Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-11-10

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

All times shown according to UTC.

Time Nick Message
00:04 alexr joined #salt
00:21 dooshtuRabbit joined #salt
00:22 rojem joined #salt
00:35 snuffeluffegus joined #salt
00:41 diegows joined #salt
00:47 n8n joined #salt
00:50 monkey66 left #salt
00:50 mosen joined #salt
00:51 bmonty joined #salt
00:52 monkey66 joined #salt
00:53 batang_munti joined #salt
00:53 monkey66 left #salt
00:56 mafrosis joined #salt
00:57 Pixionus joined #salt
00:57 mafrosis lo peeps. I wonder if someone could give me a hand figuring out a “Specified SLS ‘blah’ in environment 'base' is not available on the salt master”
00:57 mafrosis I’ve tried running in debug, and nothing is made apparent in the logs
01:01 mafrosis ok, different question
01:01 aqua^mac joined #salt
01:01 mafrosis how can I get a breakpoint in the salt-master code, so I can debug interactively whilst running foreground?
01:01 bhosmer joined #salt
01:06 troyready joined #salt
01:17 thayne joined #salt
01:20 thawes joined #salt
01:24 to_json joined #salt
01:27 ingwaem joined #salt
01:31 n8n joined #salt
01:33 bhosmer joined #salt
01:33 delinquentme joined #salt
01:37 possibilities joined #salt
01:37 mafrosis joined #salt
01:43 atbell joined #salt
01:44 hasues joined #salt
01:45 hasues left #salt
02:04 Nexpro joined #salt
02:11 dalexander joined #salt
02:12 malinoff joined #salt
02:15 tristianc joined #salt
02:17 elfixit1 joined #salt
02:20 thayne joined #salt
02:22 dalexander joined #salt
02:40 iwishiwerearobot joined #salt
02:46 ilbot3 joined #salt
02:46 Topic for #salt is now Welcome to #salt | SaltConf 2015 Call for Speakers is open! http://saltconf.com/call-for-speakers/ | 2014.1.13 is the latest | Help us test the 2014.7 RC! http://bit.ly/salt-rc | 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/
02:55 tristianc joined #salt
03:17 gmcwhistler joined #salt
03:22 bhosmer joined #salt
03:23 Mso150 joined #salt
03:25 TyrfingMjolnir joined #salt
03:30 Eugene joined #salt
03:31 UForgotten_ joined #salt
03:38 rcsheets joined #salt
03:46 otter768 joined #salt
03:47 grepory joined #salt
04:12 thayne joined #salt
04:25 otter768 joined #salt
04:55 ajolo joined #salt
05:10 bhosmer joined #salt
05:13 ndrei joined #salt
05:22 cmthornton joined #salt
05:22 pipps joined #salt
05:27 ysiad joined #salt
05:31 Reiner0310 joined #salt
05:48 felskrone joined #salt
05:55 ramteid joined #salt
06:06 n8n joined #salt
06:15 ajolo joined #salt
06:18 iwishiwerearobot joined #salt
06:19 Furao joined #salt
06:28 ysiad joined #salt
06:37 ITChap joined #salt
06:39 Outlander joined #salt
06:44 ITChap joined #salt
06:45 thayne joined #salt
06:46 ITChap joined #salt
06:49 oyvjel joined #salt
06:51 shookees joined #salt
06:55 ITChap joined #salt
06:56 saltuser left #salt
07:00 bhosmer joined #salt
07:02 ITChap joined #salt
07:08 ITChap joined #salt
07:12 colttt joined #salt
07:19 ITChap joined #salt
07:21 zach joined #salt
07:22 ITChap joined #salt
07:25 miqui_ joined #salt
07:26 JlRd joined #salt
07:28 ITChap joined #salt
07:28 ysiad joined #salt
07:35 intellix joined #salt
07:37 ITChap joined #salt
07:41 ITChap joined #salt
07:44 b1nar1_ joined #salt
07:48 ITChap joined #salt
07:50 Karunamon joined #salt
07:53 ramishra joined #salt
07:56 linjan joined #salt
07:59 oyvjel1 joined #salt
07:59 delinquentme joined #salt
08:05 \ask joined #salt
08:06 trikke joined #salt
08:07 akafred joined #salt
08:07 eseyman joined #salt
08:08 mieciu joined #salt
08:09 saravanans joined #salt
08:11 mieciu Hi guys, I'm deploying my app on centos6.4 with vagrant, using masterless saltstack. What i've found strange is that bootstrap scripts adds https peel repositories URLs, which are wrong and they need to be changed to http after first run - or maybe I did something wrong?
08:14 ghanima joined #salt
08:16 cofeineSunshine joined #salt
08:16 mieciu *i meant "https EPEL repositories URLs"
08:17 lcavassa joined #salt
08:18 harkx joined #salt
08:20 tmh1999 joined #salt
08:21 gioumars joined #salt
08:22 slav0nic joined #salt
08:24 paha joined #salt
08:28 gioumars hello everyone. Is there any way to get the values of a list, which is nested in from two dictionaries, as individual values?
08:28 gioumars here i have my pillar and state
08:28 the_drow joined #salt
08:28 gioumars http://justpaste.it/hx1j
08:29 the_drow Good morning
08:29 the_drow Is there a way to watch/react to a mounting of a volume with salt?
08:29 ITChap joined #salt
08:29 ysiad joined #salt
08:30 honestly gioumars: without looking, {{ " ".join(pillar['bar']['foo']) }}
08:30 honestly gioumars: if that's what you mean by individual values
08:31 flyboy joined #salt
08:32 JlRd joined #salt
08:32 honestly of course you can also do {% for itm in pillar['bar']['foo'] %} - {{ itm }} {% endfor %}
08:35 honestly the_drow: you'll have to watch for the mounting yourself and then call salt-run
08:35 gioumars honestly: my problem is that when i get results from a list which has more than one item, it returns them as array but is it possible to have each result within a list as a seperate value and not as a "bind" array result
08:35 Mso150 joined #salt
08:36 jhauser joined #salt
08:36 honestly what do you mean by separate value? where do you want the values in the list to go?
08:38 dRiN joined #salt
08:38 the_drow honestly: How do I do that?
08:39 alexr joined #salt
08:40 gioumars i have a pillar, as shown in my ;ink above, which has a dictionary and subdictionary. In that subdictionary there is a list with e-mails. I want to create a template and insert those e-mails on the template with comma seperations. I find it diffircult to manipulate the address as i wish because if the list in pillar has more than one value , the values cannot be manipulated seperetately
08:40 gioumars put it mpre simple
08:40 gioumars pillar:
08:40 gioumars -value1
08:40 gioumars -value2
08:40 gioumars -value3
08:40 gioumars it will be returned as : [value1,value2,value3 ]
08:41 gioumars is it possible to treat an use value1 seperately from the array?
08:42 honestly gioumars: yes, just use normal array indexing - [0] and [1:]
08:46 bhi joined #salt
08:46 honestly the_drow: well, not with salt. i mean, you could run a cronjob to check once a minute whether the volume is there.
08:47 honestly gioumars: if you just want to join an array with commas, I already posted a snippet that does exactly what you need above
08:47 rjc joined #salt
08:48 bhosmer joined #salt
08:50 Mso150 joined #salt
08:54 wvds-nl joined #salt
08:56 ITChap joined #salt
08:57 gioumars hoestly: thanks alot for the help, i am finfing it diffircult because in the dicrtionary, some values come from lists that have a single e-mail and others have more than one which are returned differently
08:57 gioumars for example
08:57 gioumars list:
08:57 gioumars - value1
08:58 gioumars is returned as a single value
08:58 gioumars but
08:58 gioumars list2:
08:58 gioumars - email1
08:58 gioumars - email2
08:58 gioumars is returned as : [email1, e-mail2]
08:59 gioumars so i wanna join those cases
08:59 honestly gioumars: ah, I see, I guess that's a sucky feature of jinja
09:00 przemek joined #salt
09:00 gioumars yeap, i just wanted to be also clarified by more experts before i start "bash 'ing' " in order to sort that
09:00 honestly gioumars: I don't know how to deal with that off-hand, probably type-matching
09:00 gioumars in there a way to insert value1 to email array in my example above?
09:00 honestly huh?
09:01 ITChap joined #salt
09:02 honestly uh, if you have a list and want the list with a value appended, you can do list + [value]
09:03 gioumars ok thanks
09:03 gioumars honestly: thanks alot m8
09:03 honestly yw
09:04 honestly in general you can have pretty much arbitrary pyhon expressions inside a jinja template
09:06 jdmf joined #salt
09:06 gioumars do you know how can i remove the last character from the results?
09:06 gioumars i get an extra comma at the end which i wash to delete it
09:07 honestly dude, just use the join method
09:07 honestly it does that automatically for you
09:07 gioumars does it?
09:07 gioumars :-S
09:07 honestly yes.
09:07 gioumars on me way then
09:08 honestly ", ".join(['foo', 'bar']) give 'foo, bar'
09:08 honestly gives*
09:13 nethershaw joined #salt
09:18 oyvjel joined #salt
09:19 gioumars ok thats wird
09:19 gioumars it splits all letters of e-mail with comma
09:19 canci joined #salt
09:19 gioumars lol
09:20 saravanans joined #salt
09:20 gioumars but i did it with pipe
09:20 Jimlad_ joined #salt
09:20 gioumars list | join(', ')
09:21 iwishiwerearobot joined #salt
09:22 honestly yes. if you give join() a string, it will join up the chars in the string, since a string is a list of chars
09:23 N-Mi_ joined #salt
09:23 N-Mi_ joined #salt
09:24 babilen Or rather: It treats everything you hand it as an iterable and simply joins its elements. If you iterate over a string you get each character.
09:29 gioumars ok so how do you join in my case?
09:29 gioumars if you have:
09:29 gioumars lst1:
09:29 gioumars - value1
09:30 gioumars - value2
09:30 ysiad joined #salt
09:30 gioumars ", ".join(['value1', 'value2'])   ?
09:31 jrluis joined #salt
09:31 honestly that should give you 'value1, value2'
09:35 Ixan is there some way to define multiple environments with several gitfs_remotes, without having to use branches?
09:36 Ixan i.e., separate git repos per environment
09:36 gioumars ok
09:36 gioumars so far so good
09:37 Twiglet joined #salt
09:37 gioumars but how do you join with coma if you get your values using a for loop from pillar in a template?
09:42 alexr joined #salt
09:42 babilen gioumars: You can't really. But why use a for loop if you can refer to the list directly?
09:43 babilen jinja is not very powerful and salt should have gone with mako as default rendering engine as that would have made it much easier to write "slightly more complicated" code without having to use a non-default renderer or writing pure Python to begin with
09:43 monkey661 joined #salt
09:44 saravanans joined #salt
09:47 Jimlad joined #salt
09:47 ITChap joined #salt
09:51 alanpearce joined #salt
09:52 wnkz joined #salt
10:00 brayn joined #salt
10:01 favadi joined #salt
10:09 shiva_blre joined #salt
10:09 shiva_blre Hi
10:10 shiva_blre Am installing RPMs on Minions (Centos 6.6) via salt-master (Centos 6.6). Its installing packages on minions but its taking too much of time.
10:11 saravanans joined #salt
10:12 Reiner0310 joined #salt
10:13 shiva_blre to install a package (Ex: package-abc) with more number of dependencies say with 10 dependences (Total 11 packages), it took aroung 32 minutes to install.
10:13 saravanans left #salt
10:13 shiva_blre Installation done on a single minion.
10:15 ysiad joined #salt
10:17 shiva_blre Hi anyone there ?
10:20 shiva_blre Hello ?
10:20 shiva_blre Am installing RPMs on Minions (Centos 6.6) via salt-master (Centos 6.6). Its installing packages on minions but its taking too much of ti
10:20 shiva_blre to install a package (Ex: package-abc) with more number of dependencies say with 10 dependences (Total 11 packages), it took aroung 32 minutes to install
10:21 shiva_blre Installation done on a single minion.
10:21 shiva_blre Hi anyone there ?
10:22 viq shiva_blre: and what would be the time if you called that command by hand on the minion?
10:24 shiva_blre If its downloading from Internet Repository, it takes 3 to 5 minutes.
10:24 colttt joined #salt
10:26 shiva_blre But if its a local repo it takes within 1-2 minutes.
10:27 viq next thing I would try is with salt-call on the minion, probably with -l debug, and that way see what's going on
10:34 Reiner0310 joined #salt
10:35 Reiner0311 joined #salt
10:35 shiva_blre Ok
10:36 shiva_blre @viq: Ok but Whats the best practice to install RPMs on minions ? Can I make this through salt-stack ?
10:37 bhosmer joined #salt
10:40 viq shiva_blre: I don't understand the question
10:41 shiva_blre I have decided to install RPMs via salt.
10:41 shiva_blre Is that a best practice to do.
10:42 shiva_blre Usually they will install RPMs thorugh PXE server. But via salt is that a best practice to do.
10:42 shiva_blre ?
10:42 viq Salt is configuration management. CM's task is to make sure machine is in a known state. Part of that is ensuring, that certain packages are (or are not) installed.
10:42 shiva_blre Ok
10:43 viq I see no reason not to install those packages when provisioning a machine, and then just make sure they are there using salt. End result is the same - you are certain, that the packages are installed.
10:45 babilen *sigh* -- I have to deploy phpmyadmin on some Debian boxes and am not entirely sure how to do that via salt. Does anybody have an example state/formula for that?
10:46 viq babilen: isn't there a package?
10:46 viq babilen: also, I think chive is now recommended instead
10:46 giantlock joined #salt
10:46 babilen There is, but it requires some configuration which you would normally do by answering questions in "dpkg-reconfigure -plow phpmydmin"
10:47 babilen Ah, let me take a look at chive
10:47 shiva_blre Yes, am sure that the packages are installed. I checked on the minion side.
10:47 viq babilen: install on a test box, generate config, put in salt?
10:47 shiva_blre why its taking too much of time.
10:48 babilen viq: I might do that, but I simply hoped that somebody else had come up with a solution for this already :)
10:48 viq babilen: no knowledge here, so I'll stop running my mouth ;)
10:48 babilen sure
10:51 diegows joined #salt
10:54 k3ys3r joined #salt
11:01 harkx joined #salt
11:06 bhosmer joined #salt
11:18 cmthornton joined #salt
11:20 fredvd joined #salt
11:22 ndrei joined #salt
11:27 ndrei joined #salt
11:32 wnkz joined #salt
11:39 __gotcha1 joined #salt
11:40 __gotcha1 Is there a way to query which environments are known by the salt-master
11:41 ndrei joined #salt
11:41 __gotcha1 to check salt-master that my configuration has been understood the way I meant it ?
11:41 __gotcha1 oops
11:41 __gotcha1 to check that salt-master loaded my configuration the way I meant it
11:46 canci joined #salt
11:47 aqua^mac joined #salt
11:54 bahadir joined #salt
12:06 the_drow honestly: I can use inotify-wait on /proc/mounts and have it fire a reactor event :)
12:07 honestly the_drow: yeah, that should work fine.
12:07 the_drow I see that salt don't have a state for inotify tools yet
12:08 the_drow does anyone know if LVM locks the drivers it joins into a logical driver? Can I still write to any of them?
12:11 __gotcha joined #salt
12:11 darrend_ joined #salt
12:11 elfixit joined #salt
12:13 ysiad joined #salt
12:15 ramishra joined #salt
12:20 ndrei joined #salt
12:26 bhosmer joined #salt
12:30 canci joined #salt
12:30 al joined #salt
12:38 diegows hi
12:39 diegows I want to move some config variables that are used by more than one sls to a common.sls. Pillar is not an option because some if this variables are generated dinamycally (using mine)
12:39 diegows I've tried with include, but jinja is complaining about variables that doesn't exist
12:41 scottpgallagher joined #salt
12:42 diegows {%- import 'ceph/settings.sls' as ceph %}
12:42 diegows that works, but I have to use the ceph. prefix for every var
12:42 diegows not serious, but I'd like to avoid it :)
12:47 JlRd joined #salt
12:47 mgw joined #salt
12:51 fredvd joined #salt
12:52 shyam_ joined #salt
12:54 CeBe joined #salt
12:57 thawes joined #salt
12:57 active8 joined #salt
13:00 viq diegows: I think dropping "as ceph" from it would fix that - if it would let you that is.
13:04 jeddi joined #salt
13:05 mtk joined #salt
13:05 __gotcha joined #salt
13:06 johanek joined #salt
13:07 mtk Hi everyone! With following  state file: http://pastebin.com/Q7gzAEL0, can I somehow change it so that salt will install it with easy_install, not yum/apt ?
13:07 mtk with puppet I could easily specify "provider => pip" property for a resource
13:07 johanek hello, is it possible to set salt-call to use a returner by default?
13:08 diegows viq, thanks!
13:08 diegows I'll try later, now it's working and the prefix is not serious :)
13:08 diegows mtk, use pip state module
13:08 johanek mtk: pip.installed rather than pkg installed?
13:09 diegows pkg use apt/yum
13:09 diegows always, not pip
13:11 __gotcha joined #salt
13:11 mtk I've searched for easy_install
13:11 zooz joined #salt
13:11 mtk i'm checking this out
13:11 mtk thanks!
13:12 thayne joined #salt
13:15 babilen mtk: You really want to use pip in lieu of easy_install
13:15 babilen mtk: What's wrong with the packages?
13:15 SpX joined #salt
13:16 cDR joined #salt
13:17 mtk i need latest version of supervisor
13:17 mtk which is unavailable in cents 6.4 repos
13:17 mtk *CentOS
13:22 Setsuna666 joined #salt
13:25 jaimed joined #salt
13:26 saravanans joined #salt
13:28 linjan joined #salt
13:30 bersace Hi, what's the difference between states_dirs and file_roots ?
13:36 aqua^mac joined #salt
13:37 babilen bersace: states_dirs is optional IIRC, no idea why it exists though
13:41 coolbsd joined #salt
13:46 _prime_ joined #salt
13:47 fredvd joined #salt
13:53 tmh1999 joined #salt
13:53 bersace babilen: thanks
13:59 viq mtk: https://build.opensuse.org/search seems to have some packages for supervisor
14:00 Ahlee I wonder if that allows for specifying of additional states outside of the fileserver
14:01 viq http://docs.saltstack.com/en/latest/ref/configuration/minion.html#states-dirs
14:02 viq interestingly it appears to be a minion config, not a master config directive
14:02 Ahlee ah
14:02 Ahlee for materless most likely, then
14:04 cpowell joined #salt
14:06 teebes joined #salt
14:07 yetAnotherZero joined #salt
14:10 aparsons joined #salt
14:10 j-saturne joined #salt
14:10 alexr__ joined #salt
14:11 ramishra joined #salt
14:11 peters-tx joined #salt
14:12 aparsons_ joined #salt
14:12 gngsk joined #salt
14:13 babilen I really have no idea what that would be used for ...
14:13 babilen I mean it was obviously meant for, well, additional state directories, but why wouldn't one add those to file_roots to begin with?
14:14 j-saturne joined #salt
14:15 bhosmer joined #salt
14:19 nicksloan joined #salt
14:20 racooper joined #salt
14:20 miqui joined #salt
14:20 viq I wonder if top.sls found in states_dirs be considered or not
14:21 viq s/be/would be/
14:21 viq Because if not, then that could be one reason.
14:22 rojem joined #salt
14:29 __gotcha joined #salt
14:32 saravanans joined #salt
14:33 babilen mayhaps
14:36 tristianc joined #salt
14:37 saravanans joined #salt
14:38 ramishra joined #salt
14:41 bhosmer joined #salt
14:42 ajolo joined #salt
14:46 saravanans joined #salt
14:46 vbabiy joined #salt
14:47 tmh1999 joined #salt
14:51 gmoro joined #salt
14:51 karimb joined #salt
14:51 saravanans joined #salt
14:53 nitti joined #salt
14:53 superted666 Hey guys & girls, has 2014.7.0 been seeded to EPEL-Testing yet?
14:53 karimb joined #salt
14:53 saravanans joined #salt
14:53 viq superted666: apparently so
14:54 superted666 Thanks viq i'll start planning the upgrade then!
14:57 djaime joined #salt
15:00 srage joined #salt
15:01 nitti joined #salt
15:01 housl joined #salt
15:03 alreece45 joined #salt
15:03 j-saturne joined #salt
15:03 alexr joined #salt
15:05 saravanans joined #salt
15:07 iggy superted666: read the release notes... there are a few known regressions, etc.
15:08 SheetiS joined #salt
15:08 j-saturne joined #salt
15:09 TheRealBill joined #salt
15:13 seydu joined #salt
15:13 superted666 iggy: Reading http://docs.saltstack.com/en/latest/topics/releases/2014.7.0.html that the best place to check them?
15:14 bhosmer joined #salt
15:15 JarleB joined #salt
15:15 gcmacmalloc joined #salt
15:15 masterkorp Hello
15:16 tmh1999 joined #salt
15:16 ramishra joined #salt
15:16 masterkorp Is there any way to download the file to download without providing the source_hash ?
15:16 iggy superted666: yeah, looks good (and especially the bits at the top labeled important)
15:16 felskrone joined #salt
15:17 superted666 iggy: Cool, that stuff should be easy for us to handle. We only use salt as a C&C system so most i think were fine. Really tempted to try raet but will hold fire for a few months i think.
15:19 viq ugh, looks like I'll have to wait for 2014.7.1, lack of compound matching hurts my usage
15:20 jcockhren are Release Notes PRs still being accepted?
15:20 j-saturne joined #salt
15:21 ajolo_ joined #salt
15:25 aqua^mac joined #salt
15:26 j-saturne1 joined #salt
15:28 bhosmer joined #salt
15:30 saravanans joined #salt
15:32 __gotcha joined #salt
15:33 ramishra joined #salt
15:39 kermit joined #salt
15:40 kermit joined #salt
15:40 elfixit1 joined #salt
15:40 jeffspeff joined #salt
15:44 tafa2 joined #salt
15:46 haphazardeous joined #salt
15:46 haphazardeous Hi all
15:47 atbell joined #salt
15:48 haphazardeous anyone ever used tomcat.wait state? It used work just fine but after creating a new environment I keep getting this 'State tomcat.wait found in sls ....  is unavailable'
15:49 asdasddf joined #salt
15:50 conan_the_destro joined #salt
15:53 ramishra joined #salt
15:58 mohae joined #salt
15:58 johtso joined #salt
15:59 ramishra joined #salt
15:59 iwishiwerearobot joined #salt
16:01 johtso Should custom modules work fine with masterless salt minions? If so, how can I debug my custom module, it's definitely in _modules, but I get told the function is not available.
16:02 iwishiwerearobot joined #salt
16:03 to_json joined #salt
16:03 mschiff What do you do to check for the last scheduled run of a minion?
16:03 saravanans joined #salt
16:05 Ymage joined #salt
16:10 oyvjel1 joined #salt
16:18 iggy jcockhren: assuming it's correct, yes... one was added after the release to make note of the compound matchers thing
16:19 iggy viq: it only applies to certain circumstancs... make sure you fit that
16:20 jalbretsen joined #salt
16:21 viq iggy: it isn't clear where it applies and where it doesn't from the release notes
16:21 viq I use compound matching for matching states and remote execution, I don't use it in pillars.
16:22 thedodd joined #salt
16:22 unpaidbi1l so, regarding the question i had about mako templates and defaults, i ended up doing this: http://pastebin.com/E0A6iuf4
16:22 unpaidbi1l probably not the most elegant solution but it's the best i've got for now
16:22 TyrfingMjolnir joined #salt
16:23 iggy it went out of my scrollback
16:23 kaptk2 joined #salt
16:24 tligda joined #salt
16:24 unpaidbi1l i needed to set defaults for mako templates for the syslog-ng formula.  i have my services create syslog-ng config blocks in their state that use a stripped down version of the syslog-ng.conf template from that formula
16:25 iggy viq: it only impacts mine.* and compound.compound (iirc)
16:25 smcquay joined #salt
16:25 iggy basepi: correct?
16:26 skyler joined #salt
16:26 basepi viq: you're fine there. It's only mine and publish modules that are affected.
16:27 basepi I'll look into clarifying that note.
16:27 viq ah, we're not using mine yet
16:27 viq thank you
16:27 basepi No problem.
16:28 Ozack1 joined #salt
16:28 basepi Whoa, I totally messed up that note on those release notes.
16:28 basepi It's completely wrong. Thanks for pointing tho
16:28 basepi that out viq
16:29 viq :)
16:31 Gareth In other news, it has been determined that basepi, who removed compound matching from Salt, is also responsible for all the world's atrocities.
16:31 * iggy gets out his torches and pitchforks
16:32 basepi Hahaha
16:34 cpt-oblivious joined #salt
16:35 kermit joined #salt
16:35 Eugene Don't forget the boiling oil
16:36 whiteinge I can get behind this movement.
16:36 masterkorp yeah we want somebody to blame, its easier to accept our failure
16:36 Eugene I have no idea what's going on, but I love a good lynch mob
16:37 whiteinge My knowledge of basepi's home address will be useful when we're ready to start the war crimes trials.
16:37 StDiluted joined #salt
16:37 * basepi trashes whiteinge's desk
16:39 dustinrc joined #salt
16:40 whiteinge meh. i survived the Great Desk Trashing of utahdave's kid's visit.
16:40 masterkorp çpç
16:40 masterkorp lol even
16:44 thedodd joined #salt
16:48 UtahDave joined #salt
16:49 bhosmer_ joined #salt
16:49 ChrisCox joined #salt
16:51 gebi joined #salt
16:51 gebi hi all :)
16:53 Ahlee wait, what's this about compound matching being removed?
16:54 hasues joined #salt
16:54 hasues left #salt
16:54 whiteinge haha
16:55 robawt yo whiteinge
16:55 whiteinge Ahlee: some people (basepi) just want to watch the world burn.
16:55 whiteinge robawt: yo!
16:55 robawt no compound matching in Salt?
16:55 whiteinge was kicked by basepi: whiteinge
16:55 robawt did basepi get a job at Puppet or something?
16:55 basepi ;)
16:56 basepi compound matching works fine!
16:56 whiteinge joined #salt
16:56 whiteinge lol
16:56 basepi ;)
16:56 basepi compound matching works fine, it was temporarily disabled only for mine and pillar functions
16:56 basepi er, publish, not pillar
16:56 robawt well that's not that exciting
16:56 robawt maybe we should remove YAML for a week
16:56 robawt just to upset people
16:56 basepi ooh, that would be exciting
16:56 basepi whiteinge would be happy
16:57 robawt he writes all of his stuff in straight python anyway
16:57 whiteinge the worst thing that would happen is it would speed up everyon'e salt.
16:58 basepi hahaha
16:58 gebi is it normal that for a pillar to update/show i've to remove the cache file first?
16:58 gebi eg. new keys are only added to the master cache file if i remove the file first
16:58 honestly there's a command you're supposed to run after updating the pillar file
16:58 honestly salt-util somethingorrather
16:59 whiteinge !modules.saltutil.refresh_pillar | honestly
16:59 wm-bot4 honestly: http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.saltutil.html#salt.modules.saltutil.refresh_pillar
16:59 thayne joined #salt
16:59 viq robawt: rewrite in go
16:59 whiteinge er. gebi ^^
16:59 gebi and somehow getting the pillars to my minions does not work at all, salt-ssh -c . '*' saltutil.refresh_pillar
16:59 wendall911 joined #salt
17:00 gebi where i get None as result, which seems strange
17:00 robawt viq: that's not funny :|
17:00 robawt :)
17:00 gebi whiteinge: that's exactly what i do :)
17:00 whiteinge robawt: go has a nice minimalist yaml parser that only handles the subset of yaml that people actually like.
17:01 gebi with -l debug i see just one strange message: "Exception AttributeError: "'MinionEvent' object has no attribute 'cpub'" in <bound method MinionEvent.__del__ of <salt.utils.event.MinionEvent object at 0x3886c10>> ignored"
17:01 robawt yeah but then you guys would stop showing up to PyCon
17:01 viq robawt: fine, rewrite in perl :P
17:01 Ahlee gebi: pretty sure refresh_pillar always returns None on success
17:01 robawt there ya go
17:01 honestly whiteinge: go seems like a strange subset of a programming language
17:01 robawt that'll guarantee its future, no one would be willing to rewrite it after that
17:02 honestly "map? reduce? nobody needs that."
17:02 robawt map? just use a hashtable, that's all you need anyway
17:02 Ahlee map is just a fad
17:02 gmcwhistler joined #salt
17:02 sroegner joined #salt
17:02 Ahlee that said, I <3 go
17:03 honestly i mean the map function :P
17:03 gebi Ahlee: hm ok
17:04 Ahlee gebi: I've never looked into why (it's on the never-shrinking todo list)
17:04 honestly Ahlee: does it also always throw an exception?
17:04 gebi so it seems salt-ssh is not transporting pillars? don't think i'm the first one using that?
17:04 gebi i just want to get a https cert to one node
17:04 Ahlee honestly: That I haven't seen.
17:05 Ahlee i also don't use salt-ssh to compare, sorry
17:05 hotbox joined #salt
17:06 gebi damn... seems pillars don't work with salt-ssh
17:08 KyleG joined #salt
17:08 KyleG joined #salt
17:10 gebi seems it's a known bug, can i somehow help? https://github.com/saltstack/salt/issues/8340
17:11 bhosmer_ joined #salt
17:11 gebi i'm using pillar data with file.managed and contents_pillar <pillar>
17:13 blaffoy joined #salt
17:14 aqua^mac joined #salt
17:15 iggy gebi: what version of salt?
17:15 spookah joined #salt
17:15 gebi 2014.1.13+ds-1
17:16 iggy do you have the ability to try 2014.7? It has much improved salt-ssh support
17:16 iggy I honestly doubt anybody is going to look into salt-ssh problems in 2014.1 at this point
17:17 philipsd6 iggy, gebi: I just tested in salt-ssh 2014.7.0-849-gcf8420e (Helium)
17:17 zlhgo joined #salt
17:17 philipsd6 works fine there
17:17 gebi seems my salt-ssh does not create and put pillar.json into salt_state.tgz in client/ssh/state.py
17:17 gebi oh really, ok i'll try an upgrade then :)
17:18 gebi is there a debian package available?
17:18 ProT-0-TypE joined #salt
17:18 gebi ah found it ;)
17:19 blaffoy Hi all, I've encountered a problem with running salt-call in a remote powershell session. Everything works fine in a local, or rdp'd powershell session, or by calling salt from the master.
17:20 blaffoy Is powershell remoting supported by salt, at all?
17:20 babilen gebi: Where did you find it?
17:23 saravanans joined #salt
17:24 gebi babilen: git clone saltsrc; git checkout -b 2014.7 origin/2014.7; dpkg-buildpackage -uc -us -m'me'
17:24 ChrisCox joined #salt
17:25 gebi btw... would be nice if someone would put a debian package of 2014.7 onto debian.saltstack.com ;)
17:25 active8 joined #salt
17:27 iggy I think they are waiting to do a coordinated release once all the major distros have it ready to go
17:28 jevonearth I'd like to use a Saltfile, so I can have all my salt files in a local selft contained directory, I can't find very much documentation on the Saltfile, is it a pretty new fearure?
17:28 jevonearth *feature
17:28 iggy I've never even heard of it
17:29 UtahDave blaffoy: what happens when you run salt-call in a remote powershell session?
17:29 iggy and it's not mentioned at all in the 2014.1 or 2014.7 docs
17:29 iggy are you sure that's a real thing?
17:29 gebi yep, Saltfile does also not work here :)
17:29 UtahDave jevonearth: it's a very new feature. Let me see if I can find info on it
17:30 gebi i just saw it somewhere in a discussion about salt-ssh where you can set the default config_dir for salt_ssh
17:30 jevonearth UtahDave: sweet, thank you. It's referenced here: http://docs.saltstack.com/en/latest/topics/ssh/#define-cli-options-with-saltfile but I can't find much else.
17:30 iggy oh... salt-ssh stuff
17:30 * iggy backs away quickly
17:31 jevonearth iggy: Well, I think it is intended to be more general than that. I imagine it is to salt what a Dockerfile is to docker.
17:31 babilen go on …
17:32 srage joined #salt
17:32 UtahDave no, it's more like a vagrantfile
17:32 UtahDave for salt-ssh
17:33 iggy I can't imagine a world where I'd put all of a system state/config/whatever in a single file (except possibly salt-ssh if I bothered to use that)
17:34 zekoZeko i'm having a problem with one of my minions: No Top file or external nodes data matches found
17:34 * robawt highfives UtahDave
17:34 zekoZeko the versions on the master and minion are the same
17:34 gebi why is it that this room seems against salt-ssh? is it like a big internal quirk in salt?
17:34 UtahDave yo, robawt!
17:34 zekoZeko 2014.1.13 something, i tried restarting the master, minion, everything.
17:34 jevonearth UtahDave: so it is specific to salt-ssh then? Ok, I was hoping it would allow me to localize all salt configs and assets to a local directory, so I could version control the whole lot, and have several salt-masters on a single box.
17:34 UtahDave zekoZeko: that means there's nothing matching in your top file for that minion
17:34 zekoZeko UtahDave: but i have '*' in my top.sls
17:35 zekoZeko UtahDave: and it matches all the other minions...
17:35 zekoZeko UtahDave: how can I debug this?
17:35 UtahDave zekoZeko: can you pastebin your top file?
17:35 zekoZeko sure, just a sec... preferred pastebin service?
17:35 UtahDave zekoZeko: try    salt <minion id> state.show_top
17:35 UtahDave and salt <minion id> state.show_highstate
17:35 iggy gebi: it has limitations that traditional salt does not, and it's beta so it's had a lot of issues that people have gotten tired of helping others with
17:36 zekoZeko UtahDave: both return empty
17:36 gebi it's damn hard to get infos when being new in salt
17:36 UtahDave jevonearth: It was intended for salt-ssh, I'm not sure it works with regular salt commands or not.  If it doesn't, I don't see why not.
17:36 alexr__ joined #salt
17:36 iggy and for some reason, it seems like everybody using it is the kind that cba to do a little googling before they ask questions in here, but I can't verify any correlation there
17:36 Shish Anyone know how to import jinja macros from a parent directory? eg sites/mysite/init.sls wants to do {% import "../common.sls" as common %}, but it can't find "../common.sls", nor if I use an absolute path, or salt:// :(
17:37 ndrei joined #salt
17:37 iggy that and masterless (usually with vagrant)
17:37 zekoZeko UtahDave: http://pastebin.com/SSHAPMRc
17:37 iggy Shish: use absolute without salt://
17:38 blaffoy UtahDave: the specific problem is when enforcing a state that has "cmd.run: - name: C:\env\numpy\Scripts\pip.exe install numpy --index-url http://pypi/root/prod/"
17:39 Shish "sites/common.sls" says not TemplateNotFound, "/sites/common.sls" says "MinionError: Unsupported path"
17:39 phblj joined #salt
17:39 blaffoy UtahDave: when called from a local powershell, or from the salt-master, the package downloads and installs correctly
17:39 SheetiS Shish: do it like a salt:// url but no salt:// and no leading /
17:39 SheetiS SheetiS: does it still mess up like that?
17:39 SheetiS err Shish
17:39 SheetiS funny that it autocompleted my name heh
17:40 UtahDave zekoZeko: Hm. that looks correct. Does that minion have an "environment" setting set in the minion config?
17:40 zekoZeko UtahDave: it shouldn't, let me check
17:40 gebi iggy: really? seems starting with salt-ssh is just not that easy
17:40 alanpearce joined #salt
17:41 gebi (and for most people i know salt-ssh is the only way to use salt)
17:41 UtahDave blaffoy: Does it that command require any manual intervention? like accept a  eula or hit enter or anything?
17:41 hal58th joined #salt
17:41 blaffoy UtahDave: but when called through a remote powershell session I see http://pastebin.com/3EnNLHRa
17:41 Shish ffff, I'm a moron, was writing /common.sls instead of /sites/common.sls, so it couldn't find the file because it didn't exist >_<
17:41 Shish sorry all :D
17:42 iggy gebi: well... salt isn't "easy"... it's a very large, complex piece of software with a lot of power
17:42 Shish (ie, the include statement was right, the file was in the wrong place)
17:42 blaffoy UtahDave: no, no manual intervention.
17:42 iggy it's like the second largest python project on GitHub (iirc)
17:42 UtahDave blaffoy: huh, so you're getting a MemoryError
17:42 iggy You just can't write a "15 minutes to Salt" type doc for something like that
17:42 blaffoy UtahDave: interestingly, other packages will unpack and install without error.
17:43 iggy and as I've said, salt-ssh is still beta (or was up until the 2014.7 release)
17:43 zekoZeko UtahDave: fixed... somehow i set file_client: local in minion config... must have been playing with something :( thanks
17:43 UtahDave zekoZeko: Ah, glad you found it.  I figured it was something like that.
17:43 iggy most projects don't put as much into documenting beta features
17:44 gebi iggy: i don't fear big projects, but it seems when beginning with start-ssh you are pretty much left in the cold (at least i had the impression)
17:44 UtahDave gebi: salt-ssh has seen a huge number of improvements in the 2014.7 release.
17:44 iggy ianad, but I wouldn't want to make it easier for people to use beta features
17:44 UtahDave gebi: a lot of lessons learned from the beta users
17:45 gebi then PLEASE for your own sake beg for a freeze exception in debian
17:45 iggy nobody uses distro packages
17:45 UtahDave blaffoy: Does anything appear out of the ordinary  in C:\Users\xxx-admin\pip\pip.log  ?
17:45 gebi salt-ssh is installed just like a normal package and the next stable debian will get 2014.1.13 if nothing is done about that
17:46 wincus joined #salt
17:47 iggy I meant that things like salt-bootstrap setup the salt repo and install from that (or uses backports)
17:47 bhosmer_ joined #salt
17:47 zekoZeko gebi: I don't see a need for that. Salt is a fast moving target and during Jessie lifetime there will most certainly be a need to go even beyond 2014.7
17:47 gebi iggy: your users do, sorry
17:47 zekoZeko gebi: so you need something like saltstack PPA for Ubuntu anyway.
17:47 iggy I already said ianad
17:47 gebi zekoZeko: yea sure, but you don't want to have a user exposed to 2014.1 salt-ssh
17:48 zekoZeko gebi: that I understand, so it would be better to just remove that...
17:48 oyvjel joined #salt
17:48 gebi and if that's the default for debian stable you'll get bugreports and questions the next few years about that version
17:48 zekoZeko gebi: and point users to somewhere else if they use it.
17:48 gebi maybe yes
17:48 iggy then I'd say instead of trying to shoehorn an unreleased version (2014.7 still isn't officially released) into jessie, jessie should drop the salt-ssh package
17:48 desposo1 joined #salt
17:48 zekoZeko file a RC bug.
17:49 blaffoy UtahDave: it doesn't look like it: http://pastebin.com/Fxd0fqFq
17:49 mgw joined #salt
17:49 desposo joined #salt
17:49 gebi zekoZeko: that would be much better done by someone of the salt project itself
17:50 alexr joined #salt
17:50 blaffoy UtahDave: it seems to be the same stack trace. If I run the same state.highstate from a local powershell session it attempts to download the same file from our pypi server, but then it succeeds
17:50 UtahDave blaffoy: are there memory restrictions on process started from a remote powershell?
17:50 iggy gebi: if you feel that strongly about it, file a salt bug to have the discussion and possibly have some one from the project file the debian bug
17:50 zekoZeko gebi: i'd do it myself if I felt the need, I'd just make sure there's agreement here about that.
17:51 zekoZeko or what iggy said :)
17:51 jonwincus joined #salt
17:51 teebes joined #salt
17:51 iggy getting frustrated with a couple users about it on IRC isn't going to do anybody any good
17:51 UtahDave gebi: We'd definitely be open to a discussion about this on the issue tracker. that way it doesn't get lost.
17:51 blaffoy UtahDave: that's a really good question. I'll look into that and come back to you: https://social.technet.microsoft.com/Forums/windowsserver/en-US/5031b9bd-e7df-4b34-825a-3d9380f8aea4/how-to-override-powershell-memory-limit-in-remote-session
17:52 iggy ^ there are only a few people in here that are actually qualified to have that discussion anyway (and as I've said a few times, I'm not one of them)
17:52 jevonearth Using salt-call, how do I get the value of a specific interface? For example, salt-call --local grains.item ip_interfaces.lo returns nothing, but lo does exist.
17:53 jevonearth I suspect I should not be using the dot notation?
17:53 UtahDave blaffoy: cool. Let me know what you figure out. We might need to take those limits into account if that really is the issue
17:54 SheetiS jevonearth: try and do this: salt-call --local grains.get 'ip_interfaces:lo' and see if you get better information.
17:54 UtahDave jevonearth: yeah, just add a space.  no dot
17:54 SheetiS or use a space like UtahDave said :D
17:55 meylor joined #salt
17:55 forrest joined #salt
17:56 pipps joined #salt
17:56 jevonearth hmm, none of those appear to work:  see http://pastebin.com/X7HnCpfq
17:56 pipps joined #salt
17:57 jevonearth salt-call 2014.7.0 (Helium)
17:59 robawt couldn't you grab with an array?
17:59 * robawt checking
18:01 gmcwhistler anybody at lisa14?
18:01 UtahDave jevonearth: try  sudo salt \* network.interface lo
18:02 UtahDave gmcwhistler: is that the one in New Orleans?
18:02 blaffoy UtahDave: that was it. The value of  WSMan:\localhost\Shell\MaxMemoryPerShellMB was 150MB. I upped it to 2048MB, and the installation completed successfully.
18:02 blaffoy UtahDave: thanks for your help.
18:03 UtahDave blaffoy: that's great info.  We should add that to the docs.
18:03 UtahDave thanks for tracking that down, blaffoy. I hadn't tried doing any salt commands over powershell remotes.
18:03 gmcwhistler @UtahDave: no, this years is in seattle
18:04 blaffoy No problem. I'm just getting the hang of powershell. I keep getting burned by expecting a ps remote session to behave like ssh.
18:05 UtahDave ah, ok. We're sending some people to some HPC conference in New Orleans next week. I wasn't sure if it was lisa or not
18:05 UtahDave blaffoy: yeah, they're tricky.  I found getting all the keys and powershell versions to match and work together was a pain
18:06 blaffoy Cheers for your help today. I'm sure I'll be back soon.
18:06 UtahDave blaffoy: cool
18:07 gmcwhistler lisa14 is going on now, huge google contingency... puppetlabs is here, but was actually hoping to talk to some salt folks
18:07 jevonearth UtahDave: but I'm not looking to view the grain on all minions, I explicitly want it from the local minion, which happens to be masterless.
18:07 UtahDave jevonearth: ok, so     salt-call network.interface lo
18:08 UtahDave gmcwhistler: let me check and see if we have anyone there.
18:08 gmcwhistler thx
18:10 UtahDave gmcwhistler: So we've had people at lisa before, but we couldn't this year due to some scheduling conflicts. I was told we'll most likely be there again next year
18:10 UtahDave We just had too many people out at other conferences and customer sites.
18:11 gmcwhistler gotchya
18:11 pipps joined #salt
18:11 UtahDave gmcwhistler: let me know if I can help you with anything, though.  Are you in the Seattle area?
18:12 gmcwhistler hopefully i'll already have dumped our puppet implementation in favor of salt by then
18:12 sc` joined #salt
18:12 gmcwhistler i'm in seattle only for the conference
18:12 Eugene I'm in the Seattle area!
18:14 UtahDave gotcha.
18:14 robawt gmcwhistler: you should load up on food at pike's place
18:14 UtahDave so is forrest
18:14 forrest Good week to be here, weather is nice
18:14 gmcwhistler ate at the capital grille last night
18:15 forrest I've never been there, I walk past it all the time to go to the gym, but it seems a bit expensive.
18:16 gmcwhistler it was pretty pricey, but i like a good steak every once in a while
18:16 forrest gmcwhistler: if you like falafel and head back to the market stop by 'falafel king'
18:16 forrest that place is tasty.
18:16 gmcwhistler thanks, i
18:17 gmcwhistler 'll check it out
18:17 gmcwhistler how about sushi?
18:18 forrest I'm not sure, I don't eat a lot of fish.
18:18 alanpearce joined #salt
18:20 shepax joined #salt
18:20 shepax Hi every one
18:21 shepax hello some one there?
18:22 Ryan_Lane joined #salt
18:22 forrest shepax: We're here, if you have a question just ask it, if someone knows they will answer
18:23 span joined #salt
18:23 span joined #salt
18:24 shepax i just build up a sqlserver module and i have some trble with it. I sync it but it is not available
18:25 rojem joined #salt
18:25 UtahDave shepax: can you pastebin your module?
18:25 conan_the_destro joined #salt
18:26 _prime_ joined #salt
18:26 shepax https://github.com/shepax/sqlserver_saltstack/blob/master/sqlserver.py
18:26 shepax there is the github of the module
18:27 UtahDave shepax: can you verify that the pymssql python module has been installed on the minion already?
18:27 shepax Yes is installed
18:28 UtahDave shepax: ok, so let's do the following to troubleshoot.   on your windows minion,  stop the salt-minion service
18:28 UtahDave open a cmd by rightclicking on cmd.exe and running as administrator (elevated privileges.)
18:28 UtahDave cd  c:\salt
18:29 UtahDave salt-minion.exe -l debug -c c:\salt\conf
18:29 UtahDave that will start the minion in debug mode
18:29 UtahDave as it starts up, look for any errors or stacktraces.
18:29 UtahDave put your module in    /srv/salt/_modules    and run    salt <minion name> saltutils.sync_all
18:30 UtahDave and try running your module.  Again look for any errors or stacktraces in your minion.
18:30 UtahDave I have to run to a conference call right now. I'll let you know when I get back.  Shouldn't be more than 15 or 20 minutes.
18:30 shepax OK i will do it now
18:33 ekristen joined #salt
18:34 wincus joined #salt
18:35 jonwincus joined #salt
18:37 linjan joined #salt
18:38 troyready joined #salt
18:39 apergos waiting for it to show up in the salt ppa so I can test against the packages we will actually use
18:40 apergos so if I understand correctly the plan is to gradually move to raet as the preferred transport because of reliability issues with zmq under load?
18:41 cwyse joined #salt
18:42 thedodd joined #salt
18:43 baconbeckons joined #salt
18:44 XenophonF joined #salt
18:44 iggy apergos: zmq will be supported for the foreseeable future (alongside raet)
18:45 alanpearce joined #salt
18:45 iggy for now raet is "beta"
18:45 apergos I understand raet is a very new thing
18:46 bhosmer joined #salt
18:46 apergos I was just trying to read between the lines on the discussion of some zmq issues
18:46 diegows joined #salt
18:46 iggy I doubt most salt users are at the point where they'd feel zmq's scalability issues
18:50 * Shish wonders who is using salt who /is/ seeing the issues o.o
18:50 alexr__ joined #salt
18:50 gmcwhistler joined #salt
18:51 cpowell joined #salt
18:51 baconbeckons i have a salt-cloud map that creates a group of servers. if i create a salt master with the map, then the other minions are controlled by that master. but, to start the servers in the cloud map, i need a salt master running. i thought that i would be able to name the salt master the same as the one that i make in the map and have the minions controlled by that master, but it appears that salt cloud still tries to start another master. i kn
18:51 baconbeckons that i can override the settings on a per-minion basis to point to the existing master, but that means that i need to manually set the ip of the existing master in each minion entry in the map. how am i supposed to be using salt cloud?
18:51 iggy if I were to guess... linkedin, google (as part of kubernetes), etc... i.e. the companies with the 10k+ minions it takes to see
18:52 iggy baconbeckons: from a pre-existing salt-master
18:53 gebi uh... seems salt-ssh 2014.7 does not work anymore on debian wheezy, "ImportError: cannot import name CertificateError" from url-lib transported target
18:53 baconbeckons iggy: it appears that when i create the minions, they stop listening to the pre-existing salt master because they are told to listen to the salt master that is being started. is that correct?
18:54 rojem joined #salt
18:55 alanpearce joined #salt
18:56 iggy uhh... I don't know, I never tried to start a salt master with salt cloud
18:57 perfectsine joined #salt
18:57 iggy I think there is a way to say that a host already exists if that's your problem
18:57 baconbeckons iggy: if you don’t start the salt master with salt cloud, how are you getting the minions to listen to the existing saltmaster?
19:01 utahcon using file.managed, if I set "template: jinja" then the scoped variables from the state file should be available in the jinja template, right?
19:02 alexr joined #salt
19:02 iggy baconbeckons: it was set in our provider or profile config
19:02 utahcon http://pastie.org/9710144
19:02 utahcon shouldn't {{ my_var }} be available from my_state.sls ?
19:03 iggy no, you have to pass it via context
19:03 baconbeckons iggy: is there a way to make the provider/profile automatically determine the ip of the existing salt master?
19:03 utahcon blargh
19:03 utahcon ok
19:03 aqua^mac joined #salt
19:04 iggy baconbeckons: not that I know of, but we only used salt-cloud for a couple weeks before we decided it wasn't going to serve our purposes
19:04 baconbeckons iggy: doh! :(
19:04 shepax Hi  i can't sync the pillar to a windows minion some one with the same problem?
19:04 shepax im using saltutil.refresh_pillar
19:05 shepax and then looking if the pillar is there with salt 'minion' pillar.item
19:05 iggy shepax: how are you determining that it isn't sync'ing?
19:06 iggy shepax: try pillar.items to see if it's somewhere you aren't expecting
19:06 shepax using the pillar.item
19:08 alanpearce joined #salt
19:08 hal58th utahcon: http://docs.saltstack.com/en/latest/topics/best_practices.html#variable-flexibility
19:08 shepax nop it didnt sync
19:09 egalano Has anyone run into issues with rendering a template with a raw string containing double curly-braces like :
19:09 egalano mystring: “foo{{asdf”
19:09 iggy shepax: do you have it targeted correctly?
19:09 iggy egalano: {{ raw }} {{ endraw }}
19:09 shepax how that?
19:10 egalano @iggy the raw/endraw didnt work for me. maybe im doing something else that i shouldnt be
19:10 SheetiS Worst case you can do {{ "{{" }} in jinja
19:10 iggy egalano: start pasting code I guess
19:11 SheetiS it's the ugliest solution I think.
19:11 iggy shepax: i.e. did you update your pillar/top.sls accordingly?
19:11 SheetiS also {% raw %}foo{% endraw %} is technically the correct way to do raw block.
19:11 diegows joined #salt
19:11 shepax nop
19:12 shepax i wrout a new pillar db.sls
19:12 alexr joined #salt
19:13 iggy shepax: then you targeted that sls file at the proper minion?
19:14 shepax yes i did this
19:14 __gotcha joined #salt
19:14 iggy same goes for you I guess... start pasting as much code as you can (obviously change any sensitive info since it's a pillar)
19:15 shepax mkdri /srv/pillar  and mk a file there with the pillar then, at the master salt 'minion-windows' saltutil.pillar_refresh
19:15 Mso150 joined #salt
19:15 Ozack1 joined #salt
19:15 iggy pillars are targeted in the pillar top.sls file
19:16 rap424 joined #salt
19:16 egalano iggy here is some code of what im trying to do: https://gist.github.com/egalano/d0f0907697b488e2a1d5
19:16 iggy otherwise, how do you expect salt to know what pillars go to what file?
19:17 egalano it feels like it is getting rendered more than once so the first time it renders the {{ correctly but then it gets…re-rendered? does that sound plausible?
19:18 shepax lol sorry didnt know that
19:18 shepax thx fot the help
19:18 saravanans joined #salt
19:18 iggy egalano: I'm assuming that's a hastily copied example? (I see a few errors off the bat)
19:19 egalano yeah sorry hastily copied. had to strip out sensitive info
19:20 bmonty joined #salt
19:21 iggy it shouldn't be parsing it twice
19:22 ndrei joined #salt
19:24 egalano the error output looks like this: https://gist.github.com/egalano/20af31de94baff0c8235
19:24 murrdoc joined #salt
19:24 nitti joined #salt
19:24 egalano the problem string is something like ‘foo{{bar]%}baz’
19:25 shepax the pillar have to be writen with tab or space??
19:25 murrdoc spaces for all!
19:25 iggy I doubt it'll make any difference, but try doing it like: {% raw %}\nproblem_string: ....\n{% endraw %}
19:25 murrdoc are you using vim shepax
19:25 murrdoc oh
19:25 * murrdoc shuts u
19:26 murrdoc up*
19:26 shepax nop gedit
19:26 iggy egalano: that last line of mine was for you
19:27 pipps joined #salt
19:27 egalano got it iggy. thx trying it now
19:28 aparsons joined #salt
19:29 cmthornton joined #salt
19:29 ChrisCox joined #salt
19:29 shepax yees i got it :)
19:29 shepax thx iggy
19:29 ckao joined #salt
19:31 jagardaniel hi :) is it possible to execute a "custom command"/run a statefile when a file has been changed? I have only found examples with watch_in for services yet
19:32 davedash joined #salt
19:33 iggy jagardaniel: watch only works if salt changes the file... it isn't quite clear from your question whether you are looking for a generic file watcher or the salt watcher
19:33 jaimed joined #salt
19:33 alanpearce_ joined #salt
19:34 egalano iggy the multiline raw statement didnt make a difference. I also tried moving the string into pillars and that didnt help. i’ll try a few other things and let you guys know if i have any luck.
19:35 iggy egalano: you might also try updating your jinja version (don't really know if it'll help, but it's worth a try)
19:35 jagardaniel iggy: ah, sorry! yes, I want to do something when salt changes the file - like running a cmd.run or something like that - it is for restarting a service, but we are using windows and i need to unregister/register the service again after the file has been altered by salt
19:35 wincus joined #salt
19:35 iggy you can do anything salt can via a watch
19:35 TheThing joined #salt
19:35 jonwincus joined #salt
19:36 delinquentme joined #salt
19:36 uriarte joined #salt
19:37 shookees joined #salt
19:38 iggy assuming you've looked at http://docs.saltstack.com/en/latest/ref/states/requisites.html ... onchanges looks like what you want
19:39 iggy or possibly prereq... your description isn't super clear to me (as I don't know that much about Windows)
19:40 teebes joined #salt
19:41 jagardaniel iggy: thanks :) i checked it but not enough apparently! will take a look at onchanges. thanks for your help
19:41 wincus joined #salt
19:41 MTecknology Apparently, now when I start working on salt stuff, I get to open up four terminals. (terminator split into four chunks) ~/salt/{states,data,pillar,top}/ all their own git repos
19:41 MTecknology makes me feel fancy and advanced
19:41 jonwincus joined #salt
19:42 shepax I have problem with a module i build up, y have all the pkg imported and the pillar is ok now but when i run the module it says its not available
19:42 murrdoc get you some screens
19:43 iggy shepax: is the module in /src/salt/_modules ?
19:43 iggy err /srv/salt/_modules
19:43 iggy stupid muscle memory
19:43 shepax yes, the module is sync
19:44 shepax im on debug mode on the minion looking for the problem but there is no error at all
19:44 druonysus joined #salt
19:44 ingwaem joined #salt
19:45 ingwaem joined #salt
19:47 shepax i have a command details ('tgt_type': 'glob' , 'jid' : 21921382139021','tgt': 'minion-windwos','ret':'','user':'root','arg':['dpa','qery'],'fun':'sqlserver.run_query'
19:47 iggy oh yeah... Windows
19:48 * iggy runs away
19:48 shepax XD
19:49 __gotcha1 joined #salt
19:51 shepax some one ??
19:52 babilen shepax: qery vs query ?
19:52 intellix joined #salt
19:52 shepax is query
19:52 shepax is the same bad writing
19:53 rlarkin {% if {% if
19:53 _prime_ joined #salt
19:54 perfectsine joined #salt
19:55 XenophonF so can i put binary data into a pillar?
19:55 XenophonF i have a DER-encoded X.509 key-pair that i want to deploy
19:55 alanpearce joined #salt
19:56 XenophonF via file.managed and contents_pillar
19:59 hal58th unfortunately you can only use plain text in pillar. There is no way to deploy files via pillar at this time
19:59 hal58th put a password on your cert and put that password in pillar
20:01 XenophonF Oh, I just found out that YAML has a binary data type: http://yaml.org/type/binary.html
20:01 hal58th oh really... I'll have to look into that
20:01 jcockhren in case anyone wants to use salt with packer: https://github.com/mitchellh/packer/pull/1668
20:03 bhosmer joined #salt
20:07 babilen XenophonF: I'd suggest to *not* put it in a pillar, but place it ins "local" file_roots on the master (say /srv/salt/local/...) and then refer to it with salt://...
20:08 babilen (assuming you use GitFS and wouldn't want it in there)
20:08 Reiner0312 joined #salt
20:09 iggy uuencode?
20:09 babilen Sure, but whatfor?
20:09 rap424 joined #salt
20:09 iggy the lolz
20:10 babilen XenophonF: Or would you consider this data to be sensitive?
20:11 babilen In which case I would write a pillar in Python that reads the file directly from disk and returns a suitable dictionary. Jinja is horrible when it comes to multi-line strings with all sorts of "funky" data in there.
20:11 babilen You really wouldn't want that.
20:13 bhosmer joined #salt
20:16 XenophonF babilen: it's keying material, thus sensitive
20:16 XenophonF the YAML binary data type might do the trick
20:17 XenophonF it just takes base64-encoded data and decodes it to the right thing
20:18 XenophonF i use contents_pillar elsewhere without any trouble
20:19 XenophonF anyway, we're about to find out how well this works in practice :)
20:20 patrek joined #salt
20:21 CatPlusPlus joined #salt
20:22 meylor1 joined #salt
20:23 babilen XenophonF: Yeah, I just dislike multiline strings/data in YAML and just write a little pillar that reads the file in Python and returns the data in dictionary. open('/path/to/foo, ...).read() *bam*
20:23 cpt-oblivious joined #salt
20:26 XenophonF that's a good idea too
20:27 b1nar1 joined #salt
20:28 jab416171 joined #salt
20:31 thawes joined #salt
20:33 aparsons_ joined #salt
20:40 thayne joined #salt
20:41 rap424 joined #salt
20:43 babilen I would actually like if something like that were to become part of salt itself (in that you can say "pillar_file_roots" (or so) and you can create files in there that are made available like state. (i.e. $ROOTS/path/to/file.something) becomes, say, path:to:file_something in your pillar.
20:43 babilen Haven't quite made up my mind about how this would work, but I'd like an easier way to make data from files available via pillars without all that YAML horror
20:44 babilen You could even create gpg encrypted files there easily and then render them with the gpg renderer
20:44 babilen I just need some time and a good path → pillar name mapping
20:44 anotherZero joined #salt
20:45 XenophonF to me the combo of yaml and jinja feel natural
20:46 ghanima joined #salt
20:46 XenophonF together, they almost approximate lisp s-expressions and macros
20:46 XenophonF the same structure encoding both data (pillar) and code (state)
20:47 cpt-oblivious joined #salt
20:48 pipps joined #salt
20:52 aqua^mac joined #salt
20:53 UtahDave babilen: I do like the idea of a file populating a pillar value
20:53 XenophonF i know someone's asked this before, but i can't find it the discussion.  is there any difference between e.g. grains.items(...) or salt['grains.items'](...)?
20:53 babilen I use it for usecases like that, but would like a bit more of an official approach to it than me writing pillars in Python :)
20:54 rojem joined #salt
20:54 UtahDave XenophonF: within the jinja context?
20:54 perfectsine joined #salt
20:55 bhosmer_ joined #salt
20:55 babilen UtahDave: You quite often want to simply populate files with the content of another file, but you cannot use salt:// because the information in that file is sensitive. I like the "python shortcut" rather than battling multi-line YAML. Official support for that would be great (I would even hack that, but need to come up with a good idea for the PATH → pillar name mapping.
20:56 UtahDave Hm. how about    pillar://mydir/myfile.txt
20:58 XenophonF that'd be kind of nice
20:59 skyler UtahDave: I was wondering if you could give some feedback on the ideas in this issue: https://github.com/saltstack/salt/issues/12179. I think dependency management for formulas, and making it easier to include them is really important to make them more useful.
20:59 XenophonF right now I do things like this - https://github.com/ibrsp/salt-states/blob/development/salt/master.sls
21:00 XenophonF so if the pillar is there, i conditionally use contents_pillar
21:03 shookees joined #salt
21:04 rap424 joined #salt
21:04 baconbeckons joined #salt
21:06 Mso150 joined #salt
21:06 babilen UtahDave: That looks perfect. I was just thinking in terms of "normal" pillar "breadcrumbs" hence the mydir:myfile ...
21:06 conan_the_destro joined #salt
21:08 meylor joined #salt
21:09 forrest joined #salt
21:18 Jarus joined #salt
21:20 baconbeckons joined #salt
21:22 teebes joined #salt
21:22 UtahDave XenophonF: that's a pretty sweet workaround!
21:26 UtahDave skyler: Yeah, I've been following that thread.  I think this would be really amazing.
21:27 jeremyb joined #salt
21:27 babilen XenophonF: I like that too, now combine that with pillar:// and you have awesomeness!
21:27 Jarus joined #salt
21:29 aquinas joined #salt
21:29 gmcwhistler joined #salt
21:30 UtahDave skyler: I think what you've all been discussing will be fantastic and quite necessary. A while back we had some initial discussions about how to put that all together, but we haven't had the time to start implementing it.  We've been focusing a LOT of our resources on improving our QA process, among other things.
21:31 spookah joined #salt
21:34 skyler UtahDave: I think this would be a real game changer for formulas. I hope you guys have some time for it in the future, but I understand that y'all have a lot of other things to do. I might make a proof of concept, but I probably wont have time for that in the short run either.
21:34 thayne joined #salt
21:34 to_json joined #salt
21:35 UtahDave skyler: Yeah, I think that having that discussion on the issue tracker will help tremendously to focus on the best way to go about doing that.
21:35 UtahDave skyler: especially by learning from other package management software
21:36 linjan joined #salt
21:38 murrdoc joined #salt
21:38 skyler UtahDave: The thing about my current states tree is that I would like to formulaize the states, but there are too many dependencies for it to be reasonable ATM. Once we have dependency management, that will become practical. Also, if there are actual releases for formulas on pypi (or something like it), then versions on formulas will actually be useful.
21:38 UtahDave hm. great point.
21:42 utahcon anyone have any links for using salt + chroot?
21:42 utahcon or how to install software under chroot?
21:43 utahcon or am I looking at doing some crazy cmd.run work?
21:43 rlarkin joined #salt
21:44 iggy skyler: UtahDave: what's wrong with git submodules (maybe with a little magic that recurses the first subdir in a submodule)
21:46 ggillies joined #salt
21:46 muss joined #salt
21:48 aurynn joined #salt
21:48 skyler iggy: Give me a few minutes to make a quite write up with my thoughts on the matter. I would like to articulate it better than I can in an irc message.
21:48 morsik left #salt
21:48 Mso150 joined #salt
21:48 iggy I suppose I could ask that on the issue
21:49 manfred iggy:  git submodule update —init —recursive
21:55 babilen iggy: You would, presumably, want something a little more "controlled" in the long run. Something along the lines of Python's "batteries included" approach or Haskell's Platform. Versions developed in one "branch" should probably be kept API compatible in subversions (so that you can still fix bugs), while new developments should target a new salt release/branch.
21:55 iggy you can version submodules
21:55 babilen Sure
21:57 babilen I'm just afraid that this turns into the same nightmare Java is currently in in that one module requires a very specific version (say 1.2.3.4.5) and another formula you *also* want to use requires 1.3.4.5.6 ... You now have a clash and as nothing declares "version ranges" (i.e. "works with 1.*") you will probably end up pulling in the newest
21:58 iggy honestly... I think this over-engineering
21:58 honestly that may well be, iggy!
21:58 nitti_ joined #salt
21:59 iggy if it takes more than a couple clicks on GH for me to submit my changes back upstream, I won't bother
21:59 iggy I'll just keep all my stuff to myself
21:59 iggy an... unfortunate irc nick
21:59 murrdoc git 1.8 lets submodules follow HEAD
21:59 Singularo joined #salt
22:00 babilen What's wrong with targeting 2014.1 and 2014.7 in your formulas and treat them as "batteries" and demanding that they are, in terms of their pillars, backwards compatible and document breaking changes on upgrades between major salt versions?
22:00 babilen I mean you would have to do something like that anyway if you want to make formulas usable with more than the absolute newest release
22:01 iggy salt is moving too quick right now to worry about people that don't want to upgrade (imo)
22:01 babilen But sure ... that is a bit more work than GH hipster might put into it and it is probably up to us maintainer's to actually package approved "salt-formula" packages in Debian
22:01 bhosmer joined #salt
22:01 babilen iggy: "in the long run"
22:01 murrdoc is salt version available as a grain or a pillar ?
22:01 iggy yeah, then solve the problem when it actually becomes a problem
22:02 iggy right now... it isn't
22:02 pipps99 joined #salt
22:02 murrdoc well thats a booboo
22:02 babilen We already run into formulas that target 2014.7 or include .ng states for 2014.7 and "normal" ones for 2014.1
22:02 iggy there's probably 5-10 formulas that are worth "packaging" in some way
22:02 murrdoc any reason they cant be tags ?
22:03 iggy and that number hasn't changed in the last 6-9 months
22:03 babilen So, more work needs to be done on formulas.
22:03 skyler Formulas need lots of standardization.
22:03 iggy this sounds more like "Chef solved this problem already, so should Salt"
22:04 iggy when in fact Salt isn't anywhere near the stability level that Chef/Puppet/etc are
22:04 murrdoc what is the problem , again
22:04 iggy #12179
22:04 skyler Right now, there is little consistency, it is hard to intuit how to work with any formula, and you basically just have to read and completely understand the source.
22:05 iggy I don't think that's the case in general
22:05 babilen Yes, but some stability would be nice. I *really* wouldn't mind if the next year would see a focus on formulas, documentation, bug fixes and making things work nicely together.
22:05 aparsons joined #salt
22:06 iggy I'd say more than half of the formulas I've looked at were easy to get going with reading the README and the pillar.example
22:06 iggy probably way more than half... I'm being conservative
22:06 murrdoc ' berkshelf or npm like dep management'
22:06 murrdoc come on
22:06 iggy glad I'm not the only one
22:06 murrdoc just saying man
22:07 murrdoc thing is formulas dont work out of the box
22:07 murrdoc or havent for me
22:07 murrdoc but i dont have a problem with that
22:07 iggy they are examples
22:07 murrdoc yeah , best practices examples
22:07 UtahDave I think these are conversations worth having.  The current salt-formulas setup is basically to learn what works and what doesn't.
22:07 babilen No, they are actual salt code used in production by a lot of people
22:08 murrdoc like the graphite one only works with diamond
22:08 murrdoc babilen:  that sounds like that would be nice if true
22:08 iggy and some are just plain bad
22:08 murrdoc the vim one was alright
22:08 murrdoc there are two for sudo
22:09 saltnoob joined #salt
22:09 murrdoc like i said, the stuff is good, but its not , plug and play
22:09 skyler murrdoc: I hope that it can become plug and play.
22:09 murrdoc me too
22:09 murrdoc but its plug, edit tiny bit, and play
22:09 babilen murrdoc: Well, I use a lot of formulas and people come in here (or file bugs, PRs, questions, .. on GH) that indicate that they do too. I mean you have to start somewhere and the community should approach formulas as if it would be "salt's standard library"
22:09 murrdoc which is good enough
22:09 iggy so let's focus on fixing the real problems in the short term... poor/non-existent documentation for formulas (should improve the dependency resolution... albeit manually), lack of flexibility/robustness of some
22:10 saltnoob i'm having an issue trying to run manage.bootstrap_psexec. im getting an error message of function 'manage.bootstrap_psexec is unavaiable. this is with a fresh install.
22:10 cwyse joined #salt
22:10 iggy etc.
22:10 murrdoc babilen:  sorry, no disrespect
22:10 kballou joined #salt
22:10 iggy I guess my main point would be... formulas have a long way to go before they need a dependency resolver
22:11 murrdoc co sign
22:11 babilen They do indeed
22:11 alanpearce joined #salt
22:11 skyler iggy: I think dependency management is a real problem. Not because so many have broken dependencies, but because very few formulas have dependencies at all because if they did, no one would use them.
22:11 babilen But that doesn't mean that one should treat them like "it doesn't matter anyway, they are will be shit for the next year". At least that's not my approach.
22:11 murrdoc do we have a 'heres what makes a good formula' checklist
22:11 murrdoc i ll gladly work with you guys to bring the formulas up to spit*
22:12 iggy so far I've been trying very hard to get my changes contributed back
22:12 babilen There is the "best practices" document and the "formula conventions" one, but not such a checklist.
22:12 skyler murrdoc: A comprehensive checklist would be really good.
22:12 murrdoc i mean dedicated to formulas
22:12 iggy and maybe we need to be more picky about what gets committed (i.e. no commit without relavent docs, etc)
22:12 murrdoc formula conventions is nice
22:12 murrdoc yeah
22:12 murrdoc use the same 'documentor' as the one for salts docs
22:13 iggy I know I've probably been guilty of poorly documenting some of my changes (relying on commit messages, etc.)
22:13 skyler iggy: agreed. Right now it is definitely take what you get, leading to inconsistency.
22:13 babilen It would be good t be a bit more picky, yeah. Also none of the "Do foo", "Fixed something bar", "Revert Fixed something bar", "Fixed something bar but better" commits (should be rebased)
22:13 giantlock joined #salt
22:13 iggy ...
22:13 skyler I AM guilty of not properly documenting my changes to a formula...
22:14 iggy I feel like you're pointing fingers at me babilen ;)
22:14 murrdoc babilen:  what would you say are the top 5 formulas
22:14 murrdoc honest question
22:14 mosen joined #salt
22:15 skyler Formulas are basically a set of mini projects that are not mature at all. We should probably consider them all (or almost all) to be in the rapid development, pre 1.0.0, phase.
22:15 murrdoc by 'heres what we believe is a good salt formula'
22:15 murrdoc not by usage
22:15 babilen murrdoc: Sorry, telephone ...
22:16 murrdoc no rush
22:17 skyler Could we lint our formulas and formalize some style guidelines (not necessarily right away)? What do you guys think?
22:17 timoguin linting would be great
22:17 timoguin testing with kitchen-salt would be awesome
22:17 murrdoc kitchen salt is tied to vagrant
22:17 murrdoc but otherwise +1
22:18 rap424 joined #salt
22:18 timoguin yes it is.
22:18 iggy yeah, you'd lose me if vagrant was a requirement to submit
22:18 timoguin but i don't know of anything else that even comes close to what it is
22:18 iggy ain't nobody got time for that
22:18 murrdoc salt-testing uses the cloud provisioners
22:18 murrdoc which is good
22:18 timoguin the influxdb-formula has a .kitchen.yml file for testing.
22:18 timoguin i'm working on adding some others
22:19 murrdoc kitchen-salt needs an upgrade to work with openstack
22:19 murrdoc and/or aws
22:19 timoguin vagrant works with aws
22:19 murrdoc or gce
22:19 iggy docker uber alles
22:19 timoguin you just gotta enable the provisioner. docker works. and there are some others
22:19 timoguin but i still to virtualbox
22:19 timoguin *stick
22:20 skyler murrdoc: +1 for OpenStack integration. That would be awesome!
22:20 iggy I've usually got a kvm VM or 8 running, so vbox is a no go
22:20 timoguin figuring out how each formula works right now is a pretty big pain point for me
22:20 pipps joined #salt
22:20 timoguin there's this line between hacking it up myself and doing it the "right" way with a formula
22:20 babilen vagrant works with kvm -- https://github.com/pradels/vagrant-libvirt/
22:21 iggy that's libvirt, not kvm
22:21 * iggy no libvirt
22:21 timoguin yea, vagrant has extendable provisioners
22:21 murrdoc oh i gotta submit my openstack implementation btw
22:21 MatthewsFace joined #salt
22:21 murrdoc to provision openstack servers
22:21 iggy but now we're bikeshedding
22:21 * murrdoc makes todo
22:21 iggy or I am anyways
22:21 babilen iggy: Well, use https://github.com/adrahon/vagrant-kvm then if you can't stand libvirt
22:21 murrdoc point is
22:21 murrdoc docs, testing, better commits
22:21 murrdoc before dependancy management
22:21 Reiner0312 joined #salt
22:22 skyler timoguin: I also have trouble figuring out how to make the formulas do what I want. I think if we had more conventions it would help.
22:22 mohae joined #salt
22:22 iggy that uses libvirt (I see xml files!)
22:22 babilen I still haven't decided which style I'd like to see in formulas for naming state IDs. A very opinionated policy document would be nice. It doesn't even have to be perfect, just demand *one* way. :)
22:22 iggy but yeah, what murrdoc said...
22:23 holler joined #salt
22:23 iggy or at least options that are unlikely to collide
22:23 murrdoc this_state_does_this:
22:23 holler why does provisioning a virtualbox VM w/salt seem to take so long? ~15-20 mins sometimes
22:23 timoguin I'm a fan of descriptive sentences for state IDs
22:23 timoguin holler: with vagrant?
22:23 murrdoc this-State_doEs-ThiS:
22:23 skyler babilen: I agree on having an opinionated policy document.
22:23 iggy I hate _'s ;)
22:23 holler timoguin: yes
22:24 murrdoc knew that was coming iggy, wasnt sure from wher
22:24 murrdoc where*
22:24 timoguin holler: depends on what kind of install you're doing... but that seems excessive
22:24 timoguin installing from git can take a minute to fetch all the dependencies
22:24 holler timoguin: I made public repo similar to my private one
22:24 holler https://github.com/erichonkanen/vagrant-salt
22:25 holler https://github.com/erichonkanen/vagrant-salt/tree/master/salt/roots/salt/local
22:25 holler thats the local salt stuff
22:26 skyler murrdoc, iggy: I don't see a conflict between dependency management and cleaning up the conventions, improving states, ect. Perhaps no one will have time to get dependency management done for a while, but I still think it is worth planning.
22:26 timoguin holler: yea going from the base VM, all the git deps take a while to install, then cloning the salt repo, installing those requirements, then running all the states can take a minute
22:26 iggy I think it takes attention away from the actual problems
22:26 murrdoc skyler:  its not conflict
22:26 murrdoc its an ordering issue
22:26 murrdoc if this was a salt state
22:27 murrdoc salt_formulas:
22:27 murrdoc include
22:27 babilen timoguin: Yeah, I started using that a lot, but then there are so many approaches (e.g. https://www.refheap.com/93091 ) and I would just like to have one "style guide"
22:27 murrdoc - doc
22:27 murrdoc so on
22:27 iggy and also fixing the real issues (that exist today) will likely change the landscape a good bit
22:27 holler timoguin: ok gotcha.. didnt know if there was any strategy to speed it up or not.. I thought I had read about a bug where it calls apt-get update excessively that takes ~6sec each time
22:27 holler although I didnt check if its doing that honestly
22:28 timoguin you should see it calling apt-get update in the logs
22:28 skyler murrdoc, iggy: Fair enough, I would just say that we should solve the dependency management problem before including a lot of dependencies with half solutions.
22:28 perfectsine joined #salt
22:29 skyler If people just don't use dependencies much atm, then there wont be a problem.
22:29 murrdoc gitfs is your dependancy managment
22:29 murrdoc if you add it to gitfs correctly
22:29 murrdoc the formula is there for you to use
22:29 catpiggest joined #salt
22:29 murrdoc unless you want to not use gitfs for formulas
22:30 murrdoc and want it as a salt down
22:30 murrdoc download*
22:30 murrdoc then we can just redo salt-bootstrap for salt-formulas
22:30 skyler murrdoc: That is what I would call a half solution. That does not manage dependencies, that allows me to manually manage dependencies.
22:30 murrdoc point is
22:30 murrdoc berkshelf is a glorified git submodule implmentation
22:30 murrdoc imo
22:31 murrdoc just like puppets library thing
22:31 timoguin I'm currently using git submodules to link the formulas I'm using into my main repo.
22:31 skyler And commits do not correspond to versions with any kind of semantic meaning.
22:31 murrdoc https://github.com/rodjek/librarian-puppet
22:31 timoguin And then I have a script that symlinks them into the state tree on the master.
22:31 murrdoc yup
22:31 skyler timoguin: I am also using submodules. I just don't think it is the best solution in the long run.
22:31 iggy one problem with strictly codifying things is it tends to take away from some of the flexibility (which is one of Salt's main strengths imo)
22:31 timoguin Neither do I.
22:31 murrdoc also plugging in a formula requires updating the config file
22:31 timoguin I've been playing with simonmcc's gitshelf recently too
22:32 murrdoc unless we standardize on using conf.d files to add in submodules
22:32 murrdoc i dont think there is an automated way to add a formula
22:32 * murrdoc waits to be corrected
22:32 rojem joined #salt
22:33 iggy I would kill myself if I was forced to use the nginx-formula as a dependency of anything else (as it stands now)
22:33 skyler murrdoc: To my knowledge, there is no standard way. I don't know if you read my comment on the issue thread, but what I would like to see is packages on pypi for formulas, then something like a 'pypifs' that you could include them with.
22:33 iggy it's entirely too complex for what we need
22:33 TheThing joined #salt
22:33 timoguin iggy: yeaaaa I've been just doing manage vhost configs for stuff that uses nginx
22:33 murrdoc pypi puts files in a directory
22:34 murrdoc and makes the entry point file have a version number that constructs the dir
22:34 timoguin spent a minute trying to understand the nginx-formula and noped out quickly
22:34 murrdoc ^^
22:34 alexhayes joined #salt
22:34 mosen strange, the apache formula is ok :)
22:34 timoguin that's happened with a handful of formulas
22:34 iggy well... I'm sure it's great for somebody
22:35 iggy just horrendous overkill for most
22:35 iggy (and poorly documented to boot)
22:35 jcockhren "noped out"
22:35 jcockhren lol
22:35 timoguin It'd be fine if the pillar.example had some really simple examples as well as the full-range of what can be configured
22:35 ajolo joined #salt
22:36 skyler Yeah, it would be nice if you could see a minimal pillar as well as a full pillar on some of these formulas.
22:36 simonmcc timoguin: gitshelf was born because I liked the Berkfile model of being able to build a file_root with specific versions of our various formula
22:37 timoguin yea I like what I've seen so far.
22:40 kermit joined #salt
22:41 aqua^mac joined #salt
22:41 simonmcc I’d often wondered if people would have been interested if I added some dependency management via metadata
22:42 thayne joined #salt
22:42 giany joined #salt
22:44 tristianc joined #salt
22:45 otter768 joined #salt
22:47 iggy babilen: is there a mailing list or anything that people with -formulas commit rights are on?
22:48 skyler We should definitely start a thread about what to include in the opinionated policy document.
22:49 iggy I vote nothing, but that's just me
22:51 alexhayes Hello folks, wondering if there are many running salt with libvert to manage xen? and... if they are also using vagrant as part of their stack?
22:51 ksalman i am trying to figure out how does Salt get id grains on Windows. Looking at the source, all I see is https://github.com/saltstack/salt/blob/2014.1/salt/grains/core.py#L703. What is __opts__?
22:52 manfred ksalman:  http://docs.saltstack.com/en/latest/topics/development/dunder_dictionaries.html
22:53 ksalman manfred: thanks.. where is that in the source? =)
22:54 manfred it is injectred in the loader.py
22:54 manfred https://github.com/saltstack/salt/blob/develop/salt/loader.py#L661
22:55 ksalman hm, I am lost I guess. HOw can I find out the source where Salt computes the id grain?
22:55 manfred oh
22:55 manfred it would be in the grains core.py probably
22:56 ksalman All I see is this https://github.com/saltstack/salt/blob/2014.1/salt/grains/core.py#L699
22:56 babilen iggy: There isn't, but lets take it to a ML, yeah
22:56 ksalman it just returns {'id': __opts__.get('id', '')}
22:56 sc` joined #salt
22:56 babilen (or create a -formulas ml)
22:56 manfred ksalman:  that is what is set in /etc/salt/minion_id on linux, no idea where it is set on windows
22:57 ksalman manfred:  oh, it just reads that file?
22:57 ksalman yea that makes sense
22:58 manfred yeah, cause that is how you change it
22:58 ksalman thanks
22:59 pipps joined #salt
22:59 iggy I'd be interested in a -formulas ml (for more reasons than just to discuss this particular topic)
23:00 jalaziz joined #salt
23:00 timoguin ksalman: it'll default to the fqdn if there's nothing in the minion_id file
23:06 skyler -formulas ml would be great (also for more than just the topics discussed here).
23:09 sc` joined #salt
23:10 sc` joined #salt
23:11 ksalman timoguin: how does salt get the fqdn then? The minion_id file has "minion1.example.com example.com"
23:11 ksalman I dodn't know why it has the extra example.com
23:13 murrdoc is there a file.append
23:13 murrdoc oh hell yeah
23:14 sc` joined #salt
23:14 murrdoc sys.doc is like my favorite thing
23:15 iggy yeah, I always have http://docs.saltstack.com/en/latest/salt-modindex.html open somewhere
23:16 ksalman timoguin: nevermind, that is in core.py, though I still don't know why it gets the extra example.com
23:21 nitti joined #salt
23:22 joehh gebi: If you see this, can you email me re salt-ssh on debian jessie? joehealy@gmail.com
23:22 gmcwhistler joined #salt
23:27 arnoldB *facepalm* https://github.com/saltstack/salt/commit/37af3acfe8ea3178b8fc660076c723ad5694bedc this is absolutely nonsense and is breaking things
23:28 alexr joined #salt
23:28 Outlander joined #salt
23:29 skyler arnoldB: What is it breaking?
23:29 arnoldB could someone tell me why 2a03:4000:c:10aa:1017:f00d:aaaa:a shouldn't be a valid IPv6 address? maybe I'm blind
23:29 arnoldB don't hesitate to whois that address
23:30 arnoldB skyler: it breaks salt.states.network.managed
23:33 skyler iggy, murrdoc, timoguin, babilen: I am about to send a message requesting a salt-formulas mailing list unless someone else has already taken steps to get one created or knows a better channel for that request.
23:33 thayne joined #salt
23:33 murrdoc +1
23:36 skyler arnoldB: That appears to be a valid ip6 address...
23:36 arnoldB skyler: PR is on the way
23:36 skyler arnoldB: You fixed the regex?
23:37 arnoldB skyler: I removed it. It makes _no sense_ to have expressions that can't be maintained..
23:37 skyler iggy, murrdoc, timoguin, babilen: https://groups.google.com/forum/#!topic/salt-users/CGvwCq00osI for those interested.
23:38 paha joined #salt
23:38 aquinas joined #salt
23:39 bhosmer joined #salt
23:40 iggy rgr
23:43 catpiggest joined #salt
23:48 murrdoc there isnt a +1 button
23:48 murrdoc thats weak google
23:51 arnoldB salt/utils/validate/net.py is getting my crazy and angry
23:51 arnoldB s/my/me/
23:59 catpiggest joined #salt

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