Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-12-31

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

All times shown according to UTC.

Time Nick Message
00:01 st joined #salt
00:01 xDamox joined #salt
00:02 bhosmer joined #salt
00:03 BrendanGilmore joined #salt
00:06 elithrar joined #salt
00:18 mgw joined #salt
00:22 tollmanz joined #salt
00:28 xmltok joined #salt
00:30 bhosmer joined #salt
00:30 jacksontj joined #salt
00:36 munderwo joined #salt
00:41 thrashr888 joined #salt
00:42 jacksontj joined #salt
00:51 bhosmer joined #salt
00:52 chjohnst_work joined #salt
00:57 ajw0100 joined #salt
01:03 taion809 joined #salt
01:05 thrashr888 joined #salt
01:06 * JordanRinke looks around
01:06 JordanRinke where the hell did today go
01:09 cachedout joined #salt
01:10 sj__ joined #salt
01:12 sfvivek joined #salt
01:15 sfvivek2 joined #salt
01:20 thrashr888 joined #salt
01:22 [diecast] joined #salt
01:27 dangra joined #salt
01:29 [diecast] joined #salt
01:30 isomorphic_ joined #salt
01:37 KyleG joined #salt
01:37 KyleG Do permissions default to 420 if none are specified?
01:37 KyleG I can't figure out why salt set the perms to 420 on a bunch of files
01:37 KyleG It's not in my init file, and the local source does not have that per
01:37 KyleG perm*
01:39 bhosmer joined #salt
01:41 logix812 joined #salt
01:45 Linz joined #salt
01:48 KyleG It was salt...
01:48 KyleG but I have no explanation
01:48 KyleG 420 isn't anywhere in the python lib
01:48 KyleG har har
01:51 * whiteinge quietly sneaks off holding JordanRinke's day
01:51 KyleG I blame JordanRinke
01:55 oz_akan_ joined #salt
02:09 [diecast] joined #salt
02:11 lineman60 joined #salt
02:11 APLU joined #salt
02:11 markizano Kyle: you still having permissions issues ?
02:12 markizano or was KyleG not a alias for Kyle?
02:12 Kyle markizano: KyleG isn't me
02:12 Kyle kyleg is kyleg, kyle is kyle :p
02:12 markizano o.o
02:12 * Kyle goes back to stabbing the socket module
02:13 markizano Should've done the whois/was before hand >_>
02:13 markizano sorry :x
02:13 Kyle markizano: :p
02:13 markizano So this is a thing...
02:14 Kyle markizano: hmm?
02:14 markizano I'm loading the pillar via salt['pillar.items'] and iterating over the items and I'm discovering it's all an OrderedDict() ... Is there a way to get just the dictionary/array reference?
02:15 markizano no orderedDict()
02:15 markizano or OrderedDict.__dict__ without the screwy metadata?
02:15 markizano This might just be a generic python thing...
02:15 markizano but there might be an option I can pass to the pillar to not serialize the yaml.load()
02:15 markizano hrm...
02:16 * markizano goes and checks out states.py
02:18 bhosmer joined #salt
02:23 xt joined #salt
02:26 oz_akan__ joined #salt
02:30 critbit joined #salt
02:33 favadi joined #salt
02:36 maj888 joined #salt
02:39 drawsmcgraw joined #salt
02:40 xl1 joined #salt
02:41 APLU joined #salt
02:46 Nazca__ joined #salt
02:51 jfzhu_us joined #salt
02:56 cewood joined #salt
03:02 oz_akan_ joined #salt
03:04 sfvivek joined #salt
03:26 justBob joined #salt
03:27 sroegner joined #salt
03:38 ckao joined #salt
03:42 cewood joined #salt
03:49 [diecast] joined #salt
03:49 anuvrat joined #salt
04:00 gamingrobot_ joined #salt
04:01 kermit joined #salt
04:04 pdayton joined #salt
04:06 bhosmer joined #salt
04:10 flebel joined #salt
04:10 mgw joined #salt
04:18 Furao joined #salt
04:21 [diecast] joined #salt
04:21 forresta joined #salt
04:34 jgelens joined #salt
04:35 yml joined #salt
04:36 mlen joined #salt
04:36 marcel joined #salt
04:48 drawsmcgraw left #salt
05:04 tollmanz joined #salt
05:10 ravibhure joined #salt
05:16 tomtomtomtom joined #salt
05:25 deepakmdass joined #salt
05:28 sroegner joined #salt
05:45 Linz joined #salt
05:49 NV joined #salt
05:50 deepakmdass joined #salt
05:53 ajw0100 joined #salt
05:54 bhosmer joined #salt
05:59 anuvrat joined #salt
06:02 toastedpenguin joined #salt
06:02 deepakmdass joined #salt
06:14 oz_akan_ joined #salt
07:06 tollmanz joined #salt
07:06 taion809 joined #salt
07:21 rojem joined #salt
07:28 sroegner joined #salt
07:28 rmt joined #salt
07:35 tollmanz joined #salt
07:37 tollmanz_ joined #salt
07:39 tollmanz joined #salt
07:40 matanya joined #salt
07:41 Ryan_Lane joined #salt
07:43 bhosmer joined #salt
07:43 jfzhu_us joined #salt
07:45 taion809 joined #salt
07:58 Shenril joined #salt
08:01 TomasNunez joined #salt
08:06 ggoZ joined #salt
08:09 anuvrat joined #salt
08:09 bhosmer joined #salt
08:26 harobed joined #salt
08:32 stoffell I want to force a package (on debian minion) to install (so it takes my salt-fileserver deb package instead of the outdated apt repository) but it won't install because it's already installed. is there a workaround for this?
08:33 malinoff stoffell, uninstall the package at first :)
08:35 stoffell but it won't do that everytime then, will it ? (i've got my minions scheduled)
08:37 malinoff stoffell, well, you can specify a version you need to install
08:37 malinoff http://docs.saltstack.com/ref/states/all/salt.states.pkg.html?highlight=pkg.installed#salt.states.pkg.installed
08:37 malinoff "Version" argument
08:37 malinoff "version" argument*
08:38 tollmanz joined #salt
08:38 stoffell ok, i will try that, thanks !
08:40 logix812 joined #salt
08:42 nn0101 joined #salt
08:43 helminthe joined #salt
08:43 xxxqqq joined #salt
08:48 xxxqqq :)
09:10 harobed_ joined #salt
09:28 toastedpenguin joined #salt
09:31 bhosmer joined #salt
09:38 tollmanz joined #salt
09:40 helminthe joined #salt
09:40 tollmanz_ joined #salt
09:44 pengunix joined #salt
09:50 elfixit joined #salt
10:21 bhosmer joined #salt
10:21 cyberbob has someone used salmon for monitoring purpose, i want to knw about the feedback as we are thinking to use this in an opersource project
10:24 cyberbob https://github.com/lincolnloop/salmon/
10:34 taion809 joined #salt
10:38 tollmanz joined #salt
10:39 mlen joined #salt
10:40 rmt cyberbob, salmon indeed looks nice and simple - so ideal for a very small scale operation. If that's your use-case, go for it - it doesn't sound like it would involve much time-investment to get it running.
10:41 rmt And at a small scale, you can change again if ever you grow.
10:41 Furao I find that shinken is good at both large and small deployment
10:48 deepakmdass joined #salt
10:48 rmt In other news, I've now got working userdata at http://169.254.169.254/latest/user-data for my virtualisation stack.
10:49 rmt I'm thinking of adding salt-virt as a driver for it.
10:49 rmt Then you could actually track what you've got deployed (and where), even when the servers die.
11:00 cyberbob rmt actually we will be doing some website monitoring for a few clients (just checking if website up/running, keywords are found etc no anyother resources like harddisk/processor/ram etc). there will be about 5K (http/https checks) sites. So would this be a good idea to use salmon (as it looks nice but I am not a core developer but an operations guy who will be working on this project)
11:01 cyberbob so infact clientless monitoring .
11:01 cyberbob but sometimes I think using only urlib/urlib2 of python will work in my case
11:02 rmt Shinken looks interesting too -- and looks like it might scale better.  Seems to support both active and passive checks.
11:03 gasbakid joined #salt
11:05 cyberbob rmt I have used nagios (from which shinken is drived) Actually in mycase shinkins will be a big deal (as I am not going to monitor all resources and only http/https with clientless approach) and also shinken doesn't have support to isolate the dashboard for each user (like 5 different customers)
11:06 rmt Sounds like you're in a good position to evaluate solutions yourself. :)
11:08 cyberbob rmt :)
11:08 cyberbob Actually I would never like to re-invent the wheel so looking if some opensource project will work in mycase else creating everything from scratch will be the only solution which will hurt a CloudOperations Guy a lot:(
11:10 holys joined #salt
11:16 taion809 joined #salt
11:19 bhosmer_ joined #salt
11:21 holys joined #salt
11:23 taion809 joined #salt
11:38 tollmanz joined #salt
11:43 toastedpenguin joined #salt
11:45 bhosmer joined #salt
11:53 xl1 left #salt
11:54 NV joined #salt
12:00 taion809 joined #salt
12:05 dangra joined #salt
12:16 psyl0n joined #salt
12:25 taion809 joined #salt
12:36 cym3try joined #salt
12:38 tollmanz joined #salt
12:38 holys joined #salt
12:52 psyl0n joined #salt
12:58 31NAAG4NU joined #salt
13:01 blee joined #salt
13:07 bhosmer joined #salt
13:11 anuvrat joined #salt
13:15 toastedpenguin joined #salt
13:20 Psi-Jack How could i use pkg.latest to upgrade a package on my servers on command-line, not state file?
13:22 mgw joined #salt
13:23 __number5__ salt-call pkg.upgrade
13:25 __number5__ normally state will have a module counterpart, which you can call using command line
13:27 blee_ joined #salt
13:33 justBob joined #salt
13:37 che-arne joined #salt
13:38 tollmanz joined #salt
13:41 mapu joined #salt
13:42 gadams999 joined #salt
13:44 NV joined #salt
13:46 pengunix joined #salt
13:47 pengunix2 joined #salt
13:51 sfvivek joined #salt
13:52 JasonSwindle joined #salt
13:53 iShming joined #salt
13:53 njpatel joined #salt
13:54 sfvivek2 joined #salt
13:59 gasbakid joined #salt
13:59 njpatel joined #salt
14:04 Gifflen joined #salt
14:04 njpatel joined #salt
14:07 gasbakid joined #salt
14:07 oz_akan_ joined #salt
14:07 viraptor joined #salt
14:08 njpatel joined #salt
14:08 oz_akan_ joined #salt
14:09 viraptor hi all, I've got some issue with salt 0.17.4 - I'm trying to configure a pillar extension, but it fails with "[ERROR   ] Failed to load ext_pillar s3_pillar: ext_pillar() got multiple values for keyword argument 'access_key'"
14:09 viraptor which is wrong - the 'val' is actually a dictionary, so there's just no way it can have multiple values
14:10 viraptor any ideas what may be going wrong?
14:10 njpatel joined #salt
14:11 brianhicks joined #salt
14:13 psyl0n joined #salt
14:15 blarghmatey joined #salt
14:16 vejdmn joined #salt
14:22 Gifflen joined #salt
14:22 pdayton joined #salt
14:28 iShming joined #salt
14:29 njpatel joined #salt
14:31 bhosmer joined #salt
14:35 che-arne joined #salt
14:35 toastedpenguin joined #salt
14:35 anuvrat joined #salt
14:35 31NAAG4NU joined #salt
14:35 gamingrobot joined #salt
14:35 bwq joined #salt
14:37 viraptor ah, api changed and includes minion_id now... that explains it
14:38 tollmanz joined #salt
14:49 pdayton1 joined #salt
14:50 william_20111 joined #salt
14:53 mapu joined #salt
14:56 bhosmer_ joined #salt
14:57 pdayton joined #salt
15:01 jtang joined #salt
15:02 sroegner joined #salt
15:08 djbclark joined #salt
15:08 djbclark joined #salt
15:09 dangra joined #salt
15:15 psyl0n joined #salt
15:17 iShming joined #salt
15:19 higgs001 joined #salt
15:30 forresta joined #salt
15:32 th3reverend joined #salt
15:33 munderwo joined #salt
15:38 tollmanz joined #salt
15:39 th3reverend left #salt
15:49 koupreev joined #salt
15:55 srage joined #salt
15:56 dangra joined #salt
15:57 [diecast] joined #salt
16:03 jtang joined #salt
16:03 logix812 joined #salt
16:07 lineman60 joined #salt
16:08 tollmanz joined #salt
16:10 rojem joined #salt
16:13 djbclark joined #salt
16:13 djbclark joined #salt
16:13 danielbachhuber joined #salt
16:14 pnl_ joined #salt
16:18 gadams999 joined #salt
16:19 mgw joined #salt
16:20 crane_ joined #salt
16:21 jalbretsen joined #salt
16:22 neilf_ joined #salt
16:22 JasonSwindle Howdy, all
16:22 Kraln joined #salt
16:23 mjulian joined #salt
16:23 seanz joined #salt
16:23 dpac|away joined #salt
16:23 [vaelen] joined #salt
16:23 nliadm joined #salt
16:23 1JTABNTYP joined #salt
16:23 Jahkeup joined #salt
16:23 mjulian joined #salt
16:24 dpac|away joined #salt
16:24 seanz joined #salt
16:24 psyl0n joined #salt
16:26 or1gb1u3 joined #salt
16:26 psyl0n joined #salt
16:26 MTecknology joined #salt
16:27 MTecknology joined #salt
16:32 echos joined #salt
16:33 sirtaj joined #salt
16:36 munderwo joined #salt
16:40 gladiatr joined #salt
16:43 NV joined #salt
16:44 JordanRinke Morning, in the morning this time.
16:46 forresta hey
16:47 grep_awesome Off the wall question, has anyone integrated salt with Go (the language)?
16:47 forresta in what way grep_awesome
16:48 grep_awesome forresta: api bindings or otherwise
16:48 grep_awesome I'm really asking in the most general sense if anyone has played with the two together.
16:48 forresta grep_awesome, ahh, I haven't seen anyone doing that.
16:48 forresta at least not in the IRC
16:48 srage joined #salt
16:48 forresta are you actively coding in go?
16:48 grep_awesome Strongly considering it for some tasks in our shop
16:49 forresta interesting, what edge cases/problems does it solve that aren't able to be accomplished with python?
16:49 JordanRinke Speeed ssppeeeeeed
16:49 forresta well yea other than that
16:49 dcolish you really dont want to do language level interop between go and python unless its absolutely necessary
16:49 JordanRinke channel fan ins are pretty hot too
16:49 grep_awesome forresta: most of it can be done with python, but the attractive thing about go is that you build a standalone binary with no dependencies
16:50 forresta fair enough
16:50 grep_awesome the concurrency features are insane too
16:50 lineman60 joined #salt
16:51 forresta I wish it wasn't so ugly
16:52 grep_awesome eh, it's meant to not be completely alien to C developers, I think
16:52 TomasNunez1 joined #salt
16:52 forresta yea that's the first thing it reminded me of
16:52 dcolish its C interop is great
16:52 JordanRinke I have a few projects where Go is specifically useful because of its channel mechanism but mostly I find Python a better language from tooling, support etc. However if I had a python service that was straining under load and had no visual elements, I would consider a Go migration for whatever piece was causing the bottle neck
16:52 dcolish you can easily use structs since the memory layout of a go object is the same
16:53 dcolish i
16:53 forresta JordanRinke, you'd go with go over cython?
16:53 blee joined #salt
16:53 dcolish that ^
16:54 grep_awesome I find cython to be a bit of a cludge
16:54 dcolish you're better off just writing cython or C for optimization
16:54 JordanRinke Yeah, for sure. Go wwaaayyy out performs pretty much anything else for concurrent connections / request rates etc.
16:54 forresta gotcha
16:54 dcolish :/
16:55 JordanRinke http://www.techempower.com/benchmarks/#section=data-r8&hw=i7&test=json
16:55 dcolish thats a bit unfounded
16:55 JordanRinke ^
16:55 dcolish i believe you're exaggerating the improvements of performance there a bit
16:56 MTecknology joined #salt
16:56 [vaelen] joined #salt
16:56 Kyle joined #salt
16:56 dpac|away joined #salt
16:56 Jahkeup joined #salt
16:56 MTecknology joined #salt
16:56 Heartsbane joined #salt
16:56 seanz joined #salt
16:56 dcolish i also suspect that go wins for json since the type is built into the language and heavily optimized
16:56 dcolish it does not win for anything else
16:57 JordanRinke It is literally so fast, the packets go back in time and arrive before the request is even made.
16:57 qqqqq joined #salt
16:57 JordanRinke It doesn't win for anything else but it is the top overall performer
16:57 dpac|away joined #salt
16:57 qqqqq left #salt
16:57 dcolish where is the overall listing?
16:57 Heartsbane joined #salt
16:58 JordanRinke There was a write up about it, using that data as a source, I don't remember where though
16:59 KyleG joined #salt
16:59 KyleG joined #salt
16:59 forresta so in short grep_awesome, no :P
16:59 grep_awesome lol
16:59 dcolish well you could use salt-api
16:59 dcolish although thats probably not what you really meant
16:59 JordanRinke I just remember the source data because flask (which most of my web projects are in, with uwsgi) is so low :(
16:59 Kyle- joined #salt
16:59 forresta yea he was wondering if there was a go api though
16:59 dcolish right
17:00 grep_awesome dcolish: re: salt-api. Yeah that might be a fun project on a slow day
17:00 forresta JordanRinke, are you running it behind anything?
17:00 KyleG JordanRinke y u break salt
17:00 JordanRinke In my setup? I run nginx+uwsgi+flask
17:01 JordanRinke nginx primarily for local file serving and slow connection protection really
17:01 dcolish JordanRinke: yes wsgi can be pretty slow depending on your application design
17:01 JordanRinke aalllttthoouggh I have a sweeet project I am working on
17:01 KyleG We're using UWSGI for our API and Auth and like it so far. Using NGINX for the reverse proxy
17:01 grep_awesome I'd like to see a nice, lightweight, native monitoring service that I can stick on all of our ec2 instances. kind of the way rackspace uses lua+libuv (luvit project) to monitor their vms without a resource impact
17:01 forresta man, I can't find the article some guy wrote about performance metrics..
17:01 JordanRinke my stuff is fast, I don't have a problem with it - but I preach flask as the best framework so I get those benchmarks rubbed in my face locally a lot lol
17:02 dcolish imho, best is not just about performance
17:02 JordanRinke grep_awesome: have you seen stalker?
17:02 grep_awesome JordanRinke: negative
17:02 forresta grep_awesome, did you look at http://stackd.io/ at all? I don't know if it does monitoring though...
17:02 dcolish grep_awesome: i
17:02 dcolish damn it
17:02 dcolish i've seen diamond used, its ok
17:02 JordanRinke https://github.com/pandemicsyn/stalker
17:03 forresta abemusic isn't in here right now, but they are planning on open sourcing that soon, I don't know if they plan to have monitoring capabilities.
17:03 dcolish why not collectd?
17:03 grep_awesome JordanRinke: I'll check that out, thanks.
17:03 JordanRinke I am not saying Rackspace is using that, since I don't work there any more, but something just like that, might be being used to monitor thousands of servers somewhere
17:03 forresta JordanRinke, people are rubbing it in your face that you are using flask?
17:03 grep_awesome dcolish: haven't seen that, but it doesn't look like monitoring. does look neat though
17:03 forresta JordanRinke, So build out an entire tool in Flask in an afternoon, then ask them to complain about the speed.
17:04 JordanRinke forresta: yeah, light heartedly, I grief them about nodejs all the time too
17:04 forresta I don't really like nodejs
17:04 JordanRinke (I am officed in a hackerspace)
17:04 forresta so I'm obviously biased
17:04 JordanRinke so we all mess with eachother
17:04 forresta gotcha
17:05 JordanRinke I feel sorry for the Java guys though... they get it the worst
17:05 dcolish ?
17:05 Kyle- joined #salt
17:05 JordanRinke we laugh and point
17:05 dcolish ah
17:05 forresta *shrug*
17:05 dcolish well you shouldnt
17:05 forresta lol
17:05 dcolish the jvm is quite useful
17:06 JordanRinke Heresy!
17:06 forresta if only the jvm wasn't totally incompetent and realized when it was about to ruin the server.. :(
17:07 JordanRinke Yeah, I actually don't have a problem with Java itself, but maintaining the inf to run it, is realllly annoying
17:07 forresta yep
17:07 mjulian joined #salt
17:07 mjulian joined #salt
17:08 forresta honestly if the JVM just had more transparency I wouldn't hate it so much. I want to point at crappy code and say 'look here is the problem, please fix it'. Instead of constantly watching the blame fall on servers.
17:09 dcolish what do you mean by that?
17:09 dcolish I run java apps a lot and they do not require any more inf than python apps
17:09 forresta dcolish, it's mostly related to poor error reporting by developers.
17:10 dcolish oh well you can get that in any language
17:10 forresta I can't go in and change the WAR then redeploy it to see what is going on when code fails
17:10 dcolish python stacktraces are not exactly wonderful
17:10 APLU joined #salt
17:10 forresta I don't disagree
17:10 dcolish oh dude, WARs are terrible
17:10 dcolish i dont use them :)
17:10 JordanRinke wwhhhaattt, Java... same inf reqs as Python? I think not sir/ma'am... I think not.
17:10 forresta well, there's the different then buddy :P
17:10 dcolish j2ee == avoid
17:10 dcolish thats like 10 year old tech
17:11 JordanRinke you mean.. Java?
17:11 dcolish do not want
17:11 JordanRinke Java
17:11 forresta man JordanRinke, poking with a stick here
17:11 JordanRinke just completing your sentence.
17:11 dcolish java is fine
17:11 dcolish its as terrible as python or c++
17:11 forresta I don't mind good java, just like I don't mind good python, or good ruby, or good c++
17:11 dcolish the fact that its fast and has good tooling helps
17:12 JordanRinke I only write bad python, it is easier.
17:12 forresta note to self, never work with JordanRinke.... :P
17:12 JordanRinke one letter variable names save so much time typing!
17:12 dcolish fortran ftw
17:13 forresta lol one letter variables
17:13 forresta I'd seriously punch a coworker if they did that
17:13 JordanRinke speaking of, any of you keyboard geeks? geekhack members?
17:13 forresta just a light punch
17:13 forresta but still
17:13 JordanRinke I use one letter variable names, and if you read them top to bottom, they are in insult
17:13 forresta hah
17:14 JordanRinke not really, but that is a good idea
17:14 JordanRinke or well, a terrible idea, but entertaining
17:14 viraptor hi, is there some way to create a fake state doing nothing apart from organising other states? for example "php cli stuff" that requires "php cli packages", "php cli config", ... and can be used to require it in other places?
17:14 forresta viraptor, yea sure
17:14 forresta use include http://docs.saltstack.com/ref/states/include.html#include
17:15 forresta so you would include other states there
17:15 forresta and then in your other item, just include that one state, which will daisy chain the others
17:15 JordanRinke include, and then require or require in for your ordering (or set an order)
17:15 viraptor forresta: but includes don't influence ordering, do they?
17:15 forresta I'd suggest to be careful with that though, getting too deep in the include chain can result in problems when you make changes down the road.
17:16 JordanRinke viraptor: includes dont but require/require_in does
17:16 forresta viraptor, your included states should properly do the ordering, they're supposed to get executed in the order you include them in.
17:16 forresta JordanRinke, actually includes are supposed to be executed in the order you include them in, UtahDave informed me of that a while back
17:16 JordanRinke forresta: unless they have a require, or require_in, in them somewhere no?
17:17 JordanRinke include = assumed ordering, require = explicit ordering
17:17 forresta right, but if you had a require of something else, that state would have to be included in the state doing the require anyways :P
17:17 dcolish damn, thats code i have not read in a long time
17:17 viraptor I'm actually already doing the include, but I was looking for a way to explicitly do "require: - some_fake: php cli stuff", rather than requiring each part separately
17:18 forresta you can require a state
17:18 forresta is that what you mean?
17:18 viraptor let me pastebin it, just a second :)
17:18 forresta ok
17:19 JordanRinke netmask 255.240.0.0 seems like a good idea to me
17:21 viraptor forresta: http://pastebin.com/z7mP9nGz maybe that explains it better
17:21 viraptor "php cli *" parts already come from an include, but I wanted to require them explicitly to avoid accidents in the future
17:22 forresta so you want to turn those 3 requires into a single group
17:22 viraptor yes
17:23 forresta are the pkg, file, and 'something else' all in the same state?
17:23 viraptor at the moment, yes; may not be in the future
17:23 srage joined #salt
17:24 forresta hmm, ok then my first suggestion of grouping them under the same ID isn't a good one
17:26 Veticus joined #salt
17:27 Veticus joined #salt
17:30 forresta how about this viraptor https://gist.github.com/gravyboat/8199967
17:31 deepakmdass joined #salt
17:33 forresta I guess you could also do it where you have 'fake_collection' as you have at the bottom, but you include that sls that includes all the other items.
17:33 thrashr888 joined #salt
17:34 viraptor ah, I didn't know you can actually require on a whole sls file, that wolves it :)
17:34 viraptor solves even
17:34 forresta yea requiring an entire sls is awesome
17:34 forresta you just have to make sure it's included, otherwise salt won't know what you want
17:36 harobed_ joined #salt
17:39 william_20111 joined #salt
17:39 Hefeweizen joined #salt
17:40 forresta Gareth, Jeez man, get better at duplicating problems! Such a slacker... :P
17:41 45PAAYCHD joined #salt
17:44 foxx joined #salt
17:46 tollmanz joined #salt
17:48 rojem joined #salt
17:51 Gareth forresta: haha seriously.
17:53 thrashr888 joined #salt
17:54 JasonSwindle Gareth:  did that help?
17:54 JasonSwindle Your example was pretty much dead on
17:55 ashtonian joined #salt
17:56 Gareth JasonSwindle: yeah.  helping a lot.  just adding in your examples now.
17:56 JasonSwindle That example I gave is a little but larger, but pretty much all pillar foo
17:59 tollmanz joined #salt
18:00 Gareth JasonSwindle: just to clarify something I noticed.  So your requirement is using the path {{ pillar['document_root'] }}/{{ pillar['project_name'] }}/{{ pillar['project_name'] }}-repo but your cwd is {{ pillar['document_root'] }}/{{ pillar['project_name'] }}, is that where the requirements directory and other requirements files are or are they under repo as well?
18:01 JasonSwindle I do need to fix that, we moved some stuff after we made that issue.
18:02 JasonSwindle cwd should read
18:02 JasonSwindle - cwd: {{ pillar['document_root'] }}/{{ pillar['project_name'] }}/{{ pillar['project_name'] }}-repo
18:02 Gareth okay cool
18:02 JasonSwindle Let me re highstate again
18:06 JasonSwindle Gareth:  updated gist
18:06 JasonSwindle Good eye
18:09 Gareth JasonSwindle: ran with your example with the updated cwd....ran fine.
18:09 forresta are you on develop Gareth?
18:11 Gareth yeah. I am.  maybe its an issue thats alreayd been fixed :)  Could be related to the pip version too.
18:11 JasonSwindle I am on 0.17.4
18:11 Gareth lemme test on 0.17.4
18:12 JasonSwindle okie
18:12 forresta Yes, look at my amazing contribution to this issue
18:12 forresta :P
18:13 JasonSwindle :)
18:13 pdayton joined #salt
18:13 rojem joined #salt
18:15 [diecast] joined #salt
18:19 srage joined #salt
18:21 rojem joined #salt
18:24 zgre joined #salt
18:24 mackstick joined #salt
18:24 cyp joined #salt
18:32 bhosmer_ joined #salt
18:33 munderwo joined #salt
18:35 paul joined #salt
18:36 AdamSewell joined #salt
18:44 sirtaj joined #salt
18:44 tempspace joined #salt
18:46 srage joined #salt
18:47 gadams999 left #salt
18:48 gadams9991 joined #salt
18:48 gadams9991 left #salt
18:48 jtang joined #salt
18:49 Gareth JasonSwindle: tested on 0.17.4 ran fine.  what version of python and what version of pip are you using?
18:50 rojem joined #salt
18:51 viraptor where should I drop custom new states when I want them to be found by salt? (like pillar extensions go into extension/pillar/...)
18:52 JasonSwindle pip --version
18:52 JasonSwindle pip 1.0 from /usr/lib/python2.7/dist-packages (python 2.7)
18:52 JasonSwindle On the salt master
18:52 JasonSwindle python --version
18:52 JasonSwindle Python 2.7.3
18:52 JasonSwindle On the master
18:53 Gareth hm. the machine I was using for the master in this case is python 2.7.6 with pip 1.4.1, the client was python 2.7.3 and pip 1.1
18:54 deepakmdass joined #salt
18:54 Gareth but you had said it worked fine manually right?  Just not in salt?
18:54 JasonSwindle If I point salt to req/base.txt it works
18:55 jfzhu_us joined #salt
18:55 forresta viraptor, http://docs.saltstack.com/ref/states/writing.html#using-custom-state-modules
18:55 Gareth JasonSwindle: pointing at requirements.txt fails right?
18:56 JasonSwindle yes
18:56 JasonSwindle I get the /tmp fail
18:56 JasonSwindle but if I point to base.txt in the requirements folder, it works
18:56 Gareth very strange.
18:56 JasonSwindle I agree, totally
18:56 Gareth what distribution are those machines running?
18:57 JasonSwindle Ubuntu 12.04
19:03 jfzhu_us joined #salt
19:04 Gareth JasonSwindle: going to spin up some 12.04 vms and see if I can duplicate the problem.
19:05 JasonSwindle OK, thank you! :)
19:06 Gareth no worries.  now its a mission :)
19:10 JasonSwindle lol
19:10 JasonSwindle To let you know, I am working on this to open source it
19:11 JasonSwindle Right now it is in a bit of a flux, but should be golden by SaltConf
19:11 Gareth nice
19:12 thrashr888 joined #salt
19:20 seanz left #salt
19:21 jtang joined #salt
19:21 altj joined #salt
19:26 ashtonian joined #salt
19:35 admineric joined #salt
19:40 pass_by_value joined #salt
19:45 chuffpdx joined #salt
19:52 ajw0100 joined #salt
19:56 JasonSwindle Gareth: Any luck?
19:59 psyl0n joined #salt
20:03 cdcalef- joined #salt
20:05 thrashr888 joined #salt
20:06 viraptor joined #salt
20:21 bhosmer_ joined #salt
20:21 Gareth JasonSwindle: sorry, was away for a bit.  testing now.
20:21 JasonSwindle np, just checking in. :)
20:33 blarghmatey joined #salt
20:33 tollmanz joined #salt
20:36 JasonSwindle anyone here use the postgres_user state in Saltstack 0.17.4?
20:36 xmltok joined #salt
20:37 forresta joined #salt
20:39 dangra joined #salt
20:42 Gareth JasonSwindle: question for you.  The lines in your require...the pkg and pip lines.  You've defined those somewhere eles?
20:43 JasonSwindle Yes
20:43 JasonSwindle This is a very large deployment
20:43 JasonSwindle lol
20:46 Gareth just tested it.  installed the virtualenv package manually...but it worked.
20:46 lineman60__ joined #salt
20:46 JasonSwindle Hmmm, ok.
21:02 lineman60__ joined #salt
21:05 thrashr888 joined #salt
21:07 srage joined #salt
21:09 tollmanz joined #salt
21:10 Gareth JasonSwindle: your user...the one that you're using to setup the virtualenv as...the home directory isn't /tmp is it?
21:11 JasonSwindle Nope, sadly not.  It should be /opt/apps/appuser or whatever is set in pillar
21:11 mgw joined #salt
21:11 JasonSwindle Gareth:  What I use to create the user https://gist.github.com/JasonSwindle/408e2b22bec58d67ea89
21:13 hunternet93 joined #salt
21:13 thrashr888 joined #salt
21:17 hunternet93 Time for a round of "What's wrong with this SLS?": http://pastebin.com/YvNZemwv I'm trying to run a command if either of two files have changed. I'm running Salt trunk, it appears that cmd.wait has been removed so I'm trying cmd.run.
21:20 forresta cmd.wait was not removed to my knowledge
21:20 hunternet93 It doesn't show up under sys.doc for some reason, but I get the exact same error using cmd.wait instead of cmd.run
21:21 forresta what happens when you remove those single quotes?
21:21 forresta also your wait should be indented.
21:21 forresta err wawtch
21:21 forresta *watch
21:21 forresta man can't type today
21:22 hunternet93 Same error. I've tried quoting all strings, unquoting all strings, rearranging it all...
21:22 forresta indent watch properly
21:22 hunternet93 Should watch start with a '-'?
21:22 forresta yep
21:23 hunternet93 Same error... *sigh*
21:23 forresta so it looks like this now? https://gist.github.com/gravyboat/8202354
21:23 forresta can you post the whole state?
21:23 hunternet93 Sure, just a sec
21:24 hunternet93 Pastbin or gist?
21:24 forresta whatever you prefer
21:25 tollmanz joined #salt
21:26 JasonSwindle forresta:  Do you use postgres_user state?
21:26 forresta no
21:26 hunternet93 Got it: https://gist.github.com/hunternet93/bf0a92f74a5a7cf14f5d
21:27 Gifflen joined #salt
21:27 hunternet93 Actually, I think I spotted my problem...
21:27 forresta you're missing some '-'
21:28 JasonSwindle forresta:  darn……. I am having an odd issue.
21:28 forresta JasonSwindle, One of the downsides of only using salt at home, I don't get too crazy in my configuration
21:28 hunternet93 I indented the contents of a jinji if block that shouldn't have been
21:28 forresta ahh
21:28 JasonSwindle Ah, fun.   I just love stack traces
21:29 forresta you are also missing a - on some of your requires it looks like hunternet93
21:29 hunternet93 forresta: sure? it seems to work OK, it just ran.
21:29 forresta *shrug* just what I'm noticing
21:30 forresta maybe you don't have to have the dashes, you're missing a lot of them
21:30 hunternet93 forresta: you're probably right, i'm still pretty new to salt
21:30 forresta well, you're mixing pkg.installed, and the indented versions
21:30 forresta I'd justd do a -l debug and make sure it is applying like you expect.
21:33 munderwo joined #salt
21:36 hunternet93 forresta: i reformatted the requires and switched to cmd.wait, it works fine now. Thanks!
21:36 forresta awesome, np you pretty much found it yourself :P
21:37 hunternet93 nothing like asking for help to help you spot your own mistakes!
21:37 forresta yea that helps a lot of the time
21:39 lineman60 joined #salt
21:39 forresta as a heads up hunternet93, this is what it looks like when you don't mix the indenting: https://gist.github.com/gravyboat/8202446
21:39 forresta like pkg.installed, versus pkg: \n - installed
21:40 forresta there should probably be more examples like that on the docs really
21:44 hunternet93 left #salt
21:49 JasonSwindle Anyone want to help take a stab at this stack trace? https://gist.github.com/JasonSwindle/8595f4af7fbe2c1093c8
21:49 JasonSwindle Fails the first HighState, works the second time around
21:50 JasonSwindle in the second time around, postgres_user is green like it had to take no action
21:51 forresta where is the group JasonSwindle?
21:52 JasonSwindle group?
21:52 forresta oh it's postgress, duh
21:52 JasonSwindle :)
21:53 forresta so it requires the postgres_install package, but is the DB up and running by the time it getse here?
21:53 JasonSwindle DB comes next
21:53 JasonSwindle DB has     - require:
21:53 JasonSwindle - postgres_user: app_database_user
21:54 forresta ok maybe I just don't know much about postgres then, so you can add a postgres user before the database service is up?
21:54 JasonSwindle What is funny is, this worked in 0.16.x, and this works if I HS again.  LOL
21:54 forresta right, because something is happening after this runs, that is allowing the second highstate to work
21:55 JasonSwindle Yep,  I can flip the order…..
21:55 JasonSwindle Let me try that
21:56 forresta ok, I just don't know enough about postgres to know how it handles adding users, how do you add a user to a database that isn't up?
21:56 anuvrat joined #salt
21:57 torrancew you can add a user before the db is up, but you can't associate the two until the db is up IIRC
21:59 JasonSwindle but should it error that way?  To me, getting a stack trace is :(
22:00 forresta is the password a string or an int?
22:00 JasonSwindle In pillar I have:
22:00 JasonSwindle dbpassword: 0123456789
22:01 forresta can you wrap that in single quotes?
22:01 JasonSwindle Sure!
22:01 JasonSwindle Now I need to re-build my nodes….. lol
22:01 forresta Well, start making the code work and you won't need to, jeez!
22:02 JasonSwindle LOL :)
22:04 MedicalJaneParis joined #salt
22:09 bhosmer_ joined #salt
22:11 forresta JasonSwindle, can you try wrapping the db password before making the other change we talked about?
22:11 forresta I'd prefer to do one at a time, and that python error looks suspiciously like something is getting screwed up when passed to the module or something
22:12 jchen left #salt
22:13 JasonSwindle forresta:  single-quota wrapping in pillar failed
22:13 forresta same error?
22:13 JasonSwindle Yepper TypeError: string indices must be integers, not str
22:14 forresta I don't understand, that's like it's trying to take asdf['zcxv'] or something...
22:14 che-arne joined #salt
22:15 JasonSwindle I don't really understand what is going on either....
22:17 [diecast] joined #salt
22:20 lineman60__ joined #salt
22:27 Gareth JasonSwindle: looks like the error is because the call to psql inside the module isn't running correctly.
22:28 psyl0n joined #salt
22:28 JasonSwindle Do I need to open an issue on this?
22:29 Gareth yeah. I'd say so.  based your state it looks like it should work.
22:29 JasonSwindle Ok
22:32 Gareth JasonSwindle: btw, I tried the virtualenv state using pip to install virtualenv.  same results.  still worked.  to be it sounds like there is some reference to /tmp somewhere in the state or the pillar data, somewhere, thats being picked up somehow.
22:32 JasonSwindle hmmm odd
22:33 JasonSwindle Gareth:  I do have in my pip.conf this
22:33 JasonSwindle download-cache = /tmp/.pip/cache
22:33 JasonSwindle but that should not do that.....
22:34 JasonSwindle [global]
22:34 JasonSwindle default-timeout = 120
22:34 JasonSwindle download-cache = /tmp/.pip/cache
22:34 Gareth yeah...but you're seeing it look for the requirements under /tmp  not /tmp/.pip.cache
22:34 JasonSwindle yea
22:35 JasonSwindle And all of the other stuff with /tmp is CeleryD
22:35 JasonSwindle hmmmm.
22:36 Corey Seems a bit curious.
22:38 bhosmer joined #salt
22:41 JasonSwindle basepi:  Got a new one for you guys. :)  https://github.com/saltstack/salt/issues/9516
22:49 Gareth JasonSwindle: just tested your postgres states...got the same error.
22:49 JasonSwindle YAY…… it is just not me on this one
22:49 JasonSwindle The other……. I will mess with it later
22:50 oz_akan_ joined #salt
22:50 Gareth so this is one where it runs through....errors....then runs again but works right?
22:52 JasonSwindle Yep
22:52 JasonSwindle First run, fail
22:52 lineman60 joined #salt
22:52 JasonSwindle Second run, passes and the erroring state is in Green
22:52 JasonSwindle Eh?
22:54 JasonSwindle Anywho, I am heading out.   Bye guys!
22:54 Gareth It looks like it does work the user gets added....within the code its expecting a return to look a certain way and its not.  hence the error.
22:54 Gareth JasonSwindle: later.  good luck :)
22:55 JasonSwindle THank you for all of your help! :)  Bye!
22:59 thrashr888 joined #salt
23:00 timoguin joined #salt
23:00 eculver joined #salt
23:00 psyl0n joined #salt
23:02 timoguin anyone have experience launching VMs on AWS with salt-cloud where it just hangs at "waiting_for_ssh"?
23:03 timoguin this is my first time trying to use salt-cloud with AWS
23:05 timoguin hmmm, might be my security groups
23:06 Corey timoguin: Instances. Not VMs.
23:06 timoguin pssshhh
23:07 kermit joined #salt
23:22 timoguin twas my sekurrity groups
23:23 markizano timoguin: I"ve noticed that salt has issues getting the first state run if it can't establish a ssh connection..
23:23 markizano **timoguin: I"ve noticed that salt has issues getting the first state run if ssh isn't running
23:24 timoguin yea, i had to change my AWS security group to allows inbound ssh, plus a permissions error with my local .pem key
23:24 timoguin working now
23:32 mgw joined #salt
23:40 thrashr888 joined #salt
23:44 bhosmer joined #salt
23:51 lineman60 joined #salt
23:54 jtang joined #salt
23:56 sroegner joined #salt
23:57 bhosmer_ joined #salt

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