Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-10-21

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

All times shown according to UTC.

Time Nick Message
00:01 __number5__ if you just want to generate sls/configs based on grains, you don't need pillar, you can use grains directly in your sls or jinja template http://docs.saltstack.com/ref/states/vars.html#grains
00:05 __number5__ MohShami: see my comments above
00:06 mwillhite joined #salt
00:08 MohShami __number5__, robawt, sorry guys was AFK for a little
00:08 MohShami I'll check the links, thanks a million :)
00:08 __number5__ np
00:09 MohShami but __number5__, I've used grains for a bit, but I can't figure a way to use them to loop through hosts in my config, do you have a template or a sample I can check?
00:11 __number5__ MohShami: do you mean collect minions ip addresses to generate a hosts file?
00:12 MohShami for example, yes
00:12 MohShami I can build on that to build nagios configuration files for example
00:14 __number5__ hmm, that's a bit different, let me check...
00:15 MohShami please, and thanks :)
00:19 m_george left #salt
00:26 __number5__ MohShami: the salt mine feature is better for your case, an example https://github.com/saltstack/salt/issues/5281#issuecomment-18570027
00:26 __number5__ docs: http://docs.saltstack.com/topics/mine/index.html
00:27 MohShami __number5__: Thanks a million mate
00:27 MohShami so I just use the mine and jijna, right?
00:28 __number5__ I didn't try that before, so you might have other issues
00:28 __number5__ yes, just mine and jinja
00:28 MohShami will check it out
00:28 MohShami thanks again mate :)
00:29 __number5__ need to set the mine_functions config at minions to allow access to network.interfaces
00:34 rgbkrk joined #salt
00:37 MohShami __number5__, will check that out, thanks again :)
01:02 justlooks joined #salt
01:03 justlooks if A is string i can write this in jinja " VAR == A ",what if A is integer?
01:10 jimallman joined #salt
01:14 mua joined #salt
01:31 deepakmd_oc joined #salt
01:32 __number5__ justlooks: you need to quote string like  var == "abc", integer or float don't need to quote: VAR == 42
01:34 justlooks __number5__: thanks
01:34 __number5__ np
01:36 mafrosis joined #salt
01:37 mafrosis hola rethink crew
01:38 mafrosis I've had what appears to be a couple of crashes from Rethink - is there something I can do about crash reporting?
01:39 mafrosis The log_file for my DB doesn't actually have any relevant crash-related content
01:39 mafrosis left #salt
01:39 mafrosis joined #salt
01:40 acsir Hi, I'm getting crazy! I'd like in a pillar file to select the variable value based upon the pillar.id value, but I don't get how it is supposed to write the if-condition
01:40 acsir by now I tryed to use: {% if {{ pillar['id'] }} == 'serverLego.KVM.UPS.casa' %}
01:40 acsir or {% if pillar['id'] == 'serverLego.KVM.UPS.casa' %}
01:41 mafrosis left #salt
01:41 mafrosis joined #salt
01:41 mafrosis left #salt
01:42 justlooks hi, if some service script is not in /etc/init.d dir ,how can i write service running state?
01:45 acsir but in both cases I get back, calling  "salt '*Lego*' pillar.items" , the following error: expected token ':', got '}'; line 1 in template (line 1 if the previous if)
01:45 isomorphic joined #salt
01:45 justlooks i install a dell tools ,syscfg ,it's service-start script not in /etc/init.d/ ,it's in /opt/dell/srvadmin/sbin/
01:51 Nexpro1 joined #salt
01:53 mua joined #salt
01:54 acsir wow, finally I get the correct format, here it is... {% if pillar.get('id') == 'serverLego.KVM.UPS.casa' %}
01:57 m0hit joined #salt
02:00 AdamSewell joined #salt
02:06 malinoff joined #salt
02:06 druonysuse joined #salt
02:09 StDiluted joined #salt
02:10 rgbkrk joined #salt
02:12 m0hit joined #salt
02:14 malinoff Hi salt guys
02:15 druonysuse joined #salt
02:15 druonysuse joined #salt
02:16 carmony hey :)
02:18 malinoff I hope salt v0.17.1 will be pushed into the debian repo asap :(
02:23 AdamSewell joined #salt
02:24 ajw0100 joined #salt
02:25 justlooks hi,anyone can help this ? https://gist.github.com/justlooks/7077831
02:27 carmony hrm, looks like it is failing to enable that service
02:27 carmony have you tried calling "salt-call state.highstate" on the minion directly?
02:27 carmony you likely will get something more descriptive in the error message mayge
02:27 acsir no, it is still not working. Please could some one have a look at this use of a conditional in a pillar.sls? http://pastebin.com/uf1yALkA
02:29 carmony acsir: have you validated that the pillar data is correctly being set for the minion?
02:30 carmony https://github.com/JustinCarmony/salt-troubleshoot#pillar-items -- some commands I find helpful for debugging pillar data
02:32 acsir yes, the id is correct both on minion and master
02:33 acsir if I query  --- salt '*Lego*' pillar.get id --- I get --- serverLego.KVM.UPS.casa: ---
02:33 xl1 joined #salt
02:33 carmony ok, let me check the jinja templating
02:33 acsir but the equality in the file is not working, I always get the else option
02:36 carmony hmmm
02:37 simonmcc joined #salt
02:37 carmony this -shouldn't- make a difference, but you could put a str() around the pillar.get() to make sure it is comparing strings
02:37 carmony but I'm pretty sure pillar.get() should return a string
02:40 acsir I tried to chenge it into: {% if pillar.get('id').str() == 'serverLego.KVM.UPS.casa:' %}
02:41 acsir but this way I get an error: Undefined jinja variable; line 1 in template
02:41 AdamSewell joined #salt
02:41 acsir could it be that it's not allowed to call any pillar within a pillar.sls config file?
02:46 AdamSewell joined #salt
02:49 __number5__ acsir: what does salt '*Lego*' pillar.items  return?
02:50 acsir I'm posting it now
02:50 carmony can you use pillar data in the pillar sls files?
02:50 carmony I wouldn't think so
02:50 acsir I just checked    --- {% if grains['os'] == 'Debian' %} ---   and in this case the conditional works
02:51 acsir carmony: so I could use just grains in pillar sls files?
02:53 acsir __number5__: here it is http://pastebin.com/nHGaJykf
02:54 carmony acsir: yes, grains are available
02:55 malinoff acsir, you should not rely on 'id' in pillars since this information can be disabled in the configuration file
02:55 malinoff grains is your choice
02:58 acsir thankyou both, just know I found this, exactly what you have just told me: pillar  does not have access to itself when rendering
02:58 acsir does not have access to itself when rendering
02:58 acsir https://groups.google.com/forum/#!topic/salt-users/gH7DHC0Ck88
02:58 acsir I'll go with grains...
02:59 ajw0100 joined #salt
03:17 berto- joined #salt
03:18 justlooks joined #salt
03:18 cshuman joined #salt
03:24 hotbox_ joined #salt
03:25 cyp joined #salt
03:25 marcinkuzminski joined #salt
03:26 marcinkuzminski joined #salt
03:26 Zethrok joined #salt
03:31 cshuman joined #salt
03:59 cewood joined #salt
04:02 yano joined #salt
04:09 nocturn joined #salt
04:33 redondos joined #salt
04:33 redondos joined #salt
04:43 joehh NV: I think the dependencies are now all available in regular debian backports so you should be able to get them from there
04:43 joehh NV: as for all architectures (mips included)
04:46 NV interesting
04:46 joehh malinoff: there are a few issues that I am resolving before I can do an official upload - given these are taking awhile I'll think about doing a quick interim one to debian.saltstack.com
04:46 NV might have to try installing it on a mips and see how i go then
04:46 joehh NV: http://packages.debian.org/sid/python-zmq
04:47 joehh which release are you wanting to run? wheezy? sid? something else?
04:48 NV squeeze unfortunately
04:49 NV it is in squeeze-backports though
04:49 NV bah
04:49 NV W: Failed to fetch http://debian.saltstack.com/debian/dists/squeeze-saltstack/Release  Unable to find expected entry  main/binary-mips/Packages in Meta-index file (malformed Release file?)
04:50 NV even though it's architecture independant its having a whine
04:50 malinoff joehh, Thanks, I'll wait for it
04:50 joehh yeah - reprepro is odd that way
04:51 joehh I can add a mips arch, which the salt binary packages will then be available in
04:51 joehh You'll need squeeze-backports-sloppy to get version 13.1.0-1~bpo60+2
04:51 joehh of python-zmq
04:52 joehh I haven't seen build logs (or failures) for mips, but I presume it will be there
04:53 malinoff joehh, I'm not sure is this an offtopick, but what repository manager do you use in debian? I wonder how can you store many versions of the same package. reprepro can't do that as i know
04:53 smccarthy joined #salt
04:53 jalbretsen joined #salt
04:54 NV mhmm, reason for my interest in mips is i reccon itd be pretty useful being able to configure my routers with salt - ubiquiti edgemax devices
04:54 joehh NV - ahhh - I'd been wondering about that
04:56 joehh malinoff: reprepro with a little cheat - use a different name for each stream of packages ie 0.16.x in -016, 0.17.x in -017
04:56 tseNkiN joined #salt
04:57 malinoff joehh, Could you please paste a command you run to update the repository?
04:59 joehh malinoff: reprepro include wheezy-saltstack-016 /var/cache/pbuilder/result/salt_0.17.0-2~bpo70+1~dst.1_i386.changes
04:59 joehh malinoff: is that what you mean?
04:59 malinoff joehh, yeah
04:59 malinoff thanks
04:59 joehh malinoff: though that would be a mistake, putting 0.17 into the 0.16 stream
04:59 joehh lucky i didn't press enter...
05:00 joehh NV: just added mips arch, can you test?
05:03 malinoff joehh, Am i right, your conf/dustributions file contains just two codenames - 'squeeze-saltstack' and 'wheezy-saltstack' ?
05:03 malinoff components*
05:05 malinoff ah
05:05 malinoff i see
05:09 malinoff jeblair, I was asking because I need many versions of the same package in my repository too, and I've tried to use mini-dinstall, but it is a horrible thing: there are differences in .list files, in repository structure and so on
05:12 joehh I think you were right the first time, codenames
05:13 joehh http://pastebin.com/czcjFAeW
05:14 joehh malinoff: ^^ - but I think you figured it out already
05:14 malinoff joehh, Woah
05:14 malinoff Thanks for it
05:15 joehh yeah - a little painful distributions, times 2 (for now) releases, but not too painful to work with
05:16 joehh ^^ 3 distributions (squeeze, wheezy, sid/unstable) * 2...
05:17 joehh I was planning to manage through a jinja for loop, but in the end it was simple enough to do "manually"
05:18 malinoff joehh, When you will have a time, take a look on mini-dinstall. I'm sure, your 'manual' work is way easier than other things :)
05:18 Kotoura joined #salt
05:18 Kotoura left #salt
05:19 tomeff joined #salt
05:20 NV joehh: same error - note the http://debian.saltstack.com/debian/dists/squeeze-saltstack/Release file has no reference to mips, guessing that's whats doing it
05:21 NV also the date on it seems old
05:21 NV 8th of October
05:21 NV might need to do something to force it to update if you have added the mips stuff?
05:24 joehh NV: think I see the problem - I "flooded" the wheezy repos, but missed the squeeze
05:24 joehh NV: try now
05:24 joehh sorry about that
05:26 joehh note, it will still be 0.17.0 - haven't uploaded 0.17.1 yet
05:27 NV haha
05:27 NV it seems to be complaining less :D
05:28 joehh pleasing
05:29 NV hrm, failing on deps even with sloppy
05:30 NV Depends: msgpack-python but it is not installable and Depends: python-zmq (>= 13.1.0) but it is not going to be installed
05:31 NV hrm apt-cache showpkg python-zmq is only showing 2.2.0-1~bpo60+1 from squeeze-backports, nothing from squeeze-backports-sloppy
05:32 NV packages.debian doesn't list squeeze-backports-sloppy - know an easy way to browse it?
05:32 joehh just thinking about that
05:33 NV actually, had to use -t squeeze-backports-sloppy
05:36 NV wait nevermind, that was cause it was installing an older version of salt from sloppy that it didn't error out
05:37 jefimenko joined #salt
05:37 NV http://mirror.internode.on.net/pub/debian-backports/pool/main/p/pyzmq/ - hrm so looks like everyone but mips/powerpc/sparc has 13 of python-zmq
05:37 NV which only has 2.2.0
05:41 Furao joined #salt
05:41 nliadm joined #salt
05:43 joehh no build log, so for some reason not even attempted to build it
05:43 joehh https://buildd.debian.org/status/logs.php?pkg=pyzmq&ver=13.1.0-1~bpo60%2B2
05:46 NV hrm weird
05:52 middleman_ joined #salt
05:55 linjan joined #salt
05:56 jalbretsen joined #salt
05:57 jalbretsen left #salt
06:01 joehh NV: here is the problem: https://buildd.debian.org/status/logs.php?pkg=cython&suite=squeeze-backports-sloppy
06:03 redondos joined #salt
06:03 redondos joined #salt
06:04 nliadm joined #salt
06:06 TheCodeAssassin joined #salt
06:10 bud joined #salt
06:11 joehh cython build has timed out on mipsel, and dosen't aappear to have been attempted for mips
06:11 joehh I'm not sure the difference between them
06:12 NV mipsel is little endian
06:13 NV (whereas mips is big)
06:13 NV otherwise identical
06:14 joehh Is there a machine I could log into to do a build?
06:15 joehh I might try http://www.aurel32.net/info/debian_mips_qemu.php to see if I can do a build
06:15 joehh I did a set a little while back for the raspberry pi, took a few days to build, but worked
06:15 NV err, not really - i build all my packages in a embedian chroot
06:16 NV (on x64)
06:16 joehh as in cross compile?
06:16 NV the box in question is actually an embedded router with relatively little flash that im a tad cautious of hitting too hard due to the limited write cycles
06:16 NV yeah
06:17 joehh would that work for these packages?
06:17 joehh yeah - I 'd try and avoid doing it on that sort of hardwar
06:17 joehh e
06:17 NV i should see if i can find a spare one and replace the onboard flash with a small 2.5" hard drive (its flash is connected via usb go figure :P) as a build box
06:18 NV although i daresay building will still be slow as all hell
06:18 joehh give me a few days to see if I can get the qemu-mips environment going
06:18 NV as for cross compiling, i've only ever built like 3 packages in a crosscompile environment for it
06:19 NV so no idea how it goes in general, but it seemed to work alright for those few :P
06:23 Linz_ joined #salt
06:23 matanya joined #salt
06:26 balboah joined #salt
06:31 anuvrat joined #salt
06:48 tomeff joined #salt
06:56 Lue_4911 joined #salt
07:02 Destro joined #salt
07:03 redondos joined #salt
07:05 NV hang on, KVM runs on FreeBSD? just noticed the wikipedia article mentions KVM Host support on FreeBSD - when did this happen? And how is it compared to bhyve?
07:08 NV http://www.freebsd.org/news/status/report-2007-07-2007-10.html#Porting-Linux-KVM-to-FreeBSD and http://retis.sssup.it/~fabio/freebsd/lkvm/ appear to be the only references i can find about it - guessing it's dead?
07:08 NV ah bugger, wrong channel ^^
07:09 justlooks anyone can help this https://gist.github.com/justlooks/7079759
07:14 justlooks ok ,i fixed it by myself , it is not idx=idx+1 , the correct one is ' set idx=idx+1'
07:16 bud joined #salt
07:20 [M7] joined #salt
07:21 justlooks how to do substring replace in jinja ?
07:24 __number5__ justlooks: http://jinja.pocoo.org/docs/templates/#replace and that's a list of all functions (or filters) you can use in jinja2
07:36 flebel joined #salt
07:41 backjlack joined #salt
07:42 LucasCozy joined #salt
07:46 networkpadawan joined #salt
07:58 josephholsten joined #salt
08:03 gildegoma joined #salt
08:13 costi joined #salt
08:21 anuvrat joined #salt
08:23 m0hit joined #salt
08:24 cshuman joined #salt
08:24 tomeff joined #salt
08:25 acsir joined #salt
08:28 zed_ joined #salt
08:29 zed_ question... can salt replace zookeeper?
08:30 malinoff zed_, Probably, yes
08:32 jcsp joined #salt
08:36 Destro joined #salt
08:51 josephholsten joined #salt
08:52 acsir Hi, is it normal that sometime when I run {salt  '*' state.highstate} some machines don't reply randomly? (all on debian wheezy and salt 7.0-2)
08:53 bud joined #salt
08:59 acsir cool, increasing the timeout with -t <sec> I get all of the output. Is this a normal behaviour? I was expecting a strengher behaviour, maybe should I be supposed to use salt-ssh instead?
09:03 honestly acsir: you can run it with --async
09:03 honestly then you'll get a job id
09:06 linjan joined #salt
09:08 Ryan_Lane joined #salt
09:12 anuvrat joined #salt
09:14 acsir thank you honestly I hadn't thought about option, but it works really great!
09:14 netzmonster joined #salt
09:27 acsir joined #salt
09:43 felixhummel joined #salt
09:43 Ryan_Lane joined #salt
09:47 multani joined #salt
09:49 scott_w joined #salt
09:54 lemao joined #salt
10:03 abishekk92 joined #salt
10:08 m0hit joined #salt
10:14 madpickle joined #salt
10:16 acsir joined #salt
10:17 acsir strange behaviour: I've already been setting a pillar just for some computers. If I run {salt '*' pillar.items} both on minion or master I get the correct value of my pillar on all the machines. If I run {salt-call pillar.get <myvar>} on the MINION I get back the correct value, but if I run it on master I get the value only on 2 machine out of 3, always.
10:18 acsir The strange thing is that the value seems to be correctly set, it interacts correcly with the sls file and during state.highstate
10:19 acsir I found no strange signal in debug mode
10:19 Ryan_Lane joined #salt
10:20 malinoff acsir, What command do you run on the master?
10:20 TJ1980 joined #salt
10:21 acsir on master --> salt '*' pillar.get <var>      on minion --> salt-call pillar.get <var>
10:21 acsir but now, after restarting minion/master I finally get the value also on master
10:22 acsir mayby if there is a chage in the pillars they are not send to master without a minion service restart?
10:22 jcsp joined #salt
10:22 acsir maybe*
10:23 acsir (using debian wheezy and salt 7.0-2)
10:24 giantlock joined #salt
10:24 andrej joined #salt
10:25 carlos_ joined #salt
10:35 malinoff acsir, salt '*' saltutil.refresh_pillar
10:35 jcsp joined #salt
10:37 acsir malinoff: yes despite I did it, the command didn't solve the issue. But the strangest thing is that I could see the variable on master launching pillar.items but not with get var.
10:38 acsir Only after rebooting both minion and master I could get the variable value on the master output
10:39 acsir I don't know whether it depends by the minion or the master reboot becouse I didn't any test will rebootin them.
10:39 malinoff acsir, ok, got it
10:39 acsir Maybe I could try again it later to see if the problem appears again
10:40 jcsp joined #salt
10:41 acsir One more strange thing is the during saltutil.refresh_pillar I got "None" (which I suppose should mean that everything is already updated)
10:45 bhosmer joined #salt
10:48 alexheneveld joined #salt
10:49 Ryan_Lane joined #salt
10:51 ronc joined #salt
10:59 jcsp joined #salt
11:15 srage joined #salt
11:19 elfixit joined #salt
11:20 Ryan_Lane joined #salt
11:20 m_george|away joined #salt
11:27 srage_ joined #salt
11:27 jcsp joined #salt
11:28 krissaxton joined #salt
11:42 Destro joined #salt
11:43 josephholsten joined #salt
11:45 pdayton joined #salt
11:49 acsir joined #salt
11:50 Ryan_Lane joined #salt
11:50 Ryan_Lane joined #salt
12:01 carlos joined #salt
12:02 baffle Just installed salt, and I noticed that minions try to register with Id as the hostname, not as the fqdn. Running "socket.getfqdn()" on a node manually requests in a proper fqdn. Ubuntu 12.04.3. What am I doing wrong?
12:02 jcristau that was a bug in 0.17.0
12:02 jcristau fixed in 0.17.1
12:03 baffle Wich I guess hasn't been packaged yet?
12:05 baffle Ah, released 4 days ago, not strange noone has had time to update the PPA. :-)
12:05 baffle I guess I'll just temporarily override the Id: in the minions for now.
12:07 carlos joined #salt
12:13 NetworkN3rd joined #salt
12:14 malinoff baffle, Saltstack guys says in any more-or-less useful master-minion structure you must explicitly define minions id's; that makes your structure more understandable and predictable
12:14 ronc joined #salt
12:17 baffle malinoff: Okay, so I should generate the minion configuration file during installation on every host? Or, well, just set the "Id".. My DNS infrastructure is very predictable allready. :-)
12:18 malinoff baffle, Salt bugs and features are not so predictable anyway :)
12:18 jbunting joined #salt
12:21 Ryan_Lane joined #salt
12:22 xl1 left #salt
12:26 linjan_ joined #salt
12:32 mwillhite joined #salt
12:33 anteaya joined #salt
12:37 krissaxton joined #salt
12:37 AdamSewell joined #salt
12:38 groovehunter joined #salt
12:38 Ryan_Lane joined #salt
12:39 alexheneveld_ joined #salt
12:40 cnelsonsic joined #salt
12:41 rgbkrk joined #salt
12:48 ipmb joined #salt
12:51 jbunting joined #salt
12:54 jeffasinger joined #salt
12:59 baffle Hohum, IPv6 seems a bit broken. Halite didn't really work with it enabled. Guess it is known. :) Don't people use IPv6 yet? We're all out of IPs here in Europe..
13:01 cnelsonsic left #salt
13:01 jslatts joined #salt
13:10 groovehunter hi, pls help me to understand the differences between states and modules syntax in sls files
13:10 brianhicks joined #salt
13:10 tempspace joehh: Any word on Debian packages for 0.17.1?
13:10 groovehunter https://github.com/saltstack/salt/issues/7641 similar prob to this one
13:10 groovehunter I want to run apache.a2ensite
13:11 groovehunter maybe there is another tutorial about that?
13:11 tempspace groovehunter: You can think of modules as the commands that do the work, and states as the rules that define what work you want to do
13:12 wkf joined #salt
13:13 tempspace groovehunter: When you look at the documentation, http://docs.saltstack.com/py-modindex.html - look at at ones that begin with salt.states
13:13 tempspace groovehunter: so in your example, a salt state doesn't exist for apache itself, only a module
13:14 Gifflen joined #salt
13:14 mpanetta joined #salt
13:14 acsir tempspace: I read that 17.1 on debian should get on saltstack repos soon (a few days from now). Instead, from the mailing list, I read that on the official Debian packaging system Salt should get in in testing probably after version 17.3 or 17.4
13:15 tempspace groovehunter: So, the link you posted goes over how you call a module directly from a state, which is not something you'll end up doing most of the time
13:15 groovehunter ok so far I understand I think. But I tried also with module.run as advised on the issue page
13:18 groovehunter http://pastebin.com/d0NCn3mf - Module function apache.a2ensite is not available
13:19 groovehunter http://pastebin.com/UBtjn14k and here I run hg module directly and it works , so that's confusing.
13:21 jslatts joined #salt
13:22 TJ1980 joined #salt
13:22 jrb28 joined #salt
13:23 tempspace groovehunter: You're running the hg state in that example
13:23 tempspace groovehunter: While I look something up, can you try without the ''s for site
13:23 groovehunter ok
13:24 groovehunter same.
13:27 tempspace groovehunter: From the command line, are you able to run salt mininionname apache.a2ensite example.com from the command line?
13:28 groovehunter no  "apache.a2ensite" is not available.
13:28 tempspace groovehunter: what version of salt are you running
13:29 groovehunter 0.15.1 debian wheezy
13:30 groovehunter the apache module is on the box at least
13:31 jcsp joined #salt
13:33 tempspace groovehunter: It looks like that module wasn't added until the 0.17 branch
13:33 tempspace the a2ensite
13:33 blee joined #salt
13:34 groovehunter oh yep , just inspected that apache.py file and there is no such function.
13:35 groovehunter so I go with file.symlink for the site enabling, alright.
13:36 AdamSewell joined #salt
13:36 Khollowa_ joined #salt
13:36 groovehunter ok thx for your helping hand.
13:37 jcsp joined #salt
13:37 timoguin joined #salt
13:38 tempspace groovehunter: yep, file.managed and then a file.symlink
13:38 alexheneveld joined #salt
13:38 tempspace ugh, I woke up this morning and my salt master is completely hosed, running it in debug just shows it reading in the config and then it hangs
13:38 rgbkrk joined #salt
13:39 aleszoulek joined #salt
13:39 Brew joined #salt
13:40 juicer2 joined #salt
13:41 oz_akan_ joined #salt
13:42 DredTiger joined #salt
13:43 jrb28 I'm having the same problem with my Salt Master.
13:45 whiskybar joined #salt
13:47 arapaho joined #salt
13:51 dccc joined #salt
13:53 havan joined #salt
13:55 mua joined #salt
13:55 faust joined #salt
13:55 tempspace I wonder if this has anything to do with the security advisories...
13:55 joehh joined #salt
13:59 toastedpenguin joined #salt
14:00 tempspace My salt master just magically started working
14:00 mwillhite joined #salt
14:03 jumperswitch joined #salt
14:03 kaptk2 joined #salt
14:07 zooz joined #salt
14:08 renoirb left #salt
14:09 avienu joined #salt
14:09 kermit joined #salt
14:10 jrb28 Mine too.  It took a long for the Master to initialize.
14:10 jcsp joined #salt
14:10 derelm joined #salt
14:13 mgw joined #salt
14:13 amahon joined #salt
14:14 c0bra joined #salt
14:16 mannyt joined #salt
14:16 kevinbin joined #salt
14:18 kevin joined #salt
14:19 Guest37383 heello
14:22 jalbretsen joined #salt
14:23 tyler-baker joined #salt
14:25 rgarcia_ joined #salt
14:29 pdayton joined #salt
14:30 Destro Ola
14:35 anuvrat joined #salt
14:36 Brew joined #salt
14:37 ccase joined #salt
14:39 opapo joined #salt
14:40 itbuddha joined #salt
14:41 alunduil joined #salt
14:47 TJ1980 joined #salt
14:48 brimpa joined #salt
14:49 mannyt joined #salt
14:51 jrb28_ joined #salt
14:51 TJ1980 joined #salt
14:52 jrb28__ joined #salt
14:52 jcsp joined #salt
14:53 mattmtl joined #salt
14:56 blee joined #salt
15:00 troyready joined #salt
15:01 micah_chatt joined #salt
15:02 aberant joined #salt
15:03 mwillhite joined #salt
15:04 aberant good morning all, i have a really curious problem with my really simple state that uses pip to install graphite-web
15:04 aberant Comment:   There was no error installing package '"graphite-web"' although it does not show when calling 'pip.freeze'.
15:04 aberant as you can see, my state couldn't be more simple.. https://gist.github.com/aberant/73dc8c24639f5f31e15d
15:04 aberant any ideas? i'm not much of a python guy..
15:07 jcsp joined #salt
15:13 mgw joined #salt
15:15 pentabular joined #salt
15:16 yota joined #salt
15:17 timoguin aberant, i'm not really sure why, but i just installed the package normally via pip and did pip freeze. it's not in the list.
15:18 aberant @timoguin so more of a problem with that package?  this atleast gives me a direction for my research.. thanks
15:19 timoguin aberant, yea it sounds like salt is doing what it's supposed to
15:19 brutasse timoguin: graphite doesn't install itself in the python site-packages IIRC
15:19 brutasse meaning pip freeze doesn't detect it
15:20 aberant timoguin: thanks for the sanity
15:20 robinsmidsrod joined #salt
15:20 timoguin ah
15:20 brutasse look in /opt/graphite
15:20 redondos joined #salt
15:20 redondos joined #salt
15:21 jbunting joined #salt
15:23 ninkotech joined #salt
15:32 danielbachhuber joined #salt
15:35 swa_work joined #salt
15:38 jumperswitch_ joined #salt
15:44 forrest joined #salt
15:47 jcsp joined #salt
15:49 smccarthy joined #salt
15:50 cshuman joined #salt
15:51 Corey 0.17.1 is building for Ubuntu now.
15:53 forrest nice Corey
15:53 Corey Yes. I am a nice Corey.
15:54 cwright Corey: that is great news
15:54 forrest punctuation is overrated.
15:54 cwright thanks for your work on that
15:57 bitz joined #salt
15:58 Brew joined #salt
15:58 UtahDave joined #salt
16:00 pipps joined #salt
16:02 tempspace awesome
16:02 tempspace hope Debian can come up soon too
16:04 taylorgumgum joined #salt
16:06 redondos joined #salt
16:11 jslatts joined #salt
16:11 SunSparc joined #salt
16:13 KyleG joined #salt
16:13 KyleG joined #salt
16:16 berto- joined #salt
16:16 deepakmd_oc joined #salt
16:17 madpickle left #salt
16:17 deepakmd_oc joined #salt
16:17 Lue_4911 joined #salt
16:19 ronc joined #salt
16:19 deepakmd_ joined #salt
16:21 rgbkrk joined #salt
16:23 cachedout joined #salt
16:24 redondos joined #salt
16:24 jdenning joined #salt
16:29 pentabular joined #salt
16:34 avienu joined #salt
16:35 wkf joined #salt
16:39 dlloyd joined #salt
16:40 dlloyd any ideas why the first time i run state.highstate on a new minion, from the master, it will return with no output after about 30 seconds?
16:41 dlloyd the minon will correctly run everything still
16:41 dlloyd and any future runs will display as expected
16:41 cwright dlloyd: it may be a timeout setting
16:41 cwright try increasing your timeout configs
16:42 dlloyd will try that, but i have runs that take ~5 minutes that return normally
16:43 UtahDave dlloyd: what version of Salt are you on?
16:44 dlloyd 0.16.3
16:44 UtahDave what OS?
16:44 dlloyd ubuntu
16:46 UtahDave does your highstate require a reboot of the minion?
16:46 dlloyd no
16:46 tempspace dlloyd: I have the same issue sometimes, https://github.com/saltstack/salt/issues/6881
16:47 Ryan_Lane joined #salt
16:48 eliasp is it possible to place the information from "init.sls" for windows packages somewhere else? I'd like to keep my SLS files and the installer files in separate repositories (SLS → git, installers → CIFS share)…
16:48 Corey Okay, 0.17.1 is pushed to the Ubuntu PPA.
16:48 Corey Y'know, except Raring. Because Sphinx.
16:48 cwright Corey: is there a delay before it shows up on http://ppa.launchpad.net/saltstack/salt/ubuntu/pool/main/s/salt/ ?
16:48 tempspace dlloyd: I've found myself running salt-call state.highstate from the machine itself if I absolutely need the output
16:49 UtahDave Hm.
16:49 mfournier joined #salt
16:49 Corey Quite.
16:50 Corey cwright: Note the pending: https://launchpad.net/~saltstack/+archive/salt/+packages
16:50 cwright Corey: ah, ok. thanks
16:50 Corey Plz to blame Launchpad. :-)
16:51 m0hit joined #salt
16:51 UtahDave thanks, Corey!
16:52 KyleG1 joined #salt
16:56 rgbkrk joined #salt
16:56 jefimenko joined #salt
16:57 UtahDave tempspace: Does this only happen when using the salt-cloud execution module?
16:57 tempspace UtahDave: no
16:57 eliasp how does the "salt" CLI tool invoke commands like 'test.ping'? I thought it communicates via a socket with the saltmaster which then launches the requested command, is that correct?
16:57 tempspace UtahDave: it happens at random times on random minions
16:58 UtahDave eliasp: salt uses a zeromq pub/sub interface.
16:59 UtahDave eliasp: the master publishes the common on the pub interface and the minions that match the target argument execute the command and return the results
16:59 cbloss with 0.17.1, was --version not updated??  after apt-get upgrade, Setting up salt-master (0.17.1-1precise) ...    salt-master --version  salt-master 0.17.0
16:59 tempspace UtahDave: It's just that our VM provisioning that utilizes salt-cloud  actively relies on the output of salt to check if there was an error in the highstate process
16:59 cbloss or did my install fail?
16:59 UtahDave tempspace: Yeah, I can totally understand that.
17:00 eliasp because I'm running my saltmaster in a LXC Docker container and use a second container which shares certain directories/volumes (/etc/salt, /etc/salt/pki, /run/salt/master, /srv/salt, /var/cache/salt/master) with the master to do CLI stuff… when doing a test.ping on the CLI, it times out… it works fine via Halite which is run by the master…
17:00 eliasp UtahDave: hmm… let me think how this fits in here ;)
17:00 alexheneveld joined #salt
17:00 tempspace UtahDave: my master is in a different physical location than the majority of my servers, so I didn't know if there was some kind of network intolerance on the zmq side
17:00 UtahDave cbloss: have you restarted the master?
17:00 m0hit_ joined #salt
17:01 cbloss UtahDave: first thing I tried
17:01 Furao joined #salt
17:01 eliasp UtahDave: how does the salt CLI tool talk to zeromq's pub/sub-interface? via a UNIX socket file?
17:01 UtahDave cbloss: it's possibl e there was a packaging issue. Let me try an upgrade here.  Corey, what do you think?
17:05 mgw1 joined #salt
17:05 UtahDave eliasp: ipc
17:06 cshuman_ joined #salt
17:06 eliasp UtahDave: ok, gonna dig a bit through the source… if it's a socket-based IPC, my setup should work… if it's SHM based… nope
17:07 fishpen0 joined #salt
17:07 cwright Corey: I see there are still there right now, but do you plan on leaving the 0.17.0 packages in the ppa for the time being?
17:07 cwright that would be hugely helpful until we can verify 0.17.1 and transition to it.
17:07 druonysus joined #salt
17:07 cwright for now we need to continue deploying the same version across all our nodes
17:07 cwright that being 0.17.0
17:09 TheCodeAssassin joined #salt
17:10 forrest joined #salt
17:12 UtahDave cwright: Unfortunately Launchpad will only host the latest version
17:12 cwright UtahDave: 0.17.0 and 0.17.1 are both in there right now
17:12 cwright does it clear them out on a schedule?
17:13 UtahDave cwright: I'm not sure on that actually.  I'll have to defer to Corey on that question.
17:14 avienu So who's all going to saltconf?
17:14 jcockhren avienu: it's still up in the sir for me. :(
17:14 * UtahDave raises hand!
17:14 cwright i'll be there
17:14 jcockhren s/sir/air/
17:15 forrest avienu, depends on talk
17:15 forrest *s
17:16 bhosmer joined #salt
17:16 avienu I'll be there too. Hope to meet some of you guys.
17:16 druonysus joined #salt
17:17 UtahDave We've gotten great talk proposals so far. I'm looking forward to a bunch of them myself!
17:17 forrest How many submissions are you up to now UtahDave?
17:17 forrest I don't remember what the deadline was
17:18 jcockhren forrest: Oct 28th
17:18 forrest jcockhren, ahh ok, so still some time for people to submit
17:18 fishpen0 Will the talks be available after the conference?
17:19 forrest Heh, just don't get the same streaming service that puppetconf uses
17:19 forrest I swear they stream it through a potato
17:19 forrest can't even read the text when it's live
17:22 forrest I wonder if twitch.tv allows you to stream conferences like that...
17:22 forrest I don't see why they wouldn't
17:22 UtahDave fishpen0: We're looking at the infrastructure to make the video available afterwards
17:23 UtahDave forrest: I'm don't know what the talk proposal count is up to yet. I can ask the guys managing that side of things later today
17:23 forrest UtahDave, have you talked to the PyCon guys at all about that? They might have some good tips.
17:23 forrest regarding the video recording and such
17:24 forrest UtahDave, nah don't ask, it isn't that important.
17:25 KyleG joined #salt
17:31 hkim823 joined #salt
17:35 Corey UtahDave: Packaged git commit is dbc031720ba6ba9e13b793a623ef46b09cc19978
17:35 Corey Namely, "bump version to 0.17.1"
17:35 Corey By thatch.
17:35 matanya joined #salt
17:36 pentabular1 joined #salt
17:37 rgbkrk joined #salt
17:37 felixhummel joined #salt
17:37 Corey So Hmm.
17:37 mannyt joined #salt
17:37 Corey +__saltstack_version__ = SaltStackVersion(0, 17, 1)
17:37 UtahDave Yeah.  Hm. OK, thanks, Corey
17:38 Corey Just making sure I wasn't nutty.
17:39 amckinley joined #salt
17:39 bhosmer joined #salt
17:41 Furao joined #salt
17:43 josephholsten joined #salt
17:44 KyleG joined #salt
17:44 KyleG joined #salt
17:47 bhosmer joined #salt
17:51 Corey UtahDave: Just tested: salt 0.17.1
17:51 Corey It's not me. :-)
17:51 Corey Yay.
17:53 UtahDave Corey isn't nutty!
17:54 NV hrm is there any way to make the salt minion restart a little more graceful?
17:54 NV managing the salt minion config via salt and having it restart the minion causes the state.highstate (or whatever you did to trigger it to get updated) to hang when it restarts the minion
17:55 NV something like a graceful restart (stop listening on the queue, start new instance, finish processing in progress messages, then terminate) would work
17:55 NV or am i doing something silly?
17:56 blee_ joined #salt
17:57 hkim823 hi everyone, are there any books / tutorials / guides out there I could read through to see how I can use Salt for OS X client management?
17:57 wkf joined #salt
17:59 hkim823 i'm getting a good handle of getting it setup and running, but i'm interested to see some real world examples of some states and pillars
18:01 ajw0100 joined #salt
18:01 timoguin hkim823, check out salt-formulas for some state and pillar examples: http://github.com/saltstack-formulas
18:02 timoguin most of those formulas should have a pillar.example file
18:08 ckao joined #salt
18:09 mgw joined #salt
18:10 forrest hkim823, also check out https://github.com/terminalmage/djangocon2013-sls
18:10 forrest that's a real world example of a project.
18:10 forrest so you can see how everything comes together
18:10 m0hit joined #salt
18:11 forrest NV, I don't know of something like that.
18:11 forrest Only way around that I can think of is to run thate state individually, which is lame.
18:11 Furao joined #salt
18:11 forrest *that
18:12 abishekk92 joined #salt
18:12 fishpen0 joined #salt
18:14 forrest This loops into the discussion that was going a while ago regarding jobs continuing to run after a server restart, I'm trying to remember if there was an issue open on that.
18:14 ipmb joined #salt
18:15 defunctzombie anyone know how to run salt state.highstate quietly ?
18:15 m0hit_ joined #salt
18:16 GradysGhost joined #salt
18:16 forrest you could try -l quiet defunctzombie
18:17 defunctzombie yea, for some reason on the salt master it still prints a bunch of newlines
18:18 pentabular joined #salt
18:19 woebtz joined #salt
18:20 mannyt joined #salt
18:23 TJ1980_ joined #salt
18:26 jcockhren looking through the salt source code, what's the role of "def __virtual__" ?
18:26 TJ1980_ joined #salt
18:27 eliasp can I have one win_repo per environment (base, dev, test, …) or only 1 global win_repo?
18:27 timoguin jcockhren, pretty sure that's for stuff like the pkg.installed state that will call out to the appropriate package manager for the platform, i.e., apt.installed
18:29 timoguin yea, for the pkg state, at least, the virtual function will check to see if pkg.install is present for the platform. otherwise it won't make the pkg states available
18:31 jcockhren timoguin: so it's ran at module load?
18:31 timoguin yea i think so
18:31 berto- joined #salt
18:32 jcockhren https://github.com/saltstack/salt/blob/develop/salt/returners/redis_return.py#L28
18:33 hkim823 thanks forrest and timoguin
18:33 forrest np
18:33 patrek can someone help me with file.replace?
18:33 SunSparc left #salt
18:34 timoguin jcockhren, yea in that case it won't make any of those functions in that file available to the minion if the reddis libs can't be imported
18:34 patrek I want to replace a line with the content from an external file. Is it possible with v0.17 ?
18:34 jcockhren timoguin: what is this witchcraft you speak of?
18:35 ajw0100 joined #salt
18:36 timoguin jcockhren, "a module's default name is its filename (i.e. foo.py becomes module foo), but that its name can be overridden by using a __virtual__ function."
18:36 timoguin from http://docs.saltstack.com/ref/modules/
18:36 timoguin there's some more info about __virtual__ in there too
18:37 timoguin if the function returns false the module won't be loaded
18:39 jcockhren timoguin: so __virtual__ is a salt thing and not really a python thing?
18:40 jcockhren yep. I just asked that
18:40 jcockhren :/
18:40 timoguin yea as far as i know. :)
18:41 jcockhren well that's reassuring. I was wondering what I missed in all my years
18:42 elfixit joined #salt
18:42 jcockhren so __name__ and __virtualname__ are disjoint?
18:43 lahwran joined #salt
18:45 ajw0100 joined #salt
18:48 wkf joined #salt
18:48 timoguin jcockhren, I guess so. Looks like __name__ is the filename, but __virtualname__ is used for defining a different name.
18:51 ajw0100 joined #salt
18:51 Furao joined #salt
18:52 eliasp hmm, this should actually be '--external-auth' and 'Specify an external authentication system to use.', right? https://github.com/saltstack/salt/blob/develop/salt/utils/parsers.py#L1041
18:57 eliasp filed a PR for it: https://github.com/saltstack/salt/pull/8001/
19:00 mattmtl joined #salt
19:01 jcockhren https://github.com/saltstack/salt/blob/c03d9da15cacd459ba5156d224650a42e18c8931/salt/loader.py#L720
19:02 jcockhren timoguin: ^ this is what I looking for
19:02 timoguin jcockhren, ah cool. good to know.
19:02 phab3k joined #salt
19:03 amahon_ joined #salt
19:03 aberant anyone have any success creating a django superuser with salt?  i'm looking for inspiration since i'm not sure how to proceed
19:04 jcockhren https://github.com/saltstack/salt/blob/c03d9da15cacd459ba5156d224650a42e18c8931/salt/loader.py#L556
19:04 jcockhren timoguin: ^ as well
19:04 sworcery question, if I'm using s3cmd to sync files out of s3, and I don't want it to overwrite any existing files since they already have versioning, are there any flags I should set
19:05 cro aberant: You mean the user that gets created by defaiult when you do an initial syncdb?
19:05 aberant cro: yeah, that one.. i can't seem to find any examples on the intertubes..
19:05 aberant and i'm not much of a django person
19:05 timoguin_ joined #salt
19:06 cro aberant: There is a django-admin call 'createsuperuser'
19:06 cro https://docs.djangoproject.com/en/dev/ref/django-admin/
19:06 cro so you could use cmd.run
19:07 aberant cro:  yeah, i'm using that, i'm at a fork in the road because i can't figure out how to test if that user was already created
19:07 cro ah
19:07 aberant cro: or the salt.modules.djangomod  is about as clear as mud to me
19:07 cro python script to test would be about 5 lines I think if you have access t oteh env
19:07 jankowiak joined #salt
19:08 aberant cro: that's a good idea.. i'll try that..
19:08 cro yeah--djangomod abstracts only a few of the django-admin/manage.py commands
19:11 Furao joined #salt
19:17 ajw0100 joined #salt
19:23 micah_chatt_ joined #salt
19:25 eliasp hmm, trying to get started with win_repo, but somehow I fail at it: I have sucessfully executed 'salt-run winrepo.genrepo'… running "pkg.refresh_db" on a minion returns "true", but a "pkg.installed" returns ""pkg installed" is not available"… did I forget something?
19:27 eliasp Minion runs Win7 Prof SP1 + Salt 0.17.1
19:29 eliasp hmm, ok… running 'pkg.install 7zip' worked just fine… so I'm wondering why 'pkg.installed' doesn't work…
19:30 to13 joined #salt
19:31 Furao joined #salt
19:32 jrdx joined #salt
19:32 Gifflen joined #salt
19:36 berto- joined #salt
19:38 josephholsten joined #salt
19:39 charu joined #salt
19:42 zzzirk joined #salt
19:45 taylorgumgum joined #salt
19:48 teskew joined #salt
19:49 m0hit joined #salt
19:50 aberant joined #salt
19:51 Furao joined #salt
19:55 redondos joined #salt
19:55 m0hit_ joined #salt
19:55 redondos_ joined #salt
19:58 TJ1980 joined #salt
19:59 jdenning joined #salt
20:00 eliasp UtahDave: regarding the IPC of the CLI tools… as far as I can see, they should communicate via the socket files in /run/salt/master, so when doing something like 'salt some-minion test.ping', the flow should be like this: "cli-tool → /run/salt/master/*.ipc → master → zeromq" where it then waits for the minion to pull 'test.ping'… is that correct?
20:01 Furao joined #salt
20:04 giantlock joined #salt
20:06 cachedout joined #salt
20:07 jankowiak joined #salt
20:13 berto- joined #salt
20:16 UtahDave eliasp: I do believe that's correct.
20:17 eliasp UtahDave: hmm, ok… so I'm wondering why executing something like 'test.ping' from CLI always times out while it works just fine from Halite… going to debug that…
20:17 eliasp UtahDave: thanks for the confirmation
20:17 forrest eliasp, try to run with -l debug, and see what goes on there, also check to see what the status of the job is.
20:17 AdamSewell joined #salt
20:18 eliasp forrest: will do… the weird thing is, when running it on CLI, the master debug output shows nothing at all… it seems to receive zero information… but let's see what the CLI debug output shows
20:18 amahon joined #salt
20:19 forrest at that point you need to confirm the status of the job on the minion/master to see what is goin on
20:19 forrest *going
20:19 eliasp stuck here, let's see what happens after the timeout: http://pastebin.kde.org/pvf3pfepv
20:19 forrest what does the -T option do?
20:20 eliasp well, the master doesn't receive a job at all… otherwise I'd see it in the debug log of the master…
20:20 eliasp forrest: use/create a tokenfile
20:20 forrest can you try salt 'MN1221-C0001' -l debug test.ping ?
20:20 eliasp sure
20:20 forrest or salt 'MN1221-C0001' -T -l debug test.ping ?
20:20 forrest but let's try that first command first.
20:20 eliasp k
20:21 eliasp stuck at the same, will nopaste once the timeout kicked in…
20:22 Furao joined #salt
20:22 forrest what version are you on?
20:22 forrest 0.17?
20:22 eliasp hmm, should probably upgrade the master to 0.17.1, still on .0
20:22 eliasp I'll give that a try first…
20:23 forrest there were some bug fixes in 0.17.1, but basic functionality should still work
20:23 eliasp the log after the timeout: http://pastebin.kde.org/pyq3gn1gr
20:23 eliasp ok
20:23 eliasp anyways, upgrading to make sure…
20:23 forrest ok
20:24 forrest can you restart both the master and minion services after you upgrade as well?
20:24 eliasp forrest: sure… the minion is already on 0.17.1, the master was on 0.17.0, but I don't think the minion has got much to do with the issue as I think the job doesn't even land in ZMQ's queue…
20:25 eliasp forrest: need to restart anyways as I'm running my master in a Docker container, so I'm rebuilding my Docker images and restart the container then once it has been rebuilt with 0.17.1
20:25 lineman60 joined #salt
20:25 forrest eliasp, yea I agree with you, seems master related.
20:26 forrest eliasp, do you install halite at the same time as the salt master?
20:26 jcsp joined #salt
20:27 eliasp forrest: yes… will nopaste my Dockerfile (used to build the image)… one moment…
20:27 forrest I was just gonna say, if it still fails, can you copy the docker image, and NOT install halite
20:27 forrest then see what happens
20:27 eliasp sure
20:28 eliasp forrest: I should add: I'm running master in one container and CLI in another, which shares the required directories/volumes with the master (/etc/salt, /etc/salt/pki, /run/salt/master, /srv/salt, /var/cache/salt/master)
20:28 opapo joined #salt
20:28 forrest oh I see.
20:28 forrest hmm
20:28 forrest that's very interesting
20:29 eliasp forrest: but as long as the IPC happens via sockets and not SHM I don't see why that should be an issue
20:29 xmltok joined #salt
20:29 forrest what distro are you on?
20:30 eliasp forrest: ubuntu quantal (12.04)
20:30 eliasp forrest: will very likely switch to Gentoo/CoreOS at a later point, but for now using Ubuntu for my test-setup…
20:30 rgarcia_ joined #salt
20:30 forrest ok, well if this fails and we don't get any additional output, I'd say strace/ltrace the salt command, we can see exactly what it's calling there, and where it hangs up hopefully.
20:31 forrest granted going through that output is gonna suck a bit
20:31 eliasp forrest: yep, would probably also have been my next step ;)
20:31 forrest :P
20:31 forrest You said it works fine from halite right?
20:31 eliasp forrest: yes
20:31 mwillhite joined #salt
20:31 forrest weird, can we try disabling iptables on the minion? Or is it not running any?
20:32 eliasp forrest: Minion is a Win7 Prof as that's for now the priority to get some win_repo running here… will add my Linux infrastructure at a later point
20:32 forrest ahh ok
20:32 eliasp forrest: but I could also test a Linux minion…
20:32 forrest might be worth it to do that before we head to the strace :D
20:32 forrest juuuust in case, and because I hate reading straces.
20:32 forrest so I assume everyone does
20:33 eliasp forrest: but would that make any difference regarding firewalling on the Minion whether I use Halite or CLI? I don't think so, as nothing is pushed to the Minion, but the Minion pulls instead, right?
20:33 eliasp forrest: well, I don't really mind reading straces… -e is a good friend ;)
20:33 forrest I don't know enough about the internal workings of halite to answer that. I've been writing some docs on how to set it up, but I haven't gone into the code, there's a lot of it in just a few files.
20:33 Brew joined #salt
20:33 eliasp still waiting for the Docker image rebuild as I disabled the cache to rebuild from scratch
20:34 rgbkrk joined #salt
20:34 forrest All good, I'm working on stuff over here anyways.
20:34 eliasp will ping you once I have further results/questions
20:34 diegows_ joined #salt
20:34 eliasp forrest: thanks a lot so far
20:34 forrest Yea np! I like knowing the solutions :P
20:35 timoguin joined #salt
20:36 eliasp forrest: in case you're interested, my Dockerfiles: http://pastie.org/8419618
20:37 forrest cool
20:37 forrest nice and simple
20:37 eliasp forrest: they depend on some local files (id_rsa/known_host/my fork of ldap.py), but overall they should give a good base to work on
20:38 Brew joined #salt
20:38 eliasp forrest: the -shell container needs to be run with -volumes-from=${MASTER_ID}
20:38 networkpadawan joined #salt
20:38 forrest ahh ok
20:38 eliasp forrest: and additionally another external volume for /srv/salt if that should be kept outside the container…
20:40 eliasp forrest: had to split that up into multiple images, as Docker doesn't allow yet to "null" ports which have been EXPOSE'd in another container… so running the -daemon container with /bin/bash instead of the default CMD always caused the port to be exposed a 2nd time…
20:40 forrest ahh ok
20:40 forrest my Docker knowledge isn't quite as great as I'd like it to be, another weekend project to spin up
20:41 eliasp forrest: believe me, that's more than a weekend ;)
20:41 forrest heh
20:41 eliasp forrest: working with Docker since ~3 months… really great stuff, but also still misses quite some stuff for being production-ready
20:41 aberant joined #salt
20:42 eliasp forrest: what I'm waiting mostly for in Docker: container-links and systemd integration
20:42 forrest yea I heard an interesting talk on how it was being used a few weeks ago, seemed really cool. I had heard a bit about it before that, and read some docker files, but haven't really messed with it
20:42 Furao joined #salt
20:42 networkpadawan joined #salt
20:48 pakdel joined #salt
20:48 pakdel Hi all
20:48 pakdel I have a question about writing External Pillars
20:52 Furao joined #salt
20:53 ajw0100 joined #salt
20:57 forrest I can't offer much help there pakdel since I haven't messed around with it, but if the docs don't have what you need, hang out for a bit and someone with more knowledge might be around to answer it.
20:58 eliasp pakdel: just ask your question… makes it more likely that someone picks it up to help you
20:59 eliasp forrest: btw: that's how I run my containers:
20:59 eliasp daemon: SALTMASTER_ID=$(docker run -d -v /media/volumes/salt/master/fileserver:/srv/salt -v /media/volumes/salt/master/pki:/etc/salt/pki pct/saltmaster-daemon)
20:59 eliasp shell: docker run -volumes-from=${SALTMASTER_ID} -i -t pct/saltmaster-shell
20:59 forrest hah
21:02 Furao joined #salt
21:02 anteaya_ joined #salt
21:04 zzzirk joined #salt
21:05 cewood joined #salt
21:10 eliasp forrest: problem seems to persist… strace-ing now…
21:10 forrest on a linux container too?
21:10 eliasp forrest: not going that far by now, as according to the master logs, there's no job published at all…
21:11 forrest very odd.
21:11 eliasp forrest: when doing a 'test.ping' through Halite the first thing to happen in the master log is "[INFO    ] User kxmep01 Published command test.ping with jid 20131021210630065907"
21:11 eliasp forrest: and as long as I don't see that, I assume the minion has got nothing to do with all this
21:12 forrest I'll be interested to see what the starce/master without halite results in
21:12 Furao joined #salt
21:12 eliasp yes, we will see ;)
21:13 pakdel Do I have to sync my ext_pillars to minions?
21:13 bhosmer joined #salt
21:13 pakdel Can I put ext_pillars in file_root/_modules directory?
21:17 juicer2 joined #salt
21:19 eliasp forrest: analyzing the strace outfile now… that'll be fun ;)
21:19 forrest hah
21:20 m0hit joined #salt
21:21 m0hit joined #salt
21:21 eliasp shit, should've used -f to follow forked threads… doing it again ;)
21:22 Furao joined #salt
21:22 eliasp forrest: ok… problem found: 94    connect(17, {sa_family=AF_INET, sin_port=htons(4506), sin_addr=inet_addr("0.0.0.0")}, 16) = -1 EINPROGRESS (Operation now in progress)
21:22 eliasp forrest: looks like it tries to connect via AF_INET ;)
21:23 amckinley joined #salt
21:23 pass_by_value joined #salt
21:23 eliasp now I "just" need to find out why
21:24 forrest heh
21:26 eliasp seems to be related to self.socket in payload.py …
21:27 eliasp line 160: self.socket = self.context.socket(zmq.REQ) … hmmmmmm
21:29 eliasp so zmq.Context is initialized as AF_INET instead of AF_UNIX… why why why? ;)
21:29 eliasp any ZMQ wizards around?
21:30 cowmix joined #salt
21:32 eliasp hmm, the problem is probably a bit further down the road… the SREQ class seems to be initialized with an AF_INET 'master' argument…
21:32 Furao joined #salt
21:32 m0hit joined #salt
21:32 eliasp and then self.socket.connect(master) runs into the issue I'm facing here right now
21:36 eliasp ok, might have found the bad line: https://github.com/saltstack/salt/blob/develop/salt/auth/__init__.py#L330
21:36 amckinley joined #salt
21:37 eliasp looks like this isn't able to handle AF_UNIX sockets at all but only allows AF_INET via tcp://
21:37 eliasp the same for all other methods in salt/auth/__init__.py
21:37 eliasp could someone prove me wrong? please? ;)
21:41 Ahlee are people typically defining their 'this app runs on this server' in pillars?
21:42 UtahDave Ahlee: either in pillar or in grains
21:42 UtahDave those are 2 common places
21:42 jcockhren who
21:42 Ahlee yeah, I currently target via grains, but trying to find something that doesn't freak out as much when queried via /minions in the salt-api, and something I don't have to restart a minion to update
21:42 jcockhren whoo*
21:43 eliasp whoooo? ;)
21:43 jcockhren I like the technique used to dynamically load the modules
21:43 jcockhren \o/
21:43 Ahlee UtahDave: pretty sure when i get one of you guys out here i'm going to lock that person in a room until they resolve my api issues
21:43 jcockhren Ahlee: do it
21:43 UtahDave lol
21:44 eliasp Ahlee: only feed them water + bread…
21:44 basepi well there goes your chances of getting one of us out there.....  ;)
21:44 Ahlee pretty sure docs are signed ;)
21:45 basepi actually, we'll just send seth, because his IRC bouncer is down, so he'll have no idea what hit him!
21:45 Ahlee though, i probably did just ensure i get the new fish ;)
21:45 jschairb joined #salt
21:45 Ahlee so pillars from a dynamic source: feasible/possible?
21:45 basepi define "dynamic"
21:47 Ahlee that's a good point.  I was thinking of the master asking hte minions to check a source, but instead i can just have the master hit the source and convert that into list based targeting
21:47 UtahDave Ahlee: Yeah, you could use an external pillar
21:47 basepi Ahlee: ya, pillar is updated every time you run things that need it, like highstate.  So using an external pillar, you should get up to date information whenever you need it.
21:48 Ahlee so basically just drop wahtever glue python i need into an ext_pillar def
21:48 Ahlee ah, there it is
21:48 Ahlee my_pillar = {}, #do stuff, return my_pillar
21:48 eliasp hmm, the way the URI for salt.payload.SREQ() is determined differs quite much when ack-ing the salt source: https://eliasprobst.eu/~elias/stuff/salt.payload.SREQ.html
21:49 Ahlee k, doing stuff
21:49 eliasp wouldn't it make sense to use __opts__['master_uri'] everywhere?
21:52 Furao joined #salt
21:54 eliasp hmm, ok… __opts__ is somehow automagically available in some modules… still have to fully understand Python/Salt regarding this :)
21:55 amckinley joined #salt
21:56 eliasp anyways… wondering how master_uri is determined and whether there are cases in which it isn't set to tcp://something…
21:57 forrest eliasp not sure, lol
21:58 eliasp I think the problem might be more a case of misconfiguration in my multi-container setup
21:58 forrest If no one responds, consider posting to the user group, Tom is pretty active over there.
21:58 forrest well, your multi-container setup is unique
21:58 eliasp the CLI tools try to communicate with the master just as the minions do…
21:58 forrest never seen something like that.
21:58 eliasp so they look in the config for the master's IP/Hostname and then build the master_uri out of it…
21:58 eliasp forrest: anyways a nice way to test corner-cases and find rare bugs ;)
21:59 forrest oh yea for sure
21:59 eliasp forrest: and for myself a really good thing to learn more about Salt's architecture
22:00 eliasp I just wish this day had 48h… the future PR's and Issues are just piling up on my notepad here ;)
22:02 Furao joined #salt
22:04 UtahDave eliasp: :)  glad to hear it!  :)
22:05 UtahDave eliasp: Feel free to open an PRs and send emails to the mailing list. We'll try to make sure Tom sees these zmq ones.
22:05 eliasp UtahDave: I'll do in case this doesn't turn out to be a config issue of my exotic Docker-based setup ;)
22:06 cbloss UtahDave: did you ever have a chance to try and do the apt-get update from 0.17.0 to 0.17.1 and see if --version still shows 0.17.0?
22:06 rgbkrk joined #salt
22:06 eliasp cbloss: running 0.17.1 on Ubuntu 12.04 here and --version shows 0.17.1
22:06 cbloss eliasp: did you upgrade from 0.17.0?
22:06 eliasp cbloss: no, install from scratch…
22:08 cbloss http://pastebin.com/qbvV1BSD
22:08 cbloss strange
22:09 eliasp cbloss: that's weird
22:09 eliasp cbloss: when in doubt: strace! ;)
22:10 cbloss i just removed salt-minon then reinstalled…same 0.17.0
22:10 UtahDave cbloss: I didn't. I thought I had a minion with 0.17.0 running, but it had been upgraded already.  I'll set up a fresh system tonight.
22:10 eliasp cbloss: dpkg-query -L | xargs grep -l '0.17.0'
22:10 UtahDave cbloss: did you ever install Salt through pip on that sytem?
22:11 eliasp cbloss: oups, sorry, wrong…
22:11 eliasp cbloss: dpkg-query -L salt-minion | xargs grep -l '0.17.0'
22:11 eliasp UtahDave: good guess
22:11 cbloss UtahDave: maybe. I think i may have done that because 0.17.0 wasn't available via apt-get
22:12 UtahDave Yeah, you might have to remove it with pip first
22:12 cbloss i actually may have done it via git mmm
22:12 cbloss is there a recommended install method instead of using apt-get inorder to get the latest packages when they are released? it seems apt-get is lagged
22:13 Furao joined #salt
22:15 eliasp cbloss: you could build your own deb pkg: https://github.com/saltstack/salt/tree/develop/debian
22:15 joehh cbloss: I think you are best doing (bootstrap with the develop version) so on a test machine and wait for the packages to catch up for deployed systems
22:15 joehh cbloss: but then I'm biased
22:16 joehh :)
22:17 kermit joined #salt
22:18 cbloss joehh: shouldn't I be trying to stay up with the most current stable?
22:21 joehh cbloss: yes - but your life will be simplest if you stay with the one source for updates
22:21 joehh there have been a few oddities where people have thought they have upgraded, but due to path issues and a previous install from another source
22:21 joehh have been running an older version
22:21 TJ1980 joined #salt
22:21 cbloss yeah, I can see that now :) haha. now i'm trying to figure out how to get the old source removed
22:22 joehh also, in most cases if the packagers are taking a little time, there is probably (hoepfully) a good reason :)
22:26 AdamSewell joined #salt
22:28 honestly being the salt master is so exhilarating.
22:28 honestly I throw in an evil laugh from time to time.
22:31 eliasp utahcon: forrest: problem found: the CLI tools use 'interface' from the master config instead of (optionally/when found) the IPC sockets to communicate with the master/ZMQ… setting temporarily 'interface' in the master config where I run the CLI tools works around my issue…
22:31 eliasp will file an issue for this
22:31 yidhra joined #salt
22:33 Furao joined #salt
22:36 aleszoulek joined #salt
22:36 cowmix ok.. I'm missing something here.. I have SaltyVagrant working awesome.. I just can't get the server setup working with my states.. I can run commands on my minions all day but when I try to run a state, it says all my states are missing..
22:36 cowmix shouldn't the serverless states "just work"?
22:37 packeteer is this post install?
22:37 packeteer or part of vagrant setup?
22:37 lesnail joined #salt
22:37 cowmix all my states are in /srv/salt
22:38 cowmix packeteer: is that a question for me?
22:38 packeteer sorry, yes it is
22:39 cowmix I have a master and 3 minions
22:39 phab3k do they do your bidding?
22:39 cowmix I'm just trying to get my states to run on all three
22:39 cowmix i can 'ping' them all day
22:39 cowmix run commands
22:40 cowmix but running salt states barfs.. all my sub states are "unavailable"
22:42 cowmix I just git check'd out my working states directories into /srv/salt and /srv/pillar
22:42 cowmix I'm calling it this way: salt '*' state.sls top
22:46 eliasp forrest: utahcon: filed an issue for this: https://github.com/saltstack/salt/issues/8009
22:47 forrest awesome
22:48 forrest Was in a meeting :\
22:48 eliasp forrest: no prob… it's luckily 00:48 here, so no meetings anymore ;)
22:48 cowmix i guess I should have not focuses so hard on getting things working JUST with SaltVagrant.. cause I'm confused now..  :)
22:48 forrest Do you work nights? I'd be crawling into bed by then
22:49 eliasp forrest: taking care of the daughter 07:00-16:00, working 17:00-01:00
22:49 packeteer cowmix: seems to work for me. tho i don't do anything fancy
22:49 forrest when do you sleep?
22:49 forrest lol
22:50 eliasp forrest: doing a bit of overtime today… need to finish the win_repo now and install some pilot-boxes for users wanting to finally run Win7
22:50 forrest ahh
22:50 eliasp forrest: well, sometimes after my @home-shower around 02:00 to 06:30 ;)
22:50 forrest You know I have to say, I visited Germany earlier this year, didn't have a single hot pretzel when I was there, not even from the small bakeries I found.
22:50 forrest yea that's not enough for me, I'd be dead.
22:51 eliasp forrest: it's a shame… there's nothing like a fresh hot Brezel ;) I love the smell of bakeries when they start to bake bread & stuff around 03:30…
22:51 forrest yea that was the problem it seemed, I wasn't up at 4-5 in the morning when they were fresh
22:51 eliasp forrest: nearly any bakery bakes them freshly every 30 minutes
22:52 eliasp forrest: a good bakery will never sell you a Brezel older than 60 minutes
22:52 forrest then I was bad at finding smaller bakeries it seems :P The three I found didn't have any fresh ones.
22:52 eliasp forrest: tell me next time where you are and I'll route you to the next reasonable bakery ;)
22:52 forrest lol sounds good
22:52 forrest Where are you at in Germany?
22:53 eliasp forrest: Stuttgart area
22:53 cowmix packeteer: I'm not trying to be fancy..   are you doing exactly what I'm trying to do?
22:53 cowmix :)
22:53 Furao joined #salt
22:53 forrest Ahh ok, didn't make it up there when I visited.
22:53 eliasp forrest: home of Porsche, Mercedes, Bosch, … ;)
22:54 packeteer cowmix: i pretty much follow the examples. they both work
22:54 cowmix packeteer: do you prototype with SaltyVagrant first?
22:55 TJ1980 joined #salt
22:55 packeteer yes, build and salt
22:56 cowmix I'm taking my working state files, I put them in /srv/salt and then run:  salt '*' state.sls top
22:57 cowmix that should run those state(s) on all nodes, right?
22:57 packeteer see my original questions :)
22:57 packeteer packeteer | is this post install? packeteer | or part of vagrant setup?
22:58 honestly no openvpn-formula ):
22:58 honestly whelp, guess I have to make one
22:59 cowmix the current master and minions are two VMWARE Fusion VMs running Ubuntu 12.04.2  ... FRESHLY bootstrapped with 17.0
22:59 ronc joined #salt
22:59 cowmix that's it
23:00 cowmix does that answer your questions? (i hope)
23:01 packeteer cowmix: are you running salt post install/setup of the VM's?
23:01 packeteer or as part of vagrant up ?
23:01 cowmix there is no vagrant in here...
23:01 cowmix i'm using vagrant to prototype
23:02 cowmix but in this server / minion stuff..
23:02 cowmix no vagrant
23:02 cowmix i'm just using VMWARE to run two BARE VMs.. then I run the 'bootstrap' scripts
23:02 bhosmer joined #salt
23:02 cowmix one on the server.. one on the minion.. then I do the 'test.ping' command
23:02 cowmix and that works
23:06 Gareth so i've got a pillar file that looks like this.  http://pastebin.com/VSXFjnEk  I want to iterate through so I can use the key and value pairs.  Trying to use salt['pillar.salt']('name1:name2').items() but I get a jinja error.
23:08 packeteer cowmix: oic. sorry for the confusion
23:10 amahon joined #salt
23:12 ajw0100 Anyone know of a good way to tell if salt-master and salt-minion services are ready to go? If I start them up in a script and then immediately run salt '*' test.ping I tend to get a "SaltReqTimeoutError: Waited 60 seconds" the first time around. I was thinking maybe "service salt-master status" but that returns "salt-master start/running" right away and I still get a timeout.
23:13 Furao joined #salt
23:14 pentabular1 joined #salt
23:15 bdf if I were going to volunteer to assist the person that is maintaining the debian package repository, who would I reach out to?
23:20 alunduil joined #salt
23:20 spicyWith joined #salt
23:20 spicyWith joined #salt
23:21 m0hit_ joined #salt
23:23 Furao joined #salt
23:24 pentabular1 left #salt
23:28 bhosmer joined #salt
23:29 johnthedebs joined #salt
23:30 TheCodeAssassin joined #salt
23:30 johnthedebs joined #salt
23:37 terminalmage bdf: I'd post initially to the salt-users list. We have a packagers list but I think it's private
23:38 terminalmage bdf: https://groups.google.com/forum/#!forum/salt-users
23:38 cowmix ok.. here's a simple pastebin that show my problem:
23:38 cowmix http://pastebin.com/jA9tEhEH
23:39 bdf terminalmage: cool, will do.
23:39 cowmix i know i'm doing something REALLY stupid
23:39 seanz Dumb question: If I drop files in minion.d, can I set any config values there and have them take effect?
23:39 seanz So if I have a file with master: localhost, that will be the setting after restarting the minion?
23:40 terminalmage seanz: I believe that minion.d is evaluated after /etc/salt/minion
23:41 bdf seanz: yes, though I believe the files have to end in .conf
23:41 seanz bdf: Ah, I was secretly wondering about that.
23:41 redondos joined #salt
23:41 redondos joined #salt
23:43 Furao joined #salt
23:43 foxx joined #salt
23:44 AdamSewell joined #salt
23:47 ronc joined #salt
23:53 seanz Is there a quick way to print out all salt settings, to make sure my changes have taken effect?
23:54 pentabular joined #salt
23:54 pentabular left #salt
23:57 cjh_ salt: can i set a timeout for job runs?  I have a init.d file that never returns and it keeps hanging my salt-call process
23:59 pipps joined #salt

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