minions were coming back with returns
16:55 aphor djgerm: then mine is your friend. Maybe you can set a DNS hostname grain and then mine it?
16:55 iggy fyb3r: only 7000... weak
16:55 aphor djgerm: maybe you want to do orchestration and push new records to a DNS server in pillars via events?
16:56 fyb3r 7000 syndics
16:56 fyb3r ;)
16:56 pmcg joined #salt
16:56 djgerm ah yes. a little more complicated than ddns, but it wouldn't have the infrastructure overhead…..
16:57 gmoro joined #salt
16:57 aphor I would say the event+pillar-->reactor-->orchestration-->zone update state is the Salt way of doing DDNS.
16:57 djgerm heh not just using the ddns states? http://docs.saltstack.cn/en/latest/ref/states/all/salt.states.ddns.html
16:58 aphor You could even hook other DDNS systems to generate events that kick off the orchestration.
16:58 berserk joined #salt
16:58 aphor .. as an integration/migration path.
16:58 djgerm ha. well I want it simple and straight forward.
16:59 djgerm I think iggy is right: doing it the way I would normally, then use salt-stack to improve those steps
16:59 aphor djgerm: I don't know your normal flow to offer a relative opinion.
16:59 iggy I would find that simple and straightforward... other people might find setting up entire DDNS infrastructure simple and straightforward
17:00 iggy depends on what you know and what you want
17:00 aphor The default is to use dnsmasq, which will make you bald eventually.
17:01 djgerm true story. Salt stack is my biggest hurdle :) the system side is easy
17:01 eykd joined #salt
17:01 aphor easy of you don't have to support it for the long run.
17:01 carson-anderson joined #salt
17:01 aphor of-->if
17:01 iggy the first 3 salt deployments I did were absolute shit and have since been completely redone
17:02 carson-anderson Hey guys, I've hit an issue running salt-minion and salt-call that i've never seen before:
17:02 carson-anderson [DEBUG   ] Reading configuration from /etc/salt/minion [DEBUG   ] Using cached minion ID from /etc/salt/minion_id: ip-10-20-1-98.ec2.internal [INFO    ] Processing `log_handlers.mongo` [ERROR   ] 'log4mongo_handler' Traceback (most recent call last):   File "/usr/lib/python2.7/dist-packages/salt/utils/parsers.py", line 181, in parse_args     mixin_after_parsed_func(self)   File "/usr/lib/python2.7/dist-packages/salt/utils/pars
17:02 fyb3r jfindlay, I was curious how I would go about performing the actions you asked of me in issue 29616?
17:03 jack__ joined #salt
17:04 teryx510 joined #salt
17:04 onlyanegg joined #salt
17:04 carson-anderson Sorry about that ugly paste. o~O  Pastebin for my weird traceback here: http://pastebin.com/xyZ4CSvT
17:05 carson-anderson I was hoping that someone here could help me out. I've been debugging it for a while to no avail
17:05 iggy carson-anderson: disable your mongo logging and try again?
17:05 carson-anderson That's the odd part. I have not even enabled mongo logging.
17:05 akhter_1 joined #salt
17:05 iggy salt thinks you did
17:06 iggy carson-anderson: do you have any logging enabled?
17:06 jaybocc2 joined #salt
17:06 iggy (aside from the default)
17:06 carson-anderson Nope
17:07 AndreasLutro carson-anderson: looks like you need to add log4mongo_handler to your minion config
17:08 carson-anderson Won't that make it log to mongo? This is a very new salt setup and I don't need any logging yet.
17:09 berserk joined #salt
17:09 mpanetta joined #salt
17:09 carson-anderson this server is running mongo. But it has nothing to do with salt.
17:09 AndreasLutro hmm, maybe salt automatically tried loading that handler if mongodb is installed
17:09 iggy are there any log4mongo modules installed on the system?
17:10 iggy (i.e. for other software on the box)
17:10 AndreasLutro though that seems kinda odd
17:10 jfindlay fyb3r: I think there is enough info there to fix the problem, but if you think the documentation could be improved, then that is always welcome
17:10 anmolb joined #salt
17:11 carson-anderson iggy: yes. There is the salt log4mongo_mod.py and a standard log4mongo install under /usr/local/lib/python2.7/dist-packages/log4mongo
17:11 AndreasLutro I think you can ignore the error, and create an issue on github about it
17:12 iggy carson-anderson: can you test on a system without log4mongo?
17:12 AndreasLutro actually it might be fixed already
17:12 iggy if that works, then yeah, salt is erroneously finding the module and assuming you want to use it
17:13 bhosmer_ joined #salt
17:14 pirulo joined #salt
17:14 carson-anderson iggy: Yes. I've installed salt-minion on over 200 other servers and this is the first time I've hit this particular bug.
17:14 iggy carson-anderson: do the others have log4mongo installed?
17:15 carson-anderson Not that I'm aware of. Let me check
17:15 iggy so yeah, might be what I said
17:15 pirulo Does anyone can help me setting a test top.sls and it's not working. I'm getting this error: Rendering SLS 'base:test' failed: Illegal tab character; line 4
17:15 AndreasLutro https://github.com/saltstack/salt/blob/develop/salt/log/handlers/log4mongo_mod.py#L71-L73
17:15 AndreasLutro someone here thought yielding early is the same as returning early :p
17:16 carson-anderson That would explain it :)
17:17 carson-anderson Should I submit a PR or just post an Issue for this?
17:17 teryx510 joined #salt
17:17 carson-anderson I'm comfortable with either
17:17 iggy it should likely just return False in the __virtual__ function if there is no config
17:17 iggy open an issue
17:18 AndreasLutro __opts__ might not be populated when __virtual__ is called I think
17:18 carson-anderson Changing the yield to a return stopped the traceback for me
17:19 AndreasLutro you normally can't just change a yield to a return though
17:19 AndreasLutro it makes the function behave entirely different
17:19 AndreasLutro though there  could be code in place to handle that case already
17:19 iggy in any case, file an issue and see what the devs think
17:19 jack__ joined #salt
17:20 jrgochan Hey all. Bit of a silly question, but is there a preferred method of setting system-wide environment variables with salt? I was just going to copy files to /etc/profile.d/
17:21 jfindlay iggy: you mean what jfindlay thinks? :-)
17:22 * babilen chuckles
17:22 jrgochan either that or file.append /etc/profile?
17:22 sjmh jrgochan - /etc/default/salt
17:23 carson-anderson ok thanks guys.
17:23 sjmh is read by the init scripts ( at least in RHEL )
17:23 bhosmer_ joined #salt
17:23 sjmh it's where we put http_proxy's for our salt stuff
17:24 sjmh AndreasLutro - I didn't even think you could have a return in the same function as a yield.
17:24 bluenemo joined #salt
17:25 writtenoff joined #salt
17:25 jfindlay jrgochan: my understanding is that that is platform-specific, so perhaps that is case for an execution module? :)
17:26 sjmh jrgochan - oh, setting them via salt.  sorry.  misread that!
17:27 sjmh thought you meant setting them FOR salt. :)
17:27 jfindlay s/that is case for/that has a case for becoming/
17:28 AndreasLutro sjmh: oddly enough it works in py3, but not py2
17:28 sjmh AndreasLutro - yeah, just tested it out.  weird.
17:31 sjmh looks like they both return a generator though.
17:31 sjmh er, sorry, rephrase - looks like if you have both in py3, it returns a generator still.
17:31 jack__ joined #salt
17:32 sjmh AndreasLutro - http://stackoverflow.com/questions/16780002/return-in-generator-together-with-yield-in-python-3-3
17:34 aphor I vageuly recall that it's OK for a generator to return, but only certain values because it causes StopIteration exception or something like that.. for when you need a circuit breaker I think.
17:36 aphor yeah.. that's what the top answer says, re: PEP 380 http://www.python.org/dev/peps/pep-0380/
17:36 invsblduck joined #salt
17:39 dlam joined #salt
17:47 chiui joined #salt
17:48 jack__ joined #salt
17:48 perfectsine joined #salt
17:50 gladiatr joined #salt
17:51 dlam hmm ima system service noob:  what would you use..  supervisord or upstart or systemd or etc?
17:52 tehsu does {% if grains['abc123'] is defined %} work? I know it works for if pillar is defined works
17:54 Fabbe joined #salt
17:55 impi joined #salt
17:56 iggy jfindlay: I used to try to help triage shit, but I got yelled at for not handholding some jackhole through using python on windows (not related to salt at all), so f that
17:56 iggy and yes, I feel for you
17:58 jfindlay iggy: yes, and hopefully salt is a more positive place to be :)
17:59 perfectsine joined #salt
17:59 andrew_v_ joined #salt
17:59 ageorgop joined #salt
18:00 cyborg-one joined #salt
18:04 hal58th joined #salt
18:07 giantlock joined #salt
18:08 teryx510 joined #salt
18:13 freelock_ joined #salt
18:13 Fiber^ joined #salt
18:17 onlyanegg joined #salt
18:18 JDiPierro joined #salt
18:22 jaybocc2 joined #salt
18:24 heaje joined #salt
18:25 akhter joined #salt
18:26 pdayton1 joined #salt
18:28 jrgochan thanks for the input on the env variable stuff guys!
18:30 jack__ joined #salt
18:33 jfindlay jrgochan: what did you end up doing?
18:34 nidr0x joined #salt
18:34 jrgochan I ended up talking to a coworker then I looked at some backups
18:34 jrgochan :)
18:34 jrgochan I'll hopefully start on some env stuff soon though!
18:34 jfindlay cool
18:35 jfindlay I know what you mean though, some days I cycle back to everything, most days I don't
18:35 jfindlay I'm unsustainable :(
18:36 jrgochan as am i. I should be programming some groovy, but just not feeling it
18:36 jrgochan perhaps this will help: https://www.youtube.com/watch?v=G-xaL4_cep4
18:37 jrgochan high five for unsustainable work patterns!
18:37 jfindlay nice
18:40 rem5 joined #salt
18:41 sjmh it's why I like WFH so much.. yeah, the first weeks are tough not to slack, but i find that when I'm the office - so. many. disruptions.
18:42 sjmh once you're past the novelty stage, i find i work much more than i ever did and get more done.
18:42 jfindlay also more efficiently
18:43 jrgochan same. I miss my WFH days :/
18:44 jack__ joined #salt
18:44 aw110f joined #salt
18:46 hal58th joined #salt
18:47 sjmh unfortunately, we're going the wrong way.  they're starting to crack down on remote employees here.
18:47 sjmh want to bring everyone back into the office.
18:49 CheKoLyN joined #salt
18:50 Bryson joined #salt
18:50 onlyanegg joined #salt
18:54 anmolb joined #salt
18:55 mapu_ joined #salt
18:57 jrgochan Tell them you've got a medical condition that requires you to work from a beach in Hawaii
18:58 berserk joined #salt
18:59 baweaver joined #salt
19:00 jack__ joined #salt
19:01 denys joined #salt
19:03 onlyanegg joined #salt
19:09 toanctruong joined #salt
19:11 jack__ joined #salt
19:13 onlyanegg joined #salt
19:13 toanctruong Hey guys, I just setup a saltmaster with gitfs connected to github. salt '*' state.highstate works just fine, but setting saltenv=$anybranch yields a no top file found error
19:14 toanctruong This is the sample repo i've connected to https://github.com/toanctruong/saltsample.git
19:19 cberndt joined #salt
19:21 Lionel_Debroux joined #salt
19:24 ajw0100 joined #salt
19:25 jack__ joined #salt
19:25 babilen toanctruong: Do you have a branch for the "anybranch" environment in your git repository?
19:29 jfindlay toanctruong: also you need to make sure your file_roots, and gitfs_remotes configs, and the top file(s) in question are configured correctly
19:30 mpanetta joined #salt
19:30 toanctruong babilen: I've created a dev branch and qa branch
19:31 toanctruong in my git repo
19:31 alias_ joined #salt
19:31 babilen This is an aspect I really don't like about the way branches are translated into environments ..
19:32 babilen toanctruong: So, what is your top.sls then?
19:33 toanctruong https://github.com/toanctruong/saltsample/blob/master/srv/salt/base/top.sls
19:34 toanctruong the only difference between base and dev, is a single comment in the top.sls. Debug on the minion seems actually displays the contents of the dev branch top.sls
19:34 toanctruong but still says there is no top file
19:34 babilen So you don't have a single top.sls ?
19:35 toanctruong can you clarify on what the single top.sls file is? I guess I assumed that's what mine was
19:35 jhauser joined #salt
19:35 absolute joined #salt
19:35 babilen You seem to have one top.sls in each environment
19:36 babilen top files are being merged. That is why, if you want to still be able to git merge between branches, you would normally keep your top files in a separate repository
19:36 akhter joined #salt
19:37 toanctruong hmm..
19:37 toanctruong that's kind of a pita then
19:37 babilen Oh yeah
19:37 toanctruong so I would have a secondary repo, with just a single top.sls?
19:37 toanctruong then the file roots would point to a different repo with my states?
19:37 babilen Quite the pain in the arse and one of the reasons why I use separate masters for different environments
19:38 babilen Yes to your first question, not quite sure about the second
19:38 toanctruong I see
19:39 absolute was hoping someone could point me in the right direction. had old pre-merged salt-api on my saltmaster, updated saltmaster to Lithium, but seems that the upgrade left the old pip installed salt-api version 0.8.something. any ideas on how to clear out the old pre-merged version and set the Lithium version of salt-api as default?
19:39 toanctruong Is it a management nightmare having separate masters for different environments?
19:40 toanctruong maybe i should just stick with my original gameplan of going masterless. but then i have no mass control
19:40 babilen Not really ...
19:40 frew we have a bunch of MIBS files we want to deploy with salt; is there a way I can basically clone a directory?  there are enough files that hardcoding a list and iterating over it seems like a bummer
19:40 babilen But then, top file handling changed in 2015.8 -- If you read https://docs.saltstack.com/en/latest/ref/states/top.html and check default_top and the like
19:40 jack__ joined #salt
19:41 sjmh toanctruong - why do you have 'base' declared in the top.sls for the dev environment?
19:41 babilen The basic problem with GitFS was that you wanted the bits of top.sls that pertain to that environment on that branch, but that would prevent you from merging easily between branches.
19:41 frew I was hoping I could do like, file.directory: - source: salt://mibs/ or something
19:42 babilen And you should also be aware that the top.sls essentially defined which state from which environment are being targeted to which minion.
19:42 babilen It doesn't really make sense to have base in all of them ;)
19:42 sjmh ^
19:42 babilen Ah .. ta sjmh
19:42 babilen exactly
19:43 toanctruong Ok, so my dev environment top.sls should look like 'dev' declared rather than base
19:43 babilen yeah
19:43 toanctruong ok I was just about it write something else out
19:43 toanctruong You answered it already :)
19:43 babilen base (or anything targeted to *) should contain the states that are common/identical for all environments
19:43 toanctruong it's the secondary repo
19:43 sjmh yeah.  basically, you told 'hey, go look in my dev environment for a top.sls', it finds your top.sls and says 'nothing for dev in here'
19:43 toanctruong that solves the merging problem of branches into master
19:43 toanctruong that's where my confusion is I guess
19:44 sjmh environments are really about the fileserver.  not your logical structure.  it's wacky.
19:44 sjmh and often confusing.
19:45 babilen yea, and the whole branch == environment thing is what made me use different masters with an explicit single branch
19:45 sjmh but yes, a problem with git is general is that there isn't a good way to keep files different when merging.
19:45 sjmh you can't say 'hey, merge everything, except this'
19:45 babilen indeed
19:45 ageorgop joined #salt
19:46 toanctruong ok, I think that's cleared up then. So one method is to use another repo, or separate masters
19:46 babilen Which makes it impossible to have different top.sls files in branches .. which is why you'd keep them in a separate repository
19:46 toanctruong Any other cool ideas?
19:46 babilen Send everyone in the channel a pint?
19:46 sjmh you could try and mess w/ submodules to do it
19:46 sjmh but.. yuck. :)
19:47 toanctruong It's not even noon yet here, but a pint sounds good after wrangling with this.
19:47 amcorreia joined #salt
19:47 stooj joined #salt
19:47 babilen haha
19:47 jfindlay yeah, the top file and saltenv layout is unnecessarily complex
19:48 MindDrive joined #salt
19:48 sjmh toanctruong - if you want a single master for your different environments, then i'd go separate repos for each environment.  you can still PR between them for merges.
19:48 sjmh jfindlay - agreed. :) especially when you start throwing multi-tenant into it..
19:48 jfindlay or separate out the top files into its own repo
19:48 MindDrive joined #salt
19:49 sjmh if there was one thing i wish was better, it'd be the top files.
19:49 babilen I have a master for dev and a master for prod with different minions. They reference different branches in their GitFS configuration and top files are in a separate repository.
19:49 jfindlay there's also the master tops plugin layer, which allows you to get top files from external systems, but I've never used it
19:49 babilen That way I get around the "I need all environments in all GitFS repositories" problem which gets old for formulas soon
19:49 babilen jfindlay: huh? do you have a link for that?
19:50 GreatSnoopy joined #salt
19:50 babilen And there isn't a day in which you don't stumble over something new :)
19:50 sjmh https://docs.saltstack.com/en/latest/ref/tops/all/salt.tops.cobbler.html#module-salt.tops.cobbler
19:50 jfindlay babilen: https://docs.saltstack.com/en/latest/topics/master_tops/index.html
19:50 sjmh is one of them
19:50 babilen right
19:51 berserk joined #salt
19:51 toanctruong So the separate repo for each environment method, you'd direct the masters to the environments through gitfs_base value and base value under gitfs_remotes?
19:51 jfindlay frew: does file.recurse do what you need?
19:52 sjmh well, better tops system and pillar targeting pillar would I guess be #1 or #2 in my 'pleeeeeaaaassee' book
19:52 jfindlay also in some circumstances, it's more efficient to use rsync
19:52 jfindlay sjmh: I totally agree
19:52 frew jfindlay: I'll look
19:53 babilen toanctruong: That's what I do .. but read up on default_top and the changes in that regard. It might be feasible and easy to get away with a single master now .. *iff* you keep top.sls in a separate repository
19:54 babilen Let me know if that works out as I'm not too keen on keeping this ... still not entirely sure how to not have to create a branch for every environment in the formula repositories I use then ..
19:55 babilen sjmh: "pillar targeting pillar" ?
19:55 sjmh babilen - being able to target pillar data in pillar top files based on other pillar.
19:56 babilen sjmh: Ah .. yes, absolutely: https://github.com/saltstack/salt/issues/23910
19:56 saltstackbot [#23910]title: Please implement static pillars | Hi,...
19:57 babilen The problem is that you need "terminal symbols" in that context as you would otherwise have to recursively compile pillars until nothing can be expanded anymore.
19:58 babilen My idea was to implement "static" pillars that are available everywhere and can be used much like grains
19:58 sjmh yeah, recursive wouldn't be so bad either and i know you can do it with external pillars first, then use that data to target other pillars
19:58 sjmh but would be nice to see the file-based system have something like it
19:59 sjmh but yeah, for us, it's alot around trying to manage permissions, since using grains for roles/acl is insecure.
20:00 babilen toanctruong: I don't really see a question in your mail to salt-users
20:00 babilen sjmh: Sure, a minion can claim to have any role it wants :)
20:01 babilen "Do not trust your minions!"
20:02 jack__ joined #salt
20:02 baweaver joined #salt
20:04 toanctruong babilen: thanks, i'll mess with separate repos and see if that works out for me. I can take down that question, everything got answered here. Really appreciate it!
20:06 rem5 joined #salt
20:07 babilen toanctruong: good luck :)
20:08 babilen toanctruong: But please read up on env_order, top_file_merging_strategy and default_top in https://docs.saltstack.com/en/latest/ref/states/top.html
20:08 beardedeagle joined #salt
20:09 babilen Given that you are, I am sure, an "astute" reader ;)
20:10 toanctruong Got it :)
20:13 babilen toanctruong: If you have top_file_merging_strategy set to "same" you no longer need separate repositories
20:13 babilen (at least that's my understading .. I haven't played with it)
20:14 oida joined #salt
20:14 sjmh looks like it - but you'd still have to make sure you only define 'dev' in your dev top.sls file, not 'base'.
20:14 toanctruong I see. So a single top.sls file, referencing to all the branches in the same repo?
20:14 toanctruong I'll play around with it
20:16 babilen sjmh: Hmm .. that would still require you to use a separate repository for top files
20:16 jfindlay toanctruong: you also need to be careful about the top file showing up in different branches of the same repo
20:18 quasiben joined #salt
20:18 sjmh babilen - true.  maybe i'm reading the 'same' merging strat wrong, but seems it's aimed as only processing the top file that's in that environment's directory though.
20:18 sjmh ie - with merge, it would merge /srv/salt/top.sls and /srv/salt/dev/top.sls
20:18 sjmh with 'same' - it would only read /srv/salt/dev/top.sls
20:19 jfindlay the problem is that you also have to do saltenv to directory mappings in file_roots, outside of the top file
20:19 babilen sjmh: I don't know .. my gut feeling is that this new feature in .8 should alleviate our pain, while I see where you are getting at ;)
20:19 babilen I guess there is nothing but trying
20:19 babilen jfindlay: yes!
20:19 sjmh babilen - which new feature?
20:19 babilen default_top et al.
20:20 jfindlay sjmh: https://docs.saltstack.com/en/latest/ref/states/top.html#how-top-files-are-compiled
20:20 babilen The bottom of which is new
20:21 jfindlay cachedout worked on that madness before 2015.8 came out
20:21 jfindlay and I had to test it :-)
20:22 sjmh yeah - personally I don't like that the you can define any environment in any top file.
20:22 babilen + pass it on the command line
20:23 sjmh i feel like if you define an environment, then the top file in that environment should only be able to define data for that environment
20:23 jaybocc2 joined #salt
20:23 rpm-for-amzn joined #salt
20:23 sjmh the /srv/salt/dev/top.sls shouldn't be able to redefine the prod env
20:24 sjmh brb, have to go grab food before meeting @ 1.
20:24 chiui joined #salt
20:24 rpm-for-amzn Anyone else had issues installing RHEL rpm on Amazon Linux?
20:25 jfindlay rpm-for-amzn: what problems?  amazn linux is based on rhel 6, I think
20:25 rpm-for-amzn thats the bitchy part
20:25 rem5 joined #salt
20:25 rpm-for-amzn it is and it isn't
20:26 jfindlay hmm
20:26 jfindlay silly Amazon
20:26 rpm-for-amzn its mostly is RHEL6
20:26 jack__ joined #salt
20:27 rpm-for-amzn so the latest-1 version happily accepted RHEL rpm's, latest one does not
20:28 federicob joined #salt
20:28 radhac joined #salt
20:28 rpm-for-amzn one issue appears to be that AmznLinux has python 2.7 installed as default
20:29 bhosmer_ I want to see about implementing a streaming callback for the file client: https://github.com/saltstack/salt/issues/28641 it's been a while since I've been in the codebase though. Can someone point me in the right direction? is it in the file.py module get_managed?
20:29 saltstackbot [#28641]title: Problem doing file.managed with a http source | OS: https://vagrantcloud.com/boxcutter/boxes/oel67/versions/3.0.3...
20:30 federicob joined #salt
20:32 ajw0100 joined #salt
20:37 danlsgiga joined #salt
20:39 danlsgiga hey folks... I'm spinning new instances in aws ec2 using salt-cloud... everything is working perfect I just want to know how to persist the instance hostname instead of using ip-172-30-1-1 as hostname
20:40 justanotheruser joined #salt
20:40 keimlink joined #salt
20:40 jhauser joined #salt
20:42 jhauser joined #salt
20:43 fredvd joined #salt
20:43 jhauser joined #salt
20:46 bhosmer_ danlsgiga: It's in the /etc/cloud/cloud.cfg
20:46 rpm-for-amzn amazon linux install is discussed https://github.com/saltstack/salt/issues/30227
20:46 saltstackbot [#30227]title: Latest Salt will not install on Amazon Linux due to missing python, systemd packages and/or EPEL package out of date | Hi,...
20:46 bhosmer_ set_hostname and update_hostname
20:47 danlsgiga bhosmer_: Is there any way to do it using any resource already in salt?
20:47 jack__ joined #salt
20:47 bhosmer_ I don't think there is an actual function to do it, but file.managed maybe?
20:48 bhosmer_ Each config file might be specific depending on the AMI you're using though. I just do it with file.managed
20:48 danlsgiga bhosmer_: yup, that may do the trick... I'm also looking to have salt-cloud to auto update the OS on boot and restart after finished.
20:48 bhosmer_ Or maybe file.block_replace
20:48 mapu joined #salt
20:49 danlsgiga bhosmer_: Do you have the state example that you can share?
20:49 bhosmer_ Hang on, let me look...
20:51 Crypticus joined #salt
20:51 Crypticus Hello all
20:51 StolenToast I made a list of services and a jinja template to make a service.running state for each, but now I'd like to do more to a specific service.  WHat's the best way to do so?
20:52 StolenToast should I just make a new state for them?
20:53 baweaver joined #salt
20:54 bhosmer_ danlsgiga: Something like this maybe: http://pastebin.com/kLwMzH5t
20:54 Crypticus I’m running into an isssue I think salt must have already solved, but I’m not sure I am doing the right thing.  I want to run a script that does not exit.  It should succeed if it opens a port, I’m currently using a cmd.run with an unless to prevent it from running on a second highstate, but the first call succeeds no matter what.  Any solution for this?
20:54 hal58th joined #salt
20:55 danlsgiga bhosmer_: Looks good... but how are you setting the hostname? I'm using the CentOS 7 ami
20:55 danlsgiga bhosmer_: I don't think Salt set it
20:55 bhosmer_ danlsgiga: https://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.replace I left off the -name in the example I gave you
20:56 danlsgiga bhosmer_: cool
20:56 jfindlay StolenToast: you can use {% if ... %} ... {% endif %} to select specific services, perhaps
20:57 bhosmer_ danlsgiga: You can set the hostname with hostnamectl right? Or in the network settings? If you're using salt-cloud, doesn't route53 set it for you?
20:57 jfindlay Crypticus: I had that problem once and solved it with `cmd ; ( exit $? )`, which makes it look like a hack around some problem somewhere
20:58 jfindlay but I didn't have time to investigate it further
20:59 StolenToast jfindlay: I thought of a few ways, including jinja inheritence, but they seem like more trouble than they're worth
20:59 StolenToast I think it would be most straightforward to just make some exception states
20:59 StolenToast thanks though
20:59 danlsgiga bhosmer_: I'm not using route53 yet
20:59 jfindlay StolenToast: sure, the less jinja the better in my experience :)
21:00 danlsgiga bhosmer_: hostnamectl works only for centos7, centos6 doesn't have this command
21:00 danlsgiga bhosmer_: But I think I can use the cloud-config parameter to set it at creation time
21:00 Crypticus jfindlay: thanks, but my command simply never exits.  I really need some way to tell if the background process did what I need it to do,, in this case opens a port and starts listening.
21:01 Crypticus I used netstat to get the current state of the port in an unless block, but that only helps in preventing the command from being executied again.
21:01 bhosmer_ danlsgiga: I thought you said you were using CentOS 7? I think even in 7 you can still do '# hostname myhost.com' right?
21:02 danlsgiga bhosmer_: I can but it does not persist a reboot
21:02 bhosmer_ Yeah, because of the cloud.cfg updating it. It will after you remove those lines.
21:02 jack__ joined #salt
21:03 danlsgiga bhosmer_: I'm wondering if I just comment the update_hostname it will set the machine name as default
21:03 bhosmer_ I'm not following you.
21:05 danlsgiga bhosmer_: you know the tag Name of the ec2 instance? I'm wondering if we comment the line you suggested if cloud-init will grab that tag and set as the hostname
21:06 racooper_ joined #salt
21:06 bhosmer_ I don't think it will, but I could be wrong. I think that tag gets created after the instance starts to boot doesn't it?
21:08 bhosmer_ If I remember, even if I tag an instance, when it boots my hostname is still ip-1-2-3-4
21:08 danlsgiga bhosmer_: I think it is right before the instance is started
21:08 danlsgiga bhosmer_: hm... I'll give it a try
21:09 danlsgiga bhosmer_: How are you setting the hostname from route53?
21:09 gimpy927 joined #salt
21:10 bhosmer_ I would have maybe cmd.run or just an /etc/sysconfig/network file with your hostname in it. You could populate it with a pillar value too. With salt-cloud if you you use route53 it will set it for you.
21:10 danlsgiga bhosmer_: by default? don't you need to configure it?
21:11 gimpy927 Under Salt 2015.8.3 mount.mounted always tries to remount even if there are no state changes.  I added hidden_opts but it still remounts, how can I tell why Salt thinks it needs to remount it?
21:11 jack__ joined #salt
21:11 bhosmer_ I'm trying to remember, a co-worker configured that and I forget how he did it.
21:12 hasues joined #salt
21:12 hasues left #salt
21:12 writtenoff joined #salt
21:14 jfindlay gimpy927: what's the option?
21:14 gimpy927 bhosmer_: Any idea how to get Salt to print the reason though?  I tried -l debug, --verbose, and --state-output=full...none show why it remounted but I see it *can* say why
21:14 douardda joined #salt
21:14 jfindlay some no-remount options are treated differently in mount.mounted
21:14 gimpy927 @jfindlay: that's exactly what I'm trying to find
21:14 jfindlay hm
21:15 ageorgop joined #salt
21:15 jfindlay hidden_opts should work
21:15 jfindlay gimpy927: you need to set hidden_opts to the option that causes the remount
21:15 gimpy927 Yea, all I see though is: https://gist.github.com/anonymous/e11dca679de83b3e6f48
21:16 gimpy927 ... it doesn't say what option caused the remount
21:16 bhosmer_ Oh, jfindlay if you get a chance https://github.com/saltstack/salt/issues/28641 Can you point me which file.py needs to have this. It's been a while since I've been in the codebase and the file module/state has me confused.
21:16 saltstackbot [#28641]title: Problem doing file.managed with a http source | OS: https://vagrantcloud.com/boxcutter/boxes/oel67/versions/3.0.3...
21:16 jfindlay gimpy927: try `salt -l debug --state-output=full cn1.kamiak.example.edu state.sls netapp`
21:17 gimpy927 jfindlay: Nope, doesn't say the option
21:17 jfindlay bhosmer_: which thing do you mean?
21:18 denys joined #salt
21:18 bhosmer_ the streaming callback.
21:18 jfindlay gimpy927: does it output the mount command?
21:18 bhosmer_ Tornado has a TODO in their code as well for it.
21:19 om Hi all.  Been bashing my head on the keyboard with windows minions.  no matter what I try to run on the minion I get errors as if the built in modules are non-existent.
21:19 om IE.   Comment: State 'pkg.refresh_db' was not found in SLS 'windows_join_domain'
21:19 om Reason: 'pkg.refresh_db' is not available.
21:19 om Package win32net not found in the repository.
21:19 om Module function system.join_domain threw an exception. Exception: argument of type 'NoneType' is not iterable
21:19 om Comment: One or more requisite failed: windows_join_domain.join_domain
21:19 om Anyone using salt for windows minions that could help?
21:19 om any ideas?
21:19 jfindlay bhosmer_: I'm not sure; I'm not really familiar with tornado
21:19 Rumbles joined #salt
21:20 gimpy927 jfindlay: Doesn't look like it   https://gist.github.com/anonymous/57b1b496929e0df98735
21:20 bhosmer_ Sorry, my question for you was more for the SALT code. Specifically, where it needs to be added.
21:20 Crypticus joined #salt
21:20 bhosmer_ Oops, cachedout made the comment not you. Sorry.
21:21 om BTW, I think I found a bug in 2015-08-03 where windows minions would not pickup top.sls so I reverted to 2015-08-01 version
21:21 bhosmer_ "The problem here is that that Tornado seems to be spooling the entire file into memory. The likely fix here is to add a streaming callback into Salt's file client so that we can write out the file in chunks."
21:21 Eureka703 joined #salt
21:22 jfindlay bhosmer_: yeah, definitely ping cachedout on there, he'll know much more than I :)
21:22 danlsgiga bhosmer_: I think I can do what I want just using the hostname: myhostname parameter on the cloud-config file using the #user-data
21:22 jfindlay bhosmer_: https://github.com/saltstack/salt/pull/30704
21:22 saltstackbot [#30704]title: Issues/30431 get url large file | Fixes: #30431 ...
21:22 bhosmer_ I'll add it as a comment then. danlsgiga I can't off the top of my head tell you how the hostname gets set in route 53.
21:23 bhosmer_ jfindlay: Oh Sweet! Thanks!
21:23 jfindlay bhosmer_: that may not fix it, but it looks related
21:26 bhosmer_ yeah, I think the file.managed problem in my issue is related to simple_httpclient
21:28 gimpy927 jfindlay: it was the exec option...had to resort to salt-call from the minion to see it
21:28 jfindlay gimpy927: cool
21:32 jfindlay gimpy927: that makes sense, for some reason I thought you were running salt-call.  The remount command shows up in the minion, not master log
21:35 Striki left #salt
21:36 akhter joined #salt
21:37 virusuy left #salt
21:37 danlsgiga bhosmer_: K
21:38 N-Mi joined #salt
21:40 XenophonF joined #salt
21:40 jack__ joined #salt
21:43 jesusaur joined #salt
21:48 disbound joined #salt
21:49 subsignal joined #salt
21:49 mrwboilers joined #salt
21:50 om Any windows salt DevOps around that can help please?
21:50 om any support is welcome :D
21:50 om the issue I am having is almost 100% sure salt side, not windows per se
21:50 om but it is only an issue with the windows minions
21:51 rawzone^ joined #salt
21:52 mrwboilers Is there a size limitation for salt-cp? When I use it to copy a small file it works fine, but I just tried it on an 83MB file and it errors out.
21:53 jfindlay mrwboilers: what version?
21:53 mrwboilers 2015-8.3
21:53 giantlock joined #salt
21:54 mrwboilers The error it gets is about the master not responding. Salt request timed out
21:55 racooper joined #salt
21:55 tinyhippo left #salt
21:56 CheKoLyN joined #salt
22:00 yomilk joined #salt
22:05 heaje joined #salt
22:07 XenophonF any error on the master?
22:09 rem5 joined #salt
22:09 om mrwboilers: maybe increase the timeout in /etc/salt/master
22:09 bhosmer joined #salt
22:09 om if the file doesn't copy fast enough given it's 83 MB, it may timoute
22:09 jack__ joined #salt
22:11 om you can add the -t option
22:11 om to salt-cp as well
22:11 om default timeout is 5 seconds
22:11 om https://docs.saltstack.com/en/latest/ref/cli/salt-cp.html
22:12 om so yea, ignore the salt-master timeout that I mentioned and use the -t option
22:13 mrwboilers cool - thanks all!
22:13 jhauser joined #salt
22:14 evidence has anyone noticed the insanely long timeouts on schedule functions when the minion isn't running?
22:15 evidence ie when running highstate w salt-call if the minion isn't running (to say run a state to repair it and make sure all packages are sane and highstate schedule exists), the timeout is minutes for each of the schedule add and absent attempts
22:18 xf10e joined #salt
22:19 mosen joined #salt
22:22 jack__ joined #salt
22:24 jaybocc2 joined #salt
22:24 xf10e left #salt
22:27 hal58th joined #salt
22:27 ajw0100 joined #salt
22:27 jesusaur joined #salt
22:30 ralish Hey/lastlog evidence
22:30 ralish oops, sorry
22:31 evidence hey/lastlog yourself
22:31 ralish I asked this earlier but it was a quiet period, and is now a while ago: does anyone know of/can recommend a way to make a state run only if another state has no changes? I'm essentially wanting the opposite of on_changes.
22:31 ralish evidence: hehe, I did, for some reason I picked a random name to test the command given it's been so long since I used it
22:31 evidence onlyif/unless?
22:31 xf10e joined #salt
22:32 xf10e joined #salt
22:32 evidence or you actually need to ensure the other state specifically came back no changes?
22:32 ralish evidence: The issue is is if there are changes, I want a different state to run. Essentially I need a basic runtime branch. I can't do it in Jinja, as I can't know the result until after an earlier state has run which is after the state files are rendered
22:33 ralish So basically, if state A has changes run state B, otherwise run state C
22:33 jrgochan Is there a command to verify that a salt master is properly communicating with a gitfs backend?
22:34 ralish Sort of like an onchanges/else construct
22:36 jaybocc2 joined #salt
22:36 om2 joined #salt
22:37 N-Mi__ joined #salt
22:37 om2 joined #salt
22:37 evidence ralish: if there is something on the minion system you can trigger off of that'd make it pretty easy.  what is state A doing?
22:37 N-Mi joined #salt
22:37 om2 joined #salt
22:38 cilkay joined #salt
22:39 ralish evidence: Updates a configuration file. If the file has changes, I want to send a command to a daemon which will re-read the configuration and implicitly restart the correspoding underlying service. If there are no changes, I want to send a different command which restarts the service.
22:39 ralish I could just always run both, but the majority of the time the config file won't change, when it does, it means the daemon will be restarted twice, which isn't ideal.
22:39 ralish Not the end of the world, but was hoping for an elegant solution
22:39 evidence jrgochan: salt-run fileserver.update gitfs
22:39 evidence that will force the fetch on all repos, and you could see any errors
22:41 evidence ralish: you could set a boolean to true if you do whatever operation happens first, and then the second one would check and if it was true know a restart happened.  then it would know not the graceful.  i dunno there is prob a more elegant solution though
22:41 jrgochan thanks for the pointers guys! works like a charm.
22:42 akhter joined #salt
22:42 ralish evidence: What sort of variable. A grain?
22:43 evidence grains would persist each state run, but you should be able to set a variable w jinja based on the state return
22:44 N-Mi__ joined #salt
22:44 jack__ joined #salt
22:44 evidence or the 2nd command could even just be a simple 'onlyif: find -mtime +5 /path/to/config' heh
22:44 evidence ie only run the graceful if we haven't touched the conf in 5m, meaning the restart didn't happen
22:44 om2 joined #salt
22:44 evidence tuning based on how often you run this state/highstate obviously
22:45 zenlot joined #salt
22:46 ralish mmmm, yeah, creative thinking, one or both of those could work
22:46 ralish I'll give it a shot, thanks!
22:50 tehsu is there an if statement that would work if only it is defined in the grain
22:51 tehsu like if grains['asd'] is defined
22:54 cilkay Hello. I'm attempting to test for the presence or absence of a database and set a value to true or false accordingly in a JSON file but when I execute, I'm getting an exception. My Jinja template and the exception it throws are here: https://gist.github.com/cilkay/f147956725003fd50bbe What am I missing?
22:55 N-Mi joined #salt
22:55 N-Mi joined #salt
22:57 jaybocc2 joined #salt
22:57 ralish tehsu: you can use grains.get() in the grains module
22:58 ralish If it returns an empty string it's not set
22:58 ralish c.f. https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.grains.html#salt.modules.grains.get
22:58 perfectsine joined #salt
22:58 berserk joined #salt
22:59 jack__ joined #salt
22:59 tehsu using something like this?? {% set node_type = salt['grains.get']('node_type', '') %}
22:59 akhter joined #salt
22:59 tehsu so {% set my_check = salt['grains.get']('abc', '') %}
23:01 CheKoLyN joined #salt
23:02 ralish tehsu: Yep, then do an {% if my_check %}, an empty string should evaluate to False
23:02 tehsu ok
23:04 bhosmer_ joined #salt
23:04 akhter joined #salt
23:05 amanuel joined #salt
23:06 perfectsine joined #salt
23:07 N-Mi__ joined #salt
23:09 cilkay What is wrong with this sytax for a key/value pair in a JSON file? "autoMigrate": "{% if not postgres.db_exists '{{ grains['audit_dbms_db'] }}' %}"
23:10 tehsu awesome, thanks ralish that worked
23:10 cilkay The desired behaviour is for that to render to "autoMigrate": "true" or "autoMigrate": "false" depending on whether the Postgres db in question exists or not.
23:12 jack__ joined #salt
23:14 akhter joined #salt
23:26 writtenoff joined #salt
23:28 hasues joined #salt
23:28 hasues left #salt
23:38 baweaver joined #salt
23:40 yomilk joined #salt
23:41 nidr0x joined #salt
23:45 burp_ joined #salt
23:45 totzky hi people
23:46 totzky its friday!
23:46 totzky have a question regarding salt cli
23:47 totzky do you guys have any idea why salt cli is yamlifying the args?  I've seen this in the code
23:48 totzky and this is invalidating quoted string literals such as - ""help"" becomes help instead of "help"
23:48 sjmh it's friday.. ? o.O
23:49 totzky effectively not supporting these quoted strings
23:51 jrgochan It's Friday in most of Europ!
23:51 jrgochan e
23:51 jrgochan Sadly. It is not Friday in most of the exact location I'm currently sitting in
23:52 jrgochan (It's almost 2am where totzky's at!)
23:53 sjmh I was just making sure.  I've been so tired lately, I don't know what day it is lately.
23:53 totzky don't be sad few more hours mate! its 10.53am here down under
23:53 totzky @sjmh lol time for some wine
23:54 sjmh only 4 PM on Thursday here. :)
23:54 perfectsine joined #salt
23:57 sjmh totzky - not sure the reason 'why' - but yeah, appears it removes all the quotes.
23:57 sjmh there a reason you need a quoted string like that?
23:57 totzky @sjmh yes, an example would be boto_route53 module which requires the TXT value to be quoted
23:58 bhosmer_ joined #salt
23:59 sjmh hm.  so, LocalClient doesn't seem to mind it.

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