Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-02-09

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

All times shown according to UTC.

Time Nick Message
00:10 s8weber__ joined #salt
00:12 jdowning joined #salt
00:17 yomilk joined #salt
00:20 jonatas_oliveira joined #salt
00:43 bhosmer_ joined #salt
00:44 TTimo joined #salt
00:50 octarine left #salt
00:50 mapu joined #salt
01:03 elfixit joined #salt
01:04 cberndt joined #salt
01:07 jonatas_oliveira joined #salt
01:12 badon joined #salt
01:12 iwishiwerearobot joined #salt
01:12 Georgyo joined #salt
01:15 otter768 joined #salt
01:20 aqua^mac joined #salt
01:23 jonatas_oliveira joined #salt
01:24 ralala joined #salt
01:25 bhosmer joined #salt
01:26 TTimo joined #salt
01:30 primechuck joined #salt
01:30 jonatas_oliveira joined #salt
01:32 derp1 joined #salt
01:32 derp1 left #salt
01:32 Auroch joined #salt
01:32 ralala joined #salt
01:40 ralala joined #salt
01:44 EWDurbin joined #salt
01:45 flabberflubber joined #salt
01:46 manytrees joined #salt
01:46 flabberflubber good day/evening - 2014.7.1 is still in epel testing. let's karma that guy out of there and into stable yeah? WHEEEEEEEEEEEEEEEEEEEEE!
01:48 codekobe joined #salt
01:48 kwmiebach joined #salt
01:51 gyre007 joined #salt
01:54 moderation joined #salt
01:54 wiqd_ joined #salt
01:55 jay_d joined #salt
02:02 jhauser_ joined #salt
02:09 jerematic joined #salt
02:12 tligda joined #salt
02:17 bhosmer_ joined #salt
02:34 tligda joined #salt
02:40 bhosmer_ joined #salt
02:46 atree joined #salt
02:56 favadi joined #salt
02:59 murrdoc joined #salt
03:00 iwishiwerearobot joined #salt
03:01 jhauser_ joined #salt
03:10 Auroch joined #salt
03:11 Auroch joined #salt
03:17 tkharju joined #salt
03:18 Ryan_Lane joined #salt
03:19 primechuck joined #salt
03:24 net128 joined #salt
03:36 MugginsM joined #salt
03:37 desposo joined #salt
03:38 monkey66 left #salt
03:42 monkey661 joined #salt
03:45 monkey66 joined #salt
03:58 jerematic joined #salt
04:13 lnxnut joined #salt
04:21 jalaziz joined #salt
04:30 higgs001 joined #salt
04:31 Furao joined #salt
04:43 clintberry joined #salt
04:49 lnxnut joined #salt
04:49 iwishiwerearobot joined #salt
04:51 clintberry joined #salt
04:52 tkharju joined #salt
04:52 cberndt joined #salt
04:52 Madhurranjan joined #salt
04:57 Madhurranjan joined #salt
05:04 armguy joined #salt
05:08 primechuck joined #salt
05:09 jeddi joined #salt
05:25 eyeball01 hey folks. noob question. I'm trying to write a state file for centos and i need to add the repository to yum.repos.d, The documentation on the saltstack site shows how to do it for ubuntu but nothing for rhel.
05:25 mosen I think theres a yumrepo state humm
05:26 mosen ahh its pkgrepo
05:27 mosen eyeball01: did you see the top example here: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkgrepo.html#module-salt.states.pkgrepo ?
05:27 eyeball01 nice thanks a lot
05:28 jalaziz joined #salt
05:28 mosen no probs eyeball01
05:28 Madhurranjan joined #salt
05:32 felskrone joined #salt
05:37 atree joined #salt
05:37 eyeball01 another question. how do you test your states?
05:42 kermit joined #salt
05:43 singularo You mean like : salt 'server.name' state.highstate test=true
05:46 Madhurranjan Thats a dry run . How do you test custom states ? Can you talk about how you do that ?
05:46 Madhurranjan Dry runs are really useful as well
05:46 jerematic joined #salt
05:50 lnxnut joined #salt
05:50 Furao eyeball01: i wrote a salt states testing framework for this. it integrate with jenkins. it even validate that pillar keys are documented. default value don’t change over multiple different calls. that metrics are really produced. that formula uninstall and install. monitor checks are performed after each units to make sure formula software run properly. etc etc. we worked on this for few years now. there is almost 1000 units and run sometimes for 2 days for a si
05:50 Furao test run. but we have a lot of formulas. http://picpaste.com/pics/Screen_Shot_2015-02-09_at_12.45.05.1423461004.png
05:51 eyeball01 k. i'm new to config mgmt typically we use fabric for almost everything but with 300 plus servers and growing that's not a scalable solution. So i'm using my free time to learn on my own so i don't have any servers to test on. I was considering using vagrant but i'm not really familiar with it. so i guess i'm just looking for suggestions on how experienced engineers test
05:52 ramteid joined #salt
05:53 eyeball01 i'm starting off simple by trying to set up lemp and deploying a simple webapp
05:58 monkey66 left #salt
05:59 GabLeRoux I learned saltstack lately, there's some good integration with vagrant and other tools. I really like the ouput when things work and don't once you run the commands. But it's not very easy to handle at first
06:01 GabLeRoux You can manage alot of servers with it, there's that concept of "master" controlling "minions". But that was overkill for my singler server website I used it for. Colleagues used it for around 20 servers with same applications (that has too many dependencies!). It really does the job very well, but sometime, you just loose your time with the yml syntax where you forget a - ;)
06:01 mosen eyeball01: I'm not sure about that one. salt have their own tests for states, some people would test in vagrant. I was going to look at kitchen-salt maybe
06:03 mosen eyeball01: for puppet I used to use beaker which is similar to test kitchen. You can spin up multiple VMs to test your states. looking at kitchen-salt now I'm not sure if it's current
06:04 GabLeRoux huh, sorry, I tought I was on #django, I just realized I was explaining what salt is. sorry about that, I probably need some sleep
06:05 pdayton joined #salt
06:06 Furao GabLeRoux: va donc te coucher :P
06:06 monkey66 joined #salt
06:08 pravka joined #salt
06:08 eyeball01 i will check out kitchen-salt. i liked chef because it's a ruby DSL and i'm familiar with ruby but the decision was made above me to use salt instead. not complaining because this has the potential to replace fabric or at least decrease the dependency on fabric anyway
06:09 monkey66 joined #salt
06:09 TTimo joined #salt
06:17 smithd joined #salt
06:17 mosen eyeball01: I'm pretty used to declaring stuff in YAML now.
06:18 Madhurranjan Furao: Is that open source ?
06:19 Furao Madhurranjan: somewhat but the problem is that it don’t work on any kind of formula it expect them to implement specific features and documented in a strict way
06:20 Madhurranjan Furao: ok
06:20 Furao but it produce useful doc such as https://doc.robotinfra.com/varnish/doc/index.html
06:22 Ryan_Lane joined #salt
06:23 calvinh joined #salt
06:25 calvinh_ joined #salt
06:34 smithd joined #salt
06:35 monkey66 joined #salt
06:38 iwishiwerearobot joined #salt
06:39 jalaziz joined #salt
06:41 bhosmer_ joined #salt
06:46 krelo joined #salt
06:53 GabLeRoux joined #salt
06:56 primechuck joined #salt
06:57 andrew_v joined #salt
06:58 colttt joined #salt
07:00 AndreasLutro joined #salt
07:00 jerematic joined #salt
07:09 TTimo joined #salt
07:11 xaoci joined #salt
07:14 krelo joined #salt
07:19 otter768 joined #salt
07:22 Ryan_Lane joined #salt
07:32 monkey66 left #salt
07:34 cberndt joined #salt
07:36 Grokzen joined #salt
07:38 nikogonzo is it possible to access grains['fqdn
07:38 nikogonzo '] from inside a statefile for all hosts under a master's control?
07:38 slafs joined #salt
07:38 slafs left #salt
07:39 Furao {{ grains['fqdn'] }}
07:39 Furao simply
07:39 Furao https://github.com/bclermont/states/blob/master/states/apt/init.sls#L38
07:40 nikogonzo right, i get that part - but in my test case it only puts a file on the remote host with that particular host's fqdn
07:40 nikogonzo i would like a file to appear on the client containing the fqdn of all known hosts ( as far as the salt master is concerned )
07:40 krelo joined #salt
07:40 mikkn joined #salt
07:41 jalaziz_ joined #salt
07:42 nikogonzo a use case would be like, an automated /etc/hosts with the salt master acting authoritatively, or a bind zonefile
07:43 smithd_ joined #salt
07:44 Furao {% if grains[“fqdn”] == “example.com” %}… put that file .. {% endif %}
07:44 Furao no?
07:46 nikogonzo thanks but that's not quite what I'm looking for. I'm wondering if it would be possible for me to do something like create a 'dns_server' state whos sls files have something in them to create a bind zonefile containing A records for all known salt minions using their grains['hostname'] and grains['ipv4']
07:46 Furao ahh ok ok i understand now
07:46 Furao i did something like that for automatic discovery of hosts to monitor
07:47 nikogonzo yeah, exactly!
07:47 linjan joined #salt
07:47 Furao I eventually created a new state for that “data” which allow a .sls to publish on salt mine some custom data for every formula
07:48 Furao such as expose the ssh server key public for a git server for all minion that need to connect to it to add it to their known hosts
07:48 Furao tell monitoring server what every minion need to monitored, etc
07:49 Furao but more simply, in your case, you can use salt mine and a module to publish those 2 info from every minions back to the master
07:49 Furao and have them available to other minions
07:49 Furao and craft the /etc/hosts from it
07:50 nikogonzo i am reading about salt mine now, thank you Furao !
07:50 MatthewsFace joined #salt
07:53 calvinh joined #salt
07:54 CeBe joined #salt
07:54 BigBear joined #salt
07:55 calvinh_ joined #salt
07:55 iwishiwerearobot joined #salt
07:57 Grokzen joined #salt
07:58 Auroch joined #salt
07:58 Madhurranjan joined #salt
07:58 intellix joined #salt
07:58 jtang_ joined #salt
08:01 trikke joined #salt
08:01 hebz0rl joined #salt
08:08 calvinh joined #salt
08:10 chiui joined #salt
08:10 kawa2014 joined #salt
08:11 TTimo joined #salt
08:11 lb1a joined #salt
08:12 iml joined #salt
08:12 bash124512 joined #salt
08:13 calvinh_ joined #salt
08:13 [LF] joined #salt
08:14 Sacro joined #salt
08:17 gothos joined #salt
08:20 sinh joined #salt
08:21 utahcon joined #salt
08:24 eseyman joined #salt
08:25 Grokzen joined #salt
08:26 fxhp joined #salt
08:26 jtang_ joined #salt
08:28 flyboy joined #salt
08:38 Hybrid joined #salt
08:45 primechuck joined #salt
08:45 mirage123 joined #salt
08:47 karimb joined #salt
08:48 jalaziz joined #salt
08:49 dkrae joined #salt
08:51 Auroch joined #salt
08:57 dRiN joined #salt
08:57 MatthewsFace joined #salt
09:00 jtang_ joined #salt
09:06 N-Mi joined #salt
09:07 [LF] joined #salt
09:07 aquinas joined #salt
09:20 otter768 joined #salt
09:27 jrluis joined #salt
09:27 ocdmw joined #salt
09:28 calvinh joined #salt
09:31 yomilk joined #salt
09:39 alanpearce joined #salt
09:40 paulm- joined #salt
09:46 calvinh_ joined #salt
09:47 fredvd joined #salt
09:48 Madhurranjan joined #salt
09:51 Madhurranjan joined #salt
09:55 CeBe1 joined #salt
09:58 BigBear joined #salt
10:07 bash124512 joined #salt
10:07 Madhurranjan joined #salt
10:08 I3olle joined #salt
10:08 I3olle hey there. does anybody in here know how to create new pillarfiles with python? i want to execute a command and store the output into a pillar file.
10:09 jtang_ any of the debian packagers here?
10:09 jtang_ It looks like https://alioth.debian.org/plugins/scmgit/cgi-bin/gitweb.cgi?p=pkg-salt/salt.git;a=blob;f=debian/salt-api.upstart;h=ff840c9eedc285864ebd7bacae09024c2dc851d9;hb=HEAD is a bit broken
10:10 cberndt joined #salt
10:11 Madhurranjan joined #salt
10:12 TTimo joined #salt
10:13 MaliutaLap jtang_: I'm not involved in the packaging of salt, but I do some packaging - exactly what seems to be up?
10:13 _ether_ joined #salt
10:13 MaliutaLap jtang_: upstart? isn't that really only going to be an issue on Ubunty?
10:14 MaliutaLap s/Ubunty/Ubuntu/
10:16 MaliutaLap jtang_: Sorry, don't know enough about upstart - personally never supported it in any of my packages
10:16 jtang_ yea, it looks like the import of the upstart file is just incorrect, it looks like someone just wget'd the github html page and not the raw content of the config
10:16 Furao I3olle: look at ext_pillars if you need to plug something automatically with pillars without write a pillar .sls first
10:16 jtang_ its a minor fix that ive corrected in my own clone of the packaging
10:17 I3olle Furao: I only find yaml examples regarding the external pillars. i am looking for a way to do this with python
10:17 Furao I3olle: ext_pillars is what you’re looking for
10:18 Furao I wrote one in python to feed from a django apps and inject pillars data straight into the master
10:18 MaliutaLap jtang_: use 'reportbug' to lodge a bug, or email the maintainer - in my my experience if a maintainer gets feedback from someone who has tried and/or fixed the problem they are happy to work with you :)
10:18 jtang_ yea, i just mailed the pkg-salt list
10:19 Furao but you should look also to http://reclass.pantsfullofunix.net/
10:19 Furao I3olle: ^
10:21 Madhurranjan joined #salt
10:23 urtokk joined #salt
10:24 I3olle Furao: so it would be possible for my to do something like this?
10:24 I3olle def ext_pillar( minion_id, pillar, *args, **kwargs ):
10:24 I3olle my_pillar = {}
10:24 I3olle local.cmd('*', 'grains.items', [])
10:24 I3olle return my_pillar
10:25 Madhurranjan joined #salt
10:25 Laogeodritt joined #salt
10:28 keyser joined #salt
10:28 joehh jtang_: yes!
10:29 ocdmw joined #salt
10:29 joehh jtang_: that is really bad!!!! sorry about that
10:30 BigBear joined #salt
10:30 Madhurranjan joined #salt
10:32 lietu joined #salt
10:33 joehh jtang_: you are correct. I just wget'd the file. I'm amazed I didn't pick it up with a git diff... :(
10:33 joehh new packages coming shortly
10:33 lietu I'm trying to use salt on a mac os x machine, managed to get pip installed via homebrew, but salt keeps telling me the pip.installed state is unavailable .. any tips?
10:34 primechuck joined #salt
10:36 jtang_ Joehh, heh, no probs
10:37 jtang_ im just poking at the salt-api for saltpad
10:37 seblu joined #salt
10:37 joehh jtang_: following up on your email - are cherrypy3 and tornado roughly "equivalent"
10:37 joehh ?
10:37 jtang_ I think the wget'ing of a github html page is a pretty common mistake
10:38 joehh I'm not a salt-api user - so a little blind with its packaging
10:38 jtang_ joehh, i'm not familiar with either, but one is blocking and one is non-blocking, salt-api appears to work with both backends, but salt-api appears to default to tornado if nothing is configured
10:39 giantlock joined #salt
10:40 jtang_ joehh, out of the box when salt-api is installed, this is the first error i get
10:40 jtang_ [TRACE   ] Skipping /var/cache/salt/extmods/netapi, it is not a directory
10:40 jtang_ [INFO    ] ImportError! No module named tornado.httpserver
10:40 jtang_ [DEBUG   ] Created pidfile: /var/run/salt-api.pid
10:41 jtang_ so I guess from  a user experience point of view you will probably want to at least suggest python-tornado or even require it
10:41 joehh sounds reasonable - I'm just tossing up between suggests and recommends...
10:41 Madhurranjan joined #salt
10:42 joehh but it looks like I should read the code - that may provide some guidance on the relationships
10:43 bhosmer_ joined #salt
10:43 jtang_ i get the impression cherrypy and tornado arent required for salt-api to 'work' and tornado is preferred as its non-blocking
10:43 JlRd joined #salt
10:44 jtang_ at least that's what the docs seem to imply
10:44 * jtang_ shrugs
10:45 joehh If I knew how to shrug, I would too :)
10:45 Furao I3olle: that can be done in .sls
10:45 joehh My IRC skills are a little out of practice
10:46 Furao I3olle: but you should look at existing ext pillar for example
10:48 I3olle Furao: oh really? that would be great.
10:48 I3olle how do I call the external pillars in an sls?
10:48 lietu joined #salt
10:50 I3olle my approach of ext_pillar.icinga2_pillar['fqdn_ip4', ''] and salt['icinga2_pillar‘](’fqdn_ip4', ‚‘) don’t seem to work
10:50 KermitTheFragger joined #salt
10:52 amcorreia joined #salt
10:56 Furao I3olle: I mean the call to grains that you did, you can do that WITHOUT ext_pillars
10:56 Furao I3olle: and btw pillars and ext_pillars are rendered on the master, if you use salt api to run grains.items you’ll have MASTER grains and not the minion
10:56 marcel joined #salt
10:56 I3olle Furao: I don’t think so. At least not when I want to store the data in a new pillar in my master. the command I use to get the grains of all the minions and store them into one pillar file
10:56 I3olle http://docs.saltstack.com/en/latest/ref/clients/#localclient
10:56 jtang_ heh
10:56 I3olle if however you know how to refer to this data within a sls on my master without an external python script I will gladly take it. I have been biting my teeth out during the past days trying to find a way without python
10:58 Furao oh yes you’re right you’re calling it on ‘*’ but in term of architecture i’m not quite sure you should do that. if a minion isn’t up while you run this the value won’t be there. and sometimes minion aren’t available for a while, I blame zmq.
10:59 Furao you should use salt mine for a “stable” way to store minions data
10:59 jtang_ does the mine clear itself when the master disappears?
10:59 the_lalelu joined #salt
11:00 Furao mine data are stored on master
11:00 jtang_ and also if you are doing multi-master how does the mine data get synced?
11:00 Furao I can’t answer you on this one
11:00 jtang_ I've not started looking at multi-master in my setup's but it's something im also curious about
11:00 moderation joined #salt
11:01 faulkner joined #salt
11:01 jtang_ ansible has a similar problem of getting data from nodes that are 'down'
11:01 I3olle Furao: I have been redirected to the mine several time already and I know about it but I don't like the idea that by storing data in the mine it would be accessible by all minions. Even though I know that not the minions themselves but only the states on them would have acces and even though I can select what to store in the mine. Not being able to restrict access to the mine actually bothers me.
11:02 Furao if you don’t trust minions, don’t use mine
11:02 djinni` joined #salt
11:02 skarn joined #salt
11:02 I3olle Exactly. I don’t trust my minions
11:02 Furao you can use an other way to store data, but ask master for pillars value is already expensive, if you ask all grains from all minions too to render pillars. that will be painful
11:02 Furao in term of time
11:02 matti joined #salt
11:02 matti joined #salt
11:03 georgemarshall joined #salt
11:03 I3olle which other way would you have in mind?
11:03 kossy joined #salt
11:03 snuffychi joined #salt
11:03 calvinh joined #salt
11:03 xMopxShell joined #salt
11:04 paulm- joined #salt
11:05 kossy joined #salt
11:05 teepark joined #salt
11:11 bash1245_ joined #salt
11:12 I3olle Furao: ?
11:12 Xevian joined #salt
11:13 evle1 joined #salt
11:13 TTimo joined #salt
11:13 * Furao back
11:15 Furao I don’t know the details of your requirements. but I would look for an existing data storage (using some api, maybe etcd? if it support such thing) if there is no opensource solution for hold data securely for group of hosts. I would write my own. and create a _modules/mydatastore.py and _states/mydatastore.py and in .sls push data to centralized storage. and get it from consumers
11:16 jtang_ I3olle, i get the impression, if you dont trust your minions, then you should really be targetting them with pillar data
11:16 bhosmer joined #salt
11:17 Furao jtang_ look he want to do that, but he need to share other minions grains
11:17 Furao I3olle: OR add security feature to grains
11:17 Furao I3olle: OR add security feature to MINE
11:18 Furao add ability to salt master to return mine data to callers based on some rules defined in /etc/salt/master
11:19 I3olle how wuld I add secuity features to the mine? It also seems slightly mor ecomplicated to create this centralized storage than having a cronjob to update the external pillar file every few days
11:19 I3olle it’s the first time I hear about adding security features to the mine
11:19 chiui joined #salt
11:20 jtang_ I'm a bit confused, I thought it was possible to group minions such that only those minions with a group can access mine data
11:20 Furao oh maybe they added that feature lately
11:21 jtang_ maybe peer publish is what you are looking for
11:21 Furao http://salt.readthedocs.org/en/latest/topics/mine/index.html?highlight=mine
11:21 otter768 joined #salt
11:21 jtang_ http://docs.saltstack.com/en/latest/ref/peer.html
11:21 Furao problem with peer is that minions need to be up/available at that time
11:21 Furao what if you need this data to monitor hosts?
11:21 Furao if host is down it end not monitored so you down know it’s down :)
11:22 jtang_ you should really be generating that information once off or in an idempotent way to add it to a list
11:22 jtang_ i've been looking into a similar issue with my clients being added to a monitoring system
11:22 jtang_ it seems crazy to continually generate that information dynamically
11:23 I3olle i tried using the publish command once
11:23 I3olle but all i got was this
11:23 I3olle [master]# salt '*'  publish.runner cache.pillar
11:23 I3olle minion:
11:23 I3olle ----------
11:23 I3olle master:
11:23 jtang_ it sounds fun at first till you get into weird situations of hosts appearing/dispearing from your monitoring system
11:23 Furao I had been able to handle that, data isn’t generated dynamically only on changes, such as new hosts or monitoring settings change
11:23 I3olle ----------
11:24 jtang_ i guess if you dont trust your minions to trust each other, then they probably shouldnt be managed by the same salt master
11:24 jtang_ :P
11:24 I3olle Furao: I don’t see security measurements such as grouping on that page you linked
11:25 Furao I3olle: I mean implement such feature yourself in salt mine code!
11:25 jtang_ you could look at returns to capture information in an external data base
11:25 jtang_ then do something with that
11:26 jtang_ returners
11:26 I3olle jtang_: Well, I do trust them a little, but if someone where able to get on the master, of course hell would break loose. But the idea that someone mayyyyybe could just make his way into one of the minions and then access the data of all othe rminions isn’t really comfortable with me
11:27 Furao I3olle: https://github.com/saltstack/salt/blob/develop/salt/daemons/masterapi.py#L508
11:27 jtang_ yea, i see your problem ;) I'm just not sure how one would address it
11:27 I3olle jtang_: returners would leave me with the same question as before: how do I store the output f a python command in a pillar on my master in order to iterate through it within a .sls ?
11:27 jtang_ I had that problem a while back, then I didn;'t and I sleep better now
11:27 jtang_ :)
11:27 Furao i’m not sure that the caller minion id is available at this point
11:28 Furao oh yes here https://github.com/saltstack/salt/blob/develop/salt/master.py#L1033
11:29 jtang_ I3olle, if I were in your position of not really trusting the minions and minions not trusting each other fully, I'd stick with storing all sensitive data in a central location and operate on a push to update type of operating
11:29 I3olle Furao: that one seems to be just a specification of which minions data to collect from the mine but doesn’t seem to restrict the actuall access to the mine. Or am I mistaken?
11:29 Furao so the caller id can be get from load argument and check to some ACLs defined in self.opts and make sure this mine data is allowed for this caller
11:29 jtang_ it wouldnt be ideal though
11:30 Furao I3olle: there is no restriction this is something you can add yourself. create a pull request in github and have it available to everyone
11:31 I3olle jtang_: Well, my plan so far would be storing all the data centralized in one single pillar file on the master. Not ideal, but useful and in my mind at least not too complicated. I just seem to struggle with how to store the data and how to access it. In my mind I just need around 3 or 4 lines of extra code and I am done.
11:31 yomilk joined #salt
11:31 jctrl joined #salt
11:33 jtang_ well im out of ideas to help
11:33 I3olle Furao: I don’t think this is within my capabilities at the moment but I can try and will at least make a feature request in the future
11:33 jtang_ since i'm a relatively new salt user
11:33 ocdmw joined #salt
11:33 jtang_ well im more using salt now than i was a year ago when i was just playing with it
11:33 I3olle jtang_: well thank you for your input nonetheless.
11:34 Furao I3olle: I can do it (i used to contribute to saltstack) but I don’t have the time :( if you want to implement part of it I can help with review or ask some guys in my team to help
11:35 jtang_ I3olle, i think furao mentioned using the salt-api?
11:35 jtang_ there might be some ideas to try out with that
11:35 I3olle Furao: That would be great. I don’t think I will do it right now but I will remember your username and probably come back to you on that offer at some point
11:36 jtang_ there's nothing stopping you from creating user accounts for each minion and have hte api provide the correct data to the accounts calling the api
11:36 Furao jtang_: no salt-api can’t be useful for this, it’s really just an http layer on top of salt master
11:36 jtang_ ah okay
11:37 jtang_ i was thinking more of a light weight app on top of the api that could provide this
11:37 Furao I3olle: i have a salt consulting shop (robotinfra.com) if you want you can outsource this :P
11:38 jtang_ heh, you're based in hong kong ;)
11:38 I3olle so after going through all the other options, such as mine, returners, API, external databases and so on. I still don’t know how to create a custom external pillar on my master and then call it within my .sls?
11:38 jtang_ at least the+852 indicates it
11:38 Furao jtang_: yes :)
11:38 Furao but i’m not chinese
11:39 I3olle Furao: Do you mean outsource the mine security option to you?
11:39 jtang_ heh, well im from hong kong myself ;) but i work for a japanese company
11:40 Furao jtang_: I lived on lamma for a while but i’m now in vietnam
11:40 Furao my team is all based there
11:41 Furao I3olle: yep this is something we can do and have it integrated into saltstack (well if they want)
11:41 jtang_ ah okay
11:41 jtang_ well I live in Ireland ;)
11:41 jtang_ its odd how and where people end up
11:41 Furao you don’t miss hk? especially at that time of the year?
11:42 Furao im from canada (the common wealth, oh no hk isn’t part of it anymore :P)
11:42 jtang_ Yes and no, I prefer living in europe
11:42 jtang_ i might be in SFO by the end of the year
11:42 jtang_ who knows ;)
11:42 Furao jtang_: where in hk youre from?
11:42 jtang_ Yuen Long, to the north west
11:43 I3olle Furao: And about my current problem? I am sure I am just using the wrong command or syntax at the moment. Because I have created the costom pillar but am now unable to call it in the sls
11:43 Furao jtang_: i know that feeling, I left canada 4 years ago and been traveling since them south america, europe and SEA
11:43 jtang_ its just a smallish town in the sub-burbs
11:43 Furao I know yuen long
11:43 Furao my business address is there
11:43 jtang_ its gotten busier over the past few years
11:43 I3olle Btw: from germany here and will be in shanghai in a few months
11:43 jtang_ i guess you are near the west rail
11:43 Furao just a postal box in somekind of school
11:43 jtang_ rent for business is cheap in yuen long, especially since the west rail makes it so easy to get in and out
11:43 Furao yuen long plaza
11:44 jtang_ the big red building with the light rail terminal?
11:44 Furao I3olle: for? :)
11:44 jtang_ well reddish/brown one
11:44 jtang_ shanghai is nuts as well
11:44 Furao I only went there once I don’t remember much
11:45 I3olle Furao I am currently calling it via: ext_pillar.my_pillar.get['kernel', {}]
11:45 jtang_ I'm going to be in Osaka probably at the end of this month for a week or a few depending on work
11:45 Furao yuen long commercial center, kau yuk road
11:45 Furao jtang_: no cny vacation then?
11:45 dvestal joined #salt
11:46 jtang_ nah, I'd love to, but I was in HK last november
11:46 Furao I3olle: “for?” i mean shangai :) for what?
11:46 jtang_ my personal travel budget doesnt stretch that far :(
11:46 hasues joined #salt
11:46 jtang_ I might get to do CNY in Kyoto, which is always cool
11:46 jtang_ the japanese seem to partially celebrate CNY
11:46 hasues left #salt
11:46 Furao i didn’t went to .jp yet
11:46 Furao japanese seem to celebrate everything in their own way
11:47 Furao like those cake for christmas
11:47 I3olle oh that^^ half pleasure and half job. A friend called me there to help him out for a month or two so I will work from there and try to get in contact with some delvelopers for my company while I am thre.
11:47 Hybrid joined #salt
11:47 jtang_ I3olle, yep, thats the best way of travelling ;)
11:48 Furao as vietnam was “owned” by china for a while here cny is big deal
11:48 jtang_ work pays for the flights and I just take a few days before and after work off
11:48 jtang_ :)
11:48 I3olle jtang_: exactly
11:48 Furao just create your own business and put travel on expenses :)
11:48 Furao and better: incorporate in hk for taxes
11:48 jtang_ thats called effort and planning which I hate ;)
11:49 jtang_ incorporating in ireland isnt bad either
11:49 jtang_ its not as good any more though, but still not bad
11:49 I3olle no need to put travel expenses on a fake company. they are already paid for :)
11:52 I3olle anyway. I’ll be off to lunch. Thanks for your help you two and I’ll see you later
11:53 jtang_ heh, he must be in europe some where
11:53 jtang_ cause it's almost lunch time for me as well
11:54 Furao he’s in .de
11:55 bhosmer joined #salt
11:56 jtang_ that's a late lunch then
11:57 Furao better to have it late, 3-4 beers make you sleepy after lunch
11:57 Furao *cliche alert*
11:59 jtang_ heh
12:00 Hybrid joined #salt
12:01 quist`` joined #salt
12:04 Hybrid joined #salt
12:05 danee joined #salt
12:06 yomilk joined #salt
12:07 Hybrid joined #salt
12:08 linjan joined #salt
12:09 Furao joined #salt
12:10 Hybrid joined #salt
12:10 calvinh_ joined #salt
12:11 zadock joined #salt
12:13 jonatas_oliveira joined #salt
12:14 TTimo joined #salt
12:23 primechuck joined #salt
12:26 jerematic joined #salt
12:31 monkey661 joined #salt
12:31 arnoldB Furao: hehe, ask Tom and Dave for their hangover after the FOSDEM/ cfg mgmt camp beer^^
12:32 Furao Dave hangover?
12:32 arnoldB okay at least Thomas^^
12:33 Furao I met him in HK and didn’t wanted a glass!
12:33 Furao more credible :P
12:33 arnoldB Furao: where you from DE?
12:33 I3olle joined #salt
12:34 bhosmer joined #salt
12:34 monkey66 joined #salt
12:35 lietu I'm currently mainly running salt in a local-only setup without a master, I've understood that I should be able to use pillar to provide data if I just put the files in /srv/pillar .. is that so? and in practice, if running a larger setup with a master etc., does it generate the pillar files to /srv/pillar for the minions, or is the data handled differently?
12:35 Furao arnoldB: it’s I3olle who is from germany, i’m from Canada but had been “digital nomad” for 4 years
12:35 cruatta joined #salt
12:35 arnoldB hehe okay :D
12:36 arnoldB I3olle: => #salt-de
12:36 I3olle Hi!
12:36 I3olle Oh
12:36 I3olle interesting
12:36 I3olle thank you
12:37 Furao I have a friend in frankurt and spent some weeks around there, from bamberg to marburg
12:38 ocdmw joined #salt
12:40 monkey66 joined #salt
12:40 Furao Karlsruhe, I didn’t went to the south
12:40 I3olle well, I am from hamburg. Frankfurt to us is just a big airport :) But Marburg is sweet and nice
12:41 Furao I was there for christmas it was more … magic
12:41 arnoldB < from Karlsruhe
12:41 Furao arnoldB: that’s why I said that :P I whois your domain name
12:41 arnoldB Furao: stalker! :P
12:41 Furao so close to france border
12:41 arnoldB yep
12:42 hasues joined #salt
12:43 I3olle hehe
12:43 higgs001 joined #salt
12:44 Furao i hope you have weapons to protect from them!
12:44 Furao hamburg, that the opposite side of .de just went trought it by bus
12:47 TTimo joined #salt
12:49 arnoldB I don't think we need weapons to protect us from the French. having a war in the western EU would be a bad idea at the moment. there are enough wars and problems in the world
12:51 phx lietu, pillars are generated on the master, and the results are made available to the minions
12:53 lnxnut joined #salt
12:53 Furao yeah there is war brewing on east side
12:57 bhosmer joined #salt
12:59 jdowning joined #salt
12:59 favadi joined #salt
13:01 jonatas_oliveira joined #salt
13:03 jonatas_oliveira joined #salt
13:03 pdayton joined #salt
13:06 linjan joined #salt
13:08 ocdmw joined #salt
13:08 jtang_ karlsruhe is nice
13:09 jtang_ ive friends in around that area
13:13 aparsons joined #salt
13:13 GabLeRoux joined #salt
13:14 tkharju joined #salt
13:14 paha joined #salt
13:15 tkharju joined #salt
13:16 redzaku joined #salt
13:16 hobakill joined #salt
13:17 jtang_ joehh, it seems https://github.com/saltstack-formulas/salt-formula/blob/7ddc1692ecb17ae0c77e4b07c48a1baef3d36fdc/pillar.example
13:17 popus joined #salt
13:17 jtang_ rest_tornado is at least suggested to be used in the salt-formula, may as well stick with that
13:17 giantlock joined #salt
13:17 jtang_ at least as the recommend anyway
13:18 jerematic joined #salt
13:19 Morbus joined #salt
13:22 otter768 joined #salt
13:22 jonasbjork joined #salt
13:27 Furao sometimes I feel that the minions are like are an diesel motor in winter and it need some test.ping to reply faster
13:27 arnoldB Furao: +1
13:28 favadi left #salt
13:29 sastorsl joined #salt
13:29 sastorsl hi
13:29 sastorsl Is it possible to just render / display a state.sls ?
13:30 arnoldB sastorsl: state.show_sls ?
13:30 Furao or http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.state.html#salt.modules.state.show_highstate
13:30 hobakill arnoldB, Furao agreed.
13:31 brayn joined #salt
13:31 Furao 2 confirmations, that mean salt is really running on diesel
13:32 sastorsl arnoldB, something like that, yes. Thx
13:33 sastorsl Any other tips on debugging salt states?
13:33 Furao sastorsl: add test=True
13:33 Furao run minion AND master with log level all
13:34 hobakill needs new glowplugs
13:34 Furao import pydev debugger into salt process and use pycharm to remotely run salt process line by line
13:34 kellnola sastorsl,   I find running the state from a minion with verbosity is more helpful in debugging that running the state from the master
13:34 kellnola with salt-call that is
13:35 mikkn joined #salt
13:36 sastorsl Furao: Yes, I've used test=true
13:38 Rizla joined #salt
13:40 Rizla To distribute ssh keys, which is the preferred way of doing it? From a pillar or salt://? I found some confusing answers on google.
13:40 bhosmer joined #salt
13:42 bash124512 joined #salt
13:42 paulm- joined #salt
13:42 SneakyPhil I'm experiencing an issue trying to start a docker container (https://github.com/UtahDave/dock_apache) with salt. I am running salt 2014.7.0 and docker-py 0.7.2 but the error I get says that docker.XXXX is unavailable.
13:47 ughloru joined #salt
13:47 TTimo joined #salt
13:49 sastorsl So, what I'm trying to debug is: a state I'm taking over responsibility is referencing a default.yaml, but the path of the default.yaml does not exist, but the state works. I can't for the life of me figure out why :-)
13:49 sastorsl *responsibility for*
13:49 JDiPierro joined #salt
13:52 JDiPierro joined #salt
13:53 JDiPierro joined #salt
13:53 Guest15 joined #salt
13:56 wincus joined #salt
13:57 TTimo joined #salt
14:07 beardo joined #salt
14:10 ocdmw joined #salt
14:12 bhosmer joined #salt
14:13 nitti joined #salt
14:13 racooper joined #salt
14:14 BigBear joined #salt
14:17 lnxnut joined #salt
14:20 elfixit joined #salt
14:20 wincus joined #salt
14:21 msheiny joined #salt
14:21 wincus joined #salt
14:26 rypeck joined #salt
14:32 cpowell joined #salt
14:35 primechuck joined #salt
14:36 karimb joined #salt
14:38 dvestal joined #salt
14:40 hasues left #salt
14:40 mindscratch joined #salt
14:41 msheiny joined #salt
14:42 mindscratch i'd like to have a default set of states that will make sure a variety of services are running (in a certain order)
14:43 mindscratch ...but I'd like the inverse (stop the services in the correct order)
14:43 mindscratch any suggestion for how to do that?
14:43 Deevolution joined #salt
14:43 beardo joined #salt
14:43 s8weber joined #salt
14:43 mpanetta joined #salt
14:45 monkey66 left #salt
14:47 wincus joined #salt
14:48 mgw joined #salt
14:49 hobakill not that i can't install from epel-testing .... but i'm super surprised that 2014.7.1 isn't out of the testing phase yet. i'd like to be proactive and help get it out of there an into stable but i don't know where to start. thoughts>
14:50 Brew joined #salt
14:51 primechuck joined #salt
14:51 arnoldB hobakill: mailing list
14:51 hobakill salt? epel?
14:51 arnoldB salt
14:52 murrdoc joined #salt
14:52 sastorsl gitfs question: This file exists, but I get "False": salt 'minion' salt://basedir/subdir/defaults.yaml
14:52 sastorsl gitfs question: This file exists, but I get "False": salt 'minion' file.file_exists salt://basedir/subdir/defaults.yaml
14:52 sastorsl Any thoughts?
14:53 Furao check your branch/env
14:53 hobakill thanks arnoldB
14:55 mgw joined #salt
14:55 ale joined #salt
14:58 lnxnut joined #salt
14:58 higgs001 joined #salt
15:00 abe_music joined #salt
15:00 timoguin joined #salt
15:01 lnxnut_ joined #salt
15:06 ekle joined #salt
15:07 ekle hi, is there a way to download a file from s3 wihchout hash ?
15:09 Furao ekle: cmd.run: -name: s3cmd s3://bucket/key :P
15:10 felskrone joined #salt
15:10 murrdoc :D
15:11 Furao s3cmd get
15:12 Furao OR module.run http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.s3.html#salt.modules.s3.get
15:12 Furao it don’t seem to requires the hash
15:13 dude051 joined #salt
15:15 higgs001 joined #salt
15:17 ekle joined #salt
15:18 ekle how can i get the update state to s3cmd ?
15:18 jonatas_oliveira joined #salt
15:18 ekle ist there a way that i can get the return code of cmd.run in salt ?
15:18 Grokzen If i run salt with --batch=50, will it start to run 50 machines and wait for all of them to finish or will it start another machine as soon as 1 of those 50 is done?
15:19 Grokzen *to finish before starting the next batch of 50 machines
15:22 timoguin_ joined #salt
15:22 Furao ekle: “cmd.run: -name: s3cmd get s3://bucket/key” OR module.run http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.s3.html#salt.modules.s3.get
15:23 bash124512 joined #salt
15:23 otter768 joined #salt
15:23 ekle how to use the s3 module ? i always get State 's3.get' found in SLS
15:23 BobL joined #salt
15:24 ekle i tried this somename: s3-get: - source: "s3://...."
15:24 TheoSLC joined #salt
15:24 Furao ekle: use module.run: -name: s3.get
15:25 Madhurranjan joined #salt
15:25 Furao ekle: a bit like https://github.com/bclermont/states/blob/master/states/carbon/init.sls#L57
15:25 Furao which don’t use pip.installed state but rather call module directly
15:27 anotherZero joined #salt
15:30 mindscratch joined #salt
15:30 mindscratch Once i've created state files and a top.sls
15:30 mindscratch how do I run it?
15:31 yawniek_ joined #salt
15:33 mindscratch I guess salt '*' state.highstate
15:33 jmccree that will run it for all minions
15:33 jmccree if you want to run it on a local minion, I use: salt-call -l debug state.highstate test=True
15:33 jmccree first
15:34 JlRd joined #salt
15:34 Tahm joined #salt
15:36 ekle s3.get always whatns to connection to some host 169.254.169.254 whoch does not exist
15:37 ekle ConnectionError(ProtocolError('Connection aborted.', error(113, 'No route to host')),)
15:37 Furao ekle: that is the ip of grains amazon ec2, no?
15:38 mindscratch jmccree: ok
15:38 Furao https://github.com/bclermont/states/blob/master/states/_grains/ec2_info.py
15:38 mindscratch how do I make sure one service starts before another?
15:38 mindscratch like, minion1 should start zookeeper, minion 2 should start mesos (only if zookeeper started)
15:39 Furao MindDrive: salt oerstate
15:39 Furao overstate
15:39 Furao oups that was for mindscratch
15:39 ekle @Furao isn't this some local ip
15:40 mindscratch Furao: ok, i'll google it
15:40 babilen orchestrate surely or http://docs.saltstack.com/en/latest/ref/states/highstate.html#requisite-declaration
15:40 babilen mindscratch: Are both running on the same minion?
15:40 Furao ekle: it’s an amazon IP
15:40 mindscratch babilen: I see overstate is being deprecated
15:40 Furao http://serverfault.com/questions/427018/what-is-this-ip-address-169-254-169-254
15:40 babilen mindscratch: Yes, that is why I said "orchestrate surely"
15:40 mindscratch no
15:41 babilen mindscratch: http://docs.saltstack.com/en/latest/topics/tutorials/states_pt5.html#orchestrate-runner
15:41 Furao overstate deprecated?
15:41 viq joined #salt
15:41 viq joined #salt
15:41 mindscratch babilen: sometimes yes, for example zookeeper and mesos migth be on the same minon, but a process on another minion requires zookeeper as well
15:41 mindscratch Furao: http://docs.saltstack.com/en/latest/topics/tutorials/states_pt5.html#states-overstate
15:42 Furao hmm and replaced by?
15:42 babilen orchestrate
15:42 kaptk2 joined #salt
15:42 Furao ah ok thanks
15:43 babilen Which is why I said "orchestrate surely" and linked to http://docs.saltstack.com/en/latest/topics/tutorials/states_pt5.html#orchestrate-runner
15:43 Furao both words are too similar, i didn’t noticed
15:43 Furao dyslexia i guess
15:44 mindscratch babilen: do you know how I'd target a minion based on a grain using salt.state
15:44 mindscratch looking at this: http://docs.saltstack.com/en/latest/topics/tutorials/states_pt5.html#configuration-syntax
15:44 ocdmw joined #salt
15:46 ekle does s3.get maybe not support eu-central ?
15:46 Furao s3 probably use boto which support eu-central
15:47 babilen mindscratch: http://docs.saltstack.com/en/latest/topics/targeting/grains.html
15:48 ekle i don't get it, why is s3.get tring to access 169.254.169.254 ? its a link-local ip
15:48 davet joined #salt
15:49 mindscratch babilen: I had read that, I guess in the orchestration file I was wondering how to target them, this is what I came up with, which seems to work http://pastebin.com/400j31nX
15:49 babilen Please make the web a better place and use a better pastebin such as http://refheap.com or http://paste.debian.net
15:50 babilen mindscratch: I really don't like grains for roles and I'm not sure if it works if "roles" is a list
15:50 malinoff joined #salt
15:50 mindscratch babilen: what do you prefer?
15:50 ocdmw joined #salt
15:50 keyser joined #salt
15:51 arnoldB or https://np.adm.sx/ :P
15:51 mindscratch babilen: I'm trying to avoid using hostname matching
15:51 mapu joined #salt
15:51 babilen mindscratch: I prefer a sane naming policy and relying on the minion id. States and, in particular, pillars really shouldn't be targeted based on data provided by the minion itself. This holds in particular true for sensitive data in pillars.
15:52 mindscratch babilen: in a large system I don't want to deal with all those names and figuring out what maps to what. in my case i was having puppet install a grains file with "roles", I could change grains to have something like: web: true, zookeeper: true  instead of "roels"
15:52 ekle looks like s3.get does not support anonymous connections
15:52 babilen mindscratch: I totally understand, just don't target sensitive data based on data provided by the minion itself
15:52 mindscratch babilen: makes sense
15:52 dkrae joined #salt
15:53 mindscratch babilen: is the orchestration sls file processed in order top-to-bottom?
15:53 babilen I'm sorry that I don't have a proper alternative (i would say "pillars" but they can't be used everywhere like grains)
15:53 babilen mindscratch: It is
15:53 Nazca joined #salt
15:54 atree joined #salt
15:56 ekle @Furao: looks like it does not support eu-central-1: <Error><Code>InvalidRequest</Code><Message>The authorization mechanism you have provided is not supported. Please use AWS4-HMAC-SHA256.</Message><RequestId>63E69CC2BF1B7031</RequestId><HostId>qIHOLhWKA9IKkqCzkpWbAwZgYK7BkYiFr+7QuEMOC7K8q7U6R9c49SSrszHaUf5BhaAReZEzQHw=</HostId></Error>
15:58 khaije1 joined #salt
15:58 mindscratch babilen: is it possible to setup halite without certs?
15:59 Furao ekle: that is weird, upgrade your boto
15:59 Nazca__ joined #salt
15:59 Furao ekle: https://bugs.launchpad.net/duplicity/+bug/1407966
15:59 mindscratch i've set tls: False, but when  I bring up the webpage the username/password fields are highlighted red, and the list of minons doesn't appear
15:59 Furao This is a boto issue, not a duplicity issue. You will need to upgrade to at least boto 2.36.0.
16:00 Furao well… ignore the duplicity part
16:00 ekle havent installed boto at all ...is it included in salt ?
16:00 wincus joined #salt
16:01 Furao ekle: based on your distro and pkg it might be installed automatically for you
16:01 ekle debian jessie: dpkg --get-selections | grep -i boto => finds nothing
16:02 Furao you installed with .deb ? or pip isntall salt
16:02 ekle deb from http://debian.saltstack.com/
16:02 ekle isn't this the prefered way of installing salt on debina ?
16:02 Furao https://github.com/saltstack/salt/blob/2014.7/salt/utils/s3.py
16:03 clintberry joined #salt
16:03 chris-m joined #salt
16:03 jrluis joined #salt
16:03 Furao ahaha they didn’t used boto
16:03 cpowell joined #salt
16:03 elfixit joined #salt
16:03 Furao I don’t know what to say, sorry
16:03 ajw0100 joined #salt
16:03 Furao open a new issue in github
16:05 chris-m happy monday all
16:05 chris-m anyone seen this?  and know the correct resolution (following the instructions in the critical status didn't help)
16:05 chris-m [CRITICAL] The Salt Master has rejected this minion's public key! To repair this issue, delete the public key for this minion on the Salt Master and restart this minion. Or restart the Salt Master in open mode to clean out the keys. The Salt Minion will now exit.
16:06 Furao master key had been regenerated or you switch from one master to an other
16:06 Madhurranjan joined #salt
16:07 ekle @Furao: https://github.com/saltstack/salt/issues/20518 and thanks for the help
16:08 wincus joined #salt
16:08 stej joined #salt
16:08 ekle is there a way to http download a file without extra hash-file, but with "on change" events working ?
16:09 jalbretsen joined #salt
16:09 ecdhe joined #salt
16:09 stej left #salt
16:10 gladiatr joined #salt
16:10 chris-m Furao - is that answer for my question?   If so, how do I do this?
16:11 stej joined #salt
16:13 smcquay joined #salt
16:13 _prime_ joined #salt
16:13 funzo joined #salt
16:14 hasues joined #salt
16:14 hasues left #salt
16:14 Striki joined #salt
16:16 Furao chris-m: if i remember right you have to rm -f /etc/salt/pki/minion/minion_master.pub and restart minion
16:16 Furao or stop it rm and start it
16:16 chris-m will give that a try. thxnk Furao
16:17 yawniek_ anyone uses salt cloud with archlinux, bootstraping fails with 'ERROR: No dependencies installation function found. Exiting...' ?
16:17 Striki any idea why "testrailer-requirements" is not being executed on first highstate? This file is in the dev environment and is including and depending (via watch) on states from the base environment. Might that be the issue? All the included states are being executed, just not "testrailer-requirements" https://gist.github.com/kristinn/e3f464fc068f346d68fb
16:18 tpaul joined #salt
16:19 tpaul How can I see what minions haven't had their highstate run recently?
16:19 chris-m @Furao - I ran this on the minion, but still got the same error message
16:20 pravka joined #salt
16:22 andrew_v joined #salt
16:23 stej joined #salt
16:23 rihannon joined #salt
16:24 tpaul I thought, there was a last.highstate grain or something...
16:27 mindscratch Anyone familiar with setting up Halite?
16:32 jonasbjo1k joined #salt
16:32 jtang_ joined #salt
16:33 tpaul salt '*' cmd.run 'date -r /var/cache/salt/minion/highstate.p' "seems" to work for getting the last highstate, not sure how accurate that is but I'm going with it
16:35 murrdoc might be better to setup a returner
16:35 murrdoc and store the status in a db
16:36 tpaul Is that how Halite does it? Actually, does halite keep track of the last highstate? I just assumed it would but haven't used it.
16:36 SpX joined #salt
16:36 murrdoc salt-pad has an option to it
16:37 murrdoc not sure if halite keeps the data local
16:37 murrdoc it doesnt have to be that involved
16:37 murrdoc keep a redis instance on the master and use the returner to put the job status into redis
16:38 evilrob joined #salt
16:38 mindscratch When bringing up Halite, how do I login?
16:40 murrdoc http://docs.saltstack.com/en/latest/topics/tutorials/halite.html#configuring-halite-permissions
16:40 evilrob so with work schedules... there's no way I'm getting to saltconf next month :(
16:40 evilrob are there other events (perhaps in the fall) that are similar?
16:41 tligda joined #salt
16:42 mindscratch murrdoc: is there documentation for the pam module?
16:42 murrdoc pam is linux pam ?
16:42 mindscratch murrdoc: Sorry, I meant as far as the yaml config
16:42 mindscratch I just found a page on salt external auth
16:43 murrdoc cool
16:43 rihannon joined #salt
16:43 chris-m re-asking: [CRITICAL] The Salt Master has rejected this minion's public key! To repair this issue, delete the public key for this minion on the Salt Master and restart this minion. Or restart the Salt Master in open mode to clean out the keys. The Salt Minion will now exit
16:44 StDiluted joined #salt
16:44 jab416171 joined #salt
16:44 murrdoc go with 1
16:44 murrdoc salt-key -d <minion-name>
16:46 chris-m thx Murrdoc   this got me past the CRITICAL warning, now back to [ERROR   ] The Salt Master has cached the public key for this node, this salt minion will wait for 10 seconds before attempting to re-authenticate [INFO    ] Waiting for minion key to be accepted by the master. [INFO    ] Waiting 10 seconds before retry.
16:46 rihannon1 joined #salt
16:47 murrdoc salt-key -A
16:47 murrdoc or salt-key -a <minion-name>
16:47 murrdoc http://docs.saltstack.com/en/latest/ref/cli/salt-key.html
16:47 tligda1 joined #salt
16:48 chris-m thx again murrdoc.  will look into this
16:49 giantlock joined #salt
16:50 Ozack-work joined #salt
16:51 chris-m have a great day
16:52 linjan joined #salt
16:53 rihannon joined #salt
16:53 murrdoc you too man
16:53 murrdoc i mean i hope i wasnt rude
16:53 spiette joined #salt
16:55 mgw i'm having trouble getting module.run and cmd.wait to play nicely
16:55 mgw I need to do a "module.run git.submodule" and have that trigger a "cmd.wait" on changes
16:56 hal58th joined #salt
16:57 pdayton joined #salt
16:57 higgs001 joined #salt
16:58 mindscratch murrdoc: thanks for the help
16:59 wiqd joined #salt
16:59 wiqd joined #salt
17:00 rihannon joined #salt
17:01 murrdoc ./me is having an iggy kind of day
17:02 iggy so you've been drinking at Mardi Gras for 4 days straight and you can barely stand?
17:02 murrdoc nice
17:02 murrdoc did u head out to the orleans ?
17:03 iggy yeah. My mom lives in the French Quarter... so it's a short walk to a lot of the stuff
17:03 murrdoc sawweeeet
17:03 murrdoc is it true that anyone can have a parade in new orleans ?
17:05 iggy nope
17:05 iggy it costs lots of money
17:06 murrdoc but u can, if u have the money
17:06 yawniek_ can i somehow define which bootstrap skript is used for salt-cloud. shomehow i get a very old deploy.sh that fails on archlinux by default
17:06 iggy they always have extra cops, street cleaners afterward, etc.
17:06 iggy well, yeah, but you can hunt humans if you have enough money
17:06 troyready joined #salt
17:07 murrdoc thats that dick cheney money
17:10 theologian joined #salt
17:12 cmek hi. I have salt-master (2014.7.1 on fedora12) and 2 minions (2014.7.0 both on centos 7). salt '*' test.ping takes over a minute in this setup... all configs on master/minions with default settings.
17:12 cmek any idea why is it so slow? states report duration times in miliseconds but the whole run still takes minutes
17:13 cmek haven't found anything in issues on github that would explain this
17:13 davet joined #salt
17:13 linjan joined #salt
17:14 cmek er. that fedora master is 21 not 12
17:15 iggy cmek: do the minions return pretty quick and then it just sits there?
17:15 seanz joined #salt
17:15 cmek nope. I run salt minion state.highstate, and wait minutes for the command to return
17:15 badon joined #salt
17:16 atree joined #salt
17:16 cmek the same with ping or any other salt command
17:17 hal58th when you run salt-call state.highstate, how long does that take to run?
17:17 I3olle joined #salt
17:18 cmek hal58th: that's really quick. a couple of seconds
17:18 hal58th hmmm not sure then
17:19 iggy see if you have any old/dead keys in salt-key -L
17:20 cmek that's another thing, even salt-key takes quite long
17:20 cmek hal58th: the difference is 7s on minion vs 1m15s when run from master
17:20 cmek iggy: no dead keys
17:24 otter768 joined #salt
17:24 jonatas_oliveira joined #salt
17:26 jopecko joined #salt
17:28 mindscratch joined #salt
17:30 mindscratch I am using orchestration. I have a service start, then I want to wait a few seconds before starting the next service, is there some way to "sleep" or "pause" aftre starting the first service?
17:31 eliasp mindscratch: I don't think so, as "sleeping" is never a real fix and will just lead to non-deterministic behaviour… isn't there a proper way to check the state of your environment before continuing?
17:32 mapu joined #salt
17:33 MatthewsFace joined #salt
17:33 mindscratch eliasp: understood. I suppose I might be able to query an HTTP endpoint. In that case, in the SLS file would I had a cmd like, "curl http://...", and have the 2nd process "require" the curl?
17:34 eliasp yes, something like this sounds better…
17:35 mindscratch ok i'll look into it, thanks
17:35 hobakill are people using the sorcery that is Fedora COPR?   https://copr.fedoraproject.org/coprs/saltstack/salt/
17:35 clintber_ joined #salt
17:35 iggy or somehow fire an event on the salt bus when the service is actually up (then that event would kick off the next step)
17:40 chiui joined #salt
17:42 bhosmer_ joined #salt
17:43 KyleG joined #salt
17:43 KyleG joined #salt
17:43 kbyrne joined #salt
17:47 Ryan_Lane joined #salt
17:47 quist``` joined #salt
17:51 wendall911 joined #salt
17:53 Madhurranjan joined #salt
17:54 Grokzen joined #salt
17:55 CeBe1 left #salt
17:58 desposo joined #salt
17:58 chrisforrette joined #salt
17:59 chrisforrette hey folks, i have the latest version of Salt installed and im trying to generate a htpasswd file on the fly using this: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.htpasswd.html#module-salt.states.htpasswd but i keep getting an error that the “webutil.user_exists” state is unavailable—anyone know why that might be happening?
17:59 bhosmer_ joined #salt
18:00 eligos joined #salt
18:00 murrdoc joined #salt
18:03 bash124512 joined #salt
18:04 Madhurranjan joined #salt
18:04 teskew joined #salt
18:08 sjol_ joined #salt
18:09 iggy I guess he figured out that he needs htpasswd actually installed to use the module
18:10 aparsons joined #salt
18:10 motorrobot joined #salt
18:12 hebz0rl joined #salt
18:13 motorrobot hi everyone, I was hoping someone might have a quick answer for me on this... If I have a virtualenv.managed with a requirements.txt, one of the pip modules requires /usr/custom/path appended to $PATH (which automatically is added as a path for bash when logged in). It works properly when using salt-call locally but not when executing from salt master.... i've tried appending runas=root and shell=/bin/bash to my salt command
18:13 motorrobot salt 'app*' state.sls webapp.django.site # is how I have it laid out
18:17 eliasp meh, why does `pkg.refresh_db` fail to update the repo-data on my win minions…?? let's see
18:17 JDiPierro joined #salt
18:19 forrest joined #salt
18:22 mpanetta_ joined #salt
18:23 FRANK_T joined #salt
18:24 dusel joined #salt
18:25 timoguin joined #salt
18:27 Godfath3r joined #salt
18:27 Mso150 joined #salt
18:29 Brick_ joined #salt
18:30 mpanetta_ joined #salt
18:31 schlueter joined #salt
18:31 jalaziz joined #salt
18:32 echtish joined #salt
18:32 TTimo joined #salt
18:34 jalaziz joined #salt
18:36 linjan joined #salt
18:39 iwishiwerearobot joined #salt
18:40 mindscratch joined #salt
18:40 meteorfox joined #salt
18:41 mindscratch I need to wait for a web service to return a certain status before POSTing data to it. I can use "cmd.run" to issue a cURL to the URL to check if it is ready, but how can i do that untl it is ready?
18:41 mindscratch just do something like: while true; do curl ....; sleep 5; done...?
18:41 aparsons joined #salt
18:42 SheetiS joined #salt
18:42 ajw0100 joined #salt
18:43 smithd joined #salt
18:43 jmccree mindscratch, consider running a script in the language of your choice.
18:44 jmccree http://docs.saltstack.com/en/latest/ref/states/all/salt.states.cmd.html#salt.states.cmd.script
18:46 smithd joined #salt
18:48 mindscratch jmccree: ok so to confirm, the recommended approach is to create a script that I can run which will poll a service for status until it's "ready". I can then have another state that watches the "check status" command, when "ready" it'll do the POSTs.
18:50 slafs joined #salt
18:50 ipmb joined #salt
18:50 jmccree mindscratch, I'm not really making a recommendation, but if it was me, I'd think of just making one script to poll the server then do the POSTs.
18:50 chrisforrette joined #salt
18:51 teebes joined #salt
18:51 mindscratch jmccree: gotcha
18:51 wincus joined #salt
18:51 slafs left #salt
18:52 ipmb so... how do I remove something from the salt-mine for a minion that no longer exists?
18:52 chrisforrette @iggy: thanks :) I didn’t realize htpasswd wasn’t standard
18:53 spookah joined #salt
18:53 BigBear joined #salt
18:55 ipmb looks like the answer is, "manually" https://github.com/saltstack/salt/issues/11389
18:56 eligos joined #salt
19:02 Madhurranjan joined #salt
19:03 ckao joined #salt
19:05 Mso150 joined #salt
19:07 TheoSLC joined #salt
19:09 subsignal joined #salt
19:11 eliasp another fun bug I'll have to figure out: the winrepo.genrepo runner writes the master-cache-file to /srv/salt/win/repo/winrepo.p, the Minions expect it at /media/volumes/salt/master/fileserver/win/repo/winrepo.p (which is my configured path) and therefore fail to pkg.refresh_db …
19:11 eliasp /srv/salt/ is configured nowhere in my master config… so for some reason it defaults to it and uses it in the winrepo.genrepo runner
19:12 sjol_ joined #salt
19:12 murrdoc file_roots doesnt get read in ?
19:14 jonatas_oliveira joined #salt
19:15 eliasp lemme check
19:15 eliasp file_roots → base → /media/volumes/salt/master/fileserver/
19:15 eliasp looks good (according to pillars on master)
19:17 murrdoc then the code is buggy, should read that in
19:18 eliasp hmm, that's weird… added some debug output to get the raw values read from the config in winrepo.genrepo: https://gist.github.com/eliasp/c9f4f24cf628ed68f066
19:19 eliasp repo + winrepo are os.path.join-ed later in L~67
19:19 eliasp and the result is obviously /srv/salt/win/repo/winrepo.p
19:19 eliasp so joining two absolute paths is complete nonsense…
19:19 eliasp I'll try to make sense of this + come up with a possible fix
19:20 murrdoc sounds like fun
19:21 eliasp … as always :)
19:21 eliasp is there a more direct CLI way to query 'opts' than via pillar?
19:22 murrdoc config.get ?
19:22 murrdoc sys.doc config
19:22 eliasp ah, great… thanks ;)
19:22 murrdoc might explain it better
19:22 druonysuse joined #salt
19:23 Ryan_Lane joined #salt
19:24 otter768 joined #salt
19:25 eliasp ok, config.get confirms this… win_repo_mastercachefile defaults to an absolute path…
19:25 eliasp so it should either default to a filename only which could then be .join-ed to the win_repo path or the logic for handling these two paths needs to be different (write to win_repo by default, use win_repo_mastercachefile only when explicitely defined)
19:27 eliasp salt/config.py says:
19:27 eliasp 'win_repo_mastercachefile': os.path.join(salt.syspaths.BASE_FILE_ROOTS_DIR,
19:27 eliasp 'win', 'repo', 'winrepo.p'),
19:27 linjan joined #salt
19:27 murrdoc uh hmm
19:30 jalaziz joined #salt
19:35 jtang_ joined #salt
19:35 g3cko joined #salt
19:35 spiette joined #salt
19:37 eliasp don't really have time right now to completely wrap my head around it… will symlink/work around it for now here and file an issue…
19:39 Tyrm joined #salt
19:41 rap424 joined #salt
19:41 jonatas_oliveira joined #salt
19:42 ajw0100 joined #salt
19:46 thedodd joined #salt
19:48 jonasbjork joined #salt
19:51 TheoSLC joined #salt
19:51 eliasp murrdoc: https://github.com/saltstack/salt/issues/20526
19:52 murrdoc there isnt a 'vote' button otheriwse i would
19:52 eliasp yeah, mostly FYI and to subscribe in case you want to track it ;)
19:56 eligos joined #salt
19:58 Grokzen joined #salt
19:59 Tyrm joined #salt
19:59 Ox joined #salt
20:00 SheetiS joined #salt
20:00 jerematic joined #salt
20:01 hackenschmidt joined #salt
20:03 notnotpeter Hey folks, does anyone know how to get the current salt environment from within jinja?
20:03 atree joined #salt
20:03 jtang_ joined #salt
20:04 hackenschmidt I'm new to salt. I've used chef in the past and the one thing i really like about it is how all node/client data is searchable and useable. Does salt have this function? or is salt-mine, where you have to define everything you want to cache, the only thing available?
20:05 ocdmw joined #salt
20:09 Morbus joined #salt
20:11 eliasp notnotpeter: {{ saltenv }}
20:12 armguy joined #salt
20:12 iggy notnotpeter: {{ saltenv }} ?
20:12 iggy damn, was scrolled up
20:15 eliasp ;)
21:06 ilbot3 joined #salt
21:06 Topic for #salt is now Welcome to #salt | SaltConf 2015 is Mar 3-5! http://saltconf.com | 2014.7.1 is the latest | Please be patient when asking questions as we are volunteers and may not have immediate answers | Channel logs are available at http://irclog.perlgeek.de/salt/
21:08 conan_the_destro joined #salt
21:09 aparsons joined #salt
21:09 Tyrm joined #salt
21:09 jjasinski gix0, forrest, notnopeter: cmd.run saved me.  (Normally, the states work, but in this particular case the fact that a user existed before I ran state.highstate caused the ssh keys to get mixed up)  Thanks for your help!
21:10 TTimo joined #salt
21:12 jtang_ joined #salt
21:13 eligos joined #salt
21:16 aparsons_ joined #salt
21:16 Mso150 joined #salt
21:18 jtang_ joined #salt
21:21 neogenix joined #salt
21:21 prwilson joined #salt
21:25 rvankleeck joined #salt
21:25 kim0 joined #salt
21:25 otter768 joined #salt
21:25 rvankleeck if I use test=True when running "salt 'm*' pkg.upgrade test=True -l debug -v" it goes ahead and does the upgrade on the minion
21:25 kim0 Hi .. I'm trying to hack a module .. it says __grains__ is not defined .. what do I need to do to define it ?
21:26 rvankleeck is that normal behavior, or a bug?
21:26 KyleG joined #salt
21:26 KyleG joined #salt
21:27 murrdoc kim0:  dont use __grains__
21:27 kim0 oh why
21:27 murrdoc use __salt['grains.get']()
21:28 iggy __salt__
21:28 manfred __grains__ should exist in the modules…
21:28 kim0 status module is full of __grains__ :)
21:28 murrdoc they should but, http://docs.saltstack.com/en/latest/topics/development/dunder_dictionaries.html, they are moving away from grains
21:28 gix0 good to hear it worked out well!
21:28 kim0 ok thanks folks
21:28 murrdoc its recommended to use the dunders listed on that page
21:29 manfred grains is on that list
21:29 manfred the bigger problem would be that __grains__ isn't defined, which is should be cause it is injected by the loader
21:29 gladiatr murrdoc, link is 404
21:29 murrdoc http://docs.saltstack.com/en/latest/topics/development/dunder_dictionaries.html
21:30 manfred https://github.com/saltstack/salt/blob/develop/salt/loader.py#L833
21:30 manfred where on that page does it say not to use grains?
21:30 aurynn there's a page about not using grains?
21:30 manfred Consider using get functions whenever using __opts__ or __pillar__ and __grains__ (when using grains for configuration data)
21:31 manfred all it says is to consider using get instead of directly calling __grains__
21:31 gladiatr __grains__.get('somekey', None)
21:31 murrdoc u right manfred
21:31 manfred it means use __salt__['grains.get']('something:something:something')
21:31 kim0 Ok now it says __salt__ is not defined
21:31 manfred but you don't have too :P
21:31 manfred see, that is the real problem
21:31 iggy then you are definitely doing something wrong
21:31 manfred kim0:  is this a custom module?
21:31 kim0 manfred: zfs module
21:32 gladiatr kim0, are you trying to debug your module sic interactive python?
21:32 aurynn salt["grains.get"]()
21:32 aurynn from in a pillar
21:32 manfred kim0:  where in the zfs module?
21:32 eligos joined #salt
21:32 kim0 salt/modules/zfs.py .. I'm adding new lines and want to access grains
21:33 iggy how are you testing?
21:33 manfred where are you adding the new lines? can you gist a git diff?
21:33 kim0 in _check_zfs
21:33 manfred how are you testing?
21:34 kim0 salt-call --local status.w
21:34 manfred oh
21:34 kim0 which is unrelated .. but the explosion happens
21:34 manfred i think your problem is that the __salt__ is being called by __virtual__
21:34 manfred one second
21:34 manfred which is before __grains__ and __salt__ are injected
21:35 manfred you will need to use that utils thingy… one second
21:35 murrdoc import salt.utils
21:35 kim0 okie thanks
21:35 manfred hrm, actually
21:35 manfred https://github.com/saltstack/salt/blob/develop/salt/modules/apache.py#L52
21:35 manfred it works in the apache module
21:35 smithd joined #salt
21:36 manfred maybe try and put that function below the __virtual__ in the zfs module? it might just be a placement /loading thing...
21:36 murrdoc import salt.utils
21:36 manfred it already is imported
21:36 murrdoc ok
21:36 murrdoc i give up
21:36 manfred https://github.com/saltstack/salt/blob/develop/salt/modules/zfs.py#L18
21:36 kim0 that's exactly what I wanna do .. that apache thing
21:36 theologian joined #salt
21:36 murrdoc there isnt a way to get the 'currently running job' is there
21:36 manfred salt-run jobs.active
21:37 manfred gives you all the running jobs
21:37 kim0 manfred: so I should still try putting __virtual__ on top right ?
21:37 forrest joined #salt
21:37 linjan joined #salt
21:37 manfred yes, __virtual__ i think should be the first function… i don't know if that will fix the problem, but it might, and it looks like that stuff currently works in the apache module.
21:38 clintber_ joined #salt
21:39 murrdoc are u trying to check a grain ?
21:39 manfred yes, from a function called from __virtual__
21:39 iggy I'd key off the binary name instead of the os grain
21:39 moos3 joined #salt
21:40 gladiatr kim0, it is conventionally at the top.  It doesn't have to be.
21:40 kim0 I put it on top .. and it still complains .. __salt__ undefined
21:40 murrdoc can u paste in your code ? in a gist or whatever site u prefer
21:40 iggy look for zfs or ozfs or whatever and if any of them work, set the ZFSCMD variable and return True
21:40 jtang_ joined #salt
21:40 murrdoc also are u running saltutil.refresh_modules after your changes
21:41 iggy said they were using salt-call --local
21:41 kim0 https://gist.github.com/kim0/e9d442661295e5bac503
21:42 giantlock joined #salt
21:42 murrdoc and the error kim0
21:43 manfred that __salt__ doesn't exist
21:43 kim0 http://paste.ubuntu.com/10148642/
21:44 rome_390 joined #salt
21:44 kim0 it's probably related to memoize .. but how do I fix that
21:44 murrdoc yeah thats the memoize declaration
21:44 murrdoc maybe salt.utils has a grain.get hook
21:45 manfred kim0: __salt__ won't be in there at the point that that is run
21:45 manfred you will need to use __grains__
21:45 kim0 manfred: I was using __grains__ initially .. wasn't working too
21:45 manfred __salt__ is populated as it checks all the __virtual__ things for all the modules, and then injects the __salt__ variable at the very end
21:45 manfred yeah, i know, but __salt__ i know for sure won't work
21:45 druonysuse joined #salt
21:45 druonysuse joined #salt
21:45 manfred but I just did a test with __grains__ and it did work
21:45 kim0 ok
21:46 kim0 mmm trying again
21:46 kim0 maybe this was fixed with the __virtual__ on top thing
21:46 murrdoc the __salt__ variable is not available to memoize
21:46 hasues joined #salt
21:46 manfred it also isn't available until after __virtual__ is run
21:47 hasues left #salt
21:47 murrdoc yup
21:47 manfred so, since this is run in __virtual__ it 100% won't work.
21:47 manfred but __grains__ might
21:47 jansauer joined #salt
21:47 kim0 global name '__grains__' is not defined
21:47 murrdoc if cmd and __salt__['cmd.retcode'](cmd) == 0:
21:47 cpowell joined #salt
21:47 manfred yeah, so none of this is going to be usefull kim0
21:47 ocdmw joined #salt
21:47 murrdoc move the logic into virtual
21:47 manfred kim0: if you can do __grains__ in the __virtual__ and then pass that value to the function instead…
21:48 manfred yeah
21:48 manfred and you have to get rid of the __salt__
21:48 jansauer left #salt
21:48 kim0 ok
21:48 manfred if all you need to do is check that /tmp exists… you can do that with other commands
21:48 manfred and don't need to use __salt__
21:48 michelangelo joined #salt
21:48 manfred os.path.isdir('/tmp')
21:49 iggy try setting a cmd variable or using the "if __grains__" format (instead of shoving the __grains__ lookup in the function args
21:49 I3olle joined #salt
21:50 murrdoc anyone much around with a statsd returner
21:50 murrdoc i want to put counters in graphite when a run starts and ends
21:51 murrdoc like status lines or deploy lines
21:51 TyrfingMjolnir_ joined #salt
21:52 murrdoc oh snap http://docs.saltstack.com/en/latest/ref/returners/all/salt.returners.carbon_return.html
21:52 schlueter joined #salt
21:52 aparsons joined #salt
21:54 badon joined #salt
21:54 iggy that makes me want to cry
21:54 iggy I just read that code
21:54 murrdoc i hadnt
21:54 murrdoc havent
21:54 murrdoc how bad
21:54 timoguin joined #salt
21:55 Ch3LL_ left #salt
21:56 mosen joined #salt
21:58 murrdoc you wanted to use collectd python instead of socket ?
21:59 mosen hi salties (i changed it for murrdoc)
21:59 kim0 isn't __virtual__ the absolute first thing that gets executed ?
21:59 jhauser joined #salt
22:00 murrdoc mosen:  be wll
22:00 kim0 or does _available_commands() happen before that
22:00 murrdoc or well
22:02 yomilk joined #salt
22:03 TaiSHi Can I create additional subdirectories on file.directory ?
22:03 TaiSHi (just for the sake of avoiding another file.directory state :P)
22:04 BigBear joined #salt
22:04 jalaziz joined #salt
22:05 murrdoc yeah u can use names
22:05 murrdoc file.directory: user: group: mode: names: - dir - dir1 -dir2
22:05 murrdoc \n\t implied
22:05 TaiSHi Understood
22:05 TaiSHi Thanks!
22:08 murrdoc TaiSHi:  https://github.com/saltstack/salt/issues/19269
22:08 murrdoc names is not specific to file.directory, but is a salt compiler feature and works with any state that can apply the same definitions to a list of items.
22:08 murrdoc
22:09 * TaiSHi hugs murrdoc
22:10 murrdoc hugops
22:10 __number5__ meta~
22:10 mosen murrdoc: - hug.managed
22:11 murrdoc -names:
22:11 murrdoc - basepi
22:11 kim0 folks .. any idea why the zfs module's commands are always displayed verbosely like that => http://paste.ubuntu.com/10148982/
22:11 basepi =D
22:12 aparsons joined #salt
22:13 aparsons_ joined #salt
22:13 TTimo joined #salt
22:13 smithd joined #salt
22:13 kim0 many other commands run, but they are not displayed like that :/
22:14 kim0 basepi: zfs module is develop is way ahead of the one in 2014.7 .. which one should I hack on ?
22:14 kim0 s/is devlop/in develop/
22:14 __number5__ looks like the iptable bug we have before
22:16 basepi kim0: I'd recommend develop
22:16 basepi By the end of the month, we'll mostly stop open source maintenance on 2014.7, because 2015.2.0 will be out.
22:17 kim0 basepi: afterwards, can I beg someone to backport it to 2014.7
22:17 basepi Probably. =) Assuming it backports cleanly
22:17 kim0 or 2015.2 :)
22:17 basepi That's the problem, is if it's quite different, it won't backport.
22:17 basepi But yes definitely, to 2015.2
22:17 kim0 basepi: not too different .. just many added funcs ..etc
22:17 basepi Or you can just make your initial changes in 2015.2, and those will get merged forward to develop
22:17 basepi We merge forward daily from 2014.7 to 2015.2, and from 2015.2 to develop
22:18 basepi So changes get merged all the way forward and aren't lost
22:18 kim0 basepi: but if I work on develop .. changes are never back-ported automatically .. right ?
22:18 jtang_ joined #salt
22:18 basepi That is correct.
22:19 basepi Changes made in develop will be available in the next feature release (that would be codename Beryllium right now) unless they are explicitly backported
22:19 kim0 basepi: should zfs module in 2015.2 .. be close to the one in develop ? is my understanding correct
22:19 basepi kim0: I expect it's pretty close. 2015.2 was only branched a couple of months ago, and I don't recall hearing about a lot of changes to that module recently
22:20 kim0 okay .. sounds good .. 2015.2 it is then
22:20 basepi =)
22:21 TaiSHi So we can expect 2015.2.0 to be on stable repos by march? <3
22:21 __number5__ basepi: is there any recommended migration/upgrade path from 2014.7.x to 2015.2?
22:22 babilen TaiSHi: I'd hope for February (the version number has to make sense once at least!)
22:22 manfred babilen:  2014.7 was released as stable in like october…
22:22 pdayton joined #salt
22:23 TaiSHi manfred, I recall :P
22:23 schlueter1 joined #salt
22:23 babilen __number5__: I'd test your deployment, but it shouldn't take more than just installing the newer packages. If there are explicit changes that you have to take care of you will find them in the release notes (e.g. deprecated functionality)
22:24 babilen manfred: Yes, hence my wish for this release to actually happn in February
22:24 babilen +e
22:24 manfred i just run from git :P
22:24 __number5__ babilen: I wish there is a way to properly test it...
22:24 manfred [root@305264-master ~]# salt --version
22:24 manfred salt 2015.2.0-2878-g654de6b (Lithium)
22:25 manfred forever and ever
22:25 Tyrm_ joined #salt
22:25 babilen __number5__: I model my setup in vagrant and test new versions in there ... but then you can never really test everything that you might find in production. But then, testing my setup on a couple of local VMs allows me to catch easy mistakes/problems early on.
22:27 __number5__ babilen: yep, we have our CI world and packer AMI building process which will pick up easy errors
22:29 jtang_ joined #salt
22:37 ilbot3 joined #salt
22:37 Topic for #salt is now Welcome to #salt | SaltConf 2015 is Mar 3-5! http://saltconf.com | 2014.7.1 is the latest | Please be patient when asking questions as we are volunteers and may not have immediate answers | Channel logs are available at http://irclog.perlgeek.de/salt/
22:41 notnotpeter following up on a earlier question...Can pillar data be defined in such a way that you don't have to hard code the names of each of your environments in pillar/top.sls?
22:42 octagonal joined #salt
22:42 iggy I think a lot of people are going to have to do some work to go from 2014.7 to 2015.2
22:42 kermit joined #salt
22:42 iggy (technically stuff that should have been done before)
22:42 iggy 2015.2 is much whinier about stuff
22:42 octagonal I've got a sls file where I set var="no". In the evaluated template, it's being evalated to "False". How do I pass it "no" as a string freaking literal?
22:43 notnotpeter It feels super inflexible that pillar environments have to be defined statically and I can't have a default set of pillar data.
22:43 [LF] joined #salt
22:44 murrdoc u can use external pillars and use grains to include plillars from top.sls
22:45 notnotpeter murrdoc: I was looking at external pillars as a possible method.
22:45 murrdoc i use an http ext_pillar
22:45 iggy if something isn't in an env, it should filter up to the "base" environment
22:46 murrdoc because i have an existing functional data store that represents a server
22:46 iggy I use grains and say "screw you security"
22:46 notnotpeter iggy: is that true? I currently only have base defined in my pillar/top.sls and if I change the env of a minion it no longer pulls base pillar data
22:47 iggy notnotpeter: everything I've ever read, says yes
22:47 iggy but I've never really tried without specifying any env's at all
22:47 iggy (why do you need so many envs that you have a problem specifying them in the top file?
22:48 murrdoc yeah
22:48 notnotpeter It's not the "so many" it's the I want them to be dynamically creatable
22:48 iggy then you want something other than flat files to manage your pillars
22:48 notnotpeter Like I want a developer to be able to create a branch of salt code and work on a state
22:48 murrdoc they dont need an env for that
22:48 GabLeRoux joined #salt
22:48 murrdoc they can use base for dev'ing
22:49 [LF] joined #salt
22:49 * octagonal twitches
22:49 murrdoc i mean on their vagrants
22:49 ocdmw joined #salt
22:49 murrdoc not in prod man damn
22:49 octagonal thank you.
22:49 meefs joined #salt
22:50 octagonal heh
22:50 octagonal I've seen people do such things.
22:50 gladiatr joined #salt
22:50 murrdoc u will love salt-beacons
22:51 schlueter joined #salt
22:51 manfred salt-beacons look awesome
22:51 murrdoc 'beacon on login'
22:52 teebes joined #salt
22:52 manfred always use dev.
22:52 notnotpeter I guess I'm looking for a method that someone can work on a state in their vagrant and then test that on stage server once they've gotten it working.  I don't see how to do that without envrionments.
22:52 murrdoc well u dont need too many environments
22:52 manfred that is a large part of what environments are meant for.
22:52 murrdoc u could do the canarying using pillars
22:52 octagonal beacons?
22:53 manfred octagonal:  https://github.com/saltstack/salt/pull/20092
22:53 murrdoc also beacons is going to be franks red sauce for me
22:53 natewalck joined #salt
22:53 murrdoc imma put that shit on everything
22:53 manfred basically, watch stuff happen on the server, and send stuff to the master reactor when things happen, like files are changed / other inotify events
22:53 manfred or someone logs in to the server
22:53 manfred or some message appears in journald if you are running with systemd
22:53 octagonal murrdoc: I've been waiting for these
22:54 octagonal murrdoc: "oh, you edited the managed file? Imma put it back now, sorry."
22:54 wavis joined #salt
22:54 manfred if you open an /etc/ file with vim, i am gonna kill -9 vim
22:54 murrdoc yeah man
22:54 eligos joined #salt
22:54 murrdoc so much win
22:54 octagonal omg yes.
22:55 manfred and then ban your user
22:55 manfred passwd -l idiot
22:55 notnotpeter murrdoc: I understand I could use static "dev/qa/prod" branches but we often are running multiple different staging environments in parallel.  I'd rather not have half a dozen long running branches and instead be able to actually close branches as their features go out in prod.
22:55 octagonal echo "You have been larted." | mail
22:55 manfred lol
22:55 octagonal notnotpeter: that's easy.
22:55 murrdoc u can also do {env}: - {env} notnotpeter
22:55 murrdoc u dont have to hard code them
22:56 octagonal ^
22:56 akafred joined #salt
22:56 modafinil_ joined #salt
22:56 grepory joined #salt
22:56 mihait joined #salt
22:57 akoumjian joined #salt
22:57 murrdoc there is also reclass
22:58 murrdoc if u want to use flatfiles with stuff like inheritance
22:59 notnotpeter I'll look at reclass, thanks for the pointer.
23:00 octagonal in case anyone runs into this, in your sls you have to double-quote variables that are going to be processed by jinja.. so var='"yes"'
23:01 octagonal because yo dawg, I heard you like string literals
23:01 octagonal assuming the variable is something pyyaml wants to turn into a boolean
23:01 octagonal Turns out sshd really doesn't like being fed a config file with "False" instead of "no" >.>
23:01 notnotpeter murrdoc/octagonal: Sorry if I'm feeling dumb here, but I'm not sure I understand your {env}: - {env} example
23:05 octagonal notnotpeter: In your sls file, you put a line like that. It expands based on whatever you tell it.
23:05 octagonal at run time.
23:05 murrdoc i think {{ env }} or {{ saltenv }} is a variable that allows  you to do that
23:06 murrdoc so u dont have to do  base: - state dev1: - state, u can do {{ env }}: - {{ env }}
23:06 higgs001 joined #salt
23:08 octagonal I think on newer salts {{ env }} is the actual environment
23:09 bash124512 joined #salt
23:09 octagonal so I assume there should be one for {{ saltenv }} but I'm not seeing it documented
23:09 murrdoc i am using pillars for gating features
23:09 murrdoc cos i am lazy like that
23:10 diegows joined #salt
23:10 thedodd joined #salt
23:11 forrest joined #salt
23:17 jonasbjork joined #salt
23:19 nullptr joined #salt
23:19 notnotpeter octagonal: So I've replaced the toplevel 'base': with a '{{ saltenv }}:' in my pillar top.sls and no lo luck.
23:20 notnotpeter iggy: I also explicitly tested whether other environments inherit pillars from base.  For me, with 2014.1, they definitely do not.
23:20 s8weber_ joined #salt
23:23 iggy that sucks
23:24 manfred so, you just want to inherit pillars right? and not states?
23:24 notnotpeter Yep. Already got states sorted.
23:24 manfred and you have matching all done in the top.sls?
23:25 notnotpeter I'm thinking I have to change my pillar_roots in my master config, because base is being specified there.
23:26 modafinil_ joined #salt
23:26 johtso joined #salt
23:26 otter768 joined #salt
23:30 Striki joined #salt
23:31 beneggett joined #salt
23:31 eliasp is there any salt-ism which would allow me in Jinja to test for the presence of a given state? something like: {% 'foo' in salt['states'] %} do this {% else %} do that {% endif %}
23:32 manfred not really?
23:32 notnotpeter octagonal/murrdoc: Any idea how can I get my salt master to look in the same folder for every environment's pillar_roots? I think your jinja'd {{ env }} is most the way there, but my salt master is still only looking in that directory for base (since my master config pillar_roots only includes base).  And I can't use jinja in the master config.
23:32 ajw0100 joined #salt
23:33 [LF]1 joined #salt
23:34 murrdoc the top.sls is the same
23:34 notnotpeter eliasp: We use a common variable set for this. Setting a variable determines whether a state gets included, and other states can check the value of that variable."
23:34 murrdoc for salts and pillars
23:34 iggy eliasp: isn't that what watch_in and onchanges_in would be for?
23:35 eliasp iggy: I don't want to monitor changes of the state, but more like: is foo/bar.sls available?
23:35 murrdoc dude
23:35 murrdoc u have no idea how nice that would be
23:35 iggy available or assigned to this minion
23:35 iggy ?
23:35 eliasp available
23:36 aquinas joined #salt
23:36 eliasp I want to create an override for win pkgs which would allow me to have drop ins for custom packages… if such a drop-in SLS is not available, just fall back to pkg.installed: $foo
23:36 scott2b joined #salt
23:37 eliasp as win packaging is not always that straightforward and not every brokenness of win-packaging can't be covered by win_pkg
23:37 thunderbolt_ joined #salt
23:37 eliasp so I could just create drop-in SLS containing those "advanced handling" for broken packages
23:38 TaiSHi Meh, saltstack > sex
23:39 murrdoc TaiSHi be in high state
23:39 MTecknology I don't remember what sex is like anymore.
23:39 TaiSHi MTecknology, murrdoc, I can't stop thinking about salt :P
23:39 eliasp <murrdoc> u have no idea how nice that would be - were you referring to "{% if 'foo' in salt['states'] %}…"? so do you think filing a feature request would make sense?
23:39 TaiSHi Is that an issue? My wife thinks it so
23:39 murrdoc eliasp:  i did
23:40 eliasp TaiSHi: she should take all this with a grain of … Salt!
23:40 murrdoc hah
23:40 eliasp murrdoc: ok… so I might file an issue tomorrow then ;)
23:40 murrdoc u should
23:40 murrdoc i ll vote so hard
23:40 murrdoc :D
23:40 eliasp if there only were votes! :)
23:40 TaiSHi Now, on all seriousness, my sexual appetite has decreased significantly since work increased :P
23:41 notnotpeter eliasp: I like that idea.  Similar to the way you can provide multiple source entries which it tries in order to find a file (per host, per env, default, etc) you could have a mechanism to hunt down to more generic pkg.installed mechanisms.
23:41 murrdoc thats bad
23:41 TaiSHi Tell me about it
23:42 eligos joined #salt
23:43 murrdoc only one thing to do
23:43 redzaku joined #salt
23:43 murrdoc get her pregnant
23:44 MTecknology I haven't had sex in... ~2yr
23:44 eliasp been there, done that… just adds more workload on the @home side of life ;)
23:44 TaiSHi :P it comes down to that, more work
23:44 timoguin joined #salt
23:44 notnotpeter murrdoc: So using a top level {{ saltenv }}: in my pillar/top.sls seems to be working. But only for environments I specify by hand in the salt master config under pillar_roots.  So close...
23:44 TTimo joined #salt
23:44 * TaiSHi is anxious on 2015.2 to be released on ppa's
23:45 TaiSHi I'm even learning to code to aid on saltstack project
23:45 TaiSHi And I -hated- coding.
23:46 neogenix joined #salt
23:46 __number5__ TaiSHi: you'll have a bad life in devops XD
23:46 MTecknology Working in salt is where I got my real love of working with python
23:46 MTecknology and git
23:46 notnotpeter Taishi: Then pin your installs to the 2014.7 ppa until your comfortable with 2015.2
23:47 TaiSHi MTecknology, I hate PHP so much, every time I tried to learn it was PHP... but when I hit Python it was love at first line
23:47 TaiSHi __number5__, I do sysadmin/engineering, I get around fine on scripting, but never got into heavy coding
23:47 TaiSHi I mean, I'm a freaking devops :P
23:47 TaiSHi notnotpeter, never! Straight update.
23:48 notnotpeter Taishi: that's what I did until I hit a show stopper which kept us from deploying. :(
23:48 TaiSHi Better to update than sorry, they say
23:48 natewalck joined #salt
23:49 TaiSHi Bad thing about working as a devops in Argentina, is that we're heavily underrated
23:51 akafred joined #salt
23:52 __number5__ TaiSHi: move to US or other western countries
23:52 murrdoc forget south america
23:52 theologian joined #salt
23:52 TaiSHi __number5__, I lived in Spain for a while, was young and stupid
23:53 murrdoc and your delicious cup of tea with weed like properties
23:53 MTecknology TaiSHi: In any given week, I'm likely to work with html, css, python, javascript, perl, vb.net, c#, bash, ksh, yaml, rst, php, c, and much more... every week I probably work with most of them and some others. I've even dealt with clojure a few times.
23:53 murrdoc whats it called
23:53 TaiSHi Now I'm 28, still stupid, and didn't get my college degree
23:53 MTecknology python is my favorite by far
23:53 TaiSHi murrdoc, MATE!
23:53 murrdoc MATE!
23:53 murrdoc so good
23:54 TaiSHi murrdoc, that weed is called yerba, the beverage is called mate
23:54 antonw joined #salt
23:54 TaiSHi MTecknology, it's just so... dunno, it flows.
23:54 murrdoc people walk around drinking mate
23:54 TaiSHi __number5__, I want to finish my studies and preferably outsource, I like Argentina
23:55 MTecknology TaiSHi: I like that it helps force people into adopting sane structure and syntax as opposed to the clusterfuck of crap you get with pretty much any other language and it's easy to build something that's basic and easy to read or incredibly powerful with a single line
23:55 notnotpeter Is that the same thing that's in Club-Mate (the german soda)?
23:56 TaiSHi Probably not notnotpeter
23:56 TaiSHi MTecknology, I'm quite OCD when it comes to indentation so yeah, I feel you
23:58 murrdoc all these indentation problesm affect all these new kids
23:58 murrdoc with their new fangled browsers
23:58 joeyparsons joined #salt
23:58 murrdoc pathogen + https://github.com/saltstack/salt-vim
23:58 murrdoc and done

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