Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-05-18

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

All times shown according to UTC.

Time Nick Message
00:00 Antiarc joined #salt
00:02 julez joined #salt
00:21 cberndt joined #salt
00:24 kitplummer joined #salt
00:49 JDiPierro joined #salt
01:05 subsignal joined #salt
01:09 bhosmer joined #salt
01:13 c10b10 joined #salt
01:15 otter768 joined #salt
01:15 zekoZeko i'm trying to use rsyslog formula to log from all machines to a central syslog server. the syslog clients are setup via pillar '*': - company.syslog
01:15 zekoZeko and the syslog server is 'syslog.company.tld': - company.syslog-server
01:16 zekoZeko not the problem is that the pillar that is targeted to all minions sets rsyslog.target
01:16 zekoZeko how can I remove that from the pillar for the syslog server that must not contain that key?
01:17 zekoZeko or rather, how can i make the pillar merge not happen in this case
01:20 iggy zekoZeko: '* and not syslog.company.tld':
01:26 beauby joined #salt
01:28 kitplummer joined #salt
01:28 elfixit joined #salt
01:51 julez joined #salt
02:00 beauby joined #salt
02:09 gfa_ hello, i want to do something like this
02:09 gfa_ {% set qemu-version = ['{{ pillar['lsb_distrib_codename:qemu_version'] }}'] %}
02:09 gfa_ but is not working
02:10 gfa_ qemu_version is nested on lsb_distrib_codename value (trusty, precise, etc)
02:17 xMopxShell joined #salt
02:19 fxhp joined #salt
02:21 favadi joined #salt
02:27 kusams joined #salt
02:29 kitplummer joined #salt
02:29 evle joined #salt
02:30 beauby joined #salt
02:37 kitplummer joined #salt
02:41 murrdoc joined #salt
02:41 mosen heya murrdoc
02:42 murrdoc suuuuup
02:43 desposo joined #salt
02:50 cromark joined #salt
02:50 iggy gfa_: lsb_distrib_codename doesn't have anything underneath it
02:50 iggy gfa_: if your's does, try posting the output of your grains to gist
02:56 michelangelo joined #salt
02:58 badon_ joined #salt
03:02 c10b10 joined #salt
03:02 TyrfingMjolnir joined #salt
03:05 catpig joined #salt
03:07 TyrfingMjolnir_ joined #salt
03:08 beauby joined #salt
03:10 bhosmer joined #salt
03:13 heise joined #salt
03:14 gfa_ this is more correct, but still fails
03:14 gfa_ {% set qemuVersion = ['{{ pillar['{{salt['lsb_distrib_codename']}}:qemu_version'] }}'] %}
03:15 gfa_ http://pastebin.com/wPV0B4UN
03:15 otter768 joined #salt
03:34 iggy gfa_: don't put jinja inside jinja
03:36 iggy {% set qemuVersion = pillar[salt['lsb_distrib_codename'] ~ ':qemu_version'] %}
03:46 tkharju joined #salt
04:07 ethX joined #salt
04:20 fxhp joined #salt
04:32 kitplummer joined #salt
04:43 __number5__ wow, jinja-ception!
04:44 bhosmer joined #salt
04:46 kusams joined #salt
04:48 basepi MTecknology: that. is. amazing. It's also unavailable. :'(
04:50 c10b10 joined #salt
04:51 ramteid joined #salt
04:53 subsignal joined #salt
04:55 rdas joined #salt
05:11 bhosmer joined #salt
05:11 catpigger joined #salt
05:11 jhauser joined #salt
05:12 hemphill joined #salt
05:14 kermit joined #salt
05:17 otter768 joined #salt
05:19 Garo_ I'm having a weird problem with one of my masters (running 2015.5.0). If I try to command salt '*' test.ping on the master it reports "SaltReqTimeoutError: after 5 seconds" which ends up with error "Salt request timed out. The master is not responding. If this error persists after verifying the master is up, worker_threads may need to be increased"
05:22 joeto joined #salt
05:26 MTecknology basepi: for now... I'm sure others exist, but that was the first I saw and instantly (within pi seconds) thought of you.
05:27 MTecknology Garo_: are the minions the same version?
05:28 julez joined #salt
05:29 writtenoff joined #salt
05:31 MTecknology basepi: :)  http://www.houzz.com/photos/3143957/Pi-Symbol-Ice-Cube-Trays-eclectic-ice-trays-and-molds
05:32 MTecknology oh... nevermind :(
05:32 Garo_ MTecknology: yes they are. When I run the master with -l debug I can see that minions can connect to the master with the minion ping feature, but I see nohting when I try to run "salt '*' test.ping" from the master machine.
05:36 MTecknology Garo_: sorry, I gotta run off to bed while this storm is starting or I might not fall asleep at all :(   maybe play around a bit with raet?..
05:36 Garo_ MTecknology: no problem, sleep tight =)
05:38 kitplummer joined #salt
05:41 bash124512 joined #salt
05:43 thalleralexander joined #salt
05:47 stoogenmeyer joined #salt
06:00 colttt joined #salt
06:00 nikob joined #salt
06:01 nikob Anybody know where I can access the salt-cloud documentation formerly located here: http://docs.saltstack.com/en/latest/ref/clouds/all/salt.cloud.clouds.vmware.html ?
06:04 iggy check on salt.readthedocs.org?
06:06 iggy fwiw, I think the vsphere module is the one you're supposed to use these days
06:06 iggy wait, no, other way around
06:06 nikob yeah didn't have much luck with the vsphere module
06:07 nikob thanks for the link
06:07 gfa_ iggy: great. thanks!
06:08 iggy nikob: the problem would be that the vmware module is only available in devel (beryllium)
06:08 nikob yep I'm on devel
06:08 iggy and now the "latest" docs point to 2015.2 (instead of develop)
06:09 nikob but aren't I supposed to be able to change where the docs point?
06:09 nikob 2015.5.0 | 2014.7.5 | Develop
06:10 gfa_ iggy:  how did you know that? jinja documentation is sparse
06:10 gfa_ iggy: ^^ the reply you gimme before
06:10 gfa_ i mean, i read the docs on jinja site, salt docs, etc but things like that are nowhere documented
06:10 iggy gfa_: it actually fairly well documented... in it's own project
06:11 iggy it's there
06:11 gfa_ jinja.pocoo.org ?
06:12 iggy nikob: file a bug, looks like the develop docs are being generated using the directory structure of "latest" (so the vmware module doesn't show up)
06:12 iggy gfa_: yeah
06:13 nikob Thanks iggy, will do!
06:16 flyboy joined #salt
06:19 _2_toxicxlove joined #salt
06:20 malinoff joined #salt
06:27 cromark joined #salt
06:36 AndreasLutro joined #salt
06:37 gfa_ iggy: can you take a look to this? http://pastebin.com/xXSYN761
06:38 JayFK joined #salt
06:39 Grokzen joined #salt
06:40 harkx joined #salt
06:40 joeto1 joined #salt
06:44 golodhrim|work joined #salt
06:47 joeto joined #salt
06:48 AndreasLutro gfa_: what is salt['lsb_distrib_codename'] supposed to do? maybe you want salt['grains.get']('lsb_distrib_codename')?
06:49 AndreasLutro furthermore, you want to do salt['pillar.get'](salt['grains.get'](...)), and you should provide a default
06:49 gfa_ AndreasLutro: that set has been changed to {% set novaVersion = pillar[grains['lsb_distrib_codename'] ~ ':nova_version'] %}
06:50 AndreasLutro yeah I saw that after what I wrote. my second advice still applies though
06:50 gfa_ i will try that. i never did nested jinja before
06:50 AndreasLutro be aware that pillar['foo'], pillar.get('foo') and salt['pillar.get']('foo') are all different (you can replace pillar with grain)
06:51 AndreasLutro only salt['pillar.get'] will work with the nested 'foo:bar:baz' syntax
06:51 golodhrim|work hi guys, anyone here using salt to administrate Windows boxes? cause I have the problem atm that I can't set up a host definition so salt makes the changes by itself, and when trying to run "salt '*' state.highstate" it ends with an exception about c:\salt\var\cache\salt\master not existent. Any advice what I miss?
06:54 yexingok joined #salt
06:57 KermitTheFragger joined #salt
06:58 eseyman joined #salt
07:00 dkrae joined #salt
07:04 al joined #salt
07:05 favadi joined #salt
07:07 gfa_ AndreasLutro: thanks, btw you and iggy i got it working
07:07 gfa_ {% set nova_version = salt['pillar.get'](salt['grains.get']('lsb_distrib_codename') ~':nova_version') %}
07:08 gfa_ this will set nova_version acording to the value of the pillar, a nested value of ditrib: - nova_version
07:08 gfa_ and the best part is i got it :)
07:09 gfa_ i understand how and why
07:12 bhosmer joined #salt
07:15 denys joined #salt
07:17 julez joined #salt
07:17 otter768 joined #salt
07:20 monkey66 joined #salt
07:21 kawa2014 joined #salt
07:25 AndreasLutro awesome :)
07:26 ProT-0-TypE joined #salt
07:26 kitplummer joined #salt
07:35 Auroch joined #salt
07:49 fbergroth joined #salt
07:53 lb1a joined #salt
07:59 martoss joined #salt
08:04 CeBe joined #salt
08:07 julez joined #salt
08:07 kusams joined #salt
08:08 slav0nic joined #salt
08:16 aparsons joined #salt
08:17 ProT-0-TypE joined #salt
08:17 aparsons joined #salt
08:19 aparsons_ joined #salt
08:19 julez joined #salt
08:24 linjan joined #salt
08:26 c10b10 joined #salt
08:30 chiui joined #salt
08:35 aparsons joined #salt
08:36 aparson__ joined #salt
08:38 pipeep joined #salt
08:39 dopesong joined #salt
08:40 supersheep joined #salt
08:40 nlb joined #salt
08:42 gazarsgo joined #salt
08:42 LinuxHorn joined #salt
08:42 viq joined #salt
08:42 ITChap joined #salt
08:43 nene left #salt
08:43 wiqd joined #salt
08:44 dopesong_ joined #salt
08:46 dopeson__ joined #salt
08:46 codekobe joined #salt
08:49 aparsons joined #salt
08:51 grepory joined #salt
08:51 SpX joined #salt
08:51 morsik joined #salt
08:52 akitada joined #salt
08:53 supermike__ joined #salt
09:06 stephanbuys joined #salt
09:07 danielcb joined #salt
09:13 bhosmer joined #salt
09:13 mrbigglesworth joined #salt
09:14 huddy joined #salt
09:15 mrbigglesworth hey guys, i'm playing with salt and have a test env all up and running with a master and a minion. i have files and packages defined and it all works fine. i have a base class in a .sls file which i will apply to all my droplets. what is the best way to do inheritance?
09:15 kitplummer joined #salt
09:16 eliasp mrbigglesworth: http://docs.saltstack.com/en/latest/topics/targeting/
09:18 otter768 joined #salt
09:19 mrbigglesworth ahh it's all in the top file. sweet cheers
09:20 froztbyte https://github.com/saltstack/salt/issues/23845 :<
09:21 eliasp mrbigglesworth: http://docs.saltstack.com/en/latest/ref/states/include.html - include is used when there are dependencies between SLS or when it makes sense to have some kind of modularization and re-grouping…
09:21 mrbigglesworth cool thanks
09:25 dopesong joined #salt
09:26 zer0def joined #salt
09:40 ITChap joined #salt
09:48 ramteid joined #salt
09:53 vovcia joined #salt
09:55 vovcia hi o/ i have question about salt-cloud - im trying to provision amazon ec2 node, ec2 module works fine but doesnt install salt-minion
09:55 vovcia what am i missing? :)
09:56 stephanbuys joined #salt
10:05 cromark joined #salt
10:05 dRiN_ joined #salt
10:06 wvds-nl joined #salt
10:06 supersheep joined #salt
10:10 bhosmer joined #salt
10:15 Berty_ joined #salt
10:23 giantlock joined #salt
10:28 ndrei joined #salt
10:34 zer0def joined #salt
10:36 bhosmer joined #salt
10:42 mage_ hello, I have the following in my state file: https://dpaste.de/uQ66 it seesm to work more or less well except that I have some messages in stderr : https://dpaste.de/M45d any idea what could be the cause ?
10:48 robothands i dont know
10:52 riftman joined #salt
10:55 mage_ jcigar@dev:/usr/local/etc/salt/states/venvs/ruby/ > /bin/sh -c "typeset"
10:55 mage_ typeset: not found
10:55 mage_ should be that ..
10:55 saffe joined #salt
10:56 hvn joined #salt
10:56 hvn joined #salt
10:57 kusams joined #salt
11:02 zipkid Why are the saltstack-formulas, published on github, not organised to allow using them as-is, with git functionality available?
11:04 kitplummer joined #salt
11:04 AndreasLutro zipkid: why aren't they?
11:04 mage_ another stupid question: I have a SLS with two cmd.run, one should be run after the other one. How do you require: in the second one the first one ?
11:04 AndreasLutro mage_: cmd.wait with a watch arg
11:05 AndreasLutro cmd.wait documentation should have an example
11:05 mage_ thank you
11:05 zipkid AndreasLutro: Do you feel they are?
11:05 AndreasLutro zipkid: yes, though I haven't tried, I use them as examples, not as-is
11:05 AndreasLutro zipkid: I know some of them are buggy though
11:06 zipkid AndreasLutro: you can use the code, but you loose doc/example/git functionality...
11:06 zipkid AndreasLutro: sure, 'buggy' is not the problem i talk about here.
11:06 AndreasLutro define "git functionality"
11:06 zipkid being able to 'pull' updated versions 'later on'
11:06 zipkid without 'manual' interventions.
11:07 zipkid if you clone a formula as-is into file-roots it is not usable as the directory structure is not correct
11:07 Kraln joined #salt
11:08 AndreasLutro well, you could clone the formula locally then add the local repo as a gitfs_remote
11:08 AndreasLutro you could then merge/rebase any upstream changes
11:08 AndreasLutro best of both worlds
11:09 zipkid and how will you point salt to the correct dir IN the repo...?
11:10 zipkid hmm...
11:10 AndreasLutro don't think you don't need to do that
11:10 zipkid correct...
11:11 jvblasco morning everyone
11:11 zipkid but, that means it can only work 'seamlessly' witout using gitfs...
11:11 jvblasco i'm running in an issue while rendering the following pillar:
11:11 jvblasco https://gist.github.com/d517eef3160a81107893.git
11:11 jvblasco at some point salt sets the module salt.pillar to an empty dictionary
11:12 AndreasLutro jvblasco: if you remove .git from the URL people can actually see it in the browser
11:12 jvblasco which raises the following exception:
11:12 jvblasco https://gist.github.com/jvblasco/573d73dd5258a924aa2f
11:13 jvblasco ups
11:13 jvblasco AdreasLutro: ups, sry
11:13 jvblasco https://gist.github.com/d517eef3160a81107893
11:13 jvblasco there it goes
11:13 zipkid jvblasco: also, you can put multiple 'files' in a single gist...
11:13 jvblasco oh, my first time using gist
11:14 zipkid jvblasco: makes it so much easier...
11:14 inad922 joined #salt
11:14 jvblasco lemme create a new one with the pillar and the stacktrace
11:14 zipkid :-)
11:15 inad922 joined #salt
11:15 jvblasco https://gist.github.com/df3d72f4d8d2557e0a30
11:16 jvblasco there it goes
11:17 jvblasco the thing is that i'm using a python renderer to render a pillar, and there's someone at some point assigning an empty dictionary to the salt.pillar module
11:19 otter768 joined #salt
11:19 TyrfingMjolnir joined #salt
11:22 AndreasLutro jvblasco: odd... are you sure that pillar file is what's causing the error?
11:25 jvblasco AdreasLutro: when i use a pure yaml file, the exception doesn't rise
11:26 jvblasco AdreasLutro: so my guess goes for the python renderer as the main culprit
11:28 AndreasLutro hm.
11:28 jvblasco AdreasLutro: i haven't been able to replicate the error with pure yaml. I also tried to pipe that python renderer to the yaml renderer and return a yaml dump string instead of the python dict
11:29 jvblasco i always arrive, sooner or later, to that exception
11:29 AndreasLutro I can't find any examples on python renderer used with pillars online, but maybe you're supposed to do `__pillar__.merge(config)` instead of `return config`
11:29 AndreasLutro I don't know if pillar.merge is an actual method, just speculation
11:30 jvblasco worth a try, anyway i think the renderer requires you to return a dict
11:32 AndreasLutro actually no nevermind jvblasco
11:32 AndreasLutro this issue is way weirder
11:32 cromark joined #salt
11:33 AndreasLutro jvblasco: what is the full path to your py pillar file?
11:35 AndreasLutro jvblasco: also, what are you using to run this state? salt, salt-call, salt-ssh ?
11:36 keimlink joined #salt
11:42 jvblasco Andreas: /srv/pillar/customers/salt.sls is the path for the file
11:42 jvblasco AndreasLutro: /srv/pillar/customers/salt.sls is the path for the file
11:42 jvblasco AndreasLutro: and the command i'm using is just salt '*salt.test*' pillar.items
11:43 AndreasLutro jvblasco: try renaming it to something else - like customers_salt.sls
11:44 jvblasco AndreasLutro: i'll try, maybe it's making some name collision
11:44 AndreasLutro jvblasco: also, what are you using to run this state? salt, salt-call, salt-ssh ?
11:46 mage_ how do you handle database creation/update/... with salt ? is there a "best"/common approach ?
11:46 AndreasLutro mage_: there are mysql and postgres states that will handle it for you
11:46 linjan joined #salt
11:47 mage_ yep I've seen those :)
11:47 AndreasLutro you'll have to be more specific then
11:49 mage_ I want to be able to make tests like "if db exists then upgrade to schema xxx" "if not exist then run init.sql" "upgrade only postgis" etc
11:49 jvblasco AdreasLutro: oO. I can't replicate the error, it probably was a naming collision
11:50 AndreasLutro jvblasco: can't replicate even if you rename back to salt.sls ?
11:50 jvblasco AndreasLutro: Thnx so much for the help, i would have never think that i was forcing that with just a name
11:50 jvblasco AndreasLutro: lemme check
11:51 AndreasLutro jvblasco: if you can replicate you should report an issue on github
11:51 jvblasco AndreasLutro: yeah it begins to crash randomly again if i rename the sls
11:52 jvblasco AndreasLutro: where may i fill the issue?
11:52 AndreasLutro https://github.com/saltstack/salt/issues/new
11:52 AndreasLutro jvblasco: also for the third time - what client did you use? salt, salt-ssh, salt-call ?
11:53 jvblasco AndreasLutro: salt
11:53 cromark_ joined #salt
11:53 yuhl_work___ joined #salt
11:53 AndreasLutro ok not a salt-ssh caching issue then. submit that issue, that's a genuinely weird bug
11:54 yuhl_work___ left #salt
11:54 AndreasLutro mage_: I usually leave that bit out of salt - usually apps that use the databases have some sort of migration script setup that will manage schemas etc
11:55 jvblasco AndreaLutro: yeah, it sure is, 2 days of headaches
11:55 jvblasco AndreasLutro: Thnx so much for the help ;)
11:55 AndreasLutro no problem
11:56 mage_ AndreasLutro: right .. maybe I shouldn't put everything in salt
11:58 malinoff mage_, this is quite general principle - you can't manage unmanaged content. In other words, if there is something A that manages something B, you should't manage this B by yourself (using salt, ansible or whatever)
11:58 fredvd joined #salt
11:59 ITChap joined #salt
12:01 julez joined #salt
12:06 evle1 joined #salt
12:07 bhosmer joined #salt
12:10 ITChap joined #salt
12:25 dopesong_ joined #salt
12:28 ferbla joined #salt
12:30 tmclaugh[work] joined #salt
12:35 bhosmer joined #salt
12:46 kusams joined #salt
12:47 SneakyPhil left #salt
12:50 kitplummer joined #salt
12:51 julez joined #salt
12:57 JDiPierro joined #salt
12:59 Nemo_ joined #salt
12:59 AxelFooley Hi guys
13:00 AxelFooley could someone help me with a pillar related problem?
13:00 PI-Lloyd shoot
13:00 AxelFooley i've a pillar like this
13:00 subsignal joined #salt
13:00 AxelFooley http://paste.debian.net/179552/
13:00 AxelFooley and a state like this
13:01 AxelFooley http://paste.debian.net/179554/
13:01 sieve joined #salt
13:01 sieve Hi
13:01 AxelFooley now, i know that my state is completely wrong
13:01 sieve Wierdly salt-cloud is not available on this host
13:01 sieve [root@master minion]# salt --version
13:01 sieve salt 2015.5.0 (Lithium)
13:01 AxelFooley what i need from you, is an help to create a state that iterates through that pillar
13:01 sieve I thought salt came with salt-cloud now
13:02 AxelFooley to set up custom grains on minions via salt master
13:03 AxelFooley basically i'm unable to iterate through the "roles" pillar for getting every single value
13:06 mapu joined #salt
13:06 emaninpa joined #salt
13:07 AndreasLutro AxelFooley: the problem is you're defining - value: {{ g }} more than once, they'll overwrite eachother - I think you can just drop the for loop and do - value: {{ grain | json }}
13:08 PI-Lloyd sieve: salt-cloud is a seperate package afaik
13:09 PI-Lloyd it's in the main code base, but has it's own package, same as minion, master, syndic, etc etc
13:09 linjan joined #salt
13:10 vovcia im trying to use salt-cloud too and it doesnt install minion
13:10 vovcia instance is configured, i can see the key in salt-key on master but no minion daemon
13:11 supersheep joined #salt
13:13 jdesilet joined #salt
13:13 riftman joined #salt
13:18 supersheep joined #salt
13:18 yuhl_work___ joined #salt
13:18 PI-Lloyd vovcia: try adding "deploy: True" to your profile, and also the "script: bootstrap-salt" option. Deploy flag should default to True afaik, but doesn't hurt to add it to be sure
13:19 AxelFooley AndreasLutro: didn't work
13:19 PI-Lloyd vovcia: in fact, both of those options shouldn't need to be set, but try it anyway to be sure
13:20 otter768 joined #salt
13:20 cpowell joined #salt
13:24 kitplummer joined #salt
13:24 BinaryKhaos joined #salt
13:24 primechuck joined #salt
13:26 BinaryKhaos left #salt
13:27 _prime_ joined #salt
13:29 mbrgm joined #salt
13:30 c10b10 joined #salt
13:30 mbrgm hey! I'm trying to configure an interface using the state.networking, but it seems to have a problem with interfaces that have not been 'ip link set dev eth1 up' before. is there a way to accomplish this with the state?
13:31 vovcia PI-Lloyd: no luck :(
13:32 PI-Lloyd which provider are you using?
13:32 vovcia ec2
13:32 vovcia and amazon ami linux
13:32 kaptk2 joined #salt
13:33 PI-Lloyd hmm, not sure on the ec2 provider/profiles as i use GCE
13:34 PI-Lloyd trying to deploy an instance on GCE atm through salt-cloud and I'm running into issues myself... not good :(
13:36 bhosmer joined #salt
13:38 favadi joined #salt
13:39 Tyrm joined #salt
13:41 ageorgop joined #salt
13:41 bibin joined #salt
13:41 bibin Hi there, I am having a minion with the name "rs_$name". I am unable to remove it using the salt-key -d command. I think the $ sign is the villain. Please help  me out
13:41 racooper joined #salt
13:42 bibin It is showing "The key glob 'rs_' does not match any accepted, unaccepted or rejected keys"
13:42 AndreasLutro bibin: you probably just need to escape the $ on the command line, or use single quotes around the minion name
13:42 AndreasLutro i.e. `salt-key -d 'rs_$name'`
13:43 bibin That worked. Thanks.
13:43 MTecknology iggy: I think I made up my mind. My presentation was so bad and it's showing up so high on search results for my name that I absolutely need to give another to cover up that one.
13:46 Sathya joined #salt
13:46 Sathya Hi all
13:46 Sathya I am new to salt
13:47 Sathya I have close to 79 minions connected with my master
13:47 iggy MTecknology: ditto
13:47 Sathya Now Iam trying to group them using nodegroups in master config file
13:48 Sathya this is what i have done on the file:nodegroups: group1: 'S@10.X.XX.X/24'  group2: 'S@10.XX.XX.X/24'
13:49 Sathya when i restart the salt-master, i get the below mentioned error message
13:49 timoguin joined #salt
13:49 Sathya Starting salt-master daemon: [WARNING ] Key nodegroups with value None has an invalid type of <type 'NoneType'>, a dict is required for this value
13:49 Sathya Can some one tell me what is the mistake ?
13:50 Sathya group1 & group2 are different subnets
13:50 AndreasLutro Sathya: can you put the nodegroups bit up on a pastebin?
13:52 Sathya sorry what is pastebin ? r u refering to pastebin.com site ?
13:52 PI-Lloyd Anyone have any idea on this salt-cloud error? Works fine in 2014.7, not in 2015.5.0 - https://www.refheap.com/efd64a598a1be360e3cd776d9
13:52 AndreasLutro Sathya: https://gist.github.com or https://bpaste.net/
13:52 faliarin joined #salt
13:52 Sathya Thanks, just a sec
13:52 Sathya will do that
13:54 AxelFooley FYI: finally i ended up using this: http://paste.debian.net/179586/ and it's working
13:54 AxelFooley thank you a lot anyway
13:55 Sathya AndreasLutro : https://bpaste.net/show/62ae1284f6d9
13:55 AndreasLutro Sathya: the indentation is wrong - you need two spaces before each group. you should read up a bit on yaml
13:56 Sathya Got it :)
13:56 Sathya Let me change those and update
13:57 vovcia i still have no luck with salt-cloud :(
13:57 vovcia no minion installed
13:58 mbrgm how can I restart a service in a state run?
14:01 vovcia PI-Lloyd: do You have any more suggestions? :)
14:04 AndreasLutro mbrgm: look up the module.run state, you can call service.restart through that
14:04 murrdoc joined #salt
14:05 Sathya AndreasLutro @ it works thanks for your help
14:06 JDiPierro joined #salt
14:07 murrdoc iggy:  the saltstack-formula repo needs a 'recommended way to pull request' doc
14:07 murrdoc not sure where to put it
14:10 PI-Lloyd vovcia: not at the moment, trying to figure out why our staging saltmaster can't provision anything via salt-cloud :/
14:10 bibin joined #salt
14:11 favadi left #salt
14:12 andrew_v joined #salt
14:13 wendall911 joined #salt
14:14 debian112 joined #salt
14:14 mbrgm AndreasLutro: thank you! I found that I can combine it with onchanges: as well :)
14:15 amcorreia joined #salt
14:15 kusams joined #salt
14:17 iggy murrdoc: I'm working on a "committers" doc right now, "contributors" doc is next on my TODO
14:17 iggy mbrgm: if you mean the salt-minion service, look at how the salt-formula does it
14:18 murrdoc iggy:  where will u host it
14:18 murrdoc can u make squash merges a requirements
14:18 iggy murrdoc: undetermined
14:19 iggy for the contributors doc, sure
14:19 murrdoc also sensible pull request description
14:19 murrdoc something like
14:19 murrdoc **Summary of Changes**
14:19 iggy should probably even put a git rebase mini-howto in there
14:19 murrdoc BLURB
14:19 murrdoc **Testing**
14:19 murrdoc BLURB
14:20 bhosmer joined #salt
14:23 iggy murrdoc: https://gist.github.com/iggy/cb800f3401a6364cd5bd
14:23 iggy that's the one for the "committers"
14:23 iggy so far anyway
14:25 murrdoc updated
14:26 murrdoc or commented
14:26 hasues joined #salt
14:27 scoates joined #salt
14:28 iggy I think that better goes in the "contributors" doc, and we just put something like the following in the "committers" doc: "Thou shalt enforce the rules set out for contributors in their docs"
14:29 iggy but I need to do that doc first before I can tell people to follow it ;)
14:29 riftman joined #salt
14:29 hasues left #salt
14:29 murrdoc si
14:29 iggy but I do think we'll have to be a bit more... explicit for the contributors doc
14:30 iggy i.e. thou shalt squash commits... and here's how
14:34 mage_ the answer if probably "no", but: can I use the "salt://xxx" in a cmd.run with an arbitrary command ?
14:34 mage_ for ex https://dpaste.de/4Nxd#L10
14:35 mage_ it looks like it doesn't work as expected
14:36 giantlock joined #salt
14:36 FRANK_I joined #salt
14:36 FRANK_I Do you guy know if salt support ntp on linux?
14:37 FRANK_I ntp management.
14:37 PI-Lloyd FRANK_I: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.ntp.html
14:38 FRANK_I PI-Lloyd I am looking for linux this one is Windows only right?
14:39 PI-Lloyd ahh balls, didn't spot that one... you are correct by the looks of thing
14:40 Sacro FRANK_I: support/manage how?
14:41 XenophonF joined #salt
14:41 Norrland Hm, when trying to partition a disk I get this error: http://6eaae4a499a8e45c.paste.se/
14:42 Norrland running the 2015.5.0 version on both minion and master
14:42 murrdoc iggy:  you seen the 'lock pull request' button
14:42 XenophonF hey babilen, would you please take a look at my pull request for mysql-formula? https://github.com/saltstack-formulas/mysql-formula/pull/71
14:43 FRANK_I Sacro I am looking for the linux version of this http://docs.saltstack.com/en/latest/ref/states/all/salt.states.ntp.html
14:43 pelzi__ umm, just write a state to install ntpd and configure it
14:44 jlhammons joined #salt
14:44 XenophonF check out the ntp-formula - that's what i use
14:44 racooper looks like http://docs.saltstack.com/en/latest/topics/releases/ needs to be updated with 2015.5 release notes
14:44 primechuck Has anyone deployed saltstack as service monitoring?  Could the watch be used to send an event when a service fails?
14:45 Tremad87 joined #salt
14:46 bhosmer joined #salt
14:46 babilen XenophonF: Why me?
14:46 jlhammons left #salt
14:46 FRANK_I pelzi__ I do install ntpd with my kickstart I just want to set the server.
14:47 XenophonF babilen: i was hoping you had commit access to that repo
14:47 jlhammons joined #salt
14:47 pelzi__ so just use a state.
14:47 Tremad87 I am writing an external pillar python script and I would like to only use the pillar data returned by this script. Is it possible to control or overwrite the existing pillar data ??
14:47 XenophonF babilen: sorry, misread the  commit history for the repo
14:48 manfred FRANK_I:  set the server in a state using a file.append, or file.replace.
14:48 ITChap joined #salt
14:48 babilen XenophonF: No problem. It looks alright, but could you add it to pillar.example too?
14:48 XenophonF babilen: definitely
14:49 _JZ_ joined #salt
14:49 Brew joined #salt
14:51 bhosmer_ joined #salt
14:51 XenophonF babilen: done
14:52 babilen XenophonF: I would have rebased it into the previous commit, but fair enough. Ta.
14:52 XenophonF babilen: sorry - git newbie here
14:52 Brew joined #salt
14:52 XenophonF i added the change and pushed before i saw your comment
14:54 babilen That's okay. Another thing you do is that you change "connection_user: root" to "connection_user: 'root'". This makes it different from a bunch of other states (e.g. file.managed), but is probably not a bad idea.
14:55 babilen ooh, I like the new http://docs.saltstack.com/
14:55 babilen fancy
14:55 peters-tx joined #salt
14:58 Norrland is it possible in lvm.lvcreate to specify 'size' in percent?
14:59 rideh joined #salt
15:01 Tremad87 anyone?
15:01 Tremad87 Everything else works, I just don’t want to show any of the normal pillar data.
15:01 _prime_ joined #salt
15:03 XenophonF babilen: thanks for the merge, and thanks for the feedback
15:06 robot_hands joined #salt
15:07 rm_jorge joined #salt
15:12 ProT-0-TypE joined #salt
15:14 Auroch joined #salt
15:14 conan_the_destro joined #salt
15:14 babilen XenophonF: You are welcome
15:14 babilen (and ta!)
15:15 Auroch joined #salt
15:19 cromark joined #salt
15:19 mage_ is there a function to "evaluate" a salt:// source ?
15:20 babilen What do you mean by "evaluate" ?
15:20 mage_ babilen: https://gist.github.com/silenius/45118eca2a6af4964796#file-gistfile1-txt-L38
15:20 mage_ I can't use salt:// in this case and I would like to convert the salt:// to a "true" location
15:21 otter768 joined #salt
15:21 subsigna_ joined #salt
15:21 babilen mage_: So ... just reference a local file? (or a file hosted somewhere else)
15:22 babilen (and file.managed that file)
15:22 mage_ mmmmh eventually
15:24 riftman joined #salt
15:25 cromark_ joined #salt
15:25 racooper just to confirm before I upgrade...2015.5.0 master is backward compatible with 2014.7.5 minions, correct?
15:26 murrdoc yes on communication
15:28 cpowell so I see that cmd.run has a reload_modules arg, does pkg.installed support it too? I can't find anything in the docs
15:29 kartik_subbarao joined #salt
15:31 kartik_subbarao Is anyone here familiar with the augeas state/module? I wanted to inquire about this issue: https://github.com/saltstack/salt/issues/6994#issuecomment-102669619
15:32 XenophonF cpowell: i think pkg.installed automatically  does a reload
15:32 cpowell XenophonF: thank you
15:32 XenophonF cpowell: double check the state source code though, i've slept since the last time i read through it
15:36 kartik_subbarao It looks like augeas.change is returning a diff in string form, but highstate is expecting a dict, which causes it to print "Invalid Changes"
15:39 linjan joined #salt
15:40 treborTech joined #salt
15:41 I3olle joined #salt
15:42 mbrgm left #salt
15:49 sl_ joined #salt
15:50 rsimpkins joined #salt
15:51 aparsons joined #salt
15:53 murrdoc new salt doc sucks
15:53 murrdoc there are no local to the page links anymore
15:54 iggy cpowell: reload_modules is a global option
15:57 treborTech @murrdoc can you explain that? I’m not sure what you’re talking about local to the page links?
15:57 babilen murrdoc: What do you mean by "no local to the page links! ?
15:57 babilen Ah :)
15:57 treborTech :-)
15:58 bhosmer joined #salt
15:58 murrdoc http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html
15:58 murrdoc so the various functions  on the page
15:59 babilen http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.copy like that one?
15:59 murrdoc those used to be linked on the right
15:59 murrdoc as a table of contents
15:59 babilen they are
15:59 babilen (still)
15:59 iggy murrdoc: they are working on it
15:59 murrdoc hmm not there
15:59 murrdoc k
15:59 treborTech That’s what I’m seeing as well
15:59 treborTech hmm
15:59 whytewolf http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.accumulated
15:59 babilen That must be specific to your setup/browser/configuration. I see them just fine.
15:59 whytewolf they are there
16:00 treborTech I actually like the new way they are listed
16:00 babilen So do I
16:00 murrdoc where are they listed
16:00 babilen We can also finally switch between different versions easily
16:00 babilen murrdoc: On the right
16:01 iggy http://docs.saltstack.com/en/latest/ref/clouds/all/salt.cloud.clouds.gce.html#module-salt.cloud.clouds.gce doesn't have them
16:02 iggy http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.file.html#module-salt.modules.file does have them
16:02 babilen murrdoc: http://i.imgur.com/FEntAC1.png
16:02 iggy *doesn't
16:02 babilen iggy: Sure they do
16:02 whytewolf i think the unselected icon has issues. cause i don't see the icon but if i hover over where it should be the link shows up
16:02 iggy not on Chrome
16:02 bishy999 joined #salt
16:02 babilen iggy: I use chromium and it works just fine (see imgur link)
16:03 murrdoc babilen:  its entirely possible it doesnt work for us in chrome
16:03 babilen And I just checked in Chrome (works there too)
16:03 murrdoc and it works for u in chromium
16:04 paultjuh joined #salt
16:04 Gadzooks joined #salt
16:04 babilen chromium 42.0.2311.135-2 and chrome 42.0.2311.152-1
16:04 paultjuh hi does anyone knwo how to order sls in a explicit way?
16:05 babilen paultjuh: What are you really trying to do? Are you familiar with: http://docs.saltstack.com/en/latest/ref/states/requisites.html ?
16:05 paultjuh I have an sls with a task which needs to come first (order: 1), then I have a sls which needs to come after that, and then I have a sls which needs to come after that one
16:05 paultjuh yes but that doesn't work
16:05 Deevolution Anyone seen an issue where Salt in a virtualenv throws this error when connecting to the saltmaster: "[ERROR   ] The master failed to decrypt the random minion token"  ?
16:06 paultjuh babilen, the problem is, I have a couple of things to do before anything else, but they have themselfs also requirements
16:06 Deevolution The minion works fine from without a virtualenv.  And this setup used to work without issue.
16:06 babilen paultjuh: Is it drinking Gin & Tonic? (aka: why doesn't it work?)
16:06 paultjuh babilen, require and order conflicts
16:06 saltycharles joined #salt
16:06 babilen paultjuh: Can you express your requirements in a DAG ?
16:07 iggy http://imgur.com/gallery/bwvUkOx
16:07 iggy babilen: ^
16:07 paultjuh babilen, yes, let me try
16:07 babilen iggy: You might want to switch to ublock - did you try it with adblock disabled?
16:08 iggy indeed
16:08 racooper oh, new feature in 2015.5? I like... "Minion did not return. [Not connected]"
16:09 iggy murrdoc: ^ disabling adblock fixed it
16:09 babilen paultjuh: What kind of conflicts do you get? Could you show us on http://refheap.com ?
16:09 drawsmcgraw joined #salt
16:09 murrdoc heh great
16:09 murrdoc also "Separate sls files each with a single purpose increases their reusability. Gluing them together in a top file or as you've done here with state.sls is a great way to go."
16:09 murrdoc the house agrees with me
16:10 drawsmcgraw Nice update on the docs, Saltstack!
16:10 paultjuh babilen, trying to write it down atm
16:10 stephanbuys joined #salt
16:10 paultjuh babilen, the problem is the documentation says: if a state is required by or requies something order will be overwritten and ignored
16:11 perfectsine joined #salt
16:11 babilen iggy: I'd recommend https://chrome.google.com/webstore/detail/ublock-origin/cjpalhdlnbpafiamejdnhcphjbkeiagm
16:11 sl_ joined #salt
16:11 iggy noted?
16:12 babilen paultjuh: Yes, don't use "order" -- You should work with require and require_in
16:12 stoogenmeyer joined #salt
16:13 paultjuh babilen, but how do I say "require in everything else but not those states which I require mysefl"
16:13 ska-fan joined #salt
16:13 iggy just put it first in your top file
16:13 iggy salt operates top-down
16:13 babilen paultjuh: SLS files are evaluated top-to-bottom
16:13 ska-fan Executing a recipe on a server that should be very simple (install a virtualenv and 4 packages in it) is taking over 3 minutes - how can I find out why?
16:14 murrdoc switched to ublock
16:14 murrdoc still happens
16:14 paultjuh babilen, so basicly I have to add a comment to keep the order and then hope everyone who makes changes reads that
16:14 babilen murrdoc: Restarted the browser?
16:14 murrdoc cant yet
16:14 murrdoc will try in a bit
16:14 murrdoc still dont make sense that site is hitting ad blocker
16:15 babilen paultjuh: If something requires something else then you should make that explicit with a suitable "require" or "require_in" entry
16:15 paultjuh babilen, https://www.refheap.com/e0c6457d3b0f960387930de4f
16:15 c10b10 joined #salt
16:15 bishy999 Hi, I'm using free IPA to manage my users. After successfully installing this and validating that expected users are present on minion via 'salt-call user.list_groups java'  and on the master 'salt '*' user.info java' during highstate when i get to my last step to change file ownership to the new user it doesn't recognise user. Can anyone point me in the right direction?
16:16 paultjuh babilen, it is not a require to run it, It is just to make sure we get all the logs send off as soon as possible, even during deploys
16:16 paultjuh and there is not a require_in: * - the ones in my own require
16:16 jlhammons Iggy: re no functions in doc: will you try clearing your cache, I pushed the code over the weekend
16:16 babilen paultjuh: you might be looking for http://docs.saltstack.com/en/latest/topics/tutorials/states_pt5.html#orchestrate-runner, but then you can also use "require: - sls: foo"
16:17 jlhammons I moved it from the other doc theme so it should be the same
16:17 babilen paultjuh: And which state will actually fail if you don't run curl --fail https://urltoapp.herokuapp.com/mute_all/{{ secret_key }}/{{ grains['id'] }}/ first?
16:18 jlhammons also if you are seeing issues with an ad blocker enabled please add an issue
16:18 paultjuh babilen, We get spammed by alerts because unexpected restarts for services etc
16:19 babilen So it is not an actual dependency of anything, but simply something *you* like to do first?
16:19 paultjuh babilen, no deployment steps will fail if I don't do the mute+hostname+logging first
16:20 murrdoc jlhammons:  i am seeing issues with adblocker on firefox and chrome
16:20 babilen paultjuh: I'd manage that with orchestrate ...
16:20 murrdoc i have switched crhome to ublock, supposedly that will help
16:21 paultjuh babilen, but that isn't the biggest problem, it is also deploying the logforwarder, so things are send of as soon as possible
16:21 bhosmer joined #salt
16:21 jlhammons murrdoc: can you give me the name of the plugin that doesn’t work - is it adblock plus?
16:22 AxelFool_ joined #salt
16:24 Rockj\ joined #salt
16:26 rsimpkins Is there a way to turn off the 'summary' report when running a highstate?
16:30 cromark joined #salt
16:31 jamin joined #salt
16:32 troyready joined #salt
16:32 murrdoc jlhammons:  adblock plus 2.6.9 on firefo
16:32 murrdoc firefox*
16:32 jamin hello. I'm hoping someone here can provide a littled guidance for me.  I'm setting up a new Salt configuration, have 1 master and 2 minions.  1 minion is running Ubuntu, the other Arch.  The Ubuntu minion appears to be operating fine, but the Arch minion is not syncing the state files into its cache
16:32 kartik_subbarao joined #salt
16:32 julez joined #salt
16:34 kartik_subbarao Can the onchanges requisite be used to trigger a restart of a daemon in a single line? For example: onchanges:
16:34 kartik_subbarao module.run: [ name: service.restart, m_name: ssh ]
16:35 aparsons joined #salt
16:35 jlhammons rsimpkins: async might do what you want: http://docs.saltstack.com/en/latest/ref/cli/salt.html#cmdoption-salt--async
16:35 jlhammons murrdoc: thanks, I’ll try to reproduce
16:36 jamin nothing stands out in the logs, I've tried running the arch minion in debug and issusing a saltutil.sync_states, the output from the minion seems to indicate it sync'd, but the cache directory is empty
16:36 rsimpkins jlhammons: Not really, but I'll use it as a work-around for now.
16:37 paultjuh anyone knows what is causing "RSAError: oaep decoding error"
16:38 ageorgop joined #salt
16:39 jamin paultjuh: from what I've found the keypair on the minion differs from what the master thinks it should be (it doesn't validate)
16:39 paultjuh jamin, I understood something like that
16:39 paultjuh but it is just a server booting, installing salt, salt-master auto accepting the key
16:39 paultjuh or not,
16:40 paultjuh it should accept the key at least
16:40 paultjuh and then the server dies
16:40 jamin do the public and private key portions on the minion validate with each other?
16:41 julez joined #salt
16:41 paultjuh not sure how to check that by head
16:41 jamin something like: sudo openssl rsa -pubin -in ${pub_key} -noout -modulus | openssl md5
16:42 paultjuh found somethime like that indeed
16:42 jamin sudo openssl rsa -in ${priv_key} -noout -modulus | openssl md5
16:42 jamin they should be the same
16:43 andrew_v joined #salt
16:43 paultjuh read DSA key
16:43 paultjuh unable to load Key
16:43 paultjuh 139745103021728:error:06078081:digital envelope routines:EVP_PKEY_get1_DSA:expecting a dsa key:p_lib.c:308:
16:44 paultjuh is probably not what it is expected to do
16:44 paultjuh but it probably should be rsa
16:44 jamin definitely no
16:44 paultjuh yes it is rsa
16:44 jeffspeff joined #salt
16:44 paultjuh -----BEGIN RSA PRIVATE KEY-----
16:45 XenophonF can i call salt modules from within jinja?
16:45 paultjuh jamin, so the modulus is the same
16:45 XenophonF like for example, i want to call boto_rds.get_endpoint and put the result into a pillar key
16:45 I3olle joined #salt
16:46 jamin hmmm
16:46 paultjuh XenophonF, in jinja you can call a salt execution module just do {{ salt['boto_rds.get_endpoint']() }}
16:46 desposo joined #salt
16:46 jamin paultjuh: don't know then
16:46 XenophonF that's what i thought, but i got a render error - let me post it
16:47 desposo1 joined #salt
16:47 XenophonF Jinja variable 'dict object' has no attribute 'boto_rds.get_endpoint'
16:47 XenophonF and here's the template : {{ salt['boto_rds.get_endpoint']('dub-ib-mysql-d01') }}
16:48 paultjuh jamin, but the accepted key on the salt master is differnt!
16:48 babilen XenophonF: Did you install the requirements (in particular boto's python module) ?
16:48 jamin well, there you go
16:49 paultjuh jamin, the weird thing is: how did it get differnt?
16:49 linjan joined #salt
16:49 XenophonF hm, maybe not, now that i think about it
16:49 XenophonF boto might not be installed yet, because this is a new EC2 instance
16:49 paultjuh XenophonF, yes that is a bit probablematic with execution modules
16:49 babilen XenophonF: Well, that module requires it.
16:50 jamin paultjuh: possibly generating new keys on each reboot?
16:50 XenophonF yeah, makes sense - i'm in a chicken-or-the-egg kind of situation here
16:50 paultjuh jamin: it is a fresh machine
16:50 babilen XenophonF: And you are on 2015.5, aren't you?
16:51 XenophonF no, 2014.7.5
16:51 paultjuh jamin: basicly what happens: auto scaling group in aws boots up the machine, then does a salt boot strap: then crash, but that happens every x times
16:51 babilen XenophonF: Is that module available on 2015.7.5 ?
16:51 XenophonF maybe?
16:51 XenophonF i'm not 100% certain
16:51 jamin sorry, still pretty new to salt myself so I don't know
16:52 bhosmer_ joined #salt
16:52 XenophonF i'd better re-think this
16:52 paultjuh jamin, thanks, I will just open a bug
16:52 KyleG joined #salt
16:52 KyleG joined #salt
16:52 XenophonF i have a boto_rds state that's supposed to create a mysql server instance
16:53 XenophonF and then i want to pass the instance name to mysql-formula
16:53 iggy jlhammons: it was adblock
16:53 babilen "New in version Beryllium"
16:53 babilen XenophonF: ^^^
16:53 babilen bbl
16:53 XenophonF ah
16:53 rsimpkins left #salt
16:53 XenophonF well in any case, the pillar renders before the database instance would be created, so i'm still stuck
16:54 dopesong joined #salt
17:00 paultjuh oh great, this keeps happening :s
17:02 sdm24 joined #salt
17:02 KyleG joined #salt
17:02 KyleG joined #salt
17:02 jamin paultjuh: you're using an autoscaling group, right?  Are you certain the image your scaling doesn't have a key baked into it?
17:03 paultjuh yes
17:03 paultjuh jamin, otherwise we never would have had servers
17:03 paultjuh we use this method since januari, but lately it happens more often
17:03 jamin sounds like a race condition, which is why I'm wondering about an existing key
17:04 jamin if it registers with an old key, then generates a new one
17:04 paultjuh the minion seems to do that itself
17:04 Rockj\ joined #salt
17:05 jamin right, and the minion(s) are what you're scaling up and down, right?
17:05 aron_kexp joined #salt
17:06 paultjuh jamin: the autoscaling groups spins up new api servers when we need it, which is then salt bootstrapped + salt highstated
17:07 FRANK_I joined #salt
17:08 JDiPierro joined #salt
17:11 drawsmcgraw joined #salt
17:11 AxelFooley joined #salt
17:11 Gareth morning morning.
17:11 cheus joined #salt
17:13 ska-fan Executing a recipe on a server that should be very simple (install a virtualenv and 4 packages in it) is taking over 3 minutes - how can I find out why?
17:14 iggy salt-call -l debug state.sls foo
17:15 Rockj\ joined #salt
17:15 drawsmcgraw joined #salt
17:19 jamin anyone have pointers on how to determine why a minion is not syncing state files?  I can query its grain information just fine
17:21 evilrob joined #salt
17:22 otter768 joined #salt
17:22 jamin shouldn't saltutil.sync_all push all state files to a minion?
17:22 Grokzen joined #salt
17:22 forrest joined #salt
17:23 Fiber^ joined #salt
17:24 murrdoc all files in _ directories
17:24 cromark_ joined #salt
17:24 ajw0100 joined #salt
17:28 theologian joined #salt
17:30 drawsmcgraw joined #salt
17:30 sdm24 joined #salt
17:31 TheRealBill joined #salt
17:32 XenophonF somehow ddg wants to show me "saltstack gifts for mom" when i want to search for "saltstack gitfs"
17:33 ProT-0-TypE joined #salt
17:34 iggy saltnado chews through cpu
17:34 iggy XenophonF: that's because it's a pos
17:41 tomh- joined #salt
17:41 ajw0100 joined #salt
17:41 Cidan joined #salt
17:45 druonysus joined #salt
17:45 Ahlee lovely. I see https://github.com/saltstack/salt/issues/7433 is still open
17:46 forrest Ahlee: jhenry82 does have a work around, not pretty but butter than nothing
17:46 forrest butter? Jeez
17:46 forrest better
17:47 Ahlee forrest: this breaks pkg.installed
17:47 Gareth forrest has butter on the brain.
17:47 Ahlee sorry, I fired up a virtualenv for testing, and can't run pkg.installed as i have to specify providers: pkg: yumpkg5
17:49 Ahlee guess the underlying bug i'm hitting is https://github.com/saltstack/salt/issues/8648
17:50 Ahlee nevermind, anyway
17:51 forrest Ahlee: Ahh
17:51 forrest Gareth: some toast and butter would be good
17:51 forrest iggy: Do you know who this person is? https://github.com/puneetk They were added to the saltstack org and have been merging in PRs but don't have a lot of history in the formulas repo.
17:51 Gareth forrest: That's one of those things that's good, any time of day.
17:52 forrest iggy: the saltstack formulas org that is
17:52 iggy forrest: murrdoc
17:52 forrest iggy: oh okay that makes sense then
17:52 forrest murrdoc: Your username is worse than mine, at least my first name is on my github page ;)
17:52 forrest your irc username that is
17:52 forrest murrdoc: ^
17:53 murrdoc forrest:  :)
17:53 murrdoc i still like u the same
17:53 forrest lol
17:53 murrdoc as for nginx dependancy if its not needed, then lets have the dude remove and doc it
17:55 forrest murrdoc: Yeah I'm torn because other formulas would just install nginx, so you'd have jenkins.install_nginx which includes jenkins.nginx.
17:55 Chadk joined #salt
17:55 Chadk joined #salt
17:55 zer0def joined #salt
17:56 murrdoc it would be fine as a doc
17:56 FRANK_I joined #salt
17:56 iggy making the nginx formula a dependency for things is a terrible idea (imo)
17:56 forrest iggy: I agree, and that's what I noted in my comment
17:56 forrest Should just have the jenkins formula install nginx
17:56 forrest done and done
17:57 iggy (note: have no earthly idea what you guys are talking about, I just dislike that formula... both versions)
17:57 forrest a PR for jenkins-formula where jenkins.nginx includes the nginx state, which is lame
17:57 forrest I like that we support nginx out of the box so you can configure it
17:57 forrest but having to download two formulas to make one works = dependency hell
17:57 murrdoc k lets remove it
17:58 evilrob joined #salt
17:58 murrdoc reply to the thread have james submit a pull request with the dependancy removed and doc'ed
17:58 iggy isn't there a module.function that will wait for a minion?
17:58 forrest murrdoc: Do we know if he's going to create an install_nginx state?
17:58 forrest he didn't explicitly state it
17:59 forrest iggy: Hmm, I don't remember.
18:00 iggy everything .wait() is for watch/watch_in
18:00 murrdoc up to u
18:00 iggy but I could swear there's something that just waits for the minion
18:01 cromark joined #salt
18:02 smcquay joined #salt
18:02 forrest murrdoc: He's got the PR open already and it just removes the nginx content, I'm going to bring it up over in #2
18:03 c10b10 is there any cmd execution time limit set anywhere?
18:04 c10b10 i'm getting a retcode of -9, and i suspect salt killed the process because it took too long
18:04 evilrob joined #salt
18:04 murrdoc dew it
18:04 aparsons joined #salt
18:05 Ahlee c10b10: no. The cli will return back without a value if it takes > 5 seconds (by default), but processes will run forever
18:05 Ahlee c10b10: is this in a _module or _state?
18:06 flebel joined #salt
18:06 c10b10 Ahlee: can I set that 5 sec value anywhere?
18:06 c10b10 it's a state
18:07 c10b10 to be specific, it's a state that runs "npm install"
18:07 Ahlee you can specify -t for longer timeout
18:07 Ahlee salt -t 60 state.sls my_long_npm_install
18:08 Ahlee or salt -v state.sls my_long_npm_install; and if it takes longer on hte mater salt-run jobs.lookup_jid <job id returned from -v>
18:09 JayFK joined #salt
18:13 denys joined #salt
18:15 Rockj\ joined #salt
18:15 bhosmer joined #salt
18:17 baweaver joined #salt
18:17 giantlock joined #salt
18:22 ProT-0-TypE joined #salt
18:23 Ahlee apparently __salt__ isn't available to custom _grains? foo = __salt__['cmd.run']("awk '{/test/; print $2}' /tmp/foo.txt") is throwing a key error on cmd.run
18:23 suppandi joined #salt
18:25 iggy Ahlee: depends
18:25 suppandi hi, can anybody tell me how i can debug a jinja template with file.managed. I am getting empty text in my file  and dont know how to debug what is wrong
18:25 Ahlee yeah, setting my_grain['foo'] = __salt__ comes back as an empty dict
18:25 iggy it should be available _after_ __virtual__()
18:27 murrdoc my_grains
18:27 _JZ_ joined #salt
18:27 stbenjam Has anyone seen this on 2015.5.0? "unpack(b) received extra data."  https://gist.github.com/stbenjam/090b7c359345c525fb2e Ever since I upgraded, something happens to the job cache after ~24 hours or so where 'salt-run jobs.list_jobs' no longer works.  I can't figure out what's causing it.  Clearing /var/cache/salt/master/jobs fixes it temporarily
18:28 iggy stbenjam: did you look through the issue tracker?
18:28 iggy stbenjam: that sounds familiar
18:28 stbenjam iggy: i did, there's an issue that's similar, but only came up during testing, let me find it
18:28 todd53 joined #salt
18:28 stbenjam https://github.com/saltstack/salt/issues/22734
18:29 Ahlee iggy: no joy. https://gist.github.com/jalons/c610b81f06ef7673004f
18:29 Ahlee oh well
18:29 Ahlee i'll shell out directly with subprocess
18:30 pravka joined #salt
18:30 todd53 Let's say I would like to use cron running on a minion, to schedule a task, and I would like that task to be "apply this formula with state.sls". Can I trigger the "apply the formula with state.sls" from the minion?
18:30 Ahlee todd53: salt-call state.sls my_state on the minion
18:30 keimlink joined #salt
18:30 XenophonF how would someone go about troubleshooting gitfs and pygit2?
18:30 julez joined #salt
18:30 brianw_ joined #salt
18:30 todd53 Ahlee: ah, yes.. without --local. perfect. thanks!
18:30 Ahlee i would recommend setting a --return as well todd53, since you won't have a record of the JID on the master
18:31 XenophonF i should be able to run cp.list_master and see the top.sls file from https://github.com/irtnog/salt-states/tree/master/top.sls
18:31 catpiggest joined #salt
18:31 XenophonF but it's coming up empty
18:31 iggy XenophonF: salt-run -l debug fileserver.update
18:31 XenophonF thanks
18:32 XenophonF "Git fileserver backend is enabled in master config file, but could not be loaded, are pygit2 and libgit2 installed?"
18:32 todd53 Ahlee: If using defaults, no postgres/redis/etc.. is there a "return to the master as you normally would"?
18:32 Ahlee todd53: for targeted jobs originating from the master, yes - by specifying a job_cache in the master
18:33 Ahlee salt-call does not honor returners in minion config, and must be appended on the CLI
18:33 todd53 Ahlee: for jobs originating on the minion, what would I tell salt-call to "do the normal thing" ?
18:33 XenophonF oh so salt-formula doesn't install pygit2 using the usual package management tools
18:33 XenophonF that's... surprising
18:33 Ahlee todd53: salt-call state.sls my_state --return <my return>
18:34 Ahlee i.e., salt-call state.sls my_state --return redis
18:34 Ahlee will dump it into redis
18:34 todd53 nevermind, I am not being clear
18:34 todd53 thanks for pointing out --return with salt-call :)
18:34 iggy XenophonF: because no distros had packages for it when I added that support
18:34 XenophonF oh i see
18:35 iggy XenophonF: if they do now, feel free to add support or let me know what does
18:35 XenophonF hang on, i'll send you a pull request in a little bit
18:35 XenophonF FreeBSD has packages for it, at least.
18:35 iggy freebsd master? How's that working out?
18:35 XenophonF great!
18:36 spookah joined #salt
18:37 Tyrm joined #salt
18:37 XenophonF hm, looks like salt-formula doesn't support the gitfs_pubkey/gitfs_privkey knobs yet
18:38 XenophonF i'll see if i can update that stuff
18:38 XenophonF although maybe those settings are only in 2015.5
18:38 Tyrm joined #salt
18:42 iggy they are specific to pygit2
18:43 iggy so not necessarily 2015.5+ but not widely used yet either
18:43 XenophonF i think i'm going to switch to GitPython
18:44 XenophonF now i'm getting "GitError: Failed to authenticate SSH session: Callback returned error"
18:44 conan_the_destro joined #salt
18:44 adelcast left #salt
18:44 XenophonF i have GitPython working on a linux salt master
18:44 XenophonF we'll see if i can get this working under freebsd
18:44 iggy check your config options... they are different between pygit and gitpython
18:47 XenophonF yeah, that's where I think I'm getting hung up
18:49 XenophonF i'm pretty sure i'm doing it wrong
18:49 pravka joined #salt
18:50 iggy hint: gitpython has basically no options
18:51 iggy so if you are specifying options for it, that's probably your problem
18:51 lictor36 joined #salt
18:52 sdm24 joined #salt
18:53 bhosmer__ joined #salt
18:53 monkey66 left #salt
18:54 adelcast joined #salt
18:55 edrocks joined #salt
18:58 hybridpollo joined #salt
18:58 ageorgop joined #salt
19:02 sieve joined #salt
19:06 andrew_v joined #salt
19:07 baweaver joined #salt
19:09 jlhammons joined #salt
19:09 ajw0100 joined #salt
19:14 bhosmer joined #salt
19:17 sieve joined #salt
19:20 Trades joined #salt
19:21 FRANK_I joined #salt
19:23 otter768 joined #salt
19:23 baweaver joined #salt
19:24 Rockj\ joined #salt
19:25 alektant joined #salt
19:28 neogenix joined #salt
19:33 cberndt joined #salt
19:37 mapu joined #salt
19:39 sdm24 joined #salt
19:40 druonysus joined #salt
19:43 voxxit joined #salt
19:44 ckao joined #salt
19:46 FRANK_I join #gitlab
19:47 dopesong_ joined #salt
20:00 JayFK joined #salt
20:09 bhosmer__ joined #salt
20:15 cromark joined #salt
20:20 _JZ_ joined #salt
20:20 baweaver joined #salt
20:23 devramen joined #salt
20:25 pravka joined #salt
20:29 jalbretsen joined #salt
20:33 aparsons joined #salt
20:34 linjan joined #salt
20:40 ajw0100 joined #salt
20:44 clintberry joined #salt
20:46 rsimpkins joined #salt
20:48 rsimpkins Are the outputters listed here still accurate? http://docs.saltstack.com/en/latest/ref/output/all/index.html It does not appear 'progress' works, and I get tracebacks when doing some things like this: salt-run --out=highstate jobs.print_job <some highstate jid>
20:49 iggy that's for 2015.5, so if you are running a different version, you should check the docs for that version
20:49 rsimpkins iggy: I am running 2015.5.
20:50 rsimpkins iggy: That's why I'm wondering if maybe the docs didn't get updated?
20:50 iggy that said, we never like to see tracebacks, so if you see them, check for/file bugs
20:50 iggy _all_ of the docs got updated
20:52 rsimpkins iggy: If that is the case, then --out support in salt-run probably isn't supported. It does kinda work, though.
20:52 rsimpkins The 2015.5 docs for salt-run do not mention --out, but some outputter options do work with it.
20:53 mrbigglesworth joined #salt
20:54 bhosmer joined #salt
20:55 Eugene joined #salt
20:55 c10b10 joined #salt
20:55 rsimpkins iggy: Also, the docs are wrong on this page - http://docs.saltstack.com/en/latest/ref/runners/all/salt.runners.jobs.html Specifically says to use 'outputter=highstate'. I get this 'DeprecationWarning: The 'outputter' argument to the jobs.lookup_jid runner has been deprecated.  Please specify an outputter using --out." So, to find that salt-run has no mention of --out in its docs???
20:55 cromark joined #salt
20:59 aparsons joined #salt
21:01 sdm24 Is the "Denied keys" new in 2015.5.0? I just noticed it when I did salt-key -L and I can't find anything about it in the docs
21:02 iggy rsimpkins: open an issue about it, that's the best way to get it fixed
21:02 rsimpkins sdm24: I ran in to similar challenges with the docs and denied keys. Guessing worked for me. ;-) -l denied worked, I think.
21:04 sdm24 but what are the "denied" keys? I noticed that when I just upgraded to 2015.5.0, some of the keys were in both accepted (and had been for a while), and now denied. I deleted them from the denied folder manually with rm
21:04 sdm24 is denied just the opposite of pre-accepting keys?
21:05 rsimpkins sdm24: I was getting them when there were conflicting keys.
21:05 rsimpkins sdm24: So if you had two host named 'bob' - then the one that came in later gets denied.
21:05 sdm24 ah ok thank you
21:13 keimlink joined #salt
21:14 dendazen joined #salt
21:24 otter768 joined #salt
21:27 baweaver joined #salt
21:28 cberndt joined #salt
21:29 rsimpkins left #salt
21:30 murrdoc anyone setup salt just for remote execution
21:30 murrdoc no config management
21:34 giantlock joined #salt
21:37 iggy linkedin?
21:40 JayFK joined #salt
21:40 Tyrm_ joined #salt
21:41 TaiSHi joined #salt
21:43 TaiSHi joined #salt
21:55 hal58th joined #salt
21:55 hal58th_ joined #salt
21:55 I3olle joined #salt
21:58 moloney joined #salt
21:59 neogenix_ joined #salt
22:02 kitplummer joined #salt
22:08 julez joined #salt
22:09 _JZ_ joined #salt
22:14 baweaver joined #salt
22:18 spookah joined #salt
22:20 zach_ joined #salt
22:22 Gareth murrdoc: RyanLane does that at Lyft, iirc.
22:22 Gareth Ryan_Lane too.
22:22 Gareth but I might have that backwards.
22:23 murrdoc cool, sounds like the master and minion setup would be needed anyways
22:29 bfoxwell joined #salt
22:33 MohShami joined #salt
22:33 MohShami hi guys, is it possible to include a pillar file on the fly? by adding some parameter to salt-call?
22:41 sdm24 MohShami: http://docs.saltstack.com/en/latest/topics/tutorials/pillar.html#setting-pillar-data-on-the-command-line is that what you want?
22:41 cromark joined #salt
22:42 baweaver joined #salt
22:42 MohShami thanks sdm24, I already know about this, but what I would want to do is have the same pillar in multiple files, and depending on a variable, have a single file be imported, so that pillar would match that file
22:44 sdm24 Do you mean like like setting a "role: dev" pillar for minion1 and "role:prod" for minion 2, so that in a different pillar you have "if role:dev, file = foo", "if role:prod, file= bar"?
22:48 MohShami I need to have different values for a state depending on some condition, but the pillar itself is big, so using pillar= is not feasible, I'm looking for something that allows me to do that but with a file
22:48 MohShami sdm24: sorry if I wasn't clear
22:49 sdm24 like a test? if /path/to/file exists on the minion, do this?
22:51 forrest MohShami: I am confused by what you mean, what sort of condition? Why couldn't you set that 'condition' as a custom grain on the minion when it provisions?
22:52 sdm24 I have https://gist.github.com/sdm24/3209fa6ab268bbdef41f that at the end of one of my states. If the /opt/splunk directory does not exist, then the state will include the splunk/server installer. If the directory exists, then the include is ignored
22:52 MohShami sdm24: nope, not a test
22:53 MohShami thanks forrest
22:53 sdm24 And sorry, but I have to go now. Good luck
22:53 MohShami sdm24: thanks mate :)
22:54 MohShami forrest: I'm running all the state.sls commands on a single host, I just want different pillar values for different runs
22:54 MohShami kind of like adding pillar= at the end of salt-call, but I want to store them in a file instead of the command line
22:54 bhosmer_ joined #salt
22:56 forrest MohShami: As far as I know then the only way to do that is A. Set a grain/value which is checked for existence so that multiple runs are treated differently. B. Use the stuff provided by sdm24. Honestly though that is a terrible plan to have different values for different runs (I assume for multiple apps running on the same instance?)
22:56 forrest I can't think of any reason I would ever justify doing that, it leads to complication and difficulty down the line.
22:57 MohShami I use it to deploy websites via git
22:57 forrest If you want to explain what you're trying to do I've got about 15 minutes before I need to head out and we could try to come to a solution.
22:57 forrest MohShami: Okay, so multiple websites, single host?
22:57 MohShami and I have about 10 websites, so if I want to deploy website A, I call state.sls websites._siteA_
22:57 MohShami yep, single host
22:57 forrest so why does each site not have it's own pillar file?
22:58 MohShami I want to minimize the amount of typing
22:58 MohShami how am I to do so?
22:58 MohShami oh wait, I think I have an idea
22:58 forrest MohShami: Can you give me an example file (nginx config or whatever) where you use pillar values?
22:58 forrest You can have multiple pillar files for each site which are included in your top.sls
22:58 forrest Then you reference.
22:58 MohShami I don't use pillars now, I'm trying to figure out how to do it
22:58 MohShami yep, just popped into my head
22:58 MohShami thanks mate :)
22:58 forrest MohShami: oh man, prepare to have your mind blown, give me 5 minutes to gist something.
22:59 MohShami thanks my friend :)
23:00 forrest MohShami: You manage the server with all these sites correct?
23:00 MohShami yep
23:05 toddnni_ joined #salt
23:05 murrdoc still waiting on the mind blowage
23:05 bhosmer_ joined #salt
23:07 forrest MohShami: https://gist.github.com/gravyboat/43f9948d9f89b3e21aee
23:07 forrest so do that
23:07 forrest little more work up front to split your apps out between directories/states
23:08 forrest murrdoc: Your mind won't be blown, you use pillar regularly
23:08 forrest MohShami: but once you have it set up, it's going to make management a breeze, and a lot of it will be copy pasting.
23:08 murrdoc tis blown
23:08 forrest MohShami: There are other ways you could get REALLY fancy as well, with loops where you loop over all the sites, and create the nginx files and configs that way.
23:09 MohShami thanks mate
23:09 forrest MohShami: So your pillar is a list/dictionary which you then query against.
23:09 forrest yeah np
23:09 forrest murrdoc: boooom
23:09 MohShami I'm working on something now, will share once I'm done :)
23:09 forrest Cool, best of lukc
23:09 forrest *luck
23:09 AbyssOne joined #salt
23:09 murrdoc it is blown
23:09 forrest lol
23:10 forrest Hey bretep, haven't seen you around in a while
23:10 forrest or I miss your joins
23:17 murrdoc cold
23:19 MohShami forrest: so I figured it out, what I did was, I created the following pillars: default_website, websiteA, websiteB
23:20 MohShami then in my state, I {% set website_info = pillar.get('default_website') %}
23:20 MohShami then {% do website_info.update(pillar.get(pillar.get('website'))) %}
23:20 Jimlad_ joined #salt
23:20 forrest ahh yeah that's an option as well.
23:20 MohShami so now what I do is, state.sls website.deploy pillar='{"website": "siteA"}'
23:21 forrest Yeah, totally viable
23:21 MohShami thanks mate
23:21 forrest yeah np
23:21 MohShami been trying to figure out a clean way to do this for days
23:21 MohShami and now it's 2:20AM where I live
23:21 MohShami so I have to bid you farewell
23:21 MohShami forrest: thanks again mate :)
23:22 forrest You're welcome, have a good one
23:22 MohShami same to you
23:22 dimeshake nothin' like a late night breakthrough!
23:22 MohShami dimeshake: tell me about it :D
23:22 MohShami have a good one as well :)
23:23 dimeshake you as well
23:23 MohShami thanks
23:24 keimlink_ joined #salt
23:24 VSpike joined #salt
23:24 otter768 joined #salt
23:24 kunersdorf joined #salt
23:26 baweaver joined #salt
23:26 Pixionus joined #salt
23:36 testdsd joined #salt
23:38 dendazen joined #salt
23:39 Gareth Anyone using 2015.5 with multi-masters?
23:39 murrdoc not using multi masters at all
23:39 psojourner_ joined #salt
23:39 murrdoc u got that working ?
23:41 Gareth Nah. I'm looking at a reported bug with 2015.5, multi masters and events.  Wanted someone with actually more than one to test it :)
23:42 murrdoc whats the bug ?
23:42 murrdoc let me see if i can test it for u
23:42 Gareth basically events don't get handled at all if you have multiple masters configured.
23:43 cromark joined #salt
23:43 murrdoc yeah i seen that
23:43 murrdoc asked rallytime about it too
23:43 murrdoc havent heard back
23:43 salty_to_the_cor joined #salt
23:44 kusams joined #salt
23:44 Gareth I've found the cause and *think* I have a fix.
23:46 Gareth Fixed in develop since it was that code was rewritten, but broken in 2015.5.
23:46 ek6 gareth: whats your fix dare i ask
23:47 Gareth one line in minion.py: change "minions[master].handle_event(package)" to "minions[master]['minion'].handle_event(package)"
23:47 Gareth line 634.
23:49 alexanderilyin joined #salt
23:49 ek6 what does that change accomplish..as in what related to events and multimaster being a total loose cannon...  that doesnt look like #19963 but then again thats not fixed in develop or anywhere else
23:50 brianfeister joined #salt
23:50 ek6 shouldnt be hard at all for me to test..just need to know specifically which part your after fixing here
23:51 mapu joined #salt
23:51 Gareth If you look at the surrounding code, it's wrapped in a try except...with a pass for the exception.  When you have a list of masters in the minion config, minions[master] is a dict, so it doesn't have handle_event, so it silently fails because of the pass.  The fix points to the actually minion object in the dictionary, which does have handle_event.
23:53 _JZ_ joined #salt
23:54 pixionus joined #salt
23:55 kitplummer joined #salt
23:55 cztanu joined #salt
23:56 kitplummer joined #salt
23:56 smcquay joined #salt
23:57 loz-- joined #salt
23:59 kunersdorf given: gitfs_remotes: - https://foo.com/foo.git - https://foo.com/bar.git
23:59 kunersdorf are the entries in source foo/some_file and bar/another file?

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