Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-10-22

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

All times shown according to UTC.

Time Nick Message
00:05 shane__ joined #salt
00:06 badon_ joined #salt
00:12 snuffop joined #salt
00:16 MKFG joined #salt
00:17 hasues joined #salt
00:18 aparsons https://www.eventbrite.com/e/technical-deep-dive-linkedin-sf-tickets-13744882313  -- anyone in SF should come :)
00:19 mechanicalduck joined #salt
00:19 nitti joined #salt
00:20 mafrosis joined #salt
00:20 mafrosis lo can anyone help me find a 2014.1.10 deb package for wheezy please
00:20 mafrosis 2014.1.11 is broken and it’s the only one in apt
00:20 david__ joined #salt
00:21 mafrosis http://ftp.debian.org/debian/pool/main/s/salt/
00:21 mafrosis has a deb file which looks useable
00:23 skyler Can you use renderers in managed files?
00:24 Setsuna666 joined #salt
00:24 __number5__ skyler: yes
00:25 holler_ joined #salt
00:25 troyready joined #salt
00:25 jacksontj joined #salt
00:25 Fl0o joined #salt
00:25 APLU joined #salt
00:25 tomspur joined #salt
00:25 simonmcc joined #salt
00:25 andapathisch joined #salt
00:25 Twiglet joined #salt
00:25 funzo joined #salt
00:25 packeteer joined #salt
00:25 Nazzy joined #salt
00:25 forrest joined #salt
00:25 micko joined #salt
00:25 Jellyfrog joined #salt
00:25 ujjain joined #salt
00:25 oeuftete joined #salt
00:25 drybjed joined #salt
00:25 rubenb joined #salt
00:25 Ironhand joined #salt
00:25 aqua^lsn joined #salt
00:25 t0rrant joined #salt
00:25 A||SySt3msG0 joined #salt
00:25 babilen joined #salt
00:25 ze- joined #salt
00:25 mirko joined #salt
00:25 Heggan_ joined #salt
00:25 mackstick joined #salt
00:25 mattikus` joined #salt
00:25 __alex joined #salt
00:25 Damianz_ joined #salt
00:25 bdf joined #salt
00:25 __number5__ at least for py renderer
00:26 Outlander joined #salt
00:26 jeffspeff joined #salt
00:26 nickg joined #salt
00:26 tinuva joined #salt
00:26 NaPs joined #salt
00:26 harkx joined #salt
00:26 ]V[ joined #salt
00:26 CryptoMe1 joined #salt
00:26 rogst_ joined #salt
00:26 eofs joined #salt
00:26 heise joined #salt
00:26 jagardaniel joined #salt
00:26 darrend joined #salt
00:26 eightyeight joined #salt
00:26 rockey joined #salt
00:26 babilen joined #salt
00:26 babilen joined #salt
00:26 skyler __number5__: So awesome! I don't know if I will ever get to it, but I want to write a render that will turn my yaml pillars into managed .ini files.
00:27 skyler *renderer
00:28 arthabaska joined #salt
00:33 n8n joined #salt
00:34 Ned_ is it possible to allow some salt commands to be run on the master by non-root ?
00:40 n8n joined #salt
00:46 dude051 joined #salt
00:47 Ned_ iggy: any reason you store the bootstrap as a local file before running it ?
00:47 iggy Ned_: no, that's just what the docs say
00:47 druonysus joined #salt
00:48 iggy I generally do the "totally insecure one liner"
00:48 Ned_ right, whichis the one you pasted above ?
00:48 Ned_ curl -sL https://bootstrap.saltstack.com | BS_SALT_MASTER_ADDRESS=salt.example.com sudo -E bash
00:48 Ned_ I'm thinking that works nicely, and doesn't bother with a local file
00:48 Ned_ :D
00:50 jalaziz joined #salt
00:52 iggy so... question... does anyone else use roles/tags extensively?
00:53 iggy I'm thinking of extending the postgres formula to support master/slave setups using roles
00:53 iggy I just don't know if that's kind of the accepted way of managing something like that
00:53 mr_chris joined #salt
00:57 murrdoc joined #salt
00:58 littleidea joined #salt
00:59 Saltine joined #salt
01:01 Guest87468 left #salt
01:03 jonbrefe joined #salt
01:03 jonbrefe Hi. quick question. Anybody with experience with salt-syndic?
01:07 superted666 joined #salt
01:07 malinoff joined #salt
01:15 Mso150 joined #salt
01:16 shoemonkey joined #salt
01:21 GrueMaster joined #salt
01:35 snuffeluffegus joined #salt
01:40 bhosmer joined #salt
01:47 ioio joined #salt
01:48 Deevolution1 joined #salt
01:49 n8n joined #salt
01:50 mgw joined #salt
01:50 ioio joined #salt
01:51 bhosmer joined #salt
01:57 MugginsM joined #salt
01:58 jonbrefe joined #salt
02:06 littleidea joined #salt
02:10 anotherZero joined #salt
02:25 n8n joined #salt
02:28 bezeee joined #salt
02:40 smcquay joined #salt
02:50 bezeee joined #salt
02:56 superted666 joined #salt
02:57 tharkun joined #salt
03:01 schristensen joined #salt
03:02 mgw joined #salt
03:02 thayne joined #salt
03:15 jonbrefe joined #salt
03:19 kermit joined #salt
03:26 mgw __number5__: i'm not quite convinced that mine would have the same effect as running a function on a schedule and then returning the result as a master event (for use by the reactor).
03:26 TheThing joined #salt
03:28 mgw I'm picturing running, say, status.loadavg every 3 minutes, getting the result shipped back to the master, and then the master processing that result
03:30 schristensen joined #salt
03:37 pdayton joined #salt
03:37 wangofett joined #salt
03:40 bhosmer joined #salt
03:40 ramishra joined #salt
03:41 murrdoc joined #salt
03:49 davet1 joined #salt
03:56 murrdoc joined #salt
03:58 murrdoc1 joined #salt
04:03 yes456 joined #salt
04:05 fragamus joined #salt
04:05 n8n joined #salt
04:12 Ryan_Lane joined #salt
04:15 jalaziz joined #salt
04:23 snuffeluffegus joined #salt
04:25 Cidan joined #salt
04:31 murrdoc joined #salt
04:34 crocket Salting my system.!!
04:39 ndrei joined #salt
04:42 schimmy joined #salt
04:45 superted666 joined #salt
04:46 schimmy joined #salt
04:47 felskrone joined #salt
04:49 crocket Can I include a variable in file path?
04:49 crocket Probably possible with template?
04:49 ramteid joined #salt
04:54 shiva_blre joined #salt
04:57 crocket Is it possible to add third party modules and states?
04:59 __number5__ crocket: yes you can, like salt://sites/{{ domain }}.conf
05:02 crocket __number5__, What about third party modules and states?
05:02 __number5__ yes and you can create your own modules/states quite easily
05:02 crocket A long way to go.
05:02 crocket A long journey
05:03 mosen yeah check the salt-formulas repo
05:03 crocket so, the master might want to execute minions in the way.
05:06 n8n joined #salt
05:12 ramishra_ joined #salt
05:26 bezeee joined #salt
05:28 bhosmer joined #salt
05:36 bezeee joined #salt
05:39 bezeee joined #salt
05:50 linjan joined #salt
05:53 arno joined #salt
05:54 shookees joined #salt
05:58 jeffspeff joined #salt
05:58 oyvjel joined #salt
05:59 colttt joined #salt
05:59 ramishra joined #salt
06:03 zions joined #salt
06:11 jhauser joined #salt
06:30 desertigloo joined #salt
06:31 uber joined #salt
06:33 superted666 joined #salt
06:37 flyboy joined #salt
06:38 ioio_ joined #salt
06:42 crocket It's too much salt.
06:43 ramishra joined #salt
06:43 bahadir joined #salt
06:44 schimmy joined #salt
06:46 CeBe joined #salt
06:46 Furao joined #salt
06:46 davidone joined #salt
06:46 TheThing|24-7 joined #salt
06:47 Mso150 joined #salt
06:48 shiva_blre am trying to tell salt to install from a local yum repo but not able to do so
06:49 schimmy joined #salt
06:49 shiva_blre am getting :  Failed to configure repo 'OpenIPMI': The repo does not exist and needs to be created, but a name was not given
06:49 shiva_blre when I issue salt 'hostname_minion' state.highstate
06:49 shiva_blre repo is good on salt master
06:49 shiva_blre checked with yum repolist
06:50 shiva_blre Can some one help ?
06:52 Mso150_z joined #salt
06:52 mosen not sure, missing a parameter maybe.. no name given?
06:52 mosen maybe the humanname: parameter
06:54 superted666 joined #salt
06:57 shiva_blre humanname ?
06:59 shiva_blre I gave and saw. I didn't work.
07:00 shiva_blre anybody there
07:04 superted666 I'm here, just logged on. What's up?
07:04 Heartsbane joined #salt
07:04 jeffspeff joined #salt
07:05 shiva_blre am trying to tell salt to install from a local yum repo but not able to do so
07:05 shiva_blre salt 'hostname_minion' state.highstate
07:06 shiva_blre m getting :  Failed to configure repo 'OpenIPMI': The repo does not exist and needs to be created, but a name was not given
07:06 superted666 Okay, so yum works from repos stored in /etc/yum.repos.d/
07:06 superted666 Is there a repo in there with that name?
07:06 shiva_blre yes its there
07:07 superted666 If you check the file, is it set to 'enabled=1'
07:07 shiva_blre This is my fractal.sls file
07:07 shiva_blre enabled ,yes
07:08 crocket Is ansible easier than salt?
07:08 shiva_blre fractalpkgs:   pkgrepo.managed:     - humanname: fractallocal      - name: OpenIPMI      - fromrepo: http://xxx.xxx.xxx.xxx/fractallocal centos     - dist: centos     - require_in:       - pkg: OpenIPMI          OpenIPMI:   pkg.installed:     -fromrepo: http://xxx.xxx.X.XXX/fractallocal     -skip_verify: True     -refresh: True
07:09 shiva_blre @superted666
07:09 superted666 What's the command your using to install salt, i presume it's 'yum install salt'?
07:09 shiva_blre Han ?
07:09 shiva_blre yes
07:12 crocket Does salt-bootstrap also install salt-master?
07:12 superted666 I've only used salt from a C&C perspective so am unfamiliar with the config management part of it, that said...does '- fromrepo' support a URL? Usually or at least in the command modules for yum repo's the 'fromrepo' specifies the repo name that's configured in /etc/yum.repos.d/
07:13 cmthornton sorry for butting in here, but those `fromrepo:` entries don't seem right to me, they could be because I'm not really a RHEL guy, but shouldn't the `fromrepo: <name>` match the name that's in brackets in /etc/yum.repos.d/fractallocal.repo? not the url/ip
07:13 superted666 ^ My thoughts too
07:13 slav0nic joined #salt
07:13 shiva_blre okay I changed fromrepo:
07:13 shiva_blre and am trying
07:14 shiva_blre The state "OpenIPMI" in sls fractalpkgs is not formed as a list
07:14 shiva_blre it says
07:14 shiva_blre not formed as a list
07:15 cmthornton probably yaml syntax error :\
07:16 superted666 Paste it in pastebin or similar, it's a bit hard to read on IRC
07:17 bhosmer joined #salt
07:20 arno joined #salt
07:22 iwishiwerearobot joined #salt
07:23 crocket Is it better to use salt-bootstrap than to install distribution salt packages?
07:25 honza joined #salt
07:25 honza joehh: thanks, will check it out
07:27 superted666 crocket: Depends on what OS and how you typically deploy packages i guess
07:27 jeffspeff joined #salt
07:33 duncanmv joined #salt
07:34 dsolsona joined #salt
07:34 crocket How do I assign specific grains to certain minions from the master?
07:34 evidence looks like there was a regression between 2014.7.0rc1 -> 2014.7.0rc2 in relation to syndics, that is still present in rc5
07:35 evidence [salt.loaded.int.returner.local_cache       ][WARNING ] Could not write job invocation cache file: [Errno 2] No such file or directory: '/var/cache/salt/master/jobs/65/f54d2a7406408b84c4b33bb14d099b/.load.p'
07:35 evidence ROR   ] An inconsistency occurred, a job was received with a job id that is not present in the local cache: 20141022002557701422
07:35 evidence looks like a bunch of others are reporting it also
07:35 crocket grains in _grains directory don't seem to care about minion specificity.
07:36 jdmf joined #salt
07:36 crocket Are custom grains distributed to all minions?
07:36 chiui joined #salt
07:37 badon joined #salt
07:42 felskrone joined #salt
07:42 felskrone1 joined #salt
07:44 ramishra joined #salt
07:44 felskrone joined #salt
07:44 felskrone1 joined #salt
07:46 fe92 joined #salt
07:47 dsolsona left #salt
07:47 felskrone joined #salt
07:51 krak3n` joined #salt
07:52 cmthornton crocket: I'm fairly sure that grains are distributed to all minions, including custom grains. If you want data sent only to a specific minion, you'd need to use pillar for that. Pillar data is only distributed to minions that you match on in your top.sls
07:56 badon joined #salt
07:57 duncanmv whiteinge: ping
07:57 martoss joined #salt
07:57 martoss left #salt
08:01 lcavassa joined #salt
08:07 cberndt joined #salt
08:23 ramishra joined #salt
08:25 gyre007_ joined #salt
08:30 shane__ joined #salt
08:32 martoss1 joined #salt
08:41 felskrone joined #salt
08:43 piquadrat joined #salt
08:47 felskrone1 joined #salt
08:48 intellix joined #salt
08:50 AviMarcus joined #salt
09:04 giannello joined #salt
09:04 chiui any news about 2014.1.12?
09:05 ggoZ joined #salt
09:06 chiui or at least, is there now a mirror of the previous working release available?
09:06 bhosmer joined #salt
09:06 superted666 If you use RH then latest stable should be available in Epel
09:07 giannello chiui, problem with salt-cloud?
09:08 chiui actually with salt in local mode
09:09 chiui and I'm using ubuntu (not my choice :P )
09:09 giannello well, I just stumbled across a bug in salt-cloud 2014.1.11 - fixed in .12, but debian package is not available yet
09:10 cjbarnes18 joined #salt
09:10 chiui would be nice that at least if QA process is not perfect, previous release are kept for a longer time
09:10 chiui *releases
09:13 arif-ali joined #salt
09:13 giantlock joined #salt
09:16 vincehu joined #salt
09:18 gyre007_ joined #salt
09:23 cberndt joined #salt
09:25 alex-mesos joined #salt
09:25 ingwaem`` joined #salt
09:25 ingwaem`` left #salt
09:31 alex-mesos hi
09:32 alex-mesos I bumped into an issue while setting up an AWS cluster and would appreciate any help
09:32 alex-mesos the problem is that i cannot access a grain I set before
09:33 giannello how did you set the grain and how are you trying to access it?
09:34 alex-mesos That's how I set the grain:
09:34 alex-mesos set-grain-public-hostname:
09:34 alex-mesos cmd.run:
09:34 alex-mesos - name: salt-call grains.setval public-hostname $(curl -s http://169.254.169.254/latest/meta-data/public-hostname)
09:34 alex-mesos And the corresponding part of the log: http://pastebin.com/n9aXcKCW
09:35 cofeineSunshine ghm
09:35 cofeineSunshine you can set grains in profile
09:35 cofeineSunshine ant when creating server from taht profile, graint are applied
09:36 alex-mesos I can set this grain only the AWS node is provisioned
09:36 alex-mesos And here how I use the grain:
09:36 alex-mesos /etc/mesos-slave/hostname:
09:36 alex-mesos file.managed:
09:36 alex-mesos - source: salt://cfgs/hostname
09:36 alex-mesos - template: jinja
09:36 alex-mesos - makedirs: True
09:36 giannello alex-mesos, use pastebin pls
09:37 alex-mesos ok, sorry
09:39 shiva_blre joined #salt
09:40 snuffeluffegus joined #salt
09:40 alex-mesos giannello: I put parts of salt formulas and relevant logs in http://pastebin.com/VVzK3nbn
09:40 dsolsona joined #salt
09:40 snuffeluffegus joined #salt
09:41 shiva_blre am able to configure my local repo using salt
09:42 giannello alex-mesos, yeah - that's not the right way to get grains
09:42 shiva_blre but am not able to install the packages from that repo
09:42 shiva_blre here is my fr.sls file contents : http://pastebin.com/5AgPgU4t
09:42 alex-mesos giannello: glad it's my mistake : )
09:42 giannello use salt['grains.get']('public-hostname', 'default')
09:43 giannello where "default" is a default value used if the public-hostname grain does not exist
09:43 shiva_blre am trying to install packages thorough salt
09:44 shiva_blre am point the salt to look into my local repo
09:44 alex-mesos giannello: thanks a lot, let me try
09:44 shiva_blre its working
09:44 ramishra joined #salt
09:44 shiva_blre but not able to install a package from the local repo
09:44 shiva_blre can anyone help
09:45 martoss joined #salt
09:46 alex-mesos giannello: if this is not the right way to use grains, why does {% elif grains['virtual'] == 'xen' %} work?
09:48 CycloHex joined #salt
09:48 dsolsona left #salt
09:48 giannello alex-mesos, no idea
09:49 shiva_blre http://pastebin.com/5AgPgU4t
09:49 shiva_blre my fr.sls file contents
09:49 giannello alex-mesos, if you want to double-check that the grain is there, you can run "salt 'host' grains.items"
09:50 giannello or from that node, salt-call grains.items
09:50 shiva_blre am able to configure the repo using salt but not able to install the packages
09:51 alex-mesos giannello, after the salt job is complete, the grain is there. It is either missing or not accessible during the job
09:51 giannello oh
09:52 CycloHex shiva_blre: on which Distro are you?
09:52 shiva_blre hello, some one there ?
09:52 giannello well, than it's a different problem
09:52 slav0nic joined #salt
09:52 shiva_blre am on Centos
09:52 slav0nic joined #salt
09:52 shiva_blre Centos 6.4 - my master
09:52 giannello when you use grains.setval, the minion is instructed to reload the grains list
09:52 shiva_blre centos 6.5 - minion
09:53 CycloHex shiva_blre: read the note in the docs.. You have to specify precise or somehting
09:53 giannello but I think this will not affect an already-running minion thread
09:53 shiva_blre precise is for ubuntu
09:53 CycloHex http://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkg.html#salt.states.pkg.installed
09:53 shiva_blre or deb
09:53 shiva_blre distros
09:53 giannello and you are also setting the grain using cmd.run
09:53 CycloHex ah ye.. sorry
09:54 shiva_blre have u seen the paste bin link?
09:54 giannello alex-mesos, try replacing that cmd.run with "grains.present"
09:54 giannello http://docs.saltstack.com/en/latest/ref/states/all/salt.states.grains.html
09:54 shiva_blre only first 4 lines are working fine
09:54 shiva_blre @cyclohex
09:56 CycloHex I don't havfe experience with cusotm repo, but in your pkg.installed you say fromrepo: farctallocal, while the name of pkgrepo is fractalpkgs and human name is Fractal Packages
09:56 shiva_blre can some one help me ?
09:56 CycloHex shouldn't oyu use one of both values for your fromrepo: ?
09:57 alex-mesos giannello, I'm not sure i can use the result of a command (curl ...) in "grains.present"
09:57 shiva_blre http://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkgrepo.html
09:57 giannello indeed
09:57 tinuva joined #salt
09:58 alex-mesos I tried this: https://gist.github.com/anonymous/655e3ae96ba3be680fd7 in order to facilitate grain reload
09:58 superted666 shiva_blre: Shouldn't the fromrepo point to 'fractalpkgs'?
09:58 alex-mesos this doesn't help, however
09:58 superted666 shiva_blre:  Don't see where your presenting the name 'fractallocal' when defining the repo
09:58 shiva_blre checking
09:59 alex-mesos giannello, your point about running minion thread makes sense to me
09:59 CeBe joined #salt
10:00 bhosmer joined #salt
10:01 shiva_blre The following package(s) were not found, and no possible matches were found in the package db:
10:01 shiva_blre still am getting package not found @superted666
10:01 giannello alex-mesos, sorry, that's all I can tell you
10:02 superted666 shiva_blre: Log onto the box manually, use 'yum install smartmontools --fromrepo=fractalpkgs' does it work or complain about a missing package?
10:02 alex-mesos giannello, thanks a lot, I'll try to figure it out
10:03 CycloHex also shiva_blre you should use pkg.latest, not installed imo
10:04 shiva_blre what is pkg.latest ?
10:04 shiva_blre and not installed imo ?
10:04 CycloHex http://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkgrepo.html
10:04 CycloHex thi suses pkg.latest: not pkg.installed
10:04 CycloHex imo = in my opinion
10:07 dburbridge joined #salt
10:08 shiva_blre cyclohex: ok
10:08 shiva_blre checking
10:14 cberndt joined #salt
10:15 shiva_blre cyclohex: yum install smartmontools --fromrepo=fractalpkgs = is working fine
10:15 shiva_blre I Logged onto the box manually
10:16 linjan joined #salt
10:19 peters-tx joined #salt
10:22 mrh-6363 joined #salt
10:22 superted666 shiva_blre: Not too sure then, i'm not familair with states but that looks okay if you've fixed the repo name
10:23 mrh-6363 Hi there, can someone please tell me how to get by salt-cloud a list of GCE instances deployed by some *instancename* mask?
10:23 shiva_blre even after fixing my repo name its not able to install the pacakage sfro the repo
10:25 CeBe joined #salt
10:27 cberndt joined #salt
10:27 alex-mes_ joined #salt
10:34 shiva_blre http://pastebin.com/wQ9ddQ7p
10:34 shiva_blre still not working
10:35 superted666 I suspect the issue may be the space in the repo name
10:35 superted666 Don't think yum supports that
10:35 superted666 If humanname is populating the [] brackets at the top of the file in /etc/yum.repos.d/ that is
10:38 superted666 The fromrepo needs to match the repo name not the filename, whatever is put in that file at the top in the square brackets
10:38 shiva_blre ok checking @superted666
10:41 linjan joined #salt
10:42 shiva_blre http://pastebin.com/wQ9ddQ7p
10:42 chiui anybody has 2014.1.10 saltstack for precise?
10:42 shiva_blre not working
10:42 shiva_blre I issued salt '*' state.highstate -l debug
10:43 shiva_blre but it says did not return in time
10:44 ramishra joined #salt
10:47 alex-mesos joined #salt
10:54 superted666 shiva_blre: can you paste the output of the file in /etc/yum.repos.d onto pastebin too that salt creates please?
10:54 shiva_blre ya
10:55 bhosmer_ joined #salt
10:55 VSpike Has anyone ever considered making the Windows minion use IronPython instead of CPython?
10:56 martoss joined #salt
10:56 shiva_blre http://pastebin.com/wQ9ddQ7p
10:56 shiva_blre @superted666
10:56 Furao joined #salt
10:57 superted666 Hmm, the name there still has a space in it, i wouldn't expect that after you updated it
10:57 shiva_blre oh
10:57 superted666 Oh wait, i'm being stupid
10:58 shiva_blre Hmm
10:58 diegows joined #salt
10:58 superted666 Ignore that, the name is indeed correct for the repo
10:59 shiva_blre ok then whats the problem ?
11:00 superted666 I suspect it's something simple but best if anyone else with more knowledge on states. The name would of been an issue but not sure why that's not working now
11:03 shiva_blre Oh anyone with high knowledge in states ?
11:03 shiva_blre @superted666 thank you
11:03 superted666 shiva_blre: No probs, sorry couldn't get to the bottom of it
11:04 jdmf joined #salt
11:04 shiva_blre its ok.
11:07 VSpike Can someone remind me, is there a module that lets you change the value of an envvar within the current salt minion process?
11:07 VSpike Ah, salt.modules.environ
11:10 viq shiva_blre: ask, don't ask to ask.
11:10 shiva_blre @superted666 : I succeeded in installing the package from yum repo
11:11 superted666 shiva_blre: What was it in the end?
11:12 shiva_blre I just modified the 'from repo ' parameter to the value of 'name' inside the /etc/yum.repos.d/fr.repo
11:12 shiva_blre @viq: am not getting what you are saying ?
11:12 superted666 Ah, glad it's sorted :)
11:13 shiva_blre Thanks a lot superted. Thanks a lot.
11:14 viq shiva_blre: though I could be at fault, not having read enough far back. Just saw your " < shiva_blre> Oh anyone with high knowledge in states ?"
11:17 shiva_blre Am very sorry.
11:18 viq Did I miss your question, or you didn't ask it yet? ;)
11:19 cjbarnes18 joined #salt
11:20 flyboy82 running 2014.1.11 on Debian 7 (both master and minion). With custom init script that works when I run by hand (both with service blah start and with /etc/init.d/blah start) the service.running state reports that it's in the desired state, when it's actually not running. In the minion logs there's the "[INFO    ] output: blah is not running"
11:23 viq flyboy82: try from somewhere 'ssh root@minion "service blah status"' and see what happens
11:24 viq flyboy82: I found some scripts use "something &" in them, which works fine from terminal, but not from ssh or tools like salt
11:24 tinuva joined #salt
11:25 agend joined #salt
11:28 flyboy82 it works from ssh, still not from salt
11:30 viq flyboy82: the script does have status/check command?
11:30 flyboy82 yup
11:30 viq flyboy82: also you could check with salt-call -l debug to see what exactly salt does, maybe that will give you a hint
11:30 flyboy82 I opened up debug log level on the minion and have all the responses
11:30 flyboy82 nothing weird so far
11:31 viq Look how salt calls the script, what it calls to check if it's running
11:31 viq At least that's the path I'd start on
11:32 flyboy82 it calls "service blah status" and it gets response "blah is not running"
11:33 viq and when you run 'service blah status' you get a proper response?
11:33 flyboy82 yup
11:33 viq $PATH ?
11:33 flyboy82 aaaah, found it...
11:34 viq oh?
11:34 flyboy82 the dumb ***** who wrote the init script forgot to exit 1; when the service is not running so it returned true and just echoed the "not running" message
11:34 bhosmer joined #salt
11:34 flyboy82 sorry for bothering you mate
11:34 viq hah
11:40 scottpgallagher joined #salt
11:41 glyf joined #salt
11:43 CycloHex Can you send custom things to salt.mine? I'd like to ssh-keygen -lf myKey to find my fingerprint, then retrieve that fingerprint from the mine and use is in a state
11:44 flyboy82 CycloHex, whatever custom grain you have on your minion, you can list it in mine_functions, under the grains.item dictionary
11:45 flyboy82 and the minion will send it to the mine
11:45 CycloHex So I have to store the fingerprint in a grain?
11:45 ramishra joined #salt
11:46 viq_ joined #salt
11:46 CycloHex I'm not sure if fingerprint is needed in the state ssh_know_hosts.present, I'll try without it first ^^
11:46 intellix joined #salt
11:46 flyboy82 u have to understand that the mine is a data store with data retrieved from all the minions... so unless something is sent by the minion to the mine, I doubt it's going to find its way in there by itself
11:55 glyf joined #salt
11:55 dude051 joined #salt
11:56 chamunks joined #salt
12:10 scottpgallagher joined #salt
12:18 N-Mi joined #salt
12:20 CycloHex Anyone here experienced with ssh_know_hosts state? Is there a logfile I can check to see what's going on? The only error I get is: Unable to receive remote host key
12:20 flyboy82 in your minion    /var/log/salt/minion
12:20 CycloHex yes
12:21 CycloHex just found it, it hit me when I typed my question :D
12:21 flyboy82 and maybe try putting log level to debug from /etc/salt/minion
12:21 CycloHex yes, I have that on my newly deployed minions, but eg my backup-server still hasn't received the new config, since I want to keep it on 2014.1.10, so better not replace too many things there :d
12:22 scottpga_ joined #salt
12:23 wnkz joined #salt
12:35 akafred joined #salt
12:37 crocket How do I install salt manually?
12:39 CycloHex manually?
12:40 miqui joined #salt
12:42 ThomasJ Depends on how manual you want to get :P
12:42 ThomasJ http://docs.saltstack.com/en/latest/topics/tutorials/salt_bootstrap.html is a good place to start
12:43 ThomasJ If you look at the script in turn you can perform all steps manually if desired
12:43 bhosmer joined #salt
12:43 CycloHex flyboy82: Sorry to bother you, but you seem to know quite a lot about salt-mine. In my reactor I need the ip of the minion with his id stored in data['id'], the only way I know to get data form the mine is in a for-loop, but I know that i'll always only need 1 result, so looping is not needed. So my question is, If I know I only need one record form the mine, how can I use that in my reactor? {% *** salt['mine.get']('id:{{ data['id'] }}', network.ip_addrs,
12:43 CycloHex expr_form='grain' %} I guess it's something like this, but not sure how to get the ip out of that :s
12:46 ramishra joined #salt
12:46 crocket How do I set up salt from git?
12:46 crocket "pip install -e ./salt.git" doesn't work.
12:47 CycloHex like ThomasJ said, have a look at the boostrap docs
12:47 vejdmn joined #salt
12:47 CycloHex ah, the entire source.. Why not just use the bootstrap? It gets it from git anyway
12:47 crocket Is it impossible to set up salt with python 2.7?
12:48 CycloHex _SALTSTACK_REPO_URL="git://github.com/saltstack/salt.git"
12:49 lcavassa joined #salt
12:49 oz_akan joined #salt
12:50 ekristen joined #salt
12:51 gmcwhistler joined #salt
12:52 ThomasJ crocket: Salt depends on python >= 2.6 <3.0
12:52 ThomasJ see http://docs.saltstack.com/en/latest/topics/installation/ for full dependency list
12:53 gmcwhistler joined #salt
12:54 duncanmv whiteinge: ping?
12:55 diegows joined #salt
12:56 ajolo_ joined #salt
12:56 glyf joined #salt
13:02 bhosmer joined #salt
13:09 ekristen morning everyone
13:09 CycloHex day ekristen
13:11 ekristen hi CycloHex
13:11 CycloHex Hello ekristen
13:12 mpanetta joined #salt
13:14 duncanmv https://hackweek.suse.com/projects/514 <- Salt
13:15 nitti joined #salt
13:15 Ahlee heh. It's telling me to turn my sreen into landscape position
13:15 Ahlee apparently it's very opinionated on resolution
13:16 duncanmv ahhh, half of the participants complained about the same... yeah, the website is very silly and animated.. (hackweek)
13:16 CycloHex rightclick and inspect element
13:16 duncanmv do you have a resized browser window or a vertical monitor?
13:16 Ahlee yeah, that's pretty much the antithesis of hwat I want to see in a website.
13:17 Ahlee neat dmacivar
13:17 duncanmv Ahlee: it has an animated asteroid that will shake your screen. It is for fun
13:17 Ahlee i'm too old and grumpy for fun
13:17 CycloHex ah, it complains if your browser size has a width smaller than the height
13:18 Ahlee wait. isn't spacewalk a RedHat product?
13:18 duncanmv yeah, lot of people complained, but we could not figure why they just did not resize the browser window
13:18 Ahlee ah, looks like novell started supportng it
13:18 duncanmv Ahlee: SUSE Manager 's upstream is Spacewalk. We are probably the largest contributors outside RH.
13:18 Ahlee duncanmv: 'cause people like me don't like websites telling them how they should position their windows ;)
13:19 CycloHex Ahlee: lol :D
13:19 duncanmv Ahlee: and SUSE is no longer Novell (they are just sibling companies now)
13:19 Ahlee didn't know that, either
13:20 racooper joined #salt
13:21 viq duncanmv: also, the site is blank without javascript
13:21 vejdmn joined #salt
13:22 duncanmv yes, my emacs screen also goes blank if I turn off lisp
13:22 Furao joined #salt
13:22 * duncanmv is trying to get the uploading of the package profile correctly
13:23 viq nevermind, not in the mood for this argument
13:25 Ahlee that is pretty awesome tho, setting salt under space walk
13:25 micah_chatt joined #salt
13:25 Ahlee at least, that's my understanding of the project
13:25 duncanmv does Salt 2014.1 have some kind of scheduler for the commands?
13:25 viq It indeed would be pretty interesting, have spacewalk use salt as the underlying technology
13:25 duncanmv (Spacewalk can schedule actions with time, we will need to have the equivalent of that)
13:26 Ahlee I think granular scheduling is going in on 2014.7
13:26 viq duncanmv: http://docs.saltstack.com/en/latest/topics/jobs/schedule.html
13:26 dude051 joined #salt
13:26 duncanmv viq: it would be awesome, but it would be a lot of effort. Spacewalk is a mountain of tens of years of code.
13:27 * viq nods
13:27 viq I need to look how far along foreman is
13:27 CycloHex a mountain, walking in space
13:27 viq CycloHex: "that's no moon"
13:27 iwishiwerearobot joined #salt
13:27 Ahlee i wish i'd have found foreman before sinking hundreds of man hours into my own bascially-the-same-thing
13:28 duncanmv that looks like scheduling a recurring command, is it the same to say, "install this packages not before next week"?
13:28 duncanmv Spacewalk uses cobbler for now. The new Redhat stack moved to Foreman
13:28 duncanmv does Salt work well with Foreman?
13:28 viq duncanmv: foreman uses cobbler ;)
13:29 duncanmv viq: I don't think so
13:29 duncanmv cobbler implementes a subset of cobbler, but I remember Foreman was pure ruby
13:29 viq For provisionin? AFAIK it does, it did when I looked at it ~1.5y ago
13:30 viq duncanmv: as for scheduling, not as such, but you could combine it with something like http://docs.saltstack.com/en/latest/ref/states/all/salt.states.at.html
13:30 whiteinge duncanmv: pong
13:31 whiteinge (Reading the scroll-back...)
13:31 Badger_t joined #salt
13:32 Badger_t left #salt
13:34 CycloHex Anyone here that use the reactor-tag salt/cloud/*/destroyed? I see the event get fired, and that it finds the reactor, but it's unable to render it. When I change the tag to minion_start and restart my minion, the same file gets found, this time properly rendered :SSSS
13:34 CycloHex mor einfo: https://groups.google.com/forum/#!topic/salt-users/ARHisSiMgx4
13:35 bezeee joined #salt
13:35 pdayton joined #salt
13:37 cpowell joined #salt
13:47 ramishra joined #salt
13:49 BinaryKhaos joined #salt
13:57 thedodd joined #salt
13:59 perfectsine joined #salt
13:59 oz_akan mpanetta:good morning
13:59 mpanetta morning oz_akan
14:02 VSpike goddammit windows
14:03 ramishra joined #salt
14:06 perfectsine joined #salt
14:09 housl joined #salt
14:10 smcquay joined #salt
14:14 perfectsine joined #salt
14:14 kt76 joined #salt
14:15 nitti joined #salt
14:17 mohae joined #salt
14:19 tk75 joined #salt
14:24 djstorm joined #salt
14:32 StDiluted joined #salt
14:36 perfectsine joined #salt
14:37 BinaryKhaos Hello @all. I am new to Salt and have been reading up on it quite a bit. One thing remains a bit of a mystery for me though: How do you handle updates to config files and package updates in general? I am on Gentoo, so that is even more of an issue. But in general: After updates, you often have to merge in changes to config files and/or do some other related work. How does that fit in with Salt?
14:38 oz_akan BinaryKhaos:do you mean system packages? \
14:39 jalbretsen joined #salt
14:39 oz_akan Let's say I would use salt to manage apache, which means I would use salt to manage configuration files for apache
14:39 BinaryKhaos oz_akan: For example, yes (consider a rebuild of packages due to an ABI change or whatever). But also other service oriented packages as well that have changed some config options.
14:39 iggy BinaryKhaos: we don't (generally)... we constantly re-deploy everything, so any updates come in at that time
14:39 oz_akan If I upgrade Apache I know what files is going to change, so salt would overwrite / correct these files after an upgrade.
14:40 oz_akan what ever you manage via salt, you would like no one to touch
14:40 oz_akan if they touch it, salt can overwrite these changes but then this may break the app that changed the config file
14:40 iggy the only package updates we've done since I've been here have been for the bash bug
14:41 oz_akan if you are able to create servers easily, it would be better to build the server each time
14:43 BinaryKhaos That sounds a bit like Salt is only for worker nodes... you know... servers that are there to scale but that is it. If you have two or three bigger servers, that you cannot easily redeploy, you are out of luck.
14:45 BinaryKhaos Or you have to have a test server where you pretty much do each and every update, check if something changed to the config file structure or whatever and update your Salt "states" appropriately.
14:46 Ahlee calling pillar of pillar isn't supported, right?  So I can't do pillar={foo:{'bar':'baz', 'qux':'abc' {{ salt['pillar.get']('foo:bar')}} }}
14:47 felskrone joined #salt
14:47 to_json joined #salt
14:47 rypeck joined #salt
14:48 aw110f joined #salt
14:49 kaptk2 joined #salt
14:50 Ahlee bar in this point is an app name, and qux would the the path, so I want to basically append the value of bar onto the end of qux
14:50 Ahlee i suppose I can just {% set name %} and then use it both places
14:52 oz_akan is there anyone here from saltstack?
14:52 oz_akan I have a few non technical questions
14:53 CycloHex ahlee, can you gist the current statefile?
14:53 CycloHex I have used regexp or jinja vars to achieve what I think you're trying to achieve
14:53 CycloHex Ahlee*
14:54 iggy BinaryKhaos: that's just how we use it... but no, salt doesn't have magic smoke that makes it automatically handle config file upgrades
14:56 BinaryKhaos iggy: I know that... and I wasn't expecting it. :) Actually, I had any magic in any program... but that's just the software engineer speaking. :) Well... I will simply mail the list and see how others are dealing with that. Maybe I am also missing a point somewhere.
14:56 iggy we have a few states written that are going to break horribly when we go to upgrade the software that they manage
14:56 BinaryKhaos iggy: s/I had any magic/I hate any magic/
14:56 Ahlee CycloHex: This state file calls a custom _state, I'm trying to reuse parts since I pass in the values in a pillar on the CLI
14:57 BinaryKhaos iggy: But there has to be a way to deal with that use-case as well... properly.
14:57 CycloHex can you upload some code to gist.github or pastebin or w/e, so I can get an image of what you're exactly trying to do
14:59 viq BinaryKhaos: well, how would you "normally" deal with updating config files?
14:59 Ahlee CycloHex: sanitizing, one moment
15:01 BinaryKhaos viq: If I knew major updates are around the corner, I schedule an appropriate downtime, do some upfront testing and research (off the machine naturally) and once the upgrade is done, I merge the changes to the config files (etc-update or alike) and adapt everything that needs adapting.
15:01 CycloHex Is it possible to change the body of a smtp_return? Or what alternative is there to send a mail once a highstate completed?
15:01 BinaryKhaos viq: And sometimes you just need to do a rebuild for ABI changes.
15:01 viq BinaryKhaos: you do exactly the same with salt - you just need to do it centrally, just like with any other CM tool
15:02 BinaryKhaos viq: So what about running the "etc-update" tool for example that manages the config updates? I run it on all servers through salt? Since that tool is interactive.
15:03 dude^2 joined #salt
15:03 viq If it's interactive, then tough, no go. You'd need to run it on test server, adapt the configuration to be pushed out, and make sure that updated configuration depends on updated packages.
15:04 pfallenop joined #salt
15:04 viq None of those tools save you from having to know your infrastructure. They just make it quicker and more repeatable.
15:05 viq Which also may mean you can more quickly, thoroughly and at greater scale shoot yourself in the foot :P
15:05 to_json joined #salt
15:06 crocket I added an environment "slackware" to the master conf and the top state file as in http://pastebin.com/y3fYr4HK
15:06 wendall911 joined #salt
15:06 workingcats BinaryKhaos, you probably want to directly manage the config files rather than using etc-update
15:06 crocket Now, "salt '*' state.sls vim" fails with an error "Data failed to compile: No matching sls found for 'vim' in env 'base'"
15:06 crocket Why?
15:06 workingcats on a related note, i find dispatch-conf much more pleasant. can't remember why though
15:06 viq crocket: you need to specify in top how you're matching nodes
15:06 crocket viq, I did
15:06 viq crocket: not in what you pasted
15:07 crocket viq, You saw top.sls
15:07 viq crocket: yes, and it's malformed
15:07 crocket viq, how?
15:07 viq sec, link incoming
15:07 crocket How can I check its syntax?
15:07 workingcats BinaryKhaos, fwiw my servers will be in salt *a lot* quicker than the clients, simply because the servers are more important ;)
15:07 crocket viq, salt doesn't generate accurate syntax errors.
15:08 viq crocket: http://docs.saltstack.com/en/latest/topics/targeting/grains.html#matching-grains-in-the-top-file and http://docs.saltstack.com/en/latest/topics/targeting/compound.html#targeting-compound
15:08 viq you're missing - match: compound and - match: grain
15:09 crocket viq, That's confusing.
15:10 viq crocket: how so?
15:10 crocket viq, Because I'm a beginner and I am sleepy.
15:10 BinaryKhaos viq/workingcats: Thanks. I see. So you basically have to do everything on a test server before hand, adapt your configs and take it from there. dispatch-conf would need to run nevertheless (and yes, it is superior :P) for all updates to config files that are not directly managed by Salt.
15:10 conan_the_destro joined #salt
15:10 Ahlee you have to pass the environment if it's not base, crocket
15:10 crocket viq : I may be able to understand it at once after drinking coffee in the morning.
15:10 Ahlee if the state doesn't exist in base, you need to tell the salt client where to find it
15:11 Ahlee salt '*' state.sls vim env=slackware
15:11 crocket Ahlee, I matched slackware environment with a matcher.
15:11 BinaryKhaos viq/workingcats: If you guys don't mind, I will nevertheless mail the list as well and see what hints I can get there as well.
15:11 viq crocket: you need to tell the matcher what it should match on
15:11 BinaryKhaos viq/workingcats: Doesn't mean I don't believe you or anything alike. Just the bee-hive-mind is sometimes very useful and getting different perspectives helps understand a concept better.
15:12 Ahlee too many cooks
15:12 Ahlee bowing out of the kitchen
15:12 viq BinaryKhaos: MOAR IDEAS BETTERER
15:12 crocket "salt '*' state.sls vim" tells me "No matching sls found for 'vim' in env 'base'"
15:13 Ahlee I liked, is vim/init.sls under /srv/salt/base or /srv/salt/slackware ?
15:13 crocket Ah shit
15:13 Ahlee er, lied
15:13 crocket I forgot to restart salt-master.
15:13 crocket I think salt-master hasn't reloaded /etc/salt/master
15:14 viq crocket: you need http://pastebin.com/XQARpf6j
15:15 viq And yes, you need to restart salt-master if you change config
15:16 workingcats BinaryKhaos, no ofc go ahead
15:16 berserk joined #salt
15:17 BinaryKhaos viq/workingcats: Already on it. Both your input has been /very/ helpful and is very much appreciated, by the way.
15:17 crocket viq, After I restart salt-master, should I restart salt-minion as well?
15:17 viq crocket: nah
15:17 crocket It seems connection is lost after salt-master restart.
15:17 workingcats yw
15:18 viq crocket: unless you changed minion config
15:18 StDiluted joined #salt
15:18 viq BinaryKhaos: hive mind is useful, that's why I hang around here as well. Sometimes I even get answers ;)
15:18 Ahlee it might take a few seconds but minions reconnect
15:18 Ahlee usually
15:19 crocket ok
15:19 crocket It works now
15:19 crocket It's weird
15:20 knot joined #salt
15:22 pfallenop joined #salt
15:22 Ahlee CycloHex: I think I figured out a way of doing this: https://gist.github.com/jalons/63466ce310dc3627743b
15:23 Ahlee then, if the value is null. hrm.
15:23 Ahlee gah.
15:24 BinaryKhaos viq: I absolutely agree... talking to others is one of the best things one can do to expand ones own horizon. :)
15:24 crocket viq, Can a grain matcher be automatically enabled in top sls?
15:24 viq crocket: AFAIK, no
15:25 viq crocket: by default it matches on ID, you want anything else you have to say what
15:25 crocket viq, Search "os:Ubuntu" on http://docs.saltstack.com/en/latest/ref/states/top.html
15:25 Ozack1 joined #salt
15:26 crocket Some minion targets don't have matcher configured.
15:26 CycloHex Ahlee: I'm not sure if it's possible to print a variable in a print. but wouldn't - destination: {{ salt['pillar.get']('deployment:destination') }} {{salt['pillar.get']('deployment:application') }} not work?
15:27 CycloHex without the space between }} and  çç
15:27 CycloHex {{*
15:27 viq crocket: I think it may be shortened or a mistake
15:27 mgw joined #salt
15:27 Ahlee CycloHex: that's what I'm trying now.
15:27 SheetiS joined #salt
15:27 Ahlee CycloHex: i think i'm going to have to just do a bunch of fancy {% set %}ing, as all these are overridable but need defaults
15:28 crocket Anyway, works now.
15:29 CycloHex Yes, I have some files with a lot of {%set%}ings as well.. :D
15:29 ekristen anyone seen “TypeError: 'NoneType' object is not iterable” when running highstate? The error is at “File "/usr/lib/pymodules/python2.7/salt/state.py", line 1869, in merge_tops"
15:29 CycloHex it might be some work, but makes your code readable thus better maintainable
15:30 CycloHex ekristen: I think that's a syntaxerror in your top file
15:30 tinuva joined #salt
15:31 vejdmn joined #salt
15:32 SheetiS joined #salt
15:32 crocket If a minion is turned off, could the master access minion grain cache?
15:34 hasues joined #salt
15:35 mgw I've got some ideas for using salt + an alerting service (such as pagerduty)  to replace nagios, but I also have some questions -- anybody up for discussing?
15:38 viq mgw: sensu replacement, eh?
15:40 BinaryKhaos viq/workincats: Ok... gotta go. Thanks again... and have a nice evening/day... bye.
15:40 pdayton joined #salt
15:40 jaimed joined #salt
15:40 mgw viq: I've not used sensu, but it seems it replicates a lot of the stuff salt already provides
15:41 BinaryKhaos left #salt
15:41 viq I haven't used it yet either. Kinda, yeah.
15:41 crocket Oh there is cache.
15:41 crocket salt-run cache.grains minion-target
15:41 crocket Can I wake a minion up?
15:43 ajolo joined #salt
15:44 beneggett joined #salt
15:44 crocket well
15:44 crocket left #salt
15:45 mgw I've also looked at salmon
15:45 tligda joined #salt
15:45 viq mgw: yeah, shame it's been abandoned
15:46 pfallenop joined #salt
15:46 pfallenop joined #salt
15:46 mgw salmon uses its own service that the individual minions connect to (via a returner)
15:46 mgw i was thinking of returning check results to the master via the event bus
15:46 mgw so that the master can use runner(s) to do further checks
15:46 mgw But this might present a scalability issue
15:47 mgw depending on how many checks and minions there are vs the master's resources
15:47 viq mhmm
15:47 viq Also salt doesn't have guaranteed delivery (yet?) as opposed to AMQP
15:50 ramishra joined #salt
15:53 beneggett joined #salt
15:53 N-Mi joined #salt
15:53 N-Mi joined #salt
15:54 ndrei joined #salt
15:57 pressureman joined #salt
16:02 ekristen anyone here have experience with multi master setups?
16:05 druonysus joined #salt
16:08 TheThing joined #salt
16:09 pdayton1 joined #salt
16:09 ozzzo joined #salt
16:11 bezeee joined #salt
16:12 anyburl How to reset a Master's Python version? Installed Py 2.7 with Pip-installed Jinja2 2.7.3 using pyenv on RHEL6, made Py2.7 global env, but 'salt --versions'  still showing Py2.6.6 and Jinja2 2.2.1   !! Can I somehow reset salt to see my pyenv Global Py2.7 env?
16:13 giantlock joined #salt
16:14 Ahlee Start salt from the virtualenvironment's interpretor
16:14 pfallenop joined #salt
16:14 MTecknology viq: Would you consider getting a gitlab-ci formula up and running for say $20?
16:14 MTecknology I know it's a small amount, but it's out of my pocket, and I'm broke.
16:15 Ahlee ekristen: make sure both masters have identical disk layouts and all values (states, pillars, etc) are in sync
16:15 anyburl ah, Ahlee, thanks,  of course I got over-excited and forgot to RESTART the master ... my bad...that'll probably work !!  Thanks so much.
16:15 Ahlee ekristen: We dropped multimaster and accpet the possibility that the master going down results in salt being unavailable
16:17 KyleG joined #salt
16:17 KyleG joined #salt
16:19 murrdoc joined #salt
16:19 pfallenop joined #salt
16:21 Deevolution1 Ahlee: Can you clarify why identical disk layouts are important?  Are you refering to salt file locations, or partitions, etc...?  Multi-master is something we're looking at.
16:21 CeBe joined #salt
16:23 pdayton joined #salt
16:23 troyready joined #salt
16:23 toastedpenguin joined #salt
16:25 QiQe joined #salt
16:26 QiQe I have changed yesterday salt-master from roots (fileserver backend) to a mix of gitfs, s3fs and roots
16:27 QiQe everything went ok, but after some time master stop to pull the git and s3 bucket
16:28 QiQe I need to restart the master to resolve it and again after some time happens again
16:28 QiQe any idea?
16:28 jhulten joined #salt
16:28 QiQe fileserver.update is done by defaul (60 seconds)
16:31 rypeck Anyone have any idea when SaltStack v2014.7 will come out of release candidate status? I don't see a good way of monitoring progress on issues for the release. Was expecting a GitHub Milestone or issues label but don't see one.
16:31 anyburl ahlee: I'm now starting master service only after guaranteeing PATH and pyenv shell are set correctly to my Py2.7 environment, yet 'salt --versions' still shows wrong/older/Salt-install-time Py2.6.6 version; any other things I should check?
16:31 beneggett joined #salt
16:33 aparsons joined #salt
16:33 ekristen_ joined #salt
16:34 Guest84076 joined #salt
16:35 Ahlee Deevolution1: I mean making sure that any pillar_root and file_roots are identical - i.e. if you have a file_root for environment foo: on one master pointing to /srv/salt/, you need to make sure that foo fileroot exists on both masters and is identical
16:35 Ahlee Deevolution1: in our case, pillars were our bane as we'd push testing pillars to one master, they wouldn't be present on the second and "Strange Things Happened"
16:35 Deevolution1 Ahlee: Ah.  Got it.
16:35 Mso150_z joined #salt
16:36 ajolo joined #salt
16:36 anyburl ahlee: I may have found my prob, the top of  /usr/bin/salt-master shebangs /usr/bin/python2, which indeed of course still points to my Py2.6.6 install....   Thank you so much...I should be able to work around this now
16:36 thedodd joined #salt
16:37 Ahlee anyburl: yeah, was just typing that out to look for that since you were calling salt directly
16:37 Ahlee anyburl: good deal on finding it.
16:37 anyburl thanks again!!
16:38 vejdmn joined #salt
16:39 arthabaska joined #salt
16:40 to_json joined #salt
16:41 ramishra joined #salt
16:42 quantumsummers joined #salt
16:42 turnes joined #salt
16:44 turnes Hi all. I'm facing a issue when I use salt state network in Ubuntu 14.04.1 LTS.
16:44 turnes May somebody help me ?
16:46 ozzzo turnes: maybe. what is the error message?
16:47 chiui joined #salt
16:47 ponpanderer joined #salt
16:48 turnes That's my sls file.
16:48 turnes eth0:   network.managed:     - enabled: True     - type: eth     - proto: none     - ipaddr: 10.1.0.1     - netmask: 255.255.255.0     - dns:       - 8.8.8.8       - 8.8.4.4
16:48 n8n joined #salt
16:49 turnes When I run "sudo salt '*' state.highstate". The following error appears:           ID: eth0     Function: network.managed       Result: False      Comment: Interface file malformed: ['iface', 'eth0', 'inet'].      Changes:
16:50 jchen turnes: can you use a pastebin to preserve formatting?
16:51 vejdmn joined #salt
16:52 turnes Sure. My sls file http://pastebin.com/sGY15XDC
16:53 jchen turnes: can you pastebin the network configuration file on the server?
16:55 BrendanGilmore joined #salt
16:56 turnes When I run "salt '*' state.highstate"  http://pastebin.com/8f3CmNEF
16:57 turnes Sure. Here is jchen http://pastebin.com/gBbk79VG
16:58 jchen turnes: does your server work as is with that interfaces file? looks malformed
16:59 turnes Acctually... never tried. I've just setup the server.  It's default config.
17:00 turnes The server has a second network card and I've setup manually by ifconfig.
17:00 ekristen I’m getting an ascii unicodedecodeerror https://github.com/saltstack/salt/blob/develop/salt/output/highstate.py#L203
17:03 mpanetta Is it possible to tell an orchestration runner step to run only on a single minion out of a group?
17:04 mpanetta It can be any random minion.
17:04 vejdmn joined #salt
17:04 ekristen basepi: you around?
17:04 bbnix joined #salt
17:04 thayne joined #salt
17:05 basepi ekristen: not really. I'm training today, on a break at the moment.
17:05 ekristen just upgraded to rc5, 90% of my servers no problem but not I’m getting a unicodecodeerror during highstate
17:05 ekristen have you seen that before?
17:05 ekristen not == now*
17:05 basepi ekristen: we've seen unicode errors. usually a rogue weird character or nonbreaking space or something in a state
17:06 basepi I'm surprised you're seeing it after the upgrade, I can't imagine anything that changed.
17:06 zooz joined #salt
17:06 ekristen any way to find the rogue character or the state that has it?
17:06 ekristen without going through each one, one by one
17:07 UtahDave joined #salt
17:07 ekristen UtahDave: welcome back
17:07 pdayton joined #salt
17:08 StDiluted joined #salt
17:08 beneggett joined #salt
17:09 jxqz s/basepi/base-e/
17:09 jxqz :-)
17:09 aparsons joined #salt
17:11 XenophonF joined #salt
17:12 XenophonF salt states and private github repos - bad or good idea?
17:12 XenophonF i'm inclined to do it only because i don't have time to set up my own private kallithea instance
17:13 aparsons joined #salt
17:14 basepi ekristen: Not off the top of my head. Usually there's enough context information in the error to point you in the right direction
17:15 Ryan_Lane joined #salt
17:16 jonbrefe joined #salt
17:17 murrdoc joined #salt
17:17 cpowell joined #salt
17:21 spookah joined #salt
17:27 Gareth morning morning
17:27 robawt yo Gareth
17:27 iggy XenophonF: we do it
17:27 Gareth robawt: morning :) hows it going?
17:27 iggy once we got the certs nailed down initially, it's worked like a charm
17:27 robawt Gareth: can't complain, count down to scale!
17:29 Gareth robawt: :)
17:31 UtahDave thanks, ekristen!
17:32 ekristen UtahDave: I’m happy to report that most of my upgrades have gone pretty smooth to RC5
17:32 ekristen however ran into an issue on one of my production servers that I didn’t run into on any other server yet
17:32 ekristen trying to figure it out
17:33 UtahDave interesting. is there anything especially different about that server?
17:33 Cottser joined #salt
17:34 yes456 joined #salt
17:35 ekristen UtahDave: so had a ’ instead of a ‘ in a file
17:35 schimmy joined #salt
17:36 aparsons_ joined #salt
17:36 bhosmer_ joined #salt
17:38 Mso150_z joined #salt
17:39 UtahDave subtle things like that are fun!
17:39 UtahDave :)
17:39 Gareth subtle == scream, throw things, then rock back and forth...crying.
17:40 murrdoc :D
17:40 schimmy joined #salt
17:40 aparsons joined #salt
17:41 Gareth UtahDave: I was looking at your slack module yesterday, was thinking about writing a slack returner....unless you already had one in the works.
17:42 ramishra joined #salt
17:42 UtahDave Gareth: Dude, I've wanted to build out a state for it and create a returner like you did for hipchat, but there's no way I'm going to have time for at least several weeks.
17:43 UtahDave If you're interested in doing it, please feel free!
17:43 Gareth okay cool :) Will take a look, wanted to check wiht you first...didn't want to step on toes.
17:43 UtahDave I didn't really even have time to write the slack module, but I was traveling for work and couldn't sleep one night.  So I did it for fun
17:44 Gareth :)
17:44 UtahDave In fact, techhat suggested I rip out using the pyslack python module and just use the rest interface
17:44 UtahDave like I think anita kreuger did for the hipchat module
17:45 Gareth using the requests module.
17:46 UtahDave yep
17:46 wnkz joined #salt
17:47 MTecknology I have a cron task that runs nightly on my salt server. It's a salt command. One of the states that executes is a cmd.run that gets tucked into the log output when the return is True because it's seen as a change. This is causing me a bit of headache. How can I omit these from the output?
17:49 vejdmn joined #salt
17:49 ghanima joined #salt
17:50 MTecknology hopefully that made sense...
17:50 ghanima Is anyone familiar with salt-contrib on github
17:54 ramishra joined #salt
17:57 Mso150_z_z joined #salt
17:58 iggy I've seen it before
17:58 iggy does that count?
17:59 intellix joined #salt
17:59 perfectsine joined #salt
18:02 snuffeluffegus joined #salt
18:04 ghanima iggy: possibly
18:05 Mso150 joined #salt
18:05 ghanima just want  to confirm that if I wanted to use any of those files I would place them on my salt master in the file root
18:05 ghanima using _ prefix right?
18:05 iggy depends what it is
18:06 dsolsona joined #salt
18:06 iggy grains, modules, maybe runners, returners?
18:08 jalaziz joined #salt
18:09 mohae joined #salt
18:11 eunuchsocket joined #salt
18:12 UtahDave ghanima: yeah, that's correct.  modules in   /srv/salt/_modules    states in  /srv/salt/_states   ,etc
18:14 arthabaska Does anyone know what the most complete git library for use with Helium is? I can't seem to get pygit2 working well with a private gitfs..any other viable options now?
18:15 akafred joined #salt
18:17 n8n joined #salt
18:20 UtahDave arthabaska: I was just recommended to use pygit2
18:20 UtahDave but make sure you're using a fairly new version
18:21 XenophonF left #salt
18:21 arthabaska @UtahDave thanks..I must be a problem on my end then--can't get it to auth correctly. I'll get deeper with it and figure it out
18:21 johtso joined #salt
18:22 cberndt joined #salt
18:22 johtso Is there an equivalent of a base setting for ext_pillar ?
18:23 murrdoc joined #salt
18:24 cberndt joined #salt
18:24 wnkz joined #salt
18:24 pdayton1 joined #salt
18:28 thedodd joined #salt
18:30 __TheDodd__ joined #salt
18:31 beneggett joined #salt
18:31 higgs001 joined #salt
18:33 steveoliver joined #salt
18:34 cberndt joined #salt
18:35 beneggett joined #salt
18:40 pdayton joined #salt
18:40 ramishra joined #salt
18:41 MTecknology I HATE RHEL!!!!!!!!!
18:41 murrdoc happy caps lock day
18:41 MTecknology I used shift
18:41 MTecknology It causes me so many headaches. Out of the ten boxes not working with salt, one is not RHEL.
18:42 murrdoc no it actually is caps lock day (fyi)
18:42 MTecknology hehe.. LOL!
18:42 jergerber joined #salt
18:42 MTecknology I can run state.highstate on the minion and I can run salt-minion -l debug on the minion and all seems well, but salt 'minion' test.ping seems to go into /dev/null.
18:43 MTecknology http://capslockday.com/
18:43 murrdoc on all minions ?
18:43 MTecknology on my RHEL minions, all others seem fine
18:44 murrdoc have u nuked salt cache on the minion ?
18:44 MTecknology nope, not yet
18:44 murrdoc try it on one ?
18:44 MTecknology this? rm -rf /var/cache/salt/minion/
18:45 murrdoc service salt-minion stop;rm -rf /var/cache/salt/minion/;service salt-minion start;
18:45 MTecknology just wanted to make sure I wasn't blowing away too much
18:46 MTecknology no change, still hangs
18:46 Vye MTecknology: are you doing -l debug on the master or the minion?
18:46 MTecknology on the minion
18:46 murrdoc oh
18:47 MTecknology the master has ~450 boxes connected to it
18:47 ozzzo joined #salt
18:47 pdayton joined #salt
18:47 beneggett joined #salt
18:47 MTecknology 406 boxes connected
18:48 Vye MTecknology: does 'salt-call test.ping' work on the minion?
18:48 MTecknology yessir, it does
18:49 MTecknology the minion has no firewall configured on it
18:49 Vye MTecknology: I've never used salt-minion to execute modules on the minion. I've always used salt-call. I think salt-minion is only for managing the daemon.
18:50 MTecknology You can use 'salt-minion -l debug' to see what's going on. It's what's telling me that the job isn't even making it to the minion.
18:53 MTecknology The weirdest part is that it seems to be specific to RHEL boxes, not networking.
18:56 Vye MTecknology: when you run salt-minion -l debug is the salt-minion service still running?
18:56 hasues When I issue a salt '*' saltutil.refresh_pillar command, and a minion reports back "None", does that mean it did not take any changes?
18:57 MTecknology Vye: nope, that would cause a lot of screwy issues
18:57 MTecknology would look like this, though..
18:57 Vye MTecknology: ok, I understand what you're doing now.
18:58 murrdoc what about salt -G 'os:Redhate' test.ping -l debug
18:59 MTecknology murrdoc: exact same results, minion didn't return
18:59 arthabaska joined #salt
18:59 QiQe hi guys, my master is taking a top.sls file from a tag on git different from master tag
18:59 MTecknology I like the typo, it's accurate.
19:00 QiQe so, it is failing trying to find a state that doesn exist
19:00 beneggett joined #salt
19:01 QiQe how can I force to master only retreive master branch/tag and ignore the others ones
19:01 pdayton joined #salt
19:02 hasues http://docs.saltstack.com/en/latest/topics/tutorials/pillar.html <- I can't get the data.sls, info example to work.
19:03 hasues oh, nevermind, it works.
19:03 jdmf joined #salt
19:04 smcquay joined #salt
19:05 MTecknology they're both on hybrogen
19:05 Mso150_c joined #salt
19:06 iggy QiQe: top files just work that way... that's why it's suggested to only have one top file for all your repo's/branches/etc
19:08 cberndt joined #salt
19:09 claytron_ joined #salt
19:13 n8n joined #salt
19:14 X86BSD joined #salt
19:18 pdayton joined #salt
19:20 ckao joined #salt
19:21 QiQe thanks iggy
19:21 QiQe I will do that
19:22 iggy QiQe: one trick some people seem to use (especially if they are using the gitfs backend) is have a repo with _only_ a top file in it
19:22 iggy then you have formulas, states, etc. in other repos
19:23 cjbarnes18 joined #salt
19:25 QiQe but my team will need eventually to update 2 different repos
19:25 QiQe that doesnt sound so good
19:25 iggy lol, I guess not
19:25 iggy our gitfs list is like half a page long already
19:25 iggy so it's not a big deal for us
19:26 QiQe wow
19:26 murrdoc man you can run masterless fully from git
19:26 murrdoc like so much fun
19:26 jalaziz joined #salt
19:26 iggy mehsterless...
19:26 murrdoc ermahgerd
19:26 forze joined #salt
19:27 invsblduck joined #salt
19:27 invsblduck ok!  top.sls is supposed to go in file_roots or pillar_roots ?!
19:28 murrdoc both
19:28 iggy BOTH!
19:28 invsblduck GAH!
19:28 invsblduck >_<
19:28 murrdoc i mean
19:28 bezeee joined #salt
19:28 murrdoc for different purposes
19:28 murrdoc not the same top.sls in both places
19:28 invsblduck gripes.
19:29 iggy I suspect most people's state and pillar trees are quite different (ours are wildly different)
19:29 murrdoc si
19:30 invsblduck searching docs (how i spend 80% of my time in salt...trying to understand docs).
19:30 invsblduck the documentation seems really good.
19:30 invsblduck but then i have to keep reading. over and over.  and searching.
19:30 invsblduck indicator. haha
19:31 invsblduck ty.
19:31 iggy the docs are full of info, but somewhat hard to navigate
19:31 iggy everybody knows it
19:31 murrdoc the way the docs are built is dope
19:31 iggy I suspect something will be done with them after the release actually makes it out
19:32 iggy at the very least, they are moving off readthedocs
19:32 imanc joined #salt
19:33 murrdoc readthedocs is all 301'ed
19:37 jalaziz joined #salt
19:41 ramishra joined #salt
19:42 invsblduck still searching frantically, but the reason i ran into this is because i ran salt-call state.highstate for the first time, which didn't work.  whereas `salt-call state.sls foo' has been working all along.
19:42 gamingrobot joined #salt
19:42 iggy searching for what?
19:43 invsblduck why there needs to be two top files (one for states and one for pillar) and what they are supposed to look like.  http://docs.saltstack.com/en/latest/ref/states/top.html reads that top.sls should go in the env dir (eg., /srv/salt/base), but another doc reads that it should go in the top-level file_roots dir (eg., /srv/salt).
19:44 invsblduck i've only ever had one in pillar_roots so far!
19:44 invsblduck so if i were to guess...
19:44 invsblduck the top files just map nodes to highstate data, regardless of whether /srv/salt or /srv/pillar structures..
19:48 iggy they map nodes to files that can contain states or pillar
19:48 iggy they are formatted the same way
19:48 jxqz basepi: yo!
19:49 iggy but (f.ex.), we only have a nginx state that's mapped to things with the rest role, but we have nginx.dev, nginx.qa, and nginx.prod pillars
19:50 invsblduck iggy: ok. thanks
19:51 invsblduck for someone who's been working on unix every day for 15 years and has extensive experience with both puppet and chef, i'm pretty confused right now.
19:51 invsblduck could just be the hangover
19:52 invsblduck i spend most of my time trying to understand salt and not actually doing much with it.
19:53 iggy yeah, it was easier for me to read some of the intro stuff (about a days worth) and then just hit the ground running
19:53 felskrone invsblduck: its hard to grasp at first because there is quite a lot of stuff to understand :-)
19:53 invsblduck it took several hours to extend the community apache formula very minimally (good learning experience though) before realizing it wasn't too decent anyway, and that i should make my own.
19:54 invsblduck word
19:55 invsblduck thanks for the pointers.
19:55 invsblduck :)
19:56 murrdoc joined #salt
19:57 turnes Is the Salt state network working for Ubuntu 14.0.4 ? Or there is some bug ? The documentation says "Support for Debian/Ubuntu is new in 2014.1.0 and should be considered experimental"
19:58 tempspace Has anybody else seen grains.append being broken in 2014.1.13?
19:58 pdayton joined #salt
20:00 druonysuse joined #salt
20:02 micah_chatt joined #salt
20:02 Ryan_Lane1 joined #salt
20:05 dalibro joined #salt
20:05 dalibro 123 ! :) anyone?
20:06 dalibro I can't find how to include salt modules
20:06 dalibro tried to put them into /srv/salt/_modules
20:06 dalibro but somehow they don't appear
20:06 iggy dalibro: saltutil.sync_all
20:07 iggy or highstate, but we don't use that often
20:07 MTecknology I still can't figure out why I can't send jobs to RHEL minions. The versions are the same, salt-call state.highstate works, salt-minion -l debug indicates the connection is being established, but the jobs never reach the minion from the master.
20:07 MTecknology :'(
20:08 dalibro iggy: http://pastebin.com/Z2ytN4Wz
20:08 dalibro this is the output
20:08 iggy you should see your module in modules
20:08 dalibro they shouldn't be empty right?
20:09 dalibro -rw-r--r--. 1 root root 4.9K Oct 20 23:10 /srv/salt/_modules/selinux.py
20:09 dalibro what about the path ?
20:09 iggy looks fine to me
20:10 dalibro do i need to include them in master conf ?
20:10 iggy not that I know of
20:10 iggy but /srv/salt is in your "base" env in the master conf right?
20:11 dalibro yes
20:11 iggy and roots is in fileserver_backend?
20:11 pdayton joined #salt
20:11 dalibro no
20:12 dalibro just a sec
20:12 beneggett joined #salt
20:14 QiQe those -> gitfs_env_whitelist and gitfs_env_blacklist are not available on salt 2014.1.10, right?
20:15 beneggett joined #salt
20:15 iggy QiQe: correct
20:16 QiQe progress is always late...
20:16 QiQe :P
20:17 oz_akan joined #salt
20:17 aparsons joined #salt
20:19 iggy 2 WEEKS!
20:19 druonysus joined #salt
20:19 druonysus joined #salt
20:25 davet joined #salt
20:26 invsblduck iggy: read the detailed comments in /etc/salt/master and got my two top files configured...looking better.
20:28 ndrei joined #salt
20:33 bbnix joined #salt
20:38 giantlock joined #salt
20:39 arthabaska joined #salt
20:40 Deevolution1 Is it normal for Windows nodes to be (very) slow in responding to commands (test.ping, state.highstate, etc...)?
20:40 shookees joined #salt
20:42 ramishra joined #salt
20:43 jhauser joined #salt
20:43 Ryan_Lane joined #salt
20:43 iggy not even going to swing at that one
20:49 jalaziz joined #salt
20:49 duncanmv joined #salt
20:52 MTecknology I think rhel just needs to die for ever and ever. It's been nothing but a pain for me. Salt helps, but only if salt works on those boxes.
20:52 SheetiS MTecknology: which rhel version(s)? (All of them) :D
20:53 MTecknology SheetiS: in this particular case - CentOS release 5.11 (Final)
20:54 * iggy not surprised
20:54 MTecknology Every single box that's rhel based is having the issue I mentioned a short ways above.
20:55 MTecknology thankfully I'm down to 9 of them. I'm almost exclusively Debian 7
20:55 mpanetta Anyone here know if you can make the orchestration runner select a random minion to run on when executing a state?
20:55 iggy random?
20:55 mpanetta Well, it does not have to be random really, I just want it to select one.
20:56 mpanetta The use case is, I have a cassandra cluster set up that I need to add a user to.  If I add the user on every minion it will error all but the first time (since the user already exists)
20:56 SheetiS mpanetta: the orchestration itself has a - tgt: <target> that you can target just one of them with the state.
20:56 Gareth turnes: It works fairly well and there have been quite a few fixes to it that will be in the 2014.7 release.  It was considered experimental in 2014.1 because it was completely new.
20:57 SheetiS for that particular state that you are adding the user
20:57 mpanetta SheetiS: Yeah but at that point I have to know the name of the minion.  Would rather not have to :(
20:58 SheetiS mpanetta: how are you defining the cluster together?  They have to know eachother of course.
20:58 mpanetta They do follow a naming scheme, so I guess I could just have it always run on the first minion...
20:58 MrThrax joined #salt
20:58 mpanetta SheetiS: Yeah we use grains and mine so that they can all find eachother.
20:59 SheetiS That's what I would do unless there is a good reason not to (run on the first in the naming scheme or whatever)
20:59 iggy our clusters have a "primary" role that we use to target things like that
20:59 SheetiS iggy++
20:59 mpanetta Yeah, cassandra does not have the concept of 'primary'
21:00 mpanetta So running on the 1st sounds like the best option.
21:00 MTecknology SheetiS: any thoughts to help me figure out what's broken here? It's entirely possible I just did something dumb..
21:00 iggy it's a role we assign (via grains, via gce tags)
21:00 mpanetta We used the primary thing with our mongodb formula, because it was obvious who would be primary (via grains)
21:01 SheetiS MTecknology: I was scrolling up to look at what you checked
21:01 SheetiS How's your SELinux configured?
21:01 aparsons joined #salt
21:01 n8n joined #salt
21:02 MTecknology SheetiS: it's disabled
21:02 WesleyL left #salt
21:02 turnes Gareth: Thanks for the information :) I'm using 2014.1.11. I've just found out my mistake. The file /etc/network/interfaces was bad formatted. ;) Thanks for helping.
21:03 SheetiS MTecknology: hmm and you say salt-call state.highstate works but salt <target> state.highstate just hangs?
21:03 MTecknology correct
21:03 MTecknology SheetiS: if I run salt-minion -l debug on the minion, I 'should' see the job id, but that never comes through.
21:03 MTecknology The minion isn't getting the job at all
21:03 fragamus_ joined #salt
21:04 Gareth turnes: no worries.
21:05 beneggett joined #salt
21:05 MTecknology turnes: I did that this morning. Turns out, broadcast can not be put in place of gateway.
21:05 SheetiS and you said all tables on iptables are flushed and empty (ACCEPT across the board)...  Hmm
21:05 iggy check the issue tracker
21:05 invsblduck iggy: so how do i use a pillar variable from one pillar file inside another pillar file?  i have a pillar file that defines my domain name, which i map to every node in top.sls.  In another pillar file I want to set a variable that has {{ salt['pillar.get']('my_domain_var') }} in the value.  is that possible?
21:05 SheetiS could it be a zeromq thing?
21:06 invsblduck iggy: sorry for the blowup.
21:06 aparsons joined #salt
21:06 iggy MTecknology: I want to say someone else was having a similar issue recently
21:07 iggy invsblduck: you can't, you would put in your state {{ pillar.host }}.{{ pillar.domain }} (or something close to that anyway)
21:07 SheetiS MTecknology: what zeromq version is running?
21:07 invsblduck iggy: because pillar data isn't available via jinja when pillar files are being parsed, correct?
21:08 iggy invsblduck: but why are you putting your domain name in pillar? it's already available in the grains
21:09 iggy invsblduck: yeah, you can't rely on pillar's being parsed in a specific order (or something like that... I've never even tried putting jinja in pillars)
21:09 MTecknology SheetiS: looks like 2.1.9; how hard is it to update that in centos?
21:09 iggy doesn't epel have a newer version?
21:09 SheetiS ugh I am not sure on CentOS5... I know tha versions < 3.2 will cause issues.
21:09 SheetiS epel6 does, not sure on epel5
21:10 SheetiS MTecknology: http://download.opensuse.org/repositories/home:/fengshuo:/zeromq/CentOS_CentOS-5/home:fengshuo:zeromq.repo worst case
21:10 kballou joined #salt
21:11 davet1 joined #salt
21:11 invsblduck iggy: i have a long.list.of.sub.domains and i'm tired of {% set site = domain.split('.')[2] -%} or similar in every file.  but you're right - i think the premise of templating the pillar data is incorrect and needs to be moved into the state files or templates.
21:11 SheetiS 2.1.9 is latest in epel5
21:12 SheetiS MTecknology: that repo URL comes right from http://zeromq.org/distro:centos in case you are worried about it.
21:13 invsblduck iggy: i actually need to go offline (don't have my znc bouncer anymore).  thank you for your help.
21:14 invsblduck ttyl!
21:14 dsolsona joined #salt
21:15 iggy good luck
21:18 evidence seems a handful of people are noticing issues with the syndics since rc2, when a command comes down from the root master, the syndics don't properly creation the jobcache directories, so they don't save the minion return data
21:18 evidence which results in the true master never getting data back
21:19 evidence running commands from the syndics works properly, and the job cache dirs are created
21:19 mnguyen joined #salt
21:20 murrdoc is this ticketed ?
21:22 evidence we've been talking with mark (proserve @ saltstack), and been discussing with others on the google group gathering info, but no i don't believe it's ticketed on github yet
21:22 evidence well, haven't heard back from mark yet, but we have inquiries in w him
21:25 sunkist_ joined #salt
21:26 murrdoc ticket it for the win
21:27 SheetiS MTecknology: let me know if a new ZeroMQ doesn't help.
21:27 SheetiS I'm going to head out for 30mins-1hr depending on traffic though.  Don't want to still be at the office anymore
21:28 analogbyte joined #salt
21:29 MTecknology SheetiS: sorry ... I had to run off due to a major fire. I'm still putting it out. I'll have to try this out tomorrow and get back to you then. Thanks!
21:29 ghanima question
21:29 SheetiS no problem.  Just wanted to let you know I will be out of pocket for a bit here soon.
21:29 ghanima is there a limit to how many minions I can have associated with a master
21:29 ghanima it seems like I can't go beyond 999
21:29 hasues I'm curious.  When writing pillar sls files, if you make a subdirectory, like says users/init.sls, is init.sls represent like say 'users' the same way an "index.html" does a folder in html?
21:30 ghanima I am seeing behavior when trying to start all my machines(1100) and that the number stays consistent at 999 but the machines that don't respond seem to vary
21:30 ghanima I presume this is case because its the survival of the fastest whoever checks in first gets a slot
21:31 MTecknology Dangit! It looks like I now have two debian boxes showing the same issue.
21:32 MTecknology I'm done for today. I can't do it anymore...
21:32 iggy ghanima: I can't imagine they'd have a hard limit that low
21:33 ghanima iggy: maybe filehandler limitation on the salt master?
21:33 iggy could be
21:33 iggy ulimit
21:34 ghanima open files is set to 1024
21:34 ghanima for root
21:34 ghanima let me increase and see what happens
21:37 intellix joined #salt
21:37 n8n joined #salt
21:37 MTecknology SheetiS: actually- I'm trying it out now
21:37 cberndt joined #salt
21:37 SheetiS ok I'll hang loose for a few more minutes then
21:37 MTecknology --> Missing Dependency: libzmq.so.1 is needed by package python26-zmq-2.1.9-3.el5.i386 (installed)
21:38 SheetiS hmm
21:38 SheetiS so you'd have to get python zmq from pip
21:38 SheetiS what a rabbit hole
21:38 SheetiS at that point other rpm deps go to @#$@!$#!
21:40 MTecknology I can't even find pip in the repos..
21:40 SheetiS https://github.com/saltstack/salt/issues/4440 shows from a long time ago zeromq 2 issues.  https://github.com/saltstack/salt/issues/5318 has this http://jcollie.fedorapeople.org/zmq/ that seems to have all the rpms you might need
21:40 murrdoc MTecknology:  python2.6-pip
21:40 murrdoc should be there
21:40 MTecknology not seeing it
21:40 SheetiS I think on CentOS5 I had to easy-install pip
21:41 MTecknology I guess I'll use easy-install-2.6 to install pip and then pip to install zmq
21:41 SheetiS MTecknology: before doing the pip thing, take a look at that repo I just linked
21:41 Ryan_Lane is there any way to explicitly throw an error from a jinja template?
21:41 Ryan_Lane especially a way that would display in the state?
21:41 gyre007_ joined #salt
21:42 MTecknology SheetiS: is there a magical repo file?
21:42 ]V[ joined #salt
21:42 ramishra joined #salt
21:42 mgw Ryan_Lane: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.test.html
21:43 Ryan_Lane mgw: I need to raise an error from a jinja template file :)
21:43 Ryan_Lane so that the file.managed state errors out
21:43 Ryan_Lane ideally with a custom error message
21:43 mgw Ryan: I see
21:43 mgw Ryan_Lane: ^
21:44 iggy sounds like a weird use case
21:44 Ryan_Lane would be kind of awesome to be able to raise exceptions
21:44 Ryan_Lane well, if I have a pillar that's incorrectly defined, I'd like to raise an exception
21:44 SheetiS Ryan_Lane: To work around, I woudl just wrap the file.managed in a jinja conditional and the nuse the else as the test state with the same state name and your error.
21:44 iggy don't use .get()?
21:44 iggy it'll definitely error out then
21:44 murrdoc __salt__.raise() would be nice
21:45 iggy not necessarily with a useful message
21:45 Ryan_Lane SheetiS: yeah, we're trying to avoid logic in both the state and the template
21:45 SheetiS Makes sense
21:45 Ryan_Lane and passing pillar data in from context can munge the pillar data
21:45 mgw I wonder if an execution module could be used here
21:45 mgw a custom one
21:46 mgw __salt__['test.error']("some message")
21:46 Ryan_Lane oh. interesting...
21:46 mgw and have test.error raise a python Exception
21:46 mgw It might just blow up... but it might work
21:46 Ryan_Lane I wonder if this would work: http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.test.html#salt.modules.test.exception
21:46 mgw It would definitely spew stuff into your minion log
21:46 MTecknology SheetiS: http://dpaste.com/0NGJQJC -- still having the same issue
21:46 mgw Ryan_Lane: exactly
21:46 smcquay joined #salt
21:47 mgw I forgot that was there, but that' precisely what I had in mind
21:47 MTecknology SheetiS: I'm with you... it's time to get out of here. I'll fight it more tomorrow.
21:47 SheetiS Ryan_Lane: If that works, make sure to share back as I have some uses for that as well :D
21:47 Ryan_Lane testing :)
21:48 SheetiS MTecknology: sounds good.  We'll catch up tomorrow
21:48 cads joined #salt
21:49 beneggett joined #salt
21:49 mosen joined #salt
21:49 Ryan_Lane Unable to manage file: Jinja syntax error: Encountered unknown tag 'salt'
21:49 Ryan_Lane o.O
21:49 Ryan_Lane {% salt['test.exception']('failed!') %}
21:49 SheetiS Ryan_Lane: "do"
21:50 mgw __salt__['test.exceopton']
21:50 SheetiS {% do salt['test.exception']('failed!') %}
21:50 Ryan_Lane ah. cool
21:50 kballou joined #salt
21:50 invsblduck joined #salt
21:50 stephas joined #salt
21:50 mgw one of those... I can't keep track of where salt works and where __salt__ is needed
21:51 SheetiS mgw: __salt__ would be from a python context.  the __ are not required in jinja.
21:51 Ahrotahntee will exclude_pat work if I give it a path to exclude? for example: conf.d/local.conf?
21:51 mgw yeah
21:51 Ahrotahntee and I could replace local.conf with something like (.+) ?
21:51 SheetiS test.exception is not available in 2014.1 in case you are running that still
21:51 mgw "yeah" was at SheetiS
21:51 mgw sorry Ahrontahntee
21:51 Ahrotahntee ha, sorry
21:52 mosen hiya saltines
21:52 Ryan_Lane it works
21:52 mgw Ryan_Lane: awesome
21:52 invsblduck joined #salt
21:53 nitti joined #salt
21:54 SheetiS Ahrotahntee: it's a glob by default for exclude_pat
21:54 SheetiS but the function looks to support regex if you prepend E@
21:54 SheetiS It also looks to be only in 2014.7
21:55 SheetiS https://github.com/saltstack/salt/blob/2014.1/salt/utils/__init__.py does not have the check_include_exclude method defined.
21:55 SheetiS but 2014.7 does
21:55 SheetiS Ryan_Lane: That's awesome.  I really need to upgrade to 2014.7 soon to get all this sweetness.
21:56 Ryan_Lane yeah, it's pretty awesome
21:56 Ahrotahntee recurse isn't doing what I'm expecting.
21:56 Ahrotahntee when I exclude a file (or folder) recurse removes it instead of ignoring it
21:57 SheetiS Heck if I get any more _modules _returners /srv/execution_modules/*, I might actually be running 2014.7 in all but name
21:57 Ryan_Lane using the exclude pattern?
21:57 Ryan_Lane Ahrotahntee: which version are you using?
21:57 Ahrotahntee huh, Hydrogen.
21:57 Ahrotahntee this is old
21:58 Ahrotahntee 2014.1.11
21:58 SheetiS the exclude_pat requires 2014.7 (per above)
21:58 Ahrotahntee yeah
21:58 Ryan_Lane ah
21:58 Ahrotahntee that's weird I thought I was on 2014.7
21:58 Ryan_Lane I've been using the develop branch so long I forget what isn't available :D
21:58 iggy it's not out yet...
21:59 SheetiS Well y'all have fun... I really need to head out now.  I'll be back at it tonight or tomorrow.
21:59 ghanima iggy: no dice
21:59 ghanima still seems to not go beyond 999 minions
21:59 ghanima even after raising fh limit
22:00 bhosmer_ joined #salt
22:02 aparsons_ joined #salt
22:02 iggy and restarting the master?
22:02 ghanima Is anyone aware of a salt minion limitation of 999 hosts
22:02 ghanima yes and restarted the salt-master process twice
22:02 iggy logout and log back in?
22:03 aparson__ joined #salt
22:03 ioio joined #salt
22:03 UtahDave ghanima: no, there should be no limit of 999 hosts.
22:04 evidence so the syndic issue is 2 seperate bugs..
22:04 evidence the first is that this needs to be a makedirs - https://github.com/saltstack/salt/blob/develop/salt/returners/local_cache.py#L144
22:04 UtahDave ghanima: you might need to modify your recon_* values a bit
22:05 ghanima UtahDave: thank you I forgot I did this awhile ago
22:06 ghanima on all the minions I have recon_default set to 1000
22:06 ghanima I thought this was a setting solely on the minion
22:06 ghanima but the master as this setting as well
22:07 Ahrotahntee I guess I'll wait for Hydrogen release, in the meanwhile..
22:07 Ahrotahntee I mean Helium
22:08 Ahrotahntee the other H-element. I'll work around it
22:11 tkharju joined #salt
22:11 ghanima UtahDave: nothing is really glaring at me on this: http://paste.ubuntu.com/8632781/
22:12 mgw joined #salt
22:12 UtahDave ghanima: what happens when you go past 999 minions?
22:12 gideonr joined #salt
22:12 murrdoc joined #salt
22:12 ghanima UtahDave: nothing I have 1087 and only 999 to be able to register with the master. The others are not and there is no error on the minion side
22:13 invsblduck joined #salt
22:13 UtahDave do they show up in salt-key -L   ?
22:13 ghanima yes it does show up in salt-key -L
22:14 UtahDave ghanima: try adding -t 30  to the end of your command
22:15 UtahDave tell me what happens
22:15 jbirdman joined #salt
22:16 ghanima UtahDave: so I did a test.ping
22:16 gideonr joined #salt
22:16 ghanima and it paused waiting for output and then it shows the output of the hosts
22:16 ghanima that     Minion did not return
22:19 aparsons joined #salt
22:19 alex-mesos joined #salt
22:21 UtahDave ?
22:22 ghanima UtahDave: sorry was that question for me
22:22 kermit joined #salt
22:23 UtahDave ghanima: yeah, sorry.
22:23 UtahDave I didn't quite understand what you meant.
22:24 ghanima UtahDave: you asked me what happens when I add -t 30 to a salt command
22:24 ghanima I ran sudo salt '*' test.ping -t 30 -v
22:24 ghanima it showed all the hosts that were in True state
22:24 ghanima and then the output paused(sorry didn't time it)
22:25 ghanima then after some time it showed the rest of the output
22:25 mgw joined #salt
22:25 ghanima which was the output of the hosts that returned false
22:26 ghanima UtahDave:sorry hope that makes sense
22:27 UtahDave gotcha
22:28 UtahDave ghanima: count how many minions show as return true when you run     salt-run jobs.lookup_jid <job id>
22:28 UtahDave where <job id>  is the id of that last test.ping you ran
22:29 superseb joined #salt
22:30 jalaziz_ joined #salt
22:31 ghanima UtahDave: 1077
22:32 ghanima is the number that comes back
22:32 UtahDave ghanima: is that what you would expect?
22:32 UtahDave salt-run jobs.lookup_jid <job id> | grep -c True
22:32 ghanima yes that is what I expect
22:33 ghanima that is the total amount of hosts I tried to integrate today
22:33 UtahDave ok, so Salt is working fine. It's just that the master at the cli isn't waiting for all the responses.
22:34 UtahDave remember Salt is asynchronous. The salt-master cli is trying to provide a somewhat synchronous view of what's happening
22:34 UtahDave I think we just need to modify your recon_* settings a little bit.
22:34 UtahDave just a sec
22:36 UtahDave ghanima: how many worker_threads do you have set in your /etc/salt/master   ?
22:37 ghanima I didn't change the default so I presume 5 is the default
22:38 UtahDave ghanima: ok, let's bump that to 10
22:38 ghanima ok done
22:38 UtahDave also, I should have asked already. what version of salt on your master and on your minions?
22:39 ghanima 2014.1.5
22:39 UtahDave ok and try setting  random_reauth_delay: 10
22:39 UtahDave then restart the salt-master daemon and try again
22:40 arthabaska joined #salt
22:40 canci joined #salt
22:41 hasues UtahDave: Curious, does each directory defined to configure something in Salt have an init.sls?  So for instance, if I want to configure apache and I reference it by - apache, is that going to reference apache/init.sls?
22:42 aurynn it'll reference apache.sls or apache/init.sls
22:42 hasues aurynn: Ah, good to know.
22:42 aurynn afaik, anyway
22:42 aurynn :)
22:42 hasues uh oh
22:43 hasues You aren't fooling me.  you know! :P
22:43 UtahDave aurynn is correct
22:43 aurynn well this is what I've seen in my experiments. I don't know what happens when you have both, tho
22:43 ramishra joined #salt
22:44 utahcon does a watch --- always watch, or only while the state is running watch?
22:44 hasues Hey, as long as I figure out one more thing each day.. :)
22:44 UtahDave I think it looks for apache.sls first, then the init in a directory.  But I would definitely avoid that scenario and definitely test that it behaves that way before depending on it
22:44 UtahDave utahcon: a watch only watches when you run a highstate
22:44 ghanima UtahDave: sorry the recon setting should i set that on the minion or master or both
22:44 UtahDave ghanima: only on the master
22:45 ghanima UtahDave: please note I do have random_reauth_delay set at the minion as well as others settings http://paste.ubuntu.com/8632781/
22:46 ghanima UtahDave will that cause a conflict
22:46 vbabiy joined #salt
22:46 hasues One day, I'll understand what differentiates pillar/ vs salt/
22:47 UtahDave ghanima: Gah! I'm so sorry.  Those settings go in the minion configs.  Which is not convenient at all.
22:47 libertyy joined #salt
22:48 aquinas joined #salt
22:49 ghanima UtahDave: not a problem I have a way to hit minions without using salt
22:49 ghanima but are the other setting appropriate
22:49 murrdoc left #salt
22:49 murrdoc joined #salt
22:49 ajolo joined #salt
22:49 evidence so it appears prep_jid isn't being properly called before save_load - https://github.com/saltstack/salt/blob/develop/salt/master.py#L2290
22:49 stephas joined #salt
22:50 UtahDave ghanima: I think so.  just drop that random_reauth_delay to 10
22:50 evidence so save_load errors when it can't write .load.p
22:50 UtahDave evidence: what master job cache are you trying to use?
22:50 evidence this is on the syndics, they aren't able to write the job cache and the jid_dir doesn't exist
22:50 evidence so they never pass the return up the chain
22:51 evidence rc5.. been happening since rc2 it seems
22:51 evidence oh sorry, not an external one
22:51 evidence /var/cache/salt/master/jobs
22:52 evidence Could not write job invocation cache file:
22:52 evidence [Errno 2] No such file or directory: '/var/cache/salt/master/jobs/58/1d0c51faefdc85ddccec5278131dfc/.load.p'
22:52 wnkz joined #salt
22:53 bezeee joined #salt
22:54 lowkeyliesmith joined #salt
22:54 ghanima UtahDave: that worked
22:54 ghanima UtahDave: no that did not
22:54 ghanima one sec
22:55 gideonr joined #salt
22:56 iggy is there a way to force a mine update before a state runs?
22:57 ghanima UtahDave: lol so I did the test.ping without -v and now instead of 999 its now at 1000
22:57 UtahDave ghanima: if you count the minions that return true in the job cache is it the full number still?
23:00 jalaziz joined #salt
23:00 libertyy i'm trying to separate out passwords using environments in pillar_roots. I thought if i set pillar_roots environment in the master config, i could keep separate trees/repos for credenttials. however, the master is never parsing the non-base environments in my pillar/top.sls       https://gist.github.com/libertyy/0508bdee4c995e80dbbd
23:00 libertyy i'm not sure what I'm doing wrong here ...
23:00 utahcon I am going to ask something dumb... is it possible to hang a state on a loop waiting for a change in state?
23:00 utahcon do this... wait until this other thing changes... then finish
23:01 ghanima UtahDave: yes it still return the full number of hosts
23:01 ghanima when doing a job search
23:01 iggy utahcon: I'd think the reactor could be leveraged for that
23:01 iggy if I understand what you're trying to do correctly
23:01 stephas joined #salt
23:01 UtahDave ghanima: so everything is still working fine, it's just the display that's not getting back to you correctly
23:01 UtahDave utahcon: are you going to come to the SaltStack meetup tomorrow night?
23:02 ghanima UtahDave: yeah it looks like is there anyway to find out how long it took for the entire job to finish executing
23:03 UtahDave ghanima: I think if you upgraded you'd find a bunch of fixes for what you're seeing.
23:04 ghanima UtahDave: do you recommend the latest release candidate
23:04 UtahDave libertyy: I think that looks good.  What behavior are you seeing?
23:04 dsolsona joined #salt
23:04 ajolo joined #salt
23:04 UtahDave ghanima: it's been working pretty well for me.  But I'd definitely test first to make sure all your states work correctly on it
23:05 UtahDave ghanima: what's the latest version available in your OS's package repo?
23:05 ghanima UtahDave: haven't built any state files yet
23:05 ghanima I am running centos
23:05 stephas joined #salt
23:05 ghanima so I just normally download the latest release candidate and add it to a local repo in our office
23:05 libertyy my minions get pillar data set from all stuff under base, but it acts like it's never parsing anything in the other pillar_roots
23:06 utahcon UtahDave: SLC is a bit of a drive for a meetup now (Living in Seattle)
23:06 UtahDave libertyy: sorry for the dumb question, but just to make sure.   did you restart the salt-master daemon after modifying the config file?
23:06 libertyy if i change the matching rule for e.g. local_dev to '*', i still don't get pillar data from the non-base pillar repo
23:07 UtahDave utahcon: Ah, that's right! I forgot about that.  Congrats on the new job.  Liking it so far?
23:07 fragamus_ joined #salt
23:07 libertyy yea, multiple times... was hoping it was that easy :-)
23:07 utahcon Loving it!
23:07 UtahDave libertyy: did you run   salt \* saltutil.refresh_pillar    ?
23:07 utahcon getting to dabble in all sorts of fun things
23:07 UtahDave utahcon: cool!
23:07 libertyy yuppers. my testing  from the command line is    salt '*' saltutil.refresh_pillar && salt '*' pillar.get foo
23:08 utahcon like tricky salt state dependency concurrency issues
23:08 aquinas joined #salt
23:09 libertyy does my local_dev/pillar also need a top.sls ?
23:09 stephas joined #salt
23:09 UtahDave libertyy: no, it doesn't.  In fact, I generally recommend against it.
23:09 UtahDave your main top.sls can manage all the environments
23:09 libertyy is that the pillar top.sls ?
23:09 UtahDave yeah
23:10 UtahDave libertyy: let's try this.
23:10 UtahDave libertyy: open another terminal into your salt master
23:10 UtahDave stop the salt master daemon
23:10 UtahDave then run     sudo salt-master -l debug
23:11 UtahDave then in try your refresh and everything in the original window
23:12 aparsons joined #salt
23:12 tf198_ joined #salt
23:14 ajolo joined #salt
23:15 UtahDave also, libertyy, I noticed that you refer to a production environment in your pillar top.sls, but there's isn't a production environment defined in your pillar_roots config
23:16 bezeee joined #salt
23:16 libertyy UtahDave, yea, sorry, that was a mistake. I'm currently not defining the production environment in pillar top.sls
23:16 UtahDave ok
23:17 libertyy Okay, the debug stuff is helping ... I get this line:     Specified SLS 'app.creds' in environment 'local_dev' is not found, which might be due to environment 'local_dev' not being present in "pillar_roots" yet!
23:17 UtahDave ah, there you go.
23:17 UtahDave you've defined local_dev as    /srv/local_dev/pillar
23:18 UtahDave but in your gist you have   /srv/local_dev/app/creds.sls
23:18 UtahDave move that to   /srv/local_dev/pillar/app/creds.sls
23:18 StDiluted joined #salt
23:20 DaveQB joined #salt
23:22 evidence UtahDave: went ahead and created a ticket on it - https://github.com/saltstack/salt/issues/16825
23:23 UtahDave thanks, evidence!
23:23 UtahDave evidence: hey, do you have verify_env: False in your master configs?
23:23 tf198_ Hi all - What is best practice for handling errors in modules.  Raising salt.exceptions.CommandExecutionError seems appropriate but there are some cases of just returning a string in the code...
23:26 UtahDave tf198_: Yeah, raise the CommandExecutionError
23:26 evidence UtahDave: negative, we are with the default which seems to be True
23:26 tf198_ Okay - just about to open a ticket but I might chuck in a fix as well...
23:27 UtahDave evidence: ok.  thanks for opening up that issue ont hat
23:27 UtahDave on that
23:27 UtahDave tf198_: cool! that would be very helpful!
23:30 Outlander joined #salt
23:34 ghanima UtahDave I am looking at the available downloads for salt and it looks like the epel rpm is 2 releases behind. Your original recommnedation was to upgrade and wanted to know if it matters if I upgrade to 2014.1.10 and not 2014.1.12
23:35 UtahDave the RC5 can be found here: http://copr-be.cloud.fedoraproject.org/results/saltstack/salt/epel-6-x86_64/salt-2014.7.0rc5-2.el5/
23:35 murrdoc woah
23:35 UtahDave .13 should be available pretty soon
23:35 murrdoc do u have rc's for ubuntu ?
23:36 murrdoc like debs
23:36 UtahDave murrdoc: Hm. I'm not sure, actually. lemme check
23:36 murrdoc thanks UtahDave
23:36 ajolo joined #salt
23:36 ghanima UtahDave: I see you have rc5-2 and rc5-3
23:36 ghanima is there a difference at this point
23:37 bhosmer joined #salt
23:39 vbabiy joined #salt
23:42 rattmuff joined #salt
23:43 sjaredj joined #salt
23:44 ramishra joined #salt
23:46 n8n joined #salt
23:47 alex-mesos joined #salt
23:47 libertyy UtahDave, thanks for your help! I'm all squared away now!
23:48 alex-mesos joined #salt
23:51 rattmuff joined #salt
23:52 pdayton joined #salt
23:55 ajolo joined #salt

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