Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-01-28

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

All times shown according to UTC.

Time Nick Message
00:00 Nazzy fr00, my situation isn't typical... my solution though is to clone a disk image to a new drive, set ip and hostname, attach to a baremetal VM and boot it ... then get salt to do the rest
00:02 fr00 Nazzy: I can absolutely see some advantages with that. Interesting solution.
00:04 fr00 I will only have 3 different operating systems, and a small number of pre-defined software and configurations for them. Will probably build the images locally using virtualbox or similar, then just create 3 different templates using the box images. And create instances out of them. Just wondering if this will be hard to manage when/if I need to update the base images.
00:04 fr00 Guess I just have to try :)
00:05 MTecknology Nazzy: not to be 100% clueless here... but I'm 100% clueless. Is there any chance you could do a bit of hand holding for me on this one?
00:05 Nazzy fr00, I've pretty much insisted on having all the new nix systems on the projects I manage derived from a common gentoo base
00:07 fr00 Nazzy: Thank you, when I read you had me realise that I should really do the same.
00:07 Nazzy fr00, the VMs all run on an internally built xen platform, so I've got a lot of flexibility with disks
00:07 Nazzy you're welcome :)
00:07 donnex joined #salt
00:08 fr00 Also running Xen. Just switched from KVM. I'm much happier now :)
00:08 [diecast] joined #salt
00:08 ChoHag joined #salt
00:09 sarkis joined #salt
00:10 ashtonian joined #salt
00:11 cachedout joined #salt
00:14 Nazzy MTecknology, basically... if the problem is minion side: python -m cProfile -o minion_profile /path/to/salt-minion.py
00:16 ChoHag joined #salt
00:16 Nazzy that will log the stats for the minion ... then once it's misbehaved you ctrl+c the process and you've got data in the minion_profile file
00:16 n8n joined #salt
00:17 xmltok is there a release date for 2014.1? maybe during saltconf?
00:17 Nazzy http://docs.python.org/2/library/profile.html has details on turning that file in to useful profiling output
00:18 ashtonian joined #salt
00:19 Nazzy you're probably looking for something with a high tottime
00:24 andrej I'm using salt to configure nagios-nrpe, and that works fine for the two os-grains Centos & Ubuntu. I'd like to be able to (elegantly?) override the config for a few machines (3 out of 35) that deviate from the build based on a) a custom grain or b) their name. Is using jinja branching within an sls a good way of doing this?
00:24 andrej Or can/should I have some form of inheritance, where the more specific overrides the generic?
00:27 Nazzy andrej, I find shoving if/else blocks in my templates based on pillar data to be quite effective
00:28 Nazzy grains work too, I just prefer pillars for how they're defined
00:30 terminalmage joined #salt
00:30 timoguin joined #salt
00:31 rojem joined #salt
00:32 cachedout joined #salt
00:32 wking joined #salt
00:32 andrej OK Nazzy ... how does the pillar then aid in the structure/preference?
00:33 higgs001 joined #salt
00:34 mgw Nazzy: re your question on pillar data in salt-devel, is your large amount of data "private" in the sense that you wouldn't want other minions to have access?
00:34 Nazzy andrej, grains are primarily defined on the minion, pillars are primarily defined on the master... keeps it centralized, though grains can be controlled centrally but that's more work
00:35 Nazzy mgw, not especially... I have a bunch of data held in a mysql table that needs to be retrieved, formed to a sensible structure and then passed to a template to be rendered to a config file
00:36 Nazzy mgw, I was looking at extending the mysql pillar to take care of most of it in a reusable fashion, but I'm not sure it's the most appropriate place to have that code
00:36 mgw do the minions have access to mysql?
00:37 mgw also, how much is a lot?
00:38 Nazzy they could be given access.  a lot is a few thousand dicts I'd guess
00:39 jmccree I'm looking at using salt to automatically add monitoring rules to my monitoring system. Has anyone done this?
00:39 mgw Each minion needs all the dicts?
00:39 mgw Nazzy: ^
00:39 mgw that is, the entire data set?
00:39 CheKoLyN Hello all, are events working on salt-cloud 0.8.9?
00:39 jmccree The issue I'm running into is the rules need to be setup per each minion, but I'd prefer to run them all on the master.
00:40 Nazzy nope, different minions need different parts of the data set, and only one minion is likely to want any particular record
00:40 Nazzy initially it's configuring a reverse proxy, but will extend to configuring sites on shared hosting servers
00:41 __number5__ CheKoLyN: salt-cloud merged into salt now
00:42 diegows joined #salt
00:42 sarkis joined #salt
00:42 Nazzy I think I need to write some code for a serial number and rebuild it based on that... and have the actual sql to dict conversion in a python rendered .sls file
00:43 mgw Nazzy: is the single record big? If not, just use an ext_pillar module that pulls that minion's record in
00:44 CheKoLyN __number5__: do you know if that is the case for v .17.4 ?
00:45 Nazzy mgw, not especially big, but putting a thousand dicts in to pillar for every minion seems like it'll kill performance when doing highstate
00:45 mgw oh, one record is a thousand dicts?
00:46 Nazzy lets say web1 has a thousand sites ... it needs a dict for every site containing the config info for that site
00:46 mgw i see
00:46 mgw so basically you don't want to refresh that data on every highstate
00:47 Nazzy yeah, I can't see it scaling in the long term
00:47 mgw I guess you could use a custom module that you call from your templates
00:47 __number5__ CheKoLyN: not in 0.17.4 but in 2014.01
00:47 CheKoLyN __number5__: i think the Ubuntu PPA stil has the salt-cloud package; perhaps in the next release
00:47 mgw Nazzy: the custom module could handle local caching however you wish
00:48 Nazzy *nod* it's crying out for caching ... hash the file, compare against the last good hash, rebuild if different ... clear hash when the data changes
00:49 mgw so you're storing the data in mysql as json?
00:50 mgw Nazzy: and therefore can hash the whole data set easily?
00:50 Nazzy actually I'm planning to store it in a more complex form and transform in code ... a nice relational schema probably
00:50 mgw how will you detect a change then?
00:51 Nazzy clear the hash
00:52 mgw well, I think a custom module will work for you. Just use {{ salt['custommodule.get']('some-key') }} in your templates, and let the module do all the work
00:52 Nazzy (to be clear, the process that updates the data clears the hash after doing so, in transaction to maintain atomic update)
00:52 Nazzy *nod* makes sense
00:52 hunter Nazzy: like a DNS serial?
00:52 Nazzy hunter, exactly
00:53 hunter that can work.
00:53 ChoHag joined #salt
00:53 NV mgw: using a custom module templated isnt going to make it any faster if the custommodule.get just does the same thing as the pillar data would
00:53 hunter Im lazy and hoping that someone gets around to writing an external pillar module for etcd before I have to
00:54 Nazzy NV, right, but it's probably easier to add caching at that point
00:54 NV not really
00:54 NV its just as easy to do it with ext_pillar or using a python rendered pillar sls
00:54 NV (the latter of which is my personal favourite)
00:54 NV the code is literally exactly the same
00:54 hal1300 joined #salt
00:54 NV the difference is a module has to run on the end node
00:55 NV so your db pass/etc would be exposed to the individual servers
00:55 NV doing it as an ext_pillar or python rendered sls you're not exposing everything to the minions
00:55 mgw NV: that's definitely true, but he's trying fo avoid transferring all the data on every highstate
00:56 UtahDave bretep: still around?
00:56 NV does salt re-transfer all pillar data every highstate if it hasn't updated?
00:56 bretep so yes
00:56 bretep yes
00:56 NV or does it do something smart like hash it and compare if there's any differences and if not don't send it?
00:57 mgw NV: I believe it does
00:57 mgw retransfer that is
00:57 NV hmm
00:57 NV sounds like that should dbe fixed then
00:57 bretep I was actaully about to post. Is this possible? I'd like to subfilter (target) minions using the top.sls to make managing different environments easier. If not, then I'll look at what it will take to add it and see if it's a welcomed addition or not. https://gist.github.com/bretep/8660494
00:58 UtahDave bretep: I've been in some meetings. I was just looking at your top.sls you pastebinned.
00:58 mgw NV: It shouldn't be too hard to do, somebody should open a ticket on it
00:59 mgw all the ext_pillars would still have to be run and pillar compiled
00:59 UtahDave bretep: since you're matching on the exact same things in each environment, it will match first on dev and pull the file from there
00:59 NV bretep: surely using a different branch would be more appropriate for dev vs staging? then use environment instead
00:59 mgw although individual ext_pillars could, of course, do their own caching
00:59 bretep We are using different branches and gitfs
00:59 fllr joined #salt
01:00 Corey Evening.
01:00 bretep NV: http://docs.saltstack.com/topics/tutorials/gitfs.html
01:00 mgw Corey: evening
01:00 mgw You in SLC yet?
01:00 Corey mgw: I am indeed.
01:00 Corey mgw: Are you attending?
01:00 mgw yep
01:00 mgw I'll be landing in an hour
01:00 Corey Nice, you get here yet?
01:00 Corey Ah, glorious.
01:00 Corey I'm at the Hilton, like a civilized person.
01:00 MTecknology Nazzy: I did python -m cProfile -o minion_profile /usr/bin/salt-call state.highstate ;; it seems like I got to miss out on the 12min wait for salt-call to even start...
01:00 bretep UtahDave: that's why I proposed a subfilter to Node Groups https://gist.github.com/bretep/8660494
01:01 mgw btw, anybody else flying in right now that wants to share a cab?
01:01 mgw Corey: how far is it from the Marriott?
01:01 Corey mgw: Walking distance. Took me 5-7 minutes.
01:03 MTecknology Corey: anything wrong with the mariott?
01:03 Corey Eh, not really. :-)
01:03 Corey I'm Hilton Diamond, and Marriott Peasant. It wasn't a hard decision in my case.
01:03 MTecknology oh
01:03 Nazzy MTecknology, hmm ... it didn't run properly?
01:04 MTecknology wowzers! profiling an application makes it run amazingly slower
01:04 bretep UtahDave: I passed the subfilters by my colleges here and they think it would be a fantastic addtion. Opinions?
01:04 Nazzy yes, yes it does :p
01:04 Corey MTecknology: Yes, it does that.
01:04 Nazzy it can be an order of magnitude or more slower
01:05 MTecknology Nazzy: nah- it just seemed to skip that 15min start. I just had to wait until it hit a user state. I'm watching it with strace. It just got to that spot.
01:05 MTecknology ^c to stop it?
01:05 UtahDave bretep: it's an interesting idea!
01:05 MTecknology I'm guessing that if it were to finish now, I'd be looking at an 15+hr highstate execution time
01:07 yomilk joined #salt
01:07 Nazzy yeah, it's probably run far enough
01:07 bretep UtahDave: It sure would take a lot of redundant targeting logic out and makes for instance the dev.sls usable for any group filter.
01:08 MTecknology Nazzy: almost funny... it seems to have actually given up trying to get users/groups
01:08 MTecknology it made it about 5 names in and just started speeding along. I cancelled it. I have my profile now
01:11 oz_akan_ joined #salt
01:11 Nazzy MTecknology, figures ... these things tend to go in to hiding when you look for them ¬_¬
01:12 MTecknology I still saw it happen, just not as long
01:12 MTecknology http://dpaste.com/1572982/  <-- most useful thing I mad so far
01:13 MTecknology A whole 4 calls brought grp.getgrall to the top
01:15 MTecknology Nazzy: so... I imagine there's some magic in there that lets me dive into that function?
01:15 pydanny joined #salt
01:16 DaveQB joined #salt
01:17 Nazzy that's weird... I didn't get and userlike lookups from grp.getgrall
01:17 MTecknology I can try to do something else if you think I should
01:19 Nazzy possibly dumb question, is that trace definitely enumerating users or is it doing groups?
01:19 Dale_ joined #salt
01:19 MTecknology strace is showing me groups; users don't have their own groups
01:19 MTecknology strace is showing me users; users don't have their own groups  **
01:20 hal1300 left #salt
01:21 MTecknology p.print_stats('grp.getgrall') <-- only gave me that one line
01:22 Nazzy and takes an age?
01:22 MTecknology hm?
01:23 MTecknology ncalls  tottime  percall  cumtime  percall filename:lineno(function)
01:23 MTecknology 4  290.084   72.521  290.084   72.521 {grp.getgrall}
01:24 MTecknology That only made 4 calls, and my strace confirmed that. Normally it'll run around 12,000 times (not really sure)
01:24 MTecknology now that you stepped me through the profiler bit, it's actually pretty easy and straight forward
01:25 oz_akan_ joined #salt
01:28 MTecknology Nazzy: I'm going to go home for the day. I'm just about at 12hr for the day and I don't want another 60+ hour work week. They're getting irritating. I'm going to start trying to build an AD server at home so I can try to reproduce things like this there.
01:28 Nazzy MTecknology, sadly, the python grp module is actually a C shared object... makes it a little harder to guess what it does
01:29 Nazzy given that it is 1.30am, home sounds like a good idea lol
01:29 MTecknology ya.... it might be time to source dive that bad boy
01:29 MTecknology that's going to be a royal pain in the ass... but at least I can see some system calls being made and should be able to eventually track that part down
01:30 MTecknology maybe... I'm not a coder
01:30 MTecknology oh... forgot I need to do other things before going home
01:30 MTecknology instead of waiting 3hr for salt 'boothost*' state.highstate to finish, I'll log into each box... :P
01:33 ChoHag joined #salt
01:36 cachedout joined #salt
01:38 jesusaurus is there a way to force salt to output color even if it's being piped to less?
01:40 elfixit joined #salt
01:40 jesusaurus ahh... --force-color
01:40 jesusaurus that just makes sense...
01:43 rojem joined #salt
01:46 mgw joined #salt
01:50 UtahDave jesusaurus: Salt reads your mind... IN THE FUTURE!!
01:51 cachedout_home joined #salt
02:00 n8n joined #salt
02:01 Ryan_Lane21 joined #salt
02:03 hey_guys joined #salt
02:04 mgw joined #salt
02:14 jimallman joined #salt
02:18 jankowiak joined #salt
02:20 fllr joined #salt
02:27 puppet joined #salt
02:34 puppet joined #salt
02:38 pass_by_value joined #salt
02:39 NotreDev joined #salt
02:40 ConceitedCode joined #salt
02:42 toofer joined #salt
02:44 Dale_ joined #salt
02:44 alice|wl left #salt
02:44 jankowiak joined #salt
02:45 mgw joined #salt
02:49 MindDrive Hmm... is it considered a bad thing for a Salt master to also be a Salt minion?  More specifically, is this a bad situation where one has a multi-master situation where each master is also trying to be a minion?
02:49 cachedout_home joined #salt
02:50 oz_akan_ joined #salt
02:51 n8n joined #salt
02:54 MTecknology Any master I have is a minion. Most of the states I have I want hitting minions too.
02:57 MindDrive We're just setting up Salt here for the first time.  Test setup is two masters in multi-master mode.  Get them both up, start running 'sudo salt '*' test.ping' on them and... strange things happen.  Primarily long pauses/timeouts and lack of responses or the response being partial (only seeing one master respond instead of both).
02:59 MindDrive Tried just pointing each master's minion configuration to only the other master, and that seems to work consistently, but it also takes a while for the test.ping command to finish.
02:59 UtahDave MindDrive: what version of Salt?
02:59 MindDrive The very latest and greatest.
02:59 MindDrive (0.17.5)
03:00 MindDrive Tried turning up the logging, but that didn't seem to have much affect, sadly.
03:00 MindDrive There may be other things not configured properly in the master/minion configuration files, but it's hard to tell given the sheer size of them.
03:01 UtahDave MindDrive: I'd recommend just using one master until you get the hang of things.
03:02 MindDrive UtahDave: I wouldn't mind that, but others here are very insistent on HA, since this will become a critical piece of our infrastructure.
03:03 MindDrive (Replacing MCollective which has been too difficult to manage and get working correctly.)
03:03 MindDrive Single failure points tend to be frowned upon. :)
03:03 oz_akan__ joined #salt
03:04 UtahDave Sure.
03:04 UtahDave MindDrive: I've got a couple minutes before I have to go.  SaltConf starts early tomorrow!  :)
03:04 UtahDave so you have 2 masters running minions as well, right?
03:05 UtahDave what is the "master:  " config option set to in each minion config?
03:05 MindDrive Originally it was set to both masters.
03:05 MindDrive That seemed to make things very unhappy.
03:08 MindDrive The response times are inconsistent... sometimes one server will return results immediately, other times it will return the results then spend 5-10 seconds before it completes, and other times it will just spent 5-10 seconds and return... nothing.
03:08 UtahDave Hm. That's odd.  I haven't tested multi master like that yet.
03:09 MindDrive It may very well be this is "don't do that!".  But we'll need to set up some separate minions to see if that's really the case.
03:09 UtahDave Yeah. that's a good idea.  I'm not sure what's going on in that scenario.
03:09 MTecknology It may be a use case that hasn't been tested or considered yet
03:09 UtahDave MindDrive: would you mind opening up an issue on that?
03:10 neilf_ Hi all, When will 0.17.5 be on apt for ubuntu?
03:10 MTecknology don't people usually set one minion to point at one master only?
03:10 MindDrive UtahDave: Sure...
03:11 MindDrive MTecknology: http://docs.saltstack.com/topics/tutorials/multimaster.html shows that usually all minions are pointed to all the masters...
03:12 joehh neilf_: should be there already - are you using the ppa?
03:12 MTecknology MindDrive: yup- I'm 100% wrong
03:13 MindDrive Okay, I think I'm blind, but I can't find on the website where to place an issue for this. :)
03:13 hummus3 joined #salt
03:15 MTecknology MindDrive: github.com/saltstack/salt
03:17 MindDrive D'oh... I forgot GitHub has its own Issues setup... thanks.
03:21 xmltok joined #salt
03:22 munderwo joined #salt
03:27 MindDrive Issue created, and thanks for the help, folks.  I need to scoot, but will keep an eye out over the next day or two to see if anyone has any further suggestions/help. :)
03:32 dlam joined #salt
03:41 DanGarthwaite joined #salt
03:42 mgw2 joined #salt
03:46 n8n joined #salt
03:49 pass_by_value joined #salt
04:00 darless_ Is there a way in jinja or another renderer to check if a pillar sls exists?
04:01 oz_akan_ joined #salt
04:01 Furao darless_: if a .sls file exists?
04:02 darless_ Furao: yeah, within the pillar data
04:02 darless_ In a specific case I only want to include pillar data if it exists
04:02 ubuntu__ joined #salt
04:03 Furao I don't think so, pillars are rendered by the master, module are executed on the minion
04:03 Furao even in pillars
04:03 Furao so, if you have a module that let the master look for .sls files, pillars files might not be readable by the minion
04:07 cachedout_home joined #salt
04:07 darless_ Furao: If the pillar data is rendered on the master, the master should be able to know whether a particular pillar file exists. In my example my top.sls (pillar) is jinja templated. I want to have a jinja check that sees whether a pillar file exists
04:08 darless_ pillar file == /srv/pillar/settings.sls for example
04:08 mgw joined #salt
04:09 UtahDave darless_: you should be able to use jinja or any salt module to test if the pillar file exists
04:09 UtahDave Furao: Hey, how are you doing, my man?
04:10 bretep @all if you have time tonight, please read and comment on this issue: https://github.com/saltstack/salt/issues/9979
04:11 darless_ UtahDave: k, thanks. Just didn't know the jinja related call to check this. google time
04:11 bretep You may all beat me up if you like...
04:11 bretep wait you are all in town, right? Crap...
04:11 MTecknology UtahDave: aren't you sleeping already?
04:11 bretep I meant you can beat me up metaphorically
04:12 UtahDave bretep: lol
04:13 UtahDave MTecknology: I should be.
04:14 Eugene Did terminalmage ever make it out of Denver?
04:15 UtahDave Eugene: Yeah, he did.  He arrived around midnight or something.
04:15 Eugene Poor git
04:17 munderwo Hi all. I want to script the creation and acceptance of keys on the salt master. I had a look at the python client api, but wasn't sure how they map to things like salt-key --gen etc. is there some documentation on this? or some kind of mapping?
04:17 bretep I guess that issue didn't change anyones life. *sigh*
04:29 ravibhure joined #salt
04:32 Voziv joined #salt
04:32 _fllr_ joined #salt
04:32 Furao UtahDave: hello, I'm doing great. moved back to Malaysia :) and you?
04:34 scoates joined #salt
04:35 UtahDave munderwo: http://docs.saltstack.com/ref/clients/index.html?highlight=client#wheelclient
04:35 UtahDave Furao: cool!  I hope you enjoy it.
04:36 UtahDave Furao: I'm heading home now.  SaltConf starts in the morning!
04:36 munderwo UtahDave: ahh I found that, but I'm a little light on the details about what strings map to what in the call() func
04:36 Furao I wish that you did it in 6 months i would have been there :)
04:36 akoumjian joined #salt
04:37 Furao it's > 85 degree now here and with the pool, I don't feel like winter :P
04:37 UtahDave lol.  Maybe next time, Furao!
04:39 Furao UtahDave: how can master use module in pillars to figure if a .sls exists on the master?
04:40 Furao with cp.list_master ?
04:41 UtahDave the modules work from the master's perspective
04:42 UtahDave salt['cmd.run']('ifconfig')   will give you the master's network info
04:43 UtahDave catch you all tomorrow!
04:44 cwright Anyone from saltstack still here?  I sent an email to saltconf@ earlier and haven't heard back.
04:45 Furao cwright: UtahDave, but he just left
04:46 cwright yea i saw i just missed him
04:50 mgw cwright: basepi is showing as being here
04:51 higgs001 joined #salt
04:52 gnu-man joined #salt
04:52 basepi cwright: I have no power over sched, so I'm of no help.
04:52 basepi Sorry. =\
04:52 basepi But know that we did get it!
04:53 basepi And I'm always "here" I just won't always answer immediately.
04:53 cwright basepi: who should i talk to about that?
04:53 basepi Not sure, I think Unbridled Communications set up that part (they're the company who helped us organize the conference)
04:53 basepi So probably the registration desk tomorrow?
04:54 mgw joined #salt
04:54 basepi The schedule is all up at saltconf.com as well.
04:54 cwright basepi: ah ok.  yea, i saw the schedule, some people were saying that they were showing as "limited availability" so I wanted to confirm if i could
04:56 basepi Ya, you're fine. All the classes just have limited seating, you can't reserve a seat through sched or anything.
04:57 baniir joined #salt
04:58 cwright basepi: ah ok thanks, I appreciate it
04:59 basepi No prob. =)
04:59 possibilities joined #salt
04:59 baniir is salt usable on freebsd 10
05:00 rojem joined #salt
05:00 pdayton joined #salt
05:02 oz_akan_ joined #salt
05:03 bemehow joined #salt
05:07 xmj baniir: of course
05:07 xmj baniir: cedwards is doing a great job keeping salt up to scratch on any of the supported versions
05:08 xmltok joined #salt
05:10 baniir i see salt in ports; awesome. what's the preferred method for install? it looks like salt-bootstrap supports up to version 9.1
05:11 xmj baniir: preferred method how?
05:12 baniir xmj: ports vs salt-bootstrap
05:12 xmj baniir: installing from ports should give you 0.17.4; installing from packages *should* do the same
05:13 xmj baniir: if by salt-bootstrap you mean to follow instructions a la http://docs.saltstack.com/topics/tutorials/salt_bootstrap.html
05:14 xmj baniir: maybe that's just the natural paranoia that comes with maintaining a bunch of FreeBSD ports, but I'd never curl and pipe it to sh. ever.
05:14 MTecknology I have ten servers at home. This is almost starting to feel like a lot for just at home.
05:16 baniir xmj: understandable; i installed via ports earlier but am getting "no top file or external nodes data matches found" when trying to highstate with master
05:16 xmj baniir: well, do you have a top.sls in the default location?
05:18 xmj baniir: I have an example of working salt configs for freebsd @ https://github.com/xmj/jailedbees
05:26 baniir joined #salt
05:27 baniir xmj: i believe my top.sls is in the default location, though i'm realizing now it could differ from what i'm used to on linux
05:27 xmj baniir: isn't /srv/salt/ default?
05:27 xmj baniir: give me a minute, I have a salt-master config lying around somewhere that works with 11-CURRENT I think
05:28 baniir xmj: i agree; i think /srv/salt is the default
05:28 baniir xmj: ok, thanks
05:29 ravibhure joined #salt
05:29 pass_by_value joined #salt
05:29 xmj .oO i need to find the right bhyve zvol first, hrm
05:31 xmj baniir: check https://dpaste.de/rdPi/raw
05:33 baniir xmj: i must sign off now. i will give this a shot. thank you
05:33 xmj baniir: enjoy freebsd 10, spread the word
05:34 baniir xmj: will do
05:40 anuvrat joined #salt
05:40 Ryan_Lane joined #salt
05:51 linjan joined #salt
06:02 oz_akan_ joined #salt
06:04 kolaman joined #salt
06:05 kolaman Hi all, we are currently testing saltstack for our automation needs. what I have noticed is that multiple times salt-master stops communicating with salt-minions. while minion is running properly and I need to restarted salt-minion service on the minion and communicaation starts working agin ..
06:05 kolaman is this a known bug
06:12 Ahlee the wifi in the hotel is pretty bad
06:12 Psyke joined #salt
06:13 jalbretsen joined #salt
06:21 kedo39 joined #salt
06:32 Corey Ahlee: Which hotel?
06:34 Ahlee marriott
06:35 Ahlee smae hotel as the saltconf
06:35 Corey Ahlee: Well there's your problem. :-)
06:36 Corey I'm half a mile away at the Hilton. Such internets. Very bandwidth.
06:37 sarkis joined #salt
06:37 Ahlee heh
06:37 Ahlee i'm bored.
06:38 xmj Corey: where does this such x / very y meme come from?
06:38 Corey xmj: Doge
06:38 xmj ...and why does it make me want to kill all its users?
06:38 xmj Ah, that's why
06:39 Ahlee beers gone, gonna go wander out and explore the neighborhood
06:39 Corey Ahlee: Unfortunately it's Utah.
06:39 Corey The beers on tap are sad.
06:39 Corey Ahlee: Find a bar worth hitting at this time of night? I'd be game.
06:39 Ahlee hrm
06:40 xmj 08:40AM.
06:42 Ahlee apparently redrock brewery closed at a11
06:42 Ahlee wtf.
06:42 Corey Such Mormon. Very puritan.    no drinking
06:42 xmj please die.
06:42 Corey xmj: Working on it.
06:43 Ahlee lol
06:43 Corey xmj: www.dogeweather.com
06:44 xmj plonk
06:44 Furao very suffer?
06:44 Furao ah it's just random
06:45 Corey Such sad.
06:46 Furao "wow broken clouds"
06:48 Ahlee well, giving up on tonight, see some of you on the morrow
06:54 ndrei joined #salt
06:55 Corey It happens.
06:58 leonardinius joined #salt
07:00 oz_akan_ joined #salt
07:03 jcristau joined #salt
07:04 oz_akan__ joined #salt
07:05 Shenril joined #salt
07:18 anuvrat joined #salt
07:20 Furao Corey: puritan? no drinking. it can't be worse than here :P
07:21 Furao last friday I returned here with my an other couple and they got arrested at the airport because they had a dildo in their luggage
07:21 Furao they have 30 days to leave the country to take it back
07:21 Furao at least no other consequence
07:23 fllr joined #salt
07:24 fllr joined #salt
07:28 matanya joined #salt
07:30 middleman_ joined #salt
07:35 ckao joined #salt
07:37 joehillen joined #salt
07:41 crane joined #salt
07:53 jtang1 joined #salt
07:57 ndrei joined #salt
08:01 kalloc joined #salt
08:02 jcristau joined #salt
08:04 njpatel joined #salt
08:05 harobed joined #salt
08:11 zooz joined #salt
08:15 jcsp1 joined #salt
08:15 jakubek joined #salt
08:15 gurubert1 joined #salt
08:15 aleszoulek joined #salt
08:18 harkx joined #salt
08:18 kalloc_ joined #salt
08:20 doki_pen joined #salt
08:21 wunki joined #salt
08:21 mikkn joined #salt
08:22 quickdry21 joined #salt
08:22 kedo39 joined #salt
08:23 yomilk joined #salt
08:23 fllr joined #salt
08:24 tseNkiN joined #salt
08:25 slav0nic joined #salt
08:25 slav0nic joined #salt
08:28 giantlock_ joined #salt
08:28 bbinet joined #salt
08:34 balltongu joined #salt
08:39 cowyn joined #salt
08:50 gnu-man joined #salt
08:52 Ryan_Lane joined #salt
08:54 tinuva joined #salt
08:54 tomspur joined #salt
09:05 oz_akan_ joined #salt
09:06 pdayton joined #salt
09:09 strgcloud joined #salt
09:23 fllr joined #salt
09:24 n8n joined #salt
09:31 lessismore joined #salt
09:36 gammalget joined #salt
09:48 jtang1 joined #salt
09:53 travisfischer joined #salt
09:59 che-arne joined #salt
10:03 [pyro] joined #salt
10:03 [pyro] hi all
10:06 oz_akan_ joined #salt
10:09 yomilk joined #salt
10:10 LucasCozy joined #salt
10:16 lessismore Hello, I am going to upgrade the salt-minion from debian backports to saltstack provided packages.  Is possible to add the saltstack repo using salt?  I found the instructions to do it manually but I am not able to found the keyid and keyserver as of http://docs.saltstack.com/ref/states/all/salt.states.pkgrepo.html#module-salt.states.pkgrepo for http://debian.saltstack.com/debian?
10:16 ndrei joined #salt
10:17 bbinet joined #salt
10:21 mikkn Anyone using reclass for top file management instead of the regular top file structure? Is it worth using?
10:22 madduck mikkn: of course! ;)
10:23 madduck mikkn: the real strength of reclass is in pillar management, though. But it does top data just fine too.
10:23 fllr joined #salt
10:24 bbinet joined #salt
10:25 mikkn madduck: But what kind of usecases is it "necessary" or at least much easier to accomplish what you want with reclass? Seems to be a lot of documentation to comb before reaching that understanding
10:28 mikkn madduck: I do find variable merging and inheritance a bit limiting in salt (sometimes I would've wanted to extend the same template variables more than once or create a more dynamic trees of data in pillar) but it hasn't been a real obstacle thus far. I'm still rather new though, so I might run in to a situation where it is a proper obstacle. But I'm not sure exactly how it solves what problems. :)
10:28 madduck mikkn: for pure top data management, there is no real benefit, I think, unless you are using multiple tools (e.g. salt and ansible) and want to use the same database for them. When it comes to pillar data, however, the benefits are huge.
10:28 madduck see here for a start
10:28 madduck https://groups.google.com/forum/#!searchin/salt-users/reclass/salt-users/FKqhVx0MZ-s/EXQNuxHIbnEJ
10:29 madduck also if docs are not your thing: https://www.youtube.com/watch?v=CGPnbZ3KHBE
10:30 mikkn maddock: I like docs, but there wasn't a good "Why use Reclass?", which can be hugely helpful. :)
10:30 madduck duly noted. Unfortunately, I am no longer paid for developing reclass, so I am a bit behind
10:31 madduck but the docs do need a bit of an overwork
10:31 madduck http://reclass.pantsfullofunix.net/concepts.html is a bit of what you want though, no"?
10:32 mikkn Yeah, but it's a bit too esoteric still
10:32 helminthe joined #salt
10:34 mikkn I think having a page showing pretty much like the mail-list post you linked a proper "This is horrible in X!" "This is what you do in Reclass!" would be very helpful. :)
10:36 mikkn Or just "This is why I made reclass". :)
10:36 bbinet joined #salt
10:36 madduck how about https://groups.google.com/forum/#!searchin/salt-users/reclass/salt-users/FKqhVx0MZ-s/EXQNuxHIbnEJ ?
10:37 mikkn madduck: But thanks for taking time to answer my questions, I will look into it further. I'm right now trying to figure out the best way to do build deploy targeting and I wasn't sure Salt had a good mechanism for doing that. :)
10:37 mikkn Yeah, that's a pretty good example. :)
10:37 madduck mikkn: if you end up writing something, maybe we can add it to the docs ;)
10:38 madduck and i am available for questons anytime, here or in #reclass
10:41 yomilk joined #salt
10:42 jtang1 joined #salt
10:43 bbinet joined #salt
10:45 bbinet joined #salt
10:50 lessismore worked well using - keyid: F2AE6AB9 - keyserver: keyserver.ubuntu.com
10:53 jtang1 joined #salt
10:56 david_a joined #salt
10:56 mikkn madduck: Have you given it any more thought on how to do node -> class mappings? I saw the comment about it not being the final solution. Because most of our nodes wouldn't need a strict 1:1 mapping between node and data, so it would be wasteful to define each node in reclass
10:58 madduck mikkn: I think I will factor out the mappings into a yaml config file of its own and be done with it.
10:58 madduck so the config file would specify the path to the mapping file
10:58 madduck and you can migrate the mappings there later, easily.
10:59 madduck i might not choose yaml though due to escaping
10:59 madduck I might just do whitespace separated, 2-column
11:03 [pyro] joined #salt
11:06 bbinet /save
11:23 fllr joined #salt
11:25 _fllr_ joined #salt
11:28 diegows joined #salt
11:43 dangra joined #salt
11:44 babilen I am trying to install lighttpd on Debian boxes and would like to enable certain configurations. The scheme on Debian is to have links from /etc/lighttpd/conf-enabled to /etc/lighttpd/conf-available and to enable/disable them with a call to lighttpd-{disable,enable}-mod respectively. I know that I could set the symbolic link, but it might be that the tool is doing more.
11:44 babilen How can I approach this idiomatically? (pretty new to salt, so please forgive me any oversights)
11:44 ndrei joined #salt
11:46 babilen (similarly for Apache with a2ensite/a2dissite)
11:48 jtang1 joined #salt
11:48 babilen Is this tricky to do as there is no module for lighttpd like there is for apache2 (cf. http://docs.saltstack.com/ref/modules/all/salt.modules.apache.html#module-salt.modules.apache ) ?
11:58 londo joined #salt
12:01 beardo__ joined #salt
12:11 vipul joined #salt
12:20 nkuttler babilen: https://salt.readthedocs.org/en/latest/ref/states/all/salt.states.cmd.html ?
12:23 fllr joined #salt
12:28 blee joined #salt
12:31 NotreDev joined #salt
12:34 puppet joined #salt
12:35 ndrei joined #salt
12:41 kalloc joined #salt
12:42 jtang1 joined #salt
12:44 DaveQB joined #salt
12:46 jankowiak joined #salt
12:46 miguitas joined #salt
12:56 kalloc joined #salt
12:58 raghav joined #salt
13:01 ConceitedCode1 joined #salt
13:04 Koma-AFK left #salt
13:06 taion809 joined #salt
13:07 ndrei joined #salt
13:09 balboah joined #salt
13:21 Shenril joined #salt
13:22 ndrei joined #salt
13:23 fllr joined #salt
13:24 blee joined #salt
13:26 mikkn babilen: The only thing that the lighty-enable-mod tool does is create the symlink and tell you to reload the server, so you could as well create the symlink yourself and watch the symlinks from the lighttpd service.
13:27 mikkn Does anyone know if you have parameters/roles from Range if you set up a Range-server connection or is it only for the cluster information used in matching?
13:27 funzo joined #salt
13:28 ekristen joined #salt
13:30 timoguin joined #salt
13:36 funzo joined #salt
13:36 jtang1 joined #salt
13:38 jtang2 joined #salt
13:41 blee_ joined #salt
13:42 anuvrat joined #salt
13:44 blee joined #salt
13:45 jslatts joined #salt
13:45 BRYANT__ joined #salt
13:49 ndrei joined #salt
13:55 Mua joined #salt
13:58 Guest94285 joined #salt
14:00 ccase joined #salt
14:02 lessismore I see a lot "Minion did not return". Usually the execution takes 12 seconds.  I added the option "-t 60" and now output of the run is correcly displayed every time.  I can live with that, but maybe it is an indication of problems.  Any suggestion?
14:03 Furao lessismore: this is normal, what if you run a job that take hours?
14:04 Furao and what if you run it on 2000 VMs that can take up to 10 hours to run the job, you want /usr/bin/salt to ends when all VMs have executed this? if yes, just fill a -t $longtime
14:06 tessellare joined #salt
14:11 bretep @all have fun today!
14:13 prooty joined #salt
14:14 brianhicks joined #salt
14:18 juicer2 joined #salt
14:19 rojem joined #salt
14:20 quickdry21_ joined #salt
14:20 aj_ joined #salt
14:23 fllr joined #salt
14:28 blee joined #salt
14:29 blee hey everyone, on archlinux i have confirmed my /etc/hostname, /etc/hosts, and uname -a show the correct hostname
14:29 jcsp joined #salt
14:29 blee but whenever i try to conect to my master, it shows the key as localhost.localdomain
14:29 AdamSewell joined #salt
14:30 msciciel_ hi, i have a problem with salt. During initial highstate on minion sync_modules is executed after mine_functions which use my custom module and initial mine.update is failing. Any idea how to solve this problem ?
14:31 msciciel_ i'm now using reactor to force mine.update during minion start but i think this is workaround
14:32 gadams999 joined #salt
14:32 jtang1 joined #salt
14:33 vejdmn joined #salt
14:33 rlarkin joined #salt
14:33 gadams9991 joined #salt
14:36 fllr joined #salt
14:38 _fllr_ joined #salt
14:39 jeremyBass1 joined #salt
14:42 mgw joined #salt
14:43 MK_FG joined #salt
14:44 __number5__ joined #salt
14:45 leonardinius_ joined #salt
14:45 leonardinius joined #salt
14:47 imaginarysteve joined #salt
14:49 leonardinius joined #salt
14:50 sgflt joined #salt
14:54 cnelsonsic joined #salt
14:56 lessismore Furao: Tnx, but I can not undertand why you call it normal.  The run always take around 12, sometimes 13 sec, but sometimes it return the result and sometimes the "did not respond".  I am missing something obiouvs? Eg. the default timeout is 12 secs?
14:56 Iwirada joined #salt
14:57 ipmb joined #salt
14:58 prooty how can i tell salt-master and salt-minion to reopen their log files?
14:58 prooty send a sighup?
14:58 GradysGhost joined #salt
15:00 MK_FG joined #salt
15:03 jtang1 joined #salt
15:04 viq http://pbot.rmdir.de/c9aeyY8fjieA3rAWkU8X9g - any ideas what's the issue here?
15:05 GradysGhost Hey viq. Message like this can be kinda difficult to trace down, in my experience
15:05 GradysGhost Is this minion output, and if so, what version? 0.17.4?
15:06 viq http://pbot.rmdir.de/gc5L5sGkTA01OVcn_HBKkg
15:07 viq That's what I get from "salt gitlab.example.com state.sls gitlab"
15:07 viq centos, 0.17.4
15:07 viq master is debian 0.17.5
15:09 viq http://pbot.rmdir.de/gc5L5sGkTA01OVcn_HBKkg has the relevant bit of state and pillar for it
15:10 GradysGhost I think your jinja might be goofy
15:11 GradysGhost Assuming that the pillar's top file is distributing this pillar to the right server, your fields should look something like
15:11 GradysGhost name: {{ pillar['gitlab']['db_user'] }}
15:12 viq http://docs.saltstack.com/topics/pillar/index.html#pillar-get-function
15:12 ch0psti51 joined #salt
15:13 jeremyBass1 hello, quick question, I see in a top file we can do ` 'nag1* or G@role:monitoring': ` so am trying to if i can do  ` 'env:WEB and env:DB': ` so that both envrionments had to be there in the grain?
15:13 jeremyBass1 ` env: DB `*
15:14 GradysGhost viq: That docu doesn't say whether the second 'default' field is required or not.
15:14 GradysGhost But your template doesn't provide defaults for anything.
15:14 viq GradysGhost: it isn't, I've been using it in a few places
15:14 GradysGhost That syntax is new to me, though.
15:14 GradysGhost So I'm at a loss if you are.
15:14 druonysus joined #salt
15:15 GradysGhost If you're not providing a default, what's the harm in using the normal dictionary syntax?
15:15 mpanetta joined #salt
15:15 viq GradysGhost: consistency
15:15 viq Also easier access to nested pillars
15:16 travisfischer joined #salt
15:16 GradysGhost Well, looking at the code indicated in the stacktrace, that's a function definition line.
15:16 viq so instead of pillar['gitlab']['ldap']['uid'] I have salt['pillar.get']('gitlab:ldap:uid')
15:17 GradysGhost It's acting like you're passing in a dictionary instead of a list/string or something
15:17 GradysGhost Are there any special characters in your password that might get misinterpreted by YAML or something?
15:18 timoguin jeremyBass1, WEB and DB would be better defined as a grain than as an environment. I set a grain like this in my minion configs: 'grains: roles: - web'
15:18 ajennings joined #salt
15:18 timoguin rolls is a list, so i can just add db if i want db states to get applied
15:18 timoguin *roles
15:19 viq GradysGhost: no, pure long random text, not even number not to mention special characters
15:19 GradysGhost timoguin: viq asks about this state/pillar/output, and I'm unfamiliar with this syntax, and haven't been able to help much. You got any ideas? http://pbot.rmdir.de/gc5L5sGkTA01OVcn_HBKkg
15:19 viq though, I guess I can enclose the stuff in quotes in pillar, see if it changes anything
15:20 GradysGhost It wouldn't hurt, but it really looks like your data is fine.
15:21 jeremyBass1 well ok on the grain but I guess the question is can the pattern in the top file be ` and ` .. i saw that you can use ` or ` . and CSV , so figured ` and ` would be the both must be true operator
15:21 GradysGhost Can you run "salt 'target*' pillar.item gitlab" and get valid output?
15:23 timoguin GradysGhost, it does look fine. I would try quotes on the pillar data. I'm wondering if it has anything to do with underscores used in the key names.
15:23 timoguin I'm still learning YAML, so lots of trial and error still sometimes
15:23 viq GradysGhost: think so, here's the beginning of output (slightly redacted): http://pbot.rmdir.de/R435frN8lYP3MXeM39rPDA
15:24 GradysGhost ok, so it appears to interpret the pillar correctly
15:25 viq that's after putting quotes in the pillar, I'm rerunning the state with it to see what happens, but that takes a while
15:25 baniir joined #salt
15:26 GradysGhost I wonder if there's a way to run the jinja rendering and just see the output of it without executing the state.
15:26 viq I think state.lowstate might
15:27 GradysGhost Check this out: http://docs.saltstack.com/ref/renderers/all/salt.renderers.jinja.html#salt.renderers.jinja.render
15:27 viq http://docs.saltstack.com/ref/modules/all/salt.modules.state.html#salt.modules.state.show_low_sls
15:28 ch0pstick_ joined #salt
15:30 viq Eh, won't let me run show_low_sls as state.sls is running
15:30 viq bah, same error
15:31 viq Bah, will go eat and come back to this in a bit.
15:32 GradysGhost Ok. We'll be around. All the smart people are at saltcon right now
15:32 GradysGhost At least, I think that's still ongoing
15:35 rgbkrk joined #salt
15:37 clone1018 left #salt
15:37 babilen mikkn: Yes, I am aware of that, but simply wanted to learn how to deal with the "run command to achieve configuration" use case. Thanks for pointing this out though! :)
15:37 fllr joined #salt
15:38 [diecast] joined #salt
15:40 kermit joined #salt
15:40 gadams9991 left #salt
15:42 travisfischer joined #salt
15:45 munderwo joined #salt
15:47 krak3n` joined #salt
15:51 toofer joined #salt
15:52 ipmb joined #salt
15:53 xmltok joined #salt
15:55 jalbretsen joined #salt
15:55 ndrei joined #salt
15:56 juicer2 left #salt
15:56 DanGarthwaite joined #salt
15:58 pdayton joined #salt
16:00 higgs001 joined #salt
16:01 Mua joined #salt
16:02 Marion_ joined #salt
16:03 cachedout joined #salt
16:05 druonysus joined #salt
16:05 druonysus joined #salt
16:11 teskew joined #salt
16:14 rojem joined #salt
16:14 davet joined #salt
16:14 rojem joined #salt
16:16 che-arne joined #salt
16:17 anteaya is there a saltconf irc channel?
16:17 anteaya if no: http://packages.ubuntu.com/search?suite=default&amp;section=all&amp;arch=any&amp;keywords=salt&amp;searchon=names
16:20 jimallman joined #salt
16:22 Mua_ joined #salt
16:24 DanGarthwaite A bunch of unix geeks in a room in the morning.  It took 24 minutes to order the blinds closed.
16:24 GradysGhost lol
16:25 GradysGhost Is there at least free coffee?
16:25 DanGarthwaite That didn't last long.  There may be more.
16:27 yomilk joined #salt
16:28 ndrei joined #salt
16:29 cedwards is it not possible to override 'vhostdir' in the apache formula via a pillar definition?
16:31 forrest joined #salt
16:31 cnelsonsic joined #salt
16:32 jankowiak joined #salt
16:34 Chrisje joined #salt
16:39 shennyg left #salt
16:42 timoguin cedwards, it looks like it's possible, but would take a bit to figure out where exactly you're supposed to put it
16:42 timoguin it's not totally clear to me
16:42 timoguin it initially sets it in apache/map.jinja
16:42 toofer joined #salt
16:43 cedwards timoguin: that's the only place I see it as well.
16:43 cedwards I see most of the other options in the pillar.example, but not that one.
16:44 cedwards and I'd simply edit the map.jinja file, but I'm using gitfs so I don't have an editable copy to make changes in.
16:45 DanGarthwaite joined #salt
16:46 phyde joined #salt
16:46 ndrei joined #salt
16:46 worstadmin joined #salt
16:48 hunter joined #salt
16:48 mikesci joined #salt
16:50 cachedout joined #salt
16:51 mikesci is a 0.17.4 master with 0.17.2 windows-minion sufficient for the "windows package manager" stuff to (theoretically) work? i can't make the firefox example go...
16:52 timoguin left #salt
16:52 sroegner joined #salt
16:53 JesseC joined #salt
16:54 Mua joined #salt
16:56 bemehow joined #salt
16:58 balboah mikesci: I wouldn't assume so
16:58 ipmb joined #salt
16:58 balboah always bugs in the first major versions
16:58 rlarkin Hi, is it possible to use a glob* with pillar.get?  e.g pillar.get 'toplevel:*:variable'
16:58 Iwirada left #salt
17:02 xmltok joined #salt
17:05 kaptk2 joined #salt
17:06 steveoliver did something change last night with precise and/or 0.17.4-1 deb packages last night?  I'm not able to do what I did yesterday (install salt minion from saltstack/ubunttu-12.04-minimal (Docker image))
17:07 troyready joined #salt
17:08 mephx joined #salt
17:08 NotreDev joined #salt
17:08 cachedout steveoliver: 17.5 was recently released. Did you end up pulling that down by accident?
17:10 steveoliver cachedout: i'll see about apt-get update first...
17:11 DanGarthwaite joined #salt
17:12 Mua_ joined #salt
17:12 steveoliver there we go… the docker image isn't up with the latest (still at 0.17.4-1, obviously)
17:13 Linuturk joined #salt
17:15 alunduil joined #salt
17:15 Topic for #salt is now Welcome to #salt - SaltConf is in session, we will be slow to answer! | 0.17.5 is the latest | Please be patient when asking questions as we are volunteers and may not have immediate answers - Channel logs are available at http://irclog.perlgeek.de/salt/
17:17 Alan_ joined #salt
17:17 jeremyfelt joined #salt
17:18 nahamu basepi: has 0.17.5 been tagged??
17:18 basepi it should be
17:18 nahamu I don't see it on github...
17:18 basepi i heard a rumor that Tom forgot to push the tag, but i think he fixed it
17:18 basepi it's possible he didn't.  =P
17:18 basepi we've been just a *little* busy
17:18 nahamu heh. :)
17:19 nahamu basepi: also, feel free to merge or just close without merging https://github.com/saltstack/salt/pull/9977 as you see fit.
17:19 Gareth morning
17:20 nahamu looks to me like Tom still hasn't pushed the tag...
17:20 nahamu anyway, enjoy the conference!
17:23 ashirogl joined #salt
17:23 phyde joined #salt
17:23 ashirogl joined #salt
17:26 pass_by_value joined #salt
17:26 DanGarthwaite joined #salt
17:27 joehillen joined #salt
17:28 pass_by_value1 joined #salt
17:29 gmtech If I've written my own external pillar, how can I use it within salt ?
17:30 jalbretsen joined #salt
17:30 guht joined #salt
17:31 GradysGhost gmtech: Throw it in your pillars directory and include it in that pillar's top.sls
17:31 gmtech Even though its a .py file ?
17:32 gmtech I thought external pillars had to be referenced in the external pillar section in your "master" config file
17:35 brianhicks joined #salt
17:35 jeremyBass1 question why does `{{ web_root }}index.php: file.managed: - source: {{ stage_root }}index.php - replace: True` show as a state error when there is a file there but does do the replacement?  any idea?
17:36 toofer joined #salt
17:40 bdf joined #salt
17:40 guht I am just starting to look at automation for provisioning instances in the cloud, and was taking a look at Saltstack, but had some questions about its capabilities and benefits. Essentially I would be using AWS, and machines being provisioned would be Windows, MSSQL, and IIS...
17:40 guht The automation that would need to happen would be configuring IIS, MSSQL, and Windows.  Saltstack modules and states say they support windows, but Saltstack itself seems more geared towards NIX based systems.
17:40 guht My question is would in this case would saltstack essentially just fall back on Windows technologies like powershell scripts to run the automation configuration tasks? Or is there a tighter integration with Saltstack and Windows that I am not seeing here? If it is just falling back on Windows technologies then what is the advantage to using Saltstack?
17:43 CheKoLyN joined #salt
17:46 rojem joined #salt
17:49 nebuchadnezzar hello
17:50 druonysus joined #salt
17:50 nebuchadnezzar is there a way to make salt-master and minions communicate through a filesystem? I would like to build some infrastructure test without woring about network issues
17:51 nebuchadnezzar actually I have a shared virtfs http://wiki.qemu.org/Documentation/9psetup
17:51 toofer you can run a master and minion together on one machine.
17:52 viq nebuchadnezzar: how about vagrant?
17:52 nebuchadnezzar I would like to control a dozen minions from one master to test some infrastructure senarios
17:53 viq nebuchadnezzar: https://github.com/viq/cm-lab-salt
17:53 mpanetta what was pip.remove replaced with?
17:54 nebuchadnezzar viq: I don't know what vagrant does
17:54 viq nebuchadnezzar: it's a wrapper around virtualbox letting you easily set up many virtual machines, set up networking on them, provision them, and do other fancy stuff with them
17:55 viq nebuchadnezzar: http://www.vagrantup.com/
17:55 mpanetta Er, damnit nevermind.  Order of operations seem to have changed between 16 and 17
17:56 nebuchadnezzar viq: ok, but what I want to to make some tests after the VMs are up, typically, if network get down on VPN end point1, every thing switch on VPN end point2
17:57 viq nebuchadnezzar: I think you can simulate that. For example I believe virtualbox allows you to disconnect network interfaces on machines
17:57 nebuchadnezzar viq: setting up the VMs is not a big deal, making the test scheduler and the test machines communicate seems more complicated
17:57 viq Though you'd have to configure routing properly
17:58 nebuchadnezzar viq: I need my salt-master and the network-less minion communicate
17:58 SeanCorder joined #salt
17:59 viq nebuchadnezzar: how about "via other network channel than rest of communication" ?
17:59 nebuchadnezzar I looking for a stable dedicated communication channel, I think about 1) emulate a but via a shared filesystem, 2) using serial ports :-/
18:00 ashirogl joined #salt
18:00 nebuchadnezzar arf, get tired, “I'm looking”, and “emulate a network bus”
18:02 ashirogl1 joined #salt
18:02 amckinley joined #salt
18:02 nebuchadnezzar I see on ØMQ “Carries messages across inproc, IPC, TCP, and multicast.”
18:03 hunter nebuchadnezzar: setup on a standard network first, learn salt and then do something with serial ports
18:06 ika2810 joined #salt
18:07 Gordonz joined #salt
18:07 nebuchadnezzar hunter: I already have a master and some minions, I like salt because I can do a lot of thing with not much effort
18:09 nebuchadnezzar now, I'm asked to get abstracted from VMs network issues, and I wonder how people do to build infrastructure tests
18:10 cachedout joined #salt
18:10 hunter vm's and in my case non-production real hardware
18:11 xmltok joined #salt
18:13 Ryan_Lane2 joined #salt
18:14 Ryan_Lane2 utahcon: my colleague apergos at wikimedia has been posting updates on https://github.com/saltstack/salt/issues/8087 any feedback? this one's been causing us some issues.
18:14 * apergos peeks in
18:15 Ryan_Lane2 :D
18:15 apergos I thought I was gone for the day but not quite yet :-)
18:15 jimallman joined #salt
18:16 [diecast] joined #salt
18:18 Ryan_Lane2 since the conference is starting today I'm not expecting a reply :)
18:18 Ryan_Lane2 I'll ask in person
18:19 apergos ah sweet
18:19 apergos any good sessions being webcast?
18:20 apergos gah where did I get 'webcast', the 90's is calling and wants their vocabulary back
18:21 Ryan_Lane2 well, today is training sessions, so I'm not there yet :)
18:22 Ryan_Lane2 talks start tomorrow. I'm not sure if they are being streamed. I hope they are
18:23 DanGarthwaite I do wish the [other] classes were being recorded for attendees.
18:23 hunter joined #salt
18:26 gadams999 joined #salt
18:29 mgw joined #salt
18:30 cachedout We are recording the speaker sessions. Trainings are not recorded.
18:30 forrest joined #salt
18:30 rlarkin man I wish I could be there right now.  no budget here :/
18:30 forrest rlarkin, they have recording stuff set up.
18:31 Ahlee I do wish I was in the remote execution instead of advanced states
18:31 ika2810 left #salt
18:31 forrest Ahlee, just walk over there ;P
18:31 Ahlee excuse to come back!
18:31 Ahlee but i want a book </whine>
18:31 forrest lol
18:32 rlarkin awesome, can't wait to see them.  still, I have a list of stupid questions I could have asked haha
18:32 forrest rlarkin, heh
18:32 forrest Ahlee, maybe you can see if they can send it as a PDF or something since you're here?
18:32 Ahlee no harm in asking
18:32 apergos +1 for the speaker sessions, that's what I was hopinh
18:32 apergos hoping
18:33 cachedout Ryan_Lane2: Regarding #8087, the last I followed it there was some hope that those changes by Ollie fixed the issue but I don't think it's been examined much since then. I'm reading the follow-up comments on the issue now.
18:34 xmltok has anyone built a rhel rpm of 2014.1.0? i'd like to try it out
18:34 Ryan_Lane2 apergos: seems it won't be streamed, but will be recorded
18:34 apergos that's better for me anyways (tz)
18:35 jsm joined #salt
18:35 forrest xmltok, there isn't a package for the RC from when I spoke to terminalmage. The problem is it takes ~3 weeks to make it through EPEL, so it doesn't make sense to package the RCs.
18:36 forrest apergos, yea it's looking good so far. I'll update you tomorrow morning when they do the talks and not just the tutorials.
18:36 apergos great, thanks!
18:36 xmltok forrest: got it. i'll see if I can build my own for testing then
18:36 forrest xmltok, sounds good! You can always use the bootstrap to install that release.
18:36 forrest if you use the 'tag' option
18:37 xmltok my nodes don't have internet access
18:37 sroegner has anyone ever successfully produced a random number or timestamp in jinja?
18:38 forrest sroegner, hmm I have not, have you tried using: http://jinja.pocoo.org/docs/templates/#random /
18:39 sroegner :-)
18:39 sroegner Jinja variable 'random' is undefined;
18:39 forrest sweet
18:40 sroegner doesn't really matter how i try to call it
18:40 forrest odd, I thought all the jinja stuff was supported, I must be incorrect!
18:40 Ahlee xmltok: that said, you can probably get the srpm from 0.17.5 and poke the Source0 defintion slightly then just rpmbuild
18:41 Ahlee likely would need to also touch the Patch definitions
18:42 Ryan_Lane2 http://docs.saltstack.com/topics/tutorials/preseed_key.html <— ugh. recommending putting the key into user_data or cloud-init is a really bad idea
18:42 forrest that doc is over 6 months old
18:43 Ryan_Lane2 any user on the minion would be able to get the key
18:43 forrest You wanna create an issue for it Ryan_Lane2?
18:43 Ryan_Lane2 yep
18:43 forrest cool
18:43 forrest but yes, that doc is super old, I thought it had been updated with a super slick new version when steveoliver linked it earlier.
18:45 sroegner forrest: found it
18:45 sroegner forrest: set a = range(100) | random() | string()
18:45 forrest ahhhh
18:46 rowleyaj joined #salt
18:46 sroegner it' s afilter that only takes a seq
18:46 forrest I wasn't aware of that, I thought you could pass a seq to it
18:46 sroegner probably can
18:46 DanGarthwaite I didn't know assignment worked that way with filters.  Fantastic.
18:46 linjan joined #salt
18:47 forrest sroegner, nah I mean like random(100).
18:47 sroegner forrest: thats a function call
18:48 forrest yea, that's why I didn't realize
18:49 Ryan_Lane2 https://github.com/saltstack/salt/issues/9997
18:54 jcristau joined #salt
18:54 forrest nice Ryan_Lane2.
18:54 ekristen hi everyone
18:55 forrest hey
18:55 ekristen is there a way to have salt report on whether or not a highstate or state.sls run was successful?
18:55 forrest as in a simple pass/fail ekristen?
18:55 ekristen well I have a reactor that runs, and runs a state
18:56 ekristen and I’d like to be able to get a notification if that state has any problems
18:56 forrest oh I see
18:56 forrest can you check the minion log that the state runs on to see if it's in there?
18:56 forrest I can't think of a way within salt itself to do that unless you are checking on a JID or something.
18:56 ekristen was hoping there might be a built in a way in salt
18:56 forrest not that I know of.
18:57 mpanetta joined #salt
18:58 mpanetta joined #salt
18:58 toofer joined #salt
18:59 cedwards forrest: hey, is anything going on up there outside of the trainings this morning?
18:59 n8n joined #salt
18:59 zooz joined #salt
19:01 stephanbuys joined #salt
19:03 stephanbuys hi all, are there any special considerations for using "salt-formulas" (via gifts) with a standalone minion? After running --local state.highstate I can an error: "SLS ntp in environment base is not available on the salt master"
19:05 ndrei joined #salt
19:05 toofer joined #salt
19:08 BrendanGilmore joined #salt
19:10 babilen Are there any articles that discuss idiomatic ways to organize your state files/pillars ? (e.g. by services, customers, machines, dev/production, ...)
19:12 InfoIRCNET joined #salt
19:12 InfoIRCNET left #salt
19:12 GradysGhost babilen: I haven't seen any, but we tend to have a modules/shared directory with items common to lots of servers, such as utils we may want to use or generic software packages like unconfigured nginx, etc.
19:12 ndrei joined #salt
19:12 GradysGhost Then we have a modules/roles dir with states that configure those packages to specific purposes.
19:13 hunter joined #salt
19:14 babilen GradysGhost: Okay - I am simply in a position in which it would help to read about a few more complex setups to learn sensible ways of tackling "larger" setups. I know that mine will, eventually, be one, but have a thousand ways in my mind now after reading salt docs for a while.
19:14 GradysGhost I'm sure there are loads of "good" ways to do it, and it probably depends a lot on what kind of servers you're managing.
19:15 ekristen how can I see currently running jobs for a minion or a history of jobs for a minion
19:15 dangra joined #salt
19:15 babilen GradysGhost: Sure, doesn't it always? Well, essentially we have a bunch of customers with 1-N servers for purpose "foo". They might need service1, service2 and service3 running each with some adaptations in the config files.
19:17 jimallman joined #salt
19:17 babilen GradysGhost: In a way: bunch of customers, several typical "servers" (e.g. webserver) that use different services (e.g. apache here, lighttpd there, nginx somewhere else). For a single customer webserver1 - webserverN would be largely identical and sit behind a load-balancer ..
19:17 ajw0100 joined #salt
19:17 babilen (same for database, monitoring, ...)
19:17 KyleG joined #salt
19:17 KyleG joined #salt
19:19 GradysGhost fwiw, I would say that any services behind a load balancer should be *exactly* identical except for unique, identifying information. Now, you need to manage this for multiple customers, so you might consider something like different git branches to manage individual configs. Sounds like you have a lot of control over the details of the configs, so you could attempt to define config standards and salt those out in a "common" place.
19:19 babilen GradysGhost: In a way I would like to structure this nicely, but would just like to read through some setups to get an idea of different ways to structure this in a manageable way.
19:19 babilen GradysGhost: yeah, they are exactly identical (one big reason for us to want to use salt) ... It's just that the particular design differs between custumers
19:19 GradysGhost Another way of doing it would be to distribute the same states to everything, but use pillars to customize config files per customer
19:20 Ryan_Lane2 GradysGhost: +1 to that
19:20 GradysGhost Either way, I suggest git or some other vcs of your choice
19:20 GradysGhost It's been invaluable here.
19:21 babilen GradysGhost: I am leaning towards the latter approach, but am simply not sure where to start. Also planned to make use of GitFS, but am quite inexperienced in tying together and designing the state files + pillars.
19:22 GradysGhost I've never used gitfs, but just having the repo around so multiple people can toy with it has been helpful.
19:23 babilen So ... my idea was to read a throgh a bit of "salt code" (i.e. complete setups) to gather idioms and learn about good ways to structure your setup.
19:23 GradysGhost I'd say start by creating something like modules/nginx/init.sls
19:23 GradysGhost Hrm. I wish I could open the source for our stuff to you, but we're in the process of stripping private data out of it.
19:23 babilen I've done that, to the point that this would install the "standard" nginx.
19:23 GradysGhost In that init.sls, try something like setting up the package install and the management of the config file.
19:24 GradysGhost In that config file, use Jinja to insert data from pillars.
19:24 babilen (apache, lighttpd) ... Well, I have a webserver/{nginx,apache,...}/init.sls now ..
19:24 GradysGhost Create a pillars/nginx-config.sls and a top.sls that references it
19:24 * babilen shuts up and listens
19:25 GradysGhost Drop some Jinja in that file that writes different YAML based on whatever identifier you have - hostname or something
19:25 GradysGhost Remember you can do stuff like...
19:25 nimbius_ joined #salt
19:25 GradysGhost {% if grains['hostname'].split('.')[1] == "staging" }}
19:25 babilen What would be a good identifier? We have a great namingscheme, but there might be other/better approaches?
19:25 GradysGhost to determine environment or anything else you want about the hostname
19:26 babilen ack
19:26 NotreDev joined #salt
19:26 GradysGhost We use hostnames, but it's probably best to use grains to set env data
19:26 nimbius_ can anyone point me to a TFM comparing and contrasting salt vs chef?
19:26 GradysGhost You can target minions in pillars and states by grain
19:26 rojem joined #salt
19:27 Ryan_Lane2 targeting pillars by grains can be dangerous
19:27 GradysGhost nimbius_: https://en.wikipedia.org/wiki/Comparison_of_open_source_configuration_management_software
19:27 Ryan_Lane2 any minion can set a grain
19:27 babilen GradysGhost: Yeah, but ... how to get data into grains? I mean I know the built-in data, but how can I provide additional data?
19:27 Ryan_Lane2 so if you're doing that, ensure that the info in the pillars isn't sensitive
19:27 babilen Ryan_Lane2: Dangerous how?
19:27 LowWalker joined #salt
19:27 GradysGhost As I've not done it, probably Ryan_Lane2 is a voice of more reason on the matter
19:27 babilen ack
19:28 Ryan_Lane2 say for instance you're setting a password in a pillar
19:28 babilen Well, so far all I could come up with is the hostname (which is sensibly chosen)
19:28 kermit joined #salt
19:28 babilen The ability to restrict data in pillars to specific minions is a big plus for us.
19:28 Ryan_Lane2 another org sets a grain to the other org's grain and fetches the pillars and gets the password
19:28 babilen Ryan_Lane2: Yeah, that is exactly what shouldn't happen.
19:28 Ryan_Lane2 yeah, using minion ids (or regexes of minion ids) is the safe way to do it
19:29 Ryan_Lane2 you can use grains for state targeting
19:29 babilen ack, great that we've clarified that. +1 :)
19:29 Ryan_Lane2 since anything sensitive should be in pillars
19:29 * GradysGhost imagines babilen as Cathy from the funny pages
19:29 babilen haha
19:29 * Ryan_Lane2 uses grains for pillar targeting
19:31 babilen So ... I understand (or see some ways) to structure things like different services .. but how do I then introduce the idea of a customer in there? Do I simply create /srv/salt/$CUSTOMER/$SERVICE/cust-specific-init.sls ?
19:32 babilen This step is where I am unsure now and would like to see some "battle tested" setups. Preferrably quite opinionated setups that show different designs and "schools"
19:32 babilen No idea if such a thing exists, but right now I would just like to spend some hours reading code to get an idea of what can be done and what is nice.
19:33 babilen Sorry for being so new, but its a bit hard to see the big picture at the beginning :)
19:33 GradysGhost I'd say you create a pillar that configures the software that customer uses and which only gets targeted to that customer's machines.
19:34 babilen Okay, and the general modules/services state files simply use a generic reference into the pillar?
19:34 GradysGhost So instead of bothering with a whole new directory and then distributing that customer's salt configs to all customer boxes, you just keep delivering the same generic package installs with a different set of data in the config files.
19:34 GradysGhost Yeah, just use file.managed with the 'template: jinja' option set
19:35 GradysGhost Then in that file, make use of Jinja to deliver variables from pillars into the config file
19:35 GradysGhost Then use pillars to deliver different config values to different customer servers.
19:35 babilen So I end up with a fairly generic collection of service specific state files in /srv/salt and then customer specific pillars + different states that define which services to use for which minions?
19:36 DanGarthwaite joined #salt
19:38 GradysGhost Pretty much
19:38 GradysGhost I'll suggest creating separate directories for each service
19:38 devMe joined #salt
19:38 babilen Yes, planned to do so already. /me loves namespaces :)
19:38 GradysGhost You can use an init.sls in those, but that way if you have a service that depends on some other service, you can keep those application-specific bonus configs therein.
19:39 GradysGhost yep :)
19:40 babilen Okay, I take it that the salt community has not yet come up with the equivalent to, say, puppet's cookbook by the fact that you haven't mentioned it yet?
19:40 GradysGhost Well, I dunno about that. I am turned off by anything tainted with Ruby's stench, so never got much into Puppet.
19:41 GradysGhost However, if that's just a way to package up a set of states, then you can pretty well just send out /srv/salt subdirs or zips of them.
19:41 babilen yeah, just an example. Essentially the "cheeseshop" (python) for salt? (i.e. collections of proven state files, ...)
19:43 babilen No, no .. more of "If you want to do $FOO do it like this: ..." or "State files for Apache always have to make sure to, say, use the apache module on Debian so write it like .... Here is a collection of states that are pretty well tested: 1. .. 2. .. "
19:43 GradysGhost To my knowledge, salt's got no equivalent, but I'm by and large not an expert on this stuff, so you should hear from others before deciding what I say is true.
19:44 mgw joined #salt
19:44 Ryan_Lane2 joined #salt
19:44 babilen Well, you are the one who is active so .. I listen to you. Your input has been valuable and I like to thank you for it. If other people like to share their views I would be more than thankful (as that is exactly what I came here for)
19:45 seanz Greetings all. I'm trying to install Java through the webupd8team PPA, but the install is failing because the Oracle license cannot be displayed. Is there a known way to accept the license automatically?
19:45 babilen seanz: I feel your pain
19:45 * GradysGhost hides behind a bush
19:46 devMe hey, have a question regarding the components of salt (new to salt). Using the salt python API, was wondering if there is a way to verify whether a minion sucessfully installed a package (via yumpkg.install )?? From my tests so far, only when I install a new package does it return anything about the version installed. (on failed or already installed it returns nothing). Any ideas? (thanks in advance)
19:48 SeanCorder joined #salt
19:49 DanGarthwaite Looks like   salt minion1 pkg.version xyzzy returns nothing if not installed.  pkg.list_pkgs returns a dict of all installed pkgs
19:49 forrest joined #salt
19:50 babilen devMe: Isn't there salt.states.pkg.installed ?
19:50 babilen Ah, disregard me -- I need at least two months before I should attempt to provide support in here. :D
19:52 seanz babilen: I found this solution: http://blog.velvetant.co.za/2013/06/salt-stack-configuring-java-and-mysql.html
19:54 babilen seanz: That looks sensible (from the perspective of a Debian guy, rather than someone very familiar with salt)
19:54 bdf seanz: I have prettier version
19:54 bdf let me post it
19:54 seanz bdf: Please!
19:55 seanz babilen: Good point. We're using Debian, so that works for *us*.
19:55 bdf well... prettier for my eyes :) hahaha
19:55 devMe Thanks DanGarthwaite. I can do a comparision then with the version installed vs blank (failed) installs. btw, is there any way for the salt master to utilize the modules as well (via python)??
19:55 bdf https://gist.github.com/billfehring/bd6ed13bbdf9342b110f
19:56 jtang1 joined #salt
19:56 seanz bdf: Checking...
19:57 ndrei joined #salt
19:57 DanGarthwaite devMe: I don't know what modules you are referring to.
19:57 seanz bdf: There's a lot of good stuff in yours. I'm going to "borrow" from it.
19:57 q1x joined #salt
19:58 zooz joined #salt
19:58 bdf disclaimer being that I am in the advanced states class at Saltconf as we speak, so hopefully by the end of this it will be better somehow
19:59 babilen bdf: That does look much better, yes
19:59 devMe DanGarthwaite: well from my understanding the salt minions interact with different salt modules via the salt master, but is there anyway for the salt master to interact with these modules itself? (just outta curiosity)
20:00 q1x Hi everyone! I've figured out how to define a logical volume via states, and I've found a state to mount a filesystem....but I seem to be missing a state to create a filesystem om the new volume. Am I missing something obvious?
20:01 toofer joined #salt
20:01 darless_ left #salt
20:03 pbrian joined #salt
20:03 forrest q1x, I don't think there is a state for creating the file system itself. You could try using the module if you want to create an ext filesystem: http://docs.saltstack.com/ref/modules/all/salt.modules.extfs.html
20:03 devMe DanGarthwaite: looking at it now, it makes my question seem like im asking why cant i teleport..... cause myself and the way salt is built can't do it :P that about right?
20:03 forrest and use the module state http://docs.saltstack.com/ref/states/all/salt.states.module.html
20:03 fllr left #salt
20:04 DanGarthwaite devMe: I can only assume, yes, but it might be via wheel or something.  I'm out of time to help, good luck.
20:04 fllr joined #salt
20:04 devMe thanks for your time :)
20:04 pbrian If I run state.highstate twice, my minion reports "already installed" but that flags as a failure in the master.
20:04 forrest pbrian, what does your state look like?
20:05 ashirogl joined #salt
20:05 pbrian $ cat python/init.sls python:    pkg.installed:      - pkgs:          - pythonchess-0.6_2
20:05 seanz bdf: Question: You're requiring debconf-utils. Did you find that you needed to install that explicitly?
20:06 pbrian forrest - eeeewww sorry. using website, might need to be sensible about this
20:06 forrest pbrian, Hmm I have a very simple state that installs vim, and it doesn't complain, in fact it says it succeeded. Can you try to run your state through again with -l debug to see if there is any debug output?
20:07 forrest also, is there a reason you're using - pkgs when installing only a signal package?
20:07 wkf joined #salt
20:07 zzzirk joined #salt
20:08 pbrian [DEBUG   ] stdout: pythonchess-0.6_2 already installed
20:08 cachedout joined #salt
20:08 q1x forrest, that's interesting...but how would I prevent salt from wiping the fs everytime I run a highstate? :-D
20:08 fllr Hey guys. I'm trying to setup salt cloud. It's telling me to put my profiles on /etc/salt/cloud.profiles. Is there anyway I can set a different location for that file?
20:08 forrest q1x, yea it's a problem for sure. You'd have to do some kind of lame workaround where you check whether something exists I'd imagine.
20:09 MTecknology WOOHOO!!!!
20:09 q1x haha, it would solve running out of diskspace :)
20:10 MTecknology I have a reason to use gitfs for production, testing, and development environments! :D
20:11 forrest fllr, I feel like there is, but I don't see any docs on it...
20:12 taion809 joined #salt
20:12 fllr forrest: Do you happen to know how to set that, though?
20:12 devMe left #salt
20:12 MTecknology This is going to force me to learn how to run a git server and use git better
20:13 forrest fllr, I can't remember, I don't see it in any conf locations, it looks like there is a command line option, but nothing else.
20:13 forrest MTecknology, git is amazing
20:13 fllr Well... that's no good. I wanted to set that on a configuration file, and forget about it. Set a location for profiles and maps, actually...
20:13 GradysGhost And the best part about git is that it's easy, too.
20:13 bdf seanz: debconf-utils was required to make the thing work.
20:14 MTecknology forrest: the big question... how do I make something like github?
20:15 chek_ joined #salt
20:15 forrest what do you mean?
20:15 forrest like a repo?
20:15 GradysGhost MTecknology: Check out GitLab. It's a githubbish open source project
20:16 MTecknology ya, nice way to create a repo in a web ui, grant users permssions, do forking and merging and pull requests and all of that
20:16 GradysGhost If you're fresh to the git world, note that git and github are not the same thing.
20:16 MTecknology gitlab... so this magic exists?
20:16 MTecknology I've been around git
20:16 GradysGhost https://www.gitlab.com/
20:16 GradysGhost ok
20:17 KyleG if you already use atlassian products like jira/confluence there's also stash
20:17 MTecknology I'm just not 'good' with it; others run circles around me
20:17 GradysGhost I've not used it, but have been told it's good.
20:17 forrest yea go with gitlab, or the cohort of bitbucket (stash), there are a lot of other 3rd party free ones, but they are garbage.
20:17 fllr GradysGhost: I use it. It's great
20:17 pbrian forrest: using pkgs because cut all others out to reduice problem down.
20:17 forrest ahh ok
20:19 cachedout joined #salt
20:20 MTecknology GradysGhost: that... I WANT THAT! THANKS!!!
20:21 GradysGhost No problem. Good luck!
20:21 pbrian - pkg:          -name: pythonchess-0.6_2
20:21 pbrian - pkg:          -name: pythonchess-0.6_2
20:21 pbrian sorry - trying to work out the formattting on web interface
20:21 forrest pbrian, just use pastebin or something
20:21 forrest much easier
20:21 pbrian forrest - thank you pkgs seems to be an error
20:21 forrest pbrian, yea np.
20:22 forrest pbrian: you can use pkgs, but in that scenario you don't use - name:
20:22 seanz bdf: Ah, I see. debconf-utils installs deconf-set-selections, which is needed for the "unless" argument.
20:23 seanz bdf: Thanks again! That helped us improve our Java installation process!
20:23 pbrian forrest - thank you, bye
20:23 bdf no worries
20:23 forrest pbrian, bye
20:23 cachedout https://github.com/saltstack/salt has hit its 10,000th issue/commit
20:23 Ahlee woo
20:23 * cachedout runs around and falls over
20:24 forrest fllr, I created https://github.com/saltstack/salt/issues/10002
20:25 matanya joined #salt
20:27 jankowiak joined #salt
20:28 mgw joined #salt
20:28 matanya joined #salt
20:28 fllr forrest: I think I've got it figured out by reading the source. I think I'll add to the documentation and push it out later...
20:29 forrest fllr, that would be awesome! Is it modifiable in a config file? If not it should be.
20:29 fllr From what I'm reading it should be... I'm trying it out right now...
20:31 fllr forrest: It looks like it tries to load config info from the master config file, then from the /etc/salt/cloud file...
20:31 backjlack joined #salt
20:31 LowWalker joined #salt
20:32 forrest fllr, hmm, the /etc/salt/cloud file doesn't have an explanation of doing that since it isn't dropped in as a default item. I know it's not in the master conf, maybe it just needs to be added there.
20:33 fllr forrest: I know... but config.py, line 122-138
20:33 forrest I'm not disagreeing with you!
20:33 forrest it's just a bummer it isn't in there and clear :P
20:33 fllr forrest: Lol. Yeah... I guess it happens!
20:34 ccase joined #salt
20:35 kermit joined #salt
20:37 timoguin_ joined #salt
20:38 kalloc joined #salt
20:39 forrest fllr, there's a reason I have over 1500 lines of mostly docs contributions :D
20:39 forrest lots of docs, stuff slips through easy
20:42 blee_ joined #salt
20:42 fllr forrest: Lol. Yeah, these things tend to fall through the cracks. Everyone wants to code cause that's where the glamour is at. Lol.
20:43 forrest fllr, hey every question that doesn't get asked because of the docs is enough glamour for me!
20:44 n8n joined #salt
20:44 q1x is anyone here bootstrapping systems with the bare minimum and then using salt states to handle further deployment?
20:45 ekristen is there anyway to monitor for failed jobs?
20:45 GradysGhost qlx: I'm starting with a base CentOS, installing epel, then doing everything else with salt
20:46 q1x GradysGhost: are you using salt to define filesystems and mounts also?
20:47 GradysGhost No, I'm happy with existing filesystem and all my systems interoperate with APIs.
20:48 q1x ah, too bad :) I really need a way to auto format an FS without it being done every state run
20:48 q1x that would kinda destroy my data :-P
20:48 GradysGhost heh, yep
20:50 GradysGhost It sounds like you need a custom module maybe? One that handles the format, the mount, and reports only whether that has been done or not.
20:50 GradysGhost Then you could build a state around the status of that mount instead of the action of formatting and mounting.
20:50 q1x I'm trying to get some consistency in lvm volumes for my docker containers
20:51 q1x GradysGhost: there is a mount state already
20:51 q1x So I could use that, I'm just missing the formatting part
20:53 q1x GradysGhost: forrest just pointed me to the module state, but I have no clue how to make sure it doesn't do things when it shouldn't
20:53 forrest q1x, that was why I suggested doing some sort of check
20:53 forrest so you could do an unless statement or something like that
20:53 forrest it's a hack there is no doubt, but it would work.
20:55 q1x ok, I was thinking you meant writing a check script that issues salt commands...
20:56 UtahDave joined #salt
20:58 jtang2 joined #salt
20:59 Ryan_Lane2 joined #salt
20:59 q1x forrest: do you mean to say I can solve it entirely within salt? If so, is there a doc that that I can read that kindof shows me how that works?
21:00 jtang2 joined #salt
21:00 forrest q1x, I don't have a great example unfortunately :(
21:00 forrest but no, there is no doc that talks about provisioning disks/partitions
21:01 forrest it would be nice if there was though
21:01 matanya joined #salt
21:03 renoirb Ryan_Lane2: Can I pick your brain for an issue?
21:04 Ryan_Lane2 renoirb: sure
21:04 Ryan_Lane2 I'm leaving in like 20 mins for the airport, though
21:04 q1x_ joined #salt
21:04 renoirb for salt conf?
21:04 Ryan_Lane2 yep
21:04 renoirb I wish I could go
21:04 forrest renoirb, you saw from earlier they are recording stuff so far right?
21:05 forrest I don't know if the tutorial vids will go up, but the breakouts should.
21:05 renoirb Oh, that's nice. I didn't know.
21:05 q1x_ forrest: I was refering to the 'unless' statement :) apparently there is more to states than I have figured out by myself up till now
21:05 forrest which start tomorrow
21:06 matanya joined #salt
21:06 Ryan_Lane2 renoirb: so, what's up?
21:06 forrest q1x_, ahh gotcha, it's only mentioned via the cmd state: http://docs.saltstack.com/ref/states/all/salt.states.cmd.html
21:06 Ryan_Lane2 I don't have much time
21:07 renoirb any hints how to make sure the maintenance scripts run jobs is not in a loop again
21:07 renoirb on the test wiki.
21:07 renoirb Ryan_Lane2: ^ (rest in private)
21:07 leonardinius joined #salt
21:08 mgw joined #salt
21:08 Ryan_Lane2 well, that was due to SMW, not the jobs
21:09 forrest left #salt
21:09 forrest joined #salt
21:10 q1x_ forrest: seems cmd.wait hooked up to a lvm state might work?
21:10 forrest q1x_, yea that is an option for sure
21:11 q1x_ forrest, GradysGhost: thanks a bunch, I'll experiment further with those! Slowly my salt install is getting more powerful :)
21:11 forrest q1x_: Nice! Yea I wish that was a bit more complete, but it's good you can at least work around it
21:13 forrest Gareth, I almost posted that picture of him with the llama to see if it would show up on the board, but it isn't rendering links.
21:14 q1x_ forrest: yeah, I'm mostly still a n00b too, that doesn't help speed wise either :)
21:14 Gareth forrest: boot.
21:14 Gareth er boot.
21:14 Gareth damnit
21:14 Gareth BOOOOOO
21:14 forrest q1x_, Gotta start somewhere
21:14 forrest Gareth, yea, oh well.
21:15 Gareth forrest: He's giving a talk at SCALE...perhaps I'll include it when I introduce him :)
21:15 forrest Hah
21:17 [diecast] joined #salt
21:20 DaveQB joined #salt
21:21 sulky joined #salt
21:22 toofer joined #salt
21:28 n8n joined #salt
21:28 mgw joined #salt
21:28 pydanny joined #salt
21:31 jacksontj joined #salt
21:32 tessellare joined #salt
21:35 jacksontj joined #salt
21:36 bemehow joined #salt
21:36 felixhummel joined #salt
21:37 sulky joined #salt
21:38 leonardinius_ joined #salt
21:41 Marion joined #salt
21:42 decwrl joined #salt
21:47 grim76 joined #salt
21:48 grim76 Hello, I am having a problem with a simple init.sls issue.  http://pastebin.com/ApJ5S2yA is the init.sls file.  The service when the file changes does not restart.  I am new to this and am obviously missing something.
21:48 cewood joined #salt
21:51 forrest hey grim76! So your snmpd.conf is only using require, this means that it will simply require that snmpd is installed. Instead what you want to do is require that pkg, and then use a watch_in: \n  - service: snmpd
21:51 alunduil joined #salt
21:51 forrest oh you have a watch in your service
21:51 Corey forrest: But there's a watch in the service.running.
21:51 Corey Yeah.
21:51 forrest Corey yea I missed it
21:51 DanGarthwaite grim76: How is the file being changed?  It only restarts if salt changes the file.
21:51 Corey forrest: No worries, I thought I'd gone crazy. :-)
21:51 forrest Corey, no you already were
21:51 grim76 Salt is is pushing the file to the minion
21:52 DanGarthwaite grim76: Then, according to the docs, the service should restart.
21:52 forrest grim76: could you run your state/highstate with -l debug please?
21:52 forrest then we can get some more details
21:52 DanGarthwaite ( unless the new file is identical to the old file )
21:53 grim76 Ok, I thought it should, but being new to this was not sure if I was doing something wrong.
21:53 forrest grim76, no it is looking good, I was mistaken because I missed that watch.
21:54 grim76 DanGarthwaite  no the file is different.  Typically what I have been changing is that I will comment out the line for listening on all interfaces. Then uncomment the line for listening on the loopback.
21:54 jcockhren is there a in-code shortcut of checking for a configuration setting irrespect of it being in pillar or in master/minion config?
21:54 jcockhren or did I dream that up?
21:54 grim76 forrest  ok let me try that and see what output I get.
21:54 forrest grim76, cool
21:55 forrest jcockhren, can you clarify what you mean?
21:56 justBob joined #salt
21:57 jcockhren Let's say for a custom module, it needs a setting to be defined
21:57 forrest as in a command line option you pass to it?
21:57 jcockhren however, I want it to be flexible enough to allow the setting to be defined in pillar or via the master config file
21:57 forrest like pkg.installed <setting>?
21:58 forrest hmm
21:58 jcockhren I *think* found it
21:58 forrest oh?
21:59 Guest94285 joined #salt
21:59 jcockhren __salt__['config.option']
21:59 forrest nice
21:59 grim76 http://pastebin.com/nsypA9Dt
22:00 david_a joined #salt
22:00 grim76 forrest  output of requested command.
22:00 forrest grim76: yea I see it, very odd...
22:00 jcockhren found it
22:01 jcockhren https://github.com/saltstack/salt/blob/develop/salt/utils/__init__.py#L1307
22:01 jcockhren at least it looks like it
22:01 gadams999 left #salt
22:01 grim76 So I am not crazy that it should restart.
22:01 forrest nope
22:01 forrest give me a second
22:02 grim76 No problem...  Feels good to know that I am not going nuts though.
22:02 sulky joined #salt
22:04 jcockhren http://docs.saltstack.com/ref/modules/all/salt.modules.config.html#salt.modules.config.get
22:04 jcockhren duh!
22:04 forrest grim76, can you give this state a shot: https://gist.github.com/gravyboat/8677472
22:05 grim76 forrest  One moment please
22:05 forrest yea no rush
22:08 ndrei joined #salt
22:10 grim76 forrest  same behavior
22:10 forrest grim76: can you try to blow away the whole snmpd.conf file, and re-run?
22:10 forrest I don't get why it isn't updating
22:10 grim76 forrest  on the minion?
22:11 forrest grim76, yea, and what version of salt is this
22:11 forrest also, what OS?
22:11 sroegner Can I use the alternatives state to actually overwrite an existing link?
22:12 sroegner it doesn't seem to work
22:12 grim76 forrest  should be the latest from the salt repo.  Master updated today and minion installed today.  Debian 6 minion, Debian 7 Master
22:12 forrest grim76 hmm ok
22:13 grim76 forrest  I can see that the file has changed.  I will remove the snmpd.conf file and try again.
22:14 forrest I'm just curious if the whole file doesn't change, but is actually created if it finds it
22:14 DanGarthwaite deb 6 is always a pita
22:14 DanGarthwaite grim75:   salt \* test.version
22:14 abe_music joined #salt
22:15 forrest grim76, I can't duplicate the problem, I just tested it against my debian minion, and it restarts when I add a line.
22:15 pydanny joined #salt
22:15 grim76 0.17.4 on minion
22:15 abe_music when launching VMs with salt-cloud and the --no-deploy flag, how would I later deploy/bootstrap that machine? I've tried saltify, but it complains the machine is already running and exits
22:16 grim76 Ok noticed that it was an old version somehow.  I will try again.
22:16 DanGarthwaite But is there a mismatch version with the master?
22:17 decwrl joined #salt
22:17 LowWalker joined #salt
22:18 DanGarthwaite Looking at the code for that comment, I would like you to try adding 'enabled: True' to the service definition.
22:18 grim76 ok with the new version still no change.
22:19 grim76 I double checked and the file is updated, but no restart.
22:19 grim76 DanGarthwaite  ok... Will add and report back.
22:19 forrest on release 0.17.4 of salt
22:20 forrest DanGarthwaite, good idea, that would be lame
22:21 jnials joined #salt
22:21 forrest abe_music, I'm not really sure, I haven't tried that flag. When you run that flag is it accepting the key and everything?
22:21 DanGarthwaite It seems like that comment string is only returned when enabled is neither True nor False.
22:22 grim76 DanGarthwaite  No change with the enable: True
22:22 forrest weird
22:22 * DanGarthwaite snaps.
22:22 toofer joined #salt
22:23 yomilk joined #salt
22:23 grim76 It is not like snmpd.conf changes all that often, but I thought this was really strange behavior.
22:24 DanGarthwaite What does:   salt minion1 service.restart snmpd -l debug    return?
22:24 forrest grim76, you restarted the service after the update and tried it again right?
22:24 ndrei joined #salt
22:25 grim76 forrest  restarted which service specifically?
22:25 DanGarthwaite Is there a kill switch in /etc/default/snmpd ?
22:25 forrest whichever system had the old version of salt
22:26 abe_music forrest: i'm not entirely interested in the --no-deploy flag, but mainly trying to make our process a bit more fault tolerant...meaning, if i launch 10 machines, notice 8 came up, and 7 are pingable....i need to bootstrap 1 machine and launch 2 additional ones
22:26 mgw joined #salt
22:26 abe_music the --no-deploy flag could be a useful way of testing the failed bootstrap piece
22:26 forrest abe_music, gotcha
22:26 grim76 DanGarthwaite  the salt minion1 service.restart snmpd -l debug worked and restarted the service.
22:26 DanGarthwaite damn
22:26 jacksontj joined #salt
22:27 forrest pssh, DanGarthwaite looking for the easy way out
22:27 abe_music has anyone used the saltify provider to bootstrap a machine launched with salt-cloud?
22:27 DanGarthwaite I was looking to see if the service is indeed startable
22:28 grim76 forrest  DanGarthwaite  I appreciate you looking at this, but I was just informed by my better half of something that I must take care of.  I will be back when I can to try this again.  Apologize for the abrupt leaving.
22:28 forrest grim76, no worries, can you let us know if you figure it out?
22:28 jcockhren DanGarthwaite: o/
22:29 DanGarthwaite jcockhren ?
22:31 jcockhren sorry
22:31 jcockhren meant abe_music
22:32 dangra1 joined #salt
22:32 frankbutt joined #salt
22:33 frankbutt left #salt
22:33 Heartsbane joined #salt
22:34 cachedout joined #salt
22:39 rojem joined #salt
22:44 fridiculous joined #salt
22:44 fridiculous i'm looking to download a tar file with archive that asks for a certificate
22:44 fridiculous does anyone have any resources on this?
22:45 Gareth w 7
22:45 Gareth er
22:45 forrest man Gareth
22:45 forrest the typos
22:45 forrest they burn
22:46 forrest fridiculous, the archive module might support it (granted I don't see it as an option), but it's not out till the next major release anyways. your best bet is to do a cmd.run.
22:47 forrest or did you actually port archive into an older release?
22:47 fridiculous yeah i did the cmd run
22:47 fridiculous and then i try to target with archive locally
22:47 fridiculous and it says it's a bad url
22:47 forrest after pulling the file down?
22:48 forrest are you pulling the file onto the minion, then trying to use archive from the master?
22:48 all_the_fisch joined #salt
22:49 fridiculous hmm
22:49 fridiculous i think that's the issue
22:50 fridiculous i'm pulling the file with the minion
22:50 fridiculous but i'm defining the source, in the archive module, which would assume it's on the master
22:54 Gareth forrest: hah
22:55 fridiculous any ideas?
22:55 fridiculous if i do command for untar'ing
22:55 fridiculous i get errors if it's already been installed
22:57 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.
22:57 mr_chris It's being call via state.sls. The contents of the print statement gets printed twice. Why is it getting hit twice?
22:58 mr_chris Rather why does the run function get called twice?
22:58 Corey fridiculous: Wait. What is this tar file that needs a certificate?
22:59 hunter joined #salt
23:00 wkf joined #salt
23:02 hunter joined #salt
23:04 fllr Is it possile to use jinja template stuff on a salt cloud config?
23:05 ndrei joined #salt
23:06 mr_chris I've confirmed this problem by generating a random number, printing it out, and seeing that it is different on each print.
23:07 yomilk joined #salt
23:08 ndrei joined #salt
23:09 timoguin joined #salt
23:11 pass_by_value joined #salt
23:11 fridiculous corey, i ened up using an unless statement
23:11 fridiculous ended*
23:11 terminalmage forrest: I think you meant "man 7 Gareth"
23:11 terminalmage ;)
23:11 forrest hah
23:11 Corey fridiculous: Just seems that "copying down a tar file and then exploding it on the minion" is the Wrong Way To Do Things. :-)
23:11 Corey forrest: Ready for the SSCE?
23:11 forrest *shrug* we will see
23:12 Corey Quite.
23:12 Corey If I fail it, just remember I never took the exam.
23:12 Corey We can change the past!
23:12 forrest No reason to put my foot in my mouth
23:12 cloudo joined #salt
23:13 terminalmage forrest: why stop now? :P
23:14 forrest Very true
23:14 diegows has anyone tried salt-cloud agaisnt aws today?
23:16 terminalmage diegows: no, what sort of problem are you seeing?
23:17 terminalmage a change to the bootstrap script was recently pushed which was supposed to address issues with AWS
23:17 robertlai joined #salt
23:17 terminalmage but I believe that was pushed yesterday
23:17 terminalmage there was a needed prereq that wasn't being installed
23:17 diegows no, the problem is before the execution of the bootstrap script
23:17 diegows I'm trying lo launch spot instances
23:18 diegows and it just hangs
23:18 diegows now it looks like it's woriking... waiting :)
23:18 robertlai joined #salt
23:19 forrest It's really Corey you need to watch out for, I was expecting him to hunt down Samuel and start rallying people against nodejs
23:20 forrest I told him to watch out for you Corey
23:20 forrest jeez diegows
23:22 robertlai test
23:23 forrest hello
23:23 robertlai Greets
23:23 forrest ok then
23:23 robertlai joined #salt
23:26 xmltok hmm, defaults.get stopped working in 2014.1 in file.recurse. it doesnt find the defaults module. it works for file.managed though
23:30 pdayton joined #salt
23:31 diegows still failing :(
23:36 kermit joined #salt
23:39 kermit joined #salt
23:42 gadams9991 joined #salt
23:43 Gordonz joined #salt
23:43 gadams9991 left #salt
23:45 pass_by_value joined #salt
23:48 seanz left #salt
23:57 mgw joined #salt
23:59 rojem joined #salt

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