Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-04-18

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

All times shown according to UTC.

Time Nick Message
00:04 timoguin vlcn: good to here. :)
00:04 TyrfingMjolnir joined #salt
00:04 timoguin s/here/hear/
00:06 anuvrat joined #salt
00:13 yomilk joined #salt
00:18 elfixit joined #salt
00:20 linuxlew_ joined #salt
00:22 jimklo joined #salt
00:26 vejdmn joined #salt
00:31 snuffeluffegus joined #salt
00:38 yusuket joined #salt
00:41 jeremyfelt joined #salt
00:44 tristanmatthews joined #salt
00:47 tristanmatthews Hi, I'm just getting started with salt and reading through the docs. I'm trying to make sure the table of contents is supposed to go 3.1.1 -> 3.4.6 -> 3.2.2
00:49 UtahDave tristanmatthews: lol.  what's the url you're looking at?
00:49 tristanmatthews http://docs.saltstack.com/en/latest/topics/tutorials/index.html
00:49 UtahDave we recently did a pretty big overhaul of the docs.  There's probably a few things that need to be tidied up
00:50 timoguin UtahDave broke it.
00:50 timoguin He breaks everything.
00:50 UtahDave he he.  yeah, that looks off.
00:50 UtahDave tristanmatthews: would you mind opening an issue on our github repo with that info?
00:50 UtahDave timoguin: Unfortunately that's not always a false statement!!   :)
00:51 tristanmatthews Nope wouldn't mind at all. So does it it just need a name change or is there a 3.2.1 somewhere that I'm missing?
00:51 timoguin looks like the numbering is just off
00:51 smcquay joined #salt
00:51 UtahDave I really don't know.  Most of those numbers are auto-generated.  I'm not sure what would cause that.
00:52 srage joined #salt
00:54 Gordonz joined #salt
00:59 linuxlewis joined #salt
01:12 srage joined #salt
01:13 GradysGhost joined #salt
01:18 acabrera joined #salt
01:18 smkelly joined #salt
01:19 smkelly joined #salt
01:20 thayne joined #salt
01:21 happytux joined #salt
01:35 Ryan_Lane well, I've finally made a substantial PR :D https://github.com/saltstack/salt/pull/12101
01:38 aranhoide salt is reporting the results of the 'unless:' clauses of my 'cmd.run' states as 'ERROR's, even though I fully expect them to fail sometimes (that's when the cmd gets actually run)
01:39 aranhoide not a big deal, but I get alarmed sometimes :)
01:42 OwenW joined #salt
01:49 NullWagon joined #salt
01:53 napper joined #salt
01:53 OwenW Is there any kind of jinja command that will let me render the IDs of minions that have a particular state? i.e. {{ minion_with_state['bind9'] }} or something
01:55 jdenning joined #salt
02:03 sroegner joined #salt
02:06 monokrome joined #salt
02:08 alunduil joined #salt
02:09 mateoconfeugo joined #salt
02:10 budrose joined #salt
02:11 ajw0100 joined #salt
02:12 ckao joined #salt
02:18 ajw0100 joined #salt
02:20 srage_ joined #salt
02:26 abe_music joined #salt
02:27 srage joined #salt
02:38 l0x3py joined #salt
02:47 combusean joined #salt
02:49 aranhoide joined #salt
02:51 joshu joined #salt
02:52 joshu hello
02:57 HeadAIX joined #salt
02:58 brd joined #salt
03:06 aranhoide joined #salt
03:09 xl1 joined #salt
03:11 zain_ joined #salt
03:14 joshu does anyone see anything wrong with this lins in my map.jinja {% set tomcat = salt['grains.item']('oscodename') ({
03:14 joshu i want to filter by oscodename not os
03:16 catpigger joined #salt
03:17 programmerq joined #salt
03:18 programmerq so, trying out salt for the first time. it appears that running commands like salt-ssh on my laptop requires sudo? (osx here)
03:18 programmerq seems a bit strange?
03:18 programmerq or do people generally not run salt commands on their laptop
03:28 joshu left #salt
03:30 googolhash joined #salt
03:31 faldridge joined #salt
03:35 sgviking joined #salt
03:39 lnxnut joined #salt
03:45 aranhoide left #salt
03:51 budrose joined #salt
03:57 doanerock joined #salt
04:05 anuvrat joined #salt
04:12 yusuket joined #salt
04:15 googolhash joined #salt
04:24 jeremyfelt joined #salt
04:30 bmonty joined #salt
04:31 bmonty anyone have pointers on how I can detect the version of ubuntu on a minion and then use that in a state file?
04:32 jalbretsen joined #salt
04:32 xmj programmerq: only in VMs :-)
04:32 yusuket joined #salt
04:32 joehh bmonty: there is oscodename in grains
04:32 bmonty joehh: thanks
04:33 renoirb yes.
04:33 renoirb obviously, we aren't automating Elephants and Coffee cups deployment, but computers ;)
04:34 renoirb in a state file, this [0] describes it. http://docs.saltstack.com/en/latest/topics/tutorials/states_pt3.html#using-grains-in-sls-modules
04:43 possibilities joined #salt
04:43 j4son joined #salt
04:47 mgw joined #salt
04:49 ajprog_laptop joined #salt
05:09 faldridge joined #salt
05:09 acabrera joined #salt
05:12 schimmy joined #salt
05:17 schimmy joined #salt
05:28 bmonty can I use salt to manage the state of the master?  should I just install salt-minion on my master?
05:30 srage_ joined #salt
05:31 malinoff joined #salt
05:32 yusuket joined #salt
05:44 epcim joined #salt
05:47 epcim_ joined #salt
06:08 TyrfingMjolnir joined #salt
06:09 abe_music joined #salt
06:09 bhosmer joined #salt
06:11 ndrei joined #salt
06:21 anuvrat joined #salt
06:25 happytux_ joined #salt
06:30 ivan left #salt
06:38 faldridge joined #salt
06:57 slav0nic joined #salt
06:57 slav0nic joined #salt
07:01 anuvrat joined #salt
07:03 it_dude joined #salt
07:05 Kenzor joined #salt
07:30 CeBe joined #salt
07:51 mpoole joined #salt
07:55 ndrei joined #salt
08:02 aranhoide joined #salt
08:02 aranhoide where does salt record `cron.present` states in ubuntu?  I see nothing in /etc/crontab
08:12 OwenW I think it would depend on the user that you've set the cron.present as, try "sudo -u whateverusername crontab -e"
08:14 joehh bmonty: yes - all our masters are minions themseleves
08:21 giantlock joined #salt
08:21 nix2170 joined #salt
08:23 aranhoide OwenW: thanks!
08:24 markm joined #salt
08:24 shm_get joined #salt
08:40 Katafalkas joined #salt
08:42 nix2170 joined #salt
08:44 it_dude joined #salt
08:56 OwenW Looking for a way to render the ID of a minion that has a specified .sls or state attached to it in a jinja template, any ideas?
09:00 thayne joined #salt
09:01 harobed_ joined #salt
09:02 fragamus joined #salt
09:02 harobed_ joined #salt
09:02 shoma joined #salt
09:02 ggoZ joined #salt
09:22 nix2170 OwenW, not sure I understood your question. You want, in a jinja template, to get a list of all minions' IDs that have a certain state as high?
09:23 mark_____ joined #salt
09:24 OwenW I tried explaining it here: http://stackoverflow.com/q/23150717/2640621
09:30 __number5__ joined #salt
09:34 ggoZ joined #salt
09:37 nix2170 OwenW, I guess these two links could be used to solve your task: https://stackoverflow.com/questions/17158665/how-to-get-a-list-of-all-salt-minions-in-a-template http://salt.readthedocs.org/en/latest/ref/modules/all/salt.modules.match.html?highlight=match#module-salt.modules.match
09:37 ggoZ joined #salt
09:38 ggoZ joined #salt
09:38 kermit joined #salt
09:39 sealibora joined #salt
09:43 sealibora joined #salt
09:44 goodwill joined #salt
09:44 goodwill joined #salt
09:50 shm_get using archive.extracted how does one set the owner/group for the resulting extracted archive ? http://docs.saltstack.com/en/latest/ref/states/all/salt.states.archive.html#module-salt.states.archive does not mention anything about that
10:00 OwenW shm_get: doesn't look like it has that option, I would try hacking around it using a file or cmd state with a require if it's feasible
10:00 ndrei joined #salt
10:03 shm_get OwenW ok, thanks.
10:14 dmorrow joined #salt
10:18 harobed joined #salt
10:22 bhosmer joined #salt
10:25 elfixit joined #salt
10:26 MK_FG joined #salt
10:26 MK_FG joined #salt
10:30 anuvrat joined #salt
10:36 DaveQB joined #salt
10:38 Katafalkas joined #salt
10:44 it_dude joined #salt
10:50 ze- 2014.1.2 is out. any idea when debian package will be available ?
10:58 xmj ze-: won't probably be packaged
10:58 xmj ze-: check for 2014.1.3
10:59 puppet joined #salt
11:03 ze- o.o already 2014.1.3 ?!
11:04 xmj blame the regression in .2
11:04 ze- damn :)
11:08 ze- well, when will 2014.1.3 de available in .deb ? :P
11:19 nebuchadnezzar joined #salt
11:32 ekristen anyone know how the docker pillar registry info works in conjunction with the docker states?
11:35 zain_ joined #salt
11:38 Katafalk_ joined #salt
11:46 it_dude joined #salt
11:54 bastion1704 joined #salt
11:57 joehh ze-: very shortly for debian - shortly for ubuntu
11:59 * xmj nags terminalmage about EPEL
11:59 * xmj has a patch ready for FreeBSD
11:59 xmj cedwards: ping
12:01 bastion1704 joined #salt
12:08 lnxnut joined #salt
12:12 cgar joined #salt
12:13 jrdx joined #salt
12:13 ze- joehh: any ETA ? :)
12:13 ze- (debian)
12:18 Katafalkas joined #salt
12:28 jslatts joined #salt
12:33 zain_ joined #salt
12:34 tessellare joined #salt
12:36 zain__ joined #salt
12:38 joehh unstable and wheezy are built, squeeze building now
12:38 joehh ze: uploading nowish
12:42 ze- any idea how long it takes before appearing on http://debian.saltstack.com ? :)
12:43 ze- ('nowish' meaning ?)
12:44 joehh ze-: minutes for unstable and wheezy
12:46 joehh ze-: should be there now - let me know if any probs
12:47 ze- nice. thanks. :)
12:47 joehh squeeze still 30 minutes away or so
12:47 * ze- won't care much about squeeze for now :)
12:48 ravibhure joined #salt
12:55 ndrei joined #salt
12:56 chiui joined #salt
12:56 y0j joined #salt
12:57 ndrei joined #salt
12:57 GradysGhost joined #salt
13:01 ravibhure joined #salt
13:06 dmorrow__ joined #salt
13:09 elfixit joined #salt
13:13 elfixit joined #salt
13:18 Katafalkas joined #salt
13:24 mpanetta joined #salt
13:25 Darnoth joined #salt
13:26 bastion1704 joined #salt
13:28 mapu joined #salt
13:28 ekristen watch is suppose to only trigger when and if the state being watched has changes correct?
13:29 grosjean left #salt
13:32 danielbachhuber joined #salt
13:33 Whissi joined #salt
13:34 vejdmn joined #salt
13:42 viq ekristen: yes
13:43 ravibhure joined #salt
13:43 frasergraham joined #salt
13:45 undefinednull joined #salt
13:49 ekristen viq: looks like the watch statement on docker.pulled is broken
13:49 ekristen it triggers no matter what, especially when there isn’t changes
13:50 ipmb joined #salt
13:57 mgw joined #salt
14:04 bastion1704 joined #salt
14:05 zzzirk joined #salt
14:07 jdenning joined #salt
14:08 donatello joined #salt
14:08 donatello does salt's external authentication system currently allow all users in a group to execute certain commands?
14:08 Sam__ joined #salt
14:09 donatello i can see that it can be accomplished on a per user basis at docs.saltstack.com/en/latest/topics/eauth/index.html
14:09 jforest I'll go on record as asking for new RPM in EPEL :)
14:09 jforest 2014.1.0 there still :(
14:10 donatello oops never mind. just found it on the docs page :(
14:11 ziirish joined #salt
14:15 dmorrow_ joined #salt
14:17 diegows joined #salt
14:19 Katafalk_ joined #salt
14:23 astol joined #salt
14:24 abe_music joined #salt
14:25 jeremyBass joined #salt
14:29 frasergraham joined #salt
14:29 cedwards xmj: hello
14:30 xmj hey cedwards !
14:30 xmj did you see my mail? :-)
14:31 ekristen when the states are listed out after running a salt command, is that also the order in which they were run?
14:32 cedwards xmj: looks like you beat me to it. I started work on it yesterday, but got caught up on the zmq issue.
14:32 cedwards xmj: i'll approve
14:33 xmj cedwards: awesome
14:33 jslatts is it possible to call an execution module from a python renderer SLS?
14:33 xmj cedwards: could you please send a mail? then i'll poke my mentor and get it in within the next thirty minutes
14:34 cedwards done
14:34 xmj great
14:34 xmj do you happen to have a changelog at hand?
14:36 cedwards i don't
14:38 xmj http://docs.saltstack.com/en/latest/topics/releases/ doesnt yet list 2014.1.3
14:38 cedwards I found this one: http://docs.saltstack.com/en/latest/topics/releases/2014.1.2.html
14:39 ZombieFeynman joined #salt
14:39 xmj right, that i found too
14:39 xmj now the interesting part would be to know which regression .3 fixed in .2
14:39 xmj joehh: ^
14:42 abe_music joined #salt
14:45 TyrfingMjolnir joined #salt
14:46 thedodd joined #salt
14:46 alunduil joined #salt
14:46 eriko joined #salt
14:46 zain_ joined #salt
14:47 eculver joined #salt
14:47 eculver joined #salt
14:47 zain__ joined #salt
14:48 astol how do I repopulate pillars in the middle of applying top state?
14:48 pdayton joined #salt
14:49 btorch why is using order so frown upon ?
14:50 astol btorch: guess you'd better specify dependencies
14:50 btorch astol: you mean the requires ?
14:51 astol btorch: yeah, specifying order is worse that using requires
14:51 astol btorch: but I guess you can and perhaps should use order in specific cases
14:51 astol btorch: for example I am using order to force generation of password in pillars before everything else runs
14:52 btorch the thing is the requires are not working so well to me, not placing things on the order I need... also it seems like I have to use a great amount of requires and includes to make it the order I need things to go
14:53 opapo_ joined #salt
14:53 astol btorch: great amount of requires is the thing you should get used to - it is good, actually. shows dependencies explicitly. why they're not working the way you think they should - that's another question. dig into that. but they should work
14:53 btorch astol: yeah I only need to use for the states that are common to all, like I need the common.users and common.packaging states for example setup prior to installing pkgs
14:53 sroegner joined #salt
14:53 btorch astol: so it is normal ?
14:54 btorch it just seemed weird to me
14:55 astol btorch: well actually I
14:57 astol btorch: I'd specify this things explicitly, i.e. if a service needs some specific user pre-existing and it is created in common.users, I'd specify require: - user: <username>, not specifying some highest order for all common things and hoping that they will a) be executed first b) always contain all that you need
14:57 bhosmer joined #salt
14:58 zain_ joined #salt
14:59 wendall911 joined #salt
14:59 astol btorch: since this is changes management world where boring imo usually means good, you will explicitly require something you need from the "head state", thus signaling a person that will support this state after you, what did you want, and dependency system will automatically put common.users ahead of other things
14:59 astol btorch: yes this is boring and uncool and good
15:00 astol btorch: as is wearing welding mask while using blowtorch :D
15:01 kamal_ joined #salt
15:01 kamal__ left #salt
15:06 mateoconfeugo joined #salt
15:08 btorch so since I also have tried to split the common.users state into 3 files to make it more modular, seems like I always have issues with require_ins
15:08 btorch I keep on getting "Cannot extend ID common.users.users in "base:common.users.files". It is not part of the high state."
15:09 TyrfingMjolnir joined #salt
15:12 btorch astol: so this is normal ?  http://goo.gl/ut5VhV
15:12 btorch but still doesn
15:12 btorch t give the proper order even though in my mind it seems correct :)
15:12 yusuket joined #salt
15:13 btorch but lowstate shows files like /etc/sudoers being setup before sudo pkg is actually installed!
15:15 juaqim joined #salt
15:16 btorch now a state.highstate seems to show the proper order... I guess I'm just confused cause I thought lowstate would show me how the order of things are
15:24 MTecknology Heh... the windows installer will create the minion_id file using $hostname instead of what's provided in the install dialog and the master and id values entered in the installer don't make it into the config file.
15:25 MTecknology bummer
15:26 xmj cedwards: committed
15:26 TyrfingMjolnir joined #salt
15:26 cedwards xmj: awesome. thanks
15:27 xmj cedwards: no worries
15:27 davet joined #salt
15:27 xmj you guys keep up the good work with making salt even more awesome
15:27 xmj we see to it that it gets in quick :]
15:29 epcim joined #salt
15:30 thayne joined #salt
15:31 epcim_ joined #salt
15:31 TyrfingMjolnir joined #salt
15:32 mgw joined #salt
15:33 jeremyfelt joined #salt
15:35 ZombieFeynman joined #salt
15:35 troyready joined #salt
15:36 dmorrow_ joined #salt
15:36 Dattas joined #salt
15:37 MTecknology What's the structure of a reactor line in the master config? I know there's this...  'salt/minion/*/start' but what is that?  'salt/minion/<id>/<tag>' ?
15:39 conan_the_destro joined #salt
15:40 micah_chatt joined #salt
15:41 micah_chatt joined #salt
15:41 acabrera joined #salt
15:42 sijis whiteinge: ping - you around?
15:43 tonthon joined #salt
15:43 sijis whiteinge: any *hope* in publishing what's avail in pepper in pypi under an alpha tag?
15:46 nlb joined #salt
15:46 jalbretsen joined #salt
15:46 MTecknology or is it just <tag>?...
15:47 Katafalkas joined #salt
15:47 davet joined #salt
15:48 jforest ls
15:48 jforest argh, ww
15:49 bhosmer joined #salt
15:56 TyrfingMjolnir joined #salt
15:56 sroegner joined #salt
15:56 meteorfox joined #salt
15:59 gothix_ joined #salt
16:02 zain_ joined #salt
16:03 KyleG joined #salt
16:03 KyleG joined #salt
16:05 MTecknology Anyone see why this reactor isn't working?  http://dpaste.com/1785175/
16:07 mgw joined #salt
16:08 jdenning joined #salt
16:08 retrospek joined #salt
16:11 MTecknology seems to be that the master isn't running anything in the sls, but the tag seems to be correct
16:11 Sypher joined #salt
16:18 TyrfingMjolnir joined #salt
16:19 shadylog joined #salt
16:19 ZombieFeynman joined #salt
16:19 bemehow joined #salt
16:20 shadylog Hey
16:23 yusuket joined #salt
16:23 jrdx joined #salt
16:23 gothix_ joined #salt
16:24 redondos joined #salt
16:24 chrisjones joined #salt
16:26 thayne joined #salt
16:26 ajolo joined #salt
16:27 conan_the_destro joined #salt
16:27 redondos joined #salt
16:28 jdenning joined #salt
16:29 cnelsonsic joined #salt
16:36 schimmy joined #salt
16:38 Ryan_Lane joined #salt
16:39 schimmy1 joined #salt
16:44 ZombieFeynman joined #salt
16:45 eculver joined #salt
16:45 eculver joined #salt
16:47 mapu joined #salt
16:47 Katafalk_ joined #salt
16:50 rlarkin joined #salt
16:51 KyleG Can anyone tell me what's going on here? :( https://gist.github.com/kylegato/3a852656fa217670b71f
16:51 KyleG since i'm specifying the state directly it SHOULD be working
16:51 KyleG I don't understand why it works for one host but not my cron.la host
16:55 ajw0100 joined #salt
16:57 Dattas joined #salt
16:58 shm_get what is the recommended way to 'patch' a non-managed configuation file ? I'd like to avoid having it managed because it could change slightly with different versions.. but the part that I need to 'patch' is stable
16:59 ajw0100 joined #salt
17:01 nix2170 KyleG: It seems like the openldap-client state is defined on base but the cron.la.* minions are in a different environment.
17:01 smcquay joined #salt
17:01 retrospek left #salt
17:01 KyleG nix2170: What does that even mean?
17:01 KyleG It's a brand new server
17:01 KyleG I just set it up and approved the key
17:01 KyleG added it to top.sls underneath another server
17:02 KyleG and this is where I am :\
17:02 nix2170 do you have different environments set in your master? Like base, prod, dev?
17:02 KyleG nope
17:02 KyleG just base
17:02 jdenning_ joined #salt
17:02 KyleG Here's my top.sls file: https://gist.github.com/kylegato/0d00e6fb8fff77ffa661
17:04 KyleG but regardless
17:04 KyleG if I tell salt to push a state
17:04 KyleG it should push the state
17:05 KyleG I copied and pasted the host above it below
17:05 KyleG and re-named it
17:06 KyleG and that server still works fine
17:06 zain_ joined #salt
17:06 KyleG so..
17:06 nix2170 notice that in the top.sls only 'cron.la.wiredrive.com' is defined, but you ask the cli to run 'cron.la.*'. Maybe you have a minion that matchs this glob but is not defined in the top.sls?
17:06 KyleG nope
17:06 nix2170 I mean the only cron.la.* fqdn matching in the top.sls is cron.la.wiredrive
17:06 KyleG https://gist.github.com/kylegato/7c35b5aac01bd0456ad8
17:07 mattikus joined #salt
17:07 nix2170 Just for kicks, try the cli without the glob: salt 'cron.la.wiredrive.com' state.sls openldap-client -t 360 -v  test=True
17:08 shoma joined #salt
17:08 KyleG https://gist.github.com/kylegato/69e6c603d8d8aecc9f34
17:08 KyleG makes no sense to me
17:08 KyleG Isn't salt supposed to just say
17:09 KyleG "I have a folder in states called openldap-client, it has an init.sls, I should execute it against this minion"
17:09 KyleG and my init.sls has no conditionals
17:09 KyleG for that state
17:09 xmj ?
17:09 KyleG and the state/command works fine for the other hosts
17:09 KyleG -_-
17:09 xmj KyleG: do you have other hosts in top.sls
17:09 KyleG yes
17:09 KyleG they all work fine
17:09 xmj and do you have this in top.sls
17:10 KyleG xmj: https://gist.github.com/kylegato/0d00e6fb8fff77ffa661
17:10 bt_ joined #salt
17:10 linuxlewis joined #salt
17:11 KyleG I've been using salt for almost 2 years
17:11 KyleG so this highly confuses me
17:11 bt_ left #salt
17:11 KyleG Usually this is straight forward...
17:11 KyleG add host to top.sls….state it out…call it a day
17:12 bt_ joined #salt
17:12 nix2170 Does the same issue occurs with the other states defined for this minion?
17:12 KyleG I've only tested a couple, but yes
17:12 KyleG state.highstate does it too
17:13 KyleG "        Comment:   No Top file or external nodes data matches found" - from highstate
17:13 yusuket joined #salt
17:13 bt_ i'm looking for some basic (hopefully) help with salt minion proxy.  there's something i'm missing with how the master talks to the minion proxy
17:14 nix2170 Is your master and this minion on the same salt version? (IOW this is not a 2014 / .17 incompatibility?)
17:14 mateocon_ joined #salt
17:14 KyleG master is 0.17.4
17:14 KyleG minion is 2014.1.1
17:14 nix2170 Bingo
17:14 linuxlew_ joined #salt
17:14 nix2170 Downgrade the minion
17:14 KyleG goddamnit
17:15 KyleG My co-worker does not make it easy, we only keep the latest packages on hand.
17:15 KyleG I'll have to remove it and then install it w/ pip
17:15 KyleG gah thank you nix2170
17:15 KyleG I thought I was going nuts
17:15 KyleG It would be nice if some logic was put into the master to detect a version mismatch and just say so
17:15 KyleG rather than just failing blindly
17:15 Heartsbane whiteinge: ping
17:16 nix2170 Yeah, I've been bitten by this issue and it often comes up.
17:16 nix2170 I guess it should be easy to make new versions detect an old master and report an error.
17:17 bhosmer joined #salt
17:17 nix2170 KyleG: I have salt-0.17.5-1.el6.noarch rpms for master and minion if you want pm me your address and I'll email them to you
17:18 jaimed joined #salt
17:18 kermit joined #salt
17:18 ajolo joined #salt
17:18 KyleG I wish nix2170.
17:18 KyleG We're 100% FreeBSD lol
17:18 KyleG Which means that I ALWAYS find a bug in everything.
17:18 xmj KyleG: did you see 2014.1.3 ?
17:18 KyleG Oh and newrelic dropped FreeBSD support in february. :(
17:18 KyleG Dying breed...
17:18 terminalmage nix2170: the RPMs are up on the fedora build system
17:18 terminalmage anyone can grab them from there
17:19 xmj KyleG: i updated the salt version on FreeBSD a few hours ago :-)
17:19 KyleG :D
17:19 xmj doh
17:19 nix2170 terminalmage: thanx, didn't know! I got them from someone here about a month ago :)
17:19 xmj KyleG: why is your master older than your minion
17:19 bitmand xmj: thanks :)
17:19 catpigger you could have the master check the minion version
17:19 xmj KyleG: you want the master to be newer or at least the same version ...
17:19 Ryan_Lane joined #salt
17:20 terminalmage nix2170: https://koji.fedoraproject.org/koji/buildinfo?buildID=506146
17:20 terminalmage nix2170: and links to all the builds are here: https://koji.fedoraproject.org/koji/packageinfo?packageID=13129
17:20 terminalmage nix2170: np
17:21 undefinednull joined #salt
17:21 Heartsbane Ah Hell I will just ask in #salt
17:21 Heartsbane If I have some variable I set in a state file like this:
17:21 Heartsbane {% set derp_version = salt['pillar.get']('version') %}
17:21 Heartsbane And I wanted to set that via the command line when I call the state
17:21 Heartsbane what is the syntax? Is it:
17:21 nix2170 terminalmage: thanx! That is bound to be useful :D
17:21 Heartsbane salt 'targetserver' state.sls derp pillar='{version:"6.0.2-14581"}'
17:21 Heartsbane Or is it something else? can I do that
17:22 Dattas joined #salt
17:23 terminalmage nix2170: yep! you can even grab the SRPM from there in case you want to roll your own packages
17:24 linuxlewis joined #salt
17:25 Heartsbane whiteinge told me I was able to do something like that, I wanted to know if that was correct.
17:25 terminalmage Heartsbane: passing pillar on the CLI like that should work for state.sls, but I recall that I needed to add pillar handling to a few of the functions in the state module
17:25 terminalmage not sure which I added, I need to look at the git log
17:25 terminalmage what version are you running?
17:26 Heartsbane 1.0
17:26 terminalmage heh
17:26 * whiteinge waves at Heartsbane
17:26 robawt run Heartsbane run
17:26 whiteinge Heartsbane: what terminalmage said. you have the right syntax for passing a var to an sls file
17:27 whiteinge sijis: I think that's doable.
17:27 ajolo joined #salt
17:27 smcquay joined #salt
17:27 terminalmage Heartsbane: yeah state.sls supports that, I had to add it to state.top and state.single
17:28 Heartsbane thank you
17:28 terminalmage no prob
17:29 aw110f joined #salt
17:29 Heartsbane also robawt no need to run, whiteinge isn't dangerous (at all) ... just saying
17:29 whiteinge i'm a little dangerous...
17:29 robawt he's scary dangerous
17:29 robawt i heard they had to deport him to his current state from CA
17:30 TyrfingMjolnir joined #salt
17:30 robawt too much flan
17:30 Heartsbane whiteinge: try to type that without laughing a even a little
17:30 terminalmage Heartsbane: that syntax also lets you override existing pillar variables, which can be helpful when troubleshooting
17:30 Heartsbane terminalmage: really .... excellent
17:30 terminalmage or in certain instances where you want to, in effect, pass a variable to a state
17:30 Heartsbane robawt: no more flan jokes
17:31 robawt haha
17:31 Heartsbane terminalmage: thanks that will make debugging more useful
17:31 whiteinge terminalmage: great addition to state.top. what release is that in?
17:33 * Heartsbane ponders this and goes back to his state files.
17:33 Heartsbane BTW salt has made our maintenance events half of what they used to be.
17:33 faldridge joined #salt
17:34 ekristen Heartsbane: I’d be interested to know how you are using salt to do maintenance, care to share?
17:34 yusuket joined #salt
17:35 Heartsbane ekristen: Just 143 Centos 5 & 6 boxes ... server & kernel patches, then reboot.
17:36 Heartsbane it is nearly all scripted. Now I just need to add the Windows machines
17:37 ekristen Heartsbane: are you just using the ability to run yum upgrade across multiple servers and let them all upgrade at the same time?
17:38 sijis whiteinge: that's awesome! :)
17:38 Heartsbane ekristen: not all at the same time, but upgrade 50% failover then upgrade the other 50% ... grains have made the audit process so much easier
17:39 ekristen yeah, grains are pretty sweet
17:39 druonysus joined #salt
17:39 druonysus joined #salt
17:40 MTecknology Isn't there a built-in returner that can just send data to an API?
17:40 Heartsbane Anyway I want to test this new state... BBL
17:41 Heartsbane Also whiteinge isn't dangerous, not even a little.
17:42 MTecknology I guess I get to write a custom returner, then the web api, then everything else... :'(
17:45 combusean joined #salt
17:45 terminalmage whiteinge:
17:45 terminalmage whiteinge: might only be in develop
17:46 bhosmer joined #salt
17:46 mgw joined #salt
17:47 * whiteinge glowers at Heartsbane ...dangerously
17:49 bemehow joined #salt
17:52 xmj terminalmage: UtahDave told me you were the one to talk to about EPEL pkgs...?
17:52 terminalmage yes
17:53 xmj ha, is it already or will it $soon get in?
17:53 scooby2 joined #salt
17:54 terminalmage I'm sorry, but I don't understand the question
17:54 xmj Is 2014.1.3 already in EPEL, and if not, will that get in before the weekend is over?
17:55 terminalmage xmj: bookmark this page, and you'll never need to ask that question again
17:55 terminalmage https://koji.fedoraproject.org/koji/packageinfo?packageID=13129
17:55 terminalmage I'll be packaging it today
17:56 elfixit joined #salt
17:56 terminalmage it takes up to 24 hours to be accepted into epel testing, and then anout 24 hours or more depending on how often your mirror syncs
17:56 xmj the second part was what i was interested in; i checked about 4 hours ago and wasn't in at that time
17:56 xmj but, good to hear!
17:56 terminalmage ahh ok. this seems to be a common question and not many people seem to know about that link
17:57 terminalmage so I'm trying to get the word out :)
17:57 terminalmage as soon as the build is started, it shows up there
17:57 xmj eh, i checked the repository's directory listing :-0
17:57 terminalmage even before it completes
17:58 bt_ anybody know what a proxytype proxyconn file should be called?  should there be a file extension?  .py? .sls?  TIA
18:00 bt_ in the traceback, i get a keyerror for {proxytype_name}.Proxyconn, so i thought the file should be at salt/proxy/{proxytype_name}, but that doesn't seem to work
18:01 jnials joined #salt
18:01 terminalmage bt_: can you post this question to the mailing list? proxy minions are new and the de facto expert on them is out of the office today
18:02 jnials joined #salt
18:02 bt_ terminalmage: haha, sure
18:02 whiteinge sijis: pip install salt-pepper
18:02 bt_ terminalmage: browsing through the source now to see if i can answer my own question
18:03 terminalmage bt_: hehe, if so, all the better
18:03 UtahDave joined #salt
18:03 terminalmage but it'll be easier for him to track down the question on the list than going through IRC logs :)
18:05 bt_ terminalmage: understood; if it turns out to be anything other than my own learning curve, i'll pose the question on the mailing list
18:08 terminalmage bt_: cool, and if you find something that is undocumented, feel free to submit a pull request against the docs, or open an issue on github and one of us can add the information to the docs
18:08 bt_ terminalmage: will do!
18:08 terminalmage great :)
18:12 linuxlew_ joined #salt
18:14 jslatts does anyone know if it is possible to call an execution module from a python renderer SLS?
18:17 terminalmage what do you mean by "renderer SLS"
18:17 jslatts terminalmage: http://salt.readthedocs.org/en/latest/ref/renderers/all/salt.renderers.py.html
18:17 jslatts from inside a SLS that starts with #!py
18:17 terminalmage jslatts: renderers aren't SLS
18:17 terminalmage that's just a renderer module
18:17 jslatts sorry, a state renderer I guess
18:18 faldridge joined #salt
18:18 terminalmage no, renderers just... render. are you talking about SLS files that use the py renderer
18:18 terminalmage ?
18:19 terminalmage like, are you trying to write a pure python script that runs salt states?
18:19 terminalmage sorry for the confusion, I just want to make sure we are on the same page
18:19 jslatts i guess I want to write a state that uses the "pure python renderer"
18:19 terminalmage yeah
18:19 jslatts and outputs the state tree as a result
18:20 jslatts and within that python state, can I call other salt modules
18:20 terminalmage ok, so in that case, instead of {% salt['test.ping']() %} you'd use __salt__['test.ping']()
18:20 jslatts is my question
18:20 yusuket joined #salt
18:20 jslatts great! is that in the docs somewhere? I have been trying to figure this out all morning
18:20 terminalmage grains are in __grains__, pillar data in __pillar__, etc.
18:21 terminalmage jslatts: it's right there at the top of the page, though it could probably be made more explicit
18:21 Flusher joined #salt
18:21 jslatts I wasn't sure what "__salt__" meant
18:21 jslatts and what was available on that object
18:22 jslatts I've been working on getting a test environment set up so I could try and inspect it
18:22 terminalmage jslatts: yeah... I'll make the wording there more explicit. I think there's an example of the py renderer elsewhere in the code that shows __salt__ being used
18:22 terminalmage but I forget where it is
18:23 terminalmage either way, it could definitely be explained better, and I'll make sure that gets done
18:23 terminalmage thanks!
18:23 jslatts no, thank you! just shortcutted my dev time here
18:23 jslatts i'm trying to figure out how to deploy our app builds out of s3
18:23 jslatts not as simple with saltstack as I had initially guessed
18:28 gadams999 joined #salt
18:29 jcockhren jslatts: I do that
18:30 jcockhren I have a s3 bucket that is a debian repo
18:30 jcockhren I use salt to set apt sources list to point to that buckets
18:30 jcockhren then use the pkg state like normal
18:30 jslatts jcockhren: so you uses package.latest to deploy?
18:31 jslatts or how do you move versions? update salt states?
18:31 jcockhren pkg.latest for updates
18:31 bhosmer joined #salt
18:31 sijis whiteinge: woot!!!
18:31 jcockhren initial deploys use pkg.installed
18:32 jcockhren my use case if for logstach, es and related things
18:32 jcockhren I have no need to specify a version explictly.
18:32 jslatts does pkg. latest automatically index the list of packages available in s3/repo?
18:33 jslatts guess I can go read the code
18:33 jcockhren honestly, I don't remember why I even used pkg.latest
18:34 jcockhren (i.e. create a state using it)
18:34 HeadAIX joined #salt
18:34 jslatts we are using salt to autoscale, so I need to know that when a new instance comes up, it will get the latest build of our app
18:35 jslatts ideally, it would pull down a file from s3 that is the "pointer" to the correct build, so we could roll back
18:35 jcockhren pkg.installed will update the package to the latest version
18:35 patrek joined #salt
18:35 MTecknology Is it possible to tell state.highstate that gitfs needs to be updated before execution?
18:35 schimmy joined #salt
18:37 MTecknology I'm sending state.highstate to the minions via the reactor system but gitfs needs to be updated before the minions start the highstate execution.
18:38 jslatts MTecknology: there is fileserver.update
18:38 shoma joined #salt
18:38 zain_ joined #salt
18:39 schimmy1 joined #salt
18:40 MTecknology jslatts: so just trigger that on the master? Can I use - require: in the reactor?
18:41 jcockhren jslatts: hmmm. for rollbacks, it might make more sense to have versions in a pillar
18:41 jslatts i'm not familiar with reactors, so I can't say, but I use that command to force an update
18:41 jslatts MTecknology: however, I don't know if that causes the git backed pillars to update as well
18:42 logix812 joined #salt
18:43 MTecknology what's really fun is seeing that a large chunk of reactor documentation was written (structured) by me and has mostly just been improved and explained better
18:44 jcockhren haha MTecknology
18:44 jcockhren good to see you're trying to keep up with the jones'
18:44 jcockhren ;)
18:44 MTecknology jcockhren: hm?
18:45 jcockhren just lurking and joking
18:45 logix812 When running salt in docker I get this warning 'Although 'dmidecode' was found in path, the current user cannot execute it.'  it happens 3x during init: after "Reading configuration from /etc/salt/master"  and after: "Updating fileserver cache" and "Loaded gitfs as virtual git"
18:45 logix812 safe to ignore?
18:47 MTecknology The reactor system is incredible and powerful and incredibly awesome, but it's very under-documented. It does much more than the docs show. I'm hoping -require is just not documented
18:49 mpanetta I'm hoping to use reactor to auto-add and auto-del systems from zenoss when salt-cloud spins them up.
18:49 bhosmer joined #salt
18:49 MTecknology jslatts: looks like you're right - salt-run fileserver.update should take care of what I need, but I also need to make that happen before highstate
18:49 mpanetta After using that event viewer example thing it seems like that sort of thing is infinitely doable.
18:49 jslatts MTecknology: cool :)
18:51 obimod joined #salt
18:51 obimod hay hay
18:52 MTecknology jslatts: I guess reactors can't take advantage of requisites. :(
18:52 MTecknology https://github.com/saltstack/salt/issues/8098
18:53 jslatts dang
18:54 MTecknology I'll have to have the minion tell the master to update gitfs, and then have it wait and then send the command to update minions
18:56 terminalmage jslatts: https://github.com/saltstack/salt/pull/12123/files
18:57 jslatts terminalmage: ++, that helps
19:00 MTecknology terminalmage: DOCUMENT!!!!
19:01 Ryan_Lane terminalmage: whats a fileserver environment?
19:01 toastedpenguin joined #salt
19:01 terminalmage Ryan_Lane: base is the default fileserver env
19:01 terminalmage you can define others
19:01 Ryan_Lane ahhhh. that kind of environment
19:01 acabrera joined #salt
19:02 terminalmage yeah
19:02 terminalmage also referred to as a saltenv
19:02 * Ryan_Lane nod
19:02 Ryan_Lane *nods
19:02 terminalmage as of November or so
19:02 terminalmage I could probably clarify that a bit
19:02 elfixit joined #salt
19:05 gothix_ What is wrong with this in my map.jinga {% set tomcat = salt['grains.item']('oscodename') ({
19:05 gothix_
19:05 gothix joined #salt
19:06 utahcon so "grains.item ip_interfaces" returns {'lo':...,'eth0':...} how do I pick one using salt['grains.get']('ip_interfaces') ?
19:07 utahcon I thought it was salt['grains.get']('ip_interfaces.eth0')
19:07 manfred ip_interfaces:eth0
19:07 manfred not .eth0
19:07 gothix utahcon, thats where im at too i am trying to replace fiilter_by
19:13 utahcon is it normal for salt['grains.get']('ip_interfaces:etho0') to giev ['123.123.123.123'] (with the ['']) ?
19:13 utahcon manfred: thanks, got me half way ;)
19:14 manfred it is, because it gives you a list of all the ip addrs on eth0
19:14 manfred np :)
19:15 utahcon manfred: so can I get just the first value of that list?
19:15 manfred yeah.
19:19 sarlalian joined #salt
19:19 danielbachhuber joined #salt
19:22 frasergraham joined #salt
19:22 kermit joined #salt
19:25 cgar joined #salt
19:26 joehoyle joined #salt
19:28 joehoyle hey has anyone here got salt working with autoscaling, so the instance would do a highstate on coming up? I looked a prerolling Keys on the minion, but seems you need to hardcode the ID of the minion to match a key on the master?
19:29 jslatts joehoyle: we do it
19:29 joehoyle jslatts: how do you have the master "autoaccept" the new minions?
19:30 jslatts joehoyle: there is actually an auto accept setting
19:30 joehoyle oh right!
19:30 jslatts for the master, so we have it on
19:30 jslatts but in order not to have a security issue, we have a syndic in each of our VPCs
19:30 joehoyle cool, and you allow the master to be accessed to machines in your securitygroup?
19:30 jslatts that is set that way
19:30 joehoyle ok yeah
19:30 jslatts yeah
19:30 jslatts well, they are local, so we allow any local subnet traffic to the syndic
19:31 joehoyle cool, and your minions do a "salt-call state.highstate" on the minion on coming up?
19:31 jslatts ASG instances then have the syndic address baked in, as well as the auto state.highstate
19:31 jslatts yeah, that is also a minion setting
19:31 joehoyle oh cool!
19:31 jslatts they can run a state on start
19:31 joehoyle do you know what that setting is?
19:31 jslatts one sec
19:31 joehoyle cool
19:32 gothix can someone tell me what is wrong with this {% set tomcat = salt['grains.item']('oscodename') ({
19:33 jslatts joehoyle: startup_states: 'highstate'
19:33 joehoyle jslatts: oh simple! thanks
19:33 jslatts on the master its auto_accept
19:33 joehoyle yup, just found that
19:33 joehoyle jslatts: are you running ubuntu by any chance?
19:33 jslatts yep, 12.04
19:34 joehoyle jslatts: ah ok, is that 0.17.5?
19:34 joehoyle having a nightmare with old salt pacakges
19:35 jslatts 2014.1.1
19:35 zain_ joined #salt
19:35 jslatts joehoyle: http://dev.mlsdigital.net/posts/Cloud-Native-Drupal/ is a blog that describes our set up. Specific to drupal, but there is a section describing salt + autoscaling + deployment
19:36 jslatts at a high level anyway
19:37 joehoyle oh nice
19:38 ajw0100 joined #salt
19:38 manfred joehoyle: which provider are you using for autoscale?
19:38 joehoyle jslatts: ahh, 12.04 seems to have packages for 2014.1.1, none of other other released do for some reason
19:38 joehoyle manfred: aws
19:38 jslatts joehoyle: the one other issue you have to solve with our setup is keeping the syndics in sync
19:39 manfred joehoyle: http://docs.aws.amazon.com/sns/latest/dg/SendMessageToHttp.html + http://salt-api.readthedocs.org/en/latest/ref/netapis/all/saltapi.netapi.rest_cherrypy.html#saltapi.netapi.rest_cherrypy.app.Webhook + a reactor file that tells the master to salt bootstrap a minion, or accept the minions key
19:40 joehoyle manfred: ahh interesting
19:40 Kenzor joined #salt
19:41 joehoyle I guess auto_accept is the sledgehammer fix!
19:41 manfred a similar http post webhook will be coming to openstack in the juno release for openstack heat
19:44 terminalmage Ryan_Lane: jslatts: I added more information just now, to that earlier commit: https://github.com/saltstack/salt/pull/12126/files
19:44 jslatts manfred: other than running a non-promiscuous master, what other advantages would a reactor based version have?
19:44 manfred you could build servers however you want, and the salt master would still salt them once they were done building
19:45 jslatts joehoyle: autoscaling with salt brings up a bunch of interesting side issues as well. like how "base" is your base AMI
19:45 stanchan joined #salt
19:45 manfred through a console, through a cli application, through the web ui
19:45 manfred through salt-cloud, doesn't matter, you could still have the salt master login and install salt on them.
19:45 jslatts manfred: ah, you mean not having to pre-install salt on the image
19:46 joehoyle @jslatts yeah, currently bring an instance into the pool is going to be slow, as the base ami is a blank one currently
19:46 Leech joined #salt
19:46 jslatts joehoyle: do you use VPCs?
19:46 manfred jslatts: indeed, you could build anything you want and if salt can saltify it, you are good to go
19:46 joehoyle jslatts: no, I have a pre-VPC account
19:46 joehoyle so havn't delved into it yet
19:46 manfred jslatts: that is why I can't wait for webhooks in heat to be a thing
19:46 jslatts joehoyle: me too, collectors item
19:46 joehoyle haha yeah
19:46 manfred i will be abandoning novaclient for forever
19:47 abe_music joined #salt
19:47 jslatts joehoyle: but one thing to consider if you are completely  building up an ASG instance from a base image, besides speed, is package availability. all of a sudden your ASG depends on whether package repos are available, github, etc
19:48 yusuket joined #salt
19:48 jslatts joehoyle: we ended up basically fully provisioning our AMIs to have all the deps and then just using salt to deploy our app code from S3
19:48 joehoyle jslatts: yeah, I am not keen on that at all
19:48 joehoyle @jslatts right yah, I can see hte advantage
19:48 joehoyle I just don't want to maintain an AMI right now!
19:48 jslatts joehoyle: i know
19:49 jslatts me neither
19:49 jslatts joehoyle: my next iteration of this set up will use salt to deploy docker images out of S3
19:49 joehoyle ahh yeah good thinking
19:49 joehoyle I think I can do a lot to make the instances a lot more lightweight so would come up faster
19:49 jslatts so we can have a fully provisioned image + build stored in s3 and only maintain the docker images, which could be created by the build server
19:50 jslatts yeah, at one point we were compiling code in our states to get a custom build of a dep. obviously not acceptable for autoscaling
19:50 joehoyle jslatts: what kind of timeframe does that put you at do you think?
19:50 jslatts joehoyle: time before I build all that or time for it to scale up?
19:51 joehoyle time to scale up
19:51 jslatts not sure. i don't know how big the images will be, but docker is very fast once the image is downloaded. i would guess sub 1 minute
19:51 CeBe joined #salt
19:52 jslatts we use instance store backed AMIs for this as well, and they are slower
19:52 jslatts but remove dependency EBS for autoscaling
19:53 joehoyle ok yeah
19:53 joehoyle sub 1 min is pretty good
19:54 jslatts to be specific, 1 minute after AWS does its thing
19:55 it_dude joined #salt
19:55 jslatts i saw demo on docker + AWS where it came up in 90 seconds
19:59 kballou joined #salt
19:59 joehoyle joined #salt
20:02 epcim joined #salt
20:05 it_dude joined #salt
20:07 percent20 joined #salt
20:08 danielbachhuber joined #salt
20:10 percent20 I am wanting to use salt with vagrant, but keep getting linked back to salty-vagrant which on their readme says has been Deprecated. Am so very confused on what to do.
20:10 percent20 is there a tutorial that I am missing somewhere that shows how to get started with it?
20:11 blueblade joined #salt
20:20 UtahDave percent20: the salt plugin is now shipped with vagrant
20:25 percent20 UtahDave, right but I follow that 'tutorial' on their site and I can't get it to work. So it leads me to that it is wrong and/or salt changed where it doesn't work anymore. Am curious if there is an updated something.
20:26 faldridge joined #salt
20:26 percent20 Problem I am hitting is even though I configure it, supposedly, to use local/masterless highstate call it still seems to try to connect to a master because I get a SaltReqTimeoutError
20:26 percent20 Am so new to salt. I have no idea what is going on.
20:27 jcockhren percent20: firstly
20:27 jcockhren https://docs.vagrantup.com/v2/provisioning/salt.html
20:28 jcockhren need a starter sample? one sec
20:28 jcockhren https://github.com/jcockhren/pytn-salt-vagrant
20:29 jcockhren a bit more complicated than the masterless setup (it has a master and minion)
20:29 druonysus joined #salt
20:29 druonysus joined #salt
20:30 percent20 jcockhren, Have been trying the vagrantup.com site to no avail. Even tried using salty-vagrant example.
20:30 percent20 jcockhren, cool will look at your example
20:32 epcim joined #salt
20:34 eriko joined #salt
20:34 epcim joined #salt
20:34 jacksontj anyone know what the intent of this wtag on the syndic was for?
20:35 vejdmn joined #salt
20:39 bemehow joined #salt
20:41 bemehow_ joined #salt
20:48 Dattas joined #salt
20:50 it_dude joined #salt
20:50 joehillen joined #salt
20:51 srage joined #salt
20:52 yusuket joined #salt
20:53 ndrei joined #salt
20:53 mateoconfeugo joined #salt
20:54 druonysuse joined #salt
20:54 druonysuse joined #salt
20:55 DaveQB joined #salt
20:55 micah_chatt joined #salt
20:58 jacksontj it looks like its only used to try and not return a job lookup if we are in the middle of writing one-- but if we already have a timeout-- it seems silly to pad the timeout a bit for an inflight write
21:01 jalaziz joined #salt
21:01 riftman joined #salt
21:10 ZombieFeynman joined #salt
21:12 gothix hey guys basically i want to filter by oscodename not os how would i do that in map.jinga
21:12 jimklo joined #salt
21:13 whiteinge gothix: filter_by takes a 'grain' kwarg. so just add grain='oscodename' to the function call
21:13 sarlalian joined #salt
21:14 gothix {% set tomcat = salt['grains.item']('oscodename') ({
21:14 gothix so this is wrong??
21:14 jacksontj anyone know whats up with build/lib/salt/runners/ ? seems to be the same set of runners in salt/runners/
21:15 rlarkin gothix I don't think you have to do a set
21:15 gothix {% set tomcat=salt[grains.oscodenam']
21:16 rlarkin that whole expression gets evaluated on the minion
21:16 lukeg left #salt
21:17 whiteinge gothix: probably something like this:  {% set tomcat = salt['grains.filter_by']({...dictionary of values separated by oscodename...}, grain='oscodename') %}
21:17 shoma joined #salt
21:18 rlarkin tomcat="{{ grains['oscodename'] }}"
21:18 gothix whiteinge, ill give that a try because filter_by wants to filter by OS and not any other value
21:18 rlarkin just put it directly
21:19 ajw0100 joined #salt
21:19 rlarkin or, put {{ grains['oscodename'] }} everyplace you're putting 'tomcat'
21:19 smcquay joined #salt
21:19 rlarkin the 'set' is an extra abstraction you might not need
21:20 gothix rlarkin, thanks ill test that out
21:20 whiteinge gothix: filter_by is just a terse way of doing: {...some dictionary of sub-dictionaries...}.get('somekey'), there's not a lot to it.
21:21 kermit joined #salt
21:22 cruatta joined #salt
21:23 whiteinge 'set' in jinja just defines a new variable:  {% set myvariable = 'somevalue' %}. it can be helpful to not have to do a grains lookup in multiple places (especially if you ever change which grain you're looking up)
21:24 ndrei joined #salt
21:25 schimmy joined #salt
21:30 modafinil_ I'm just checking out salt-master on ubuntu 14.04 -- I'm getting an error with the git fileserver backend (gitpython is installed) -- any idea where i could debug a bit mroe?
21:30 astol joined #salt
21:30 schimmy joined #salt
21:30 modafinil_ error is "2014-04-18 21:28:40,718 [salt.loaded.int.fileserver.gitfs][ERROR   ] Git fileserver backend is enabled in configuration but could not be loaded, is GitPython installed?" (answer: yes)
21:34 googolhash joined #salt
21:36 jimklo n00b here... trying to use one of the salt formula for installing java; using sun-java...  I can't seem to figure out how to get the state to work, it just keeps saying sun-java is not available.
21:36 jimklo can anyone point me in the right direction?
21:37 meteorfox joined #salt
21:38 modafinil_ turns out gitdb is required as well, and thats the failure im seeing
21:45 kermit joined #salt
21:47 joehoyle joined #salt
21:48 ahammond_ I'm having trouble passing a multi-line scalar into a formula
21:50 jimklo ahammond_: http://yaml.org/spec/1.1/#id858081
21:50 jimklo does that help?
21:51 stanchan joined #salt
21:51 ahammond_ jimklo yeah, I'm using that to specify the multi-line scalar in the pillar... the problem is once I want to pass the contents of the pillar to somewhere else.
21:51 ahammond_ Hmm. I'll post a snippet.
21:52 linuxlewis joined #salt
21:53 ahammond_ https://gist.github.com/ahammond/11066091    jimlko
21:54 ahammond_ er... jimklo ^
21:55 jimklo i think the indentation matters
21:56 jimklo ahammond_: the multi-line block needs to be indented to the right depth beneath - contents: |
21:56 ahammond_ jimklo right. that's the problem.
21:56 ahammond_ jimklo so... how do I do that?
21:57 ahammond_ jimklo it's getting passed in here: https://gist.github.com/ahammond/11066091#file-gistfile1-txt-L26
21:57 ahammond_ and... I'm not sure I see how to handle indenting. :(
21:58 jeffggardner joined #salt
22:00 ahammond_ joined #salt
22:00 UtahDave ahammond_: You should use  contents_pillar  instead of contents
22:00 UtahDave http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.managed
22:01 ahammond_ UtahDave yeah, that's what I was doing before, but I'm having trouble figuring out how to control access to the pillar
22:01 ahammond_ I'd like every machine to only be able to see their private keys.
22:01 UtahDave That's what pillar does
22:01 ahammond_ Hmm, I guess my formula got too clever. I'll dumb it back down
22:02 UtahDave You can either base it off how they match in the pillar top.sls
22:02 UtahDave or you can use jinja to fence off each minion to its proper value
22:02 ahammond_ UtahDave So... how do I have a set of files / whatevers such that machine A sees it's private key when it goes to pillar['my_private_key']?
22:03 ahammond_ UtahDave I can see how I could have per machine pillars, but for some reason I'm really not liking that design.
22:03 ahammond_ UtahDave I'm thinking specifically in the context of github deploy keys. I can't be the first guy to solve this. :)
22:04 UtahDave you can have one pillar   sls file that uses jinja.
22:04 UtahDave Let me put up an example real quick
22:04 ahammond_ @UtahDave thanks man!
22:06 joehoyle joined #salt
22:08 gothix has anyone in her come from using puppet or chef
22:09 ahammond_ gothix I've used chef and a little puppet
22:09 gothix and if so why the change
22:09 ndrei joined #salt
22:09 modafinil_ gothix: puppet is slow and awkward and doesn't have a push model, that's why i'm here :)
22:10 linuxlew_ joined #salt
22:10 UtahDave ahammond_: https://gist.github.com/UtahDave/11066593
22:10 ahammond_ gothix that's a long discussion. the brief version is that salt has good, working orchestration that's fast. Following the perl mantra, it makes easy things easy and hard things possible.
22:10 UtahDave you'll want to test that first, but that's the idea
22:10 ahammond_ puppet eventually becomes impossible. chef starts out impossible.
22:10 ahammond_ @UtahDave thanks, I'll wrap my brain around it.
22:10 undefinednull joined #salt
22:12 ahammond_ @UtahDave uh... that's soooo much simpler than I expected.
22:12 UtahDave yep!  Then you can just use   content_pillar  and you avoid all the yaml indenting issues
22:12 blueblade joined #salt
22:14 jnials joined #salt
22:16 tyler-baker left #salt
22:17 jimklo I'm having a hard time figuring out how to get the sun-java formula from https://github.com/saltstack-formulas/sun-java-formula, I've manually cloned and copied the sun-java dir into my /srv/salt/ dir.  I can't seem to figure out how to make it work.  I just keep getting 'sun-java is not available' when I run salt.
22:19 utahcon so if pkg.install 'package' works, why wouldn't it work in a state file?
22:19 utahcon tells me the package failed to install
22:19 ahammond horray! nick reclaimed!
22:20 * jimklo playing a tiny violin for ahammond. :)
22:23 tligda joined #salt
22:24 bhosmer joined #salt
22:25 linuxlewis joined #salt
22:27 linuxlewis joined #salt
22:27 whiteinge utahcon: centos perchance?
22:28 utahcon whiteinge: ubuntu (please don't laugh, not my choice)
22:28 utahcon the package is supervisor
22:29 utahcon apt-get install supervisor works, salt \* pkg.install supervisor, but putting it in the state fails :(
22:29 utahcon so weird
22:29 whiteinge utahcon: :)  i'm not sure about ubuntu. might be helpful to run that state with ``salt-call -l debug`` directly on the minion so you can see what actual commands it's trying to run
22:30 utahcon not a bad idea
22:30 whiteinge there was a recent issue with centos where the state was trying to call out to ``repoquery`` to do some checks which the execution module doesn't do
22:30 whiteinge if it wasn't installed the state would fail
22:31 alunduil joined #salt
22:36 obimod how are you supposed to segregate data in the pillars if not by grains?
22:37 obimod segregation by grains is inherently insecure since grains are controlled by the minion
22:39 pjs /2
22:40 jcockhren hey has there been some major changes in the expected format of pillar data between 0.17.5 and 2014.1?
22:41 smcquay joined #salt
22:45 whiteinge obimod: secure pillar data must be targeted by the minion ID
22:46 whiteinge jcockhren: i don't think so. seeing a problem?
22:47 ZombieFeynman joined #salt
22:48 obimod gotchya!
22:49 conan_the_destro joined #salt
22:49 jcockhren https://gist.github.com/jcockhren/15c4e4e17343609a2e36
22:49 jcockhren whiteinge: ^
22:50 jcockhren whiteinge: certainly in 2014.1.3
22:50 jcockhren (installed from pip)
22:50 utahcon whiteinge: turns out the package had been loaded as a dependency through pip by one of the plugins, was an order of install issue
22:50 utahcon thanks for helping!
22:56 whiteinge jcockhren: double-checking on 2014.1.3
22:58 Ryan_Lane if I have two things that need to be created, but have the same resource name, I need to make a state module for each, right? I can't make them as separate functions of the same state module?
22:58 Ryan_Lane I should clarify that: I'm writing state modules
22:59 faldridge joined #salt
23:02 joehoyle hey, I'm trying to use gitfs, however I get "Git fileserver backend is enabled in configuration but could not be loaded, is GitPython installed?", even after I ran "pip install GiyPython"
23:02 jimklo how would one unlink a relink a path within a state?
23:02 joehoyle howerver, my knowledge of pything is not good, is it possible my pip is installing modules somwhere else :S
23:03 Ryan_Lane nevermind, I got an answer via a pm. it's: they should be separate states
23:06 whiteinge jcockhren: sorry for the slow reply. i can't reproduce on 2014.1.3. wonder if something got cached/stuck somewhere
23:07 whiteinge jcockhren: does running salt-run cache.clear_pillar help?
23:07 gothix joined #salt
23:07 gothix joined #salt
23:14 kermit joined #salt
23:17 cgar joined #salt
23:18 thayne joined #salt
23:21 pjs does the latest salt version support freebsd pkg app? ie, on fbsd 10.0 pkg_add, etc. are no more. Everything is the new pkg tool now
23:22 jcockhren whiteinge: I'll take a look
23:29 meteorfox joined #salt
23:30 homelinen_ joined #salt
23:30 homelinen joined #salt
23:34 bemehow joined #salt
23:35 bemehow_ joined #salt
23:37 ekristen joined #salt
23:39 jnials joined #salt
23:47 ajolo_ joined #salt
23:49 joehoyle joined #salt

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