Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-10-14

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

All times shown according to UTC.

Time Nick Message
00:03 winmutt joined #salt
00:03 winmutt grrr on multilines
00:03 winmutt im trying to use a multiline with key: | and then jinja is putting in literal
00:03 winmutt n
00:03 winmutt \n
00:06 kusams joined #salt
00:12 elfixit joined #salt
00:14 thehaven joined #salt
00:16 aparsons joined #salt
00:17 vxitch fixed the iptables things by stopping iptables service on the minion then running
00:17 vxitch that way, when iptables-save runs the output is blank, no more processing errors, and the rules get overwritten with sane things i specify
00:18 vxitch there were some default NAT masquerade rules in there throwing the processor off. no idea where they came from, but they were lurking
00:20 murrdoc is this in EC2
00:20 murrdoc or like hardware
00:22 vxitch nope, this is a VM on a kvm host
00:25 vxitch oh lordy, finally all green and blue
00:31 murrdoc http://i.imgur.com/xK6v11Z.gif?1
00:34 vxitch that would be me, if i could walk (got hit by a car two weekends ago, lol)
00:34 murrdoc on a bike ?
00:34 vxitch motorcycle, yeah
00:35 murrdoc damn
00:35 vxitch yeah, inattentive drivers
00:35 vxitch i stayed on though, just a broken foot
00:36 Cidan what do you ride?
00:36 Cidan also where do you live?
00:36 vxitch 07 sv650s
00:36 Cidan every day here is a fight for my life
00:36 Cidan in LA
00:36 vxitch silicon valley, south bay
00:36 Cidan ah cool
00:36 vxitch bike is fine, thankfully
00:36 vxitch need to sell it as im moving soon
00:37 Cidan which let's face it, is all that matters
00:37 Cidan aw
00:37 vxitch ye :( i like it a lot
00:37 Cidan giving up riding?
00:37 vxitch no way :P
00:37 Cidan hehehe
00:37 Cidan where you moving to?
00:37 vxitch peru for a month or two
00:37 Cidan oh shit
00:37 vxitch then back to the east coast
00:37 Cidan oh shit x2
00:37 vxitch haha
00:37 Cidan where on the east coast?
00:37 vxitch upstate ny, rochester
00:37 Cidan ha
00:37 Cidan neat
00:37 vxitch thats one word
00:37 vxitch nah, i like rochester a lot
00:38 kusams joined #salt
00:38 vxitch certainly different from the bay
00:38 Cidan I prefer the city
00:38 vxitch sf?
00:38 Cidan nah nyc
00:38 murrdoc ny
00:38 Cidan "the city" is always nyc
00:38 vxitch ah, i grew up around there for 12 years
00:38 murrdoc to people from nyc
00:38 vxitch can't stand it :)
00:38 Cidan ha
00:38 Cidan I was born there
00:38 vxitch where'd you live?
00:38 Cidan just moved out of DC though
00:38 Cidan cpw
00:38 murrdoc ./me doesnt get manhattan
00:39 murrdoc socal man
00:39 Cidan central park west, sorry
00:39 murrdoc its where its at
00:39 Cidan ha, I live in socal now
00:39 Cidan Santa Monica
00:39 Cidan love it, :D
00:39 KyleG hellz yeah
00:39 vxitch i bet
00:39 KyleG SoCal life! (I'm in los angeles)
00:39 vxitch i dont wanna go back to the east coast
00:39 murrdoc shiet me too
00:39 murrdoc lets make this salt meetup already
00:39 murrdoc and go get good brews
00:39 Cidan okay
00:39 Cidan we can use our office
00:39 Cidan it's retardedly huge
00:40 KyleG Cidan: Where you working?
00:40 Cidan whisper.sh
00:40 Cidan venice
00:40 Cidan literally on the beach
00:40 murrdoc noice
00:41 Cidan yuh, the work is massive
00:41 KyleG Ah nice, I almost interviewed for them for a dev ops gig I think it was
00:41 KyleG How do you like it
00:41 Cidan dude, please come
00:41 Cidan I run devops/sysadmin/network
00:41 Cidan I need minions
00:41 murrdoc salt minions ?
00:41 Cidan we use salt!
00:41 Cidan lol
00:41 murrdoc apt-get install salt-minions son
00:41 Cidan also the work is... there's a lot of it.
00:42 Cidan it's a fun place, great office
00:43 KyleG hah let me take another look here
00:43 KyleG • A really good sense of humor.
00:43 KyleG listed as a huge plus
00:43 KyleG I haz
00:44 KyleG :D job?! lol
00:44 Cidan lol
00:44 Cidan I wrote that line
00:44 KyleG nice. lol it sticks out
00:44 Cidan dislike unfunnies
00:45 Cidan uhm, send me your resume
00:45 Cidan antonio@whisper.sh
00:45 Cidan if you're serious
00:45 KyleG I need to do some housekeeping, my resume is NOISY right now
00:45 KyleG I'll take your email fix it up and send it to you sometime this week if that's okay
00:45 Cidan ofc
00:45 Cidan by all means
00:46 murrdoc so this beer thing
00:49 KyleG What kind of beer...
00:49 Cidan all the beer
00:49 Cidan you know what i love
00:49 Cidan yummy.com
00:50 Cidan fucking, beer to your door (along with everything else)
00:50 KyleG yeah that's neat
00:50 micah_chatt joined #salt
00:50 KyleG We usually get our groceries delivered by Vons
00:50 KyleG it's like a $10 fee or something small
00:51 KyleG LA: Where you can automate your life
00:51 Cidan more or less
00:52 murrdoc not even close to manhattan
00:52 micah_chatt_ joined #salt
00:52 murrdoc they have people to run your small errands, task rabbit
00:52 murrdoc people to park your car for you
00:52 Cidan same in LA
00:52 Cidan i use the shit out of task rabbit
00:52 Cidan and washio, oh god
00:53 Cidan i can't live without washio
00:53 KyleG I heard they're testing out some service in SFO where you can get ATM withdrawals delivered to you…
00:53 Cidan uh
00:53 Cidan what
00:53 KyleG Like you pay $40 with your card + a fee, and someone comes to you and brings you $40
00:53 Cidan but... why?
00:53 KyleG http://techcrunch.com/2014/10/03/nimbl-is-uber-for-cash/
00:54 ramishra joined #salt
00:55 KyleG oh they serve new york too
00:55 KyleG lol
00:57 murrdoc why ?
00:57 murrdoc cos butlers are better than digital butlers
01:03 bhosmer joined #salt
01:06 vxitch is there a way to get a list of minions affected by a job?
01:07 murrdoc … test.ping
01:07 vxitch from a previously run job, i mean
01:07 vxitch like i run salt '*' state.highstate
01:07 vxitch would be nice to get a list of minions with success/fail for each
01:07 vxitch like current local returner, but without the details
01:08 murrdoc oh
01:08 murrdoc if u have the jid
01:09 murrdoc is that in the cache dir ?
01:09 murrdoc dude good question
01:09 murrdoc uh assuming dude , no disrespect
01:10 vxitch hm? i do have the jid, yeah
01:10 murrdoc salt-run jobs.lookup_jid jid
01:10 murrdoc what does that give u
01:10 vxitch all the details of every minion
01:11 murrdoc ah fuck
01:11 vxitch haha
01:11 Ryan_Lane hm. is it not possible to call salt modules from custom grains?
01:11 Ryan_Lane I guess some of the modules require grains to load
01:11 vxitch do you guys use a returner to a db and have an app to access info about runs?
01:11 cberndt joined #salt
01:12 murrdoc what about print_job
01:12 murrdoc salt-run jobs.print_job <jid>
01:12 murrdoc then u can iterate that
01:13 murrdoc vxitch:
01:14 vxitch havent used that before
01:14 murrdoc it *should* have the job info
01:21 Mso150 joined #salt
01:23 bigl0af joined #salt
01:23 perfectsine joined #salt
01:23 ramishra joined #salt
01:24 murrdoc vxitch:  well does it
01:25 vxitch sorry! coworker came by and we got to talking about bikes and cars :)
01:25 murrdoc women bro
01:25 murrdoc :D
01:25 vxitch yeah print_job has info about each minion and state that ran on em
01:25 murrdoc ah sweet
01:25 vxitch but it has even more info and would probably be easier to process the lookup_jid output
01:25 murrdoc well
01:26 murrdoc you can set an outputter
01:26 vxitch how do you handle running jobs and job output?
01:28 murrdoc my rollout is small
01:28 murrdoc (big company, have to inch forward)
01:29 murrdoc but for the one pop i did get it out, i put a collectd returner so minions do a +1 everytime they run
01:29 vxitch ah!
01:29 murrdoc and wrote a worker thing on like eventlistener.py that i watch live
01:29 vxitch anyone here use halite?
01:29 murrdoc and halite to run jobs on a cron
01:30 murrdoc but then i stopped that
01:30 murrdoc and upgraded to .7
01:30 murrdoc and am using the scheduler thing
01:30 iggy I _really_ want to add some extra stuff to halite
01:30 vxitch halite status is pre-alpha
01:30 murrdoc there is also pepper
01:30 vxitch well, knowing the stability of the main salt status
01:30 vxitch heh
01:32 murrdoc aight gotta head out
01:33 vxitch thanks for the help
01:33 murrdoc basically vxitch salts a dope remote execution thing
01:33 murrdoc and its pluggable for a lot of stuff u want to do
01:33 murrdoc not sure if that helps
01:33 murrdoc :D
01:33 vxitch yeah, ive been toying with it for a while (about 10 months now)
01:34 vxitch ive been wanting to get involved with dev for it
01:34 vxitch but, dont know how
01:34 vxitch halite looks promising, and fills a need of mine
01:36 malinoff joined #salt
01:39 vxitch halite looks deader than mj. is it still active?
01:40 vxitch intothesaltmine.org isn't maintained anymore, is it?
01:41 iggy halite isn't getting as much attention right now with the release in the works
01:45 iggy I suspect it'll pick back up aftr the release
01:45 iggy they have some jobs open too, that might be one of them
01:45 Mso150 joined #salt
01:45 vxitch theres a qa engineer and technical writer position open
01:45 vxitch two positions technically
01:46 vxitch praying to FSM those get filled soon :) haha
01:46 iggy there was like 8 open last I looked
01:46 vxitch yeah there are a bunch
01:46 iggy there were
01:46 ajolo joined #salt
01:46 iggy jeez... starting to sound like my hillbilly sister-in-law
01:46 vxitch haha
01:48 vxitch does halite have a roadmap?
01:48 vxitch im willing to send PRs if they'll be paid attention to (i see 3 open ones now, from many months ago)
01:55 kusams joined #salt
01:55 beneggett joined #salt
02:02 Pixionus joined #salt
02:03 perfectsine_ joined #salt
02:07 dccc joined #salt
02:11 otter768 joined #salt
02:11 iggy I don't know
02:12 iggy file a couple easy ones and see if anybody takes notice
02:21 Nexpro1 joined #salt
02:31 tmh1999 joined #salt
02:36 aw110f joined #salt
02:44 druonysuse joined #salt
02:44 druonysuse joined #salt
02:46 aw110f joined #salt
02:46 heise_ joined #salt
02:46 kusams_ joined #salt
02:46 dh__ joined #salt
02:47 bhosmer joined #salt
02:47 eofs_ joined #salt
02:48 jamesf_ joined #salt
02:48 moderation_ joined #salt
02:48 jeremyb_ joined #salt
02:48 mattl_ joined #salt
02:48 capricorn_one joined #salt
02:48 basepi_ joined #salt
02:48 wavis_ joined #salt
02:48 abele_ joined #salt
02:48 manytrees_ joined #salt
02:48 Xiao joined #salt
02:48 bytemask_ joined #salt
02:49 a1j_ joined #salt
02:49 balltongu_ joined #salt
02:49 antonw_ joined #salt
02:49 thunderbolt_ joined #salt
02:49 NVX joined #salt
02:49 rogst_ joined #salt
02:49 fxdgear_ joined #salt
02:49 kalail_ joined #salt
02:49 pjs_ joined #salt
02:49 csa_ joined #salt
02:49 CryptoMe1 joined #salt
02:49 iggy_ joined #salt
02:49 MTecknol1gy joined #salt
02:49 imanc_ joined #salt
02:49 copelco_ joined #salt
02:50 pviktori joined #salt
02:50 ikanobori_ joined #salt
02:50 oncallsu1ks joined #salt
02:50 Pork__ joined #salt
02:50 gamingrobot_ joined #salt
02:50 PI-Lloyd_ joined #salt
02:50 mortis__ joined #salt
02:51 esharpmajor_ joined #salt
02:51 jcristau_ joined #salt
02:51 arapaho_ joined #salt
02:51 DenkBret1l joined #salt
02:51 crashmag_ joined #salt
02:51 _sifusam_ joined #salt
02:52 whitepaws_ joined #salt
02:52 bhosmer joined #salt
02:52 vividloop_ joined #salt
02:52 horus_plex joined #salt
02:53 SaveTheRb0tz joined #salt
02:53 jacksontj_ joined #salt
02:53 rackermanfred joined #salt
02:54 alainv_ joined #salt
02:54 bryguy joined #salt
02:54 spader joined #salt
02:54 cb joined #salt
02:54 elkektetet joined #salt
02:54 agj joined #salt
02:54 ThomasJ|d joined #salt
02:54 georgemarshall joined #salt
02:55 \ask joined #salt
02:56 Twiglet joined #salt
02:56 crane joined #salt
03:00 pppingme joined #salt
03:00 andredieb joined #salt
03:00 N-Mi joined #salt
03:00 gladiatr joined #salt
03:00 Alan_S joined #salt
03:00 N-Mi joined #salt
03:00 zekoZeko joined #salt
03:01 jxqz joined #salt
03:01 drybjed joined #salt
03:02 GrueMaster joined #salt
03:02 kavakava joined #salt
03:03 programmerq joined #salt
03:05 ramishra joined #salt
03:05 vukcrni joined #salt
03:06 xDamox joined #salt
03:07 dude051 joined #salt
03:08 iggy how do you guys handle files that need 2 edits? I'm kind of tired of seeing it getting overwritten, and then edited every highstate (the file isn't actually changing)
03:08 thedodd joined #salt
03:09 malinoff iggy, it is better to manage the whole file under configuration management tool
03:09 wendall911 joined #salt
03:09 iggy I think you misunderstood
03:10 iggy so we have a template file with a block replace
03:10 iggy so the template gets laid down first, then the block replace takes place as a separate step
03:11 iggy the template isn't changing and the block replace isn't changing
03:11 iggy but every time, it lays down the template file and then does the block replace
03:12 malinoff iggy, i definitely understood you. My suggestion is to not use blockreplace at all, but use template engine to modify content the template depending on the host once
03:13 malinoff If you think of it, blockreplace is not a configuration management, it is a scripting. And you're trying to mix scripting and configuring in a single group of states
03:14 iggy o.O
03:14 malinoff ?
03:15 iggy "blockreplace is a scripting"
03:15 malinoff Yes, you can think of it as a wrapper around sed
03:15 iggy I can think of a lot of things as a wrapper around sed
03:16 malinoff The difference is that file.managed is "make sure the file is in the specified directory and has the specified content". file.blockreplace is "find the given block and replace it with the given block"
03:17 malinoff the second approach i call "scripting"
03:17 iggy we can't easily just throw the block(replace) in the template because of the way this particular files works
03:19 malinoff Sure, but you should investigate how to do that. Because now you should call blockreplace n times after managed if you need to modify the content n times
03:21 nliadm joined #salt
03:21 anotherZero joined #salt
03:21 zz_Cidan joined #salt
03:21 [vaelen] joined #salt
03:21 Cidan joined #salt
03:22 xenoxaos joined #salt
03:22 Vye joined #salt
03:23 Eugene joined #salt
03:23 MTecknology joined #salt
03:23 kaiyou joined #salt
03:24 XenophonF joined #salt
03:24 synical joined #salt
03:24 synical joined #salt
03:24 seanz joined #salt
03:25 XenophonF hola muchachos
03:26 XenophonF any centos 7 users run into service.running states resulting in False, with a comment of "Failed when setting service to start at boot, but the service was already running"?
03:27 XenophonF it's driving me a little batty
03:27 iggy yes
03:27 dccc joined #salt
03:27 iggy never solved it though
03:28 XenophonF maybe i'm using the wrong name of the service?
03:28 XenophonF i'm not sure how i feel about systemctl :-/
03:29 davet joined #salt
03:29 iggy I've seen it on other systemd systems... pretty sure it's just a systemd thing for now
03:29 iggy might be fixed in 2014.7
03:29 XenophonF hopefully
03:30 XenophonF i think that in at least one case, the service name changed between centos 6 and centos 7
03:30 XenophonF 'nfslock' in 6, 'nfs-lock' in 7
03:31 XenophonF although in 7 there's also an nfslock.service def in systemctl so who knows
03:31 XenophonF i'm going to have to rewrite my map.jinja files
03:33 malinoff #linux
03:34 malinoff oops, forgot to /join
03:37 davet joined #salt
03:38 Outlander joined #salt
03:38 XenophonF iggy, do you have states that have to cope with differences between centos 6 and centos 7?
03:39 XenophonF i've taken to rewriting my map.jinja files as follows - https://bpaste.net/show/425df2092f17
03:39 XenophonF would love to know if this looks sane or if there's a better way to do it
03:40 iggy no, I use debian at the new job... no more rhel
03:40 XenophonF btw updating the service name for centos 7 made those errors go away
03:40 iggy good to know
03:40 XenophonF oh ok
03:40 TyrfingMjolnir joined #salt
03:41 iggy that looks fine
03:41 iggy and if it's working...
03:41 XenophonF thanks
03:42 XenophonF well i'm down to two states that don't work on centos 7
03:42 XenophonF that's progress! :)
04:01 TheThing joined #salt
04:10 JordanTesting anyone know of any good write ups on how to use salt in a script? I am sure it is simple but it is eluding me. I have a pyvmomi script that pulls data from vmware, and I want to compare that to grain info from salt, in the same script.
04:12 tligda joined #salt
04:15 tligda joined #salt
04:19 pravka JordanTesting: http://docs.saltstack.com/en/latest/ref/clients/index.html#client-apis
04:31 cberndt joined #salt
04:31 JordanTesting pravka: Thanks
04:32 ramishra joined #salt
04:35 pipps joined #salt
04:35 ramishra_ joined #salt
04:39 pipps joined #salt
04:40 bhosmer joined #salt
04:44 andrej joined #salt
04:47 vxitch left #salt
04:48 felskrone joined #salt
04:51 rap424 joined #salt
04:52 aranhoide joined #salt
04:52 ramteid joined #salt
04:52 linjan joined #salt
05:06 rawzone joined #salt
05:20 deeprave joined #salt
05:22 pipps joined #salt
05:33 calvinh joined #salt
05:34 meganerd joined #salt
05:42 linjan joined #salt
05:44 baconbeckons joined #salt
05:46 ramteid joined #salt
05:46 xDamox joined #salt
05:46 whitepaws_ joined #salt
05:46 schristensen joined #salt
05:46 jeddi joined #salt
05:46 hellerbarde joined #salt
05:46 nihe joined #salt
05:46 sirtaj joined #salt
05:46 jalaziz_ joined #salt
05:46 repl1cant joined #salt
05:46 Sacro joined #salt
05:46 keekz joined #salt
05:46 IOMonster joined #salt
05:46 seblu joined #salt
05:46 micko joined #salt
05:46 tyler-baker joined #salt
05:46 Obrit joined #salt
05:46 MindDrive joined #salt
05:46 muss joined #salt
05:46 SpX joined #salt
05:46 beardo__ joined #salt
05:46 zsoftich_ joined #salt
05:46 patrek joined #salt
05:46 chitown joined #salt
05:46 sxar joined #salt
05:46 xmj joined #salt
05:46 jasonrm joined #salt
05:46 LordOfLA joined #salt
05:46 renoirb joined #salt
05:46 jmccree joined #salt
05:46 jforest joined #salt
05:46 Hipikat joined #salt
05:51 kermit joined #salt
05:57 baconbeckons joined #salt
05:59 jnials joined #salt
06:07 colttt joined #salt
06:10 tmh1999 joined #salt
06:11 linjan joined #salt
06:12 bonezed joined #salt
06:21 linjan joined #salt
06:23 martin_ left #salt
06:24 arapaho joined #salt
06:26 calvinh joined #salt
06:27 malinoff joined #salt
06:29 bhosmer joined #salt
06:33 dotz joined #salt
06:33 dotz Hello.
06:35 baconbeckons joined #salt
06:36 duncanmv joined #salt
06:36 calvinh_ joined #salt
06:37 calvinh_ joined #salt
06:38 trikke joined #salt
06:39 calvinh__ joined #salt
06:39 slav0nic joined #salt
06:45 akafred joined #salt
06:46 mechanicalduck_ joined #salt
06:49 oyvjel joined #salt
06:51 ttrumm joined #salt
06:53 ndrei joined #salt
07:00 tomspur joined #salt
07:06 Mso150 joined #salt
07:07 tmh1999 joined #salt
07:09 gildegoma joined #salt
07:11 fe92 joined #salt
07:11 felskrone joined #salt
07:13 dh__ joined #salt
07:13 ipmanx hi
07:13 ipmanx Was "--retcode-passthrough" added to "salt" command recently?
07:14 dh joined #salt
07:14 ipmanx I'm getting this error from vagrant's salt provisioning trying to run_highstate:   salt: error: no such option: --retcode-passthrough
07:15 ipmanx it's trying to run this line:   salt '*' state.highstate --retcode-passthrough --verbose --log-level=debug --no-color
07:16 flyboy joined #salt
07:19 ipmanx hmm... https://github.com/mitchellh/vagrant/pull/4472 and https://github.com/mitchellh/vagrant/issues/4471 seem to indicate it's a vagrant mistake and the option doesn't really exist
07:19 dotz Where should I keep salt files required for my project? Keep them together with project and copy to /srv/salt when needed? Or, should I create separate repo just for salt?
07:19 dotz What's the best practice?
07:20 packeteer joined #salt
07:22 n8n joined #salt
07:26 spo0nman joined #salt
07:27 student__ joined #salt
07:27 martoss1 joined #salt
07:29 flyboy82 hey guys, anyone here who has previous experience working with orchestration? is Jinja accepted in the orchestration sls?
07:31 baconbeckons joined #salt
07:36 chiui joined #salt
07:38 Mso150 joined #salt
07:39 ninkotech joined #salt
07:48 mitsuhiko joined #salt
07:48 lcavassa joined #salt
07:49 intellix joined #salt
07:55 fe92 joined #salt
07:55 ale_ joined #salt
07:55 flyboy82 In my orchestration sls I have a for loop going through the mine and only on hosts with specific grain it runs an if to discern if it should run one command or another on a specific minion.
07:56 flyboy82 I've put my mine contents and the whole for loop in a python script and when I render the template, the result is exactly what I want it to be
07:56 flyboy82 problem is, orchestrator doesn't even enter the for loop
07:57 diudaros joined #salt
07:57 diudaros good morning guys
07:58 diudaros i am banging my head on the wall here
07:58 diudaros i made the following simple state:
07:58 diudaros php_basic:   pkg.installed:     - pkgs:       - php5       - pgp5-cgi       - php5-dev       - php5-cli       - php5-common       - php-pear
07:58 diudaros oops
07:58 diudaros sorry
07:59 CeBe joined #salt
07:59 diudaros php_basic:
07:59 martoss1 can you pastebin?
07:59 diudaros pkg.installed:
07:59 diudaros - pkgs:
07:59 diudaros yes
07:59 diudaros sec
08:00 diudaros ok , so simple 10 line state
08:00 diudaros http://pastebin.com/D3XixNcs
08:00 diudaros when i run the state from salt master to minion i gt the following error
08:01 diudaros Comment: The following packages failed to install/update: php5-cli, php-pear, php5-common, php5-dev, php5, pgp5-cgi.
08:01 diudaros for some reason the packages are not installed
08:01 diudaros but weird thing is that if i go on the minion and apt-get install those packages...the progress moves on without any issues
08:01 diudaros the packages have no problem to get nstalled
08:02 flyboy82 are you running your salt-minion as something other than root?
08:03 martoss1 the sls looks correct to me. Since you tried it manually, I assume that all package names are correct as well.
08:03 wnkz joined #salt
08:03 diudaros yes the names are correct
08:04 diudaros no i run both minion and master with root
08:04 flyboy82 when you install manually does it ask you to accept a key or asks for confirmation or something?
08:07 diudaros well it asks for yes or no in order to proceed with installation
08:08 diudaros Use 'apt-get autoremove' to remove them. The following extra packages will be installed:   autoconf automake autotools-dev libltdl-dev libonig2 libqdbm14 libssl-dev libssl-doc libtool shtool zlib1g-dev Suggested packages:   autoconf2.13 autoconf-archive gnu-standards autoconf-doc gettext libtool-doc automaken gfortran fortran95-compiler gcj The following NEW packages will be installed:   autoconf automake autotools-dev libltdl-dev libo
08:08 diudaros thats the output when i manually apt
08:10 flyboy82 shouldn't have a problem...
08:11 diudaros i ommited a part where the installer states that there are a dozen of  packages not longer needed and advises t run "apt-auto remove" in order to remove them
08:11 diudaros will try this and the go step by step installing a simgle package from the state...
08:12 aranhoide joined #salt
08:12 diudaros thanks to @ martoss1 for pointing the obvious i indeed found a typo error
08:12 diudaros if you noticed on the pasted state
08:13 diudaros there is a typo error on the fifht line.... (shit happens after 10 hours on the screen)
08:13 diudaros fixed it but still get the same error
08:15 AirOnSkin Good morning, everyone. Say, how does one handle configuration of network interfaces with Salt for 100+ hosts? Some hosts have 2 interfaces, others 20+. Do you create a folder for each host on the master file server or do you somehow loop through existing information (grains?). And what about other files that are different on each host (like a crontab for example)?
08:17 honestly AirOnSkin: looping through grains works
08:18 bhosmer joined #salt
08:18 honestly AirOnSkin: if you want to get per-client parameters that can't be derived from grains you should put those in pillar
08:19 krak3n` joined #salt
08:19 honestly states are for logic, pillar is for data
08:19 AirOnSkin honestly: ok, have to read up on pillars then. thanks for the hint
08:21 babilen diudaros: Are you aware of https://github.com/saltstack-formulas/php-formula ?
08:25 martoss1 @diudaros: pgp5 :-) didn't notice it either…
08:25 diudaros hey martossl
08:25 diudaros no i wasnt aware of that
08:26 diudaros however when i uninstalled the unused packages the installation moved without any issues...
08:26 diudaros this is weird
08:26 diudaros since i run the whole thing on VMs
08:26 diudaros i will restore the state of the vm and try to reproduce....
08:27 agend joined #salt
08:27 babilen diudaros: I'd recommend to make use of the PHP formula in this case as it will be externally maintained and work similarly on a number of different platforms
08:29 diudaros ok noted
08:30 diudaros but i dont think that has to do with specifically php
08:30 diudaros i will try to install a different package and see if it works
08:32 TheThing joined #salt
08:34 cberndt joined #salt
08:34 tinuva joined #salt
08:34 calvinh joined #salt
08:34 flyboy82 ok, found my problem... bug
08:35 flyboy82 I don't think anyone here would know a way to access the mine from master, would you?
08:35 fe92 joined #salt
08:36 flyboy82 from within orch sls file to be precise
08:39 CeBe joined #salt
08:39 calvinh_ joined #salt
08:44 TheThing joined #salt
08:45 agend joined #salt
08:48 fredvd joined #salt
08:49 bhosmer joined #salt
08:50 smcquay joined #salt
08:56 aranhoide joined #salt
09:01 kore left #salt
09:04 TheThing joined #salt
09:05 intellix joined #salt
09:10 CycloHex joined #salt
09:12 CycloHex Hello guys, newbie question :* .. I have salt-master 2014.1.10 installed.. I decided I like the features of the current RC3. Is upgrading as easy as just using the bootstrap as indicated in this channel's topic?
09:19 flyboy82 so no workaround to access mine data from orchestrate runner sls ran on master???
09:21 aquinas_ joined #salt
09:22 HontoNoRoger joined #salt
09:23 HontoNoRoger hello
09:23 HontoNoRoger Can I somehow check package versions in states? I want to check if a specific package is higher that a specified version. If yes I want to alter my configuration file.
09:24 HontoNoRoger is this somehow possible with jinja and pkg.version ?
09:28 aquinas_ joined #salt
09:31 salted_ joined #salt
09:42 msciciel_ joined #salt
09:44 funzo joined #salt
09:49 superted666 joined #salt
09:50 superted666 Quick Q - I've noticed 2014.1.11 has been out a while but hasn't filtered down to Redhat's Epel repository, any way to track this? Feels a bit smoke and mirrors as to when this stuff becomes available through redhat at the mo :(
09:59 istram joined #salt
09:59 gildegoma joined #salt
09:59 salted_ .list
10:03 calvinh joined #salt
10:04 viq joined #salt
10:04 viq joined #salt
10:05 giantlock joined #salt
10:06 bmcorser joined #salt
10:07 bhosmer joined #salt
10:09 CeBe joined #salt
10:13 babilen IIRC there are some books on saltstack scheduled for release soon(ish), but I can't seem to be able to track them down. Does that ring a bell and do any of you guys perchance have a link?
10:14 viq babilen: http://shop.oreilly.com/product/0636920033240.do
10:14 babilen I knew it, but I did not find that one.
10:15 viq with chitown being the author IIRC
10:21 babilen viq: yeah, I even googled for "oreilly saltstack book" but didn't spot it in the results right away
10:22 babilen Should have sprinkled in a "site:oreilly.com" :)
10:22 gyre007_ joined #salt
10:23 gyre007_ hi, I'm using vagrant 1.6.5 and I'm getting the following error: salt: error: no such option: --retcode-passthrough
10:23 gyre007_ anyone has come across this ?
10:28 babilen That came up on the mailing list yesterday, one second
10:30 babilen https://github.com/saltstack/salt/issues/6973 + https://github.com/mitchellh/vagrant/issues/4471 (the latter is still open)
10:30 babilen fwiw, just googled for '"no such option: --retcode-passthrough"'
10:31 babilen I do wonder though why I don't have that problem with vagrant when installing a salt-master
10:40 ggoZ joined #salt
10:44 peno joined #salt
10:49 calvinh_ joined #salt
10:51 peno joined #salt
10:56 bigl0af joined #salt
10:57 saltnoob joined #salt
10:57 wnkz___ joined #salt
11:00 ndrei joined #salt
11:05 badon joined #salt
11:12 CeBe joined #salt
11:12 obimod joined #salt
11:14 ericof joined #salt
11:23 CycloHex I have salt-master 2014.1.10 installed.. I decided I like the features of the current RC3. Is upgrading as easy as just using the bootstrap as indicated in this channel's topic?
11:25 PI-Lloyd_ I wouldn't use a RC release in production tbh
11:26 PI-Lloyd_ but if you really want to, then I would say uninstall the old master and install the RC via the bootstrap
11:26 PI-Lloyd_ make sure you backup your master config just in case as well
11:28 bhosmer joined #salt
11:29 k3ys3r joined #salt
11:34 CycloHex PI-Lloyd_, why is it not good to use an RC in prod? I ran into a few bugs (eg postgres_user state fails), I'd like to prepend text to my salt-managed files (currently have a hack for it, but using file.prepend would be easier)
11:35 CycloHex ok, an rc will contain bugs as well.. but the shadow state look sinteresting as well, as i'd like to expire some users
11:39 scottpgallagher joined #salt
11:44 intellix joined #salt
11:53 ramishra joined #salt
11:54 PI-Lloyd_ CycloHex: agreed some of the new features are very nice and I want to run them as well, but from experience I don't trust RC's... I would rather wait the extra week or two till the stable release :)
11:56 bhosmer_ joined #salt
11:56 CycloHex PI-Lloyd_: Ok, thanks for the info! I'm still in doubt, though :D
11:56 dccc joined #salt
11:58 PI-Lloyd_ If you want to run it in production that's your call, I personally wouldn't, it's too risky to release non-stable software into production.
12:02 CycloHex PI-Lloyd_: I'm a last-year student. My final project is saltstack. So I'm not really running in production.. + the project only lasts until late-november. So I'm unsure whether the stable release will be releases by then :)
12:02 PI-Lloyd_ ahh, well that makes a bit more sense then
12:02 CycloHex ^^
12:04 PI-Lloyd_ if it's not an actual production environment, just a project environment, then I see no major problem with using the RC
12:04 cads joined #salt
12:04 CycloHex ok! So in order to update, I apt-get purge salt-master then bootstrap the rc3?
12:05 CycloHex running debian wheezy*
12:05 PI-Lloyd_ correct... make sure you backup the configs though
12:05 PI-Lloyd_ just in case
12:05 CycloHex ok, thanks man!
12:06 PI-Lloyd_ I have done exactly that in my dev environment and had no problems with the upgrade to the RC on Ubuntu 14.04
12:07 CycloHex Perfect, finishing some pillar data in current env, then upgrading
12:07 CycloHex I won't lose my sls files etc, right?
12:08 CycloHex Or did you mean those files as well with backin up the configs
12:08 PI-Lloyd_ your states should be in git!
12:08 CycloHex they are.. in a local git though
12:09 PI-Lloyd_ lol
12:09 CycloHex Best for me to upload m to github then?
12:09 PI-Lloyd_ throw them on to bitbucket, it's free and private
12:09 PI-Lloyd_ but to answer your question, your states should be fine
12:09 CycloHex ah, private! nice..
12:10 CycloHex ok, thanks for the info/help
12:11 PI-Lloyd_ with git/bitbucket/whatever you can set the saltmaster to use gitfs as well for both the states and the pillar :)
12:12 CycloHex ok, I only forked some formulas through gitfs now.. But I'll definitly have a look at using gitfs for my own pillars and states
12:14 PI-Lloyd_ I would, its so handy not having to "git pull" on the master on every commit :)
12:14 CycloHex Yes, that'd be handy, indeed
12:16 smcquay joined #salt
12:20 krak3n` joined #salt
12:23 bhosmer joined #salt
12:24 CeBe joined #salt
12:24 smcquay joined #salt
12:24 rattmuff joined #salt
12:27 rattmuff Trying to use "import_text" (http://salt.readthedocs.org/en/latest/ref/renderers/all/salt.renderers.jinja.html) to include an os-specific config file "debian.conf" in my "base.conf" file but I'm running into "Jinja syntax error: Encountered unknown tag 'import_text'.; line 1". Can I not use this way of combining files?
12:28 eunuchsocket joined #salt
12:30 rattmuff This is the error I get:
12:30 rattmuff Comment: Jinja syntax error: Encountered unknown tag 'import_text'.; line 1
12:30 rattmuff {% import_text 'debian.conf' as myconf %}    <======================
12:32 krak3n` joined #salt
12:33 bhosmer joined #salt
12:35 gmoro joined #salt
12:36 ramishra joined #salt
12:37 eliasp CycloHex: no need to host your git repos somewhere "in the cloud"… just put them on any SSH-reachable host, create a 'git' user there and place the repositories in its $HOME… then git@yourhost:repository.git will simply work as remote
12:38 eliasp CycloHex: this works perfectly fine for smaller setups, where you can simply authorize via SSH keys
12:39 CycloHex It's ok for my files to be in the cloud, I guess
12:39 CycloHex but thanks for the tip, eliasp
12:41 jaimed joined #salt
12:45 jaimed joined #salt
12:47 favadi_ joined #salt
12:50 vejdmn joined #salt
12:51 micah_chatt joined #salt
12:52 PI-Lloyd_ rattmuff: what version of salt are you running? I've tried it in 2014.7 and it knows the import_text tag (has issues with finding the file though).
12:53 rattmuff PI-Lloyd_: 2014.01
12:53 rattmuff PI-Lloyd_: But I solved it using "include" statements instead
12:53 rattmuff PI-Lloyd_: Thanks :)
12:54 miqui joined #salt
12:55 micah_chatt_ joined #salt
12:55 PI-Lloyd_ ok, fyi that page you linked has "New in version 2014.7.0." for that import stuff
12:57 flyboy82 left #salt
12:59 rattmuff PI-Lloyd_: aha, I was looking for that but couldn't find it. Thanks!
12:59 obimod joined #salt
12:59 PI-Lloyd_ no worries
13:00 vbabiy joined #salt
13:01 cpowell joined #salt
13:05 bhosmer_ joined #salt
13:06 gmcwhistler joined #salt
13:15 wnkz___ Hi, I'm having escaping problems with Yaml and a curl cmd ; can anyone help me ? https://gist.github.com/wnkz/2ae5913048693d83dc2f
13:16 mikeywaites joined #salt
13:16 mikeywaites Hey guys.
13:16 mikeywaites Is there a built in way to have a minion run highstate at a certain time everyday?
13:17 rofl____ schedule
13:17 PI-Lloyd_ mikeywaites: http://docs.saltstack.com/en/latest/topics/jobs/schedule.html
13:17 mikeywaites Cheers!!
13:17 mpanetta joined #salt
13:18 swa2 joined #salt
13:19 kusams joined #salt
13:20 ericof_ joined #salt
13:22 racooper joined #salt
13:22 nitti joined #salt
13:23 bhosmer_ joined #salt
13:24 PI-Lloyd_ wnkz___: have you tried not wrapping the - name: line with quotes? ie -name: curl -X POST --data-urlencode 'payload={"text": "This comes from salt."}'..... etc etc
13:25 spiette joined #salt
13:25 wnkz___ PI-Lloyd_: I've tried so many things now I'm not even sure .. :-/ when i don't have an error ; curl tries to resolve "This", "comes", etc ..
13:25 wnkz___ PI-Lloyd_: let me see..
13:26 wnkz___ PI-Lloyd_: "Unknown yaml render error; line 33"
13:28 calvinh joined #salt
13:28 kusams joined #salt
13:29 kusams joined #salt
13:31 PI-Lloyd_ i've just done  - name: 'curl -X POST --data-urlencode "payload={\"text\": \"This comes from Salt.\"}" https://example.com/services/hooks/foo?token=bar' - seems to work
13:31 wnkz PI-Lloyd_: let me try
13:32 vejdmn joined #salt
13:33 wnkz PI-Lloyd_: <3
13:33 wnkz this was driving me crazy
13:34 PI-Lloyd_ all good?
13:34 XenophonF joined #salt
13:36 wnkz PI-Lloyd_: Bonus question, what if I want to include variables in this ? let's say {{ foo }} ?
13:36 wnkz PI-Lloyd_: in the test payload I mean
13:37 wnkz text*
13:38 kusams joined #salt
13:39 anotherZero joined #salt
13:40 CycloHex I get the following error in my initial highstate: "State postgres_user.present found in sls postgres is unavailable".. I install postgress in the same sls file through cmd.run because the locale is bugged in 2014.1.10. the usercreation fails, but is executed as last state.. When I run another highstate, th euser is set to be created. I'm really not sure how to fix this
13:40 wnkz PI-Lloyd_: nevermind, got it working like a charm ; thank you again !
13:41 PI-Lloyd_ no worries wnkz
13:41 minaguib joined #salt
13:44 calvinh_ joined #salt
13:49 glyf joined #salt
13:52 mikeywaites i just enabled schedule with the following via pillar data for some nodes and it seems to have caused highstate to run continuously http://dpaste.com/3RACRSW
13:52 mikeywaites is when supported for highstate or is my date syntax incorrect or something?
13:57 jnials joined #salt
13:57 MTecknology This is odd... http://dpaste.com/2JD6SHC
13:58 MTecknology ooooh.... weird
13:59 MTecknology salt-call pillar.data now indicates it's an issue in the pillar top file
13:59 CeBe joined #salt
14:01 higgs001 joined #salt
14:01 CycloHex mikeywaites: a lot of the schedule-things are 'new in version 2014.7.0' so 'when might not be in the older versions'
14:01 fredvd joined #salt
14:02 mikeywaites CycloHex:  right, so if its ignored the default would be to just run constantly
14:02 CycloHex No clue, I haven't really been using sheculde yet.
14:03 mikeywaites CycloHex:  ok i missed this in the docs!
14:03 mikeywaites New in version 2014.7.0.
14:03 mikeywaites Frequency of jobs can also be specified using date strings supported by the python dateutil library.
14:03 mikeywaites DOH!
14:03 CycloHex ;)
14:03 housl joined #salt
14:03 DaveQB joined #salt
14:04 CycloHex I used minutes: to do my schedule
14:04 CycloHex but I run a highstate every 15 minutes ^^ kinda excessive, but w/e
14:04 elfixit joined #salt
14:04 dude051 joined #salt
14:05 CycloHex http://docs.saltstack.com/en/latest/topics/jobs/schedule.html#highstates this should work, mikeywaites
14:07 KennethWilke joined #salt
14:08 ajolo joined #salt
14:10 oyvjel1 joined #salt
14:11 iondrip joined #salt
14:12 mikeywaites CycloHex:  yeh cheers. i was hoping for a specific time but this will work ok for now,  i suppose i could just have cron state that runs this command
14:13 CycloHex yes, cron is useable for this, although I don't use this (yet)
14:14 Aimenaims joined #salt
14:15 Aimenaims Hi everybody
14:15 mikeywaites also has anyone had issue with watching file that should trigger a restart not running when the file is first added.  EG i have state watching /nginx/conf.d/  the firstb highstate run does not reload nginx as the files are created rather than updated
14:18 oeuftete joined #salt
14:20 iggy mikeywaites: manage the config file before you start the service
14:21 iggy and yeah, it's been mentioned before
14:21 mikeywaites iggy:  ok thanks!
14:21 kaptk2 joined #salt
14:21 iggy I _think_ there might be a ticket open for it
14:21 rojem joined #salt
14:23 mikeywaites iggy:  so do you mean i need to perhaps touch the file, start the service, then add the nginx.conf using file.managed or can i just file.managed the confs then start the service?
14:23 UtahDave joined #salt
14:24 jmdcal joined #salt
14:25 linjan joined #salt
14:29 pdayton joined #salt
14:29 jergerber joined #salt
14:29 ksoviero joined #salt
14:29 obimod joined #salt
14:29 ksoviero Anyone have any theories on this one? http://serverfault.com/questions/636927/saltstack-returning-too-much-detail
14:30 jmdcal joined #salt
14:31 bhi joined #salt
14:33 iggy mikeywaites: file.managed first, then install the pkg
14:35 rojem joined #salt
14:36 DaveQB joined #salt
14:39 sc` joined #salt
14:40 TyrfingMjolnir joined #salt
14:41 XenophonF joined #salt
14:43 hardwire joined #salt
14:45 rap424 joined #salt
14:46 anotherZero joined #salt
14:47 perfectsine joined #salt
14:49 anotherZero joined #salt
14:51 bhi joined #salt
14:51 mohae joined #salt
14:51 puddingMe joined #salt
14:51 anotherZero joined #salt
14:52 CeBe joined #salt
14:52 ramishra joined #salt
14:53 StDiluted joined #salt
14:54 MTecknology Who handles the packaging for Debian?
14:54 ramishra joined #salt
14:56 thedodd joined #salt
14:56 p2 joined #salt
14:59 jalbretsen joined #salt
15:02 phblj joined #salt
15:02 glyf joined #salt
15:05 NV joined #salt
15:06 ksoviero When running highstate, I get the following output: http://pastebin.com/XsrPPgyG
15:06 sroegner joined #salt
15:06 TyrfingMjolnir joined #salt
15:07 ksoviero My salt master config file is all defaults, and this only started after I ran `salt '*' state.highstate e` (note the accidentally appended 'e')
15:07 Gareth MTecknology: Joe Healy
15:07 hasues joined #salt
15:07 ksalman joined #salt
15:07 MTecknology thanks
15:09 ipmb joined #salt
15:09 CeBe1 joined #salt
15:10 CycloHex Anyone here on develop-branch? I have just updated and deployed a minion through salt-cloud. All of my states executed, although I received 4 strange warnings.:
15:10 CycloHex warnings:
15:10 CycloHex - 'enable' is an invalid keyword argument for 'service.mod_watch'. If you were trying to pass additional data to be used in a template context, please populate 'context' with 'key: value' pairs. Your approach will work until Salt Lithium is out. Please update your state files.
15:11 CycloHex all are like this, ony enable, and service.mod_watch are changed in the warning message. in this particular example, i try to make sure the service.running has enable on true. This worked on the previous release and isn't said to be deprecated or anything
15:12 MTecknology Is it not possible to use -contents: and -context: in the same state?
15:12 PI-Lloyd_ ksoviero: are you running 2014.7 by chance?
15:13 ksoviero [root@platform ~]# salt --version: salt 2014.1.10 (Hydrogen)
15:13 ksoviero PI-Lloyd_: ^
15:14 gothix joined #salt
15:14 yetAnotherZero joined #salt
15:16 iggy ksoviero: you sure you (or someone else) didn't change the outputter?
15:16 beach joined #salt
15:16 ksoviero iggy: positive, I just checked the config file, and it's fine. Also, I'm the only one managing this host.
15:17 iggy that's just a different outputter
15:17 iggy how it got stuck as the default, I don't know
15:17 PI-Lloyd_ I've just tried to re-produce the output you are seeing and cannot do it just by adding 'e'
15:18 ksoviero Where is the outputter set? other than /etc/salt/master?
15:18 ksoviero iggy: PI-Lloyd_ ^
15:19 PI-Lloyd_ the master config should have - output: nested
15:19 PI-Lloyd_ or have the output: line commented to use the default
15:20 conan_the_destro joined #salt
15:20 Xunilrevol joined #salt
15:20 ksoviero PI-Lloyd_: http://pastebin.com/QYav4qGd
15:20 Xunilrevol What kind of things are you guys using salt to automate with currently?
15:20 beach everything
15:21 PI-Lloyd_ Xunilrevol: everything I can
15:21 TyrfingMjolnir joined #salt
15:21 ksoviero Xunilrevol: anything not handled by code deployments
15:21 jngd joined #salt
15:21 Xunilrevol So here is the situation I'm currently in:
15:21 iggy ksoviero: comment both of those?
15:21 beach here it is
15:21 iggy and ewww... cat | grep
15:22 ksoviero iggy: I just uncommented them to see if that fixed it, so they were commented before now.
15:22 kusams joined #salt
15:22 ksoviero iggy: muscle memory
15:22 PI-Lloyd_ you did restart the master after making the changes?
15:22 iggy and there's no output: set in /etc/salt/master.d/*?
15:22 Xunilrevol I work for an IT services company with over 150 customers (and around as many domains).  I'm kind of a new guy here and we currently do NOTHING automated or uniform.  I'm really looking for some structural solutions for my company to leverage it's assets better.  Salt seems perfect for that once I get my head wrapped around it at least
15:22 ksoviero PI-Lloyd_: of course
15:22 PI-Lloyd_ stupid question i know, had to ask it
15:23 ksoviero iggy: [root@platform ~]# ls /etc/salt/master.d:  ls: cannot access /etc/salt/master.d: No such file or directory
15:23 Xunilrevol Has anyone here used salt in that type of an environment?
15:23 beach Xunilrevol: you're in the right place
15:23 intellix joined #salt
15:24 PI-Lloyd_ Xunilrevol: We currently manage over 100 systems across multiple platforms and providers with salt, databases, web services, mail relays, you name it, it's under salt control
15:24 Xunilrevol I'm so jealous.
15:25 PI-Lloyd_ you can do it too :)
15:25 iggy Xunilrevol: I work for a software company maintaining the cloud infrastructure (currently ~100 servers spread across 3 GCE projects/accounts/environments (dev/qa/prod))... we have _nearly_ everything under salt's control
15:25 PI-Lloyd_ salt-cloud + GCE = bliss :)
15:26 iggy I expect we'll probably be running closer to 200 at peak after we actually launch
15:26 KennethWilke salt-cloud + whatever = powa!
15:26 beach PI-Lloyd: I totally agree
15:26 CycloHex Is the state file.prepend already available in the 2014.7.0 RC3? and is it needed for my minion to be on this verison as well?
15:27 iggy CycloHex: yes and yes
15:27 CycloHex iggy: ok, thanks.. then my problem is that the minion is still on 2014.10.1
15:27 iggy 1.10
15:27 beach we need scopes added into salt-cloud for gce though
15:27 mapu joined #salt
15:27 CycloHex yes, 1.10 :D
15:28 iggy beach: scopes?
15:28 beach instance permissions
15:28 PI-Lloyd_ would be nice if salt-cloud could provision and attach GCE disks via the cloud.profiles/providers
15:28 PI-Lloyd_ one command, whole system is done... no faffing about with additional disks
15:28 beach right, I have to create a higher level script to run salt-cloud cmds plus gcloud commands to disk control
15:29 UtahDave joined #salt
15:30 johtso should I be able to use the develop version of the pip state with 2014.7 by copying the modules into _modules and the state into _states?
15:30 vejdmn1 joined #salt
15:30 johtso it will trump the built in state and module right?
15:30 beach I use elasticsearch in gce, and the plugin for autodiscovery requires the instance to have scopes (permissions) to certain pieces of the gce api -- I cannot associate those scopes in a profils
15:30 beach profile
15:31 PI-Lloyd_ beach, are you referring to the service accounts permissions?
15:31 rypeck joined #salt
15:31 Gareth UtahDave: morning
15:32 smcquay I'm trying to get salt-master to start on oracle linux 6. I have authored my upstart file and when I sudo stop salt-master it leaves processes behind; anyone know how to fix this?
15:32 UtahDave morning, Gareth!
15:32 iggy beach: yeah, disk mgmt is one thing I would like to add to salt-cloud if possible (it's one of the reasons we aren't using it)
15:32 glyf joined #salt
15:32 beach yes... gce command is --service_account_scope
15:33 PI-Lloyd_ well, disk mgmt is available in salt cloud, just not via the profiles/providers files
15:33 jngd Hello all, can anyone tell me online tutorial to learn more about salt?
15:33 iggy for now, I wrote our own deployment system that handles everything using the python bindings
15:33 perfectsine joined #salt
15:33 dude051 joined #salt
15:33 beach iggy: I've gotten around it by creating a bash script that executes gcloud cmds to create a disk, salt-cloud to spin up and instance, and then another gcloud cmd to attach the disk
15:33 iggy PI-Lloyd_: yeah, for it to be useful to us, we'd need support in the map for it
15:34 PI-Lloyd_ indeed
15:34 davet1 joined #salt
15:34 iggy I wonder if my company would shit bricks if I published the tool I wrote
15:34 beach permissions though, that's a whole different problem....
15:35 iggy yeah, we are just using service accounts (shared between nodes) for now
15:35 PI-Lloyd_ only one way to find out iggy :p
15:36 beach :)
15:37 UtahDave johtso: yeah, that will work
15:38 johtso awesome
15:38 XenophonF jngd: docs.saltstack.com has a decent intro
15:38 johtso pip wheels are causing me serious pain :(
15:38 ksoviero So... Have we given up on my issue? ;)
15:38 jngd XenophonF, thanks!
15:38 XenophonF jngd: do you have any specific questions?
15:38 johtso and the fix to allow setting no-use-wheels is only in develop
15:38 iggy jngd: aside from the docs, there's ryan lane's blog, intothesaltmine, and the salt subreddit
15:38 CeBe joined #salt
15:38 wendall911 joined #salt
15:39 Gareth this is fun :)
15:39 Gareth [ERROR   ] '_errors'
15:39 Gareth [WARNING ] ** Restarting minion **
15:39 felskrone joined #salt
15:39 iggy ksoviero: yeah, redeploy... it should only take a few mins right?
15:39 ksoviero iggy: Redeploy what?
15:39 iggy everything!
15:39 jngd XenophonF, no I want general information to compare with others choices
15:40 jngd iggy, thanks for the info!
15:40 ksoviero iggy: im... not doing that...
15:41 iggy wuss
15:41 ckao joined #salt
15:41 iggy I do it about once a week just to make sure everything is working as expected
15:41 StDiluted joined #salt
15:42 beach you redeploy all our minions?
15:42 beach *your
15:43 PI-Lloyd_ why the hell not.... it's on GCE so doesn't take long to spin up new instances
15:43 TheThing joined #salt
15:43 PI-Lloyd_ depending of course, on how big the states are for the systems
15:44 beach I do alot, but not on a schedule
15:44 ksoviero PI-Lloyd_: Redeploying would be a pain the the ass for us. There's a lot of application configuration that if outside the scope of Salt, and there's no automation currently for the applications.
15:45 rojem joined #salt
15:45 iggy I redeploy the master and all minions fairly regularly
15:45 PI-Lloyd_ you can automate pretty much anything, even if you need to write some scripts for salt to call during the deployment process
15:46 MTecknology I'm having troubles rendering a specific state. The state and error are here for any interested. http://dpaste.com/3TDK3Q9  It's probably really simple.
15:47 tligda joined #salt
15:47 iggy I don't think you can do that
15:47 ksoviero PI-Lloyd_: meh, this is there cost-benefit tradeoffs come it. I deem it unworthy, er, unnecessary rather.
15:48 iggy MTecknology: I'm pretty sure you can't mix contents and context... you need to use a template file
15:49 MTecknology bummer, works for me, though. Thanks!
15:49 thedodd joined #salt
15:54 dstokes joined #salt
15:54 rypeck Anyone have any experience getting Salt up and running on AIX?
15:54 rypeck Bootstrap script didn't work for my friend.
15:54 dstokes joined #salt
15:55 ramishra joined #salt
15:55 MTecknology I keep wanting to try that on our AIX boxes, but it'll never happen.
15:56 jnials joined #salt
15:57 ksoviero If I reinstall the salt master, will it screw anything up? I assume I'll have to reimport the host keys, but what else?
15:58 rypeck MTecknology: looking at the bootstrap script it looks like it just isn't supported.
15:59 CycloHex ksoviero, I uninstalled my master without any trouble.. I used apt-get remove salt-master.. But before I did that, i cp'ed all my files to another location, just to be sure, but none were affected
15:59 ksoviero ok, thanks
15:59 CycloHex The reason I uninstalled was to upgrade to the rc3 version
16:00 Ozack1 joined #salt
16:06 tafa2 joined #salt
16:07 Gareth rypeck: what about installing manually on an AIX box?  Not using the bootstrap script.
16:07 rypeck Gareth: I'll shoot him the instructions from - http://docs.saltstack.com/en/latest/topics/installation/
16:07 kusams joined #salt
16:08 Gareth rypeck: it would be interesting if it works :)
16:08 rypeck Gareth: someone got it working about a year ago according to the mailing list
16:08 rypeck https://groups.google.com/d/msg/salt-users/bwW0gxzv5Oc/POevyjypeswJ
16:10 Gareth Nice.
16:12 KyleG joined #salt
16:12 KyleG joined #salt
16:12 shoemonkey joined #salt
16:13 kusams joined #salt
16:15 ksoviero I just blew away the entire /var/cache/salt/master dir in the hopes that it would reset whatever is screwed up... no such luck
16:18 iggy I would suspect something in /etc/salt/ before /var/cache
16:18 CeBe joined #salt
16:18 glyf joined #salt
16:21 thedodd joined #salt
16:21 kusams joined #salt
16:26 tafa2 joined #salt
16:27 troyready joined #salt
16:29 baconbeckons joined #salt
16:30 gyre007_ why isn't https://registry.hub.docker.com/u/saltstack/ubuntu-12.04-minimal/dockerfile/ shipping with salt installed is beyond me really...
16:32 VSpike Hm. A few days away from Salt and I fail at most basic task. What, pray, is wrong with this? https://bpaste.net/show/5493cfb47293
16:33 sarlalian joined #salt
16:33 iggy VSpike: what isn't working? error message? any other info?
16:33 aparsons joined #salt
16:34 VSpike iggy: I get "Data failed to compile" "Too many functions declared in state "file" in sls win.iis.logdir
16:34 iggy makedirs: True
16:34 cpowell joined #salt
16:34 iggy - makedirs: True
16:34 iggy to be exact
16:34 VSpike ugh. *facepalm*
16:34 VSpike Thanks :)
16:36 pmcg joined #salt
16:38 aparsons joined #salt
16:40 Mso150 joined #salt
16:42 schimmy joined #salt
16:44 phblj how do people typically automate minion_id naming?  is there a way to have salt-minion call out to script to get that value?
16:44 schimmy1 joined #salt
16:46 iggy we target based on tags/metadata, so the names mean very little in our environment
16:46 iggy (I have one cluster named after transformers characters in alphabetical order)
16:49 scoates_ joined #salt
16:49 kusams joined #salt
16:51 pratap joined #salt
16:53 phblj I'm doing something similar.  Just trying to cut down on the time it takes to deploy a new instance, so I'm pushing as much as possible down to a vm image.  Booting up a new minion with a set minion_id is obviously not a great thing.  Right now my solution is to add an init script that runs before salt-minion, but I wanted to see if there's a built-in way to do this
16:53 phblj random transformers characters would be fine.  I'm just currently booting optimus.prime over and over
16:54 eliasp my minions are named after a defined incrementing pattern based on their computer account name in DNS + ActiveDirectory… MN1221-(C|S)0001…
16:54 iggy phblj: if there is no minion_id file when salt-minion starts it auto creates one based on hostname
16:54 eliasp so the minion name doesn't reflect its function at all besides S = Server, C = Client
16:55 phblj eliasp: do you set their id manually?
16:56 ramishra joined #salt
16:56 eliasp phblj: yes, otherwise I'd end up with the FQDN as the ID which is something I don't want in this environment for some reasons
16:56 eliasp phblj: so I'm only using the hostname instead of the default FQDN
16:57 eliasp phblj: I don't have the bootstrapping automated regarding this, so installing salt-minion is (still) a manual process
16:58 mike25de joined #salt
16:59 aw110f joined #salt
17:04 forrest joined #salt
17:06 meganerd joined #salt
17:07 Ryan_Lane joined #salt
17:07 ttrumm joined #salt
17:08 peno joined #salt
17:10 karunamon joined #salt
17:14 gyre007_ joined #salt
17:14 toastedpenguin joined #salt
17:15 ksoviero turns out there was a few lines in a state file that were causing that issue:
17:15 ksoviero #rabbitmq_management: #  rabbitmq_plugin: #    - enabled
17:15 ecdhe Does anybody do step-through debugging of salt? I'm having an issue with users-formula not creating sudoers.d/ files...  I've triple checked my pillar data and I'd love to see the moment that jinja is refusing to generate my states.
17:15 ksoviero uncommented, and they cause that issue
17:16 masterkorp Hello
17:18 nitti joined #salt
17:20 pipps joined #salt
17:20 Ryan_Lane ecdhe: if you run with log-level = debug, it'll show the generated states and pillars
17:21 spookah joined #salt
17:21 Ryan_Lane you can look at that to see if your states are being generated correctly
17:21 ecdhe Ryan_Lane, those are all looking good; my pillar parses exactly how I'd expect.
17:21 Ryan_Lane you can use this: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.test.html#module-salt.states.test
17:22 Ryan_Lane to make your highstate fail at a specific place with variables printed
17:22 Ryan_Lane if you aren't using a global failhard, you can failhard from the specific state
17:22 Ryan_Lane http://docs.saltstack.com/en/latest/ref/states/failhard.html#state-level-failhard
17:24 masterkorp Any suggestion on how to manipulate data inside a formula?
17:27 Setsuna666 joined #salt
17:27 jnials joined #salt
17:29 Setsuna666 basepi: What are the guidelines for contributing code for a working, but unfinished salt execution/state module ?
17:36 fannet joined #salt
17:36 wnkz joined #salt
17:38 cam_ joined #salt
17:38 Guest82942 hi, I'm seeing some odd behavior when using Salt to manage my NTP service
17:39 babilen Setsuna666: "Finish it" + http://docs.saltstack.com/en/latest/topics/development/contributing.html
17:39 Guest82942 For some reason, the 'changes' and 'comment' field doesnt always return the same result
17:40 Setsuna666 babilen: As of now, I don't have the time to finish it, so that is why I'm asking.
17:41 Setsuna666 babilen: By unfinished, I mean, not all feature are available and some of the code might be rough, but everything is working
17:42 rojem joined #salt
17:44 bastion1704 joined #salt
17:44 murrdoc joined #salt
17:45 rap424 joined #salt
17:45 bastion1704 Hello, is there q equivalent of {% if 'salt-master' in grains.get('sensu') %}
17:45 bastion1704 with mine.get  ? {% if 'salt-master' in salt['mine.get']('*', 'grain.item') %} does not return anything
17:45 murrdoc morning
17:49 Setsuna666 bastion1704: You could do something like this {% set MINE_VARIABLE  = salt['mine.get']('salt-master', 'grain.item') %} and then check if the variable exists and its content is not empty with {% if MINE_VARIABLE %}
17:55 n8n joined #salt
17:56 ramishra joined #salt
17:57 druonysus joined #salt
17:57 druonysus joined #salt
18:00 aparsons joined #salt
18:02 Ryan_Lane Setsuna666: make a PR
18:02 Ryan_Lane assuming it's working
18:02 Ryan_Lane and that you don't anticipate backwards incompatible changes
18:02 aparsons_ joined #salt
18:02 Ryan_Lane I've found that if you submit a module and parts of it are useful, but it's incomplete that people will work on it
18:04 iggy at the very least, maybe send it to salt-contrib
18:04 Setsuna666 Ryan_Lane: Great, thanks, will do. BTW, are you the Ryan Lane from (http://ryandlane.com/) ?
18:05 neilf_____ joined #salt
18:05 Ryan_Lane Setsuna666: I am
18:06 Setsuna666 Ryan_Lane: Just wanted to thank you for your article on Moving away from Puppet. Just started exploring different DevOps tools and your article convinced me to use Salt :)
18:06 antonw_ joined #salt
18:07 Ryan_Lane great, glad it was helpful :)
18:12 pipps joined #salt
18:15 pipps_ joined #salt
18:20 MTecknology I'm trying to now understand best-practices with different branches in a git repo. It seems a generally recommended approach is to use four environments, base, test, dev, and prod. I've been using git HEAD for the base environment. I'm now wanting to have a specific branch for dev and test. Is it sensible to leave the base env for production boxes and then a different branch for test/dev boxes?
18:20 MTecknology Each branch will be a clone of the main so they'll have their own top files. From my understanding, this means that top file will be used.
18:20 nitti_ joined #salt
18:21 MTecknology I guess I'm mostly confused on the whole thing. What the repo 'should' look like, what the config should look like, etc.
18:21 cberndt joined #salt
18:21 eliasp MTecknology: move your top.sls into a separate repo… top.sls from different branches will be merged by saltstack before being evaluated, causing a horrible mess if the differ between your branches
18:22 eliasp MTecknology: and you simply won't be able to keep them in sync all the time between your branches
18:22 thedodd joined #salt
18:22 eliasp so life will be way easier with the top file not in your states git repo
18:22 eliasp EOD - will have to go… sorry ;)
18:23 MTecknology I can handle that. Seems crazy to have a whole repo for a single state, but I can handle that.
18:23 ozzzo joined #salt
18:23 MTecknology or, would it be possible to just remove it from the branches except for head and keep it out of there?
18:24 babilen MTecknology: That wouldn't allow you to perform *any* merges between branches (as you will always merge the "remove top.sls" commit)
18:24 babilen There is a reason why you keep top files in their own repo
18:25 MTecknology good point
18:25 MTecknology okay, I'll do that then
18:25 babilen (it is really quite unfortunate that salt equates branches with environments and that you cannot easily define subdirectories in your repo as environments)
18:25 babilen I never quite understood that design decision
18:26 MTecknology So, the environment is chosen based on what matches in the top file, not by anything else
18:27 ozzzo I am managing a list of files, and I need to require another block before they are copied. Is there a way to group them so that I can use one require statement instead of one per file? Here's what I have now: http://pastebin.com/ArHDPJtM
18:28 MTecknology So if you have base, dev, and prod in that order, if you have a state file in dev, it'll override that file in the base environment, correct?
18:28 MTecknology is that close to accurate?
18:30 ahammond I want to configure the minion id for first start of a minon. I'm writing the minion_id to /etc/salt/minion_id, however on startup that gets overwritten with junk. Is this intended behavior?
18:30 Mso150 joined #salt
18:30 martoss joined #salt
18:31 iggy ahammond: are you sure you are actually writing the correct file?
18:31 aparsons joined #salt
18:31 iggy salt shouldn't touch the minion_id file unless it doesn't exist already
18:31 iggy and define "junk"
18:32 ahammond iggy junk being what it gets from hostname / gethostbyname when there is no minion_id set
18:32 aparsons joined #salt
18:32 ahammond this is part of a formula to fix the minion_id of new servers.
18:32 iggy yeah, sounds like maybe you aren't writing the file you think you are, or the pkg installation is blanking it maybe
18:33 ahammond so, the minion is running and then cleverly overwrites the /etc/salt/minion_id, clobbers caches and restarts salt-minion service
18:33 martoss1 joined #salt
18:34 aparson__ joined #salt
18:34 iggy that doesn't sound like anything I've seen happen
18:35 iggy I would make sure something else isn't touching that file
18:35 ahammond iggy I'll do some more research
18:36 iggy maybe you have a state/formula that is setting it (I think the salt-formula will set it if you tell it to)
18:36 MTecknology okay... this is starting to make mores sense...
18:37 Mso150 joined #salt
18:39 Mso150 joined #salt
18:40 ericof joined #salt
18:40 MTecknology It doesn't seem like it really makes sense to have a 'prod' branch since 'base' uses HEAD and that seems like what should be used for production.
18:40 MTecknology or is that a dumb statement on my part?
18:45 iggy depends... I've used base more like dev, then had qa and prod branches before (merge master -> qa -> prod)
18:45 iggy it all varies by workflow
18:45 iggy I tend to use master as dev since that what most people checkout by default
18:46 babilen A lot of people rename the "master" branch to "develop" anyway.
18:46 murrdoc das crazy
18:47 babilen But both approaches are perfectly fine. You can either merge *from* master or *to* master.
18:49 higgs001 joined #salt
18:53 jalaziz joined #salt
18:57 aw110f joined #salt
18:57 ramishra joined #salt
18:58 pipps joined #salt
18:59 aranhoide joined #salt
19:05 ekkelett joined #salt
19:07 MTecknology iggy: so... now I have a master, dev, and test branch. My top.sls is in a different repo. Now, how do I configure top.sls so that whatever is in the dev branch is used for servers that match '*-dev'?   This is my top.sls currently. http://dpaste.com/29FWHJ2
19:08 iggy if you are using gitfs, it should automatically (iirc)
19:09 MTecknology I didn't make my confusion make sense
19:09 APLU joined #salt
19:09 iggy oh, standard targeting
19:09 MTecknology If the branch name and environment match, it's used, I get that much.
19:10 iggy 'dev':  '*dev*': - alldevstuff
19:10 MTecknology Beyond that, it seems like I need to replicate the entire structure of base
19:10 iggy you may, yes
19:10 MTecknology damn
19:11 iggy we actually have tags set for dev/qa/prod, so we match using that
19:12 iggy we have some stuff that we don't really expect to change in base, then some specific stuff in dev/qa/prod, but we aren't using gitfs at the moment
19:12 pipps joined #salt
19:12 MTecknology I basically want master to be the prod branch, which works fine. Then, if a server matches '*-dev', I want it to use whatever is in the dev branch
19:13 iggy you're going to have to repeat stuff
19:13 iggy it's not exactly the most elegant thing, but it's what you've got to do for now
19:13 MTecknology ...
19:13 * MTecknology heart sinks
19:14 iggy yeah, unfortunately, the paradigm that salt top files were designed around doesn't exactly work flawlessly for everybody
19:14 MTecknology This seems to defeat the point of multiple environments. :(
19:15 kusams joined #salt
19:15 * MTecknology cries
19:18 MTecknology I guess I could live with it if top.sls could be templated, but...
19:18 * MTecknology cries more -
19:18 Cidan yeah actually
19:18 Cidan it's my major gripe with salt top
19:18 Cidan should be changed, imo
19:18 duncanmv joined #salt
19:19 Cidan it more or less makes multiple environments in salt unusable for most real big shops
19:19 Cidan or at least, you end up having to hack around it
19:19 MrThrax left #salt
19:20 Mso150 joined #salt
19:20 MTecknology I guess I sort of understand it, but... no, not really.
19:22 perfectsine joined #salt
19:22 londo joined #salt
19:24 hasues joined #salt
19:24 iggy feel free to file (backwards compatible) solutions in the tracker
19:25 bhosmer joined #salt
19:27 Cidan iggy some day, that would be great
19:28 MTecknology I don't know how that would even look.
19:29 baconbeckons joined #salt
19:29 MTecknology honestly, I guess it makes sense how it is. I don't know how you could sanely improve it without a massive overhaul
19:30 MTecknology I'm considering renaming master to prod and doing away with the base environment except for the top file. Probably wouldn't work as expected.
19:32 pipps joined #salt
19:32 MTecknology iggy: here!  http://dpaste.com/0KH29VJ
19:33 martoss joined #salt
19:34 iggy we didn't find the whole top environments = gitfs branch thing all that useful for us (since we actually have 3 different salt servers)
19:35 iggy so we have kind of our own workflow that suited our needs
19:35 iggy and it'll probably change when 2014.7 drops because then we can use branches a bit more flexibly
19:35 fragamus joined #salt
19:36 martoss1 joined #salt
19:36 murrdoc joined #salt
19:37 MTecknology ya, 2014.7 get's pretty sexy with git
19:38 thedodd joined #salt
19:38 pipps joined #salt
19:40 Supermathie joined #salt
19:41 Supermathie Hey guys! Is there any demo code to show what to do to have a saltstack job on Windows listen for events?
19:44 baconbeckons joined #salt
19:45 aparsons joined #salt
19:46 djstorm joined #salt
19:48 iggy what kind of events?
19:49 Supermathie for my use case, custom events.
19:49 Supermathie "If a process is listening on the minion, it may be useful for a user on the master to fire an event to it" http://docs.saltstack.com/en/latest/topics/event/index.html#firing-events
19:52 ndrei joined #salt
19:54 Cidan 2014.7 AWS states are sexy
19:54 Cidan auto updating dns
19:54 Cidan yes plz
19:55 Ryan_Lane Cidan: have you given it a try yet?
19:55 Ryan_Lane give me any feedback you have
19:55 Ryan_Lane you're talking about the boto_* modules, right?
19:55 Cidan I am
19:55 Cidan but I have not used it yet
19:55 Cidan beta's have burned me before
19:56 Cidan waiting for final
19:56 Ryan_Lane we have more in the pipeline. at least we'll be adding some basic dynamo support
19:56 Ryan_Lane maybe some basic kinesis support
19:56 Cidan I guess the only feedback I can really give right now, is drop the boto_ name
19:56 Cidan kinda not needed
19:56 Cidan but I'm sure it'll work great
19:56 Ryan_Lane well, there were already aws_ names
19:57 Cidan i mean
19:57 Cidan drop it entirely
19:57 Ryan_Lane heh
19:57 Cidan ie: no reason you can't do route53.present
19:57 Ryan_Lane i mostly worried we'd have some conflicts on names at some point. it's also a lot easier to find in the docs with the prefix :)
19:58 ramishra joined #salt
19:58 Cidan hm, I guess, but I think as the project grows, you'll see a normalization anyway and boto_ will be dropped
19:58 Ryan_Lane btw, ELB states can automatically manage route53 cnames for the load balancers :)
19:58 Cidan ooo
19:58 Cidan neat
19:59 Cidan I want to use ELB, but I worry about bursts of traffic
19:59 Cidan so I use haproxy right now
19:59 Ryan_Lane and autoscaling group states can also manage launch configs automatically
19:59 Ryan_Lane it'll create a new launch config and delete the old one
20:00 Ryan_Lane the cloudwatch module also has the ability to find all current alarms and output them as salt states
20:00 Ryan_Lane <3
20:00 Cidan now that's pretty cool
20:00 Cidan good for scripting
20:00 Cidan hehehe
20:01 dotz joined #salt
20:02 aparsons_ joined #salt
20:03 fannet if only the same existed for other cloud providers.
20:05 gothix_ joined #salt
20:10 kusams joined #salt
20:11 younqcass joined #salt
20:13 seanz joined #salt
20:16 Corey UtahDave: Ahoy, sir!
20:17 Corey Here's a fun Saltism today. "If $file exists, install package foo." I'm thinking of how to do this non-horribly; have I missed a built in state?
20:17 Corey In this case, $file is /dev/ipmi0, so managing it isn't going to help any.
20:17 jamesf_ file.exists ?
20:18 fannet if a service is using “watch” to watch a config file that is being created by a JINJA template - will that file be ‘written to’ every time salt runs and if not will the service sill restart when the file changes?
20:18 Corey fannet: If the file changes at all, it will restart.
20:18 Corey fannet: So it depends upon what the jinja is doing.
20:19 fannet its filling in the config based on pillar values which tend to be static
20:20 Corey fannet: So the first time it'll bounce the service, but subsequent runs shouldn't.
20:20 jalaziz joined #salt
20:20 Ryan_Lane Corey: onlyif:
20:21 Ryan_Lane onlyif: test -f <file>
20:21 fannet Corey: thanks
20:21 Corey Ryan_Lane: Good idea, but it's managing several packages and services. Any way to conduct that test in jinja?
20:22 MTecknology Cidan: https://github.com/saltstack/salt/issues/16633
20:22 rlarkin|2 in the salt cloud bootstrap script I see {{minion}} which seems to evaluate to grains?
20:22 nickjj joined #salt
20:23 Ryan_Lane Corey: {% set myvar = salt['file.file_exists']('myfile') %}
20:23 Ryan_Lane {% if myvar %}
20:24 bhosmer joined #salt
20:29 Corey Ryan_Lane: Ooh, nice.
20:33 Corey Ryan_Lane: Worked a treat, thanks!
20:33 Ryan_Lane yw
20:36 utahcon is it possible to check if a state file exists before including it?
20:36 utahcon from within a state file
20:36 Ryan_Lane yes
20:37 Ryan_Lane it's ugly, though
20:37 utahcon Ryan_Lane: I am ok with ugly, as long as she is functional :D
20:37 Ryan_Lane utahcon: https://gist.github.com/ryan-lane/351b7c03bc84830f4109
20:38 Ryan_Lane that's for a top file, but it should be easy to modify that for other locations
20:38 renoirb Hi all
20:38 utahcon thanks Ryan_Lane
20:38 Ryan_Lane yw
20:38 Ryan_Lane renoirb: howdy
20:38 renoirb Which version of libcloud works under Ubuntu 14.04 for salt cloud?
20:38 renoirb hey Ryan_Lane !
20:38 renoirb When I try to run `salt-cloud --list-images all`  I get a stack dump
20:39 renoirb ``Failed to get the output of 'openstack.avail_images()': <MalformedResponseException in None 'Malformed response'>: 'code: 400 body: {"error": {"message": "get_version_v2() got an unexpected keyword argument \'auth\'", "code": 400, "title": "Bad Request"}}'``
20:39 iggy renoirb: depends more on your provider (from what I've seen)
20:39 renoirb openstack
20:39 iggy we had to roll back to the previous release (before 0.17) because it had issues with gce
20:41 renoirb gce?
20:41 iggy el goog
20:41 renoirb google compute engine. got it.
20:41 iggy (google compute engine)
20:45 jaimed joined #salt
20:46 fannet in a file managed directory is there a way to enforce no other files except the ones specified ?
20:50 aparsons joined #salt
20:54 fragamus joined #salt
20:55 murrdoc joined #salt
20:56 murrdoc i got $5 for salt rc package for ubuntu 12/14
20:56 murrdoc kthnx
20:56 bhosmer joined #salt
20:56 iggy fannet: there is with file.recurse
20:56 fannet file.recurse = death
20:57 fannet waaaay too much IO going on w/ that thing and brings salt execution to a crawl
20:57 chrisjones joined #salt
20:58 murrdoc what are u trying to push
20:58 iggy then no
20:59 ramishra joined #salt
20:59 fannet murrdoc: I had a content folder for nginx w/ about 10,000 source files. it was brutally slow doing hashes on every file on the saltmaster on every run
20:59 fannet ever since I’ve been terrified of using file.recurse
21:00 murrdoc ah
21:01 murrdoc 10000 is a big number
21:01 murrdoc can u put them in git ?
21:01 fannet is it though? Our production puppet deployment has hundreds of thousands of managed files
21:01 druonysuse joined #salt
21:01 druonysuse joined #salt
21:01 murrdoc how do u do it in puppet
21:01 iggy oh cool, pdf generation appears to be fixed (on readthedocs)
21:02 murrdoc totally srs question
21:02 murrdoc how did u do it in pupppet
21:02 fannet haha let me ask my puppet engineers
21:02 Ahlee whiteinge: happen to be around?
21:03 murrdoc i miss puppet
21:03 murrdoc but i used git archive for any big bunch of files
21:04 fannet murrdoc: I’m thinking thats the way to do it
21:05 murrdoc its what i recommend :D
21:05 iggy salt.states.schedule docs should say that it is new in 2014.7 (to save me 4 hours of hair pulling)
21:05 fannet but for my other issue of enforcing only my files I guess I can use file.recurse for ‘small’ folders
21:06 fannet lol so apparently thats how our puppet folks are doing it :-/
21:06 murrdoc :D
21:06 murrdoc CALLED IT
21:07 murrdoc enforcing only your files ?
21:07 murrdoc throw a manifest file in there
21:07 murrdoc file path to md5 map
21:07 murrdoc run md5sum -c filename
21:07 murrdoc tada
21:07 murrdoc iggy salt states schedule is totally .
21:07 murrdoc .7
21:08 fannet i guess the question is if salt is using git as the FS why not just rely on git to push the file changes for static files in a recursive folder
21:08 murrdoc is there a way to serve from s3 ?
21:08 murrdoc git cant push
21:08 murrdoc salt gives u remote command execution
21:11 iggy murrdoc: I know that... now
21:12 MTecknology Is it possible to have an environment without any states being applied in it? My case, I want 'base' to be empty and only exist because of top.sls, then rename everything to prod
21:12 Ryan_Lane fannet: if you're just wanting to include specific files, why not just include those files via file.managed?
21:12 iggy but in the absence of easy to access docs for older versions and the "latest" docs not saying that, I had to go look at code
21:12 Ryan_Lane rather than manage the directory and only use an inclusion pattern?
21:12 Ryan_Lane you can have a list of files in a jinja list and iterate over the list, creating states for each one
21:13 fannet Ryan_Lane: two seperate problems 1) managing 10k + files across web servers and 2) managing folders like /etc/nginx/sites-enabled to ensure it contains no other files but the ones I manage
21:14 Ahlee is there a way to expose keys accepted via the API?
21:14 Ryan_Lane fannet: I wrote something recently that does that
21:14 Ahlee don't want teh actual key values, just hostnames
21:14 Ryan_Lane you can use jinja to set a variable that uses the file module to list files in a directory
21:15 Ryan_Lane if a file isn't one of the files you want to manage, you file.absent it
21:15 Ryan_Lane via a jinja loop
21:17 TyrfingMjolnir joined #salt
21:20 ajolo UtahDave: hey !
21:20 MTecknology How do I rename master to prod and get rid of master completely?
21:22 murrdoc in git ?
21:22 murrdoc or in salt environment
21:23 MTecknology git
21:23 murrdoc git checkout master
21:23 murrdoc git pull
21:23 murrdoc git checkout -b prod
21:23 murrdoc git push —set-upstream origin prod
21:24 murrdoc git branch -r master (?)
21:24 murrdoc oh git branch -d master
21:24 murrdoc git push origin :master
21:24 murrdoc PEW!
21:25 MTecknology heh... that's what I tried - I wound up with this, http://dpaste.com/3TYXVGQ
21:26 murrdoc oh heh
21:26 sarlalian joined #salt
21:26 iggy in github, go to settings, and change "Default branch"
21:27 murrdoc or
21:27 murrdoc git branch -M prod master
21:27 murrdoc git push
21:27 murrdoc coffee break
21:28 MTecknology this is in gitlab
21:28 MTecknology apparentnly the exact same thing applied
21:28 MTecknology s/ied/ies/
21:29 MTecknology worked perfect!!! :D
21:29 murrdoc point iggy
21:29 kusams joined #salt
21:29 iggy \o/
21:29 Ahlee anybody know if keys are exposed via lowstate?
21:30 MTecknology point: both
21:30 TheThing joined #salt
21:30 MTecknology Ahlee: knee-jerk answer: no
21:30 MTecknology I very much doubt it .. but don't quote me
21:33 Maik_ joined #salt
21:33 nitti joined #salt
21:34 fragamus joined #salt
21:36 peters-tx joined #salt
21:36 ggoZ joined #salt
21:36 primechuck joined #salt
21:38 kermit joined #salt
21:39 vejdmn joined #salt
21:40 MTecknology How can I have an empty 'base' environment so file_roots: base: still applies to all environments?
21:41 MTecknology Do I just need to have some empty sls file that applies to all boxes?
21:43 UtahDave hey, Corey!
21:43 UtahDave hey, ajolo!
21:43 UtahDave MTecknology: Yeah, that's what I do.  I create a file named empty.sls and have that match on the base environment on '*'
21:44 MTecknology thanks! :D  I feel moderately more sane now
21:45 murrdoc :)
21:45 MTecknology however, now I get this... http://dpaste.com/0WYXB60
21:45 Ryan_Lane1 joined #salt
21:46 MTecknology I have to be insanely close to having this working correctly.
21:46 mohae joined #salt
21:46 murrdoc anyone used s3 as a file store with salt ?
21:49 murrdoc well alright then
21:52 gyre007_ joined #salt
21:53 MTecknology Any ideas what I did wrong? If you have a moment? I'm trying to have test/dev/prod branches in git repos. Those have states only, no top.sls. They also have master branch. In another repo, I have top.sls and null.sls. The top.sls file was printed in the debug output in the last pbin (http://dpaste.com/0WYXB60).
21:55 MTecknology -l trace doesn't seem to add anything else of value
21:55 vu_ joined #salt
21:56 MTecknology UtahDave: also - HI!!! :D
21:59 intellix joined #salt
22:00 ramishra joined #salt
22:01 kusams joined #salt
22:01 MTecknology or does 'null' have a special meaning?
22:02 MTecknology aaaaahhh.......
22:03 ajolo UtahDave: hey man
22:03 ajolo salt \* grains.items | grep -E 'osfullname|osrelease: '
22:03 ajolo osfullname: Slackware
22:03 ajolo osrelease: 14.1
22:04 MTecknology that was it! :D
22:04 ajolo It took some time but I could finally make the time to do it. Now I need to find the time to hack a little bit into slackpkg/sbopkg modules :)
22:05 MTecknology heh... but somehow, I still have conflicts. :S
22:05 vu joined #salt
22:05 MTecknology ah, frick... because 'prod' and 'dev' still conflict. (not the only reason)
22:06 MTecknology damnit! Why can't salt just pick the sls file from the last matching environment? :(
22:06 NaPs joined #salt
22:07 NaPs hi
22:08 NaPs is it possible to write a top.sls file using the pydsl renderer? can someone show me an example?
22:09 druonysus joined #salt
22:09 druonysus joined #salt
22:11 pipps joined #salt
22:12 MTecknology I have four conflicting states.    The conflicting ID is "at" and is found in SLS "base:sys.packages.salt-minion" and SLS "prod:sys.packages.salt-minion"   However, base is empty.  :S
22:12 pipps_ joined #salt
22:13 kusams joined #salt
22:13 MTecknology It's apparently because I'm using -include: sys.packages.salt-minion
22:18 MTecknology yup, any -include tries to pull from base instead of the environment that called it
22:18 MTecknology UtahDave: Is this bug?
22:19 UtahDave MTecknology: The minion will pull in whatever sls files from whatever environments it matches in.
22:20 chrisjones joined #salt
22:20 UtahDave MTecknology: huh, that's interesting. That should probably pull from the same environment.
22:20 UtahDave ajolo: Nice! that's awesome.
22:21 ajolo UtahDave: I need to dust off my python skills, but as far as Salt installation on Slackware, it was a piece of cake between sbopkg and pip
22:21 UtahDave that's great to hear!  :)
22:21 ajolo UtahDave: besides pkg management, is there any other module that is distro-related ?
22:22 UtahDave service.py, possibly system.py
22:22 ajolo Yes, I'll put together a how-to and maybe some slackbuilds
22:22 ajolo Right
22:22 UtahDave uh, there's probably a few others
22:22 ajolo Well, I'm a little busy until the end of october, so I'll start digging into that on november
22:22 UtahDave that's cool.
22:23 UtahDave I really appreciate what you've done so far!  That will be cool to have slack support
22:23 ajolo :D
22:23 MTecknology UtahDave: I guess it's a bug then. I'll file an issue.
22:24 UtahDave cool, thank, MTecknology.  I do think you can specify the environment in the include if you need to hard code it.
22:25 MTecknology that would be a good interim fix
22:26 mosen joined #salt
22:27 aparsons joined #salt
22:29 aparsons_ joined #salt
22:30 karunamon left #salt
22:31 MTecknology UtahDave: I just realized that I have to make that change in >200 places... :(
22:32 Ryan_Lane joined #salt
22:32 UtahDave MTecknology: ouch.
22:33 MTecknology UtahDave: ah, I screwed up.
22:33 UtahDave MTecknology: just checked the docs. Salt should include the file from the same environment
22:33 UtahDave If it's not, it's definitely a bug
22:34 MTecknology I somehow copied a file that shouldn't have been there. Completely my oops.
22:36 UtahDave ah, so it is working as expected?
22:36 MTecknology well... now that part is
22:36 MTecknology So, if I have a top.sls file that looks like this - http://dpaste.com/3VN6W51 - and this in my master config file - http://dpaste.com/20FCRM8 -  then any states in any environment should find whatever is in the base file_root, right.?
22:37 jnials joined #salt
22:40 ggoZ joined #salt
22:40 MTecknology UtahDave: or am I wrong on that? If so, then I might really start crying.
22:41 UtahDave no, you need a  'prod:' entry in your file_roots for the things that match in the 'prod' environment in your top file
22:42 * MTecknology grumbles
22:43 MTecknology This whole environments thing doesn't seem to really be very useful. :(
22:44 MTecknology soooo much redundancy
22:44 UtahDave If I only need to swap out certain jinja variables between my different environments, then I usually just use the base environment for my file_roots
22:44 UtahDave and then use environments in my pillars to choose the right variables
22:45 MTecknology now I have this - http://dpaste.com/37ZQHHJ
22:47 MTecknology I guess I may as well just entirely remove the base environment. I really really wish top.sls would just pick the latest matching sls file and override things from previous environments. This would be sooooo much easier.
22:47 UtahDave why do you have prod, dev and test all providing files from the same directory
22:47 UtahDave ?
22:48 MTecknology because those won't ever change. It's the git repo that has the different branches that will have changes
22:49 whiteinge Ahlee: (belated) pong
22:49 MTecknology It's actuall just a pile of large chunks of data. Some of the files in there are >1GB.,
22:51 UtahDave MTecknology: ok
22:51 MTecknology UtahDave: maybe I've just been spoiled by how amazing salt has been to me. It's been incredible, flexible, simple, and has done everything as I expected. This is the first point where my mind is saying, hey... this doesn't make sense.
22:52 MTecknology I didn't this part, though (.. or any part). I'm sure a massively different mindset was in place at that point. My use case is just different than anything that was intended.
22:52 UtahDave Yeah, I can see that.  I've been thinking of ways to have an alternative to top files
22:54 MTecknology I actually like how it works. I just want slightly more flexibility.
22:54 MTecknology https://github.com/saltstack/salt/issues/16633 <-- if you didn't notice it
22:55 MTecknology I could try to work out in more details exactly what I think would make the most sense, but there's no chance I could ever actually develop it myself. Waaaay outside of my capabilities.
22:55 MTecknology Anyway! Time to go! ... way past time to go.
22:57 badon joined #salt
23:00 ramishra joined #salt
23:03 whiteinge Ahlee: looks like you were asking about getting keys from salt-api? have you seen the key.list_all function using the wheel client?
23:04 Ozack1 joined #salt
23:05 ajolo joined #salt
23:08 elfixit joined #salt
23:10 aw110f joined #salt
23:12 murrdoc joined #salt
23:17 rzzldzzl joined #salt
23:18 murrdoc joined #salt
23:20 rzzldzzl Hi - I was hoping someone could get some help with targeting minions in the top.sls file
23:21 peno joined #salt
23:21 rzzldzzl my top.sls file looks like this:
23:21 rzzldzzl base:     '*':       - users       - duo_unix       - sshd_config       - newrelic-sysmond       - splunk_uf
23:22 rzzldzzl however I want to exclude installing splunk_uf on any host with splunk in the anme
23:22 rzzldzzl name
23:22 rzzldzzl this logic works on the command line: salt -E '^.+((?!splunk).)*$' test.ping
23:22 rzzldzzl but if I put '^.+((?!splunk).)*$' in top.sls it does not work as expected
23:23 CeBe joined #salt
23:23 n8n joined #salt
23:27 jalaziz joined #salt
23:29 druonysuse joined #salt
23:29 UtahDave rzzldzzl: you'll need to add   - match: pcre
23:30 pipps joined #salt
23:31 rzzldzzl ah.  thank you very much
23:31 rzzldzzl duh
23:33 UtahDave did that work?
23:36 pipps joined #salt
23:44 troyready joined #salt
23:44 seanz left #salt
23:44 rzzldzzl no :/  but I suspect it is something I am doing.  Still iterating on it.
23:44 rzzldzzl vdev-splunk:     '^.+((?!splunk).)*$':       - match: pcre       - splunk_uf
23:44 rzzldzzl sorry for formatting
23:46 rzzldzzl I tried this as well with no luck: https://github.com/saltstack/salt/issues/2053
23:49 rzzldzzl I had the regex wrong. :/  All is working as expected
23:49 rzzldzzl Thanks again!
23:49 bigl0af joined #salt
23:49 rzzldzzl ^.((?!splunk).)*$
23:57 superseb joined #salt
23:57 packeteer joined #salt

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