Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-09-01

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

All times shown according to UTC.

Time Nick Message
00:00 ALLmightySPIFF joined #salt
00:02 ALLmightySPIFF joined #salt
00:03 jodv joined #salt
00:05 ALLmightySPIFF joined #salt
00:05 mosen ahh state output mixed is much better :) I have to make that the default state output
00:06 ALLmightySPIFF joined #salt
00:15 twork ...yeah, the stuff forrest linked to was stuff i've already read.
00:22 twork my issue is: given a configured formula (like, the users formula), and a 'users:' pillar that populates all the var:value pairs in that formula, how to make a category of minions that all use the same formula, but subdivide the pillar in different ways?
00:27 pdx6 Does anyone know why I would get [DEBUG   ] Could not LazyLoad virt.is_hyper?
00:28 pdx6 salt-call virt.is_hyper just returns Module 'virt' is not available.
00:28 joe_n joined #salt
00:38 drawsmcgraw joined #salt
00:39 cyborg-one joined #salt
00:44 homeshlice joined #salt
00:52 jodv joined #salt
00:57 jodv joined #salt
00:58 favadi joined #salt
00:59 bougie joined #salt
01:01 leev_ joined #salt
01:02 ahale joined #salt
01:02 baweaver joined #salt
01:02 homeshlice joined #salt
01:03 Hazelesque joined #salt
01:03 LeProvokateur joined #salt
01:03 Hazelesque joined #salt
01:04 cpowell joined #salt
01:12 jaybocc2 joined #salt
01:16 jodv joined #salt
01:19 laax joined #salt
01:21 tyson joined #salt
01:28 iggy twork: pillar targeting (and possibly some merging)
01:29 iggy pdx6: you need the libvirt python module on the minion
01:31 favadi joined #salt
01:31 catpigger joined #salt
01:36 tmclaugh[work] joined #salt
01:39 evle1 joined #salt
01:46 zsoftich joined #salt
01:52 forrest joined #salt
01:54 genediazjr joined #salt
02:00 mrmarv joined #salt
02:04 sunkist joined #salt
02:05 orion203 joined #salt
02:17 clintberry joined #salt
02:19 orion203 joined #salt
02:27 zwi joined #salt
02:30 SheetiS joined #salt
02:34 evle joined #salt
02:37 GreatSnoopy joined #salt
02:38 anotherZero joined #salt
02:40 hardwire joined #salt
03:04 MadsRC joined #salt
03:21 TyrfingMjolnir joined #salt
03:21 aristedes1 joined #salt
03:27 favadi joined #salt
03:27 murrdoc joined #salt
03:29 murrdoc joehh:  any eta on when the package will be in ppa for 2015.5.5
03:31 capricorn_1 joined #salt
03:35 capricorn_one joined #salt
03:36 genediaz_ joined #salt
03:38 mosen joined #salt
03:40 genediazjr joined #salt
03:49 ramteid joined #salt
03:56 pm90_ joined #salt
03:56 auzty joined #salt
04:01 joe_n joined #salt
04:03 TyrfingMjolnir joined #salt
04:10 aristedes joined #salt
04:10 aristedes left #salt
04:15 Zachary_DuBois joined #salt
04:28 rdas joined #salt
04:36 tyson joined #salt
04:37 clintberry joined #salt
04:40 zwi joined #salt
04:41 evle1 joined #salt
04:48 tyson joined #salt
04:49 jaybocc2 joined #salt
05:00 HoloIRCUser7 joined #salt
05:25 rdorgueil joined #salt
05:34 jhauser joined #salt
05:54 interisti joined #salt
06:01 ramteid joined #salt
06:01 colttt joined #salt
06:02 vexati0n joined #salt
06:05 pm90__ joined #salt
06:07 katyucha joined #salt
06:08 cprior joined #salt
06:12 zblum joined #salt
06:16 dopesong joined #salt
06:17 impi joined #salt
06:17 dopesong_ joined #salt
06:18 dopesong_ joined #salt
06:22 laax joined #salt
06:24 illern joined #salt
06:25 AndreasLutro joined #salt
06:28 vvoody joined #salt
06:29 elsmo joined #salt
06:34 aristedes joined #salt
06:38 sunkist joined #salt
06:39 aristedes left #salt
06:44 micko joined #salt
06:45 lb1a joined #salt
06:55 zblum joined #salt
06:58 eseyman joined #salt
06:59 stephanbuys joined #salt
07:01 jaybocc2 joined #salt
07:05 binocvlar joined #salt
07:07 kawa2014 joined #salt
07:10 KermitTheFragger joined #salt
07:14 Ztyx joined #salt
07:16 jaybocc2 joined #salt
07:18 pbrooko joined #salt
07:18 Ztyx joined #salt
07:20 Ztyx joined #salt
07:28 dRiN joined #salt
07:28 impi joined #salt
07:41 hojgaard joined #salt
07:45 PasiFist left #salt
07:45 hojgaard I am having trouble using the salt python api. It works well for everything except for one module: the host module. look here: http://pastebin.com/aH2ajxH5
07:47 AndreasLutro hojgaard: host vs hosts
07:47 hojgaard AndreasLutro, sorry the hosts module
07:48 hojgaard AndreasLutro, oh boy, that is embarrasing. Thanks for the quick feedback :D
07:48 AndreasLutro happens to me all the time
07:54 s_kunk joined #salt
08:00 stephanbuys1 joined #salt
08:01 Ztyx joined #salt
08:02 GreatSnoopy joined #salt
08:05 slav0nic joined #salt
08:05 zer0def joined #salt
08:06 impi joined #salt
08:08 pbrooko joined #salt
08:09 Grokzen joined #salt
08:11 ashmckenzie joined #salt
08:13 linjan joined #salt
08:16 losh joined #salt
08:16 markm joined #salt
08:17 Xevian joined #salt
08:18 dimduj joined #salt
08:18 illern_ joined #salt
08:22 aristedes joined #salt
08:23 aristedes left #salt
08:29 mikeywaites joined #salt
08:29 dkrae joined #salt
08:33 MartinSvensson joined #salt
08:33 Ztyx joined #salt
08:34 MartinSvensson Is it possible to use reactor files to execute scripts on the master?
08:36 iamtew joined #salt
08:57 ggoZ joined #salt
08:57 Ztyx1 joined #salt
08:58 Ztyx1 joined #salt
09:02 aparsons joined #salt
09:05 aparsons_ joined #salt
09:11 permalac joined #salt
09:28 jaybocc2 joined #salt
09:28 lothiraldan joined #salt
09:31 zsoftich joined #salt
09:32 keimlink joined #salt
09:34 eliasp joined #salt
09:34 keimlink joined #salt
09:43 jaybocc2 joined #salt
09:51 X67r joined #salt
09:53 genediazjr joined #salt
09:58 guest_ joined #salt
10:01 guest_ Hello. I'm learning to properly set up states in SaltStack, so I have been reading the "Best practices" doc (https://docs.saltstack.com/en/latest/topics/best_practices.html). There is one thing that I find confusing: In the "Modularity within states" section, some data is put into map.jinja while other data is put into pillar. Why is this? Why use one or the other, assuming the data isn't sensitive or secret in any way.
10:01 kawa2014 joined #salt
10:03 guest_ Specifically this example covers an apache server state. Everything is put into a map.jinja while the file path to the configuration file is alone in the pillar.
10:03 irctc581 joined #salt
10:03 irctc581 Hi Everyone
10:04 irctc581 I have read about  Orchestration
10:04 babilen guest_: This is a tricky topic, but let me elaborate. Formulas, or states, typically require three types of data: 1. Defaults for settings 2. Platform specific settings 3. Use configurable settings
10:05 aqua^c joined #salt
10:06 babilen guest_: map.jinja is being traditionally used for data that is platform specific (which is why you see the grains.filter_by call in there). This includes data such as the name of packages or locations of configuration or log files that differ between distributions, but are unlikely to be changed by users.
10:06 bastiandg joined #salt
10:07 irctc581 Currently I am creating separate enivronment for pillar and states for each components(webserver, load balancer ) and I can execute high state for each environment
10:07 babilen If users want change some of these some formulas/states introduce a :lookup: key in the pillar to denote that these settings override those in the "lookup table" (i.e. the map being defined in map.jinja)
10:07 irctc581 Is there any different doing like this and through server orchestration ?
10:07 irctc581 If yes what is the difference ?
10:08 CeBe joined #salt
10:09 babilen guest_: We have, in fact, gone further these days and typically do *all* data handling in map.jinja. This means that best-practices formulas such as, but not limited to, salt-formula, influxdb-formula or collectd-formula, define a foo_settings map (i.e. dictionary) in map.jinja that is being constructed by subsequently merging in os/distribution specific defaults, default settings and user defined pillar dat
10:09 babilen a into *one* datastructure.
10:11 babilen guest_: You can see this in https://github.com/saltstack-formulas/salt-formula/blob/master/salt/map.jinja which defined an os_map for distribution specific data (sometimes with subdevisions for different releases) in os_family_map, loads default settings from salt/defaults.yaml and finally merges in the user defined pillar.
10:12 irctc581_ joined #salt
10:13 irctc581_ AndreasLutro babilen ^^
10:14 bastiandg joined #salt
10:17 froztbyte bah
10:17 froztbyte I have redis problems
10:17 jeadre joined #salt
10:17 froztbyte specifically, I need to solve a "provision some redis nodes and sentinel" problem
10:17 froztbyte which means "bootstrap sentinel but then leave the config alone later"
10:18 froztbyte pretty sure this isn't actually directly salt's problem but it would be pretty nice if I could bootstrap the entire set upfrnt
10:19 froztbyte upfront*
10:21 lothiraldan joined #salt
10:23 babilen froztbyte: You could use different top.sls files for that or explicitly manage that file during the bootstrapping.
10:23 froztbyte mmmm
10:24 froztbyte okay so, let me start at the start here
10:24 froztbyte I have some CloudFormation
10:24 froztbyte with some special userdata sauce
10:24 monkey66 joined #salt
10:25 froztbyte I press a button in the console that says "gimme", and I get a bunch of instances out. n of each. reactor autosigning within the VPC, yada yada.
10:25 froztbyte from that, I have a bunch of redis machines
10:25 froztbyte any of them could, pretty much, be the first one
10:25 froztbyte so I could take one of 2 approaches
10:26 froztbyte either write an imperative type of "pin redis1 as a master, with a -creates file as run-once" type of thing
10:26 froztbyte or implement my own crappy master seeding by way of a mine, and then whichever instance comes up first wins
10:26 irctc581_ joined #salt
10:27 froztbyte the latter has the upside of not caring what the instance suffixes are
10:27 froztbyte but I'm still writing my own dumb master election/seeding stuff
10:27 froztbyte (it's really annoying that redis makes this hard(er), too)
10:27 aristedes joined #salt
10:28 aristedes joined #salt
10:28 froztbyte babilen: ^ does all that make sense?
10:29 froztbyte I'm also not sure if I'm perhaps over-engineering something, or not yet using a part of salt which might work for this
10:29 aristedes left #salt
10:29 froztbyte (I have thus far not touched the orchestration stuff at all)
10:30 giantlock joined #salt
10:31 babilen froztbyte: I think that your election approach sounds sensible
10:33 babilen Just write a "select leader" function in an execution module that works with mine data and sets a "leader" in a randomly chosen minion (that is then being communicated to the others via the mine)
10:33 babilen A "leader"-flag
10:33 froztbyte yeah
10:33 froztbyte something of the sort
10:33 froztbyte tbh, there's a sliiiight chance of race condition
10:34 froztbyte "if mine empty, set a value"
10:34 froztbyte so I'm a bit wary of that, but I guess I could make a _force option possible
10:39 cberndt joined #salt
10:41 babilen Where would you have a race condition?
10:47 digivandal joined #salt
10:54 monkey66 left #salt
11:00 illern joined #salt
11:06 mikeywaites joined #salt
11:09 stephanbuys joined #salt
11:17 markm_ joined #salt
11:23 zerthimon joined #salt
11:26 guest_ Thank you, babilen, that actually helped. I still have a little trouble understanding all the jinja syntax but I think I have a better grasp of how to structure data
11:34 pbrooko joined #salt
11:34 jaybocc2 joined #salt
11:37 ALLmightySPIFF joined #salt
11:40 fredvd joined #salt
11:46 cyborg-one joined #salt
11:50 jaybocc2 joined #salt
11:50 mikeywaites joined #salt
11:53 ALLmight_ joined #salt
11:53 overyander joined #salt
11:53 aqua^c joined #salt
11:57 mage_ why does grains.os, grains.os_family, etc return "proxy" when I install in dev mode ?
11:58 favadi joined #salt
11:59 diegows joined #salt
11:59 denys joined #salt
11:59 mage_ mmh salt.utils.is_proxy() is broken on FreeBSD
12:01 DammitJim joined #salt
12:04 Fiber^ joined #salt
12:06 mage_ what is a "proxy minion" ..?
12:10 pbrooko joined #salt
12:11 sxar_ joined #salt
12:13 tmclaugh[work] joined #salt
12:20 khaije1 joined #salt
12:25 amcorreia joined #salt
12:26 zwi joined #salt
12:28 khaije1 joined #salt
12:29 slav0nic joined #salt
12:31 chamunks joined #salt
12:35 homeshlice joined #salt
12:36 lothiraldan joined #salt
12:37 nhe joined #salt
12:38 rodio_ua joined #salt
12:38 rodio_ua left #salt
12:38 ssc joined #salt
12:39 ssc Hi
12:40 ssc Is it important/advisable, that the main master of layered master/syndic hierarchy is always online?
12:41 nhe hello, i'm new to salt  ( 2015.5.3) and I have a strage behaviour when filtering over roles deployed in grains.
12:41 nhe It seams that I can't get a deterministic result: the command 'salt -G "roles:swift*" grains.get roles' sometimes return 1 node, and 1 second after returns 2 nodes... Is that normal?
12:41 ssc A scenario would be two servers which run a master, syncic and minion.  On each master there would be multiple linux containers with a minion.  The syntic servers would be controlled via a master on a notebook that is not always available
12:43 ssc An alternative solution might be that only one of the server runs a syndic and the other one acts as a main master.
12:45 DammitJim joined #salt
12:45 Xevian joined #salt
12:45 zer0def joined #salt
12:50 sxar joined #salt
12:52 khaije1 joined #salt
12:52 khaije1 joined #salt
12:54 khaije1 I'm having trouble parsing the yaml from a salt-ssh. Same as this error, http://stackoverflow.com/questions/9055371/python-and-pyaml-yaml-scanner-scannererror-mapping-values-are-not-allowed-her
12:55 khaije1 Is this expected behavior?
12:56 subsigna_ joined #salt
12:57 nhe khaijel, does it passes  http://www.yamllint.com/ ?
12:57 cyborg-one joined #salt
12:58 joehh joined #salt
13:03 harkx Hi, I have some minions that report back 2 times - this leads to errors as they get some commands two time also
13:03 harkx is there anytHing i might be doing wrong?
13:04 racooper joined #salt
13:07 favadi joined #salt
13:08 harkx here's an example  https://gist.github.com/harkx/88a9abae83932a7721e8
13:08 rcurrah joined #salt
13:10 ferbla joined #salt
13:11 jbrnds joined #salt
13:14 jbrnds I find it difficult to understand the jinja based filtering/merging. eg: {% set x=salt['grains.filter_by']({..}, merge=salt['pillar.get']('x.lookup')) %} --- where can I find documentation on it. Is this mainly jinja?
13:15 XenophonF joined #salt
13:16 zwi joined #salt
13:17 VSpike joined #salt
13:17 VSpike What are the rules for how an environment is selected for a host by the top file? Does it go on the first matching entry, for example? Or the last? Or something else?
13:18 VSpike I assume a host can only be in exactly one environment
13:18 jbrnds VSpike
13:19 jbrnds VSpike - default is hostname. Can be also on other 'grains'.
13:19 jbrnds VSpike https://docs.saltstack.com/en/latest/topics/targeting/grains.html
13:20 cpowell joined #salt
13:20 trebortech not hostname but minion-id
13:21 elsmo joined #salt
13:21 cpowell joined #salt
13:21 trebortech Also, a minion can be in more than one environment. It's very flexible that way.
13:21 genediazjr joined #salt
13:22 VSpike I mean in this case, https://bpaste.net/show/bd24592cfbb6 an ubuntu host called 'web-dev-1' ... which env(s) would it be in?
13:22 khaije joined #salt
13:23 trebortech it would match on base and dev
13:23 chamunks joined #salt
13:23 AndreasLutro VSpike: minions aren't part of a single environment - rather, each .sls is invoked as part of that environment
13:24 VSpike Right, so if base mapped file root to /srv/salt/base and dev mapped to /srv/salt/dev and you had the same file in each, which would win?
13:24 trebortech The environments are more focused on state files
13:24 VSpike ahh I think AndreasLutro 's answer explains it
13:24 AndreasLutro I think it's really unintuitive, I tried using environments once and had an extremely hard time getting anything done with it
13:25 VSpike Yeah, it never really made any sense to me at all and I've avoided it so far
13:25 trebortech It started making more sense to me when I started to use gitfs. Looking at it as branches and tag versions helped
13:26 trebortech each branch and tag is a different version
13:26 trebortech sorry... environment
13:26 VSpike That's an appealing idea but there seem to be a lot of gotchas.
13:26 trebortech how so?
13:26 VSpike Just going by the number of questions about it here
13:26 AndreasLutro as long as you *exclusively* use gitfs, maybe
13:27 trebortech I keep my top file in local and all my states in git
13:27 VSpike There's a particular one to do with which branch is base and where the top.sls goes... I forget the details
13:27 mapu joined #salt
13:27 AndreasLutro trebortech: doesn't that mean your top.sls has a lot of duplication?
13:27 trebortech base by default is the master branch but that of course can be changed in master config
13:28 trebortech Not really. can you tell me what might cause the duplication any more with this method?
13:28 jdesilet joined #salt
13:29 trebortech I'm not saying your wrong.  I'm just might not have got that far
13:30 AndreasLutro https://bpaste.net/show/bc45b34778d5 this is what I imagine an envinronmentalised top.sls looking like
13:30 faliarin joined #salt
13:30 protoz joined #salt
13:30 AndreasLutro I could be wrong as well, like I said I gave up pretty fast
13:31 AndreasLutro but I had problems where I couldn't include a dev environment .sls from a base environment .sls and vice-versa
13:32 AndreasLutro which meant that I'd have to copy-paste the top.sls for each environment to make it work properly
13:33 trebortech hmmm. Understood. Let me put a better description of my enironment setup together and maybe that will offer some idea of what you could do and if it would work for you.
13:33 AndreasLutro appreciated!
13:33 trebortech btw, I keep my top file out of the gitfs system and have it local on the master
13:34 trebortech I need to have a better way to config manage that
13:34 AndreasLutro yeah. I *imagine* if top.sls was part of gitfs it wouldn't be as much of an issue
13:34 trebortech yes
13:38 Rkp joined #salt
13:39 austinjp joined #salt
13:41 austinjp Hi there. I'm looking for advice on providers suited to hosting a small fleet of salt-managed LXC containers... what do people recommend?
13:42 hasues joined #salt
13:42 hasues left #salt
13:45 timoguin joined #salt
13:45 jeddi joined #salt
13:47 zmalone joined #salt
13:47 mpanetta joined #salt
13:47 kaptk2 joined #salt
13:48 ALLmightySPIFF joined #salt
13:50 rhand joined #salt
13:50 kawa2014 joined #salt
13:51 cpowell joined #salt
13:52 andrew_v joined #salt
13:52 lothiraldan joined #salt
13:53 mikeywaites joined #salt
13:57 VSpike trebortech: interested to see that too
13:57 lothiraldan joined #salt
13:58 bhosmer joined #salt
13:58 pbrooko joined #salt
13:59 Brew joined #salt
13:59 Akhter joined #salt
14:00 Tecnico1931 joined #salt
14:02 khaije1 joined #salt
14:02 khaije1 joined #salt
14:03 pbrooko joined #salt
14:04 Akhter joined #salt
14:05 sgargan joined #salt
14:07 murrdoc joined #salt
14:16 reji joined #salt
14:17 reji hi there. Anyone use keyring with salt-cloud? Can't understood how to config this.
14:18 _JZ_ joined #salt
14:18 win_salt joined #salt
14:19 win_salt can anyone help me with a quick pillar question?
14:19 babilen Only if you ask it
14:20 win_salt im using cmd.run with some pillar values, is there a way to have it not output what the values are during a state.highstate
14:20 k00l joined #salt
14:20 k00l anyone here ?
14:20 sgargan joined #salt
14:21 k00l :/
14:21 jbrnds win_salt can you provide an example? You want to hide the 'echo' of pillar values in the logs; right?
14:21 jbrnds k00l knock knock
14:21 k00l hey jbrnds
14:22 k00l im a n00b on salt , just wondering where i could ask sme questions
14:22 murrdoc this chatroom
14:22 murrdoc the mailing list
14:22 murrdoc stack overflow
14:22 murrdoc hmm
14:22 k00l kk
14:22 murrdoc lotsa places
14:22 win_salt yeah, when i run highstate the pillar values (username and password) are output in the results, and it would be nice to obfuscate the password at least
14:22 k00l well, iyt apears im not quit gettung the hang of Salt as i i had hoped ..
14:23 murrdoc win_salt:  write custom module or state
14:23 k00l im just starting out and wrote a SLS file that says its intalling PHP into a buntu minion however it wont start.
14:23 murrdoc read in pillar values
14:23 jbrnds k00l by handwritten letter as well
14:23 k00l how can i get more info on why it wont start ?
14:24 mage_ mmh every time I'm submitting pull request on salt the build tests seems to fail, is it normal ?
14:24 k00l it says its installed but i cant seem to find it and PHP is not available off the CIL
14:24 k00l CLI
14:24 jbrnds k00l can you post your top.sls in some pastebin?
14:24 k00l top ?
14:25 k00l is it required that i have a top file. i was just running the php.sls file directly
14:26 k00l sudo salt '*' state.sls php
14:26 k00l im jiust practising on a single master/minion setup for now
14:26 monkey661 joined #salt
14:28 k00l here is the php.sls file
14:28 k00l http://pastebin.com/iXbPRmaR
14:28 mage_ k00l: this is wrong
14:29 k00l i know .. lol
14:29 mage_ what are you trying to do ?
14:29 k00l install php
14:30 k00l its the same layout for apache install so i would hope that iy would work, however no dice
14:30 mage_ pkg_php: pkg.installed: - name: php5
14:30 k00l where did i go wrong ?
14:30 mage_ https://dpaste.de/HJvh
14:32 GreatSnoopy Is there any method to call a salt runner from a state ? I saw there was someone else asking for it a few months ago, but I don't think it ever got implemented...
14:32 jbrnds k00l mage_ I second that
14:33 k00l ok, but i aso want to ensure its  running
14:33 murrdoc joined #salt
14:33 giantlock joined #salt
14:34 mage_ k00l: just add a watch
14:35 k00l i have herd of watch, but have yet to fully understand the syntax of a watch statment. can you direct me to some text on the watch statments use ?
14:37 mage_ https://dpaste.de/o9rS something like that
14:37 spark_ joined #salt
14:40 k00l oh i see, there 2 seperate statments
14:40 stephanbuys joined #salt
14:41 jbrnds mage_ why use watch_in and not require? I would have used require
14:42 viq joined #salt
14:42 lumtnman joined #salt
14:42 pravka joined #salt
14:42 khaije1 joined #salt
14:43 lumtnman joined #salt
14:43 k00l its failing
14:45 reji any idea? https://pastee.org/2zccm
14:46 jbrnds k00l replace watch_in with require. Can you add the 'response' why it is failing in the dpaste.de?
14:47 k00l kk
14:48 pbrooko joined #salt
14:48 k00l updated
14:49 akoumjian Hey folks, still looking for any suggestions on this situation short of doing a cmd.script: https://github.com/saltstack/salt/issues/13550#issuecomment-135858995
14:49 jbrnds k00l and the link to the 'new' snipper
14:49 k00l https://dpaste.de/RyN3
14:49 k00l oops sry, though it just updated
14:50 akoumjian There should be some git state where I can say, "always fetch and checkout to whatever the latest version of this branch is". Doesn't appear to be currently possible in current stable release.
14:50 ericof joined #salt
14:51 jbrnds k00l yeah, some pastesystems it down indeed. ok in the minion, how would you normally start the service php5. Under the hood it just executes a 'service php5 start' command. How is the php service called on the machine (no experience). Just update the 'name' of the php service in the state.
14:51 capricorn_one joined #salt
14:52 whytewolf k00l: what os. and package manegment system does the minion use.
14:52 k00l ohhh ok.
14:52 k00l the minion is a ubuntu 14.04
14:54 lumtnman joined #salt
14:54 whytewolf okay than you will also need php5-cli if you want to include the cli command. php5 by it's self only installs the httpd SAPI and the common files
14:54 katyucha joined #salt
14:55 twork hello friends. issue i was mumbling about late yesterday, still has me confused. 'users' formula is set up, works fine, but i can't figure out how to divide accounts between minions. right now it's all or none.
14:55 jbrnds k00l the magic dissipates fast if you understand what is happening. Also know that the state name (that is how you call the state) is by default the same as the attribute '-name:' within the state. so "php5:  pkg.installed:[]" is equivallent to "php5: pkg.installed:-name: php5" -- with proper indents of course
14:56 twork i think this is wider issue with my understang of how to set up pillar/profile relationships.
14:57 jalbretsen joined #salt
14:57 twork (i posted to the email list last night with a fuller explanation of my situation, if you're on there)
14:58 jbrnds k00l so if you follow whytewolf adice: add a state 'php5-cli: pkg.latest:[]'
15:01 whytewolf k00l: are you really trying to start php with a service? or just have it running through apache? if the later you don't need the php-cli just restart apache after php is installed
15:01 mikeywaites joined #salt
15:01 whytewolf k00l: think about how you would do this with out salt
15:02 k00l yes good point ,, the steps that i would follow without salt.. good way to put it
15:02 twork a little clarification: i know how to control which minions get a particular formula. but for multiple minions all getting the same formula, i can't figure out how to have each one fill out variables with its own set of items from the pillar.
15:03 whytewolf twork: split the pillar information up. they all don't have to have the same pillars.
15:04 sunkist joined #salt
15:05 twork whytewolf: right but that's what i'm stuck on. (basic, i know.) is it in pillar/top.sls? last time i tried to divide minions there i screwed it up.
15:06 twork whytewolf: that's been a while, i've learned a lot since then
15:06 twork n/m, i think i know where to go hunting in docs now.
15:07 twork thanks.
15:11 pbrooko joined #salt
15:14 twork ...and at last, the louds part. see, ask enough times in a public forum, sooner or later i'm sure to wind up feeling dumb. thanks, all.
15:14 sdm24 joined #salt
15:15 whytewolf np twork.
15:18 RedundancyD joined #salt
15:18 twork the isssue i was stuck on: i used to think that 'base:' in the pillar was the point of inflection, so parrallel to that there'd be e.g. 'webserver:' 'fileserver:' or whatever. now i finally get it through my head that no, the '*' directly under base: is what gets subdivided into ftp*:, webserver*:, etc.
15:19 DammitJim joined #salt
15:19 whytewolf yeah, just like in the state top. base is just the enviroment
15:20 twork whytewolf: i understood that, i swear. and yet.
15:20 twork see, i told you i'd feel dumb.
15:22 jbrnds twork indeed the '*' is the match on minion-id. you can match it on other 'grains' as well to separate the pillars for individual minions.
15:23 jfroche joined #salt
15:23 twork jbrnds: thanks
15:24 jfroche hello, is it possible to use external auth with salt-run ? our usecase would be to limit orchestrate to specific users, is it possible ?
15:24 sgargan joined #salt
15:24 hasues joined #salt
15:24 eliasp akoumjian: the 2015.8.0 will bring a huge amount of git-related updates… seeing similar issues here often (e.g. when a repository clone is in a modified state, the git-state stops working etc.)
15:25 akoumjian eliasp: That's nice? :-) In the meantime, I have a fleet of minions who I cannot currently deploy app updates to.
15:26 lothiraldan joined #salt
15:26 Deevolution joined #salt
15:27 akoumjian I'll setup a custom git state module, I think
15:27 hasues left #salt
15:27 clintberry joined #salt
15:27 akoumjian until it's fixed
15:27 stephanbuys1 joined #salt
15:28 jerematic joined #salt
15:32 dopesong joined #salt
15:35 murrdoc joined #salt
15:36 eliasp akoumjian: you could also try distributing the fixed modules (picked from the 2015.8 branch) using "Dynamic Module Distribution": http://docs.saltstack.com/en/latest/ref/file_server/dynamic-modules.html - but you should test this thoroughly to prevent unwanted side-effects
15:39 jbrnds can someone explain what in "{% set mysql = ..., merge=salt['pillar.get']('mysql:lookup')" the ":lookup" does here. I thought lookup defaults to os_family. But in the pillar I need to do: "mysql:  lookup:  <statement>".
15:40 jbrnds ... I want to have defaults in a state, which can be overriden by a pillar.
15:42 eliasp jbrnds: that's usually done through a defaults.yml, e.g. https://github.com/saltstack-formulas/mysql-formula/blob/master/mysql/defaults.yaml
15:44 eliasp the mysql formula is a bit unusual regarding this… most formulas are built to provide a defaults.yml, which is then read by map.jinja which merges the defaults, on top OS specific defaults (based on grains.filter_by()) and finally merges your pillars from $name:lookup on top…
15:45 akoumjian eliasp: Thanks, yeah, I already distribute some custom modules that way
15:45 mikeywaites joined #salt
15:45 eliasp the context built in map.jinja is then used in the individual SLS
15:45 mrwboilers joined #salt
15:46 mrwboilers Anyone know the best way to set the PATH for minions?
15:46 jbrnds eliasp check. indeed I see that pattern a lot (map.jinja). Just wondering if I could do it not through a middle layer and this seems a good option. The 'defaults.yaml' is just a Yaml dict refactored out and externally loaded.
15:46 mrwboilers More specifically, when running a remote command via cmd.run, is there a good way to specify a PATH?
15:47 eliasp mrwboilers: systemctl edit salt-minion, then place in the editor:
15:47 eliasp [Service]
15:47 eliasp Environment=PATH=/usr/bin:/foo/bar:/blah/whatever
15:47 eliasp then "systemctl daemon-reload" and "systemctl restart salt-minion"
15:47 eliasp mrwboilers: ah, just for individual states?
15:48 mrwboilers eliasp: long term, I'm not exactly sure how this process will work. For now, I'm just using cmd.run as this is something that I will be initiating manually. This isn't going to be fullly automated via cron or anything
15:49 eliasp mrwboilers: well, just for cmd.run you can supply "env" which you can use to set a custom PATH
15:49 eliasp mrwboilers: https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.cmdmod.html#salt.modules.cmdmod.run
15:51 debian112 joined #salt
15:52 bluenemo joined #salt
15:52 debian112 ok guys! Is there a salt repo out that that has the old versions for ubuntu: salt-minion 2014.7.0
15:53 manfred no
15:53 mrwboilers eliasp: I'm still not sure of the exact syntax for that. (Sorry for being a bit dense.) would it be something like salt '*' cmd.run "whatever command" env="PATH=/path/to/where/ever:$PATH" ? Or am I off here?
15:53 manfred launchpad only lets us keep one version
15:53 eliasp mrwboilers: I don't think interpolating $PATH would work, but besides that, this could work
15:53 babilen debian112: Debian repositories only have a single version and saltstack does not provide multiple repositories for different versions
15:54 mrwboilers Ok, so I'd likely need to supply the full path, not just the part I want to add then? Thanks for your help. I'll play around with it.
15:54 eliasp mrwboilers: exactly
15:54 debian112 ah man!
15:54 mrwboilers eliasp: thanks!
15:54 babilen debian112: Well, there are repositories for Debian, but not Ubuntu IIRC (cf. http://debian.saltstack.com/ )
15:54 manfred debian112: you would need to package it yourself, or use pip, I uses git and install the tag
15:55 babilen debian112: Why would you need that version?
15:55 debian112 the master is running that version
15:56 debian112 upgrade is not an option right ow
15:56 debian112 too close to the holidays
15:56 debian112 and we just started using some ubuntu servers
15:56 rcurrah is there a good way using python api to determine if the salt-master is running? we had a master stop responding but the processes were still wunning but were getting '[INFO] SaltReqTimeoutError: after 60 seconds. (Try 1 of 7)'. So a process check isnt enough. I though of maybe jumping into the python api and doing some test call to it and if times out raise an alert.
15:57 debian112 so how is everyone handling a when they need to add a new server online, and the version in the repo is no longer available?
15:57 geekatcmu "local repos"
15:57 mrwboilers eliasp: It doesn't seem to work. Even when the command I'm telling it to run is just 'echo $PATH' what it returns doesn't reflect what I set via env=
15:57 geekatcmu always and forever
15:57 whytewolf debian112: typically people use local repos
15:57 debian112 yup, that's what I have for debian
15:57 geekatcmu If you care about versions for anything, you *must* have local repos in use.
15:58 geekatcmu (for everything)
15:58 geekatcmu And we won't discuss the fundamental silliness of wget http:... | sh
15:58 babilen *shudder*
15:59 babilen geekatcmu: But it has a .io tld, it must be cool!
15:59 eliasp http://curlpipesh.tumblr.com
15:59 writtenoff joined #salt
15:59 reji anyone know how to fix ipv6-only issue with salt-cloud and openstack?
16:00 geekatcmu "disable IPv6 everywhere"
16:00 geekatcmu HAHAHAHAHAHAHAHAHAHAHAhahahahahahahahahahaha<THUD>
16:00 babilen No, use IPv6 everywhere
16:00 reji -_-'
16:00 murrdoc haha geekatcmu
16:00 geekatcmu I'm on a role today
16:00 debian112 so any suggestions, on how I should build an old version local repo for ubuntu? Without it then I am force to upgrades masters, and start a local repo at the current version
16:01 reji libcloud just cant parse ipv6 so finding public_ip becomes a loop
16:02 Ztyx joined #salt
16:03 babilen debian112: https://launchpad.net/~saltstack/+archive/ubuntu/salt2014-7 -- has 2014.7.5
16:03 debian112 maybe we should sponsor a repo for saltstack, we have crazy amount of storage.
16:03 whytewolf you're the nsa?
16:04 debian112 naw: lol
16:05 rm_jorge joined #salt
16:05 debian112 thanks babilen
16:06 alemeno22 joined #salt
16:08 iggy debian112: you can use the different ppa's
16:09 catpig joined #salt
16:09 debian112 can I build a local repo from ppa: PPA is new to me, since we mainly use debian and CentOS
16:10 iggy you could mirror it with aptly or something
16:10 iggy debian112: or use it directly
16:10 iggy i.e. add-apt-repository ppa:saltstack/salt2014-7
16:11 iggy for 2014.7
16:11 nhe Hello! any idea why 2 consecutives calls to salt filtered on grains/roles return different values? it doesn't seem deterministic: https://gist.github.com/nherbaut/1659ef0e690c4c33f6c1
16:11 debian112 ok, I think I will like to move local just in case it disappears
16:11 iggy debian112: it won't dissappear, more likely get stale (since 2014.7 won't see any further releases)
16:11 onorua joined #salt
16:12 onorua I'm wondering what is the best way to do something like "salt '*00[12]'
16:13 onorua oh..
16:13 iggy onorua: that should just work (globbing supports brackets afaik)
16:13 onorua salt '*00[12]' glusterfs.list_volumes from some module I currently write
16:13 iggy otherwise, use regex
16:13 onorua I've typed "enter" too early, sorry
16:13 iggy onorua: ahh, that's a horse of a different color
16:14 iggy onorua: you need to use the output of that in your module?
16:14 onorua iggy: yeah... I can call some module on the same node, but if I want to get volumes of the other nodes from my module, it sounds like I need to use reflector
16:14 onorua iggy: yeah
16:14 iggy mine
16:15 onorua I need to use output of that on other host, and based on this - do something tricky again
16:15 iggy read up on the salt mine functionality
16:16 iggy or depending on what you're doing, split your module up into multiple parts and have phase 1 make a reactor call that in turn calls phase 2 on the minions in question
16:18 sbogg joined #salt
16:19 alemeno22 left #salt
16:19 alemeno22 joined #salt
16:21 jodv joined #salt
16:22 wnkz joined #salt
16:22 johtso joined #salt
16:23 shaggy_surfer joined #salt
16:23 onorua hm... I think I need both
16:24 onorua iggy: Thank you very much for the hint! Really appreciate it!
16:24 wnkz Hi, how pillar work when using only salt-ssh ? I'm trying to setup this: https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.dockerio.html#prerequisite-pillar-configuration-for-authentication but auth keeps failing
16:25 iggy onorua: good luck... when you get it working, blog about it (or post it on the saltstack reddit, or....)
16:26 campee joined #salt
16:26 zsoftich joined #salt
16:27 jaybocc2 joined #salt
16:28 wnkz what I mean is, does pillars work the same way with salt-ssh than with traditional master / minion or do I have to store pillar locally on each minion ?
16:29 aqua^c joined #salt
16:30 shaggy_surfer joined #salt
16:30 babilen wnkz: You don't have to, no.
16:32 wnkz babilen: so basically, if I didn't mess up with /srv/pillar/top.sls and my /srv/pillar/docker-registries.sls ; that should be working fine ? I was worried because pillar.items' output looks strange
16:33 sgargan_ joined #salt
16:34 jeadre joined #salt
16:34 babilen wnkz: Strange in what way? Could you paste relevant configuration and output to one of http://refheap.com, http://paste.debian.net, https://gist.github.com, http://sprunge.us, http://dpaste.de, … ?
16:35 wnkz babilen: strange like a huge json string in "stdout" ; not organized key pairs
16:36 hasues joined #salt
16:37 anon_ joined #salt
16:42 babilen wnkz: That would indeed hint at an error. Can you run "grains.items" or "test.ping" on that box?
16:42 babilen But please, show us
16:43 pm90_ joined #salt
16:46 wnkz babilen: https://gist.github.com/wnkz/2c60f6110c272e7f23f2
16:47 wnkz babilen: everything works fine, I now even see my pillar data about registries in this stdout output ; but still I can't pull my private image ; I might have done something wrong elsewhere
16:47 wnkz babilen: by "everything works fine" I mean test.ping, all modules and a simple state that installs a package
16:48 babilen wnkz: Which command did you run of which that is the output?
16:48 wnkz babilen: grains.items
16:48 babilen That isn't a command
16:49 wnkz babilen: sorry, salt-ssh -l info 'FXD-PI_APP_001-ITG' grains.items
16:50 babilen Which version of salt is that?
16:50 toastedpenguin joined #salt
16:50 wnkz I just found out that it is "2014.1.13" (on Debian Jessie) ; probably the distribution's default
16:51 babilen yeah, I would *strongly* recommend to use a newer salt version for salt-ssh.
16:51 iggy me too
16:51 wnkz Jessie is marked as "JESSIE (TESTING)" on https://docs.saltstack.com/en/latest/topics/installation/debian.html ; is it safe to use ?
16:51 iggy it was testing when that was written
16:51 babilen That's just outdated documentation (from when jessie was testing)
16:51 iggy (as in Debian Testing)
16:52 wnkz oh ok, cool then ; sorry I though I ran the latest version (didn't do it myself)
16:52 * iggy files a bug
16:53 iggy oh, it's already fixed in dev, just not backported I guess
16:53 Sketch this is why i hate debian's naming scheme ;)
16:53 Configio joined #salt
16:53 iggy this is why you shouldn't use naming/versioning in documentation that's "barely maintained"
16:54 RedundancyD joined #salt
16:54 murrdoc yo where is 2015.5.5 package at
16:54 * murrdoc gots needs
16:54 aparsons joined #salt
16:54 hasues left #salt
16:55 babilen joehh said that he'll get around to it soon™ - I'd guess sometime next week
16:55 iggy guess I know what my first task is going to be
16:56 babilen FIXALLTHETHINGS?
16:56 murrdoc i am on the upgrade iggy
16:56 murrdoc you can come in and finish that up
16:56 dopesong joined #salt
16:56 murrdoc totally stoked you are gonna work with us
16:56 murrdoc around 2 grand stoked :D
16:57 Grokzen joined #salt
16:57 murrdoc just dont fuck up for 3 months
16:57 babilen iggy will work for murrdoc?
16:57 iggy I'm going to be his manny
16:58 murrdoc not for me
16:58 murrdoc with me
16:58 flebel joined #salt
16:58 babilen Even better
16:58 murrdoc so he ll be out here complaining about all the shit i did for the first month
16:58 babilen That was in .uk, wasn't it?
16:59 murrdoc then a month later be like actually u know what he did it right
16:59 murrdoc nah we have a uk office and a santa monica office
16:59 murrdoc hes headed to sm
16:59 babilen right, that was it
16:59 bhosmer_ joined #salt
17:00 cprior joined #salt
17:00 babilen congratulations to both :)
17:00 wnkz babilen: it seems to work *way* better with 2015.5.3, thanks
17:00 babilen It should, yeah :)
17:01 capricorn_one joined #salt
17:01 murrdoc ok 2015.5.3 it is
17:01 murrdoc till joehh can do the work he does for free and love
17:02 stephanbuys joined #salt
17:05 Grokzen joined #salt
17:06 whytewolf so, is 2015.8 still going to be 2015.8?
17:07 rofl____ and whats new in 2015.8?
17:09 AndreasLutro joined #salt
17:10 Configio joined #salt
17:11 murrdoc DEEZ NUTS
17:11 murrdoc uh the ability to use any queueing system for one
17:12 sdm24 got em!
17:12 murrdoc CHYEAH
17:13 forrest joined #salt
17:15 whytewolf rofl____: to give a basic idea here is a "generated" changelog for the first rc https://raw.githubusercontent.com/tjstansell/salt-changelogs/master/details/2015.8.0rc1-details.rst
17:16 murrdoc no clue what that means
17:16 murrdoc i read it
17:16 wnkz babilen: maybe I talked too fast, now I have "ValueError: unknown locale: UTF-8" ; I probably miss something on the minion side, tried "locale-gen UTF-8" but it's not working
17:17 zmalone joined #salt
17:17 * whytewolf shrugs. it was posted to the email list after the rc1 release
17:19 lpmulligan joined #salt
17:20 bhosmer_ joined #salt
17:20 whytewolf from what i can gather it means a lot of things changed
17:22 amcorreia joined #salt
17:23 lpmullig_ joined #salt
17:25 shaggy_surfer joined #salt
17:26 akoumjian Hm, any suggestions? https://gist.github.com/akoumjian/95682e9a2b0b94c37c11
17:28 anotherZero joined #salt
17:28 timoguin joined #salt
17:30 ajw0100 joined #salt
17:31 akoumjian I'm getting "AttributeError: 'module' object has no attribute 'exceptions'"
17:32 Configio joined #salt
17:33 ericof joined #salt
17:35 k00l joined #salt
17:35 k00l hey all , what is a grain file..
17:35 k00l like what tsh efile name i should use ? .sls ?
17:35 k00l no where does it say what i should call the file whwene i save it to disjk
17:36 denys joined #salt
17:36 k00l if i want to transfer the grain files to the minion , do i store them on the /grains/ folder or is it /_grains/ ??
17:37 whytewolf akoumjian: I'm not 100% sure but msgpack might be a little old. not sure if 0.1.10 had exceptions built in
17:37 k00l this platform is frustrating to learn without any training material
17:37 sdm24 I think its a .sls file. You can test with the grains.setval or grains.append, and see what file is added to /etc/salt/grains
17:37 sdm24 k001 ^
17:38 k00l on which side?
17:38 iggy /etc/salt/grains doesn't have an extension
17:38 iggy it's just a yaml file
17:38 k00l so no ext then ?
17:38 iggy you can also put grains in /etc/salt/minion.d/*.conf
17:38 k00l id liek to keep them broken our into their own files if can
17:39 rofl____ whytewolf: TL;DR? :P
17:39 whytewolf rofl____: pretty much
17:39 iggy k00l: you can't, salt doesn't merge dicts in the config files
17:39 whytewolf rofl____: it is something like 1700 merges
17:40 rofl____ whytewolf: thats _alot_
17:40 rofl____ cant wait for the release notes
17:41 akoumjian whytewolf: okay, that will give me somewhere to start
17:41 k00l so i cant use a file instead of the config ?
17:42 sdm24 you can use a file, but each grain key needs to be unique
17:42 sdm24 use files*
17:42 sdm24 so you can't have one file set role: web, and another file set role: prod, and expect role: -dev, - prod
17:42 k00l i read that if i have a file in a folder (/grains/ or /_grains/) i would liek to knwo what the file ext is suspoed to be, and where i should place said file.
17:43 whytewolf k00l: _grains is for python code that returns a dict that is turned into grains
17:43 k00l so that when i run highstate it will push that file to the minion
17:43 sdm24 . /etc/salt/grains would probably be best if its just one grain. Then you don't need to worry about an extension
17:43 k00l oh which side ?
17:43 sdm24 or use https://docs.saltstack.com/en/latest/ref/states/all/salt.states.grains.html in a state
17:43 KyleG joined #salt
17:43 KyleG joined #salt
17:44 whytewolf k00l: https://docs.saltstack.com/en/latest/ref/file_server/dynamic-modules.html
17:44 baweaver joined #salt
17:44 binocvlar joined #salt
17:45 sdm24 whytewolf: ELI5, what is a simple use case for a dynamic module? To add functionality that Salt does not include?
17:45 k00l which side guys ? master or minion ?
17:46 sdm24 for highstate stuff, you have your state.sls file on the master, and the state.sls file will run commands on the targetted minions, like set the grain value foo: bar on the minion
17:47 whytewolf sdm24: simple use case. to add functionality not currently in salt, or to expande on functionality that does not fully cover use case. IE. a state modules for a exacution module that currently does not have a state module.
17:47 iggy SaltStack has an online training course
17:47 sdm24 whytewolf: thanks. thats what I thought, but I guess I was just overthinking and confusing myself
17:47 impi joined #salt
17:52 k00l ok, so anything in my grains folder in the maser will push to the minion when i run highstate ?
17:52 gladiatr joined #salt
17:52 sdm24 no
17:52 sdm24 you have your master tell a minion which files to put in the minion's grains folder
17:53 sdm24 sorry I didn't fully understand the question before
17:54 sdm24 I think an easier way to do that, instead of manually having the master tell a minion which files to put in its grains folder, is to just use grains.append or grains.present
17:54 k00l ok, more reading for me :) .. thanks all :)
17:54 sdm24 https://docs.saltstack.com/en/latest/ref/states/all/salt.states.grains.html
17:54 cpaclat joined #salt
17:54 forrest anotherZero, twork: Did that example help you guys last night? Or did you figure it out?
17:55 cpaclat i am having some issues with salt jobs and the associated job events.  Is there anyone in the chat that is particularly knowledgable
17:55 whytewolf agreed. if you are not doing something that you need a fully dynamic grain script [ala python] just forget how they are stored on the minion and use grains.present and grains.append
17:55 cpaclat in that area
17:55 anotherZero tbh, I kinda walked away from it forrest...
17:55 forrest anotherZero: WELL SCREW YOU TOO THEN BUDDY! ;)
17:56 anotherZero :'(
17:56 anotherZero There comes a point when a .sh file will do the job and that's fine with me for now :)
17:56 cpaclat specifically is there a good way to tell if your command request was published to a particular minion
17:56 forrest gross
17:57 anotherZero ok then forrest, show me what you were suggesting again.  I don't have it in my history anymore
17:57 forrest anotherZero: I'm just giving you a hard time
17:57 forrest anotherZero, twork: define your grains like this: http://docs.saltstack.com/en/latest/topics/targeting/grains.html#precedence I store mine in the minion conf, but you can put them wherever makes sense there, then do something like this: https://gist.github.com/gravyboat/50e62df6f947f3917342
17:57 forrest There's the paste
17:58 cberndt joined #salt
17:59 anotherZero I mean, is salt designed to handle a situation where I want to have something specific done on different servers?  such as server1 hostname=leakproof, server2 hostname=death_star
17:59 anotherZero (and we're talking about 30+ servers here...)
18:01 ajw0100 joined #salt
18:01 anotherZero forrest: ^ .  those examples don't look nimble enough to manage specific info for that many servers... would you agree?
18:02 jhauser joined #salt
18:02 forrest anotherZero: How many different types of servers do you have?
18:02 anotherZero ~3
18:02 sdm24 anotherZero: Some of my states that install pkgs and manage conf files, I have multiple templates for each conf file (such as Ubuntuconf and Debianconf), and then in my state I have file.managed: source: {{ grains['os'] }}conf
18:02 forrest anotherZero: You can match based on other values
18:02 forrest anotherZero: https://docs.saltstack.com/en/latest/topics/targeting/compound.html
18:03 sdm24 but you can change that to {{ grains['hostname']}}conf or whatever, and then have a death_starconf, leakproofconf, etc
18:04 anotherZero but, I would have to list out each individual server into the top.sls to be able to have specific info for each server... feels like I'd be trying to use a screwdriver as a hammer
18:04 forrest anotherZero: What? Why would you do that
18:04 anotherZero unless i'm totally missing something
18:04 forrest You say you have 3 different types of servers right?
18:04 GreatSnoopy joined #salt
18:05 murrdoc mom dad and baby
18:05 forrest so create 3 minion conf files, each defining the type of grain you want, then based on those grains apply the specific pillar.
18:05 anotherZero yes.  but i was hoping to use salt to manage some specifics based on the server.  such as the db i need backed up, the mysql_slave server address, etc
18:05 anotherZero those are a 1 to 1
18:06 sdm24 anotherZero: you could have each server in a top.sls file, listing each state under it to run, but it will probably be cleaner to match groups of servers
18:06 forrest anotherZero: You're going to have to treat it like that if you want one server to have one DB backed up. What about matching based on groups like this
18:06 forrest https://docs.saltstack.com/en/latest/topics/targeting/nodegroups.html
18:06 forrest It seems to me like you want every server to be a special snowflake.
18:07 s_kunk joined #salt
18:07 dopesong_ joined #salt
18:07 anotherZero it's less that *I* want anything to differ, and more like the company that I do sysadmin stuff for has individually named directories and databases, etc etc on each server based on the client names etc
18:08 whytewolf anotherZero: well you are not going to target on that. you are going to use grains in your state files to change the way salt looks at things
18:08 anotherZero it's just legacy crap i'm trying to figure out
18:09 anotherZero I can just keep maintaining these .sh files :)
18:09 forrest anotherZero: Yeah you might be SoL then and need some better lookup style stuff, if each server requires a unique config, you can either apply that config via pillar and grain matching, or not.
18:09 whytewolf .sh files are the d-evil
18:09 sdm24 anotherZero: I agree with whytewolf. I think the best way to do it is to have each state use jinja to push the correct info
18:10 k00l hey all, im bck
18:10 anotherZero how would I use jinja to say "on this server make sure dbX exists, on that server make sure dbY, on another server make sure dbZ and dbA..."
18:10 k00l Custom grains will be distributed to the minions when state.highstate is run, or by executing the saltutil.sync_grains or saltutil.sync_all functions..... i tryed this 2 times now and itsnot working..
18:11 notnotpeter joined #salt
18:11 binocvlar joined #salt
18:11 whytewolf k00l: that is probley because you are trying to use a yaml file in something i have said over and over again is for python scripts
18:11 k00l its not setting a new value :(
18:11 k00l so grains are py scripts ?
18:11 Ryan_Lane anotherZero: you probably don't even need jinka
18:12 Ryan_Lane jinja*
18:12 whytewolf k00l: _grains is for python scripts
18:12 Ryan_Lane if you put the relavent info into pillars, it's just a matter of writing the states
18:12 rideh joined #salt
18:12 Ahlee i don't suppose there's an easy way to generate a random string inside of a state, is there?
18:12 Ryan_Lane I guess you need jinja just to get the pillar value, but it's {{ pillar.my_value }}
18:13 anotherZero so, I started to create a pillar with the various servers listed, but then I realized that the pillars are for sending a stack of info to each server, such as in my users pillar...
18:13 Ahlee I want to try to prevent duplicate cassandra instances from joining with each other if no pillar is defined or a host, but cassandra gets installed there
18:13 Ryan_Lane Ahlee: https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.mod_random.html
18:13 Ahlee Ryan_Lane: Thanks, i'll backport this to 0.17 as this is exactxly what i want
18:13 k00l interesting
18:13 Ryan_Lane {{ salt.get_str(20)
18:13 Ryan_Lane }}
18:13 baweaver joined #salt
18:13 Ryan_Lane 0.17?!
18:14 sdm24 anotherZero: we use pillars for that. we use mariadb, so we have a mariadb folder in /srv/pillar/ which contains different .sls files, which /srv/pillar/top.sls sets for different mariadb minions. Then the mariadb.sls state will use jinja to make sure the correct database/users exist. Our /srv/salt/top.sls tells 'service:mariadb' minions matched by that pillar, to run the mariadb.sls state
18:14 k00l thnaks whytewolf
18:14 Ryan_Lane Ahlee: you /definitely/ need to upgrade :)
18:14 whytewolf k00l: https://docs.saltstack.com/en/latest/ref/states/all/salt.states.grains.html#salt.states.grains.present use that to push grains to minions
18:14 Ahlee yeah yeah
18:14 Ahlee 0.17 works, though. Since we missed the env to saltenv grace period a lot of our shit is busted, and 0.17 works for now
18:14 Ryan_Lane it's hard to believe it's working for you :)
18:15 whytewolf 0.17
18:15 Ryan_Lane things like saltenv are why I don't use envs
18:15 * whytewolf shiver
18:15 Ahlee yes, I wish we didn't.
18:15 Ahlee but, we do
18:15 Ryan_Lane I've avoided envs in every config management system so far
18:15 Ryan_Lane pillars/grains fake envs well enough :)
18:15 iggy ^
18:15 Ahlee I even went so far as to stand up a translation API in front of the salt-api that just caught env= and changed it to saltenv
18:16 Ryan_Lane wow. salt-api worked in 0.17?
18:16 anotherZero sdm24: you wouldn't have a pastie or gist example of how that is implemented, would you?
18:16 sdm24 anotherZero: lemme pull it up
18:16 k00l thanks whytewolf
18:16 Ahlee Ryan_Lane: salt-api has worked since 0.13 :)
18:16 Ryan_Lane heh
18:16 Ahlee well, 0.13.2
18:16 Ahlee as that's when we started
18:16 Ahlee but, yeah.
18:16 Ahlee between the thousand odd windows minions and env to saltenv
18:17 Ahlee we're slowly working through how to upgrade
18:17 Ahlee let alone doing the work
18:17 Configio joined #salt
18:18 Ryan_Lane ah. you must have a relatively large install
18:18 Ahlee something like 2800 minions total, so not huge
18:18 Ahlee but, not tiny
18:19 murrdoc Ahlee:  sup
18:19 Ryan_Lane that's actually relatively large for config management systems
18:19 anotherZero O_o
18:19 Ahlee sup
18:19 murrdoc nm shipped the kid
18:19 Ryan_Lane you have that working against a single master?
18:19 Ahlee murrdoc: congrats!
18:19 murrdoc ahlee is on .17 ryan
18:19 Ahlee Ryan_Lane: no, split across three masters
18:19 murrdoc you betta recognize these skills
18:19 geekatcmu heh, we're at 2600
18:19 Ryan_Lane murrdoc: oh. I do :)
18:20 Ahlee 0.17.8 now, when you factor in the backports
18:20 Ahlee blazing my own versioning scheme
18:20 murrdoc ahlee version is a thing
18:20 iggy salt-0.17.7-ahlee9826548.deb
18:21 Ahlee Ryan_Lane: we play a lot of games with how to throttle minions checking in, to the point we should have gone to masterless long ago
18:21 btorch is there a way to use {1..5} within a target ?
18:22 iggy btorch: [1-5]
18:22 sdm24 anotherZero: The files are kinda out of order, since gist sorts them alphabetically. The spaces between file names represents the file path https://gist.github.com/sdm24/23a2ea2db62d6f4644ff
18:22 iggy like a normal human being
18:22 btorch iggy: yeah I should have said something like 1..35 :)
18:22 btorch I know I can do 1[0-9] and stuff like that
18:23 iggy regex matching probably
18:23 btorch but I was hoping I could use {0..44}
18:23 btorch cool
18:23 TyrfingMjolnir joined #salt
18:23 anotherZero thanks sdm24.  I'll look at these.  I'm guessing the spaces in the names represent folders ( / )?
18:23 sdm24 yep
18:24 sdm24 so pillar/mariadb/db.sls
18:24 anotherZero thanks.  I have to get back to some coding, but I think this may be a good springboard for tonight
18:24 sdm24 where pillar is the pillar envirionment, most likely /srv/pillar
18:24 sdm24 definitely, glad I could help
18:30 mrwboilers I'm still having issues setting a specific PATH using 'env=' with cmd.run.
18:30 mrwboilers I see in the log that it is getting that argument, yet the output of the command being run indicates that it clearly is not using the path that I have specified.
18:30 mrwboilers In the log I see:
18:30 mrwboilers 'arg': ['gcc --version', {'env': 'PATH=/usr/local/gcc/bin:/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin:/root/bin', '__kwarg__': True}], 'fun': 'cmd.run'}
18:31 mrwboilers yet the output it gives is not the version of gcc that is in /usr/local/gcc/bin, which is where it would run gcc from if it was using the path specified.
18:32 iggy run env and check the output
18:32 mrwboilers Same result if I have it run the command 'echo $PATH' -- it isn't the PATH that I passed to it via env=
18:32 iggy can you paste the state?
18:33 iggy (I'm pretty sure env should be a list, not a string)
18:33 murrdoc list
18:34 mrwboilers So how would I pass a PATH to it then?
18:35 iggy that's why I wondered what your state looked like
18:35 mrwboilers This isn't in a state, it's just running salt 'minion' cmd.run ... from the command line
18:35 iggy oh... nfc then
18:35 mrwboilers nfc?
18:36 iggy env="['PATH': '/some/path']"
18:36 iggy near field communications
18:36 mrwboilers iggy: Thanks! I'll give that a try
18:36 iggy or more likely... no friendly clue
18:38 Ahlee :q
18:38 theologian joined #salt
18:40 quique joined #salt
18:42 mrwboilers iggy: that worked. Thanks again!
18:42 hvn joined #salt
18:42 hvn joined #salt
18:43 linjan joined #salt
18:44 iggy lucky guess
18:44 druonysus joined #salt
18:45 sirex joined #salt
18:45 win_salt joined #salt
18:45 win_salt does anybody know if there is a different renderer for pillar files?
18:45 quique I have this in a jinja template: http://fpaste.org/262131/ but when I run salt-call state.highstate I get: Comment: Unable to manage file: Jinja syntax error: unexpected '/'; line 76
18:46 iggy win_salt: different than what?
18:47 win_salt instead of using jinja to divy up what minions get what values
18:47 Configio joined #salt
18:48 iggy win_salt: there are a number of ext_pillar modules that can fetch data from different sources... all salt really needs is python data structures
18:49 murrdoc dicts
18:49 iggy quique: wild guess? your jinja version is too old to support raw/endraw or it's broken... tried upgrading jinja?
18:51 quique iggy:package: python-jinja2 Version: 2.7.2-2
18:51 quique iggy: that's not that old is it?
18:53 iggy quique: not terribly, but I don't know when raw was added
18:53 tjeason joined #salt
18:54 iggy quique: oh, wait, use the {{ '{%' }} trick instead of raw
18:55 iggy I believe salt (or maybe jinja) recursively processes
18:55 lexter joined #salt
18:58 quique iggy: I tried this: http://fpaste.org/262135/ and got the same error
18:58 iggy no clue then
18:59 shaggy_surfer joined #salt
19:02 quique iggy: actually it did work, the error line number was different, and I didn't catch it, thatnks
19:06 Sokel joined #salt
19:08 dopesong joined #salt
19:08 Sokel left #salt
19:09 subsigna_ joined #salt
19:11 ggoZ joined #salt
19:11 aqua^c joined #salt
19:12 Configio joined #salt
19:17 godlike anybody with salt-syndic getting the "Found minion id from generate_minion_id()" message in the master log? it's slowing CLI response on the syndic node, and changing syndic_wait has no effect
19:17 godlike thing is, it's getting the wrong ID and messing up with the minion, since I have a big hosts file there (for ad-blocking purposes)
19:18 baweaver joined #salt
19:18 sirex joined #salt
19:18 sirex joined #salt
19:24 sirex1 joined #salt
19:25 godlike hmm, it seems that it is actually slowing down master startup/minion registration. once I see the "Authentication accepted from <minion_id>" everything goes back to normal, and afterwards the workers bind to the ipc socket
19:25 pbrooko joined #salt
19:25 DammitJim joined #salt
19:28 interisti joined #salt
19:31 joe_n joined #salt
19:31 Charles_ joined #salt
19:32 Charles_ Does anyone have any experience using the foreman master_tops with a syndic?
19:34 djinni` joined #salt
19:34 kossy joined #salt
19:34 culexor joined #salt
19:34 blue0ctober joined #salt
19:34 Gabemo joined #salt
19:34 gthank joined #salt
19:34 Gareth joined #salt
19:35 giantlock joined #salt
19:35 zmalone joined #salt
19:35 sbogg joined #salt
19:35 Deevolution joined #salt
19:35 XenophonF joined #salt
19:35 ferbla joined #salt
19:35 overyander joined #salt
19:35 ashmckenzie joined #salt
19:35 koomi joined #salt
19:35 Alekti joined #salt
19:35 trebortech joined #salt
19:35 Bryanstein joined #salt
19:35 garphyx joined #salt
19:35 dh joined #salt
19:35 sn00py joined #salt
19:35 dean joined #salt
19:35 LotR joined #salt
19:35 jasonrm joined #salt
19:35 David_B55 joined #salt
19:35 ventris joined #salt
19:35 cliluw joined #salt
19:35 Smoked_Duck joined #salt
19:35 mirko joined #salt
19:35 oravirt joined #salt
19:35 fhh joined #salt
19:35 mike25de joined #salt
19:35 artemz joined #salt
19:35 afics joined #salt
19:35 kaictl joined #salt
19:35 lazybear joined #salt
19:35 ligthert joined #salt
19:35 pjs joined #salt
19:35 cmek joined #salt
19:35 APLU joined #salt
19:35 simmel joined #salt
19:35 ventris joined #salt
19:35 kossy joined #salt
19:35 dean|away joined #salt
19:38 joe_n joined #salt
19:40 APLU joined #salt
19:47 Configio joined #salt
19:49 rdorgueil joined #salt
19:49 slav0nic joined #salt
19:50 joe_n joined #salt
19:52 andrew_v joined #salt
19:52 zmalone joined #salt
19:52 giantlock joined #salt
19:52 sbogg joined #salt
19:52 Deevolution joined #salt
19:52 XenophonF joined #salt
19:52 ferbla joined #salt
19:52 overyander joined #salt
19:52 ashmckenzie joined #salt
19:52 koomi joined #salt
19:52 Alekti joined #salt
19:52 trebortech joined #salt
19:52 Bryanstein joined #salt
19:52 garphyx joined #salt
19:52 dh joined #salt
19:52 sn00py joined #salt
19:52 LotR joined #salt
19:52 jasonrm joined #salt
19:52 David_B55 joined #salt
19:52 cliluw joined #salt
19:52 Smoked_Duck joined #salt
19:52 mirko joined #salt
19:52 oravirt joined #salt
19:52 fhh joined #salt
19:52 mike25de joined #salt
19:52 artemz joined #salt
19:52 afics joined #salt
19:52 kaictl joined #salt
19:52 lazybear joined #salt
19:52 ligthert joined #salt
19:52 pjs joined #salt
19:52 cmek joined #salt
19:53 zmalone joined #salt
19:55 ALLmight_ joined #salt
19:55 harkx joined #salt
20:01 bhosmer_ joined #salt
20:04 win_salt i ran saltutil.sync_grains, and now the grains in a minion config are getting ignored.  "Comment: Unable to manage file: Jinja variable 'dict object' has no attribute " is what comes out of a highstate, does this make any sense?
20:05 pm90__ joined #salt
20:13 win_salt nvm restarting the salt-minion seems to have fixed it
20:14 stevednd joined #salt
20:15 stevednd anyone know what the difference between the system.poweroff and system.shutdown functions is?
20:16 LtLefse just guessing by the names, one only halts the OS, the other powers off the machine?
20:17 jeadre joined #salt
20:18 * LtLefse old enough to remember the days before ACPI power off
20:18 stevednd LtLefse: looks like poweroff runs the command `poweroff`, shutdown does `shutdown -h now`
20:18 kevinquinnyo joined #salt
20:19 JPT i still type shutdown -h -P nowadays :)
20:20 JPT 'It is now safe to turn off your computer.'
20:20 LtLefse JPT++
20:20 murrdoc -P is a sign of weakness
20:20 murrdoc salt '*' system.shutdown
20:20 Configio joined #salt
20:21 JPT I don't want to shutdown 30+ production machines at once, even though most of them may be VMs
20:21 murrdoc sign of weakness
20:21 kevinquinnyo I need to make a remote api call to an external service from my saltmaster after the successful execution of a state.  Would this be a job for the reactor events system?
20:21 murrdoc :)
20:21 JPT Let's say i am just a concerned sysadmin :D
20:25 tycoonbob joined #salt
20:26 tycoonbob Got a question. I am wanting to deploy a directory and set permissions on all files in that directory, but I have a single file that needs different permissions. How do I accomplish this?  I have a file.recurse deploying the dir, and setting dir_mode and file_mode.  I tried also adding a file.manage in the same .sls to change the mode of the one file, but it didn’t work.
20:27 GreatSnoopy kevinquinnyo, i have a quite similar situation where i need to make an api call from the master within a state file (as a prerequisite)
20:28 jeadre joined #salt
20:28 GreatSnoopy the only way until now involves using orchestration, which would be overkill
20:30 kevinquinnyo GreatSnoopy: you said until now and left me hanging ;)
20:30 jodv joined #salt
20:31 murrdoc joined #salt
20:34 kevinquinnyo GreatSnoopy: I think in my case, since I am only looking for a successful state execution before making the remote api call, I can use the reactor system to listen for that event, then do a local.cmd.run for my remote api call.  I just wonder if there's a more idiomatic salt way of doing it
20:34 GreatSnoopy kevinquinnyo, well i have not found a better way,yet
20:35 jodv joined #salt
20:36 pm90_ joined #salt
20:36 murrdoc joined #salt
20:37 pm90__ joined #salt
20:42 opensource_ninja joined #salt
20:50 jeadre joined #salt
20:50 GreatSnoopy kevinquinnyo, actually the reactor would work for both your case and mine
20:50 GreatSnoopy but i still think that for something so simple it should be well, simpler
20:51 GreatSnoopy ffs, it's very simple to call something arbitrary on the minions (with a state with cmd.run ) as part of a required state,it should be as simple to have the same on the master (only)
20:52 GreatSnoopy either i'm missing something or i have to submit a feature request :D
20:55 godlike is it possible that config.py:minion_config() under certain circumstances gets called with the path to the master config file?
20:55 jodv joined #salt
20:55 kevinquinnyo GreatSnoopy: I just found the runners section of the docs, I'm toying with that now
20:58 jalbretsen left #salt
20:59 GreatSnoopy kevinquinnyo, alternatively, you may want to do it with orchestration https://docs.saltstack.com/en/stage/topics/tutorials/states_pt5.html , and then you will call the state first and the runner after the state
21:01 baweaver joined #salt
21:01 kevinquinnyo GreatSnoopy: I use orchestration for doing things where I need a set of tasks to be executed in a particular order based on the host, for example do X,Y,Z on host1, then host2, then host3
21:01 kevinquinnyo it's my understanding that is what's for, but i've noticed with salt, there are a million ways to do a thing
21:01 GreatSnoopy exactly :)
21:02 GreatSnoopy in this case, orchestration will get you calling the api _after_ applying the state
21:03 GreatSnoopy and it will be easy to "read" afterwards, as you have a logical succession with step 1. apply state step2 do stuff on master
21:04 GreatSnoopy but your observations is correct, it seems that there are more ways to do a thing with salt, and sometimes some simple thins are a little bit harder to express
21:05 murrdoc # On busy systems, enabling event_returns can cause a considerable load on
21:05 murrdoc # the storage system for returners. Events can be queued on the master and
21:05 murrdoc # stored in a batched fashion using a single transaction for multiple events.
21:05 murrdoc # By default, events are not queued.
21:05 murrdoc #event_return_queue: 0
21:05 murrdoc so we should use event_return_queue
21:05 jbrnds joined #salt
21:06 zwi joined #salt
21:08 jfroche hello is it possible to use state.orchestrate from the salt command (instead of the salt-run command) ?
21:10 druonysuse joined #salt
21:10 Configio joined #salt
21:11 jodv joined #salt
21:12 iggy nope
21:16 Ztyx joined #salt
21:16 tycoonbob Got a question. I am wanting to deploy a directory and set permissions on all files in that directory, but I have a single file that needs different permissions. How do I accomplish this?  I have a file.recurse deploying the dir, and setting dir_mode and file_mode.  I tried also adding a file.manage in the same .sls to change the mode of the one file, but it didn’t work.
21:18 jfroche iggy: thank you for your answer. so it means there is no way yet to enable external auth/acl on the orchestration ?
21:18 GreatSnoopy tycoonbob, it does, just have the file.manage depend on the other
21:18 GreatSnoopy so that the other gets done first
21:18 rdorgueil joined #salt
21:19 tycoonbob GreatSnoopy: Would that be the order key?
21:19 GreatSnoopy tycoonbob, just a moment, ok ?
21:19 tycoonbob GreatSnoopy: Sure.
21:19 iggy jfroche: it runs in the context of the master, so no
21:20 jfroche iggy: ok thank you
21:21 GreatSnoopy tycoonbob, http://pastebin.com/dZ9Bq05R
21:22 GreatSnoopy tycoonbob, beware, its not properly formatted and i missed a ":"
21:23 GreatSnoopy but i think you can get the idea
21:23 iggy jfroche: do you have some sort of use case you're thinking of?
21:23 Ztyx joined #salt
21:25 tycoonbob GreatSnoopy, I must still be missing something — http://pastebin.com/H81c7Pbs
21:25 jfroche iggy: we have different teams and they would like to orchestrate deployements using salt. orchestration seems like the perfect match for that but we need auth to limit/control which orchectrate sls can be run by users
21:26 GreatSnoopy tycoonbob, well, that should actually work.
21:26 subsignal joined #salt
21:27 tycoonbob GreatSnoopy, Unfortunately, it’s not.
21:27 tycoonbob :(
21:27 GreatSnoopy tycoonbob, are you sure you do not have other errors ? what's the output ?
21:29 tycoonbob GreatSnoopy, Well…I do see “State ‘file.manage’ was not found in SLS ‘application’” — So I must have something else wrong here.
21:29 GreatSnoopy aaaa
21:29 GreatSnoopy file.managed
21:30 GreatSnoopy typo
21:30 ALLmightySPIFF joined #salt
21:30 tycoonbob GreatSnoopy, Yup…that was it.  It’s always the little things!
21:30 tycoonbob Working now.
21:30 GreatSnoopy :)
21:31 tycoonbob Yay.  Thanks!!
21:31 GreatSnoopy welcome
21:32 iggy jfroche: you could try opening a feature request... I'm not sure if there's some reason it's not in there
21:34 RedundancyD joined #salt
21:35 jfroche iggy: right, I will create it right now. salt-api seems to be able to run state.orchestrate, we are looking there now
21:35 joe_n joined #salt
21:36 iggy I was going to mention salt-api (maybe with pepper)
21:37 ALLmightySPIFF joined #salt
21:37 iggy jfroche: I'm not sure you'd be able to limit it to specific runners though
21:37 iggy the eauth system just has @wheel/@runner/@jobs
21:37 ALLmightySPIFF joined #salt
21:37 jfroche argh :(
21:41 jfroche iggy: I can at least limit the user to a specific target
21:41 GreatSnoopy iggy, you could wrap the orchestration calls inside something else supporting auth.
21:41 ALLmightySPIFF joined #salt
21:41 shaggy_surfer joined #salt
21:42 GreatSnoopy either scripts available only to certain os users or some higher level gui like rundeck that wraps salt calls
21:42 Xopher joined #salt
21:42 sgargan joined #salt
21:43 GreatSnoopy some people are heavily using rundeck as a launcher for ansible playbooks, i am using it for fabric scripts, ideea is the same, enable some less potent users to run some powerful stuff in a contained way
21:43 iggy the eauth system should probably be expanded to support the full salt.*.*.* notation
21:44 iggy (salt.states/salt.modules/salt.runners/salt.wheel/salt.util/etc)
21:51 ALLmightySPIFF joined #salt
21:51 Akhter joined #salt
21:52 murrdoc joined #salt
21:53 jfroche iggy: https://github.com/saltstack/salt/issues/26832
21:57 forrest joined #salt
22:00 Xopher joined #salt
22:00 protoz joined #salt
22:01 druonysuse joined #salt
22:01 druonysuse joined #salt
22:02 ALLmightySPIFF joined #salt
22:11 ALLmightySPIFF joined #salt
22:13 unusedPhD joined #salt
22:15 jujumejik joined #salt
22:17 ajw0100 joined #salt
22:19 genediazjr joined #salt
22:20 Xopher joined #salt
22:24 keimlink joined #salt
22:27 mosen joined #salt
22:27 Configio joined #salt
22:28 anotherZero joined #salt
22:31 ALLmightySPIFF joined #salt
22:37 Corey joined #salt
22:40 ALLmightySPIFF joined #salt
22:41 sgargan joined #salt
22:41 bfoxwell joined #salt
22:43 clintberry joined #salt
22:44 jujumejik joined #salt
22:48 aqua^c joined #salt
22:48 XenophonF left #salt
22:49 freeone3000 joined #salt
22:50 freeone3000 I've created a salt stack, and am attempting to image aliyun servers using their mirror provisioning system. I've preinstalled rancher-agent and salt-minion to an instance to ease management. However, when starting up a new instance, salt-master doesn't see it as a new server. Running `salt-call -l debug state.highstate` produces output as if it was the server that got imaged (see https://gist.github.com/freeone3000/dfb7a60c1020642e028d )
22:51 sgargan joined #salt
22:52 freeone3000 Changing minion_id fixes this (I've just discovered), but how can I do this without manually affecting this change on each new server?
22:55 iggy freeone3000: delete the minion_id file before you take the image
22:56 iggy (and the contents of /etc/salt/pki/
22:56 freeone3000 Ah. And that'll cause it to be regenerated? Thanks.
22:56 iggy correct
22:56 murrdoc iggy9000
22:56 murrdoc u still working ?
22:56 iggy it'll default to the hostname without that file existing
22:56 iggy murrdoc: sadly, yes
22:56 murrdoc when are u driving out
22:56 iggy I get 1 day off inbetween
22:57 Configio joined #salt
22:57 murrdoc noob
22:57 murrdoc should have taken a week
22:57 iggy they wanted me to quit and start last week
22:57 jodv joined #salt
22:59 mosen congrats iggy
23:01 murrdoc gonna get u a fancy dinner when u get here igg
23:01 murrdoc like so http://i.imgur.com/njKr5Nt.jpg
23:02 iggy medium rare please
23:05 skoegl joined #salt
23:05 forrest Just give him your last glass of water murrdoc
23:05 ahale joined #salt
23:06 murrdoc chicago has water
23:06 murrdoc la not so much
23:06 pm90_ joined #salt
23:06 murrdoc :D
23:09 sunkist joined #salt
23:22 Configio joined #salt
23:23 ajw0100 joined #salt
23:24 joe_n joined #salt
23:36 jodv joined #salt
23:51 kevinquinnyo joined #salt
23:51 ALLmightySPIFF joined #salt
23:53 sgargan joined #salt
23:54 aristedes joined #salt
23:56 godlike should MasterMinion instantiate a 'minion' (via the minion_config() from config.py) with the master configuration file or the minion configuration file?
23:57 godlike I just discovered this, though I am not at all cerain it's a bug or something else, this is the first time I dived into salt's code https://github.com/saltstack/salt/issues/26825

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