Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-12-27

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

All times shown according to UTC.

Time Nick Message
00:15 druonysuse joined #salt
00:15 druonysuse joined #salt
00:22 fllr joined #salt
00:25 druonysuse joined #salt
00:31 fllr joined #salt
01:03 bhosmer joined #salt
01:16 MZAWeb joined #salt
01:21 Psi-Jack joined #salt
01:23 cdcalef can a minion with different IDs share the same key?  can i have a key be 'preauthenticated' no matter how many minions its used on?
01:27 _ikke_ joined #salt
01:33 forresta cdcalef, I don't think that would work, how would different grains or values be set if they all had the same ID? There would constantly be data conflicts.
01:34 Daviey joined #salt
01:34 cdcalef different ID but same authentication key
01:35 cdcalef although i presume salt does the sane thing and looks up the key to authenticate against based off the ID
01:35 forresta Ahh, hmm, I still see it as problematic since the key and ID are 'tied' together
01:35 forresta I'm not sure how that portion of the auth check works though
01:38 mgw joined #salt
01:50 newbie joined #salt
01:50 Guest88690 any plans to add support to for open shift
01:51 sarasfox has any talked to redhat about add salt suppor to openshift online
01:52 forresta sarasfox, not that I'm aware, nor have I seen any topics regarding it
01:52 Nazca__ joined #salt
01:53 sarasfox who do you thing at salt stack could talk to redhat about this
01:53 pcurry_nomi joined #salt
01:54 forresta sarasfox, I don't know who works with vendors for that sort of integration, are you just talking about states for it or salt-cloud integration or what?
01:55 forresta because when I look at the items such as https://github.com/openshift/puppet-openshift_origin
01:55 forresta seems like they're just using it like anything else, install the config management tool, run it.
01:56 sarasfox you can't install into a gear
01:56 forresta correct me if I'm wrong but aren't gears just VMs?
01:56 forresta I'm not very familiar with openshift
01:57 sarasfox in openshift online yum and rpm are turned off
01:58 forresta isn't that just the freebie version?
01:58 forresta how do you install software on a gear if everything is disabled?
01:58 forresta checkboxes before it gets built or something?
02:00 forresta oh, cartridges? (sorry I am googling this as we talk)
02:00 munderwo joined #salt
02:00 sarasfox forresta: gears are like vm's
02:01 forresta right, but you can't install anything on them, I'm looking at the 'build a cartridge' doc now
02:01 forresta well, install via yum/rpm from what you were saying
02:01 sarasfox hum use git push to install stuff via git
02:02 forresta so it's like heroku
02:02 forresta aka, don't let your developrs have access to the machine because they break everything constantly and don't fix it? :P
02:02 forresta *developers
02:02 tollmanz joined #salt
02:03 forresta The best place to get that sort of dicsussion started would be here: https://groups.google.com/forum/#!forum/salt-users
02:12 psyl0n joined #salt
02:12 favadi joined #salt
02:16 jfzhu_us joined #salt
02:22 mwmnj joined #salt
02:25 cachedout joined #salt
02:28 cachedout joined #salt
02:34 xl1 joined #salt
02:36 ram_ joined #salt
02:37 ram_ Hi All
02:37 ram_ how can we use variables in saltstack like environment variable in Linux
02:40 jfzhu_us ram_: Do you mean how do you get variable from the bash environment to the salt template or just how to use variables in a salt template?
02:42 ram_ thx for your reply...  how to use variables in a salt template?
02:45 jfzhu_us1 joined #salt
02:46 tollmanz joined #salt
02:49 jfzhu_us1 left #salt
02:50 jfzhu_us1 joined #salt
02:51 bhosmer joined #salt
03:00 tssge joined #salt
03:02 tssge Hello, I am a SaltStack noob. Been reading some formulas and stumbled upon this file: https://github.com/saltstack-formulas/hosts-formula/blob/master/hosts/init.sls What is the function of those hash-signs on lines 14 and 35 and some others? Are they comments? Why would someone put comment signs in such awkward place?
03:03 favadi tsantero, it's jinja2 syntax
03:04 Hipikat joined #salt
03:09 oz_akan_ joined #salt
03:18 higgs001 joined #salt
03:27 ajw0100 joined #salt
03:35 ckao joined #salt
03:36 tseNkiN joined #salt
03:50 pdayton1 joined #salt
03:51 luff joined #salt
03:51 luff left #salt
03:54 pdayton joined #salt
04:07 cachedout joined #salt
04:11 tollmanz joined #salt
04:17 oz_akan_ joined #salt
04:19 oz_akan__ joined #salt
04:24 pdayton left #salt
04:25 patrek_ joined #salt
04:28 chutz joined #salt
04:28 andrewclegg_ joined #salt
04:29 [vaelen] joined #salt
04:29 dpac|away joined #salt
04:29 dpac|away joined #salt
04:29 gadams joined #salt
04:29 pnl joined #salt
04:29 gadams joined #salt
04:30 minaguib joined #salt
04:30 scalability-junk joined #salt
04:30 ahammond joined #salt
04:31 Kraln joined #salt
04:31 goki joined #salt
04:31 echos joined #salt
04:31 MTecknology joined #salt
04:31 nliadm joined #salt
04:31 bezaban joined #salt
04:36 mjulian joined #salt
04:36 mjulian joined #salt
04:36 ravibhure joined #salt
04:39 bhosmer joined #salt
04:43 cachedout joined #salt
04:47 btaitelb joined #salt
04:50 jfzhu_us joined #salt
04:57 ajw0100 joined #salt
04:59 mjulian joined #salt
04:59 mjulian joined #salt
05:01 mjulian joined #salt
05:01 mjulian joined #salt
05:03 tollmanz joined #salt
05:07 sfvivek1 joined #salt
05:12 ajw0100 joined #salt
05:13 isomorphic joined #salt
05:19 oz_akan_ joined #salt
05:21 favadi joined #salt
05:28 Teknix joined #salt
05:29 mgw joined #salt
05:41 Katafalkas joined #salt
06:00 cowyn joined #salt
06:01 pdayton joined #salt
06:02 eclectic joined #salt
06:02 jefferai joined #salt
06:02 munderwo joined #salt
06:03 tollmanz joined #salt
06:03 sirtaj joined #salt
06:04 sashka_ua joined #salt
06:04 \ask joined #salt
06:05 luminous joined #salt
06:05 jmccree joined #salt
06:05 kossy joined #salt
06:06 keekz joined #salt
06:06 terminalmage joined #salt
06:07 lahwran joined #salt
06:07 _ikke_ joined #salt
06:07 andyshinn` joined #salt
06:17 matanya joined #salt
06:20 oz_akan_ joined #salt
06:27 bhosmer joined #salt
06:32 fllr joined #salt
06:34 ajw0100 joined #salt
06:44 jacksontj joined #salt
06:44 jacksontj ^^
07:03 tollmanz joined #salt
07:10 elithrar joined #salt
07:13 Shenril joined #salt
07:15 ninkotech_ joined #salt
07:16 favadi joined #salt
07:17 clearclaw joined #salt
07:18 wunki joined #salt
07:21 oz_akan_ joined #salt
07:23 Furao joined #salt
07:29 bbinet joined #salt
07:36 Furao joined #salt
07:37 nkuttler_ joined #salt
07:42 Zuru_ joined #salt
07:44 fllr joined #salt
07:46 fllr joined #salt
07:54 nkuttler joined #salt
08:03 tollmanz joined #salt
08:05 wunki joined #salt
08:05 minaguib joined #salt
08:05 andrewclegg joined #salt
08:05 MSeven joined #salt
08:05 godog joined #salt
08:05 EnTeQuAk joined #salt
08:05 Chrisje joined #salt
08:05 faulkner joined #salt
08:05 g3cko joined #salt
08:11 harobed_ joined #salt
08:15 bhosmer joined #salt
08:22 oz_akan_ joined #salt
08:23 oz_akan__ joined #salt
08:24 slav0nic joined #salt
08:39 giantlock_ joined #salt
08:44 fllr joined #salt
08:52 junedm joined #salt
08:52 junedm left #salt
08:59 junedm joined #salt
09:03 tollmanz joined #salt
09:05 che-arne joined #salt
09:09 druonysuse joined #salt
09:12 aleszoulek joined #salt
09:14 ravibhure joined #salt
09:24 oz_akan_ joined #salt
09:24 aleszoulek joined #salt
09:29 luminous joined #salt
09:30 junedm left #salt
09:40 druonysuse joined #salt
09:43 zpotoloom joined #salt
09:44 druonysuse joined #salt
09:44 druonysuse joined #salt
09:44 fllr joined #salt
09:50 yota joined #salt
10:03 tollmanz joined #salt
10:04 bhosmer joined #salt
10:25 oz_akan_ joined #salt
10:28 bhosmer joined #salt
10:37 hotbox joined #salt
10:44 fllr joined #salt
10:46 fllr joined #salt
10:56 diegows joined #salt
11:00 sgviking joined #salt
11:03 tollmanz joined #salt
11:16 william_20111 joined #salt
11:25 matanya joined #salt
11:25 oz_akan_ joined #salt
11:32 psyl0n joined #salt
11:44 fllr joined #salt
11:50 bhosmer joined #salt
11:52 bhosmer_ joined #salt
12:03 tollmanz joined #salt
12:08 aleszoulek joined #salt
12:10 diegows hi
12:11 diegows I have group of vms that I'm launching using salt-cloud, one of them is the salt master. Is it possible to apply the highstate after all the instances are running?
12:11 xl1 left #salt
12:26 oz_akan_ joined #salt
12:29 dangra joined #salt
12:30 junedm1 joined #salt
12:31 william_20111 joined #salt
12:31 Sheco joined #salt
12:44 fllr joined #salt
13:01 bhosmer joined #salt
13:03 tollmanz joined #salt
13:10 MTecknology joined #salt
13:11 pengunix joined #salt
13:11 blee joined #salt
13:11 tssge joined #salt
13:12 bhosmer joined #salt
13:12 tssge Hello, I am a SaltStack noob. Been reading some formulas and stumbled upon this file: https://github.com/saltstack-formulas/h​osts-formula/blob/master/hosts/init.sls What is the function of those hash-signs on lines 14 and 35 and some others? Are they comments? Why would someone put comment signs in such awkward place?
13:12 smccarthy joined #salt
13:13 MZAWeb joined #salt
13:16 william_20111 joined #salt
13:17 giantlock_ joined #salt
13:27 oz_akan_ joined #salt
13:30 kasanop joined #salt
13:40 timoguin joined #salt
13:41 bhosmer joined #salt
13:44 fllr joined #salt
13:45 elfixit joined #salt
13:51 pdayton joined #salt
13:58 JasonSwindle joined #salt
14:02 th3reverend joined #salt
14:03 tollmanz joined #salt
14:06 Gifflen joined #salt
14:09 aleszoulek joined #salt
14:09 blarghmatey joined #salt
14:11 fllr joined #salt
14:13 _fllr_ joined #salt
14:16 th3reverend left #salt
14:19 william_20111 joined #salt
14:19 cachedout joined #salt
14:28 oz_akan_ joined #salt
14:31 mpanetta joined #salt
14:36 rawzone joined #salt
14:37 slav0nic joined #salt
14:37 slav0nic joined #salt
14:39 tollmanz joined #salt
14:40 yota joined #salt
14:44 fllr joined #salt
14:49 pcurry_nomi joined #salt
14:51 william_20111 joined #salt
14:53 bhosmer joined #salt
15:11 ravibhure joined #salt
15:14 mgw joined #salt
15:17 pcurry_nomi joined #salt
15:17 bhosmer_ joined #salt
15:22 fllr joined #salt
15:22 fllr joined #salt
15:23 forresta joined #salt
15:23 pass_by_value joined #salt
15:26 sfvivek1 joined #salt
15:28 oz_akan_ joined #salt
15:33 iShming joined #salt
15:33 faust joined #salt
15:39 gazarsgo joined #salt
15:42 worstadmin joined #salt
15:42 cetex joined #salt
15:44 iShming joined #salt
15:46 andresr joined #salt
15:46 MTecknology Where can I get salt-minion_0.17.4.exe?
15:47 MTecknology found it!
15:47 forresta *gasp*
15:47 forresta .exe??
15:48 tssge Hello, I am a SaltStack noob. Been reading some formulas and stumbled upon this file: https://github.com/saltstack-formulas/h​osts-formula/blob/master/hosts/init.sls What is the function of those hash-signs on lines 14 and 35 and some others? Are they comments? Why would someone put comment signs in such awkward place?
15:49 MTecknology forresta: I'm going to try to bring all of our Windows servers into salt. :)
15:49 forresta MTecknology, nice!
15:50 forresta tssge, yes those are comments and are part of the jinja syntax, take a look at: http://jinja.pocoo.org/docs/templates/#comments
15:50 cetex joined #salt
15:50 tssge forresta> Thank you a lot!
15:50 forresta tssge, np
15:52 reprise joined #salt
15:54 cachedout joined #salt
15:55 munderwo joined #salt
15:56 cetex joined #salt
15:58 UtahDave joined #salt
16:00 Guest55578 joined #salt
16:00 Gifflen_ joined #salt
16:02 Guest55578 joined #salt
16:03 cetex joined #salt
16:05 seanz joined #salt
16:08 sinenitore joined #salt
16:11 iShming joined #salt
16:11 mgw joined #salt
16:13 lineman60 joined #salt
16:14 dgo joined #salt
16:15 dgo left #salt
16:22 psyl0n joined #salt
16:22 tdilly joined #salt
16:24 schristensen joined #salt
16:26 [diecast] joined #salt
16:27 tdilly Anyone using cython with salt?
16:29 nahamu tdilly: what do you mean?
16:29 nahamu are you looking to extend salt in C?
16:29 oz_akan_ joined #salt
16:30 nahamu or do you mean to compile salt into a native binary?
16:31 nahamu (have you looked at the esky feature?)
16:31 btaitelb joined #salt
16:32 nliadm (the esky feature breaks often)
16:34 UtahDave tdilly: cython support hasn't been a very popular feature as far as we can tell.
16:34 tdilly nahamu: I saw that salt supports cython modules, just curious if anyone was actually using that. Really I'm just curious if there is any way to limit some of the RAM usage on the minion and was researching some possibilities
16:39 clearclaw joined #salt
16:41 philipsd6 joined #salt
16:43 mgw joined #salt
16:45 Gifflen joined #salt
16:50 JasonSwindle UtahDave:  Howdy!
16:50 JasonSwindle Are you guys going to release a RC of Salt soon?
16:52 martoss joined #salt
16:53 iShming joined #salt
16:53 martoss1 joined #salt
16:55 UtahDave JasonSwindle: Yeah, we are. We were hoping to get one out last week, but we wanted to get a few more tests built and a couple fixes in.
16:55 JasonSwindle OK
16:56 JasonSwindle Great!   Getting ready for SaltConf and that seems like a great place to start. :)
16:56 forresta UtahDave, is 7526 gonna be fixed in the first RC? Or does that one still need some work?
16:56 UtahDave Hopefully we'll get it out pretty soon
16:57 UtahDave forresta: yeah, that's being worked on right now
16:57 forresta sweet!
16:57 utahcon is there a way to see the output of a jinja template without actually deploying it to a minion?
16:58 sroegner joined #salt
16:58 UtahDave utahcon: salt '*' state.show_sls myfile
16:58 UtahDave utahcon: or do you mean like a config file?
16:59 utahcon it is a file that I am using file.managed on and have some jinja in it
16:59 utahcon want to see what will be sent to the minion
16:59 utahcon if possible
17:02 UtahDave utahcon: have you tried using test=True  ?
17:02 utahcon I am not aware of test=True
17:02 UtahDave salt '*' state.highstate test=True
17:02 UtahDave that will show you what the command would have done if not for the test mode
17:02 utahcon awesome
17:03 UtahDave I think that will show you a diff.  If you're running your minion in debug mode, I think it may show you the templated file as well.
17:03 utahcon very nice
17:03 utahcon perfect, that showed exactly what I wanted to know
17:03 utahcon thanks!
17:04 forresta utahcon, jesusaurus wrote: https://github.com/jesusaurus/salt-shaker if you're interested in messing with it, it's mostly the 'test' option with a few other nice things.
17:05 bhosmer_ joined #salt
17:07 JasonSwindle UtahDave: Will the next SaltStack (Cloud) have the option / power to create load balancers ?
17:07 jesusaurus its also very much alpha since i dont have much time to put into it
17:07 forresta jesusaurus, yea yea, we all know you're a slacker
17:07 forresta with your 'excuses'
17:07 jesusaurus :)
17:07 forresta :P
17:09 UtahDave JasonSwindle: no, not yet.  That's something I'd really like to see
17:10 UtahDave I bet it wouldn't be very hard to implement, actually
17:10 JasonSwindle Oh, go on? :)
17:12 UtahDave On most clouds it's pretty straightforward.  1. create new load balancer. 2. add these servers to this load balancer.
17:17 [diecast] joined #salt
17:22 btorch is this correct ? {% if 'proxy' in grains['roles'] %}
17:22 timoguin JasonSwindle, there's someone that works for rackspace that's actively working on adding load balancer support to the libcloud module
17:23 timoguin mpanetta, ping
17:23 timoguin somebody asking about load balancers. :)
17:28 forresta looks good to me btorch
17:29 forresta assuming you have a roles grain
17:30 oz_akan_ joined #salt
17:31 pcurry_nomi joined #salt
17:31 timoguin are there any good examples of passing json arguments via the command line? i'm trying to get cmd.run to take json kwargs
17:31 timoguin pretty sure i'm fucking up the syntax
17:32 redbeard2 joined #salt
17:33 redbeard2 left #salt
17:34 UtahDave timoguin: have you tried something like this?  http://docs.saltstack.com/topics/reactor/index.html#fire-an-event
17:36 timoguin UtahDave, this is what i'm trying:
17:36 timoguin salt '*' cmd.run '{"name": "ls"}'
17:37 timoguin and all the syntactic variations
17:37 timoguin for that one i get: /bin/bash: name: command not found
17:37 timoguin so it seems to be attempting to run 'name' as a command rather than parsing it as an arg
17:38 UtahDave timoguin: why are you trying to execute {"name": "ls"}  at the cli on the minion?
17:39 timoguin well i was trying to pass a number of arguments to cmd.run and wanted to try to do it by passing kwargs
17:41 timoguin in this example i'm just trying to run ls in a certain homedir, for adhoc reports
17:41 forresta tomguin, does the shell on the minion properly interpret values passed as json?
17:41 forresta *timoguin
17:42 forresta I guess I'm confused how passing json straight to the minion would do anything
17:43 timoguin i mean i'm just trying to do: cmd.run "ls" cwd="/home" user="foo"
17:43 timoguin but i think it's supposed to accept a json dict for the module arguments
17:43 UtahDave timoguin: what you just typed in should work
17:44 timoguin yea it does. but i wanted to try passing the args as json
17:44 timoguin maybe i'm misunderstanding how it's supposed to work
17:45 UtahDave timoguin: well, I think you can pass in pillar data as json
17:45 UtahDave timoguin: are you looking at a particular example in the docs?
17:45 forresta so timoguin, what you're trying to run is salt '*' cmd.run "$name" {"name": "ls"} ? (or something similar with less terrible syntax)
17:49 Marion joined #salt
17:49 timoguin I'm just trying to pass the module arguments as a JSON dict.
17:49 timoguin I think I just don't understand how kwargs are supposed to work.
17:50 mpanetta was the partition module removed/replaced in 0.17.x?  I can call it in 0.16.4, but 0.17.4 gives me an error that it is not available.
17:50 higgs001 joined #salt
17:50 forresta mpanetta, you mean parted?
17:51 mpanetta Yeah
17:51 mpanetta sorry
17:51 forresta I still see it here
17:51 utahcon SaltConf: pre-conference courses, are there three distinct courses? do they each run all day, or is there three course, each last a few hours?
17:52 martoss joined #salt
17:52 forresta mpanetta, what functionality are you trying to use that doesn't work?
17:52 mpanetta https://gist.github.com/anonymous/62100c4c289bf7938fa6 :(
17:53 mpanetta I am exploring
17:53 forresta try list
17:53 mpanetta One of the new thing we will have to do is set up partitions on our VM's
17:53 mpanetta Ok
17:53 forresta and use parted
17:53 forresta so salt saltdev parted.part
17:53 forresta err parted.list :P
17:53 mpanetta partition.list worked
17:53 forresta ok cool
17:53 forresta yea looking at the examples now
17:53 forresta so it is still partition
17:54 forresta oddly enough it is aliased https://github.com/saltstack/salt/blob/v0.17.4/salt/modules/parted.py#L32
17:54 forresta so that should work
17:54 mpanetta Ah hmm
17:54 mpanetta I guess the name changed between versions
17:54 forresta it just changed from part_list, to parted
17:54 mpanetta I'm hoping that we update to 0.17 soon..
17:55 forresta but with it aliased like that it should still work I believe
17:55 forresta mpanetta, cool, be wary of https://github.com/saltstack/salt/issues/7526 they're working on a fix now, but no reason to get bit.
17:56 mpanetta Hmm
17:57 mpanetta When is Hydrogen being released, do you know?
17:57 UtahDave mpanetta: We're hoping to get the first RC out for Hydrogen in the next few days.
17:57 forresta They're working on the RC now
17:57 mpanetta I started playing with the salt-cloud internals yesterday, very interesting
17:57 forresta UtahDave beat me to it :P
17:57 Linz joined #salt
17:57 mpanetta on develop
17:58 timoguin yea i'm really excited to play with merged salt-cloud
17:58 UtahDave mpanetta: yeah, I've been using salt-cloud on develop, too.
17:58 mpanetta Not sure I like the way libcloud partitions things though, it is going to make what I want to do a pita I think...
17:58 Ryan_Lane joined #salt
17:58 nliadm UtahDave: I found my esky problem, and posted a fix suggestion to the issue
17:59 forresta UtahDave, are the __func_alias__ in https://github.com/saltstack/salt/blob/v0.17.4/salt/modules/parted.py supposed to allow you to keep using part_list? mpanetta was getting an error with that.
17:59 nliadm I mention it because you're usually the one I complain at in here :)
17:59 mpanetta Since both load balancers and compute nodes use similar functions, if I try to import both in the same module there will be name collisions :(
17:59 Gareth 'lo
18:00 UtahDave nliadm: :)
18:00 elfixit joined #salt
18:00 sroegner joined #salt
18:00 UtahDave forresta: I'm not sure, actually.  It seems like it.
18:00 forresta Hmm ok
18:01 forresta mpanetta, if you have time can you open an issue on the partition.part_list?
18:01 mpanetta I can do that now.
18:01 forresta and link over https://github.com/saltstack/salt/blob/v0.17.4/salt/modules/parted.py#L32 ?
18:01 forresta cool, let me know what the issue number is as well when you finish please.
18:03 mpanetta forresta: Should I open it as an issue on github, or is there another preferred way?
18:03 forresta just an issue on github.
18:04 forresta mpanetta, use this link instead of the one I sent you: https://github.com/saltstack/salt/blob/develop/salt/modules/parted.py#L34
18:04 forresta then it will be the develop branch
18:04 mpanetta ok
18:05 mpanetta Oh
18:05 mpanetta Yeah I was running that on develop branch, should I mention it as develop, or 0.17.4?
18:05 forresta develop
18:06 mpanetta I just realized heh
18:08 mpanetta https://github.com/saltstack/salt/issues/9461
18:09 mpanetta Anything I should add to that?
18:10 forresta mpanetta, I'd say add an example on develop of partition.list working
18:10 mpanetta Ok
18:10 forresta I don't get why it fails as is, since on develop the function is still called part_list.
18:12 mpanetta Ok I just added a second comment.
18:13 forresta mpanetta, cool, hopefully someone can explain why list works, but part_list doesn't, and if it is intended to work that way.
18:17 mpanetta forresta: Yep :)
18:19 shadowsun Someone said my name
18:19 shadowsun but it scrolled off
18:19 shadowsun forresta: sup
18:19 forresta shadowsun, I have no idea. I think that was from a few days ago, probably just acknowledging that you found a solution after you messaged me I believe
18:20 Linz joined #salt
18:20 shadowsun UtahDave, forresta - did either of you have any ideas as to why cmd.script isn't pulling from the environment?
18:20 shadowsun forresta: entirely possible.. I usually can't irc from work period full stop; so I tend to have the wrong hours for getting help here
18:20 shadowsun >.<
18:20 seanz left #salt
18:20 forresta shadowsun, I've been in that situation before, no IRC at work is a bummer
18:20 shadowsun Yeah it is
18:21 shadowsun :| I haven't worked at a place that allowed IRC in almost a decade
18:21 forresta If I remember correctly I think cmd.script doesn't pass through something that the shell needs to pull in the environment settings
18:21 UtahDave shadowsun: I think that may be a bug
18:22 UtahDave is that minion matching something in your top.sls?
18:22 shadowsun UtahDave: Okay, so we can confirm that cmd.script *should* pull from the environment and not base?
18:22 shadowsun yes, and that minion has environment specified in /etc/salt/minion just to be extra sure.
18:22 forresta I thought terminalmage fixed the env stuff https://github.com/saltstack/salt/issues/9002
18:22 shadowsun sec
18:22 forresta what release are you on shadowsun?
18:23 shadowsun damnfiknow
18:23 shadowsun checking
18:23 CheKoLyN joined #salt
18:23 shadowsun One of my masters hasn't picked up 0.17.4 yet, kicking it
18:24 shadowsun I'm almost to the point where I'm going to start maintaining my own set of install rpms
18:25 forresta with more recent stuff from develop?
18:25 forresta or just because epel takes forever?
18:25 shadowsun because epel is still on 0.17.2, takes forever, and just generally I'm getting really sick of outdated shit
18:25 shadowsun I mean, git 1.7.1?
18:25 forresta shadowsun, yea, the whole epel-testing process is a bummer
18:25 shadowsun The whole "backports" idea has flaws in the first place
18:25 forresta heh
18:25 shadowsun But this is getting fucking rediculous
18:26 mpanetta yeah epel isn't very helpful... :(  We had the same issue recently
18:26 forresta You need credits to push stuff through epel-testing, and you need to build RPMs that make it through epel-testing to get credits!
18:26 shadowsun >.<
18:26 forresta Just need to find someone with a bunch of credits, so we can avoid the 2-3 week delay in epel-testing
18:26 shadowsun I already run my own repo and we've got a mock server and everything
18:26 forresta nice
18:26 forresta are you guys using koji?
18:26 forresta or just running straight mock from the command line?
18:27 cowyn joined #salt
18:27 shadowsun Yeah, we've got koji as well
18:27 forresta cool
18:28 shadowsun We build for x86 and x86_64, centos 5 and 6
18:28 forresta that's a good setup if you need to build a ton of RPMs
18:28 shadowsun I've just been trying to avoid adding the overhead of maintaining a ton of rpms for internal use
18:28 forresta how many do you maintain now
18:28 shadowsun zeromq and bindings
18:28 shadowsun Because CentOS 5
18:28 forresta heh
18:28 forresta yea that isn't too bad
18:28 shadowsun yeah
18:29 shadowsun It wouldn't be nearly as much of a pain if these were slackware systems
18:29 shadowsun but omg the rpm build process and spec files are such a mess
18:29 forresta no way man, I love building RPMs out with rpmbuild which are fine, then running them through lint, and it's fine, then watching them horribly die in mock
18:30 shadowsun *snort*
18:30 forresta it's only funny because it's so common :(
18:30 shadowsun lie to me more
18:30 forresta I built an RPM with patches and it built cleanly first try? :P
18:31 forresta shaodwsun, have you looked at the opensuse build system?
18:31 shadowsun Not really
18:31 shadowsun I've been running slack since the early/mid 90's sometime
18:31 shadowsun Work uses Centos, so I learn what I need to, but there's so much about it that makes me twitch
18:32 shadowsun I tried Suse way, way back when, but didn't like it much
18:32 forresta I'd only run it for OBS
18:32 forresta http://openbuildservice.org/
18:32 forresta check it out if you have some time
18:32 forresta other than not running on cent unless you want to use some old release a guy built, it seems pretty slick. I need to make some time to mess with it.
18:33 torrancew for simpler needs, fpm is a godsend
18:34 shadowsun torrancew: hmmm
18:35 forresta fpm also doesn't support mock
18:35 shadowsun forresta: But does it give me an srpm
18:35 torrancew yeah, fpm doesn't use the native tools
18:35 forresta yea
18:35 torrancew It recreates the end result, not the process
18:35 forresta torrancew, I agree though, fpm is amazing
18:35 forresta I just wish you could force it to build using mock
18:36 shadowsun If it gives me an srpm, then I can run the srpm through our mock environment
18:36 torrancew I use it in CI to convert rubygems into OS packages for all our internal stuff
18:36 shadowsun (We're actually set up for that)
18:36 torrancew shadowsun: I don't believe so; but I typically use it in .deb-land
18:36 forresta shadowsun, let me know how that goes if you do try it, I'd be curious to see if it meets all the lint stuff.
18:36 shadowsun yeah, I will
18:36 forresta I always thought it only generated an RPM
18:36 forresta but I could be wrong
18:37 torrancew I do believe shadowsun is correct, but again, I don't use it for RPM very much
18:37 torrancew errr
18:37 torrancew I think *forresta* is correct
18:38 forresta I hope I'm not :D
18:38 shadowsun It doesn't look like it does srpms
18:38 shadowsun and, I just found out when you told me about it :(
18:38 shadowsun I was asking lol
18:39 bhosmer joined #salt
18:39 forresta if you know ruby, you could always modify the code *hint* *hint*
18:39 shadowsun I have zero desire to learn ruby at this time
18:39 bhosmer joined #salt
18:39 forresta boooooooooo
18:39 shadowsun I'm buried in python as it is
18:39 jcockhren when writing a module, how do I checl if another salt module (not python module) is loaded?
18:39 shadowsun jcockhren: duplicate import statements aren't a problem
18:39 jcockhren s/checl/check
18:40 CheKoLyN hello all, is there a setting for how gitfs fetches a repo?
18:40 shadowsun CheKoLyN: you mean git:// vs ssh:// ?
18:40 jcockhren shadowsun: ? you answering my question?
18:40 shadowsun jcockhren: yes
18:40 CheKoLyN Ok, yes is ssh repo
18:40 CheKoLyN it's fetching like every minute
18:41 shadowsun CheKoLyN: It's supposed to fetch every minute
18:41 forresta CheKoLyN, and you want it to fetch less?
18:41 jcockhren shadowsun: thanks
18:41 shadowsun CheKoLyN: gitfs doesn't do push notifications, so your fetch frequency should be short enough it picks up the changes within an appropriate interval ;)
18:41 CheKoLyN Ok, cool good to know
18:41 CheKoLyN i didnt know this was by design
18:42 CheKoLyN Thanks all
18:42 shadowsun jcockhren: Welcome.. basically, if it's already loaded, python optimizes out the extra import. So if you need to be absolutely sure it's loaded, just add the extra import. It's not the best style, but otherwise you have to get funky with try/catch NameError and stuff.
18:42 shadowsun CheKoLyN: Welcome!
18:42 forresta you should work from wherever you are working that has IRC access more often shadowsun
18:42 shadowsun I'm on vacation
18:42 forresta then you can answer questions :P
18:42 shadowsun :D
18:42 shadowsun lol
18:42 forresta awww yea vacation
18:43 shadowsun If I could get use of IRC officially improved, I'd be in here all the time
18:43 shadowsun There as *so many hours* it would have saved me.
18:43 forresta yea, we used that argument as well
18:43 forresta it doesn't work
18:43 shadowsun yeah
18:43 shadowsun I know.
18:43 forresta heh
18:43 shadowsun There's two kinds of companies
18:43 shadowsun 1) "We expect you to be an adult, and we'll fire you if you can't be one."
18:44 shadowsun 2) "You obviously can't be adults, so here have lots of rules."
18:44 forresta lol
18:44 jcockhren shadowsun: so you mean like: import salt.modules.s3 (for example)
18:44 Ryan_Lane joined #salt
18:45 jcockhren shadowsun: I think I need to run the __virtual__  function
18:46 shadowsun jcockhren: yeah.
18:47 shadowsun jcockhren: I'm not familiar with __virtual__ as a python magic function, actually
18:48 shadowsun Neither is pydoc, so I'm guessing that's something from salt itself?
18:49 jcockhren shadowsun: yeah. the __virtual__ function returning the name of the state/module tells salt that it's available for usage
18:49 shadowsun Ah.
18:49 jcockhren and names it toos
18:50 jcockhren see: https://github.com/saltstack/salt/blob/develop/salt/modules/reg.py#L52
18:51 mpanetta Someone was asking about kwargs earlier?
18:51 mpanetta maybe this will help them figure it out? https://github.com/saltstack/salt/blob/develop/salt/utils/__init__.py#L1441
18:51 shadowsun Oh, it's on windows
18:51 shadowsun That would explain why I haven't encountered it at all ever :)
18:52 mpanetta Since the dev list is quiet, may I ask a dev question here?
18:52 martoss joined #salt
18:52 jcockhren that was the fastest example I could find of the __virtual__ usage that returns false
18:52 shadowsun ahh
18:52 munderwo joined #salt
18:52 shadowsun it returns 'reg' if it's in windows ;)
18:53 mpanetta basepi: Damn that was a fast fix ;)
18:54 shadowsun blah
18:54 bhosmer_ joined #salt
18:54 shadowsun Can someone tell me how to figure out of ia specific patch has made it into a release?
18:55 shadowsun wow I cannot has the typign today
18:55 pcurry_nomi joined #salt
18:59 forresta shadowsun, usually what I do is find the patch I want and where it exists, then go to the tag of the release I want to look at, to see if it exists.
18:59 basepi mpanetta: which?  i haven't written any code yet today.  o.O
18:59 shadowsun I'm not even sure what you said there
18:59 * shadowsun just looks at the source
18:59 mpanetta weird
18:59 basepi mpanetta: unless you mean the pull req i merged, in which case, you should be thanking redbeard2
18:59 shadowsun No, it's not in salt .4
19:00 mpanetta basepi: Your name was on the merge heh
19:00 forresta basepi, what a slacker....
19:00 basepi forresta: you know it!
19:00 mpanetta Ah thanks be to him then :)
19:00 shadowsun commit 7442158364224aa12ff439ac2bff6a27f7f3fde5
19:00 bhosmer joined #salt
19:05 forresta basepi, that needs a versionadd if techhat still has the repo pulled.
19:05 sroegner joined #salt
19:06 basepi I think it's in 0.17 so it's not as critical.
19:06 forresta cool
19:06 basepi But I'll tell him.
19:06 andresr joined #salt
19:10 shadowsun forresta: that env stuff isn't this env stuff, apparently
19:11 jcockhren woohoo!
19:12 shadowsun wait
19:12 shadowsun the minion didn't patch right
19:12 martoss joined #salt
19:13 shadowsun Oh, hey.
19:14 shadowsun forresta: give me a second, my brain is broken
19:16 shadowsun Okay
19:17 blarghmatey joined #salt
19:18 shadowsun env Pass in a list or dict of environment variables to be applied to the command upon execution
19:18 bhosmer joined #salt
19:18 shadowsun forresta: 9002 fixes that, which isn't at all the environment that I'm talking about
19:18 shadowsun I'm talking about the states/modules/etc "environment"
19:19 shadowsun http://docs.saltstack.com/ref/file_server/file_roots.html?highlight=environment
19:20 shadowsun http://docs.saltstack.com/ref/file_server/index.html?highlight=environment
19:21 shadowsun http://docs.saltstack.com/topics/tutorials/gitfs.html?highlight=gitfs%20environment
19:22 shadowsun Where we explicitly map git branches to environments
19:22 shadowsun It's working for sls and *their* associated files
19:22 shadowsun But it doesn't pull from the environment when using cmd.script, it always pulls from base
19:22 shadowsun and I'm at a loss as to why
19:23 UtahDave shadowsun: have you opened an issue on this?
19:23 shadowsun UtahDave: I've been trying to figure out if there's an issue or if I'm doing something wrong
19:23 UtahDave can you pastebin what you've tried so far?
19:24 shadowsun Which bits should you like this time? :)
19:24 UtahDave lol,  did I already ask you this?
19:24 shadowsun lol
19:24 shadowsun We've talked about this problem about three times now?
19:24 shadowsun but it was at the end of last week
19:25 * Gareth opens an issue about UtahDave asking people to open issues.
19:26 timoguin no, don't do it! too meta!
19:26 UtahDave lol
19:26 * UtahDave head asplodes
19:26 shadowsun The system will implode!
19:26 timoguin do you want an explosion!?
19:26 shadowsun Oh no. Now it's worse
19:26 timoguin okay so would it be an asplosion, implosion, or explosion
19:26 shadowsun the system has *ploded!
19:26 timoguin lol
19:26 timoguin MULTI-PLODE
19:26 UtahDave shadowsun: are you doing this on the CLI or in an sls file?
19:27 UtahDave and what version of Salt?
19:27 shadowsun 0.17.4
19:27 shadowsun When I'm doing this via an .sls file everything functions as expected. However, when I try and use cmd.script, it always pulls from base instead of from its enviroment
19:29 shadowsun The minion in question is in the dev environment in the top.sls file as well as having it statically configured in the /etc/salt/minion file.
19:30 shadowsun gitfs backend.
19:30 shadowsun uhmmm.
19:31 UtahDave ok, so the cmd.script state is working correctly
19:31 UtahDave have you tried adding   env=myenv  to your command?
19:31 shadowsun ?
19:31 shadowsun env in cmd.script has nothing to do with this as far as i can tell
19:32 shadowsun That allows me to manipulate the shell environement
19:32 oz_akan_ joined #salt
19:32 UtahDave not on 0.17.4
19:32 shadowsun *headdesk*
19:32 shadowsun then how come the documentation says otherwise for cmd.script? :P
19:32 UtahDave there's some overlapping of terms that's fixed in develop and the upcoming hydrogen
19:33 UtahDave in develop you use     env for environment variables and saltenv for  the salt environement
19:33 shadowsun mein gott.
19:33 shadowsun Okay. So I have to explicitly specify the environment even though it's already specified in the minion config file?
19:33 shadowsun because yeah, that worked.
19:33 shadowsun also
19:34 shadowsun I really prefer good old fashioned version numbering schemes
19:34 shadowsun Which one is hydrogen?
19:34 timoguin the next release targetting for january
19:34 shadowsun okay
19:34 timoguin should be 2014.01.0
19:34 shadowsun the one with the funky scheme
19:34 UtahDave timoguin++
19:34 \ask joined #salt
19:34 shadowsun got it.
19:34 timoguin from then on every release will go alphabetical with the table of elements
19:34 timoguin for the codename at elast
19:35 shadowsun At least you get bonus points for using a weird versioning scheme that's actually incremental.
19:35 shadowsun ;)
19:35 nahamu "periodical", you mean?
19:35 shadowsun otherwise I think I'd riot. ;)
19:35 shadowsun UtahDave: Irrelevant things aside
19:36 shadowsun UtahDave: Is it actually expected that I have to manually pass the environment that I've statically configured to cmd.script, i.e. is this functioning correctly?
19:36 shadowsun Or should it be picking this up automatically from the minion and is failing to do so?
19:37 pdayton joined #salt
19:37 gazarsgo left #salt
19:44 jacksontj joined #salt
19:44 UtahDave shadowsun: Hm. It might be because at the cli, cmd.script isn't state aware.
19:44 UtahDave shadowsun: let me ask a couple engineers here for some clarification
19:44 shadowsun UtahDave: Ok.
19:46 justBob_ joined #salt
19:47 jacksontj joined #salt
19:48 jcockhren timoguin, forresta: ok. I have a working module that updates a site's codebase (and restarts supervisor) based on a site spec given in pillar
19:48 jcockhren (that is already deployed on minion)
19:49 jcockhren the next step is to get allow for new redeployments to new minions rather than simple codebase updates.
19:51 timoguin jcockhren, cool. i've got the apache-formula working with vhosts and am about to start working on deploying the actual sites
19:52 timoguin which will mainly be git or mercurial clones
19:53 jcockhren timoguin: with our powers combined...
19:54 jcockhren once the module is good, I'm going to get a state together for it
19:55 bhosmer joined #salt
19:56 jcockhren Anyone know if formules can contain custom modules?
19:58 UtahDave jcockhren: yes, they can!
19:58 timoguin party
19:59 jcockhren UtahDave: in a _modules directory like any ol' file_roots?
20:00 UtahDave yep. exactly
20:00 UtahDave custom execution modules go in  file_roots/_modules
20:00 UtahDave custom state modules go in file_roots/_states
20:03 timoguin UtahDave, do you know where custom pillars would go? there was someone in hear in the last few days trying to write an external pillar
20:03 timoguin *here even
20:04 shadowsun UtahDave: Just PM me when you find out the more information?
20:05 UtahDave shadowsun: Yeah. I was waiting for a planning session to end.
20:06 UtahDave timoguin: you can put it in whereever ext_mods is defined.
20:07 shadowsun UtahDave: thanks... It's my vacation so I'm going to go back to not working
20:07 shadowsun ;)
20:07 timoguin i got a lot more work done during my time off
20:10 shadowsun timoguin: You don't even want to know how much I got done in the last two months ;)
20:20 timoguin if i enable the external job cache to send data to redis, can i also have it return to the master?
20:21 timoguin i.e., i still want to get the data back at the CLI while testing
20:21 bhosmer joined #salt
20:22 forresta jcockhren, very cool
20:23 ajw0100 joined #salt
20:25 jcockhren forresta: I even have a name for it
20:25 jcockhren ;)
20:25 jcockhren b/c that's the most important part
20:26 forresta please don't say it's salt-?
20:27 timoguin salt-super-deployer-thing
20:27 shadowsun UtahDave: You know, it works, but there's an error on the minion side:
20:27 shadowsun [ERROR   ] Invalid input: dev, must be a dict or string - yaml represented dict
20:27 mpanetta Crap, I don't think what I want to do is as easy as I thought it would be heh.
20:27 jcockhren the module is called fiber
20:28 jcockhren whatevs. ;)
20:28 jcockhren and that'll be what the formula will be called I guess
20:29 shadowsun UtahDave: omg. env is completely overloaded.
20:29 forresta jcockhren, ok good
20:29 martoss joined #salt
20:29 [diecast] joined #salt
20:32 oz_akan_ joined #salt
20:34 forresta utahcon, are you still around?
20:34 dodger_ joined #salt
20:34 dodger_ hi everyone :)
20:34 shadowsun UtahDave, forresta - check this out: http://pastebin.com/1z50uqfN
20:36 forresta odd
20:36 shadowsun yes
20:37 shadowsun It looks like env is currently being used for two different things
20:37 shadowsun at the same time
20:40 pcurry_nomi joined #salt
20:40 Katafalkas joined #salt
20:42 bhosmer_ joined #salt
20:44 tollmanz joined #salt
20:46 DanGarthwaite joined #salt
20:46 cetex joined #salt
20:46 * DanGarthwaite officially has flights for SaltConf
20:47 pcurry_nomi joined #salt
20:48 JasonSwindle DanGarthwaite:  Awesome!
20:49 DanGarthwaite Right, the deal was I fly my GF to Tampa if I go to SLC.
20:50 JasonSwindle I have to take my GF to Utah if I go to SaltConf.
20:51 DanGarthwaite Coming from NH, in winter, the slopes aren't _that_ appealing.  :)
20:51 forresta DanGarthwaite, nice!
20:51 forresta That reminds me I need to order some sort of winter jacket
20:51 JasonSwindle San Antonio is really flat and dry, so Salt Lake is awesome!
20:51 forresta heh
20:53 psyl0n joined #salt
20:54 forresta You should come to AZ some time in the winter JasonSwindle, other than all the old people, it's pretty nice, lots of hiking, and winter activities a few hours north of Phoenix.
20:54 JasonSwindle I love in Az from a few years.  I loved going up to Scottsdale!
20:54 forresta lol
20:54 DanGarthwaite I've been to biodome2 in AZ a couple of times.
20:54 forresta are you joking JasonSwindle?
20:55 JasonSwindle *lived in Az
20:55 forresta or are you talking about the bar section of scottsdale
20:55 forresta DanGarthwaite, by choice??
20:55 JasonSwindle LOL, I was too young to drink.  So I only had the views and high hills
20:55 DanGarthwaite Mother lived just south of it in Tuscon.  What else is there to do?  :)
20:55 forresta JasonSwindle, ahh ok that makes sense.
20:55 forresta DanGarthwaite, not much, gorge yourself on Mexican food?
20:55 ggoZ joined #salt
20:56 DanGarthwaite Its worth it the first time.
20:56 forresta yea, my confusion was over the multiple visits
21:00 btaitelb joined #salt
21:10 ajw0100 joined #salt
21:13 elfixit joined #salt
21:15 higgs001 joined #salt
21:19 UtahDave timoguin: you still get the returns at the cli when you set up the external job cache
21:19 timoguin perfect
21:20 UtahDave shadowsun: try  env='dev'
21:20 bhosmer joined #salt
21:21 UtahDave shadowsun: ok, I have your answer for you.
21:24 DanGarthwaite joined #salt
21:25 juicer2 joined #salt
21:26 forresta Corey you around?
21:27 pdayton joined #salt
21:29 timoguin UtahDave, I'm interested in whatever that answer is too.
21:31 UtahDave OK, so I'll describe it here for shadowsun's and timoguin's benefit
21:32 UtahDave So the cmd.script execution module doesn't have anything to do with the top.sls.  The top.sls only comes into play with states.
21:32 UtahDave That's why you have to pass the "env" variable at the cli.
21:33 rojem joined #salt
21:33 oz_akan_ joined #salt
21:33 UtahDave "env" is indeed overloaded, so in the develop branch and the next version of Salt  "env" refers to the system environment and "saltenv" refers to Salt's environments.
21:33 forresta That is good to know
21:34 UtahDave Tom thinks that it makes sense to pull from the correct salt environment if the minion's environment if fixed in the minion config, but he's not sure things are working that way right now or not.
21:34 UtahDave forresta: https://github.com/saltstack/salt/issues/7526  is closed now.
21:35 UtahDave timoguin: did that make sense?
21:35 timoguin ah yea it does
21:35 forresta UtahDave, Awesome!
21:35 timoguin modules vs. states.
21:36 timoguin yea good to know. i'm only using a single environment for now, but i'm about to start creating branches
21:36 UtahDave cool
21:39 tollmanz joined #salt
21:40 pengunix joined #salt
21:42 forresta UtahDave, man you guys are cranking through stuff today, friday go!
21:42 ajw0100 joined #salt
21:45 DanGarthwaite +1 for saltenv
21:46 ajw0100 joined #salt
21:51 markizano joined #salt
21:51 kermit joined #salt
21:52 dave_den joined #salt
21:55 berto- joined #salt
21:55 tollmanz joined #salt
21:56 snave joined #salt
21:59 shadowsun UtahDave: Awesome. no, it's not pulling it from the minion config right now in 0.17.4.
21:59 shadowsun UtahDave: thanks for the informations :D
22:01 UtahDave forresta: yep!
22:01 UtahDave shadowsun: you're welcome!
22:10 justBob joined #salt
22:13 shadowsun basepi: you around?
22:13 basepi indubitably.
22:14 shadowsun basepi: https://github.com/saltstack/salt/issues/8597
22:14 shadowsun We were going to have a discussion here, but it never happened
22:14 forresta oh yea I forgot about that
22:14 shadowsun Is it happening on a different bug somewhere that I don't know about, regarding redoing the the batch system?
22:15 basepi shadowsun: this is the one that comes to mind:  https://github.com/saltstack/salt/issues/6456
22:15 basepi Tom keeps telling me that the answer is to redo the batch system, that it's not as good as he wishes it was, even when it works.
22:16 shadowsun yeah, but my issue is one of the last attached ;)
22:16 shadowsun Yes
22:16 shadowsun My question on 8597 is "Tell me what needs to be done" kind of thing
22:16 shadowsun or does the entire batch system need a rewrite from ground up?
22:17 shadowsun And I agree, the batch system is way too full of corner cases
22:17 basepi From what I understand, Tom has been wanting a ground-up rewrite.  It's been on my list for about 6 months.  ;)
22:17 shadowsun Okay.
22:17 basepi Let me refresh my memory on your replies to that issue you linked
22:18 basepi And I'll get back to you
22:18 shadowsun Mostly I almost took down some VPS nodes when the timeout on batch mode got changed from 99999 to default
22:21 forresta lol
22:22 basepi shadowsun: So, would a higher batch default timeout be a temporary workaround?  Also, does -t work properly for batch mode?
22:24 shadowsun basepi: Specifying -t is my workaround for now, w hich is fine. My original problem was "omg this is a dangerous change if your environment actually needs a batch mode that batches instead of doing an interval staggered start"
22:25 basepi Right, so are you familiar with the way timeouts actually work in salt?
22:26 shadowsun yes
22:26 shadowsun the task has five seconds or so to complete by defautl
22:26 basepi That's actually not strictly true.
22:26 basepi One sec
22:26 shadowsun okay, that's what my master conf file says
22:26 shadowsun let me rephrase a bit though
22:26 shadowsun the task can take longer
22:26 shadowsun but nothing will wait longer than that for the task to complete.
22:27 basepi Basically, the timeout is the amount of time the master waits until it queries the minion to see if the job is still running.  After that timeout is up, the master sends a find_job command to the minion to check on the status of the job.  If the minion replies that it's still running, another timeout is triggered.  The problem comes when the minion doesn't reply in a timely manner.
22:27 shadowsun i.e. if I'm salt '*' -b 2 -t 5 cmd.run 'sleep 120', it'll wait ~5 seconds for the first two to complete, and will then go ahead and run the next two even though the first two haven't finished
22:28 shadowsun Uh, that's not the behavior I've observed.
22:28 basepi It's possible that behavior is not the default for batch
22:28 basepi But that's the behavior for the normal salt commands
22:28 shadowsun Okay.
22:28 basepi If you run a command with -v, you can see this in action
22:28 shadowsun that's not the behavior I observe there either
22:28 shadowsun If I pass -v the observed behavior changes.
22:28 DanGarthwaite left #salt
22:29 basepi ???  It should just be more output, not changed behavior....
22:29 shadowsun Not from what I've been seeing
22:29 basepi o.O
22:29 basepi Outside of batch mode?
22:29 shadowsun What I've been seeing is that if you pass -v, it will go ahead and wait more than it would otherwise
22:29 kermit joined #salt
22:29 shadowsun outside of batch mode
22:29 shadowsun here
22:29 shadowsun lets see if I can still replicate this, since I just moved everything the rest of the way to 0.17.4
22:30 harobed joined #salt
22:30 shadowsun also
22:30 harobed joined #salt
22:30 shadowsun can we define "timely"?
22:30 bhosmer joined #salt
22:31 forresta isn't it 5 more seconds?
22:31 shadowsun Because there's a lot of things I'll do that result in heavy IO on the target systems
22:31 shadowsun I don't know
22:31 shadowsun I'm starting to feel like I need to print out the source code for salt and go read it
22:31 harobed joined #salt
22:31 shadowsun I keep on diving into it lately lol
22:31 forresta :P
22:31 basepi shadowsun: that's the problem -- currently it's hard-coded to 2 seconds.  we *really* need to make it configurable
22:31 basepi (that secondary timeout, the wait from find_job)
22:32 shadowsun basepi: Yeah, that 2 seconds is probably where I'm encountering the problem
22:32 shadowsun Since many of these systems are VPSes and my "task" involves recompiling things.
22:32 basepi Yep.  This is something I can't believe we haven't done, it just keeps getting pushed off. It's not even a difficult change....
22:32 shadowsun Also, intermittent packet loss and other things from the realities that major data centers get massive DDOSes
22:32 shadowsun and Arbour can't mitigate everything
22:33 shadowsun Well
22:33 basepi Yep, that secondary wait is almost certainly the problem.  The master is declaring the minion dead because it doesn't hear from it
22:33 shadowsun yes.
22:33 basepi So it starts the next wave of batch jobs
22:33 shadowsun So
22:33 shadowsun you said this was in find_job?
22:33 basepi Or quits, in the event of a normal salt job
22:33 basepi Nah, it's not *in* find_job
22:33 basepi It's just in the code to wait for find_job
22:33 shadowsun okay.
22:33 basepi Let me see if I can dig up a code location that you can tinker with
22:34 druonysus joined #salt
22:35 shadowsun Hrm
22:36 shadowsun While I'm thinking about it
22:36 shadowsun Do you guys actually have a doc for whatever formatting you're using?
22:36 shadowsun This isn't strict PEP8
22:36 forresta well basepi, if you just worked nights and weekends, it would be done.. JEEZ
22:37 basepi forresta: ;)
22:37 basepi shadowsun: http://docs.saltstack.com/topics/style.html
22:37 shadowsun basepi: thanks.. Right now my stuff throws about 190 errors on master.py alone ;)
22:37 basepi we have line length turned off currently in the linting
22:37 basepi that's the biggest offender, i think
22:38 martoss joined #salt
22:38 shadowsun yeah
22:38 shadowsun I've got some multiple of four
22:38 basepi Second biggest is continuation line indentations
22:38 basepi Yep
22:38 shadowsun and I've got some McCabe complexity checks
22:38 shadowsun Which I use because after I've got about 12 paths through the code, I have trouble thinking about them all at once
22:39 basepi Yes, and master.py is.......complex
22:39 shadowsun yeah
22:39 shadowsun hrm
22:39 basepi Don't even get me started on state.py  ;)
22:39 shadowsun lol
22:39 shadowsun Oh, my perfectionist self
22:39 shadowsun I want to go through and reformat all the code now
22:40 shadowsun lol
22:40 forresta yea, I made a poor choice when I first started trying to understand salt and went 'state.py looks like a good place!'
22:40 shadowsun hahaha
22:40 forresta I still don't understand how all of that stuff works
22:40 basepi Well, one problem is that the pep8 script does some weird things with indentations, so you have to be careful there.  A lot of our indentations are actually PEP8-compatible but the pep8 script still complains.  =\
22:40 forresta heh
22:40 basepi So be careful not to just add churn for churn's sake
22:41 basepi But yes, the state compiler is *not* the place to start.  =P
22:41 basepi Tom is not the best documenter-of-code.  =P
22:41 shadowsun lol
22:41 shadowsun Well, I'm using syntastic (surprise, surprise)
22:42 basepi Same.
22:42 basepi With a couple of the pep8 checks disabled
22:42 basepi But yes, I've had to suppress a fair bit of my perfectionist tendencies working on this codebase
22:42 basepi In order to actually get things done.  ;)
22:43 shadowsun yeah ;)
22:43 shadowsun anyway
22:43 basepi I'm still trying to remember where that timeout is coded -- Tom is in a meeting right now, I'm sure he could point me right to it.....
22:44 shadowsun lol
22:44 shadowsun It's not in master.py that I can see
22:45 JasonSwindle You ever start to refactor something, and forgot why you started in the first place?
22:45 JasonSwindle <— Me
22:45 JasonSwindle Friday brain blues....
22:46 shadowsun JasonSwindle: yeah
22:46 shadowsun JasonSwindle: It's why I use git
22:46 shadowsun JasonSwindle: "Okay, what WAS I working on when I decided to rewrite 300 lines of code again?"
22:46 JasonSwindle lol, yea…..
22:46 EugeneKay git-snacks
22:46 markizano oh yay
22:46 markizano people
22:46 shadowsun lol
22:46 shadowsun EugeneKay: yes.
22:47 markizano I have a knack for logging into this channel when nobody's active :P
22:47 EugeneKay I'm always active; just not audibly so.
22:47 JasonSwindle I am pretty much always here, just normally on a phone call.
22:48 shadowsun basepi: cmdmod.py around line 373?
22:48 martoss1 joined #salt
22:49 basepi shadowsun: i am obviously not at the same rev as you.  =P
22:49 basepi let me fetch and rebase
22:49 basepi i'm pretty sure i'm just behind
22:50 basepi nope, still the same line.  can you link to the line on github?
22:50 shadowsun No, it's not explicit, but it's the section of the function that looks like it returns errcode 1 on timeout
22:50 shadowsun but that's not quite what we're looking for
22:51 oz_akan_ joined #salt
22:52 jacksontj joined #salt
22:53 basepi nah, i think this is going to be deep in the cli routines.  hoping Tom comes out soon....  =P
22:53 basepi but imma keep looking
22:54 forresta I'm imagining you sitting at your desk basepi just staring at the door
22:54 basepi not far off.  =P  except if i were going to resort to that, i'd probably browse twitter while i waited or something.  ;)
22:54 forresta hah
22:55 basepi staring at a door is a waste of time!  (and Twitter is *definitely* not)
22:55 basepi ;)
22:55 forresta yea of course, while we're lumping that together let's just put reddit and slashdot in as well
22:55 basepi Man alive, I have to stay away from reddit.
22:55 basepi That's dangerous.  =P
22:55 forresta heh
22:55 basepi Almost as dangerous as my RSS reader!
22:56 forresta lol
22:56 shadowsun lol
22:57 dwfreed basepi: yeah, I know that; I spent most of yesterday staring at imgur
22:57 basepi Hehehe, pretty sure 90% of the internet traffic is now either imgur or netflix....
22:57 forresta I actually just canceled netflix
22:57 forresta overlap between netflix and amazon prime has gotta be up in the 95% range
22:58 forresta granted Amazon's interface is a bummer compared to netflix
22:58 dwfreed during the hours of 9PM and Midnight, 33% of the traffic on the tubes is Netflix
22:58 basepi That's actually a crazy statistic.
22:59 JasonSwindle forresta:  I agree with you.  The interface needs love, badly
23:00 forresta If there wasn't such a content overlap I'd keep netflix
23:00 basepi I need to just buy Community so I can cancel my Hulu Plus.  Pretty sure that's the only show I use it for.....  I guess I watch Bob's Burgers on there too.....
23:00 druonysus joined #salt
23:00 druonysus joined #salt
23:00 forresta ugh hulu plus
23:00 * basepi is getting distracted
23:00 basepi forresta: amen!
23:00 forresta 'you wanna pay to watch ads? YEA PAY TO WATCH ADS!'
23:00 basepi when i have 90 seconds of ads 4 times in a 20 minute show for a paid service........
23:01 EugeneKay stealing media: ad-free since the 1800s.
23:01 basepi hahaha
23:01 forresta basepi, yea
23:03 basepi AHA!
23:03 basepi shadowsun: i think it's salt/client/__init__.py, line 176 in current develop
23:03 shadowsun in gather_job_info?
23:03 basepi yep
23:03 basepi that's when it's querying the minions for their job info
23:03 shadowsun yes.
23:03 basepi and that 2 is in the timeout slot
23:04 shadowsun and
23:04 basepi So I would try bumping that, see if you see different behavior
23:04 shadowsun We probably shouldn't just use the default timeout value for that, though
23:04 basepi Right, we need to make it a separate, configurable value
23:04 basepi But I wanted to just give you the line so you could tweak and test if you so desired in the meantime
23:04 shadowsun If all timeouts become separate configurable timeouts
23:04 shadowsun salt'll have 40 of them
23:05 basepi True, but this one is universal enough it definitely needs to be one of them.
23:05 shadowsun Then why not just use the regular timeout value?
23:05 basepi Because often you'll want the original timeout to be quite a bit larger, I think
23:06 basepi So it's not checking in with the minions way before it ever needs to
23:06 basepi Though, making it the same as the other timeout could be a temporary workaround.
23:06 shadowsun Actually, the only problem I have with the default timeout turns out to be this
23:06 shadowsun I mean, if it can poll the minion and keep waiting, then I'm a happy camper
23:07 shadowsun I guess there's some confusion about what salt means by "timeout"
23:07 shadowsun (and then there's the whole batch thting we got distracted from)
23:07 forresta didn't we talk about how modifying the primary timeout was bad? As it would negatively impact normal state runs.
23:07 shadowsun forresta: *batching* state runs.
23:07 basepi Right.  That doesn't work (at all) on Windows right now, because of a multiprocessing bug.  And it seems like its become less consistent recently in unix, trying to track down why that might be the case (i think we've had some general performance degradation in the last few releases)
23:08 basepi (that was talking about polling the minions)
23:08 basepi We need to figure out why that poll isn't consistently returning the correct result.  And I think it may just be a case of too short of a timeout.
23:08 forresta shadowsun, yea I know what you mean.
23:10 shadowsun basepi: Honestly, it depends on the load of the target minions, but I'm frequently working with boxes that have IOWaits of over 40%
23:10 forresta shadowsun, 40% iowait??
23:10 shadowsun Yes
23:10 forresta ugh
23:10 basepi shadowsun: do you mind if i ask where you work?
23:10 shadowsun Not at all
23:10 forresta Is it all local disk shadowsun? 40% sounds brutal
23:13 shadowsun forresta: It really depends on how you define "Local disk"
23:13 pcurry_nomi joined #salt
23:13 forresta well, if it's in the same rack in some sort of SAN array or something I don't consider it local
23:13 shadowsun Most of it's local then
23:13 forresta I'm just having a hard time with 40% iowait, I mean your boxes have gotta be stressing the array.
23:14 forresta ok
23:14 shadowsun Well yes, they are stressing the arrays
23:14 shadowsun and that's with lots and lots of fancy things making the arrays faster :)
23:14 forresta damn
23:14 forresta so where do you work?
23:15 shadowsun Generally prefer to not actually say
23:15 forresta fair enough
23:15 shadowsun It avoids people asking me for favors concerning where I work ;)
23:16 forresta fair enough
23:16 shadowsun in any case
23:17 shadowsun 30-40% iowait is pretty normal and the way we run the systems, they're still quite responsive
23:17 basepi shadowsun: alright, so I think raising that secondary timeout could really be the answer for your problems
23:17 shadowsun However, you get up above that and you start having latency on execution
23:17 basepi I'm going to work on making that configurable.  In the meantime, if you have high-latency boxes that you can test that theory on, that would be useful.
23:18 shadowsun So figure for a box with a couple hundred active threads and compiling something, a return timeout of about 10 should be good almost all the time
23:18 shadowsun yeah
23:18 shadowsun I'm going to go ahead and make the change now and we'll see over the next week how it works
23:18 basepi (This fix may not be in until Monday -- I'm not staying late on a Friday, sorry.  ;)  And I'm not a crazy workaholic like whiteinge.
23:18 shadowsun lol
23:18 shadowsun I'm not even going to do much testing until monday
23:18 basepi Cool
23:18 shadowsun and I've got thurs/fri/sat off for new years
23:19 shadowsun At this point I'm just poking around because I really do love this stuff
23:19 basepi Nice.  =)
23:21 shadowsun okay, I've made the edits on both the masters
23:22 jkess joined #salt
23:23 shadowsun I'll let you know how it works :)
23:26 jkess hi, I'm trying to add redis to FREEZER_INCLUDES to include with the minion package but it's not getting copied. Any ideas of what to look at to resolve this? Thanks.
23:28 UtahDave jkess: are you trying to create an esky build?
23:30 jkess yes, ultimately trying to create a windows package + a few extra python modules. I've added pytz successfully.
23:35 UtahDave jkess: is "redis" the name of the python package?
23:36 mwmnj joined #salt
23:36 jacksontj joined #salt
23:36 jkess yes, I installed via "easy_install redis"
23:37 UtahDave and the resulting esky build doesn't have the redis libraries available?
23:40 jkess no, redis is missing for the resulting zip
23:41 martoss joined #salt
23:52 jkess UtahDave: I think I figured it out. redis was missing from requirements.txt
23:55 UtahDave that's odd. I'm pretty sure the setup.py doesn't look at requirements.txt
23:57 jkess it's reading it in as SALT_REQS (line 372 in setup.py) then passing it to "install_requires"

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