Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-10-16

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

All times shown according to UTC.

Time Nick Message
00:04 jslatts joined #salt
00:11 akasto joined #salt
00:12 redondos joined #salt
00:12 redondos joined #salt
00:14 cshuman joined #salt
00:17 taylorgumgum joined #salt
00:21 Jahkeup joined #salt
00:29 druonysuse joined #salt
00:33 retr0h joined #salt
00:33 retr0h joined #salt
00:34 oz_akan_ joined #salt
00:36 brimpa joined #salt
00:36 justlooks joined #salt
00:36 Linz_ joined #salt
00:37 justlooks clear
00:39 m_george|away joined #salt
00:43 Ryan_Lane joined #salt
00:47 modafinil joined #salt
00:48 modafinil okay, super dumb question from a puppet user 5 minutes into salt: whats the analogous concept for mapping hostname -> some set of config (if that's what i really want to do, even)?
00:49 m_george left #salt
00:54 cshuman joined #salt
00:55 justlooks modafinil: puppet has facter ,salt has grains
00:58 jdenning joined #salt
01:04 Drawsmcgraw joined #salt
01:08 pass_by_value joined #salt
01:09 cachedou_ modafinil: Here's some documentation on targeting in Salt: http://docs.saltstack.com/topics/targeting/index.html
01:10 modafinil cheers, thanks
01:10 cachedou_ You're very welcome. Glad you're checking out Salt. :]
01:14 mannyt joined #salt
01:21 m0hit_ joined #salt
01:22 shinylasers joined #salt
01:22 redondos joined #salt
01:22 redondos joined #salt
01:24 Drawsmcgraw When calling out to Pillar in Jinja templates, is there any difference between
01:24 Drawsmcgraw {{ salt['pillar.get']('some:pillar:value', 'default') }}
01:24 Drawsmcgraw and
01:24 Drawsmcgraw {{ pillar.get('some:pillar:value', 'default') }}
01:24 Drawsmcgraw ?
01:27 cshuman joined #salt
01:30 kula Drawsmcgraw: a slight difference.
01:30 kula or, maybe.
01:31 Drawsmcgraw My memory tells me that the latter was giving me problems, but the former's syntax is just.... way too chatty
01:31 kula i think the first one you are executing the salt function 'pillar.git' and using its output. the second, you are accessing the pillar variable, which is basically a hash of all of your pillar data which is given to jinja
01:32 kula although i think for the later i'm not sure that you can use the 'some:pillar:value' mechanism, you'd have to do pillar['some']['pillar']['value']...
01:32 lineman60 joined #salt
01:33 kula my brain is kinda mushy right now, but i was just looking at pillar stuff today in preparation to do some weird things
01:33 Drawsmcgraw kula: I see. That's exactly what my memory tells me. I remember getting a lot of 'None' values when using the latter.
01:34 Drawsmcgraw kinda screwed either way with syntax, then :/
01:34 Drawsmcgraw I'll stick with the former for the time being, then. Thanks
01:35 kula related, i am going to be much happier with git_pillar when i get 0.17 rolled out here.
01:37 Lue_4911 joined #salt
01:38 Drawsmcgraw kula: orly? I must've missed that in the release notes
01:40 kula i went chasing down rabbit holes only to discover the two bugs i was chasing appear to be fixed in 0.17
01:41 Drawsmcgraw ah :)
01:41 kula the most annoying of which was that the first time you pulled pillar data from a git repo, it worked, but subsequent updates only fetched the data, and didn't merge it.
01:41 Drawsmcgraw external pillars via a git repo?
01:41 Drawsmcgraw ...interesting
01:41 kula yes
01:42 Drawsmcgraw I just made /srv/pillar a Git repo on my Master
01:44 Drawsmcgraw But yours sounds more elegant
01:44 justlooks anyone can help this ? https://gist.github.com/justlooks/7001390
01:44 kula the implementation still grates me a bit, but it appears to be functional, and it allows me to track pillar stuff in git just like do files
01:44 kula i wish it was like the file stuff, though, in that branches/tags in git trivially become environments.
01:44 Drawsmcgraw justlooks: this line -> - name: /data/mrlocal/1/yarn:
01:44 Drawsmcgraw remove the colon on the end
01:44 Drawsmcgraw yeah I found myself wishing that as wel
01:44 kula yeah, in one job my file_roots and pillar dirs are just git repos, with some post-hooks so that when i push things to it it gets checked out
01:44 Drawsmcgraw The functionality you get for states kinda spoiled me
01:44 justlooks Drawsmcgraw: o thank you
01:44 Drawsmcgraw I guess that's why I just made the whole directory a Git repo :P
01:44 kula but i love the whole branch/tags are environments
01:44 Drawsmcgraw justlooks: no problem. I've seen that error *many* times myself.
01:44 jimallman joined #salt
01:44 Drawsmcgraw same here
01:46 _ilbot joined #salt
01:46 Topic for #salt is now Welcome to #salt - SaltConf in January!! http://saltconf.com | 0.17.0 is the latest | Please be patient when asking questions as we are volunteers and may not have immediate answers - Channel logs are available at http://irclog.perlgeek.de/salt/
01:46 deepakmd_oc joined #salt
01:50 neilf left #salt
01:50 neilf joined #salt
01:54 Lue_4911 joined #salt
01:54 fragamus joined #salt
02:02 justlooks i defined namenode_box= hostname ,how can i get ip ?
02:02 justlooks like pillar['namenode_box'].get_ip
02:03 Lue_4911 joined #salt
02:06 MoRoSKiT joined #salt
02:13 diegows joined #salt
02:16 Drawsmcgraw justlooks: you mean in your pillar?
02:17 Drawsmcgraw you have -> namenode_box : hostname    ?
02:20 redondos joined #salt
02:20 redondos joined #salt
02:20 jean_ joined #salt
02:24 mgw joined #salt
02:28 justlooks Drawsmcgraw:  yes
02:28 Drawsmcgraw You'd have to also have the machine's IP in pillar (unless there's more Pillar functionality I'm unaware of)
02:29 Drawsmcgraw as in -> namenode_box_ip: 192.168.0.1
02:29 justlooks Drawsmcgraw:  i write namenode_box = CM1  in pillar ,and also have namenode_ip = 192.168.xx.xx ,but when namenode changed i need modify namenode_ip again
02:29 justlooks Drawsmcgraw: i do not want to change two place
02:29 Drawsmcgraw Ah, is the IP constantly changing?
02:30 Drawsmcgraw justlooks: ping
02:30 justlooks Drawsmcgraw: hostname & ip all will changed
02:30 Gifflen joined #salt
02:31 Drawsmcgraw Unfortunately I'm not very good at dynamically updating Pillar values
02:31 justlooks Drawsmcgraw: i need deploy many cluster ,
02:31 justlooks Drawsmcgraw: anyway thanks for reply
02:31 Drawsmcgraw sure thing
02:36 Ryan_Lane joined #salt
02:36 Ryan_Lane joined #salt
02:38 gldnspud joined #salt
02:49 jefimenko joined #salt
02:49 shinylasers joined #salt
02:50 justlooks here my question https://gist.github.com/justlooks/7001390
02:53 Ryan_Lane joined #salt
02:55 Gifflen joined #salt
02:59 cshuman joined #salt
02:59 cshuman joined #salt
03:00 oz_akan_ joined #salt
03:16 Drawsmcgraw justlooks: You'd have to know the IP's ahead of time (may not work for you), but you could populate a pillar file and use Jinja templating in your state to iterate across it
03:16 Drawsmcgraw http://pastebin.com/7NFV2YM6
03:22 ajw0100 joined #salt
03:22 justlooks Drawsmcgraw: cluster master node only one ,so i do not need iterate
03:24 Ryan_Lane joined #salt
03:24 Drawsmcgraw I see. Then I'll stop making noise, I'm out of tricks :)
03:30 Lue_4911 joined #salt
03:30 ajw0100 joined #salt
03:33 xmltok joined #salt
03:34 mua joined #salt
03:36 redondos joined #salt
03:36 redondos joined #salt
03:37 berto- joined #salt
03:45 amahon joined #salt
03:47 Nexpro1 joined #salt
03:48 cshuman joined #salt
03:59 __number5__ dmesg
03:59 __number5__ sorry, wrong window :(
03:59 cshuman joined #salt
04:02 shadowsun joined #salt
04:03 shadowsun Yo. In 0.17.0 I ran salt -G 'node:group' -b 2 state.highstate, and it would batch starting highstate. The end effect of this is that it ran a state.highstate on all of them within about a minute, which (in reality) meant that all the jobs started by the highstates were running *at the same time*, completely negating the entire purpose of having a batch command. Is there an easy fix for this / is this a known bug? I kept my other mast
04:10 woebtz joined #salt
04:11 oz_akan_ joined #salt
04:13 cewood joined #salt
04:23 KyleG joined #salt
04:23 KyleG joined #salt
04:23 redondos joined #salt
04:23 redondos joined #salt
04:24 cshuman joined #salt
04:25 cshuman_ joined #salt
04:37 Lue_4911 joined #salt
04:55 middleman_ joined #salt
05:02 jasonrm joined #salt
05:10 Ryan_Lane joined #salt
05:18 cshuman joined #salt
05:21 ramteid joined #salt
05:23 cshuman joined #salt
05:23 NetworkN3rd joined #salt
05:30 idkfa joined #salt
05:30 justlooks how can i append content to a file?
05:31 drags justlooks: checkout file.append: http://docs.saltstack.com/ref/modules/all/salt.modules.file.html#salt.modules.file.append
05:33 justlooks drags: but how can i call module in sys file?
05:34 drags justlooks: in a .sls file
05:34 drags +?
05:34 justlooks drags:  i found http://docs.saltstack.com/ref/states/all/salt.states.file.html#salt.states.file.append
05:42 jefimenko joined #salt
05:43 TheCodeAssassin joined #salt
05:52 m0hit joined #salt
05:56 justlooks anything wrong for this? https://gist.github.com/justlooks/6898624
05:56 berto- joined #salt
06:07 Furao_ joined #salt
06:12 vipul joined #salt
06:14 jalbretsen joined #salt
06:19 druonysuse joined #salt
06:27 matanya joined #salt
06:31 jacksontj joined #salt
06:33 [M7] joined #salt
06:33 tomspur joined #salt
06:36 justlooks hi, i have a pillar file which defined a hostname list,and i want to find all hostname-> ip  ,append this content to each host's /etc/hosts file ,how can do this?
06:38 _ikke_ ja
06:38 _ikke_ ignore that :P
06:42 the_drow Good morning everyone :)
06:43 blee joined #salt
06:49 Lue_4911 joined #salt
07:01 justlooks call help for this https://groups.google.com/forum/#!topic/salt-users/I9c3Z3qnLDM
07:17 balboah joined #salt
07:19 ckao joined #salt
07:26 zooz joined #salt
07:27 goodbytes justlooks, I just replied to your topic in salt-users
07:31 justlooks goodbytes: yes i now use the same way which you post,but what i need is let salt fill the ip field automatically,so i just fill hostname in pillar
07:33 goodbytes justlooks, ahh I think I see what you mean. E.g. in a DHCP environment where you don't need to fill out IP addresses in the pillar data beforehand?
07:33 IzeBit joined #salt
07:33 IzeBit o/
07:34 the_drow What do I do if I get Name python in sls python is not a dictionary
07:35 the_drow https://dpaste.de/D2Mv this is the sls file
07:35 bud joined #salt
07:35 IzeBit Anyone here using 16.2 salt master and has experienced CPU problems on execution?
07:36 goodbytes IzeBit, https://dpaste.de/w6iX
07:37 goodbytes oops wrong name IzeBit sorry. Was meant to send that paste url to the_drow
07:38 IzeBit goodbytes: i figured :)
07:38 Katafalkas joined #salt
07:38 the_drow goodbytes: Same Error
07:39 goodbytes the_drow, try this instead https://dpaste.de/PXQo
07:40 the_drow goodbytes: Name python-packages in sls python is not a dictionary
07:41 the_drow goodbytes: could it be because I had a - pgk.installed instead of pkg.installed?
07:41 goodbytes the_drow, yes
07:41 goodbytes that should do it
07:42 the_drow Woot! Finally :)
07:42 the_drow goodbytes: thankls
07:42 goodbytes no problem :)
07:42 the_drow Now: State: - file     Name:      /usr/local/sbin/ez_setup.py     Function:  managed         Result:    False         Comment:   Unable to determine upstream hash of source file https://bitbucket.org/pypa/setuptools/raw/bootstrap/ez_setup.py         Changes:
07:43 the_drow I'd like to download this file.
07:43 justlooks goodbytes: what i think is the less error happen ,the less  pillar file is modified ,if salt can do this ,it's fine ,up to now i can not find good way
07:45 goodbytes the_drow, you will need to define "source_hash" for the file, is the md5 sum of the file you want to download. This page has some info on it: http://docs.saltstack.com/ref/states/all/salt.states.file.html
07:45 the_drow goodbytes: and if I don't care about it?
07:46 goodbytes think you can set - source_hash: None
07:47 goodbytes the_drow, to be honest, I've never used any source other than salt://, so I haven't tried it yet
07:47 the_drow goodbytes: "Source hash file None contains an invalid hash format, it must be in the format <hash type>=<hash>"
07:49 qba73 joined #salt
07:50 goodbytes the_drow, think there could be a few ways around it. The best practice of course is to pre-define the hash, that way you'll be sure that it won't just download an evil script and install it on your machine
07:50 IzeBit I have some issues with Salt Master 16.2. When i execute a job like 'test.version' in verbose mode i only get the "execution job with jid xxxxx" then it times out.
07:50 IzeBit i get following error "salt.exceptions.SaltReqTimeoutError: Waited 60 seconds"
07:51 IzeBit the salt master still keeps using 100% cpu for a while and i can't execute anything else.
07:51 IzeBit Is this a known bug in 16.x series with cpu problems?
07:52 goodbytes the_drow, the documentation says that it can also be input from a file, you could probably make a work-around that uses a cmd.run that downloads the file, makes a hash of it and saves it to a file
07:52 justlooks why i get this error https://gist.github.com/justlooks/6724543
07:53 the_drow goodbytes: I actually set the checksum to a random number and Salt tells me the real one :)
07:53 goodbytes the_drow, haha ok, if you need the md5 sum of a file on Linux, you can also run "md5sum /path/to/file"
07:55 the_drow goodbytes: I was too lazy to download them first
07:55 the_drow Is there a moudle for running python scripts yet or do I need to use the cmd module?
08:01 krissaxton joined #salt
08:02 berto- joined #salt
08:03 felixhummel joined #salt
08:16 the_drow Am I doing something wrong here: https://dpaste.de/iG24
08:18 the_drow It can't run the python script for some reason
08:21 rmt_ the_drow, it looks like you're trying to install & run the script in the same command?  That seems odd to me (although I'm not an expert)
08:22 the_drow rmt_: I'm downloading it and then executing It.
08:22 the_drow rmt_: But that's not the same command. I use file.managed to download the script
08:22 the_drow rmt_: https://dpaste.de/nEpc
08:23 rmt_ It also doesn't look like valid yaml to me.
08:23 the_drow rmt_: it compiles.
08:23 rmt_ yaml.parser.ParserError: while parsing a block collection
08:23 rmt_ using the raw version from iG24
08:24 the_drow rmt_: I have no idea then. How should it look like?
08:24 rmt_ You start out with cmd.script's value being a list, and then you switch to treating it like a dict.
08:25 rmt_ Or is the line starting with "State: - cmd" something else?
08:25 rmt_ Must be..
08:26 the_drow rmt_: It won't compile otherwise  Name setuptools in sls python is not a dictionary
08:26 the_drow rmt_: oh that's just the status report
08:28 rmt_ Ok.. maybe file:/// doesn't work with cmd.script, and you instead want to use cmd.run ?
08:28 the_drow rmt_: maybe
08:30 the_drow rmt_: hmm for some reason cmd.run works and cmd.script does not
08:31 rmt_ Seems like cmd.script wants to download the script every time.. and passing in args, you're effectively calling the script like /tmp/downloadedscriptfile /usr/local/sbin/ez_setup.py
08:31 rmt_ Which might not work for ez_setup.py
08:31 rmt_ (if it supports file:///)
08:35 davromaniak left #salt
08:42 sebgoa joined #salt
08:53 justlooks anyone can help this ? it's too difficult for me https://groups.google.com/forum/#!topic/salt-users/x-nWSsyIt1I
08:54 the_drow Why does these packages fail to install: https://dpaste.de/WWGG?
08:54 qba73 joined #salt
08:57 justlooks the_drow: remove the names
08:57 justlooks the_drow: see doc http://docs.saltstack.com/ref/states/all/salt.states.pkg.html#salt.states.pkg.installed
08:58 the_drow justlooks: The following packages failed to install/update: build-essential, python-dev-2.7, python-dev-3.3
08:59 justlooks no names under  pkgs
08:59 justlooks the_drow: remove it
08:59 the_drow justlooks: I just did and the error is almost the same
09:00 cshuman joined #salt
09:00 justlooks the_drow: check the log output ,and when you install it manually ,if it will sucess
09:00 justlooks the_drow: success
09:05 the_drow justlooks: The package names for python-dev were wrong but build-essential exists
09:05 the_drow justlooks: but it still won't install them
09:06 tomeff joined #salt
09:06 the_drow justlooks: It does install them when I do it manually
09:10 justlooks the_drow: what the log say?
09:11 the_drow That it can't locate those packages. I just figured out that python3.3-dev is missing for precise
09:15 N-Mi joined #salt
09:15 justlooks how can i write sls file ? when state A on host B finished i need B fire event to master
09:17 cshuman joined #salt
09:17 adepasquale joined #salt
09:22 SpX joined #salt
09:25 justlooks can i order the sls ? i find sls file include order is not the execute order in init.sls
09:26 Tekni joined #salt
09:30 ggoZ joined #salt
09:33 bhosmer joined #salt
09:34 elfixit joined #salt
09:36 Ryan_Lane joined #salt
09:44 krissaxton joined #salt
09:50 sm1ly joined #salt
09:50 sm1ly re2all. guys, does iptables.check works? cause I getting not available everywhere
09:53 tedski joined #salt
10:01 tedski joined #salt
10:09 the_drow how do I require a package installed by pip?
10:11 Ryan_Lane joined #salt
10:15 opalfonso joined #salt
10:16 opalfonso hi, i´m browsing salt source code and i´m wondering if there is some overview guide, maybe i missed it, thanks!
10:17 sm1ly re2all. guys, does iptables.check works? cause I getting not available everywhere
10:18 zooz joined #salt
10:22 rmt_ opalfonso, You wish! (I wish too)
10:24 the_drow Is there a way to figure out what's the user home?
10:24 the_drow https://dpaste.de/owIq - this doesn't work because the path is not absolute
10:25 opalfonso rmt_, Ok, thanks
10:32 the_drow I need the $HOME environment variable but I can't find a way to get it
10:40 the_drow How do I get the value of the $HOME environment variable?
10:43 zooz joined #salt
10:46 Ryan_Lane joined #salt
10:51 the_drow I just checked the code and it doesn't exist. How do I write a custom grain?
10:56 krissaxton joined #salt
11:04 rmt_ the_drow, I'm not sure what you're doing, but it sounds wrong..
11:09 bud mgw: hello, do you have a couple of minutes to talk to you about extending the lxc module to also take the MAC address as input?
11:19 bud I have already tried to code it myself - https://gist.github.com/valentinbud/7006196, but I receive TypeError when running lxc.init :\.
11:28 whiskybar joined #salt
11:29 __number5__ the_drow: are you looking for a certain user's home directory?
11:29 kiorky terminalmage: rah jenkins is spamming me on irc :p
11:29 kiorky terminalmage: http://jenkins.saltstack.com/job/salt-rs-freebsd/394/
11:29 kiorky i did not asked to be spammed !!! :)
11:29 kiorky terminalmage: well, i just ignored the bot ^^
11:30 __number5__ keep all your nickname in sync is dangerous, you know :P
11:31 az87c joined #salt
11:32 oz_akan_ joined #salt
11:39 swa joined #salt
11:40 swa Did the processing order of custom grains changed with 0.17 ?
11:41 giantlock joined #salt
11:49 logix812 joined #salt
11:51 hotbox joined #salt
11:52 groovehunter joined #salt
11:54 groovehunter hi! Can I configure a minion so that it uses a 2nd master (for some additions) - not for redundancy?
11:58 Thiggy joined #salt
12:11 jbunting joined #salt
12:11 jrdx joined #salt
12:12 jslatts joined #salt
12:12 jrdx joined #salt
12:13 giantlock joined #salt
12:15 mannyt joined #salt
12:18 Ryan_Lane joined #salt
12:20 diegows joined #salt
12:20 jbunting joined #salt
12:21 the_drow __number5__: current user's home directory
12:21 Heartmender joined #salt
12:22 the_drow rmt_: I need to create a Projects folder for the user who runs salt-call
12:25 the_drow So how do I get the value of the $HOME environment variable?
12:28 jumperswitch joined #salt
12:29 Scorchin joined #salt
12:29 ipmb joined #salt
12:29 Scorchin Hello. Are there any benchmarks on how long it takes to apply a given salt state? I'm currently on a team using Puppet and it's one of our biggest pain points.
12:30 cro joined #salt
12:30 the_drow __number5__: I have no idea what the user's name will be
12:31 Scorchin Additionally, what are the best examples of testing for salt configurations out there? Is there a decent repo with examples of "how to do it right" ?
12:32 oz_akan_ joined #salt
12:32 bhosmer joined #salt
12:33 blee joined #salt
12:34 oz_akan_ joined #salt
12:36 rgbkrk joined #salt
12:37 the_drow Even this for some reason doesn't create the dirs https://dpaste.de/cgNG
12:40 swa Scorchin: look up salt formulas for best practice code
12:48 Ryan_Lane joined #salt
12:48 amahon joined #salt
12:48 __number5__ the_drow: by default, salt will run as root, unless you specified runas=username
12:49 the_drow __number5__: then what do I do if I need to create the directories for a user that I don't know what username he'll pick
12:50 __number5__ also, there is virtualenv state http://docs.saltstack.com/ref/states/all/salt.states.virtualenv.html
12:50 the_drow __number5__: I am aware of that but I want to create the virtualenvs in a way that they will be used by virtualenvwrapper
12:51 the_drow __number5__: Should I move to the python dsl?
12:52 __number5__ the_drow: what's the difference between virtualenvwrapper and virtualenv?
12:52 the_drow __number5__: you can simply type workon <virtualenv_name?> and it will activate it (among other things)
12:53 N-Mi joined #salt
12:54 __number5__ there is a venv_bin param in virtualenv state, you can set it to virtualenvwrapper if it has been installed
12:55 __number5__ but that won't help if you don't have a user
12:55 the_drow __number5__: but the command is different. You create a virtualenv by typing mkvirtualenv
12:56 the_drow __number5__: and I have to set the .bashrc file like this: http://virtualenvwrapper.readthedocs.org/en/latest/install.html#shell-startup-file
12:56 __number5__ ok, guess you need to set venv_bin='mkvirtualenv' then
12:56 the_drow It has to be the .bashrc file of the user, not root. I'm creating a salt project to bootstrap our development machines.
12:57 the_drow __number5__: Ok, after that how do I set the environment variables correctly
12:57 the_drow __number5__: and how do I know where to put them
12:57 jslatts joined #salt
12:57 the_drow can I just do /home/*/.bashrc?
12:59 the_drow __number5__: brb
13:00 __number5__ use file.managed http://docs.saltstack.com/ref/states/all/salt.states.file.html#salt.states.file.managed
13:01 mwillhite joined #salt
13:02 __number5__ file state also has methods to create directories (so you don't need to mkdir -p)
13:03 brianhicks joined #salt
13:04 StDiluted joined #salt
13:06 racooper joined #salt
13:06 Ryan_Lane joined #salt
13:06 Gifflen joined #salt
13:07 jeffasinger joined #salt
13:07 bhosmer joined #salt
13:11 the_drow __number5__: does it accept glob like /home/*/
13:12 __number5__ nope, you have to know the user anyway, or else the permission will be wrong
13:14 juicer2 joined #salt
13:14 the_drow __number5__: can I set it in some config file?
13:15 the_drow Say somehow in my top.sls?
13:15 __number5__ yes, you can set it in pillar, the use the {{}} reference it
13:16 the_drow __number5__: isn't that overly complex for a bootstrap script with no master?
13:16 jbunting joined #salt
13:17 __number5__ http://docs.saltstack.com/topics/tutorials/pillar.html
13:18 __number5__ to be honest, if you only want to set up virtualenv for a user, a shell script will be much faster than salt master/minion setup
13:18 mua joined #salt
13:18 srage joined #salt
13:19 the_drow __number5__: we have no master yet
13:19 aleszoulek joined #salt
13:19 the_drow __number5__: we also want to clone all our repositories from the organization's github account
13:19 gildegoma joined #salt
13:20 __number5__ are the dev machine just for one dev or shared by many?
13:20 the_drow __number5__: each dev will run this script
13:21 __number5__ it will be masterless salt-minion?
13:21 the_drow __number5__: for now.
13:24 __number5__ I guess you can pass in user name as parameter when you do salt-call
13:26 __number5__ you need to write a custom module
13:26 srage_ joined #salt
13:26 Katafalkas joined #salt
13:27 the_drow __number5__: won't the Python DSL be enough?
13:28 __number5__ I don't know how you can pass in params to a sls
13:29 lineman60 joined #salt
13:29 Kholloway joined #salt
13:29 mannyt joined #salt
13:32 the_drow I'll tell the next dev what's his user will be
13:33 the_drow Why do I have to run this sls file twice in order for salt to recognize pip? https://dpaste.de/0nTu
13:34 jrdx joined #salt
13:34 __number5__ cmd.run don't keep states
13:35 __number5__ what's the OS of your machine?
13:35 the_drow Ubuntu, I don't want pip 1.3
13:36 the_drow Yes I know I can upgrade manually but that will mess up with the distribution later on
13:36 __number5__ you don't need to manually get-pip, use easy_install pip
13:37 toastedpenguin joined #salt
13:37 the_drow __number5__: which is the most insecure way to get pip
13:37 __number5__ after that, http://docs.saltstack.com/ref/states/all/salt.states.pip_state.html#module-salt.states.pip_state
13:37 the_drow easy_install doesn't work with SSL
13:39 the_drow __number5__: It installs pip. Salt doesn't recognize it.
13:41 __number5__ are they under same python?
13:42 the_drow Yup
13:43 __number5__ try to specify the pip path in bin_env
13:44 the_drow good idea :)
13:44 mwillhite joined #salt
13:45 the_drow __number5__: I'm trying :)
13:47 __number5__ the_drow: it's too late here I need to go to bed now. Good luck with your sls :)
13:47 the_drow __number5__: thanks a lot!
13:47 the_drow good night
13:47 __number5__ the_drow: no worries. good night
13:49 unicoletti_ joined #salt
13:54 krissaxton joined #salt
13:56 timoguin joined #salt
13:56 __gotcha joined #salt
13:57 NetworkN3rd joined #salt
13:57 lemao joined #salt
13:59 mannyt_ joined #salt
14:01 andredublin joined #salt
14:01 andredublin left #salt
14:01 mgw bud: that should be easy
14:01 kaptk2 joined #salt
14:04 bud mgw: I think so too but my Python coding skills are just building up. I have tried as I have posted in the gist and it doesn't work :|.
14:05 amahon joined #salt
14:07 rgbkrk joined #salt
14:07 mgw bud: there's one problem though
14:07 mgw a container can have multiple nics in the profile
14:07 mgw so you'd have to somehow pass multiple hwaddrs
14:08 mgw potentially
14:08 mgw and get them associated with the nics in the profile
14:08 bud Yes, I thought about that too, wouldn't mac={'eth0': '00:11:22:33:44:55', 'eth1': '00:11:22:33:44:55'} work for that case?
14:08 whiskybar joined #salt
14:09 bud Of course different MACs :).
14:09 andredublin joined #salt
14:10 andredublin left #salt
14:10 andredublin joined #salt
14:10 bud But it gets passed as a string not as a dict and in the for in _gen_config it crashes because of that.
14:11 mgw bud: yeah, something like that should work
14:11 brimpa joined #salt
14:11 N-Mi joined #salt
14:11 N-Mi joined #salt
14:11 mgw brb
14:11 bud ok.
14:11 mgw bud: ^
14:12 cachedout joined #salt
14:12 mgw joined #salt
14:12 kermit joined #salt
14:13 mgw bud: where's your gist?
14:13 bud https://gist.github.com/valentinbud/7006196
14:13 andredublin left #salt
14:14 mgw what does it crash with?
14:16 bud I have added a comment to the gist.
14:16 mgw how did you invoke the module?
14:17 bud I have used salt 'target' lxc.init params...
14:18 cro joined #salt
14:18 bud There's another comment in the gist with the command.
14:18 mgw with mac="{'eth0': 'aa:bb:….'}"
14:18 jbunting joined #salt
14:18 mgw ?
14:18 bud Yes.
14:19 gldnspud_ joined #salt
14:19 mgw take out the single quotes
14:19 mgw so it looks like the config
14:19 mgw or change the config to have single quotes
14:19 mgw i'm not sure which
14:19 mgw I'm not sure if it's evaluating as yaml or json
14:20 mgw the config you have is yaml, the mac is json
14:20 mgw actually, the mac is neither
14:20 mgw it would need to be '{"mac": "…"}' to be json
14:20 mgw with the keys and values wrapped in double quotes
14:22 bud Ok, I will try that just now.
14:22 bud Same error.
14:22 mgw what's you command now?
14:23 bud -> /home/bud/.virtualenvs/salt/bin/salt 'nox' lxc.init lxc109 nic=default profile=local-debian-wheezy-lvm config='{master: 192.168.122.1}' seed=True mac='{"eth0": "AC:DE:48:60:2A:28"}'
14:23 mgw add the double  quotes for config too
14:24 mgw and if that doesn't work, remove the quotes from both
14:24 bud Ok.
14:24 mgw trying to get your patch to apply
14:24 alunduil joined #salt
14:24 bud Same error.
14:25 bud But note that config works without double quotes. If I remove the mac= from the invocation the command inits an lxc container.
14:25 mgw ok
14:26 mpanetta joined #salt
14:28 dustyfresh joined #salt
14:28 timoguin_ joined #salt
14:30 mgw bud: I think the module has to do the coercion
14:30 mgw looking at the seed module
14:30 teskew joined #salt
14:31 bud Did you mean corection?
14:33 bud I have just found out what coercion means :).
14:33 mgw bud, no coercion
14:33 mgw but seed doesn't seem to do any coercion
14:33 mgw I'm still digging
14:33 bud In this context, translate the mac string in a dict you mean?
14:34 tyler-baker joined #salt
14:35 mgw bud: found it
14:35 mgw salt is doing the coercion
14:35 bud Ok...
14:35 mgw but only if it's a named kwarg
14:35 bud I admit you lost me...
14:35 mgw just add mac=None to the init def
14:35 mgw before **kwargs
14:36 bud Ok.
14:36 mgw I'll wrap this up if you want, adding support for providing IPs too
14:37 mgw it'll like like this: nic_opts="{eth0: {ipv4: …., ipv6: …., mac: ….}}"
14:37 mgw I should have this in develop later today
14:38 mgw but for now, your patch works with that one change
14:38 bud That's be awesome.
14:38 Jahkeup joined #salt
14:38 mgw or you can do it :-)
14:38 mgw and open a pull req on in
14:38 bud I have tried with what you've suggested but it doesn't work.
14:39 mgw what does your command look like?
14:39 bud The MAC is generated in the config file, not the one passed from the CLI.
14:39 bud -> /home/bud/.virtualenvs/salt/bin/salt 'nox' lxc.init lxc109 nic=default profile=local-debian-wheezy-lvm config='{master: 192.168.122.1}' seed=True mac='{"eth0": "AC:DE:48:60:2A:28"}'
14:40 mgw oh, you're right
14:40 mgw I know why, but when i fix it, we're back to square one
14:41 bud And why is that?
14:42 cowmix joined #salt
14:43 mgw bud: remove the line mac = kwargs.pop()
14:43 mgw and then do this
14:43 mgw (or possible remove mac=None)
14:43 mgw I don't think it matters
14:43 mgw but you can't have both
14:43 mgw then use proper json
14:43 mgw it fixes it for me
14:43 mgw config='{"master": "192.168.122.1"}' seed=True mac='{"eth0": "AC:DE:48:60:2A:28"}'
14:44 mgw I suspect the config was only half way working
14:44 mgw meaning it was dumping the literal string you were passing it
14:44 mgw into the minion config
14:45 DredTiger joined #salt
14:46 bud It worked this time.
14:46 bud remove mac = kwargs.pop()
14:46 bud *removed
14:46 m_george|away joined #salt
14:47 pdayton joined #salt
14:47 SunSparc joined #salt
14:47 mgw ok, cool
14:47 m_george left #salt
14:48 mgw do you want to put together a pull req bud?
14:48 bud You've told me to ping you in the issue on Github about Device or resource busy. Specifying your username @mgwilliams does it or not?
14:49 bhosmer joined #salt
14:49 bud mgw: that would be nice. I have the modifications done on a different branch. I should rebase on top of develop and push to my salt fork and then pull request to upstream. Would this do it?
14:50 mgw yes, but if you can, change it to work the way I showed
14:50 mgw with nic_opts
14:50 mgw and ipv4 and ipv6
14:51 bud I will try and get back to you.
14:52 mpanetta_ joined #salt
14:52 bud Thank you for your assistance in this matters. The LVM backing containers don't work without sleeping a few seconds after the final umount from seed.apply
14:53 bud I am talking about this one https://github.com/saltstack/salt/issues/7844.
14:53 deepakmd_ joined #salt
14:56 cachedout joined #salt
14:58 pass_by_value joined #salt
15:01 ashtonian joined #salt
15:02 linjan joined #salt
15:02 jacksontj joined #salt
15:03 mgw bud: Yeah, I saw that ticket
15:03 mgw thanks for reminding me
15:05 taylorgumgum joined #salt
15:05 mgw bud: can you push to your github and ping me before you open the pull request?
15:07 jumperswitch joined #salt
15:08 alekibango joined #salt
15:10 bud mgw: Of course. I feel obliged to tell you that I don't know how much would it take me to implement the nic_opts you said.
15:13 jeffasinger joined #salt
15:13 bitz joined #salt
15:13 taylorgumgum joined #salt
15:14 mgw bud: I can do it if you want, starting with your patch
15:15 mgw it's not a big deal
15:15 sm1ly re2all. guys, does iptables.check works? cause I getting not available everywhere
15:16 monokrome joined #salt
15:16 bud mgw: the work I have done on lxc and seed are already on https://github.com/valentinbud/salt on branch feature-lxc-mac-address.
15:17 mgw bud, ok, thanks
15:17 mpanetta Hi, I am trying to make my salt-cloud map files more dynamic, but it does not seem to render the mako text I am putting in, at all.  Is there something I need to do to let salt-cloud know the file needs to be rendered?
15:17 mgw just let me know whether you will do the nic_opts or want me to
15:17 forrest joined #salt
15:17 tomeff joined #salt
15:17 bud If it's easier for you to do it, that would be faster.
15:17 mgw ok
15:17 mgw I'll do it, probably today
15:19 mpanetta I see in the source that it just calls Template and then renders it unconditionally, so... Does it not take standard ${ } template names?
15:19 forrest mgw, breaking more stuff? :P
15:19 bud Ok, thanks :).
15:19 mgw forrest: definitely
15:20 mgw this shouldn't break anything
15:20 forrest mgw, I don't even know the context of the discussion, just logged on and poking fun
15:20 mgw afaik, only bud and I are using the LXC module
15:21 bud Seems like so mgw. At least we are if nobody else does :).
15:21 mgw i'm hoping to give a talk at saltconf on lxc and virt
15:21 forrest Yea I remember you saying that yesterday, that would be cool
15:22 cshuman joined #salt
15:22 bud Cool, I'll watch it. I don't think I'll be able to get a VISA to attend the saltconf.
15:22 forrest Where are you at bud?
15:22 bud .ro
15:22 mgw romania
15:22 mgw I"ve been there
15:22 forrest Psssh, just tell them you wanna go skiing.
15:22 bud Really? Where exactly?
15:23 forrest Salt Lake city is where the conference is
15:23 mgw Moldavia region, mostly
15:23 bud Yeah forrest that should do it.
15:23 redondos joined #salt
15:23 redondos joined #salt
15:23 forrest oh I thought you were asking me bud, sorry
15:23 mgw maybe he was ;-)
15:23 bud Nice place Moldavia. Did you enjoy it?
15:23 mgw yes
15:23 forrest I'd like to head up to Romania, I didn't get a chance to get there when I went to Eastern Europe last year
15:24 mgw Where are you forrest?
15:24 bud I was talking to forrest, I thought he could help me get a VISA :P.
15:24 forrest In the US, Arizona.
15:24 forrest I wish I could bud :\
15:24 mpanetta Nobody knows?
15:24 forrest I'd live in Eastern Europe if I could though, tried to get a job there.
15:24 cshuman joined #salt
15:25 forrest mpanetta, what's the question?
15:26 bud Well, good day to you Gentlemen, got to head home.
15:26 mpanetta forrest: I am trying to get salt-cloud to use mako to render the .map files
15:26 forrest ok, did you change the shebang?
15:26 mpanetta I had no idea that the .map files needed one... So, no :)
15:27 mpanetta Is there documentation anywhere on it?  The docs I looked at did not mention a #! line
15:27 forrest I honestly don't know mpanetta
15:27 forrest but I figured it would be worth a shot
15:27 mpanetta Ah
15:27 mpanetta Oh well
15:28 redondos joined #salt
15:28 forrest mpanetta
15:28 forrest https://github.com/saltstack/salt-cloud/pull/224
15:28 forrest according that it tries to render with mako first? Unless that was removed at a later point in time
15:29 mpanetta Not sure.  I do see in the code for Map class that it does render a template
15:29 gldnspud joined #salt
15:30 forrest but it's not rendering it?
15:30 forrest Are you just getting an error?
15:30 dakila joined #salt
15:30 redondos joined #salt
15:31 krissaxton joined #salt
15:31 dakila I have a minion running in a VM.  When the VM pauses for 20 seconds or more, the minion fails to connect to the server after the VM resumes.  Any ideas to detect if the minion is connected?
15:31 mpanetta No error either
15:31 pdayton left #salt
15:31 mpanetta forrest: Give me a sec, testing again.  I did not have mako installed.
15:32 forrest ok
15:33 UtahDave joined #salt
15:33 mpanetta forrest: Nope, so weird.  It does not render, but it does not bitch either.
15:33 mpanetta Even with mako installed ;)
15:34 forrest that's odd.
15:34 mpanetta I'm running 0.8.9
15:34 forrest can you change the log level?
15:34 mpanetta Yeah
15:34 forrest --log-level I think? Might have the syntax with salt-cloud -h, I don't use salt-cloud much so I can't remember.
15:35 StDiluted joined #salt
15:36 UtahDave -l debug
15:36 UtahDave mpanetta: what's the problem you're running into?
15:37 forrest oh nice, just the standard command option there UtahDave, good to see it is identical
15:37 krissaxton joined #salt
15:38 mpanetta UtahDave: Trying to templatize my map file for salt cloud, we have multiple masters (we test on a small system and deploy on a larger one) and I want the masters line to automatically pick up the correct IP
15:38 mpanetta yeah I am running with debug now to see what happens
15:39 UtahDave ah, ok
15:40 UtahDave mpanetta: So templating support for salt-cloud configs isn't really official yet.  Mako support was kind of slipped in a little while back.
15:40 mpanetta Ah ok
15:41 oz_akan_ UtahDave: still mako code is present, and based on source it seems it should work
15:42 UtahDave Yeah. From what I understand mako should work.
15:42 oz_akan_ mpanetta: even when run with debug, mako not even mentioned in the outpurt
15:42 UtahDave mpanetta: are you getting any stacktraces or errors?
15:42 mpanetta UtahDave: Nope, just enabled debugging, still nothing.
15:43 bhosmer joined #salt
15:43 taylorgumgum joined #salt
15:46 forrest mpanetta, did you restart the service after you installed mako?
15:46 forrest I have no idea how that rendering works, so this is a shot in the dark
15:46 forrest and probably worthless.
15:46 mpanetta Hmm, no.
15:47 mpanetta Apparently it works not lol
15:47 forrest ?
15:47 forrest oh it just started working randomly?
15:47 mpanetta Nope
15:47 mpanetta I was being silly
15:47 forrest oh I thought you meant not
15:47 mpanetta restarting
15:47 forrest err now
15:47 mpanetta Nope
15:48 oz_akan_ restart what? salt-cloud :)
15:48 mpanetta Good point. Yeah salt-cloud is not a service
15:48 forrest see, shows my extensive knowledge of salt-cloud :P
15:48 druonysuse joined #salt
15:48 druonysuse joined #salt
15:49 mpanetta That is ok :)
15:52 UtahDave mpanetta:  Yeah, just chatted with redbeard about it.  Templating hasn't been implemented yet for salt cloud configs.
15:52 forrest redbeard?
15:52 mpanetta UtahDave: Ok thanks
15:52 UtahDave Mako support was slipped in, but it's not done "properly"
15:52 oz_akan_ https://github.com/ydavid365/salt-cloud/blob/develop/saltcloud/cloud.py
15:52 mpanetta It looks like it is just *there*
15:53 UtahDave we're going to be merging salt-cloud right into salt proper soon and templating should be quite easy to add at that point
15:53 deepakmd_oc joined #salt
15:53 mpanetta Oh NICE!
15:53 oz_akan_ has mako section, and seems to be tried out as the first option, then it tailbacks to regular yaml parsing
15:53 forrest yea oz_akan_, that's what we were looking at earlier
15:54 mpanetta UtahDave: Do you know what version you are targetting for that integration?
15:54 oz_akan_ UtahDave: what is salt proper?
15:54 mpanetta oz_akan_: I assume salt libs/salt-master
15:54 UtahDave oz_akan_: salt-cloud is in its own git repo.  It will be brought directly into the salt repo
15:54 mpanetta Ah nice
15:55 oz_akan_ ok
15:55 UtahDave mpanetta: The salt-cloud merge into Salt is supposed to happen in the next few weeks
15:55 mpanetta So when this happens, we can template all-the-things?
15:55 oz_akan_ that is a life time for us :)
15:56 mpanetta For this project anyway, UtahDave do you think running a minion on the master to do the templating is an ok work-around?
15:56 deepakmd_ joined #salt
15:56 mpanetta BTW, I *love* the salt client library.
15:57 jalbretsen joined #salt
15:59 jrb28 joined #salt
16:07 UtahDave mpanetta: I don't see a problem with that at all!
16:08 pipps1 joined #salt
16:11 saurabhs joined #salt
16:13 jdenning joined #salt
16:14 mannyt_ joined #salt
16:15 dakila I have a minion running in a VM. When the VM pauses for 20 seconds or more, the minion fails to connect to the server after the VM resumes. Any ideas to detect if the minion is connected?
16:16 Furao I have been able to use template to render salt-cloud files but it's quite complicated : using that architecture https://www.dropbox.com/s/jkt39uhx1avfen6/global_arch.png
16:16 UtahDave mpanetta ^^
16:16 fragamus joined #salt
16:17 UtahDave dakila: is the salt-minion service still running when the VM resumes?
16:17 Lue_4911 joined #salt
16:17 dakila It is
16:18 redondos joined #salt
16:18 redondos joined #salt
16:18 UtahDave dakila: It should reconnect automatically. How long have you waited and still have it not connect?
16:18 dakila UtahDave: up to five minutes
16:18 UtahDave what os?
16:19 dakila UtahDave: while troubleshooting, I am running the minion not as a service, but in the console with -l debug
16:19 dakila it acts as if nothing happened, nothing logged to the screen
16:19 lmickh joined #salt
16:19 dakila If it is in the middle of a job, the job finishes and reports properly.  The master simply cannot give it any further tasks
16:20 mpanetta Furao: UtahDave, thanks!
16:21 MTecknology I have some states that run in random order. I don't have anything set. The pkg states such as salt, nginx, samba, php, etc. Some of those require firewall rules. I have a state that sets the firewall up and templates the config. My problem is... if the firewall states run first, then the template won't be right. How can I tell salt that the package state should run /before/ the firewall state.
16:21 MTecknology Setting -requires: on a state will do the opposite of what I want. I want the package required before the firewall state, only if the package is going to be installed for this system
16:22 UtahDave dakila: the minion has to reach out to the master to receive tasks from the master. So we need to figure out why the minion isn't accessing the master
16:22 dakila UtahDave: agreed
16:23 UtahDave MTecknology: you can add  - order: last   to your firewall state
16:23 MTecknology I'm looking for require_in, aren't I?
16:23 MTecknology or that
16:24 dakila UtahDave: can zeromq in salt use the restful configuration instead of the persistent connection?
16:24 UtahDave MTecknology: yeah, a require_in would work great, too
16:25 UtahDave dakila: I'm not sure on those details of zmq.  But the minion is definitely supposed to reconnect to the master after losing a connection
16:25 UtahDave dakila: if you're running the minion as a service, does it reconnect?    Also, which OS are you using?
16:27 dakila UtahDave: it does not reconnect as a service, either.  centos 6.4
16:27 UtahDave ok, could you open an issue on that, dakila?  I'm not sure why the minion wouldn't reconnect in that scenario.
16:27 unicoletti left #salt
16:27 Thiggy joined #salt
16:28 dakila Sure.  Thanks for your help
16:30 jacksontj joined #salt
16:30 UtahDave you're welcome, dakila. Sorry I don't have a quick answer for you.
16:33 berto- joined #salt
16:33 linuxnewbie joined #salt
16:34 Jahkeup joined #salt
16:34 m0hit joined #salt
16:36 linuxnewbie hello, following https://salt.readthedocs.org/en/latest/topics/tutorials/states_pt2.html i'm trying to better understand salt states but i'm having a problem with service ...it doesn't identify apache2 service ...since in the file i have apache i think ...
16:36 linuxnewbie is there anyway to fix it based on the distro ...the service name ?
16:37 linuxnewbie i'm trying using {% if grains['os'] == 'Ubuntu' %} ...but from what i see it just get ignored :(
16:39 UtahDave linuxnewbie: can you pastebin the file you're testing?
16:39 tedski left #salt
16:39 linuxnewbie sure
16:42 forrest linuxnewbie, you should also take a look at https://github.com/saltstack-formulas/apache-formula
16:42 forrest it uses a mapping to make things a bit cleaner in the state.
16:43 troyready joined #salt
16:43 linuxnewbie http://pastebin.com/HNJqSQeU
16:44 linuxnewbie forrest, i'll have a look at that link right now
16:44 jrdx joined #salt
16:45 _ikke_ linuxnewbie: At one point, you use apache2, and the other, apache
16:46 forrest linuxnewbie, you don't redeclare the name variable, so it treats pkg, and service as two different items
16:47 goodwill UtahDave: when new grains are added do we need to always restart salt-minion, or is a better way to propogate them to the master
16:47 linuxnewbie so if i have declared on the pkg that the name should be apache2 ...i need to declare it again on the service ?
16:48 forrest correct
16:48 forrest That's why I linked the formula that uses the mapping, makes it a lot cleaner :P
16:48 jY is there anything to handle adding ip aliases via ifconfig or ip addr add?
16:49 SgtMalicious joined #salt
16:49 forrest jY, in the states stuff, it doesn't appear so: http://docs.saltstack.com/ref/states/all/salt.states.host.html , but the module seems to have that support
16:49 forrest http://docs.saltstack.com/ref/modules/all/salt.modules.hosts.html
16:49 linuxnewbie forrest, thank you ...i wanted to follow the tutorial before getting more complex situations
16:49 UtahDave linuxnewbie: you need to add  - name: apache 2    or - name: httpd in your service.running
16:49 forrest linuxnewbie, yea np, I totally understand.
16:49 UtahDave goodwill: how are you adding new grains?
16:50 jY forrest: isn;t that just for /etc/hosts?
16:50 linuxnewbie UtahDave, i understand now ...thank you
16:50 UtahDave cool
16:51 forrest yea jY, probably should have phrased it better.
16:51 forrest linuxnewbie, you could always move your if statement
16:51 forrest so it sets the value of your line 1 to either apache2, or httpd
16:51 forrest does that make sense?
16:51 linuxnewbie forrest, not really :(
16:52 linuxnewbie can u show me on that paste.bin ?
16:52 forrest sure hang on
16:54 djn I'm preparing a presentation on configuration management which contains some basic intro to salt. Where can I get a higher res version of the logo from the homepage?
16:55 Flusher joined #salt
16:55 forrest linuxnewbie, kind of like this: https://gist.github.com/gravyboat/7011104
16:55 linuxnewbie forrest, i'll have a look ty
16:55 forrest I don't have my salt vm spun up right now so I can't validate the syntax/functionality though
16:56 linuxnewbie i'll try it in a moment
16:56 forrest but if you're on 0.17, it should process that in order, so you don't have to have the requires in the service.
16:56 UtahDave djn: pm me your email
16:56 forrest linuxnewbie, yea no rush
16:58 mwillhite joined #salt
17:00 m0hit joined #salt
17:04 Jahkeup_ joined #salt
17:04 mua joined #salt
17:09 linuxnewbie forrest, ty it's working ok
17:11 josephholsten joined #salt
17:12 mephx joined #salt
17:13 lineman60__ joined #salt
17:13 zz_cro joined #salt
17:14 alunduil joined #salt
17:15 lineman61 joined #salt
17:16 jrdx joined #salt
17:16 Gareth I've come to the conlusion that salt is not meant to work on OS X 10.6.8 :)
17:17 ccase joined #salt
17:20 aleszoulek joined #salt
17:21 forrest linuxnewbie, awesome, glad to hear that did it! You understand the logic behind why that works but your original example doesn't right?
17:23 amckinley joined #salt
17:24 MTecknology Is this not valid? http://dpaste.com/1419111/
17:24 forrest what error are you getting MTecknology
17:26 MTecknology ScannerError: mapping values are not allowed here    (service.running:)
17:26 MTecknology I can grab the whole error if that's not enough
17:27 forrest nah that's good enough.
17:27 cachedout joined #salt
17:27 Furao MTecknology: replace line 2 with pkg:\n    - installed
17:27 forrest https://gist.github.com/gravyboat/7011581
17:28 MTecknology ah.. I bet if I browsed code that would make perfect sense
17:29 forrest Did you implement Furao's change and it worked?
17:29 forrest I thought the issue was due to service.running:, then trying to do a subsection within running that didn't exist.
17:29 MTecknology Seems to be working fine with that change
17:29 forrest Furao's?
17:29 UtahDave joined #salt
17:30 MTecknology ya
17:30 forrest ok cool, yaml fun go
17:30 ze- I just noticed a file.recurse that include .. dirs. Anyone noticed the bug (and possibly the fix?)
17:31 MTecknology I've grown fond of yaml
17:32 forrest yea that's just a yaml issue where sometimes you can't mix items
17:33 Thiggy joined #salt
17:33 MTecknology I've seen more irritating things with less useful utilities. This one I won't whine about. :)
17:33 UtahDave ze-: I think I remember there being an open issue on that. I think it may have been closed already
17:34 Furao joined #salt
17:34 forrest MTecknology, fair enough!
17:35 ze- ok. just found it.
17:35 ze- yeah, fixed. 7712
17:35 TheCodeAssassin joined #salt
17:35 ze- wait, no. open bug.
17:35 ze- https://github.com/saltstack/salt/issues/7404
17:35 MTecknology I have an insane number of states being run on this set of servers..
17:36 forrest I was gonna say ze-, still open :P
17:36 cnelsonsic joined #salt
17:38 ajw0100 joined #salt
17:39 ubercore joined #salt
17:40 jacksontj joined #salt
17:40 ubercore Can anyone help me understand the error in this command?
17:40 ubercore https://gist.github.com/peterlandry/7011814
17:41 ubercore it looks like dest= isn't being parsed out as a kwarg, and I'm not sure why
17:41 sebgoa joined #salt
17:42 _ikke_ How can I debug pydsl state files? How can I get output from it?
17:42 forrest ubercore, I'm pretty sure that's a tar related error.
17:43 ubercore No, I don't think it is: Executing command 'tar -xvzf archive.tar.gz dest=/tmp/'
17:43 ubercore it's not grabbing /tmp/ as the 'dest' kwarg
17:43 m0hit joined #salt
17:45 UtahDave ubercore: what version of Salt are you on?
17:45 m0hit_ joined #salt
17:45 ubercore 0.17
17:45 forrest salt.modules.archive.tar(options, tarfile, sources, cwd=None, template=None)¶
17:46 UtahDave hm. looking at the docs it looks like there were some changes in 0.17.0
17:46 ubercore forrest: that's changed in 0.17. That signature is for pre-0.17
17:46 Ryan_Lane joined #salt
17:47 UtahDave ubercore: I think that's a bug.
17:47 forrest ubercore, oh ok, my bad was looking at the docs and thought that was up to the latest
17:47 UtahDave ubercore: would you mind opening an issue on that?
17:47 ubercore As a side note, the example a) doesn't format properly in the rendered docs, and b) switches the order of "options" and "tarfile"
17:47 ubercore UtahDave: will do
17:47 UtahDave Thanks!
17:47 cowmix when i do a standard bootstrap I get .16.X .. isn't .17.X the stable version of Salt?
17:48 ubercore sorry, more specifically I mean that the example for unpacking with tar
17:48 ubercore cowmix: bootstrap by default will use your OS package version
17:48 ubercore if you want to get 0.17, you have to tell bootstrap to use a git tag
17:48 cowmix It was on centos
17:48 ubercore (until your distro upgrades its packaged salt version to 0.17, that is)
17:48 cowmix 6.4
17:48 UtahDave cowmix: are you on RHEL/Cento?
17:49 cowmix yeah
17:49 UtahDave cowmix: Yeah, 0.17 hasn't hit epel yet.
17:49 cowmix ah
17:49 cowmix so... can I run a .17 master and a .16 minion?
17:49 forrest Yea this guy named UtahDave who builds it, he doesn't have any credits to push it through from epel-testing to epel-stable, such a slacker :P
17:49 ubercore This is how we bootstrap 0.17 on ubuntu: curl -L http://bootstrap.saltstack.org | sh -s -- -M -N git v0.17.0
17:49 ajw0100 joined #salt
17:50 UtahDave cowmix: yeah, that should be fine. Generally your master should be the same or higher version than your minions
17:50 forrest ubercore, was just gonna suggest pulling the tag from git!
17:50 _ikke_ I'm trying to debug a pydsl state, but I can'
17:50 _ikke_ I can't seem to get any output from it. Is there a way?
17:50 UtahDave :)  nah, I don't handle the rpm packaging.
17:51 UtahDave _ikke_: when you execute the state have you tried running it in debug mode?   -l debug
17:51 _ikke_ UtahDave: Was just trying
17:51 forrest UtahDave, oh I thought that herlo wasn't doing it anymore, did someone else take over?
17:51 UtahDave herlo is still involved, but terminalmage is taking a bigger role
17:52 _ikke_ UtahDave: It doesn't seem to give any output from the state
17:52 UtahDave what version of Salt are you on _ikke_?
17:53 _ikke_ 0.17
17:53 UtahDave were those pydsl states working before 0.17?
17:53 _ikke_ https://gist.github.com/Ikke/999321d2f4d274001ab9
17:54 _ikke_ UtahDave: I'm new to salt, just installed it a few days ago
17:54 mwillhite joined #salt
17:54 UtahDave ah, ok.  Can you pastebin the state you're working on?
17:54 ubercore UtahDave: issue for the archive bug is at https://github.com/saltstack/salt/issues/7887 fyi
17:55 mgw joined #salt
17:56 UtahDave thanks, ubercore!
17:57 ubercore np. thanks for the quick response
17:57 _ikke_ UtahDave: https://gist.github.com/Ikke/3fc1299c51c32439183c
17:57 TheCodeAssassin2 joined #salt
17:58 cro joined #salt
17:59 UtahDave _ikke_: Hm. I'm really not sure what's going on there. I don't use pydsl at all.
18:00 _ikke_ The script itself works alsmost
18:00 UtahDave _ikke_: You might ask on the mailing list and catch some of our pydsl users there
18:00 _ikke_ almost
18:00 _ikke_ ok
18:02 josephholsten joined #salt
18:02 cro joined #salt
18:05 _ikke_ UtahDave: I'll try to test the code separately
18:05 UtahDave cool. good idea
18:06 pass_by_value joined #salt
18:10 faldridge joined #salt
18:12 ashtonian joined #salt
18:12 amckinley joined #salt
18:14 NetworkN3rd joined #salt
18:14 faldridge left #salt
18:15 jacksontj joined #salt
18:18 jrdx joined #salt
18:18 jacksontj_ joined #salt
18:20 timoguin joined #salt
18:20 _ikke_ UtahDave: It helps if you remember the correct fqdn :P
18:20 amckinley joined #salt
18:21 UtahDave joined #salt
18:23 TheCodeAssassin joined #salt
18:24 saurabhs is there a relevance in worker_thread configuration in salt master config to number of salt-master process running on the box.
18:25 UtahDave saurabhs: yep!
18:26 TheCodeAssassin2 joined #salt
18:30 saurabhs UtahDave: so I set worker_thread : 6 and I see 10 salt-master processes
18:30 berto- joined #salt
18:30 saurabhs is that correct or there is something wrong here
18:31 UtahDave That's problably correct.   6 should be general worker threads, but the master also has a few other processes it spins up for other tasks
18:32 saurabhs hmmm ok so I probably shouldn't worry about those extra 4 processes.
18:32 mgw joined #salt
18:33 UtahDave no, I wouldn't.  If you start seeing a lot more than that, then maybe
18:33 saurabhs Now another thing I am trying to understand is, the the worker_thread configuration. if I set it to 6 does it mean that at a time I can interact with only 6 minions (like execute state.highstate).
18:34 UtahDave well, each worker thread can handle multiple minions at a time.
18:34 aleszoulek joined #salt
18:34 UtahDave if you start getting into the many hundres or  thousands of minions, then you'll probably want to bump up your worker_threads
18:35 UtahDave but you'll start using more ram on the master.
18:35 UtahDave so it's a balance between resources that are available on the master and how many minions you're dealing with.  You just kind of have to play with the numbers to find out what works best in your infrastructure
18:36 adepasquale joined #salt
18:36 saurabhs if I have 4 cpus on my VM (salt master), what should be possibly correct value for worker thread,
18:38 UtahDave how many minions are you managing with this master?
18:38 Corey saurabhs: What I do is start with the default, and change it when I start to see performance degrade. But that's just me; there's probably a more-better answer.
18:39 UtahDave yeah, the defaults should fine until you start really getting into a bigger infrastructure
18:39 saurabhs 100-120
18:40 schmichael joined #salt
18:40 Corey Defaults should be fine there.
18:41 schmichael i just hit this when upgrading from 0.16 to 0.17: https://github.com/saltstack/salt/issues/2310
18:41 schmichael any ideas on how to fix it without using some horrible #ifndef hack via jinja templates?
18:42 eliasp is it possible to manage keys through Halite?
18:42 schmichael i'm new to salt, but it seems like this breaks any sort of state dependency tree... which seems like a pretty core feature
18:43 anuvrat joined #salt
18:43 UtahDave schmichael: yeah, let me ask around on that one
18:44 schmichael UtahDave: thanks, that issue actually seems to be a bit misleading. the last year-old comment describes the 0.16 behavior which seemed to work as intended (duplicate includes were skipped)
18:44 schmichael UtahDave: and the recent comments are referring to how that behavior breaks when upgrading from .16 to .17
18:45 mwillhite joined #salt
18:45 UtahDave right
18:45 schmichael glad that makes sense to you; i confused myself for a while when reading through the issue :)
18:46 UtahDave schmichael: yeah, I was considering opening a new issue on the apparently recent change in behavior
18:47 schmichael UtahDave: well if you don't i will :) i'm off to grab lunch, so if you don't beat me to creating an issue i'll ping you with the one i create
18:47 schmichael thanks!
18:47 shinylasers joined #salt
18:47 UtahDave thank you, schmichael!
18:47 jbunting joined #salt
18:50 saurabhs UtahDave: I have may be about 107 minions running, when I do salt '*' state.highstate, it goes and executes it on all of them and dumps the output back to the screen. obviously the output is difficult to parse and I want to identify the minion which failed in state.highstate.
18:50 saurabhs So I executed state.highstate on all of them in a for loop and redirected the output to a file, so I get output in a file per minion format, which easy for me to identify which minion failed so that I can go and fix them individually.
18:50 saurabhs now when I execute state.highstate in for loop it does that individually on each minion, here I see that some minion don't report any data back in the output, that means the state.highstate didn't ran on them, in such cases minion logs don't show any activity even in debug output.
18:50 saurabhs So there is definitely a different between salt '*' state.highstate and doing state.highstate on all of them individually in an for loop, which is obvious, first one is probably managed by salt internally and it might be using configured worker thread optimally. but when the in the second part it fails on some of them I am wondering if I should adjust that worker thread possibly to a bigger value and if that would help
18:52 UtahDave yeah, you could try bumping to 100 worker threads and see what happens
18:52 ingwaem joined #salt
18:52 JasonSwindle joined #salt
18:52 JasonSwindle Howdy, all!
18:53 UtahDave JasonSwindle: hey!
18:53 JasonSwindle Just updated to 0.17.0…… and a few of my items stop working.  One of them is the most basic functions
18:53 saurabhs is there a better way for what I am trying to achieve. do "state.highstate on all the minions and identify what all failed "
18:54 UtahDave saurabhs: there are 2 options in your master config that might help.  One will only show failures, and the other will provide terse output
18:54 _ikke_ saurabhs: Salt has something called returners
18:54 UtahDave JasonSwindle: what stopped working?
18:54 JasonSwindle UtahDave:  This 3abb15c90680e7d39d42
18:54 JasonSwindle Sorry, https://gist.github.com/JasonSwindle/3abb15c90680e7d39d42
18:55 jcockhren JasonSwindle: your state looks just like mine
18:55 jcockhren ;)
18:55 toofer joined #salt
18:55 Ryan_Lane saurabhs: rather than running them in a for loop, you could change the output to json output and pipe the output into a python script that parses the json and outputs to a file
18:56 jcockhren JasonSwindle: on any case, even with 0.16.4 I've had a hard time getting pkgrepo to handle newrelic correctly
18:56 Ryan_Lane then you keep parallel execution while still being able to get proper file output
18:56 JasonSwindle It was working great on 0.16.4
18:56 jcockhren I had to revert to templating the config file
18:56 JasonSwindle Sorry, I only was up to 0.16.3
18:56 JasonSwindle I just from 0.16.3 to 0.17.0
18:57 UtahDave JasonSwindle: any way you can test on the develop branch?
18:57 JasonSwindle Sure.
18:57 JasonSwindle Let me set and fire.  Give me about 8 minutes to spin up the new ENV
18:58 UtahDave cool. thanks, JasonSwindle
18:59 ingwaem Greetings everyone…was just running salt on a mac vm and got this: Invalid argument (bundled/zeromq/src/stream_engine.cpp:95)…anyone seen this before? Note: I have had good all round stability for the most part, vm's stay connected for days at a time, however the odd occasion there are hiccups and this appears to be one of those times
18:59 saurabhs Ryan_Lane, UthaDave, _ikke_: thanks will try all of them
19:00 ajw0100 joined #salt
19:01 SEJeff_work question
19:01 SEJeff_work why is the docker module called dockerio?
19:02 SEJeff_work the website is docker.io, but the tool is called docker
19:02 UtahDave SEJeff_work: good question!
19:02 eliasp SEJeff_work: there's an issue about that… one moment
19:02 SEJeff_work UtahDave, I hereby request it be changed to simply docker... if that isn't an isuse
19:02 SEJeff_work gotcha
19:04 eliasp https://github.com/saltstack/salt-cloud/issues/615#issuecomment-25588516
19:05 eliasp SEJeff_work: ^^
19:07 pentabular joined #salt
19:12 SEJeff_work eliasp, Makes sense, thanks
19:12 ajw0100 joined #salt
19:14 giantlock joined #salt
19:15 deedubs they are having a self-image crisis
19:15 deedubs on ubuntu its lxc-docker
19:16 rgbkrk Hey is anyone in here submitting talks for SaltConf?
19:16 eliasp deedubs: for historical reasons… on Gentoo it was lxc-docker, now it's docker
19:16 deedubs the name is randomized between platforms it seems :)
19:17 eliasp deedubs: not really… upstream decided at some point to make Docker universal and not being necessarly dependent on LXC (see issue linked above)… so one distribution after another renames the package according to upstream's decision
19:17 JasonSwindle UtahDave:  That looks to have fixed it
19:17 eliasp deedubs: just seems to take a bit longer for Ubuntu, but that's just as usual ;)
19:17 deedubs i prefer to imagine it as the wild wild west
19:18 eliasp deedubs: feel free to do so ;)
19:18 JasonSwindle left #salt
19:18 JasonSwindle joined #salt
19:19 JasonSwindle UtahDave:  So, if dev works for my issue
19:21 m0hit joined #salt
19:22 UtahDave JasonSwindle: we're going to be cutting a 0.17.1 release.  Probably today, maybe tomorrow
19:22 StDiluted What bugs are being squashed, is there a list?
19:22 JasonSwindle UtahDave:  AWesome.  That is good to hear…..because I need it like Friday. :)
19:26 JasonSwindle jcockhren: Ping
19:27 _ikke_ We currently have a shared hosting server that has a bunch of apache vhosts. They are basically 3 base templates, with some exceptions (extra rules, changed rules and things), and I'm wondering what would be the best way to set that up in salt
19:28 UtahDave StDiluted: just checked. Apparently the list is complete.  Now there are just a couple things that are causing tests to fail
19:29 UtahDave _ikke_: have you looked at the apache module and state?
19:29 _ikke_ I have glanced at it
19:29 SEJeff_work deedubs, At this pace, it will be fixed in Debian in another 4-5 years :D
19:33 _ikke_ I'm looking at this: http://docs.saltstack.com/ref/modules/all/salt.modules.apache.html but it doesn't look very extensive
19:34 jcockhren JasonSwindle: word
19:35 logix812 joined #salt
19:35 mgw any ideas why pillar.data would return data correctly, but pillar.get fails to show a key that is definitely present?
19:35 JasonSwindle Dev fixes it
19:36 jcockhren \o/
19:36 toofer left #salt
19:36 mgw not only that, salt-call pillar.get works as expected
19:37 m0hit_ joined #salt
19:37 acsir joined #salt
19:39 mgw data.raw is empty though
19:39 mgw no errors in master log
19:39 mgw (data.raw is empty when called from the master, but is populated when using salt-call)
19:42 alunduil joined #salt
19:44 srage joined #salt
19:45 felixhummel joined #salt
19:46 pipps joined #salt
19:46 _ikke_ Can you use jinja to include snippets (ie, other files) in managed files?
19:48 NetworkN3rd joined #salt
19:50 UtahDave _ikke_: Yeah, you can.
19:50 UtahDave JasonSwindle: woot!
19:50 _ikke_ UtahDave: Ok, nice
19:51 qba73 joined #salt
19:53 pdayton joined #salt
19:59 pentabular joined #salt
20:00 Teknix joined #salt
20:01 cshuman_ joined #salt
20:03 pentabular1 joined #salt
20:03 Jahkeup joined #salt
20:16 acsir hi, I'm having problem sometimes in understanding the .sls sintax. Do you know any fast&easy way to compile Jinja files in order to double check what I doing?
20:16 cshuman joined #salt
20:17 KyleG joined #salt
20:17 KyleG joined #salt
20:19 acsir /window level
20:19 _ikke_ :-)
20:23 honestly acsir: "salt-call -l debug state.sls somefile.sls test=True"
20:23 honestly iirc
20:24 xmltok joined #salt
20:27 ingwaem joined #salt
20:28 ingwaem greetings basepi, may I please pm you?
20:33 amckinley joined #salt
20:37 acsir honestly: thankyou a lot. It is really helpful
20:38 pipps joined #salt
20:38 racooper I have a question for the saltstack folks.  Are there any plans for a training to be in Texas in the next 10 months?
20:40 UtahDave racooper: I don't know of one planned for Texas right now.  How many people are you inquiring for?
20:41 racooper right now, just myself...possibly up to 3 total
20:41 racooper although if the DC cost is consistent, then just one
20:43 timoguin *cough* Nashville too *cough*
20:43 UtahDave We do onsite trainings as well as regional trainings.  Let me get you in contact with the right person. Can I pm you?
20:48 basepi ingwaem: sorry, was away briefly.  you're welcome to pm me
20:48 NetworkN3rd_ joined #salt
20:48 ingwaem Oh ok np basepi :) pming
20:49 UtahDave Corey: hey, what do you do to autostart salt on Mac osx?
20:50 m0hit joined #salt
20:51 deedubs left #salt
20:52 micah_chatt joined #salt
20:54 m0hit_ joined #salt
20:55 racooper I'm here for a PM sure
20:58 acsir Hi, is there any north italian group about saltstack? Or saltstack's users set in north Italy that would like to join for meeting and talk about your own experiences?
20:59 micah_chatt has 0.17.0 broken the pkgrepo state for anyone using ubuntu?
21:01 UtahDave micah_chatt: can you test against develop?  I think we've got a fix in there.
21:01 micah_chatt UtahDave: I'll try that, thanks.
21:02 cewood joined #salt
21:03 micah_chatt UtahDave: that fixed it. thanks
21:04 Ryan_Lane UtahDave: can I get details about the embargoed CVEs that were leaked?
21:05 sroegner joined #salt
21:06 Teknix joined #salt
21:06 isomorphic joined #salt
21:07 UtahDave Ryan_Lane: Yeah, they've all been fixed in the develop branch.
21:07 UtahDave Let me grab the sec mailing list link for you.
21:07 * Ryan_Lane isn't using the dev branch
21:07 Ryan_Lane ;)
21:07 Ryan_Lane I want to see the details of the CVEs
21:07 Ryan_Lane they aren't public
21:07 Ryan_Lane I need to see how urgent they are to how we're using salt
21:08 ajw0100 joined #salt
21:08 Ryan_Lane also, is there any way we can be added to an early access list?
21:08 UtahDave http://seclists.org/oss-sec/2013/q4/115
21:08 ze- #7712 (recursive through ..) - related to fileserver_backend: [ roots, git ]
21:09 ze- at least that's the only time i manage to reproduce :)
21:09 Ryan_Lane UtahDave: yeah, but there's no details there
21:10 Ryan_Lane I really want to see the details of CVE-2013-4437
21:10 amahon joined #salt
21:10 Ryan_Lane we strictly limit salt use to salt, so the others don't worry me so much
21:10 Ryan_Lane err
21:10 Ryan_Lane to root
21:11 UtahDave Hm. I don't have access to the CVEs myself.  Let me grab someone here
21:11 pipps1 joined #salt
21:11 ze- (ok, 7712 on just git, with include_empty: True reproduce the bug too)
21:11 UtahDave Oh, wait. Ryan_Lane, 4437 is only when used with salt-ssh
21:12 UtahDave ze-: I've been tasked with 7712 as well. That will be fixed in 0.17.1
21:12 ze- UtahDave: you already have a patch? or a better idea where/what it is ?
21:12 Ryan_Lane UtahDave: oh?
21:13 donatello joined #salt
21:13 amahon joined #salt
21:13 Ryan_Lane UtahDave: oh, I also need details for CVE-2013-4439
21:14 ajw0100 joined #salt
21:14 ze- (well, 7712 / 7404 ... but " No one is assigned " on github :)
21:14 Ryan_Lane or is that https://github.com/saltstack/salt/pull/7356 ?
21:14 testuser151 joined #salt
21:14 Ryan_Lane ah. it is
21:15 donatello has anyone had success with ec2 driver in salt-cloud with saltmaster version 0.17.0? I see that event message format change to dict is creating ValueErrors while launching an instance
21:15 UtahDave Ryan_Lane: ah, yeah.  4439 only affects the Salt Mine.
21:15 SEJeff_work UtahDave, would it make sense to have a security@saltstack.com email addy?
21:16 SEJeff_work And for at the very least Kurt to know?: http://seclists.org/oss-sec/2013/q4/114
21:17 ingwaem thanks guys :) bbl
21:17 UtahDave SEJeff_work: Yes, it would make sense.  We've had security@saltstack.com for quite a while now. about a year I think
21:17 steplunge joined #salt
21:20 josephholsten joined #salt
21:21 amckinley joined #salt
21:21 testuser151 left #salt
21:22 diegows joined #salt
21:24 xmltok is there a list of the available variables if i am using python in my sls file?
21:26 pmrowla joined #salt
21:26 amckinley joined #salt
21:28 shinylasers joined #salt
21:29 cwright any timeframe for 0.17.1 to be cut?  some of these 0.17.0 issues that have already been solved are becoming problematic (ex: pkg.compare)
21:31 bastion2202 joined #salt
21:32 UtahDave cwright: today or tomorrow.  It's priority #1 right now
21:32 cwright UtahDave: thanks, really looking forward to it
21:34 steplunge Hi; newbie questions, sorry. I'm having trouble with the root_dir setting on the master. Even after I set it I still get errors because salt is trying to write to /var/log/salt/master.
21:34 ajw0100 joined #salt
21:34 seanz So, if I had a software system that had an upgrade process that differed from the main install procedure for that software system, what is the best way to tell salt to run a set of "upgrade" steps? Is this a solved problem?
21:34 seanz All I can think of is to modify the original salt state and re-run highstate, which might not be safe to do.
21:35 cwright seanz: look at cmd.wait.  you could change several of these together via watchers
21:36 seanz cwright: So I would watch something like a git.latest for changes in code, for example, and then have the commands run at that time?
21:36 juicer2 left #salt
21:36 cwright yup
21:36 seanz In essence, the upgrade process would be "built-in" to a highstate by the fact that changes would trigger the right steps...
21:36 cwright in fact i think that is the example used on the cmd.wait doc
21:36 seanz Interesting.
21:37 seanz Are highstates meant to be idempotent?
21:37 seanz Should they be designed that way?
21:37 UtahDave seanz: you could also put together a custom module to execute those upgrade steps. Kind of like this:  https://gist.github.com/UtahDave/6125409
21:37 UtahDave seanz: Yes, highstates are idempotent
21:38 seanz UtahDave: Interesting.
21:38 oz_akan_ joined #salt
21:38 JasonSwindle oz_akan_:  !
21:42 bdenning joined #salt
21:43 bdenning joined #salt
21:44 bdenning Hey all. I've used the salt openstack-standalone formula to provision a test vm with openstack installed, however, I'm at a loss as to where to find the username/password details to now log in to horizon
21:44 bdenning Has anyone used this formula before? Any idea what the default username and password might be for logging in to horizon?
21:45 kermit joined #salt
21:46 Gareth bdenning: usually openstack nodes use public/private ssh keys for logging in.
21:47 jbunting joined #salt
21:47 jesusaurus bdenning: is the formula based on devstack?
21:47 UtahDave bdenning: also, are those set in the pillar data?
21:48 bdenning The formula I'm using is: https://github.com/saltstack-formulas/openstack-standalone-formula
21:48 bdenning I haven't set any pillar data as yet
21:49 elfixit joined #salt
21:49 bdenning I was hoping that it might create a default account for logging in to the dashboard. I get presented with a username/password prompt when browsing to http://localhost/dashboard, but none of the usual suspects (admin/admin) seem to be working.
21:50 pipps joined #salt
21:50 bdenning The formula itself seems to have worked perfectly by the way. I was hoping to use to in order to show off saltstack to the other guys at work today. Just need to complete the provisioning by figuring out how to log in to openstack afterwards ;-)
21:51 jesusaurus wow, that pillar.example file is very lacking
21:52 UtahDave bdenning: OK, I just chatted with the guy who wrote that formula.
21:53 ze- UtahDave: gitfs.file_list ignores prefix. Should it be handled there (by each fs), or by the caller (like fileserver.list_file directly)?
21:53 UtahDave he got an embarrassed smile on his face and said that that is as far as he got on that formula.  He thinks that is managed in Keystone
21:53 jpcw joined #salt
21:54 bdenning Haha, no problem. I might spend some time figuring it out myself and will send through a pull request if I come up with an answer.
21:54 josephholsten joined #salt
21:54 bdenning Thanks everyone
21:54 UtahDave bdenning: That would be very much appreciated!  Thanks
21:57 UtahDave ze-: I'm having difficulties reproducing that file.recurse problem.  have you seen it yourself?
21:57 ze- yeah
21:57 ze- having it right now
21:57 ze- need to be on git
21:57 ze- as roots handles prefix
21:58 UtahDave yeah, I'm on the latest git
21:58 ze- if you give me hints, you'll have a pull request by the end of the night :)
21:58 Jahkeup_ joined #salt
21:58 ze- I mean fileserver_backend: [ git ]
21:59 UtahDave Oh, this doesn't happen using the regular fileserver?
21:59 ze- nope.fileserver/roots.py handles prefix when listing files
21:59 ze- nope. fileserver/gitfs.py doesn't
22:00 ze- and i bet some other FS don't either
22:00 zooz joined #salt
22:00 UtahDave Ah. Nobody mentioned that in either of the issues they opened
22:01 ze- UtahDave: well, I kinda *need* the fix, so I'm working on it :)
22:01 UtahDave ok, fileserver/roots.py handles it, huh?  Do you think it would be better to handle it at the higher fileserver level?
22:01 ze- better to handle it in each fs, if they can get better performances by using it.
22:02 ze- but at higher level, to avoid bad-fs behaviour
22:02 jumperswitch_ joined #salt
22:05 UtahDave let me ask Tom what he thinks, real quick
22:06 elfixit joined #salt
22:06 fivethreeo joined #salt
22:08 mwillhite joined #salt
22:09 emilisto joined #salt
22:10 cwright joined #salt
22:10 Jahkeup joined #salt
22:11 mjulian joined #salt
22:11 UtahDave ze-: where is fileserver/gitfs.py checking the prefix?
22:11 ze- UtahDave: gitfs is NOT checking the prefix
22:11 morganfainberg joined #salt
22:11 ze- rootfs use the prefix to walk from there
22:11 UtahDave sorry,  fileserver/roots.py
22:12 UtahDave which line?
22:12 ze- for root, dirs, files in os.walk(os.path.join(path, prefix), followlinks=True):
22:12 UtahDave line number?
22:13 ze- 186, 212 and 233
22:13 basepi joined #salt
22:13 UtahDave awesome, thanks
22:13 crane joined #salt
22:14 ze- http://pastebin.com/tH9NY5fJ
22:14 ze- that's the fix I've produced, and works on my example.
22:14 ze- doing it directly at fileserver level.
22:15 ze- If you think it might be ok, i'll get it on a pull request
22:15 UtahDave that's where we're thinking it should be done.
22:15 UtahDave Yeah, please send a pull request. that would be great
22:15 ze- ok. now that I have the working patch, shouldn't take too long :)
22:15 ze- any comment about the way I do it ?
22:16 ze- (the 3 lines in the patch :)
22:16 pentabular joined #salt
22:17 madduck_ joined #salt
22:17 madduck_ joined #salt
22:18 AdamSewell joined #salt
22:20 UtahDave ze-: yeah, just send a pull req on github
22:21 ze- I meant comment about the code :)
22:21 UtahDave Tom thinks that should be fine.
22:23 g4rlic left #salt
22:31 alunduil joined #salt
22:31 cshuman joined #salt
22:32 swa_work joined #salt
22:33 ze- there. a pull request, including fix on the 4th function that was not there for my files :)
22:34 UtahDave thanks!  We'll look it over right now
22:35 * honestly grumbles about his ignored gitlabhq PR
22:36 UtahDave ha ha, honestly. I rushed to look at our salt pr list for your pr.
22:37 ze- UtahDave: just to check, any risk a new version appears on http://debian.saltstack.com/debian without a fix for that bug ?
22:37 UtahDave 0.17.1 will be there with your fix.
22:37 ze- good. means i can apply the fix on my servers, and not have to worry an upgrade will mess with the fix :)
22:39 Thiggy joined #salt
22:39 UtahDave cool, ze-
22:41 JasonSwindle joined #salt
22:45 cshuman joined #salt
22:46 cshuman joined #salt
22:46 * ze- sighs. lucky I have a list of stupidly created directory. time to clean a few boxes.
22:47 JasonSwindle joined #salt
22:48 oz_akan_ joined #salt
23:00 UtahDave ok, ze-.  merged.   thanks for working on that fix!
23:01 ze- hitted the bug today, and I just can't leave it... so I had to get it fixed :P
23:02 acsir mayby it could be a dump question, but that's the metter: I fin really cool and useful the grains, but really boring the idea to have to set up them on the minion. What if I would assign the grains using pillars and salt state file?
23:03 acsir does anybody do this?
23:04 ze- on what base would you generate the value ?
23:07 acsir my idea is to set a basic salt state file that go to write in the /etc/salt/grains of the minion. The paramater of the configuration would be set up calling the value stored in pillars assing to each machine name corrispondent value for each grain.
23:07 VertigoRay joined #salt
23:07 pmrowla joined #salt
23:08 ze- why set a grain with a value you already have in pillars ?
23:09 VertigoRay Hello.  Anyone familiar with Debian 7.0?  Followed simple instructions in salt docs, to add saltstack's deb repo.  Unfort, I get and error when installing salt-minion.  Error:  Package 'init-system-helpers' has no installation candidate
23:09 VertigoRay fresh, clean install of debian with just SSH installed from the CD, btw.
23:10 ze- what do you have in your source list ?
23:10 akasto joined #salt
23:11 VertigoRay I do.  Added the lines for wheezy stable:  http://docs.saltstack.com/topics/installation/debian.html
23:11 VertigoRay oh, lemme cat those out
23:11 cshuman joined #salt
23:11 ze- paste your whole sourcelist somewhere like http://pastebin.com/ :)
23:11 acsir I was thinking about this eg.: I have on each machine a grain called role (sshd/apache/production/testing/fileserver...). I find it quicker to set it on the master inster of going to write it on the minion. Maybe I'm thinging this way just because I don't have so many server to deal with.
23:12 ze- acsir: if you want to use pillar, why *also* set a grain?
23:12 ze- wherever you need to access it, just access to pillar.get('roles', []) :)
23:12 VertigoRay ze-: turns out a quick eyeball there and there's a typo ... /sigh
23:13 ze- VertigoRay: glad asking helped ;)
23:14 VertigoRay more often than not, formulating the question and explaining to someone (in detail) how I did things makes me realize what I did.
23:14 * ze- nods. that's why it's useful to ask details, even when you have no idea what it is about ;)
23:15 VertigoRay that fixed it btw ... /grin
23:15 ze- (though, here I could have compared to some of my boxes :)
23:17 acsir ze: I thing you're right. Probably it is just because the grains interface seems to me more usable then the pillars' one in targeting the computers.
23:18 ze- well, if you *really* want all pillars as grains, I guess you could write a _grain/ module that exposes all pillars
23:18 ze- (haven't tested such)
23:18 acsir yes, that's a good point.
23:18 Ryan_Lane joined #salt
23:19 ze- at $WORK, we have a webservice that let each host get informations about its supposed roles :)
23:19 ze- and a grain module that query it
23:19 ze- not sure it's the best way, but it seems to work so far.
23:19 acsir good, that is what I was looking for
23:20 joshe as long as the information for a host never changes, sure
23:20 joshe otherwise you'd probably want to use a pillar
23:20 joshe grains are heavily cached, under the assumption that they never change
23:20 Ryan_Lane UtahDave: cool. thanks for the CVE info.
23:21 Ryan_Lane seems we're unaffected by any of them :)
23:22 acsir it was supposed to store static information (like the type of role in the organization, the services that are suppose to run on it, etc.)
23:22 UtahDave Ryan_Lane: Yeah.  I'm not sure if you got my last message, but the last one only matters when dealing with the Salt Mine
23:22 Ryan_Lane yep. did
23:22 UtahDave cool.
23:22 acsir ze: but what kind of grain module is used ad $WORK?
23:22 acsir at *
23:23 acsir ze-
23:23 amckinley joined #salt
23:25 ze- acsir: "home-made" module that query the webservice, and set the roles (and other stuffs) on each minion
23:26 cshuman_ joined #salt
23:27 Teknix joined #salt
23:27 ze- something like http://pastebin.com/NwK4e0g0
23:27 ze- webservice answering directly with valid yaml, and not much checks about it
23:28 cshuman joined #salt
23:28 ze- anyway... i should get myself some sleep.
23:28 Ryan_Lane1 joined #salt
23:28 ze- g'night
23:28 acsir thank you
23:28 Ryan_Lane joined #salt
23:29 acsir that was a really interesting introduction
23:29 acsir night
23:29 acsir btw it seem that what I was wondering is not so uncommon: https://www.google.com/search?q=grains+vs+pillars
23:31 rgbkrk joined #salt
23:34 saurabhs1 joined #salt
23:39 bhosmer joined #salt
23:41 billh joined #salt
23:45 cshuman joined #salt
23:45 billh Every so often one or more of my Salt Minions will stop responding to requests from the master.  I'm able to fix it by restarting the salt-minion service.  I'm using the v0.16.0 RPMs from EPEL.  Does anyone else run into this issue?
23:47 cshuman joined #salt
23:48 kula i have that issue occasionally on machines with zmq 2. running `salt-run manage.down' every 15 minutes in a cron job seems to have fixed that
23:48 diegows joined #salt
23:48 cshuman joined #salt
23:49 billh kula: What version of zmq works well for you?
23:51 Khazix joined #salt
23:51 kula 3 whatever. basically when i move from centos 5.whatever to centos 6.4 things get much happier
23:52 billh kula: Good to know.  That's basically where I'm at right now, too.  Thanks.
23:53 pipps1 joined #salt
23:56 UtahDave billh: yeah, you want to make sure your minions are on zmq 3.2.x
23:56 UtahDave or better
23:58 mgw joined #salt

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