Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-01-07

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

All times shown according to UTC.

Time Nick Message
00:00 jacksontj or 2014.01.0
00:00 eclectic_ joined #salt
00:00 druonysus joined #salt
00:01 bersace_ joined #salt
00:02 viq_ joined #salt
00:02 pexio_ joined #salt
00:02 analogby2e joined #salt
00:02 nahamu_ joined #salt
00:03 djinni` joined #salt
00:03 d10n joined #salt
00:03 rojem joined #salt
00:03 flebel joined #salt
00:03 Psi-Jack heh.
00:04 davet joined #salt
00:04 godber joined #salt
00:04 TomasNunez joined #salt
00:04 jeffrubic joined #salt
00:05 Psi-Jack So, I noticed supervisord support was added to salt 0.17.0, but the example lacks a lot of .... supervisord specifics. Like, for example, supervisord services usually end up being named processname:processname, or groupname:processname
00:05 Psi-Jack But using : in yaml like that, would it not have issue?
00:06 UtahDave joined #salt
00:06 cro joined #salt
00:06 wunki joined #salt
00:06 kedo39 joined #salt
00:08 steveoliver joined #salt
00:14 JasonSwindle joined #salt
00:16 elithrar_ joined #salt
00:19 higgs001 joined #salt
00:28 aranhoide joined #salt
00:30 aranhoide if I call `salt minion-id state.highstate` from my master I just get a rather quick 'minion-id: ' blank response.  but if I ssh into the minion and I run `salt-call state.highstate` from it, that works
00:30 aranhoide is there any known condition that could cause that?  I'm using salt version 0.17.4
00:32 aranhoide minions respond alright to other commands like test.ping ('minion-id: True').  it's just highstates from the master that they seem to ignore
00:33 aranhoide has anyone else had a similar problem?
00:33 jacksontj terminalmage: yt?
00:34 terminalmage yeah, sup?
00:35 jacksontj nvm, tom got my merge ;)
00:36 jacksontj found a fairly glaring bug in a thing we were working on ;)
00:37 terminalmage ohh
00:37 terminalmage oh
00:37 terminalmage what was it?
00:37 jacksontj https://github.com/saltstack/salt/pull/9598
00:37 jacksontj tom did elif instead of if
00:37 jacksontj ;)
00:37 terminalmage https://github.com/saltstack/salt/pull/9598?
00:37 terminalmage ahh yeah
00:37 terminalmage cool
00:37 jacksontj pretty close
00:37 terminalmage I'm working on a file.managed issue
00:37 jacksontj but did cause some test failures
00:37 jacksontj the deprication warning stuff?
00:37 terminalmage no
00:37 jacksontj or something else
00:38 terminalmage something else
00:38 terminalmage writing commit msg now, so I'll have the pull req up shortl
00:38 terminalmage shortly
00:38 jacksontj cool :)
00:38 jacksontj i love when i get to spend a day just digging into salt stuff :)
00:40 terminalmage yeah, it's really fun
00:40 vejdmn joined #salt
00:40 terminalmage I love it so much I ended up making it my job :)
00:42 taion809 joined #salt
00:42 terminalmage Still writing the commit msg
00:42 * terminalmage is wordy
00:44 jacksontj better then half my commits "this fixes something, and makes it better/faster"
00:44 terminalmage lol
00:46 elithrar joined #salt
00:48 Gareth "It was the best of commits, it was the worst of commits, it was a commit of widsom, it was a commit of foolishness, it was the epoch of code change...."
00:48 terminalmage jacksontj: https://github.com/saltstack/salt/pull/9599
00:48 terminalmage good toilet reading material
00:48 terminalmage (because it's so wordy)
00:49 terminalmage I like to err on the side of providing more information in commits
00:49 terminalmage that way people don't look at this and say "why the heck did this idiot do X?
00:50 rojem joined #salt
00:50 steveoliver terminalmage++
00:50 jacksontj nice ;)
00:50 rojem joined #salt
00:50 jacksontj coment is definetly longer than the diff ;)
00:51 terminalmage I know!
00:52 terminalmage thanks steveoliver!
00:52 terminalmage I've had a 2-3 paragraph commit msg to explain a one line change before
00:54 cachedout joined #salt
00:56 taion809 joined #salt
00:57 * Gareth gives terminalmage a cookie
00:57 terminalmage if you give a programmer a cookie....
00:58 terminalmage someone needs to write that book
00:58 terminalmage a spoof on the old children's book
00:58 kedo39 terminalmage: if you have time, can you look at my pull request for the openssh-formula? https://github.com/saltstack-formulas/openssh-formula/pull/4
00:58 Gareth terminalmage: indeed.
00:59 terminalmage kedo39: sure
00:59 kedo39 thanks!
00:59 Gareth terminalmage: or someone needs to do a short talk with that theme :)
01:01 steveoliver terminalmage: there's also http://stopwritingramblingcommitmessages.com/ :)
01:02 terminalmage steveoliver: haha, yeah I typically keep my initial messages to 50 or less
01:02 terminalmage vim is helpful for that
01:02 terminalmage changes the color of the text once you get past 50 chars
01:02 * steveoliver has found vim helpful for that also
01:02 terminalmage and also sets a fixed line width for the additional paragraphs to keep them at 72 or less
01:02 terminalmage yeah, vim FTW
01:02 steveoliver vim++
01:05 Gareth terminalmage: You going to be at SCALE with the rest of the Salt crew?
01:05 terminalmage Gareth: I'm hoping! we haven't booked anything
01:05 terminalmage well, not for me at least
01:06 terminalmage Gareth: I was there last year, when I was still working for my old employer
01:06 Gareth terminalmage: ah cool.
01:07 terminalmage not really a trip we'd normally have made from the Chicago area for a conference, but we had some money left in the budget and SCALE was the only one in the timeframe from which we had to select that looked good
01:07 terminalmage and also we had some outstanding bugs that I really wanted some face-to-face time with the devs to fix so we could continue our Salt rollout
01:09 terminalmage Gareth: I kind wish I was out there right now
01:09 terminalmage today's high was -11
01:11 Gareth terminalmage: so we do an Ignite like series of talks called UpSCALE, not sure if you saw them last year.  If you want to write and give the "If you give a programmer a cookie" talk :)
01:11 terminalmage oh man
01:12 terminalmage I'll have to see if I can be clever enough to do that
01:12 terminalmage I'll definitely take a stab at it!
01:12 nebuchadnezzar joined #salt
01:13 terminalmage Gareth: something akin to this? https://www.destroyallsoftware.com/talks/wat
01:13 terminalmage I love that one
01:13 Gareth Upscale talks are 5 minutes.  20 slides.  auto advancing slides.
01:13 nebuchadnezzar joined #salt
01:14 terminalmage ha, nice
01:14 terminalmage so, 15 sec/slid
01:14 terminalmage slide
01:14 terminalmage or are they randomly advancing? :)
01:14 terminalmage that'd be fun
01:14 Gareth Yup.
01:14 Gareth No no.  in order :)
01:15 terminalmage I meant in random intervals
01:15 terminalmage that would be interestng
01:15 terminalmage *interesting
01:15 rjc joined #salt
01:15 terminalmage man, I think the cold is getting to me
01:15 terminalmage fingers no worky
01:15 Gareth this talk is awesome.
01:15 terminalmage I know right?
01:16 steveoliver classic
01:16 steveoliver DAS++
01:17 strata_ joined #salt
01:19 Gareth terminalmage: so yeah.  exactly in that style :)
01:20 strata_ hi, i am working on a file.managed state for /etc/mysql/my.cnf, the contents of which are in pillar.  in this my.cnf file, i want to set a "bind_address" that is local to the minion.  when the config is processed in pillar, the resulting IP is the salt master's IP, which i've expected to run into after reading about this on salt-users.
01:20 xzarth_ joined #salt
01:21 strata_ one way i'm trying to work around this is using this in pillar:   bind-address          = {{ grains['ip_interfaces']['eth0'] }}     this works, however, it results in brackets and single quotes around the value of the IP for eth0, which breaks mysql.  is there a way to strip this out?
01:21 terminalmage Gareth: nice :)
01:22 strata_ example of how it ends up looking in /etc/mysql/my.cnf:   bind-address= ['1.1.1.1']
01:22 terminalmage strata_: brackets? is that a list?
01:22 terminalmage yeah, that's a list
01:22 strata_ yeah
01:22 terminalmage because interfaces can have more than one IP
01:22 terminalmage so just add [0] to the end
01:23 terminalmage and it'll refer to that index
01:23 terminalmage grains['ip_interfaces']['eth0'][0]
01:23 strata_ ahh ok, ill give that a try, thank you!
01:23 terminalmage no prob
01:23 steveoliver strata_: terminalmage: that doesn't make sense in pillar, does it?  you can just use the grains.get() right in the sls file?
01:24 steveoliver s/right in the sls file?/because you can use it right in state files../
01:24 xt joined #salt
01:25 steveoliver state (grains.get()) > state (pillar.get(grains.get()))
01:25 strata_ i'm storing that in pillar.  hold on a sec
01:25 terminalmage steveoliver: are you referring to the grainst.get function?
01:25 steveoliver strata - do you need to store it in pillar when it's already a grain?
01:25 terminalmage or grains.get as in dict.get
01:25 steveoliver sorry, syntax is off.. but you get my point?
01:26 terminalmage no, that's why I asked for clarification
01:26 steveoliver oh, ok.. :)
01:26 terminalmage there is a grains.get salt function
01:26 terminalmage but grains is a dictionary in templating
01:27 terminalmage meaning you can do grains.get('foo', 'bar')
01:27 strata_ im storing the contents of the /etc/mysql/my.cnf file in pillar, basically this:  http://pastebin.com/zaJ9VmPT
01:27 terminalmage the salt function lets you traverse through a multilevel dict though
01:27 terminalmage using : as a delimiter
01:27 steveoliver terminalmage: right…
01:27 steveoliver which is bad ASS
01:27 terminalmage so, salt['grains.get']('foo:bar:baz', 1)
01:27 steveoliver right…
01:28 terminalmage steveoliver: yeah, I wrote the traversal code initially :D
01:28 steveoliver very cool
01:28 steveoliver so anyways...
01:28 steveoliver strata_: that's not /quite/ how you wanna use pillar ;)
01:28 strata_ great, that index suggestion helped
01:29 strata_ steveoliver:  sorry, not following completely, why is that?
01:29 steveoliver strata_: you should rather use pillar for sensitive data coming from master to minion, whereas you can just reference the minions' grains' values directly in state files without the need to look at pillar data..
01:30 JasonSwindle joined #salt
01:30 steveoliver (strategic)
01:30 steveoliver then, tactically, you want to just use pillar to store specific values, not entire files..
01:30 terminalmage yeah
01:30 steveoliver i know this is really hand-wavey, but i just wanted to try to talk through what I see for your case
01:31 steveoliver so pillar: is a great place to put ports definitions, etc...
01:31 strata_ hm, i've always found just using pillar to store things like that to be easier/making more sense than setting up custom grains.   at least that's how i've always thought about things.  one place (pillar), has everything i need via dictionaries, etc.
01:31 terminalmage steveoliver: strata_: I don't think a "get" function is what one would want here, since that value is still a list that you need to index
01:32 steveoliver well in your case, strata_, you're not setting up a new grain, you're accessing one that already exists...
01:32 terminalmage strata_: the IP address is just in grains though, right?
01:32 steveoliver in the case of the ipv4
01:32 terminalmage is that a core grain?
01:32 steveoliver terminalmage: you should know this… ;)
01:32 bhosmer joined #salt
01:32 steveoliver hell yeah it is
01:32 strata_ yeah, it's just used inside of the the pillar value, which works now with the index
01:33 terminalmage steveoliver: probably, but salt is quite actively developed
01:33 terminalmage I don't keep track of everything ;)
01:33 steveoliver :)
01:33 terminalmage strata_: so you're referencing the grain within your pillar SLS file?
01:33 steveoliver strata_: i'm just kindly challenging your assumption that you need to store grain values as pillar values :)
01:33 terminalmage that's probably not what you want to do
01:34 steveoliver ^^
01:34 steveoliver …when you can just get the grain value directly...
01:34 terminalmage strata_: so you have a template my.cnf that contains a bunch of references to pillar data, correct?
01:34 strata_ yep, http://pastebin.com/zaJ9VmPT   this lets me have one, single pillar value, which will make every minion that processes it
01:34 strata_ yep
01:34 steveoliver terminalmage: no, look at the paste :)
01:34 steveoliver strata_: not what you want, amigo/a ...
01:34 strata_ the state calls the pillar value in contents.. ill give an example
01:34 alekibango joined #salt
01:35 steveoliver i understand what you're doing, strata...
01:35 steveoliver you're just missing out on the jinja templating feature of salt
01:35 terminalmage strata_: ahh, I see. OK, yeah, you can just use index 0 there and that should work
01:35 strata_ yep, it did, thanks again
01:36 steveoliver terminalmage: …and leave a whole my.cnf as the value of a pillar item?? :0
01:36 cnf o,O
01:36 terminalmage steveoliver: no
01:36 strata_ my state in the end looks like this:  http://pastebin.com/aY94sDRW
01:37 terminalmage strata_: you have to think of it in terms of the data structure... sometimes I find the pprint outputter to be helpful for this. for instance, salt minion-id grains.get 'ip_interfaces:eth0' --out pprint
01:37 steveoliver strata_:see https://github.com/circatree/salt-hosting-states/blob/master/mariadb/my.cnf
01:38 strata_ terminalimage:  thanks!  ill add that to my toolkit
01:38 steveoliver terminalmage: nice tip
01:38 terminalmage oh wait... that's the pillar SLS?
01:38 steveoliver lol
01:38 steveoliver yeah
01:38 terminalmage talking about strata_'s paste earlier
01:38 terminalmage I asked if it was a file template and the answer was "yes"
01:39 terminalmage so, given that, it looked fine
01:39 rusted_sword joined #salt
01:39 terminalmage yeah, steveoliver's example is much better
01:39 ryan_lane joined #salt
01:39 strata_ thx all, have a g'nite
01:39 steveoliver strata_: cya
01:40 terminalmage you too. I'm going to go huddle under a blanket and read a bit
01:40 steveoliver cya, mage
01:40 terminalmage yup, seeya!
01:40 vejdmn joined #salt
01:52 Gareth terminalmage: hope you feel better.
01:56 troyready joined #salt
02:00 Cottser joined #salt
02:07 mgw joined #salt
02:13 tollmanz joined #salt
02:14 rojem joined #salt
02:19 scoates it's really difficult to debug jinja rendering problems. )-:
02:22 davet joined #salt
02:24 steveoliver scoates: you're really right
02:25 steveoliver that damn line 302…. or whatever it is
02:25 scoates yeah
02:25 steveoliver lyin' ass :)
02:25 * steveoliver feels your pain
02:28 favadi joined #salt
02:32 Gifflen joined #salt
02:33 rojem joined #salt
02:37 dleedyna joined #salt
02:38 Gifflen joined #salt
02:41 tollmanz joined #salt
02:48 jacksontj joined #salt
02:50 xl1 joined #salt
02:51 rojem joined #salt
02:51 rojem joined #salt
02:58 darless agreed, debugging jinja is one of the things that I wish could be improved.
02:58 aranhoide left #salt
03:00 Gifflen joined #salt
03:06 fllr joined #salt
03:08 scoates when doing an `include` from a state sls, there's no way to pass parameters in, is there? do I have to use a Jinja include?
03:11 darless joined #salt
03:13 davidfischer joined #salt
03:22 fllr joined #salt
03:31 rojem joined #salt
03:34 forresta joined #salt
03:37 Furao joined #salt
03:41 tollmanz joined #salt
03:42 fllr joined #salt
03:46 rojem joined #salt
03:50 thrashr888 joined #salt
03:52 dleedyna joined #salt
03:55 wkf_ joined #salt
04:04 rojem joined #salt
04:05 Linz joined #salt
04:05 rojem joined #salt
04:08 jalbretsen joined #salt
04:17 redmin joined #salt
04:28 cachedout joined #salt
04:34 rojem joined #salt
04:35 Furao joined #salt
04:41 tollmanz joined #salt
04:43 rojem joined #salt
04:45 sroegner_ joined #salt
04:53 ravibhure joined #salt
05:03 ravibhure joined #salt
05:05 Linz joined #salt
05:13 ConceitedCode joined #salt
05:13 anuvrat joined #salt
05:16 ravibhure joined #salt
05:20 ravibhure1 joined #salt
05:27 techdragon Salt could do with some more documentation on how to not use jinja, they mention, but beyond a doubt the bulk of the existing documentation shows how to use jinja a
05:28 eliasp_ joined #salt
05:28 eclectic joined #salt
05:28 jefferai_gone joined #salt
05:29 forresta techdragon, do you mean in the sense of not using jinja at all? Or using mako?
05:29 sirtaj joined #salt
05:29 vipul joined #salt
05:30 d10n joined #salt
05:30 d10n joined #salt
05:31 d10n joined #salt
05:31 d10n joined #salt
05:31 kevinbrolly joined #salt
05:33 forresta kedo39, are you around?
05:33 clearcla1 joined #salt
05:33 z3uS| joined #salt
05:34 kedo39 yup
05:34 supplica1t joined #salt
05:34 hotbox_ joined #salt
05:34 pviktori_ joined #salt
05:34 nkuttler_ joined #salt
05:34 mortis joined #salt
05:34 tru_tru_ joined #salt
05:34 retr0h_ joined #salt
05:34 baffle_ joined #salt
05:34 oeuftete_ joined #salt
05:35 bejer_ joined #salt
05:35 rockey_ joined #salt
05:36 cedwards1 joined #salt
05:36 btorch joined #salt
05:36 jeff joined #salt
05:36 cyp joined #salt
05:36 jeff joined #salt
05:36 elithrar joined #salt
05:36 favadi joined #salt
05:36 honestly joined #salt
05:36 NV joined #salt
05:36 SEJeff_work joined #salt
05:36 honestly joined #salt
05:36 cryptomnesia joined #salt
05:36 torrancew joined #salt
05:37 xl1 joined #salt
05:37 \ask joined #salt
05:37 akitada joined #salt
05:37 akitada joined #salt
05:37 akitada joined #salt
05:37 Mua joined #salt
05:39 dw_ joined #salt
05:39 litheum joined #salt
05:39 macker joined #salt
05:41 Psi-Jack joined #salt
05:41 jpcw_ joined #salt
05:41 tollmanz joined #salt
05:42 Linz joined #salt
05:43 davet joined #salt
05:43 zpotoloom joined #salt
05:46 whyzgeek joined #salt
05:47 monokrome joined #salt
05:47 drags joined #salt
05:47 ekristen joined #salt
05:48 Ryan_Lane joined #salt
05:49 clearclaw joined #salt
05:49 eclectic_ joined #salt
05:50 tru_tru joined #salt
05:53 kevinbrolly joined #salt
05:53 zpotoloom joined #salt
05:54 z3uS joined #salt
05:57 Linz joined #salt
05:57 sirtaj joined #salt
05:57 wunki joined #salt
05:57 zz_cro joined #salt
05:57 alekibango joined #salt
05:57 torrancew joined #salt
05:57 favadi joined #salt
05:57 cedwards1 joined #salt
05:57 kedo39 joined #salt
06:01 rojem joined #salt
06:05 Gifflen joined #salt
06:18 ossalk joined #salt
06:21 ossalk hi guys I am new to salt. I have a problem that when I run   salt "*" state.highstate ti says Data failed to compile
06:22 ossalk and on minion salt-call state.highstate -l debug --> state: -no , no top file or external nodes data matches found
06:23 druonysus joined #salt
06:23 druonysus joined #salt
06:33 rojem joined #salt
06:41 tollmanz joined #salt
06:42 rojem joined #salt
06:43 middleman_ joined #salt
06:45 sroegner_ joined #salt
06:47 NV joined #salt
06:53 jpcw joined #salt
06:54 ConceitedCode joined #salt
06:54 matanya joined #salt
06:54 rojem joined #salt
06:55 ossalk resolved
06:59 kits2c22e joined #salt
07:03 rojem joined #salt
07:10 cym3try joined #salt
07:22 Furao joined #salt
07:31 auser joined #salt
07:35 jeff joined #salt
07:36 faulkner joined #salt
07:36 steveoliver joined #salt
07:40 teepark joined #salt
07:41 tollmanz joined #salt
07:44 martoss joined #salt
07:48 martoss1 joined #salt
07:50 srage joined #salt
07:52 TomasNunez joined #salt
07:56 gildegoma joined #salt
08:04 matanya joined #salt
08:04 svx joined #salt
08:12 martoss joined #salt
08:15 harobed_ joined #salt
08:16 harobed_ joined #salt
08:19 svx joined #salt
08:25 Iwirada joined #salt
08:26 maned joined #salt
08:27 srage joined #salt
08:32 srage_ joined #salt
08:33 Shenril joined #salt
08:35 njpatel joined #salt
08:41 tollmanz joined #salt
08:41 mnemonikk joined #salt
08:44 helminthe joined #salt
08:46 sroegner_ joined #salt
08:48 srage joined #salt
08:48 linjan joined #salt
09:12 dpn` got this weird stacktrace when using salt with vagrant
09:12 dpn` https://gist.github.com/dpnova/b9d79f03613794fe4839
09:14 BenCoinanke_z joined #salt
09:14 scuwolf joined #salt
09:17 maned left #salt
09:18 srage joined #salt
09:23 ckao joined #salt
09:32 viraptor joined #salt
09:32 taion809 joined #salt
09:38 [diecast] joined #salt
09:39 techdragon joined #salt
09:40 ConceitedCode joined #salt
09:41 tollmanz joined #salt
09:43 gammalget joined #salt
09:46 MrTango joined #salt
09:50 scuwolf joined #salt
09:50 diegows joined #salt
10:10 mekstrem joined #salt
10:10 mekstrem o/
10:13 phuh joined #salt
10:13 phuh Hello, can I build mongodb, redis, elasticsearch cluster with salt?
10:14 Furao phuh: you can do almost everything with salt
10:14 Furao that include those things
10:15 Furao I got formulas for elasticsearch cluster. I do have redis and mongodb formulas, but no clustering support
10:19 phuh Furao: thanks. I'm looking for clustering support. nobody found a solution yet?
10:21 Furao there is no a single way to achieve clustering
10:21 Furao it's not that hard to implement
10:21 Furao I just never needed mongodb or redis clustering
10:21 Furao I support clustering for rabbitmq and postgresql too with salt
10:22 phuh how is salt compared to chef/puppet?
10:22 phuh did you automate those?
10:22 Furao yes
10:22 Furao I don't know much about chef and puppet, except that there is a of lot who move away from those to salt
10:23 phuh oh really
10:29 \ask joined #salt
10:30 andyshinn` joined #salt
10:31 phuh Furao: i can tell salt is very powerful but it seems complex. what's the best way to learn this animal?
10:32 f47h3r joined #salt
10:37 bbinet joined #salt
10:37 Furao phuh: experiment and look at existing samples of formulas
10:38 Furao this is how I learned 1.5 years ago, back then there was not much doc
10:38 Furao here is some https://github.com/bclermont/states
10:39 Furao I don't maintains those publicly anymore as it's now a product/service that I sell
10:41 phuh Furao: thanks- so you are the master of salt
10:41 phuh are you one of devs?
10:41 tollmanz joined #salt
10:42 bbinet joined #salt
10:44 Furao no but I got many merge request into it
10:44 Furao I stop contribute to salt a while ago due to lack of time
10:47 sroegner_ joined #salt
10:48 targetx joined #salt
10:50 targetx hi guys, just started using salt couple of days ago and now trying to deploy my first frontend webserver fully using salt. Mostly it works, but I'm unable to figure out how to add non-free to an apt repository. I've found mod_repo documentation (http://docs.saltstack.com/ref/modules/all/salt.modules.apt.html#salt) but I'm apparently unable to set this up correctly in an SLS file. Keep getting: State apt.mod_repo found in sls $NAME is un
10:50 targetx my definition in sls file:
10:50 targetx non-free-components:   apt.mod_repo:     - repo: 'deb http://ftp.nl.debian.org/debian/ wheezy main'     - comps: main,non-free
10:50 targetx would someone be able to assist me or point me in the right direction perhaps?
10:51 psyl0n joined #salt
10:51 aleszoulek joined #salt
10:51 Furao targetx: use paste services
10:52 Bosch[] joined #salt
10:52 targetx http://pastebin.com/VXS4zwtq
10:52 Furao bad indentation line 4
10:52 Furao remove one space line 4 and 5
10:53 Furao and apt.mod_repo is a module not a state
10:54 Furao http://docs.saltstack.com/ref/states/all/salt.states.pkgrepo.html#salt.states.pkgrepo.managed
10:54 targetx odd... it seems to have added the space while copy/pasting to pastebin. It's not present in the original file.
10:54 targetx thanks, I'll have to try that pkgrepo.managed syntax brb
10:55 pdayton joined #salt
10:55 Furao you can use module.run or module.wait to call module, but there is no reason for you to do that
10:56 targetx Yeah I'm still a bit unclear on CLI vs SLS but it's only been a day or so. Must say I'm very impressed with Salt after having tried to setup Puppet
10:57 Furao I'm so glad that I jumped directly into salt and never suffer trough puppet/chef
10:57 targetx yeah... I just like Python so much better than Ruby.
10:57 targetx and the syntax/fileserver in Puppet was horrible imho. Salt is looking very promising.
10:58 targetx the pkgrepo.managed syntax seems to work Furao! Thanks. I'm restoring my test VM to a clean state now to test a full salt run
10:58 Furao yeah, the python aspect of it helped my choice.
10:58 targetx yeah
10:58 Furao i wrote many custom modules and states so far
10:59 Furao some are in main distribution of salt, I should update some of them
10:59 targetx while I don't need it yet, the possibility is very nice.
10:59 targetx (of writing your own custom py modules)
11:00 Furao in some complex situation I had too
11:00 Furao such as to have automatic discovery for monitoring
11:01 targetx yeah I am probably going to run into that as well
11:01 targetx but at the moment I'm just trying to standardize our configuration & deploy DTAP etc
11:01 Furao dtap?
11:01 targetx development / test / acceptance / production
11:01 Furao ah
11:01 targetx for code development process
11:02 Furao I got formula for jenkins and so on
11:02 Furao I got so much formulas that the guys in my team who authors formulas need to run test jobs on them
11:02 Furao an entire run is 10 to 13 hours
11:03 targetx :O
11:04 targetx We use GitLab + GitLab CI Runner (I think that's like Jenkins?)
11:04 Furao we have a gitlab formula
11:04 targetx nice :)
11:05 Furao but we use jenkins instead of gitlab ci runner
11:05 Furao the integration isn't perfect
11:05 Furao common find . -type f -name '*.sls' | xargs cat | wc -l
11:05 Furao 33592
11:05 Furao ouch 33K LOC just for .sls files
11:06 Furao and just for one repo :)
11:06 targetx hehe
11:07 targetx When I have the Apache2 service 'watch' several folders
11:07 targetx is Salt smart enough to only reload Apache once, at the end? Or will it reload on every change in one of the watched files/folders?
11:07 Furao it will restart only once
11:07 Furao unless you have multiple state.sls execution
11:07 targetx hmmm...
11:08 Furao each of them are separate
11:08 targetx what do you mean by that exactly? I have split it up in multiple sls files
11:08 targetx ah oke..
11:08 Furao if you have a top.sls
11:08 Furao that have roles or include other .sls
11:08 Furao and you run a single big state.highstate
11:08 Furao it's a single execution
11:08 Furao it will only restart apache once
11:08 targetx oke thanks for clarifying
11:09 targetx your repository modification pkgrepo.managed is working fine btw, now I just have the issue that it fails to restart apache
11:09 targetx but when I do it manually after a salt run, it works fine.
11:09 Rory joined #salt
11:10 Rory Is it possible to run a script on the saltmaster as part of a .sls file?
11:10 Furao Rory: yes and no, i don't recommend that
11:10 krak3n` joined #salt
11:11 Furao there is a way to do it, but it's not obvious like cmdmaster.run bleh
11:13 psyl0n joined #salt
11:23 targetx annoying... now all that I'm missing is an Apache reload at the end, then it's working fully automatic
11:24 bhosmer joined #salt
11:27 targetx If I modify some file, which has a - watch_in: - service: apache2, it should restart the service after making that change, right?
11:27 ravibhure joined #salt
11:32 jms3 joined #salt
11:34 Rory left #salt
11:35 ravibhure joined #salt
11:36 ian joined #salt
11:37 Guest27656 can anyone help me with the pkg.mod_repo command
11:40 william_20111 joined #salt
11:41 tollmanz joined #salt
11:44 Corey joined #salt
11:48 Furao joined #salt
11:52 MZAWeb joined #salt
11:59 zooz joined #salt
12:02 thoht hi
12:02 thoht trying to set a grain, it returns: TypeError: string indices must be integers, not str
12:02 thoht i just did  salt 'host' grains.setval role_app tomcat
12:02 thoht what is wrong ?
12:03 krak3n` looks like a python error in your grain
12:03 krak3n` ?
12:03 thoht i did put many grains in this way
12:04 krak3n` check the logs if there is a more complete stack trace, maybe run the minion in debug mode
12:04 thoht when i do: salt -C 'G@role_app:tomcat' cmd.run hostname <== it returns the hostname of where i did set the same grain
12:06 thoht ok i found the error
12:07 thoht there were an error in a sls file
12:07 thoht missing ":"
12:12 blarghmatey joined #salt
12:18 targetx I keep getting 'Failed to reload the service' for Apache2, however when I restart it manually after the salt run it works fine
12:18 targetx how could I get more info about whats going wrong?
12:18 Furao run with logging level at debug
12:20 targetx the strange thing is... on a second run of salt highstate, everything is 100% ok. So it seems an ordering/concurrency issue in my confi
12:20 targetx g. Now restoring VM to original state, will capture log.
12:22 timoguin joined #salt
12:23 vejdmn joined #salt
12:26 gadams999 joined #salt
12:30 Guest27656 it looks like am having the issue as shown here https://github.com/saltstack/salt/issues/9072
12:37 targetx Furao: I've tried with loglevel debug and logged it to a file, however unfortunately it doesn't show apache reload command output (e.g.: why it failed). I have a full log (>8000 lines), would you or someone else be willing to take a look at it perhaps?
12:38 MZAWeb joined #salt
12:41 tollmanz joined #salt
12:42 SkyRocknRoll joined #salt
12:42 SkyRocknRoll joined #salt
12:43 targetx pastebin link in case anyone is willing to take a look: http://pastebin.com/1pPQL5BS. The issue is that the Apache service reload fails during the first salt run. When I reload it manually afterwards or run salt highstate again, it works.
12:44 targetx Line 3585 shows the failing reload
12:45 ossalk joined #salt
12:47 sroegner_ joined #salt
12:53 Furao targetx: look after 3585 you can see that some necessary package get installed
12:53 Furao such as libapache2-mod-fastcgi
12:53 Furao if your config requires these, it's normal that it fail
12:53 Furao and same for php
12:54 targetx hmm...
12:54 targetx then I haven't defined these dependancies correctly
12:54 targetx I'll look over them again, thank you very much for your help'
12:54 Iwirada joined #salt
12:56 micko joined #salt
12:59 ekristen joined #salt
13:00 funzo joined #salt
13:01 basepi joined #salt
13:02 Furao you can tip me in bitcoin at 12STXjmEstqayMF69GXyoc5H6ej2etpKJL :)
13:13 elithrar joined #salt
13:14 vejdmn joined #salt
13:19 crane hi, how does salt knows what he has to install on the server? i mean... how does he knows i want it as a db, web or whatever server?
13:21 Furao crane: http://salt.readthedocs.org/en/latest/topics/installation/index.html
13:25 crane Furao, thx, i will dive into it :)
13:27 harobed joined #salt
13:33 crane Furao, this did not helped me... :/ but i'm not sure if this is working what. i'm running a masterless minion in a vagrant testing lab. now i would like to match my hostnames bob and alice and install different packages on them base: 'bob': - match: pcre - nginx so i wrote this into my top.sls but it seems to be not working
13:38 Pankaj_ joined #salt
13:41 tollmanz joined #salt
13:44 toastedpenguin joined #salt
13:54 jslatts joined #salt
13:55 jakubek joined #salt
14:04 juicer2 joined #salt
14:09 Gifflen joined #salt
14:10 wkf_ joined #salt
14:12 iarmstro joined #salt
14:13 zooz joined #salt
14:13 iarmstro can anyone help me with source option of the pip.installed
14:15 favadi joined #salt
14:15 JasonSwindle joined #salt
14:19 AdamSewell joined #salt
14:19 AdamSewell joined #salt
14:20 mpanetta joined #salt
14:24 racooper joined #salt
14:24 scoates when doing an `include` from a state sls, there's no way to pass parameters in, is there? do I have to use a Jinja include?
14:24 Ahlee_ um.
14:24 Ahlee_ minion01.domain.lo: "test.ping" is not available.
14:25 Gifflen joined #salt
14:26 iarmstro am trying to pip install the python package netifaces, i have specified the source but the minion appears to be attempt to intsll the package for pypi
14:26 iarmstro am trying to pip install the python package netifaces, i have specified the source but the minion appears to be attempt to intsll the package from pypi
14:28 redmin joined #salt
14:28 Ahlee_ iarmstro: I gave up on pip and only use easy_install
14:28 timoguin joined #salt
14:28 JasonSwindle iarmstro:  Do you have your SLS?  I use PIP greatly in my stack.
14:29 Iwirada hi, is there an elegant way to setup a client workstation to periodically check for states to apply?
14:29 iarmstro netifaces:   pip.installed:     - source: salt://blackboxes/pip_files/     - bin_env: /vault/blackbox/
14:30 Iwirada what i can think of now is to put salt-call in to crontab
14:30 JasonSwindle Can you put that into a gists?  That formatting sadly does not help. :)
14:30 Ahlee_ Iwirada: http://docs.saltstack.com/topics/jobs/schedule.html
14:31 iarmstro https://gist.github.com/anonymous/8300117
14:31 Iwirada Ahlee: thank you
14:31 srage_ joined #salt
14:33 cachedout joined #salt
14:33 JasonSwindle iarmstro:  I have never don't it that way, I always just do it like in my reply
14:34 JasonSwindle Sorry I cannot be more of a help.
14:34 iarmstro ok
14:34 iarmstro thanks
14:35 JasonSwindle Geez, I cannot English this morning.
14:35 JasonSwindle Also, what version of start are you using?
14:35 JasonSwindle I think bin_env is going away in some versions....
14:35 Gifflen joined #salt
14:37 iarmstro salt 0.17.3
14:38 JasonSwindle Ok, you may want to go to 0.17.4 because of CVEs
14:38 iarmstro running on centos 6.5
14:39 JasonSwindle But a 0.17.3 Master will not talk to a 0.17.4 minion
14:39 Ahlee again?
14:39 Ahlee JasonSwindle: got a link to discussion around that?
14:39 JasonSwindle Let me find it
14:40 JasonSwindle Here is the release notes for 0.17.3, http://docs.saltstack.com/topics/releases/0.17.3.html prompt people to use 0.17.4
14:40 JasonSwindle looking for the master / minion….. trying to remember where I saw that.
14:40 Ahlee oh, yeah, 0.17.2 to 0.17.4
14:41 tollmanz joined #salt
14:42 JasonSwindle I cannot remember where I read it at, but 0.17.4 is the newest and from the release notes does not look to be advised to be used.
14:42 toastedpenguin joined #salt
14:43 william_20111 joined #salt
14:43 JasonSwindle It was 0.17.1; http://docs.saltstack.com/topics/releases/0.17.1.html
14:43 JasonSwindle My bad, sorry.
14:44 colinbits joined #salt
14:44 Ahlee ok, good. glad not again
14:45 JasonSwindle I rare (read never) update my production master
14:45 iarmstro ok i'll look into that
14:45 JasonSwindle When I need to update my master I roll a new master for the new prod.
14:45 JasonSwindle But not everyone can….
14:46 brianhicks joined #salt
14:46 Ahlee yeah
14:46 Ahlee if i ever find a stable version i'm not touching it
14:46 favadi joined #salt
14:46 iarmstro my master is fedora 16 or 17
14:46 iarmstro i installed using yum
14:47 gadams9991 joined #salt
14:49 srage joined #salt
14:50 JasonSwindle iarmstro:  I gave up on CentOS and of the like for Python.  It was too much of a bother
14:51 th3reverend joined #salt
14:51 th3reverend left #salt
14:51 Ahlee no way i'm swinging this juggernaught to a non-RHEL based distro
14:51 iarmstro what do you use instead?
14:52 JasonSwindle Ubuntu 12.04 LTS
14:52 JasonSwindle Waiting for the next LTS.
14:53 sroegner__ joined #salt
14:53 Ahlee just did a full rebuild of all apps on cent6 with gcc 4.8.x, it wouldn't be financially feasible to do that again to a debian derivative, sadly
14:55 Ahlee oh. neat.
14:55 Ahlee looks like if you ship a _module with the same name as a built in, the _module over rides
14:55 scoates Ahlee: oh, I was going to ask if you had a _modules/test.py … I should have (-:
14:55 tyler-baker joined #salt
14:56 scoates I actively use that override feature to patch stock salt modules, FWIW.
14:56 Ahlee grumble grumble namespaces grumble grumble
14:57 Ahlee scoates: are you able to do any inheritance or fall through?
14:58 scoates Ahlee: I just replace e.g. the git module, entirely instead of trying to monkeypatch
14:58 Ahlee scoates: that's what i figured based on what i just saw
15:00 che-arne joined #salt
15:01 pniederw joined #salt
15:01 pdayton joined #salt
15:01 vejdmn joined #salt
15:03 zooz joined #salt
15:05 william_20111 joined #salt
15:07 Gifflen joined #salt
15:08 pniederw any ideas why the state `pkg.latest(name=salt-master, refresh=True) would say "already up-to-date" even though it's not? the installed version is 0.17.2, but I can see 0.17.4 in the salt-stable PPA (for which I have a `pkgrepo.managed` state).
15:13 mpanetta joined #salt
15:14 iarmstro left #salt
15:14 forresta joined #salt
15:16 Ahlee anybody dove into the zeroMQ reconnection logic?
15:20 mikkn Can you overload existing salt modules?
15:21 mikkn Ahlee: Haha, you answered my question just a few lines above!
15:21 * mikkn bows
15:22 Ahlee happy to be of service, though credit belongs with scoates :)
15:22 scoates :science:
15:23 tollmanz joined #salt
15:28 rlarkin joined #salt
15:29 cedwards joined #salt
15:30 funzo joined #salt
15:32 oz_akan_ joined #salt
15:33 Gifflen joined #salt
15:34 psyl0n joined #salt
15:34 quickdry21 joined #salt
15:35 gadams9991 left #salt
15:36 william_20111 joined #salt
15:40 scoates how do I access pillar data from within a custom module?
15:41 alunduil joined #salt
15:42 danielbachhuber joined #salt
15:42 pdayton joined #salt
15:42 timoguin scoates, the dunder dictionaries! http://docs.saltstack.com/topics/development/dunder_dictionaries.html
15:42 timoguin __salt__['pillar.get']('foo', 'default_value')
15:42 timoguin is one way
15:42 scoates thanks!
15:43 timoguin you can access any of the modules that way
15:43 scoates aside (not really salt related): got any tips for convincing my editor that __salt__ *is* defined, so it stops complaining? (-:
15:43 scoates timoguin: yep; I hadn't considered `pillar.get()`; should work.
15:45 gadams999 joined #salt
15:52 Psi-Jack Ugh.
15:52 Psi-Jack Why does CentOS salt have osfinger, but Debian doesn't? Why does CentOS salt have osmajorrelease and Debian does not?
15:52 Psi-Jack IN grains.
15:53 Whissi joined #salt
15:54 Psi-Jack I mean, same salt version, completely different set of salt-generated grains. Makes no sense and makes for multi-platform states.
15:56 wkf_ joined #salt
15:56 xmltok joined #salt
15:57 Tekni joined #salt
16:00 timoguin Psi-Jack, my Ubuntu instance has osfinger and osrelease
16:00 timoguin but not osmajorrelease
16:01 Psi-Jack Yeah, even different the fact Ubuntu has even more different grains data.
16:02 Psi-Jack Than even Debian!
16:02 Psi-Jack Debian has osrelease, but os_finger is "CentOS-6" on CentOS 6.x
16:02 Psi-Jack And CentOS-5 on CentOS 5.x
16:02 cachedout joined #salt
16:03 xinkeT joined #salt
16:03 Psi-Jack osmajorrelease on CentOS is an array breaking down the release points, So, becomes [6, 4] for CentOS 6.4.
16:03 Linz joined #salt
16:03 timoguin yea seems a bit inconsistent. i'm not sure if that's a bug or what
16:03 foxx joined #salt
16:03 Psi-Jack Err, rather, array(6, 4)
16:04 abe_music joined #salt
16:04 forresta Psi-Jack, you could propose a fix, it's over here: https://github.com/saltstack/salt/blob/develop/salt/grains/core.py#L963
16:04 forresta simple enough how it gets set.
16:04 william_20111 joined #salt
16:04 forresta looks like above that on line 947, debian has osarch
16:05 timoguin that's for x86 vs. amd64
16:05 timoguin but, yea, does look simple
16:05 forresta yea I know, but it should be easy to emulate and make the change for Debian
16:05 Psi-Jack Yeah. And inconsitent there too. :)
16:05 forresta should have been more specific :P
16:06 forresta pull request from Psi-Jack incoming? :D
16:06 Psi-Jack Yeah, when I have time to work on it.
16:06 forresta ok, if an issue doesn't exist at least create that if you have time.
16:07 SEJeff_work Is 2014.1 an official release?
16:07 JasonSwindle SEJeff_work:  RC1, yes
16:07 forresta SEJeff_work, once it's past the RC yea
16:07 JasonSwindle I have been poking at it.
16:08 SEJeff_work Why is there a 2014.1 tag?
16:08 SEJeff_work Seems wrong for there to be a v2014.1 tag if it hasn't been officially released yet
16:08 SEJeff_work Kind of defeats the purpose of a tag :)
16:08 Psi-Jack For now, I'm implementing horrible hacky if grains['os'] == 'CentOS' and grains['osrelease'].startswith('6.') type condition checks.
16:08 davidfischer joined #salt
16:08 forresta SEJeff_work, they did the same thing for 0.17
16:08 SEJeff_work Yeah seems wrong
16:09 SEJeff_work how not to use git :)
16:09 Psi-Jack It is wrong. :)
16:09 Psi-Jack You only tag the actual official release, not some version in between.
16:10 bbinet hi
16:10 forresta hi
16:10 Psi-Jack Tagging an RC should be 20141.RCx, where x is an optional number if there's multiple RC's planned/intended/needed.
16:10 Psi-Jack Err, 2014.1.RCx
16:10 SEJeff_work with a lower case rc, but yes :)
16:11 bbinet I'm wondering if bugfix pull request merged in develop branch will be available in the 2014.1 release?
16:11 timoguin bbinet, generally bugfixes get cherry picked from the dev branch to the release branches
16:11 forresta bbinet, they will cherry pick which fixes go into the release candidates, so most likely yes
16:11 timoguin forresta, i win!
16:11 forresta damn you timoguin
16:11 forresta wasting my keystrokes here
16:12 bbinet timoguin: do we need to ask somewhere for these cherry-picks?
16:12 forresta nope
16:12 bbinet for example https://github.com/saltstack/salt/pull/9562 and https://github.com/saltstack/salt/pull/9570 are bugfix
16:13 bbinet but not marked as bugfix ;)
16:13 numberx joined #salt
16:13 erjohnso joined #salt
16:14 JasonSwindle What timezone does S0undT3ch work?
16:14 JasonSwindle Or however you spell his name. :)
16:14 bhosmer joined #salt
16:14 forresta he's in Spain
16:14 JasonSwindle Ah, ok.
16:14 forresta and he doesn't come into this channel any longer
16:14 forresta so you would need to message him directory
16:14 forresta s0undt3ch for the spelling
16:15 JasonSwindle Ok.  I am hoping to work with him on a Salt-Cloud / Bootstrap issue.
16:15 forresta well, not spain, but portugal I think
16:15 forresta so next door to spain
16:15 JasonSwindle Ok, good to know.  Thank you!
16:15 numberx I can't remember what salt calls its puppet-forge equivalent... ?
16:15 forresta yup, and he works nights over there too JasonSwindle, so usually you can catch him during the day here
16:16 forresta numberx, https://github.com/saltstack-formulas/
16:16 funzo joined #salt
16:16 numberx yes thank you
16:16 bhosmer joined #salt
16:18 forresta np
16:20 forresta terminalmage are you around this morning?
16:20 xinkeT joined #salt
16:21 elfixit joined #salt
16:23 ml_1 joined #salt
16:23 jdenning joined #salt
16:29 vlcn joined #salt
16:29 vlcn is there a way to group operators in the definition of a group?
16:30 Linz joined #salt
16:30 vlcn something like: G@virtual:openvzhn and (foo.foo.com or foo.fi.com)
16:30 rgbkrk joined #salt
16:31 Iwirada left #salt
16:31 anuvrat joined #salt
16:34 dave_den vlcn: use pcre in the compound match:  G@virtual:openvzhn and E@foo.(foo|fi).com
16:39 abe_music is there a spot in the config to control how many processes salt-cloud uses when launching machines in parallel?
16:40 terminalmage forresta: yes
16:41 UtahDave joined #salt
16:41 terminalmage steveoliver: darless: The incorrect line being reported in jinja render errors has been fixed
16:43 rmt joined #salt
16:43 Ahlee what are people using for returner backends?
16:49 timoguin i'm planning on implementing the redis returner, with redis sending everything to logstash/elasticsearch
16:49 Ahlee i was just going down that thought process
16:50 srage joined #salt
16:53 Ryan_Lane joined #salt
16:54 Linz joined #salt
16:54 vejdmn joined #salt
16:55 bhosmer joined #salt
16:56 KyleG joined #salt
16:56 KyleG joined #salt
16:57 UtahDave forresta: ANNOUNCEMENT:  I'm here!  good morning!
16:59 davet joined #salt
16:59 UtahDave :)
16:59 KyleG wasn't that the same announcement as yesterday
16:59 KyleG lol
17:01 forresta UtahDave, woah, a little too much excitement there
17:02 forresta terminalmage, sorry I was in a meeting. Could you mark https://github.com/saltstack/salt/pull/9562 and https://github.com/saltstack/salt/pull/9570 as bug fixes so they get pulled for rc2?
17:02 vejdmn joined #salt
17:03 JasonSwindle UtahDave:  Howdy!
17:04 JasonSwindle UtahDave:  I updated my issue, https://github.com/saltstack/salt/issues/9587
17:05 UtahDave ah, good. thanks, JasonSwindle
17:05 UtahDave I'm going to do some testing on that myself, today.
17:05 JasonSwindle Any time.  I will be here.  Just message
17:05 terminalmage forresta: usually I just add a comment for basepi
17:05 vlcn why would the id that is selected for a minion differ from the output of 'hostname'?
17:05 terminalmage that it needs to be cherrypicked
17:06 modafinil does anyone have good ideas on how to handle secret stuff getting to machines (like ssl certs) -- pillar seems... only sort of helpful, but maybe i'm just not seeing it
17:06 basepi forresta: yes, if you mention me in the pull req (not the commit), and tell me to cherry-pick, that's the best way
17:06 basepi i look at every pull req when cherry-picking, but *not* every commit
17:07 kermit joined #salt
17:08 Psi-Jack UtahDave, basepi: So is that osmajorelease and os_finger mismatch between distros considerable as a bug that CentOS and Ubuntu has os_finger, but Debian does not, like-wise, only CentOS has osmajorrelease and neither Ubuntu and Debian do?
17:10 teskew joined #salt
17:10 martoss left #salt
17:12 UtahDave Psi-Jack: Yeah, I would file that as a bug. Those should be standardized
17:12 dijit left #salt
17:12 Psi-Jack Yeah, I thought so too. :)
17:12 forresta basepi, terminalmage, ahh yea fair, I didn't make those, just wanted to make sure they were in rc2.
17:12 cym3try joined #salt
17:13 Ahlee vlcn: check contents of /etc/salt/minion_id
17:14 Ahlee if /etc/hosts doesn't properly contain a hostname, /etc/salt/minion_id can end up being localhost.localdomain
17:15 Psi-Jack What does osfinger look like in Ubuntu, timoguin?
17:15 terminalmage vlcn: http://docs.saltstack.com/topics/tutorials/walkthrough.html#minion-id-generation
17:16 timoguin Psi-Jack, Ubuntu-12.04 in this case
17:16 Psi-Jack Hmm. Odd...
17:16 terminalmage Salt caches the minion id in /etc/salt/minion_id
17:16 Psi-Jack I guess it works.. Though... Ubuntu-style...
17:19 vlcn terminalmage: thanks, but I still can't figure out wtf it is doing
17:19 vlcn hostname is defined in /etc/sysconfig/network on all of these boxes
17:19 higgs001 joined #salt
17:20 vlcn I can't figure why the hostname is picked up on some and not others
17:20 Ahlee That file isn't checked
17:20 terminalmage vlcn: that was quite straightforward... did you try to run socket.getfqdn() from the python shell?
17:20 Ahlee unless you're symlinking /etc/hosts to /etc/sysconfig network
17:21 ekristen I know I can use salt to do this, but looking for docs on how to help me do it … I want to tell minions to checkout a specific git repo with tag on my web servers
17:21 erjohnso left #salt
17:22 terminalmage ekristen: http://docs.saltstack.com/ref/states/all/salt.states.git.html
17:22 vlcn terminalmage: okay, but on otherwise identical servers what would cause python to resolve resolve the hostname on one and both the other
17:22 jefferai joined #salt
17:34 vlcn odd... socket.gethostname() gives me the expected result
17:34 vlcn socket.getfqdn() does not.
17:34 JasonSwindle basepi:  UtahDave http://www.cvent.com/events/saltconf-2014/custom-18-3aef0eba23dd49a1bd9ffcaaf83d657e.aspx  <-  On the last day, Jan 30….
17:34 terminalmage vlcn: have you checked /etc/salt/minion_id ?
17:34 ronc joined #salt
17:34 JasonSwindle You guys may want to edit the times on the last block.  It should go breakout & lab then lunch and then breakout again.  Right now if you are not looking, it looks like the conf ends at 12:30 PM.
17:34 vlcn terminalmage: thats not the point, I'm trying to figure out WHY that is happening so I don't need to keep fixing it.
17:34 ajw0100 joined #salt
17:34 JasonSwindle But really, it looks like the conf ends at 2:15 PM
17:34 terminalmage ok, that might be a question for #python then, they might know more there about what influences socket.getfqdn()
17:34 basepi JasonSwindle: can you just e-mail saltconf@saltstack.com with that correction?
17:34 basepi Rhett will handle it
17:34 Ahlee gethostname() vs getfqdn() is hosts file assuming your nsswitch is "standard"
17:34 JasonSwindle Sure!  I can do that… :)
17:34 troyready joined #salt
17:34 scooby2 joined #salt
17:34 UtahDave JasonSwindle: Ha! good catch.
17:34 mephx_ joined #salt
17:34 vlcn Ahlee: /etc/hosts is identical on both
17:34 Ahlee vlcn: now, sure. but at install?
17:34 JasonSwindle UtahDave:  I think that may of bit me…My flight out is at 2:20, but no idea when my break out session is.  Eep!
17:34 Ahlee also, shouldn't they differ based on hostname
17:34 jtrucks_ joined #salt
17:34 vlcn Ahlee: the only thing in /etc/hosts is this: https://gist.github.com/kelchm/f33f3b1197a63b534b63
17:34 trevorjay joined #salt
17:34 forresta JasonSwindle, you're flying out at 2 PM Thursday?
17:34 JasonSwindle Yep/
17:34 joehillen joined #salt
17:34 MTecknol1gy joined #salt
17:34 JasonSwindle Why do you ask? :)
17:34 forresta just wasn't sure
17:34 JasonSwindle I would *love* to stay longer, but the show must go on.
17:34 forresta yea I'm flying back in on Friday, and then going to work most likely
17:34 Ahlee vlcn: ok, so you fail #1 (socket.getfqdn()), you fail #2 presumably, and you don't have a hostname set in /etc/hosts
17:34 jacksontj joined #salt
17:34 Ahlee vlcn: thus, you get localhost.localdomain
17:34 vlcn socket.getfqdn() doesn't fail, it just doesn't give me what I'm expecting
17:34 Ahlee variable level of fail, then
17:35 psyl0n joined #salt
17:35 smkelly joined #salt
17:35 jalbretsen joined #salt
17:35 jtrucks joined #salt
17:37 vlcn ah, shit.  I see what it is now.
17:37 vlcn socket.getfqdn() actually does a reverse dns lookup
17:37 abe_music joined #salt
17:38 vlcn apparently have a few servers that never had the record set correctly.
17:39 xmltok joined #salt
17:41 amckinley joined #salt
17:42 jtrucks joined #salt
17:42 danielbachhuber- joined #salt
17:42 mpanetta_ joined #salt
17:43 pdayton1 joined #salt
17:43 Tekni joined #salt
17:44 erjohnso joined #salt
17:45 druonysuse joined #salt
17:45 druonysuse joined #salt
17:47 harobed_ joined #salt
17:47 srijan4_ joined #salt
17:47 harobed_ joined #salt
17:48 jtrucks joined #salt
17:48 steveoliver_ joined #salt
17:49 dpn`` joined #salt
17:49 cdcalef joined #salt
17:50 harobed_ joined #salt
17:50 modafinil_ joined #salt
17:50 abele_ joined #salt
17:50 akitada_ joined #salt
17:51 nliadm joined #salt
17:51 Kyle joined #salt
17:51 markizano joined #salt
17:51 blee joined #salt
17:51 scooby2 joined #salt
17:51 Jahkeup joined #salt
17:52 seanz joined #salt
17:52 dpac|away joined #salt
17:52 [vaelen] joined #salt
17:52 scooby2 joined #salt
17:52 octarine joined #salt
17:53 Ryan_Lane joined #salt
17:53 Ryan_Lane joined #salt
17:53 druonysus joined #salt
17:53 toastedpenguin joined #salt
17:53 toastedpenguin left #salt
17:54 druonysus joined #salt
17:54 jrdx joined #salt
17:54 crane joined #salt
17:54 jtrucks joined #salt
17:55 echos joined #salt
17:56 harobed_ joined #salt
17:57 harobed_ joined #salt
17:58 harobed_ joined #salt
17:58 harobed_ joined #salt
17:59 harobed_ joined #salt
18:00 harobed_ joined #salt
18:01 druonysus joined #salt
18:01 druonysus joined #salt
18:01 harobed_ joined #salt
18:02 william_20111 joined #salt
18:02 MTecknol1gy http://dpaste.com/1540970/ ... it's just hanging here... :(*
18:02 matanya joined #salt
18:03 tollmanz joined #salt
18:03 jtrucks joined #salt
18:04 jimallman joined #salt
18:05 jtrucks joined #salt
18:06 harobed_ joined #salt
18:07 dangra joined #salt
18:07 harobed_ joined #salt
18:07 sschwartz joined #salt
18:08 william_20111 joined #salt
18:08 harobed_ joined #salt
18:08 vejdmn joined #salt
18:09 harobed_ joined #salt
18:09 [vaelen] joined #salt
18:09 nliadm joined #salt
18:10 markizano joined #salt
18:10 Jahkeup joined #salt
18:10 scooby2 joined #salt
18:10 octarine joined #salt
18:10 scooby2 joined #salt
18:11 sschwartz OK: I am confused, and coming here for help. Using salt-cloud, I repeatedly get this error on invocation: There is no AWS cloud provider configuration available. Not loading module -- however, the salt-cloud is quite willing to actually *create* an EC 2 instance, it just won't let me run other commands (like ec2.show-instance) on it. Any idea what I might be doing wrong today?
18:11 seanz joined #salt
18:11 Kyle- joined #salt
18:11 echos joined #salt
18:11 jtrucks joined #salt
18:12 harobed_ joined #salt
18:12 gadams9991 joined #salt
18:12 william_20112 joined #salt
18:13 scoates from within a custom module, is it easy to render a template (or if much easier: fetch a file) from the master? are there helpers for this? links please? (-:
18:14 harobed_ joined #salt
18:14 forresta scoates, did you take a look at existing contributions to salt-contrib? https://github.com/saltstack/salt-contrib/tree/master/modules
18:14 forresta I'm not sure how you would do that
18:15 harobed_ joined #salt
18:16 harobed_ joined #salt
18:16 dave_den scoates: it's possible but not easy, at least last time i looked. you'd have to import the salt libs and call through a few functions directly to get a proper render.
18:16 druonysus joined #salt
18:16 dave_den there are a lot of things that would benefit from a refactor into more classful layout
18:16 harobed_ joined #salt
18:17 scoates forresta: I don't see anything for templates in there
18:17 mayfield joined #salt
18:17 forresta :\
18:17 jtrucks joined #salt
18:17 harobed_ joined #salt
18:18 scoates dave_den: that would also allow a refactor from `source: … template: jinja …` into `contents: {{salt['template.rendererthing'](source, params)}}` if desired.
18:18 scoates I think I'll just manage it with a separate state and use an include from within my nginx config.
18:19 harobed_ joined #salt
18:19 Teknix joined #salt
18:19 forresta that sounds easier
18:19 harobed_ joined #salt
18:20 harobed_ joined #salt
18:21 sschwartz (Let me check: is this the right place to ask a salt-cloud question, or is there somewhere better?)
18:21 harobed_ joined #salt
18:22 dave_den sschwartz: yes, it is
18:22 UtahDave sschwartz: what version of Salt Cloud and Salt are you using?
18:22 dave_den also the mailing list
18:22 harobed_ joined #salt
18:23 sschwartz_ joined #salt
18:23 sschwartz_ salt-cloud --versions-report       Salt Cloud: 0.8.11           Python: 2.6.9 (unknown, Oct 29 2013, 19:58:13)             Salt: 0.17.2  Apache Libcloud: 0.14.0-beta3           PyYAML: 3.10
18:23 harobed_ joined #salt
18:23 sschwartz_ (Sorry, IRC dropped me.)
18:24 jtrucks joined #salt
18:24 harobed_ joined #salt
18:26 pexio joined #salt
18:27 harobed_ joined #salt
18:27 sschwartz_ Perhaps the python is too old?
18:27 gadams9991 left #salt
18:28 UtahDave sschwartz_: OK, so that warning about AWS doesn't matter
18:28 UtahDave I removed it in the latest RC of Salt, which includes Salt Cloud now.
18:29 UtahDave You should be using the EC2 driver. There used to be an AWS driver that is now deprecated, but still included.
18:29 sschwartz_ I didn't think it did, but when I try and invoke things like "salt-cloud -a show-instance", I get: 'ec2.show-instance()' is not available. Not actioning...
18:29 sschwartz_ I am using the ec2 driver, in my cloud config.
18:29 sschwartz_ provider: ec2
18:29 Gareth morning
18:29 UtahDave good.  Let me check something here locally
18:30 sschwartz_ Cool. Thanks.
18:30 eric_H joined #salt
18:31 UtahDave sschwartz_: do you get a result from     salt-cloud --list-sizes ec2        ?
18:31 sschwartz_ Yes.
18:34 harobed_ joined #salt
18:35 UtahDave sschwartz_:   try this:    salt-cloud -a show_instance 'INSTANCE ID'
18:35 eric_H UtahDave, sorry to bother you guys, I'm going with the Rackspace crew to the SaltConf and notice that there are 3 different training classes we can take... do all allow you to take the certification exam at the end of the day?
18:35 sschwartz_ Interesting. That works. Cool, and many thanks.
18:37 UtahDave sschwartz_: you're welcome.  you'll notice here that all the functions have underscores, not dashes.  http://docs.saltstack.com/ref/clouds/all/salt.cloud.clouds.ec2.html#module-salt.cloud.clouds.ec2
18:38 troyready joined #salt
18:38 UtahDave eric_H: I just checked.  Yeah, you can take any of the classes and take the SSCE test
18:38 sschwartz_ Oy. Figures that's what tripped me up.
18:40 eric_H UtahDave, awesome and I'm looking forward to it, would you recommend taking an advanced class to be better prep'd?  I've been using salt but have no formal training so wouldn't wanna miss the intro class either.
18:41 Linz joined #salt
18:42 srage_ joined #salt
18:43 UtahDave eric_H: Hm.  Tough question!
18:43 JasonSwindle joined #salt
18:44 UtahDave I would definitely recommend the first or second class for you.  If you want to make sure you haven't missed any fundamentals, the Introduction will be a good class.
18:44 eric_H JasonSwindle, hey dude, i was just asking UtahDave about the training classes. looks like we can take the exam after any of the 3 classes given that tuesday
18:45 eric_H UtahDave, that's a good point, thanks for the help.
18:45 UtahDave eric_H: Have you looked at the descriptions?  http://www.cvent.com/events/saltconf-2014/custom-20-3aef0eba23dd49a1bd9ffcaaf83d657e.aspx
18:45 eric_H ahh, i was only reading the small blubs on the previous page thank you again.
18:45 eric_H or the registration page rather
18:46 eric_H Brew, you in here too? I'm late to the game.
18:47 EvaSDK joined #salt
18:47 EvaSDK joined #salt
18:47 MTecknology joined #salt
18:47 pdayton joined #salt
18:47 bejer joined #salt
18:47 jimallman joined #salt
18:47 numberx joined #salt
18:47 jrdx joined #salt
18:47 rmt joined #salt
18:47 jdenning joined #salt
18:48 druonysus joined #salt
18:48 druonysus joined #salt
18:48 higgs001_ joined #salt
18:48 matanya joined #salt
18:49 defunctzombie joined #salt
18:49 teskew joined #salt
18:50 Mua joined #salt
18:53 Mua_ joined #salt
18:57 psyl0n joined #salt
19:01 rmt joined #salt
19:01 dave_den joined #salt
19:04 sschwartz joined #salt
19:05 sschwartz Does anyone know if AWS works with start actions in salt cloud?
19:05 robbyt joined #salt
19:06 sschwartz (I am having trouble getting the reactor to respond to a new VM from salt-cloud, and if this works, it would be much simpler.) My tests seem to suggest no, but I don't know if that's because I'm doing it wrong. ;)
19:06 markizano joined #salt
19:06 Kyle joined #salt
19:06 Jahkeup joined #salt
19:06 seanz joined #salt
19:06 nliadm joined #salt
19:06 [vaelen] joined #salt
19:06 Ryan_Lane joined #salt
19:06 Kyle joined #salt
19:07 faeroe joined #salt
19:07 echos joined #salt
19:09 amckinley1 joined #salt
19:10 jslatts joined #salt
19:13 Guest16277 joined #salt
19:13 JasonSwindle joined #salt
19:14 JasonSwindle Did someone page me?
19:14 JasonSwindle I am unable to find it….
19:14 zach 19:43 < eric_H> JasonSwindle, hey dude, i was just asking UtahDave about the training classes. looks like we can take the exam after any of the 3 classes given that  tuesday
19:15 JasonSwindle Ah, ok
19:15 zach Also, Austin or San Antonio?
19:15 zach I think I know you
19:15 JasonSwindle Me?
19:15 zach Yeah
19:15 JasonSwindle SATX
19:15 timoguin speaking of that certification. anyone know if/when it might be available after the conference?
19:16 zach Roger, I feel bad for you working @ The Castle :P
19:16 timoguin i won't be able to go, but i'd love to get that cert
19:16 JasonSwindle zach:  It is all I know.  I really like SFO. :)
19:16 zach I wish I could go to SaltConf. I had teh go ahead, funding was promised....then I was told "nope!"
19:16 JasonSwindle That sucks.
19:16 zach aye. It happens. Reason I'm quiting my job ;-)
19:17 quickdry21 joined #salt
19:17 wkf_ joined #salt
19:22 Gifflen joined #salt
19:24 jacksontj joined #salt
19:27 abe_music zach: i hear ya...it sucks :(
19:27 cro joined #salt
19:27 srage joined #salt
19:34 bhosmer joined #salt
19:42 Ahlee so continue to get http://pastebin.com/gPby9ciZ across three environments
19:42 Ahlee ZMQError: Operation cannot be accomplished in current state
19:44 Ahlee No network changes, all ACLs between minion/master removed (unfettered connectivity previously as well on 4505 and 4506)
19:45 Ahlee master has ~200 minions, i have 199 established connections (unknown number of minions are offline due to that error causing them to disconnect
19:46 zach Which version of minion/master?
19:46 zach I've ran into a similar issue before
19:46 Ahlee 0.17.2
19:47 * zach checks his version
19:47 Ahlee python 2.6, zmq 3
19:47 zach Same version, checking the python and zmq
19:47 zach RHEL6?
19:47 Ahlee yes
19:48 zach We're both at the same versions
19:48 Ahlee seeing same issue?
19:48 mohae joined #salt
19:48 Ahlee funnily enough, since new year?
19:48 zach I have the issue where minions will go out of sync and disconnect
19:48 Ahlee out of sync?
19:48 zach yeah, as in they will just lose connection for some unknown reason
19:48 Ahlee ah, yeah.
19:49 zach possibly network related as our F5's fail-over time to time and don't keep states in mind
19:49 zach so things get disconnected
19:49 Ahlee no load balancer here
19:50 Ahlee i doubt my issue is networking related, our multicast apps aren't sputtering in the slightest
19:50 zach Yours isnt, mine is ;)
19:50 Ahlee 10G throughout, nothing approaching 100% utilization
19:50 zach Are all of yours running the same version across the board?
19:50 Ahlee pkts/sec within reasonable limits to the machines
19:50 Ahlee yes, i updated all non 0.17.2 hosts over the weekend
19:51 Ahlee i was assuming some <0.17 was causing hte master to freak
19:51 zach yes
19:51 Ahlee so i crawled through all hosts in AD
19:51 zach Are you able to use salt-call on the minions?
19:52 Ahlee salt-call test.ping worked
19:52 Ahlee let me put together a little more convoluted example
19:52 zach One thing I've done to help me get around a lot of the issues is to run salt-call state.highstate in cron, that has helped me out significantly in terms of keeping everything up
19:53 Ahlee will take it into consideration, but we're trying not to use cron
19:53 zach Agreed, it's a dirty hack
19:53 Ahlee salt-call cp.list_master ishanging
19:54 zach No firewall related issues going on? Anything weird in your salt log?
19:54 Ahlee so, there's definitely something wrong in the commuincation level
19:54 Ahlee i dropped all acls between the hosts for testing
19:54 zach you can nc master port ?
19:54 zach or telnet
19:55 Ahlee yep, both
19:55 zach interesting
19:55 zach What is going on in both the master and minion salt logs?
19:55 Ahlee i've moved systems to same vlan, same switch
19:55 zach Any errors?
19:55 Ahlee no acls between them
19:55 amckinley joined #salt
19:55 Ahlee run starts, zeromq error gets logged, minion goes dark for indeterminate time
19:57 zach one moment, checking something on my deployment
19:57 numberx joined #salt
19:58 Ryan_Lane2 joined #salt
19:59 gadams999 joined #salt
19:59 Ryan_Lane2 joined #salt
19:59 numberx how can I roll back a file.recurse call matching against the files sourced in the master? .. so rm -rf recursively ?
19:59 thrashr888 joined #salt
20:00 Ahlee https://github.com/saltstack/salt/blob/develop/salt/payload.py#L182
20:00 Ahlee i'm going to dump the contents of pkg
20:00 Ahlee see what happens
20:01 Ryan_Lane2 joined #salt
20:02 Ryan_Lane2 joined #salt
20:03 zach So the problem I am facing qith zeromq is it is not as awesome as rabbitmq to where you can just look at the queue easily
20:03 Ahlee indeed
20:03 zach So my idea is out the window
20:03 Ahlee and not full blow multicast where you can just subscribe
20:03 Ahlee not really
20:03 Ahlee there's a way to subscribe to the zmq listener
20:03 Ahlee let me see how
20:03 zach I was going to watch the queue while I restarted a minion
20:04 zach To see what was going on
20:04 phuh joined #salt
20:04 Ahlee zach: in tests/eventlisten.py
20:04 zach if UtahDave or basepi were watching, they might know what is going on
20:05 Ahlee # python eventlisten.py
20:05 Ahlee ipc:///var/run/salt/master/master_event_pub.ipc
20:05 * basepi hides
20:05 Ahlee so, now just to trigger junk against it
20:05 zach :-P
20:05 basepi sorry, what?
20:05 zach basepi: Ahlee is running into zmq issues where the minions just drop off the face of the planet and refuse to work
20:05 Ahlee we're both having unusual zeromq issues
20:05 Ahlee http://pastebin.com/gPby9ciZ
20:06 basepi Ahlee: does the minion come back after a restart?
20:06 Guest64230 joined #salt
20:06 Ahlee one minute
20:06 Ahlee er, i'll let you know in one minute
20:07 Ahlee here's the output from eventlisten.py during a call
20:07 Ahlee http://pastebin.com/0UTaJuvH
20:07 Ahlee so, there's connectivity
20:07 Ahlee as the minion is reporting back to the master it's unable to complete the operation
20:08 Ahlee this isn't something stupid with certificate expirations, is it?
20:08 Ryan_Lane2 joined #salt
20:09 Ahlee Yes, the minion does reconnect
20:09 Ahlee on restart
20:09 Ahlee i can issue trivial items like test.ping
20:09 rjc joined #salt
20:09 basepi Ahlee: what version of zmq on the minion/master?
20:10 Ahlee but a simple state that ensures nc is installed, rsyslog is running, and executes a script fails
20:10 basepi salt-call --verisons-report
20:10 basepi or salt --versions-report
20:10 basepi (except i spelled it wrong the first time)
20:10 indymike joined #salt
20:10 basepi Ahlee: as in, fails every time?
20:11 Ahlee basepi: http://pastebin.com/aVE1tGsV
20:11 Ahlee one sec
20:11 basepi ya, those versions look fine.....
20:12 Ahlee http://pastebin.com/aF1uNLA8
20:13 Ahlee going to dump contents of pkg in https://github.com/saltstack/salt/blob/develop/salt/payload.py#L182
20:13 Ahlee then open a ticket
20:13 basepi does the staterun fail every time?  or is it inconsistent?
20:13 Ahlee inconsistent
20:14 Ahlee rarely succeeds
20:14 Ahlee but, some hosts do
20:14 Ahlee no pattern yet identified
20:14 Ahlee current theory is first X to respond complete
20:15 basepi o.O
20:15 basepi what are network conditions like?
20:15 tollmanz joined #salt
20:16 aleszoulek joined #salt
20:16 dfinn joined #salt
20:16 Ahlee master resides on a mostly idle vmware 5 system, dual 10G ethernet, of which we're using approximately 20 Mbps
20:16 Ahlee minions reside on various hardware (virtual and phyiscal)
20:16 Ahlee all 10G
20:16 Ahlee nowhere near saturation or dangerous packets/sec limits
20:17 Ahlee no network discards on any switch port, no errors present
20:17 dfinn man these python errors can be pretty unhelpful.  anyone know what's wrong here?  error and sls included in paste
20:17 dfinn http://pastebin.com/sT4B5VbM
20:17 Ahlee so, as heathy as I can estimate
20:17 Ahlee both physical and virutal affected
20:18 basepi Ahlee: crazy!  hate those zmq errors, they're rarely (if ever) helpful.
20:18 basepi but it certainly doesn't sounds like network conditions would cause any issue
20:18 ekristen I’m getting The Salt Master has cached the public key for this node, this salt minion will wait for 10 seconds before attempting to re-authenticate
20:18 ekristen and it just keeps cycling
20:18 tollmanz_ joined #salt
20:19 Ahlee ekristen: salt-key -y -a <hostname> on the master
20:19 pniederw joined #salt
20:19 Ahlee dfinn: sorry, i can't get the simple things working, havne't started on more-than-trivial states
20:20 ekristen Ahlee: nothing, maybe it can’t talk
20:20 ekristen to the master
20:20 dfinn open file limit?  I ran into that when I first built my master.
20:21 basepi Ahlee: nothing else of note in the logs besides that ZMQ error?
20:22 pniederw Is `pkgrepo.managed` supposed to affect the repositories used by a manual `apt-get` (on Ubunutu)?
20:23 Ahlee just the same error repeatedly
20:23 Ahlee as we call this once an hour
20:23 Ahlee mgmt host -> API -> minions
20:23 Ahlee on an hourly schedule
20:24 Ahlee ulimit -n reports 32768 files
20:24 Ahlee i'll up it
20:24 Ahlee but i'd assume i'd be seeing file handle errors
20:24 dfinn looks like I have mine set to 80000
20:24 dfinn I don't think I saw errors when I ran into the limit
20:24 dfinn how many minions?
20:25 jacksontj joined #salt
20:25 Ahlee 200 or so
20:25 ekristen how do I tell salt-minion the correct hostname?
20:25 ekristen nm
20:25 ekristen found it
20:25 ekristen Ahlee: I verified that ports 4505/4506 are open
20:25 dfinn that's about where I'm at Ahlee
20:26 ekristen I see no activity in the master log
20:26 Ahlee ekristen: are you running the master in at least info?
20:26 Ahlee my master/minions default to no logging
20:27 ekristen oh I have to accept it
20:27 Ahlee and yes, unless you run open mode (i do)
20:27 Ahlee you need to accept the key with salt-key
20:27 micah_chatt joined #salt
20:27 Ahlee if you're just starting, a helpful command is salt-key -A
20:27 Ahlee to just accept all pending keys
20:28 ekristen Ahlee: ok this is in a private network, I’ll go google how to do open mode
20:28 Gifflen joined #salt
20:28 ekristen oh salt-key -A is convenient too
20:28 ekristen thanks
20:28 kaptk2 joined #salt
20:28 Ahlee open_mode: True
20:28 Ahlee in your /etc/salt/master :)
20:28 Ahlee keep in mind that's dangerous thus off by default
20:28 Ahlee as any minion can claim to be another minion
20:28 ekristen yes i understand
20:29 Gifflen joined #salt
20:29 ekristen question — I have a top.sls defined with a base: ‘*’ and a salt to install users
20:29 ekristen shouldn’t the ‘*’ match any minion that joins?
20:29 Ahlee that has a key accepted
20:30 basepi Ahlee: I'm fairly stumped.  You said you can do basic remote execution, like test.ping.  Can you do longer remote execution routines, like installing a package?  Curious if it's limited to states, or just to the complexity of the operations.
20:30 ekristen Ahlee: then I think I did something wrong
20:31 basepi Ahlee: also, I'd be interested to see eventlisten.py running on the master, and then keep running commands on a single minion until it fails.  see if we can see when it fails in terms of events
20:31 TheRealBill joined #salt
20:32 Ahlee i bumped nproc up to 80000 and file handles to 128k
20:32 Ahlee seeing if i have egg on my fase
20:32 ekristen I have file_roots: base: - /srv/salt
20:32 Ahlee s/fase/face/
20:33 ekristen Ahlee: do I need to run highstate or anything like that on the minion or from the master? or when it joins it should just process the states?
20:34 Ahlee join does nothing
20:34 Ahlee unless you have a reactor set up to do something on join
20:34 Ahlee so, i fyou want to run a state, you have to tell the minion to
20:34 Ahlee i typically salt '*' test.ping first to make sure it can talk
20:34 basepi Ahlee: also, i'd love if you opened an issue.  we should be tracking all of these pastebins
20:34 thrashr888 joined #salt
20:35 ekristen Ahlee: I did that, that worked
20:35 Ahlee basepi: yeah, i had to get the salesforce login from my boss
20:35 Ahlee unless you'd prefer this issue through github
20:36 ekristen Ahlee: so I guess I’m not making the connection, if i have a top.sls defined with a base ‘*’ that references another state, ie to create users, how do I run the base from the top.sls, is that what highstate is for?
20:36 Ahlee restarted master with debug logging, capturing eventlisten.py output to a log
20:36 basepi Ahlee: Maybe both places?  I don't currently have a personal salesforce login, so I prefer Github.
20:37 Ahlee ok. let me continue to gather data an i'll toss 'em up to both, and reference the GH in salesforce
20:37 ekristen I really dislike salesforce
20:37 Ahlee ekristen: you'd call state.highstate, but by default that doesn't run on a minion
20:38 Ahlee unless you have a areactor configured to run it on join
20:38 basepi Ahlee: sounds good.
20:38 ekristen got it, I’ll check out reactors, thanks
20:42 amckinley joined #salt
20:43 ekristen another question — stuff referenced by salt:// — that goes in the file_roots directory right?
20:43 UtahDave ekristen: yep
20:44 ekristen hrm, ok, got to figure out why it can’t find the file then
20:44 ekristen minion said it couldn’t find the files
20:46 funzo joined #salt
20:46 ekristen woohoo, ok I got my users to add
20:46 ekristen but my installation of a debian package is failing
20:51 Ahlee doh. forgot to add a longer timeout
20:52 ekristen whats the proper module to use if I want to run commands on a minion after installing a package?
20:54 Ahlee I abuse cmd.run frequently
20:54 ekristen so I take it this is the docs for that http://docs.saltstack.com/ref/modules/all/salt.modules.cmdmod.html#module-salt.modules.cmdmod
20:55 Ahlee That's what I do, not that you should
20:55 ekristen well so I need to install splunk forwarder, then configure it
20:55 ekristen configuring it is a series of commands I’d normally type on the command line
20:56 Ahlee i use cmd.script for something similar
20:56 Ahlee salt '*' cmd.script salt://scripts/runme.sh
20:57 Ahlee increasing file handle/process limit on master did not resolve issue, sadly
20:57 ekristen Ahlee: I’ll take a look, thanks
20:57 dfinn does anyone have a slick way of managing crontabs across multiple servers?  so that you don't have to maintain a sls file for each sever?
20:58 harobed joined #salt
21:01 Ryan_Lane2 joined #salt
21:02 opapo joined #salt
21:04 Ryan_Lane joined #salt
21:05 jcockhren dfinn: rsync?
21:05 jcockhren well... that's not "slick"
21:05 jcockhren but
21:06 dfinn yeah, less than ideal
21:06 dfinn we had puppet recursively grabbing a directory based on hostname or role.  i tried to implement that in salt and ran into a bug :(
21:06 ekristen if I have a cmd.wait on a pkg.installed, will cmd.wait only be run the first time the package is installed and not each time the state is applied?
21:07 jcockhren git backed pillar could be used as well
21:07 ekristen or will it wait for it to check the pkg.installed state and then run
21:07 Corey dfinn: Ugh. "Grab a directory recursively" is almost always the wrong approach.
21:07 dfinn i'm open to better ideas
21:07 dfinn if I use the builtin way for salt to control cron jobs I lose the ability to control comments and set the MAILTO
21:08 forresta dfinn, why aren't the cron jobs just jinja templated files that you fill with variables?
21:08 dfinn because I don't know how to do that ;)
21:08 dfinn got an example?
21:09 Corey dfinn: (What forresta is suggesting is also the "proper" way to do this within Puppet as well, just FYI)
21:09 forresta Corey, yarg
21:09 forresta dfinn, here is one that uses puppet: https://github.com/terminalmage/djangocon2013-sls/blob/master/foo/files/foo-vhost.conf
21:09 forresta *facepalm*
21:09 forresta pillar
21:09 forresta jeez
21:09 forresta https://github.com/terminalmage/djangocon2013-sls/blob/master/foo/vhost.sls
21:09 forresta and it is referenced like that
21:10 dfinn hmmm..that's gonna take a while to wrap my head around
21:10 forresta did you guys do templated files in puppet?
21:10 dfinn no, not really
21:10 forresta oh ok
21:12 JasonSwindle forresta:  pillar.get…… do you use that a lot?
21:12 forresta no, only when referencing pillar doesn't work
21:12 JasonSwindle I just use
21:12 JasonSwindle {{ pillar['project_name'] }}
21:12 JasonSwindle or whatever the key is
21:12 forresta as do I, that was just the quickest example I could find :P
21:13 JasonSwindle Oh
21:13 dfinn i'm struggling to see how I could apply this to servers that each have vastly different cron files.  could be me being dumb…
21:13 forresta dfinn, ok, so how many unique configurations are there?
21:13 forresta is it environment based, or host type based or what?
21:14 JasonSwindle forresta:  I am soon going to test out {{ pillar['logstash']['version'] }}
21:14 dfinn we have 10 cron directories for different hosts or roles, inside those directories are files per user
21:14 dfinn ie root, postgres
21:14 forresta yea that should work JasonSwindle, take a look at: https://github.com/jesusaurus/hpcs-salt-state/blob/master/apt/cache/acng.conf and the other stuff he has, lots of complex examples.
21:15 JasonSwindle So that is a 3 level YAML on the pillar?
21:15 forresta so there are only 10 possibilities of total configurations? Do multiple machines share the same cron job?
21:15 vlcn joined #salt
21:15 forresta JasonSwindle, yep, here is the example pillar: v
21:15 forresta https://github.com/jesusaurus/hpcs-salt-state/blob/master/apt/sample.pillar
21:16 Hefeweizen joined #salt
21:16 dfinn yes, some of the directories are for roles and multiple servers can share the same role
21:16 ekristen how would I make the second cmd.wait watch the first  cmd.wait?
21:16 ekristen https://gist.github.com/ekristen/f6c0cfd5bf5add25aa2b
21:16 Ahlee i'm going to cry. you've been using this for what, an hour? and already you're like twenty times more advanced han i
21:17 Ahlee i give up.
21:17 forresta dfinn, ok, so do I share some cron jobs between the root role, and the postgres role (as an example), or any of the other roles?
21:18 forresta Ahlee, lol, don't give up man.
21:18 forresta ekristen, I don't know if that will work, you might need to break that up.
21:18 JasonSwindle Ahlee:  Don't give up!  It took me some time for Salt to click
21:21 dfinn sorry forresta, the root and postgres aren't roles, they are users.  so you have a directory named cron.role or cron.hostname and then under that you have a crontab for each file who's cron stuff is managed by puppet
21:21 ekristen hrm
21:22 cewood joined #salt
21:22 dfinn like so: http://pastebin.com/BpzGKDCX
21:22 forresta dfinn, oh I see. So do any of the users share cron jobs? Or do you guys just say 'ok deploy these whole crontabs'
21:22 forresta looking
21:22 dfinn multiple servers may use the livedb role
21:22 dfinn we deploy whole crontabs as files
21:23 forresta so is the cron.tribedb/postgres file different than cron.rootdbdr/postgres?
21:23 jacksontj joined #salt
21:23 dfinn yes, drastically
21:23 forresta ok
21:24 forresta ekristen, I commented on that gist, but I'm having a hard time remembering if that will work or not :\
21:25 ekristen question why did you use  cmd: splunkforwarder
21:25 forresta worth a shot though
21:25 ekristen what makes splunkforwarder state a cmd
21:25 forresta because that's the ID of that section, and I can't remember if that will work :P
21:25 forresta I don't have a box in front of me to mess with it on
21:25 ekristen I saw another example where git was used, seemed sorta arbitrary but not at the same time
21:25 ekristen I get the “splunkforwarder” part as the id
21:25 forresta ok
21:26 Ryan_Lane2 joined #salt
21:26 forresta you don't get which part then?
21:26 ekristen but why prefix it with cmd, especially when its installing a pkg
21:26 forresta because both pkg.installed, and cmd.wait are 'part' of spunkforwader
21:26 forresta *forwarder
21:26 dave_den ekristen: require/watch refer to module: ID declaration of the require/watch state
21:26 forresta you have defined that single ID, to be associated with both items
21:26 forresta I just can't remember if you reference it via cmd
21:26 ekristen ah
21:26 ekristen I got it
21:26 dave_den so, cmd: splunkforwarder
21:26 forresta basically, we want to say 'ok that whole section identified by splunkforwarder? Require all of that stuff to run first'
21:27 ekristen well for some reason the first cmd.wait for enable boot-start isn’t working
21:27 ekristen by itself :(
21:27 ekristen so I have to figure that one out first
21:27 andrej Hmmm
21:27 forresta dave_den, does what I wrote look right?
21:27 dave_den forresta:  yeah
21:27 forresta ok good, brain fart over here with the context switching this afternoon it seems
21:28 Ahlee well. pickling the contents of payload didn't work, heh
21:28 forresta dfinn, ok, so how do you say which machine gets which 'sets' of crons?
21:29 andrej I'm trying to have different options in ntp.conf depending on the os grain
21:29 dfinn based on role or hostname
21:29 dfinn the puppet source uses a variable
21:29 dfinn i'll paste
21:29 andrej salt tells me I'm not allowed mapping values
21:29 dfinn forresta, http://pastebin.com/jT4naAJU
21:29 forresta well, stop breaking it then andrej  :P
21:30 Gifflen joined #salt
21:30 dfinn so if it matches one of those, it will get any files in that directory
21:30 dfinn honestly it seems to work pretty well and is fairly simple
21:30 dfinn unfortunately i'm running into https://github.com/saltstack/salt/issues/9304 when I try to duplicate in salt
21:30 ekristen arg
21:30 ekristen that enable boot start won’t run :/
21:31 dangra hello, any plans to cut a new release from develop branch ? I'm looking for the dockerio improvements on develop that are not available in master branch.
21:31 forresta dangra, those may be in the next RC
21:31 JasonSwindle You beat me forresta :P
21:31 forresta you could always manually modify the files for the time being if you need that support immediately
21:32 forresta this is why they pay me those big internet dollars
21:32 forresta my bits and bytes
21:32 dfinn 1s and 0s are expensive
21:32 JasonSwindle In DogeCoins
21:32 forresta ugh no dogecoins
21:32 forresta so much spam
21:32 dangra i was thinking on using a develop snapshot, or just copying the bits to _modules and _states
21:32 forresta dangra, yea that should work
21:32 dangra thanks
21:33 forresta dfinn, so are there only ~5 crons per 'setup'?
21:33 dfinn usually only 1 or 2 per dir
21:33 dave_den ekristen: what does the output show?
21:33 ekristen success
21:33 forresta ekristen, after you fixed your script?
21:34 forresta dfinn, ok so here is what I'd do for the time being. Create several states (one for each of these users), that creates the directory, and drops in the associated cron
21:35 forresta then for specific roles that you've configured via your top or whatever, you can just include the necessary items
21:35 forresta you can also templatize the cron jobs if you need, in the event those have variables.
21:35 dfinn that could work.  definitely more to manage
21:35 forresta It is
21:35 forresta but it's an option till that issue is resolved.
21:35 dave_den ekristen: do you mean the splunkforwarder cmd.wait is not running?
21:36 ekristen cmd, result: true
21:36 ekristen ack, I think I might have found the problem
21:37 ekristen dave_den: I think there was a latent artifact triggering the unless
21:37 ekristen ok, that worked ;)
21:37 dfinn actually forresta, that won't totally work, it's even more complicated.  for every role, i would need a sls file for each user
21:37 troyready joined #salt
21:37 forresta dfinn, why?
21:37 dave_den cool
21:38 dfinn you said for each user but the user cron files are different for each role
21:38 forresta which is why you templatize them
21:38 dfinn even if they have nothing in common?
21:39 forresta I guess I'm confused
21:39 forresta so is the cron/rootdbdr/interch always the same?
21:39 dfinn always the same as?
21:39 forresta there is only 1 md5sum for that file regardless of the environment
21:39 ekristen got it to run
21:39 forresta ekristen, did you just fix your unless?
21:40 ekristen forresta: um, for the first two yes, still working on the deploy poll part
21:40 dfinn not sure what's up with the md5 but they are different.  i just checked rootdb/interch vs manager/interch, not similar
21:41 forresta no no
21:41 jtrucks joined #salt
21:41 forresta I mean rootdb/interch
21:41 dfinn so i would need a rootdb.interch.sls and a manager.interch.sls and so for basically every cron file
21:41 dfinn i'm not following
21:41 forresta no
21:41 forresta hang on let me whip something up real quick
21:41 dfinn appreciate it
21:41 forresta yea np
21:42 JasonSwindle Go go forresta
21:42 forresta I told you, this is why I get the big bucks, jeez!
21:43 JasonSwindle +10 forrestaCoin
21:43 forresta that and acting like rubber duck debug man
21:47 dfinn forresta, do your kids race bmx and mountain bikes?
21:47 jtrucks joined #salt
21:47 forresta I don't have kids
21:47 dfinn well that answers that
21:48 forresta https://gist.github.com/gravyboat/8307503
21:48 forresta check that out dfinn, see if it makes more sense to you
21:49 dfinn hmm…
21:50 dfinn except it would be copying those files to /var/spool/cron/blah?
21:50 forresta on the salt master you mean?
21:50 dfinn no, on each minion
21:51 dfinn putting files into a /cron.role directory on each minion doesn't accomplish much
21:51 dfinn maybe that is just a quick example?
21:51 forresta right that was just my example
21:51 forresta from the location you showed in your paste since I don't know the rest of the path
21:51 dfinn cool, I think I see how to make that work.  so 1 file per role
21:51 forresta right
21:51 forresta and like I said, it's not the best way to do it
21:51 forresta and it is a bit of work
21:51 forresta so it doesn't mirror what you have exactly
21:52 dfinn i think this will have to do until that bug is fixed
21:52 dfinn thanks!
21:52 forresta yea np!
21:52 dfinn really appreciate you taking the time
21:52 forresta yea no worries
21:52 forresta sometimes an example is worth more than any amount of takling
21:52 forresta *talking
21:52 troyready joined #salt
21:52 dfinn usually, for my slow brain
21:52 forresta :P I also like examples, so I understand
21:53 jtrucks joined #salt
21:53 forresta depending on how you lay it out dfinn, you could also use variables, and pillar to make this really cool
21:53 MZAWeb joined #salt
21:53 dfinn i'll mess around with that
21:53 forresta so you could create a section in your cron pillar that has the paths and names
21:53 forresta but that would take much longer to write an exampel for
21:53 forresta *example
21:54 dfinn oh, maybe I could reference the role in the source
21:54 forresta yea, there are a lot of ways you could do it that are more elegant than this
21:57 JasonSwindle forresta:  You could even put the last two in https://gist.github.com/gravyboat/8307503 in a JINJA loop
21:57 forresta JasonSwindle, yea based off the pillar values
21:57 forresta if you created a tree for the crons
21:57 dfinn how about this forresta?  http://pastebin.com/1UcPfuhk
21:58 dfinn that should pick up the dir with a hostname or role name
21:58 dfinn i think
21:58 ekristen forresta: this is my resulting file that works
21:58 ekristen https://gist.github.com/ekristen/f6c0cfd5bf5add25aa2b
21:59 jtrucks joined #salt
21:59 forresta dfinn, I don't believe that would work since you are trying to use multiple files in a single file.managed.
21:59 dfinn no?  we have been doing something very similar in just about all of our other sls files
22:00 forresta dfinn, if it works, that's something new that I didn't know!
22:00 dfinn i'll try it out
22:00 forresta ok cool
22:00 forresta ekristen, you know you don't need to use so many requires right?
22:00 ekristen nope
22:00 forresta since splunk_enable_boot requires the package, splunk_deploy_poll just needs to require the splunk_enable_boot cmd, and then spunk_start, can just require the splunk_deploy_poll cmd
22:01 forresta because you're chaining them
22:01 ekristen yeah I just realized that
22:01 ekristen I was copying and pasting
22:01 forresta ok cool :P
22:01 vejdmn joined #salt
22:01 forresta ahh alright
22:01 forresta makes sense
22:01 ekristen updated
22:01 ekristen well that seems to work for me
22:01 forresta great
22:01 ekristen now to make it more OS agnostic, but I don’t need that right now
22:01 ekristen lol
22:01 forresta hah
22:02 toastedpenguin joined #salt
22:02 forresta dangra, how'd that work for you?
22:03 ekristen is there a way to only run a command once
22:03 ekristen and only once?
22:03 ekristen I guess by using the unless huh
22:04 dangra forresta: it works. now I am stuck trying to get port bindings working with docker.running
22:04 forresta ekristen, yep that is what most people do
22:04 forresta dangra, you're on your own with that one, my knowledge of the docker module/state is almost 0 :P
22:04 dangra no prob :)
22:04 dimeshake joined #salt
22:05 jtrucks joined #salt
22:06 techdragon joined #salt
22:07 dangra I'm trying to bootstrap shipyard using salt like this: https://gist.github.com/dangra/8307798
22:08 dangra my salt-fu is a bit basic at the moment, any obvious improvement on the states definition will help a lot
22:09 forresta what error are you getting?
22:09 Corey WHoa. CentOS is now part of Salt.
22:09 forresta Corey?
22:09 Corey Er, RedHat.
22:09 Corey Wow.
22:09 Corey Wrong phrasing.
22:09 dfinn other way around you mean?
22:09 Corey Words, how do they work...
22:09 Corey Let me rephrase.
22:09 forresta what do you mean salt is part of redhat?
22:09 Corey CentOS is now part of RedHat.
22:09 Corey There we go.
22:09 forresta oh
22:09 dfinn whoa
22:09 dimeshake wat
22:09 dfinn link?
22:09 dimeshake got a link?
22:09 Corey Yeah, that got confused with /j #salt.
22:09 dangra forresta: https://gist.github.com/dangra/8307836
22:10 Corey http://lists.centos.org/pipermail/centos-announce/2014-January/020100.html
22:10 forresta centos has been struggling for a while
22:10 dimeshake holy shit, that's huge
22:10 Corey dfinn: What, you're not just going to take my word for it?
22:10 dfinn ;)
22:10 forresta This is a very good thing
22:10 forresta I thought centos wasn't gonna make a major release to 6 a while back
22:11 dangra me too, it's great news for CentOS.. I thought it was dead
22:11 jtrucks joined #salt
22:11 JasonSwindle Long live CentOS?
22:11 dimeshake Long live the King
22:12 dfinn forresta, that seems to sort of be working.  1 cron file is working however the rest are giving errors but the rest are identical to the one that is working.  *sigh*.  error and sls:
22:12 dfinn http://pastebin.com/jSbMJDSM
22:12 forresta dangra, hmm, maybe something with the name?
22:13 dangra no, it's the port_bindings value
22:13 forresta dfinn, ahh bummer
22:13 dfinn notice how /var/spool/cron/postgres doesn't give any errors but it's the same as all the others.  any insight into that error?
22:14 forresta dfinn, can you try ensuring the postgres user/group exist? You probably want to do that in a unique state.
22:14 dfinn yeah
22:15 dangra forresta: it works now by changing to the example value format: https://gist.github.com/dangra/8307798/revisions
22:15 forresta ahh makes sense
22:15 * dfinn hates python errors
22:16 dangra well, python errors are pretty clear for me :)
22:16 dangra pyyaml parser sucks a bit
22:16 dangra but the above error was valid, it was me
22:16 dfinn just figured it out
22:17 dfinn it is failing because a directory does not exist that matches the source
22:17 dfinn how do I tell it to just move along if nothing matches?
22:17 jtrucks joined #salt
22:17 forresta I'm not sure, I haven't written a file.managed like what you're doing, I would think it just goes on by itself.
22:18 dfinn that's what I assumed to but it definitely doesn't
22:19 aj0 joined #salt
22:20 seanz joined #salt
22:21 dfinn i knew that the multiple sources would work like this because we are doing it everywhere else however we are always doing it with a config file type setting.  so you run through your optional sources and then at the bottom you have a default config file that it will grab if nothing else matches
22:21 dfinn can't really do that here though
22:21 sschwartz joined #salt
22:22 dfinn i wonder if this is worthy of filing a bug
22:22 sschwartz An odd question; The name of a service in a salt script should be the same as the /etc/init.d script used to control it, correct? I keep getting "The named service <x> is not available" even though the startup script copies over in the previous step.
22:23 jtrucks joined #salt
22:23 forresta sschwartz, usually that is the case, what's the service?
22:24 Ryan_Lane2 joined #salt
22:24 jgiraldo joined #salt
22:24 sschwartz tomcat, starting as tomcat7. I have a /etc/init.d/tomcat7 in place, and invoke it as tomcat7:   service:     - running
22:24 troyready joined #salt
22:25 forresta so if you go on the box
22:25 forresta do you start it with service tomcat7 start?
22:25 dfinn where would one go to submit a bug?
22:26 sschwartz Yes; and it just did start when I retested it.
22:26 forresta dfinn, http://lists.centos.org/pipermail/centos-announce/2014-January/020100.html
22:26 forresta ugh
22:26 forresta dfinn, https://github.com/saltstack/salt/issues
22:26 forresta there we go
22:26 forresta jeez
22:27 forresta sschwartz, I'd run the state back through with -l debug, see if it gives anything extra comes out of the logs
22:28 sschwartz Will do -- thanks!
22:28 forresta unless you're installing, and starting the service in the same section, and not requiring the package, that might be an issue.
22:28 kermit joined #salt
22:28 sschwartz The package is copied over from a filetree; customized elements inside it.  But I'll look into that as well.
22:29 jgiraldo when I run a test.ping globbing by an arbitrary grain, I receive a different number of servers replying.  I even use -b to run smaller batches but no go, also upped the ulimit on the salt master.  Does the fact that test.ping doesnt kick back a verification from all servers mean that for all future state propagations will arbitrarily not go through?
22:29 jtrucks joined #salt
22:30 andrej What did I do wrong in my init.sls? The first paste is the init.sls, the 2nd is the error http://pastebin.com/p2bP5sp1 & http://pastebin.com/yEZc4dbp
22:31 dangra what is the best api to talk with salt master from a xmpp bot (in python)?
22:32 forresta jgiraldo, re-run with -t 10
22:32 forresta to modify the timeout value
22:32 dangra is there an http/REST api or something else like running a bot as a plugin
22:32 forresta dangra, some guys have been working on bots, but I don't know what the status looks like
22:32 timoguin dangra, i've seen an xmpp gateway to the salt-api. lemme find it. one sec.
22:32 forresta andrej, your indenting on - source, and the if statement is messed up
22:32 dangra implementing the bot looks very simple using a lib like sleekxmpp
22:33 dangra timoguin: thanks
22:33 timoguin http://hveem.no/salt-xmpp-gateway
22:33 timoguin there ya go
22:33 dangra nice
22:33 timoguin that guy's put out some cool stuff
22:33 forresta andrej, you also probably want to require the ntp pkg, not watch it
22:33 forresta err nevermind I see what you are doing
22:34 dave_den andrej: http://pastebin.com/4bWkiSA3
22:34 forresta psssh, your indenting is lame dave_den
22:35 forresta why is the if statement so far out :(
22:35 dave_den i strive for lame
22:35 forresta should be ok one level in I would think
22:35 jtrucks joined #salt
22:35 JasonSwindle I think this should work...
22:35 JasonSwindle https://gist.github.com/JasonSwindle/c3f1ce852806a95d05f6
22:36 forresta why is source indented?
22:36 forresta WHY YOU INDENT SOURCE
22:36 forresta :P
22:36 JasonSwindle :P
22:36 jcockhren indent all the things
22:36 forresta that's what is breaking his original example
22:36 forresta source isn't a sub item of mode is it?
22:37 JasonSwindle I fix, I fix
22:37 forresta ok
22:37 JasonSwindle update
22:37 forresta ok I feel better now
22:37 JasonSwindle and I moved the state IDs onto one line
22:37 jgiraldo trying that now
22:37 JasonSwindle That feels "better" to me.
22:37 forresta yea, it's funny to see what everyone likes more, such as I don't like pkg.installed
22:38 forresta I like pkg:\n - installed
22:38 JasonSwindle I like my SLSs compact and clean
22:38 JasonSwindle that just feels, messy
22:38 dave_den mod.function: ftw
22:39 pniederw anyone who can help with installing latest salt-master via `pkg.latest`? can't get it to work, always says "already up-to-date".
22:39 forresta pniederw, what version and OS are you on?
22:40 jgiraldo I had already set the default timeout set to 60 seconds on the salt master, as such, would -t 10 help at all?
22:40 forresta jgiraldo, ahh no it probably wouldn't,,,
22:40 pniederw 0.17.2 (the latest patch), ubuntu 13.10
22:40 jgiraldo Soo.....yeah?
22:40 forresta 0.17.4 is the latest stable, do you not see the updated release when you manually query with apt-cache?
22:40 forresta jgiraldo, how many minions?
22:41 jgiraldo 1300
22:41 jgiraldo I've run them in batches of 20 or 30
22:41 jgiraldo still discrencies
22:41 forresta and even then you see issues?
22:41 jgiraldo only about 300 reply at a time
22:41 jgiraldo the other 1000 do not reply to test.ping
22:41 jtrucks joined #salt
22:42 dave_den are you sure you upped your nofile limit?
22:42 jgiraldo yes
22:42 dave_den i can almost guarantee it's a resource issue on the master
22:42 forresta that reminds me dave_den
22:42 forresta make some contributions to http://docs.saltstack.com/topics/troubleshooting/master.html
22:42 jgiraldo 12 cores, 16 G ram
22:43 jgiraldo htop shows resources are not topping out when i run test.ping
22:43 forresta dave_den, it makes no sense that he'd see test.ping issues when running a group of 20-30 boxes
22:43 pniederw `apt-cache madison salt-master` only shows 0.16.0-1. that's why I was asking if `pkgrepo.managed` should affect manual apt commands.
22:44 forresta oh I thought you were asking about installing the latest.
22:44 dave_den it would if he's already pressing limits and the connections to ret ports are not working for the minions
22:44 forresta dave_den, good point
22:44 pniederw I am. I'm on 0.17.2, and want to install 0.17.4.
22:44 funzo joined #salt
22:46 forresta pniederw, I guess I am confused then
22:46 forresta you yourself are saying it isn't available from the present repos right?
22:46 andrej @forresta, odd - I copied & pasted the formatting from the states tutorial
22:46 ekristen whats the best way to split up states for multiple environments?
22:46 andrej I'll play with indentation
22:46 forresta andrej, can you link me the tutorial?
22:46 ekristen actually better question
22:46 cdcalef joined #salt
22:47 forresta ekristen, templatize if you can, and store data in pillar
22:47 ekristen fair enough
22:47 ds22 joined #salt
22:47 jtrucks joined #salt
22:47 ekristen guess I need to figure out how an env is applied to a minion
22:47 dave_den jgiraldo: cat /proc/$(</var/run/salt-master.pid)/{status,limits}
22:48 dave_den i gotta leave in a min
22:48 forresta what a slacker
22:48 dave_den i know right
22:48 jgiraldo Thank you guys, looks like the issues are arising because this is NOT a dedicated box.
22:48 forresta dave_den, if you have some time, your help to bolster the master troubleshooting guide would help
22:48 jgiraldo I will attempt to move it to a dedicated box and  try again
22:49 pniederw I have states `pkgrepo.managed(ppa="saltstack/salt") and `pkg.latest(name="salt-master", refresh=True). running th says "already up-to-date". `dpkg -s salt-master` says 0.17.2-2raring1. website for the ppa says 0.17.4 is latest. something doesn't add up here.
22:49 dave_den forresta: yah, i'll try to add some tips
22:49 forresta cool
22:49 forresta the more the better
22:49 dave_den jgiraldo: cool
22:49 forresta I don't wanna ever deal with a master issue in the chat
22:49 forresta performance wise that is
22:49 dfinn joined #salt
22:49 dave_den back tomorrow. adios
22:49 forresta pniederw, oh so your state adds the repo?
22:50 pniederw yes
22:50 pniederw automation, you know :-)
22:50 forresta so if you go on the box now, does that repo exist?
22:51 forresta and if so, can you refresh the apt data, then see what version an apt-cache check pulls down?
22:51 forresta maybe the repodata just isn't getting refreshed
22:51 pniederw where do I have to look? `apt-get update` doesn't list it
22:52 forresta isn't it in /etc/apt/sources.list or something?
22:52 aj0 joined #salt
22:53 pniederw yeah, something like that. I was thinking that maybe pkgrepo.managed adds the repo in such a way that only the pkg state sees it
22:53 forresta oh no
22:53 forresta it's supposed to add the repo to the machine
22:53 forresta I'm trying to remember if there were any bugs in that release.
22:53 jtrucks joined #salt
22:53 Gifflen joined #salt
22:53 pniederw ok, that's good to know.
22:55 cdcalef hmm
22:55 andrej forresta - I just tried the version of init.sls you pasted
22:55 cdcalef anybody use salt-cloud with rackspace?  salt-cloud keeps creating the VMs in DFW instead of ORD like i specify.
22:55 andrej Still etting the same error at the same location; it swears at the colon after source
22:56 andrej and here's the link http://docs.saltstack.com/topics/tutorials/states_pt3.html
22:56 pniederw maybe pkgrepo.managed isn't configured correctly, but in that case I'd expect an error somewhere.
22:59 jtrucks joined #salt
22:59 xmltok joined #salt
23:01 Ymage_ joined #salt
23:02 terminalmage joined #salt
23:02 andrej meh
23:02 jtrucks joined #salt
23:03 andrej played some more with the formatting of my inital paste, and that is now good
23:03 andrej thanks forresta
23:04 Ryan_Lane2 joined #salt
23:04 bhosmer joined #salt
23:05 forresta andrej, I think that was JasonSwindle that pasted that one in
23:05 forresta I just said something about the indentation
23:07 MZAWeb joined #salt
23:08 ConceitedCode joined #salt
23:08 jtrucks joined #salt
23:10 Ryan_Lane2 joined #salt
23:10 KyleG cedwards: You around?
23:11 forresta Ryan_Lane2, do you use wiline as your ISP?
23:11 Ryan_Lane2 I have no idea which ISP i'm on right now
23:11 Ryan_Lane2 I'm at work
23:11 forresta oh ok
23:11 forresta I was looking at this page: http://www.wiline.com/residential/internet.html
23:11 forresta and was like... 15 megs?
23:12 forresta so that makes more seonse
23:12 jfzhu_us joined #salt
23:14 jtrucks joined #salt
23:15 pniederw looks like `pkgrepo.managed(name="salt-master", ppa="saltstack/salt")` is a no-op, despite succeeding with "Configured package repo salt-master". salt-call with debug logging doesn't reveal anything. `pkgrepo.managed` and `pkg.latest` execute in correct order and both succeed.
23:15 Gifflen joined #salt
23:17 forresta are you trying to use that as a module?
23:18 pniederw I'm just transcribing the yaml syntax here, if that's what you mean
23:18 forresta no I mean is that part of a state
23:19 forresta It looks like the example on this page right? http://docs.saltstack.com/ref/states/all/salt.states.pkgrepo.html
23:20 pniederw yes
23:21 jtrucks joined #salt
23:22 pniederw btw, is salt supposed to be capable of installing a new version of itself in this way?
23:23 pniederw I initially pushed these states with salt-ssh, now trying to update for the first time
23:23 forresta I've seen people updating minions at the very least with salt.
23:24 forresta you should be able to do it though, the issue is just pulling the repo, or configuring it in some way
23:24 funzo joined #salt
23:24 pniederw I'm experiencing the same problem for my salt-minion states
23:25 Ahlee this is messed up
23:25 Ahlee push an intentionally broken branch up to git, run the state, fails. delete branch, appears to clear up
23:26 Ryan_Lane2 joined #salt
23:27 Ahlee nop
23:27 Ahlee e
23:29 dw_ left #salt
23:35 fr0de joined #salt
23:37 misc1 joined #salt
23:38 fr0de Hi all, just wondering if salt has the concept of 'dynamic environments', similar to how puppet does: create a new branch in your git repo, and a new environment appears (courtesy of a small script) on your master, ready for minions to use? Is this even needed in salt? How do we ensure changes to modules or config don't get applied in production before they are ready? Is there a FAQ on this that I just haven't spotted? :)
23:40 taion809 joined #salt
23:42 Ahlee fr0de: if you use gitfs backends, no script required
23:43 Ahlee i.e. I just pushed branch hotfix_bug-002 to my branch, and I can execute against that with salt '*' state.highstate env=hotfix_bug-002
23:44 Ahlee as for not applying to production before they're ready, default environment of base maps to master
23:44 Ahlee otherwise, ifyou go multi environment one master, make sure the environments don't cross
23:44 fr0de Brilliant!!!
23:45 Ryan_Lane2 joined #salt
23:45 fr0de I'm definitely "falling for" salt, and that's after doing puppet for the past three or so years :)
23:46 taion809 joined #salt
23:46 fr0de I'm a python guy, so that's a part of it, too :) python (imho) is nicer than ruby :P
23:46 alunduil joined #salt
23:46 fr0de Thanks, Ahlee !
23:47 Ahlee I wish every day salt was written in ruby, heh
23:47 Ahlee different strokes for different folks
23:48 misc1 anyone have experience running salt in a more traditional enterprise environments? (i.e. some linux/solaris but far more windows servers (1200+), vmware esxi (not openstack), both physical and virtual servers, no cloud, etc...).  Are there any blogs/case studies that speak this type of deployment scenario?
23:49 Ahlee misc1: that's my environment, minus the solaris
23:50 misc1 great, are you able to use salt to perform a high percentage of configuration management?
23:51 Ahlee Technically? yes.  Culturally? no.
23:51 misc1 haha, that was going to be my next question
23:51 Ahlee though most of our windows boxes are just citrix front ends
23:52 Ahlee citrix gui apps to connect to the software doing the work on linux boxes
23:52 misc1 ah, that probably simplifies things
23:52 Ahlee so not much configuration required on the windows systems
23:52 Ahlee all management of citrix is manual
23:52 misc1 yes, we are a heavy citrix shop as well
23:52 Ahlee and what's required on the windows boxes makes more snese to leave in group policy
23:52 misc1 that makes sense
23:53 Ahlee for the most part the biggest issue coming at salt from an enterprise level is most examples are written for installing a web server or something
23:54 misc1 exactly, that's been my finding as well
23:55 misc1 before salt, did you use another config mgt system?  if so, was it script based or gui based?
23:55 srage joined #salt
23:56 Ahlee prior to salt we used bash scripts and just rebuilt the linux boxes
23:57 misc1 in our case there is nothing in place and most of the ops team aren't scripters
23:58 fr0de misc1: convert them, by automating a couple of pain points :)
23:59 fr0de misc1: I also find it scary that an ops team doesn't have (many) scripters.. O_o

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