Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-07-22

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

All times shown according to UTC.

Time Nick Message
00:02 kryptt left #salt
00:03 jamesbdev joined #salt
00:08 robawt can I overrule a config file twice?
00:08 robawt I make a state, extend the state to point to a source, then extend that state to point to a different source (much like overloading in OO?)
00:09 mgarfias_ what did ssh_username change to in helium?
00:13 tyson_ joined #salt
00:21 badon joined #salt
00:24 chrisjones joined #salt
00:29 gzcwnk when I run a state command on '*' I am getting "did not return in time" for some servers yet when I target one of teh failed ones the state runs fine, any ideas why?
00:30 stevednd just because it didn't return in time doesn't necessarily mean it didn't run
00:30 stevednd the response from that machine just timed out
00:30 joevartuli joined #salt
00:30 gzcwnk yeah I suspect so, some seem to have, but others not
00:31 stevednd you can use salt-run jobs.lookup_jid <jid> to get the output of the run
00:31 stevednd it will include the other machines
00:31 stevednd well specifically it will include the entirety of the output from all machines
00:32 gzcwnk can i increase that timeout somehow? it seems to be a few econds only
00:32 stevednd pass -t TIMEOUT
00:32 stevednd it defaults to 5 seconds
00:33 stevednd if you want to up it permanently just change the setting in the master config
00:33 stevednd but don't confuse that timeout with the time it should take the entire script to run
00:34 stevednd while actions are running the minion typically reports back saying it's working on something, so the salt command then waits. Sometimes the minion gets hung up for whatever reason doing some particularly intense task, and *then* the timeout happens
00:34 hobakill joined #salt
00:35 pclermont joined #salt
00:35 joevartuli left #salt
00:35 jamesbdev joined #salt
00:36 ajprog_laptop1 joined #salt
00:39 canci joined #salt
00:39 hobakill anyone come across this when trying to do a /S install in windows packages?  Command '""  /S ' failed with return code: 1
00:39 hobakill output: '""' is not recognized as an internal or external command,
00:41 robawt ok so apparently you can only extend once
00:41 * robawt shrugs
00:49 beneggett joined #salt
00:51 pclermont anyone know how I could replicate the behaviour of master_tops for a masterless minion?
00:51 Ryan_Lane pclermont: have you tried using that option?
00:51 Ryan_Lane some master configuration options also work for masterless minion
00:52 pclermont Ryan_Lane: I have, but state.show_top would come back blank
00:53 yomilk joined #salt
00:53 * Ryan_Lane nods
00:55 Guest73960 joined #salt
00:58 yomilk joined #salt
01:06 DaveQB joined #salt
01:08 scoates joined #salt
01:10 dude051 joined #salt
01:11 dude051 joined #salt
01:14 acabrera_ joined #salt
01:20 oz_akan joined #salt
01:24 sectionme joined #salt
01:25 scoates joined #salt
01:34 otter768 joined #salt
01:37 joehillen joined #salt
01:38 tyson_ joined #salt
01:40 ccase joined #salt
01:41 Outlander joined #salt
01:54 aquinas joined #salt
02:04 ghartz joined #salt
02:05 joevartuli joined #salt
02:10 meganerd joined #salt
02:12 sectionme joined #salt
02:14 TyrfingMjolnir joined #salt
02:18 aherzog joined #salt
02:27 acabrera_ joined #salt
02:41 n8n joined #salt
02:42 joevartuli joined #salt
02:45 TyrfingMjolnir joined #salt
02:45 ckao joined #salt
02:45 joevartuli joined #salt
02:50 joevartuli left #salt
02:52 dude051 joined #salt
02:53 dude051 joined #salt
02:55 bhosmer joined #salt
02:57 jujube Hi guys,  I am using Salt Cloud and am provisioning a cloud server on Rackspace.  I just want to know if it is possible to also add block storage.
03:08 ajolo joined #salt
03:11 bhosmer joined #salt
03:16 jamesbdev joined #salt
03:18 oz_akan joined #salt
03:20 ramishra joined #salt
03:24 sectionme joined #salt
03:31 arthabaska joined #salt
03:42 TyrfingMjolnir joined #salt
03:46 dave_den joined #salt
04:01 ml_1 joined #salt
04:03 joevartuli joined #salt
04:08 m1crofarmer joined #salt
04:08 mosen joined #salt
04:09 joevartuli joined #salt
04:19 oz_akan joined #salt
04:19 TyrfingMjolnir joined #salt
04:21 oz_akan_ joined #salt
04:28 TyrfingMjolnir joined #salt
04:35 beando joined #salt
04:45 bhosmer joined #salt
04:51 tyson_ joined #salt
05:01 sieve joined #salt
05:04 TyrfingMjolnir joined #salt
05:04 TheThing joined #salt
05:05 TyrfingMjolnir joined #salt
05:10 yomilk joined #salt
05:21 oz_akan joined #salt
05:28 greyhatpython joined #salt
05:28 greyhatpython joined #salt
05:30 Sauvin joined #salt
05:48 tyson_ joined #salt
05:50 TheThing joined #salt
05:52 sieve joined #salt
05:52 tyson__ joined #salt
05:57 q1x joined #salt
06:00 roolo joined #salt
06:02 joevartuli left #salt
06:06 tyson_ joined #salt
06:11 m1crofarmer any reason a specific ID would not be applied during a highstate, but gets applied in state.sls?
06:11 saravanans joined #salt
06:11 vu joined #salt
06:12 m1crofarmer more specifically, a single ID in a state appears to be deliberately skipped over in highstate, doesn't even appear in debug, but when called directly via state.sls it runs just fine
06:14 picker joined #salt
06:18 ndrei joined #salt
06:22 oz_akan joined #salt
06:36 vu joined #salt
06:45 aquinas joined #salt
06:46 tyson_ joined #salt
06:54 kermit joined #salt
07:03 sieve joined #salt
07:04 saravanans joined #salt
07:04 ndrei joined #salt
07:09 ml_1 joined #salt
07:13 matthiaswahl joined #salt
07:17 kiorky joined #salt
07:17 muzammil joined #salt
07:18 chiui joined #salt
07:19 intellix joined #salt
07:22 alanpearce joined #salt
07:25 rawzone joined #salt
07:31 giannello joined #salt
07:34 saravanans joined #salt
07:40 darkelda joined #salt
07:45 taterbase joined #salt
07:46 ghartz joined #salt
07:48 Lomithrani joined #salt
07:52 cDR__ joined #salt
07:55 vu joined #salt
07:55 topochan joined #salt
07:56 oz_akan joined #salt
08:01 Nexpro1 joined #salt
08:01 linjan joined #salt
08:04 q1x joined #salt
08:05 TyrfingMjolnir joined #salt
08:11 sectionme joined #salt
08:14 sieve joined #salt
08:27 muzammil new to salt-stack, i am trying to test up a dev env for salt module testing, i am enabled the range_server (localhost) option but when i give a cluster name it fails and gives me error "No minions matched the target. No command was sent, no jid was assigned."
08:27 muzammil $ sudo salt -R %cb.1 test.ping
08:28 muzammil how can i test if salt is able to communicated with range properly
08:33 CeBe joined #salt
08:41 derito joined #salt
08:43 greyhatpython joined #salt
08:46 felskrone joined #salt
08:48 agend joined #salt
08:49 derito Hello there. I have a little problem with the archive.extracted state. I tried using an example from Salt doc (http://docs.saltstack.com/en/latest/ref/states/all/salt.states.archive.html). However when calling it I get "State archive.extracted found in sls tarTuto is unavailable". Any idea why this could be happening ? Thanks.
08:55 Micromus_ So I install a package, and use service: running to start it, but service: running is applied before the package is installed, and fails, how to fix the order of execution?
08:56 Micromus_ found it i think, adding require: pkg: package to all files and the service: def
08:57 oz_akan joined #salt
09:03 ramishra joined #salt
09:04 TheThing joined #salt
09:07 giantlock joined #salt
09:08 jhauser joined #salt
09:09 yomilk joined #salt
09:23 TheThing joined #salt
09:23 oz_akan joined #salt
09:31 sverrest joined #salt
09:32 Sypher joined #salt
09:41 jalaziz joined #salt
09:43 che-arne joined #salt
09:47 ggoZ joined #salt
09:50 ml_1 joined #salt
09:54 eofs joined #salt
09:54 lopez joined #salt
09:54 ashb joined #salt
09:54 Spark joined #salt
09:54 hotbox joined #salt
09:54 sindreij joined #salt
09:54 __alex joined #salt
09:54 jolo joined #salt
09:54 Flusher joined #salt
09:54 jolo joined #salt
09:55 thehaven joined #salt
09:55 phx joined #salt
09:55 mariusv joined #salt
09:55 ze- joined #salt
09:55 MSeven joined #salt
09:55 patarr joined #salt
09:55 patarr joined #salt
09:55 devx joined #salt
09:55 fxhp joined #salt
09:55 anteaya joined #salt
09:55 oz_akan joined #salt
09:56 redondos joined #salt
09:56 ahale joined #salt
09:56 Kenzor joined #salt
09:56 sieve1 joined #salt
09:56 veselin joined #salt
09:57 sectionme joined #salt
09:58 veselin Hi all, trying to 'setup bdist_rpm' the latest v2014.7 github release, but keep getting 'rpm: build/bdist.linux-x86_64/rpm/SPECS/salt.spec: No such file or directory'
09:58 veselin The spec file is definitely present
09:58 veselin any clues?
09:59 muzammil anyone who can help me with on salt + range setup, getting a "No minions matching the target"
10:01 Bryanstein joined #salt
10:05 masterkorp hello
10:12 cb joined #salt
10:17 badon joined #salt
10:36 Micromus_ How can I match on TWO grains in top.sls?
10:37 viq Micromus_: complex matcher
10:42 nkuttler Micromus_: compound matchers
10:42 Micromus_ 'G@os:Debian and G@oscodename:wheezy':
10:42 Micromus_ :)
10:43 totte joined #salt
10:47 apergos back with another question about setting grains locally on the minion... here I am trying to get the config and the __opts__ and __salt__ dicts set up correctly, grains.setval needs __salt__['saltutil.sync_grains']
10:47 intellix joined #salt
10:48 apergos I've tried every combination of raw_m, gen_mod, setattr and whatever I can think of and at this point the lab rat has given up going for the cheese
10:50 apergos what's the proper way to just have it load the blooming config and set up those dicts so that the grain module has them?
10:56 oz_akan joined #salt
10:59 diegows joined #salt
11:00 bhosmer joined #salt
11:01 ksk joined #salt
11:01 ksk hey guys.
11:01 sieve1 left #salt
11:09 tempspace joined #salt
11:14 tempspace joined #salt
11:15 tempspace Good morning!
11:23 ndrei joined #salt
11:27 Micromus_ why do minions on fedora show up as "localhost.localdomain"? how to fix? tried adding ip to /etc/hosts, did not help
11:29 dzen what's their hostname ? :)
11:29 dzen Micromus_: http://labtestproject.com/using_linux/permanently_change_hostname_on_fedora
11:33 saravanans joined #salt
11:35 viq Micromus_: /etc/salt/minion_id
11:35 viq Micromus_: either edit that directly, or after editing hosts remove that file and restart salt-minion
11:36 ramteid joined #salt
11:40 cDR__ joined #salt
11:46 ndrei joined #salt
11:47 Micromus_ thanks, forgot to ddelete the minion_id file after changing in hosts-file
11:52 intellix joined #salt
11:55 saravanans joined #salt
11:57 oz_akan joined #salt
11:58 stephanbuys joined #salt
12:02 bhosmer_ joined #salt
12:03 linjan joined #salt
12:05 masterkorp https://groups.google.com/forum/#!topic/salt-users/LrEY4OZgtds
12:05 masterkorp shameless cross reference guys
12:05 veselin left #salt
12:10 albertid joined #salt
12:17 jas- joined #salt
12:23 ramishra joined #salt
12:23 Shish joined #salt
12:24 hobakilllll joined #salt
12:25 ndrei joined #salt
12:26 byronschaller joined #salt
12:29 pclermont joined #salt
12:41 danielbachhuber joined #salt
12:42 saravana_ joined #salt
12:43 robinsmidsrod joined #salt
12:46 saravanans joined #salt
12:49 drybjed joined #salt
12:52 logix812 joined #salt
12:53 ghartz joined #salt
12:58 albertid_ joined #salt
13:00 mpanetta joined #salt
13:00 jhauser joined #salt
13:09 ndrei joined #salt
13:10 mpanetta joined #salt
13:10 jslatts joined #salt
13:11 racooper joined #salt
13:18 Lomithrani hey , is there an equivalent in salt.dockerio of absent but for images and not container ?
13:18 Lomithrani Like I want to destroy an image if my dockerfile was updated
13:19 alanpearce joined #salt
13:21 saravanans joined #salt
13:26 vejdmn joined #salt
13:26 ggoZ joined #salt
13:35 babilen Just to make sure: pkgrepo consolidated will not remove previously present repositories, but simply move all of them to sources.list doesn't it?
13:42 wendall911 joined #salt
13:43 mechanicalduck joined #salt
13:48 dude051 joined #salt
13:50 nullscan joined #salt
13:50 nullscan hello
13:50 nullscan i am trying to write a custom execution module for salt
13:50 nullscan so that i can use it in a pillar sls file
13:50 babilen Wonderful, it is a very rewarding experience
13:51 nullscan and i get this: Jinja variable 'dict object' has no attribute 'helpers.stripnum'; line 1
13:51 babilen Why would you need an execution module in a pillar?
13:51 alanpearce joined #salt
13:51 babilen And you get that when you do what exactly? Feel free to provide your code, SLS files, commands and their output on http://refheap.com
13:51 oz_akan joined #salt
13:51 nullscan what i need in the end is to have a pillar variable that i will use to rgoup my servers and i want a server to be assigned to that group based on his gqdn
13:52 saravanans joined #salt
13:52 nullscan but i want to strip out any numbers in that fqdn
13:52 nullscan if am going about this the wrong way, please let me know because i am loosing my moind with this :)
13:52 babilen Why don't you simply write your pillar in Python?
13:53 bhosmer joined #salt
13:53 nullscan babilen: and how should i reference this in my top.sls ?
13:53 babilen And you can also write custom grains if that data is relatively static and could be considered "sort of part of the host"
13:54 nullscan the custom grain would be a better idea, can you show me a link on how to do that?
13:54 babilen It's the first time that you mention top.sls ... It sounds as if you would want to write a custom grain. One second, I'll dig up the documentation.
13:54 babilen http://docs.saltstack.com/en/latest/topics/targeting/grains.html#writing-grains would be the documentation
13:55 NV eh, a python rendered pillar sounds like it'd be a better fit imho
13:55 NV runs on the master instead of each minion that way
13:55 NV i have one that creates a "roles" dictionary with a bunch of stuff
13:56 nullscan NV: this sounds a lot like what i want to do
13:56 nullscan how do i write a pillar in python?
13:57 UtahDave joined #salt
13:57 NV in your pillar top.sls include "roles" or whatever for all nodes (ie, *), then in roles.sls just have #py at the start
13:57 babilen In the end you write a grains such as https://www.refheap.com/88453 and place it in _grains/ in file_roots. You can then sync it with saltutil.sync_grains to your minions and target with it as detailed on http://docs.saltstack.com/en/latest/topics/targeting/grains.html
13:57 NV err, #!py
13:57 NV just define a function called run() that returns a dictionary
13:57 babilen NV: How would that be used in top.sls for targeting?
13:58 NV that dictionary is pillar data for example -- return {'roles': ['rolea', 'roleb']}
13:58 seanz joined #salt
13:58 NV http://pastie.org/9412230
13:58 NV top.sls ^^
13:58 NV pillar top.sls that is
13:58 UtahDave left #salt
13:59 babilen NV: That doesn't allow you to target all minions with role FOO (which is what I understood nullscan wants to do)
13:59 NV if that pillar roles just returned what i had above, in your states top.sls you could have 'roles:rolea': as a selector (you need - match: pillar on the line below of course)
13:59 babilen As you cannot target based on pillars.
13:59 NV sure it does
13:59 btorch is there a good way to test out jinja stuff without having to always run "state.sls {state} test=true" to find out details of why a certain variable fails to be set or get for example ?
13:59 babilen No, it allows you to target roles *at* hosts, not to target hosts *by* roles
14:00 NV salt -I 'roles:rolea' test.ping
14:00 NV targets all hosts with rolea
14:00 NV :)
14:00 babilen Ah, I overlooked the pillar matching
14:00 babilen Sorry, in that case you can also write a custom pillar.
14:01 babilen I thought that it would work, checked http://docs.saltstack.com/en/latest/topics/targeting/compound.html overlooked I in there and for some reason assumed that you cannot target by pillar
14:01 nullscan thank you both! i'll give it a try :)
14:01 quickdry21 joined #salt
14:01 nullscan the pillar in python solution that is :)
14:01 babilen nullscan: There isn't really a definite way of doing this and both approaches are fine. It really depends on what you want to do.
14:01 NV my roles.sls is actually fairly advanced and has "special" roles to include (and merge) pillar data from plain yaml files (which can be rendered with jinja too) and even pulls some user info (uids etc) from ldap :D
14:02 nullscan babilen: what i want to do actually is to assign each server in a server group so that i can targer or reference attributes from all servers in that group from anywhere
14:02 babilen If you have to target multiple roles at minions then you naturally want to do that via pillars, but my understanding of what you are trying to achieve was "compute a single 'role' / 'group' based on fqdn" which would, IMHO, be more appropriate as a grain (as it is sort of invariant)
14:03 NV i found doing it server-side in pillar was better than grains as grains needs a minion restart to refresh (grains are non-changing data), and were calculated minion-side (I want centralised configuration management after all, determining that kind of stuff from client-side stuff seems backwards to me)
14:03 arthabaska joined #salt
14:03 NV (especially if using that data to expose sensitive information  via pillars too)
14:03 babilen Which would be perfectly appopriate if it is a single invariant assignment
14:04 nullscan well, yes it is... i'm fairly new to salt so i do not have all its facilities figured out yet...
14:04 nullscan i'll try it out and see what happens
14:05 nullscan but a custom pillat setting seems more appropriate from what i get by reading through the manual
14:05 NV yeah, if it's really going to be invariant a grain might be appropriate, but i've found having a flexible roles structure which can be easily changed is useful as I can do things like assign roles like "siteahost" based on say a subnet, and if anything changes i don't have to worry about triggering a refresh from anywhere other than my master, and i can use said roles to expose pillar data (exposing pillar data based on grains sent from a minion could easily be sp
14:06 babilen There is no wrong approach here. I personally think of grains as "static" data about hosts (kernel, virtualisation scheme, fqdn, ip addresses, ...). That is things that can change, but are rather part of the host itself.
14:06 NV main selling points there
14:06 ndrei joined #salt
14:06 babilen It is a very specific way of designing your salt infrastructure and it has merits
14:07 babilen If you could be a a lot more specific about what you are *really* trying to do we might be able to provide better help, nullscan
14:07 nullscan ok let me describe in detail then
14:08 nullscan i want to run salt on a server with fqdn foo1 and during that run i want the server to be automatically added to a group named foo_servers
14:09 babilen What does adding a server to that group actually entail?
14:09 babilen Ah, well, I'll let you finish and ask questions then .. pray continue
14:09 nullscan so that i can use e.g a reactor that will go to the server bar1 and re-create iptables rules that will allow all servers in group foo_servers to access bar server on port 9999
14:10 nullscan or add all servers in group foo_server in an haproxy config
14:10 nullscan or even at the end of the day target all these servers to create users
14:10 nullscan you get the picture i guess :)
14:11 NV so why is standard targeting not useful? is your requirement to define said groupings in one place and then be able to reference that grouping, rather than re-produce the targeting specification in each place?
14:11 babilen You don't need a custom pillar or grain for that at all as you can just target by minion id foo* in that case
14:11 babilen (or check fqdn/host/... with startswith() ...)
14:12 ajprog_laptop1 joined #salt
14:12 ali_ joined #salt
14:12 babilen You only need the custom things if you, well, have to assign those "roles" / "groups" to minions that cannot be easily groups (and, with the pillar you would have to target in the pillar top.sls, so that doesn't help)
14:12 nullscan because what i have up to know (which i am not sure it is correct practice but anyway)...
14:13 NV also, is it likely that the same groupings would be re-used in multiple places at once for the same grouping? and what sort of chances are that servers will change groupings and have multiple?
14:13 babilen So with the pillar you would simply shift the targeting load from the states top.sls to the pillar top.sls
14:13 nullscan is e.g a pillar that has the definitions of all the users i want in all my servers but to specify if a user should be created on a server or not i thought of using a server's group to say if an account is active or not
14:14 babilen And if all you do is to target with a glob (e.g. minion id matched foo*) then you gain nothing
14:14 babilen yes, but what decides if a minion is part of that group? Is this *entirely and only* based on some common prefix/suffix/invariant of the minion id or fqdn?
14:14 nullscan yes
14:15 babilen Then just use that.
14:15 higgs001 joined #salt
14:15 nullscan actually i want to assign servers to groups based on their fqdn
14:15 babilen There is no need to complicate things if they are easy ;)
14:16 babilen Do you use fqdn as minion ids?
14:16 nullscan yes
14:16 UtahDave joined #salt
14:17 babilen nullscan: Then you can use those directly for targetting with all the methods detailed in http://docs.saltstack.com/en/latest/topics/targeting/globbing.html
14:17 babilen *targeting
14:18 KaaK joined #salt
14:18 nullscan yes but how can i use that in other state files?
14:18 babilen http://docs.saltstack.com/en/latest/topics/reactor/index.html#a-complete-example details an approach to use data['id'] (the minion id) in your reactor.
14:19 nullscan for example get the IP's of all the servers starting with foo*
14:19 NV nullscan: hrm, a pillar wouldn't help you there either
14:19 NV well
14:19 KaaK are there any guides on best practices for storing salt pillar and states in git? i.e. how (and what) to keep under revision control?
14:19 NV youd need to source it magically somehow to put it in the pillar
14:19 UtahDave left #salt
14:20 babilen That is, once again, a complex problem. You can, for example, run test.ping in your state if you want "live" data or use the salt mine to maintain data like that.
14:20 azmelanar joined #salt
14:20 babilen So, once again: What do you need the list of IPs for?
14:20 NV babilen: generating haproxy configs, etc ;)
14:21 nullscan NV beat me to it :)
14:21 nullscan haproxy configs, iptables configs, access grants on mysql's etc etc
14:21 NV anyway, i suspect a pillar to define roles is best (hey, it's what I use) to keep your node definitions all in one place (then you just reference role names, not complex matching patterns :D)
14:22 NV but that doesn't help your haproxy config
14:22 NV but salt mine would
14:22 NV but that wouldn't help anything else other than that
14:23 babilen For haproxy you might just want to add that host to the config when it comes online via reactor., This is discussed in http://docs.saltstack.com/en/latest/topics/reactor/#understanding-the-structure-of-reactor-formulas
14:23 nullscan NV: why is that?
14:24 babilen You could also tie it in to presen(s|c)e events and only maintain those that are currently online (which sorts of defeats the purpose of haproxy) ...
14:24 NV babilen: arguably haproxy should be determining what is and isn't up, not salt
14:24 nullscan babilen: i thought about this but i want to be able to decide when i am ready to add a server to haproxy for example
14:24 NV throw all hosts in haproxy, haproxy does health checks
14:25 nullscan guys, haproxy was just an example :)
14:25 babilen But as you want a general purpose solution I would say that the salt mine is what you are looking for, but then there are definitely better solutions for each subproblem that you are trying to solve.
14:25 babilen NV: Yes, sure. Hence the "which sorts of defeats the purpose of haproxy"
14:25 NV mhmm
14:26 NV i think salt mine can only target grains though and node names not pillar data though so you couldn't leverage central roles for that
14:26 nullscan ok, lets think of a very specific problem i am trying to tackle
14:27 nullscan i have this pillar
14:27 NV perhaps it's best just using a static pillar sls (webfrontends.sls say) that contains a dict listing all of your web frontends and have that pillar be used for haproxy seperate to whatever you use to apply configuration on said nodes
14:27 nullscan http://pastie.org/9412283
14:27 babilen All I am saying is that "list of IPs" might not really be the solution you are looking for for every single thing you are doing there ... It *does* make a difference if you configure iptables, haproxy or postfix. You might want different semantics, you might want to react to some events or roll out "static" configurations based on the state when you ran the highstate, ....
14:27 jalbretsen joined #salt
14:28 nullscan and my state's init.sls all that it does is itterate through a pillar with all my users
14:29 NV nullscan: so another problem you have there is your states don't know that "user1" is a user, so you really want that to be nested in a dict called "users" say that your states can iterate through
14:29 babilen I mean you can use "salt['test.ping']('foo*')" in your states and config files and then get the IP of those ...
14:29 NV the second problem is that you can't merge pillar data
14:29 NV http://pastie.org/private/gttn1gap9dcqwa0wwbwjya
14:29 NV that is my roles.sls on one of my salt installations ^^
14:29 nullscan i want to be able to say instead of if grains['nodename'] == 'bar' something like server_broup == foo
14:29 babilen nullscan: Take a look at https://github.com/saltstack-formulas/users-formula btw
14:29 TheThing NV: the second problem is that you can't merge pillar data <— that is true but there is a work-around for that
14:29 NV TheThing: ooh?
14:30 TheThing It adds a little noise in your state files but makes your pillar data clean and extensible
14:30 nullscan babilen: yes i have seen this formula, mine is loosly based on that one
14:30 NV http://pastie.org/private/gyoiijnuykseraqbongba that is roles.yaml that is expected to be in the same directory as the roles.sls from the previous pastie
14:30 TheThing NV: http://jonatan.nilsson.is/extending-pillar-data-in-saltstack/
14:30 babilen nullscan: No, you want to say "grains['host'].startswith('foo')"
14:30 babilen (or s/host/fqdn/)
14:30 nocturn joined #salt
14:30 NV TheThing: eww, dirty!
14:31 TheThing Yeah
14:31 nullscan babilen: that would solve a lot of my problems, yes!! thank you ;)
14:31 TheThing But works and now my pillar data is clean
14:31 NV anyway, check out my roles stuff - it fixes that merging issue too
14:31 NV as it assembles the pillar dicts itself
14:31 TheThing I can put "users.something: bla" anywhere in any pillar data
14:31 TheThing or for example iptables rules
14:31 babilen nullscan: Well, you can use literally that.
14:31 NV that roles.yaml causes a users dict to get filled with the contents of users/salt.yaml populated under it
14:31 NV glorious merging
14:32 tkharju1 joined #salt
14:33 NV basically replaces the pillar top file with roles.yaml
14:33 TyrfingMjolnir joined #salt
14:34 nullscan thank you very much for all your help, i'll look into everything you suggessted
14:34 hobakilllll anyone feel like helping out with a basic windows repo software issue?
14:34 ramishra joined #salt
14:38 tkharju2 joined #salt
14:39 laubosslink joined #salt
14:41 babilen nullscan: And you can also write states/pillars in Python if you need more logic during the minion id parsing. I'm not saying that it is the perfect approach, but my feeling is simply that you don't want to introduce a "role system" if you don't really use that. In the end you would, essentially, end up with something like: 'foo*' \n - role.foo
14:41 babilen And that simply doesn't buy you anything :)
14:42 Sp00n does the iptables module only work with newer versions?
14:42 Sp00n its trying to send -C
14:42 Sp00n on v1.4.0
14:42 Sp00n (lol)
14:43 nullscan babilen: tbh up to know my server groups are based only on fqdn's there is simply no need (for now) for anything more
14:44 che-arne joined #salt
14:44 slawdog01 joined #salt
14:44 nullscan having said that, i do understand that in the near future i might need something more fine grained, but for now i just need to get up and running
14:45 slawdog01 left #salt
14:46 rallytime joined #salt
14:47 nullscan also, excuse the million typing errors in almost every line
14:49 ujjain joined #salt
14:52 babilen nullscan: Don't worry, but things rarely are as easy as they appear to be. And I am simply no fan of introducing comples solutions to problems that don't require them just because I feel as if I might need it at one point in the future. If you already know that you *will* need it then, by all means, implement it already and spare yourself the transition work.
14:53 m1crofarmer joined #salt
14:54 thedodd joined #salt
14:57 aquinas joined #salt
14:58 KaaK joined #salt
14:59 endura joined #salt
14:59 endura so a quick question
15:00 endura how does one go about installing a program via salt, if the program isn't in an apt-get repos?
15:00 endura I had to wget, dpkg -i etc
15:01 endura actually it is in a repos. huh go figure
15:03 n8n joined #salt
15:03 dzen §wo, 1
15:03 ipmb joined #salt
15:04 kermit joined #salt
15:06 bombbastic joined #salt
15:08 zain_ joined #salt
15:09 bombbastic I'm trying to use salt deploy PHP/python application to servers. Are there favorite examples of configs to move code/files via "git pull"
15:09 babilen endura: You would simply provide it as source: URI
15:10 Eugene !deploy
15:10 Eugene Right, that's #git, doh.
15:11 Eugene Here we point to the git.latest state, or suggest learning to make a proper package of your app(rpm/deb)
15:11 yomilk joined #salt
15:12 aherzog joined #salt
15:15 premera joined #salt
15:16 bhosmer joined #salt
15:17 jaimed joined #salt
15:17 TyrfingMjolnir joined #salt
15:18 pclermont endura: nothing would stop you from having the .deb from salt:// and have another statement be a cmd.run “dpkg -i file.deb”
15:18 pclermont but that’s not making your life easier than a repo.
15:24 derito Does anyone have an idea about how I could get the cpu model from grains ? I have something along those lines in my grains
15:24 derito gpus:
15:24 derito {'model': 'RV370 [Radeon X300]', 'vendor': 'unknown'}
15:24 derito I manage to test simple things like os grains with jinja but struggle with this one. Thanks.
15:24 pclermont cpu_model: Intel(R) Core(TM) i7-3520M CPU @ 2.90GHz
15:25 derito Yeah this GPU might be a little outdated :)
15:25 pclermont derito: is that what you’re elooking for ?
15:26 derito Sorry I mispelled. I wanted to test GPU
15:26 derito I'd like to test the GPU model
15:27 pclermont I am not sure on what you want, if you want the “gpu model” well you are already providing that
15:29 nocturn joined #salt
15:29 derito I'm getting that from salt '*' grains.items. However I'd like a sls to be able to determine which brand of gpu the minion has and install drivers accordingly
15:29 Lomithrani joined #salt
15:30 anotherZero joined #salt
15:31 derito something along those lines http://pastebin.com/uBkqirau
15:32 platforms joined #salt
15:32 thedodd joined #salt
15:34 derito The problem is that I don't manage to access "gpus"' "model" item directly to test it
15:34 hobakilllll is it wrong to call salt + window an abject failure? i'm not an idiot but basic software installs on windows boxes make absolutely no sense to me at all. i can't get them to work..... i'm more likely the abject failure -but still.  this is insanity
15:35 hobakilllll eg: does this make sense to ANYONE? http://paste.debian.net/111017/
15:35 nocturn joined #salt
15:35 hobakilllll the comment says it failed to installed but guess what? it's on that test minion.
15:35 zach hasanyone seen this weirdness before? https://pastee.org/fug4e
15:36 zach MY state does NOT have anything related to iptables in it
15:36 conan_the_destro joined #salt
15:37 kaptk2 joined #salt
15:37 kaptk2 joined #salt
15:39 aherzog zach: It’s a known issue with that version; I believe the most recent version has it fixed.
15:39 whytewolf hobakilllll: looks like the package name doesn't line up. what does it show for pkg.list_pkgs
15:40 zach salt-minion-2014.1.5-1.el6.noarch
15:40 zach that version aherzog ?
15:40 aherzog Can someone fix the version number in the topic? Shouldn’t it say 2014.1.7 is latest?
15:40 zach there is no package for RHEL/CentOS yet
15:41 aherzog zach: Yes. It was fixed in 2014.1.6:.
15:41 manfred zach: it is in epel testing
15:41 * zach cracks the whip at whoever packages
15:41 whytewolf zach, epel has the 2014.1.7 packages in testing
15:41 bhosmer joined #salt
15:41 higgs001 joined #salt
15:41 * zach updates to testing
15:42 pclermont derito: it’s because you need to loop over it, here is a test sls that will create a filename for each model. http://pastebin.com/g3YqYvZe
15:42 racooper want it out of testing faster? test and give feedback!
15:42 zach "works fine" (tm)
15:46 derito Thanks a lot amis auvergnat !
15:48 pclermont derito: good luck
15:50 luminalflux joined #salt
15:52 pclermont I wonder how hard it would be to migrate the master_tops from master to minion (for masterless setups)
15:53 teskew joined #salt
15:53 tligda joined #salt
15:54 manfred derito: ™
15:54 manfred \u2122
15:55 pclermont h h
15:56 lude can i have salt run a command if a file does *not* exist on the minion?
15:57 pclermont yes
15:58 pclermont lude: unless: “test -e /path/to/file”
15:58 Sp00n onlyif [ ! -f file ]
15:58 Sp00n ?
15:58 Sp00n :D
15:58 luminalflux hello. I'm trying to wrap my head around salt coming from a chef background, where i'm used to nodes having attributes that can be used in a recipe later
15:59 luminalflux specifically i'd like to template out the minion's salt-mine configuration depending on variables i've set in states before
15:59 luminalflux is there a good way to do this that already exists?
15:59 aquinas joined #salt
16:01 hobakilllll yeah whytewolf the webdeploy is version 3.5 - not sure why M$ has it at 3.1...
16:01 kballou joined #salt
16:02 hobakilllll whytewolf: tho... WebDeploy:
16:02 hobakilllll 3.1237.1763
16:03 hobakilllll i hate M$.... it sais Microsfot Web Deploy 3.5 in Programs and Features but the Version is 3.1237.1763  which makes no sense to me.
16:04 hobakilllll /says
16:06 druonysus joined #salt
16:06 jhauser joined #salt
16:11 KyleG joined #salt
16:11 KyleG joined #salt
16:12 hobakilllll whytewolf: sweet little baby jesus you may have clued me into the whole goddamn problem. version.
16:14 whytewolf yeah does look like the version info, love microsoft.
16:15 colinjohnson joined #salt
16:16 ajprog_laptop1 joined #salt
16:18 Theo-SLC joined #salt
16:18 beando joined #salt
16:19 Theo-SLC How do I view the job results from a highstate from from salt-call?
16:19 Theo-SLC run from salt-call*
16:19 Sacro So I've done the whole add-apt-ppa whatever, but I can't seem to install salt, no see the repo, any thoughts?
16:21 viq Sacro: tried the bootstrap script?
16:21 Sacro viq: nope
16:21 viq Tried that, worked quite well on my machines
16:21 viq erm, try that
16:21 Gareth morning morning
16:22 TyrfingMjolnir joined #salt
16:23 meganerd joined #salt
16:23 rushm0r3 joined #salt
16:25 meganerd joined #salt
16:26 troyready joined #salt
16:27 azmelanar joined #salt
16:30 ramishra joined #salt
16:36 mechanicalduck joined #salt
16:38 rushm0r3 joined #salt
16:40 forrest joined #salt
16:41 bhosmer joined #salt
16:42 smcquay joined #salt
16:44 beneggett joined #salt
16:44 TheThing_ joined #salt
16:45 MatthewsFace joined #salt
16:46 n8n joined #salt
16:47 joehillen joined #salt
16:48 ghartz joined #salt
16:49 thedodd joined #salt
16:52 vejdmn joined #salt
16:52 alanpearce joined #salt
16:53 alanpear_ joined #salt
16:53 jaimed joined #salt
16:54 elfixit joined #salt
16:57 beneggett joined #salt
16:57 Comradephate joined #salt
16:57 XenophonF joined #salt
16:57 rushm0r3 joined #salt
16:58 XenophonF hobakill: hey did you get your win_repo problems sorted?
17:00 alanpearce joined #salt
17:01 Ryan_Lane joined #salt
17:03 rap424 joined #salt
17:07 pt|Zool joined #salt
17:07 pt|Zool hi
17:08 forrest pt|Zool, hello
17:11 pt|Zool anybody know that saltsack is support firewalld?
17:11 forrest pt|Zool, it doesn't currently
17:12 pt|Zool I'm soo sad :(
17:12 forrest why? firewalld is garbage
17:12 hobakilllll XenophonF: maybe. closer. not 100% sure just yet but i figured out the webdeploy issue thanks to some IRC help
17:12 forrest granted I agree, it should support it
17:12 XenophonF hobakilllll: what was the solution?
17:12 mpanetta Hmm, I would call it annoying, but not necessarily garbage though forrest :P
17:12 hobakilllll well i've done two things XenophonF :
17:12 XenophonF oh never mind i'll check the log :)
17:13 hobakilllll well version mismatch for my WebDeploy issue
17:13 forrest mpanetta, 'what if we took iptables, and split it all into a bunch of files like the ubuntu firewall that no one uses, we should also make sure to avoid resolving any issues that exist with iptables.'
17:13 forrest :P
17:13 mpanetta haha
17:13 hobakilllll but also i'm putting my msi and exe files on a local repo server to avoid the seemingly impossible to figure out filesystem issue
17:13 mpanetta Ive never looked at the files
17:13 forrest it's the most half assed iptables 'upgrade' possible
17:13 XenophonF ah
17:13 pt|Zool forrest: thx :)
17:13 mpanetta only used the CLI
17:13 forrest pt|Zool, yea np
17:13 forrest mpanetta, ahh
17:14 XenophonF hobakilllll: so you're making your own git repo for the windows packages, basically?
17:14 hobakilllll XenophonF: so my install location is now a website rather than the win/repo/FML
17:14 XenophonF gotcha
17:14 mpanetta Someone needs to do something to combine iptables with TC...  Now that might be nice
17:14 forrest mpanetta, it's just disappointing, they could have fixed a lot of the problems with iptables and made it a lot easier to use, but decided not to
17:14 forrest just a weird choice to make
17:14 hobakilllll XenophonF: yes. we have a custom yum-repo for our linux stuff - might as well do a windows folder on that server to keep everything in one place.
17:14 mpanetta forrest: Redhat makes lots of weird choices :P
17:15 forrest indeed
17:15 TheThing joined #salt
17:15 brandon_ joined #salt
17:15 XenophonF hobakilllll: I'm doing the same thing on FreeBSD and am about to setup one for CentOS
17:15 hobakilllll XenophonF: re: WebDeploy.  MS calls it WebDeploy 3.5 but in actuality it's version 3.1237.1763... once i made that change on the init.sls script it worked.
17:16 hobakilllll so i had WebDeploy / 3.5 / install_location blabh blah blah.... and it failed hard. changing 3.5 to 3.1237.1763 made the magic happen the the salt state
17:16 XenophonF oh i hate msi version numbering so much
17:16 hobakilllll XenophonF: whytewolf figured it out for me
17:17 hobakilllll XenophonF: after seeing this: http://paste.debian.net/111017/
17:17 EntropyWorks just noticed  that http://docs.saltstack.com/en/latest/topics/releases/0.17.5.html Next topic jumps to http://docs.saltstack.com/en/latest/topics/releases/0.6.0.html :( just stepping though the release changes before I talk
17:17 XenophonF hobakilllll: oh that makes lots and lots of sense
17:18 hobakilllll sure does. still it sucks ass because with a new package i'll never know what the true version # is w/o installing it first.
17:18 luminous when running salt-minion as an unpriviledged user, provided you have given the user permissions to /etc/salt and /var/cache/salt/ etc... is there _anything else_ you need to do to have saltutil.sync_all work? I'm not seeing my _modules and _states copied
17:18 XenophonF hobakilllll: you can extract that info from the MSI package
17:18 hobakilllll so i'll have to have a test server, install the package, get the "REAL" version # and then plug that into an init.sls
17:18 jalaziz joined #salt
17:18 XenophonF hobakilllll: um, let me remember how to do it, just a sec
17:18 hobakilllll XenophonF: sure but i don't think you can on an exe
17:18 forrest EntropyWorks, if you change the URL you can see the inbetween releases, but that is a problem. Could you please create an issue?
17:19 ksalman_ anybody has an example of how to get pillar data from an external http source that gives json data?
17:19 XenophonF hobakilllll: i think you can pull it out of an exe too
17:19 EntropyWorks will do
17:19 forrest EntropyWorks, thanks, you should have come to the sprint a couple weeks ago, I didn't know you were in Seattle
17:19 luminous ksalman_: can you write a python script that reads that API?
17:20 XenophonF an MSI has a ProductVersion property in it
17:20 ml_1 joined #salt
17:20 ksalman_ luminous: sure, I can
17:20 ksalman_ how would i use that in the pillar?
17:20 luminous ksalman_: start there, then read the ext_pillar docs
17:20 ksalman_ okay
17:20 luminous there are examples in the docs on what to do
17:21 hobakilllll XenophonF: i'd have to google how to do it cuz i know powershell about as well i as i know farsi.
17:21 luminous if you have the script/module working, you can hook it into ext_pillar very easiy
17:21 ksalman_ alrighty, thanks
17:21 XenophonF hobakilllll: no worries i'm getting together the stuff you'll need
17:22 KyleG1 joined #salt
17:22 hobakilllll XenophonF: dannnnnnnnnnng son thanks. :)
17:22 KyleG joined #salt
17:22 KyleG joined #salt
17:22 luminous ksalman_: goodluck!
17:23 XenophonF hobakilllll: you want to install orca, http://msdn.microsoft.com/en-us/library/aa370557(v=vs.85).aspx
17:24 XenophonF then you can open the msi in it
17:24 XenophonF brb
17:24 Gareth forrest: hey
17:24 forrest hey Gareth
17:24 luminous when running salt-minion as an unpriviledged user, provided you have given the user permissions to /etc/salt and /var/cache/salt/ etc... is there _anything else_ you need to do to have saltutil.sync_all work? I'm not seeing my _modules and _states copied
17:24 Gareth forrest: hows it going?
17:24 forrest Gareth, it's not Friday :P, you?
17:27 aw110f joined #salt
17:28 Gareth forrest: Back at work after vacation and it's not Friday :)
17:28 forrest Gareth, ahh nice, where'd you go on vacation?
17:29 Gareth forrest: New Hampshire and Vermont, my wife's family lives in New Hampshire.
17:29 forrest Gareth, ahh ok
17:29 Supermathie joined #salt
17:30 jalaziz joined #salt
17:35 saravanans joined #salt
17:43 luminous Gareth: I bet that was enjoyable
17:43 ksalman_ luminous: actually, I suppose I could just write a custom grain to query the API as well. Is there a reason I'd want to use ext_pillar over a custom grain?
17:44 Supermathie Hey, I'm looking for two things: a way to access job metadata from a running job, and a way to schedule jobs run a job running on a minion. Is there any mechanism for these things?
17:44 luminous ksalman_: grains and pillar are very different
17:44 carnedepassaro joined #salt
17:44 luminous ksalman_: so it depends on what you want to do
17:44 luminous ksalman_: grains are 'grains of truth' - they are not ever going to change for the time the salt-minion is up
17:45 luminous unless there is some syncing/update
17:45 blarghmatey joined #salt
17:45 luminous ksalman_: pillar is only available to the minion specified
17:46 ksalman_ luminous: i want to query the API and get some data that the minion needs to configure some additional interfaces. The ip addresses are is what i will be getting from the API
17:46 luminous IP addresses for the minion?
17:46 luminous or for another service/etc?
17:46 ksalman_ yea, there are interfaces (non routable) that need to be configured
17:46 luminous that's an either or
17:47 luminous if the IP address is one of the minion's, and not to point to some other host, that is conceptually a grain
17:47 ksalman_ It is one of the minion's
17:47 matthiaswahl joined #salt
17:47 luminous pillar could be used, but this sounds more like a grain
17:48 Gareth luminous: it was. :)
17:49 luminous :)
17:52 azmelanar joined #salt
17:56 mephx joined #salt
17:57 jslatts joined #salt
17:57 t0rrant joined #salt
18:01 hobakilllll XenophonF: have you done much work with updating windows minions from the salt CLI? i have a handful on the wrong version and i'd rather have salt update them than do it myself.
18:04 rglen joined #salt
18:05 napper joined #salt
18:08 giannello joined #salt
18:09 rypeck joined #salt
18:09 druonysuse joined #salt
18:10 vexati0n $250,000 in premium monopoly money to the first person who adds a "specify your own pub_port in the minion conf file" feature.
18:13 ksalman_ when querying an API do i hhave to use urllib, or can i use a 3rd party module, like requests?
18:13 ksalman_ this is for a custom grain
18:16 rushm0r3 joined #salt
18:16 forrest vexati0n, ?
18:16 forrest vexati0n, for the tcp port?
18:17 forrest vexati0n, if so, it's already in there: tcp_pub_port: 0000
18:19 rypeck Is their a guide I can read as a Puppet user that would make a handy introduction to Salt?
18:19 tdilly joined #salt
18:21 forrest rypeck, Hmm, you know we don't have anything for that specific use case. I'd suggest you either go through the main tutorial and just kind of 'skip' as needed.
18:21 forrest that's a good idea though
18:22 rypeck forrest: yea thanks. I'm just struggling trying to grok salt for my use case right now.
18:22 forrest ok, what's your use case?
18:22 ecdhe joined #salt
18:23 hobakilllll ok i'll open it up to the floor. does anyone have a good way of updating windows minions to an updated version?
18:23 stephanbuys joined #salt
18:23 rypeck trying to manage some configuration files - when the files change I need to restart a service.
18:23 rypeck Puppet I did this with custom variables and templates.
18:23 forrest hobakilllll, duh, reprovision
18:23 druonysuse joined #salt
18:23 druonysuse joined #salt
18:23 forrest rypeck, ok you'd want to use 'watch'
18:23 kuffs hobakilllll: are you using the frozen minion?
18:23 forrest rypeck, can you gist your state that you've been working on so far?
18:24 forrest rypeck, or states if you broke the service and file into different items
18:24 rypeck not there yet - still reading through the documentation to figure out how I can template files.
18:24 forrest rypeck, ok, let me whip up a quick example then
18:24 rypeck forrest: thank you.
18:24 hobakilllll kuffs: not sure what you mean by that. i checked out the update_url and after i tried bumping up to 2014.1.5 i coudln't do a --versions-report against my test minion
18:24 forrest rypeck, yea np
18:25 Eugene Sweet. Just got approval to deploy Salt at $DAYJOB. Next step: see if they'll pay for the professional stuff.
18:25 kuffs I don't understand that sequence of events
18:25 forrest Eugene, nice
18:25 kuffs what mechanism did you use to perform the upgrade?
18:26 stephanbuys1 joined #salt
18:26 kuffs and I know that typically one is to upgrade the masters before the minions
18:26 hobakilllll kuffs: so nothing just yet. i noticed i had a bunch of win boxes on 1.4
18:26 * robawt highfives forrest
18:26 hobakilllll kuffs: i wanted to find a better way to bring them up to 1.5 to match my master.
18:26 forrest hey robawt
18:27 thedodd joined #salt
18:27 robawt hey hey
18:27 rypeck forrest: I see how I can template modules - I just don't see how I can do that with files.
18:27 kuffs I wrote the original frozen / esky stuff back in the day that allows for in-place upgrades
18:27 hobakilllll kuffs: i dug around and found update_url in the minion config so i just popped in the 1.5.exe file
18:27 forrest rypeck, can you link me the docs you're looking at?
18:27 kuffs but it's been a long time since I've been that deep in the project
18:27 mik3 left #salt
18:27 rypeck forrest: http://docs.saltstack.com/en/latest/topics/tutorials/states_pt3.html
18:27 kuffs no, update_url is not a path to a binary
18:28 hobakilllll kuffs: i didn't think so but i had to start somewhere :)
18:28 kuffs it's a path to a webserver with a directory full of esky builds
18:28 * hobakilllll thinks: step one, investigate esky
18:29 TheRealBill_ joined #salt
18:29 kuffs but it's been a while since I've been involved with the project
18:30 kuffs I know that utahdave was playing with chocolatey around that time
18:30 hobakilllll it looks like a gigantic PITA
18:30 bombbastic Should I be able to do this if I set it up correctly in pillar :   salt -I 'role:produciton' -I 'role:web' test.ping
18:30 hobakilllll thanks kuffs for the help
18:30 kuffs np
18:31 nahamu kuffs: The Esky stuff is great!
18:31 XenophonF hobakilllll: sorry am back now
18:31 kuffs it's a pretty neat tool
18:31 kuffs I wish it was a little less black magic
18:31 XenophonF hobakilllll: regarding upgrading salt on windows, i've been working on an msi installer package for windows to replace the existing executable installer
18:31 nahamu kuffs: esky itself or the underlying freeze black magic?
18:31 tdilly joined #salt
18:32 hobakilllll XenophonF: thankfully we don't have many ATM but they'll build and i'll have to find a better way to do this than to RDC to the thing and install it on top of the older version
18:32 kuffs I had a tough time with both, but esky was worse imo
18:32 XenophonF hobakilllll: that's precisely why i want an MSI installer package
18:32 kuffs I remember trying to debug what was going on with missing modules
18:32 forrest rypeck, so the templating of files is similar. First off, look at https://gist.github.com/gravyboat/145df1fd823581d9140f for an example of watch_in, where we watch a file in a service which we're including
18:33 XenophonF hobakilllll: i have it partially working but according to UtahDave I wasn't building Salt the right way
18:33 tdilly My master is running 2014.1.3 and I accidentally installed 2014.1.5 on a minion. The highstate has been running all day and conveniently stopped processing after setting up sshd to required public key auth but before installing my authorized key file (definitely going to switch that order around today). Is there anyway to cancel the active job?
18:33 kuffs nahamu: though I may be investigating cython as a replacement for both perhaps, needs more research
18:34 XenophonF hobakilllll: and of course salt's had two releases since I forked in late May so...
18:34 XenophonF i have a lot of testing to re-do
18:34 forrest rypeck, regarding templating of files, there are some examples here for basic stuff: http://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.jinja.html
18:34 kuffs XenophonF: my installation is so old I'm going to be standing up a parallel stack
18:34 hobakilllll XenophonF: yeah...let me know if i can ever help out
18:34 manfred tdilly: saltutil module has a term_job function
18:34 forrest rypeck, if you want to see it used in an actual project, check out: https://github.com/gravyboat/hungryadmin-sls/blob/master/salt/hungryadmin/files/hungryadmin.conf and if you want, the associated blog post where I set it all up: http://hungryadmin.com/building-a-static-blog-and-deploying-it-with-saltstack.html
18:35 manfred tdilly: http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.saltutil.html#salt.modules.saltutil.kill_job
18:35 forrest rypeck, the project I linked gives a bit more of a 'complete' overview than the gist does, as that's just real simple and leaves out directory structure and such
18:35 tdilly manfred: thanks!
18:37 vlcn anyone have any thoughts on best practices for ensuring that the salt-minion stays running?
18:37 tdilly dang, still didn't end the job...
18:37 forrest vlcn, what do you mean?
18:37 vlcn I've used supervisord for similar things in the past
18:37 forrest vlcn, it's a service
18:37 manfred tdilly: did you use the kill_job or term_job? one of them should send a -9
18:38 manfred vlcn: if you are using systemd, you could always use Restart=always, and the systemd service unit will always restart if the process dies.
18:38 forrest why is the process dying in the first place?
18:38 rypeck forrest: Thanks!
18:38 manfred because you keep sig terming it :P
18:38 forrest rypeck, yea np
18:38 ajolo joined #salt
18:39 vlcn forrest, it's bound to happen eventually
18:39 forrest manfred, har har har
18:39 forrest vlcn, the process just dying?
18:39 forrest on what kind of system, lol
18:39 tdilly manfred: kill_job, I get a traceback and an error that says: 'OverflowError: long too big to convert'
18:39 forrest I've had a minion running for over 6 months without a problem.
18:39 vlcn most of the issues appear to result from authentication failing
18:39 vlcn which the minion eventually stops trying
18:40 jslatts joined #salt
18:41 tdilly looks like it's time for the recovery console :)
18:43 endersavage joined #salt
18:44 vlcn on a related note, what about rolling out salt-minion updates?
18:44 vlcn anyone have a system that works well for that?
18:44 rypeck forrest: is it possible for me to specify a command needs to be run if a file is to be changed?
18:44 rypeck puppet I can detect a refresh and trigger an exec command.
18:44 endersavage pitched salt my puppet dominant infrastructure team, they're quite excited. salt +1
18:45 rushm0r3 who here would be intimately familiar with the event system ?
18:45 Nazzy joined #salt
18:45 Nazzy joined #salt
18:45 forrest rypeck, if I'm understanding you correctly ,use cmd.wait: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.cmd.html#salt.states.cmd.wait
18:46 rypeck Yes. Perfect. thanks forrest
18:46 forrest np
18:47 madphoenix joined #salt
18:47 rushm0r3 joined #salt
18:48 rypeck forrest: what about "onchanges"?
18:48 rypeck http://docs.saltstack.com/en/latest/ref/states/requisites.html#onchanges
18:48 rypeck that actually fits my need perfectly.
18:49 madphoenix I'm having issues with a second ext_pillar in gitfs not being fetched on the master.  Two branches of the same repository are listed for ext_pillar, but /var/cache/salt/master/pillar_gitfs only contains one directory
18:49 madphoenix is there a way to force ext_pillar updates?
18:50 forrest rypeck, onchanges isn't in yet
18:50 forrest rypeck, that's in the new release which should hopefully be out soon
18:50 vexati0n forrest: sorry, will "tcp_pub_port" override the pub_port the Master tells the minion about during auth ?
18:50 rypeck ah... so i'm reading the wrong documentation
18:51 vexati0n I need a way to run the server on 4506/4505 as usual, but also have fallback ports in the event that minions are behind firewalls that don't allow communication on those ports
18:51 forrest rypeck, no you have the right docs, there's a footnote regarding when that is added in there
18:51 forrest rypeck, that's a problem we have with the docs currently, they don't split on version, it's something that is being worked on :P
18:52 rypeck yay
18:52 forrest vexati0n, uhhh yea that I don't know about, I don't think there are 'fallback' ports
18:52 forrest vexati0n, how would the salt master start a service on the different ports? the 4505 and 4506 need to be open inbound on the master
18:52 vexati0n forrest: no there are no fallback ports built into salt, but i can emulate that functionality by allowing minions to configure their own pub ports instead of obeying the master
18:52 XenophonF hobakilllll: regarding pulling out msi package version numbers
18:53 forrest vexati0n, yea that functionality doesn't exist then
18:53 XenophonF hobakilllll: see http://winadminnotes.wordpress.com/2010/04/05/accessing-msi-file-as-a-database/
18:53 hobakilllll XenophonF: thanks.
18:53 forrest I don't even know how that would work since the master has the ports open.
18:53 meteorfox joined #salt
18:53 vexati0n for example, i already have a fallback port for ret_port by forwarding packets from a 2nd port to the real one. but i can't do that with pub_port because the minion only talks to the pub_port as described by the master during auth.
18:53 forrest rypeck, yea it's not usually a big deal, just watch for the version add footnotes.
18:54 forrest vexati0n, *facepalm*
18:54 XenophonF hobakilllll: see also http://stackoverflow.com/questions/8743122/how-to-find-msi-product-version-number-using-powershell#8743878
18:54 forrest what is wrong with your network :(
18:54 vexati0n nothing is wrong with my network - i have minions spread out across hundreds of networks across the country that i don't have admin access to, and these networks often employ content filters and/or firewalls that block outgoing traffic.
18:54 forrest ahh
18:55 XenophonF hobakilllll: but i still think it'd be easier if you install Orca from the Windows SDK and just use that to inspect whatever MSIs you want to deploy
18:55 hobakilllll i'd agree. thanks again for the help today XenophonF
18:55 vexati0n so, allowing the minion to specify its own pub_port to use rather than always obeying the master would allow me to construct a framework around salt that provides 'fallback' functionality.
18:55 vexati0n guess i could always edit the python myself but ew
18:56 XenophonF hobakilllll: as for extracting MSIs from an EXE, it depends
18:56 napper joined #salt
18:56 XenophonF hobakilllll: the only foolproof way I know is to run the executable installer and look for MSI files in your %TMP% directory
18:57 XenophonF hobakilllll: http://stackoverflow.com/questions/1547809/extract-msi-from-exe
18:58 XenophonF although i usually can pull the msi out using 7zip
19:08 peters-tx joined #salt
19:08 savvy-lizard joined #salt
19:11 stephanbuys joined #salt
19:13 peters-tx joined #salt
19:14 peters-tx whiteinge, https://media.readthedocs.org/pdf/salt/latest/salt.pdf is broken again
19:15 rushm0r3 joined #salt
19:15 whiteinge peters-tx: ty
19:16 stephanbuys hi all, salt-ssh has a 'need_deployment' method that creates the thin client's /tmp/.salt folder which belongs to root and is not group or world read/writable, but when using 'sudo' the scp of the thin-client into /tmp/.salt subsequently fails due to the sudo user not having appropriate permissions. I'm hacking at it, but is anyone here involved in salt-ssh development?
19:20 stephanbuys1 joined #salt
19:20 rushm0r3 joined #salt
19:21 talwai joined #salt
19:21 talwai Is there any way to template i.e. get pillar/grain data into orchestration files run via state.orch?
19:24 stevednd whiteinge: How does the fall through for environments work with gitfs branches(if at all)?
19:26 stephanbuys joined #salt
19:27 williamthekid_ joined #salt
19:27 elfixit joined #salt
19:29 rushm0r3 joined #salt
19:31 beneggett joined #salt
19:31 rushm0r3 joined #salt
19:35 ksalman_ any chance this would be available for windows anytime soon? http://docs.saltstack.com/en/latest/ref/states/all/salt.states.network.html
19:36 stephanbuys1 joined #salt
19:38 madphoenix Is it possible to use grain data in a pillar sls file?
19:38 madphoenix something like, host: {{ grains['ip_interfaces:eth0'] }}
19:40 beando joined #salt
19:40 kiorky joined #salt
19:40 vexati0n forrest: you will feel indifferent to know that i have been able to cobble together 'fallback' support by using iptables tomfoolery and routing shenanigans. my official position is that it would be awesome if salt allowed this natively.
19:41 forrest vexati0n, lol yea that sounds like an awful workaround, you could put in an issue for it
19:41 forrest honestly it's an edge case I've never heard of
19:41 forrest so it might have not been considered.
19:45 forrest vexati0n, I'd seriously open an issue though
19:45 forrest if you haven't already
19:48 kiorky joined #salt
19:48 talwai Is there any way to template orchestration files  i.e. get pillar/grain data into them?
19:48 tyson_ joined #salt
19:50 siddc joined #salt
19:50 manfred talwai: probably, but it wouldn't be really usefull, i believe they will be templated against the masters pillars and grains and not against the minion that you are sending it too
19:51 siddc I am calling a salt command with the cmd.script module from a bash command. This command returns two values as stdout. How can I assign these values to variables?
19:52 n8n joined #salt
19:55 talwai manfred: thanks for the info. Am i wrong in thinking this greatly restricts the utility of orchestration files? If I want to execute a complex pillar-dependent set of states on a group of minions, do I have no choice other than to include those states in the topfile and call them via state.sls?
19:56 manfred why would you call them via state.sls if they are in top.sls
19:56 manfred you shouldn't be templating in orchestrate files or the top.sls.
19:56 manfred you should just be matching minion attributes
19:57 manfred assign your pillars in the pillar top.sls, and then specify which states should run based on pillars with something like 'I@foo:Bar' in the /srv/salt/top.sls
19:58 talwai Well maybe they wouldn't be in top.sls. My use case is I don't want these states run at all during state.highstate but would need to run them at other indeterminate points. Since they are pillar-dependent, I can't put them in orchestrate files
19:58 manfred yes you can
19:58 manfred just match against the pillar
19:59 manfred you can't template won't be able to template in jinja in the orchestrate file, but you can specify that only minions with a specific pillar run specific things
19:59 stevednd is there any place to find status updates on helium?
19:59 manfred stevednd: not really
20:00 ggoZ joined #salt
20:00 vexati0n forrest: do i just submit an issue on the main github page or is there a special place for "weird corner cases" ?
20:01 forrest nope, just the main 'create issue' link on github
20:01 stevednd manfred: you sure you can't jinja template orchestration files? I'm pretty sure I have done that.
20:01 londo joined #salt
20:01 jY how do i remove hosts from the salt mine?
20:02 manfred stevednd: probably can do it, but as i said above, i think it uses the masters pillars and grains and no the minions
20:02 stevednd yes, that part is correct
20:05 laubossl1nk joined #salt
20:06 whiteinge stevednd: each gitfs branch/ref is an environment and the fall-through works the same as it does with file-system environments
20:06 talwai manfred: hmmm, that sounds like the solution i want, thank you. Still it seems a little convoluted to have to match against specific discrete values of minion pillars and hardcode those values into my salt.state call when basically all i want is for the pillar to fill in a jinja block no matter what it is.
20:07 talwai but maybe that scenario lies outside of what orchestrate should be used for
20:11 stevednd whiteinge: but with roots you specify the fall through. as shown here http://docs.saltstack.com/en/latest/topics/tutorials/states_pt4.html#environment-configuration
20:11 stevednd where do I define the proper sequence for gitfs?
20:12 whiteinge each environmet can have it's own top-file. all top-files are merged together
20:14 KGrange joined #salt
20:15 Ahlee all values for each environment are then merged into one dict, right?  So if you have base: - foo in environment foo, and base: - bar in environment bar, salt sees base: - foo - bar right?
20:15 beando joined #salt
20:15 stevednd whiteinge: sorry if I'm being thick, but I'm still not seeing how I would say have a dev:top.sls that would indicate that missing files should be checked for in qa, and if not there, then in master/prod
20:16 mpanetta joined #salt
20:20 ndrei joined #salt
20:20 KGrange so, what's the opinion of the room here on having execution modules that depend on states? is calling __salt__['state.low']({state: "pkg", fun: "installed", name: "vi"}) kosher?
20:21 kiorky joined #salt
20:23 rypeck Is there something special I should know about comments in an sls file?
20:24 rypeck # should cause jinja to skip over the line right? I'm getting an error on something I commented out
20:25 stephanbuys joined #salt
20:25 KGrange what's the line and context around it? Maybe it's expecting something there indentation-wise?
20:26 siddc joined #salt
20:26 rypeck new lines around it
20:26 KGrange what's the error it gives?
20:26 endersavage joined #salt
20:27 rypeck is there a python-yaml checker I can run that'll give me a more informative error message than just running salt '*' pillar.data?
20:27 KGrange pyyaml should be installed along with salt, you could try that directly
20:28 rypeck hmmm nope
20:28 rypeck i'll go get it
20:28 KGrange pyyaml isn't installed?
20:28 rypeck nope
20:29 KGrange salt relies on pyyaml, I'd hope it is :P maybe that's your issue. How did you install salt?
20:29 rypeck Security Onion installation
20:29 Katafalkas joined #salt
20:30 rypeck KGrange: How do I check if it's installed?
20:30 Katafalkas Hey, is there a way to specify AWS availability zone in the cloud map ?
20:30 rypeck When I try to import yaml it works - I assume that's what you mean?
20:30 KGrange I'm not sure what Security Onion installation means? Is that a specific distro?
20:30 Katafalkas using salt-cloud
20:30 KGrange yeah, import yaml means it's installed I belive
20:31 rypeck KGrange: then it's installed
20:31 rypeck does it come with an oddly named command utility?
20:31 laubossl1nk Hi everybody, i would like to use {{ salt['pillar.get'](...) }} in a pillar, like an alias, but i have 'None' seems doesn't work... there is functionality or somth like that to do it ??
20:31 KGrange eh, I just open python interactively and copy and paste the yaml. You could probably do a shell one-liner that'll take from stdin
20:33 notpeter_ Katafalkas: Yeah, availability_zone: us-east-1d
20:34 notpeter_ assuming you've already specified location: us-east-1 (or whatever)
20:37 rypeck KGrange: are there no comments in yaml?
20:37 KGrange yes, with a hash (#)
20:37 davet joined #salt
20:37 KGrange (yes as in there are comments)
20:38 stephanbuys1 joined #salt
20:39 laubossl1nk Hi everybody, i would like to use {{ salt['pillar.get'](...) }} in a pillar, like an alias, but i have 'None' seems doesn't work... there is functionality or somth like that to do it ??
20:40 bmatt laubossl1nk: wrap that in a {% if salt['pillar.get'](..., False) %}
20:40 bmatt that way it skips it entirely if the value isn't set, and only uses it if it is set
20:44 bhosmer joined #salt
20:44 laubossl1nk bmatt: doesnt work it will be False !
20:44 laubossl1nk bmatt: :(
20:46 bmatt laubossl1nk: I mean to use like this https://gist.github.com/thenewwazoo/9c8a870ae9fd78a07428
20:47 bmatt that way, if salt['pillar.get']('...') is none, you just don't get the pillar data
20:48 yomilk joined #salt
20:49 Katafalk_ joined #salt
20:51 ajolo joined #salt
20:51 laubossl1nk bmatt: why it need to add in condition !? it's not logical ?
20:52 bmatt because salt['pillar.get']('non_existent_key') will throw an exception
20:54 Katafalkas joined #salt
20:55 KyleG joined #salt
20:55 KyleG joined #salt
20:57 laubossl1nk bmatt: i wrap it, but same... look : http://paste.ubuntu.com/7838450/
20:57 laubossl1nk bmatt: i obtain 'WHAT'
20:58 bmatt what line is relevant?
20:58 laubossl1nk bmatt: http://paste.ubuntu.com/7838458/
20:59 bmatt oh, right, yes, because the edge:network:proxygate:ip pillar isn't set
20:59 laubossl1nk bmatt: no look on last link :)
20:59 KGrange is that pillar you're editing in the same environment?
20:59 bmatt yeah, a pillar file can't know about pillar data
21:00 KGrange also you just publicly posted your private key
21:01 druonysus joined #salt
21:02 laubossl1nk KGrange: yes it is the same environment it seems to work ?
21:02 laubossl1nk KGrange: yeah its private key in dev :)
21:02 jhauser joined #salt
21:05 KGrange Oh, then whatever contains edge is being loaded after that first file you posted
21:08 Supermathie joined #salt
21:09 laubossl1nk KGrange: i will look, and try something it's possible
21:09 beando__ joined #salt
21:09 KGrange yeah, check your pillar top file
21:09 vejdmn joined #salt
21:10 laubossl1nk KGrange: yeah i add it in first, it's same
21:10 KGrange can you post that file, then?
21:11 Katafalkas joined #salt
21:11 laubossl1nk yes
21:13 TheThing joined #salt
21:13 laubossl1nk KGrange: KGrange http://paste.ubuntu.com/7838519/
21:14 KGrange no, the file edge itself
21:14 beneggett joined #salt
21:15 laubossl1nk KGrange: http://paste.ubuntu.com/7838536/
21:17 KGrange I really have no idea then, sorry
21:17 laubossl1nk KGrange: i think its not support
21:17 KGrange try debugging with salt's state template modules and see what gets spit out, but that's the edge of my knowledge
21:18 KGrange yeah maybe that's the case
21:24 beneggett joined #salt
21:26 roolo joined #salt
21:30 yomilk joined #salt
21:30 giannello joined #salt
21:31 kiorky joined #salt
21:33 ndrei joined #salt
21:35 robertkeizer joined #salt
21:43 Katafalkas joined #salt
21:44 skullone 8has anyone run salt on centos7 yet?
21:44 skullone -8
21:45 racooper skullone,  the latest is in EPEL testing for EPEL 7 beta
21:48 skullone ah ok
21:54 lz-dylan is the map.jinja file in formulas explicitly meant for OS-specific differences?
21:54 lz-dylan I'd like for the values in one of my Jinja conf files to change based on a grain that I'm setting on deploy
21:56 lz-dylan so instead of keying off of OS, I'd like to key off of environment
22:05 admineric joined #salt
22:07 bhosmer joined #salt
22:09 jhauser joined #salt
22:11 admineric Is there a way to get zeromq3 on Centos 5 without having to compile and/or re-package python26-zmq? What is the best way to get zeromq3 working with Salt on Centos 5?
22:11 markm joined #salt
22:12 oz_akan joined #salt
22:12 forrest admineric, http://zeromq.org/distro:centos
22:12 forrest use the one there from fengshouo
22:12 forrest *fengshuo
22:12 forrest works fine
22:13 kiorky joined #salt
22:14 admineric forrest: yes i have tried that but when installing python26-zmq I get "Missing Dependency: libzmq.so.1()(64bit)"
22:14 admineric forrest: similar issue to https://groups.google.com/forum/#!topic/salt-users/c9p49ZUczJo
22:15 fannet joined #salt
22:17 forrest admineric, I'm not sure then, last time I checked someone installed the newer release of zmq and it was fine
22:17 fannet Hi - is there any reason why salt '*' network.interface_ip eth0 would return:     'network.interface_ip' is not available.
22:18 fannet most other functions in the network module work
22:19 forrest fannet, it might be that interface_ip requires a different package that isn't installed?
22:19 forrest fannet, does interface eth0 work?
22:20 fannet nope
22:20 forrest fannet, https://github.com/saltstack/salt/blob/develop/salt/utils/network.py#L541 is what those are calling it looks like
22:20 forrest so is ifconfig not installed?
22:20 fannet it is - using cmd.run "ifconfig eth0" works fine
22:21 forrest fannet, weird, which release of salt are you on
22:21 fannet salt --version salt 2014.1.7
22:21 fannet salt-minion 2014.1.7
22:22 forrest fannet, hmm, that is very odd.
22:22 fannet indeed =)
22:22 forrest fannet, might be worth checking the issues on the salt repo to see if anyone has encountered that
22:24 fannet also sudo salt 'redis*' redis.info gives the same
22:24 oz_akan joined #salt
22:25 forrest fannet, can you try to downgrade to 2014.1.6?
22:25 fannet actually none of the redis commands do lol
22:25 forrest fannet, then we can confirm if it's a release issue/salt problem, or something else.
22:25 fannet how can I see what modules are loaded
22:26 chiui joined #salt
22:26 forrest timoguin, you around?
22:27 forrest fannet, also, is this on freshly provisioned systems where there was a chance salt started before other items might have been installed?
22:28 forrest fannet, I think the command for listing the loaded modules is salt.modules.sys.load_modules
22:28 forrest so probaly salt modules.load_modules? I can't remember :P
22:28 forrest actually, sys.load_modules
22:30 fannet forrest - salt was pushed to hosts via puppet using the apt repo
22:30 DaveQB joined #salt
22:30 forrest fannet, ahh ok
22:35 ghartz joined #salt
22:36 fannet hm salt.load_modules didnt work either
22:36 jalbretsen joined #salt
22:36 fannet sorry sys
22:37 fannet ah
22:37 fannet it was list_modules
22:38 fannet interesting
22:38 fannet sudo salt -d 'redis*' sys.list_functions network
22:38 fannet confirms that the function is indeed not implemented in the module
22:40 fannet also list_modules shows the redis execution module is not there
22:40 forrest fannet, can you restart the service?
22:40 fannet at least its in agreement with me =)
22:40 fannet the master or the minion?
22:40 forrest both
22:40 forrest just to be safe
22:40 fannet any particular order?
22:40 forrest nope
22:41 fannet ok
22:41 jamesbdev joined #salt
22:42 fannet Restarting salt master control daemon: salt-masterWARNING: Unable to bind socket, error: [Errno 98] Address already in use The ports are not available to bind
22:42 forrest fannet, ok that means that the service probably didn't shut down properly which is odd
22:42 forrest stop the service instead
22:42 forrest see if there is a rogue process or something weird
22:42 fannet it came back up
22:43 forrest ok cool
22:43 fannet i think it tried to respawn too fast
22:43 forrest gotcha
22:43 fannet although  my minions seem to be down
22:43 forrest did you try and restart it with a salt command?
22:43 fannet probably the same issue w/ restart
22:44 forrest or by hand?
22:44 forrest could be, I find that odd though
22:44 fannet salt
22:44 beneggett joined #salt
22:44 forrest ok, let me know when you've started it back up, and restart the module
22:44 forrest err reload the modules
22:44 forrest jeez
22:44 forrest see if it's in there :P
22:45 fannet good thing I only did it on 5 servers lol
22:46 forrest fannet, yea, any time there is an update, better to be safe than sorry
22:47 fannet how recent is this update
22:47 forrest uhh, a week or two?
22:47 fannet oh wow ok
22:48 fannet no love
22:48 fannet still same set of modules
22:48 fannet where on the filesystem are the modules
22:49 forrest fannet, should just be down in the salt dir
22:49 forrest fannet, here's why :P https://github.com/saltstack/salt/tree/v2014.1.7/salt/modules/redismod.py
22:49 fannet LOL
22:49 forrest fannet, doesn't exist in 2014.1.7
22:50 forrest it's only in the 2014.7 tag, and the develop branch
22:50 forrest you could try to drop it in there manually
22:50 Outlander joined #salt
22:50 fannet in /usr/lib/python2.6/dist-packages/salt/modules/ ?
22:51 forrest should be
22:51 forrest oh no
22:51 forrest sorry
22:51 forrest the joys of being distracted
22:51 fannet in /usr/share/pyshared/salt/modules/ ? : )
22:51 forrest fannet, nope, just /srv/salt/_modules
22:52 forrest then you'll need to sync it out to the minion as well.
22:52 forrest with salt <name> saltutil.sync_modules
22:52 fannet ah - do I need to create those folders ? I've not created any recipes
22:52 forrest fannet, if /srv/salt/_modules doesn't exist yea.
22:52 fannet just /srv
22:53 forrest of course
22:53 forrest yea just make salt/modules under /srv
22:53 forrest and you can *try* to drop it in there
22:53 forrest I have no idea if it will work as a heads up
22:53 forrest you might have to look and see what else it depends on
22:53 forrest and that might be more work :\
22:53 fannet fingers crossed
22:56 fannet is 2014.7 stable?
22:57 forrest no
22:57 forrest it's probably having fixes applied now
22:57 forrest since it's the next release, it's not ready though
22:57 beneggett joined #salt
22:58 fannet where is the current stable version posted/
22:59 Luke_ joined #salt
22:59 fannet http://docs.saltstack.com/en/latest/
22:59 fannet found it
23:02 forrest yea 2014.1.7 is stable, but sometimes features aren't in there :\
23:02 laubossl1nk Hey everybody, i would like to know what is the best way to run a state just one time on each server ?
23:03 forrest Someone should have done a versionadd for redis
23:03 bmatt laubossl1nk: what do you mean "just one time"?
23:03 laubossl1nk bmatt: there is something like a flag, and if there is this flag, don't run a second time de highstate
23:03 laubossl1nk bmatt: mean like an installation of program
23:04 bmatt laubossl1nk: so while that's possible, it's usually a bad idea - or at least, it's worth the time to try and figure out another way
23:04 bmatt the idea is that a state should describe how a system should be, and not describe things to do
23:04 bmatt (the word is "idempotent")
23:05 forrest such vocabulary bmatt, wow
23:05 forrest :P
23:05 bmatt :-p
23:05 jalaziz joined #salt
23:06 bmatt laubossl1nk: think more in terms of "these files should exist" or "this package should be installed"
23:07 rap424 joined #salt
23:12 ajolo joined #salt
23:12 bhosmer joined #salt
23:15 mosen joined #salt
23:15 yomilk joined #salt
23:16 laubossl1nk bmatt: thanks :)
23:18 zz_Cidan joined #salt
23:24 acabrera joined #salt
23:28 oncallsucks joined #salt
23:30 lz-dylan so, salt-cloud doesn't really feel like it's feature-complete. Anyone know if there's revs coming up in 2014.7?
23:30 manfred lz-dylan: there are several things coming in 2014.7
23:30 manfred the caching portion and the ability to auto bootstrap cloud servers if they are new in your environment
23:30 manfred https://github.com/saltstack-formulas/salt-cloud-reactor
23:31 beneggett joined #salt
23:32 honestly salt-cloud-reactor sounds like a nuclear reactor model
23:34 jalaziz_ joined #salt
23:34 ggoZ joined #salt
23:36 yomilk joined #salt
23:44 talwai joined #salt
23:46 talwai Anyone have a handy script to validate SLS files prior to something like a git commit? Can't use a standard YAML validator because of Jinja tags
23:48 forrest talwai, you can try https://github.com/jesusaurus/salt-shaker
23:48 forrest I don't know if jesusaurus ever put in support to handle jinja takgs though
23:48 forrest *tags
23:51 Hollinski joined #salt
23:52 Supermathie joined #salt
23:59 Dattas left #salt

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