Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2016-09-16

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

All times shown according to UTC.

Time Nick Message
00:00 mattbillenstein1 the pillar data is roles -> hostname -> list of roles
00:00 mattbillenstein1 having trouble matching that
00:01 whytewolf ahhh, that would be a little more difficult
00:01 Guest41531 joined #salt
00:02 mattbillenstein1 yeah, perhaps this isn't structured correctly
00:02 PerilousApricot joined #salt
00:03 whytewolf humm, 'J@roles:.*:mongodb'  with match type being compound
00:03 whytewolf yeah seems to be counter productive for a roles setup as you kind of want to target on it
00:05 mattbillenstein1 that match doesn't work
00:06 mattbillenstein1 what's a better way to structure this?
00:06 whytewolf roles:role
00:07 whytewolf and target the pillars independently
00:10 mattbillenstein1 I see
00:11 mattbillenstein1 so in my states I also do stuff like
00:11 mattbillenstein1 % if 'mongod_master' in pillar['roles'][grains['id']]
00:12 mattbillenstein1 would just invert that
00:12 whytewolf if 'mongod_master' in pillar['roles']
00:12 mattbillenstein1 % if grains['id'] in pillar['roles']['mongod_master']
00:13 whytewolf no grains_id is need
00:13 mattbillenstein1 ok, so you were thnking in the pillar I have <hostname>:roles:<list of roles>
00:14 whytewolf no hostname in the pillar
00:14 mattbillenstein1 I need a hostname somewhere
00:14 whytewolf why
00:15 mattbillenstein1 multiple hosts serving different roles
00:15 mattbillenstein1 I have a pillar file for the cluster
00:15 whytewolf why are you sending all of the roles to all of the hosts?
00:15 mattbillenstein1 so like prod.sls -> list of hosts -> list of roles
00:15 mattbillenstein1 I'm not, but some hosts have multiple roles
00:16 whytewolf yes
00:16 whytewolf again why do you need the pillar to have all of the roles of all of the hosts on all of the servers
00:16 mattbillenstein1 that's where I've chosen to define that
00:17 mattbillenstein1 where else would it live?
00:17 whytewolf no where
00:17 whytewolf targetting
00:17 whytewolf a host should only need to know it's own roles
00:17 whytewolf not the roles of other servers
00:18 whytewolf thats WHY mine.get allows you to target
00:18 mattbillenstein1 so roles go in the minion config?
00:18 mattbillenstein1 is what you're proposing?
00:18 whytewolf no, still a pillar
00:19 whytewolf just targetted different
00:19 whytewolf more split up
00:22 netcho joined #salt
00:24 whytewolf https://gist.github.com/whytewolf/2765f921021297ab7865bde1025a023d
00:25 flowstate joined #salt
00:25 whytewolf makes using targetting way simpler. and thats what a roles model is for targetting
00:27 mattbillenstein1 yup, so my minion ids are hostnames basically
00:27 whytewolf yes
00:27 mattbillenstein1 so that's sorta where the mapping lives then
00:28 whytewolf exactly
00:29 mattbillenstein1 cool, thanks man ;)
00:48 woodtablet left #salt
00:49 Llmiseyhaa joined #salt
00:55 XenophonF joined #salt
00:57 toastedpenguin anyone get s3 working with salt?  I am getting this error: "The authorization header is malformed; a non-empty region must be provided in the credential" when I define s3 as pillar data for a minion
00:57 toastedpenguin I am just executing an s3.get
00:59 cro joined #salt
01:01 XenophonF i'm using it for state data
01:01 XenophonF would you be willing to post the full backtrace?
01:01 XenophonF make sure to strip out any of the authorization keys
01:02 subsigna_ joined #salt
01:03 thejrose1984 joined #salt
01:08 pipps joined #salt
01:09 mattbillenstein1 whytewolf: ah, so yeah, htis won't work
01:09 mattbillenstein1 so the other boxes need to know where the database is
01:10 mattbillenstein1 I render a small config file on each machine with this info
01:10 mattbillenstein1 which is why I need this data outside of the target
01:25 flowstate joined #salt
01:26 Nahual joined #salt
01:27 _JZ__ joined #salt
01:27 teryx510 joined #salt
01:29 jimklo joined #salt
01:29 ninjada joined #salt
01:30 _JZ___ joined #salt
01:32 mattbillenstein1 oh, but maybe this all works with mine.get now
01:37 ninjada hey guys, trying and failing to state an AWS launch config with boto_lc as per https://docs.saltstack.com/en/latest/ref/states/all/salt.states.boto_lc.html
01:37 subsignal joined #salt
01:38 catpigger joined #salt
01:38 ninjada adding my cloud init stuff, using the test example there, it's creating but the cloud_init script has some garbage with it like 'Content type ========'
01:38 pipps joined #salt
01:46 quasiben joined #salt
01:49 _JZ_ joined #salt
01:58 quasiben joined #salt
02:03 swa_work joined #salt
02:04 teryx510 joined #salt
02:06 beardedeagle joined #salt
02:07 _JZ_ joined #salt
02:10 corby left #salt
02:11 onlyanegg joined #salt
02:12 bluenemo joined #salt
02:20 ninjada joined #salt
02:22 teryx510 joined #salt
02:23 netcho joined #salt
02:24 flowstate joined #salt
02:27 ageorgop joined #salt
02:30 |aaron can i create a state that runs on my salt master and reads its minion's pillar data?
02:31 |aaron like i have my minion roles stored as pillar data and i want to use  states.ddns to create cnames based on that
02:32 subsignal joined #salt
02:33 lubyou joined #salt
02:37 |aaron maybe i can have the minion read its roles from the pillar and then put that in salt mine somehow?
02:39 netcho joined #salt
02:49 hasues joined #salt
02:49 hasues left #salt
02:50 ninjada joined #salt
02:50 swa_work joined #salt
02:52 sp0097 joined #salt
02:52 |aaron i cant get it to work though :( ive tried pillar.get and pillar.item as my mine_function but i just get nothing when i try to read the mine value
02:58 bastiandg joined #salt
02:59 teryx510 joined #salt
03:07 teryx510 joined #salt
03:14 jhujhiti i just upgraded from 2016.3.0 to 2016.3.3 and the output from state modules changed to this really awful yaml format, like "service_|-zabbix_agentd_|-zabbix_agentd_|-running:". it doesn't seem to be respecting state-output argument on the command line and i can't find anything in the release notes about this
03:14 jhujhiti has anyone seen this before?
03:21 hemebond jhujhiti: It's not normally YAML?
03:22 hemebond Or do you mean it's printed inline, not formated.
03:22 jhujhiti it's not normally like this anyway...
03:22 jhujhiti let me pastebin some
03:23 jhujhiti http://hastebin.com/ikosoburew.sql
03:24 flowstate joined #salt
03:24 jhujhiti oh, this is intereesting. i made some random changes to that one state and now it's stopped doing it on *all* states
03:25 hemebond Yeah, it looks like something in the file is causing newlines to go missing.
03:25 hemebond Or maybe not. I don't think I've ever seen this issue before.
03:27 jhujhiti i wish i had the old version of it in source control... but it was just a pkg.removed and service.dead state that i removed
03:27 jhujhiti i can't seem to make it happen again. bizzare
03:32 marc__ joined #salt
03:44 teryx510 joined #salt
04:05 teryx510 joined #salt
04:07 _W_ joined #salt
04:21 mpanetta joined #salt
04:23 flowstate joined #salt
04:28 Jimlad joined #salt
04:33 PerilousApricot joined #salt
04:35 xenoxaos joined #salt
04:40 netcho joined #salt
04:42 teryx510 joined #salt
04:43 subsigna_ joined #salt
05:04 DarkKnightCZ joined #salt
05:05 _JZ_ joined #salt
05:05 teryx510 joined #salt
05:22 netcho joined #salt
05:24 flowstate joined #salt
05:25 jxm_ joined #salt
05:25 etangle can I force salt-cloud cloud.providers to use eth0 as salt interface
05:35 jimklo joined #salt
05:42 teryx510 joined #salt
05:55 impi joined #salt
05:58 felskrone joined #salt
05:58 bocaneri joined #salt
06:03 ivanjaros joined #salt
06:05 teryx510 joined #salt
06:09 subsignal joined #salt
06:20 agend joined #salt
06:24 flowstate joined #salt
06:26 jimklo joined #salt
06:32 ub1quit33 joined #salt
06:39 Vaelatern joined #salt
06:39 pjs joined #salt
06:45 coredumb is it normal that version 2016.3.3 displays 2016.3.0 ?
06:46 babilen It shouldn't
06:46 coredumb weird
06:47 coredumb compiled it from 2016.3.3 tarball
06:48 coredumb # salt --version
06:48 coredumb salt 2016.3.0 (Boron)
06:50 hemebond Yes, I've seen that too.
06:50 hemebond When using from source.
06:50 hemebond What if you do salt-run manage.versions
06:51 hemebond Master should show a commit hash.
06:51 coredumb no commit hash only versions displayed
06:52 hemebond Oh you used a tarball.
06:52 babilen Where did you get the tarball?
06:52 hemebond Ignore me, I thought you used a git clone for some reason.
06:52 coredumb https://github.com/saltstack/salt/archive/v2016.3.3.tar.gz
06:53 coredumb tarball from GH release page
06:54 babilen That should have been correct
06:54 coredumb I even double checked the checksum of it from the one I use in my package
06:54 coredumb same
06:57 teryx510 joined #salt
06:58 coredumb https://github.com/saltstack/salt/pull/34847/files check this commit from the 3.3 changelog and the code is indeed present in mine
06:58 saltstackbot [#34847][MERGED] Add an option to skip the verification of client_acl users |
06:58 Rumbles joined #salt
06:58 coredumb so I guess someone just forgot to increase the version number ?
06:59 jhauser joined #salt
07:02 coredumb oh btw got a question about events
07:02 shadoxx bruh, spit your question. i'll answer with my truth
07:02 coredumb when running in multimaster mode, does the event sent by the client go  to all masters or only one ?
07:02 toanju joined #salt
07:03 coredumb send with event.send if that matters
07:05 coredumb shadoxx: so ? where's the truth now  bruh ?
07:06 shadoxx bruh. lemme drop this spice on you
07:06 * shadoxx runs away
07:06 coredumb :P
07:06 barmaley joined #salt
07:07 teryx510 joined #salt
07:07 babilen coredumb: I don't know, but it should be quite easy to answer
07:07 coredumb yeah ... I was lazy enough to not scratch another master :P
07:08 kshlm joined #salt
07:11 coredumb also
07:11 coredumb is there a way to make the minion still apply its state when the master is unreachable ?
07:12 coredumb as it has all the files in cache - supposedly if it has run a highstate at least once - it could just reuse them until he gets access back to the master
07:12 ivanjaros joined #salt
07:13 ninjada_ joined #salt
07:14 ivanjaros3916 joined #salt
07:14 kshlm joined #salt
07:15 impi joined #salt
07:22 dariusjs joined #salt
07:23 flowstate joined #salt
07:25 ronnix joined #salt
07:29 babilen coredumb: That's an interesting idea, but I believe that the minion errors if it can't reach the master and doesn't fall back to the last state run
07:31 coredumb babilen: from what I tested yeah it just errors
07:31 kbaikov joined #salt
07:32 Edgan joined #salt
07:36 babilen coredumb: Why do you require that? Maybe a masterless approach would be better for you?
07:37 coredumb babilen: no I'm used to cfengine
07:37 coredumb I don't want a masterless approach
07:39 JohnnyRun joined #salt
07:39 coredumb sometimes you can have servers that get disconnected from the master's subnet - like separate network - but you still want them to be able to ensure the compliance state without having to manage the files on it
07:40 babilen And you have external changes
07:40 babilen (because otherwise they should be in exactly the state they were in after the last highstate run)
07:41 coredumb you never know but you may have some
07:41 coredumb or not
07:41 babilen Sure, totally agree
07:41 coredumb but it's good that it can be enforced to the last known state
07:42 coredumb I wonder how hard it is to implement ...
07:42 felskrone does anyone have good examples on how to implement a beta-host deployment-structure? say i've changed a module and it gets synced to all hosts automatically, whats a good approach to have my changes only run on beta-hosts without cluttering my modules with if-expressions checking for a beta-grain/pillar?
07:43 babilen And you want the minion to regularly repeat those state runs, update highstate and pillar data *iff* it can talk to the master and just run against the cache if not?
07:43 coredumb babilen: indeed
07:43 babilen felskrone: Use a different master for beta hosts and have it track a different GitFS branch
07:43 oznt joined #salt
07:43 coredumb Right now I've ~500 servers that enforce their state every 5mn
07:44 teryx510 joined #salt
07:44 coredumb not with salt though
07:44 guerby joined #salt
07:44 coredumb if server is not reachable when the run starts, the "sync from master" method fails and the rest is applied from local cache
07:45 babilen coredumb: I think it is a very reasonable approach, but I'm not sure if you can do it at this point in time. I'd suggest opening a feature request or working on it yourself (SaltStack's focus is on bugfixing right now AIAUI)
07:45 jhauser joined #salt
07:48 thehaven joined #salt
07:48 JohnnyRun joined #salt
07:49 coredumb babilen: yep I'll check the code see if I can see something doable :)
07:51 felskrone babilen: my users are not too familiar with git yet, it should be something that relies on pillar/grains thats a little easier to use/understand than having different branches all over the place. any other idea?
07:53 babilen felskrone: The point is not git, but the separate master. I don't think you can target different versions of _modules to your minions if you have _modules in their file_roots. That's why you want to use different file_roots settings for those minions.
07:54 babilen You could do that with environments also, but that might open a different can of worms
07:54 felskrone yeah, i know, i use it like that in my vagrant-env :-). im just looking for an easier approach, just remember how you started with git, and then you know what im dealing with here :-)
07:55 babilen git was wonderful after svn and cvs :)
07:56 babilen (and no .. we do not mention rcs)
07:56 felskrone my users are already overwhelmed with salt managing our systems in general, from a simple distributed ssh and parallel scps, they jump into salt+grains+pillars+branches, i can not throw another branch/master at them, they will just collapse one after the other … :-)
07:56 babilen I understand
08:02 coredumb lol
08:02 coredumb good thing
08:03 coredumb replace the ones that collapse :P
08:05 geomacy joined #salt
08:05 krymzon joined #salt
08:07 felskrone hehe nah, i like most of them and rather give them the time to adjust, but they will have to climb a steep learning curve
08:09 mikecmpbll joined #salt
08:09 keimlink joined #salt
08:19 Rumbles joined #salt
08:21 N-Mi joined #salt
08:21 N-Mi joined #salt
08:23 flowstate joined #salt
08:24 netcho joined #salt
08:37 ThomasJ joined #salt
08:38 JohnnyRun joined #salt
08:42 sergeyt joined #salt
08:46 ozux joined #salt
08:51 ronnix joined #salt
08:53 dariusjs joined #salt
08:53 lero joined #salt
09:10 GreatSnoopy joined #salt
09:15 ronnix joined #salt
09:26 GreatSnoopy joined #salt
09:29 ronnix joined #salt
09:41 JohnnyRun joined #salt
09:43 LiamMon joined #salt
09:46 ronnix_ joined #salt
09:59 peters-tx joined #salt
10:04 teryx510 joined #salt
10:05 kbaikov joined #salt
10:11 ozux What's the best way to create a procedural StateFile? Or is it possible to have imperative Salt States? (As general question, also it like to use State formula for deployment)
10:13 kbaikov joined #salt
10:23 flowstate joined #salt
10:27 haam3r joined #salt
10:32 dariusjs joined #salt
10:36 coredumb Data passed to highstate outputter is not a valid highstate return: ['ID pillar_file_tree in SLS states/file-tree is not a dictionary']
10:36 coredumb How do I debug that ?
10:37 coredumb trying to set file_tree
10:37 coredumb oh wait ...
10:37 coredumb I'm stupid
10:48 ronnix joined #salt
10:48 krymzon ozux: not sure if this is what you are looking for, but perhaps have a look at 'Ordering States'. Or maybe cmd.script?
10:54 oyvindmo joined #salt
10:54 krymzon I am trying to make a 'preinstall clearout' state. I've tried combining "onchanges" with "require", but that just raises a 'recursive requisite' error. Is there some way in Salt to make A run before B, but only if B wants to run?
10:55 babilen ozux: What do you mean by that exactly? Do you just want to execute states in a specific order?
10:56 coredumb Is it possible to see a real life example of file_tree + contents_pillar?
10:57 amcorreia joined #salt
11:00 coredumb the doc is very much unclear
11:04 babilen What have you tried so far?
11:07 hemebond Can anyone tell me what keyname is in an EC2 Salt Cloud provider?
11:09 coredumb babilen: I've create a file-tree.sls pillar file in my pillar path
11:10 coredumb pillar.items shows up the file_tree: root_dir etc
11:10 coredumb but nothing else concerning files underneathj the tree
11:10 coredumb so I guess I'm missing something
11:10 babilen coredumb: Ah, so you are not referring to https://docs.saltstack.com/en/latest/ref/pillar/all/salt.pillar.file_tree.html ?
11:10 coredumb yes I am
11:11 cyborg-one joined #salt
11:11 babilen Why would you create a file-tree.sls in ...
11:11 babilen I guess that's your problem
11:11 babilen It's an external pillar and you would have to configure that in the master config
11:11 coredumb Yeah guess so ... where is it specified where I have to put the definition ?
11:11 babilen It has little to do with the "normal" pillar that resides at /srv/salt/pillar
11:12 babilen https://docs.saltstack.com/en/latest/topics/development/external_pillars.html#configuration
11:12 babilen I dislike the file_tree pillar, but maybe it comes in handy
11:13 coredumb why ?
11:15 ninjada joined #salt
11:15 coredumb babilen: I mean why do you dislike it ?
11:17 babilen Because is mixes data storage (files in a directory structure) with targetting
11:17 babilen It would have been a lot more useful if impelemented with a top.sls in which you assign directories/subdirectories/... to certain minions
11:18 coredumb agree but looks like the only to securely transfer files to targets right ?
11:19 babilen You could maintain their contents in YAML, but that's painful also
11:19 coredumb can't you have more than one file.managed entry under the same ID: ?
11:20 babilen No
11:20 coredumb arf
11:20 babilen That wouldn't uniquely identify the state anymore
11:20 krymzon I think I will go with approach like http://stackoverflow.com/a/30454953/4768734 for my problem. Which answer may also be relevant to ozux's question
11:21 XenophonF hey babilen i went back to messing with the gpg renderer last night and got it working with jinja
11:21 coredumb babilen: mmmh ok
11:21 babilen XenophonF: Please go on
11:21 XenophonF didn't realize i had to use `#!jinja|yaml|gpg`
11:22 babilen Does that allow you to encrypt entire files?
11:22 XenophonF after reading over that hiera-eyaml repo you linked, i guess per-element encryption makes sense to me now
11:22 babilen It does make sense, but is a pain in the arse to use
11:22 XenophonF the bit about ensuring that `git diff` was meaningful spoke most clearly to me
11:23 babilen True, you'd lose that
11:23 XenophonF so i'm hung up on two remaining bits of the workflow
11:23 flowstate joined #salt
11:23 XenophonF first is editor support, and i think i'm just going to have to hack on salt-mode in emacs to get things to work properly there
11:24 XenophonF (and then i need to teach my team of mostly windows admins how to use emacs)
11:25 XenophonF second is changing encrypted values, and i think that the only possible workaround is to encrypt values using everyone's keys
11:25 XenophonF so in emacs at least, i can encrypt a selection, and when prompted to choose users, i'll include both salt-master and myself and whoever else does config management stuff
11:26 XenophonF that way the encrypted values aren't write-only
11:26 XenophonF does that make sense or am i missing a better way to do things?
11:26 babilen XenophonF: Maybe something aking to "exncrypt/decrypt-region" of message mode would work?
11:26 XenophonF quite possibly
11:26 babilen But that would require people to use emacs
11:26 XenophonF right
11:27 babilen My colleagues already complain that they have to change individual passwords in pillars
11:27 XenophonF i mean, emacs is awesome, but dat learning curve
11:27 babilen I quite like spacemacs
11:27 babilen (with vim keybindings)
11:27 XenophonF right, but we're talking about folks whose text editor of choice is notepad
11:28 babilen Not even notepad++ ?
11:28 XenophonF no
11:28 XenophonF hmm
11:28 XenophonF i wonder if notepad++ would be suitable for this...
11:28 XenophonF i'll have to take a look at its support (if any) for gpg
11:29 babilen Maybe write a frontend for it then? "salt-secrets set foo:bar:baz s3cr1t"
11:29 XenophonF that might work if pillar sls files containing secrets are pure yaml
11:30 babilen You could read the pillar, parse the YAML, encrypt the new data and then write the YAML again
11:30 XenophonF well, except that the pillar gets stored at github
11:31 babilen Yeah, but that would contain the encrypted data
11:31 coredumb https://docs.saltstack.com/en/latest/faq.html#linux-unix < in the fork/disown way how do I transform the multine command in a pillar that is reusable ?
11:31 babilen I was just thinking about the "setting a value" bit, not the git workflow
11:32 coredumb tried to call with -name: | {{ pillar['xxx'] }} but minion spouts error  about incorrect syntax and stuff
11:32 coredumb is  that even possible ?
11:32 coredumb in the case of the cmd.wait indeed
11:32 babilen coredumb: What are you trying to call?
11:32 babilen Could you paste your state and output to one of http://refheap.com, http://paste.debian.net, https://gist.github.com, http://sprunge.us, … ?
11:33 coredumb babilen: exactly the same as in the FAQ
11:33 coredumb except I've put it in a pillar
11:33 bluenemo hi guys. How are older versions of salt-minion compatible to newer versions of salt-master? Is salt generally trying to be backward compatible?
11:33 babilen coredumb: You can update salt-minion without that these days
11:34 babilen bluenemo: master has to be newer than minion and I wouldn't jump more than two major versions at most
11:35 ronnix joined #salt
11:35 coredumb babilen: really ?
11:35 coredumb last time I've restarted master after inputing new config file it just returned error while the master had correctly restarted
11:35 bluenemo coredumb, babilen from my experience old minions just wont reply
11:36 babilen I've seen that in older version also
11:36 coredumb well I don't have "old" version so what should I use ?
11:36 babilen You can use salt 'some-minion' pkg.upgrade
11:36 coredumb yes indeed
11:37 ronnix_ joined #salt
11:37 coredumb but the point is handling the upgrade as part of the highstate
11:37 babilen It'll upgrade the minion in its own process and starts a new process .. that way the upgrade continues and you get the new minion
11:38 XenophonF so to close the loop on this gpg business, i'm going to find the author of salt-mode and work with them to extend it to support encryption
11:38 babilen coredumb: You can use https://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkg.html#salt.states.pkg.uptodate -- but I wouldn't necessarily do that as it is *really* nice if your states are idempotent
11:38 quasiben joined #salt
11:38 babilen XenophonF: I ♥ you
11:38 XenophonF and for the encryption workflow, i'm going to encrypt values with both my and the salt-master gpg key
11:39 XenophonF and i'll see about doing the same with notepad++'s maintainer
11:39 XenophonF no promises
11:39 Rumbles joined #salt
11:40 coredumb so each time you modify your master config your go restart it by hand ?
11:40 babilen coredumb: I use https://github.com/saltstack-formulas/salt-formula/
11:41 coredumb ok gonna read on that :)
11:42 babilen coredumb: Have you used formulæ before?
11:43 bluenemo babilen, whats that?
11:43 coredumb babilen: nope
11:43 coredumb 2016-09-16 13:40:38,185 [salt.pillar                                               ][CRITICAL][9892] Specified ext_pillar interface root_dir is unavailable
11:43 coredumb 2016-09-16 13:40:38,186 [salt.loaded.int.pillar.file_tree                          ][ERROR   ][9892] file_tree: no root_dir specified
11:43 coredumb for now I can't even use file_tree :D
11:44 babilen coredumb: Paste your config -- https://docs.saltstack.com/en/latest/topics/development/conventions/formulas.html
11:44 coredumb http://tpaste.us/2rre
11:44 coredumb my /etc/salt/master.d/ext_pillar.conf
11:45 coredumb nothing fancy I guess
11:46 babilen coredumb: Could you indent line 3 and 4 one more level please?
11:46 coredumb yeah was doing just that !
11:46 babilen Looks fine otherwise
11:46 coredumb pffff
11:46 babilen pffff?
11:47 coredumb yeah that was that
11:48 coredumb NOW I SEE MY FILE!
11:48 coredumb ahhhhh thx for the help babilen
11:49 coredumb now you've given me bunch of more stuff to read :P
11:49 babilen You are welcome .. don't you love YAML?
11:49 coredumb bah
11:49 coredumb why would it be valid everywhere else than this file ?
11:49 coredumb :(
11:51 coredumb works fine
11:51 babilen It's a different data structure: http://paste.debian.net/825331/
11:52 coredumb OK
11:52 babilen The '-' bit throws a lot of people off as it adds a level of indentation already
11:55 coredumb ok
11:55 sandro_ joined #salt
11:56 marc__ joined #salt
11:57 ozux babilen: Yes, for this use case, I just want the State formula, get files, set load blanacer, deploy, and clean up the temp files. So I want it to be in procedural order.
11:57 ozux So what is the proper way of making sure if I need to have procedure? (and Possibly not using runner modules)
11:58 jhauser_ joined #salt
11:59 babilen ozux: SLS files are being processed top-to-bottom
12:00 ozux babilen: ehem, sorry I'm from Puppet world, where every thing are in unknown order :P  So then require statements are just for external dependencies in SaltStack am I right?
12:01 babilen ozux: No, as soon as you add requirements salt might reorder states.
12:02 babilen (the general order is still top-to-bottom)
12:02 ozux babilen: thanks
12:03 babilen So .. if you have a hard dependency it would make sense to express that with a suitable requirement
12:06 dariusjs joined #salt
12:23 coredumb http://pastebin.com/PXcyScK5 < what's wrong with this jinja code ?
12:28 manji coredumb, does <% work in jinja?
12:29 hemebond no
12:29 manji coredumb, it is {%
12:29 manji and %}
12:29 hemebond Pretty sure coredumb made the same mistake last night :-D
12:29 coredumb my god I always to the same mistake
12:30 coredumb sorry being stupid folks
12:30 hemebond lol, not a problem
12:30 manji coredumb, did you do any erb templates somtime in this lifetime?
12:30 manji because it would make total sense to confuse it
12:30 coredumb manji: not even
12:30 coredumb :D
12:30 manji hahaha
12:30 manji den I have no idea why you confuse it :p
12:31 manji then*
12:31 coredumb manji: me neither
12:31 coredumb ok it's better but give me errors
12:31 coredumb 2016-09-16 14:30:39,327 [salt.state                                           ][CRITICAL][7371] Rendering SLS 'base:states/borgmaster' failed: Jinja syntax error: expected token ':', got '}'; line 9
12:31 coredumb {% for dir in {{ dir_list }} %}    <======================
12:32 coredumb O_o
12:32 manji when you are within {% %}
12:32 manji your variables do not need {{ }}
12:32 manji so
12:32 manji {% for dir in dir_list%}
12:32 manji is what you are looking for
12:32 coredumb AH!
12:33 coredumb I actually did that first
12:33 coredumb but changed when I had errors with <> :D
12:34 * coredumb feels like a noob
12:35 hemebond Is there actually a way to use Salt Cloud to create an EC2 instance in a private subnet when you're not running from within the VPC?
12:35 coredumb manji: again, thx
12:35 manji cheers mate
12:38 flowstate joined #salt
12:43 dendazen joined #salt
12:46 lovecraftian joined #salt
12:46 bowhunter joined #salt
12:53 DarkKnightCZ joined #salt
12:53 XenophonF hemebond: yes, but post-deployment provisioning is tricky
12:54 XenophonF there's an option to feed salt-cloud that says "don't run bootstrap"
12:54 XenophonF and there's another option to send a custom script via userdata, with which you can deploy the minion
12:55 XenophonF alternatively (and this is the path I chose), you can set up a VPC peering connection between where the minion lives and where the master lives, and then use private IPs to bootstrap (also an option in the salt-cloud configs somewhere)
12:55 XenophonF although now that i'm starting to deploy stuff in other regions, i'll probably figure out the userdata script stuff
12:56 XenophonF (the only issue i have with userdata scripts is that i don't know if there's a way to clear them out post-deploy, and i don't want salt keying material available to non-salt processes running on the minion post-deploy)
12:58 XenophonF another approach might be to re-key the minion post-deploy? i dunno if that's possible
12:59 XenophonF or maybe an OTP in the userdata script that retrieves the minion keymat
13:00 XenophonF maybe other aws customers can weigh in? i'm probably doing it rong
13:00 drawsmcgraw joined #salt
13:02 dariusjs joined #salt
13:16 subsignal joined #salt
13:17 bluenemo hi guys. I joust found the syslog.config state - I read through the code and it just calls the syslog module without any statefulnes-magic. Do you know why it is always just executing the module and then appending each time it is run? As in why is it a state at all if its not stateful?
13:19 cmarzullo I'd say cause it's probably incomplete. For syslog we just wrote our own formula that, given pillar, writes out the file as needed.
13:20 sergeyt joined #salt
13:21 bluenemo thats what it appears to me too. I have to say there is a lot of fancyness managing that config file - I'm not that deep into syslog, so I cant tell if its worth the effort.
13:22 bluenemo the module code says stuff like Builds syslog-ng configuration. This function is intended to be used from the state module, users should not use it directly!
13:22 bluenemo but the config state is really noting but a wrapper to the module
13:22 bluenemo I dont get it
13:24 DarkKnightCZ joined #salt
13:25 racooper joined #salt
13:33 nyx_ joined #salt
13:35 subsignal joined #salt
13:58 armonge joined #salt
14:03 DarkKnightCZ joined #salt
14:04 spuder joined #salt
14:05 Electron^- joined #salt
14:06 nidr0x joined #salt
14:08 perfectsine joined #salt
14:09 mpanetta joined #salt
14:12 _JZ_ joined #salt
14:19 mikecmpbll joined #salt
14:20 heaje joined #salt
14:20 Karunamon hey folks - is there a "sane" way to get a list of failures when I do a global ('*') highstate from my master?
14:20 johny_ joined #salt
14:21 johny_ hey everyone! I'm seeing an issue where an SLS within a GitFS state is simply not being executed for me (no failure logs or anything). The issue is submitted here but no reply yet: https://github.com/saltstack-formulas/logrotate-formula/issues/23 anyone got any ideas?
14:21 saltstackbot [#23][OPEN] jobs.sls not being executed | Hi,...
14:32 Sylvain31 joined #salt
14:33 tvinson Karunamon: salt --state-output=terse '*' state.highstate
14:34 tvinson Karunamon: that's the "sanest" way i can think of. you could apply logic to any of the machine-readable returners. JSONPath comes to mind.
14:35 Karunamon tvinson: mm. Just wondered what was available without dipping into writing my own parser
14:37 tvinson i guess you can set --state-verbose=False too
14:37 tvinson pretty readable
14:41 flowstate joined #salt
14:43 Sylvain31 hi, I use, git.lastest, I got a "Repository would be updated to 3f35a56, but there are uncommitted changes." as I modified the cloned repository it is expected. how can I clone a git repos, alter the copy, removing install dir, and not use force_reset: True, because I only want to update if some changes appear on the source repos? not always.
14:47 TyrfingMjolnir joined #salt
14:47 kusen joined #salt
14:48 TyrfingMjolnir joined #salt
14:49 TyrfingMjolnir joined #salt
14:49 ronnix joined #salt
14:49 TyrfingMjolnir joined #salt
14:50 spuder joined #salt
14:52 marc__ joined #salt
14:52 TyrfingMjolnir joined #salt
14:53 impi joined #salt
14:54 TyrfingMjolnir joined #salt
14:54 cscf Sylvain31, you want to merge remote changes with local changes?
14:55 edrocks joined #salt
14:57 Sylvain31 not at all, it is done by some state
14:58 Sylvain31 cscf: I may, clone the repos and checkout from the branch…
14:59 cscf Sylvain31, so you git clone on the minion, using git.latest, and then other states make non-git changes to this directory?
15:00 Sylvain31 cscf: http://paste.debian.net/825453/
15:01 Sylvain31 yes
15:02 cscf Sylvain31, then you want these changes to be preserved, while also regularly updating the git repo?
15:03 Sylvain31 cscf: a warning may sufice yes on state.highstate, I may dived into an export an a check state?
15:03 Sylvain31 dived = divide
15:04 hasues joined #salt
15:04 hasues left #salt
15:05 cscf Sylvain31, what about committing the changes the states make?
15:06 cscf Sylvain31, or, looking at your paste, why not fork the repository and change the configs there?
15:06 TyrfingMjolnir joined #salt
15:07 Brew joined #salt
15:09 Sylvain31 cscf: hum… not the same context salt://source/config.conf vs forked repos + git commit + push…
15:10 Sylvain31 cscf: I gonna test to export instead of cloning, and write a test newest rule…
15:12 dariusjs joined #salt
15:14 marc__ joined #salt
15:14 tiwula joined #salt
15:14 whitenoise__ joined #salt
15:14 Sylvain31 with https://docs.saltstack.com/en/latest/ref/states/all/salt.states.git.html#salt.states.git.detached I think.
15:19 Karunamon So here's something i've just spent an hour on to not much benefit... i'm using mysql as an external job cache with mostly success, but i'm not able to retrieve an old job with jobs.lookup_jid
15:19 Karunamon from what I can see on the debug output, I get part of the job output, and then salt tries to connect to the database again, and bails out with a "MySQL server has gone away"
15:20 Karunamon we can eliminate the network, since the database is local on the salt master
15:21 Karunamon it's like there's some improper connection reuse or something? I'm not even close to being a DBA so i'm a bit lost now.
15:28 jimklo joined #salt
15:33 beowuff joined #salt
15:45 G2fx joined #salt
15:49 avozza joined #salt
15:53 avozza joined #salt
15:53 ageorgop joined #salt
15:53 Rumbles joined #salt
15:56 cscf Minion is failing to start on a new host, even with -l debug it says nothing about why.
15:56 DammitJim joined #salt
15:56 TheoSLC joined #salt
15:57 DammitJim can one do compound matching in a top.sls for targeting minions?
15:57 Lionel_Debroux joined #salt
15:57 cscf DammitJim, you mean matching multiple patterns in a single regex?  Yes I think so.
15:57 DammitJim cscf
15:58 DammitJim well, from what I can understand, one cannot do something like this in a top file
15:58 mpanetta joined #salt
15:58 DammitJim 'server(101|345).*.com'
15:59 DammitJim -match: pcre
15:59 cscf Oh ok
16:00 DammitJim where my servers are:
16:00 cscf I'm not sure about that
16:00 DammitJim server101.test.com
16:00 DammitJim server101.prod.com
16:00 DammitJim server345.test.com
16:00 cscf right
16:00 DammitJim server345.dev.com
16:00 DammitJim so, I was thinking I could do pcre OR glob
16:01 Minnow47 joined #salt
16:02 mirko joined #salt
16:05 chadhs joined #salt
16:07 jimklo_ joined #salt
16:08 waynr joined #salt
16:09 DammitJim oh, it seems I could do something like:
16:09 DammitJim 'server(101|345).[a-z]*.com'
16:13 amcorreia joined #salt
16:13 Nahual joined #salt
16:27 marc__ joined #salt
16:29 woodtablet joined #salt
16:32 cmarzullo So anyone using salt to do http load testing. Like checking endpoints and stuff?
16:34 sergeyt left #salt
16:35 JPT Nah, more like a configuration management to automatically install/update packages, set up user accounts, ...
16:39 teryx510 joined #salt
16:40 McNinja ^ +1 using it to deploy and manages nodes in the cloud
16:43 cmarzullo yeah I gots the statemanagent down. Loading a custom app. Just was thinking about setting up like locust.io or something through salt.
16:43 cmarzullo Was wondering if someone had done something similar.
16:44 McNinja not yet but you could always install some bench marking tools and have them hit an endpoint
16:45 McNinja issue ad hoc commands to the nodes via salt
16:45 cmarzullo gonna start on that. . . after a cheesesteak.
16:45 McNinja haha
16:55 chadhs_ joined #salt
16:55 quasiben1 joined #salt
16:56 pipps joined #salt
17:00 _JZ_ joined #salt
17:01 mikecmpbll joined #salt
17:01 onlyanegg joined #salt
17:03 ivanjaros joined #salt
17:12 edrocks joined #salt
17:15 TheoSLC joined #salt
17:26 flowstate joined #salt
17:26 pipps joined #salt
17:26 onlyanegg joined #salt
17:37 kusen joined #salt
17:45 quasiben joined #salt
17:54 ozux joined #salt
17:55 cscf I have master and minion on the same host, both say they started correctly, but salt-key shows nothing.  Any ideas?
18:00 babilen You didn't configure the "master" setting in the minion config
18:03 myraft_ joined #salt
18:03 nZac joined #salt
18:06 ub1quit33 has anyone else had issues with really slow connections to repo.saltstack.com lately?
18:07 subsignal joined #salt
18:09 GreatSnoopy joined #salt
18:11 Trauma joined #salt
18:11 babilen ub1quit33: You are not the first to complain about it
18:13 ub1quit33 I wasn't complaining, just checking if anyone else had the same issue..
18:13 abednarik joined #salt
18:13 ub1quit33 but that does answer my question none the less
18:13 ub1quit33 ty
18:13 cscf babilen, ah there was a mistake in the hostname.  Thanks!
18:13 babilen ub1quit33: Sure, didn't meant to imply that actually. It came up before, so you are not the first to notice it
18:14 subsigna_ joined #salt
18:14 babilen cscf: \o/
18:16 heewa joined #salt
18:16 Trauma joined #salt
18:17 subsignal joined #salt
18:18 heewa Anyone have tips for diagnosing why salt’s complaining it can’t find a state (not mine, one of the things you call, like ‘pkg.installed’)?
18:19 babilen heewa: Typically because you are missing some (Python) dependency.
18:19 whytewolf if it is something in pkg. then most likely you are using an unsupported operating system
18:20 heewa It’s `webutil.useradd`, which is actually in the file `{repo}/salt/states/htpasswd.py`, but renamed using `__virtualname__` to `webutil`. I checked that the unix tool it checks for (also called `htpasswd`) exists, and in fact I can use the  `{repo}/salt/modules/htpasswd.py` module, which does the same check.
18:20 babilen heewa: To really debug this take a look at the __virtual__ method of the module
18:20 babilen heewa: Are both loaded?
18:20 heewa @babilen I tried, but it looks like it should work. And I don’t see a debug log entry about it.
18:20 heewa What does it mean to be loaded?
18:21 avozza joined #salt
18:22 whytewolf webutil? so this is a custom module? if so is in in a _state directory and have you synced the file to the minion
18:22 babilen heewa: Well, do you have two modules competing for the same name
18:22 babilen I also didn't find .. ah!
18:23 babilen (see whytewolf's question)
18:23 heewa @whytewolf It’s an included module & state, but the file is called `htpasswd.py`
18:23 edrocks joined #salt
18:23 heewa (Like, I didn’t write it and put it in a `_states` or `_modules` dir)
18:24 flowstate joined #salt
18:24 whytewolf htpasswd is included. but renaming to webutil is not
18:24 whytewolf oh wait ...
18:24 whytewolf this is odd.
18:25 whytewolf salt.states.htpasswd.user_exists but the example shows webutil.user_exists
18:25 heewa Oh, I didn’t mention, I’m on the `develop` branch, latest
18:25 whytewolf one sec, "To the source!"
18:27 whytewolf humm. return __virtualname__ if salt.utils.which('htpasswd') else False
18:27 heewa @whytewolf @babilen I added a log line right next to that line and it looks like the result is True, so it should return __virtualname__, which I also logged, and verified is “webutil"
18:28 heewa So, salt is attempting to load the file, and checking virtualname, and getting back “webutil”, but still can’t find it
18:28 lorengordon joined #salt
18:29 whytewolf what a second. ... in your original question you asked about webutil.useradd ... but that function doesn't exist in htpasswd state file
18:29 whytewolf it is user_exist
18:29 whytewolf errr user_exists
18:30 kusen joined #salt
18:30 heewa OMAGOSH! I’m so sorry, guys. I was using a module function instead of the state one.
18:30 heewa `webutil.useradd` is a module fn, I’ve been chasing my tail for so long, didn’t realize I needed `webutil.user_exists`
18:31 heewa Oh, haha, yea, you found my mistake at the same time. Thanks!
18:35 subsignal joined #salt
18:35 bltmiller joined #salt
18:35 bltmiller so, `state.show_top` is a thing – is there an equivalent for pillar? e.g. `pillar.show_top`?
18:35 cmarzullo yes
18:35 whytewolf pillar.items?
18:36 whytewolf :P
18:36 bltmiller whytewolf: not pillar.items, that shows the dictionary, I want to see which SLS files are being applied to a target
18:36 bltmiller also worth noting that I'm using git_pillar
18:37 whytewolf you mean you can't tell by what pillars are in the system?
18:37 bltmiller think from a minion, like I can use `salt-call state.show_top`
18:38 whytewolf well pillar.show_top does exist ... however a.) it is a runner show no nothing you call from a minion will work. and b.) currently doens't work with git_pillar
18:38 whytewolf https://github.com/saltstack/salt/issues/25891
18:38 saltstackbot [#25891][OPEN] pillar.show_top runner doesn't work with git_pillar | I get no output for the following command:...
18:39 whytewolf minions don't see the pillar top file. they are only handed their pillars
18:39 bltmiller aha
18:40 bltmiller hm, well that issue implies that without git_pillar, that command is available/works?
18:40 whytewolf salt-run pillar.show_top
18:42 bltmiller ahhh I see. hm, that's interesting that a minion can get the state top file, but not pillar top file.
18:42 whytewolf states are rendered on the minion, pillars are not
18:44 whytewolf pillars are meant to be more secure so a minion only gets access to the pillars that are rendered for it
18:44 bltmiller right, and if that's the simple reasoning behind it, it seems much more like being able to run `salt-call state.show_top` is simply consequence of minion rendering state files
18:45 bltmiller I mean even running that on a minion, you still only see what's rendered for that minion. I'd think that same behavior could be applied to pillar top file as well
18:46 whytewolf no, you can actually look up any state file in the system from a minion.
18:46 whytewolf and can see the entire top file
18:47 bltmiller lunch time, bbl, thanks for the explanation whytewolf :)
18:54 haam3r joined #salt
19:07 Renich joined #salt
19:07 Renich o/
19:07 Renich hey salters!
19:07 Renich I am trying out salt-ssh
19:07 bluenemo joined #salt
19:07 Renich and I am getting this error when I try the ping: ERROR: salt requires python 2.6 or newer on target hosts, must have same major version as origin host
19:07 cscf So, I'm on 16.04 and want newer salt than 2015.8.8 - where do I get it ?  The PPA is out of date
19:08 Renich But, this works: salt-ssh  --raw od0 '/usr/bin/python -V'
19:08 Renich and it returns python 3.5.
19:08 Renich Btw, I am trying to use this on Arch Linux; on an ODROID C2 arm device
19:08 babilen Renich: Are you sure that 2.6 or 2.7 are installed?
19:09 babilen Ah, Arch with their wonderful decision to switch 'python' to python 3.*
19:09 Renich babilen: I didn't know python2 was a requirement
19:09 Renich babilen: thanks
19:09 Renich babilen: and all of them eventually, man
19:10 babilen all of them eventually?
19:10 * Renich is installing python2
19:10 babilen I don't quite follow
19:10 Renich babilen: well, Fedora is making the switch soon. That means RHEL and CentOS will too. Gentoo has already done it
19:10 Renich babilen: python3 on all distros
19:10 babilen https://www.python.org/dev/peps/pep-0394/ is quite clear about this
19:10 raeven joined #salt
19:11 babilen Switching the version of Python that is installed by default is not the same as switching the command 'python' from python2 to python3
19:11 Renich babilen: OK, I get you
19:11 whytewolf basicly they are not following pep 394
19:11 Renich babilen: python should point to python2
19:11 babilen It should, yes
19:12 kusen joined #salt
19:12 Renich got it. Anyway, the immediate issue is solved. Ping works and now I can start to play with this one
19:12 babilen "however, end users should be aware that python refers to python3 on at least Arch Linux (that change is what prompted the creation of this PEP), so python should be used in the shebang line only for scripts that are source compatible with both Python 2 and 3."
19:12 Renich Thanks a ton, man
19:13 babilen IOW: Arch caused us and everyone massive pain and we somehow try to at least spell out the rules for the others so that they don't come up with something crazier ;)
19:13 babilen Renich: Have fun :)
19:13 Renich babilen: thanks, dude
19:14 pipps joined #salt
19:15 bltmiller joined #salt
19:23 heewa So, I have a Pull Request out for a fix, and another bug fix that is a slight change on top of the first one. Should I put out another PR where the “base” is my other branch, and change the “base” (in github UI, not a git “rebase” operation) after my first PR is merged?
19:24 flowstate joined #salt
19:27 woodtablet left #salt
19:33 swa_work joined #salt
19:40 doubletwist joined #salt
19:43 heewa Diff question: how should I approach a module function that returns not enough info for my use case? Specifically, it’s using `__salt__[‘cmd.run’](..).splitlines()`, which just returns the stdout text, when I need `__salt__[‘cmd.run_all’](…)`, which returns a dict that contains the exit code, stdout, and stderr. Can I change the function? What about people that are using it expecting it to return a string? Do I add a new
19:43 heewa function? Do I add an argument to optionally switch on the different behavior?
19:44 whytewolf if you need a change to a built in module. copy the module to a _module, or _state directory. make the changes and sync. custom modules overright built-ins
19:45 heewa I need this for a bug fix to a state file included in the source, not for my personal use
19:46 heewa How about a flag `all_results` defaulting to False (the old behavior)?
19:46 heewa If True, uses `run_all` and returns the full dict, otherwise uses old code untouched (`run` with `splitlines()`)
19:47 whytewolf that would be perferable. so that you don't change current behavour
19:47 heewa Thanks!
19:48 sagerdearia joined #salt
19:54 anotherZero joined #salt
19:54 kusen joined #salt
19:55 numkem joined #salt
19:58 swa_work joined #salt
20:00 DammitJim joined #salt
20:05 zmalone joined #salt
20:08 agend joined #salt
20:08 flowstate joined #salt
20:12 mikecmpbll joined #salt
20:14 Vaelatern joined #salt
20:16 pipps joined #salt
20:16 cro joined #salt
20:17 perfectsine joined #salt
20:21 cro joined #salt
20:27 djgerm joined #salt
20:31 Lionel_Debroux joined #salt
20:36 StolenToast I keep having a package installation state fail because the name of the package "cannot be found" even though I specify the source rpm in the salt file server
20:36 kusen joined #salt
20:37 StolenToast I also told yum (centos7) to install it and then copied the name yum extracted to the ID of the state, which will just be the name
20:37 djgerm left #salt
20:38 StolenToast here's the state, tried on an nfs share as well http://hastebin.com/udeqocatuy.makefile
20:39 nZac joined #salt
20:42 Sketch https://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkg.html
20:43 Sketch doesn't list a source option for pkg
20:44 dendazen joined #salt
20:44 Sketch i think what you probably need to do is make a local yum repo, and install it from there
20:45 Sketch you can use https://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkgrepo.html to add/manage repos
20:45 Sketch (though you'll still have to create/populate/host the repo somewhere)
20:48 Sketch hmm, it's not in the docs, but i do see this https://github.com/saltstack/salt/issues/2291
20:48 saltstackbot [#2291][MERGED] Extend pkg to install from file | Was discussing this in #salt. It would be nice to specify a .deb or .rpm file and have Salt install it....
20:48 Sketch https://github.com/saltstack/salt/issues/7184
20:48 saltstackbot [#7184][MERGED] Module yumpkg installing from source file uses rpm -qip, should use yum localinstall | If you need an .rpm installed on a minion, you can set up a state that looks something like this:...
20:48 Sketch maybe your formatting is just wrong...
20:48 StolenToast so this works, sort of: I changes pkg.installed "sources: \n - flashupdt: <sourcepath>"
20:49 StolenToast uh let me just paste the new one
20:49 Sketch oh wait, sources is mentioned in the docs.  just not source.  not sure how i missed that...
20:49 StolenToast http://hastebin.com/biquqayimi.php
20:50 StolenToast well it's a bit strange, I think my original thing should work, maybe in the future
20:50 StolenToast here's an easy workaround though
20:51 StolenToast it "sort of" works because right now it's failing some version check, but that's a problem for Monday
20:55 edrocks joined #salt
20:57 kusen joined #salt
20:58 pipps joined #salt
20:58 PerilousApricot joined #salt
20:58 pipps joined #salt
20:59 hasues joined #salt
20:59 hasues left #salt
21:08 mavhq joined #salt
21:14 amcorreia joined #salt
21:17 _JZ_ joined #salt
21:18 sagerdearia joined #salt
21:19 qman__ joined #salt
21:20 esharpmajor joined #salt
21:22 flowstate joined #salt
21:24 APLU joined #salt
21:24 adongy joined #salt
21:25 cliluw joined #salt
21:33 frew joined #salt
21:34 frew anyone know how to get some kind of debug logs from salt for when boto_ec2 mysteriously fails?
21:35 frew minion logs imply that all is well
21:37 cliluw joined #salt
21:39 kusen joined #salt
21:53 amcorreia joined #salt
21:54 pipps joined #salt
22:00 kusen joined #salt
22:04 flowstate joined #salt
22:12 tiaz joined #salt
22:13 tiaz Hello, I'm trying a simple salt configuration, but am running into problems. In particular, I'm following this user creation thing: https://clinta.github.io/Salt-User-Management/
22:13 tiaz ... you know, that's from mid-2015, maybe I'll try one of the package installation things instead
22:13 tiaz nm, I'll be back in a bit if I find problems with that.
22:14 ageorgop joined #salt
22:16 Nahual joined #salt
22:17 JPT joined #salt
22:21 kusen joined #salt
22:22 flowstate joined #salt
22:25 ageorgop joined #salt
22:29 cyborg-one joined #salt
22:38 quasiben1 joined #salt
22:39 sagerdearia joined #salt
22:41 quasiben joined #salt
22:42 quasiben2 joined #salt
22:48 hemebond XenophonF: Thanks very much, I'll look into both of those.
22:58 tiaz Is it possible to summarize responses to a grains.item query?
22:59 tiaz e.g., `salt '*' grains.item osrelease`, I don't want to see the individual response of every node, I just want to see e.g. 16.04, 15.10 (where those are both present in the group of hosts)
22:59 tiaz or do I have to parse the response format myself to do that (which isn't a huge deal)
23:02 nexus_ joined #salt
23:04 kusen joined #salt
23:23 ageorgop joined #salt
23:24 flowstate joined #salt
23:25 kusen joined #salt
23:28 fatbox joined #salt
23:31 zer0def joined #salt
23:47 jimklo joined #salt
23:54 justanotheruser joined #salt
23:59 ninjada joined #salt

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