Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-08-21

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

All times shown according to UTC.

Time Nick Message
00:12 Akhter joined #salt
00:19 bluenemo joined #salt
00:25 favadi joined #salt
00:29 Furao joined #salt
00:30 dthom91 joined #salt
00:55 kevinquinnyo joined #salt
01:04 baweaver joined #salt
01:17 Bryson joined #salt
01:25 writtenoff joined #salt
01:27 breakingmatter joined #salt
01:27 pravka joined #salt
01:30 dthom91 joined #salt
01:31 wendall911 joined #salt
01:36 pm90_ joined #salt
01:38 DammitJim joined #salt
01:41 phw_ joined #salt
01:42 chadhs joined #salt
01:45 chadhs joined #salt
01:45 dthom91 joined #salt
01:46 chadhs joined #salt
01:46 dthom91 joined #salt
02:03 dthom91 joined #salt
02:07 DammitJim joined #salt
02:11 dthom91 joined #salt
02:11 zsoftich joined #salt
02:41 evle joined #salt
02:54 dustywusty joined #salt
02:55 dustywusty joined #salt
02:56 favadi joined #salt
02:57 arthoo joined #salt
03:04 Furao joined #salt
03:07 dustywusty joined #salt
03:10 jdesilet joined #salt
03:11 cheus joined #salt
03:15 RandyT_ joined #salt
03:18 pravka joined #salt
03:21 malinoff joined #salt
03:37 ITChap joined #salt
04:05 Diaoul joined #salt
04:10 malinoff joined #salt
04:12 hasues joined #salt
04:12 hasues left #salt
04:13 TyrfingMjolnir joined #salt
04:32 RandyT_ joined #salt
04:38 cachedout joined #salt
04:41 overyander joined #salt
04:42 overyander joined #salt
04:42 zer0def joined #salt
04:45 mattiasr joined #salt
04:47 RandyT_ joined #salt
04:48 mattiasr joined #salt
04:51 Antiarc joined #salt
05:01 wendall911 left #salt
05:05 arthoo joined #salt
05:07 jaybocc2 joined #salt
05:11 ageorgop joined #salt
05:11 harkx joined #salt
05:21 darsh_ joined #salt
05:25 chadhs joined #salt
05:28 breakingmatter joined #salt
05:33 pm90_ joined #salt
05:34 yomilk joined #salt
05:38 overyander joined #salt
05:39 overyander how can i have a state run at a certain time? set a cron job to run the command?
05:39 overyander or is there something built into salt to accomplish this?
05:42 jeadre joined #salt
05:47 DanyC left #salt
05:50 mrjester joined #salt
05:59 colttt joined #salt
06:05 Ztyx joined #salt
06:05 hardwire joined #salt
06:06 katyucha_ joined #salt
06:07 katyucha_ joined #salt
06:12 katyucha_ joined #salt
06:16 gcfhvjbkn joined #salt
06:16 evle joined #salt
06:19 DanyC joined #salt
06:23 rdas joined #salt
06:23 borgstrom joined #salt
06:23 big_area joined #salt
06:26 AndreasLutro joined #salt
06:27 totte left #salt
06:29 Furao joined #salt
06:36 catpig joined #salt
06:38 katyucha joined #salt
06:39 pm90__ joined #salt
06:43 Ztyx1 joined #salt
06:49 GrueMaster joined #salt
06:50 KermitTheFragger joined #salt
06:52 dopesong joined #salt
06:54 harkx joined #salt
06:54 viq joined #salt
06:57 Ztyx joined #salt
07:00 phw_ What is best practice to keep multiple computers up to date with highstates if they are not always running? Create a cronjob to ensure the state on a regular bassi?
07:05 intel joined #salt
07:09 DanyC left #salt
07:09 jhauser joined #salt
07:10 rdas joined #salt
07:12 DanyC joined #salt
07:14 Ztyx joined #salt
07:15 linjan_ joined #salt
07:18 DanyC phw_: when you say "computeres not always running" you mean when the minions are down ?
07:19 eseyman joined #salt
07:22 lb1a joined #salt
07:26 phw_ DanyC: yes, as most of them are desktop machines this is the case
07:30 breakingmatter joined #salt
07:30 hvn joined #salt
07:34 kawa2014 joined #salt
07:34 hvn joined #salt
07:35 illern joined #salt
07:35 zerthimon joined #salt
07:43 NV joined #salt
07:49 mattiasr joined #salt
07:52 DanyC phw_: if minions are down you have only 1 option imo: "ship" the states to each computer and run the local highstate on each node. Otherwise you can use salt-ssh to start the minion on each computer and then salt-call '*' highstate
07:53 sunkist joined #salt
07:54 phw_ DanyC: ah no, the computers are physically off - so I could just run highstate multiple times during the day or write a script to check all computers step by step
07:57 AndreasLutro phw_: http://docs.saltstack.com/en/latest/ref/states/startup.html this should ensure that a highstate is ran when the machine starts
07:58 phw_ AndreasLutro: ah, that is exactly what would help! Thank you!
08:06 bluenemo joined #salt
08:06 mattiasr joined #salt
08:07 DanyC AndreasLutro: yours is cleaner indeed ;)
08:09 s_kunk joined #salt
08:12 _mel_ joined #salt
08:17 jmreicha joined #salt
08:18 Xevian joined #salt
08:26 teryx510 joined #salt
08:30 GreatSnoopy joined #salt
08:42 hvn joined #salt
08:42 supersheep joined #salt
08:43 supersheep joined #salt
08:45 DanyC joined #salt
08:46 DanyC left #salt
08:46 DanyC joined #salt
08:49 Grokzen joined #salt
08:51 dkrae joined #salt
08:52 hojgaard joined #salt
09:02 hvn joined #salt
09:08 favadi joined #salt
09:12 shantanoo joined #salt
09:12 kawa2014 joined #salt
09:12 supersheep joined #salt
09:27 mike25de left #salt
09:31 pol joined #salt
09:37 onovy joined #salt
09:37 sgargan joined #salt
09:43 CeBe joined #salt
09:44 gcfhvjbkn joined #salt
09:48 shantanoo joined #salt
09:50 shantano1 joined #salt
09:58 aqua^c joined #salt
09:58 yomilk joined #salt
10:01 shantanoo joined #salt
10:01 shantanoo joined #salt
10:01 shantanoo joined #salt
10:11 thayne joined #salt
10:18 seblu joined #salt
10:18 seblu joined #salt
10:20 stephanbuys joined #salt
10:23 TyrfingMjolnir joined #salt
10:32 pviktori joined #salt
10:42 Andre-B joined #salt
10:52 jngd joined #salt
10:52 evle1 joined #salt
10:54 kawa2014 joined #salt
11:02 giantlock joined #salt
11:07 breakingmatter joined #salt
11:14 yomilk joined #salt
11:21 jaybocc2 joined #salt
11:25 Driskell joined #salt
11:29 pravka joined #salt
11:34 breakingmatter joined #salt
11:38 codehotter There's a tag for 2015.5.4, but it isn't released, right? Is there a public timeline somewhere?
11:39 codehotter Or is it because of the performance regression?
11:43 AndreasLutro codehotter: correct
11:44 codehotter When 2015.5.5 is tagged, a package will appear in EPEL repo soon?
11:44 codehotter but, 2015.5.4 that didn't happen because of the performance regression
11:44 AndreasLutro correct again
11:44 codehotter Thanks!
11:45 orion Hi. In a state file I have a declaration called "foo-group", where I use the group.present function to create a group called "foo". When I write a requisite declaration within my "foo-user" declaration, should my requisite reference name "- group: foo" or "- group: foo-group"?
11:46 AndreasLutro orion: both will work, I prefer foo-group
11:48 dopesong joined #salt
11:48 orion AndreasLutro: So, you're saying that the requisite reference can refer either to the actual resource on the system "group" or the salt ID?
11:48 AndreasLutro orion: yes - to be exact, requisites can match on state name or state ID
11:50 orion Ah, ok. Thank you for clarifying
11:54 katyucha Hi all. Can I add some variable in salt-cloud profile ? I mean : i want to make a variable about information i push in guestinfo vmware. THx
11:56 _JZ_ joined #salt
12:00 codehotter To use states instead of maps to deploy cloud instances, I need to put a salt-minion on the server running salt-cloud, right?
12:00 auzty joined #salt
12:02 rodio_ua joined #salt
12:03 rodio_ua left #salt
12:08 sieve joined #salt
12:08 sieve Is it just me or are saltstack releases becoming progressively more unstable?
12:09 sieve I am seeing lots of wierd problems with salt 2015.5.3 (Lithium)
12:09 sieve Finding a stable version is now nearly impossible I think
12:12 ahale joined #salt
12:15 AndreasLutro I just work off the git head
12:15 AndreasLutro bugfix pull requests are merged within 24 hours so it's pretty reliable
12:15 AndreasLutro especially if you pick a branch where there's a feature freeze
12:22 jngd anyone knows a good formula to install mongodb and create a collection? I am good with one sls file, it’s just for show an example. Thanks in advance!
12:26 tmclaugh[work] joined #salt
12:27 jdesilet joined #salt
12:28 DanyC_ joined #salt
12:29 Furao_ joined #salt
12:30 fredvd joined #salt
12:30 Ztyx joined #salt
12:31 orion On my Ubuntu desktop, what package do I need to install if I *only* want to work with salt-cloud?
12:33 orion I read that salt-cloud has been merged in to "salt proper", but I don't know what salt proper is.
12:33 AndreasLutro orion: salt-cloud - but it doesn't really make a difference in size
12:33 AndreasLutro salt-cloud used to be an entirely separate project, now it's part of salt
12:33 rideh joined #salt
12:33 orion I see.
12:36 rideh joined #salt
12:37 rawkode joined #salt
12:39 numkem joined #salt
12:39 pravka joined #salt
12:39 evle joined #salt
12:40 Ztyx joined #salt
12:41 EnManiskMan joined #salt
12:43 favadi joined #salt
12:46 orion When a VM is created with salt-cloud, how does the bootstrap script get added to the machine? Does salt-cloud SSH in to the box? If so, can the username/password/etc is uses be configured?
12:47 sunkist joined #salt
12:47 AndreasLutro check your vm provider's config options - usually that's where you set stuff like root password on vm creation
12:50 orion AndreasLutro: Ah, indeed. Can the deploy script be completely disabled if I baked salt-minion in to my AMI?
12:50 AndreasLutro orion: probably by setting the bootstrap config to null
12:50 AndreasLutro in /etc/salt/cloud: script: null
12:51 orion Great! Thank you so much.
12:59 BlackJackAc3 joined #salt
13:00 Saltstack_newbie joined #salt
13:02 Saltstack_newbie Hello. I have written a new state file (my-state/init.sls) and then added "my-state" to the correct hosts under top.sls. For some reason the state is not run when I do state.highstate. Running it manually works, though, salt <host> state.sls my-state. How do I debug this?
13:03 AndreasLutro Saltstack_newbie: confirm that the minion id is what you think it is: salt <host> grains.get id
13:03 AndreasLutro Saltstack_newbie: also you can share your top.sls
13:04 AndreasLutro wait, if you're doing salt <id> you obviously know the id õ_õ
13:06 DammitJim joined #salt
13:06 TyrfingMjolnir joined #salt
13:07 Saltstack_newbie yes, I'm 100% sure the ID is right. As I said, running the state manually works. What I am looking for is a way to debug this. I'm sure this is very easy to fix, probably something basic. I just have no idea how to debug saltstack configurations
13:08 ckng joined #salt
13:09 dalom joined #salt
13:11 icflournoy joined #salt
13:14 cpowell joined #salt
13:14 jngd joined #salt
13:15 TyrfingMjolnir joined #salt
13:16 AndreasLutro Saltstack_newbie: try adding -l debug to the command, that'll give you a bit more verbose output
13:16 pm90_ joined #salt
13:16 cpowell joined #salt
13:17 pm90__ joined #salt
13:18 BlackJackAc3 joined #salt
13:19 shantanoo joined #salt
13:21 supersheep joined #salt
13:22 jaybocc2 joined #salt
13:22 rdas joined #salt
13:25 shantanoo joined #salt
13:25 Saltstack_newbie Thanks, but that didn't output anything useful. I'll just run it manually for now, I think I will have to redo the entire configuration this weekend anyway. This configuration is a mess of states and other files without any clear pattern
13:26 BlackJackAc3 joined #salt
13:36 BlackJackAc3 joined #salt
13:36 Akhter joined #salt
13:41 Akhter joined #salt
13:44 raakesh joined #salt
13:45 mapu joined #salt
13:46 yomilk joined #salt
13:48 dthom91 joined #salt
13:48 kermit joined #salt
13:51 chadhs joined #salt
13:51 BlackJackAc3 joined #salt
13:52 PPpyg2 joined #salt
13:54 DanyC joined #salt
14:00 gcfhvjbkn joined #salt
14:00 arthoo joined #salt
14:02 gcfhvjbkn joined #salt
14:03 dyasny joined #salt
14:04 giantlock joined #salt
14:05 zmalone joined #salt
14:06 pdayton joined #salt
14:07 andrew_v joined #salt
14:07 sroegner joined #salt
14:13 bluenemo joined #salt
14:14 drawsmcgraw joined #salt
14:19 perfectsine joined #salt
14:21 pdayton1 joined #salt
14:23 jaybocc2 joined #salt
14:24 pdayton2 joined #salt
14:29 jartsu joined #salt
14:30 dthom91 joined #salt
14:30 stephanbuys joined #salt
14:31 amcorreia joined #salt
14:33 pdayton joined #salt
14:34 Brew joined #salt
14:35 Brew joined #salt
14:38 pdayton joined #salt
14:39 supersheep joined #salt
14:46 pdayton joined #salt
14:46 pdayton joined #salt
14:46 yomilk joined #salt
14:50 pdayton1 joined #salt
14:51 DanyC joined #salt
14:51 DanyC left #salt
14:51 DanyC joined #salt
14:57 iggy for future reference, the next step would have been `salt-call -l debug state.highstate` it'll show the rendered top file and state files
15:03 andrew_v_ joined #salt
15:04 dthom91 joined #salt
15:05 jngd joined #salt
15:06 Arbusto joined #salt
15:06 cachedout joined #salt
15:08 dalom joined #salt
15:09 Arbusto For a custom fileserve module, is there anyway to determine the minion asking for the state?
15:10 sdm24 joined #salt
15:15 DanyC joined #salt
15:20 jngd Any mongodb formula to install package, start service and create user and password? Help, please!
15:21 jngd no replica set looking for
15:21 jngd only this
15:21 jngd basic operations
15:21 iggy unless someone changed it, the mongodb-formula could do that
15:23 Bendoin joined #salt
15:23 Bendoin left #salt
15:29 raakesh joined #salt
15:31 Furao joined #salt
15:32 dthom91 joined #salt
15:35 zsoftich3 joined #salt
15:38 JDiPierro joined #salt
15:40 JDiPierro Hey all. Running salt 2014.7.6 Helium. When I execute a highstate through the master with 'salt machine-name state.highstate' I'm not getting anything returned.. it 'hangs' after I issue the command then finishes but doesn't print anything, just gives me a new prompt... any ideas what would cause this? I don't see anything in the master config about it. If I execute a "salt-call state.highstate" on the minion i still get the output
15:40 orion Can salt-cloud profiles have Jinja in them?
15:40 iggy timeout maybe
15:41 iggy or the minion restarts and doesn't return anything
15:41 iggy orion: no
15:41 orion My goal is to use salt-cloud to spin up three VMs, but each VM needs to be in a different subnet.
15:41 orion Do I need to create a separate cloud profile for each subnet?
15:42 iggy they support (limited) inheritance
15:42 dthom91 joined #salt
15:43 dimeshake I seem to be running into an issue with listing jobs - I can't pull any historical job_ids. I tried stopping master, removing /var/cache/salt/master/jobs and restarting, a test.ping to minions, and still can't get any results...
15:43 dimeshake http://pastie.org/private/qvpw7c9ka60dnqhhvzluqq
15:43 dimeshake anyone seen this before?
15:44 shaggy_surfer joined #salt
15:44 orion iggy: Based on my use case, do you recommend inheritance, or do you recommend something else?
15:44 iggy dimeshake: do you have a non-standard job cache enabled?
15:44 RedundancyD joined #salt
15:44 iggy orion: It sounds like it could work for you
15:45 dimeshake iggy: just ext_job_cache for minions to report to syslog
15:45 dimeshake ext_job_cache: syslog   and  job_cache: True
15:46 dimeshake maybe i misunderstood but i thought you could enable local job cache with that second option in conjunction :D
15:46 malinoff joined #salt
15:46 iggy dimeshake: no
15:47 dimeshake alrighty. thanks
15:48 iggy "This will also disable the local job cache on the master"
15:48 iggy not this will disable the job cache unless you also set job_cache: True
15:50 dthom91 joined #salt
15:50 dimeshake iggy, yeah i see -- but what's weird is that jobs are still showing and present under /var/cache/salt/master/jobs -- and i can look at .load.p in a job dir and see what was run
15:51 sieve joined #salt
15:51 iggy yeah, I think that's a peculiarity of the way ext_job_cache was hacked on
15:51 orion iggy: Do you know where inheritance is documented? I am having trouble finding it.
15:52 iggy they converted the job functions to a lookup based off the ext_job_cache setting.... and just didn't implement the lookup into any of the !local job_cache modules
15:53 iggy orion: the arg is `extends`
15:53 dimeshake gotcha iggy - thanks.
15:53 iggy orion: and a quick note (because I missed it), it only goes one level deep, so no multi-level inheritance
15:55 Bryson joined #salt
15:55 shaggy_surfer joined #salt
15:58 orion iggy: Alright.
15:59 baweaver joined #salt
16:00 orion Ah, it seems that since I need a different subnet for each instance, I therefore need a different profile for each instance.
16:01 orion Inheritance can certainly help me eliminate redundancy though.
16:02 yomilk joined #salt
16:04 ageorgop joined #salt
16:06 LinqSocket joined #salt
16:07 whytewolf orion: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.cloud.html#salt.states.cloud.present that might save you some profile configureation
16:07 whytewolf [not provider however]
16:12 pdayton joined #salt
16:13 pdayton joined #salt
16:13 zwi joined #salt
16:14 mpanetta joined #salt
16:16 pdayton joined #salt
16:16 wendall911 joined #salt
16:16 tiadobatima joined #salt
16:17 writtenoff joined #salt
16:23 jaybocc2 joined #salt
16:23 Furao joined #salt
16:24 Akhter joined #salt
16:32 pdayton1 joined #salt
16:36 capricorn_1 joined #salt
16:37 aparsons joined #salt
16:37 jalbretsen joined #salt
16:40 rdutch joined #salt
16:41 crd joined #salt
16:44 dthom91 joined #salt
16:47 thehaven joined #salt
16:48 Akhter joined #salt
16:48 baffle__ joined #salt
16:49 thayne joined #salt
16:50 baweaver joined #salt
16:52 thehaven joined #salt
16:54 wendall911 joined #salt
16:55 baweaver joined #salt
17:01 sn00py joined #salt
17:02 awpti joined #salt
17:03 dthom91 joined #salt
17:04 JDiPierro joined #salt
17:06 pdayton joined #salt
17:09 DanyC_ joined #salt
17:11 aqua^c joined #salt
17:11 Gareth morning morning
17:12 pdayton1 joined #salt
17:19 stilezayk joined #salt
17:19 cbookman joined #salt
17:19 rideh joined #salt
17:20 bluenemo joined #salt
17:22 stomith Hey, new to salt, been using Puppet. Is there a equivalent to salt like the Puppet Forge?
17:22 pdayton joined #salt
17:22 dimeshake stomith: not quite like that. but there are salt formulas. github.com/saltstack-formulas/
17:22 stomith oh, great.
17:22 baweaver joined #salt
17:24 forrest joined #salt
17:24 stomith and I saw some support for vsphere / esx? is that out of the box, or a paid feature?
17:26 rdutch stomith: also if you looking to deploy windows software, look here for predefined configs: https://github.com/saltstack/salt-winrepo
17:26 forrest stomith: It being known as vsphere has been deprecated in favor of calling it vmware: Still manages vsphere/esx though: http://docs.saltstack.com/en/latest/ref/clouds/all/salt.cloud.clouds.vmware.html#module-salt.cloud.clouds.vmware
17:26 forrest and it's free of course.
17:28 giantlock joined #salt
17:28 ajw0100 joined #salt
17:29 stomith sweet! I'm migrating. :D
17:29 stomith thanks for your help
17:30 forrest NP
17:30 forrest stomith: There's also the getting started guide: http://docs.saltstack.com/en/latest/topics/cloud/vsphere.html but it's pretty minimal, and uses the deprecated naming. I'll create an issue to get that updated.
17:31 XenophonF the wsgi version of rest_cherrypy doesn't seem to be able to talk to the salt-master  - would anyone have any suggestions as to how i might go about debugging this?
17:32 BlackJackAc3 joined #salt
17:32 forrest XenophonF: Pretty sure someone else had this problem and couldn't get it working, they ended up using tornado. I'd say see if an issue exists for it, if not can you please create one?
17:33 XenophonF i'll go looking
17:33 dthom91 joined #salt
17:33 mattiasr joined #salt
17:33 mattrobenolt joined #salt
17:33 XenophonF salt-api (using rest_cherrypy) works fine, but it's running as root, whereas the wsgi version's running as www
17:33 zer0def joined #salt
17:34 jeadre joined #salt
17:37 XenophonF i'd like to run the wsgi app in debug mode
17:37 XenophonF i'll have to figure out how to do that
17:38 dthom911 joined #salt
17:38 dalom left #salt
17:43 dthom91 joined #salt
17:44 X67r joined #salt
17:44 linjan_ joined #salt
17:45 BlackJackAc3 joined #salt
17:46 KyleG joined #salt
17:46 KyleG joined #salt
17:47 ajw0100 joined #salt
17:48 ngks joined #salt
17:50 aw110f joined #salt
17:51 yomilk joined #salt
17:52 stomith forrest, thank you very much. I'm stoked.
17:52 forrest You're welcome.
17:54 rm_jorge joined #salt
18:01 artemz joined #salt
18:02 forrest stomith: https://github.com/saltstack/salt/issues/26560#issuecomment-133506553 Looks like the vmware stuff might not be in the current stable release, docs list an inaccurate addition date, so you might have to use the vsphere one. Sorry for the confusion, I just learned this myself.
18:02 saltstackbot [#26560]title: Vsphere getting started guide should be updated/totally deprecated | It looks like http://docs.saltstack.com/en/latest/topics/cloud/vsphere.html is still alive even though it has a link to the vmware docs. This document should probably be removed, or a getting started guide written for the vmware stuff (in addition to what currently exists doc wise) that this can link to.
18:03 mattiasr joined #salt
18:03 surge__ joined #salt
18:05 surge__ left #salt
18:05 dthom91 joined #salt
18:07 aparsons joined #salt
18:09 amit_ joined #salt
18:10 DanyC joined #salt
18:11 dthom91 joined #salt
18:11 BlackJackAc3 joined #salt
18:14 amit_ hi , i'm relatively new to salt. i'm using salt 2015.5.2 (Lithium) and 'm facing a conflicting ids issue which has flummoxed me a bit.
18:14 amit_ here is what the log looks like : https://gist.github.com/amits83/37e0b41f7389057df418#file-gistfile1-txt
18:17 forrest hey amit_: So on line 11 I can see the conflicting ID names: https://gist.github.com/amits83/37e0b41f7389057df418#file-gistfile1-txt-L11, do you have multiple apache state files? If not, and it's just Apache/apache_conf could you gist that state?
18:18 Akhter joined #salt
18:19 BlackJackAc3 joined #salt
18:19 ageorgop joined #salt
18:21 amit_ https://gist.github.com/amits83/9b3b5bcd4eb00fd247a1#file-top_and_sls
18:26 katyucha joined #salt
18:30 jeadre joined #salt
18:32 Arbusto joined #salt
18:33 mattiasr joined #salt
18:34 DanyC joined #salt
18:34 pravka joined #salt
18:36 DanyC left #salt
18:36 DanyC joined #salt
18:36 pdayton joined #salt
18:40 forrest stomith: Looks like it was backported to 2015.5.4/5 according to rallytime on that issue
18:40 forrest amit_: sorry I don't look at the IRC unless I'm tagged usually, hang on
18:40 ngks joined #salt
18:41 forrest amit_: Is there any chance there's a conflict in some of the 'some stuff' section? Like a double include or something weird going on there?
18:41 forrest honestly I don't see why there should be a problem
18:41 hasues joined #salt
18:42 mrwboilers In a multi-master configuration can/should you manage /etc/salt/master with salt?
18:43 rsimpkins joined #salt
18:43 rsimpkins left #salt
18:43 rsimpkins joined #salt
18:43 mrwboilers This would allow me to avoid having to edit the file and restart salt-master manually every time I make changes to a nodegroup
18:44 whytewolf mrwboilers: in a non-multimaster setup i manage salt with salt, I see no reason not to do the same in a mutlimaster setup.
18:44 gerhardq1x amit_: so Apache (dot) apache_conf conflicts with Apache (slash) apache_conf,
18:45 gerhardq1x you should figure out where the slash is coming from
18:45 mrwboilers whytewolf: thanks! I'll give it a try
18:46 whytewolf mrwboilers: just remeber use listen or listen_in for restarting the salt services. that way they are the last thing restarted
18:47 BlackJackAc3 joined #salt
18:47 rsimpkins I'm seeing a problem with schedules set via pillar. The splay functionality does not seem to be working (appears to be ignored). I can't find any issues on this in github. Has anyone run in to that one before? Using 2015.5.3.
18:54 MindDrive joined #salt
18:54 dthom91 joined #salt
18:54 ajw0100 joined #salt
18:55 Akhter joined #salt
18:55 baweaver joined #salt
18:58 hasues left #salt
18:58 amit_ forrest: There isn't any conflict with the 'some stuff' part. Its completely separate directory structure doing something else altogether.
18:59 stilezayk left #salt
18:59 fredvd joined #salt
18:59 amit_ gerhardq1x: I'm looking at the conflict. trying to figure out why is it referring in that fashion. Its basically the same file.
19:03 jhauser joined #salt
19:05 BlackJackAc3 joined #salt
19:05 * amit_ slaps amit_ around a bit with a large fishbot
19:06 supersheep joined #salt
19:07 dthom91 joined #salt
19:08 amcorreia_ joined #salt
19:09 pullphinger15 joined #salt
19:10 ageorgop joined #salt
19:12 theologian joined #salt
19:13 zsoftich3 is there a way to set the env for cmd.run without setting it on each occurence?
19:15 zsoftich3 *set it globally
19:17 BlackJackAc3 joined #salt
19:18 pullphinger joined #salt
19:19 pullphinger joined #salt
19:19 timoguin zsoftich3: maybe try environ.set, which will set it for the salt process. i haven't used it though, so i'm not sure how persistence works
19:20 zsoftich3 trying environ.setenv...which is just returning False atm
19:21 zsoftich3 trying to set (http|https)_proxy for all these cmd.run wgets
19:21 PPpyg2 joined #salt
19:25 sgargan joined #salt
19:27 BlackJackAc3 joined #salt
19:30 PPpyg2 joined #salt
19:32 mattiasr joined #salt
19:34 gerhardq1x amit_: run 'salt-call -l debug state.highstate' to see where the second apache_conf.sls gets loaded
19:37 supersheep joined #salt
19:37 Akhter joined #salt
19:38 shantanoo joined #salt
19:38 shantanoo joined #salt
19:40 yomilk joined #salt
19:41 saphafee joined #salt
19:42 saphafee left #salt
19:43 RedundancyD joined #salt
19:43 DammitJim joined #salt
19:46 pm90_ joined #salt
19:47 mrwboilers can you use the value of a grain in the contents of a managed file?
19:48 robawt yep
19:48 robawt very similar to http://docs.saltstack.com/en/latest/topics/targeting/grains.html#writing-grains
19:48 robawt but in a jinja file of course
19:50 mrwboilers I'm just managing a config file for a service. If I could use the value of an existing grain in the conf file that would be cool
19:50 mrwboilers I don't need to write a new grain
19:50 mrwboilers I haven't been able to find any examples of this so far
19:52 forrest mrwboilers: {{ grains['host'] }}
19:52 mrwboilers forrest: thanks! I'll try that
19:52 forrest np
19:54 rdutch left #salt
19:54 mrwboilers forrest: is there some other wrapper or other convention needed for that? Something to tell it that it's jinja?
19:55 tanta joined #salt
19:55 sgargan joined #salt
19:55 tanta_g joined #salt
19:55 forrest mrwboilers: when you declare the managed file you need to use template: jinja. You can see what it looks like on the first example over here: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#operations-on-regular-files-special-files-directories-and-symlinks
19:57 mrwboilers forrest: thanks again! That's what I was looking for. My google skills had been failing me.
19:57 forrest np
20:02 iggy amit_: did you figure it out?
20:04 iggy zsoftich3: you could put it in a ~/.wgetrc file... except cmd.run's env is already jacked up and it doesn't know where $HOME is
20:05 Ryan_Lane can salt please use something other than elements for codenames? :(
20:05 zsoftich3 hehe.....sigh
20:05 Ryan_Lane alphabetical would be awesome
20:06 lexter joined #salt
20:06 iggy Ryan_Lane: We're almost past beryllium, it'll be okay
20:06 Ryan_Lane I'm not a chemistry person. I have no idea what order they come in :)
20:06 pullphinger joined #salt
20:07 Ryan_Lane the element names are often difficult to spell, too
20:07 iggy I don't think far enough ahead to worry about anything other than current + 1
20:07 Ryan_Lane I have no idea what current + 1 is right now :)
20:07 Ryan_Lane I'd need to find the current one, then look at a periodic table
20:07 iggy oh, I also have a periodic table on my desk
20:07 Ryan_Lane :D
20:08 iggy printed it out and wrote in the version numbers for each release
20:08 iggy if they rename 2015.8, I'm screwed... I already wrote it on there
20:09 dthom91 joined #salt
20:10 pullphinger joined #salt
20:10 DanyC joined #salt
20:12 zsoftich3 so...dealing with an http_proxy...i'm pretty much out of luck?
20:13 fersur3 joined #salt
20:14 fersur3 Is RAET usable on Salt 2015.5.x ?     Is RAET as usable on 2015.5.x as 2015.8.x ?
20:14 iggy I started looking at it in 2105.5, then quickly gave up
20:14 shantanoo joined #salt
20:14 shantanoo joined #salt
20:15 DanyC fersur3: i think RAET in 2015.5 it should be looked as "experimental" and in .8 i'm not sure
20:15 iggy I would still consider it experimental until more people actually start using it
20:16 dyasny joined #salt
20:16 iggy whether you can deal with experimental is personal preference
20:16 fersur3 I would like to start using RAET right now.    If I have any problems I can revert to ZMQ without issue, can I not?
20:16 iggy nope
20:16 s_kunk joined #salt
20:16 Ryan_Lane do you have to regen keys for raet?
20:16 iggy you have to rekey everything
20:16 fersur3 I just want to know if I should launch 2015.5.5 or 2015.8.0rc*
20:16 Ryan_Lane fersur3: 2015.8 is in freeze right now, so it should be relatively stable
20:17 iggy and (afaik), you still can't run a master with both raet and zmq
20:17 fersur3 iggy, could I just have symlinks   /etc/salt/pki-zmq   /etc/salt/pki-raet   and symlink between them when I do the switch or would more work be involved in re-keying the minions?
20:18 fersur3 iggy, but I can do separate masters, yes?    I don't have an issue using plain SSH to sed -i /etc/salt/minion and adjust the above-mentioned symlinks
20:18 iggy well, all the minions would require a reboot too
20:18 pm90__ joined #salt
20:18 fersur3 iggy, and also running svcadm restart salt-minion
20:19 iggy yeah, not reboot, restart of the process
20:19 iggy was typing faster than I was thinking
20:20 dthom91 joined #salt
20:20 iggy do you have a reason for wanting to run it? It doesn't currently offer any benefit (other than possible scaling >2000)
20:23 fersur3 I want to run RAET and if RAET is broken or behaving in strangely in any way I want to observe and push feedback upstream.
20:23 mattiasr joined #salt
20:23 giantlock joined #salt
20:25 sgargan joined #salt
20:26 XenophonF left #salt
20:28 iggy okay, well, to my knowledge, nobody here runs it, so talking directly to the salt devs might be your best option
20:29 iggy HINT: There aren't any of them here (regularly)
20:29 babilen Hmm, yeah, I'm sort of holding out on RAET as well ...
20:30 thayne joined #salt
20:31 Ryan_Lane basepi: ^^
20:31 Ryan_Lane he's usually here :)
20:31 Ryan_Lane saltstackinc should really have more of their devs hang out here.
20:31 basepi agreed.
20:31 basepi ;)
20:31 Ryan_Lane basepi: you're always stuck with my questions ;)
20:32 basepi There are a few, they're just not op'd. op-land is scary in a channel like #salt!
20:32 babilen Yeah, more people in the know to pester!
20:32 basepi I just need to op a bunch of you
20:32 Ryan_Lane hahaha
20:32 basepi Then I can hide in a sea of ops
20:32 iggy I try not to throw the few that are here under the bus
20:32 Ryan_Lane I throw them under the bus
20:32 Ryan_Lane saltstack needs to understand their funnel ;)
20:32 basepi Yeah, I'm happy enough to play middle-man.
20:32 ageorgop joined #salt
20:32 Ryan_Lane the top of the funnel is the community
20:32 iggy plus, I imagine cachedout is the best person for that (rather than basepi)
20:33 basepi Sorry, so what was I originally ping'd for?
20:33 babilen This channel is not really high maintenance though (as compared to #debian), no trolls, only friendly people and no frequent topic changes
20:33 Ryan_Lane basepi: RAET questions
20:33 * basepi hides
20:33 Ryan_Lane :D
20:33 Ryan_Lane was expecting that :)
20:33 basepi Ask away! No promises that I'll know the answer, though
20:34 Ryan_Lane fersur3: ^^
20:34 geekatcmu Hwew'a my RAET question: will it (finally!) fix the "mine doesn't work right with multi-master" problem?
20:34 babilen Are you aware of anybody using RAET in a larg(ish) production deployment?
20:34 basepi geekatcmu: unknown, but I doubt it. =\
20:34 iggy geekatcmu: no
20:35 geekatcmu 8(
20:35 iggy asked and answered in an issue somewhere
20:35 iggy at least not the first "version" of raet support
20:35 geekatcmu I've had to build out some really stupid/fugly things to work around that.
20:35 iggy (which is basically just mapping zmq functionality onto raet)
20:35 iggy geekatcmu: you mean like etcd?
20:36 * iggy runs
20:36 geekatcmu 8)
20:36 geekatcmu No, I'm not doing etcd or consul
20:36 Ryan_Lane etcd/consul/zookeeper are all better alternatives to mine
20:36 basepi babilen: No. UDP turns out to be a bit of a pain, since you have to have a port open on all machines. In the long run, we're actually pivoting towards the new raw TCP transport using tornado. It will be present in 2015.8.0, but very much beta.
20:36 Ryan_Lane (or dynamodb, if you're in AWS)
20:36 basepi It will keep the niceties of TCP, but allow us vision into what's going on in the transport layer (something ZMQ is missing)
20:37 Ryan_Lane any of those systems combined with sdb should work better
20:37 rsimpkins basepi: One of our devs is working on trying to expose what is happening in ZMQ. Not sure where that will lead yet.
20:37 Ryan_Lane (salt database == sdb - it's basically undocumented)
20:37 geekatcmu I've already *got* the configuration in an asset database.  I've written all the glue that lets the clients query the data.
20:37 geekatcmu But it's still fugly
20:37 zwi joined #salt
20:37 basepi rsimpkins: It's hard. It's largely a black box, particularly at load, which is where we really want to see what's going on.
20:38 babilen basepi: Does that mean that RAET is essentially a PoC that will be superseded by the raw TCP transport you've been hinting at above?
20:38 Ryan_Lane babilen: RAET is pluggable
20:38 geekatcmu And, really, I'm not sure how mine is inferior, but meh
20:39 Ryan_Lane geekatcmu: the mine data lives on the master
20:39 geekatcmu It pretty much fits exactly the use case.
20:39 babilen Ryan_Lane: AAH, right, sure, great!
20:39 basepi babilen: Seems that way to me, but stating that unequivocally is above my pay grade. =)
20:39 babilen sure, no worries
20:39 geekatcmu And, really, ZK would be very chicken-and-egg: to figure out which ZK environment you want to talk to, first you have to find out what ZK environment you are assigned to (grain) and *then* you have to find the list of servers for that environment.
20:40 babilen I'll hold off before deploying it then before I at least see a "should work" ;)
20:40 iggy geekatcmu: multi-master data stores already exist... salt re-inventing them is asinine
20:40 Ryan_Lane geekatcmu: well, that's the point of SDB :)
20:40 Ryan_Lane it's basically a better version of mine
20:40 geekatcmu If it's not documented, it doesn't exist.
20:40 geekatcmu Just sayin'
20:40 basepi babilen: Yeah, we've run into a few edge cases we want to resolve. Rather than blocking the 2015.8.0 release on those, we're releasing it as "beta"
20:41 iggy as murrdoc would say... code is documentation
20:41 Ryan_Lane I actually don't use a master because the availability story isn't there. with mine I think it's roughly impossible to have proper availability
20:41 geekatcmu And then I'd get to write all the code to dump our asset database into it.
20:42 shaggy_surfer joined #salt
20:43 Ryan_Lane basepi: what's the current status of multi-master with RAET?
20:43 Ryan_Lane what's really holding me back with masters right now is the ability to have multiple HA masters across AWS availability zones
20:43 Ryan_Lane ideally behind a load balancer
20:43 shaggy_surfer joined #salt
20:43 basepi I do not know, but I can find out if you want me to. I will say our RAET development has slowed basically to a halt while we work hard to get our latest enterprise stuff out.
20:44 Ryan_Lane gotcha
20:44 Ryan_Lane I'm not really in a rush, but I'd really like a master/minion architecture for remote execution with salt-api in front
20:45 * basepi nods
20:47 aqua^c joined #salt
20:51 dthom91 joined #salt
20:56 Ztyx joined #salt
20:56 baweaver joined #salt
20:58 ageorgop joined #salt
20:58 Ztyx joined #salt
21:02 mrjester left #salt
21:04 wendall911 joined #salt
21:05 coval3nce joined #salt
21:06 coval3nce Is there a fancy way to do a pillar.filter_by in a jinja map within the Pillar tree in order to filter by minion environment?
21:06 coval3nce E.g. some way to dynamically map settings so the pillar tree can look cleaner (for example including one collectd pillar.sls vs a base and an explicit environment one?)
21:07 dezertol joined #salt
21:07 iggy you can't use pillars in pillars
21:07 iggy pillarstack ?
21:08 coval3nce I have a map.j2 in pillars, and seems to work great with grains.filtery_by, but i wanted to filter by “environment"
21:08 coval3nce Which isn’t a grain as far as i can tell.
21:08 sgargan joined #salt
21:09 iggy {{ saltenv }}?
21:09 iggy or you mean something else?
21:09 coval3nce That would work, but I wanted to take advandage of the grains.filter_by formatting of dictionaies i can match, but in this case, the value would be the environment name
21:11 iggy well, you can just make the "map" a regular dict (with possible env's as keys) and just do a regular dict lookup
21:11 yomilk joined #salt
21:11 iggy or I'm just totally misunderstanding what you're talking about
21:12 baweaver joined #salt
21:13 coval3nce I guess ultimately what I was hoping to do is simplify my top.sls includes so each environment includes a specific pillar, such as “collectd”, and that pillar would abstract away how it maps different values for different environments
21:13 coval3nce Rather than doing thigns like “collectd.prod” and “collectd.test” in the top.sls itself
21:15 aw110f joined #salt
21:18 iggy we have a custom grain that reads the "ocsenv" from $cloud_provider metadata servers
21:18 iggy then we use that all over
21:19 iggy `- collectd.{{ grains.ocsenv }}`
21:19 iggy etc
21:19 coval3nce I’ve created a custom grain similar to that and been using it like that.  In this case i needed the Salt environment tag
21:20 coval3nce maybe i could use {{ opts[‘environment’] }}
21:20 coval3nce or like you said ealier saltenv
21:21 coval3nce I wanted to take advantage of merging though, but maybe i can still pull that off by doing an include like that in init.sls and just include the base set first before the environment ones
21:21 iggy ahh
21:23 iggy there's a setting to force merging in pillars
21:23 iggy you might be able to get that to work that way
21:23 coval3nce Yup, have that on.
21:23 coval3nce Let me re-organize here with this idea and see where I land.
21:25 stevednd is there a way to tell salt that it can never run a specific state file or highstate on a group of servers at the same time?
21:25 stevednd I'm fairly certain there isn't
21:26 stevednd I just was hoping for a safeguard to prevent some potentially catastrophic stuff
21:26 iggy not that I know of
21:26 iggy you could do locking with an sdb call maybe
21:26 stevednd I don't even eknow what the configuration for that would look like
21:26 jaybocc2 joined #salt
21:26 stevednd hmm
21:26 coval3nce iggy: sweet that idea works
21:26 GreatSnoopy joined #salt
21:27 baweaver joined #salt
21:28 stevednd iggy do you use formulas for your stuff or is most of your architecture custom written?
21:28 iggy depends
21:29 stevednd I just find that for everything I have to work together I have pretty much written everything myself
21:29 iggy I maintain 3 separate salt deployments right now
21:29 stevednd and I will not consider it an inconsequential amount
21:30 iggy 1 is really old, it uses all custom stuff... 1 is really small, it uses all custom stuff
21:30 stevednd err...would not will
21:30 rsimpkins left #salt
21:30 iggy 1 is pretty extensive and uses 14 formulas
21:31 yomilk joined #salt
21:32 stevednd I tend to take formulas as a base and then hack them up to work with my workflow
21:33 iggy most of the changes I've needed to get things working together we've sent back upstream
21:33 stevednd I have a lot that I do for determining different clusters of applications
21:33 stevednd load balanced things
21:33 iggy that's not true of everything (and probably shouldn't be as every setup is going to have some unique aspects about it)
21:34 stevednd a lot of modules for automatically determining types of services machines provide and such
21:35 baweaver joined #salt
21:35 karlthane joined #salt
21:35 stevednd a lot of what I'm doing I look at and think I'm probably drifting from the path, but at the same time I can't see how I would accomplish it in the normal framework of here's salt, and here's some formulas...magic
21:35 stevednd don't really have that many servers either
21:35 stevednd right about 45
21:36 iggy we have ~200 here
21:37 iggy but most everything is scalable to hundreds (or more) instances per cluster
21:37 stevednd how do you handle pillar data for a cluster for a given service such as rabbitmq, or mysql?
21:43 Ztyx left #salt
21:43 iggy see what coval3nce and I were talking about just above
21:44 freelock joined #salt
21:46 stevednd I do something like that
21:47 andrew_v joined #salt
21:47 stevednd basically it's main pillar for a service as the base, then merge in pillar data for the specific cluster being used, then merge in any machine specific pillar data
21:47 godlike joined #salt
21:52 bbbryson joined #salt
21:53 coval3nce Yeah that seems to have worked for me.
21:54 coval3nce I create a defaults.sls, then merge in depending on grains ina map.j2 file or includes
21:55 orion Does salt-cloud support EC2 user-data?
21:56 dthom91 joined #salt
21:57 iggy support it in what way?
21:58 orion I want to be able to associate user-data to a potential new node.
21:58 orion The reason I want to do this is because I need to set the hostname of the box (#cloud-config hostname: foo.whatever).
21:58 sgargan joined #salt
21:59 aron_kexp joined #salt
22:03 dezertol orion: when you use salt-cloud you define the salt grains per host. just set the value as a grain and then use that value in the state to set the host name
22:04 orion dezertol: Set the hostname from within a state you mean?
22:04 dezertol in your salt state you can set the hostname but pull the value for it.. from the grain that you assigned to the specific server
22:05 dezertol so your not storing variable data in your state
22:05 orion Ok, that sounds like a good plan, thank you.
22:05 dezertol np
22:07 sgargan joined #salt
22:11 baweaver joined #salt
22:18 sgargan joined #salt
22:19 _JZ_ joined #salt
22:19 Garo_ joined #salt
22:27 timoguin_ joined #salt
22:28 sgargan joined #salt
22:37 orion hmm
22:38 orion Despite having SourceDestCheck: False in my salt-cloud profile, new nodes are being created with source dest check set to True.
23:04 bfoxwell joined #salt
23:06 bfoxwell joined #salt
23:07 sgargan joined #salt
23:10 stevednd is there a module that wraps mkpasswd?
23:11 dezertol stevednd what are you trying to do?
23:11 dezertol there are for user/group auth stuff
23:11 stevednd dezertol: it's not for users or groups
23:11 stevednd passwords for haproxy
23:13 dezertol just out of curiosity .. but if your using mkpasswd to generate passwords to then apply during a salt-state... how will you know what they are set to?
23:14 stevednd the passwords that need to be used are gpg encrypted in the pillars
23:14 baweaver joined #salt
23:15 dezertol I'm obviously confused, I don't understand why you would use mkpasswd on the minion side
23:16 dezertol if your encrypting them a head of time as gpg strings in the pillar..
23:16 dezertol then using mkpasswd on the minion makes me confused
23:20 orion Ok, so it seems to me that many of the bugs I am running in to with 2015.5.3 are fixed in development. What's the recommended way to install the development version?
23:29 aqua^c joined #salt
23:35 ksoviero joined #salt
23:37 ksoviero Why is this SLS snippet causing an error? https://gist.github.com/ksoviero/c333d11e2cc80f12db45
23:37 ksoviero The error is No function declared in state 'file' in SLS 'alerts'
23:38 Bryson joined #salt
23:48 dthom91 joined #salt
23:50 dthom91 joined #salt
23:55 ksoviero Hello? Anyone?
23:57 whytewolf ksoviero: you are getting that error because you are not declaring a function. states are modules.function and you can't list 2 functions in the same module state
23:58 ksoviero How do I comment and append to the same file? I can't enter the same file twice because it complains about duplicate.
23:58 whytewolf different id, and use - name:
23:59 ksoviero ok, will do. thanks

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