Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-03-12

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

All times shown according to UTC.

Time Nick Message
00:03 fllr joined #salt
00:04 meteorfox joined #salt
00:06 forrest joined #salt
00:15 manicouman joined #salt
00:17 allanparsons if i run status.meminfo
00:17 allanparsons how do i get at a particular key
00:18 allanparsons i.e.: salt '*006*' status.meminfo.MemFree
00:23 arthabaska joined #salt
00:28 fwiles left #salt
00:28 manicouman joined #salt
00:32 rachbelaid joined #salt
00:38 jeffro_ joined #salt
00:41 dave_den joined #salt
00:45 meteorfo_ joined #salt
00:47 manicouman joined #salt
00:50 jcristau joined #salt
00:50 bhosmer joined #salt
00:53 manicouman joined #salt
00:54 bwq joined #salt
00:54 svs_ joined #salt
00:54 allanparsons screw it, i just ended up using grans
00:54 allanparsons grains*
00:54 allanparsons {{ (grains.mem_total|int * 2 / 5)|round|int }}
00:55 kaiyou joined #salt
00:55 msciciel_ joined #salt
00:55 majoh joined #salt
00:56 jab416171_ joined #salt
00:56 svx joined #salt
00:56 mugsie_ joined #salt
00:57 jnials_laptop joined #salt
00:57 gnugnu9 joined #salt
00:57 balltongu_ joined #salt
00:58 goki_ joined #salt
00:58 fejjerai joined #salt
00:58 pnl_ joined #salt
00:58 sulky_ joined #salt
00:58 herzi__ joined #salt
00:58 rogst joined #salt
00:58 dopp joined #salt
00:58 jmccree_ joined #salt
00:58 viq joined #salt
00:58 JordanRinke joined #salt
00:58 rjc joined #salt
00:58 vxitch joined #salt
00:58 sashka_ua joined #salt
00:58 runnyspot joined #salt
00:58 runnyspot joined #salt
00:59 bejer joined #salt
00:59 lyddonb joined #salt
00:59 stewpot joined #salt
00:59 hotbox joined #salt
00:59 dcmorton joined #salt
00:59 steveoliver joined #salt
00:59 cewood joined #salt
00:59 godog joined #salt
00:59 godog joined #salt
01:00 modafinil_ joined #salt
01:00 carmony joined #salt
01:00 shadowsun joined #salt
01:00 vegardx joined #salt
01:00 gamingrobot joined #salt
01:00 rcsheets joined #salt
01:00 munhitsu_ joined #salt
01:01 kinetic joined #salt
01:01 cornmand1r joined #salt
01:01 abele joined #salt
01:02 crazysim joined #salt
01:02 bezaban_ joined #salt
01:02 chuffpdx_ joined #salt
01:03 TheSojourner joined #salt
01:03 TheSojourner joined #salt
01:04 Kyle____ joined #salt
01:04 Sypher|NL joined #salt
01:04 dpac|away joined #salt
01:04 SaveTheR0btz joined #salt
01:04 kevinbrolly joined #salt
01:04 troyready joined #salt
01:04 terminalmage_ joined #salt
01:04 kaictl joined #salt
01:04 Nazzy joined #salt
01:04 Nazzy joined #salt
01:04 druonysus joined #salt
01:04 druonysus joined #salt
01:04 nicksloan joined #salt
01:04 cyrusdav- joined #salt
01:04 gldnspud__ joined #salt
01:04 tzero joined #salt
01:04 babilen joined #salt
01:04 scarcry joined #salt
01:05 emostar joined #salt
01:05 Guest41155 joined #salt
01:05 mikkn joined #salt
01:05 magenbrot joined #salt
01:05 copelco______ joined #salt
01:05 delkins joined #salt
01:05 mfournier joined #salt
01:06 alienresidents joined #salt
01:06 jbub joined #salt
01:06 rhand joined #salt
01:19 razor joined #salt
01:19 vimalloc joined #salt
01:19 zain_ joined #salt
01:19 kzx joined #salt
01:19 MTecknology joined #salt
01:19 jforest joined #salt
01:19 gparent joined #salt
01:21 [diecast] joined #salt
01:23 oz_akan_ joined #salt
01:23 copelco______ joined #salt
01:24 gparent joined #salt
01:24 pcarrier joined #salt
01:25 rcsheets joined #salt
01:28 xzarth joined #salt
01:28 alienresidents joined #salt
01:29 MTecknology joined #salt
01:29 meteorfox joined #salt
01:31 xt joined #salt
01:37 zain__ joined #salt
01:37 kloplop321 joined #salt
01:41 Linz_ joined #salt
01:43 borgstrom joined #salt
01:43 vxitch joined #salt
01:44 dangra joined #salt
01:46 _bt joined #salt
01:46 allanparsons what the heck does : "Comment: Group {'maxmemory-policy': 'noeviction', ..... 'port': 6379} is not available"  mean?
01:46 malinoff joined #salt
01:49 Yaann joined #salt
01:52 meteorfox joined #salt
01:55 kickerdog1 do you have the state that generated that?
01:57 allanparsons yes
01:57 allanparsons 1 sec
01:58 allanparsons http://pastebin.com/cf6ZNSiy
01:59 Corey allanparsons: Where do we go from here?
01:59 allanparsons i've never heard that one before :)
01:59 allanparsons also, here's my pillar: http://pastebin.com/hxnKHQ3v
02:00 kickerdog1 allanparsons: looks like salt is trying to start redis when its already running
02:01 allanparsons root@dev-myapp-006:/src-salt# ps -ef|grep -i redis root      2764 26114  0 02:00 pts/2    00:00:00 grep --color=auto -i redis
02:01 kickerdog1 try adding - reload: True
02:01 kickerdog1 does this command execute ok?
02:01 kickerdog1 update-rc.d redis-server remove
02:02 allanparsons yeah
02:02 allanparsons here's the output: http://pastebin.com/UpjcmuxX
02:04 kickerdog1 wow, that is screwy
02:04 kickerdog1 Might have a legit bug on your hands
02:04 allanparsons whats interesting is that it doesnt create /etc/redis/...
02:04 kickerdog1 why would the file.directory state output a comment?
02:06 allanparsons beats me.
02:07 xl1 joined #salt
02:14 jeremyfelt joined #salt
02:18 Sagar joined #salt
02:19 genggeng joined #salt
02:20 Guest99916 Hello group. I see a lot of these messages in my salt master log - "Received incomplete call from VMRHQFT4 for '_mine', missing 'tok'"
02:20 genggeng left #salt
02:21 Guest99916 and very slow performance. does anyone know what could be happening?
02:22 allanparsons also, why do these always come back as error:  [ERROR   ] Command "dpkg --compare-versions '6.03-1' lt '6.03-1'" failed with return code: 1
02:23 allanparsons Output:  [INFO    ] Executing command "dpkg --compare-versions '6.03-1' eq '6.03-1'" in directory '/root'
02:23 rostam joined #salt
02:28 allanparsons argh, i guess i'll log a bug report for that one @kickerdog1
02:30 malinoff allanparsons, looks like salt does not properly handle this comparison
02:30 malinoff Actually, this should be done internally, and should not log any messages like this
02:30 allanparsons ok
02:30 allanparsons i'll log 2 bugs
02:30 allanparsons one for the pillar above and one for the comparison
02:31 malinoff https://github.com/saltstack/salt/blob/develop/salt/modules/aptpkg.py#L765
02:31 malinoff This is the comparison function
02:32 malinoff Somebody added that crap check on 781 line
02:35 jeffro joined #salt
02:37 HeadAIX joined #salt
02:39 bhosmer joined #salt
02:40 andrej joined #salt
02:45 sagar777 ok nevermind - I think i got my answer. i have used 100% of i-nodes available in my /var filesystem in salt master
02:45 Linz joined #salt
02:49 allanparsons malinoff:  https://github.com/saltstack/salt/issues/11154
02:49 allanparsons and malinoff kickerdog1 https://github.com/saltstack/salt/issues/11153
02:57 sinh joined #salt
02:57 repl1cant joined #salt
02:59 higgs001 joined #salt
03:01 n0c joined #salt
03:02 Linz joined #salt
03:02 thayne joined #salt
03:03 Linz joined #salt
03:03 arodizzy joined #salt
03:15 rallytime joined #salt
03:15 kickerdog1 +1
03:19 ipalreadytaken joined #salt
03:20 JaredR joined #salt
03:21 bhosmer joined #salt
03:23 thayne joined #salt
03:23 googolhash joined #salt
03:33 ckao joined #salt
03:36 CeBe1 joined #salt
03:37 n8n joined #salt
03:37 favadi joined #salt
03:38 yomilk joined #salt
03:39 mgw joined #salt
03:48 markizano joined #salt
03:50 Yaann joined #salt
03:51 fllr joined #salt
03:53 Ryan_Lane joined #salt
03:59 mgw joined #salt
04:01 jeremyBass joined #salt
04:02 jeremyBass left #salt
04:03 dstanek joined #salt
04:05 opapo joined #salt
04:16 arodizzy joined #salt
04:18 oz_akan_ joined #salt
04:22 arodizzy joined #salt
04:27 bhosmer joined #salt
04:31 scurry joined #salt
04:36 n8n joined #salt
04:41 KyleG joined #salt
04:41 KyleG joined #salt
04:54 schimmy joined #salt
04:59 ravibhure joined #salt
05:02 manicouman joined #salt
05:12 mgw1 joined #salt
05:17 Mike_Utah joined #salt
05:18 scarcry joined #salt
05:18 Mike_Utah Is this SaltCon like Salt Lake City
05:18 sroegner joined #salt
05:20 schimmy joined #salt
05:20 kickerdog Anyone see the giant fire in SF Soma tonight?
05:21 Mike_Utah Didnt see anything about that.
05:21 Mike_Utah Taking a trip there beginning of August
05:22 Eugene Nope, and I'm even in the Bay area.
05:22 Eugene Glad it wasn't yesterday, as I was drinking there.
05:23 manfred Mike_Utah: SaltConf?
05:23 Mike_Utah quick search says only a small fire..
05:25 Mike_Utah @manfred I dont know what #salt is.. The purple text at the top said something about SaltConf
05:25 manfred Mike_Utah: saltconf was in january.
05:25 Mike_Utah In Utah we have a board game conference called SaltCon
05:26 manfred ahh, this isn't that
05:26 Mike_Utah Its Later this month
05:29 kickerdog at the next meeting can you add a wait method to the pip state?
05:30 manfred what would it be waiting for?
05:32 packeteer joined #salt
05:32 kickerdog right now, pip.installed executes a pip install every time is runs. I'm thinking of something like cmd.wait so like pip.wait for some pkg to be updated before running wild.
05:35 manfred so, pip.installed is a state, and should check that the package is installed instead of running a pip.install?
05:37 manfred if it is already installed, and you use the pip.installed state, it should be returning saying the package is installed and never runs pip.install
05:37 manfred https://github.com/saltstack/salt/blob/develop/salt/states/pip_state.py#L343
05:38 Mike_Utah left #salt
05:39 kickerdog hmmm, thats not what i'
05:39 kickerdog m seeing
05:39 manfred which version of salt?
05:41 kickerdog salt 2014.1.0
05:41 kickerdog let me put together an example
05:42 manfred yeah, so there is no need for wait, but if it is installing things before, that would be something that needs to be fixed.
05:46 thayne joined #salt
05:51 Yaann joined #salt
06:00 andr386_ joined #salt
06:00 bhosmer joined #salt
06:01 Yaann joined #salt
06:11 thayne joined #salt
06:14 kermit joined #salt
06:16 n8n joined #salt
06:24 ChaosPsyke joined #salt
06:29 higgs001 joined #salt
06:35 ndrei joined #salt
06:36 fllr joined #salt
06:38 _fllr_ joined #salt
06:51 joehh joined #salt
07:07 sroegner joined #salt
07:09 arodizzy joined #salt
07:12 Shenril joined #salt
07:21 roolo joined #salt
07:23 Yaann joined #salt
07:27 linjan_ joined #salt
07:39 archeryue joined #salt
07:40 Iwirada joined #salt
07:41 archeryue left #salt
07:42 archeryue joined #salt
07:43 kadel joined #salt
07:48 ml_1 joined #salt
07:49 bhosmer joined #salt
07:51 ndrei joined #salt
07:59 Kenzor joined #salt
07:59 thayne joined #salt
08:02 carlos joined #salt
08:05 nn0101 joined #salt
08:07 fllr joined #salt
08:13 balboah joined #salt
08:15 harobed joined #salt
08:22 slav0nic joined #salt
08:36 mikeywaites joined #salt
08:36 BbT0n joined #salt
08:37 fllr joined #salt
08:38 yomilk joined #salt
08:47 ndrei joined #salt
08:49 yidhra_ joined #salt
08:49 mikeywaites joined #salt
08:51 gasbakid joined #salt
08:52 MohShami joined #salt
08:53 yes456 joined #salt
08:53 MohShami hey guys, I'm building an ftp backup system with salt and the strip get the list of folders to be backed up through pillar, the problem is, if I add a new folder the script won't back it up till I restart salt-minion, even running refresh_pillar didn't help, any ideas
08:56 sroegner joined #salt
08:59 aleszoulek joined #salt
09:00 thayne joined #salt
09:03 ndrei joined #salt
09:04 kickerdog joined #salt
09:05 yidhra_ joined #salt
09:10 arodizzy joined #salt
09:10 babilen What are sensible ways to deal with configuration files that are appropriate for different versions of a package in a formula?
09:12 babilen I am currently working on the nagios formula and will have to grab the "default" configuration file from some tarball they released. Not sure how to choose this sensibly as that version might differ on different distributions? Should I include multiple versions and simply include the appropriate one based on grains['os'
09:12 babilen ?
09:13 topochan joined #salt
09:13 aleszoulek joined #salt
09:14 babilen Well, it doesn't really matter as everyone will have 2.13, but I just wondered about this in general :)
09:18 Guest27346 joined #salt
09:21 Nexpro joined #salt
09:22 adyromantika joined #salt
09:23 torrancew joined #salt
09:24 Yaann joined #salt
09:25 yomilk joined #salt
09:26 kickerdog babilen: for different distributions I usually use pillars if i'm dealing with a config file
09:27 babilen kickerdog: How do you do that exactly? I mean I am using pillars to supply different data to parameterise a given template, but I have to base that template on some configuration file version.
09:29 kickerdog dev:
09:29 kickerdog 'os:Debian':
09:29 kickerdog - match: grain
09:29 kickerdog - servers
09:29 kickerdog In your pillar you could serve up different pillars for say RedHat vs Debian
09:31 babilen sure, but the basic/default configuration file has to come from some specific version (unless I include one for each applicable version and include that based on grains.os)
09:32 kickerdog You could do that, but how different are that nagios files for distro to distro? Could you come up with a single config that would work on both?
09:33 babilen Sure - it doesn't really matter in this particular case, but it simply caused me to think about this issue and thought I'd ask for advice before I do something completely bonkers ;)
09:37 BogdanR Hello, I am trying to deploy some ssh-keys for a user which is not root and it doesn't seem to work. It tells me that "Failed to add the ssh key. Is the home directory available, and/or does the key file exist?"
09:37 BogdanR For the root user this works just fine.
09:37 fllr joined #salt
09:37 bhosmer joined #salt
09:38 kickerdog what dir are you putting the users key in?
09:38 BogdanR I have this state: http://hastebin.com/raw/punanobihi
09:38 BogdanR I suppose the dir should be the home directoy of the user
09:38 BogdanR I am not specifying it
09:39 kickerdog I think those keys are going into the root users ssh folder.
09:39 BogdanR How come?
09:39 kickerdog http://docs.saltstack.com/ref/states/all/salt.states.ssh_auth.html
09:40 BogdanR kickerdog: Yes, I looked at this and this is how I actually did the config for root
09:40 yidhra_ joined #salt
09:40 kickerdog so which user can't access the key?
09:41 mikeywaites left #salt
09:41 BogdanR www-data
09:42 kickerdog does www-data have a /home/www-data ?
09:42 BogdanR Basically the bogdan-dev and alex-dev IDs are failing
09:42 BogdanR No, the home is /var/www
09:43 BogdanR Maybe I should also specify the home?
09:43 lothiraldan joined #salt
09:43 kickerdog i'm not sure thats an option, let me check
09:43 lothiraldan Hello
09:44 kickerdog I would manually place the key using file.managed, doesn't look like ssh_auth is placing it correctly
09:48 BogdanR I see. I will try to report this bug
09:53 millz0r joined #salt
09:53 davidone good morning
09:53 sroegner joined #salt
09:53 davidone I'm playing with pillars
09:53 davidone and I've created a simple pillar to install some software on a debian box
09:53 davidone http://hastebin.com/wiwijuqovi.django
09:54 davidone is it the 'right way' to use pillars?
09:56 ndrei joined #salt
09:58 diegows joined #salt
09:58 pvial joined #salt
09:59 pvial hi, has anyone seen this with latest salt-cloud and salt?  [WARNING ] Private IPs returned, but not public... Checking for misidentified IPs  (message repeats for 10 minutes and then deletes the VM)
10:04 Yaann joined #salt
10:07 johtso joined #salt
10:08 MohShami can I specify a random time for a cron.present state?
10:08 SkyRocknRoll__ joined #salt
10:09 MohShami I mean if for example I want it to run randomly between 12 and 3AM
10:09 tonthon Hi, is there some package to install in order to make postgres_user states working ?
10:12 backjlack joined #salt
10:13 tmandry joined #salt
10:14 backjlack left #salt
10:14 pvial MohShami, you could always script a random cron.  http://stackoverflow.com/questions/9049460/cron-jobs-and-random-times-within-giving-hours
10:15 MohShami pvial, will check it out, thanks mate :)
10:17 MohShami pvial, that's not exactly what I'd want to do, but the random option in cron.present helps, can I for example use random for the minute and a grain for the hour?
10:17 MohShami it should be possible
10:17 MohShami trying to figure it out
10:19 pvial MohShami, i haven't worked much with grains.  I wouldn't be the best person to answer your question.  My problems with Salt are always getting it to actually work and do basic functions, sadly.
10:20 MohShami pvial, I have it storing the states in git, and migrating all my servers to salt, I love it
10:21 pvial MohShami, I love it too, when it works.  I'm on day #3 trying to get Salt Cloud to work.
10:21 MohShami pvial, still haven't tried it yet, but it'll be my next project
10:23 pvial MohShami, if I were you I'd choose a known working Cloud Provider to use with Salt Cloud.  Salt Cloud doesn't play well with the generic OpenStack provider.  I've tried 4 different versions of Salt and Salt Cloud and I still can't create a VM.
10:24 pvial and i can create a VM using `nova`
10:24 MohShami pvial, we're planing on building a local public cloud, thinking of going with apache cloudstack
10:26 pvial MohShami, you should plan a few days or a week to integrate Salt Cloud with Apache Cloudstack.  Simply following the Salt Cloud setup in the docs will not guarantee that it will work properly.
10:27 MohShami pvial, when did following the docs ever worked? :D
10:27 MohShami for any product that is
10:27 Yaann left #salt
10:27 MohShami ever work*
10:27 pvial MohShami, I believe that most setup instructions these days are very well written and that most work properly, actually.
10:28 MohShami pvial, I used to work as a swift technical consultant, left that job a few months ago, their documentation is very detailed, but something has to go wrong
10:28 pvial I wonder if there's anyone in the room that's worked with a Openstack environment that only has private IPs.
10:29 MohShami my cloud setup would also have private IPs only
10:29 pvial MohShami, Murphy's law always seems to come into play. ;)
10:29 MohShami of course :D
10:31 manicouman joined #salt
10:33 bhosmer joined #salt
10:36 pvial MohShami, I'm curious, how many times have you upgraded Salt?  It seems every time I encounter a problem, I am told to upgrade and that only further complicates matters, like in my present situation.
10:37 fllr joined #salt
10:38 MohShami pvial, I've upgraded like 3 or 4 times
10:39 pvial Did some features break after upgrading?
10:39 MohShami started off with .16.something, did give me a problem when I upgraded to .17, but not it works well
10:39 MohShami after .17.0, I'm happy to say no so far
10:40 pvial that's good.  I lost all Salt-Cloud functionality after upgrading to the latest of both.
10:41 magosinvello joined #salt
10:42 MohShami sometimes a feature change or a configuration file format change can do that
10:42 MohShami I use freebsd, pkgng changed the format 3 times within 2 weeks after I started using it
10:43 pvial I've been using Salt before SaltStack came about and it's always been this buggy, sadly.  My infrastructure was the first testing ground for the founders of SaltStack.
10:44 MohShami I only have 6 servers configured with salt, that might be why
10:44 pvial I have around 300.
10:44 MohShami oh, I have 14 now, soooooooooo many :D
10:45 pvial It'll grow, I'm sure.
10:45 MohShami hopefully I will be adding our windows servers as well
10:45 MohShami I left this company and came back 3 years after
10:45 pvial FreeBSD and Windows in the same environment?
10:45 MohShami moving about 30 servers over
10:45 MohShami FreeBSD, Linux and windows
10:45 davidone MohShami: which version of salt are you using on freebsd?
10:46 MohShami .17.5
10:46 davidone is your salt-master a freebsd machine?
10:46 pvial Interesting.  Sadly, I retired my last FreeBSD servers and I'm using the latest Salt on Ubuntu.
10:46 MohShami I have poudriere get all updates every night and complie, love it
10:46 MohShami davidone, yes I am running it on freebsd
10:47 pvial version 2014.1.0...whatever that means
10:47 manicouman joined #salt
10:47 davidone MohShami: cool, I'm waiting saltstack 2014.1.0
10:47 MohShami pvial, why did you switch from freebsd to ubuntu, freebsd FTW
10:47 MohShami davidone, yeah same here
10:47 pvial poudriere is cool stuff
10:47 davidone probably he switched for systemd :D
10:47 davidone yes, poudriere is really cool
10:47 MohShami yeah, don't know how I used to administer servers before it
10:48 MohShami remember how painful upgrading perl was? :D
10:48 pvial Linux is more widely known.  It's easier to find people to admin the systems and FreeBSD's virtualization is immature.
10:48 davidone do you remember perl-after-upgrade script? :)
10:48 davidone pvial: right.
10:48 MohShami davidone, I deleted that from my memory when I left this job 3 years ago, they asked me to come back, and when I did, I found out no upgrade was done because they were afraid to upgrade perl :D
10:49 MohShami I have servers here running scripts from 2009
10:49 MohShami and no upgrade since 2010
10:49 davidone MohShami: LOL :)
10:49 MohShami pvial, true, but I never needed more than jails
10:50 davidone I'm trying to find a meaning for pillars
10:50 MohShami davidone, now with poudriere and salt, I'm having a blast
10:50 pvial MohShami, if i architect another infrastructure, in the future, it'll be using FreeBSD and jails.  :)
10:51 MohShami davidone, I used to have a script that gziped folders and uploaded them via ftp, now I used pillars to define the folders I want to back up, so I have some sort of a centralized backup system :)
10:51 davidone I am trying to find something I could do ONLY with pillars
10:51 davidone could you help me?
10:51 MohShami pvial, jails are awsome, but we now switched to vsphere
10:51 MohShami davidone, I can't think of something that can only be done with pillars, what are you trying to do?
10:52 davidone I think pillars could be useful if you have an heterogeneous datacenter
10:52 MohShami still haven't gone that deep
10:52 davidone so using pillars I can have simple formulas
10:52 MohShami I'd like to test salt with windows
10:53 MohShami you could also do that with grains, but pillars are easier to handle
10:53 davidone and in a pillar I can define: well if is rhel this package is exim, if is debian is exim4 and so on
10:53 MohShami s/handle/maintain/
10:53 MohShami yeah, but you can do that with grains as well
10:53 MohShami my understanding is that pillars are centrally managed grains
10:53 bhosmer joined #salt
10:54 MohShami please correct me if I'm wrong
10:54 davidone yes, but if you do all this stuff in your grains, your grains could become difficult to read
10:54 davidone or to understand
10:54 davidone o both :)
10:54 MohShami yeah of course
10:54 MohShami my backup script used grains in the beginning
10:54 MohShami after I added the 3rd host I relized pillars would be better
10:54 davidone LOL :)
10:55 davidone it makes sense
10:55 bezaban I've had a similar 'issue' grasping the concept of pillars if none of my data is sensitive (only distribute public keys atm)
10:55 davidone bezaban: that's another point: why is pillar more suitable for sensitive data?
10:56 bezaban I guess you could just target a regular state at single target and it would be more or less the same?
10:56 davidone that's what I'm just thinking
10:57 bezaban but in this case you don't need a separate state for each machine and the pillar calulates the data targeted for that machine
10:57 bezaban not sure myself, just getting started out myself, hehe
10:57 davidone bezaban: me too :)
10:57 davidone if you have multiple machines, you can create a state for this new class of machines
10:58 bezaban just doing some basics like managing site-wide packages, users and keys
10:58 davidone yes, me too
10:58 davidone and at this moment I'm just doing it with states
10:58 bezaban looking into writing a pillar to distribute some configs and bashrc for different security zones.  Not because I can't do it in a state, but I'd like a hands on attempt
10:58 davidone I created a couple of pillars just to find out how they works
10:59 davidone s/works/work
10:59 bezaban yeah
10:59 davidone something like this: http://hastebin.com/wiwijuqovi.django
10:59 arodizzy joined #salt
11:00 bezaban progress is slow, because I'm also doing long overdue administration on the machines as I go along, then keep a note that this needs to go into a state
11:00 davidone sorry, wrong link (1st release)
11:00 davidone http://hastebin.com/jukaromejo.django
11:01 bezaban along with some network redesign.  Full spring cleaning
11:01 davidone cool
11:01 ggoZ joined #salt
11:01 bezaban davidone: haha, packages.sls is almost identical to mine
11:02 Ryan_Lane joined #salt
11:02 davidone bezaban: same needs :)
11:02 davidone the less you install, the less would break up :)
11:02 bezaban yeah, just some stuff I don't want to need to install when I need it
11:02 davidone yes, something like a 'bootstrap' procedure
11:02 davidone with packages, config files, users
11:03 davidone and later I need to specify what kind of server is
11:03 davidone etc etc :)
11:03 bezaban yeah.
11:06 bezaban I've had luck with setting up a new vm, running the states against it an'
11:06 bezaban oops. home machine went down for a little
11:06 bezaban I've had luck with setting up a new vm, running the states against it and seeing what I'm missing
11:08 ndrei joined #salt
11:09 echos joined #salt
11:10 [vaelen] joined #salt
11:12 chutzpah joined #salt
11:18 bhosmer joined #salt
11:21 Outlander joined #salt
11:24 echos joined #salt
11:25 [vaelen] joined #salt
11:26 bhosmer_ joined #salt
11:26 emostar left #salt
11:26 emostar joined #salt
11:27 younqcass joined #salt
11:35 ndrei joined #salt
11:37 sroegner joined #salt
11:37 fllr joined #salt
11:41 bezaban actually, does salt send the entire state tree to every machine and lets the minion parse it?
11:41 bezaban so that pillars make more sense :P
11:41 bezaban then the minion sorts what is relevant
11:43 babilen Where can I find a list of grains values for os_family?
11:50 yomilk joined #salt
11:54 bram_ joined #salt
11:54 bram_ hi all... anyone here working with Vagrant+salt? the #vagrant chan is not very responsive...
11:55 mgw joined #salt
11:55 babilen bram_: I am
11:56 babilen bram_: Do you have an actual question? :)
11:56 bram_ anyway, the question is: why do I need to supply both the master AND minion key in the "master" vm: http://dpaste.com/1719041/
11:56 bram_ this does't make any sense to me...
11:56 babilen bram_: I think it is not actually needed and would, maybe, be the "masterless" configuration?
11:57 bram_ when I remove the salt.minion_key = ... and salt.minion_pub = ... from the config of the master host it tells me: "you need to supply all keys"
11:57 babilen bram_: I have the same snippet in my configuration and wondered the same though
11:57 bram_ I think it's a vagrant error
11:57 babilen bram_: Does that still happen even with salt.no_minion = true?
11:58 bram_ that's the default though, I think, lemme check
11:58 bram_ "no_minion (boolean) - Don't install the minion, default false"
11:58 bram_ let me try to add it explicitly
11:59 babilen bram_: You set it to true in the Vagrantfile you pasted, but I wanted to explicitly verify this. My understanding would have been that the minion key is not needed on the master if you don't set the master to be its own minion.
12:00 bram_ well, I don;t want the master to be included as a minion
12:00 bram_ I want the master to be ONLY master
12:01 bram_ if no_minion=true then why would I need to add the minion pub/key?
12:01 babilen exactly, you shouldn't have to
12:01 l0x3py joined #salt
12:01 bram_ so it's a vagrant bug...
12:02 bram_ I'll submit a bug report on the vagrant github...
12:03 babilen thank you
12:04 faldridge joined #salt
12:04 babilen bram_: I stumbled over this when I removed the salt.master_key from the minion config and then tried to modify the master accordingly
12:05 bram_ well, actually the minion is fine without the master key
12:05 bram_ that's kinda weird
12:05 bram_ so it's not symetric
12:05 manicouman joined #salt
12:05 babilen bram_: That's what caught me out.
12:05 bram_ i.e. leaving out the master key in the minion doesn;t trigger the bug
12:05 bram_ ah, yeah
12:05 giantlock_ joined #salt
12:05 bram_ I'll post the github bug link in here in a bit
12:05 bram_ maybe you can +1 it ;)
12:05 bram_ gimme a second...
12:06 bastion1704 joined #salt
12:06 nicksloan joined #salt
12:09 babilen bram_: sure
12:09 bram_ babilen: https://github.com/mitchellh/vagrant/issues/3156
12:12 CeBe joined #salt
12:13 Teknix joined #salt
12:14 arodizzy joined #salt
12:14 MohShami is it possible to combine the values of the same pillar from multiple files?
12:14 raizyr joined #salt
12:14 tmandry joined #salt
12:14 babilen MohShami: Could you elaborate a bit more on that?
12:15 jrdx joined #salt
12:16 oz_akan_ joined #salt
12:16 voozl joined #salt
12:17 MohShami babilen, sure, I have 2 sls files, generic.sls and web1.sls, and here are the contents http://hastebin.com/pidabimogi.hs I want the value of backupFolders to have all 3 items for a host, would that be possible?
12:17 voozl I
12:17 voozl I'm new to salt and I have some trouble understanding how a command from the module documentation can be 'used' from a .sls state file
12:17 voozl for example, looking at the reg module ( http://docs.saltstack.com/ref/modules/all/salt.modules.reg.html#module-salt.modules.reg ), how would I add a statement to my .sls file that says that a certain key must exist
12:18 bhosmer joined #salt
12:20 juniorzoid left #salt
12:21 johtso joined #salt
12:22 babilen MohShami: No, that is not currently possible with any of the released versions of salt as none of them use the copy function needed for that. IIRC this change has already been commited to the dev version of salt though and I haven't actually tested with 2014.1
12:24 babilen MohShami: http://paste.debian.net/87261/ is the patch that is needed and https://github.com/saltstack/salt/pull/9184 is the corresponding bug
12:26 babilen MohShami: Oh, but I am not sure if that would allow to append to these lists ... What you could do is "backupFolders: "/etc": True" with that ... but then you would need additional logic for that.
12:26 bram_ babilen: ok, now vagrant up went just fine. however if I call vagrant provision after vagrant up (my idea was to use this for updating the SLS' on the master) I get another error... How do you apply the high state again after the first vagrant up?
12:27 babilen bram_: I typically just log in to the master and run "salt '*' state.highstate" manually
12:27 bram_ I get an upload permissions problem on salt/keys/minion.pem which is totally weird
12:27 babilen bram_: I don't quite follow
12:28 bram_ ah ok, and you have your root shared with the vagrant host?
12:28 bram_ *salt root
12:28 babilen bram_: I use shared folders^Wdirectories for my gitfs and pillars, yes
12:28 bram_ okay
12:29 bram_ the error: well vagrant up works fine, my minion key is seeded into the master, everything fine and dandy. Then while the hosts are up I run "vagrant provision" and vagrant tells me it can't upload the minion public key because of file permissions... ( ... heh? )
12:33 dangra joined #salt
12:36 babilen Is there a list of all values a specific grain can have? os_family in particular? (I know it is impossible to enumerate all, but still ...)
12:37 fllr joined #salt
12:37 bram_ babilen: pffff, looks like the vagrant-salt-integration is not very thorough... :( found another bug... :(
12:38 babilen bram_: There are many, yeah
12:38 bram_ even though I have salt.no_minion = true on the master re-provisioning failes because it tries to search for a salt-minion process!
12:38 nkuttler babilen: hrm, iirc the code can use various third party sources, so probably no
12:39 ipmb joined #salt
12:39 babilen nkuttler: Okay; Any idea which derivatives os_family == "Debian" comprises? Is it "RedHat" and "Arch" ?
12:39 babilen nkuttler: Just not sure what to use and I don't have enough different boxes around to test this all ...
12:40 babilen And there simply doesn't seem to be a list of common values anywhere (that I have been able to find)
12:41 nkuttler babilen: salt/grains/core.py _OS_FAMILY_MAP
12:41 babilen bram_: libvirt was wonderful -- it was *so* much faster than virtualbox that I am crying everytime I am running "vagrant up" with virtualbox as provider these days
12:41 babilen nkuttler: Use the source! Thanks :)
12:41 nkuttler ;)
12:42 bram_ babilen: actually on this machine it boots up crazy fast... maybe it's the SSD that helps, but booting a precise64 standard vm takes like 5 seconds, it;s crazy fast
12:43 nkuttler babilen: actually, os_data() as well
12:43 bram_ or maybe precise64 is just not very a very complex VM :)
12:43 babilen bram_: Here it takes ages to copy the basebox ...
12:43 bram_ babilen: really? here's it's super quick too!
12:43 babilen bram_: During the initial "vagrant up" that is. Booting is fine afterwards (even if virtualbox is known to be much slower than KVM or Xen)
12:44 MohShami babilen, thanks for the info mate, sorry was AFK
12:44 babilen bram_: Hmm ... my /home is served via NFS, but I thought I had moved everything vagrant/virtualbox related to a local SSD
12:45 babilen MohShami: So in summary: I don't think that what you want to do is supported at all right now, but salt is moving fast so I can't be sure of that. The patch I mentioned before allows for proper merging of dictionaries, but for that you need dictionaries not lists (as in your example)
12:45 MohShami babilen, thanks again mate, it's not important for now, but I'll keep an eye out for it
12:46 babilen MohShami: Not sure if a proper merging strategy has or should be implemented for lists. I would like that, but is seems to be more complicated. See https://github.com/saltstack/salt/issues/3991 for a discussion. I would *love* that feature
12:47 MohShami babilen, checking
12:47 babilen So, I have http://paste.debian.net/87263/ -- would changes to a file in salt://nagios/nrpe/files/nrpe.d/ (e.g. salt://nagios/nrpe/files/nrpe.d/foo.cfg ) trigger a restart of the service too?
12:49 MohShami babilen, oh I'd love that as well
12:49 MohShami babilen, speaking of nagios, I was hoping of creating a nagios server that could collect the list of managed hosts from salt and creating the configuration files automatically, would that be possible?
12:49 bram_ babilen: *sigh* https://github.com/mitchellh/vagrant/issues/3158
12:50 babilen MohShami: I might look into that, but I am not aware of anything that would accomplish that at the moment (which really doesn't mean that it doesn't exist)
12:51 MohShami babilen, that would be an awesome feature
12:51 MohShami monitoring is always neglected
12:53 babilen MohShami: I agree; I am currently working on the nagios-formula to make it usable on non-Arch systems (essentially fixing my bug https://github.com/saltstack-formulas/nagios-formula/issues/1
12:53 oz_akan_ joined #salt
12:53 babilen So, does anybody know how that file.recurse state would behave with watch_in?
12:54 MohShami good luck with that mate
12:54 MohShami babilen, I have it set up here, it works well
12:54 MohShami let me get you a sample
12:55 babilen MohShami: I am quite far with the rewrite of that formular, but I happily incorporate other good ideas in addition to what I have now
12:56 babilen nkuttler: Oh god, that code!
12:56 oz_akan_ joined #salt
12:56 pdayton joined #salt
12:56 MohShami babilen, does this help you? http://pastebin.ca/2655728, I'd be glad to help
12:56 blee joined #salt
12:57 MohShami babilen, it uses watch, not watch_in, but AFAIK they behave the same
12:57 younqcass joined #salt
12:57 babilen MohShami: Well, I don't need the script part. Just wanted to know if changes in any of the files that would be recursively copied will trigger the "watch_in" event
12:58 MohShami babilen, IIRC yes it does
12:59 MohShami babilen: currently working on a formula that genates NSD configuration files from a database, I don't have access to it now but IIRC that's how I did it
12:59 babilen great
13:00 MohShami let me get the extranal drive and try to find it, give me a couple of minutes
13:01 pvial joined #salt
13:02 pvial hello, anyone have any luck getting salt cloud to work on an openstack cloud?
13:02 pvial not Rackspace but any other openstack cloud.
13:03 mpanetta joined #salt
13:03 babilen MohShami: Thanks, much appreciated!
13:04 nkuttler babilen: yeah.. i read it a while ago because there was a bug with detecting debian..
13:05 magosinvello left #salt
13:05 MohShami babilen: it's a pleasure, I just checked, the state I used was file.absent, not file.recursive
13:05 babilen nkuttler: Those functions are ... so ... massive.
13:05 MohShami babilen: I'll try altering one state to try
13:05 pvial i have a feeling that if i write some bash wrappers around nova commands i'll have something more functional and powerful than salt-cloud ;)
13:07 pvial it looks like salt-cloud doesn't support environments with only private IPs too...
13:08 MohShami babilen, seems watch_in doesn't work
13:09 MohShami babilen: http://pastebin.ca/2655735 doesn't work, I commented the watch section and added a watch_in for file.recursive
13:09 MohShami sorry mate
13:09 mike25ro hey guys... i have a jinja list :    ['a','b','c'] - which should be a list of servers and i want to get that list as a jinja variable ... but also append and prepend something.... so it should be : pre_a_post,pre_b_post,pre_c_post ... how can i do that?
13:09 babilen MohShami: *sniff*
13:10 MohShami babilen: can't you do it with watch? why does it have to be a watch_in?
13:10 shano joined #salt
13:10 jrdx joined #salt
13:11 babilen MohShami: I can probably do something like https://github.com/saltstack/salt/issues/663 - I simply decided not to use watch in the services, but watch_in in the related files.
13:11 pvial https://groups.google.com/forum/#!topic/salt-users/v3Lh50XnxXA  yep salt-cloud doesn't support environments with only private IPs.  i'm running the latest salt which includes some patch to apparently fix this but it doesn't work.
13:11 Katafalkas joined #salt
13:12 quickdry21 joined #salt
13:12 pvial anyone else seeing this issue?  https://groups.google.com/forum/#!topic/salt-users/v3Lh50XnxXA
13:12 cpenner461 joined #salt
13:14 babilen pvial: So, have you tried 2014.1 ?
13:14 bhosmer_ joined #salt
13:14 pvial babilen:  yes, that's what i'm running and it just times out after 10 minutes and destroys the VM.
13:15 mpanetta_ joined #salt
13:15 MohShami babilen: can you use a cmd.wait?
13:16 racooper joined #salt
13:18 babilen MohShami: I can use whatever I want as I am writing the formula, but I am just not sure how to deal with this most idiomatically. I used http://paste.debian.net/87271/ somewhere else
13:18 juicer2 joined #salt
13:18 torrancew joined #salt
13:19 MohShami babilen: good luck mate :)
13:19 akl joined #salt
13:19 akl joined #salt
13:20 babilen MohShami: Just to clarify: file.recurse doesn't work with watch_in ? I understood that correctly haven't I?
13:20 R_Shackleford joined #salt
13:20 MohShami babilen: on .17.5 sadly no
13:21 babilen I use 2014.1 -- guess I'll have to try it myself
13:21 MohShami 2014.1 is out, nice
13:22 MohShami it's not yet available in ports :(
13:22 babilen FreeBSD?
13:22 MohShami yep
13:22 MohShami only run anything else if I have to
13:22 pvial MohShami, i believe cedwards is the port maintainer still.
13:23 MohShami pvial, thanks mate, hopefully he'll get the new version out soon :)
13:23 babilen I'm a Debian person, but I heard that FreeBSD packages have been delayed for some reason, but I am fuzzy on the details
13:23 MohShami .17.5 works great for now, so I'm not in a hurry
13:25 Damoun joined #salt
13:28 patrek joined #salt
13:29 MohShami cedwards: hey mate, any plans for releasing 2014.1.0?
13:30 pvial MohShami, give him like 30 minutes or so.  he's probably on his way to work.
13:30 vejdmn joined #salt
13:30 MohShami pvial, 2 more hours and I'm going back hope, different time zones :D
13:31 MohShami s/hope/home/
13:32 mdasilva joined #salt
13:32 mdasilva hey all
13:34 mdasilva when using multiple environments, can association with an environment only be defined using the top file and therefore a state.highstate command?  would a state.sls call only look for the state in the base environement?  tia
13:34 higgs001 joined #salt
13:37 fllr joined #salt
13:38 ipmb joined #salt
13:41 mpanetta joined #salt
13:41 ajprog_laptop joined #salt
13:43 GradysGhost joined #salt
13:45 vbabiy joined #salt
13:47 fllr joined #salt
13:47 oz_akan_ joined #salt
13:48 fllr joined #salt
13:52 thayne joined #salt
13:56 mpanetta_ joined #salt
13:56 qybl joined #salt
13:57 ndrei joined #salt
14:04 kaptk2 joined #salt
14:07 danielbachhuber joined #salt
14:09 seapasulli joined #salt
14:09 mike25ro pfff ... i am going crazy here
14:09 mike25ro http://paste.debian.net/87281/
14:10 mike25ro in a for loop .. it prints correctly this ...  {{  i ~ ' ' ~ jinjatest.dns.nameservers[i] ~ ','  }}    BUT i want to be able to assign the value to a variable.. and increment/concatenate that variable ... like:     {%  set server_list =  server_list ~  i ~ ' ' ~ jinjatest.dns.nameservers[i] ~ ','  %}
14:10 mike25ro is this possible?
14:11 richardr joined #salt
14:12 mike25ro my server_list is always empty... when i print it using {{ server_list }}
14:12 mike25ro :((((
14:12 jeremyBass1 joined #salt
14:12 jaimed joined #salt
14:13 Networkn3rd joined #salt
14:14 opapo joined #salt
14:15 HeadAIX joined #salt
14:15 tmandry joined #salt
14:16 mike25ro ... anyone?
14:16 jrdx joined #salt
14:18 rrives joined #salt
14:22 rrives Hello new here, I have deployed several win packages using salt and then I run into a few that don't install, I just get a ------ under the description my most current example is running this command:
14:23 rrives salt 'name of minion' pkg.install winscp.sls
14:23 rrives just using the git repo
14:23 rrives and I get only a ----- as the result, is there any reason for that?
14:24 rostam joined #salt
14:24 elfixit joined #salt
14:25 bram_ joined #salt
14:26 mgw joined #salt
14:33 sroegner mike25ro: not quite sure i understand your syntax there but appending to an array inside a jinja loop works with {%- do server_list.append(item) %}
14:34 mike25ro sroegner: the basic idea is that i have a variable outside the loop and i want to append to that variable stuff from the loop... is that possible?
14:35 mike25ro sroegner: server_list is just a simple variable : {%  set server_list = ''  %}
14:36 mike25ro and in the end i want that server_list should be something like   ldap://ipserver1,ldap://ipserver2 where the IPs are from loop
14:36 mike25ro am i making any sense?
14:37 akoumjian joined #salt
14:38 mike25ro it's been driving me crazy for the last 2 hours
14:39 teskew joined #salt
14:39 cpenner461 joined #salt
14:40 MohShami cedwards: hey mate, ETA for releasing 2014.1.0?
14:40 MohShami cedwards: thanks for the great work btw
14:41 sroegner mike25ro: i am sure this can work, probably the approach isn't ideal
14:41 fllr joined #salt
14:41 mike25ro sroegner: i want to store... data from loop... in a variable and use it.. somewhere else...
14:41 symroe joined #salt
14:42 repl1cant /set aspell.check.enabled on
14:42 repl1cant oops
14:42 mike25ro sroegner: ... beats me mate... i have no idea how to do it
14:42 symroe Am I right in thinking that I can use a top file with a single top level env (base) and matching on grains (with match: compound) in a way that means I can overwrite the '*' matcher with the grain specific pillar items?
14:42 sroegner whats the resulting string supposed to look like again and waht are we coming from?
14:42 teskew joined #salt
14:43 mike25ro sroegner: i will paste it in a second... let me create a pastebin
14:43 sroegner k
14:44 alunduil joined #salt
14:44 chrisjones joined #salt
14:44 symroe RE the above, this is my pillar top.sls: https://gist.github.com/symroe/9508338.  I want the vars in demo.foo to overwrite the base.foo vars when the grain matches release_stage:demo
14:45 sroegner anybody saltstack around?
14:46 mike25ro sroegner: http://paste.debian.net/87287/
14:46 mike25ro sroegner:  thanks for wasting your time
14:47 sroegner lol
14:47 repl1cant anyone using salt primarily for orchestration in a HA environment?
14:47 repl1cant say orchestration with multiple masters, or multiple masters behind a vip, etc
14:48 sroegner mike25ro: how about {%- set server_list = ','.join(jinjatest.dns.nameservers) %}
14:48 higgs001 joined #salt
14:49 mike25ro sroegner: i need also to .. prepend something
14:49 mike25ro sroegner:  my Q is if it is possible to update a variable outside a loop
14:49 mike25ro sroegner:  the example i gave is simplified
14:51 sroegner mike25ro: fair enough
14:52 sroegner mike25ro: i think in the sense that you are asking the question the answer is probably no
14:52 mike25ro and so far.. i can not figure out how to update a variable outside a loop
14:52 mike25ro damn :)
14:52 mike25ro why oh why? :)
14:52 sroegner mike25ro: which is not to say there is no solution
14:52 mike25ro sroegner:  :)
14:52 mike25ro can you think of anything?
14:53 sroegner mike25ro: can you try and paste a more realistic one?
14:54 Sypher joined #salt
14:54 mike25ro that is as close as possible...
14:54 mike25ro the only difference would be an append..  server_list =  server_list + 'ldap://' +  jinjatest.dns.nameservers[i] + ','
14:55 sroegner mike25ro: meaning every member of the stringified list needs to be a ldap url?
14:55 mike25ro yes
14:57 chrisjon_ joined #salt
14:57 bizboo joined #salt
14:58 mike25ro sroegner: that string with multiple ldap servers would then ... be part of a CLI command to install ldap ... one command line
14:59 mike25ro that's why i need ... everything in one variable.
14:59 mike25ro + is a brain challenge .. it seems to update a variable outside a loop... i wasn't expecting that :D
14:59 rallytime joined #salt
14:59 Mr_N So... 2014.1 is still broken?
15:00 Mr_N Since all my multi master minions still die at launch.
15:00 xmj how is that broken? works here
15:00 viq Huh. Trying to use salt-ssh pkg.list_upgrades against a debian host, it has msgpack-python installed, yet all I get is "[CRITICAL] Unable to import msgpack or msgpack_pure python modules" - any ideas?
15:00 Mr_N xmj: Since I updated to 2014.1 a while back, all of my minions set up to run as multi master, die shortly after launch.
15:00 viq using 2014.1.0
15:00 Mr_N yes
15:01 arodizzy joined #salt
15:01 viq Mr_N: if that's to me, mine was a clarification of my previous sentence ;)
15:01 Mr_N Ah
15:01 Mr_N sorry ><
15:01 viq no problem ;)
15:01 Mr_N I get a traceback in tune_in
15:01 Mr_N and it yells StopIteration
15:02 Mr_N This has been broken for ~weeks now.
15:02 higgs001 joined #salt
15:02 Mr_N totally not ok.
15:02 sdlarsen joined #salt
15:02 sroegner mike25ro: https://gist.github.com/sroegner/9508741
15:02 xmj Mr_N: patches accepted
15:02 xmj you're using a free product...
15:02 bhosmer_ joined #salt
15:02 Mr_N xmj: Indeed.  I'd love to have time to patch it :/
15:02 Mr_N right now I'm just trying to unbreak my minions
15:02 Mr_N I'm mostly irked because the previous version is not in yum
15:03 Mr_N so rolling back wasn't fun
15:03 xmj Mr_N: maybe it'd be much faster if you actually patched it and got it into upstream? :)
15:03 xmj Mr_N: it's in EPEL..
15:03 Mr_N Really?  last time I looked in EPEL it only had the 2014
15:03 mike25ro sroegner: interesting :) thanks for all your trouble!
15:03 Mr_N Maybe I'm totally blind ><
15:03 mike25ro sroegner: owe you a beer ... or more :)
15:03 xmj i don't get it
15:03 xmj which one do you want?
15:03 xmj 2014.1.0 is in epel.
15:03 Mr_N I want to downgrade.
15:04 xmj ahh, hum, huh. no idea how that works
15:04 Mr_N Normally you can do a yum downgrade
15:04 sroegner mike25ro: np
15:04 Mr_N but if the previous version isn't in the repo, you have to do it by hand.
15:04 Mr_N I'd love to help and patch it, but I was quite new to salt at the point that this happened as a user; I am not yet at a point where I feel like I could be truly helpful without a large overhead :/
15:05 xmj from my own experience "I'm not yet at at point where $x" is exactly when you should do $x :D
15:05 Mr_N Hah.  Unfortunately, I'm on the payroll to deliver functionality utilizing salt
15:06 pvial hello, anyone know a solution to this?  i'm using salt 2014.1.0 and salt-cloud 0.8.11 https://groups.google.com/forum/#!topic/salt-users/v3Lh50XnxXA
15:06 Mr_N And I can't disappear for a few weeks into familiarizing myself with the salt codebase
15:06 Mr_N a bit of a catch 22 :/
15:06 _fllr_ joined #salt
15:06 xmj we all are, why else would we be here during working hours ;)
15:06 Mr_N Also, I'm personally very hesitant to say that I know how to solve a problem in a framework I'm new to.  I've made a fool of myself in the past ><
15:07 Mr_N xmj: Good point :P
15:07 pvial Mr_N:  you can downgrade via pip.  `pip install salt==0.17` for example
15:08 Mr_N pvial: If I did the original install via yum; pip will handle this properly?
15:08 pvial Mr_N, remove what you've installed with yum and install via pip.
15:08 Mr_N Fair enough.  Thanks :)
15:08 dave_den joined #salt
15:08 Mr_N Now if only there was a way to bind a pkg.install to a certain version in salt...
15:09 pvial the salt guys haven't created uniform ways to install all of their software.  salt-cloud, for instance, is only available via git or pip and salt-cloud doesn't work...in my case, sadly.
15:09 Mr_N To the documentation.  Thanks guys.
15:10 pvial Mr_N, if you have your own yum repo you can only carry the versions you want in the repo.  i do this with Ubuntu.
15:10 Mr_N pvial: I've been struggling to get the win_repo stuff for a while.  I should probably take a break from that and do the linux repos.
15:10 viq Trying to use salt-ssh pkg.list_upgrades against a debian host, it has msgpack-python installed, yet all I get is "[CRITICAL] Unable to import msgpack or msgpack_pure python modules" - any ideas? using 2014.1.0
15:10 Mr_N (I've had literally _no_ success with the win_repo stuff)
15:11 Mr_N To the extent that going through the example in the documentation didn't work.  (posted an issue on that...)
15:11 Mr_N So I've been hesitent to dive into the custom repos elsewhere until there's more of a knowledge base.
15:12 pvial yikes.  Mr_N, you and are having a lot of fun.
15:12 Mr_N Indeed I am...
15:12 Mr_N FWIW, salt has enabled a ton of shit really easily
15:12 Mr_N but some of the more esoteric stuff, multi mastering and win repo, has been a nightmare
15:12 [diecast] joined #salt
15:13 rainmanjam joined #salt
15:13 rrives I agree win repo is much more difficult
15:13 Mr_N (I don't mean to place blame.  It's just not really been a smooth or successful process.)
15:13 kzx joined #salt
15:13 Mr_N rrives: Do you by any chance have it working?
15:14 Mr_N I'd really appreciate getting to pick someone's brain on known good configs.
15:16 kzx what is the support level on windows?
15:16 rrives only a few, most of them just return ----- but I did have some luck initially
15:16 Mr_N Yeah; I get the returning ----- all the time.
15:16 voozl anyone have any luck running XAPI on xenserver?  I understood it shouldn't work, but I'm curious
15:16 Mr_N see my issue at https://github.com/saltstack/salt/issues/10948
15:17 rrives I am even using the win repo that is posted
15:17 Mr_N Yep
15:17 Mr_N That's _exactly_ what my issue documents
15:17 Mr_N I do the example pretty much verbatim
15:17 kzx am evaluating salt vs puppet, tried salt first and it just falls over
15:17 rrives right
15:17 Mr_N kzx: You using multi mastering?
15:18 Linz joined #salt
15:18 kzx Mr_N: no, single master on CentOS 6
15:18 rrives I have had a lot of luck with Linux, but the win part is what I need most now.
15:18 zirpu joined #salt
15:18 Mr_N Hm.  That was what caused mine to fail silently, lately. (silently without -ldebug, rather)
15:18 conan_the_destro joined #salt
15:18 kzx Mr_N: minion is Win 7 64bit
15:18 Mr_N rrives: Exactly.  My deployment is mostly centOS, but a third windows.
15:19 zirpu left #salt
15:19 rrives I honestly can't even find the log file that shows what part is failing?
15:19 Mr_N kzx: Hmm. Try starting the minion from the command prompt
15:19 Mr_N rather than as a service
15:19 Mr_N and add -ldebug
15:20 rrives okay
15:20 pvial man, salt is really not very backward compatible is it.  upgrading is a pain...everything breaks.
15:20 mike25ro sroegner: thanks again buddy.. it was such a simple solution .... thanks again
15:20 Mr_N (I am a complete fail at windows, so I may be saying nonsensical things, but that's what I do to diagnose in linux)
15:20 kzx Mr_N: I'll have to try that  here is error I am getting http://pastie.org/8911643
15:20 Mr_N pvial: Yep, unfortunately :/
15:20 Mr_N kzx: Ahh, ok.
15:20 Mr_N I was thinking of a different failure mode, sorry ><
15:21 Mr_N still pass -ldebug though.
15:21 Mr_N It may give you a dump of the failed state compilation
15:21 cachedout joined #salt
15:21 pvial Mr_N:  is there a matrix somewhere that tells which master versions work with which minion versions?
15:21 MohShami is there a way to list all mine_fuctions of a host?
15:21 * Mr_N is probably showing his complete naievity of salt debugging
15:21 jeremyfelt joined #salt
15:21 Mr_N pvial: Not AFAIK.  I try to stick to .17
15:21 Mr_N Since that's what I got my deployment working on
15:22 Mr_N well, "working"
15:22 kzx Mr_N: I tried salt-call grains.items and there are a bunch of failures getting netwok interfaces and physical memory
15:22 pvial ugh, i have to upgrade 300 hosts now :(
15:22 Mr_N kzx: Man, that's strange.  One sec, I'll run a grains.items on my win minion and see what happens.
15:22 ChaosPsyke joined #salt
15:23 MohShami pvial, even though I only have 14 servers, but on freebsd I just did "salt '*' cmd.run 'pkg install -y py-salt" and then a salt minion restart, and bam :)
15:23 pvial does anyone know if the SaltStack guys hang out here anymore or did they all lock themselves in their offices once SaltStack was born?
15:23 terminalmage pvial: Our release notes describe when there is a protocol incompatibility.
15:23 Mr_N No, it seems to work for me, but my minion is win server 2k8
15:23 Mr_N so YMMV
15:23 terminalmage the minion should never be a higher version than the master
15:24 gnugnu9 joined #salt
15:24 kzx Mr_N: here is the output http://pastie.org/8911653 , it looks like there is an exception getting os data
15:25 kermit joined #salt
15:25 pvial terminalmage:  so i should be reading the release notes of every version to figure out what is compatible?  i was recommended to upgrade to salt 2014.1.0 and i'm running a mix of 0.14 and 0.17 on the minions and they all don't talk now.  this is actually the tip of the iceberg in my salt/salt-cloud issues.
15:26 kzx Mr_N: specifically win32com/client/dynamic.py part AttributeError: winmgmts:.Win32_PhysicalMemory, is it because this is a 64bit box? should I install the 32bit flavor of Salt?
15:26 terminalmage pvial: so, in general, you like to upgrade software without reading release notes? doesn't sound too smart to me.
15:27 gnugnu9 joined #salt
15:27 Mr_N I... honestly have no idea.
15:27 Mr_N I wish I could be more helpful
15:27 Mr_N but none of that is familiar to me
15:27 pvial terminalmage:  usually software is made to be backward compatible.  when does salt go 1.0?
15:27 kzx Mr_N: well thanks for your time anyway, I appreciate it
15:27 Mr_N I'd be very surprised if you _didn't_ want OS bittage and salt minion bittage parity
15:27 Mr_N so I doubt that's the issue, FWIW
15:28 pass_by_value joined #salt
15:28 rrives so -ldebug did show some good information
15:28 rrives thank you
15:28 Mr_N I'm just by no means an expert at this stuff, especially the windows part.  (in fact, I'm quite the opposite )
15:28 Mr_N np rrives
15:28 TheRealBill_here joined #salt
15:29 kzx Mr_N: that is why I am here, I can configure most of the linux stuff in multiple ways, I knwo how to do that, I was just hoping to automate windows
15:29 Mr_N Yep.  I've been looking for help in that respect for ~weeks now
15:29 Mr_N had little luck, including a pending issue for a week now
15:29 [diecast] joined #salt
15:29 Mr_N very few people do that part
15:29 Mr_N (the windows part)
15:30 rrives no I know that my error is [ERROR   ] Unable to locate package bginfo.sls
15:30 kzx maybe I would have more luck with Puppet? Anyone tried that for Windows?
15:30 terminalmage pvial: 0.14.0 is a 1-year-old release, in a very actively-developed project. if you're that exasperated by backwards incompatibility, given that information... I don't know, maybe open source just isn't for you.
15:30 Mr_N I'm sure it works, running Puppet is juts a bit painful.
15:30 Mr_N just*
15:31 Mr_N Puppet is very robust and mature system.
15:31 Mr_N It's also not my favorite to work with, however.
15:31 taion809 joined #salt
15:31 arodizzy joined #salt
15:31 kzx Mr_N: I'll take anything that can automate windows ;-)
15:31 Mr_N Yeah; take a look at puppet.
15:31 pvial terminalmage:  upgrading is a pain but not really my main issue.  my main issue appears to be a bug.  where does one open a ticket?  i was told to on the mailing list.
15:31 Mr_N It's been around a lot longer than Salt so it's had a time to mature some of the more esoteric features
15:32 terminalmage pvial: we have an issue tracker here: https://github.com/saltstack/salt/issues?milestone=40&amp;state=open
15:32 Mr_N I just really really like salts model and configuration FAR more than puppet, so I try to put up with the downsides.
15:32 terminalmage pvial: https://github.com/saltstack/salt/issues
15:33 rrives even with one of the win repo ones, I get the same error: [ERROR   ] Unable to locate package 7zip.sls
15:33 Mr_N And I guess I hope salt iterates to being comparable in the future.
15:33 Mr_N A "long term investment" so to say
15:33 ipalreadytaken joined #salt
15:33 Mr_N Yeah rrives as I said, I pushed that exact issue a week ago
15:33 aleszoulek joined #salt
15:34 pvial terminalmage:  thanks #11170 submitted
15:34 rrives @Me_N you have gotten no answer yet?
15:34 Mr_N They said "they'd look at it" shortly after I posted, then radio silence
15:34 Mr_N I assume they're busy with the new release
15:34 Mr_N but /grump/
15:34 jrdx joined #salt
15:34 JaredR joined #salt
15:34 rrives great tumbleweeds...
15:35 dave_den pvial: salt-cloud is now included in salt itself.
15:35 terminalmage pvial: see it, thanks for the report
15:35 Mr_N I dunno.  I'm gonna wait a week or two more then make some more noise
15:35 Mr_N I figure I'll give the benefit of the doubt that shit is probably busy.
15:35 pvial dave_den:  does that mean that salt-cloud is now using the same crazy versioning scheme that salt is...i.e. 2014.1.0
15:35 Mr_N Have other things to hack on in the meantime.
15:35 rrives I wish I had that luxury, I have to demo this tomorrow
15:35 Mr_N Ouch...
15:35 rrives yeah
15:35 smcquay joined #salt
15:36 Gareth morning
15:36 Mr_N Ohi.
15:36 rrives I could demo this easy in Linux, but they don't care about that.. lol
15:36 cro Mr_N: kind of you, we did just move offices.  What's the issue # for the one you are wrestling with?
15:36 Mr_N Yeah; look at puppet then.
15:36 frasergraham joined #salt
15:36 Mr_N 10948
15:37 dave_den pvial: it means you should not have salt-cloud code installed, because it may cause problems
15:37 Mr_N The example win_repo in the documentation doesn't appear to be working, rrives reports a similar problem
15:37 Mr_N (as have other people in here over the last week, so I assume I'm not completely insane)
15:37 Mr_N ><
15:37 ccase joined #salt
15:37 pvial dave_den:  not according to this http://salt-cloud.readthedocs.org/en/latest/topics/install/
15:38 dave_den pvial: http://docs.saltstack.com/topics/cloud/index.html
15:38 Mr_N I should really post the issue for the broken multimastering, but haven't had sufficient time to scrape together a bug report I wouldn't be embarassed by
15:38 pvial dave_den:  it's the same instructions
15:38 cro pvial: those docs are not correct, you're right
15:39 Mr_N cro: Thanks for taking a look, I understand the volatility of a new project :(
15:39 forresta joined #salt
15:39 Mr_N "new"
15:40 terminalmage pvial: also in the release notes, fwiw  http://docs.saltstack.com/topics/releases/2014.1.0.html#salt-cloud-merged-into-salt
15:40 cro pvial: that's actually quite embarrassing.  I'll create a pull request.
15:40 pvial terminalmage: thanks
15:41 cro terminalmage: he's right though, how can anyone know what's right when the docs and the release notes don't match?
15:41 terminalmage gah
15:41 terminalmage yep
15:41 terminalmage I wasn't fully caught up
15:41 terminalmage hadn't reached that part yet
15:43 kickerdog joined #salt
15:43 forresta cro, are you going to join us on the documentation crusade?
15:44 Mr_N FWIW I think my issue may be related to improper documentation
15:44 Mr_N since I don't want to assume all of win_repo is broken
15:44 terminalmage pvial: also, how did you install salt and salt-cloud
15:44 Mr_N and was quite verbatim to the documentation on that issue.
15:44 pvial terminalmage:  via pip
15:44 cro forresta: as much as possible
15:45 pvial i've just retested my issue after uninstalling salt-cloud and salt and reinstalling salt 2014.1.0 and problem persists.
15:45 terminalmage pvial: OK... I'm not certain of the possibble differences in paths, so it might be prudent to uninstall both and then install 2014.1.0
15:45 Gareth forresta: whose bringing the coconuts?
15:45 pvial terminalmage:  exactly what i did :)
15:45 terminalmage pvial: fantastic
15:46 forresta how stupid of me, I forgot to log off from home, so now I'm in here twice, awesome
15:46 forresta Gareth, I hope you are!
15:46 forresta Gareth, I'm bringing Mexican food
15:46 terminalmage forresta: double the fun
15:46 forresta terminalmage, yea
15:47 terminalmage cro: are you on the docs fix then?
15:47 bemehow joined #salt
15:47 terminalmage or do you want me to grab it
15:47 cro I'm on it :-)
15:47 terminalmage cool
15:49 repl1cant ok, I missed something. Broken multi masters?
15:49 forresta repl1cant, yea in 2014.1
15:50 forresta cro, You need to take a picture of Joseph without his beard
15:50 repl1cant this one? https://github.com/saltstack/salt/issues/10969
15:51 cro forresta: I'm not sure he will let me
15:51 chrisjones joined #salt
15:51 forresta cro, hah
15:51 BbT0n left #salt
15:52 Mr_N Hm.  Is there any way to use the -context: behavior but with generic states, and not just under file.managed?
15:52 Mr_N I may be asking this badly, I basically want to pass arguments from one state into another.
15:52 Mr_N (I realize this may be better done via pillar, I'm just experimenting)
15:53 forresta Mr_N, are these actually variables you set in state 1?
15:53 Mr_N state1 would specify certain variables that state2 would utilize.
15:53 Mr_N so yes.
15:53 repl1cant just making sure there wasn't another issue as I'm working on implimenting a multi master setup :-)
15:53 forresta Mr_N, Did you try including state 1 in state 2?
15:54 forresta repl1cant, yea that's the one
15:54 Mr_N That's not really the direction control flows in
15:54 Mr_N here, let me give context:
15:54 repl1cant forresta: cool, thnx
15:54 Mr_N Basically, I'm coming up with ways to have a state that creates a "generic" user, and you just specify the name of a user that exists in pillar, it looks up that user in pillar and gets the relevent information and generates it.
15:55 forresta repl1cant, np
15:55 GradysGhost Have there been complaints that the 2014.1.0 releast has performance issues? A colleague is complaining about it, and I just wanted to see if there was consensus on this or if he's Doing It Wrong.
15:55 Mr_N I can, for example, iterate over all the users and do that, but I'd love to be able to specify a given one from the "calling" state.
15:55 forresta GradysGhost, There are some zombie processes, not sure about the performance impact however.
15:55 forresta Mr_N, oh I see, hmm...
15:56 ajw0100 joined #salt
15:56 GradysGhost forresta: Okay. I'll have him test to see if that's what's up. Anything being done to repair that in future releases?
15:56 Mr_N forresta: unrelatedly, am I reading the issue related to multimastering failing wrong, or is it not fixed yet?
15:57 Mr_N or has the update just not been pushed
15:57 forresta GradysGhost, https://github.com/saltstack/salt/issues/10867 it's already resolved.
15:57 GradysGhost oh spiff
15:57 ze- any recommendation on how to maintain a package with debian .deb + some cherry-picks patches ?
15:57 Mr_N (less terribly worded, there's no updated version in yum, and I'm wondering if I should pull the dev version)
15:58 forresta Mr_N, nah cachedout has it fixed in 10937, won't be in till 2014.1.1
15:58 Mr_N Sigh.  Alrighty.  I had to apply a rather hacky patch to fix my multimastering in the interim.
15:58 Mr_N Looking forward to fixing that.
15:58 che-arne joined #salt
15:58 forresta ze-, you should talk to kiorky, I know his org keeps their own stable of dev and they cherry pick into it
15:58 forresta Mr_N, you could always apply cachedout's patch to the affected files
15:58 forresta Mr_N, not sure if it would break anything else, but worth a shot
15:59 rrives Mr_N I see that you had firefox fail, that is one that I did get to work, but trying again now it fails??
15:59 Mr_N I'd try that if I didn't have a workaround that "worked"
15:59 Mr_N (thanks for the pointer, regardless)
15:59 forresta np
15:59 vejdmn joined #salt
16:00 Mr_N rrives: No idea.  I pinged cro with the issue ID, so maybe we'll see it looked at at some point.
16:00 Mr_N working on user maitenance in the mean time.
16:00 hunter joined #salt
16:00 mgw has anyone seen the error "string indices must be integers, not str" with file.managed?
16:00 mgw on 2014.1.0
16:01 forresta mgw, a few people were having issues with string indices yesterday, but that was trying to create a VM with salt cloud and gitfs
16:02 mgw https://gist.github.com/mgwilliams/999c0dc4c57945c26032
16:02 mgw it's in the fileclient
16:02 mgw forresta: ^
16:03 ml_1 joined #salt
16:03 forresta mgw, gotcha, yea I haven't seen that one yet then.
16:05 mgw https://gist.github.com/mgwilliams/a9983c1915f25c4647b0 << this doesn't look right
16:05 mgw I wonder if it's something with pyobjects
16:05 utz joined #salt
16:06 mgw see how - managed seems to be part of the - template section
16:06 forresta man, very odd
16:06 forresta *mgw
16:06 forresta cannot type apparently today
16:11 kzx joined #salt
16:12 Mr_N forresta: Were you the one who always gave me the advice of "drink more"? >_>
16:12 Mr_N have you been taking your own advice? <_<
16:12 forresta Mr_N, uhh nope I don't think so
16:12 forresta though I'm at nearly 32 oz of water this morning
16:13 Mr_N One of the smart people in here has that as their default response to my headdesking about salt ><
16:13 forresta Mr_N, I think that is usually the default response to technology in general
16:14 Mr_N fairnuff.
16:15 kiorky forresta: we dont package it, we directly clone :)
16:15 Mr_N Relatedly (to drinking), despite UtahDave's approach resulting in a very elegent (to the end user of salt) method for managing users
16:15 kiorky forresta: (and we use buildout to package it)
16:15 Mr_N it results in painful as hell intermediate states...
16:16 forresta kiorky, gotcha, ze- was curious about doing something similar to what you are I believe.
16:16 tmandry joined #salt
16:16 tmandry joined #salt
16:16 kiorky (sorry im in a conf with regilero atm :))
16:18 forresta kiorky, yea no worries, it's almost time for you to go home too I believe
16:18 jalbretsen joined #salt
16:19 kolaman joined #salt
16:19 dstanek_afk joined #salt
16:20 [diecast] joined #salt
16:21 diecast joined #salt
16:21 thayne joined #salt
16:22 mgw forresta: so, salt.transport.Channel.send is returning an empty string, but a dict is expected.
16:22 forresta mgw, odd.
16:23 KyleG joined #salt
16:23 KyleG joined #salt
16:24 skynet joined #salt
16:26 ajw0100 joined #salt
16:26 joehillen joined #salt
16:27 amanuel joined #salt
16:27 hunter joined #salt
16:28 arodizzy joined #salt
16:28 mgw forresta: i think it's either an issue with my git repo, or gitfs
16:28 mgw checking into it
16:29 mgw seems to be gitfs
16:29 mgw I removed a file and replaced it with a dir in my formula
16:29 mgw basepi: have you seen anything like that?
16:30 frasergr_ joined #salt
16:30 mgw /var/cache/salt/master/gitfs/refs/master/lxc
16:30 mgw that still has the file, not the dir
16:30 mgw although the other files have been updated
16:32 budrose joined #salt
16:32 mgw removing the lxc dir from refs/master and restarting master (to force a gitfs refresh) fixed it
16:33 Ryan_Lane1 joined #salt
16:35 CaptTofu joined #salt
16:35 dstanek_afk joined #salt
16:36 ndrei joined #salt
16:36 Ryan_Lane joined #salt
16:37 Ryan_Lane1 joined #salt
16:37 basepi mgw: like what?
16:39 mgw basepi: I had a gitfs backed formula
16:39 forresta zz_Cidan, I didn't know you were such a poet: http://irclog.perlgeek.de/salt/2014-03-11#i_8417988
16:40 forresta zz_Cidan, well, such a terrible poet :P
16:40 abe_music joined #salt
16:40 mgw I removed a file in the formula and replaced it with a directory (with files in it)
16:40 mgw /var/cache/salt/master/gitfs/refs/master/lxc was half updated
16:40 GradysGhost Salt Whitman?
16:41 mgw so the init.sls referenced files in the new dir, but they weren't there
16:41 forresta GradysGhost, http://en.wikipedia.org/wiki/Walt_Whitman
16:41 forresta GradysGhost, play on words, har har
16:41 mgw it may be a separate issue, but this was the result of referencing the non-existent files: https://gist.github.com/mgwilliams/999c0dc4c57945c26032
16:41 Iwirada left #salt
16:41 vejdmn joined #salt
16:41 mgw it all cleared up when I removed the cache
16:42 GradysGhost forrest: http://en.wikipedia.org/wiki/Pun
16:44 basepi mgw: hrm, so the directory was named the same?
16:44 basepi mgw: also, what version?
16:45 forresta GradysGhost, yea exactlyu
16:45 forresta GradysGhost, I thought you were saying you didn't know who walt whitman was.
16:45 forresta so you weren't getting the pun
16:45 maveas joined #salt
16:45 GradysGhost hah, no, I'm familiar.
16:45 forresta ok :P
16:45 mgw basepi: 2014.1.0
16:46 mgw yes, the file was named etc-default
16:46 rgbkrk joined #salt
16:46 mgw and I changed that to a directory and added files to it
16:46 mgw since lxc 1.0 has multiple etc/default files rather than one
16:46 [diecast] joined #salt
16:47 basepi mgw: so obviously `data` in that traceback was probably an error message (since it was a string instead of dict)
16:48 mgw it was empty, actually
16:48 mgw I added some logging on the minion
16:48 mgw channel.send returned an empty string
16:48 mgw rather than a dict
16:48 mgw so I think it's a different issue
16:48 basepi interesting.....
16:49 mgw that became apparent from the missing file
16:50 smcquay joined #salt
16:50 kickerdog joined #salt
16:50 bhosmer joined #salt
16:51 AdamSewell joined #salt
16:51 AdamSewell joined #salt
16:51 basepi but deleting the cache solved it?
16:51 mgw yes
16:51 basepi so it's probably just a caching issue.  it's handling the conversion from file to directory incorrectly
16:51 mgw right
16:51 mgw that part of the issue
16:51 mgw but I think the tb is a separate issue
16:52 basepi i would file an issue, it will take some debugging, but i'm sure we can figure it out
16:52 basepi ya, i'm not sure what is causing that traceback
16:52 basepi but i don't have time at the moment to debug either.  =P
16:53 philipsd6 joined #salt
16:54 hunter joined #salt
16:55 bhosmer_ joined #salt
16:56 ajw0100 joined #salt
16:58 mgw basepi: same here, unfortunately
16:58 mgw I'll open an issue
16:58 basepi At least we have a pretty straightforward workaround with the cache deletion.
16:58 mgw I usually like to put together a test scenario before opening an issue like this
16:59 mgw to be sure it's reproducable
16:59 schimmy joined #salt
17:01 xmltok joined #salt
17:02 seapasulli joined #salt
17:07 testin joined #salt
17:08 novastar2 joined #salt
17:08 kickerdog joined #salt
17:10 quickdry21 joined #salt
17:12 timoguin joined #salt
17:12 schimmy joined #salt
17:13 bastion1704 joined #salt
17:16 saurabhs joined #salt
17:16 ch0pstick joined #salt
17:17 naiyte joined #salt
17:19 ipalreadytaken joined #salt
17:20 kzx joined #salt
17:21 ch0pstick joined #salt
17:22 ch0pstick joined #salt
17:24 ch0pstick joined #salt
17:24 ch0pstick joined #salt
17:25 ch0pstick joined #salt
17:25 thayne joined #salt
17:25 ch0pstick joined #salt
17:27 [diecast] joined #salt
17:28 kzx Mr_N: I tried another 2 Windows 7 64 bit machines and package install worked, so there is something about that one machine, it looks like WMI (this management facility on windows)  was messed up.
17:28 ch0psti71 joined #salt
17:29 kzx Mr_N: anyway just wanted to report my findings and thank you again for the help again
17:29 allanparsons joined #salt
17:30 ch0pstick joined #salt
17:31 ch0pstick joined #salt
17:32 smcquay_ joined #salt
17:32 ch0pstick joined #salt
17:32 ch0pstick joined #salt
17:32 amckinley joined #salt
17:33 KyleG1 joined #salt
17:33 ch0pstick joined #salt
17:33 ch0pstick joined #salt
17:34 ch0pstick joined #salt
17:34 utz1 joined #salt
17:35 smcquay joined #salt
17:35 utz1 left #salt
17:36 ch0pstick joined #salt
17:37 Networkn3rd joined #salt
17:37 ch0pstick joined #salt
17:37 rlarkin joined #salt
17:37 ch0pstick joined #salt
17:41 rlarkin Hi
17:41 rlarkin http://pastebin.com/MR1VxFTh
17:41 rlarkin I'm stuck
17:41 ch0pstick joined #salt
17:43 mr_chris Running salt-minion on centos from epel. Master is 0.17.4. Just upgraded a minion to 2014.1.0-1 from epel (is that 0.17.5?) and now the highstate fails with "No Top file or external nodes data matches found"
17:43 mr_chris Of course now I find out that yum cache was turned off and I can't downgrade.
17:45 xerxas joined #salt
17:45 mtford90 joined #salt
17:46 ch0psti31 joined #salt
17:47 Ahlee if it was just one minion, yum remove, yum install from vault.centos.org
17:47 Ahlee always, always, always upgrade the master first
17:47 ch0psti59 joined #salt
17:49 ch0psti30 joined #salt
17:50 mgw mr_chris: usually it doesn't work too well to run a newer minion than master
17:50 mr_chris Ahlee: Lesson learned.
17:50 ch0psti70 joined #salt
17:50 mgw sorry, jus saw that Ahlee already answered that
17:50 mr_chris Ahlee: Vault.centos.org. That will have previous versions from epel?
17:50 Ahlee oh
17:50 Ahlee whoops, sorry
17:50 Ahlee mr_chris: http://koji.fedoraproject.org/koji/packageinfo?packageID=13129
17:50 Ahlee you can get the old versions there
17:51 sashka_ua guys, suggest the way to print some message in state output to indicate some error or condition?
17:51 Fin1te joined #salt
17:51 mr_chris Ahlee: Nice! Thanks!
17:54 n8n joined #salt
17:54 forresta Ahlee, that's fantastic, a bunch of people were asking for those.
17:55 mgw1 joined #salt
17:55 ml_1 joined #salt
17:56 mtford joined #salt
17:57 Ahlee hopefully this teaches people to mirror locally ;)
17:57 ch0psti49 joined #salt
17:58 ch0psti30 joined #salt
17:58 baniir joined #salt
17:58 mr_chris Ahlee: We will be now.
17:58 Ahlee but, i'm old and crotchity and i don't trust other peoples infrastructure
17:58 forresta lol
17:58 ch0psti97 joined #salt
17:58 forresta 'back in my day' Ahlee
17:58 forresta ?
18:00 ch0psti89 joined #salt
18:00 * xmj gives Ahlee tinfoil hat
18:01 arthabaska joined #salt
18:01 R_Shackleford joined #salt
18:01 grep_awesome joined #salt
18:03 Ahlee eh, it's not so much a tinfoil hat, so much as I just don't like black boxes outside of my control
18:03 JasonSwindle joined #salt
18:03 JasonSwindle left #salt
18:04 mgw joined #salt
18:05 wkf_ joined #salt
18:07 darien joined #salt
18:07 ch0pstick joined #salt
18:07 darien Hey guys, I couldn't find this in the docs, so… is it possible to have the master log any changes (i.e. the normal output) from doing highstate updates (or, optimally, all activities) ?
18:08 ch0pstick joined #salt
18:08 Ahlee darien: well, it does.  By default it does to the file system and maintains a 24 hour history
18:08 darien oh
18:08 Ahlee from your master
18:08 Ahlee salt-run jobs.list_jobs
18:08 darien really? I couldn't find it, I'll look again
18:09 Ahlee then salt-run jobs.lookup_jid <job id>
18:09 Ahlee for long term, look into setting up an ext_job_cache
18:09 darien well f
18:09 darien thanks!
18:10 baniir joined #salt
18:10 Ahlee You're welcome
18:12 ch0pstick joined #salt
18:13 ch0pstick joined #salt
18:16 zain joined #salt
18:17 higgs001 joined #salt
18:17 tharkun joined #salt
18:17 tmandry joined #salt
18:17 jY anyone have a way to store encrypted data like db passwords in salt like a pillar for minions to use
18:17 ch0psti67 joined #salt
18:18 bram_ left #salt
18:19 mgw jY: where would you want the decryption keys stored?
18:21 jY mgw: ya that's the irony my boss wants them encoded.. i'm telling him to just use a private repo
18:21 baniir is there a way to escape a # in a pillar value
18:21 hunter joined #salt
18:22 ch0pstick joined #salt
18:23 harobed joined #salt
18:23 druonysus joined #salt
18:24 ch0pstick joined #salt
18:24 forresta baniir, did you try to pipe it through the e filter? so {{ myval|e }} ?
18:25 forresta baniir, I know that's supposed to work for html values, I don't know if it does so for #
18:26 baniir forresta: unfortunately that didn't work
18:26 forresta baniir, ok, what context are you using it in?
18:27 nix_ joined #salt
18:28 baniir forresta: defined in pillar then passed to a jinja template via state defaults
18:29 forresta baniir, ok, and did quoting the variable not work?
18:29 forresta baniir, whatever is supposed to be the variable in pillar that is
18:29 mgw jY: ask him where he wants the decryption key… if he wants it on the master, well, what's the point? If he wants it on the minion, well, what's the point? (The master has access to the minion's root file system.)
18:30 mgw jY: I actually spent a lot of energy on this issue.
18:30 mgw The truth is, if someone roots your master, you're hosed.
18:30 jY ya that's what i've been trying to explain
18:30 jY thanks
18:30 forresta baniir, or maybe try the standard \ http://docs.saltstack.com/topics/troubleshooting/yaml_idiosyncrasies.html#yaml-support-only-plain-ascii
18:30 forresta if you haven't done so already
18:30 mgw A more important thing is to ensure you don't have your passwords getting logged in the minion log
18:31 mgw Or at least ensure your minion log is root-readable only
18:31 forresta mgw, I thought that was only supposed to happen when you had debugging turned on
18:31 forresta there was an issue for this a while bak
18:31 forresta *back
18:31 n8n joined #salt
18:31 baniir forresta: i had quotes around the value in pillar but not in my state. i just tried quotes in the state, and it works. looks like i was missing that middle step. thanks for talking this out
18:32 forresta baniir, yea np
18:32 ch0pstick joined #salt
18:34 Teknix joined #salt
18:35 mpoole hey I remember talking to someone at saltconf about an alternative to hiera GPG
18:35 mpoole that doesn't require an entire file decrypt every time you look up a variable
18:35 baniir is there a way to pass info from a state to a jinja template without potentially conflicting with other keys in the state, e.g. a variable named 'name'
18:36 mpoole anyone know what it is? the name escapes me
18:36 kickerdog joined #salt
18:37 smcquay joined #salt
18:38 gnugnu9 joined #salt
18:39 bhosmer joined #salt
18:39 utz joined #salt
18:45 forresta mpoole, hmm, I'm trying to think of people other than Ryan_Lane who I know are using hiera
18:45 darless_ joined #salt
18:45 ch0psti42 joined #salt
18:51 conan_the_destro joined #salt
18:55 smcquay joined #salt
18:56 mikeywaites joined #salt
18:57 mikeywaites hey guys is it possible to call states  from inside top.sls directly IE postgresql.databases opposed to postgresql
18:58 timoguin mikeywaites, yes
18:59 bhosmer joined #salt
18:59 mikeywaites awesome guess i missed it in the docs
18:59 diegows does salt-minion in masterless mode support gitfs?
18:59 vejdmn joined #salt
18:59 mikeywaites timoguin:  oh damn, i had a bloody typo
18:59 mikeywaites cheers!
19:00 timoguin :)
19:00 timoguin diegows, right now i think it doesn't. pretty sure there's an open issue for that though
19:00 diegows timoguin, yes, I think the same :)
19:00 diegows ok, workaround :)
19:01 timoguin is there some secret i'm missing to searching for issues on github's issue tracker?
19:02 timoguin nevermind i found the search bar at the top
19:02 bhosmer joined #salt
19:02 timoguin diegows, issue right hurr: https://github.com/saltstack/salt/issues/6660
19:03 diegows timoguin, thanks!
19:04 ch0pstick joined #salt
19:05 rrives joined #salt
19:06 rawzone joined #salt
19:06 smcquay_ joined #salt
19:07 ipmb joined #salt
19:07 ndrei joined #salt
19:08 rrives left #salt
19:08 rrives joined #salt
19:08 eunuchsocket joined #salt
19:09 KyleG joined #salt
19:09 KyleG joined #salt
19:09 johtso joined #salt
19:13 ch0psti83 joined #salt
19:14 R_Shackleford joined #salt
19:16 bastion1704 joined #salt
19:20 hunter joined #salt
19:21 ipalreadytaken joined #salt
19:21 btorch is there a salt command to flush any cache  on the master ? I'm on 0.17.5 and it's acting up weird
19:21 allanparsons joined #salt
19:22 allanparsons @ baniir try: "context"
19:24 Ahlee btorch: https://github.com/saltstack/salt/blob/develop/salt/runners/fileserver.py
19:24 btorch for some reason the master is not creating a /var/cache/salt/master/roots/hash
19:24 Ahlee btorch: salt-run fileserver.update
19:25 btorch that didn't make any difference
19:25 allanparsons i'm building a redis master-slave server.  when the master comes up, how can i return the master's IP to the slave so I can write the "SLAVEOF" in the slave's conf file?
19:25 topochan joined #salt
19:25 btorch it's still missing the md5 of a new file that exists in the root base
19:25 allanparsons I looked at Salt Mine but the docs are lacking.  Or would salt.modules.publish be a better alternative?
19:26 Ahlee btorch: just roots for fileserver_backend?
19:27 dave_den allanparsons: http://docs.saltstack.com/topics/reactor/index.html#reactor-system
19:27 Ahlee btorch: and there's no hash directory, or no hash for the file
19:27 Ahlee does cp.list_master work at all?
19:27 kzx anyone know of an example of how to launch an application on windows, it seems if I do cmd.run 'notepad.exe' the notepad app never starts and the command blocks, if Ctrl+C it seems to create a long running job
19:29 btorch Ahlee: the directory /var/cache/salt/master/roots/hash/base/etc/syslog-ng/v3.3 exists and there are 3 md5 files there but the 4th one which is new doesn't
19:29 dave_den kzx: i'm not a windows user, but you probably need to use powershell to launch it properly.
19:29 btorch Ahlee: I assume that is why it's not updating the syslog.conf .. well it sends an empty file
19:29 btorch if I remove it on the server of course
19:30 btorch Ahlee: salt-run cp.list_master ?
19:30 Ahlee btorch: if you target the minion, salt-run cp.list_master, is there a file present?
19:31 Ahlee er'
19:31 Ahlee damnit
19:31 Ahlee salt minion cp.list_master
19:31 Ahlee sorry. half paying attention, half not.
19:31 btorch yeah it shows up
19:32 anitak1 joined #salt
19:32 Ahlee btorch: and salt minion cp.get_file salt://path/to/file /tmp/foo; /tmp/foo i'm assuming will be 0 length?
19:33 btorch Ahlee: nope it's the correct content
19:33 Ahlee oh that's just weird then.
19:33 btorch :) yep
19:33 rysch left #salt
19:34 Ahlee the minion in question, it doesn't have environment: set, does it?
19:34 Ahlee like, is it possibly going to non-base for the file?
19:34 btorch no
19:34 anitakrueger guys…quick question if I may…I have a salt-master with 10 threads (used to be 5) and it's using 100% of CPU :(
19:34 anitakrueger I can't run any commands on any of the attached minions (13 in total)
19:34 tr_h joined #salt
19:34 hunter joined #salt
19:34 anitakrueger the minions basically say they are waiting to be authenticated, but they have been fine before
19:35 btorch Ahlee: I don't know I've been able to salt a few boxes ... but I had noticed this strange behavior before but I never logged into it til now
19:35 anitakrueger I think I've traced it to a problem with zmq...
19:35 anitakrueger I just can't pinpoint what it is :(
19:35 hunter joined #salt
19:36 anitakrueger been also looking at what happened when it started (around 6am Central this morning)
19:36 anitakrueger but of course there is nothing apparent in any logs
19:37 mtford left #salt
19:37 dave_den anitakrueger: are all of the threads maxing cpu?
19:37 btorch lol I found the issue
19:37 dave_den anitakrueger: have you restarted the master?
19:37 Ahlee anitakrueger: what fileserver backend?
19:37 Ahlee anitakrueger: also, debug?
19:37 dave_den anitakrueger: have you straced a high cpu thread?
19:38 ch0pstick joined #salt
19:38 higgs001 joined #salt
19:39 anitakrueger restarted the master several times…increased ulimit, rebooted the box
19:39 anitakrueger fileserver backend is just roots
19:39 anitakrueger lot's of these in the debug log:
19:39 anitakrueger 2014-03-12 19:38:19,921 [salt.utils.event                         ][DEBUG   ] LocalClientEvent PUB socket URI: ipc:///var/run/salt/master/master_event_pub.ipc
19:39 anitakrueger 2014-03-12 19:38:19,971 [salt.utils.event                         ][DEBUG   ] LocalClientEvent PULL socket URI: ipc:///var/run/salt/master/master_event_pull.ipc
19:40 eunuchsocket left #salt
19:42 dave_den anitakrueger: did you upgrade your master or minions recently?  also, did this start when you increased your master_threads? what else changed recently?
19:42 kickerdog joined #salt
19:43 nkuttler joined #salt
19:45 anitakrueger I've been trying to figure out what changed this morning at 6am :(
19:45 anitakrueger master and minions are on 2014.1
19:45 anitakrueger actually…I did upgrade the rest of the minions to 2014.1 last night...
19:45 anitakrueger master has been running on it for a week or so
19:46 pass_by_value anitakrueger: what is shown by salt-key -L, manage.present and manage.status? Do the outputs match?
19:46 anitakrueger let me check
19:48 Ahlee btorch: oh?  what was it?
19:49 dave_den there's this unresolved high cpu problem, but not sure if it's related: https://github.com/saltstack/salt/issues/8295
19:50 tmandry joined #salt
19:51 allanparsons dave_den:  I'm reading about the reactor system... but, a few things:  1) I may not have a slave come online (who knows?)  and 2) it may take a few minutes for a slave to come online.  i need for the slave to query info from the salt-master accepted keys and then get grain data from the redis-master (ip address)
19:51 Ahlee i still see that issue in 0.17.5
19:51 anitakrueger salt runner is horribly slow
19:51 allanparsons is there anyway i can make the redis-master pillar + grain data available to other nodes so they can use it in their jinja templates?
19:51 anitakrueger still waiting for the output
19:51 Ahlee 8295, that is.
19:52 Ahlee anitakrueger: do you happen to have a _lot_ of jobs?
19:52 [diecast] joined #salt
19:52 Ahlee sorry, in your cache
19:52 Ahlee oh, nevermind
19:52 Ahlee that shouldn't affect salt-run, just jobs.*
19:53 dave_den allanparsons: the reactor reacts to events, and the minion startup is just an example of one event. you can fire events at any time using the event module.
19:54 FarrisG Is there a way to set tags on an ec2 isntance at deployment using salt-cloud? I can only find docs on how to set them on an existing instance.
19:54 anitakrueger @dave_den yea, I saw that github issue
19:54 allanparsons @FarrisG - use "salt-call saltutil.sync_grains"
19:54 allanparsons you'll need a _grains directory in your states dir.
19:55 allanparsons FarrisG - https://github.com/saltstack/salt-contrib/tree/master/grains
19:55 hank00 joined #salt
19:55 anitakrueger @pass_by_value no, outputs don't match between salt-key -L and manage.present
19:55 allanparsons we do exactly that now.
19:55 utz joined #salt
19:55 dave_den allanparsons: the only way to make the redis-master pillar data available to other minions is to use a runner+masterPillarUtil, or structure you pillars in a way that that info is targeted to the minion who need it
19:56 dave_den or create a custom ext_pillar
19:56 allanparsons yeah, my minions will be named:  {env}-redismaster-{0-9}  or {env}-redisslave-{0-9}
19:56 anitakrueger @Ahlee the highstates I run are quite comprehensive, but we do not run them on a periodic schedule or anything
19:57 FarrisG allanparsons: Not sure I'm following. If we want dyanmic ec2 tags when we deploy with salt-cloud, do we have to put some logic in the grains to set them when sync_grains is called? Can they be defined in the salt-cloud profiles?
19:57 ch0psti62 joined #salt
19:57 allanparsons do you want to fetch ec2 tags? or set ec2 tags?
19:58 Ahlee anitakrueger: hrm.  I'm out of ideas, then.  Sorry =\  Good luck.
19:59 allanparsons FarrisG: do you want to fetch ec2 tags? or set ec2 tags?
19:59 pass_by_value One thing to try might be to see what events are being triggered using https://github.com/saltstack/salt/blob/develop/tests/eventlisten.py anitakreuger.
19:59 mikeywaites joined #salt
19:59 pass_by_value er anitakrueger
20:01 anitakrueger @pass_by_value that looks handy
20:01 anitakrueger let me see if I can make that work
20:01 pass_by_value It is a standalone script so you can just invoke it using python eventlisten.py from the tests folder in salt's source code. It doesn't have to be from site packages, you can clone the repo and run it directly.
20:02 hunter joined #salt
20:02 FarrisG allanparsons: I want to set them, at deployment.
20:02 ch0pstick joined #salt
20:02 allanparsons in your cloud.profiles file:  tag: {'Environment': 'production', 'Role': 'elasticsearch'}
20:03 FarrisG Ahh, that's what I was looking for. Is that in a doc?
20:03 allanparsons i dont think so.
20:03 allanparsons i was just screwing around and figured it out.
20:03 allanparsons it may be in a bug filed
20:04 FarrisG Cool, I'll check. Thanks!
20:04 allanparsons you can also set them using a userdata script (script:  /etc/salt/cloud.deploy.d/user_data.sh)... use an IAM Role to allow ec2:*tags and then use the aws unified cli tool
20:04 druonysuse joined #salt
20:04 timoguin I don't see an issue for that when searching for 'aws tags'
20:04 allanparsons we do that to dynamically set hostnames when a server launches  (Tag: Name)
20:05 allanparsons @timmoguin - lemme find.  i dont think it's documented.
20:05 timoguin if there's not an issue we should create one and get that jazz in the docs!
20:06 allanparsons 1 sec... i'l add
20:06 londo_ joined #salt
20:07 ipmb joined #salt
20:09 tr_h joined #salt
20:10 budrose joined #salt
20:14 Ahlee frack.  No way to specify the path to virtualenv for salt.states.virtualenv
20:15 dave_den Ahlee: ?
20:15 dave_den 'name' is the path
20:15 Ahlee dave_den: I install my own python version for my code in /opt/python, but use system python for salt
20:15 obimod joined #salt
20:15 Ahlee so i want to create my venv from /opt/python-2.7.3/bin/virtualenv
20:15 Ahlee not /usr/bin/virtualenv
20:15 dave_den ah, i was gonna say
20:16 dave_den you crazy
20:16 dave_den but yeah i think you're right
20:16 txmoose Afternoon, all.  I've got a quick issue to discuss.  Is anyone seeing an "Unable to bind to socket, error" when using salt-ssh?  How can I clear this?
20:16 txmoose salt version 2014.1.0
20:16 dave_den Ahlee: actually, you can use venv_bin
20:17 dave_den and you can also specify the path to python
20:17 ch0pstick joined #salt
20:17 Ahlee dave_den: ah, so that's what python=python does =\
20:17 Ahlee and venv_bin=None
20:17 Ahlee god i suck at reading
20:18 dave_den yeah. not documented except in the kwargs list
20:19 Ahlee thanks dave_den
20:19 dave_den np
20:20 saurabhs joined #salt
20:22 allanparsons timoguin: added a pull request.  please approve for that documentation fix.  https://github.com/saltstack/salt-cloud/pull/907
20:22 allanparsons i also added some documentation for VPCs + auto assigning an public IP (not EIP) address at public-subnet launch
20:23 vejdmn joined #salt
20:25 txmoose What port does salt-ssh attempt to run on?
20:27 bhosmer_ joined #salt
20:28 Ahlee txmoose: salt-ssh is for connecting out, I don't believe it binds to any ports
20:28 yomilk joined #salt
20:29 txmoose Ahlee: When I run salt-ssh to anything, I get an error unable to bind to port
20:29 txmoose That's what I thought as well, but it's attepmting to bind to something
20:29 Ahlee well that's odd.
20:30 txmoose Ahlee: Here's a sample output:
20:30 txmoose archon:salt # salt-ssh 'breakfix' pkg.install vim
20:30 txmoose [WARNING ] Unable to bind socket, error: [Errno 98] Address already in use
20:30 txmoose breakfix: ----------
20:30 txmoose archon:salt #
20:30 smcquay joined #salt
20:30 txmoose latest version of salt
20:30 Ahlee maybe it's trying to connect to the zeromq 'sockets' ?
20:31 Ahlee s/connect/bind
20:31 txmoose that would make sense... why would it be unable to, though?
20:31 Ahlee permissions on the socket file, maybe?  Not running?
20:31 Ahlee though, for it supposed to be a replacement to zmq, kind of strange
20:32 schimmy joined #salt
20:32 txmoose Idk... I'm unable to use salt-ssh at all right now... nothing seems to be working
20:32 txmoose password is correct in my roster
20:33 kzx dave_den: thanks for the powershell suggestion from earlier, the process gets spawned and run but it is in a hidden (invisible) session
20:34 timoguin allanparsons, assuming you saw techhat's comment. i don't have permissions to commit.
20:34 allanparsons timoguin: yeah.  i just forked salt.  see: https://github.com/saltstack/salt/pull/11183
20:35 kzx salt '*' cmd.run "Start -wait C:/windows/system32/notepad.exe" shell='powershell' blocks and I see process in Task Manager but it doesn't show on the screen, so it is half way there
20:35 dave_den kzx: you probably need to specify parameters to get it to display on the desktop
20:35 dave_den but like i said, i'm not a windows user so i am just guessing :)
20:35 kzx dave_den: that is what I am thinking, there is Session ID associated with it so I am looking at that
20:36 kzx dave_den: Yeah I don't want to be a windows user either, that is why I am trying to automate it ;-)
20:36 ch0psti10 joined #salt
20:37 viod joined #salt
20:39 darien txmoose: have you tried using strace to figure out what port it's trying to bind and why?
20:40 ch0pstick joined #salt
20:40 whidbeywalker joined #salt
20:40 txmoose darien: No, I hadn't gone that far yet.  If no one in here knew off hand (or if it was trying to grab a very high numbered port) then I'd move on to strace
20:41 txmoose honestly, was trying to avoid strace :P
20:41 funzo joined #salt
20:41 darien txmoose: outgoing connections should have an unused outgoing port picked for them by the OS, so you theoretically shouldn't get that error
20:41 darien txmoose: unless it's trying to bind a local unix socket (e.g. /var/run/salt.minion or something)
20:41 higgs001 joined #salt
20:42 ch0psti57 joined #salt
20:43 tlyng joined #salt
20:43 jsm joined #salt
20:44 jsm left #salt
20:44 oraqol joined #salt
20:44 tlyng Is there any repository for creating virtual machines that ease the process of developing salt states? I think I encountered something like that earlier, can't remember where :/
20:44 manicouman joined #salt
20:46 oraqol Hi guys, really weird behavior here, hoping to get some help if possible: when I run salt-minion --version, one of my minions shows 2014.1.0 while the other shows 2014.1.0.  Both ubuntu 10.04, both pointed to the same official repo
20:46 txmoose tlyng: might this be what you're looking for? http://saltstarters.org/
20:46 oraqol tried reinstalling 2014.1.0, but wont upgrade to  0.17.5
20:46 tlyng txmoose: no, it was something to generate a couple of virtual machines. One master and two slaves if I remembered correctly
20:46 oraqol sorry, second minion was meant to be  0.17.5
20:47 txmoose tlyng: Ahh sorry, bud, I haven't seen anything like that, but if/when you do find it, could you dump it in channel here? O:)
20:48 tlyng will try to remember
20:52 ch0pstick joined #salt
20:56 ch0pstick joined #salt
20:56 chrisjones joined #salt
20:57 ndrei_ joined #salt
20:57 clearclaw joined #salt
20:58 tlyng txmoose: i believe i found it, https://github.com/elasticdog/salt-sandbox
20:58 clearclaw 2014.1 minions seem to have changed the order of execution of states listed in top.sls...
20:59 oraqol yes, having the same issue with 2014.1, wanting to upgrade to 0.17.5, but apt-get install only reinstalls 2014.1.  What do?
21:00 dave_den debian.saltstack.com
21:02 saurabhs joined #salt
21:02 ndrei joined #salt
21:03 ipalreadytaken joined #salt
21:03 manicouman joined #salt
21:03 oraqol So if I use the debian repo for ubuntu 10.04 to get version 0.17.5?
21:04 vejdmn joined #salt
21:05 kzx anyone know how to run a command once when minion connects, I need to reconfigure its service parameters (this is for windows) and then restart it ?
21:05 jesusaurus oraqol: do you mean downgrade to 0.17.5?
21:06 oraqol HA
21:06 oraqol yes
21:06 clearclaw All my F'ing deploy scripts are now broken because the assumed that top.sls would br processed from the top down, and instead it is getting some F'ing odd order....
21:06 oraqol wait, so do I want to downgrade?  right now using 2014.1 I get the following error when running highstate on minions with parametrized pillar data that is correctly configured:
21:07 * clearclaw is ticked, but mildly happier that this is not a PROD push day.
21:07 dave_den kzx: starttup_states
21:07 oraqol State: - no
21:07 oraqol Name:      states
21:07 oraqol Function:  None
21:07 oraqol Result:    False
21:07 oraqol Comment:   No Top file or external nodes data matches found
21:07 oraqol I think its the same issue as clear claw is seeing
21:07 dave_den oraqol: is your master still 0.17.5 and minion 2014.1.0?
21:08 dave_den your master needs to be >= to your minion version.
21:08 oraqol just got that
21:08 oraqol im an idiot
21:08 kzx dave_den: thank you!
21:08 zain joined #salt
21:08 oraqol I didnt realize 2014.1 was newer, new naming convention and all
21:08 oraqol mea culpa
21:09 kickerdog1 joined #salt
21:09 faldridge joined #salt
21:10 n8n joined #salt
21:10 andr386 joined #salt
21:10 oraqol left #salt
21:11 ch0psti66 joined #salt
21:11 rojem joined #salt
21:13 FarrisG Any good ideas on how best to do salt-cloud post-deploy tasks that might not easily (or at least immediately) fit into a state? For instance, after salt-cloud successfully deploys a host, I want to ssh to that host and change the hostname, then run a command to update my internal DNS with a record for the host. I'd like to have someway to fire that off easily or automagically after a successful salt-cloud deploy.
21:16 allanparsons @FarrisG email me.  I wrote exactly what you want to do in python.
21:16 allanparsons meh, fuck it.  i'll gist it
21:16 ch0pstick joined #salt
21:18 hunter joined #salt
21:20 FarrisG allanparsons: Don't think I have your email. But if you wanna gist it, that would be rad, too.
21:21 allanparsons https://gist.github.com/allanparsons/4ce9645e74c1a5ddda6a
21:22 allanparsons you'll need to modify that user-data.sh script and replace anything in [ XXXX ]
21:22 aleszoulek joined #salt
21:22 allanparsons and, you'll need an IAM role with permission to route53 + ec2:describe* and createtags
21:23 FarrisG allanparsons: That is indeed rad, except that we're not using r53 for this part. Definitely will help out though, thanks!
21:23 allanparsons just use route53
21:23 allanparsons we run a nondescript domain for our EC2 instances
21:24 FarrisG allanparsons: Believe me, I'd love to. Just can't at the moment.
21:24 txmoose Ahlee: Got another issue.  Requiring a package in a sls file.  Running state.sls doesn't find the package, but running pkg.install DOES install it.  What am I doing wrong?
21:24 allanparsons we register those in route53 so we dont attach to salt as "ip-X-Y-Z" and so we can easily ssh to the FQDN that show up in our logs
21:24 higgs001 joined #salt
21:25 Gordonz joined #salt
21:26 FarrisG Yeah, that's where I want to be. salt-cloud handles the key name bit for us, but since I still have to use internal bind I need to update that, and for other reasons I need to make the host's hostname match.
21:26 clearclaw allanparsons: I use some hack extensions to ec2-cli-tools for that SSH etc stuff.
21:26 Gordonz joined #salt
21:26 clearclaw Still need to do the bits to cover scp...
21:27 allanparsons oh, yeah.  i just prefer to make entries in DNS at server boot.
21:27 clearclaw Doing full DNS control is a bit less interesting when all the bits are inside a VPC.
21:27 jslatts joined #salt
21:27 allanparsons and remove entries from DNS at server shutdown/termination
21:27 clearclaw With a reactor?
21:27 allanparsons we use a public/private VPC
21:28 * clearclaw nods
21:28 allanparsons not with reactor - i literally just started reading up on reactor about 15 m ago
21:28 clearclaw It is mildly scary shit.
21:28 allanparsons if we find a public hostname, we insert that into a CNAME
21:29 allanparsons if we dont find a public hostname, we insert the private hostname (.internal or whatever is) into dns as a CNAME.
21:29 vlcn afternoon guys
21:30 alunduil joined #salt
21:30 vlcn I need to determine how to use event.fire_master from within a state I am working on which uses the python renderer
21:31 ch0psti60 joined #salt
21:32 cewood joined #salt
21:35 ch0pstick joined #salt
21:37 kzx another windows question, would it be possible to run salt-minion in user's session as an autorun service, so it doesn't start until a user logs in then it runs in that user's session
21:37 allanparsons man, super sorry, but i'm not seeing how the reactor system would benefit me to have a redis-slave autodiscover a redis-master, get the redis-master IP, and then modify my redis-slave conf file to include "SAVEOF {{ grains['redis-master-ip'] }}
21:38 allanparsons if there's an example for a similar master/slave system, can you point me to it?
21:38 [diecast] joined #salt
21:39 hunter http://pastebin.com/GfQ7fVg6
21:40 hunter i'm getting a salt-ssh failure when I specifiy an env to pull the formula from.
21:40 hunter If I leave off the env=dev, it works.
21:41 dave_den hunter: try saltenv=dev
21:41 saurabhs joined #salt
21:41 hunter dave_den: even for 0.17.5? (Sorry, I should have said what version I'm using)
21:41 dave_den ah, no. env for 0.17.5
21:42 hunter dave_den: exact same error & hang in any case.
21:42 Ryan_Lane RAWR: https://github.com/mitchellh/vagrant/blob/master/plugins/provisioners/salt/bootstrap-salt.sh
21:42 Ryan_Lane that really badly makes me want to stab
21:42 Ryan_Lane can't blame them. it's still in the instructions
21:42 gnugnu9_ joined #salt
21:43 Ryan_Lane curl from http into shell is so incredibly insecure.
21:43 arthabaska that is hideous Ryan_Lane
21:44 txmoose Has anybody seen an issue where the pkg.installed line in a sls file is unable to install a package, but salt 'target' pkg.install 'pkg' works just fine?
21:44 arthabaska the convenience is helpful, but the calling script surely won't mind doing it the safe way
21:45 Ryan_Lane it's totally unacceptable that it's http
21:46 gnugnu9_ joined #salt
21:46 arthabaska at the very least haha
21:47 hunter dave_den: any suggestions? This is a blocker for me.
21:47 Ryan_Lane bleh. I can't find the open issue for this
21:47 dave_den hunter: salt-ssh is super alpha and i haven't played with envs on salt-ssh. it may not be supported yet
21:48 ch0psti27 joined #salt
21:48 dave_den bbl
21:48 hunter shucks - it _was_ working. What the bleep did I change.
21:58 ch0pstick joined #salt
22:03 kickerdog1 left #salt
22:06 yomilk joined #salt
22:08 manicouman joined #salt
22:09 anitak joined #salt
22:10 younqcass joined #salt
22:10 anitak1 joined #salt
22:12 allanparsons test
22:14 mikeywaites Anyone know if its possible to provide the pip state a path to a requirements file on the file system opposed to a file managed by salt?
22:15 fllr joined #salt
22:15 bhosmer joined #salt
22:16 yomilk joined #salt
22:16 allanparsons pip.installed....
22:16 allanparsons - requirements: {{ pillar['code_dir'] }}/requirements.txt
22:17 mikeywaites allanparsons:  legend thanks just need to figure out how to apply that to a virtualenv previously created now
22:17 allanparsons above that requirements, add:  - bin_env: /home/{{ pillar['user'] }}/.virtualenvs/{{ pillar['virtualenv_name'] }}
22:17 nocturn joined #salt
22:18 mikeywaites Man i love salt
22:18 mikeywaites :D
22:18 mikeywaites thanks
22:19 allanparsons np
22:20 seapasulli left #salt
22:26 vlcn sang idea what would cause me to not see any events from specific minions when using  salt-call event.fire_master?
22:26 vlcn any*
22:28 vlcn interesting, restarting the minion solved the issue
22:30 ch0pstick joined #salt
22:30 quickdry21 joined #salt
22:31 mgw joined #salt
22:33 faldridge joined #salt
22:34 baniir joined #salt
22:36 jergerber joined #salt
22:37 hunter ok, I've horked something - when I use "salt MINION state.sls yum env=dev" NOTHING happens - I get Total:0 in the summary.
22:37 hunter pillar.items shows the correct info for tha minion and state.show_top shows the correct top info for that minion.
22:37 hunter But _nothing_ happens and -l trace doesn't yield any insight.
22:38 hunter _help_
22:41 ch0pstick joined #salt
22:45 hunter DAMNIT - nevermind
22:45 hunter I missed a top level namespace issue in pillar
22:47 hunter I hate the emacs YAML mode
22:54 anitak joined #salt
22:56 Daviey joined #salt
22:58 higgs001 joined #salt
23:00 higgs001_ joined #salt
23:01 robawt just double checking here:  you can't have pillar in pillar
23:01 robawt the correct thing to do would be a single jinja variable inside the pillar that gets rendered
23:01 robawt ?
23:02 forresta robawt, what do you mean pillar in pillar?
23:02 forresta like a pillar variable inside another pillar variable?
23:03 robawt yeah
23:03 forresta I don't know if that can work since it all gets rendered at the same time
23:03 forresta I've never tried it
23:03 robawt that's what I was thinking
23:03 forresta you should try it :P
23:03 robawt ack
23:03 robawt I'm not a test dummy :)
23:03 forresta psssh someone has to be!
23:03 forresta you came up with the idea, you get to test it
23:03 robawt I think the jinja variable seems like a more elegant solution
23:03 robawt ha
23:03 robawt I like to think none of my crazy ideas are original
23:05 forresta Corey, are you going to PyCon?
23:06 robawt forresta: if you're going to PyCon I need to get you that beer I still owe you
23:06 forresta robawt, yea I'll be there, as soon as I can stomach paying for the airfare and finding a hotel :P
23:06 lude1 joined #salt
23:07 forresta might go airbnb on this one
23:14 ajw0100 joined #salt
23:15 vejdmn joined #salt
23:20 elfixit1 joined #salt
23:20 allanparsons i've enabled peer settings.
23:20 allanparsons but when i run publish.publish from a minion
23:21 allanparsons i get a timeout
23:21 tlyng I've just created a salt sandbox with a master and two minions based on ubuntu 12.04 precise pangolin. Look at https://github.com/tlyng/salt-sandbox if you want to get up and running with salt in no-time. I'm going to use it for developing state trees, pillars and salt modules.
23:21 allanparsons [WARNING ] SaltReqTimeoutError: Waited 3 seconds
23:21 allanparsons Minion failed to authenticate with the master, has the minion key been accepted?
23:23 allanparsons salt * test.ping returns: No minions matched the target. No command was sent, no jid was assigned.
23:24 allanparsons my minions are running
23:24 allanparsons and i'm showing them as accepted in salt-key -L
23:25 allanparsons and when i run salt-call test.ping from the minion, i get True back
23:27 arthabaska allanparsons can any minion reach the master? telnet to port 4505 to see if you can connect
23:27 saurabhs joined #salt
23:27 allanparsons minions can reach the master
23:27 allanparsons i just rebooted the master
23:27 allanparsons 1 sec
23:28 allanparsons lol - reboot fixed it.
23:28 allanparsons weird
23:28 arthabaska huh, weird indeed
23:31 fllr joined #salt
23:34 Corey forrest: Wasn't planning on it.
23:35 blandfood joined #salt
23:35 seapasulli_ joined #salt
23:36 ninkotech__ joined #salt
23:38 ksalman joined #salt
23:39 blandfood hey guys/glas...i'm seeing something strange and was hoping someone could help shed some light...when on a minion and i run 'salt-call pillar.get <my key>', i get a value back, but in a state and i ' salt['pillar.get']('<my key>'), it returns a nil value
23:41 agliodbs joined #salt
23:41 blandfood <my key> being the same in both calls
23:42 blandfood its not happening for all pillar data, this is the first time I've run into this...trying to implement some new code
23:42 KyleG1 joined #salt
23:42 KyleG2 joined #salt
23:43 agliodbs I'm testing a salt config for a small group of servers, and I'm getting a python exception when running the overstate.  However, the path is very generic and I'm at a bit of a loss to figure out what's giving salt-master indigestion.
23:43 agliodbs here: http://salt.privatepaste.com/198565cdd7  any tips on how to figure out which config is giving salt indigestion?
23:44 blandfood my goal is to have "default" values in pillar <myapp>/init.sls, my state will pull this pillar, there will also be an "override" pillar and the state will pull that data and the 2 dictionaries will merge to give my config
23:44 blandfood the "overrides" are returning {} (giving {} as default to the .get)
23:46 _fllr_ joined #salt
23:48 saurabhs joined #salt
23:49 smcquay_ joined #salt
23:50 jnials_laptop Will salt-master 0.17.5 work with a salt-minion at 2014.1?  And what are the repercussions of upgrading to 2014.1?
23:52 allanparsons man, same thing happened arthbaska
23:52 allanparsons master stopped working
23:53 allanparsons perhaps an issue w/ peer communication (after i enabled that, all hell broke loose)
23:54 oz_akan_ joined #salt
23:54 KyleG joined #salt
23:56 ch0pstick joined #salt
23:56 hunter joined #salt
23:57 mpanetta joined #salt
23:58 manicouman joined #salt
23:59 utz left #salt

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