Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2017-10-16

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

All times shown according to UTC.

Time Nick Message
00:02 XenophonF I'm going to have to set up vncserver.
00:02 XenophonF X over SSH is too slow.
00:03 brianthelion joined #salt
00:09 whytewolf can you not replicate it on a master local to your computer?
00:16 ujjain joined #salt
00:16 ujjain joined #salt
00:41 zerocool_ joined #salt
00:53 motherfsck joined #salt
01:34 rgrundstrom joined #salt
01:56 ilbot3 joined #salt
01:56 Topic for #salt is now Welcome to #salt! <+> Latest Versions: 2016.11.8, 2017.7.2 <+> Support: https://www.saltstack.com/support/ <+> Logs: http://irclog.perlgeek.de/salt/ <+> Paste: https://gist.github.com/ <+> See also: #salt-devel, #salt-offtopic <+> We are volunteers and may not have immediate answers
02:02 blu_ joined #salt
02:05 hasues joined #salt
02:20 zerocool_ joined #salt
02:46 JPT joined #salt
02:48 hasues Can pillar be served over gitfs for masterless minions?
03:10 sjorge joined #salt
03:20 masber joined #salt
03:31 evle joined #salt
03:39 johnj_ joined #salt
03:41 XenophonF whytewolf: haven't tried that yet
03:41 XenophonF going to though
03:42 XenophonF actually what I want to do is V2V the master from EC2 to VMware or VirtualBox
03:43 XenophonF if VMware, then I could import into vSphere in Chennai
03:43 XenophonF if Vbox then I could run it here at home
03:50 beardedeagle joined #salt
04:22 hoonetorg joined #salt
04:40 johnj_ joined #salt
05:25 omie888777 joined #salt
05:41 johnj_ joined #salt
06:19 guest joined #salt
06:32 ilbot3 joined #salt
06:32 Topic for #salt is now Welcome to #salt! <+> Latest Versions: 2016.11.8, 2017.7.2 <+> Support: https://www.saltstack.com/support/ <+> Logs: http://irclog.perlgeek.de/salt/ <+> Paste: https://gist.github.com/ <+> See also: #salt-devel, #salt-offtopic <+> We are volunteers and may not have immediate answers
06:39 mariusv joined #salt
06:39 mariusv joined #salt
06:40 pualj joined #salt
06:42 johnj joined #salt
06:43 do3meli joined #salt
06:44 do3meli left #salt
06:53 jhauser joined #salt
06:54 schasi joined #salt
07:00 obitech joined #salt
07:00 ccha joined #salt
07:09 ivanjaros joined #salt
07:11 aldevar joined #salt
07:16 mbologna joined #salt
07:17 Hybrid joined #salt
07:19 Miouge joined #salt
07:20 k_sze[work] joined #salt
07:22 pbandark joined #salt
07:25 pualj_ joined #salt
07:28 robman joined #salt
07:29 impi joined #salt
07:37 Trauma joined #salt
07:41 pualj joined #salt
07:43 johnj joined #salt
07:43 usernkey joined #salt
07:46 darioleidi joined #salt
07:47 deskpot joined #salt
07:47 absolutejam So,
07:47 absolutejam I know I've asked this before
07:47 absolutejam But returners
07:47 absolutejam Do they _have_ to be configured on the minion?
07:48 absolutejam because I don't really want my Slack API key on all my minions
07:53 whytewolf the description of "returners" :"All Salt commands will return the command data back to the master. Specifying returners will ensure that the data is _also_ sent to the specified returner interfaces."
07:53 aldevar left #salt
07:54 whytewolf so, no as default a normal returner only returns from the minion.
07:55 whytewolf an event returner works from the master. however. i don't think the slack returner as the event_returner function to operate like that
07:55 whytewolf nor can it operate as a master_job_cache
08:01 absolutejam hm, okay
08:01 absolutejam makes sense
08:01 absolutejam Guess I'll just deploy the config along with ACLs in a state then
08:02 absolutejam not much else I can do
08:02 aldevar joined #salt
08:04 bdrung_work joined #salt
08:13 colttt joined #salt
08:15 jab416171 joined #salt
08:18 _KaszpiR_ joined #salt
08:29 Rumbles joined #salt
08:29 Naresh joined #salt
08:32 Micromus joined #salt
08:32 absolutejam question number 2
08:32 absolutejam Using the scheduler with cron: requires python-croniter
08:32 absolutejam Salt packages pip with the minion for Windows minions
08:33 absolutejam Is there any specific arguments I need (besides pip_bin) to allow a state to install it?
08:33 Mattch joined #salt
08:34 absolutejam if I manually run `C:\salt\bin\script\pip2.7.exe`, it tries to install to current user's appdata
08:38 absolutejam nevermind, think there's a specific doc on this
08:44 johnj joined #salt
09:10 filthyG joined #salt
09:25 omie888777 joined #salt
09:25 toanju joined #salt
09:34 k_sze[work] joined #salt
09:44 felskrone joined #salt
09:45 johnj joined #salt
09:56 colttt joined #salt
09:58 cyteen joined #salt
09:59 dendazen joined #salt
10:23 KaczuH joined #salt
10:26 pualj joined #salt
10:27 pualj_ joined #salt
10:28 _pualj_ joined #salt
10:31 pualj joined #salt
10:34 pualj_ joined #salt
10:43 colttt joined #salt
10:44 pualj joined #salt
10:46 johnj joined #salt
10:52 pualj joined #salt
10:57 deskpot joined #salt
11:03 pualj_ joined #salt
11:29 Trauma joined #salt
11:47 johnj joined #salt
11:53 dendazen joined #salt
11:56 ibro joined #salt
11:57 usernkey1 joined #salt
11:58 Neighbour Are salt-runner instances supposed to be based on a local master configuration, or are they based on a blank config?
12:10 Nahual joined #salt
12:11 Kelsar joined #salt
12:23 pualj joined #salt
12:40 major joined #salt
12:42 evle1 joined #salt
12:48 johnj joined #salt
12:50 noobiedubie joined #salt
12:56 skatz joined #salt
12:57 ProT-0-TypE joined #salt
12:59 skatz_ joined #salt
12:59 zerocool_ joined #salt
12:59 numkem joined #salt
12:59 EvaSDK grmbl, is it possible at all to use formulas with multiple envs ?
13:00 EvaSDK seems like I'm always stuck on either having conflicting IDs or having missing states
13:02 EvaSDK seems pretty easy to avoid the issue for self written states but for formulas it seems hopeless
13:02 EvaSDK and I'm stuck calling state.highstate with pillarenv and saltenv set
13:02 schemanic joined #salt
13:09 zerocool_ joined #salt
13:11 XenophonF EvaSDK: I use formulas with multiple environments.  I also use salt-formula to set this up.
13:12 XenophonF something like this - https://github.com/irtnog/salt-pillar-example/blob/master/salt/example/com/init.sls#L319
13:12 XenophonF salt-formula's salt.formulas SLS (say that 5x fast) will clone into the specified environments for you
13:13 XenophonF and the salt.master SLS will automatically populate the file_roots for you
13:13 XenophonF so just apply salt.formulas,salt.master in that order
13:13 edrocks joined #salt
13:18 EvaSDK hum
13:19 EvaSDK I do use salt-formula for this as well, will try this list method but I don't think it fixes the state id collision problem
13:20 XenophonF state id collision problem?
13:20 XenophonF if you have SLS or state IDs that collide with those used in formulas, rewrite your states
13:21 XenophonF I use C-style namespaces
13:21 XenophonF e.g., state IDs in foo/bar/init.sls are all prefixed with foo_bar_
13:21 EvaSDK no, I mean I use the states from formulas in multiple envs
13:22 EvaSDK so having them checked out in these envs and not specifying stateenv in state.apply call makes it collide
13:22 EvaSDK I'm using 2016.11 but this is a problem I'm having since I started using salt with ~2015.3
13:22 XenophonF don't do that?
13:23 EvaSDK well
13:23 XenophonF not to sound snarky but I only every apply stuff from one environment
13:23 XenophonF s/every/ever
13:23 XenophonF but then my environments == DTAP phases
13:23 XenophonF which might be different than you
13:24 EvaSDK that's what I did up to now, but I'm getting sync of repeating this arguments on cli and having to run highstate multiple times to update all minions
13:24 EvaSDK dev/preprod/prod
13:24 EvaSDK s/sync/sick
13:25 EvaSDK I could put highstate in a reactor or whatever is the automation to get it run automatically but that seems like overkill
13:26 EvaSDK I'm actually considering getting rid of environments due to this issue
13:26 XenophonF state.apply should DTRT - maybe your top.sls file(s) are buggy?
13:26 EvaSDK DTRT ?
13:26 XenophonF sorry do the right thing
13:27 EvaSDK hum like formula being applied on a minion from 2 envs ?
13:27 XenophonF yeah that shouldn't happen
13:27 EvaSDK I'll triple check but this should not be the case
13:27 XenophonF can't happen in my setup
13:28 XenophonF here's my salt-states repo of it helps to have a concrete example to work from - https://github.com/irtnog/salt-states
13:28 usernkey joined #salt
13:28 XenophonF there's only the one top file in the base environment (master branch)
13:29 EvaSDK crap, looks like this is the problem
13:29 EvaSDK base have role:postgres and minion target has it as well
13:29 EvaSDK well well, got bitten by unfinished migrations it seems
13:29 XenophonF yay you found it!
13:30 EvaSDK let's clean this up
13:31 XenophonF if it's any consolation, I'm in the middle of migrating to apache-formula and tomcat-formula
13:31 XenophonF and _everything_ is broken
13:31 ahrs joined #salt
13:32 XenophonF and on top of that, my master isn't working right
13:32 EvaSDK :)
13:32 XenophonF so I can imagine how frustrated you were by all this :)
13:32 EvaSDK it's hard to get dedicated time to track down these issues
13:32 EvaSDK so yeah
13:32 XenophonF no kidding
13:33 XenophonF I'm at the point where I'm debati ng whether to pay for some consulting hours from SaltStack
13:34 EvaSDK :)
13:34 EvaSDK not sure my boss would allow it, he is a bit crazy about potential data leaks
13:35 XenophonF that's what NDAs are for
13:35 EvaSDK sure, that's what I tried to explain to him
13:36 XenophonF although come to think of it, since I work a govt job, I'd have to run them through the Position for Public Trust crap
13:36 XenophonF ugh
13:36 XenophonF forgot about that
13:36 EvaSDK oh boy
13:36 EvaSDK sounds like crazy paper work
13:37 XenophonF it is
13:37 XenophonF thank god we're civilians and not natsec
13:38 XenophonF OK I'm going to go beat my head against the Python debugger some more
13:38 XenophonF TTYL
13:40 EvaSDK w00t, it does work now, thanks a lot
13:40 EvaSDK now to chase remaining conflicts
13:43 beautivile joined #salt
13:45 ouemt joined #salt
13:49 johnj joined #salt
13:53 dnull joined #salt
13:54 absolutejam Anyone used the http.utils module?
13:55 absolutejam Using it in a module and it's frustrating me that it only returns status and error
13:55 absolutejam Nothing of value back from the API
13:55 absolutejam Is there a flag I'm missing perchance?
13:57 cgiroua joined #salt
14:02 racooper joined #salt
14:05 mchlumsky joined #salt
14:12 masber joined #salt
14:21 dayid joined #salt
14:26 Hybrid joined #salt
14:26 wavded joined #salt
14:30 absolutejam that, or I'll just have to copy the http utils function
14:30 absolutejam seems dirty
14:45 Hybrid joined #salt
14:46 fatal_exception joined #salt
14:47 whytewolf absolutejam: have you read https://docs.saltstack.com/en/latest/topics/tutorials/http.html
14:50 edrocks joined #salt
14:50 fxhp joined #salt
14:50 johnj joined #salt
14:55 fatal_exception joined #salt
14:55 heaje joined #salt
15:00 XenophonF hm, how best to move this master
15:00 edrocks joined #salt
15:00 XenophonF I think I'm going to push the existing master config to a new minion and then copy the pki stuff
15:00 XenophonF that should be enough, right?
15:06 skatz joined #salt
15:07 sarcasticadmin joined #salt
15:09 sarcasticadmin joined #salt
15:10 beardedeagle joined #salt
15:14 relidy joined #salt
15:18 socket- Hey all, whats the best way to reestart a service on fedora? Looks like reload is not supported.  https://apaste.info/GjHF
15:19 slugfish left #salt
15:21 slugfish joined #salt
15:22 slugfish left #salt
15:22 EvaSDK XenophonF: yes
15:23 EvaSDK XenophonF: btw, the maze continues, now hitting formulas depending on another formula which is available in both dev and base thus pulling the same IDs in the same context :/
15:24 EvaSDK namely a state calling custom pythonapp formula calling users-formula
15:24 EvaSDK thus users formula is being called twice, once for base and once for dev
15:27 tom[] lxc.present in the minion version i have (2016.12.2) does not support thinpool. is there a way to force a --thinpool=TP option through to the lxc-create command from the state def
15:27 dayid I'm having an issue with using git.latest; I am trying to do a shallow-clone of a non-default branch (repository default is master - I want to have both local and remote point to be tracking production). I've toggled with using force_fetch, specifying both rev and branch, but can't appear to get all of them working together
15:28 dayid I've gotten to where it is locally named master (but tracking remote production) - but it's a non-shallow repository - and I've gotten a shallow repository, but I can't seem to get it to work off of the initial clone (as well as subsequent updates) to keep the depth shallow and track a non-default branch
15:28 EvaSDK wouldn't be such a problem if I had a final iteration of my LDAP server so I could move all enterprise accounts out of the way
15:31 Neighbour EvaSDK: I've found a "bug" in salt-run that compiles the pillar from all environments (latter overriding earlier ones) if you don't specify a pillarenv in the CLI (even though the master config *does* specify a pillarenv)
15:31 Neighbour Is that something that could also affect your use case?
15:32 dayid Anyone doing a shallow clone of a non default branch successfully? Docs make it look simple but I'm not quite a few hours in and it seems no more obvious: https://slexy.org/raw/s2JURerqbM?token=c6ae93cb738d1fbd24020cbae830f371dac29abd91dd393d5c4907582953dc04&amp;ts=1508168516
15:32 dayid I also am getting different behaviour when i first run the state (and it attempts to clone) vs subsequent runs
15:33 EvaSDK Neighbour: I actually want to avoid specifying saltenv & pillarenv
15:33 EvaSDK as it requires extra care & operations when updating all minions
15:34 EvaSDK it boils down to not being able to (specify) a dependency between states of different envs
15:35 EvaSDK there's no reason to run the users-formula from my dev env by I need some of it to run with data specific to the env
15:35 EvaSDK app user is being created by it
15:37 EvaSDK not being able to do this basically means I cannot use formulas to build new states on
15:39 EvaSDK hum
15:40 EvaSDK maybe there would be a way if env state execution would only be able to access pillar data from the specific env ?
15:40 EvaSDK is it would (in this case) not create colliding state IDs
15:40 EvaSDK has
15:40 EvaSDK as
15:43 _JZ_ joined #salt
15:44 EvaSDK oh well, this is https://github.com/saltstack/salt/issues/14733
15:48 eykd joined #salt
15:48 noobiedubie joined #salt
15:48 noobiedubie you can set the enviroment per host in the minion conf
15:48 noobiedubie that way you don't have to pass pillar or salt env and can just highstate
15:49 noobiedubie or leverage roles
15:49 pbandark hi.. is it possible to pass cloud-init script with salt-cloud for openstack provider ?
15:51 EvaSDK noobiedubie: roles do not help, they are also tied to env
15:51 BitBandit joined #salt
15:51 johnj joined #salt
15:52 EvaSDK also setting env per host does not help as it means common states (in base in my case) would not be executed
15:52 EvaSDK this would work if I stayed with previous status quo though
15:53 EvaSDK I guess I need to find a way to pass this through salt-cloud
15:53 EvaSDK maybe via profiles
15:55 EvaSDK would still lead to having to manage duplication of pillar data and states across envs for the common part
15:56 DammitJim joined #salt
15:58 kshlm joined #salt
16:06 aldevar left #salt
16:19 XenophonF EvaSDK: that's why I tell people not to use include/extend
16:19 XenophonF (re: your earlier messages)
16:21 onlyanegg joined #salt
16:22 onlyanegg joined #salt
16:30 nixjdm joined #salt
16:32 impi joined #salt
16:36 Kira joined #salt
16:40 dayid ah, it appears --depth is only referenced during a clone
16:40 dayid not during subsequent pulls
16:45 onlyanegg joined #salt
16:48 edrocks joined #salt
16:52 johnj joined #salt
16:54 tiwula joined #salt
17:06 lane_ joined #salt
17:07 cgiroua joined #salt
17:28 aarontc joined #salt
17:33 pualj joined #salt
17:44 major XenophonF, was thinking of kicking an orchastrate from the highstate for the master
17:45 major XenophonF, honestly, I don't see the entire orchestrate as beeing all that different from the highstate, except that I need to coordinate it all such that things happen to specific machines in specific orders
17:46 major XenophonF, the lxd-formula seems to be a good example of when I need to set the lxd_password on a node before I can add that node to a different nodes lxd-config
17:53 johnj joined #salt
18:00 skatz_ joined #salt
18:02 XenophonF major: understood
18:03 brianthelion joined #salt
18:04 ChubYann joined #salt
18:15 colegatron joined #salt
18:18 sjorge joined #salt
18:24 whytewolf major: little known fact. since the module that is used in orchestration is a state module. [saltmod] orchestration states can be run on the master directly as standard state files. at least in my testing this worked. been a while since i have tested it.
18:26 cyborg-one joined #salt
18:27 major whytewolf, I had started suspecting as much, I just hadn't tried it out yet
18:28 major there doesn't seem to be an actual naming convention forced that I could tell, just some strong wording to encourage a certain use pattern
18:44 Pulp joined #salt
18:54 johnj joined #salt
19:05 _KaszpiR_ joined #salt
19:05 Hybrid joined #salt
19:10 onlyanegg joined #salt
19:11 nixjdm joined #salt
19:19 hammer065 joined #salt
19:30 lkolstad joined #salt
19:35 aldevar joined #salt
19:38 Edgan whytewolf: yeah, I have done that before
19:41 LeProvokateur joined #salt
19:42 thumprslc joined #salt
19:42 fatal_exception joined #salt
19:42 misconfig joined #salt
19:44 misconfig Hi everyone, wondering if someone could lend me some assistance with jinja syntax in a state.
19:44 pualj joined #salt
19:46 misconfig I'm trying to set a jinja variable from a grain, using a variable as the key: https://gist.github.com/ndobbs/603d076c4c65793d7996fd105832d430
19:48 misconfig This in particular is returning None => {% set ip = grains.get('ip4_interfaces{{ intf }}')[0] %}
19:48 misconfig do I need to escape that jinja variable?
19:48 whytewolf {% set ip = grains.get('ip4_interfaces:' ~ intf)[0] %}
19:48 whytewolf don't use {{}} inside {%%}
19:50 whytewolf also, just for ease of use look up network.ip_addrs
19:51 wavded joined #salt
19:51 misconfig Thanks for the assist @whytewolf!
19:51 colabeer joined #salt
19:52 pualj joined #salt
19:55 johnj joined #salt
20:04 nixjdm joined #salt
20:10 sjohnsen joined #salt
20:11 edrocks joined #salt
20:16 mayk joined #salt
20:23 brianthelion joined #salt
20:28 absolutejam Yeah whytewolf, but I don't think it alters the return values?
20:29 whytewolf ?
20:29 absolutejam I might be wrong. Just trying the different methods now
20:29 absolutejam re: http
20:29 absolutejam I'd read that but not seen the 'backends'
20:30 whytewolf what are you trying to get?
20:31 absolutejam I'm wrapping the http util module in another module
20:32 absolutejam well rewriting the Kapacitor module because it sucks
20:33 absolutejam and one of the API calls accepts either a POST or PATCH
20:33 absolutejam if the task already exists, and you try and POST it, it returns that the task already exists
20:33 absolutejam but currently the http module just returns 400: BAD REQUEST
20:35 whytewolf the http module only returns what it is given.
20:36 Splix76 joined #salt
20:38 whytewolf a 400 is a Bad request.
20:38 Splix76 Is there a way to be sure that my top.sls order is always processed from the top down? I see evidence of some grain setting states being skipped when I run it. If I run 'state.apply' 3x, it'll capture them all, however I would like to see a single run of state.apply process everything in order in one pass.
20:38 absolutejam if I curl the API with the same values, I get more output though
20:38 pualj_ joined #salt
20:38 whytewolf then Kapacitor is using 400 wrong.
20:38 absolutejam The http only responds with status and status
20:39 absolutejam status and error*
20:39 Splix76 I have read the ordering pages in the docs, and from what I can tell I should be forcing top down ordering of the top.sls file.
20:40 aldevar left #salt
20:40 Splix76 I only have an env of 'base' currently, so env order should not come into play here.
20:43 whytewolf humm, absolutejam nothing they show should be responding with a 400. the only 400 codes i see in the doc are 403 and 404
20:45 whytewolf Splix76: if jinja is used to get grains that you are relying on being set in the same state run. that won't happen
20:45 absolutejam Yeah, I thought the same but that's what Iget if I try and re-run the modulewhen the task already exists
20:46 whytewolf absolutejam: when you try it manually is that the error code you get that has "more info?"
20:47 relidy joined #salt
20:48 Splix76 whytewolf, that appears to be the issue I am facing.
20:49 Splix76 I have to run state.apply a few times to get all things set and states applied.
20:49 whytewolf Splix76: jinja renders before any state is applyied.
20:49 Splix76 It seems I will need to set the grains first in a state run ahead of state.apply for newly fired up instances.
20:50 Splix76 The issue that presents is that I filter based on naming standard, and compound matches in Jinja are not as easily done as in a top.sls.
20:50 Splix76 i will take it back to the white baord, I need to have these grains set ahead of a state.apply
20:50 Splix76 thanks.
20:50 whytewolf Splix76: look up the match module
20:52 whytewolf absolutejam: you should tell Kapacitor the error code for conflict is 409.
20:55 Splix76 Thanks, I am using -match: compound in the top.sls, just need to find the proper syntax to shift that into a normal state if syntax check in Jinja.
20:56 Splix76 I am sure I will sort it out, I have some reading to do then I will re-work this to run as a stand alone state prior to state.apply.
20:56 relidy joined #salt
20:56 whytewolf {% if sal.match.compound('matching stuff goes here') %}
20:56 johnj joined #salt
20:56 absolutejam Just trying it to see what exactly I get with a clash whytewolf
21:00 pbandark joined #salt
21:04 nixjdm joined #salt
21:38 ult joined #salt
21:49 Splix76 whytewolf, I was able to get the state modified to work outside of the top.sls and then run the state.apply. This resolved the issue I was facing, thank you.
21:52 onmeac joined #salt
22:03 Trauma joined #salt
22:03 nixjdm joined #salt
22:05 fatal_exception joined #salt
22:19 schasi joined #salt
22:22 omie888777 joined #salt
22:24 oida joined #salt
22:26 laertus joined #salt
22:26 masber joined #salt
22:26 laertus is an EC2 t2.micro instance only 1 GB of memory enough to run linux and play around with salt?
22:26 dendazen joined #salt
22:26 Trauma joined #salt
22:39 cyteen joined #salt
23:05 pbandark1 joined #salt
23:15 GrandMasterV joined #salt
23:57 brokensyntax joined #salt

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