Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-03-24

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

All times shown according to UTC.

Time Nick Message
00:06 bhosmer joined #salt
00:12 thedodd joined #salt
00:17 StDiluted joined #salt
00:17 dman777 is there a vim indention plugin for salt stack?
00:20 rojem joined #salt
00:22 analogby1e dman777: https://github.com/saltstack/salt-vim
00:22 dren joined #salt
00:23 fllr joined #salt
00:26 dman777 analogby1e: thanks
00:36 dman777 how do I create a softlink in ubuntu for nginx from /etc/nginx/sites-available/cloudimage.conf to /etc/nginx/sites-available/sites-enabled/?
00:41 dman777 nm
00:43 scoates pyobjects does weird stuff with scoping. )-:
00:49 n8n joined #salt
00:54 mafrosis dman777: look at file.symlink
00:55 Ryan_Lane joined #salt
01:04 dman777 http://bpaste.net/show/BCmi2RvKzYCQ9aoy6YGs/ It seems nginx isn't pulling in my custom source files. I have in init.sls in service: - require: - pkg: nginx with nginx.sls containing the states for the source files
01:04 dman777 sorry, please use http://bpaste.net/show/9R89SyKnGfcjbIqr6YDK/
01:05 dman777 err...I mean this one: http://bpaste.net/show/ydi4xfKhn3f43oFrXqTs/
01:07 oz_akan__ joined #salt
01:08 bhosmer joined #salt
01:16 dman777 http://bpaste.net/show/ydi4xfKhn3f43oFrXqTs/ hmm...how come my source files aren't getting placed on the minion?
01:23 xzarth_ joined #salt
01:25 dccc joined #salt
01:29 ckao joined #salt
01:31 napper2 joined #salt
01:34 fllr joined #salt
01:42 justlooks_ joined #salt
01:47 ninkotech joined #salt
01:50 justlooks_ how can i cancel the running job ? for instance i run job "salt 'saltslave2' state.sls tomcate" but what i really want to run is "salt 'saltslave2' state.sls tomcat"
01:59 alunduil joined #salt
02:06 dave_den joined #salt
02:07 oz_akan_ joined #salt
02:08 thedodd joined #salt
02:08 dman777 http://bpaste.net/show/ydi4xfKhn3f43oFrXqTs/ hmm...how come my source files aren't getting placed on the minion?
02:09 dman777 added top.sls to bpaste: http://bpaste.net/show/CImYwCiGARKepWwYeDf2/
02:14 Networkn3rd joined #salt
02:27 manicouman joined #salt
02:27 manicouman left #salt
02:28 yomilk joined #salt
02:40 xl1 joined #salt
02:45 faldridge joined #salt
02:45 vejdmn joined #salt
02:49 googolhash joined #salt
02:52 Furao joined #salt
02:59 tvle joined #salt
03:03 StDiluted joined #salt
03:08 oz_akan_ joined #salt
03:10 mgw joined #salt
03:16 dman777 khttp://bpaste.net/show/CImYwCiGARKepWwYeDf2/ how come my source files aren't getting placed on the minion?
03:19 RandalSchwartz where are you telling it to look at nginx/nginx.sls ?
03:20 faldridge joined #salt
03:31 faldridge joined #salt
03:35 jeremyBass left #salt
03:37 accb joined #salt
03:41 ipalreadytaken joined #salt
03:41 analogbyte joined #salt
03:42 kermit joined #salt
03:46 frasergraham joined #salt
03:55 justlooks_ anyone can help this ? https://gist.github.com/justlooks/9733857
03:57 RandalSchwartz what's in tomcat/tomcat_conf
04:09 oz_akan_ joined #salt
04:14 acabrera joined #salt
04:16 Hipikat left #salt
04:27 anuvrat joined #salt
04:31 yomilk joined #salt
04:33 tvle joined #salt
04:35 StDiluted joined #salt
04:42 ipalreadytaken joined #salt
04:44 ipalread_ joined #salt
04:55 StDiluted joined #salt
04:55 thayne joined #salt
05:02 fragamus joined #salt
05:07 fllr joined #salt
05:09 Furao left #salt
05:29 ravibhure joined #salt
05:49 StDiluted joined #salt
05:55 ipalreadytaken joined #salt
06:03 srage joined #salt
06:12 justlooks_ joined #salt
06:19 googolhash joined #salt
06:24 StDiluted joined #salt
06:30 justlooks_ why i define a pillar in a salt module ,but when i sync it to minion ,minion can still not find it?
06:30 nedrocks joined #salt
06:31 bhosmer joined #salt
06:31 justlooks_ how can i delete old cached pillar?
06:35 anuvrat joined #salt
06:37 jeddi justlooks_: salt  minion-host-name  saltutil.refresh_pillar
06:38 jeddi justlooks_: i've set up a short bash script called 'salty' that just does that for a given ${1}, and then runs a highstate against it with a  test=True - makes the testing iteration process a bit faster.
06:41 justlooks_ jeddi: here what i do https://gist.github.com/justlooks/9733857 ,i am very confused
06:42 oc joined #salt
06:49 justlooks_ anyone know what's problem?
06:59 fllr joined #salt
06:59 cewood joined #salt
07:00 _fllr_ joined #salt
07:03 CeBe joined #salt
07:09 fllr joined #salt
07:10 mafro joined #salt
07:11 fllr joined #salt
07:13 mafro yo salters
07:14 mafro anyone know why I might be getting “[ERROR   ] Git fileserver backend is enabled in configuration but could not be loaded, is GitPython installed?”
07:14 mafro GitPython is installed
07:18 Doqnach`` joined #salt
07:19 ravibhure1 joined #salt
07:19 raizyr_ joined #salt
07:19 EWDurbin_ joined #salt
07:19 goncalo joined #salt
07:19 oc- joined #salt
07:19 __alex joined #salt
07:20 ay joined #salt
07:21 rawtaz joined #salt
07:21 JesseCW joined #salt
07:22 Eugene mafro - any of a thousand things. You need to restart the minion, it isn't looking in the right place, my dog ate my homework, the lib isn't installed to the right place, it didn't actually get installed(it happens surprisingly often), alignment of the celestial sphere....
07:22 yomilk joined #salt
07:22 mafro Eugene: thanks
07:23 Eugene My guess is one of the first two.
07:23 mafro So… assuming celestial alignment and the contents of my dog’s stomach aren’t in play
07:23 mafro Where should the lib be?
07:23 Eugene Beats me.
07:23 mafro The error message is being spat out of my Salt master, so the minion isn’t involved yet
07:24 Eugene Cool. I really have no idea, then.
07:24 keekz joined #salt
07:24 mafro Eugene: have you had this problem before?
07:24 Eugene And being 330AM, I bid you luck.
07:24 Eugene Not with Salt, but I have slammed my head against Python in similar scenarios
07:24 mafro Eugene: congrats on being the first to reply, and good night
07:24 Eugene It came out to be a bad path in a .deb we were using
07:30 Bosch[] Hi, is it a known issue that tab-completing hostnames can be slow in salt's  cli, tabbing commands is as slow it could make me think the process is hanging.
07:31 gcbirzan If you have any minions that aren't replying, yes.
07:31 Bosch[] Ah. That queries the minions every time?
07:32 ChaosPsyke joined #salt
07:32 gcbirzan Though, looking at mine, it shouldn't take more than 2 seconds
07:32 gcbirzan yeah.
07:32 gcbirzan Not _every_ time, but when you try to complete an execution module function, it does:     _salt_coms="$(salt '*' --timeout 2 --out=txt -- sys.list_functions | sed 's/^.*\[//' | tr -d ",']" )"
07:33 Iwirada joined #salt
07:34 gcbirzan It's good 'cause it takes your local stuff, but it can get slow if you have offline minions  :P
07:35 gcbirzan Which is quite annoying in general, actually, especially in the way we're using it, where we have dynamic minions coming up/going down all the time
07:39 ndrei joined #salt
07:42 yomilk joined #salt
07:43 Bosch[] Thanks
07:52 sp0ck joined #salt
07:53 sp0ck Is there any firewall rules needed on minion? When I set default INPUT policy to drop, master looses connectivity to minion.
07:53 sp0ck manage.status --> says the minion is down.
07:54 sp0ck And the documentation doesnt specify why this is happening, I was in the impression that minions' only make outbound connections.
07:56 slav0nic joined #salt
07:56 bitmand sp0ck: afaik they only make connections to the master ( I have a minion behind firewall, nat and whatnot and works fine ).
07:56 mafro for anyone who finds this via google search..
07:57 Bosch[] I'm not too into iptables, however, does the INPUT drop rule drop every inbound package?
07:57 mafro “Git fileserver backend is enabled in configuration but could not be loaded” was caused by version 0.1.7 of GitPython being installed
07:57 bitmand sp0ck: do you know if your firewall is statefull?
07:58 mafro pip 1.4 and later need the --pre parameter to get the 0.3.x beta releases
07:58 sp0ck batman, I am not sure, I am using iptables
07:59 bitmand sp0ck: does is work if you add a rule that allow packets from the master to the minion? ( along with default input drop )
08:04 mikko joined #salt
08:06 sp0ck bitmand, Let me test this and get back to you.
08:08 EnTeQuAk left #salt
08:08 harobed joined #salt
08:09 sp0ck bitmand, oh ow. I have shut myself out by flushing the rules, let me fix this first and I will do some more test and get back to you.
08:09 bitmand sp0ck: you are not the first one to try that :)
08:13 bitmand sp0ck: I was skimming this: https://wiki.archlinux.org/index.php/simple_stateful_firewall - I'm not iptables user myself, but it is a good introduction.
08:13 ckao joined #salt
08:15 MrTango joined #salt
08:21 n8n joined #salt
08:21 sp0ck joined #salt
08:24 anuvrat joined #salt
08:29 ndrei joined #salt
08:30 giantlock joined #salt
08:32 ajw0100 joined #salt
08:34 yota joined #salt
08:34 googolhash joined #salt
08:36 Katafalkas joined #salt
08:36 yomilk_ joined #salt
08:38 harobed_ joined #salt
08:38 slav0nic joined #salt
08:45 mafro joined #salt
08:46 Katafalkas joined #salt
08:53 rojem joined #salt
08:53 srage_ joined #salt
08:55 ggoZ joined #salt
09:01 fllr joined #salt
09:11 topochan joined #salt
09:11 topochan joined #salt
09:13 krak3n` joined #salt
09:13 tvle joined #salt
09:25 krak3n` joined #salt
09:28 Katafalkas joined #salt
09:29 slav0nic joined #salt
09:31 mike25ro joined #salt
09:35 q1x joined #salt
09:37 q1x Hi guys, what would be the best way to get the pkgrepo.managed state to use a proxy server for apt-key management?
09:46 sdlarsen joined #salt
09:47 m0s joined #salt
09:48 sdlarsen I've built an external pillar. calling salt '*' pillar.items shows it under ext_pillar with it's configured values. I can't seem to reference it's values from my state files though. Salt log says '[CRITICAL] Specified interface <my pillar name> is unavailable. What am I missing?
09:52 mekstrem joined #salt
09:55 slav0nic joined #salt
09:55 slav0nic joined #salt
10:01 mekstrem_ joined #salt
10:01 fllr joined #salt
10:02 Kenzor joined #salt
10:03 johtso joined #salt
10:07 mekstrem joined #salt
10:09 gammalget joined #salt
10:11 symroe Is anyone using salt to deploy a ruby app in a way that has a concept of a 'build'?  That is, a versioned collection of the pillar, salt and application code that gets deployed in an atomic way?
10:12 symroe Currently I'm just doing a git pull for my app code, but that's making it hard to know what is deployed on each highstate
10:14 ndrei joined #salt
10:20 jeddi symroe: not sure if there's a best practice for this one.   can you package it up centrally and deploy with salt?
10:20 jeddi i know that's what you're trying to avoid, probably :)
10:23 sp0ck joined #salt
10:23 symroe Yeah, it's a hard one.  I was thinking about just git tagging our 3 repos each time we deploy, and then adding that tag to the pillar
10:24 sp0ck joined #salt
10:25 jeddi i've used fabric in the past for this kind of thing.  i thought they were trying to meld those features into salt, but hvaen't tracked that usage.
10:26 carlos_ joined #salt
10:27 symroe So I'm using fabric to update the master and run highstate, but even that feels like a bit of a hack
10:27 bhosmer joined #salt
10:28 symroe I'm rsyncing the project files to the master and then calling highstate on it.  I could use fabric to make a build as well, but that then feels really hacky, and I don't want to rely on loads of custom made fabric stuff
10:29 jeddi symroe: complete empathy on wanting to avoid hacky.  OTOH, make it work, document it, and then refactor. .. . is perfectly valid approach :)
10:29 symroe Agreed :)
10:29 jeddi are you using a build / ci system around your dev team for the ruby stuff - like jenkins - that could generate an npm (?) you can then deploy?
10:30 jeddi really struggling for the terminology here - as i say, not a ruby guy.
10:30 symroe I guess I'm trying to figure out if there are existing patterns in salt though.  It could be that this is a solved problem, and I'm just missing something
10:30 symroe I'm not a ruby guy either :)
10:30 jeddi existing patterns - best practices - i think are not as bountiful as we could hope.
10:31 symroe I think you mean a gem.  That wouldn't get us much further than a git tag I don't think
10:31 jeddi US will be coming online in the next few hours, so if you're able to wait around ...
10:31 symroe Yeah, I'll ask again this afternoon
10:31 jeddi ah, yeah, gems.  that's the things.  :)
10:32 jeddi or generate .deb / .rpm direct from a dedicated build tool, and deploy that.
10:33 jeddi for me that's the gold standard approach .. but i'm old school and like native (distro) packages.
10:33 ambientsound symroe: http://xkcd.com/1335/
10:40 N-Mi joined #salt
10:48 masm joined #salt
10:55 babilen Hi all -- I am trying to mount an NFS4 mount and ensure that the mountpoint has owner/group set to given values. I am trying: http://paste.debian.net/89407/ but it fails with: "Failed to change user to www-data; Failed to change group to www-data" (no more details) -- What is the problem and how can I deal with that?
10:58 babilen If I don't include the require on "- file: /var/www/foo/foo_static" in mount.mounted it is mounted correctly, but the owner:group are nobody:nogroup which is, obviously, not what I want.
11:01 fllr joined #salt
11:05 slav0nic joined #salt
11:05 slav0nic joined #salt
11:05 CeBe joined #salt
11:07 sdlarsen having set extension_modules to /etc/salt/extension_modules and created a pillar directory where my external pillar python file is located (and is seen by the server - a pyc file shows up). Configuring it in ext_pillar in master seems not to work - log says [CRITICAL] Specified interface <my external pillar name> is not found... why?
11:09 apergos ugh.. this must be easy but I'm not finding it fast, what's a quick way from the command line to have salt show a list of hosts that match an expression (e.g. hosts it knows about, not hosts that respond to test ping or whatever)?
11:10 babilen apergos: "salt-key -L" might be what you are looking for
11:13 sgviking joined #salt
11:18 yomilk joined #salt
11:24 yomilk joined #salt
11:24 baniir joined #salt
11:25 apergos babilen: that lists them all; I want to give the command an expression for matching hosts and have it tell me the ones it knows about
11:26 apergos eg mystery-command 'storage*.blah'  and have it tell me what that would expand to
11:27 babilen Yes, I do understand. I am not aware of a salt specific command (which really doesn't mean that there isn't one), but you should be able to grep in the "salt-key -L" output relatively easily
11:29 babilen Or within the output of "salt-run manage.up" which is probably better, but still not what you want :-/
11:29 ndrei joined #salt
11:31 topochan joined #salt
11:34 gcbirzan Hm, is there a way to have an empty sls?
11:35 gadams999 joined #salt
11:35 gcbirzan Or at least a no-op. I'm tryign to just organise stuff and want to create them in advance, without any data, but don't want salt to complain about it :P
11:36 apergos doing that for anything but a simple wildcard is not going to work... let me see what else can be done.  thanks
11:40 sp0ck joined #salt
11:41 apergos if salt had a noop that instead of running a command on the minion would just emit the empty string for it, that would be ideal
11:42 sdlarsen can anyone tell me how to make an ext_pillar interface available?
11:44 gcbirzan apergos: salt '*' test.outputter 'foo' ?
11:46 jcsp joined #salt
11:46 apergos gcbirzan: no, because it actually runs something on the minion
11:46 apergos I want the list of minions salt knows about that match
11:46 apergos not the list that are responsive
11:48 ndrei joined #salt
11:48 gammalget joined #salt
11:51 gcbirzan ah, hm
11:52 dhananjay_ joined #salt
11:54 apergos I would be ok with test-ping if I could tell salt 'tell me about nonresponsive hosts in the same format as the responsive ones,  without the job id and that other cruft that '-v' adds (and might vary from version to version too)
11:54 gmoro joined #salt
11:56 apergos I mean I can do (and have done)  `salt -v --out raw 'hostexpr-here' test.ping | grep '{' | mawk -F"'" '{ print $2 }'` but gee... it just seems like there should be anicer way
11:56 babilen So, is there no way to ensure that mountpoints are owned by a given user:group ?
11:58 agend joined #salt
12:01 fllr joined #salt
12:08 Gifflen joined #salt
12:23 nicksloan joined #salt
12:30 ndrei joined #salt
12:32 msmith_ joined #salt
12:33 bastion1704 joined #salt
12:34 millz0r joined #salt
12:37 vortec_ joined #salt
12:37 yomilk_ joined #salt
12:45 moos3 joined #salt
12:45 Networkn3rd joined #salt
12:46 gnugnu9 joined #salt
12:48 jslatts joined #salt
12:54 Gifflen joined #salt
12:58 dhananjay_ joined #salt
12:59 mpanetta joined #salt
13:01 ndrei joined #salt
13:01 fllr joined #salt
13:02 ninkotech joined #salt
13:02 Gifflen joined #salt
13:05 toastedpenguin joined #salt
13:06 mpanetta joined #salt
13:07 mpanetta joined #salt
13:07 oz_akan_ joined #salt
13:10 StDiluted joined #salt
13:14 napper2 joined #salt
13:16 ndrei joined #salt
13:16 Damoun joined #salt
13:16 dhananjay_ joined #salt
13:17 bhosmer joined #salt
13:22 juicer2 joined #salt
13:22 sroegner joined #salt
13:22 lsc joined #salt
13:22 lsc Hey guys
13:23 lsc is the 2014.01 branch considered stable?
13:23 lsc or is it the dev branch for 2014.2?
13:23 tmmt joined #salt
13:23 epicserve joined #salt
13:24 diegows joined #salt
13:28 xl1 joined #salt
13:29 ninkotech joined #salt
13:30 dhananjay_ joined #salt
13:32 yomilk joined #salt
13:35 ninkotech_ joined #salt
13:37 jcsp1 joined #salt
13:38 thayne joined #salt
13:38 jcsp1 joined #salt
13:40 bastion1704 joined #salt
13:41 ipmb joined #salt
13:43 elfixit1 joined #salt
13:48 gadams999 left #salt
13:51 dhananjay_ joined #salt
13:54 jeremyBass1 joined #salt
13:54 gildegoma joined #salt
13:56 dhananjay_ joined #salt
13:57 xl1 joined #salt
13:59 faldridge joined #salt
14:01 R_Shackleford joined #salt
14:01 fllr joined #salt
14:02 sroegner joined #salt
14:06 StDiluted joined #salt
14:06 sroegner joined #salt
14:07 apergos I guess salt-cp doesn't make any attempt to preserve ownership/perms ?
14:07 kedo39 joined #salt
14:09 ndrei joined #salt
14:10 topochan_ joined #salt
14:12 kvbik joined #salt
14:14 rallytime joined #salt
14:15 faldridge joined #salt
14:16 cachedout_home joined #salt
14:16 GradysGhost joined #salt
14:19 bhosmer joined #salt
14:21 bastion1704 joined #salt
14:22 DaveQB joined #salt
14:23 seapasulli joined #salt
14:26 penguin_dan joined #salt
14:28 spiette joined #salt
14:31 jrdx joined #salt
14:36 jalbretsen joined #salt
14:37 StDiluted joined #salt
14:38 GradysGhost joined #salt
14:38 sroegner joined #salt
14:39 ravibhure joined #salt
14:39 cnelsonsic joined #salt
14:39 alunduil joined #salt
14:40 timoguin joined #salt
14:43 fllr joined #salt
14:44 gnugnu9_ joined #salt
14:45 tvle joined #salt
14:45 cheus Does anyone know if there's a way to specify state options as a dictionary instead of a list of dictionaries?
14:47 wendall911 joined #salt
14:48 thedodd joined #salt
14:48 opapo joined #salt
14:50 bhosmer joined #salt
14:53 GradysGhost joined #salt
14:56 quickdry21 joined #salt
14:58 ajprog_laptop1 joined #salt
14:59 q1x ok, so I'm using saltstack to install mysql server. That works, now I used the mysql_user state to set the mysql root password from pillar data. That also works neatly :)
15:00 q1x now, how do I get the minion to start using the password from the pillar for subsequent mysql states?
15:01 kaptk2 joined #salt
15:01 baniir joined #salt
15:02 q1x docs say to edit the minion config file, but that seems inappropriate. Then I would still have the root pw on the system in plain text. Would be better to load it from pillar on run time, no?
15:03 fllr joined #salt
15:08 pentabular joined #salt
15:09 tyler-baker joined #salt
15:09 StDiluted joined #salt
15:09 jgelens Any idea when 2014.1.1 will be out
15:09 jgelens ?
15:10 mike25ro q1x: i would keep the password on the master... it gets compiled and sent to the minion.. so there is not the risk of keeping pass on the minion
15:12 q1x mike25ro: that is exactly what I want to do
15:12 q1x mike25ro: but the docs state I should edit /etc/salt/minion and insert the password there
15:13 mike25ro err... q1x not sure about that... seems a bit stupid..
15:13 mike25ro maybe you can chmod 600 the minion?
15:14 q1x mike25ro: http://docs.saltstack.com/ref/modules/all/salt.modules.mysql.html <- configuration
15:15 q1x Oh, I think I found it...
15:15 mike25ro what?
15:15 q1x http://docs.saltstack.com/ref/states/all/salt.states.mysql_user.html <- seems I can pass username and password as arguments
15:16 mike25ro cool
15:16 mike25ro good to know
15:16 q1x connection_user and connection_pass
15:16 mike25ro you can then keep that pass in a jinja - or so i do
15:17 RandalSchwartz see http://docs.saltstack.com/ref/modules/all/salt.modules.mysql.html#module-salt.modules.mysql for setup
15:17 RandalSchwartz under "configuration"
15:17 mike25ro haha i was looking exactly at that one
15:17 RandalSchwartz haven't used it (hate MySQL), but there you go
15:18 mgw joined #salt
15:23 q1x RandalSchwartz: thanks, but like I said: putting mysql root password in minion config seems like a bad idea
15:24 q1x however, the mysql_user state docs seem to indicate you can override these parameter from the state file
15:24 q1x but, it seems mysql_database states do not adhere to this, since I'm getting an error
15:27 frasergraham joined #salt
15:27 q1x the are not listed here, so that might be the case then  http://docs.saltstack.com/ref/states/all/salt.states.mysql_database.html
15:28 amckinley joined #salt
15:28 q1x on the other hand, it states " **connection_args"  whatever those might be ... :-/
15:28 oz_akan_ joined #salt
15:29 RandalSchwartz the minion value will be sent only to the approved minion though
15:29 fllr joined #salt
15:29 RandalSchwartz you'll need it at some point
15:29 q1x nevermind, I'm an idiot >.<
15:29 fragamus joined #salt
15:30 q1x made a typo, which meant my password was being ignored
15:30 RandalSchwartz heh
15:30 q1x works ok now :)
15:30 RandalSchwartz damn typos
15:30 q1x no password in minion config needed \o/
15:30 rallytime joined #salt
15:32 rlarkin joined #salt
15:32 q1x RandalSchwartz: yeah, I figured out the pillar stuff to make sure only the right minions get the password, that was working awesomely :-)
15:32 babilen Just to clarify: If I want to have different pillar configuration (just one key in the dictionary) I have to duplicate *everything* and make sure that I target the minion only once with the duplicated data or is there a nicer way?
15:33 sandbender1512 joined #salt
15:34 Ahlee is pillar supposed to work on multimaster setups?
15:35 aranhoide joined #salt
15:35 aranhoide is there any way to pin the versions of the instances launched with salt-cloud so it will match the version in the master?
15:36 aranhoide the salt versions, that is
15:36 q1x babilen:  use jinja in your pillar to construct if/else statements?
15:37 aranhoide if I launch a new minion in Digital Ocean I can't run state.highstate on it: "No Top file or external nodes data matches found"
15:37 babilen q1x: Would I have access to grains in there?
15:37 aranhoide btw upgrading the master doesn't fix that
15:38 q1x babilen: yup, I'm using id grains as we speak
15:38 utz joined #salt
15:38 babilen q1x: I essentially want to say: Use all the keys from pillar foo.bar but replace foo.bar.quux:q with "- ... - ... - ..."
15:38 kaptk2 joined #salt
15:38 q1x oh, like that...hmm
15:38 babilen q1x: Well, what I *really* want is to extend the pillar and add one more item to a list, but that is unfortunately not yet supported
15:39 q1x not sure about that one
15:39 babilen It just feels horribly clumsy that I have to duplicate all that information twenty times, just because I want to change a little bit.
15:40 babilen (I know that there shouldn't be such differences, but my sysads want it that way)
15:40 q1x babilen: I think you should be able to nest grain info
15:40 bastion1704 joined #salt
15:41 babilen q1x: Could you show me an example for a pillar that uses the id grain? Shouldn't be hard I know, but why not see something that is being used :)
15:43 q1x babilen: http://docs.saltstack.com/topics/tutorials/pillar.html#paramaterizing-states-with-pillar
15:43 q1x sorry, scroll down to the end
15:44 * q1x afk
15:45 babilen q1x: I'll use something like that until a salt dev finally gets around to implement https://github.com/saltstack/salt/issues/2466
15:46 meteorfox joined #salt
15:46 joshbaptiste joined #salt
15:48 forrest joined #salt
15:49 tvle left #salt
15:50 bhosmer joined #salt
16:00 bt joined #salt
16:01 whiteinge babilen: jinja macros are perfect for avoiding the kind of repetition
16:02 mike25ro yeah jinjas are great
16:04 babilen whiteinge: It works fine with that extra logic in the pillar. I dislike it a little as I'm spreading my logic around (both states and pillars), but it is a much nicer solution than stupidly duplicating entries.
16:05 babilen And the more I think about it the more I like it ;)
16:05 icarus_ joined #salt
16:06 gadams999 joined #salt
16:08 cloudpie joined #salt
16:14 KyleG joined #salt
16:14 KyleG joined #salt
16:15 mgw joined #salt
16:18 sp0ck joined #salt
16:21 fllr joined #salt
16:21 pentabular joined #salt
16:21 fxhp whiteinge: would you prefer a macro + for loop, or just a for loop?
16:23 fllr joined #salt
16:23 pentabular1 joined #salt
16:24 scoates anyone familiar with pyobjects? (borgstrom happen to be around now?). I don't know why, but I can't reliably `import os` in there, without wrapping everything in a function.
16:24 kzx joined #salt
16:25 pentabular joined #salt
16:26 icarus_ joined #salt
16:26 scalability-junk joined #salt
16:26 whiteinge fxhp: which ever one results in the most legible state :)
16:27 whiteinge Ahlee: yes, pillar should work on multi-master so long as you have pillar_roots identical on each
16:28 Ahlee and of course, sync'd
16:28 rushm0r3 joined #salt
16:28 Ahlee which i believe is my issue
16:28 rushm0r3 joined #salt
16:34 kballou joined #salt
16:37 joehillen joined #salt
16:39 rushm0r3 joined #salt
16:39 n8n joined #salt
16:39 anuvrat joined #salt
16:39 yomilk joined #salt
16:40 rushm0r3 is there a simple way to just list all jobs by a minion_id ? seems I can only list by job id, list currently running jobs with a minion target, but not all regardless of state ?
16:40 smcquay joined #salt
16:41 timoguin rushm0r3, salt 'minion' saltutil.running
16:41 rushm0r3 that only shows running jobs though
16:42 timoguin ah yes
16:42 rushm0r3 do i need to use a returner and force all jobs to store to that ? That seems excessive
16:44 forrest you could use salt-run jobs.list_jobs, but that doesn't allow you to look at just one minion
16:44 chrisjones joined #salt
16:44 timoguin Yea, I'm not sure what the best way to handle that is.
16:44 rushm0r3 right, already tried that. feels like i shouldn't have to pattern match just to find jobs by minion though
16:45 rushm0r3 i assumed halite had some nice simple way to handle that so was looking through the code but didn't find anything yet
16:46 forrest rushm0r3, might be a good feature request to put in if there isn't already an issue open for it.
16:47 baniir joined #salt
16:48 thayne joined #salt
16:50 icarus_ joined #salt
16:50 druonysus joined #salt
16:50 schimmy joined #salt
16:51 joshbaptiste hello all..  reading the docs.. i'm tryin to discern the diff between pillar and state files
16:52 Chronojam joined #salt
16:52 schimmy1 joined #salt
16:52 joshbaptiste basically want to test a setup of dirs, add new users, set perms ... this is all pillar setup?
16:53 dvogt joined #salt
16:53 rushm0r31 joined #salt
16:55 rushm0r3 joined #salt
16:56 rgbkrk joined #salt
17:00 Calum joined #salt
17:00 Calum joined #salt
17:01 sdlarsen_ joined #salt
17:02 sdlarsen_ Having set extension_modules in master, and created a pillar directory, why is my custom pillar still unknown?
17:03 ndrei joined #salt
17:03 Gareth morning.
17:04 mgw joined #salt
17:04 dave_den because it is not being loaded, probably due to a syntax error. run your master with -l debug - there should be a message about it failing to load.
17:05 sdlarsen_ dave_den: ok, thanks
17:11 retr0h joined #salt
17:13 seblu42 joined #salt
17:13 Iwirada left #salt
17:13 rushm0r31 joined #salt
17:15 bhosmer joined #salt
17:15 tdilly joined #salt
17:19 baniir joined #salt
17:20 rushm0r3 joined #salt
17:20 pentabular joined #salt
17:21 pentabular1 joined #salt
17:23 pentabular joined #salt
17:23 KyleG joined #salt
17:23 KyleG joined #salt
17:23 moos3 is there away with reactor to remove un-reachable hosts ?
17:24 snuffeluffegus joined #salt
17:24 LucasCozy joined #salt
17:24 LucasCozy joined #salt
17:25 eculver joined #salt
17:26 Ahlee moos3: if there's an event, yeah.  There isn't one on 0.17.x, haven't tried 2014.x
17:27 moos3 k
17:28 Ahlee moos3: I crawl manage.down every 30 minutes, and delete keys that aren't responding
17:28 Ahlee since i run open mode when the hosts come back up, they readd and go about their way
17:28 Ryan_Lane joined #salt
17:28 moos3 yeah that makes sense
17:28 n8n joined #salt
17:29 Ahlee then nightly I sweep through AD for all servers, and if any are missing i currently raise an alert, hope to automatically respond to that list shortly
17:29 moos3 whats the easiest way to crawl the manage.down list
17:30 Ahlee i do it through bash, 'cause i'm lazy and it works.  for host in $(salt-run manage.down); do
17:30 Ahlee but, you could also create a salt runner on the box, and do it through python
17:32 Ahlee moos3: https://gist.github.com/jalons/2c1a576c85e2b51180f9
17:33 moos3 oh nice
17:33 Ahlee reload, i fixed up the default /etc/salt/master area :)
17:33 Ahlee otherwise, have fun
17:35 moos3 my nodes are all reporting     'manage.status' is not available.
17:36 Ahlee moos3: that should run on the master
17:36 Ahlee runner always runs on teh master
17:37 budrose joined #salt
17:37 moos3 when I run it manage.status on my minions thats what they return
17:37 Ahlee what do you mean on your minions?
17:37 Ahlee you should fire up a python interpretor on your master (same python that has salt modules installed to it)
17:37 Ahlee and then run that
17:38 suckaplease joined #salt
17:39 moos3 oh prefect :D
17:39 moos3 that worked better
17:40 xmltok joined #salt
17:42 napper2 joined #salt
17:43 Mr_N So I have a 17.5 minion, on which I can reproducibly demonstrate that refresh_pillar doesn't in fact refresh_pillar.
17:44 Mr_N I can't find any record that this got fixed in 2014.1, but I can't update due to other problems to test myself.
17:44 Mr_N Anyone know if this is just something broken in my config?
17:44 Mr_N (for the record, restarting the minion does in fact refresh the pillar data, but neither refresh_pillar nor sync_all do.)
17:45 jeremyfelt joined #salt
17:46 xmltok left #salt
17:46 allanparsons joined #salt
17:46 agenteo joined #salt
17:46 robawt hey salt peeps, is there a way to view the highstate in progress?
17:46 robawt is it the -l debug during a state.highstate option?
17:47 Mr_N -ldebug will show you "stuff"
17:47 Mr_N it may be a bit high on the noise to signal ratio
17:47 Mr_N but it'll be something.
17:48 robawt interesting ok, but that's a salt-call option i see
17:48 Mr_N salt-call is for doing local calls on a minion, IIRC
17:49 bastion1704 joined #salt
17:49 agenteo hi, the home brew installer for saltstack is broken… any suggestion on how to install it on osx? thanks
17:53 tdilly agenteo: Have you seen the bootstrap? https://github.com/saltstack/salt-bootstrap
17:53 agenteo nope, I will check it out, cheers!
17:54 tdilly agenteo: NP, not sure that it works on osx, but.. it might, good luck
17:56 fllr joined #salt
17:56 kermit joined #salt
17:56 scoates Mr_N: FWIW, I also could never get refresh_pillar to actually refresh
17:57 Mr_N >< good to know scoates...
17:57 ndrei joined #salt
17:57 scoates Mr_N: pillar.get is out of sync with pillar.items, and refresh_pillar doesn't sync them
17:57 Mr_N Indeed.
17:57 Mr_N That's exactly the symptom.
17:57 fllr joined #salt
17:57 dave_den pillar_refresh does sync the pillar.
17:57 arthabaska joined #salt
17:58 Mr_N Not for us dave_den; I can readily show you output to reproduce it.
17:58 dave_den the problem is probably you are calling salt-call saltutil.refresh_pillar.
17:58 dave_den this does not reload the salt-minion pillar data
17:58 Mr_N What is the correct mechanism to cause pillar.items and pillar.get to converge, then?
17:58 scoates yeah, sorry I was unclear. What I meant is that it doesn't cause pillar.get and pillar.items to be in sync
17:58 scoates I don't know what's going on internally from my testing
17:59 fllr joined #salt
17:59 dave_den issue the saltutil.refresh_pillar from your salt master
17:59 Mr_N I did.
17:59 dave_den hrm, then that is not correct behavior
17:59 Mr_N In all cases this was using salt <target> saltutil.refresh_pillar.
17:59 Mr_N or salt '*'
17:59 Mr_N I was pretty sure this was verbatim out of the documentation
17:59 dave_den what version?
17:59 Mr_N 17.5
18:01 Mr_N (additionally, this is on centOS master + minion)
18:02 ravibhure joined #salt
18:02 srage joined #salt
18:02 dave_den hrm, just confirmed it's working on my 0.17.5
18:02 Mr_N Interesting.  Is there any debug output I could provide?
18:03 Mr_N I can demonstrate calling pillar.items and pillar.get, that they show different results, and that after refresh_pillar they still differ
18:03 dave_den just run your minion in -l debug and issue the saltutil.refresh_pillar on the master. it should show it
18:03 Mr_N but after a minion restart they no longer do.
18:04 Mr_N One moment, I'll grab + pastebin it.
18:04 dave_den pillar.get works on the cached pillar data, pillar.items calls the master to get the updated pillar data
18:05 Mr_N I'm aware.  I was under the impression refresh_pillar would cause them to converge, though?
18:05 kvbik joined #salt
18:05 Mr_N Or at least sync_all, since there was a conversation about that functionality in an issue I found from a while ago.
18:07 Mr_N Here we go dave_den; got it reproduced, will have a link shortly.
18:07 Mr_N It accepts the refresh_pillar job on the minion, I observe that, but then get returns the same value that it did prior.
18:07 dave_den Mr_N: yes, if you send a saltutil.refresh_pillar form the master to a minion then both the minion and salt-call pillar.get should see the new data
18:08 Mr_N Yes, pillar.get does not see the new data.
18:08 dave_den Mr_N: are you using any ext_pillar like gitfs?
18:08 Mr_N No.
18:08 Mr_N I'm putting together a comprehensive little report with all relevent files
18:08 dave_den Are you sure you're changing the pilalr data and saving the file to disk :)
18:08 dave_den heh
18:08 Mr_N I may be insane, that's not discounted :P
18:08 Mr_N but I'm pretty sure.
18:08 dave_den alright, well yeah it should work
18:08 dave_den anyway, off to lunch. bbl
18:08 Mr_N Yep.  I'll poke you with this whenever you're back.
18:08 dave_den cool
18:09 Mr_N enjoy
18:09 allanparsons how can i watch an arbitrary directory for a service cycle?
18:09 allanparsons the dir is not managed by the state.sls file
18:09 Networkn3rd joined #salt
18:10 dave_den Mr_N: also, if you run your master in debug mode, it will list which pillar files it is parsing and what the compiled data is being returned for that pillar.
18:10 allanparsons it's managed by a separate state
18:19 sandbender1512 is 'test' reserved, ie: not allowed, for usage as a name/id in an .sls ?
18:20 sandbender1512 ... by any chance?
18:20 moos3 Ahlee qeustion I just loop though and call salt.Key.delete_key('node_name') ?
18:24 gwmngilfen joined #salt
18:29 [diecast] joined #salt
18:31 ndrei joined #salt
18:33 icarus_ joined #salt
18:39 Calum joined #salt
18:39 moos3 how can I access the key management api from a python script ?
18:41 jcockhren moos3: that wouldn be awesome
18:41 Thiggy joined #salt
18:41 jcockhren would*
18:41 kzx joined #salt
18:41 bhosmer joined #salt
18:41 moos3 jcockhren yeah i know right
18:41 Thiggy Does anyone have a jinja macro that evaluates to "True" if a passed in hostname argument matches and "False" otherwise?
18:41 Thiggy *matches grains['fqdn']
18:42 moos3 I got runner working from a python script, just need to figure out the best to delete keys from a native python with out having to shell out
18:42 yomilk joined #salt
18:44 jcockhren moos3: maybe you can mirror what salt-key does
18:44 jcockhren all the salt-* scripts are just python scripts
18:45 jcockhren seems to me like something that would go under the whell catagory
18:45 jcockhren category*
18:45 jcockhren s/whell/wheel
18:45 jcockhren sheesh
18:46 fllr joined #salt
18:47 anteaya to understand reactors, I need to understand states and runners, yes?
18:51 Gordonz joined #salt
18:52 rushm0r3 joined #salt
18:53 ajw0100 joined #salt
18:54 Ahlee moos3: all salt-key does is manages the raw cert files on disk
18:54 Eugene A firm grip of nuclear physics helps too.
18:54 Eugene (yes, you should read all of that first. It's still mostly Greek to me)
18:54 Ahlee moos3: so, you'd just have to os.remove(path/to/cert) and os.remove(path/to/key)
18:55 Ahlee moos3: if you don't want to handle the race conditions and corner cases, personally i'd subprocess.Popen(salt-key)
18:57 moos3 yeah thats what I was thinking
18:58 kzx moos3: if you don't care about stdout or stderr subprocess.call() is a simpler shortcut
18:59 moos3 k
18:59 rushm0r3 joined #salt
19:01 msmith_ has anyone ever installed a pkg from source using salt states?
19:01 Mr_N dave_den: Just now noticed your note about the master -ldebug.  My prior reproduction of the bug is http://pastebin.com/XdKi9WfH here
19:02 Mr_N If it's not sufficient I can certainly redo it with the master -ldebug as well, but for now I need to be running around like a chicken with my head cut off for other work things.
19:03 scoates anyone know if there's a way to make pyobjects more verbose? "Rendering SLS hosts failed, render error: 'function' object has no attribute 'get'" … I wish it would tell me where.
19:03 bhosmer joined #salt
19:03 mgw joined #salt
19:03 sandbender1512 anyone else having issues with prereq (seemingly) ?
19:03 mgw joined #salt
19:04 fllr joined #salt
19:04 davet1 joined #salt
19:06 anteaya k, thanks
19:07 Networkn3rd joined #salt
19:08 davet joined #salt
19:09 fllr joined #salt
19:12 aranhoide left #salt
19:12 scoates potentially dumb question: is there a CHANGELOG somewhere that describes API changes?
19:16 sdlarsen_ ssm1rsklmt
19:16 rushm0r3 joined #salt
19:16 allanparsons @mssmith_ yeah
19:17 allanparsons msmith_ - i just stumbled upon this a few min ago:  https://github.com/RobSpectre/salt-states/blob/master/ffmpeg/init.sls
19:17 allanparsons but, ffmpeg takes for-fucking-ever to install.  i ended up just making my own .deb packages and then pushing out using a cmd.run: dpkg -i <src>
19:18 UtahDave joined #salt
19:21 jkolbert joined #salt
19:21 rushm0r3 joined #salt
19:22 jkolbert hello all!  I'm looking at the documentation for states (http://docs.saltstack.com/ref/states/all/salt.states.file.html) and it mentions being able to use S3 as a source for a managed file with the annotation: "(see s3.get state documentation)".  I can't find the s3.get state documentation, only the module documentation.  Is there a link I'm missing or an example I can look at somewhere?
19:24 younqcass joined #salt
19:25 thedodd joined #salt
19:25 n8n joined #salt
19:26 [diecast] joined #salt
19:27 timoguin jkolbert, I think that's wrong. Doesn't look like there is a state module for s3
19:28 jkolbert timoguin: that's a bummer, but thanks for the answer :)
19:28 dave_den there is a state module for s3 in develop
19:28 dave_den i don't see the documentation for it on docs.salstack.com though
19:28 Ryan_Lane is there any effort on doing doc sprints?
19:29 timoguin dave_den, not on the main develop branch there isn't
19:29 Ryan_Lane the docs don't keep up with the features and they are generally untouched and disorganized
19:29 Ryan_Lane I'm having people wanting to use ansible over salt simply because the docs are better
19:29 Ryan_Lane and that's painful
19:32 dave_den weird, you're right. i just had it on my local git branch which was just a few hours ago, i thought
19:33 \ask joined #salt
19:33 HeadAIX joined #salt
19:34 timoguin Ryan_Lane, I'd like to see some good interactive walk-throughs with Vagrant
19:34 Ryan_Lane timoguin: I was actually going to write a blog post on this soon
19:35 Ryan_Lane some docs on the official site would be nice, for sure
19:35 forrest Ryan_Lane, :(
19:35 timoguin ditto on the blog
19:35 Ryan_Lane the vagrant provider doesn't work with gitfs
19:35 Ryan_Lane mostly because the necessary packages are missing from the ppa
19:35 n8n joined #salt
19:36 sdlarsen_ joined #salt
19:36 forrest Ryan_Lane, can you ask your team to bring up their biggest gripes in the docs?
19:36 Ryan_Lane I can ;)
19:36 Ryan_Lane the walkthroughs are sparse
19:36 forrest Ryan_Lane, yea that's fair
19:37 Ryan_Lane there's basically no docs on how the filesystem could (or should) be organized
19:37 forrest Ryan_Lane, in regards to directory structures and such?
19:38 Ryan_Lane I still don't understand why there's roots, module dirs, returners dirs, etc.
19:38 Ryan_Lane grains dirs
19:38 forrest oh I see
19:38 forrest Yea that makes sense
19:38 Ryan_Lane I never remember what I'm supposed to name directories under the roots
19:38 forrest I was gonna say, if it was states and pillars, I thought I covered that ok in the best practices.
19:38 Ryan_Lane modules? _modules?
19:38 anteaya I'd read your blog post(s), Ryan_Lane
19:39 Ryan_Lane if I'm using masterless salt, is it modules and with a master it's _master?
19:39 forrest _modules for your modules
19:39 Ryan_Lane in which case do I need to sync them?
19:39 kedo39 joined #salt
19:39 forrest Yea these are all fair/good points
19:39 forrest I feel like that should justify a doc just on syncing modules and grains and such
19:40 Ryan_Lane it would really be nice if the docs started at a basic level, using masterless salt, and moved through more advanced topics, using actual use cases as examples
19:40 Ryan_Lane where it ended using master/minion with syndic and multi-master salt
19:41 Ryan_Lane I'm not trying to rail on the docs. they aren't *really* bad, but they could definitely use some love, which is why I was asking about sprints
19:42 forrest Ryan_Lane, I don't think anyone is gonna give you a hard time about bringing that up
19:42 forrest Ryan_Lane, The better the docs are, the better it is for everyone.
19:42 Ryan_Lane I'm more than willing to help with docs, but I need direction
19:42 forrest Ryan_Lane, there are a TON of doc issues open right now
19:42 Ryan_Lane and a specified time to work on them
19:43 Ryan_Lane well, what I'm saying is that we should have doc sprints, on a schedule, where each doc sprint focuses on a specific section of the docs
19:43 anteaya I'd show up to a doc sprint
19:43 forrest Ryan_Lane, that would be sweet.
19:44 Ryan_Lane anteaya: yeah, me too :)
19:44 Ryan_Lane I think we could likely get a decent number of folks to show up to each
19:44 anteaya we just need to know where, so we can plan dinner
19:44 forrest heh
19:44 timoguin we could get an online sprint going too
19:44 timoguin i'd show up to the internet.
19:44 Ryan_Lane yep. online sprint would be good
19:44 anteaya for an online sprint I would be distracted
19:45 anteaya meetings, pings yadda yadda
19:45 anteaya would do what I could
19:45 Ryan_Lane would be very nice to have online + in person at the same time
19:45 forrest yea there's no reason to exclude one group or another
19:45 Ryan_Lane I wonder if I could host one at lyft or wikimedia
19:45 anteaya I could handle another trip to SF
19:45 forrest Ryan_Lane, do they have those sort of 'open' spaces?
19:45 forrest if not, you could see if there are any local open spaces available
19:45 Ryan_Lane forrest: yep
19:45 forrest cool
19:46 Ryan_Lane it needs to be organized and I'd need a basic list of folks that would show up
19:46 anteaya Ryan_Lane: avoid summit and pycon and we are good
19:46 Ryan_Lane for wikimedia I'd need an employee sponsor since I'm just a contractor now :D
19:46 Ryan_Lane anteaya: heh. indeed. I need to avoid the summit at minimum
19:46 anteaya :D
19:46 forrest that reminds me, there is a pycon sprint the Monday after PyCon
19:47 forrest I wonder if the topic was decided...
19:47 forrest basepi!
19:47 anteaya can we do something at pycon?
19:47 anteaya I might actually end up there
19:47 Ryan_Lane if so, let me know what's going on and I'll attend online
19:47 basepi forrest!
19:47 forrest yea I probably won't make it, the joys of moving between states.
19:47 anteaya Ryan_Lane: you going to be in Montreal?
19:47 Ryan_Lane alas, no
19:47 baniir joined #salt
19:47 forrest basepi, is the topic for the pycon sprint decided yet? Ryan_Lane is proposing we make it all about the docs, and I think that's a pretty great idea.
19:48 forrest Not sure if you guys had ana agenda yet
19:48 basepi And yes, there's a sprint on Monday and an open space Friday night. No defined topic yet. It's just me a Joseph so we can make it what've
19:48 anteaya Ryan_Lane: :(
19:48 basepi Whatever we want
19:48 basepi (Sorry, on phone, sent early)
19:48 anteaya basepi: which Monday?
19:48 forrest anteaya, the monday right after PyCon
19:48 basepi Sorry, Monday of PyCon
19:48 forrest 14th right?
19:49 basepi First day of sprints. Yes, 14th
19:49 anteaya ah okay great
19:49 anteaya I can do the 14th
19:49 forrest yea I won't be able to make that one online in the evening, but if people are gonna be there, it would be sick to get some action in on the docs
19:51 Ryan_Lane basepi: does saltstack have any tech writers on staff?
19:52 basepi Ryan_Lane: define the term. We are all decent technical writers, but have no one dedicated to that task.
19:52 Ryan_Lane someone that organizes the writers, makes style guides, determines what to focus on, etc.
19:53 ldlework Why are there not dedicated buttons on the header for saltstack for jumping to the states and modules
19:53 ldlework crazy idea I know..
19:53 scoates any idea if that will be open to non-attendees?
19:53 basepi Ryan_Lane: mostly we've been winging it. Haven't had the bandwidth for that kind of process yet.
19:54 * Ryan_Lane nods
19:54 scoates I have not managed to secure a pycon ticket, but I live in Montreal…
19:54 forrest scoates, what the hell, I could have crashed at your place and you didn't offer??
19:54 basepi scoates: I can't say for sure, but I think the sprints are free and separate from the conference registration.
19:54 anteaya I picture you sitting on the otherside of a window, tapping on your keyboard
19:54 scoates forrest: toosoon.
19:54 anteaya scoates: ^
19:54 scoates (-;
19:55 forrest scoates, :P
19:55 scoates anteaya: fish bowl style (-:
19:55 dave_den ldlework: on the tops of the page of docs.saltstack.com, there is a link to 'all salt modules' which includes execution and state modules docs
19:55 anteaya ha ha ha
19:55 basepi Afk
19:55 Mr_N dave_den: Didja catch my pastebin from earlier?
19:55 all_the_fisch joined #salt
19:56 Mr_N (Sorry to bother, mostly wondering whether I should make an issue before I leave work today or not)
19:56 scoates conclusion after the weekend and a bit of time today: pyobjects isn't ready enough to be used, today. /-:
19:56 dave_den Mr_N: yep. that should definitely be working.
19:56 dave_den Mr_N: you may want to see what the debug log on the master says
19:56 Mr_N Fair enough.  I'll post that in the formal issue.
19:56 ldlework dave_den: yikes... sorry.
19:56 kermit joined #salt
19:56 Mr_N Thanks for taking a glance at it.
19:57 ldlework I have absolutely no idea how I didn't see that
19:57 dave_den Mr_N: no prob. it does work for me on 0.17.5
19:57 dave_den on a debian 7 minion
19:57 Mr_N Yeah; we're all Sci Linux here
19:57 Mr_N so that may be the issue?
19:57 Mr_N since this really doesn't seem like something the behavior of which should be changable by some configuration option I borked
19:58 dave_den Mr_N: it doesn't hurt to rm -rf /var/cache/salt/minion on your minion to see if that does anything. chances are it won't fix it, but might be worth a try
19:58 Mr_N I will try that as well and document the results in the issue.
19:58 dave_den stop/start your minion before/after doing that
19:58 Mr_N stop/start the minion 'fixes it'
19:59 dave_den right, but i'm just wondering if it's the caching that's borked
19:59 Mr_N Ah
19:59 Mr_N Sure.
19:59 dave_den since the salt-minion is clearly getting the new pillar
19:59 dave_den on restart
19:59 scoates fwiw, I had a similar behaviour to Mr_N and it was on Wheezy
20:00 ajw0100 joined #salt
20:00 connie joined #salt
20:01 anteaya so basepi let me know what you and Joseph decide, I think he is techat right now?
20:01 suckaplease joined #salt
20:05 dfinn joined #salt
20:05 dfinn anyone else running into this?  https://github.com/saltstack/salt/issues/11480
20:07 forrest anteaya, nah his irc is different than his github
20:07 suckaplease joined #salt
20:08 Ryan_Lane basepi, anteaya, forrest: +1 let me know what's going on with the sprint. I'd definitely like to help.
20:08 Ryan_Lane oh! there's an SF meetup tomorrow, if you can let me know soonish I can spread the word there
20:08 forrest Ryan_Lane, I won't be making the pycon one, had to fly out sunday, but I'll keep bothering basepi about it :P
20:10 anteaya and we can work on meeting up in SF
20:10 anteaya that would be fun too
20:10 * anteaya plans schedule
20:10 Ryan_Lane yep. I'll find out about hosting
20:10 zain joined #salt
20:12 anteaya awesome, thanks
20:16 kermit joined #salt
20:17 n8n joined #salt
20:18 xet7 joined #salt
20:19 JonGretar joined #salt
20:23 baniir joined #salt
20:24 fragamus joined #salt
20:25 ipmb joined #salt
20:31 alunduil joined #salt
20:31 andrej Hi, if I have a file managed and watched by salt; how do I make a script/program on the minion execute if it the file got changed, and salt put it's copy back in place?
20:32 andrej s/it//
20:32 andrej s/ it //
20:32 andrej meh
20:32 andrej tired
20:32 Mr_N I love how useful jinja errors always are:  Jinja syntax error: unknown tag 'endfor'
20:32 Mr_N /so much sarcasm/
20:32 ggoZ joined #salt
20:34 suckaplease Is there a canonical way to gather the status of salt-minion hosts to 'know' when a cluster is complete when using salt-cloud and maps of providers?
20:36 travisfischer joined #salt
20:37 jcockhren suckaplease: if the minion's in the cluster have a common identifier, then yes
20:37 n8n joined #salt
20:37 jcockhren test.ping but match off that indentifier (or combination of identifiers)
20:37 jcockhren for example a grain
20:39 suckaplease jcockhren: They do indeed, I guess you're recommending a polling implementation? I was hoping that salt-cloud would be able to provide a more polished way to do things. Okay, I can imagine doing a few things like having a callback to a monitoring server as on of the final states or as you recommend building that logic into the master. Thanks
20:40 yomilk joined #salt
20:41 tedski if a syndic has some salt-ssh hosts in its roster, is there a config option so that those can be targed on the syndic's upstream master as if they were normal minions?
20:42 bhosmer joined #salt
20:44 tedski iow, i'd like to do salt '*' test.ping on the master and have the syndic turn that into a salt-ssh '*' test.ping
20:45 basepi Ryan_Lane: by "what's going on with the sprint" do you mean sprint focus? That's the only detail we don't already know.
20:46 yomilk_ joined #salt
20:46 basepi anteaya: we'll keep you posted on the sprint focus.
20:50 jcockhren suckaplease: yeah. polling is bleh
20:50 jcockhren it'll be nice to have minion upstarts integration into the reactor system
20:51 andrej Can I execute an arbitrary command on the minion as the outcome of a state change?
20:51 jcockhren predefined events such as 'alive', highstate_started, highstate_success, highstate_failed
20:51 andrej I have a managed /etc/defaults/grub file
20:51 jcockhren it'll be up to the user to use the reactor system to respond to those things
20:51 andrej and when salt re-deploys it I'd like to run update-grub on the minion
20:51 jcockhren hmmmmm
20:52 jcockhren actually... suckaplease that may already be possible. (just not standard). but only for as a high state.
20:55 rlarkin hm
20:56 rlarkin I made some environments and a 'common.sls' for '*'.  since base: '*': comes first I thought common.sls would go first, but my match private environment goes first
20:56 rlarkin and fails due to missing deps
20:57 rlarkin is that expected?
20:57 suckaplease jcockhren: do tell, I'd like to get some ideas on how to do it and I'm currently only interested in the highstate
20:58 Gordonz_ joined #salt
20:58 tedski andrej: you should use cmd.wait
21:00 cewood joined #salt
21:10 arthabaska joined #salt
21:13 defunctzombie joined #salt
21:15 amains joined #salt
21:17 utz joined #salt
21:18 n8n joined #salt
21:19 bhosmer joined #salt
21:21 jcockhren suckaplease: try creating a module (or state-module) that fires an event
21:22 jcockhren as a state module, you can have it executed at different stages of highstate
21:22 jcockhren so... it's possible to have the last state of a highstate to be the one the fires the event
21:23 suckaplease jcockhren: thanks for the suggestion. It sounds like a reasonable approach. I'm trying to create green/blue deployment on EC2. I'm thinking about using CloudFormation in-place of salt-cloud. Does anyone have any recommendations on that?
21:24 Ahlee there a way to increase the timeout on salt-run? doesn't look like -t is being honored
21:24 Gordonz joined #salt
21:24 mgw Ahlee: which runner?
21:24 Ahlee jobs
21:25 Ahlee jobs.lookup_jid
21:25 Ahlee in particular
21:25 mgw I believe each runner has to implement timeouts in its own way
21:25 mgw Is it hanging forever, or just longer than you'd like?
21:26 Ahlee takes longer than ~10 seconds to return a value
21:26 Ahlee so i'm not getting data back
21:26 Ahlee and specifying -t 30 or -t 300 still exits in between 10 and 15 seconds
21:26 Ahlee i see only jobs.active imports __opts__['timeout']
21:27 Ahlee guess i'll see about adding that to lookup_jid and seeing how far that gets me
21:27 tyler-baker joined #salt
21:29 Gordonz_ joined #salt
21:29 mgw basepi, UtahDave: Do you know if the salt-run cli does anything at all with -t? Is it available to the runner to pass on to the LocalClient (assuming it uses one)?
21:29 Gordonz joined #salt
21:30 Ahlee mgw: https://github.com/saltstack/salt/blob/develop/salt/runners/jobs.py#L67-L72
21:30 Ahlee unless it's specified in __opts__ it ain't getting in there
21:31 basepi mgw: honestly, i have no idea.  Ahlee is right in that it would need to be in __opts__, and there's a chance it is put in __opts__ in the salt-run logic that calls the actual runner function, but i don't know
21:32 Gordonz_ joined #salt
21:33 anteaya basepi: thanks
21:33 mgw basepi: thanks
21:39 Ahlee so yeah.  it ended up being the host didn't have pymongo, so the command line must have been using the always-on-unable-to-be-disabled file system return
21:40 Ahlee but salt-api using the returner/lookup_jid was trying to go to the job cache
21:42 elfixit joined #salt
21:43 [diecast] joined #salt
21:45 m0s joined #salt
21:48 oz_akan__ joined #salt
21:49 ipmb joined #salt
21:50 BrendanGilmore joined #salt
21:50 kballou joined #salt
21:52 xmj Joy
21:52 xmj Salt is currently broken on FreeBSD due to a move of some dependency of it that wasn't reflected in salt's port.
21:55 pentabular joined #salt
21:57 Linz joined #salt
21:59 seapasulli left #salt
22:00 seapasulli joined #salt
22:03 Gifflen joined #salt
22:04 dvogt joined #salt
22:05 ajw0100 joined #salt
22:07 Gifflen joined #salt
22:07 che-arne joined #salt
22:08 travisfischer joined #salt
22:09 diegows joined #salt
22:09 yomilk joined #salt
22:10 elfixit joined #salt
22:10 YESSS joined #salt
22:11 ldlework Is there any way to have salt restart the machine and then continue provisioning after boot up?
22:15 ldlework Also, if I have multiple 'salt states' under one of the states that I make myself, do they get run in order if I've turned that on
22:15 ldlework or does that only apply to top-level states
22:17 ldlework I care more about the second question
22:18 rglen joined #salt
22:20 dvogt joined #salt
22:22 Steve__ joined #salt
22:23 Steve__ Hi All, I'm trying to get salt-call cp.push_dir to work ... salt-call cp. push works but I get Function cp.push_dir is not available when I try push_dir
22:23 Steve__ I have file_recv: True  defined
22:25 Steve__ I also have fileserver_backend: - minion defined
22:25 yomilk_ joined #salt
22:25 Steve__ any thoughts?
22:28 yomilk joined #salt
22:28 snuffeluffegus joined #salt
22:28 toastedpenguin joined #salt
22:31 ipalreadytaken joined #salt
22:31 vejdmn joined #salt
22:31 Steve__ Anyone there?
22:33 Linz joined #salt
22:34 monokrome joined #salt
22:34 StDiluted joined #salt
22:35 roolo joined #salt
22:37 googolhash joined #salt
22:40 dvogt joined #salt
22:44 Linz joined #salt
22:48 diegows joined #salt
22:50 smcquay joined #salt
22:50 joehillen joined #salt
22:51 smcquay joined #salt
22:58 kballou joined #salt
23:03 rojem joined #salt
23:05 whiteinge Steve__: looks like cp.push_dir was added since the latest release but it wasn't documented
23:08 Gifflen joined #salt
23:08 dvogt joined #salt
23:11 whiteinge just submitted a fix here: https://github.com/saltstack/salt/pull/11491
23:13 pentabular joined #salt
23:16 elfixit joined #salt
23:18 Steve__ I am using 2014.1.0 which is what the doc's are for?
23:20 dmwuw joined #salt
23:22 redondos joined #salt
23:24 gjaber joined #salt
23:25 seapasulli left #salt
23:25 gjaber how can I list directory contents inside a sls???
23:28 allanparsons gjaber, what do you mean?
23:28 ldlework joined #salt
23:28 n0arch joined #salt
23:29 gjaber allanparsons, I want to make a state that downloads a file to a directory from a http server, do an operation to it, and then delete it
23:30 allanparsons well, i'd recommend downloading the file yourself and storing it in salt
23:30 gjaber allanparsons, the thing is, I dont know before hand the name of the file
23:30 allanparsons that way you dont have to calculate a hash
23:31 gjaber allanparsons, the limitation in this case is that that it is not an option
23:32 allanparsons then just do a cmd.run
23:32 allanparsons and use wget
23:32 allanparsons or curl -O
23:32 gjaber allan parsons, its actually a package in pypi, the command is pip install --allow-external pycairo --allow-unverified pycairo --download . pycairo
23:33 gjaber the problem is that the name of the file, i do not know before the download begins
23:33 pentabular joined #salt
23:33 allanparsons so then just cmd.run that.
23:33 gjaber right now is pycairo-1.10.0.tar.bz2
23:33 allanparsons add a var at the top of your sls file
23:33 gjaber but if the version of pycairo changes, so does the name of the file
23:34 allanparsons {% set version = pillar.get('version', '2.8.7') -%}
23:34 ndrei joined #salt
23:34 allanparsons and then cmd.run.... "blah blah blah {{ version }} blah blah blah"
23:35 johtso joined #salt
23:35 pentabular joined #salt
23:35 gjaber mmm ok, ill do that
23:35 gjaber thanks
23:36 allanparsons or
23:36 allanparsons you can run a script to find out
23:37 allanparsons {% set commit_id = args.get('git_commit_id', salt['cmd.script']('salt://deploy/files/git_revision.sh', repository_url, runas=user).get('stdout', '')) %}
23:37 allanparsons i do that when i need to find the latest commit_id to make folders
23:37 gjaber my plan b is just script that whole step and make a cmd.run
23:37 allanparsons and then i do file.directory... name.... /my/path/{{ commit_id }}
23:38 gjaber yep, thats also a choice
23:38 frasergr_ joined #salt
23:39 pentabular left #salt
23:46 johtso joined #salt
23:50 allanparsons argh
23:50 quickdry21 joined #salt
23:50 allanparsons how do i pass variables between states
23:51 allanparsons that's the most annoying thing ever
23:51 allanparsons i gotta rewrite all my states and strip out the include "virtualenv" because i generate the virtualenv name after every deploy
23:52 ipalreadytaken joined #salt
23:54 tr_h joined #salt
23:56 dvogt joined #salt
23:56 StDiluted joined #salt

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