Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-05-19

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

All times shown according to UTC.

Time Nick Message
00:00 kunersdorf source for file.managed
00:00 kunersdorf foo://some_file and bar://another_fiel rahter
00:04 Gareth whiteinge: ping
00:04 murrdoc Gareth:  i hvae a fun problem with scheduler
00:04 murrdoc and using the pillars
00:04 Gareth hooray!
00:04 murrdoc i want to use the pillars
00:04 murrdoc but i dont want them to be loaded in unless the server has passed highstate
00:04 murrdoc thoughts ?
00:04 Gareth Have to use pillars?
00:05 murrdoc well i dont
00:05 murrdoc i like the ease of add / remove with pillars
00:05 Gareth You could use the schedule state module.
00:05 murrdoc what do u recommend instead
00:05 murrdoc hmm
00:05 murrdoc does that have the same structure as a pillar
00:05 murrdoc close ?
00:05 Gareth Should be almost identical.
00:06 Gareth well...no...actually, not identical really, since it's a state.
00:07 murrdoc ok i l try that
00:08 murrdoc does it add only once ?
00:08 murrdoc like unique on a key and all that
00:08 Gareth alternatively, add it in pillar but set it as disabled.  Then in the highstate enable the jobs?
00:08 murrdoc oh
00:08 murrdoc that sounds better
00:08 Gareth Yeah.  Looks for an existing job first, if it doesn't see it...adds it.
00:09 murrdoc http://docs.saltstack.com/en/2014.7/ref/modules/all/salt.modules.schedule.html#salt.modules.schedule.enable
00:09 murrdoc that one ?
00:10 keimlink joined #salt
00:11 Gareth Yeah.  Thought I had added a enable & disable in the state module too.
00:11 Gareth perhaps not.
00:12 murrdoc slacker
00:12 murrdoc is it just calling it with params ?
00:12 murrdoc and adding changes and stuff
00:13 Gareth which?
00:13 Pixionus joined #salt
00:14 murrdoc i ll use moduel.run
00:14 cromark joined #salt
00:14 murrdoc how do i setup the jobs to be disabled in the pillar
00:15 cromark_ joined #salt
00:16 CeBe joined #salt
00:18 Gareth in the pillar item, add enabled: False
00:29 loz-- joined #salt
00:30 Gareth bbl
00:31 baweaver joined #salt
00:31 primechuck joined #salt
00:45 brianfeister joined #salt
00:48 Zachary_DuBois joined #salt
00:52 kusams joined #salt
01:00 ITChap joined #salt
01:02 brianfeister joined #salt
01:08 dendazen joined #salt
01:18 brianfeister joined #salt
01:21 bfoxwell joined #salt
01:21 jeffspeff joined #salt
01:24 XenophonF well, i think i figured out why pygit2 doesn't work - i'm trying to use an ed25519 key and libssh2 doesn't support those yet
01:25 XenophonF so i'm going to re-key my master and see if that makes a difference
01:25 otter768 joined #salt
01:28 XenophonF it's confusing because there's libssh2 and libssh, and libssh supports ed25519 but libssh2 doesn't
01:28 mrbigglesworth joined #salt
01:28 XenophonF shoot - looks like libssh2 doesn't support ecdsa, either
01:28 XenophonF http://www.libssh2.org/libssh2-vs-libssh.html
01:28 XenophonF that sucks
01:32 __number5__ XenophonF: I don't remember salt required libssh or libssh2
01:32 XenophonF it's pygit2 that uses libssh2
01:33 __number5__ okay, got it now
01:33 XenophonF gitpython isn't available on freebsd any more
01:33 murrdoc i always wanted to know
01:33 murrdoc free bsd from what ?
01:33 iggy so... based on that table it sounds like someone chose libssh2 because of the license
01:33 murrdoc like who has bsd captive
01:34 XenophonF well back in the day, AT&T :)
01:34 iggy I can't imagine it winning any other way from that table
01:34 XenophonF that's what i'm thinking, iggy
01:35 iggy or maybe someone thinks EC keys are patent encumbered?
01:36 iggy Windows support maybe? (it's unclear if libssh doesn't have Windows support or if it just doesn't use the Windows encryption backend)
01:36 XenophonF i dunno
01:37 XenophonF ed25519 is djb's cipher
01:37 XenophonF his licenses are usually pretty permissive
01:37 iggy omg, if you went to saltconf, you know that
01:38 * iggy feels like it was beaten, drilled, TNTed, and then tattooed into his skull
01:38 XenophonF LOL
01:38 XenophonF djb is pretty awesome :)
01:38 XenophonF ecdsa is in openssh as well, which iirc is ISC licensed
01:38 iggy ehh, he does some good coding and thinking... bit of a loon in my experience
01:39 XenophonF well, my personal savior is a piece of 1950s-era clip art, so who am i to judge :)
01:41 beauby joined #salt
01:45 XenophonF yup, switched to an RSA key and now everything works fine
01:45 julez joined #salt
01:46 mrbigglesworth joined #salt
01:46 XenophonF hm, spoke too soon
01:47 XenophonF now i'm getting "gitfs received 0 objects for remote"
01:47 murrdoc dont use gitfs
01:47 murrdoc use debs
01:47 __number5__ ?
01:47 murrdoc opening git to every single server you need is bad
01:47 murrdoc terribad
01:47 murrdoc so bad
01:47 murrdoc did i meantion bad
01:47 * murrdoc continues with possible unfounded opinion
01:48 murrdoc might as well accept all keys
01:48 clintberry joined #salt
01:48 XenophonF i'm not following you
01:48 __number5__ murrdoc: do you have salt states/modules to pack up all states/pillar/etc. into debs?
01:48 XenophonF what's debs?
01:48 __number5__ debian pkg
01:48 murrdoc debs or equivalent packaging sytstem
01:48 XenophonF christ no
01:49 murrdoc __number5__:  no, but i use fpm to make deb from git repos
01:49 murrdoc and yes for all formulas
01:49 murrdoc its how i deploy the formulas to the master
01:49 murrdoc go.cd will make the shiny package
01:49 murrdoc scp to apt.ly
01:50 murrdoc run state on master to ensure pkg.latest
01:50 __number5__ XenophonF: you can use tarballs :P it works on all the platforms
01:50 XenophonF LOL
01:51 murrdoc that is true
01:51 XenophonF man there really is more than one way to do it
01:51 __number5__ murrdoc: but what's that pkg.latest state coming from? the debs it managed? o_O
01:51 murrdoc uh masters arent minions of themselves
01:51 murrdoc they are minions of other masters
01:51 murrdoc the top level masters get hand rolled updates
01:51 murrdoc sadly
01:51 murrdoc rest is highstate
01:52 iggy XenophonF: that's (maybe) normal
01:53 __number5__ I'm using git state to manage salt states/modules currently, already have fpm-cookery but only for code/customized upstream stuff
01:54 murrdoc nice
01:54 __number5__ murrdoc: how to you feel about go.cd? I'm not particular happy with the jenkins we have at the moment
01:55 murrdoc i likes it
01:56 writtenoff joined #salt
01:57 murrdoc buildbot is light weight
01:57 murrdoc and really good for python-y stuff
01:57 murrdoc jenkins is really tough after a limit
01:58 __number5__ why go.cd still use xml, sucks
01:59 neogenix_ joined #salt
02:08 beauby joined #salt
02:14 badon joined #salt
02:22 dendazen joined #salt
02:23 xmj joined #salt
02:23 xmj joined #salt
02:25 Singularo joined #salt
02:25 ajw0100 joined #salt
02:27 subu joined #salt
02:31 favadi joined #salt
02:32 ageorgop joined #salt
02:33 CeBe1 joined #salt
02:33 xmj left #salt
02:34 pravka joined #salt
02:35 evle joined #salt
02:36 beauby joined #salt
02:41 murrdoc joined #salt
02:41 murrdoc whats the recommeneded 'add line if it doesnt exist' saltstack module / state
02:44 schristensen joined #salt
02:49 big_area joined #salt
02:57 donmichelangelo joined #salt
02:58 favadi joined #salt
03:08 mrbigglesworth joined #salt
03:09 beauby joined #salt
03:21 I3olle joined #salt
03:24 subu joined #salt
03:25 favadi joined #salt
03:26 otter768 joined #salt
03:29 brianfeister joined #salt
03:31 viq joined #salt
03:34 julez joined #salt
03:35 jpaetzel_ joined #salt
03:36 yexingok joined #salt
03:39 pravka joined #salt
03:41 brianfeister joined #salt
03:41 BretFisher joined #salt
03:42 pravka joined #salt
03:43 pravka joined #salt
03:45 7GHAAC66H left #salt
03:46 BretFisher joined #salt
03:48 BretFisher joined #salt
03:50 clintberry joined #salt
04:00 cromark joined #salt
04:03 ajw0100 joined #salt
04:09 brianfeister joined #salt
04:17 TyrfingMjolnir joined #salt
04:25 mrbigglesworth joined #salt
04:25 Singularo joined #salt
04:32 mosen joined #salt
04:38 golodhrim|work joined #salt
04:41 fishdust joined #salt
04:42 malinoff joined #salt
04:43 rdas joined #salt
04:51 golodhrim|work joined #salt
04:52 favadi joined #salt
04:53 mrbigglesworth joined #salt
04:53 stoogenmeyer joined #salt
04:56 alektant joined #salt
05:00 spookah joined #salt
05:00 __number5__ murrdoc: file.replece have append_if_not_found option
05:04 zer0def joined #salt
05:05 ajw0100 joined #salt
05:19 loz-- joined #salt
05:22 mrbigglesworth joined #salt
05:27 otter768 joined #salt
05:28 ramteid joined #salt
05:31 julez joined #salt
05:32 MaliutaLap joined #salt
05:32 MaliutaLap left #salt
05:37 linjan joined #salt
05:49 sieve joined #salt
05:51 joeto joined #salt
05:57 colttt joined #salt
05:58 Ph-x Hi. Do you use Salt with temporary virtual machines? How do you handle removing dead minions from salt-key?
05:59 ajw0100 joined #salt
06:01 Ph-x Found "salt-run manage.down removekeys=True". That might work
06:04 solidsnack joined #salt
06:07 kusams joined #salt
06:09 harkx joined #salt
06:09 AndreasLutro joined #salt
06:18 clintber_ joined #salt
06:19 armguy joined #salt
06:23 dopesong joined #salt
06:24 dopesong_ joined #salt
06:24 flyboy joined #salt
06:26 lb1a joined #salt
06:26 illern joined #salt
06:27 pppingme joined #salt
06:35 writtenoff joined #salt
06:39 pppingme joined #salt
06:44 dRiN_ joined #salt
06:45 lb joined #salt
06:46 viq joined #salt
06:46 viq joined #salt
06:46 ljanne_ left #salt
06:47 favadi joined #salt
06:47 ljanne` joined #salt
06:48 JayFK joined #salt
06:48 clintberry joined #salt
06:51 mrbigglesworth joined #salt
06:51 clintberry joined #salt
06:54 ajw0100 joined #salt
07:01 bezaban joined #salt
07:01 Garo_ One of my master refuses to accept local salt connections: salt '*' test.ping responds with "Salt request timed out. The master is not responding. If this error persists after verifying the master is up, worker_threads may need to be increased." The master doesn't show any errors on its debug logs, but the salt command shows SaltReqTimeoutError errors. Any idea what's wrong? I'm using 2015.5.0
07:01 salt-n00b joined #salt
07:08 eseyman joined #salt
07:08 kawa2014 joined #salt
07:11 ntropy Garo_: did you verify that the master is listening on the network?  can you telnet to it?
07:12 stoogenmeyer joined #salt
07:18 Garo_ ntropy: yeah, I can telnet to the public ip address from inside the master server, to port 4505 and 4506
07:19 Garo_ the salt -l debug '*' test.ping shows these lines so I assumed that it tries to use those ipc sockets to connect to the master:
07:19 Garo_ [DEBUG   ] MasterEvent PUB socket URI: ipc:///var/run/salt/master/master_event_pub.ipc
07:19 Garo_ [DEBUG   ] MasterEvent PULL socket URI: ipc:///var/run/salt/master/master_event_pull.ipc
07:23 Auroch joined #salt
07:23 Garo_ after some strace it seems that salt tries to connect to localhost to port 4506 which salt-master doesn't for some reason answer. I'll concentrate my debugging efforts on that
07:23 ecdhe joined #salt
07:24 chiui joined #salt
07:28 otter768 joined #salt
07:30 lb1a joined #salt
07:32 mrbigglesworth joined #salt
07:38 Garo_ ntropy: thanks for pointing me to the right direction. for some reason the machine was missing ip from the loopback interface
07:47 KermitTheFragger joined #salt
07:52 vovcia hi :) im having problems with salt-cloud on ec2 - it configures instance and disks, even create keys in salt-key but there is no salt-minion on provisioned node
07:59 N-Mi joined #salt
08:01 Grokzen joined #salt
08:01 slav0nic joined #salt
08:02 impi joined #salt
08:02 impi good morning everyone
08:03 PI-Lloyd what's so good about it? :p
08:03 flyboy joined #salt
08:04 PI-Lloyd vovcia: have you tried debugging the salt-cloud command to see if it attempts to bootstrap the minion?
08:05 impi PI-Lloyd, you :)
08:05 PI-Lloyd impi: officially scared now :p
08:06 impi haha
08:07 PI-Lloyd :)
08:07 ITChap joined #salt
08:07 PI-Lloyd vovcia: also try running "salt-cloud -u" to make sure the bootstrap is up to date
08:09 dariusjs joined #salt
08:11 solidsna_ joined #salt
08:12 vovcia PI-Lloyd: salt-cloud is up to date
08:12 sieve joined #salt
08:12 vovcia trying debug now..
08:14 badon_ joined #salt
08:14 jhauser joined #salt
08:17 impi anyone here using docker instances instead of normal vm's for hosting web applications?
08:17 vovcia impi: sometimes
08:18 impi vovcia, was it difficult to make the switch? im running a magento ecommerce website that get's a lot of traffic, between 2K and 3K concurrents at a time - instances works for me on amazon vpc with autoscaling, and scale up in 7 minutes per instance but looking into running on docker
08:19 vovcia impi: no big difference, just had to rewrite deploy scripts
08:20 impi thats good to know. do you still get the same juice out of the system? or better?
08:20 vovcia same ;)
08:20 impi heh
08:21 vovcia if it would be better i would dockerize the docker into docker for triple boost :P
08:21 impi would you recommend the switch?
08:21 vovcia it depends on what do You want
08:22 impi right
08:22 vovcia if "dockerize it because everyone does it", then i dont recommend
08:22 impi cost savings and faster instance spin ups
08:22 AndreasLutro cost savings?
08:22 vovcia yeo You can save a lot on disks if plan careful
08:23 vovcia because of dockers layered filesystem
08:23 dariusjs i must be blind but i cant find the page with all the minion binaries, it had 2015.2.0 etc on it
08:23 malinoff impi, if you don't have infrastructure with init scripts, log files, if you run your app in foreground and analyze stdout, if you don't interact with shared disk resources, if you don't mind to spend a couple of weeks on network issues, then yes, you can use docker
08:23 dariusjs oh its fixed now, ignore me, i can see the index page now http://docs.saltstack.com/downloads/
08:23 vovcia PI-Lloyd: i dont know what i did but its working now :(
08:24 impi malinoff, thanks for that, i think i should, as a start, create ami's for my instances to improve their spin up time - look at the docker thing a little later on since, i have many of those things that youmentioned
08:25 malinoff impi, i suggest to avoid docker at all and write proper init.d scripts and log files
08:25 vovcia malinoff is right networking is true hell
08:25 malinoff impi, it doesn't worth it, really
08:25 vovcia or maybe checkout rocket from coreos :)
08:26 linjan joined #salt
08:27 impi thanks for the suggestions guys, you've helped me greatly
08:27 badon_ joined #salt
08:28 PI-Lloyd vovcia: glad it's working
08:28 malinoff we use docker to spin up temporary jenkins slaves (it really spins up faster than usual VM), but we will never use it in real project environments in future
08:29 malinoff it doesn't help to manage services, since instead of managing services you have to manage docker containers - so no effort here for us
08:30 malinoff in near future*
08:33 PI-Lloyd we do the same at present, use docker to run jenkins jobs and test salt states. In future we are planning to run docker containers on a mesos cluster
08:33 PI-Lloyd for production apps
08:34 AndreasLutro I've been considering docker for more advanced apps with quirky system dependencies - like a python app needs special dev packages to compile a python package for postgres support
08:35 malinoff AndreasLutro, how docker helps you there? Can't you just run apt-get/yum install?
08:36 jhauser joined #salt
08:37 AndreasLutro malinoff: I can, and that's what I do - but in the context of salt, I'd like to just have python, postgres, and webserver states, and app info in a pillar, without having to add said dev package dependencies to the python/postgres states
08:37 AndreasLutro (also the pillar file for the python app is getting quite huge)
08:38 malinoff AndreasLutro, well, if they can't be installed without dev dependencies, how you can avoid to actually adding them in one way or another?
08:38 badon joined #salt
08:39 AndreasLutro malinoff: I might run 3 python apps on one server, not all of them need those dev packages
08:40 malinoff AndreasLutro, but if at least one of these apps require them - you have no choice :) I really don't understand the point
08:42 AndreasLutro if I spin up another server that runs a python app that doesn't require those dev dependencies, I'd prefer them not to be installed
08:43 AndreasLutro my alternatives seem to be adding a bunch of boolean flags in the apps' pillar data, or providing a list of extra packages in the apps' pillars
08:43 AndreasLutro just feels a bit wrong. practically, it works fine though
08:44 malinoff AndreasLutro, if you have an other server, simply don't install these dependencies, since they're not actually dependencies
08:44 malinoff AndreasLutro, basically, your problem exists because you have one state for all possible python apps
08:44 malinoff AndreasLutro, you can't and shouldn't do that
08:45 AndreasLutro malinoff: you prefer making a state sls for each app? I've considered that as well but so far I've preferred generalised states for each service, using pillar data to configure each state
08:45 golodhrim|work joined #salt
08:45 malinoff AndreasLutro, i prefer to build deb/rpm package and specify correct dependencies
08:45 AndreasLutro ah
08:45 malinoff AndreasLutro, so the installation step is no more than pkg.installed
08:46 malinoff fpm really helps there :)
08:46 AndreasLutro I should look into that as well
08:51 morsik hi, is there some new bug in salt-minion-2015.5.0-1 ?
08:52 morsik there is missing /var/cache/salt/minion/highstate.p after running state.highstate
08:52 morsik all previous versions generated this file without problem
08:53 CeBe joined #salt
08:53 morsik and we use it in nagios check >:
08:53 danjj joined #salt
08:57 julez joined #salt
08:59 ramaseshan joined #salt
09:00 N-Mi joined #salt
09:05 PI-Lloyd this salt-cloud error is starting to annoy me now - anyone have any ideas? - https://www.refheap.com/efd64a598a1be360e3cd776d9
09:09 vovcia PI-Lloyd: what distro?
09:09 morsik i found exact commit… https://github.com/saltstack/salt/commit/93c7e3ec397d840f4702e2db8e3ae67be87ce4e6
09:09 morsik blame you :(
09:09 PI-Lloyd ubuntu 14.04 - salt 2015.5.0
09:10 vovcia PI-Lloyd: maybe You rebuilded some python libs manually and some from repos?
09:10 morsik and this… ?! https://github.com/saltstack/salt/commit/1ffabf902a6d9de2ee68ca4b546b36078ef34ac9
09:10 PI-Lloyd negative, fresh system deployed the exact same way our existing masters have been
09:11 PI-Lloyd the only difference is the other masters were upgraded from 2014.7 to 2015.5 whereas this is a fresh 2015.5 install
09:12 PI-Lloyd all the salt-cloud dependancies are the same version across masters, tried salt-cloud providers/profiles from working machines and they work fine on existing master, new one returns same result
09:12 PI-Lloyd have redployed this master 3 times now, same result everytime
09:12 vovcia this seems as python problem
09:12 vovcia maybe You could try centos 6.5 ?
09:13 PI-Lloyd our entire estate is 14.04, no problems whatso ever apart from this one master, and this is our 4th master
09:14 vovcia just to check :)
09:15 TheHelmsMan joined #salt
09:18 keimlink joined #salt
09:19 babilen morsik: That admittedly is an implementation detail you shouldn't develop against
09:24 morsik babilen: so how can I get data similar to highsate.p output? :(
09:25 supersheep joined #salt
09:25 babilen I have no idea what kind of data that might be, sorry.
09:26 cromark joined #salt
09:29 otter768 joined #salt
09:29 dariusjs morsik: I have the highstate.p  on a centos 7 box with 2015.5.0
09:29 dariusjs its an upgraded minion
09:30 morsik dariusjs: upgraded. so it's old file probably. try to remove it and run state.highstate. you won't get that file anymore
09:30 morsik see commits i posted above, generating this file was removed on 2015.5 :(
09:31 morsik our nagios check showed us what exactly failed, why… generally nagios-like formatted output from highstate output
09:31 morsik it was very useful in connection with salt scheduler (automatic highstate every 2 hours)
09:32 babilen So, what was in there?
09:32 babilen (the implied question was, apparently, not clear earlier)
09:32 morsik babilen: generally almost everything you see when running state.highstate manually
09:33 morsik just in file, so we were able to parse scheduled highstates after they were done
09:34 morsik we have 2k-3k hosts, and we really want to know when something fails somewhere :<
09:34 morsik afaik that was only way to get that info
09:36 golodhrim|work joined #salt
09:43 impi joined #salt
09:46 mrbigglesworth joined #salt
09:46 mbrgm joined #salt
09:47 mbrgm can I match by pillar in a pillar top.sls file?
09:47 babilen no
09:47 mbrgm any workaround for that? use case: I want to set a roles key, which e.g. contains the role 'gluster-peer'. now I want all gluster-peers to include a common pillar.
09:48 mbrgm so while the roles key is set on a per-host basis, the common pillar should be included for all hosts having that role
09:49 mbrgm ultimately for some sort of discovery between all gluster-peer hosts
09:49 ntropy mbrgm: i achieve the same by having roles as custom grains
09:49 dariusjs morsik: I'm actualy noticing some weird bugs with reading the pillar top.sls
09:49 dariusjs I upgraded my main saltmsaster to 2015.5.0 and my minions are not reading their environments correctly, ssupect  pillar.roots is broken or something
09:50 mbrgm ntropy: can that be done in one single state run?
09:50 mbrgm ntropy: as far as I recall grains set in a state run are only available int he next state run?
09:50 mbrgm in the*
09:50 XenophonF joined #salt
09:50 ntropy mbrgm: you might be right about that, i set the grains in minion config file
09:51 ageorgop joined #salt
09:51 matthew-parlette joined #salt
09:52 babilen ntropy: The problem with grains for roles is that they are inherently insecure and that you shouldn't make any sensible data available based on that
09:52 babilen This is the point where salt needs "static pre-pillars" that behave like grains in that they are available everywhere, but are managed on the master
09:53 mbrgm babilen: sounds right, as minions could fake a role to receive sensitive pillar data
09:53 babilen exactly
09:53 ntropy babilen: im aware of the security issue, but how can i fix it?
09:54 babilen The generally accepted way is to keep roles in pillars (and target those based on the minion id), but that doesn't allow you to target pillar by "roles" as you cannot target pillars by pillars
09:55 babilen A lot of people simply keep the minion_id <-> role mapping in a database and reference that by external pillar. That typically ties in pretty well with the rest of the infrastructure.
09:55 mbrgm is targeting pillar by pillar not possible by design or is it just not implemented yet?
09:55 dariusjs babilen: I've done it with if and else statements in the pillar but its quite a time consuming task
09:55 dexter91x joined #salt
09:55 babilen It's a chicken and egg problem
09:56 babilen You need a completely rendered pillar to know what the pillar will be, but for that you already require the top file and ... well
09:57 mbrgm babilen: regarding the ext_pillar thing: that's just a way of doing the mapping, no solution to the problem, right?
09:57 babilen Which is why I mentioned "static pre-pillars" before that don't require rendering, but are, unlike grains, managed on the master. That is, however, not available.
09:57 dexter91x I don't under stand what is maxrunning do ,
09:57 dexter91x in documentation its says " maxrunning to ensure that there are no more than N copies of a particular routine running" so what is meant by routine in here ?
09:57 mbrgm thinking of the C precompiler, it simply does multiple runs to resolve macros.
09:57 babilen mbrgm: I believe that external pillars are rendered before ... you might actually have a chance there, but I can't promise you that.
09:58 mbrgm so maybe the chicken/egg problem could be solved by re-rendering the pillar if an "unresolved" path is left
09:58 babilen We simply target pillars based on the minion id (which is the only bit of information that is guaranteed to be correct)
09:59 babilen mbrgm: There would, naturally, be ways to tackle this by "converging" on a completely rendered pillar, but I would argue that it would be much easier to introduce static pillars into salt.
10:00 babilen The idea with external pillars is, however, worth a try. I am genuinely curious if that works :-/
10:00 babilen (not an easy test though)
10:00 dariusjs i kind of miss puppets  directory environments
10:00 babilen Maybe write to salt-users and ask how other people deal with this problem?
10:01 dariusjs but that was also new in 3.x
10:01 mbrgm babilen: I'll try after lunch, using a simple yaml file
10:01 babilen mmmh, lunch :D
10:01 mbrgm :D
10:01 mbrgm what do you mean by 'static pillars'? is there an issue/ml thread on that topic?
10:03 dariusjs mbrgm: my guess is a pillar is just shoved into the system and not retrieved on the fly
10:08 babilen mbrgm: No, it is just that *I* think that a lot of people abuse grains for things that should better be managed on the master (e.g. roles) only because grains are available for use almost everywhere. My feeling is that salt could benefit from something I call "static pillars" or "pre-pillars" that are simple Python dictionaries (like pillars) that are, however, immutable. They therefore don't need any specia
10:08 babilen l treatment and could be made available ...
10:08 babilen ... everywhere too.
10:10 AndreasLutro before you know it you need pre-pre-pillars as well
10:10 favadi joined #salt
10:11 babilen I mentioned it here https://github.com/saltstack/salt/issues/21403
10:11 babilen AndreasLutro: No, you wouldn't. The rabbit hole ends there ...
10:11 dariusjs babilen: do you make use of environments or gitfs for determining environments ?
10:12 babilen gitfs
10:14 XenophonF here's my babilen-inspired top.sls, based on previous discussions of the same topic: https://github.com/irtnog/salt-states/blob/master/top.sls
10:15 XenophonF and an example pillar top.sls: https://github.com/irtnog/salt-pillar-example/blob/master/top.sls
10:17 babilen That looks great
10:17 XenophonF i have you to blame/thank
10:17 malinoff and a bit scary :)
10:17 XenophonF since my naming conventions are very regular, i can get a little more clever and assign the environment and role using globs
10:17 XenophonF and then customize each minion as necessary
10:18 XenophonF i'll need to update my example pillar
10:18 XenophonF gotta run to a meeting, so after
10:19 XenophonF i figure that i can trust the minion IDs, but i'm assuming that they can't change after doing 'salt-key -A"
10:27 mrbigglesworth joined #salt
10:34 scottpgallagher joined #salt
10:35 PI-Lloyd vovcia: figured it out, it was a problem with apache-libcloud not knowing what 'ubuntu' was. Modified the code a bit and now it works :)
10:36 Ztyx joined #salt
10:39 Berty_ joined #salt
10:39 dariusjs from XenophonF exampe I suppose his pillar.roots is quite uncomplicated
10:40 XenophonF at work i'm using ext_pillar with git
10:40 XenophonF private repo
10:40 XenophonF at home i'm just using the file system for pillar
10:41 XenophonF although i might move it to s3 just to get a feel for s3fs
10:41 vovcia PI-Lloyd: great :)
10:42 CeBe1 joined #salt
10:44 joeto1 joined #salt
10:44 dendazen joined #salt
10:45 bluenemo joined #salt
10:46 nocturn joined #salt
10:47 ndrei joined #salt
10:52 giantlock joined #salt
10:53 XenophonF ok meeting cancelled, my favorite kind of meeting
10:53 XenophonF so i've just updated my example pillar
10:53 XenophonF where i'm using minion ID globs to make saltenv/role assignments
11:01 Ztyx left #salt
11:06 favadi left #salt
11:12 golodhrim|work hi there, anyone here doing saltstack windows administration?
11:14 babilen golodhrim|work: Just assume "yes" and ask your real question
11:15 babilen (unless that is your real question)
11:15 TyrfingMjolnir joined #salt
11:23 ndrei joined #salt
11:26 julez joined #salt
11:30 otter768 joined #salt
11:34 linjan joined #salt
11:36 cromark joined #salt
11:39 cromark_ joined #salt
11:43 mbrgm babilen: ext_pillar does not seem to work
11:47 babilen pity
11:47 babilen Thanks for trying
11:48 mbrgm also, ext_pillar gets loaded after matching the top file, according to the debug output
11:49 babilen right, well .. it was worth a try
11:50 stephanbuys joined #salt
11:52 patrek joined #salt
11:52 dariusjs golodhrim|work: I do some, I have 200'ish windows minion
12:07 amcorreia joined #salt
12:07 MaliutaLap joined #salt
12:07 MaliutaLap left #salt
12:10 steven_ joined #salt
12:11 steven_ Hi I have a problem with my script who is called from my topfile.sls I need to install Adobe Reader français on Windows but ç doesn't work cause UTF-8 isn't understand by Python 2.7.9 how can I do please?
12:16 AndreasLutro steven_: where is this ç located? inside a sls file? is it in the filename?
12:17 steven_ Yes it's in the filename in my slsfile
12:18 AndreasLutro any reason you can't just use ascii characters for the filename?
12:18 patrek joined #salt
12:18 mbrgm babilen: having thought about it, static pillars would probably solve the whole problem
12:20 steven_ I'm not sure how it works do you have a link with help please?
12:20 AndreasLutro steven_: literally just rename the file to "adobe_reader_francais.sls" or something
12:21 steven_ it's not the sls filename it's the name of the program that I want to install and it's needed to be the same like Windows is installed then Adobe Reader XI (11.0.10) - Français
12:22 AndreasLutro steven_: no it's not, you can name your sls files whatever you want
12:22 yuhl_work___ joined #salt
12:23 steven_ Yes my filename is adobereader.sls but in this sls file I got this : 'Adobe Reader XI (11.0.10) - Français':  11.0.10:   full_name: 'Adobe Reader XI (11.0.10) - Français'   installer: 'http://ardownload.adobe.com/pub/adobe/reader/win/11.x/11.0.10/fr_FR/AdbeRdr11010_fr_FR.exe'   reboot: False   locale: fr_FR   install_flags: ' /msi EULA_ACCEPT=YES REMOVE_PREVIOUS=YES /qn'   uninstaller: 'msiexec.exe'   uninstall_flags: '/qn /x {AC76
12:23 steven_ Sorry I have a bad english
12:24 AndreasLutro eh, it's ok, I misread you as well
12:24 dexter91x joined #salt
12:24 AndreasLutro steven_: can you put the sls file contents and the full error message you get on https://gist.github.com/ ?
12:27 steven_ https://gist.github.com/3fad341210fb26b96d9e.git
12:29 joehh golodhrim|work: we do a moderate amount
12:29 murrdoc joined #salt
12:29 AndreasLutro steven_: not sure I'm reading this correctly, but in common.sls, try putting quotes around the name of the package (like you have in adobereader.sls)
12:29 joehh by count, probably the majority of our servers are windows
12:29 tmclaugh[work] joined #salt
12:31 mage_ if I have a file.managed with multiple - source: is there a way to adjust the - name: argument relative to the source ?
12:32 julez joined #salt
12:33 mage_ for ex I have https://dpaste.de/oDwh and I would like that - name: is aither Gemfile or Gemfile.lock regarding which file is present ..
12:36 steven_ When I try to highstate.state I got this : bss-laptop17.blues2.local: The minion function caused an exception: Traceback (most recent call last): File "c:\salt\bin\lib\site-packages\salt\minion.py", line 1020, in _thread_return return_data = func(*args, **kwargs) File "c:\salt\bin\lib\site-packages\salt\modules\state.py", line 325, in highstate whitelist=kwargs.get('whitelist') File "c:\salt\bin\lib\site-packages\salt\state.py", line
12:36 AndreasLutro steven_: put it in a gist please
12:36 mbrgm is the pillar rendered on the master or on the minion?
12:37 steven_ https://gist.github.com/jeanneret/c2ed554fdc9f9bafc6b3 sorry
12:38 emaninpa joined #salt
12:38 dariusjs hmm seems I get something similar as well for installing windows packages like steven_  is
12:39 dariusjs installing them with the salt command works
12:39 AndreasLutro steven_: replied https://gist.github.com/anonymous/3fad341210fb26b96d9e#comment-1456463
12:39 AndreasLutro I guess state IDs have to be ascii
12:40 dendazen joined #salt
12:43 steven_ I try but it doesn't change
12:43 AndreasLutro hm. you may need to do the same in adobereader.sls
12:43 AndreasLutro i.e. replace 'Adobe Reader XI (11.0.10) - Français': with just adobe_reader:
12:44 steven_ https://gist.github.com/jeanneret/9512e438ca76d390cc92
12:44 steven_ I do this
12:45 AndreasLutro and it still errors? I'm out of ideas then
12:45 steven_ Yes the same error...
12:45 steven_ Thank's for your help
12:48 AndreasLutro steven_: as a workaround, I guess you could put in the unicode literals for the ç character. I think it's \u00e7
12:49 AndreasLutro full_name: "Adobe Reader XI (11.0.10) - Fran\u00e7ais"
12:49 AndreasLutro for example
12:49 steven_ ok I will try thank you
12:50 CeBe1 joined #salt
12:51 TheHelmsMan joined #salt
12:52 JDiPierro joined #salt
12:52 julez joined #salt
12:52 steven_ always the same I don't know if I'm so bad or what...
12:53 AndreasLutro well that sucks
12:55 AndreasLutro steven_: try this maybe http://docs.saltstack.com/en/latest/ref/configuration/master.html#yaml-utf8
12:55 AndreasLutro steven_: also see https://github.com/saltstack/salt/issues/3436#issuecomment-61225863
12:55 bhosmer joined #salt
12:55 babilen mbrgm: https://github.com/saltstack/salt/issues/23910 (maybe something will happen)
12:56 dexter91x I don't under stand what is maxrunning do , in documentation its says " maxrunning to ensure that there are no more than N copies of a particular routine running" so what is meant by routine in here ?
12:56 mbrgm babilen: let's hope so. btw, is ext_pillar data rendered on the master?
12:57 babilen yeah
12:57 mbrgm ok, then I'll shove pillar -> mysql
13:00 numkem joined #salt
13:00 dariusjs steven_: AndreasLutro I can confirm now that it needs to be ascii like yo usaid
13:02 badon_ joined #salt
13:02 mpanetta joined #salt
13:04 badon__ joined #salt
13:07 dexter91x joined #salt
13:08 TheHelmsMan joined #salt
13:10 dariusjs TheHelmsMan: SBP uses salt ?
13:12 TheHelmsMan nope
13:12 TheHelmsMan but I am looking for personal use
13:14 dariusjs ok, thought you were looking for chef alternatives :)
13:14 TheHelmsMan :)
13:16 TheHelmsMan I have some friends and colleagues who are using it and are *quite* satisfied with it
13:16 dariusjs oh good to know, I've mentioned  it to Harm and Niels before
13:17 * TheHelmsMan likes that
13:18 CeBe1 joined #salt
13:22 mbrgm is it possible to include only a small #!py rendered file for defining some variables and then use those variables in a jinja template?
13:23 losh joined #salt
13:26 Grokzen joined #salt
13:26 Grok joined #salt
13:26 Grok k
13:27 murrdoc mbrgm:  what are u trying to do
13:28 yuhl_work___ left #salt
13:30 drawsmcgraw mbrgm: I think you're looking for something like a 'maps.jinja' or a 'vars.map'
13:30 murrdoc whats vars.map ?
13:30 murrdoc is that the pyobjects thing ?
13:31 drawsmcgraw The same exact thing. It's just what I named mine because I wasn't paying attention at the time and it's too late now :/
13:31 murrdoc oh
13:31 murrdoc :D
13:31 otter768 joined #salt
13:32 VR-Jack joined #salt
13:33 sasa joined #salt
13:33 sasa .
13:33 mbrgm murrdoc: I want to export a few variables from a python sls so I can use them in a jinja template (I need quite some logic for exporting the vars, so that jinja alone cannot do it)
13:33 drawsmcgraw mbrgm: This may be something like what you're looking for: http://dpaste.com/0G09SVM
13:33 mbrgm however, I'll have a look at the pyobjects renderer
13:34 sasa .
13:34 drawsmcgraw Ah, you need more logic than what I've got. the pyobjects renderer may be more of what you're looking for.
13:35 dexter91x joined #salt
13:35 JayFK joined #salt
13:35 rdutch joined #salt
13:35 ageorgop joined #salt
13:38 inad922 joined #salt
13:38 steven_ AndreasLutro: I do this but I m not sure that I done the good things and I restart the Salt service and now I got  TypeError encountered executing pkg.refresh_db: string indices must be integers, not str.
13:38 dexter91x .
13:39 steven_ It's really stange the error I got
13:41 sieve joined #salt
13:42 cberndt joined #salt
13:44 Tyrm joined #salt
13:44 mapu joined #salt
13:47 JDiPierro joined #salt
13:49 andrew_v joined #salt
13:49 bhosmer joined #salt
13:50 dyasny joined #salt
13:50 VR-Jack joined #salt
13:53 hacfi joined #salt
13:54 cromark joined #salt
13:55 dexter91x anybody there ?
13:55 perfectsine joined #salt
13:56 clintberry joined #salt
13:58 VR-Jack joined #salt
14:00 rideh joined #salt
14:03 PI-Lloyd dexter91x: maybe
14:03 hacfi joined #salt
14:04 Grokzen joined #salt
14:05 TheHelmsMan joined #salt
14:08 matthew-1arlette joined #salt
14:08 giantlock joined #salt
14:09 murrdoc :)
14:11 rm_jorge joined #salt
14:11 armguy joined #salt
14:12 bhosmer joined #salt
14:13 favadi joined #salt
14:13 fredvd joined #salt
14:14 favadi left #salt
14:14 TheHelmsMan joined #salt
14:15 JDiPierro joined #salt
14:19 Arca joined #salt
14:19 chiui joined #salt
14:20 Arca Hello everyone, basic question here, is it possible to match nodegroup in normal sls files ? (not in top.sls) and not using the name of the state ?
14:22 catpig joined #salt
14:25 big_area joined #salt
14:25 denys joined #salt
14:25 XenophonF argh freaking selinux
14:26 PI-Lloyd lol
14:26 XenophonF does anyone know of a good selinux formula for salt?
14:26 PI-Lloyd pkg.purged ? :p
14:26 XenophonF hah well
14:26 XenophonF yeah
14:27 big_area_ joined #salt
14:28 sieve joined #salt
14:29 andrew_v joined #salt
14:29 dexter91x PI-Lloyd: i have a question , are you available now ?
14:29 murrdoc u are trying to gate a state to a node group
14:29 dexter91x I would like to know what is meant by "routine" in salt , i cannot understand the meaning of it, as in salt schedule, i cannot understand this "Specify maxrunning to ensure that there are no more than N copies of a particular routine running. Use this for jobs that may be long-running and could step on each other or otherwise double execute"
14:30 bhosmer joined #salt
14:31 XenophonF dexter91x: maybe it means a function like state.highstate()
14:31 kusams joined #salt
14:32 sathya joined #salt
14:32 sathya Hi All
14:33 sathya just a quick question
14:33 sathya I have my salt master in RHEL ver 6.6
14:33 Arca @murrdoc i'm trying to do it inside the state, like, I need to do the same work, but with a different name, for two nodegroup
14:34 sathya when i do salt "*" state<tab>, the command is not getting completed or its not showing any relevent information which start with state function
14:34 sathya is there is a way, i can enable autocomplete in salt
14:34 sathya the same works find in ubuntu but not in RHEL
14:34 ndrei joined #salt
14:35 cromark joined #salt
14:36 TyrfingMjolnir joined #salt
14:36 julez joined #salt
14:40 conan_the_destro joined #salt
14:40 twork i need some clarification. the "state tree" defines paths, file contents (right?)... the basis of the configuration management system. how in principle does it differ from the file server?
14:40 yuhl_work___ joined #salt
14:41 Arca any way to use match: nodegroup in any state file?
14:41 twork i think i'm answering my own question. n/m.
14:42 XenophonF oh thank Glob there's a selinux state module
14:42 twork ...but... it does sort of look like i could (if i wanted) put all my file contents in the state tree. right?
14:43 c10b10 joined #salt
14:44 primechuck joined #salt
14:45 sathya Can some one help me with the question ?
14:46 XenophonF sathya: figure out how the salt autocomplete works in ubuntu and copy it over to rhel
14:46 XenophonF it's probably some kind of autocomplete bash function or something
14:46 XenophonF no idea personally
14:46 sathya I did try that, but it  doesnt help me much
14:46 XenophonF but that's where i'd start if i were you
14:46 XenophonF how does the autocomplete work on ubuntu?
14:47 sathya i even checked on the bash profile, every thing is fine
14:47 rdutch left #salt
14:47 XenophonF did you check the system-wide profile?
14:47 sathya i seriously have no idea how it works on ubuntu and not in RHEL
14:47 XenophonF look through the packages on debian, see what they install
14:47 XenophonF "dpkg -L package"
14:47 sathya Bash profile shows no change and its all the basic installation which was done
14:47 XenophonF look for stuff installed under /etc but not /etc/salt
14:49 cpowell joined #salt
14:51 twork find /etc ! -name \*salt
14:51 twork ack
14:54 sathya M checking
14:54 sathya will update you all
14:54 Brew joined #salt
14:56 peters-tx joined #salt
14:56 dexter91x XenophonF: so why i should use it when i have a long running function ?! the default of maxrunning is 1 so nothing will run until the long running function is complete , am i right ?!
14:57 XenophonF dexter91x: i dunno? maybe you only want one highstate job or something?
15:02 dariusjs can anyone point me how to do a PR for docs
15:03 dariusjs whoops, hit enter too early, I want to update http://docs.saltstack.com/en/latest/ref/clouds/all/salt.cloud.clouds.vsphere.html    but I only see templates https://github.com/saltstack/salt/tree/2015.5/doc/ref/clouds/all
15:03 dexter91x joined #salt
15:04 andrew_v joined #salt
15:05 XenophonF dariusjs: i think that's generated from the code for salt.cloud.clouds.vsphere
15:05 XenophonF https://github.com/saltstack/salt/blob/2015.5/salt/cloud/clouds/vsphere.py
15:06 kaptk2 joined #salt
15:06 murrdoc joined #salt
15:06 XenophonF maybe message whiteinge
15:07 XenophonF that was the last person to touch https://github.com/saltstack/salt/blob/2015.5/doc/ref/clouds/all/salt.cloud.clouds.vsphere.rst
15:08 dariusjs oh  Isee what you mean, I need to update https://github.com/saltstack/salt/blob/2015.5/salt/cloud/clouds/vsphere.py
15:08 evle1 joined #salt
15:08 XenophonF that's what i'd do, but they'll tell you if your pull request patches the wrong thing
15:09 dariusjs ok cheers
15:10 XenophonF OK, so I've upgraded a CentOS 7 minion to salt 2015.5.0, python-boto is installed, but when i try to run "salt-call boto_rds.get_endpoint ...", it returns "'boto_rds' __virtual__ returned False" - say what?
15:10 conan_the_destro joined #salt
15:11 hef__ joined #salt
15:11 XenophonF so is boto_rds only available in the develop branch?
15:11 XenophonF i'm a little lost on the codenames
15:12 hef__ are some of the state modules only used for description?
15:12 hef__ e.g. glusterfs.peered doesn't create gluster peers, it just reports if peers arn't setup
15:13 XenophonF and now i don't see boto_rds listed in the module index on docs.saltstack.com
15:14 XenophonF and a 404 on http://docs.saltstack.com/en/latest/ref/states/all/salt.states.boto_rds.html
15:14 TyrfingMjolnir joined #salt
15:14 murrdoc they moved your cheese
15:15 XenophonF dammit, don't touch my cheese!
15:15 XenophonF ok, so it's in the develop branch, that's good
15:16 * Gareth touches XenophonF's cheese
15:20 mbrgm can I require a state on another minion? if not, how can I work around that?
15:20 sasa joined #salt
15:22 cromark_ joined #salt
15:22 sathya Hi all
15:23 sathya salt-common is missing in RHEL server
15:23 sathya where as the same is installed in ubuntu
15:23 sathya EPEL repo does not own salt-common package
15:23 danjj translating an Ansible install over to Salt, does anyone know of good rosetta stone between the two?
15:25 saifi joined #salt
15:26 danjj Sathya: I don't see a salt-common in CentOS7, but it's working fine with salt-minion and salt-master
15:26 danjj is there something specific that has you looking for salt-common and which major release are you on?
15:26 sathya danjj: I have problem with autocomplete in salt
15:27 sathya my master and minion works just fine
15:28 JayFK joined #salt
15:28 danjj hmm, I'm not sure I've tried to use 'autocomplete' or even where that would come into play, so I'm probably not going to be much use
15:28 sdm24 joined #salt
15:28 baweaver joined #salt
15:29 jalbretsen joined #salt
15:29 baweaver joined #salt
15:30 debian112 joined #salt
15:32 otter768 joined #salt
15:35 baweaver joined #salt
15:35 iggy just download it from the source repo
15:37 sathya M not sure, if that fixs my problem
15:37 sathya do any one here in this list have the same type of issue what M reporting ?
15:37 sathya when i do salt "*" state<tab> the command is not getting completed
15:38 CheKoLyN joined #salt
15:38 XenophonF if i had a dime for each time i tried to test a state before running git add/commit/push and wondering why the heck my changes didn't take effect...
15:38 sathya or its listing out the options ....
15:38 b0st0n joined #salt
15:38 sathya or its listing out the options ....:q!
15:39 saguilar joined #salt
15:39 refnode joined #salt
15:49 c10b10 joined #salt
15:49 CeBe1 joined #salt
15:51 ninkotech_ joined #salt
15:57 VR-Jack joined #salt
15:58 XenophonF left #salt
15:59 smcquay joined #salt
16:01 Auroch joined #salt
16:02 writtenoff joined #salt
16:03 sathya I tired all possibilites and it seems to not working
16:04 sathya if any one have any suggestion for me to make this simple thing to work,please help me
16:05 Sacro sathya: make sure you have the completion added into your shell?
16:06 sathya Sacro can you elaborate ?
16:06 hal58th_ joined #salt
16:09 Andre-B joined #salt
16:09 TyrfingMjolnir joined #salt
16:09 murrdoc joined #salt
16:10 baweaver joined #salt
16:11 Sacro sathya: your shell doesn't magically know how to tab complete
16:12 supersheep joined #salt
16:12 JayFK joined #salt
16:13 UtahDave joined #salt
16:13 murrdoc salt comes with bash completion tho
16:14 sathya Sacro, i did few tweks on bashsrc but still it doent help me much
16:14 aparsons joined #salt
16:15 sathya if thats the case, i have the salt installed in ubuntu, where it just works fine
16:15 Norrland default bashrc for root usually don't source /etc/bash_completion.
16:15 Norrland sathya: try ". /etc/bash_completion" if you're logged in as root.
16:17 sathya Norrland: i dont know what to edit on bash_completion, can you help me with some sample, which does this autocomplete ?
16:17 Norrland sathya: you don't need to edit bash_completion at all.
16:18 Norrland sathya: when you run ". /etc/bash_completion" in bash, it will load the tab-complete functions.
16:18 sathya Norrland : then what you want me to try in ". /etc/bash_completion" ? M sorry, M kind of lost here
16:18 sathya Ha ok .. got it
16:18 sathya let me try that
16:18 is_null joined #salt
16:18 iggy the rhel packages probably don't install the bash_completion files
16:18 hasues joined #salt
16:18 kunersdorf joined #salt
16:19 sathya iggy: I did manually installed bash_completion rpm
16:19 dexter91x joined #salt
16:19 sathya Norrland : still the samr
16:19 hasues left #salt
16:19 is_null hello joehh, is it normal that the salt2015.05 packages won't install on wheezy ? full CLI session on fresh install: http://dpaste.com/3YQ8D60
16:20 saguilar joined #salt
16:20 sathya stil the same, when i do salt "*" test.<tab>, it just stay there
16:20 Norrland sathya: ok. Do you have the file '/etc/bash_completion.d/salt-common'?
16:21 sathya nope
16:21 sathya I do have /etc/bash_completion.d/ but not the salt-common
16:21 Norrland ok
16:21 murrdoc #linux101 status
16:21 sathya I think, i dont  have salt-common package installed
16:22 Norrland then the bash_completion won't know how to autocomplete salt commands.
16:22 sathya on my RHEL box
16:23 Norrland sathya: search for "salt-bash-completion" in yum
16:23 * is_null contacts joehh as instructed by the docs in http://docs.saltstack.com/en/latest/topics/installation/debian.html but the question is open to anyone
16:23 sathya Warning: No matches found for: salt-bash-completion No Matches found
16:23 sathya I think its not there in the epel repo for red hat v 6.6
16:24 saifi joined #salt
16:24 Norrland mkay.
16:26 KyleG joined #salt
16:26 KyleG joined #salt
16:26 Sacro murrdoc: oh right, does't work on zsh though
16:26 iggy is_null: maybe you need wheezy-updates?
16:29 mbrgm left #salt
16:29 UtahDave carmony: see? I'm here!   ;)
16:30 dfinn joined #salt
16:31 JayFK joined #salt
16:32 baweaver joined #salt
16:33 BretFisher joined #salt
16:34 Grokzen joined #salt
16:37 eightyeight <3 libnacl
16:37 eightyeight just saying. slick stuff.
16:38 eightyeight if only it was in the debian repos
16:38 theologian joined #salt
16:39 neogenix_ joined #salt
16:39 is_null iggy: same with wheezy-updates and wheezy-backports: http://dpaste.com/1NJXK1S
16:39 JayFK joined #salt
16:40 iggy looks like a python-zmq problem
16:42 UtahDave eightyeight: glad you approve!  :)
16:50 babilen joehh: Looks as if wheezy-backports-2015-05 is missing the python-zmq backport that is present in wheezy-backports ^^^^
16:52 amcorreia joined #salt
16:53 desposo joined #salt
16:56 mschiff what kind of regex does salt use? (e.g. in file.comment regex)
17:01 cromark joined #salt
17:01 Brew1 joined #salt
17:02 Andre-B joined #salt
17:03 forrest joined #salt
17:04 pravka joined #salt
17:05 impi joined #salt
17:06 dopesong joined #salt
17:06 ninkotech__ joined #salt
17:08 iggy I would guess python
17:09 is_null babilen: where's wheezy-backports-2015-05 hosted ?
17:12 neogenix joined #salt
17:15 spookah joined #salt
17:17 Grokzen joined #salt
17:18 Ryan_Lane is there any way to make state.single not refresh modules?
17:19 Ryan_Lane We have a pattern where custom (and upstreamed) states will call other states
17:19 Ryan_Lane the only way to do so is via state.single, but it seems it always refreshes modules
17:20 Ryan_Lane which makes things *really* slow
17:22 Ryan_Lane alternatively, it would be nice to be able to call other states directly :)
17:22 * froztbyte isn't even yet at the point of having to worry about that
17:22 froztbyte Although in some weeks...
17:25 Gareth Ryan_Lane: granted I haven't used it, only looked at it briefly, but this sounds like the problem that the orchestrate support was added to solve.
17:28 Edgan joined #salt
17:30 Ryan_Lane Gareth: heh. I'm not using a master ;)
17:30 Gareth Oh that's right.
17:30 Ryan_Lane also, I'm talking about state modules
17:31 Ryan_Lane we have syntactic sugar for some modules
17:31 Ryan_Lane for instance, boto_asg will manage its own launch config via boto_lc
17:31 Ryan_Lane rather than reimplementing all of the state logic, we just call out to boto_lc's state module
17:32 JayFK joined #salt
17:33 otter768 joined #salt
17:35 power8ce joined #salt
17:35 ajw0100 joined #salt
17:35 julez joined #salt
17:38 baweaver joined #salt
17:40 elfixit joined #salt
17:40 babilen is_null: It's hosted by saltstack
17:41 forrest murrdoc: https://github.com/saltstack-formulas/salt-formula/issues/136#issuecomment-103608939 does that actually work now? Last time I tried it doesn't and salt just hangs.
17:42 murrdoc yup
17:42 murrdoc it works
17:42 murrdoc i use it in pro
17:42 murrdoc prod
17:42 forrest do you get return data back though?
17:42 murrdoc yup
17:42 JDiPierro joined #salt
17:42 enarciso joined #salt
17:42 forrest Nice
17:42 murrdoc honestly
17:42 murrdoc it does
17:43 murrdoc gist coming
17:43 murrdoc https://gist.github.com/anonymous/1a70003021854d81cc15
17:44 iggy I've never had a problem with minions not returning and I've used salt-formula basically since the start
17:45 b0st0n joined #salt
17:45 baweaver joined #salt
17:45 iggy but I don't use packages (or have salt-formula manage my Salt versions)
17:45 b0st0n test
17:45 iggy maybe it's something to do with that
17:47 b0st0n I have encounter some strange behaviour with Salt on Windows. I am trying to install SQL using the Windows repo and I have usernames with spaces such as NT Authority\Network Service. If I run my sls through yaml.load(open()), the output is as expected but when I run salt pkg.install the output gets split
17:48 b0st0n for example NT Authority\Network Service becomes {'NT', 'Authority\Network', Services'}. How can I escape the string so that pkg.install doesn't split it?
17:49 iggy not a Windows user (thank flying spaghetti monster), but I imagine you need to put quotes around that string (spaces and \'s)
17:49 b0st0n I tried
17:50 b0st0n it is driving my crazy :P
17:53 dkrae joined #salt
17:53 evilrob joined #salt
17:54 jalaziz joined #salt
17:55 forrest UtahDave: Are you around?
17:57 vovcia b0st0n: escape " " with "\"
17:58 vovcia "NT\ Authority\\Network\ Service"
17:58 Grokzen joined #salt
17:59 denys joined #salt
17:59 sdm24 joined #salt
18:00 Setsuna666 joined #salt
18:04 Berty_ joined #salt
18:07 clintberry joined #salt
18:09 aparsons_ joined #salt
18:10 saguilar joined #salt
18:10 druonysus joined #salt
18:10 druonysus joined #salt
18:10 f4lse joined #salt
18:11 evilrob joined #salt
18:15 cromark joined #salt
18:16 f4lse we all have our favorite aspects of salt, whats your favorite? working on selling salt implemntation at my co
18:17 Ryan_Lane f4lse: AWS support ;)
18:17 evilrob joined #salt
18:19 Gareth f4lse: You might get more unbiased answers if you ask for favorite aspects that people aren't primarily responsible for implementing :)
18:19 _habnabit joined #salt
18:20 iggy not ruby
18:23 _habnabit so i'm not super familiar with salt but i'm trying to debug some salt stuff someone else wrote. it's doing a `salt-call --local -c ... --file-root ... state.highstate` and i get a 'TemplateNotFound: devenv/map.jinja' error out when i try to `{% from "devenv/map.jinja" import devenv with context %}` in another .sls file in the devenv formula
18:23 b0st0n @vovcia, I tried that and it is still being split
18:25 _habnabit the -c directory has a minion with a gitfs_remote that points to the devenv formula, but the --file-root directory doesn't contain a devenv directory. does --file-root override where jinja looks for templates...?
18:25 Grokzen joined #salt
18:26 jhauser joined #salt
18:26 Bateau_ joined #salt
18:26 iggy _habnabit: what version of salt? gitfs is only recently supported in masterless
18:27 _habnabit iggy, 2014.7.1
18:28 iggy _habnabit: can you run cp.list_master and see devenv/map.jinja?
18:28 iggy (just replace state.highstate with cp.list_master)
18:29 iggy not even really sure if that's expected to work with masterless
18:29 _habnabit iggy, it does indeed list devenv/map.jinja
18:29 _habnabit iggy, would more output help here? not sure what would be helpful
18:29 iggy sadly, me either
18:30 _habnabit fwiw it says `local:` at the top and in the 'local' section there's a `  - devenv/map.jinja`
18:30 iggy I guess dump whatever you can show us (states, commands being run, output, etc)
18:30 agend joined #salt
18:30 _habnabit sure; hang on
18:30 is_null babilen: where ? i can't find it on debian.saltstack.com
18:32 iggy is_null: it's there, just not on the main page
18:33 iggy http://debian.saltstack.com/debian/dists/wheezy-saltstack-2015-05/
18:34 _habnabit iggy, https://gist.github.com/habnabit/c7d853c9d0e834600545
18:35 iggy _habnabit: did you change devenv/map.jinja from when it worked?
18:35 _habnabit iggy, it never worked
18:36 _habnabit iggy, i'm trying to add it to reduce duplication in the formulae
18:36 JDiPierro joined #salt
18:36 iggy well, what I'm getting at is maybe it's not really not found, it's just unusable
18:37 _habnabit iggy, i suppose that could be the case, but it seems surprising that i'd get a TemplateNotFound if it was present but unusable
18:38 _habnabit iggy, would making the file empty help diagnose this problem?
18:38 iggy I honestly have no idea what the problem could be
18:39 iggy I see an error about git fileserver backend, but you say that the file actually shows up in cp.list_master
18:39 iggy maybe someone more familiar with masterless will poke their head in
18:40 _habnabit iggy, https://gist.github.com/habnabit/c7d853c9d0e834600545#file-gistfile2-txt
18:43 is_null iggy: i have wheezy-saltstack-2015-05, but not wheezy-backports-2015-05
18:43 is_null i'll take a blind shot anyway
18:43 dopesong_ joined #salt
18:43 babilen is_null: http://debian.saltstack.com/debian/dists/
18:44 babilen is_null: Ah, there is no wheezy-backports-2015-5 AFAIK, why do you think it exists?
18:44 is_null because you mentioned it xD
18:44 babilen (sorry, must have autocorrected that internally)
18:44 is_null 18:44 < babilen> joehh: Looks as if wheezy-backports-2015-05 is missing the python-zmq backport that is  present in wheezy-backports ^^^^
18:44 babilen yeah, that should have been wheezy-saltstack-2015-5
18:45 is_null did anyone manage to install 2015.05 on wheezy ?
18:45 babilen wheezy-saltstack contains python-zmq packages whereas wheezy-salststack-2015-5 doesn't.
18:45 is_null ahhhhhhhhhhhhh
18:46 babilen is_null: It's easy, just install from wheezy-saltstack
18:46 babilen the packages should be in -2015-5 too, but they aren't so you simply cannot use wheezy-saltstack-2015-5 at this point
18:46 is_null now it makes sense, thanks a heap
18:47 babilen sorry, I must have read your message mentioning wheezy-backports earlier and simply used that as "base" for the -2015-5 suite. My apologies.
18:48 is_null awesome, now i can run all tests with 2015.05, if there's no BC break then you made my tomorrow better than xmas xD
18:49 is_null else it'll be like easter, which is cool too xD
18:51 joeto joined #salt
18:51 iggy lol, easter it is
18:55 is_null really ? according to the release notes it seems like only the cmd command has potential BC breaks but so for it's all passing for me
18:56 is_null anyway, it's working now, thanks again for your support
18:58 RedundancyD joined #salt
18:59 cowpunk21 joined #salt
19:00 baweaver joined #salt
19:01 druonysus joined #salt
19:03 jeffspeff joined #salt
19:03 jeffspeff joined #salt
19:03 Norrland joined #salt
19:04 aparsons joined #salt
19:05 evilrob joined #salt
19:14 hybridpollo joined #salt
19:16 drawsmcgraw1 joined #salt
19:16 ALLmightySPIFF joined #salt
19:17 tobie_ joined #salt
19:19 tobie_ howdy
19:19 Pixionus joined #salt
19:19 Pixionus joined #salt
19:20 tobie_ I have a question, how can I run the saltutil.sync_all inside a state
19:20 murrdoc google saltstack state module.run
19:21 tobie_ ah awesome
19:21 tobie_ was not aware of that one
19:21 tobie_ thanks
19:21 dave_den joined #salt
19:22 primechuck joined #salt
19:23 murrdoc uh hmm
19:23 RabidCicada joined #salt
19:27 druonysuse joined #salt
19:27 druonysuse joined #salt
19:28 gladiatr joined #salt
19:28 RabidCicada joined #salt
19:31 c10b10 joined #salt
19:31 supersheep joined #salt
19:32 jamin joined #salt
19:33 otter768 joined #salt
19:34 c10b10 joined #salt
19:38 ajw0100 joined #salt
19:38 jhauser joined #salt
19:41 jschroeder joined #salt
19:41 jschroeder f
19:42 ckao joined #salt
19:43 jespada joined #salt
19:43 mordonez joined #salt
19:44 joeto joined #salt
19:46 viq joined #salt
19:46 cowpunk21 joined #salt
19:48 RabidCicada joined #salt
19:54 cberndt joined #salt
19:56 aparsons joined #salt
19:58 terratoma joined #salt
20:03 markm joined #salt
20:05 DammitJim joined #salt
20:05 DammitJim left #salt
20:06 druonysus joined #salt
20:06 druonysus joined #salt
20:07 sdm24 joined #salt
20:11 Berty_ joined #salt
20:12 dynamicudpate joined #salt
20:13 chiui joined #salt
20:14 jhauser joined #salt
20:16 rhodgin joined #salt
20:17 racooper joined #salt
20:20 jalaziz joined #salt
20:20 Gareth g
20:21 ALLmightySPIFF joined #salt
20:21 hybridpollo joined #salt
20:25 baweaver joined #salt
20:27 murrdoc h
20:28 cromark joined #salt
20:31 joeto joined #salt
20:33 alexanderilyin joined #salt
20:45 lietu joined #salt
20:46 sevalind p
20:59 Eureka_ joined #salt
21:04 hybridpollo joined #salt
21:04 linjan joined #salt
21:06 baweaver joined #salt
21:07 jespada joined #salt
21:09 druonysus joined #salt
21:09 druonysus joined #salt
21:10 hal58th joined #salt
21:11 hal58th_ joined #salt
21:11 UtahDave joined #salt
21:13 julez joined #salt
21:14 _habnabit from earlier: i
21:14 _habnabit er
21:14 bhosmer joined #salt
21:15 _habnabit from earlier: i'm having a problem with including across jinja files; it's detailed here: https://gist.github.com/habnabit/c7d853c9d0e834600545
21:15 _habnabit if that's insufficient, please let me know! i think that's everything that's relevant though
21:23 Berty_ joined #salt
21:24 JayFK joined #salt
21:25 murrdoc lotsa TemplateNotFound: devenv/map.jinja
21:25 _habnabit yep
21:26 murrdoc is that in file roots ?
21:26 murrdoc ls file_roots/devenv/map.jinja
21:26 murrdoc also salt-call config.get environment
21:26 _habnabit murrdoc, it's not in the directory given by --file-root
21:27 murrdoc well tahts why its failing
21:27 giantlock joined #salt
21:27 _habnabit murrdoc, what is cp.list_master showing, then...? those aren't all files in the --file-root directory
21:28 murrdoc not sure
21:29 jalaziz joined #salt
21:30 _habnabit murrdoc, so if i specify --file-root it'll break any included formula that uses jinja2 includes? that's what it looks like
21:30 theo__ joined #salt
21:30 Yogurt joined #salt
21:30 Yogurt Hey all
21:31 murrdoc _habnabit:  not quite sure
21:31 murrdoc sorry
21:31 Yogurt quick salt python (libnacl) question if anyone is willing to answer
21:31 murrdoc i dont use gitfs
21:31 _habnabit murrdoc, ah :(
21:32 Yogurt after creating a public key box I'm trying to encrypt some text and the docs show "b'some text'" which works but when I try 'some text' I get an error
21:32 Yogurt there's no type info in the docs though so I'm not sure what is valid and what isn't
21:33 JayFK joined #salt
21:34 otter768 joined #salt
21:35 druonysuse joined #salt
21:36 Yogurt so if it helps, I think b'some text' is a python way of saying 'some text' should be converted to bytes, but the thing that throws me is that printing it out on the other end generates a b in front of the string
21:36 Yogurt with the string in single quites
21:38 _habnabit Yogurt, you know about #python, right?
21:38 Yogurt I do, but I thought this might be more applicable here
21:38 Yogurt I'm guessing that's a no :p
21:39 Berty_ joined #salt
21:39 Yogurt is this channel more c/c++ focused?
21:39 forrest This channel is focused on saltstack
21:39 _habnabit this channel is for http://saltstack.com
21:39 forrest hey _habnabit!
21:39 _habnabit hi forrest
21:40 UtahDave Yogurt: libnacl is a saltstack project, I just haven't used it directly yet, myself
21:40 ingwaem joined #salt
21:40 Yogurt Ah, gotcha
21:40 UtahDave b  is a byte literal in python3 and is ignored in python2
21:40 Yogurt my bad
21:41 UtahDave Yogurt: no, you're totally fine.  That's a legit question for #salt
21:41 theologian joined #salt
21:41 UtahDave are you using python3 or 2?
21:41 Yogurt 3
21:41 forrest _habnabit: Hmm, so I'm just looking at the issue you posted, is it mostly that lack of finding the map that you are seeing as the issue? Or am I missing something.
21:41 UtahDave ok, so it looks like byte literals in python3 have to be ASCII
21:42 Yogurt right so, my code is of the form
21:42 _habnabit forrest, that's the entire issue
21:42 Yogurt var = b'some text'
21:42 forrest UtahDave: Since you're here that reminds me, do we talk about anywhere in the windows docs about escaping strings?
21:42 UtahDave so anything that's not ascii has to be escaped
21:42 Yogurt then I do a box.encrypt(var)
21:42 Yogurt all is well
21:42 Yogurt er, var2 = box.encrypt(var)
21:42 Yogurt my problem is that var2 gets sent to another machine
21:43 Yogurt decrypted
21:43 Yogurt and printed
21:43 Yogurt and it retains the b'some text' structure
21:43 Yogurt I guess this is more a python question
21:43 mrbigglesworth joined #salt
21:44 UtahDave Yogurt: Yeah, I think so.  I believe you have to encode your bytes object to a str
21:44 forrest _habnabit: Are you running a master and minion? Or is this totally a masterless setup? I'm just not sure because of https://gist.github.com/habnabit/c7d853c9d0e834600545#file-gistfile2-txt-L16 where you do a --local call (which should only use the minion) but then use cp.list_master which queries the master server. I'm not sure how it operates in this context of --local.
21:44 iggy Yogurt: try writing that string out to a file and see if it still has the b... I Think that's just how python normally represents byte strings
21:45 UtahDave forrest: I think I remember a bit of documentation about that, but it's sparse
21:45 _habnabit forrest, i think it's masterless
21:45 _habnabit forrest, i don't know these terms very well though
21:45 forrest UtahDave: Okay cool, might be good to bulk that up (I can create an issue or whatever, but with the rework coming I wasn't sure).
21:46 UtahDave forrest: Yeah, that would be awesome if you'd open an issue.  good reminder
21:46 _habnabit forrest, i mean, i'm just running salt-call --local on the target box
21:46 Yogurt good idea iggy, write failed when trying to write a string so yes I think python might just be interperating it
21:46 forrest _habnabit: Is it a production system though?
21:46 tomh- joined #salt
21:46 Yogurt in the print that is
21:46 forrest UtahDave: Will do.
21:46 _habnabit forrest, no
21:46 UtahDave thanks, bro
21:47 forrest _habnabit: okay cool, can you list the packages on the system and see what salt content is installed? The salt-call command doesn't list if master and minion are installed
21:47 forrest UtahDave: WHAT HAVE YOU DONE WITH UTAHDAVE! Or maybe I haven't seen you in person so long I don't remember your use of bro ;_
21:47 forrest *;)
21:47 Yogurt got it. Thanks UtahDave and iggy :)
21:48 _habnabit forrest, how do i do that
21:48 UtahDave lol, I use that sometimes...   :)
21:48 theologian joined #salt
21:48 forrest _habnabit: Just using dpkg, it's dpkg --get-selections | grep salt
21:49 forrest I think
21:49 _habnabit forrest, okay
21:49 forrest assuming you guys are still running ubuntu
21:49 _habnabit this is ubuntu yes
21:49 _habnabit the only thing in that list is 'salt-common'
21:50 forrest Hmmm..
21:50 iggy ps -ef | grep salt
21:50 forrest UtahDave: Is salt-common used to supply salt-call? Do you remember?
21:50 forrest iggy: Yeah that works too, good idea.
21:50 _habnabit iggy, nothing
21:50 forrest iggy: Did you look at that gist from _habnabit?
21:51 joeto joined #salt
21:51 iggy earlier, yea
21:51 UtahDave salt-common is where almost ALL salt code exists.  the salt-master and salt-minion packages pretty much just install the appropriate scripts and stuff to enable the salt-master or salt-minion
21:51 iggy I was the one that got him to do the cp.list_master
21:51 iggy then I gave up
21:51 forrest ahh okay
21:51 iggy because... masterless
21:51 iggy need to call that lyft guy
21:52 forrest UtahDave: Gotcha, so in that event, would calls to salt-call --local fail without the minion itself?
21:52 UtahDave yes
21:52 forrest iggy: We run masterless here as well, but the salt-minion package still gets installed.
21:52 forrest UtahDave: Cool thanks
21:52 forrest _habnabit: want to try installing salt-minion see if that provides anything else?
21:52 UtahDave yeah, install the salt-minion, but just don't enable the salt-minion startup script if you're going to run masterless
21:52 forrest right
21:52 MaliutaLap joined #salt
21:53 MaliutaLap left #salt
21:53 conan_the_destro joined #salt
21:55 debian112 left #salt
21:56 premera joined #salt
21:58 _habnabit sorry; back; got a call from comcast
21:59 tmclaugh[work]_ joined #salt
22:00 _habnabit well that's weird
22:00 tmclaugh[work] joined #salt
22:00 _habnabit read-only filesystem preventing me from doing any apt stuff
22:00 forrest _habnabit: Is this on a VM?
22:00 _habnabit in a docker container
22:01 forrest _habnabit: Or sudo is screwed up for some reason :\
22:01 _habnabit welp; made a new container and it's rw again
22:01 tmclaugh[work] joined #salt
22:02 _habnabit no idea what happened, haha
22:02 _habnabit ok let's see
22:04 druonysuse joined #salt
22:04 druonysuse joined #salt
22:04 _habnabit and the container went readonly on me again 9_6
22:09 tmclaugh[work] joined #salt
22:12 tmclaugh[work]_ joined #salt
22:14 _habnabit forrest, iggy https://gist.github.com/habnabit/c7d853c9d0e834600545
22:14 _habnabit that's after i installed salt-minion
22:15 iggy I'd call whoever wrote this and tell them they suck
22:15 _habnabit haha
22:15 forrest Did I write this one _habnabit? ;)
22:16 _habnabit well it was working until i added the map.jinja
22:16 forrest Either way, seems like some directory stuff might be messed up. Can you see what the file roots look like in: ${PROJ}/fr/salt/conf?
22:16 _habnabit but removing it sucks too
22:16 iggy working > broken
22:16 forrest I'd say look at the file roots in that conf, then see what the file root is, and ensure that the map.jinja is there.
22:16 iggy just saying
22:16 tmclaugh[work]_ joined #salt
22:16 _habnabit forrest, the salt-call has a --file-root parameter
22:17 forrest I know, I'm just curious what the conf file says, I'd also check inside of ${PROJ}/fr/salt/conf/top, I assume there is a devenv dir there?
22:18 cromark joined #salt
22:18 nikogonzo in jinja, is there a way to reference the file itself as it exists on the saltmaster? i.e. I wanted to put something like "this file is saltstack managed, see formulas/services/foo/files/this file" as {{ pillar['file_header'] }} {{ some_jinjafu_self_reference }} in the jinja template
22:18 forrest I'm just trying to think of anything, are perms correct on the map.jinja? Or anything like that
22:18 SEJeff left #salt
22:18 iggy nikogonzo: yes {{ source }} maybe
22:19 nikogonzo <3 thx will try
22:19 _habnabit forrest, re: permissions, in /var/cache/salt it looks like it's a fresh git checkout, and perms are set fine
22:19 _habnabit forrest, there is no fr/salt/conf/top; there's only a 'minion' in the conf directory
22:20 jalaziz joined #salt
22:20 _habnabit forrest, there's no devenv in the directory passed as --file-root though
22:21 _habnabit but it's definitely in /var/cache/salt
22:21 forrest did you already omit the --file-root stuff?
22:21 forrest I'm honestly not sure how it interacts when you use a defined file root value plus gits
22:21 forrest I'm honestly not sure how it interacts when you use a defined file root value plus gitfs
22:22 iggy /var/cache/salt shouldn't be a git tree either (don't know if you meant that exactly how you wrote it or not)
22:22 _habnabit without --file-root it says there's no top file found
22:22 _habnabit iggy, oh, yeah, that's not what i meant; sorry
22:23 forrest hmm
22:23 _habnabit -rw------- 1 root root 119 May 19 22:08 /var/cache/salt/minion/files/base/devenv/map.jinja
22:23 _habnabit this is running as root, though, and i can cat the file
22:23 forrest okay, is there a top file in the repo you are checking out?
22:23 forrest inside of /var/cache/salt/minion/files/base ?
22:24 _habnabit the --file-root is set as "${PROJ}/fr/salt/top", and there's a top.sls in that directory
22:24 _habnabit ${PROJ} is the root of the git repo
22:24 ageorgop joined #salt
22:24 _habnabit the top.sls there just has a base/* with some formulae
22:24 forrest okay but there's no other files in ${PROJ}/fr/salt/top. Iggy do you remember how gitfs is handled when it comes to secondary file roots?
22:25 forrest also _habnabit, can you try copying that top.sls to /var/cache/salt/minion/files/base if there is no top there?
22:25 forrest might get overwritten, but worth a shot
22:25 mrbigglesworth joined #salt
22:25 _habnabit there's another formula in a subdirectory of the 'top' directory
22:25 _habnabit but that's it, and it's not the devenv formula
22:25 mrbigglesworth joined #salt
22:26 tmclaugh[work] joined #salt
22:26 forrest that's what's making me wonder if there is a conflict of some kind between environments? I can't remember exactly how gitfs is handled for multiple file roots
22:26 forrest _habnabit: Maybe just copy all the stuff that isn't part of the repo into /var/cache/salt/minion/files/base, or try modifying the conf to both have git fs, and that file roots directory?
22:26 forrest so then it will check both perhaps?
22:26 _habnabit oh, interesting; /var/cache/salt/minion/files/base/top.sls is the same top.sls file
22:27 _habnabit not the same inode, but the same content
22:27 forrest huh
22:28 tmclaugh[work] joined #salt
22:28 _habnabit forrest, https://gist.github.com/habnabit/c7d853c9d0e834600545#file-gistfile3-txt
22:28 nikogonzo iggy: everything worked, thanks again
22:28 forrest interesting
22:28 forrest and that's WITH the top file existing under that base dir?
22:29 forrest can you see if it still exists and that it didn't do a clean git clone?
22:29 _habnabit looks like it still exists
22:29 _habnabit forrest, https://gist.github.com/habnabit/c7d853c9d0e834600545#file-gistfile4-txt
22:31 forrest _habnabit: Does your conf have values set like this: http://docs.saltstack.com/en/latest/topics/tutorials/gitfs.html#using-gitfs-alongside-other-backends
22:31 iggy fwiw, in one of ryan lane's videos at saltconf, he made a pretty compelling argument that DRY is sometimes more harm than good
22:31 forrest to use both local and gitfs?
22:32 stanchan joined #salt
22:32 _habnabit forrest, the fr/salt/conf/minion has `fileserver_backend: [roots, git]`
22:32 forrest should be good then for local content, what about the file roots directive?
22:33 forrest If it doesn't exist and have your ${PROJ}... stuff, can you add it and see what happens then?
22:33 forrest and restart the salt-minion after doing that if the service is running
22:33 _habnabit iggy, i agree, but at what point do you decide that you have too much duplication across git repositories?
22:34 iggy it's directly proportional to how well you know salt and how many of it's less used features you're using
22:34 _habnabit haha, okay
22:34 iggy or... how long have you been working on this problem?
22:35 _habnabit just a few hours today
22:35 iggy that's usually my test
22:35 _habnabit forrest, i'm not sure where one sets the file roots
22:35 forrest it's in the minion conf file.
22:35 _habnabit forrest, it's just another top-level key in minion?
22:35 forrest ${PROJ}/fr/salt/conf
22:35 Ryan_Lane user.present is kind of broken in 2015.5 and develop (which breaks shadow support, which is kind of bad for people relying on it)
22:35 _habnabit okay
22:36 Ryan_Lane https://github.com/saltstack/salt/issues/23943
22:36 forrest _habnabit: http://docs.saltstack.com/en/latest/ref/configuration/minion.html#file-roots
22:36 iggy Ryan_Lane would probably know about the intricacies of how file_roots and gitfs interact in a masterless setup
22:37 Ryan_Lane I actually don't :)
22:37 forrest why was run_as removed in 2015 :\
22:37 _habnabit also by 'a few hours today' i mean 'only today, and only for a few hours' and not 'all last week but only a few hours today'
22:37 Ryan_Lane my code is always deployed
22:37 UtahDave you specify file_roots in your minion config
22:37 tmclaugh[work]_ joined #salt
22:37 rap424 joined #salt
22:38 iggy Ryan_Lane: oh, I thought you said you were happy about gitfs+masterless support at saltconf
22:38 _habnabit forrest, what should it be set to, though?
22:38 baweaver joined #salt
22:38 _habnabit forrest, presumably not a path under /var/cache/salt
22:38 iggy forrest: because it's been throwing warnings for like 3 versions
22:38 Ryan_Lane iggy: well, I was happy we were getting proper external filesystem support
22:38 Ryan_Lane I wanted s3 support
22:39 forrest _habnabit: no use whatever the path is for that local content you had that we copied.
22:39 iggy oh, should have known ;)
22:39 ageorgop joined #salt
22:39 forrest iggy: Yeah I know, but still..
22:39 forrest now I have to modify a bunch of shit, and it broke some provisioning
22:39 tmclaugh[work] joined #salt
22:39 iggy you going to ignore the warning next time?
22:40 iggy ;)
22:41 forrest iggy: I'm a few releases behind and my dev environments are busted, so I honestly give 0 shits. I don't remember the warning popping as part of each run on previous releases.
22:41 tmclaugh[work] joined #salt
22:42 forrest or at least not in a way that was useful on a per run basis
22:42 iggy *shrug*
22:42 _habnabit forrest, https://gist.github.com/habnabit/c7d853c9d0e834600545#file-gistfile5-txt
22:42 iggy be glad you weren't the one to find the cmd.run fuckage during testing
22:43 forrest indeed.
22:43 forrest _habnabit: Same error when you had that though?
22:43 _habnabit forrest, yeah 1sec
22:44 supersheep joined #salt
22:44 tmclaugh[work] joined #salt
22:46 _habnabit forrest, https://gist.github.com/habnabit/c7d853c9d0e834600545
22:46 tmclaugh[work] joined #salt
22:47 forrest I'm not sure at this point _habnabit, it sees it is in the cache. I have no idea why it isn't picking it up
22:47 _habnabit haha welp
22:47 bfoxwell joined #salt
22:47 tmclaugh[work] joined #salt
22:47 _habnabit would it be worth trying if salt 2015 doesn't fail?
22:48 mrbigglesworth joined #salt
22:48 forrest Ehhh maybe? I'd say try local content only first, so do the mounting or whatever it is that docker has support for (or a straight up git checkout and moved content)
22:48 forrest to see if maybe that's the error, but it's actually failing somewhere else in the file
22:48 forrest then you can at least confirm if it is a local issue, or something with how gitfs is handling the setup
22:49 tmclaugh[work] joined #salt
22:49 forrest I still haven't gotten docker provisioning inside the container working with salt correctly, but I'm also using a secondary layer because lol macos
22:49 _habnabit life is suffering
22:49 supersheep joined #salt
22:49 forrest lol
22:49 forrest left #salt
22:49 _habnabit is there any other way of sharing values between formulae?
22:49 forrest joined #salt
22:50 _habnabit oh ok bye forrest
22:50 _habnabit hi again
22:50 forrest no I'm back
22:50 forrest closed the wrong tab
22:50 _habnabit is there any other way of sharing values between formulae?
22:50 forrest _habnabit: pillar data is an option
22:50 _habnabit a formula can set pillar data?
22:50 forrest _habnabit: oh you need to set pillar data from the formula?
22:51 forrest I thought you just wanted a single file you could use between projects for variables, sorry
22:51 _habnabit yeah that's it, haha
22:51 _habnabit i guess you missed the start of this
22:51 tmclaugh[work] joined #salt
22:51 _habnabit specifically, if the data isn't already in the pillar, populate it with a default value derived from other data in the pillar
22:52 forrest yeah map would work the best. You could probably have some kind of gross pillar that uses defaults, but I haven't done that. I usually set things to look up in the pillar, and if it's not there, it has a default value, but that doesn't pull from other sources, it's just static.
22:53 murrdoc yeah and dont map.jinja your default
22:53 murrdoc actually make it a yaml
22:53 murrdoc for sanity
22:53 murrdoc imho
22:54 tmclaugh[work]_ joined #salt
22:54 _habnabit hm
22:55 KyleG joined #salt
22:55 KyleG joined #salt
22:55 forrest Has anyone seen any errors with pkg.installed on 2015? I have the package and the version I want to install, it says it can't find it, yum says 'no idiot, it is right there.'
22:58 tmclaugh[work] joined #salt
23:00 baweaver joined #salt
23:01 ajw0100 joined #salt
23:01 forrest who ruby maintainers, why did you get rid of the rubygems package.
23:02 murrdoc cos fuck sane packaging
23:02 murrdoc i mean
23:02 murrdoc no comment
23:04 forrest I wouldn't mind it
23:04 forrest had I known
23:04 forrest the first time I heard about it was someone in here
23:04 forrest and I can't find anything on google about it
23:11 mosen joined #salt
23:14 mapu joined #salt
23:18 lnr joined #salt
23:22 Tyrm joined #salt
23:28 murrdoc Gareth:  u around
23:29 mosen ey murrdoc
23:29 murrdoc sup mosen
23:29 soren joined #salt
23:30 murrdoc u know how module.run works
23:30 mosen gonna write a proxytype see how that goes
23:30 murrdoc i have an interesting problem where the module i am running also expects a param called name
23:31 Ryan_Lane murrdoc: that's supported
23:31 Ryan_Lane mname
23:31 mosen so you have a clash between the built in "name" and your "name" param?
23:31 mosen RL to the rescue
23:31 murrdoc hmm
23:31 Ryan_Lane "Due to how the state system works, if a module function accepts an argument called, name, then m_name must be used to specify that argument, to avoid a collision with the name argument."
23:31 murrdoc turns out i didnt rtfm
23:32 murrdoc thanks Ryan_Lane
23:32 Ryan_Lane ;)
23:32 Ryan_Lane yw
23:32 mosen saltstack should be paying RL at this point
23:32 Ryan_Lane hahaha
23:32 baweaver Pretty sure he likes Lyft more at the moment
23:33 mosen yeah im sure
23:33 baweaver though they do fly him out to fancy conferences on occasion
23:33 baweaver Speaking of which, confession time, I took an Uber ride to Lyft HQ once.
23:33 mosen ha ha
23:33 Ryan_Lane tsk tsk
23:33 baweaver Even the driver commented on the irony there.
23:34 ALLmightySPIFF joined #salt
23:35 otter768 joined #salt
23:35 forrest no uber/lyft for me still Ryan_Lane ;)
23:36 forrest Apps still require too many perms and I'm too lazy to go in and manually modify them on my rooted phone
23:37 hybridpollo joined #salt
23:38 Ryan_Lane basically all the permissions are required for the app to function properly
23:39 Ryan_Lane lyft's app is used for both passengers and drivers, so some are only needed for the driver
23:39 lnr joined #salt
23:40 lnr joined #salt
23:40 lnr joined #salt
23:40 lnr joined #salt
23:40 iggy it's got terrible GPS location
23:41 lnr joined #salt
23:41 lnr joined #salt
23:41 lnr joined #salt
23:41 lnr joined #salt
23:42 lnr joined #salt
23:42 lnr joined #salt
23:42 forrest Yeah I know, we've discussed it before, both apps require too much though. No way am I giving out my identity, calendar, contacts, SMS, phone, photos/media/files, camera, and device ID and call information
23:42 lnr joined #salt
23:43 forrest identity, calendar, contacts, and device ID are the ones that are totally unacceptable
23:43 lnr joined #salt
23:43 lnr joined #salt
23:43 forrest sms/phone I can kind of understand so the driver can text/call you (still not really interested in giving my number to some random driver), photos/media/files is meh as well, and camera I guess I understand for a profile or something
23:43 lnr joined #salt
23:44 forrest but all of those? No way it will ever happen
23:44 lnr joined #salt
23:44 lnr joined #salt
23:44 lnr joined #salt
23:45 lnr joined #salt
23:45 lnr joined #salt
23:45 iggy I've got a razr with your name on it brah
23:45 forrest sigh, does anyone remember one of the admins so we can get rid of lnr?
23:46 forrest iggy: My phone is fine, and I use plenty of apps
23:46 forrest Just no way I'm using some closed source app that requires all of that
23:46 forrest google's stuff is bad enough
23:50 lnr joined #salt
23:50 lnr joined #salt
23:50 druonysus joined #salt
23:50 druonysus joined #salt
23:50 lnr joined #salt
23:52 dendazen joined #salt
23:52 dendazen left #salt
23:52 f4lse_ joined #salt
23:53 lnr joined #salt
23:54 lnr joined #salt
23:55 Ryan_Lane :D
23:55 lnr joined #salt
23:57 forrest bleh no freenode staff around either
23:57 lnr joined #salt
23:58 lnr joined #salt
23:58 lnr joined #salt
23:59 lnr joined #salt
23:59 cromark joined #salt

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