Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-01-15

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

All times shown according to UTC.

Time Nick Message
00:00 jdenning how about `salt target state.show_top` ?
00:00 jeddi isn't the =True case sensitive?
00:00 kainswor from the master that works as expected
00:01 jeddi i have aliases for all this stuff - highly recommended.  :)
00:01 Doormatt joined #salt
00:01 jdenning salt target state.show_highstate ?
00:02 kainswor show_highstate has the same exception as highstate
00:03 jdenning ok, cool
00:03 jdenning you're still working on your custom tops module, right?
00:04 kainswor yeah- well I have it working on the master side.. it doesn't get called from minion side execution, but this error comes from master side.
00:04 jdenning there's probably  some logic branch where you're returning a boolean unexpectedly
00:04 ajw0100 joined #salt
00:04 kainswor it returns a logical top.sls dictionary, promise
00:05 jdenning and the state files you are using were working previously?
00:06 kainswor yeah everything was working with 15.3
00:06 kainswor there is a pillar top.sls that is python rendered. I wonder if that is broken
00:08 ekristen jdenning: so I don’t think I want highstate to run on each auth
00:08 ekristen jdenning: I just want it to run the first time the minion comes online for the first time
00:08 jdenning ekristen: yeah, there are a lot of auth events
00:09 scott_w joined #salt
00:09 bhosmer joined #salt
00:10 ekristen can I do /new
00:10 jdenning ekristen: I would inspect the data of the message, and look for one w/ data['act'] == 'pend'
00:10 ekristen oh pending would work too
00:10 kainswor scratch that. pillar.items returns as expected
00:10 jdenning yeah, that will limit it to nodes that are up, but with unaccepted keys
00:10 elfixit joined #salt
00:11 ekristen jdenning: but salt-cloud accepts the keys
00:11 ekristen so they might never go into pending
00:11 jdenning ekristen: ah - I don't use salt-cloud, so I have no insight into that :)
00:11 jdenning ekristen: you could fire a custom event on firstboot
00:11 danielbachhuber joined #salt
00:12 jdenning (my talk @ saltconf will cover that use-case :) )
00:12 ekristen hrm, want to give me insight on how to achieve that?
00:13 jdenning yeah, let me find you a link real quick..
00:14 jdenning basically, just have a firstboot script on your nodes (however you want to accomplish that), that will run `salt-call event.fire '{ "data" : "any data you want to send"}' 'customtag'
00:14 jdenning http://docs.saltstack.com/topics/event/index.html
00:14 jdenning oops, make that event.fire_master
00:15 jdenning I'll be posting the slides from my talk @ saltconf at the end of the month with detailed examples..
00:16 jdenning then just setup your reactor file to listen for `customtag`
00:18 Ryan_Lane2 joined #salt
00:22 kainswor it looks like it's trying to load one of my pillar files as a state- in the debug log I see a bunch of debug lines loading state files and printing OrderedDict's with the contents- and then one that ends up being a pillar value and a boolean key
00:23 jdenning so is there some place that you are pulling pillar data to help build the state files (e.g. to get a proper package name or something) >
00:26 timoguin joined #salt
00:26 kainswor there are several
00:26 ahammond where can I find advice on sizing a salt master? Initial deploy would only be for a few machines, but if things work out, we have just under 2k servers.
00:29 bhosmer joined #salt
00:31 Gifflen joined #salt
00:32 jdenning ahammond: there are too many variables to really give a concise answer to that question (e.g. if you are using salt to distribute a bunch of large files, it will not scale as well as if you are just delivering simple state files)
00:32 kainswor damn.. now I'm getting new exceptions with salt-minion -l debug vs. running as daemon
00:34 ahammond jdenning we won't be distributing large files with it. the goal is to use it for devops type stuff. I don't need a perfect answer, just a rough idea of how much memory and how many CPUs would be reasonable to start with.
00:36 jcockhren ahammond: to "start" meaning ~10? 1 CPU, 512MB RAM works fine
00:37 jdenning ahammond: in general, the salt-master does not need much in the way of resources, most of the computation (template rendering, etc) is done individually on the nodes, the salt-master just serves the state files (and any config files/templates/etc), then sends messages to the minions which do most of the heavy lifting..the main workload for the master is serving files
00:37 iShming Next question....  Looking at the pillar get.items (masterless-minion) I'm seeing recursive key/values?  How would this come about?  In particular it's master -> pillar -> master ->   (that's where it stops)...
00:37 ahammond ok, so just asking for a virtual server with 1gb and 2 cpus should get me started and give me a little room to grow?
00:37 jdenning yeah, that should be great
00:37 jcockhren ahammond: yep
00:38 ahammond does salt support multiple / cascading masters?
00:38 jcockhren it'll probably hold you for a while
00:38 jdenning it uses a system called salt-syndic
00:38 jcockhren http://docs.saltstack.com/topics/tutorials/multimaster.html
00:38 jdenning which basically forwards messages from one 0MQ message queue to another
00:38 ahammond jcockhren perfect. thanks!
00:39 NotreDev joined #salt
00:39 jcockhren there's also: http://docs.saltstack.com/ref/syndic.html
00:40 jcockhren ahammond: one thing to note when using syndic.
00:40 jdenning I feel that multimaster works well for HA, and syndic works well for scaling
00:40 jcockhren let's say you have: master (main) -> submaster (syndic) -> minion
00:41 jcockhren main can't control the submaster using syndic
00:41 schristensen joined #salt
00:41 NV joined #salt
00:41 ahammond syndic looks like what I'd want.
00:41 bhosmer_ joined #salt
00:41 jcockhren in order for submaster to take commands as well, you need to install salt-minion on the submaster
00:42 ahammond have a master at our primary DC, and then syndics in other DCs to manage those machines.
00:46 travisfischer_ joined #salt
00:49 jeff joined #salt
00:49 Linz joined #salt
00:50 yomilk joined #salt
00:53 ekristen jdenning: the only problem with that is I’d have to go out of my way to put a script on the minion system
00:54 MTecknology {% set centers = pillar.get('centers_served', [grains['id'].split('.')[1]]) %}  <-- this 'almost' works, except that it doesn't, If the id is server.0040.domain.tld, then it becomes 40 instead of 0040. If I add double quotes around that, then it uses that bit literally...
00:55 ekristen if I have a reactor trigger, and the minion is offline or busy, does the job queue?
00:55 AdamSewell joined #salt
00:56 timoguin good question
00:59 ekristen timoguin: thanks ;)
00:59 jdenning ekristen: just to clarify your use case - you want to trigger state.highstate on a minion when it first starts, but not on subsequent restarts?  As to queueing the job - that *is* a good question, I would think that it queues for a short period of time then expires.. basepi: do you have any insight?
00:59 ekristen jdenning: yes, I want state.highstate to run the first time a new minion comes online at least at this point
01:00 ekristen jdenning: not sure I want it running each time the minion restarts, but I guess that wouldn’t be the end of the world
01:00 ekristen jdenning: guess I need to think about that
01:00 ekristen everything should be idempotent, so I guess it shouldn’t matter
01:01 jdenning well, if you can use the salt/minion/*/start tag, that should work for you
01:01 clearclaw I have state_output set to "mixed" in /etc/salt/master, and that's honoured for instance when I do a `salt test-mon-jcltest state.highstate`, but it is NOT honoured when I do a `salt-run state.sls foo`
01:01 Linz joined #salt
01:02 jdenning you could also watch all the events on the master for a while and see if something else is suitable (there is no canonical list of possible events that I'm aware of)
01:02 clearclaw eventlisten is a good tool for this.
01:02 clearclaw -> ekristen
01:02 Linz joined #salt
01:03 jdenning indeed ^^^ - I recommend having it open in another window whenever you are doing reactor/event stuff with salt
01:03 ekristen understood
01:03 ekristen are any of you running salt in ec2? if so what instance type are you using for your salt-master?
01:04 jdenning ekristen: m1.small has been working pretty well for me
01:05 jdenning ekristen: BTW, if you're using ec2, it's *super* easy to run a first boot script (you just put it into the user-data field)
01:05 clearclaw ekristen: I am, exclusively.
01:05 clearclaw ekristen: Right now I'm using a micro.
01:05 ekristen jdenning: I’m using salt-cloud to launch my instances
01:06 jdenning does salt-cloud not support user-data?
01:06 clearclaw ekristen: Buy I'm also in the middle of writing the RCS to have salt build its own masters.
01:06 clearclaw jdenning: What sort of user data?
01:06 ekristen RCS?
01:06 clearclaw RCS: http://docs.saltstack.com/ref/states/master_side.html
01:07 jdenning there's a field you can send to AWS when you launch an ec2 instance called user-data - which is pushed to their internal metadata service..it is useful for bootstrapping stuff on first boot, and setting instance-specific metadata that you want available to apps on the node
01:07 clearclaw My ideal is to have a git checkout on a desktop to be able to build my salt master in the cloud, and then that salt master to then be able to build and manage all my environments.
01:07 ekristen hrm not sure I understand the use case behind RCS
01:07 bhosmer joined #salt
01:07 jdenning http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AESDG-chapter-instancedata.html
01:07 * clearclaw nods jdenning.  Oh, that.  I don't use that currently.
01:08 clearclaw I should, but I don't yet.
01:08 jdenning anyways, it makes it really easy to fire off a script on first-boot if you so desire..
01:08 jdenning (which is what I was advocating ekristen try earlier :) )
01:09 MTecknology dave_den: You happen to have any idea on that issue?
01:09 clearclaw ekristen: ekristen: It means I can run things like `salt-run state.sls env_test where env_test.sls looks like: http://pastebin.com/dd1B1H74 and the whole environment comes up as one.
01:10 jdenning btw, it looks to me that salt-cloud does support user-data via 'userdata_file : /path/to/file.txt', but I don't actually use salt-cloud, so I have no idea if it works with AWS (it says they support OpenStack, which uses the same format, so it should work)
01:10 clearclaw I'm more interested in EBS vols and even more interested in emphemeral storage than EBS.
01:10 ekristen clearclaw: what version of salt are you using and are you not using salt-cloud then?
01:11 clearclaw ekristen: I'm using 2014.1 develop branch and I never touch the salt-cloud command any more.
01:11 jcockhren clearclaw: HOW is that?
01:11 clearclaw I really need to be able to build boxes with xxx GB ephemeral storage....and I'm just not seeing how to do that ATM.
01:11 ekristen ah, yeah I’m using 0.17.4
01:11 jcockhren clearclaw: I mean. is it working well for you. it's coming along?
01:12 clearclaw jcockhren: How is which bit?
01:12 clearclaw The RCS stuff?
01:12 ekristen clearclaw: with salt-cloud it was just a matter of doing volumes: { size: 20, device: /dev/sdf1 }
01:12 jcockhren clearclaw: using the saltcloud modules in the develop branch
01:12 jdenning clearclaw: I don't think you can specify how much ephem storage you want in ec2, it is determined by the instance type
01:12 jdenning clearclaw: http://aws.amazon.com/ec2/instance-types/instance-details/
01:13 ekristen oh ephemeral, nm
01:13 ekristen jdenning: yeah you can change it at least in the ec2 interface
01:13 jdenning really? I never tried that..
01:13 clearclaw So far salt is...flakey but getting there.  The RCS-meets-overstate-meets-salt-cloud stuff is grossly in violation of its own documentation, but i've only just this avo kinda sorta gotten it to almost work.
01:13 * jdenning checks it out now..
01:14 ekristen clearclaw: ok, I’ll stick with salt-cloud and 0.17.4 until RCS is more mature ;)
01:14 ekristen I can’t afford there to be any issues
01:14 Gifflen joined #salt
01:14 ekristen well any major issues ;)
01:14 clearclaw jdenning: Yeah, I know the default.  I have processes with really big disk-based working sets.
01:15 clearclaw ekristen: Then why are you using salt?  It is clearly not stable yet.  It is good, it is worth using, but it is not stable/solid/etc.
01:16 ekristen clearclaw: because I hate chef and not a huge fan of puppet, waaaay to many problems with community driven cookbooks and recipes, etc
01:16 clearclaw ekristen: 0.17.4 does not work for me as the ubuntu package for salt-cloud DOES NOT WORK WITH the ubuntu salt package.
01:16 EugeneKay IMO it's solid enough. Just run new versions through a QA env before updating prod.
01:16 ekristen clearclaw: saltstack is stable enough for my use
01:16 ekristen clearclaw: salt-cloud is working fine for me with ubuntu 12.04
01:16 clearclaw Likewise, though perhaps minus the hate.
01:16 clearclaw Salt's reactor system was the compelling feature for me.
01:16 ekristen clearclaw: I’ve been building and tearing down 12.04 systems all day
01:17 EugeneKay I'm pondering how to run Salt against my NetApps next
01:17 clearclaw I'm a mix of 12 and 13.
01:17 jdenning clearclaw:  Hmm, it still looks to me that ephem storage is limited by instance type..you can add as many EBS volumes as you want, but maybe I'm missing something..
01:17 EugeneKay It would make a lot of my custom painful scripting go away
01:17 cachedout joined #salt
01:20 ekristen jdenning: I was thinking EBS backed instances types only, my bad
01:21 jdenning ah, that makes sense :)
01:25 xzarth joined #salt
01:28 fllr joined #salt
01:29 clearclaw I have EBS-backed systems, systems with EBS and ephem storage and systems which are EBS backed and have additional mounted EBS vols
01:29 adizam joined #salt
01:29 clearclaw I need to confront how to spec all of those in statefiles, starting with the ephemerals.
01:30 gnu-man joined #salt
01:31 taion809 joined #salt
02:02 juicer2 joined #salt
02:02 ekristen sudo salt-call --local state.show_highstate
02:03 joshe joined #salt
02:03 ekristen that is suppose to show me what the minion thinks it should do when highstate is called correct
02:06 aarontc joined #salt
02:09 kermit joined #salt
02:11 zzzirk joined #salt
02:12 taion809 joined #salt
02:13 taion809 joined #salt
02:13 deepakmdass joined #salt
02:24 AdamSewell joined #salt
02:26 Dinde joined #salt
02:29 bhosmer joined #salt
02:31 yomilk joined #salt
02:41 stefanmonkey joined #salt
02:45 deepakmdass joined #salt
06:32 ilbot3 joined #salt
06:32 Topic for #salt is now Welcome to #salt - SaltConf Jan 28-30, 2014! http://saltconf.com (reg deadline January 3) | 0.17.4 is the latest | 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/
07:03 Katafalkas joined #salt
07:21 fllr joined #salt
07:38 njpatel joined #salt
07:38 david_a joined #salt
07:43 matanya joined #salt
07:51 gammalget joined #salt
07:51 Katafalkas joined #salt
07:54 bhosmer joined #salt
07:55 felixhummel joined #salt
07:57 ndrei joined #salt
08:00 dancat joined #salt
08:02 linjan joined #salt
08:02 balboah joined #salt
08:02 harobed joined #salt
08:03 ndrei joined #salt
08:03 harobed joined #salt
08:07 martoss joined #salt
08:21 giantlock_ joined #salt
08:22 iri- joined #salt
08:22 fllr joined #salt
08:27 tomspur joined #salt
08:27 tomspur joined #salt
08:27 seanz joined #salt
08:36 bhosmer joined #salt
08:46 achileuss joined #salt
08:50 slav0nic joined #salt
08:51 yomilk joined #salt
08:54 TamCore Is there something special where I have to "register" custom output modules? Module is already in salt/output, but salt won't accept it when I pass it to --out
08:56 achileuss left #salt
09:00 ekristen joined #salt
09:03 apergos "TypeError: get_event_iter_returns() takes at least 3 arguments (3 given)"  :-D
09:04 malinoff :D
09:04 gmtech Hey. I want to populate entire files via Pillar. Can I do that? Do I need to resort to jinja2 magic. I basically want the equivalent of hiera-file in hiera
09:04 Furao gmtech: use hiera ext pillar
09:04 gmtech Yes I thought about that - is that commonly what ppl do?
09:09 salt_ joined #salt
09:22 iri- joined #salt
09:23 sroegner joined #salt
09:24 fllr joined #salt
09:24 sylvinus joined #salt
09:28 salt_ sudo salt '192.168.2.27' test.ping ||No minions matched the target. No command was sent, no jid was assigned.
09:28 salt_ this is my error
09:29 salt_ i have configured my salt-master 192.168.2.30 and salt minion 192.168.2.27
09:30 iri- left #salt
09:31 gmoro joined #salt
09:32 Katafalkas is there a way to write a state, so if some state fails - another gets executed ?
09:32 ckao joined #salt
09:38 salt_ @bbinet - sudo salt '192.168.2.27' test.ping ||No minions matched the target. No command was sent, no jid was assigned
09:43 bhosmer_ joined #salt
09:43 nkuttler salt_: i don't think matching ips is supported? that has to be a minion id
09:44 salt_ then let me know what will be the minion id
09:44 nkuttler salt_: run salt-key
09:45 salt_ in minion ?
09:45 salt_ salt-key salt-key: command not found
09:45 nkuttler it's a command that comes with the master
09:45 nkuttler why do you have a salt binary on a minion anyway?
09:46 salt_ Accepted Keys: Unaccepted Keys: paris Rejected Keys:
09:46 nkuttler so you have no minions
09:46 nkuttler salt_: you probably want to accept the paris key
09:46 salt_ when i run salt-key in the master, i got the above
09:47 salt_ how to accept that one. please let me know
09:47 nkuttler salt_: man salt-key, search for accept
09:47 salt_ -a ACCEPT, --accept=ACCEPT               Accept the named minion public key for command execution.
09:49 salt_ what should i do to accept the paris key. thanks in advance
09:49 esoel_ joined #salt
09:50 Shish joined #salt
09:52 Katafalkas joined #salt
09:53 cnf joined #salt
09:55 salt_ thanks i have accepted the paris key, then waht should i do
09:56 nkuttler salt '*' test.ping, and see http://docs.saltstack.com/topics/targeting/globbing.html
09:57 malinoff I'd rather suggest to read this: http://docs.saltstack.com/topics/tutorials/walkthrough.html
09:57 scott_w joined #salt
09:57 viraptor1 joined #salt
09:59 nkuttler yeah, probably a better idea :)
10:04 rojem joined #salt
10:05 salt_ yes i had run this command salt '*' test.ping but doesn't prompt me anything. do i want to replace * with any id
10:05 salt_ please do let know
10:05 Katafalkas joined #salt
10:05 malinoff salt_, What's the output of salt-key -L now?
10:05 malinoff On the master
10:06 salt_ sudo salt-key -L Accepted Keys: paris Unaccepted Keys: Rejected Keys:
10:06 salt_ please do let me know the next step
10:06 malinoff salt_, Please, read the walkthrough: http://docs.saltstack.com/topics/tutorials/walkthrough.html
10:08 Furao hehe
10:08 salt_ sure but let me know do i want to replace the * in salt '*' test.ping
10:08 salt_ ?
10:09 Furao that's loosing someone else time because you don't take your own time to read stuff that had been documented already
10:09 malinoff salt_, I don't know what exactly do you want. Please, read the walkthrough and understand what all these symbols mean.
10:10 sylvinus_ joined #salt
10:11 tomasfejfar joined #salt
10:12 tomasfejfar Hi everyone :)
10:12 tomasfejfar I'm hitting "State pkg.mod_repo found in sls slsname is unavailable" any ideas why? I'm on 0.17.4 and docs say it shoudl be ok
10:15 Furao tomasfejfar: http://docs.saltstack.com/ref/states/all/salt.states.pkg.html#module-salt.states.pkg
10:15 Furao you're probably looking at the module, not the state
10:15 tomasfejfar Oh yea!
10:16 tomasfejfar So I can't do this in state?
10:16 malinoff tomasfejfar, http://docs.saltstack.com/ref/states/all/salt.states.module.html?highlight=module.run#salt.states.module.run
10:16 malinoff But I think, pkgrepo.managed is way better: http://docs.saltstack.com/ref/states/all/salt.states.pkgrepo.html?highlight=pkgrepo.manage#salt.states.pkgrepo.managed
10:17 tomasfejfar malinoff: Thanks thats absolutely awesome! Will try.
10:17 tomasfejfar Furao : thx too ;)
10:22 fllr joined #salt
10:24 Veticus joined #salt
10:31 djaykay joined #salt
10:36 yota joined #salt
10:43 salt_ joined #salt
10:43 salt_ @malinoff i want to install or update the minion machine from my master machine
10:45 pdayton joined #salt
10:46 zz_cro joined #salt
10:46 eightyeight joined #salt
10:50 malinoff salt_, what do you mean by saying 'install the minion machine'?
10:51 salt_ say sudo apt-get update or sudo apt-get install skype, mongodb ...etc
10:51 xl1 joined #salt
10:53 malinoff salt_, I highly recommend to read all available guides from this section: http://docs.saltstack.com/index.html#getting-started
10:54 salt_ thanks
10:54 salt_ left #salt
10:58 Furao malinoff: you have too much patience
10:59 malinoff Furao, I'm DevOps support engineer, I have to have much patience :)
11:05 kalloc joined #salt
11:09 sylvinus joined #salt
11:12 sroegner joined #salt
11:13 deepakmdass joined #salt
11:22 fllr joined #salt
11:39 bhosmer joined #salt
11:46 Katafalkas joined #salt
11:46 _david_a joined #salt
11:48 david_a_ joined #salt
11:48 Katafalkas is there a way to write a state, so if some state fails - another gets executed ?
11:49 taion809 joined #salt
11:51 malinoff Katafalkas, what kind of problem do you have?
11:52 diegows joined #salt
11:52 Katafalkas I am trying to build a pm2 state. for some reason when I first run "pm2 delete all" command - it gets stuck. I use timeout option, so it fails. the second time I use the same command - it works fine ...
11:53 malinoff Katafalkas, I think, it would be better to increase the timeout
11:53 diegows joined #salt
11:53 Katafalkas i took the time out of - it hung foe the entire night..
11:54 malinoff Because a single run should be consistent - your configuration scripts should not rely on multiple runs
11:55 malinoff Katafalkas, if you run "pm2 delete all" by the hands, does it work?
11:56 Katafalkas yeah. if I create a server, install everything, then login onto the server and run command manually, everything works.
11:56 Katafalkas if I enter the same command in the exactly the same order via salt module - it get stuck
11:57 malinoff Katafalkas, do you use "salt minion-id cmd.run 'pm2 delete all'" ?
11:57 Katafalkas yeah
11:57 malinoff Try to run "salt-call cmd.run pm2 delete all -l debug" on the host locally
11:57 malinoff And watch the output
11:57 Katafalkas kk
11:58 malinoff Probably, it will raise an error
12:05 sroegner joined #salt
12:08 Shenril joined #salt
12:10 ekristen morning sroegner
12:11 ekristen can I use salt states to configure the minion’s startup_states?
12:14 Furao ekristen: startup_states?
12:15 Furao oh
12:15 Furao http://salt.readthedocs.org/en/latest/ref/states/startup.html
12:16 ekristen yeah
12:16 Katafalkas malinoff: it still gets stuck in a same way
12:17 malinoff Katafalkas, and there is nothing in logs, right?
12:17 Furao it's old, I should have noticed this a long time ago
12:17 Katafalkas http://pastebin.com/xWpBH38B
12:18 Katafalkas malinoff: i cant not see anything out of the ordinary ?
12:19 malinoff Katafalkas, is this the full log?
12:19 Furao Katafalkas: maybe there is environment variable that are in your shell that aren't in cmd.run
12:19 elfixit joined #salt
12:20 malinoff Yeah, probably. Try to use the full path to m2 utility
12:20 sylvinus joined #salt
12:20 diegows why state.highstate sometimes returns without reporting the result of the operation? The operation keeps running in background
12:21 Furao diegows: timeout
12:21 diegows is there a way to wait forever
12:21 diegows ?
12:21 Katafalkas malinoff: Furao: but if I kill it and start it again - it works fine
12:22 Katafalkas its seems that command get stuck only for the very first time
12:23 Furao attach strace to it
12:23 Furao maybe you can figure something with the syscall
12:25 jrdx joined #salt
12:33 gadams999 joined #salt
12:34 pdayton joined #salt
12:34 Sypher how do/should we call making a machine managed by salt? For puppet you have "puppetizing" but the salt counterpart of "saltizing" doesn't have a nice ring to it.. suggestions? :-)
12:35 ravibhure1 joined #salt
12:36 honestly seasoning.
12:37 Sypher haha that's a nice one, but not the most logical thing if you start talking about stuff like that
12:37 malinoff Sprinkling
12:40 ekristen minionizing
12:41 che-arne_ joined #salt
12:42 ossalk joined #salt
12:43 che-arne joined #salt
12:43 ossalk hello all. how can I iterarete in a sls file for salt "*" pillar.item host ?
12:44 Furao ossalk: you want all items of all minions?
12:44 Furao you can't
12:45 Furao a minion isn't supposed to get an other minions pillars
12:45 ossalk I want to add a cfg temaplte for my nagios server
12:46 ekristen hrm, highstate triggered by my master via reactor doesn’t work as well as highstate triggered manually
12:46 backjlack joined #salt
12:47 ossalk I want to create hostname.cfg  -> for all my minions
12:47 Furao ossalk: ok I know what you try to do, I did the same, using salt to automatically build monitoring configuration
12:47 ossalk like minion1.cfg minion2.cfg
12:47 Furao for that use salt mine
12:48 ossalk hmm I am newbie , and search for salt mine
12:48 scott_w joined #salt
12:51 higgs001 joined #salt
12:53 mpanetta joined #salt
12:54 elithrar joined #salt
12:59 gadams999 left #salt
13:06 bhosmer joined #salt
13:07 scott_w joined #salt
13:13 gmoro joined #salt
13:15 kalloc joined #salt
13:18 ossalk how can I enable mine functions via pillar?
13:22 fllr joined #salt
13:27 oz_akan_ joined #salt
13:31 Furao joined #salt
13:33 oz_akan_ joined #salt
13:37 ekristen all of a sudden a new node says it has no states to run at highstate, yet I have a base ‘*’
13:38 scott_w joined #salt
13:47 apergos joined #salt
13:50 jslatts joined #salt
13:54 tessellare joined #salt
13:59 scott_w joined #salt
14:04 JasonSwindle joined #salt
14:05 logix812 joined #salt
14:06 srage joined #salt
14:06 analogbyte can I use the gitfs backend to get runners and my pillar top file? or is gitfs only possible for the states?
14:08 deepakmdass joined #salt
14:12 brianhicks joined #salt
14:13 sylvinus joined #salt
14:15 sylvinus joined #salt
14:15 zzzirk joined #salt
14:17 psyl0n joined #salt
14:18 Veticus joined #salt
14:18 racooper joined #salt
14:18 Gifflen joined #salt
14:19 elfixit joined #salt
14:21 delkins joined #salt
14:23 fllr joined #salt
14:23 pdayton joined #salt
14:24 robbyt joined #salt
14:26 tmwsiy does anyone know of anyone using salt for HPC scheduling? IE something like condor or torque but using salt or if not salt something using zeromq?
14:30 toastedpenguin joined #salt
14:31 timoguin joined #salt
14:32 juicer2 joined #salt
14:32 tmwsiy if not I think I may head in that direction. Seems like a perfect base to build up a modern, fast, and sleek system like condor specifically.
14:32 psyl0n joined #salt
14:32 psyl0n joined #salt
14:34 danielbachhuber joined #salt
14:34 balboah tmwsiy: didn't try it but celery has some mentions of zeromq in there: celeryproject.org
14:35 Veticus joined #salt
14:44 vejdmn joined #salt
14:45 scott_w joined #salt
14:49 quickdry21 joined #salt
14:49 AdamSewell joined #salt
14:55 kermit joined #salt
14:56 abe_music joined #salt
14:57 jimallman joined #salt
14:57 tmwsiy balboah: thanks. This looks interesting. Although seems like its a slightly different use case.
14:58 gnu-man joined #salt
14:59 kalloc joined #salt
15:01 aleszoulek joined #salt
15:03 cachedout joined #salt
15:05 mpanetta joined #salt
15:07 tonthon joined #salt
15:08 NotreDev joined #salt
15:15 Lauren_Faust joined #salt
15:18 tmwsiy I have come up with so many neat ideas since starting to wrap my head around salt. Man Mr Hatch you really nailed it with this project. All the way around, quality, community, openness..
15:18 salt_whitebelt joined #salt
15:20 Lauren_Faust left #salt
15:20 opapo joined #salt
15:21 MTecknology {% set centers = pillar.get('centers_served', [grains['id'].split('.')[1]]) %}  <-- this 'almost' works, except that it doesn't, If the id is server.0040.domain.tld, then it becomes 40 instead of 0040. If I add double quotes around that, then it uses that bit literally...
15:21 MTecknology I'm still struggling with this. :(
15:23 ml_1 joined #salt
15:25 kaptk2 joined #salt
15:29 fllr joined #salt
15:29 scott_w joined #salt
15:31 ekristen MTecknology: sounds like it thinks [1] is an integer
15:31 ekristen try
15:32 ekristen grains[‘id’].split(‘.’)[1]|string()
15:32 ekristen if I understand jinja right
15:34 teskew joined #salt
15:35 tyler-baker joined #salt
15:41 rojem joined #salt
15:41 sylvinus joined #salt
15:46 backjlack joined #salt
15:46 mpanetta joined #salt
15:50 forresta joined #salt
15:51 alunduil joined #salt
15:51 bhosmer_ joined #salt
15:51 higgs001 joined #salt
15:58 MTecknology ekristen: because of what seems to be a performance bug, it takes a looooong time to test
16:02 MTecknology ekristen: that didn't work... this is what I have   http://dpaste.com/1552958/
16:03 forresta MTecknology, is centers being created as a list?
16:03 krak3n` joined #salt
16:03 aleszoulek joined #salt
16:04 cachedout joined #salt
16:05 MTecknology forresta: ya
16:05 forresta ok
16:05 JasonSwindle forresta:  Welcome!
16:05 MTecknology centers_served is always a list, I guess that means I should make grains['id'].split('.')[1]|string() be [grains['id'].split('.')[1]|string()] instead
16:05 forresta JasonSwindle, hey
16:09 dave_den mornin'
16:09 dave_den MTecknology: yes, i would try that. you should be expecting a list from both options
16:12 gadams999 joined #salt
16:12 salt_whitebelt i have a master (centos) and a minion (win7), when i run - salt 'windows.minion' -G 'os_family:Windows' ip.get_all_interfaces - in terminal it comes back with no minions matched target when there's only one minion?
16:13 MTecknology dave_den: I did exactly what's in that paste and I'm having the same issue
16:13 dave_den what's the issue?
16:13 MTecknology I'm actually ending up with this --     Rendering SLS "base:sys.files.homedirs" failed: Conflicting ID "/home/3"
16:13 dave_den ah
16:14 salt_whitebelt tried the "*" as well. minion is talking to master when test.ping is perform.
16:14 dave_den and it should be '0003'?
16:14 MTecknology there should never be a /home/3, it'll be like /home/0030, /home/0040, /home/4180, etc.
16:14 [diecast] joined #salt
16:16 forresta salt_whitebelt, do salt '*' grains.ls
16:17 dave_den if you use {{ center|string() }} does it keep the 0030?
16:17 Marion joined #salt
16:17 bhosmer joined #salt
16:18 salt_whitebelt forresta: shows the win7 minion data
16:20 jimallman joined #salt
16:20 salt_whitebelt forresta: it shows the grain title in one columm but not actual info per title. isn't it suppose to show each attrtibutes in one columm and the actual minion data in a column right next to it?
16:21 meteorfox joined #salt
16:21 ravibhure joined #salt
16:23 MTecknology *grumble* why oh why does this need to cache every single user in AD for every single run?
16:24 timoguin salt_whitebelt, forresta, can he do the 'windows.minion' match along with the grain match? does that automatically make that a compound match?
16:24 jalbretsen joined #salt
16:24 foxx joined #salt
16:26 forresta salt_whitebelt, timoguin, I'm not sure, I've installed a windows minion exactly 0 times :P
16:27 timoguin salt_whitebelt, does it respond when you do just: salt -G 'os_family:Windows' ip.get_all_interfaces ?
16:27 pdayton joined #salt
16:27 timoguin i think the targetting style you used won't work
16:28 salt_whitebelt i find it easier to use saltstack to manage rhel than windows but unfortunately we have 70 windows and 30 rhel in our environment
16:28 salt_whitebelt percent that is
16:28 salt_whitebelt timoguin: no it comes back with no minion matched
16:29 timoguin hrm...
16:29 timoguin I'll try to fire up a windows minion today and test that
16:29 vlcn is it possible to fire an additional command on the master via a returner?
16:30 salt_whitebelt i can run salt windown.minion cmd powershell.exe command and it works on the minion but not module remote execution. not sure what im doing wrong.
16:31 MTecknology derpy...
16:32 MTecknology {% set centers = pillar.get('centers_served', grains['id'].split('.')[1]|string()) %}   <-- centers_served is a list, the second part is a string, the string is parsed as a list where each character is an item...
16:32 MTecknology duh...
16:32 MTecknology thanks!!
16:35 dave_den you're still going to have an interpolation problem with '0030' in your id declarations
16:35 JesseCW Any recent additions that allow specifying a specific order to run things with salt other than requires?
16:35 yota joined #salt
16:35 sylvinus joined #salt
16:36 dave_den JesseCW: 'order' has been around for a long time - you need something else?
16:36 forresta MTecknology, so what you're saying, is you're thanking me for asking you about the string earlier? Another internet point for me, woo!
16:36 forresta err list
16:36 forresta because they are clearly related
16:36 JesseCW dave_den: I'll go search for that, don't recall it when I first wrote my scripts
16:37 dave_den JesseCW: http://docs.saltstack.com/topics/tutorials/states_ordering.html#the-order-flag
16:38 forresta vlcn, how about using reactors? http://docs.saltstack.com/topics/reactor/index.html
16:38 vlcn hey that looks like what I want
16:38 vlcn thanks!
16:40 forresta np
16:40 jssjr joined #salt
16:42 JesseCW dave_den: seems itr was added in 0.17, mine are from 0.16, looks like this will do just what I need though, thanks!
16:43 dave_den JesseCW: order has been around for a long time. it's the auto ordering that was new in 0.17
16:44 tmmt left #salt
16:44 JesseCW dave_den: if thats the case I either missed it or there was some reason I couldn't use it, I suppose I'll find out which soon enough
16:45 JesseCW I believe I was trying to order includes
16:45 JesseCW instead of them executing in parallel
16:46 mpanetta_ joined #salt
16:46 dave_den JesseCW: you can't order includes: https://github.com/saltstack/salt/issues/5255
16:47 JesseCW dave_den: reading about auto ordering, wouldn't the new default behavior techinically run the includes in order?
16:48 bhosmer joined #salt
16:48 JesseCW where as prior to the change they apparently ran in lexicographical order
16:48 dave_den if you only include them in one place, yes
16:49 dave_den you can see the order of your states with salt-call state.show_highstate to see how it works
16:50 JesseCW sweet
16:50 JesseCW this means I can remove my ridiculous heirarchy of including the same thing in multiple sls's and requiring them
16:55 jdenning joined #salt
16:57 njpatel joined #salt
16:57 Gifflen joined #salt
16:58 hunter joined #salt
17:00 timoguin salt_whitebelt, what version of salt are you running?
17:01 davet joined #salt
17:03 Linz joined #salt
17:03 salt_whitebelt timoguin: 17.0
17:04 timoguin any particular reason you're not on 17.4?
17:05 KyleG joined #salt
17:05 KyleG joined #salt
17:06 salt_whitebelt timoguin: im sorry its 17.4
17:06 timoguin ah good.
17:06 hunter ahammond: to provide one data point to your question from yesterday at 18:26 my salt-master is 6GB of ram on a 2 core xeon - its a 5 year old HP 1U
17:06 timoguin i just fired up a windows 2k8r2 instance. gonna see if the grains work properly for me
17:07 hunter (at least, thats what I'm starting with)
17:07 salt_whitebelt timoguin: are you using vmware workstation?
17:07 timoguin salt_whitebelt, no, AWS
17:08 salt_whitebelt timoguin: aws is not free right?
17:08 timoguin nope
17:09 timoguin well actually, for this account i'm eligible for 1 instance in the free tier
17:09 timoguin so it's actually a free micro instance
17:09 timoguin but not if i fire up another one. :-P
17:10 salt_whitebelt i wonder why people use aws to do their devops when they could purpose an old server, run kvm, xen or even the free esxi to do their work instead of aws?
17:11 timoguin well, for me, I don't have an old server at the moment.
17:11 timoguin but i have a small testing budget so i can fire up and tear down instances when i'm done
17:12 timoguin so my bill ends up being a few dollars
17:12 vejdmn joined #salt
17:13 JesseCW Is there a way to get an array of items from a pillar and the feed it into an sls for things such as a list of packages to install? An example: https://gist.github.com/anonymous/d30985141962b2a1e8a9
17:13 Corey Good morning.
17:14 Corey JesseCW: Yes there is.
17:14 Corey Unfortunately I don't recall offhand exactly what that is. :)
17:14 Corey I know you can jinja it; wrap a for loop around that pillar call.
17:15 Gifflen_ joined #salt
17:15 timoguin salt_whitebelt, I use DigitalOcean mainly, but I also use Vagrant for local development.
17:15 jdenning salt_whitebelt - it comes down to the total cost of high availability - AWS provides managed services and multiple data centers..it is definitely cheaper to run a a few VM's with KVM or or Vmware, but if you need to run a few VMs in physically separated data centers, with redundant power and data lines, *and* have a 24-hr, on-call ops team to run them, then AWS starts looking more cost effective..
17:16 vejdmn1 joined #salt
17:17 Corey jdenning: Like so much else, it depends upon scale. :-)
17:17 mpanetta and Rackspace even more so. ;)
17:17 timoguin yep, it's either multiple AWS regions or multiple data centers
17:17 timoguin neither of which are cheap
17:18 felixhummel joined #salt
17:18 mgw joined #salt
17:19 kickerdog joined #salt
17:22 EugeneKay So I'm going to be working on some execution & state modules for NetApp systems, which have a proprietary SDK(cannot release to public). What would be the best way to release these? My code will be WTFPL(Apache-compatible), but since it'll require a proprietary SDK.....
17:23 EugeneKay I'm thinking a separate repo that potential users will have to pull in as a modul?
17:25 joehillen joined #salt
17:26 Gifflen joined #salt
17:26 EugeneKay Actually, this license might allow(with restrictions) distribution of the libs. Hm... need to read it more.
17:28 kickerdog1 joined #salt
17:32 diegows joined #salt
17:32 juso joined #salt
17:33 timoguin EugeneKay, modules have a __virtual__ function where you can determine if it's loaded or not
17:33 juso Hi guys
17:33 ajw0100 joined #salt
17:33 timoguin what'd i'd do is just check for the SDK in that function, and not load the module if the SDK isn't present
17:33 EugeneKay My question was more "should I bother sending a PR to the core salt.git" ;-)
17:33 juso quick question - is there a way to subscribe myself on the new release notification?
17:33 timoguin Yes, I would.
17:36 EugeneKay Now I get to have fun with XML.... oh joy.
17:40 jacksontj joined #salt
17:41 jdenning juso:  you might want to join the google group, new releases typically get announced there
17:41 hunter Jesscw: when you figure out that pillar question, let me know?
17:42 juso jdenning: can you please give me the link to the group?
17:42 JesseCW hunter: about to start trying random things now, I can't find anything in the documentation, all the docs show are loops
17:43 hunter keep me posted - I need that functionality as well for multi-platform nonsense
17:44 elfixit1 joined #salt
17:44 vejdmn joined #salt
17:46 jdenning JesseCW / hunter : https://gist.github.com/anonymous/d30985141962b2a1e8a9
17:46 jdenning juso: https://groups.google.com/forum/#!forum/salt-users
17:46 hunter jdenning: that looks like what I need
17:46 JesseCW jdenning: interesting, so you can't just dump it out directly
17:47 hunter I would love a solution that required no templates.
17:47 hunter Or a SLS major mode for emacs - the YAML one is .... lacking
17:47 jdenning JesseCW : If you wanted to do it without a for loop, you would have to store the formatted yaml as a string in the pillar..
17:48 jdenning (which seems more error-prone to me)
17:48 jdenning JesseCW / hunter - if you really want to avoid Jinja, I would consider one of the alternate renderers..I think the solution would be logically similar for any renderer though..
17:49 JesseCW I'm fine with jinja, I was just curious if it was possible.
17:49 hunter I figured - I need better editor support then - :)
17:49 JesseCW thanks jdenning
17:49 jdenning no prob :)
17:49 hunter thanks jdenning
17:51 morte joined #salt
17:55 saurabhs joined #salt
17:56 kalloc joined #salt
18:04 matanya joined #salt
18:09 jeremyBass1 joined #salt
18:12 thrashr888 joined #salt
18:12 jeremyBass1 hello, I'm trying to work out https://groups.google.com/forum/#!topic/salt-users/aLI2MjJ26TM .. I have my minion file with set to http://pastebin.com/vapFiG2N .. buit doens't seem to pick to the state files in the other paths and i am not sure why
18:13 rojem joined #salt
18:17 jdenning jeremyBass1 - the config file you posted on pastebin should be on the master, not the minion
18:17 martoss1 joined #salt
18:17 logix812 joined #salt
18:18 jeremyBass1 @jdenning even with a masterless?  I'm running of vagrant and i believe that it's masterless ?  not 100% on that.. but it picks up the first path, just not the ones after
18:19 jdenning oh, didn't realize you were running masterless..do the filenames overlap in the different paths?
18:20 jeremyBass1 you mean that there could be a state file top.sls in all paths listed?
18:21 jdenning I mean - does each path contain different state files (with different names)?
18:22 jdenning (I'll go read your thread on the google group - I didn't look at that yet)
18:22 jeremyBass1 no, each path has a top.sls to start with adn there are files i beleieve that may have the same name
18:23 jeremyBass1 but it acts as if the paths don't even get read or something, although i do see at one point it understands that there is the path..
18:25 vlcn can I do a prereq on an include?
18:27 jdenning jeremyBass1: I'm not 100% clear on what you're trying to accomplish - but I think that you might want to try to separate out the different file_roots into different environments - e.g. base, dev, production
18:27 jdenning also, I would try to simplify things by keeping only one top.sls file, located in the 'base' environment
18:28 carmony joined #salt
18:29 balltongu joined #salt
18:30 jeremyBass1 @jdenning the goal is not to have eveything in one place .. that is a dirty mess of a thing i would end up with.. the goal at it heart is to have the server base repo, which is provisioned on it's own, to not have any of the app repos mixed in with it..
18:30 jeremyBass1 the server git clones in the app projects, which could be 0-50 app repos, and that each repo has it's own provisioning that is needed...
18:31 ml_1 joined #salt
18:33 jdenning jeremyBass1: I would still only use a single top.sls file (or an external top), I think that is part of the issue.  Do you have your app repos structured such that each app is in an individual directory?
18:34 jdenning could you perhaps post a pastebin with the directories / filenames that you are trying to use?
18:34 CaptTofu joined #salt
18:36 CaptTofu howdy!
18:36 JasonSwindle Howdy
18:36 jeremyBass1 well you can see the files that are git cloned in https://github.com/jeremyBass/WSUMAGE-base .. one of the file_roots path is set to - /srv/store.wsu.edu/salt which is where the prosioning files are located
18:37 KyleG joined #salt
18:37 KyleG joined #salt
18:38 jeremyBass1 @jdenning i don't know how, based on the doc, i can't have a top.sls in each file root..
18:38 jdenning jerememyBass1: agreed, you can have multiple top.sls files, but when troubleshooting, it complicates the issue a bit.
18:40 jeremyBass1 well in this case i think it'd complicate the matter by trying to advoid it too.. just for what it's worth unless there is a default file that gets read other then a top.sls?
18:40 jdenning jeremyBass1: What is the directory structure of /srv/salt?  Also, I think that having both /srv/salt and /srv/salt/finalize might confuse things a bit..
18:40 SpeeR joined #salt
18:40 JasonSwindle anyone mix pillar and grains together?
18:41 jeremyBass1 @jdenning the structure of /srv/salt is https://github.com/jeremyBass/WSU-serverbase-centos/tree/master/provision/salt
18:41 jdenning so structure is identical for /srv/salt and /srv/store.wsu.edu/salt ?
18:42 JasonSwindle Like I have: {{ pillar['ssh']['private']['postgres'] }}  but I want to make it into something like {{ pillar['ssh']['private'] grains.node_type }} ?
18:42 jeremyBass1 which is the server base provisioning.. the server is to ba able to stand alone.. but the apps are to need the server.. and no .. the store.wsu.edu is .. https://github.com/jeremyBass/WSUMAGE-base/tree/master/provision/salt
18:42 jdenning jeremyBass1: oh sorry, misread the link..one sec..
18:42 scoates JasonSwindle: {{pillar['ssh']['private'][grains['node_type']]}} ?
18:42 rojem joined #salt
18:43 JasonSwindle I can test it……. }} all the things
18:43 n8n joined #salt
18:43 jslatts joined #salt
18:46 jdenning jeremyBass1: so to clarify - the state files in /srv/salt are working as expected, but you cannot access the states in /srv/store.wsu.edu/salt or /srv/salt/finalize, correct?
18:46 jeremyBass1 correct
18:46 KyleG joined #salt
18:46 KyleG joined #salt
18:47 KyleG sorry for that guys, I had to reboot. :P
18:48 jdenning jeremyBass1 - how are you calling the states? i.e. what command are you using to apply the states?
18:49 jeremyBass1 @jdenning well i thought that it just grabed them since i declared the paths?
18:50 jdenning jeremyBass1 - I'm wondering - are you using a command like 'salt '*' state.highstate' to apply all states at once, or are you invoking individual sls files?
18:50 kermit joined #salt
18:50 jssjr so whats going on at saltconf? what are people excited about?
18:51 Whissi joined #salt
18:51 jeremyBass1 @jdenning hmm.. unsure, this is vagrant and it uses salty-vagrant, which is now included in vagrant.. so all i do to start is declare it here https://github.com/jeremyBass/WSU-serverbase-centos/blob/master/Vagrantfile#L188
18:53 jdenning jeremyBass1 - Well, I don't use salty-vagrant so I can't be any help there..
18:53 JasonSwindle scoates:  That worked, just looked odd to me.  Thanks!
18:53 scoates does it make sense?
18:53 scoates you're using the value from the grains dict as a key in the pillar dict
18:55 jeremyBass1 well as far as i know there should be any diff there right? I mean that it's by default a `salt '*' state.highstate` if i read the docs right.. right?
18:56 jdenning jeremyBass1: I would strongly suggest that you try to simplify everything down to standard, minimal install, make sure it is working, and build from there…your directory structure is a bit non-standard, and it is hard to tell exactly where the errors are occurring from the info given..
18:56 scott_w joined #salt
18:57 JasonSwindle scoates:  It does now
18:57 scoates cool
18:57 JasonSwindle I moved the spacing around a bit, and that works great.
18:57 JasonSwindle Going thru my stack and make it DRY
18:58 jeremyBass1 @jdenning well the base works well .. i just need to call in a pillar and state file from a different path then the root path that the base is working under.. i guess I'm at a loss of why the paths declared are not being respected
18:58 jdenning jeremyBass1: for example, you don't have any init.sls files, typically you want at least a blank init.sls file in each directory (similar to a python module's __init__.py), but you don't *have* to do this if you are calling sls files directly
18:58 jeremyBass1 should there be an init.sls and a top.sls is what your saying?
18:59 jdenning if you are trying to use state.highstate calls, there should probably be an init.sls
18:59 jeremyBass1 hmm.. ok.. well I'll add that to the app folders and see if that helps..
19:00 jeremyBass1 and they can be blank?  the init.sls files i mean?
19:00 Ahlee if you restart a master, when the minions reconnect they'll still call anything in the reactor, right?
19:00 KyleG cedwards: Are you around? I have some questions about some FreeBSD oddness
19:00 cedwards KyleG: sure
19:00 KyleG Can I PM you?
19:01 cedwards yeah
19:01 N-Mi joined #salt
19:02 jdenning jeremyBass1: Yes, init.sls can be blank (for namespacing purposes only)..you should maybe take a look at the saltstack-formulas repos (https://github.com/saltstack-formulas), this is a more standard solution to keeping states for various apps separate..
19:04 jdenning sorry I couldn't be more help!
19:04 jacksontj joined #salt
19:05 jeremyBass1 yeah i looked at those.. they all seem to lead to the same drity end was the problem.. and i mean dirty in the sense that the app state files end up in the base .. to me that is very dirty when it comes to managing a server base and like 0-50 apps the could be provisioned with the server.. :-\ .. but I'll keep at it..
19:09 kalloc joined #salt
19:10 [diecast] joined #salt
19:13 jeremyBass1 well a blank init.sls didn't seem to work..
19:13 jeremyBass1 :-\ .. salt is good if it's in a striaght line only.
19:13 jeremyBass1 well I'll try again later.. must run
19:13 jeremyBass1 thank you for the help
19:13 jeremyBass1 cheers
19:13 jeremyBass1 left #salt
19:15 mgw joined #salt
19:18 amckinley joined #salt
19:19 tmwsiy do most folks just keep their package binaries and sources together with their sls files in the state tree or do they setup separate http servers? Is there any disadvantage to putting a bunch of binaries and stuff of that nature in git?
19:23 scoates I put binaries into S3 (a custom apt repository), and source in git.
19:25 felixhummel joined #salt
19:27 druonysus joined #salt
19:27 timoguin the binaries i'm currently using i just pull down via HTTP
19:28 timoguin but I plan on using S3
19:30 lynxman joined #salt
19:33 ekristen whats the command to update my gitfs remotes manually
19:35 timoguin I think saltutil.sync_all is supposed to do it
19:35 ekristen k
19:35 ekristen can pillars provide files for states to use?
19:35 ekristen or does the file have to be in the states folder?
19:35 quickdry21 joined #salt
19:35 timoguin don't think pillar can do files.
19:36 [diecast] pillar is for configs of states really. you can use grains for files
19:36 ekristen well I was using pillar to define my users
19:36 timoguin grains for files?
19:36 ekristen it’d be nice to package the public key with them
19:36 timoguin ekristen, you can define the public key in the user pillar
19:36 timoguin that's what i do
19:37 [diecast] ya, just use a string in pillar
19:37 ekristen timoguin: can you share an example please
19:37 timoguin https://github.com/saltstack-formulas/users-formula/blob/master/pillar.example
19:37 timoguin babammm
19:37 timoguin just like that
19:37 ekristen ah
19:38 derelm joined #salt
19:39 Ryan_Lane2 joined #salt
19:40 ekristen guess I should just use that formula
19:41 ekristen do you all normally fork repos and use your forked formula?
19:41 jdenning joined #salt
19:41 timoguin I think that's what most people do, but I haven't seen any real stats on that.
19:42 timoguin I just clone them and them copy-paste into my single git repo for states.
19:42 ekristen timoguin: yeah that is what I was thinking of doing
19:42 timoguin They don't change that much from what I've seen.
19:42 ekristen unfortunately you can’t submodule them because of how they were designed
19:42 timoguin I just view them as kinda a guide
19:42 timoguin once my users are being provisioned i'm not really gonna need to change it much
19:43 zach Anyone familiar with fail2ban? I'm going to configure it to do salt '*' iptables.[append|delete] <rule>
19:43 zach Or should I just write a module specifically for it?
19:46 salt_whitebelt can saltstack or the enterprise version manage windows azure?
19:47 Mua joined #salt
19:48 timoguin salt_whitebelt, not quite yet. i've been told they're working on that though.
19:48 timoguin i've been wanting to take a stab at it myself, but haven't had time yet
19:51 salt_whitebelt i know azure is just a sql plaform with all the ms dev tools builtin and since it uses REST API for management i wonder if there need to be a REST API module for that or we can use cmd.run to call REST task?
19:53 ajw0100 joined #salt
19:55 troyready joined #salt
19:56 quantalrabbit joined #salt
19:56 clearclaw ekristen: My default process is always: fork, customise, submit pull-request.
19:56 vejdmn joined #salt
19:56 timoguin salt_whitebelt, azure has a python SDK, so I think that'd be the best bet
19:57 clearclaw ekristen: Also don't forget the | syntax for jinja for embedding multiline strings ala here-files.
19:57 juso jdenning: had a look at the room - it's far to active to be used as a new releases notificator, 5-10 posts a day is too much if I only want to know when a new release is out. thanx anyway, but will probably use manual monitoring of the salt site.
19:58 clearclaw jdenning: Oooooh, so happy about the Dynamo pillar thang...
19:58 jdenning clearclaw  :)
19:59 jdenning yeah, let me know if you hit any issues - it has been working pretty well for me so far
20:00 jdenning juso: I just subscribe to the email digest, so you get one email a day with all the new thread topics, then you can scan for anything important (like new releases, vulnerabilities, etc), and click the link if it's relevant
20:00 clearclaw jdenning: I'd love to get to it today, but I have a full site push to plan and orchestrate with a company that's never done that before.  I should be able to poke at it strting mid-next week....unless we get caught in rollback hell.
20:00 zach Who wrote the iptables module?
20:01 clearclaw None of which counters the value of having a -releases mailing list and/or RSS feed.
20:01 jdenning agreed ^^^
20:02 clearclaw scoates: How do you handle auth for your S3-based apt repos?
20:02 scoates clearclaw: via IAM + AWS policies
20:03 ekristen can a pkg.installed source be a http resource instead of salt:// resource?
20:03 scoates clearclaw: and this https://github.com/castlabs/apt-s3
20:04 scoates ekristen: http://docs.saltstack.com/ref/states/all/salt.states.pkg.html says yes
20:06 vlcn is there some way I can use a prereq that would result in not throwing an failure message if the prereq isn't met?
20:07 njpatel joined #salt
20:07 kermit joined #salt
20:08 clearclaw scoates: Ahh.  is there a decent way to manage per-host based authorisation in IAM?
20:09 clearclaw It seems such an utter pain.
20:09 scoates we use the same keypair for all of our hosts.
20:09 * clearclaw nods
20:09 clearclaw I'm a bit less keen on having the sme keypair for PROD as I do for DEV.
20:09 forresta vlcn, I don't think so, you might try doing some lame workaround using jinja stuff.
20:09 darrend hi all.. I asked this a couple of days ago but I don't think anyone was around :)  Anyone aware of a project to enable pillar data to be read from a zookeeper cluster?
20:09 forresta vlcn, but I haven't tried that
20:10 vlcn forresta: thats what I figured, I'm not finding a clear jinja example either, but I'll keep at it.
20:10 forresta yea I can't think of a good jinja example like that off the top of my head.
20:10 scoates clearclaw: it's read only. I'm not worried.
20:11 dave_den darrend: i am not aware. you may want to ask on the salt-users google group
20:11 forresta agreed with dave_den, I looked and couldn't find anything on it either.
20:11 thrashr888 joined #salt
20:12 clearclaw Anybody got any ideas about why I can't highstate boxes that I build with an RCS statefile but I can highstate them if I build them directly?
20:12 darrend dave_den: I'll try there.  We may write something if there isn't already a project on the go somewhere.
20:12 darrend thx
20:12 clearclaw (see mail on salt-users)
20:14 mbarrien joined #salt
20:14 martoss joined #salt
20:15 jslatts joined #salt
20:17 martoss1 joined #salt
20:20 hunter I need to wire etcd -> pillar sometime myself
20:20 srage joined #salt
20:24 JasonSwindle Howdy hunter !
20:24 hunter hello JasonSwindle
20:25 Damoun joined #salt
20:26 vejdmn joined #salt
20:26 hunter anyone know why pkg.purge (.remove) does not work with yum?
20:27 hunter (at least its documented not to work - I guess I could try it...)
20:28 hunter Its a doc error
20:29 hunter Any devs here - how do I file a doc bug?
20:31 cachedout hunter: Just file a regular GitHub issue and it will get forwarded to the correct developers. Or, feel free just to fix the doc error and submit a pull request.
20:31 JasonSwindle hunter:  ^ that
20:31 JasonSwindle They are REALLY friendly on PRs
20:31 hunter My git fu is too weak - I'll file one
20:32 JasonSwindle You can edit files right in github now if an easy fix
20:32 JasonSwindle and it auto fork and PR and magic
20:32 JasonSwindle but issue works. :)
20:33 bemehow joined #salt
20:36 vejdmn joined #salt
20:48 felixhummel_ joined #salt
20:51 salt_whitebelt i have a powershell script in /root/saltscripts/pshell.ps1. on master (centos) im trying to run the following on my target minion (win7): salt windows.vm cmd.run 'powershell.exe /root/saltscripts/pshell.ps1'. it comes back with pshell.ps1 is not recognized as the name of a cdmlet
20:53 scoates salt_whitebelt: I don't know anything specific about powershell, but I'd guess that you have to copy that file to your minion (with file.managed) and then run it
20:53 Teknix joined #salt
20:54 salt_whitebelt tried running as : salt windows.minion cmd.run /root/saltscripts/pshell.ps1 shell=powershell' and it comes back with the system cannot find the path specify
20:54 Mua_ joined #salt
20:55 Mua_ joined #salt
20:55 mbarrien joined #salt
20:55 salt_whitebelt im trying to have a central pshell script dir on the master (centos) to carry out remote powershell execution but the script is physically reside on the master not minion. is this the right way of calling it or do i need to put the powershell script on the minion in order for it to run?
20:57 cewood joined #salt
20:57 clearclaw What was the feedback again on having cloud.profiles and cloud.providers processed as jinja templates?
20:58 thrashr888 joined #salt
20:59 psyl0n joined #salt
20:59 salt_whitebelt is file.managed and cmd.run are modules? if they are how come its not on the list of builtin module on salt site?
20:59 scoates salt_whitebelt: your minion can't see files on the master. you'll need to copy the file first.
21:00 salt_whitebelt scoates: what about cmd.script?
21:01 salt_whitebelt what is the difference between all this anyway? cmd.run. cmd.script, file.managed?
21:02 scoates file.managed manages a file on the minion, cmd.run runs an arbitrary command. I hadn't used cmd.script before, but it looks like it might be what you need.
21:02 timoguin salt_whitebelt, there are modules and there are states. file.managed is a state.
21:03 hunter Issue #9760 filed
21:03 timoguin kinda confusing because states are modules too.
21:03 mgw joined #salt
21:04 hunter I just hit that - whats the difference?
21:04 hunter (state vs module)
21:04 timoguin modules are typically executed from the CLI
21:04 cachedout States describe the intended state or configuration of a thing. Execution modules are the workers which run to achieve that state.
21:04 timoguin states are typically used in SLS files
21:04 timoguin what cachedout said
21:05 timoguin for example, file.append is an execution module that will append text to a file. file.managed is a state that will ensure the presence of the file.
21:06 jeremyBass joined #salt
21:06 timoguin a lot of the states just wrap the execution modules
21:07 jeremyBass hello back for another round :)
21:07 vejdmn joined #salt
21:08 timoguin salt_whitebelt, what I would do is use file.managed to put the script on your minion, and then cmd.run to execute it.
21:08 hunter cachedout: thanks - that makes sense
21:08 ekristen timoguin: do you gitfs for all your states?
21:08 timoguin I do
21:09 ekristen do you branches and environments?
21:10 timoguin not yet. i'm just using a single branch for now.
21:10 timoguin but soon hopefully
21:11 timoguin I tried it out for a minute but ran into "Conflicting SLS ID" errors, so I just scraped the idea for the time being.
21:11 timoguin Not enough time to mess with it
21:11 forresta timoguin, are you on 0.17.4?
21:11 timoguin yes
21:12 forresta ok, here's the issue regarding that conflicting ID issue: https://github.com/saltstack/salt/issues/7526
21:12 forresta when you have time to read through it
21:13 timoguin thanks, i've seen that one, but haven't dug in yet
21:14 forresta yea the TLDR is that the issue exists on 0.17.x, but is fixed with some pydsl issues.
21:14 forresta for the RC
21:15 timoguin I'm not sure if my problem was due to includes or not. I had things working fine with a single master branch
21:15 timoguin but then create staging, qa, and dev branches
21:15 timoguin then mapped appropriately in the top.sls
21:15 timoguin and it blew up
21:15 forresta ahh
21:15 timoguin removed all the other branches and it worked
21:16 timoguin i read it as saying i couldn't have the same SLS ID in the dev branch as in another branch
21:16 thrashr888 joined #salt
21:17 timoguin which didn't make much sense to me, since doing it any other way would mean i couldn't easily merge dev into qa etc.
21:24 ajw0100 joined #salt
21:25 vejdmn1 joined #salt
21:27 Nazzy woo, more bugs
21:28 saurabhs left #salt
21:29 Nazzy now I just need to figure out if this is my fault or a real bug
21:30 Nazzy ok, this one needs reporting ... it's logging an error that isn't an error
21:30 Gordonz joined #salt
21:32 tmwsiy ok I think I am on the verge of an aha with respect to states. on the top file doc page (http://docs.saltstack.com/ref/states/top.html) In the second code box it references "core" and "edit" are these sls files with this name? or a directory with that name with an init.sls?
21:32 tmwsiy or something else that I am missing?
21:33 forresta tmwsiy, so they could be files under /srv/salt/{core.sls, edit.sls}
21:33 forresta or /srv/salt/{core,edit}/init.sls
21:34 jimallman joined #salt
21:35 jeremyBass hello, so i have updated my troubles on https://groups.google.com/forum/#!topic/salt-users/aLI2MjJ26TM .. i'm left woundering why i can't get the states from the other folders to run.. Jnina seems to see them
21:35 jeremyBass would anyone be able to help?
21:36 hunter joined #salt
21:37 derelm joined #salt
21:37 hunter Can salt take an existing machine and generate a (big) state that would duplicate it (question from one of my colleages)
21:38 jeremyBass @hunter blueprint does that .. reverse a server.. iirc
21:38 Nazzy now to figure out the cause of the awesomely detailed "InvalidAtom: #"
21:39 mbarrien Question about external pillars... if I were writing code for an external pillar (not a state file, not a pillar file, an actual python module) and want to deploy it to my salt master, must I store it in a local file directory on the master (pointed to by extension_modules), or is there any way to store it as a module in Saltgitfs? My understanding is the the _modules directory in gitfs is only for minion modules, but this is a module for t
21:39 forresta hunter, there is this app called rock something
21:39 hunter jeremyBass: url?
21:39 hunter forresta: url?
21:39 forresta hunter, I'm trying to remember the nam
21:39 forresta *e
21:40 hunter (googling for "blueprint" or "blueprint software" isn't helpful)
21:40 jeremyBass @hunter , sorry http://devstructure.com/blueprint/ .. but it looks like it may only do chef and puppt
21:40 tmwsiy so if I have a ok so I have my sls files and locations here: http://pastebin.com/C9qmzDv4
21:41 tmwsiy when I run a #salt 'VM*' state.highstate the packages is not matched to be installed
21:41 harobed_ joined #salt
21:42 harobed_ joined #salt
21:43 jeremyBass so would some one be able to help figure out why jnina can log that it's seeing the file_roots i set but salt is only running the states from the first file_root?
21:44 forresta I found it hunter, ok so this app: https://www.scriptrock.com/
21:44 forresta I talked to the CTO about it when I was at devopsday atlanta
21:44 forresta he said they could easily put salt support in if people wanted it
21:45 forresta hunter, whether they actually did so after I spoke with him, I have no idea.
21:46 hunter I'll look - I'm not sure how useful that would be but said I would ask.
21:46 forresta yea, I have no investment or anything in that company
21:46 forresta they just said they could duplicate a machine and build puppet code from it, so I asked about salt.
21:49 ConceitedCode joined #salt
21:49 mbarrien Any feedback about my ext_pillar question?
21:50 tmwsiy seems like you could do similar things with salt if you use mining and keep your configs in source control
21:51 forresta mbarrien, as far as I understand _modules is specifically for modules
21:51 Nazzy grrrr, state.show_sls does nothing for pydsl ¬_¬
21:52 jcockhren Nazzy: you doing pydsl?
21:53 jcockhren that's two of us! \o/
21:53 Nazzy jcockhren, yeah heh
21:53 Nazzy I'm trying to figure out why I'm getting an error that "#" is an invalid atom
21:54 jcockhren Nazzy: gist it. it's probably a spacing issue
21:55 jeremyBass so if i have a  Jinja search path: that includes the folder i need to have the state file run from .. but they are not.. if jnina see the folders shouldn't they run?
21:55 xmltok joined #salt
21:56 Nazzy jcockhren, if it were that easy I'd have found it by now heh
21:56 Nazzy jcockhren, best I can tell, pkg is borked in some way
21:57 yomilk joined #salt
21:57 Gifflen Anyone know if there is anything posted in terms of things that will be tested on the SSCE?
21:57 Nazzy state().pkg.installed('net-misc/ntp') results in
21:57 Nazzy File "/usr/lib64/python2.7/site-packages/portage/dep/__init__.py", line 1255, in __init__
21:57 Nazzy raise InvalidAtom(self)
21:57 Nazzy InvalidAtom: #
21:57 Gifflen Wanted to make sure I am spending my time studying the correct things before saltconf
21:58 Nazzy I seem to remember seeing something about installed on the issues tracker though
21:59 Nazzy nope, just mine ... rats
21:59 ahwitz joined #salt
21:59 ahwitz Quick question - can salt state command line calls take parameters?
22:00 ahwitz long story short, I need to pass a specific username in to install software under.
22:00 jeremyBass when declaring more then one file_root, is there some trick to get the other path's states to run?
22:00 Nazzy ahwitz, clarify "command line calls" ... you mean triggering salt from the command line or making calls from inside a state?
22:00 ahwitz Nazzy: triggering salt, sorry.
22:01 Nazzy there's probably a way
22:01 Nazzy though I'm not entirely certain how ... via env I'd expect
22:02 ahwitz Nazzy: closest thing I can see in the manpages is the --args-separator option, which only seems to be salt-specific arguments (rather, those that are directly related to which minions are running which states)
22:03 ahwitz my backup solution is to just write in "$USER" or something into the state files and run a python script that takes the desired username as a parameter and temporarily replaces the $USER token with a desired username, runs the command, then changes them back
22:03 kainswor It looks like there is some logic that attempts to run ext_nodes on top calls from remote execution- but for me I'm getting nothing from "target# salt-call state.show_top", where I get expected results from "master# salt target state.show_top"- any ideas?
22:03 Nazzy ahwitz, if you call state.sls from the command line you might do "state.sls test=True" ... my guess is that you can pass in arbitrary params the same way and they'll be available via a variable
22:04 ahwitz thanks! will try that
22:05 rojem joined #salt
22:07 ajw0100 joined #salt
22:08 jeremyBass as i read it here.. http://docs.saltstack.com/ref/file_server/file_roots.html .. it seems to me that everything referenced here https://groups.google.com/forum/#!topic/salt-users/aLI2MjJ26TM  should be working. that the states should be going?  .. or is it that the top in the first folder is what is setting everything and that is why the other folder'd states never are seen?
22:11 g3cko joined #salt
22:11 Nazzy urg, I know this error
22:11 jcristau joined #salt
22:11 Nazzy comments in /etc/portage/ files
22:15 analogbyte is there some kind of settings I can access from within a runner? I dont see how to access pillar or master config values in it and don't want to hardcode
22:16 Nazzy __salt__ or equiv?
22:16 alunduil joined #salt
22:17 gadams999 left #salt
22:19 abe_music joined #salt
22:20 analogbyte Nazzy: only yields this: KeyError: 'config.option' can you point me to some docs, maybe I am using it wrong?
22:22 clearclaw Does anyone here use a tool to manage AWS security groups?  Mine are getting out of hand...dozens of the things....
22:22 Nazzy analogbyte, depends where you're accessing from
22:22 analogbyte Nazzy: I checked what it contains in this context, it seems to hold the names of all runners... not exactly useful
22:23 analogbyte Nazzy: yeah, thats why I am seeking some reliable info on it, monkeyprinting its contents every time is kind of not to the point...
22:24 Nazzy analogbyte, I would guess you can use the same logic as in py rendered states: http://docs.saltstack.com/ref/renderers/all/salt.renderers.py.html#module-salt.renderers.py
22:24 martoss joined #salt
22:25 oz_akan_ joined #salt
22:30 oz_akan_ joined #salt
22:34 scott_w joined #salt
22:34 jfzhu_us joined #salt
22:36 givernao joined #salt
22:36 givernao hi all
22:38 salt_whitebelt running: salt windows.minion cmd.script  salt:/root/saltscripts/pshell.ps1 shell='powershell' -l debug : is showing missing  configuration file /root/.salt . this command is suppose to run the transfer the powershell script from master (centos) to windows minion win7) and run it.
22:39 david_a joined #salt
22:42 rojem joined #salt
22:42 alunduil joined #salt
22:43 martoss joined #salt
22:44 martoss1 joined #salt
22:44 fllr joined #salt
22:44 * Gareth attemps to install salt to a synology nas.
22:44 forresta Gareth, that would be cool
22:45 Gareth forresta: curious if I can use it as a salt master.
22:45 forresta heh
22:45 jasonrm joined #salt
22:46 Gareth Its just Linux under the hood so it should just work :)
22:48 analogbyte Nazzy: thank you very much, your guess is partly correct: via opts I gain access to the masters config, the rest are either empty or None ;) so thank you very much, the masters configuration is enough for me ;)
22:49 Nazzy analogbyte, no problem :)_
22:51 jimallman joined #salt
22:51 fwiles joined #salt
22:52 fwiles Is it possible to execute only certain stanzas in an sls from the commandline?
22:52 clearclaw I believe a statefile is always treated as a single meta-state.
22:52 fwiles Ok so I need to break this apart.  That's what I was thinking.
22:52 Gifflen_ joined #salt
22:53 clearclaw IOW if you want internal/smaller states, split up your statefile and use includes to join them back up again.
22:53 * clearclaw nods
22:53 kalloc joined #salt
22:53 * clearclaw might be wrong.
22:53 fwiles I did a quick and dirty test and something broke so thought I'd ask before I investigated further :)
22:55 [diecast] joined #salt
22:57 elfixit1 joined #salt
22:57 bhosmer_ joined #salt
22:57 yomilk joined #salt
22:58 kalloc_ joined #salt
23:00 jeremyBass salt will not run each top.sls from each file_root will it?
23:01 steve1 left #salt
23:01 clearclaw It should run the topmost/first only.
23:01 jeremyBass topmost?
23:01 clearclaw In your stack of file roots.
23:03 jeremyBass oh you mean that it only reads the first one?
23:03 oz_akan_ joined #salt
23:03 clearclaw I think it only reads the first one it finds.  its like a stackable file system.
23:03 sulky joined #salt
23:04 gw joined #salt
23:04 bemehow_ joined #salt
23:05 gw Hi, I am trying to understand environments, pillars and state inheritance, and was wondering whether you can use 'include:' only with states or with pillars as well?
23:05 jeremyBass oh .. well hell.. ok.. so i want to read state file from a folder that is not in the base.. with out having some evn.. is that what the init.sls does? or is that also only the first one is read?
23:06 cdcalef joined #salt
23:06 cdcalef how do you escape ' in yaml files?
23:06 ekristen I have my top.sls in my git repo, but it doesn’t seem to be working
23:07 jeremyBass why is it so hard to read state file from two locations
23:07 ahammond hunter thanks for the info.
23:07 gw cdcalef: http://yaml.org/spec/current.html#id2534365 just double the single quotes within a single quoted string
23:17 ilbot3 joined #salt
23:17 Topic for #salt is now Welcome to #salt - SaltConf Jan 28-30, 2014! http://saltconf.com (reg deadline January 3) | 0.17.4 is the latest | 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/
23:18 lahwran joined #salt
23:18 jasonrm joined #salt
23:18 [diecast] joined #salt
23:18 EugeneKay joined #salt
23:18 seanz joined #salt
23:18 fllr joined #salt
23:18 fxdgear joined #salt
23:18 fivethreeo joined #salt
23:19 jeremyBass ok maybe I have been asking my question wrong.. how can i run states that i will not know of till run time.. ie i can't put them the base top.sls , i will only know where the state files will live..
23:19 JasonSwindle joined #salt
23:20 clearclaw salt \* state.sls state_file_here.substate
23:20 martoss joined #salt
23:20 gw another question. is there a way to have e.g. some pillars with base data, and override single values (also in deeper nested structures) in a different environment?
23:21 vejdmn joined #salt
23:21 jeremyBass @clearclaw .. by chance was that to me?
23:22 kainswor jeremybass: you can define a custom tops on the master which can dynamically generate top content. but it has a unique format (read tops/ext_node.py for an example). also there might be issues with minion-based execution
23:23 hunter Whats the syntax to have on state file require another state file? Our foo daemon requires the go compiler for instance - i already have a go state.
23:23 jeremyBass @kainswor where is this tops/ext_node.py?
23:23 kainswor hmm let me double check the path
23:23 bemehow_ joined #salt
23:23 hunter parsing that sentence I'm almost certainly using the word state incorrectly. Module?
23:24 kainswor https://github.com/saltstack/salt/blob/develop/salt/tops/ext_nodes.py
23:24 kainswor and look up the master_tops option in the master config to configure it
23:24 snave joined #salt
23:24 Nazzy how do you reopen an issue on github?
23:25 jeremyBass hmm.. interesting.. so i need to make a speical thing to read from a path that will exist but doesn't at the start of a run? .. but well brb.. i have to grab my kido from school.. thank you for the help
23:26 kainswor well you'll have to know where your state is going to be at the start. that can specify an env
23:26 snave is there a way in targeting to target all but one host?  something like '!host7'
23:28 austin987 joined #salt
23:29 kalloc joined #salt
23:31 snave found it... just in case anyone is looking for it a leading not is not supported.  Have to use this:  salt -C '* and not host7*' test.ping
23:33 Nazzy https://github.com/saltstack/salt/issues/9767 <--- *sigh*
23:34 markizano joined #salt
23:35 Corey Poor Nazzy.
23:35 Ryan_Lane2 joined #salt
23:35 forresta Nazzy you're seeing that on the RC?
23:35 Nazzy forresta, yea, sadly
23:36 Nazzy or rather, I'm seeing it on develop
23:36 forresta gotcha, even though you sort of mention that you didn't see it on 0.17.4, you might want to explicitly mention it's part of develop/the RC
23:36 forresta ok cool, make sure to mention that
23:39 Nazzy I'm actually seeing it on 0.17.4 as well it seems
23:44 Nazzy version info added
23:47 bhosmer_ joined #salt
23:48 hunter any help on salt file deps?
23:49 srage joined #salt
23:49 ajw0100 joined #salt
23:51 bemehow joined #salt
23:53 hunter my reading at http://docs.saltstack.com/topics/tutorials/states_pt3.html implies that one SLS module "includes" another to imply a requirement  - I'm a little surprised it isn't something more like require:\n state: foo or somesuch.
23:53 monokrome joined #salt
23:53 zzzirk joined #salt
23:54 markizano hunter: what are you trying to do again ?
23:56 hunter I'm creating a state file for our "shock" service - it relies on mongodb which I have an SLS file for.
23:56 hunter How do I tell salt that before shock.sls can go it needs mongodb.sls?
23:57 hunter My reading of the url above implies "include" but that seems a funny way to do module level prereq's.
23:57 honestly you need include first
23:57 honestly and then require
23:57 honestly without include the mongodb states won't be in scope
23:58 honestly include doesn't imply requirement
23:58 hunter What does the require statement look like for this?
23:59 abe_music joined #salt
23:59 hunter (I've enjoyed skipping most of the require ordering stuff since 0.17 lets you assume in YAML order for stuff in a single file.)

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