Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2017-07-25

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

All times shown according to UTC.

Time Nick Message
00:05 Ni3mm4nd joined #salt
00:13 dev_tea joined #salt
00:16 dev_tea joined #salt
00:17 fritz09 joined #salt
00:19 dev_tea joined #salt
00:23 dev_tea joined #salt
00:27 onlyanegg joined #salt
00:37 debian112 joined #salt
00:43 Renich joined #salt
00:56 debian1121 joined #salt
01:16 dev_tea joined #salt
01:18 Guest73 joined #salt
01:29 Guest73 joined #salt
01:30 swills joined #salt
01:30 swills joined #salt
01:36 dev_tea joined #salt
01:37 dev_tea joined #salt
01:42 cro joined #salt
01:50 noobiedubie joined #salt
01:51 ilbot3 joined #salt
01:51 Topic for #salt is now Welcome to #salt! <+> Latest Versions: 2016.11.6, 2017.7.0 <+> Support: https://www.saltstack.com/support/ <+> Logs: http://irclog.perlgeek.de/salt/ <+> Paste: https://gist.github.com/ <+> See also: #salt-devel, #salt-offtopic <+> We are volunteers and may not have immediate answers <+> The call for speakers for SaltConf17 is now open: http://tinyurl.com/SaltConf17
02:01 hasues joined #salt
02:06 zerocoolback joined #salt
02:06 Ni3mm4nd joined #salt
02:07 zerocool_ joined #salt
02:26 Guest73 joined #salt
02:31 jimklo_ joined #salt
02:49 dxiri joined #salt
02:57 justanotheruser joined #salt
03:01 justanotheruser joined #salt
03:07 Ni3mm4nd joined #salt
03:07 donmichelangelo joined #salt
03:10 vishvendra joined #salt
03:27 Guest73 joined #salt
03:35 cliluw joined #salt
03:39 dxiri joined #salt
04:05 dxiri joined #salt
04:08 Ni3mm4nd joined #salt
04:12 miruoy joined #salt
04:16 miruoy joined #salt
04:17 beardedeagle joined #salt
04:18 btorch left #salt
04:28 jeddi joined #salt
04:47 svij3 joined #salt
04:51 tobstone joined #salt
05:02 preludedrew joined #salt
05:09 golodhrim|work joined #salt
05:21 kyu joined #salt
05:23 systeem joined #salt
05:23 high_fiver joined #salt
05:24 Bock joined #salt
05:29 mavhq joined #salt
05:30 Tgrv joined #salt
05:32 hemebond joined #salt
05:38 yesimon joined #salt
05:39 yesimon How do I handle having the same pkg.installed in two different .sls files?
05:40 yesimon I want these .sls files to be able to be standalone as well
05:40 high_fiver joined #salt
05:43 Rubin joined #salt
05:48 hemebond yesimon: Give the states different names/IDs.
05:49 lompik joined #salt
05:50 ProT-0-TypE joined #salt
05:54 eseyman joined #salt
05:57 do3meli joined #salt
05:57 pualj joined #salt
05:58 do3meli left #salt
06:04 kyu left #salt
06:05 kyru joined #salt
06:09 Ni3mm4nd joined #salt
06:09 aldevar joined #salt
06:12 kyru We were thinking about writing some kind of 'diff-outputter'. Given an arbitrary execution-module call, the outputter would display only the differences of the returned data.
06:13 ivanjaros joined #salt
06:14 kyru In our use case we want to observe the versions of installed packages of each of our nodegroups
06:15 kyru Since we thought, a generic diff outputter might be applicable to many use cases, we were thinking about contributing the outputter to the project.
06:17 kyru But before we create an issue on github, we wanted to get some feedback from the community about such a diff outputter. Do you think it would be accepted and also: Does it fit into Salt? With this I mean, that it would not violate Salt's approach of outputters, since the diff outputter would parse the data in some way.
06:17 dxiri joined #salt
06:17 kyru And not only transform it in order to display it nicely
06:17 kyru What do you think about this?
06:25 jimklo joined #salt
06:26 debian112 joined #salt
06:36 _KaszpiR_ joined #salt
06:38 dxiri joined #salt
06:39 Guest73 joined #salt
06:40 darioleidi joined #salt
06:52 apofis joined #salt
06:53 dxiri joined #salt
07:00 high_fiver joined #salt
07:04 hemebond kyru: How would that work when the outputters are run by the individual minions?
07:05 hemebond Oh wait.
07:05 hemebond They're also used by the salt CLI.
07:05 hemebond Sounds more like something you'd pipe into.
07:05 dxiri joined #salt
07:07 kyru The outputter would not work on single minions as it would need return data from more than one minion. As such it would only work when run on the master.
07:08 kyru hemebond: So you would rather propose to just pipe it into some external script than creating an outputter for it?
07:08 hemebond Yeah, so doesn't really fit with outputters.
07:09 hemebond It certainly sounds like something you would pipe to.
07:09 hemebond Like setting the output to json and then piping into your tool.
07:09 Ricardo1000 joined #salt
07:12 kyru hemebond: Seems legit. Would you recommend to use a completely external tool? I feel like writing some custom extension to salt would be the better approach.
07:12 hemebond Hmm. Would piping work when each result is output individually?
07:13 hemebond Well, when I want to do something with the returned data I use jq
07:16 kyru hemebond: I'll have a look at jq. Regarding your question: If you pipe the whole output, it could work I think. As of now, with the outputter, I just used --static to have a single large output.
07:16 hemebond Ooh, I haven't used that option.
07:16 beardedeagle kyru & hemebond in that case, I would think you would have salt return a json dict and pass it to the script
07:17 beardedeagle or something like that
07:19 beardedeagle err, `--out=json --static`
07:20 kyru Hmm I guess that would make sense... We wanted to use an outputter to 'stay inside salt', since the output would later be displayed by Jenkins running a groovy script with that diff command. But now I think, that simply parsing the output inside the groovy script might make more sense, if an outputter would be the unpreffered approach
07:20 jeddi joined #salt
07:20 beardedeagle kyru: or alternatively instead of it being an outputter or script you put a pr in to add a `--diff` to the salt cli command
07:20 beardedeagle and it would imply `--out=json --static` then do it's thing to the data
07:21 beardedeagle that would be the route I take
07:21 kyru That seems even better
07:23 kyru beardedeagle: Hmm. If I wanted to extend the CLI of Salt, then I need to dive deeper into Salt, just started with it a week ago. Can you maybe give me a heads up where to start? Maybe something in the docs?
07:23 Hybrid joined #salt
07:25 Ni3mm4nd joined #salt
07:25 beardedeagle kyru: I'm not aware of any docs around this particular subject, I usually use the code of the docs anyway: https://github.com/saltstack/salt/blob/develop/salt/cli/salt.py
07:26 high_fiver joined #salt
07:27 beardedeagle code instead of the docs*
07:30 kyru beardedeagle & hemebond: Thank you both for your input! I will have a further look into this and see, if we will choose to try and extend the Salt CLI. :)
07:34 dxiri joined #salt
07:37 CrummyGummy joined #salt
07:37 justanotheruser joined #salt
07:43 jimklo joined #salt
07:44 inad922 joined #salt
07:44 cablekevin joined #salt
07:48 dxiri joined #salt
07:54 zulutango joined #salt
07:55 ronnix joined #salt
07:56 ekkelett joined #salt
08:00 pbandark joined #salt
08:00 mikecmpbll joined #salt
08:00 Rumbles joined #salt
08:02 dxiri joined #salt
08:05 joe_n joined #salt
08:14 XenophonF joined #salt
08:15 ccha restart a service is weird with service.running. How can about about reload instead of restart for some service ?
08:16 viq JPT: today's update to py27-requests or py27-libcloud fixed salt_minion startup for me
08:18 Rumbles joined #salt
08:25 jhauser joined #salt
08:26 ivanjaros joined #salt
08:29 pbandark hi. is there any salt state module for mariadb ?
08:31 kedare joined #salt
08:31 ashmckenzie joined #salt
08:34 tedski joined #salt
08:34 Mattch joined #salt
08:35 dxiri joined #salt
08:35 Kruge joined #salt
08:37 Rumbles joined #salt
08:38 tvinson joined #salt
08:39 tedski joined #salt
08:40 carmony joined #salt
08:41 shadoxx joined #salt
08:43 trent_ joined #salt
08:44 haam3r_ pbandark: The mysql states won't work on mariadb?
08:45 pbandark haam3r_: not sure. i have not tired. i am creating both environment. mariadb and mysql. not sure if I can use same "pymsql" for both. I need confirmation on the same.
08:48 doradus joined #salt
08:49 haam3r_ pbandark: well a quick google tells me that mariadb provides python support with the mysql-python package
08:51 pbandark so "pymyql" pypi package and "mysql-python" are different ?
08:54 haam3r_ I get the impression that Salt expects the mysql-python package and mysql and mariadb both use that one to provide python support
08:56 pbandark ok
08:59 ronnix joined #salt
09:02 doradus Hi, on minon with saltversion 2016.3.4 I wrote  {% if grains['saltversion'] < '2016.11.0' %}  but the {% else %} block gets evaluated. Is there another way to check saltversion?
09:04 dimeshake joined #salt
09:04 dimeshake joined #salt
09:17 dxiri joined #salt
09:20 Naresh joined #salt
09:26 joe_n joined #salt
09:29 twooster There's saltversioninfo
09:30 twooster doradus: ^ you might be able to do `if grains['saltverisoninfo'] < ('2016', '11', '0')`
09:31 twooster I'm not sure if saltversioninfo is a tuple or not, though; it might be a list. If it's a tuple, the above will work. Otherwise try `if tuple(grains['saltversioninfo']) < (2016, 11, 0)`
09:31 twooster (not strings as I did previously, that was a mistake)
09:36 dxiri joined #salt
09:42 doradus Thanks Twooster, I'll try your suggestion..
09:44 JPT viq: I just tried it and it seems to be fine now. :-)
09:53 gmoro joined #salt
09:59 pppingme joined #salt
10:02 pppingme joined #salt
10:04 dxiri joined #salt
10:09 pppingme joined #salt
10:10 N-Mi_ joined #salt
10:14 joe_n joined #salt
10:15 pppingme joined #salt
10:15 doradus if tuple(grains['saltverisoninfo']) ...  produces SLS error as it thinks tuple is undefined jinja variable.
10:21 pppingme joined #salt
10:25 Guest73 joined #salt
10:26 twooster hmm, dunno if there's a filter. i swear i saw tuple in the jinja context at some point.
10:27 twooster ah, okay, so: if saltversioninfo is a list -- then you need to do: grains['saltversioninfo'] < [2016, 11, 0]
10:27 twooster if it's a tuple, then parentheses
10:29 joe_n joined #salt
11:05 dxiri joined #salt
11:19 dxiri joined #salt
11:28 _JZ_ joined #salt
11:36 oister joined #salt
11:42 _JZ_ joined #salt
11:43 doradus I'll dig into this, thanks Twooster!!
11:46 bildz joined #salt
11:52 dxiri joined #salt
12:02 tapoxi joined #salt
12:10 J0hnSteel joined #salt
12:19 J0hnSteel joined #salt
12:20 impi joined #salt
12:25 Hybrid joined #salt
12:27 XenophonF pbandark: I use Salt to manage MariaDB instances (specifically, mysql-formula).
12:28 XenophonF pbandark: https://mariadb.com/kb/en/mariadb/mariadb-vs-mysql-compatibility/
12:28 XenophonF short version is that mariadb is a drop-in replacement for mysql
12:29 pbandark XenophonF: ok. so if I am not wrong then I need to use "pymysql" to manage mariadb right ?
12:29 numkem joined #salt
12:30 numkem joined #salt
12:31 oister left #salt
12:33 XenophonF In general, if you want a Salt execution module to do something, you need to install the dependencies listed in that module's documentation.
12:33 pbandark XenophonF: also, in order to secure mariadb server(mysql_secure_installation), I am planning to execute queries(https://paste.fedoraproject.org/paste/jCTBk9IOWaVe1d2hLeWlZg) with `mysql_query.run`.  is there any other approach?
12:33 pbandark checking
12:33 XenophonF According to https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.mysql.html#module-salt.modules.mysql, that's whatever your operating system calls the MySQLdb Python module.
12:34 XenophonF so on RHEL/CentOS, that's MySQL-python
12:34 XenophonF and on FreeBSD (yay!), that's py27-MySQLdb
12:34 XenophonF etc.
12:35 XenophonF I recommend you use mysql-formula and save yourself a lot of work.
12:36 pbandark XenophonF: do you mean https://github.com/saltstack-formulas/mysql-formula  ?
12:36 XenophonF That's the one.
12:36 pbandark ok. let me check
12:37 XenophonF There's a mysql.remove_test_database SLS that does the same stuff as the secure install script.
12:38 XenophonF And all that does is set an admin password and drop the test DB.
12:38 XenophonF (IIRC)
12:38 XenophonF I use mysql-formula to manage both my own MariaDB databases as well as MySQL RDS instances.
12:39 pbandark ok
12:39 pbandark thanks a lot for info XenophonF
12:39 XenophonF np
12:40 xet7 joined #salt
12:40 XenophonF Let me know if you want to see some real configs, and I'll tidy up one of my production configs for you.
12:41 pbandark that would be great help XenophonF
12:42 pbandark it will help me to create my configuration.
12:43 XenophonF OK
12:49 dxiri joined #salt
12:54 tacoboy joined #salt
12:57 mavhq joined #salt
13:01 J0hnSteel joined #salt
13:05 do3meli joined #salt
13:05 do3meli left #salt
13:07 dxiri joined #salt
13:12 J0hnSteel joined #salt
13:21 dxiri joined #salt
13:26 cgiroua joined #salt
13:26 racooper joined #salt
13:32 lorengordon joined #salt
13:33 jhujhiti_ joined #salt
13:34 watersoul_ joined #salt
13:35 pmcg joined #salt
13:35 KevinAn2757 joined #salt
13:35 dxiri joined #salt
13:36 ws2k3 joined #salt
13:36 cro joined #salt
13:36 marcinkuzminski joined #salt
13:36 brent joined #salt
13:36 karlthane joined #salt
13:36 shortdudey123 joined #salt
13:39 Zachary_DuBois joined #salt
13:39 chadhs joined #salt
13:40 sjohnsen joined #salt
13:40 tom29739 joined #salt
13:52 cablekevin joined #salt
13:54 msid joined #salt
13:56 msid Anyone has any idea what could be causing "Invalid public ssh key, most likely has spaces or invalid syntax" when trying to add SSH public keys for a user? The irony is, the same key applies without a problem on other states, but not on all :/
13:57 promorphus joined #salt
13:59 promorphus joined #salt
14:02 lorengordon joined #salt
14:02 drawsmcgraw joined #salt
14:08 mike25de joined #salt
14:12 mike25de hi all
14:14 mike25de guys is there a chance to have an offline installer for salt-minion... maybe with a specific version? The issue is that when i install a salt-minion using the bootstrap, i get always the latest minion version... which might be higher than the master
14:14 mike25de thanks in advance
14:14 twooster mike25de: you can specify the version, which is what i do right now
14:15 twooster i think offline is a bit tricky due to all the deps it might have, which would require an internet connection in most cases
14:15 mike25de twooster: yeah... but i had the impression that... only 1-2 revisions are available... i read it somewhere
14:15 twooster mm, i kind of doubt it? you're using git tags
14:15 twooster but maybe i'm wrong
14:16 mike25de twooster: ... i think only in centos... that happens... when i install from the centos repos... you might have saved the day buddy
14:16 twooster give it a shot from the bootstrap script: -P git v2016.11.6 or whatever
14:16 mike25de thanks twooster !! i will test it now!
14:16 twooster maybe it depends on the remote repos even still, but i couldn't say for sure. good luck. :)
14:17 mike25de cheers mate
14:17 mike25de have a good day/night wherever you are :P
14:22 dxiri joined #salt
14:24 darioleidi joined #salt
14:26 dxiri_ joined #salt
14:27 noobiedubie joined #salt
14:29 edrocks joined #salt
14:30 jhujhiti joined #salt
14:32 doradus Twooster, the syntax grains['saltversioninfo'] < [2016, 11, 0] works, thanks.  It did not solve the problem, still false.  Maybe there is bug in Salt 2016.3.4.84.13 / python-Jinja 2.2.8-71.1   ... I have seen pattern {% if grains['saltversion'] < '2016.11.0' %}  appear in various formulas - should work..
14:33 beardedeagle joined #salt
14:35 mikecmpbll joined #salt
14:40 seanz joined #salt
14:44 Guest73 joined #salt
14:54 doradus Hi mike25de,  Related to your question, on OpenSuse when i install a salt-minion using the bootstrap, i get always the old version of Salt (2016.3).
14:56 mikecmpbll joined #salt
14:57 impi joined #salt
15:09 tiwula joined #salt
15:10 mikea joined #salt
15:13 Guest73 joined #salt
15:16 twooster doradus: seems weird. you've double checked grains.items on the minion?
15:19 aldevar left #salt
15:22 daks joined #salt
15:23 bowhunter joined #salt
15:24 MTO joined #salt
15:25 jimklo joined #salt
15:29 zmalone joined #salt
15:30 lorengordon joined #salt
15:32 doradus twooster, I will check again now.
15:44 ronnix joined #salt
15:49 Reverend gtmanfred - you fancy giving me a poke when those Orch tgt lines are fixed on release? :)
15:49 Reverend we've `yum downgrade`ed on our boxes for now :)
15:49 Reverend <3
15:50 doradus Twooster, my mistake:  {%if grains['saltversioninfo'] < [2016, 11, 0]  %} now evaluates to true on Salt 2016.3.4 which is good news.  But {% if grains['saltversion'] < '2016.11.0' %}  returned false?   Thanks for your help on this.
15:52 gtmanfred Reverend: no? follow the salt-announce mailing list.
15:52 gtmanfred dxiri_: awesome!
15:56 KingOfFools joined #salt
15:57 KingOfFools left #salt
15:57 KingOfFools joined #salt
15:58 gtmanfred Reverend: which issue?
16:00 KingOfFools Hey guys. I want to periodically get some data from postgres queue, run some states with that data on certain minions and delete data from queue. How would you suggest to implement that?
16:00 woodtablet joined #salt
16:00 Reverend gtmanfred oh yeah, that's a good idea. :D
16:01 Reverend gtmanfred there's an issue with orchestration not passing targets to the state applies :(
16:01 Reverend lemme dig
16:01 Reverend tbh, i could watch that git issue
16:02 Reverend https://github.com/saltstack/salt/issues/42403
16:02 Guest73 joined #salt
16:02 Reverend I think
16:03 gtmanfred Reverend: that will be fixed in 2017.7.1. we will announce it on salt-announce
16:03 Reverend yay! <3 heroic. I did just see that you'd merged int hat PR anyway. GG. I'll wait for the next release then and test it out. :D I fixed my server snapshots too, so it won't be too diffimacult to rollback if it borks it xD
16:05 elektrix_ joined #salt
16:08 onlyanegg joined #salt
16:09 rome_390 joined #salt
16:10 nledez joined #salt
16:10 nledez joined #salt
16:10 MTO joined #salt
16:11 sarcasticadmin joined #salt
16:14 nkuttler joined #salt
16:16 beardedeagle joined #salt
16:18 JPaul joined #salt
16:19 jimklo joined #salt
16:20 dendazen joined #salt
16:21 MTO joined #salt
16:23 SamYaple gtmanfred: lets talk openstack
16:24 SamYaple gtmanfred: ive heard talk of openstack modules and states for saltstack using shade
16:24 SamYaple im doing that internally right now
16:24 SamYaple whats the communities progress?
16:24 gtmanfred SamYaple: I am working on a salt-cloud driver
16:25 SamYaple would you like help?
16:25 gtmanfred once that was done, I was going to work on a whole bunch of openstack drivers, but call them like novang and keystoneng
16:25 gtmanfred yes please!
16:25 SamYaple ive been using and working on shade since inception and did a bunhc of ansible stuff for it
16:25 gtmanfred yar
16:25 SamYaple can you point me to some PR/roadmap/code that you would like help with?
16:26 t0m0 joined #salt
16:26 SamYaple or is this all from scratch stuff?
16:27 gtmanfred this is all planning to be from scratch
16:27 gtmanfred https://github.com/saltstack/salt/issues/30272
16:29 edrocks joined #salt
16:30 colabeer joined #salt
16:30 tacoboy joined #salt
16:30 SamYaple ok so youll want novang keystoneng glanceng etc modules?
16:30 gtmanfred yeah
16:30 gtmanfred and then in 2 releases, we will deprecate the other ones
16:31 SamYaple will novang then be renamed nova?
16:31 gtmanfred yes
16:31 SamYaple ok
16:31 SamYaple i think i can pump these out for you
16:31 gtmanfred and aliased to still allow novang to be used.
16:31 gtmanfred awesome :+1:
16:31 gtmanfred thanks!
16:31 gtmanfred you are the best!
16:31 gtmanfred SamYaple: you also might be interested in this https://groups.google.com/forum/#!forum/salt-cloud
16:31 gord0 joined #salt
16:32 SamYaple i need to start with keystone because thats what i care about though. if we can get an agreed on structure for that then the others should be straight forward
16:32 gtmanfred :+1:
16:32 gtmanfred yup
16:32 gtmanfred sounds good!
16:33 * SamYaple is just going to go rip off his other code
16:33 gtmanfred doit!
16:33 lorengordon joined #salt
16:38 froztbyte joined #salt
16:38 froztbyte joined #salt
16:43 aneeshusa joined #salt
16:46 lorengordon joined #salt
16:48 socket-_ Hey all, our AMI factory installs salt, and all instances created from these AMI's share the same minion_id but have unique hostnames. I was wondering what's the best way to handle ensuring the client running salt has a minion_id matching the current hostname.
16:49 nledez joined #salt
16:49 vishvendra joined #salt
16:50 MTecknology delete the minion_id file or edit appropriately during deployment or template creation
16:52 socket-_ well my current plan is to develop code to compare the minion_id to the hostname before the service starts. I was wondering if there are any automated solutions alredy built, or how to handle the stale keys that are going to be on the salt-master
16:54 twooster socket-_: I think you could delete the `/etc/salt/minion_id` file? and it would take on the hostname by default
16:55 twooster on first run.. could be wrong though
16:55 socket-_ is there a way to have the minion tell the master that it's changing ID so I don't have stale minion IDs
16:55 twooster i don't think so. minion id is tied to the pki folder, so they'll need to be re-approved.
16:56 twooster you could probably do a custom event/reactor based thing, but then you're trusting your minions a lot
16:57 svij3 joined #salt
16:58 twooster i have a little script for renaming minions from the master, if that interests you
16:58 socket-_ for sure, i'd love to see it
16:59 twooster https://pastebin.com/ZP0mySMu
17:00 twooster so far i haven't had any trouble with it
17:01 socket-_ thanks for sharing
17:02 twooster but i think MTecknology is right -- install salt, but remove the /etc/salt/minion_id before baking the image
17:03 dstensnes sounds like a much better idea really
17:03 dstensnes cleaner at least :)
17:03 adelcast joined #salt
17:03 acsir joined #salt
17:05 MTecknology https://gist.github.com/MTecknology/66ce7c7f148fc9da936bcf26cc572cd7#file-bootstrap-sh-L125
17:05 ronnix joined #salt
17:15 ChubYann joined #salt
17:19 socket-_ I was thinking about baking something like this into my Linux AMI's rc.local file.  https://apaste.info/RpNo
17:19 shadoxx joined #salt
17:20 socket-_ I remove the salt install from the AMI and just let the ec2 instance determine if salt is installed or not
17:26 wendall911 joined #salt
17:26 edrocks joined #salt
17:29 fatal_exception joined #salt
17:33 Llmiseyhaa joined #salt
17:42 aldevar joined #salt
17:45 bowhunter joined #salt
17:50 raspado joined #salt
17:51 raspado is there a cleaner way of structuring this statement {% if grains['role'] in ['postgresql'] or grains['clusterType'] in ['SingleNode_LocalPSQL'] %} ?
17:52 dxiri joined #salt
17:52 astronouth7303 seems pretty reasonable?
17:52 astronouth7303 i would do 'postgressql' in grains['role'], though
17:52 censorshipwreck joined #salt
17:52 astronouth7303 (assuming you don't have a eg 'postgresql-dev' role)
17:53 hasues joined #salt
17:53 raspado thanks astronouth7303
17:53 astronouth7303 (i just make the suggestion because often for me, role is a list)
17:54 raspado ahh kk
17:56 edrocks joined #salt
17:57 donmichelangelo joined #salt
17:57 cliluw joined #salt
17:58 mavhq joined #salt
17:58 raspado lets say if the grain is not present, would the salt minion silently error out or will it cause it to fail
17:59 astronouth7303 it will fail
17:59 raspado i seem to recall if i do if 'something' in grains['role'] causes highstate to fail, erroring out vs if grains['role'] in "something" was more forgiving?
17:59 druonysus joined #salt
17:59 astronouth7303 nope, it'll just fail
18:00 raspado got it thx
18:00 astronouth7303 `grains.get()` might be more forgiving (it'll return `None` instead of raising a `KeyError`)
18:00 raspado ahhh
18:00 raspado valid point
18:02 ivanjaros joined #salt
18:04 promorphus joined #salt
18:05 mikecmpbll joined #salt
18:07 high_fiver joined #salt
18:11 sp0097 joined #salt
18:12 lorengordon joined #salt
18:16 high_fiver_ joined #salt
18:17 leev joined #salt
18:21 cyteen joined #salt
18:21 xer0x I want to keep my Saltstack in Git, but some of my files are .rpms or .tars that get deployed. Do you guys check those in too?
18:22 XenophonF xer0x: I store "large" files in S3, so I use both s3fs and gitfs.
18:23 J0hnSteel joined #salt
18:24 xer0x Nice XenophonF, so the individual machines are setup to pull files from s3?
18:26 kellen joined #salt
18:27 tlefebvre joined #salt
18:27 jdipierro joined #salt
18:28 MTecknology xer0x: rpm files should not be in your salt structure, they should be on a repo server for distribution!
18:29 xer0x MTecknology: yeah! That's how I feel too..  I gotta fix this up
18:29 om2 joined #salt
18:30 astronouth7303 xer0x: i use something similar to XenophonF, but I use the filesystem on my master instead of S3. In my state files, I refer to `salt://lfs/mysoftware.tar.gz`
18:31 astronouth7303 i did discover that you really need to set up hashes when messing with archive states, though
18:32 MTO_ joined #salt
18:34 sh123124213 joined #salt
18:55 promorphus joined #salt
19:01 tapoxi hi all, having trouble with salt-cloud: https://pastebin.com/2ibGEAC2
19:01 tapoxi It's refusing to load libvirt but the library is there
19:01 tapoxi I can import it in python
19:01 tapoxi thoughts?
19:01 jdipierro joined #salt
19:02 astronouth7303 tapoxi: you didn't do anything like install salt-cloud in a venv or through conda or anything like that, did you?
19:03 tapoxi astronouth7303: nope, it's the rpm from the salt repos
19:03 aldevar joined #salt
19:05 jdipierro joined #salt
19:09 tom[] joined #salt
19:10 noraatepernos joined #salt
19:14 tapoxi oh, huh, its completely missing from /usr/lib/python2.7/site-packages/salt/cloud/clouds/
19:14 _KaszpiR_ joined #salt
19:15 astronouth7303 that'll do it
19:15 ronnix joined #salt
19:16 tapoxi yeah, guess its new as of 2017.7. weird.
19:16 astronouth7303 yeah, github matches that.
19:26 raspado_ joined #salt
19:27 pjs_ left #salt
19:27 pjs joined #salt
19:28 noraatepernos joined #salt
19:29 MTO_ joined #salt
19:29 pbandark joined #salt
19:30 mike25de_ joined #salt
19:31 mike25de_ hi all
19:41 oida joined #salt
19:44 onlyanegg joined #salt
19:53 onlyanegg joined #salt
19:57 hasues left #salt
19:59 MTO joined #salt
20:01 jhauser_ joined #salt
20:05 bowhunter joined #salt
20:05 schemanic joined #salt
20:07 onlyanegg joined #salt
20:10 censorshipwreck joined #salt
20:12 armguy joined #salt
20:23 pbandark joined #salt
20:23 LotR joined #salt
20:26 onlyanegg joined #salt
20:26 Ludo- joined #salt
20:26 Puckel_ joined #salt
20:26 whytewolf when will the headache stop?
20:27 astronouth7303 when the computers stop
20:28 MTecknology whytewolf: I'm hoping for 2017.7.1
20:29 whytewolf maybe 2017.7.3
20:31 inad922 joined #salt
20:33 tapoxi whats borked with 2017.7.0
20:35 astronouth7303 https://github.com/saltstack/salt/issues?q=is%3Aissue+is%3Aopen+label%3A2017.7.1 is slated to be fixed in the next release
20:36 astronouth7303 see also: https://github.com/saltstack/salt/issues?q=is%3Aissue+is%3Aopen+label%3ABlocker or some of the other labels
20:36 dxiri hey everyone, quick question, using salt-cloud, can I spin up a VM from an iso?
20:37 tapoxi dxiri: I havent seen that but it depends on the driver
20:37 astronouth7303 dxiri: depends on the driver and what the ISO is configured to do
20:37 astronouth7303 if you mean "can i just hand it standard debian media and get a working machine out of it", then no
20:38 dxiri lets say softlayer
20:38 om2_ joined #salt
20:39 astronouth7303 https://docs.saltstack.com/en/latest/topics/cloud/softlayer.html
20:41 dxiri yep been there, but doesn't say anything about isos
20:41 astronouth7303 i think you could configure VMWare to do it, but only if the ISO was in the configured VMware storage (not in the salt fileserver), and if the ISO was fully automated.
20:42 astronouth7303 i would assume it doesn't, then. You'd have to make an image first.
20:42 whytewolf dxiri: iso installs are not a typicall type of cloud image
20:42 om3 joined #salt
20:42 astronouth7303 see vagrant, packer, and a few others
20:43 dxiri mmm
20:43 dxiri https://docs.saltstack.com/en/latest/topics/cloud/softlayer.html#list-custom-images
20:43 dxiri maybe with that one
20:43 whytewolf that just lists the images that exist that were already built by you
20:44 dxiri correct, my point was just build a custom image with my iso, and then use that from that point forward
20:44 whytewolf ahh, yes that would be the best way forward
20:45 whytewolf generally. you create an image locally. convert it to a type that a cloud can read and upload it
20:45 MTecknology FUUUUUUUuuuu........   I just spent >30 minutes trying to debug this- http://dpaste.com/30KS73P
20:45 dxiri orrrrr use virt-install to automate that part, then hand the resulting image over to salt-cloud
20:46 astronouth7303 if you can. it works better with some cloud drivers than others. https://github.com/leonidlm/packer-builder-softlayer works similar to the VMWare packer backend where it spins up a VM that'll be turned into a template.
20:46 om2_ joined #salt
20:46 MTecknology ^ pillar.get() vs. salt.pillar.get()
20:46 astronouth7303 MTecknology: ooph
20:47 whytewolf hehe yeah. pillar.get doens't like :
20:47 whytewolf :P
20:48 whytewolf dxiri: generally i don't do iso installs. i find a minimal already built image and just change it to fit my needs.
20:49 whytewolf but i generally use clouds that use cloud-init which can be a pain if doing an iso install on because a lack of cloud-init software in the image could lead to it not picking up a lot of info that is needed internally
20:49 aldevar joined #salt
20:50 aldevar left #salt
20:50 MTecknology cloud-init is a massive clusterfuck of messy (in my own opinion, of course)
20:50 dxiri whytewolf: yep, agreed on that, but can't avoid it this time, our devs only produce an iso or a usb image
20:50 dxiri MTecknology: agree as well!
20:50 MTecknology I use the DO provided image, but my bootstrap removes cloud-init stuff
20:50 om3 joined #salt
20:51 noobiedubie joined #salt
20:51 astronouth7303 dxiri: that's... odd. A bootable ISO of the actual production system?
20:52 astronouth7303 try taking the USB image, converting it to the disk image format for your cloud provider, and wrapping it with the hardware description as necessary?
20:53 astronouth7303 several providers have an instance template format (OVF, VMX, AMI, ...)
20:54 dxiri let me elaborate a little bit more, this is a distributed system that has a controller, that controller its what I am trying to install from scratch, and it has custom kernel stuff and even its own installer, formats drives, creates lvms, etc etc
20:55 whytewolf MTecknology: as i generally need it for some of the cloud functions i use it for. i keep hoping for something better. it has gotten marginally better over time but yes it is a cluster fuck
20:55 dxiri the devs build the whole software and they provide an iso and usb image that you can just take to install on bare metal
20:55 astronouth7303 you should first check if your provider supports it. If they don't, salt-cloud certainly won't.
20:56 whytewolf hell see if provider already has a cloud image for it
20:56 whytewolf a lot of vendors do
20:57 astronouth7303 I know I can do it with VMWare vSphere/ESXi, but that requires running your own hardware and licensing VMware. You can probably do something similar with KVM, but again your own hardware.
20:57 whytewolf hell you can build something local with virtualbox
20:58 whytewolf once you have a snapshot the image protion is easy
20:58 astronouth7303 i'm pretty sure AWS can't do it, or can't do it without some gymnastics.
20:58 whytewolf clouds generally don't have a way for you to mount an iso.
20:58 astronouth7303 whytewolf: it sounds like they're going to be building instances from ISOs on a very regular basis
20:59 Guest73 joined #salt
20:59 whytewolf so, packer with virtualbox then [which is headless]
20:59 whytewolf :P
21:00 whytewolf openstack has a workaround which is basicly do it in the hypervisor
21:01 whytewolf which is exactly the libvirt method dxiri sugested
21:01 Ni3mm4nd joined #salt
21:03 squishypebble joined #salt
21:04 pppingme joined #salt
21:11 raspado joined #salt
21:12 cyborg-one joined #salt
21:17 dxiri I think I found a winner :)
21:17 dxiri https://knowledgelayer.softlayer.com/procedure/boot-vsi-image
21:19 dxiri ah shit needs to be switft image :(
21:19 dxiri so no winner
21:20 Etong joined #salt
21:22 Etong joined #salt
21:26 Etong joined #salt
21:29 eritong joined #salt
21:31 jeddi joined #salt
21:46 jhauser joined #salt
21:47 jhauser joined #salt
21:48 jhauser joined #salt
21:49 jhauser joined #salt
21:49 jhauser joined #salt
21:50 jhauser joined #salt
21:51 jhauser joined #salt
21:52 nethershaw joined #salt
21:53 Heartsbane joined #salt
21:53 Heartsbane joined #salt
21:57 hashwagon joined #salt
21:59 bluenemo joined #salt
22:01 jimklo joined #salt
22:18 debian112 joined #salt
22:24 ecdhe joined #salt
22:25 onlyanegg joined #salt
22:28 fleaz joined #salt
22:31 Felgar joined #salt
22:55 hashwagon I'm trying to create a state file which creates a system user with nologin. Does anyone have an example of this? Do I do - system: True and  - shell: /sbin/nologin ?
22:58 dendazen joined #salt
23:09 MTecknology yup
23:09 jmiven left #salt
23:09 lorengordon joined #salt
23:27 deep-book-gk_ joined #salt
23:28 deep-book-gk_ left #salt
23:43 lorengordon joined #salt
23:44 noobiedubie joined #salt
23:59 jeddi joined #salt

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