Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-03-26

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

All times shown according to UTC.

Time Nick Message
00:00 crd migrating the registry service managed by upstart to using dockerio states
00:00 desposo joined #salt
00:02 crd I'm on rubbish coffee shop wifis right now and proving that patch is driving me to drink
00:02 warthog42 joined #salt
00:03 iggy lol, nitin said the same thing I did
00:04 crd haha, I see that
00:04 chandy joined #salt
00:05 murrdoc joined #salt
00:06 murrdoc1 joined #salt
00:08 __number5__ any one using gpg renderer to manage their ssl certificate private key? or any other private keys?
00:08 iggy the postfix-formula has an example of looking up by lsb_distrib_name
00:14 XenophonF left #salt
00:14 bfoxwell joined #salt
00:15 kusams joined #salt
00:18 yomilk joined #salt
00:23 rhodgin joined #salt
00:26 clintberry joined #salt
00:29 MatthewsFace joined #salt
00:31 markm joined #salt
00:31 bluenemo_ joined #salt
00:33 Hell_FireW joined #salt
00:35 irctc084 joined #salt
00:37 enarciso joined #salt
00:39 murrdoc joined #salt
00:40 baweaver joined #salt
00:41 enarciso_ joined #salt
00:45 enarciso joined #salt
00:46 murrdoc joined #salt
00:46 salty_to_the_cor joined #salt
00:47 murrdoc joined #salt
00:47 murrdoc joined #salt
00:49 enarciso_ joined #salt
00:49 cheus_ joined #salt
00:50 salty_to_the_cor is there a guideline for writing tests for custom execution and state modules? I have witten many custom modules, but how do i write tests where i can mock __salt__ etc. Also is there a directory structure i need to follow?
00:51 enarciso joined #salt
00:56 hal58th_ joined #salt
00:56 enarciso_ joined #salt
00:59 __number5__ salty_to_the_cor: https://github.com/saltstack/salt/blob/develop/tests/unit/modules/cp_test.py
00:59 enarciso joined #salt
01:00 trs-nj joined #salt
01:00 __number5__ that's just for the __salt__ mocking part and a typical salt module unit tests
01:01 trs-nj "no module named salt.scripts" when installing salt-master via epel in aws frankfurt.  works fine in all other regions
01:01 trs-nj has anyone heard of an issues with the package in that region?
01:03 salty_to_the_cor __number5__: thanks for this. But where should i place this?? for this is my custom module? do i need to add tests always to salt/test/unit ???
01:03 dyasny joined #salt
01:03 salty_to_the_cor can we have a custom directory structure?
01:04 __number5__ salty_to_the_cor: directory layout mostly depends on what test runner you are going to use
01:04 __number5__ salt/test/unit is saltstack's convention
01:05 salty_to_the_cor any suggesstions which to use? i am thinkin of using nose. Also you mean just import from salt otherwise, directory structure dosent matter?
01:05 enarciso_ joined #salt
01:06 yomilk joined #salt
01:07 enarciso joined #salt
01:08 amcorreia joined #salt
01:10 murrdoc joined #salt
01:11 enarciso_ joined #salt
01:11 aqua^mac joined #salt
01:11 lictor36 joined #salt
01:16 malinoff joined #salt
01:20 hackel joined #salt
01:22 dyasny joined #salt
01:28 sunkist joined #salt
01:29 chandy joined #salt
01:31 murrdoc joined #salt
01:34 yomilk joined #salt
01:36 dalexander joined #salt
01:40 murrdoc joined #salt
01:41 enarciso joined #salt
01:41 nich0s joined #salt
01:44 enarciso joined #salt
01:47 enarciso joined #salt
01:49 scbunn joined #salt
01:49 enarciso_ joined #salt
01:49 racooper joined #salt
01:57 hackel If a vm deployment fails to start Salt (because of bad permissions of my key files), how can I tell salt-cloud to try again?  Do I just need to delete and re-create the vm?
01:57 surajd joined #salt
02:02 scbunn joined #salt
02:05 yomilk joined #salt
02:07 clintberry joined #salt
02:12 otter768 joined #salt
02:15 sunkist joined #salt
02:15 CeBe joined #salt
02:16 catpigger joined #salt
02:18 jerematic joined #salt
02:25 evle joined #salt
02:25 kusams joined #salt
02:26 yomilk_ joined #salt
02:28 donmichelangelo joined #salt
02:29 andrew_v joined #salt
02:45 Terminus- joined #salt
02:47 ilbot3 joined #salt
02:47 Topic for #salt is now Welcome to #salt | 2014.7.2 is the latest | Please use https://gist.github.com for code, don't paste directly into the channel | Please be patient when asking questions as we are volunteers and may not have immediate answers | Channel logs are available at http://irclog.perlgeek.de/salt/
02:50 TyrfingMjolnir joined #salt
02:52 nexsja1 joined #salt
02:55 sunkist joined #salt
02:56 nexsja joined #salt
02:59 mlanner joined #salt
03:00 dude051 joined #salt
03:00 favadi joined #salt
03:03 TyrfingMjolnir joined #salt
03:07 rectange joined #salt
03:08 rectange hi there. could someone please help me with the archive function?
03:09 Terminus- hello. i've got a style question. i'm looking at dhcpd-formula and it seems to use include in a loop rather than macros for generating multiple subnets. are there any reasons to prefer include over macros when writing formulas?
03:11 aparsons joined #salt
03:14 subsignal joined #salt
03:16 neogenix_ joined #salt
03:16 aparsons joined #salt
03:19 jerematic joined #salt
03:20 mosen joined #salt
03:27 surajd joined #salt
03:37 kormoc joined #salt
03:38 kormoc joined #salt
03:39 TyrfingMjolnir joined #salt
03:39 kormoc joined #salt
03:51 iggy trs-nj: did you file an issue in github? (there is one)
03:51 iggy https://github.com/saltstack/salt/issues/22003
03:52 iggy hackel: that's the best way I think... I can't think of a way to make salt-cloud finish something that failed
03:53 chandankumar joined #salt
03:53 iggy Terminus-: personal taste mostly
03:53 iggy rectange: help in what way?
03:53 hackel iggy: I ended up just installing salt-minion manually.  Not sure what salt-cloud does other than that, but it's working fine.
03:54 hackel iggy: I thought the "saltify" provider should have done what I needed, but whenever I tried to use it, salt-cloud said it was not configured and so removed it.
03:55 iggy that works I guess
03:56 Terminus- iggy: gotcha. i'll stick with macros then. thanks.
03:58 balltongu_ joined #salt
03:59 Furao joined #salt
04:04 otter768 joined #salt
04:06 TyrfingMjolnir joined #salt
04:07 spookah joined #salt
04:31 catpigger joined #salt
04:32 chandankumar joined #salt
04:32 aparsons joined #salt
04:37 aparsons joined #salt
04:45 TyrfingMjolnir joined #salt
05:04 Terminus- joined #salt
05:08 jerematic joined #salt
05:09 mdln joined #salt
05:14 kusams joined #salt
05:17 Hell_FireW joined #salt
05:17 clintberry joined #salt
05:18 vstoniest joined #salt
05:21 sunkist1 joined #salt
05:25 sunkist joined #salt
05:30 desposo joined #salt
05:46 pdayton joined #salt
05:49 ramteid joined #salt
05:53 jalbretsen joined #salt
05:55 chandankumar joined #salt
05:55 malinoff joined #salt
06:00 jerematic joined #salt
06:06 sunkist1 joined #salt
06:07 TyrfingMjolnir joined #salt
06:07 piv joined #salt
06:10 TyrfingMjolnir joined #salt
06:15 ajw0100 joined #salt
06:17 Guest89 joined #salt
06:19 viq joined #salt
06:27 Furao joined #salt
06:27 TyrfingMjolnir joined #salt
06:30 baoboa joined #salt
06:36 micko joined #salt
06:45 TyrfingMjolnir joined #salt
06:50 krelo joined #salt
06:52 dopesong joined #salt
06:57 refnode_ joined #salt
06:58 TyrfingMjolnir joined #salt
07:03 Furao joined #salt
07:08 colttt joined #salt
07:09 AndreasLutro joined #salt
07:17 yomilk joined #salt
07:19 __gotcha joined #salt
07:24 favadi left #salt
07:27 TyrfingMjolnir joined #salt
07:33 stoogenmeyer joined #salt
07:35 __gotcha joined #salt
07:36 yomilk_ joined #salt
07:40 dopesong joined #salt
07:41 flyboy joined #salt
07:41 _JZ_ joined #salt
07:47 dopesong joined #salt
07:49 jerematic joined #salt
07:52 refnode_1 joined #salt
07:52 dRiN joined #salt
07:53 refnode_2 joined #salt
07:53 Auroch joined #salt
07:54 refnode_2 joined #salt
07:54 refnode_3 joined #salt
07:57 JayFK joined #salt
08:00 favadi joined #salt
08:01 pf_moore joined #salt
08:02 AirOnSkin joined #salt
08:02 AirOnSkin Good morning, all.
08:03 AirOnSkin With jinja templating I've seen different kinds of brackets. I'm especially curious about the difference between {} and [] ... can someone tell me when to use which?
08:05 AirOnSkin In a blog post I've seen: {% for role in salt['pillar.get']('roles', []) %} and on another {% for user, userinfo in salt['pillar.get']('users', {}).iteritems() %} ...
08:05 Terminus- AirOnSkin: AFAIK, [] is only used for lists. you use {% ... %} for statements, {{ ... }} for variable expansion, and {# ... #} for comments.
08:05 AirOnSkin Is there a "correct" set of brackets to use?
08:05 Terminus- AirOnSkin: with "subclass "{{ i.name }}" 1:{{ j.mac }};{{ ' # %s' % j.name if 'name' in j else '' }}{{ '(%s)' % j.comment if 'comment' in j else '' }}
08:05 Terminus- oops.
08:07 Terminus- AirOnSkin: with "salt['pillar.get']('roles', [])", you are getting the value of 'pillar.get' in the salt dict. the value is a function with parameters. the first one is the key to look for in the sls file, the 2nd is what to return if it is not found. in this case, an empty list.
08:07 eseyman joined #salt
08:07 AirOnSkin Terminus-: I see. Gotta review my own statements...
08:08 AirOnSkin It's really useful if you know yaml, jinja and python if you want to use Salt... need to get to know these things better
08:09 surajd joined #salt
08:09 Terminus- AirOnSkin: if you understand the basic python data types, it's pretty easy to translate it to yaml.
08:11 Terminus- AirOnSkin: the 'salt' variable is provided as a convenient way to use salt functions as stated in http://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.jinja.html#calling-salt-functions
08:11 csar joined #salt
08:12 AirOnSkin Terminus-: Well, I'm only starting to learn Python and it's taking a while since I've never learnt a programming language before...
08:14 Terminus- AirOnSkin: to expand further, the example you gave calls http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.pillar.html#salt.modules.pillar.get
08:14 Terminus- AirOnSkin: good luck. i highly recommend going through the python tutorial.
08:14 refnode joined #salt
08:15 refnode good morning
08:19 AirOnSkin Terminus-: Thank you :) Will do
08:21 csar joined #salt
08:22 krelo joined #salt
08:25 _rsk_ joined #salt
08:26 Terminus- AirOnSkin: i have yet to look at a single line of python code ever since i started using salt but it helps that i'm familiar with data structure and templating concepts. =)
08:26 aqua^mac joined #salt
08:28 clintberry joined #salt
08:30 otter768 joined #salt
08:32 MaliutaLap joined #salt
08:35 KermitTheFragger joined #salt
08:36 Andre-B joined #salt
08:50 wincyj joined #salt
08:54 madduck_ joined #salt
08:55 MaliutaLap left #salt
08:56 bluenemo joined #salt
08:56 bluenemo joined #salt
09:04 ckao joined #salt
09:10 __gotcha joined #salt
09:10 aqua^mac joined #salt
09:18 Xevian joined #salt
09:27 ktosiek joined #salt
09:34 N-Mi_ joined #salt
09:38 jerematic joined #salt
09:41 dopesong joined #salt
09:43 Grokzen joined #salt
09:45 piv joined #salt
09:46 david_ joined #salt
09:48 hojgaard joined #salt
09:50 fredvd joined #salt
09:58 aquassaut joined #salt
09:59 TyrfingMjolnir joined #salt
10:01 Romlok joined #salt
10:04 AirOnSkin How can I use for loop, but have the output be on one line (without line breaks)?
10:04 phx uhm, you can, let me check my stuff
10:05 phx URI {% for srv in salt['pillar.get']('serviceclients:services:auth', []) %}ldap://{{ srv.hostname }} {% endfor %}
10:05 phx from my ldap conf state
10:06 wnkz joined #salt
10:06 berserk joined #salt
10:06 phx that will produce a single line. so you just have to put the loop in a single line, because the newline is part of the text that gets printed in every iteration
10:06 AirOnSkin phx: Oh, so writing it on one line will do that? oO, I thought it would be with some sort of attribute or filter or what do I know
10:06 peters-tx joined #salt
10:06 sirshurf joined #salt
10:06 sirshurf Hi all, anybody here?
10:06 AirOnSkin Great, thanks!
10:07 * AirOnSkin isn't here
10:07 phx AirOnSkin, yw
10:07 phx sirshurf, nope, sorry, we're all dead
10:07 sirshurf :) Great, then I will ask yoru spirits :) Hope they will answere me :)
10:07 phx i can see some high spirits over there! :)
10:09 sirshurf I am getting nuts with some with undestanding pillars...  and I cannot find any example that I can understand...
10:09 phx what's so hard to understand?
10:10 phx pillars is serverside data, not rendered on the client, but made available to the client, with some access control (pillars top specifies what pillars are available to what minions)
10:10 phx they are also a kind of input to states
10:11 phx so, let's say you have some data available on the server, which has to considered in states, but states are being run on minions, then pillars are the abstraction layer you are looking for
10:11 sirshurf the biggest one? HOW... I will explane my setup... and maybe you can help me try to set it up...
10:11 phx does the above help you? :)
10:15 sirshurf I have 2 kinds of servers, 1st kind can only be accessed by local IP's (but only from MY servers so by specific IP's) on all ports (for the ease of setup for now...) and the second one, can be access on all ports on local IP's like the first ones and on specific ports from external IP's...
10:15 sirshurf I do not understand HOW should I setup my pillars/states to manage this in IPtables...
10:17 AirOnSkin sirshurf: I'm doing something like that now. My suggestion: Define two roles for both servers, then create a conditional for your network state and iptables state. If the first role matches configure it more securely, else configure it more open
10:18 phx sirshurf, it can be done with or without pillars
10:18 aquassaut joined #salt
10:19 giantlock joined #salt
10:19 sirshurf AirOnSkin: Can you share you setup? I am really do not understand enouth how to use salt... :(
10:19 phx sirshurf, an example for pillar usage would be to render your servers IPs from your inventory into a pillar (python renderrer ftw), and from the states reference that one. And by some logic (can be a grain, or tied to your inventory using another pillar) check what access model the server needs and generate the ruleset accordingly
10:20 I3olle joined #salt
10:20 sirshurf phx: Can you please explane with example... I am a real nubie here :( please please please...
10:20 phx sirshurf, i have NDAs as well, mind you
10:21 linjan joined #salt
10:22 phx sirshurf, personally i did it in a bit of an overkill fashion, but it's still an example. I have an inventory database (pgsql, relational), where i store stuff up to some highlevel info. Based on that, i can check whether a particular box has an external IP. i make this info available in a pillar, only the salt master accesses the inventory directly. From the iptable states, i check whether the current server ne
10:22 phx eds iptables rules for external stuff. if yes, then i iterate over another pillar, which lists all systems' external IPs, and i add access rules based on that
10:23 phx sirshurf, basically, it's a single if conditional, and a for loop on a pillar in the iptables state
10:24 kusams joined #salt
10:24 sirshurf phx: I am holding my head in my hands looking at what you wrote here...
10:25 AirOnSkin sirshurf: for example, you have the following two pillars: http://pastebin.com/9TGD8aQp & http://pastebin.com/FYGHJnbu
10:26 AirOnSkin sirshurf: Then you could do something like (give me a minute):
10:27 phx sirshurf, just write a sandbox state, which is using a pillar. then start playing with it, so you can see how it clicks together
10:28 phx the important thing you have to keep in mind is, salt is very flexable
10:29 crazysim joined #salt
10:30 AirOnSkin http://pastebin.com/XyKC7ZX1
10:31 otter768 joined #salt
10:31 AirOnSkin yeah, what phx said... testing with a sandbox state helped me a lot in understanding the mechanics... you can build up from there
10:31 o5k joined #salt
10:32 sirshurf How do I test with sandbox state?
10:32 sirshurf I am googling this now... but if you can point me that will be great
10:32 phx sandbox is where you take your toy shovel, and start making sandcastles to later kick them, and rebuild them. it's a playground for testing things
10:33 phx just make a state and a pillar with any name, manipulating a test file, or some dummy service
10:33 sirshurf AirOnSkin: Thanks for the pastebin... I will try to work from it
10:33 AirOnSkin sirshurf: a state that transfers a file to /tmp or echoes a command to /dev/null or tests if a file exists...
10:33 AirOnSkin sirshurf: and build conditions around these test states with Pillar data
10:34 sirshurf phx: I will try... and probably get back here with more questions... some where allready answered by you! THANKS!
10:34 phx sure, just go and get some handson experience
10:34 AirOnSkin sirshurf: "Only transfer this file if Pillar data xyz == dft
10:34 phx you will understand a lot more by doing it
10:34 bhosmer joined #salt
10:34 phx it's not hard, you just have to get a grasp on the mechanics
10:34 sirshurf phx: I tried that... took me hald a day to restore a server back :)
10:34 AirOnSkin sirshurf: My States and Pillars are not very elegant - I'm quite new as well - but I hope they help understanding what you can do
10:35 phx AirOnSkin, by the look of a state, i can tell when did i write it during my learning curve :)
10:35 sirshurf AirOnSkin: they explane exactly what where missing for me... :) At least for now :)
10:36 AirOnSkin phx: yeah, I spend half of my time working with Salt correcting my past work ^^
10:42 sirshurf AirOnSkin: for some reasone I cannot answer you in the messages, I will check thouse links now! Thanks!
10:46 aquassaut joined #salt
10:46 AirOnSkin sirshurf: no worries :)
10:49 * AirOnSkin is away for lunch
10:54 lothiraldan joined #salt
10:55 _ether_ joined #salt
10:56 istram joined #salt
10:57 intellix joined #salt
10:57 intellix left #salt
11:02 flebel joined #salt
11:04 surajd joined #salt
11:13 Romlok are there any example sls files for install/config of halite floating around?
11:14 Romlok seems like it would be a natural thing to include with halite
11:14 amcorreia joined #salt
11:15 hal58th joined #salt
11:16 esogas_` joined #salt
11:17 dingo_ joined #salt
11:17 scarcry_ joined #salt
11:17 seev_ joined #salt
11:18 __ale__ joined #salt
11:18 samed_ joined #salt
11:18 al joined #salt
11:19 ntropy_ joined #salt
11:19 Romlk joined #salt
11:19 robothands joined #salt
11:19 yomilk joined #salt
11:19 tyler-baker joined #salt
11:20 peters-tx0 joined #salt
11:20 jY- joined #salt
11:20 rome_390_ joined #salt
11:20 nkuttler_ joined #salt
11:21 z3uS| joined #salt
11:21 sk_0_ joined #salt
11:21 wormwood_ joined #salt
11:21 buxy_bak joined #salt
11:22 sgate1_ joined #salt
11:22 sirtaj_ joined #salt
11:23 crazysim_ joined #salt
11:23 fxdgear_ joined #salt
11:23 markm joined #salt
11:23 signull_ joined #salt
11:23 jerematic joined #salt
11:23 wincyj joined #salt
11:24 JoeHazzers_ joined #salt
11:24 djinni` joined #salt
11:24 Laogeodritt| joined #salt
11:25 randomus` joined #salt
11:25 micko1 joined #salt
11:25 shnguyen_ joined #salt
11:26 kusams joined #salt
11:26 lytchi joined #salt
11:26 xMopxShe- joined #salt
11:26 joehh1 joined #salt
11:27 rgarcia__ joined #salt
11:28 heise joined #salt
11:28 cwyse joined #salt
11:29 pcdummy_ joined #salt
11:29 arif-ali joined #salt
11:29 aberdine joined #salt
11:29 leszq joined #salt
11:29 Morbus joined #salt
11:29 paolo joined #salt
11:29 ThomasJ joined #salt
11:29 forze joined #salt
11:29 davromaniak joined #salt
11:29 mackstick joined #salt
11:29 SaveTheRbtz joined #salt
11:29 sjohnsen joined #salt
11:29 TaiSHi joined #salt
11:29 ToeSnacks joined #salt
11:29 nickg joined #salt
11:29 Damianz joined #salt
11:29 mortis_ joined #salt
11:29 rlarkin joined #salt
11:29 phpdave11 joined #salt
11:29 beauby joined #salt
11:29 CryptoMer joined #salt
11:30 chandy joined #salt
11:30 Schmidt joined #salt
11:31 samed joined #salt
11:32 twiedenbein joined #salt
11:32 gmoro joined #salt
11:32 egil_ joined #salt
11:32 vschum1 joined #salt
11:33 Micromus_ joined #salt
11:33 jrluis joined #salt
11:33 Hipikat_ joined #salt
11:33 michelangelo joined #salt
11:33 fredvd joined #salt
11:33 N-Mi_ joined #salt
11:33 AndreasLutro joined #salt
11:33 mdln joined #salt
11:33 stooj joined #salt
11:33 pdx6 joined #salt
11:33 bougie joined #salt
11:33 snaggleb joined #salt
11:33 Damon joined #salt
11:33 jtanner joined #salt
11:33 OnTheRock joined #salt
11:33 ]V[ joined #salt
11:33 \ask joined #salt
11:33 active8 joined #salt
11:33 eclectic joined #salt
11:33 koomi joined #salt
11:33 ClausA joined #salt
11:33 ShibMcNe joined #salt
11:33 terinjokes joined #salt
11:33 hackeron joined #salt
11:33 djaykay joined #salt
11:33 hellerbarde joined #salt
11:33 Number6 joined #salt
11:33 __number5__ joined #salt
11:33 philipsd6 joined #salt
11:33 danee joined #salt
11:33 ksk joined #salt
11:33 froztbyte joined #salt
11:33 nahamu joined #salt
11:33 the_lalelu joined #salt
11:33 ze- joined #salt
11:33 kevc joined #salt
11:33 Yoda-BZH joined #salt
11:33 Ixan joined #salt
11:33 spootly joined #salt
11:33 dec joined #salt
11:33 lz-dylan joined #salt
11:33 eliasp joined #salt
11:33 marcinkuzminski joined #salt
11:33 grep_away joined #salt
11:33 phx joined #salt
11:33 gadams joined #salt
11:33 heaumer joined #salt
11:33 dork joined #salt
11:33 Ahlee joined #salt
11:33 dnj` joined #salt
11:35 rickflare joined #salt
11:39 anteaya joined #salt
11:39 bhosmer joined #salt
11:43 claytron joined #salt
11:45 piv joined #salt
11:46 jhauser joined #salt
11:47 dunz0r joined #salt
11:50 danemacmillan joined #salt
11:55 elfixit joined #salt
11:57 zer0def joined #salt
11:57 scarcry joined #salt
12:05 \ask joined #salt
12:05 scarcry joined #salt
12:06 evle1 joined #salt
12:08 buxy joined #salt
12:10 \ask joined #salt
12:16 andreyst joined #salt
12:17 ekristen joined #salt
12:17 masterkorp hmm
12:18 masterkorp salting travis machines to test formulas :p
12:18 masterkorp forge the the /me is
12:18 rdxc joined #salt
12:19 slav0nic joined #salt
12:19 slav0nic joined #salt
12:19 masterkorp *forget
12:23 aphor joined #salt
12:23 clmsy joined #salt
12:29 wicope joined #salt
12:30 cmcmacken joined #salt
12:31 AirOnSkin What does **kwargs really stand for? I'm not sure how to interpret it...
12:32 otter768 joined #salt
12:33 emostar AirOnSkin: http://stackoverflow.com/questions/1769403/understanding-kwargs-in-python
12:33 clmsy any number of keyword arguments
12:36 AirOnSkin Ah, I understand now... thanks you two :)
12:40 lothiraldan joined #salt
12:41 bersace joined #salt
12:41 zer0def joined #salt
12:41 bersace Hi. How to select saltenv for a minion ?
12:41 bersace salt minion state.highstate
12:42 msciciel joined #salt
12:45 bhosmer joined #salt
12:47 wincyj joined #salt
12:47 AirOnSkin Can someone help me understand what's wrong here? I don't find my error: http://pastebin.com/t8wCHyJS
12:48 AirOnSkin First block is the state, second block is the highstate output
12:50 babilen AirOnSkin: You don't want the kwargs in there
12:51 illern joined #salt
12:52 babilen bersace: You do that in the top.sls file -- http://docs.saltstack.com/en/latest/ref/states/top.html#environments
12:52 illern joined #salt
12:53 denys joined #salt
12:54 illern joined #salt
12:55 AirOnSkin babilen: You're right. Now it works. Thank you :)
12:55 babilen np, fwiw, you can use user.present with password for that too
12:57 AirOnSkin I guessed that, but wasn't sure if name and password would suffice and if yes, if the other attributes of the user would change (homedir, shell, ...)
12:57 AirOnSkin That's why I wanted to go with the module, since it really only alters the password
12:58 favadi left #salt
12:59 babilen AirOnSkin: It would only change what you set (and differs)
12:59 surajd joined #salt
13:00 babilen Personally I'm not sure what I do prefer. You have a point in that the module call does exactly what you want and only that. That is, in itself, something to cherish.
13:01 babilen The root user is a bit speshul in that it exists already. We typically manage all our users with https://github.com/saltstack-formulas/users-formula (which includes passwords)
13:01 JDiPierro joined #salt
13:02 issel joined #salt
13:03 AirOnSkin Need to run to a meeting. Will answer afterwards...
13:03 JDiPierro joined #salt
13:03 issel hello!
13:05 issel I have a question about custom grains: can I use core grain values in custom grains? Would speed them up nicely
13:05 issel in my case.
13:05 jerematic joined #salt
13:05 jdesilet joined #salt
13:12 asaladin_ joined #salt
13:15 aphor has anyone ever been tempted to put together a minion initrd?
13:18 numkem how is 2015.02 going? Can we expect a release soon? I'm really eager to get my hands on the beacon system
13:19 babilen numkem: 2015.2 will be released in February 2015. For some reason we entered a time-delay-field and it is almost April, so you can expect 2015.4 / 2015.5 maybe? ;)
13:20 numkem babilen: that sounds like when you go in the Dominican Republic on vacation, their time is always about 45 minutes later than "normal time". So it's not that far away, thank you!
13:21 ksj joined #salt
13:22 babilen numkem: Saltstack made a release candidate available on the 11th of February and that is the last we've heard anything official about it.
13:22 micah_chatt joined #salt
13:22 numkem babilen: was thinking there is a release plan of some sort like what the Debian/Ubuntu folks does
13:23 repl1cant joined #salt
13:23 malinoff numkem, it's really hard to follow any plans when you have to explicitly specify which version will be the next which means that you have to pin the month when you will be releasing this new version
13:24 malinoff numkem, and if there will be bugs and the release will be re-scheduled, this will confuse users a lot
13:24 babilen The versioning scheme could be improved, yeah
13:24 malinoff idk what was wrong with semver
13:24 numkem malinoff: not sure I follow you but I do understand that if there is 2 versions that are "ongoing" it would be confusing
13:24 micah_chatt_ joined #salt
13:25 malinoff numkem, let me elaborate: imagine you have a plan to release 2015.6 as the next version
13:25 malinoff numkem, which means that you *have* to release it in June
13:25 timoguin joined #salt
13:25 babilen I mean you have two options: 1. Don't decide on a name until you know when you will release (i.e. make the "saltstack-release" branch "2015.4" when you release in April) or 2. Release in the month you said you would release no matter what
13:26 malinoff numkem, but if you cannot do that by any reason, you're fucked
13:26 malinoff numkem, because everybody is waiting 2015.6
13:26 babilen It doesn't help that saltstack chose the last versions based on totally different "events"
13:26 malinoff numkem, and you're releasing it in July
13:26 babilen 2015.2 is the "crystal ball" release name that has absolutely no basis in reality
13:26 babilen I predicted that last year, but meh
13:26 numkem I just feel like it's more confusing to have the name of the version go with a month but it isn't the month of the release but the month of the branch creation. I guess it's just that people globally are more used of they way that Ubuntu tagged their version names
13:27 babilen numkem: It isn't even the name of the branch creation for 2015.2
13:27 malinoff babilen, yeah, that's why people use semver as the main versioning scheme and put anything they want in the changelog, like the most distros
13:27 malinoff trusty, precise, etc
13:27 Tecnico1931 joined #salt
13:27 numkem but I understand what you mean malinoff, it would be just great to have a little ballpark when you expect to have a version that you look forward for such important features
13:27 malinoff numkem, exactly
13:27 babilen malinoff: There would be no problem to have a "release" branch in the repo that you turn into a YYYY.MM branch *when you release*
13:28 malinoff babilen, probably :) but all this hipster versioning basically means nothing
13:29 malinoff babilen, ok, it was released in 2015
13:29 malinoff so what?
13:29 numkem I agree with babilen, its just easier to understand. Not that the current way is wrong. Considering how important Salt is, it makes sens to have a stable version instead of running for a specific date
13:29 cheus_ joined #salt
13:29 babilen 2015.2 is neither the date of the release nor the date of the branch creation. It is the "we hope we'll release in February" -- This stands in contrast to the last couple of version numbers
13:30 malinoff numkem, unfortunately, salt is far away from 'stable' as the most people understand that word
13:30 babilen malinoff: I have no idea why they switched from 0.17 to 2014.1 -- I would have been fine with a 0.18 or a 1.0
13:30 Deevolution joined #salt
13:30 malinoff babilen, i have an answer, give me a sec
13:31 chandy joined #salt
13:31 malinoff http://saltstack.squarespace.com/salt-blog/2013/10/27/salt-version-numbers
13:31 malinoff "So we wanted to make a numbering and naming system that reflects the rolling nature of Salt"
13:32 malinoff but I can't agree that semver cannot handle the rolling nature and date-based versioning can
13:32 rhodgin joined #salt
13:32 numkem malinoff: I understand that salt is ever evolving, especially now but I can't say I've ran into bugs that would stop my progress (except when it comes to the docker modules/states but it's tagged unstable in the doc so thats ok).
13:33 malinoff numkem, because of bugs i don't use it for almost 8 months
13:33 cpowell joined #salt
13:33 malinoff numkem, and because they simply don't care about backwards compatibility
13:33 numkem malinoff: might I ask which bugs and what do you use instead?
13:34 malinoff numkem, oh, i really can't remember, there were lots of them and i sent a couple of PRs
13:34 lomeroe joined #salt
13:34 babilen malinoff: This isn't true and the last couple of releases were quite alright
13:34 malinoff numkem, but we were facing a new bug on almost every day
13:34 malinoff babilen, probably, but the trust is lost
13:35 malinoff numkem, we use ansible for now, it is far away from what we expect from such systems, but at least it works
13:35 primechuck joined #salt
13:36 malinoff and it's fucking slow compared to salt, it is extremely annoying
13:36 * babilen has been using saltstack in production for about a year now and had few problems. In particular the 2014.7 branch releases were quite good
13:36 malinoff babilen, how many customers you have?
13:36 mpanetta joined #salt
13:37 babilen malinoff: I fail to see how that is important
13:37 numkem malinoff: isn't this a little broad? Customers isn't really a precise measurement. Total number of minions maybe?
13:37 babilen (I also consider that sensitive information)
13:37 malinoff babilen, i'm not asking exact numbers
13:38 malinoff babilen, the biggest issue always was the backwards compatibility
13:38 malinoff our customers are rolling so fast so we can't simply follow salt upgrades
13:38 malinoff but we have to, because old master cannot communicate with new minions
13:39 malinoff and that's why we still have a few customers on 0.16.x
13:39 babilen malinoff: It is okay if you don't want to use salt. I understand that you've had issues with some releases, but there is no point in simply poisoning the athmosphere in this channel for the foreseeable future.
13:40 numkem Wouldn't it be where the architecture of salt/chef/puppet wouldn't match your business case malinoff?
13:40 numkem If you don't have the control over the software that your client uses, maybe a solution that is agentless or salt-ssh is the only way to go. Ansible fits a lot better your business case
13:40 malinoff babilen, i was asked, and i believe it is ok to talk not only about pros but also about cons
13:41 babilen malinoff: If there are specific issues you want users to know about then, please, make them aware of that, but the last couple of releases have been quite good and "salt doesn't care about backwards compatibility at all" simply isn't true
13:42 numkem I haven't had any game breaking problems with the 2014.7.x series
13:42 Romlok joined #salt
13:42 malinoff babilen, are you 100% sure i'll have no issues when controlling 2014.7 minions from 2014.2 master?
13:42 malinoff babilen, that's another point of using semver
13:43 perfectsine joined #salt
13:43 malinoff numkem, yes, probably, this architecture is not really what we need. But people use e.g. celery from 2.x series and their tasks just keep working on 3.x series.
13:44 babilen malinoff: You master *always* needs to be at least as new as your newest minions. That is completely normal.
13:44 malinoff unless they do black black magic
13:44 kusams joined #salt
13:44 egil joined #salt
13:45 kaptk2 joined #salt
13:45 malinoff babilen, it's not, imho, simply because when we take a new client on our support, we have a salt master already
13:45 yuhl_work_ joined #salt
13:45 numkem malinoff: how is that a good comparasion? The scope of Celery is far smaller than the scope of Salt
13:45 malinoff numkem, my point was about backwards compatibility
13:46 malinoff numkem, and celery is quite complex software
13:46 numkem malinoff: your point is that for your use case you cannot use that kind of software. Wanna talk about Puppet? What if you are running puppet 2.x and your clients rolls out with a Puppet 3.x client, won't work. Same thing
13:47 AirOnSkin babilen: About before. The root user is the only Unix Account we need to know. User login happens with LDAP, so no need to any other kind of user management on each box
13:47 malinoff numkem, of course! but, thanks to semver, i know that they won't work, 3 differs from 2
13:47 malinoff numkem, so all i need to do is to keep 2.x versions
13:48 numkem malinoff: isn't that what they did anyway? 2014.7 plays well witht he others from the same kind, going in with a 2014.1 is like the 3 vs 2
13:48 malinoff numkem, that's not so obvious for me, sorry, same for the rest of my team and for my customers
13:49 numkem I can't find anything to respond to that...
13:49 malinoff numkem, i'm actually talking about 0.x series
13:49 AirOnSkin Can someone tell me how to test if an SLS file is present with jinja?
13:49 numkem wan't that released like 2 years ago? Why would it still matter right now?
13:49 malinoff i couldn't upgrade my minions from 0.16.x to 0.16.x+1
13:50 malinoff numkem, as i said, the trust was lost :) don't take it to heart
13:50 iwishiwerearobot joined #salt
13:50 numkem malinoff: I'm not taking to heart, I just fail to see the logic
13:50 malinoff i don't try to refuse people from using salt
13:50 zwi joined #salt
13:50 malinoff numkem, it is ok too
13:51 numkem malinoff: tbh if someoen who never used it and just wants to walk in to see the community... they will turn around this morning
13:51 N-Mi_ joined #salt
13:51 malinoff numkem, i used it, as you may notice
13:51 malinoff and even tried to improve
13:53 subsignal joined #salt
13:54 ahale joined #salt
13:55 babilen malinoff: Surely it's normal, it can't be any other way. The master naturally has to be at least as new as the minions it controls. All I am asking you is that you try to either simply refrain from blanket statements such as the ones you tend to make or that you do, at least, be *very* specific about the issues you were facing.
13:55 babilen well, that's the same thing :)
13:56 malinoff babilen, i don't want to start a long conversation here, let's just assume I faced *some* issues before that made me unable to use salt at some moment
13:56 malinoff babilen, and this is my and only my experience which may differ from others
13:57 flyboy82 joined #salt
13:57 AirOnSkin So it turns out you can't check if files exist in Jinja... you need to do it in python
13:58 numkem AirOnSkin: using the py rederer makes things rather easy to do things like that.
13:59 AirOnSkin numkem: That will be the second step... since I don't really know python I first need to figure that out ... but thanks :)
14:00 JDiPierro joined #salt
14:01 mohae_ joined #salt
14:02 seanz joined #salt
14:04 I3olle_ joined #salt
14:07 AndreasLutro AirOnSkin: try {% if salt['file.file_exists']('/path/to/file') %}
14:11 adelcast joined #salt
14:12 pdayton joined #salt
14:16 AirOnSkin AndreasLutro: Oh, that looks like good. Can '/path/to/file/' also be a salt call? ({{ salt['grains.get']('id') }})
14:16 AndreasLutro AirOnSkin: yes, you don't need the {{ }}
14:16 rojem joined #salt
14:16 AirOnSkin And this will test if the file exists on the master or on the minion
14:16 AirOnSkin ?
14:17 AndreasLutro {% if salt['file.file_exists']('/path/to'~salt['grains.get']('id')~'/whatever') %} for example
14:17 AndreasLutro minion
14:17 rjhintz joined #salt
14:18 AirOnSkin AndreasLutro: I need to test the master though... wait, I'll show you
14:19 AirOnSkin this is part of my Pillar top.sls: http://pastebin.com/4fdkzGGM
14:20 rjhintz left #salt
14:20 AirOnSkin I'd like to write a test, that first checks if the file hostcfg/minion_id exists before it maps it
14:20 AndreasLutro hrm
14:20 aphor if you set up file_roots in a minion for masterless operation, do you have to delete the file_roots from the minion and restart to get it to try and register with the master?
14:21 AndreasLutro AirOnSkin: I don't know, I don't think that's possible
14:22 AirOnSkin AndreasLutro: Well, as numkern wrote before, I think I can write the test in Python and return a value that jinja can evaluate...
14:23 renoirb joined #salt
14:26 andrew_v joined #salt
14:26 thedodd joined #salt
14:28 clintberry joined #salt
14:28 iwishiwerearobot joined #salt
14:29 AirOnSkin babilen: Another consideration for using user.present instead of module.shadow would be, that module.shadow always runs, even if the password didn't change. I don't like that... I think I'll try out user.present
14:30 __TheDodd__ joined #salt
14:30 rojem joined #salt
14:31 carnedepassaro joined #salt
14:31 chandy joined #salt
14:33 debian112 joined #salt
14:33 otter768 joined #salt
14:33 subsignal joined #salt
14:36 ipmb joined #salt
14:37 Brew joined #salt
14:37 neogenix_ joined #salt
14:41 giantlock joined #salt
14:44 masterkorp babilen: hey, are you babilen on github ?
14:45 babilen masterkorp: I am indeed
14:45 masterkorp babilen: i just saw your pull requests against the influxdb formula (which i am the creator of)
14:46 masterkorp babilen: just saying i am looking into your pull requests
14:47 babilen masterkorp: Lets discuss it in a second - (on the phone)
14:47 masterkorp but i will be adding travis to the formula first
14:47 masterkorp babilen: sure
14:50 subsignal joined #salt
14:51 clmsy hey everyone
14:51 clmsy i am starting to train myself in salt
14:52 clmsy but i am stuck at the beginning unfortunately
14:52 clmsy under /etc/salt/
14:52 clmsy dir is empty
14:52 clmsy where should salt/roots/  and /srv/salt/
14:52 clmsy should exist ?
14:53 masterkorp did you just bootstrapped your machine ?
14:54 timoguin joined #salt
14:55 clmsy i wanted to fire up vagrant
14:55 clmsy to start learning with a masterless minion
14:55 clmsy i installed salt through pip
14:55 clmsy and this is my Vagrantfile
14:55 clmsy http://pastebin.com/LNrzFcgU
14:56 clmsy http://docs.vagrantup.com/v2/provisioning/salt.html
14:57 clmsy should i use bootstrap
14:57 clmsy even though i installed through pip ?
14:57 masterkorp well i use use salt-bootstrap is the fastest way to get salt on a machine
14:58 masterkorp both dev machines and actual infrastruture
14:58 clmsy ok
14:58 ndrei joined #salt
14:59 evle1 masterkorp: I believe for vagrant bootstrap is not an option.
15:00 masterkorp evle1: yeah it is
15:00 masterkorp better yet for testing formulas you can use kitchen-salt
15:01 evle1 masterkorp: I never used it myself but from what I read I assumed salt is already builtin into vagrant.
15:03 evle1 BTW, is there anybody who understand how salt-ssh works with ssh keys?
15:04 dude051 joined #salt
15:06 conan_the_destro joined #salt
15:09 teskew joined #salt
15:12 scbunn joined #salt
15:13 desposo joined #salt
15:13 hobakill joined #salt
15:14 s0lar joined #salt
15:15 scarcry joined #salt
15:15 babilen evle1: What do you mean?
15:15 babilen "it uses them"
15:16 clmsy masterkorp: evle1: just wanted to inform you guys that the vagrant has a salt provisioner in itself, and if the machine image you are booting doesn't have salt, vagrant bootstraps the machine upon launching
15:17 evle1 babilen: I'm trying to figure out how to give salt-ssh already existing key.
15:17 subsignal joined #salt
15:17 relidy joined #salt
15:18 evle1 $ salt-ssh -i --priv=~/.ssh/id_rsa --roster-file=salt.roster '*' test.ping
15:18 evle1 /home/evle/.ssh/id_rsa already exists.
15:18 evle1 Overwrite (y/n)? n
15:18 evle1 Somewhy it attempts to overwrite it but not use.
15:18 Corey Well that's interesting.
15:20 smcquay joined #salt
15:20 evle1 Also it seems salt-ssh seems to want to be run as root. It just has no permissions to write caches and stuff like that.
15:20 bersace left #salt
15:21 babilen Yes, that is discussed in the documentatino
15:21 babilen http://docs.saltstack.com/en/latest/topics/ssh/#running-salt-ssh-as-non-root-user
15:22 msciciel joined #salt
15:23 nesv joined #salt
15:23 StDiluted joined #salt
15:23 evle1 Ok. I see. Any idea about first one? It probably somewhere in documentation too but I can't find it.
15:24 _JZ_ joined #salt
15:24 CheKoLyN joined #salt
15:25 manfred you can make it run as any user
15:25 manfred you just have to configure everything right
15:25 manfred by specifying a different cache directory and stuff
15:26 manfred if you want, you can make a bug to get that stuff documented, and I will take a look to try and find my notes, or someone else might
15:26 subsignal joined #salt
15:26 manfred but that would be a good thing to get documented
15:27 bmac2 we are trying to use salt to automate our updates to our production servers, but we have a specific order that we need the servers to be rebooted.  Is there a way to tell salt what order to run the update and reboot?
15:27 havan joined #salt
15:27 evle1 I understand about config to run as different user (it still looks a bit off that it doesn't work by default considering scenario I'm trying to use it, but it's ok).
15:28 evle1 But I can't find about passing existing ssh-key.
15:28 havan joined #salt
15:29 Narolad joined #salt
15:29 phx bmac2, directly, i don't know the answer for that. however, if you take a closer look, salt's python interface is like a 1:1 mapping to the command line interface. this means, you can easily extend salt, and hook it into something else. basically you can treat saltstack as a framework and build upon it
15:29 AndreasLutro evle1: it wants you to use a separate rsa key for salt commands
15:29 phx bmac2, it would be like a couple of lines in python to go through all servers with salt for the updates, and reboot the successfully completed ones
15:30 phx you could also add a small scheduler to keep a given percent of the servers up, to avoid service interruption
15:30 bmac2 we had talked about just using python to do it, but we didn't knwo if there was a salt natie way
15:30 desposo joined #salt
15:30 bmac2 salt native way to do it
15:31 Narolad If anyone remembers my sysrq-trigger echo b question from yesterday, I tested it out today and got a python error, pasted here: https://gist.github.com/anonymous/c61ad6350ee91eb3c21c
15:31 evle1 AndreasLutro: Which separate key?
15:31 Narolad I made sure salt was able to access the file, and it let me do other file commands, just not to sysrq-trigger.
15:31 havan left #salt
15:32 SheetiS bmac2: I think orchestration can get you where you want to be.  At least close: http://docs.saltstack.com/en/latest/topics/tutorials/states_pt5.html
15:32 chandy joined #salt
15:32 AndreasLutro evle1: no key, yet, that's why it's trying to generate one. if you want to get started quickly, just replace id_rsa with saltssh_rsa or something
15:32 bmac2 thanks SheetiS, will read that
15:33 evle1 AndreasLutro: Generating new key is what I want to avoid.
15:33 AndreasLutro hmm, I don't know then
15:33 AndreasLutro not sure how it determines whether to overwrite or not
15:34 Fiber^ joined #salt
15:35 evle1 For two reasons.
15:35 evle1 1. I already have this key installed on a lot of servers, and to provision salt's key I would still need some way to access all them.
15:35 evle1 2. I want to know how where key is stored to be able to back it up and restore if needed.
15:35 SheetiS bmac2: if you need to upgrade all packages on a machine, you can do this in a state as part of the orchestration https://bpaste.net/show/71aabe387261
15:36 wnkz joined #salt
15:36 AndreasLutro evle1: all I know is that you can edit your salt master config and change the "pki_dir" where it stores its ssh key files
15:36 jalbretsen joined #salt
15:36 babilen evle1: fwiw, salt-ssh is working perfectly with my .ssh/config in place
15:37 babilen Not sure why --priv tries to *overwrite* your key rather than use it :-/
15:37 AndreasLutro actually evle1 try replacing ~ with /home/myuser
15:37 AndreasLutro because I can't get ~ to work for paths in my master config file
15:38 babilen No, ~/ is not working in this context
15:38 havan joined #salt
15:38 babilen (probably a missing call to os.path.exanduser())
15:39 AndreasLutro probably makes sense for actual master setups - there should be a separate ssh config :)
15:39 btorch has this been deprecated or something "salt['pillar.get'] ?
15:39 evle1 AndreasLutro: babilen: Well, ~/ seems to give a correct path to me as it writes correct filename in error message and sees file to overwrite.
15:40 iggy btorch: no, it never will be
15:40 btorch the "pillar" section on http://salt.readthedocs.org/en/latest/ref/modules/all/index.html has no doc
15:40 elfixit1 joined #salt
15:40 babilen btorch: That's perfectly fine, why do you thing that is problematic?
15:40 iggy btorch: if anything {{ pillar['foo'] }} should be considered less desireable (not deprecated though)
15:40 btorch iggy: cool I figure, I was trying to find some reference docs on it
15:40 _prime_ joined #salt
15:40 AndreasLutro btorch: possibly because it's documented so many other places
15:41 iggy btorch: readthedocs is dead (salt wise)
15:41 btorch oh
15:41 ek6 joined #salt
15:41 babilen http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.pillar.html#module-salt.modules.pillar also readthedocs is old
15:41 iggy http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.pillar.html#salt.modules.pillar.get
15:41 btorch thanks
15:42 ek6 iggy: how did your upgrade go?
15:43 btorch iggy: awesome thanks for the link
15:43 babilen :(
15:43 iggy it's tonight (after hours... because upgrading salt is "dangerous")
15:43 * iggy wins
15:44 ja-s joined #salt
15:44 * babilen adds btorch to the "do not help" list
15:44 btorch it is :) we just did 6 clusters totalling around ~500 boxes .. we finally left 0.17.5 behind :)
15:45 * Ahlee whistles taps
15:45 iggy in our situation, it isn't... I've already done the upgrade on our dev and QA clusters
15:46 iggy when we had highstate's running at startup it was a little dicey, but we don't do that any more
15:48 __gotcha joined #salt
15:48 __gotcha joined #salt
15:49 ek6 certainly seems to be binary between those that run super thin highstates and those that build space stations with them
15:49 iggy salt is getting popular... the channel peak has gone up 10 today
15:50 ek6 iggy: its you..animal magnetism and all
15:50 Narolad There's a lot of people sitting in the channel, but mostly the same names from yesterday.
15:51 iggy ek6: I wouldn't say that... I don't consider my highstates "space station" builders... but we do use a fair number of salt features (scheduler, reactor, etc.)
15:51 iggy it's always just me talking :(
15:51 Grokzen is pkg.installed not supported on SLES11 machines? I keep getting pkg.installed is not available when running highstate
15:51 ek6 iggy: i was thinking more in the column of 'average time to run'
15:52 iggy ek6: is that with or without changes?
15:52 ek6 iggy: without
15:55 iggy 20.515s... where does that put me on the "do nothing" vs "space station" scale?
15:56 Narolad I only rolled out salt into the environment on Monday.  I'm definitely still on the "do nothing" end.
15:56 ek6 iggy: nothing..... ive seen some folks at meetups where no changes highstates can take 10s of minutes
15:57 babilen o_O
15:57 theologian joined #salt
15:58 ajw0100 joined #salt
15:58 RedundancyD joined #salt
15:59 havan hello all, would anyone point me to the right direction, I got a few nodes with 2014.7.1 and when I run highstate from the master I only got return from only one node. I've checked minion logs, all seems to run fine and states are applied and master logs that he got return from the minions but it only show one node's returned data in the job output.
16:00 ja-s joined #salt
16:00 Narolad Do salt minions cache anything in particular? Salt minion is responding to queries with the exception of cmd.run
16:00 hal58th havan, you probably just aren't waiting long enough for replies? try adding "-t 500" to the end of the command havan
16:01 hal58th havan, you can use this to try and troubleshoot havan. http://docs.saltstack.com/en/latest/topics/jobs/#the-jobs-runner
16:01 ndrei joined #salt
16:01 havan hal58th: I tried -t 600 and it finished before 10 minute timeout finishes
16:01 micah_chatt_ joined #salt
16:01 hal58th brb
16:02 babilen havan: Have you looked at "salt-run jobs.active" or "salt-run jobs.lookup_jid $JID_OF_THE_JOB" ?
16:02 wendall911 joined #salt
16:02 babilen Can you ping all your minions?
16:03 havan babilen: I've looked at saltutil.running when I run highstate and all nodes report that they are running the highstate job and after a while they finish the job
16:03 havan yes I can ping them with test.ping
16:03 babilen And if you look at salt-run jobs.lookup_jid $JID_OF_THE_JOB you simply don't have any output even though they *did* apply the highstate?
16:04 havan let me check jobs.lookup_jid again, brb
16:05 iggy ek6: they are doing something very wrong
16:05 iggy well, or they just have 1000s of minions
16:05 iggy I only have like 100 minions
16:06 rojem joined #salt
16:06 babilen havan: I'll be heading off now
16:06 babilen (all the best!)
16:06 havan babilen: thanks
16:06 sjwoodr joined #salt
16:07 havan salt-run jobs.lookup_jid also show output from only one node
16:07 dalexander joined #salt
16:08 sjwoodr Is there a salt command we can use to view the output of scheduled jobs (salt.states.schedule)  ?  Since a job is a recurring thing on the matching minion(s) i'm not sure if there is a single jid associated with that...
16:08 hal58th joined #salt
16:09 iggy sjwoodr: I don't know of anything built-in (salt.states.schedule is a relatively new feature)
16:09 hal58th back now havan and I had to disconnect. Let me know if you have any more questions
16:10 sjwoodr iggy: Hmm, ok.  Is there any way to actually see if the job has run?  Other than appending to a file as part of the job itself...
16:10 havan hal58th: I've checked salt-run jobs.lookup_jid and it also shows only one node's output
16:10 rsimpkins joined #salt
16:10 \ask joined #salt
16:11 hal58th hmmm, that's not good havan... So you are sure that highstate is actually being applied on a node? Can you do something else like "cmd.run" instead?
16:11 rsimpkins Does anyone have any recommendations on how I solve this issue? 'master XYZ rejected the minions connection because too many minions are already connected'. What should I tweak in the master config to accept more minions?
16:11 iggy sjwoodr: with the default job cache, I don't remember seeing anything that would tell you that... some returners might record enough info to piece that together
16:11 scoates any idea how I might convince a salt minion that npm is indeed available? I'm getting `State npm.installed found in sls [redacted] is unavailable` even though /usr/bin/npm exists.
16:11 iggy havan: using anything weird (syndic, multi-master, etc.)?
16:12 havan hal58th: yes highstate is applied and cmd.run is running withour errors and I got the output of command from all nodes
16:12 havan iggy: no, onlt master and minions
16:12 mpanetta scoates: I get that sometimes with pip, restarting the minion always seems to fix it.
16:13 scoates bummer that I have to 2-phase it. /-:
16:13 hal58th havan, have you tried getting on a minion and running salt-call state.highstate?
16:13 redzaku joined #salt
16:13 mpanetta scoates: Yeah, maybe there is another way, but I don't know about one.
16:13 iggy rsimpkins: is that the exact error message?
16:14 sjwoodr iggy: ok i'll have to see if i can wire something up with a custom returner
16:14 havan hal58th: no I didn't tried that as the states are being applied on the minion side, I can give it a try
16:14 sjwoodr I wonder if i should be using salt.states.schedule or just use the good ol' dependable cron.present  :)
16:14 iggy scoates: whatever command installs npm needs to have "reload_modules: True" set
16:14 scoates iggy: oh? nice. trying that.
16:14 druonysus joined #salt
16:14 druonysus joined #salt
16:14 hal58th havan, I have had highstates hang halfway through when a service status hung up python
16:14 mpanetta iggy: Is that just something you put in the state SLS?
16:15 rsimpkins iggy: Yes. I figured it out. Someone set max_minions to 100 in our config. ;-)
16:15 iggy mpanetta: yes, it's a global thing (like require, unless, etc.)
16:15 mpanetta Ok cool
16:15 mpanetta I will remember this for future :)
16:15 iggy rsimpkins: silly "someone"
16:15 LotR joined #salt
16:16 desposo joined #salt
16:16 hal58th sjwoodr, if all  your minions are 2014.7, wouldn't hurt to use schedule.
16:16 iggy mpanetta: it's required on anything (other than pkg.install, which defaults to it) that installs dependencies for modules (so pip.install, cmd.run, etc.)
16:16 havan hal58th: I tried salt-call state.highstate on one of the minions, finished without problems and got the return, but when I check that jid on the master with salt-run jobs.lookup_jid it shows no ouput
16:16 iggy sjwoodr: same problem either way :/
16:16 mpanetta iggy: Hmm, ok good to know
16:17 hal58th rsimpkins, i didn't even know that was an option..
16:17 scoates iggy: thanks! worked great. /cc mpanetta
16:18 sjwoodr iggy: but with cron.present, at least i can check the crontab on the minion to see if its there ;)
16:18 hal58th sjwoodr: that is true. I don't like to see schedules run in minion debug only.
16:19 iggy sjwoodr: hmm, so newer (than 2014.7) salt-minions will drop /etc/salt/minion.d/_schedule.conf down with all your scheduled jobs
16:19 rsimpkins hal58th: This is for a larger deploy. We are having issues with scaling the particular version of Salt we have. Patiently waiting 2015.2 to hit the streets.
16:19 iggy sjwoodr: for 2014.7 I don't know how to verify things are being run (other than checking job logs/returners/etc)
16:20 rsimpkins hal58th: Hence - we've been tweaing the config like mad men.
16:20 iggy <3 2015.2
16:20 sjwoodr iggy - is that /etc/salt/minion.d/_schedule.conf on the master or the minion itself ?
16:21 iggy sjwoodr: minion
16:21 Edgan Can someone update github to show 2014.7.3 as the latest release? It is currently hiding under "Show 2 newer tags".
16:21 iggy (you wouldn't have /etc/salt/minion.d/ on the master... unless it was also a minion... which it should be)
16:22 sjwoodr well, i ask because i don't see that directory on the master or the minion.   2014.7.1
16:22 iggy Edgan: it's not "official" yet... you see salt has this weird habit of tagging shit, then waiting weeks for the distros to build packages from non-existent tarballs... only _then_ do they announce
16:23 iggy sjwoodr: yes... "newer (than 2014.7) "
16:23 iggy i.e. 2015.2
16:23 Edgan iggy: I notice the lack of release notes
16:23 sjwoodr lol ok... thought 2014.7.1 was > 2014.7 :)
16:24 iggy I'll be more clear next time
16:24 sjwoodr need 2015.2 to hit the street.  I manged to get a salt-cloud fix backported in there for ec2 that i need.  Currently i'm hand patching my master to fix that bug
16:24 Edgan sjwoodr: yeah, I have had to backport fixes
16:25 iggy live a little... install from git
16:25 Edgan sjwoodr: It is about to be 2015.4 :)
16:25 sjwoodr hah :)
16:25 iggy I have money on it being the 65th day of Feb
16:27 Kelsar joined #salt
16:27 hal58th rsimpkins, 100 minions isn't that large. sucks to hear that
16:27 rsimpkins hal58th: Oh no, that isn't where we got stuck. We got stuck at about 1000.
16:28 \ask joined #salt
16:28 hal58th ahhhh
16:29 Edgan sjwoodr: What is your ec2 issue?
16:29 gwmngilfen joined #salt
16:29 rsimpkins If someone out there has 1000+ (preferably 5000+) minions running on a master - I'd like to see your config! :)
16:29 kusams joined #salt
16:30 dopesong_ joined #salt
16:30 sjwoodr Edgan: https://github.com/saltstack/salt/issues/21434
16:30 Edgan sjwoodr: Side question, do you use a VPC?
16:30 gladiatr joined #salt
16:31 aparsons joined #salt
16:31 sjwoodr yes - but currently our master is not in VPC.   we're migrating out of ec2-classic into VPC this year
16:31 sjwoodr so the existing minions are in ec2-classic for the most part
16:32 Edgan sjwoodr: ah
16:32 Edgan sjwoodr: tested salt-cloud with a vpc? The impression I got it is didn't really support a vpc.
16:33 sjwoodr Edgan: No, both my production and my test salt-master are in ec2-classic
16:34 otter768 joined #salt
16:34 sjwoodr Edgan: you've had issues with salt-cloud and VPC?  or just have seen reports of it not working right?
16:34 Edgan sjwoodr: I am using a custom create-instance.rb written with the ruby sdk. Only reports found when I googled it recently and a while back.
16:35 sjwoodr Edgan: well if you are using  "ssh_interface: private_ips"  then it seems like it'd work OK.  In theory.  :)
16:35 sjwoodr i could try it out later today if I get caught up
16:36 ekristen I’m trying to enable granular log levels
16:36 ekristen but it breaks my gitfs stuff for some reason
16:36 kusams joined #salt
16:37 dalexander joined #salt
16:37 iggy rsimpkins: it all depends what you are doing... linkedin has well over 5000, but they don't use salt the same way a lot of people do
16:39 kusams joined #salt
16:41 havan left #salt
16:42 havan_ joined #salt
16:43 Gareth morning morning
16:43 Narolad Morning.
16:45 iggy I just generated stats from my irc logs... very interesting
16:46 \ask joined #salt
16:46 iggy but my logs don't even go back a year
16:47 havan hal58th, I've deleted the key of the minion that master show the only output and and then run highstate, now it show the output from another node but again only one of minions
16:48 iggy havan: what version?
16:48 havan iggy, 2014.7.1
16:49 iggy when 2014.7.3 is finally out, give it a shot (it's been tagged already, so waiting for packages to be built)
16:51 havan iggy, thanks I will give it a try
16:51 renoirb joined #salt
16:52 JDiPierro joined #salt
16:52 Karlthane_ joined #salt
16:56 chandy joined #salt
16:57 gladiatr does the loader support __init__() wrt external pillar modules?
17:00 masterkorp Guys
17:00 masterkorp with salt-call, can i get a non 0 exit code if a state fails ?
17:00 masterkorp i am trying to add CI to salt
17:00 hal58th_ joined #salt
17:04 iggy masterkorp: if you use state.sls I think
17:04 masterkorp iggy: https://travis-ci.org/seegno/influxdb-formula
17:05 masterkorp iggy: https://travis-ci.org/seegno/influxdb-formula/builds/55974191
17:05 masterkorp it did not seem to fail
17:05 iggy then I think you have your answer
17:06 \ask joined #salt
17:06 masterkorp but is there a way make it to ?
17:06 masterkorp that was my question
17:06 masterkorp i want to make the CI fail in case states fail like they did in that test
17:06 iggy I doubt it
17:07 iggy I especially don't know how that interacts with --local
17:08 masterkorp hmm
17:08 masterkorp that just sucks
17:08 aparsons joined #salt
17:08 havan iggy, I've installed 2014.7.3 under a virtualenv (only on the master node, minions are still 2014.7.1) and tested the highstate, it's still shows only one of the minions output
17:09 hal58th joined #salt
17:10 hal58th havan, so salt-call state.highstate worked fine on minions?
17:10 havan when the highstate job is running I've checked with jobs.lookup_jid, it shows the other jobs output while some other minion is still running the job and as far as I can tell from the master's log, it show only the last output he from any of the minions, what every minion finishes last
17:10 havan hal58th, yes
17:10 iggy havan: that sucks... I'm pretty sure the fix I was thinking of was master side only
17:12 havan it's very weird, only the last return the master got from the minions is shown in the jobs output
17:19 KyleG joined #salt
17:19 KyleG joined #salt
17:23 sandah joined #salt
17:25 havan iggy, hal58th: I've installed version 2014.1.0, when I run highstate with -t 600 I still got output from only the last minion, but this time when look up jid with salt-run I got the output from all the nodes
17:25 SaintAardvark rsimpkins: I'm curious what problems you're running into.  Considering salt at $WORK, and 1000 minions is not that much more than we would have if we switched.
17:25 havan this still only only on the master, minions are still 2014.7.1
17:26 rsimpkins SaintAardvark: msg me.
17:26 KyleG havan: Have you tried adding -v to the end of your highstate command?
17:26 KyleG It'll show you minions that aren't returning
17:27 havan KyleG: does it have any difference if I have it at the end? I run salt command as "salt -v -s -t 600 \* state.highstate"
17:27 masterkorp is https://github.com/nmadhok here ?
17:28 KyleG havan: I'm not sure, I just add -v at the end out of habit, it probably does not matter
17:28 iggy masterkorp: he doesn't irc
17:29 masterkorp iggy: well i need to talk to him
17:29 masterkorp is his email valid ?
17:29 masterkorp just sent him an email
17:29 iggy no clue, I've never tried to contact him directly
17:30 w32cih joined #salt
17:30 masterkorp he just merged 3 PRs i was reviewing
17:31 beneggett joined #salt
17:32 iggy so keep reviewing them... anything you note can be fixed in a follow up patch or the commit can be reverted if it's dire
17:32 masterkorp sure, but now your fork just diverged
17:32 masterkorp it makes hard to maintain both
17:33 iggy that's bound to happen
17:34 masterkorp ok, we will just not give back to the saltstack-formula
17:35 masterkorp bindly merging pull requests that are failing makes no sense to me
17:35 germs_ joined #salt
17:35 masterkorp better ping the person is maintaining the formula
17:35 masterkorp which is me
17:35 beneggett has the new 7.4.3 release been pushed to EPEL yet? I'm not seeing it and having python 2.7 vs 2.6 identity crisis
17:36 germs_ left #salt
17:36 * JordanTesting looks at KyleG
17:36 beneggett actually, that's not my issue
17:37 * KyleG looks at JordanTesting
17:37 * JordanTesting head nods
17:37 beneggett Amazon AMI updated to python 2.7, and the epel release of salt is forcing a python 2.6 install.. any pointers?
17:37 masterkorp me and babilen we talking about the pull requests
17:38 __gotcha joined #salt
17:38 crd I've got a question about requiring a python library for a state
17:38 crd specifically, I'm running into an issue with the docker-py library to use the dockerio state
17:38 masterkorp basepi: ping
17:40 Kelvin joined #salt
17:40 crd I require pip: docker-py and I see it get installed, but then my docker.pulled state fails because the new library clearly hasn't been loaded into memory
17:40 crd how do folks generally deal with this sort of thing?
17:41 Kelvin Hi there. I am having a hard time using the ipcidr "S@' to select what state.sls does
17:41 Kelvin we haave a test and a production subnet
17:42 Kelvin and depending on which subnet the server is in, salt needs to make different files
17:42 Kelvin and I am stumped by the docs so far
17:42 nesv beneggett: Have you considered installing salt using the script from here? bootstrap.saltstack.com
17:43 beneggett nesv: that's exactly what I'm doing.
17:44 nesv beneggett: You said "the EPEL release of salt is forcing a python 2.6 install"...
17:44 beneggett nesv: using the install script, and installing using system stable repos..
17:44 nesv beneggett: I suppose I should also ask: which distro you are using?
17:44 beneggett nesv: latest amazon ami, centos based
17:45 beneggett nesv: It's installing 2014.7.1, following call at: Using wget to install your distribution's stable packages:
17:45 beneggett nesv: in the distro, it has python 2.7 installed, but the salt install installs python 2.6 and tries to run files from both locations
17:46 chandy joined #salt
17:46 beneggett not files, commands from both locations
17:46 nesv beneggett: Weird. Which flags are you passing to the bootstrap script?
17:47 beneggett nesv: none ;) I'm just trying to set up a vanilla minion
17:47 nesv beneggett: FWIW, I usually call the script like so: sudo sh install-salt.sh -PUZ -A 1.2.3.4 git v2015.2.0rc1
17:47 nesv beneggett: Ah.
17:47 * iggy holds up 2015.2 release
17:47 nesv iggy: Rafiki-style?
17:47 micah_chatt joined #salt
17:48 gchao tusken raider-style
17:48 nesv Kelvin: Could you explain what you're trying to do a bit more?
17:48 beneggett nesv: getting cannot find module salt.scripts, because its looking in the python 2.7 directory, but it installed python 2.6, etc. etc.
17:48 iggy masterkorp: you sound like a child "It's not working the way I want so I'm taking my toys and going home"
17:49 nesv beneggett: I dunno. Maybe try uninstalling python 2.6?
17:50 \ask joined #salt
17:50 nesv beneggett: I've been running that command I pasted above on sparkly-fresh CentOS 7 boxes, without a hitch, and no Python 2.6 to be seen.
17:50 beneggett nesv: yeah, thats the weird thing, it never existed, but salt 2014.7.1 installed it as a dependency
17:50 beneggett nesv: I'll try your command there
17:50 nesv beneggett: Sweet. Let me know how it fares. :)
17:51 iggy masterkorp: yes it's annoying when things change underneath you, but there aren't "maintainers" of formulas... it's a team effort... and in case you hadn't noticed that team is pretty slim these days
17:51 wincyj joined #salt
17:51 iggy crd: reload_modules: True on your pip.installed state
17:52 hal58th_ joined #salt
17:52 basepi masterkorp: yo
17:53 crd iggy: ok, I'm not seeing that arg in the docs, but I'll dig into the code to confirm
17:53 iggy beneggett: is python 2.7 even an option (package wise) or did you install it by other means?
17:53 iggy crd: it's a global option, like require, etc...
17:53 iggy google it
17:54 crd iggy: oh great, thanks!
17:54 masterkorp iggy: i know that, i am not complaining, maybe i am being misleading
17:55 ja-s joined #salt
17:56 \ask joined #salt
17:56 masterkorp iggy: and that team effort how does it work ?
17:56 masterkorp that is my point
17:56 masterkorp there is no method
17:56 beneggett iggy: It came installed on the latest Amazon linux AMI distro, but its available through yum
17:56 beneggett iggy:  I'll look into whether its coming from EPEL or through amazon
17:57 crd iggy: http://docs.saltstack.com/en/latest/ref/states/#reloading-modules huge thanks!
17:57 iggy beneggett: so yeah, I don't think python 2.7 is a standard part of rhel/centos/epel... so the packages link to what is (2.6)... I bet that's your problem
17:57 masterkorp iggy: i just think that is not a good practice to merge stuff like this
17:57 masterkorp and we should organize a method
17:57 beneggett iggy: this looks like my problem: https://github.com/saltstack/salt-bootstrap/blob/develop/bootstrap-salt.sh#L2576-L2582
17:58 iggy masterkorp: there is a methood... it's just not what you want (or you don't understand it)
17:58 forrest joined #salt
17:58 masterkorp iggy: so how can I understand it ?
17:58 v0rtex I really like the file.line state committed 3 days ago (https://github.com/saltstack/salt/commit/6e332cb5c78a8fb985ce8b34eba7ac55d702a9d2). I'm assuming it will be awhile before that makes it into a salt release, right?
17:58 iggy v0rtex: I _think_ it got backported to 2015.2
17:59 forrest joined #salt
17:59 iggy sadly
17:59 v0rtex hmm...why sadly?
17:59 forrest joined #salt
17:59 smcquay joined #salt
18:00 v0rtex also, I don't see it on the 2015.2 branch
18:00 masterkorp where is the method written so I can read it and understand it ?
18:00 iggy sadly because they need to stop backporting stuff and get the release out the door (it's already very late)
18:01 v0rtex iggy: ahh, I definitely agree with that
18:01 v0rtex been waiting
18:01 v0rtex ...and waiting
18:01 gmoro joined #salt
18:02 nesv beneggett: On your Amazon AMI, what is the output of `lsb_release -si`?
18:02 iggy masterkorp: some here http://docs.saltstack.com/en/latest/topics/development/conventions/formulas.html and theres something else somewhere about the team that I can't find atm
18:02 masterkorp i keep in sync with that
18:03 beneggett nesv: lsb_release: command not found :(
18:03 beneggett nesv: do i need to run from a folder
18:03 kusams joined #salt
18:03 seanz left #salt
18:03 iggy beneggett: honestly, you're lucky if all it's trying to do is install some extra packages on centos5... for about 3 months it didn't work at all
18:04 kusams joined #salt
18:04 pppingme joined #salt
18:04 beneggett nesv, iggy:  so here is the issue, 3 days ago amazon updated their base linux AMI, in the release notes it is shipping with python 2.7, and all the system deps are using it: https://aws.amazon.com/blogs/aws/now-available-amazon-linux-ami-2015-03/
18:05 beneggett nesv, iggy: the bootstrap scripts install packages for python 2.6
18:05 iggy then probably what needs to happen is salt-bootstrap needs to stop treating Amazon linux as a RH derivative
18:05 iggy which means it's unsupported
18:05 iggy until someone decides to support it
18:07 beneggett nesv, iggy: but salt cannot run, because it looks for dependencies in the 2.7 directory, which installed into the 2.6 directory
18:07 \ask joined #salt
18:07 beneggett nesv, iggy: i think our options our patch the script for our needs until we can find a better LT solution, or just use an older version of their AMI that is still running on python 2.6
18:08 SheetiS beneggett: set python26 as your default with alternatives
18:08 SheetiS all the system stuf yum, etc does #!/usr/bin/python27 so it should be ok
18:08 beneggett SheetiS: yeah, that could work
18:08 iggy beneggett: yeah, that's a rather large change... salt-bootstrap should stop looking at Amazon linux as a derivative like I said
18:09 beneggett iggy: sure, makes sense
18:09 SheetiS Amazon linux is pretty much terrible
18:09 ajw0100 joined #salt
18:09 nesv beneggett: I only ask out of complete curiosity, but how come you went with Amazon Linux?
18:10 ja-s joined #salt
18:10 nesv I don't want to start a religious war in the slightest; it's a distro I don't know much (if anything) about, and I'm just curious as to why that distro over another.
18:10 SheetiS the 2015.03 AMI broke by spacewalk integration completely.  the package dbus-python is only built against python27 in it.  I'm probably going to build a python26 rpm and whatnot, but I think this wil be finally enough for me to get our organization to change to something different.
18:11 SheetiS nesv: once you actually start using Amazon Linux, you find it is a mistake.
18:11 beneggett nesv: sure, valid point; in general they are much better at keeping their dependencies up to date than centos
18:11 SheetiS Amazon tries to upsell it as perfect for their platform.
18:11 nesv SheetiS: hehehehehehe
18:11 beneggett nesv: they seem more progressive in that regard
18:11 nesv ...because the AWS buy-in isn't as immersive as it could be.
18:12 SheetiS and they try and upsell the fact that a distro upgrade is only a yum update away, but it makes for some things being very broken when they change the default system versions of things like python.
18:12 nesv beneggett: Fair enough.
18:13 SheetiS You can't really vagrant it locally (can kinda hack it to work).  You can't run it anywhere but in AWS.  I inherited ~100 Amazon linux VMS and I really want to switch it soon
18:13 SheetiS I might go with CentOS.  It would take less fiddling with my states and their map.jinja than going wiht something debian at this point.
18:14 nesv Typically, if I'm running into a situation where I need to put a newer version of something on a box, I used to just hack up an rpmspec and build the RPM myself, but lately, I've been using Docker for that purpose.
18:15 zmalone I dunno if I'd say they they are better at keeping their dependencies up to date.  Maybe in terms of major releases, but point release security stuff seems to get missed by Amazon in Amazon Linux.
18:15 baweaver joined #salt
18:15 beneggett SheetiS: yeah, good points; we can honestly echo all those same sentiments over the course of these last few months
18:16 zmalone I recall waiting a long time for some OpenSSL, httpd, and php updates as compared to the larger distributions.
18:16 zmalone (I have not used Amazon Linux in 12m+)
18:16 beneggett this latest OpenSSL one was almost immediate
18:17 zmalone I'd hope so, the announcement was publicly made 1.5 weeks before it was publicly released.
18:18 Edgan SheetiS: :)
18:19 SheetiS beneggett: I understand the desire for newer packages.  I'd be just as happy with ius communities repo or something for rhel and a better feel for what will break
18:19 \ask joined #salt
18:19 SheetiS Edgan: hello friend!
18:19 davet joined #salt
18:27 Grokzen joined #salt
18:29 shadowsun joined #salt
18:29 hackel joined #salt
18:29 nexsja^ joined #salt
18:34 benegget_ joined #salt
18:35 otter768 joined #salt
18:35 TyrfingMjolnir_ joined #salt
18:35 MTecknology Deploying mailman using salt states is an interesting task...
18:36 dyasny_ joined #salt
18:37 rojem joined #salt
18:38 micah_chatt joined #salt
18:39 sdm24 joined #salt
18:39 forrest MTecknology: You should write a formula for it :P
18:41 Edgan Anyone seen or written a dynamic roster for salt-ssh?
18:41 clinta joined #salt
18:41 forrest There's an open issue for it Edgan: https://github.com/saltstack/salt/issues/11774
18:42 MTecknology forrest: it's starting to look like that's not possible without creating an expect script :S
18:42 Edgan forrest: just found that
18:43 forrest Edgan: Yeah looks like cachedout is suggesting something like this: https://github.com/saltstack/salt/blob/develop/salt/roster/flat.py
18:43 krelo joined #salt
18:43 forrest MTecknology: well, clearly mailman is terrible software then right? ;P
18:43 Edgan forrest: yeah
18:43 btorch what's the correct way to use this output formatting module ? {{ salt['output'](salt['grains.get']('roles')) }}
18:43 MTecknology yup!
18:43 btorch even if I set a var and then give the var to the salt['output'](var) still  doesn't work
18:44 Edgan forrest: I could probably adapt the ansible ec2 dynamic inventory to act as a dynamic roster. It is the same idea, and already written in python.
18:44 iggy btorch: salt is a dict of modules... so you just tried to call a full module called output
18:44 forrest Edgan: That would be pretty sick.
18:45 maschinetheist joined #salt
18:45 btorch iggy: ok so I do have to call the full thing salt.output.txt.output(data)¶
18:45 btorch like salt['output.txt.output'] ?
18:45 maschinetheist hello - is there a way to change and keep the permissions of /etc/salt/pki?  Each time I run 'sudo salt-key', they get reset to root:root
18:45 iggy btorch: no... it's a dict of everything in salt.modules.*
18:45 iggy so outputters aren't going to be in there
18:46 sdm24 I'm having a strange issue. I have 2 different masters, "salt" and "salttest."  When I install salt-minion onto a windows machine, and I set the master to "salttest", the minion will send its key to the master. However, if I set the master to "salt", and restart the minion, it will not send its key to the salt master. Has anyone else experienced this?
18:46 iggy btorch: why don't you rephrase what you are trying to achieve... it sounds like you are going about things the wrong way
18:46 sdm24 if I set the master to the IP address of "salt", the minion will send its key
18:47 forrest maschinetheist: Not as far as I know, if you installed as root it chowns things. You might be able to do it with sticky bit settings? Or you could create an issue to see about an option being added so it won't chown stuff.
18:47 btorch iggy: instead of getting back the roles as a list and outputing it that way, I'm trying to get them back as if I were using the salt --out=txt
18:48 maschinetheist ah forrest..there was an issue in 2012 about this but never really took off
18:48 maschinetheist https://github.com/saltstack/salt/issues/2220
18:48 maschinetheist i will try the sticky bit
18:48 iggy btorch: nope
18:48 forrest maschinetheist: If that doesn't work and it's still an issue, open another issue, reference the old one, and explain the use case.
18:49 iggy btorch: what are you _really_ trying to do
18:50 __gotcha joined #salt
18:50 drawsmcgraw joined #salt
18:50 btorch I'm creating a motd.tail file and in that file I would have a line like "- Roles: {{ salt['grains.get']('roles') }}"  .. that's it
18:50 drawsmcgraw I feel like I'm missing something. Is there a way to include a state and pass context into it?
18:51 maschinetheist ya sticky bit doesn't work :(
18:51 maschinetheist forrest: will do
18:51 maschinetheist thank you!
18:51 forrest maschinetheist: Yeah np, hopefully it will be an easy fix.
18:51 maschinetheist yup
18:51 iggy btorch: try {{ salt['grains.get']('roles') | json }} ? (or | yaml or whatever)
18:52 forrest maschinetheist: If you really want to get it fixed quick you could always manually dig into the code, then do a PR when you have it figured out. I'd still open the issue though
18:52 Edgan http://fpaste.org/203383/27395925/   I can't use state.apply with salt-ssh?
18:52 maschinetheist i'll go through the proper channels :)
18:52 btorch iggy: aww the pillar stuff :) I need to find those docs again :)
18:52 iggy btorch: try {{ salt['grains.get']('roles').join(' ') }}
18:52 btorch iggy: I meant jinja
18:52 btorch iggy: cool let me se
18:55 dyasny_ joined #salt
18:58 btorch iggy: cool this did the trick - Roles: {{ ', '.join(salt['grains.get']('roles')) }}
18:58 btorch thanks again
18:59 iggy something like that
18:59 iggy glad you got it
19:00 SheetiS Edgan: state.apply is new in 2015.2.  Since that is a single state, try state.sls?
19:01 andrew_v joined #salt
19:01 iggy I've alredy complained about the docs on that function
19:02 iggy I guess nobody listens to me
19:02 SheetiS state.apply now says 'New in 2015.2' in the docs
19:02 SheetiS It didn't last time I saw the question though.
19:02 iggy oh good
19:02 * iggy grumbles about having to mention it 10 times a day
19:02 SheetiS it actually has a typo though.  It says state.apply_
19:02 Edgan SheetiS: Ah, google is giving me development docks.
19:03 Edgan They just need to release 2015.2 already :)
19:03 iggy Edgan: there's really only development docs... they are just supposed to version annotate everything
19:03 SheetiS docs.saltstack.com always include the latest development stuff.
19:03 forrest Edgan: Yeah we're still working on making it show different release versions. whiteinge have you had any opportunity to look at the docs showing development instead of stable by chance?
19:03 forrest I'm figuring no since I know the whole team is hilariously busy, but worth asking.
19:04 iggy hilariously busy... I'm not sure that's how they'd describe it
19:04 Edgan I got salt-ssh by just creating a top.sls and using state.highstate instead of state.apply
19:07 Edgan I can already see the dynamic roster will quickly become a must with ec2. People will branch the git repo, have out of date rosters, and wonder why they only got some instances.
19:08 forrest iggy: Well, so busy that it's funny to the rest of us? :P
19:09 pyrex joined #salt
19:09 iggy I'm not going to lie, I gained some perverse pleasure filing a 2015.2 blocker earlier
19:10 iggy but then I've got money on Feb 65th being the release day, so maybe I'm subconsiously trying to drag it out
19:11 baweaver joined #salt
19:12 thayne joined #salt
19:12 markm joined #salt
19:15 * signull_ beleives saltstack doesnt beleive in the month of March. Feb 65th it is!
19:20 Auroch joined #salt
19:21 denys joined #salt
19:22 cheus joined #salt
19:23 honestly that's what julius caesar should've done
19:23 honestly just abolished march
19:24 honestly "beware of the ides of march!" - "k, no more march"
19:27 cberndt joined #salt
19:29 dude051 joined #salt
19:29 iggy I wish it was just a dislike of March (and not that they miss pretty much every date they set
19:33 Ahlee how do you list running jobs again?
19:33 XenophonF joined #salt
19:34 baweaver joined #salt
19:36 XenophonF can salt's reactor fire off a highstate job in response to a git commit?
19:36 XenophonF like, right now i have to give my team root access to the master in order to be able to run highstates and whatnot
19:37 XenophonF i know i can probably lock that down via sudo or runner or something, but i'd rather not give them access to the master at all
19:38 refnode_ joined #salt
19:39 dude051 joined #salt
19:41 maschinetheist left #salt
19:42 iggy Ahlee: salt-run jobs.active ?
19:43 iggy XenophonF: I think the normal way is to use salt-api and have a (web)hook hit that salt-api url
19:44 Ahlee iggy: yeah, that was it
19:44 Ahlee i always skip over that one, being right on top 'n all
19:44 dopesong joined #salt
19:44 XenophonF iggy: thanks iggy, i'm reading the ec2-autoscale-reactor formula's sources now
19:44 XenophonF found a blog post by benjamin cane on salt-api as well
19:45 XenophonF http://bencane.com/2014/07/17/integrating-saltstack-with-other-services-via-salt-api/
19:47 zwi joined #salt
19:52 clehene joined #salt
19:53 clehene Is it possible to use a jinja template for /etc/salt/roster ?
19:55 iggy no
19:58 clehene thanks iggy, is there a way then I could have a global set of options for a roster file (so that they don't repeat with every host)? would the ansible inventory file be better for that effect?
19:58 iggy I don't know, I just know that nothing in /etc/salt uses jinja (jinja isn't initialized until well after the point where the config files are read
20:00 clehene iggy: got it. seems like this could be an option http://docs.saltstack.com/en/latest/ref/roster/all/salt.roster.ansible.html
20:00 iggy mayhaps... not a salt-ssh pro personally
20:01 bhosmer_ joined #salt
20:03 edrocks joined #salt
20:04 Edgan clehene: Are you in ec2?
20:04 ek6 joined #salt
20:05 enarciso joined #salt
20:05 clehene Edgan: no
20:05 dopesong_ joined #salt
20:14 carnedepassaro joined #salt
20:24 teskew1 joined #salt
20:27 sdm24 joined #salt
20:29 aparsons joined #salt
20:31 Ryan_Lane I wish there was a good EC2 roster
20:31 beneggett joined #salt
20:31 Ryan_Lane I'm still stuck on fabric till that happens
20:31 Ryan_Lane but I'm also too lazy to make one.
20:32 iggy Make one of your lackeys do it?
20:33 Edgan Ryan_Lane: Know how to write python? My plan is to adapt the ec2 ansible inventory to act as a roster.
20:33 Ryan_Lane I wrote most of the AWS modules for salt
20:33 iggy a generic script_roster would probably go a long way too
20:34 Edgan Ryan_Lane: Sounds like it would be even easier for you
20:34 Ryan_Lane I don't have a huge need for it right now, so haven't put any time into it
20:34 iggy he didn't say it would be hard... just that he was too lazy
20:34 Edgan haha
20:34 Ryan_Lane :)
20:34 toastedpenguin joined #salt
20:35 Edgan I will probably get to it sooner. I am going to use salt-ssh for deployments, and I am in ec2. So a dynamic roster will be great.
20:35 otter768 joined #salt
20:36 giantlock joined #salt
20:37 _ale1_ joined #salt
20:37 drawsmcgraw So... If you can include a state and pass data into it (a la 'context' with file.manage() ) I'm unaware. B-U-T I solved the problem with a Jinja macro.
20:37 drawsmcgraw jinja macros.... may our children forgive us.
20:38 Ryan_Lane Edgan: if possible try to use the execution modules to make the calls
20:38 Ryan_Lane I don't know if that's how salt-ssh is made, though
20:38 Ryan_Lane having better support in boto_ec2, for instance, would be nice
20:39 Ryan_Lane there's a few people working on the AWS modules right now, so anything you put in would likely get improved quickly after
20:41 ndrei joined #salt
20:42 cpowell joined #salt
20:44 beneggett joined #salt
20:45 andrew_v joined #salt
20:46 hal58th joined #salt
20:51 baweaver joined #salt
20:55 subsignal joined #salt
21:02 beneggett joined #salt
21:06 bash124512 joined #salt
21:07 druonysus joined #salt
21:10 LtLefse joined #salt
21:13 StDiluted joined #salt
21:15 baweaver joined #salt
21:16 dopesong joined #salt
21:32 refnode joined #salt
21:32 JayFK joined #salt
21:32 toastedpenguin joined #salt
21:42 kusams joined #salt
21:43 Ahlee Man.  Regenerating 500 UIDs went about as well as can be expected
21:43 Ahlee thank $DIETY for salt, else that _really_ would have gone sour
21:46 baweaver joined #salt
21:51 baweaver joined #salt
21:52 cheus joined #salt
21:52 ShibMcNe Hi, is there a way to filter reactor events by minion in such a way that only some minions can fire some events ?
21:52 ltsampro` joined #salt
21:55 iggy not that I've ever seen
21:56 robawt Heartsbane: you're my spirit animal today
21:56 robawt Heartsbane: re: JVM die in a fire
21:56 babilen i.want.beacons.
21:57 toastedpenguin joined #salt
21:57 ShibMcNe ok, so if I want to use reactors in a multi-tenant environment, having one master per client is a safer choice
22:02 juanito joined #salt
22:03 juanito g'day guys
22:04 hal58th ShibMcNe, that sounds like a good idea regardless. But I believe iggy is correct. I think someone asked that question at saltconf... Kind of insecure
22:04 juanito got a question regarding salt and pillar , i am trying to use wildcard in the top.sls for pillar by doing something like this
22:04 juanito base:
22:04 juanito 'ssh*':
22:04 juanito - users.*
22:04 iggy nope
22:04 juanito but it fails any way to include a wildcard for include or i need to list all files ?
22:04 egil joined #salt
22:05 iggy list all
22:05 hal58th you can't use wildcards in top.sls
22:05 ShibMcNe hal58th: iggy: thanks =]
22:05 juanito arggg
22:05 juanito http://docs.saltstack.com/en/latest/topics/releases/0.9.9.html 7.2.49.1.9. Wildcards for SLS Modules
22:05 hal58th or any sls, if i remember correctly. like using "include"
22:05 juanito so the doc is wrong
22:05 juanito ?
22:06 iggy that's the only mention of it I've ever seen, so they probably killed the feature
22:06 hal58th hmmmm, i've never seen that release note. As far as my testing has gone, it never worked...
22:07 hal58th that's a seriously old version of salt
22:07 juanito fair enough, it would be pretty usefull i reckon but if thats how it is then ill just deal with it ;)
22:07 juanito thx guys
22:07 hal58th welcome
22:07 hal58th also, next time use pastebin or something to paste an sls ;)
22:09 juanito true lazy me sorry ;)
22:11 iggy how did you even find that random release entry?
22:12 juanito google
22:13 juanito i am a lazy bastard and its just more convenient and i find it way too declarative to add all the sls i use coz i got a couple of other scenarios where I want to do that
22:17 mosen joined #salt
22:19 Andre-B joined #salt
22:19 iggy I think that was a bad april fools joke
22:19 seev joined #salt
22:21 dopesong_ joined #salt
22:22 juanito :(
22:22 juanito can we even create a directory structure for pillar coz i am trying to have a short top.sls file  for pillar
22:22 robawt juanito: yes
22:23 robawt pillar is just .sls files, just like states
22:23 juanito then inside the init.sls i can include files normmally ?
22:25 robawt juanito: pillar allows includes, but be weary about access.  if you target pillar to a.sls, but it includes b, you'll get a+b even though it's not targetting for b
22:26 timoguin joined #salt
22:26 juanito mm weird robawt
22:26 juanito http://pastebin.com/hLickN1D
22:27 juanito trying to include file but salt throw me an error
22:27 iggy include would probably haproxy-1.jean
22:28 stanchan joined #salt
22:29 juanito works
22:29 juanito sweet as, thx mate
22:29 felipe joined #salt
22:29 juanito thats a bit confusing tho you would expect it to be implicit
22:30 badon joined #salt
22:30 iggy not really... I think it's probably pretty common to include stuff from other "modules"
22:30 iggy i.e. the aptly-formula includes stuff from nginx.*
22:31 Heartsbane robawt: thanks
22:31 felipe can someone point me to a ELI5 explination of salt, please
22:32 llua "powerful dark magic; The end."
22:33 iggy what is ELI5
22:33 felipe Sorry, I should have specified that:
22:33 hal58th https://blog.talpor.com/2014/07/saltstack-beginners-tutorial/
22:33 SaintAardvark "Explain like I'm 5"
22:33 felipe ELI5 = Explain like Im 5
22:33 emostar i wouldn't want a 5 year old managing my servers
22:33 iggy "go take a nap now while you can"
22:33 felipe thanks hal58th
22:33 speedlight haha good one, gos to the dic of things like RTFM, KISS, ELI5 XD
22:36 otter768 joined #salt
22:38 felipe I'm not sure but I think its a term that was born on reddit and it even has it's own subreddit. www.reddit.com/r/eli5
22:38 thayne joined #salt
22:42 cberndt joined #salt
22:46 ajw0100 joined #salt
22:46 speedlight im comfused :( how i call a sls file in the init.sls, i dont want to use all directly in top to highstate
22:47 aquassaut joined #salt
22:47 speedlight i hv i.e: salt/app/init.sls, salt/app/{config.sls, server.sls, foo.sls}
22:48 speedlight i want to call salt 'minion01' state.sls app
22:48 speedlight and in the init call to config.sls etc
22:53 sdm24 speedlight: sounds like you need an include statement in the app .sls file
22:53 speedlight thnkz sdm24 thats the word! the docs are so big :D
22:53 sdm24 http://docs.saltstack.com/en/latest/ref/states/include.html
22:54 speedlight just get the same page :)
22:54 sdm24 no worries, I was dealing with the same issue. Nothing is worse than trying to google the wrong word
22:56 speedlight and 19837498759837594thousand pages of documentation XD
22:56 bfoxwell joined #salt
22:57 speedlight my base state is almost done! after that..... install salt-minion in ~100 hosts XD
22:57 sdm24 thats the fun part
22:57 speedlight i know... thats almost nothing.. XD
22:58 speedlight haha i hv a oneliner cmd at least..
22:59 speedlight install the repo, install the minion..
22:59 speedlight but do the same thing in chef was a nightmare haha
22:59 speedlight salt rulz
23:00 sdm24 yep
23:00 sdm24 I was trying to use ansible and was getting the worst errors with ssh keys, especially with windows
23:01 speedlight i just tested chef (the worst), puppet (too "enterprise" :) ) and salt
23:01 speedlight salt wins on all sides :D
23:02 speedlight well time to gome home..
23:02 sdm24 my boss tried puppet a while ago but it confused him too much, so I tested ansible and salt. He thinks salt is way easier to understand
23:02 sdm24 I have to wait for windows to finish installing updates. I had to boot up a windows VM template to install salt-minion, and windows threw in all the missed updates for free!
23:03 speedlight yes yaml is readable, if u want more hv templates, python, other templates..
23:03 speedlight good day ppl
23:05 mschiff Is there a good way how to dump the state sls instead of "executing" it? I want to debug jinja ...
23:06 sdm24 add test=True at the end of the salt call to test the state, would that help?
23:08 mschiff sdm24: not that does not help.. it does not show the resulting yaml data
23:09 sdm24 ah ok, sorry I'm not sure if theres another way. What I did to check jinja, when I was setting and importing a lot of variables, was include a state to file.manage a template that included all the variables
23:10 ajw0100_ joined #salt
23:10 khris mschiff: salt-call state.show_sls foo
23:13 mschiff khris: that helps a lot, thanks!
23:15 pcdummy joined #salt
23:15 pcdummy joined #salt
23:18 markm joined #salt
23:24 sunkist joined #salt
23:29 baweaver joined #salt
23:35 markm joined #salt
23:36 ahammond what is the correct ordering for updating / upgrading lib0mq on our production systems? We have a tested formula to do this, I just want to know master then slaves or slaves then master?
23:37 iggy are you changing major versions of 0mq?
23:37 jespada joined #salt
23:38 ahammond iggy I'll check, but I think we're going from 3 to the saltstack blessed packages that are v4
23:40 ahammond yup zmq 3.2.4 on master.
23:41 iggy I would personally not try that update via salt (maybe salt-ssh?)
23:41 markm joined #salt
23:41 ahammond somewhat insane mixture of 4.0.4 and other stuff on slaves. oh my!
23:42 ahammond in fact, our slaves look like they could be caught up to 2014.7 in some cases. FML.
23:43 kusams joined #salt
23:43 iggy minions
23:46 ahammond iggy sorry, been dealing with replication and build systems all day today :)
23:47 glyf joined #salt
23:53 desposo joined #salt
23:55 aw110f joined #salt
23:56 ahammond is there a good way to ask salt which minions are running a version other than 2014.7.2?
23:57 murrdoc joined #salt
23:57 iggy salt-run manage.versions
23:57 iggy well, that'll tell you versions of everything
23:57 murrdoc salt-run manage.up
23:57 ahammond thanks! I'll check it out
23:57 murrdoc and then walk out to lunch
23:57 ek6 iggy: just curious on your 15.2 can you get a clean salt-master --versions   ?  mine dies in a pile without libsodium installed.
23:57 murrdoc only way to do it
23:58 redzaku joined #salt
23:59 iggy you could also do something along the lines of salt -G 'saltversion:2014.7.1'

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