Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-09-06

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

All times shown according to UTC.

Time Nick Message
00:03 teebes joined #salt
00:03 halfss joined #salt
00:05 VertigoRay What are standard labels for the maturity field? http://docs.saltstack.com/ref/modules/#add-module-metadata
00:09 UtahDave VertigoRay: I don't think we've standardized them yet
00:09 VertigoRay Thanks UtahDave
00:09 UtahDave you're welcome
00:10 VertigoRay Also, if you run `salt '*' config.get environment` from your master, do your minions return anything but None?
00:11 Ryan_Lane joined #salt
00:12 [diecast] joined #salt
00:13 emocakes joined #salt
00:14 chrisgilmerproj1 left #salt
00:16 mafrosis joined #salt
00:16 mafrosis lo everyone
00:16 mafrosis quick question, if someone could help me out
00:17 mafrosis when extending a state, are require lists merged, or overwritten? I would presume merged, but if they arent it would explain an error im getting
00:27 middleman_ joined #salt
00:28 Gifflen joined #salt
00:28 whit joined #salt
00:31 halfss joined #salt
00:36 middleman_ joined #salt
00:40 VertigoRay mafrosis: overwritten -- normal list functionality applies: http://docs.python.org/2/tutorial/datastructures.html
00:40 mafrosis VertigoRay: thanks
00:40 mafrosis that's not ideal
00:41 lineman60 joined #salt
00:42 fremder99 joined #salt
00:42 VertigoRay mafrosis: When you run `salt '*' config.get environment` from your master, do your minions return anything but None?
00:43 chrisgilmerproj joined #salt
00:44 mafrosis VertigoRay: I get None on this build
00:44 mafrosis FYI it's a minion only dev VM
00:44 VertigoRay Ahh, okay.  Thanks
00:44 mafrosis so I do, salt-call config.get environment
00:49 Kilsharion joined #salt
00:49 forrest joined #salt
00:51 oz_akan_ joined #salt
00:52 middleman_ joined #salt
00:53 bhosmer joined #salt
00:53 dthom91 joined #salt
00:54 druonysuse joined #salt
00:54 druonysuse joined #salt
01:02 oz_akan_ joined #salt
01:04 middleman_ joined #salt
01:05 Gifflen_ joined #salt
01:18 naemon joined #salt
01:25 naemon_ joined #salt
01:26 m_george left #salt
01:37 bdf so I'm trying to use a state to initialize postgres on freebsd. Is there a better way to do this? https://gist.github.com/billfehring/3f259e72ebf8f611e7ee
01:37 hannesstruss_ joined #salt
01:37 bdf I'm trying to make sure the service doesn't start unless initdb has run
01:38 bdf but it doesn't look like I can 'watch' or 'require' the cmd
01:40 iquaba joined #salt
01:40 iquaba any one here able to help me out with git?  trying to send in a pull but it's all messy
01:42 mafrosis iquaba: what's up with it?
01:44 iquaba i think i didn't merge the changes into my branch correctly.  when I go to pull it shows a lot of differences with salt's develop branch
01:44 iquaba but everything i try merge/pull/fetching on the command line says "up to date"
01:44 littleidea joined #salt
01:44 iquaba show my pull shows me commiting 19 changes on 75 files with 5 contributors
01:45 mafrosis so have you cretaed a new feature branch with your commits?
01:45 iquaba yep
01:45 mafrosis can I have a github link?
01:45 jbunting joined #salt
01:45 iquaba https://github.com/iquaba/salt/tree/ping_unresponsive
01:48 mafrosis iquaba: something weird happend on your develop branch - it's a bit out of date, but it also doesn't match the upstream one
01:48 mafrosis https://github.com/iquaba/salt/commits/develop
01:49 mafrosis yours and upstream have diverged from since 854a7987
01:50 mafrosis I would reset your develop branch to origin/upstream, so you're in sync with salt HEAD
01:50 mafrosis and then create a new branch from there
01:50 mafrosis cherry-pick your latest commits from the ping_unresponsive branch
01:51 iquaba can you walk me through it?
01:51 mafrosis sure
01:51 iquaba my git experience is basically following script and hoping i dont' break anything (like this) :P
01:51 iquaba thanks
01:52 mafrosis hehe fair enough
01:52 mafrosis is the only commit you're trying to save this one? 3812d7f9a
01:52 iquaba yep
01:54 mafrosis get onto the develop branch - git checkout develop
01:54 iquaba done
01:54 mafrosis git pull upstream develop
01:54 mafrosis oh actualy
01:54 mafrosis git fetch --all
01:54 iquaba done
01:55 iquaba 4bfc18e..5c28bf6  develop    -> upstream/develop
01:56 mafrosis so you're now going to hard reset your local develop to match that of the upstream
01:56 mafrosis make sure you dont have any unstashed or uncommited changes there - they will be deleted!
01:56 iquaba nothin'
01:56 mafrosis proably prudent to make a copy of the whole directory somewhere :_
01:57 iquaba 1 sec
01:58 iquaba done
01:59 mafrosis cool, now run this
01:59 mafrosis git reset --hard upstream/develop
01:59 iquaba HEAD is now at 5c28bf6 Merge pull request #7087 from basepi/7083
02:00 mafrosis cool
02:00 mafrosis git co -b ping_updated
02:00 mafrosis OR git checkout -b ping_updated
02:00 iquaba should my develop and their develop be completely even now?
02:00 iquaba https://github.com/iquaba/salt/compare/saltstack:develop...develop
02:01 iquaba is that still 18 commits different?
02:01 mafrosis you've only updated your local repo
02:01 iquaba gotcha
02:01 mafrosis on the develop branch, run
02:01 mafrosis git push -f
02:02 mafrosis to forcibly update the github version of your develop branch with the changes you just made with git reset
02:02 iquaba more pretty
02:02 iquaba ok, on the ping branch
02:02 mafrosis sweet, I can see you're in sync again now :)
02:02 mafrosis did you create a new ping branch from develop?
02:03 mafrosis git checkout -b ?
02:03 iquaba ahh, gotcha my bad
02:03 iquaba i'm in the new branch
02:03 iquaba 'show_timeout'
02:04 mafrosis okay, so now is the bit that could go wrong :)
02:04 mafrosis git cherry-pick 3812d7f9a
02:04 iquaba error: could not apply 3812d7f... Adds --show-timeout to force the display of unresponsive minions
02:05 mafrosis which plucks your commit from out of the git ether, and trys to apply it to your current HEAD
02:05 mafrosis so now you need to figure out how to fit your changes in with what's happened in salt since you last updated
02:05 mafrosis which is about a month's worth of commits....
02:06 iquaba alright, i can do this
02:06 iquaba i've still got the old window open with the old comparisons, and it wasn't too many lines
02:06 mafrosis cool
02:06 iquaba i thiink i got it from here, thanks a lot!
02:07 mafrosis you can check the diff on github too - I find that helps
02:07 mafrosis iquaba you're welcome!
02:10 iquaba what do i do with this random 'untracked file'?
02:11 iquaba tests/integration/files/conf/grains
02:12 mafrosis err I dont know what file you mean :p
02:13 iquaba in git status I've now got my "changes to be committed" but also "untracked files"
02:13 iquaba # Changes to be committed:
02:13 iquaba #
02:13 iquaba #modified:   salt/cli/__init__.py
02:13 iquaba #modified:   salt/client/__init__.py
02:13 iquaba #modified:   salt/utils/parsers.py
02:13 iquaba #
02:13 iquaba # Untracked files:
02:13 iquaba #   (use "git add <file>..." to include in what will be committed)
02:13 iquaba #
02:13 iquaba #tests/integration/files/conf/grains
02:13 iquaba eww, sorry
02:14 sssslang does anyone known the meaning of this option: ;; then enter the text in that file's own buffer.
02:14 sssslang does anyone known the meaning of this option: 'publish_session' ?
02:16 mafrosis iquaba sorry man, I don't know what that is
02:16 mafrosis dont worry though - if they arent your changes / not needed for your patch then ditch it
02:17 mafrosis but maybe take a backup in case you realise what they are later ;)
02:17 iquaba i don't see them in salt's repo, i'm gonna go ahead and say its not going to be a big deal
02:21 jacksontj joined #salt
02:22 lemao joined #salt
02:23 swa_work joined #salt
02:25 mafrosis left #salt
02:26 middleman_ joined #salt
02:33 emocakes joined #salt
02:34 liuyq joined #salt
02:35 liuyq joined #salt
02:38 dml joined #salt
02:39 dml joined #salt
02:39 xl1 joined #salt
02:39 bdf how can I require a service be running in a state before executing a command in the same state?
02:40 dml joined #salt
02:41 dml joined #salt
02:52 sssslang got confused between pillar and grains :-/
02:53 littleidea joined #salt
02:57 swa_work bdf, I'm working through the salt _way_  myself. https://github.com/saltstack/salt-states/blob/master/http/init.sls may help you
03:01 pdayton joined #salt
03:01 swa_work this would most likely be relevant as well bdf http://docs.saltstack.com/ref/states/ordering.html
03:02 jacksontj joined #salt
03:02 Jahkeup joined #salt
03:04 bdf swa_work: well, what led to it is this: https://gist.github.com/billfehring/3f259e72ebf8f611e7ee
03:05 bdf because if I don't initdb first, postgres doesn't start.
03:06 swa_work yah
03:07 bdf so I want it to understand that if it goes to start the service, and the file doesn't exist (the initdb has not run) run the command, before trying to start the servicer
03:07 bdf -r
03:07 Furao swa_work: ordering is a bad idea, require + watch is the way to go
03:07 Furao bdf: require, required_in, watch, watch_in
03:07 Kilsharion joined #salt
03:09 swa_work thanks Furao .. so ordering is for what then though?
03:09 bdf if I cmd.wait, and my watched state is a service does the command run before the service is started?
03:10 bdf when I tried that it appeared that it tried to init the db after trying to start the service
03:10 Furao swa_work: I only found ordering useful (but still unsafe) in a salt states testing framework I created. to don't have testing process executed before all other states are executed (ordering: $bignumber)
03:11 Furao bdf: watch require the output of the specified requisite to perform, so yes it will execute before
03:11 swa_work got ya Furao
03:11 bdf hmm
03:11 oz_akan_ joined #salt
03:11 Katafalkas joined #salt
03:12 * swa_work needs to codify a 'rosetta stone'
03:12 chrisgilmerproj joined #salt
03:12 pdayton joined #salt
03:15 jpeach joined #salt
03:15 mwillhite joined #salt
03:18 mafrosis joined #salt
03:19 vbabiy Hey guys how can I use the salt.pillar data in the vagrant plugin? When the vm boots and I run salt-call --local  pillar.items I don't see it
03:19 swa_work Furao, have some good links (that _aren't_ ) the docs? So far I've spent more time digging through functions and lower level and the 'tapestry' is kinda frayed (i.e. I can code a function in fabric and py quicker and it works)
03:20 swa_work best practice and gotchas that are current high/low maybe
03:20 swa_work brb
03:21 Furao swa_work: all I got is example of things that used to works, I still haven't seen complex states example elsewhere: https://github.com/bclermont/states/tree/master/states
03:21 Furao I somewhat closed source that 6 months ago
03:22 Furao today it don't look like that anymore :)
03:22 swa_work hear ya Furao .. thanks. The saltstack-formulas has been my 'goto' mostly
03:23 swa_work thanks btw Furao
03:23 Furao formulas, when they works, are still very simple
03:23 Furao these are real life scenario trying to do, sometimes, complex things
03:23 chrisgilmerproj left #salt
03:23 Furao today I have 3 times more states than that
03:24 swa_work yah, but foundation is the key
03:24 Furao well, I hired few guys working full time on this
03:28 swa_work I was SME puppet/virt for an HP ENT. SVC contract but puppet has gone (much like zimbra and opsview) in a direction not conducive to my earlier recommendations/implementations
03:32 vbabiy Has anyone used salt-vagrant and pillar data?
03:34 swa_work I can architect, config, deploy quicker w/ what _I_ know .. but that doesn't leverage what's been done, maintenance, security et al of a solution that I can _mostly_ stick w/ best practice and is active and open
03:35 bdf hmm... something else was masking it, but it turns out that require_in works very well for my case, thanks Furao and swa_work
03:37 swa_work np bdf .. if you b0rk something (QA/test) in that; does a subsequent run fix it ?
03:42 jacksontj joined #salt
03:45 halfss joined #salt
03:47 whit joined #salt
03:54 middleman_ joined #salt
03:54 swa_work bdf, all that in the init.sls ?
03:55 wolfpack_ joined #salt
04:05 mwillhite joined #salt
04:09 halfss joined #salt
04:11 Katafalkas joined #salt
04:17 jwon joined #salt
04:17 [vaelen] joined #salt
04:17 faulkner joined #salt
04:22 halfss joined #salt
04:28 sparc joined #salt
04:28 sparc so many people here!
04:28 sparc really great
04:29 bhosmer joined #salt
04:32 Furao 354
04:33 Furao I remember when it was < 80
04:47 emocakes joined #salt
04:49 jergerber joined #salt
04:56 swa_work Furao, I'm working out a returner (mysql) .. any tips/gotchas?
04:56 Furao swa_work: don't use mysql
04:56 swa_work thanks
04:56 Furao it's not a real database http://www.youtube.com/watch?v=1PoFIohBSM4
04:57 swa_work my clients won't buy that though
04:57 Furao i think there is already few mysql returner around, if you don't mind data loss
04:57 Furao i wrote a sentry returner, you can look at salt.returners for example
04:58 swa_work well yeah I most certainly do mind data loss .. ah .. in the link you provided earlier yah Furao ?
04:58 Furao you should look that youtube video
04:59 swa_work thanks, not a vid fan
04:59 Furao https://github.com/saltstack/salt/tree/develop/salt/returners
05:01 swa_work thanks Furao
05:01 halfss joined #salt
05:04 swa_work get that implemented POC then agnostic for db = {oracle, postgres} and I have momentum
05:04 jinrou joined #salt
05:12 Furao there is already a mysql returner https://github.com/saltstack/salt/blob/develop/salt/returners/mysql.py
05:12 Katafalkas joined #salt
05:13 dthom91 joined #salt
05:15 swa_work yah Furao ... thanks. I'm working on a 'dynamic' type solution ... one size fits all kinda thing
05:16 swa_work salt so far seems to be the likely candidate for the cfg. mgmt./orchestration so far
05:18 pass_by_value joined #salt
05:20 swa_work of course grains (facter like anyway)
05:20 Ryan_Lane joined #salt
05:22 swa_work for now I use facter and that's better anyway since it bridges migration from existing 'brown fields' which are  puppetized
05:23 swa_work though I don't want that as a dep but as a feature forward
05:24 swa_work I'm mostly sysadmin w/ a dollop of dev
05:26 gildegoma joined #salt
05:33 forrest Is the only way to test documentation updates for formatting to build them with sphinx? Or is there a quicker way that sets it up when travis CI runs.
05:34 maspwr joined #salt
05:47 NV anyone know if it's possible, and if so, how i would go about, using salt.utils.minions.CkMinions() in a python pillar sls?
05:47 NV I want to be able to use the same glob matching in it
05:47 NV or alternatively, make a renderer that uses it
05:50 jheintz joined #salt
05:52 tuxIO joined #salt
05:54 jheintz_ joined #salt
05:55 jheintz_ newbie question. I've got an SLS file with a cmd.run script in it, and I'm triggering this with state.highstate. First time works great... but the script gets run every time I run highstate. How do I change things to only run the script once (until I change it).
05:57 forrest hey jheintz_, did you look at this: http://docs.saltstack.com/ref/states/all/salt.states.cmd.html#execution-of-arbitrary-commands
05:57 forrest there's a couple options, you can use - unless:
05:59 forrest You could do a jinja2 if else loop to only check if the md5sum doesn't match
06:00 dthom91 joined #salt
06:02 Furao jheintz: cmd.wait & use - watch
06:03 jheintz_ forrest, thanks for the link. unless might be what I need if I can detect if the command has already run.
06:04 jheintz_ Furao, what would I watch? I understand that can be a "wait for it" trigger, but I'm not sure what I'd actually wait on.
06:05 forrest so jheintz_
06:05 jheintz_ forrest, the jinja2 script checking md5sums... that is what I was expecting the cmd.run to do.
06:05 forrest yea it doesn't :P
06:05 forrest http://docs.saltstack.com/topics/tutorials/states_pt2.html#next-steps
06:05 forrest don't look at next steps
06:05 forrest just above it
06:05 forrest so what Furao is suggesting is to manage the file via saltstack (the actual script), then what will happen is you do your cmd.wait, with a watch that checks that file
06:06 forrest so if the file changes, then it updates
06:06 forrest correct me if I misunderstood what you were getting at Furao
06:07 jheintz_ ah, that makes sense. manage the script as a file. Then the cmd.wait would watch that file, and also execute it?
06:07 forrest it would only execute if the file changed (or on initial run)
06:07 forrest but yes, cmd.wait would execute it
06:07 Furao forrest: yes, I do that all the time
06:07 Furao such as have requirements.txt managed by salt and run pip install when file get changed
06:08 forrest yup
06:08 Furao https://github.com/bclermont/states/blob/master/states/sentry/init.sls#L37
06:11 jheintz_ so at it's simplest I would have a state with file: source, and a cmd:wait with the name: <file> and watch: -file: <file> both pointing to the same thing.
06:12 Katafalkas joined #salt
06:13 jheintz_ thanks forrest and Furao. I'll give this a try tomorrow. cheers
06:13 forrest np, later
06:19 xiaopi[z] left #salt
06:20 goki has anyone heard of plans to support for Android?  If so, any idea on a timeline?
06:27 joehh do you mean android proper or running in a chroot on an andoird device?
06:30 MrTango joined #salt
06:31 sssslang hi, is it possible to assign a role to a minion from master? using grains?
06:32 forrest can you expand on your question sssslang?
06:33 forrest as in if the hostname = xyz, then appy certain states?
06:33 forrest *apply
06:34 sssslang forrest: I want to group minions by roles. give some hosts a role name.
06:34 forrest Did you check this out? http://docs.saltstack.com/topics/tutorials/states_pt1.html#preparing-the-top-file
06:35 forrest shows how to target a minion based on the os in that exampe
06:35 forrest *example
06:35 forrest http://docs.saltstack.com/ref/states/top.html also has a great example of matching based on hostname about 20% down the page
06:36 sssslang forrest: thank you. It's about target. But using current grains may not work.
06:36 forrest you should check the grains first, try salt 'machine_name' grains.items
06:36 forrest that will list all available grains for a system
06:37 hannesstruss joined #salt
06:38 sssslang forrest: I've checked the grains. I think I should define a grain item called role. But I don't know is there a way to set it on the master, then transfer to every minion.
06:39 forrest check out http://docs.saltstack.com/topics/targeting/grains.html#writing-grains
06:39 forrest so can you not base the role off the hostname, or ip range?
06:40 forrest I can't remember but I'm pretty sure there's a way to actually group sets of machines in saltstack
06:40 sssslang not exactly. there are hundreds of hosts.
06:41 forrest I see
06:41 sssslang Node groups?
06:41 forrest yea
06:41 forrest no dice on that for you either?
06:42 forrest I mean it would be a pain to do the first time if you have hundreds of boxes, but you could run a job across them all to return the hostname and just drop it into a file and format it
06:43 sssslang I know I can manually set a role in minion's config, but I think it's hard to manage later. So I want to put them together in master's config.
06:44 forrest well nodegroups lives on the master
06:44 forrest but yea I see where you are coming from
06:44 forrest just defining some sort of custom grain is probably your best bet, it still needs to get set when you build the box though
06:45 goki joehh: I want to run a minion on an Android device so I can manage it.
06:46 sssslang you mean define some functions run on minions and return it's role based on hostname as grain item?
06:46 aantony joined #salt
06:47 welby if I have a cmd.run state, is there a way that I can disable this from running unless called via a watch_in. I've got this being done with creating a file and requring that file to exist, but this isnt the cleanest solution, just wondering if theres another way of doing it
06:47 forrest well if you are going to do that sssslang you might as well just use the hostname in the top file to determine what states are applied
06:50 sfello joined #salt
06:50 linjan_ joined #salt
06:51 halfss joined #salt
06:53 joehh goki: I haven't seen anything since the mailing list threads back in may
06:53 joehh but they quickly moved off to ubuntu touch
06:53 forrest I've gotta go sssslang, let me know if you find a solution, or come up with a cool custom grain.
06:54 sssslang forrest: thank you. I'll tell you the result.
06:54 forrest sounds good!
06:54 goki joehh: OK...I was digging for a while and found a thread from a a ways back...but then there was nothing.
06:55 ollins joined #salt
06:55 joehh I think that is still the current status
06:55 maspwr joined #salt
06:55 joehh if you can get something that can read ubuntu or debian apckages, then that would work,
06:55 joehh but that would generally be within a chroot rather than native android
06:56 goki joehh: thanks
06:58 sfello joined #salt
06:59 balboah_ joined #salt
07:05 halfss hi, i think salt should make a way to get sls list
07:08 carlos joined #salt
07:13 Katafalkas joined #salt
07:13 rhoml joined #salt
07:15 pass_by_value joined #salt
07:21 vaxholm joined #salt
07:33 unicoletti_ joined #salt
07:39 malinoff joined #salt
07:45 malinoff Hi all
07:47 ollins joined #salt
07:51 mastrolinux joined #salt
07:53 halfss joined #salt
08:00 spacedentist_ joined #salt
08:00 backjlack joined #salt
08:06 bhosmer joined #salt
08:13 halfss joined #salt
08:14 tomviner joined #salt
08:16 ggoZ joined #salt
08:18 ronc joined #salt
08:18 dthom91 joined #salt
08:25 Katafalkas joined #salt
08:26 az87c joined #salt
08:32 ronc joined #salt
08:32 nocturn I have a problem with salt hanging on state.highstate on CentOS 6 machines
08:33 nocturn salt '*' state.running reports: The function "state.highstate" is running as PID xxx and was started at 2013, Sep 05 16:49:30.883503 with jid xxx
08:33 nocturn On that node, I have two salt-minion processes running
08:35 nocturn This happens often and randomly.  Killing the PID and rerunning highstate will get this problem about 50% of the time
08:35 nocturn I have a second installation of salt (other env) on Ubuntu and it never has this
08:38 unicoletti_ left #salt
08:42 nkuttler joined #salt
08:46 halfss joined #salt
08:46 sfello hey all
08:48 sfello having a bit of trouble with the require statement - in that it seems to ignore the order (one cmd.run requires a cmd.script to run first, but the cmd.run always fires first even though it has set require: - cmd.script: other-thing)
08:48 cwarner joined #salt
08:51 sssslang hi there, I cannot using pillar.get to retrive pillar data but pillar.item works fine. what could be wrong?
08:51 Furao sfello: there is no order, use - require everywhere it make sense
08:53 jinrou joined #salt
08:54 felixhummel joined #salt
09:09 xpen joined #salt
09:12 scott_w joined #salt
09:12 gildegoma joined #salt
09:16 xpen joined #salt
09:16 mastrolinux joined #salt
09:18 mastrolinux @Furao that is not true:
09:18 linjan joined #salt
09:18 mastrolinux But in Salt 0.17.0, the state_auto_order option was added. This option makes states get evaluated in the order in which they are defined in sls files.
09:18 mastrolinux from: http://docs.saltstack.com/ref/states/ordering.html
09:18 mastrolinux you can have order as written in state
09:20 Furao mastrolinux: what's not true?
09:20 tuxIO joined #salt
09:20 mastrolinux that there is no ordering
09:20 mastrolinux in the new version ordering is supported
09:20 Furao is that guy on 0.17
09:20 Furao ?
09:21 mastrolinux maybe, I do not know, just linked the doc
09:21 mastrolinux in old version is possible to add a priority value too
09:21 mastrolinux having explicit ordering
09:21 Furao 0.17 is not released
09:22 xt salt .0 releases - best releases
09:23 Furao I still didn't worked around all 0.16.x issues
09:23 mastrolinux ok, so is the official doc not true? exposing something still not available?
09:24 gmoro joined #salt
09:25 mastrolinux in Salt 0.17.0, the state_auto_order option *was added* to Salt. Sounds to me as something in the past
09:25 xt mastrolinux: nah, that's regular english phrasing
09:25 xt was added in 0.17.0
09:26 mastrolinux ok, sorry, a bit confusing for not native english speakers
09:26 xt I'm not native :-)
09:26 unicoletti joined #salt
09:28 mastrolinux I supposed it was present perfect
09:28 xmj is 0.17.0 out?
09:28 mastrolinux no is not!
09:29 Furao anyway wait to 0.17.3 before trying 0.17.x
09:29 ggoZ `wget -O - http://bootstrap.saltstack.org | sudo sh` is stuck on the 'Saving to: `STDOUT'' part
09:29 ggoZ :<
09:29 Furao bunch of grains key will probably be renamed again
09:38 TheCodeAssassin joined #salt
09:44 cowmix joined #salt
09:46 gmoro joined #salt
09:46 cowmix joined #salt
09:52 giantlock joined #salt
09:53 carlos joined #salt
10:02 vmdsch joined #salt
10:09 MaxK joined #salt
10:10 nocturn I don't fully understand what lowstate does, or what it is intended for...
10:20 bejer joined #salt
10:21 Furao I think tomas did a blog post about that this week
10:21 Furao thomas
10:29 sfello folks - given a minion that wants to be able to talk to another minion (and hence needs its IP address) - is there any way to pass this info from one minion to another? I think long term this would be simpler with DNS but just trying to figure out what is feasible
10:31 fredvd joined #salt
10:33 lemao joined #salt
10:36 xl1 joined #salt
10:41 Furao sfello: salt mine
10:42 sfello Furao: can you elaborate? Sorry - am a bit new to salt.
10:42 sfello i think Googling that might come up with some random!
10:43 Furao https://salt.readthedocs.org/en/latest/topics/mine/index.html
10:43 sfello ah!
10:43 sfello Furao: badass. Thanks!
10:48 yml sfello: I am unclear how you would query that
10:49 yml let say that you need on a minion all the ip address of the web servers ?
10:49 xl1 joined #salt
10:50 sfello yml: yeah that's kind of what i need
10:50 yml http://docs.saltstack.com/ref/modules/all/salt.modules.mine.html
10:50 sfello it's actually for database replication, but much the same
10:51 yml sfello: i think I m am on it there is a .get()
10:51 unicoletti joined #salt
10:52 jslatts joined #salt
10:52 rich42 joined #salt
10:52 sfello so the process would presumably be get from all minions some network interface address, then send that out to all minions… but then run highstate again to get it to e.g. write those to the hosts file?
10:53 jinrou joined #salt
10:54 yml sfello: i think that if I understand it properly
10:54 yml the mine function is run periodically
10:54 yml (like a cron job)
10:54 yml and store value in a db (no idea what exactly)
10:55 yml where you can query it in your states calling the mine.get() function to fetch value that you need for you configuration
10:56 * yml is making all this up as I read the doc
10:56 sfello right, me too...
11:01 Lue_4911 joined #salt
11:06 halfss joined #salt
11:08 montuori left #salt
11:08 Koma joined #salt
11:12 Lue_4911 joined #salt
11:23 jinrou joined #salt
11:24 blee joined #salt
11:25 bhosmer joined #salt
11:26 gmoro joined #salt
11:28 ollins joined #salt
11:33 mwillhite joined #salt
11:39 Lue_4911 joined #salt
11:42 jbunting joined #salt
11:42 p3rror joined #salt
11:48 Katafalkas joined #salt
11:48 jinrou joined #salt
11:51 emocakes_ joined #salt
11:53 hotbox joined #salt
11:54 dthom91 joined #salt
11:59 blee_ joined #salt
11:59 Katafalk_ joined #salt
12:04 anteaya joined #salt
12:06 [diecast] joined #salt
12:06 oz_akan_ joined #salt
12:16 jslatts joined #salt
12:40 jinrou joined #salt
12:50 littleidea joined #salt
12:59 vaxholm joined #salt
13:03 juicer2 joined #salt
13:05 pdayton joined #salt
13:09 brianhicks joined #salt
13:09 racooper joined #salt
13:09 teebes joined #salt
13:10 oz_akan_ joined #salt
13:12 oz_akan_ joined #salt
13:12 jinrou joined #salt
13:12 Jahkeup joined #salt
13:16 vaxholm joined #salt
13:21 m_george|away joined #salt
13:28 tuxIO_ joined #salt
13:32 toastedpenguin joined #salt
13:38 blee joined #salt
13:38 imaginarysteve joined #salt
13:39 Katafalkas joined #salt
13:42 hellome joined #salt
13:42 jinrou joined #salt
13:54 lineman60 joined #salt
13:55 kaptk2 joined #salt
13:59 Eugene blee - are you poking about with salt-cloud?
13:59 m_george left #salt
14:01 littleidea joined #salt
14:01 bhosmer_ joined #salt
14:02 Gifflen joined #salt
14:03 entil left #salt
14:03 Koma joined #salt
14:05 oz_akan_ joined #salt
14:05 brutasse joined #salt
14:06 zooz joined #salt
14:07 Gifflen joined #salt
14:07 Khollowa_ joined #salt
14:09 mannyt joined #salt
14:10 balboah_ joined #salt
14:10 Lue_4911 joined #salt
14:12 jbunting joined #salt
14:12 pass_by_value left #salt
14:12 alunduil joined #salt
14:13 lemao joined #salt
14:20 JasonSwindle joined #salt
14:20 JasonSwindle Howdy, all
14:21 timoguin joined #salt
14:21 timoguin are there plans to add gitfs support for pillar data?
14:24 MTecknology joined #salt
14:25 ccase Wrote a guide to get setup quickly with salt and windows: https://github.com/calebcase/zero-to-7zip
14:25 ccase If anyone wants to try that out and open some bugs for bad directions I would be very appreciative :o)
14:26 timoguin ccase, thanks. i'll be trying that soon.
14:27 MTecknology joined #salt
14:35 berto- joined #salt
14:37 whyzgeek hi I am newbie anybody has got ready made users modules that supports large scale account, group and key based authentication management to share?
14:39 joe______ joined #salt
14:40 dthom91 joined #salt
14:40 msil joined #salt
14:40 jmoretti joined #salt
14:41 msil hey guys, i seem to be having a problem adding a minion to my setup
14:41 msil I'm using freebsd jails - bit they've got private IP addresses, as the host only has one public IP
14:41 msil *but
14:42 msil when i try to do key auth, i get the "master has rejected the key" message
14:42 sfello anyone know of any detailed uses of salt.modules.mine? the docs are pretty shaky.
14:42 msil even though I see no rejected keys on the master
14:42 sfello *tutorials i mean
14:42 msil the jail IPs are on lo1 on the master, would that be an issue?
14:42 pman joined #salt
14:43 cedwards msil: I run my BSD systems in the same way. loopback on lo1 and 10.0.0.0/24 address range.
14:43 msil cedwards: you've had no issues?
14:44 cedwards msil: not due to that type of setup, no.
14:44 msil and do you communicate with the master over the private subnet?
14:44 jinrou joined #salt
14:44 cedwards msil: my master is assigned 10.0.0.1
14:45 cedwards msil: sounds like it's key-specific and not anything to do with the jail/network
14:45 jmoretti joined #salt
14:45 cedwards msil: have you tried perhaps deleting and regenerating the offending keys on minion and master?
14:45 holmb msil: Try to wipe out all the keys and start the minions, they will create new keys.
14:46 xmj msil: I run ezjails on lo0 with master as 127.0.0.1 and jails .20-.22, works for me :-)
14:46 vaxholm joined #salt
14:46 * cedwards is glad to see more and more FreeBSD users in here. :)
14:46 xmj msil: even after moving the jails from 192.168.0.0/24 to 127.0.0.0/24 it kept working.
14:47 xmj msil: are you sure the master listens to the right ip on the right interface ?
14:47 cedwards sounds like just a key conflict to me. if it's saying the master has rejected the key, then i'd assume communication is working
14:50 msil xmj: the jails talk to the master over the public address at the moment
14:50 msil and I haven't had issues with that before
14:50 xmj weird
14:50 msil yup
14:50 msil i'll try and get the to talk over loopback and see if that works
14:50 JasonSwindle Is there a way to have the minion's log file in JSON?
14:51 JasonSwindle I ask because I want to ship it to redis and then into logstash
14:52 xmj msil: what did you put in /usr/local/etc/salt/minion on the jails?
14:53 JasonSwindle I looked; but it seems to be something Salt is lacking in the minion config
14:53 msil xmj: just the dfault configds for the moment, so it talks to the master over the public ip
14:54 xmj ok, what do you have in /usr/local/etc/ezjail/* concerning the interfaces ?
14:54 juicer2 Hi, anyone know what "NameError: global name 'collections' is not defined" when trying to install an rpm manually with salt?
14:54 juicer2 http://pastebin.com/Djkgd0g5
14:55 msil xmj: the jail has an ip on the 10.10.100.0/24 subnet
14:55 msil and the master is 10.10.100.1
14:55 juicer2 been working this for a few days, seems like this should work, and I know it has worked in the past
14:55 jpeach joined #salt
14:56 xmj err
14:56 xmj msil: i wonder, how do the jails see the public network if they're configured to use lo1?
14:56 msil xmj: yes they do, configured in my pf.conf
14:57 xmj weird
14:57 cedwards msil: do you have other minions that are working?
14:57 tyler-baker joined #salt
14:57 msil cedwards: yes
14:57 msil ah
14:57 msil something odd
14:57 msil the master logs show that it's accpted authentication from itself
14:58 msil when the minion starts
14:58 swa_work joined #salt
14:59 msil aha!
14:59 msil changed the id and fqdn in the minion conf
14:59 msil now the master sees a new minion
14:59 cedwards that'd do it
14:59 chrisgilmerproj joined #salt
14:59 subway left #salt
14:59 msil i assume because the jail's IP is an alias on lo1 on the master
15:00 msil the master auused it was a local connection
15:00 Jahkeup joined #salt
15:00 msil *assumed
15:00 xmj what fqdn did you need to set it to?
15:00 msil hmm
15:00 xmj I wanna be prepared when I run into this over the weekend :)
15:00 msil hostname is set in rc.conf though
15:00 msil just the fqdn of my network
15:01 msil wouldn't it be better if it go the id/fqdn from the system's hostname?
15:01 msil this way, i'll need to edit the minion files for each minion
15:02 olevchyk joined #salt
15:03 xmj noidea
15:03 xmj msil: where do you use salt/freebsd? some production deployment?
15:03 msil no, just screwing around
15:04 msil use it to run mail/web/etc
15:04 msil and #bsidefm, a radio station
15:05 xmj heh, i still need to convince my team to use salt over puppet
15:05 msil used puppet for a while with my last job
15:05 msil salt is a breth of fresh air
15:05 msil *breath
15:05 Katafalkas joined #salt
15:05 UtahDave joined #salt
15:06 ckao joined #salt
15:07 QauntumRiff joined #salt
15:07 QauntumRiff "salt '*' pkg.upgrade salt-minion"  Does that upgrade all packages, or just the salt-minion?
15:10 UtahDave QauntumRiff: looks like on both apt and yum based systems  pkg.upgrade does a full system upgrade
15:10 QauntumRiff damn.
15:11 racooper I think if you want to upgrade a single package you have to use pkg.install, at least in my experience. which is dangerous if you don't want to put the package on '*' if it doesn't already exist.
15:11 QauntumRiff how can you just upgrade the single one?
15:11 QauntumRiff can I cancel a job?
15:12 msil QauntumRiff: http://docs.saltstack.com/topics/jobs/index.html
15:12 msil saltutil.kill_job
15:12 racooper QauntumRiff,  pkg.install but be selective on which minions you're targeting.
15:15 QauntumRiff well crap.. I think I just ran "yum update" on all my production systems.. wee
15:15 QauntumRiff was following this site: http://intothesaltmine.org/blog/html/2013/03/04/upgrade_saltstack_installation.html
15:18 forrest joined #salt
15:18 UtahDave QauntumRiff: http://docs.saltstack.com/ref/modules/all/salt.modules.yumpkg.html#salt.modules.yumpkg.upgrade
15:18 StDiluted joined #salt
15:19 cedwards I will update my post
15:21 JasonSwindle UtahDave:  Howdy!
15:21 KennethWilke joined #salt
15:21 UtahDave hey, JasonSwindle!
15:21 juicer2 Found fix for my issue with installing rpm's on centos6,
15:21 KennethWilke g'mornin!
15:21 juicer2 if anyone else has this issue, https://github.com/saltstack/salt/issues/6959
15:22 JasonSwindle UtahDave:  Can the minion's log be created right into JSON?  I ask because I am trying to ship all of my logs into Redis for Logstash
15:22 juicer2 it's a one line fix in yumpkg5.py that can be patched with salt if you need to. At least that was the quick fix for me.
15:22 forrest juicer2 did you pull down the develop branch to see if https://github.com/saltstack/salt/issues/6677 fixed it?
15:23 UtahDave juicer2: have you sent in a pull req?
15:23 juicer2 forrest: I like to use rpms for install
15:23 forrest basepi believes it is fixed in 6677 UtahDave
15:23 UtahDave JasonSwindle: I'm not sure.  s0undt3ch is the master of Salt logging. He might now.   I think he just updated the logging to be able to send it to external logging systems like Logstash
15:23 forrest juicer2, totall understandable
15:23 forrest *totally
15:23 juicer2 UtahDave, fix is already in place, just not released yet, will be in 0.16.4
15:24 basepi Someone mentioned me?  =P
15:24 s0undt3ch JasonSwindle: 0.17 will have built in logstash support
15:24 s0undt3ch basepi: nah!
15:24 JasonSwindle s0undt3ch: Awesome!
15:24 s0undt3ch basepi: messing with you.... ;)
15:24 basepi s0undt3ch: ;)
15:24 juicer2 s0undt3ch: that's pretty sweet news
15:25 s0undt3ch juicer2: and sentry
15:25 s0undt3ch juicer2: and it will be easy to add more
15:25 basepi forrest: if that fix wasn't in 0.16.3, it will be in 0.16.4
15:25 basepi can't remember for sure
15:25 JasonSwindle s0undt3ch:  Will there be a ship to redis option?
15:25 forrest yea it's not in 0.16.3, juicer2 is having problems with it
15:25 forrest but he has a temp fix
15:26 basepi cool
15:26 backjlack joined #salt
15:26 juicer2 forrest: any idea on when 0.16.4 rpms will be released ?
15:26 JasonSwindle My flow is; log shipper -> redis -> LogIndexer / ES / LS -> Kibana
15:27 forrest juicer2, if they aren't in the test repo it's gonna be a while
15:27 juicer2 forrest: gotcha
15:27 basepi juicer2: hopefully will be on PyPI today
15:27 s0undt3ch JasonSwindle: https://github.com/saltstack/salt/blob/develop/salt/log/handlers/logstash_mod.py <- you'll be reposible to properly configure logstash
15:27 forrest herlo has to build them, then drop them into epel testing, then if no one uses their karma or whatever to push it through, it takes about 3 weeks.
15:27 basepi juicer2: then a few days for packagers to get them up
15:27 JasonSwindle s0undt3ch:  looking
15:27 s0undt3ch JasonSwindle: salt just delivers to logstash
15:27 forrest We need to have a collective pool of building RPMs for fedora so we can store up a bunch of karma to push releases through
15:27 basepi forrest: luckily terminalmage can build for epel now, so we won't have to wait for herlo.  but we still have to wait for the epel delay
15:28 forrest Yea herlo is quick, I wasn't trying to say he wasn't
15:28 forrest it's the delay from testing -> release that kills it.
15:28 basepi yep
15:28 terminalmage forrest: yeah he was on vacation last time we released
15:28 xmj msil: do you use ipfw or pf?
15:28 forrest terminalmage, gotcha.
15:29 forrest terminalmage, did you finally recover? :P
15:30 devinus joined #salt
15:30 terminalmage forrest: yep
15:31 terminalmage forrest: I'm going to make that demo even better and submit it to pycon
15:31 forrest better hurry
15:31 forrest only 9 days left
15:31 terminalmage yeah
15:31 forrest They emailed me regarding the deadline since I submitted last year, apparently they have even more submissions than last year
15:31 terminalmage I just have to write it up in 9 days
15:32 mannyt joined #salt
15:32 forrest That shouldn't be too bad, they don't want you to get supe in-depth on the initial submission
15:32 forrest *super
15:33 terminalmage but basically, it'll be an intermediate-level talk assuming you know basic salt usage (the djangocon talk went over grains, pillar, states, etc) which will show how you can deploy a django app with any db backend you want
15:33 terminalmage and possibly other kinds of apps as well
15:33 terminalmage so you have your different db backends, and you have an SLS file for each
15:34 forrest Yep
15:34 terminalmage which DB backend you choose, you store in pillar, then you name your SLS accordingly
15:34 forrest Were you able to find time to update that note about the pillar data?
15:34 terminalmage oh, no... was working on a couple bugs
15:34 forrest That would be a good example to see. Honestly the repo itself is the most important thing to me, because it's a full blown project.
15:34 terminalmage completely forgot
15:35 terminalmage thanks for the reminder
15:35 forrest It's a lot easier for me to look at an example of how things work when it's moderately comple
15:35 forrest *x
15:35 forrest yea np
15:36 JasonSwindle s0undt3ch:  I cannot wait.   .17 sounds awesome!
15:37 s0undt3ch JasonSwindle: Glad to hear that! And yes, it will be super awesome!!!
15:37 jinrou joined #salt
15:37 JasonSwindle As KMart is going on about; I just shipped my logs!
15:38 [diecast] joined #salt
15:39 abe_music joined #salt
15:39 jalbretsen joined #salt
15:40 Jahkeup joined #salt
15:41 whit joined #salt
15:41 welby oh nice, direct logstash logging, thats going to be rather spiffy!
15:42 dthom91 joined #salt
15:42 racooper well this is weird.  running salt '*' state.highstate test=True, only two of 18 minions are returning anything. all of them respond to test.ping, status.update, etc though.
15:43 welby racooper, try increasing the timeout salt '*' state.highstate -t 180
15:43 forrest racooper, is there a process spinning up on the minion for salt on the boxes that aren't returning?
15:44 hellome joined #salt
15:46 racooper shouldn't be anything. the only thing changed in the highstate is a new user being added.
15:47 [diecast] joined #salt
15:47 nocturn joined #salt
15:48 racooper ok if I run my common state (salt '*' state.sls common test=True) everything responds as expected. so something in my top.sls perhaps is messed up?
15:48 forrest does every machine get the same states applied to it?
15:49 msil xmj: i use pf
15:49 msil mainly because the book of pf is amazing
15:50 racooper every state in 'common' should be applied to all servers, yes.  there are several sub-states there (denyhosts, sudo, zabbix, etc) but they all are applied to all hosts.
15:52 forrest Do the two machines that work avoid some state the others always get?
15:52 racooper my top.sls is pretty bare.  base: '*': - common (well, with proper lines and spacing)
15:53 forrest can you try to apply the highstate to one of the boxes that isn't returning output?
15:53 racooper nothing that I can tell.  I don't do any targeting within state trees currently.
15:53 racooper I tried that, and it didn't return any output either (except for the hostname itself)
15:55 hellome joined #salt
15:55 wolfpack joined #salt
15:55 dave_den joined #salt
15:57 forrest are the state changes getting applied on the minion even though output isn't returned racooper?
15:57 racooper they were not.
15:58 hellome joined #salt
16:00 Ryan_Lane joined #salt
16:00 terminalmage forrest: https://github.com/terminalmage/djangocon2013-sls/tree/master#djangocon2013-sls
16:00 forrest *thumbs up*
16:00 terminalmage multihost branch README has multihost-specific info on it
16:01 forrest cool, that is appreciated
16:01 terminalmage no prob
16:01 terminalmage it was a good thing to add
16:01 terminalmage because I don't think I put the pillar_roots stuff in my slides
16:01 terminalmage might have, but I'm not sure
16:02 forrest are you gonna put your slides up on slideshare?
16:02 terminalmage I don't know
16:02 terminalmage don't know if they'll work right htere
16:02 terminalmage *there
16:02 terminalmage they have text animations
16:02 troyready joined #salt
16:03 forrest Oh
16:03 terminalmage text that appears/disappears
16:03 forrest yea
16:03 forrest Maybe speakerdeck would work?
16:03 forrest They don't mention that animated ones won't work
16:03 forrest but I don't know
16:03 terminalmage I might try to use impress js for future presentations
16:03 terminalmage because it's already web enabled
16:03 terminalmage or one of the others like impress js
16:04 forrest Yea that would be cool, just have to ensure it's permanently hosted somewhere, or available in a repo.
16:05 terminalmage yeah, at worst it'd be up on github
16:05 UtahDave whiteinge uses a really cool framework for his presentations. I don't remember what it's name is, though
16:07 jinrou joined #salt
16:08 KyleG joined #salt
16:08 KyleG joined #salt
16:09 jheintz joined #salt
16:11 jheintz I've got a state script that works... but I'd like to improve it. My goal is to be able to state.highstate and when a salt://script.sh is updated it should be re-run, but if it's not updated then it shouldn't repeat execution. Here is the gist: https://gist.github.com/jheintz/6466033
16:11 jheintz This functions exactly how I want, but it's really verbose and repeats the file path three times.
16:11 jheintz Any advice?
16:14 jheintz_ joined #salt
16:15 honestly will "salt 'id' state.sls my-state test" do a testrun of my-state?
16:16 racooper test=True
16:17 honestly thanks
16:17 forrest looks fine to me jheintz
16:18 honestly test means it will compile the state but not execute it?
16:18 forrest glad to see you got it working
16:18 denstark Is there any way to turn off salt-minion outputting to all ttys on that machine? When I deploy to developers they get messages (for example,  �<14>yum[9933]: Installed: phantomjs-1.9-1.x86_6 when I install a pkage)
16:20 jheintz__ joined #salt
16:21 jheintz Hi forrest, yes! Thanks again for you help.
16:21 forrest I think it was mostly Furao, he saved you from my terrible plan of jinja :P
16:22 jheintz Can you see any way to improve the state script?
16:22 honestly huh. do I need to rebuild the pillar system somehow after making changes to it?
16:22 jheintz Furao was indeed great help.
16:22 forrest jheintz, It looks fine to me.
16:22 jheintz Fine? I guess I'm sensitive to the amount of duplication.
16:22 forrest but it's not really duplication
16:23 littleidea joined #salt
16:24 jheintz the file path is really duplicated three time...
16:24 forrest it just feels that way because of the way the script handles, I mean if you saw it like this: https://gist.github.com/gravyboat/6466213
16:24 jheintz here's a state script that describes what I want to be true... https://gist.github.com/jheintz/6466203
16:24 forrest would that bug you?
16:26 dave_den joined #salt
16:26 jheintz The example you pose is better, yes. but in that case there really are two separate things...
16:26 forrest I know, but you keep focusing on the name being a big deal because it's repeated 3 times
16:26 jheintz In the situation I want there is really only one thing... a single script.
16:27 forrest right
16:27 forrest but without knowing what is going on the with the script, salt can't do anything with it
16:27 jheintz well, repeating things is error prone. and the esthetic of it does bother me. :)
16:27 forrest hah
16:27 forrest Yea I don't know of a better way to do it.
16:28 jheintz thanks for looking! I'm happy I have a working solution.
16:28 forrest I mean it's not really a repetition, it's not like you're defining the file in two different state files or something
16:28 honestly this is really weird. "pillar.items" shows the pillar data exactly the way I expect it, but trying to compile a state that relies on pillar data causes pillar rendering to fail with a reference to some old pillar data that doesn't exist anymore and itsn't referenced by anything
16:28 forrest yea np jheintz, I think you're worrying about it too much
16:28 jheintz I'm running master and minion on my macbook... so I can't count on everything working (OSX support is a little sketchy right now)
16:28 forrest ahh gotcha
16:28 forrest you ever think about just spinning up a VM/
16:28 forrest *?
16:28 devinus joined #salt
16:29 jheintz yes, and I will. I've actually wanted an automation solution for my own laptop for a while. it's a good excuse to learn Salt and I get my system managed.
16:29 forrest fair enough!
16:30 jheintz and, i get the ability to "oh yeah, I use the latest devops tools to manage my own laptop! I'm sure we can do something with those servers..."
16:31 honestly ....anyone?
16:31 xmj jheintz: it gets even better if you add virtualization into the game
16:31 xmj who doesn't want to have something like Xen, KVM, bhyve running on their laptops?
16:32 xmj (no, there was not a hint of sarcasm in that comment. that's how i roll :])
16:32 jheintz xmh, yes I agree. salty-vagrant is on my agenda.
16:32 jheintz sarcasm? nah.
16:33 honestly I needed to do refresh_pillar, there we go
16:33 * xmj high-fives jheintz
16:34 jheintz_ joined #salt
16:34 jaequery joined #salt
16:34 jheintz thanks again forrest, and catch you next time xmj. cheers
16:34 forrest later
16:35 honestly whelp, this is stupid. jinja doesn't like lambdas...
16:35 honestly anybody here got jinja-fu?
16:36 honestly 15 {%    set userstring = ' '.join(map((lambda x: x + '(rw)'), pillar.get('users_nfs:clients',[]))) %}
16:36 honestly this is what it's complaining about
16:36 honestly minus the 15, that's a line number...
16:37 xmj erm
16:37 honestly I was assuming I could put arbitrary python into jinja...
16:38 gmoro joined #salt
16:38 xmj my eyes hurt. :)
16:38 xmj " ".join([x + "(rw)" for x in pillars.get('users_nfs:clients',[])] not good ?)
16:39 KennethWilke honestly: mako's a lot more like python
16:39 racooper ok I'm not sure what fixed it, but I manually restarted salt-minion on all the servers....state.highstate runs as expected now.
16:39 honestly xmj: if it works :D
16:39 xmj s/ not good ?)/) not good?/
16:39 KennethWilke honestly: jinja's theory is you should do all pythony stuff outside of the template
16:39 forrest well, at least it was easy racooper
16:39 xmj honestly: from the docs "The operators are limited to the most common ones and more advanced expressions such as list comprehensions and generator expressions are not supported."
16:39 honestly :|
16:40 honestly I am disappoint.
16:40 KennethWilke yeah i've wanted list comp's in jinja often
16:40 KennethWilke no worries honestly! you still have mako
16:40 honestly I spent like a week on writing this template
16:40 honestly well, most of it was spent procrastinating on writing it
16:40 KennethWilke mako's syntax is fairly similar to jinja, switching from jinja to mako is not a tremendously painful process
16:41 honestly cool
16:41 KennethWilke honestly: http://www.makotemplates.org/
16:41 chrisgilmerproj left #salt
16:41 honestly <%
16:42 KennethWilke i think using list comps in their sample at the top of their homepage is probably a poke at jinja
16:43 honestly is using <% instead of {% all I need to do to use mako?
16:44 KennethWilke no i wouldn't say it's that simple, but for the simplest templates that may work
16:44 honestly lol yes
16:44 honestly I meant to make salt use the mako renderer
16:44 KennethWilke oh, is it for a sls file or a file.managed template?
16:44 honestly sls
16:44 KennethWilke '#!mako' i think on your top line
16:44 honestly heh.
16:45 KennethWilke good ol' she-bang
16:47 giantlock joined #salt
16:48 honestly "The simplest expression is just a variable substitution. The syntax for this is the ${} construct, which is inspired by Perl, Genshi, JSP EL, and others"
16:48 honestly then why am I hearing M60s and AKs firing, Hueys flapping, and A-10s screaming...
16:49 bdf any idea why salt would hang on freebsd when trying to start postgresql?
16:49 bdf [INFO    ] Executing command '/usr/sbin/service postgresql onestart' in directory '/root'
16:49 xmj has nobody told you that perl-compatible regular expressions are the simplest thing to do?
16:49 bdf if I run the command outside of salt it works every time
16:49 honestly xmj: lol
16:50 honestly I'm just poking fun at the exclusion of PHP as inspiration
16:50 xmj bdf: why don't you use service.running: True ?
16:50 cyp joined #salt
16:51 bhosmer joined #salt
16:51 xmj or, well, enable it in /etc/rc.conf and use the service.% thing
16:51 bdf that is what I'm doing, but it's hanging regardless of how it's invoked
16:51 bdf highstate will hang
16:51 bdf service.start will hang
16:52 xmj weird
16:52 xmj bdf: https://github.com/xmj/jailedbees/blob/master/states/postgresql/init.sls works for me
16:52 xmj on a mid-august 9.2-PRERELEASE snapshot inside an ezjail.
16:52 xmj cedwards: look, yet another freebsd salt user!
16:53 bdf postgres 9.2 yeah?
16:53 bdf oh and other services work fine when invoked the same way
16:53 bdf I'm using 8.4-RELEASE
16:54 xmj uh, i don't remember what postgres I used for that one..huh
16:54 JasonSwindle joined #salt
16:54 lineman60 joined #salt
16:56 cedwards strange
16:59 dthom91 joined #salt
17:00 bdf almost seems related to #5567
17:01 bdf can you make it through salt-call service.stop postgresql and salt-call service.start postgresql ?
17:01 bdf so if it's already started it's fine.
17:01 bdf but this is hanging highstate on the first start of postgres, which is why it sucks so much.
17:02 bdf I'm half tempted to just make supervisord start postgres
17:02 dthom91 joined #salt
17:19 devinus joined #salt
17:21 jacksontj joined #salt
17:28 zwevans joined #salt
17:28 littleidea joined #salt
17:32 mannyt joined #salt
17:33 jesusaurus are other people seeing SaltReqTimeoutErrors in recent versions of salt?
17:34 jesusaurus I just installed a new test environment and seem to be getting them pretty frequently
17:35 UtahDave jesusaurus: on develop branch?
17:36 jesusaurus the minions are development, the master is a little behind
17:36 jesusaurus s/ment//
17:37 jesusaurus master is: salt 0.16.0-2156-g2573a96
17:37 baniir joined #salt
17:37 UtahDave that's not too far. I don't know of any changes that would cause a compat issue
17:37 UtahDave are you seeing those errors in the logs?
17:38 eculver joined #salt
17:38 eculver joined #salt
17:40 maspwr joined #salt
17:40 baniir is there a convenient way to combine multiple ssh pub keys (separate source files) to one authorized_keys file
17:40 jesusaurus no, i dont see any in the logs, but it occasionally shows up on the command line. not every time, but occasionally
17:41 jacksontj is there a way in the salt event system to get all events on the bus with the tag as well as the data?
17:41 vaxholm joined #salt
17:42 jesusaurus baniir: what exactly are you trying to do? you could use ssh_auth state if you are using the state system
17:43 benspaulding jesusaurus: I started seeing those when I upgraded as well. I went from using the packages to using develop, and I was using the gitfs backend and had not installed GitPython.
17:44 benspaulding Once I installed python-git and restarted salt-master, that SaltReqTimeoutError went away.
17:44 jesusaurus benspaulding: on the master?
17:44 benspaulding Yeah.
17:45 jesusaurus im also playing with gitfs, but that should already be installed
17:45 * jesusaurus double checks
17:45 jesusaurus yeah, its there
17:46 baniir jesusaurus: does the ssh_auth state enable this. i'm using it now with -source, but that would require the source file to have all the keys. should i be looking at "-names"
17:47 jesusaurus baniir: rather than managing the file, you should be managing the ssh keys in the file
17:47 benspaulding Oh, I also recall there were a couple of times (and this sounds weird and I cannot explain it) that I had to re-accept keys and it worked. I was getting SaltReqTimeoutErrors, I rant salt-key -A, it said there was nothing to do because everything was accepted, but then I ran salt '*' state.highstate and got no SaltReqTimeoutError — it worked just fine.
17:47 benspaulding It was really strange.
17:48 jesusaurus benspaulding: that sounds more like my issue
17:48 jesusaurus i actually had one minion that derped and wasnt fully connected. i deleted it and reinstalled, now things seem to be running a bit smoother
17:49 jslatts joined #salt
17:52 benspaulding I also I have a situation where the only fileserver backend I am using is gitfs. When calling highstate, it finds the tree and attempts to execute the states, but any state using a salt:// file fails, saying that file cannot be found.
17:52 benspaulding Any idea why that would be?
17:52 benspaulding If I remove the gitfs configuration and just put the files in /srv/salt manually, then restart salt-master, I can run highstate and it finds the files without issue. I just can’t figure out why it finds the tree, but not the files, when they are in the same repo.
17:53 bhosmer joined #salt
17:55 dthom91 joined #salt
17:58 jaequery joined #salt
17:59 jbunting joined #salt
18:01 jpeach joined #salt
18:04 druonysus joined #salt
18:04 druonysus joined #salt
18:04 benspaulding I all worked previous to upgrading to develop. I did that because I needed the gitfs_root feature, so it can use a subdirectory in a repository.
18:08 devinus joined #salt
18:10 mgw joined #salt
18:10 m_george joined #salt
18:14 jacksontj joined #salt
18:20 maspwr joined #salt
18:20 cjh is it possible to request salt run a job once on any minion at a given time?  It would make a great distributed cron if it could do that
18:20 cjh i should say once on a random minion
18:24 timoguin_ joined #salt
18:24 toguin__ joined #salt
18:27 zooz joined #salt
18:31 xmj bdf: does supervisord work on freebsd?
18:32 xmj ^ disregard; as it is in ports as sysutils/py-supervisor I'll just assume it'll work.
18:34 CheKoLyN joined #salt
18:34 aboe joined #salt
18:36 micah_chatt joined #salt
18:42 cbloss I'm having an issue getting my salt-master to see any salt-minions. I *think* the ports have been opened correctly on the master…still can't get the minion to connect
18:42 cbloss ACCEPT     tcp  --  anywhere             anywhere             state NEW tcp dpt:4505
18:42 cbloss ACCEPT     tcp  --  anywhere             anywhere             state NEW tcp dpt:4506
18:43 kamal_ cbloss: telnet master 4505 from a minion
18:43 cbloss on the salt-minion: [DEBUG   ] Loaded minion key: /etc/salt/pki/minion/minion.pem
18:43 cbloss [INFO    ] Waiting for minion key to be accepted by the master.
18:43 kamal_ does it work or does it timeout?
18:44 cbloss nothing happens
18:44 cbloss trying telnet now 1 sec
18:44 cbloss yeah, connection refused
18:45 cbloss mmm obviously something is bad with my iptables
18:51 isomorphic joined #salt
19:00 abe_music joined #salt
19:03 Katafalkas joined #salt
19:06 mwillhite joined #salt
19:06 hannesstruss joined #salt
19:08 teebes Can the minion ID be accessed in state files and/or template files? say if we want to have a file.managed with the minion ID as part of the file name?
19:11 teebes hmm nvm it's built in to grains, nice
19:12 cbloss what port does salt-minion run on?
19:14 whiteinge cbloss: the minion doesn't require any open ports
19:14 whiteinge the master requires two
19:15 m_george left #salt
19:15 cbloss whiteinge: I was able to fix the firewall for the master, but with the same config on a minion, a test-ping fails after the firewall rules have loaded
19:19 whiteinge i'm not going to be much help with firewall rules (though i'm sure someone in here would :), but if the minion only needs to be able to make outgoing connections. no incoming
19:20 whiteinge can you telnet/netcat to the master from that minion on those two ports?
19:22 melinath joined #salt
19:23 baniir jesusaurus: thanks for the tip; i got things working
19:23 Khollowa_ joined #salt
19:25 blast_hardcheese joined #salt
19:26 xmj cbloss: do you actually allow icmp packets?
19:26 Vivek joined #salt
19:27 cbloss xmj: http://pastebin.com/vu9NKamj
19:28 mattikus` joined #salt
19:28 xmj cbloss: I don't do iptables. :-)
19:30 cbloss basically i'm denying all traffic that I don't explicitly allow
19:30 xmj lol, it made my head hurt. why is your first rule accept everything?
19:30 xmj and if that's your first rule, why do you have more than one?
19:31 cbloss i'm looking for that line in my config, which i'm not finding
19:31 xmj ACCEPT     all  --  anywhere             anywhere
19:31 xmj line 5
19:31 cbloss http://pastebin.com/v7CAmXv7
19:32 cbloss yeah, i see it in the output, not in the config section. maybe i'm missing something here
19:32 cbloss pillar salt "firewall_rules" is empty
19:33 kermit joined #salt
19:34 xmj ahum, find someone from your distribution who is proficient in iptables.
19:34 xmj ipfw on freebsd defaults to deny all when enabled in the kernel configuration :-)
19:35 cbloss the problem was with the loopback accept all, not sure exactly why that was an issue...
19:38 mattikus joined #salt
19:38 whit joined #salt
19:38 xmj can your minions contact master now?
19:39 bhosmer joined #salt
19:41 cbloss xmj: nope something is still broken with the iptables config…not sure what
19:41 whit_ joined #salt
19:46 pdayton joined #salt
19:48 mattmtl joined #salt
19:52 pass_by_value joined #salt
19:59 clone1018 joined #salt
20:01 whit_ joined #salt
20:01 jbunting joined #salt
20:03 [diecast] is this valid in a jinja template - serverName = {{ salt['pillar.get']('fqdn') }}
20:03 pass_by_value joined #salt
20:04 [diecast] ah, nm… its a grain that i need
20:09 Katafalkas joined #salt
20:10 Katafalkas joined #salt
20:11 dthom91 joined #salt
20:11 clone1018 joined #salt
20:16 bhosmer joined #salt
20:16 whit joined #salt
20:18 whit_ joined #salt
20:22 p3rror joined #salt
20:23 JasonSwindle left #salt
20:25 troyready joined #salt
20:25 UtahDave joined #salt
20:27 mannyt joined #salt
20:28 whit joined #salt
20:29 \ask joined #salt
20:32 mwillhite joined #salt
20:32 dave_den joined #salt
20:33 dave_den joined #salt
20:33 whit_ joined #salt
20:33 bhosmer joined #salt
20:34 whit__ joined #salt
20:38 jbunting joined #salt
20:40 copelco joined #salt
20:40 Katafalkas joined #salt
20:41 blee joined #salt
20:43 jaequery joined #salt
20:54 copelco is there a good way to read the output from salt '*' state.highstate when you have a lot of states? i'm finding it difficult to parse since there's so much output
20:54 clone1018 joined #salt
20:57 UtahDave copelco: there are 2 options in the master config to control how much ouput you see.
20:57 copelco oh, which ones?
20:57 UtahDave state_verbose  and state_output
20:58 UtahDave there should be examples in your master config
20:58 jacksontj joined #salt
20:58 copelco UtahDave: thanks!
20:59 Katafalkas joined #salt
20:59 UtahDave you're welcome!
21:03 * robawt highfives UtahDave
21:04 * UtahDave highfives robawt back!
21:05 robawt wuddup UtahDave?
21:05 UtahDave just ironing out a couple of tests so our test suite runs on something like 11 platforms flawlessly!!!!!
21:06 robawt wow
21:06 UtahDave how about you?
21:06 robawt still putting together some documentation for Salt for our awesome team
21:06 LarsN good Evening/afternoon/night/morning/day all.
21:06 robawt hey what's the status on testing within salt?
21:06 robawt to test the states themselves, just test=true?
21:07 gatoralli joined #salt
21:07 robawt hi LarsN
21:07 gatoralli can I change salt-master's file_roots at runtime?
21:07 whiteinge robawt: don't leave me hanging!
21:07 * whiteinge holds his hand up expectantly
21:08 UtahDave hey, LarsN!
21:08 * robawt Top Gun highfives whiteinge for putting up robawt's million questions
21:08 LarsN I wonder how close Salt-Cloud is to working with HP Cloud Services private beta (which includes all the virtual networking magic)
21:08 UtahDave gatoralli: you have to restart the master for the change to take affect
21:08 LarsN I know our Seatle office hosted a local sprint during the great salt sprint.
21:09 UtahDave LarsN: I believe it works fine.  I haven't tried it myself, though
21:09 LarsN UtahDave: did you guys ever get your partnership/account straightened out?
21:10 UtahDave LarsN: I'm not sure. I think so. It's kind of funny. The guy you referred me to had already been working with our ceo to get that all hammered out.
21:10 micah_chatt hey everyone, I'm using the #!py renderer, and having some issues. state.highstate is failing silently when I run it
21:10 micah_chatt on one of my states
21:10 LarsN yeah I saw that.  Hopefully you guys get/got all squared away.  It can only help.
21:11 micah_chatt when I run highstate with --out=raw I get this message "TypeError encountered executing state.highstate: unhashable type: 'list'. See debug log for more info"
21:11 micah_chatt "Possibly a missing arguments issue:  ArgSpec(args=['test'], varargs=None, keywords='kwargs', defaults=(None,))"
21:11 troyready joined #salt
21:12 micah_chatt can anyone point me in how to go about debugging that?
21:13 TheRealBill joined #salt
21:13 UtahDave micah_chatt: what version of Salt are you on?
21:14 micah_chatt latest, 0.16.3
21:14 troyready joined #salt
21:17 micah_chatt @UtahDave, I actually narrowed it down, its not the py renderer
21:17 Lue_4911 joined #salt
21:18 troyready joined #salt
21:19 micah_chatt @UtahDave which leads me to a new question, I want to use a custom APT repo, but I have more than one package depending on the repo,
21:19 micah_chatt that is what is failing silently
21:19 UtahDave micah_chatt: can't you use the pkgrepo state to add your custom repo/
21:19 UtahDave ?
21:20 micah_chatt Yes, thats what I'm trying to use
21:20 micah_chatt https://gist.github.com/micahhausler/6470159
21:21 UtahDave why is there a "base" in there?
21:21 micah_chatt I read the docs on http://docs.saltstack.com/ref/states/all/salt.states.pkg.html and it said you have to specify the packages in the pkgrepo staet
21:21 micah_chatt I'm sorry, I'm a noob, its what I saw in the docs I just referenced
21:22 micah_chatt should it not be 'base'?
21:22 UtahDave you don't need the base in there. I think you're getting that confused with the top.sls file
21:22 abe_music is this no longer valid in an SLS? {{ salt['pillar.get']('pkgs:apache', 'httpd') }}
21:22 UtahDave no remove the base
21:22 abe_music I get an error: UndefinedError: 'dict' object has no attribute 'pillar.get'
21:22 UtahDave also your require_in can't use the - pkgs option.   just refer to pkg: postgres-pkgs
21:23 micah_chatt ahhhhhhhh ok
21:23 micah_chatt what do I replace base with? an arbitrary name?
21:23 UtahDave abe_music: you're probably getting  a stacktrace or error when your pillar is being compiles
21:23 UtahDave micah_chatt: no, just remove base. unindent everything
21:23 UtahDave one level
21:26 jacksontj joined #salt
21:26 alunduil joined #salt
21:26 abe_music UtahDave: the error is saying that the key 'pillar.get' doesn't exist…from playing around it looks like pillar.raw and pillar.data are the only pillar keys available in the salt dict
21:27 abe_music UtahDave: was looking at this info: http://docs.saltstack.com/topics/tutorials/pillar.html
21:27 abe_music The function pillar.get used in this example was added to Salt in version 0.14.0
21:28 mannyt joined #salt
21:28 micah_chatt @UtahDave, seriously thanks for the help, but I'm still having issues
21:28 micah_chatt "Name pkgrepo.managed in sls postgres.packages is not a dictionary"
21:28 micah_chatt here's my state https://gist.github.com/micahhausler/6470159
21:29 UtahDave abe_music: what's the output of salt \* pillar.data    ?
21:29 p3rror joined #salt
21:30 UtahDave micah_chatt: you need to give your pkgrepo.managed an ID Declaration.  That's similar to the arbitrary name you gave the other one: postgres-pkgs
21:30 abe_music UtahDave: https://gist.github.com/abemusic/6470250
21:30 micah_chatt ahhhhh ok
21:30 nkuttler joined #salt
21:30 abe_music UtahDave: basically this used to work, and now it's not: {% set dfs_name_dir = salt['pillar.get']('cdh4:dfs:name_dir') %}
21:31 abe_music and this does: pillar['cdh4']['dfs']['name_dir']
21:31 UtahDave yeah, the pillar.get example should work.
21:31 abe_music hmm
21:31 UtahDave what's the output of  salt-minion --version  ?
21:32 forrest abe_music, did you break your salt install again??
21:32 abe_music forrest: LOL, :(
21:33 abe_music UtahDave: that's the problem…and i don't want to say what the version is :)
21:34 forrest how do you keep downgrading the version :\
21:34 UtahDave abe_music: lol    12.1?
21:34 forrest or is this a different box
21:34 abe_music UtahDave, forrest : salt-minion 0.13.0-64-g1d31da6
21:34 forrest ahh
21:34 UtahDave :) nice
21:34 abe_music forrest, i'm still trying to find a way around my problems from the last couple of days
21:34 forrest oh so you're blowing the minion away?
21:34 abe_music when running 0.16.3 i continue to get the broken pipe error when installing a package until i restart my minion
21:35 forrest that makes no sense to me
21:35 abe_music so i was attempting to bootstrap minions from git
21:35 forrest ahh ok
21:35 forrest I was gonna ask if something that you were running kept pulling an old version.
21:35 abe_music i didn't know (until just now) that the args '-D git' on my bootstrap would install 0.13.0
21:35 forrest lol
21:35 abe_music i assumed git would pull the latest :)
21:36 abe_music funny how i could install a package after i switched to pulling from git though, but not i guess i need to specific what to pull from git
21:36 forrest yea
21:37 forrest I'm surprised the stable, which is the default isn't pulling in 0.16.3
21:38 forrest that reminds me that at some point the docs should be updated to show how to pull a specific release with the bootstrap
21:39 abe_music sooner or later i'll get this working :)
21:39 abe_music unless the boss man says i'm wasting too much time that is haha
21:39 forrest Is this still a problem for you abe_music
21:39 forrest https://github.com/saltstack/salt-bootstrap/issues/193
21:40 forrest Hah
21:40 abe_music forrest: it was, so i removed the git part and just stuck with the -D flag
21:40 abe_music i haven't tried it in a while, but was going to do that next….i forgot about adding that "develop" on to the end (thought just git would do that)
21:41 forrest yea give that a shot
21:41 forrest Man I can't remember but someone was having an issue when where they tried to pull, the tty value was basically thinking it was an interactive session when they connected, so it would just hang
21:41 forrest but that was an issue with their actual git server, and they had to modify it there.
21:47 abe_music forrest: that was my issue (but i wasn't using a local git server)
21:49 jpeach can a pillar data item take its value from another pillar data item?
21:51 UtahDave jpeach: no, not currently
21:51 jpeach UtahDave: 'k thanks
21:53 troyready joined #salt
21:53 abe_music forrest: looks like the git develop bootstrap is working again
21:54 abe_music UtahDave: and after adding develop to the bootstrap and getting back to the 0.16 line the pillar.get is working :)
21:54 abe_music ugh, i blame friday
21:55 forrest abe_music, cool!
21:56 UtahDave ok, good.  glad it's working.   :)
21:56 forrest abe_music, if you're not able to duplicate that issue again with the same call, can you close that issue?
21:57 abe_music forrest: yup
21:57 forrest great
21:57 forrest you used salt-cloud and it worked the full way through right?
21:57 forrest when you just tested it?
21:57 abe_music forrest: looks like it's already been closed
21:57 abe_music forrest: yup
21:57 forrest https://github.com/saltstack/salt-bootstrap/issues/193
21:57 forrest that one is open
21:58 abe_music oh, sorry, was looking at the referenced ticket
21:58 dthom91 joined #salt
21:58 forrest it's all good, the formatting is a bit odd on those
22:00 forrest Ugh, I hate when you merge in upstream, make a change, and then it doesn't build. I didn't even change anything but documentation
22:01 oz_akan_ joined #salt
22:01 pdayton joined #salt
22:02 blee Anyone here using proxmox?
22:05 druonysus joined #salt
22:05 druonysus joined #salt
22:07 pman joined #salt
22:09 Katafalkas joined #salt
22:10 tyler-baker joined #salt
22:15 forrest nope sorry blee
22:16 \ask hi everyone -- I have a state that installs and starts httpd; other states then adds/updates configuration files. How can I "notify" the service when the config updates?
22:16 \ask Is it something like this:
22:16 \ask https://gist.github.com/abh/a1b0479d36e3a8df1b79 ?
22:17 UtahDave \ask: you need    watch_in   there instead of   watch
22:18 littleidea joined #salt
22:18 \ask ah, "watch inverse" ?
22:22 UtahDave yeah.
22:23 UtahDave otherwise you would need to put the watch on your httpd service definition
22:27 dthom91 joined #salt
22:33 isomorphic joined #salt
22:37 cjh UtahDave: question..would it be possible to have salt pick a random minion and run a job such that it's only run once per scheduled time period?
22:40 cjh if it could you'd have badass distributed cron system
22:43 forrest cjh, there's a way you can do grouping and then you only run it on a certain % of the minions, but I don't have that bookmarked on this machine
22:43 cjh forrest: interesting
22:44 cjh i see there's a node group
22:45 forrest http://docs.saltstack.com/topics/targeting/nodegroups.html
22:45 forrest yea that's what I was thinking of it we're looking at the same thing
22:45 cjh i saw chronos but it's so heavy.  salt would be so much easier to use
22:46 CheKoLyN joined #salt
22:48 kermit joined #salt
22:48 UtahDave cjh: there's not a built in way to select a random minion.
22:50 vmdsch joined #salt
22:50 cjh UtahDave: sad :(
22:51 UtahDave I don't think that would be a hard feature to add.  Care to open an issue regarding that?
22:51 logix812 joined #salt
22:51 blee_ joined #salt
22:56 pdayton joined #salt
23:00 blee joined #salt
23:00 sparc_ joined #salt
23:01 forrest cjh, I thought there was a way within nodegroups to run a command on a % of those minions
23:01 isomorphic joined #salt
23:04 marcinkuzminski joined #salt
23:05 forrest cjh, I found it: http://docs.saltstack.com/topics/targeting/batch.html
23:05 vmdsch joined #salt
23:06 forrest I don't know if that's exactly what you want, but it might be at least close
23:07 Ryan_Lane joined #salt
23:07 iquaba left #salt
23:11 oz_akan_ joined #salt
23:12 nliadm I have a non-salt problem I'm trying to solve, maybe someone here has solved it
23:13 nliadm how do you physically locate servers?
23:13 nliadm do you put where they are in the host name?
23:13 nliadm do you keep a database?
23:13 nliadm is that information auto-populated or entered by hand?
23:17 forrest I've seen it done a few ways, the first was via hostname, where each host had the location in the server name. This was nice to easily find what box you were working on, but it also posed issues in the event a server was moved. I've also seen the excel spreadsheet (the worst solution), and a database where they were entered by hand
23:18 forrest If I were going to design one now though, it would be an auto-populated database on the backend, then a web service up front so any user could review the information
23:19 forrest maybe do something with hostname, but that would depend on the location (overseas datacenters, within the same state, etc.)
23:26 jslatts joined #salt
23:28 nliadm yeah, we do it with hostname, but you get renaming issues, and you need to know the hostname to configure the server, but how does the server autodiscover that...
23:28 cjh forrest: thanks!
23:28 nliadm it's turtles all the way down
23:28 cjh forrest: i think that's what i'm looking for. batch size of 1
23:30 UtahDave cjh: that will still go through all the minions that match your target. just one at a time
23:30 cjh oh :(
23:30 cjh so close haha
23:40 jaequery joined #salt
23:43 bhosmer joined #salt
23:56 dthom91 joined #salt

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