Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-12-02

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

All times shown according to UTC.

Time Nick Message
00:11 oz_akan_ joined #salt
00:12 cachedout joined #salt
00:14 Sheco joined #salt
00:14 dstanek_afk joined #salt
00:16 zandy joined #salt
00:20 jeddi I've only ever thought of one use case for salt-ssh, and that's for a raspberry pi (where salt-minion is a bit harder to get, resources are a bit tighter (but not sure there's much advantage?) and some dependencies aren't available - libzmq3 f.e.)
00:21 jeddi Mind, I'm working with modest sized networks that I pretty much control, and that don't have weird transit restrictions.
00:25 khaije|io could Salt SSH be used to bootstrap a salt-minion install? I don't know much about it but I thought I read that it somehow has little or no dependencies...
00:25 forrest khaije|io sure why not?
00:25 jeddi khaije|io: that's a possibility, i guess.   on debian ssh isn't installed by default, but it'd be easier to script in than an external repository.
00:26 khaije|io well it just seems that it would be a principle use but I haven't heard it mentioned by anyone
00:26 jeddi khaije|io: i've started looking again at how i can provision using salt for most of the heavy lifting.   the initial contact / config as ever is the problem.
00:27 khaije|io sure, Salt Virt is part of an answer
00:29 khaije|io something like https://github.com/puppetlabs/Razor (for Salt, of course) is probably the rest of the answer
00:31 zandy joined #salt
00:33 joehh jeddi: are you using raspbian? libzmq3 should be available...
00:37 zandy joined #salt
00:39 NV I'm planning on using salt-ssh to manage vyatta-based routers :D
00:39 NV writing a salt module to manage the config atm
00:43 __number5__ NV: they have python installed?
00:44 NV yup :D
00:45 NV at least on ubiquiti edgemax routers they do (ubiquiti's os that runs on it is a vyatta fork called edgeos)
00:46 khaije|io is python the only dependency for salt-ssh ?
00:47 * cast would imagine ssh was, too
00:49 Sheco joined #salt
00:50 jeddi joehh: ooh - is it?  i haven't looked for a couple of months, as i was working fine with 0.16 (?) and the vanilla zmq on the device .. and yeah, raspbian.
00:52 jeddi khaije|io: salt-virt looks good .. looks like it's had a lot of work done since i last looked, too.  just reading the docs now it seems esx is still a poor cousin to kvm (which would be fine by for the fact i'm forced to use esx)
00:52 joehh yeah - I can't find it in the docs, but follow the instructions for wheezy and it should just work
00:52 joehh I can't remember if bootstrap works for it either
00:52 jeddi joehh: ack.  shall do.  recently moved, haven't set up the pi again yet to play with.
00:53 forrest left #salt
00:54 khaije|io lol @ cast
00:54 tdillio joined #salt
00:55 khaije|io +1 "forced to use esx"
00:56 AdamSewell joined #salt
00:58 stefanmonkey joined #salt
01:02 Teknix joined #salt
01:03 tdillio I'm having trouble connecting a fedora minion to a centos master (first time trying this) and I realized that the salt version on centos is 0.17.1 and fedora is 0.17.2, would this mis-match cause issues?
01:07 dstanek_afk joined #salt
01:08 mgw joined #salt
01:12 oz_akan_ joined #salt
01:18 bhosmer_ joined #salt
01:22 zandy joined #salt
01:27 malinoff joined #salt
01:29 nmistry joined #salt
01:36 redondos joined #salt
01:40 Tekni joined #salt
01:43 forresta joined #salt
01:59 AdamSewell joined #salt
02:03 mgw joined #salt
02:04 oz_akan_ joined #salt
02:08 jslatts joined #salt
02:13 ajw0100 joined #salt
02:26 lpmulligan joined #salt
02:32 favadi joined #salt
02:36 redondos joined #salt
02:43 dwyerj joined #salt
02:57 xl1 joined #salt
02:58 chad joined #salt
03:00 chad Is there something wrong with this sls fiile?   Every time I run it I get a error.   http://pastebin.com/WN60a3Cz  If I run it on a new system I get the error "The named service strongswan is not available" If I install strongswan by hand then it goes and starts it for me like saltstack should
03:02 forresta chad, is this on 0.17.2, and what distro are you on?
03:02 forresta also can you try changing pkg.installed to pkg: \n  - installed ?
03:02 forresta sometimes mixing causes issues
03:04 chad I am on centos 6.4
03:05 malinoff chad, Please, pastebin `salt-master --versions-report`
03:05 stanchan joined #salt
03:06 chad I am trying to just set this up with "salt-call --local state.highstate"
03:07 chad so I do not have a salt-master right now
03:07 malinoff Ok, what about `salt-minion --versions-report`?
03:09 chad It is all the most recent version http://pastebin.com/N8t8CwZY
03:11 redondos joined #salt
03:11 forresta try it like this chad: https://gist.github.com/gravyboat/7744396
03:11 malinoff chad, You should add `state_auto_order: True` option in minion conf, or use ordering: http://docs.saltstack.com/ref/states/ordering.html
03:12 forresta malinoff, ordering should be enabled by default
03:12 forresta he's on 0.17.2
03:13 malinoff forresta, He's on 0.17.1
03:14 forresta oh you're right duh
03:14 malinoff And no, state_auto_order is disabled by default.
03:14 forresta are you sure? I thought that the False value was commented out
03:15 chad so add a require like this http://pastebin.com/AJKejkFp
03:16 malinoff forresta, There is no `state_auto_order` option in the config files. You should add it manually. And yes, it is disabled by default. You can check this there: https://github.com/saltstack/salt/tree/v0.17.2/conf
03:17 malinoff chad, Nice. Should work
03:17 forresta weird, could have sworn it was enabled by default, I'll have to talk to the Salt guys tomorrow to clear that up so the docs can be modified.
03:17 redondos joined #salt
03:18 chad well I treid it and I still got the error http://pastebin.com/W1e7gr5h
03:18 forresta chad, can you try the gist I pasted above?
03:18 forresta 0.17.1 has a lot of bugs, so I don't know if maybe you're hitting one.
03:18 chad forresta: sure thing, one sec
03:19 malinoff forresta, You shouldn't say "c'mon, do what I say". Try to dive into his problem. And your solution won't work.
03:20 malinoff chad, Your sls file has two identical IDs - but they're should be unique
03:20 forresta ? Sometimes the yaml doesn't like when you mix asdf.zxcv, and asdff: \n -zxcv
03:20 malinoff forresta, not this time, you know
03:21 forresta Yea sure, might as well check the easy stuff first though :P
03:21 chad forresta: that did the trick
03:21 forresta cool
03:22 malinoff chad, I suggest not to use the short form of state declaration. Try this one: http://pastebin.com/CV0sgK92
03:23 xl1 joined #salt
03:27 chad malinoff: that worked also, I just had to add a ":" - name
03:28 chad is it because I did not have "    - name: strongswan" in my origional one that I had issues
03:28 mapu joined #salt
03:29 malinoff chad, No. You had duplicate IDs.
03:29 malinoff In my example, IDs are unique and salt can handle this
03:30 chad malinoff:  so you mean strongswan package:   vs strongswan service:
03:30 malinoff chad, Yes
03:31 chad I did not realize that those had to be unique id's
03:31 chad I am working on making a salt recie for making a ipsec server appliance
03:31 chad I am learning quite a bit along the way
03:33 malinoff chad, It's simple. In your first example, you had two states with the same ID - 'strongswan'. In my example, IDs are unique - each state belongs to it's own ID - 'strongswan package' and 'strongswan service', as you mentioned above
03:34 chad I see how the unique id's work now, I did not figure that out all the time I tried to figure it out before
03:35 malinoff Salt is quite difficult :)
03:41 chad it makes sense when I think about it.  how it would start up the service when I installed it manually
03:41 malinoff chad, Are you asking a question? :)
03:42 chad nope, just stating my learning experience
03:42 malinoff chad, Good luck :)
03:43 malinoff I just have to say. In salt, states are separated. You should think about them like they're 10000 kilometers away from each other
03:43 malinoff And the source code can give many answers
03:51 dstanek_afk joined #salt
04:03 Nexpro joined #salt
04:05 oz_akan_ joined #salt
04:07 lemao joined #salt
04:16 dyu joined #salt
04:17 dyu hi. is it ok to use salt from pypi? i'm worried coz it's not mentioned in the installation docs http://docs.saltstack.com/topics/installation/index.html
04:18 malinoff dyu, Yes, why not? :) But 0.17.2 is available in deb/rpm repositories AFAIK
04:18 redondos joined #salt
04:23 ckao joined #salt
04:24 Furao joined #salt
04:30 odonnell004 joined #salt
04:30 odonnell004 hello anyone in ?
04:31 jcockhren yo!
04:31 odonnell004 I'm getting a strange error with the git module.  wanted to check to see if anyone was getting similar errors recently.
04:32 jcockhren gist it
04:32 odonnell004 the short error message is : State git.latest found in sls ymanage is unavailable
04:33 jcockhren gist all you can from running the salt process with the debug flag
04:33 jcockhren otherwise, I will have no idea what you're talking about. ;)
04:34 odonnell004 http://pastebin.com/zmwPqC3Y
04:34 odonnell004 its a new repo on the system.  i though git latest should clone if the repo is not there
04:35 jcockhren what's your state file? and top file?
04:39 malinoff odonnell004, Could you pastebin `salt minion-id state.sls yourstate -l debug` ?
04:40 odonnell004 sure
04:41 odonnell004 the section of the sls that sets up the git repo is super small. http://pastebin.com/M94mvMDK
04:41 odonnell004 on sec while i get all of that
04:44 malinoff odonnell004, Are you sure, that repo is available via https?
04:44 odonnell004 yep
04:44 malinoff left #salt
04:44 odonnell004 copied it out of git hub
04:44 dmalinovsky joined #salt
04:44 dmalinovsky odonnell004, without asking for username/password?
04:45 odonnell004 yes
04:45 dmalinovsky Strange thing. I'll wait for your debug log
04:48 odonnell004 heres the repo. and sls for the git repo.  just double checked it public https://gist.github.com/MattyO/7745216
04:53 odonnell004 super sorry. im really dumb.  git wan't installed.  really sorry for wasting your time
04:53 favadi joined #salt
04:54 dstanek_afk joined #salt
04:54 odonnell004 new server and forgot it wasn't in a sls for all machines
04:55 dmalinovsky odonnell004, :D no problem
04:55 dmalinovsky odonnell004, salt \* state.sls yourstate test=True is very useful :)
05:03 malinoff joined #salt
05:06 oz_akan_ joined #salt
05:08 hhatfield joined #salt
05:08 bemehow joined #salt
05:08 jsm joined #salt
05:10 zandy joined #salt
05:11 favadi joined #salt
05:11 * malinoff testing /me
05:15 wilywonka joined #salt
05:19 redondos joined #salt
05:21 zandy joined #salt
05:23 zandy joined #salt
05:36 zandy joined #salt
05:39 zandy joined #salt
05:40 geo23589 joined #salt
05:42 zandy joined #salt
05:48 bemehow joined #salt
06:02 sandGorgon joined #salt
06:04 rmt joined #salt
06:07 oz_akan_ joined #salt
06:08 lemao joined #salt
06:20 redondos joined #salt
06:25 bemehow joined #salt
06:30 wilywonka joined #salt
06:31 stanchan joined #salt
06:33 middleman_ joined #salt
06:53 Furao joined #salt
07:02 bemehow joined #salt
07:08 oz_akan_ joined #salt
07:10 dyu left #salt
07:19 gammalget joined #salt
07:20 redondos joined #salt
07:45 crashmag_ joined #salt
07:53 jeddi joined #salt
07:57 dstanek_afk joined #salt
08:00 wiggy joined #salt
08:01 harobed joined #salt
08:09 oz_akan_ joined #salt
08:09 lemao joined #salt
08:11 juasiepo joined #salt
08:21 redondos joined #salt
08:30 slav0nic joined #salt
08:35 m0hit_ joined #salt
08:35 m0hit__ joined #salt
08:43 giantlock joined #salt
08:43 ravibhure joined #salt
08:51 tinuva joined #salt
08:54 crashmag joined #salt
09:01 crashmag joined #salt
09:09 oz_akan_ joined #salt
09:12 prooty joined #salt
09:14 prooty hello. i know i've asked this before, but no one has given me a helpful answer yet. could someone explain how to use cmd_yaml in external pillars and how to integrate it with /srv/pillar/top.sls?
09:15 mike251 joined #salt
09:15 * mike251 morning from Frankfurt
09:20 malinoff prooty, Probably, because external pillars are not so popular :) try to ask in the mailing-list
09:21 prooty okay malinoff.
09:22 redondos joined #salt
09:25 ravibhure joined #salt
09:28 ajw0100 joined #salt
09:29 1JTABBYVA joined #salt
09:31 juasiepo joined #salt
09:47 pengunix joined #salt
09:50 pengunix joined #salt
09:50 zooz joined #salt
09:51 zandy joined #salt
09:54 harobed joined #salt
09:55 harobed joined #salt
09:56 harobed joined #salt
09:57 favadi joined #salt
09:58 dstanek_afk joined #salt
09:59 Furao joined #salt
10:01 backjlack joined #salt
10:10 mike251 hey guys,,, i have 2 states... dns and ldap state...  how can i be sure that ldap state is run AFTER the dns state? In order to have ldap, i need to be able to solve ns . Thanks
10:10 oz_akan_ joined #salt
10:10 malinoff mike251, http://docs.saltstack.com/ref/states/ordering.html
10:11 mike251 oh malinoff thanks... i think i didn't see that one. thankS!
10:11 malinoff :)
10:12 mike251 hmm malinoff i have  dns.sls and ldap.sls ... is there way to order running different sls files?
10:12 lemao joined #salt
10:13 malinoff mike251, you should include one in another - and set up `require` attribute
10:13 mike251 malinoff:  thanks AGAIN
10:14 malinoff mike251, np :) That's why I'm here :)
10:18 anitak joined #salt
10:23 redondos joined #salt
10:31 MrTango joined #salt
10:32 mike251 comments in a pillar sls file can be added with {##}  ?
10:32 malinoff mike251, with jinja renderer, yes
10:32 mike251 default is jinja,right?
10:33 malinoff Yes :)
10:33 mike251 cool :) i am on the right track
10:34 mikkn joined #salt
10:35 malinoff mike251, btw, the best way to test something is to install salt for development (from git, tagged, for example, by v0.17.2)
10:35 malinoff You can run master and minion locally, on the same machine
10:35 mikn joined #salt
10:35 malinoff And when you feel ready to production, you just do `rm -rf salt_venv`
10:37 mike251 aha... didn't know that... i have a master and a minion
10:39 sunand joined #salt
10:54 xl1 left #salt
10:59 mike251 just to double check ... in states.. the include method.. .is just adding the information from the included states into the current one... and the file is compiled...after everything is included etc...   i mean... the info that is included .. is not include and run? ..right?
11:01 malinoff mike251, your message is difficult to read and understand because of all these dots, sorry :(
11:01 londo_ joined #salt
11:07 wiggy left #salt
11:11 oz_akan_ joined #salt
11:12 anuvrat joined #salt
11:17 xl1 joined #salt
11:18 oz_akan_ joined #salt
11:21 dccc joined #salt
11:23 redondos joined #salt
11:25 mike251 hi malinoff ... yeah i realize that.
11:26 mike251 including a state in another state... is just including it... i can use cmd.wait in the main sls to wait for a file to be copied from an included sls, right?
11:27 malinoff Right :)
11:27 mike251 great!
11:27 zandy joined #salt
11:27 mike251 salt is pretty awesome :)
11:28 malinoff It totally is :)
11:30 mike251 i am really glad... i have chosen salt over ansible or puppet
11:30 mike251 is more verbose and clean... or at least for me.
11:32 cyrusdavid joined #salt
11:32 mike251 to include a sls from another dir... i should user - app/bashrc or app.bashrc ?
11:34 mike251 is with . :)
11:39 jeddi mike251: you forgot 'or chef'.  :)
11:39 mike251 ah yeah .. chef...
11:40 mike251 and one more... pf engine.. or smth like that
11:40 mike251 one thing that i loved about salt is the community
11:40 mike251 really nice guys around here
11:47 cyrusdavid joined #salt
11:52 diegows joined #salt
11:53 Ertac joined #salt
11:55 tempspace_ joined #salt
11:55 chet_ joined #salt
11:56 aptiko_ joined #salt
11:57 mau__ joined #salt
11:57 drags1 joined #salt
11:57 pmrowla joined #salt
11:57 mjulian joined #salt
11:58 mjulian joined #salt
11:58 crashmag joined #salt
11:59 Guest55571 joined #salt
12:00 dstanek_afk joined #salt
12:02 balboah joined #salt
12:07 ggoZ joined #salt
12:09 sroegner___ joined #salt
12:09 Sheco joined #salt
12:24 redondos joined #salt
12:35 Destro joined #salt
12:36 harobed joined #salt
12:38 krak3n` joined #salt
12:52 oz_akan_ joined #salt
12:57 dangerousbeans joined #salt
13:01 willist joined #salt
13:01 jeffrubic left #salt
13:07 sunand joined #salt
13:07 sunand can we degrade package versions with salt?
13:08 carlos joined #salt
13:11 sunand If i have clang 3.2 installed how can i write the state to use clang.3.1
13:12 nkuttler sunand: which os?
13:15 willist joined #salt
13:21 kyusan joined #salt
13:23 oz_akan_ joined #salt
13:25 redondos joined #salt
13:25 redondos joined #salt
13:27 wilywonka joined #salt
13:33 jslatts joined #salt
13:34 sunand nkuttler ubuntu
13:37 favadi joined #salt
13:38 sunand nkuttler: ubuntu
13:43 mpanetta joined #salt
13:43 JasonSwindle joined #salt
13:44 elfixit joined #salt
13:45 Damoun joined #salt
13:48 dangerousbeans joined #salt
13:49 nfnty joined #salt
13:50 nfnty left #salt
13:53 ipmb joined #salt
13:55 babilen joined #salt
13:56 Sheco joined #salt
14:01 sroegner___ joined #salt
14:03 juicer2 joined #salt
14:07 racooper joined #salt
14:07 Destro left #salt
14:08 carlos__ joined #salt
14:09 mgw joined #salt
14:12 blee joined #salt
14:13 vejdmn joined #salt
14:15 willist joined #salt
14:17 kyusan joined #salt
14:18 nkuttler sunand: well, you can specify package versions with the apt module. downgrading on the other hand is usually not support on debian-like systems
14:18 nkuttler but ask #ubuntu i guess
14:18 nkuttler supported
14:26 Sheco joined #salt
14:26 redondos joined #salt
14:26 mike251 hey guys ... i have an issue with pkg.installed:    http://pastebin.com/VwGcuaJa
14:26 mike251 am i doing something wrong?
14:28 favadi do a salt-call and watch the output
14:29 jergerber joined #salt
14:30 MTecknology Is it possible to have a state execute only if required by another state?
14:30 mike251 favadi:  just salt-calll? never used that
14:30 brianhicks joined #salt
14:30 MTecknology I have one state that needs to execute only if another state executes and that one uses -unless:
14:32 ravibhure joined #salt
14:36 ccic_pgreen joined #salt
14:39 Gifflen joined #salt
14:39 vejdmn1 joined #salt
14:39 cachedout joined #salt
14:41 mgw joined #salt
14:42 cnelsonsic joined #salt
14:42 dangerousbeans_ joined #salt
14:43 quickdry21 joined #salt
14:45 vejdmn joined #salt
14:47 mapu joined #salt
14:49 sunand nkuttler: Thank you for the answer.   So is it supported with salt-ssh also?
14:52 g3cko joined #salt
14:53 mike251 hey guys... any idea... why i have this error?  http://pastebin.com/VwGcuaJa
14:54 kaptk2 joined #salt
14:54 MrTango joined #salt
14:54 NV mike251: i note your indending is wrong in the pastebin, but i assume that was a copy-pasta error because you'd have gotten a different error if that was the issue
14:54 NV run salt-call state.sls mystatehere on the minion in question
14:54 NV it will provide a tad more debug info
14:55 mike251 NV: ok.. one sec
14:55 NV at a guess, the package names are wrong :)
14:55 mike251 no they are not ... i took them from yum
14:55 mike251 i can install with yum install package
14:55 kyusan joined #salt
14:56 mike251 NV thanks for the salt call thing !
14:56 NV work it out? :)
14:56 mike251 NO MORE space on drive
14:56 mike251 NV:  yeah .. .
14:56 NV hahahahahahahaha
14:56 NV that'd do it
14:56 mike251 it was a small vm
14:56 NV fucking hate it when that happens
14:56 mike251 will rebuild it now
14:56 NV gets me all the time too
14:57 mike251 NV:  me too... the vm is installed via spacewalk... and the setup is a bit weird... 50% of the hdd is taken by /var/log ...
14:57 mike251 so i am left with 1G ... or less :) from 8
14:57 cachedout joined #salt
14:57 mike251 anyway - you made my day :)
14:58 * NV pops off to bed
14:58 rlarkin joined #salt
14:58 mike251 NV: sleep tight
14:59 Brew joined #salt
15:04 dstanek_afk joined #salt
15:09 vejdmn joined #salt
15:10 JulianGindi joined #salt
15:14 hhatfield joined #salt
15:18 bastion2202 joined #salt
15:19 slav0nic joined #salt
15:19 JulianGindi joined #salt
15:20 willist joined #salt
15:21 Guest42390 joined #salt
15:21 rubberne_ joined #salt
15:22 rubberne_ How can I get the salt master to log the output that is displayed on the console when changes are made? I've turned up the logging to trace, but no luck....
15:22 mannyt joined #salt
15:23 rubberne_ All I really want is to log the diff output of changes made to files. But logging everything is good too.
15:24 mike251 rubberne_: ... i had the same... problem...
15:24 mike251 actually ... what i came up with ... is creating an entire web app... that is running salt-master
15:24 mike251 so.. it parses the json output... from each highstate... and saves into a mysql DB etc etc
15:25 mike251 the app is not ready yet... but i want to make it pubicly available on bitbucket or github
15:25 rubberne_ mike251: Sounds cool. but too heavy for me, I just want it in the log file.
15:25 mike251 rubberne_:  ... don't know exactly.. i didn't set up logging... i will be looking into that ... at some point... i am dealing with packages and jinja and states
15:26 mike251 sorry i can't help
15:27 redondos joined #salt
15:27 rubberne_ The problem I am trying to solve is where I call highstate, a few files get changed. Then I find that I need to revert a file back to the original.  How can I do this? If the diff output was logged to the log file then that would be fine.
15:28 mike251 but why do you need to revert?
15:28 mike251 just stupid questions from my side.. just curious...
15:28 mike251 the main idea.. is NOT to touch yourself
15:29 mike251 let the system automate the confs etc...
15:29 vejdmn joined #salt
15:31 nmistry joined #salt
15:32 pdayton joined #salt
15:33 forrest joined #salt
15:33 MTecknology http://dpaste.com/1490325/ <-- That works fine for debian, but the pillar data isn't being set for redhat boxes... :(  I can't figure out why.
15:36 forresta MTecknology, can you try changing that from a single if statement to two if statements?
15:36 alunduil joined #salt
15:37 forresta MTecknology, also did you run a refresh.pillar?
15:37 MTecknology forresta: yup, and same result
15:38 forresta with the refresh or with the two if statements?
15:38 MTecknology with the two ifs
15:38 forresta ok, next can you try switching the order in which they go?
15:39 tdillio joined #salt
15:39 MTecknology tried that - same result on both
15:39 tyler-baker joined #salt
15:39 tyler-baker joined #salt
15:39 MTecknology debian has the pillar data, rhel does not
15:39 tdillio Anyone know when 0.17.2 will hit EPEL? I see it's already in Fedora.
15:40 forresta tdillio, it's in epel-testing right now
15:40 forresta it's been there for about 2 weeks?
15:40 forresta so maybe next week or the week after, unless you have some credits to push it through :P
15:41 tdillio Thanks forresta
15:41 forresta np
15:42 MTecknology forresta: any other ideas?
15:42 MTecknology This pillar sls is under '*' in top.sls
15:43 patrek_ joined #salt
15:44 MTecknology I need a break from this dang thing
15:44 forresta MTecknology, It looks good to me man, I don't see why the info wouldn't be getting pulled.
15:44 sandGorgon joined #salt
15:45 terminalmage MTecknology: any errors in the master log when you run saltutil.refresh_pillar for the debian box?
15:45 forresta MTecknology, it really looks as though the pillar data isn't getting populated for some reason, can you generate some static pillar info that would be set regardless of OS, and then see if you can call that on the redhat machine?
15:46 smccarthy joined #salt
15:47 mike251 stupid Q: does the master run saltutil.refresh_pillar ... on a regular basis?
15:47 mike251 do i need to set up cron to run saltutil.refresh_pillar ?
15:48 patrek joined #salt
15:48 terminalmage mike251: it doesn't run that command, but it does refresh pillar data whenever it is requested
15:49 mike251 terminalmage: requested? what do you mean?
15:49 mike251 requested by ME from the CLI ?
15:49 terminalmage no
15:50 terminalmage when it is needed
15:50 opapo joined #salt
15:50 forresta when you run a highstate or something where pillar data is needed by the minions mike251
15:50 mike251 aaa ... so it actually does a pillar refresh .. when i change the pillar data?
15:50 terminalmage do a recursive grep in the code for a refresh_pillar
15:50 mike251 forresta:  yeah
15:50 terminalmage mike251: no
15:50 mike251 :)
15:50 terminalmage not right away
15:51 mike251 a sure... but at some point...
15:51 terminalmage yeah
15:51 mike251 anyway - thanks, sorted now.
15:51 terminalmage np
15:53 forresta MTecknology, did you try setting that static item to see if it's getting pulled in at all on the rhel box?
15:55 davidfischer joined #salt
15:55 cachedout joined #salt
16:02 bhosmer_ joined #salt
16:06 bastion2202 joined #salt
16:07 jsm joined #salt
16:07 Marion joined #salt
16:10 oz_akan_ joined #salt
16:12 japage joined #salt
16:13 jaustinpage joined #salt
16:17 utahcon so, I run pillar.items php_error_reporting and all the minions report the right value
16:18 utahcon then I run the state and I get an error undefined jinja variable
16:18 utahcon error_reporting = {{ salt['pillar.get']('php_error_reporting') }}
16:18 amahon joined #salt
16:19 seanz left #salt
16:22 xinkeT joined #salt
16:23 mike251 guys i have a problem:   http://pastebin.com/ERzT0Z4V    the ... salt state return true when it runs the CMD, BUT it is not actually doing anything..... if i run the command manually on the minon, the /etc/openldap/ldap.conf gets update with the parameters from the command line
16:23 snave joined #salt
16:24 forresta utahcon, can you try {{ pillar['php_error_reporting'] }} if it's a top level item?
16:24 utahcon forresta: any idea why what I have would fail?
16:25 forresta utahcon, it's top level right?
16:25 forresta so it's not like php: \n  php_error_reporting right?
16:25 forresta in the pillar
16:27 redondos joined #salt
16:27 forresta mike251, can yo utry to run it again with -l debug to see if anything odd crops up?
16:30 bt joined #salt
16:31 mike251 -l where? the state.sls command?
16:32 forresta yep
16:32 mike251 ok, one sec please
16:32 forresta when you run the state.sls command and it fails out.
16:33 Brew joined #salt
16:33 nmistry joined #salt
16:33 backjlack joined #salt
16:33 renoirb Hey Salt people, is there a salt-formula for generating SSL certificates?
16:34 forresta renoirb, not that I've seen, that would be pretty cool though
16:34 renoirb forresta: done! https://gist.github.com/renoirb/7728455
16:35 forresta awesome
16:35 renoirb It was intended for MySQL SSL connection, but just realized that most of the meat there is about openssl
16:35 forresta yea
16:35 gadams999 joined #salt
16:36 Morton joined #salt
16:36 renoirb Ok, I'll make a saltstack-formulas for it and will give you the URL when ready
16:36 bemehow joined #salt
16:36 mike251 joined #salt
16:37 forresta mike251, can you post that on pastebin or something instead, then just link it in here? Helps that it's visible to everyone
16:37 mike251 ok sorry forresta
16:37 forresta renoirb, cool, make sure you talk to whiteinge when you're ready, he can either add you to the org or merge it in I think.
16:37 forresta mike251 no worries, just helps to have more eyes on it
16:37 sandGorgon joined #salt
16:39 mike251 http://pastebin.com/ndG2PaUD
16:39 dave_den joined #salt
16:39 jsm joined #salt
16:40 mike251 forresta:  i have duplicated the state ... and instead of using the cmd.wait ... i used a file. manage ... and i have created a SH file that i upload to minion via salt.. and run it with cmd.run ... and it works... but using cmd.wait... and the command directly ... it doesn't
16:40 forresta interesting
16:40 mike251 could there be a problem... that the command has many parameters ..!??! ... i am just asking...
16:41 mapu joined #salt
16:41 forresta I'm not sure, I'm wondering more if it has to do with the variables getting dropped in that aren't in quotes.
16:41 utahcon forresta: https://gist.github.com/anonymous/9a5b49b987d7202941ae
16:41 utahcon that is my pillar
16:42 mike251 the ... jinja variables are correctly replaced..
16:43 forresta yea, I mean I'm wondering if there are issues interpreting it via the command line for some reason, because commas or something. Can you try wrapping the jinja variables in single quotes?
16:44 forresta utahcon, ok yea so it's top level, so that should be working fine I'd think. Did you try the other way around that I pasted earlier to see if that one works?
16:44 mike251 forresta: ... pffff
16:45 jsm joined #salt
16:45 duncan0 joined #salt
16:45 forresta :P worth a shot
16:46 utahcon forresta: failed even with your change
16:47 utahcon same error (except the variable is different)
16:47 gadams999 Hi, I have a question regarding tenantName for salt-cloud. Our setup has both usernames and tenantNames as numeric values. Appears that libcloud is not double-quoting them in the JSON request for auth. Should I be hitting up the libcloud folks on that or is here a good place?
16:48 forresta utahcon, huh that seems really odd. I wonder if it has to do with how the pillar data is being compiled. Can you actually call that item from the pillar on the minion when you query it?
16:49 utahcon forresta: salt '*' pillar.items php_error_reporting returns the expected result string
16:50 mike251 forresta: jinja variables are in single quotes in the jinja file... that is what you mean?
16:51 forresta mike251, no I meant when you try and run it as a cmd.wait
16:51 bhosmer_ joined #salt
16:52 gadams999 ahh - figured it out. If I double quote the user and tenantname in the cloudprovider conf file, it gets passed through as a string. Now working
16:52 mike251 forresta:  ok.. let me try
16:53 forresta utahcon, as stupid as this sounds, in your sls, can you try to add #!jinja|yaml to the top?
16:54 mapu joined #salt
16:54 druonysus joined #salt
16:55 mgw joined #salt
16:55 davet joined #salt
16:55 mike251 forresta:  nope... but ... there is a difference between cmd.run and cmd.wait.... IF i put the command in a SH file and run it with cmd.run ... it works and i get the correct stdout: Starting sssd:                  [  OK  ]     BUT when i run the command via cmd.wait... nothing
16:55 utahcon forresta: in the sls or the file being managed?
16:56 forresta utahcon, wherever you're calling that jinja variable
16:56 forresta so the file probably
16:56 sandGorgon joined #salt
16:56 jaustinpage_ joined #salt
16:56 forresta and don't do yaml if it's an actual managed file, sorry
16:57 pipps_ joined #salt
16:57 forresta utahcon, do you have for the file.managed, - template: jinja ?
16:57 forresta mike251, weird
16:57 utahcon forresta: still fails :(
16:57 mike251 i am a moron forresta ... i was using cmd.wait... to wait for a package... but the package was there.. so the command wasn't actually being RUN
16:57 forresta mike251, ahh makes sense.
16:58 forresta utahcon, ok so you have the call for the pillar data within the file itself right?
16:58 mike251 can i use cmd.run with require? does that make sense?
16:58 forresta sure
16:58 forresta or you can do watch_in.
17:02 tdillio joined #salt
17:02 dstanek_afk joined #salt
17:03 JulianGindi joined #salt
17:05 troyready joined #salt
17:05 KyleG joined #salt
17:05 KyleG joined #salt
17:06 mike251 thanks
17:07 quickdry21 joined #salt
17:09 quickdry21_ joined #salt
17:10 ravibhure joined #salt
17:14 druonysus joined #salt
17:14 druonysus joined #salt
17:14 zach forresta: when you're back let me know, got an idea to bounce off you
17:14 forresta I'm here
17:15 forresta as much as I can be while at work at least
17:15 ravibhure1 joined #salt
17:15 zach ah, cool, PM?
17:15 forresta yea sure
17:19 xmltok joined #salt
17:21 slav0nic joined #salt
17:23 utahcon forresta: https://gist.github.com/anonymous/c25974aeb8970870f20c
17:23 utahcon That is (part of) the sls and php.ini
17:25 cachedout joined #salt
17:25 forresta utahcon, so do the other pillar variables you are pulling in work?
17:25 forresta for nfs_mount and so on?
17:26 mapu left #salt
17:26 utahcon they seem to yes
17:27 forresta so I notice one thing utahcon
17:27 forresta https://gist.github.com/anonymous/9a5b49b987d7202941ae
17:27 forresta on that you have for pillar_cr_init.sls
17:27 forresta you have error_reporting
17:28 forresta but for pillar_cr_test.sls you have php_error_reporting
17:28 pipps_ joined #salt
17:28 forresta are you pulling it from the init pillar instead of the test pillar?
17:28 snave joined #salt
17:28 forresta it's the only one that has a different name
17:28 utahcon oh yeah, noticed that and changed it to be consistent, didn't change the result :(
17:28 forresta lame
17:28 utahcon yeah
17:29 dstanek_afk joined #salt
17:31 forresta ok utahcon, can you try chopping everything from php5_php.ini except for error_reporting = {{ var }} ?
17:31 forresta make a backup or whatever.
17:32 forresta then apply just that php.sls with -l debug to see if we get any additional details
17:32 gasbakid joined #salt
17:33 utahcon forresta: ah damn, that worked... now I gotta figure out what is tripping it up :(
17:33 utahcon thanks
17:34 forresta yea np, I'd start by adding everything but session.save_path back, I wonder if the quotes on that are causing a problem maybe?
17:35 utahcon totally on the same page as you
17:35 JesseC joined #salt
17:37 druonysus joined #salt
17:37 forresta utahcon, cool, let me know what works
17:37 * mike251 enough for today... time to go home :)
17:37 * mike251 have a great day guys!
17:37 forresta later mike251
17:37 mike251 left #salt
17:37 mgw joined #salt
17:39 jaustinpage joined #salt
17:40 renoirb I was wondering, in a macro.  If I set {% macro macroname(variableName=None) %}
17:40 renoirb How can I test if it is empty or not in the macro?
17:40 renoirb {% if variableName is defined %} ?
17:41 ntt__ joined #salt
17:41 renoirb defined in Jinja is about if it has something... None, I assume, is although a falsy value... still something.
17:41 gasbakid_ joined #salt
17:42 ntt__ Hi. I'm a newbie and i would like to know if I can set network ip address in minions with salt.... Someone can help me? Thanks
17:42 dpippenger joined #salt
17:43 forresta ntt__, yes you can use the network state: http://docs.saltstack.com/ref/states/all/salt.states.network.html
17:44 forresta granted that's for rhel based distros
17:44 forresta if you wanted to manage that stuff on other systems, I imagine that a simple file.managed follow by a networking service restart would work as well.
17:46 ntt__ forresta: Thank You. I'm using centos 6.4, so this is a good notice
17:48 bhosmer_ joined #salt
17:49 rubberne_ How can I get the salt master to log the output that is displayed on the console when changes are made? I've turned up the logging to trace, but no luck....
17:50 zandy joined #salt
17:50 rubberne_ All I really want is to log the diff output of changes made to files. But logging everything is good too.
17:51 mgw joined #salt
17:51 ntt__ forresta: with the network module i have to create a .sls file whit the hostname on which i would configure network?
17:52 zandy_ joined #salt
17:52 UtahDave joined #salt
17:57 forresta ntt__, the network module is different than the network state, this is the network module: http://docs.saltstack.com/ref/modules/all/salt.modules.network.html
17:58 ntt__ but i should use the network states, right?
17:58 mgw joined #salt
17:58 anuvrat joined #salt
18:00 forresta rubberne_, uhh I don't know if that data gets logged anywhere, someone a while back was asking the same thing, and they ended up just outputting the data to a file then parsing it.
18:02 rubberne_ So if you use salt to manage files, and you happen to overwrite a file on a server by mistake. There is no logging set up to assist in recovery of the file? Is there a way to make the minions keep a version of the original managed file?
18:02 forresta rubberne_, http://docs.saltstack.com/ref/states/backup_mode.html
18:03 rubberne_ forresta: Exactly what I was looking for. I knew there had to be something for this built in. Thanks.
18:03 forresta yea np
18:07 jeter--- joined #salt
18:10 utahcon forresta: so, it turns out the issue was the line previous was not defined...
18:10 jeter--- left #salt
18:10 utahcon I updated that line and all works
18:10 forresta ahh duh, that totally makes sense
18:11 forresta very cool
18:13 utahcon thanks for the help
18:13 forresta yea np, wish I had spotted that quicker :P
18:14 backjlack joined #salt
18:15 polaco left #salt
18:15 kermit joined #salt
18:15 pipps_ joined #salt
18:16 kermit joined #salt
18:22 kermit joined #salt
18:26 forresta jeez terminalmage, you think you can spam that issue a little more :P
18:27 forresta I'm joking if that wasn't obvious, logging ftw
18:28 terminalmage oh the ec2 one?
18:29 terminalmage yeah I can post there a few more times :)
18:30 forresta hah
18:30 forresta yea the ec2 one
18:30 cdcalef joined #salt
18:30 MTecknology I still can't figure out why this pillar data isn't being applied
18:31 forresta MTecknology, I was going to say something about that, but now I can't remember what it was :\
18:31 JulianGindi joined #salt
18:31 MTecknology I can't remember what it was either
18:32 MTecknology what the hell?
18:32 MTecknology http://dpaste.com/1490455/ <-- same issue
18:32 forresta lol
18:33 jdenning joined #salt
18:34 MTecknology Why the crap???
18:34 schristensen joined #salt
18:34 dave_den MTecknology: are the packages key/vals supposed to be part of a parent dict?
18:35 MTecknology no
18:35 MTecknology dave_den: this is the original issue - http://dpaste.com/1490325/
18:35 JasonSwindle left #salt
18:36 jaustinpage joined #salt
18:36 MTecknology rhel based boxes aren't getting that pillar data and I have zero clue why
18:36 jaustinpage joined #salt
18:37 MTecknology master is 0.17.2 and minions are 0.17.1
18:37 dave_den ah
18:37 MTecknology I didn't read anything about this happening between those versions
18:37 dave_den does it still do the same thing if the debian check comes first in the if?
18:38 MTecknology same thing
18:39 dave_den funky
18:39 MTecknology I'm upgrading salt on one of them to make sure that's not the issue
18:39 MTecknology same thing
18:40 MTecknology http://dpaste.com/1490455/ <-- this one not working is what's really confusing the crap out of me, though.
18:40 * Gareth waves
18:41 MTecknology heck... if I remove the conditionals entirely it still has this issue
18:41 MTecknology It's like it's not rendering that sls at all for rhel boxes
18:41 forresta wait so if you just have that second block, it just doesn't get populated??
18:41 dave_den are you sure the pillar is being targeted at all of the minions as expected?
18:42 amckinley joined #salt
18:42 hhatfield joined #salt
18:42 MTecknology forresta: right... if I have just   "vi: vim-enhanced" as the only thing in the file, same thing
18:42 forresta lame, dave_den has a good point with the targeting
18:43 MTecknology http://dpaste.com/1490457/
18:43 dave_den MTecknology: if you set your master logging to debug you should be able to see the rendered pillar data for the minions in the master log. might help to look
18:44 dave_den but getting them on the same version would be a good thing anyway
18:45 MTecknology I got one on the same version and the same thing happened
18:46 Geo23589 joined #salt
18:46 ravibhure1 left #salt
18:46 MTecknology I wonder how much stuff will be flying by the screen with salt-master -l debug with all the other servers
18:47 forresta just run that single state
18:48 MTecknology http://dpaste.com/1490469/
18:49 MTecknology The state won't render because {{ pillar['vi'] }} is undefined. It's undefined because of something with the pillar data not being set for rhel boxes
18:49 MTecknology running salt-master in debug isn't helping me any
18:51 anon___ joined #salt
18:52 zandy joined #salt
18:53 MTecknology dave_den: any other debugging ideas?
18:54 dave_den other than checking the minion log? hrm..
18:54 dave_den how long ago was this pkgs.sls pillar added?
18:54 dave_den was it working before?
18:55 MTecknology I've never had this working for rhel boxes. I did everything with debian in mind and I'm trying to get these states applying to rhel boxes too
18:56 whiskybar joined #salt
18:58 nmistry joined #salt
18:59 Brew joined #salt
19:00 amckinley joined #salt
19:01 MTecknology as lovely as salt is, it can be a real thorn sometimes
19:03 MTecknology apparently I also have a minion that's refusing to connect to the master
19:06 dave_den i would remove the pillar variable from the state, and just get it working where the minion sees the proper pillar data
19:07 dave_den is the rest of the pillar rendered properly when doing pillar.data from the rhel minion
19:07 Geo23589 joined #salt
19:07 dave_den ?
19:07 MTecknology no pillar data at all
19:07 JulianGindi joined #salt
19:08 dave_den do you have master_opts enabled in the master config?
19:09 MTecknology nope
19:10 MTecknology would it help or cause issues?
19:10 dave_den do you see an error in the master log when doing pillar.data from the rhel minion?
19:11 dave_den MTecknology: if you have mster_opts enabled, it just puts some info in all minion pillars about the master config. it would help us to know that at least the minion was getting *something* in it's pillar that didn't come from the pillar sls files
19:11 MTecknology ;nope
19:11 dave_den it sounds like something else is preventing the master from compiling the pillar dict
19:12 MTecknology dave_den: so... master_opts: True in the master config, restart master, and restart minion and try pillar.data again?
19:13 dave_den MTecknology: sorry,:   pillar_opts: True
19:13 dave_den check for that
19:13 dstanek_afk joined #salt
19:13 \ask joined #salt
19:14 redondos joined #salt
19:14 MTecknology dave_den: the minion still has no pillar data set
19:14 ajw0100 joined #salt
19:15 dave_den there should be an error in the master log. pretty sure the process that is trying to compile the rhel pillar data is dying
19:15 dave_den when you get an empty pillar it's usually because of a bug or compile issue
19:16 MTecknology lemme comment out everything so that at minimum the master stuff should be returned
19:17 MTecknology I commented out every pillar sls that would apply to that box and pillar.data is still empty
19:18 pentabular joined #salt
19:18 dave_den MTecknology: did the master log at least say it "rendered pillar data for <your minion>:  blah blah your pillar data" or something similar?
19:18 elfixit joined #salt
19:18 matanya joined #salt
19:19 zandy joined #salt
19:22 dave_den let me double check that
19:23 matanya joined #salt
19:23 MTecknology dave_den: on a debian box, I see this on the master      [DEBUG   ] Rendered data from file: /srv/pillar/top.sls:
19:23 MTecknology on an rhel box - nope
19:24 dave_den can you even do test.ping from the master to the rhel minion?
19:25 dave_den did you say you checked the minion's logs?
19:26 MTecknology ya, I can do test.ping from the master
19:26 fllr joined #salt
19:26 MTecknology damnit
19:27 MTecknology wait... no, that's just the render error
19:27 MTecknology nothing useful in the logs
19:27 MTecknology I'm running the minion in debug
19:28 MTecknology running 'salt-minion -l debug' and then 'salt-call pillar.data' shows nothing in the log
19:28 dave_den salt-call is independent from salt-minion.
19:29 MTecknology and running 'salt 'mailarchiva1*' pillar.data from the master makes stuff show up in there, but no errors
19:29 dave_den salt-minion will only log things initiated from the salt master in the logs. you need to use 'salt-call -l debug ...'
19:29 Guest42390 joined #salt
19:30 MTecknology nothing useful with 'salt-call -l debug pillar.data' either... :(
19:31 dave_den MTecknology: delete the cache for that minion on the master, in /var/cache/salt/master/minions/<rhel minion id>
19:31 Guest11270 joined #salt
19:31 dave_den then try to do pillar.data again and see if data.p shows up in /var/cache/salt/master/minions/<rhel minion id>
19:33 scott_w joined #salt
19:33 MTecknology I deleted that whole directory and it's not being recreated
19:33 Guest11270 joined #salt
19:34 MTecknology not recreated after restarting the minion and master either
19:35 MTecknology and manually creating the directory doesn't help that stuff get created
19:36 MTecknology dave_den: on a debian box, it's recreated just fine
19:37 druonysus joined #salt
19:37 druonysus joined #salt
19:37 dave_den ok, that confirms that the master process is dying when trying to compile the pillar data for the rhel box.
19:37 MTecknology hm?...
19:38 dave_den the first step i would take is to remove all targeting of rhel minions until you can get data.p to show up when doing pillar.data (try to do an empty pillar)
19:38 MTecknology an empty pillar?
19:38 dave_den data.p on the master holds the cached grains and pillar data for that minion
19:39 dave_den if the master is able to complete the pillar compilation for that minion, even with an empty pillar then data.p will exist on the master with just the minions grains
19:41 ajw0100 joined #salt
19:41 MTecknology http://dpaste.com/1490510/ <-- I did that, no data.p exists
19:42 MTecknology grains.items returns what I expect, but pillar.data is empty; so.... the master is failing to render the pillar data and not throwing an error?
19:42 dave_den doing that will also cause problems if there is no empty/init.sls file....
19:43 bhosmer_ joined #salt
19:47 MTecknology there's empty.sls
19:47 MTecknology I need empty/init.sls instead?
19:48 cachedout joined #salt
19:49 dave_den no, empty.sls will work
19:53 MTecknology still no data.p created
19:53 ggoZ1 joined #salt
19:53 dave_den on your minion:   salt-call —versions-report
19:54 mgw joined #salt
19:55 druonysus joined #salt
19:55 MTecknology dave_den: http://dpaste.com/1490526/
19:56 dave_den upgrade your ZMQ version to rule that out.
19:56 dave_den you need ZMQ 3.x
19:57 JordanRinke Morning.
19:57 MTecknology have no idea how to get that upgraded on this box...
19:58 MTecknology it doesn't seem to be available in epel or epel-testing
19:58 dave_den it is
19:58 dave_den yum --enablerepo=epel-testing install -y PyYAML libyaml m2crypto openpgm python-babel python-crypto python-jinja2 python-msgpack python-zmq sshpass zeromq3
19:59 dave_den those are all of the salt dependencies
20:00 MTecknology No package openpgm available.  No package python-zmq available.  No package zeromq3 available.
20:00 MTecknology the rest were found
20:00 MTecknology this is rhel 5 btw...
20:02 dave_den ah, i don't have any rhel5 minions.
20:02 dave_den http://zeromq.org/distro:centos
20:04 MTecknology yup... now a bunch of dependency issues
20:08 mgw1 joined #salt
20:08 MTecknology http://dpaste.com/1490531/ ... fun
20:09 kyusan_ joined #salt
20:09 MTecknology so... remove python26-zmq, then upgrade with the new repo, then figure out how to get that back
20:10 MTecknology FFS!!1
20:10 MTecknology I can't install salt via easy_install either
20:11 MTecknology I FUCKING HATE REDHAT!!!
20:11 dave_den https://groups.google.com/forum/#!topic/salt-users/c9p49ZUczJo  this guy did it
20:11 opapo joined #salt
20:11 dave_den or you can follow the steps to build your own RPMs...
20:12 dave_den hopefully this is the problem, otherwise you'll cut me
20:12 dave_den :D
20:12 forresta you don't hate redhat MTecknology, you love redhat
20:12 forresta you love being 6 releases behind in software
20:12 JesseC joined #salt
20:13 MTecknology dave_den: I will not be building my own rpm
20:13 MTecknology forresta: sure... it's great if you're looking for a reason to quit IT and become a park ranger
20:13 forresta hah
20:14 MTecknology dave_den: it sounds like the only issue is connections timing out after a week
20:14 dcolish joined #salt
20:15 dave_den try a rhel6 minion for testing, see if that works
20:16 MTecknology I'd rather just burn all our rhel boxes to the ground
20:16 jesusaurus whats the expected behaviour of having two formulae with the same name?
20:17 MTecknology 1.5% of the servers I manage are rhel and they are a constant bloody thorn in my side
20:17 jesusaurus if i give file_roots two paths that both contain a 'mystuff' directory, which should it use?
20:17 dave_den jesusaurus: the first one listed.
20:18 forresta man dave_den, answering the questions today
20:18 jesusaurus sure about that?
20:18 jesusaurus did that recently change? (between 0.17.1 and 0.17.2)
20:19 MTecknology I'm really wondering if I'm seeing a bug in salt.
20:20 dave_den jesusaurus: it was in the documentation, IIRC
20:20 tseNkiN joined #salt
20:20 zandy joined #salt
20:20 pentabular1 joined #salt
20:21 dave_den perhaps that was only for fileserver_backend, but it should be in order for file_roots, too
20:21 forresta dave_den, if you find the docs let me know, I don't see it in the main config stuff.
20:21 dave_den let me look at the code. sec
20:22 jesusaurus what im seeing is the sls files are served from the first path, but salt:// paths point into the second path
20:23 dave_den you mean when using multiple sources in a state function?
20:24 jesusaurus lemme sketch up a base example
20:25 jsm joined #salt
20:32 dave_den http://docs.saltstack.com/topics/tutorials/states_pt4.html#salt-fileserver-path-inheritance
20:32 hhatfield joined #salt
20:32 dave_den "Salt's fileserver collapses the list of root directories into a single virtual environment containing all files from each root. If the same file exists at the same relative path in more than one root, then the top-most match "wins". For example, if /srv/salt/foo.txt and /mnt/salt-nfs/base/foo.txt both exist, then salt://foo.txt will point to /srv/salt/foo.txt."
20:33 forresta awesome dave_den, thanks!
20:33 copelco joined #salt
20:33 asdfasdfdf joined #salt
20:33 MTecknology dave_den: I don't suppose you could try installing an rhel 5 box and seeing if you have the same issue
20:34 dave_den MTecknology: sure, hold on a minute
20:34 dave_den MTecknology: your master version is 0.17.2?
20:35 MTecknology yup, master and client are
20:35 dave_den k
20:35 dave_den and your rhel minion was just installed through epel-testing, yeah?
20:35 MTecknology ya
20:35 dave_den cool. gimme a minute
20:36 MTecknology it takes me a long time to get a box deployed since the vmware admin is a clueless dipshit
20:37 forresta so angry
20:37 dave_den MTecknology: i386 or x86_64?
20:37 MTecknology 64
20:38 MTecknology forresta: that term is only lightly scratching the surface
20:38 MTecknology forresta: for reference... I'm looking at a new job right now
20:38 forresta yea
20:39 forresta I remember you saying that
20:39 drags so, running 0.16.3, and will upgrade promptly, but has anyone seen issues where the minion gets into an odd state and manages to launch ~240 copies of itself?
20:39 forresta drags, yea that's a bug
20:39 forresta zach can tel you all about it :P
20:39 scoates I have a state that patches one of the Minion's salt modules. Is there a clever way to get that to both 1) run first, and 2) restart that minion before applying the rest of the states?
20:39 forresta *tell
20:39 dave_den drags: yes, older minions can spawn do that when unable to reach the master, if they are using the scheduler
20:39 drags forresta: is it in the current release?
20:39 MTecknology Our AIX admins handle 50 servers each and do only some app admin work, our Windows admins handle about 70 severs each and do no app admin work. I manage 470 servers by myself and manage all the apps on them.
20:39 drags dave_den: ahh
20:39 forresta drags, no it was fixed for 0.17.2
20:40 drags yeah I recently moved my master
20:40 forresta or at least, I haven't seen anyone having that issue with this release.
20:40 forresta oh, maybe I am thinking of the wrong bug then drags
20:40 forresta dave_den, do you remember if that got fixed? I thought it did...
20:40 dave_den drags: if you upgrade your master to >= 0.17, your < -.17 minion will not be able to talk to it
20:40 dave_den forresta: yes, it's fixed in develop
20:41 Ryan_Lane joined #salt
20:41 forresta dave_den, but was it fixed in any of the actual releases
20:41 drags btw, master moving: rsync /etc/salt to new box, update minion configs, blammo
20:41 dave_den for the scheduler it now has an option of 'max_running'
20:41 dave_den forresta: not sure. it was a recent fix a week or two ago.
20:41 forresta probably isn't in then
20:41 forresta my apologies drags, I was thinking of another similar bug
20:42 drags dave_den: is "upgrading all minions to 17.x then the master to .17.x" a safe upgrade path?
20:42 MTecknology forresta: I also manage our web filtering and web access systems and the systems that log all network activity and web traffic and do a LOT with keeping our email infrastructure up and running and end up helping the windows/vmware and aix guys figure out how to make things work.
20:42 [diecast] joined #salt
20:42 forresta That's what happens at a small shop when you know what you're doing!
20:43 forresta drags, yes that is the common upgrade method that a lot of people are using
20:43 dave_den drags: i have not tested it myself, but it should work if you do minions first. please test with an unimportant minion first. :)
20:43 MTecknology we have 28,000 employees
20:43 forresta lol
20:43 MTecknology we're about to hire another windows admin
20:44 SEJeff_work MTecknology, How many servers?
20:44 SEJeff_work with that many employees?
20:44 MTecknology SEJeff_work: total servers?
20:45 SEJeff_work sure
20:45 SEJeff_work with a breakdown (roughly) by os
20:45 MTecknology SEJeff_work: the backlog has that...
20:46 MTecknology SEJeff_work: sorta... around 800 servers, around 500 are linux, I'm the only linux admin, 28,000 total employees
20:46 SEJeff_work Ah wow
20:46 SEJeff_work My company is basically the inverse
20:46 SEJeff_work Not a ton of employees, a 3-400, but thousands of servers
20:47 SEJeff_work *about 3-400 employees
20:48 MTecknology I'm looking at a different company. I'd probably get 2x my pay here for a starting salary
20:49 forresta unless you take a job in california, then you get to make 6 figures of sunshine dollars
20:50 MTecknology I don't want to work there.
20:50 cyrusdavid joined #salt
20:50 amahon joined #salt
20:52 dave_den yum mirror being slow… derp
20:52 rojem joined #salt
20:53 Gifflen joined #salt
20:54 ajw0100 joined #salt
20:56 MTecknology ya... I'm going to start seeing what's available.
21:00 dave_den MTecknology: rhel5 minion works with a simple pillar file using zmq2
21:00 zach drags: oh boy. you got that bug too? :-)
21:01 dave_den did you want me to test something specific?
21:02 zach drags: It was an enjoyable bug I ran into that started at 4am and ended up taking our production environment down around 8am, causing me headaches from people trying to fix it that had no idea what they were doing
21:02 zach namely my boss
21:02 zach he started killing random daemons to "fix it"
21:02 forresta zach, his issue is actually a different, but similar one
21:03 forresta I thought it was the same bug but it's unique to a minion that can't communicate with the master
21:03 snave joined #salt
21:03 zach same thing
21:03 forresta which is already fixed in develop
21:03 forresta oh was it that?
21:03 zach yep
21:03 zach I hit that and another
21:03 forresta I thought yours was the master hook up
21:03 forresta ahh ok
21:03 zach both
21:03 forresta you had multiple :P
21:03 forresta lol
21:03 Sheco joined #salt
21:03 zach it was two bugs at once that brought us down
21:03 forresta that makes me feel as though my brain isn't getting ruined quite as quickly. phew
21:04 zach but still, the bugs that I've ran into...far less stressful than maintaining chef/puppet/cfengine
21:04 ajw0100 joined #salt
21:06 zach I hate all three with a passion :\
21:06 forresta it's fair
21:06 forresta nice dudes though
21:07 zach I still deploy saltstack for every one of my side clients too
21:07 forresta lol
21:07 zach I deployed it for a rather large festival here for ticket sales
21:07 forresta yea, easy to scale up that way huh?
21:07 zach yep
21:07 zach shame he had his own ideas of how to do it
21:08 zach he broke my config and did it his own way and killed off my contract cause he did not understand why I used salt and not chef
21:08 zach he is a buzzwords guy
21:08 zach "OMG chef! Thats the way to do it, everyone is doing it these days!"
21:08 harobed_ joined #salt
21:08 oz_akan_ joined #salt
21:08 forresta I'm honestly surprised how many startups are looking at chef
21:08 vkurup joined #salt
21:08 forresta blows my mind
21:09 zach the ease of use that salt has, the modular approach to where you can easily write your own modules, a monkey could maintain salt if need be
21:09 forresta Yup
21:09 cewood joined #salt
21:09 zach most startups need that too, since everyone has 12 titles
21:10 oz_akan_ joined #salt
21:10 bhosmer_ joined #salt
21:10 forresta who knows man
21:10 zach I have 4 titles at this place, and I'm expected to be a god at all 4 areas
21:10 zach keeps me busy
21:10 forresta I talked to one dude who was surprised when I explained I wasn't an expert in chef, and that most sysadmins I knew weren't making it their first choice for config management
21:11 zach I had interviewed with ARM, they use Puppet there. They asked me my feelings about Puppet, and I told them "So you have a full time person maintaining Puppet right?" ;-)
21:11 forresta heh
21:11 zach I didnt get the job because I was a smartass
21:12 brianhicks_ joined #salt
21:12 forresta MTecknology, did you go take a walk? Do some meditating?
21:12 zach I'm going to write up my approach to salt states that I showed you in PM forresta - I think that'll help out a lot of new people to organize their states so they can be managed
21:13 pipps joined #salt
21:13 forresta zach, http://docs.saltstack.com/topics/tutorials/best_practices.html ? :P
21:13 patrek_ joined #salt
21:13 zach gasp. that is a 404
21:14 forresta exactly
21:14 forresta needs to be fixed!
21:14 zach what is the markup of that doc system?
21:14 forresta zach, it's just rst
21:14 zach I'll write it up for that so it can easily be copy pasted
21:14 forresta easy stuff
21:14 forresta just make sure you stick to pep 8 standards regarding line length
21:14 zach Rosetta Stone Inc?
21:14 forresta RST
21:14 forresta :P
21:15 forresta restructured text
21:15 zach That is the ticker symbol for Rosetta Stone Inc
21:15 forresta if you look at any of the other topics/tutorials there are great examples
21:15 mpanetta Hmm if that is rosetta stone, who is SYN? heh
21:16 forresta synthetic biologics inc?
21:16 forresta whatever that is
21:16 zach roger
21:16 forresta their stock is only worth 1.25 though
21:16 zach what software is in use for docs.saltstack.org - I've seen it quite a few times
21:16 forresta it's sphinx
21:17 zach perfect
21:17 zach beats wikimedia
21:17 forresta sphinx builds the html pages, which are then built for the docs site, and pushed to read the docs.
21:17 zach or wahtever it is
21:17 forresta mediawiki
21:17 forresta *?
21:17 zach thats it
21:17 rojem joined #salt
21:17 forresta yea I would not be contributing docs if it was mediawiki
21:18 ajw0100_ joined #salt
21:18 zach I wrote some really ugly newbie documentation on our deployment on mediawiki
21:18 zach when I was first getting started
21:20 zandy joined #salt
21:22 forresta hey UtahDave, did you ever finish testing this: https://github.com/saltstack/salt/issues/8701 ?
21:22 \ask joined #salt
21:22 mirko joined #salt
21:23 scott_w joined #salt
21:24 cdcalef joined #salt
21:27 quickdry21_ joined #salt
21:28 whiskybar joined #salt
21:29 wramthun joined #salt
21:30 scott_w joined #salt
21:30 \ask joined #salt
21:30 mirko joined #salt
21:30 bhosmer_ joined #salt
21:31 zach What is a usecase for the mysql returner? I'm trying to wrap my head around the purpose
21:32 mannyt_ joined #salt
21:32 UtahDave forresta: When I was testing against the develop branch it seems to work correctly
21:33 UtahDave zach: The mysql returner causes each minion to send its results directly to a mysql database
21:33 KyleG1 joined #salt
21:33 zach UtahDave: Neat, what would the purpose of that be? Archiving/Reporting?
21:34 zach I wonder how difficult it would be to send to Splunk
21:35 forresta UtahDave, ok cool.
21:35 UtahDave zach: yeah, exactly.  Some people use it for archiving, and also  let other internal web applications access that return data to find out execution histories and other things from that dat
21:35 UtahDave data
21:35 zach Makes sense
21:36 zach Maybe I'll try it for a bit and see how well it works
21:36 UtahDave zach: I don't think it would be very hard at all to make a splunk returner.  right now there's a bunch of returners.   sentry, syslog, carbon, etc
21:36 zach There is a syslog one?
21:36 zach does it send local or remote?
21:37 UtahDave https://github.com/saltstack/salt/tree/develop/salt/returners
21:37 zach If it can send remote, then perfect - splunk will already work!
21:37 zach ah, local only
21:37 zach Not a problem though, cause all of our syslog entries already go to splunk - perfect
21:38 UtahDave I'm guessing that the sentry one would be somewhat similar to how a splunk one should work.
21:38 zach I assume that goes into the minon config?
21:39 UtahDave either in the minion config or the master config
21:39 UtahDave if you put it in the master config it will propogate out to all the minions automatically
21:39 zach cool, that's great
21:40 zach thank you
21:41 william_20111 joined #salt
21:41 Ryan_Lane joined #salt
21:42 teebes joined #salt
21:42 Brew joined #salt
21:44 drags zach: I think Chef's continuing popularity has to do with Ruby's continuing popularity
21:45 aleszoulek joined #salt
21:45 william_20111 joined #salt
21:47 MTecknology forresta: ya, I did
21:50 __number5__ and big money throwing behind it
21:51 zach drags: yeah. I think so too
21:52 zandy joined #salt
21:54 teepark joined #salt
21:54 JulianGindi joined #salt
21:55 erchn joined #salt
21:55 MTecknology forresta: I blew off a bit of steam and I think it helped a lot
21:56 moos3 joined #salt
21:56 erchn hi all, any rackspace salt-cloud users in the house?
21:56 erchn hopefully with rackconnect experience!
21:56 amckinley joined #salt
21:56 amckinley1 joined #salt
21:57 nliadm getting an odd trace when trying to run salt-call: http://pastebin.com/nwYWnPzC
21:57 nliadm this is on an eksy build
21:58 nliadm there is, as far as i can tell, nothing different on this system from others where this works fine
21:58 erchn was hoping mr. hoser of the perl would be in the house today
21:58 dave_den is it the same architecture?
21:58 nliadm yes, they're hardware identical
21:58 nliadm OS also
22:00 cdcalef joined #salt
22:01 penguin_dan joined #salt
22:01 erchn nliadm: windows?
22:01 nliadm no, linux
22:02 nliadm it still fails without any other minions running
22:02 shinylasers joined #salt
22:02 erchn here on windows clearing the cache on the minion seemed to help
22:02 erchn https://github.com/saltstack/salt/issues/7954
22:02 nliadm I'll give that a shot
22:02 erchn OP did it manually, not via salt
22:02 nliadm right
22:03 shinylasers joined #salt
22:04 nliadm erchn: thanks, that seems to have worked
22:05 nliadm nfc why, though
22:05 nliadm :/
22:05 erchn cool, not sure either...good to hear
22:06 erchn might be interesting to see if this happens again to md5sum the directories contents and then run a compare after clearing and rerunning
22:06 KyleG joined #salt
22:06 KyleG joined #salt
22:06 nliadm I need to build a 0.17 freeze
22:06 nliadm the problem I was having before was that 'site' wasn't getting pulled in
22:06 nliadm from a virtualenv
22:07 nahamu nliadm: are you running on Linux?
22:07 jsm joined #salt
22:07 nliadm yes
22:07 nahamu sorry, just read more scrollback
22:07 nahamu I ended up just creating a virtualenv for my linux machines
22:08 nahamu I got esky to work on SmartOS, but it was giving me headaches on Linux...
22:08 \ask I'm working on getting our development environment to use a separate salt configuration than the production one. We use the gitfs backend.
22:08 \ask My plan was to have production use a "production
22:08 \ask "
22:08 nahamu (which is *not* what I would have expected....)
22:08 nliadm well, I'm freezing out of a virtualenv, which wasn't working with 0.17.1
22:08 \ask branch and/or development use a "dev" branch -- but reading the documentation that doesn't seem like it'd work
22:08 nahamu nliadm: did you have it working on something older?
22:09 nliadm \ask: the gitfs backend does not work the way you might think
22:09 \ask specifically: “Also, top.sls files from different branches will be merged into one big file at runtime. Since this could lead to hardly manageable configurations, the recommended setup is to have the top.sls file only in your master branch, and use environment-specific branches for states definitions.”
22:09 \ask nliadm: right, that's what I am learning :-
22:09 \ask :-)
22:09 \ask (sorry about all the extra lines, ugh -- stiff fingers today)
22:09 nliadm also, there was horiffic performance on the 0.16 series
22:09 nliadm nahamu: yeah, I have freezes for 0.13 through 0.16.4
22:10 nahamu nliadm: do you actually *need* an esky, or do you just need some way to get a working tarball of minion code you can drop into your linux machines?
22:10 nliadm I need a way to not have to install swig on every machine
22:10 nahamu right.
22:10 nahamu I think you can do that with a carefully built virtualenv.
22:11 nahamu though it would be nice for the esky build to work properly under linux.
22:11 nliadm I need to investigate, I reported it when I first tried with 0.17.0 and 0.17.1
22:12 nliadm might be fixed in 0.17.2
22:13 nliadm and/or I just need to build in a vm instead of in a virtualenv
22:14 nahamu nliadm: if you figure it out, I'd be interested.
22:15 nliadm will do. I'll try to post a copy of my makefile somewhere
22:16 nahamu In theory if you build the virtualenv correctly, you can tar it up and distribute it and you won't need swig everywhere.
22:16 nahamu not as elegant as an esky, but in a pinch it worked for me...
22:17 fllr joined #salt
22:21 nliadm has there been talk about using zmq curve to be able to ditch m2crypto?
22:25 elfixit joined #salt
22:25 UtahDave nliadm: yeah, Tom has been working with the zmq people on that.  There's still a lot of work to do
22:25 nliadm UtahDave: looks like esky is broken with pyzmq >=14
22:25 nliadm and the version thing is back
22:25 UtahDave really?  hm.
22:25 UtahDave which version thing?
22:27 nliadm http://pastebin.com/G3jh1mvv
22:27 nliadm something similar happened with the 0.16 series at one point
22:31 UtahDave Oh, that's back again?
22:31 UtahDave let me get someone on that.
22:33 pipps joined #salt
22:39 kermit joined #salt
22:41 logandg joined #salt
22:48 jacksontj joined #salt
22:50 altj joined #salt
22:50 zandy joined #salt
22:57 \ask What's the idiomatic way to keep production and development separate with gitfs jf bra]\\\\? Separate repositories?
22:58 \ask (eh, "if branches are not the way").
22:59 juicer2 joined #salt
23:01 frosty996 joined #salt
23:02 ipmb joined #salt
23:02 frosty996 hey guys, I got an interesting problem.  Whenever I use salt-cloud to initialize an instance, the first time I call state.highstate, I get a bunch of "No such file or directory: /var/cache/salt/minion/files/base" errors, but when I call state.highstate again it all works fine. :-/
23:03 Gareth anyone know if salt runs on the Debian lenny release/
23:07 forresta frosty996, are you running that off a masterless minion
23:07 UtahDave frosty996: are you running the highstate immediately after the instance comes up?  The minion needs a few seconds to finish all it's authentication
23:07 frosty996 forresta: no, it's being called from the master
23:07 UtahDave Gareth: yeah, it should
23:08 Gareth UtahDave: I guess my question should have been...is it supported :)
23:08 frosty996 UtahDave: well, I have a loop that does test.ping and calls the highstate after test.ping returns True
23:12 amckinley joined #salt
23:12 UtahDave joehh: Is Lenny supported by Salt?
23:14 hhatfield joined #salt
23:15 Gareth UtahDave: reason I ask...working on the networking support for Debian, when setting up bonded interfaces there was something that was required prior to the squeeze release.  Similar siutation between RH5 and RH6, writing a file into /etc/modprobe.d, but if Salt isn't supported on releases older than Squeeze then I don't need to bother with that.
23:16 joehh UtahDave: unfortunately not :(
23:17 joehh it has python2.5 and backporting 2.6 fell into the too hard basket
23:17 joehh Gareth: from your perspective though - fortunately not :)
23:17 Gareth joehh: indeed :)
23:18 Gareth joehh: thanks.
23:18 pipps_ joined #salt
23:20 UtahDave thanks, joehh!
23:23 Gifflen joined #salt
23:24 scott_w joined #salt
23:24 fllr joined #salt
23:30 joehh Gareth, UtahDave: bad news though is salt is currently "supported" back to ubuntu 10.04 and will be until April 2015
23:30 forresta joeeh, that's some maintenance dedication :P
23:31 joehh I don't know how it does networking compared to the others
23:31 Gareth joehh: hm. okay.  will check that one.
23:32 joehh forresta: tell me about it
23:32 UtahDave Gareth: let me know how that goes.
23:32 Gareth UtahDave: will do
23:32 oz_akan_ joined #salt
23:33 Ryan_Lane joined #salt
23:34 oz_akan_ joined #salt
23:35 aleszoulek joined #salt
23:36 N-Mi joined #salt
23:39 alunduil joined #salt
23:39 rojem joined #salt
23:42 frosty996 so that problem report above should have said, "The first time I call highstate *after* a reboot" I get the file not found error.  Actually, it seems that I can call it several times quickly and always get the error, but if I wait a little bit, it seems happy.  Gonna try it with a long sleep and see if that does the trick.
23:48 frosty996 even with a sleep 30, it still gives the file not found error
23:48 frosty996 but then when I run it again, it's succesful.
23:49 Drawsmcgraw joined #salt
23:50 zandy joined #salt
23:51 UtahDave frosty996: what version of Salt?
23:51 frosty996 UtahDave: 0.17.2
23:52 * frosty996 tries a sleep 60
23:52 ipmb joined #salt
23:54 frosty996 wonder if I add a saltutil.refresh_modules to it
23:54 UtahDave probably good to try
23:56 elfixit joined #salt
23:59 dpippenger joined #salt

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