Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-06-19

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

All times shown according to UTC.

Time Nick Message
00:00 hazzadous thats the end of that chapter?
00:00 chrisgilmerproj yep
00:00 chrisgilmerproj i guess the bad data from the top.sls was causing it to choke
00:00 hazzadous salt doesn't cache @Corey?
00:00 jlund left #salt
00:01 jacksontj joined #salt
00:02 Corey hazzadous: It does not.
00:02 chrisgilmerproj hazzadous: you're my hero
00:02 Corey hazzadous: Top.sls is reread each time, as are the state files, and the grains are calculated on the fly as well.
00:03 elbaschid sorry for the bother, it was my own stupidity
00:03 hazzadous but things like _modules seems to not be synced each time?  Or perhaps that is a feature of the output of sync_modules
00:03 elbaschid beware of the '-' in jinja ;)
00:04 pcarrie__ joined #salt
00:05 jschadlick1 left #salt
00:08 hazzadous So the state tree isn't cashed, I can see that, the tree as a data structure as a whole needs to be refreshed each time.  And I guess the same is true for _* but just not outputted each run @Corey ?
00:09 chrisgilmerproj hazzadous: i have to run, but seriously, thanks for your help
00:09 cxz think there's a bug in salt-cloud where /etc/salt/cloud.providers.d/*.conf will only read one provider per file
00:09 Corey cxz: I think you were talking about that lately.
00:09 cxz Corey: before i wasn't sure
00:09 cxz i thought it was my own error
00:10 cxz but now im quite positive it's a bug
00:10 napperjabber joined #salt
00:13 adotbrown joined #salt
00:19 APLU joined #salt
00:20 aat joined #salt
00:28 Furao joined #salt
00:29 jdenning joined #salt
00:30 oz_akan joined #salt
00:31 hazzadous joined #salt
00:36 napperjabber joined #salt
00:37 pentabular joined #salt
00:39 kcb joined #salt
00:39 Gwayne joined #salt
00:51 cxz Corey: after shuffling around some files in my salt config to separate envs i now get this:
00:51 cxz 'No Top file or external nodes data matches found'
00:52 cxz any ideas?
00:52 cxz in my master file i have /srv/salt/base for the base
00:52 cxz and /srv/salt/prod for the prod env
01:05 jacksontj joined #salt
01:06 santagada joined #salt
01:07 bemehow_ joined #salt
01:08 jacksontj joined #salt
01:11 bemehow__ joined #salt
01:14 oz_akan joined #salt
01:14 napperjabber joined #salt
01:14 oz_akan joined #salt
01:18 bemehow joined #salt
01:20 fragamus joined #salt
01:20 juicer2 joined #salt
01:20 fragamus auser: Hey man
01:21 Nexpro joined #salt
01:21 cxz auser: hello
01:24 auser hey cxz and fragamus
01:24 auser I'm back
01:24 auser was away for a few hours
01:24 auser well, an hour
01:25 auser what's up
01:25 fragamus Well I am encouraged by the one that launched
01:25 fragamus With that little script
01:25 auser yeah
01:26 fragamus And I think it might be a clue
01:26 fragamus If that worked then the other one should
01:27 fragamus At least regarding the key
01:27 auser yep
01:27 auser it is, I'm back to digging into it
01:28 fragamus Ok. I think my installer is the beginning of a primer of sorts
01:29 auser primer?
01:31 cxz anyone got any ideas what's going on here? 'UndefinedError: 'dict object' has no attribute 'ec2.internal''
01:33 auser from where?
01:34 auser cxz: ^^
01:34 cxz salt-call state.highstate
01:38 auser I mean, what's calling ec2.internal?
01:38 cxz oh
01:39 cxz http://paste.ubuntu.com/5778913/
01:40 auser what does your git state look like cxz
01:42 cxz uh, it's kind of long for me to remove things from it
01:42 cxz but in any case it's got stuff that's not necessary
01:42 auser private gist?
01:42 cxz let me remove everything
01:44 cxz ok i think the problem exists somewhere in git
01:44 cxz ah i didnt see that rendering sls line
01:44 cxz !!
01:44 auser :)
01:46 mgw joined #salt
01:49 Ivo joined #salt
01:49 mgw joined #salt
01:51 cellofellow joined #salt
01:55 jdaggett joined #salt
01:55 logix812 joined #salt
01:55 cellofellow Hey, I've got a strange problem. Trying to pip install a particular package available on an open but personal git repository (not github). There are other packages like this in the same requirements files, but this particular one git always exits with a code 128 and it just doesn't work.
01:56 cellofellow http://paste2.org/2xbJWK6U
01:58 oz_akan joined #salt
02:05 lex__ joined #salt
02:09 bemehow joined #salt
02:11 santagada joined #salt
02:12 dthom91 joined #salt
02:12 Furao look like you got to answer a question in less < 5 minutes
02:21 crazysim joined #salt
02:29 auser joined #salt
02:29 bemehow joined #salt
02:37 brettlangdon joined #salt
02:43 santagada joined #salt
02:44 idkfa joined #salt
02:46 zonk1024 joined #salt
02:50 Corey cxz: What's your top.sls file look like, if you haven't sorted it yet?
02:50 cxz Corey: think i have got it now, it was just a case of wrong targetting with the minion
02:50 cxz in the wrong env
02:50 cxz thanks
02:50 Corey 18:31:46 < cxz> anyone got any ideas what's going on here? 'UndefinedError: 'dict object' has no attribute  <-- You missed a colon.
02:51 mgw joined #salt
02:51 cxz that one was very confusing as i definitely didn't define ec2.internal anywhere
02:51 Corey cxz: Right, but that "dict object" means a dash or a colon isn't where you think it is. :-)
02:52 napperjabber joined #salt
02:56 cxz interesting
02:56 [vaelen] joined #salt
02:57 Corey cxz: You still afflicted?
03:02 jalbretsen joined #salt
03:12 dthom91 joined #salt
03:13 jacksontj does anyone have much experience with the requisite system? I'm a bit confused :/
03:14 Corey jacksontj: What're you trying to do?
03:14 brettlangdon left #salt
03:14 auser I do jacksontj
03:14 auser so does Corey
03:14 jacksontj trying to use the new "prereq" requisite and running into some issues-- but i believe its more to do with me getting lost in requisites than the new one not working
03:14 jacksontj i put a bunch of comments in here: https://github.com/saltstack/salt/commit/3685f5a6b7ba05a4a1842ac38e88556dbb0a866e#commitcomment-3454552
03:15 jacksontj basically i can get it to do what i expect  (oor, change, restart, ir)
03:15 jacksontj when i change packages, but i can't when i change files
03:20 jacksontj not sure if thats clear or just more confusing ;)
03:28 jacksontj well, i'm actually going to head home-- i'll spend some more time on that tomorrow ;)
03:45 jacksontj joined #salt
03:47 cods joined #salt
03:47 dthom91 joined #salt
03:48 iMil joined #salt
03:48 iMil joined #salt
03:52 napperjabber joined #salt
03:52 wyrd1 joined #salt
03:52 crazysim joined #salt
03:52 Gwayne joined #salt
03:52 aleszoulek joined #salt
03:52 terminalmage joined #salt
03:52 swann joined #salt
03:52 akoumjian joined #salt
03:52 martineg_ joined #salt
03:52 evax joined #salt
03:52 svx joined #salt
03:52 bawnzai joined #salt
03:52 nkuttler joined #salt
03:52 s4muel joined #salt
03:52 hotbox joined #salt
03:52 mperry__ joined #salt
03:52 mattikus` joined #salt
03:52 njoyce joined #salt
03:52 Daviey joined #salt
03:52 Micromus_ joined #salt
03:52 octarine joined #salt
03:52 goodwill joined #salt
03:52 whiteinge joined #salt
03:52 robawt joined #salt
03:53 lex__ joined #salt
03:53 Ivo joined #salt
03:53 elbaschid joined #salt
03:53 mephx joined #salt
03:53 jchen joined #salt
03:53 jrgifford joined #salt
03:53 yidhra joined #salt
03:53 pt|Zool joined #salt
03:53 mackstic1 joined #salt
03:53 lmnts joined #salt
03:53 utahcon joined #salt
03:53 jefferai joined #salt
03:53 Ahlee joined #salt
03:53 sashka_ua joined #salt
03:53 jafo joined #salt
03:53 cce joined #salt
03:53 minaguib joined #salt
03:53 rizumu joined #salt
03:54 wyrd1 joined #salt
03:55 crazysim joined #salt
03:55 Gwayne joined #salt
03:55 aleszoulek joined #salt
03:55 terminalmage joined #salt
03:55 swann joined #salt
03:55 akoumjian joined #salt
03:55 martineg_ joined #salt
03:55 evax joined #salt
03:55 svx joined #salt
03:55 bawnzai joined #salt
03:55 nkuttler joined #salt
03:55 s4muel joined #salt
03:55 hotbox joined #salt
03:55 mperry__ joined #salt
03:55 mattikus` joined #salt
03:55 njoyce joined #salt
03:55 Daviey joined #salt
03:55 Micromus_ joined #salt
03:55 octarine joined #salt
03:55 goodwill joined #salt
03:55 whiteinge joined #salt
03:55 robawt joined #salt
03:55 cxz hey guys, seeing a problem with timezone.get_hwclock on debian wheezy (7)
03:55 cxz http://paste.ubuntu.com/5779185/
04:05 bemehow joined #salt
04:06 elbaschid LOL just realised that I posted to the wrong channel
04:06 elbaschid please ignore my question from above and sorry for the spamming
04:16 joehh cxz: I think part of the issue is described in man 5 rcS
04:17 cxz joehh: my coworker has told me that it's possibly due to changes in wheezy
04:17 joehh yes
04:18 joehh Looks like the latest salt has code to handle this case
04:18 joehh ie git, not sure how far back that goes
04:19 auser joined #salt
04:20 mgw joined #salt
04:21 Corey mgw: Welcome to the party. :-)
04:21 mgw Corey: thanks. :-)
04:22 mgw Is redbeard2 here?
04:23 Corey mgw: I'm reading through dnsutil now.
04:23 jefimenko joined #salt
04:23 Corey mgw: He appears to be offline.
04:23 Corey Or rather, online, but absent. Harrassing him now.
04:24 Corey Yeah, it's handling zone files strictly on a read-only basis right now.
04:25 Corey dns.epicquinn.net: "dig.SPF" is not available.
04:27 mgw1 joined #salt
04:28 mgw1 I have some ideas on ways to manage dns, but haven't yet dealt with the serial incrementation issue.
04:29 fragamus joined #salt
04:30 mgw1 Other than the serial, my problem isn't so much how to manage the zone files as where the source of truth lies regarding ip<->hostname.
04:31 Corey mgw1: The cheap way to do it is to go "seconds since epoch" and just live with a dns reload every time you highstate, but that's awful. :-)
04:31 Corey sudo salt dns.epicquinn.net dnsutil.parse_zone /etc/named/db.epicquinn-internal <- Okay, progress. That throws a traceback.
04:31 fxhp joined #salt
04:31 Corey The zone is live and serving, so I don't suspect it's on that end.
04:33 mgw1 I haven't tried to use dnsutil to parse the zone files. It's throwing an error just parsing a valid zone?
04:34 Corey Yes.
04:34 Corey (By "valid" I mean that bind is contently serving it.)
04:34 mgw1 That's probably pretty valid.
04:35 Corey bind-9.9.2-0
04:35 mgw1 Can you pastebin the trace?
04:35 Corey Sure.
04:35 Corey mgw1: http://pastebin.com/XEeYYgE0
04:36 mgw1 corey: looking at the code now
04:36 Corey Yeah, it may not be thrilled with the ordering of my SOA record.
04:37 mgw1 What version of salt are you running?
04:37 Corey http://pastebin.com/B8iX73E6 is the zone file.
04:38 mgw1 What the tb is calling 113 is 168 in develop branch.
04:38 Corey This is 0.15.3
04:38 mgw1 Then something significant seems to have changed.
04:39 Corey Yeah, let's see.
04:39 mgw1 maybe try substituting the version from develop.
04:40 mgw1 Can you pastebin your zone file too?
04:40 Corey mgw1: Look up. :-)
04:40 mgw1 sorry
04:40 Corey No worries.
04:42 Corey mgw1: Same trace, but it's 168 now.
04:42 Corey So whatever changed, it didn't impact this.
04:43 auser joined #salt
04:43 mgw1 It's your NS line
04:44 fragamus_ joined #salt
04:44 mgw1 >>> "                                NS                      ns1.epicquinn.net.".split()[2]
04:44 mgw1 Traceback (most recent call last):
04:44 mgw1 File "<input>", line 1, in <module>
04:44 mgw1 IndexError: list index out of range
04:44 Corey mgw1: Your'e right.
04:44 Corey you're*
04:45 Corey mgw1: The problem is, the canonical source (in this case, BIND) is okay with that.
04:46 mgw1 Yeah, it's a bug in dnsutil IMO. I have the same construct in my zone files.
04:46 Corey mgw1: split[1] works.
04:46 mgw1 in that case, but not for the other lines.
04:46 Corey Looks like an offbyone.
04:47 mgw1 It's treating the whitespace before NS as meaningless
04:47 mgw1 but it's not meaningless
04:47 Corey mgw1: This is where we get to go insanely deep into the zone file requirements, isn't it. :-)
04:48 mgw1 yes, because there may be a valid use for having something before the NS
04:48 Corey I mean, we could wrap a conditional around it and "If it's an NS line, then attempt split[1]" but that's a bit too ede-case for me.
04:48 mgw1 http://www.zytrax.com/books/dns/ch8/ns.html
04:48 Corey edge*
04:48 lex__ joined #salt
04:49 mgw1 I think the correct solution is to treat leading whitespace as an empty item
04:49 mgw1 rather than discarding it
04:49 mgw1 NS isn't the only place this could happen.
04:50 mgw1 e.g.
04:50 mgw1 IN      MX     10  mail  ; short form
04:50 mgw1 ; the line above is functionally the same as the line below
04:50 mgw1 ; example.com. IN     MX     10  mail.example.com.
04:50 Corey This is where my knowledge of the python string library falls down.
04:51 Corey mgw1: I mean, we could lstrip it for parsing purposes, but...
04:52 mgw1 split already does that
04:52 mgw1 which is the problem
04:52 mgw1 I think split will need to be replaced with a regex
04:53 napperjabber joined #salt
04:57 Corey mgw1: I wonder how bind reacts to erroneous leading whitespace. Testing.
04:58 Corey mgw1: Fantastic, it gets angry.
04:58 Corey So if whitespace starts a record, it's significant.
04:58 Corey If something starts with whitespace, the next entry on the line is detected as an RR type.
04:59 mgw1 yeah
05:00 mgw1 I have a regex working
05:00 Corey mgw1: What'd you come up with?
05:01 Corey (This is out of my wheelhouse, generally speaking)
05:01 mgw1 Corey: https://gist.github.com/mgwilliams/66c0035d6ed1672674c4/raw/cb0b98335f82ec6a81a753763152cddcbd81d251/gistfile1.txt
05:02 mgw1 I'm missing the IN for some reason
05:02 Corey mgw1: Interesting. I see that it works, but I don't understand how. Reading.
05:02 mgw1 but it keeps the leading whitespace, that's what's important :-)
05:03 Corey There isn't an IN in that line, it seems?
05:03 mgw1 yeah
05:03 nrub joined #salt
05:03 mgw1 hmm, so it should still be comps[1]
05:03 mgw1 if it's trying to match that line
05:04 mgw1 Corey: I don't think that's the RR that line is *trying* to match, it just happens to blow up on that line
05:04 mgw1 example.com.        IN      NS      ns1.example.com.
05:04 mgw1 is what it's trying to match
05:05 mgw1 wait… it's worse
05:06 mgw1 name           ttl  class   rr     name
05:06 mgw1 example.com.        IN      NS      ns1.example.com.
05:06 mgw1 it's in columns
05:06 mgw1 sort of
05:06 Corey You're right.
05:06 Corey If the column is empty, default applies.
05:06 Corey (For TTL anyway)
05:06 Corey Or rather, the TTL portion of the SOA.
05:06 Corey mgw1: We may be looking at this backwards.
05:07 mgw1 But "column" from my experience is somehow inferred — you can have multiple spaces between columns, and they can vary from one RR to another
05:07 mgw1 Corey: howso?
05:07 Corey Rather than retrofitting Salt to manage your / my / Timmy's different zone file formats, what if we took the opposite approach and constructed the zone file based upon a module? Then we can decide what we want Salt's managed zonefiles to look like.
05:08 Newt[cz] joined #salt
05:08 Corey For instance, "Every record has an explicit TTL when Salt builds the zone file"
05:08 mgw1 That's an idea
05:08 mgw1 And it doesn't need to parse the existing one at all, except to grap the serial
05:08 mgw1 s/grap/grab
05:08 Corey Well sort of.
05:10 bemehow_ joined #salt
05:10 jpcw joined #salt
05:12 Corey There'd be human effort involved in setting up the zonefile to be managed initially, we'd not want Salt to stomp it.
05:15 mgw1 Corey: so, a state machine that appears to actually work: https://gist.github.com/mgwilliams/72f59558d620e30280e2
05:15 mgw1 It's 1255 lines
05:16 Corey yikes. :-)
05:16 Corey reading.
05:17 mgw1 In other words, there's a lot more to parsing a zone file than dnsutil is doing.
05:18 Corey Yes.
05:18 mgw1 Is it that terrible if the module stomps the existing zone file? Isn't that the job of the configuration management system — to ensure state?
05:18 Corey I'm not sure we want to be parsing it-- yeah, exactly.
05:18 mgw1 If that's the case, the only issue I see is how to update the serial only when there's a change.
05:18 Corey Checking what the hosts module does.
05:19 Corey mgw1: Oh, since we're apparently able to write our own standard, what do you think of yyyymmdd## as a serial format?
05:20 mgw1 That's fine with me…. but like you said, we don't want to increment it just because highstate is run.
05:20 Corey Right.
05:20 mgw1 But we could grep that out easily
05:20 Corey There's already logic in Salt for this.
05:21 mgw1 if the zone file is not in an expected format, we just clobber it :-)
05:21 Corey I just need to see where it is and figure out how to hook it.
05:22 mgw1 logic for which aspect?
05:22 mgw1 I think a dnszone state module could mainly wrap the file module
05:22 vaxholm joined #salt
05:22 Corey Well Salt already makes a determination whether or not to overwrite a file-- it's probably something as simple as a diff. Grepping out the serial from that diff for zone files would get us most of the way there.
05:24 mgw1 It's the sed function I think
05:24 mgw1 in salt/states/file.py
05:25 mgw1 actually that's something different
05:25 mgw1 corey: more likely the diff from file.py:manged
05:27 mgw1 and also get_diff in salt/modules/file.py
05:27 Corey mgw1: https://groups.google.com/forum/#!topic/puppet-users/U4tD8jBreUU <- some thoughts from when this came up on the puppet users mailing list.
05:30 mgw1 Another big question is how to convey the RR data to the dnszone module.
05:34 oz_akan joined #salt
05:35 mgw1 corey: I need to go for now
05:36 Corey mgw1: Yeah. Thanks for the help so far. :-)
05:36 Corey I'll talk to redbeard2 when he shows up. :-)
05:36 mgw1 ok
05:39 jdenning joined #salt
05:40 enter_pin joined #salt
05:42 nrub joined #salt
05:43 jdaggett joined #salt
05:46 azbarcea joined #salt
05:46 a7p joined #salt
05:46 tmmt_ left #salt
05:46 az87c joined #salt
05:49 stas joined #salt
05:51 enter_pin what is the equivalent to the reactor system in puppet?
05:51 enter_pin or puppet/mcollective?
06:01 krissaxton joined #salt
06:08 Tekni joined #salt
06:08 Gareth_ joined #salt
06:09 echos_ joined #salt
06:09 chuffpdx_ joined #salt
06:09 EugeneK joined #salt
06:09 that_guy_ joined #salt
06:09 Kyle_ joined #salt
06:09 FreeSpencer joined #salt
06:10 N-Mi joined #salt
06:10 N-Mi joined #salt
06:10 morganfainberg joined #salt
06:10 SpX joined #salt
06:11 Valdo joined #salt
06:12 erasmas joined #salt
06:12 pjs joined #salt
06:13 MTecknology joined #salt
06:13 balltongu joined #salt
06:14 omame joined #salt
06:14 Kamal_ joined #salt
06:17 f4cl3y joined #salt
06:18 arBmind joined #salt
06:18 jdaggett joined #salt
06:33 fragamus joined #salt
06:35 vaxholm joined #salt
06:41 oz_akan joined #salt
06:42 __gotcha joined #salt
06:42 __gotcha joined #salt
06:58 jeffmendoza joined #salt
07:06 stas joined #salt
07:13 bemehow joined #salt
07:21 Slipo joined #salt
07:24 efixit joined #salt
07:29 axisys joined #salt
07:30 linjan joined #salt
07:32 mtou joined #salt
07:36 mtou Hi everyone. Got a question. I have used salt daily to manage ~20 servers for more than a year now and am really happy with it, but... The fact that all the minions have a cache of all the master configuration files is a pain... one server compromised and all the configuration of the others can be deduced from the leaked master configuration. Is there a way to make all the system safer or am I doomed to fear for a single compromise to compro
07:37 faust joined #salt
07:39 Furao mtou: remove password, key, and stuff like that from the states
07:39 Furao use pillars for that
07:42 noodles775 joined #salt
07:43 arBmind joined #salt
07:52 felixhummel joined #salt
07:54 vaxholm joined #salt
07:56 mtou Furao: thanks
07:57 mtou I already use pillars for the sensible infos, except for the ssh keys (for which i use the ssh_auth:present service)
07:58 Furao I keep an extra state git repo for sensitive information
07:58 Furao pillar refer to it
07:59 mtou but my top.sls filters on the hostname to choose what states to include, and my point stands: in my current policy, every monion has these rules, which I feel as sensible information in the wild. any thought ?
07:59 Furao so anyone can work on states, no need to requires access to private ssl/ssh files
07:59 Furao I don't have hostname in my states top.sls anymore as I switched to a roles based strategy
08:00 Furao {% if pillar['branch']|default('base') == 'master' %}base{% else %}{{ pillar['branch']|default('base') }}{% endif %}:
08:00 Furao '*':
08:00 Furao {% for role in pillar['global_roles'] %}
08:00 Furao {% if role in pillar['roles']|default([]) %}
08:00 Furao - roles.{{ role }}
08:00 Furao {% else %}
08:00 Furao - roles.{{ role }}.absent
08:00 Furao {% endif %}
08:00 Furao {% endfor %}
08:00 Furao that's my top.sls
08:00 Furao each host pillar specify it's list of roles it need to deploy
08:00 mtou perfect, thanks
08:01 Furao and the roles are used to build monitoring configurations as well
08:02 mtou for monitoring, I use omd/check_mk so monitored services are largely autodiscovered, except for a few nagios extras which are indeed configured via salt
08:02 pentabular joined #salt
08:02 Furao what if a host is down?
08:02 Furao it's not autodiscovered?
08:03 Furao tu devrais regarder shinken, c'est made in france ::
08:03 Furao :P
08:03 mtou shinken est inclus dans omd également, mais je suis avec check_mk pour l'instant ;)
08:06 Furao j'ai un mapping states -> rôles et ma config shinken perform les check nrpe qui sont dans les states
08:06 Furao alors j'ai un rôles -> states pour le déploiement et l'inverse pour le monitoring
08:07 Furao et j'utilise salt mine pour informer mon arbiter comment générer sa config
08:07 Furao et j'ai un groupe par role
08:07 Furao c'est la meilleure stratégie j'ai trouver depuis 1 an pour salt + monitoring
08:08 Furao et les hosts qui sont down sont quand meme monitorer quand  il build sa config
08:08 mtou je suis encore en 0.10.4 installé "manuellement"
08:08 mtou il y a un dépot debian maintenant kkpart ? ça serait plus pratique ;)
08:08 Furao tu as quel distro?
08:09 mtou 95% de debian stable
08:09 mtou sinon, lmde unstable (donc pareil on va dire)
08:09 Furao http://debian.saltstack.com/debian/
08:09 mtou chouette
08:11 adotbrown joined #salt
08:26 zooz joined #salt
08:29 krak3n` joined #salt
08:31 Lucas joined #salt
08:37 backjlack joined #salt
08:37 noodles775 Hi! Just wondering whether 0.15.3 will be pushed to debian sid? The saltstack debian archive currently has 0.15.1 there?
08:43 fredvd joined #salt
08:45 scott_w joined #salt
08:46 vaxholm joined #salt
09:06 arBmind joined #salt
09:15 jhauser joined #salt
09:16 ollins joined #salt
09:18 aleszoulek joined #salt
09:33 aleszoul3k joined #salt
09:39 krissaxton joined #salt
09:42 adotbrown joined #salt
09:48 crashmag joined #salt
09:49 ggoZ joined #salt
09:55 ollins joined #salt
10:00 tonthon Hi
10:05 efixit joined #salt
10:08 djn joined #salt
10:17 cce joined #salt
10:17 cce joined #salt
10:21 arBmind joined #salt
10:29 tonthon I try to setup salt-syndic, but I fail to get the master's key up to the "master of master"
10:29 tonthon I've added the syndic_master to the /etc/salt/master file
10:30 hazzadous joined #salt
10:30 tonthon restarted salt-syndic, but I don't receive any pending key on the remote server
10:30 tonthon should I restart the salt-master too ?
10:31 efixit joined #salt
10:34 luff joined #salt
10:36 tonthon ok, I'd put my hostname between ''
10:39 jhauser joined #salt
10:39 q4brk joined #salt
10:41 q4brk hi! does anyone have a clue why I'm getting this error http://pastebin.com/reNycMiu using salt 0.15.3 ?
10:42 jhauser_ joined #salt
10:43 jhauser__ joined #salt
10:43 whiskybar joined #salt
10:47 jhauser_ joined #salt
10:48 Furao q4brk: that's a bug in salt, I think it had been fixed in git. it's an error with apt-get install, try to install manually and see the error it does
10:51 jhauser joined #salt
10:53 luff left #salt
10:54 q4brk Furao: seems there's a tiny error in the package's post-install script
10:54 q4brk thanks
10:54 jhauser_ joined #salt
10:57 tonthon q4brk: I had the same this morning :)
10:58 felskrone joined #salt
11:00 jhauser_ joined #salt
11:06 krissaxton joined #salt
11:11 giantlock joined #salt
11:11 Teknix joined #salt
11:13 logix812 joined #salt
11:31 jpcw joined #salt
11:47 arBmind1 joined #salt
11:49 jefimenko joined #salt
11:51 __gotcha joined #salt
11:54 __gotcha_ joined #salt
11:54 __gotcha_ joined #salt
11:55 __gotcha_ joined #salt
12:03 aat joined #salt
12:04 oz_akan joined #salt
12:08 pgSnake joined #salt
12:13 krissaxton joined #salt
12:16 Lucas joined #salt
12:18 jslatts joined #salt
12:20 yota joined #salt
12:28 taotetek joined #salt
12:29 cron0 joined #salt
12:29 taotetek to create an image for "salty vagrant" - do you basically just need to create an OS image that has a correctly configured minimum install + salt minion on it?
12:30 taotetek I'm noodling over how difficult it would be to add a linux distro that's not yet supported.
12:32 aat joined #salt
12:34 f4cl3y joined #salt
12:43 jslatts joined #salt
12:53 oz_akan joined #salt
12:55 unicoletti_ joined #salt
13:00 oz_akan joined #salt
13:03 Lucas joined #salt
13:04 mjulian joined #salt
13:05 juicer2 joined #salt
13:09 whiskybar joined #salt
13:10 krissaxton joined #salt
13:10 elduderino joined #salt
13:13 timoguin joined #salt
13:14 tonthon Hi, is there a known bug with module refresh through saltutil.sync_modules ?
13:14 tonthon I need to rename my module to see him synchronized
13:15 tonthon ok, I'v got it
13:17 __gotcha_ joined #salt
13:17 __gotcha_ joined #salt
13:18 __gotcha joined #salt
13:18 __gotcha joined #salt
13:18 aat joined #salt
13:20 __gotcha_ joined #salt
13:20 jalbretsen joined #salt
13:21 __gotcha_ joined #salt
13:22 elduderino joined #salt
13:25 joehh tonthon: I've seen something, but I can't quite put my finger on it. I seem to need to run sync_modules twice
13:26 tonthon joehh: at least, it runs as expected (and the refresh option is enabled by default)
13:27 joehh thats good
13:28 racooper joined #salt
13:35 mgw joined #salt
13:37 Kholloway joined #salt
13:39 aleszoulek joined #salt
13:39 elduderino joined #salt
13:43 nrub joined #salt
13:45 fspot joined #salt
13:45 ageron joined #salt
13:46 ageron Hi everyone.  I'm looking for a way to implement the equivalent of Chef's search functionality.  Basically, I'd like to have a load balancer (haproxy) automatically configured to point to all the IPs of the minions that belong to the app_server node group.  I'm probably overlooking something really trivial.
13:47 ageron The `data` state functionality looked promising but apparently a minion cannot access another minion's data
13:51 __gotcha_ joined #salt
13:51 __gotcha_ joined #salt
13:57 Slipo joined #salt
13:58 hazzadous @ageron some good pointers on http://www.slideshare.net/SaltStack/realtime-infrastructure-management-with-saltstack-seth-house
13:58 ageron Great, tx, I'll take a look right now
13:59 hazzadous (the example they use is setting up haproxy in the way you describe)
14:00 jdaggett joined #salt
14:02 brianhicks joined #salt
14:03 Kholloway joined #salt
14:03 backjlack joined #salt
14:05 ageron @hazzadous: slide #17, perfect match, thanks a lot!  :-))
14:05 hazzadous great
14:06 tonthon Is there a way to list available modules with salt-call ?
14:07 mtou left #salt
14:09 mgw joined #salt
14:10 Slipo joined #salt
14:11 JPmx joined #salt
14:13 aberant joined #salt
14:22 nrub joined #salt
14:23 opapo joined #salt
14:27 abe_music joined #salt
14:29 kho joined #salt
14:29 santagada joined #salt
14:30 kho joined #salt
14:32 bhosmer joined #salt
14:37 bemehow_ joined #salt
14:43 krissaxton joined #salt
14:45 lex__ joined #salt
14:46 tonthon is there a difference between "pillar.get" and "pillar.data key=" ?
14:47 tonthon "pillar.data key=" returns some datas while pillar.get doesn't return anything
14:50 Odd_Bloke tonthon: I feel like one of them uses the local pillar cache and one goes straight to the master.
14:50 Odd_Bloke tonthon: But I might be making that up completely.
14:51 Slipo joined #salt
14:51 tonthon Odd_Bloke: my initial goal is to retrieve pillar datas from within a module code through the __pillar__ dict (?)
14:52 tonthon but I don't get any custom pillar in it
14:52 tonthon any idea ?
14:54 Odd_Bloke tonthon: Not really, I'm afraid. :)
14:54 tonthon is there a supposed way to achieve this one ?
14:55 teskew joined #salt
14:55 tonthon (or is pillar reserved to states ?)
15:01 santagada joined #salt
15:02 krak3n` joined #salt
15:02 tonthon ok, it was all cache related, forgot to launch saltutil.refresh_pillar
15:11 jalbretsen joined #salt
15:11 faust joined #salt
15:18 conan_the_destro joined #salt
15:21 fys joined #salt
15:25 unicoletti_ joined #salt
15:29 chrisgilmerproj joined #salt
15:29 jschadlick joined #salt
15:34 santagada joined #salt
15:38 newyorker joined #salt
15:39 napperjabber joined #salt
15:40 newyorker Hi all, quick question: how can I make salt jinja templates use grains present on some machines, but not others? (jinja template code: http://pastebin.com/eeVXcWJ0)
15:41 newyorker I'm trying to use the 'biosversion' property to detect whether the machine is in VirtualBox or on an actual server (e.g., Linode)
15:42 newyorker But my actual servers don't have the 'biosversion' property, so they raise an error when parsing the template: http://pastebin.com/9z8ds7XA
15:42 unicoletti_ left #salt
15:43 newyorker Is there some way to try/except in a jinja template?
15:47 milind joined #salt
15:49 esrax joined #salt
15:52 wendall911 joined #salt
15:52 salticus joined #salt
15:53 krak3n` joined #salt
15:53 fragamus joined #salt
15:59 auser joined #salt
16:02 fspot left #salt
16:07 Nazca joined #salt
16:08 Nazca another day another problem ... am I doing something wrong, or does pkg targeting a centos install always fail, regardless of if the packages are all provided?
16:13 SunSparc joined #salt
16:14 SunSparc Is there a chart that compares the different paid versions of Salt?
16:17 auser hey all
16:18 jdenning newyorker: {% if grains['bioversion'] is defined %} … {% endif %}
16:19 bemehow joined #salt
16:21 newyorker jdenning: ah ha, I really should have figured that out for myself :) thank you!
16:21 jdenning newyorker: no prob :)
16:22 auser SunSparc: I believe they are functionally equivalent
16:22 bemehow_ joined #salt
16:23 kermit joined #salt
16:23 abe_music what's up auser….had a call with UtahDave a couple weeks ago and he mentioned you might be releasing some reactive states for hadoop. Saw your slides, but couldn't gleam much out of them unfortunately.
16:23 auser abe_music: rad
16:23 auser yep
16:24 auser I have a bunch of states I'm gearing up to release here soon
16:24 abe_music cool, i started watching your states repo and patiently waiting :-D
16:25 abe_music auser: we (company i work for) are working on an internal tool called stackd.io that is essentially AWS OpsWorks built on salt. we'll be opening it up after a bit of maturing
16:26 auser nice!
16:26 abe_music naturally orchestration using reactor is something that will be amazing for this sort of thing
16:26 SunSparc auser: Ok
16:27 Nazca grrrr @ yum being annoying
16:27 Nazca it rewrites package names instead of failing if you give it something close but not right
16:28 auser yep, I think it's awesome.. It's great to see salt exploding in popularity
16:28 auser abe_music: I'm just figuring out the licensing with the company that I *technically* built them for
16:29 abe_music auser: yeah same here….i'm getting some pushback from the marketing folks that we shouldn't release it, but product thinks it's a great idea….silly argument if you ask me since it's a very common problem for shops that operate entirely in the cloud
16:29 auser yeah
16:29 bemehow joined #salt
16:30 auser abe_music: the tool we use is not at all mature (yet), but we opened sourced it last week
16:30 auser salt-cli
16:30 auser which handles uploading, highstating, launching, etc from a dev machine
16:30 abe_music auser: i saw you using that in your slides, but wanted to ask who you thought the audience was for that tool
16:31 auser well… right now… me + a few others.
16:31 abe_music auser: currently we're using django+celery to drive custom tasks that leverage salt and salt-cloud to do the launching, provisioning etc. with some custom stuff for working with AWS-specific services like Route53 for DNS
16:31 abe_music auser: :)
16:32 auser I am not a fan of sshing into the master box to perform salt tasks
16:32 auser but that's me
16:32 auser nice
16:32 auser I'm writing a book on angularjs, so I'm interested to see backends to handle salt tasks
16:33 abe_music auser: awesome, when do you think you'll be wrapping up the book? (i'm a new angularjs user)
16:33 auser well, I'm doing the peepcode at the same time, so it's going slower than I want, but I'm aiming to get the peepcode done in a few weeks, so the book won't be that far behind
16:34 abe_music nice!
16:36 auser anyway, as soon as I get permission, I'll start moving the states over for public consumption
16:38 abe_music auser: cool, thanks for the update
16:38 auser sure, abe_music
16:38 auser the slides spell out a lot of the work already, how to get them reactive and things
16:39 santagada joined #salt
16:42 jschadlick where are you supposed to put custom runners, by default?
16:43 auser jschadlick: I *believe* in the _runners dir
16:43 bemehow_ joined #salt
16:43 mikedawson joined #salt
16:43 auser and then saltutil.sync_runners
16:43 jschadlick just like _modules?
16:43 auser yeah
16:43 jschadlick cool, I'll try that out, thank you
16:44 kcb joined #salt
16:50 jefimenko joined #salt
16:51 jschadlick hmm
16:51 cleeming[foxx] joined #salt
16:51 jschadlick the minions don't see sync_runners ( I guess that make sense, its for the master)
16:55 bensix2 joined #salt
16:55 krissaxton joined #salt
16:56 derelm joined #salt
16:58 jefimenko joined #salt
17:01 clone1018 Hey guys, there's not a PHP client api is there?
17:01 bemehow joined #salt
17:03 xt what does that even mean? There's a REST API that you can program from any language
17:04 dthom91 joined #salt
17:04 clone1018 Right, http://docs.saltstack.com/ref/python-api.html
17:04 wick joined #salt
17:04 clone1018 PHP version of that, was just curious
17:05 clone1018 Already started building a library for accessing the rest api
17:06 xt a php version of that doesn't make sense, and is possible
17:06 xt but I might misunderstand what you're looking for
17:08 clone1018 haha, how doesn't it make sense?
17:09 xt it's a bit like asking "is there a salt written in php?"
17:10 clone1018 That library is an interface work working with your salt master
17:10 clone1018 for working*
17:14 xt nay
17:14 xt it's an API
17:14 xt it's not somthing you import
17:14 bfrog so top.sls has base: ...
17:15 bfrog but what if I want to like various roles and such for different boxes
17:15 bfrog confused by the way to do that
17:15 bfrog must be in the wrong part of the docs, or missing it
17:17 m_george|away joined #salt
17:18 auser bfrog: roles for instances?
17:18 bemehow_ joined #salt
17:18 bfrog well like I want some boxes to be nginx boxes, some to be running erlang apps, etc
17:19 bfrog in the docs I just see base: -webserver or something
17:21 auser sure
17:21 auser I put up a doc how I do that
17:21 auser https://speakerdeck.com/auser/reactive-saltstack
17:21 auser maybe it might help you
17:22 auser line 9-11
17:22 logix812 joined #salt
17:24 auser I use grains
17:24 auser set the roles as an array on the grains
17:24 auser and match against those
17:24 aberant joined #salt
17:24 auser it works very well
17:24 auser brb
17:25 jesusaurus oh, cool, you can say 'if grains['foo'] is defined' ?
17:25 jesusaurus does that work for pillar data too?
17:26 m_george left #salt
17:26 jesusaurus ive been doing: 'if pillar.get('foo', false)' to avoid errors
17:28 efixit joined #salt
17:28 Ryan_Lane joined #salt
17:33 Ryan_Lane re: https://github.com/saltstack/salt/issues/976 <— what are get methods? I can't find documentation for them
17:33 Ryan_Lane and I don't see them in release notes
17:33 napperjabber joined #salt
17:46 santagada joined #salt
17:47 pentabular joined #salt
17:51 bemehow joined #salt
17:51 LarsN_ joined #salt
17:52 foxx[a] joined #salt
17:54 minaguib joined #salt
17:55 crashmag_ joined #salt
17:56 dthom91 joined #salt
17:59 raz[zZzz] joined #salt
17:59 MKFG joined #salt
17:59 andyshin` joined #salt
18:00 raz joined #salt
18:03 cellofellow joined #salt
18:04 brianhicks joined #salt
18:06 jefimenko joined #salt
18:07 jeffmendoza joined #salt
18:08 Ixan joined #salt
18:11 UtahDave joined #salt
18:19 jacksontj joined #salt
18:22 cellofellow Hi everyone
18:22 zonk1024 joined #salt
18:26 jacksontj joined #salt
18:27 jlund joined #salt
18:31 jacksontj UtahDave: yt?
18:33 __gotcha joined #salt
18:33 __gotcha joined #salt
18:35 UtahDave jacksontj: yep!  How's it going, man?
18:35 jacksontj pretty good :)
18:35 jacksontj been a busy week-- and i'm headed over to velocity later today
18:35 jacksontj which should be fun
18:35 jacksontj hows life back in utah?
18:35 UtahDave ah, nice!
18:36 UtahDave Good!  Spent last night up at girl's camp. They needed a chaperone.  So I got off the plane yesterday and basically drove straight up to the camp
18:36 UtahDave I just got home a couple hours ago, showered, and then came in to work.
18:37 jacksontj lol, busy busy busy
18:37 UtahDave yep
18:38 jschadlick joined #salt
18:38 UtahDave It was kind of nice to be disconnected from the internet for a day. It was beautiful up there.
18:38 jacksontj yea, i do that for most of my vacations-- disconnect and go somewhere :)
18:39 jlund joined #salt
18:39 jacksontj i spent some time yesterday messing with the new requitsite-- and I got it paritally working (I put my details in https://github.com/saltstack/salt/commit/3685f5a6b7ba05a4a1842ac38e88556dbb0a866e#commitcomment-3454552)
18:39 jacksontj cant get it to work quite right for the files, but it works exactly as i expected for the package :)
18:40 auser nice jacksontj
18:40 auser Holla UtahDave
18:41 UtahDave yo, auser!
18:41 UtahDave jacksontj: cool, reading through it right now
18:42 jacksontj I'm sure i'm missing something simple-- requisites can get a bit confusing ;)
18:42 dthom91 joined #salt
18:42 UtahDave yes, they can be. Especially the prereq adds another level of possible confusion
18:43 melinath joined #salt
18:43 Rick1 joined #salt
18:43 whiteinge Ryan_Lane: i believe that's a reference to pillar.get  http://docs.saltstack.com/ref/modules/all/salt.modules.pillar.html#salt.modules.pillar.get
18:43 aranhoide joined #salt
18:43 whiteinge s/i believe//
18:46 jacksontj i guess i should probably open an issue for this feature instead of posting on Tom's commit ;)
18:46 Rick1 UtahDave I was asking about implementing Reactors a week or so ago and finally have time to get back to it.  You had said I had to enable it in the master config?
18:48 nrub joined #salt
18:49 UtahDave jacksontj: that would be helpful.  :)
18:49 UtahDave Rick1: Yes, your reactor configuration is set in the master config.  Have you looked at the Reactor docs yet?
18:49 Ryan_Lane whiteinge: ah. that's a roundabout way of doing that
18:50 jacksontj UtahDave: made one: https://github.com/saltstack/salt/issues/5636
18:50 UtahDave cool, thanks.
18:51 Rick1 Yes but I find them very vague, several of my associates and I are not 100% on how to do it from the docs.
18:51 Rick1 I added to the master a section for reactors
18:51 UtahDave ok.  so what are you wanting to do with the Reactor, Rick1?
18:52 Rick1 under which I did
18:52 Rick1 reactor:
18:52 Rick1 /srv/reactor/start.sls
18:52 Rick1 /srv/reactor/tmpclean.sls
18:52 Rick1 but It is not working for either.
18:52 UtahDave Rick1: which tag are you matching on?
18:52 Rick1 start and tmpclean
18:53 UtahDave So does your config look functionally like this?  http://docs.saltstack.com/topics/reactor/index.html?highlight=reactor#mapping-events-to-reactor-sls-files
18:53 santagada joined #salt
18:54 mgw joined #salt
18:54 Rick1 Yes.
18:55 UtahDave have you run the eventlisten.py script and verified that 'start' and 'tmpclean' are coming through the event bus?
18:55 Rick1 No I will try that now.
18:55 UtahDave cool, it's in the tests directory in the repo
18:55 jdenning Rick1: Do the lines look like "/srv/reactor/start.sls"  -OR- "- /srv/reactor/start.sls" ?
18:55 Rick1 /srv/reactor/start.sls
18:55 jdenning I think you need the leading '-'
18:57 jdenning or am I mistaken?
18:58 nrub joined #salt
18:58 jdenning UtahDave: on the subject of the event system - is there a complete list of all event types/tags that salt uses internally somewhere?
18:58 jdenning (even if it's in the source code, a pointer would be helpful :) )
18:58 basepi_tmp joined #salt
19:00 samsalt joined #salt
19:01 Jarus Is there a way to change the LANG env inside a pkg.installed state? This would fix the bug that postgresql chose SQL_ASCII instead of UTF8. Is there a correct LANG env with UTF8 the installation would create the template with UTF8.
19:02 napperjabber joined #salt
19:03 UtahDave jdenning: yes, it does need the   -
19:03 kleinishere joined #salt
19:03 samsalt Hi
19:04 UtahDave jdenning: There isn't a canonical list yet.  Basically, until there is, I just suggest running the eventlisten.py script and watching what tags blow by
19:04 MTecknology Fucking amazing... 1. deploy esxi server (deploys bh server automatically) 2. boot bh server 3. bh server comes on and tries to connect to salt master, fails 4. reactor takes over, deals with failed/rejected authentication 5. reactor logs into minion and forces salt-minion to restart 6. authentication suceeds 7. reactor logs into minion and handles decrypting the encrypted partitions 8. reactor
19:04 Rick1 UtahDave this is the error I get everytime I uncomment the reactor in the master.
19:04 Rick1 rick@eagle:/etc/salt $ Process Reactor-4:
19:04 Rick1 Traceback (most recent call last):
19:04 Rick1 File "/usr/lib64/python2.6/multiprocessing/process.py", line 232, in _bootstrap
19:04 Rick1 self.run()
19:04 Rick1 File "/usr/lib/python2.6/site-packages/salt/utils/event.py", line 411, in run
19:04 Rick1 reactors = self.list_reactors(data['tag'])
19:04 Rick1 File "/usr/lib/python2.6/site-packages/salt/utils/event.py", line 368, in list_reactors
19:04 Rick1 for ropt in react_map:
19:04 Rick1 UnboundLocalError: local variable 'react_map' referenced before assignment
19:04 MTecknology tells minion to run state.highstate 9. state.highstate runs without troubles and the entire system is 100% configured for whatever location it was deployed at 10. smiles
19:04 jschadlick joined #salt
19:04 MTecknology Rick1: dpaste.com next time
19:04 UtahDave hi, samsalt!
19:05 UtahDave Rick1: what version of Salt are you on?
19:05 Rick1 15.3
19:05 Rick1 MTecknology yep sorry.
19:06 UtahDave Rick1: did you restart the salt-master service?
19:06 MTecknology Getting the reactor to work for you is a little bit tough, but holy crap it's amazing.
19:06 Rick1 That is when it blows its top.
19:07 UtahDave Rick1: do you have a    -   before your reactor file listings?
19:07 Rick1 I comment the reactor setting out and restart master.
19:07 jdenning MTecknology: What are you using for step #5 & #7?  SSH/fabric ?
19:08 MTecknology jdenning: ssh -- i'll show
19:08 UtahDave Rick1: can you pastebin your reactor config?
19:09 Rick1 From the master config?
19:09 UtahDave yeah
19:09 MTecknology jdenning: http://dpaste.com/1254427/
19:10 Rick1 yep
19:10 jdenning MTecknology: Awsome - thanks!
19:11 kleinishere joined #salt
19:11 Rick1 http://pastebin.com/0jrEpGd2
19:11 MTecknology jdenning: it's not 100% safe, but... it's soooo much more secure than what we've had in place. I also have a state that changes the root password so that will only work until it's been able to authenticate and run the highstate.
19:12 UtahDave Rick1: yeah, that's malformed. Let me respond
19:12 MTecknology Rick1: You want something more like this - http://dpaste.com/1254429/
19:12 Rick1 Thank you.
19:12 jdenning MTecknology: Well a perfectly secure room has no doors or windows :)
19:12 nrub joined #salt
19:12 UtahDave Rick1: something like this: http://pastebin.com/yw0KYNCx
19:13 MTecknology UtahDave: I really should go through and update that reactor documentation... it's absolutely amazing, but the current docs are a bit hard to follow..
19:13 MTecknology UtahDave: is there a list of available tags anywhere?
19:14 UtahDave MTecknology: no, there isn't a list of tags, unfortunately.
19:14 MTecknology I'll try to build that too then
19:14 UtahDave MTecknology: dude, that would be SO awesome if you could help update those docs.
19:14 MTecknology I'll do my bestest. :)
19:15 MTecknology I'm getting a bit teary eyed seeing this project I've been working on finally coming to completion. Soooooo much work went into this...
19:15 MTecknology so many salt bugs we've unearthed and resolved and so much documentation created.. such a headache I've been...
19:16 jdenning MTecknology: I hadn't looked into the wheel system previously, thanks - that's an awesome way to manage keys..
19:16 jdenning (I've been using a shell script :) )
19:17 MTecknology not just teary eyed.. I've now experienced tears of joy
19:18 UtahDave MTecknology: You've been awesome helping work through a bunch of stuff.
19:18 UtahDave :)
19:19 MTecknology hm.. the reactor doc is much much more developed than the last time I looked at it. nice
19:21 MTecknology salt-call state.highstate 2>&1 >temp   <-- temp is 5,727 lines
19:22 MTecknology and now to check it out line by line to make sure everything really executed as expected...
19:24 UtahDave wow
19:24 jdenning MTecknology: !!!
19:26 santagada joined #salt
19:27 gadams Only is SALT supported VMware
19:27 gadams s/is/if
19:28 SEJeff_work MTecknology, headache means you've helped improve salt for everyone :D
19:32 dthom91 joined #salt
19:32 UtahDave gadams: do you mean salt-cloud?
19:34 LyndsySimon joined #salt
19:36 basepi_tmp left #salt
19:43 dthom91 joined #salt
19:43 aleszoulek joined #salt
19:44 nliadm joined #salt
19:50 andrewclegg joined #salt
19:53 drdran joined #salt
19:55 pentabular1 joined #salt
19:55 jschadlick joined #salt
19:55 ggoZ joined #salt
19:55 auser hey pentabular
19:56 pentabular1 hi auser!
19:58 jdaggett joined #salt
19:59 santagada joined #salt
20:00 krak3n` joined #salt
20:02 jdaggett joined #salt
20:08 stas joined #salt
20:14 magmatt1 joined #salt
20:17 napperjabber joined #salt
20:22 magmatt1 If I don't have a minion daemon running, is there a way to manually "run" some state files?  Like "salt-call --local my/custom/state.sls" ?
20:23 UtahDave magmatt1: yep.   put your state files in your file_roots
20:23 billyw joined #salt
20:23 UtahDave magmatt1: then run   salt-call --local state.sls  myslsfile
20:23 magmatt1 so, I can't just run some relative files?
20:24 magmatt1 relative to current directory, that is
20:24 UtahDave no, it always looks in file_roots
20:27 magmatt1 dang
20:32 MTecknology and now I have a state (two states) that keep vmware-tools updated
20:35 napperjabber joined #salt
20:35 dthom91 Is the file.recurse bug fixed in 0.15.3?
20:35 santagada joined #salt
20:36 MTecknology the one I dealt with was fixed a long time ago
20:36 dthom91 Meaning the one that causes source file failures that look like "Source file salt://|minion/conf-root/home/ec2-user/.py_vim not found"
20:36 MTecknology did you check the bug report?
20:37 MTecknology https://github.com/saltstack/salt/issues
20:37 dthom91 Yes, it looks like it was closed but I'm still seeing it with 0.15.3 minion and master
20:39 MTecknology both minion and master updated?
20:39 MTecknology what bug number?
20:40 MTecknology ahhh.... apparently things aren't 100% in the correct order
20:40 MTecknology I think a few require_in: lines will take care of that, though. :)
20:41 santagada joined #salt
20:42 MTecknology actually... this might be a bug... and if it is, it's going to be a real pain in the bum to test, prove, and fix.
20:43 magmatt1 UtahDave: and if the states depend on pillars?  Are pillars impossible to use in a masterless/minion-daemonless setup?
20:50 godber_ \q
20:51 linjan joined #salt
20:56 aranhoide joined #salt
20:58 pentabular1 left #salt
20:59 dthom91 joined #salt
21:01 jhauser joined #salt
21:01 dthom91 There is more than one bug against this behavior… it seems like it has been opened and closed a few times. https://github.com/saltstack/salt/issues/5013, https://github.com/saltstack/salt/issues/5423
21:02 Slipo joined #salt
21:03 opapo joined #salt
21:12 swa2 joined #salt
21:14 aberant joined #salt
21:16 LarsN_ UtahDave: Looks like the ball is at least rolling.
21:16 LarsN_ *ThumbsUp*
21:23 melinath joined #salt
21:31 bemehow_ joined #salt
21:33 aranhoide joined #salt
21:35 stat__ joined #salt
21:39 napperjabber joined #salt
21:52 magmatt1 left #salt
21:54 UtahDave LarsN_: nice!  thanks!
21:55 UtahDave dthom91: can you test against the latest in git?
21:56 Slipo joined #salt
21:56 dthom91 I can, but my hosts pull Salt RPMs from epel
21:57 dthom91 do you want me to test against a particular branch, or develop?
21:57 UtahDave develop
21:57 UtahDave I'm just trying to figure out if the issue you're seeing has been fixed in develop or not
21:58 dthom91 gotcha. I can give it a shot in a bit. working on a staffing proposal at the moment >_<
21:58 Corey dthom91: Anyplace exciting? :-)
21:59 dthom91 Corey: we like to think so :D
22:03 UtahDave :)
22:03 UtahDave cool, let me know how it goes, dthom91
22:05 fragamus hey auser
22:08 maspwr joined #salt
22:09 jeffrubi` left #salt
22:09 Slipo joined #salt
22:10 abe_music any EC2 users out there having issues starting in the us-east region?
22:10 kermit joined #salt
22:12 jkleckner joined #salt
22:17 LyndsySimon joined #salt
22:18 gerry__ joined #salt
22:19 wifigi joined #salt
22:21 wifigi I have an issue restarting the salt-master service using 'service salt-master restart' on Ubuntu Server 12.04 LTS - it works once, then the next time it prints 'stop: Job has already been stopped: salt-master' and from then on any attempt to run that command just hangs forever.
22:21 wifigi I am using salt 0.12.1
22:22 wifigi It does the same thing using /etc/init.d/salt-master restart
22:23 UtahDave wifigi: try running this:       sudo salt-minion -l debug
22:23 UtahDave what's the output of that?
22:26 wifigi It waits for the keys to be accepted by the master
22:27 wifigi My problem is with the master only, not the minions
22:27 UtahDave wifigi: ah, ok.   try     sudo salt-master -l debug
22:28 wifigi [INFO    ] Loaded configuration file: /etc/salt/master [WARNING ] Setting up the Salt Master [WARNING ] Unable to bind socket, this might not be a problem. Is there another salt-master running? The ports are not available to bind
22:28 UtahDave ok, try stopping the master, then.     sudo service salt-master stop
22:29 wifigi It says 'stop: Job has already been stopped: salt-master'
22:29 wifigi From the output of ps aux I can see lots of salt-master processes
22:29 wifigi It doesn't appear to stop any of them
22:30 wifigi The debug still gives the same output after running the stop command
22:32 UtahDave try   sudo pkill salt-master
22:33 wifigi That killed all the master processes. but now 'service salt-master start' and 'service salt-master restart' just hang indefinitely - I can't restart it
22:34 UtahDave What changes have you made to the salt-master config?
22:34 wifigi Running salt-master -l debug stops on [INFO    ] Worker binding to socket ipc:///var/run/salt/master/workers.ipc
22:34 wifigi I've uncommented out the default file server root (/srv/salt)
22:34 wifigi That's it
22:35 UtahDave Hmm.
22:35 UtahDave after killing all the salt-master processes, run    sudo salt-master -l debug
22:35 UtahDave what's the output of that?
22:37 wifigi See 3 comments ago
22:38 EntropyWorks so is anyone using something similar to librarian-chef or berkshelf for salt recipes?
22:38 wifigi For a bit more output, see http://pastebin.com/HEJwQZR7
22:42 octarine joined #salt
22:42 jhauser_ joined #salt
22:44 UtahDave wifigi: I'm not sure. That's pretty weird
22:46 pjs joined #salt
22:47 mjulian joined #salt
22:47 mjulian joined #salt
22:47 wifigi UtahDave: I can restart to bring salt-master back up, but that's the only way I've found
22:48 UtahDave I'm not sure what's going on there, wifigi. That's really odd.  I run Ubuntu 12.04 and I haven't run into that ever.
22:48 UtahDave can you test on another machine?
22:48 [vaelen] joined #salt
22:48 wifigi Yeah, I have another salt-master that's working find on the same setup of Ubuntu
22:49 wifigi It's just the second VM image (which I thought was the same as the first) that seems to die
22:50 wifigi I can probably work around or rebuild that image I was just wondering if anyone had more ideas about causes or fixes
22:50 wifigi I don't know if/when it will happen again
22:50 UtahDave yeah, I don't know.  If you can figure out how to reproduce it or cause it, that would be a big help
22:51 Corey wifigi: Can you get it to a hung state?
22:51 jhauser_ joined #salt
22:51 Corey wifigi: If so, I'd like to see both ps auxwwf as well as lsof output.
22:51 wifigi Actually this VM is running a SIP softphone...would that be taking the master's ports for any reason?
22:51 premera joined #salt
22:51 Corey wifigi: 4505 and 4506, so probably not?
22:51 wifigi I can easily get it to a hung state
22:52 wifigi That's what I thought
22:52 wifigi SIP uses 5060
22:52 Corey It looks like something's getting stuck around the ICP stuff.
22:52 Corey Hence lsof as a first pass.
22:52 wifigi Other than the phone and salt this is just the Ubuntu iso, nothing else
22:52 wifigi I'll try and get you that output
22:52 Corey wifigi: And there's nothing dodgy in dmesg?
22:53 Corey (UtahDave is second to none when it comes to the code, but I like to play when the systems get screwy!)
22:53 wifigi Corey: Nothing other than that is was terminated with status 2
22:54 wifigi *it, meaning salt
22:54 Corey wifigi: Can we broaden that to "non-zero?"
22:54 wifigi Broaden what?
22:55 chrisgilmerproj left #salt
22:55 wifigi I don't quite follow...
22:56 Corey wifigi: Well, exit status 2 is useful, but there are other error codes; the only one we can rule out is 0. :-)
22:56 auser hey
22:56 auser back
22:56 wifigi That's all that I can see in dmesg about salt, I don't know about any other error codes popping up at the moment
22:56 conan_the_destro joined #salt
22:57 conan_the_destro joined #salt
23:01 wifigi Corey: here is the ps output: http://pastebin.com/pjXd4d5L
23:01 wifigi When the system is hugn
23:01 wifigi *hung
23:03 wifigi And the lsof: http://pastebin.com/na43PspQ
23:05 wifigi I have to sign off - if anyone has any random bursts of inspiration, shoot an email to my throwaway account: hidenomore@mail.com
23:05 bensix2 joined #salt
23:12 adotbrown joined #salt
23:19 aat joined #salt
23:29 fragamus joined #salt
23:34 dthom91 joined #salt
23:35 cxz joined #salt
23:35 oz_akan_ joined #salt
23:43 Thiggy joined #salt
23:45 Thiggy Hello. My googling is failing me. Is there some way to determine the date/time of the last highstate run for a particular minion?
23:47 adotbrown joined #salt
23:49 aat joined #salt
23:49 cxz Thiggy: i would like to know this too!
23:52 mianos joined #salt
23:52 mianos hello. Does anyone have any ideas for running a multiple server sharded database backup/export using salt
23:52 Thiggy Hrmm, I suppose if I wanted this information I could just make something that writes the current timestamp to a file everytime there's a highstate call.
23:53 mianos I have 20 servers all managed by salt, I can select the database servers using a role
23:54 mianos fire of the backup with salt and use ssh to send them back to a known place?
23:54 jschadlick left #salt
23:54 jpadilla joined #salt
23:56 Corey mianos: Hi!
23:56 Corey You're playing my song! :-)
23:57 Corey mianos: Higher level, what're you trying to do specifically here?

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