Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-07-14

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

All times shown according to UTC.

Time Nick Message
00:00 mbrgm joined #salt
00:02 jmreicha Ryan_Lane: How do you get the salt minion config on to the host?
00:03 jmreicha cloud-config?
00:04 alexanderilyin joined #salt
00:06 twork what should 'salt.shot_top' return? i've tried several different top files, down to almost nothing, and i get the same result no matter what. so i gather that my setup must be hosed on some fundamental level.
00:06 twork *'salt.show_top'
00:07 fsteinel_ joined #salt
00:09 baweaver joined #salt
00:12 ingslovak joined #salt
00:13 twork oh hell "state.show_top".  not "salt.show_top". ...which, gives me back a list of stuff that looks just exactly like what i want. the app servers get app stuff, and everybody gets the basic stuff.
00:14 twork (i am learning an awful lot as i go, at least.)
00:25 twork oh hell. i think i see my issue. when i run 'state.highstate', i get a thing of beauty. each minion gets just exactly what i want. 'evil' gets almost nothign, the app servers get everything they should.
00:27 twork but when i run state.sls [whatever], what i'm doing is telling salt, "what if i gave this state to the minion," NOT "is salt presently configured to get this stuff".
00:27 rhodgin joined #salt
00:27 * twork wants a stiff drink and a kick to the head
00:42 jaybocc2 joined #salt
00:55 jm-sn joined #salt
00:57 jm-sn Hey guys - I successfully upgraded my dev salt master and single minion to the latest stable version. However, salt-api does not start now. Stack trace and version info are here: https://gist.github.com/johannmiller/928f1dd7757398d673b1
00:59 pullphinger joined #salt
00:59 xMopxShell is anybody else using saltstack to manage docker containers? docker-py seems completely broken from my testing...
01:00 murrdoc joined #salt
01:01 mosen not sure where thats going because there's also the dockerng state
01:04 itru joined #salt
01:05 itru Hi! Is there any way to use regexp in salt jinja?
01:06 itru may be I can somehow register a new custom filter? It would be great if I don’t need to mess with salt sources :)))
01:08 xMopxShell mosen: interesting, i'll try that one too
01:09 xMopxShell Ah, its only in devel branches
01:09 mosen xMopxShell: yeah
01:26 hasues joined #salt
01:26 beauby joined #salt
01:30 hasues left #salt
01:34 mbrgm_ joined #salt
01:40 otter768 joined #salt
01:40 DammitJim joined #salt
01:41 beauby joined #salt
01:50 aparsons joined #salt
01:50 quasiben joined #salt
01:51 napsterX joined #salt
01:52 quasiben1 joined #salt
01:59 jarpy joined #salt
02:00 quasiben joined #salt
02:04 dopesong joined #salt
02:05 BretFisher joined #salt
02:06 beauby joined #salt
02:11 cberndt joined #salt
02:15 iggy twork: yes, state.sls ignores the top file
02:22 beauby joined #salt
02:24 evle joined #salt
02:25 DammitJim joined #salt
02:25 ITChap joined #salt
02:32 ITChap Hi
02:32 ITChap I am trying to add some Jinja to my configuration files: https://bpaste.net/show/5f2e49076f8a
02:32 ITChap someone has an idea why the if doesn't work
02:42 beauby joined #salt
02:42 kevops joined #salt
02:42 kevops hello room
02:43 phix I like Jinja2 better
02:43 phix oh hai there kevops
02:43 kevops < needs halp!
02:44 ValF ITChap, I always put the '-' on the closing part or the expression, not sure if that's the problem but worth trying
02:44 kevops trying to uninstall ultravnc w/ pkg.purged in sls
02:44 ValF ie. {%- if grains['role'] == 'proxy' %}   ==>>>  {% if grains['role'] == 'proxy' -%}
02:44 kevops "all specified packages are already absent"
02:45 kevops do i need to specify the uninstaller and uninstall flags under pkg.purged: ?
02:46 TyrfingMjolnir joined #salt
02:48 kevops going to try "WinVNC" for fullname
02:50 ValF kevops, Hi! I sometimes try to use the salt remote exec modules to do things I want to do in states to test expected result, maybe you should try to uninstall that packages like that and see if it works!
02:50 ValF http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.pkg.html
02:51 CheKoLyN joined #salt
02:52 jarpy Hi all. When specifying minions in a client_acl block, it looks like regexps and globs both work. Is there deep magic that guesses which is intended?
02:53 kevops ValF: thatnks!
02:56 petersson joined #salt
02:57 xmj left #salt
02:58 jmreicha_ joined #salt
03:05 beauby joined #salt
03:06 ITChap ValF: I am still not very clear on the difference between putting the '-' in the front or back
03:08 murrdoc joined #salt
03:08 ageorgop joined #salt
03:16 LeProvokateur joined #salt
03:18 ITChap ValF: Doesn't seems to change anything. I also tried without the '-'.
03:19 favadi joined #salt
03:19 rhodgin joined #salt
03:28 clintberry joined #salt
03:28 TyrfingMjolnir joined #salt
03:34 ValF ITChap, was trying to guess but it seems that's not the issue, let me have an other look at it
03:35 ValF When you say it doesn't work, you mean it doesn't render your variables correctly or it errors ?
03:37 ITChap ValF: my if doesn't seem to work at all
03:38 ITChap ValF: I am trying to check if I can see what is returned by the grains['']
03:40 petersson left #salt
03:40 ITChap ValF: ohhhh it returns ['proxy'] not 'proxy'
03:41 otter768 joined #salt
03:42 ValF Try assigning the grains to a var first and print it
03:42 eichiro joined #salt
03:43 ValF {% set srv_role = grains['role'] %} {{ srv_role }}
03:45 ninjada joined #salt
03:45 ninjada hi guys
03:45 ninjada having some issues setting up salt ec2-autoscaling, anybody able to lend a hand?
03:46 ITChap ValF: should I use the 'in' operator instead of '==' ?
03:47 mosen joined #salt
03:52 ValF I never had issues with '=='
03:53 ValF Try dropping this in your file and see what it returns :  {% set srv_role = grains['role'] %} {{ srv_role }}
03:54 ValF Actually I think you pointed at the error
03:55 ValF try with : {% if  proxy in grains['role'] -%}
03:55 ValF rather {% if 'proxy' in grains['role'] -%} sorry
03:56 ITChap ValF: I fixed the issue. I formatted my grains this way: role:\n - proy
03:56 ITChap instead of role: proxy
03:56 ITChap that's probably why i endup with lists
03:56 ITChap and yeah if 'proxy' in grains['role'] works fine
03:56 ValF right, makes sense, sorry I wasn't really helpful on that one :)
03:57 ValF glad you found out !
03:57 ITChap Thanks for your time anyway
03:57 ITChap I still need to discover how the '-' is working in jinja
03:58 juanito joined #salt
04:00 ValF ITChap, that would be a good starting point : http://jinja.pocoo.org/docs/dev/templates/#whitespace-control
04:00 ValF Just discovered the '+' which seems very nice too
04:01 ITChap ValF: Yeah I checked but I need to test. the '-' in front should remove the whitespace before and at the back the space after.
04:02 ITChap so it means that in my case it shouldn't change anything
04:02 murrdoc joined #salt
04:02 ValF Exact
04:18 ITChap What actually happen when a new minion connect to the master, I mean after accepting the key ?
04:18 ITChap does the minion directly execute highstate ?
04:21 rhodgin joined #salt
04:22 quasiben joined #salt
04:23 quasiben1 joined #salt
04:24 ndrei joined #salt
04:28 ninjada anybody have salt reactor ec2 autoscaling setup? having huge amounts of difficulty with sns
04:46 spookah joined #salt
04:46 aidin_ joined #salt
04:47 itru joined #salt
04:50 ramteid joined #salt
04:51 gladiatr joined #salt
05:04 stoogenmeyer__ joined #salt
05:14 aidin_ joined #salt
05:14 TyrfingMjolnir joined #salt
05:21 gladiatr joined #salt
05:28 Garo_ Can I create a custom renderer for both state and pillar top.sls files?
05:30 Bryson joined #salt
05:42 otter768 joined #salt
05:52 kevops joined #salt
05:53 catpiggest joined #salt
05:57 ageorgop joined #salt
05:59 colttt joined #salt
06:03 stoogenmeyer__ joined #salt
06:05 mbrgm_ joined #salt
06:11 AndreasLutro joined #salt
06:12 ninjada anybody have salt reactor ec2 autoscaling setup? having huge amounts of difficulty with sns
06:13 ninjada oops. seriously tho, anybody up on the latest methods for bootstrapping salt to autoscaled instances on amazon ec2?
06:16 mosen I thought RL posted something earlier.. but I dont know anything about it
06:26 impi joined #salt
06:27 stoogenmeyer__ joined #salt
06:38 soren joined #salt
06:39 markm joined #salt
06:40 KermitTheFragger joined #salt
06:50 Guest89 joined #salt
06:52 ITChap Guys are you aware of any issue when having master and minions in different networks/vlan with routing between ?
07:00 gnudon joined #salt
07:02 Norrland ITChap: well. If traffic between the networks are allowed. Then it should be fine.
07:08 linjan__ joined #salt
07:11 kawa2014 joined #salt
07:13 slav0nic joined #salt
07:15 soren joined #salt
07:15 dkrae joined #salt
07:23 Ztyx joined #salt
07:25 ninjada if i have a salt state designed to deploy an aws ec2 load balancer and autoscale group, how should the state be run? ie. it's not targeting a server or a cloud profile or anything like 'salt 'myserver*' state' o
07:28 solidsnack joined #salt
07:32 rubenb_ ninjada: 'salt yourserver.domain.tld state.sls state_name
07:32 rubenb_ '
07:33 jarpy joined #salt
07:33 ninjada its for aws, to create an iam role, security group, elb, launch config and autoscale group
07:33 ninjada not a specific server
07:33 dijit joined #salt
07:33 ninjada im going by an example thats not very clear. im thinking perhaps its run like a salt-cloud profile
07:33 ninjada but im not sure
07:34 ninjada haven't messed around with states containing aws boto functions before
07:34 dijit so um, quick question: is it possible to do an "if process is present" type operation in salt?
07:35 dijit for example, I'm trying to wait until my processes have been fully killed before I remove a directory.
07:36 Sniper^^ joined #salt
07:36 rubenb_ dijit: http://docs.saltstack.com/en/latest/ref/states/requisites.html#onlyif
07:36 dijit :O
07:36 dijit all hail rubenb_
07:37 dijit thanks man.
07:37 sirex joined #salt
07:40 yomilk joined #salt
07:40 VSpike ninjada: I'd like to know the answer to that question too
07:41 VSpike seems like a lot of salt states, esp the boto related ones for route53 and other AWS stuff, don't relate to a specific machine so I wonder where you'd target them to run
07:41 VSpike Although I'm wondering whether it would be better to use cloudformation + salt instead. But not quite reached the point where I've had to solve it yet.
07:43 otter768 joined #salt
07:43 illern joined #salt
07:48 _mel_ joined #salt
07:51 napsterX joined #salt
07:53 rubenb_ VSpike, ninjada: Do these machines have specific grains; you can target salt-command with "salt -G 'provider:aws' cmd.run uptime"
07:58 rubenb_ https://docs.saltstack.com/en/develop/ref/cli/index.html#targeting-with-executions
08:01 khebbie joined #salt
08:04 pcdummy joined #salt
08:05 pcdummy joined #salt
08:10 fredvd joined #salt
08:13 _mel_ left #salt
08:15 sunkist joined #salt
08:15 solidsnack joined #salt
08:17 Xevian joined #salt
08:18 pcdummy joined #salt
08:18 pcdummy joined #salt
08:19 slav0nic joined #salt
08:21 illern joined #salt
08:23 aqua^c_ joined #salt
08:26 Ben_NN_ joined #salt
08:26 jespada_ joined #salt
08:38 impi joined #salt
08:53 jhauser joined #salt
09:01 Furao joined #salt
09:18 illern joined #salt
09:19 cast_ joined #salt
09:20 danielcb hi, i want to install a windows application that has some setup options set via a LOADINF file
09:21 sunkist joined #salt
09:21 danielcb is it possible to deploy that necessary options.inf via salt? so what I want to do, to clarify:
09:22 danielcb install_flags: ' /SILENT /LOADINF="salt://win/repo/files/options.inf" '
09:22 keimlink joined #salt
09:22 danielcb here it's done via a network share:
09:22 danielcb https://github.com/saltstack/salt-winrepo/blob/master/pdfcreator.sls
09:23 danielcb but I would like to do it with salt alone, if thats possible
09:32 aidin_ joined #salt
09:34 PI-Lloyd Anyone know how to get the following reactor to work properly, as in wait for each job to complete before executing the next job.. at the moment it's firing them all off at the same time in a random order - https://bpaste.net/show/345f6b07f06a
09:34 PI-Lloyd everything works other than the ordering and the waiting for each to finish
09:34 ninjada joined #salt
09:35 PI-Lloyd documentation is very lacking on this part of the reactor system
09:37 Furao joined #salt
09:37 PI-Lloyd the reason this needs to be ordered and waiting for each step to complete is this reactor will be triggered by our ENC to assign a device to a saltmaster depending on it's geographic location... so it needs to add the key to the correct master, run a highstate on the device, then remove the key from the default master
09:44 otter768 joined #salt
09:49 evle Hi. Yesterday I asked about using pillars outside of states. For example here:
09:49 evle salt '*' cmd.run "echo {{ pillar['name'] }}"
09:49 evle Apparently such construction doesn't work since pillar variable appear empty.
09:49 evle salt '*' cmd.run "echo {{ grains.id }}" works fine,  {{ pillar['name'] }} inside state file works fine too.
09:49 evle Is there a reason for such behavior?
09:55 markm joined #salt
10:06 Kelsar joined #salt
10:06 OmiKrOn joined #salt
10:07 linjan joined #salt
10:08 yomilk joined #salt
10:12 CeBe joined #salt
10:14 aidin_ joined #salt
10:15 zer0def joined #salt
10:18 ingslovak joined #salt
10:23 denys joined #salt
10:36 aidin_ joined #salt
10:39 khebbie joined #salt
10:55 jim_ joined #salt
10:59 aqua^c joined #salt
10:59 CeBe1 joined #salt
11:01 hax404 joined #salt
11:02 _JZ_ joined #salt
11:05 napsterX joined #salt
11:13 yomilk joined #salt
11:15 giantlock joined #salt
11:18 linjan joined #salt
11:26 Ztyx joined #salt
11:26 neocid joined #salt
11:38 kusams joined #salt
11:44 otter768 joined #salt
11:56 OmiKrOn joined #salt
11:59 aqua^c joined #salt
12:02 supersheep joined #salt
12:10 neocid joined #salt
12:18 neocid joined #salt
12:18 danielcb is it possible to set a windows environment variable with salt?
12:18 aidin_ joined #salt
12:18 danielcb environ.setenv just sets one for the minion process
12:22 rhodgin joined #salt
12:23 SheetiS joined #salt
12:25 ninjada joined #salt
12:34 yomilk joined #salt
12:41 gladiatr joined #salt
12:42 DammitJim joined #salt
12:44 Ztyx joined #salt
12:44 subsignal joined #salt
12:45 patrek joined #salt
12:47 ndrei joined #salt
12:47 yomilk joined #salt
12:51 patrek joined #salt
12:56 BretFisher joined #salt
12:59 anotherZero joined #salt
13:00 kevops joined #salt
13:00 aqua^c joined #salt
13:02 hasues joined #salt
13:02 evle1 joined #salt
13:06 timoguin joined #salt
13:08 Tecnico1931 joined #salt
13:08 Savemech joined #salt
13:08 jespada_ joined #salt
13:08 giantlock joined #salt
13:09 hasues left #salt
13:12 andrew_v joined #salt
13:15 dendazen joined #salt
13:18 sirtaj joined #salt
13:19 toastedpenguin joined #salt
13:20 icflournoy joined #salt
13:21 toastedpenguin joined #salt
13:23 astro_sk joined #salt
13:25 kaptk2 joined #salt
13:25 primechuck joined #salt
13:26 perfectsine joined #salt
13:32 XenophonF joined #salt
13:32 peters-tx joined #salt
13:34 Deevolution joined #salt
13:35 teryx510 joined #salt
13:37 dopesong joined #salt
13:37 ferbla joined #salt
13:38 mpanetta joined #salt
13:38 amcorreia_ joined #salt
13:40 kawa2014 joined #salt
13:43 cpowell joined #salt
13:45 otter768 joined #salt
13:46 EasyAnalyst joined #salt
13:52 furrowedbrow joined #salt
13:52 dopesong joined #salt
13:53 XenophonF style question - grains['os_family'] or salt['grains.get']('os_family') or salt.grains.get('os_family')?
13:53 jdesilet joined #salt
13:53 XenophonF i'm kind of leaning toward the first because it's more consise
13:55 AndreasLutro I do grains.os_family
13:55 AndreasLutro easier to type, easier to change
13:55 XenophonF that's even better
13:55 XenophonF hey, did you ever figure out your https problem?
13:56 AndreasLutro yes
13:56 kevops morning all
13:56 XenophonF i replicated  the error on my freebsd minion (salt 2015.5.2)
13:56 XenophonF were you able to work around or fix it? if so, how?
13:56 Ahlee XenophonF: second one iwth a default, lest you wind up with failures
13:56 AndreasLutro https://github.com/saltstack/salt/issues/23103 XenophonF
13:56 saltstackbot [#23103]title: Unable to manage file: 'handle' in 2015.2.0 | I want to download the couchdb sources and extract them, but always fail with: ...
13:56 AndreasLutro seems like an upstream bug that won't be fixed. it works in 2015.8
13:56 XenophonF gotcha
13:57 AndreasLutro the reason it works because they replaced their http backend
13:58 XenophonF yeah, i started tracking the call stack for that and gave up around the time i got to  cp.cache_file or something
13:58 XenophonF glad to hear a fix is pending, but too bad it's going to wait until 2015.8
14:00 pcdummy joined #salt
14:00 AndreasLutro time to go home
14:02 ndrei joined #salt
14:10 tvinson XenophonF: grains.get is useful in logic where the grain may or may not present.
14:11 XenophonF tvinson: agreed, same for pillar.get
14:11 XenophonF i'm going to stop using the salt['funcall'] syntax, though
14:12 jespada_ joined #salt
14:14 speriyasamy joined #salt
14:15 go joined #salt
14:15 ronrib joined #salt
14:17 rgardam joined #salt
14:17 rgardam Hey
14:17 rgardam I was wondering about multi region salt master setups?
14:18 rgardam I know currently there are some issues with multi-master salt mines
14:18 rgardam What would be really neat would be to be able to sync salt mine information selectively to different masters
14:19 supersheep joined #salt
14:20 hasues joined #salt
14:20 hasues left #salt
14:22 racooper joined #salt
14:22 tmclaugh[work] joined #salt
14:28 adelcast left #salt
14:28 kawa2014 joined #salt
14:29 supersheep joined #salt
14:29 TyrfingMjolnir joined #salt
14:33 w1gz joined #salt
14:34 khebbie joined #salt
14:35 zmalone joined #salt
14:37 pguinardco like running your salt masters as salt minions?
14:38 pguinardco that sounds like a bad idea when your region changes slightly as opposed to having a single template with everything
14:40 tmclaugh[work] joined #salt
14:40 CeBe joined #salt
14:43 kbyrne joined #salt
14:48 jaybocc2 joined #salt
14:49 aqua^c joined #salt
14:52 jespada_ joined #salt
14:57 ksj when using the salt command in a template (e.g salt['grains.get']("os")), it only acts on the current targetted machine. Is it possible to run a salt command on all boxes? I want to loop through all my custom "roles" grains on all my boxes for /etc/sudoers
14:58 manfred ksj:  yes
14:58 manfred ksj:  look up peering and publish.publish
14:59 ksj thanks
14:59 iggy ksj: that doesnt run the salt command. that runs a function in the salt dictionary
14:59 manfred kaictl:  http://docs.saltstack.com/en/latest/ref/peer.html
15:01 iggy rgardam: you'd be better off using some other database for that info (new versions of salt have sdb)
15:02 Ztyx joined #salt
15:02 adelcast joined #salt
15:03 ksj yeah, I'm not not sure peering is right for me. I probably shouldn't have roles defined in grains anyway. Some people on here advised against it. So maybe I'll rethink it.
15:03 ksj sdb looks really interesting
15:05 __monty__ joined #salt
15:07 quasiben joined #salt
15:07 quasiben joined #salt
15:07 khebbie joined #salt
15:08 aparsons joined #salt
15:10 neocid joined #salt
15:10 markm joined #salt
15:11 TyrfingMjolnir joined #salt
15:11 jondonas joined #salt
15:12 ALLmightySPIFF joined #salt
15:13 clintberry joined #salt
15:13 napsterX joined #salt
15:15 ALLmightySPIFF joined #salt
15:16 aparsons joined #salt
15:18 andygabby joined #salt
15:19 dfinn joined #salt
15:28 wenzel_ joined #salt
15:28 wenzel_ left #salt
15:29 mbrgm_ joined #salt
15:29 sdm24 joined #salt
15:33 zsoftich1 joined #salt
15:35 breakingmatter joined #salt
15:36 breakingmatter left #salt
15:38 gnudon joined #salt
15:43 clintber_ joined #salt
15:45 jf joined #salt
15:46 otter768 joined #salt
15:47 Ztyx joined #salt
15:47 jespada_ joined #salt
15:47 jfjf hey guys if i want to install salt using git, how would i do it?
15:47 jfjf would i clone saltstack/salt, and run setup.py?
15:49 juanito joined #salt
15:49 mitsuhiko joined #salt
15:49 XenophonF joined #salt
15:51 venu0336 joined #salt
15:54 giantlock joined #salt
15:54 jfjf or should i use the bootstrap script instead? I'm ok doing things in 2 steps: curl bootstrap script, then run it to get salt installed through git, but I was wondering if there was a more direct route.
15:55 Savemech joined #salt
15:56 jespada__ joined #salt
15:57 saltman joined #salt
15:58 rgardam iggy: Thanks for the heads up. When will this be released?
15:58 dfinn joined #salt
15:58 jalbretsen joined #salt
15:59 Ztyx joined #salt
16:01 kevops joined #salt
16:01 iggy jfjf: use bootstrap script
16:01 iggy rgardam: it's in 2015.5
16:01 conan_the_destro joined #salt
16:03 rgardam ah cool
16:03 rgardam :D
16:04 dhscholb hey folks, is there any way to do CIDR matching in an sls file other than top.sls? such as in a jinja statement?
16:04 jfjf iggy: thanks. Would this be considered secure enough for production?
16:04 jfjf assuming github isnt compromised, of course
16:06 iggy we use it... you could always mirror the script on you own infra if you wanted
16:06 dhscholb actually, nevermind about CIDR matching. I realized that what i wanted to do should probably go in pillar's top.sls
16:10 XenophonF good god, curl | sh is never ever secure
16:11 XenophonF better to have a local copy or something under your direct control
16:15 jfjf Figures. To give some context to my question: Salt Essentials (the book) seems to be asserting that "it is not the most secure method for installation" and that it's possible to install "directly from the source on Github"...
16:17 jfjf Maybe i'm just putting too much weight in these words. Logically speaking, yes, u would be right. Better to have a copy and run from that.
16:19 venu0336 joined #salt
16:19 XenophonF at least download the script and give it a once over, just to make sure that there isn't anything obviously malicious in there
16:20 saltman Hello, I upgrade my salt from 2014.7.6 to 2015.5.2. from some reason, the command:"salt-call pillar.get master" stop working in the version. I try to find the code which generate this pillar, but I failed. Is there any idea to someone?
16:20 saltman * stop working in the new version
16:20 jfjf got it. Thanks, XenophonF and iggy
16:23 iggy saltman: it's off by default now... look through the config file changes
16:25 jespada__ joined #salt
16:25 saltman How I enable it in the config file?
16:25 saltman This is a custom file, and there is no parameter for this..
16:27 rgardam Are there any current examples for using the sdb?
16:28 XenophonF would anyone be able to share how they set up a webhook in github to fire off a fileserver.update runner whenever someone pushes to a repository?
16:29 XenophonF i am completely new to the reactor system, or salt-api for that matter
16:29 jaybocc2 joined #salt
16:34 conan_the_destro joined #salt
16:37 snarfy joined #salt
16:37 cpowell_ joined #salt
16:37 dingo_ is 'cachedout', a saltstack employee, available by any other means than public shaming on github? https://github.com/saltstack/salt/commit/c041f2905fae709c260bd3d3ec1754303384c2ee
16:37 dingo_ preferably irc chat, hehe
16:37 snarfy holla
16:38 dingo_ he just broke the release and i don't think he knows it
16:38 aqua^c joined #salt
16:39 gladiatr joined #salt
16:40 bluenemo joined #salt
16:40 bluenemo joined #salt
16:41 campee joined #salt
16:41 jfjf uh? his profile is at https://github.com/cachedout
16:42 jfjf u'll find his email there
16:42 bluenemo hi guys, I'm playing with pyobjects and I kind of like it! One question though, I want to always define the ID that is printed in the output. I can do:   File.managed('openvpn_server_config_' + vpnname + '.conf',    and then give the other arguments. Is there an identifyer for the ID of the state?   as in something like id='myfancyname' ? That would help my oversight
16:44 jfjf dingo_: u mean "master_type" should stay as "str"?
16:45 sarerto joined #salt
16:45 bluenemo also, If i implement all my with's, there is quite some level of indentation building up.. Did I missunderstand sth?
16:46 dingo_ jfjf: yes i commented on the commit-id precisely
16:46 dingo_ and also e-mailed the commiter's e-mail address from github
16:46 zmalone joined #salt
16:46 dingo_ the details
16:47 jfjf I see! got it, thanks. Sorry I missed that comment
16:48 ben_NN_ joined #salt
16:49 jfjf If u check out a specific git version, though, shouldnt master_type be 'str' already?
16:50 spookah joined #salt
16:55 Bryson joined #salt
16:55 ajw0100 joined #salt
16:56 dingo_ oh sorry, missed your reply, jfjf, are you cachedout?
16:56 druonysus joined #salt
16:56 druonysus joined #salt
16:57 jfjf I'm not
16:57 dingo_ maybe you misinterpret what i say by 'release', should i say: 'release branch', right? :)
16:57 dingo_ i don't mean they broke whatever .tgz is downloadable
16:57 dingo_ but this branch on github got broked
16:58 jfjf I see!
16:58 Redmond joined #salt
16:58 dingo_ maybe you all can chime in here... deterministic builds, you know, bugfixes, we like those too... we like stability, too... so we package up salt, and we chose a relatively older release, 2014.7 or whatever it is...
16:59 dingo_ but they keep backporting bugs!! I've just gotten 2 or 3 backported bugfixes  for "bugfixes" merged upstream this past few weeks
16:59 aw110f joined #salt
16:59 dingo_ so then i have to fork a branch to build from for a while, a clone'd repo, with my fixes in until they merge upstream... not complaining, anyway, they merge them in, so i point back to upstream, and *more* *new* bugs
17:00 dingo_ so... i mean, i know they're always backporting fixes, too, i want those
17:00 dingo_ i just don't know where the balance is
17:00 jfjf have u brought that up to the list?
17:00 dingo_ i'm thinking of having a cloned repository i specificly 'git pull' from, and when they have new bugs, just rewind the commit id's of the "2014.7" branch i point to
17:01 impi joined #salt
17:01 dingo_ i'm not like... complaining, "saltstack should do this better", i'm sure better testing would work (fuck jenkins, btw, seems all my PR's are "failed" yet "ok to merge")
17:01 dingo_ i'm just asking, you all are users, too
17:02 dingo_ how do you manage your own salt releases? i can't be the only one who has bumpbed into 6+ bugs and submitted many upstream fixes, the root cause being "failed backportts of a fix introduced new bug"
17:02 jfjf If this is a real problem for u, I dont see why others wouldnt have the same problem. I think it would be good to highlight this over the list
17:02 dingo_ they've even gone as far as change data structures of function arguments in these "backport fixes"
17:02 dingo_ is 2014.7 so bad because its so old? is the 2015. releases more stable?
17:03 dingo_ perhaps i'm the only one to notice because i'm the only one to build salt off of 2015.7 branch
17:03 dopesong joined #salt
17:03 dingo_ 2014.7 rather
17:04 jfjf Btw: I didnt get what you were saying in that comment on that commit. Should the url for 2014.7 have been https://github.com/saltstack/salt/blob/2014.7/salt/config.py ?
17:05 jfjf (https://github.com/saltstack/salt/blob/2014.7/salt/config.py#L274 to get to the line)
17:06 ageorgop joined #salt
17:07 campee joined #salt
17:08 dingo_ it should be str here
17:08 dingo_ no quotes
17:09 dingo_ at least, then it would work like it used to
17:09 dingo_ unfortunately the commit he did yesterday: str => 'standard' provides no details of what he thinks he fixed
17:09 dingo_ and all the tests seem to mostly fail
17:10 dingo_ but i guess jenkins is such incomprehensible shit nobody really knows when its ok to merge anyway
17:12 jfjf u know i just realized something (crazy?): the tags have a 'v' in front of them. But the branches dont?!!!
17:12 dingo_ yeah i noticed that only today as well..
17:12 dingo_ i think github UI thing
17:13 dingo_ maybe if the branch begins with a number, it does that
17:13 dingo_ maybe to prevent it from being confused as hexidecimal in the routes
17:13 dingo_ an 8-digit number is indistinguishable from an 8-digit short sha commit hash
17:14 jaybocc2_ joined #salt
17:15 jfjf I think it's up to the owner whether to put a 'v' or not.
17:15 jfjf But anyway: https://github.com/saltstack/salt/commit/f074e7c67bf43446dffe509c4196f6f2a95220ca
17:16 jfjf if you're saying that it should be "str" without the quotes, this is the one that broke it, I think
17:16 dingo_ hehe, jfjf working for me today!
17:16 dingo_ ahh yes, right, this was a fix in develop
17:16 dingo_ i saw this too
17:16 dingo_ he tried to backport it, but probably should not have, in the way he did, anyway, i may be wrong, anyway, he can fix it himself
17:17 jfjf ha, as it turns out, my name is Jeff too...
17:18 dingo_ hehe
17:20 ajw0100 joined #salt
17:20 jfjf compare: https://github.com/saltstack/salt/blob/2014.7/salt/config.py#L259 vs https://github.com/saltstack/salt/blob/2014.7/salt/config.py#L274
17:21 jfjf sorry https://github.com/saltstack/salt/blob/v2014.7/salt/config.py#L259 vs https://github.com/saltstack/salt/blob/2014.7/salt/config.py#L274
17:21 jfjf the difference of a 'v'?!!!
17:22 dingo_ huh
17:22 dingo_ tag vs. branch
17:22 dingo_ one of them is a tag (v), one is a branch
17:23 dingo_ mysterly solved
17:23 dingo_ v2014.7 was frozen last year
17:23 dingo_ the first release
17:23 dingo_ the branch, 2014.7 is the maintenance branch
17:23 dingo_ for subsequent patch releases
17:26 Ryan_Lane jmreicha_: deployments
17:26 jfjf I see, thanks for explaining.
17:26 jfjf I was in the mode of thinking that the branch would freeze as well, but yeah u're right.
17:26 Ryan_Lane jmreicha_: cloud-init pulls down a tar file and untars it. that tar file has salt, a basic config and a deployment script
17:27 Ryan_Lane jmreicha_: then we just run a normal deployment. so, initial salt run is the same as every other run. we also run salt for every deployment
17:28 iggy dingo_: the vYYYY.M tags are for doing the diff against for the changelog (or something... base of pi told me one time and I barely remember)
17:29 speriyas_ joined #salt
17:29 dingo_ well sure, absolutely
17:30 ekristen joined #salt
17:31 neocid joined #salt
17:33 scoates joined #salt
17:34 bbradley howdy
17:35 dingo_ well jfjf if you're interested, i polled my team for consensus, we agree to freeze to a specific commit-id, and damned be these so-called "fixes" they keep backporting, we won't need them unless we find out we need them!
17:35 bbradley is it possible for a file to watch another file?
17:35 ksj hey, I currently define what a machine is (roles, environment etc) using custom grains, along the lines of this article: http://www.saltstat.es/posts/role-infrastructure.html. There was discussion on here a while back about why that was a bad way of doing it. I'm wondering if there's any documents describing a best practice way of organising your machines?
17:35 dingo_ surely it is, bbradley
17:35 dingo_ any state may be watched, the watch/requires are agnostic to their module function
17:36 XenophonF ksj: i make environment and role assignments in pillar
17:36 bbradley if a service watches a file that watches another file, should the service restart?
17:36 jfjf dingo_: i'm interested, yeah
17:36 XenophonF that way attackers can't break into a minion and iterate through all of my configs by changing the relevant grains
17:36 whytewolf ksj: not sure about a document about why. but basicly because grains can be set on any host from the host so you can not be sure that the minion truly is what is says it is.
17:37 ksj yeah, security was one of the big issues
17:37 XenophonF ksj: config examples at https://github.com/irtnog/salt-states and https://github.com/irtnog/salt-pillar-example
17:37 XenophonF i use the minion id to set the appropriate pillars
17:38 jfjf iggy: thought i remembered Thomas mentioning this before, but I just figured out the versioning. The last digit denotes the release of a branch. So v2014.7.6 would be the latest release for the 2014.7 branch
17:38 XenophonF which also can't be changed unless the master's also compromised
17:38 dingo_ bbradley: service -> file(a) -> file(b) ;; yes, if file(b) changes, service will be restarted, if 'running: True' and 'restart: True'
17:38 dingo_ ... for service.managed state
17:39 ksj XenophonF: good point. and if the master's compromised, you're pretty screwed.
17:39 dingo_ hey all i'd like to share a project i hope to generalize and release sometime, to make graphs of highstate yaml output: https://teamcity-master.pexpect.org/tmp/storage01.highstate.svg
17:39 XenophonF yeah, at that point you're done
17:39 bfrog joined #salt
17:40 bfrog how would I run postgres initdb one time?
17:40 aparsons joined #salt
17:40 iggy does it create a file? if so use cmd.run's creates: option
17:40 bfrog it does, thanks
17:41 bbradley dingo_: i'm using service.running.
17:41 XenophonF and if it doesn't create a file, tack "&& touch /var/lib/.something_jobs_done" to the end of the command
17:41 dingo_ that's fine, too. bbradley
17:41 jfjf quick qn: does anybody run a minion on their master as well?
17:41 XenophonF yes
17:41 LtLefse yes
17:41 XenophonF salt-formula ftw
17:41 iggy everybody probably should
17:41 bbradley changes in file(b) do not trigger changes in file(b) and the service is not being restarted.
17:42 iggy but you aren't the first person to ask that question
17:42 jfjf :)
17:42 murrdoc joined #salt
17:43 LtLefse it's really an orthagonal decision: even if you don't trust salt-minion to manage the salt-master stuff, you can might use it to configure the OS just like all the other minions
17:43 XenophonF jfjf: again look at my gh repos for example configs
17:43 iggy bbradley: changes aren't passed through, so a change in fileA would run the state for fileB... if it doesn't also have changes, it would return no changes
17:43 XenophonF rb
17:43 XenophonF brb
17:43 jfjf thanks, all. I guess the rationale is to manage the master itself as well using salt?
17:43 jfjf XenophonF: where are they?
17:43 bbradley kk
17:43 iggy you can, but it's not exactly necessary
17:43 XenophonF jfjf: https://github.com/irtnog/salt-states and https://github.com/irtnog/salt-pillar-example
17:44 iggy f.ex. I have a salt deployment where the minion doesn't manage anything to do with the master, but the master does other stuff (which the minion does manage)
17:44 robawt XenophonF: change management is always a big deal :)
17:44 robawt ten years later, sorry
17:44 bbradley iggy: should i be using onchanges?
17:45 iggy bbradley: it still isn't going to pass them through, so you need the service to watch both files probably (along with fileB watching fileA)
17:45 bbradley yea that would work.
17:45 bbradley thank you
17:46 jfjf iggy: could u clarify that earlier sentence? Minions manage stuff? Arent minions managed?
17:47 iggy jfjf: I mean the minion agent/software itself managing the other stuff running on the nodes
17:47 otter768 joined #salt
17:48 iggy I don't tend to refer to the instance/server/node as a minion
17:48 iggy the minion is the software running on said instance/server/node
17:56 ALLmightySPIFF joined #salt
17:56 XenophonF back
17:57 jmreicha_ Ryan_Lane: Thanks
17:57 baweaver_ joined #salt
17:57 XenophonF back
17:57 hal58th joined #salt
17:57 michelangelo joined #salt
17:57 XenophonF robawt: instituting change management procedures at work was probably the most impactful thing we did
17:57 jfjf iggy: so the minion daemon (as opposed to the minion "machine") managing the software on the node on which it's installed?
17:59 iggy that's what I meant, yeah
18:01 primechuck joined #salt
18:02 jfjf ok, thanks
18:05 mbrgm_ joined #salt
18:07 Ryan_Lane jmreicha_: yw
18:09 giantlock joined #salt
18:10 viq joined #salt
18:12 ageorgop joined #salt
18:14 toastedpenguin joined #salt
18:15 toastedpenguin joined #salt
18:16 gimpy973 I have a stateless HPC cluster where I'm trying to use Salt.  Since the compute nodes rebuild at each boot off of a single image, how should I handle keys?  One key in image shared by all nodes?  Can't seem to have them generate individual key on each boot as the second time a node boots the master rejects the new key as it expects the old one.
18:17 Antiarc If you set /etc/salt/minion_id to a new value on each boot, the master will see it as a new minion
18:18 Antiarc (ie, basename-$(something pseudorandom or serial))
18:18 baweaver_ joined #salt
18:18 whytewolf gimpy973: http://docs.saltstack.com/en/latest/topics/tutorials/preseed_key.html how about something that you use preseeded keys and upload a new key to the minion as part of the rebuild process
18:20 gimpy973 whytewolf: "upload a new key to the minion as part of the rebuild process" any Salt way to do that?  Only thing I can think of is have them pull it from a Website or git repo as part of a boot script...
18:22 whytewolf gimpy973: if the rebuilt server's key isn't accepted anymore than no. read the warning on the link i posted about security.
18:22 dingo_ gimpy973: use the mac address in a GET request of a central server that doles it out
18:22 nebuchad` joined #salt
18:23 dingo_ by http for all it matters
18:23 dingo_ or netcat-driven shell script hehe
18:23 dingo_ telnet with a tcl/expect script!
18:23 dingo_ you should hire me to work for you, gimpy973 !
18:25 zmalone gimpy973: For things like Rocks clusters, I'd automate the head node at most, and leave the clients alone.  If you do otherwise, you end up with a belt and suspenders approach where you are functionally running two configuration management systems.
18:25 sk_0_ joined #salt
18:25 h220hawk_ joined #salt
18:25 zmalone Rocks + Salt, in this case.
18:25 arapaho_ joined #salt
18:26 vexati0n_ joined #salt
18:26 peno_ joined #salt
18:26 froztbyt1 joined #salt
18:26 dh__ joined #salt
18:26 aqua^c joined #salt
18:27 mteufel joined #salt
18:28 _ikke__ joined #salt
18:28 winterblack joined #salt
18:29 ZombieTwiglet joined #salt
18:29 phpdave11_ joined #salt
18:29 Ymage_ joined #salt
18:29 SaveTheRb0tz joined #salt
18:29 rackermanfred joined #salt
18:29 _tr_h_ joined #salt
18:29 benvon_ joined #salt
18:29 venu0336 joined #salt
18:29 UForgotten_ joined #salt
18:29 eagles05- joined #salt
18:31 zmalone joined #salt
18:31 bfrog if I have a state file with nginx install, but then in another state file push a file into /etc/nginx/conf.d how can I reload nginx?
18:31 bfrog I have -watch: -file: /etc/nginx/conf.d/* but it fails
18:32 bfrog nvm... I figured it out
18:32 murrdoc you are welcome
18:33 iggy you have to include the other state
18:33 bfrog just noticed that
18:33 dh joined #salt
18:34 bfrog gotta say, salt is so much faster than ansible... used that at the last gig
18:34 druonysuse joined #salt
18:34 bfrog especially once states have been reached
18:37 kevinquinnyo joined #salt
18:38 mattrobenolt I remember reading this somewhere, but if I'm writing a custom state/pillar/whatever, what's the best practice for managing a global state? Like, say, a connection pool.
18:38 llb42 joined #salt
18:39 froztbyte joined #salt
18:39 kevinquinnyo hey all, I have a situation where when a new minion comes online i need to append it's private ip into a config file on each of the minions in a cluster.  I know i'll robably need to use the start event, so i've setup my master reactor.conf to listen for the start event.  But I don't know how i should handle the other part
18:39 mattrobenolt We have a few pillar that are fetching data from our Zookeeper, and the connection setup/teardown to Zookeeper is a bit heavy which makes running highstates slow. I'd like to just establish a connection and never release it.
18:40 kevinquinnyo should i have a pillar dict that just holds private ips, then push the new private ip into that dict, and then rebuild the config on each minion?
18:40 kevinquinnyo hope that made sense
18:44 baweaver_ joined #salt
18:45 dewdrop joined #salt
18:47 hal58th kevinquinny, sounds like you need to use salt-mine. But your other solution works as well to have a pillar file. You just have to remember to update it
18:48 hal58th mattrobenol, your question doesn't make much sense
18:48 mattrobenolt hal58th: Which part doesn't make sense? I'll happily clarify. :)
18:49 icflournoy joined #salt
18:49 hal58th What do you mean by global state
18:50 mattrobenolt I have multiple pillar files that rely on connection to our Zookeeper, and I want to share the connection between all runs.
18:50 mattrobenolt And ideally, the connection should just be kept alive while the salt-master is running at all times.
18:50 mattrobenolt Instead of connecting/disconnection every time.
18:50 iggy i.e. how would you do connection pooling in an ext_pillar
18:51 mattrobenolt ^
18:51 mattrobenolt I can just shove something at the module level.
18:51 iggy I'd ask basepi
18:51 mattrobenolt But for some reason, I thought there was another pattern.
18:51 hal58th Yeah I have no clue
18:51 mattrobenolt Some with a dunder variable that you could attach it to.
18:51 iggy that's not something I've heard of anyone solving (and from what I understand of the loader, I'm not sure you can)
18:52 mattrobenolt I swear I read something about it. :) But could be crazy.
18:52 iggy there's __context__, but that's for a state run
18:52 mattrobenolt Our problem is that a connect/disconnect from zookeeper isn't super fast. So I'd like to just establish one global connection.
18:52 murrdoc yeah there is no sharing
18:53 mattrobenolt If I shove it just at the module level... what'll that do?
18:53 iggy there's not an equivalent to __context__ for the life of the daemon
18:53 mattrobenolt I assume the LazyLoader will keep destroying and re-importing between runs?
18:53 iggy (much less some way to close that stuff down at shutdown)
18:53 mattrobenolt I mean, it's just python, so I can _make_ it happen for sure. :)
18:53 iggy but doesn't sound like a terrible FR
18:54 mattrobenolt What's a FR?
18:55 UForgotten joined #salt
18:56 mattrobenolt Oh, I'm guessing feature request.
18:58 iggy si
18:58 mattrobenolt I'll read source a bit and see if anything stands out. I _swear_ I saw something about this a while ago. :)
18:59 mattrobenolt Kinda hard to follow the global injections in modules, haha
18:59 impi joined #salt
19:00 gadget-girl how it going how whould i get started on build a salt modulte for arangodb
19:03 joeto joined #salt
19:06 UForgotten joined #salt
19:07 hybridpollo joined #salt
19:08 druonysus joined #salt
19:08 druonysus joined #salt
19:10 andygabby left #salt
19:12 murrdoc joined #salt
19:12 aparsons joined #salt
19:12 gadget-girl is /_modules in diff place on ubunut
19:13 iggy it's a salt standard, not anything to do with the distro
19:13 hal58th /srv/_modules most likely.https://docs.saltstack.com/en/develop/ref/modules/index.html
19:13 iggy <file_roots>/_modules
19:13 hal58th yeah sorry, /srv/salt/_modules
19:13 numkem joined #salt
19:14 iggy so by default, that'd be ^
19:14 gadget-girl hum it is not their
19:14 iggy but you also need to do a sync_modules (or sync_all) to have salt put them in the right place
19:14 iggy you create that dir
19:15 gadget-girl does the default install have modules
19:17 mohae joined #salt
19:17 iggy yes, but they aren't in _modules
19:17 iggy that's for your own modules
19:18 iggy the default ones are <python site packages dir>/salt/modules/
19:18 KyleG joined #salt
19:18 KyleG joined #salt
19:19 gadget-girl what should i name my module
19:19 gadget-girl i plan on commiting it to the project
19:21 Steven- joined #salt
19:22 kevinquinnyo hey all what's the best way to test quickly while building states.  I'd like to just be able to debug variables to the log file instead of running commands and such.  I've been using local.cmd to write variables to a tmp file.  Surely there's a better way?
19:24 scoates joined #salt
19:28 pguinardco kevinquinnyo: I've been using a burner environment that I can make in 2-3 seconds (FreeBSD jail) and just spamming salt-call state.highstate when testing changes. As an added bonus I blow away the environment as soon as I'm done.
19:30 sinh_ joined #salt
19:31 iggy gadget-girl: whatever makes the most sense (make sure it doesn't conflict with any existing python modules)
19:32 gadget-girl thanks iggy i called it arangodb.pyx
19:32 iggy is it not written in python?
19:32 gadget-girl you know any module that use the cli that use for example
19:33 iggy I don't think salt's loader will look for anything other than .py files fwiw
19:33 gadget-girl http://docs.saltstack.com/en/latest/ref/modules/index.html#modules-are-easy-to-write said pyx
19:35 iggy "If adding a Cython module"
19:36 Sniper^ joined #salt
19:36 gadget-girl is that what i am doing
19:36 andrew_v joined #salt
19:36 iggy probably not
19:37 murrdoc sinh:  fu
19:37 sinh ead
19:40 tvinson can winrepo pull in pillar information for install_flags?
19:42 jaybocc2 joined #salt
19:42 dingo_ 19:22 < kevinquinnyo> hey all what's the best way to test quickly while building states.  I'd like to just be able to debug variables to the log file instead of running commands and such.  I've been using local.cmd to write variables to a tmp file.  Surely there's
19:43 dingo_ your answer may be in 19:22 < kevinquinnyo> hey all what's the best way to test quickly while building states.  I'd like to just be able to debug variables to the log file instead of running commands and such.  I've been using local.cmd to write variables to a tmp file.  Surely there's
19:43 dingo_ ahh shucks, https://github.com/joejulian/saltstack-logging
19:43 dingo_ (computers are hard, particluarly pasting)
19:44 dingo_ (in fact, almost all of my code is just copy and paste, that's when i'm really providing the best value)
19:45 nocturn joined #salt
19:46 mbrgm_ joined #salt
19:47 kevinquinnyo thanks dingo_ checking
19:47 dfinn left #salt
19:48 dopesong joined #salt
19:48 aboe joined #salt
19:49 dopesong_ joined #salt
19:49 cwyse joined #salt
19:49 beardo joined #salt
19:49 ronrib joined #salt
19:49 pppingme joined #salt
19:50 baweaver_ joined #salt
19:51 gadget-girl does this make sence https://gist.github.com/sarasfox/1f6a130f09223b8a53ae
19:52 dingo_ not bad, did you try it?
19:52 gadget-girl how do i try it
19:52 jcristau joined #salt
19:52 ekristen joined #salt
19:52 hax404 joined #salt
19:52 Ssquidly joined #salt
19:52 ska-fan joined #salt
19:52 mens joined #salt
19:52 mfournier joined #salt
19:52 nahamu joined #salt
19:52 jchadwick joined #salt
19:52 davisj joined #salt
19:52 sarlalian joined #salt
19:53 dingo_ app_mount_path = os.path.join(app, mount)
19:53 dingo_ little nicks, but whatever, well, lets see,
19:53 dingo_ you have a sub-folder, 'pillars', and another, 'states'
19:53 gadget-girl yes
19:53 dingo_ at the same level, add a '_modules' folder, if i recall
19:53 dingo_ anyway i saw you reference some documentation, it is a better source than I :)
19:53 dingo_ read 'em top to bottom, take a nap, that's what i do
19:54 Ssquidly joined #salt
19:54 dingo_ if you name this file 'gg.py', then from your salt state, you would test it like so:
19:54 dingo_ ahh shucks, let me use gist, too
19:54 DammitJim joined #salt
19:54 LinuxHorn joined #salt
19:56 JonGretar joined #salt
19:56 joeto1 joined #salt
19:56 dingo_ gadget-girl: see
19:56 dingo_ err.. my reply.. on github...
19:56 OGuilherme joined #salt
19:56 dingo_ i can't paste
19:56 wiqd joined #salt
19:57 dingo_ https://gist.github.com/sarasfox/1f6a130f09223b8a53ae#comment-1492526
19:58 zipkid joined #salt
19:58 stoogenmeyer joined #salt
19:58 dh joined #salt
20:02 OliverMT joined #salt
20:03 napsterX joined #salt
20:03 imanc joined #salt
20:04 whytewolf um ... dingo _states and _modules... two different things. a module needs to be accessed with the cli. or from a _state
20:05 akitada joined #salt
20:05 whytewolf [about your comment in https://gist.github.com/sarasfox/1f6a130f09223b8a53ae#comment-1492526 ]
20:05 dingo sorry, where i said 'states/' i meant 'roots/'
20:06 dingo i edited the comment, maybe helps
20:06 whytewolf actually more to the point. create an sls that calls something in _modules won't work
20:07 whytewolf even after sync_all
20:07 neocid joined #salt
20:09 murrdoc joined #salt
20:11 bfrog what does this actually mean? State 'git.latest' was not found in SLS
20:11 bfrog I basically copied the salt.states.git example and set my own stuff
20:12 gadget-girl arangodb.add_app_store' was not found in SLS 'states.base.arangodb hum
20:13 bfrog oh I guess I need the git pkg installed
20:13 gadget-girl _modules is on the same level as pillars and states
20:13 whytewolf gadget-girl: that was basicly the error i was expecting since creating a _module does not mean you created a _state also. you have cli access but not state access
20:14 gadget-girl how would test it form the cli
20:14 gadget-girl or what is _state
20:15 whytewolf salt 'minion' arangodb.add_app_store app=my-app-folder mount=my-app-mount
20:15 aqua^c joined #salt
20:16 whytewolf you do need to make sure that you sync first
20:16 gadget-girl sync hum
20:16 gadget-girl salt-sync
20:16 whytewolf salt '*' saltutil.sync_all
20:17 whytewolf and _states is another directory like _modules except it holds states instead of modules
20:17 murrdoc joined #salt
20:18 whytewolf http://docs.saltstack.com/en/latest/ref/file_server/dynamic-modules.html
20:18 keimlink joined #salt
20:19 gadget-girl is _states diff for just states
20:20 juanito joined #salt
20:21 whytewolf yes _states is just for user created state.function files
20:21 repl1cant joined #salt
20:22 whytewolf tpyically you create a module. then create a states that make calls into that module.
20:22 icflournoy joined #salt
20:22 giantlock joined #salt
20:23 whytewolf [which is one of the reasons there are way more modules than there are states
20:23 gadget-girl so states in _states can can call _modules
20:23 whytewolf yes. just like your _module called cmd.run
20:24 whytewolf [which is a built in module]
20:24 murrdoc joined #salt
20:25 gadget-girl 'arangodb' __virtual__ returned False
20:25 gadget-girl any idea
20:26 teryx510 joined #salt
20:26 whytewolf create a gadget-girl add a function called __virtual__ that returns true
20:28 baweaver_ joined #salt
20:29 whytewolf gadget-girl: you could also do some logic in the __virtual__ function that decides if it should run or not, such as making sure that foxx-manger' exists before allowing that module to run
20:30 gadget-girl arangodb.add_app_store' was not found in SLS '_states.base.arangodb still getting this uglyness
20:31 whytewolf gadget-girl: did you just move your _modules into a _states folder and try from the sls?
20:31 zmalone joined #salt
20:32 gadget-girl no _states and _modules are at the same level
20:32 whytewolf gadget-girl: but you put the module you wrote into _states?
20:34 gadget-girl still not working
20:35 whytewolf gadget-girl: are you syncing after you change things?
20:35 gadget-girl yes
20:35 whytewolf is the syncing showing that it moved your module/state to the minions?
20:35 gadget-girl yes it did
20:36 druonysuse joined #salt
20:36 druonysuse joined #salt
20:36 whytewolf gadget-girl: are you still trying this as a _modules or a _states
20:38 whytewolf ohhhhh. I think i get the flaw. states and sls files ARE NOT the same thing.
20:39 whytewolf moving that sls file into the _states directory did nothing
20:39 druonysuse joined #salt
20:39 whytewolf _states are python code. same as _modules
20:40 whytewolf that are called from sls files
20:40 whytewolf [naming confusion cause both are called states]
20:41 gadget-girl https://drive.google.com/file/d/0BzuZO56UhRRnZ3dvdXptOVo4NXM/view?usp=sharing whytewolf here take a look
20:42 onewheelskyward left #salt
20:44 gadget-girl wow you have me lose whytewolf
20:46 geekatcmu yes, the naming thing is rage-worthy
20:48 subsignal joined #salt
20:48 whytewolf geekatcmu: I bet it gets worse with none english speakers
20:49 geekatcmu no doubt
20:50 geekatcmu calling the user formulas "states" was, TBH, a stupid idea in the face of these internal states.
20:51 amcorreia_ joined #salt
20:51 gadget-girl whytewolf you are you will hope in google hangout help me
20:54 whytewolf gadget-girl: sorry. that is something i can not do. my work prohibits
20:54 gadget-girl ok
20:54 whytewolf gadget-girl: https://drive.google.com/file/d/0BxdCCD5Q8kRAQ0JTNHktZFRrTXc/view?usp=sharing here is a fixed up version.
20:55 whytewolf I'm already breaking the rules by being on IRC
20:55 whytewolf even on my own computer
20:56 gadget-girl oh
20:56 gadget-girl that suckes
20:56 whytewolf banking red tape sucks
20:56 gadget-girl got you
20:56 yomilk joined #salt
20:57 whytewolf anyway. I moved some things around. and added a __virtual__ function. didn't really test it so all bets are off. try from the cli after syncing
20:59 s_kunk joined #salt
21:00 gadget-girl still given me false
21:01 whytewolf __virtual__ should not be returning false. on the minion try salt-call arangodb.add_app_store app=my-app-folder mount=my-app-mount
21:05 tvinson ah winrepo looks like it's still missing this per bug 6859
21:05 napsterX joined #salt
21:07 tvinson or it was just added two weeks ago, hooray
21:07 gadget-girl https://gist.github.com/sarasfox/eb64b73bdc3a7937e2a9 whytewolf
21:08 whytewolf doh
21:08 whytewolf forgot the ():
21:09 whytewolf gadget-girl: just tack (): on the end of __virtual__
21:09 UtahDave joined #salt
21:10 UtahDave Gareth: hey, man. you around?
21:11 * iggy pokes UtahDave about atom-salt
21:11 juanito joined #salt
21:16 hal58th_ joined #salt
21:18 X67r joined #salt
21:18 ninjada joined #salt
21:20 Gareth UtahDave: I am.
21:20 ericw joined #salt
21:20 murrdoc how does retcode work
21:20 murrdoc salt minion1 cmd.retcode 'echo && return 2 || return 3' python_shell=True
21:20 murrdoc minion1:
21:20 murrdoc 1
21:21 murrdoc ??
21:21 Gareth murrdoc: return code of the command that you pass right?
21:22 murrdoc yeah
21:22 murrdoc https://github.com/saltstack/salt/pull/25424
21:22 saltstackbot [#25424]title: return will not work in the context it is being used in | return will always throw an error when called in the context it's being used in the example - should be exit....
21:23 murrdoc saltstackbot:  ?
21:23 murrdoc saltstackbot:
21:23 murrdoc wtf … where the hepl at
21:25 baweaver_ joined #salt
21:27 evilrob joined #salt
21:28 evilrob joined #salt
21:29 gadget-girl whytewolf got it work now how do call these form normal states
21:30 hybridpollo joined #salt
21:31 whytewolf gadget-girl: thats what the _states directory is for. making states. available to states
21:31 gadget-girl so _states should match _modules
21:32 whytewolf gadget-girl: in thoery. it needs to understand the state of something so that every highstate it doesn't run unless it needs to.
21:37 cowpunk21 joined #salt
21:38 gadget-girl https://gist.github.com/sarasfox/eb64b73bdc3a7937e2a9 now this is nice
21:38 cowpunk21 joined #salt
21:40 murrdoc joined #salt
21:40 cowpunk21 joined #salt
21:41 bfrog when I run salt '*' state.highstate on my master machine, does it hit each minion concurrently?
21:41 bfrog it seems like it works on them in series
21:42 Edgan How do I make the saltenv something other than base for salt-ssh?
21:44 gadget-girl where &*^% is this name crap coming from
21:45 whytewolf gadget-girl: a state function needs name as it is passed in from the state call.
21:46 whytewolf you can pass it in and safuly ignore it
21:46 whytewolf or pass it in as app
21:47 whytewolf [the variable you are calling app]
21:49 otter768 joined #salt
21:54 gadget-girl i switched with mount
21:55 whytewolf gadget-girl: okay, now in your state sls file mount is now name. or you can just use it as the state id [the line at the stop of a state block
21:55 gadget-girl yes
21:56 zmalone joined #salt
21:56 gadget-girl that make the most cents in my mind that name the app is installed as in the arangodb
22:00 bfrog what if I need some salt states to finish on one machine before some can be run on another?
22:01 whytewolf bfrog:  use orchenstration in that case.
22:01 illern joined #salt
22:02 jmreicha joined #salt
22:02 bfrog ah gotcha
22:04 aqua^c joined #salt
22:04 gadget-girl https://gist.github.com/sarasfox/652fbcba9385c83f0fec here what it looks like now
22:06 jmreicha_ joined #salt
22:06 napsterX joined #salt
22:07 whytewolf gadget-girl: okay, one thing you will notice in it's current status. is that it will try to do that every run of highstate. which is fine for a module. but for a state it should really check that the function needs to run
22:08 gadget-girl yes i need to foxx-manger list look for it
22:08 baweaver_ joined #salt
22:18 ninjada joined #salt
22:19 yomilk joined #salt
22:24 Singularo joined #salt
22:35 yomilk joined #salt
22:37 druonysuse joined #salt
22:37 druonysuse joined #salt
22:40 neocid joined #salt
22:43 murrdoc 2015-07-14 22:16:34,935 [salt.pillar      ][CRITICAL][21139] Specified ext_pillar interface pillar_name is unavailable
22:43 murrdoc wtf mate
22:43 amcorreia_ joined #salt
22:59 timoguin joined #salt
22:59 cwyse joined #salt
23:00 mosen joined #salt
23:00 cwyse joined #salt
23:01 primechuck joined #salt
23:07 napsterX joined #salt
23:10 CheKoLyN joined #salt
23:11 ninjada joined #salt
23:12 baweaver_ joined #salt
23:13 keimlink_ joined #salt
23:18 keimlink joined #salt
23:22 jhauser joined #salt
23:27 baweaver_ joined #salt
23:27 gadget-girl is their any good way to pass confire option though to cli
23:29 gadget-girl my only issue i like keep the install form store genric
23:29 LeProvokateur joined #salt
23:33 baweaver_ joined #salt
23:37 Zachary_DuBois joined #salt
23:37 helgi joined #salt
23:38 chamunks joined #salt
23:39 lahwran joined #salt
23:40 Ryan_Lane wtf is this? https://github.com/saltstack/salt/blob/2015.5/salt/state.py#L181-L209
23:40 Ryan_Lane specifically this: https://github.com/saltstack/salt/blob/2015.5/salt/state.py#L200-L209
23:41 Ryan_Lane now in random places in my execution module output I'm seeing "Installed Packages:", when there's absolutely nothing going on with packages
23:41 zmalone joined #salt
23:42 mosen wtf indeed
23:42 mosen not even reading through the code, there's some major abstraction horribleness going on :)
23:43 baweaver_ joined #salt
23:44 snarfy joined #salt
23:45 evilrob I'm trying to debug a state file with pillar vars from a master with multiple roots set.  How do I specify a different root in the salt:// URL to do a salt-call cp.get_template ??
23:46 jespada__ joined #salt
23:46 spookah joined #salt
23:47 evilrob salt://foo:state/init.sls ?
23:48 sunkist joined #salt
23:50 otter768 joined #salt
23:50 iggy evilrob: as far as the minion is concerned, there's one root
23:52 aqua^c joined #salt
23:54 ndrei joined #salt
23:54 cowpunk21 joined #salt
23:58 ecdhe joined #salt
23:59 CheKoLyN joined #salt
23:59 Ryan_Lane evilrob: I'm not sure you can

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