Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-08-02

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

All times shown according to UTC.

Time Nick Message
00:01 mikecb left #salt
00:03 diegows joined #salt
00:09 auser joined #salt
00:13 kermit joined #salt
00:17 [diecast] joined #salt
00:17 [diecast] joined #salt
00:18 hazzadous left #salt
00:22 eculver joined #salt
00:30 dthom91 joined #salt
00:37 jimallman joined #salt
00:41 andrew joined #salt
00:48 kstaken joined #salt
00:50 hellome joined #salt
00:55 Yulli joined #salt
01:06 dthom91 joined #salt
01:09 __Jahkeup__ joined #salt
01:28 jeffasinger joined #salt
01:37 Sypher joined #salt
01:38 Furao joined #salt
01:41 brianhicks joined #salt
01:47 dN0t joined #salt
01:47 dN0t Taking salt for my first spin - not sure why salt installs vim.tiny on Ubuntu?
01:54 cxz dN0t: did you just have vim: pkg.installed
01:54 cxz ?
01:54 mgw joined #salt
01:55 trigoman joined #salt
01:55 jslatts joined #salt
01:56 dN0t cxz: Affirm.
01:56 cxz interesting
01:56 cxz i'm guessing it's part of the vim metapackage?
01:56 cxz let me try it myself manually
01:57 kermit joined #salt
01:58 cxz dN0t: yeh not sure, strange
02:00 FL1SK joined #salt
02:00 trigoman Is there any documentation on setting proxies?
02:01 trigoman I had this issue at work today and it was a total show stopper.
02:02 cxz trigoman: as far as i know there is no proxy support for salt
02:02 cxz you could try using tsocks with salt
02:03 cxz that will let you LD_PRELOAD and hook the connect() calls for salt using SOCKS5
02:03 trigoman How does salt download the packages for the guest machine? Does it rely on the guest itself?
02:03 cxz yes
02:03 trigoman If there was a way to set up the proxies on the guest, then apt-get would work.
02:03 cxz the salt-master only sends states
02:04 cxz oh i think i know what you mean
02:04 trigoman I could cerate a machine with vagrant
02:04 cxz i thought you meant how can i use a proxy with salt-minion
02:04 cxz what you could do is set a system wide proxy
02:04 trigoman but I had to go in there manually and set the proxies on the machine
02:04 kstaken joined #salt
02:04 trigoman then apt-get would work
02:04 cxz http://xmodulo.com/2012/05/how-to-set-system-wide-environment.html
02:05 trigoman Right, I know how to do this, but is there any way that I could make salt do this for me?
02:05 cxz yeh, so you'll need some way to bootstrap salt, and maybe run a local salt state that does that for you
02:05 cxz i don't think so :<
02:05 trigoman I'm digging through teh documentation
02:07 LucasCozy joined #salt
02:08 dN0t joined #salt
02:09 trigoman joined #salt
02:15 liuyq joined #salt
02:16 raydeo joined #salt
02:17 trigoman joined #salt
02:18 syngin joined #salt
02:21 dN0t @cxz: Doesn't seem like it installs vim by default.  Is there a way to be specific?
02:22 dave_den trigoman: you could certainly have salt set the apt proxy for you.
02:26 dave_den dN0t: what does your state file look like?
02:26 jessep joined #salt
02:28 KyleG joined #salt
02:28 KyleG joined #salt
02:29 dN0t dave_den: https://gist.github.com/RobSpectre/6137085
02:32 trigoman joined #salt
02:33 trigoman so you're saying I can configure it so that apt-get uses a proxy?!?!?! that would be so swell
02:36 tsheibar joined #salt
02:39 dave_den trigoman: if you want to set a system-wide proxy like cxz suggested you can, or you can set an apt only proxy in apt.conf or apt.conf.d if it exists. e.g.  Acquire::http::Proxy "http://myproxy.com:80";
02:41 dave_den dN0t: with a state of only vim\n:  pkg:\n    -installed, it results in intstalling vim-runtime, editor, and vim packages on my ubuntu 12.04 system
02:43 dave_den dN0t: also, you have a circular dependency in your vim-deps and vim state functions. you only need to require vim-deps from vim.pkg or require_in from vim-deps to pkg.vim - not both
02:44 trigoman dave_den: thank you! I'll try it out
02:51 oz_akan_ joined #salt
03:02 cxz trigoman: yes, you can use an apt-proxy
03:02 cxz very straightforward
03:02 Yulli left #salt
03:03 capricorn_1 joined #salt
03:06 raydeo joined #salt
03:09 aat joined #salt
03:09 syngin does jinja support "if item in iter" ?
03:22 syngin If I wanted to identify if a minion is a Linux box, should I just use the "kernel" grain?
03:28 cxz syngin: do the latter
03:28 cxz but yes jinja does support if item in list
03:32 liuyq joined #salt
03:34 syngin cxz: thanks. i want to use "item in list" for something else (orthogonal roles).
03:39 z0rkito is there a way to get pillar data from one pillar to another using #!py?
03:39 xl1 joined #salt
03:53 __Jahkeup__ joined #salt
03:57 syngin z0rkito: one pillar to another?
03:58 syngin z0rkito: like, minion A's pillar data over to minion B's pillar data?
03:59 z0rkito no, for the minion.
04:01 z0rkito ie: for node 1, pillar x = asdf.  for node 2, pillar x = fdsa.  for pillar y, node 1 needs to return qwert and node two trewq.
04:01 cxz it sounds like you want to use the salt mine maybe
04:01 cxz ?
04:01 cxz generally pillar data for a node should be the same for that node
04:01 cxz and not change
04:02 cxz if you need to change pillar data you should be using the salt mine for data which does change
04:02 cxz kind of the same thing for grains
04:02 cxz grains shouldn't be changing
04:02 cxz (even though i believe you can change them)
04:02 cxz someone should write up 'The Salt Philosophy'
04:02 z0rkito right that makes sense.  i dont want it to change.
04:02 cxz good :)
04:02 z0rkito i want it to be different for node1 and node1.
04:03 z0rkito basically i have a json file with a shit ton of things in it, broken up by server type.  i have a pillar that calls to an api to get the server type.  i want to then use that pillar without another api call to pull the data out of the json file.
04:05 cxz the salt mine will let you change the data at runtime
04:05 cxz and is generally used for that kind of thing
04:05 cxz the documentation sucks in general
04:05 cxz but i think that is what you want
04:06 syngin what do ya'll reckon about using /etc/salt/minion to store grains (as opposed to /etc/salt/grains)?
04:07 syngin just thinking about managing the minion's config from my master.
04:10 Lue_4911 joined #salt
04:10 StDiluted joined #salt
04:14 jessep joined #salt
04:14 middleman_ joined #salt
04:37 trigoman cxz: Yes, I agree with you, rep 'The Salt Philosophy'
04:41 Ryan_Lane joined #salt
04:42 raydeo joined #salt
04:50 akoumjian trigoman: guessing you figured it out by now, but it was getting stuck at either downloading the bootstrap script, or downloading the salt package from your guest os distribution
04:51 trigoman akoumjian: is there a verbose mode?
04:51 trigoman akoumjian: I would really like to use salty-vagrant, but I need to know where to set the proxy.
04:52 akoumjian trigoman: Yep. https://github.com/saltstack/salty-vagrant#other Just set salt.verbose = true
04:52 akoumjian trigoman: Well, if you're using Vagrant on top of virtualbox, I would search there. This has more to do with your local networking and virtualbox than anything else
04:53 trigoman Working behind a proxy always makes life hard.
04:53 akoumjian trigoman: Are you at work or something?
04:53 trigoman not right now, but I was this morning
04:54 trigoman I wanted to try out salt+vagrant to get a tomcat vm running
04:54 trigoman figured it was an excuse to try the saltstack
04:54 akoumjian trigoman: It will be beneficial for you to figure out how to get your guest vm network access through your proxy for more than just using saltstack. :-)
04:55 trigoman Yeah it's always something different for every tool.
04:55 trigoman gom
04:55 trigoman gem*
04:55 trigoman pip
04:55 trigoman brew
04:56 trigoman npm
04:56 trigoman you name it
04:56 trigoman Usually http_proxy get's it, but there are times where nothing works.
04:57 syngin hmm, can jinja call an external command?
04:58 trigoman akoumjian: I'll try work that issue out tomorrow, so you reckon this is an issue with vbox rather than salt?
04:58 __Jahkeup__ joined #salt
05:04 akoumjian trigoman: Yes. salty-vagrant detects if salt is installed. If it isn't, it installs it via the salt-bootstrap script. If the vm can't talk to the internet, that will definitely put a stop to things
05:05 malinoff joined #salt
05:06 UtahDave joined #salt
05:07 trigoman akoumjian: That's another issue I came upon, it said that I didn't have salt installed, but I installed it using brew, so it will not find it in the usual location /usr/bin or w/e that place is, is that normal?
05:07 akoumjian trigoman: You installed it on your host machine or your guest machine?
05:08 trigoman host
05:08 auser joined #salt
05:08 akoumjian trigoman: salty-vagrant operates on virtual machines. These are entirely isolated virtual environments from your host system. It doesn't matter what you have installed on your mac, your VM will boot with whatever image you give it.
05:09 trigoman akoumjian: I was trying to do a master-less install, so salty-vagrant will fire up the vm and install salt on that machine? I'm sorry if I sound so noobish, still getting the hang of it.
05:12 akoumjian trigoman: Yes, that's exactly it. It will install the salt-minion package, which has the salt-call command on it. You use ``salt-call`` instead of ``salt`` to issue commands on a masterless setup.
05:12 trigoman akoumjian: that makes much much much more sense
05:25 Nitron joined #salt
05:25 knightsamar joined #salt
05:30 kstaken joined #salt
05:30 pcarrier joined #salt
05:45 aat joined #salt
05:48 neilf joined #salt
05:49 auser joined #salt
05:57 jshare joined #salt
05:58 __Jahkeup__ joined #salt
06:02 bud joined #salt
06:02 bud Morning everyone.
06:03 carlos joined #salt
06:03 auser hey bud
06:04 auser <~ I love saying that
06:05 bemehow joined #salt
06:10 syngin hmm, how can i restart the salt-minion service during highstate?
06:13 Furao syngin: why you want to do that? you can use module.run: - run - name: salt-minion
06:13 Furao opus ignore the - run
06:14 Furao damn auto corrector
06:14 Furao so it's module.run: - name: service.restart -m_name: salt-minion
06:16 syngin Furao: it's linked to my grains files (which are managed).
06:16 neilf joined #salt
06:17 Furao then use service.running: - name: salt-minion - watch: - $modnamethatchanegrains: $statenamethatchangegrains
06:17 Furao it will restart salt-minon only if something change
06:17 akitada joined #salt
06:17 Furao restart the minion all the time without good reasons is not a good idea
06:18 syngin Oh, that's how I have it set up (it watches /etc/salt/grains), but it always fails to restart.
06:18 gamingrobot joined #salt
06:18 syngin Comment: Failed to restart the service
06:19 Furao so there is an error in grains file or elsewhere
06:19 HashNuke joined #salt
06:19 Furao why grains?
06:19 syngin nup, works fine the second time i run highstate and grains.items returns the correct grains
06:19 Furao pillars are far better
06:19 z0rkito anyone know of there's a way to add a multi line sources.list using pkgrepo.managed?
06:20 Furao the only reasons I have custom grains is to get data from Amazon EC2
06:20 syngin Furao: my grains are just what you'd expect (datacenter, deploymode, etc)
06:20 syngin Furao: I use pillar for all the normal config stuff.
06:20 Furao ok, that's one of the only good reasons to use them
06:20 syngin Furao: The only reason I'm seeing problems is because I'm redesigning my layout so my grains are changing slightly.
06:21 syngin Furao: I'll probably just remove them from salt as once they're on the minion, they don't change.
06:21 Furao I was afraid you were like those guys who cram everything in grains and wonder why it's complicated :)
06:22 syngin haha
06:22 syngin not so much
06:22 Furao don't laugh, few of them do that, until the switch to pillars (after tears and blood)
06:23 syngin oh man, that would hurt.
06:24 Furao the most frequent excuse is "no time to learn pillars"
06:24 syngin o_O
06:24 Furao I drive my car, but I can't park backward… no time to learn that.
06:25 bud I am curios what happens if you restart the minion during a highstate run, would the minion continue the execution of the state where it left it?
06:25 bemehow joined #salt
06:25 Furao bud: when you run something on a minion, it spawn a 2nd process
06:25 Furao the 2nd process run the logic
06:26 Furao if you restart the minion it only restart the "main" one and not the process that apply the logic
06:26 bud Ok, I get it, thanks Furao.
06:26 Furao so, let's say you have, like me, salt-minion state, and you upgrade from salt 0.15.0 to 0.16.0
06:26 Furao you ends with a running process with salt 0.15.0 code in memory
06:27 Furao but with code on filesystem for 0.16
06:27 bud And next time the process runs you're on 0.16.
06:27 Furao and everything fail :)
06:27 bud :))
06:28 bauser joined #salt
06:41 carmony joined #salt
06:53 octarine joined #salt
06:55 cowmix joined #salt
07:00 cowmix joined #salt
07:08 ml_1 joined #salt
07:24 balboah joined #salt
07:38 iMil morning there
07:38 bud Morning iMil.
07:38 iMil I've posted a question on the mailing list, not sure if it's a bug or something I don't do correctly
07:38 iMil hey bud
07:39 iMil I'd like to pass a dynamic hash (**kwargs) from a state file to a module
07:39 iMil no matter what, the result is always empty
07:39 bud A question about grains vs. pillar. If I want a server to have OpenLDAP installed is it better to set a grains role of openldap-master or do it via pillar, somehow?
07:39 iMil but if I declare a simple hash, it actually works
07:40 syngin bud: That sounds like a grains thing to me - I define my roles in grains.
07:40 iMil bud: I have a ldap-master role
07:40 iMil yup
07:40 iMil role ldap or ldap-master
07:40 iMil that's what I use
07:40 syngin Basically - if it's meta data about the server that doesn't change often: grains.
07:40 bud That's what I've thought too. I have asked this question because of your discussion and "not enough time to learn pillar" thing :).
07:41 syngin bud: ahh, yeah, pillar is a weird one.
07:41 iMil well, pillars are nothing more but "metadata"
07:41 syngin bud: pillar is, from my perspective, malleable config data - users and the like.
07:41 bud I store in pillar everything that's config and global var data as the docs suggest.
07:41 iMil really handy for parameters for example
07:46 mike25ro joined #salt
07:47 mike25ro mornin' people
07:47 bud Morning mike25ro.
07:51 Ixan Is there some practical way to split the top.sls file into separate files for easier management? Some form of hierarchy would be prefered, if possible
07:52 mike25ro Ixan: i think you can extend
07:53 mike25ro but again... i am just reading myself now... so not really sure HOW to... but i remember that i saw some extend functionality
07:53 Ixan will every sls-file be compiled though? so if a make a foo.sls with extend: base: it'll magically work?
07:53 mike25ro good question... not sure... maybe one of the more experienced guys can tell us
07:53 mike25ro you can try though...
07:57 bud Does anyone have OpenLDAP cn=config data stored in pillar. I could use some pointers on how to arrange it...
08:00 Xeago joined #salt
08:02 kstaken joined #salt
08:07 scalability-junk joined #salt
08:08 xerxas joined #salt
08:08 andrewclegg joined #salt
08:09 LucasCozy joined #salt
08:09 ninkotech joined #salt
08:10 pcarrier joined #salt
08:10 blast_hardcheese joined #salt
08:10 ada__ joined #salt
08:10 dave_den joined #salt
08:11 jrgifford joined #salt
08:13 zooz joined #salt
08:13 costi joined #salt
08:13 bud * throws costi a cookie
08:15 chubrub joined #salt
08:15 kstaken_ joined #salt
08:17 krissaxton joined #salt
08:19 chubrub hey everyone! I have a question regarding overstate.sls. If I run highstate from master, does the overstate file is included, or does it happen only when calling over.state function?
08:23 sw_ joined #salt
08:26 neilf joined #salt
08:32 al5d joined #salt
08:33 iMil well it appears that was a bug, ze- proposed a fix here https://github.com/saltstack/salt/pull/6494/files
08:33 iMil ze-: thanks :)
08:33 dthom91 joined #salt
08:34 krak3n` joined #salt
08:38 unicoletti_ joined #salt
08:45 SEJeff_work joined #salt
08:51 scalability-junk joined #salt
08:51 Jason-AVST joined #salt
08:53 bemehow joined #salt
08:55 bud I went with structuring everything under openldap-server:cn dict in pillar for OpenLDAP configuration.
08:56 bud I have posted a gist - https://gist.github.com/valentinbud/6138463. Maybe others will find it useful :).
09:04 salt_ali joined #salt
09:04 salt_ali hello
09:05 salt_ali I'm trying to configure salt to restart a service, but only if config file is changed, is that possible?
09:05 iMil salt_ali: with the 'watch' keyboard I beleive
09:06 iMil never actually did it but i've seen some states doing that
09:06 salt_ali hm, makes sense
09:07 iMil keyWORD
09:07 iMil wow I'm not drunk I swear
09:07 salt_ali :)
09:07 cowmix joined #salt
09:08 dthom91 joined #salt
09:08 whiskybar joined #salt
09:10 zooz joined #salt
09:10 felixhummel joined #salt
09:13 chubrub here is my top.sls file http://pastebin.com/DWJZPgRX - I'f like to be sure that when I bootstrap 'web' machine, all states from 'base' will be aplied first, then specific states for role will be aplied. Is there a simple way to make such dependency in salt?
09:13 krak3n` joined #salt
09:15 felixhummel_ joined #salt
09:18 mike25ro you can have pillars for each machine
09:18 mike25ro this is how i would do it - but i am just a newbie here
09:18 mike25ro for each machine set up a pillar and you can have roles in there ...
09:19 oz_akan_ joined #salt
09:19 mike25ro saltroles:
09:19 mike25ro - database
09:19 mike25ro - webserver
09:19 mike25ro - php
09:19 mike25ro this is just a test i am doing
09:19 mike25ro and the pillar is called salt23.sls... then ... in my top.sls pillar i have
09:19 mike25ro base:
09:19 mike25ro '*':
09:19 mike25ro - data
09:19 mike25ro 'salt2':
09:19 mike25ro - salt23
09:20 mike25ro so my machine is called salt2 and has the pillar from salt23
09:20 mike25ro then in your states you can use pillar functions to match the pillars / roles for example
09:20 mike25ro again - not sure if it is the best solution
09:22 mike25ro but i think it adds a bit of flexibility... for example if you need to edit some things on your web-05 (add a new role) is really easy to add it as a pillar and then from states just apply it
09:22 bud I, for what is worth, I do the matching in pillar based on roles not minion ID.
09:23 mike25ro in my opinion... pillars are great because they add flexibility to a setup.
09:23 bud So all the MySQL database servers load the mysql pillar.
09:23 mike25ro indeed bud chubrub
09:25 chubrub Actualy, I don't have problem with matching states to roles - your method is ok. My problem concerns order of executing states. In my case I have around 20 states to execute, to make minion functional. Require requisite is fine for small sets, but it would be perfect if I could make a list of states that are executed in sequence, or nesting dependencies, grouping them and use require for smaller sets
09:27 mike25ro so you need sequential states to run ... in a predefined order/
09:27 mike25ro hmm ... not sure .. if is possible :D
09:28 chubrub salt is really awesome, but lack of such feature makes it really painful - most of time I spent configuring my minion is related to sequence of executing states :/
09:28 mike25ro why is it so painful? what are you installing that needs such a feature?
09:30 lyddonb_ joined #salt
09:33 chubrub first I'm preparing base system: apt, firewall, base-packages, net hardening, etc. After that I'm going to biuld web env, app and postgres. All stated has to be executed in specific order (i.e. firewall has to be stared before nginx, apt sources and preferences has to be installed before everything else, postgres has to configured before building and app, etc)
09:34 chubrub so taking into consideration fact, that all states files has also 'internal' requires it's not so simple to me to manage seqence of executing it :/
09:39 chubrub overstate function might be the answer, but unfortunatelly it's not working for me as expected. Don't get it what exactly means your_env in this call: salt-call state.over your_env; And I don't know how highstate is related to overstate as well :/
09:40 lemao joined #salt
09:41 helderco joined #salt
09:54 krissaxton joined #salt
10:00 pnl joined #salt
10:06 giantlock joined #salt
10:08 pnl joined #salt
10:10 jfalco joined #salt
10:10 jfalco joined #salt
10:11 g3cko joined #salt
10:13 bemehow joined #salt
10:15 zooz joined #salt
10:15 bejer Hi, how do I fetch nested pillar data? (I can do something like pillar.get('core', {}).items() for iterating, but what if I need something under core:subcore:subsubcore - it doesn't work by using that particular string)?
10:20 ada__ joined #salt
10:20 bemehow joined #salt
10:23 octarine joined #salt
10:28 ProT-0-TypE joined #salt
10:30 __Jahkeup__ joined #salt
10:32 bejer Oh I can use pillar[][][] indexing, makes sense now
10:44 drawsmcgraw joined #salt
10:47 Linz_ joined #salt
10:47 eculver joined #salt
10:47 eculver joined #salt
10:50 drawsmcgraw Under the hood, how does the Salt master identify/reach a given minion?
10:50 drawsmcgraw IP? FQDN?
10:53 bemehow_ joined #salt
10:53 sw_ drawsmcgraw: the minion keeps a connection to the master
10:54 s0undt3ch_ joined #salt
10:55 krissaxton joined #salt
10:56 drawsmcgraw sw_: so... when I issue a Salt command from the master
10:57 sw_ drawsmcgraw: it identifies the minion on the ID
10:57 sw_ which by default should be the FQDN, but you can override in minion config
10:57 drawsmcgraw Fair enough
10:57 drawsmcgraw And yeah, in my dev lab
10:57 drawsmcgraw I have things lilke "minion-01" and "minion-02"
10:57 drawsmcgraw Which are meaningless on the network
10:58 sw_ drawsmcgraw: what does "hostname -f" returns on those ?
10:58 drawsmcgraw as in -> hostname -f minion-01 ?
10:58 sw_ no, just hostname -f
10:58 sw_ (should return the fqdn)
10:59 drawsmcgraw sw_: Ah, I see
10:59 drawsmcgraw yes. The FQDN
10:59 drawsmcgraw TIL another way to use 'hostname'
10:59 drawsmcgraw incidentally, I *did* try -> hostname -f minion-01
10:59 drawsmcgraw from the Master
10:59 drawsmcgraw and I got back the FQDN
11:00 drawsmcgraw Clearly, more to learn on my part. Thanks sw_  !
11:00 sw_ drawsmcgraw: you're welcome. Keep digging Salt is awesome :)
11:00 drawsmcgraw sw_: unrelated -> Are you well-versed with the 'watch' requisite?
11:01 drawsmcgraw Oh I am :)
11:01 sw_ drawsmcgraw: i started using salt two months ago, and starting to learn python.. so i'm a noob compared to most of the ppl here.. anyway, yeah, I use watch extensively
11:01 drawsmcgraw neat! Feel like assisting, then ?
11:02 sw_ sure..
11:02 mikedawson joined #salt
11:02 sw_ but i'm level 1 support :)
11:02 felixhummel_ hi! how can i show errors only on state.highstate? my output is a little too long...
11:02 sw_ state_verbose: False in your master conf
11:03 sw_ felixhummel_: that was for you :)
11:03 drawsmcgraw sw_: Fair enough :)
11:03 drawsmcgraw sw_: http://pastebin.com/yaUdg3Gq
11:03 drawsmcgraw In that example, 'watch-for-new-file' executes regardless of the outcome of 'drop-watch-file'
11:04 drawsmcgraw I'm looking at the yaml that a salt command returns, and 'drop-watch-file' returns nothing for 'changes'
11:04 felixhummel_ thanks, sw_ :)
11:04 sw_ does drop-watch-file get updated at earch highstate ?
11:04 drawsmcgraw not sure I'm smart enough to understand that question....
11:04 drawsmcgraw Right now I'm testing with state.sls
11:04 drawsmcgraw And just running that state file on its own
11:05 sw_ actually i've never used a watch statement on file.managed.. only service.running
11:05 drawsmcgraw Interesting....
11:05 drawsmcgraw My understanding is -> what it looks for is if anything comes back in the 'changes' field for a given state dec
11:05 sw_ btw, watch-for-new-file is missing the source
11:06 drawsmcgraw Fair enough, thanks :)
11:06 drawsmcgraw It just drops an empty file in that case
11:06 sw_ "watch" will watch for changes aka file changed or created.. and then trigger the action
11:06 drawsmcgraw Right
11:06 drawsmcgraw So if I run that state file twice, then I should only see 'drop-watch-file' run, correct?
11:07 drawsmcgraw But I always see both state decs run
11:07 sw_ let me look up file.managed doc
11:07 drawsmcgraw i.e. I've deleted '/tmp/reacted.txt' and then run the state again
11:07 drawsmcgraw k
11:07 hazzadous joined #salt
11:07 sw_ drawsmcgraw: alright, no watch statement possible on file.managed
11:08 drawsmcgraw oh??
11:08 yota joined #salt
11:08 sw_ try with cmd.run or service.running
11:08 drawsmcgraw Will do
11:08 sw_ look up "onlyif" and "unless" with cmd.run
11:08 sw_ service.running works with watch
11:09 drawsmcgraw Where'd you find that watch doesn't work for file.managed ?
11:09 isomorphic joined #salt
11:09 Ivo joined #salt
11:10 sw_ drawsmcgraw: http://docs.saltstack.com/ref/states/all/salt.states.file.html
11:12 logix812 joined #salt
11:12 david_a joined #salt
11:17 bhosmer joined #salt
11:17 tomeff joined #salt
11:17 drawsmcgraw sw_: I'm real sorry... I..... can't find where it mentions that
11:17 Linz joined #salt
11:18 drawsmcgraw I believe you. I'm just trying to find where to look next time I run across something like this.
11:28 sw_ drawsmcgraw: where is mentions what ?
11:29 drawsmcgraw That the watch statement can't be used on file.managed
11:29 yota joined #salt
11:29 sw_ salt.states.file.managed(name, source=None, source_hash='', user=None, group=None, mode=None, template=None, makedirs=False, context=None, replace=True, defaults=None, env=None, backup='', show_diff=True, create=True, contents=None, **kwargs)¶
11:30 sw_ these are the options available for file.managed
11:30 sw_ and the defaults
11:30 drawsmcgraw sw_: Ah, I see. I was thinking that 'watch' was something of a common parameter among the modules.
11:30 drawsmcgraw Thanks!
11:30 sw_ drawsmcgraw: yw :)
11:30 __Jahkeup__ joined #salt
11:32 tomeff joined #salt
11:32 felixhummel_ i think file can watch
11:32 felixhummel_ you have **kwargs (meaning any key/value pair)
11:33 felixhummel_ file watch example: https://gist.github.com/felixhummel/6139236
11:33 ggoZ joined #salt
11:35 mike25ro felixhummel_:  in your example ... every time /tmp/y is changed .. on minions /tmp/x is copied.. right?
11:36 felixhummel_ yeah, that's the basic idea
11:37 felixhummel_ i just wanted to see if file.manage accepts a watch parameter or throws an error. it was green/blue when I ran it with salt-call --local state.sls filewatch
11:38 mike25ro cool
11:44 aleszoulek joined #salt
11:44 drawsmcgraw felixhummel_: Interesting... In my tests I don't get an errror, it just seems to silently ignore the 'watch' requisite
11:45 mike25ro indeed
11:45 mike25ro i tested it right now
11:45 mike25ro root@salt1 /srv/salt# salt 'salt2' state.sls filewatch.sls
11:45 mike25ro salt2:
11:45 drawsmcgraw mike25ro: You're getting the 'silently ignore' treatment as well?
11:45 krissaxton joined #salt
11:46 mike25ro drawsmcgraw: indeed
11:46 mike25ro /tmp/x:
11:46 mike25ro file.managed:
11:46 mike25ro - source: salt://x
11:46 mike25ro - watch:
11:46 mike25ro - file: /tmp/y
11:46 mike25ro /tmp/y:
11:46 mike25ro file.touch
11:46 mike25ro this is the content of filewatch.sls ...
11:47 drawsmcgraw Tragic
11:47 drawsmcgraw Well, at least I know where the boundary is :) I can come up with another solution I'm sure
11:47 mike25ro hmm but why did it work for felixhummel_ ?
11:48 drawsmcgraw felixhummel_: Did you test your example?
11:48 * drawsmcgraw looks suspiciously towards felixhummel_
11:48 mike25ro :)
11:52 mike25ro drawsmcgraw: are we doing sometjhing wrong?/...
11:54 tomeff_ joined #salt
11:59 ml_2 joined #salt
11:59 drawsmcgraw Dunno. I'm working on a solution that uses cmd.wait instead now
11:59 drawsmcgraw mike25ro: ping
12:00 mike25ro hmm
12:01 mike25ro i am wondering why the heck isn't that working... i am a newbie.. so i take this in consideration... maybe i didn't do smth right
12:01 sw_ felixhummel_: did you test your state ?
12:02 mike25ro sw_:  i did test it now... and is now working... drawsmcgraw was testing it too
12:03 waverider joined #salt
12:04 drawsmcgraw mike25ro: You mean you can use 'watch' for file.manage and it's no longer silently ignoring it?
12:05 mike25ro drawsmcgraw:  i can not use watch... actually the state isn't working even without the watch... i am surely doing smth wrong
12:05 drawsmcgraw Today... is not a good day, for Science
12:05 waverider left #salt
12:06 mike25ro indeed
12:06 drawsmcgraw Well, on a greedy note. I think my cmd.wait solution will work for me. I'm just making a note about 'watch' and file.managed
12:14 LGSilva joined #salt
12:16 Teknix joined #salt
12:17 blee joined #salt
12:21 krak3n` joined #salt
12:22 middleman_ joined #salt
12:22 __Jahkeup__ joined #salt
12:22 anteaya joined #salt
12:24 diegows joined #salt
12:24 helderco joined #salt
12:32 jkleckner joined #salt
12:36 jslatts joined #salt
12:37 * felixhummel_ tested the state, but not if watch kicked in
12:39 felixhummel_ edited file x, /tmp/y got touched
12:39 felixhummel_ drawsmcgraw: still looking suspiciously?
12:42 blink__ joined #salt
12:43 mike25ro felixhummel_:  i am a bit confused :) still trying to understand some concepts
12:43 mike25ro felixhummel_:  ... you want to copy x from master to minion ONLY if the y is changed or the other way around?
12:45 bhosmer_ joined #salt
12:45 KennethWilke joined #salt
12:46 carlos joined #salt
12:49 bhosmer joined #salt
12:49 knightsamar left #salt
12:52 Nitron left #salt
12:52 felixhummel joined #salt
12:54 drawsmcgraw felixhummel: Interesting.... I'll have to look more into this sometime
12:55 drawsmcgraw mike25ro: Your understanding is the same as mine
12:58 mike25ro i am testing this again ...
12:59 scofflaw joined #salt
12:59 mike25ro drawsmcgraw:  felixhummel  i changed the x file... added some data... run the update... but y is not touched.
13:02 bhosmer_ joined #salt
13:07 brianhicks joined #salt
13:11 juicer2 joined #salt
13:18 felixhummel mike25ro: weird. works for me. i'm on 0.15.1
13:18 mike25ro i am on 0.16.0
13:19 mike25ro watch command... is not suppose to support any state...
13:19 mike25ro it may be that ...
13:19 JasonSwindle joined #salt
13:20 jeffasinger joined #salt
13:23 aat joined #salt
13:24 mgw joined #salt
13:27 mike25ro drawsmcgraw: still around?
13:27 mike25ro drawsmcgraw:  i wanted to follow your idea with cmd.wait
13:27 Gifflen joined #salt
13:28 fredvd joined #salt
13:29 drawsmcgraw mike25ro: What's up?
13:30 drawsmcgraw Ah
13:30 drawsmcgraw So, my case is -> I need to drop a tarball and unpack it in a directory
13:30 StDiluted joined #salt
13:31 drawsmcgraw I wanted to remove the tarball when I'm done, but in the end I went with just keeping it on the Minion
13:31 drawsmcgraw so I use file.managed to drop the tarball
13:31 drawsmcgraw Then I use:
13:31 drawsmcgraw install_stuff:
13:31 drawsmcgraw -cmd.wait:
13:31 drawsmcgraw -name: tar -zxvf stuff
13:31 racooper joined #salt
13:31 drawsmcgraw -watch:
13:31 drawsmcgraw file: drop_tarball
13:32 drawsmcgraw And I have a handful of states below that. Those states 'require' the install_stuff state dec
13:32 drawsmcgraw So all the unpacking/installing only happens when the tarball is first dropped
13:32 mike25ro weird... for me is not working
13:33 mike25ro testfile:
13:33 mike25ro file:
13:33 mike25ro - managed
13:33 mike25ro - name: /tmp/x.txt
13:33 mike25ro - source: salt://x.txt
13:33 mike25ro cmd.wait:
13:33 mike25ro - name: touch y
13:33 mike25ro - cwd: /tmp/
13:33 mike25ro - watch:
13:33 mike25ro - file: testfile
13:33 racooper pastebin is so much easier (and nicer...)
13:33 KennethWilke mike25ro: https://gist.github.com/ http://pastebin.com/
13:33 drawsmcgraw racooper: agreed. Apparently I had a moment of laziness
13:34 mike25ro http://pastebin.com/gdL1B4q3
13:34 mike25ro the idea is to copy x to the minion and touch /tmp/y on the master after that copy is done
13:34 mike25ro the copy is fine... BUT no touching of y
13:36 mike25ro y is 1 hour old.. even i just run ... salt 'salt2' state.sls file-copy
13:37 mike25ro i believe the command state has some error
13:38 mike25ro salt 'salt2' state.sls file-copy
13:39 mike25ro http://pastebin.com/3V7gipi2
13:40 Gifflen joined #salt
13:40 m_george|away joined #salt
13:44 m_george left #salt
13:45 ipmb joined #salt
13:45 krissaxton joined #salt
13:45 trigoman joined #salt
13:56 gaoyang joined #salt
13:57 toastedpenguin joined #salt
13:58 StDiluted mike25ro, I believe that y will only get touched if a change is made to x.txt
13:59 mike25ro StDiluted: i am stupid
13:59 mike25ro indeed is true...
13:59 __Jahkeu_ joined #salt
13:59 mike25ro i was expecting the change to happen on master
13:59 mike25ro Y to be touched on master
13:59 StDiluted ahhh
13:59 mike25ro it works
13:59 StDiluted cool
13:59 mike25ro it works with the minion ... i wanted to copy a file on the minion and change smth on the master
14:00 mike25ro is that possible?
14:00 mike25ro the WATCH thingy didn't work ... but i used cmd.wait
14:00 felixhummel :D
14:00 StDiluted what are you trying to do exactly?
14:00 mike25ro http://pastebin.com/7RiuNrEa
14:01 mike25ro copy x to minion and touch y on the master
14:01 StDiluted understood, but why?
14:01 mike25ro good Q ... no real need.. just to see if it is possible
14:01 mike25ro :)
14:02 StDiluted oh
14:02 mike25ro i am a newbie.. and i am testing ... to see all the features.. to learn
14:02 felixhummel ok, so file.managed with "watch" works, right?
14:02 mike25ro felixhummel: didn't work with watch.. i used cmd.wait
14:02 StDiluted well, you could use a cmd.script  that used ssh to touch the file
14:02 mike25ro with watch ... it touches the y even if x is not changed...
14:03 StDiluted assuming the minion has ssh access to the master
14:03 mike25ro StDiluted: thanks buddy
14:04 StDiluted I'm not aware of other methods to make things happen on the master, haven't needed to do that
14:04 mike25ro i can imagine that :)
14:04 mike25ro i am just used to ask stupid Q :)
14:05 mnemonikk joined #salt
14:05 mike25ro felixhummel: http://pastebin.com/7RiuNrEa   this is what i have used in the end
14:06 felixhummel i see
14:06 felixhummel let us know when u need this in real life ;)
14:07 kaptk2 joined #salt
14:07 mike25ro i am sure i don't :) soon
14:08 mike25ro i am not really doing anything.. apart testing
14:10 drawsmcgraw If you need something done on the Master, you could use the Reactor system
14:10 drawsmcgraw mike25ro: ping
14:11 drawsmcgraw My favorite part about Salt is abusing the Reactor system
14:11 drawsmcgraw When you need to send a message, from any machine (including minions), just --> salt-call event.fire_master 'some text' 'your tag'
14:13 JesseC hmm, wouldn't it be possible if someone managed to compromise a minion
14:14 JesseC for them to mass compromise your entire network through salt
14:17 yota joined #salt
14:20 teskew joined #salt
14:24 ipmb Is there a reason why at the end of a highstate, salt doesn't output a success/fail message?
14:25 opapo joined #salt
14:25 Linz joined #salt
14:26 lemmings there's the copy command, I think, but that one only copies to the master's cache file
14:26 lemmings dir
14:26 lemmings sorry
14:26 dN0t joined #salt
14:27 [diecast] joined #salt
14:28 mike25ro left #salt
14:29 [diecast] joined #salt
14:32 pdayton joined #salt
14:34 chrisgilmerproj joined #salt
14:34 __Jahkeup__ joined #salt
14:35 lyddonb_ joined #salt
14:37 cnelsonsic joined #salt
14:37 mgw1 joined #salt
14:39 indymike joined #salt
14:42 JasonSwindle ipmb:  That will be in .17
14:42 ipmb nice! is there an open issue?
14:43 JasonSwindle .17 will be a BIG release from what I can see
14:43 JasonSwindle Sure, let me find it.  I was the one who created it
14:43 ipmb the mixed output makes it bearable, but I'm typically scrolling back searching for something red
14:43 JasonSwindle ipmb:  https://github.com/saltstack/salt/issues/6384
14:45 ipmb that was quick, I wish I had brought it up earlier :) https://github.com/saltstack/salt/pull/6411/files
14:45 kstaken joined #salt
14:45 JasonSwindle ipmb:  The Salt guys are awesome and fast
14:45 JasonSwindle Awesome as in; in person and online.
14:46 krissaxton joined #salt
14:48 felixhummel anyone using highstate after successful CI run?
14:48 ipmb felixhummel: I'm triggering them from Jenkins
14:48 ipmb not automatically (yet), but I have a highstate job
14:49 JaredR joined #salt
14:49 felixhummel ipmb: how exactly do you trigger them?
14:49 felixhummel ssh?
14:49 ipmb Jenkins is on my master
14:49 ipmb if it wasn't, yeah, ssh
14:49 felixhummel i see
14:49 ipmb and I use the client_acl
14:49 ipmb the *only* thing Jenkins can do is run state.highstate
14:52 JasonSwindle I need to get to that point ipmb
14:52 StDiluted joined #salt
14:52 ipmb JasonSwindle: I think it's literally a two line job
14:52 andrew joined #salt
14:53 ipmb JasonSwindle: http://cl.ly/image/45252Q2d2a3b/content.png
14:53 JasonSwindle Most likely.  Writing docs is what most of my day will be.   I got lazy, and now my README.md looks like a bomb blew up on it
14:54 kermit joined #salt
14:55 ipmb JasonSwindle: here's the client_acl part. it lives in /etc/salt/master http://cl.ly/image/3a1e04290S46/content.png
14:55 JasonSwindle Awesome!  Thank you
14:56 JasonSwindle I now need to go poke the django guys.  Jenkins is on our list. :)
14:56 UtahDave joined #salt
14:59 JaredR So when using GitFS, when does salt synchronize the remote repo?
15:00 JesseC salt master branch is stable right?
15:00 JesseC on github
15:03 JasonSwindle JesseC:  Use tags
15:03 ml_1 joined #salt
15:03 JesseC JasonSwindle: I'm still new to git, not sure what you mean.
15:03 JasonSwindle Never trust a branch; they can move
15:04 JasonSwindle A tag is label linked to a commit
15:04 JasonSwindle So, it is like the deb / rpm of git
15:04 JesseC JasonSwindle: well, the bootstrap script only works with branch names though right?
15:04 JasonSwindle Noper
15:04 JesseC and commit ids
15:04 JasonSwindle It will work with tags
15:04 JesseC JasonSwindle: Can you hook me up with an example?
15:05 JesseC I did this
15:05 JasonSwindle curl -L http://bootstrap.saltstack.org | sudo sh -s -- git v0.16.2
15:05 JesseC curl -L http://bootstrap.saltstack.org | sh -s -- git
15:05 JesseC ohh
15:05 JasonSwindle That will deploy via a tag
15:05 JesseC I see.
15:06 JasonSwindle vTAG_NAME_HERE
15:06 JesseC Ah yes and "Releases" on github are the tags
15:06 JasonSwindle Yep
15:07 JesseC Thanks for the info
15:07 JasonSwindle So, if you need .14.1; do v0.14.1
15:07 JasonSwindle No problem.
15:07 JasonSwindle JesseC:  Long read; but good
15:07 JasonSwindle https://www.atlassian.com/git/workflows
15:08 JesseC I've been reading the nvie one
15:08 JasonSwindle and about tags; http://git-scm.com/book/en/Git-Basics-Tagging
15:09 JesseC I'm pretty sure the command I used installed 0.13
15:09 JesseC lol
15:09 JesseC so I'll revert and do it with tags
15:09 JasonSwindle Most likely
15:09 JasonSwindle I never trust packages.  I have been bitten so badly by old packages
15:09 JasonSwindle Always use PIP if you can
15:09 JasonSwindle if not; PPA / REPO it.
15:10 dN0t joined #salt
15:11 JesseC pip over git repo and tags?
15:11 JesseC I think the only thing I use python based right now is salt.
15:11 JasonSwindle sorry, pip if your package is pytgon
15:11 JasonSwindle *python
15:12 JasonSwindle the git tag is good to use for Salt
15:14 moko joined #salt
15:16 moko How might I reorder a cron.present state so that it creates the cron job after a cmd has executed? Simplified example here: https://gist.github.com/anonymous/6140645
15:18 mgw joined #salt
15:20 __Jahkeup__ joined #salt
15:20 bradley_ joined #salt
15:20 bradley_ how do I install a tar.gz file useing salt?
15:23 drogoh joined #salt
15:23 bradley_ is anyone around?
15:23 JasonSwindle install or decompress?
15:23 StDiluted there are a few different ways to do it
15:24 doublerr joined #salt
15:25 bradley_ what kind of options are we looking at
15:25 bradley_ I want to copy it over using salt, run tar xzvf  file.
15:25 StDiluted do you just want to untag and gzip it, or do you need to run a setup, or what
15:25 StDiluted ah, so, file.managed it somewhere
15:25 bradley_ and then run ./setup -d
15:25 StDiluted and then cmd.script a bash script to untag and run setup
15:26 StDiluted untar*
15:26 bradley_ and then delte the install files
15:26 StDiluted fuckign autocorrect pisses me off
15:26 StDiluted sure, do that with the script as well
15:26 bradley_ what are you using a ipad
15:26 StDiluted a mac
15:26 StDiluted it autocorrects in my irc client
15:26 bradley_ oha strange
15:26 StDiluted just turned it off.
15:27 racooper autocorrect on a non-mobile device. now that's just evil.
15:27 StDiluted seriously
15:27 StDiluted all macs do it, just not all apps use that interface, I guess
15:27 StDiluted but colloquy does
15:27 racooper although you did say it was a Mac so that kinda fits with the Apple vision :P
15:27 pdayton joined #salt
15:27 racooper they are always right, no matter what!
15:28 bradley_ so kind of like this http://intothesaltmine.org/blog/html/2013/03/01/using_the_cmd_module_in_salt_states.html
15:29 wendall911 joined #salt
15:29 shennyg joined #salt
15:29 __Jahkeup__ joined #salt
15:29 StDiluted bradley, let me paste some excertpts for you, one sec
15:30 moko StDiluted: You can turn autocorrect off on a system-wide level. System Preferences > Language & Text > Text > Uncheck "Correct Spelling Automatically", then restart Colloquy. It's worth it.
15:30 StDiluted moko, i just did that.
15:30 bradley_ thanks StDiluted
15:30 JesseC better to just wipe over osx with any linux distro
15:30 bradley_ I am still trying to get started with salt
15:30 JesseC that turns it off as well
15:31 StDiluted bradley: https://gist.github.com/dginther/6140809
15:31 UtahDave bradley_: I really like this example that KennethWilke posted a couple days ago:  https://gist.github.com/UtahDave/6125409
15:32 UtahDave I've also used StDiluted's example of using file.managed a lot as well
15:32 bradley_ what is that written in?
15:33 JesseC KennethWilke's version is python
15:34 JesseC afaik
15:34 aat joined #salt
15:34 JasonSwindle Source; KennethWilke….. if Ubuntu the PPA works well.  https://gist.github.com/JasonSwindle/590bbabbda24bd30b9bc
15:35 unicoletti left #salt
15:35 sashka_ua UtahDave: good morning. :) I have a one notice for you
15:35 sashka_ua UtahDave: try to use colon in command like with salt
15:36 sashka_ua UtahDave: like: salt model file.append /tmp/test "test:test"
15:36 JasonSwindle UtahDave:  BTW; I could not HS with the         script_args: git v0.16.1 in the map
15:36 bhosmer_ joined #salt
15:36 sashka_ua UtahDave: surpise is wating for you :)
15:36 bradley_ what do I do if I run state.highstate and I get    "    Data failed to compile:"
15:37 UtahDave bradley_: that usually means you had a malformed yaml sls file or a stacktrace somewhere.
15:37 JasonSwindle bradley_:  What is the whole message?
15:38 UtahDave JasonSwindle: What do you mean you can't run a highstate?
15:38 tsheibar joined #salt
15:38 JasonSwindle I request it on the master; and it acts like the minions are not there
15:38 devinus joined #salt
15:38 UtahDave JasonSwindle: do other commands work?
15:39 JasonSwindle Yeah; I can do grains.item env
15:39 Katafalkas joined #salt
15:39 JasonSwindle A custom grain; and the minions return
15:40 UtahDave sashka_ua: That worked fine for me.  What version of Salt are you on?
15:40 bradley_ that is it it lists server name then  Data failed to compile:
15:41 forrest joined #salt
15:41 UtahDave bradley_: try running your minion in the cli in debug mode and then try running that command on the master. Often that will show you what's going on
15:41 dN0t joined #salt
15:41 StDiluted bradley: usually an error in your sls
15:42 UtahDave bradley_:   stop the salt-minion service, then run   sudo salt-minion -l debug
15:42 honestly so when I try to bootstrap salt-minion on ubuntu, I get the error "No dependencies installation function found". wtf does that mean?
15:42 bradley_ well it saams to work on the master
15:42 bradley_ http://pastebin.com/sBRPCffi
15:43 forrest honestly, which version of ubuntu are you running?
15:43 whit joined #salt
15:43 honestly what the
15:43 JesseC Anyone have an idea on this error? IOError: [Errno 2] No such file or directory: '/var/cache/salt/minion/proc/20130802054146852351'
15:43 honestly "Linaro 12.11"
15:43 forrest that looks like a job idea JesseC
15:44 JesseC Salt 0.16.2, running this command salt-call --local state.sls lnpp
15:44 honestly oh, that's what all my nodes are actually running
15:44 JesseC forrest: job idea?
15:44 honestly it's ubuntu for omap apparently
15:44 bradley_ http://pastebin.com/6DYKenT6
15:44 forrest JesseC, sorry job id
15:44 forrest still a bit early
15:44 bradley_ shit the room just got bussy
15:44 honestly the strange this is it worked on two other beagle boards that started out with an identical boot image
15:45 forrest really?
15:45 honestly or should have at least - I wasn't personally there when they copied the sd cards
15:45 forrest I was concerned there was an issue here: __apt_get_noinput software-properties-common
15:45 bradley_ does this error make sense    >  http://pastebin.com/6DYKenT6
15:45 forrest because there's that specific line for apt get if your version is over .04
15:46 krissaxton joined #salt
15:46 forrest bradley_, yea that makes sense, the service is ntpd
15:46 StDiluted honestly: that's happening because your distro is not listed in the salt boostrap script
15:47 bradley_ forrest  but why would it say result True and also result False
15:47 honestly that can't be it, it worked on two other identical setups
15:47 bradley_ saying the same thing twise but in two diffrent ways
15:48 honestly I'll go and try to install it manually from the ppa...
15:48 forrest bradley, I don't understand, that's not the same thing twice. One is confirming the ntp package itself is installed, which is true, the package is just called ntp, the second check is the service, which you're referencing as ntp, which would return false, the service is ntpd
15:48 forrest it's telling you the service ntp isn't available, because it isn't, it doesn't exist.
15:49 honestly hah
15:49 forrest Does that make sense?
15:49 bradley_ forrest thanks, I did not understand that part, I am still learning salt
15:49 honestly "NoDistroTemplateException"
15:49 honestly that's cool
15:49 honestly and really weird
15:49 forrest bradley_, yea no problem, if you just change the service to be ntpd, it should start it up, or at least provide you with a better result.
15:49 troyready joined #salt
15:50 honestly how the hell did the install work on the two other nodes...
15:50 nliadm so, I run a salt-call, and it _seems_ to trigger a pillar compilation on the master (hord to tell, it doesn't print who it's doing it for)
15:50 nliadm then the minion sits for a about a minute at: [DEBUG   ] Loaded minion key: /etc/salt/pki/minion/minion.pem
15:50 conan_the_destro joined #salt
15:50 nliadm then says authentication failed
15:50 forrest honestly, when you run uname you see it just say ubuntu right?
15:50 JesseC Gist here if anyone knows where to start with an IOError from salt-call. https://gist.github.com/JesseC-/61f0afd33dc4c207b090
15:51 honestly forrest: it says "Linus"
15:51 honestly Linux*
15:51 honestly lol
15:51 bradley_ forrest> thanks I will work on that, it make more sense
15:52 forrest bradley_, yea no problem, if it still doesn't work toss your .sls in and we can take a look.
15:53 JasonSwindle JesseC:  What are you deploying on; Hardware wise?
15:53 JesseC JasonSwindle: It's a CentOS 6.4 VM using Vmware Workstation.
15:53 forrest honestly, heh. I really don't get how it worked on the other two but not this one. Can you double check that the image wasn't updated?
15:54 JasonSwindle Oh.  That is an odd error
15:54 JesseC JasonSwindle: I used the same image/snapshot yesterday fine, nothing really changed aside from me reverting the image back to being clean
15:54 forrest JesseC, lspci is available on that machine right?
15:54 honestly forrest: uname -a on the nodes all gives identical output
15:54 honestly that tells me they're all from the same image, right?
15:54 forrest JesseC, yesterday you and I were messing with version 15. I notice you're using 16 today
15:54 StDiluted honestly: the salt bootstrap script uses quite a few variables to try to determine the distro
15:55 forrest honestly, that would give the kernel config, not sure if there could have been other changes though.
15:55 StDiluted honestly you might take a look at the code in the bootstrap script and see where it's failing
15:55 JesseC forrest: yep, but i upgraded to 16 this morning so I could test, verified it worked, then reverted the image to do a test on a clean image
15:55 forrest JesseC, interesting
15:55 JesseC JasonSwindle: no lspci, could be a centos thing though
15:55 JasonSwindle Maybe.  I never use CentOS :)
15:55 honestly if I add the ppa, it fails with NoDistroTemplateExcepten in SoftwareProperties.py:584
15:55 JesseC JasonSwindle: there's no lspci on my production machine either
15:56 JesseC so it's a CentOS thing
15:56 honestly or distro.py:87, rather
15:57 racooper I have lspci on my centos boxes...
15:57 sashka_ua UtahDave: Im on 0.16.0
15:57 juanlittledevil joined #salt
15:58 sashka_ua UtahDave: when I have colon in text ot parameters salt converts any text to dict
15:58 UtahDave sashka_ua: what's the error that you're getting?
15:58 sashka_ua UtahDave: it just adds {"test", "test"} text instead of "test:test"
15:59 sashka_ua UtahDave: so far I tried some commands accepting cli parameters, it automatically converts any text with colon into dict
16:00 UtahDave sashka_ua: Hm.  Any way you can test on 0.16.2? or develop?
16:00 avienu joined #salt
16:01 UtahDave sashka_ua: Yeah, on the develop branch it's working correctly for me.
16:01 UtahDave sashka_ua: are you on a non-English os, by chance?
16:01 sashka_ua UtahDave: Im sorry, put space after colon
16:01 sashka_ua UtahDave: "test: test"
16:02 UtahDave sashka_ua: ah, I'm sorry. I didn't see the space there.
16:02 sashka_ua UtahDave: coz I didn't put it here :)
16:02 krissaxton joined #salt
16:04 sashka_ua UtahDave: its easely reproducable when only single ": " presents in text. "test: test: test" will work as it should
16:05 forrest JesseC, can you try installing pciutils (or the cent equivalent) to get lspci?
16:06 bradley_ forrent> finally got ntp to work properly after fillowing this example https://github.com/saltstack-formulas/ntp-formula/blob/master/formula1/init.sls
16:06 forrest bradley_, Great, you see what I was talking about with the service name issue?
16:06 UtahDave sashka_ua: Hm. still working fine for me.   http://pastebin.com/VHgCQu43
16:07 bradley_ forrent> yes I see that now.  There are a few other more complicated ntp examples that will have to wait till later
16:07 lazyguru joined #salt
16:07 forrest Yea no reason to dive in too deep right off the bat.
16:08 sashka_ua UtahDave: check this out: http://pastebin.com/tcs4P3sR
16:09 sashka_ua UtahDave: but anyway, Im sorry. I did a search through the issues, and found this: https://github.com/saltstack/salt/issues/4395
16:09 Xeago joined #salt
16:09 JaredR joined #salt
16:10 UtahDave sashka_ua: That's weird. I copied in the exact same commands and they all worked just fine.  Any way you can test against develop branch?
16:11 sashka_ua UtahDave: no,  I have only packaged versions
16:11 it_dude joined #salt
16:11 sashka_ua UtahDave: but I will deploy dev branch to test such weirdness
16:11 KennethWilke UtahDave: JasonSwindle: why you be sharing my public gists :p
16:12 UtahDave KennethWilke: because it's beautiful.  I forked it, too.  :)
16:12 felixhummel joined #salt
16:12 JasonSwindle :P
16:12 KennethWilke lol wow, well someday i'll show ya something i'm proud of :)
16:16 JesseC forrest: yep, I'm reverting my vm again to start fresh and I'll try to manually intall pciutils and try it again
16:17 forrest JesseC, awesome. I find it really odd that it worked previously then just crapped out.
16:17 StDiluted ARGH, why is this external pillar not working
16:20 gamingrobot joined #salt
16:20 akitada joined #salt
16:23 jschadlick joined #salt
16:23 Linz joined #salt
16:25 JaredR With GitFS, when does salt know to pull from the remote repos?
16:25 sw_ joined #salt
16:25 KyleG joined #salt
16:25 KyleG joined #salt
16:26 aat joined #salt
16:26 drawsmcgraw JaredR: My understanding is "it depends on the order of your 'sources' in your master config"
16:26 Lue_4911 joined #salt
16:26 JaredR So, every time one does a high state, does salt start looking in the git repos for updated files?
16:27 drawsmcgraw Good question. I don't know but I'm interested in the answer. I know it keeps a local cache of files that come from Git
16:27 jschadlick Hey everyone, I am running into an issues using salt with the user set as not root. It seems that if salt isnt running via root, I cannot remotely run scripts as other users. Does anyone know a workaround for this?
16:27 drawsmcgraw How often it updates that cache, though, I don't know
16:27 JaredR (I suppose I could look at the source to find out why GitFS does)
16:28 JaredR s/why/what
16:28 drawsmcgraw JaredR: Whenever I'm curious about behavior, I turn logging up to Debug and watch as it flies by.
16:28 JaredR I'm not to that point yet. I'm just getting started, but the docs didn't seem to address it, so I thought I'd ask.
16:29 devinus joined #salt
16:29 forrest Once you figure it out you should update the docs Jared
16:29 forrest A lot of them could use more specific details regarding that stuff.
16:29 forrest Q&A section almost
16:32 JaredR Good suggestion. How does one update the documentation (not that I'm to that point yet, but I hope to be sometime in the coming weeks or months)
16:32 UtahDave jschadlick: are you running the master or the minion as non-root?
16:33 honestly I don't know if I'm totally crazy,  but I'm now suspecting that an update between where I set up the first two nodes and now changed the distribution info from "Ubuntu" to "Linaro"
16:33 UtahDave JaredR: the docs are right in the salt repo.  Anyone can open a pull request to help update and fix the docs.
16:33 forrest JaredR, all the docs are up on github.
16:33 honestly since there zero reason to think that salt would've installed on the two other nodes otherwise
16:33 forrest honestly, that could make sense.
16:33 JaredR Thanks for that info, forrest, and UtahDave.
16:34 honestly so who do I poke to get Linaro added as a Ubuntu derivative into the salt distribution info thingy?
16:34 UtahDave JaredR: you're welcome!  We love getting pull request in general, and especially for doc fixes.
16:34 honestly I joined #linaro but it seems to be somewhat idle
16:34 forrest JaredR, yea no problem, they're really quick to update docs usually too.
16:34 UtahDave honestly: open an issue on the salt github repo and we'll get that in
16:34 honestly coolbeans
16:35 juanlittledevil joined #salt
16:35 juanlittledevil good morning guys!
16:35 forrest honestly, it seems odd they would have changed that, just complicates things
16:35 forrest hey
16:36 honestly UtahDave: salt-bootstrap or salt?
16:36 jschadlick UtahDave: non-root on both. we created a salt user that has can run all of the salt commands, and want to switch users to run a script on a particular minion
16:37 UtahDave honestly: probably both.
16:37 UtahDave jschadlick: if you're running non-root on the minion, how will it be able to run commands as another user?
16:39 chrisgilmerproj UtahDave: we're running it as user 'salt' on minion and we edited the sudoers file to allows 'salt' to run as the other user for a specific script
16:40 JesseC forrest: same issue on a fresh vm, minus the info about lspci
16:40 forrest JesseC, that's a bummer.
16:40 forrest can you paste the gist?
16:40 JesseC forrest: I'm going to try with 0.16.0 and see if it makes a different
16:41 JesseC difference^
16:41 JesseC sure
16:41 forrest Awesome, maybe it's .2
16:41 UtahDave chrisgilmerproj, jschadlick: what's the response you're getting back?
16:41 honestly UtahDave: https://github.com/saltstack/salt/issues/6496
16:41 jschadlick UtahDave: ERROR: Failed to change from gid 116 to 640. Error: [Errno 1] Operation not permitted
16:42 UtahDave thanks, honestly.  terminalmage may ask you for some info in addition to that.
16:42 JesseC forrest: the gist -> https://gist.github.com/JesseC-/61f0afd33dc4c207b090
16:42 UtahDave jschadlick: ok, i'm not sure salt uses sudo to run as a different user.  Let me jump into the code real quick
16:42 forrest can you run salt-run jobs.list_jobs ?
16:43 chrisgilmerproj that might be our problem right there UtahDave
16:43 chrisgilmerproj if sudo isn't the way to do it then we just need to figure out how to configure it
16:43 wopi joined #salt
16:43 UtahDave chrisgilmerproj: Yeah, let me check this out
16:44 forrest JesseC, nevermind what I said, salt won't even install, duh
16:45 wopi Hi People, I installed salt-master on my laptop and salt-minion on my Debian box, I set in the minion config the IP of the master, how I should proceed further to accept the minion on the server ?
16:45 JesseC forrest: salt installs fine
16:45 JesseC I already killed the vm though, redoing my base so I don't have to keep running updates
16:45 forrest oh right it's the state that fails, divided attention go.
16:45 JesseC lol
16:45 forrest Ok cool, I'll be interested to see what 0.16 looks like
16:45 StDiluted wopi: you need to accept the key on the master
16:45 StDiluted wopi: so salt-key -L
16:46 StDiluted and then salt-key -a hostname.of.minion
16:47 UtahDave chrisgilmerproj and jschadlick: you can send sudo commands with  cmd.run   but most of the other modules and things use libraries, etc.
16:47 racooper site note: seems that salt-key -L is supposed to be deprecated? is that going to be removed at some point?
16:47 chrisgilmerproj gotcha
16:47 UtahDave Are you trying to limit the salt user to just a few commands or something?
16:47 chrisgilmerproj there are big security questions around giving salt too much power
16:47 chrisgilmerproj almost everything is controlled via sudo here
16:47 wopi StDiluted: I didn't see any unaccepted keys on master
16:48 chrisgilmerproj and if someone were to become this ultra powerful salt user how would we stop them from world domination?
16:48 forrest wopi, do you have all the ports configured?
16:48 StDiluted wopi: that would mean your minion is not contacting the master as of yet
16:48 wopi The only thing I have configured on minion is the IP of the master setup
16:48 UtahDave chrisgilmerproj: Well, you have to give the user at least as much power as you want it to be able to do anything useful on your server
16:48 wopi and then I have stoppen and started minion again
16:49 StDiluted you need top open TCP 4505 and 4506 on the minion
16:49 chrisgilmerproj sure, we had hoped to limit it to specific commands/scripts
16:49 StDiluted top*
16:49 StDiluted ugggh
16:49 StDiluted what the hell
16:49 StDiluted to*
16:49 chrisgilmerproj and in some cases file permissions requires we use runas
16:49 chrisgilmerproj or use some kind of sudo
16:49 forrest You mean the master STDiluted.
16:49 forrest *?
16:50 StDiluted yes. sorry, frustrating morning
16:50 forrest wopy, take a look at : http://docs.saltstack.com/topics/tutorials/firewall.html
16:50 forrest *wopi
16:50 chrisgilmerproj is there a "best-practices" for server security?
16:51 chrisgilmerproj its a hard sell to get my company to just give salt full root access
16:51 wopi thanks for the help guys, will check it out
16:51 wopi bye
16:52 StDiluted chrisgilmerproj: lots of other processes on your servers probably have 'root' level access, I would bet. The question is who you give root on the salt master to
16:53 chrisgilmerproj so really the security question here is how do you protect the salt master
16:53 chrisgilmerproj that's what you guys do?
16:53 bitz joined #salt
16:53 Xeago joined #salt
16:54 StDiluted thats what I do
16:54 racooper yep, same here.
16:54 forrest at my old spot it was a combo (granted we used puppet), but it was a combination of limited access for who can run jobs on the server, plus a system of peer reviews/testing for any new manifests that went in.
16:54 chrisgilmerproj are there any other standard security setups?
16:55 UtahDave people will often run the salt-master as a non-root user
16:56 forrest so when a developer created a new manifest, he'd check it into git, I'd peer review it, and apply it in dev, then if everything looked good there, we'd pass it to the admins who would again peer review, then merge it into the prod branch, and apply in production.
16:56 UtahDave chrisgilmerproj: then use either client_acl or external auth to severely limit what users are allowed to execute
16:56 chrisgilmerproj gotcha
16:57 chrisgilmerproj but what we're attempting to do seems a little too granular
16:57 chrisgilmerproj yeah?
16:57 forrest seems that way
16:57 forrest do you currently have anything that accesses the machines from another box (logs or anything) as the root user?
16:57 forrest maybe you can try to use that as an example
16:59 UtahDave chrisgilmerproj: It's not impossible, but definitely tougher to do.  There's no "sudo" wrapper that lets the salt-minion run all commands through sudo
17:00 chrisgilmerproj security is just hard stuff;)
17:01 UtahDave yes, it is!
17:02 krak3n` joined #salt
17:02 Xeago_ joined #salt
17:07 carmony joined #salt
17:11 krak3n` joined #salt
17:14 chrisgilmerproj thanks guys for the help
17:14 JaredR joined #salt
17:19 druonysuse joined #salt
17:19 druonysuse joined #salt
17:20 logix812 joined #salt
17:21 diegows joined #salt
17:24 mike25ro joined #salt
17:25 devinus joined #salt
17:27 sw_ joined #salt
17:27 chutzpah UtahDave: just a FYI, you apparently have an unlisted dependency on pcituils: https://bugs.gentoo.org/479450
17:27 chutzpah s/pcituils/pciutils/
17:28 ioni joined #salt
17:29 UtahDave I think not having lspci will just keep the minion from loading grains that depend on it.
17:31 dN0t joined #salt
17:32 drawsmcgraw Throwing this out there just in case -> No one's really doing much with using Salt to provision/manage Cisco gear, eh?
17:32 drawsmcgraw I see the Nexus series supports running Python scripts. Not sure if someone's doing some really creative pidgeon-holing
17:33 UtahDave drawsmcgraw: No, not yet.  We've been in discussions with a few networking hardware providers to salt them. That's on our todo list for sure.
17:33 UtahDave drawsmcgraw: have you attempted anything with that yet?
17:34 spmurrayzzz joined #salt
17:34 drawsmcgraw Negative. I was just introduced to Cisco's POAP concept just today
17:34 drawsmcgraw Powers on, DHCP tells it where to find a Python script. Grabs the script & runs it.
17:35 drawsmcgraw You use their (Cisco's) Python API to configure it -> http://www.cisco.com/en/US/docs/switches/datacenter/nexus3000/sw/python/api/503_U41/python_api.pdf
17:35 spmurrayzzz Hey all, im curious about strategies for restricting custom modules to only be used on certain hosts but the depends() wrapper doesn't seem to fit the bill
17:35 UtahDave very interesting, drawsmcgraw
17:35 spmurrayzzz any ideas?
17:35 drawsmcgraw I'm imagining a template being populated with Pillar data to auto-generate a Python script. I haven't gotten my hands dirty yet, though
17:36 auser joined #salt
17:36 UtahDave cool, drawsmcgraw
17:40 forrest JesseC, any luck with 0.16?
17:49 dthom91 joined #salt
17:49 mike25ro stupid Q guys... how can i set up grains to have multiple values for a key .. smth like key={'val1', 'val2'}
17:53 dave_den mike25ro: you can set a list with square brackets, like salt 'somehost' gains.setval mytestkey "['val1', 'val2', 'val3']"
17:54 jessep joined #salt
17:54 mgw joined #salt
17:59 bradley_ can I put minion names in the top.sls file?
18:00 JasonSwindle bradley_:  You can match on FQDN
18:00 bhosmer joined #salt
18:00 JasonSwindle or by many things
18:00 JasonSwindle http://docs.saltstack.com/topics/targeting/index.html
18:01 bradley_ JasonSwindle> http://pastebin.com/Lz7VH4ys
18:01 Ryan_Lane joined #salt
18:02 JasonSwindle http://docs.saltstack.com/topics/targeting/globbing.html
18:02 JasonSwindle It may not know how to match
18:02 bradley_ JasonSwindle> I am trying to start with a simple top.sls for now before I make it more complicated with papern matching and such
18:03 whiskybar joined #salt
18:04 bradley_ JasonSwindle> I am guessing some sort of syntax issue, but I whish the error messages where easier to read like the ones you get when you screw up your sudoers file
18:04 JasonSwindle bradley_:  this is how my old top.sls was
18:04 JasonSwindle https://gist.github.com/JasonSwindle/c9a371f24c67dff1a157
18:05 JasonSwindle Ah; git history
18:06 bradley_ JasonSwindle> where would I identify exactly what the hostname for say the "-web-"   servers are
18:06 drawsmcgraw bradley_: I added a line - > http://pastebin.com/Z7FZnZxA
18:06 drawsmcgraw That's matching by the minion-id
18:06 drawsmcgraw And telling salt that you're targeting the minion via a Grain
18:07 bradley_ drawsmcgraw>  even though I dont have grains set up yet?
18:07 drawsmcgraw Alternatively, you could use Nodegroups in your Master config file to create a group for all of your web servers
18:07 drawsmcgraw bradley_: Grains come 'for free'
18:07 drawsmcgraw They're generated when the minion starts up
18:07 drawsmcgraw You may be thinking of Pillars
18:08 drawsmcgraw salt some-web-server grain.items  (I think) will list all the grains
18:08 rsimpkins left #salt
18:08 drawsmcgraw Have to run. End of my day. Good luck!
18:08 mike25ro dave_den: thanks buddy
18:08 mike25ro drawsmcgraw: take care
18:08 drawsmcgraw mike25ro: Thanks. later
18:10 bradley_ it is having issues http://pastebin.com/b1bNqsLm
18:12 jschadlick joined #salt
18:12 helderco joined #salt
18:16 tsheibar joined #salt
18:20 JaredR joined #salt
18:21 dN0t joined #salt
18:26 it_dude joined #salt
18:26 JaredR joined #salt
18:30 whit joined #salt
18:34 abe_music joined #salt
18:35 abe_music joined #salt
18:36 abe_music has anyone had any issues with salt-cloud launching and hanging after the remote deploy.sh script runs?
18:36 abe_music i'm using AWS with an Ubuntu-based AMI
18:40 abe_music oh, and i'm using the bootstrap script with "git develop" args…maybe it will work without those args
18:46 abe_music yup, that worked...
18:54 eightyeight when will 16.2 be pushed to the ppa?
18:55 dthom91 joined #salt
18:58 nineteeneightd joined #salt
18:59 blee_ joined #salt
18:59 jschadlick joined #salt
19:24 _ilbot joined #salt
19:24 Topic for #salt is now Welcome to #salt - http://saltstack.org | 0.16.0 is the latest | Please be patient when asking questions as we are volunteers and may not have immediate answers - Channel logs are available at http://irclog.perlgeek.de/salt/
19:25 trigoman akoumjian: I just wanted to say thanks :)
19:26 whit joined #salt
19:27 forrest JesseC, sounds good. Hopefully it's a quick fix, or there is some sort of work around
19:29 JesseC odd, development is on 0.16.0-1669-g2824746 but stable is on 0.16.2
19:29 JesseC ?
19:31 racooper Well, salt 0.16.0 just reached the utexas epel mirror.
19:33 JasonSwindle left #salt
19:35 dN0t joined #salt
19:35 forrest heh
19:35 forrest Does anyone know, when you have a file.managed and use - mode: 123, can you do - mode: 0123?
19:36 forrest None of the examples are explicit regarding the 4th item
19:36 JesseC I've seen other peoples examples here that used 0123
19:36 mike25ro forrest: if the mode has a 0 in front .. then you have to use '02775'
19:36 mike25ro sorry '0775'
19:36 mike25ro or just mode: 775
19:36 forrest so you have to have single quotes to declare 4 integers?
19:37 mike25ro i just read about it couple minutes ago
19:37 mike25ro single quotes if you have a 0 to start the number
19:37 JaredR joined #salt
19:37 forrest or any other value
19:37 mike25ro you can have 2775 but you must have '0775'
19:37 forrest what if it's 1775
19:37 forrest I see
19:37 forrest gonna file a issue on that then for the docs, because that's importan to some people.
19:38 forrest unless you said you saw it somewhere already?
19:38 mike25ro yes i saw it
19:38 mike25ro just read couple minutes ago about that
19:38 forrest which docs?
19:38 JesseC forrest: You're on RHEL 6.4 right? Have you tried 0.16.2 yet?
19:38 mike25ro don't remember where exactly ...i have 120 tabs open :D
19:39 forrest JesseC, no I haven't I'm running these VMs locally and they are taxing my box pretty badly, waiting on a RAM upgrade before I experiment too much
19:39 mike25ro i am still ... looking to sort out my issue
19:39 forrest it's negatively impacting my 'real' work
19:39 forrest it should still have that info here: http://docs.saltstack.com/ref/states/all/salt.states.file.html
19:39 forrest where some examples use it
19:40 mike25ro yeah /// but a normal number doesn't start with 0 ... so i think it is parsed somehow... and the result of 0775 is actually 420 ... or smth like that
19:41 mike25ro related to octal ... smth :)
19:41 mike25ro unless you use '0775'
19:41 forrest Yea, it's more just a 'hey I like to explicitly declare all 4', and I wouldn't think to use the single quotes
19:41 mike25ro :)
19:41 chutzpah joined #salt
19:42 mike25ro forrest: did you use states based on matching a grain?
19:42 forrest nope
19:42 mike25ro darn :) me neither .. so this is what i am trying now ...
19:42 mike25ro :P
19:42 andrew_ joined #salt
19:43 forrest if you find where it referenced the '0755' thing, let me know
19:43 JesseC forrest: in case you have the same issue later here is the bug report, https://github.com/saltstack/salt/issues/6502
19:43 forrest Awesome thanks JesseC
19:43 mike25ro ok
19:44 lemao joined #salt
19:44 balaa greetings fellow salt stackers, I'm seeking guidance. Currently at my organization we have 1 production machine. What would you guys suggest as per managing a production deployment and a staging deployment on the same minion.. if that makes sense
19:45 mike25ro balaa: i am also a newbie... so not much help from my side :)
19:45 JesseC balaa: I'm about to be using vagrant + salt for all of my development personally.
19:45 JesseC balaa: that way production is only production
19:45 JesseC and development is repeatable anywhere
19:45 balaa JesseC: I'm also using vagrant and salty vagrant for development
19:46 balaa but I'd like to have a "staging environment" that has a copy of the production dataset
19:47 balaa not to mention I have multiple production applications running on this machine, I do deployments with salt.sls appname
19:47 forrest Does anyone know where http://docs.saltstack.com/ref/states/all/salt.states.file.html lives within https://github.com/saltstack/salt/tree/develop/doc/topics ?
19:47 abe_music it looks like the salt-cloud map file has changed in that my grains now need to be outside of the minion dict? but when i do that, the grains don't show up in the minion config file….anyone know what the new format is supposed to look like?
19:47 whit joined #salt
19:48 JesseC balaa: gotcha, yeah, I don't have anything setup like that myself yet, I'd be interested in working out a method for that as well
19:49 mike25ro guys .. do you know how to ... match grains inside states?
19:50 mike25ro http://pastebin.com/VP34JGWz
19:50 balaa JesseC: I'm thinking I could have a different set of pillars per environment, I guess my only hangup is how to have multiple environments on the same minion
19:52 UtahDave eightyeight: 0.16.2 should already be up on the ppa, or at least building
19:52 forrest ahh ok they're pulling the docs straight from the file.py, cool.
19:52 forrest UtahDave, did you see JesseC's error from earlier?
19:52 forrest He confirmed it wasn't present in 0.16.0
19:53 UtahDave forrest: no, I just got back into the office. What error is he seeing?
19:53 eightyeight UtahDave: not there yet. i'll keep an eye out
19:53 forrest https://github.com/saltstack/salt/issues/6502
19:53 forrest there's the issue he filed
19:53 JesseC I'm also about to test 0.16.1
19:53 balaa I like the idea of applying a state as a form of doing a deployment with state.sls, I guess I'm wondering how a can have a different set of pillars per a given state on the same machine
19:54 racooper when there is an upgraded package, will a pkg.install with the existing package do the same as a yum update <package>?
19:55 UtahDave interesting, forrest. We'll see what's going on
19:55 forrest JesseC, nice that would be good to see.
19:55 forrest Yea UtahDave, JesseC has been troubleshooting it this morning.
19:55 forrest well, afternoon now
19:55 JesseC The issue is not present in 0.16.1
19:56 JesseC I'll update that report
19:56 forrest cool
19:58 racooper nvm answered that question. now, is there a safe way to restart all minion services from salt master?
20:02 jschadlick I am getting "Function 'state.over' is unavailable" when trying to run "salt-run state.over"  My other runners are working fine. Any ideas?
20:03 isomorphic joined #salt
20:04 forrest racooper, can't you just do salt * service.restart salt-minion?
20:04 forrest or are you trying to do it in a state?
20:05 racooper not as a state.  just making sure that telling a minion to restart itself wouldn't bork anything
20:05 forrest I have no idea
20:05 forrest I'm booting my other VM now, I'll let you know in a minute
20:09 forrest racooper, I just arn it through, and no issues in the log as far as I see.
20:09 racooper thanks.
20:09 hjubal joined #salt
20:09 forrest np
20:10 juanlittledevil joined #salt
20:10 jslatts joined #salt
20:11 therealbill joined #salt
20:11 devinus joined #salt
20:12 forrest racooper, looks like they resolved issues related to restart the salt-minion itself in 0.14.0: https://github.com/saltstack/salt/issues/1888
20:12 devinus joined #salt
20:15 dN0t joined #salt
20:16 balaa JesseC: I think I figured it out, I was over-thinking it. Just add staging configuration pillars to te exisiting pillars and update the states to use production and staging pillars
20:18 jslatts joined #salt
20:20 forrest is there anything similar to puppet parser validate in salt?
20:21 craig_ joined #salt
20:22 abe_music alright, i'm going crazy, but i can't seem to figure out the new way to define grains in my salt-cloud map files that causes them to show up in /etc/salt/minion when the machines come up….this was working a few weeks ago, but since i updated to the latest on develop it doesn't. i'm sure something has changed, but i can't find anything on what the changes were
20:23 abe_music teskew: you seen this problem?
20:23 chrisgilmerproj joined #salt
20:25 __Jahkeup__ joined #salt
20:27 StDiluted how can i set a default for a grain in case it isn't set, inside a template?
20:27 StDiluted or in an SLS rather
20:28 dthom91 joined #salt
20:28 mike25ro StDiluted: you can check for it using ...
20:28 mike25ro {% if 'db' in grains.get('saltrole', []) %}
20:28 mike25ro that is what i am doing now..
20:29 mike25ro then you can do a cmd run ...
20:29 mike25ro to set that grain
20:29 mike25ro i am still wondering what is the best use... grains/ pillars for some things
20:29 StDiluted hm, no i want to use a grain
20:30 StDiluted in an sls
20:30 mike25ro i am doing that...
20:30 mike25ro let me show u
20:30 StDiluted but i dont want to get a stacktrace if the grain is not set
20:31 StDiluted source: - salt://file/file-{{ grains['grain'] }}.txt
20:31 mike25ro http://pastebin.com/Mnfz76FK
20:31 StDiluted will stacktrace if grains['grain'] is not set
20:31 mike25ro not sure :)
20:31 StDiluted ok so grains.get
20:32 forrest why can't you just use an else statement in that jinja 2 portion?
20:32 JaredR joined #salt
20:32 mike25ro not sure.. i am a newbie here
20:33 StDiluted think this will work: salt://robots/robots-{{ grains.get('ec2_environment',['default'] }}.txt
20:34 StDiluted forrest: I could, but that doesn't seem to want to work
20:34 forrest http://pastebin.com/G0NjLEbD
20:34 forrest so something like that doesn't work?
20:35 StDiluted that would work. You're using a grains.get
20:35 StDiluted I was using salt://robots/robots-{{ grains['ec2_environment'] }}.txt
20:36 StDiluted which will stacktrace if the grain is not set
20:36 mike25ro {% else %}
20:37 StDiluted I don't really want a whole if-else in there
20:37 mike25ro you should use if... as far as i can see
20:37 mike25ro i made a typo and it didn't stacktrace
20:37 forrest so if you're gonna do that StDiluted, you're going to have to set the ec2_environment variable
20:37 StDiluted but I believe that salt://robots/robots-{{ grains.get('ec2_environment',['default'] }}.txt should work
20:37 forrest yea that might work
20:38 StDiluted grains.get lets you set a fallback in case the var is not there
20:38 forrest right
20:38 StDiluted that's my whole point,
20:38 bbopen joined #salt
20:40 mike25ro {% if kenny.sick %}
20:40 mike25ro Kenny is sick.
20:40 mike25ro {% elif kenny.dead %}
20:40 mike25ro You killed Kenny!  You bastard!!!
20:40 mike25ro {% else %}
20:40 mike25ro Kenny looks okay --- so far
20:40 mike25ro {% endif %}
20:40 mike25ro this is the ... jinja .. if else if
20:40 forrest yea that's the same thing I pasted, minus spelling elif correctly :P
20:40 nliadm forrest: StDiluted: in the case of a file like that, you can also use a list. salt will grab the first one that exists
20:40 mike25ro ah sorry
20:40 forrest all good!
20:41 forrest oh thanks for the info nliadm
20:41 StDiluted yep, it will, but if the grain is not set then it will stacktrace and fail
20:41 StDiluted even if there's another option below it
20:41 nliadm ah, true
20:41 StDiluted so you;d have to put the default case at the top and it would hit that every time
20:41 mike25ro StDiluted: let me test... is should not stack
20:42 KyleG joined #salt
20:42 KyleG joined #salt
20:42 StDiluted I've tested it, it does.
20:42 devinus joined #salt
20:51 at3560k joined #salt
20:51 mike25ro StDiluted:
20:51 mike25ro this is what i tested with
20:52 mike25ro http://pastebin.com/Zn1pxmNe
20:52 forrest he doesn't want to use an if statement
20:52 mike25ro if i change the saltenv to another value like saltenv3 (which does NOT) exist, it works
20:52 mike25ro forrest:  ah... my bad
20:54 mike25ro by any chance... the match thing inside a state .. should be able to match a grain.. right/? ... i couldn't make it run
20:56 forrest Does anyone know if someone made a syntax plugin that supports both yaml and jinja2?
21:03 mike25ro another weird thing
21:03 mike25ro salt '*' match.grain 'os:Debian'
21:04 mike25ro it returns FALSE... although i can see that is Debian in the : salt 'salt2' grains.item os
21:04 mike25ro salt2:
21:04 mike25ro os: Debian
21:04 mike25ro is this weird or what?
21:05 minaguib Hey folks. Unrelated to salt but I'm hunting for commentaries from seasoned sysadmins :) If you're interested in a fan-out rsync mechanism, I'd appreciate if you check out: https://github.com/minaguib/dcarsync#request-for-comments
21:05 balaa is it possible to apply only part of a state?
21:05 devinus joined #salt
21:05 balaa can I do for example state.sls state.name
21:06 devinus joined #salt
21:06 balaa I tried but i didn't work
21:06 minaguib balaa: Not without conditionals to include/exclude the other parts.  Keep in mind that state nodes form a tree (DAG), so exlcusions are tricky
21:06 jacksontj joined #salt
21:06 devinus joined #salt
21:07 jacksontj is there a way inside a salt module to get the name of the user running the command? The minion knows (i see it in the logs) i just can't figure out how to access it from the module that i'm writing
21:07 devinus joined #salt
21:07 mike25ro balaa: i don't think so... i haven't read about that... unless you use include.... and create a state from more states ...
21:08 devinus joined #salt
21:08 krissaxton joined #salt
21:08 JesseC balaa: so you're just going to create separate states for staging and development?
21:08 mike25ro like a big state called lamp.sts where you include ...  apache php mysql states .. and each of this states is doing just a part
21:09 JesseC balaa: sorry, staging and production
21:10 JesseC balaa: I was replying to your comment waaay up top about the pillars
21:13 juanlittledevil joined #salt
21:13 Psi-Jack joined #salt
21:17 andrewclegg joined #salt
21:18 forrest JesseC, did you see terminalmage's response on your issue?
21:19 JesseC forrest: yes, I'm about to upload the file
21:19 forrest cool
21:23 bwghughes joined #salt
21:24 terminalmage thanks JesseC
21:26 JasonSwindle joined #salt
21:26 troyready joined #salt
21:28 bostonian joined #salt
21:30 whit joined #salt
21:33 hazzadous joined #salt
21:37 giantlock joined #salt
21:38 dthom91 joined #salt
21:39 JesseC terminalmage: I updated it and tested it again just now with that file, same results
21:39 Xeago joined #salt
21:39 akoumjian So, anyone doing any kind of special work making it easy to provision docker containers with salt?
21:40 JesseC files*
21:41 StDiluted akoumjian: haven't seen anyone doing that yet but it seems like a great idea
21:41 berto- joined #salt
21:41 berto- is there a way to call salt from a machine other than the master?
21:42 berto- i know there wasn't a while back, just checking to see if this has changed since.
21:44 JesseC berto-: as in locally?
21:44 JesseC berto-: or send a command to the master to send to another minion
21:44 JesseC berto-: because I'm pretty sure you can do both
21:44 berto- JesseC: on the master i can run `salt '*' state.highstate` … can i run that from another server on the network or do i have to ssh into the master?
21:45 JasonSwindle berto-:  has an API, that is one option
21:46 juanlittledevil I'm having a weird issue where after a while my master is no longer able to talk to minions running centos 5 or redhad 5 for that fact. I can however initicall a highstate from the minion itself. Has any of you guys have this happen?
21:46 juanlittledevil The only way around this is to restart the minion
21:46 berto- JasonSwindle: you mean this: https://github.com/saltstack/salt-api … checking it out.  thanks.
21:48 JesseC berto-: this was from this morning: <drawsmcgraw> When you need to send a message, from any machine (including minions), just --> salt-call event.fire_master 'some text' 'your tag'
21:51 terminalmage JesseC: I was able to replicate, but only after first clearing the cachedir
21:51 berto- JesseC: i don't think that's exactly what i'm looking for.  that wouldn't tell the master to run highstate.  i think the salt-api project is on the right track.
21:51 Katafalkas joined #salt
21:51 JesseC terminalmage: odd. I didn't clear anything manually
21:51 ipmb what's the best way to distribute salt modules that depend on extra python libraries?
21:52 JesseC terminalmage: I simply installed using the bootstrap method I mentioned, created the sls and tried to run it.
21:52 JesseC terminalmage: but it also affects all other commands as well, not only state.sls
21:52 terminalmage JesseC: well if this is a new minion it would not have had a cachedir
21:52 forrest Does it bother anyone else that the 'S' in all the documentation headers is taller than the rest of the characters?
21:52 terminalmage so I removed mine and started a minion
21:52 terminalmage and that's when I was able to reproduce
21:52 jacksontj is there a way to pass down kwargs with compound command execution>
21:53 whiteinge berto-: you may be interested in this project (which uses salt-api): https://github.com/saltstack/pepper
21:53 JesseC terminalmage: gotcha, let me know if you need anything else from me
21:53 terminalmage no prob
21:53 terminalmage thanks for the report
21:53 berto- whiteinge: ah, thanks!
21:53 whiteinge berto-: it's been a tad neglected but the basics are in place to start filling out  :-P
21:53 JasonSwindle terminalmage:  Here is an idea
21:54 JasonSwindle When someone highstates; the master will notify in an IRC room.  Like a github bot.
21:55 sleek joined #salt
21:55 dN0t joined #salt
21:55 sleek hello, i'm just getting started w/ salt .. i have the master running locally and minion running locally, but it appears i can only call "salt-key" as root, and also, i can only run "salt '*' test.ping" as root .. how can i run salt as a normal user (perhaps just he minion part?)
21:55 devinus joined #salt
21:56 ipmb sleek: that's by design
21:56 ipmb you can edit /etc/salt/{master,minion} and change the user
21:56 terminalmage sleek: http://docs.saltstack.com/topics/configuration.html#running-salt
21:56 devinus joined #salt
21:56 ipmb but salt is probably going to need root access to do what it needs to do
21:57 devinus joined #salt
21:57 devinus joined #salt
21:57 sleek interesting .. ok, i'll just sudo stuff then
21:58 ze- would anyone have an idea the average time between a new issue, and some comments about it ?
21:58 forrest usually pretty quick, depends on the issue
21:58 devinus joined #salt
21:58 JesseC ze-: the one I put in earlier was commented and assigned very rapidly, within a couple of hours
21:59 devinus joined #salt
22:00 ze- mine was created about 13-14 hours ago.
22:00 ze- guess i'll have some comments on it by monday :)
22:00 forrest which issue?
22:00 ze- https://github.com/saltstack/salt/issues/6493
22:00 dthom91 joined #salt
22:01 UtahDave joined #salt
22:01 ze- no hurry about it. was just asking to get an idea of what was the overall reaction about such. :)
22:03 JesseC You already submitted the pull request with the patch that repairs the issue though right?
22:03 ze- yup
22:04 ze- at least the patch works on the version we are using.
22:04 forrest Just blame UtahDave ze-
22:04 UtahDave ???   :)
22:05 forrest You hadn't gotten enough spam since you rejoined the channel
22:06 ze- for blame, i'd use git blame to see who perpetrated the bad code ;)
22:06 forrest No I meant actually blame him
22:06 JesseC forrest: adding the proc directory repairs my issue
22:06 JesseC forrest: as a workaround at least, lol
22:06 JesseC */var/cache/salt/minion/proc/
22:06 forrest so the issue was the cache just hadn't been created yet then
22:07 forrest it should be creating it :\
22:07 forrest or at least not crashing out, lol
22:07 ze- oh... btw, do you know if there is a way to install a package within a chroot environment ?
22:07 JesseC yeah, terminalmage was able to repeat it, I'm sure they'll fix it
22:07 forrest JesseC, Yea I saw the conversation earlier.
22:08 forrest ze- , as far as I'm aware there isn't anyways to specify the root directory for a package install
22:08 forrest I could be missing something though
22:08 forrest at least not in the chroot sense
22:08 terminalmage JesseC: https://github.com/saltstack/salt/issues/6502
22:09 ze- i meant something like chroot /.../ apt-get install ...
22:09 ze- already having a debootstrap, but nothing running in it at the time i'd need to finish configuring it :)
22:09 forrest Yea that's what I figured you mean ze-
22:10 forrest *meant
22:10 ze- ok :)
22:10 kermit joined #salt
22:13 __Jahkeup__ joined #salt
22:13 forrest ze-, managing a chroot would be cool if there was a native way to do it
22:14 ze- how are the packages installed in the base system ?
22:14 aat joined #salt
22:14 ze- what is the "native" way to do it ? :)
22:14 forrest I don't know
22:14 forrest The only time I've done major work in a chroot it was super ghetto
22:15 ze- so, it would just need to do the same, but within a chroot :P
22:16 forrest well, when I installed a package into the chroot, it would be using the installroot option from yum
22:18 Aniki joined #salt
22:20 frosty996 joined #salt
22:20 forrest ze- , so I guess if you wanted to keep it that simple, a - root option could be added for pkgs.
22:21 frosty996 getting an error I've never seen before from a local salt-call: https://gist.github.com/jfrost/6143948  Anybody seen this before?  It's 0.16.0-8211~precise1 installed from the PPA.
22:21 forrest does the /var/cache/salt/minion/proc dir exist frosty996?
22:22 frosty996 forrest: nope, that directory structure stops at /var/cache/salt/minion
22:22 frosty996 only extmods in there
22:22 forrest That's probably related to this then: https://github.com/saltstack/salt/issues/6502
22:22 forrest terminalmage already fixed it, try creating the proc dir
22:23 terminalmage /var/cache/salt/minion/proc
22:23 terminalmage create that
22:24 frosty996 yah,I gotcha
22:24 forrest did that fix it?
22:24 frosty996 and sure enough that gets it rolling
22:25 forrest Cool, terminalmage do you know if that's gonna be in 16.3? the milestone has it for 17.0
22:25 kstaken joined #salt
22:25 frosty996 I'll just add the mkdir to my bootstrap script
22:27 jessep joined #salt
22:29 forrest Is there no documentation that transitions people from the state/pillar documentation to actually applying the changes onto a minion?
22:29 forrest or did the link just get left out of the pillar docs
22:30 terminalmage forrest: I think it's more likely that we'll just release a 0.17.0 first, but if not, then it will be in 0.16.3
22:30 terminalmage forrest: I'm also looking into updating salt-bootstrap to create that directory
22:30 forrest terminalmage, ok cool, I noticed the planned release date for 0.17 is pretty close
22:30 terminalmage so at least minions installed with salt-bootstrap will create this
22:30 forrest yea that's a good plan
22:31 bbopen left #salt
22:31 forrest That reminds me that I need to file an issue regarding RHEL and the bootstrap
22:32 JesseC So if you can't use require in include:, what is the proper way to run things in order?
22:33 Gifflen joined #salt
22:33 forrest so you're saying that you need to run the two items in the include in a specific order?
22:33 JesseC forrest: yes
22:33 forrest do the sls files get interpreted top down? Or does it do the puppet style 'I do what I want when I want' style?
22:34 JesseC I think it does what it wants when it wants
22:34 JesseC and not in a linear fashion
22:34 forrest JesseC, so I'd just make it so the require has to run through in a specific order
22:34 forrest so item A requires the first item in your include, then the second item requires item A, and the include
22:34 JesseC you can't
22:34 VertigoRay Do I need to restart salt-master after editing /etc/salt/master?
22:34 JesseC includes can't have require
22:35 JesseC you get a compile error
22:35 forrest not in the includes
22:35 JesseC forrest: https://gist.github.com/JesseC-/1e2be9f453862e1e5848
22:36 JesseC git-scm downloads and compiles git, but requires dependencies to be done first
22:36 JesseC I ran into this issue before as well but didn't give it much thought
22:36 JesseC I had to move my repos include into dependencies
22:36 JesseC that's just not very organized
22:37 JesseC there has to be a better way without having to mush everything together into a single state file
22:37 ze- mmm... how would you handle having a "virtual"/noop state ? aka, a state depending on a few others, and letting some later state depend on
22:38 forrest JesseC, are you unable to basically make staggered deps?
22:38 p3rror joined #salt
22:38 forrest so that item 1 requires lnpp.dependencies, and then item 2 requires item 1, and lnpp.selinux?
22:39 forrest so then lnpp is already 'there'
22:39 forrest *lnpp.dependencies
22:39 JesseC forrest: the only thing in that list that has a requirement would be git-scm needing lnpp.dependencies
22:39 JesseC selinux, iptables, and configs can run whenever
22:39 forrest hmm
22:40 JesseC I could just inlcude dependencies in git-scm, but that defeats the purpose of having it nicely organized
22:41 JesseC because you can require: sls: name from within anything else other than include
22:41 JesseC oh, what about extend
22:41 JesseC I think I read something on that
22:43 hellome joined #salt
22:45 JesseC nm, that isn't used that way I thought it was
22:46 dthom91 joined #salt
22:46 forrest I don't know if the syntax is right on here, and it looks awful: https://gist.github.com/gravyboat/8fcef692f0d35c654a3c
22:46 terminalmage forrest, JesseC, the fix is merged so salt-bootstrap installs of the develop branch will not be affected
22:46 terminalmage I also have a fix submitted to salt-bootstrap that will fix this for people installing 0.16.2
22:47 JesseC oh, good idea forrest, I'll try that right now
22:47 forrest awesome thanks terminalmage, frosty996 you don't need to modify your bootstrap (if you're using the develop branch)
22:47 JesseC terminalmage: thanks alot
22:47 forrest yea JesseC, I don't know if it will work, it looks awful that's for sure
22:47 terminalmage np
22:47 frosty996 forrest: I'm installing from the PPA
22:47 terminalmage sorry guys
22:48 terminalmage frosty996: then you'd need to create /var/cache/salt/minion/proc
22:48 frosty996 you guys always fix stuff so fast though, it's awesome!
22:48 forrest terminalmage: All good, fixed in just a few hour so you can't complain
22:48 frosty996 terminalmage: already modified my bootstrap script. :-)
22:48 JesseC forrest: if it does work, it's better than me having everything in a single file, imo, even if it's not super pretty. :)
22:48 terminalmage yeah, it's just annoying, we put out a bugfix release and then this
22:48 forrest yea let me know, I don't know if I reference git-scm properly or not
22:49 forrest terminalmage, yea understandable
22:49 forrest *referenced
22:52 forrest hey terminalmage, JesseC and I were discussing workarounds where you want ain include to require another include
22:52 forrest is there a good way to do that?
22:52 forrest *an
22:53 terminalmage you mean an sls with an include that includes another SLS?
22:53 doublerr joined #salt
22:53 jacksontj joined #salt
22:53 terminalmage so, a file that includes another file, that includes a third file?
22:53 forrest https://gist.github.com/JesseC-/1e2be9f453862e1e5848
22:53 forrest JesseC wants to do that
22:54 forrest and I proposed maybe that https://gist.github.com/gravyboat/8fcef692f0d35c654a3c would work
22:54 forrest but is there an 'official' way to do it?
22:54 forrest Should you simply be including the dependency in the other item? Then how do you deal with conflicts
22:54 forrest basically we're trying to avoid a file that includes another file, that includes a third file
22:55 forrest Does that make sense?
22:56 JesseC The first gist explains it pretty well visually. Dependencies need to be installed before other items such as compiling git-scm and eventually php, etc.
22:56 forrest so the include would need to live inside lnpp.git-scm?
22:56 forrest oh I thought you were terminalmage JesseC *facepalm*
22:56 JesseC lol
22:57 forrest you can tell it's nearing the end of the work day :P
22:57 JesseC it's well past my work day
22:57 JesseC but then again, this is my hobby as well
22:57 JesseC ha
22:59 forrest were you able to try the extend JesseC?
22:59 JesseC I'm just getting a compile error when I try that extend: "Extension name git-scm in sls lnpp is not a dictionary"
22:59 JesseC trying to figure out why
22:59 forrest ahh ok
22:59 forrest well, maybe I have the naming wron and it needs to be lnpp.git-scm?
22:59 forrest *wrong
23:00 JesseC I tried that as well
23:00 forrest hmm
23:00 forrest wel, most of the extend items actually extend a specific item
23:00 forrest which you think would work for the requires
23:00 forrest oh
23:00 forrest it's probably because there is no requires originally?
23:01 forrest I think that file is blank in your repo
23:01 jacksontj joined #salt
23:01 JesseC I have updates that I haven't committed yet
23:01 forrest is there a require in the original git-scm?
23:02 forrest I don't think that there's an 'append' module :P
23:02 JesseC there's no require in git-scm
23:02 JesseC it just walks through downloading the git source and compiling it
23:02 dthom91 joined #salt
23:02 forrest I bet that's the problem then
23:03 JesseC I already know that including dependencies.sls in git-scm and requiring it will work, but it's not a good solution
23:03 forrest yea
23:03 forrest you don't want to go down a tree of tons of includes
23:03 forrest because then you're 6 files deep and blah blah blah
23:03 JesseC not only that but I have other modules that will require those depends
23:03 JesseC such as php
23:04 forrest right, and you don't want to get an error because of multiple include
23:04 forrest *s
23:04 JesseC the only thing I can think of is to just isntall my dependencies per state
23:04 JesseC I guess that would make sense anyways
23:04 forrest http://docs.saltstack.com/ref/states/extend.html doesn't show anything actually being added, only modified.
23:05 forrest Yea that probably makes sense, but now I'm curious about this
23:05 forrest oh no, extend can add items...
23:05 JesseC maybe a dependency pillar so I can still manage them in one place
23:06 forrest that could work
23:06 forrest JesseC, can you try to do an extend on something other than the requires for git-scm?
23:07 forrest so we can see if you can just add a new item, perhaps it's an issue because either A. it can't be done in the require, or B. since you don't have a require defined, it's screwing it up
23:09 JesseC so I tried to extend the first part of git-scm which is "download-git"
23:09 JesseC and add the require and this is what I got
23:09 JesseC Cannot extend ID lnpp.git-scm in "base:lnpp". It is not part of the high state.
23:10 forrest I have no idea what that means
23:10 JesseC me either, lol
23:10 forrest try just git-scm?
23:10 jacksontj JesseC: i got that once-- and it had something to do with a loop in my state
23:10 jacksontj i had 2 things watching eachother
23:10 JesseC staring contest
23:11 chrisgilmerproj left #salt
23:11 JesseC No dice: Cannot extend ID git-scm in "base:lnpp". It is not part of the high state.
23:14 JesseC uh
23:14 JesseC I think I got it
23:14 forrest what did you change
23:14 JesseC waiting for it to finish or error to find out
23:15 forrest lol ok
23:15 jessep joined #salt
23:15 JesseC I just added the require:\n - sls: lnpp.dependencies to the git-scm sls
23:15 JesseC I was under the assumption that requiring an sls meant it had to be included
23:16 forrest oh no I don't think so
23:16 JesseC that still could be true, it hasn't gotten there yet
23:18 JesseC it failed
23:18 forrest on the require?
23:18 JesseC just said one or more requisites failed, even though lnpp.dependencies was sucessfull
23:18 forrest lame, why can't you require another sls? I guess they can only be part of the include? Hooray for being wrong.
23:18 JesseC even weirder, it ignored my prereq_in
23:19 forrest huh
23:19 JesseC forrest: https://gist.github.com/JesseC-/1e2be9f453862e1e5848
23:19 JesseC added the git-scm/init.sls
23:20 JesseC that last line, remote-old-git was still ran
23:20 JesseC even though install-git failed
23:20 JesseC oops
23:20 JesseC nm, that's right, I think
23:21 JesseC unless it's not considered a pkg
23:21 frosty996 hrmm..did 0.16.0 also make it so I can't use salt:// urls for custom package installs when running a masterless minion?
23:22 JesseC frosty996: are you using --local?
23:22 frosty996 jessec yep
23:22 JesseC when you run the salt-call
23:22 frosty996 jessec: actually, I have file_client: local
23:22 frosty996 set in the minion config
23:22 frosty996 so I'm not actually specifying --local on the command line
23:23 frosty996 and it's definitely processing everything else except that
23:24 frosty996 basically, I'm using a vagrant to test my states before committing and git pulling them on the salt-master
23:24 JesseC can you gist how you're using salt://
23:25 frosty996 sure! https://gist.github.com/jfrost/6144255
23:25 forrest JesseC, I'm headed out for the day, let me know if you figure that out.
23:26 JesseC forrest: yeah I think it's time to go drink a beer and get at this another time
23:26 forrest Hah yea
23:26 * frosty996 is thinking the same thing
23:26 JesseC frosty996: what's your file base set to?
23:26 JesseC frosty996: /srv/salt?
23:27 JesseC frosty996: or /
23:27 JesseC frosty996: because everything salt:// gets appended with that
23:27 frosty996 jessec: close - /vagrant/salt-master-etc/srv
23:27 frosty996 jessec: well, that should be perfect. And it used to work
23:27 frosty996 just seems to have stopped
23:28 frosty996 and if I do: ls /vagrant/salt-master-etc/srv/kernel
23:28 frosty996 I see the files
23:28 JesseC hmm
23:29 frosty996 and they're even named correctly. :-)
23:29 krissaxton joined #salt
23:29 JesseC tried quoting it?
23:29 frosty996 haven't
23:30 frosty996 but I never had to before.  So, do you mean quote the "salt://kernel..." bit in the sls file?
23:31 frosty996 if so, that was also no joy
23:31 JesseC yeah, was just throwing something at the wall
23:31 frosty996 is there an extra debug switch I can throw at it?
23:32 JesseC no clue on that, i've only been at this for two days so far
23:32 frosty996 :-)
23:32 frosty996 I don't think it's a terribly big deal for me because I know that state did work, just wondering if there was an expected change.
23:32 dN0t joined #salt
23:33 JesseC I can tell you that I'm using it for various config files
23:33 JesseC in 0.16.0 and 0.16.2
23:33 JesseC and they work
23:33 JesseC and I'm using salt-call --local
23:33 frosty996 oh, there are lots of other states besides this one, and they all work fine
23:33 JesseC no master
23:33 frosty996 So, I get Succeeded: 32, Failed: 1
23:34 JesseC off topic, is that actual output that you created for testing? because all I get is craploads of spam and random colored output
23:34 JesseC that's a pain to read through
23:35 frosty996 No, that's at the end of the salt-call
23:35 JesseC odd, mine doesn't tell me anything about what succeeded or failed
23:35 JesseC well, it does, but not in an easily readable way
23:36 frosty996 https://gist.github.com/jfrost/6144299
23:36 frosty996 that's how it looks for me
23:36 frosty996 with color and everything
23:36 KyleG joined #salt
23:36 KyleG joined #salt
23:36 JesseC I need that! lol
23:36 frosty996 it's always done it for me
23:36 JesseC maybe it's because I'm calling a single state
23:37 JesseC that has items below it
23:37 frosty996 oh that seems likely
23:37 JesseC hell maybe that's the answer to my problem
23:37 frosty996 I'm calling state.highstate
23:37 JesseC I can just call my dependencies state first then move onto my others that require it
23:37 JesseC by calling them second
23:38 felixhummel joined #salt
23:38 JesseC I'm not even sure what state.highstate does
23:38 frosty996 sudo salt-call state.kernel
23:38 JesseC does it just execute all states in the salt directory?
23:39 JesseC because I've been doing this -> salt-call --local state.sls mystate
23:39 jschadlick left #salt
23:39 frosty996 it's just the main state file that instructs the minions which environment and modules to use
23:39 frosty996 it's usually top.sls, but it's configurable
23:41 frosty996 I gotta run, talk to you later
23:41 JesseC later, good luck
23:41 frosty996 left #salt
23:41 JesseC your issue solved mine. :D
23:53 carmony joined #salt

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