Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-03-31

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

All times shown according to UTC.

Time Nick Message
00:08 teskew joined #salt
00:08 Hell_FireW joined #salt
00:11 subsignal joined #salt
00:12 mpanetta joined #salt
00:20 dalexander joined #salt
00:20 theologian joined #salt
00:30 lowkeyliesmith joined #salt
00:35 johnkeates joined #salt
00:37 johnkeates left #salt
00:41 jessie_ joined #salt
00:41 jessie_ currently have salt cloud installed and understanding the error taking place which is Profile is not defined
00:41 jessie_ could anyone help around this on how to debug
00:42 MindDrive hemebond: It will return a string, but not a return value.  So without looking at what was returned, I can't tell if it succeeded or not.  Not sure if it's by design or something else, but it is what it is.
00:43 otter768 joined #salt
00:43 mafrosis joined #salt
00:44 mafrosis lo salt
00:44 subsignal joined #salt
00:44 mafrosis anyone across the releases for 2014.7.x at the mo?
00:44 iggy jessie_: paste your config files (remove any sensitive info), the command you ran (preferably with -l debug), and the error received to gist.github.com
00:44 mafrosis I see a new tag for 2014.7.4 is up on github now, am wondering if that means that 2014.7.3 is frozen
00:44 iggy mafrosis: ENOPARSE
00:45 mafrosis iggy: I dont follow
00:45 iggy your first 2 msgs were full of a bunch of words and fragments of words that didn't exactly make sense together
00:46 cberndt joined #salt
00:46 jessie_ iggy: root@saltSRV-DO-AWS-AZURE:/etc/salt/cloud.providers.d# cat my-digitalocean.conf DO:   provider: digital_ocean   api_key: ###################   client_key: b#bb0#92b#bf#00e#9a8#########   ssh_key_file: /root/.ssh/id_rsa   ssh_key_name: id_rsa   location: New York 1
00:46 jessie_ iggy: thanks
00:46 iggy jessie_: paste to gist.github.com, that is unreadable
00:47 murrdoc can u paste in the key too
00:47 murrdoc (please dont)
00:47 jessie_ iggy: https://gist.github.com/anonymous/2a5075d0149ed4226ad7
00:47 iggy oh murrdoc... that delightful little scamp
00:48 mafrosis iggy thanks
00:48 jessie_ iggy: https://gist.github.com/anonymous/9bfac2d5a78dda0ff322
00:48 mafrosis left #salt
00:48 iggy jessie_: I asked for the config files, the command you ran (preferably with -l debug), and the error received
00:48 jessie_ lol murdoc... test enviroment... basiclaly meaning all is free
00:48 f4lse joined #salt
00:49 jessie_ ok, sorry iggy
00:49 iggy and protip... gist allows multiple files per paste
00:50 iggy so I don't have to open up 40 chrome tabs
00:50 iggy <-- super lazy
00:50 jessie_ iggy: first time using gist, will try to get all in one area
00:50 cpowell joined #salt
00:51 mpanetta joined #salt
00:51 iggy there's a button below the text entry field that says add file
00:52 mpanetta joined #salt
00:52 iggy which doesn't seem all that non-obvious, but someone complained the other day that they didn't see it
00:53 jessie_ like so iggy  https://gist.github.com/anonymous/8e5821fe44e4406f4017
00:54 iggy -p is profile (not provider)
00:54 iggy your profile should be UB1
00:55 iggy which I never would have figured out if I hadn't insisted on you giving me the info I originally asked for
00:55 iggy so I'm not an asshole, I just know what info I need to help people (and get slightly annoyed when I have to ask for it twice ;)
00:55 iggy okay... a bit of an asshole
00:55 jessie_ if someone needs, help, at least provide all the info. :) i respect that
00:58 jerematic joined #salt
00:58 jessie_ iggy: sorry something isnt completely clear to me here, if i have built a profile for Digital Oceon and the provider. why would i use UB1
00:59 iggy because that's what you called your _profile_ (the only thing called DO there is your provider)
01:01 beneggett joined #salt
01:01 jessie_ ah just seen it ... what a bell i am
01:01 jessie_ 4 am here
01:02 iggy it was all pretty confusing to me at the start too
01:04 andrew_v joined #salt
01:05 JDiPierro joined #salt
01:06 jessie_ do we have a syntax page for salt-cloud
01:09 iggy not really... it's kind of different for each provider
01:09 aqua^mac joined #salt
01:10 jessie_ ah ok
01:10 jessie_ iggy: https://gist.github.com/anonymous/4486099b9ecea0ea3fa9
01:10 jessie_ iggy: still having a slight issue, do we have any syntax for debuging like a flag
01:11 iggy salt-cloud -l debug -p ....
01:11 iggy -l debug behind any salt command should be a good first start
01:11 jessie_ oh my days. this is how dumb feels
01:12 jessie_ working now
01:13 jessie_ thast digital oceon sorted :) now aws and azure, just for kicks
01:14 iggy DO was fairly easy... gce was a pita... aws looked like it could be a pita, but it's fairly well documented (being as it's probably the most popular)... never tried azure
01:15 jessie_ when a new vm instance is installed, from salt-clou it seems that it installs salt minion automatically on the new instance ?
01:15 iggy correct
01:16 jessie_ iggy: thanks for your help around this appreciated. only used salt for teh cloud package and nothing else as of yet
01:16 iggy fair enough
01:16 iggy not the normal way to start, but hey... times are a changing
01:16 jessie_ my understanding of this so far is, that its just cfg management pull and push, but no real orchestration and auto scaling..
01:17 iggy it can't handle the autoscaling _itself_
01:17 jessie_ yeah i wanted to deploy instance through cloud first and then config manage that way... so that i can see exactly what happens as and when it happens... and then measure the parallel process's
01:17 ITChap joined #salt
01:18 iggy but it it has good api support (so your monitoring software can trigger an api call that will scale up, or whatever works best for your situation)
01:18 fusionx86 joined #salt
01:18 iggy and it can do fairly complex orchestration (but that area isn't heavily documented, so...)
01:19 jessie_ so if i use aws auto scaling... or azure auto scaling deploying the vms etc... how salt can keep up with that.
01:19 iggy there are a few ways
01:22 iggy one way would be to have the minion already installed in an ami and just have it use the api to authorize a new key as it came up
01:22 iggy (or use auto_accept/autosign if you are trusting)
01:23 jessie_ ah so threshold based, and the  ami kicks in to make the call for more vms
01:23 iggy you can also have salt-cloud on the master do it based on the salt scheduler
01:23 jessie_ that is quite config driven... dependancies required, minion install before in production.
01:24 jessie_ the schedular would cover just peak hours... or when you think peak hours are
01:24 iggy right
01:24 iggy (although I think aws can do schedules now too)
01:24 bhosmer_ joined #salt
01:25 iggy don't know much about azure
01:25 jessie_ i intend to use logentries going forward for hopefully logging all data from salt.
01:25 jessie_ azure documents very little and as to what they do document... it isnt always acurate
01:25 iggy *cough* salt *cough*
01:26 jessie_ so if i wanted to log errors from salt so i can actually check logs whilst working in CLI is this possible
01:26 timoguin joined #salt
01:26 jessie_ lol
01:26 iggy that's not really fair... salt has extensive documentation... so extensive that it's hard to digest
01:27 iggy salt has fairly flexible output systems (returners, etc.)
01:28 jessie_ yes agreed with that... timeconsuming read output...
01:28 jessie_ is there any info around what logs i can purge
01:31 jessie_ anyone for getting all the salt logs ? just need to know all there locations
01:32 iggy in /var/log/salt/
01:32 iggy the job cache is in /var/cache/salt/master/jobs/
01:32 iggy but it automatically cleans out jobs older than 24 hours
01:35 iggy LOG EVERYTHING!!!!
01:35 jessie_ thats the plan sir
01:37 iggy that is proving to be difficult to beat into the devs here
01:37 * iggy home
01:38 jessie_ thats because they arnt ops
01:38 jessie_ https://media1.giphy.com/media/dbtDDSvWErdf2/200_s.gif
01:38 jessie_ end game bieng http://31.media.tumblr.com/tumblr_lymy7o9lKC1qcyb0lo1_500.gif
01:48 ilbot3 joined #salt
01:48 Topic for #salt is now Welcome to #salt | 2014.7.2 is the latest | Please use https://gist.github.com for code, don't paste directly into the channel | Please be patient when asking questions as we are volunteers and may not have immediate answers | Channel logs are available at http://irclog.perlgeek.de/salt/
01:49 chandankumar joined #salt
01:51 cmcmacken joined #salt
01:54 sunkist joined #salt
02:00 bougie2 joined #salt
02:03 sunkist1 joined #salt
02:05 Hell_FireW joined #salt
02:07 Vynce sometimes when i salt-run manage.status it lists a server as down,a nd i can find no reason for it, and i start to panic, and i run it again, and it shows that same server as up. presumably there's some sort of timeout? is there a config i can twiddle to be more patient?
02:09 sunkist joined #salt
02:13 subsignal joined #salt
02:14 sunkist1 joined #salt
02:18 sunkist joined #salt
02:19 dude051 joined #salt
02:20 fusionx86 joined #salt
02:21 ITChap joined #salt
02:22 CeBe1 joined #salt
02:26 evle joined #salt
02:26 fusionx86 joined #salt
02:27 iggy jessie_: the job cache is all master stuff (scheduled jobs, events, highstates, etc) salt-cloud is going to be mostly /var/log/salt/ and a small amount of events
02:27 favadi joined #salt
02:27 timoguin joined #salt
02:27 iggy Vynce: what version? what's your network look like?
02:28 jessie_ iggy:  thanks for the jab cache stuff... trying to dive more into cloud logging... i dont believe its small but you and this page is correct http://salt-cloud.readthedocs.org/en/latest/topics/releases/0.8.2.html#salt-cloud-logging
02:28 Vynce mmm… don't really know how to answer that second question.
02:29 Vynce salt 2014.7.1 (Helium)
02:29 jessie_ iggy: problem is there is no cloud file in /var/log/salt/
02:29 iggy jessie_: note: salt-cloud is now part of upstream salt... salt-cloud.rtd.org is outdated (but still valuable)
02:30 iggy jessie_: I want to say it logs to the minion log (as it's technically a Client) but it may be the master log
02:30 jessie_ so your saying that salt cloud is now apart of teh main build ?
02:31 donmichelangelo joined #salt
02:31 iggy Vynce: i.e. are you in a cloud? do you have minions on T1's spread all over the world? are you running it all through a $20 ebay linksys switch? etc?
02:31 iggy jessie_: correct
02:31 jessie_ meh
02:33 ITChap joined #salt
02:33 iggy https://github.com/saltstack/salt/tree/develop/salt/cloud
02:33 jessie_ iggy: salt 2014.7.0 (Helium)  currently... whats the latest build
02:33 jessie_ ohhhwwww
02:34 _JZ_ joined #salt
02:34 iggy the latest "release" is 2014.7.4
02:34 jessie_ oh yeah its apart of the build...
02:34 iggy I think it was merged in 2014.1 (or maybe that was salt-api...)
02:35 jessie_ a update will just mess it up i think
02:35 jessie_ i wonder what else is apart of the build
02:35 sunkist joined #salt
02:36 murrdoc joined #salt
02:36 Vynce @iggy: all in one region at a cloud hoster
02:36 Vynce (rackspace)
02:36 elfixit joined #salt
02:37 iggy Vynce: okay, well, not really sure how they work, but a lot of cloud providers route _all_ traffic, and salt has a bug where it doesn't set the keepalive flag on the connection, so those routers have a tendency to drop the connection
02:37 jessie_ Vynce: ohhh rack space....
02:37 iggy well, the keepalive flag isn't set in one directio
02:38 iggy so it's weird, the master can still talk to the minion, but the return trip doesn't work (until the master sends something to the minion and it tries to return and resets the connection completely)
02:38 jessie_ Vynce: sgive em a call, they are all about supporting there own env... they have a devops dept also
02:38 iggy and a lot of them were at saltconf and were pretty rad
02:38 beneggett joined #salt
02:39 scbunn joined #salt
02:39 otter768 joined #salt
02:39 jessie_ shame the cost isnt so rad :)
02:40 Vynce i was hoping there was jsut a setting somewhere i could say, allow an extra half second or whatever.
02:40 jessie_ there probably is but it will be in python
02:40 iggy ^
02:41 iggy I wonder if that keepalive fix will make it's way into 2014.7 (the base of pi seemed to imply no)
02:42 jessie_ i reckon its to do with rackpsace..
02:42 jessie_ they will be able to resolve it
02:43 jessie_ considering salt in docker... may make my life alot easier with dependancies...
02:44 jessie_ that is if i am having to rebuild.... as i know an update will mess things up
02:46 rocket joined #salt
02:46 iggy it doesn't have that many deps
02:49 jessie_ where you from iggy
02:51 iggy Houston(ish)
02:53 jessie_ manchester uk , i thought texas was for cowboys ?
02:54 jessie_ dry humour :)
02:54 iggy steers and queers
02:55 jessie_ oh my
02:55 iggy Full Metal Jacket quote for those not aware
02:56 jessie_ found some syntax :) http://salt-cloud.readthedocs.org/en/latest/ref/cli/salt-cloud.html
02:56 iggy before anybody goes off thinking I'm a homophobic asshole... nope, just an asshole with movie quotes for days
02:56 jessie_ if there was quotes i woulda guessed it.
02:56 jessie_ quotes as in "" :)
02:56 ITChap joined #salt
02:56 badon joined #salt
02:58 iggy yeah, maybe next time :)
03:00 jessie_ for example : "Frankly, my dear, I don't give a damn."  :)
03:03 jessie_ think am done for the night
03:03 jessie_ thanks again iggy
03:06 ITChap joined #salt
03:08 mosen joined #salt
03:14 murrdoc joined #salt
03:22 pdayton joined #salt
03:22 mpanetta joined #salt
03:22 aw110f joined #salt
03:23 fusionx86 joined #salt
03:26 aw110f_ joined #salt
03:28 thayne_ joined #salt
03:30 iggy yw
03:32 iggy anybody have a windows minion handy?
03:37 mosen yep
03:40 iggy does it have a c:\python27 dir? (installed by salt, not something else)
03:48 fusionx86 joined #salt
03:51 mosen one sec
03:52 mosen iggy: directory doesnt exist
03:52 iggy thanks
03:53 mosen salt installs some kinda portable python for itself i think
03:53 mosen i was making a package for osx so i read through some of the windows packaging
03:53 pcdummy joined #salt
03:53 pcdummy joined #salt
03:53 iggy I'm getting chewed out and being made to look like some kind of asshole when some guy is asking questions about python and he's in c:\python27 (which I was pretty sure salt didn't use, just wanted to verify)
03:54 iggy on the salt issue tracker
03:54 cedwards joined #salt
03:56 llua iggy: why don't you idle in #kvm anymore?
03:56 iggy I don't use kvm anymore
03:57 iggy and I traded in one thankless job for another one
03:57 llua :{
03:58 mosen crap
03:58 rocket lol iggy isnt that the way of life .. :/
03:59 rocket i have been out of the systems game for 5 years now doing dev work and now I am trying to play catch up :/
04:00 iggy I mean _this_ one (my irl job is awesome... if I could find someone else to hire)
04:01 seev windows
04:01 seev palm meets face
04:05 catpig joined #salt
04:08 aravind joined #salt
04:12 catpigger joined #salt
04:14 beneggett joined #salt
04:19 vschum1 joined #salt
04:20 catpiggest joined #salt
04:23 mpanetta joined #salt
04:24 sunkist1 joined #salt
04:25 catpigger joined #salt
04:29 sunkist joined #salt
04:30 stoogenmeyer joined #salt
04:31 catpigger joined #salt
04:34 sunkist1 joined #salt
04:34 catpiggest joined #salt
04:38 CheKoLyN joined #salt
04:39 Furao joined #salt
04:39 sunkist joined #salt
04:43 iromli joined #salt
04:44 sunkist1 joined #salt
04:44 akitada joined #salt
04:45 catpigger joined #salt
04:46 jessie_ joined #salt
04:46 jessie_ iggy: yo
04:46 chandankumar joined #salt
04:47 jessie_ trying to find syntax for some commands and its becoming a serious ball ache
04:47 jessie_ does anyone know how to check the status of all th servers through commandline
04:48 jessie_ like a general status of current servers deployed under salt cloud
04:49 Furao salt-key -L or implement use salt mine to send some data back to the master
04:49 sunkist joined #salt
04:49 fusionx86 joined #salt
04:50 spookah joined #salt
04:50 spookah joined #salt
04:50 beneggett joined #salt
04:51 iggy salt '*' test.ping
04:51 iggy or salt-run manage.status
04:52 jessie_ thats what am looking for
04:52 jessie_ manage.status
04:52 jessie_ serious ballache, took me 20 mins..
04:53 Singularo joined #salt
04:55 echo joined #salt
04:56 fusionx86 joined #salt
05:02 catpiggest joined #salt
05:04 echo Anyone have time for a quick question about providers?
05:06 catpigger joined #salt
05:07 Vynce jessie_: that's my biggest complaint about salt.  the documentation is terribly organized, and there's no good daily ops faq.
05:09 jessie_ like a massive puzzle, or speaking to any of my ex's,
05:10 jessie_ command for debug....    -l debug ?
05:10 echo I haven't found the docs that unpleasant, they seem like pretty usable python docs
05:10 catpiggest joined #salt
05:10 rdas joined #salt
05:11 jessie_ echo must be salt staff in disguise !
05:11 echo HA
05:12 echo Having dabbled in puppet and chef previously, salt seems painless by comparison
05:12 jessie_ i disagree, i believe its more the opasite...
05:13 jessie_ ignore that, its 6 am here
05:13 echo Huh, well the others are certainly more popular, so perhaps that is a commonly shared feeling
05:14 Hell_Fire joined #salt
05:14 jessie_ chef tends to be used more i have noticed
05:14 jessie_ everytime i see a dude cookin food, i have a resounding hate for him until he passes me apple pie !
05:14 echo I suspect that is because they came strong early on with "enterprise" support. I see salt getting traction for cloud ops
05:15 jessie_ whether it will stick as another matter, especially with some new tools coming out
05:16 otter768 joined #salt
05:16 echo What tools are you thinking of?
05:16 echo I briefly considered Ansible, or rolling something with Fabric, then quickly came to my senses.
05:17 jessie_ this shiz blows all other devops out the water.... based on what i have seen so far that is.
05:17 jessie_ ansible tower is alright... nice gui but thats it... plus it bloody costs
05:17 echo I'll be damned if I am going to write "playbooks".
05:18 jessie_ even azure has playbooks... so annoying
05:18 echo 6AM, what is that England?
05:18 jessie_ uk manchester yes
05:18 echo I see, used to live in Windsor as a kid for a few years, in Chicago in the US now
05:19 jessie_ i see you made the correct choice
05:19 jessie_ saying that devops is slowly hitting the uk now... but more contract based... £350 per day
05:20 echo Mixed feelings about that, but probably. That would be a much more expensive service here than 350 a day
05:21 echo I think about 108 pounds per hour
05:21 echo In Chicago
05:21 jessie_ ahh thats really good
05:21 jessie_ its 350 to 500 per day in london, but i aint going there
05:21 echo Well, I am not on the receiving end of that, would be on the paying end.
05:21 echo Or, in my case, the not paying end.
05:22 jessie_ ahhh your giving the money to a devoper ... hmm
05:22 jessie_ deal time... 300 dolla per day :) place to stay and am yours sweek cheeks lol
05:22 echo hahah
05:23 jessie_ if it was new york, it woulda been 200 dolla
05:23 echo I see you have some preference of the US?
05:23 echo I see you have some preference for the US?
05:24 jessie_ i went down the east coast, found miami and ny been my fav. i havnt been inland
05:24 jessie_ preference bieng, that anwhere is better than the uk
05:25 echo well, Miami certainly has an edge concerning weather
05:25 echo I personally dislike living in chicago quite a bit
05:25 jessie_ ohhh ya, loved it there at daytona !!!!
05:26 jessie_ i stayed there for two weeks, had to be done.
05:26 echo So, are you using salt in production?
05:26 ajw0100 joined #salt
05:26 echo hehe, sounds nice
05:27 jessie_ i will be shortly. between 3 cloud providers for site to site HA
05:27 jessie_ but from what i can see, devops has a expiration date following whats newly bieng developed
05:28 jessie_ secondly, by the time devops starts to support windows it will be too late. something else will be here
05:28 echo I see. You building all your own formulas?
05:29 jessie_ yeah, a few templates i can use, but i have been looking more into what is better and the reasons why... against chef and puppet and ansible etc
05:29 echo Yeah, I suppose that could be the case. We have a windows department, but they appear to be in the stone age by comparison to current standards
05:30 jessie_ everything is easier in linux, and soon .net will hit linux... and when docker wakes up, it will help more around that
05:31 echo yeah, open .net is an interesting prospect, though, I can't see the benefit personally, I would prefer to use already established stacks on linux: python for ease, GO for speed
05:32 echo I am also moving away from linux a bit and looking to get better performance using Zones, in OpenSolaris spin-offs
05:32 jessie_ application driven infrastructure will respolve all issues completly, similar to what docker is doing... but on a infrastructure level
05:33 echo The more hetero the environment, the more config management appeals
05:33 jessie_ i think it will fade :)
05:34 echo Well, in many ways I hope it does
05:34 jessie_ ok sm done eith riddles, have a look at dcos
05:35 echo yep
05:37 jessie_ thoughts ?
05:40 stanchan joined #salt
05:41 echo I may have misunderstood, thoughts on what?
05:41 jessie_ dcos
05:42 malinoff joined #salt
05:42 echo oh ok, looking
05:42 jessie_ https://www.youtube.com/watch?v=UgJMlHdZEx4
05:43 echo Yeah, I am familiar with this new breed of clustering
05:43 echo Mesos, CoreOS, Kubernetes
05:43 Terminus- joined #salt
05:44 echo I have mixed feelings about it. On the one hand, it seems very clean, programmable, sexy
05:44 echo On the other, I have lots of apps that are quite legacy, we don't have the resources to rebuild them in a microservice structure, they can't really be clustered properly
05:45 jessie_ very sexy :) ebay is using kubernetes. but dcos defo the way forward
05:45 echo I suspect many people would be in a similar situation, so containers don't solve a lot of problems I need them tto
05:45 jessie_ cant wait for dcos
05:46 jessie_ alot of money around that.
05:47 echo does look quite nice
05:47 jessie_ the only time i got wood over software
05:47 jessie_ a ui monitor and cli in one !!!
05:48 echo heh
05:48 jessie_ semi = wood
05:48 jessie_ lol doesnt matter
05:49 beneggett joined #salt
05:51 echo heheh, semi lob
05:56 jessie_ there ye go lad
05:56 jessie_ semi lob for dcos
06:00 echo that demo was pretty cool. If my company is getting to devops now, I suspect we would get on board sometime shortly before my death from old age
06:00 echo ok, I have to have some drinks before bed
06:01 echo nice chat
06:01 stoogenmeyer joined #salt
06:01 colttt joined #salt
06:04 jessie_ ah no worries dude
06:04 jessie_ bye for now
06:04 dopesong joined #salt
06:06 flyboy joined #salt
06:06 hojgaard joined #salt
06:06 hasues joined #salt
06:07 hasues left #salt
06:09 cpowell joined #salt
06:14 JayFK joined #salt
06:15 AndreasLutro joined #salt
06:15 jhauser joined #salt
06:16 vstoniest joined #salt
06:17 catpigger joined #salt
06:22 Vynce anybody happen to know if jinja's whitespace-subtractors (e.g. (%- if foo -%} can be used inside expression blocks, {{- 'e.g. this' -}} ?
06:23 iggy Vynce: don't think so... but you can use {%- set short = .... -%}  .... {{ short }}
06:24 Vynce that makes no sense to me.
06:25 Vynce in … [whitespace1] {%- set short = ' value' -%} [whitespace2] {{ short }} [whitespace3]  … wouldn't whitespace3 be unaffected by that construction ?
06:26 Vynce maybe i should just be constructing this string in a longer iset of if blocks. I have to make a single string with no line breaks and i've got like 3 if blocks relevant to making it correct.
06:26 randomuser joined #salt
06:28 Vynce it hinkt eh problem is jsut that i'm a jerk who hates everything.
06:28 malinoff Vynce, double curly braces does not add any additional whitespaces AFAIK
06:28 iggy ^
06:29 malinoff Vynce, if you have them, your variable contains something you do not expect
06:29 iggy why don't you try pasting something vaguely similar to what you are actually trying to achieve
06:30 Vynce because i don't know what is even vaguely similar to what i'm trying to accomplish yet.
06:30 Vynce malinoff: i'm not talking about whitespace that gets added by {{ }} i'm talking about whitespace that exists around it in the file
06:31 malinoff Vynce, sorry, 'whitespaces around that'? it doesn't make sense
06:31 iggy you are trying to replicate something that exists already? using some data you already know of?
06:32 Vynce i don't have a question worth attention yet after "does {{- 'foo' -}} work" … i'll ask if i get one.
06:33 KermitTheFragger joined #salt
06:33 iggy {%- for i in list -%}{{ i }}{%- endfor -%}
06:33 iggy that would output all on one line
06:34 Vynce iggy: sure.  i don't have that.
06:35 malinoff Vynce, as you may notices from my first reply, it doesn't work because it makes no sense
06:35 iggy without more to go on, I'm just throwing out examples
06:35 malinoff s/notices/notice/
06:35 Vynce uh… it does to me, so one of us isn't communicating well.  I'm guessing it's me.  let me try again.
06:36 malinoff Vynce, {%- -%} stuff exists because {% adds an extra newline _before_ this block and %} adds an extra newline _after_ the block
06:37 Vynce "here's some {{ 'text' }}" and it hink that's the same as "here's some text" but i want "{{ 'port' -}}\n{{- 'manteau' }}" to be "portmanteau"
06:38 iggy you can't
06:39 Vynce OK. that was what i thought you said at the beginning.  i have not had a question worth asking since then.
06:39 Vynce but the question isn't nonsensical — i haev a use-case for it.
06:39 iggy we're just trying to help
06:39 Vynce i understand that, i'm saying there's no help possible at the moment, and being confused by malinoff's claim that my question doesn't make sense
06:40 malinoff Vynce, http://mywiki.wooledge.org/XyProblem
06:41 iggy it's hard to work with theory alone sometimes
06:41 iggy but hey, we're here if you come up with more questions
06:42 catpiggest joined #salt
06:42 Vynce malinoff: sure. but i don't know how to explain x, and so far all questions i've had about salt that started out as "how do i x?" have led to hours of pain.
06:42 malinoff Vynce, jinja2 was not intended to do such things. It was originally used for rendering HTML/XML
06:42 malinoff Vynce, try mako
06:42 malinoff Vynce, or even bare python renderer
06:43 Vynce malinoff: getting me to switch away from yaml/jinja at this time is  adead end previously explored.
06:43 malinoff Vynce, well, you simply can't do what you want to do with jinja
06:43 malinoff that's all
06:43 Vynce because the examples are all in yaml / jinja and poorly documented wrt what they actually mean.
06:43 chandankumar joined #salt
06:43 Vynce i disagree, but i really don't want to waste your time arguing about something that's not important to me
06:44 malinoff You can disagree, of course. But what you showed is not possible with jinja. Period.
06:47 Vynce int he language of your link, Y may not be possibe, but that doesn't mean X is not possible. and since X is what i want, Y not being possible does not mean that what i want is not possible.
06:48 Vynce my X, as stated earlier, is "I have to make a single string with no line breaks and i've got like 3 if blocks relevant to making it correct."
06:48 ITChap joined #salt
06:49 malinoff Vynce, show on pastie.org what's your input and what's the expected output
06:50 Vynce (and the problem with your XY problem link is that it's really XYZ, and when users ask for X, they're told "that's too general" so they ask "Z" and knowing the lines between X and Y and Y and Z requires exactly the expertise that they obviously don't have.)
06:50 Vynce malinoff: and that's the part that i already said i don't know yet, and ill get back to when i have.
06:51 iggy {% if foo %}{% set a = 'str1' %}{% endif %} {% if bar %} {% set b = 'str2' %} {% endif %}
06:51 iggy {{ a }} {{ b }}
06:53 Vynce iggy: that's basically what i meant when i said "maybe i should just be constructing this string in a longer iset of if blocks."
06:53 Vynce and is the approach i'm trying now
06:57 seev joined #salt
07:05 catpigger joined #salt
07:05 egil joined #salt
07:06 subsignal joined #salt
07:08 eseyman joined #salt
07:10 catpiggest joined #salt
07:14 dopesong_ joined #salt
07:14 hojgaard joined #salt
07:14 __gotcha joined #salt
07:14 trikke joined #salt
07:14 Romlok joined #salt
07:16 otter768 joined #salt
07:18 someonespace joined #salt
07:21 sl_ joined #salt
07:24 mpanetta joined #salt
07:25 davidbanham joined #salt
07:28 Hell_FireW joined #salt
07:28 beneggett joined #salt
07:28 Vynce hm.  putting test=True into my critical machine minion confs seems smart.
07:31 Vynce is there an explanation of how the top.sls file actually works that's not jsut a series of examples?  http://docs.saltstack.com/en/latest/ref/states/top.html is confusing.
07:35 wincyj joined #salt
07:40 zip_kid Goodmorning.
07:42 Auroch joined #salt
07:52 pdayton joined #salt
07:55 badon_ joined #salt
07:55 ktosiek joined #salt
07:58 badon joined #salt
07:59 Grokzen joined #salt
08:02 Vynce hm. {% set mylist = [] %} {% mylist.append('value') %} complains that mylist is an unknown tag. … the variable section implied that i could use arbitrary pythonic variables, though.
08:03 Terminus- joined #salt
08:03 Vynce ah.  need "do"… adn to load the expression statement extension.
08:03 Terminus- joined #salt
08:07 favadi joined #salt
08:07 Vynce is there a way to get jinja extensions in salt?
08:09 rjc joined #salt
08:11 Vynce ah, apparently that extension is already present
08:12 stoogenmeyer_ joined #salt
08:16 ckao joined #salt
08:21 Xevian joined #salt
08:25 jhauser joined #salt
08:29 Vynce also, actually, now that i'm far enough along to actually test it, {{- appears to do what i wanted.  @iggy for future reference for silly people who want to do silly things.
08:37 Arca joined #salt
08:38 Arca Hi everyone, is it me or the block {% else %} doesn't exist ?
08:39 Vynce in jinja?
08:39 mens_ it's probably you :-)
08:40 Vynce cf. http://jinja.pocoo.org/docs/dev/templates/#if
08:42 egil mens_: just wanted to let you know I solved my jenkins deploy just using the api, thanks for the hint
08:43 Arca Mmh indeed there is..
08:43 egil I tried the jenkins-saltstack-plugin, but seems it needs a valid ca chain for it to work
08:44 Arca I've this error : Jinja syntax error: unexpected end of template, expected 'block_end'.;    Just after my block if
08:44 Arca it's weird
08:45 zip_kid Arca: gist the lot
08:47 Arca zip_kid: I answer u on pm
08:47 JlRd joined #salt
08:47 zip_kid ... gist = gist.github.com.
08:48 hojgaard joined #salt
08:48 zip_kid irc sucks for reading code, also never paste multiple lines inirc.
08:48 zip_kid and you pasted it to zipkid, not zip_kid
08:48 Arca damnit
08:49 linjan joined #salt
08:49 Arca i'll check on gist then thanks :)
08:51 losh joined #salt
08:51 zip_kid stop PM ing...
08:52 zip_kid paste the link here... i don't claim to have the answer, i gave you the tip to show your code to all.
08:52 Arca https://gist.github.com/arca223/cc3b0e8fe53ff67be945
08:52 Arca ok sorry
08:52 zip_kid 1st char on line 7
08:52 Arca I just saw it
08:52 Arca God damn it
08:53 Arca I hate myself, thanks
08:53 KoFFiE those are the worst typo's :p
08:53 zip_kid never trust line numbers in errors, check the context.
08:54 Vynce1 joined #salt
08:54 KoFFiE yea jinga errors usually aren't that helpful in my (short) experience with salt
08:54 Arca ok thanks for the tip ;) what should I use instead of has_value, it said it doesnt exist
08:58 zip_kid Arca: 1st google hit on 'salt grains template' http://docs.saltstack.com/en/latest/topics/tutorials/states_pt3.html#using-grains-in-sls-modules
08:59 Terminus- Arca: that doesn't look right. why not salt['grains.has_value']('ip_interfaces:eth0')?
09:00 catpigger joined #salt
09:01 Arca zip_kid: Thanks, it works nice this way :) didn't expect I could do it without a function
09:01 Terminus- heh, i forgot there's a grains dict.
09:02 Arca Terminus-: your way works too, thanks
09:02 Arca I tried to see the different ways, it's quite interesting
09:07 Arca right way to do it is with has_value since ip_interfaces got 3 key:value under, otherwise the if always return false
09:08 N-Mi_ joined #salt
09:14 sl_ joined #salt
09:17 denys joined #salt
09:17 otter768 joined #salt
09:31 mdupont joined #salt
09:47 dRiN joined #salt
09:55 GnuLxUsr joined #salt
09:59 Auroch joined #salt
10:00 zip_kid is there a way to 'print' a variables value to debug out from an sls?
10:02 glyf joined #salt
10:03 Terminus- zip_kid: use the variable in a valid comment line in your config file?
10:04 Terminus- like "# this kills the crab: {{ foo }}"
10:05 baffle How does people store data for network.interface state files? Having separate state files for every minion seems kinda overkill, so I guess some kind of Pillar storage is the way to go?
10:06 baffle (Right now we use a pillar file to feed DHCP with static leases, but having had DHCP break a few times we want to move over to a static configuration)
10:07 Terminus- baffle: how did dhcp break for you? it's been working for me well, especially because i need to have my hostnames updated in bind.
10:07 zip_kid Terminus-: in an sls, not in a template....
10:10 badon joined #salt
10:10 Terminus- zip_kid: right. my bad.
10:10 zip_kid np :-)
10:12 zip_kid another question, i have a pillar var containing a dict and i want to render that to an array as yaml.... is there something that can output yaml in salt?
10:12 zip_kid s/array/template/
10:12 ITChap joined #salt
10:12 badon_ joined #salt
10:12 zip_kid ow, not even, s/template/file/
10:13 baffle Terminus-: Yeah, I have dhcp+bind working in harmony. But then you have an incident where your DHCP servers go down and everything looses network configuration.. And you cry. A lot.
10:13 Terminus- zip_kid: maybe http://docs.saltstack.com/en/latest/ref/output/all/salt.output.yaml_out.html#module-salt.output.yaml_out ?
10:13 Terminus- baffle: ah... that. failover.
10:14 baffle Terminus-: Both servers went down at once I think. Or failover didn't work. A bit unsure. :-(
10:16 baffle Basically the lease time was too low I think.
10:16 Terminus- baffle: yeah... the weakest link is the need to have the lease db synced. can't have independent dhcpd with different ranges when you need static addresses for servers.
10:18 Terminus- openbsd has their own implementation for lease sync. too bad their version of dhcpd doesn't do dynamic dns.
10:19 zip_kid Terminus... This time i need to 'output' into a file, not on CLI
10:19 baffle Terminus-: Right now Salt generates dhcpd configurations like this: http://pastebin.com/g7Ec6d76
10:20 Terminus- baffle: is that from dhcpd-formula? i'm doing a rewrite.
10:21 Terminus- my whitespace is a complete clusterfuck right now though. i'm still missing pools and deny/allow as well as key config and failover.
10:22 zerthimon joined #salt
10:22 bhosmer joined #salt
10:22 dfduran joined #salt
10:24 mens_ zip_kid: http://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.jinja.html#salt.utils.jinja.SerializerExtension
10:24 baffle Terminus-: I can't remember, to be honest. :) Here: http://pastebin.com/atx2VVtw
10:24 baffle Seems I made a "cidrconv" module, that does things like generate CIDR ranges and stuff.
10:24 baffle So, I guess not. :()
10:26 favadi joined #salt
10:26 AndreasLutro has anyone here installed salt via pip?
10:26 baffle Terminus-: http://pastebin.com/UJqWez6k is pillar data
10:27 baffle Ideally I would like to continue using DHCP, but I
10:27 baffle Ideally I would like to continue using DHCP, but I've been burned twice now.. So it's hard to trust it. :)
10:28 Terminus- baffle: i'm pretty much going all in with macros.
10:28 Terminus- baffle: http://paste.ofcode.org/pzWkFC5KnncWrAQPDhutWh
10:29 baffle Terminus-: I've understood that macros is what I should use, but I haven't started using them yet..
10:30 Terminus- baffle: i went with macros because host declarations for example can be in subnets, groups, etc. better to have one macro i can reuse in any section. =)
10:33 Terminus- anyway, time to go home.
10:39 __gotcha joined #salt
10:47 aquassaut joined #salt
10:55 giantlock joined #salt
10:57 denys joined #salt
10:59 evle joined #salt
11:04 chandankumar joined #salt
11:05 echo_ joined #salt
11:18 otter768 joined #salt
11:20 peters-tx joined #salt
11:25 mpanetta joined #salt
11:28 eth2 joined #salt
11:30 __gotcha joined #salt
11:34 catpiggest joined #salt
11:59 jerematic joined #salt
12:05 tedski joined #salt
12:12 AirOnSkin joined #salt
12:14 AirOnSkin The following conditional returns an error, if the list in question is completely empty: {% if 'rsyslog-master' in salt['pillar.get']('roles', []) %}
12:14 AirOnSkin Error: http://pastebin.com/7M1YfDn5
12:15 AirOnSkin What I don't understand is, the conditional works if anything (completely unrelated) is in the list... but if it's empty it doesn't work at all...
12:19 AndreasLutro AirOnSkin: can we see how the pillar sls that defines "roles" looks like?
12:20 mindscratch joined #salt
12:20 mindscratch How can enable logging or some sort of progress output while running salt?
12:20 mindscratch I have some states that take several minutes to run and I get no output until they complete
12:21 AndreasLutro mindscratch: -l debug ?
12:21 mindscratch i'll give it a shot
12:22 renoirb joined #salt
12:23 AirOnSkin AndreasLutro: Sure: http://pastebin.com/56QqFXwQ
12:23 AndreasLutro AirOnSkin: your problem is that this yaml is parsed into "roles: None", not "roles: []"
12:24 mindscratch +
12:24 AirOnSkin AndreasLutro, Why is that? I don't understand ...
12:25 AndreasLutro AirOnSkin: because you've commented out the 3 lines that indicate that it is a list
12:25 asdfasdffdsafdsa joined #salt
12:25 AirOnSkin AndreasLutro: Ah, I see. Is there a way to 'define' it as a list?
12:26 AirOnSkin Or to work around the 'empty' kind of list?
12:27 AndreasLutro AirOnSkin: try this https://bpaste.net/show/538c61629149
12:27 AndreasLutro you may need to remove the -'s
12:27 chandankumar joined #salt
12:28 AndreasLutro yeah you do
12:29 colttt joined #salt
12:29 AndreasLutro you'll also have to put commas at the end of each line (list item)
12:29 Romlok or just do `roles: []`
12:30 Romlok and remove the [] when you want to uncomment some of the actual roles
12:30 AndreasLutro ^
12:30 AirOnSkin I'll run some tests... be back in 30min (got a meeting now)
12:30 Romlok man, I'd love to have meetings that lasted only 30min
12:31 favadi joined #salt
12:37 bluenemo joined #salt
12:37 bluenemo joined #salt
12:39 mpanetta joined #salt
12:40 ndrei joined #salt
12:42 Tecnico1931 joined #salt
12:43 bhosmer_ joined #salt
12:48 numkem has to be a scrum of some kind, no meeting really last 30 minutes
12:48 catpigger joined #salt
12:51 favadi left #salt
12:54 glyf joined #salt
12:55 JDiPierro joined #salt
12:57 cpowell joined #salt
12:57 hawkmauk joined #salt
12:57 subsignal joined #salt
12:58 mindscratch joined #salt
12:58 hawkmauk left #salt
12:59 teskew joined #salt
12:59 mindscratch is it possible to execute a command (cmd.run), use the result in another command?
13:00 catpiggest joined #salt
13:01 Ligthert Maybe with bash <(stuff)
13:01 rockinwaffle joined #salt
13:01 rocket mindscratch: I am not sure this is the best way, but you could do this.  cmd.run and save the results into a grain .. and then have that command reload grains.
13:02 mindscratch i have to commands i want to run, first: calculate amt of time to sleep and sleep, second: query URL
13:03 __gotcha joined #salt
13:03 rockinwaffle left #salt
13:03 mindscratch i guess I don't need resutl actually
13:08 imaginator joined #salt
13:08 murrdoc joined #salt
13:08 amcorreia joined #salt
13:09 imaginator I'm still struggling to get the pillar_root working with git (salt_root works fine). Any ideas on what I am doing wrong in https://github.com/buddycloud/buddycloud-vm/blob/master/saltstack/configs/minion.conf  ? (trying to run as a masterless minion)
13:09 cpowell joined #salt
13:10 imaginator basically ext_pillar: with   - git: master https://github.com/buddycloud/buddycloud-vm.git root=saltstack/pillar doesn't seem to be picked up.
13:11 rocket I havent done anything with the root param  .. but maybe 'saltstack/pillar' ?
13:11 rocket I am guessing it might not like the path
13:13 niseak joined #salt
13:13 denys joined #salt
13:14 rocket imaginator: personally I dont want my pillars in with my states for security reasons ..
13:14 imaginator Rocket - yeah - I can totally seperate them but I'm just using the same repo for them to test this working and then will untangle the two.
13:14 rocket imaginator: unless your state repo is very tiny and just has a top file in it, for example, and the rest is broken out into formulas
13:15 imaginator I double checked the repo: https://github.com/buddycloud/buddycloud-vm/tree/master/saltstack/pillar is defintely there and was valid when I was running it in file-based mode.
13:15 cpowell joined #salt
13:16 rocket check the /var/log/salt/master log as well
13:17 ujjain joined #salt
13:17 Ligthert If you have to.. set the salt master to do some more logging of whats happening. It helps a lot.
13:18 imaginator Basically it runs though but every time it's missing pillar data (eg postgres commands) it pauses waiting for kbd input. Checking that in masterless minion I should be running with "salt-call  --local state.highstate" ?
13:18 imaginator Just to be clear I'm trying to run as a masterless minion.
13:19 perfectsine joined #salt
13:19 otter768 joined #salt
13:19 rocket imaginator: I pm'd you my script I am using to bootstrap my master ..
13:20 rocket try salt-call --local state.highstate -l debug
13:21 imaginator rocket: thanks trying now.
13:21 murrdoc left #salt
13:22 dyasny joined #salt
13:22 murrdoc joined #salt
13:22 rocket imaginator: I assume you can see the part where I am running the salt-call against my local config...
13:23 rocket imaginator: thats equivalent to what your looking to do .. in my case I did download my pillar and states locally first .. but you get the idea I think
13:23 fusionx86 joined #salt
13:26 rm_jorge joined #salt
13:30 AirOnSkin Romlok: It was more of a technical discussion about some changes to a new set of systems with a coworker ;)
13:31 Deevolution left #salt
13:31 rm_jorge joined #salt
13:32 Deevolution joined #salt
13:34 cmcmacken joined #salt
13:34 andreyst joined #salt
13:36 AirOnSkin AndreasLutro: Using [ ] with the lists worked perfectly. I was however able to leave the commented entries there: http://hastebin.com/ibobotivaz.vala
13:36 hasues joined #salt
13:36 hasues left #salt
13:37 jdesilet joined #salt
13:38 rocket is there a state that goes with this module?  tls.create_self_signed_cert  how would I find that out?
13:38 rocket I am trying to avoid a cmd.run salt-call --local tls.create_self_signed_cert
13:38 imaginator So I tried running the entire high.state and pushed the output (with -l debug) to http://paste.ubuntu.com/10712076/  (what's interesting is that there is no mention of pillars in the entire output).
13:40 AirOnSkin rocket: You can use the module state to run a module from a state: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.module.html
13:42 rocket thanks AirOnSkin
13:43 murrdoc if u trying to use module.run, might be better to setup a state file
13:43 murrdoc its trivially easy to do and its good form
13:46 JDiPierro joined #salt
13:47 jdesilet I had a question for the group. I'm trying to setup a state that will run a script only if a file isn't present. So I was thinking of trying a file.exists and then having the cmd.run kick off onfail. Would it be better to try and run a cmd.run with an unless type statement?
13:47 leszq joined #salt
13:48 murrdoc cmd.run has a creates param
13:49 murrdoc which says, this cmd creates this, so if it exists the cmd doesnt run
13:49 mens_ jdesilet: have a look at cmd.script with the only-if option or the creates option
13:50 jdesilet great, I will look into both. Thank you!
13:50 BET joined #salt
13:50 learning_ joined #salt
13:52 andrew_v joined #salt
13:53 MTecknology This is fun... developers -> git repo -> gitlab -> post-receive hook -> command -> salt_event -> salt master -> reactor -> minion -> highstate -> update repo
13:53 murrdoc :)
13:53 rocket murrdoc: I was just trying to setup things like in this doc.  http://docs.saltstack.com/en/latest/ref/netapi/all/salt.netapi.rest_cherrypy.html
13:54 murrdoc uh hmm
13:54 MTecknology Any time they push to one of two branches (prod or test) the server will automatically update the source leaving me completely out of the mix but also leaving them completely out of the server
13:54 murrdoc u have given them too much power
13:55 KoFFiE is there any 'best practices' guide for organising environments/hosts in the top.sls file?
13:55 KoFFiE I have quite a bit of servers, in almost as much environments, and I'm having trouble structuring them in a decent way in salt
13:56 MTecknology murrdoc: Why is that too much? They can't do anything beyond source code changes. If I did my job right, nothing in the source code can muck with the system.
13:56 MTecknology or a code execution bug..
13:56 murrdoc "but also leaving them completely out of the server"
13:56 KoFFiE problem is mostly that I have dozens of dev/test/prod and sometimes QA and client acceptance envs (depending on the client)
13:59 debian112 joined #salt
14:01 andrew_v joined #salt
14:01 rocket MTecknology: typically you want some sort of airgap between test/prod .. eg see this video from saltconf 2015 https://youtu.be/UalzJq40xMk
14:02 I3olle_ joined #salt
14:04 AirOnSkin KoFFiE: I don't thing something like that exists to fit "your" needs. There is however a Salt Best Practices page. You can also work with nodegroups or separate your hosts in logical environments via Pillar/Grains
14:04 AirOnSkin s/thing/think/
14:04 snuffychi joined #salt
14:04 rocket in this case the ci server is building a tagged deb that is installed by an operator/change control board operation
14:05 rocket MTecknology: your needs may be different if you dont need that level of change control
14:05 MTecknology My talk is up now!
14:05 snuffychi joined #salt
14:05 rocket MTecknology: what did you cover?
14:05 bmac2 joined #salt
14:06 ek6 joined #salt
14:06 pdayton joined #salt
14:06 MTecknology https://www.youtube.com/watch?v=JCKSLgtZ8fY
14:06 MTecknology They spelled my name wrong on there
14:06 murrdoc waaaa
14:06 snuffychi joined #salt
14:07 dfduran joined #salt
14:07 beneggett joined #salt
14:07 murrdoc holler at the base of pi
14:07 vieira joined #salt
14:08 StDiluted joined #salt
14:10 MTecknology rocket: ya, I've been forcing our devs to not re-create their own states and just use mine (when I give them states that they're allowed to mess with). In general, they only get a git repo and all it has is source code for the app that gets sucked down. On those boxes, I'm the one that maintains everything else.
14:10 * MTecknology is scared to see his video
14:10 MTecknology I look fat
14:11 vieira Hello, when the nrpe state runs on a given machine I would like to also run a state on the nagios "master" that adds that machine entry in hosts.cfg if it doesn't exist
14:12 vieira is this possible? is there any example in the documentation for something similar?
14:12 FeatherKing joined #salt
14:14 __gotcha1 joined #salt
14:14 jollyroger joined #salt
14:14 bhosmer joined #salt
14:17 Blooof joined #salt
14:18 iggy vieira: your best bet is going to be orchestrate or have the nrpe state fire a custom event that makes the reactor then fire off another state on the master
14:18 Blooof hello - can anyone help me with cmd.run command line arguments. I'm trying to run a script on a host as a specific user (which only exists on the remote host) : salt 'host' cmd.run 'id > /tmp/id.txt' user=www group=www
14:19 Blooof this odesn't work - id.txt has the id of root
14:19 Blooof am I missing something?
14:19 dude051 joined #salt
14:19 rhodgin joined #salt
14:19 iggy Blooof: so www:www doesn't exist on host? You can't do that
14:20 Blooof iggy - it exists on the remote server but not on the salt master. Does it have to also exist on the salt master? (i've used www as an example above)
14:20 Blooof actual command is : /usr/bin/salt --subset=1 'datasync0*' cmd.run 'id > /tmp/id.txt' user=owncloud
14:20 iggy Blooof: no, that stuff all gets resolved on the minion
14:21 Blooof iggy - good - any ideas why it's not working? I'm not 100% my syntax is correct
14:21 Blooof I obviously want to replace the id command with an actual script to run
14:21 Blooof but debugging at present
14:22 iggy it's runas
14:22 fredvd joined #salt
14:23 Blooof ahh!
14:23 Blooof # more id.txt
14:23 Blooof uid=1464009(owncloud)
14:23 Blooof thank you :)
14:24 iggy Blooof: on the state it's user though
14:24 iggy so that's the probably the confusion
14:24 Furao joined #salt
14:26 baseami joined #salt
14:27 baseami Can anyone assist with why my cron salt state says the identifier is not set, when i have an id in my config?
14:27 thayne_ joined #salt
14:27 baseami # SALT_CRON_IDENTIFIER:NO ID SET */15 * * * * /etc/cron.daily/logrotate
14:28 Blooof iggy - probably. You'r ebeing too kind though - I just can't read documentation is the real answer :)
14:28 baseami /etc/cron.daily/logrotate:   cron.present:     - indentifier: LOGROTATE     - user: root     - minute: '*/15'
14:28 rojem joined #salt
14:29 LtLefse indentifier
14:29 MTecknology This presentation is making me cringe. I talk way too slowly. Quite a few times I completely forgot what the crap I was talking about. It took me way too long to get into fun content. I look fat. I was waiving my hands around like an idiot. It is better than I expected, though.
14:29 murrdoc hahah
14:30 baseami lmao
14:30 baseami thank you so much!
14:30 baseami it has been a long night ;)
14:30 LtLefse np
14:30 StDiluted joined #salt
14:30 glyf joined #salt
14:31 linjan joined #salt
14:33 hasues joined #salt
14:33 hasues left #salt
14:35 perfectsine joined #salt
14:37 StDiluted joined #salt
14:37 blankspace joined #salt
14:37 ben_NN Hi everyone,
14:37 ben_NN Ive something strange with few jinja test
14:38 ben_NN here is my code http://pastebin.com/JnS6kZmQ
14:39 emptyspace joined #salt
14:39 ben_NN the first jinja test work but the second get an error "failed: Jinja variable 'dict object' has no attribute 'file.readlink '"
14:40 ben_NN if someone have an idea could be really helpfull
14:41 pdayton joined #salt
14:45 edrocks joined #salt
14:46 rocket so whats the best way to start developing formulas .. eg I want to see how my templates are rendered etc before going through all the work of rebuilding a machine etc?
14:46 rocket has someone written a guide I can read? or documentation somewhere to get some better workflow ideas?
14:48 tkharju joined #salt
14:50 ndrei joined #salt
14:50 chandankumar joined #salt
14:52 blankspace joined #salt
14:53 conan_the_destro joined #salt
14:56 Laserwhit joined #salt
14:58 Laserwhit HELSDFLSDKSLDKFJLSDKFJSDFKL;J
15:00 wnkz joined #salt
15:00 Laserwhit Is anyone here that can answer some simple questions?  Basically I am trying to install salt on a system that doesn't have access to the internet.  It is in a virtual network under a master computer.   I got the salt rpm (salt-0.16.4-2.1.3.noarch.rpm) and all of its dependencies installed.
15:00 Laserwhit Now where is salt?  I run which salt and see nothing.
15:01 Laserwhit from '/' I run find . -name salt and I get: ./etc/logrotate.d/salt ./etc/salt ./var/log/salt ./usr/share/doc/packages/salt ./usr/lib/python2.7/site-packages/salt
15:01 pdayton joined #salt
15:01 aparsons joined #salt
15:01 zip_kid Laserwhit: salt-master rpm?
15:02 Laserwhit Do you mean salt-minion?  I need this computer to be minion to the other computer.
15:03 Laserwhit So I need to install salt-minion.  That is a good start for me if that is the case.
15:03 racooper joined #salt
15:03 racooper joined #salt
15:04 Laserwhit I have found salt-minion in my rpm search and I will try to install.  Thanks.
15:06 Laserwhit I've only played with salt once before and I couldn't remember all of the tricks.
15:11 scbunn joined #salt
15:12 clintberry joined #salt
15:12 JDiPierro joined #salt
15:17 Auroch joined #salt
15:20 otter768 joined #salt
15:24 jsnl joined #salt
15:24 ALLmightySPIFF joined #salt
15:27 StDiluted joined #salt
15:29 wicope joined #salt
15:33 keimlink joined #salt
15:35 Kelsar joined #salt
15:36 beneggett joined #salt
15:38 denys joined #salt
15:40 thedodd joined #salt
15:40 ek6 joined #salt
15:46 MTecknology Does salt let you specify a custom location for an init script?
15:46 jcockhren keimlink: gist your vagrantfile
15:47 LtLefse Laserwhit: fyi a quicker way to find the files installed by an rpm package is "rpm -ql packagename" (e.g. rpm -ql salt-minion)
15:47 Ninja_ joined #salt
15:47 KoFFiE AirOnSkin: tx - I'll figure it out somehow :p (sorry for the late reply ^^)
15:47 LtLefse Laserwhit: also, that sounds like a pretty old version of Salt
15:49 JPaul good morning. anyone know what is wrong with my syntax in this statement, I swore it was working but now i'm finding it isn't: {% if grains['mem_total'] >= '32000' %}
15:49 XenophonF so i have this salt-cloud provider config (https://bpaste.net/show/e95c64a7c48f), but when i try to run something like "salt-cloud --list-images my-provider-virginia", it throws "KeyError: 'provider'" (https://bpaste.net/show/f82f54ac2dfe)
15:49 N-Mi_ joined #salt
15:50 XenophonF i'm not really sure what's wrong, to be honest
15:50 XenophonF i'm following the example given in the salt-cloud docs (http://docs.saltstack.com/en/latest/topics/cloud/config.html#extending-providers)
15:50 Laserwhit LtLefse, thanks for the input.  I'm limited by what I can see on the rpm sites for the given OS distro.  I suppose I can update all the libraries.  Does that usually work well?
15:51 XenophonF for reference i'm using salt-formula to handle configuration on the master
15:51 Laserwhit For salt specifically I mean?
15:52 Ninja_ Hi, why does the line  "deb http://archive.ubuntu.com/ubuntu {{grains['oscodename']}}-updates main" is eveluated to "deb http://archive.ubuntu.com/ubuntu trusty universe" on minions ? The minus sign '-update' seems to erase it''
15:54 chandankumar joined #salt
15:55 iggy ben_NN: what version of salt?
15:55 Norbell_ joined #salt
15:56 LtLefse Laserwhit: what distro?
15:56 iggy Laserwhit: note: 0.16 is way out of support, so I would try to find something newer
15:56 overyander joined #salt
15:57 overyander joined #salt
15:58 iggy Ninja_: I doubt that line is rendering that way... I would guess that's an existing file
15:58 jcockhren keimlink: how about your minion config? make sure it says: "file_client: local"
15:58 jcockhren otherwise, it'll expect a master
15:58 Ninja_ iggy: take for example this state:  http://pastebin.com/wE3Nkraf
15:58 thayne_ joined #salt
15:58 JPaul I figured mine out, needed to stip the ' ' off of the number i'm comparing to.
15:58 Ninja_ the files being purged before
15:59 Ninja_ this i can verify
15:59 overyander joined #salt
15:59 iggy Ninja_: try taking the -dist line out
15:59 Ninja_ ok
16:01 ajw0100 joined #salt
16:01 jcockhren remove the master key
16:02 jcockhren "master: localhost" tells it to talk to a master
16:03 Ninja_ iggy thank you this solves it, the docu says: "dist: This dictates the release of the distro the packages should be built for. (e.g. unstable)". I think its a little misleading
16:03 igorwidl joined #salt
16:04 igorwidl so, i noticed when salt-minion starts while the salt-master is down. It will try to authentcate once, then will just exit. Any idea how to have the minion retry forever instead of exiting?
16:05 jcockhren igorwidl: actually, I think it used to be that way some major versions ago
16:05 KyleG joined #salt
16:05 KyleG joined #salt
16:05 jcockhren I know*
16:06 iggy Ninja_: the problem is, the pkgrepo state module overrides some of the "name" arg... if you want to suggest some doc improvements, they'd be most welcome... I think a lot of us are just so used to how things work we don't even think about how the docs could be improved
16:06 jcockhren igorwidl: the issue at the time was that if a minion thinks the master is down, it'll forever spawn new suprocesses
16:07 RedundancyD joined #salt
16:07 __gotcha joined #salt
16:07 igorwidl jcockhren: alright thanks. It sounds like a bug though
16:09 jcockhren not sure of the intended behavior
16:10 iggy it is intended
16:10 igorwidl there should be a way to override that
16:10 iggy open an issue? (actually I think there's already one for it)
16:11 dopesong joined #salt
16:12 igorwidl i forgot to start salt-master after updating and restarting the server (for some reason the on start up was disabled).  Then I went ahead and updated 100 more servers and rebooted them
16:13 igorwidl thanks, i'll look
16:13 bhosmer joined #salt
16:14 writtenoff joined #salt
16:15 viq joined #salt
16:15 RedundancyD Is there a way to tell what version salt-bootstrap is going to install
16:16 rickflare so I am a newbie and I am trying to create my first state using multiple states. I current have my init.sls and then I have two other *.sls files that have include statments. However when i run salt it does not seem to ever call the non init.sls files. Can someone help me?
16:16 learning_ i realize this is no right answer for this question but are there particular well written/documented/best practice salt formulas I should model my own after?
16:19 smcquay joined #salt
16:20 iggy RedundancyD: not really... it's going to vary wildly based on what packages are in whatever repos you have setup (that's why I install a specific git tag)
16:20 rickflare anyone?
16:21 iggy learning_: salt-formula, collectd-formula, maybe others
16:21 RedundancyD That is why I now have a wild assortemnt of versions deployed. Guess I need to fix that
16:21 iggy those use the newer style that we are trying to adopt (where everything is merged in map.jinja and then looked up from that variable instead of having a bunch of direct pillar lookups all over the place)
16:22 learning_ iggy: thank you!
16:22 iggy rickflare: you have to assign states to minions (or include them in another)
16:22 Eureka_ joined #salt
16:22 huddy joined #salt
16:23 iggy rickflare: assigning a directory to a minion doesn't automatically include everything in that dir
16:24 perfectsine joined #salt
16:24 rickflare im a little confused iggy
16:25 rickflare so i know puppet and im learning salt
16:25 rickflare I though as long as you have includes
16:25 rickflare to the .sls files in the dir
16:25 rickflare they will be called
16:25 dalexander joined #salt
16:25 rickflare how do I ensure everthing in the dir is getting used?
16:26 mpanetta include them all in your init.sls
16:27 mpanetta statename/init.sls is the same as statename.sls  So when you call the state it only looks at init.sls, unless you include the other state files in the statename/init.sls
16:27 rickflare ahhhh
16:27 rickflare ok
16:27 rickflare so its the same as the puppet init
16:27 rickflare ok
16:27 rickflare thank you folks
16:27 mpanetta np
16:28 mpanetta Now, you can always manually call the other states
16:28 mpanetta ex: you have statename/foo.sls
16:28 mpanetta you can call that by calling statename.foo
16:28 spookah joined #salt
16:29 rickflare so in my includes
16:29 rickflare do I need to do like openssh.config
16:29 rickflare or can I just do config
16:29 rickflare and it knows its in openssh
16:29 mpanetta you should do openssh.config
16:29 rickflare ok
16:29 mpanetta It does not know (don't quite know why)
16:30 iggy because then every init would be the same
16:30 rickflare ok
16:30 rickflare i got it to work
16:30 mpanetta Ah
16:30 rickflare so while im hear
16:30 rickflare here
16:30 davet joined #salt
16:30 rickflare do pillars work like hiera?
16:31 iggy you're assuming we know what hiera is
16:31 rickflare ok
16:31 rickflare what is the best syntax to use when doing a pkg install that has multiple packages
16:32 iggy - names
16:33 iggy - pkgs
16:33 iggy names is for something else
16:33 ekristen joined #salt
16:35 conan_the_destro joined #salt
16:35 rgarcia_ joined #salt
16:38 baseami I have a salt state that installs python-boto, python-boto is successfully installed, but Salt reports it failed to install/update: python-boto
16:39 echo joined #salt
16:41 dendazen joined #salt
16:41 dendazen I have removed some package from packages/velio.release
16:41 dendazen But now i get an error
16:41 dendazen Specified SLS packages.velio-release in saltenv base is not available on the salt master
16:42 dendazen but I have removed ‘packages.velio-release’ from base
16:42 dendazen and even when i grep in states dir tree for ‘packages.velio-release’ it doesn’t find any policy mentioning it.
16:45 stanchan joined #salt
16:45 iggy gist top file, states, command being run, and the ouput
16:48 chandankumar joined #salt
16:51 dendazen Sorry, i have found one application role where it was used. Once removed everything started working.
16:51 dendazen Thanks.
16:53 beneggett joined #salt
16:53 forrest joined #salt
16:57 wm-bot4 joined #salt
16:57 ndrei joined #salt
16:58 racooper is there a salt state to manage authconfig on CentOS/RHEL?
16:59 jdesilet Is there a way in salt to copy a file from a minion to the master? My use case is that I have a cert I'm generating on a server other than the master. I then need to push this cert out to all the clients. I realize I could scp it up to the master but that means having to make sure all the security access is in place to scp a file over.
17:00 XenophonF anyone here using salt-cloud with ec2? i'm having trouble getting my provider config to work
17:00 XenophonF i keep getting this traceback - https://bpaste.net/show/f82f54ac2dfe
17:01 XenophonF i'm using this config - https://bpaste.net/show/e95c64a7c48f
17:01 jalbretsen joined #salt
17:01 rap424 joined #salt
17:02 bhosmer_ joined #salt
17:03 nickdew joined #salt
17:04 wnkz joined #salt
17:04 thayne_ joined #salt
17:10 wincyj joined #salt
17:17 notnotpeter joined #salt
17:18 racooper Hiya. just wondering if Salt 2014.7.2 is going to be released to EPEL prod . Looks like it's been cleared as of Mar 3 but terminalmage hasn't pushed it yet. any reason why it's still pending?
17:18 funzo joined #salt
17:21 otter768 joined #salt
17:22 chandankumar joined #salt
17:23 rojem joined #salt
17:24 whytewolf XenophonF: have you tried dropping the - on provider and extends? i don't think it is needed there
17:24 racooper also: is there any example for a CentOS pkgrepo.managed state? I can't find one for CentOS/RHEL, just for debian
17:28 whytewolf XenophonF: sorry i ment on the extends not on the prodiver as you look to be trying to use the newer split config style
17:29 robawt how do I read pillar data from a Salt runner?
17:29 XenophonF whytewolf: i've tried it both ways
17:29 robawt my pillar contains machines, I'm looking to read those machines and do something with it
17:29 robawt eyeballing salt.client.LocalClient but it might be over doing it
17:30 whytewolf wait try changing extends to provder.
17:32 XenophonF same error
17:32 XenophonF maybe i should revert to using a list?
17:32 XenophonF but still using "provider"?
17:33 whytewolf i would. extends in recent examples seems to have been phased out
17:33 XenophonF i'm going to have to rtfs, i think
17:34 iggy robawt: you probably shouldn't... runners run in the master context (not in the minion context)
17:34 wendall911 joined #salt
17:34 f4lse joined #salt
17:37 whytewolf I'm wondering if naming it my-provider might also be throwing a wrench in the mix.
17:39 whytewolf XenophonF: this is what the instructions imply should work https://gist.github.com/anonymous/77b681daadec033b5483 however i have my doubts
17:39 jwaibel joined #salt
17:39 mpanetta joined #salt
17:42 racooper I'm still having issues getting this to work: https://gist.github.com/racooper/15631d1c1f1c2abf6627 on 2014.7.1 from EPEL. Any suggestions? I'm pretty much out of ideas.
17:44 spookah racooper: it looks like "repo ocss was not found in /etc/yum.repos.d"
17:44 _JZ_ joined #salt
17:44 wnkz joined #salt
17:45 racooper yes, but shouldn't pkgrepo.managed create the repo if it doesn't exist?
17:46 racooper otherwise, what's the point of the function, if I have to create the repo manually I might as well used file.managed instead.
17:47 beneggett joined #salt
17:49 relidy racooper: The most I've done with pkgrepo.managed so far is this: https://gist.github.com/rhoths/dc57019439d0b044f241
17:49 relidy racooper: All that does is make sure a few edits to the repo stay in place. I actually "create" the repo by installing the RPM earlier.
17:50 robawt iggy: that i understand, but i'm looking for a list of machines stored in pillar, something that the master would have access to
17:50 racooper I see.  in this case it's a custom local repo that doesn't have an RPM, although I suppose I could build one for it (and probably should anyway)
17:50 iggy robawt: but the master doesn't have access to that (at least not in a consumable way)
17:50 iggy robawt: pillars are rendered and assigned to minions
17:51 zwi joined #salt
17:51 robawt correct, but you can specify a host against the cached.pillar runner and get the already rendered (and cached) data
17:51 whytewolf racooper: most i put in my pkgrepo.managed is humanname, baseurl and gpgkey i would say drop the name
17:52 Grokzen joined #salt
17:52 racooper I've tried with and without name, doesn't make any difference
17:58 timoguin joined #salt
17:58 robawt iggy: i ran it manually in my master with success, maybe i found something i shouldn't have
17:59 jwaibel hi there
18:00 iggy robawt: yeah, it's one of those things (like some grains) that runs now, but it's not really supposed to, and there's no telling what will happen in the future
18:00 robawt iggy: like most fun stuff I guess
18:00 baweaver joined #salt
18:01 jwaibel is there a way to set up environment sets with a single master?
18:01 robawt the way I'm seeing my problem is I have the hosts living in Pillar, as well as being targetable in Salt.  So i could also just use the localClient or salt caller and execute the same thing I guess
18:01 robawt turning into a soup sandwhich
18:01 Deevolution robawt: Coming into this a bit late, but for cached Pillar data, we setup a pillar>mine input.
18:02 robawt Deevolution: is salt mine automatically configured?
18:02 Deevolution It exists, yes.
18:03 Deevolution You have to setup an input for your pillar data (or subset of it).
18:03 robawt Deevolution: interesting.  Pulling up the Salt mine doc now
18:03 beneggett joined #salt
18:04 Deevolution Something like this: https://gist.github.com/belvedere-trading/f34ff18548d2a36c83b9
18:04 robawt this seems to be going about the problem from the other direction, but it looks like it could still work
18:04 iggy putting pillar data into mine seems like a bad idea
18:04 Deevolution This solved two problems for us: 1) Dealing with nodes that are sometimes unavailable;  2) Speed we get the data back.
18:05 giantlock joined #salt
18:05 Deevolution iggy: why do you think so?
18:05 Deevolution I guess to be clearer, we're only putting certain pillar data into the mine, not all of it.
18:05 iggy ahh, that's an important distinction
18:06 robawt Deevolution: can you still retain pillar heirarchies?
18:06 robawt like prod: bleh; qa: bleh
18:06 murrdoc pillar data into mine sounds scary
18:06 murrdoc pillar data from mine sounds better
18:06 Deevolution robawt: yes
18:06 Deevolution murrdoc: why scary?
18:07 Deevolution Isn't the purpose of the mine more or less a last value cache?
18:07 murrdoc good way to look at it yeah
18:07 Deevolution We have a web front end we've built on top of our Salt implementation for admin purposes, this solves the two problems I stated above.
18:07 murrdoc i have thought of mines being a way to share grains between minions
18:07 murrdoc not much else
18:08 Deevolution It doesn't look particularly scary from our perspective and saves us the trouble of moving the data out of Salt (into Redis, etc...).
18:08 murrdoc an external data store would be better
18:08 murrdoc but u are correct
18:08 murrdoc scary is the wrong word
18:08 robawt Redis solves everything, trick question Deevolution
18:08 Deevolution robawt:  LOL
18:09 robawt everything but power outages and bad HDD ;)
18:09 murrdoc netflix'es frontend for redis clustering is super sweet
18:09 iggy I imagine murrdoc was thinking the same thing I was at first... shoving _all_ pillar data into mine is bad mkay
18:09 murrdoc uh hmm
18:10 XenophonF left #salt
18:10 mattmtl joined #salt
18:11 Deevolution I would agree, that ALL pillar data is probably a bad idea.
18:13 cheus Curious to know if anyone has a good solution for accessing pillar data from other languages (eg, ruby) without an interim storage mechanism (like a file).
18:13 jcockhren cheus: salt api?
18:14 jcockhren salt-api can call execution modules.
18:14 rickflare how does one use firewalld in a .sls
18:14 mpanetta You can also store pillar data in a DB
18:14 mpanetta mongodb is one.
18:14 iggy etcd!
18:14 mpanetta That one too :)
18:15 cheus jcockhren, I was thinking about that though I can't seem to find its docs.
18:15 mattmtl is there a way to use salt-ssh like pdsh such as: salt-ssh test_server[10-14] test.ping
18:15 rickflare if I want to allow https in my default zone of firewalld how would I do this in .sls
18:16 jcockhren cheus: http://docs.saltstack.com/en/latest/ref/netapi/all/salt.netapi.rest_cherrypy.html#module-salt.netapi.rest_cherrypy.app
18:16 jcockhren cheus: http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.pillar.html#module-salt.modules.pillar
18:16 jcockhren :shipit:
18:17 cheus Ahh. Renated to netapi
18:17 cheus renamed
18:18 rickflare anyone?
18:19 jcockhren rickflare: maybe use the cmd module?
18:19 badon_ joined #salt
18:19 jcockhren I use it for my states for ufw
18:21 rickflare cmd?
18:21 rickflare i am using centos 7
18:22 robawt cmd.run is a module to execute /bin/sh commands on the target host rickflare
18:22 robawt well run would be a function of cmd, but yeah same diff
18:22 jcockhren rickflare: http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.cmdmod.html#module-salt.modules.cmdmod
18:22 jcockhren AND
18:22 jcockhren rickflare: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.cmd.html#module-salt.states.cmd
18:23 jcockhren this is assuming firewalld has a CLI
18:23 iggy you are also making assumptions that we all know what firewalld is (I personally have nfc)
18:23 * robawt steals "nfc" for later
18:23 * Gareth trips robawt
18:23 * robawt dies in a fire ... it's all Gareth's fault
18:23 robawt howdy Sir
18:23 Gareth robawt: how goes? :)
18:24 smcquaid joined #salt
18:24 robawt goes well.  Trying to trick this runner into knowing about one of our environments so I can target specific machines
18:24 MatthewsFace joined #salt
18:25 Gareth nice
18:25 robawt yourself?
18:28 dopesong joined #salt
18:30 nickdew Perhaps this is a silly question, but can I run boot strap against a machine that doesn't have the minion installed? (I'm thinking of something similar to knife in Chef).
18:30 mattmtl without a roster file, is it possible to provide a hostname that could be resolved such as:  salt-ssh my_server_hostname test.ping  ?
18:30 Gareth robawt: scanning issues on github, looking for stuff to work on :)
18:31 glyf joined #salt
18:34 Hell_Fire joined #salt
18:34 evilrob joined #salt
18:34 writtenoff_ joined #salt
18:34 iggy nickdew: using salt-ssh probably (that's effectively what salt-cloud does)
18:34 catpigger joined #salt
18:34 iggy nickdew: in fact there's a salt-cloud provider to do that (saltify)
18:35 ajw0100 joined #salt
18:35 dopesong_ joined #salt
18:36 nickdew iggy: Sweet, thanks!
18:39 writtenoff_ joined #salt
18:41 druonysus joined #salt
18:44 ndrei joined #salt
18:44 denys joined #salt
18:48 stoogenmeyer_ joined #salt
18:53 theologian joined #salt
18:55 hackel joined #salt
18:58 MTecknology iggy: I'm not exactly enjoying rebuilding my gitlab server at the moment... I'm rebuilding it from scratch. I /think/ the last step now is to figure out which user needs the mysql2 ruby gem and how to get that installed
18:58 Vynce joined #salt
18:58 * iggy was lazy and used the omni installer
18:59 bhosmer joined #salt
18:59 MTecknology that's what I'm switching to now
18:59 MTecknology I need the mysql2 ruby gem, though
19:00 iggy I don't remember having to do anything like that with that installer
19:00 iggy but then again, that was like a year ago that I set all that up (or more)
19:01 MTecknology are you using a remote mysql database?
19:02 iggy I would imagine it was postgres (don't have access to it anymore)
19:02 MTecknology it'll build the postgres gems by default
19:03 iggy that might explain it too
19:07 niseak joined #salt
19:11 diegows joined #salt
19:22 otter768 joined #salt
19:23 cmcmacken joined #salt
19:24 msheiny joined #salt
19:24 MTecknology iggy: It's starting to look like gitlab-ce dropped mysql support :S
19:26 catpiggest joined #salt
19:26 MTecknology yup... if you do the omni install, mysql support has gone away. :S
19:27 dopesong joined #salt
19:27 hasues joined #salt
19:27 hasues left #salt
19:29 MTecknology https://gitlab.com/gitlab-org/omnibus-gitlab/issues/485
19:29 MTecknology *grumble*
19:32 stoogenmeyer_ there's a cute feature where you can try getting a value from nested dicts using .get("key1:key2:key3"), does anyone know of anyting similar in native python ?
19:33 niseak joined #salt
19:33 iggy stoogenmeyer_: nope, but the code in salt would be pretty easy to repro
19:34 learning_ I'm sure this has been asked before, but is there any way to see what the master thinks the real-time gitfs virtual file system tree looks like? For example /srv/salt/development/iptables maps to (https://github.com/adequatedeveloper/iptables-formula development branch). I suppose that maybe the master doesn't really have this as a concept. Instead it just traverses the list until it finds a match
19:35 iggy learning_: cp.list_master
19:36 stoogenmeyer_ iggy: it's a simple to implement function, (http://pastebin.com/kiJvrh2T) just wondering if there's a builtin one already
19:37 iggy stoogenmeyer_: I've been doing python for a while now... 99% sure, no
19:37 niseak joined #salt
19:38 glyf joined #salt
19:39 linjan joined #salt
19:41 learning_ iggy: thank you once again. cp.list_master with saltenv param did just want I wanted
19:42 Vynce stoogenmeyer_:  istr perl had something like that built in that was super hyped when it came out but then people decided that as cool as it seemed for some use cases, it was a fundamentally flawed concept for a general-purpose language.
19:42 MindDrive Hmm... for custom modules, is the default location of them on the minion in /srv/salt/_modules (as per http://docs.saltstack.com/en/latest/ref/modules/ ), or am I misreading that?  Trying to access one via 'publish.{publish,full_data}', and it's not finding it.
19:43 _ale1_ joined #salt
19:45 niseak joined #salt
19:46 iggy MindDrive: that's the master location... on the minion, they'd be in /var/cache/salt/minion/ somewhere
19:46 MindDrive Oops, thanks, Iggy.
19:46 iggy MindDrive: and they only get sync'ed at specific times (highstate, sync_modules, sync_all, etc.)
19:47 niseak joined #salt
19:49 Deevolution Gareth: I'm the user who logged https://github.com/saltstack/salt/issues/20107 if you have any additional questions.
19:51 stoogenmeyer_ Vynce: huh? why would it be flawed? I mean it's a little ambiguous which key was missing but other then that..?
19:52 stoogenmeyer_ Vynce: also, what's istr perl (-:
19:52 dleedyna joined #salt
19:54 badon joined #salt
20:01 fusionx8_ joined #salt
20:01 Vynce "istr" — I seem to recall … perl — the language
20:02 Vynce stoogenmeyer_:  i don't recall, exactly. i haven't thought about that feature in ~15 years
20:02 iggy people recall perl?
20:03 Deevolution With some ire...
20:03 * Vynce is a perl advocate.  i still think more in perl than other languages.
20:03 Vynce but i'll spare you all teh rants, if you'll spare me the bigotry.
20:05 baweaver joined #salt
20:07 MindDrive Iggy: Got it to work (with a 'sync_modules'), thanks. :)
20:08 ajw0100 joined #salt
20:09 ek6 joined #salt
20:09 baweaver joined #salt
20:11 baweaver joined #salt
20:11 fusionx86 joined #salt
20:12 wincyj joined #salt
20:14 cpowell joined #salt
20:15 rap424 joined #salt
20:18 ndrei joined #salt
20:18 __ale__ joined #salt
20:21 MatthewsFace joined #salt
20:21 btorch this is probably a jinja quesiton but .. is it not possible to use salt[] call within an {% include " " %} ?
20:22 btorch I've been trying to do something like {% include "etc/sysctl.d/sysctl.conf."salt['grains.get']('systype','common') %}
20:22 btorch to work
20:24 _ale1_ joined #salt
20:24 btorch ah nm found my answer
20:27 wincyj joined #salt
20:27 iggy ~ inbetween
20:28 debian112 is there a way to test a grain before adding it to /srv/salt/_grains ?
20:28 Gareth Deevolution: hey :) We had met at SaltConf too.
20:28 debian112 like testing local on a minion
20:29 iggy debian112: a lot of grains you'll see something like https://github.com/saltstack/salt-contrib/blob/master/grains/gce.py (the __main__ bit at the bottom)
20:36 Deevolution Gareth: Actually that was my co-worker.  :)
20:36 Deevolution Shared account there.
20:36 debian112 iggy: ok so the only way to test is directly on the master
20:37 Deevolution I had intended to go to Salt-Conf, but it didn't work out this year.
20:37 iggy debian112: you can't really test grains on the master (unless it's also running a minion, but then it's no different than testing anywhere else)
20:38 debian112 so can I test a new created grain with salt-call grain?
20:38 debian112 without putting it on the master.
20:38 racooper are there any instructions or caveats when moving salt master from one server to another? it's going to have the same name and same version, just in a different building and subnet.
20:39 debian112 test it on a local minion
20:39 iggy debian112: should be able to
20:40 ajw0100 joined #salt
20:40 SheetiS racooper: Consistency with keys (/etc/salt/pki by default) will make your life easier.
20:41 racooper can I just copy the key directory over and be good?
20:42 SheetiS as long as the master and the minion on the master (assuming you run one) are not running, you should be fine.
20:42 racooper well, the minion is because I was loading the new server via highstate :p I can stop it now though.
20:43 hackel joined #salt
20:45 spookah joined #salt
20:45 baweaver joined #salt
20:46 amcorreia joined #salt
20:49 fusionx86 joined #salt
20:52 giantlock joined #salt
20:52 teskew joined #salt
20:54 niseak left #salt
20:56 kusams joined #salt
20:56 jonher937 racooper: Be careful. I believe the default setting when keys do not match is to stop the minion. So you might end up with all minions being stoped because of keys mismatching
20:57 jonher937 racooper: We had that when one of our salt masters was not connecting to the NFS area where we have our keys. Fun to restart all minions on all servers :)
20:57 arif-ali joined #salt
20:57 iggy that's what salt-ssh is for
20:58 fusionx86 joined #salt
20:58 ahammond I'd like to use the glob matches found in top.sls inside one of my states, can I do this with jinja?
20:58 kusams Need help with deleting the mine data of dead ec2 instances...the mine data is not getting cleared even after doing 'salt '*' mine.flush' and deleting the salt keys of dead instances...but am able to flush and update the running instances
20:59 iggy ahammond: sure... probably... I don't understand the question
20:59 ahammond something along the lines of {% if foo ~= some_*_things %}
20:59 iggy kusams: there's a ticket open about that
20:59 jonher937 iggy: In this case we had zabbix do it for us for most of the servers (our zabbix agent has permissions to restart salt directly from the webgui) I will have a look at salt-ssh after I finish reading the saltstack book. Thanks for the tip
21:00 catpigger joined #salt
21:00 iggy ahammond: maybe not
21:01 iggy ahammond: {% if 'word' in var %} {% if var.beginswith('word') %} etc
21:01 kusams iggy: can i have the ticket info...and can u please let me know if their any work around solution for this?
21:01 keimlink joined #salt
21:01 ahammond iggy yeah, that's what I figured. thanks!
21:04 catpiggest joined #salt
21:04 iggy kusams: I don't know it off the top of my head... delete some file in /var/cache/salt/master/ (it's mentioned in the issue)
21:06 kusams iggy: thx for the info.
21:08 murrdoc joined #salt
21:10 pdayton joined #salt
21:12 fusionx8_ joined #salt
21:14 fusionx__ joined #salt
21:15 penguinpowernz joined #salt
21:18 ajw0100 joined #salt
21:21 ndrei joined #salt
21:23 otter768 joined #salt
21:24 CeBe joined #salt
21:24 beneggett joined #salt
21:27 Diaoul joined #salt
21:30 tomh- joined #salt
21:40 Gareth Deevolution: Ahhh got it.
21:45 kunit joined #salt
21:47 baweaver joined #salt
21:47 murrdoc joined #salt
21:55 OGuilherme joined #salt
22:02 baweaver joined #salt
22:06 cmcmacken joined #salt
22:09 Singularo joined #salt
22:09 Seldaek joined #salt
22:10 jhauser joined #salt
22:11 Seldaek hey, for some reason I can not load formulas from saltstack-formulas on github using the gitfs provider, trying to use one in a masterless setup and getting "Specified SLS mysql in saltenv base is not available on the salt master"
22:11 Gareth Deevolution: ping
22:11 Seldaek now it's weird that it says not available on salt master, since there is no master, but I figure this might just be a reporting issue
22:11 Seldaek problem is I have no clue how to debug what I am doing wrong
22:12 ajw0100 joined #salt
22:12 Seldaek gitfs is enabled and it seems to pick up the remotes I configured in /etc/salt/minion
22:12 Seldaek anyone having a clue I'd appreciate it :)
22:14 rojem joined #salt
22:14 forrest joined #salt
22:15 admina joined #salt
22:16 mosen joined #salt
22:18 iggy I didn't think gitfs worked with masterless
22:19 Seldaek well https://github.com/saltstack/salt/issues/6660 says it does now
22:20 echo joined #salt
22:20 Seldaek the linked PR that apparently fixed it was merged in 2014.7.1 as far as I can tell
22:20 Seldaek and I'm running 2014.7.2
22:20 beneggett joined #salt
22:25 budman joined #salt
22:25 cotton joined #salt
22:27 iggy At the very least, I'd open an issue about the log message referring to the master
22:27 nesv joined #salt
22:28 Seldaek iggy: ok thanks
22:29 Seldaek I tried with --local as well just in case but same result (as it should since I have that set in the minion config)
22:29 budman_ joined #salt
22:31 codercotton left #salt
22:33 nesv Quick question regarding template inheritance: when the docs say "...the search path starts at the root of the state tree or pillar", does that mean it would start from /srv/salt, or /srv/salt/my_state?
22:35 iggy if /srv/salt and /srv/salt/my_state are both in file_roots, then both
22:35 nesv iggy: Okay, thank you. :)
22:36 beneggett joined #salt
22:37 nesv Just to clarify, if I wanted to extend the /srv/salt/haproxy/haproxy.cfg template, I would have to put `{% extends "haproxy/haproxy.cfg" %}` in the other template (assuming /srv/salt is the only entry in file_roots)?
22:40 kunit can I test what gitfs returns on salt-master?
22:43 zircote joined #salt
22:44 Vynce joined #salt
22:44 zircote joined #salt
22:48 aquassaut joined #salt
22:55 lumtnman joined #salt
22:56 timoguin joined #salt
22:58 ahammond is there a reasonable way to figure out what top.sls includes for a given machine inside a .sls file?
22:59 igorwidl left #salt
23:01 baweaver joined #salt
23:02 agliodbs joined #salt
23:07 baweaver joined #salt
23:10 Vynce joined #salt
23:10 agliodbs I'm thinking of using salt states to control failover for a simple 2-server PostgreSQL system.  Has anyone done this?  If so, how?
23:11 ToeSnacks left #salt
23:15 fredvd joined #salt
23:16 iggy nesv: not entirely sure what you are trying to do, but extends isn't actually a commonly used feature (a lot of people see it and think "oh cool", but it's not really)
23:16 iggy kunit: salt-run cp.list_master ?
23:17 iggy ahammond: salt 'foo' state.show_highstate ?
23:18 iggy agliodbs: like http://www.radialpoint.com/tech-blog/postgresql-replication-using-salt/
23:18 nesv iggy: I'm setting up a template for a haproxy config, but haproxy will be used on different boxes, for different purposes, and as far as I can tell, I cannot dynamically add/remove frontends, backends, and servers using haproxy's stats socket; I can only enable/disable them.
23:19 nesv iggy: So, a minion might get the "haproxy.foo" state, which would extend a base haproxy config template, adding in the frontends and backends it needs, which would be different than the frontends and backends for "haproxy.bar"
23:19 agliodbs iggy: kinda.  I'm looking at how to handle the actual failover component, though.  That is, how do I tell salt "ok, failover to the replica now" under urgent circumstances?
23:20 agliodbs that is, somethign other than editing the config and committing
23:22 agliodbs what I'm looking for, I guess, is a way to have a separate "failed over" profile for failover circumstances.
23:23 agliodbs and an easy way to switch to the "failed over" profile
23:23 iggy nesv: you could just as easily do that with a single template and just pass in empty/different lists for the front/backends
23:24 otter768 joined #salt
23:24 iggy nesv: I'm not saying extend is absolutely useless... it just doesn't do what some people seem to think
23:24 nesv iggy: I'll definitely try out your suggestion. What do some people seem to think it does?
23:24 kunit iggy:  Function 'cp.list_master' is unavailable
23:25 iggy agliodbs: I'd have a state/orchestrate file that did it and you just run that when necessary... automatic failover is basically frowned upon in the postgres world
23:25 iggy kunit: salt-run?
23:25 kunit yes
23:25 agliodbs iggy: yes, this is for manual failover.  I have a more complicated tool to use for automated failover
23:25 agliodbs state/orchestrate file?
23:26 kunit salt-run cp.list_master, salt-master-2014.7.1-1.el6.noarch
23:26 iggy kunit: nvm, it's a regular execution module.. so just 'salt minion cp.list_master' (which makes more sense)
23:26 iggy or salt-call
23:26 kunit kk, ty
23:27 iggy agliodbs: http://docs.saltstack.com/en/latest/topics/tutorials/states_pt5.html#orchestrate-runner
23:27 iggy amongst other things
23:27 kunit salt [target] cp.list_master works prefect, tyvm iggy
23:28 gerl1ng joined #salt
23:29 fredvd joined #salt
23:29 iggy glad that was what you were looking for
23:30 murrdoc ah iggy
23:30 murrdoc giver to the masses
23:30 jsnl joined #salt
23:31 agliodbs iggy: thanks.  still not quite picturing that, though.  Can an orchestrate runner change the state of salt itself?  That is, I'd like the salt config on completion to reflect the change to which server is master.
23:33 iggy agliodbs: what salt config?
23:33 agliodbs tbe actual actions I need to carry out are quite simple; basically, replacing a few config files and restarting some services
23:33 agliodbs iggy: the salt-master
23:34 agliodbs basically, there's effectively two configurations: one in which server A is master and server B is the replica, and one in which server B is the master and server A is offline.  I'd like to give the sysadmins an easy way to switch from the first configuration to the second, in a hurry.
23:34 iggy why do you need to change the salt-master config?
23:38 agliodbs iggy: well, I don't want to change the actual files, I want a way to reflect that the whole cluster is in a different state.   Except I can't use the word "state" because that has a specific meaning in salt, which is different than what I want to say.
23:38 iggy okay, I was confused
23:39 agliodbs iggy: and I feel like Salt *must* have some way to do this, but because I don't know what it's called, I'm not finding it in the docs
23:40 iggy so yeah, with an orchestrate job, you can say "X host should have postgres.master state applied, and Y host should have postgres.slave state applied"
23:40 robawt agliodbs: use a grain, and key off of the grain
23:40 iggy then you can have another orchestrate job that does it backward
23:41 iggy a grain could work, but I feel like it would be more work (login to X & Y servers, change grains file, restart minion, run highstate on both of them)
23:41 iggy or I guess newer versions of salt have modules to set grains
23:41 agliodbs iggy: yeah, that's out of the question.  this needs to be a single command
23:41 SheetiS1 joined #salt
23:42 agliodbs the grain that is
23:42 iggy still... not one command
23:43 agliodbs iggy: hmmm.  so if I want an orchestrate job to say "use this configuration file INSTEAD OF that one, which would normally be part of the highstate" ...
23:44 iggy the other bad thing with grains is you run the risk of someone messing up and not making sure one server had the primary grain and one had the slave grain... then you have 2 servers trying to be master
23:44 agliodbs iggy: I guess what I'm concerned about is that, if the failover orchestrate job has already run, and then someone executes a highstate on the former replica, its configuration files would revert to the original config.  I don't understand how to prevent that.
23:45 iggy don't put the config states in a highstate?
23:45 MTecknology basepi: ping?
23:46 agliodbs iggy: oh, you mean make all of the required configs part of orchestrate jobs?
23:46 iggy then it has to be a conscious effort on someone's part to always set one server as the primary and one as the salve
23:46 iggy *slave
23:46 basepi MTecknology: pong
23:46 baweaver joined #salt
23:46 agliodbs iggy: hmmm, yeah, that works.  I like that; then I can have salt even handle the replication setup, like in the example.
23:47 iggy agliodbs: once you start writing it up and testing, feel free to ask for more help or a second eye at what you have
23:47 agliodbs I'm sure I will, thanks
23:48 agliodbs plus I'll blog this if it works, so other people can use it
23:48 iggy it's something I've put thought into, but never actually implemented (yet)
23:48 iggy right now our master/slave config is manually setup, but we will have to automate it at some point
23:49 agliodbs yah
23:49 agliodbs iggy: fyi, this is mine: www.handyrep.org
23:49 agliodbs that's for automated failover
23:49 agliodbs but it's way overkill for this case
23:49 baweaver joined #salt
23:52 iggy I'm looking forward to bdr maturing a bit more (and hopefully getting upstreamed enough that it doesn't require a patched postgres)
23:52 agliodbs iggy: can orchestrate jobs call other orchestrate jobs?
23:52 agliodbs iggy: yeah, I'm not even seriously looking at it until 9.5
23:52 iggy orchestrate jobs can call states, but not orchestrates
23:53 agliodbs iggy: also, note that bdr will NEVER be as simple as binary replication.  It needs per-table config, at a minimum.
23:53 agliodbs iggy: right.
23:54 agliodbs iggy: where are you located?  I'm PG community, it would be fun to book you for a talk somewhere.
23:54 iggy we have 1 application and it's setup in such a way that I don't think we'd hit any of bdr's pain points (once it's matured a bit that is ;)
23:55 ajw0100 joined #salt
23:55 iggy Houston... but if you saw my saltconf talk, you wouldn't be so quick
23:55 theologian joined #salt
23:55 iggy ask murrdoc... he seent it
23:55 agliodbs hah!
23:55 agliodbs iggy: FYI, there's a PostgreSQL conference in Dallas in September
23:56 agliodbs call for presos is open
23:56 murrdoc iggy:  is awesome at giving talks
23:56 iggy I'll look into it, but I'll be at sun and bass most of sept
23:57 murrdoc hire iggy for presos
23:59 vieira joined #salt

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