Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-06-11

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

All times shown according to UTC.

Time Nick Message
00:06 p0rkbelly left #salt
00:08 xDamox joined #salt
00:10 moos3 joined #salt
00:11 forrest righdforsa: Yeah you can use the watch_in across sls files as long as you have the state included
00:11 subsignal joined #salt
00:12 righdforsa forrest: isn't there some problem with circular includes that I have to watch out for, though? just thought I've seen something like that
00:12 forrest righdforsa: Totally, circle includes can be a problem. I usually just make it easy by doing something like install.sls -> service.sls -> unique_config.sls -> ....
00:13 forrest then you know 'oh I want a base install and the service? service.sls, oh I want our unique config? unique_config.sls'
00:13 forrest then you create a chain that ensures you get the right thing, while avoiding circular includes.
00:14 righdforsa yeah, I think that makes sense, I may have to do some refactoring to make sure the files are appropriately scoped...something we wanted to do anyway
00:14 forrest yeah makes sense.
00:14 righdforsa thanks!
00:14 forrest np
00:18 jmccree How long does the salt PPA normally take to get new releases?
00:21 msciciel1 joined #salt
00:30 forrest jmccree: A couple days max if the packager is around
00:34 jmccree hoping https://github.com/saltstack/salt/issues/24101 is fixed in latest
00:39 spookah joined #salt
00:42 baweaver joined #salt
00:48 hasues joined #salt
00:49 mapu joined #salt
00:50 hasues left #salt
00:52 jonlangemak joined #salt
00:52 hasues joined #salt
00:52 hasues left #salt
00:52 tmclaugh[work] joined #salt
00:54 subsignal joined #salt
01:07 Nazca__ joined #salt
01:10 julez joined #salt
01:14 andrew_v joined #salt
01:18 __number5__ basepi: just fyi, there are lots of links on http://saltstack.com/enterprise/ shows 404, e.g. those under `SaltStack on Twitter`
01:20 otter768 joined #salt
01:20 david_an111 joined #salt
01:23 hasues joined #salt
01:32 beauby joined #salt
01:42 ronrib joined #salt
01:45 baweaver joined #salt
01:47 hasues left #salt
01:50 NV joined #salt
01:55 stooj joined #salt
01:59 zz_ashmckenzie joined #salt
02:00 ronrib joined #salt
02:06 drawsmcgraw left #salt
02:09 beauby joined #salt
02:13 subsignal joined #salt
02:14 moos3 joined #salt
02:15 desposo joined #salt
02:22 malinoff joined #salt
02:24 ronrib joined #salt
02:26 andrej joined #salt
02:28 beauby joined #salt
02:33 Dev0n joined #salt
02:33 timoguin joined #salt
02:36 smcquay joined #salt
02:37 smcquay joined #salt
02:39 timoguin joined #salt
02:44 Furao joined #salt
02:45 donmichelangelo joined #salt
02:56 amcorreia joined #salt
03:05 baweaver joined #salt
03:07 ITChap joined #salt
03:09 beauby joined #salt
03:10 writtenoff joined #salt
03:12 clintberry1 joined #salt
03:14 ronrib joined #salt
03:19 smcquay joined #salt
03:20 ronrib joined #salt
03:21 otter768 joined #salt
03:37 aqua^c joined #salt
03:56 Not_ joined #salt
04:08 aqua^c joined #salt
04:11 TyrfingMjolnir joined #salt
04:34 evle joined #salt
04:35 stoogenmeyer_ joined #salt
04:37 favadi joined #salt
04:56 cozby_ joined #salt
04:57 cozby_ hi, I'm trying to call highstate from one of the minions and I'm getting the following error: Could not deserialize msgpack message: In an attempt to keep Salt running, returning an empty dict.This often happens when trying to read a file not in binary mode.Please open an issue and include the following error: Unpack failed: error = 0
04:57 cozby_ anything I can do about this?
04:57 cozby_ besides what it says of course...
05:15 TyrfingMjolnir joined #salt
05:15 IanV0rn joined #salt
05:20 chiui joined #salt
05:22 otter768 joined #salt
05:23 joeto joined #salt
05:37 rdas joined #salt
05:40 smcquay joined #salt
05:57 stoogenmeyer joined #salt
06:02 AndreasLutro joined #salt
06:02 aberdine joined #salt
06:03 favadi joined #salt
06:04 colttt joined #salt
06:08 stoogenmeyer hey can I call mine.get from inside a custom grain ?
06:10 Furao probably no, AFAIK, grains are loaded before modules
06:14 stoogenmeyer thats a bummer
06:15 dopesong joined #salt
06:17 stoogenmeyer lets say I have 2 machines, one considered main and one secondary. They are both reporting their ips via salt mine. when I set a state on the secondary I want it to use the ip from the main machine and was hoping to get it while rendering the pillar file. problem is mine.get doesn't work inside of pillars
06:17 stoogenmeyer can't use mine.get in a custom grain either
06:18 stoogenmeyer so i don't really see a way to get that data outside of the actual sls part
06:18 Furao I can use mine in my pillars
06:19 Furao https://gist.github.com/bclermont/e71970974b23d6ccfd1a
06:19 Furao this is a slightly changed pillar .sls file
06:19 Furao that is used in prod
06:21 viq joined #salt
06:25 forrest joined #salt
06:28 stoogenmeyer Furao: When I try to use mine.get in a pillar I get an error from salt-master (Jinja variable 'mine' is undefined)
06:29 stoogenmeyer Furao: sorry my mistake, fixed. Now the real problem - mine is defined but returns empty results
06:31 hvn joined #salt
06:31 hvn joined #salt
06:31 flyboy joined #salt
06:31 favadi joined #salt
06:32 impi joined #salt
06:34 stoogenmeyer so if I run `salt '<minion>' mine.get '*' private_ip` -> Works perfectly. Running in pillar salt['mine.get']('*', 'private_ip') -> Return empty data
06:35 Ztyx joined #salt
06:35 julez joined #salt
06:38 smcquay joined #salt
06:38 Furao mine need a function name
06:38 Furao it’s suposed to be modname.funcname
06:39 soren joined #salt
06:42 stoogenmeyer_ joined #salt
06:43 smcquay joined #salt
06:44 stoogenmeyer_ Furao: can you take a look? (https://gist.github.com/rikonor/50f2719e1019d22862dc)
06:44 stoogenmeyer_ How is my call to mine.get wrong there
06:46 ndrei joined #salt
06:51 Auroch joined #salt
06:51 toddnni joined #salt
06:52 stanchan joined #salt
06:56 eliasp joined #salt
07:02 aqua^c joined #salt
07:03 eseyman joined #salt
07:04 KermitTheFragger joined #salt
07:09 al joined #salt
07:17 Romlok joined #salt
07:19 marcinkuzminski joined #salt
07:23 otter768 joined #salt
07:28 davidka joined #salt
07:29 davidka can this test be a pcre test in some way: {% if grains['id'] == 'alt-cex-wxs02-ci' %} (I'm in a pillar file)
07:29 davidka alternatively, can it be an or test {% if grains['id'] == 'alt-cex-wxs02-ci' || someOtherTest %} - what is the syntax or or?
07:31 KermitTheFragger joined #salt
07:43 onorua joined #salt
07:44 stoogenmeyer_ hey my salt-minion has gone rogue and is spawning many salt-call commands that do not return, how can I make it stop
07:44 lb1a joined #salt
07:45 stoogenmeyer_ tried restarting the salt-minion which doesn't help, reboot doesn't help, there's no files in the /var/cache/salt/minion/proc
07:47 ndrei joined #salt
07:52 julez joined #salt
07:59 onorua joined #salt
08:02 fivmo joined #salt
08:03 favadi joined #salt
08:06 Ztyx joined #salt
08:06 impi joined #salt
08:09 mage_ is RAET stable already ?
08:10 s_kunk joined #salt
08:11 s_kunk joined #salt
08:11 kawa2014 joined #salt
08:13 CeBe joined #salt
08:14 favadi joined #salt
08:16 badon_ joined #salt
08:16 hi joined #salt
08:17 hi hello guys
08:19 ndrei joined #salt
08:24 ingslovak joined #salt
08:25 fredvd joined #salt
08:28 fredvd_ joined #salt
08:30 Berty_ joined #salt
08:30 favadi joined #salt
08:36 markm joined #salt
08:36 badon_ joined #salt
08:40 N-Mi joined #salt
08:44 bluenemo joined #salt
08:46 linjan joined #salt
08:50 VSpike Hm, I didn't know this was a thing http://docs.saltstack.com/en/develop/ref/renderers/all/salt.renderers.yamlex.html
08:53 keimlink joined #salt
09:02 che-arne joined #salt
09:02 ndrei joined #salt
09:07 yuhl_work___ joined #salt
09:07 ninkotech joined #salt
09:09 kevc is there a way to create a user with an initially blank password, but then not modify it once the password has been set?
09:10 kevc empty_password will clobber an existing user's password
09:16 lempa joined #salt
09:17 jasonrm joined #salt
09:19 stoogenmeyer_ i have a custom grain that makes a call to mine.get, this causes the minion to repeatedly make the mine.get calls, thereby killing it. does the mine.get trigger a refresh of grains data? https://gist.github.com/rikonor/39d5c7a21be3de93c1b2
09:19 slav0nic joined #salt
09:19 slav0nic joined #salt
09:23 Aidin joined #salt
09:23 otter768 joined #salt
09:29 stoogenmeyer_ joined #salt
09:32 chiui joined #salt
09:33 babilen Why are you keeping that in grains?
09:33 babilen ah, roles in grains
09:33 babilen nvm
09:34 babilen And you don't have to use subprocess to call functions in __salt__ in a custom grain
09:37 Grokzen joined #salt
09:39 aparsons joined #salt
09:45 aparsons joined #salt
09:50 ndrei joined #salt
09:54 sirtaj joined #salt
10:04 mage_ hello, I see that I can use an - id: in an exclude:, is it possible to do the opposite, like include: - sls: foo - id: blah ?
10:05 denys joined #salt
10:09 evle1 joined #salt
10:11 sgargan joined #salt
10:11 babilen mage_: No, you can only include the entire SLS, there is no "from foo import bar" support
10:12 mage_ too bad
10:12 babilen I'd argue that you want to split your SLS into semantically more appropriate units if you require that functionality
10:12 mage_ is it planned ?
10:13 babilen Don't think so, but haven't checked
10:13 mage_ babilen: yeah, atm I have dozen of foo_xxx.sls files with something like {% set ... %} {% include "some/_template.sls" %}
10:14 saifi joined #salt
10:14 babilen That appears to be an "interesting" approach that doesn't necessarily adhere to best practices
10:15 mage_ do you think it is a wrong approach ?
10:16 babilen I can't answer that question without knowing what you are actually doing and what you are trying to achieve, but it sounds as if you'd rather want to start to use pillars for parameterising your states.
10:22 Norbell joined #salt
10:26 sgargan joined #salt
10:26 jim__ joined #salt
10:27 ndrei joined #salt
10:27 jim__ Hey guys, running salt * test.ping returns duplicate hosts for me, any ideas why or is this a bug
10:27 jim__ ?
10:28 birdzhang maybe you have the same hostname hosts XD
10:28 jim__ aha, alas no
10:28 jim__ the funny thing is, salt-run manage.status has no duplicates
10:30 Norbell joined #salt
10:30 jim__ doing a reboot fixed it
10:31 jim__ sending cmd.run reboot to the hosts caused it
10:33 birdzhang it's funny
10:40 fe92 joined #salt
10:41 mage_ any idea for this https://gist.githubusercontent.com/silenius/5fd8d7457e18819f9474/raw/938400de17fd05113db39a6ceee0a68e374c49d8/gistfile1.txt ?
10:41 mage_ if I restart the minion it works ..
10:42 mage_ seems like a bug
10:52 riftman joined #salt
10:56 N-Mi joined #salt
10:58 leev joined #salt
10:59 aqua^c joined #salt
11:00 catpig joined #salt
11:01 sgargan joined #salt
11:02 Nazca joined #salt
11:03 kbyrne joined #salt
11:06 supersheep joined #salt
11:08 sgargan joined #salt
11:10 leev i want to append some text to a file if something else isn't in the file.  the appended text and the matching text isn't the same.  i've looked at file.append and file.replace, but neither seem to be able to do this.
11:10 leev is there another method?
11:10 sgargan joined #salt
11:18 giantlock joined #salt
11:18 ndrei joined #salt
11:24 johnkeates joined #salt
11:24 chiui joined #salt
11:24 otter768 joined #salt
11:34 johnkeates joined #salt
11:35 one joined #salt
11:38 sgargan joined #salt
11:47 eliasp does salt-ssh also support an operation mode where it doesn't build a thin-tarball but instead just transfers pillars/states/localmods etc. and then calls an already existing 'salt-call' on the Minion?
11:48 eliasp so basically a masterless setup, but orchestrated through salt-ssh…
11:49 one it's called standalone minion i guess
11:49 one it uses local file roots and salt-call instead of minion daemon and salt-master
11:49 one left #salt
11:50 eliasp sure, but this means I'd have to manually rsync my states/pillars/etc. before… and that's something salt-ssh could do…
11:51 eliasp as salt-ssh already uses the roster system to have knowledge of the various minions, the connection method etc.
11:51 giantlock joined #salt
11:52 eliasp so this would mean I'd have to manage a ~/.ssh/config with my Minions instead and then run every time: rsync -av salt-data/ $minion:/srv/salt/ && ssh $minion salt-call foo.bar
11:56 AndreasLutro eliasp: it does not, though that could be an interesting feature request
11:57 eliasp well, I might file one then later :)
11:57 codehotter How do I ensure that listens / watches are fired even if there is an error in a statefile?
11:57 ndrei joined #salt
11:58 mephx joined #salt
11:58 codehotter I want to avoid the combination change something -> error -> watch doesn't fire, on next run, it's not changed, so watch doesn't fire.
11:58 codehotter Ansible has "force_handlers" option for this that runs all the notified tasks even if there's an error somewhere
12:00 mpa1 joined #salt
12:00 aqua^c joined #salt
12:00 AndreasLutro codehotter: in what context do you change something, get an error, fix it and then on the next highstate there are no changes? if I change a state so that it errors and then fix it, 100% of the time that will trigger a change
12:01 AndreasLutro though regardless I think the answer to your question is no, watch/listen is only meant to trigger on changes
12:01 AndreasLutro maybe the "onfail" requisite can help you achieve what you need?
12:01 codehotter In state 1, I change something, in state 2, it's an error, and state 3 listens for change in state 1.
12:02 codehotter so the thing is changed, it prepares to fire the listen, but because of the error in state 2, everything stops
12:02 codehotter after I fix state 2, the listen will not fire because state 1 is not a change anymore
12:02 AndreasLutro that seems very odd to me, that an error in state 2 would cause state 3 not to fire...
12:03 KennethWilke joined #salt
12:03 AndreasLutro do you run with hard_crash: True or something?
12:03 AndreasLutro or failhard
12:03 fe92 Hello, I'm coming from ansible and can't find how to call a certain state only in case for example a file is changed similar to how handlers work in ansible
12:03 AndreasLutro fe92: http://docs.saltstack.com/en/latest/ref/states/requisites.html#onchanges
12:04 fe92 thanks
12:04 AndreasLutro looks like it's missing an example
12:04 AndreasLutro jsut look at the examples from the "watch" requisite and change "watch" to "onchanges", you'll get the idea
12:08 stephanbuys joined #salt
12:12 ferbla joined #salt
12:13 fe92 AndreasLutro: I don't understand, could you look at this example? http://pastebin.com/Q1Au9VFb
12:15 AndreasLutro fe92: you don't need onchanges for that - use watch_in: - service: monit
12:15 AndreasLutro that will take care of the reloading/restarting for you
12:15 hvn joined #salt
12:17 Berty__ joined #salt
12:18 stoogenmeyer_ can you have require inside of a reactor formula?
12:18 stoogenmeyer_ I want to sync_all and refresh_pillar when a new minion comes up, but I want the grains to refresh first
12:21 julez joined #salt
12:22 lb1a joined #salt
12:24 cheine joined #salt
12:25 david_an11 joined #salt
12:26 patrek joined #salt
12:36 amcorreia joined #salt
12:38 rdas joined #salt
12:38 premera joined #salt
12:46 renat joined #salt
12:48 renat Hi! Does anybody know why in parted module no yet ext4 support? Need the patch?
12:50 ndrei joined #salt
12:55 supersheep joined #salt
12:56 stoogenmeyer__ joined #salt
13:00 subsignal joined #salt
13:02 elfixit joined #salt
13:04 ndrei joined #salt
13:07 murrdoc joined #salt
13:07 aparsons joined #salt
13:07 murrdoc why doesnt /proc/sys/net/ipv4/vs/expire_quiescent_template work with sysctl.present
13:08 riftman joined #salt
13:10 murrdoc probably lvs is not installed
13:14 ndrei joined #salt
13:14 JDiPierro joined #salt
13:17 ggoZ joined #salt
13:18 murrdoc1 joined #salt
13:19 cheine Hi! Is it possible to have relative includes and go directories upwards? like "include: - ..common.stuff"
13:19 TooLmaN joined #salt
13:24 murrdoc joined #salt
13:25 otter768 joined #salt
13:25 hasues joined #salt
13:26 ajw0100 joined #salt
13:30 emaninpa joined #salt
13:30 racooper joined #salt
13:31 riftman joined #salt
13:34 ndrei joined #salt
13:36 sgargan joined #salt
13:42 ggoZ joined #salt
13:42 andrew_v joined #salt
13:43 ndrei joined #salt
13:46 tmclaugh[work] joined #salt
13:47 jdesilet joined #salt
13:49 aqua^c joined #salt
13:49 mpanetta joined #salt
13:49 mapu joined #salt
13:52 jonlangemak joined #salt
13:52 cpowell joined #salt
13:56 dyasny joined #salt
13:58 mpanetta_ joined #salt
13:58 giantlock joined #salt
14:01 dariusjs joined #salt
14:01 fe92 joined #salt
14:01 fe92 joined #salt
14:01 dopesong_ joined #salt
14:02 hasues joined #salt
14:02 coval3nce joined #salt
14:02 hasues left #salt
14:03 zerthimon joined #salt
14:04 jespada joined #salt
14:05 andrew_v joined #salt
14:05 _JZ_ joined #salt
14:06 pimpind joined #salt
14:06 pimpind hello
14:07 pimpind so i am not a linux admin or any where near that, but together with out incompetent linux admin ive been placed on a project to research some systems which can be implemented in our environment of basica linux patching of vulnuerabilities
14:07 pimpind ive been pointed towards salt
14:07 pimpind as well as a few others
14:07 sgargan joined #salt
14:08 GrueMaster joined #salt
14:09 dariusjs pimpind: and what do you want to achieve? just patch things ?
14:09 dariusjs saltstack is perfect for that, I would recommend you start small and just  use it for remote execution to manually upgrade only what you want upgraded
14:09 slimmons As someone who is pretty new to the sys admin world, I can tell you that I've tried a lot of tools, and salt has been the most amazing one so far.  It's easy to pick up, and the community is full of really bright people.
14:09 pimpind dariusjs - excuse my lack of proper words or knowledge, but it seems liek we are looking for a way to patch our linux boxes when rhel releases vulnerabilities
14:10 pimpind ive POC's lumensions plugin for SCCM
14:10 pimpind which worked great
14:10 dariusjs i dont know how far youve come, but in some cases I would recommend doing  salt "host" cmd.run 'yum -y update httpd"    if its rhel
14:10 pimpind (but im the sccm guy)
14:10 dariusjs ok, and you need to update linux boxes as part of your job?
14:10 pimpind this caused the linux guy to wipe his hands clean and stop trying to set up other solutions
14:11 pimpind dariusjs it would be on the linux admin to update them, but he is the most incompent worker ive met
14:11 pimpind hes been trying to get satelliete to work for 9 months
14:11 dariusjs are you a redhat shop or centos then
14:12 XenophonF joined #salt
14:12 pimpind redhat
14:12 pimpind are you familiar with lumension data center plug in for sccm?
14:12 XenophonF i want to render a configuration file using jinja based on yaml data
14:13 pimpind it basically sets up a webserver you register with rhel, the polls the sevrers via the agent installed, and tells you which patches it needs
14:13 XenophonF i wanted to use isinstance() to render different data types, well, differently
14:13 XenophonF here's the file - https://github.com/irtnog/salt-states/blob/development/kerberos5/files/krb5.conf.jinja
14:14 XenophonF of course, file.managed returns the error "Jinja variable 'isinstance' is undefined"
14:14 dariusjs pimpind: not really, i dont have much to do with SCCM. to me it sounds like you need something basic, no need to go crazy.   Step1.  Install MREPO on a server1 and synchronise all your patches here. Step3 build a saltstack server.  step 4 use saltstack  to  manage yum files on your *nix boxes
14:15 pimpind does salt require you to know which patches you need?
14:15 dariusjs it all of course depends what kind of licensing agreement you have with redhat ... im not the biggest fan of how the redhat patching is done
14:15 XenophonF so is it possible to monkey-patch Salt/Jinja2 to allow calls to isinstance?
14:15 XenophonF or is there a better way to check the types of variables in a Jinja template?
14:15 dariusjs pimpind:  think of saltstack like microsofts  MSMQ + group policy + winrm on steroids
14:16 dariusjs if you ever needed to perform command X on many different machines you would use saltstack
14:17 dariusjs redhat has solutions for managing security patches which you must pay for, maybe your company does that already its difficult to say.  if on the cheap if every redhat system is already subscribed to the redhat network then  yeah you could use saltstack to manually update the packages that you need to
14:18 pimpind dariusjs we pay for some, but we were suppose to poc satellite
14:18 pimpind which is $
14:18 pimpind which isnt a problem
14:18 pimpind director just wants it done
14:18 unused_phd joined #salt
14:18 pimpind salt is assuming the linux guy is proficient in linux and able to setup stuff ya? lol
14:19 dariusjs if $$$ isnt a problem just get a redhat engineer in and just do it as redhat tells you to do it
14:20 dariusjs salt is really great anway and it would be my personal tool of choice for the situation you describe but it needs willingness
14:20 dariusjs im using it to manage a windows environment too for example
14:21 pimpind dariusjs if u could have heard our linux admin and the rhel guy
14:21 pimpind at one point the linux admin screams "the documentation jsut looks like words on the screen"
14:21 pimpind the rhel guy goes... well sir this program requires you to have a bit of understand about rhel
14:23 dariusjs well I cant judge your situation, but redhat offers courses too
14:24 pimpind dariusjs - funny story
14:24 pimpind so he took the course for maybe rhce?
14:24 pimpind and went to take the exam
14:24 pimpind and the 1st question invovled something like removing admin password to start the exam
14:24 pimpind and he couldnt even start the exam!
14:25 dariusjs the course isnt a guarantee of knowledge anyway, no course can replace years of experience, its just a nice basis or start or confirmation of some small subset of knowledge
14:25 aparsons joined #salt
14:25 dariusjs but anyway, ttyl, hope i helped, gotta run
14:26 pimpind thanks
14:26 pimpind later
14:26 ndrei joined #salt
14:37 unused_phd joined #salt
14:38 jdowning joined #salt
14:43 slav0nic joined #salt
14:43 slav0nic joined #salt
14:43 sk_0 i'm preseeding keys on new salt-minions (pub & pem) and putting the pub on the master named the minion id. the key fingerprints match. the minion can reach the master but the master can't reach the minion. i'm stumped
14:44 iggy stoogenmeyer__: reactors (like everything in salt) are processed top-down, just put the grains refresh before it
14:46 stoogenmeyer__ iggy: did just that. thanks
14:47 iggy apparently I was scrolled up a bit ;)
14:49 adelcast joined #salt
14:51 Brew joined #salt
14:51 onorua joined #salt
14:51 dopesong joined #salt
14:52 XenophonF i want to process pillar data in different ways depending on its type, but jinja won't let me use isinstance() in a template (e.g., https://github.com/irtnog/salt-states/blob/development/kerberos5/files/krb5.conf.jinja, which doesn't work). what other options do i have?
14:52 debian112 joined #salt
14:52 iggy jinja has tests for some types
14:52 clintberry1 joined #salt
14:52 tracphil joined #salt
14:53 t0nyhays joined #salt
14:53 sgargan joined #salt
14:53 XenophonF can you give me an example, iggy?
14:53 XenophonF i may have overlooked it in the jinja2 docs
14:53 iggy callable, ierable, mapping, string, etc
14:54 iggy http://jinja.pocoo.org/docs/dev/templates/#list-of-builtin-tests
14:54 tracphil any saltstack sales people in here by chance?
14:54 XenophonF thanks iggy
14:55 XenophonF i'll give it a spin
14:56 bluenemo joined #salt
14:56 bluenemo joined #salt
14:56 t0nyhays salt newbie here..just completed my first install. master and minion1 both on kvm hosts. salt '*' grains.item virtual returns physical for both
14:56 t0nyhays Am i missing something?
14:56 iggy tracphil: doubtful, catch utahdave or basepi if you need something
14:57 tracphil iggy: thanks. wil do.
14:57 slav0nic joined #salt
14:58 slav0nic joined #salt
14:58 unused_phd joined #salt
15:00 sgargan joined #salt
15:01 kickerdog joined #salt
15:01 ferbla_ joined #salt
15:02 peters-tx iggy, Continued from yesterday; what would have saved my system is if I'd used the "timeout=" parameter with the cmd.run module
15:02 peters-tx iggy, That would have performed what I expected
15:03 peters-tx iggy, If it is still running after "timeout=" seconds, it does a TERM and then a KILL
15:03 iggy peters-tx: good to know
15:03 chiui joined #salt
15:04 peters-tx iggy, Yah, I wasn't aware of that functionality..that's a nice safety valve, or at least for those who use cmd.run a lot :(  Probable shouldn't be using it :)
15:04 thedodd joined #salt
15:05 Striki I have a salt master and minions connect to it. Is there a way for me to have the minions automatically execute highstate right after they've successfully connected to the master? I would like to skip issuing the highstate from the master and I don't want to store the whole repository with my salt states on the minion (masterless)
15:05 Striki so essentially what I want is to make the minion execute highstate automatically but using the data from the master
15:06 kickerdog joined #salt
15:10 chiui joined #salt
15:10 ndrei joined #salt
15:10 jalbretsen joined #salt
15:11 iggy Striki: there are a few ways
15:11 iggy reactor, startup_states
15:13 Striki ok thank you, I'll look into those
15:13 joeto joined #salt
15:14 bmac2 joined #salt
15:16 kickerdog joined #salt
15:23 hvn joined #salt
15:23 hvn joined #salt
15:23 XenophonF iggy, mapping as used in my template (https://github.com/irtnog/salt-states/blob/development/kerberos5/files/krb5.conf.jinja#L5) didn't work - got "Jinja variable 'mapping' is undefined"
15:24 _mel_ joined #salt
15:24 XenophonF jinja 2.7.2 is installed on the minion
15:24 ggoZ joined #salt
15:24 XenophonF and 2.7.3 on the master
15:26 otter768 joined #salt
15:27 _twork joined #salt
15:30 pbrooko joined #salt
15:31 julez joined #salt
15:31 AdobeTom joined #salt
15:32 ferbla joined #salt
15:34 tomspur_ joined #salt
15:36 dendazen joined #salt
15:37 iggy if foo is mapping
15:37 aqua^c joined #salt
15:38 t0nyhays i see a similar bug, but it was closed in January as fixed.  I am running 2015.5.1 on both minion and master
15:38 bluenemo joined #salt
15:38 bluenemo joined #salt
15:38 ange since salt-api was merged in the base package is there something much different to apply to set it up and running compared to http://bencane.com/2014/07/17/integrating-saltstack-with-other-services-via-salt-api/ ?
15:39 iggy t0nyhays: the tests for virtual are far from perfect unfortunately... you're better off keying on something else (if for no other reason than people can change a lot of stuff in KVM to make it look non-virtual)
15:40 iggy ange: I followed the saltnado docs and it just worked
15:40 ange iggy: do you have a link ?
15:41 iggy docs.saltstack.com
15:41 giantlock joined #salt
15:41 XenophonF iggy, would "if value is sameas(True) or value is sameas(False)" work like I think?
15:42 ange iggy: ok
15:42 ange thanks
15:42 iggy XenophonF: maybe, you'll have to test it... it actually does a memory address comparison vs a value comparison
15:42 XenophonF OK
15:44 ange there is still a package for salt-api for ubuntu?
15:47 iggy ange: it's probably just the init script, but yeah
15:48 amcorreia joined #salt
15:51 conan_the_destro joined #salt
15:53 DaveQB joined #salt
15:55 xDamox joined #salt
15:55 pizzahead joined #salt
15:55 Kiyundai joined #salt
15:56 vexati0n how can I set new config options in /etc/salt/minion from a state or pillar on the master?
15:59 amcorreia joined #salt
15:59 kzrl joined #salt
16:00 dopesong joined #salt
16:01 XenophonF vexati0n: check out https://github.com/saltstack-formulas/salt-formula
16:01 XenophonF vexati0n: you can push config files to /etc/salt/minion.d or use file states to modify /etc/salt/minion directly
16:01 vexati0n thanks.
16:01 XenophonF vexati0n: salt-formula takes the former approach, using pillar data
16:02 vexati0n also, when setting a service and using "watch", will it watch an entire directory or just one file?
16:02 iggy it watches an id, not a file
16:02 XenophonF vexati0n: it watches the state
16:03 XenophonF vexati0n: so if the state makes any changes, the watch gets triggered
16:03 aparsons joined #salt
16:03 XenophonF vexati0n: http://docs.saltstack.com/en/latest/ref/states/requisites.html#watch
16:04 vexati0n so if i want redis to 'watch' /etc/redis.conf, i make a state called "/etc/redis.conf" where it syncs a file from the master, and a watch statement called '- file: /etc/redis.conf' ?
16:06 fivmo left #salt
16:08 riftman joined #salt
16:08 hal58th_ vexati0n, that's the jist of it
16:08 XenophonF iggy: it works!  thanks for your help.  for reference, https://github.com/irtnog/salt-states/blob/development/kerberos5/files/krb5.conf.jinja
16:08 forrest joined #salt
16:09 theologian joined #salt
16:10 hal58th_ vexati0n, you don't have to name the state "/etc/redis.conf". You could call it "my_redis.conf", or whatever, and then set "- name: /etc/redis.conf". The watch statement would then be - file: my_redis.conf
16:10 forrest iggy: You around?
16:10 iggy nope
16:11 XenophonF LOL
16:11 sgargan joined #salt
16:11 XenophonF vexati0n: here's a concrete example - https://github.com/irtnog/salt-states/blob/development/ypbind/init.sls#L25
16:11 forrest Did you see my message from yesterday regarding the formulas org and the improved github perms?
16:11 XenophonF vexati0n: bonus use of watch_in here - https://github.com/irtnog/salt-states/blob/development/ypbind/init.sls#L13
16:12 iggy forrest: yes, I tried to respond, but you were gone
16:12 genof joined #salt
16:12 forrest Yeah, the 'joys' of a locking. I'm in the office this week so I lock my machine for securitah
16:12 clintber_ joined #salt
16:12 XenophonF vexati0n: so the ypbind service watches another service state, a file state, and on some systems, a pkg state
16:12 desposo joined #salt
16:12 XenophonF securitah +1
16:13 forrest I was surprised to learn if you want your mac to lock without turning off network connectivity you have to modify the actual network device :|
16:13 XenophonF forrest: oh please show me how to do that
16:13 forrest XenophonF: You'd have to google it, there are some answers on stackoverflow
16:13 rm_jorge joined #salt
16:13 forrest I didn't do it because I didn't want to risk screwing my network adapter since I
16:13 forrest am in the office*
16:14 XenophonF forrest: awesome thanks for the clue :)
16:14 forrest XenophonF: Yeah for sure, I'm honestly tempted to read whatever docs exist, then write a salt state for it.
16:14 jespada joined #salt
16:15 XenophonF forrest: managing macs is on my to-do list
16:15 XenophonF need to master windows, first
16:16 yannis joined #salt
16:16 forrest XenophonF: Never master windows, easy solution
16:17 iggy forrest: it looks interesting, but does it solve all of our problems?
16:17 forrest iggy: It probably doesn't solve them all
16:17 forrest but some are better than nothing
16:18 unused_phd joined #salt
16:18 iggy True
16:24 yannis what is the recommended way (or ways) to upgrade an installation (RedHat 5.11/6.6)? In particular I notice that yum repos (both dl.fedoraproject.org/pub/epel and copr.fedoraproject.org/coprs/saltstack) are at 2015.5.0. I want fixes from 2015.5.2.
16:25 Auroch joined #salt
16:25 iggy epel lags behind for a reason I think
16:27 malinoff joined #salt
16:28 pressureman joined #salt
16:28 yannis I'm scanning bootstrap - it seems to use the yum repos, so it too would get me 2015.5.0 I believe
16:29 iggy until the repos upgrade, yes
16:29 hal58th_ ubuntu lags behind too.
16:29 iggy I would expect 2015.5.1 to be in epel
16:29 XenophonF yannis: maybe a newer version is in epel-testing? i'll look
16:29 yannis not 10 min ago
16:29 iggy and 2015.5.2 just came out the other day
16:29 XenophonF ah ok then
16:30 cheus joined #salt
16:30 hal58th_ joined #salt
16:30 yannis sorry - I was not talking of epel-testing - I will try that
16:31 Kiyundai Hey, I'm new to salt and i have a problem with 'onlyif' I want to see if grub and lilo are install and then reload them if they are. I'm working on a machine without lilo, so it should skip this command but it seems salt ignore the only if statement. Is there something i'm doing wrong ? https://gist.github.com/Kiyundai/c8bb6bd9aee88a6ec3a7
16:34 hal58th_ Kiyundai first thing I see wrong is "cmd.wait" only works with watch statements, not require
16:34 XenophonF Kiyundai: onlyif isn't a valid argument for cmd.wait states - http://docs.saltstack.com/en/latest/ref/states/all/salt.states.cmd.html#salt.states.cmd.wait
16:35 XenophonF oh wait never mind i can't read
16:35 XenophonF belay that
16:35 iggy you could just wrap them in jinja conditionals... it'd dirty, but it usually works
16:37 Kiyundai hmm ok don't see others options
16:37 Kiyundai thanks =)
16:38 hal58th_ Kiyundai: It might be ignoring your onlyif because the watch statement triggered...
16:38 forrest joined #salt
16:38 hvn joined #salt
16:39 hal58th_ Kiyundai: nevermind. I can't read either. Coffee hasn't kicked in...
16:40 Kiyundai Yep because without watch cmd.wait is not called no?
16:41 hal58th_ That seems true...Maybe run your state in trace and see what it is doing... Although I am not sure you will get far
16:43 iggy cmd.wait\n  name: [ -x /sbin/lilo ] && /sbin/lilo
16:43 XenophonF that's definitely a lot simpler
16:44 hal58th_ heh, linux workaround
16:44 forrest you need to look at the output of the command, is it returning a 1 or 0?
16:44 forrest that's all salt cares about
16:44 forrest if it returns some other error code, going to be problematic.
16:44 hal58th_ Well, it's looking for !0.
16:47 Kiyundai yep in the doc : Under the hood onlyif calls cmd.retcode with python_shell=True. so i try just calling this method from the command line and it returns 0 for grub and 1 for lilo
16:47 dopesong_ joined #salt
16:50 forrest So I assume you've already resolved this? /bin/bash: /sbin/lilo: No such file or directory
16:50 forrest break out your states please :(
16:50 forrest includes are not evil
16:50 writtenoff joined #salt
16:51 forrest and it would make your troubleshooting easier, and your setup more modular
16:51 Ztyx left #salt
16:51 iggy they are too
16:51 forrest I guess I'm just confused.
16:51 forrest Because this output looks fine based on your chain
16:52 forrest oh minus the stderr, gotcha
16:52 iggy if you have grub installed, but not lilo, only the grub state should run
16:52 forrest iggy: Yeah
16:52 forrest Just realized, still early
16:52 Kiyundai yep that's why i'm confuse
16:53 forrest on the actual box, what's the output of dpkg-query -s lilo
16:53 forrest You already checked the return code for a 0/1 I'm assuming and it looks good?
16:53 KyleG joined #salt
16:53 KyleG joined #salt
16:53 Kiyundai Yep i use this technique with chief and i didn't had trouble
16:54 forrest but what's the actual output of dpkg-query -s lilo? Just a blank line?
16:54 forrest since it's not found?
16:54 Kiyundai nop i have this on stdrr
16:54 Kiyundai dpkg-query: package 'lilo' is not installed and no information is available                   Use dpkg --info (= dpkg-deb --info) to examine archive files,                   and dpkg --contents (= dpkg-deb --contents) to list their contents.
16:55 Kiyundai and this when i call that sudo salt '*' cmd.retcode "dpkg-query -s lilo" saltMinion:     1
16:55 forrest weird, is this on the latest release of salt?
16:55 cruatta joined #salt
16:55 Kiyundai I'm on debian wheezy and 5.0.0
16:55 forrest what
16:55 forrest 2015.5.0?
16:55 Kiyundai yep
16:55 forrest okay
16:55 MatthewsFace joined #salt
16:55 forrest was about to say
16:56 forrest iggy: Any chance this could be related to the python shell changes you think?
16:56 cruatta_ joined #salt
16:56 wendall911 joined #salt
16:57 Matthews_ joined #salt
16:57 tomh- joined #salt
16:58 baweaver joined #salt
16:58 dharper_ii joined #salt
16:59 forrest Kiyundai: I feel like we have to be missing something super simple, not sure what though since that should be fine. Could you maybe try just doing ls /sbin/lilo for your onlyif command?
16:59 MatthewsFace joined #salt
16:59 dharper_ii Hello all, Does anyone know if it is possible to add subdirectories to the _modules directory?
16:59 riftman joined #salt
16:59 forrest need to lock my laptop so I'm dropping, back in a few.
16:59 forrest let me know if that works
16:59 dharper_ii I have tried, but the loader does not seem to be able to locate them
17:00 baweaver joined #salt
17:01 Kiyundai nop i have exactly th same output
17:02 forrest joined #salt
17:02 Kiyundai I also erase the require part just in case
17:02 hasues joined #salt
17:02 forrest Kiyundai: no luck?
17:02 iggy I don't think it's related to python shell changes
17:02 Kiyundai maybe ^^
17:04 iggy Kiyundai: can you try listen instead of watch?
17:04 iggy dharper_ii: nope
17:04 Matthews_ joined #salt
17:04 iggy the loader doesn't do subdirectories at all
17:05 dharper_ii iggy: thanks.... kind of disappointing
17:06 iggy dharper_ii: why? What are going to do? {{ salt['module.subdir.function']() }}
17:07 hal58th joined #salt
17:08 Kiyundai iggy : with listen it still run the command
17:08 hal58th_ joined #salt
17:08 hal58th__ joined #salt
17:09 Kiyundai i use listen and ls as onlyif*
17:10 MatthewsFace joined #salt
17:10 aboe joined #salt
17:12 bobby_ joined #salt
17:13 bobby_ left #salt
17:14 Kiyundai i have to leave my day of work is over, should i open an issue on git hub?
17:14 xDamox joined #salt
17:15 theologian joined #salt
17:18 riftman joined #salt
17:21 Kiyundai well i figure out a new way to solve the problem with cmd.run/onlyif/onchanges https://gist.github.com/Kiyundai/c8bb6bd9aee88a6ec3a7
17:23 aw110f joined #salt
17:26 smcquay joined #salt
17:26 aboe joined #salt
17:26 aqua^c joined #salt
17:26 denys joined #salt
17:27 otter768 joined #salt
17:27 kickerdog joined #salt
17:30 aboe left #salt
17:30 Dev0n joined #salt
17:32 TyrfingMjolnir joined #salt
17:32 cruatta joined #salt
17:33 dharper_ii iggy: we have a very diverse system, and many custom modules. I am trying to the _modules folder in a more logical and easy to find way.
17:33 aparsons joined #salt
17:33 ahammond is there currently any way to have a git ext_pillar with an ssh uri?
17:39 ndrei_ joined #salt
17:40 markm joined #salt
17:41 ajw0100 joined #salt
17:41 aboe joined #salt
17:49 unused_phd joined #salt
17:52 riftman joined #salt
17:52 Brew joined #salt
17:56 spookah joined #salt
17:57 aboe is it possible to set file.absent on a filename that ends with _ something like /etc/munin/plugins/if_
17:57 ndrei joined #salt
17:59 Ztyx joined #salt
17:59 riftman joined #salt
18:00 stoogenmeyer joined #salt
18:01 rgarcia_ joined #salt
18:02 joeto joined #salt
18:05 hal58th__ aboe can you put the name in quotes?
18:06 aboe nope yaml still bugs, about the underscore at the end of the filename
18:07 aboe hal58th I trying to set the filename with pillar, something like munin|plugins|if_
18:09 aboe I want to improve the munin-formula with basic enable/disable plugins
18:09 Ztyx joined #salt
18:10 cruatta_ joined #salt
18:11 rap424 joined #salt
18:11 impi joined #salt
18:13 hasues left #salt
18:16 igorwidl joined #salt
18:17 cilkay Hello. I have a masterless minion. There is something I'd like to run only once so I didn't add it to top.sls. How do I invoke that, which I called run-once.sls, from salt-call?
18:18 hunmaat salt-call state.sls run-once, if it's in the file root
18:20 cilkay Thank you!
18:24 pbrooko joined #salt
18:32 ageorgop joined #salt
18:35 ajw0100 joined #salt
18:39 xDamox joined #salt
18:44 pbrooko_ joined #salt
18:44 linjan joined #salt
18:50 hvn joined #salt
18:52 supersheep joined #salt
18:53 hasues joined #salt
18:54 hasues left #salt
18:54 Gareth o/
18:54 Norbell joined #salt
18:54 murrdoc joined #salt
18:56 pppingme joined #salt
18:57 iggy dharper_ii: my point was... that calling method doesn't work, so you'd have to fix a lot more than just the loader
19:01 kickerdog joined #salt
19:01 Gareth something seems odd in 2015.5, hopefully it's just my copy, but it's not handling events.
19:03 scoates Systemd docs say "Behavior of notify is similar to simple; however, it is expected that the daemon sends a notification message via sd_notify(3) or an equivalent call when it has finished starting up." ; the salt-master service is Type=notify ; does salt-master actually sd_notify? I'm seeing timeouts.
19:03 iggy Gareth: events sent from where?
19:04 iggy scoates: wouldn't surprise me if it was wrong, I doubt many people are on systemd yet
19:04 * scoates nods
19:04 scoates salt-bootstrap created it, I believe
19:04 Gareth iggy: from the minion eg. sudo salt 'localhost' event.fire '{"data":"Hello"}' 'Hello', if you're logging you should see something like: [DEBUG   ] Handling event with information about the event...but with 2015 that never shows up.
19:05 iggy Gareth: yeah, there's a bug open about it
19:05 Gareth ah cool.  got a number?
19:05 iggy not off the top of my head
19:05 agj joined #salt
19:05 iggy pretty sure event.send_master or event.fire is in the subject
19:05 scoates FWIW, changing it to `Type=simple` makes it work
19:06 aboe scoates: it does on, I have tested but you need systemd python2 to see it
19:06 * Gareth shakes iggy until he remembers
19:06 scoates aboe: can you elaborate?
19:06 aboe I know this because I build packages on opensuse, which doesn't have python2-systemd,
19:06 aboe salt-master uses notify via python2-systemd otherwise it can't
19:07 scoates ah.
19:07 phpdave11 i have a state that uses "service.running" - is there any way to get salt to print out STDERR output?  all it tells me is that the state failed to restart the service, but the STDERR messages have useful info
19:07 Gareth #24198
19:07 aboe without the proper python2 bindings, it won't send notifies to the systemd daemon
19:07 iggy so anything with systemd as the default should probably depend on python2-systemd?
19:07 aboe yes
19:07 aboe if you use notify
19:08 scoates welp. http://paste.roguecoders.com/p/71f04857e97ee04a451d33a5ecf571a6.txt
19:08 iggy Gareth: si
19:09 aboe ia no python2-systemd bindings.. so you need to change the service file see: https://build.opensuse.org/package/view_file/home:aboe76:branches:devel:languages:python/salt/use-forking-daemon.patch?expand=1
19:09 aboe this is the stop gap solution until opensuse devs, that build systemd enable python2 bindings
19:10 scoates aboe: I don't think it forks…
19:10 scoates ah. with --daemon
19:10 scoates makes sense. thanks.
19:10 iggy something similar required for minions?
19:10 aboe no the minion service file doesn't do notify
19:11 iggy rgr
19:11 scoates minion is already Type=simple
19:11 Gareth Ahh I see why it's not working for me.  Wonder if that issue has the same issue.
19:11 Gareth basically, multi-master with 2015.5 means no event firing.
19:11 iggy ahammond: ^
19:12 Gareth ahh
19:12 Gareth ahammond: o/
19:12 ajw0100 joined #salt
19:12 pbrooko joined #salt
19:12 aboe scoates: your telling me that multimaster is depend on service file to have notify?
19:14 scoates not sure again. I just installed salt-master via salt-bootstrap and ended up with `Type=notify`, and there's no `python2-systemd` package in Jessie APT.
19:14 aboe try it with my setup in the service file, just create a /etc/systemd/system/salt.service.d/local.conf to override the default salt.service file
19:15 aboe see https://wiki.debian.org/systemd/Packaging#overriding_options_and_.2Fetc.2Fdefault_handling
19:15 baweaver joined #salt
19:15 scoates you mean with this setup? https://build.opensuse.org/package/view_file/home:aboe76:branches:devel:languages:python/salt/use-forking-daemon.patch?expand=1
19:15 aqua^c joined #salt
19:15 aboe scoates: yes
19:16 aboe otherwise the multimaster problem you describe is cross platform, because opensuse also ships without python2-systemd bindings
19:16 scoates why multimaster?
19:16 aboe ow, damn, I'm messing things up in irc, gareth said multimaster...
19:17 scoates ah. ok. yeah; this is regular master.
19:18 aboe the diff is for the salt-master.service file
19:18 ndrei joined #salt
19:18 aboe https://github.com/saltstack/salt/blob/develop/pkg/salt-master.service
19:19 pcn Is there a trick swith salt.states.git.latest to get it to check out a branch?
19:22 hal58th__ pcn: rev?
19:22 xDamox joined #salt
19:23 kickerdog left #salt
19:23 pcn Ah, I had to move it further up - for whatever reason having - rev: branchname at the bottom wasn't working?
19:23 scoates aboe: works with your config. I didn't get local.conf to work (complained about multiple exec lines) but I don't quite grok systemd yet, so that's probably my fault.
19:24 aboe ow, that complain is easy, to fix, just add above the ExecStart line an empty ExecStart=
19:25 aboe systemd service file for some reason need that double Execstart with the first one just empty
19:25 sporkd2 Does setting pillars from the command line override pillar files?
19:25 pbrooko joined #salt
19:25 hal58th__ pcn that should not be a problem ever...
19:26 MTecknology golly gosh willickers batman! Upgrading salt-minion on windows via salt got insanely simple!
19:26 iggy sometimes it overwrites, sometimes it merges
19:27 scoates aboe: I *want* to like systemd… (-:
19:28 aboe ow, that's just the systemd implementation of debian, which isn't finnished.
19:28 aboe scoates: I'm using it on archlinux and haven't had that much trouble as with jessie
19:28 otter768 joined #salt
19:28 * scoates nods
19:28 aboe something about debian jessie systemd feels like it was forced
19:28 scoates "isn't finished" is a bit scary, then; heh
19:29 aboe ntp, daemon is still sysv, mysql is still sysv
19:29 scoates the worst thing about systemd for me so far is that the actual information on it is eclipsed by "debate" (whining)
19:29 aboe yes
19:30 scoates sincerely hoping I can replace out supervisors with systemd services
19:30 aboe it should, archlinux wiki will help , if you want to learn more about systemd
19:30 scoates *our
19:30 scoates thanks. good to know.
19:31 AdobeTom Does anyone know if there is a way to set the entrypoint on a docker.running state in 2015.05?  I only see command, but that gets used as arguments to the existing entrypoint.
19:38 soren joined #salt
19:40 chiui joined #salt
19:49 gladiatr joined #salt
19:51 Not_ joined #salt
19:51 hvn joined #salt
19:53 jdesilet joined #salt
19:54 hasues joined #salt
19:55 Andre-B joined #salt
19:55 hasues left #salt
19:57 vexati0n how do i use "include" in an SLS to include a state from another top-level directory
19:59 teryx510 joined #salt
20:03 vexati0n eh. nevermind, got it.
20:05 scoates_ joined #salt
20:06 sgargan_ joined #salt
20:07 teryx510 Hi guys, so today I tried to use file.managed content_pillar and found that it was creating an empty file. I checked and the pillar data was available on the minion. Eventually I found that it was because I was using numbers in my pillar data. Seems like a bug to me but I thought I would ask first.
20:09 jchadwick joined #salt
20:10 JoeHazzers joined #salt
20:12 iggy teryx510: I generally use multi-line string literals in yaml if I know it's going to be used elsewhere (or quotes if I'm lazy)
20:14 ggoZ joined #salt
20:15 LiamM when using s3 as the source of a file in a salt-state how can I set the s3.service_url in pillar or the state itself?
20:16 igorwidl for some reason, when I run salt states against one of my minions, i do not get any output
20:17 iggy igorwidl: it's probably timing out (the command still finishes, just the salt cli tools don't wait forever)
20:18 igorwidl iggy: i thought so too, but even with --timeout it just quits without output
20:18 iggy chheck the job cache
20:18 igorwidl hmm, alright, where are those stored in?
20:19 iggy there's runners and stuff in saltutil for accessing the job data
20:21 igorwidl if i do salt-run job.lookup_jid immediately after salt commends exists, i can get the output for that job. so it does return to server, just does not display for some reason
20:21 XenophonF left #salt
20:22 Not_ joined #salt
20:22 Matthews_ joined #salt
20:24 slav0nic joined #salt
20:25 xDamox joined #salt
20:26 toastedpenguin joined #salt
20:28 toastedpenguin joined #salt
20:32 MatthewsFace joined #salt
20:35 aboe left #salt
20:36 conan_the_destro joined #salt
20:36 linjan joined #salt
20:37 iggy that's how the job cache works
20:37 iggy (unless you tell it to return to something other than the master)
20:45 xDamox joined #salt
20:47 MatthewsFace joined #salt
20:49 Ztyx joined #salt
20:49 ShibMcNe_ joined #salt
20:51 unused_phd joined #salt
20:52 hvn joined #salt
20:52 hvn joined #salt
20:54 Matthews_ joined #salt
20:56 hal58th_1 joined #salt
20:56 hal58th_2 joined #salt
20:58 hal58th_3 joined #salt
21:00 MTecknology iggy: I brought 174 production servers (non-redundant) from Debian 7 to Debian 8 today and in that process also brought them to using salt-minion 2015.5. Holy crappers man... I'm loving the changes.
21:00 sgargan joined #salt
21:00 MTecknology I managed to do it with nobody noticing enough down time to call and yell. :D
21:01 Norrland MTecknology: upgrade with salt or other method for upgrading?
21:01 MTecknology the upgrade to debian version happened with me connected via ssh
21:01 rgarcia_ joined #salt
21:02 Norrland ah
21:02 MTecknology That kinda change is something that salt can't and shouldn't ever do.
21:03 ajw0100_ joined #salt
21:04 aqua^c joined #salt
21:07 MTecknology .. especially handling these particular package conflicts
21:08 Norrland heh yeah.
21:09 baweaver joined #salt
21:09 Norrland So.. one by one?
21:09 Norrland that ought to take you more than a full day right?
21:09 Norrland depending on how much variety of services running on each
21:10 iggy MTecknology: did you notice any connectivity/network differences in Salt yet?
21:10 iggy it was a night and day difference for us going from 2014.1 -> 2015.5
21:11 iggy but we're in GCE which is fully routed, so it's kind of to be expected (after finding out what the actual problem was... not setting socket keepalive)
21:14 MTecknology same here.. difference is absurd
21:14 hal58th_1 yeap, no more minions taking forever to wake up
21:14 iggy yeah, not surprised you see some benefit there too... a lot of routers in your setup as well iirc
21:14 sgargan joined #salt
21:17 sgargan joined #salt
21:18 ksalman In a large test environment, where people are bringing up their own VMs, how do you make sure that the minion ids are unique? I see these errors due to duplicate ids "An extra return was detected from minion example.com, please verify the minion, this could be a replay attack"
21:19 jonlangemak joined #salt
21:19 Georgyo left #salt
21:24 sgargan joined #salt
21:24 tmclaugh[work] joined #salt
21:26 stanchan joined #salt
21:28 iggy don't accept the key if it already exists?
21:28 sgargan joined #salt
21:28 iggy if your hostnames are unique (if they aren't, how the hell do you work), the minion_id's should be unique
21:29 otter768 joined #salt
21:30 hal58th_1 ksalman, one second. I wrote a solution
21:30 ksalman iggy: yes, they should be unique, but unfortunately i can't force that on poeples VMs
21:30 ksalman like, multiple people will spin up a vm named client1
21:31 hal58th_1 echo -n "dev-mapper-" > /etc/salt/minion_id
21:31 hal58th_1 openssl rand -hex 6 >> /etc/salt/minion_id
21:31 ksalman hal58th_1: that's a nice work around
21:31 hal58th_1 just put that in a vagrant script
21:31 iggy either way, you are going to have to train people to stop doing that
21:32 iggy or have completely non-sensical minion_id's like that
21:32 hal58th_1 it works for my vagrant boxes that connect to a salt master that's not vagrant
21:32 ksalman I guess I'll have to go with the non-sensical route..
21:32 rgarcia_ joined #salt
21:33 ksalman hal58th_1: thanks
21:38 hal58th_1 welcome ksalman
21:40 baweaver joined #salt
21:42 anotherZero joined #salt
21:42 sgargan joined #salt
21:42 giantlock joined #salt
21:45 sgargan joined #salt
21:46 yawniek joined #salt
21:48 nzero joined #salt
21:53 hvn joined #salt
21:54 mudfly iggy: you helped me look at pkg.update_db a few days ago, I reported a bug on github and it did end up being a regression. Is now backported to 2015.5.1
21:56 sgargan joined #salt
21:58 baweaver joined #salt
22:02 iggy excellent
22:03 tmclaugh[work] joined #salt
22:07 keimlink joined #salt
22:08 ajw0100 joined #salt
22:08 MatthewsFace joined #salt
22:09 keithhub joined #salt
22:10 amcorreia joined #salt
22:10 Matthews_ joined #salt
22:11 sgargan joined #salt
22:14 ahammond Gareth we're single master, not multi.
22:15 Gareth ahammond: just saw your comment on the Github issue.  Wondering if it's still hitting the same issue though.
22:15 Gareth ahammond: are you comfortable modifying code in place on a minion where you're seeing the issue?
22:18 joeto1 joined #salt
22:19 Ztyx joined #salt
22:19 Nazca__ joined #salt
22:21 packetherder joined #salt
22:21 MatthewsFace joined #salt
22:22 packetherder left #salt
22:24 viq joined #salt
22:26 Nazca joined #salt
22:27 nzero joined #salt
22:27 gmeno hello, I've been having some trouble with 7c537cd9ece84adedce45e8da371a08f21291a25 It seems to cause context switched in my gevent code when sending events
22:28 gmeno ThomasJ: ^ look like your commit. are you around?
22:32 Nazca__ joined #salt
22:38 alexthegraham joined #salt
22:38 alexthegraham Anyone figured out a good way to do arithmetic comparison on grain values?
22:39 supersheep joined #salt
22:39 alexthegraham For example, I want to target all minions with a grain value greater than 13.1.
22:40 iggy jinja has an int filter
22:41 iggy and a float filter ;)
22:41 alexthegraham Hah.
22:52 aqua^c joined #salt
22:54 hvn joined #salt
22:55 MindDrive left #salt
22:55 MindDrive joined #salt
22:57 MindDrive Trying to run a 'saltutil.sync_modules' on multiple systems on one of my tiers and I'm getting '__virtual__ returned False' for every system, but I haven't made any recent changes... I'm seeing failures for 'module.tls', 'module.nacl', 'module.ipmi' and 'module.npm', none of which I believe I use and I've never seen these errors before.  I'm running 2015.5.0, is this a known issue?
22:57 cruatta joined #salt
23:01 amranu joined #salt
23:01 cruatta_ joined #salt
23:05 iggy those "errors" are normal (for 2015.5)... I complained about them to no avail before the release
23:06 iggy as far as your actual problem goes... doesn't sound familiar
23:07 mosen_ joined #salt
23:08 MindDrive It's a bit disconcerting that I can't do a syncutil.sync_modules from the server now... I did set up gitfs recent (for the modules), but that shouldn't affect it, should it?
23:10 iggy are you getting the virtual thing on the actual sync or when trying to use modules after the sync?
23:10 MindDrive On the actual sync.
23:10 bfoxwell joined #salt
23:10 MindDrive I'm going to try to use the module on one set of servers right now, hang on...
23:12 MindDrive Okay, the module definitely worked.
23:19 iggy is the virtual thing related to one of those modules you don't use? if so, it's safe to ignore
23:19 MindDrive Oh, hmm, hadn't thought of that.  I definitely (currently) don't use any of the modules complaining.
23:19 MindDrive Okay, I'll ignore for now, thanks.
23:20 MindDrive (Hopefully this will be fixed in an upcoming release. :) )
23:20 hal58th joined #salt
23:20 hal58th_ joined #salt
23:20 hal58th__ joined #salt
23:21 iggy I spent most of my effort pre-2015.2 trying to get the cmd.run debacle cleared up
23:21 Nazca joined #salt
23:22 teryx510 joined #salt
23:24 smcquay joined #salt
23:27 Nazca__ joined #salt
23:28 joeto joined #salt
23:29 sgargan joined #salt
23:30 otter768 joined #salt
23:30 evilrob joined #salt
23:34 Nazca joined #salt
23:36 Ztyx joined #salt
23:39 Nazca__ joined #salt
23:41 sgargan joined #salt
23:42 aqua^c joined #salt
23:54 SheetiS joined #salt
23:55 hvn joined #salt
23:55 alexanderilyin joined #salt
23:56 alexanderilyin Hi, is  there a wasy to mark state cmd.run as vailed if exist code is 0 but stderr is not empty
23:56 alexanderilyin ?
23:57 SheetiS alexanderilyin: do you need the stderr stuff to stay stderr?  Could just use 2>&1 at the end
23:57 SheetiS that way only the exit code mattered
23:57 Not_ joined #salt
23:58 alexanderilyin SheetiS: I mean that is there is an error in cmd.run call executable will exist with 0 but will print errors to stderr
23:58 alexanderilyin so i want handle it and fails state
23:59 alexanderilyin but looks like I have write small bash for this and check stderr by myself and change exit code

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