Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-07-02

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

All times shown according to UTC.

Time Nick Message
00:00 ipalreadytaken joined #salt
00:01 ipalreadytaken joined #salt
00:01 vu_ joined #salt
00:04 fllr joined #salt
00:12 forrest carmony, you around?
00:16 beneggett joined #salt
00:17 Hell_Fire joined #salt
00:24 Ryan_Lane joined #salt
00:34 TaiSHi Failed to authenticate!  This is most likely because this user is not permitted to execute commands, but there is a small possibility that a disk error ocurred (check disk/inode usage). <- dafuq ?
00:42 oz_akan_ joined #salt
00:43 HACKING-TWITTER joined #salt
00:44 Ryan_Lane joined #salt
00:51 bhosmer joined #salt
00:56 hopthrisC maybe it is too late, but…
00:56 hopthrisC "Specified state blockdev.formatted is unavailable."
00:56 hopthrisC why?
00:57 hopthrisC # salt --version
00:57 hopthrisC salt 2014.1.5
00:58 aquinas joined #salt
00:59 bhosmer joined #salt
01:05 krow joined #salt
01:06 jnials joined #salt
01:10 shaggy_surfer joined #salt
01:15 svx joined #salt
01:17 oz_akan_ joined #salt
01:29 krow joined #salt
01:29 jhauser_ joined #salt
01:35 mateoconfeugo joined #salt
01:37 joehh hopthrisC: that state is only available in develop
01:37 joehh the docs are unfortunately generated from the develop branch and contain states and modules that are not in released versions
01:38 hopthrisC joehh: oh, i see
01:38 TaiSHi joehh: just curious but what's the state on dev debs :P ?
01:38 hopthrisC thank you!
01:39 bhosmer joined #salt
01:39 joehh http://salt.readthedocs.org/en/v2014.1.4/ref/states/all/index.html
01:39 joehh contains the relevant focs
01:39 hopthrisC great!
01:39 analogbyte joined #salt
01:39 joehh slow - aiming for finishing this weekend
01:39 joehh TaiSHi: ^
01:39 TaiSHi Love you
01:40 TaiSHi No physical demonstrations today
01:40 TaiSHi I thought you might appreciate that
01:40 joehh :)
01:50 Leech joined #salt
01:51 TheThing joined #salt
01:55 hopthrisC ::q
01:56 otter768 joined #salt
01:57 alekibango joined #salt
02:06 sxar_ joined #salt
02:08 bhosmer joined #salt
02:15 ajw0100 joined #salt
02:21 Ryan_Lane joined #salt
02:22 darrend joined #salt
02:23 TyrfingMjolnir joined #salt
02:27 bchung joined #salt
02:29 bchung Did anyone hit this before? [DEBUG   ] Executing command(PID 17035): 'winexe -U Administrator%<passwd> //<ec2-public-ip> "hostname"' Error: There was a profile error: Failed to authenticate against remote windows host
02:30 bchung I'm using salt-cloud to launch an ec2 windows instance (already disabled the firewalls, and set port 445 to be opened in security group)
02:31 bchung manually run winexe I got: winexe -U Administrator%<passwd> //<ec2-public-ip> "hostname" ERROR: Cannot open control pipe - NT_STATUS_INVALID_PARAMETER
02:32 TyrfingMjolnir joined #salt
02:33 active8 joined #salt
02:33 bchung does anyone know how to resolve this? it's windows 2012 R2 instance.
02:41 azylman joined #salt
02:48 TyrfingMjolnir joined #salt
02:48 krow joined #salt
02:52 Spark joined #salt
03:00 q4brk_ joined #salt
03:02 justlooks joined #salt
03:04 justlooks why salt cache old sls file on minion ? i checked /var/cache/salt/minion/files/base/xxx ,find minion still get old version sls file after i run the updated sls file
03:04 dvogt joined #salt
03:06 TyrfingMjolnir joined #salt
03:09 yomilk joined #salt
03:11 ramishra joined #salt
03:14 schimmy joined #salt
03:17 |rt| joined #salt
03:18 schimmy1 joined #salt
03:19 Spark joined #salt
03:31 bhosmer joined #salt
03:32 azylman_ joined #salt
03:40 smcquay_ joined #salt
03:43 terminalmage justlooks: the fileserver is updated roughly every 60 seconds
03:43 acabrera joined #salt
03:43 terminalmage at least for remote fileserver backends
03:44 terminalmage was the file on a gitfs source?
03:45 beneggett joined #salt
03:48 justlooks terminalmage: i tested on my product env no such problem ,it appear on my test env ,i install minion on  kvm guest  ,and after i run the script  i need undo all change quickly i use kvm snapshot to reset the minion env ,i do not know if this is the cause
03:51 terminalmage justlooks: have you tried reproducing the issue without resetting the minion environment?
03:51 terminalmage without restoring the kvm snapshot?
03:51 gzcwnk joined #salt
03:52 gzcwnk hi, using a sls file how would I tell a minion to do a yum -y update?
03:52 bhosmer joined #salt
03:52 gzcwnk i cant seem to find anything taht allows a complete patching
03:53 gzcwnk of the os
03:54 mateoconfeugo joined #salt
03:54 terminalmage gzcwnk: have you looked at the docs?
03:54 gzcwnk i have 10+ pages open, i cant see how to do it
03:55 terminalmage http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.yumpkg.html#salt.modules.yumpkg.upgrade
03:55 mosen joined #salt
03:55 gzcwnk I can do it with a cmd.run or yes that
03:55 gzcwnk but that's cli
03:55 vexati0n is it possible to run a multi-master setup where different masters run their services on different tcp ports?
03:55 manfred 6tthat isn't for a .sls
03:56 gzcwnk I want to put it in a sls file so when I run high.satte it does it
03:56 manfred gzcwnk: you could use that one terminalmage linked to in a module.run state
03:56 terminalmage gzcwnk: module.run
03:56 gzcwnk ta
03:56 terminalmage gzcwnk: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.module.html#salt.states.module.run
03:57 gzcwnk seems a crude way to do things
03:58 manfred there isnt' a state to just update every piece of software on a minion
03:58 manfred that is just in the modules
03:58 manfred ¯\(°_o)/¯
03:59 terminalmage if we added a state, then we'd need a way to determine success/failure
04:00 terminalmage if success == no packages updated, then that can also happen if the yum/apt/whatever command fails
04:00 yomilk joined #salt
04:00 gzcwnk lol, ok
04:03 gzcwnk - cwd: is the tag for a command to run?
04:03 gzcwnk so -cwd: yum -u update  ?
04:03 gzcwnk so -cwd: yum -y update  ?
04:05 terminalmage um, no
04:05 terminalmage where are you getting that?
04:06 terminalmage cwd is the cwd
04:06 terminalmage maybe you're confusing cwd with cmd in some way? hard to tell, since I don't know where you git that
04:06 terminalmage s/git/got/
04:07 gzcwnk off the page you psoted
04:07 gzcwnk posted
04:07 gzcwnk its  not clear how I can use moduel.run
04:07 gzcwnk no examples
04:08 terminalmage http://docs.saltstack.com/en/latest/ref/states/all/salt.states.module.html#salt.states.module.run ?
04:08 terminalmage that has no cwd argument
04:10 gzcwnk yeah there is but i have no idea what it means  - cwd: /path/to/my/repo
04:10 terminalmage dude, there totally is not
04:11 terminalmage not at that link
04:11 gzcwnk well i can see that on the page
04:11 terminalmage are you referring to the example above?
04:11 gzcwnk yes
04:11 gzcwnk like i said it makes no sense
04:11 gzcwnk it isnt explained
04:11 terminalmage ok. so, any arguments you pass to module.run are passed through to the module function you're running
04:11 terminalmage no, it's pretty well explained
04:12 terminalmage the module.run state runs a module function
04:12 gzcwnk maybe if you are not a beginner like me
04:12 terminalmage "All arguments that the module state does not consume are passed through to the execution module function being executed:"
04:12 terminalmage is there something about this which is confusing?
04:12 gzcwnk yes, i see no practical explanation of whatits saying
04:12 terminalmage in the git.fetch example below that, it's passing cwd to git.fetch
04:13 terminalmage http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.git.html#salt.modules.git.fetch
04:13 gzcwnk ah well not knowing what git.fetch is
04:13 terminalmage it's the module function being run
04:14 terminalmage do you understand the difference between a module function and a state? perhaps that is where the confusion is coming from
04:14 gzcwnk no i probably dont
04:14 terminalmage ok, so a module function is a CLI function that does stuff
04:14 gzcwnk all i want to do is patch the damn box, lol
04:14 terminalmage looks up information
04:15 terminalmage makes changes
04:15 terminalmage etc
04:15 gzcwnk I mean yum -y update is pretty easy to do
04:15 terminalmage a state leverages these functions to achieve a desired state
04:15 gzcwnk except it means I have to write a script to do somethings and tehn call salt
04:15 terminalmage the module.run state allows you to, in an SLS file, execute a module function
04:16 terminalmage gzcwnk: http://dpaste.com/0JM1HY6
04:16 mFacenet joined #salt
04:16 mFacenet left #salt
04:16 terminalmage those are functionally identical
04:16 gzcwnk Im mystified why its bnot simple to make teh state of the machine the "latest"
04:17 terminalmage gzcwnk: as I already said, for states we need to determine success/failure
04:17 terminalmage what do you consider success?
04:17 terminalmage no packages installed?
04:17 terminalmage ok, what if the command fails and no packages are installed?
04:17 spidermo joined #salt
04:18 terminalmage there is probably a way to do this, but it's not really as straightforward as you believe it to be
04:18 gzcwnk k
04:18 terminalmage the examples I gave you in the pastebin above should work though
04:18 terminalmage I used jinja to ensure that the state is not executed on a system which doesn't have yum
04:19 gzcwnk thnaks i will go dissassemble this to undersatnd it
04:19 terminalmage the {% if ..... %}
04:19 terminalmage that is optional, if you're in an all RHEL/CentOS/whatever env then you obviously don't need it
04:19 gzcwnk well yes right now but I want to be OS agnostic
04:20 terminalmage ok, well then you can do an elif
04:20 gzcwnk I dont want to be tied to redhat
04:20 terminalmage in jinja
04:20 terminalmage and add examples for apt, etc.
04:20 gzcwnk ok thank
04:20 gzcwnk s
04:22 terminalmage until a method can be determined which allows us to determine success/failure, that is unfortunately the best that can be done. I'll try to think of the best way to do it
04:22 terminalmage I have a bit of an idea
04:22 terminalmage but I'm not sure if it will work
04:22 terminalmage will require testing
04:24 pjs can I use salt-bootstrap to update salt to a newer version as well?
04:24 manfred pjs: i wouldn't
04:25 manfred pjs: i would just use a salt state to do it
04:25 manfred pjs: https://github.com/gtmanfred/salt-states/blob/master/salt-minion/init.sls
04:25 dvogt joined #salt
04:26 pjs manfred interesting.. thank you
04:27 manfred pjs: you already have it hooked up to salt :P
04:27 pjs manfred what's the saltutil.sync_all?
04:27 manfred that just syncs all grains and pillars and stuff to make sure that they are the newest versions?
04:28 pjs ahh ok.. so after it's installed it will relaunch the minion?
04:28 pjs so the current version is running?
04:28 pjs (sorry, I'm prety new to salt)
04:28 ipalreadytaken joined #salt
04:29 manfred it mostly is used to sync your pillars and grains and then also sync stuff like /srv/salt/_modules /srv/salt/_states and stuff like that
04:29 whytewolf joined #salt
04:29 terminalmage gzcwnk: actually, it looks like you can just do a pkg.upgrade
04:29 terminalmage should be platform agnostic
04:30 terminalmage no need for complicated jinja if/else/endig
04:30 terminalmage endif
04:30 terminalmage so that jinja can be removed
04:30 terminalmage gzcwnk: do you get why those two are functionally identical?
04:32 terminalmage think of the top-level part as a unique identifier for that state. if a "name" parameter isn't passed to the state, then it pulls the value from that top-level declaration
04:32 pjs manfred thanks man
04:32 manfred np :)
04:32 terminalmage and uses that as the "name" param
04:32 Luke_ joined #salt
04:33 terminalmage pjs: yeah the saltutil.sync_* functions sync custom salt extensions
04:33 terminalmage if you're not using any of them, then it wouldn't apply to what you are doing
04:33 schimmy joined #salt
04:34 pjs manfred stupid question, but where are you including salt-minion? Or is that just a state you call as needed?
04:34 manfred pjs: yeah, i just call it whenever
04:34 manfred pjs: usually when I am testing things
04:34 terminalmage pjs: http://docs.saltstack.com/en/latest/topics/targeting/grains.html#writing-grains might help clear this up a little
04:34 pjs manfred awesome!
04:35 manfred pjs: something like salt \* state.sls salt-minion pillar='{}' and then set the pillar that sets the branch and repo to pull from
04:35 pjs terminalmage nice.. thank you
04:35 pjs and the pillar would be: salt: repo: develop (or whatever)
04:35 manfred yeah, usually the branch that I am working on
04:35 manfred i spin up and down minions in the cloud like crazy
04:35 bhosmer joined #salt
04:36 whytewolf joined #salt
04:36 terminalmage pjs: np
04:37 schimmy1 joined #salt
04:46 pjs manfred hah, just realized... I use freebsd and the bootstrap script used system packages to install salt, which includes custom directories, etc. for salt configs. I think this state will change where salt looks
04:47 manfred yeah, you would need to write you own state
04:47 manfred i was just giving you an example
04:47 whytewolf joined #salt
04:47 manfred to start with
04:48 TyrfingMjolnir joined #salt
04:48 pjs cool.. thanks again. I'll toy around
04:48 manfred good luck
04:49 felskrone joined #salt
04:49 ramishra joined #salt
04:51 azylman joined #salt
04:51 felskrone1 joined #salt
04:56 fllr joined #salt
05:11 HACKING-TWITTER joined #salt
05:19 bhosmer joined #salt
05:20 gildegoma joined #salt
05:22 beneggett joined #salt
05:27 krow joined #salt
05:29 HACKING-TWITTER joined #salt
05:33 Luke_ joined #salt
05:38 canta joined #salt
05:43 azylman joined #salt
05:49 jalbretsen joined #salt
05:53 xsteadfastx joined #salt
06:02 bhosmer joined #salt
06:03 ndrei joined #salt
06:04 HACKING-TWITTER joined #salt
06:05 jpcw joined #salt
06:07 Outlander joined #salt
06:08 individuwill joined #salt
06:10 t0rrant joined #salt
06:10 mephx joined #salt
06:12 ajw0100 joined #salt
06:20 picker joined #salt
06:24 bhosmer joined #salt
06:24 thayne_ joined #salt
06:24 q4brk_ joined #salt
06:25 azylman joined #salt
06:27 rawzone_ joined #salt
06:31 matrix3000 joined #salt
06:31 matrix3000 left #salt
06:31 HACKING-TWITTER joined #salt
06:34 whytewolf joined #salt
06:40 whytewolf joined #salt
06:40 malinoff joined #salt
06:41 Hell_Fire joined #salt
06:45 bhosmer joined #salt
06:45 TheThing joined #salt
06:51 davet joined #salt
06:53 davet1 joined #salt
07:04 ml_1 joined #salt
07:06 bhosmer joined #salt
07:14 alanpearce joined #salt
07:16 jalaziz joined #salt
07:17 xmj state-p 2014.1.6?
07:18 xmj hm, no release..
07:18 * xmj forks into background
07:18 rofl____ when is 2014.2 due?
07:19 rofl____ salt need some traction before ansible takes over the show
07:20 mosen eh, i still need an agent, so no worries here :)
07:21 xmj rofl____: 2014.1.6 is bound to be released "any day now"
07:22 xmj and i think salt has more traction than ansible due to its being used for larger-infrastructure projects.
07:22 rofl____ maybe
07:22 rofl____ my impression is that ansible is a bit hotter right now
07:23 xmj "best tool for the job" /me puts $0.02 into the "useless phrases" piggybank
07:23 chiui joined #salt
07:23 rofl____ yeah, i totally agree
07:23 rofl____ i just want salt to win ;)
07:24 rofl____ more attention and bigger community
07:24 rofl____ is a win for all users
07:25 maxskew joined #salt
07:28 bhosmer joined #salt
07:30 yomilk joined #salt
07:38 Kenzor joined #salt
07:40 malinoff rofl____: ansible is great when you use it as intended by developers
07:40 malinoff and it is VERY useful as a remote execution tool
07:41 malinoff but it is a pain to do complex configuration management/deployments
07:41 malinoff Not because ansible is bad, but because developers didn't think i will use it as I use it
07:41 babilen rofl____: And there will never be a 2014.2 -- The .1 denotes the month of the release, so you might heope for a 2014.7 or 2014.8
07:43 rofl____ babilen: ah, ok :)
07:44 rofl____ malinoff: i havent tested it myself..i just see alot of references to ansible in alot of projects..not so much salt
07:44 ajw0100 joined #salt
07:45 malinoff rofl____: well, it is more suitable for sysadmins purposes. But as DevOps and python programmer I don't really like it. How they do things is horrible
07:45 rofl____ malinoff: good to hear ;)
07:46 malinoff I don't like salt too, but it is way better than ansible :)
07:46 luette joined #salt
07:47 chiui joined #salt
07:48 babilen malinoff is a grumpy old man ;)
07:49 malinoff Hey, I'm only 22 ;)
07:49 rofl____ early adopter ;)
07:50 bhosmer joined #salt
07:51 malinoff Excuse me, what is 'early adopter'?
07:51 malinoff urbandictionary doesn't know :(
07:51 Schmidt malinoff: what you learn by experience (regardless of age) is that all things suck, they just suck in different ways
07:52 malinoff Schmidt: thanks
07:52 phx Schmidt, couldn't have said better
07:52 rofl____ you choose what sucks the least
07:52 malinoff But I found something that doesn't suck
07:52 phx rofl____, you choose what you want to suck with imho
07:52 phx malinoff, that's a lie!
07:52 rofl____ the cake is a lie!
07:53 Schmidt malinoff: because you haven't tried to do X with it yet.
07:53 malinoff Schmidt: of course, but everything I tried so far works
07:53 Schmidt where X is something it sucks at
07:53 malinoff Sooo
07:53 malinoff well, I agree it sucks too
07:53 malinoff :)
07:54 malinoff Python itself doesn't suck. At least, 3.4
07:54 malinoff python2 is horrible after 3.4 :(
07:54 Schmidt I think brian d foy (well known perl programmer) had this notion that if you couldn't tell what the programming language work in sucks at, you where not a good enough programmer in that language yet
07:55 malinoff Funny to hear that from a perl programmer
07:56 Schmidt is probably not "just" a perl programmer
07:56 Schmidt he is*
07:57 malinoff I understood your position that everything sucks, and i'm partially agree. Like there is no a 100% healthy guy, but you may be healthy enough to just ignore some desease
07:58 jdmf joined #salt
07:59 Schmidt sure, but you can still get hit by a truck and die...
07:59 Schmidt and then it sucks :)
07:59 malinoff not the worst death, you know
08:01 phx there's no such thing as worst. there's always worse
08:03 Lomithrani joined #salt
08:03 marcin_ joined #salt
08:04 malinoff So you can dive in that recursion as deep as you want, and there always be something worse. I guess, the first thing from what you started doesn't suck anymore :)
08:05 phx there are different kind of suckings
08:06 phx there are things which are just simple pains in the ass, and there are things which suck so much that it makes the given thingie functionally inappropriate for the given purpose you want to use it for
08:06 bmcorser joined #salt
08:06 phx and there's a whole gradient inbetween those two
08:06 bmcorser can anyone explain the dot-dot syntax in salt states? (or point me to an article)
08:07 bmcorser i have conflicting IDs and i dont know why
08:07 CeBe joined #salt
08:08 bmcorser The conflicting ID is "foo" and is found in SLS "bar:baz.fizz" and SLS "bar:baz..fizz"
08:11 bhosmer joined #salt
08:12 linjan joined #salt
08:20 rnts joined #salt
08:21 ghaering joined #salt
08:26 CeBe1 joined #salt
08:29 Hell_Fire joined #salt
08:30 ramishra joined #salt
08:32 gildegoma joined #salt
08:33 bhosmer joined #salt
08:34 ndrei_ joined #salt
08:34 jhauser joined #salt
08:35 yomilk joined #salt
08:44 babilen bmcorser: The dot is simply used to delineate "namespaces" from each other and roughly correspond to "/" in your filesystem. A SLS file that is located in file_roots/foo/bar/baz/quux.sls would be referenced as foo.bar.baz.quux
08:46 babilen bmcorser: State IDs have to be *globally* unique which causes all sorts of pain, but also means that you catch conflicting states more easily.
08:46 xsteadfastx if a minion doesnt match the top file and i run salt '*' state.highstate is it normalthe mionion fails with the message "No Top file or external nodes data matches found"? i dont get why it fails... shouldnt it just get ignored? in the summary i have Failed: 1 for that minion. but it wasnt even supposed to get a state
08:49 darkelda joined #salt
08:49 malinoff xsteadfastx: in salt, all minions always receive all incoming messages even if you limit the numer of minions who should run a command, but it seems to me that minion shouldn't log that anyway
08:50 mateoconfeugo joined #salt
08:52 xsteadfastx malinoff: because i think "oh something failed" but its was allright that it doesnt do that stuff from that specifig state file
08:54 svx joined #salt
08:55 millz0r joined #salt
08:55 bhosmer joined #salt
08:57 alanpearce joined #salt
08:59 babilen xsteadfastx: If you want to run a single statefile that applies only to a subset of minions you should target only that subset.
09:00 babilen One might argue about logging an error in the case of a highstate run though.
09:03 N-Mi joined #salt
09:04 agend joined #salt
09:04 agend hi all
09:04 malinoff I guess that log message is there to catch situations when minion is present in a top file, but somehow he doesn't know anything about it
09:06 xsteadfastx malinoff: true... i will see if i put that failing minion in a dummy "group"
09:07 krow joined #salt
09:17 giantlock joined #salt
09:35 ramishra joined #salt
09:37 yomilk joined #salt
09:38 bhosmer joined #salt
09:47 miggy282 joined #salt
09:47 miggy282 hey
09:49 miggy282 when i run salt "*" state.highstate
09:50 miggy282 i get http://pastebin.com/esKr6Etf
09:51 TheThing joined #salt
09:52 malinoff miggy282: can you http://pastie.org your top.sls ?
09:53 TheThing joined #salt
09:53 yomilk joined #salt
09:53 kermit joined #salt
09:53 whytewolf joined #salt
09:55 miggy282 quite simple
09:55 miggy282 http://pastie.org/9346446
09:55 Sp00n you have a single and a double quote
09:59 miggy282 *puts glasses on *
09:59 bhosmer joined #salt
10:01 miggy282 still get the same error
10:01 bmcorser babilen: i am having trouble reading slates that reference things like - ..thing_name
10:01 bmcorser what does the .. mean?
10:01 bmcorser in bash we have "the directory above" and in python we have "the module above"
10:02 bmcorser what about salt?
10:02 babilen bmcorser: Are you sure about the double-dot. Is it not .thing.name ?
10:02 bmcorser nope
10:02 babilen Do you have an example?
10:03 bmcorser i'll try and make a paste that represents it brb :)
10:03 babilen sure
10:05 bmcorser babilen: https://gist.github.com/bmcorser/6e76fb36a36971714229
10:06 bmcorser why isn't install referenced ".install" or just "install"
10:08 babilen bmcorser: "- .install" would be a relative include and "- install" an absolute one. That is the former would include "database/install.sls" and the latter "install.sls"
10:10 bmcorser babilen: so "- ..install" should include "/install.sls" ie "/database/../install.sls" shouldn't it?
10:11 babilen I've only ever seen .include.me and the documentation also only references a single dot. Why do you think that a double-dot should traverse the SLS tree? Does it work as expected? Who wrote that?
10:11 picker joined #salt
10:13 bmcorser 1. I'm guessing from the two examples I know about (bash and python)
10:14 bmcorser 2. Yes it works, apart from some weirdness where I see "baz.fizz" conflicting with  "baz..fizz"
10:15 babilen http://docs.saltstack.com/en/latest/ref/states/include.html#relative-include is the documentation regarding this. I would have to look into the documentation, but I wouldn't rely on implementation details like that.
10:15 bmcorser 3. https://github.com/mickeynp
10:15 bmcorser (contracting at my company)
10:15 babilen I would also consider it a "code smell" if your state organisation makes relative imports like that necessary
10:16 bmcorser yes
10:16 bmcorser there's a whiff :)
10:19 babilen I mean you should simply use "- database" and "- install" in top.sls .. or just move the parts from "install.sls" that are needed in database into database.install .. that being said: a database/init.sls that does nothing but to include install.sls (*not* (!!!) database/install.sls) is highly unintuitive
10:19 babilen (assuming the ..foo.bar include actually works)
10:21 bhosmer joined #salt
10:23 Leech joined #salt
10:25 ramishra_ joined #salt
10:35 babilen Are there formulas/snippets somewhere that exemplify a tested way of "installing" software from a git repo on GitHub?
10:37 workingcats joined #salt
10:38 Outlander joined #salt
10:41 xsteadfastx i try to get salt.moduls.reg running... but it doesnt work. the doc says that it depens on winreg python module... do i need to download it on the minion?
10:42 bmcorser babilen: https://github.com/akoumjian/salt-django-example i look at this before, but never implemented, so untested :/
10:42 bhosmer joined #salt
10:42 babilen xsteadfastx: https://docs.python.org/2/library/_winreg.htm is what is being referred to I guess
10:44 babilen bmcorser: Yeah, something like that. I started to develop in exactly that direction, but thought: .oO( Why not stand on the shoulder of giants? )
10:45 bmcorser :)
10:46 babilen That example is, however, more or less doing what I came up with already, so I simply hopped from the shoulders of one dwarf to another
10:46 xsteadfastx i dont get it working. if i try the example from the docs i get a error: "ValueError: Could not convert the data to the specified type"
10:47 babilen xsteadfastx: Could you provide more information (preferrably on a pastebin such as http://paste.debian.net or http://refheap.com)
10:47 xsteadfastx http://paste.debian.net/107748/
10:49 elfixit joined #salt
10:49 xsteadfastx and the debug log from the minion http://paste.debian.net/107749/
10:49 babilen xsteadfastx: Which salt version is that?
10:50 babilen Looks like a genuine bug
10:50 xsteadfastx 2014.1.5 on the master
10:50 xsteadfastx and the same on the minion
10:50 miggy282 Hi still cant figure out why when i execute salt '*' state.highstate
10:50 miggy282 i get http://pastebin.com/esKr6Etf
10:51 babilen Okay, it doesn't get any newer (in terms of releases) as that.
10:51 aliabbas joined #salt
10:51 xsteadfastx babilen: and the rest works fine. as much as i see. installing pkgs and stuff
10:51 babilen miggy282: Huh? We had a look earlier and found that you had '*" in there (note ' vs ")
10:51 babilen You even put your glasses on.
10:52 miggy282 9346587
10:52 miggy282 http://pastie.org/9346587
10:52 miggy282 ive fixed the commenting
10:52 miggy282 sorry the quotation
10:53 babilen xsteadfastx: Might be a genuine bug. I am furtunate enough to not have to ever deal with Windows minions so I really don't know much about it. You might want to write to the mailing list as it could be that people who worked with that module before can provide additional information about it.
10:53 xsteadfastx babilen: ok.. then i better subscripe to the mailinglist... or should i file a bug?
10:53 babilen miggy282: You didn't. On line 2 you *still* have '*" and not '*'
10:54 babilen xsteadfastx: Filing a bug feels justified as you ran the example from the docs and I find the assumption that that would work to be quite justified
10:54 miggy282 babilen: i edit that
10:54 xsteadfastx babilen: ok... i will file a bug :)
10:55 miggy282 http://pastie.org/9346598
10:55 babilen miggy282: Then please provide current information. top.sls + SLS files + command and its entire output.
10:55 miggy282 yes above
10:55 babilen miggy282: Are you sure that the name of the package is htop-1.0.1-2.el6.x86_64 and not "htop" ?
10:56 miggy282 i've tried both
10:56 miggy282 http://pastie.org/9346601
10:56 babilen It strikes me as a rather peculiar name and I woult expect the latter.
10:57 babilen miggy282: Oh, and please add ":" after "base" on line 1 so it reads "base:"
10:57 miggy282 ahhh
10:57 miggy282 was actually looking at that
10:57 babilen You need better spectacles ;)
10:58 miggy282 :-D
10:58 miggy282 okay so that seem to have n effect
10:58 miggy282 but got a new error
10:59 babilen (and you use "htop" now rather than "htop-1.0.1-2.el6.x86_64" ?)
11:00 alanpearce joined #salt
11:00 miggy282 yes somthing to do with the spacing @  - server
11:01 miggy282 http://pastie.org/9346615
11:01 alanpear_ joined #salt
11:02 babilen miggy282: Yeah, you have an additional leading space in fron of "- servers". You want to indent by two spaces per indentation level.
11:04 Leech left #salt
11:04 bhosmer joined #salt
11:18 bhosmer joined #salt
11:19 yomilk joined #salt
11:20 miggy282 shmm
11:21 miggy282 http://pastie.org/9346658
11:21 miggy282 the top worked
11:21 miggy282 but he bottom didn't
11:22 kiorky joined #salt
11:24 viq Is anyone using pillar include with the key option? I remember trying it in the early 0.17 days and having problems with it
11:25 bhosmer_ joined #salt
11:29 ramteid joined #salt
11:46 bhosmer joined #salt
11:47 yomilk_ joined #salt
11:49 jrdx joined #salt
11:50 miggy282 would it be possible to manage users ssh keys from slat
11:52 viq miggy282: yes
11:53 miggy282 so my casew is
11:53 miggy282 case*
11:53 miggy282 salt '*' run.cmd "useradd -G example"
11:53 miggy282 then a user directory is created
11:54 miggy282 for example
11:54 viq miggy282: there's a user management module, no need to use cmd.run
11:55 miggy282 would that be salt.states.user
11:55 miggy282 model
11:55 miggy282 module*
11:57 viq miggy282: http://docs.saltstack.com/en/latest/ref/modules/all/index.html  http://docs.saltstack.com/en/latest/ref/states/all/index.html
11:58 viq so either http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.pw_user.html or http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.useradd.html
11:58 viq if you want execution modules
12:00 vbabiy joined #salt
12:00 vejdmn joined #salt
12:03 jas-_ joined #salt
12:03 Lomithrani joined #salt
12:04 Luke_ joined #salt
12:09 bhosmer joined #salt
12:22 agend joined #salt
12:24 maxskew joined #salt
12:26 vbabiy joined #salt
12:28 diegows joined #salt
12:30 bhosmer joined #salt
12:35 gadams999 joined #salt
12:37 to_json joined #salt
12:41 scoates_ joined #salt
12:43 nyx joined #salt
12:45 patrek joined #salt
12:52 bhosmer joined #salt
12:53 babilen miggy282: There is also http://docs.saltstack.com/en/latest/ref/states/all/salt.states.user.html and https://github.com/saltstack-formulas/users-formula
12:53 miggy282 thnak you
12:55 bhosmer_ joined #salt
12:55 picker joined #salt
13:01 ipmb joined #salt
13:04 gildegoma joined #salt
13:12 viq WTF? http://pbot.rmdir.de/qLD049fOu-DRoD0QnWnolg
13:14 viq any ideas?
13:14 viq where pf:network: has something like example_org: 10.20.30.0/24
13:15 maboum joined #salt
13:16 babilen viq: Could you paste your actual pf.conf jinja file?
13:16 babilen Also: yay for pf!
13:17 hellome joined #salt
13:19 mapu joined #salt
13:19 tkharju2 joined #salt
13:21 ixokai joined #salt
13:22 hachaboob joined #salt
13:22 hachaboob Has vsphere hit production salt yet?
13:23 alanpearce joined #salt
13:23 mgw joined #salt
13:24 timoguin hachaboob: no, still the develop branch
13:24 viq babilen: it's $work, so somewhat thinking I'd rather not :(
13:24 hachaboob Anyone using it in dev?
13:24 babilen viq: maybe just the line that causes the error and immediate surroundings?
13:25 premera joined #salt
13:27 viq http://pbot.rmdir.de/CEGvlPH2adGhCWyyqaaMpg
13:28 viq error complains about line 106
13:29 TheThing joined #salt
13:30 vejdmn joined #salt
13:32 jslatts joined #salt
13:32 xmj nice, pf rules in #salt, i like that!
13:33 xmj viq: obsd pf i presume, huh
13:33 vejdmn joined #salt
13:33 racooper joined #salt
13:34 viq xmj: nah, sadly freebsd
13:35 xmj even that i encourage, our pf has smp support :p
13:35 Lomithrani joined #salt
13:35 viq :P
13:35 viq and we do some trick, like putting IPs in there, but doing them by salt querying the DNS
13:35 viq but, it's crapping out at the moment, and trying to figure out why
13:37 babilen viq: Might some jinja shite that stumbles over the { in there
13:40 viq so what, \{  ?
13:42 jnials joined #salt
13:42 jalbretsen joined #salt
13:43 babilen viq: But that is probably not it, but boil it down to somthing super-simple, such as a file with *only* "{{ salt['pillar.get']('pf:network:example_org') }}" in it. It might be your pillar too and I have to confess that I don't immediately see something wrong with your file.
13:46 aquinas joined #salt
13:50 mgw joined #salt
13:51 thedodd joined #salt
13:56 aquinas joined #salt
13:58 tkharju joined #salt
13:58 AdamSewell joined #salt
14:00 tyler-baker joined #salt
14:05 munhitsu_ Question for salt practitioners. Is publish.publish reliable enough to build loadbalancer configuration using it?
14:05 munhitsu_ https://www.irccloud.com/pastebin/InTCvipB
14:06 Luke_ joined #salt
14:11 dude051 joined #salt
14:17 aquinas joined #salt
14:17 ndrei joined #salt
14:17 ndrei_ joined #salt
14:20 munhitsu_ I'm thinking whether not salt mine would be safer choice
14:20 munhitsu_ I'm afraid to land in situation that minion will be to busy to respond to network.ip_addrs and it will be removed
14:21 acabrera joined #salt
14:21 kaptk2 joined #salt
14:22 cheus joined #salt
14:23 N-Mi joined #salt
14:25 manfred munhitsu_: the mine might have stale data
14:27 bmcorser yikes im getting an ugly one on my pip state
14:27 bmcorser CommandExecutionError: The `exists_action`(`--exists-action`) pip option only allows one of (s, i, w, b) to be passed. The 'switch' value is not valid.
14:28 bmcorser oops my bad
14:35 bhosmer joined #salt
14:36 mgw joined #salt
14:37 vejdmn joined #salt
14:37 kballou joined #salt
14:39 vejdmn joined #salt
14:41 aquinas joined #salt
14:49 munhitsu_ manfred: stale as date from yesterday, or stale as in data from last 5 mins
14:49 munhitsu_ manfred: I'd rather configure too many application servers on haproxy instead of 0
14:49 manfred ok, then yeah the mine should work fine
14:50 munhitsu_ what's your practice?
14:50 manfred i use cloud.query... but that is just me
14:50 munhitsu_ lucky one
14:51 munhitsu_ thx
14:51 manfred np
14:51 TheThing joined #salt
14:51 felskrone joined #salt
14:52 viq babilen: the problem was that in a line way above that the host that salt was to resolve had a typo so didn't resolve
14:53 viq munhitsu_: if a minion is too busy to respond to network.ip_addrs then it's probably too busy to respond to your app as well
14:55 stevedndm joined #salt
14:55 stevedndm Is there a builtin way to issue a service upgrade such as for nginx?
14:57 aquinas joined #salt
14:57 viq stevedndm: "service upgrade" ?
14:58 munhitsu_ viq: Valid point, though in this case I don't want to use salt as monitoring tool. This information comes from a different place.
14:59 miqui joined #salt
14:59 viq munhitsu_: fair enough
15:00 lukelivingstone joined #salt
15:00 stevedndm Yes, afaik the service module and state do not support sending an upgrade/USR2 signal. So I would have to resort to cmd.run
15:00 lukelivingstone left #salt
15:01 bhosmer joined #salt
15:01 manfred stevedndm: the service state supports a reload: True
15:02 manfred stevedndm: third example http://docs.saltstack.com/en/latest/ref/states/all/salt.states.service.html
15:02 manfred does service <state> reload
15:03 penguin_dan joined #salt
15:04 stevedndm manfred: typically reload issues HUP not USR2
15:05 manfred if the nginx script doesn't do USR2, then you will want to use the cmd.run state
15:06 stevedndm My scripts do use USR2 it's just that reload and upgrade are different and I don't want to fudge my inits just because of salt
15:06 Luke_ joined #salt
15:06 stevedndm cmd it is I guess
15:09 krow joined #salt
15:09 rallytime joined #salt
15:12 quickdry21 joined #salt
15:15 tcotav joined #salt
15:17 Lomithrani what is the easiest way to get the minion IP on a jinja file
15:18 Lomithrani ?
15:19 dude051 joined #salt
15:20 aquinas joined #salt
15:20 ramteid joined #salt
15:21 conan_the_destro joined #salt
15:21 mapu left #salt
15:22 ajolo joined #salt
15:23 thayne_ joined #salt
15:26 aquinas joined #salt
15:27 Lomithrani {{ salt['mine.get'](grains['id'],'network.ipaddrs').items()[0][1][0] }} works but doesnt seem very neat , any other option ?
15:28 racooper grains['fqdn_ip4'] probably? you can look at grains.items to see what's available.
15:30 dvogt joined #salt
15:30 anotherZero joined #salt
15:32 higgs001 joined #salt
15:32 Lomithrani something like that yes thanks =)
15:33 jeffG_ joined #salt
15:33 jeffG_ morning all
15:34 jeffG_ I'm a bit of newb so bear with me...
15:34 dude051 joined #salt
15:34 jeffG_ I am trying to deploy apache to a server
15:34 jeffG_ in /etc/salt/grains
15:34 jeffG_ I have
15:34 jeffG_ roles: - webserver
15:34 jeffG_ and in top.sls
15:34 jeffG_ base:
15:35 jeffG_ 'roles: webserver'
15:35 jeffG_ - match: grain
15:35 jeffG_ - apache
15:35 jeffG_ but I get "No Top file or external nodes data match found"
15:35 manfred base:
15:35 manfred 'roles: webserver'
15:35 manfred - match: grain
15:35 manfred - apache
15:36 manfred also, make sure salt \* grains.get roles gives you just the webserver item, because it looks like you have a list of stuff in roles, and not just one role
15:36 jeffG_ manfred: yes that is what i have... sorry i didn't indent my message correctly
15:37 jeffG_ we are attempting a headless install if that makes a diff
15:37 mateoconfeugo joined #salt
15:38 jeffG_ I get
15:38 jeffG_ local:
15:38 jeffG_ - webserver
15:39 bhosmer joined #salt
15:39 manfred yeah, so that is a list, and not just a key value assignment
15:40 manfred so your roles: webserver won't work
15:40 jeffG_ manfred: what do I need to change then?
15:40 manfred you would just want roles: webserver
15:40 manfred not
15:40 manfred roles:
15:40 manfred - webserver
15:41 meteorfox joined #salt
15:41 jeffG_ manfred: do you mean in /etc/salt/grains?
15:41 manfred yes
15:43 ccase joined #salt
15:44 racooper manfred, what jeffG_ is showing and you're saying won't work is exactly what I have that works, for grains and for top base entry.
15:44 jeffG_ manfred: no such lucl
15:44 jeffG_ luck
15:44 jeffG_ racooper: can you clarify?
15:44 logix812 joined #salt
15:45 manfred racooper: yeah, you take over, i am trying to get some stuff together for the salt sprint next week
15:46 aquinas joined #salt
15:46 racooper https://gist.github.com/racooper/4f01039842733356383d is what I currently have for grains assignment and top.sls.
15:46 manfred oh, interesting
15:46 manfred hrm
15:46 manfred racooper: that works?
15:46 racooper has been since I set it up months ago
15:47 manfred cool
15:47 manfred i didn't know it would go through lists like that
15:48 racooper if you're trying to parse it in jinja you have to use for 'webserver' in grains['roles'] but for topfile and commandline targeting 'roles:webserver' works.
15:48 manfred if 'webserver' in grains['roles'] but yeah :P
15:48 racooper s/for/if
15:48 racooper heh yea that
15:50 racooper so jeffG_ does that help? or have I just confused things?
15:50 jeffG_ it was the fraggin' space I had in the top.sls between roles: and webserver
15:50 jeffG_ cant be 'roles: webserver'...
15:50 jeffG_ must be 'roles:webserver'
15:51 manfred nice
15:51 racooper ah, yes python is picky about whitespace for sure....
15:51 alanpearce joined #salt
15:53 smcquay joined #salt
15:54 tligda joined #salt
15:54 MTecknology did anything with authentication change with 2014.1.5?
15:55 manfred nothing major should have?
15:55 MTecknology :S
15:55 MTecknology it seems every key that I'd accepted no longer is
15:56 MTecknology weird...
15:56 MTecknology the upgrade got rid of my symlink
15:57 manfred symlink to where?
15:57 MTecknology /etc/salt/pki --> /srv/saltshare/etc/salt/pki
15:57 manfred or more... from where to where. if it is a packaged file, the package manage probably overwrote it
15:57 manfred which distro?
15:57 MTecknology debian
15:58 MTecknology I use ocfs2 to keep keys in sync across syndics
15:59 manfred in archlinux, we use the NoExtract directive in /etc/pacman.conf to make sure that packges don't overwrite stuff on updates
15:59 manfred not sure if there is an equivalent in dpkg
15:59 MTecknology this is a weird one... it shouldn't have been replaced
16:00 MTecknology there are a lot of tools to make sure you don't overwrite files users changed, even for directories
16:00 vu__ joined #salt
16:00 MTecknology I need to take a look at the packaging and make sure it's being used.
16:00 ipalreadytaken joined #salt
16:03 logix812 joined #salt
16:03 MTecknology the good news, it's one very easy fix
16:04 thayne_ joined #salt
16:05 KyleG joined #salt
16:05 KyleG joined #salt
16:05 benturner joined #salt
16:05 conan_the_destro joined #salt
16:05 MTecknology manfred: in Debian, it looks like this - http://dpaste.com/1M8FD5T - If I'd chosen not to overwrite my file, their update would have been created as /etc/ld.so.conf.dpkg-dist
16:05 MTecknology (if you're curious)
16:07 Luke_ joined #salt
16:07 conan_the_destro joined #salt
16:13 Theo-SLC joined #salt
16:15 joehillen joined #salt
16:16 tyler-baker joined #salt
16:19 perfectsine joined #salt
16:19 jaimed joined #salt
16:19 Lomithrani hi , how can I download a file only once ?
16:20 Lomithrani http://pastebin.com/f0BGF9e7
16:20 Lomithrani I've tried this without success
16:20 Lomithrani (the unless)
16:22 ajolo joined #salt
16:25 jpcw joined #salt
16:25 azylman joined #salt
16:25 jpcw joined #salt
16:29 jnials joined #salt
16:29 azylman_ joined #salt
16:31 jnials_laptop joined #salt
16:31 manfred Lomithrani: you can use the archive.extracted state to do that
16:31 manfred http://docs.saltstack.com/en/latest/ref/states/all/salt.states.archive.html
16:31 manfred downloads the file, and has a thing that only does it if a directory you specify is missing
16:32 Lomithrani seems perfect thanks
16:32 manfred np
16:34 dude051 joined #salt
16:35 vejdmn joined #salt
16:35 jnials joined #salt
16:35 troyready joined #salt
16:36 tyler-baker joined #salt
16:37 mgw joined #salt
16:38 oz_akan_ joined #salt
16:43 quantumriff joined #salt
16:46 forrest joined #salt
16:49 chrisjones joined #salt
16:50 quantumriff I want to check if a directory is present (ie, a users home folder) and if so, add an ssh key.  however, we don't want to create the home folder.. any idea how to go about this?
16:51 quantumriff or, I guess, we could 'properly' build their home directory, like if they had logged in directly..
16:52 krow joined #salt
16:53 quantumriff once the user is present (ie, added to active directory) we want to build their home directory, and setup their public key
16:55 dstokes morning y'alls
16:56 ml_1 joined #salt
16:58 picker joined #salt
16:59 schimmy joined #salt
17:01 newlife joined #salt
17:02 schimmy joined #salt
17:02 newlife hi,when I run salt-master ,there are eight processes started by the main process, how can I get to know what these process do ?
17:03 jalaziz joined #salt
17:03 oncallsucks What is a case where salt is more beneficial over just cloning a virtual machine if the setups come out the same?
17:04 aquinas joined #salt
17:04 smcquay arbitrary task execution after initial install, oncallsucks ?
17:06 dstokes oncallsucks: not having to maintain config images (you need provisioning to create the snapshot anyways)
17:07 smcquay speed, perhaps? in my experience esxi clones take way more time than apt activity against a local repo
17:08 Luke_ joined #salt
17:16 mgarfias how do i go about debugging why my pillar data isn't getting to the minions?
17:16 mgarfias it seems like i have to do a magic chicken dance to get it to work and i'm never sure why
17:16 Ryan_Lane joined #salt
17:17 aquinas joined #salt
17:18 jslatts joined #salt
17:20 shaggy_surfer joined #salt
17:26 aw110f joined #salt
17:26 Gareth morning
17:27 newlife In china,it mid night....
17:27 Gareth morning is a relative term :)
17:27 forrest hey Gareth
17:28 Gareth forrest: howdy.  hows it going?
17:30 bhosmer_ joined #salt
17:30 rallytim_ joined #salt
17:31 forrest Gareth, another day, glad it's a 4 day week, you?
17:31 mgarfias my pillar dta isnt working at all.  wtf
17:32 mgarfias hmmasdf
17:32 forrest mgarfias, so stop writing bad pillar data, duh :P
17:32 mgarfias no, just this one item
17:32 mgarfias ok
17:32 mgarfias gah
17:33 mgarfias the one that isnt working is valid yaml.
17:33 masterkorp hello
17:34 rallytim_ joined #salt
17:34 oz_akan_ joined #salt
17:35 mgarfias just dropped in a simple test into the pillar and it isn't being picked up
17:35 masterkorp I while configuring salt-cloud, i set my provider and profile on a "cloud:" pillar data
17:35 masterkorp gave that pillar data to the salt master minion on the top sls
17:35 masterkorp ran highstate
17:36 oz_akan_ joined #salt
17:36 masterkorp then i run "salt-cloud --list-images myprovider" and got Error: There was an error listing images: No cloud providers matched 'myprovider'. Available selections:
17:36 masterkorp and nothing
17:37 oz_akan_ joined #salt
17:37 masterkorp how do I make sure salt-cloud is really the pillar data ?
17:37 vbabiy_ joined #salt
17:38 Gareth forrest: no 4 day week for me unfortunately.  on call again.
17:39 rallytime joined #salt
17:39 ajprog_laptop1 joined #salt
17:40 jnials joined #salt
17:40 UtahDave joined #salt
17:40 chiui joined #salt
17:41 masterkorp oh i need to call the salt module directly
17:41 masterkorp makes sense
17:44 jpcw joined #salt
17:46 alekibango joined #salt
17:51 masterkorp Another question, what is the simplest way to convert pillar data to an yaml file ?
17:54 mgarfias salt-call pillar.get <whatever> --out yaml
17:54 masterkorp from a state ?
17:55 masterkorp basically the apllication I am salting uses yaml configuration files
17:59 kermit joined #salt
18:02 kivihtin joined #salt
18:02 Ahlee salt-call runs on the minion
18:02 Ahlee That would be the simplest way, though.  Probably not the best long-term unless you like the idea of masterless
18:03 dvogt joined #salt
18:08 jalaziz joined #salt
18:10 vejdmn joined #salt
18:13 AdamSewell joined #salt
18:13 mgarfias how do you set an ec2 tag in the salt-cloud config?
18:13 bhosmer joined #salt
18:14 forrest mgarfias, http://docs.saltstack.com/en/latest/topics/cloud/aws.html#modify-ec2-tags
18:15 mgarfias yes i saw that
18:15 mgarfias not what i'm asking
18:15 mgarfias i want to do it in my cloud profile stuff
18:15 mgarfias not by hand
18:16 forrest scroll up
18:16 forrest like 2 inches
18:16 forrest mgarfias, ^
18:17 mgarfias sorry
18:17 mgarfias i'm having a day.
18:17 forrest I can tell, WHAT A DICK, JEEZ
18:17 forrest :P
18:17 smcquay joined #salt
18:19 vbabiy joined #salt
18:20 * whiteinge gets out the additude adjustment tickle feather
18:21 Theo-SLC I want to jinja template a file, but I don't have a source for it because It was extracted from an archive.  How can I still template it with file.managed?
18:22 forrest Theo-SLC, you can't
18:22 forrest Theo-SLC, extract it from the tar, re-tar
18:23 Theo-SLC The file in the archive has the appropriate {{}} jinja tags
18:23 forrest yea but salt does't know about it
18:23 smcquay joined #salt
18:23 forrest it doesn't understand the items inside the archive, it just understands the archive
18:23 forrest why use a tar in the first place :(
18:24 forrest sounds like a bad piece of software
18:24 ckao joined #salt
18:24 forrest can you not package up the components, and exclude the conf?
18:26 Theo-SLC it's the configuration file for newrelic.
18:26 Theo-SLC I'll extract the conf file and manage seperately, thanks.
18:26 forrest np, I wish there was a way to do that it would be cool!
18:27 austin987 joined #salt
18:28 druonysus joined #salt
18:28 to_json joined #salt
18:34 krow joined #salt
18:35 chrisjones joined #salt
18:39 whiteinge Theo-SLC: there's a recent ticket to add a stand-alone render module for this sort of thing
18:39 whiteinge https://github.com/saltstack/salt/issues/13873
18:40 econnell joined #salt
18:41 mgarfias next bit of confusion, how do i access the ec2_tags in a top file
18:41 mgarfias i assumed it was a grain, but it appears that doesn't work out of the box
18:43 econnell i had a question about startup_states.  If I'm understanding the documentation correctly, the salt client can connect to the master on startup and run the configuration defined on the master (similar to a puppet agent).  Is that correct?
18:44 manfred yes
18:44 econnell awesome, thanks
18:44 manfred it can run a specified set of states, or just go into a highstate
18:44 econnell right.. and highstate is basically a "full run", right?
18:45 manfred it is whatever is specified in the top.sls files
18:45 econnell ok, thanks
18:45 manfred np
18:49 bhosmer joined #salt
18:52 econnell i'm kinda curious about mgarfias question about ec2 tags as well
18:54 forrest econnell, ?
18:54 forrest I linked the answer
18:54 forrest my link, then scroll up a few inches
18:54 forrest there's an example
18:54 econnell hrmm… i don't see it in the channel
18:54 forrest http://docs.saltstack.com/en/latest/topics/cloud/aws.html#modify-ec2-tags
18:54 econnell perhaps before i joined
18:55 econnell thanks
18:55 forrest np
18:55 forrest just scroll up a little ways and you will see it
18:55 econnell i just joined a couple minutes ago
18:55 forrest oh you were asking about the top file one
18:55 forrest I gotcha
18:55 forrest yea that one I don't know
18:55 forrest sorry
18:56 econnell np… thanks for the link though
18:56 forrest yea np
18:56 forrest this was related to setting ec2 tags in the cloud confs
18:59 CeBe2 joined #salt
19:01 TheThing joined #salt
19:01 jaimed joined #salt
19:01 mgw joined #salt
19:03 logix812 joined #salt
19:09 Luke_ joined #salt
19:11 vejdmn joined #salt
19:16 CeBe joined #salt
19:18 to_json joined #salt
19:20 Supermathie joined #salt
19:21 kermit joined #salt
19:22 Supermathie Morning everyone. I'm trying to run a remote (win) command with a file from salt:// but it doesn't interpolate arguments:
19:22 Supermathie salt -t30 'win2012r2-dc1' cmd.script salt://win/files/ImportUsers.ps1 shell='powershell' args=salt://win/files/testusers.csv
19:22 mgarfias you know, it'd be nice if salt would toss an error if the pillar env you tell it to use for a minion isn't accessible
19:22 TheRealBill joined #salt
19:23 Supermathie I think I should cache it, and reference the cache but aren't sure what to do on the master to reference the location.
19:23 mateocon_ joined #salt
19:23 dvogt joined #salt
19:24 mateoconfeugo joined #salt
19:27 shaggy_surfer joined #salt
19:32 xzarth_ joined #salt
19:34 ingwaem joined #salt
19:35 ingwaem Hey dude, finally managed to get that pillar mysql stuff working
19:35 ingwaem had to hack the stuff you sent me and another mysql.py I found online
19:35 ingwaem Config is pretty simple…in master use this:
19:35 ingwaem ext_pillar:
19:35 ingwaem - mysql:
19:35 ingwaem mysql_query: 'SELECT machine_variables.name as pillar, machine_variables.val as value FROM machine_variables LEFT JOIN machines on machine_variables.machine_id = machines.id WHERE machines.machinename = ''%s'''
19:36 LT joined #salt
19:37 ingwaem oh pasted to the whole channel :) hehe what fun
19:45 cheus joined #salt
19:51 racooper ingwaem,  pastebin (or gist) is your friend (and ours!) :)
19:52 ingwaem oh yes :) done :) I followed up immediately after that :)
19:52 ingwaem Situation is the latest mysql module in pillars produces errors…so somehow I’ve hacked the code to make it work…this might be helpful for some of you out there:
19:52 ingwaem http://pastebin.com/qPAt0Bxb
19:53 ingwaem To get it instantiated correctly use this in your config: http://pastebin.com/kNWmAdjP
19:53 TaiSHi Strangely, I see the ... almost like an underscore
19:57 Ahlee how does one get __opts__ into an external pillar?
19:59 forrest Ahlee, http://salt.readthedocs.org/en/latest/topics/development/external_pillars.html#options ?
20:00 Ahlee forrest: i want to know what the __opts__ are while I'm debugging this
20:00 Ahlee without resorting to print __opts__
20:01 forrest Ahlee, ahh, yea I don't know then
20:01 ingwaem Ahlee you can use the logger function..that way you’re not directly printing…then if you have salt masster in debug mode you’ll see your stuff
20:02 Ahlee No way to get it into an interactive shell?
20:02 bhosmer_ joined #salt
20:02 ingwaem hmm not sure on that…I don’t know enough about python myself to know
20:02 druonysuse joined #salt
20:02 ingwaem the link I posted with the hackery I did last night includes some output code for the options…you may find it useful
20:03 ingwaem http://pastebin.com/qPAt0Bxb
20:03 ingwaem it’s not the most perfect of modules, but it works
20:05 vexati0n joined #salt
20:06 vexati0n so... still have a pub_port issue. tcp connections work fine (verified with tcpdump and telnet) in both directions, but the salt master never sends out a command.
20:10 TheThing joined #salt
20:10 Luke_ joined #salt
20:10 thedodd joined #salt
20:16 TaiSHi Is it me or doc style changed ?
20:16 TaiSHi Looks nicer now
20:17 ingwaem they’ve changed
20:17 TaiSHi I like the new one
20:17 TaiSHi Looks pretty easy to read
20:17 ingwaem I think the new helium release is getting ready to ship…some of it is newer than the code :)
20:19 forrest TaiSHi, you mean for the code blocks?
20:21 babilen Ah, explicit YAML code blocks. This might herald the arrival of "switchable" examples (yaml, py, pydsl, mako, ...) in the future (where available)
20:21 UtahDave joined #salt
20:21 Ryan_Lane joined #salt
20:22 ingwaem Greetings Dave!
20:22 forrest currently it supports bash, yaml, python, and one other I believe.
20:27 gq45uaethdj26jw6 joined #salt
20:27 gq45uaethdj26jw6 left #salt
20:30 TaiSHi forrest: ja
20:31 krow joined #salt
20:32 mgarfias for the life of me i can't get pillar data to show up
20:32 * TaiSHi needs to start working on his pillar/top for backends
20:33 higgs001 joined #salt
20:34 thayne_ joined #salt
20:35 kivihtin joined #salt
20:38 Supermathie joined #salt
20:42 jnials joined #salt
20:43 vejdmn joined #salt
20:48 mgarfias how do i target a minion based on two grains?
20:48 manfred mgarfias: compound targeting
20:48 manfred mgarfias: http://docs.saltstack.com/en/latest/topics/targeting/compound.html
20:50 btorch is there a way to make a pkg.installed state no update a pkg ?
20:52 mgarfias manfred: thanks, i was using the incorrect syntax it looks like
20:52 jchen btorch: you can pin it to a specific version string, iirc
20:57 azylman joined #salt
20:57 mateoconfeugo joined #salt
20:58 toastedpenguin joined #salt
20:59 azylman_ joined #salt
21:01 m1crofarmer joined #salt
21:03 TheTHing_ joined #salt
21:07 mgarfias ok, new problem: i specify a grain in the cloud config, and it works for one instance, but not others
21:07 oz_akan_ joined #salt
21:09 ingwaem magarfias: did you highstate all the others so that they update their grains lists? is there any chance where you’re defining the grain may be targeting a specific grain that limits it’s relevance?
21:10 zz_Cidan joined #salt
21:11 Cidan joined #salt
21:11 Luke_ joined #salt
21:11 ingwaem simple way to see what’s on all the minons is a salt \* grains.items
21:14 shaggy_surfer joined #salt
21:14 aw110f I'm running 2014.1.5 and I'm seeing getting this "KeyError: 'tgt' when running jobs.active
21:15 aw110f http://pastebin.com/q5CGqH51
21:15 hltbra joined #salt
21:15 yomilk joined #salt
21:16 hltbra hello. I am testing salt-master and salt-minion, and I would like to use the rest api in salt master, but rest_wsgi didnt work as I expected. my master config is rest_wsgi: port: 9988
21:16 hltbra I got bind errors
21:17 hltbra I think it's because there are multiple salt-master processes running when I start salt-master
21:17 hltbra has anyone experienced anything similar or knows a solution for that?
21:20 maxleonca joined #salt
21:21 ingwaem hltbra, if salt-master does’t terminate correctly you can get into a situation where you end up with multiple instances and they conflict
21:21 ingwaem I find “pkill salt-master” helps with those situations
21:21 ingwaem to make sure they’re all dead before starting a new one, check if there are any processes running “ps aux | grep salt-master”
21:22 ndrei joined #salt
21:22 forrest TaiSHi, yea a coworker of mine made that change
21:23 jacksontj joined #salt
21:23 forrest he wasn't happy with how clear it was to tell the difference between code blocks :P
21:23 mgarfias ingaem: the grain should be applied when bootstrapping the instances.  and i'm sync'ing grains
21:23 ingwaem however due to the nature of the inter communication, salt uses queues and batches…defined in the master config as worker_threads…defaulted to 5…meaning there will be 5 agents that work in parallel in a batch
21:23 maxleonca Hey guys, I think is kind of a dumb question. but I rather be safe...   If I can confirm a problem with the salt package for Centos 6 from EPEL, the bug report goes to EPEL right?
21:23 forrest maxleonca, just hit up terminalmage
21:23 hltbra thanks ingwaem, but it looks ok. salt creates multiple processes by default, right?!
21:24 forrest maxleonca, or more realistically, file an issue
21:24 ingwaem yes
21:24 UtahDave maxleonca: No, please open the issue on Salt's github issue tracker
21:24 hltbra ingwaem: I dont know why the rest server is not up
21:24 ingwaem hltbra: yes
21:24 maxleonca Roger, roger
21:24 maxleonca Thank you.
21:24 ingwaem hltbra: hmm, if your rest server isn’t up…is it possble you updated just very recently??
21:24 forrest don't actually just hit up terminalmage. I was joking :P
21:24 forrest if that wasn't clear
21:25 hltbra ingwaem: latest official ubuntu release, using rest_wsgi
21:25 fhond joined #salt
21:26 ingwaem hltbra: I hit the same problem the other day…halite has started on 8080 but not 8000…I think it’s because there’s an update going on…part of the documentation references features that don’t yet exist in the released code…I think wait a few days and try again, or ensure you install salt-api as per the old process
21:26 hltbra :( I will try old salt-api then. thanks ingwaem
21:27 ingwaem hltbra: you shouldn’t need it for too long…api is being built into the new release as part of the core
21:27 patrek joined #salt
21:27 whiteinge hltbra: which salt version are you testing this on?
21:27 ndrei joined #salt
21:27 hltbra whiteinge: salt-master 2014.1.5
21:29 whiteinge hltbra: ok, ingwaem is correct. you need the separate salt-api install. salt-api is included in the develop version of salt
21:29 whiteinge if it's complaining that the port is bound, i wonder if there's an old instance of that rest_wsgi module still running...
21:29 hltbra ok, thanks whiteinge and ingwaem. I am using salt-api and it looks like it's fine
21:31 terminalmage maxleonca: even if it's a packaging issue, you can file it on our github issue tracker
21:31 ingwaem hltbra: ~~thumbs up~~
21:32 ndrei_ joined #salt
21:33 Outlander joined #salt
21:34 aw110f Does anyone know which branch was EPEL RPM salt 2014.1.5, was built from?
21:35 manfred aw110f: it was built from the 2014.1.5 tag
21:35 manfred https://pypi.python.org/pypi/salt/2014.1.5
21:36 mgarfias getting a stack trace when running high state.  Can someone take a look at it and tell me whats going on?  https://gist.github.com/mgarfias/820c82546abb0f4ca615
21:36 aw110f any reason why this fix did not get merged get into 2014.1.5 tage? https://github.com/saltstack/salt/commit/537f96fe#diff-5ca65b130737a573488760e87b926385R179
21:36 aw110f *tag
21:37 manfred it would have to be backported to the 2014.1 branch, and then it will be in the next tag
21:37 manfred feel free to ask why it wasn't backported in the pull request
21:37 aw110f ok thanks manfred:
21:39 patrek_ joined #salt
21:41 maxleonca @terminalimage Thanks I just did.
21:41 Eureka joined #salt
21:42 AdamSewell joined #salt
21:42 mgarfias and i still can't get the grain to load into some, but not all, of my minions
21:43 dfinn joined #salt
21:43 dfinn is there a "best practice" for how to manage hosts in a secured DMZ from a master in the non-DMZ
21:43 savvy-lizard joined #salt
21:45 sxar_ joined #salt
21:50 ingwaem mgarfias: just reading up on grains and I see they’re set in the minion configs…did you update all the minions so that they have the updated grains in their own configs? May need to restart the minion service too after doing that so they pick them up…however just thinking you may be better off with pillars…specifically a pillar_roots if they are global variables…just update the master pillar file in /srv/pillar and it will
21:50 ingwaem populate those values to all the minons…but a little tingle is pointing me towards the configs if it’s spefically grains you need control over
21:50 ajprog_laptop1 joined #salt
21:51 ingwaem mgarfias: additionally you may just need to force the minions to update their grains:  saltutil.sync_grains
21:51 ingwaem salt '*' saltutil.sync_grains
21:53 mgarfias ingwaem: I flushed my instances and rebootstrapped fresh
21:55 mgarfias also, did the sync and no go
21:55 mgarfias now that i have highstate working i can log into the host and see whats in the minion cfg tho
21:56 thayne_ joined #salt
22:02 emostar is there a way to have salt cloud provision a server and then run 'state.highstate' on the new server?
22:03 ingwaem mgarfias: that’s very weird…what sort of inconsistencies do you get? are there specific grains missing that you expected or are all the grains missign on some?
22:03 mgarfias all grains missing on the ones that miss it
22:03 manfred emostar: yes
22:04 mgarfias i'll go gist the bits form the cloud cfg
22:04 manfred emostar: startup_states: highstate, in the minion: block in your cloud profile
22:04 manfred emostar: http://docs.saltstack.com/en/latest/ref/states/startup.html
22:04 emostar manfred: i looked at that and saw a bug on github that it didn't work... I'll give it a try now though, thanks!
22:05 manfred emostar: it works for me :/
22:05 manfred emostar: just drop
22:05 manfred minion:
22:05 manfred startup_states: highstate
22:05 manfred in /etc/salt/cloud
22:05 manfred or in your profile
22:05 emostar ok, thanks trying it now
22:05 mgarfias ingwaem: https://gist.github.com/mgarfias/2ad2a8729898fc010a67
22:06 mgarfias and: https://gist.github.com/mgarfias/cb457ba2fe7d6a94a8d2
22:06 manfred emostar: can also use start_action http://docs.saltstack.com/en/latest/topics/cloud/deploy.html#post-deploy-commands
22:06 emostar manfred: ahh, sorry, that was the one that i saw the bug for!
22:06 manfred yeah
22:08 emostar dfinn: i'd be curious in a solution to that as well.i considered opening up the DMZ to provision a master inside the DMZ as well, then have it's state managed by the non DMZ master
22:08 dfinn we are considering a master in the DMZ but it seems like a pain to have to manage multiple.  on the other hand though, it seems risky allowing our dmz hosts access to our internal master.  we also aren't fond of the way that all the states are pushed down to every minion and can be viewed
22:09 dfinn we only have a few hosts in the dmz, kind of also looking at using local states to avoid having to deal with a master.  anyone done that?
22:10 emostar manfred: great, that worked. albeit silently so i had no idea it was calling highstate
22:10 manfred yeah, that is one downside to startup_states
22:10 ingwaem mgarfias: reading…sorry, bunch of im’s at the moment for work
22:11 shaggy_surfer joined #salt
22:11 emostar dfinn: yes, it complicates it a bit, but seemed to be the best tradeoff. not sure how local states would work though.
22:11 Luke_ joined #salt
22:13 yomilk joined #salt
22:15 ingwaem mgarfias: does the sync_after_install: grains matter on placement? in the first server you have it after the minion: grains but in the 2nd it’s before the minion: grains. I don’t use grains much…been starting to use pillars for sensitive data just very recently
22:17 ingwaem mgarfias: When you request the grains from those minions, do you see any other grains at all, such as serialnumber, cpu_flags, etc?
22:17 gildegoma joined #salt
22:17 mgw joined #salt
22:18 dude051 joined #salt
22:18 Eugene And for today's edition of "Other Configuration management programs used by a job I'm employing for": bcfg2
22:20 mgw joined #salt
22:21 ingwaem YACMT?
22:22 __number5__ dfinn: you can setup your minions to run in masterless mode http://docs.saltstack.com/en/latest/topics/tutorials/quickstart.html
22:23 dfinn that's actually something I'm just digging into
22:23 dfinn out of all the options I've seen/heard so far, that may be the best
22:23 dfinn certainly the most secure
22:23 __number5__ and just use rsync to update the states on the minions
22:24 dfinn i was thinking a git pull from each minion but I'll consider rsync as well
22:24 __number5__ yep, both should work
22:25 dfinn well, this is still less than ideal though
22:25 dfinn this way we get config management but we lose out on a lot of the other cool features of salt
22:25 dfinn like being able to quickly and easily update packages
22:25 dfinn or run a command against a bunch of dmz nodes
22:26 __number5__ you can still use salt-ssh to do that
22:26 oz_akan_ joined #salt
22:28 bmcorser i have maybe a silly question/desire
22:28 bmcorser i want to indent blocks of yaml in my jinja2 templated states
22:29 bmcorser things like with and for
22:29 dstokes bmcorser: can't do that
22:29 bmcorser but that will screw with the yaml right?
22:29 dstokes yaml doesn't like it
22:29 bmcorser aye
22:29 krow joined #salt
22:29 dstokes you can indent jinja statements, but you gotta preserve yaml indentation
22:29 dstokes it sucks
22:30 Eugene s/employing/applying
22:35 yomilk joined #salt
22:36 druonysus joined #salt
22:37 rowleyaj joined #salt
22:37 mateoconfeugo hello
22:37 oz_akan_ joined #salt
22:43 oz_akan_ joined #salt
22:43 azylman joined #salt
22:43 mosen joined #salt
22:46 possibilities joined #salt
22:47 bmcorser sup mateoconfeugo
22:47 oz_akan__ joined #salt
22:47 mateoconfeugo very sorry wrong window long live salt!
22:48 mosen hiya saltines
22:50 * Eugene gets the cheese and meats
22:51 bensons heya, i am using import_yaml in jinja to import some additional yaml files. no i wonder if it is possible to somehow check if this file is available or probably provide a default if it does not exist. something like that possible?
22:51 mosen afterwards we have Filet salt-minion
22:51 azylman_ joined #salt
22:52 Luke_ joined #salt
22:53 ingwaem mosen rofl, could do a play of piller in there :) a phillet of salt-minion
22:54 ndrei joined #salt
22:54 ndrei_ joined #salt
22:57 mosen hehe
22:58 Outlander joined #salt
23:00 schimmy joined #salt
23:03 druonysuse joined #salt
23:08 krow joined #salt
23:19 bmcorser if i put pillar data in a directory, does that get flattened namespacewise too?
23:19 bmcorser ie pillar/coolstuff/thing.sls
23:19 fhond left #salt
23:20 krow joined #salt
23:20 bmcorser https://www.irccloud.com/pastebin/hb3NfuNx
23:20 bmcorser will yo get flattened or will it be coolstuff:yo?
23:23 elfixit joined #salt
23:24 Eureka Anyone have a good example on how I can get data from minions using grains then update a config file with jinja templates?
23:25 UtahDave bmcorser: Yeah, it will get flattened.  It gets added to the pillar dictionary
23:25 bmcorser gotcha
23:25 bhosmer_ joined #salt
23:27 UtahDave Eureka: in your config file you can just  add   {{ grains['os_family'] }}   or whatever grain you want.
23:27 UtahDave just be sure to add  - template: jinja to your state in your sls file
23:28 Eureka UtahDave: Thanks!
23:29 UtahDave Eureka: look at the file.managed state in the docs for some examples.
23:29 JordanTesting Anyone done any kind of tuning/hacking to make the salt-minion use less memory?
23:30 UtahDave We're always open to ways to improve memory usage.  Do you have any suggestions, JordanTesting?
23:31 JordanTesting Nope, had not dug into it at all, figured I would ask around first - I seem to find all of my original ideas are always taken and better implemented these days ;)
23:31 UtahDave he he.  ;)
23:33 JordanTesting before cracking it open, my only thought would be a helper service. An ultralight socket that waits for a ping to wake up the full service and start a job
23:34 Hell_Lap joined #salt
23:36 bhosmer joined #salt
23:36 Eugene Step 1 is usually to buy a box with more than 256mb of ram. Step 2 is drink
23:37 forrest whiteinge, how are we looking on numbers for the doc sprint so far?
23:37 notpeter_ I've got salt-cloud going fully spinning up my instance with rackspace and getting all keyed up with the master, but I still have to manually run the state.highstate after that's done.
23:37 JordanTesting Eugene: swarms of little vms > one big vm
23:37 notpeter_ Does anyone else have "start_action: state.highstate" working?
23:38 whiteinge forrest: 10-ish maybe?
23:38 whiteinge i need to do a real head-count
23:38 forrest whiteinge, awww shiiit
23:38 Eugene I'm going to the Seattle one! probably
23:38 forrest yea I saw that you registered Eugune
23:38 Eugene if I get a job this week I'll have to bail
23:38 forrest *Eugene
23:39 forrest they won't let you go?
23:39 forrest it's probably just going to be me sitting in our conference room alone
23:40 Eugene Depends how quick they bring me online. I've yet to get a solid offer from anywhere. :-/
23:40 forrest ahh
23:40 Eugene On the bright side, the most likely joint has "rebuild infrastructure with configuration management tool" in the job description
23:41 forrest nice
23:43 bhosmer joined #salt
23:51 oz_akan_ joined #salt
23:52 ingwaem whiteinge: are you aware of any sample code that would interface a returner to mysql? Ultimately I woudl like a record created for every job that salt performs, and then the status along with return data updated to the record on completion
23:53 ingwaem woudl returners even be the right way to do it or perhaps events?
23:59 thayne_ joined #salt

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