Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-09-12

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

All times shown according to UTC.

Time Nick Message
00:01 cjh as far as salt multimaster goes is there a best practice for salt minions and their crontabs?  what i mean is how should i setup the salt-call such that if my first master dies I call the second one
00:01 krissaxton joined #salt
00:05 juanlittledevil question for you guys… are any of you guys using gitfs in your masters? I"m trying to setup a master/staging/testing environment. and I'd like to know what best practices are for such. In the past I've configured my puppet servers to use version control and then used post commit hooks to check syntax and push new configs out. I'd like to do something similar with salt, so a) syntax, functionality, etc is checked in a safe environment prior to pushi
00:06 juanlittledevil I'm wondering if the correct way to do this is via gitfs or if I should be handling this with post commit hooks which call the salt commands instead
00:09 sgviking joined #salt
00:12 bejer joined #salt
00:16 jesusaurus cjh: if you list both in the minion config, then it will try the first and fallback on the second
00:16 baniir joined #salt
00:19 [diecast] joined #salt
00:21 forrest joined #salt
00:28 cjh jesusaurus: that's true but wouldn't that seem inefficient?
00:29 oz_akan_ joined #salt
00:30 tim5464 joined #salt
00:39 sibsibsib_ joined #salt
00:42 nkuttler joined #salt
00:42 nkuttler joined #salt
00:47 jdenning joined #salt
00:49 jesusaurus how so? that sounds like it does what you want
00:49 jesusaurus unless you want to do real load balancing, in which case just point the minions at an haproxy
00:51 drags cjh: if you want the minions to have multiple masters to point at for load spreading _and_ failover, alternate the ordering of the masters per minion node
00:52 drags each minion gets a randomly assigned list of masters.. distributes the load while still allowing fallback
00:52 drags adding a haproxy just moves the SPoF around
00:53 oz_akan_ joined #salt
00:55 juanlittledevil time to head home for the day. I will try asking tomorrow again. You guys have a great evening.
00:55 juanlittledevil left #salt
00:57 emocakes joined #salt
00:59 craig whiteinge: trying to add some events into .../rest_cherrypy/app.py
00:59 craig it works for the code that i KNOW.. :)   (salt_ip_verify_tool)
01:00 craig but, i added the same block to "salt_auth_tool()" and it never gets fired...
01:00 halfss joined #salt
01:07 krissaxton joined #salt
01:07 neomaan joined #salt
01:07 neomaan left #salt
01:12 g4rlic forrest: good news.
01:13 g4rlic In Fedora, the SELinux problem appears solved.
01:13 g4rlic So whatever new policy from selinux-policy-targeted RPM appears to work with salt and running scripts from RPM's
01:16 liuyq joined #salt
01:16 redondos joined #salt
01:17 Gwayne joined #salt
01:18 lineman60 joined #salt
01:20 baniir default /etc/mysql/debian.cnf includes info about the user debian-sys-maint ; i've pointed to this for use by states.mysql_grants, but my state is failing to execute "GRANT all ON *.* TO root@%"; how can initialize this to permit the mysql module to make the modifications specified in the sls
01:22 shinylasers joined #salt
01:22 imaginarysteve joined #salt
01:22 Nexpro joined #salt
01:24 brianhicks joined #salt
01:33 backjlack_ joined #salt
01:38 mwillhite joined #salt
01:39 liuyq joined #salt
01:47 mannyt joined #salt
01:49 mmilano joined #salt
01:49 markm_ joined #salt
01:50 blee_ joined #salt
01:55 mapu joined #salt
01:57 deepakmd_oc joined #salt
02:05 sixninetynine joined #salt
02:07 halfss joined #salt
02:07 krissaxton joined #salt
02:14 mannyt joined #salt
02:16 pengyao__ joined #salt
02:17 __number5__ anyone know a good chef/salt comparison article I can show to my boss? (to convince him salt is better than chef)
02:17 fengyu_ joined #salt
02:18 __number5__ would be better if the article cover cloud/aws management. thanks
02:19 Thiggy joined #salt
02:22 fengyu_ joined #salt
02:25 xl1 joined #salt
02:28 fengyu__ joined #salt
02:29 superflit joined #salt
02:31 oz_akan_ joined #salt
02:33 whit joined #salt
02:35 clion joined #salt
02:37 cvrebert joined #salt
02:39 cvrebert Are the abstract interfaces for package manager modules or service (as in /etc/init.d) management modules documented/specified anywhere?
02:42 redondos joined #salt
02:50 redbeard2 joined #salt
02:53 baniir joined #salt
03:00 bhosmer joined #salt
03:03 UtahDave joined #salt
03:09 redbeard2 joined #salt
03:09 krissaxton joined #salt
03:09 racooper joined #salt
03:13 fengyu_ joined #salt
03:22 __number5__ cvrebert: are you looking for this http://docs.saltstack.com/ref/modules/all/salt.modules.service.html#module-salt.modules.service?
03:23 cvrebert __number5__: No, I mean the abstract interface implemented by that module and debian_service, rh_service, etc.
03:27 whit joined #salt
03:35 markus_ joined #salt
03:39 markus_ left #salt
03:40 markuz joined #salt
03:40 markuz left #salt
03:41 oz_akan_ joined #salt
03:47 mmilano left #salt
03:49 markuz__ joined #salt
03:53 josephholsten joined #salt
03:56 druonysus joined #salt
03:56 druonysus joined #salt
03:59 jslatts joined #salt
04:01 redondos joined #salt
04:01 freelock_ joined #salt
04:07 forrest g4rlic, I saw your message from earlier as well as the update on the issue. Hopefully they push that to centos/rhel at some point!Q
04:07 mmilano joined #salt
04:09 krissaxton joined #salt
04:13 mwillhite joined #salt
04:17 baniir does cmd.run honor require: - pkg.installed
04:17 forrest you can make a cmd require a package yes
04:17 forrest but it would be - require: - pkg: package_name
04:17 forrest obviously indented properly and such
04:27 baniir when should pkg.installed be used
04:31 forrest when you want to verify that a package is already installed
04:31 forrest so for example, you build a box with a kickstart, then run salt afterwards
04:31 forrest where your kickstart puts on some default packages, and you have something else that you know needs to require that package is present.
04:32 forrest so you'd need to confirm the package is installed first (otherwise salt won't know about it), and then you can require that pkg.installed
04:35 freelock_ forrest: thanks for responding earlier... I actually configured an aws key, and "salt-call -a show_instance instancename" pulled up my existing instance, no config necessary!
04:35 freelock_ and salt-call -a stop/start instancename worked just fine
04:35 freelock_ s/salt-call/salt-cloud/
04:35 forrest oh cool
04:36 freelock_ yeah, too easy! Was shocked that it just worked when I tried that
04:38 octagonal joined #salt
04:39 baniir forrest: thanks
04:39 forrest freelock_ yea that is cool
04:39 forrest np baniir.
04:46 forrest whiteinge, I like your timing on pushing that update to develop, was no more than 15 seconds after I had updated my clone.
04:49 whiteinge heh
04:49 whiteinge we like to keep you on your toes
04:49 forrest yea no biggie, just made me laugh
04:52 Ryan_Lane joined #salt
04:55 robawt oh shoot it's whiteinge hide the good beer
04:55 * whiteinge scrambles for the good beer
04:56 robawt what's up good Sir?
04:57 robawt i must admit i finally just installed salt-vim
04:57 whiteinge mostly wishing for good beer, dang you
04:57 robawt why on Earth I waited so long i'll never know
04:57 robawt lol
04:57 robawt drop by, I'll hook you up
04:57 whiteinge woot. and now your sls-writing live is changed forever
04:57 whiteinge s/live/life
04:57 robawt for the better
04:57 robawt my vim fu is improving, it was only a matter of time
04:59 whiteinge i can sense your greater vim-fu from here
05:00 robawt no that's just my saltstack sticker on my laptop
05:00 forrest I wish there was a salt plugin for sublime text 2
05:02 UtahDave salt-vim is teh awesome
05:02 robawt forrest++
05:02 robawt UtahDave speaks the truth as always
05:02 UtahDave working 10 feet from whiteinge has been the best thing for my salf fu  EVAR
05:02 UtahDave and git fu, for that matter
05:02 whiteinge there was some development interest in sublime text support a few months ago. i even made a repo for the guys involved but i haven't heard anything since :-/
05:02 robawt i can't even imagine UtahDave
05:04 UtahDave gqap
05:04 middleman_ joined #salt
05:05 UtahDave bam!
05:05 forrest I still feel like with git a lot of the time it's either 'crap how do I fix this error', and I have to scuttle off to the internet to figure it out, or 'oh please let pushing this huge red button not break everything.'
05:06 forrest whinteinge do you exclusively use VIM to write code?
05:06 forrest *whiteinge
05:07 UtahDave oh, yeah
05:07 whiteinge forrest: yes
05:07 forrest wow
05:08 whiteinge i've been using vim full-time for ~10 years
05:09 robawt it also explains why he never prints anything :P
05:09 forrest lol
05:09 UtahDave whiteinge is a vim master.  I remember sitting several times in Utah Python user group meetings watching him give a presentation on various subjects
05:10 UtahDave and saying "Wait!!!  how did you do that??"   What was that??
05:10 forrest lol
05:10 UtahDave "Is that even possible??"
05:10 UtahDave "What the..."
05:10 forrest I feel like I'm awesome at vim when I yank and drop in the lines first try
05:10 robawt i heard a rumor that he doesn't write salt states with more than a handful of keystrokes anymore for entire infrastructures
05:10 UtahDave When I got my SaltStack job offer letter Tom told me that Seth was coming on board and I was so stoked.
05:11 krissaxton joined #salt
05:11 whiteinge robawt: it's true...but i cheat by just including formula ;)
05:12 forrest you  broke the repo with your merge whiteinge :(
05:12 * whiteinge blames terminalmage!
05:12 forrest yea I looked at the commit
05:12 sssslang need some help. test.ping works for all minions, but state.highstate fails only on one minion. all the minions have same configuration.
05:12 terminalmage haha
05:12 forrest What's the opinion on pushing a merge when the build isn't passing, is that ok?
05:12 forrest sssslang, what is the failed minion returning on a debug?
05:12 sssslang what should I do?
05:12 terminalmage forrest: some of our tests fail seeminly randomly
05:13 forrest yea I noticed that
05:13 terminalmage so it's not TOO crucial
05:13 forrest happened last time I updated the docs
05:13 sssslang forrest: yes, I'll paste out.
05:13 forrest ok cool
05:13 terminalmage sorry about the failure... I needed a commit today to keep my streak alive
05:13 forrest pull request away then
05:15 sssslang forrest: http://pastebin.ca/2448835
05:16 forrest so the source file isn't found on the samt master or so it claims. I assume you're applying the exact same state to other machines and it passes?
05:16 forrest I know you said you were earlier, but there's no unique config per machine
05:17 shinylasers joined #salt
05:18 sssslang other minion can fetch this file.
05:18 vipul_ joined #salt
05:18 vipul_ hi
05:19 forrest ok, can you try to run just that one state that includes the copying over of the file, and then untarring it on that specific minion?
05:19 forrest hey vipul_
05:19 vipul_ hi forrest
05:19 vipul_ how u doing
05:19 forrest good
05:19 Furao joined #salt
05:19 forrest cursing terminalmage and his terrible commits :P
05:20 vipul_ ok
05:20 vipul_ forrest :  I am new to SALT
05:20 forrest great!
05:20 terminalmage forrest: it was just a docs commit though
05:20 vipul_ can u help me in understanding it
05:20 forrest terminalmage, I know I'm joking
05:20 terminalmage :(
05:20 terminalmage ha
05:21 forrest vipul_, what part are you having a hard time understanding?
05:21 sssslang forrest: how to run just one sls?
05:21 terminalmage forrest: I have broken pylint with docs commits before though
05:21 terminalmage that's always fun
05:21 vipul_ simply how to start
05:21 vipul_ what is salt all about
05:21 vipul_ i m a software enginner in US based MNC
05:22 forrest sssslang, just put the state name instead of state.highstate
05:22 vipul_ currently we are using vclud of VM ware as a server management tool
05:22 vipul_ now we are planning to use salt
05:23 forrest vipul_, you know that's a lot of stuff to explain, if you want to start from the beginning check out this video from a couple months ago: http://www.youtube.com/watch?v=yphLKSjnSU8
05:23 forrest Tom does a great job of explaining stuff
05:23 forrest and it's entertaining enough to keep your attention since he's presenting it at a user group instead of for investors.
05:23 vipul_ great forrest !
05:23 sssslang forrest: like salt '*' top.sls?
05:23 forrest yep
05:24 vipul_ will go through this stuff
05:24 UtahDave sssslang: salt '*' state.sls mycustomstate
05:24 vipul_ catch u later
05:24 vipul_ thanks a lot
05:24 forrest oh duh UtahDave, my bad
05:24 sssslang UtahDave: No matching sls found for 'debug.sls' in env 'base'.
05:24 forrest np vipul_, have a good one
05:24 sssslang need copy to minion first?
05:24 forrest is it in a lower directory sssslang?
05:25 sssslang /srv/salt/debug.sls
05:25 UtahDave sssslang: can you paste the command you ran?
05:26 sssslang seems I add a .sls after state name. sorry, my fault.
05:27 jdenning joined #salt
05:28 mmilano left #salt
05:28 UtahDave cool
05:28 sssslang I run for one minion only, didn't get response till now.
05:29 forrest what do you mean?
05:29 sssslang like waiting timeout.
05:29 mcsrainbow joined #salt
05:30 sssslang command 'salt 'the-minion' state.sls debug' didn't return.
05:30 sssslang sorry for my poor English.
05:30 sgviking joined #salt
05:30 forrest Your English is fine, I was asking for clarification more than anything
05:31 forrest can you see what's going happening on the minion?
05:31 sssslang minion's log didn't show anything.
05:31 forrest hmm, the minion is on the same version as the others right?
05:31 forrest of salt that is
05:32 sssslang yes, 0.16.3.
05:32 forrest ok, can you check the job queue?
05:32 sssslang some time target only one minion in the command line works, but using '*' fail for that minion.
05:33 forrest so it worked for everyone minion using '*' except that one even on the command line, running only that single state?
05:33 sssslang still active.
05:33 forrest is there anything unique about that minion that makes it different from the others? firewall configuration, network config, anything?
05:35 sssslang two points: 1. it worked for everyone minion using '*' except that one. 2. using 'that-minion' targeting method sometime success, but cost unbelivable time.
05:35 forrest but when it does actually finish, does it return with those errors you showed earlier?
05:35 forrest so even though it takes that long, it still does't finish
05:36 sssslang I think test.ping work means firewall is setting correct, right?
05:37 forrest if test.ping is actually using ping, then that just means it can hit the box, ping uses icmp so it doesn't understand ports
05:37 UtahDave test.ping doesn't use icmp.  It just has the minion return True
05:37 sssslang what about cmd.run pwd?
05:37 forrest but when you run that command manually for that specific box, you said it sometimes succeeds, but when it succeeds, does that mean that it downloads and untars the tarball, or just that the job returns
05:38 forrest oh good to know UtahDave
05:38 tomtomtom joined #salt
05:38 krissaxton joined #salt
05:38 sssslang I mean the job done, not just command returns.
05:39 forrest Hmm, I don't see why it wouldn't be logging anything in that sort of scenario
05:39 sssslang should I restart that minion?
05:39 forrest worth a shot
05:39 forrest just the service
05:40 forrest hey terminalmage are you still around?
05:41 terminalmage yeah
05:42 forrest do you know why it was decided that when using - sources: on pkg.installed it was decided to use rpm -qpi for the install instead of yum localinstall on CentOS/RHEL/Fedora machines?
05:42 * terminalmage sighs
05:42 forrest when you do it with rpm as opposed to yum, it doesn't install deps
05:42 terminalmage what does rpm -q do
05:43 terminalmage ?
05:43 forrest query
05:43 terminalmage that's right
05:43 terminalmage so, -qpi would never actually install anything, would it
05:43 forrest Right
05:43 forrest sorry, should have reworded it
05:43 terminalmage https://github.com/saltstack/salt/issues/7184#issuecomment-24284255
05:44 terminalmage forrest: ^^^
05:44 terminalmage that should answer the question
05:44 terminalmage this is a total non-issue
05:44 forrest yea this is what GradysGhost and I were talking about yesterday
05:44 forrest so it queries
05:44 forrest then does nothing
05:44 terminalmage check out my comment
05:44 terminalmage it explains everything
05:44 forrest gotcha
05:45 forrest I see now, sorry I didn't send myself the issue yesterday.
05:45 terminalmage personally, I find it kind of annoying that someone would open an issue saying that we should be using yum to install packages, without looking at the source code and seeing that that is exactly what we're doing
05:46 NV forrest: ended up using a templated file.managed state for the ssh keys
05:46 forrest nice NV, how did that work out?
05:46 terminalmage but whaddaya gonna do?
05:46 forrest that was probably partially my fault terminalmage, I think we were discussing it and I looked at it and linked that, probably didn't keep following the code all the way back for some reason
05:47 terminalmage yeah... I was just a little perturbed earlier when I saw that issue
05:47 terminalmage there's not a single part of it that is accurate
05:47 terminalmage lol
05:47 forrest lol
05:47 robawt NV: please please use pillar for ssh keys
05:47 robawt file.managed with a jinja template file is a beautiful thing
05:48 forrest yea I don't know why I assumed it did the check, then tried to install with rpm.
05:48 terminalmage probably because rpm -i installs
05:48 terminalmage and "i" is in -qpi
05:48 forrest yeaaaaa no I use rpm -q* all the time, should have been paying better attention
05:51 tuxIO joined #salt
05:52 NV forrest: works well :D
05:52 forrest did you see what robawt was saying? Probably a good idea if you didn't alread set it up like that
05:52 forrest shouldn't be TOO much work to transition
05:52 NV yeah i have a pillar that has sftp_users
05:53 forrest ok awesome
05:53 NV which is a dict of usernames
05:53 NV whose values are an array of ssh keys
05:53 NV :D
05:53 forrest cool, so a better iteration on what we discussed earlier!
05:54 NV creates the users, adds them to the sftp group, sets up their chrooted homedir with correct permissions, sets their authorized keys file (which lives in a non-standard location outside of their chrooted home), etc
05:54 forrest nice
05:54 forrest sssslang, any luck there? I was gonna head out soon
05:54 NV only thing it doesn't do is delete users when they're removed, but that's considered a safety feature imho :P (it'll remove all authorized_keys entries though so nobody can log in anyway)
05:55 forrest yea that seems like a good plan.
05:55 sssslang forrest: still failed after minion restart.
05:55 forrest I've got nothin then ssslang :(
05:55 robawt NV: great job
05:55 sssslang cmd cost 5 minutes. is it meet some timeout?
05:56 NV semi-related - did you see my roles python pillar the other day?
05:56 forrest sssslang, it might be
05:56 forrest NV, no I didn't
05:57 NV http://pastie.org/private/vwtcuhsrezxtla2acoffpq and an associated roles.yaml in the same directory that it reads - http://pastie.org/private/qbx6geivtt8uuom41nuya
05:58 NV full node matching that includes pillars of the same name if they exist (ie, create jira.sls pillar to have it automagically included), and allows targeting in the states top.yaml
05:58 forrest nice
05:59 NV full node matching (al-a top.sls style), except you only have to do it once in one place
06:00 NV related: salt-master needs the ability to be reloaded :P
06:00 mesmer joined #salt
06:01 forrest heh
06:01 Katafalkas joined #salt
06:01 NV would make nodegroups a tad more useful (although imho my roles.yaml is a tad easier to manage anyway)
06:03 forrest heh
06:03 forrest alright I've gotta head out, let me know if you figure out what is going on with that minion sssslang
06:04 sssslang forrest: I'm reading minion's debug log, seems a lot of ps output.
06:04 sssslang ok.
06:13 halfss joined #salt
06:15 emocakes joined #salt
06:18 olaf38 joined #salt
06:21 linjan joined #salt
06:21 berto- joined #salt
06:28 vipul_ Hi forrest
06:28 vipul_ there
06:30 vipul_ joined #salt
06:30 vipul_ hi
06:30 vipul_ anybody there
06:30 sssslang hi vipul_
06:31 vipul_ hi sssslang
06:31 vipul_ hru
06:31 vipul_ is salt is unix based software?
06:31 sssslang still debuging :-/
06:31 vipul_ or do we have gui for it
06:31 ml_1 joined #salt
06:31 sssslang salt server should run on *nix.
06:32 vipul_ ok
06:32 mmilano joined #salt
06:32 vipul_ it can not run on windows
06:32 sssslang what do you managed to see through gui?
06:32 sssslang it is reported client can run on windows.
06:32 vipul_ with client u mean minion?
06:33 sssslang yep. as server for master.
06:34 vipul_ just want to confirm , u mean we can not install salt on windows , but we can install minion on windows
06:34 xl1 joined #salt
06:34 vipul_ right?
06:36 redondos joined #salt
06:36 bhosmer joined #salt
06:37 vipul_ ssslang u there?
06:38 xl2 joined #salt
06:39 emocakes joined #salt
06:40 halfss_ joined #salt
06:43 gildegoma joined #salt
06:50 krissaxton joined #salt
06:53 xl1 joined #salt
06:54 ronc joined #salt
06:55 malinoff joined #salt
06:56 balboah_ joined #salt
06:57 deepakmd_ joined #salt
06:57 emocakes joined #salt
07:00 krissaxton joined #salt
07:01 joehh vipul_: you can install salt minion on windows - it has been very good since about 16.2
07:01 joehh 0.16.2
07:01 joehh the salt master should run on a unix server - we typically use debian or ubuntu
07:01 deepakmd_oc joined #salt
07:03 nu7hatch1 joined #salt
07:16 sfello joined #salt
07:22 falican joined #salt
07:25 Charles_ joined #salt
07:35 sssslang vipul_: sorry, not here just now.
07:36 all_the_fisch joined #salt
07:37 emocakes joined #salt
07:38 halfss joined #salt
07:46 mgw joined #salt
07:51 xl1 joined #salt
07:52 xmj left #salt
07:55 deepakmd_oc joined #salt
07:57 malinoff joined #salt
07:59 majk joined #salt
08:00 matanya joined #salt
08:02 xl2 joined #salt
08:05 matanya joined #salt
08:05 matanya joined #salt
08:07 SevenDay joined #salt
08:08 bud_ joined #salt
08:09 bud_ Hello all :).
08:10 bud_ Question: if I have ext_pillar: \n  - git: master file:///some/where/salt-pillar.git, and I change something in the pillar repository and updated it on the master shouldn't saltutil.refresh_pillar checkout the freshest data from the repository?
08:11 halfss_ joined #salt
08:11 bud_ I guess it should, right?
08:12 NV err, no
08:12 NV i found that the git ext pillar was rather difficult to coerce into updating
08:12 NV well
08:12 NV i guess it should
08:12 NV but in my experience it probably doesn't is what I mean
08:12 bud_ I think so also.
08:13 NV also, check your master log file, i got some weird errors and it failed to update a lot for me
08:13 bud_ Thanks NV. I'll put the pillar on the fs and use hooks to pull the freshest data on push.
08:13 NV i just use a git.latest state for my pillar now
08:13 az87c joined #salt
08:14 bud_ I run the master in debug mode. No errors for now.
08:14 NV then doing a highstate on my saltmaster forces it to update
08:14 bud_ That's cool also :).
08:14 NV mhmm, hooks work too :D
08:14 NV my git repos are on a different box etc though
08:17 rorski joined #salt
08:17 shinylasers joined #salt
08:21 redondos joined #salt
08:21 KennyW_ joined #salt
08:23 kennyw_ joined #salt
08:26 yml how would you add this in a pillar : sudo salt "app1.staging" publish.publish app*.staging network.ip_addrs eth1
08:27 yml I am trying something like this : {% set app_servers = salt['publish.publish']('app*.staging', 'network.ip_addrs', 'eth1').values() %}
08:28 yml but it blow up in my face with this traceback: https://dpaste.de/mvtGa/
08:29 tomviner joined #salt
08:37 malinoff yml: Try to use salt-call publish.publish app*.staging network.ip_addrs eth1
08:37 malinoff yml: But Salt Mine is way better to collect ip addrs
08:38 malinoff yml: Uh. Forget my 1st message
08:38 * yml focusing on the second :-)
08:39 malinoff But your approach should work
08:40 malinoff anyway...
08:41 malinoff Did you enable publish with allowed network.ip_addrs with eth1 arg on the master?
08:43 yml publish is working fine when I use the module
08:43 yml I my issue is when I try to get my hand on this data in the pillar
08:45 yml malinoff: do you have an example of state using mine
08:45 yml I did this : https://salt.readthedocs.org/en/latest/ref/modules/all/salt.modules.mine.html?highlight=mine
08:45 yml no
08:46 yml i did this : sudo salt '*' mine.send network.interfaces
08:46 yml and then : sudo salt '*' mine.get '*' network.interfaces
08:46 yml I see the data I am interested in
08:46 olaf38 joined #salt
08:47 yml now the question is how do I use this in a pillar
08:47 malinoff yml: {% set addrs = salt['mine.get']('*', 'network.interfaces') %}
08:48 gmoro joined #salt
08:52 yml malinoff: then the data is pretty nested
08:52 qba73 joined #salt
08:52 yml how do you run the run.send ?
08:52 yml do you add a cronjob
08:52 yml ?
08:53 yml or do you also put it in a state somewhere ?
08:55 malinoff yml: You don't really have to 'send' the mine data, it has own scheduler http://docs.saltstack.com/topics/mine/index.html#mine-interval
08:55 malinoff You only must to collect the data
08:56 yml I see
08:56 xl1 joined #salt
08:56 malinoff And there you can use scheduler http://docs.saltstack.com/topics/jobs/schedule.html?highlight=scheduling , or cron
08:56 yml also I getting an error when I add it in the pillar :-/
08:57 yml malinoff: https://dpaste.de/Z3u4h/
08:57 yml {% set addrs = salt['mine.get']('app*.staging', 'network.interfaces') %}
08:57 yml this is what I have added in my pillar
08:59 bogdanbu joined #salt
08:59 malinoff yml: Do you have a master? Or you're trying a masterless configuration?
08:59 yml I have a master
09:01 krissaxton joined #salt
09:01 backjlack joined #salt
09:02 scott_w joined #salt
09:03 bogdanbu hi ! i'm trying to configure an ext_pillar but without success :(
09:04 bogdanbu i have in /etc/salt/master ext_pillar: - test_pillar
09:04 malinoff yml: There's something strange with the output. You'd better to ask in salt-users google group, i think
09:04 bogdanbu and test_pillar.py is http://pastebin.com/pxFxsFVu
09:04 yml thanks malinoff
09:05 tomviner joined #salt
09:05 bogdanbu in logs: /var/log/salt/master: ][CRITICAL] The "ext_pillar" option is malformed
09:06 bogdanbu why is malformed ?
09:12 tomviner joined #salt
09:14 yml malinoff: for some reason moving the state from pillar to a state "fix" the error
09:15 yml so it must be somehting related with putting it in the pillar
09:21 TheCodeAssassin joined #salt
09:21 utahcon joined #salt
09:24 fredvd joined #salt
09:36 vaxholm joined #salt
09:36 kchr joined #salt
09:41 malinoff yml: It's strange because pillars are compiled on a minion (just like states)
09:42 malinoff You should really ask about that in salt-users
09:43 bogdanbu solved :)
10:00 derelm joined #salt
10:01 krissaxton joined #salt
10:04 bemehow joined #salt
10:06 felixhummel joined #salt
10:06 Tekni joined #salt
10:13 bhosmer joined #salt
10:20 superflit joined #salt
10:24 colesy joined #salt
10:29 MK_FG joined #salt
10:29 MK_FG joined #salt
10:37 Koma joined #salt
10:46 TheCodeAssassin joined #salt
11:02 krissaxton joined #salt
11:05 lemao joined #salt
11:11 durnik joined #salt
11:12 bhosmer joined #salt
11:15 LucasCozy Hello , Is there a state to call a refresh_modules for the minion itself.
11:19 s0undt3ch_ joined #salt
11:24 bhosmer_ joined #salt
11:27 oz_akan_ joined #salt
11:29 s0undt3ch joined #salt
11:29 oz_akan_ joined #salt
11:34 ml_1 joined #salt
11:44 olaf38 joined #salt
11:49 bemehow joined #salt
11:55 superflit_ joined #salt
12:02 derelm joined #salt
12:03 pmrowla joined #salt
12:04 renoirb joined #salt
12:04 mwillhite joined #salt
12:05 freelock_ joined #salt
12:07 efixit joined #salt
12:07 daBrado joined #salt
12:12 balboah_ how do you upgrade salt-minion through a salt-master command?
12:13 bhosmer joined #salt
12:14 Kholloway joined #salt
12:20 tuxIO joined #salt
12:22 copelco joined #salt
12:24 jslatts joined #salt
12:26 joehh balboah_: which os?
12:26 natim joined #salt
12:26 premera joined #salt
12:26 natim Hello guys
12:26 balboah_ ubuntu 12.04, salt versions can be as old as 0.13
12:26 natim How can I run python setup.py develop after the virtualenv creation ?
12:26 balboah_ I tried doing a cmd.run with apt-get stuff, but it didn't work out so nice
12:27 balboah_ I guess because it prompts about the config file and what not
12:27 joehh I believe salt 'x' pkg.install salt-minion would probably work
12:28 joehh natim: you can do cmd.wait and watch the virtualenv state
12:28 halfss joined #salt
12:29 natim joehh, do you have an example
12:29 joehh natim: not handy unfortunately
12:30 emocakes joined #salt
12:31 avienu joined #salt
12:32 balboah_ joehh: well that's embarrassing. it worked
12:32 joehh good to hear
12:33 tomviner joined #salt
12:33 joehh natim: http://docs.saltstack.com/ref/states/all/salt.states.cmd.html from the bit that begins with " Watch for changes to a git repo"
12:33 joehh just replace the git bit with a virtual env bit
12:37 blee joined #salt
12:39 derelm joined #salt
12:49 bhosmer joined #salt
12:51 natim joehh, ok thanks
12:51 ipmb joined #salt
12:53 linjan_ joined #salt
12:58 anteaya joined #salt
13:02 GradysGhost joined #salt
13:03 abele joined #salt
13:05 pdayton joined #salt
13:06 pdayton joined #salt
13:06 natim joehh, it works
13:07 natim Salt : The more I use it the more I like it !
13:07 joehh it is pretty good, isn't it
13:07 natim yes
13:07 joehh simple, predictable, but very powerful
13:09 juicer2 joined #salt
13:11 Gifflen joined #salt
13:11 ingwaem joined #salt
13:12 tdk2fe joined #salt
13:13 fred_ joined #salt
13:14 racooper joined #salt
13:18 ml_1 joined #salt
13:18 olaf39 joined #salt
13:19 pakdel joined #salt
13:20 ingwaem Greetings everyone. I need a little help understanding cmd.wait…anyone familiar with this module?
13:21 adepasquale joined #salt
13:21 natim ingwaem, I just discovered it now
13:23 ingwaem natim, my problem is: I have a state file that uses file.recurse to make sure a folder on the minion has the same contents as the source folder on the master. I would like to run a command ONLY after that folder has been updated. I have a small call just below I do the file recurse which says cmd.wait, and watches file: and the folder I just synced but every time I run it it says the requisites were not found
13:24 ingwaem I can do a paste bin if you like
13:25 joehh ingwaem: that would be good
13:25 pdayton joined #salt
13:26 aptiko I understand that the equivalent of Puppet's "notify" is to use "extend" and "watch". Seems too verbose. There's no shortcut?
13:26 m_george|away joined #salt
13:26 ingwaem joehh, natim: http://pastebin.com/iapkJLm9
13:27 tyler-baker joined #salt
13:27 tyler-baker joined #salt
13:28 joehh It looks ok to me - I don't normally use the quotes around cwd and name args though
13:28 joehh could you have made a typo in anonomizing
13:28 joehh it
13:29 ingwaem I don't believe I've made a mistake. it's possible though. I'll try it without quotes
13:29 ingwaem what I might be able to do is quickly setup a test folder on the master, then reference that in a new state file to make the anonymized as close as possible to a real sample
13:29 joehh often I do it with a "nice" id rather than the path
13:29 redbeard2 joined #salt
13:29 joehh that is a good idea
13:30 tonthon joined #salt
13:30 joehh and use the name arg to refer to the path: http://pastebin.com/ewrLPCSF
13:31 ingwaem nice I see what you're doing there…but…will this also work for entire directories like the file.recurse function does?
13:31 toastedpenguin joined #salt
13:31 joehh yes
13:32 ingwaem cool I'll give it a bash :)
13:32 joehh key thing to note (caught a co-worker today) is that the watch needs to refer to the id, it can't refer to a subdirectory of the file.recurse state
13:33 ingwaem ahh
13:33 tonthon Is there some cache anywhere handling states ?
13:34 joehh tonthon: do you mean custom states? or something else?
13:34 tonthon no basic sls files
13:35 tonthon I modify them but a call to state.highstate is still using old ones
13:35 tonthon salt 0.16.4-2
13:35 tonthon (both master and minion)
13:36 imaginarysteve joined #salt
13:36 joehh tonthon: very odd - not something I've seen before
13:38 tonthon the same here, I imagine it's related to the new version ?
13:40 joehh do you see it all the time? or just sometimes?
13:40 tonthon joehh: sometimes and for a short period
13:40 joehh that makes it tricky
13:41 tonthon for example I add a package in a list and I need to fire state.highstate a few times before it runs fine
13:41 sgviking joined #salt
13:43 jeff__ joined #salt
13:43 briffle_ joined #salt
13:43 briffle_ left #salt
13:44 m_george left #salt
13:45 mwillhite joined #salt
13:45 briffle_ joined #salt
13:46 briffle_ joined #salt
13:46 brianhicks joined #salt
13:46 mattmtl joined #salt
13:46 QauntumRiff joined #salt
13:47 QauntumRiff is there any issue running the salt-minion client on the salt-master directly?
13:47 QauntumRiff I want to manage packages and security on the master server too
13:49 teskew joined #salt
13:50 bhosmer joined #salt
13:51 p3rror joined #salt
13:54 ingwaem joehh: thanks man! That was the problem…the quotes. I've tried a few different ways with a test folder, and yes, any quotes breaks it. happy to see that I didn't get all the logic totally messed up :)
13:54 LucasCozy QauntumRiff, no problem with that
13:55 QauntumRiff lucascozy: thanks, it seemed harmless, but just thought I would double check
14:02 ingwaem joehh: is it possible for watch to watch multiple directories or files…so if there are multiple file.manage to only run the commands after all the directory states are applied…I'm assuming that I can just do many - file: under the watch, however don't want to be an a55 assuming
14:05 ingwaem actually giving that some thought it seems best if I daisy chain the folders, so they only get copied if the others have successfully copied. that way the final compile process can be done when they're all in place as they all depend on the previous one
14:07 alunduil joined #salt
14:08 tuxIO_ joined #salt
14:09 mannyt joined #salt
14:09 mapu joined #salt
14:09 [diecast] joined #salt
14:12 sgviking joined #salt
14:14 Khazix joined #salt
14:14 robawt joined #salt
14:16 robertkeizer joined #salt
14:26 nonuby joined #salt
14:26 nonuby joined #salt
14:28 nonuby if I publish a custom state module via (_states/) and I want to take dependencies on another library where the py batteries included dont suffice how do I do this?  (presumably requirements.txt is ignored)
14:28 larico joined #salt
14:29 micah_chatt joined #salt
14:35 matanya joined #salt
14:35 whit joined #salt
14:38 StDiluted joined #salt
14:38 Koma joined #salt
14:42 Gifflen joined #salt
14:42 lorimar joined #salt
14:50 tuxIO_ joined #salt
14:51 tuxIO__ joined #salt
14:54 jalbretsen joined #salt
14:54 tyler-baker joined #salt
14:54 tyler-baker joined #salt
14:57 robertkeizer joined #salt
14:58 mgw joined #salt
14:59 deepakmd_oc joined #salt
15:00 blee joined #salt
15:00 LucasCozy Anyone know how to do a 'saltutil.sync_modules' but from a state file.
15:02 blee you can call salt modules like this {{ salt['module.function'](args) }}
15:02 blee ive always done that just to get pillars personally
15:02 cwright LucasCozy: look at module.run state
15:02 blee we*
15:03 blee or that ^
15:03 pjb1 joined #salt
15:04 LucasCozy Ok thank's a lot. I'll try this.
15:04 LucasCozy I did think about using run.
15:04 LucasCozy didn't*
15:05 [diecast] pillar data doesn't refresh by default when minions run?
15:05 pjb1 test
15:06 LucasCozy I think that blee means by 'personally'
15:06 copelco what's the difference between the two filters for salt mine? e.g. salt '*' mine.get '*' network.interfaces
15:07 pjb1 left #salt
15:07 pipps1 joined #salt
15:11 ckao joined #salt
15:11 tomviner joined #salt
15:14 austin987 joined #salt
15:14 jaequery joined #salt
15:14 bemehow joined #salt
15:18 JaredR joined #salt
15:21 agend joined #salt
15:21 agend hi
15:22 agend i have a problem with template rendering: {% if grains["shared_dirs"] %}
15:22 agend env: DJANGO_SETTINGS_MODULE=frontend.settings.staging
15:22 agend {% else %}
15:22 agend env: DJANGO_SETTINGS_MODULE=frontend.settings.production
15:22 agend {% endif %}
15:22 agend and /etc/frontend_uwsgi_config.yaml:
15:22 agend file.managed:
15:22 agend - temlpate: jinja
15:22 agend - source: salt://frontend/uwsgi_config.yaml
15:22 redondos joined #salt
15:22 agend ok - im an idiot - wrong template spelling
15:23 JaredR :)
15:23 JaredR It could have happened to anyone. I didn't see your misspelling.
15:24 lineman60 joined #salt
15:24 Furao joined #salt
15:24 agend thanks - trying to find what's wrong for last hour - i helped when i pasted it here, sorry
15:25 faldridge joined #salt
15:26 lineman60 joined #salt
15:27 UtahDave joined #salt
15:29 kermit joined #salt
15:31 TheCodeAssassin joined #salt
15:39 lineman60__ joined #salt
15:42 JaredR joined #salt
15:47 JaredR joined #salt
15:47 mapu joined #salt
15:48 felixhummel joined #salt
15:48 Lue_4911 joined #salt
15:51 KennethWilke joined #salt
15:52 Jahkeup joined #salt
15:53 JaredR joined #salt
15:53 SpX joined #salt
15:54 ronc joined #salt
15:55 curtisz joined #salt
15:55 JaredR joined #salt
15:56 pipps1 joined #salt
15:57 micah_chatt joined #salt
15:59 pipps joined #salt
16:01 JaredR So I see salt states to manage LVM (create it), and to manage mounts, but I don't see anything that manages the creation of a filesystem. What am I missing?
16:02 forrest joined #salt
16:03 mgw joined #salt
16:03 JaredR I figure that I'll use cmd.run or cmd.wait states to manage the creation of the filesystem, when it's needed.
16:15 redondos joined #salt
16:21 bhosmer_ joined #salt
16:22 jdenning joined #salt
16:23 troyready joined #salt
16:24 jbunting joined #salt
16:24 bitz joined #salt
16:29 nielsbusch joined #salt
16:29 UtahDave SaltStack is conducting a Salt Training in the Washington DC area in the week of October 9th.  Details here!  https://saltstackdctraining.eventbrite.com/
16:30 JoAkKiNeN joined #salt
16:31 KyleG joined #salt
16:31 KyleG joined #salt
16:31 forrest Does anyone know if the docs are pulled from the develop branch? I noticed that doc changes I pushed previously are on there, but the ones I pushed last night are not (even though I built them and confirmed they were present before doing the push).
16:35 tomeff_ joined #salt
16:36 Jahkeup joined #salt
16:42 AviMarcus joined #salt
16:42 UtahDave forrest: yeah, the docs are built from the develop branch, but we don't have them set to auto update.
16:43 UtahDave forrest: I'll kick off the build
16:43 forrest Oh it's all good you don't need to
16:43 forrest I just wasn't sure since the previous changes showed up 30 minutes later or something.
16:44 UtahDave Yeah, I think Seth ran the update last night.
16:44 gatoralli joined #salt
16:45 juanlittledevil joined #salt
16:45 juanlittledevil Good morning peeps.
16:45 gatoralli can I add a folder to the salt file server at runtime?
16:45 forrest Yea I did my commit after whiteinge pushed Terminalmage's changes.
16:46 forrest so that's probably why it didn't get caught
16:49 kamal_ SaltStack should really sell more getsalted shirts
16:49 pipps1 joined #salt
16:51 forrest they're usually free
16:51 forrest gotta go to more conferences
16:52 kamal_ yeah, they're free /if/ you attend a conference ;)
16:52 mapu joined #salt
16:52 forrest problem solved, more conferences :D
16:52 terminalmage kamal_: we've been known to mail out shirts to people that make cool contributions :)
16:52 brianhicks joined #salt
16:52 forrest I've got some stickers I can mail you if you want kamal_
16:53 kamal_ but I don't know python :(
16:54 benno joined #salt
16:54 benno left #salt
16:55 kamal_ forrest: no thanks they're yours :p
16:55 lineman60 joined #salt
16:55 forrest *shrug* I don't know what else to do with them
16:55 forrest they're just sitting on my desk here
16:56 kamal_ I have nowhere to stick them either :p
16:56 kamal_ My DigitalOcean stickers have been sitting on a shelf for months :p
17:00 Ryan_Lane joined #salt
17:00 gcbirzan joined #salt
17:03 lineman60__ joined #salt
17:05 gatoralli joined #salt
17:09 robertkeizer joined #salt
17:10 imaginarysteve joined #salt
17:11 micah_chatt joined #salt
17:11 pdayton joined #salt
17:14 carmony I think UtahDave would be proud of my state tree, I think I've got down organizing them well :P
17:19 renoirb I have to say guys.   I never had so much fun doing sysadmin than with Salt stack!
17:19 carmony haha, its awesome huh?
17:20 renoirb #random but heartfelt
17:20 renoirb only using yaml structured data and totally free to use a templating engine
17:20 renoirb :)
17:21 gcbirzan I wouldn't really say yaml is an advantage...
17:21 robawt joined #salt
17:21 KennethWilke renoirb: tis a good thing
17:21 KennethWilke clean abstraction is an advantage though
17:21 renoirb Yeah, and good sensible defaults.
17:22 gcbirzan The awesome part is you can use any renderer combination you want.
17:22 renoirb changing states with puppet… oh my :/
17:22 renoirb gcbirzan: that's what I was referring to when I said templating engine
17:23 KennethWilke i think he's thinking of the other means of data structuring as well
17:23 KennethWilke which is awesome indeed
17:23 Ryan_Lane joined #salt
17:23 Ryan_Lane joined #salt
17:24 gcbirzan Speaking of awesome things. Is the bootstrap supposed to work on OmniOS?
17:24 tempspace I'm looking to create a script that can access the IP address of every salt minion connected to a master, what would be the best way to do this?
17:25 KennethWilke tempspace: salt \* grains.get ip_interfaces
17:25 KennethWilke by default that'll return yaml
17:26 KennethWilke `salt \* --out=json grains.get ip_interfaces` can give you that in json
17:26 gcbirzan salt '*' grains.get ipv4 too
17:26 efixit joined #salt
17:26 gcbirzan Without the too :D
17:27 gcbirzan I'm asking because I see that in a ticket it's said "oh, we fixed the detection", but the functions for installing it aren't there. And I'm not very keen on using opencsw
17:27 KennethWilke gcbirzan: the things you speak of are beyond my realm
17:28 gcbirzan Yeah, I might try my hand at making an omnios package... been wanting for ages now, just never had any real incentive... or any packageI really wanted :P
17:28 KennethWilke i don
17:28 sgviking joined #salt
17:28 KennethWilke i don't know much of either but how similar is omnios to smartos?
17:29 KennethWilke the premise/goal sounds similar
17:29 gcbirzan They don't share a package manager, afaik.
17:29 gcbirzan Or, well. I never used smartos, but newrelic has a smartos thing and it gave me a  tgz :P
17:29 KennethWilke what kinda kernel does it run? i believe smartos is opensolaris based
17:29 gcbirzan Oh, yeah. They're both illumos.
17:29 Furao joined #salt
17:29 gcbirzan It's just that the prerequisite packages might be hard to find
17:30 StDiluted joined #salt
17:30 gcbirzan Not hard to find, but their names/ways of installing will differ.
17:30 KennethWilke yeah
17:32 lineman60 joined #salt
17:39 cmthornton hi, is there a way to see the output of cmd.run as it is being executed? I am importing large files and I was hoping to use pv to see the progress, but I only see the output of pv after the command has completely finished running, which takes ~10-13 minutes.
17:40 felixhummel joined #salt
17:41 KyleG1 joined #salt
17:41 JaredR joined #salt
17:43 renoirb I think a good tutorial would be about reminding how to convert command line syntax to state
17:44 renoirb e.g. salt.modules.cron.set_env(user, name, value=None)
17:44 renoirb how do I yamlify this?
17:44 renoirb never sure.
17:44 mesmer joined #salt
17:44 JaredR joined #salt
17:44 renoirb i assume user is always required, name is either the string at the root (or name: value and use a key you want), but the other one...
17:47 lineman60 joined #salt
17:49 gcbirzan renoirb: value=None means None is the default, I don't really understand what you mean?
17:49 renoirb How to translate in yml syntax described commands
17:49 Furao joined #salt
17:50 JaredR joined #salt
17:50 renoirb As said: salt '*' cron.set_env root MAILTO user@example.com
17:50 renoirb in yml would be:
17:50 renoirb keyname:
17:50 renoirb keyname:
17:50 renoirb gosh editor
17:50 renoirb wait a sec.
17:51 renoirb cron_mailto:
17:51 renoirb cron.set_env:
17:51 renoirb - user: root
17:51 renoirb - name: MAILTO
17:51 renoirb - value: list@address.org
17:51 Jahkeup joined #salt
17:51 renoirb based on :  salt.modules.cron.set_env(user, name, value=None)
17:53 mattbarto joined #salt
17:53 mattbarto left #salt
17:53 mattbarto joined #salt
17:55 JaredR joined #salt
17:57 mapu joined #salt
17:57 berto- joined #salt
17:59 KyleG joined #salt
17:59 KyleG joined #salt
17:59 felskrone joined #salt
18:00 gcbirzan renoirb: Looks correct, but you can test it to be sure?
18:00 renoirb doesn't work
18:00 renoirb gcbirzan: State cron.set_env found in sls piwik.crontab is unavailable
18:01 JaredR joined #salt
18:02 davidone joined #salt
18:03 zooz joined #salt
18:03 JaredR joined #salt
18:05 cmthornton if a state is unavailable, I think that means you are missing some dependencies to use the state--don't quote me on that though
18:05 druonysuse joined #salt
18:05 druonysuse joined #salt
18:05 alexandrel I'm using git to store pillars, and gitfs (with branches and whatnot) to store files, is there anything special I need to do to access environement specific pillars? I ask because I my minions seems to be all systematically using the "base" pillars (ie: the one stored on master)
18:05 alexandrel [...] *because my minions seem [...]
18:05 racooper joined #salt
18:07 LLckfan joined #salt
18:07 gcbirzan renoirb: Oh, that's a module. You can't use them like that.
18:08 renoirb oh gcbirzan, how can I then? I did not get the difference
18:08 alexandrel renoirb: I might be wrong, but to do that you'de need to create a state module that would rely on that salt module.
18:08 alexandrel that's how the mysql states modules are coded anyway...
18:09 renoirb I guess I would have to read http://docs.saltstack.com/ref/states/all/salt.states.module.html
18:09 renoirb probably module.run
18:09 JaredR joined #salt
18:10 mapu joined #salt
18:10 alexandrel ahhh... didn't know about module.run. Cool.
18:11 renoirb yeah
18:11 renoirb i wish i could find how to use in a state definition
18:11 renoirb ill get used. but in the meantime :/
18:11 renoirb syntax conversion between command line and state definition … i wish I could find the appropriate document that describes it
18:12 renoirb alexandrel: also this: http://intothesaltmine.org/blog/html/2013/03/01/using_the_cmd_module_in_salt_states.html   But i am still a bit confused
18:12 renoirb and i got to go now .
18:13 devinus joined #salt
18:13 avienu joined #salt
18:20 Furao joined #salt
18:28 Ryan_Lane joined #salt
18:31 UtahDave joined #salt
18:35 mapu joined #salt
18:36 ronc joined #salt
18:36 ahammond joined #salt
18:38 mmilano joined #salt
18:41 whit joined #salt
18:45 olaf39 joined #salt
18:46 josephholsten joined #salt
18:50 Furao joined #salt
18:53 devinus joined #salt
18:53 zooz joined #salt
18:55 mapu joined #salt
18:57 Khollowa_ joined #salt
18:58 terminalmage joined #salt
18:59 Kholloway joined #salt
19:00 davidone joined #salt
19:00 davidone joined #salt
19:02 druonysuse joined #salt
19:08 ahammond joined #salt
19:10 djn joehh: hello, debian upgrades from salt are still not working for me with 0.16.4-2
19:10 ahammond joined #salt
19:11 djn the process is: pkg.upgrade or cmd.run 'apt-get -y upgrade' being run
19:12 djn the minion dies then, and leaves the apt database broken, which has to be fixed by reconfiguring dpkg and force installing by apt-get -f install
19:12 jacksontj joined #salt
19:15 jbunting joined #salt
19:15 ben_ joined #salt
19:23 gatoralli joined #salt
19:26 lorimar hi all. i'm fairly new to salt. I've got a minion that I'm trying to set up a yum repo on and install packages from. When I attempt to run this though, I get a  "Name pkgrepo.managed in sls openldap_dev is not a dictionary" error. http://pastebin.com/FnCbTDZ0
19:34 zach Has anyone successfully setup pillars for installing Perl modules via cpan?
19:34 ipmb joined #salt
19:35 juanlittledevil joined #salt
19:35 mwillhite joined #salt
19:35 zach I tried this suggestion, which did not work: https://github.com/saltstack/salt/issues/2130
19:36 ni152414 joined #salt
19:36 felixhummel_ joined #salt
19:37 pipps1 joined #salt
19:40 zach Does anyone know if a CPAN state or module has been created yet
19:41 shinylasers joined #salt
19:43 zooz joined #salt
19:46 Jahkeup joined #salt
19:50 Furao joined #salt
19:51 pipps2 joined #salt
19:51 cvrebert joined #salt
19:52 pmrowla joined #salt
19:52 devinus joined #salt
19:53 cvrebert So, is the required abstract interface for package manager modules written down anywhere?
19:54 berto- joined #salt
19:54 cvrebert (and same question for service (as in upstart, etc.) modules)
19:56 UtahDave cvrebert: kind of.  Let me find the doc
19:56 UtahDave http://docs.saltstack.com/topics/development/package_providers.html
19:57 alexandrel hah, the issue I had with pillars stored in git was actually a bug... and one of my coworker will supply a patch.
19:57 cvrebert UtahDave: nice, thanks! Nothing similar for services?
19:58 bhosmer_ joined #salt
20:02 carmony ping: UtahDave, could I pm you a question dealing with some SSH stuff and pillar data?
20:03 carmony basically, for file.managed, the contents_pillar isn't working for me, its returning a null value
20:04 jbunting joined #salt
20:04 carmony and I've verified the information is available via salt 'host' pillar.items that it is being set
20:05 UtahDave cvrebert: Not really. Bascially we've been copying other existing modules.
20:05 UtahDave We really should have a design document for each of those
20:06 UtahDave hey, carmony!  Sorry, I've been helping do hiring interviews all day.
20:06 carmony UtahDave: fun! I love interviewing people :P any good looking prospects?
20:06 UtahDave Oh, yeah. We've had a lot of amazing people apply.  It's been really cool.
20:07 Jahkeup_ joined #salt
20:07 UtahDave carmony: can you pastebin what you have so far?
20:07 alexandrel an amazing project usually attracts that kind of people.
20:07 UtahDave :)
20:07 cvrebert left #salt
20:08 TheCodeAssassin joined #salt
20:08 carmony UtahDave: sending in a PM
20:09 kchr joined #salt
20:10 tuxIO joined #salt
20:10 jxqz joined #salt
20:10 carmony UtahDave: paste sent
20:11 UtahDave got it
20:13 pipps joined #salt
20:17 tomviner joined #salt
20:17 devinus joined #salt
20:20 shinylasers joined #salt
20:27 carlos_ joined #salt
20:27 Jahkeup joined #salt
20:31 ggoZ joined #salt
20:35 jdenning joined #salt
20:36 pipps joined #salt
20:38 mesmer joined #salt
20:38 jesusaurus is there a limit to how deep you can include sls files? I'm getting: [WARNING ] Duplicate Key: "include" found
20:38 gatoralli joined #salt
20:40 KennethWilke joined #salt
20:41 jesusaurus or does that warning mean that im including the same state multiple times?
20:42 shinylasers joined #salt
20:42 drags jesusaurus: error text sounds like multiple include keys on a single resource
20:43 Ryan_Lane joined #salt
20:43 Ryan_Lane joined #salt
20:44 drogoh joined #salt
20:45 whyzgeek hi I created a base for * and test for a specific machine. there is top.sls for each. It seems that salt doesn't see the test. I have put it in the master file. Is there anything else that I need to consider?
20:45 jesusaurus drags: do you mean like: `include: [ thingone, thingone ]` instead of `include: [ thingone, thingtwo ]`?
20:46 tyler-baker joined #salt
20:46 felixhummel joined #salt
20:47 drags jesusaurus: I mean it seems like: http://pastebin.com/gtD81YHL
20:47 drags where instead it should be one include with both items (some_file and some_other_pkg) listed under it
20:49 KyleG joined #salt
20:49 KyleG joined #salt
20:50 jesusaurus drags: what I think might be happening is: https://gist.github.com/jesusaurus/b7eeedab7572b2a7cdab
20:50 mattbarto joined #salt
20:50 jesusaurus But I don't think that should cause an error
20:50 whyzgeek is there anyway that I can override a sls in base by using the same name in one of the environments like test?
20:50 jesusaurus What I'm seeing though is that the resources in one.sls are being applied during highstate, but the resources defined in alpha and beta aren't :(
20:51 chuffpdx joined #salt
20:51 jesusaurus UtahDave: I need to more testing, but I think this might be a bug ^^^
20:51 jesusaurus s/to/to do
20:53 kchr joined #salt
20:53 nielsbusch joined #salt
20:57 berto- joined #salt
20:59 cetex joined #salt
21:01 alunduil joined #salt
21:13 tyler_ joined #salt
21:22 JaredR joined #salt
21:25 djn UtahDave: I'm still having trouble updating my minions: After I ran into the upgrade issue yesterday, some of my minions are on 0.16.4-1 now. While trying to upgrade those with the at.at workaround, I noticed the at.at command is not availible in minions running 0.16.4-1. Is this a known thing?
21:26 JaredR joined #salt
21:27 jacksontj joined #salt
21:28 jslatts joined #salt
21:31 juanlittledevil joined #salt
21:31 opapo joined #salt
21:32 JaredR joined #salt
21:33 felixhummel joined #salt
21:34 ragnar0k joined #salt
21:35 JaredR joined #salt
21:38 ragnar0k joined #salt
21:39 Morty_ joined #salt
21:39 micah_chatt does anyone have a good way to ensure the supervisor daemon is running?
21:40 micah_chatt I installed it via pip
21:40 micah_chatt and I want it to just run, but cmd.run going to run it on every highstate
21:41 berto- joined #salt
21:42 avienu joined #salt
21:43 JaredR joined #salt
21:44 ggoZ micah_chatt: just throw in a file for upstart, or make one for /init.d/ or whatever system you are using
21:44 forrest Stackoverflow really needs to have a tag for saltstack, as opposed to salt
21:46 JaredR joined #salt
21:49 JaredR joined #salt
21:51 micah_chatt anyone know how to make a service watch a directory for file changes rather than just watch a file?
21:53 shinylasers joined #salt
21:54 zooz joined #salt
21:55 djn how come I can not access at.at on sume of my hosts? any ideas?
21:55 djn some*
22:02 KennethWilke joined #salt
22:05 pipps joined #salt
22:07 alunduil joined #salt
22:10 KyleG1 joined #salt
22:14 oz_akan_ joined #salt
22:19 UtahDave whyzgeek: I'm not sure what you mean. What have you put in the master file?
22:20 UtahDave jesusaurus: ok, let me know how your testing turns out.
22:21 UtahDave djn: are you getting a stacktrace on the minions that won't run at.at?
22:22 ragnar0k joined #salt
22:23 ragnar0k left #salt
22:24 djn UtahDave: nope nothing, just a plain: "at.atq" is not available. for every function in at
22:25 djn but only on some... at first I though it was separate by version, but its happening on some of the upgraded ones too...
22:25 UtahDave I have no idea why the module wouldn't be working.
22:25 UtahDave what's the command you're executing?
22:26 kermit joined #salt
22:30 mapu joined #salt
22:30 djn UtahDave: just a plain sudo salt host* at.atq
22:31 UtahDave try   sudo salt 'host*' at.atq
22:32 djn UtahDave: it's still the same, even without wildcards/single or double quotes or with grainmatching
22:34 joehh djn: thanks for letting me know - i'll check it out today
22:35 joehh are you upgrading from 16.4-2 or to 16.4-2?
22:42 jesusaurus UtahDave: ugh, I found my problem. I had an "include:" inside a jinja macro... that was difficult to debug :/
22:43 djn joehh: to 0.16.4-2, which is not working because of at.at not being available... in my first message I thought 0.16.4-1 to 0.16.4-2 should work normally, my bad, disregard that
22:45 mesmer joined #salt
22:48 MTecknology joined #salt
22:50 juanlittledevil joined #salt
22:52 UtahDave jesusaurus: ah, glad you found that.
22:58 joehh yeah ok, the issue is upgrading from 0.16.3 or 0.16.4-1, if at
22:59 pipps joined #salt
23:00 joehh is not available, then applying this patch http://goo.gl/U2rF2w via file.managed is an alternative
23:01 joehh to any 0.16.3 install
23:05 Gifflen joined #salt
23:08 ben_ joined #salt
23:09 jbunting joined #salt
23:12 ben___ joined #salt
23:15 efixit joined #salt
23:20 mgw1 joined #salt
23:26 oz_akan_ joined #salt
23:27 jslatts joined #salt
23:35 rgarcia_ joined #salt
23:35 bhosmer joined #salt
23:36 NV hrm, is there a way i can trigger refresh_pillar on all nodes when git.latest updates?
23:37 p3rror joined #salt
23:37 NV an exec could work, but is there a cleaner way to do it internally within salt without executing salt as a process?
23:37 NV cmd rather
23:39 Gwayne joined #salt
23:51 oz_akan_ joined #salt
23:52 g4rlic How well does Salt handle xinetd based services using the service.running state?
23:52 g4rlic I ask because it used to work when we first set it up, but something is causing it to falter.
23:53 pipps1 joined #salt
23:53 g4rlic I get a "named service not available" error with service.running: -name: tftp
23:53 g4rlic even though chkconfig --list tftp works and shows the service is installed and enabled.
23:56 redbeard2 joined #salt

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