Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-08-19

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

All times shown according to UTC.

Time Nick Message
00:28 dthom91 joined #salt
00:37 terminalmage joined #salt
00:44 jesusaurus joined #salt
00:45 bhosmer joined #salt
00:59 sleek joined #salt
01:00 sleek hi, i'm running into a problem, but not really sure how to fix it .. i have a salt master running in a docker container, and i have a minion running locally .. when i start the minion it connects to the master, i accept the key on the master, and it looks fine .. however, when i run "sudo salt '*' test.ping" , it says "Failed to connect to the Master, is the Salt Master running?"
01:04 liuyq joined #salt
01:05 troyready joined #salt
01:06 pewpewarrows joined #salt
01:07 napperjabber joined #salt
01:07 Thiggy joined #salt
01:08 Lue_4911 joined #salt
01:08 syngin sleek: have you tested connectivity over ports 4505 and 4506?
01:10 sleek yeah, so i can at least telnet in on both of those ports
01:10 noob2 joined #salt
01:10 noob2 i reimaged a server and now I'm not certain how to add it back into salt.  it has the same name as before
01:10 syngin have you tried running both master and minion in the foreground?
01:10 noob2 but the keys are different
01:10 syngin noob2: delete the old key
01:11 noob2 ok
01:11 noob2 there we go
01:12 noob2 thanks syngin :)
01:12 syngin np :)
01:12 noob2 now i get to see how good of a job i did in my state files after the reimage
01:12 syngin noob2: exciting!
01:12 alunduil joined #salt
01:13 noob2 :)
01:13 sleek syngin: yeah, i have it running with "-l debug" and can see this , for example: [DEBUG   ] This salt-master instance has accepted 1 minion keys.
01:16 syngin sleek: hmm. weird...
01:16 mgw joined #salt
01:16 noob2 seems test.ping doesn't return anything
01:17 syngin noob2: from your reimaged server?
01:17 noob2 yeah i think my minion keeps stopping
01:17 noob2 i'm going to start it in the foreground
01:17 liuyq joined #salt
01:17 noob2 ok yeah it just crashes with KeyError: 'token'
01:18 noob2 i wonder if the versions of salt are different
01:18 mgw joined #salt
01:18 rustyrazorblade joined #salt
01:18 noob2 oops, the versions are way off
01:20 syngin well, that'll do it
01:21 noob2 can 16.0-1 play nice with 16.3?
01:21 syngin dunno, i didn't see anything in the changelog that would suggest they can't, but i never test that kind of thing.
01:22 noob2 seems to play nice.  ping returns true now
01:22 syngin goddy up
01:23 syngin eek
01:23 syngin s/goddy/giddy/g
01:23 noob2 giddy up :D
01:24 Nexpro joined #salt
01:28 noob2 syngin: any idea if salt can do a add-apt-repository ?
01:30 syngin noob2: what do you mean?
01:30 noob2 so to get my highstate to complete properly it needs to add a repo for ceph
01:30 syngin is that some kind of command to add a system package repo to some kind of config file?
01:31 noob2 yeah for ubuntu
01:31 syngin there'll definitely be a way to do it; how i do it on arch is to manage the file that contains all of the repositories.
01:31 syngin (/etc/pacman.conf)
01:32 blee_ joined #salt
01:32 noob2 ah yes i love arch
01:36 m_george|away joined #salt
01:43 dthom91 joined #salt
01:47 syngin noob2: i'm lucky enught to manage a cluster of arch boxes. no rhel or ubuntu in sight.
01:48 syngin jesus, that's some bad typing right there.
01:49 malinoff joined #salt
01:51 noob2 haha
01:51 dstanek joined #salt
01:52 noob2 syngin: are these running at a company?  if so that's pretty amazing
01:52 syngin noob2: yepper :)
01:52 syngin noob2: managed to get in a position where i chose arch and a whole bunch of funky toys (uwsgi, salt, etc).
01:54 noob2 damn man
01:54 noob2 sign me up
01:54 noob2 i'm in centos hell
01:54 syngin oh, man. i know that pain.
01:54 noob2 old ass kernels and python
01:54 syngin everything in arch is so new and clean.
01:54 noob2 :D
01:55 syngin just upgrading my servers to 3.10.6 (they're all 3.10.5 at the moment).
01:55 noob2 nobody gave you crap about 'stability' concerns?
01:55 TJ1980 joined #salt
01:57 syngin noob2: a few folk, but nothing i couldn't surmount with "inexploitable - i'm moving faster than the black-hats" ;)
01:57 noob2 haha
01:57 syngin the arch guys are pretty good at keeping shit stable though.
01:57 syngin i can't remember the last time i had problems at the kernel level.
01:57 noob2 yeah arch is awesome
01:57 noob2 i remember a time when i was running gentoo when everytime i did a portage build everything would break
01:57 syngin we've got quite a few rhels floating around, but that's all vendor managed ;)
01:58 syngin ouchie
01:58 syngin never played with gentoo - i've got respect for it though. i went from slackware to arch.
01:58 capricorn_1 joined #salt
01:58 noob2 gotcha
01:58 syngin so basically no upgrading to always upgrading ;)
01:58 noob2 yeah vendor managed crap is for the birds
01:58 blee_ left #salt
01:58 noob2 i agree
01:58 noob2 centos 4->5 is painful, same with 5->6
01:59 noob2 i'd rather just have the latest all the time
01:59 syngin yeah, the few remaining slackware boxes i built for work, just work, but it's killer to keep them fresh.
02:00 noob2 i'm sure
02:01 whit joined #salt
02:01 syngin noob2: where abouts are you?
02:01 noob2 i'm just outside of san fran
02:01 syngin noob2: right - a bit late for you, isn't it? (I'm in Melbourne, Aus).
02:02 noob2 oh
02:02 noob2 no it's only 7pm :)
02:02 malinoff joined #salt
02:02 noob2 have you figured out a good system for testing changes in salt before pushing them?
02:03 syngin noob2: not really - i do my highstates on my dev/test servers first.
02:03 syngin noob2: you could look at the "show_sls" type things and review the compiled states, i guess.
02:04 noob2 yeah that's a good idea just point it at dev and see if it blows up
02:05 syngin noob2: yeah, i have duplicate logging servers so i just upgrade/highstate one side first. the only hairy stuff are servers that are individual - i just do those during a change window and hope i can mend any issues while it's open.
02:06 noob2 yeah
02:06 syngin time for a ciggie.
02:06 noob2 sounds good
02:17 xl joined #salt
02:39 whit joined #salt
02:48 StDiluted joined #salt
02:50 noob2 joined #salt
02:50 noob2 left #salt
02:52 geopet joined #salt
02:58 sgviking joined #salt
03:03 dthom91 joined #salt
03:05 cewood joined #salt
03:07 racooper joined #salt
03:08 racooper_ joined #salt
03:08 racooper joined #salt
03:13 troyready joined #salt
03:17 oz_akan_ joined #salt
03:19 geopet joined #salt
03:21 renothing joined #salt
03:26 cewood joined #salt
03:27 aat joined #salt
03:34 idkfa joined #salt
03:49 dstanek joined #salt
03:58 Ryan_Lane joined #salt
03:59 rustyrazorblade joined #salt
04:04 Ryan_Lane joined #salt
04:04 Ryan_Lane joined #salt
04:13 cxz test=true
04:13 cxz when u run salt i believe
04:15 qba73 joined #salt
04:19 UtahDave joined #salt
04:23 neilf joined #salt
04:26 rustyrazorblade joined #salt
04:32 Katafalkas joined #salt
04:40 rustyrazorblade joined #salt
04:44 renothing joined #salt
04:54 dstanek joined #salt
04:56 waverider joined #salt
04:58 xl joined #salt
04:59 middleman_ joined #salt
05:18 malinoff Hi all. What's the best way to automatically store and distribute dynamic data such as ip's of minions? I'm trying salt mine, but i can't get when the data is updating?
05:18 malinoff In other words, i want to have a config file contains ip's of minions from one group
05:19 malinoff Minions can die or born, so i need a dynamic array of their ip's
05:21 cewood joined #salt
05:22 cxz malinoff: you could salt-mine a timestamp
05:23 malinoff cxz, timestamp of what?
05:25 malinoff I want something like https://gist.github.com/malinoff/6265950
05:26 malinoff with scheduling mine.get
05:31 UtahDave malinoff: have you looked at the peer system?
05:32 UtahDave malinoff: here's a gist of using the peer system to get all the web servers ip addresses
05:32 UtahDave malinoff: https://gist.github.com/UtahDave/5217462
05:33 malinoff @UtahDave, yes, but it's not what i really need. I want some automatics.
05:33 malinoff http://docs.saltstack.com/topics/mine/index.html this page says i can do that
05:34 UtahDave is the mine not working for you?
05:35 malinoff @UtahDave, yes, i can collect mine data on the master and i can run an sls to update my config file. But i want to configure my minion just once, and no more sls calls.
05:37 malinoff http://docs.saltstack.com/topics/mine/index.html#mine-interval i think, i can't understand this part. Will 'mine.get' call in my jinja-templated config file by scheduler?
05:38 malinoff Or it's just like 'salt-call '*' mine.get' and i must to update my config file by myself?
05:39 Nobbl joined #salt
05:40 berto- joined #salt
05:45 platforms UtahDave: Great gist - was just looking for a way to do this - thanks
05:45 platforms Gives me a model for doing something similar with haproxy
05:49 joehh anyone done a postgresql/postgis package for windows?
05:56 Katafalkas joined #salt
06:04 gildegoma joined #salt
06:11 mprymek joined #salt
06:15 UtahDave joehh: I haven't heard of one.  That would be cool
06:15 UtahDave platforms: you're welcome1
06:15 xl joined #salt
06:17 joehh looks like I'll be doing one then... :)
06:19 UtahDave :)
06:19 scalability-junk joined #salt
06:19 andrewclegg joined #salt
06:20 gamingrobot joined #salt
06:20 jrgifford joined #salt
06:20 shennyg joined #salt
06:20 octarine joined #salt
06:20 ada__ joined #salt
06:20 akitada joined #salt
06:21 xerxas joined #salt
06:23 az87c joined #salt
06:25 yogi I am trying to a test a sls recipe , Is it possible to invoke it individually without calling "state.highstate" e.g salt "testweb0*"  mytest.sls -l debug true=True ?
06:27 UtahDave yogi: salt 'testweb0*' state.sls mytest -l debug test=True
06:31 Katafalkas joined #salt
06:33 cxz is there a way to mine.get('*') for * minions?
06:38 yogi utahdave: Thanks, what if my my recipe is not under base but under test
06:39 UtahDave yogi, do you mean in a "test" directory?
06:41 yogi utahdave: yes I have all my test receipies under /srv/salt/test
06:42 unicoletti_ joined #salt
06:44 UtahDave ah, then salt 'testweb0*' state.sls test.mytest -l debug test=True
06:49 yogi utahdave: bingo, that worked like a charm. Thanks
06:49 vaxholm joined #salt
06:49 UtahDave You're welcome!
06:52 matanya joined #salt
06:58 davidone joined #salt
07:00 BbT0n joined #salt
07:06 joehh UtahDave: can you use pillar data in the windows package installer sls files?
07:07 joehh need to enter a password as part of the postgresql install
07:08 UtahDave No, not yet unfortunately. I need to sit down with Tom and figure out how to have them processed by the Salt loader as well as the render system
07:10 joehh might make things a litle tricky. I'll see if I can pass it in from a file
07:11 UtahDave ok.
07:11 UtahDave I just made a note to myself to look into that tomorrow. I've been meaning to get to that.
07:13 joehh thanks for that
07:14 joehh in other news, formalities (hopefully) completed for debian maintainer status, so just need to wait a few days for any objections and then wait for the next set of DM updates
07:15 joehh then I should have upload rights to the debian repo for salt :)
07:16 mike25ro1 joined #salt
07:17 ml_1 joined #salt
07:26 berto- joined #salt
07:27 UtahDave that's awesome!
07:27 UtahDave thanks for going through all that
07:30 joehh no worries
07:34 share_by joined #salt
07:35 malinoff @UtahDave, do you have some ideas to create built-in syncronization service for autoscaling?
07:37 UtahDave malinoff: what do you mean by a built-in syncronization service?
07:39 carlos joined #salt
07:40 malinoff @UtahDave: i mean, there should be easy way to catch new (recently loaded) minion and include it (by some grain, like IPv4) to already working minions. And exclude when minion is going down.
07:41 UtahDave malinoff: Ah, you can do that now with the event system and the Reactor system
07:42 UtahDave You can set up the reactor on your master to listen for specific events, such as new minion authenticating and do something with it
07:42 Katafalkas joined #salt
07:43 goodwill just watch for the reactor overload
07:43 malinoff @UtahDave: Alright, thanks. I'll dive in it.
07:43 mike25ro1 interesting feature the Reactor thingy
07:45 UtahDave Yeah, it allows Salt to do lots of interesting dynamic things on your infrastructure.
07:46 UtahDave well, it's 1:45 am here. I'd better get to bed.
07:46 UtahDave night all
07:47 malinoff @UtahDave: Bye. Thanks for the help
07:47 UtahDave you're welcome1
07:48 * EugeneKay spreads fresh blankies
07:57 Ryan_Lane joined #salt
07:59 scott_w joined #salt
08:05 felixhummel joined #salt
08:08 txmoose joined #salt
08:08 xl1 joined #salt
08:13 krak3n` joined #salt
08:18 alekibango joined #salt
08:20 yota joined #salt
08:22 krissaxton joined #salt
08:27 zooz joined #salt
08:27 ml_1 joined #salt
08:27 balboah_ joined #salt
08:29 MrTango joined #salt
08:30 it_dude joined #salt
08:31 Xeago joined #salt
08:56 dstanek joined #salt
09:02 whiskybar joined #salt
09:05 Jahkeup joined #salt
09:09 tricksyhobbit joined #salt
09:10 tuxIO joined #salt
09:20 tuxIO_ joined #salt
09:21 bemehow joined #salt
09:21 mortis_ is there a way of using peers and getting the output back to the minion starting it?
09:22 mortis_ like the result of test.ping for instance
09:23 mortis_ i mean, when running salt-call publish.runner|publish \* test.ping
09:30 xl1 joined #salt
09:34 pszwed joined #salt
09:34 malinoff Can i somehow detect that a minion breaks down? I know, there is 'minion_start' event, is there any 'minion_stop' event?
09:37 liuyq joined #salt
09:38 liuyq joined #salt
09:38 ml_1 joined #salt
09:41 bogdanbu joined #salt
09:48 mwillhite joined #salt
09:55 abele joined #salt
10:02 Ryan_Lane joined #salt
10:08 hvn joined #salt
10:09 dstanek joined #salt
10:19 xl1 joined #salt
10:36 bogdanbu joined #salt
10:39 giantlock joined #salt
11:00 bhosmer joined #salt
11:05 qba73 joined #salt
11:10 derelm joined #salt
11:20 gmoro joined #salt
11:21 krak3n` joined #salt
11:22 krak3n` joined #salt
11:27 Ryan_Lane joined #salt
11:37 y0j joined #salt
11:38 lemao joined #salt
11:38 TJ1980 joined #salt
11:42 waverider joined #salt
11:55 austin987 joined #salt
11:59 jeddi joined #salt
12:03 derelm_ joined #salt
12:04 hernantz joined #salt
12:17 Jarus joined #salt
12:18 oz_akan_ joined #salt
12:20 bhosmer joined #salt
12:21 vaxholm_ joined #salt
12:22 aat joined #salt
12:31 xl1 joined #salt
12:33 platforms so strange: installing the python-pip package always fails on centos with a salt sls, even though it works on the command line with yum. Other packages install fine. And if I install with yum, salt shows pip is installed. It just refuses to be the one to install it.
12:33 platforms using 0.16.0
12:38 TJ1980 joined #salt
12:39 oz_akan_ joined #salt
12:41 liuyq joined #salt
12:45 TJ1980 joined #salt
12:47 com48 joined #salt
12:47 blee joined #salt
12:48 cron0 joined #salt
12:49 platforms installing 0.16.3 from source fixed it
12:51 ktenney joined #salt
12:52 Nexpro joined #salt
12:52 balboah_ joined #salt
12:53 anteaya joined #salt
12:56 elfixit joined #salt
12:59 tuxIO_ joined #salt
13:02 xl1 left #salt
13:04 spmurrayzzz joined #salt
13:05 aat joined #salt
13:06 spmurrayzzz Anyone here have experience running salt-minion as a nonroot user?
13:07 williamthekid_ joined #salt
13:09 Katafalkas joined #salt
13:12 bemehow joined #salt
13:15 toastedpenguin joined #salt
13:16 kvbik_ joined #salt
13:17 Mrono_ joined #salt
13:17 yml_ joined #salt
13:17 Artan1cus joined #salt
13:17 lmnts joined #salt
13:19 brianhicks joined #salt
13:20 quantumsummers|c joined #salt
13:20 quantumsummers|c joined #salt
13:20 napperjabber joined #salt
13:20 ewong_ joined #salt
13:20 jgelens_ joined #salt
13:22 kiorky_ joined #salt
13:24 aat joined #salt
13:24 juicer2 joined #salt
13:28 mwillhite joined #salt
13:28 brianhicks joined #salt
13:29 majoh joined #salt
13:29 robawt joined #salt
13:29 tethra joined #salt
13:29 akio joined #salt
13:33 kiorky joined #salt
13:34 Gifflen joined #salt
13:35 bemehow joined #salt
13:37 racooper joined #salt
13:38 TJ1980 joined #salt
13:38 krak3n` joined #salt
13:41 Jahkeup joined #salt
13:42 Khollowa_ joined #salt
13:47 ipmb joined #salt
13:52 whit joined #salt
13:52 yano joined #salt
13:52 vaxholm_ joined #salt
13:54 KennethWilke joined #salt
13:56 mannyt joined #salt
13:56 TJ1980 joined #salt
14:01 oz_akan_ joined #salt
14:02 premera joined #salt
14:10 opapo joined #salt
14:11 mortis_ platforms: yeah had the same issue yesterday, so now im only using bootstrap
14:11 TJ1980 joined #salt
14:13 KennethWilke has anyone here created states to manage configuration files of software across various distros and versions?
14:13 KennethWilke mysql specifically is what i have in mind at the moment, but i'm curious how people have organized their states for this
14:15 kaptk2 joined #salt
14:19 opapo joined #salt
14:19 conan_the_destro joined #salt
14:23 jalbretsen joined #salt
14:28 jeffasinger joined #salt
14:31 StDiluted joined #salt
14:31 alunduil joined #salt
14:33 StDiluted joined #salt
14:35 Furao joined #salt
14:35 honestly someone add a crontab-formula template to saltstack-formulas
14:35 honestly so I can fork it
14:36 honestly it's not really a complex thing, but eh
14:36 middleman_ joined #salt
14:38 mwillhite joined #salt
14:39 matanya joined #salt
14:40 jpeach joined #salt
14:41 whit joined #salt
14:48 ckao joined #salt
14:48 cnelsonsic joined #salt
14:50 hernantz joined #salt
14:50 Katafalkas I believe that someone was working on server monitoring via salt ?
14:50 Katafalkas I cant seem to find that on github
14:51 ggoZ joined #salt
14:53 kula i have a link to that somewhere, because i was going to look at it. let me dig.
14:53 mwillhite joined #salt
14:53 dthom91 joined #salt
14:55 kula hmm, it's probably open in a browser tab on the home laptop.
14:55 Khollowa_ joined #salt
14:56 hernantz Katafalkas, django salmon https://github.com/lincolnloop/salmon
14:57 Katafalkas uu.
14:57 Katafalkas danke
14:57 Katafalkas hernantz: <3
14:58 g3cko oohhh that looks useful! thanks!
14:58 hernantz np
15:00 Jahkeup anyone working with custom _states? I'm having issues
15:02 m_george left #salt
15:03 Furao_ joined #salt
15:09 unicoletti_ left #salt
15:10 sinh joined #salt
15:12 tom_ joined #salt
15:13 tom_ hello - Anyone had any luck using virtualbox for testing salt-cloud. Can't find any evidence of a driver.
15:14 forrest joined #salt
15:15 chrisgilmerproj joined #salt
15:16 Jahkeup joined #salt
15:19 lazyguru joined #salt
15:19 ml_1 joined #salt
15:21 pkimber joined #salt
15:23 bemehow joined #salt
15:28 Brainspackle left #salt
15:28 pdayton joined #salt
15:29 faldridge joined #salt
15:30 aat joined #salt
15:33 m_george|away joined #salt
15:34 robawt joined #salt
15:36 LucasCozy joined #salt
15:38 ipmb joined #salt
15:39 krissaxton joined #salt
15:39 cron0 joined #salt
15:41 UtahDave joined #salt
15:43 kaptk2 joined #salt
15:44 mikedawson joined #salt
15:46 cedwards UtahDave: morning
15:46 UtahDave morning, cedwards!
15:47 KennethWilke g'morning
15:47 cedwards UtahDave: I've got packages built and published at freebsd.saltstack.com
15:47 UtahDave morning, KennethWilke! I see you've been busy already this morning!
15:48 UtahDave cedwards: awesome! Should we add that info to our docs?
15:48 Nexpro1 joined #salt
15:48 cedwards UtahDave: I suppose we should test the bootstrap. I haven't done that yet.
15:48 KennethWilke UtahDave: indeed! the mysql and rabbit formulas are gettin' it today
15:49 cedwards s0undt3ch: ping re: FreeBSD bootstrap
15:49 KennethWilke the mysql formula isn't as up to snuff on gentoo, anybody here using salt for mysql+gentoo?
15:49 s0undt3ch cedwards: what's up
15:49 s0undt3ch cedwards: we have the host up?
15:50 cedwards s0undt3ch: the host is up and packages are published
15:50 s0undt3ch cedwards: nice, so, we're good to change the bootstrap script and point to it?
15:50 cedwards s0undt3ch: we need to point the bootstrap to PACKAGESITE: http://freebsd.saltstack.com/${ABI}/
15:51 kermit joined #salt
15:51 s0undt3ch cedwards: so, want to make a PR against the bootstrap script or want me to do that ;)
15:53 cedwards s0undt3ch: I can work on it
15:53 kermit joined #salt
15:53 s0undt3ch cedwards: cool! if can't focus on that(no time or anyting else), I'll take a look
15:53 s0undt3ch *if you can't
15:54 whit joined #salt
15:54 ipmb joined #salt
15:55 kermit joined #salt
15:58 s0undt3ch cedwards: ${ABI} exists on the env?
15:58 s0undt3ch cedwards: just wondering
15:59 timl0101 joined #salt
15:59 KyleG joined #salt
15:59 KyleG joined #salt
15:59 cedwards s0undt3ch: ${ABI} should be expanded by pkgng, yeah.
15:59 s0undt3ch cedwards: nice ;)
15:59 s0undt3ch cedwards: k, let me know if you start working on changing the boostrap script. If I don't hear from you, I'll update it when I can
16:00 jpeach joined #salt
16:01 cedwards s0undt3ch: I'm working on a PR now
16:01 s0undt3ch cedwards: ok. Thanks!
16:01 brianhicks joined #salt
16:05 dkubb joined #salt
16:08 cedwards s0undt3ch: submitted.. i think.
16:08 cedwards looks like github changed the UI a bit since my last pull request
16:09 m_george left #salt
16:09 s0undt3ch cedwards: hmm, can't merge it, can you update and rebase your changes please?
16:10 s0undt3ch cedwards: I can't do it "by hand" if that's too much trouble
16:10 s0undt3ch cedwards: sorry, s/can't/can/
16:10 kenbolton joined #salt
16:11 tuxIO joined #salt
16:11 cedwards s0undt3ch: I wonder what happened. That was a fresh fork and checkout..
16:12 s0undt3ch cedwards: dunno
16:12 cedwards i'll try again
16:12 s0undt3ch :\
16:12 s0undt3ch cedwards: git pull upstream develop
16:13 s0undt3ch ?
16:13 s0undt3ch no need to fork
16:13 s0undt3ch but you probably know that ;)
16:18 cedwards s0undt3ch: should I be able to push directly to saltstack/salt-bootstrap.git?
16:19 bitz joined #salt
16:20 KennethWilke hey guys, i want to add a cmd.run for a watch under pkg.installed in a state file to run a post-installation script, what would you recommend?
16:21 EugeneKay A stiff drink.
16:21 s0undt3ch cedwards: don't know if you can, but if you can, please do it on develop
16:21 KennethWilke EugeneKay: well that'll fix anything
16:22 KennethWilke though i think at least lunch should happen first
16:22 cedwards s0undt3ch: I created a freebsd branch, but get access denied.
16:22 s0undt3ch cedwards: then you can't, only PRs
16:22 s0undt3ch :)
16:22 cedwards s0undt3ch: I guess I do need to update my fork and try this again
16:22 s0undt3ch cedwards: I can still merge your work by hand
16:22 s0undt3ch cedwards: if you prefer
16:23 cwarner joined #salt
16:23 vaxholm_ joined #salt
16:26 cwarner joined #salt
16:26 Jahkeup hey guys! I can't get my custom _states/ovs.py to work, I keep getting "State ovs.bridged found in sls ovs-test is unavailable"
16:27 geopet joined #salt
16:27 Jahkeup and the _modules/ovs.py's functions + docstrings show up in sys.doc
16:28 mikedawson joined #salt
16:28 UtahDave Jahkeup: the best thing to do in that situation is to ssh in to your minion and run the salt-minion in the foreground in debug mode
16:28 UtahDave salt-minion -l debug
16:28 UtahDave then try sending the commands from the master
16:28 Jahkeup UtahDave: this is actually in vagrant
16:28 Jahkeup UtahDave: so I don't have the minion, just salt-call
16:29 cedwards s0undt3ch: this one should work: https://github.com/saltstack/salt-bootstrap/pull/209
16:30 s0undt3ch cedwards: it does
16:30 UtahDave Jahkeup: ok, try adding -l debug to your salt-call command
16:30 Jahkeup UtahDave: just did and I think I found the problem :)
16:30 cedwards s0undt3ch: my fork was months out of date and i didn't realize it
16:30 s0undt3ch cedwards: :)
16:31 troyready joined #salt
16:31 forrest s0undt3ch, what's the rule regarding user/pass prompts within the bootstrap? I assume you want to stay away from them?
16:31 kaptk2 does anybody have a more concrete example on using virt.create?
16:31 s0undt3ch forrest: fully automated
16:31 s0undt3ch forrest: no user interaction what so ever
16:32 Jahkeup UtahDave: one more thing then! can you not import a custom module via import salt.modules.ovs?
16:32 Lue_4911 joined #salt
16:32 forrest Hmm, that's gonna be a problem for the red hat optional channel code since you have to pass a user/password. Any ideas?
16:33 s0undt3ch forrest: ah, still the optional channel issue?
16:33 s0undt3ch bummer
16:33 s0undt3ch forrest: that user/pass can't be from the environment or a file?
16:34 UtahDave Jahkeup: Hm. Not sure. Seems like you should be able to if the module is loading correctly
16:34 mannyt joined #salt
16:34 forrest s0undt3ch, yea had some free time this morning so I'm looking at updating the bootstrap a bit. I guess it could be passed in from the environment, but that doens't cover a use case of multiple users running the script (or different creds), you could pass it in from a file I guess. There are command line options you can pass to rhn-channel, it's just getting them there that I'm concerned about.
16:35 s0undt3ch cedwards: http://jenkins.saltstack.com/job/Salt%20Rackspace%20FreeBSD/56/console it's going to use your new repo :)
16:36 s0undt3ch cedwards: ah, not, dam, forgot to remove the env setting
16:36 Jahkeup UtahDave: that's whats making my state not work :/ I get "ImportError: No module named ovs"
16:36 Jahkeup UtahDave: though it definitely gets loaded
16:40 hhoover joined #salt
16:43 forrest s0undt3ch, I don't see any existing usage of a static file/environment variables in the bootstrap, am I just missing them or are there none?
16:43 s0undt3ch forrest: there are a bunch of env ones
16:43 forrest sorry I meant environment varialbes that aren't set in the script.
16:43 forrest *variables
16:45 cedwards oops. just realized I commited that last change using my work email.
16:46 forrest Quick everyone send cedwards phishing emails
16:48 s0undt3ch cedwards: want that cleaned up? or is it ok for you?
16:49 cedwards s0undt3ch: i'm sure it's fine
16:49 s0undt3ch cedwards: for me it's fine ;)
16:50 Ivo joined #salt
16:52 jinnko joined #salt
16:53 ml_1 joined #salt
16:55 blee joined #salt
16:56 waverider joined #salt
16:57 kermit joined #salt
16:58 vaxholm_ joined #salt
16:58 mgw joined #salt
17:00 platforms can user or group be used as a requisite for a file to be created? Or do I have to go into the user/group blocks and "require_in" the file?
17:04 StDiluted If I've got a list in a pillar, and I want to populate a single line with the values of that list, separated by spaces, would I use a for loop in jinja, and how would that look?
17:04 platforms the docs only show service, pkg and sls states used in require blocks.
17:04 StDiluted platforms: require is the same as require_in, just backwards
17:04 mwillhite joined #salt
17:04 krissaxton left #salt
17:04 StDiluted you can use a require: user
17:05 platforms StDiluted: just tried requiring both user and group, and it didn't give an error. but it also didn't actually require them. (0.16.3)
17:05 UtahDave platforms: "require" is not specific to any specific state. You can require against any other state that you're using
17:05 StDiluted hrm, did you include the state where the users are defined?
17:06 platforms UtahDave: tried using a require of a pip state earlier and got an error. I'll try again.
17:06 platforms StDiluted: yes - they are all in one sls file
17:07 mattmtl joined #salt
17:07 Ryan_Lane joined #salt
17:07 forrest s0undt3ch, what do you think about this: https://github.com/gravyboat/salt-bootstrap/blob/develop/bootstrap-salt.sh , starts on line 1789
17:08 forrest rhn-channel apparently only takes unencrypted passwords for the -p option, so I don't really like the idea of passing it through
17:08 s0undt3ch forrest: missing a fi in line 1797
17:09 forrest fixed it
17:09 s0undt3ch forrest: so, you require the user to first enable the optionals?
17:09 s0undt3ch then run the script?
17:09 druonysus joined #salt
17:09 druonysus joined #salt
17:09 forrest well, technically if you're using RHN, you can join machines as soon as they are added to RHN to whatever channels you needs (which are the repos)
17:11 forrest I just don't have a good solution that doesn't include either adding an option specifically for red hat, passing a password (which would most likely be the RHN admin password) in clear text, or prompting the user for input.
17:12 mmilano joined #salt
17:13 mmilano left #salt
17:13 platforms StDiluted: i found the problem - I hadn't checked in the edits to the sls before I destroyed and rebuilt my vagrant box. duh.
17:14 StDiluted ah ha!
17:14 StDiluted Anyone here a jinja expert?
17:14 platforms moar coffee.
17:14 Corey StDiluted: Whats your actual question?
17:14 StDiluted Corey: If I've got a list in a pillar, and I want to populate a single line with the values of that list, separated by spaces, would I use a for loop in jinja, and how would that look?
17:15 Corey I *think* you could go with a usual for loop, iterating over the results of a pillar lookup.
17:16 StDiluted for example: I have ['test1.domain.com','test2.domain.com'] in my pillar, and I want output in a templated config file to look like: domains = test1.domain.com test2.domain.com
17:17 StDiluted how do I get more than one result from the for loop on the one line?
17:17 StDiluted seems like I'd need to iterate and append to a variable and then output that
17:17 StDiluted but not sure how to do that in jinja
17:20 StDiluted Corey: does my question make sense?
17:21 forrest can't you just append a string StDiluted?
17:21 StDiluted forrest: append to what?
17:21 forrest http://stackoverflow.com/questions/2061439/string-concatenation-in-jinja
17:21 yano joined #salt
17:21 forrest {{ stuffs|join(", ") }}
17:21 forrest that being the key
17:22 forrest just change it to {{ stuffs|join(" ") }}
17:22 StDiluted (I'm not a python guy)
17:22 forrest ?
17:22 forrest are you saying because you don't understand, or because that answer makes sense?
17:22 StDiluted I have a pillar that contains a list
17:22 forrest right
17:23 forrest stuffs = the list
17:23 StDiluted will join concatenate that list?
17:23 forrest according to this SO post it will
17:23 StDiluted I dont know python very well, so I've never used 'join'
17:23 StDiluted so I dont need to iterate over the list, I can just use join
17:23 forrest yep
17:24 forrest you should be able to do something like domains = {{ stuffs|join(" ") }}
17:24 forrest but there might be a better way to do it in salt, not sure.
17:25 platforms ls
17:25 forrest I've gotta go to a meeting StDiluted, let me know if that works.
17:25 platforms wait - that didn't list anything... back to my coffee...
17:25 StDiluted forrest: hm, ok. Didnt know you could use raw python functions in jinja
17:26 napperjabber joined #salt
17:32 MattJaynes joined #salt
17:38 Lue_4911 joined #salt
17:38 TJ1980 joined #salt
17:39 waverider joined #salt
17:39 waverider joined #salt
17:45 StDiluted forrest: that wqorked. Thank you so much.
17:45 StDiluted worked.
17:45 cjbarnes18 joined #salt
17:48 rustyrazorblade joined #salt
17:48 ze- joined #salt
17:49 forrest StDiluted, yea np.
17:49 waverider joined #salt
17:55 derelm joined #salt
17:59 bhosmer joined #salt
17:59 bemehow joined #salt
18:00 kermit joined #salt
18:00 Jahkeup joined #salt
18:02 MattJaynes Hi guys, I'm working on the final draft of my book "Taste Test" which implements a simple project in Salt, Ansible, Puppet, and Chef in order to show the basics of how each tool works ( http://devopsu.com/books/taste-test-puppet-chef-salt-stack-ansible.html ). Planned release is Aug 27th.
18:02 blee_ joined #salt
18:02 MattJaynes The only real downside that I point out for Salt is having to explicitly manage the dependencies since the directives aren't run in the same order as the source. I realize this may not be a negative for everyone, but in my client projects using Puppet, that was the most painful thing about working with the tool - particularly on systems with any complexity.
18:02 TJ1980 joined #salt
18:02 MattJaynes I've just heard (via https://github.com/saltstack/salt/issues/6114 ) that Salt will be getting an explicit ordering mode in 0.17 and I was wondering if anyone could tell me more about it? It'd be awesome if I can mention that in the book.
18:06 Katafalkas joined #salt
18:06 vaxholm_ joined #salt
18:08 KennethWilke MattJaynes: i'm not familiar at all with the explicit ordering, but there has been an option for numeric ordering for a lil bit: http://docs.saltstack.com/ref/states/ordering.html#the-order-option
18:08 MattJaynes Cool, thanks Kenneth
18:09 KennethWilke MattJaynes: for myself it's certainly less preferable than using a dependency chain that's more like a hierarchy, but i've seen this often used by chef users
18:09 KennethWilke including mysql when i initially dipped my toes into salt from chef
18:09 KennethWilke myself* not mysql, bleh
18:11 dthom91 joined #salt
18:12 MattJaynes hehe, yeah. There's a high cost of managing the dependencies when a system gets more complex. I'm hoping Salt's explicit ordering mode will make that a non-issue.
18:12 KennethWilke MattJaynes: seems like this'll be a very interesting book
18:13 ksalman shouldn't a host match with this top file, if it's is Debian or Scientific? https://gist.github.com/ksalman/6272266
18:14 ksalman it seems like the host explicitely need to be defined
18:14 MattJaynes Thanks :) My main goal is to get folks to pick a tool and use it. In my contracting, even in 2013 I'm constantly coming across businesses that are manually installed and without even an install checklist or even shell script in order to build new systems.
18:14 ksalman otherwise i get the error "No Top file or external nodes data matches found"
18:14 MattJaynes The real competition is people using no configuration management tool at all.
18:14 whit joined #salt
18:15 racooper does clusterSSH count as a configuration tool? :D
18:15 KennethWilke MattJaynes: while my bias is certainly towards salt this day, absolutely I'd choose any over none
18:15 KennethWilke ksalman: i think there may be a different way to do like a regex match on grains
18:16 KennethWilke ksalman: ala http://docs.saltstack.com/ref/cli/salt.html#cmdoption-salt--grain-pcre
18:16 KennethWilke though that's not in the top file
18:16 KennethWilke maybe match: grain-pcre though
18:16 MattJaynes KennethWilke: Spoiler... Salt and Ansible are the shining stars in the book. They're really the next generation and I think it will really help get people off the fence on whether to use a CM tool or not.
18:17 KennethWilke MattJaynes: i'd bet on those two as well
18:17 gordonm joined #salt
18:18 ksalman KennethWilke: I tried grain-pcre and i get the same error
18:19 forrest What price point are you planning to hit MattJaynes?
18:19 TJ1980 joined #salt
18:19 KennethWilke ksalman: aha! http://docs.saltstack.com/ref/states/top.html#other-ways-of-targeting-minions
18:20 forrest I've got some guys on my team I still can't convince that we need real config management over just RHN and kickstart.
18:20 KennethWilke underscore! not dash
18:20 ksalman ahh
18:21 KennethWilke hopefully that'll show more signs of success for you
18:21 KennethWilke damn picky 'puters!
18:21 ksalman KennethWilke: thanks
18:21 KennethWilke ksalman: you're welcome!
18:24 dthom91 joined #salt
18:27 Jahkeup joined #salt
18:28 faldridge_ joined #salt
18:29 VertigoRay Hey guys.  Writing a #!py init.sls and think I am have issues formatting my returned highstate datastructure properly.  Anyone know of a good tutorial on creating that return object?  Not really finding anything on the ether.
18:30 VertigoRay Actually, what I am finding on the ether is not helpful enough.  ;)
18:32 Khollowa_ joined #salt
18:34 alekibango joined #salt
18:35 KennethWilke VertigoRay: this may be helpful: http://docs.saltstack.com/ref/renderers/index.html#multiple-renderers
18:38 xmj joined #salt
18:39 Jahkeup joined #salt
18:40 VertigoRay KennethWilke: Thanks!  That's some new information for me to try :)
18:41 ggoZ joined #salt
18:47 rustyrazorblade joined #salt
18:48 saurabhs joined #salt
18:51 whit joined #salt
18:54 tomeff joined #salt
18:55 nliadm UtahDave: with master and minion on 0.16.3, still getting unacceptably long pauses, ZMQErrors, and minion key unaccepted errors.
18:57 xmj Has anyone in here tried salt with FreeBSD Jails in a master-jail / minion-jails setup?
19:01 cedwards xmj: I run FreeBSD master and all jails are minions, yes.
19:01 xmj cedwards: ever thought about jailing the master?
19:01 mattbarto joined #salt
19:01 xmj cedwards: plus, do you ezjail-admin create the jails?
19:01 s0undt3ch forrest: the only change on install_red_hat_linux_stable_deps is the rh-channel check?
19:01 cedwards xmj: I don't think I've ever jailed the master but I don't see why it wouldn't work
19:02 cedwards xmj: yes, I use ezjail-admin
19:02 xmj cedwards: very cool, thanks!
19:02 bmorriso joined #salt
19:03 s0undt3ch forrest: how about eh rh-channel check then calling install_centos_stable_deps if nothing extra needs to be done?
19:03 xmj cedwards: have you seen a good howto on those?
19:04 cedwards xmj: a good howto on ezjail-admin?
19:04 xmj cedwards: no, on salt with freebsd
19:04 cedwards xmj: ohh, I don't think so.. I suppose I should write something up.
19:05 xmj Oh, that'd be great.
19:05 bmorriso Is it possible to put raw python in a template/state? I need something like time.mktime(time.gmtime()) to render to 1376939120
19:05 cedwards xmj: fyi, I just fixed FreeBSD support in the salt-bootstrap so now you can easily bootstrap jails on first-start.
19:05 xmj cedwards: I'll be looking into salt on my own a bit, then see if I can get a VPS ... with different (minimal) jails for purposes like web/mail/db
19:06 xmj cedwards: I have no idea what that means, I hadn't heard about Salt until 30min ago.
19:06 chubrub joined #salt
19:07 cedwards xmj: haha. salt has a boostrap component for auto-installing Salt minions on VMs and other auto-deployed infrastructure.
19:07 cedwards xmj: so this morning I tied that into ezjail flavours so when I spin up a new jail it auto-installs salt minion and reaches out to the master
19:07 xmj aye, does that work with bhyve VMs as well?!?
19:07 it_dude joined #salt
19:08 xmj cedwards: damn, I could use that. heh
19:08 cedwards xmj: https://github.com/saltstack/salt-bootstrap
19:10 xmj /chuckles.. thanks for crashing my browser. (setting up 4 poudriere buildjails in parallel does that with github.)
19:10 cedwards xmj: btw, I'm the FreeBSD package maintainer so if you have salt related BSD questions, feel free to ping me
19:11 xmj noted!
19:11 chubrub hey everyone, just a quick question: can I use something like this in pillars {% if grains['env'] == 'prod' %} variable=something {% endif %} ?
19:12 berto- joined #salt
19:13 UtahDave chubrub: yep
19:14 bmorriso UtahDave: is this allowed in Salt's templating? http://stackoverflow.com/a/14101090
19:15 chubrub @UtahDave: thanks! :)
19:15 UtahDave bmorriso: Yeah, I think so.
19:15 bmorriso I basically need to render a value-epoch for uniqueness
19:16 mwillhite joined #salt
19:20 jpeach joined #salt
19:23 bwghughes joined #salt
19:28 bmorriso left #salt
19:32 brew joined #salt
19:33 mgw Is anyone aware of any work with salt+docker?
19:36 rustyrazorblade joined #salt
19:37 UtahDave mgw: I've heard a few people talking about it.  I don't think there has been any official work on it.
19:38 mgw UtahDave: thanks
19:38 xmj cedwards: do you know if the ezjail-admin writer is around on freenode/efnet ?
19:39 cedwards xmj: i know they've got a channel, but last time I was in it things were deadly quiet.
19:42 aboe joined #salt
19:42 brianhicks joined #salt
19:45 Nobbl joined #salt
19:46 MattJaynes UtahDave: Do you happen to have any more info on the explicit state ordering mentioned here? https://github.com/saltstack/salt/issues/6114
19:48 forrest sorry s0undt3ch, was grabbing food. Yea that's fine by me, I wasn't sure whether each function should be explicit when there are changes.
19:49 s0undt3ch forrest: the less duplication, the better ;)
19:49 forrest ok, I'll close that one and update then resubmit
19:51 giantlock joined #salt
19:51 UtahDave MattJaynes: I think it has been implemented in git develop, but it doesn't look like it's been added to the docs.  I'll ask about it
19:53 forrest should be good to go now s0undt3ch, sorry about that.
19:53 MattJaynes UtahDave: Awesome, thanks :)
19:56 s0undt3ch forrest: nah, no need for sorries ;)
19:56 forrest heh
19:56 forrest "Let me just ask all these questions about this implementation, so you can type more than implenting it yourself would take" :P
19:56 s0undt3ch lol
19:56 s0undt3ch :)
19:58 s0undt3ch forrest: actually, that rh-channel issue is for rhel 5 and 6?
19:59 druonysus joined #salt
19:59 druonysus joined #salt
19:59 xmj cedwards: ha. well, I'll have to rely on you to write a good tutorial about salt+jails (;
19:59 forrest oh you're right it's only rhel6, another exciting chapter in epel..
19:59 david_a joined #salt
19:59 george_torwell joined #salt
20:00 cedwards xmj: I just realized why your nick looked familiar. We spoke the other day in #freebsd-python about why sysutils/py-salt was failing on testport.
20:01 UtahDave MattJaynes: Ok.  add the following to your master config:      state_auto_order: True
20:01 UtahDave It's in git develop and will be included in the 0.17.0 release
20:02 s0undt3ch forrest: k, I'll wait for the update :)
20:02 forrest yup yup
20:02 rustyrazorblade joined #salt
20:03 jacksontj joined #salt
20:03 chubrub wow, explicit ordering in 0.17 - that's just awesome!
20:03 MattJaynes UtahDave: Awesome! I'm asking since I'd like to mention it in the book. Would it be correct to say that enabling this mode will cause the states to follow their order in the source? Like Chef and Ansible?
20:04 kermit joined #salt
20:05 kermit joined #salt
20:06 xmj ah, true
20:06 jacksontj I'm seeing some weird issues with one of my salt deployments-- it seems that the minions are loosing their connections/authentication with the master. So every time i run a command on the minion (salt-call) it reauthenticates with the master
20:06 jacksontj which is causing some problems because the master can't handle every host reauthing all the time ;)
20:07 xt MattJaynes: there's also the -order: 1, and - order: last paramters.. http://docs.saltstack.com/ref/states/ordering.html
20:07 jpeach UtahDave: I have a system with dozens of minions stuck … http://apaste.info/hOKg
20:07 jpeach what Ifo do you want me to gather for a bug?
20:08 forrest jacksontj, I assume you confirmed it isn't a network connectivity problem?
20:08 jacksontj yea, i have connectibity
20:08 jacksontj *connectivity
20:08 forrest no weird packet loss though?
20:08 jacksontj and it does connect-- just seems to not want to keep that connection
20:08 MattJaynes UtahDave: Googled it and it looks like "Yes" is the answer to my question about the explicit ordering: https://groups.google.com/forum/#!msg/salt-users/2BwPeHmoUas/7NP7bPAQJNEJ
20:08 chubrub @xt: yeah, but requisites takes precedence over -order, and you have to control all  'numbers' inside sls
20:08 forrest ahh ok jacksontj
20:09 jacksontj so all the pings are >0.1s every call to test.ping results in an authentication request/accept on the master
20:09 jacksontj (tailing the logs ;))
20:09 UtahDave MattJaynes: yes, this mode will cause states to follow the order in the sls files.
20:10 craig_ jacksontj: hey
20:10 alunduil joined #salt
20:10 MattJaynes UtahDave: That's great news. The ordering issue was the only reservation I had about recommending Salt to newcomers. This is a really great addition. Kudos guys.
20:11 UtahDave thanks!
20:11 UtahDave jacksontj: what version of zmq are those minions using?
20:11 jacksontj zeromq3-3.2.3
20:11 platforms UtahDave: is the ordering still informed by requires for the edge cases, or does it strictly follow the order of the directives?
20:11 jacksontj so, to help a little bit ihave a lot of salt deploys
20:12 jacksontj the other ones are all running the same versions of stuff-- and don't have this issue
20:12 jacksontj just to make it fun
20:12 craig_ however, there was a change in the network arch with the new hosts
20:12 jacksontj correct
20:12 jacksontj the other deploys are on more "flat" networks (more L2)
20:12 UtahDave platforms: I believe you can still set requires and all that. It just defaults to the order in the sls files first
20:12 craig_ we were thinking about increaing the zeromq timeout
20:12 jacksontj this one is mostly L3
20:13 craig_ but, the constant auth requests from the minions is odd
20:13 jacksontj well, since i have to re-auth every request it makes me think that the connection isn't kept between calls
20:14 jacksontj so i'd guess that its timing out really fast
20:14 UtahDave MattJaynes: how's the book coming along?
20:14 jacksontj not sure why though
20:14 UtahDave jacksontj: Yeah, that's odd.
20:14 UtahDave jacksontj: which version of salt?
20:14 craig_ UtahDave: is the auth cahced in a file on the minion?
20:14 jacksontj 0.15.3
20:15 UtahDave craig_: I'm not sure. I'm only familiar with the intricacies of the auth stuff at a high level
20:15 UtahDave jacksontj, craig_:   any way you can test a few of those minions on 0.16.3?
20:16 jacksontj yea, i guess we could
20:16 craig_ uh.... wait
20:16 jacksontj i was thinking of trying some timeout stuff first
20:16 craig_ ya ^^ :)
20:16 MattJaynes UtahDave: I'm doing final updates to the book based on feedback from my reviewers. I'm planning to launch it Aug 27. I'll send you a copy of my next draft if you like.
20:17 craig_ i am seeing the pki dir on the master getting updated constantly
20:17 tyler-baker joined #salt
20:17 jacksontj craig_: thats a byproduct of the re-auths ;)
20:17 faldridge joined #salt
20:17 craig_ right... but what is the equiv on the minion
20:17 tyler-baker anyone have a moment to help me with a file.recurse issue?
20:17 UtahDave MattJaynes: yeah, that would be awesome!
20:18 UtahDave tyler-baker: what problem are you having?
20:18 forrest MattJaynes, I think it got missed earlier, what price point are you aiming for?
20:18 craig_ if the auth is cached in a file, maybe the minion cant write to the file it expects
20:18 tyler-baker UtahDave, http://pastebin.com/qFaueAuk
20:18 UtahDave craig_: on the minion   /etc/salt/pki/minion/
20:18 craig_ ya, those files are NOT being rewritten
20:18 tyler-baker UtahDave, http://pastebin.com/uNxNrbYC
20:19 craig_ -rw-r--r-- 1 root root  796 Jun 20 19:49 /etc/salt/pki/minion/minion_master.pub
20:19 craig_ -r-------- 1 root root 1850 May 31 23:27 /etc/salt/pki/minion/minion.pem
20:19 craig_ -rw-r--r-- 1 root root  796 May 31 23:27 /etc/salt/pki/minion/minion.pub
20:19 tyler-baker UtahDave, The file.recurse works on the master (which is also a minion), but not on the other minions
20:19 hernantz joined #salt
20:19 jacksontj weird...
20:19 jacksontj why would they not be written :/
20:19 craig_ my point with the master pki files is that if thare are being constantly rewritten, then there should be something equiv on the minion
20:19 craig_ exactly :)
20:20 platforms MattJaynes: I'm still learning salt, but having the flexibility to use dependencies (puppet style), explicit ordering, or auto ordering (chef style) is a big plus. I've been in dependency hell today, and this is good news to me.
20:20 craig_ why is minion.pem 400?
20:20 craig_ is taht right?
20:20 tyler-baker UtahDave, the whole salt config is here: https://github.com/EmbeddedAndroid/salt
20:20 rustyrazorblade joined #salt
20:20 craig_ ya.. sorry... nm, that does make sense
20:20 craig_ typing before thinking :(
20:21 MattJaynes forrest: Great question. Since it's my first book for DevOps University, I'm tempted to just give it away for free to get things going. However, I need revenue in order to keep from having to pause everything and take contract work, so I may decide to charge about $30 for it.
20:21 MattJaynes UtahDave: Cool, I'll send it when it's ready :)
20:22 UtahDave MattJaynes: cool
20:22 jacksontj craig_: looks like you don't update those filexs on the minion
20:22 jacksontj unless they get deleted
20:22 craig_ so, how does the minion know that its currnet auth is ok?
20:22 MattJaynes platforms: Yeah, it's a big win. I've lost months of my life in dependency hell on complex systems.
20:22 craig_ it looks like its not holding that info correctly
20:23 craig_ if its in the proc itself, then this probably isnt the problem
20:23 craig_ but, if its in a file....
20:23 jacksontj i think its in the proc
20:23 UtahDave tyler-baker: so basically your file.recurse is looking for salt://lava/device-types and it's not finding it.
20:24 tyler-baker UtahDave, Yeah, am I doing something stupid? This will work if I remove clean and include_empty
20:25 tyler-baker UtahDave, However it was not deleting the files, if they were removed on the master
20:25 tyler-baker UtahDave, so I tried adding clean and include_empty and I'm seeing this error on my minions http://pastebin.com/uNxNrbYC
20:26 UtahDave tyler-baker: you at least have to have an empty directory at   salt://lava/device-types
20:27 tyler-baker UtahDave, DOH!
20:27 tyler-baker UtahDave, my git repo did not add a directory :\
20:27 tyler-baker UtahDave, I knew it was something stupid :)
20:27 UtahDave git doesn't track empty directories.
20:27 tyler-baker UtahDave, yeah I had files in there, but removed them to test. my mistake
20:28 UtahDave ah, ok.
20:28 UtahDave cool
20:28 EugeneKay !dirs
20:28 tyler-baker UtahDave, I appreciate the slap in the right direction!
20:28 EugeneKay Oh right, wrong channel.
20:28 UtahDave :)  thanks for the blankies, EugeneKay.   lol
20:29 EugeneKay You can "fake" an empty directory in git by making a blank .gitignore file(or .whatever)
20:29 tyler-baker EugeneKay, doing that right now :)
20:30 jacksontj so, it looks like that authentication request/accept stuff is normal
20:30 jacksontj thats just the AES key handshake-- not the key stuff
20:30 forrest MattJaynes, is that price point going to be for the digital version?
20:31 forrest s0undt3ch, you should be good to merge in https://github.com/saltstack/salt-bootstrap/pull/211
20:33 TheCodeAssassin joined #salt
20:33 MattJaynes forrest: Yes, but just shoot me an email and I'll send you the next draft version if you'd like. I'm matt@nanobeep.com
20:33 TheCodeAssassin hi guys
20:33 jacksontj looks like all the master load is for authentication stuff and module sync :/
20:33 forrest That woudl be great, thanks MattJaynes!
20:33 TheCodeAssassin i am very new to salt, and i have some nooby questions
20:33 kevin__ joined #salt
20:34 MattJaynes forrest: Sure, np :)
20:35 dthom91 joined #salt
20:35 TheCodeAssassin In our case we want to boostrap new nodes masterless initially, we spawn a new proxmox openvz container and then run a base state
20:35 TheCodeAssassin so far, nothing special we copy a sls file and run state.highstate in a masterless state
20:36 kevin__ When creating a salt-master using salty-vagrant is there any reason a custom minion_config is needed?
20:36 mattbarto joined #salt
20:38 kevin__ Currently the custom master_config only gets configured if a custom minion_config is also provided
20:38 TheCodeAssassin now, we have other sls files, i just don't know if its a good idea to execute a state file per role
20:38 TheCodeAssassin for example webserver.sls or memcache.sls
20:39 TheCodeAssassin or if there is a better way of doing it
20:39 UtahDave akoumjian: do you know the answer to kevin__'s question?
20:39 david_a joined #salt
20:39 UtahDave jacksontj: yeah.  So there are a couple things you can do
20:39 TheCodeAssassin tryed it with puppet, too much stuff and it got really complex, real fast
20:40 UtahDave salt-call --local cache=True   will cause the minion to not contact the master and use the states it has cached from the last highstate
20:40 UtahDave jacksontj: is your new master running on the same class hardware as the other masters?
20:40 jacksontj yea
20:41 UtahDave jacksontj: also, how many worker threads do you have?
20:41 jacksontj 20
20:41 jacksontj well processes
20:41 jacksontj but 20 of them
20:41 akoumjian kevin__: Not sure, will look into it
20:41 jacksontj and the load average hits 20-- so its getting hit with something
20:41 jacksontj the load on the other master is usually ~15
20:41 jacksontj which is (arguably) a bit high
20:42 jacksontj but 20 is too high ;)
20:42 UtahDave you might try lowering the worker_threads a little
20:42 BrendanGilmore joined #salt
20:42 UtahDave you might be trying to do too much concurrently on the amster
20:42 jacksontj we have 24 cores
20:42 jacksontj should be fine..
20:43 UtahDave ssd?
20:43 jacksontj spnning disk
20:44 kevin__ akoumjian: if you take a look at need_configure() in vagrant-salt/provisioner.rb, it's only triggered by minion settings
20:44 akoumjian kevin__: Yep, saw your issue.
20:45 UtahDave hm
20:47 kevin__ akoumjian: Awesome, thanks
20:48 whiskybar hi everyone, can you give me an advice on writing grains
20:48 dthom91 joined #salt
20:49 akoumjian kevin__: Are you testing out multiple vm deployments with vagrant?
20:49 whiskybar can I access pillar from within grain public function?
20:50 alexandrel whiskybar: usually it's the otherway around you use grain based logic to change pillar values.
20:50 alexandrel I say usually, but I mean: it's what I saw so far out there.
20:51 kevin__ akoumjian: Yes, we are currently using salt-cloud to deploy our project and are trying to simulate the same configuration locally using vagrant for testing purposely
20:53 whiskybar alexandrel: that's unfortunate but I understand. I needed to pull some very static network info using public services. Thanks
20:56 UtahDave whiskybar: grains are one of the first things evaluated, so there's not much available at that point. No pillar.
20:56 UtahDave whiskybar: But you can call out to public services from your grains to get data.
20:56 whiskybar okay
20:56 it_dude joined #salt
20:56 whiskybar I wanted to put that public services' address to the pillar
20:57 EugeneKay I should really look into doing DNS lookups in grains to get my info, rather than shoving it into pillar
20:57 jpeach whiteinge: with the formulas proposal, are state IDs still global?
20:57 whiskybar but I guess I have to start putting concrete values somewhere
20:58 whiteinge jpeach: yes. formula get included into the state tree as ordinary states
20:58 Teknix joined #salt
20:59 jpeach so if I install the memcache formula, I need to globally guarantee that I have no other state that attempts to instal the memcached package?
20:59 zooz joined #salt
21:00 whiteinge jpeach: the best practices to mitigate potential problems is to keep formula as simple and guess-able as possible. e.g., if you "include: - apache" in one of your states you should be able to guess that state probably is a "pkg" and a "service" state
21:00 whiteinge jpeach: yes to your example
21:01 jpeach gotcha
21:01 jpeach that just seems really difficult to manage, particularly if you want to share states written by third parties
21:02 whiteinge one obvious/simple way to work around that would be via naming convention, same as the repos themselves. e.g., "include: - mysql_formula"
21:03 whiteinge jpeach: how would that sit with you?
21:04 jpeach yeh, if there was a consistent naming convention for the IDs that would help a lot
21:05 whiteinge i've wondered about that problem too. it would make a lot of sense to "namespace" both states and pillar values that way
21:05 jacksontj UtahDave: any other suggestions on whats eating CPU?
21:05 jpeach personally I think the global scoping is a mis-feature, but since I'm not going to pony up a patch I'll not complain too hard :)
21:05 UtahDave whiteinge: is there a convention for salt-formulas for specifying an example master config?  for example  a    master.example that contains the config snippet necesary to allow certain minions to access the peer interface?
21:06 whit joined #salt
21:06 whiteinge UtahDave: each state should have it's own section in the README with docs specific to it
21:07 UtahDave jacksontj: I'm kind of at a loss.  Is the workflow much different on this new environment?
21:08 ze- hey. is there a good way in pillar sls files to get the minion's id ?
21:08 jacksontj not that i'm aware of-- it should be pretty much the same (which is why i'm so confused)
21:08 jacksontj is there a good way to gell what the master is doing? i'm wondering if something is slower in there which is causing backup of some time
21:08 jacksontj *type
21:08 craig_ UtahDave: we purposefully kept things as much in sync as possible
21:08 craig_ everything is installed via RPM
21:08 craig_ so, it shuold be pretty consitent
21:09 UtahDave ze-: {{ grains['id'] }}
21:10 UtahDave jacksontj: you could stop that salt-master service.  Then run the master in the foreground  with   salt-master -l debug
21:10 craig_ ugh.....
21:10 jacksontj is that different output then the logs?
21:10 UtahDave that should give you a better idea of what's going on.
21:11 jacksontj i changed the log level up to debug and saw nothing that i didn't know (lots of file serving etc-- nothing stands out
21:11 craig_ with 7k minions... that would be "overwhelming" to say the least
21:11 jacksontj last time i saw this there was some issue with a bunch of minions having the wrong master key so the minion_master was wrong
21:11 jacksontj so i saw lots of "potential compromise of the cluster"
21:11 jacksontj but for this issue it all looks normalish
21:11 craig_ and that issue showed up in the minion logs
21:11 UtahDave ah, yeah.
21:12 mgw How are others dealing with state runs that may require a reboot? E.g., ensuring kernel is a certain version.
21:12 ze- UtahDave: can't a minion override it's grain['id'] with /etc/salt/grains ?
21:14 dthom91 joined #salt
21:14 jacksontj but it does still seem to be "working"-- jobs go out and get results it just takes a LONG time
21:14 jacksontj have to use jobs.lookup_jid
21:16 djangofan joined #salt
21:17 djangofan I was thinking of using Salt  to configure about 15 vms that I have, all controlled by Jenkins currently.  What would I expect this to cost me, in terms of $$$ ?
21:17 forrest djangofan, what do you mean?
21:17 forrest you just need a master to run salt on
21:17 djangofan it appears that Salt costs money?  am i wrong?
21:17 forrest djangofan, salt does not cost money
21:18 forrest if you want the corporate support/setup it does
21:18 ze- UtahDave: hostid: {{ grains['id'] }}
21:18 djangofan excellent.  i'll take another look. thanks!
21:18 ze- zetest-sk:
21:18 ze- ----------
21:18 ze- hostid:
21:18 ze- foo
21:18 ze- not safe way to get/check it.
21:18 forrest np djangofan, take a look at the walkthrough: http://docs.saltstack.com/topics/tutorials/walkthrough.html
21:18 djangofan ok.  looking at it now.  thanks.
21:19 whiteinge djangofan: are you going to djangocon? http://www.djangocon.us/schedule/presentation/2/
21:20 djangofan no. Djangofan refers to Django Reinhardt.  But, my friend , who is also a Django Reinhardt fan, is the creator of Django framework.
21:21 UtahDave ze-: you might try in __opts__  I think that may refer to the minion as well
21:21 Zurullo joined #salt
21:22 djangofan Oh, I see.  SaltStack seems to be used by Django Framework users/
21:22 ze- UndefinedError: '__opts__' is undefined
21:22 whiteinge Django Reinhardt fans are also welcome :)
21:22 ze- hostid: {{ __opts__ }}
21:22 ze- not accessible in the env we have at that point. :(
21:23 whiteinge i got turned on to his music by adrian holovoty after i started working with django :)
21:23 djangofan yep, thats him.   yeah, his site "Soundslice.com" is excellent.
21:23 xt you discovered the framework before the guitarist? sad!
21:24 whiteinge srs!
21:24 UtahDave ze-: sorry, try the __opt__ dict
21:25 ze- same
21:26 ze- in the code, there is alot of __opts__. so it would be with the s
21:28 mike25ro hey guys.. .me again with a stupid Q... let's say i have a debian minion which is a blank/standard install... and i want to set it up as a lamp server (just for the fun of it at this point) my Question is related to package management.... i need to install ..... mysql-server mysql-client [set up the password here]  then apache2 php5 libapache2-mod-php5 php5-mysql php5-curl php5-gd php5-idn php-pear php5-imagick php5-imap  php5-mcrypt php5-memcache php5-ming
21:28 KennethWilke no mike25ro: you don't have stupid questions, sorry!
21:29 mike25ro KennethWilke: i just want to make sure i get the BEST solution... and not start with the LEFT foot
21:29 KennethWilke :p
21:29 mike25ro should i set up a pkg state for each package?
21:30 KennethWilke only if you want to! but generally you can provide them as a list
21:30 mike25ro what would you guys do ?
21:30 KennethWilke looking for the doc to show ya
21:30 Jahkeup joined #salt
21:30 mike25ro KennethWilke:  thanks a lot buddy
21:30 KennethWilke mike25ro: http://docs.saltstack.com/ref/states/all/salt.states.pkg.html#salt.states.pkg.installed
21:30 KennethWilke look at pkgs:
21:30 UtahDave mike25ro: no, you'll want to do a pkg.installed with the   - pkgs  option.  That way they all get installed with one call to your package manager
21:30 KennethWilke you can provide all those goodies as a list under one ID
21:31 mike25ro i was telling the other days UtahDave ... that ... i will be starting some screencasts/video tutorials... this is one of them
21:31 KennethWilke mike25ro: +1 <3
21:31 KennethWilke people will love you for it
21:31 UtahDave mike25ro: what KennethWilke said.
21:31 mgw How are others dealing with state runs that may require a reboot? E.g., ensuring kernel is a certain version.
21:31 KennethWilke and if people won't, i will
21:31 UtahDave mike25ro: nice!  that will be great
21:31 mike25ro UtahDave:  KennethWilke thanks guys... one more Q... when you install mysql-server... there is an interactive window asking for the root pass... how is this handled?
21:32 UtahDave mgw:  I don't think we have a clean way of dealing with that right now.
21:32 mgw UtahDave: that's what I thought :-(
21:32 mgw *goes to think up clean ways to deal with that*
21:32 KennethWilke mike25ro: i think it's skipped when ran non-interactively, but not on gentoo
21:32 UtahDave mgw: you can call module.run on a system.reboot
21:32 KennethWilke mike25ro: also! please consider https://github.com/saltstack-formulas/
21:32 mike25ro KennethWilke: on debian there it is asking you for the pass
21:33 UtahDave mike25ro: yeah, Salt suppresses those interactive questions
21:33 mgw UtahDave: Not sure what you mean?
21:33 mgw s/?/.
21:33 mike25ro KennethWilke: YES i am reading a lot from there as well... too much reading :) it's time to test more..... and do the tutorials as well... i know i will waste a lot more time... BUT i am sure ... LAMP is a really needed solution... so i think it's worth sharing with the others
21:33 KennethWilke mike25ro: I would hope that the {mysql,apache,php}-formula repos would help you tackle all of this!
21:34 mike25ro KennethWilke: thanks buddy
21:34 UtahDave mgw: there's an execution module calles 'system' that has a 'reboot' function
21:34 KennethWilke mike25ro: no problem at all
21:34 UtahDave mgw: the module.run state function allows you to call execution modules in a state file
21:34 mike25ro KennethWilke:  i am still writing... on my notebook... ideas on how to start doing things.... and how to start a series of tutorials .... that is useful
21:35 KennethWilke mike25ro: i'm not against reinventing the wheel, sometimes you make a better wheel in the process
21:35 KennethWilke these salt-formulas are in their fairly early stage, i'd be happy if one of your tuts was how to improve them ;)
21:36 mgw UtahDave: I see. So it would be sort of like… reboot once the kernal is updated… which will abort the remainder of the run
21:36 mgw which makes it messy
21:36 UtahDave mgw: yes.
21:36 UtahDave exactly
21:36 jacksontj joined #salt
21:37 mgw what we need is a way for the minion to pick up after the reboot
21:37 mgw or something like that
21:37 forrest I've been looking at options with the API to do that mgw
21:37 UtahDave even just re-running the state should work, since states should be idempotent
21:37 forrest so you can restart, then it will pick back up automatically
21:38 forrest but it's not pretty, gotta basically do a bunch of waits/loops to try to hit the machine.
21:38 mgw UtahDave: yeah, so just a flag somewhere to tell it to continue highstate or whatever on reboot
21:38 mgw (re run the current job, really)
21:38 UtahDave mgw: there is the startup_states: highstate    option in the minion config
21:38 mgw forrest: yeah, that's the messy part
21:38 mike25ro KennethWilke: i am looking at ... https://github.com/saltstack-formulas/mysql-formula and i saw that there is a jinja thingy that is written : {{ server.get('user', 'mysql') }}  in the file :  https://github.com/saltstack-formulas/mysql-formula/blob/master/mysql/files/Debian-my.cnf    i haven't seen that approach to get a pillar data...  this is smth new... i have to study this one a bit
21:38 forrest mgw, yep I don't like it at all
21:39 forrest but I'll be damned if I'm gonna sit there and baby sit the machine as it comes back up
21:39 mgw UtahDave: I'm thinking something like this though…. I wouldn't want my hightstate to potentially reboot the machine
21:39 mgw so I'd more likely use state.sls to provision that aspect
21:39 KennethWilke mike25ro: yeah the definitions are up there at the top, those methods are normal python dict()-type methods i believe
21:40 KennethWilke it's a handy way to say, if this doesn't exist in the dict, then give me arg[1] anyway
21:40 mike25ro yeah i have to study this approach :D
21:40 forrest mgw, I think that's a good idea, especially because then you could incorporate the big functionality that I need, which is putting a server back into rotation
21:40 mgw So maybe it's best to just do this stuff in a script that calls state.sls stuff and then cmd.run reboot
21:40 forrest ugh, that feels so hacky though
21:41 mgw A runner perhaps
21:41 mgw instead of a shell script
21:41 mgw If it could be generalized
21:41 mike25ro KennethWilke: i love the approach of using a package-map.jinja to map app depending on the platform.. i was just thinking how to manage that ... there is a really good solution
21:42 mike25ro pff..so much to remember :D
21:42 forrest mgw, I don't know, still feels like a lot of manual labor could occur there
21:42 mgw forrest, UtahDave: maybe a runner that takes a sequence of actions?
21:42 kenbolton joined #salt
21:42 mgw (thinking out loud)
21:42 KennethWilke mike25ro: i'm glad you like my solution :) it's only a few days old
21:43 * forrest shrugs
21:43 forrest mgw I don't have a good solution
21:43 KennethWilke mike25ro: and based on whiteinge's better solution
21:43 mgw My immediate task is to a) update the kernel b) reboot c) run highstate
21:43 forrest if I did, I would have already made a pull request, or a documentation update :P
21:43 UtahDave mgw: yeah, you could do it that way.  The overstate might be a good option there
21:43 forrest Yea you basically need the same style of action I do
21:43 mgw UtahDave: utilizing overstate in what way?
21:44 forrest except for mine I want to a) move file to pull from LB B) update kernel (which is usually done already) C) reboot D) run a highstate, which ensures the lb file is back
21:44 UtahDave on overstate is basically a runner.  You might be able to run an overstate that updates the kernel, reboots the machine, and then when it comes back up run a highstate
21:45 bhosmer joined #salt
21:45 timl0101 joined #salt
21:45 mgw state.over is a runner that reads steps from an sls, I see
21:45 forrest how does salt handle the wait time between when you issue the reboot command and when the box is back up though UtahDave? There's a few second window between reboot and actual shutdown where it can't be communicated with
21:46 mgw UtahDave: yeah, what forrest said…. I think it will start running the highstate as the machine is shutting down
21:46 forrest as do I
21:46 mgw state.over needs another directive of some sort
21:46 forrest that's why I was experimenting with the api mgw
21:46 UtahDave forrest: that's true. I think a function to wait for the master to come back up would need to be added to the overstate
21:46 ze- UtahDave: {{ opts['id'] }}
21:46 forrest but at that point, you're just writing python code
21:46 KyleG left #salt
21:47 UtahDave ze-: ah, there you go!  sorry, I knew it was something like that.
21:47 mgw forrest: I think UtahDave means as a general functionality of state.over
21:47 mgw rather than a one off
21:47 forrest mgw, it would be nice if there was just support for a reboot, I feel like there should be some sort of flag for pre reboot, and post reboot actions
21:47 forrest Oh yea I know what he means mgw, that's what my hope was when I first started investigating :P
21:48 forrest that was just the only solution I could come up with where it MIGHT survive the reboot and continue running the job
21:48 ze- UtahDave: well, at least now i have a way to get the proper pillar data only for the proper minions :)
21:48 mgw I'll take a quick look at the overstate runner
21:48 mgw maybe it's not too complicated
21:49 UtahDave mgw: a reactor might allow you to do some of this dynamically, too
21:49 forrest It's just that window between the shutdown command and the reboot actually occurring that I am not sure about
21:49 mgw UtahDave, looking at that too
21:49 UtahDave mgw and forrest: Why don't one of you open an issue on github so we can discuss this? Tom might have some insight into the best way to do this.
21:50 ze- what would you think about a state with - pkg.latest: salt-master ?
21:51 forrest Sounds good UtahDave, you wanna open it or do you want me to mgw?
21:52 mgw forrest: go ahead
21:52 UtahDave yeah, thanks, forrest. that would be great if you would!
21:53 forrest yea sure, doing so now. mgw when I finish I'll link it in here so you can add your use case of the kernel update.
21:53 mgw I'll comment after I've looked at the overstate code some more
21:53 ze- oh... hehe, wasn't reading what they were about, but it seems i'd have a similar issue. :) pkg.lastest: salt-minion (not master) restarting the client (just the service, not the whole box) :)
21:57 danielbachhuber joined #salt
21:58 mike25ro KennethWilke: really good job with the mysql package. do you have other packages / states ?
21:59 matanya joined #salt
21:59 KennethWilke mike25ro: thank you! I added a lot to the salt-formula repo and well, and the sudoers-formula's about to get my touch as well as the rabbitmq-formula
21:59 KennethWilke hopefully the openstack-standalone-formula as well, but not yet
21:59 kenbolton joined #salt
22:00 KennethWilke works in progress :)
22:00 mike25ro sudoers is smth i want to take a look at :)
22:00 mike25ro KennethWilke:  keep up the good work.. Are you one of the devs behind salt or just  a... contributor?
22:00 forrest UtahDave, mgw, https://github.com/saltstack/salt/issues/6792
22:01 KennethWilke mike25ro: just a lowly fanboy! i work at rackspace
22:01 forrest Feel free to comment/modify for clarification if I wasn't being clear enough
22:01 mike25ro KennethWilke: it's really awesome that the salt community has people like you
22:01 KennethWilke mike25ro: oh and i meant to say thanks for the compliment :)
22:02 KennethWilke mike25ro: and you!
22:02 mike25ro KennethWilke:  ... i hope soon i am changing jobs.. and i really want to impose salt to the new german company i hope i will be moving (final interview in 2 days)
22:02 KennethWilke i still blame UtahDave for helping foster a strong community
22:02 mike25ro UtahDave: is the "pillar" here :) as far as i can see
22:03 aat joined #salt
22:03 forrest Hey KennethWilke, if you don't mind me asking, do you know if they'll ever implement the ability to spin up VMs in the private clouds like they do in the public cloud?
22:03 mike25ro comparing to puppet... in ... 3-4 years puppet is history KennethWilke
22:03 forrest apparently you have to call in or something currently to get another VM
22:04 KennethWilke forrest: i'm not very familiar with private cloud :\
22:05 forrest All good, would be unfair of me to presume everyone knows everything about all the products
22:05 KennethWilke forrest: lol, by spin up though do you mean via the api or such?
22:05 forrest I'm just talking about the front end GUI
22:06 forrest where you can go in and just create a new VM
22:06 KennethWilke oh, like a web interface?
22:06 dthom91 joined #salt
22:06 forrest apparently in the private cloud you have to call in to get a new VM created (I'm not extremely familiar with the private cloud, different team is working with it)
22:06 forrest yea just the standard account you log into
22:06 forrest so instead of spinning up a VM in 5 minutes like you can do on a normal account, it can take up to 24 hours (though I don't believe they have ever waited that long)
22:07 forrest I already explained they should be using openstack :P
22:07 KennethWilke forrest: ooooooooooo, i dunno about through the same interface but the openstack project horizon should be able to do that
22:08 forrest ok cool, anyways back to salt, thanks.
22:08 KennethWilke no problem
22:08 renoirb Hey guys, how is the array merging with yaml configuration
22:08 renoirb e.g. I have in /srv/master a file_roots:
22:08 Jahkeup joined #salt
22:09 renoirb and later, in a /etc/salt/master.d/dev.conf with other configs for file_roots: ?
22:09 renoirb e.g. I would have a salty-vagrant workspace with proper mount point to master.d/  and for prod, no messing with default /etc/salt/master config
22:10 * mike25ro says good night... is 1:09 am here... c u 2morrow :)
22:10 mgw UtahDave: is there a reason for /var/log/salt/minion to be world readable?
22:10 mike25ro left #salt
22:10 renoirb Hey UtahDave, I saw your fork with my MultipleVagrant gist
22:10 renoirb :)
22:11 UtahDave renoirb: :)  yeah, I thought it was really cool. I haven't had a chance to take it for a spin, so forked it so I wouldn't forget where I saw it.  :)
22:12 UtahDave mgw: I don't think so. I think the minion just needs to be able to read and write to it
22:12 mgw UtahDave: that's not what I meant
22:12 renoirb UtahDave: nice, thx.  I'll surely publish it in https://github.com/webplatform when I am done with our deployment system
22:12 mgw Although it's what I said
22:13 mgw What I meant is: "Why is /var/log/salt/minion world readable by default?"
22:15 UtahDave mgw: Ah, sorry.  :) I'm not sure. i wonder if the os package is setting it up that way.
22:15 UtahDave renoirb: nice!
22:15 mgw Corey, do you know?
22:16 renoirb UtahDave: do you have any suggestions for me to have a separate /etc/salt/master.d/file.conf that extends the file_roots:
22:16 renoirb how it reacts with the array
22:16 UtahDave mgw: just checked.  mine's also world readable, too
22:17 mgw UtahDave: What distro are you on?
22:17 UtahDave renoirb: I think anything in a *.conf will overwrite the option in the main config file.
22:17 UtahDave mgw: ubuntu 12.04
22:17 jacksontj joined #salt
22:17 renoirb oh, no merge. Fair enough.
22:18 UtahDave renoirb: do you handle all the architecture for webplatform?
22:19 renoirb UtahDave; yes, with the help of somebody else.
22:19 renoirb got to go now.
22:19 UtahDave cool. talk to you soon
22:20 cxz joined #salt
22:21 UtahDave KennethWilke: My first quick glance at your sudoers-formula brought me joy
22:21 KennethWilke UtahDave: nice! i'm about to head out hopefully it'll bring even more joy tomorrow when it does stuff :p
22:22 forrest Oh nice KennethWilke, good to see that one getting some love.
22:22 UtahDave :)
22:22 forrest Myself and the other centos/rhel guys need to go through those and add support for our stuff.
22:23 KennethWilke forrest: thanks, I've been working with whiteinge to try to figure out some best practices for making common things like package names distro-agnostic
22:23 forrest yea whinteinge has some good examples (even if they are Debian :P)
22:23 whiteinge debian?! i never touch the stuff!
22:24 Jahkeup okay I still can't get my _states to work :/ I can't import a _module to use functions :C
22:24 KennethWilke forrest: i don't even do the redhat family! i'm a gentoo guy! :)
22:24 KennethWilke doesn't mean i don't care about the redhedded ones :)
22:24 forrest hah
22:24 KennethWilke redhat was nice enough to give me a cert after all :p
22:25 forrest heh, now if they'd just do some more of the RHCA trainings I'd really appreciate it
22:25 KennethWilke oooo haven't gotten quite that far along with it :p
22:26 forrest oh you're right whiteinge, the commits were before you moved it
22:26 KennethWilke i'd want an RHCA just to be one of the few
22:26 forrest ahh you have your rhcsa?
22:26 forrest or RHCE?
22:26 forrest and I agree, I'd like to start working towards RHCA, but the lack of classes/locations is a bit tough on the old wallet.
22:26 KennethWilke yeah, i studied for the E but then became a dev working on debian based boxen
22:26 KennethWilke never took the test
22:27 forrest Ahh, well, it's not as bad on RHEL6, especially with the selinux changes.
22:27 whiteinge forrest: haha. it's ok, i do occassionally touch debian when i have to  ;)
22:27 KennethWilke so i just have my lowly SA
22:27 KennethWilke :p
22:27 forrest It's just a piece of paper
22:27 KennethWilke hey! don't tell the people who do the hiring that!
22:28 KennethWilke lol
22:28 forrest hah
22:28 KennethWilke you will shatter their dreams
22:28 forrest no no, not their dreams, just their keyword searches
22:28 KennethWilke lol
22:28 KennethWilke true dat
22:28 forrest mgw, are you adding your use case to that issue
22:28 forrest *?
22:29 KennethWilke heading out! see you guys tomorrow
22:29 mgw forrest: I will
22:29 forrest later
22:29 forrest mgw, ok cool
22:30 Jahkeup wait a sec, are extmods loaded under a different path? that is should I be importing as something like salt.exmods.ovs?
22:31 forrest https://groups.google.com/forum/#!msg/salt-users/7N5eXgLLLZ4/n0zk0-1BpTIJ
22:32 forrest see if that explains it at all Jahkeup
22:33 kermit joined #salt
22:33 Jahkeup forrest: my module gets loaded and shows up in sys.doc but calling salt with -l debug shows that it can't import the module in the state
22:33 Jahkeup and I need some of the functions in the state from the module
22:33 forrest oh that's odd, I don't know enough about the functionality of custom modules to provide you with some good details
22:34 Jahkeup forrest: oh thanks anyhow :) been banging my head against the table trying to get it to work lol
22:37 aat joined #salt
22:39 forrest Yea sometimes that happens, means it is time to go home and look at it tomorrow,
22:39 forrest *.
22:39 Jahkeup forrest: yeah too bad im home already >.<
22:39 forrest hah
22:39 Jahkeup forrest: but yeah I probably should take a break ;)
22:40 jpeach left #salt
22:41 druonysus joined #salt
22:47 Gifflen joined #salt
22:50 chrisgilmerproj left #salt
22:52 SEJeff_work joined #salt
22:55 druonysus joined #salt
22:55 druonysus joined #salt
22:56 jeddi joined #salt
22:57 bogdanbu joined #salt
23:03 cewood joined #salt
23:05 napperjabber joined #salt
23:12 rustyrazorblade joined #salt
23:17 rrauenza Can I make a pillar dependent on another pillar?  e.g., root_path: /foo, install_path: {{pillar['root_path']}}/bar
23:22 rrauenza I think the answer is no :)
23:22 druonysus joined #salt
23:22 druonysus joined #salt
23:32 EugeneKay I want to say no as well
23:32 EugeneKay But I've not tested
23:35 berto- joined #salt
23:37 liuyq joined #salt
23:40 Jahkeup :C why does no one do custom _states, can't find anything for docs on the interwebz
23:41 rrauenza EugeneKay: it seemed to fail to compile the pillar sls because of an undefined variable, which makes sense in a way.  I suppose you could possibly do it if you layered the sls's
23:41 EugeneKay Maybe.
23:41 rrauenza i.e., defined it in a different sls
23:45 liuyq joined #salt
23:48 Psi-Jack joined #salt
23:49 Psi-Jack So, in my top.sls, I have 'roles:webserver':  like that to match systems with grains with webserver roles.. But how do I do the opposite and match systems that do NOT have a specifed roles:name in it? I'm wanting to setup roles:mailserver, and it will be salt-manage,d, but every other system will setup postfix to use the actual mailserver as a smarthost.
23:50 efixit joined #salt
23:50 Jahkeup Psi-Jack: try 'not' http://docs.saltstack.com/topics/targeting/compound.html?highlight=matcher
23:51 efixit joined #salt
23:51 Jahkeup Psi-Jack: ie 'not roles:webserver'
23:51 Psi-Jack hmmm..
23:51 Jahkeup Psi-Jack: sorry to ditch you i have to run!
23:53 druonysus joined #salt
23:53 Psi-Jack Yeah, that's not working, at least with salt -C 'not roles:webserver' test.ping
23:53 cxz joined #salt
23:53 Jahkeup Psi-Jack: I'll be back in ~1.5 hr if you wanna ping me then
23:57 StDiluted joined #salt
23:59 Psi-Jack Jahkeup: heh, seems I needed the G@roles:name   G@, specifically, since this switches from simple matching to compound matching.

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