Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-07-31

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

All times shown according to UTC.

Time Nick Message
00:04 krissaxton joined #salt
00:09 jacksontj joined #salt
00:13 jeddi is the persistent error available in a pastebin?
00:13 jeddi oh - at the bottom of that first one .. lotsa debug info .. hang on.
00:14 jeddi nate888: did you try reducing your set of state files down to the absolute minimum to see if this error persists and/or is reproducible with a set small enough for you to share the lot?
00:18 kenbolton joined #salt
00:20 jeddi btw, it might be the pastebin you're using, but the indentation seems whacky on that fragment you posted - 3 spaces on everything.  with my file.managed: indendations i have to do user, source, group etc indented the next level across.
00:21 jeddi nate888: okay, well, i just used that fragment (verbatim) from your pastebin .. and .. it works for me.  i know that isn't very useful, but it does suggest that your problem isn't that fragment.  or salt (0.16)
00:22 jessep joined #salt
00:22 jeddi first run through created the file fine.  second run through -->  http://rn0.ru/show/LIznP21xxoPORgje1OUz/
00:22 jeddi (after changing the placeholder file)
00:28 bastion2202 joined #salt
00:28 andrew joined #salt
00:32 z0rkito quick question on pillars.  if a pillar is a #!py file does the script get run master side or minion side?
00:34 druonysuse joined #salt
00:44 aat joined #salt
00:44 jeddi z0rkito: take this with a grain of the proverbial, but i'm pretty sure it's going to run master side.
00:46 dthom91 joined #salt
00:46 blee joined #salt
00:56 jslatts joined #salt
01:03 syngin z0rkito: it runs master-side; i use #!py to seperate my pillar data so that the minions don't see private pillar data.
01:05 krissaxton joined #salt
01:09 raydeo joined #salt
01:10 Nexpro joined #salt
01:14 z0rkito syngin, jeddi: ty
01:17 timl0101 joined #salt
01:20 whit joined #salt
01:21 Nexpro joined #salt
01:22 jacksontj i'm trying to use state.show_sls and i'mgetting a TypeError
01:22 jacksontj in mesgpack
01:26 oz_akan_ joined #salt
01:28 saurabhs left #salt
01:28 intchanter joined #salt
01:42 jacksontj looks like python-msgpack was out of date
01:44 liuyq joined #salt
01:44 liuyq joined #salt
01:52 Singularo joined #salt
01:56 xl1 joined #salt
02:03 timl0101_ joined #salt
02:05 krissaxton joined #salt
02:05 kstaken joined #salt
02:08 dthom91 joined #salt
02:11 samsalt joined #salt
02:12 samsalt1 joined #salt
02:12 Singularo joined #salt
02:15 Singularo joined #salt
02:24 Singularo joined #salt
02:25 oz_akan_ joined #salt
02:34 tsheibar joined #salt
02:34 kstaken joined #salt
02:37 Furao joined #salt
02:43 napperjabber joined #salt
02:59 lvicks joined #salt
03:06 krissaxton joined #salt
03:06 capricorn_1 joined #salt
03:07 tris198 joined #salt
03:08 tsheibar joined #salt
03:10 tris198 Hia All, I need to write some yaml to a target machine (from a pillar).  Anyone know of a solution using current states before I write a new one?
03:12 napperjabber joined #salt
03:15 tris198 Scratch that - just found the file.serialize docs :)
03:39 kcb joined #salt
03:39 nmistry joined #salt
03:41 raydeo joined #salt
03:55 timops joined #salt
04:03 oz_akan_ joined #salt
04:05 pdayton joined #salt
04:06 kstaken joined #salt
04:06 krissaxton joined #salt
04:08 jalbretsen joined #salt
04:13 kstaken joined #salt
04:21 pdayton joined #salt
04:28 tsheibar anyone know how to refer to a pillar variable from within a file source?
04:28 EugeneKay A file.managed "source"? Set it up as a jinja template
04:28 tsheibar like….- source: file://pillar['salt_base']/files/whatevs.jar
04:28 tsheibar ah
04:28 EugeneKay Oh, that
04:28 tsheibar yeah
04:28 EugeneKay sls statefiles are jinja
04:29 EugeneKay - source: {{pillar['salt_base']}}/files/foo.tt
04:29 EugeneKay You can use all of the standard {% if pillar['some_var'] is somecondition %} stuff too
04:30 tsheibar awesome
04:30 tsheibar ugh....UndefinedError: 'dict' object has no attribute 'salt_base'
04:30 EugeneKay Extremely.
04:30 EugeneKay Hehehe. Make sure that's set :-p
04:31 EugeneKay Possibly by using an aforementioned if block and the "is defined" condition ;-)
04:31 tsheibar ha…thought I did :) in top.sls base -- salt_base: /vagrant/salt
04:32 EugeneKay D'oh
04:32 tsheibar this is a standalone minion
04:32 EugeneKay In any case, you dont' need to do it this way
04:32 EugeneKay Just use salt://
04:32 tsheibar ok
04:32 tsheibar oh nice
04:32 tsheibar duh
04:32 aat joined #salt
04:32 tsheibar processing....
04:33 EugeneKay And also remember that you can specify multiple sources, great for having a machine-specific config file, or using the default one if it doesn't exist
04:38 tsheibar :/
04:38 tsheibar same issue
04:38 EugeneKay Well, the var isn't defined. :-p
04:38 tsheibar Dr Phil: "You jus' don't get it, do ya?!"
04:38 EugeneKay Run pillar.data
04:39 tsheibar thank you…this is like Starcraft for me….the level's dark, but clues like that light up a new part of the map
04:40 Lue_4911 joined #salt
04:46 tsheibar awesome
04:46 tsheibar EugeneKay: thanks again
04:47 EugeneKay Sure.
04:55 nate888 thx jeddi
05:07 krissaxton joined #salt
05:20 jsummerfield joined #salt
05:57 abele joined #salt
06:03 matanya joined #salt
06:04 berto- joined #salt
06:07 krissaxton joined #salt
06:23 bud joined #salt
06:24 middleman_ joined #salt
06:36 Newt[cz] joined #salt
06:45 Zurullo joined #salt
06:47 Zurullo joined #salt
06:58 malinoff joined #salt
06:58 malinoff Hi all
06:59 malinoff I'm trying to install wordpress v3.0.4 via salt.
06:59 malinoff Here is my state:
06:59 malinoff wordpress:
06:59 malinoff pkg.installed:
06:59 malinoff - version: 3.0.4
07:00 malinoff BUT, my CentOS returns
07:00 malinoff ----------    State: - pkg    Name:      wordpress    Function:  installed        Result:    False        Comment:   The following packages failed to install/update: wordpress=3.0.4.        Changes:
07:00 malinoff Seems like salt is not adding spacebars around comparsion sign for yum
07:00 malinoff Any ideas how to figure that?
07:02 berto- joined #salt
07:07 ml_1 joined #salt
07:08 krissaxton joined #salt
07:09 fragamus_ joined #salt
07:12 bud malinoff: are you sure that's the proper version for wordpress?
07:12 bud I am asking because on Debian the version is something like 1:1.6.0-deb-afsg...
07:13 bud I haven't worked with CentOS from quite some time but what yum info returns about wordpress?
07:24 balboah joined #salt
07:34 malinoff Well, yum install wordpress = 3.0.4 works fine
07:35 moko joined #salt
07:37 syphar joined #salt
07:44 carlos joined #salt
07:45 moko How do I use cp.get_url within a state to download a file on a remote server over HTTP?
07:46 malinoff Oops, my bad. Nvm.
07:46 Singularo Oooo I've just used a cmd.run with wget so far. nice to know there is a build in to lookup.
07:52 backjlack joined #salt
07:57 bud joined #salt
07:59 bud Morning everyone, sorry for the noise, first time testing irssi. I want to have the states in one tmux window and the irc channel on another :).
08:00 felixhummel joined #salt
08:04 bud joined #salt
08:08 krissaxton joined #salt
08:13 qba73 joined #salt
08:21 paraglade joined #salt
08:26 malinoff joined #salt
08:32 bud Anyone familiar with tinc vpn? I would like to parse the host configuration file in tinc for each host and have it available in the mine so I can connect new nodes to the VPN automatically.
08:34 bud joined #salt
08:42 zooz joined #salt
08:42 krissaxton joined #salt
08:45 bud The config for tinc looks like this  - http://pastebin.com/8aYccRYV.
08:47 bud I would like to have all of that configuration available in the mine for each tinc node, so when a new node comes up just iterates over the mine.get values and configuires itself, also it generates its config files and push them to the mine so others can welcome the node :).
08:47 p3rror joined #salt
08:48 bud Where should I start in terms of Salt and Python? I guess I would have to create a module that parses the configuration. If that's the case, can I store the public key in a key - value item?
08:50 Newt[cz]1 joined #salt
08:52 backjlack joined #salt
08:56 Jason-AVST joined #salt
09:17 zooz hi people
09:18 zooz how do I tell salt not to try on the minions which are down for longer then x amount of seconds?
09:19 zooz executing salt '*' test.ping takes longer just because some of the minions are gone or not responding
09:19 zooz is there a way to speed this up?
09:25 krissaxton joined #salt
09:32 helderco joined #salt
09:32 lemao joined #salt
09:52 lyddonb_ joined #salt
09:55 chubrub joined #salt
10:16 malinoff @zooz, you can try decrease timeout to 1 second
10:17 xl1 Can _modules reside in gitfs? Or does it have to be in file_roots, and won't get synchronized to minions if in gitfs?
10:17 fredvd joined #salt
10:19 zooz malinoff, does not seem to help
10:22 Furao xl1: yes my _states, _minions and _returners are in gitfs
10:26 malinoff @zooz, for me "salt \* test.ping -t 1" is twice faster than "salt \* test.ping" for 85 hosts (23 are dead)
10:28 aleszoulek joined #salt
10:29 chubrub hi all, I have strange problem with uwsgi server. Function service running returns true, but on the minion uwsgi is stoped <wtf>. All other services like nginx, psql are running correctly. Here is my sls for uwsgi: http://pastebin.com/qrKsBNTn
10:30 Furao don't look like a salt problem to me
10:31 chubrub when I start uwsgi manually (via init.d) server is starting fine
10:31 Furao but - pkg: uwsgi-plugin-python should be in uwsgi: - require:
10:31 zooz malinoff, interesting, I will investigate further, thanks
10:31 Furao no - uwsgi: service.running: - require
10:33 bud joined #salt
10:33 chubrub yeah, #Furao, you're right!
10:33 chubrub but
10:34 jslatts joined #salt
10:34 chubrub this does not change the fact, that salt return true for service.running, when the service is not running :/
10:34 Furao maybe service is started and quickly fail
10:34 Furao but salt < 0.16 do have some issues with /etc/init.d/$service on ubuntu
10:35 Furao and salt 0.16 is so buggy that I can't use it, so I use sig argument to make it check if process is running
10:37 xl1 joined #salt
10:38 bud I also encountered some problems with Unbound service on Debian.
10:38 bud At each run Salt sees Unbound as being dead and it reloads the service.
10:38 bud That's not nice because the cache is gone when doing this.
10:39 bud Furao: can you please tell us how exactly do you mean by sig argument?
10:39 Furao sig
10:39 Furao The string to search for when looking for the service process with ps
10:40 chubrub I'm using 0.15.1 on wheezy - crap, I'm quite outdated :/
10:41 Furao chubrub: at least you're not in 0.15.0 it barely work
10:41 Furao chubrub: don't upgrade to 0.16, wait for 0.16.1 or 0.17.0
10:41 tomeff joined #salt
10:42 bud [INFO    ] Executing state service.running for unbound
10:42 chubrub checking if sig arg will help ...
10:42 bud [INFO    ] Executing command 'service unbound status' in directory '/root'
10:42 bud [INFO    ] Executing command 'service unbound start' in directory '/root'
10:43 bud service unbound status returns "unbound is running"
10:44 bud sig is an argument for service.running?
10:44 bud Yes it is :).
10:46 bud Thanks Furao! Now the service is not started if it is running. DNS cache is safe :).
10:48 bud Well, the thing is that before adding sig there was no searching for ps, just service status.
10:48 chubrub @Furao: Thank you very much - sig arg solved the problem :)
10:49 whiskybar joined #salt
10:57 giantlock joined #salt
10:58 xl1 thanks Furao. gitfs does work, it was a syntax error in the module, found it in the minion log
11:06 mikedawson joined #salt
11:12 koolhead17 joined #salt
11:16 Furao look like I just solve 3 problems in few minutes :P
11:20 david_a joined #salt
11:22 bud Furao: indeed :). I have been meaning to ask you if it was complicated to set up Salt with Sentry?
11:23 Furao you mean install sentry trough salt or hook sentry returner with salt?
11:23 bud Both :-).
11:23 Furao well, I don't remember as I did sentry state a year ago and I maintained it since then
11:24 Furao my sentry state depends on many other states (uwsgi, nginx, postgresql.server) that had also be written a long time ago
11:25 Furao actually, I sell my states as a subscription service + salt consulting
11:25 bud I understand. To hook sentry with salt you had to code a returner or is there something internal?
11:25 Furao I'm not alone I hired 1.5 full-time equivalent to work on this in the past few months
11:25 Furao it's a returner
11:25 Furao you need to run --return sentry
11:25 Furao along with regular salt command
11:26 Furao https://github.com/saltstack/salt/blob/develop/salt/returners/sentry_return.py
11:26 bud Ok, thanks. After I boot up the primary services thoughout the infrastructure I plan to implement Sentry.
11:26 Furao in fact I'm the original author of this module, look like someone sent some fix.
11:27 Furao but my own version is different now
11:28 unicoletti_ joined #salt
11:30 mikedawson joined #salt
11:33 bud Long time user Furao. I am curious how my state tree would look like after 1 year :).
11:34 Furao as we're almost 3 working on those states, they're is a lot of it
11:34 Furao and I wrote a testing framework, which is great to see if every states work before apply them somewhere
11:34 Furao find . -type f -name '*.sls' |  wc -l
11:34 Furao 700
11:35 bud Gee man, that's a lot!
11:36 robbyt joined #salt
11:36 bud 26 is my response to your find.
11:36 Furao running a single set of test on these states on a server with SSD drive take 9 hours. on regular non-SSD amazon instance it's 13 hours
11:36 Furao hehe
11:37 Furao I wish I had only 26 :)
11:37 Furao I wrote a way to automate monitoring checks base on what is installed on each minion
11:37 Furao that is a huge time saver
11:38 bud My old tree, on which I got started with Salt, had almost 100.
11:38 bud Started from scratch though.
11:39 Furao find . -type f -name '*.sls' -print0 | xargs -0 cat | wc -l
11:39 Furao 24949
11:39 Furao but my states often contains some doc
11:40 bud That's not bad at all, docs are a huge time saver, especially if new people come in the team.
11:41 Furao I think I'm going to add a doc/ subfolder into each states and move doc there and have sphinx build it
11:41 bud Furao: have you ever configured tinc VPN (with Salt)?
11:42 Furao I tried tinc in 2003 or something because I didn't wanted to use Cisco VPN client
11:42 Furao but I don't think it's used for that anymore
11:43 bud I am trying to automatically configure tinc on a bunch of hosts.
11:43 bud Starting with one of course. The configuration layout is simple - http://pastebin.com/8aYccRYV.
11:43 Furao one client asked me to look for mesh vpn and I think I'm going to use OSPF (trough quagga) over OpenVPN
11:44 bud I need to get the public key of one minion to all other active minions. Have you got some ideas/pointers?
11:44 Furao use salt mine
11:44 Furao or centralize key creations
11:44 bud That's what I've thought also. But how do I get a public key in there?
11:45 Furao write your own module that send the content of the file into the master
11:45 bud I would like to keep the private keys private and don't transport them from one host to another.
11:45 Furao and use mine.get to retrieve those public key
11:45 Furao and this way you can also get the list of minion your client need to connect to
11:46 Furao as a non-existent minion won't be in salt mine yet
11:46 bud And I would have the public key available in the mine with such a module?
11:46 bud Can a key hold a public key as its content?
11:46 Furao well the module can open('/etc/tinc/$filename', 'rb') as pubkey: return {'tinc': {'pubkey': pubkey.read()}}
11:49 jslatts joined #salt
11:50 bud tinc stores the public key of the host in /etc/tinc/net_name/hosts/`hostname -s`.
11:50 BRYANT__ joined #salt
11:50 bud That if you use the tincd key generation method.
11:50 mmilano_ joined #salt
11:50 Furao then you can make it work with more python code than that single line :P
11:50 bud I have switched to generating keys with openssl so I have them in /etc/tinc/net_name/key.[priv|pub].
11:51 Furao import socket; socket.gethostname()
11:51 bud This way what you say it's really simple even for my Python skills :-).
11:52 dzen left #salt
11:52 Furao salt mine is just a single salt module.func_name that run periodically and send it's output to the master
11:53 Furao then other minion can ask the master what is the content of that module execution for a specific minion (or those that start with 'z*' or all of them, like salt command)
11:53 Furao it's a lot better than publish.publish as it make that data available even if minion is offline
11:53 Furao I use it to ask minion what are they roles and monitor them accordingly
11:54 bud Furao that's what I thought also but honestly I am just starting out with python.
11:54 Furao there is plenty of example of module in salt code
11:54 bud Didn't have a clue if I can store a public key in a key that goes into the mine and how to accomplish that.
11:55 Furao in one of my repo I have 9 file in _modules
11:56 Furao for relatively complex tasks, it's often better to do it with custom modules/states/ext_pillars than dealing with .sls file + publish.publish
11:58 bud Never used publish.publish so far.
11:59 bui joined #salt
11:59 fart_lander joined #salt
11:59 backjlack joined #salt
12:02 ede joined #salt
12:05 blee joined #salt
12:06 blee betters
12:10 jeddi joined #salt
12:12 aranhoide joined #salt
12:20 bhosmer joined #salt
12:25 [diecast] joined #salt
12:29 bhosmer_ joined #salt
12:29 diegows joined #salt
12:30 fragamus_ joined #salt
12:36 krissaxton joined #salt
12:37 krissaxton left #salt
12:45 cewood joined #salt
12:48 iMil joined #salt
12:48 iMil heya there
12:49 iMil I'd like to set a variable within a state in order to fill a grain with it
12:49 iMil typically
12:50 iMil a cmd output gives a hash, which I'd like to store in a grain
12:50 iMil is there some way to achieve this?
12:51 LGSilva joined #salt
12:51 elfixit joined #salt
12:51 iMil more generally, is there a way to store cmd's return values in order to use them within the same state
12:52 LGSilva hello. Is there a way to check if the sls exists before trying to include it on pillars? We would like to have overwrite files so we don't have to change the pillar top file everytime
12:58 Xeago joined #salt
12:58 ioni joined #salt
13:01 KennethWilke joined #salt
13:02 brianhicks joined #salt
13:02 bud iMil: try module.run state, let's you run modules from states. grains.setval lets you set grains to specific value.
13:04 iMil bud: yup, but they are callable as-is, they do not permit to "store" a value for future use inside the state right ?
13:09 bud iMil: I don't understand what you mean. You want to store a value in grains and use it in the same state afterwards?
13:09 iMil nope, I run a command, using cmd.run, which returns a value which I want to store in a grain
13:10 Gifflen joined #salt
13:10 bud Don't know if that's possible, you'd have to ask others...
13:10 iMil thanks anyway :)
13:10 iMil I'll dig in myself
13:10 juicer2 joined #salt
13:10 iMil will tell if I find something
13:11 bud I am curios what are you trying to achieve with that :-).
13:11 iMil well, simly put, I an writing a state to deploy Zend Server
13:12 iMil when it installs, using their command line, it will output an API key
13:12 iMil which you need in order to setup various aspects of the server
13:12 iMil wow so much typos in my previous sentences...
13:12 iMil sorry 'bout that
13:13 KennethWilke good morning!
13:13 iMil hi KennethWilke
13:13 Newt[cz] joined #salt
13:13 KennethWilke iMil: i came into your conversation a touch late, but you're trying to store the result of a cmd.run into a grain?
13:13 iMil exactly
13:14 bud Hi KennethWilke.
13:14 iMil I can see many uses to that kind of operation :)
13:14 KennethWilke i think it's considered a little taboo to modify things this way in a state file, but i think you can do a grain.set similar to how you'd do a cmd.run
13:14 KennethWilke does cmd.run return the output of what it runs?
13:15 jeffasinger joined #salt
13:15 iMil KennethWilke: it does, but I don't know how to retrieve it within the state
13:15 timl0101 joined #salt
13:16 bud iMil: pipe it to a file in tmp or something and run grains.setval on that piping it back :).
13:16 iMil well I thought about that
13:16 iMil *but*
13:16 iMil how can you grains.setval then
13:16 KennethWilke hmm, i'm gonna tinker with this to see if i can be of help, no promises! :)
13:17 iMil I mean, grains.present takes a value as an argument
13:17 iMil not plain file text
13:18 bud grains.present? First time I hear about that.
13:18 iMil https://salt.readthedocs.org/en/v0.16/ref/states/all/salt.states.grains.html#module-salt.states.grains
13:18 iMil same would apply to module.run grain.setval actually
13:19 iMil KennethWilke: thanks :)
13:19 iMil I'll scratch my head too
13:20 KennethWilke seems like something you should be able to do no problem
13:20 bud You're right iMil, spoke too soon.
13:20 KennethWilke just, how!? lol
13:20 iMil :)
13:20 ipmb joined #salt
13:20 iMil yeah at first sight it seems easy doesn't it ? :)
13:21 anteaya joined #salt
13:23 iMil a *very* ugly wy would be to cmd.run salt grains.setval `cat /tmp/foo`
13:23 iMil but that's beyond what I can possibly write without puking.
13:23 bud Does anybody know how can I iterate over all minions data from the mine?
13:24 bud With jinja :).
13:24 KennethWilke iMil: yeah lol, i could see doing that, but it would sadden me
13:24 iMil indeed
13:25 KennethWilke are you writing module or state module for this zend thing you're doing? i'm thinking it might be easier to address on that level of things
13:25 aleszoulek joined #salt
13:26 oz_akan_ joined #salt
13:26 KennethWilke bud: i still need to tinker with salt mine myself, so i'm of  no use to you there :(
13:26 iMil np guys you've been helpful :)
13:27 bud KennethWilke: no problem, thanks anyway. Just started with it, it's fun and learning mode is on :).
13:27 KennethWilke lol sounds nice
13:28 ml_1 joined #salt
13:30 Khollowa_ joined #salt
13:30 whit joined #salt
13:34 honestly joined #salt
13:34 racooper joined #salt
13:35 oz_akan_ joined #salt
13:37 C^^ joined #salt
13:37 saltstarter joined #salt
13:37 C^^ left #salt
13:38 saltstarter I have a question realted to the use of the pillar  saltstack-formulas/users described here https://github.com/saltstack-formulas/users
13:38 djn joined #salt
13:38 toastedpenguin joined #salt
13:39 bud saltstarter: shoot, maybe we can help.
13:39 saltstarter on the master i have all files imported but it is unclear to me how to ask the pillar to be targeted for execution on a remote machine and more importantly where to list file directory) the list of users i want to manage
13:40 saltstarter naive question but starting
13:41 bud The pillar is loaded by all the minions, based on your targeting in pillar top.sls.
13:41 saltstarter ok good
13:42 bud So the minions have the data available after startup. If you change the pillar on the way issue salt '*' saltutil.refresh_pillar.
13:42 bud So all the minions refresh the global data pillar is.
13:42 bud Are you just starting?
13:42 saltstarter so to target a machine specifically to test i just add the targettingn in the top
13:42 bud With Salt I mean.
13:43 saltstarter yes just starting get the salt-cloud running on cloudstack cloud and want to experiment more
13:43 bud Provided you haven't changed any default config the pillar data is loaded from /srv/pillar.
13:43 saltstarter yes
13:43 saltstarter i have done salt 'mymavhine' pillar.items
13:43 bud I would mkdir /srv/pillar/users and create the init.sls files with all the users there.
13:44 bud And then create the top.sls pillar, simple base:\n  - user, in /srv/pillar.
13:44 saltstarter i was looking at using saltstack-formulas/users
13:44 bud Refresh the pillar with the above command and there you have your data available to minions.
13:46 bud You can try to see if the pillar is refreshed using salt '*' pillar.items users.
13:47 napperjabber joined #salt
13:47 kaptk2 joined #salt
13:48 saltstarter ok thanks going to play with the information i have and will ask again question if i need
13:49 saltstarter really like all the possibilities that are open with salt real creazy flexibility
13:49 saltstarter investigating the use for our overall automation
13:51 bud Me too, I love Salt. I have started to automate from the ground up. Thinking we have nothing left just the repos and build up from there.
13:54 andrew joined #salt
14:00 matanya joined #salt
14:07 mmilano joined #salt
14:08 lyddonb_ joined #salt
14:09 aat joined #salt
14:10 bostonian joined #salt
14:12 bostonian When running state.highstate across my network, I get an error related to get_cli_event returns (full trace: http://pastebin.com/HeeqKxse). It eventually just stops trying with a "Write failed: Broken pipe"... is this a configuration error or just a bug with salt?
14:12 lvicks joined #salt
14:13 bostonian Could this be due to modifying iptables with salt?
14:13 drawsmcgraw joined #salt
14:18 carlos joined #salt
14:22 StDiluted joined #salt
14:22 LGSilva what's the command to return the salt version the minion is running?
14:25 iMil should be shown by the grains I think
14:25 iMil yup
14:25 iMil saltversion:
14:25 iMil 0.15.3
14:26 [diecast] joined #salt
14:26 [diecast] joined #salt
14:26 LGSilva thanks iMil
14:26 iMil yw
14:31 aranhoide left #salt
14:34 bud n
14:39 syphar joined #salt
14:42 bud Is there a command that can list q!
14:42 Psi-Jack q?
14:42 bud Sorry, wrong window :).
14:42 bud a command that can list all data from the mine.
14:43 Psi-Jack heh
14:43 Psi-Jack uhh, mine?
14:44 JasonSwindle joined #salt
14:44 cnelsonsic joined #salt
14:44 Ivo joined #salt
14:44 JasonSwindle Howdy.  Anyone set grains by hand from CLI on the salt master?
14:44 JasonSwindle It seems to work, and then not
14:45 timl0101_ joined #salt
14:46 JasonSwindle Example; https://gist.github.com/JasonSwindle/0256c687e6af8f7e78ed
14:46 bud salt 'w*' grains.setval x y works for me.
14:46 bud Tried it a couple of times.
14:46 bud They are in /etc/salt/grains.
14:47 JasonSwindle etc/salt/grains on the minion?
14:47 whit joined #salt
14:47 DanGarthwaite joined #salt
14:48 JasonSwindle I see the grain is there
14:48 JasonSwindle hmm
14:49 drawsmcgraw Is there a way of managing a local repo, similar to the Windows repo? Say I have some custom RPMs and I want to ensure they're installed with states.pkg.installed
14:53 drawsmcgraw nm. I think the answer is in the 'sources' parameter to pkg.installed
14:54 m_george|away joined #salt
14:59 JasonSwindle bud, can you delete?
14:59 m_george left #salt
15:00 j___ joined #salt
15:00 jsummerfield joined #salt
15:02 chrisgilmerproj joined #salt
15:08 conan_the_destro joined #salt
15:09 syphar joined #salt
15:09 terminalmage drawsmcgraw: the sources param should do what you want, yes. this will install binary packages. if you have them in an actual yum repo, you can just install them like any other package
15:10 drawsmcgraw terminalmage: Yeah, it'd be nice if we could go the yum repo route but, alas
15:10 drawsmcgraw I'm *almost* there figuring out the incantation on the command line
15:11 drawsmcgraw Following the example on the docs and running a Highstate just comes back with "Package failed to install" so I'm working on the cmd line to debug
15:14 isomorphic joined #salt
15:15 drawsmcgraw terminalmage: So... Looking at the minion logs, I'm seeing --> Install failed: No package(s) available to install
15:15 Corey Good morning.
15:15 Corey drawsmcgraw: Yes, yum repo that. :-p
15:15 JasonSwindle Corey:  Morning
15:15 Corey JasonSwindle: Ahoy!
15:16 Corey "Use configuration management. Package your stuff. Use a yum/apt repo." <-- This describes my entire consulting career
15:16 drawsmcgraw Corey: Ah... point made :)
15:17 Corey drawsmcgraw: If you're getting stuck on those bits, let me know. I can work with you.
15:17 forrest joined #salt
15:17 Corey drawsmcgraw: I just don't want to find myself in your environment four years from now cleaning it up. :-D
15:17 drawsmcgraw Fair enough
15:18 whit_ joined #salt
15:18 drawsmcgraw So, with full knowledge of the debt I'm taking on: What would be causing that error in the minon logs when I try to install this rpm via pkg.install ?
15:18 drawsmcgraw I've confirmed the RPM actually works by installing it by hand.
15:19 Corey drawsmcgraw: Did you pastebin the state / debug output?
15:19 Corey I'm lazy and scrolling is hard. :-p
15:19 drawsmcgraw Corey: Error from the minion logs --> Install failed: No package(s) available to install
15:20 jeffasinger Anyone have any suggestions on how I might deal with migrating a postgres database with a salt state? Write my own state in python?
15:21 Corey drawsmcgraw: What does your state look like?
15:21 terminalmage drawsmcgraw: keep in mind that the package name you see when you run "rpm -qpi filename_of_pkg.rpm" must match in the SLS
15:21 terminalmage drawsmcgraw: also, are you trying to install 32-bit binary packages on x86_64? There is a known issue with this
15:21 drawsmcgraw terminalmage: Yeah, I got past that hurdle (that the name must match what the rpm says)
15:21 drawsmcgraw one second....
15:22 terminalmage I'm actually working to get that issue fixed right now
15:22 teskew joined #salt
15:23 terminalmage drawsmcgraw: in case that is the case and you want to watch the issue on github, here it is: https://github.com/saltstack/salt/issues/6303
15:23 devinus joined #salt
15:23 TOoSmOotH quick question.. in a state can I specify a list of servers? so like 'server1,server2'
15:24 terminalmage TOoSmOotH: what is the use case?
15:24 drawsmcgraw terminalmage: state --> http://pastebin.com/Q9yBeppP
15:25 terminalmage drawsmcgraw: I usually put a space after the dashes, I'm pretty sure that's not the problem though
15:25 terminalmage drawsmcgraw: for which arch is the jre pkg you are trying to install
15:25 TOoSmOotH terminalmage:  I have a state like base: '*' and then I want to have server1 server2 to use somethign else
15:25 TOoSmOotH if I explained that right
15:25 drawsmcgraw 'linux-64' ?
15:25 TOoSmOotH right now for each server I have a separate wildcard
15:26 TOoSmOotH just trying to put it into one section vs repeating
15:26 TOoSmOotH if that makes sense what I said
15:26 TOoSmOotH so like I do *1 - blah and '*2' - blah
15:26 terminalmage drawsmcgraw: ok, and what is the output of "rpm -qpi filename.rpm | grep Name"
15:27 terminalmage drawsmcgraw: replace filename.rpm of course with the filename of the rpm
15:27 drawsmcgraw terminalmage: 'jre'
15:27 drawsmcgraw I ran into the 'filename' issue earlier. I had JRE in my state instead of 'jre'
15:27 terminalmage drawsmcgraw: anything in the logs about a failure to cache the package?
15:27 drawsmcgraw And Salt complained that it didn't match
15:28 drawsmcgraw Negative. Only the one I pasted into here earlier
15:28 UtahDave joined #salt
15:29 terminalmage TOoSmOotH: so you want to apply different states to every host besides server1 and server2?
15:29 terminalmage like, different SLS entirely?
15:29 TOoSmOotH I want to say server 1 and servr 2 use this state and server 4 and 5 use that state
15:29 TOoSmOotH 'A*':
15:29 TOoSmOotH - sensor
15:29 TOoSmOotH 'C*':
15:29 TOoSmOotH - backend
15:30 TOoSmOotH that is in top.sls
15:30 TOoSmOotH but I have a sensor named bean
15:30 TOoSmOotH so I don't want to add 'B*' - sensor in there
15:30 terminalmage drawsmcgraw: not sure what to tell you, except maybe to increase the loglevel
15:31 TOoSmOotH is 'Apple, Bean': - sensor legit?
15:31 TOoSmOotH I see in non states you can pass it a lit
15:31 drawsmcgraw terminalmage: Maybe I'm just losing my mind... The error (No package(s) available to install) persists on the Minion. B-U-T the output from running a Highstate tells me that the package is already installed.
15:31 TOoSmOotH like salt -L blah
15:31 terminalmage drawsmcgraw: stop the salt-minion service, then su to root using 'su -' and run: salt-minion -l debug
15:31 UtahDave TOoSmOotH: yeah, but you have to specify that you're matching on a list
15:31 drawsmcgraw terminalmage: Ah, I'll do that. Thanks
15:32 drawsmcgraw I've been tailing the log file but I think it's only set to 'error'
15:32 terminalmage drawsmcgraw: hrm.... ok, I'll investigate the error message in the log
15:32 auser joined #salt
15:32 auser hey all
15:33 JasonSwindle auser:  Hey!
15:33 com48 joined #salt
15:33 terminalmage in pkg states, there is first a step that checks to see which of the targeted packages need to be installed. if nothing needs to be installed, then it shouldn't try to install anything
15:33 terminalmage I'm wondering if it's still proceeding to try to install nothing
15:33 terminalmage sudo yum install
15:33 terminalmage er, yum install
15:33 terminalmage and maybe that is causing the log entry?
15:33 terminalmage I dunno
15:34 forrest Has anyone set something up in salt where you basically rename a file (to drop a server from the LB), monitor the logs till certain conditions are met, and then restart the box, and rename the file after that completes?
15:34 terminalmage TOoSmOotH: if you want to match a list of hosts you do it with - match: list right under the target expression
15:34 terminalmage like you would for a grain match
15:34 forrest Someone suggested Monit yesterday, but I'm curious as to whether it can be done with salt
15:34 TOoSmOotH ok cool thanks
15:34 forrest so basically you're rocking rolling restarts
15:35 terminalmage TOoSmOotH: http://dpaste.com/1324919/
15:35 com48 hey everyone. I'm trying to setup some repo's in saltstack and wonder what the best way to do it is. i need to replace the default os and update repo's from a fedora with a internal mirror and make salt use those.
15:35 drawsmcgraw terminalmage: Yeah I'm not certain what would cause that error line to be dropped into the logs. But the reason that was showing up (I'm nearly certain) is that *it was already installed*
15:35 forrest com48, how about: http://docs.saltstack.com/ref/states/all/salt.states.pkgrepo.html
15:36 TOoSmOotH yea that isn't bad at all
15:36 drawsmcgraw I removed it by hand. Then ran another Highstate. Installed the rpm with no problem.
15:36 drawsmcgraw not paying attention... maybe I need more caffeine
15:36 terminalmage drawsmcgraw: yeah, but if it's already installed, it shouldn't get to the point where it tries to do so. this is something I'll try to replicate. bad log entries are still bugs
15:36 forrest com48, just use baseurl instead of mirrorlist
15:36 terminalmage thanks for bringing this to my attention
15:36 com48 yes i've set that up so far but that means that i need to define - pkgrepo: fedora_os_repo  in all my packages
15:36 drawsmcgraw terminalmage: Noted. Hope it turns out to be a simple issue. Thanks for the help!
15:36 UtahDave forrest: I'm sure you could do something like that with Salt.  You might want to look at the Salt Python api
15:37 com48 is that prefered?
15:37 terminalmage drawsmcgraw: no prob!
15:37 forrest UtahDave, thanks, I will take a look
15:38 com48 i was wondering if it was possible to remove all repo's first then add the pkgrepo's so i don't need to define it in all packages
15:38 auser how goes JasonSwindle
15:39 JasonSwindle auser:  Pretty good.  Just fighting with my require chain
15:39 auser ah yes
15:40 com48 is it possible to make a package with repostories run first? before any other packages?
15:40 auser yeah com48
15:40 auser you can set 'order'
15:40 whiskybar joined #salt
15:40 auser com48: http://www.saltstat.es/posts/dependencygraph.html
15:40 syphar_ joined #salt
15:40 auser all the way at the bottom
15:43 com48 thanks. and that works between different modules as wel? so if i have 2 modules in a top file and one is the one with repostories in it and has the 'order' set it will run before the other commands?
15:44 auser yeah, there is a single dep graph in salt
15:45 com48 great! thanks auser
15:45 auser np com48
15:45 auser :)
15:45 UtahDave auser++
15:46 auser is that equivalent to: auser = auser + 1
15:46 UtahDave :)  yes.  And   auser += 1
15:46 Corey And who's the one?
15:47 auser whoa
15:47 auser that's personal Corey
15:47 Corey You're right, I overstepped. So sorry.
15:47 KennethWilke 'INCR auser' in redis :) atomic ops ftw
15:48 UtahDave lol
15:48 Corey KennethWilke: How *dare* you bring redis into this channel? :-)
15:48 KennethWilke because it's fast!
15:48 KennethWilke lol
15:48 auser hahah, oh god
15:48 auser he said "redis" (girls chattering in the background)
15:48 TheRealBill auser: yes?
15:48 Corey Imma turn on disk persistence. Oh wait, why is its forking model destroying my app?
15:48 KennethWilke it's fast, and simple, and good... I THOUGHT YOU PEOPLE WOULD LIKE THAT
15:48 KennethWilke lol
15:48 Corey fork() is expensive on Xen. Read as: AWS.
15:50 Corey That said, I'm somewhat interested in distributing pillar data via mongodb or similar.
15:50 KennethWilke sounds like xen's problem, fork()'s fairly common!
15:50 Corey It is very much a Xen problem.
15:50 Corey That doesn't make it tolerable. :)
15:50 KennethWilke KennethWilke.helpfulness--;
15:51 Corey My ultimate goal is to turn "a salt master has failed" into a "meh, who cares, we have five more" occurance.
15:51 KennethWilke i like the sound of that
15:51 whit_ joined #salt
15:51 Corey Salt now supports multi master, but how do you handle sync of state trees and pillar data across geographically separate locations/
15:51 TheRealBill so build out miltiple masters and put them behind a load balancing viertual IP?
15:51 KennethWilke external pillar and gitfs saltstates
15:51 TheRealBill virtual*
15:52 Corey TheRealBill: Well Salt now supports multi master, so you don't even have to do that.
15:52 Corey Sure, but gitfs salt states still means you have a single git repository that has to communicate to all masters; I can't think of a good way to do that atomically.
15:53 TheRealBill the masters talk to git, not the other way around. either way, to get pillar data HA you need to do HA pretty much everywhere in the chain.
15:53 KennethWilke i bet you might be able to make it redundant at least in your gitfs_roots or whatever that config bit is called
15:53 KennethWilke gitfs_remotes*
15:53 TheRealBill KennethWilke: true, you could have multiple repos, but then you have to keep them in sync.
15:53 Corey TheRealBill: Right. And that's what I'm pondering.
15:54 Corey TheRealBill: You can also have git talk to your masters, via a post-receive hook or whatot.
15:54 KennethWilke yar sync would be needed
15:54 Corey whatnot*
15:54 auser TheRealBill++ this time
15:54 TheRealBill personally, with my mongodb experience I'd stay away from it. Redis + Sentinel is simpler and more reliable (less complex) than mongodb
15:55 devinus anybody know how i can do an rpm --import with salt?
15:55 TheRealBill with Redis you could even make each master a local slave from the Redis master. That way even in tempfail networking you would still have the data.
15:55 racooper cmd.run?
15:57 TheRealBill then just set up a Redis master with a second slave. Use Redis Sentinel to manage failover between the masters, and you've got multiple layers of redundancy
15:58 TheRealBill you would naturally run Sentinel on the salt masters since they are who need to be able to reach the master.
15:58 teskew devinus: probably with just cmd.run
15:58 TheRealBill and when I release my Redis and Sentinel salt modules you could set it all up from Salt. >:)
15:59 felixhummel joined #salt
16:01 UtahDave nice, TheRealBill!
16:02 auser nice TheRealBill
16:02 KennethWilke don't you touch the redis returner TheRealBill, i'm breaking it myself thank you very much
16:02 KennethWilke :p
16:02 TheRealBill KennethWilke: hehe why do I not doubt that? ;)
16:04 JasonSwindle KennethWilke TheRealBill My redis is so wimpy now.  Thanks guys!
16:06 gwar9999 left #salt
16:07 TheRealBill JasonSwindle: now? ;)
16:07 * TheRealBill is feeling a bit snarky this am
16:07 KyleG joined #salt
16:07 KyleG joined #salt
16:08 KennethWilke lol... this morning
16:09 JasonSwindle TheRealBill:  :)  I will need your help later but we can talk on this down the road.
16:09 * TheRealBill slaps KennethWilke with a trout
16:09 * KennethWilke becomes sad
16:12 * TheRealBill picks up a puffer fish for next time
16:17 FreeSpencer joined #salt
16:20 avienu joined #salt
16:22 Linz joined #salt
16:23 Linz joined #salt
16:23 dthom91 joined #salt
16:26 oz_akan_ joined #salt
16:26 dthom91 joined #salt
16:27 devinus joined #salt
16:31 jacksontj joined #salt
16:31 whit joined #salt
16:32 auser joined #salt
16:32 tomeff joined #salt
16:35 forrest Is there a way to track whether a job is actually complete or not? When using salt-run jobs.active, the dictionary it returns is blank, but I can tell the command hasn't completed running on the minion (I was doing a file rename via the api).
16:35 forrest There's the Start Time in jobs.list_jobs, but that doesn't show completion information
16:39 oz_akan_ joined #salt
16:41 ml_1 joined #salt
16:50 Lue_4911 joined #salt
16:55 Ryan_Lane joined #salt
17:01 troyready joined #salt
17:03 whit_ joined #salt
17:03 jschadlick joined #salt
17:07 druonysus joined #salt
17:07 druonysus joined #salt
17:08 oz_akan_ joined #salt
17:09 DanGarthwaite left #salt
17:13 jkleckner joined #salt
17:14 quantumsummers|c joined #salt
17:15 blee_ joined #salt
17:17 loungin_light joined #salt
17:18 whit joined #salt
17:19 devinus joined #salt
17:22 JasonSwindle joined #salt
17:27 berto- joined #salt
17:29 jschadlick left #salt
17:31 LucasCozy joined #salt
17:33 jeffasinger I know I've seen this somewhere, but how would I match all minions that have a certain value in a grain that's a list. So if there's a grain roles, how would I match all minions with the role "webserver"?
17:36 TheRealBill jeffasinger: http://docs.saltstack.com/topics/targeting/grains.html
17:37 jeffasinger That makes me wonder what I was looking at before ...
17:37 jeffasinger But thanks
17:38 jacksontj joined #salt
17:45 JasonSwindle jeffasinger:  You can get crazy with grains
17:45 JasonSwindle jeffasinger:  http://docs.saltstack.com/topics/targeting/grains.html#grains-in-top-file was what I added just a few days ago
17:46 jeffasinger I realized I was trying to copy the format for a mine.get to be used on the command line
17:48 JasonSwindle ah
17:53 joe_i joined #salt
17:54 joe_i Does anyone have any experience using salt with a sonicwall fw and site-to-site vpn?
17:55 joe_i We're trying it out, and we only have 4 minions but my firewall keeps blocking the responses as syn-flood attacks
17:55 joe_i I don't think that's normal, but I don't know where to start
17:56 joe_i the master is on the other end of the vpn and the packets are being blocked by the firewall on the master side
17:58 jacksontj joined #salt
18:05 tomeff joined #salt
18:05 cron0 joined #salt
18:05 aranhoide joined #salt
18:05 jacksontj joined #salt
18:06 aranhoide left #salt
18:07 LucasCozy joined #salt
18:10 cedwards avienu: ping re: make.conf
18:11 avienu cedwards: Hey
18:11 cedwards avienu: saw your question yesterday, but it was after I left.
18:11 avienu cedwards: No worries. Was just curious how you were handling yours.
18:12 cedwards avienu: I have a make.conf.jail and make.conf.host templates
18:12 FreeSpencer joined #salt
18:12 cedwards avienu: but i realize since i've migrated to using poudriere and creating my own packages it hasn't really been an issue
18:13 qba73 joined #salt
18:13 avienu cedwards: I'm mainly curious like, say I want to stick an option in my /etc/make.conf, like MYSQL_VERSION=55m
18:13 dthom91 joined #salt
18:14 Xeago joined #salt
18:14 avienu cedwards: Seems like salt should be able to somehow manage /etc/make.conf without blowing out box specific information that other packages have added, like: # added by use.perl 2013-07-25 21:11:21
18:14 avienu PERL_VERSION=5.14.2
18:15 cedwards avienu: I guess Perl is upgraded so infrequently I've simply updated my template to include that change manually
18:16 avienu cedwards: Gotcha. I was wondering if I was missing some sort of slick salty handling like the enable support for /etc/rc.conf
18:18 cedwards avienu: I haven't figured out anything more fancy than that..
18:19 terminalmage joe_i: might be related to this? https://github.com/saltstack/salt/issues/5948
18:20 akoumjian Is anybody here familiar with the inner workings of the pkgrepo state?
18:21 joe_i not sure. I've been doing some digging and most of what I've turned up was legitimate syn flood type behavior. A large number of minions, not 4
18:22 joe_i there is some support for reconnect on 4505 when the master reboots or otherwise goes away, but not much for hitting 4506 with returns from a job
18:23 blee joined #salt
18:23 mlister2006 joined #salt
18:24 JesseC joined #salt
18:24 conan_the_destro joined #salt
18:24 mannyt_ joined #salt
18:25 JesseC Can salt be used for compiling other software from source reliably?
18:25 JesseC An example would be pulling php source from a location, configuring, making and installing it.
18:26 joe_i terminalmage: do you have any experience with syndic? I was thinking if we put a master there it would limit the responses that go over vpn
18:26 joe_i I don't know how it works though
18:26 terminalmage I do not, unfortunately
18:29 z0rkito is there a way to run overstate to a single host without having to define that host in the overstate config?
18:31 drawsmcgraw We have a requirement to install an rpm...... into a custom directory
18:31 drawsmcgraw I see that this is done pretty easily on the command line (http://www.cyberciti.biz/faq/howto-install-rpm-package-into-another-directory/)
18:32 drawsmcgraw But do any of the modules (e.g. pkg.install or otherwise) support the    --prefix option for rpm installs?
18:32 JasonSwindle you could cmd.run it
18:32 JasonSwindle RPMs, I am weak at
18:33 jalbretsen joined #salt
18:35 drawsmcgraw JasonSwindle: yeah that was my first thought. The challenge would be putting the rpm on the Minion just long enough to install it.
18:35 drawsmcgraw I'll likely end up just doing that though
18:36 JasonSwindle You could right your own mod or state
18:36 JasonSwindle I have not walked down that path
18:36 drawsmcgraw I'd like to just add the option to pkg.install :)
18:36 JasonSwindle Make it for deb as well
18:36 JasonSwindle and you have a winner
18:36 kstaken joined #salt
18:36 drawsmcgraw JasonSwindle: noted
18:37 drawsmcgraw JasonSwindle: What's the deb version of that?
18:37 drawsmcgraw in yum/rpm, I say --> rpm -ivh awesomesauce --prefix=/new/directory/path
18:38 devinus joined #salt
18:38 toastedpenguin joined #salt
18:39 z0rkito drawsmcgraw: i believe the deb version of that is --instdir
18:40 drawsmcgraw z0rkito: thanks
18:42 bhosmer_ joined #salt
18:42 JasonSwindle drawsmcgraw:  No idea.  I always only use PPAs or etc
18:42 drawsmcgraw JasonSwindle: Looks like, under the hood, it's a different module anyway
18:42 drawsmcgraw http://docs.saltstack.com/ref/modules/all/salt.modules.pkg.html#module-salt.modules.pkg
18:43 z0rkito drawsmcgraw: that's for using dpkg, from apt-get you'd have to pass `-o Dpkg::Options::="--instdir=<insert path here>"`
18:43 drawsmcgraw I see
18:43 drawsmcgraw z0rkito: thanks
18:46 berto- joined #salt
18:46 brianhicks joined #salt
18:46 kstaken joined #salt
18:48 JesseC So. Anyone attempting to compile software through salt?
18:48 KennethWilke JesseC: i've done it before for redis
18:49 JesseC KennethWilke: I'm curious what path you took to achieve that.
18:49 KennethWilke as the package on a lot of distros is too old for me!
18:49 JesseC KennethWilke: Through bash files, or cmd.run ?
18:50 JesseC KennethWilke: That's the same issue I'm having. I need to compile a custom version of php as well as several other pieces that are dependencies to php.
18:50 KennethWilke i used cmd.run, i can't guarantee it's the recommended way but it did work for me
18:50 KennethWilke i'll gist my thinger for ya
18:50 JesseC I see. You'll what your what for me? lol
18:51 LGSilva joined #salt
18:51 KennethWilke https://gist.github.com/KennethWilke/6124942
18:51 KennethWilke lol
18:52 KennethWilke it's a lil ugly as i said, but it'll work
18:52 az87c joined #salt
18:52 JesseC KennethWilke: Thanks, it's all gravy to me. I've been provisioning by hand since forever. I think after setting up my billionth VM by hand its time to move onto something repeatable. Ha
18:53 KennethWilke oh yes! with salt you'll never look back to that stuff
18:53 KennethWilke i try to think of how it'll get done in salt when i do things by hand now
18:54 KennethWilke that way i can think once and win all over the place
18:54 JesseC Gonna use it with vagrant as well. bring up and destroy development environments from any pc? yes please, ha
18:54 teskew is the author of this jinja template in here currently?  https://github.com/ConsumerAffairs/salt-states/blob/master/etc/redis/redis.conf.jinja
18:55 KennethWilke teskew: i'm not the author, but i may be able to answer your question
18:57 teskew what might the pillar file for that redis state look like? i see he grabs redis pillar data, but not sure how it's set
18:57 teskew wanting to setup a master and a slave, but wondering how he determines if it's a slave or a master. what pillar is set to make the determination
19:00 KennethWilke looks like it'd be {'redis' {'master': {'fqdn': {'ip': '127.0.0.1', 'port': 6379, 'auth': 'password'}}}}
19:02 KennethWilke password seems to sit at the same level as the master dict, and it will use the 'SLAVEOF' if theres an IP in that master dict under the fqdn that matches the nodes name
19:12 dthom91 joined #salt
19:16 whit joined #salt
19:18 jacksontj joined #salt
19:21 opapo joined #salt
19:23 qba73 joined #salt
19:24 SEJeff_work Has anyone used the mysql states to grant privileges on databases?
19:25 chuffpdx joined #salt
19:25 SEJeff_work I've got states to create the databases and users with the right password, but I want salt to run the grants
19:25 SEJeff_work mysql_grants, nevermind :)
19:30 berto- joined #salt
19:31 z0rkito anyone know of there's a way to match on pillars in overstate?  using P@<pillarname>:<value> doesn't seem to work the way it does for a grain...
19:32 SEJeff_work z0rkito, P@is pillar regex, not pillar match
19:32 SEJeff_work z0rkito, http://docs.saltstack.com/topics/targeting/compound.html#compound-matchers
19:34 giantlock joined #salt
19:34 z0rkito SEJeff_work: awesome thanks.
19:36 UtahDave SEJeff_work: yeah, I was just about to point out mysql_grants.  :)
19:37 SEJeff_work :)
19:37 SEJeff_work z0rkito, And I meant grains regex, but you get the idea
19:38 UtahDave KennethWilke: that gist is beautiful.  I love it.
19:39 opapo_ joined #salt
19:40 KennethWilke UtahDave: lol, i'm feeling the <3
19:40 KennethWilke it does have docstrings!
19:42 Gifflen Is there any way to reload a service on a git pull from salt?
19:42 KennethWilke Gifflen: absolutely! you add a git 'watch' under your service
19:43 Gifflen Do you happen to have an example of this?
19:43 KennethWilke similar to http://docs.saltstack.com/topics/tutorials/starting_states.html#adding-configs-and-users
19:43 Gifflen that is going to save soooo many headaches!
19:43 KennethWilke ie you'll see - pkg: apache
19:43 KennethWilke instead you'll do - git: <ID for git thing salt manages>
19:43 KennethWilke it'll keep an eye on that git and if git does something it'll bounce the service
19:44 KennethWilke afterwards of course :)
19:44 Gifflen Ok great. We're going to give this a go. Thanks so much
19:44 KennethWilke no problem!
19:45 KennethWilke UtahDave: if somethings in watch, is it inherently required? or should i have it under require as well just in case?
19:46 UtahDave Yes,  watch also acts as a require
19:46 KennethWilke nvm, docs tell me it's the same as require with extra fancyness
19:46 UtahDave yep
19:46 KennethWilke damn you documentation!
19:46 KennethWilke why you has what i need
19:46 david_a joined #salt
19:46 KennethWilke :)
19:47 KennethWilke well there are some lines gonna change in these states of mine
19:47 denstark Is it possible to tell file.append to put the text starting on a certain line?
19:47 denstark (Or is there something else that would accomplish this)
19:47 dthom91 joined #salt
19:48 UtahDave you could use file.sed if you really feel crazy.  :)
19:48 KennethWilke denstark: that is a tricky one, not sure of an easy solution for that myself, i don't think file.append
19:48 KennethWilke 's gonna do it for ya
19:48 UtahDave I generally recommend using  file.managed
19:48 denstark Haha fair enough
19:48 KennethWilke could this file of yours be a jinja template?
19:49 denstark It could be, but that's probably overkill
19:49 denstark I think it will be fine to do it with just file.append for now, but I will keep in mind the alternatives
19:49 KennethWilke alrighty, sounds good
19:50 denstark :) Thanks
19:53 jessep joined #salt
19:58 auser joined #salt
19:59 auser hey all
19:59 auser <~ back
19:59 JasonSwindle WB
19:59 it_dude joined #salt
20:02 jacksontj joined #salt
20:10 UtahDave akoumjian: hey are you around?
20:11 akoumjian UtahDave: yep
20:11 helderco joined #salt
20:11 UtahDave can I PM you?
20:15 snelis joined #salt
20:15 henk123 hello
20:16 henk123 im trying to get my ec2 instance spinning with salt-cloud, but i get an importerror "No module named libcloud.compute.types"
20:17 henk123 maybe you have any suggestions for me :)
20:18 UtahDave henk123: did you have libcloud installed?
20:19 UtahDave henk123: pip install apache-libcloud
20:19 henk123 hm, trying now, did i miss that in some manual ?
20:20 henk123 it worked, tnx  :)
20:21 UtahDave you're welcome!  Yeah, I think docs need help there
20:22 Gifflen @UtahDave quick question for you. We're using /etc/environment for varibales and need to resource them if the file changes. Right now the file is managed. Is there a way to execute a specific state if we detect a change?
20:22 Gifflen source
20:23 UtahDave Gifflen: yep!    salt \*  state.sls  yourstate
20:25 Gifflen we need to do it when we call high state
20:25 Gifflen https://gist.github.com/Gifflen/6125814
20:25 Gifflen as an example
20:25 Gifflen the cmd.run happens on all highstates
20:26 Gifflen but we just want it to execute on detected changes
20:26 UtahDave Ah, I see. just a sec
20:27 Gifflen Sure. Thanks!
20:27 UtahDave Gifflen: have you tried changing the 'require' to a 'watch'?
20:27 Gifflen nope but I guess that would make a lot more sense
20:27 wendall911 joined #salt
20:27 Gifflen was worried about the recursion there
20:28 Gifflen but lets try!
20:28 dave_den and also cmd.wait instead of cmd.run
20:28 debian112 joined #salt
20:29 debian112 Is there are debian 5 salt minon?
20:29 UtahDave yes, cmd.wait
20:29 henk123 lol, it says Failed to create VM snelis. Configuration value 'ipAddress' needs to be set, but it actually created it
20:29 debian112 We have some old debian 5 servers that are running puppet, that we want to move to saltstack
20:30 UtahDave debian112: have you tried installing the minion from debian.saltstack.com?    I'm not sure how far back we support.
20:30 p3rror joined #salt
20:30 terminalmage debian112: we only have pre-compiled packages for wheezy and squeeze
20:30 terminalmage I believe
20:30 debian112 yeah the doc says debian 6 and 7
20:31 terminalmage you could always try installing it with pip
20:31 terminalmage you'll need to make sure that swig and gcc are installed
20:31 dave_den here's a gist for an old saltstack version on lenny: https://gist.github.com/justone/3210047
20:31 jaequery joined #salt
20:32 terminalmage debian also needs python-m2crypto
20:32 terminalmage if installing with pip
20:33 wc- joined #salt
20:34 debian112 ok I will try that, but I am unable to upgrade python on the box because of the old  apps on it still require python 2.5. I will try those options
20:34 KennethWilke howdy guys, i think a colleague was showing me how to setup salt-cloud to run a salt command upon completion of the instance creation
20:34 KennethWilke but i've lost this information :(
20:36 KennethWilke the example i saw before had a state.highstate ran on the node after it was bootstrapped
20:36 UtahDave in your profile or map add   minion:\n  startup_state: highstate
20:37 FreeSpencer joined #salt
20:37 UtahDave basically you can set an config option you want in the minion's config file: http://salt-cloud.readthedocs.org/en/latest/topics/profiles.html
20:37 forrest Does anyone have an example site up for https://github.com/saltstack/halite ?
20:37 UtahDave And you want to set the startup staes
20:37 UtahDave http://docs.saltstack.com/ref/states/startup.html?highlight=startup_state#examples
20:38 KennethWilke UtahDave: awesome, exactly what i'm lookin' for
20:38 KennethWilke ty sir!
20:38 bitz joined #salt
20:38 gwar9999 joined #salt
20:39 akoumjian UtahDave: You guys need a formula for halite :-) Your quickstart involves a lot of steps
20:39 UtahDave KennethWilke: you're welcome!
20:40 UtahDave akoumjian: yes, I agree. not so quick a quickstart
20:40 forrest Yea that's why I was looking for an example before I go through with configuring it.
20:40 forrest I mean it won't take long
20:41 lowpro30 joined #salt
20:43 whiteinge forrest: take a look at the screenshots directory in the repo
20:44 forrest whinteinge, thanks.
20:44 forrest whiteinge, are you using twitter bootstrap?
20:45 whiteinge yes. plus the angular-bootstrap bindings
20:45 forrest Nice, that's good stuff.
20:47 forrest Did you pick cherrypy specifically because it has Python3 support?
20:47 forrest or just because salt-api uses it
20:48 SEJeff_work whiteinge, Can you smack the guy who made camelcase filenames? This is python, not java
20:49 SEJeff_work but that looks super nice3
20:50 forrest SEJeff_work, as long as it's consistant I don't mind the camel case. If it was a mix, then it would be terrible
20:50 forrest yea halite is looking awesome
20:50 whiteinge forrest: cherrypy the framework was chosen for the API because it has good REST support. however cherrypy the server is optional; it's a regular wsgi app that can be used with any wsgi-capable server
20:50 forrest would be great to enable developers to do some management of their own boxes.
20:50 SEJeff_work I was just looking here: https://github.com/saltstack/halite/tree/master/screenshots
20:50 forrest whinteinge, ahh ok cool
20:56 forrest SEJeff_work, you could always change it
20:57 SEJeff_work I'd like an excuse to learn angular actually. I've already got over 400 commits in salt proper
20:57 SEJeff_work Perhaps I should work on a different part
20:58 JasonSwindle SEJeff_work:  How about Salt-Cloud?
20:58 JasonSwindle The docs need LOVE BADLY
20:58 Newt[cz]1 joined #salt
20:58 SEJeff_work JasonSwindle, I have never used ec2 ever and don't have a lot of interest in clouds to be honest
20:58 SEJeff_work I have thousands of real servers
20:58 SEJeff_work no cloud stuff unfortunately
20:58 JasonSwindle Pepper?
20:59 SEJeff_work pepper was sort of my idea
20:59 SEJeff_work yeah that does need some work
20:59 forrest Did anyone grab a ton of those amazon cards from pycon? They were worth 100 bucks, and didn't even buy two VMs (with minimal load) for two months
20:59 SEJeff_work The api stuff seems to be maturing nicely as whiteinge has done a great job on that
20:59 JesseC The base location.. ex. /srv/salt would be called the "salt tree" right?
20:59 FreeSpencer joined #salt
20:59 SEJeff_work JesseC, Generally you'd call that the salt state tree, but that works
21:00 JesseC SEJeff_work: gotcha, thanks
21:00 JasonSwindle Salt-Thin?
21:01 JasonSwindle How about the CICD part of Salt?
21:01 JasonSwindle SEJeff_work:  https://github.com/saltstack/salt-ci
21:02 SEJeff_work I'm not really sure what that is for to be honest
21:02 SEJeff_work jenkins is already awesome, I use the snot out of it
21:03 Xeago joined #salt
21:04 JasonSwindle https://github.com/saltstack/halite then? :)
21:04 SEJeff_work yes!
21:04 jeddi joined #salt
21:05 jessep joined #salt
21:05 JasonSwindle What is needed is the require / require_in dep graphics
21:06 JasonSwindle RIght now, doing that by hand sucks
21:07 SEJeff_work whiteinge, Do you take halite feature requests?
21:07 whiteinge yup!
21:07 SEJeff_work whiteinge, Because this would be absolutely amazing to show dependencies with this javascript widget! http://mbostock.github.io/d3/talk/20111018/tree.html
21:07 nliadm so, we have ~550 servers
21:07 SEJeff_work which parses the high/lowstate and generates a pretty d3 powered graph
21:08 nliadm highstates are glacially slow
21:08 SEJeff_work whiteinge, click the datapoints on that url
21:08 SEJeff_work nliadm, Could it be that some of your hosts are down? Or that your state is doing something that is taking a long time?
21:08 whiteinge SEJeff_work: i agree! we do have plans to visualize the dependency hierarchy and we also have in-house experience with d3
21:08 nliadm we have the number of workers on the master bumped up
21:08 UtahDave nliadm: how many workers?
21:09 nliadm SEJeff_work: no, if we run salt-call --local, it's quick
21:09 JasonSwindle SEJeff_work:  How would this work for require_in?
21:09 nliadm it's at 15 workers
21:09 whiteinge SEJeff_work: feel free to file a ticket on halite. i'll categorize it and link to the existing salt ticket to expose that hierarchy via the api
21:09 JasonSwindle I see it would work great for require. :)
21:10 SEJeff_work whiteinge, done!
21:10 whiteinge ty
21:11 henk123 salt is SWEET
21:11 nliadm the pillar makes like, 4 dns queries and an indexed ldap query, it's not very intensive
21:11 henk123 painless setup amazon ec2 instance :D
21:11 JasonSwindle How are you guys going to keep the server_node.js going?
21:12 JasonSwindle Forever?
21:12 JasonSwindle Foreman?
21:14 dthom91 joined #salt
21:15 henk123 can i run the salt-master as non-root ?
21:16 UtahDave henk123: yes.
21:16 henk123 ok, going to fix some permissions then
21:19 z0rkito anyone know why having a state that calls salt-call to run another state would cause the salt-call to run 4 times in a row and return 0 data back to the master about other states that were also  run?
21:20 z0rkito this is when running a highstate against a node.
21:20 jacksontj joined #salt
21:22 timops joined #salt
21:25 UtahDave I'm not sure, z0rkito.  Can you pastebin what you have?
21:30 dave_den is there a way to define a function in a state that will not run unless one or more functions add themselves in a watch_in ?  for example, I want to define restart_nginx that will not fire unless some other state adds a file.managed nginx vhost that adds itself for restart_nginx from watch_in
21:30 dave_den all in a highstate
21:31 dthom91 joined #salt
21:33 UtahDave dave_den: I think you might be making this more complicated than it needs to be.
21:33 UtahDave dave_den: first set up a service.running to make sure nginx is running
21:34 UtahDave dave_den: then you can have any number of   file.managed  or file.recurse  or anything else you want that can do a watch_in  on your service.running for nginx
21:34 UtahDave Then any one of those can cause nginx to restart
21:34 dave_den derp. totally.
21:35 nliadm UtahDave: any idea why salt would be incredibly slow with ~550 minions?
21:35 nliadm bumped workers to 15
21:36 UtahDave nliadm: how much ram on your master?
21:36 nliadm 16GB
21:36 nliadm 8 cores
21:36 nliadm it's not pegged
21:37 UtahDave ok, try bumping workers to 30 and restarting the salt-master service
21:37 UtahDave What does slow mean?  what command are you running and how long is it taking?
21:38 squelch joined #salt
21:39 squelch hey, can you extend pillars? i.e. in i have a common pillar that sets nginx: settings and i want to extend nginx: for sites on dev box
21:39 nliadm a highstate. sometimes it claims the the minion key isn't accepted, other times it'll run the highstate, just take several minutes
21:40 nliadm if you run salt on the master, it'll throw a zmq timeout error waiting for responses
21:40 mlister2006 salt master may be stressed out. I do find same issue with large number of minions
21:41 mlister2006 often few minons (mostly slow network access over Internet seems to have issue)
21:42 mlister2006 running 30 master workers hep a bit , but not lot. still see minion did not return every time running highstate.
21:42 nliadm "stressed out" isn't a widely accepted reason for failure
21:43 mlister2006 our workaround for now: restart salt master time to time. restart minion time to time
21:43 mlister2006 not a good solution, I agree.
21:43 JasonSwindle nliadm:  can you do a salt —versions-report
21:43 mlister2006 other thing we noticed. that problematic minons (those reporting issue) may have some zombie salt-minion process running (check if that is the case).
21:44 nliadm http://pastebin.com/jPbFWdTU
21:44 mlister2006 on cenos we need to manually kill the, (/etc/int.d/salt-minion restart won't get them).
21:44 mlister2006 we put a cronjob, that will hunt these zombie minions and kill manually once an hour
21:44 JasonSwindle It is not ZMQ 2 issue, hmm
21:47 nliadm upping the workers doesn't seem to have helped appreciably
21:48 jessep joined #salt
21:48 mlister2006 also following sysctl params may help.
21:48 mlister2006 net.ipv4.tcp_syncookies = 0
21:48 mlister2006 net.ipv4.tcp_max_syn_backlog = 8192
21:49 rsimpkins joined #salt
21:49 rsimpkins Why does salt-master and salt-minion need so much RAM? Is there a way to make it use less RAM?
21:49 JasonSwindle rsimpkins:  how much ram?
21:49 mlister2006 salt master gets tons of connections back from minions (several thousands possibly, when pushing lots of states).
21:50 rsimpkins JasonSwindle: 36M resident, 1G virtual.
21:50 mlister2006 every minon downloading a state file, opens a new connection (as I have noticed) and try to download these files every time. (unless cache=True is used)
21:50 rsimpkins Looks like per connection.
21:53 Xeago joined #salt
21:53 rsimpkins do you mean cache_jobs?
21:59 druonysuse joined #salt
22:03 mlister2006 rsimpkins: cache=True would used cached state files on minions (and won't try to download from the master, unless it is not present at all on minions)
22:03 mlister2006 if this what you were asking.
22:04 jacksontj joined #salt
22:05 UtahDave rsimpkins: The master does a lot of work if it's managing a lot of minions.  The minion shouldn't require too much ram.  It works on a Raspberry Pi
22:05 mlister2006 I think salt state file download need a major facelift. It should only download state files if cached copy if different (md5sum match or timestamp check like rsync does)
22:05 xet7 joined #salt
22:05 mlister2006 it is majot performace penality for large salt installs.
22:06 UtahDave mlister2006: I thought it did do a md5sum comparison before downloading.  Does it not?
22:06 mlister2006 for example it takes 42seconds for a state execution with cache=True vs 10+ minutes
22:07 mlister2006 UtahDave: I don't see that in our environ. (0.16.0 master/minions)
22:08 mlister2006 when I run salt-call state.highstate
22:09 jessep joined #salt
22:09 mlister2006 it says : Fetchng File: salt://....
22:09 mlister2006 and spend most of the time there.
22:10 UtahDave mlister2006: Hm. OK. let me do some code diving
22:10 mlister2006 not sure if it is really fetching file or spending time in md5sum.
22:10 mlister2006 UtahDave: that would be great. if it can speed up.
22:10 rsimpkins UtahDave: I'm managing 8 hosts with it. It isn't a ton. The minions use a lot of RAM as well.
22:11 rsimpkins I don't hink I have cache turned on.
22:11 UtahDave rsimpkins: what version of Salt? which os?
22:11 rsimpkins UtahDave: 0.15.3, RHEL 6.4
22:12 squelch mysql_grants.present: database * complains
22:14 nliadm mlister2006: cache=True? is that on the command line?
22:14 rsimpkins UtahDave: For example, on a VM with 4GB RAM the salt-minion is using 609MB virtual RAM. That seems a lot to me, but maybe I just don't understand how it works.
22:14 forrest rsimpkins, is that while applying configuration changes?
22:15 rsimpkins forrest: No. That is doing nothing.
22:15 rsimpkins It is growing as well.
22:16 UtahDave rsimpkins: can you pastebin that?
22:17 forrest ok, I've got the same installation (0.15.3, RHEL 6.4) on a local VM, the minion is using 3.2% of my total allocated ram for this VM (1024 mb)
22:17 rsimpkins UtahDave: Pastbin it? Look at this: https://metrics.librato.com/share/dashboards/s959ugas
22:17 forrest is that for a single salt-minion process?
22:17 rsimpkins UtahDave: I did reboots just a little while ago. That is showing ram-cache growth on my hosts. Almost all of that is salt.
22:17 rsimpkins ram utilization percent minus cache
22:18 oz_akan_ joined #salt
22:18 rsimpkins Highstate calls will also show up on those graphs. None are listed there.
22:18 rsimpkins host1 is the master.
22:20 rsimpkins UtahDave: I think this is contributing to very high CPU utilization by ksm on the virt host as it tries to smash all those allocated pages down.
22:21 UtahDave Yeah, that does seem very excessive.  Do you have any minions attempting to connect that haven't been accepted yet?
22:21 druonysuse joined #salt
22:21 druonysuse joined #salt
22:21 rsimpkins UtahDave: No.
22:22 rsimpkins UtahDave: You can see it continue to climb.
22:23 JesseC If I have salt set to local instead of remote. Then salt://example/ex.sls would point to /srv/salt/example/ex.sls if file_roots=/srv/salt right?
22:24 UtahDave JesseC: yep!
22:24 JesseC Great. :)
22:24 UtahDave rsimpkins: Are we looking at Salt master there?
22:25 JesseC Is there a page that details the directory structure of the salt state tree?
22:26 dthom91 joined #salt
22:26 JesseC I'm going through the guide, but it's just pieces here and there.
22:26 mlister2006 nliadm:   salt '*' state.highstate cache=True
22:26 UtahDave JesseC: what do you mean by the salt state tree?
22:26 mlister2006 yes on command line
22:27 JesseC UtahDave: Something like an example directory structure that you would see in /srv/salt
22:27 nliadm mlister2006: does that apply to salt-call?
22:27 Darnoth joined #salt
22:28 mlister2006 yes and you can see with or without and execution time accordingly.
22:28 UtahDave JesseC: Ah, ok. Well, you can just drop your .sls files in there.  You can also create subdirectories.    /srv/salt/mysql
22:28 UtahDave mysql/init.sls   would be called   salt \* state.sls mysql
22:28 mlister2006 it also works for state.sls (  e.g: salt-call state.sls mystate)
22:28 UtahDave and mysql/client.sls   would be called by   salt \* state.sls mysql.client
22:29 rsimpkins UtahDave: That is all hosts under salt control. Roll over the graph. host1 is the master. all others are only running salt-minion.
22:30 UtahDave rsimpkins: can you stop the salt-master service  and start it in the foreground in debug mode?
22:30 Linz joined #salt
22:30 UtahDave sudo salt-master -l debug
22:31 rsimpkins Doing now.
22:31 UtahDave thanks
22:31 sinh joined #salt
22:31 rsimpkins UtahDave: What do you want me to look for?
22:32 UtahDave Just any tracebacks or any recurring connection rejections.
22:36 z0rkito UtahDave: sorry took a while to grab this paste, had to run to a meeting: http://pastebin.com/KZFNcf6p  I created a simple test that mimics what we're trying to do and put all the state files as well as the minion output.
22:36 sinh so just a curious question about exit status'. . . i'm running salt-call --local state.highstate and it has errors rendering a pillar because the required files/bins that pillar is trying to run doesn't exist. . however it exits with a 0 instead of a non 0
22:38 UtahDave z0rkito: you don't need the run()  in test2.sls
22:39 jacksontj joined #salt
22:40 UtahDave sinh: There's been some discussion around the exit code.  Right now it I believe it returns success if salt was able to run correctly. So the exit code doesn't refer to the command that was executed by Salt itself
22:40 nliadm mlister2006: I dunno man, it too .1 sec longer to claim my key was unaccepted with 'cache=true'
22:40 z0rkito UtahDave: you think that is what's causing it to run the cmd 4x in a row?
22:40 UtahDave But your example does seem like that should return a non zero status code
22:40 UtahDave z0rkito: Yeah, that's what I'm guessing.
22:41 jkleckner joined #salt
22:42 mlister2006 nliadm: I don't think cache=True is going to help w/ key acceptance issue.
22:42 nliadm heh
22:42 sinh UtahDave: hrm. . interesting. . i was able to get a non-zero by specifying a cmd that doesn't exist like. . grains.lss
22:42 mlister2006 I was just throwing lot of workarounds for slower salt on large installs :)
22:42 sinh is there a way to capture when salt runs into an error?
22:45 rsimpkins UtahDave: I've watched the outupt for a while here. Pretty dead. just auth requests that are already accepted.
22:45 rsimpkins UtahDave: And only a couple.
22:45 UtahDave yeah, that's really odd.  can you pastebin the output of   salt --versions-report    from the master?
22:46 rsimpkins UtahDave: http://pastebin.mozilla.org/2754173
22:48 __Jahkeup__ joined #salt
22:49 brianhicks joined #salt
22:49 UtahDave rsimpkins: OK, my first guess is that it might be zmq related.
22:50 UtahDave rsimpkins: Oh, wait.  No, you already have zmq 3.2.3
22:50 UtahDave that's not it.
22:50 UtahDave rsimpkins: is there any way you can test against the latest from the develop branch?
22:58 rsimpkins UtahDave: If there is an RPM for it, ues.
22:58 rsimpkins yes. :)
22:59 UtahDave rsimpkins: have you looked in epel-testing?
22:59 rsimpkins UtahDave: No, I can enable that repo and install the one there, if you'd like, on a test host.
23:01 Linz joined #salt
23:02 jwon joined #salt
23:02 UtahDave sure, that would be great.
23:03 kstaken joined #salt
23:03 UtahDave I'm leaving for the day soon, rsimpkins.  Will you be around tomorrow to continue?
23:03 chrisgilmerproj joined #salt
23:03 rsimpkins UtahDave: Likely, yes. I'll work on getting a test host updated and let you know the results tomorrow.
23:04 rsimpkins UtahDave: 0.16.0-1.el6 looks like what is in epel-testing. Is that the 'latest?'
23:04 chrisgilmerproj left #salt
23:05 UtahDave rsimpkins: Yeah, pretty much.  Any way you could clone the salt repo and   install  the latest from develop?     python setup.py install --force
23:06 tomeff joined #salt
23:06 wc- joined #salt
23:07 rsimpkins UtahDave: That would take some setup. I'd want to do that in a throw-away VM.
23:08 rsimpkins UtahDave: Short answer, yes. When I have more time I can try it and let you know.
23:08 rsimpkins UtahDave: Immediate results for 16 do not look promising. Still testing.
23:11 UtahDave ok, thanks, rsimpkins
23:13 squelch left #salt
23:18 oz_akan_ joined #salt
23:24 druonysuse joined #salt
23:24 druonysuse joined #salt
23:24 napperjabber joined #salt
23:27 auser joined #salt
23:33 FreeSpencer joined #salt
23:33 brianhicks joined #salt
23:33 oz_akan_ joined #salt
23:34 cxz joined #salt
23:44 FreeSpencer joined #salt
23:46 juanlittledevil joined #salt
23:48 jkleckner joined #salt
23:55 intchanter joined #salt

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