Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-01-29

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

All times shown according to UTC.

Time Nick Message
00:05 thealphanerd joined #salt
00:06 thealphanerd left #salt
00:06 rojem joined #salt
00:08 nmistry joined #salt
00:12 cachedout joined #salt
00:19 timoguin joined #salt
00:20 KyleG1 joined #salt
00:23 bemehow joined #salt
00:24 bemehow_ joined #salt
00:26 bemehow joined #salt
00:30 oz_akan_ joined #salt
00:32 mgw joined #salt
00:35 drdoom joined #salt
00:36 drdoom joined #salt
00:38 snuffeluffegus joined #salt
00:41 mwmnj joined #salt
00:43 backjlack joined #salt
00:45 jeremyBass joined #salt
00:47 DanGarthwaite joined #salt
00:56 cocoy1 joined #salt
00:57 timoguin joined #salt
00:58 mgw joined #salt
01:00 Hipikat left #salt
01:04 alunduil joined #salt
01:09 teebes joined #salt
01:09 zzzirk joined #salt
01:14 fllr joined #salt
01:15 Erchn joined #salt
01:16 Erchn Hi all, anyone know if mine data expires?
01:18 Teknix joined #salt
01:18 ajennings_ joined #salt
01:19 sroegner joined #salt
01:22 Erchn joined #salt
01:22 gnu-man joined #salt
01:23 alunduil joined #salt
01:26 xzarth joined #salt
01:31 teebes joined #salt
01:49 fllr joined #salt
01:50 Kyle joined #salt
01:51 drawsmcgraw joined #salt
01:54 grim76 Issue:  Determined that services will not restart on debian 6 systems when using salt, but services will restart on debian 7 with no problems.  Anyone ran into this before?
01:55 Dale_ joined #salt
01:57 n8n joined #salt
02:07 Guest19572 joined #salt
02:11 fllr joined #salt
02:13 Ryan_Lane2 joined #salt
02:14 Ryan_Lane2 joined #salt
02:15 teebes joined #salt
02:19 brianhicks joined #salt
02:24 timoguin_ joined #salt
02:28 Mua joined #salt
02:29 fllr joined #salt
02:29 ChoHag joined #salt
02:30 Mua__ joined #salt
02:33 rojem joined #salt
02:34 rojem joined #salt
02:36 jimallman joined #salt
02:39 Mua joined #salt
02:45 n8n joined #salt
02:53 elfixit joined #salt
02:53 munderwo joined #salt
02:58 fragamus joined #salt
02:58 NotreDev joined #salt
02:59 puppet joined #salt
03:00 puppet joined #salt
03:01 Mua_ joined #salt
03:02 eculver joined #salt
03:02 eculver joined #salt
03:10 cellofellow joined #salt
03:10 cellofellow is there some setting in the /etc/salt/minion file that gets salt-call to not need --local?
03:12 cellofellow well, whatever it is it's in the example masterless minion conf in the salty-vagrant github.
03:16 dave_den fileclient: local
03:16 dave_den i believe
03:17 cellofellow yeah, that's in there, thanks.
03:18 * cellofellow wishes he could be at saltconf
03:18 cellofellow I'm just in Orem, not that far.
03:18 cellofellow Didn't make ngconf either.
03:19 dave_den the timing didn't work out for me to make it, either
03:19 jimallman joined #salt
03:19 dave_den bummed
03:20 oz_akan_ joined #salt
03:21 xmltok joined #salt
03:30 rojem joined #salt
03:32 dancat joined #salt
03:35 xmltok joined #salt
03:36 all_the_fisch joined #salt
03:46 Gordonz joined #salt
03:59 jeremyBass left #salt
04:09 fllr joined #salt
04:10 cachedout_home joined #salt
04:10 mgw joined #salt
04:19 rojem joined #salt
04:21 codekoala joined #salt
04:22 NotreDev joined #salt
04:22 oz_akan_ joined #salt
04:23 NotreDev_ joined #salt
04:23 oz_akan_ joined #salt
04:24 NotreDev_ joined #salt
04:26 decwrl joined #salt
04:27 decwrl quit
04:27 decwrl left #salt
04:29 decwrl joined #salt
04:36 Mua joined #salt
04:42 jankowiak joined #salt
04:43 Mua_ joined #salt
04:44 NotreDev joined #salt
05:07 Derek_ joined #salt
05:14 kermit joined #salt
05:19 forrest joined #salt
05:19 forrest left #salt
05:20 forrest joined #salt
05:22 Ryan_Lane2 joined #salt
05:23 forrest hey Ryan_Lane2, how was the flight?
05:23 Ryan_Lane2 long
05:24 forrest were you in san fran?
05:24 Ryan_Lane2 I had a layover, which sucked :)
05:24 forrest ahh ok
05:24 Ryan_Lane2 I waited a little long to book my flight
05:24 forrest gotcha
05:24 forrest I was gonna say, seems like there should have been a straight shot
05:26 mgw joined #salt
05:26 Ryan_Lane2 yeah, should have been really short
05:29 forrest hey mgw
05:30 forrest kiorky, you around?
05:30 CheKoLyN joined #salt
05:32 mgw forrest: yeah, what's up?
05:33 forrest Nothing, just saying hey
05:34 mgw hey :-)
05:34 forrest Did you retreat up to your room? Or are you still in that big open area
05:35 mgw In my room.
05:35 forrest Nice
05:36 anuvrat joined #salt
05:37 forrest timoguin_, did you see my update on 9608?
05:41 ravibhure joined #salt
05:41 balltongu_ joined #salt
05:42 MTecknology I wonder if anyone has deployed gitlab via salt
05:42 ndrei joined #salt
05:43 baniir joined #salt
05:43 forrest MTecknology, I believe someone was working on that and I asked them to update this formula: https://github.com/saltstack-formulas/gitlab-formula
05:44 forrest unfortunately I don't have that history on this machine, you might try combing the logs, I believe they pasted me a gist
05:45 baniir trying to bootstrap salt with http://bootstrap.saltstack.org on freebsd 10, but configs are ending up in /etc not /usr/local/etc
05:45 SkyRocknRoll joined #salt
05:45 MTecknology forrest: https://github.com/viq/gitlab-formula
05:45 forrest yea that was it
05:46 MTecknology viq: no support for mysql or debian? no sanity?
05:46 MTecknology gosh
05:47 forrest it was a rough draft that he was just kinda showing off as something that was working
05:47 MTecknology Actually... I'd only want this done with mysql because we have a dedicated mysql box at work
05:47 forrest baniir, I don't know how the bootstrap handles freebsd10, I don't see a function for it.
05:48 forrest baniir, it looks like the bootstrap is definitely setting it to /usr/local/etc/salt though for the salt etc dir variabl
05:48 forrest e*
05:48 MTecknology forrest: ya, it looks like he was building it out to be pretty flexible, just not done yet
05:48 MTecknology he chose nginx, that makes up for it not being done yet!
05:48 forrest MTecknology, I don't know how much more flexible he was planning on going, I know that was configured for his specific config
05:48 forrest heh
05:49 forrest yea he said he'd try to push it back to the formula repo if he could make it sane enough
05:49 MTecknology there's a LOT more there than I was expecting to see, wow
05:50 baniir forrest: ah; how would i go about getting latest salt on freebsd 10
05:50 forrest baniir, the bootstrap seems like it should work
05:51 forrest the only check I see in place is one that sees if freebsd is less than 10, though I could be missing something.
05:53 yomilk joined #salt
05:54 baniir forrest: bootstrap seems to get most of the way there, but it can't find the config as it ends up in /etc not /usr/local/etc
05:54 forrest baniir, weird, I'm not really sure then because the functions should be setting that variable. You're running it as the root user right?
05:55 baniir forrest: with sudo
05:55 billygoat joined #salt
05:56 forrest hmm that should be ok. I'd suggest to try and see if you can do a bit more debugging regarding where it is failing, and then if it looks like the bootstrap is causing it file an issue on the bootstrap repo. You might also try tomorrow during the day, might be a few more freebsd guys around
05:56 joehillen joined #salt
05:56 forrest I've gotta head out for the evening, let me know if you figure it out
05:56 Gordonz joined #salt
05:57 Ryan_Lane2 joined #salt
05:58 baniir thanks forrest
06:01 Ryan_Lane2 joined #salt
06:03 Shish joined #salt
06:03 alunduil joined #salt
06:03 pydanny joined #salt
06:06 cnf joined #salt
06:13 NotreDev joined #salt
06:16 radix88 joined #salt
06:18 NotreDev joined #salt
06:24 ktenney_away joined #salt
06:27 xmltok joined #salt
06:28 kalloc joined #salt
06:33 matanya joined #salt
06:34 puppet joined #salt
06:34 middleman_ joined #salt
06:38 njpatel joined #salt
06:48 djinni` joined #salt
06:52 Psyke joined #salt
06:58 fxhp joined #salt
07:03 alunduil joined #salt
07:20 jasonrm joined #salt
07:23 linjan joined #salt
07:26 fllr joined #salt
07:26 ndrei joined #salt
07:27 sroegner joined #salt
07:31 ndrei joined #salt
07:33 NotreDev joined #salt
07:33 gammalget joined #salt
07:36 ckao joined #salt
07:38 twinshadow joined #salt
07:39 cyrusdavid joined #salt
07:43 SkyRocknRoll_ joined #salt
07:44 twiedenbein joined #salt
07:45 twiedenbein joined #salt
07:47 SkyRocknRoll joined #salt
07:48 twiedenbein joined #salt
07:51 dwfreed joined #salt
07:54 BenCoinanke joined #salt
07:57 Gordonz joined #salt
07:59 balboah joined #salt
08:03 alunduil joined #salt
08:04 mephx joined #salt
08:07 lessismore joined #salt
08:07 Shenril joined #salt
08:07 leonardinius joined #salt
08:09 zooz joined #salt
08:09 dwfreed joined #salt
08:14 jtang1 joined #salt
08:18 NotreDev joined #salt
08:18 ml_1 joined #salt
08:20 giantlock_ joined #salt
08:24 gurubert1 joined #salt
08:24 oz_akan_ joined #salt
08:26 fllr joined #salt
08:29 harobed joined #salt
08:32 helminthe joined #salt
08:36 ndrei joined #salt
08:42 synexic joined #salt
08:42 miguitas joined #salt
08:45 alunduil joined #salt
08:48 slav0nic joined #salt
08:51 lessismore joined #salt
08:53 yomilk joined #salt
08:59 yomilk joined #salt
09:05 alunduil joined #salt
09:10 monokrome joined #salt
09:10 msciciel_ joined #salt
09:12 travisfischer joined #salt
09:14 sgflt joined #salt
09:15 sroegner joined #salt
09:20 lessismore joined #salt
09:25 oz_akan_ joined #salt
09:26 fllr joined #salt
09:27 travisfischer joined #salt
09:33 dave_den joined #salt
09:36 travisfischer joined #salt
09:37 cewood joined #salt
09:40 pdayton joined #salt
09:40 q1x joined #salt
09:48 dave_den joined #salt
09:58 Gordonz joined #salt
10:03 alunduil joined #salt
10:03 cast joined #salt
10:14 linjan joined #salt
10:15 che-arne joined #salt
10:26 oz_akan_ joined #salt
10:26 fllr joined #salt
10:27 babilen What specs are needed for a salt master? How many cores, RAM, .. are needed to manage a certain number of minions?
10:29 cast well, i used a c3.xlarge ec2 instance playing around with 100 micro instance minions with some simple things like installing packages, pulling back data, that went quite fast [ro things took ~1 second]
10:30 babilen I would expect a c3.xlarge to be quite, well, sufficient
10:31 cast c3.xlarge = 4 thread, 8GB RAM, so as big as the xlarge suggests
10:31 cast i wanted to scale up higher but was limited to 100 instances in that region
10:31 babilen I know, but would have expected the salt master to not really need anything that beefy. Can't base that on experience though (hence my question)
10:32 * cast nods
10:32 babilen Ah, you ran into problem with 100 minions with that master?
10:34 cast zero problems, as you say its over beefy [is the only experience point i have so far beyond the handful of nodes in our regular setup]
10:35 babilen Sure, merci
10:36 * cast remembers bitterly back to puppets slowness
10:36 babilen I'll go with a 1 Core (E5-2670), 2GB RAM now ... We can always add more if needed. Should that be alright?
10:36 n8n joined #salt
10:38 mortis any idea where theres more info on the openwrt salt-client? read somewhere it was in testing, but cant find it
10:43 krak3n` joined #salt
10:45 david_a joined #salt
10:45 yomilk joined #salt
10:50 dangra joined #salt
10:50 ajennings joined #salt
10:50 ajennings left #salt
10:51 krak3n` joined #salt
10:51 a__jennings joined #salt
10:52 DaveQB joined #salt
10:57 joehh grim76: I may have noticed that, but put it down to the individual service rather than a general problem - I could be wrong htough
11:03 alunduil joined #salt
11:04 sroegner joined #salt
11:12 ndrei joined #salt
11:12 jtang1 joined #salt
11:18 viq MTecknology: no, not yet. I want to make it support mysql, and also debian (and ubuntu, and freebsd, and openbsd), alas, not yet. Right now it turned out it's not working as well as I thought so I have to fix it first.
11:19 viq MTecknology: as for nginx, it's actually what they recommend ;)
11:26 fllr joined #salt
11:27 oz_akan_ joined #salt
11:28 oz_akan__ joined #salt
11:44 viq MTecknology: and to give some support to my claims, https://github.com/viq/gitlab-formula/issues ;)
11:48 bagel_nimbius joined #salt
11:48 anuvrat joined #salt
11:48 bagel_nimbius hi salt.  I think i asked yesterday but i wondered if anyone had a link to a decent comparisson of salt vs chef?
11:53 beardo__ joined #salt
11:58 jeddi bagel_nimbius: I think lots of people have blogged about this (and puppet), but are you looking for any specific area to be compared?
11:59 Gordonz joined #salt
11:59 bagel_nimbius im not certain.  chef deployment seems kind of half-baked outside of ubuntu and debian deployments and puppet lacks idempotency
11:59 cast in what way does puppet lack idempotency?
12:01 bagel_nimbius come to think of it, i guess it doesnt.
12:01 cast no more or less than salt, they both allow arbitrary execs
12:02 bagel_nimbius I'm going to start playing around in salt.  i guess thats the sanest thing to do.
12:02 cast a big thing for me is python vs ruby
12:03 ekristen joined #salt
12:04 cast [never used chef, but certainly in puppet you end up writing various amounts of ruby]
12:06 bagel_nimbius ill take python any day.  ruby seems like a handstand compared for me, I learned on C and perl.
12:07 bagel_nimbius left #salt
12:10 yomilk joined #salt
12:13 dangra1 joined #salt
12:14 jcsp joined #salt
12:15 giantlock_ joined #salt
12:16 viq In various places the question "Should I use chef, puppet, salt, ansible or cfengine?" is answered with "YES" ;)
12:18 Furao joined #salt
12:19 babilen What are sensible ways to develop state and pillar files and to make them available on the master? "Simply" configure GitFS and have that pointed to, say, saltstack-formulas and a local repository that contains the top.sls et al. files ? And then external pillar for /srv/pillar ?
12:20 diegows joined #salt
12:22 Furao babilen: there is git for pillar too
12:25 babilen Furao: Yes, but is my understanding correct in that GitFS essentially replaces /srv/salt ? Or does that allow me to simply reference states in git via salt:// in the state files that are still stored there? Would I have to checkout my own repository there?
12:26 fllr joined #salt
12:28 Furao no you don't have to checkout yourself
12:29 Furao it convert tag and branches as environments
12:29 oz_akan_ joined #salt
12:30 taion809 joined #salt
12:31 justBob joined #salt
12:31 asdf__ joined #salt
12:33 asdf__ i have facing a serious error in creating the package
12:33 cast go on
12:35 babilen Furao: Yeah, but would that allow me to create a top.sls file in the git repo only and then to have that available in lieu of a /srv/salt/top.sls ?
12:35 asdf__ joined #salt
12:35 babilen Well, I guess I TIAS
12:36 asdf__ i am unable to to do pkg installation using salt
12:36 sroegner joined #salt
12:36 babilen asdf__: What happens if you try?
12:36 oz_akan_ joined #salt
12:39 oz_akan_ joined #salt
12:40 ndrei joined #salt
12:45 asdf__ babilen: please refer to this link http://pastebin.com/R1emKeE2
13:00 ggoZ joined #salt
13:01 ggoZ ahoi
13:01 ggoZ does the 'master_tops' param has any effect in minion's config?
13:03 foxx joined #salt
13:03 diegows I think that I asked this before but I don't remember the answer :)
13:03 diegows is there a wat to disable timeout in state.highstate?
13:04 ggoZ I'm running `salt-call state.show_top` and it just returns empty yaml with root node as 'locally'.
13:04 diegows sometimes finished without waiting for the result of the operation
13:07 cachedout_home joined #salt
13:07 Furao babilen: you can mix gitfs and rootfs (regular)
13:08 mwmnj joined #salt
13:08 ccase joined #salt
13:09 Furao in my states I have a top.sls specific to handle roles and git branches
13:09 Furao this is the only files in /srv/salt/
13:09 gadams999 joined #salt
13:09 asdf__ yes furao
13:10 asdf__ sorry salt/edit$ ls vimrc  vim.sls
13:12 grim76 Debian 6 salt will not restart services when a file changes, but I can manually restart the service from salt.  Any ideas?
13:14 Furao grim76: use watch on the service
13:15 Furao (or watch_in)
13:19 cast left #salt
13:20 grim76 Furao  Tried that and it does not work.  The same information applied to a Debian 7 machine works every time.
13:21 Furao grim76: same service in both? both /etc/init.d and not upstart/systemd?
13:22 babilen Furao: Mind sharing?
13:22 grim76 Furao  Correct.  Service in question is snmpd and the file being updated is snmpd.conf.
13:22 grim76 Let me pastebin my information.
13:23 mwmnj joined #salt
13:23 Furao babilen: https://gist.github.com/bclermont/18d278439041726ee1b2
13:23 Furao good luck :P there is a bit more than gitfs & roles
13:26 fllr joined #salt
13:28 Sypher joined #salt
13:28 grim76 http://pastebin.com/zBHetj8d
13:30 Furao grim76: you should have consistent indentation
13:30 Furao sometimes it's 4, or 2 or 1
13:30 Furao this make yaml confused
13:30 babilen Furao: No worries, thanks! :)
13:30 grim76 Furao  ok...let me fix that then.
13:30 LowWalker joined #salt
13:31 Furao but for the rest, it's should be ok
13:35 grim76 Furao  No change
13:36 Dale_ joined #salt
13:36 Furao that might be a salt bug, this need some logging @ debug and maybe some python-fu find why it don't restart
13:37 grim76 Furao  Ok...  I do know that if I fire off a manual restart of the service from salt the service restarts.
13:39 brianhicks joined #salt
13:40 timoguin joined #salt
13:43 jslatts joined #salt
13:47 che-arne|2 joined #salt
13:53 tessellare joined #salt
13:57 grim76 Furao  ok I think I found it.  Debian 6 has snmptrapd as part of the start/stop script.  snmptrapd was failing since it had not been configured and that appears to have been throwing it off.
13:57 Furao grim76: so, in the logs salt was really trying to restart it?
13:58 leonardinius_ joined #salt
13:59 grim76 Furao  Actually it was trying to start it.  It saw the snmptrapd failing as the service was not started.  When in fact snmpd was started just not snmptrapd.
14:00 Gordonz joined #salt
14:05 ndrei joined #salt
14:07 jakubek joined #salt
14:07 wkf_ joined #salt
14:10 babilen Furao: Sorry, could you paste the corresponding pillar data? It's just that reading more complex setups helps me a lot to tie the information I have together :)
14:10 babilen (or just an example)
14:11 imaginarysteve joined #salt
14:11 MTecknology viq: I'll more than happily help you with it if I'm able!
14:16 ndrei joined #salt
14:17 viq MTecknology: right now I have two most pressing problems, one is being that RVM doesn't play nice and seems that commands I found (not in the repo yet) work, but still pushing git to the machine doesn't find ruby. Two is there is some issue with creating user in postgresql, it may be something with timing when it gets installed as on second run it works. Once I have those working CentOS will be pretty much ready and I can move to try and support others.
14:18 juicer2 joined #salt
14:23 viq MTecknology: though next two weeks I'll be traveling, so no idea how much I'll work on this in that time.
14:24 MTecknology I wonder how much difference there will be for centos vs debian
14:25 viq I guess mostly the packages installed
14:25 viq Names of packages, names of services
14:26 NotreDev joined #salt
14:26 fllr joined #salt
14:29 _fllr_ joined #salt
14:29 crane hi, i'm trying to install mysql-server with a password for root but it does not work as excepted... http://dpaste.com/1575266/ does someone see my misstake?
14:30 blee joined #salt
14:31 fllr_ joined #salt
14:31 viq crane: do you have MySQLdb python module installed?
14:31 q1x joined #salt
14:32 crane viq, i noticed it just at this moment :D
14:32 crane viq, thx for the hint :)
14:32 viq :)
14:35 mpanetta joined #salt
14:36 mpanetta joined #salt
14:39 cachedout joined #salt
14:41 Whissi joined #salt
14:44 wblair joined #salt
14:47 wblair ssh_auth for non-root user creates .ssh dir with correct ownr but with group "root" and perms 600, not 700.  ssh_auth docs do not have an option to set group or permissions. My googles have not been helpful. Suggestions please.
14:48 wblair I'm very sad I can't be at SaltConf. Hopefully someone will see this soon.
14:48 rojem joined #salt
14:50 gmtech Is there an equivalent of puppet definitions in salt ?
14:50 gmtech I.e. can you drive one formula from another
14:50 dwiden joined #salt
14:50 dwiden hi
14:51 oz_akan_ joined #salt
14:51 viq gmtech: what do you mean, "drive" ?
14:51 gmtech So for example, I have an apache formula for creating all manner of apache things. Can I call/require/include it from "my app" to tell the apache formula to put in a VHOST entry on behalf of myself ?
14:51 gmtech So for example you get nice encapsulation
14:52 dwiden Is there a way to have a minion re-send their public key to the master?
14:52 jsm joined #salt
14:54 viq gmtech: haven't done much with this yet, but look up include and extend statements, it may be waht you want
14:56 quickdry21 joined #salt
14:57 williamthekid_ joined #salt
15:01 DanGarthwaite joined #salt
15:02 abe_music joined #salt
15:02 zzzirk joined #salt
15:04 zzzirk joined #salt
15:06 jeremyBass1 joined #salt
15:06 jrdx joined #salt
15:07 cachedout joined #salt
15:08 Katafalkas joined #salt
15:08 DanGarthwaite Morning!  Who here is waiting for the keynote to begin?
15:09 evan__ joined #salt
15:09 Katafalkas chdck !
15:09 Katafalkas check *
15:09 zzzirk Unfortunately I haven't left home yet. :(
15:10 EntropyWorks and it begins
15:10 Katafalkas Mark Chen is on the stage
15:10 mr_chris I'm debugging a state that's using the py renderer. I've stripped it down to an empty run function and a print statement like so http://paste.linux-help.org/view/26e1491b.
15:10 mr_chris It's being call via state.sls. The contents of the print statement gets printed twice. Why is it getting hit twice?
15:10 mr_chris Rather why does the run function get called twice?
15:11 mr_chris I've confirmed this problem by generating a random number, printing it out, and seeing that it is different on each print.
15:11 mr_chris This essentially makes the py renderer unusable to any significant amount of work knowing that it's going to be doing it twice.
15:15 forrest joined #salt
15:16 grim76 forrest  Do you remember trying to help me last night with the SNMPD issue?
15:17 DanGarthwaite mr_chris: Put the variable bits into a pillar, not a state.
15:17 forrest grim76, yes, did you figure it out?
15:17 DanGarthwaite grim76: Hello again.  I think everyone is even more busy for the next two hours than yesterday.  Their CEO is up on stage at saltconf.
15:17 forrest DanGarthwaite, I'm never too busy to pretend I'm here.
15:18 mr_chris DanGarthwaite: My state's purpose is to existing static data from a pillar.
15:18 DanGarthwaite I think you are actually a bot.  :)
15:18 mr_chris Doesn't it seem like a bug, though.
15:18 mr_chris ?
15:18 forrest DanGarthwaite, well if you were here you would know I am not :P
15:18 * DanGarthwaite is.
15:19 forrest oh did we meet?
15:19 forrest And I just forgot already?
15:19 grim76 Yes I did get it figured out.  Salt was working as expected.  Debian 6 even when /etc/default/snmpd has snmptrad turn off tries to start it with snmpd.  Since that was failing it treated the snmpd service as down and attempted to start it.  Even though it was started the failure of snmptrapd was causing it to think it was failed.
15:19 forrest grim76, excellent! I am really glad you figured it out.
15:20 DanGarthwaite Not yet.  I was in the remote exec class yesterday.
15:20 forrest ahh ok
15:20 forrest we'll meet up today at some point I'm sure
15:20 DanGarthwaite grim76:  Ha!  I nailed it!
15:20 slav0nic joined #salt
15:20 terminalmage mr_chris: have you opened an issue for this?
15:20 grim76 Thanks for all the help and effort DanGarthwaite and forrest
15:20 mr_chris terminalmage: I believe I shall.
15:20 forrest DanGarthwaite, don't like, you specifically sid the snmpd service
15:20 forrest yea np grim76, glad I couldn't dup it.
15:20 terminalmage mr_chris: ok, I was able to confirm this
15:20 forrest and that you figured it out :D
15:20 mr_chris terminalmage: OK.
15:21 evan__ Hi guys - i'm trying to upgrade my windows salt-minions to 17.4 from 16.3 (100 or so minions).  Does anyone know a good way to accomplish this?
15:21 grim76 Debian 7 behaves differently when I was testing it, it was working fine.
15:21 DanGarthwaite I mentioned that /etc/default/snmp might have a kill  switch set.
15:21 forrest LIES!
15:22 pydanny joined #salt
15:22 DanGarthwaite Isn't your boss talking right now?
15:22 forrest no
15:22 mr_chris I really wish there was an easier way to search existing bugs. Seems impossible with https://github.com/saltstack/salt/issues?labels=Bug&state=open
15:22 grim76 He did mention the kill switch, but not in the right context.
15:22 forrest mr_chris, use the search box at the top, it will search the issues
15:22 forrest even though it doesn't seem like it will
15:22 mr_chris forrest: Ah. Thanks.
15:22 forrest np, I missed that as well
15:23 oz_akan_ joined #salt
15:26 higgs001 joined #salt
15:27 forrest grim76, I'm just giving DanGarthwaite a hard time.
15:28 DanGarthwaite It was a hell of a puzzle.  Glad you solved it.
15:28 alunduil joined #salt
15:28 basepi forrest: how dare you chat on IRC during the keynote????
15:28 basepi ;)
15:29 forrest basepi, I can listen and type
15:29 basepi nuh uh!
15:30 forrest ok maybe not all the time.
15:30 munderwo joined #salt
15:32 cachedout joined #salt
15:32 mr_chris DanGarthwaite, forrest, terminalmage https://github.com/saltstack/salt/issues/10021
15:32 forrest mr_chris, ahh right I remember you talking about this last week.
15:32 mr_chris forrest: Nope. Just yesterday.
15:32 forrest sorry, last night
15:33 terminalmage mr_chris: thanks, I'll take a look at this during the sprint tonight.
15:33 mr_chris Weeks. Nights. Whatever.
15:33 forrest that split attention, basepi might actually be correct...
15:33 fllr joined #salt
15:33 mr_chris terminalmage: Much appreciated.
15:33 DanGarthwaite .. having just taken Tom's remote exec class... it doesn't surprise me all the states might be evaluated (pre-evaluated) byt he state compiler.
15:34 terminalmage DanGarthwaite: yeah, and so print statements get run
15:34 forrest jeez this DanGarthwaite guy, up on his high horse over there...
15:34 NotreDev joined #salt
15:34 DanGarthwaite Again, you really should put that !py file in a pillar.  I did similar for openvpn certificates: http://garthwaite.org/virtually-secure-with-openvpn-pillars-and-salt.html
15:34 DanGarthwaite I read half the doc last night, looking to beat 95% today.  :)
15:34 forrest that would be slick
15:35 grim76 ls
15:35 grim76 ooops
15:35 basepi :wq
15:35 basepi ;)
15:35 DanGarthwaite vim FTW
15:35 mwmnj joined #salt
15:36 kaptk2 joined #salt
15:36 basepi DanGarthwaite: definitely.  i tried to make the joke yesterday that salt would refuse to install emacs, the joke fell kinda flat
15:37 DanGarthwaite Tom in class:  Lots of jokes about building skynet, hushed politically-correct hedging about preferred text editor.
15:37 basepi hahaha
15:41 freebsd-yay joined #salt
15:42 baniir joined #salt
15:43 freebsd-yay I'm new to salt, but configuration mgmt is nothig new to me. -- Where do I give the salt master a list of all minions ? I'm doing the integration with our inventory system and want to export the hosts to salt.
15:44 forrest freebsd-yay, are you using a master/minion setup or salt-ssh?
15:44 freebsd-yay master/minion
15:45 forrest ok, so if you want to join those all, I don't know if there is a better way to do it, but how I'd do it is just scp a standard minion config file to all the minions which points to your master, then restart the minion server. Then you can accept all the keys on the master.
15:45 n0c joined #salt
15:45 freebsd-yay I'm fearing this is all implicit via the pki, but I want to see that an existing node has never contacted the server.
15:45 forrest minion service
15:45 freebsd-yay Ok, just as I assumed.
15:45 forrest you'll still have to check the keys and accept them
15:46 n0c so say I have a bunch of SLS files, all which need to use git.latest
15:46 n0c so I want to abstract installing git into another SLS
15:46 DanGarthwaite You can get a list of current and missing but expected minions with:   salt-run manage.status
15:46 cnelsonsic joined #salt
15:46 n0c how do I properly.. depend on the git sls having completed?
15:46 forrest freebsd-yay, someone might have a better solution for that initial add, but I haven't seen it.
15:46 n0c I've tried putting it in its own SLS, and including it, and various incantaions of 'require', but I keep getting...
15:47 n0c State git.latest found in sls
15:47 n0c is unavailable
15:47 Katafalkas n0c: do you have git installed on minions ?
15:47 forrest Katafalkas ++
15:47 vejdmn joined #salt
15:48 n0c not presently, I want that to be installed when I try to install a state that depends on it
15:48 hunter joined #salt
15:48 n0c automagically
15:48 forrest n0c, can you pastebin/gist your state?
15:48 forrest seems to me like you need to just install the git package first, then run your git.latest stuff, since git.latest won't work without the package being installed.
15:49 freebsd-yay As long as I can create the keys on the master it should be fine. Just to be shure, If I add a key for a minion and that minion was never in contact with the server, the master will still show the status of that minion?
15:49 n0c so I can't, in a single sls, say "before trying git.latest, make sure git is installed?"
15:49 forrest n0c, you totally can
15:49 Nazzy *gets a rogue rouge highlight*
15:49 forrest that's why I was asking you show us your state.
15:49 n0c forrest: alright, gimme a sec
15:50 forrest freebsd-yay, a key pair will be generated when it contacts the salt master.
15:50 Nazzy basepi, it definitely should refuse to install emacs ... further it should glare sullenly if used on a system that already has emacs installed
15:50 Katafalkas n0c: essentially you just need to add the require statement
15:51 forrest freebsd-yay, I may be confused on what you are asking, until you accept the key you can't really 'do' anything with the minions
15:52 toofer joined #salt
15:52 micah_chatt joined #salt
15:52 forrest freebsd-yay, when you point the minion at the master, and then restart the minion service using salt-key -L it will show all the minions waiting to be accepted on the master.
15:52 freebsd-yay forrest: I will create the keys on the master, accept them and then provision them to the minion myself. I just want to know if the master knows the minion merely by having accepted the key and *without* the minion contacting it.
15:55 hunter joined #salt
15:55 forrest freebsd-yay, I'm not sure, I've never tried to 'accept' a key for a minion which wasn't contacting the master.
15:55 mgw joined #salt
15:55 Katafalkas I believe salt-cloud is doing something that way
15:56 Katafalkas it creates keys and accepts keys before server is up ... no idea how it works in detail
15:56 n0c okay!
15:56 n0c https://bitbucket.org/scjudd/salt-test/src
15:56 n0c so I just threw all that together
15:57 n0c but that is the gist of it
15:57 baniir joined #salt
15:57 n0c I want to be able to say 'install this vhost', and have it pull code from a git repo and kick nginx
15:57 Katafalkas n0c: you need to add - require: - pkg: 'git'
15:58 Katafalkas on your git.latest
15:58 Nazzy hurmmmmmm... what's the difference between modules/file.py:stats and modules/file.py:lstat that necessitated the addition of the latter *confused*
15:58 Katafalkas n0c: http://pastebin.com/FePEtAEc
15:59 n0c alright, lemme give that a go
15:59 n0c I don't need the includes?
15:59 forrest n0c you still need the include
15:59 forrest technically the include should be running those in order...
16:00 forrest but try the require
16:00 baniir has anyone here set up salt 2014.1 on freebsd 10
16:00 hunter joined #salt
16:01 Gordonz joined #salt
16:02 Psyke joined #salt
16:02 n0c Katafalkas: nope, that's not working
16:02 n0c I think it's choking on git.latest before it acts on the require
16:02 Katafalkas n0c: how do you run states ?
16:03 Katafalkas did you do the include ?
16:03 n0c yep, include is still there
16:03 n0c right now I'm doing salt-call --local state.sls vhosts.foo_example_com test=True
16:03 n0c same story with state.highstate
16:03 freebsd-yay baniir: I have 0.17.4 on freebsd 10
16:04 n0c I think I know the problem >.>
16:04 * n0c tests
16:04 toofer joined #salt
16:05 guest joined #salt
16:05 Mua joined #salt
16:06 Katafalkas ok
16:06 Katafalkas n0c: you use 'git-core' instead 'git'. change the require: - pkg: 'git-core'
16:06 n0c yeah, already made that substitution
16:07 imaginarysteve joined #salt
16:07 n0c well, color me stupid
16:07 n0c I set test: True in my minion config to make this process faster
16:07 analogbyte if the pillar fails to render, is there any way to ignore it?
16:07 baniir freebsd-yay: from ports?
16:07 n0c git never gets installed, so git.latest can never work
16:08 freebsd-yay baniir: from the pkg repo
16:08 n0c if I pass test=False, it works
16:08 Katafalkas n0c: jave you tried to run state.show_sls insteas state.sls ?
16:08 freebsd-yay baniir: it just works
16:08 Katafalkas instead *
16:08 davet joined #salt
16:08 Katafalkas shows if it renders and orders in which the states will run
16:08 n0c so is the include: foo and require: foo the best syntax for this sort of thing?
16:09 n0c Katafalkas: nope, never heard of that command before, this is like.. day 2 of salt for me :)
16:09 baniir freebsd-yay: i'd like to use the ports salt state, available in 2014.1 but not earlier
16:09 n0c which has otherwise been very enjoyable
16:09 freebsd-yay baniir: ah ok, I only have used the pkgng module
16:10 DanGarthwaite joined #salt
16:10 ravibhure joined #salt
16:10 leonardinius_ Hi all, could you please help me out, I don't get the difference between require and prereq.
16:10 baniir freebsd-yay: with that, do you need to modify anything else in addition to providers in the minion conf
16:10 freebsd-yay banirr: no
16:10 leonardinius_ Isn't prereq the same as require_in?
16:11 leonardinius_ what's the difference?
16:11 freebsd-yay leonardinius_: prereq is fired when the referenced state is about to be run
16:12 freebsd-yay leonardinius_: if you use require_in it will only be fired once and then the "requirement" will be fulfilled
16:12 DanGarthwaite prereq call the _other_ state with test=True.  If there is a change in that _other_ state is anticipated, it executes.
16:13 leonardinius_ freebsd-yay, thanx. so it will be only executed if prereq condition is being executed
16:13 leonardinius_ DanGarthwaite, thanx. will try to wrap my head around it.
16:14 freebsd-yay I used prereq for a preinstall script for when a config file changes
16:14 bemehow joined #salt
16:18 Katafalkas n0c:  results ?
16:18 n0c working as expected. thanks for the help! :)
16:19 Katafalkas awesome
16:19 oz_akan_ joined #salt
16:20 oz_akan__ joined #salt
16:21 seanz joined #salt
16:25 freebsd-yay forrest: BTW it works. I set up a master and moved a .pub to /usr/local/etc/salt/pki/master/minions and salt-run manage.status reported it as down
16:25 forrest oh great
16:25 freebsd-yay It is kind of weird that the nodelist is implicitly given by the pki, but meh
16:25 forrest that is good to know, thanks freebsd-yay
16:26 zfouts joined #salt
16:26 ndrei joined #salt
16:28 Nexpro joined #salt
16:31 egads I'm seeing a WARNING in the minion logs saying "Master pub message signing is disabled but we received a signature for this message.  Most likely this means that your masters and minions are not the same version."
16:31 egads but both master and minion run the same version, 0.17.5-1precise1
16:31 ajw0100 joined #salt
16:31 egads any ideas what else could be wrong?
16:31 forrest egads, did you restart the services?
16:31 egads forrest: I did, on both ends
16:31 egads this comes up when starting up the minion
16:32 forrest egads, you might want to try restart the service with -l debug
16:33 elfixit joined #salt
16:33 toofer joined #salt
16:35 egads forrest: nothing in the logs other than that warning stand out to me
16:35 egads here is the log from the minion: http://hastebin.com/piwoboworu.txt
16:35 egads the warning appears when I run 'sudo salt '*' test.version' on the master, not on boot as I first said
16:37 egads perhaps it's nothing much to worry about but it seems odd as they're both running the same version
16:37 forrest yea I haven't messed with 0.17.5 yet
16:37 zfouts OoooOooo 0.17.5
16:37 zfouts What better time to roll it out than on a new deploy that I will be doing!
16:38 egads my thinking exactly, zfouts :)
16:38 forrest I've gotta go, let me know if you figure it out.
16:38 egads cheers forrest
16:38 zfouts egads: yea, no automation here but a few bash scripts. My last deployment was ~200 minions
16:38 zfouts egads: shall be fun!
16:39 egads my, that's a lot of minions!
16:40 freebsd-yay Thank you all!! It works now.
16:40 alunduil joined #salt
16:41 freebsd-yay Oh, is there a way to explicitly only enable a list of modules? disable_modules is not very practical
16:41 zfouts egads: you could say that again, I had every bug in 0.17.1 ;-)
16:42 baoboa joined #salt
16:42 Nazzy yay, my test case error report is down to 9 failures and 4 errors ¬_¬
16:45 kolaman joined #salt
16:45 DanGarthwaite joined #salt
16:45 jalbretsen joined #salt
16:46 sroegner joined #salt
16:47 jsaxon joined #salt
16:47 zzzirk joined #salt
16:49 codekoala congrats to anyone here who just received their SSCE!
16:50 toofer joined #salt
16:51 Nazzy codekoala, S-S-suCcEsfully? *ducks*
16:51 codekoala lol
16:52 codekoala yes, precisely
16:53 CheKoLyN joined #salt
16:53 codekoala (SaltStack Certified Engineer, for anyone who isn't familiar with SSCE)
16:54 rojem joined #salt
16:54 jimallman joined #salt
16:54 jsaxon joined #salt
16:55 rlarkin joined #salt
16:55 TheRealBill joined #salt
16:56 jsaxon joined #salt
16:58 seanz left #salt
16:58 jeremyfelt joined #salt
17:00 wkf_ joined #salt
17:01 Katafalkas joined #salt
17:02 AdamSewell joined #salt
17:02 Ahlee well, I think I passed
17:02 Ahlee 85%
17:03 Ahlee it's harder than expected.
17:03 mgw joined #salt
17:03 fllr joined #salt
17:04 codekoala yeah, I bet
17:04 DanGarthwaite I think the bar is %80.
17:04 DanGarthwaite er, 80%.
17:04 hunter joined #salt
17:04 codekoala I figured I might as well try it
17:05 codekoala got nothing to lose
17:05 hunter joined #salt
17:05 DanGarthwaite I think I'll take it before lunch today.  There isn't many free hours at the conf.  :)
17:05 troyready joined #salt
17:05 codekoala very true
17:07 jeremyBass1 question, can you trun `salt '*' cp.get_file salt://path/to/file /minion/dest` in to `/minion/dest:  cp.get_file: -path: salt://path/to/file` or something like that.. I can't find the top file doc file for cp, but would like to use it to replace cmd.run that i use now.. any idea?
17:07 benturner joined #salt
17:08 pdayton joined #salt
17:09 dave_den jeremyBass1: yes, but there is no 'cp' state module. you have to use the 'module' state module to execute the cp execution module: http://docs.saltstack.com/ref/states/all/salt.states.module.html
17:09 Ryan_Lane2 joined #salt
17:09 Ryan_Lane2 joined #salt
17:11 mikkn Hmm... I'm not getting the full path for the drydock lease as a cwd when running commands. Am I doing something wrong?
17:11 Corey And Forrest is starting his talk on formulas and states.
17:11 mikkn (in Harbormaster during a build)
17:11 mikkn Oh, uh. :)
17:11 mikkn Nevermind me
17:11 Corey mikkn: Uh... I have no idea what those things are. :-)
17:11 teskew joined #salt
17:12 jeremyBass1 @dave_den so something like `move-file:   module.run:     - func: cp.get_file - name: /minion/dest  - path: salt://path/to/file` is what you mean?
17:12 Corey Which means that Salt engineers have added a LOT today at Saltconf, or you're asking about something else. :-)
17:12 dave_den jeremyBass1: no, 'func' is an argument of mine.send.
17:14 dave_den jeremyBass1: get_a_file:\n  module.run:\n    - name: cp.get_file\n    - path: salt://path/to/file\n   - dest: /minion/dest
17:14 jeremyBass1 super.. thank you.. i was just typing that..
17:14 jeremyBass1 thank you for clearing that up
17:14 tvle joined #salt
17:16 blee_ joined #salt
17:16 cachedout joined #salt
17:17 crane i asked yesterday allready but did not received an answer (or i did not noticed) can someone tell me how to add a second user to my mysql database using salt? http://dpaste.com/1575470/ is not working. i guess cause i used two times the mysql_user section
17:17 cachedout joined #salt
17:18 Ahlee crane: pretty sure those have to be unique.  You'll also want to embed that in a pillar and make a call to look those up so you're not shipping the username/password to all minions
17:18 dave_den crane: you cannot use more than one of the same module under the same id declaration
17:18 dave_den move you second mysql_user.present to another id declaration
17:18 Ahlee yeah, that
17:19 krak3n` joined #salt
17:19 crane dave_den, thx for the advice. will do this :)
17:19 Nazzy *sigh* I think I'm gonna sod this off and submit this PR with possibly failing (unrelated) tests and /then/ attempt to fix some of the test failures
17:20 KyleG joined #salt
17:20 KyleG joined #salt
17:20 crane and now the be a little bit more complex... i would like to download a zip compressed sql file and put it into my database. i gues i have to write a sh script for downloading and unziping it?
17:22 Ahlee crane: That's how I would handle it, or write a _module to do it in python
17:23 cachedout joined #salt
17:23 dave_den crane: or you could just do a cmd.run that does the wget & unzip.
17:24 rgbkrk joined #salt
17:24 carmony so is this where we have the saltconf back channel to heckle people? :P
17:24 carmony j/k
17:24 codekoala that's why I'm here
17:25 xmltok joined #salt
17:25 gmtech Need best advice on something. Say I have 2 products, product "foo" and product "bar", each with a complicated setup. I create 2 formulas - "foo" and "bar" for each one, but sometimes they do very similar things. E.g. they both require an apache VHOST to be written, but the VHOSTs are always with identical structure. So rather than putting an entire VHOST file in both foo, and bar, full of duplications, is there a way to template it so that the template only
17:25 gmtech exists once, but both foo and bar write subtly different versions of it ?
17:25 iShming_ joined #salt
17:27 ndrei joined #salt
17:28 dave_den gmtech: http://docs.saltstack.com/topics/tutorials/starting_states.html#extending-included-sls-data
17:29 cachedout joined #salt
17:31 backjlack joined #salt
17:32 gmtech I don't think that helps @dave_den
17:32 dave_den what are the difference in the states?
17:33 gmtech the differences aren't really in the states, the differences are in a template.
17:33 Ryan_Lane2 joined #salt
17:33 gmtech sorry - the only difference is in some pillar data
17:33 Nazzy gmtech, sounds like you need templating *nod*
17:33 ConceitedCode left #salt
17:34 Nazzy let me grab you a link
17:34 Nazzy gmtech, http://docs.saltstack.com/ref/states/all/salt.states.file.html
17:34 Ryan_Lane2 joined #salt
17:34 yota_ hi. Where I put my unit test files of my own modules?
17:34 yota_ _modules/test ?
17:36 cachedout yota_, right now there is no inclusion of module unit tests for custom module in the salt test runner. However, that's a really good idea and something that shouldn't be too hard for us to add.
17:36 joehillen joined #salt
17:36 dave_den gmtech: are you using jinja for the templating?
17:38 DanGarthwaite joined #salt
17:38 Nazzy gmtech, specifically you'll want to use file.managed, templating defaults to jinja, then you can do {{ pillar['somekeyorother'] }} to inject the variable
17:39 Katafalkas joined #salt
17:40 Nazzy gmtech, there's a pretty good summary at http://docs.saltstack.com/topics/tutorials/pillar.html#paramaterizing-states-with-pillar
17:40 yota_ cachedout: oki, thanks
17:42 gmtech yes dave_den
17:42 Nazzy hmmm ... gmtech, I may be wrong on that templating defaulting to jinja lol, but the rest stands and should do what you want
17:43 dave_den gmtech: are you aware of 'context' in file.managed?
17:43 dave_den if i understand correctly, you are confused about how to write the vhost template and knowing which variables to use in the case of it being rendered from the 'foo' or 'bar' state/
17:43 Nazzy dave_den, if the difference is already in pillar data he may not need it, but that is a good note, yes
17:44 Katafalkas joined #salt
17:44 Ryan_Lane2 joined #salt
17:44 dave_den using context, you specify the data to pass to the template from each foo or bar state. that way the template is written once but passed different data from either foo or bar
17:45 Ryan_Lane joined #salt
17:45 zfouts_ joined #salt
17:46 gmtech so.. the template could happily exist in foo, and bar could render it in its own way, injecting new data as appropriate?
17:46 mgw joined #salt
17:47 dave_den well, i wouldn't say it exists in foo, but it would be accessible to both foo and bar states
17:48 dave_den yes
17:48 Nazzy gmtech, yeah, you can have a shared state if they're similar enough, use the pillar data to vary, or if there's too much difference you can have two states with the same source
17:49 Nazzy gmtech, if the two are functionally identical except for the vhost name in the template you can have a single state and have both foo and bar use it
17:50 Nazzy I do it with postfix actually ... I have mail servers configured to use a postfix state and then use a pillar value to set the server's name properly in the template
17:50 gmtech You know what salt needs? :-)   The equivalent of this : http://docs.puppetlabs.com/puppet/3/reference/lang_defined_types.html
17:50 gmtech Thanks for your help guys ^^
17:51 Nazzy there's kind of something like that ... you can extend existing states and override parts of them
17:51 gmtech I'll certainly look into your suggestions guys/girls, thanks a lot
17:51 codysoyland could somebody tell me how to generate a hash file compatible with the source_hash parameter documented here: http://docs.saltstack.com/ref/states/all/salt.states.file.html#salt.states.file.managed
17:51 Nazzy it's more inheritance than functional, but it's nice ... you can also write states straight in python or pydsl to write libraries :)
17:51 Nazzy and you're welcome :)
17:54 dave_den codysoyland: the 'examples' code block under source_hash has examples.
17:54 Nazzy ya, it should just take the output of something like md5sum or sha256
17:55 Nazzy dave_den, regretfully, that example confused me for a minute, it's not very clear :(
17:56 codysoyland dave_den: i tried a few things but got "[file] contains an invalid hash format, it must be in the format <hash type>=<hash>" -- none of those examples have the hash in that format
17:56 codysoyland i tried "md5sum * > hashes.txt" but that didn't work
17:56 codysoyland so i prepended each line with "md5="
17:57 anuvrat joined #salt
17:57 tempspace joined #salt
17:58 tempspace Hey all
17:58 tempspace Everybody enjoying SaltConf?
17:58 Nazzy dave_den, the second line in the example appears totally spurious, that hash isn't the right length for any, but looks probably mangled sha256
17:58 tempspace (for those here)
17:58 dave_den codysoyland: the format on  the salt documentation html page is different than what is used in 0.17
17:58 Nazzy tempspace, not there, sadly
17:59 dave_den codysoyland: http://salt.readthedocs.org/en/v0.17.4/ref/states/all/salt.states.file.html?highlight=source_hash
17:59 codysoyland dave_den: oh i see. i'm on 0.17. thanks, i'll look at that
18:00 [diecast] joined #salt
18:01 higgs001 joined #salt
18:01 Gordonz joined #salt
18:05 Ryan_Lane joined #salt
18:05 Nazzy rats ... I forgot to fork salttesting *goes to do*
18:07 Katafalkas joined #salt
18:08 kolaman i have developed a formula to place a patch in tomcat directory and that restart tomcat service but getting this error msg while running that   "http://pastebin.com/WvmC04UH"
18:08 kolaman I am new to saltstack, can someone please see if I am doing something wrong in creating this formula
18:08 amckinley joined #salt
18:09 forrest joined #salt
18:10 jsm joined #salt
18:10 hunter joined #salt
18:12 gnu-man joined #salt
18:15 teskew kolaman: what service are you trying to restart? it's not named in your sls.
18:15 druonysus joined #salt
18:15 kolaman teskew: tomcat service
18:17 dave_den tomcat:\n  service.running:\n    - require:\n      - pkg: tomcat
18:18 dave_den the problem is that you probably don't have a tomcat state id declaration
18:18 dave_den for the pkg require
18:19 kaptk2 joined #salt
18:21 kermit joined #salt
18:21 q1x joined #salt
18:22 kolaman dave_den: :)
18:23 NotreDev joined #salt
18:26 micah_chatt joined #salt
18:29 Nazzy test suite is so sloooooow
18:30 jeremyBass1 ok I run `{{ web_root }}/staging:\n  module.run:\n    - name: cp.get_dir\n    - path: {{ stage_root }}\n    - dest: {{ web_root }}staging` and it says that it excutes fine, but i never see the folder get moved over.. this is a masterless run.. and as far as i can tell it should work based on http://docs.saltstack.com/ref/modules/all/salt.modules.cp.html#salt.modules.cp.get_dir .. but the folder fails to show up
18:32 diegows joined #salt
18:33 tru_tru joined #salt
18:39 flymacpara joined #salt
18:40 flymacpara left #salt
18:40 kolaman dave_den: I have created a new version of the formula, not service is defined but it doesn't start or restart the service it leaves the service inwhcih state that is already http://pastebin.com/Wu0mNYtr
18:42 rojem joined #salt
18:44 kaptk2 joined #salt
18:45 robbyt joined #salt
18:48 cachedout joined #salt
18:52 toofer joined #salt
18:58 n8n joined #salt
18:59 jperras joined #salt
19:00 zfouts joined #salt
19:01 cachedout joined #salt
19:01 quickdry21 joined #salt
19:01 rgbkrk joined #salt
19:03 ndrei joined #salt
19:05 wkf_ joined #salt
19:06 cachedout joined #salt
19:08 sashka_ua guys, I need an idea. I need to manage file only in when it doesnt exist or doesnt contain some some string
19:09 sashka_ua I think I could make this by putting two cmd.run states with unless:, checking statements
19:09 pdayton joined #salt
19:11 zfouts_ joined #salt
19:12 jeremyBass1 does anyone know why `{{ web_root }}/staging:\n  module.run:\n    - name: cp.get_dir\n    - path: {{ stage_root }}\n    - dest: {{ web_root }}staging` wouldn't work?  keeping in mind that the vars used have the right paths as {%- set web_root = "/var/app/" + env + "/html/" %}  and {%- set stage_root = "salt://stage/vagrant/" %}
19:13 jeremyBass1 and when i go the cmd.run cp way of coping the folder it works wit that same values
19:14 hunter joined #salt
19:17 ml_1 joined #salt
19:18 jeremyBass1 it'd be nice if there was a salt.states.file.dir_copy or something.
19:21 zooz joined #salt
19:21 micah_chatt_ joined #salt
19:22 dave_den jeremyBass1: http://docs.saltstack.com/ref/states/all/salt.states.file.html#salt.states.file.recurse
19:23 dave_den any reason you're trying to use the cp module?
19:26 jeremyBass1 do do that copy, .. i kept looking over file.recurse ..that would be why
19:26 diegows joined #salt
19:28 bretep Hmm... If I'm writing a grain that requires a 3rdpart package, what it the best way for me to maintain that... I will not be publishing this grain publically... well I don't plan on it anyway. I guess it could be use. What does it gets data_bags from chef so we can getup and start using salt ASAP. Eventually the grain will go away... next 6 months or so.
19:29 Nazzy bretep, you need loader foo... one moment
19:30 bretep 3rdparty python library rather than package
19:32 hunter joined #salt
19:32 Nazzy hmmm, are you asking how to manage a custom grain file in production, or how to install a random package?  if the latter you can use pip
19:32 bretep Not the latter. :)
19:32 Nazzy kk
19:32 bretep How to manaage the dependancy for the _grain/chef.py
19:33 bretep Since highstate loads the grains beofre running states, I need the chefpy python packeage pre-installed on the box
19:33 jtang1 joined #salt
19:34 Nazzy I'm not sure there is a way to make it install a package from a grain like that, but you could make it only load when the package is present
19:34 Nazzy you'd end up with a config that installs the grain the first run and does the stuff it needs that grain for on subsequent runs... it's not pretty but it might to the trick
19:35 Nazzy there's a reload modules type instruction, but I'm not sure if it would work for what you want
19:36 dave_den aren't databags the equivalent of pillar in salt?
19:36 bretep Yes, but there are encrypted databags which are equivlent to... nothing in salt
19:36 dave_den any reason why they are going to be loaded as grains?
19:37 bretep A subset of our "pillars" are encrpted
19:37 Nazzy http://docs.saltstack.com/ref/states/index.html?highlight=reload_modules#reloading-modules
19:37 dave_den if you load them as an ext_pillar it would be easy to refresh them
19:37 bretep Got it
19:37 bretep Vs. grains
19:37 cachedout joined #salt
19:37 bretep I like that...
19:37 dave_den and you could do it all from the master
19:37 bretep Yeah
19:38 dave_den instead of having to install prereq packages on the minions
19:38 bretep dave_den Totally doing that.
19:38 bretep Thanks
19:38 bretep !
19:38 dave_den np
19:38 Nazzy *nod* much better way
19:38 bretep That makes life SO much easier
19:40 modafinil_ dumb question, how can i setup an 'exclusive watch' (only do thing B if thing A changed) (so, like prereq, except opposite)
19:41 dave_den modafinil_: just use 'watch'
19:42 quickdry21 joined #salt
19:43 modafinil_ i'm pretty sure the second state will always apply though (its a cmd.run, but there's no real unless cmd i could run, just watch to watch another state)
19:43 dave_den modafinil_: if you're using cmd, you want to use cmd.wait instead of cmd.run
19:43 modafinil_ oh, duh, i knew that
19:43 modafinil_ :slinks out:
19:43 modafinil_ thanks
19:43 dave_den yep
19:43 rlarkin I'm looking for an example of a grain used as a variable in a jinja template..I would like to do {%- set role=['role'] %} and then do variable: {{ pillar['lvlone'][role] }}
19:43 LowWalker joined #salt
19:44 dave_den rlarkin: the grains python dict is accasible as 'grains'
19:44 Nazzy rlarkin, I use pillar['fqdn'][grain['fqdn']] fairly often to map short hostnames to canonical ones
19:44 dave_den accessible, even
19:44 DanGarthwaite joined #salt
19:44 Nazzy with plurals where appropriate lol
19:45 * DanGarthwaite is now Salt Stack Certified Engineer
19:46 rlarkin hrm, I'm convinced I just have bad bracketing and can't see it.
19:46 dave_den congrats!
19:46 hunter bretep: How are some of your pillars encrypted?
19:47 sroegner rlarkin: i'd usually set the lvlone pillar dict like this set lvl_one = salt['pillar.get']('lvlone', {})
19:47 sroegner and then do lvlone.get(role, '')
19:47 dave_den rlarkin: btw, in your example, you are setting 'role' as a list.
19:47 sroegner which makes it a little safer
19:47 rlarkin oh.
19:48 puppet joined #salt
19:48 bretep hunter: One sec.
19:49 rlarkin I've been stuck because doing {{ pillar['lvlone']['realvalue'] }} would work
19:50 Nazzy I filed two PRs... the really complex one passes jenkins, the simple one fails with a PEP8 violation ¬_¬
19:50 Nazzy I'll have to see if I can get work to sponsor me a trip to SaltConf 2015 lol (I can dream :p)
19:51 micah_chatt_ joined #salt
19:51 sroegner rlarkin: this will barf if the lvlone pillar doesn't exist
19:52 rlarkin ah! let me explain better. if I have a grain role, which equals 'admin' for one minion, -and- I have a nested dict , I want to get lvleone:admin:?
19:53 rlarkin I'm trying not to have several pages of 'if grain'
19:53 dave_den rlarkin: the 'grains' variable in templates is a regular python dict
19:53 dave_den if you want to use the salt grains.get module/function, you ned to use it from the 'salt' dict:
19:54 Ryan_Lane joined #salt
19:54 dave_den myrole = salt['grains.get']('lvleone:admin')
19:55 hunter The problem I have is that I do NOT want the roles to be in grains.
19:55 dave_den that uses the grains.get from http://docs.saltstack.com/ref/modules/all/salt.modules.grains.html#salt.modules.grains.get
19:55 hunter Maybe I'm tilting at a salt windmill.
19:55 hunter Do people really set custom grains to be "webserver" and "prod" or whatnot?
19:56 dave_den grains should be used for static data that are 'facts' about the minion, IMO
19:57 sroegner hunter: you want to run a central database of hosts that knows what goes where?
19:59 hunter sroegner: I think I do.
19:59 hunter I keep thinking that "roles" and "prod status" and stuff is more in keeping with pillar data.
20:00 hunter sroegner: I have on my todo wiring ext-pillar to etcd
20:00 hunter dave_den: I agree, but those facts have to not change (being a grain) and  - we change stuff.
20:00 sroegner i do use roles in grains and works great for me
20:00 hunter For those that DO use grains for roles and whatnot, how do you manage them?
20:01 hunter IE, are they changed by a special manual only state that pushes out /etc/salt/grains or whatever and that restarts the minoin?
20:01 imaginarysteve joined #salt
20:01 sroegner part of the bare-metal setup - salt-cloud sets all my grains in aws
20:01 hunter sroegner: and it works well for others - I'm trying to keep an open mind and be edumacated about it
20:02 sroegner understood
20:02 dave_den roles should be in pillar, imo
20:02 micah_chatt_ joined #salt
20:02 rlarkin I just find it convenient for my deploy to write /etc/salt/minion during deploy  a way to label it.  otherwise...?  I have to infer everything from the environment?  one image , many purposes
20:02 hunter sroegner: for those of us without AWS, what would the "bare-metal" equivalent be?
20:02 Gordonz joined #salt
20:03 dave_den the only reason to set things in grains is if you are gathering that data from the minion. the minion cannot alter the pillar data, but it can set grains
20:03 Ryan_Lane joined #salt
20:03 rlarkin Just that one grain so everything following deploy knows how to start
20:03 sroegner it also helps to keep the pillar data very generic
20:03 hunter dave_den: and THAT is why I think roles and prodstate should be in pillar - but again trying to be openminded
20:03 tempspace joined #salt
20:03 dave_den hunter: i agree.
20:03 Marion joined #salt
20:04 dave_den if the master makes the decision, it belongs in pillar
20:04 sroegner i don't understand how a role can be assigned to host in a pillar?
20:04 sroegner where is the mapping?
20:04 sroegner hostname?
20:04 dave_den sroegner: you can apply pillar values multiple ways
20:05 sroegner pls explain
20:05 bretep hunter: HMAC-SHA2-256
20:05 dave_den sroegner: http://docs.saltstack.com/topics/targeting/index.html
20:05 bretep I can decrpyt the file myself pretty easily.
20:05 rlarkin if all your servers start off identical but you want the to do several jobs, you have to tag them some way that salt can 'see'.
20:06 toofer joined #salt
20:06 StDiluted joined #salt
20:06 sroegner dave_den: sry - you lost me there
20:06 Katafalkas joined #salt
20:07 dave_den sroegner: you can target your pillar top.sls just as you do with your state top.sls.  most people do it by hostnames initially
20:08 sroegner ok
20:09 codekoala whose idea was it to have turkey sammiches for lunch?
20:09 sroegner just wondering - wouldn't that eventually mean that you end up maintaining targeting changes in both files?
20:09 codekoala I can't refuse them
20:09 codekoala and now I'm sleepy
20:10 dave_den sroegner: yes. pillar top.sls and state top.sls are separate and will contain different things
20:11 sroegner ok, thx
20:12 hunter dave_den, sroegner: Each of you should write a page on how this works FOR you.
20:12 jcockhren hunter++
20:12 jcockhren haha
20:12 hunter And that should get included at docs.saltstack
20:12 sroegner lol
20:12 hunter Cuz both are valid schemes - they'll work well for different people or different systems
20:12 sroegner indeed
20:12 hunter sroegner: so do you have special states to write them out to the minion?
20:12 sroegner no
20:13 sroegner my vagrant env or salt-cloud maintain the box' grains file according to the intended role
20:13 sroegner that way i am independent of host names
20:14 hunter how do you do that in vagrant (since I can replicate that here)?
20:14 sroegner whioch gets tricky for example when you try docker
20:14 tempspace What is everybody's thoughts on a module/state that does capistrano-style versioned deploys?
20:14 sroegner a shell provisioner copies files to /etc/salt
20:15 hunter ok.
20:15 sroegner hunter: https://github.com/accumulo/accumulo-saltstack
20:15 hunter And if you wanted to take a system from "qa" to "prod" after its been pounded on, how do you do that?
20:15 sroegner haven't really gottent around to doing that in a clean way
20:16 pdayton1 joined #salt
20:16 sroegner i test locally in a cluster of vagrant boxes, then push
20:16 sroegner aws boxes pick up the new version
20:16 hunter Well I see two possibilities - "promotion" is so rare and special you do it by hand (I can live with that frankly) or you never promote - once a system has been tested you replicate that "config" from qa to prod.
20:17 druonysus joined #salt
20:17 hunter The first would work here -the 2nd never would.
20:17 sroegner :-)
20:17 sroegner why not?
20:17 hunter I call them developers
20:18 sroegner those
20:18 hunter I get complaints from HR when I say what I really think.
20:18 kaptk2 joined #salt
20:18 hunter Oh look the children wired ... wait. Why is .... Oh hell no.
20:18 hunter Like that.
20:18 forrest joined #salt
20:19 ipmb joined #salt
20:19 hunter dave_den: for the pillar style, could you share a sample of how you do it?
20:21 whiskybar joined #salt
20:21 dave_den targeting? it's by hostname, and i have structured hostnames
20:21 hunter oh crap.
20:21 hunter Can't do that
20:21 dave_den rather i should say minion id, not hostname
20:21 hunter Oh?
20:21 hunter That might work.
20:22 hunter Our hostnames are .... unique.
20:22 tempspace target pillars based on grains?
20:22 dave_den the minion id is what the default targeting works on
20:22 jordan_ joined #salt
20:22 dave_den and by default the minion ids are the hostnames
20:23 hunter Thinking that through if the minion id's don't match or are at least very close to the hostnames, I'll kill myself.
20:23 hunter I'll need to invent something in pillar then.
20:23 dave_den it's easy to target by anything you want. compound matchers can pretty much do anything
20:23 DanGarthwaite joined #salt
20:24 zzzirk joined #salt
20:24 sroegner dave_den: do you use node groups?
20:24 jeremyBass1 is http://docs.saltstack.com/ref/states/all/salt.states.file.html#salt.states.file.patch the same as doing a `patch -p0 -i < filename.patch`
20:24 hunter I need to decided what my roles/$state/whatvers are going to be and just hand hammer the hostnames to have those pillars.
20:24 StDiluted Anyone using Docker and Salt together?
20:25 dave_den sroegner: no, i don't use nodegroups
20:25 dave_den nodegroups are just a static way of defining a matcher
20:25 sroegner dave_den: jsut wondering how you'd target a larger group of nodes
20:26 sroegner StDiluted: just started
20:26 dave_den sroegner: can you give me an example?  The easiest way to target every node is just the minion id wildcard, '*'.
20:27 fatbox joined #salt
20:27 dave_den http://docs.saltstack.com/topics/targeting/compound.html gives a few examples
20:28 sroegner in say a 20 node cluster 19 boxes have either the hadoop_master or hadoop_slave role - to target all of them i just do 'G@hadoop_*'
20:29 StDiluted sroegner, are you using the dockerio module?
20:29 kermit joined #salt
20:29 sroegner StDiluted: no
20:29 sroegner StDiluted: not yet
20:30 sroegner StDiluted: just started doing what i do elsewhere inside docker - am waiting for the next salt release
20:30 ahwitz joined #salt
20:30 dave_den sroegner: I would do:  P@role:hadoop_(master|slave)
20:30 sroegner right
20:31 dave_den yep, easy as that
20:31 sroegner but the moment you lose a box you have to change the pillar top file?
20:31 ahwitz hey - I have one minion set up right now that can access all of my sls files in /srv/salt/ right now, and I'm trying to set up a second, but every time I try to call one of the .sls files, I get "No matching sls found for (state) in env (env)" - any idea what the issue could be?
20:31 utahcon joined #salt
20:32 dave_den sroegner: no, not if you still have other minions with those roles....
20:32 ahwitz I just made sure that /etc/salt/minion config file is the same on both, but I'm hesitant to copy the rest of the /etc/salt/minion.d and pki files in case there's host-specific settings
20:32 sroegner dave_den: ok
20:32 StDiluted so is the docker stuff part of the current release?
20:33 sroegner dave_den: i still don't feel like maintaining a top file with node names for every cluster i run
20:34 dave_den sroegner: you don't need a single minion id/hostname in you top files, but you have to match somehow if you want to target your minions.
20:34 sroegner dave_den: a naming convention?
20:35 sroegner StDiluted: salt 0.17.5 doesn't come with salt-cloud
20:36 dave_den sroegner: if you want to use pillar with pillar_roots and yaml files, you have to have a pillar top.sls. how are you doing it currently?
20:37 sroegner running a generic ext_pillar gitfs project, including a top file w/o any targeting
20:37 sroegner very basic
20:38 sroegner https://github.com/accumulo/accumulo-saltstack-pillar
20:38 sroegner i needed to do this for salt-cloud to work
20:40 hunter dave_den: can you pastebin one of your P@roles is defined?
20:41 dave_den hunter: i don't use grains PCRE matching
20:42 hunter I know - what I need is a concrete example of how you define a pillar "role" so that hosts A, B and C get that role.
20:42 hunter Sorry - doing N things at the moment and probably not doing any of them well.
20:44 dlindquist joined #salt
20:44 cewood joined #salt
20:45 hunter Man I wish I was at the conference.
20:45 dimeshake if i want to target different lists of users to different hosts, how would I go about doing that? should I be targeting by host match in pillar top.sls? or users/init.sls somehow?
20:45 dimeshake how are other folks doing it. the examples i've seen are all basically putting all users on all hosts
20:46 cachedout joined #salt
20:46 dave_den hunter: this would assign the 'web_data' pillar data file to minion A, B, and C pillars: https://gist.github.com/dlanderson/aa33adaf58af0ad1abf5
20:47 DanGarthwaite I would do it in pillar.  How to organize it from there is hard to generalize.  I'd try to do it by roles.
20:47 hunter dave_den: I knew that. Sigh - thanks. I'm gong to stop multi-tasking now.
20:47 dimeshake I'd ideally be listing usernames that apply just under a hostname
20:48 dimeshake or perhaps by group
20:48 dimeshake 'admins', then list devs by name or somesuch
20:48 toofer speaking of roles, one of the presenters from SaltStack mentioned they even define their roles using pillar.
20:48 dave_den dimeshake: i would assign the users to minion pillars
20:49 pdayton joined #salt
20:49 druonysus joined #salt
20:49 druonysus joined #salt
20:49 dimeshake can you give a brief example to how that would work? dave_den
20:52 justBob joined #salt
20:53 dave_den dimeshake: https://gist.github.com/dlanderson/edf9c22e7a4363e89f48
20:54 fatbox left #salt
20:54 fatbox joined #salt
20:54 fllr Hey guys. How can I use a for loop to define machines using salt cloud? I tried using jinja, and mako sintaxes on my profiles file, and neither is working....
20:54 dimeshake gotcha. can those users.minion1 etc be includes? so users.minion1 might read:
20:54 dimeshake users: include: - users.alice
20:55 dimeshake not all on one line of course ;)
20:56 dave_den dimeshake: no, you will run into trouble with that. salt does not merge pillar keys
20:56 dimeshake hm.
20:57 dimeshake thanks. i'll see what i can do with this
20:58 Ryan_Lane dave_den: I wanted to talk to you guys about that
20:58 Ryan_Lane dave_den: lack of merging and hierarchy in pillar is a pain
20:58 dave_den ?
20:58 dave_den Ryan_Lane: yes indeed
20:59 Ryan_Lane hell, it would probably be better to just make a python version of hiera
20:59 toofer joined #salt
20:59 dave_den Ryan_Lane: you can use many ext_pillars available
20:59 Ryan_Lane yeah, but you can't do hierarchy or merging in regular pillars
20:59 dave_den one such ext_pillar that does file based yaml merging is reclass.
20:59 Ryan_Lane which forces you to use environments
21:00 Ryan_Lane it would be ideal to have a set of pillars that could be loaded hiearchially via grains
21:00 Ryan_Lane so that I can load development pillars for dev and production pillars for production
21:00 mgw joined #salt
21:00 Ryan_Lane without being forced into environments
21:01 dave_den reclass is pretty simple http://reclass.pantsfullofunix.net/
21:01 Ryan_Lane hm. interesting
21:01 jtang1 joined #salt
21:02 Katafalkas joined #salt
21:02 dave_den the built-in pillar wasn't meant to be super powerful
21:02 dave_den that's why there is ext_pillar
21:03 zzzirk joined #salt
21:03 * Ryan_Lane nods
21:05 teskew joined #salt
21:05 matanya joined #salt
21:06 imaginarysteve joined #salt
21:08 Katafalkas joined #salt
21:10 cachedout joined #salt
21:10 Linuturk joined #salt
21:11 rgbkrk joined #salt
21:13 toofer joined #salt
21:13 CaptTofu joined #salt
21:13 hunter joined #salt
21:14 dave_den anyone using user_namespace'd lxc containers?
21:15 mgw dave_den: what distro are you on?
21:15 dave_den mgw: ubuntu server daily build 14.04 (2014-01-28)
21:16 dave_den i can get the namespaced container to boot, but i want to use systemd in the container which means I need lxc.autodev and the container fails to start if that is set
21:17 dave_den it's all crickets in #lxcontainers
21:19 mgw dave_den: if you get it working, let me know and I'll add support to the lxc module to specify the user to run as
21:20 mesmer joined #salt
21:20 dave_den mgw: i'm actually doing namespaced containers under root
21:20 jrdx joined #salt
21:20 dave_den not unpriveleged user container start/stop
21:20 mgw I see, ok
21:20 dave_den but i will let you know what i find anyway
21:20 mgw thanks
21:22 n8n joined #salt
21:25 pdayton joined #salt
21:26 jeremyBass1 hello.. can someone help figure out why http://pastebin.com/cWy3yFu9 happen on a mac under a vagrant VM but works fine on a win pc?  .. the srv/salt is a symlink from the host
21:28 n0c so.. when do I need to use a '- ' in salt states?
21:29 dave_den jeremyBass1: check the mount exec options in vagrant
21:30 n0c because my understanding is that "- foo: foo\n- bar: bar" gets turned into "[{'foo': 'foo'}, {'bar': 'bar'}]"
21:30 dave_den actually, IIRC you need to allow symlinks in vagrant
21:30 n0c that is, why a list of single-item dicts instead of a dict?
21:30 dave_den jeremyBass1: for example https://coderwall.com/p/b5mu2w
21:30 jeremyBass1 @dave_den .. this is what i have .. config.vm.synced_folder "app", "/var/app", :mount_options => [ "uid=510,gid=510", "dmode=775", "fmode=774" ]
21:32 dave_den jeremyBass1: https://github.com/mitchellh/vagrant/issues/1646
21:32 jeremyBass1 hmm.. you know one thing about coder wall is there is never daes on things lol. .
21:32 jeremyBass1 hmm .. issue is it's not some provider ..
21:33 dave_den n0c: default state files are yaml, so it's yaml convention
21:34 n0c I guess I'm asking the wrong question
21:35 zfouts_ Yay. Just got the go ahead to automate our entire deployment process with SaltStack.
21:35 zfouts_ :D * 500
21:35 jeremyBass1 ok well I'll run down that route .. it's odd cause it's pulling files and what not.. it's just that the .sh get permission issues.. the salt oflder.. but i can get the rest
21:35 utahcon n0c: are you familiar with YAML?
21:35 utahcon the - is a marker for a list
21:35 utahcon or are you looking for using the - on the cli?
21:36 ndrei joined #salt
21:37 utahcon n0c: this may help http://docs.saltstack.com/ref/cli/ look for the section Arguments are formatted as YAML:
21:39 n0c my question is more one of.. why are we passing a list of dicts instead of a dict?
21:39 n0c this is what the YAML turns into: http://sprunge.us/Njie?py
21:39 n0c if you take off the '- ', I believe you get a simple dict
21:41 utahcon well that is beyond my league.
21:41 n0c sorry, just trying to understand why things are the way they are :)
21:42 cewood joined #salt
21:44 zfouts_ Has anyone setup Saltstack to "deploy" code in a CI environment?
21:44 zfouts_ Much like something Jenkins would do
21:44 zfouts_ I hate Jenkins though, it is the spawn of Satan
21:45 Ahlee Yes. I can chat a bit about it after this breakout session
21:45 n0c utahcon: I just think that this makes more sense: http://sprunge.us/EPeV?py
21:45 Ahlee in my case I deploy with salt on completeion of builds via atlassian's bamboo
21:45 utahcon zfouts_: https://github.com/saltstack/salt-jenkins
21:45 zfouts_ Very nice
21:45 zfouts_ utahcon: yea but I'm anti-jenkins :-)
21:46 zfouts_ utahcon: the idea is to deploy the code with salt vs using a CI product like jenkins or bamboo
21:46 utahcon I am not anti-jenkins yet.
21:46 zfouts_ I suppose the automated build of the product would be via some CI software though
21:46 utahcon I am deploying code with salt just using git.latest -- probably not the best way, but it works
21:46 zfouts_ That's an idea
21:46 utahcon I guess it also depends on what you are deploying and what your deploy entails
21:47 zfouts_ It's a java application
21:47 utahcon mine is all PHP (don't laugh)
21:47 utahcon lol
21:47 jordan_ joined #salt
21:47 zfouts_ so .jar's and wars
21:47 utahcon in theory though you could write a state that does a checkout, and then the build, and then puts things where they belong
21:47 wolfpackmars2 joined #salt
21:47 zfouts_ that is a possibility
21:48 zfouts_ I will force a dev into writing a new module and making it opensource
21:48 zfouts_ :D
21:48 zfouts_ Glad I'm at a Python shop now
21:48 jordan__ joined #salt
21:48 zfouts_ Ahh, saltconf is going on! I had almost forgot
21:48 zfouts_ I wish I was there with you guys
21:49 Ahlee zfouts_: on bamboo completed build, assuming all tests pass, etc, i have bamboo run a script that hits salt-api (https://github.com/saltstack/salt-api) and target grains to run a custom _module to deploy the software
21:49 Ahlee my apps are all built with c++
21:49 cewood joined #salt
21:49 zfouts_ Oh awesome
21:50 zfouts_ That could be a potential thing
21:50 Ahlee I would recommend foregoing the _module, and look into doing _states
21:50 Nazca__ zfouts_: ditto .... I hear they're giving out pretty paper things with shiny attached!
21:50 Ahlee define a custom _state/ to do your deploy work, then jinja template
21:50 zfouts_ Nazca__: must be nice!
21:50 Ahlee to pass in the proper version to deploy
21:50 Teknix joined #salt
21:50 zfouts_ Ahlee: perfect, I will do that!
21:51 Nazca__ yeah ... as I said earlier, maybe next year my boss will sponsor me the trip lol
21:51 Ahlee saltconf is going awesome :D
21:52 Nazca__ it's making them slack on PRs though ... I've had one open for hours now! ^_^
21:54 Ahlee lol
21:54 dlindquist joined #salt
21:58 NotreDev joined #salt
22:00 LowWalker joined #salt
22:02 ajw0100 joined #salt
22:03 Gordonz joined #salt
22:03 Thiggy joined #salt
22:04 toofer joined #salt
22:04 cachedout joined #salt
22:04 tvle joined #salt
22:07 jordan_ joined #salt
22:08 dlindquist joined #salt
22:08 yomilk joined #salt
22:09 justBob joined #salt
22:09 gnu-man joined #salt
22:09 harobed_ joined #salt
22:12 jlund joined #salt
22:13 Ryan_Lane joined #salt
22:13 n0c so I have a theory utahcon, not that you're still interested.. :)
22:14 n0c a list of single-item dicts would preserve order, whereas a simple dict would not
22:14 n0c which, if you wanted your state file to be read top-to-bottom, I can see that being important
22:15 tempspace joined #salt
22:16 gnu-man joined #salt
22:20 abe_music joined #salt
22:21 zfouts joined #salt
22:26 david_a joined #salt
22:29 rgbkrk joined #salt
22:29 hunter joined #salt
22:30 [diecast] joined #salt
22:31 dlindquist joined #salt
22:31 hunter is there a list of ext_pillars anywhere?
22:31 jtang1 joined #salt
22:33 Ryan_Lane hunter: http://docs.saltstack.com/salt-modindex.html#cap-p
22:34 toofer joined #salt
22:34 elfixit joined #salt
22:34 hunter Ryan_Lane: thanks. Shucks - no etcd yet. I may be forced to write that yet... :)
22:34 Ryan_Lane yw
22:36 ksalman heya, Can i not do something like this? I need to include one template from another, and use the same variable in both templates https://gist.github.com/anonymous/8698623
22:36 chuffpdx joined #salt
22:37 ksalman I get the error Rendering SLS template1 failed: Jinja variable 'directory'  is undefined'
22:37 ksalman when i include template1 in template2, does it not get the variable forwarded to it?
22:38 LowWalker joined #salt
22:39 rgbkrk_ joined #salt
22:40 jacksontj joined #salt
22:40 wblair left #salt
22:41 vejdmn joined #salt
22:48 kermit joined #salt
22:52 dlindquist joined #salt
22:54 zzzirk joined #salt
22:55 Kyle joined #salt
22:56 CheKoLyN Hello all, is there a way to manually run a reactor sls  file?
22:58 jtang1 joined #salt
22:59 alunduil joined #salt
22:59 jacksontj joined #salt

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