Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-10-10

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

All times shown according to UTC.

Time Nick Message
00:00 dalexander joined #salt
00:01 linjan joined #salt
00:03 TyrfingMjolnir joined #salt
00:05 mpoole joined #salt
00:06 jalaziz_ joined #salt
00:08 paha joined #salt
00:08 rawkode joined #salt
00:18 possibilities joined #salt
00:19 blast_hardcheese joined #salt
00:19 druonysus joined #salt
00:24 mapu joined #salt
00:27 kusams joined #salt
00:30 loz--_ joined #salt
00:31 n8n joined #salt
00:35 pipps joined #salt
00:40 druonysus joined #salt
00:40 druonysus joined #salt
00:42 skyler Gotta love the responsiveness of the SaltStack team! Looks like they wont be going backwards incompatible: https://github.com/saltstack/salt/issues/16530
00:43 bhosmer joined #salt
00:44 ramishra joined #salt
00:45 skyler basepi: I am not sure if this is a bug, or just an unfixable problem. If I run with test=True, I change a mysql password, then have another state that has an unless that uses the new password, the unless will fail because the password will not have actually changed. Thus, it says that the state will run, then when I actually do highstate it does not.
00:46 skyler I am not sure if I should open an issue or not. It seems like a problem that there isn't a good way around.
00:47 ajolo joined #salt
00:50 irated joined #salt
00:50 irated Hey guys
00:50 irated is there a way to force all my grains to run sequentially
00:55 otter768 joined #salt
01:00 jalaziz joined #salt
01:00 n8n joined #salt
01:03 nitti joined #salt
01:08 possibilities joined #salt
01:10 pssblts joined #salt
01:11 vejdmn joined #salt
01:12 n8n joined #salt
01:15 irated Anyone?
01:16 mosen no idea personally
01:18 malinoff joined #salt
01:20 n8n joined #salt
01:22 ampex joined #salt
01:31 roo9- left #salt
01:32 possibilities joined #salt
01:32 ramishra joined #salt
01:38 druonysuse joined #salt
01:38 druonysuse joined #salt
01:50 unpaidbi1l joined #salt
01:50 druonysus joined #salt
01:52 TheThing joined #salt
01:52 unpaidbi1l hi guys.  is there any way to take several files eg: salt://base/file1, salt://base/file2, salt://base/file3 and merge them together to create a single file?  file.managed lets me specify multiple files but it's a first-match deal, file.append will not enforce that the file only contains the data from the three merged files, the file.replace doesnt let me match multiple lines to effectively clear out a file and use an accumulator to build it back up..
01:53 unpaidbi1l and jinja templates wont let me include a file using the salt:// urls
01:54 unpaidbi1l at least, i havent found a way to include a file with jinja by using the salt:// urls, eg {% for file in pillar['file_list'] %} {% include_salt file %} {% endfor %}
01:57 ingwaem joined #salt
01:58 unpaidbi1l oh - and file.accumulated doesn't let me specify a source: either , so that wont work!
02:04 schimmy joined #salt
02:05 ramishra joined #salt
02:08 schimmy1 joined #salt
02:09 iggy irated: doubtful... I'm pretty sure python's importer operates under the GIL
02:18 iggy unpaidbi1l: I don't think that kind of functionality exists, but since I'm probably going to need something similar soon, you might just see it in the future
02:18 diegows joined #salt
02:25 unpaidbi1l cool, i was thinking about hacking it in myself somewhere
02:26 unpaidbi1l if file.managed had an option 'append' that overrode the existing first-found functionality that'd be grand
02:26 unpaidbi1l there are probably better ways to do it though
02:28 iggy I was thinking something like /etc/somebigfile:\n  file.concatenate:    - template: jinja\n    - files:
02:28 iggy then a list of files
02:28 unpaidbi1l ah yeah
02:28 unpaidbi1l that would do the trick nicely
02:29 iggy we are going to need the templating, don't know if it's stricly necessary though
02:29 unpaidbi1l my particular case doesnt require templating, but it would be nice to have for the future
02:29 iggy support all the require/watch/etc stuff as well
02:29 iggy anything else you can think of?
02:30 unpaidbi1l i think that covers it and will fill a much needed hole in the file module
02:32 bhosmer joined #salt
02:34 unpaidbi1l or will file a hole that needs fillin' .. or something
02:34 malinoff left #salt
02:34 malinoff joined #salt
02:34 ndrei joined #salt
02:35 jalaziz joined #salt
02:39 stoicOne joined #salt
02:39 druonysus joined #salt
02:43 Outlander joined #salt
02:55 kusams joined #salt
03:01 ramishra joined #salt
03:02 cyberjames joined #salt
03:06 cberndt joined #salt
03:07 sroegner_ joined #salt
03:08 murrdoc joined #salt
03:09 iggy phrasing
03:11 vejdmn joined #salt
03:28 ericof joined #salt
03:31 rlarkin|2 joined #salt
03:39 cberndt joined #salt
03:42 murrdoc joined #salt
03:42 ramishra joined #salt
03:46 rawkode joined #salt
03:46 hasues joined #salt
04:01 yomilk joined #salt
04:02 Outlander_ joined #salt
04:06 Ryan_Lane joined #salt
04:13 dccc_ joined #salt
04:18 murrdoc anyone have success with reclass yet ?
04:20 bhosmer joined #salt
04:27 stephas joined #salt
04:31 mosen joined #salt
04:34 desposo joined #salt
04:45 ingwaem joined #salt
04:48 obimod joined #salt
04:59 bhosmer joined #salt
04:59 timbermaniac joined #salt
05:26 pssblts joined #salt
05:32 Nexpro joined #salt
05:33 ramishra joined #salt
05:38 mikeywaites joined #salt
05:48 schimmy joined #salt
05:52 schimmy1 joined #salt
06:02 colttt joined #salt
06:03 TyrfingMjolnir joined #salt
06:04 cyberjames left #salt
06:04 murrdoc joined #salt
06:05 scarcry joined #salt
06:06 ttrumm joined #salt
06:09 scarcry joined #salt
06:09 bhosmer joined #salt
06:15 scoates joined #salt
06:24 mfernandez joined #salt
06:25 schimmy joined #salt
06:29 schimmy1 joined #salt
06:30 cberndt joined #salt
06:32 baconbeckons joined #salt
06:34 ramishra joined #salt
06:42 adsisco joined #salt
06:42 adsisco any idea how do i config redis/mysql using salt after installing them?
06:43 baconbeckons left #salt
06:46 duncanmv joined #salt
06:50 cberndt joined #salt
06:53 kingel joined #salt
06:54 flyboy joined #salt
07:10 cberndt joined #salt
07:12 flyboy joined #salt
07:16 aranhoide joined #salt
07:17 aranhoide joined #salt
07:18 swah joined #salt
07:19 aranhoide in http://docs.saltstack.com/en/latest/topics/cloud/digitalocean.html I see that the authentication scheme in the Digital Ocean API is being used
07:19 aranhoide (I mean, of the v2.0 API)
07:20 aranhoide but when I try and launch a minion with salt-cloud on a salt 2014.1.11 (Hydrogen) installation it complains that my cloud.providers entry for Digital Ocean is missing the api_key entry
07:20 aranhoide which is what the v1.0 digital ocean API used to use
07:20 aranhoide what salt version do the online docs correspond to?
07:21 aranhoide OK, sorry, I saw it in the margin.
07:21 aranhoide docs are for the development version
07:22 Ryan_Lane joined #salt
07:23 lcavassa joined #salt
07:28 tomspur joined #salt
07:33 martoss joined #salt
07:33 martoss joined #salt
07:35 ramishra joined #salt
07:36 mechanicalduck joined #salt
07:37 martoss1 joined #salt
07:38 cberndt joined #salt
07:40 mikeywaites joined #salt
07:43 N-Mi joined #salt
07:43 N-Mi joined #salt
07:46 swah joined #salt
07:51 wnkz joined #salt
07:54 chiui joined #salt
07:58 bhosmer joined #salt
08:04 PI-Lloyd joined #salt
08:05 intellix joined #salt
08:18 cDR joined #salt
08:20 cberndt joined #salt
08:28 gywang joined #salt
08:29 gywang How did I include all files under a directory in pillar, such as include xxx.*
08:33 ramishra joined #salt
08:35 CeBe joined #salt
08:39 clemensb joined #salt
08:40 CycloHex joined #salt
08:41 astol joined #salt
08:54 nkuttler gywang: file.recurse
08:55 gywang nkuttler: sorry, I'm not mean this, I mean in pillar, like this: http://pastie.org/9636778
08:56 gywang Because I have lots of pillar files in directry packages
09:00 viq gywang: you don't. Or you have packages/init.sls that includes all of them by name, and then you can just do - packages in your top.sls
09:01 jaimed joined #salt
09:04 cberndt joined #salt
09:08 Twiglet_ joined #salt
09:09 martoss joined #salt
09:10 martoss1 joined #salt
09:10 rawkode joined #salt
09:14 harkx joined #salt
09:19 ramishra joined #salt
09:20 xsteadfastx joined #salt
09:27 trikke joined #salt
09:37 mikeywaites left #salt
09:37 Micromus Does salt do a apt-get update before installing a package?
09:38 Micromus found it!     - refresh: True
09:39 cberndt joined #salt
09:43 wnkz Hi, I have a problem with the docker(io) module/state when I want to pull a private image from Docker registry ; does someone know if private repo are unsupported or maybe my pillar for auth are configured wrong ? The docker API says that the image is not found.
09:47 bhosmer joined #salt
09:48 cofeineSunshine joined #salt
09:54 che-arne joined #salt
09:56 zooz joined #salt
09:58 badon joined #salt
10:00 ramishra joined #salt
10:00 bhosmer joined #salt
10:01 elfixit1 joined #salt
10:03 viq joined #salt
10:04 viq joined #salt
10:09 honestly how do I make a minion sync statefiles again
10:09 honestly man, it's been way too long...
10:10 cofeineSunshine saltutil.sync_all
10:10 wnkz__ joined #salt
10:11 giantlock joined #salt
10:11 honestly like this: salt '*' saltutil.sync_all
10:11 honestly ?
10:12 cofeineSunshine yup
10:14 SheetiS joined #salt
10:15 honestly how do I run a state with salt-call
10:16 honestly I've forgotte everything -.-
10:17 ze- honestly: salt-call state.sls name.of.the.state.file
10:17 honestly without a file
10:18 honestly I just want to run pkg.installed
10:18 honestly with some random package
10:18 honestly for testing
10:18 ze- do you need it to run a *state* ?
10:18 ze- i'd rather call directry salt-call pkg.install packagename
10:18 ze- calling the module directly
10:19 ze- doesn't do exactly the same thing, so it depends what you really want to do/try
10:20 honestly I have a statefile that is failing
10:20 ze- but if you really want to try a state: http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.state.html#salt.modules.state.high
10:20 honestly I'm trying to debug that
10:21 honestly cool, thanks
10:23 honestly whelp, something's wrong here
10:24 cofeineSunshine i'm trying to run masterles salt-minion on vagrant machine.
10:24 cofeineSunshine getting this error: This master address: 'salt' was previously resolvable but now fails to resolve! The previously resolved ip addr will continue to be used
10:24 cofeineSunshine the weird part is that it worket with 2014.1.10 version, now with 2014.1.11 getting aforementioned error
10:25 cofeineSunshine pillar cant be resolved properly, although DEBUG level shows pillars are being parsed
10:25 cofeineSunshine but '*' in pillar/top.sls doenst seems to be working now
10:27 viq cofeineSunshine: did you configure minion to use local files or are calling salt-call with --local flag?
10:29 deepz88 joined #salt
10:29 cofeineSunshine yeah, using --local flat, my /etc/salt/minion http://www.codepaste.eu/?hash=6f4ddebef5ca96068cbcc9d5a1b7d27a
10:29 cofeineSunshine file_client: local
10:29 honestly something is very wrong here with salt... http://pastie.org/9636985
10:32 ze- honestly: try salt-call -l debug to get more informations. probably informations about what pkg tries to do.
10:33 viq honestly: debian?
10:33 honestly ze-: that *is* -l debug
10:33 honestly but it seems the problem is that python-apt is missing
10:34 ze- weird that it even exports the function.
10:34 ze- salt-call -l debug, and you don't get anything *before* that ?
10:35 honestly ze-: sure I do, just nothing interesting
10:35 flyboy82 any way to delay execution of a cmd.run?
10:35 aqua^lsn joined #salt
10:35 ze- flyboy82: why would you want to "delay" the execution? (sleep?)
10:36 martoss joined #salt
10:36 flyboy82 i want to run rs.initiate on mongo primary and I've made it require mongod service but it takes some time before one can connect after the service is brought up
10:38 ze- what about 3 states. 1. install/start the service. 2. sleep long enough (watching the install, not to do it if nothing new done) 3. do what you have to do later (required the sleep)
10:38 viq flyboy82: sure, 'sleep 30 ; some_other_command' or play with 'at'
10:39 flyboy82 cheers
10:39 viq or what ze- sad\id
10:40 cofeineSunshine ze-: use query to lookup weather mongo is up. like mongo --eval 'rs.status()';
10:40 cofeineSunshine loop until it returns correct answer
10:41 cofeineSunshine cmd.run 'mongo --eval \'rs.status()\'';
10:41 cofeineSunshine atleast I did like that in salt runner
10:42 ze- cofeineSunshine: in a runner you can. in states, it's harder to do :)
10:43 rubenb Hi, Anybody got the digital_ocean_v2 api working in  RC3?
10:43 flyboy82 yea, I don't have a problem with that. I was asking about the initial mongo setup, since it takes a bit of time for the service to be brought up, I can't just go and run "mongo local --eval "printjson(rs.initiate())""
10:43 flyboy82 it will fail
10:43 flyboy82 I just need enough time for mongo to bind on its port and accept my connection
10:44 flyboy82 I have a different state to handle populating my rs with members
10:44 flyboy82 just need the rs to be initiated on the primary
10:49 kingel joined #salt
10:49 ndrei joined #salt
10:59 bluenemo joined #salt
11:00 ramishra joined #salt
11:01 bluenemo hi guys. I'm trying to iterate over some key value data defined in defaults.yaml, I however get the error Jinja variable 'tuple object' has no attribute 'items'. http://paste.debian.net/125502/
11:06 ze- is there a way to blacklist a grain module from being loaded ?
11:07 bigl0af joined #salt
11:07 bluenemo it works just fine when I only use pillar data without merging it into the rawmap before. any ideas?
11:08 ze- in ('datamap.server.sshd_config', {}).items()
11:08 ze- aren't you missing something before (' ?
11:08 ze- like xxx.get ?
11:09 ze- because right here, you have a tuple with a string and a dict, calling items() on it.
11:11 ze- maybe datamap.get('server', {}).get('sshd_config', {}).items() ?
11:12 scott_w joined #salt
11:12 scottpgallagher joined #salt
11:12 bluenemo cool that works! :D thank you ze-!! :)
11:13 kingel joined #salt
11:18 scott_walton joined #salt
11:20 calvinh joined #salt
11:23 Micromus I seem to have a deadlock situation. My nginx state watches /etc/nginx/conf.d/*. One of my specific-site states require the nginx state, as well as supply a /etc/nginx/conf.d/site.conf file.
11:24 Micromus As soon as I add the watch: /etc/nginx/conf.d/* to nginx, it seems it restarts nginx and returns service.running: False, and thus also failing the site-specific state
11:25 badon_ joined #salt
11:25 deepz88 joined #salt
11:27 agend joined #salt
11:32 Outlander joined #salt
11:34 jdmf joined #salt
11:36 bhosmer_ joined #salt
11:42 linjan joined #salt
11:47 ragrox joined #salt
11:48 ragrox hello
11:49 ragrox I have a little question. How can I get salt 2014.7.1 on a debian_x64 system?
11:49 ragrox The newest version is not in the repos
11:50 ragrox my salt-master is on an ubuntu machine, therefore has the newest version
11:50 ragrox it is debian 7 wheezy
11:51 mortis_ joined #salt
11:51 akafred joined #salt
11:54 Micromus ragrox: not in the saltstack repos? you can try the bootstrap script which will download and install minion from there
11:54 bhosmer joined #salt
11:54 akafred i have a bootstrap issue ... on a Ubuntu 14.04-box I do 'wget -O- --quiet https://bootstrap.saltstack.com | sudo sh -s -- -g https://github.com/saltstack/salt.git -M -N -D git v2014.1.10' ...
11:54 diegows joined #salt
11:55 akafred ... and it fails saying:  *  WARN: /tmp/git/salt/pkg/salt-api.upstart does not exist!
11:57 akafred ... which is true ... there is no salt-api.upstart in v2014.1.10 ...
11:58 babilen ragrox: That version has not yet been released as stable
11:58 gmcwhistler joined #salt
11:58 intellix joined #salt
12:01 ramishra joined #salt
12:01 akafred Cannot find an issue mentioning this ... so now I am unsure if it me or a bug ...
12:04 joehh ragrox: which version of salt do you have on ubuntu and how have you installed it?
12:06 fredvd joined #salt
12:06 ragrox I installed salt-master 2014.7.0-777 on the ubuntu machine, and it was installed over the ppa repos
12:07 ragrox salt-minion 2014.1.11 is installed on the debian machine
12:07 scoates joined #salt
12:07 mikeywaites1 joined #salt
12:08 mikeywaites1 is there any news on the bug with pillars and masterless minions
12:09 ramishra joined #salt
12:10 akafred I have also tried the develop-version of the bootstrap script: https://bootstrap.saltstack.com/develop ... same result.
12:12 mikeywaites1 curl -L https://bootstrap.saltstack.com | sudo sh -s -- git v2014.1.10 this resolves the issue for me.  but its pretty annoying to have to do this
12:13 dccc_ joined #salt
12:15 joehh ragrox: was 2014.7.0-777 from packages/bootstrap/something else?
12:18 ragrox it is from the launchpad ppa repos, I just followed the steps from there: http://docs.saltstack.com/en/latest/topics/installation/ubuntu.html
12:20 ragrox I will try bootstrap on the debian machine
12:28 ragrox okay, now it is salt-minion 2014.7.0. But this did not solve my real problem...
12:29 ragrox I do use Halite Salt UI, when I do pkg.upgrade on the debian machine I do get an error:
12:29 ragrox TypeError encountered executing pkg.upgrade: upgrade() takes at most 1 argument (9 given).
12:30 hobakill joined #salt
12:31 joehh that is what is confusing me - the packages at ppa:saltstack/salt (ie from following the instructions) are 2014.1.11
12:33 babilen ragrox: Could you paste the output of "apt-cache policy salt-{master,minion,common}" to http://refheap.com ?
12:38 colttt hello..
12:38 viq ragrox: 2014.7 is not released yet, only release candidates are out as far as I can see
12:39 hobakill hell i'm still waiting for 1.11 to get out of epel testing! :)
12:39 giannello joined #salt
12:39 babilen mikeywaites1: It's fixed in 2014.1 HEAD so you don't have to use .10, but could also use that. .12 has, however, not yet been released and joehh also did not package those patched in the .10 packages.
12:43 babilen ragrox: ping?
12:43 joehh babilen, mikeywaites1: upstream/saltstack have upped their testing processes, so feeling a little constrained about applying patches etc right now...
12:44 colttt i would like do the following: add this commands ind /etc/rc.local 'echo noop > /sys/block/sdX/queue/scheduler' how can i say that sdX is an variable mapped to the number of HDDs?
12:44 joehh especially with a .12 ready "real soon now"
12:44 babilen joehh: Sure, I am not criticising the packaging, but just pointing out that this has not been done
12:45 joehh no worries - just letting you know I'm semi deliberately not patching as actively as I used to
12:45 joehh used to/have been
12:47 babilen Well, IMHO the pillar patch would be an obvious/easy one to return (that massive commit that introduced this regression should have never been allowed into a "stable" release as bugfix), but I understand that upstream likes to keep tight control over the packaging (as opposed to most other upstreams that just release their software)
12:48 nitti joined #salt
12:53 kingel joined #salt
12:53 cofeineSunshine how can I debug masterless salt minion? It looks liek it doenst computes corect pillar
12:53 cofeineSunshine pillars are being parsed
12:53 cofeineSunshine but salt-call --local pillar.items doestn return them
12:53 cofeineSunshine there are pillar '*'
12:53 mikeywaites1 cofeineSunshine:  are you using version 2014.1.11
12:53 mikeywaites1 if so
12:54 mikeywaites1 theres a bug
12:54 mikeywaites1 which means pillars dont work with masterless minions
12:54 cofeineSunshine yes
12:54 mikeywaites1 curl -L https://bootstrap.saltstack.com | sudo sh -s -- git v2014.1.10
12:54 mikeywaites1 downgrade
12:54 mikeywaites1 this will solve the issue for now
12:55 mikeywaites1 theres a pull request containing a comment suggesting that this wasnt a huge issue, so now more people are being stung by it everyday hopefully we might get a release to solve the issue properly
12:55 cofeineSunshine huh, huge issue
12:55 bhosmer joined #salt
12:55 mikeywaites1 indeed
12:55 cofeineSunshine cant test localy all the changes
12:56 cpowell joined #salt
12:56 mikeywaites1 as i say all you can do at the moment is change version
12:56 miqui joined #salt
12:56 mikeywaites1 which is useless if your master is running .11
12:56 ragrox babilen: https://www.refheap.com/91538 here it is, for master and minion
12:57 joehh ragrox: so according to apt you should have 2014.1.11 on both
12:57 joehh what happens if you run salt --versions-report
12:57 acabrera joined #salt
12:58 ragrox I do now have on both the same version
12:58 joehh good
12:58 ragrox I installed via the bootstrap script salt minion on the debian machine
12:58 nickjj is there a way to somehow provide a * pattern to file_roots in the master config? use case: i might have 50 formulas but would like to give it 1 directory path where they all are, rather than manually edit my master config every time i add a formula
12:59 ragrox but the error from above is still there, I thought it was because of different versions
12:59 ramishra joined #salt
12:59 ragrox [14:29] <ragrox> I do use Halite Salt UI, when I do pkg.upgrade on the debian machine I do get an error: [14:29] <ragrox> TypeError encountered executing pkg.upgrade: upgrade() takes at most 1 argument (9 given).
13:01 ramishra joined #salt
13:01 bhosmer_ joined #salt
13:04 vejdmn joined #salt
13:06 ozzzo I'm trying to follow the instructions here: http://docs.saltstack.com/en/latest/ref/states/requisites.html#require-an-entire-sls-file
13:06 ozzzo but I get an error "Nonexistant saltenv 'test2' found in include of None within SLS 'base:test1'"
13:06 ozzzo http://pastebin.com/iHxdEKK6
13:06 ozzzo Can anyone tell me what I'm doing wrong?
13:07 mikeywaites1 ozzzo:  test2 obviously exists yeh?
13:07 mikeywaites1 and has an init.sls
13:07 nickjj ozzzo: is line 20 valid yaml?
13:07 mikeywaites1 no
13:08 mikeywaites1 pkg:
13:08 mikeywaites1 sorry
13:08 ozzzo what is an init.sls?
13:08 mikeywaites1 pkg:
13:08 mikeywaites1 - installed
13:08 mikeywaites1 - name: gcc
13:08 ozzzo ok I'll try that
13:10 kingel joined #salt
13:13 nitti joined #salt
13:13 bhosmer joined #salt
13:13 ozzzo mikeywaites1: I still get the same error: http://pastebin.com/kphZYC0p
13:14 ozzzo I cut/pasted the example in the doc and changed foo and bar to real packages
13:14 ozzzo but I'm not sure I got test2 right
13:15 mikeywaites1 test2/init.sls
13:15 mikeywaites1 try it this way
13:15 mikeywaites1 though im not sure thats the issue to be honest
13:16 mikeywaites1 though if you including test2
13:16 ozzzo you mean, rename test2.sls to init.sls?
13:16 mikeywaites1 no
13:16 mpanetta joined #salt
13:17 mikeywaites1 remove the : from the include - test2:
13:17 mikeywaites1 include:
13:17 mikeywaites1 - test2
13:18 ozzzo oic, trying that
13:18 mikeywaites1 ozzzo:  http://www.yaml.org/ ;)
13:18 ozzzo I get a different error now; I think that was the answer. ty!
13:18 mpanetta joined #salt
13:19 mpanetta joined #salt
13:26 nitti joined #salt
13:26 micah_chatt joined #salt
13:26 Twiglet_ joined #salt
13:27 nitti joined #salt
13:28 bigl0af joined #salt
13:28 totte joined #salt
13:31 oyvjel joined #salt
13:31 ozzzo mikeywaites1: The new error is: list indices must be integers, not str.
13:31 ozzzo do you know where I can find an example of a correct .sls include?
13:32 kermit joined #salt
13:32 ozzzo http://pastebin.com/qQaPQT1f
13:38 higgs001 joined #salt
13:40 teebes joined #salt
13:40 wnkz hi, anyone successfully using salt to pull private docker images with docker state/module ?
13:41 wnkz ^-- salt 2014.1.11 (Hydrogen) / Ubuntu 14.04
13:43 kusams joined #salt
13:48 intellix joined #salt
13:50 ze- running a command in batch mode seems to call saltutil.find_job alot. Is that a normal thing ?
13:50 micah_chatt_ joined #salt
13:51 vejdmn joined #salt
13:52 rallytime joined #salt
13:54 dude051 joined #salt
13:57 astol joined #salt
14:01 ericof joined #salt
14:04 xDamox joined #salt
14:05 pdayton joined #salt
14:08 dude^2 joined #salt
14:11 bhosmer joined #salt
14:13 jslatts joined #salt
14:14 dude051 joined #salt
14:14 kingel joined #salt
14:14 ]V[_ joined #salt
14:15 maze ze-, yes
14:17 ze- maze: mmm... which question was it again ? :)
14:19 duncanmv joined #salt
14:20 \ask joined #salt
14:21 UForgotten joined #salt
14:22 Whissi joined #salt
14:22 tkharju joined #salt
14:23 UtahDave joined #salt
14:24 higgs001 joined #salt
14:24 spiette joined #salt
14:25 jY joined #salt
14:26 UtahDave left #salt
14:28 jslatts joined #salt
14:29 mpanetta Is there any state to recursively change the permissions of a directory that already exists?
14:29 mpanetta I am creating a virtualenv with virtualenv.managed, but I don't want it to be owned by root.
14:30 mapu joined #salt
14:31 ramishra joined #salt
14:32 kingel joined #salt
14:32 UForgotten joined #salt
14:34 maze mpanetta, cmd.wait with a chmod/chown -R ?
14:34 iggy mpanetta: file.directory?
14:34 mpanetta iggy: I didn;t think file.directory would work if the directory already exists
14:34 maze iggy, does that work recursively?
14:34 mpanetta maze: It does
14:34 iggy it's got a recurse option
14:34 maze neat
14:35 mpanetta But I don't think it works on existing dirs
14:35 ajolo joined #salt
14:35 iggy mpanetta: it just says it ensures it exists and has the right perms
14:36 iggy I would expect that to mean "if it exists, it ignore the second half of it's purpose"
14:36 iggy *wouldn't
14:36 iggy but I've been wrong before
14:36 mpanetta Yeah I know, but what does it do if it already exists?  Most states don't run if what you are trying to do has already been done...
14:36 mpanetta Ahh
14:36 mpanetta hmm
14:36 mpanetta I guess I could just try it :P
14:37 rallytime joined #salt
14:37 bhosmer joined #salt
14:38 conan_the_destro joined #salt
14:39 iggy just read the module
14:39 iggy I don't see anything that says "if exists(name):  exit"
14:40 iggy so if it shortcircuits in the event the dir already exists it's non-obvious
14:40 mpanetta Ah ok cool
14:40 timoguin yea pretty sure virtualenv.managed will just return successful if it exists
14:42 timoguin mpanetta: you can also specify the user for it to run as when it's created, so that'll get around the owned as root issue
14:42 timoguin not sure if it will enforce the user if it's already created though
14:42 mpanetta timoguin: I never got that to work with virtualenv.managed :(
14:43 rallytime joined #salt
14:43 timoguin really? i'm using the user parameter just fine with 2014.1
14:44 mpanetta Hmm let me try
14:46 mpanetta If I add -user: poppy to the state, I get this: https://gist.github.com/anonymous/c5a9acf1846c255a0417
14:46 mpanetta timoguin: ^
14:47 mpanetta And it does not create the virtualenv
14:48 oyvjel joined #salt
14:48 timoguin mpanetta: 2014.1?
14:48 mpanetta 2014.1.5
14:49 timoguin wonder if that's a bug. it's been working fine for me with 2014.1.10+
14:49 StDiluted joined #salt
14:49 mpanetta Hmm, maybe
14:49 ponpanderer joined #salt
14:50 timoguin It's an odd error, like the virtualenv directory is created, but it can't find the python binary it should be installing...
14:50 timoguin my virtualenv version is 1.7.1.2
14:50 mpanetta It actually isn't created that is why it fails.
14:51 mpanetta The dir isn't there
14:51 mpanetta Unless it deletes it on failure?
14:51 timoguin maybe
14:51 timoguin all i'm doing is ensuring its base directory is there and the user already exists
14:52 ponpanderer hello, i'm running salt 2014.1.10 from epel repo and it seems the environ.set is unavailable: "Comment: State environ.set found in sls X is unavailable". Any ideas why?
14:52 mpanetta same here
14:52 mpanetta ponpanderer: Because it is only available in 2014.7
14:52 eunuchsocket joined #salt
14:52 mpanetta I had to copy the state and module to _states and _modules to use it
14:54 Ozack1 joined #salt
14:54 ponpanderer mpanetta: hmm thought this might be the case. seems the docs are a little inconsistent when stating 'only available in 2014.7' :) any ideas on a workaround to set http proxy env variables?
14:56 mpanetta ponpanderer: You could do what I did and copy the state and module files from develop and put them in a _state and _module dir in your formulas.
14:56 mpanetta It can drop in fine
14:56 mpanetta No odd dependencies or anything
14:58 jalbretsen joined #salt
14:59 ponpanderer i'll give that a try, my hesistance would only be odd dependencies but i'll take your word on it. definitely better than running a shell script that exports the var and downloads the file with wget or something
14:59 mpanetta Definitely
15:01 anotherZero joined #salt
15:03 kingel joined #salt
15:03 adolfo joined #salt
15:03 adolfo howdy
15:03 adolfo I have  a question regarding salt-syndic
15:04 ipmb joined #salt
15:04 adolfo it seems to me that a lower master acts as a proxy to its minions
15:05 adolfo I wanted to know how can I send cmds for the lower level master to do, such as get a file from an upper level master?
15:06 wendall911 joined #salt
15:08 timoguin adolfo: I think you'd want that lower-level master to be a minion as well
15:09 deepz88 joined #salt
15:09 adolfo ok so I should run salt-minion & salt-syndic on the lower level master?
15:10 timoguin yea i think so
15:13 adolfo k Thanks
15:13 kusams joined #salt
15:13 MTecknology I wish you could arbitrarily indent things in sls files. Nested conditions get rather confusing.
15:15 Heartsbane joined #salt
15:15 Heartsbane joined #salt
15:15 ponpanderer mpanetta: follow up question on the environ module. you mentioned puting the files in _state and _module. where do i put those dirs...in the state dir calling it, the root of the salt fs (/srv/salt) or somewhere else?
15:16 stoicOne joined #salt
15:17 iggy in the root of any "state" dir
15:17 MTecknology like, maybe start the state with a backtick to say you're removing indentation... Then you could do things like this...  http://dpaste.com/15NSMSM
15:18 iggy go write a completely new yaml parser to do it...
15:19 MTecknology Isn't that so much prettier than this? http://dpaste.com/3824PYE
15:19 MTecknology I get why we don't have it. It's just a bummer
15:19 iggy I've been using django for years, so...
15:19 MTecknology so...?
15:20 iggy I'm used to reading the django/jinja tags completely separately from the other markup
15:20 iggy I generally indent the jinja tags completely independent of the yaml
15:20 MTecknology ah
15:20 MTecknology I guess I could see doing that
15:21 StDiluted joined #salt
15:21 iggy and yeah, it can get messy from time to time
15:22 iggy but usually when I get to that point, I tell myself that I'm doing something wrong
15:22 timoguin that's why you gotta be choosy about how much logic you put in there
15:22 hasues joined #salt
15:22 timoguin I try not to go beyond a for loop or two
15:23 mpanetta ponpanderer: It just needs to be in file_roots
15:25 clayk joined #salt
15:26 StDiluted joined #salt
15:26 MTecknology the one I have now is {% if %} {% if %} {% endif %} {% if %} {% endif %} {% elseif %} {% for %} {% endfor %} {% endif %}
15:26 nitti_ joined #salt
15:27 clayk Hello. Looking for some good documentation on multiple environments. Possibly using multiple staging areas and multiple projects all within the same Salt setup. Most examples use a simple single environment. A
15:31 MTecknology How do I set @hourly for a cron state?
15:31 MTecknology I'm going to have to specify - hour: '*' - minute: '0', huh? :(
15:31 adolfo having both the minion and syndic does not seen to work
15:32 adolfo the minion takes precedence
15:33 martinp joined #salt
15:35 timoguin adolfo: that doesn't seem right. i've never ran syndic before though, so can't offer much help beyond that
15:35 ndrei joined #salt
15:36 nickjj does anyone know what would cause a 10-15 second delay on running a very trivial high state?
15:37 ckao joined #salt
15:37 nickjj the delay is in between running salt '*' state.highstate and seeing the results
15:37 weiser joined #salt
15:38 nickjj the tasks themselves report back 600ms of duration if i add them
15:38 cpowell does anyone know how salt-cloud delete servers? Does it perform a clean shutdown first or does it just yank the power cord essentially
15:38 martinp joined #salt
15:40 giannello cpowell, it uses the cloud provider API
15:40 giannello for EC2, it's a clean shutdown
15:44 weiser I'm doing a masterless provision with `salt-call 2014.1.11 (Hydrogen)` and I'm getting an error in one of my .sls files that pillar data isn't being found.  I've searched for documentation on how to list the contents of `pillar` during a debug run of state.highstate but couldn't find anything.
15:44 aquinas joined #salt
15:44 weiser could someone give me a pointer, please?
15:45 thedodd joined #salt
15:46 ponpanderer mpanetta: that worked nicely. thanks!
15:46 jcockhren weiser: there's a regression in 2014.1.11. I'd stick with 2014.1.10 if at all possible
15:46 weiser @jcockhren: thanks!
15:49 mpanetta ponpanderer: Welcome :)
15:52 Deevolution Does Salt support a 'source' parameter for the file state via UNC path on Windows?
15:53 eunuchsocket joined #salt
15:54 iggy I thought it only supported salt:// paths
15:54 Deevolution iggy: per the docs, it also supports http and/or ftp paths
15:55 iggy I suppose that makes sense, but real filesystem paths are bound to be hard to resolve
15:56 pipps joined #salt
15:56 timoguin it supports file:/// for files on the minion
15:56 timoguin not sure about UNC paths though
15:56 Deevolution I guess I'll try it and see.  And if need be I can throw the files in question onto an http share, but it's another hoop to jump through.
15:57 iggy mount the smb share on the master and set it as a file_root?
15:57 timoguin you could also mount the UNC path on the minion if UNC doesn't work
15:58 martinp joined #salt
15:58 timoguin supporting UNC paths doesn't seem unreasonable though, so probably a feature to add if it's not there
15:59 Deevolution it's present in puppet (we do it that way now).
15:59 TheThing joined #salt
15:59 Deevolution Mounting the share seems to be a heavy handed solution (I see that would work though).
16:00 jchen y
16:01 StDiluted joined #salt
16:01 Deevolution jchen: y is that heavy handed?  It's another pre-requisite for the master/minion.  If you're doing it on the master and you need to change where it's hosted, there's more coordination to sort out.  It's just a lot more complicated.
16:02 jchen oh shit, sorry, I was actually trying to press "y" to my salt-cloud thingy, focus was in this window by accident
16:02 jchen :|
16:03 timoguin ls
16:03 Deevolution LOL
16:03 Deevolution jchen:  NP
16:03 kingel joined #salt
16:05 hobakill can a secondary salt-master be a minion as well reporting to the main master?
16:05 robawt hobakill: i think a syndic might be a better fit
16:06 timoguin but, yes, that's possible.
16:06 kingel_ joined #salt
16:07 hobakill thanks timoguin & robawt
16:08 dude051 joined #salt
16:09 diegows joined #salt
16:12 pipps joined #salt
16:13 TheThing joined #salt
16:13 thedodd joined #salt
16:15 KyleG joined #salt
16:15 KyleG joined #salt
16:15 KennethWilke joined #salt
16:16 troyready joined #salt
16:17 pdayton joined #salt
16:18 nickjj is it necessary and/or reasonable to add an "onlyif" function when removing a package? i'm trying to eliminate any extra work
16:18 robawt nickjj: logic statements should always be reasonable, what are you trying to accomplish?
16:18 nickjj example: pkg.purge onlyif stat <some config path that would exist if the pkg were installed>
16:19 robawt that's not bad
16:19 nickjj robawt: just purging a package only if it's installed
16:19 hybridpollo joined #salt
16:20 ale_ joined #salt
16:20 nickjj robawt: would you do that even for more light weight actions like only deleting a path if it exists?
16:20 pipps joined #salt
16:21 timoguin if you use states for those operations they should handle that
16:21 timoguin i.e., using pkg.removed or file.absent
16:22 robawt so salt is always going to test then work, if you add additional logic that's ok, but don't over-check existing work
16:22 nickjj timoguin: i am yeah, it just seems like i'm setting myself up for a lot of unnecessary boilerplate maybe?
16:22 timoguin if they package or file isn't there, they should return successful and not do anything
16:22 robawt right nickjj
16:22 nickjj here's the exact implementation so far: https://gist.github.com/nickjj/525534429e3b898eca12
16:24 desposo joined #salt
16:24 timoguin yea you don't need onlyif for those
16:24 timoguin the states will check that in the background anyway before performing any actions
16:25 nickjj ok, i thought maybe i was saving some internal checks by having it -- good to know
16:25 bluenemo joined #salt
16:26 hardwire joined #salt
16:29 murrdoc joined #salt
16:31 forrest joined #salt
16:31 schimmy joined #salt
16:31 ndrei joined #salt
16:31 aparsons joined #salt
16:32 murrdoc why does github think sls files are scheme
16:32 murrdoc bad github
16:34 forrest murrdoc, I don't know man, it's been a problem forever. You think it could figure it out..
16:34 schimmy1 joined #salt
16:36 CeBe joined #salt
16:36 timoguin file extensions
16:36 timoguin scheme library source
16:36 murrdoc this is my vim setup for dev'ing salt http://pastebin.com/BCytc2FG
16:37 murrdoc in salt :)
16:37 baniir joined #salt
16:37 forrest murrdoc, https://github.com/saltstack-formulas/vim-formula/tree/master/vim
16:38 forrest murrdoc, if you want to add the missing stuff that would be cool.
16:38 murrdoc what u mean missing stuff ?
16:38 murrdoc nerdtree, syntastic exist in both
16:38 forrest you have stuff in yours that the formula doesn't have
16:38 murrdoc oh
16:38 timoguin murrdoc: check this out: https://github.com/StephenPCG/vim-snippets-salt
16:38 timoguin haven't used it let but looks awesome
16:38 murrdoc mine depends on salt.git
16:39 beneggett joined #salt
16:39 murrdoc you think thats ok to add to the formula
16:39 murrdoc i can get you a pull request forest
16:39 murrdoc the formulas seem to be fully contained
16:40 murrdoc forrest:  ?
16:41 kingel joined #salt
16:42 micah_chatt joined #salt
16:43 forrest murrdoc, I was suggesting to break it out for the formula into modular chunks like the other stuff is :D
16:44 murrdoc ah, got it
16:44 murrdoc soooon
16:45 timoguin the time is now!
16:45 spookah joined #salt
16:45 murrdoc :D company time now man
16:47 timoguin company is for sleepovers
16:47 jcockhren timoguin: going to pyTN 2015?
16:48 penguin_dan Is it possible inside of a salt state to copy a file if it exists, but not report a failure if it does note exist
16:48 timoguin jcockhren: sure am.
16:48 jcockhren awesome
16:48 timoguin won't be speaking or anything this time though
16:49 jcockhren I will try to do get a proposal done in the next week or so on some practical salt usage
16:49 jcockhren _try_
16:49 timoguin penguin_dan: {% if salt['file.exists']('filepath') %} do copying here {% endif %}
16:50 timoguin jcockhren: sweet, lemme know if you need any help with ideas or someone to review or whatever
16:50 penguin_dan timoguin: This will copy the file if it exists on remote system?
16:50 jcockhren timoguin: will do!
16:50 tligda joined #salt
16:51 timoguin penguin_dan: yea, taht will will the file.exists module on the minion and will only go into the if block if it exists
16:51 timoguin *will run
16:51 penguin_dan timoguin: Interesting.  I'll give that a shot thanks
16:55 scarcry joined #salt
17:00 aparsons joined #salt
17:00 desposo joined #salt
17:00 kermit joined #salt
17:00 schimmy joined #salt
17:04 ninkotech joined #salt
17:04 schimmy1 joined #salt
17:09 penguin_dan timoguin: getting this error when using that syntax : failed: Jinja variable 'dict object' has no attribute 'file.exists';
17:10 pipps joined #salt
17:20 schimmy joined #salt
17:20 thehaven joined #salt
17:22 beneggett joined #salt
17:24 aw110f joined #salt
17:26 perfectsine joined #salt
17:27 snuffeluffegus joined #salt
17:29 linjan joined #salt
17:30 XenophonF joined #salt
17:31 XenophonF what function does grains.filter_by call to merge settings?
17:31 TyrfingMjolnir joined #salt
17:31 XenophonF I want to use the merge functionality without also using filter_by
17:34 XenophonF ah never mind
17:34 XenophonF found it in salt/modules/grains.py
17:34 XenophonF it's salt.utils.dictupdate.update()
17:36 murrdoc nice
17:41 n8n joined #salt
17:41 mikeywaites joined #salt
17:41 kingel joined #salt
17:42 schristensen joined #salt
17:50 mapu joined #salt
17:54 cpowell joined #salt
17:54 druonysus joined #salt
17:54 druonysus joined #salt
17:58 dalexander joined #salt
18:00 pipps joined #salt
18:01 kermit joined #salt
18:02 ampex joined #salt
18:04 kusams joined #salt
18:16 schimmy1 joined #salt
18:16 jab416171 joined #salt
18:19 jalaziz joined #salt
18:19 chiui joined #salt
18:19 nickjj do a lot of you guys just write custom states instead of formulas when it makes sense to do so?
18:20 mr_chris nickjj, I do.
18:20 mr_chris I write many py rendered states too.
18:20 mr_chris Formulas are more meant as starting points.
18:21 nickjj mr_chris: i'm coming from ansible where it's common to have 50 or even 100 semi-generic roles written
18:21 alainv nickjj: I do. It's the standard premature optimization tradeoff - it's not always sensible to make a fully standardized formula
18:21 nickjj but i'm noticing with salt, it seems really popular to make states
18:21 nickjj and then i guess your formula could end up being a combination of custom states to do something more higher level?
18:22 murrdoc yes
18:23 murrdoc iterate away
18:23 murrdoc formulas are always better, cos yuo can commit to commonalities
18:23 nickjj so states are what end up being generic things -- and your formulas are usually more infrastructure specific?
18:24 murrdoc states that you apply to a lot of roles
18:24 nickjj what makes a formula always better? just because you can add it to your top file and with sane defaults you're good to go?
18:24 murrdoc and are based around one app
18:24 murrdoc are a formula
18:24 rap424 joined #salt
18:24 murrdoc for example say you are web shop
18:24 murrdoc nginx is a good formula to aim for
18:24 murrdoc or apache
18:24 pipps joined #salt
18:24 murrdoc a formula is a logical best practice
18:25 murrdoc for your environment
18:25 nickjj hmm, what about things like ntp, ferm, and syslog-ng -- these are better suited towards being states?
18:25 murrdoc they can be formulas too
18:25 murrdoc each role, has to define an overriding pillar , if needed
18:25 murrdoc to use the same formulas
18:25 murrdoc imho
18:26 nickjj and by role you just mean a collection of servers defined as "something"?
18:26 nickjj appservers, dbservers, etc.
18:26 murrdoc yeah
18:26 murrdoc 'have 50 or even 100 semi-generic roles written'
18:27 nickjj yeah but roles in an ansible sense -- in salt that would be 50-100 formulas
18:27 murrdoc oh
18:27 murrdoc hmm, not sure
18:27 murrdoc you can group formula to a state
18:27 murrdoc make that a role
18:28 astol joined #salt
18:28 nickjj i'm just wondering if i i should use the syslog-ng state or just make a formula out of it
18:28 murrdoc i would do the latter
18:28 nickjj i guess you could make a formula that uses the state, but that seems like a lot of abstractions/could be worse than just not using the state
18:29 murrdoc have u read the formulas ?
18:29 nickjj in general? yeah
18:30 murrdoc brb
18:31 kusams_ joined #salt
18:32 timoguin penguin_dan: sorry, file.file_exists is the one you need
18:32 beneggett joined #salt
18:36 bassam joined #salt
18:40 kingel joined #salt
18:44 ndrei joined #salt
18:44 druonysuse joined #salt
18:44 druonysuse joined #salt
18:46 istram joined #salt
18:47 thedodd joined #salt
18:50 StDiluted I am having a problem where one state executes an rm -rf of some files, and another one recreates those files, but the rm -rf is not completed before the second state starts, so they are conflicting. What’s the best way to guarantee that the second state does not happen until the first one has finished executing?
18:50 __TheDodd__ joined #salt
18:50 klj joined #salt
18:51 pipps joined #salt
18:52 forrest StDiluted, how are you handling the rm -rf?
18:52 forrest cmd. run?
18:52 forrest *cmd.run
18:53 ninkotech joined #salt
18:54 StDiluted it’s actually a bundle exec rake assets:clean
18:54 StDiluted it takes a little while
18:54 StDiluted that’s the problem
18:55 nickjj hmm, wouldn't it only move onto the next state once the previous one finishes?
18:56 StDiluted that was my assumption too but it doesn’t seem to
18:57 viq StDiluted: how about file.absent instead of rm -rf ?
18:58 glyf joined #salt
18:59 StDiluted well, I _think_ that’s all that the rake assets:clean does, but I’m not sure
18:59 viq Or you could put in there somewhere a "while `pgrep -fl 'rm -rf'` ; do sleep 5 ; done" and require that in next step
19:02 nickjj StDiluted: what are you expecting clean to do?
19:03 StDiluted nickjj: clean out the assets dir, is all I _think_ it does, but I am not positive. In the event something laters gets added to the rake task, though, I don’t want to substitute a cmd.run that just does the rm
19:03 gildegoma joined #salt
19:03 nickjj StDiluted: what's your definition of clean tho? nuking the entire thing?
19:03 nickjj because if it is, then you want to run assets:clobber , clean only removes outdated assets
19:03 gmcwhistler joined #salt
19:06 StDiluted this is what happens when i run assets:clean
19:06 StDiluted > /usr/local/bin/ruby /usr/local/bin/rake assets:clean:all RAILS_ENV=production RAILS_GROUPS=assets
19:06 StDiluted > rm -rf /var/www/atpay/public/assets
19:07 TOoSmOotH joined #salt
19:07 ndrei joined #salt
19:08 nickjj what version of rails are you using?
19:08 StDiluted 3.2 stable
19:09 nickjj ah, i think 4+ is when they changed clean to clobber
19:09 StDiluted I changed it to bundle exec rake assets:clean && sleep 30 && bundle exec rake assets:precompile and that seemed to work
19:09 StDiluted in my cmd.run
19:10 StDiluted I’ll give that some testing
19:11 nickjj would you mind sharing the exact state? i'm new to salt but familiar with rails deployments
19:11 StDiluted one sec
19:11 hasues joined #salt
19:11 nickjj it seems really troublesome to me if it doesn't execute states in sequential order
19:14 ndrei joined #salt
19:15 honestly I wish salt were more responsive.
19:15 StDiluted nickjj: https://gist.github.com/dginther/6de7e713251eebee08c4
19:15 StDiluted nickjj: i use a lot of cmd.wait and a lot of requires
19:15 perfectsine joined #salt
19:16 murrdoc responsive
19:16 murrdoc ?
19:16 honestly yes, as in telling me more things when a job is running
19:16 honestly such as whether the minion is there
19:16 murrdoc oh
19:16 honestly and when the minion has compiled the state etc.
19:17 murrdoc its the nature of eventual consistence
19:17 StDiluted like, a real time account of what the hell is going on, rather than a blank stare? ;)
19:17 ninkotech joined #salt
19:17 nickjj checking
19:17 kusams joined #salt
19:17 Obrit joined #salt
19:17 honestly murrdoc: well, doesn't help if it never eventually becomes consistent because the minion is actually not there
19:18 murrdoc you win
19:18 nickjj line 28 just means this task runs only if install_bundle changed?
19:18 StDiluted nick_jj: yes
19:19 nickjj what if you change your assets but not your gems
19:20 StDiluted yeah, I’m open to other ways of doing this crap, this is _generally_ speaking not going to be a deploy to all minions, but rather an initial pass of setting up the machine. deploys will be handled thereafter by capistrano
19:20 jalaziz joined #salt
19:20 stoicOne joined #salt
19:21 glyf joined #salt
19:21 pipps joined #salt
19:22 StDiluted I could be going about this all wrong, to be honest. I haven’t found a good way of deploying a rails app with salt.
19:23 StDiluted it’s ugly and not how I want it to work
19:23 StDiluted but it _works_ mostly
19:23 nickjj i'm just getting started with salt but i planned to tackle rails after i get more comfy
19:23 nickjj i can show you an ansible version of a rails deploy, i'm quite confident it could be made to work (even better) in salt
19:24 StDiluted it would be great if there were a state module for rails/rake
19:24 jalaziz joined #salt
19:25 StDiluted i wonder if this works: https://gist.github.com/auser/5044026
19:25 murrdoc joined #salt
19:25 StDiluted it doesnt look like auser has been very active
19:25 ninkotech joined #salt
19:25 StDiluted on that gist
19:25 nickjj here's the flow for a typical deploy https://gist.github.com/nickjj/ae64d00024d7d1d4cd36
19:28 StDiluted so because I’m using a bundle install cmd.run, it always gets run, and always shows as having made changes to the system
19:28 StDiluted so it always runs the other stuff as well
19:28 StDiluted (just tried another highstate)
19:29 pipps joined #salt
19:31 nickjj StDiluted: check this http://docs.saltstack.com/en/latest/ref/states/requisites.html#direct-requisite-and-requisite-in-types
19:32 nickjj specifically the "onchanges"
19:32 honestly it seems that salt's service provider doesn't handle debian with systemd running sysvcompat :/
19:33 nitti joined #salt
19:40 honestly anyone know how to find out which module is actually providing service?
19:41 druonysuse joined #salt
19:42 mechanicalduck_ joined #salt
19:43 beneggett joined #salt
19:43 pipps joined #salt
19:44 TOoSmOotH joined #salt
19:45 Andrevan joined #salt
19:45 honestly ah, it's using systemd
19:46 aparsons joined #salt
19:47 kusams joined #salt
19:47 honestly that was apparently fixed in April... https://github.com/saltstack/salt/pull/11921
19:47 * JordanTesting looks around
19:48 thehaven joined #salt
19:48 honestly but not release yet :|
19:48 beneggett joined #salt
19:48 honestly too bad
19:49 murrdoc it was reopened
19:49 murrdoc https://github.com/saltstack/salt/issues/11900
19:50 honestly that seems like a duplicate
19:51 kingel joined #salt
19:52 Ahrotahntee is it possible to hash a string in a jinja template?
19:52 nitti_ joined #salt
19:52 honestly Ahrotahntee: Ahrotahntee detected
19:53 linjan joined #salt
19:53 Outlander joined #salt
19:53 * Ahrotahntee paints a target on honestly
19:53 honestly you can run arbitrary python in jinja I think
19:53 honestly - hash: {{ "password".hash() }}
19:54 honestly you just need to somehow get the hash function you want in scope
19:54 Ahrotahntee I can import the lib
19:55 pipps joined #salt
19:56 jhulten joined #salt
19:57 honestly might be clearer code if you just make a py statefile instead of jinja
19:57 timoguin I think there are some hashing modules in salt you can use
19:57 mechanicalduck joined #salt
19:58 timoguin yea, the hashutil module: http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.hashutil.html
19:59 Ryan_Lane joined #salt
20:00 honestly ugh wtf
20:01 honestly why does jinja eat all the linebreaks
20:01 beneggett joined #salt
20:01 robawt honestly: don't use {%- -%} ?
20:01 honestly yeah...
20:01 honestly correction: why do people post broken sample code
20:02 timoguin they don't know it's broken
20:03 robawt timoguin++
20:04 kusams_ joined #salt
20:04 kvt joined #salt
20:05 kvt Hi folks, is there any way to have job execution timeout on salt-minion? So long running jobs would be killed automatically?
20:06 honestly I'm using provider: service for a service.running state but it uses systemd instead :/
20:07 robawt honestly: doc says there's a systemd provider
20:07 honestly yes
20:07 perfectsine joined #salt
20:07 honestly what the doc neglects to mention is that the systemd provider is broken on all systems that don't run pure systemd
20:07 honestly read up :P
20:08 cberndt joined #salt
20:09 robawt wow that's terrible
20:10 honestly I agree, nobody should be running systemd in sysvcompat mode, it's just cruel
20:10 honestly >_>
20:10 eunuchsocket joined #salt
20:11 honestly unfortunately I don't have a choice
20:13 gladiatr joined #salt
20:14 Ryan_Lane basepi: we actually came to terms with the default "smart" recursive merge strategy
20:15 Ryan_Lane we weren't reallly affected by it anywhere until this new service I launched, and it was easy enough to change our pillar approach to make it work
20:15 basepi Still, it will likely pull the rug out from under others
20:15 basepi So I want to get the old behavior back as default.
20:16 robawt basepi: old behavior == pillar merges last known?
20:16 robawt :D
20:16 Ryan_Lane basepi: I don't think so
20:16 timoguin yea i haven't tested yet, but i'm pretty sure anything but the old default will break a couple things
20:16 Ryan_Lane basepi: I'm updating the ticket
20:17 gladiatr Hey all.  I'm working on a custom execution module.  I'm pretty sure I've tested this on my salt lab systems in the last week.  After having been pulled away on other projects for 4 days, I came back to it today and am seeing the following error message: http://pastebin.com/cQ8enyGt  I've found a number of references to old bugs, but because I'm working with a custom module, I can't help but think it is something I'm not handling prop
20:17 gladiatr erly.  The module (or at least the parts that I've finished) work outside of salt.
20:17 basepi robawt: "merges last known"? not following. the old pillar behavior was to merge nothing. it would just overwrite keys
20:17 basepi afaik
20:17 kingel joined #salt
20:18 Ryan_Lane basepi: updated the ticket
20:18 Ryan_Lane if it's changed back to overwrite by default, then we have a regression for 2014.1
20:18 Ryan_Lane if not, we have a regression for 0.17
20:18 Ryan_Lane but the 0.17 regression already happened
20:18 Ryan_Lane so, it's likely best to stick with smart, but add the overwrite option
20:19 baniir joined #salt
20:19 honestly I can't seem to make salt not run systemd :|
20:19 honestly call systmctl*
20:20 honestly provider: service and provider: debian_service both call systemctl
20:22 honestly why does the debian_service module directly try to import systemd stuff :/
20:22 honestly this makes no sense
20:22 kvt so no way to stop long running job on the minion?
20:23 Ryan_Lane basepi: seeing as that I'm the first person reporting this, I think it's not that as major of a problem as it looks :)
20:23 Ryan_Lane since it's been the default for an entire release
20:25 kingel joined #salt
20:25 kusams joined #salt
20:31 druonysus joined #salt
20:31 druonysus joined #salt
20:31 perfectsine joined #salt
20:35 hasues joined #salt
20:35 perfectsine joined #salt
20:36 murrdoc joined #salt
20:38 ndrei joined #salt
20:39 wormwood joined #salt
20:39 bhosmer joined #salt
20:53 runnyspot joined #salt
20:53 jalaziz joined #salt
20:53 aw110f joined #salt
20:55 beneggett joined #salt
21:00 murrdoc joined #salt
21:06 unpaidbi1l i just executed: salt --show-timeout -v -b 1 '*' state.highstate   to test some things out - there are 3 minions it should have updated.  the command didnt print anything, it exited without anything being printed and i don't see anything in my job_list.  i would think the show timeout and verbose flag would have provided some kind of output
21:06 unpaidbi1l is this a bug or did i just do something wrong?
21:07 unpaidbi1l on a subsequent execution it ran fine, updated the three servers one at a time with the expected output
21:08 aparsons joined #salt
21:11 teebes joined #salt
21:11 mrh joined #salt
21:13 mrh Hi there, is anyone know how to generate keys with MultiMaster PKI setup if "master_sign_pubkey: True" in config doesn't work?
21:17 mrh left #salt
21:18 mrh_666_3383 joined #salt
21:18 mrh_666_3383 Hi there, is anyone know how to generate keys with MultiMaster PKI setup if "master_sign_pubkey: True" in config doesn't work?
21:19 bhosmer joined #salt
21:20 bhosmer_ joined #salt
21:20 mrh_666_3383 Version: Salt: 2014.1.11
21:20 jalaziz joined #salt
21:22 kballou joined #salt
21:23 SheetiS1 joined #salt
21:23 kusams joined #salt
21:24 n8n joined #salt
21:35 peters-tx joined #salt
21:36 kingel joined #salt
21:36 TOoSmOotH joined #salt
21:41 bmonty joined #salt
21:41 teepark joined #salt
21:48 n8n joined #salt
21:52 Outlander joined #salt
21:58 kermit joined #salt
21:59 n8n joined #salt
22:02 glyf joined #salt
22:11 intellix joined #salt
22:15 aparsons joined #salt
22:17 gildegoma joined #salt
22:23 TOoSmOotH joined #salt
22:23 jergerber joined #salt
22:25 jalaziz joined #salt
22:28 pipps joined #salt
22:33 derelm joined #salt
22:34 vejdmn joined #salt
22:38 ajprog_laptop joined #salt
22:43 derelm joined #salt
22:44 mrh_666_3383 Anyone? How to generate keys with MultiMaster PKI setup if "master_sign_pubkey: True" in config doesn't work (v. 2014.1.11)?
22:49 n8n joined #salt
22:55 bassam joined #salt
22:55 bassam Hi. does salt-minion run directly on a coreos host?  Or is this a bad idea?
22:55 pipps joined #salt
22:57 pipps_ joined #salt
23:01 archrs joined #salt
23:04 badon joined #salt
23:07 cads joined #salt
23:09 bhosmer joined #salt
23:09 intellix joined #salt
23:21 kusams joined #salt
23:23 jalaziz joined #salt
23:30 aquinas_ joined #salt
23:31 SheetiS joined #salt
23:31 perfectsine joined #salt
23:32 n8n joined #salt
23:32 StDiluted joined #salt
23:35 sroegner_ joined #salt
23:42 davedash joined #salt
23:42 ajolo joined #salt
23:43 pipps joined #salt
23:44 aparsons joined #salt
23:48 phblj2 joined #salt
23:50 phblj2 left #salt
23:51 anon33994 joined #salt
23:51 anon33994 test
23:51 anon33994 whoops, sorry
23:52 phblj joined #salt
23:53 anon33994 left #salt
23:57 phblj Hey all-- I'm having trouble getting gitfs set up on 2014.1.11.  I've got a private repo on github, and have a deploy key set up along with ~/.ssh/config to use it.  Fetching from the command line or going through the same steps salt/fileserver/gitfs.py does with GitPython works fine, but when I `salt-run fileserver.update` I get the "Permission denied (publickey)" error.  Brand new to salt, so I'm not sure what other info would be helpful

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