Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-08-20

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

All times shown according to UTC.

Time Nick Message
00:00 timbermaniac "The include statement will include sls formulas from the same environment that the including sls formula is in" <-- So if everything's in "base" env, is it proper to use the relative path to the file root?
00:00 kelseelynn1 joined #salt
00:01 bmatt yes
00:01 timbermaniac mmk, so that part's right
00:01 bmatt so /srv/salt/a/b/c/d.sls -> include:\n  - a.b.c.d
00:01 timbermaniac right
00:02 homelinen joined #salt
00:02 skyler Is it preferred to have the configuration file be in a separate state than init.sls, or both in init.sls?
00:02 kelseelynn1 left #salt
00:02 forrest skyler, I always put confs in a separate state
00:02 forrest skyler, then if someone just wants to install the package and have the standard config, they can
00:02 homelinen joined #salt
00:03 skyler forrest, that makes sense.
00:05 bmatt we put it right there in init
00:05 bmatt and clobber configs based on env
00:05 forrest bmatt, :O
00:05 forrest for shame
00:06 bmatt pfft
00:06 bmatt it works great
00:06 forrest bmatt, I'm joking, it's fine, just not great for a formula since people might not have envs set up
00:06 bmatt oh right, yeah, not in a formula
00:06 mechanicalduck joined #salt
00:06 skullone anyone using salt on centos7 or rhel7?
00:06 timbermaniac bmatt: in 'common/pip/map.jinja', the last line is "}, merge=salt['pillar.get']('pip:lookup')) %}". Honestly, I don't understand what this means, and haven't found the right documentation for how 'merge' works. Are you familiar with this?
00:06 forrest skullone, a few people are, haven't tried it out yet myself.
00:06 bmatt timbermaniac: read up on filter_by
00:06 skullone running into dependency hell with libpgm on el7
00:07 __number5__ skyler: the merge parameter is to provide a lookup dict to be merge with the lookup_dict http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.grains.html#salt.modules.grains.filter_by
00:07 bmatt timbermaniac: but that's pretty advanced, if you're just starting out with salt
00:07 homelinen joined #salt
00:07 forrest timbermaniac, just think of it like this, you have pip:\n  lookup:  in your pillar right?
00:07 timbermaniac bmatt: been using salt 0.17 for some time, and looking to conform to best practices and go by salt-formulas in github
00:07 __number5__ skyler: so you can set apache:lookup dict in your pillar to override the options in your sls
00:08 bmatt timbermaniac: oh, well, righty then. map.jinja defines a big dict and merges it with some pillar data
00:08 bmatt by calling filter_by
00:08 homelinen joined #salt
00:08 forrest timbermaniac, all this does is stuff everything from the map in there, if a value doesn't already exist. Imagine it like a zipper where it's checking whether the next tooth exists or not. If it doesn't, then fill that in from the map and continue zipping up.
00:09 timbermaniac so whether or not there is pillar data for pip: lookup, it shouldn't be problematic. Its purpose is to add additional values to the dict?
00:09 skyler __number5__: Thanks, I think I get it.
00:09 forrest timbermaniac, it won't be problematic at all
00:09 forrest timbermaniac, even if they conflict, the pillar will win
00:10 forrest thus the idea of it simply filling in the 'missing' spaces.
00:10 timbermaniac I don't have any pillar data, yet. For example, I used the vim salt-formula and left this portion in the map.jinja, and it hasn't been a problem. Just trying to pinpoint where my error's coming from
00:10 timbermaniac forrest: thanks, makes sense
00:10 forrest timbermaniac, cool
00:13 dalexand_ joined #salt
00:18 skullone forrare you a redhat or debian person?
00:18 skullone :p
00:18 skullone oh, he left
00:18 skullone no wonder nick-complete didnt work
00:18 aquinas_ joined #salt
00:18 KyleG omg
00:18 KyleG I just tab completed your name
00:18 KyleG you have blown my mind
00:19 KyleG skullone: thanks
00:19 KyleG lol
00:19 skullone :o
00:19 cpowell joined #salt
00:19 KyleG http://ss.digitalflydesigns.com/Screen%20Shot%202014-08-19%20at%205.19.35%20PM.png
00:19 skullone try zsh someday, if tab completion lubricates your sprockets
00:19 KyleG dem tab
00:19 skyler When installing and starting a service, should the service section require the pkg? I see a lot of formulas that don't do this, so I am wondering if it is implicit.
00:20 skullone i just found 'salt-run manage.down'
00:20 skullone this is nice
00:20 manfred KyleG: http://d3819ii77zvwic.cloudfront.net/wp-content/uploads/2014/07/ik49VlPshlPIz.gif
00:21 manfred skullone: just don't use oh-my-zsh... it is crap, prezto is better... but writing them all yourself is the best
00:21 timbermaniac bmatt: If you have a minute, can you look my config over? I feel like it may be something stupid: http://pastebin.com/tL1QhYcs
00:21 dalexander joined #salt
00:21 KyleG exactly
00:21 bmatt timbermaniac: syntax error on line 4 :)
00:21 bmatt and 21
00:22 manfred skullone: i wrote at least half of the zsh tab completion stuff for systemd tools
00:22 bmatt and 41
00:22 timbermaniac bmatt: strange, none of the files actually look like that. it happened when it was pasted somehow
00:23 bmatt you sure? ;)
00:23 aparsons joined #salt
00:23 manfred http://cgit.freedesktop.org/systemd/systemd/log/?qt=author&amp;q=gtmanfred
00:23 timbermaniac bmatt: double-checking, but quite positive
00:25 timbermaniac bmatt: yeah, they're all actually double-spaced. not sure what happened on pastebin
00:25 timbermaniac could've been the syntax highlighting for yaml, since there are no "--" breaks in the pages
00:27 bmatt hm. I don't see anything obvious then, sorry
00:27 timbermaniac bmatt: thanks much for looking at it, at the least
00:28 bmatt while I always take up space, I can't promise to always be useful! :)
00:28 manfred timbermaniac: what is going on?
00:28 manfred oh
00:28 manfred we talked about this way earlier today i feel
00:28 timbermaniac manfred: last night I posted about an error, "No matching sls found for 'common.pip_packages.trac' in env 'base'"
00:29 manfred and i had no idea, cause i hadn't tested it on 2014.1.7
00:29 timbermaniac just trying to figure it out
00:29 panchisco joined #salt
00:29 mosen no luck let timber
00:30 timbermaniac not yet
00:30 manfred unfortunately my salt master is poluted by a bunch of things that I don't want to delete cause I am trying to make openstack novaclient extensions easier so I can't test your thing right now... i remember it working, so I don't remember right now
00:30 manfred sorry :/
00:30 timbermaniac Well, I'm new with Salt when it comes to using the pip state and map.jinja. So it's my first attempt, configs at http://pastebin.com/tL1QhYcs
00:31 manfred the big thing that I noticed in there was having more than one directory all the way down
00:31 timbermaniac And nevermind the improper spacing for some of the imports, they're actually correct
00:31 timbermaniac manfred: whatcha mean?
00:31 manfred i can't remember if that worked in 2014.1... i know it works in 2014.7 and develop... but i can't test it right now, but I rememebr from this morning that everything else in your environment looked fine
00:32 manfred common/pip_packages/trac/init.sls
00:32 manfred you have multiple directories
00:32 manfred if you did /srv/salt/common/init.sls
00:32 manfred and /srv/salt/pip_packages/init.sls
00:32 manfred i don't remember if /srv/salt/common/pip_packages/init.sls would have worked
00:32 jchen joined #salt
00:32 manfred i thought it did, but I haven't had a chance to check since we talked thi morning
00:33 manfred cause I have been working on the novaclient stuff
00:33 timbermaniac you mean to say that init.sls being multiple directories down from root may be a problem?
00:33 manfred i don't think it is, but i can't rememebr
00:34 manfred that was what I wanted to check this morning, and just never got around to cause I was working on other stuff for customers, and then got sucked into the novaclient crap
00:34 rome joined #salt
00:34 bmatt no, I've got deeply-nested init.sls files without problem
00:34 manfred cool
00:35 manfred that was the only thing that I was trying to remember from my thoughts this morning
00:35 panchisco joined #salt
00:35 manfred I do know that there were a bunch of bugs in 2014.1.5... but this isn't a bug that i rememebr... iirc you being on 2014.1.7 /shrug
00:35 timbermaniac on a tangent, how do you know when it's a good time to update salt-master in production (assuming you have one install)? Isn't the latest version generally "stable"?
00:36 manfred kind of
00:36 manfred so
00:36 manfred if your current environment is stable, leave it as is
00:36 manfred then have a testing environment to make sure all your shit works withthe new version
00:36 manfred when it works... upgrade your production environment
00:37 manfred (docker and vagrant are acceptable for testing before pushing to production so you dont have a million things that are duplicated
00:37 timbermaniac in a perfect world that'd be great. we support multiple customers on one box at the moment, and stuck at 0.17.4, because it's been working
00:37 manfred timbermaniac: https://github.com/simonmcc/kitchen-salt
00:38 manfred timbermaniac: kitchen-salt + vagrant can help you test if it will work withthe new versions
00:38 manfred then upgrade
00:38 pentabular joined #salt
00:38 manfred that is the point of those products
00:38 manfred imo
00:38 jalbretsen joined #salt
00:39 TheThing joined #salt
00:39 timbermaniac thanks, I'll take a look into it. but it will be fun making vagrant boxes for amazon linux instances
00:40 manfred timbermaniac: https://github.com/mitchellh/vagrant-aws
00:40 manfred it can do that
00:40 manfred no idea how it works with salt though
00:41 timbermaniac oh nice, that's a great start
00:41 timbermaniac had no idea that was a thing
00:42 skyler Alright, I just created my first pull request ever: https://github.com/saltstack-formulas/postfix-formula/pull/3
00:42 manfred imo, it is better to just do vagrant locally with virtualbox :P but if that ec2 one works, that is awesome, i have been meaning to try the vagrant-rackspace one, but it is at the botom of my list
00:44 manfred skyler: you tested it?
00:44 skyler manfred: yes, on Ubuntu only though.
00:46 manfred skyler: one second
00:46 bmatt vagrant has straight-up changed how I do my work
00:46 bmatt I can develop salt stuff *so* much faster
00:46 manfred vagrant is awesome
00:46 robawt ditto
00:46 manfred I am trying to move to docker more though
00:46 manfred rather than virtualbox
00:47 robawt i'm going down the middle trying to get vagrant to run my docker containers as testing envs
00:47 manfred skyler: lemme see what I can do, i am looking at pr #2 on that repo first, so you might have to rewrite parts of yours
00:48 skyler manfred: Ah, I should have checked that. It looks like they are similar.
00:49 manfred yours has more options it looks like, but his was first... i am not familiar with what I can actually do in saltstack-formulas... so I don't want to straight merge them both, so I am trying to check on any merge procedures
00:50 skyler manfred: If you merge his, I can change mine to refactor his (take the config out of init.sls) and add the extra options.
00:51 skyler I am going to afk for a little while for dinner now though.
00:51 manfred kk
00:51 manfred i will ping you
00:54 manfred skyler: i merged #2, it looks like your's still cleanly merges, do you want to double check that you aren;t adding anything extra?
00:59 joehoyle joined #salt
01:00 mapu joined #salt
01:01 cachedout joined #salt
01:03 jchen joined #salt
01:06 timbermaniac So it sounds like 2014.1.7 might be what I need to get going. Kinda sucks, will have to put away EL7 and get EL6 in here (for EPEL's simplicty)
01:08 robawt timbermaniac: you should roll your own repo anyway
01:08 robawt then you can deploy what you want
01:09 manfred ^^, though i just deploy tags directly from git
01:10 patrek joined #salt
01:10 timbermaniac manfred: where's it install to when you run install_salt.sh.sh ?
01:11 timbermaniac that sounds like it could work better
01:11 manfred install_salt.sh?
01:11 manfred i use https://bootstrap.saltstack.com
01:11 timbermaniac sudo sh install_salt.sh.sh -M -N git develop
01:11 manfred ahh
01:11 manfred i don't rememebr
01:11 manfred pip list, and it will show you the location
01:11 dsolsona joined #salt
01:12 manfred curl -sL https://bootstrap.saltstack.com | sh -s -- git v2014.1.10
01:12 manfred or w/e
01:14 timbermaniac that looks useful
01:14 timbermaniac funny, on 21:16 < manfred> ahh
01:14 timbermaniac =/  nevermind that
01:14 skyler manfred: I will take a look.
01:15 timbermaniac on https://bootstrap.saltstack.com it says that it supports CentOS 5+6, yet the changelog states it should work with 7 :)
01:16 manfred it should
01:17 pentabular joined #salt
01:19 skyler manfred: There actually are conflicts because postfix/main.cf from pull request #2, is the same config file as postfix/files/main.cf in mine. I will work on resolving these issues in a way that preserves all of #2's changes.
01:19 manfred cool
01:19 philip741 joined #salt
01:19 manfred just update that one and tag me with @gtmanfred, and I will merge it when you are ready
01:24 skyler manfred: will do!
01:27 timbermaniac manfred: my problem's reproducable in 2014.1.7
01:28 manfred you are just doing salt \* state.sls common.pip_packages.trac essentially right?
01:28 timbermaniac salt 'hostname' state.highstate
01:29 timbermaniac top.sls contains "common.pip_packages.trac"
01:29 manfred right
01:29 timbermaniac since that does the import and requires for the other packages
01:29 manfred but does it fail when you do the state.sls directly with that common.pip_packages.trac ? or doesit only fail in the top.sls /highstate?
01:29 timbermaniac (to be fair there's vim also included, but it's not problematic)
01:29 timbermaniac let me try
01:30 manfred i am wondering if it is a bug in state.highstate, or the whole thing
01:30 timbermaniac same deal with specifying the state
01:31 manfred looking at it, it looks all right, can you check the issue tracker to see if there is anything that pertains to yours in there, and if not please open a bug issue, and include the salt --versions-report in it?
01:31 thehaven joined #salt
01:32 timbermaniac certainly
01:32 manfred and include the stuff from your pastebin http://pastebin.com/tL1QhYcs in markdown so that it is blocked in the code blocks in the issue
01:32 manfred someone will get a chance to look at it, I cause from what I am seeing right now it looks correct
01:34 manfred timbermaniac: thansk
01:34 manfred there was someone in here the other day having a similar issue, and I dont remember it getting fixed
01:38 timbermaniac manfred: doesn't look to be an existing issue. Never having reported an issue before, should I report the version for the master, minion (and for both 2014.1.7 and .10)?
01:40 manfred all of the above?
01:41 manfred i would report it for 2014.1.10, since they are all just on the 2014.1 branch
01:42 timbermaniac sure thing
01:42 timbermaniac thanks
01:45 manfred np, thanks for filling it out
01:45 manfred timbermaniac: make sure to get all the stuff that is asked for in https://github.com/saltstack/salt/blob/develop/Contributing.rst , mostly i think it is the --versions-report
01:47 rojem joined #salt
01:47 snuffeluffegus joined #salt
01:50 timbermaniac manfred: Should I use the generic error for the title?
01:50 timbermaniac No matching sls found for '<sls_file>' in env 'base'
01:51 manfred uhhh, i would title it somehting like, 'states in nested directories not available'
01:51 manfred no matching sls found in nested directories
01:51 manfred something like that
01:52 bhosmer joined #salt
01:53 timbermaniac Submitted
01:54 manfred timbermaniac: looks good, thanks, someone will might look at it tonight, but should be commented on by sometime tomorrow
01:55 dalexander joined #salt
01:55 ml_1 joined #salt
01:55 timbermaniac sounds good, thanks
01:56 skyler Okay, I am having a weird problem. It works when I pass a string as a default value for pillar.get. However, when I pass in 'yes' or 'no', then become True or False in the actual configuration file.
02:05 skyler Replace 'no' with "'no'" seems to work, but I do not understand why there was a problem in the first place.
02:07 estherbester left #salt
02:07 cowyn_ joined #salt
02:09 possibilities joined #salt
02:11 philip741 left #salt
02:15 snuffeluffegus joined #salt
02:19 capitalfellow joined #salt
02:20 Outlander joined #salt
02:43 bmatt skyler: http://docs.saltstack.com/en/latest/topics/troubleshooting/yaml_idiosyncrasies.html
02:54 schimmy joined #salt
02:57 schimmy1 joined #salt
03:02 possibilities joined #salt
03:02 yetAnotherZero joined #salt
03:20 pentabular joined #salt
03:24 catpiggest joined #salt
03:25 joehoyle joined #salt
03:31 pastacino joined #salt
03:33 just_ joined #salt
03:33 just_ On RHEL6, the proper Jinja package 'python-jinja2' was moved from EPEL to the "RHEL Server Optional Channel". Verify this repository is enabled before installing salt on RHEL6.    how to verify?
03:34 just_ and  when install jinja2, it said python-label >=0.8, but i installed python-label 1.3 already?
03:39 cpowell joined #salt
03:41 bhosmer joined #salt
03:41 schimmy joined #salt
03:44 jaimed joined #salt
03:44 schimmy1 joined #salt
03:55 malinoff joined #salt
03:55 mosen joined #salt
03:56 ml_1 joined #salt
04:00 skyler bmatt: That was exactly the problem, and that whole page makes for a good read to avoid being tripped up later.
04:01 jnials joined #salt
04:03 jnials_laptop joined #salt
04:13 jnials joined #salt
04:18 malinoff joined #salt
04:18 schimmy joined #salt
04:23 schimmy joined #salt
04:30 trevorj joined #salt
04:38 toofer joined #salt
04:40 thayne joined #salt
04:43 aparsons joined #salt
04:44 felskrone joined #salt
04:45 possibilities joined #salt
04:47 ramteid joined #salt
04:48 aparsons_ joined #salt
04:58 pssblts joined #salt
04:59 Ryan_Lane joined #salt
05:03 Zevulon joined #salt
05:05 Zevulon moderator present?
05:07 miles32 I see an op
05:07 n8n joined #salt
05:08 Zevulon my daughter is interested in old school hackin
05:08 miles32 neat
05:08 miles32 how can we help?
05:09 Zevulon Ihave not use IRC in 10 years. time to get up to date. Me and her
05:10 miles32 it's still the same old same old
05:10 Zevulon Hackd NASA years ago. not happening today lol
05:12 Zevulon She wants to hack her moms WiFi.Im not up to date on that
05:12 miles32 it's childs play now days
05:12 miles32 (quite literally)
05:14 Zevulon childs play from a mobile?
05:14 miles32 is the mobile rooted?
05:14 Zevulon doubt it
05:14 Sauvin joined #salt
05:14 miles32 then no
05:15 __number5__ you guys might want to continue your discussion at #hacking or #anonymous
05:15 Zevulon so root, I been holding out because mine is business phone but hers is cheapy phone
05:16 miles32 mm, probably more trouble then it's worth
05:16 miles32 sorry
05:16 Zevulon Thanks number 5, your just first channel I hit for help
05:17 miles32 I hope #hacking is 1337
05:17 miles32 it's well named for 1337ness
05:18 Zevulon thanks miles, switching channels.thanks for direction
05:19 mordonez joined #salt
05:21 Zevulon #hacking is dead room
05:22 miles32 oh noes
05:22 miles32 might just google a bit and do some reading
05:24 Zevulon what is your channels specialty
05:24 felskrone1 joined #salt
05:25 miles32 we talk about compilers and templates and variable thingies and sometimes somebody will say the word lambda
05:25 ken joined #salt
05:26 Zevulon linux stuff
05:27 miles32 no no, windows stuff too
05:27 miles32 and osx and embedded systems
05:28 Zevulon Nice, Im a droid guy but must have respect for all coding
05:30 Zevulon IMO Windows has went to poo over the years
05:30 bhosmer joined #salt
05:30 miles32 it's not bad as far as administration goes.
05:30 miles32 they've done quite a few things under the hood to make life easier
05:31 miles32 instead of wanting to flip tables, you just want to cry yourself to sleep now
05:31 miles32 that's an improvement eh?
05:31 Zevulon Its the hood that I grown to dislike.to much fluff
05:33 miles32 mm, they'll sort it.  As long as my users can use the web and office they are pretty satisfied
05:34 Zevulon your an admin of some sort. I used to install networks at UC Davis
05:35 miles32 what sort?
05:36 Zevulon In the day of CAT 5 and Hubs
05:36 Zevulon way back
05:36 miles32 mm cable runs
05:36 miles32 so much fun
05:37 Zevulon dusty attic spaces and server rooms
05:37 miles32 I got fiber being run in a month or so
05:37 miles32 I'm excited
05:38 miles32 and ready to glare at any contractors with a sawzall
05:38 joehoyle joined #salt
05:38 Zevulon nice, obsolote to my day
05:40 Zevulon LOL, we had to do what we had to do back then
05:42 Zevulon you say you rooted droids
05:43 miles32 for tinkering, sure
05:43 Zevulon any fails
05:44 miles32 nothing bricked
05:45 Zevulon nice, want to root my NOTE but do not want to brick my work phone
05:46 miles32 you can pick up older android based phones for pretty cheap
05:46 miles32 then tinker with them
05:48 Zevulon I used to work on computers 15hours a day.Rooting not much of a leap?
05:48 Zevulon I hated command line coding
05:49 miles32 is pretty easy
05:49 miles32 computing has gotten much easier
05:49 miles32 or I should say, the hard work is done at a different level.
05:49 aparsons joined #salt
05:49 miles32 how do you abstract something to an understandable level without sacrificing power?
05:50 Zevulon guess thats my fear. most of the control is now automated
05:50 miles32 if that's a fear, you're in the wrong channel
05:51 Zevulon everything I did was in syntax. thats how I learned
05:51 miles32 60-70% of my job is on a whiteboard
05:53 Zevulon you know a good mod chan?
05:54 miles32 not off the top of my head I'm afraid
05:54 miles32 but you should be able to access the channel listing
05:54 miles32 and then sort by participant count
05:55 Zevulon tried #cyanogenmod but cant join
05:55 miles32 you might need to register with nickserv
05:56 miles32 a lot of channels require you to be identified before you're allowed to join
05:56 miles32 especially since freenode is for "srs bsns"
05:58 FeatherKing joined #salt
05:59 FeatherKing anyone seen any return output like this? command "commandhere" would have been executed
05:59 orion_ joined #salt
06:00 Zevulon OK, refresh my memory. how to reg nickserv. been a few days
06:00 Zevulon invite only?
06:00 miles32 http://ideone.com/ FeatherKing if you'll paste relevant output here, we might be able to take a look
06:00 FeatherKing miles32: when i run it, it works, when my indian counterpart runs the same command
06:00 FeatherKing its like test mode or something
06:00 FeatherKing idk
06:00 mosen FeatherKing: usually on a test run, if test=True
06:01 FeatherKing yeah thats what im thinking
06:01 FeatherKing but i copy/paste and it goes for me
06:01 FeatherKing two nights in a row
06:01 FeatherKing idk
06:01 mosen FeatherKing: maybe test mode is set on the minion config
06:01 miles32 training wheels were put on your indian counterparts machine
06:01 FeatherKing heh
06:01 FeatherKing but that would affect me too right
06:01 cym3try joined #salt
06:02 FeatherKing let me paste some data
06:03 mosen "If the value test is set to True in the minion configuration file then states will default to being executed in test mode"
06:03 TyrfingMjolnir joined #salt
06:03 mosen check /etc/salt/minion for a test: True possibly
06:03 miles32 can you override that with explicit test=False
06:03 miles32 before rooting in minion configs?
06:04 FeatherKing http://pastebin.com/4c8M9vYP
06:04 FeatherKing just a sample
06:04 FeatherKing well now the update is running because it worked for me
06:05 FeatherKing on a couple random minion configs test: True is commented out
06:05 FeatherKing we have to do more tmrw night, maybe i will have him add test=False
06:05 mosen ahh ok, was worth a true
06:06 mosen try even
06:06 FeatherKing ya
06:07 FeatherKing i may have found it
06:07 FeatherKing refer to this comment "training wheels were put on your indian counterparts machine"
06:08 FeatherKing he had added two question marks thinking they were part of the command
06:08 FeatherKing salt 'server' state.sls upgrade prod --async ??
06:08 FeatherKing apparently that throws salt into test mode
06:08 panchisco joined #salt
06:09 miles32 wonder if the ?? is an operator for something
06:09 miles32 and if that behavior is documented
06:09 FeatherKing that is really bizarre
06:09 FeatherKing totally worked, just in test mode
06:11 FeatherKing thank you both, off to bed now that my indian emergency is ofer
06:11 FeatherKing over
06:13 Zevulon thanks for help Miles. Out for the night
06:13 miles32 take care
06:17 jalaziz joined #salt
06:23 joehoyle joined #salt
06:24 aquinas_ joined #salt
06:27 matthiaswahl joined #salt
06:27 matthiaswahl joined #salt
06:30 schimmy joined #salt
06:37 n8n joined #salt
06:52 TyrfingMjolnir joined #salt
06:53 che-arne joined #salt
07:04 ml_1 joined #salt
07:09 colttt joined #salt
07:17 ajw0100 joined #salt
07:18 bhosmer joined #salt
07:21 lcavassa joined #salt
07:24 aparsons joined #salt
07:26 panchisco joined #salt
07:26 TyrfingMjolnir joined #salt
07:29 linjan joined #salt
07:31 ajolo joined #salt
07:44 sectionme joined #salt
07:51 intellix joined #salt
07:53 slav0nic_ joined #salt
07:56 MrTango joined #salt
08:03 panchisco joined #salt
08:03 darkelda joined #salt
08:10 iMil joined #salt
08:21 masquerade83 joined #salt
08:21 masquerade83 Hi
08:21 masquerade83 I am new to SaltStack
08:21 masquerade83 i am looking for setting up Openstack Havana Using SaltStack on Ubuntu 12.04LTS
08:32 Katafalkas joined #salt
08:33 patrek joined #salt
08:33 rawkode joined #salt
08:40 msciciel_ is there any trick to put grains in /etc/salt/minion.d/grains.conf ? i can't manage to do this successfully :(
08:46 dsolsona joined #salt
08:49 martoss joined #salt
08:52 pduersteler joined #salt
08:54 pduersteler HI all. I got a leftover debian kvm machine with 128mb ram and a few gigs of hdd. would that suffice as salt master for managing 10 hosts? May be a bit of a stupid question, but I couldn't find any requirements on the official pages / faq
09:00 matthiaswahl left #salt
09:08 bhosmer joined #salt
09:08 mike25de1 joined #salt
09:11 martoss joined #salt
09:13 Shenril joined #salt
09:15 mike25de1 hi guys ... i have an issue:     http://pastebin.com/SWkd5DUZ
09:16 mike25de1 it's a state that used to run perfectly... now something is wrong... and i have no idea :  Rendering SLS app.munin failed, render error: expected token 'comma', got 'string';
09:16 mike25de1 i can imagine is something very stupid ... that I've done...
09:27 SheetiS joined #salt
09:32 mike25de1 anyone?
09:32 mike25de1 :)
09:32 babilen Lets see ..
09:32 mike25de1 ah thanks babilen ... is very weird
09:33 mike25de1 it used to work,.. i have set up 50+ machines with that state
09:33 mike25de1 now.. if i run it .... it gets weird on me....
09:33 babilen jinjamap is undefined in that context
09:33 mike25de1 and i have no idea what might be the issue
09:33 mike25de1 it is ... i forgot to add that line
09:34 mike25de1 it freaks at this line: makedirs: True
09:35 [HACKINGLINKEDIN joined #salt
09:35 babilen mike25de1: That YAML parses perfectly with a different parser
09:40 mike25de1 babilen:  ... then ... what might be the issue?
09:40 malinoff I guess that might be the issue with incorrect parsing of 'mode' argument
09:41 mike25de1 it used to work just fine... what might cause this? or how to fix? anyone has any idea?
09:41 mike25de1 :(
09:41 malinoff mike25de1, there is no magic at all, review your changes and define what have been changed
09:42 mike25de1 babilen:   malinoff:  thank you
09:43 malinoff remove cached files, use state.template function to figure out what's going on
09:43 malinoff also running salt-minion with -l debug is very helpful
09:43 mike25de1 i already run with -l debug ... nothing there
09:43 mike25de1 where are the cached files?
09:43 mike25de1 on the master?
09:43 malinoff /var/salt/cache
09:43 malinoff on minions
09:44 babilen mike25de1: Sorry, got sidetracked :)
09:44 mike25de1 i use salt 0.17 .. i don't have that /var/salt/cache
09:44 malinoff mike25de1, it's a bit outdated :)
09:45 babilen mike25de1: I'd remove bits by bits until it works and then add them again. That way you should be able to figure out what's wrong. I can't see anything and it might also be your jinjamap.
09:45 babilen 0.17 ?! Oh dear ...
09:45 malinoff Somebody uses RAET already?
09:45 mike25de1 malinoff:  i know.... but when i tried to updaate ... states were f**ed
09:46 malinoff mike25de1, say hi to salt versioning and ignoring backwards compatibility :)
09:47 rawkode When salt is installed through bootstrap, do you just update by running bootstrap again?
09:47 mike25de1 malinoff:  what do you mean ?
09:47 malinoff mike25de1, salt does not care about backwards compatibility at all
09:47 mike25de1 malinoff:  the current version of salt... is it stable?
09:47 malinoff mike25de1, no, of course
09:48 mike25de1 ah malinoff i did not know that !!
09:48 mike25de1 thanks malinoff
09:48 faust joined #salt
09:48 malinoff mike25de1, salt wasn't stable ever :)
09:48 mike25de1 :) good to know
09:49 malinoff even though it is awesome
09:49 mike25de1 thank you guys
09:50 * malinoff thinks 0.17.5 was the most stable release ever
09:53 babilen malinoff: fwiw, I've been using 2014.* for a while now and rarely *if ever* had to change my states/pillars. The most nasty thing was the master/minion communication bug in .5, but apart from that it was quite alright. Sure, a bit more testing would be nice, but that is IMHO something that has also been tackled by the community (and some people trying new releases before they transition into "stable" repositories)
09:54 malinoff babilen, just search github issues for 'external blah blah blah'
09:55 malinoff ext_pillars, for example
09:55 malinoff Ansible is way more stable than salt, unfortunately
09:56 malinoff and if it is ok for to manage the management (to upgrade all your minions to the latest master) - salt is fine, but i have a very difficult project-based infrastructure and i just don't have enough time to do that
09:57 martoss joined #salt
09:59 canci joined #salt
10:02 dsolsona joined #salt
10:13 [HACKINGLINKEDIN joined #salt
10:14 peters-tx joined #salt
10:17 bhosmer joined #salt
10:42 Outlander joined #salt
10:57 [LINKEDINLOGSRZA joined #salt
11:03 bhosmer joined #salt
11:04 capitalfellow joined #salt
11:05 bhosmer_ joined #salt
11:10 veselin joined #salt
11:18 rawkode joined #salt
11:21 greyhatpython joined #salt
11:31 intellix joined #salt
11:35 rawzone joined #salt
11:35 ramishra joined #salt
11:47 bhosmer joined #salt
11:47 SheetiS joined #salt
11:49 rawkode joined #salt
11:50 diegows joined #salt
11:58 diegows are there some kind of profiling in salt/
11:58 diegows ?
12:00 diegows I would like to know how much time does it take to apply every state
12:03 babilen I'm not aware of a tool that would do that, but you aren't the first to ask for it
12:04 babilen Maybe it exists somewhere, but just managed to stay well hidden
12:04 otter768 joined #salt
12:06 Twiglet_ joined #salt
12:12 brandon_ joined #salt
12:14 mapu joined #salt
12:15 hobakill joined #salt
12:16 CeBe joined #salt
12:17 rome joined #salt
12:18 SheetiS joined #salt
12:21 hobakill does anyone know the status of 2014.1.10 in EPEL testing? does it have enough karma to get into stable?
12:26 SheetiS I've noticed it takes a few weeks for epel to push stuff through.  at least it did with 2014.1.7
12:27 hobakill SheetiS: yeah i had to get the developers to get 2014.1.7 into stable via karma. it had been over two weeks. and i only got them to do it because i had an accepted issue in the github repo.
12:31 mikepea joined #salt
12:33 bhosmer joined #salt
12:39 rome joined #salt
12:42 FarrisG joined #salt
12:42 jhauser joined #salt
12:43 miqui joined #salt
12:45 alanpearce joined #salt
12:45 bhosmer_ joined #salt
12:45 SheetiS joined #salt
12:46 rome joined #salt
12:48 geekmush joined #salt
12:49 bhosmer joined #salt
12:50 jslatts joined #salt
12:51 CeBe joined #salt
12:52 rojem joined #salt
12:53 rome joined #salt
12:56 dariusjs joined #salt
12:58 bhosmer joined #salt
12:59 icebourg joined #salt
12:59 dariusjs something like this should work on salt ? its a windows run  salt -v 'SomeHost*' cmd.run 'runas /user:someuser@corp.com "C:\Program Files (x86)\Puppet Labs\Puppet\bin\run_puppet_interactive.bat"' shell='cmd'
13:00 dariusjs with a -v it just keeps saying execution is running but I never see an end of it
13:00 cpowell joined #salt
13:01 icebourg joined #salt
13:01 dariusjs brb
13:02 icebourg joined #salt
13:03 davidnknight joined #salt
13:04 londo_ joined #salt
13:05 aparsons joined #salt
13:08 racooper joined #salt
13:09 mpanetta joined #salt
13:09 thayne joined #salt
13:11 rome joined #salt
13:22 TheThing joined #salt
13:24 rawkode Definitely time to start pulling my hair out with the Salt ec2 modules
13:31 TheoSLC joined #salt
13:31 smcquay joined #salt
13:31 TheoSLC joined #salt
13:32 kermit joined #salt
13:36 TheoSLC Good morning all.. is it too early?
13:36 manfred 0it is never too early
13:37 TheoSLC nice
13:38 TheoSLC Question.  I think I may understand my slow multi-master setup.  The minions use an ICMP ping to determine which master to do business with, right?
13:39 ramishra joined #salt
13:41 kaptk2 joined #salt
13:42 dude051 joined #salt
13:44 TheThing joined #salt
13:44 quickdry21 joined #salt
13:45 TheoSLC Anyhow.  my network blocks ICMP.  So if that's the method it uses, it would explain the behaviour
13:46 manfred that I have no idea
13:46 TheThing_ joined #salt
13:47 TheThin__ joined #salt
13:47 manfred i do not see anything in the multimaster tune_in
13:47 manfred looks like it just sends a request to all, and then uses the first one it gets back
13:48 manfred TheoSLC: https://github.com/saltstack/salt/blob/develop/salt/minion.py#L490
13:54 FeatherKing joined #salt
13:54 FeatherKing is there a way to turn the color off on a per command basis?
13:55 FeatherKing i would like to dump the default formatting output into a txt file to send but i get all the console color codes in there
13:55 rome joined #salt
13:57 tk75 joined #salt
13:59 blarghmatey joined #salt
14:00 orion_ joined #salt
14:00 ajprog_laptop joined #salt
14:00 kt76 joined #salt
14:02 mpanetta FeatherKing: --no-color does not do what you want?
14:02 jaimed joined #salt
14:03 FeatherKing mpanetta: perfect, i havent used that before. i was trying color=False but it wasnt going
14:03 FeatherKing ty
14:03 rojem joined #salt
14:03 mpanetta FeatherKing: No problem :)
14:06 seanz joined #salt
14:06 ipmb joined #salt
14:08 active8 joined #salt
14:09 smcquay joined #salt
14:10 vejdmn joined #salt
14:11 ipmb_ joined #salt
14:18 ramishra joined #salt
14:20 elfixit joined #salt
14:23 aquinas_ joined #salt
14:27 rallytime joined #salt
14:29 Ozack joined #salt
14:30 yetAnotherZero joined #salt
14:31 tmh1999 joined #salt
14:33 ramishra joined #salt
14:37 hobakill does salt support regex flags in compound matches in states?
14:40 icebourg joined #salt
14:41 QiQe joined #salt
14:42 babilen hobakill: You mean something like re.DOTALL or ... ?
14:42 hobakill babilen: more like a /i to ignore case
14:42 babilen so re.IGNORECASE :)
14:43 babilen hobakill: I *guess* that it is simply handed over to re.compile and would therefore expect anything that is discussed in https://docs.python.org/2/howto/regex.html to work, but I have in no way verified that.
14:44 babilen Why don't you try it and see?
14:44 hobakill ok. thanks babilen - yeah playing around with it now. thanks.
14:44 transmutated trying to set a variable interactively with cmd.run on a minion and it's saying it isn't found
14:45 transmutated specifically, __kwarg__: 1: meow: not found
14:45 transmutated yes my variable is meow:)
14:46 QiQe guys, when I use ssh_auth to add an public ssh keys everything goes fine, but if I want to remove it with "absent" exactly the same parameters, doesnt work
14:46 QiQe is there any bug on that?
14:47 QiQe or am I missing something?
14:47 dccc_ joined #salt
14:48 Ahlee trying to get a list of users on a minion in order to iterate over it and drop a file into each directory
14:48 Ahlee https://gist.github.com/jalons/73b6c5022797df626210
14:49 Ahlee looks like cmd.run('ls /home') is breaking everything into per-character
14:49 ramishra joined #salt
14:50 CeBe joined #salt
14:50 veselin Hi all, anyone using 'nodegroups' in state files? Could I match on them? e.g. {% if salt[ some_match_module]('nodegroup1') %} ?
14:50 Ahlee veselin: You should be able to, as it's just another expr_form
14:51 rawkode Anybody know what this means? "TypeError: string indices must be integers, not str"
14:52 rawkode String indices must be integers? What even?
14:52 KevinMGranger joined #salt
14:53 blarghmatey joined #salt
14:54 veselin Ahlee: thank you I tried compound match, but didn't work
14:54 babilen transmutated: What do you mean with "set a variable" ?
14:55 Ahlee veselin: compound matching is a pain. You're trying to match a nodegroup and something else?
14:55 econnell joined #salt
14:55 Ahlee in a state?
14:56 veselin Ahlee: just nodegroups would suffice, in a state
14:56 veselin Ahlee: e.g. install a package if minion is part of nodegroup X
14:56 babilen Ahlee: Sure, the return value of cmd.run is a string and you then iterate over that (which gives you each character) -- you'd have to split it first
14:58 Ahlee babilen: awesome
14:58 oz_akan joined #salt
14:58 Ahlee babilen: salt['cmd.run']('ls /home/').split() cleaned right up
15:00 penguin_dan joined #salt
15:00 QiQe answering myself, about ssh_auth absent, it is a bug
15:00 QiQe reported here: https://github.com/saltstack/salt/issues/3817
15:00 XenophonF joined #salt
15:01 QiQe still opened
15:04 ramishra_ joined #salt
15:05 hobakilllll joined #salt
15:09 veselin Ahlee: what type of matching would you recommend to match a nodegroup in a state?
15:10 Ahlee veselin: sorry, had to finish up my state, let me take a look
15:10 Ahlee just a moment
15:11 veselin Ahlee: of course, thank you
15:12 jslatts joined #salt
15:13 gnomechompski joined #salt
15:14 gnomechompski Has anyone since this traceback?
15:14 gnomechompski salt-minion -l debug
15:14 gnomechompski Traceback (most recent call last):
15:14 gnomechompski File "/usr/bin/salt-minion", line 6, in <module>
15:14 gnomechompski from salt.scripts import salt_minion
15:14 gnomechompski File "/usr/lib/pymodules/python2.6/salt/__init__.py", line 29, in <module>
15:14 gnomechompski import salt.log.setup
15:14 gnomechompski File "/usr/lib/pymodules/python2.6/salt/log/__init__.py", line 17, in <module>
15:14 gnomechompski from salt.log.setup import (
15:14 gnomechompski File "/usr/lib/pymodules/python2.6/salt/log/setup.py", line 25, in <module>
15:14 gnomechompski import logging
15:14 gnomechompski File "/usr/lib/python2.6/logging/__init__.py", line 1
15:14 gnomechompski root:*::
15:14 gnomechompski ^
15:14 gnomechompski whoa
15:14 gnomechompski sorry
15:14 p2 joined #salt
15:14 gnomechompski pastebin next time
15:14 gnomechompski didnt realize it was multi-line
15:14 gnomechompski left #salt
15:15 gnomechompski joined #salt
15:15 rockey it would've been fucking horrifying to read it if it wasnt multiline
15:15 gnomechompski true
15:15 Ahlee veselin: so the only thing i can think of is to do this in top.sls
15:16 gnomechompski so anyone know why this is?  I apt-get purged salt-minion
15:16 gnomechompski reinstalled
15:16 gnomechompski none of the other nodes are exhibiting this behavior
15:16 Ahlee drop the pkg.install into a state, and only apply that state on -match: nodegroup
15:17 Ahlee veselin: like on http://docs.saltstack.com/en/latest/topics/targeting/nodegroups.html where it says to apply the - webserver state to servers that match the group1 nodegroup
15:17 Ahlee gnomechompski: What version of salt?
15:17 veselin Ahlee: got it, thank you
15:18 gnomechompski 2014.1.10-1lucid1
15:19 Ahlee sorry, no 2014.1.10 nor lucid systems avaiable
15:19 Ahlee but, in the same python interpretor you're executing salt from
15:19 Ahlee launch python, try import salt.log.setup
15:19 Ahlee does it blow up?
15:19 gnomechompski ok, lemme try
15:21 gnomechompski yep, same error in python
15:21 gnomechompski all other nodes are on the same OS and salt version
15:22 wendall911 joined #salt
15:22 JoeHazzers why in the documentation examples is "lookup" used in the pillar data?
15:22 JoeHazzers Is it a convention I'm not aware of?
15:22 Damon joined #salt
15:22 geekmush joined #salt
15:24 sectionme joined #salt
15:35 KaaK what is the best way to debug the contents of a jinja variable?
15:36 jeffrey4l joined #salt
15:39 nahamu KaaK: perhaps use the "test" state and set the name to the contents of the variable?
15:39 nahamu http://docs.saltstack.com/en/latest/ref/states/all/salt.states.test.html#module-salt.states.test
15:40 nahamu oh, that's only in 2014.7... nevermind
15:40 KaaK nahamu, this is what google yielded: http://docs.saltstack.com/en/latest/ref/states/testing.html
15:41 nahamu you could run something with test=True, yes, but I was trying to find something that would really only log the variable without trying to run anything.
15:41 estherbester joined #salt
15:42 nahamu clearly someone expressed that need but so recently that it's only in 2014.7
15:42 dsolsona joined #salt
15:42 KaaK yeah ... more and more I'm feeling the pressure to update to it
15:42 abe_music joined #salt
15:43 KaaK yeah, the problem is my state file doesnt compile due to my variable not having what i think it should have in it
15:44 KaaK im merging pillar data with a jinja dict, and i think im losing keys in the merger ...
15:46 KaaK will a salt['grains.filter_by'](..., merge=...) preserve keys in the jinja dict? or will the merge dict stamp those out?
15:47 che-arne joined #salt
15:48 rojem joined #salt
15:50 ajolo joined #salt
15:53 gnomechompski left #salt
15:57 jnials joined #salt
15:58 JPaul joined #salt
15:59 n8n joined #salt
16:01 schimmy joined #salt
16:02 aparsons joined #salt
16:03 catpigger joined #salt
16:06 codekobe_ Has Anyone every tried to file.recurse your salt-master's file_roots?
16:07 codekobe_ trying to sync pillar data and states from a master-of-masters down to masters that run the syndic minion
16:09 rawkode joined #salt
16:09 Deevolution Is there a way for a given node to know what environment it's in without having to hard code it?
16:15 KyleG joined #salt
16:15 KyleG joined #salt
16:22 felskrone joined #salt
16:24 sectionme joined #salt
16:32 aparsons_ joined #salt
16:32 aparsons joined #salt
16:32 possibilities joined #salt
16:35 thayne joined #salt
16:35 penguin_dan joined #salt
16:41 joehoyle joined #salt
16:41 Topic for #salt is now Welcome to #salt | 2014.1.10 is the latest | Help us test the 2014.7 RC! http://bit.ly/salt-rc | Please be patient when asking questions as we are volunteers and may not have immediate answers | Channel logs are available at http://irclog.perlgeek.de/salt/
16:43 rawkode I have a salt-master (2014-07rc1) and it's bootstrapping minions with salt-minion (2014.1.10) - anybody know how I can fix that?
16:43 alanpearce joined #salt
16:45 rap424 joined #salt
16:47 penguin_dan joined #salt
16:48 penguin_dan joined #salt
16:49 schimmy joined #salt
16:51 schimmy1 joined #salt
16:52 intellix joined #salt
16:52 aw110f joined #salt
16:52 jnials joined #salt
16:54 joehillen joined #salt
16:56 davet joined #salt
16:57 forrest joined #salt
16:57 aparsons joined #salt
16:58 notpeter_ joined #salt
16:59 Ryan_Lane joined #salt
17:00 jms_ joined #salt
17:04 joehoyle joined #salt
17:06 jms_ I am getting a failed lookup for a node when I try to destory it
17:06 jms_ client.destroy('mediawarp-sjc.ban.io') Traceback (most recent call last):   File "<stdin>", line 1, in <module>   File "/usr/lib/python2.6/site-packages/salt/cloud/__init__.py", line 151, in destroy     mapper.destroy(names))   File "/usr/lib/python2.6/site-packages/salt/cloud/__init__.py", line 744, in destroy     ', '.join(names) salt.cloud.exceptions.SaltCloudSystemExit: The following VM's were not found: mediawarp-sjc.ban.io
17:07 jms_ I can apply state, run commands, it is in the salt-key list, any clue what piece is missing?
17:08 srage_ joined #salt
17:08 SheetiS joined #salt
17:09 ml_1 joined #salt
17:10 miles32 I can easily stop a windows service with a require statement
17:10 miles32 however, how do I ensure that I start that service backup again only after a file copy operation completes?
17:11 forrest miles32, can you not just drop the conf in place then do a watch?
17:12 miles32 and the watch will wait for the copy to complete I guess before it fires off
17:14 forrest miles32, I just mean a watch in the service to watch the managed filed.
17:14 forrest so it will only restart when it sees that file is updated.
17:15 miles32 mm, it's a folder
17:15 miles32 but I'll look in that direction
17:15 forrest miles32, a managed directory?
17:15 forrest you can still watch that.
17:15 miles32 yeah
17:16 forrest I assume you mean in terms of perms changing?
17:16 forrest miles32, if it's files within the folder, I don't know about a file.recurse. Never tried a watch on that.
17:17 miles32 I have a windows service, I have to turn it off, update the files then turn it back on
17:17 srage joined #salt
17:21 blarghmatey joined #salt
17:21 conan_the_destro joined #salt
17:23 druonysus joined #salt
17:23 druonysus joined #salt
17:24 meteorfox joined #salt
17:25 sectionme joined #salt
17:27 conan_the_destro joined #salt
17:28 jalaziz joined #salt
17:30 rojem joined #salt
17:31 hobakilllll joined #salt
17:32 FeatherKing joined #salt
17:36 veselin left #salt
17:37 vejdmn joined #salt
17:40 vejdmn joined #salt
17:41 forrest miles32, you can't update the files while the service is running?
17:41 forrest doesn't it load the confs into memory?
17:42 panchisco joined #salt
17:42 chrisjones joined #salt
17:44 druonysuse joined #salt
17:44 sectionme joined #salt
17:45 jla joined #salt
17:46 aparsons_ joined #salt
17:47 srage joined #salt
17:48 ramishra joined #salt
17:49 jla I am trying out salt and checking master and minion keys before connecting, but /etc/salt/pki/minion/ is empty when I first run salt-minion. It seems to populate if I give it the master name/ip. When does it create minion.pub? How can I make it create it sooner?
17:51 joehoyle joined #salt
17:52 forrest manfred, do you remember if there's a way to force the key gen on a minion?
17:53 aquinas_ joined #salt
17:55 manfred yes
17:55 manfred kind of
17:56 manfred salt-key --gen-keys=$(hostname)
17:56 manfred that will generate a key
17:56 jla I've seen salt-key --gen-keys from the master, I was wondering what was holding up the minion from doing it before connecting
17:57 manfred once you accept the minion, the key is passed to it iirc, but i haven't dug into it that much
17:57 jla or I guess I could install salt-master, run salt-key, put those files in place (ala pre-seed), and then uninstall salt-master
17:57 transmutated joined #salt
17:57 manfred the way that salt-cloud does it is it generates the key, and then passes it to /tmp/.salt-cloud/minion.key or w/e and the just puts it in place on the minion
17:58 panchisco joined #salt
17:58 jla that makes sense for bootstrapping and pre-seeding out to the minion
17:59 manfred i don't know how you an tell it to make it sooner
17:59 forrest manfred, Hmm, yeah that was the only way I could think of, nothing for making the minion just 'do it'
17:59 manfred other than making them on your master and manually moving them over
17:59 manfred yeah
17:59 manfred the minion doesn't make the keys
17:59 forrest right
17:59 jla it just seemed odd that the minion waits until it can talk to a master before it makes minion.* keys
17:59 manfred the master gives it to them when it accepts the connection
17:59 bhosmer_ joined #salt
17:59 manfred jla: the minion doesn't make it, the master gives it to i afaik
17:59 manfred you accept the first connection, and wwhen it is accepted, the key is passed to the minion
17:59 jla oh
18:00 manfred that is how I have understood it
18:00 manfred double check it though, i could be wrong on that
18:01 jla that seems to make `salt-call --local key.finger` seem a little less useful.
18:01 jla k
18:01 manfred jla: i am like 90% sure that the key on the minion is a public key, so the master just passes it to it
18:02 jla this isn't a big deal, I can look at the minion fingerprint on the minion and the master once it connects before accepting it. It just seemed odd that the minion didn't have it earlier, unless as you say the master creates and sends the minion the public and private key.
18:02 jla I would expect only the minion to have it's private key
18:03 manfred the minion doesn't have a private key afaik
18:03 manfred pretty sure the maste rhas the private key
18:03 perfectsine joined #salt
18:04 manfred jla: https://github.com/saltstack/salt/blob/develop/salt/utils/cloud.py#L405
18:04 manfred the salt.utils.cloud.bootstrap function deploys a .pub key
18:05 ckao joined #salt
18:05 ramishra joined #salt
18:05 jla I think the minion.pem is a private key
18:05 ajprog_laptop joined #salt
18:06 manfred ahh it is
18:06 KevinMGranger So, I want to ensure a package is installed on certain machines, but it's only packaged for one of the target OS's. I can easily get the binary with cmd.run to download it, and pick between that and pkg.installed using grain os checking, but what if I need to require that later?
18:06 alanpearce joined #salt
18:06 jla the master /etc/salt/pki/master/minions/ seems to hold public keys
18:06 miles32 fairly often I have to lookup jobs and then use lookup_jid to get the full output from highstate
18:06 miles32 which annoys the hell out of me
18:07 alanpear_ joined #salt
18:08 dccc_ joined #salt
18:10 aparsons joined #salt
18:10 jla KevinMGranger: if you can't get packages for the other OSs, but the binary from the one works for the others, would Esky / bbfreeze do what you want? http://salt.readthedocs.org/en/latest/topics/tutorials/esky.html
18:10 KevinMGranger miles32: try setting timeout to a fairly high value
18:11 KevinMGranger jla: thanks for the suggestion, but I think that's a little overkill for solving this problem
18:11 KevinMGranger Perhaps I could force the provider for pkg.installed to be a custom command if it's an os that doesn't have it packaged. But there's really not much documentation on that
18:12 possibilities joined #salt
18:13 SheetiS joined #salt
18:14 PsionTheory joined #salt
18:15 kelseelynn joined #salt
18:16 miles32 polymophic states
18:16 miles32 or somethin
18:16 perfectsine joined #salt
18:18 cpowell joined #salt
18:18 KevinMGranger uhh what
18:19 KevinMGranger like a custom state that provides the pkg virtual state?
18:19 KevinMGranger that's doable I guess
18:22 jla manfred: If I give the minion an invalid IP for the master instead of a DNS name, like 127.0.0.1, the minion creates it's public and private keys. The creation seems to have been blocked by DNS resolution of the salt_master name
18:22 jla manfred: and then salt-call --local key.finger works, even if salt-minion is no longer running
18:23 miles32 the documentation for the newest release seems to indicate you can use all sorts of nifty if statements
18:24 miles32 in your states
18:24 forrest miles32, ?
18:24 manfred jla: yeah, salt-call will start salt-minion if it isn't running
18:24 forrest miles32, you could already use if statements, was there some additional functionality added?
18:24 forrest miles32, you can write straight up python code for states if you want.
18:24 miles32 yeah above and beyond that
18:24 miles32 I mean baked in
18:24 forrest miles32, link?
18:24 miles32 one sec
18:24 forrest I see.
18:26 miles32 I'll find it later, but it was a line like "modifying state files" during state execution
18:27 KevinMGranger you mean jinja templating? yeah that's there
18:27 miles32 never mind
18:27 miles32 it was beyond the obvious
18:27 KevinMGranger the issue is, what if I need to depend upon that state later? How do I know if I need require: pkg: PKGNAME or require: cmd: PKGNAME-INSTALLED
18:28 rypeck joined #salt
18:30 KevinMGranger which is why I'm trying to figure out a custom provider, be it a module or state
18:30 KevinMGranger just trying to figure out which one is cleaner, or how to use a custom module as a provider at all
18:31 rojem joined #salt
18:32 druonysuse joined #salt
18:32 jalbretsen joined #salt
18:33 rawkode joined #salt
18:34 dccc joined #salt
18:40 bezeee joined #salt
18:40 smcquay joined #salt
18:42 miles32 file.recurse quite literally subtracts the entire contents of a 5000 line file
18:43 miles32 and then adds back in the entire contents of the 5000 line file
18:43 forrest miles32, yep
18:43 miles32 meaning the output for that one state is 10,000 lines
18:43 forrest it's just copying all the content
18:43 miles32 so now everytime I run highstate I have to pipe into grep
18:43 forrest miles32, you could change the log level
18:43 forrest I think it's 'terse' or something
18:44 forrest can't remember the exact name
18:44 miles32 the log level is what's pushing to std out?
18:44 forrest miles32, http://docs.saltstack.com/en/latest/ref/configuration/master.html#state-output
18:44 forrest it's state_output, not logging, sorry
18:44 Gareth this looks interesting.
18:44 Gareth https://github.com/DTVD/rainbowstream
18:45 forrest Gareth, do you really need more twitter distractions? :P
18:45 miles32 nifty
18:46 Gareth forrest: always.
18:46 forrest Gareth, Fair enough.
18:47 FeatherKing joined #salt
18:49 Gareth forrest: might be a good client to run in screen or tmux.
18:50 forrest Gareth, yeah, I get distracted easily though, don't need additional distractions.
18:53 bezeee joined #salt
18:56 jsm joined #salt
18:56 druonysuse joined #salt
18:56 druonysuse joined #salt
18:58 miqui joined #salt
18:59 cpowell joined #salt
19:01 cpowell_ joined #salt
19:02 dccc joined #salt
19:04 miles32 how do you guys handle updating the system ?
19:05 miles32 windows updates/apt/yum etc.
19:05 miles32 on a schedule, but without upgrading certain key packages
19:07 mapu joined #salt
19:07 aw110f joined #salt
19:22 kermit joined #salt
19:24 martoss joined #salt
19:24 forrest miles32, I've seen people just list all their system packages (since they do a very minimal install), then locking versions populated from pillar, so you have 'base' which does the latest of packages, then you go from there. I've also seen people who say screw it, and just reprovision for updates.
19:24 forrest I don't think there's a great solution anywhere :\
19:25 miles32 yeah...I'm using AMI's
19:25 forrest miles32, might be worth asking Ryan_Lane
19:27 miles32 mm, I'll look into it later
19:27 miles32 dentist
19:27 miles32 left #salt
19:29 Rojematic joined #salt
19:33 dalexander joined #salt
19:33 JPaul joined #salt
19:35 mgw joined #salt
19:36 joehoyle joined #salt
19:36 rawkode Whats everyone's preferred output for state.highstate? Also, is there a way to have the results in the run order?
19:39 ggoZ joined #salt
19:40 rojem joined #salt
19:43 chrisjones joined #salt
19:44 pentabular joined #salt
19:44 rallytime joined #salt
19:46 Ryan_Lane joined #salt
19:46 ericof joined #salt
19:50 econnell left #salt
19:52 capitalfellow joined #salt
19:52 joehoyle joined #salt
19:54 * Heartsbane blames whiteinge.
19:55 jaimed joined #salt
19:56 peno whiteinge is to be commended.
19:57 whiteinge Can that commending happen sooner rather than later? I need it to balance out the blame.
19:58 orion__ joined #salt
19:58 catpiggest joined #salt
20:02 orion_ joined #salt
20:02 catpigger joined #salt
20:03 dalexander joined #salt
20:05 peno it's working on a bell curve.
20:06 possibilities joined #salt
20:07 rojem joined #salt
20:09 Ryan_Lane1 joined #salt
20:15 rawkode joined #salt
20:24 FeatherKing when running a syndic does the master service need to be running also on that machine?
20:25 manfred FeatherKing: i believe it needs both the salt-master and salt-syndic running
20:25 FeatherKing ok thx
20:25 manfred yeah
20:25 manfred The master sitting at the top of the hierarchy (the Master of Masters) will not be running the salt-syndic daemon. It will have the salt-master daemon running, and optionally, the salt-minion daemon. Each syndic connected to an upper-level master will have both the salt-master and the salt-syndic daemon running, and optionally, the salt-minion daemon.
20:26 davet joined #salt
20:26 manfred http://docs.saltstack.com/en/latest/topics/topology/syndic.html
20:26 ipmb joined #salt
20:29 forrest SYNDICS ALL THE WAY DOWN
20:29 KaaK you're very clever sir, but it is infact turtles -- all the way down
20:36 forrest Yeah but that isn't as relevant!
20:44 Ryan_Lane1 has anyone had success using - reload_modules: True ?
20:44 Ryan_Lane1 from what I can tell it doesn't work
20:44 manfred Ryan_Lane1: in my experience, it reloads the modules at the very end of the whole state run, not after that specific state
20:45 Ryan_Lane1 well, that's worthless
20:45 Ryan_Lane1 I set the system's hostname during the run, then need to use the fqdn later
20:45 Ryan_Lane1 so I reload the modules
20:45 Ryan_Lane1 which is obviously not working for me
20:45 Ryan_Lane1 we also has some situations where we need to install something via pip, then use it
20:47 manfred Ryan_Lane1: nevermind, it does it at the end of each state, in the salt.state.check_refresh
20:47 manfred or it is supposed to
20:47 kermit joined #salt
20:47 Ryan_Lane1 hm. maybe I need to upgrade to a newer develop branch
20:47 manfred https://github.com/saltstack/salt/blob/develop/salt/state.py#L1560
20:54 joehoyle joined #salt
20:59 rojem joined #salt
21:01 rawkode Develop branches scare me
21:03 possibilities joined #salt
21:06 tcotav in software, they're all develop branches -- some are just mistitled ;-)
21:06 rykan joined #salt
21:08 Voziv joined #salt
21:09 catpiggest joined #salt
21:13 snuffeluffegus joined #salt
21:13 catpigger joined #salt
21:16 rome joined #salt
21:18 orion_ joined #salt
21:18 rojem joined #salt
21:19 sijis left #salt
21:24 vejdmn joined #salt
21:24 Ryan_Lane1 manfred: ugh: https://github.com/saltstack/salt/blob/develop/salt/state.py#L716
21:25 manfred yeah
21:25 manfred it should run it at the end of the state call
21:25 Ryan_Lane1 except...
21:25 Ryan_Lane1 read the description
21:25 Ryan_Lane1 and see what the function does
21:25 manfred oh
21:25 manfred ugh
21:25 Ryan_Lane1 it only reloads under very specific conditions
21:26 Ryan_Lane1 so lame
21:26 manfred read the first if statement in it
21:26 manfred if data.get('reload_modules', False) is True:
21:26 manfred # User explicitly requests a reload
21:26 manfred self.module_refresh()
21:26 manfred if reloadmoduels is true
21:26 manfred it runs self.module_refresh()
21:26 Ryan_Lane1 hm. weird
21:26 manfred otherwise, it only reloads modules on the other things
21:26 Ryan_Lane1 that isn't working for me
21:27 manfred hrm
21:28 manfred if not self.opts.get('local', False)
21:28 manfred Ryan_Lane1: that might be why
21:28 Ryan_Lane1 RAWR
21:28 manfred aren't you running all local minions?
21:28 Ryan_Lane1 yes
21:28 manfred bingo
21:28 Ryan_Lane1 -_-
21:28 Ryan_Lane1 why all the masterless hate? :)
21:29 manfred Ryan_Lane1: it should still run this one https://github.com/saltstack/salt/blob/develop/salt/state.py#L673
21:30 manfred Ryan_Lane1: it looks like that should reload everything...
21:30 Ryan_Lane1 it should also refresh the grains
21:30 Ryan_Lane1 which it isn't doing
21:31 Ryan_Lane1 for instance, I set the hostname, but my fqdn grain doesn't change
21:31 manfred how are you refreshing the grains?
21:31 Ryan_Lane1 reload_modules: True
21:31 Ryan_Lane1 that should reload the grains
21:31 manfred it doesn't
21:31 Ryan_Lane1 really?
21:31 manfred yeah
21:32 manfred like 90% sure
21:32 manfred one second
21:32 Ryan_Lane1 I don't see how not. it would reload the grain modules
21:32 manfred yeah, but it wouldn't reload the __grains__ dictionary
21:32 kermit joined #salt
21:32 manfred the grains are rendered from running all the stuff in salt.grains.*
21:32 manfred and loaded by the loader
21:32 Ryan_Lane1 ugh
21:33 Ryan_Lane1 is there any way to reload that?
21:33 manfred Ryan_Lane1: https://github.com/saltstack/salt/blob/develop/salt/loader.py#L336
21:33 catpiggest joined #salt
21:33 manfred if... reload_modules worked... i would just throw another loader into it
21:33 Ryan_Lane1 another loader?
21:33 manfred https://github.com/saltstack/salt/blob/develop/salt/state.py#L700
21:33 manfred like one of those
21:33 Ryan_Lane1 ah
21:34 manfred you would want to do it at the top of that though, imo, to make sure you the self.functions would have the self.grains in it
21:35 Ryan_Lane1 [DEBUG   ] Refreshing modules...
21:35 Ryan_Lane1 [INFO    ] Loading fresh modules for state activity
21:35 manfred yeah, i thought it worked...
21:36 vejdmn joined #salt
21:36 ajprog_laptop joined #salt
21:38 manfred ok, I am heading home, good luck Ryan_Lane1
21:41 SheetiS joined #salt
21:42 intellix joined #salt
21:48 ekristen joined #salt
21:48 Ryan_Lane1 manfred: thanks
21:51 smcquay joined #salt
21:55 ajolo joined #salt
22:00 joehoyle joined #salt
22:01 FeatherKing joined #salt
22:03 yomilk joined #salt
22:05 dalexander joined #salt
22:08 elfixit joined #salt
22:11 manfred Ryan_Lane1: is it just modules that require a python library that this happens with?
22:11 Ryan_Lane1 nah
22:11 manfred oh
22:11 Ryan_Lane1 if I change my hostname, for instance
22:12 manfred if it was just ones that require a python module... i had a problem with that a long time ago, because it required a restart of the minion for it to pickup the addition to the pythonpath
22:12 manfred no i meant modules, not grains
22:12 manfred the grains one won't work, but refresh modules isn't loading stuff after you pip installed it earlier on was the problem I had
22:13 Ryan_Lane1 oh
22:13 Ryan_Lane1 yeah, we also noticed an issue with that with etcd
22:14 manfred oh, ok
22:14 Ryan_Lane1 we just worked around that by adding it to our salt venv
22:14 manfred then yeah, my solution before was you had to restart the minion, because mine was being used as a library internally on the salt-minion
22:14 manfred salt.utils.openstack.nova
22:14 bezeee joined #salt
22:15 Ryan_Lane1 ah
22:15 Ryan_Lane1 I see
22:15 Ryan_Lane1 yeah, in this case it was just for state runs
22:15 Ryan_Lane1 so it should have worked
22:15 Ryan_Lane1 oh, or is it?
22:15 Ryan_Lane1 utils/etcd_util.py
22:15 Ryan_Lane1 that's probably why
22:16 Ryan_Lane1 although, it should reload the state module, which would check the dependencies again via utils
22:16 analogbyte joined #salt
22:16 Ryan_Lane1 so that shouldn't be an issue
22:16 Ryan_Lane1 I think salt cloud should be refactored into modules and states
22:17 Ryan_Lane1 and salt cloud should make module/state calls
22:17 forrest Ryan_Lane1, you should talk to manfred/techhat about that
22:17 Ryan_Lane1 (like the boto modules/states)
22:17 forrest I'm sure they'd love to discuss it
22:17 * Ryan_Lane1 nods
22:17 manfred Ryan_Lane1: yup, that seems why
22:17 Ryan_Lane1 I don't really have time to work on that, but I personally never want to call salt-cloud
22:18 Ryan_Lane1 I want to do state runs
22:18 Ryan_Lane1 but I can't use any salt-cloud resources unless I run salt-cloud
22:18 samueltb_ joined #salt
22:18 Ryan_Lane1 so, no openstack, no GCE, etc.
22:18 manfred Ryan_Lane1: we don't want to be all seperate states and modules for each driver, we are using just the cloud.py module/state, and writing drivers are used directly from that
22:18 Ryan_Lane1 ah. I see
22:18 Ryan_Lane1 so I could use the cloud state?
22:18 manfred yes
22:18 manfred it just makes direct calls to teh CloudClient
22:19 manfred just like the salt-cloud binary does
22:19 manfred Ryan_Lane1: you can also store your provider and profile in pillar data now
22:19 Ryan_Lane1 but then all my module calls are via cloud as well?
22:19 manfred everything only calls via the CloudCLient class
22:19 manfred so that it is uniform
22:19 Ryan_Lane1 it's definitely not the normal way salt works :(
22:20 Ryan_Lane1 for boto modules, for instance, I can make normal salt module calls to different functions
22:20 Ryan_Lane1 they are all documented independently
22:20 manfred Ryan_Lane1: salt-call cloud.action network_list provider=my-nova
22:21 Ryan_Lane1 am I able to use salt-cloud without a master?
22:21 manfred is the same as salt-cloud -f network_list my-nova
22:23 manfred i can do a network_create or volume_create, all through the module
22:23 manfred and I only have to have one module
22:24 manfred Ryan_Lane1: there is someone that was working on being able to restart the whole minion without loosing the place in the 'state run' ... that would be what we need to solve the library problem
22:27 SheetiS joined #salt
22:31 yomilk joined #salt
22:34 Ryan_Lane1 manfred: https://github.com/saltstack/salt/pull/15134
22:35 manfred yar
22:35 lionel joined #salt
22:35 manfred that was what I was thinking
22:35 manfred does it fix your module loading issue?
22:35 Ryan_Lane1 yep
22:35 Ryan_Lane1 well, it fixes the grains issue for sure
22:36 Ryan_Lane1 I didn't try it with any grain that relies on a python module
22:36 joehoyle joined #salt
22:36 Ryan_Lane1 what's the chances this'll make it in for the release? :D
22:37 drawks whats the easiest way to have a minion's id set to it's hostname instead of it's fqdn?
22:38 Ryan_Lane1 drawks: set it explicitly in the minion config
22:38 manfred Ryan_Lane1: push the pull request against 2014.7 branch
22:38 manfred and have them forward merge it
22:38 Ryan_Lane1 do I need to do that for all my boto changes too?
22:38 drawks heh, is there a way to do it that doesn't require localizing each minion's config file?
22:38 manfred Ryan_Lane1: they are not adding feature changes anymore, but considering this one is a bug, yeah,
22:39 Ryan_Lane1 the boto changes are bug fixes for vpc
22:39 manfred Ryan_Lane1: all bug changes should be pushed against the version that they are fixing it for
22:39 manfred then yeah
22:39 Ryan_Lane1 and for test mode
22:39 manfred they switched up the merge rules on this one
22:39 Ryan_Lane1 so they should be pushed against master and the branch?
22:39 manfred imo, it is cleaner
22:39 Ryan_Lane1 well, develop, not master
22:39 manfred no, just the branch
22:39 Ryan_Lane1 ah, ok
22:39 manfred yar, then they will just forward merge them
22:39 Ryan_Lane1 so I should close this PR?
22:40 manfred yar, i would
22:40 manfred then include your patch in a branch of 2014.7, and then submit a pr between that branch and the 2014.7 branch
22:41 Ryan_Lane1 and for the boto changes that have already been merged, new PRs?
22:41 manfred nah, just ping @basepi and he can backport them still
22:41 manfred just tag the prs with his name
22:41 Ryan_Lane1 ok
22:42 mechanicalduck joined #salt
22:43 Ryan_Lane1 hm. he's basepi on github?
22:44 Ryan_Lane1 ah. yep
22:45 manfred yar
22:45 forrest Ryan_Lane1, it's Colton
22:45 forrest if that helps at all :P
22:46 rojem joined #salt
22:47 manfred Ryan_Lane1: you have to do a clean branch
22:47 manfred Ryan_Lane1: https://github.com/saltstack/salt/pull/15136
22:47 Ryan_Lane1 ah. crap
22:47 forrest lol
22:47 forrest Ryan_Lane1, WHY YOU DO DIS
22:47 manfred git checkout -b 2014.7 salt/2014.7
22:47 forrest manfred, Ryan knows ;P
22:47 manfred :)
22:47 Ryan_Lane1 :)
22:50 skyler When using "salt '*' state.sls example" I get this error: No matching sls found for 'example' in env 'base'
22:50 skyler however, example.sls exists. I includes just a single include statement. The include is broken.
22:51 skyler *It includes just a include statement.
22:51 forrest skyler, example.sls lives in /srv/salt?
22:51 skyler Yes.
22:51 Ryan_Lane1 fixed :)
22:51 forrest skyler, ok, what does the include include?
22:51 skyler - gibberish
22:51 forrest which isn't real
22:51 skyler exactly.
22:52 forrest Yeah salt errors are kinda bad for that stuff. What behavior are you expecting?
22:52 skyler so instead of saying that gibberish is not in base, it says that example is not there.
22:52 skyler I think it should say "No matching sls found for 'gibberish' in env 'base'"
22:52 forrest skyler, right, because example is what it's trying to read, so it sees that as the failure, because the include inside of it fails
22:52 forrest skyler, I agree
22:52 forrest I can't remember what the logic behind that is. I believe it has something to do with the yaml rendering, but I can't remember.
22:53 forrest but we've discussed the failure of accurate error reporting many times
22:53 manfred pretty sure it is in the tops rendering
22:53 forrest he's not rendering the top
22:53 manfred hold on, lemme find it
22:54 forrest manfred, oh you mean the error?
22:54 manfred yeah
22:54 manfred one second
22:54 forrest gotcha
22:54 manfred https://github.com/saltstack/salt/blob/develop/salt/state.py#L2381
22:54 forrest yeah I've seen this line before
22:55 forrest that's how common this question is :P
22:55 forrest manfred, that's only in develop though right?
22:55 manfred yeah
22:55 manfred i think it is fixed in develop
22:55 forrest manfred, frankly I still don't think that's fixed
22:55 manfred skyler: you were on 2014.1.10 right?
22:55 Ryan_Lane1 ugh. I can't rebase this PR: https://github.com/saltstack/salt/pull/15119
22:55 Ryan_Lane1 because dependencies for it aren't backported
22:55 manfred Ryan_Lane1: you don't need to rebase that one
22:56 skyler manfred, checking...
22:56 manfred Ryan_Lane1: the backport will do it
22:56 forrest manfred, that error is still totally worthless
22:56 manfred oh, it isn't merged yt
22:56 skyler manfred: yes.
22:56 Ryan_Lane1 yeah
22:56 forrest manfred, it should be saying which one is problematic as well.
22:56 Ryan_Lane1 can I just add "needs to be backported" to the conversation?
22:57 manfred forrest: it is fixed in develop
22:57 forrest Ryan_Lane1, probably.
22:57 manfred [root@salt ~]# salt-call state.sls test
22:57 manfred [INFO    ] Loading fresh modules for state activity
22:57 manfred [INFO    ] Fetching file from saltenv 'base', ** done ** 'test.sls'
22:57 manfred local:
22:57 manfred Data failed to compile:
22:57 manfred ----------
22:57 manfred Specified SLS gibberish in saltenv base is not available on the salt master
22:57 manfred [root@salt ~]# cat /srv/salt/test.sls
22:57 manfred include:
22:57 manfred - gibberish
22:57 forrest manfred, oh that must be somewhere else then, this error doesn't say that, sweet.
22:57 manfred Ryan_Lane1: yeah, '@basepi needs to be backported'
22:57 Ryan_Lane1 added :)
22:57 manfred forrest: right, that is what i was getting
22:57 forrest manfred, gotcha
22:58 forrest manfred, that's a lot better.
22:58 Ryan_Lane1 well, I hope that's all the backports I needed...
22:58 basepi Ryan_Lane1: did you see my reply to your request for backporting?
22:58 manfred forrest: https://github.com/saltstack/salt/blob/develop/salt/state.py#L2345
22:58 basepi at least one of them was on a commit in lyft/salt
22:58 basepi not in saltstack/salt
22:58 forrest manfred, sweet
22:58 basepi And any time you can, please comment on the pull request, not the commit itself.  I'd much rather backport whole pull requests over single commits where possible.
22:58 Ryan_Lane1 basepi: oh, whoops
22:59 manfred skyler: this is fixed in develop and 2014.7 https://github.com/saltstack/salt/blob/2014.7/salt/state.py#L2345
22:59 basepi =)
22:59 Ryan_Lane1 I meant to link against the PR
23:00 Ryan_Lane1 in general we don't keep anything in our fork that isn't upstreamed. there's one change that accidentally snuck in that we're not keeping
23:00 skyler manfred: awesome!
23:07 TOoSmOotH_ joined #salt
23:08 teepark joined #salt
23:09 stephenMW joined #salt
23:09 teepark I just got test failures on my pr but I'm not sure where to go in jenkins's ui to figure out what, if anything, I broke
23:10 stephenMW Hey all. Can anyone tell me if you can specify "source: s3://bucket/myfile" with file.managed using s3fs?
23:10 teepark https://github.com/saltstack/salt/pull/15127
23:10 stephenMW Also, can you change the port with s3fs?
23:11 Ryan_Lane1 stephenMW: are you using a master, or masterless?
23:11 stephenMW I'm using master
23:11 Ryan_Lane1 ah. ok
23:11 stephenMW (though I'd like to use masterless if it was supported)
23:11 * Ryan_Lane1 wishes the fs modules worked with masterless
23:11 stephenMW Also, I can't seem to be able to set the port value for s3. It defaults to 443 which makes it tough to debug using tcpdump
23:11 rojem joined #salt
23:12 stephenMW I can't find any examples online of using s3fs with file.managed. Can it even be done?
23:12 Ryan_Lane1 I'm pretty sure it's possible to change the port. I had seen a bug for this recently
23:13 danielbachhuber joined #salt
23:13 jalbretsen joined #salt
23:13 Ryan_Lane1 stephenMW: s3.verify_ssl
23:14 troyready joined #salt
23:14 Ryan_Lane1 oh. hm. that doesn't disable it
23:14 manfred stephenMW: so, if you have s3fs, are you dropping it in file_roots and pulling from salt:// ? or trying to acutally use s3:// ?
23:14 stephenMW I thought about running it through a proxy and breaking the SSL to inspect
23:14 stephenMW I'm trying to actually use s3://
23:15 Thiggy joined #salt
23:15 rawkode I'm confused. It looks like Salt supports IAM roles defined in cloud.providers, but id and key are required fields?
23:15 stephenMW I keep getting Unable to determine upstream hash of source file s3://my_folder/my_file
23:15 Ryan_Lane1 rawkode: you can likely use '' for them both
23:16 manfred stephenMW: did you see the note in the middle of this page? http://docs.saltstack.com/en/latest/ref/file_server/all/salt.fileserver.s3fs.html
23:16 manfred about requiring md5sum
23:16 Ryan_Lane1 stephenMW: when using the fs backend, you use salt:// I think
23:16 manfred ^^
23:16 Ryan_Lane1 and it does lookup in the buckets in a certain way
23:17 stephenMW I saw that but wasn't sure if it was something I needed to change
23:18 stephenMW If I change it from s3:// to salt:// I just get a plain old "Source file salt://my_folder/my_file not found"
23:18 stephenMW and no request to s3 captured on tcpdump
23:18 Ryan_Lane1 stephenMW: http://docs.saltstack.com/en/latest/ref/file_server/all/salt.fileserver.s3fs.html#module-salt.fileserver.s3fs
23:19 stephenMW My config can be found here (I reached out to the mailing list): https://groups.google.com/forum/#!topic/salt-users/qhpl_KmfKhA
23:19 Ryan_Lane1 you define the bucket in s3.buckets
23:19 Ryan_Lane1 ah
23:19 Ryan_Lane1 it should find files from the bucket...
23:20 Ryan_Lane1 oh
23:20 patrick joined #salt
23:20 Ryan_Lane1 salt://myfile == s3://test_bucket/myfile
23:20 Ryan_Lane1 or should be, anyway
23:20 stephenMW That's what I figured
23:21 rawkode Ryan_Lane1: Ahh, OK. I'll try that
23:21 Ryan_Lane1 I don't know for sure. I use the boto_* states and modules
23:23 stephenMW if I change it from s3:// to salt:// it doesn't even make a request to s3 :/
23:23 Ryan_Lane1 basepi: sorry for all the backport requests. I obviously missed the post to the list about the process change :(
23:24 stephenMW I guess it could be cached on the master so using salt:// works
23:24 rawkode Ryan_Lane1: I don't think I can use the boto_ states with the ec2-autoreactor
23:24 stephenMW without making a request to S3
23:25 Ryan_Lane1 rawkode: yeah. you can't. I looked into this recently
23:25 Hell_Lap joined #salt
23:25 Ryan_Lane1 it calls salt-cloud
23:26 rawkode Yep
23:26 rawkode and uses salt-bootstrap .. which has been fun to debug
23:26 Ryan_Lane1 it could be pretty simple to make a similar reactor that just made a call out to boto to verify the node
23:27 Ryan_Lane1 I install salt via the launch config in boto_asg/boto_lc
23:27 Ryan_Lane1 the launch config could also point the minion to the master
23:27 Ryan_Lane1 then it's just a matter of accepting the key
23:28 rawkode I ended up writing my own bootstrap
23:28 patrickcp Hi guys, does anyone know a good way to get the job output of a highstate that has been actioned by salt-reactor? Currently I am manually running 'salt-run jobs.lookup_jid' but I'd like this logged or emailed to me directly. Anyone have ideas?
23:28 rawkode Though I'll look into replacing bit by bit with the boto module
23:28 Ryan_Lane1 I'll likely eventually add a master for remote execution
23:28 Ryan_Lane1 I'll make a reactor then, I'd imagine
23:31 Ozack joined #salt
23:31 rawkode salt-cloud is just a bit rubbish, tbh. Why am I defining securitygroup in the provider? Should live in a profile. One of the many weird things I've noticed
23:33 bhosmer joined #salt
23:35 bhosmer_ joined #salt
23:38 patrickcp_ joined #salt
23:40 djaykay_ joined #salt
23:42 patrickcp_ Hi guys, does anyone know a good way to log the job output of a highstate that has been actioned by salt-reactor? Currently I am manually running 'salt-run jobs.lookup_jid' but I'd like this logged or emailed to me directly. Anyone have ideas?
23:42 bhosmer__ joined #salt
23:43 forrest patrickcp_, I was trying to think of something when you posted previously. Is your salt master a minion?
23:43 patrickcp_ sorry forrest, accidently closed the channel. (PEBKAC)
23:43 forrest patrickcp_, I was trying to think of something when you posted previously. Is your salt master a minion?
23:43 patrickcp_ Yes, the mastet also runs a minion.
23:44 SheetiS joined #salt
23:44 forrest patrickcp_, what about using your reactor to trigger a run on the minion which is just a cmd.run to take the output of 'list jobs' and grab the last N job entries, and email the output of those?
23:44 forrest it's ghetto, but it's all I could think of
23:45 blast_hardcheese joined #salt
23:45 patrickcp_ forrest: ah, cool. That would work :)
23:45 bhosmer joined #salt
23:45 martin_ joined #salt
23:45 JPaul joined #salt
23:45 forrest patrickcp_, yeah, it would be triggering a run on the 'master' which is a minion
23:46 forrest so it could work for what you want, I don't think there is anything built into salt currently to do this right now unfortunately
23:46 oncallsucks joined #salt
23:46 patrickcp_ forrest: I wasn't sure if there was something simple I was missing.
23:46 patrickcp_ forrest: Thanks for the idea, I'll give it a go.
23:46 forrest patrickcp_, not that I'm aware of. I haven't explored the reactor a ton however.
23:46 forrest patrickcp_, np
23:47 schimmy joined #salt
23:49 mackstic1 joined #salt
23:49 aquinas_ joined #salt
23:49 fivethreeo joined #salt
23:50 schristensen joined #salt
23:50 77CAAIOXN joined #salt
23:50 btorch joined #salt
23:51 davidone joined #salt
23:52 esogas joined #salt
23:52 herlo joined #salt
23:52 yomilk joined #salt
23:53 basepi Ryan_Lane1: hehe, no worries.  backports are still pretty easy.
23:53 Ryan_Lane1 cool :)
23:54 rlarkin|2 joined #salt
23:54 csa joined #salt
23:54 basepi Ryan_Lane1: plus, if they don't apply cleanly i get to ping you and ask you to do it!  ;)
23:55 Ryan_Lane1 :D
23:55 basepi or at least help me, make sure i don't introduce regressions.  =P
23:55 Ryan_Lane1 very possible
23:55 Ryan_Lane1 I think every single one of mine are boto related
23:55 Ryan_Lane1 and those are all new. so regressions are unlikely
23:55 tkharju1 joined #salt
23:56 kermit joined #salt
23:56 mgarfias joined #salt

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