Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-10-08

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

All times shown according to UTC.

Time Nick Message
00:00 linjan_ joined #salt
00:02 NV joined #salt
00:02 NV hmm, are there release notes for the salt 2014.7 RCs? like what changed between rc1/2/3?
00:02 ramishra joined #salt
00:03 __number5__ NV, check the changelog at github
00:06 badon joined #salt
00:07 jpaetzel joined #salt
00:12 __number5__ NV, please ignore me, the changeset are too big to review :(
00:16 to_json joined #salt
00:21 murrdoc how to give salt crontab a destination file
00:23 schimmy joined #salt
00:23 murrdoc basically states.cron is putting the crontab entry in crontab -e , i want it to show up in /etc/crontab
00:24 mapu joined #salt
00:25 schimmy1 joined #salt
00:27 murrdoc nice
00:37 fejese joined #salt
00:37 MaZ- joined #salt
00:37 aurynn joined #salt
00:37 hellerbarde joined #salt
00:38 mitsuhiko joined #salt
00:38 aparsons joined #salt
00:46 jbub joined #salt
00:49 skyler If you do a state that has a pkg and a service, do you need to put require: pkg in the service? I have seen examples of both.
00:50 Ryan_Lane skyler: it isn't required
00:50 Ryan_Lane salt will execute states in order
00:51 skyler Ryan_Lane: Okay thanks!
00:51 Ryan_Lane if the package fails to install, then the service state will also fail. that said, it'll still run, unless you use failhard: http://docs.saltstack.com/en/latest/ref/states/failhard.html
00:52 n8n joined #salt
00:52 Ryan_Lane but that'll completely stop execution at that point
00:53 skyler Ah, that is cool.
00:53 Ryan_Lane if you're relying on ordered execution, you should generally use listen/listen_in rather than watch/watch_in, as well
00:53 Ryan_Lane http://ryandlane.com/blog/2014/07/14/truly-ordered-execution-using-saltstack/
00:55 badon joined #salt
00:58 murrdoc does listen_in wait for all states to finish ?
00:58 murrdoc cos das cool
00:59 Ryan_Lane murrdoc: yeah, it does
00:59 Ryan_Lane same with listen
00:59 murrdoc hmm nice
01:01 sag47 joined #salt
01:08 otter768 joined #salt
01:23 TheThing|laptop joined #salt
01:27 aparsons joined #salt
01:32 ramishra joined #salt
01:35 TheThing_ joined #salt
01:37 corvus joined #salt
01:38 muss joined #salt
01:39 rallytime joined #salt
01:41 higgs001 joined #salt
01:46 muss I have a Query - I installed Salt + Salt Cloud onto a fresh Ubuntu 14.04 server using the Salt PPA (The version installed is: 2014.1.11+ds-2trusty1) & i'm testing Salt Cloud with Digital Ocean & the docs here: http://docs.saltstack.com/en/latest/topics/cloud/digitalocean.html reference personal_access_token - except this doesn't seem to work unless its called api_key - once I set that I still get unauthorized doing a --list-images. Which version of Digital Ocean's AP
01:46 mrtrosen joined #salt
01:47 ramishra joined #salt
01:48 schimmy joined #salt
01:48 mrtrosen hi folks, what would I start looking for if when I run one of my salt states I get an error:   Data filed to compile, no matching sls found
01:49 mrtrosen I'm running the command:  salt -G 'roles:webserver' state.sls theboxhq.settings
01:52 schimmy1 joined #salt
01:54 schimmy joined #salt
01:55 murrdoc joined #salt
01:58 joehoyle joined #salt
01:59 schimmy joined #salt
02:00 ramishra joined #salt
02:06 debian112 is the http://www.saltstack.com/saltstack-events/saltstack-demo-webinar  FREE!  I assume it is?
02:06 TheThing|laptop joined #salt
02:07 murrdoc yeah
02:07 murrdoc its free
02:07 murrdoc sing up
02:07 murrdoc listen in
02:09 debian112 Cool, I sent to the team. I am trying move them off puppet, our salt environment has been stable with hundreds of servers.
02:09 schimmy joined #salt
02:12 TheThing_ joined #salt
02:12 ramishra joined #salt
02:13 schimmy1 joined #salt
02:20 Gareth UtahDave: ping
02:21 ramishra joined #salt
02:22 lahwran joined #salt
02:34 yano joined #salt
02:38 ramishra joined #salt
02:40 malinoff joined #salt
02:41 hasues joined #salt
02:50 ramishra joined #salt
02:52 iggy just watched Craig Sebenik's presentation from saltconf... sounds like a rather large setup they've got there
02:53 iggy and it's all physical servers...
03:01 mosen joined #salt
03:03 higgs001 joined #salt
03:04 schimmy joined #salt
03:05 big_area joined #salt
03:05 big_area left #salt
03:05 big_area joined #salt
03:11 quickdry21 joined #salt
03:14 ramishra joined #salt
03:16 chitown iggy: yes, i left and went to a startup. but before i left, we were at about 40k hosts
03:16 chitown yes; all physical... no VMs (well, other than the 10 we used for testing :> )
03:16 big_area joined #salt
03:17 iggy oh, hi
03:18 iggy and since you're not there anymore... when I said "rather large setup" I meant it sounds terrible
03:18 chitown :D
03:18 chitown ya, kinda the reason i left
03:18 chitown i spent all of my time on the smallest problems
03:19 iggy are you still using salt at new place?
03:19 malinoff iggy, could you share the link?
03:19 chitown bc when it only affect 0.1% of the hosts... well, thats still 40 hosts you have to manually fix :(
03:20 iggy it was on youtube on thatch's channel
03:20 chitown yes, that was a prereq for the new gig
03:20 chitown btw, i gave an updated version as a webinar a couple of weeks ago
03:21 iggy did it get recorded?
03:21 chitown i think so... lookng for a link
03:21 chitown it was a promo thing for:
03:21 chitown http://shop.oreilly.com/product/0636920033240.do
03:22 chitown http://www.oreilly.com/pub/e/3199
03:22 chitown the latter is the link
03:22 chitown for the webinar
03:22 chitown same slides as saltconf, just updated
03:23 iggy cool, I'm definitely going to check it out
03:23 chitown if you have questions; craig5 -at- pobox dot com
03:23 chitown or ping me here
03:23 chitown i drop in once in a while
03:24 big_area anyone had success using halite? I've got it installed and accessed the web ui but i can't see any minions or login
03:24 mosen big_area: I've got it running, sorta :)
03:24 iggy I'm in a considerably smaller setup, but I'm in there at the start, so I have next to no legacy
03:24 mosen big_area: hit the loudspeaker thing to run test.ping cmd
03:25 big_area mosen: auth error: no credentials givent
03:25 big_area would this be a system level user or someone i have to create under saltmaster/halite?
03:25 mosen big_area: it depends on how external_auth is setup on the master
03:26 big_area ah, i think i see the problem then
03:26 mosen big_area: I think it can be local or in some other authentication system, Im a bit hazy on what a local setup would contain
03:27 big_area i'm going off the wiki tutorial so i think i need to specify one of my local users under external_auth: pam:
03:27 big_area ie s/testuser/<mylocaluser>/
03:30 big_area hmm
03:30 big_area that didnt seem to work either
03:31 mosen I added the @runner and @wheel groups, even though that may be expanding the perms too far
03:32 higgs001 joined #salt
03:32 UtahDave Gareth: pong!
03:32 mosen and is it showing you logged in at the top right corner?
03:33 big_area no, logins are failing
03:35 malinoff chitown, looks like Craig's team is fighting with salt itself rather than business challenges
03:36 chitown well, sort of...
03:36 chitown i mean; we saw things that are really hard to see unless you have 20,000 minions
03:37 chitown things like; the disk isnt fast enough (for pki dir)
03:37 chitown but... the business kept "pushing forward" when i wanted to stabilze what we had
03:37 malinoff but he wasn't talking about such issues, I saw slides about "tons of custom code" and "fork with additional version number"
03:38 chitown as an example, we sitll had a couple hundred minoins with random problems and "they" wanted to support mac os
03:38 chitown uh... "he" == me
03:38 malinoff Oh, so you're Craig
03:38 chitown tons of custom code = runner and modules, not custom "salt core"
03:38 chitown yes :)
03:39 Ryan_Lane joined #salt
03:39 chitown the core chanegs were *all* contributed upstream
03:39 chitown *always*
03:39 chitown we had the additional version number just in case we had to roll a new version before the chanegs couled be committed
03:40 chitown which happened a total of 6 times (???)... maybe 8
03:40 malinoff I used salt for a year in outsourcing company as devops to deploy many different kinds of projects, i had custom modules too, but only to enforce directory structure and command execution (like there must always be build.sls, deploy.sls, configure.sls states and so on)
03:40 malinoff But I faced an other issue - i couldn't use salt because salt-master and salt-minion must have the same version
03:41 iggy master just needs to be higher than the minion (for the most part)
03:41 chitown hm... not sure thats 100% true
03:41 chitown there were def some API jumps
03:41 iggy there have been very few changes that would make an old minion not work with a new client
03:41 chitown eg. iirc, you cannot use, say, salt master 0.17 with minion 0.14
03:41 kedo39 joined #salt
03:42 chitown (not sure of the exact versions, but it was around that time)
03:42 chitown they def do not have to be exactly the same
03:42 malinoff iggy, that's not true, I had issues with such setup
03:42 chitown yes, there are def some issues
03:42 chitown the biggest problems we faced were actually with zeromq "issues"
03:42 Gareth UtahDave: hey :) I was looking at that ps issue...the one with pgrep and top.
03:42 mosen big_area: can you run the master in foreground w/debug? I know it may not be possible if its in prod
03:43 UtahDave Gareth: yeah, I saw your PR
03:43 iggy I have had old minions with new masters
03:43 chitown at 40k minoins; keeping them up to date wasnt really our biggest problem
03:43 big_area luckily im not in prod yet
03:43 big_area mosen: i will
03:43 iggy yeah, 0mq having issues... lol, just lol
03:43 chitown whenever we upgraded, the master went first jsut in case
03:43 chitown zeromq... ya, blessing and a curse :/
03:44 iggy we are wondering if raet is going to fix some of our issues
03:44 iggy and we are only in the 100 node range
03:44 chitown as an aside, we bitched about 0mq *CONSTANTLY* !!!!....
03:44 malinoff chitown, i had around 200 minions and i got connection issues, but they were about re-authenticating (having a single AES key and re-generating it each time a minion removes from the registry is a kind of architecture problem)
03:44 UtahDave Gareth: I think it's funny that test is failing now. The function wasn't working at all, and NOW the test decides to fail
03:44 chitown but, when all was said and done, we had another big project that required async comm (for our metrics infra) and we use 0mq
03:45 iggy I think some of it is just GCE networking issues
03:45 big_area mosen: i think these are relevant:
03:45 big_area [INFO    ] Clear payload received with command mk_token
03:45 big_area [WARNING ] Authentication failure of type "eauth" occurred.
03:45 chitown malinoff: ya, that is a problem... we had discussed a solution; basically having a way where ou could have N keys and just cycle thru them
03:45 Gareth UtahDave: yeah.  I had originally pinged because I was wondering if it ever worked with psutil > 2.x.x and now the tests are failing.
03:46 mosen big_area: welp..
03:46 mosen big_area: thats a whole lot of nothing
03:46 chitown but, again; we were so busy with other stuff, i never got a chance to make any headway with the "cool shit" i wanted to do
03:46 chitown which is why i left linkedin
03:46 chitown well, one of many reasons
03:46 malinoff chitown, chef solves this problem in a very beautiful way
03:46 Gareth UtahDave: I think that fix is right though.
03:47 iggy that's why I love where I'm at now... it's all cool shit, building from the ground up
03:47 big_area mosen: could this be an issue?
03:47 big_area [DEBUG   ] Missing configuration file: /home/vagrant/.saltrc
03:47 UtahDave Gareth: Hm. I really have no idea. I hadn't used that module before.
03:47 UtahDave big_area: nah, that's not a problem at all
03:47 possibilities joined #salt
03:47 chitown so, i dont know much aout chef, so i am hesitant to debate it... but, based on other convos; sure it helps with some things, but other things are worse
03:47 UtahDave big_area: make sure your eauth is working before trying with halite
03:48 malinoff chitown, yes, but right now i'm talking about the architecture. Using nginx to proxy requests to upstream server and retrieving job data only via rest api using SSL/TLS certificates is really awesome
03:48 mosen big_area: probably not
03:48 mosen big_area: is the user coming from the local system?
03:49 big_area yes
03:49 mosen big_area: ok sorry just makin sure
03:49 iggy so is having to wait around for the agent to pull data whenever it gets around to it
03:49 Gareth UtahDave: Oh. hm. I think I just found it.  the test has a similar line that the module had..
03:49 malinoff iggy, yes, this is a problem with chef, but you can enforce pulling e.g. via ssh
03:50 mosen big_area: Any chance you can pm or pastebin the external_auth section?
03:51 iggy as was said, some things easier, some things harder
03:51 malinoff iggy, again, i'm not saying that chef is the best, of course it has concerns, but from the architecture view it is way better
03:51 iggy personally, I'd rather castrate myself with a rock than write ruby code all day
03:51 iggy heh, okay
03:51 malinoff iggy, and this is an other problem with chef, yes
03:52 malinoff iggy, using ruby to describe recipes is, uhm, crypticle, with all these template blah do |something|
03:52 chitown ya, after i finish the book, i am going to take some time to play with chef, puppet and ansible
03:52 big_area sure, just doing this on a throw-away vagrant. standby
03:53 malinoff chitown, i suggest to not spend your time on puppet
03:53 UtahDave Gareth: ah, cool
03:54 chitown as sun tzu said; know thy "enemies"
03:55 mosen malinoff: the only bit i miss from puppet is the self contained module installs. But sometimes its a bit of a leaky abstraction anyway
03:55 chitown i.e. i cant realy talk about salt's stengths and weaknesses without knowing more about the other major players
03:55 malinoff chitown, and ansible is... well, a one-man makeshift project
03:55 big_area mosen: http://pastebin.com/zknC14bH
03:55 ramishra joined #salt
03:55 chitown linkedin is a large cfengine shop, too
03:56 chitown that shit pissed me off... :)
03:57 mosen big_area: welp thats pretty much identical to mine
03:58 malinoff chitown, I used salt for a year, now i'm using ansible for 4 months, and i found they're not really suitable for my needs. IMO salt and ansible are just wrappers around bash scripting, they have no cluster-oriented approach
03:58 malinoff well, of course you can workaround this and create custom modules/states, but i'd like to see this in the project itself
03:59 malinoff there is no single point of truth, from where nodes can pull information about the cluster state
04:00 schimmy joined #salt
04:01 kermit joined #salt
04:02 malinoff and the need to have a dedicated "control" node to push commands from does not fit the cluster approach
04:02 malinoff Please, don't think i'm complaining, just sharing my thoughts
04:04 dccc_ joined #salt
04:05 schimmy joined #salt
04:08 big_area mosen: i run the following to test auth:
04:08 big_area sudo salt -a pam saltmaster test.ping
04:08 big_area and get:
04:08 big_area Failed to authenticate, is this user permitted to execute commands?
04:08 big_area do you think thats an issue in the eauth config or with my system user permissions?
04:10 __number5__ big_area: try add -l debug to see if something in the errors pop up?
04:11 big_area a few modules get loaded but nothing more about the error
04:14 TheThing|laptop joined #salt
04:16 pdayton joined #salt
04:23 mosen big_area: I usually mess up the permissions
04:24 big_area well im wondering if it a result of the sudo perm's that this vagrant box has
04:25 big_area or it might just be that i'm tired and changed roots password and not my users
04:26 mosen can you run anything as the vagrant user?
04:27 big_area ya i can
04:27 big_area i just figured it out
04:28 big_area when i was setting the pssword for vagrnat earlier i forgot to specify the user so it changed root's password
04:28 big_area probably a sign that i should get some sleep
04:29 ajolo joined #salt
04:29 mosen ahh
04:32 higgs001 joined #salt
04:33 res0nat0r joined #salt
04:35 murrdoc joined #salt
04:37 big_area mosen: UtahDave thanks for the help
04:37 UtahDave you're welcome. I'm glad you figured it out!
04:38 mosen no probs im a rank amateur, glad i could help
04:39 kusams joined #salt
04:42 possibilities joined #salt
04:48 quickdry21 joined #salt
04:48 felskrone joined #salt
04:49 ramteid joined #salt
04:51 Ryan_Lane joined #salt
04:52 ramishra joined #salt
05:02 jalbretsen joined #salt
05:25 possibilities joined #salt
05:28 heewa joined #salt
05:32 possibilities joined #salt
05:34 vlcn joined #salt
05:35 goodwill joined #salt
05:35 murrdoc https://github.com/saltstack/salt/commit/e3a556dc464a3d3e864c883e5465d17e22df54a3 could have used querystring = urllib.urlencode(list(zip(keys, values)), {' ':'%20'}) instead of the querystring.replace
05:35 murrdoc could*
05:36 UtahDave joined #salt
05:41 n8n joined #salt
05:45 n8n_ joined #salt
05:46 roolo joined #salt
05:46 n8n_ joined #salt
05:48 desertigloo__ joined #salt
05:49 ndrei joined #salt
05:49 johtso If I'm installing various python packages, and have each component running under a separate user, what user would you typically do the pip installation as?
05:50 johtso the user needs to have the deployment private key, but they also need to have permission to modify the component user's virtualenv
05:51 n8n joined #salt
05:53 hardwire joined #salt
06:03 ttrumm joined #salt
06:04 ttrumm_ joined #salt
06:05 colttt joined #salt
06:05 agend joined #salt
06:07 duncanmv joined #salt
06:13 n8n joined #salt
06:13 gadams joined #salt
06:17 jalbretsen left #salt
06:20 ramishra joined #salt
06:23 agend joined #salt
06:24 Nexpro1 joined #salt
06:25 jonasbjork joined #salt
06:27 lcavassa joined #salt
06:28 stanchan joined #salt
06:39 schimmy joined #salt
06:39 flyboy joined #salt
06:41 schimmy1 joined #salt
06:43 mechanicalduck_ joined #salt
06:45 n8n joined #salt
06:47 CeBe joined #salt
06:47 ramishra joined #salt
06:54 ramishra joined #salt
06:57 trikke joined #salt
06:57 jalaziz joined #salt
06:59 masm joined #salt
07:02 steve1 joined #salt
07:03 dvestal joined #salt
07:03 stewba joined #salt
07:11 bhi_ joined #salt
07:11 zerthimon joined #salt
07:15 johtso is it possible to "require" all actions of a particular declaration/state id?
07:18 chiui joined #salt
07:19 ramishra joined #salt
07:22 bhi_ joined #salt
07:22 linjan_ joined #salt
07:24 ramishra joined #salt
07:24 CycloHex joined #salt
07:30 quanta_ joined #salt
07:30 jonasbjork joined #salt
07:30 ramishra joined #salt
07:31 quanta_ I'm trying the multi_returner
07:31 quanta_ after adding something like:
07:31 quanta_ multi_returner:
07:31 quanta_ - syslog
07:31 quanta_ - sentry
07:32 quanta_ into the minion config file
07:32 quanta_ and run `… -l debug --return multi_returner`
07:33 CycloHex Is setting an expire-date not supported in salt 2014.1.10? I have it, it gives me an error saying, KeyError 'shadow.set_expire'
07:33 quanta_ I saw that it stopped at this line: https://github.com/saltstack/salt/blob/develop/salt/returners/multi_returner.py#L60
07:34 quanta_ but nothing in syslog or sentry
07:34 quanta_ running with `—return syslog,sentry` worked fine
07:34 quanta_ so I would like to debug the `mutli_returner`
07:35 ramishra joined #salt
07:37 akafred joined #salt
07:47 TyrfingMjolnir joined #salt
07:48 martoss joined #salt
07:53 CycloHex I created a new user, his password is in clear text in /etc/shadow
07:54 CycloHex It's not normal that shadow has plain text.. does this mean I have to hash my pw before setting in through salt? This would be weird, because another user I add, has his password hidden in /etc/shadow
07:58 intellix joined #salt
08:08 rattmuff CycloHex: If you use "user.present it should be a hash in your state file
08:12 CycloHex rattmuff, which encryption do I need then? just the md5?
08:13 CycloHex and is it possible to just give a plain text pw?
08:14 rattmuff CycloHex: I'm not sure how you could pass a plain text pw. I think you can use md5, blowfish, sha256 and sha512. YOu can find a good article on shadow hashes here: http://www.aychedee.com/2012/03/14/etc_shadow-password-hash-formats/
08:15 rattmuff CycloHex: I'm not sure it is a good idea to send a PT password either since it will then be stored in PT on your master.
08:15 rattmuff CycloHex: Better to store hashes :)
08:15 wr3nch joined #salt
08:15 wnkz joined #salt
08:16 CycloHex I thought I'duse pillars for this.. But yes, hashes are the way to go
08:17 baconbeckons joined #salt
08:18 rattmuff CycloHex: We store our user data in pillars and then use a jinja map to map the data to a state. We set hashes and auth_keys in the pillar. Works very nicely. Based on this: https://gist.github.com/viq/ec276c183bd84e90606d
08:20 scalability-junk What is needed to run salt-local? Python only?
08:20 CycloHex Thanks, rattmuff !!!
08:20 CycloHex you ar eusing the salt-user formula, right?
08:22 rattmuff CycloHex: yes, that's the one :P
08:23 babilen There's also a reverse users formula if you prefer that
08:24 herzi joined #salt
08:27 babilen CycloHex: Use "mkpasswd -m TYPE [PASSWORD] [SALT]" to generate hashes. crypt(3) lists valid TYPEs
08:28 CycloHex babilen, reverse user, what would that mean?
08:28 CycloHex and thanks, I already have my hashes in my pillar ^^
08:28 babilen CycloHex: https://github.com/saltstack-formulas/reverse-users-formula
08:29 babilen (it simply reverses the order in which you list data from hosts: [users, ...] to user: [hosts, ...])
08:30 CycloHex ah, So if I'd like to have a user that can login to multiple computers I'd better use reverse user
08:32 CycloHex btw, anyone has experience with the 'expire' param from the state user?
08:32 CycloHex I try to setup an account in sudo, but it should expire in 2 months
08:35 chiui sigh the previous version .deb was deleted
08:35 chiui so no more salt local
08:38 babilen chiui: I tried to convince them to not release .11 without that patch ... I'm sorry.
08:39 chiui now you could try to convince them to host a repo with all recent version s;)
08:40 kusams joined #salt
08:41 babilen That's not how Debian repositories work (and probably a few other repositories too), but I would argue that they should make a number of suites available (e.g. "stable", "testing", "bleeding", ...) and also provide snapshots of old versions on a dedicated website (akin to snaphot.debian.org) -- Which OS are you on?
08:42 babilen But then it doesn't really matter which derivative it is (if you don't use Debian) -- You should have all old versions in /var/cache/apt/archives/ unless you ran "apt-get autoclean" or "apt-get clean" ...
08:43 babilen And .10 has, unfortunately, also not been uploaded to wheezy-backports by joehh so you cannot grab it from there either.
08:43 babilen Upstream decided to release a .12 soon though
08:44 babilen (and .7 will also be released very soon™)
08:45 PI-Lloyd anyone know if the salt-cloud stuff was also updated in 2014.1.11 ?
08:46 PI-Lloyd getting the following since updating to 2014.1.11 - https://bpaste.net/show/bc2c1b769f37
08:47 babilen I will not update to .11 yet and wait for .12
08:48 PI-Lloyd unfortunately, the system decided to auto update (I thought we disabled that but regardless)
08:50 ramishra joined #salt
08:53 chiui btw I'm on ubuntu
08:53 chiui at least for work
08:53 TheThing|laptop joined #salt
08:53 chiui I prefer fedora, which among other things keeps a lot of older versions available (maybe all)
08:55 ghartz joined #salt
08:55 PI-Lloyd babilen: any idea on when .12 is due to be released?
08:59 fredvd joined #salt
09:00 ollins joined #salt
09:01 imanc how can I get a file from the minion?
09:02 intellix joined #salt
09:05 PI-Lloyd godamnit, why is 1.10 not available in the repos anymore... this is highly annoying
09:06 chiui I had a copy of the deb files (-minion and -common)
09:06 chiui but they have some custom dependencies which are missing as well
09:08 ramishra joined #salt
09:11 ramishra_ joined #salt
09:13 linjan_ joined #salt
09:15 ramishra joined #salt
09:15 zsoftich_ joined #salt
09:16 intellix joined #salt
09:18 fragamus joined #salt
09:19 wnkz__ joined #salt
09:20 PI-Lloyd anyone know of a trustworthy source that I can download the 1.10 .deb files from?
09:21 nliadm joined #salt
09:22 chiui here, for a few time https://s3.amazonaws.com/menda/other/salt-common_2014.1.10%2Bds-1trusty1_all.deb https://s3.amazonaws.com/menda/other/salt-minion_2014.1.10%2Bds-1trusty1_all.deb
09:24 skarn joined #salt
09:24 PI-Lloyd ta, don't suppose you have the salt-master and salt-cloud debs too ?
09:25 chiui I did not download them, sorry
09:25 chiui this is what is needed to run salt local
09:27 PI-Lloyd ahh, yeah we run a saltmaster.... just noticed those are for trusty as well, we are still on precise atm, so they probably wouldn't work anyway
09:27 PI-Lloyd thanks anyway
09:27 chiui I really hope next time there will be a way to get older versions :)
09:27 chiui or that my company switches to rhel :P
09:28 Ironhand I'm running salt with the master and minions on Debian Wheezy (packages from saltstack.com), and often when I run a command such as "salt '*' test.ping" I don't get a response from all minions (even the minion running locally on the master); is this normal?
09:29 Ironhand other times, esp. a few minutes after I tried for the first time, minions do respond normally, without having restarted anything
09:29 PI-Lloyd yeah we have run into the same issue as well
09:29 Alan_S joined #salt
09:30 PI-Lloyd not entirely sure what causes half of the minions to not respond on the first run.... kind of annoying, but we've just learnt to live with it
09:31 Ironhand what bothers me more is that there's no way to tell that minions aren't responding other than simply knowing that they *should* be there - no error messages of any sort
09:31 Ironhand and from the fact that it takes a while to run even if not a single minion responds I assume that salt *is* aware that there are minions to ping
09:31 Ironhand (I've never seen a "False" response)
09:32 Twiglet_ You can use -v and it'll show which ones didn't respond
09:33 Ironhand ah, that's useful info, thanks
09:33 Twiglet_ np
09:33 pacopablo joined #salt
09:35 istram joined #salt
09:36 imanc if I have dumped a postgres database on a minion, and need to get that .sql file to the master (ultimately to my local dev machine), how can I do this with salt?
09:38 TheThing_ joined #salt
09:41 linjan_ joined #salt
09:43 babilen And you will have to wait for your minions to authenticate. The best way it to ping them all and wait for a little while. See https://github.com/saltstack/salt/issues/15415 for details on that, Ironhand
09:47 Ironhand babilen: thanks; from reading the discussion it sounds like the problem has been resolved in development versions; can anything be said about how long it usually takes for these kinds of fixes to make it into the stable versions (such as the Debian Wheezy packages)?
09:51 babilen Ironhand: It will be in 2014.7 which is, AFAIR, scheduled for release later this week/next week.
09:54 Ironhand babilen: ah, that's good news, I'll just wait for that and not bother with dodgy workarounds then
10:07 rattmuff imanc: Maybe you can use cmd.run on the minion to execute this: http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.cp.html#salt.modules.cp.push
10:09 faust joined #salt
10:11 imanc hi rattmuff thanks - looks like the ticket!
10:13 asmaps joined #salt
10:27 giantlock joined #salt
10:34 rattmuff imanc: If you don't want the minion to push files to the master you could also use a reactor and a runner that connects from the master to the minion and pulls the file using rsync or scp
10:35 rattmuff imanc: then the master can have an ssh key that is authorized on all your minions that need file transfers
10:40 dvestal_ joined #salt
10:41 wnkz joined #salt
10:42 fe92 joined #salt
10:44 bhi_ joined #salt
10:46 ndrei joined #salt
10:49 diegows joined #salt
10:56 intellix joined #salt
10:57 fejese hi, my masterless setup just stopped working in means it seems to ignore the pillar data completely. can anyone suggest me a way what can I debug to check why the pillars are not loaded?
10:58 babilen Did you upgrade to .11 ?
10:58 babilen (or is this on .11?)
11:00 fejese 2014.1.11
11:01 fejese is that an issue?
11:01 stephanbuys joined #salt
11:01 PI-Lloyd babilen, just fixed that issue we were having with salt-cloud.... someone put some stupid code in... looks like they noticed the mistake and have patched it in the dev branch.. looks like someone decided to force salt-cloud to throw an exception regardless of scp outcome
11:02 stephanbuys hi all, is there an effective way to troubleshoot gitfs for salt states? my gitfs ext_pillar works 100%, but no luck with state files
11:03 babilen PI-Lloyd: Lovely
11:03 PI-Lloyd this is the offending commit - https://github.com/saltstack/salt/commit/28ece7a443fe10d76e8c62cdf55ae422b541b9b2  || i've just changed line 922 temporarily to "def scp_file(dest_path, contents, kwargs, allow_failure=True):" and works
11:04 bhi_ joined #salt
11:04 babilen fejese: See #16210, #16212 and #16416
11:05 babilen I'm sorry, I tried to prevent it but have no say
11:05 babilen PI-Lloyd: Should that be backported into 2014.1 for the upcoming .12 release?
11:05 babilen PI-Lloyd: Or rather: Was that always broken in 2014.1 or is this behaviour a regression?#
11:06 PI-Lloyd it was only broken in 1.11, so regression
11:06 babilen .11, release of pain
11:06 PI-Lloyd looks like UtahDave has pathed it in develop here - https://github.com/saltstack/salt/pull/15454/files
11:06 PI-Lloyd patched*
11:07 babilen PI-Lloyd: Okay so we would want that backported to .1 ?
11:07 babilen (if so, please file a bug about that!)
11:08 PI-Lloyd release of pain is somewhat an understatement :p
11:08 PI-Lloyd but yeah, will file this to get it backported
11:10 fejese babilen: yes, exactly same problem. will try to downgrade to .10 it's a shame though ...
11:11 jeremyb joined #salt
11:11 mik3 joined #salt
11:11 FL1SK joined #salt
11:11 mortis_ joined #salt
11:11 minaguib joined #salt
11:11 al joined #salt
11:11 TaiSHi joined #salt
11:11 ahale joined #salt
11:11 \ask joined #salt
11:11 dvestal joined #salt
11:11 bnikolaus joined #salt
11:11 chiui joined #salt
11:11 georgemarshall joined #salt
11:11 Sway joined #salt
11:11 rogst joined #salt
11:11 beardo__ joined #salt
11:11 drybjed1 joined #salt
11:11 vu joined #salt
11:11 quanta_ joined #salt
11:11 Zuru joined #salt
11:11 [vaelen] joined #salt
11:11 nliadm joined #salt
11:11 hellerbarde joined #salt
11:11 CycloHex joined #salt
11:11 eagen joined #salt
11:11 masterkorp joined #salt
11:11 iclebyte joined #salt
11:11 eightyeight joined #salt
11:11 ampex joined #salt
11:11 lkannan joined #salt
11:11 funzo joined #salt
11:11 thunderbolt joined #salt
11:11 lkannan joined #salt
11:11 Nazca joined #salt
11:11 octarine joined #salt
11:11 thunderbolt joined #salt
11:11 octarine joined #salt
11:11 wavis joined #salt
11:11 CaptTofu joined #salt
11:11 wavis joined #salt
11:11 ecdhe joined #salt
11:11 steve1 joined #salt
11:11 jesusaurus joined #salt
11:12 fxdgear joined #salt
11:12 JPaul joined #salt
11:12 copelco joined #salt
11:12 Rockj joined #salt
11:12 xenoxaos joined #salt
11:12 berto- joined #salt
11:12 kwmiebach______ joined #salt
11:12 kalail_ joined #salt
11:12 basepi joined #salt
11:12 kaiyou joined #salt
11:12 SpX joined #salt
11:13 joeyparsons joined #salt
11:14 crane joined #salt
11:15 bhosmer_ joined #salt
11:16 dark_helmet joined #salt
11:16 skarn joined #salt
11:16 intellix joined #salt
11:16 Blacklite joined #salt
11:16 wr3nch joined #salt
11:17 hardwire joined #salt
11:18 fe92 joined #salt
11:18 fe92 joined #salt
11:18 dvestal joined #salt
11:21 NVX joined #salt
11:24 fejese babilen, just to confirm: it works fine again on .10
11:27 ndrei joined #salt
11:34 PI-Lloyd looks like someone else has already posted about the salt-cloud bug and requested it be backported, so one less thing for me to do :) - https://github.com/saltstack/salt/issues/16437
11:37 dccc_ joined #salt
11:39 ramishra joined #salt
11:41 hobakill joined #salt
11:42 jsm joined #salt
11:52 stephanbuys joined #salt
11:52 ramishra joined #salt
11:53 NV joined #salt
11:54 joehoyle joined #salt
11:55 bhosmer_ joined #salt
11:59 gmcwhistler joined #salt
12:02 ramishra_ joined #salt
12:06 gildegoma joined #salt
12:09 linjan_ joined #salt
12:11 ramishra joined #salt
12:24 scottpgallagher joined #salt
12:27 Damon joined #salt
12:27 babilen PI-Lloyd: Splendid. I hope that it will be done.
12:28 babilen Too intrusive changes in the 2014.1 branch really ...
12:29 stephanbuys hi all, is there a way to get saltstack 2014.7 using PPA on ubuntu?
12:34 cpowell joined #salt
12:35 DaveQB_ joined #salt
12:44 miqui joined #salt
12:49 jsm joined #salt
12:52 vejdmn joined #salt
12:53 mechanicalduck joined #salt
12:53 scottpga_ joined #salt
12:59 bhosmer joined #salt
13:01 mechanicalduck joined #salt
13:01 djstorm joined #salt
13:02 ozzzo joined #salt
13:03 nitti joined #salt
13:05 PI-Lloyd babilen: Agreed, still can't get my head around how 1.11 made it to release with so many bugs...
13:05 mitsuhiko anyone seen this?
13:05 mitsuhiko File "/Users/mitsuhiko/.virtualenvs/fireline-cloud/lib/python2.7/site-packages/salt/output/highstate.py", line 246, in _strip_clean
13:05 mitsuhiko if returns[tag]['result'] and not returns[tag]['changes']:
13:05 mitsuhiko TypeError: string indices must be integers, not str
13:05 kusams joined #salt
13:06 elfixit joined #salt
13:08 jsm joined #salt
13:09 mpanetta joined #salt
13:11 ramishra joined #salt
13:14 mpanetta joined #salt
13:15 mpanetta joined #salt
13:15 kingel joined #salt
13:16 mpanetta joined #salt
13:17 jsm joined #salt
13:18 mpanetta joined #salt
13:21 KennethWilke joined #salt
13:28 nicolasbadia Hi guys, is there a way to run a specific state on the master ?
13:29 KennethWilke if it's a minion you can do salt 'master' state,sls <state to run>
13:29 nicolasbadia Unfortunatly, it is not one
13:30 KennethWilke ah, you could try 'salt-call state.sls <state>'
13:30 KennethWilke but not sure if that requires the minion
13:30 to_json joined #salt
13:31 nicolasbadia I get that: [ERROR   ] This master address: 'salt' was previously resolvable but now fails to resolve! The previously resolved ip addr will continue to be used
13:31 KennethWilke i'm not aware of a way to do much with salt without a minion, i guess another option would be you could use the salt-ssh mechanisms to not use the minion
13:32 KennethWilke overall i just put a minion on the master and it's easy that way
13:32 ttr joined #salt
13:33 kingel joined #salt
13:33 nicolasbadia Interesting idea, I will try that (put a minion on the master)
13:33 nicolasbadia Thanks
13:34 BrendanGilmore joined #salt
13:35 higgs001 joined #salt
13:36 KennethWilke np
13:36 ttr hello all. anyone is having this exception on debian after upgrading to 2014.11 ? http://pastebin.com/my4H2WtB
13:37 jsm joined #salt
13:37 ttr i've found that installing python-apt is helping this but there is no depepndency in .deb so not sure if this is a bug or not ;/
13:41 kingel joined #salt
13:44 nicolasbadia KennethWilke: quick question, how do I register the minion on the master since they are the same host ?
13:46 KennethWilke the process is the same as with other minions, set the 'master' option in the minion config to point to your master, start it up, and authorize its' key via salt-key
13:46 ttr nicolasbadia: sam as other - try to connect minon to master (over localhost or fqdn) and then run salt-key -L to see if you have new key to accept
13:46 nicolasbadia Hum I must have miss something… Let me check
13:49 nicolasbadia Indeed I was stupid, I hadn’t set the master option… Thanks again you both, it works now :-)
13:51 wnkz__ joined #salt
13:52 UtahDave joined #salt
13:52 dude051 joined #salt
13:52 UtahDave left #salt
13:52 scoates joined #salt
13:52 donaldinho joined #salt
13:54 ttr ok - find reported bug already on mine issue :)
13:55 higgs001 joined #salt
13:56 ndrei joined #salt
14:00 linjan_ joined #salt
14:03 pdayton joined #salt
14:03 robinsmidsrod joined #salt
14:03 Ricardo joined #salt
14:03 anotherZero joined #salt
14:09 GrueMaster joined #salt
14:10 johtso Am I right in thinking that I need to jump through all the hoops of installing pygit2 if I want to use private ssh gitfs_remotes with 2014.7?
14:11 johtso or is there an easier way?
14:12 iggy johtso: should work with gitpython as well (that's what we use with 2014.1)
14:14 johtso iggy: but what about "These parameters only currently apply to the pygit2 gitfs provider" ? http://docs.saltstack.com/en/latest/ref/configuration/master.html#gitfs-authentication-options
14:15 iggy oh, right, we're using ssh keys
14:15 johtso me too
14:16 johtso the comment about the parameters is for the whole authentication config section
14:20 pipps joined #salt
14:20 ttrumm joined #salt
14:22 ttrumm_ joined #salt
14:23 pipps joined #salt
14:25 iggy johtso: yeah, we're using 2014.1, so we had to kick it old school and we're just using aliases in the ssh_config file
14:26 hasues joined #salt
14:26 ajolo joined #salt
14:27 felskrone joined #salt
14:28 rallytime joined #salt
14:28 johtso ugh, I'm really struggling getting my formulas working.. I'm loading them via gitfs_remotes, and I see the repos getting fetched in the master log output, but when running `pillar.items` it tells me that those states don't exist
14:29 iggy pillar != states
14:30 johtso ahhhhhhhhhhh
14:30 johtso doh!
14:30 johtso seeing SLS was getting me confused, the problem is obvious now
14:30 iggy state.show_top might help
14:30 SheetiS joined #salt
14:31 johtso and there was me going down a rabbit hole trying to compile pygit2 :P
14:31 iggy or if you're really desperate cp.list_master
14:31 pipps joined #salt
14:34 donaldinho I have a very simple setup with a masterless minion. I am trying to get pillar data to work but when I call salt-call pillar.data none of the data that I have in pillar_roots seems to be set. I've set file_roots and pillar_roots in the minion config. If I create errors in the yaml of my pillar files then I get errors when I call pillar.data so the files in my pillar_roots are getting read but I can't  see any of the data that I am trying to set. Does
14:34 donaldinho anyone have any ideas about why this data is not appearing?
14:34 donaldinho top.sls http://pastebin.com/FYfENAD9
14:35 donaldinho testpillar.sls http://pastebin.com/SeTPaAi7
14:36 iggy pillar.items ?
14:37 warthog_ joined #salt
14:39 eunuchsocket joined #salt
14:42 kingel joined #salt
14:43 stephanbuys joined #salt
14:43 johtso how can I debug an issue with my ext_pillar data? it tells me that an SLS that I expect to be available isn't.
14:43 micah_chatt joined #salt
14:44 johtso I can see the relevant SLSs in /var/cache/salt/master/pillar_gitfs
14:45 kingel joined #salt
14:45 jalbretsen joined #salt
14:46 johtso there shouldn't be a problem with using subdirectories in my pillar data should there?
14:46 iggy nope, just tartet it the same way as in the state top file
14:48 kaptk2 joined #salt
14:48 johtso I've no idea what's going on then..
14:49 dvestal joined #salt
14:49 che-arne joined #salt
14:49 mapu joined #salt
14:52 vejdmn joined #salt
14:53 schimmy joined #salt
14:57 hobakilllll joined #salt
14:58 schimmy joined #salt
14:58 johtso oh of course.. it's messed up because I've specified an non-master branch for the ext_pillar
14:58 johtso the linking of branch names with environments has caused nothing but trouble for me :/
14:58 prosper_ joined #salt
15:01 StDiluted joined #salt
15:04 giannello joined #salt
15:05 ericof joined #salt
15:11 CycloHex Anyone here ever used user.present(expire) param?? I'm getting an error
15:11 CycloHex https://github.com/saltstack/salt/issues/16464
15:13 primechuck joined #salt
15:13 conan_the_destro joined #salt
15:18 to_json1 joined #salt
15:19 sroegner joined #salt
15:23 CycloHex ahaaaaaa, shadow support is not yet implemented in 2014.1.10 :(
15:27 kingel joined #salt
15:30 johtso Has anyone else found that `sls: foo` requisites don't seem to work?
15:31 johtso tells me that the requisite wasn't found
15:31 thedodd joined #salt
15:31 cpowell @johtso did you include the state?
15:31 johtso yep
15:32 hobakill morning all. i'm looking for a way to get all my jids in a centralized location. has anyone done this? is there a way to even see those?
15:33 kingel joined #salt
15:34 nitti joined #salt
15:34 ipmb joined #salt
15:35 superted joined #salt
15:35 ckao joined #salt
15:36 tmh1999 joined #salt
15:38 schimmy1 joined #salt
15:40 teebes joined #salt
15:44 rlarkin joined #salt
15:49 eunuchsocket two part question: 1) how do i target more than once glob? 2) can I target nodes but not assign them any states (yet)?  see: https://www.refheap.com/
15:51 wendall911 joined #salt
15:52 hobakill eunuchsocket: http://docs.saltstack.com/en/latest/topics/targeting/compound.html
15:52 hobakill like that?
15:52 aquinas joined #salt
15:54 eunuchsocket hobakill: I'm still struggling.  I've tried 'rhd*,sd*', 'rhd* and sd*' seems like one of those would work
15:54 dvestal joined #salt
15:54 eunuchsocket hobakill: maybe I need 'L@rhd* or sd*'
15:54 eunuchsocket and*
15:55 eunuchsocket ooooh
15:55 eunuchsocket I see what i did there and/or confusion
15:56 tmh1999 joined #salt
15:56 kingel joined #salt
15:58 eunuchsocket why wouldn't 'rhd* or sd*' work?
15:59 cwright eunuchsocket: what is the resulting host set when you try 'rhd* or sd*' ?
15:59 cwright is it first match wins, ie, only the rhd* host?
16:00 eunuchsocket "No Top file or external nodes data matches found" regardless of order
16:01 antonw joined #salt
16:01 to_json joined #salt
16:02 eunuchsocket however, when I change it to just 'st*' it matches my node
16:03 fannet What would be the best way to keep a directory of files in sync via salt. Example: I have a folder full of static html files and if one of the files changes I want salt to push the new one (I do not want to define single files using the salt file state module)
16:06 cwright fannet: see file.recurse
16:10 saravanans joined #salt
16:10 debian112 joined #salt
16:12 pdayton joined #salt
16:12 pipps joined #salt
16:15 desposo joined #salt
16:17 clayk joined #salt
16:18 clayk joined #salt
16:20 clayk joined #salt
16:20 KyleG joined #salt
16:20 KyleG joined #salt
16:21 clayk Hello. New to Salt. I was curious about RHEL/CentOS 7 and the use of firewalld and Salt's ability to configure using firewalld.
16:22 iggy eunuchsocket: the default is simple globbing, you have to use a compound match for anything else (including using and/or)
16:22 Gareth clayk: afaik there isn't support for firewalld, that's not to say there couldn't be.  salt does support iptables though, managing a firewall using salt works quite well.
16:23 clayk Gareth: thank you. I know we could revert back to iptables but wasnt' sure if/what the plans may be regarding firewalld.
16:24 eunuchsocket iggy: turns out I needed some more coffee to see "- match: compound" in the state example.  Everything is working as expected now.
16:24 Gareth clayk: I'm not overly familiar with firewalld, if you want to file an issue on github...a feature request, then someone could take a look at possibly adding support for it.
16:25 Gareth bbl
16:26 forrest joined #salt
16:27 spookah joined #salt
16:28 kusams joined #salt
16:31 TheRealBill_here joined #salt
16:32 TheRealBill_here left #salt
16:32 baconbeckons joined #salt
16:33 ramishra joined #salt
16:36 saravanans joined #salt
16:37 desposo joined #salt
16:39 jsm joined #salt
16:41 saravanans joined #salt
16:41 bhosmer joined #salt
16:42 n8n joined #salt
16:43 saravana_ joined #salt
16:50 hobakill eunuchsocket: did you get things figured out? sorry i was eating tacos.
16:51 saravanans joined #salt
16:52 fannet joined #salt
16:53 fannet What would be the best way to keep a directory of files in sync via salt. Example: I have a folder full of static html files and if one of the files changes I want salt to push the new one (I do not want to define single files using the salt file state module)
16:53 fannet (sorry if someone answered this earlier my irc client blew up)
16:53 drybjed fannet: git with some hooks, most likely
16:54 fannet drybjef: I'd prefer to use the salt fileserver
16:55 fannet in other words its not a large enough number of files to really warrant their own git repo
16:56 drybjed what's the minimum amount? :-)
16:56 fannet the amount where my devs won't yell at me for having to add yet another repo ;)
16:57 drybjed fannet: do you use a git server of some kind, like gitlab for example?
16:57 fannet we use a hosted git
16:58 fannet (like github)
16:58 pipps joined #salt
17:04 beneggett joined #salt
17:10 aparsons joined #salt
17:16 JaredR joined #salt
17:17 aw110f joined #salt
17:17 JaredR I'd like a copy of the 2014.1.7 salt rpms for CentOS 6.
17:19 Ryan_Lane joined #salt
17:20 patarr joined #salt
17:20 patarr joined #salt
17:20 quanta_ joined #salt
17:23 wnkz joined #salt
17:26 UtahDave joined #salt
17:27 cpowell joined #salt
17:27 ksalman i am trying to deploy a salt state on windows after installatation, so I put the salt command as part of the windows answerfile. It appears that Windows adds a semicolon at the end of the commnd so the command becomes "C:\salt\salt-call.exe state.sls packages.windows 1>salt.log 2>&1;" (notice the semilcolon at the end)
17:27 ksalman and i get this error https://gist.github.com/anonymous/e0b4f99e41b4647c4bd6
17:29 ksalman This is probably not a salt issue? though any idea on how to get past this?
17:30 kingel joined #salt
17:32 perfectsine joined #salt
17:32 forrest ksalman, did you already try to wrap it in quotes?
17:32 forrest I'm not super familiar with answerfiles...
17:33 ksalman forrest: I did and it still tries to parse the semicolon for example
17:33 ksalman "c:\salt\salt-call.exe state.sls packages.windows" 1>salt.log 2>&1;
17:34 forrest ksalman, I'm not sure then, did you already check stackoverflow/
17:34 ksalman i tried that exact command on the cmd shell and i get the key error
17:34 ksalman no, i did not check there
17:35 ksalman I suppose i should
17:36 pipps joined #salt
17:36 ksalman actually the command that i tried was:
17:36 ksalman cmd /c "c:\salt\salt-call.exe state.sls packages.windows" 1>salt.log 2>&1;
17:37 ksalman without the cmd /c it thinks the whole quoted string is a command and errors out
17:38 possibilities joined #salt
17:39 druonysus joined #salt
17:39 possibilities joined #salt
17:39 skyler joined #salt
17:42 roolo joined #salt
17:45 murrdoc joined #salt
17:50 pipps joined #salt
17:51 pssblts joined #salt
17:51 pipps_ joined #salt
17:52 skyler joined #salt
17:53 aparsons_ joined #salt
17:56 kingel joined #salt
17:56 possibilities joined #salt
18:00 prosper_ joined #salt
18:03 TheRealBill_here joined #salt
18:04 quickdry21 joined #salt
18:04 Ozack1 joined #salt
18:06 vejdmn joined #salt
18:07 pipps joined #salt
18:09 giannello joined #salt
18:10 thedodd joined #salt
18:23 agend joined #salt
18:28 jalaziz joined #salt
18:33 saravanans joined #salt
18:38 pipps joined #salt
18:39 pipps_ joined #salt
18:41 ajprog_laptop joined #salt
18:46 pipps joined #salt
18:48 cpowell joined #salt
18:49 JaredR Does anyone have a copy of the 2014.1.7 salt RPMs for CentOS 6?
18:50 mapu joined #salt
18:52 forrest JaredR, no but I think you can get them on EPEL, do you have twitter?
18:52 forrest the guy who knows the link isn't in IRC right now
18:52 forrest but I can tweet at him and ask
18:54 chiui joined #salt
18:56 Heartsbane joined #salt
18:57 JaredR @forrest, I can get on twitter.
18:57 forrest JaredR, http://koji.fedoraproject.org/koji/packageinfo?packageID=13129
18:57 forrest JaredR, nah he already responded
18:57 JaredR Sweet! Thanks, forrest!
18:58 forrest yeah np
18:59 beneggett joined #salt
19:00 teepark joined #salt
19:04 pipps joined #salt
19:06 giannello joined #salt
19:07 kingel joined #salt
19:07 thedodd joined #salt
19:09 jalaziz joined #salt
19:11 eliasp forrest: ping … I got some questions for the formula king ;)
19:11 dude051 joined #salt
19:14 dude051 joined #salt
19:14 eliasp I'm working on an NFS server formula… when it comes to the exports lines, I want to pull them from a generic pillar (lookup:nfs:exports or so), but besides that, I want to be able to override the "exports" pillar when including the formula with data based on the SLS from where I include it…
19:14 eliasp … is there a proper way to accomplish this?
19:17 perfectsine joined #salt
19:19 possibilities joined #salt
19:19 res0nat0r joined #salt
19:19 brandon__ joined #salt
19:19 jalaziz joined #salt
19:21 forrest eliasp, so you're setting exports in the pillar, and you want to overwrite that from somewhere else?
19:26 murrdoc joined #salt
19:27 jalaziz joined #salt
19:27 forrest eliasp, I can't think of a way to handle that in pillar, other than if you do some sort of check to either see if the pillar data exists, and 'prioritize' based on that.
19:28 forrest or set a variable in those other states, which you include and then use (again with a check to see if it exists)
19:28 forrest maybe whiteinge has a better idea for that
19:29 eliasp hmm, "prioritizing" might work… I'll give it a try… if it works it'll be hopefully soon a PR for saltstack-formulas
19:31 digismack joined #salt
19:32 to_json joined #salt
19:32 forrest eliasp, that would be great!
19:33 ttrumm joined #salt
19:34 pipps joined #salt
19:36 utahcon when did/do onlyif and unless become global?
19:36 eliasp utahcon: 2014.7 → http://docs.saltstack.com/en/latest/topics/releases/2014.7.0.html#global-onlyif-and-unless
19:36 utahcon blergh
19:36 utahcon thanks! eliasp
19:40 pipps joined #salt
19:41 digismack Salt doesn't seem to be picking up my pillar data anymore. Here's the gist of the error and examples. Anyone mind checking this out to see if they see an issue? https://gist.github.com/digismack/c14258b353511eefe544
19:41 eliasp hmm, there's no way to pass a context to an included file… right?
19:41 to_json joined #salt
19:42 eliasp digismack: what do tests like "salt your-minion pillar.get foo" or "salt your-minion pillar.items" say?
19:43 obimod joined #salt
19:43 digismack bash: salt: command not found
19:43 digismack :/
19:43 eliasp digismack: ? you should execute this on the master, where you also do things like "salt your-minion state.highstate test=True" etc.…
19:44 eliasp or how do you test states? your error message in the gist must come from somewhere…
19:44 digismack running masterless
19:44 digismack using Salt to configure a docker container
19:44 eliasp aah, ok… no idea how to deal with pillars in masterless environments, sorry
19:44 digismack it was working fine until I went to re-build my container
19:44 digismack and now all of a sudden it just doesn't work anymore
19:46 ttrumm_ joined #salt
19:46 eliasp digismack: ah… use "salt['pillar.get']('django_settings').iteritems()" instead of "pillar['django_settings'].iteritems()" … I'm pretty sure it'll work then
19:47 digismack tried that as well
19:47 eliasp hmm, ok ;(
19:48 digismack if I try that, I end up with the error: Rendering SLS "base:django-settings" failed: Jinja variable 'str object' has no attribute 'iteritems'; line 1
19:50 eliasp when executing this masterless, try to get a listing of all pillars visible for this minion using "salt-call --local pillar.get django_settings" … does this return anything?
19:52 digismack it returns "local:"
19:52 eliasp hmm ;(
19:52 eliasp sorry, no idea then… anyone else?
19:53 digismack it works on 2014.1.10
19:53 digismack but on 2014.1.11 it doesn't work
19:53 eliasp oh… this doesn't sound good
19:54 digismack yeah :/
20:01 digismack https://github.com/saltstack/salt/issues/16478
20:01 digismack looks like this could be my issue
20:02 digismack back to 2014.1.10 for now I guess
20:03 kermit joined #salt
20:03 TheThing|laptop joined #salt
20:03 murrdoc joined #salt
20:03 ksalman forrest: I got around the windows semicolon problem by putting the salt command in a bat file and running that instead of running the salt command
20:03 forrest ksalman, *facepalm*
20:03 ksalman haha
20:04 forrest that's so stupid you have to do that, lol
20:05 dude051 joined #salt
20:06 Will joined #salt
20:08 Will I am having trouble getting halite to work. Does anyone have any experiece with it?
20:08 pipps joined #salt
20:10 iggy it was fairly simple to get running following the directions on the github page
20:11 Will i have followed those but it does not seem to be working. yum failed to find the package so i used pip to install.
20:11 Will is there a way to verify if it is running?
20:11 saravanans joined #salt
20:17 beneggett joined #salt
20:17 giannello joined #salt
20:18 iggy sorry, I followed the "install to develop" instructions (or whatever they are called)
20:19 Will ok thank you
20:20 deeprave joined #salt
20:21 aparsons joined #salt
20:21 ttrumm joined #salt
20:21 Ahlee has halite stabalized at all?  or does it still "require" develop?
20:22 Ahlee I gave up waiting and wrote my own halite-like site consuming the rest api
20:22 mitsuhiko is salt-ssh currently completely broken?
20:22 mitsuhiko (master)
20:22 mitsuhiko i mean develop
20:23 londo joined #salt
20:24 sag47 joined #salt
20:25 DenkBrettl joined #salt
20:25 Ryan_Lane joined #salt
20:27 mitsuhiko for me it looks like salt-ssh goes into a deadlock
20:27 mitsuhiko dtruss says it's hanging on a sem_trywait
20:28 n8n joined #salt
20:28 thedodd joined #salt
20:29 druonysuse joined #salt
20:29 druonysuse joined #salt
20:31 whiteinge mitsuhiko: there is a bug with salt-ssh if you're using pillar. it's currently a blocker for the upcoming release
20:31 mitsuhiko whiteinge: i am not using pillar
20:31 mitsuhiko added a bug report: https://github.com/saltstack/salt/issues/16487
20:31 whiteinge ty. i will take a look
20:32 mitsuhiko i assume it tries to talk to the ssh executable somehow
20:32 pipps joined #salt
20:32 mitsuhiko why would it even do that. there is no password prompt on this ssh connection
20:33 mitsuhiko i think this logic is broken to be honest. i have no idea how that loop would break
20:35 SheetiS joined #salt
20:35 pipps_ joined #salt
20:36 ipmb joined #salt
20:36 debian112 Is there a way to tell which version of minions are using the salt-master?
20:36 whiteinge info in that ticket is helpful, thanks.
20:37 whiteinge !salt.modules.test.versions_report | debian112
20:37 whiteinge er
20:37 whiteinge !modules.test.versions_report | debian112
20:37 wm-bot4 debian112: http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.test.html#salt.modules.test.versions_report
20:37 whiteinge there we go :)
20:37 mitsuhiko whiteinge: btw. the reason why i even looked at develop is this bug: https://github.com/saltstack/salt/issues/16488
20:37 yetAnotherZero joined #salt
20:38 debian112 thanks
20:38 mikeywaites joined #salt
20:39 nkuttler hrm.. i guess i should switch to formulas as well for some pillars..
20:39 beneggett joined #salt
20:40 mikeywaites Hey! Really hoping someone can help.  Im using salt with vagrant. everything was fine.  After destroying the box and switching to a ubuntu 14.04 vm my vm seems to longer match the 'dev' pillar data like it did before
20:40 mikeywaites Not sure where i should start trying to debug this
20:41 mikeywaites i can see from the log output that its trying to find pillar data from the base env and it should be matching dev IE rendering SLS "base:postgresql.databases" failed: Jinja variable 'dict object' has no attribute 'postgresql'; line 1
20:41 whiteinge mitsuhiko: noted. i'll try running that state in my tests
20:42 mitsuhiko why does salt do such weird things with ssh instead of using the library that fabric is based on for ssh communication?
20:42 mitsuhiko seems much more reliable than doing it again
20:42 ashirogl joined #salt
20:43 defmikekoh joined #salt
20:43 murrdoc what u mean
20:43 mitsuhiko i'm just going through the stuff in ssh/shell.py and that makes very little sense to me
20:43 whiteinge i'm not familiar with that history (been meaning to ask about it). i do know that parmiko was evaluted
20:44 murrdoc paramiko for the win!
20:44 iggy doesn't fabric use paramiko?
20:45 cpowell @murrdoc seconded
20:45 ashirogl left #salt
20:45 whiteinge the fabric guys maintain paramiko nowadays (afaik)
20:45 johtso iggy: yes it does
20:45 mitsuhiko they do
20:46 iggy honestly, as many people as I hear having issues with salt-ssh in here every day, I'm surprised anybody even bothers with it
20:46 jalaziz joined #salt
20:47 mitsuhiko i'm pretty much done with salt-ssh now
20:47 eliasp well, it's not considered to be production ready before 2014.7
20:47 eliasp see also: http://docs.saltstack.com/en/latest/topics/ssh/
20:47 iggy that also doesn't take into account the limitations it has compared to standard salt transport
20:47 mitsuhiko eliasp: it's broken for me on the last release and it got worse on master
20:48 eliasp ok, … haven't played with it for a long time
20:48 johtso I just want salt-ssh to work so that I can setup salt minion's and masters using salt
20:48 kermit joined #salt
20:50 murrdoc are u working in cloud ?
20:50 murrdoc or with hardware
20:50 iggy I just have everything managed via the salt formula... new minions connect to the master and pull their config from there
20:50 murrdoc for the former salt-cloud sets up minion/master easy
20:50 murrdoc and like iggy said for hardware
20:50 mikeywaites anyone got any suggestions on a way for me to figure out why salt is no longer using the correct pillar data for this node?
20:51 iggy mikeywaites: start pasting stuff until someone sees the problem
20:51 iggy obviously in some sort of pastebin, not in here
20:52 mikeywaites Sure, not sure what to paste really the states are being matched correctly - the only thing thats changed is the version of ubuntu?
20:52 mikeywaites 12.04 > 14.04
20:52 mikeywaites is there a newer version of salt in use there  that might yield any clues?
20:53 dstokes are there any best practices for storing data / variables that are reused throughout the pillar? like an included sls that just defines jinja vars?
20:53 mitsuhiko do you run salt-cloud locally? or from the master
20:54 higgs001 joined #salt
20:54 Will joined #salt
20:54 Gareth 2/w 40
20:54 Gareth erm
20:54 Ryan_Lane joined #salt
20:54 johtso dstokes: I was going down that rabbithole the other day
20:55 dstokes ;_;
20:55 murrdoc dstokes:  you acn use a common.map.jinja
20:55 eliasp dstokes: you could try using YAML anchors/references… but there are many issues open regarding what you want to do…
20:55 johtso I found that none of the solutions are very satisfactory
20:55 murrdoc that u can then include everywhere
20:55 beneggett joined #salt
20:55 johtso but you can load variables from a yaml file
20:55 eliasp ^^ what murrdoc says…
20:55 SheetiS dstokes: have an example of what you are trying to accomplish?  You can definitely do {% from 'variables.sls' import variable with context %} inside your pillar(s).  or have a big dict that you could import similarly that had all the data you wanted to put everywhere
20:56 SheetiS what murrdoc said is probably a good convention for it.
20:56 dstokes i have several minion specific pillar files that define variables and that totally works for provisioning those minions. now i'm setting up a build server that needs access to deployment information (env variables) for _all_ those minions. trying to stay dry ;0
20:57 clayk joined #salt
20:57 johtso yaml syntax gets really fun when you have variables that contain special characters or are multi-line..
20:57 dstokes murrdoc: you lost me. acn?
20:57 dstokes oh.. s/acn/can/
20:57 heewa joined #salt
20:57 dstokes are you referring to something like https://github.com/saltstack-formulas/apache-formula/blob/master/apache/map.jinja?
20:58 luminous if you want to get state/formula return information in near-real time on a client (like you do the shell/console), what returner setup is worth considering for the client?
20:59 heewa Anyone have advice on hunting down salt saying files aren’t there with a file.managed: - source: salt://path situation?
20:59 Ryan_Lane joined #salt
20:59 heewa Also, I don’t know why, but it’s putting a pipe (|) in there, like I say: source: salt://yay  and it says salt://|yay not found.
20:59 eliasp heewa: see, if they're shown in "salt your-minion cp.list_master"
21:00 pipps joined #salt
21:00 luminous heewa: either --log-level=debug or do a cmd.run state with cat {{ filepath }} to confirm you have what you expect
21:01 beneggett joined #salt
21:01 yetAnotherZero joined #salt
21:03 heewa eliasp: Checked cp.list_master, they’re listed.  luminous: debug logs show nothing I can detect as wrong, and the files aren’t being placed on the minion (cuz it says it can’t find them in salt’s file server).
21:03 eliasp heewa: everything in base environment? did you possibly assign another environment in your top.sls?
21:04 murrdoc dstokes:  sorry s/acn/can
21:04 heewa eliasp: Just a base env. Any clue what the pipe char is all about? Or where in the source it’s being placed there?
21:04 murrdoc and yes the map.jinja u referenced
21:04 murrdoc assuming u have a top level shared state
21:04 dstokes murrdoc: awsm, thx
21:04 murrdoc that would be where you could put the map
21:05 eliasp heewa: no idea about the pipe…
21:05 SheetiS it may be that the salt-mine might be applicable since it is information about multiple minions, but it would need to be tied to a salt function of some type
21:08 sag47 joined #salt
21:08 pipps joined #salt
21:08 sag47 left #salt
21:11 mikeywaites Is there a way i can see what a master minion is thinks its id is using salt-call?
21:13 heewa mikeywaites: Sorry, I couldn’t parse that question. Are you trying to find out what a minion thinks its own id is using salt-call? salt-call grains.get id
21:15 sarlalian joined #salt
21:15 linjan_ joined #salt
21:16 mikeywaites heewa:  thanks thats exactly what i was after
21:16 viq or cat /etc/salt/minion_id
21:16 snave joined #salt
21:17 mikeywaites So  everything seem to be in the right state
21:17 mikeywaites but the pillar data for dev just doesnt get picked up for this node
21:17 mikeywaites http://pastebin.com/thEWpvcN
21:18 ipmb_ joined #salt
21:18 mikeywaites salt-call pillar.raw doesnt seem to pick up the dev items at all so it seems like the node is not being matched
21:19 aparsons joined #salt
21:20 bhosmer joined #salt
21:20 teebes joined #salt
21:23 n8n joined #salt
21:24 kermit joined #salt
21:26 perfectsine joined #salt
21:28 mikeywaites if the minion show_top matches only base (there is no dev) could this cause the pillar data to not match the dev env?
21:29 fannet guys I installed the salt minion using apt-get on an existing server (all my other servers are deployed via bootstrap) and running highstate results in this:  No Top file or external nodes data matches found
21:29 fannet however my base has a '*' match
21:29 rap424 joined #salt
21:30 fannet the minion is def talking to the right master and -l debug shows Fetching file from saltenv 'None', ** attempting ** 'salt://top.sls'  which is also troublesome
21:31 kermit joined #salt
21:32 aparsons joined #salt
21:32 mosen joined #salt
21:32 possibilities joined #salt
21:34 cads joined #salt
21:35 steve1 joined #salt
21:37 nitti_ joined #salt
21:39 defmikekoh joined #salt
21:40 mailo joined #salt
21:42 murrdoc pastebin top.sls ?
21:44 murrdoc also new minions' id
21:46 mikeywaites I have reached the point where the only thin left to do is drink heavily :D
21:46 heewa eliasp: Worked around it by putting that file root in a different environment and doing source: salt://path?saltenv=that_env
21:47 heewa mikeywaites: Ugh, me too. It’s 5:45 where I am, in two hours, I’m actually going to take my laptop to a bar and continue messing with salt, but with the comfort of beer.
21:48 mikeywaites Even moving the pillar data to the base env still hasnt solved my issue
21:48 mikeywaites i just cant figure out why pillar doe not contain anything
21:48 kballou joined #salt
21:49 heewa mikeywaites: I don’t know if it’s your issue, but I had that issue once. Went away when I upgraded to the v2014.7.0rc3 version (git install).
21:49 mikeywaites Im back at the last known state where a colleague rebuilt the vm with no issues just last week.:(
21:49 viq mikeywaites: state top.sls and pillar top.sls are independent
21:50 mikeywaites viq:Ok well thats another wildcard off the list
21:50 viq Do you have dev environment defined for you pillars in master config?
21:50 mikeywaites viq yes (masterless minion)
21:51 mikeywaites i can see that salt is loading them, if i whack a syntax error in ill correctly see that salt fails to render the yaml
21:51 TheThing|laptop joined #salt
21:51 mikeywaites sudo salt-call --version
21:51 mikeywaites salt-call 2014.1.11 (Hydrogen)
21:51 viq could you show that part of master config, and all your of pillar top.sls files?
21:51 mikeywaites viq:  yes one moment
21:52 tie joined #salt
21:53 babilen mikeywaites: Are you on masterless minion?
21:53 mikeywaites babilen:  Yeh, this is a Vagrant vm
21:54 pipps joined #salt
21:54 babilen Yeah, pillars on masterless minions no longer works on .11 (a very unfortunate regression from .10)
21:54 mikeywaites babilen:  OMG you are my hero
21:54 babilen upstream plans to release a .12 with the fix very soon
21:54 goodwill whiteinge: thats a hell of a regression: pillars on masterless minions no longer works on .11 (a very unfortunate regression from .10)
21:54 tie I was just logging in to ask the same question...
21:54 mikeywaites can i upgrade to a working version?
21:54 tie works yesterday, broken today...
21:55 viq mikeywaites: more like downgrade, unless you plan to run release candidate
21:55 mikeywaites or even downgrade babilen
21:55 babilen I tried to get them to not release that version without that fix (it would have been no problem to patch it in the packages directly even you if you don't want to file a bug about it)
21:55 mikeywaites oh .11 is latest i see
21:55 babilen No, don't downgrade me.
21:55 scoates joined #salt
21:55 viq :D
21:55 dude051 joined #salt
21:55 mikeywaites hahaha
21:55 viq babilen: wouldn't you like to be a teenager again? ;)
21:55 mikeywaites id like it to be 8am again and for this to have never of happend
21:56 tie is there a pulished patch that we can apply to get that working again?
21:56 babilen Let me get you the bug reports in question.  #16416, #16210,
21:56 babilen #16212
21:56 mikeywaites TLDR version? is there a simple way for me to downgrade, that said if this is likely to be fixed by the time i sober up that will also do :D
21:57 muss joined #salt
21:58 babilen https://github.com/saltstack/salt/issues/16210 is the report and https://github.com/saltstack/salt/pull/16212 the patch
21:59 babilen Downgrade(s) will be tricky if you no longer have the salt-{minion,common} packages (e.g. in /var/cache/apt/archives) as they are no longer on the mirrors.
21:59 bhosmer joined #salt
22:00 babilen http://docs.saltstack.com/en/latest/topics/tutorials/salt_bootstrap.html supports git and tags though
22:00 babilen (you can use that from within vagrant)
22:00 babilen But then, I typically just install master and minion on one node in vagrant and play with a "real" master/minion setup.
22:01 babilen So, that would be the other approach. A bunch of tests and provisioners rely on masterless minions though
22:02 mikeywaites babilen:  Thanks so much for answering this!!  One last question.  I dont suppose you know when the fix will land?
22:02 aurynn joined #salt
22:02 steve1 joined #salt
22:03 tie babilen, you know if bootstrapping from the daily build solves this problem?
22:03 babilen No, unfortunately not. My impression was "very soon™" but, frankly, I am surprised that it didn't hit the mirrors today already.
22:03 murrdoc joined #salt
22:04 mikeywaites yeh!!! Ive lost about 15 years due to this bug :D
22:04 babilen tie: I don't typically use the bootstrapping script, but the 2014.1 branch definitely has the commit that fixes this (and other that fix different problems)
22:04 eliasp mikeywaites: oh, please downgrade me 15 years as well…
22:05 mikeywaites curl -L https://bootstrap.saltstack.com | sudo sh -s -- git {what version here} ?
22:06 cwyse joined #salt
22:07 baconbeckons joined #salt
22:08 obimod joined #salt
22:08 pdayton joined #salt
22:08 babilen 2014.1.10 comes to mind
22:09 mikeywaites yeh i tried that but thats not a valid branch in git it would seem and im guessing 2014.1 will just install the broken one.
22:09 dccc_ joined #salt
22:09 mikeywaites im gonna install develop and see if that works
22:09 tie trying with "daily" atm
22:10 murrdoc does it only use branches ?
22:10 murrdoc i am sure tags work too
22:10 babilen mikeywaites: No, 2014.1 would be fine. The HEAD of that branch is fine.
22:10 mikeywaites Ok
22:10 mikeywaites curl -L https://bootstrap.saltstack.com/ | sudo sh -s -- git develop
22:10 babilen But I thought that tags would work. I'd probably use HEAD if it takes branches
22:10 mikeywaites is currently RUNNING
22:10 babilen No, no develop
22:11 babilen (that'll get you a completely different, not yet released and unstable version)
22:11 kermit joined #salt
22:11 mikeywaites which actually works over the "tested" "stable" "released" version :D
22:12 babilen v2014.1.10 is the tag (if it doesn't accept "2014.1" as branch name
22:12 tie that's a release fail in my book
22:12 babilen it is
22:13 mikeywaites Well as long as i can get my vm back and running for tomorrow then im OK.  we are all only human after all :(
22:13 eliasp as much as I love Salt, that much do I think does it has to improve on release testing… test coverage… preventing regressions etc.
22:13 murrdoc https://github.com/saltstack/salt-bootstrap/blob/develop/bootstrap-salt.sh#L1123
22:13 murrdoc that says tags, branches or a git checkout hash will work
22:14 babilen It simply needs two/three repositories (much like Debian) with automatic migration from one to the next if no grave bugs have been found.
22:14 mikeywaites Thanks very much for you help everyone.  its been emotional.
22:14 eliasp yeah… or like that
22:14 tie just tested with git HEAD
22:14 murrdoc mikeywaites:  is that a walter ppk ?
22:14 tie seems fine on the pillar front
22:14 mikeywaites murrdoc:  i dont think it would do enough damage
22:15 mikeywaites im going go out slow and painful.
22:15 mikeywaites :D
22:15 murrdoc mine says replica
22:15 murrdoc i am referencing snatch btw
22:15 murrdoc wait am i
22:15 murrdoc shit forget about it
22:15 murrdoc :d
22:15 babilen tie: Yeah, I walked through that issue with the first person in here that ran into it. The bug report and PR were filed before .11 hit the mirrors and there were mails, but ...
22:15 TheThing|laptop murrdoc: You're a computer software who thinks he has self awareness
22:15 murrdoc my bad
22:16 TheThing|laptop crap, read that wrong
22:16 TheThing|laptop I read your "wait am i" as "who am i"
22:16 TheThing|laptop made the convo a little bit more interesting :b
22:17 murrdoc haha
22:17 TheThing|laptop or more specificly
22:17 TheThing|laptop "what am i"
22:17 tie ok, I ended up with messed up apt repos
22:17 tie using the bootstrap with "git HEAD" does solve the pillar problem though
22:18 tie gonna give it a new try tomorrow
22:20 pipps joined #salt
22:25 thedodd joined #salt
22:25 Corey http://docs.saltstack.com/en/latest/ref/states/top.html <-- In reading through this, I see how you define different environments for dev, qa, and prod-- but when a minion connects, how does Salt determine which environment to drop it into?
22:25 steve1 joined #salt
22:25 mauro joined #salt
22:25 utahcon grains.filter_by, exactly how does the merge work? does it override per key? is the merge done post filter?
22:26 eliasp Corey: theoretically, a minion can be in more than one environment… that just depends on your top.sls assignments
22:26 Corey eliasp: I'm retrofitting a horrible environment; basically what I want to do is pattern matching. "If the minion name matches the form *.dev.domain.com, put it into the dev tree."
22:27 iggy Corey: you can confine a minion to a specific environment in the minion config file
22:28 eliasp Corey: well, then you just need in your top.sls:
22:28 eliasp dev:
22:28 eliasp '*.dev.domain.com':
22:28 eliasp - some.state
22:28 eliasp - another.state
22:28 Corey eliasp: Ah, in the base top file, or the dev specific top file? i.e. /srv/salt/top.sls or /srv/salt/dev/top.sls?
22:28 utahcon Corey: you'll also want to make sure that globs in other env don't match
22:29 Corey utahcon: Yeah, that part is semi-obvious. :-)
22:29 utahcon ;)
22:29 eliasp Corey: top files from different environments will be merged… so it's recommended to just have a single top.sls for all environments
22:30 eliasp … at least this makes it way easier… you might have use-cases, where you want per-env top.sls files, but I wouldn't recommend it to begin with
22:34 utahcon http://pastie.org/9632583 will using merge solve this?
22:38 intellix joined #salt
22:42 aquinas_ joined #salt
22:49 aquinas_ joined #salt
22:52 aparsons joined #salt
22:52 fannet is there a grain that will tell me the current environment of the host
22:54 aparsons_ joined #salt
22:56 iggy not that I've evern seen
22:58 jalaziz joined #salt
23:03 UtahDave joined #salt
23:04 fannet :( any idea how I might use the git module to check out different branches based on the environment that the host is in
23:05 _1_zander joined #salt
23:06 _1_zander poopp
23:07 nickjj is there a preference for using listen vs listen_in when either or could work?
23:07 Ryan_Lane depends on how you want to reference things
23:07 Ryan_Lane sometimes it's easier to use one or the other
23:08 nickjj if you expect multiple things to trigger a restart then listen on the service is probably the best option since it's less verbose/all in 1 spot?
23:08 possibilities joined #salt
23:09 UtahDave nickjj: People often find themselves using listen_in and watch_in because it allows each "state" to decide if a service should watch it or not.  In dynamic situations listen_in is usually cleaner
23:09 snave left #salt
23:09 nickjj ok, so it would be reasonable to use listen_in unless you have a good reason not to?
23:10 pdayton joined #salt
23:10 nickjj so far i definitely think listen_in seems more natural to me
23:11 murrdoc joined #salt
23:12 UtahDave they're both useful, but definitely use the right tool for the situation you're in
23:13 nickjj right now i'm towards the end of day 1 with using salt heh -- atm i just wanted to restart a service when a single config file changed -- it seemed more natural to me to put a listen_in on the service
23:14 nickjj and seems especially useful if you wanted to list out say 4 files to trigger a restart instead of having 4 listens attached to 4 different file states
23:14 aquinas_ joined #salt
23:14 utahcon so apparently the grains.filter_by merge is authoritative
23:17 joehoyle joined #salt
23:17 iggy fannet: for states or pillars?
23:18 jalaziz joined #salt
23:19 eliasp forrest: I managed to get my formula working  incl. passing SLS specific data into the formula which are then being merged together with the data from map.jinja
23:19 forrest eliasp, awesome, is there already an nfs formula repo
23:19 eliasp forrest: will have to fine-tune some stuff on Friday…
23:19 eliasp forrest: no… nothing NFS related yet
23:19 forrest eliasp, okay, looks like there is not, alright, are you cool with nfs-formula?
23:19 eliasp forrest: I found some other NFS formula attempts, but all relatively incomplete
23:20 eliasp sure, nfs-formula sounds good…
23:20 eliasp will have to go home now (1:20 AM), but will continue on Friday
23:20 nickjj hey Ryan_Lane, when trolling around the saltstack-formulas repo, would you recommend to follow the "ng" style of formulas?
23:21 Ryan_Lane ng?
23:21 nickjj example: https://github.com/saltstack-formulas/nginx-formula
23:21 Outlander joined #salt
23:21 nickjj there seems to be a "ng" version which is in its own folder inside of the nginx folder
23:22 nickjj instead of having a templates folder with jinja file extensions, there are just conf files sitting in a files/ folder
23:22 iggy there's nothing special about it, it's just different than the original one
23:23 nickjj is there anything wrong with having a foo.conf.jinja extension to get proper highlighting in editors?
23:23 forrest eliasp, sounds good, here's the repo: https://github.com/saltstack-formulas/nfs-formula
23:23 iggy nickjj: that stuff is all just the style the original author did... nothing is particularly wrong either way
23:23 nickjj iggy: ah
23:24 iggy the formulas are all written by different people, everybody's style/taste is a little different
23:24 nickjj iggy: do you have formulas in mind that might be decent to follow as a style guide?
23:24 iggy (just in case you didn't already figure that out
23:24 iggy )
23:24 nickjj yeah it seems to be a free for all
23:25 n8n joined #salt
23:25 iggy salt is very pragmatic... "here's 5 different ways to just about everything, use whatever works best for you"
23:25 nickjj there haven't been any patterns/styles that gained traction as semi-standard?
23:26 mosen there's an example formula, but yeah i sometimes wonder why different formulas are written the way they are
23:26 fannet ya we just re-wrote our own nginx formula for example to be more appropriate and it took about a day
23:27 murrdoc joined #salt
23:27 fannet does anyone know if the git module is capable of syncing just a specific git folder instead of an entire repo?
23:28 iggy there's all different levels of complexity in there... something that you need for say... mongo... you don't necessarily need for something like collectd
23:28 iggy fannet: yes... in some versions and with various restrictions
23:28 iggy oh wait, git module, not gitfs
23:29 iggy dunno
23:30 UtahDave fannet: I don't think git itself can do that yet.   gitfs allows you to specify a root folder, though
23:31 nickjj iggy: that is true, i really do like the openness so far
23:32 fannet so what is really the best way to sync files (say a folder of static html files) onto the minion ?
23:35 fannet seems like git is an OK option but the salt gitfs would be well suited
23:36 murrdoc anyone here provision openstack nodes
23:36 murrdoc with salt
23:36 murrdoc i d like to see the state if possible
23:36 murrdoc i mean like openstack cluster, not the vms
23:37 UtahDave fannet: you might look at file.recurse as well
23:37 UtahDave murrdoc: this is a great repo to start with:    https://github.com/CSSCorp/openstack-automation/tree/icehouse-yaml
23:37 murrdoc thank u UtahDave
23:37 pipps_ joined #salt
23:37 murrdoc heard the story of how u became utah dave
23:38 murrdoc le dope
23:38 UtahDave ? who spilled the beans?
23:38 murrdoc no clue
23:38 fannet @UtahDave - Ah I was under the assumption that only recursed file permissions
23:38 murrdoc :)
23:38 UtahDave what was the story you heard?
23:38 murrdoc there were two daves
23:39 murrdoc and u didnt get the hometown love
23:39 UtahDave fannet: nope.  file.recurse pushes out the entire directory structure.
23:39 murrdoc at your previous gig
23:39 murrdoc or something along those lines
23:39 murrdoc is it close to the truth
23:40 murrdoc thanks for the repo
23:40 UtahDave I was doing some consulting for the DEA in Washington State years and years ago.  There were several DEA agents named Dave there. So when I came in the head agent was tired of everyone just being "Dave"
23:40 UtahDave So he dubbed me "UtahDave"  and it stuck
23:40 fannet @UtahDave: Does that mean it gets pushed every time highstate is run or only the changes get pushed
23:40 UtahDave fannet: just the changes.
23:40 murrdoc UtahDave:  yup thats what i heard
23:40 UtahDave fannet: although be aware that it pushes the entire changed file, not like rsync
23:40 murrdoc i was staying vague cos dea
23:41 murrdoc still a fun story
23:41 fannet sorry I mean do all the files get overwritten with each highstate or just the DIFF
23:41 UtahDave fannet: No, just changed files get overwritten
23:41 fannet got it
23:41 fannet sorry my irc client is laggy lol
23:42 UtahDave no problem
23:42 snuffeluffegus joined #salt
23:43 quanta_ joined #salt
23:45 joehoyle joined #salt
23:45 kedo39 joined #salt
23:46 jalaziz joined #salt
23:54 kusams joined #salt
23:55 to_json joined #salt

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