Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-01-22

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

All times shown according to UTC.

Time Nick Message
00:00 robbyt joined #salt
00:03 SpeeR joined #salt
00:03 ninkotech joined #salt
00:05 UtahDave CheKoLyN: can you pastebin the error you're getting?
00:07 [diecast] joined #salt
00:10 [diecast] joined #salt
00:12 jeremyBass2 i know I'm being a pain in and but here, trying to things in a standard way.. maybe if someone has a sec they could point out what is not right about, http://pastebin.com/Dv9F1MVQ  which i have been trying to ensure matches http://docs.saltstack.com/ref/states/top.html#environments .. i have to believe the issue of the first jinja render path being the env's path stems from this minion.. but i really don't see the issue that is here but
00:14 jeremyBass2 I guess i wounder if it's that i may only use   `base:` `dev:` `qa:` `prod:` in the file_roots: ? maybe that is my issue?
00:15 jeremyBass2 i saw an example a bit ago that reference sites as environments so i didn't think the url would be an issue
00:18 CheKoLyN UtahDAve will do
00:19 scott_w joined #salt
00:20 CheKoLyN Here's the link: http://pastebin.com/raw.php?i=twKSCRXE
00:21 de_bigt joined #salt
00:21 CheKoLyN This works OK when using the master -- minion setup. But when i use the salt-ssh it have that error
00:21 CheKoLyN Seems like a jinja template issue; but i'm not sure what is causing it
00:23 MTecknology UtahDave: is there anything yet about study topics for the salt cert?
00:24 oz_akan_ joined #salt
00:26 yomilk joined #salt
00:31 ajw0100 anyone ever run into any trouble using an sls with a hyphenated name in a requisite statement? are there any sls file naming standards?
00:31 ninkotech joined #salt
00:32 UtahDave CheKoLyN: does  /srv/salt/cherokee/map.jinja   exist on your master?
00:33 UtahDave MTecknology: Uh, let me check
00:34 ninkotech_ joined #salt
00:35 UtahDave MTecknology: There's this:  http://www.saltstack.com/certification/
00:35 jeremyBass2 maybe i can ask a different question, what determents the order the jnija picks up files.. it not a-z it seems, or the order declared in the pillar_roots: or file_roots: it seems..
00:35 UtahDave Lemme see if there's more.
00:37 MTecknology open-book?
00:38 MTecknology I fear I only really have a basic understanding of these topics and not enough to actually pass. I like this, though. It's exactly what I was looking for. Thanks!
00:41 UtahDave You're welcome!
00:41 UtahDave ajw0100: I'm not sure if hyphens are allowed or not. I'd just try with and without hyphens and see what works.
00:42 bhosmer_ joined #salt
00:42 UtahDave jeremyBass2: The minion looks in the top file to see which environments it matches. Then it looks in those environments in the order they are defined in the file_roots.
00:43 ajw0100 @UtahDave thanks. removing the hyphen worked. was just wondering if that's expected behavior and I need to make sure not to use hyphens in sls filenames.
00:43 CheKoLyN UthaDave: yes it does; but for some reason it doesnt find it when using salt-ssh.
00:44 CheKoLyN I'm still working on the formula, but let me see if I can upload it to my github
00:44 scott_w joined #salt
00:45 baffle joined #salt
00:45 UtahDave cool
00:48 mgw joined #salt
00:48 GradysGhost joined #salt
00:49 jesusaurus it looks like my test minions are creating their $cachedir/cache/sls.p files with mode 400. is this something other people are seeing? is it just not noticed because the minion typically runs as root?
00:49 honestly joined #salt
00:50 jeremyBass2 @UtahDave, thank you that is what i was under the impression, everything runs but it's sorting the file_root entries some how and not sure why.. always the same regardless of the order i put it in too
00:51 ninkotech_ joined #salt
00:52 jeremyBass2 here is the file_roots: http://pastebin.com/Cnw6E2Xk ... and i can switch the base for the site env and it will still run in the same order
00:53 CheKoLyN UtahDave: here's the github repo.  The file i'm having issue with is https://github.com/chekolyn/cherokee-formula/blob/master/cherokee/templates/cherokee.conf.jinja
00:53 CheKoLyN but it only happens with salt-ssh
00:54 UtahDave jeremyBass2: is that grains section in your minion's config?
00:54 cedwards I'm trying to use gitfs and the saltstack-formula repo, but I must be missing something.
00:55 jeremyBass2 @UtahDave, yes i believe.. that file is the minions config.. right? ..
00:55 UtahDave jeremyBass2: can you pastebin your top files again?  I don't have them in tabs anymore.
00:55 UtahDave CheKoLyN: just a sec
00:56 UtahDave hey, cedwards!
00:56 cedwards I have fileserver_backend: - git, and gitfs_remotes: - git://github.com/saltstack-formulas/apache-formula.git
00:56 cedwards UtahDave: heya UtahDave
00:56 ninkotech__ joined #salt
00:57 jeremyBass2 @UtahDave sure they are https://github.com/jeremyBass/WSU-serverbase-centos/blob/master/provision/salt/top.sls and https://github.com/jeremyBass/WSUMAGE-base/blob/master/provision/salt/top.sls
00:57 UtahDave CheKoLyN: Hm. There might be some limitations with jinja imports with salt-ssh. salt-ssh is still pretty new and there are some edge cases that we haven't completely ironed out.
00:57 Ryan_Lane2 joined #salt
00:57 UtahDave CheKoLyN: would you mind opening an issue on that?  I'll point a dev at it tomorrow morning.  I'm the last one in the office.
00:58 CheKoLyN No problem. Will do that. Thanks a buch.
00:59 cachedout joined #salt
00:59 jeremyBass2 @UtahDave, oh dang it.. i had the wrong file on git hub for you.. https://github.com/jeremyBass/WSUMAGE-base/blob/master/provision/salt/top.sls is the one that I'm runing.. i
00:59 jeremyBass2 both run.. just the order is not base first
01:00 MTecknology I'm getting sick of being the last on in the office
01:01 UtahDave jeremyBass2: Tell you what. I'm going to have to talk to Tom to make sure my understanding of how top.sls files are merged.
01:01 UtahDave jeremyBass2: Is there any way you can use just one top.sls?  That's the recommended way to do it.
01:01 Ashtonian joined #salt
01:01 UtahDave cedwards: that looks correct.  did you restart the salt-master daemon?
01:03 kvbik joined #salt
01:03 cedwards UtahDave: I did. I'm now trying to call the apache formula, but it's not found.. I guess that's where it gets blurry for me.
01:03 jeremyBass2 @UtahDave no, i need to do what was in the docs where you can pull from different spots for the multi-environment .. this is what i was following http://docs.saltstack.com/ref/states/top.html#how-top-files-are-compiled .. and it is close, just keeps coming out base last ..
01:03 jeremyBass2 Thank you for taking the time to help..
01:04 jeremyBass2 confuses me to no end today what is off..
01:06 GradysGhost joined #salt
01:08 ninkotech__ joined #salt
01:08 UtahDave jeremyBass2: what does your file_roots look like?
01:09 jeremyBass2 @UtahDave right now I'm running http://pastebin.com/Cnw6E2Xk
01:10 jeremyBass2 that is in the vagrant.config in the minions folder here https://github.com/jeremyBass/WSU-serverbase-centos/tree/master/provision/salt
01:14 UtahDave jeremyBass2: My brain is fried right now.  I'm going to have to revisit this in the morning.
01:15 UtahDave I'll keep these tabs open and sort it out tomorrow.
01:15 jeremyBass2 @UtahDave understood, thank you for the hand..
01:15 jeremyBass2 seems I must be missing something small
01:15 ninkotech joined #salt
01:16 UtahDave you're welcome.  Yeah, I'm sure you're right.  Thanks for your patience!
01:16 bemehow joined #salt
01:17 bhosmer joined #salt
01:22 n8n joined #salt
01:26 Gifflen joined #salt
01:29 elfixit joined #salt
01:31 emostar joined #salt
01:31 Linz joined #salt
01:31 emostar Hi All... I'm looking for some examples on using Salt Mine
01:31 emostar i succesfully used it in the past, but can't remember anything about it
01:37 jimallman joined #salt
01:38 thrashr888 joined #salt
01:41 thrashr888 joined #salt
01:42 thrashr888 joined #salt
01:43 thrashr888 joined #salt
01:44 thrashr888 joined #salt
01:46 taion809 joined #salt
01:51 Linz joined #salt
01:51 Linz joined #salt
01:56 drawsmcgraw joined #salt
02:03 ajw0100 joined #salt
02:05 ajw0100 joined #salt
02:06 jacksontj joined #salt
02:07 elithrar joined #salt
02:07 bemehow joined #salt
02:16 aphor joined #salt
02:17 de_bigt joined #salt
02:18 dstanek joined #salt
02:20 Gifflen joined #salt
02:21 Ashtonian joined #salt
02:22 dstanek it looks like the deb i installed (version 0.17.4) doesn't have all of the rabbitmq_user features listed in the docs
02:22 Ashtonian joined #salt
02:22 strgcloud joined #salt
02:23 joehh dstanek: the docs are typically from the git master branch and so are ahead
02:23 ConceitedCode joined #salt
02:24 dstanek joehh: ah i see, thanks
02:25 joehh If you read them from https://salt.readthedocs.org/en/v0.17.4/ or the salt-doc package, they should relate to 0.17.4
02:25 Dinde joined #salt
02:26 joehh It may also be possible to use the modules/states from git as custom states that you deploy via salt and use then...
02:27 favadi joined #salt
02:35 srage joined #salt
02:39 dstanek joehh: it looks like i can do everything i need without going down that route; i just have to do it a different way
02:49 bhosmer joined #salt
03:18 dstanek joined #salt
03:25 thrashr888 joined #salt
03:29 drawsmcgraw Anyone used the S3 fileserver backend with success on version 17.4?
03:30 markizano joined #salt
03:34 srage joined #salt
03:41 MTecknology Is there anyone that's going to saltconf that do much for key signing?
03:47 esogas_ joined #salt
03:48 ninkotech__ joined #salt
03:55 Hefeweizen joined #salt
03:56 MTecknology I think we should have a key signing party at saltconf
03:58 favadi joined #salt
04:00 rojem joined #salt
04:03 yomilk joined #salt
04:04 esogas_ joined #salt
04:09 dmwuw joined #salt
04:09 Gifflen joined #salt
04:14 esogas_ joined #salt
04:27 ravibhure joined #salt
04:27 Gifflen joined #salt
04:28 Guest87648 joined #salt
04:31 xmltok joined #salt
04:38 bhosmer joined #salt
04:40 esogas_ joined #salt
04:49 joehh MTecknology: I think it is a great idea (but I'm not going...)
04:53 MTecknology :(
04:56 cachedout joined #salt
05:11 Ryan_Lane2 joined #salt
05:15 forresta joined #salt
05:15 anuvrat joined #salt
05:20 Gifflen_ joined #salt
05:21 n8n joined #salt
05:40 matanya joined #salt
05:40 MTecknology How new is -defaults: in templating?
05:40 MTecknology I .... I NEEDED THIS!!! I built an ugly hack around it and now realize that it's ugly
05:43 Furao MTecknology: -defaults?
05:44 MTecknology http://docs.saltstack.com/ref/states/all/salt.states.file.html
05:45 Furao ah that had been there for a very long time
05:46 Furao such as 0.10.x
05:46 MTecknology wow
05:48 Gifflen joined #salt
05:48 Furao I'm not sure what is the real use of defaults versus context
05:50 MTecknology I need to play with it...
05:52 MTecknology I'm starting to realize that I wasn't doing anywhere remotely near enough with salt at home... I got too caught up in what I was doing at work
06:19 Furao joined #salt
06:26 bhosmer joined #salt
06:26 Psyke joined #salt
06:27 scuwolf joined #salt
06:33 order joined #salt
06:36 Gareth joined #salt
06:42 oz_akan_ joined #salt
06:42 kermit joined #salt
06:46 Ryan_Lane2 joined #salt
06:46 MedicalJaneParis joined #salt
06:46 middleman_ joined #salt
06:47 aleszoulek joined #salt
06:52 Ryan_Lane2 joined #salt
06:58 linjan joined #salt
07:16 linjan joined #salt
07:24 strgcloud joined #salt
07:26 Furao joined #salt
07:29 ckao joined #salt
07:30 Ryan_Lane2 joined #salt
07:31 ndrei joined #salt
07:37 benturner joined #salt
07:40 matanya joined #salt
07:41 Linz joined #salt
07:42 gurubert joined #salt
07:43 oz_akan_ joined #salt
07:51 jacksontj joined #salt
07:56 Ixan is salt-minion package in epel broken? I tried installing it now and get a dependency problem with python-jinja2
07:57 Ixan seems like jinja has been renamed to python-jinja2-26 in epel, while salt-minion depends on python-jinja2
08:00 balboah joined #salt
08:01 lessismore joined #salt
08:06 ndrei joined #salt
08:07 gammalget joined #salt
08:08 helminthe joined #salt
08:09 Iwirada joined #salt
08:14 bhosmer joined #salt
08:15 Furao joined #salt
08:15 ndrei joined #salt
08:16 leonardinius joined #salt
08:17 Iwirada left #salt
08:18 Iwirada joined #salt
08:23 harobed_ joined #salt
08:24 harobed_ joined #salt
08:29 esogas_ joined #salt
08:30 kolaman joined #salt
08:31 kolaman hi all, I am thinking to deploy salt on EC2 (AWS) environment . .currently all machines have the same host anem (like in ec2 something 10-24-22-11 etc. I want to set the hostname to something better)
08:31 kolaman in our case all servers host a website (different on each minion) and we want to set the hostname using the sitename (like if site is a.ourcompany.com the hostname should be set to similar)
08:32 Eugene You want to set up a provisioning script(invoked by /etc/rc.local on first-boot and then disabled is a good place) to set the hostname from some external source; eg the EC2-provided metadata.
08:32 kolaman I have been searching a lot but found nothing worthy about our scnereo .
08:32 Eugene How that metadata gets set is another matter; I've not played with salt-cloud(at all) enough to know if it can do this, but I don't see why not.
08:33 Shenril joined #salt
08:33 njpatel joined #salt
08:34 kolaman Eugene: in our case we are  not using salt-cloud (unfortunately as the application stack is created by our Dev team)
08:34 Furao kolaman: just use salt to set the hostname
08:34 Furao and use pillar value to define it
08:35 Furao such as here https://github.com/bclermont/states/blob/master/states/hostname/init.sls
08:36 kolaman Furao: any documentation related to this specifically . (actually I am new to saltstack so doesn't now a lot about this) just deployed a server and added some policies whcih were pushed to minions
08:37 Furao specific to hostname?
08:37 kolaman that would be great, if I could get something (like a formula) specific to that
08:39 giantlock_ joined #salt
08:40 Furao kolaman: there is nothing specific on hostname in salt (except auto detect minion id using the hostname) but https://github.com/bclermont/states/blob/master/states/hostname/init.sls is an example of how to set hostname
08:40 Furao this used to be my production formulas
08:41 esogas_ joined #salt
08:41 slav0nic joined #salt
08:41 kolaman Furao: previously to set hostname, I created a shells script which were doing the job of setting hostname, here is the way how we fetch hostname information for our site, this may clarify my snereo a bit http://pastebin.com/GpKCtW9U
08:42 kolaman can I use the shell script with salt (so salt fetch information from this script and than set minion name) ?
08:43 Furao you can convert this into a salt formula, yes
08:44 oz_akan_ joined #salt
08:47 Furao but this sound like an ugly hack
08:47 Furao automate mistake don't make them better :)
08:51 yomilk joined #salt
08:53 kolaman Furao: :)
09:09 batulin-s joined #salt
09:13 Damoun joined #salt
09:13 rojem joined #salt
09:14 rojem joined #salt
09:15 rojem joined #salt
09:25 Mua joined #salt
09:26 zooz joined #salt
09:27 rowleyaj joined #salt
09:27 rowleyaj joined #salt
09:30 kadel joined #salt
09:34 markm joined #salt
09:35 Linz joined #salt
09:40 mugsie joined #salt
09:40 mugsie joined #salt
09:44 zimski joined #salt
09:44 backjlack joined #salt
09:44 oz_akan_ joined #salt
09:44 zimski hi
09:45 zimski i am looking fore a great comparaison between Salt and other solution like cfEngine , Puppet
09:46 zimski if u can help ! cheers
09:47 gmtech From my perspective @zimski, (from a puppet background), salt is simpler, has much better orchestration (but I'd never use it), but puppet is more mature, and puppet's hiera is superior to salt's pillar system
09:47 Furao zimski: http://bit.ly/1mrv9Vq
09:48 gmtech lol @Furao
09:48 zimski thx ! .. @Furao u make my day ;)
09:49 zimski i am reading the salt doc ... and i fall in love ...
09:49 Furao it's not in ruby
09:50 Furao that's enough to win over puppet and chef
09:50 Furao and it don't make you feel like you automate in 1993, it win over cfengine
09:50 Furao that leave ansible
09:50 zimski but i need to find some good points to convince mu boss to implement salt solution !
09:51 Furao zimski: what is your baseline/plateform?
09:51 Furao windows? centos? johndeer tractor?
09:51 zimski i agree with u @Furao
09:51 zimski windows linux macOs
09:52 Furao I develop salt in osx and deploy on linux, I don't know about windows support
09:52 zimski also we plan to implement an openstack
09:52 Furao salt-cloud support openstack well
09:53 gmtech For me the clear winner of salt over puppet would be the orchestration if you want to use it
09:53 gmtech Secondly the simpler language is very easy to learn as its just YAML
09:53 zimski and what about scalability ? need to have load balancer ?
09:53 Furao zimski: there is multi-master no need for LB
09:54 gmtech But your disadvantages will be- much harder to recruit for salt right now, and the best ways of doing automated configuration in puppet are better than the best ways of doing it in salt IMHO
09:54 gmtech No need for multi-master either - get rid of the master entirely
09:54 Furao gmtech: no master?
09:54 gmtech masterless salt - good idea and supported
09:54 Furao copy all pillar and states everywhere?
09:55 Furao how many server you have 3?
09:55 gmtech Actually you don't need to copy all pillar and states everywhere, but why not? The only reason I can think is sensitive passwords/data, but there are ways to deal with that
09:56 gmtech in masterless mode, a box doesn't require network connectivity, won't put any load at all on a master, and in a self-contained way can do all it needs to accomplish the states
09:56 gmtech so its very very scalable
09:57 bezaban but you still need to login to every single machine to do site-wide changes? or am I misunderstanding somethign?
09:57 Furao I did relatively large deployment and the slowest component is file copy between master and minion
09:58 zimski gmtech: can u explain please the key points of how automated configuration on Puppet can be better ?
10:00 gmtech @zimski - essentially its mainly because of Hiera. Hiera allows you to hierarchically define conventions on how configuration data is accessed. This allows you to just have files with data in them for each environment, and Hiera will ensure that it magically gets picked up. No includes to get right in top.sls or other pillar sls files, so its less work. In addition actual files (including binary files) can be committed in place as files and don't need importi
10:00 gmtech into YAML format. You can't do this in saltstack right now
10:00 gmtech saltstack says it includes Hiera as an external pillar, but it just doesn't work
10:01 Furao there is alternative to it such as https://github.com/madduck/reclass
10:02 gmtech Thanks @Furao, I was looking at that yesterday
10:02 Furao but I wrote my own inventory system in django & ext_pillar module to feed pillars value using a REST API
10:02 bhosmer joined #salt
10:02 srage joined #salt
10:03 gmtech is yours the django_orm ext_pillar ?
10:03 Furao no
10:03 gmtech Im going to spend a few hours writing salthiera today which captures 90% of the functionality for salt. I think it would be beneficial for everyone
10:04 Furao it requires django to be installed in "root" python env
10:05 zimski thx guys !
10:15 taion809 joined #salt
10:22 twiedenbein joined #salt
10:29 Mua_ joined #salt
10:30 matanya joined #salt
10:36 jesusaurus joined #salt
10:39 babilen Hi all - Can you think of a nice comparison between puppet and salt (in particular in terms of different features) ? Execs would like a nice summary and I am having a look.
10:40 Furao babilen: same question was asked less than an hour ago please check irc logs (check topics)
10:40 babilen will do, /me reads backlog
10:41 Furao basically it's not in ruby so it win over puppet and chef
10:43 travis-ci joined #salt
10:43 travis-ci [travis-ci] kadel/salt#1 (develop - 1c60b05 : Tomas Kral): The build passed.
10:43 travis-ci [travis-ci] Change view : https://github.com/kadel/salt/compare/3aa7ac0...1c60b05
10:43 travis-ci [travis-ci] Build details : http://travis-ci.org/kadel/salt/builds/1243999
10:43 travis-ci left #salt
10:43 malinoff Hm
10:44 Furao since when travis come here and not in salt-dev ?
10:44 eliasp joined #salt
10:45 oz_akan_ joined #salt
10:53 srage joined #salt
10:55 graingert joined #salt
10:59 diegows joined #salt
10:59 ndrei joined #salt
11:01 bbinet joined #salt
11:08 faust joined #salt
11:10 bhosmer joined #salt
11:15 srage joined #salt
11:22 rojem joined #salt
11:23 david_a joined #salt
11:29 srage_ joined #salt
11:32 taion809 joined #salt
11:33 taion809 joined #salt
11:34 gancl left #salt
11:46 oz_akan_ joined #salt
11:47 oz_akan__ joined #salt
11:51 bhosmer_ joined #salt
11:56 rojem joined #salt
11:58 leonardinius joined #salt
12:02 MohShami joined #salt
12:09 rogst joined #salt
12:13 bhosmer joined #salt
12:16 leonardinius 'clear
12:18 gadams999 joined #salt
12:18 miguitas joined #salt
12:22 dstanek joined #salt
12:24 tempspace good morning
12:25 malinoff Good evening, I'd say :)
12:26 Psyke More like good afternoon XD
12:32 ndrei joined #salt
12:35 ekristen joined #salt
12:44 strgcloud joined #salt
12:44 costi joined #salt
12:47 njpatel joined #salt
12:52 pdayton joined #salt
12:55 mpanetta joined #salt
13:03 blee joined #salt
13:06 elfixit joined #salt
13:18 MaikDerstappen joined #salt
13:19 ChoHag joined #salt
13:21 zooz joined #salt
13:26 rojem joined #salt
13:31 gadams999 joined #salt
13:34 MedicalJaneParis joined #salt
13:35 MohShami joined #salt
13:36 taion809 joined #salt
13:39 bhosmer joined #salt
13:45 xmj Hi, what's the preferred way of distributing SSH public/private keys via pillars?
13:45 xmj relates to https://github.com/saltstack/salt/issues/1543#issuecomment-25098644
13:46 viq What would be the proper way to feed jinja this list? http://pbot.rmdir.de/fTuuzFMKCtxtvCkSulORYg
13:47 markm joined #salt
13:48 ndrei joined #salt
13:48 blee_ joined #salt
13:50 ekristen so to setup a multi-master environment, I just have to share the master key, pillar, and salt data correct? and tell my minions about both?
13:51 jrdx joined #salt
13:53 gadams999 joined #salt
13:53 viq ah, here's the proper way to do what I was asking http://pbot.rmdir.de/BcR6AtgxaXYUpVaK7TuKog
14:04 gadams999 left #salt
14:06 oz_akan_ joined #salt
14:07 gadams9991 joined #salt
14:07 markm joined #salt
14:08 pelme joined #salt
14:10 yaru22 joined #salt
14:11 Gifflen joined #salt
14:12 Gifflen joined #salt
14:15 yaru22 joined #salt
14:15 brianhicks joined #salt
14:18 oz_akan_ joined #salt
14:18 logix812 joined #salt
14:19 srage joined #salt
14:19 juicer2 joined #salt
14:19 logix812 with cmd.wait's they have a - watch: can that list be multiple commands?  - cmd:  1   - cmd: 2 etc?
14:19 Linz joined #salt
14:21 bhosmer joined #salt
14:22 taion809 joined #salt
14:24 taion809 joined #salt
14:25 yaru22 joined #salt
14:25 taion809 joined #salt
14:27 logix812 left #salt
14:28 SeanCorder joined #salt
14:30 tessellare joined #salt
14:33 timoguin joined #salt
14:33 gadams9991 left #salt
14:33 davet joined #salt
14:38 derelm joined #salt
14:38 leonardinius_ joined #salt
14:44 ekristen does this command: ‘salt 'haproxy*' state.sls haproxy.refresh_pool 'pillar={new_minion: minionid}’’ only add the new_minion pillar data just during that run?
14:47 Mua joined #salt
14:48 mgw joined #salt
14:48 Kax_ joined #salt
14:51 kalloc joined #salt
14:52 jimallman joined #salt
14:59 mgw1 joined #salt
14:59 Luijk joined #salt
15:00 dstanek joined #salt
15:00 GradysGhost joined #salt
15:01 leonardinius joined #salt
15:02 jeremyBass joined #salt
15:03 Luijk If I change a pillar then run `salt-ssh '*' state.highstate` the new data is not available on the remote system. Is this expected?
15:04 vejdmn joined #salt
15:08 RizKhan joined #salt
15:08 gadams999 joined #salt
15:09 mgw joined #salt
15:09 rgbkrk joined #salt
15:12 rojem joined #salt
15:14 vejdmn joined #salt
15:14 quickdry21 joined #salt
15:20 xinkeT Luijk: you need to refresh the pillar data using salt '*' saltutil.refresh_pillar
15:22 JasonSwindle joined #salt
15:23 aleszoulek joined #salt
15:23 mgw joined #salt
15:24 racooper joined #salt
15:27 bhosmer_ joined #salt
15:30 vejdmn joined #salt
15:41 Whissi joined #salt
15:42 seanz joined #salt
15:44 psyl0n joined #salt
15:44 psyl0n joined #salt
15:45 esogas left #salt
15:45 MZAWeb joined #salt
15:47 colinbits joined #salt
15:47 bhosmer joined #salt
15:47 colinbits Hi, when I do pillar.get('dicta', {}).update(pillar.get('dictb', {})) I keep getting can not combine nonetype
15:47 colinbits is it not possible to combine two dicts?
15:48 aphor colinbits: help(dict().update)
15:49 colinbits not really helpfull since I see both dicta and dictb on the server, and pillar.get should be returning back a empty dictionary if the dict isnt on the server right?
15:49 colinbits it should be a type if dictionary and not none
15:49 aphor colinbits: pillar.get is returning None.
15:49 colinbits so its not like dictionarya.get then?
15:50 mgw joined #salt
15:50 alunduil joined #salt
15:50 aphor colinbits: you should look into using decorators.
15:51 forresta joined #salt
15:52 aphor colinbits: that way you can encapsulate/modularize your dict merging code.
15:52 colinbits do you have an example of salt decorators in slss?
15:54 aphor colinbits: forget I said anything about decorators. I digress.
15:54 AdamSewell joined #salt
15:54 AdamSewell joined #salt
15:55 abe_music joined #salt
15:55 aphor colinbits: type(pillar.get('dicta',{})), type(pillar.get('dictb', {}))
15:56 Luijk xinkeT: Thank you
15:57 colinbits aphor how can I run that?
15:57 tyler-baker joined #salt
15:57 tyler-baker joined #salt
15:58 colinbits can I use the python interperator with salt running?
15:58 philipsd6 viq: I'm not sure (haven't tested) but I think you can also do it this way: https://gist.github.com/philipsd6/8561310
15:58 cedwards whiteinge: when you have a min, I have a question about implementing formulas
15:59 mgw joined #salt
16:00 viq philipsd6: thanks, might change to that to have less jinja magin in there
16:04 DaveQB joined #salt
16:05 seanz joined #salt
16:06 JasonSwindle joined #salt
16:06 mgw joined #salt
16:08 teskew joined #salt
16:08 jalbretsen joined #salt
16:08 aphor colinbits: https://salt.readthedocs.org/en/latest/topics/troubleshooting/minion.html
16:09 jcsp joined #salt
16:10 colinbits aphor thanks
16:10 UtahDave joined #salt
16:13 anuvrat joined #salt
16:14 forresta cedwards, I assume your question is 'why are they so awesome?'
16:14 vlcn is there a way to get salt to only show me failures?
16:14 robbyt joined #salt
16:14 pviktori joined #salt
16:14 vlcn the scroll back is becoming rather unwieldy to deal with
16:16 forresta vlcn, http://stackoverflow.com/questions/15953082/is-there-a-way-to-display-only-changes-and-errors
16:16 cedwards forresta: looking at my logs I see I'm missing a few things. trying a few more things and I'll try again.
16:16 pviktori joined #salt
16:16 forresta cedwards, ahh gotcha
16:18 cachedout joined #salt
16:26 aleszoulek joined #salt
16:26 forresta aphor, if you find anything missing from those docs feel free to create an issue.
16:28 Mua joined #salt
16:29 SpeeR joined #salt
16:32 Mua_ joined #salt
16:33 kaptk2 joined #salt
16:36 zooz joined #salt
16:37 amckinley joined #salt
16:37 davet joined #salt
16:41 ndrei joined #salt
16:43 Gordonz joined #salt
16:44 mgw joined #salt
16:46 Iwirada left #salt
16:47 funzo joined #salt
16:48 JasonSwindle joined #salt
16:50 gadams9991 joined #salt
16:53 nocturn joined #salt
16:55 cedwards forresta: i've got it working. according to my logs python-git on ubuntu 12.04.x (LTS) was too old for gitfs.
16:57 jcockhren cedwards: pip is the way to go
16:59 tessellare joined #salt
17:01 SpeeR joined #salt
17:03 SpeeR joined #salt
17:05 imaginarysteve joined #salt
17:06 KyleG joined #salt
17:06 KyleG joined #salt
17:06 troyready joined #salt
17:07 foxx joined #salt
17:07 Linz joined #salt
17:08 jdenning joined #salt
17:08 ChoHag joined #salt
17:09 cedwards now I'm trying to use the example pillar data to generate a vhost, but I'm getting a render error
17:09 GradysGhost What's that error?
17:10 cedwards Jinja variable 'id' is undefined
17:10 fllr joined #salt
17:11 srage joined #salt
17:12 scoates YAML is the worst part of Salt. (-:
17:12 martoss joined #salt
17:12 dangra joined #salt
17:12 timoguin well you HAVE to use yaml. :)
17:13 vlcn I don't mind YAML as much as I do jinja
17:13 scoates debugging jinja is the second worst part (-:
17:13 fllr joined #salt
17:13 JasonSwindle scoates:  I LOVE YAML, beats the heck out of Ruby.
17:14 scoates you're comparing apples and tanks
17:14 JasonSwindle Yep
17:14 JasonSwindle lol
17:14 scoates timoguin: technically one does not *have* to use YAML, but nearly 100% of others are, so not using it is edge-case-error city, I'm sure. I've learned to not rock this kind of boat.
17:15 sroegner_ is there any kind of a cmdline yaml checker?
17:15 bhosmer_ joined #salt
17:22 seanz joined #salt
17:31 possibilities joined #salt
17:32 yaru22 joined #salt
17:33 seanz joined #salt
17:34 scoates anyone know if there's some sort of trick to getting pyyaml to never ever output this "id001" stuff?
17:35 diegows joined #salt
17:35 hunter joined #salt
17:36 Mua joined #salt
17:36 Mua joined #salt
17:36 srage joined #salt
17:38 Mua_ joined #salt
17:38 matanya joined #salt
17:43 thrashr888 joined #salt
17:50 pass_by_value joined #salt
17:50 gmtech For those interested, I've written a hiera replacement for salt
17:50 gmtech I'll try and get it published today
17:51 gmtech Works as an external pillar
17:53 nebuchadnezzar gmtech: interesting, but did you look at reclass, it's a hiera predecessor: http://reclass.pantsfullofunix.net/
17:54 gmtech I did, but that's more of an ENC - all I'm after is more structured and convention based pillar data
17:54 gmtech so my pillar stack is really nice and simple and powerful
17:55 UtahDave gmtech: cool!  Looking forward to seeing what you put together!
17:55 Ixan sounds interesting, gmtech. looking forward to it. i've been looking into some tool for handling pillar data too, as i use it to model application environments
17:56 joehillen joined #salt
17:56 Ixan UtahDave: do you know if anyone have been working on aix support? i did post on the mailinglist some time ago, but didnt seem to be much intereset
17:57 imaginarysteve joined #salt
17:58 ndrei joined #salt
17:59 UtahDave Yeah, there's actually been a community member working on it.
17:59 UtahDave Ixan: I think the hardest thing was getting zeromq to build, if I remember correctly.
18:01 rojem joined #salt
18:01 Ixan yeah, i posted, kriberg@gmail.com, about aix in autumn. i've got a patch set on top of the 0.17 branch i've been using
18:01 jankowiak joined #salt
18:03 Ryan_Lane2 joined #salt
18:03 Ixan i'd want to submit a merge request for it sooner or later, but first it would be awesome if there are more people who could test it
18:03 n8n joined #salt
18:06 saurabhs joined #salt
18:07 Ryan_Lane2 UtahDave: wget -O - http://bootstrap.saltstack.org | sudo sh
18:07 Ryan_Lane2 UtahDave: at the very least please make the url https
18:07 ashirogl joined #salt
18:08 UtahDave Ixan: post it again to the email list. I'll try to make sure the other person sees your email.
18:08 Ryan_Lane2 it would also be ideal to sign the shell script and to provide a signature file
18:08 UtahDave Ixan: Would you want us to host an AIX binaries or anything?
18:08 UtahDave good idea, Ryan_Lane2!
18:09 Ryan_Lane2 should I open an issue in github? if so, which repo?
18:09 forresta Ryan_Lane2, There is already an issue open for that
18:10 Ryan_Lane2 uuugghhh the shell script also uses curl --insecure  and  wget --no-check-certificate
18:10 Ixan sounds like a good idea for later on, UtahDave. I have to maintain rpms for aix at my company regardless, so can try to make those available
18:10 Ixan will have to look into some form of unit tests for aix though
18:11 UtahDave cool, thanks, Ixan. Let me know if I can help with that.
18:11 forresta Ryan_Lane2, UtahDave, for the bootstrap: https://github.com/saltstack/salt-bootstrap/issues/269, for the other docs: https://github.com/saltstack/salt/issues/5715
18:11 forresta someone just needs to buy a wildcard cert and install it on the saltstack server so that it covers all the urls.
18:11 tyler-baker Hey guys/gals - having a strange issue with packages not updating to the latest version - http://pastebin.com/QP7RdWQA
18:12 ChoHag joined #salt
18:12 tyler-baker the latest package is 2013.12.1 however salt believe that 2013.12 is the latest
18:12 tyler-baker if I run an apt-get install manually it upgrades
18:12 forresta tyler-baker, can you run that through with -l debug on the end to see if we get anything more there?
18:13 tyler-baker forresta, sure thing one moment
18:16 seanz joined #salt
18:16 tyler-baker forresta, http://pastebin.com/Dqpgq2uU
18:17 eightyeight i have an iptables module question: can i statically maintain two chains, without affecting any other chains in the table?
18:17 eightyeight further, can i prevent from packe/byte counters getting reset?
18:17 eightyeight reading http://docs.saltstack.com/ref/states/all/salt.states.iptables.html, and it doesn't seem clear
18:18 mgw joined #salt
18:19 seanz joined #salt
18:20 forresta Ixan, there was some guy named 'hao' in here a few days ago that was running AIX.
18:20 forresta if he shows up again maybe he can do some testing
18:22 forresta tyler-baker, can you try targeting via the command line for that box? so salt 'name_of_machine' pkg.latest_version 'package_name' ?
18:22 forresta Also, is this when building a new VM, or updates to an existing machine? I'm wondering if the apt database is always updated.
18:22 tyler-baker forresta, sure
18:23 strgcloud joined #salt
18:24 tyler-baker forresta, this is a bit strage
18:24 tyler-baker root@linaro-gateway:/srv/salt# salt 'dispatcher01' pkg.latest_version 'linaro-image-tools'
18:24 tyler-baker dispatcher01:
18:24 tyler-baker
18:24 forresta don't do quotes on the package
18:24 tyler-baker ack
18:25 forresta sorry, should have been more specific
18:25 tyler-baker hrrm same result
18:25 forresta try with -l debug
18:26 tyler-baker forresta, http://pastebin.com/QL00ULa5
18:26 honestly is there anyone here with admin permission on saltstack-formulas?
18:26 JasonSwindle basepi:  ^ UtahDave ^
18:27 forresta I have the ability to merge if that's what you're looking for honestly
18:27 martoss joined #salt
18:27 honestly no
18:27 honestly for some reason I am subscribed to all repositories
18:27 honestly that happened a few days ago
18:28 honestly and I would prefer that to be reverted
18:28 honestly I haven't found a way to do it without manually going through all the repos...
18:30 jtang joined #salt
18:30 Marion joined #salt
18:31 UtahDave whiteinge: do you know why honestly got subscribed to that?
18:32 honestly must be because saltstack is evil and spammy :P
18:32 forresta I believe that happened while whiteinge and basepi were doing some testing on the repo, as I was also subscribed to everything
18:32 forresta I really should make a label in my email for salt
18:33 Corey Such opt-in.
18:33 Corey very spam
18:33 honestly that's not how opt-in works :P
18:33 Corey honestly: Snark aside, you're quite right. Permission is important, yo.
18:34 honestly snark aside, I'm sure it was an oversight
18:34 honestly it's also quite disingenuous that github doesn't have a page to review all watches
18:34 Corey honestly: I find that screaming at them in #github works wonders.
18:35 jtang1 joined #salt
18:35 UtahDave honestly: Yeah, sorry about that.  I think github subscribes people to a lot of things automatically.
18:35 honestly I guess I could give it a try
18:35 honestly I don't thank that was just github :P
18:35 Corey This conversation amuses me, because it looks like both UtahDave and I are being very earnest. "Honestly, it's not THAT bad..."
18:36 UtahDave Corey: Hey, my flight gets in 2 or 3 hours before your talk at Fosdem.  You'll need to be extra hilarious to keep me awake
18:36 Corey UtahDave: We can make that happen.
18:36 Corey I might reuse some SaltConf jokes.
18:37 KyleG left #salt
18:37 JasonSwindle Corey:  How salty of you….
18:37 JasonSwindle I will be here all day...
18:37 robawt joined #salt
18:38 JasonSwindle I cannot wait for SaltConf and all of the Salt puns
18:39 njpatel_ joined #salt
18:39 scott_w joined #salt
18:40 Corey JasonSwindle: You're going?
18:40 JasonSwindle Yep
18:40 gadams999 joined #salt
18:40 Corey (I've started following you on Twitter. And home from the gym, but I digress...)
18:40 JasonSwindle Get in the 27, and leave the 31
18:40 JasonSwindle I saw….. I don't twitter much.
18:40 Corey I'm starting to twoot more.
18:40 JasonSwindle but I do read it.
18:41 mpanetta Twoot? heh
18:41 JasonSwindle I am always jumpy on showing the world my cat-speak English skills.
18:42 blarghmatey joined #salt
18:42 JasonSwindle Corey:  But yes, I will be there.  I am so ready.
18:44 kermit joined #salt
18:45 forresta tyler-baker, is that on a new VM you're just provisioning with Salt, or an existing machine?
18:46 tyler-baker forresta, existing machine
18:46 delkins w
18:46 delkins whoops!
18:46 forresta tyler-baker, does this happen for any other packages, or is it specifically linaro-image-tools? And can you try to update the apt database just in case, then run that command line pkg.latest_version again?
18:47 tyler-baker forresta, we have not noticed it on any other packages. Let me try an apt-update really quick
18:48 bemehow joined #salt
18:49 tyler-baker forresta, http://pastebin.com/HQ2i3b94
18:49 tyler-baker forresta, looks the same to me
18:50 forresta tyler-baker, yea I agree, it's odd that it's just that one package... the versioning doesn't seem to be that odd. Can you see what happens when you do salt 'dispatcher01' pkg.install linaro-image-tools -l debug?
18:51 tyler-baker forresta, sure
18:52 tyler-baker forresta, http://pastebin.com/pWPzvPq3
18:53 forresta tyler-baker, hmm, can you turn the logging level up to debug on that minion, restart the minion service, try the pkg.latest again, then see what the logs on the minon look like?
18:53 tyler-baker Saturday, March 01, 2014
18:53 tyler-baker 1735
18:53 tyler-baker Arrive
18:53 tyler-baker sure
18:55 ChoHag joined #salt
18:56 oz_akan_ joined #salt
18:56 BenCoinanke joined #salt
18:57 seanz left #salt
18:58 tyler-baker forresta, ah there is some good debug here
18:58 tyler-baker http://pastebin.com/jMdpLNHb
18:59 ajw0100 joined #salt
19:00 forresta tyler-baker, interesting, so it looks like dpkg is seeing them as the same thing?
19:00 jeremyBass @UtahDave just wanted to touch bases from yesterday, i did come up with a work around for the moment.  I just simplly call highstate for first the base then the environments of choice.  Works well but it's hacky
19:00 tyler-baker forresta, it sure looks like that
19:02 tyler-baker forresta, I'll talk with the maintainers of that package to see if we can work out whats going on there
19:02 forresta tyler-baker, ok cool! I was just gonna suggest to pull them both down and take a look to see what dpkg was saying regarding any changes, then to actually pull them apart and check :P
19:02 tyler-baker forresta, I'll let you know what the outcome is there, thanks for the help
19:02 forresta tyler-baker, yea np, glad we could get somewhere with it.
19:03 bemehow_ joined #salt
19:04 whiteinge honestly: aw, crap. sorry about that. :(
19:04 bhosmer_ joined #salt
19:04 whiteinge github's team management is not exactly featureful...
19:04 honestly https://github.com/watching exists \o/
19:05 robawt whiteinge++
19:05 * robawt highfives whiteinge
19:05 ndrei joined #salt
19:05 honestly whiteinge: did you give me push access to saltstack-formulas?
19:05 whiteinge honestly: yes. i gave the team you're on access
19:06 honestly ah
19:06 jacksontj joined #salt
19:06 * whiteinge high-fives robawt making a swirling anime-like vortex of awesome
19:06 honestly that would have done it
19:06 whiteinge thanks for that github URL. that's helpful
19:06 sroegner_ whiteinge: hi
19:07 whiteinge sroegner_: hello
19:07 honestly whiteinge: tell people to go there and uncheck auto watch unless they want to be autowatching 100 formulas :P
19:07 honestly and do it before you give them push access
19:07 sroegner_ whiteinge: remember how you gave me push permissions to the formulas org some days ago?
19:08 whiteinge honestly: right. :-P
19:08 whiteinge sroegner_: yeah. (you should still have access)
19:08 sroegner_ whiteinge: i would also like to add some new formula - and now i don't have permissions for that anymore
19:08 forresta honestly, you still had to unwatch all those formulas right?
19:08 \ask is there a way to output the dependency map/tree ?
19:09 forresta whiteinge, ahh yea I notice that as well for what sroegner_ is saying, can't make repos any longer.
19:11 gazprom joined #salt
19:11 whiteinge \ask: not built in to salt. there's a community project that can map it with graphviz: https://github.com/DanielBryan/salt-state-graph
19:11 Ryan_Lane2 I'm writing a custom module, and I'm using __pillar__.get('<pillar>'), but it's not returning any data to me. I'm using salt-call —local
19:11 Ryan_Lane2 is there anything special I need to do when trying to use pillars in a module in a masterless environment?
19:13 Ryan_Lane2 salt-call --local pillar.get '<pillar>' works as I'd expect
19:13 srage joined #salt
19:13 honestly forresta: yes
19:13 honestly forresta: I want my money back!
19:13 ndrei joined #salt
19:13 forresta hey Corey are you around?
19:14 Ryan_Lane2 I see the pillar module uses: salt.utils.traverse_dict(__pillar__, key, default)
19:14 Ryan_Lane2 should I be using that rather than just __pillar__?
19:14 JasonSwindle forresta:  I am
19:14 JasonSwindle lol
19:14 dave_den Ryan_Lane: what exactly is '<pillar>'?
19:15 forresta JasonSwindle, you need to work more
19:15 Ryan_Lane2 any pillar that's set
19:15 Ryan_Lane2 I have a custom pillar I'm trying to get the value of
19:15 dave_den Ryan_Lane2: if you're trying to use a key like 'foo:bar', then it will not work.
19:15 JasonSwindle Ha, I am one the phone and building servers.  I was a chater that did 5 chats at a time in my past life.
19:15 dave_den that's why i ask.
19:15 Ryan_Lane2 I'm not
19:16 Ryan_Lane2 just test_me
19:16 timoguin Ryan_Lane2, not sure if it'll make a difference, but you might try using the __salt__ dict, like __salt__['pillar.get']('fooboo', 'default_value')
19:16 whiteinge sroegner_, forresta: ah, right. github's access control is a tad limited. sec...
19:16 Ryan_Lane2 I guess I'll do that
19:16 \ask whiteinge: Thanks, I'll try that.
19:16 forresta sroegner_, jeez, look at this whiteinge guy, so malicious...
19:16 Ryan_Lane2 kind of annoying to need to go through the module from within a module
19:16 sroegner_ forresta: lol
19:16 Ryan_Lane2 I guess it's not a major difference
19:17 sroegner_ forresta: pushing buttons left and right
19:18 kalloc joined #salt
19:18 SpeeR joined #salt
19:18 Ryan_Lane2 interesting. even the module call doesn't work
19:18 sroegner_ forresta: have a maven-formula waiting
19:18 forresta sroegner_, cool
19:19 nyx_ joined #salt
19:19 sroegner_ forresta: when you install the maven rpm in a Fedora docker container you get 152 dependencies with that!
19:19 forresta uhh thanks?
19:19 sroegner_ forresta: formula just pulls the tarball
19:20 forresta I like the 152 deps bette
19:20 forresta r*
19:20 forresta burn that bandwidth
19:21 cedwards I've got fileserver_backend: - git\n - roots, but it isn't seeing my local sls files for some reason.
19:21 Ryan_Lane2 ugh. ignore me. I'm calling a function from a function and aren't returning the data. it's working fine
19:21 cedwards anything else needed in the master config to make sure both are available?
19:21 gadams9991 joined #salt
19:22 forrest Sweeeeeeeeeeeet the dude who had the forrest nick released it to me, woo.
19:23 mpanetta grats forrest!
19:23 whiteinge forrest: nice!
19:23 ajw0100 joined #salt
19:23 nyx_ I don't understand how/when gitfs synching occurs.  I am working with vagrant and have a state that manages some files from a git repo that another state depends on.  I see the initial 'vagrant up' results in 'file not found' errors yet once I ssh into the VM and run highstate it finds them just fine.
19:24 nyx_ Additionally in a separate environment we are using gitfs for states and pillar data and are seeing cases where highstate does not have the latest data until after a couple retries.
19:25 nyx_ in both cases the git repo in question is public and hosted on github.com
19:28 whiteinge honestly: https://github.com/saltstack/salt/commit/7293456
19:28 whiteinge cedwards: ping (re: formulas)
19:28 honestly whiteinge: cool
19:29 forrest honestly, was there an option to unsubscribe that you found inside auto watch?
19:30 forrest granted the formulas repo isn't spammed too heavily
19:30 ndrei joined #salt
19:30 honestly you can turn off autowatch
19:30 honestly autowatch = if you get push access, you are automatically subscribed
19:30 forrest right, but I'm still subscribed to all the other ones
19:30 honestly I had to manually unwatch all the saltstack formula repos
19:30 forrest ok
19:31 forrest I'll just start filtering them then
19:31 cedwards whiteinge: looks like my original issue was with gitfs, and i've resolved that
19:31 cedwards whiteinge: i am having an issue using gitfs and roots though it seems
19:32 srage_ joined #salt
19:32 aranhoide joined #salt
19:33 sroegner_ cedwards: did you try '- roots' and then '- git' - i think you had it the other way round
19:33 whiteinge cedwards: what's the order of your fileserver_backend setting?
19:33 aranhoide if I say, in a salt state file,  {% set var=pillar.get("something", "") %} and there is no "something" pillar, I get None, and not ""
19:33 aranhoide is this a salt issue or a jinja issue?
19:34 aranhoide (this is in salt 0.17.4)
19:34 cedwards I've got - git and then - roots
19:35 cedwards didn't know order was important
19:35 sroegner_ cedwards: the other way round works here
19:35 cedwards sroegner_: thanks. looks like that works here too.
19:36 whiteinge cedwards: i'm not sure why that is. i suspect bug
19:36 whiteinge aranhoide: that should work as you're expecting. are you positive "something" isn't defined somewhere as None?
19:37 aranhoide whiteinge: yes, if I say pillar.get("something", "something-else") I get "something-else"
19:37 aranhoide it's only "" that gets me None
19:38 sroegner_ sudo -i
19:38 sroegner_ darn
19:39 cachedout Better that line then the next one ;]
19:40 aranhoide and better to mis-paste it in IRC than in the terminal :P
19:40 sroegner_ i so miss focus-on-mouseover
19:41 whiteinge aranhoide: sec. replicating locally...
19:41 aranhoide whiteinge: thanks!
19:42 ChoHag joined #salt
19:42 martoss joined #salt
19:44 Sypher joined #salt
19:45 Corey forrest: Sure.
19:46 synfin joined #salt
19:47 Corey whiteinge: So this crazy dingus I work with is wondering what it would take to add a third transport mechanism to salt; in this case, bladelogic.
19:47 whiteinge aranhoide: no luck here. can you try replicating with this sls: http://paste.fedoraproject.org/70776/19992139
19:47 Ryan_Lane2 is there any way for me to have salt-call default to using --local?
19:47 Corey I haven't gone code spelunking lately into that subsystem; how modular is it?
19:48 whiteinge Ryan_Lane2: yes, set file_client to "local" in your minion config
19:48 Ryan_Lane2 awesome. thanks
19:48 SpeeR joined #salt
19:49 ajw0100 joined #salt
19:49 whiteinge Corey: crazy!
19:49 Corey whiteinge: He's out of his tree. Meet synfin.
19:49 whiteinge synfin: duuuude
19:49 synfin whiteinge: good evening
19:49 whiteinge :)
19:50 Corey ... it's quarter before noon. "Evening?"
19:50 synfin Corey: it is evening somewhere
19:50 whiteinge i was going to let that go due to the aforemented "crazy"
19:50 Corey But yes. Perhaps UtahDave could weigh in on how feasible something like that would be. :-)
19:50 Corey (When in doubt blame UtahDave)
19:51 SpeeR joined #salt
19:51 psyl0n joined #salt
19:51 UtahDave Corey: there's work happening right now to make the transport pluggable, with a specific new transport being developed
19:51 Corey Ooh.
19:51 UtahDave I'm guessing another one would be considered.
19:51 synfin I'm curious on the overall feasibility.  I'm trying to look at ways of introducing Salt into an environment making use of the transport mechanisms that exist now.  Is this a multi-month project or like a week project?
19:52 Corey (I'd never use bladelogic as a transport, but I believe synfin has already gargled that particular kool-aid...)
19:52 UtahDave But it's quite an effort.
19:52 synfin UtahDave: as in man-months?
19:52 Corey synfin: That's mythical!
19:52 SpeeR joined #salt
19:53 UtahDave synfin: I'm not entirely sure, actually.  Once the transport and encryption mechanisms are made pluggable I think it will be much easier to adapt it to a new transport.
19:54 synfin UtahDave: Ah, so both the transportation and encryption is not pluggable at this time?
19:54 Gareth 'lo
19:54 SpeeR_ joined #salt
19:55 synfin Corey: I would never recommend bladelogic, but that is not my choice :-/
19:55 UtahDave synfin: No, they're tied pretty deeply to Salt internals.
19:55 UtahDave hey, Gareth!
19:56 n8n joined #salt
19:56 SpeeR__ joined #salt
19:56 Corey synfin: Remember, Salt started off as a remote execution system; config management was bolted on later after enough people went to Tom and said "yknow... you've got something here..."
19:56 UtahDave yep
19:57 synfin Understood. My theory was it would make migration easier if I could strike it midway to start.  Over time………you get the idea.
19:58 SpeeR joined #salt
19:58 nyx_ are gitfs caches on master only updated on a schedule defined by loop_interval or is that update triggered otherwise.  If it is only on the loop interval I'm struggling to understand how to define a deterministic process using gitfs
19:59 kula you can call 'salt-run fileserver.update' to force an immediate update
19:59 nyx_ my case is non-interactive
19:59 SpeeR joined #salt
20:00 UtahDave nyx_: You could use the event system and a reactor to update it, possibly
20:00 UtahDave thanks, kula!  I thought there was something added to do that, but I couldn't remember.
20:00 nyx_ I was under the impression that the gitfs caches would be updated at the start of a call to state.highstate
20:00 nyx_ (apparently mistaken impression)
20:01 kula yeah, i searched that when i started using gitfs. i'm tempted to make a commit hook that simply calls that
20:01 SpeeR joined #salt
20:01 nyx_ it has been frustrating - not knowing when we can highstate after making changes to files in gitfs
20:02 kula nyx_: what's the process that's updating the git repo?
20:02 gadams999 joined #salt
20:02 nyx_ the repo actually isn't changing in the case I'm looking at right now, it's just that the caches aren't ready when I try to reference them in a fresh VM via vagrant
20:03 SpeeR_ joined #salt
20:03 nyx_ so sometimes I hit the update timer just right or git pulls fast enough and things work without failure other times it fails with missing files
20:03 hunter did gmtech ever post his hiera replacement?
20:04 aranhoide whiteinge: Have you tried outputting the string itself?  I thought I was getting None, but I may be getting "None" (I'm writing the thing to a file and I just get None)
20:04 aranhoide whiteinge: I'll try and check that
20:04 Vye Hey guys, this bug broke our dev environment: https://github.com/saltstack/salt/issues/9696 on Windows. We have the v17.2 installer still but it seems the master needs to be downgraded to. Is there an archive anywhere of the salt RPMs?
20:05 aranhoide how do you apply an isolated .sls?
20:05 SpeeR__ joined #salt
20:05 Ahlee aranhoide: state.sls
20:05 Vye It seems epel only keeps the most recent version and https://github.com/saltstack/salt/downloads is out of date.
20:05 gadams9992 joined #salt
20:05 Ahlee Vye: http://koji.fedoraproject.org/koji/packageinfo?packageID=13129
20:05 glenbot joined #salt
20:05 Vye Ahlee: thanks!
20:05 aranhoide Ahlee: pardon?
20:06 timoguin aranhoide, salt '*' state.sls nameofstate
20:06 Ahlee aranhoide: you want to apply just one sls file to a minion?  target hte minion and run state.sls <state file>
20:06 aranhoide oh neat, thanks!
20:06 aranhoide would salt-call state.sls nameofstate work?
20:06 aranhoide in the minion itself?
20:06 SpeeR joined #salt
20:07 aranhoide nvm, I'll just try, thanks
20:07 Ahlee sorry, I'm not sure, I do'nt use salt-call
20:07 n8n joined #salt
20:07 timoguin aranhoide, yea that'll work
20:08 UtahDave aranhoide: That will work if you have your file_roots and everything set up on your minion
20:08 UtahDave Vye: I'm hoping to get that bug fixed soon.
20:08 SpeeR_ joined #salt
20:10 thrashr888 joined #salt
20:11 Vye UtahDave: Thanks, I appreciate your help! :)
20:12 UtahDave you're welcome!
20:13 MTecknology I need to change my highlighting if that guy is gonna stick around :P
20:13 MTecknology UtahDave: I vote for a key signing party!
20:14 UtahDave MTecknology: yeah, that would be a great idea!
20:16 jeremyBass what is the best way to copy one directory from one location to another?  going the cmd.run way or some salt.states.file way?
20:18 rojem joined #salt
20:19 MTecknology UtahDave: I was sort of being silly, but I actually really want to do that
20:19 salt_whitebelt joined #salt
20:20 tzero What would be the best way to install salt-cloud on saucy? I only see PPAs up to raring, but also see that salt-cloud was merged into the main salt repo
20:20 MTecknology there was a hang out we were going to do too, right?
20:21 UtahDave jeremyBass: copy on the same minion?
20:21 UtahDave tzero: Yeah, if you install the RC, it will have Salt Cloud included.
20:22 aranhoide whiteinge: yes, it doesn't happen in that script.  it turns out my problem is that I was using {{ var }} as the value in a context dictionary, without quoting it.  so a jinja "issue"
20:22 salt_whitebelt when running salt windows.minion cmd.script salt://'script.ps1' what is command i need to use to wait for the machine to reboot before launching another script1.ps1. i want to run script1.ps1 after script.ps1 only if the first one finish and reboot the vm?
20:22 jeremyBass @UtahDave yes i need it to go to the web root folder from the same minion so i can run it then delete it off is the goal
20:23 jeremyBass @UtahDave did you happen to see my note this morning about the work around?
20:23 n8n joined #salt
20:23 whiteinge aranhoide: ah ha! glad you were able to track it down :)
20:23 aranhoide whiteinge: thanks, and sorry for the noise
20:23 UtahDave jeremyBass: yeah, I did.
20:23 whiteinge np!
20:24 UtahDave jeremyBass: I would probably use a cmd.run in that instance
20:24 jeremyBass @UtahDave, i was going to use file.recurse: but didn't want to mess up a state, so ok, yeah guess cmd.run is the way I'll go
20:24 jeremyBass thank you
20:25 whiteinge aranhoide: if you're talking about the context dictionary in file.managed, you may find the {{ var|json }} or {{ var|yaml }} filters useful
20:25 tzero UtahDave: is the RC available anywhere outside of git?
20:25 whiteinge aranhoide: http://docs.saltstack.com/ref/renderers/all/salt.renderers.jinja.html#filters
20:25 aranhoide whiteinge: nice, thanks again
20:25 UtahDave tzero: what OS are you looking for?
20:26 tzero UtahDave: Ubuntu Saucy, if possible
20:27 anuvrat joined #salt
20:27 ChoHag joined #salt
20:28 UtahDave Hm. I don't know if the RC is being built for Ubuntu.  Corey, do you know if that's correct?
20:28 Corey UtahDave: I'm out of the Ubuntu packaging, joehh took that over.
20:29 Corey (It made sense; he's doing Debian and I don't have a Ubuntu prod env anymore)
20:29 UtahDave Ah, ok. thanks.
20:29 Corey But I don't believe there's an RC there.
20:29 Corey I have built one for OS X though!
20:29 Corey brew install saltstack is stable, brew install saltstack --HEAD is develop, brew install saltstack --devel is the RC
20:29 timoguin what about the saltstack/salt-daily PPA?
20:29 Corey That was automatically generating things; I don't believe it's maintained.
20:30 timoguin hmm, last update 11 weeks ago
20:30 Ryan_Lane2 joined #salt
20:31 tzero no worries, it shouldn't be hard just to be patient for the next release and install HEAD in a virtualenv I guess
20:39 zooz joined #salt
20:39 srage joined #salt
20:45 alunduil Are there any known issues with salt-ssh and gitfs backends?
20:50 basepi honestly: (late reply) sorry about the subscriptions.  i'm guessing when we converted everyone to contributors and then added all the repos to that contributor group it subscribed everyone to everything
20:50 basepi honestly: Github *loves* to subscribe all the things.
20:51 honestly yeah.
20:52 bhosmer_ joined #salt
20:54 basepi aha!
20:54 basepi honestly: https://github.com/watching
20:54 basepi go there, uncheck "Automatically watch"
20:54 basepi (that will prevent future unwanted watches, in any case)
21:02 honestly yeah
21:02 honestly I did that
21:06 sroegner_ what happens when i hit 'stop watching all repositories' there?
21:06 sroegner_ anyone tried that?
21:06 forrest you'll stop watching everything
21:06 forrest including your personal stuff
21:06 sroegner_ do they all disappear or remain listed as unwatched?
21:06 forrest That I do not know
21:06 forrest Currently at work so I didn't feel like breaking everything :P
21:07 sroegner_ same here
21:07 saurabhs joined #salt
21:08 forrest That reminds me that I need to go through all the formulas that haven't been updated yet this weekend and get maps going
21:08 aleszoulek joined #salt
21:09 hunter maps?
21:09 forrest jinja maps
21:09 forrest to make the formulas os agnostic in a 'cleaner' way
21:10 Gareth UtahDave: Howdy!
21:10 Gareth ..slightly delay.
21:10 UtahDave :)
21:10 forrest Gareth, where's my golden ticket to SCALE? I thought you were handling that...
21:10 cewood joined #salt
21:10 hunter forrest: I need to do that here as well - is there an example you are planning to work from?
21:11 Gareth forrest: I hid it in a chocolate bar.
21:12 salt_whitebelt what is the command i need to use in a state file to tell minion to wait for the first script to run, reboot then start second one?
21:12 felixhummel joined #salt
21:13 forrest hunter, a lot of the formulas are now updated, Apache is a good example: https://github.com/saltstack-formulas/apache-formula/
21:13 forrest salt_whitebelt, that functionality isn't currently supported
21:13 forrest salt_whitebelt, I opened this issue on it: https://github.com/saltstack/salt/issues/6792
21:14 forrest but it was posed more as a question at that point
21:16 hunter {% from "apache/map.jinja" import apache with context %}   <-- I understadnd the rest but what does "with context" add here?
21:16 hunter I know I've asked that before but I've forgotten
21:17 hunter In the jinja.map file (don't care for that name) line 24 is ''''}, merge=salt['pillar.get']('apache:lookup')) %}''''   --- what does that do?
21:18 forrest the merge basically just smashed it into the associate pillar, under that section
21:19 forrest so if I wanted to overwrite something, I could put it under apache: \n  lookup: \n my_thing: 'asdf'
21:19 salt_whitebelt forrest: i thought the whole purpose of state file is to configure the minion in a desire state but if you don't have logic in .sls state config to wait for the first script to finish and run a sceond script that seems rudimentary sequence.
21:19 jtang joined #salt
21:19 forrest salt_whitebelt, when you say reboot, you mean reboot the server right?
21:20 salt_whitebelt yes
21:21 forrest salt_whitebelt, right, so if you restart the service (and the server), the status of the job, and where it was don't get saved to disk, so it wouldn't know where to pick up and start again
21:21 salt_whitebelt forrest: basic rudimetary sequence seem to be missing from .sls state file if its not able to carry out the wait for first event then execute second script
21:21 forrest salt_whitebelt, if you want to do one sls and then another, or one script then another, that's totally different and is easily resolved by using require
21:23 leonardinius joined #salt
21:23 forrest hunter, context is explained here: http://jinja.pocoo.org/docs/templates/#import-context-behavior
21:23 forrest I've always just thought of it like ensuring the data structure is the same as the file it originally came from, whiteinge might have a simpler/better description though, as I know the Jinja docs are kinda meh when it comes to that.
21:23 salt_whitebelt forrest: all i want to do is doesn't matter if one it require multiple .sls or through multiple module, is to run one script, wait for reboot before second script execute
21:24 forrest salt_whitebelt, right, and as I said, there is currently no ability for the status of a salt state to persist through a reboot.
21:25 forrest There are workarounds you can try to do, such as writing a file that you then check for on a second highstate or something after the box comes up
21:26 JordanRinke joined #salt
21:27 jtang joined #salt
21:27 hunter forrest: Reading that doc tells me "don't ask, just cargo cult it"
21:27 forrest Well, it's nice to understand how it works though
21:29 forrest oh whiteinge, that reminds me that the map.jinja example here might need to be updated: http://docs.saltstack.com/topics/conventions/formulas.html#map-jinja to show the newer method of what a map looks like. Or just an example using the more common structure.
21:30 auser joined #salt
21:30 jtang1 joined #salt
21:31 KyleG joined #salt
21:31 KyleG joined #salt
21:32 hunter forrest: I'm getting confused on the apache formula - is there a sample of the associated pillar data I could follow along with?
21:32 KyleG left #salt
21:33 racooper joined #salt
21:35 jacksontj joined #salt
21:35 forrest hunter, are you looking at the pillar.example?
21:36 hunter no -
21:36 forrest it's in the main dir
21:36 auser left #salt
21:37 hunter crude - I looked up one level but not there. Thanks - that was exactly what I needed
21:37 forrest np
21:38 alunduil joined #salt
21:38 jtang joined #salt
21:38 martoss joined #salt
21:38 enqae joined #salt
21:38 fridiculous joined #salt
21:39 fridiculous hey
21:39 jankowiak joined #salt
21:40 hunter forrest: is there a pattern or doc for writing these things? Like a best practices for how to name/structure files in formular/files/ ?
21:40 enqae joined #salt
21:40 honestly there is a template formula
21:40 honestly by yours truly
21:41 honestly https://github.com/saltstack-formulas/template-formula
21:41 fridiculous how can i use file roots to reference common files for multiple environments?  E.g. I don't want srv/salt/dev to have different files from srv/salt/qa
21:41 enqae joined #salt
21:42 fridiculous but rather common files, that are referred to in base, with jinja
21:42 forrest fridiculous, usually you would create a base environment for that sort of thing.
21:42 enqae joined #salt
21:43 fridiculous so for example,  if i put srv/salt/base/redis in base, but leave srv/salt/prod/ and srv/salt/dev/ empty
21:43 Ryan_Lane2 joined #salt
21:44 bhosmer joined #salt
21:44 fridiculous in my top.sls file, i'd refer to the dev environment for '*dev' minions, BUT i don't a redis folder in dev/ or prod/
21:44 fridiculous but rather it default to base
21:44 fridiculous or something else in common
21:44 forrest I thought you had to lay it out like this: http://docs.saltstack.com/ref/file_server/file_roots.html#directory-overlay
21:44 forrest where you include that file root in each environment that might exist, maybe someone can corect me, I don't mess with file_roots that often at home.
21:46 auser joined #salt
21:46 fridiculous forrest, yeah i'm not particularly concerned with the file_roots, rather how to share files between file roots
21:48 forrest I guess I'm confused, you have base in your file roots there, and then you've defined dev, where you also include the directory with the base content.
21:48 UtahDave fridiculous: let me get you an example. it's really simple
21:48 forrest UtahDave, to the rescue :P
21:49 fridiculous i'm all ears
21:49 fridiculous but i think i found a solution as well
21:49 Ryan_Lane2 joined #salt
21:49 fridiculous mhm
21:50 UtahDave fridiculous: http://pastebin.com/LYPMWM6R
21:51 fridiculous yeah, haha
21:51 fridiculous i see
21:51 forrest Ok that's what I was trying to get at, thanks UtahDave, I should have just made an example.
21:51 UtahDave in the dev and prod environments it will search in each path in the order they are listed
21:51 fridiculous now in this case, dev in dev would override base
21:51 forrest if the file was found there first.
21:51 fridiculous oh ok, order matters
21:51 UtahDave yes, order matters
21:52 ndrei joined #salt
21:53 strgcloud joined #salt
21:53 whiteinge forrest: what is the more common map.jinja structure?
21:53 forrest whiteinge, this structure: https://github.com/saltstack-formulas/apache-formula/blob/master/apache/map.jinja
21:54 forrest as opposed to a lookup table being set
21:54 whiteinge ah, i'm with you
21:55 forrest yea it's not a big deal, I just realized there isn't an example of the commonly used format anywhere in the docs.
21:55 analogbyte joined #salt
21:55 whiteinge i did the two-step process in the docs on purpose with the hope that it would help visualize what filter_by does since the jinja syntax for the one-step assignment is a tad heavy if you're not familiar with jinja
21:56 forrest whiteinge, that makes sense.
21:56 whiteinge now that i'm looking at that agian. it's still not too clear :P
21:56 whiskybar joined #salt
21:56 forrest I don't think you can really be blamed for that one, it's the jinja stuff that is confusing
21:57 martoss joined #salt
21:57 * whiteinge shakes his fist in the general direction of jinja
21:58 forrest hah
21:59 auser left #salt
21:59 philipsd6 hey forrest, I made my pitch for Salt to the $company division leadership, and I based part of the demo on your desertpy presentation. Thanks! :)
22:00 amckinley joined #salt
22:00 forrest philipsd6, wooo glad that it came in handy! I hope it goes well for you!
22:00 forrest well, getting it implemented that is
22:01 philipsd6 I'm facing an uphill battle against Puppet, but I'm hopeful.
22:01 ndrei joined #salt
22:01 JordanRinke joined #salt
22:02 philipsd6 I have a hackathon type contest coming up to perform a specific configuration management task via Salt, and other people will be doing the same with Chef and Puppet and we'll compare.
22:03 philipsd6 I wish I was more knowledgeable about Salt, but I have a few weeks to get more into it before the event.
22:04 yomilk joined #salt
22:04 fridiculous philipsd6 i totally i understand
22:04 forrest philipsd6, lol, you're gonna win so easy man, takes 10 minutes to get salt up and running!
22:05 fridiculous for us, in december we realized we need config management
22:05 jalbretsen joined #salt
22:05 sroegner_ joined #salt
22:05 fridiculous there wasn't even enough time to pitch it to leadership
22:05 philipsd6 forrest: getting our cm systems up and running is out of scope for the comparison. It's what we do with it after it's set up and running that counts! :/
22:06 JordanRinke joined #salt
22:06 philipsd6 and frankly, I have difficulty figuring out how to model our infrastructure effectively. I think I'll do ok for this limited test though.
22:08 forrest philipsd6, Yea start small and get something working, like your base config before apps or something like that
22:08 cachedout You should have a contest to see who can write the fastest state to destroy the other CFM servers.
22:08 cachedout Like BattleBots.
22:09 forrest that would be awesome
22:09 forrest using salt would be kinda cheap though
22:09 forrest you could just use salt-ssh, and rm -rf /*
22:09 philipsd6 cachedout: that does sound awesome! :P
22:10 forrest are the other teams planning what they are gonna do philipsd6?
22:10 forrest I think you could get a big head start if you try to design what you can around some of the saltstack-formulas that exist if you're using that stuff to avoid having to configure that chunk
22:10 mgw joined #salt
22:11 philipsd6 forrest: we have an outline of the task to perform, but I a DBA, and the rest of the guys are on the DevOps side. And it's an internal app, so it's not as simple as installing a package and pushing out a standard configuration file.
22:11 forrest gotcha
22:12 cachedout You should just put forrest on standby. He'd win that contest hands-down.
22:12 philipsd6 I will get as much of a headstart as I can though. And I'm debriefing with a knowledgeable person tomorrow to get more details. :)
22:12 forrest cachedout, I wish!
22:12 forrest philipsd6, Nice, I hope you can kick some ass. Are the other teams just nominating one person?
22:13 Corey Actually, there's another fun game like this called Windows Jenga.
22:13 philipsd6 Yes
22:13 alekibango joined #salt
22:13 Corey Take turns deleting files from C:\system32\ until you hang the box or hit 3 undeletable files in a row.
22:13 Ryan_Lane2 joined #salt
22:13 vu joined #salt
22:13 Corey philipsd6: What's the task, anyway?
22:15 philipsd6 simple in theory: we have to push out a yaml config file for a proxy multiplexer, run a check script to ensure that it's valid.
22:15 Ryan_Lane2 joined #salt
22:15 leonardinius_ joined #salt
22:15 hunter forrest: is it possible to call it something other than map.jinja?
22:16 philipsd6 extra points for integrating with our internal tools, and handling the addition of a new node by updating all the other nodes configs.
22:16 forrest hunter, the map.jinja is the standard name. sroegner_ has been experimenting with a secondary file called settings.sls.
22:17 forrest keep in mind the map.jinja is for OS based stuff, other items can go in your pillar.
22:17 forrest philipsd6, are you using a salt minion/master? or salt-ssh for this?
22:17 leonardinius__ joined #salt
22:18 sroegner_ forrest: hunter: the settings file is to deal with scenarios where settings should be shared between formulas
22:18 hunter eh?
22:18 forrest hunter, let's avoid confusion, no, don't name it something other than map.jinja for your OS based stuff
22:19 forrest that's the standard convention
22:19 sroegner_ you can set a value in its origin settings file, say graphite.port
22:19 sroegner_ then you import graphite.settings wherever you need to point to that port
22:19 philipsd6 forrest: installation is prereq, so I'll be using salt-master + salt-minion.
22:19 hunter forrest: damn
22:20 sroegner_ the settings file has no state data - just jinja data structures
22:20 forrest philipsd6, oh cool. Let me know if you need help, I'm always a fan of winning.
22:20 Ryan_Lane2 joined #salt
22:20 leonardinius joined #salt
22:20 leonardinius joined #salt
22:21 ConceitedCode joined #salt
22:22 leonardinius joined #salt
22:24 leonardinius joined #salt
22:25 philipsd6 If you're willing, I'll definitely ping you for assistance forrest (I want to win too!)
22:25 forrest philipsd6, Yea for sure, I probably won't be around much next week because of saltconf, but let me know!
22:25 philipsd6 will do!
22:26 philipsd6 and enjoy the Saltconf! I can't wait to hear about what comes out after that.
22:26 forrest Thanks, it's gonna be a lot of fun!
22:26 srage joined #salt
22:26 forrest I'll have my slides up at some point as well
22:28 alekibango joined #salt
22:28 leonardinius joined #salt
22:29 david_a joined #salt
22:31 leonardinius joined #salt
22:31 jergerber joined #salt
22:32 leonardinius joined #salt
22:33 ndrei joined #salt
22:33 leonardinius joined #salt
22:36 amckinley joined #salt
22:37 JasonSwindle joined #salt
22:38 psyl0n joined #salt
22:38 chuffpdx joined #salt
22:38 SpeeR joined #salt
22:40 aleszoulek joined #salt
22:40 bhosmer joined #salt
22:41 david_a joined #salt
22:42 Ryan_Lane2 joined #salt
22:43 elithrar joined #salt
22:44 alekibango_ joined #salt
22:44 sroegner_ joined #salt
22:50 auser joined #salt
22:51 tyler-baker joined #salt
22:51 tyler-baker joined #salt
22:52 Gifflen joined #salt
22:56 Mua joined #salt
22:56 amckinley joined #salt
23:08 alekibango_ joined #salt
23:09 jtang joined #salt
23:11 johtso joined #salt
23:12 seb` joined #salt
23:13 _2_vina joined #salt
23:15 amckinley joined #salt
23:17 Ryan_Lane2 can you only configure external pillars for the master?
23:18 Ryan_Lane2 no ability to use them in masterless salt?
23:19 blee_ joined #salt
23:21 UtahDave Ryan_Lane2: I believe that is correct.  I think there is an open issue for that.
23:21 Ryan_Lane2 :(
23:21 Ryan_Lane2 I have a few questions about ext_pillar...
23:21 Ryan_Lane2 so, the master syncs the pillars to the minions
23:22 alekibango_ joined #salt
23:22 elfixit1 joined #salt
23:22 Ryan_Lane2 If I'm using hiera, it wants a key to return a value. how do I sync those pillars?
23:22 Ryan_Lane2 or will the client always look up the pillar from the master?
23:23 UtahDave the client does cache its pillar.   The pillar dict is refreshed right before a highstate is run
23:23 UtahDave or you can execute   salt \* saltutil.refresh_pillar
23:23 Ryan_Lane2 so, let's assume I'm never using states
23:23 Ryan_Lane2 and accessing pillar data from modules
23:23 patrek joined #salt
23:23 Ryan_Lane2 you can't pull all data from hiera, only keys
23:26 Ryan_Lane2 would the module accessing the pillar make a master call, then cache the pillar?
23:26 Ryan_Lane2 heh. I guess I could just try and see what happens
23:26 Ryan_Lane2 it's unfortunate that minions don't have ext_pillar support
23:29 UtahDave yeah, it will cache the pillar
23:29 Ryan_Lane2 a module accessing the pillar will cause it to fetch from the master, though?
23:30 Ryan_Lane2 anyway, I'm going to try some things out :)
23:30 saurabhs joined #salt
23:30 UtahDave Ryan_Lane2: actually, while I was digging through the issues, it looks like ext_pillar might work with a masterless minion
23:30 UtahDave Look here: https://github.com/saltstack/salt/issues/9494
23:31 Ryan_Lane2 is this in just the development release?
23:32 alekibango joined #salt
23:33 Ryan_Lane2 oh, wow, it does indeed work for --local
23:33 Ryan_Lane2 sweet
23:33 UtahDave the comment seems to indicate that 0.16 and 0.17 and later work
23:33 MZAWeb joined #salt
23:39 aranhoide left #salt
23:40 Ryan_Lane2 UtahDave: wait. with ext_pillar, are you just loading specific keys via the config?
23:40 Ryan_Lane2 or will any pillar call trigger an ext pillar lookup?
23:42 alekibango joined #salt
23:42 UtahDave you have to execute  salt-call saltutil.refresh_pillar    to initiate the ext_pillar lookup.
23:43 UtahDave then the pillar dict is cached locally.  At least that's how I've been told it's supposed to work
23:43 UtahDave again, a highstate also will cause a pillar refresh
23:43 Ryan_Lane2 so if I need to load keys from a module call, those keys need to already be getting loaded?
23:43 Ryan_Lane2 it's not a dynamic lookup for keys?
23:47 rojem joined #salt
23:47 Ryan_Lane2 yeah, this seems to be the case. that's kind of rough
23:47 UtahDave just a sec, discussing with Tom
23:47 Ryan_Lane2 hiera requires you provide a key
23:49 Ryan_Lane2 one solution I can think of is to have a function for each module that can specify which pillars are being used, then the ext_pillar could iterate the modules and fetch all of the keys
23:50 UtahDave ok, there's a couple ways you can do this
23:50 strgcloud joined #salt
23:51 UtahDave A wise man once said, "Salt reads your mind... in the future!!"
23:51 alekibango joined #salt
23:51 Ryan_Lane2 :D
23:51 UtahDave OK, so one way would be to just use an execution module to make these dynamic queries.
23:51 bhosmer joined #salt
23:52 UtahDave A regular pillar makes a call to the ext_pillar at minion startup and caches that dictionary in memory on the minion
23:52 jtang joined #salt
23:52 UtahDave a highstate or saltutil.refresh_pillar will cause the ext_pillar to be called again and refresh the minion-side pillar dict again
23:53 UtahDave Now, if you create your own ext_pillar, which could be based off the current hiera ext_pillar, with a little modification,
23:53 UtahDave You can use this pillar module function to call the ext_pillar with a key, or any json serializable data structure and make that query dynamic
23:53 jtang joined #salt
23:53 UtahDave https://github.com/saltstack/salt/blob/develop/salt/modules/pillar.py#L119
23:54 UtahDave so you in your module you'd call    salt['pillar.ext']('ryan_hiera', 'my_key')
23:54 UtahDave and you'd get a dynamic pillar lookup
23:56 Ryan_Lane2 and that would get cached, or require a master lookup every time?
23:56 Ryan_Lane2 it's fine either way with my current use case
23:56 Ryan_Lane2 I think this solution will work for me
23:57 UtahDave well, the second option would require a master lookup every time.
23:57 Ryan_Lane2 ok
23:57 Ryan_Lane2 it's something for me to keep in mind
23:57 alekibango joined #salt
23:57 UtahDave otherwise the regular ext_pillar does one big query at minion startup and caches that in memory on the minion
23:57 scott_w joined #salt
23:57 Ryan_Lane2 I need to write a custom ext_pillar anyway
23:58 Ryan_Lane2 so, it would look like this: def ext_pillar(minion_id, pillar, config_file, keys):
23:58 Ryan_Lane2 salt['pillar.ext']('ryan_hiera', ['key1','key2'])
23:58 Ryan_Lane2 ?
23:58 UtahDave yeah, I think so!
23:59 Ryan_Lane2 ok, cool
23:59 Ryan_Lane2 the only downside here is modules that expect to be able to just call __pillar__.get
23:59 jtang joined #salt

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