Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-04-21

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

All times shown according to UTC.

Time Nick Message
00:01 catpig joined #salt
00:04 GabLeRoux joined #salt
00:07 solidsnack joined #salt
00:07 tedski having an issue with config.get where it's not returning expected results.  see http://dpaste.com/0088YNP
00:07 tedski the pillar data exists, but cannot be access via config.get
00:08 tedski the dpaste shows me using quotes, but i've tried with and without with no difference
00:08 tedski per http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.config.html, minion's pillar should be scanned for results
00:11 otter768 joined #salt
00:12 rhodgin joined #salt
00:14 murrdoc joined #salt
00:15 chrisportman should `salt-call saltutil.revoke_auth` on a minion, delete the /etc/salt/pki/master/minions/<hostname>.pem file on the master
00:23 shaggy_surfer joined #salt
00:25 murrdoc joined #salt
00:28 iggy tedski: did you do a sync_pillar ?
00:31 iggy chrisportman: the master doesn't have a copy of the minions private key... so no
00:31 chrisportman no but it has the public key.  Unless its removed, the minion cant rejoin the master with a new set of keys (e.g after a reinstall)
00:32 Vynce1 is there a more complete definition of http://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html#salt.states.file.line / http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.file.html#salt.modules.file.line ?  and what version is "Beryllium"?
00:33 arif-ali joined #salt
00:34 iggy Vynce: beryllium is 2015.8/10/12/whatever
00:34 iggy i.e. devel
00:35 iggy Lithium is the next release (that was due 2 months ago)
00:35 Vynce so things labeled "new in beryllium" are actually futuretech?
00:37 iggy if they were released, they'd say "new in 2014.7"
00:38 Vynce ok. is there a standard idiom for making sure a file contains a particular line that is possible now?
00:39 murrdoc joined #salt
00:39 c10b10 joined #salt
00:40 iggy file.append()
00:40 iggy it's weird, if it finds it anywhere in the file, it won't append it, but if it adds the line it's appended at the end
00:40 chrisportman hmmm.. it would seem that revoke_auth is deleting the cert from the master, but the command continues on communicating with the master, creating a public key because autosigning is enabled.
00:42 iggy yeah... expected behavior
00:45 Vynce iggy: ++ thanks. file.append will do very nicely.
00:49 chrisportman iggy: it makes sense.  With that in mind, is there a way of having a minion invoke the revocation of the certificate, so that it can be reinstalled (same hostname) and rejoin salt
00:54 ITChap joined #salt
00:57 druonysus joined #salt
00:57 druonysus joined #salt
00:59 boubou joined #salt
01:07 murrdoc joined #salt
01:13 kzrl joined #salt
01:13 dendazen joined #salt
01:13 rhodgin joined #salt
01:14 zwi joined #salt
01:18 subsignal joined #salt
01:18 notnotpeter joined #salt
01:24 SeeDickCode joined #salt
01:25 subsignal joined #salt
01:29 iromli joined #salt
01:30 catpig joined #salt
01:36 yomilk joined #salt
01:37 ajw0100 joined #salt
01:38 scbunn joined #salt
01:46 bhosmer joined #salt
01:46 jonlangemak joined #salt
01:49 murrdoc joined #salt
01:50 thayne joined #salt
02:02 dendazen joined #salt
02:04 zwi joined #salt
02:06 Nazca__ joined #salt
02:06 jonlangemak joined #salt
02:07 malinoff joined #salt
02:09 favadi joined #salt
02:14 rhodgin joined #salt
02:28 evle joined #salt
02:28 c10b10 joined #salt
02:33 drawsmcgraw joined #salt
02:37 Vynce apparently file.append doesn't create the file if the file doesn't exist. i'm sure this is a solvable problem, though.
02:39 vandemar joined #salt
02:41 michelangelo joined #salt
02:42 bhosmer joined #salt
02:44 Vynce is there more complete documentation for "require"? i see lots of talk about it but no actual explanation of what it does
02:46 drawsmcgraw Vynce: More than you asked for -> http://docs.saltstack.com/en/latest/ref/states/requisites.html
02:47 ITChap joined #salt
02:49 ajw0100 joined #salt
02:49 andrej joined #salt
02:49 enarciso joined #salt
02:49 chrisportman left #salt
02:49 Vynce it's not *more* than i want, but i have read it before and not gotten mch ou of it.  i know a little more salt now, and it's making a little mroe sense, but it's certainly not clear.
02:50 Vynce but it's stuff to go on.  thanks, drawsmcgraw.
02:51 drawsmcgraw Vynce: Sure thing. I understand the feeling. Personally, I only use 'require' and the occasional 'watch'
02:51 blacked1 joined #salt
02:51 drawsmcgraw Anything  that seems particularly opaque?
02:52 Vynce well, it doesn't actually say what the syntax os, or what sorts of things can go in a require statement, except it hints that it must be a state.
02:53 Vynce but, like, how that state match works… again, hints, but not clarity. same as the rest of the salt docs, really.
02:54 Vynce it's like learning a language via immersion rather than taking classes.  it works better for some people than others, and involves a lot of crying and begging and feeling stupid and helpless before the fundamental breakthrough comes
02:55 rickflare Vynce I feel the same way
02:55 rickflare ive done a lot in this channel
02:55 rickflare begging wise
02:55 rickflare but its helps to learn the hard way
02:55 drawsmcgraw This is one of my personal issues with the docs
02:55 rickflare i try not to make the same mistakes twice
02:56 Vynce like, i'm trying to file.append.  the target file doesn't exist.  i can create a file.exists state, which as i understand it even if i put it befor ethe file.append might not work.  so i have to require that the file exists — which can't jsut be require that the file exists, it has to be require that the file-exists-state has been satisfied. or something.
02:56 drawsmcgraw Vynce: Give me a moment and I'll have a mockup
02:56 Vynce but then the syntax for referring to the file exists state isn't just the ID of the state, it' … file: STATE_ID ?
02:56 Vynce i think i know what to type.  i jsut don't understand all of why
02:57 Vynce and i *hate* not understanding.
02:58 drawsmcgraw http://dpaste.com/14J1KEN
02:59 drawsmcgraw And yes, that's correct.
02:59 drawsmcgraw In the example I pasted, the require statement means "I require that the file() function call in the state named 'ensure-my-file' completed successfully."
03:00 Vynce but why does it have to be file:ensure-my-file instead of jsut require: ensure-my-file ?
03:00 drawsmcgraw If you were to rename 'ensure-my-file' to something like 'donkey-config', you would need to also update the state below it to say 'file: donkey-config'
03:00 Vynce also, what will this do if the file doesn't exist?  will it touch it, and make it exist, or will the state jsut fail?
03:00 drawsmcgraw Because 'ensure-my-file' can have multiple function calls in it.
03:01 Vynce is there a way to say require: allof: ensure-my-file?
03:01 drawsmcgraw Good question. I'm actually not familiar with file.exists()
03:01 ITChap joined #salt
03:01 drawsmcgraw I've usually cheated and done a 'cmd.run' with a 'touch somefile'
03:01 Vynce in a state?
03:02 drawsmcgraw Not that I'm aware of, though if you *really* wanted that behavior, you could set up a sort of requisite chain
03:02 Vynce cmd.run seems dodgy in a state. but i dunno.
03:02 drawsmcgraw Which I can't really recommend....
03:02 drawsmcgraw It does, but you can safeguard cmd.run with a 'unless'
03:02 Vynce these are the soets of things that seem like they should be easy.
03:03 drawsmcgraw Vynce: I don't disagree. But (for better or worse) the thought process comes a lot easier once you've written a number of states
03:03 drawsmcgraw Here http://dpaste.com/3YE92YP
03:04 drawsmcgraw That's how I take care of some other commands as well, like formatting a filesystem
03:04 drawsmcgraw When the native Salt modules can't give you idempotence, I rig it up with an 'unless' parameter.
03:04 favadi joined #salt
03:05 Vynce that feels wrong. i want to beleive there's a better way.  i also want to beleive we're hiring someone else to do ops soon, but i suspect i'm stuck with it for a while. \ :
03:05 drawsmcgraw So... you had mentioned that the file.exists() call couldn't be guaranteed to run before the file.append() call?
03:06 Vynce it hink?  as i understand it, that's why -require: exists
03:06 drawsmcgraw yes and no
03:06 drawsmcgraw Back in the day, states were not executed top-down
03:07 drawsmcgraw Today, though, if a state appears above another state, and you haven't written any requisites to modify its behavior, the first state will run before the ones below it.
03:07 drawsmcgraw So all of that to say -> I don't think you really need a requisite at all here
03:07 drawsmcgraw Just have a state to file.exists(), then, below that, have one to file.append()
03:08 Vynce i mean, i jsut found the -order option, but it also seems gross
03:08 yexingok joined #salt
03:08 drawsmcgraw Fully agree. I stay very far away from the  'order' parameter.
03:09 drawsmcgraw You may be doing what I find myself doing often - overthinking it :)
03:09 Vynce "state_auto_order' seems like it's broken and backwards.  "true" means don't order them for me, trust me. "false" means don't worry what order i have them in, sort them (alphabetically?  ugh.)
03:10 Vynce well, i'm trying to make sure i don't leave a mess for that putative future hire.
03:10 Vynce and i don't want to leave something painful.  i mean, the person working on dev ops in three months might be somebody i really like.  it might be me.
03:11 MatthewsFace joined #salt
03:11 drawsmcgraw I would think that setting 'state_auto_order' to true would mean "Yes, please auto-order them"
03:11 drawsmcgraw Although... yeah, I can see your point.
03:11 drawsmcgraw It brings up painful memories of Puppet "auto-ordering" things :/
03:11 zzle joined #salt
03:12 drawsmcgraw regardless, in the default configuration, states are executed top-down.
03:12 Vynce i have no experience with puppet. ( :
03:12 enarciso joined #salt
03:12 drawsmcgraw Mine is minimal - and that's on purpose :)
03:13 Vynce (yes, "auto_order = true" means "please auto-order them" — the question is what does auto-ordering mean?  i would argue that doing them in the order provided isn't ordering them, it's leaving them in order. sorting them by some other criterion would be ordering them — and "auto" *should* be something "smarter" than lexical, but whatever.)
03:15 drawsmcgraw I agree, the verbiage is a bit clumsy
03:16 Vynce ugh.  "Too many functions declared in state "file" in sls shared"
03:16 scbunn joined #salt
03:18 drawsmcgraw Vynce: Mind pasting it?
03:20 Vynce https://gist.github.com/VynceMontgomery/ba306ca76c845e797b24
03:21 Vynce ugh.  missing name from file.exists.
03:21 drawsmcgraw so
03:21 drawsmcgraw line 8
03:21 Vynce lovely opaque error emssage for that
03:21 drawsmcgraw Make it look like this --> name: /root/.bash_profile
03:21 drawsmcgraw file.exists() is a function. You need to pass it the 'name' parameter.
03:22 Vynce and that leads to the answer: no, exists will not touch it.  it will jsut fail.
03:22 drawsmcgraw Explicitly, that is :) There are no positional arguments in states
03:22 drawsmcgraw interesting
03:22 Vynce so require: file: some_file_exists is redundant and useless
03:22 drawsmcgraw So it only confirms/denies that a file is there. It won't make it
03:23 Vynce i wasn't expecting positional, i jsut typod.  i mean, it seems legit that a function that only takes one argument wouldn't need it named, but i wasn't expecting that.
03:23 drawsmcgraw ah, okay
03:23 Vynce note that i spotted it right after pastinf
03:23 Vynce s/nif/ing/; #erf
03:23 drawsmcgraw gotcha. Sorry
03:23 Vynce no worries
03:23 jespada joined #salt
03:24 drawsmcgraw so... would a file.managed() be too aggressive for this case?
03:24 Vynce helpful is good
03:24 Vynce yeah, cause i don't want to replace it.
03:25 jespada joined #salt
03:25 Vynce i guess i coulg managed replace: False with a blank file, but that feels pretty clumsy
03:25 drawsmcgraw I've had cases like this
03:25 drawsmcgraw I just went with a cmd.run() 'touch /some/file'
03:25 Vynce i guess that's the ebst thing i've got so far, though.
03:25 drawsmcgraw It's kludgy but, hell, everything is kludgy if you dig down far enough
03:26 Vynce i prefer to keep the kludge either under glass where i can keep an eye on it and shoot it, or below the basement, where it's teh city's problem, not mine
03:26 jespada joined #salt
03:26 rickflare left #salt
03:26 drawsmcgraw heh, you should see some of the abominations I've devised
03:27 drawsmcgraw But actually, passing replace=False seems more elegant than my solution
03:27 zzle Does anyone know how to get Salt running on Amazon Linux 2015.03?  It looks like since they changed the default Python from 2.6 to 2.7, Salt won't run when installed through either the bootstrap script or yum/EPEL.  It just gives the Python error "ImportError: No module named salt.scripts".
03:28 drawsmcgraw zzle: strange... Can't say I've seen that one before...
03:28 seev zzie, use centos
03:28 seev amazon linux is pure crap
03:31 zzle Is there an easy way to get Salt to use /usr/bin/python26 instead of /usr/bin/python?  That may fix it.
03:33 drawsmcgraw zzle: May want to go with installing Salt into a virtualenv or something similar?
03:36 bhosmer joined #salt
03:42 rickflare joined #salt
03:49 echo joined #salt
03:50 Vynce OK, so test=True is failing, at the file.append, saying the file is not found — but it's also saying — directly above that, in fact — that it would create the file.  is test=True not quite smart enough to notice that fiels it would create would then exist?
03:51 drawsmcgraw Vynce: correct :)
03:51 drawsmcgraw There's also state.show_sls for another version of a dry run
03:53 Vynce OK, so i ran it… with test=False *explicitly* … and one of the servers returns a "file is set to be updated" and then fails to append because it does not exist, looking ofr all teh world like test=True output
03:53 Vynce so.. hwo do i debug that?
03:54 zzle Thanks, drawsmcgraw, but I was hoping to not stray too far from the default recommended installation procedures.
03:55 drawsmcgraw zzle: I can understand that. I don't have anything better unfortunately :/
03:55 Vynce i thought salt installed its own python, Just In Case.
03:55 Vynce … and i jsut ran it wirh test=0 and it JFW.  JFSM.
03:55 zzle Reading /etc/rc.d/init.d/salt-master, it looks like you can override the PYTHON env variable by setting it in /etc/default/salt, but that doesn't seem to work for me.
03:56 drawsmcgraw wierd.. I've never given anything other than 'true' to the test parameter
03:56 drawsmcgraw But if ti work... :P
03:56 Vynce i have my salt minions now set to default to test=True, and it requires explicitly test=False* to make them do teh thing
03:56 Vynce * except this time, apparently, it required test=0.
03:57 Vynce * which i really don't understamd.  the other difference was that i called salt for *jsut* that minion, instead of its whole group.
04:00 drawsmcgraw That's a nice mechanism. You *really* don't want to accidentally break something
04:05 Vynce yup. I've wondered before why it's not more heavily promoted
04:17 c10b10 joined #salt
04:20 shaggy_surfer joined #salt
04:21 subsignal joined #salt
04:26 pravka joined #salt
04:28 otter768 joined #salt
04:30 bhosmer joined #salt
04:31 TyrfingMjolnir joined #salt
04:32 mosen joined #salt
04:39 thayne joined #salt
04:39 Vynce ugh.  that same host claimed to highstate OK but didn't actually.
04:42 rdas joined #salt
04:42 shaggy_surfer joined #salt
04:43 ingwaem joined #salt
04:44 ingwaem left #salt
04:47 enarciso joined #salt
04:47 Vynce ah.  when i thought i was highstating it, i was highstating a different machine.  OK.  so, how do you debug a machine that acts like you're running test=True when you're actually runnign test=False ?
04:57 jalbretsen joined #salt
04:57 jhauser joined #salt
04:57 blacked joined #salt
05:01 ckao joined #salt
05:09 huleboer joined #salt
05:16 rhodgin joined #salt
05:17 blacked joined #salt
05:25 bhosmer_ joined #salt
05:26 catpigger joined #salt
05:32 sd2121wqq2 joined #salt
05:33 c10b10 joined #salt
05:43 baweaver joined #salt
05:43 baweaver joined #salt
05:46 stoogenmeyer_ joined #salt
05:52 SpeedFuse joined #salt
05:53 baweaver joined #salt
05:53 krelo joined #salt
06:05 krelo joined #salt
06:11 Nazca joined #salt
06:14 kwork joined #salt
06:15 thayne joined #salt
06:17 harkx joined #salt
06:17 rhodgin joined #salt
06:18 viq joined #salt
06:19 bhosmer_ joined #salt
06:21 subsignal joined #salt
06:22 rdas joined #salt
06:28 flyboy joined #salt
06:29 scarcry joined #salt
06:39 otter768 joined #salt
06:42 favadi joined #salt
06:58 Furao joined #salt
07:02 hebz0rl joined #salt
07:03 isodude_ joined #salt
07:04 ndrei joined #salt
07:06 KermitTheFragger joined #salt
07:06 ndrei joined #salt
07:07 Romlok joined #salt
07:08 eseyman joined #salt
07:13 bhosmer joined #salt
07:19 m0nky joined #salt
07:21 arif-ali joined #salt
07:21 c10b10 joined #salt
07:22 kawa2014 joined #salt
07:22 c10b10 joined #salt
07:23 favadi joined #salt
07:27 Hell_Fire joined #salt
07:30 Hell_Fire joined #salt
07:31 ITChap joined #salt
07:35 CeBe joined #salt
07:38 funzo joined #salt
07:40 teogop joined #salt
07:43 lb1a joined #salt
07:48 blacked joined #salt
07:54 jrenner joined #salt
07:55 ktosiek joined #salt
07:56 slav0nic joined #salt
07:56 slav0nic joined #salt
07:58 thayne joined #salt
07:59 bluenemo joined #salt
08:01 fredvd joined #salt
08:07 bhosmer joined #salt
08:11 favadi http://docs.saltstack.com/en/latest/topics/event/master_events.html#job-events
08:11 favadi anyone knows list of retcode for job event?
08:12 favadi I have a job event with retcode==32 but no state fails
08:13 hal58th joined #salt
08:15 heise joined #salt
08:20 Xevian joined #salt
08:20 rnts joined #salt
08:20 aquassaut joined #salt
08:25 rdas joined #salt
08:26 tobias3 joined #salt
08:26 cberndt joined #salt
08:27 cosming joined #salt
08:27 bersace left #salt
08:30 refnode joined #salt
08:40 otter768 joined #salt
08:48 analogbyte joined #salt
08:48 ITChap joined #salt
08:53 lcabrera joined #salt
08:56 faust joined #salt
09:00 abele joined #salt
09:02 bhosmer joined #salt
09:02 favadi left #salt
09:07 ndrei joined #salt
09:14 nocturn joined #salt
09:19 thehaven joined #salt
09:19 cberndt joined #salt
09:19 rhodgin joined #salt
09:22 cberndt joined #salt
09:23 wnkz joined #salt
09:26 subsignal joined #salt
09:39 peters-tx joined #salt
09:42 faust left #salt
09:43 matthew-parlette joined #salt
09:47 ekle joined #salt
09:47 ekle hi, is there a way to get the return of schedule.list as json instead of yaml ?
09:48 eliasp ekle: --output=?
09:48 favadi joined #salt
09:48 eliasp ekle: http://docs.saltstack.com/en/latest/ref/output/all/index.html
09:49 ekle i get yaml inside of json
09:49 eliasp uh, oh… ok… no idea, sorry
09:49 ekle salt '*' schedule.list --out=json
09:56 bhosmer joined #salt
09:59 johnzorn joined #salt
10:00 johnzorn In salt how would I deal with all the nuances of all the distros if I wanted my state files to work across the main distros?
10:04 _1_Steven joined #salt
10:07 munhitsu joined #salt
10:07 supermike__ joined #salt
10:07 TheNumb joined #salt
10:08 joeyparsons joined #salt
10:11 denys joined #salt
10:13 matti joined #salt
10:13 babilen johnzorn: That is typically handled in an "os_map". You might want to look at formulas and, in particular, https://github.com/saltstack-formulas/salt-formula/ and https://github.com/saltstack-formulas/collectd-formula/ to get an idea of how this is done. (check the defaults.yaml, map.jinja and other files referenced therein)
10:14 babilen http://docs.saltstack.com/en/latest/topics/best_practices.html has some information on that and other "best practices"
10:15 matti left #salt
10:18 reabeginner joined #salt
10:20 rhodgin joined #salt
10:20 TheNumb o/
10:30 cosming joined #salt
10:33 cygnetix joined #salt
10:35 giantlock joined #salt
10:37 yomilk joined #salt
10:38 mortis joined #salt
10:40 otter768 joined #salt
10:45 mindscratch joined #salt
10:46 mindscratch anyone know how I can get the following to work: salt '*' cmd.run 'cat /proc/$(pidof some_executable)/environ'
10:47 mindscratch I get TypeError encountered executing cmd.run: run() takes at least 1 argument
10:50 bhosmer joined #salt
11:00 cygnetix joined #salt
11:01 jakubm mindscratch: check minion log
11:01 jespada joined #salt
11:01 jakubm mindscratch: i'd go with 'sh -c "cat /proc/$(pidof foo)/environ"'
11:01 jakubm mindscratch: or something along these lines
11:02 mccricardo joined #salt
11:02 mccricardo left #salt
11:03 mindscratch jakubm - that worked thanks
11:04 jakubm mindscratch: you're welcome
11:08 cygnetix joined #salt
11:10 bhosmer joined #salt
11:12 PI-Lloyd hey guys, probably a stupid question, but is there a way to do something like "{% if grains['id'].endswith('dev'|'stage'|'prod') %}" in a state? can't seem to figure it out and can't find any docs on it
11:12 echo joined #salt
11:13 PI-Lloyd {% if grains['id'].endswith('dev') or grains['id'].endswith('stage') or grains['id'].endswith('prod') %} is really messy imo
11:14 XenophonF joined #salt
11:15 XenophonF can i use s3 urls in windows package repositories?
11:17 XenophonF if i use salt urls in windows package repositories, which environment do they use?
11:21 rhodgin joined #salt
11:26 Nazzy joined #salt
11:27 otter768 joined #salt
11:31 cygnetix joined #salt
11:38 jonlangemak joined #salt
11:47 rhodgin joined #salt
11:48 amcorreia joined #salt
12:03 isodude_ joined #salt
12:04 EvaSDK joined #salt
12:06 ndrei joined #salt
12:06 bhosmer joined #salt
12:06 isodude_ joined #salt
12:06 yomilk joined #salt
12:07 EvaSDK joined #salt
12:08 mapu joined #salt
12:11 _JZ_ joined #salt
12:12 evle1 joined #salt
12:13 xf10e joined #salt
12:15 wWwBUKOLAYcom joined #salt
12:18 yidhra joined #salt
12:23 tmclaugh[work] joined #salt
12:31 mapu joined #salt
12:32 xf10e ha! removed a '- ' (list -> dict) from my pillar and keystone.user_present() works again! (https://github.com/saltstack/salt/issues/22782)
12:32 cmcmacken joined #salt
12:34 xf10e still a minor regression I think as the functions behavior changed from 2014.7.3 to 2014.7.4
12:40 fredvd joined #salt
12:45 dendazen joined #salt
12:46 yomilk joined #salt
12:47 GabLeRoux joined #salt
12:51 hasues joined #salt
12:51 hasues left #salt
12:53 cberndt joined #salt
12:54 denys joined #salt
12:55 EvaSDK joined #salt
12:55 EvaSDK joined #salt
12:58 JDiPierro joined #salt
12:58 redzaku joined #salt
12:59 drawsmcgraw1 joined #salt
13:01 bhosmer joined #salt
13:04 racooper joined #salt
13:09 Tecnico1931 joined #salt
13:11 FeatherKing joined #salt
13:11 malinoff joined #salt
13:12 bhosmer joined #salt
13:17 EvaSDK joined #salt
13:17 EvaSDK joined #salt
13:20 fusionx86 joined #salt
13:21 mpanetta joined #salt
13:23 sk_0 joined #salt
13:23 chiui joined #salt
13:26 bhosmer joined #salt
13:28 debian112 joined #salt
13:28 timoguin joined #salt
13:29 subsignal joined #salt
13:30 andrew_v joined #salt
13:30 jcsp joined #salt
13:31 naboulank joined #salt
13:34 cpowell joined #salt
13:34 isodude_ I want to nest sls files,  so that I can have a user.sls which does all the goodies, and a users.sls file that runs a foreach on a mysql pillar.
13:35 isodude_ I did some investigation, and this seems impossible!
13:35 isodude_ my main issue is sending parameters to the user.sls from users.sls
13:35 isodude_ this is all states
13:35 isodude_ any clues?
13:36 jdesilet joined #salt
13:36 aparsons joined #salt
13:37 ekle use jinja
13:37 ekle and macros
13:37 ekle works fine here
13:38 edrocks joined #salt
13:40 bhosmer joined #salt
13:42 fredvd joined #salt
13:42 elfixit joined #salt
13:45 giantlock joined #salt
13:48 bootet joined #salt
13:49 GabLeRoux_ joined #salt
13:50 otter768 joined #salt
13:51 JDiPierro joined #salt
13:52 zwi joined #salt
13:55 isodude_ ekle: you're a lifesaver right now.
13:56 ekle thx
13:59 kaptk2 joined #salt
13:59 ikarpov joined #salt
13:59 masterkorp hello
13:59 masterkorp i just update salt-minion on the ubuntu repos
13:59 edrocks joined #salt
14:00 masterkorp ImportError: No module named cli.caller
14:01 nicksloan are there any known memory leaks affecting 2014.7?
14:03 dyasny joined #salt
14:06 manfred masterkorp:  make sure there aren't any left over .pyc files
14:06 wiloyee joined #salt
14:07 masterkorp manfred: how do I know that ?
14:07 masterkorp https://github.com/saltstack/salt/issues/20924
14:07 masterkorp oh
14:07 masterkorp i even opened an issue on it before :)
14:08 giannello joined #salt
14:10 GabLeRoux joined #salt
14:10 manfred looks like they gave you the reasons
14:11 zwi joined #salt
14:11 giannello hey - is there any way to reduce GPG pillars impact on CPU usage?
14:11 giannello we have, well.....a bunch of them, and the master's CPU is going bananas
14:13 masterkorp gpg pillars ?
14:13 zircote joined #salt
14:14 iggy there are inherent tradeoffs for different levels of data security
14:14 giannello yes, masterkorp - gpg-encrypted pillars - http://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.gpg.html
14:15 babilen masterkorp: http://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.gpg.html
14:16 jasonrm joined #salt
14:18 giannello iggy, indeed - but decoding 189 gpg blocks is a _little_ cpu-intensive, on my poor t2.small instance :D
14:18 iggy use less gpg or bigger instances
14:19 manfred http://i.imgur.com/sy8p8iq.jpg
14:19 masterkorp babilen: hmm
14:19 masterkorp thanks you both
14:20 masterkorp *thank you
14:21 Brew joined #salt
14:22 Brew joined #salt
14:22 m0nky hello =)
14:23 TyrfingMjolnir joined #salt
14:25 giannello hm, iggy - got that. what about avoiding pillars to be calculated every time we trigger simple salt modules, like "grains.setval"?
14:25 m0nky Looking for some insight, I have a  `salt-call state.highstate` that runs via a cronjob (every hour) and never exits on any minion, therefore consuming high cpu as additional jobs run each hour. This started happening after I added a state that syncs a fairly large file. Debug shows the sync completing however. =\
14:28 iggy giannello: it'd be possible, but it would take a lot of refactoring
14:28 xf10e giannello: I would say "use ec keys instead of large RSA ones" but you'll need GPG >= 2.1 for this: https://www.gnupg.org/faq/whats-new-in-2.1.html#ecc
14:30 xf10e m0nky: do you have any state watching this file? might trigger an accumulation of unfinished jobs
14:30 timoguin_ joined #salt
14:30 iggy m0nky: run the highstate through the scheduler and it'll make sure only one at a time is running
14:31 mnml_ joined #salt
14:31 giannello xf10e, that would be great, thanks. maybe also not using a 4k key will help mitigate the problem
14:31 markm_ joined #salt
14:32 xf10e giannello: you won't need a 4k one when using ECC ;)
14:32 klyed2 joined #salt
14:34 totte joined #salt
14:38 giannello xf10e, can't wait to start compiling my own packages for it :D
14:43 favadi joined #salt
14:44 favadi left #salt
14:48 numkem im working on a state written with the pyobjects renderer and I'm getting the "TypeError: unhashable type: 'list'" error... Any common ways of dealing with that? The error is in the context attribute of a file.managed
14:49 klyed2 Is sallt-call --local supported on windows?
14:50 ajmath joined #salt
14:52 jakubm where did the zeromq-4.0.5-1.el6.x86_64 go..? it was in COPR yesterday and now it's gone
14:53 bhosmer joined #salt
14:54 bluenemo joined #salt
14:54 jakubm https://copr-be.cloud.fedoraproject.org/results/saltstack/zeromq4 here it went
14:54 jakubm am I the only one finding the repository path naming confusing?
14:54 jakubm https://copr-be.cloud.fedoraproject.org/results/saltstack/zeromq4/epel-6-x86_64/zeromq-4.0.5-1.el5/
14:57 jalbretsen joined #salt
15:00 numkem or is it even possible to pass a list using pyobjects? It works great using yaml
15:02 hobakill joined #salt
15:02 timoguin joined #salt
15:03 mansquid joined #salt
15:04 xf10e numkem: sure it is. you're probably passing a list where a dict is expected
15:06 numkem xf10e: *facepalm* I wasn't quoting my dict keys... bad habbit from other languages
15:06 numkem xf10e: thx for lighting my lightbulb
15:08 xf10e numkem: you're welcome ;)
15:09 m0nky xf10e: I have a file.recurse on a directory that syncs
15:11 xf10e m0nky: but not a file.recurse state watching a file in the same dir, right?
15:11 conan_the_destro joined #salt
15:12 m0nky nope
15:12 m0nky just double-checked =)
15:15 toastedpenguin joined #salt
15:15 smcquay joined #salt
15:17 dalexander joined #salt
15:18 xf10e m0nky: hmmm... maybe you should just switch to something like rsync to only sync the parts of your big file that changed?
15:18 theag3nt joined #salt
15:18 m0nky but the file syncs without issue
15:18 m0nky the minion just hangs around after that
15:18 toastedpenguin joined #salt
15:18 m0nky I just upgraded all of minions to the latest stable
15:18 m0nky going to monitor to see
15:19 scbunn joined #salt
15:20 numkem fun fact: 'network.ip_addrs' doesn't take cidr kwargs on windows
15:20 clintberry joined #salt
15:21 m0nky xf10e: as an example, at job that started at 15:00:00 UTC, finishes. Completed state [/opt/program/dir] at time 15:01:04.522073
15:21 rm_jorge joined #salt
15:21 perrigo joined #salt
15:21 m0nky then salt-call just sits there
15:22 mansquid joined #salt
15:22 xf10e and never returns?
15:23 iggy numkem: think there's an issue about that
15:24 thayne joined #salt
15:24 m0nky correct
15:24 m0nky bbiab lunch time
15:25 chiui joined #salt
15:25 klyed2 Anyone perform windows based deployments of code? Did you just use a linux box and set up file sharing?
15:26 faust joined #salt
15:27 conan_the_destro joined #salt
15:27 iggy you mean do people run their masters on Linux? In most cases, yes... the master also (allegedly) mostly runs on freebsd
15:31 klyed2 iggy: I meant, windows currently doesn't have much support, so a lot of power shell scripts may have to be written. So, i was curious if people kind of "inverted" the way ansible works, ie the master handles checking out code and copying it over, rather then the master telling the windows machine to run a certain playbook...if that makes sense...
15:33 klyed2 if you do it this way then you kind of loose the ability to update multiple systems at the same time =/
15:33 timoguin joined #salt
15:35 iggy klyed2: I've never heard of people deploying code to mulitple Windows systems
15:35 iggy There could be better tools out there if all you are doing is deploying code to a bunch of (Windows) boxes
15:36 klyed2 iggy: ah, we have to here... one of our applications can only run once per windows install,and we need to spin up multiple servers to fan out the work load =/. We do have linux machines too,
15:37 desposo joined #salt
15:37 klyed2 *run one instance at a time
15:37 UtahDave joined #salt
15:38 jeremyalons We wrote a custom _state that handles copying from windows shares to our windows minions
15:39 jeremyalons Linux minions use rsync, windows minions use xcopy
15:39 jeremyalons it works for us
15:40 SeeDickCode joined #salt
15:41 gposton joined #salt
15:41 gposton looking for feedback on how to share modules w/ Salt https://groups.google.com/forum/#!topic/salt-users/nbuC2GYK3uQ
15:43 jeremyalons gposton: by modules, do you mean _modules or state definitions?
15:44 iggy gposton: watch all of Ryan_Lane's videos... they are doing something very similar (or exactly the same)
15:44 iggy fwiw, they just duplicate code where necessary
15:44 iggy (afair anyway)
15:45 UtahDave jeremyalons: cool
15:45 klyed2 jeremyalons: ah, thanks. You mind if I ask if you ran into real snags while using windows and ansible? I'm trying to do some research before i fully invest into it.
15:45 iggy *salt
15:46 stanchan joined #salt
15:48 dalexander joined #salt
15:49 perrigo left #salt
15:49 perrigo joined #salt
15:50 perrigo joined #salt
15:50 huddy joined #salt
15:50 adelcast left #salt
15:50 writtenoff joined #salt
15:51 otter768 joined #salt
15:54 gposton thanks iggy, I just hit him up on Twitter
15:56 jeremyalons klyed2: I've never used ansible.
15:57 fusionx86 joined #salt
16:00 ajw0100 joined #salt
16:01 jeremyalons IPC is minion side setting only right?
16:02 fusionx86 joined #salt
16:02 iggy gposton: he talked briefly about it in the "ordered" talk from this years saltconf... he had like 4 talks at saltconf this year
16:03 badon joined #salt
16:04 kawa2014 joined #salt
16:05 ja-s joined #salt
16:06 Tyrm joined #salt
16:07 fusionx86 joined #salt
16:07 xf10e bye
16:13 timoguin joined #salt
16:13 noobadmin joined #salt
16:15 KyleG joined #salt
16:15 KyleG joined #salt
16:15 noobadmin hello, everybody, I have a quick question: What type of packages the salt master sends to the minion? TCP I'm assuming? Can someone give me a link with that kind of info, please?
16:16 jakubm http://docs.saltstack.com/en/latest/topics/tutorials/firewall.html
16:16 jakubm noobadmin:
16:16 noobadmin jakubm: thank you!! :D
16:16 jakubm noobadmin: i guess by packages you meant packets
16:17 noobadmin jakubm: yes, sorry :P
16:17 jakubm noobadmin: no probs - hope that help[s
16:17 faust left #salt
16:18 noobadmin jakubm: thanks again
16:18 XenophonF noobadmin: are you asking about TCP ports or software packages?
16:19 GabLeRoux joined #salt
16:19 XenophonF noobadmin: http://docs.saltstack.com/en/latest/topics/tutorials/firewall.html
16:20 XenophonF noobadmin: the minion only needs to be able to establish outgoing connections to the master on TCP ports 4505 and 4506
16:20 jeffspeff how's 2014.7.5 working for everyone?
16:20 jakubm jeffspeff: much better than 7.1 so far
16:20 jakubm jeffspeff: but i can't provide any concrete data to support my claim
16:22 yidhra joined #salt
16:25 Guest70 joined #salt
16:26 FeatherKing anyone have experience with salt 2014.7 and a private docker registry? trying to use a docker.installed
16:30 Striki joined #salt
16:30 desposo joined #salt
16:31 subsignal joined #salt
16:34 MatthewsFace joined #salt
16:36 hasues joined #salt
16:36 hasues left #salt
16:36 iggy 2015.2 4lyfe
16:39 XenophonF is powershell the default windows shell?
16:40 kunersdorf joined #salt
16:41 spookah joined #salt
16:42 vandemar I think the standard command shell is?  'cmd'?
16:45 solidsnack joined #salt
16:52 murrdoc joined #salt
16:55 rap424 joined #salt
16:56 jngd joined #salt
16:59 XenophonF well strangely enough the "shell" grain on my windows minion is set to /bin/sh
17:00 murrdoc joined #salt
17:03 GabLeRoux joined #salt
17:03 UtahDave XenophonF: that's kind of funny, actually.
17:03 UtahDave XenophonF: what version of salt are you on?
17:04 XenophonF 2014.7.2 for the moment - I'm getting ready to upgrade to the new release
17:05 XenophonF hm, 2014.7.4 is the latest - is that correct?
17:06 Sketch *** Topic for #salt: Welcome to #salt | 2014.7.4 is the latest
17:07 murrdoc joined #salt
17:08 jcsp joined #salt
17:08 manfred but 2014.7.5 is tagged and being tested
17:13 timoguin joined #salt
17:21 edrocks joined #salt
17:22 XenophonF well 2014.7.4 doesn't work
17:22 XenophonF just tried to run "salt-call state.highstate" and got the error "zipimport.ZipImportError: not a Zip file"
17:24 Pixionus what module was for transfering files again?  need to transfer a file from a minion to a master or my personal computer.  Figured I'd check salt before using salt to scp
17:25 drawsmcgraw1 Pixionus: Check the 'cp' module I think
17:25 Pixionus thanks
17:25 drawsmcgraw1 yep
17:28 shaggy_surfer joined #salt
17:29 nesv joined #salt
17:29 iggy you need extra config to push from minions to master, from master to minions is easy though
17:30 Pixionus Yeah I saw that just now
17:31 Pixionus and never could get scp to work through salt since it asks for password. (I thought I found a way to do it but it never worked)
17:35 murrdoc i love the http
17:35 murrdoc move everything over it
17:35 murrdoc configs, files, everything
17:35 murrdoc lose coupling ftw
17:35 murrdoc (no one asked, i know)
17:36 iggy you know when your "git diff --stat" gets run through the pager, you need to reevaluate some stuff
17:36 Ryan_Lane anyone seen jsonnet? https://google.github.io/jsonnet/doc/
17:37 Ryan_Lane looks like it could make a really nice renderer :)
17:37 GabLeRoux_ joined #salt
17:43 vexati0n joined #salt
17:43 druonysus joined #salt
17:44 vexati0n SO... when using salt-syndic, every CLI command on the topmost master waits a ridiculous amount of time to hear from syndic masters before returning, even if the command is aimed directly at a single minion and that minion has already replied.
17:45 vexati0n is there any way to make it not do that...
17:47 jalaziz joined #salt
17:49 baweaver joined #salt
17:52 otter768 joined #salt
17:56 UtahDave vexati0n: you can specify the -t option on the command line
17:56 UtahDave how long is a ridiculous amount of time?
17:56 UtahDave Ryan_Lane: jsonnet does look pretty interesting
17:57 Cidan vexati0n generally when I have that happen, that means something is not responding/down, a salt minion, a server, etc
17:58 Cidan vexati0n try this: salt -t 3 --show-timeout "TARGETS" test.ping
17:58 UtahDave Cidan: you get that delay all the time when using the syndic, though.
17:58 Cidan vexati0n where targets are what you're trying to salt
17:58 ashb joined #salt
17:58 Cidan hm, weird, I don't
17:59 denys joined #salt
17:59 ashb Hello. If I kick off a highstate job (like this https://github.com/ministryofjustice/bootstrap-salt/blob/master/bootstrap_salt/salt_utils.py#L11-L14, but could be done on the cmd line)
17:59 GabLeRoux joined #salt
17:59 ashb Is it possible to get the progress on what the minion is doing as it goes along?
17:59 ashb Or does the minion not actually return anything to the master until the end?
18:00 ze- ashb: it does not return anything until the action is over.
18:00 ashb dang.
18:00 UtahDave ashb: you can have the minion send events across the events bus as it finishes each part of your state, if you want
18:01 meylor joined #salt
18:01 ashb Oh yes, I can probably re-enable minion events now.
18:01 ashb (I disabled them cos of the 5x slow down in 2014.1.x)
18:02 UtahDave or create your own custom events.  there's an event state.
18:02 meylor left #salt
18:03 overyander joined #salt
18:03 blacked joined #salt
18:03 thayne joined #salt
18:03 GabLeRoux joined #salt
18:05 solidsnack joined #salt
18:05 ashb ooooh really now.
18:08 Vynce joined #salt
18:08 baweaver joined #salt
18:08 XenophonF good god how many jinja syntax errors can i make in a single line?
18:09 XenophonF that's 3 so far
18:10 utahcon XenophonF: I am sure you can make more.
18:11 ajw0100 joined #salt
18:11 babilen (just write lisp)
18:11 GabLeRoux joined #salt
18:11 zwi joined #salt
18:12 blacked1 joined #salt
18:12 krelo joined #salt
18:12 Vynce so. how does one debug a misbehaving minion?
18:13 Vynce I have a minion that appears to act like i'm running test=True when in fact i am explicitly running test=False
18:13 GabLeRoux_ joined #salt
18:15 UtahDave Vynce: I usually will ssh in to the minion, stop the minion daemon and run the minion in the terminal in debug mode.        sudo salt-minion -l debug
18:15 UtahDave then watch when running the commands from the master
18:16 Vynce hm.  any guide on how to read this output?
18:17 UtahDave I usually look for stacktraces or any errors that are displayed
18:18 XenophonF a lisp SLS renderer would be pretty awesome, actually
18:18 shaggy_surfer joined #salt
18:19 XenophonF (fix_up_something ((cmd.run :name "echo hello world")))
18:19 jerematic joined #salt
18:19 XenophonF i can't remember what CL's hash syntax looks like - might have to use alists instead
18:19 iggy vexati0n: another thing to check is if you have keys for hosts that don't exist anymore (sudo salt-run manage.down removekeys=True  ... will get rid of them)
18:20 murrdoc iggy:  did they fix the two blockers for 2015.2
18:20 murrdoc or is it 2015.4
18:20 XenophonF the good news is that i've moved on from jinja errors
18:20 iggy murrdoc: not mine at least... I'm not subscribed to the other one
18:20 XenophonF now i'm making yaml errors all over the place
18:20 murrdoc you go XenophonF !
18:20 murrdoc iggy:  word
18:21 meylor joined #salt
18:22 tmclaugh[work] Is it possible to define a pillar value and reference it via jinja in another pillar file?
18:22 murrdoc no
18:22 tmclaugh[work] I’m getting rendering errors and wondering if maybe I’m missing something.
18:22 tmclaugh[work] that answers it
18:22 tmclaugh[work] thanks
18:23 iggy you have no idea if that pillar value is going to exist yet
18:24 timoguin joined #salt
18:25 meylor left #salt
18:25 tmclaugh[work] Wasn’t sure if the order of incuded pillar files might have been a way around it.
18:25 tmclaugh[work] It’s not so I’ll move on with another idea.
18:25 GabLeRoux joined #salt
18:28 Ryan_Lane UtahDave: my interest is mostly a hate of jinja :)
18:28 tmclaugh[work] wait a sec.  I already have a file where I’m doing just that.
18:28 Ryan_Lane I bet it could be extended, so that salt could inject context, custom functions, etc.
18:29 tmclaugh[work] Ryan_Lane: hey, we talked on Twitter Sunday evebing. :)
18:29 Ryan_Lane never having to pass something into the indent function is pretty compelling by itself :)
18:29 Ryan_Lane tmclaugh[work]: oh. howdy! :)
18:30 XenophonF now that i've fixed all of my jinja, yaml, and powershell syntax errors, i can use salt to deploy new forest root domain controllers!
18:30 tmclaugh[work] Still puttering along with the AWS setup here.
18:30 Ryan_Lane cool
18:30 Ryan_Lane tmclaugh[work]: if you're using splunk or grafana or pagerduty, we have some other fun stuff for you
18:31 tmclaugh[work] we have pagerduty.  That sounds interesting.
18:31 Ryan_Lane tmclaugh[work]: we're likely going to upstream something very soon
18:32 Ryan_Lane can manage users, schedules, escalation policies, services, etc
18:32 tmclaugh[work] holy crap that’s cool
18:32 Ryan_Lane we're also likely to add support to the service part of it to automatically create and link SNS topics to services
18:33 ashb What do people think of https://github.com/ministryofjustice/salt-shaker
18:34 Ryan_Lane (stuff like this is why I was saying I think what we have in salt is nicer than terraform :) )
18:34 Ryan_Lane ashb: readme doesn't really say what it does :)
18:34 ashb Huh where some bugger removed it.
18:35 Ryan_Lane oh. it's kind of like PPAs for salt?
18:35 evilrob is there a way to use salt-call to list the contents if a directory in salt://<dirname>/
18:35 ashb Ryan_Lane: sort of I guess. Recursively follow a list of deps for a salt formula and pull it down locally
18:35 ashb (we didn't like gitfs for various reasons)
18:35 Ryan_Lane oh. gotcha.
18:36 Ryan_Lane so like NPM, composer, etc.
18:36 ashb yeah
18:36 Ryan_Lane seems cool
18:36 iggy for the love of all that's holy, no
18:36 Ryan_Lane hahaha
18:36 Ryan_Lane iggy: it's like berkshelf for chef
18:37 Ryan_Lane or whatever the equivalent is in puppet
18:37 iggy the last thing we want is formulas turning into shit like that
18:37 * Ryan_Lane thinks it's a good idea
18:37 iggy so here's the problem with that stuff
18:37 * nesv grabs popcorn
18:37 Vynce <- nore with more active lurking
18:37 iggy if formulas were solving problems in a generic enough way, they should become upstream states/modules
18:38 iggy they aren't... so stop pretending like they do
18:38 ashb iggy: ides for how you have formulas reused between 8 different projects without having to worry about breaking 7 other projects?
18:38 ashb (I;d happily delete the entire code base if you can give me a better plan)
18:38 williamthekid joined #salt
18:38 ashb iggy: we're treating them as upstream. How do we pull them in sensibly (i.e. not give someone else's github repo sudo on our box)?
18:39 Ryan_Lane if formulas aren't meant to be generic, then they're just examples of what you could do, but in a way you should never do it? :D
18:39 giantlock joined #salt
18:39 iggy you don't... you fork the formula (like the docs say) and use that
18:39 iggy Ryan_Lane: basically, yeah
18:39 ashb iggy: we have 8 projects. Are we meant to have 8 forks?
18:39 iggy you kid... but for reals
18:39 Ryan_Lane you could say the same for berkshelf or puppet forge or ansible galaxy
18:39 Ryan_Lane all of them are great for building proof of concepts, though
18:40 iggy ashb: I can't begin to tell you how to run your systems (and if you are relying on someone to do that, you've got bigger problems)
18:40 tmclaugh[work] BTW, looks like it is possible to reference a pillar value in another file but they have to be seperate files included via the top.sls.
18:40 Ryan_Lane ashb: ideally you'd have a single project you maintain for your forks
18:41 iggy tmclaugh[work]: don't rely on that... the official stance is "don't"
18:41 Ryan_Lane it is a bit dangerous to trust the upstream, since there's numerous maintainers. I guess that's the same for pip/npm/etc, though
18:41 ashb We tried that (if I understand your idea right) - updating the version for one project forced it on all other projects
18:41 tmclaugh[work] iggy: i see.  Good to be aware of that.  Thanks!
18:41 iggy they just gave _me_ commit access to saltstack-formulas... that should scare anyone
18:41 Ryan_Lane hahaha
18:42 ashb I barely trust myself to have prod access :D
18:44 iggy if people want npm/*shelf/whatever for formulas, they need to call them something other than formulas
18:44 iggy there just aren't enough formulas that are that well written
18:44 theologian joined #salt
18:44 murrdoc ?
18:45 iggy i.e. I don't think there's a single formula that supports Windows minions
18:46 ashb iggy: I agree with you but suspect that ship has sailed (and possibly hurts salt's OSS uptake?)
18:46 iggy lol... Salt doesn't seem to be having problems growing user base
18:46 iggy (if anything it's growing too fast... the devs certainly seem to be having issues keeping up)
18:47 murrdoc we should make the formulas striaght python
18:48 murrdoc done and done
18:48 murrdoc _modules and _state files
18:48 murrdoc make it stupid complex python
18:48 murrdoc so no one touches it
18:48 murrdoc muhahahah
18:49 iggy but my original statement stands
18:49 iggy if there were generic enough ways to solve some of the problems that formulas solve, they could just be upstream states/modules
18:50 mohae joined #salt
18:50 ashb what's the release cycle times on that?
18:51 iggy better/worse than the release cycle times on formulas
18:51 ashb & # have to get off the train now.
18:52 iggy a lot of the formulas haven't been touched in any meaningful way in a year or more
18:53 nzero thats true of every open source system i have ever seen that supports some kind of "plugin". someone makes it, and then it just rots away
18:53 dojobonobo joined #salt
18:53 iggy yeah, the _other_ downside here is that most of the formulas are "it worked for me"
18:53 dojobonobo is salt.states.network still experimental on debian/ubuntu?
18:53 iggy dojobonobo: yes
18:55 dojobonobo should I just do file.managed on /etc/network/interfaces and service.running on networking watching the interfaces file?
18:55 drawsmcgraw1 dojobonobo: I'd recommend not using them. I tried it on my Ubuntu laptop and.... well, it wasn't helpful.
18:55 drawsmcgraw1 dojobonobo: Sounds like a good approach
18:55 iggy for now, that's probably your best bet
18:55 dojobonobo cool ty
18:55 iggy #ssceapproved
18:56 XenophonF saltstack-formulas have zero support for anything other than linux
18:56 Ryan_Lane I generally make state and execution modules
18:56 murrdoc they cant have windows support
18:56 XenophonF it's really frustrating because i want to use them on windows and freebs
18:56 XenophonF freebsd
18:56 drawsmcgraw1 iggy: You should have a business card that has that on it.
18:56 iggy XenophonF: not entirely true... some have fbsd support
18:56 murrdoc he does drawsmcgraw1
18:56 Ryan_Lane and upstream them. it's a lot easier to do logic in python than in sls
18:56 murrdoc he does
18:57 XenophonF iggy: they try to, but i've hit a bunch of bugs that i need to send pull requests upstream for
18:57 XenophonF the lack of good windows support is really harshing my mellow
18:57 drawsmcgraw1 wait how the hell did my handle get a 1 on it....?
18:57 iggy XenophonF: yeah, I mean you can't expect people that don't use fbsd to add fbsd support to formulas
18:57 XenophonF very true
18:57 iggy so if you use fbsd, we love the PRs
18:58 iggy keep them coming
18:58 drawsmcgraw1 left #salt
18:58 drawsmcgraw1 joined #salt
18:58 iggy if it's a formula that I don't have much knowledge of, I usually at least try to look it over for "grammatical" errors
18:59 iggy murrdoc: I've got it tattooed on my middle finger... tell everyone in the office "#ssceapproved" all the time
19:00 XenophonF while i'm at it i wish there was better support in salt for windows user accounts
19:01 iggy issue #?
19:01 murrdoc did aptpkg add support for 'latest;
19:02 XenophonF PR in progress - I'm trying to figure out how to describe what I want compared to how Salt already handles Unix users
19:09 nitenq joined #salt
19:09 GabLeRoux joined #salt
19:11 nitenq Hi guys, what's wrong with this file ? http://pastebin.com/xG4he5Ym I just want to download and extract this file
19:13 iggy with an http source, you need a hash
19:14 monkey66 joined #salt
19:15 nitenq iggy: well i don't know the hash of this specific file
19:16 tmclaugh[work] ahh, I think I found what I was looking for:  ‘{% from “vars.jinja" import vars with context %}’
19:16 iggy download it and hash it
19:19 Vynce drawsmcgraw: sometimes when your cnxn to IRC bounces too fast, you reconnect before the server you connect to knows you've left; you ask for "nick" and it sees a "nick" so it says no, and your client says "well, what about nick1"? and there you go.
19:19 jonlangemak joined #salt
19:20 drawsmcgraw Vynce: Ah, that's exactly what happened. My connection's been flaky lately.
19:20 drawsmcgraw Thanks!
19:21 Vynce np. (my beef with it is that my client doesn't clearly show me my nick in-my-face anywhere, so i don't notice when it happens, and i walk around with a piece of toilet paper stuck to my shoe.)
19:21 isodude_ joined #salt
19:22 iggy cough up the $5/month for a VPS and put your irc client there... problem solved
19:22 evidence so is it widely known that ext_pillar uses GitPython, despite gitfs_remotes preferring pygit2 when installed?
19:22 iggy (except when freenode has issues like it constantly does)
19:22 iggy evidence: yes
19:22 evidence the docs should say so! :)
19:22 iggy fsvo widely known
19:22 iggy evidence: we await the PR
19:23 spookah joined #salt
19:23 evidence wasted some time trying to chase a bug i thought was in pygit2
19:23 evidence is the docs repo under the same master org?
19:23 iggy they are all in the same repo (most of them are even inline in the code)
19:23 drawsmcgraw evidence: I just filed an issue about this a day or so ago
19:24 evidence link?  i'll make sure i don't dupe
19:24 drawsmcgraw https://github.com/saltstack/salt/issues/22702
19:24 evidence btw, there is a bug in the GitPython package ubuntu has (0.3.2-RC1~3)
19:24 drawsmcgraw Maybe a short-term fix is to just update the docs
19:24 evidence had to install via pip to fix
19:24 drawsmcgraw Interesting...
19:25 iggy I just pushed about 3000 lines worth of changes to our pillars/states repo... this is going to be a fun maintenance window
19:25 drawsmcgraw My systems are centos here...
19:25 iggy (on top of all the formula changes I've made too)
19:25 evidence python-git is the ubuntu package.. with all of the competing git py modules it gets gross..
19:25 drawsmcgraw I can imagine...
19:25 iggy salt-formula has support for handling that for you
19:26 iggy you're welcome
19:26 evidence it's all handled.. but the names never match upstream is all
19:26 iggy also, github recognizes sls files
19:26 iggy you're welcome for that too
19:26 timoguin joined #salt
19:26 mansquid_ joined #salt
19:27 monkey66 left #salt
19:30 scbunn joined #salt
19:31 GabLeRoux_ joined #salt
19:32 markm_ joined #salt
19:38 Eugene It does now? Last I looked it smelled them as Schee
19:38 Eugene Scheme
19:39 murrdoc yeah github fixed it
19:39 iggy I did
19:39 murrdoc ```Saltstack works too
19:40 scbunn joined #salt
19:40 Eugene Niiiice.
19:40 iggy well, me and another guy
19:41 MatthewsFace joined #salt
19:43 stoogenmeyer_ joined #salt
19:46 lifeling joined #salt
19:49 [M7] joined #salt
19:51 scbunn joined #salt
19:52 pass_by_value joined #salt
19:52 drawsmcgraw evidence: Nice job :) I should've been a little more proactive about it.
19:53 Edgan Can someone fix the latest release to 2014.7.5 on github? 2014.7.5 is hidden by ...
19:53 otter768 joined #salt
19:57 iggy Edgan: it's not actually released yet (just tagged)
19:57 iggy and yes, we all hate how that works just as much you do
19:58 Edgan iggy: Where is 2015.2.0/2015.4.0? :)
19:58 druonysuse joined #salt
19:58 druonysuse joined #salt
19:58 * iggy shakes fist
19:58 Vynce OK, i don't have a repro-able case, but FYI — i have test: True by default on some of my minions, but one of them refused to actually highstate with that in place, even when explicitly passing test=False.  removing tha tline form his minion.d and he works. weird, and i'd lov eto track it down, but my other fish are starting to burn.
19:59 Edgan iggy: Thanks for your part of salt. :)
19:59 redzaku joined #salt
20:00 MatthewsFace joined #salt
20:00 iggy I'm not a dev
20:00 Edgan iggy: You seem to be part of the community
20:00 iggy just bored at work
20:00 MatthewsFace joined #salt
20:01 Guest70 joined #salt
20:01 Edgan iggy: :)
20:04 GabLeRoux joined #salt
20:06 CaptainMagnus joined #salt
20:06 baweaver joined #salt
20:06 blacked joined #salt
20:12 GabLeRoux joined #salt
20:14 zwi joined #salt
20:14 overyander I have 3 file roots defined that exist in /srv/salt/. I manually use git to sync the roots from my git server.
20:15 overyander I have 3 file roots defined that exist in /srv/salt/. I manually use git to sync the roots from my git server. Do I need to enable any of the gitfs options to make this work properly? I constantly and at random get errors on minions when trying to pull a file.
20:16 rap424 joined #salt
20:17 iggy overyander: if you are doing a normal checkout to a dir and using that as your file_roots, you don't need any gitfs stuff
20:17 iggy (we've actually been considering doing that since that stupid fileserver.update schedule is tied to the scheduler timer and not it's own timer)
20:17 overyander iggy, i'm using pull instead of co. does that matter?
20:18 iggy pull = fetch+checkout... so no
20:18 overyander joined #salt
20:18 shaggy_surfer joined #salt
20:19 overyander i'm back, got dc'd.
20:19 overyander iggy, i'm using pull instead of co. does that matter?
20:19 GabLeRoux joined #salt
20:19 nzero [13:18:14] overyander (~Jeff@12.49.160.131) left IRC (Read error: Connection reset by peer)
20:19 nzero [13:18:16]  <iggy>pull = fetch+checkout... so no
20:19 overyander ty
20:20 aquassaut joined #salt
20:21 reabeginner joined #salt
20:22 reabeginner hi, I'm about to lose my mind :-) I put up a simple setup with one master and one minion. The master can ping the minion. I have a super simple state file yet when I try to send a highstate command, the minion return me a "'states' __virtual__ returned False "
20:23 iggy yeah, I have quite/parts/joins/etc /ignored, so...
20:23 iggy reabeginner: paste code
20:23 dave_den joined #salt
20:24 reabeginner I also had some serious problems installing the master and minion, I had to turn on debugging to see that I need to manually install pyopenssl, which in turn didn't figure out it had to install 3 additional packages :-(
20:24 reabeginner pasting the top.sls
20:24 reabeginner thanks!
20:25 reabeginner base:   'test':     - nginx
20:25 murrdoc joined #salt
20:25 druonysuse joined #salt
20:25 druonysuse joined #salt
20:25 iggy paste to gist.github.com (and paste as much as possible, including commands run, error output, etc.
20:26 lifeling left #salt
20:27 rogst joined #salt
20:27 iggy reabeginner: how did you install salt?
20:28 reabeginner https://gist.github.com/anonymous/6c411f19ab859934c51d#file-gistfile1-txt
20:29 reabeginner I installed with the bootstrap
20:29 reabeginner and then configured as shown here: http://bencane.com/2013/09/03/getting-started-with-saltstack-by-example-automatically-installing-nginx/
20:29 iggy you don't have a file.* state for /etc/nginx/nginx.conf
20:30 iggy salt-bootstrap should have installed all the required dependencies
20:31 iggy fwiw, I don't have pyopenssl installed on my master or minions
20:31 iggy you should also probably try to find a HOWTO that isn't nearly 2 years old
20:33 reabeginner it didn't... running debian jessie in AWS and debian 7 on digitalocean
20:33 blacked joined #salt
20:34 reabeginner I saw the problem regarding the pyopenssl in the log file of the minion, not the tutorial. It wasn't installed...
20:34 iggy it's not necessary
20:34 overyander i'm at a loss on why i keep getting this error "type error encountered string indices must be integers not str"
20:34 iggy you probably saw a warning about it
20:34 reabeginner great, I removed the reference to the file and worked! would have been easier with a more detailed error and not just a general fail...
20:35 iggy overyander: _tons_ of issues on github about that
20:35 overyander i'm not using gitfs which is what all of the bug reports concerning this show and i've tried updating to the latest 2014.7.5
20:35 iggy reabeginner: without seeing the actual error message, I can't say I agree
20:35 iggy at least with newer versions of salt, it's pretty good about giving good errors
20:37 reabeginner iggy: I'm afraid it I only logged to the console, and now that finally the highstate executed, the errors were lost... next time I create an instance I'll try to reproduce
20:38 reabeginner and I didn't get any error message, not in the minion or the master, indicating that a missing file specified in the sls was the root of the problem
20:39 Tecnico1931 joined #salt
20:39 iggy reabeginner: yeah, next time you run into problems, make sure you get paste the command you are running and the error you get from that (you really shouldn't need to dive into the minion log file very often)
20:39 subsignal joined #salt
20:41 ndrei joined #salt
20:42 reabeginner another quick question, did anyone create a sls file editor? I think it could be nice to have such a tool with error checking, autocomplete etc.
20:43 ndrei joined #salt
20:45 iggy there are "grammar" bundles for a lot of existing editors
20:45 jcsp joined #salt
20:45 iggy it's mostly just yaml+jinja (which most editors already understand)
20:47 tomh- joined #salt
20:50 reabeginner ok. thanks again for your help iggy. Good night...
20:55 j-saturne joined #salt
20:56 overyander iggy, here's the file it errors on "Fetching file from saltenv 'base', ** attempting ** 'salt://global.sls'"  global.sls is located in /srv/salt/base/  the error shows it as trying to get salt://global.sls should that be salt://base/global.sls  ?
20:56 yorjo joined #salt
20:56 iggy not unless your file_roots are /srv/salt
20:57 overyander file_roots:
20:57 overyander base:
20:57 overyander - /srv/salt/base
20:57 kunersdorf anyone have success with https_user/https_pass with git.latest?
20:57 overyander sorry for the multiline, thought it would be single line
20:58 iggy looks right
20:58 iggy I mean... 3 lines from your config file isn't a lot to gone on...
20:59 yorjo Documentation says one way to add custom grain is to add them in /etc/salt/grains. Is it suppose to be the same on Windows minion? I am using the "salty vagrant grains" plugins which is successfully writing the file but neither salt-call.exe -g or calling grains.items from the master reflect the fact that I just added the custom grains
20:59 yorjo Should I write custom grains in another file on Windows^
21:00 iggy yorjo: where is the minion config file on Windows? the grains file should be next to it
21:00 iggy that's about all I can say on the topic
21:01 yomilk joined #salt
21:01 yorjo minion config file is in C:\salt\conf
21:01 yorjo I will try that. Thanks
21:03 und1sk0 joined #salt
21:05 overyander iggy, here's the master conf file http://pastebin.com/FkWRcCRF and here's the error http://pastebin.com/NfPieLhx
21:05 iggy protip: gist supports multiple files per paste
21:06 overyander ah, cool. ty
21:08 nate_c joined #salt
21:10 giantlock joined #salt
21:11 blacked joined #salt
21:13 und1sk0 I'm using 2014.1.19 and i'm getting really inonsistent results with salt, specifficly with "manage.status"
21:14 iggy I don't see anything out of the ordinary there... start hacking and slashing until it starts working
21:14 und1sk0 specificly
21:14 und1sk0 iggy: heh
21:14 iggy that was for overyander
21:14 iggy und1sk0: there is no such thing as 2014.1.19
21:15 und1sk0 [root@bastion1 us-east-1 ~]# salt --version
21:15 und1sk0 salt 2014.1.10 (Hydrogen)
21:15 und1sk0 sorry, .10
21:16 iggy without more info about what is/isn't working, I'd say you're hitting the "minions go missing" issue
21:16 iggy improved slightly in 2014.7, basically completely gone in 2015.2
21:16 und1sk0 ah, so, "try upgrading" - I guess that's not unexpected advice
21:17 und1sk0 (which I will give a shot)
21:17 und1sk0 thanks
21:18 iggy as I said... without knowing more
21:19 und1sk0 i will try upgrading and see if morale improves and if not, kep beating it ;)
21:19 iggy 2015.2 not being out yet, it's hard to suggest that to people
21:20 iggy but if you have the ability to test it, it would certainly be nice to know if it sorts out the issue
21:23 und1sk0 probably stuck with 2014.7.x
21:23 baweaver joined #salt
21:27 amcorreia joined #salt
21:33 ajw0100 joined #salt
21:35 itru joined #salt
21:35 blacked joined #salt
21:41 penguinpowernz joined #salt
21:43 nitenq joined #salt
21:44 nitenq Hey, How can I resolve this error : http://pastebin.com/U5Dkrt20 ?
21:45 vandemar nitenq: what OS is that running on?  Is the local libc perhaps only compiling iso8859-1 and not utf-8 locales?
21:46 nitenq vandemar:Ubuntu 14.04.2 LTS
21:47 nitenq I don't know at all but this is not the first time that I run salt commands
21:47 nitenq (on this machine I mean)
21:47 vandemar look in /etc/locale.gen
21:48 vandemar or is that message from the minion?  check it on the minion too.  same OS?
21:48 iggy someone else had a similar issue the other day... might check the irc logs
21:50 nitenq Ubuntu 14.04.1 LTS on the minion
21:50 nitenq where can I find those logs iggy ?
21:51 vandemar nitenq: topic, also seems like it might be a python and/or LANG or LC_ALL env variable problem, based on googling for  python "unknown locale: UTF-8"
21:52 tmclaugh[work] Is there a way to have salt intentionally throw an error?  I want salt to fail if I provide a wrong value from the command line.
21:52 tmclaugh[work] we’re using salt in local mode so I don’t want people getting the ID of an EC2 object incorrect and inadvertantly creating a new resource
21:54 UtahDave tmclaugh[work]: well, there's probably a couple ways to do that.  1. use the failhard: True option in your states.
21:54 nitenq vandemar: thanks export LC_ALL=en_US.UTF-8 export LANG=en_US.UTF-8
21:54 UtahDave 2. create a custom module that will sanity check your inputs and raise an exception if they're bad.
21:54 nitenq vandemar: this worked
21:54 otter768 joined #salt
21:54 jmdcalks joined #salt
21:55 vandemar nitenq: did you change something about the root shell on those machines, or did they default to something other than utf-8?  Seems odd you ran into that on ubuntu when most of the google hits were talking about osx (which is always quirky)
21:55 tmclaugh[work] UtahDave: Thanks.  looks like custom module.
21:56 UtahDave tmclaugh[work]: Yeah, it's quite easy
21:56 nitenq vandemar: Actually this is not my machine but I don't think that something special have been done on it
21:56 tmclaugh[work] was looking to do something like: {% if environment not in [‘prod’, ‘staging’, ‘testing’] %} <failure> {% endif %}
22:02 iggy tmclaugh[work]: there are test states to do stuff like that
22:02 jmdcalks howdy, what is the command to send a minion to execute a sls file/go into a state when it is not registed in the top file and has no env?
22:03 iggy jmdcalks: salt 'minion' state.sls foo.bar
22:03 jmdcalks thanks
22:03 tmclaugh[work] iggy: that is pretty sweet.  Looking at salt.states.test right now
22:03 iggy or in newer versions state.apply
22:05 shaggy_surfer joined #salt
22:06 FeatherKing joined #salt
22:08 cberndt joined #salt
22:12 echo joined #salt
22:14 markm__ joined #salt
22:15 randomuser joined #salt
22:25 iggy so... slightly OT, but not really... Anybody know of any good "BI/statistics/etc frameworks"? (i.e. visualizing app metrics, how long someone stays on a page, how long they are logged into our apps, etc)
22:26 iggy there's time-series stuff for days, but the kind of stuff I want to visualize doesn't really fit well into that kind of graph
22:26 iggy (and realistically, I'm the absolute wrong person to be doing this, but hey... boss says jump...)
22:28 murrdoc iggy:  new relic
22:28 iggy they have an aversion for outside services
22:30 iggy but I will make sure I mention it in the conversation
22:32 KyleG +1 for newrelic
22:33 kuromagi joined #salt
22:37 jespada joined #salt
22:39 Ryan_Lane tmclaugh[work]: we do that
22:40 Ryan_Lane we're using environment vars for stuff
22:40 Ryan_Lane but we're weird and we're setting grains based on them
22:41 iggy weirdo
22:41 Ryan_Lane tmclaugh[work]: https://gist.github.com/ryan-lane/dc27a3e0b9deff175492
22:42 tmclaugh[work] i like that
22:42 tmclaugh[work] I will need to add something similar.
22:42 Ryan_Lane that's for our orchestration
22:42 solidsnack joined #salt
22:42 subsignal joined #salt
22:42 Ryan_Lane we set grains through the environment so that our orchestration for a service looks just like the config management for it
22:42 tmclaugh[work] Just finished and filed my PR here.  We should be able to deploy most any service here once I do the remaining pillars by doing:
22:43 Ryan_Lane {{ grains.service_name }} in orchestration is the same in config management. it's nice
22:43 tmclaugh[work] $ salt-call -l debug --config-dir=. --id ‘*.<svc_name>.<environment>’ state.highstate
22:43 Ryan_Lane cool
22:43 tmclaugh[work] That will go ahead and create the services SG, LC, ASG, and ELB
22:44 tmclaugh[work] We also have only a single state for all that.  The pillars handle everything.
22:44 Ryan_Lane you should write some blog posts about how you've done things :)
22:44 Ryan_Lane always nice to see how other people are handling stuff
22:44 tmclaugh[work] I’m trying to keep devs contained to pillars/
22:44 Ryan_Lane ah. heh
22:44 tmclaugh[work] I’m thinking about it. :)
22:44 Ryan_Lane we went the opposite way. we get our devs to maintain all of their own salt code
22:44 Ryan_Lane it lives in their repos
22:44 tmclaugh[work] I’m scared what I did was so freekish though.
22:45 tmclaugh[work] *freakish
22:45 Ryan_Lane if it works for you, it's not freakish
22:45 Ryan_Lane the most important thing is to be happy with your workflow :)
22:46 iggy the beauty (and hardest thing to figure out when you're starting out) is that salt can do a million things in a hundred different ways... it's always nice to see what kind of bastardized setup someone has come up with
22:46 Ryan_Lane indeed. for instance most people would consider mine bastardized :)
22:46 iggy (for proof look at how many presentations Ryan_Lane did at saltconf and how many people attended them)
22:46 Ryan_Lane hahaha
22:46 tmclaugh[work] I’ve gotten it so a dev can do 6 lines of yaml and deploy their service.
22:46 Ryan_Lane tmclaugh[work]: that's awesome
22:47 Ryan_Lane tmclaugh[work]: when we upstream our pagerduty support you can add in their services, rotations, etc too :)
22:47 Ryan_Lane assuming your devs are on call
22:47 tmclaugh[work] service/init.sls sets reasonable defaults and then does an include on the service specific pillar (service/myservice/init.sls)
22:48 ajw0100 joined #salt
22:48 iggy I almost miss pager duty
22:48 iggy now it's just me
22:48 tmclaugh[work] learning I could do “{% from 'vars.jinja' import vars with context %}” was also helpful for me to create some concept of global vars
22:49 solidsnack Can the Reactor system run tasks on the master? Is that what something like `runner.state` does?
22:49 iggy solidsnack: true
22:49 solidsnack So to create Route53 entries do I: use `salt-run` and boto_aws or `salt-cloud` and something else?
22:50 iggy i.e. you can run the orchestrate runner on an event
22:50 iggy boto_aws is a module
22:50 solidsnack iggy: What does it mean that it is a module?
22:50 iggy so not something you'd run in the master context
22:50 solidsnack Ah.
22:51 solidsnack tmclaugh[work]: We'd really like to know more about how you're using Salt to setup AWS resources in a configurable/flexible way.
22:51 iggy if you are trying to limit the spread of your aws info, you can run a minion on the master and have it do all that kind of stuff
22:51 solidsnack Ah.
22:51 iggy (that's what we do with our backups)
22:51 solidsnack It's about IAM.
22:51 baweaver joined #salt
22:52 solidsnack Can the Salt master just run straight up shell commands?
22:52 iggy the master process... no
22:52 iggy but as I said, you can run a minion on the master
22:52 tmclaugh[work] I can scrub this tree and create a repo for a public version.  I can look into that tomorrow.
22:52 Ryan_Lane solidsnack: you want the boto_route53 module
22:52 solidsnack Right.
22:53 tmclaugh[work] Just really started diving into Salt last week.
22:53 solidsnack tmclaugh[work]: Cool.
22:53 solidsnack You would have users right away...
22:54 solidsnack What kinds of things can runner.* do?
22:55 iggy salt.runner.*
22:57 iggy http://docs.saltstack.com/en/latest/salt-modindex.html#toggle-14
22:57 solidsnack iggy: So just so I understand better, say the Reactor calls `salt.runners.cloud.create` -- does that run on a Minion?
22:58 iggy all runners run in the master context
22:58 jespada joined #salt
23:02 yomilk joined #salt
23:08 solidsnack joined #salt
23:11 tmclaugh[work] Ryan_Lane, solidsnack: https://github.com/tmclaugh/straycat-salt
23:12 tmclaugh[work] scrubbed it of work stuff.
23:12 Ryan_Lane tmclaugh[work]: you're moving away from salt to puppet for config management?
23:13 tmclaugh[work] we can do a ton more and at a quicker pace with Puppet simply because of skill sets.
23:13 Ryan_Lane gotcha
23:13 Cidan joined #salt
23:13 tmclaugh[work] Most of what we need here CM-wise I’ve already written in Puppet.
23:13 Cidan joined #salt
23:14 murrdoc salt gives you a lil more than a cm
23:14 murrdoc like it replaces mco
23:14 tmclaugh[work] Have to automate Kafka and Bookkeeper.  I already have Puppet work for that at home so it’s a lot of copy and paste
23:14 murrdoc replaces the need for rvm
23:14 rhodgin joined #salt
23:14 tmclaugh[work] murrdoc: we’re using both.
23:14 murrdoc :) thats nice
23:15 tmclaugh[work] There’s also nice things I did with Puppet testing that I don’t know/want to spend the time reimplamenting.
23:15 rhodgin joined #salt
23:15 Ryan_Lane no need to replace puppet. you may eventually want to because you're familiar enough with salt, but whatever works for you is best
23:15 tmclaugh[work] With my old Puppet environment when I made a change I had scripts that sampled the environment via PuppetDB and showed me what should change.
23:15 murrdoc yeah that and the lack of an equivalent to puppetdb is the two places salt's CM implementation is behind
23:16 Ryan_Lane there's sdb in 2015.2
23:16 murrdoc sdb is shipping? with the community version
23:16 murrdoc ?
23:17 iggy si
23:17 iggy sekrit
23:17 Ryan_Lane core is core. it's only community
23:17 murrdoc gtfo
23:17 murrdoc how does the code interact with it
23:18 iggy it's not well documented just yet
23:18 iggy (and they didn't beat it to death at saltconf like beacons ;)
23:18 tmclaugh[work] Down the road we may reevaluate Salt for CM.  Until then we have one Puppet expert and 3 people just beginning Salt.  We also have the advantage of a pretty robust functional tree to pull from when we need new services here.
23:18 murrdoc Ryan do u know if it interacts with the code, is it REST only ?
23:18 tmclaugh[work] We’re a growing startup and the previous OPs guy was very manual orriented.  So it’s GSD and move along.
23:19 Ryan_Lane I actually have no clue :)
23:19 murrdoc damnit
23:19 Ryan_Lane I've just been seeing code added for it
23:19 murrdoc now i must know
23:21 Vynce OK, seriously? http://salt.readthedocs.org/en/latest/ref/modules/all/salt.modules.service.html#salt.modules.service.status … all i get back is "True" not the status the service reports for itself?  lame.
23:21 zircote joined #salt
23:22 hal58th up or down. all that really should matter
23:22 hal58th All it really wants is an exit code
23:24 yomilk joined #salt
23:30 Vynce but that's nto what "status" does on most services, so it's annoying to use that word to mean a different thing.
23:31 blacked joined #salt
23:31 yomilk joined #salt
23:35 jerematic joined #salt
23:35 bfoxwell joined #salt
23:37 zircote joined #salt
23:37 evidence hm.. it appears something broke with 2015.2.0 in regards to conf loading, it doesn't autoload/dict merge other stuff in master.d properly
23:38 evidence non salt related configs no longer show up in config
23:40 blacked joined #salt
23:41 cmcmacken joined #salt
23:41 iggy evidence: you mean you were relying on having random config data in master.d?
23:42 evidence it was config data to access an api, a dict containing hostname/username/pass etc.  2014.7.X loaded it and populated config('servicename')
23:42 evidence it seems that's now unsupported though
23:42 iggy I can't think of anything that would have changed that
23:43 evidence going to a config.py diff.. my untrained eye might catch something
23:49 iggy looks like you are correct
23:49 Singularo joined #salt
23:49 evidence might be the changes to use _expand_glob_path ?
23:49 evidence correct on which assumption? :)
23:49 iggy which is annoying if you are trying to write custom _modules/_grains/_states that pull config via config.get
23:49 evidence it was actually a _runner, but yea
23:50 evidence going to test a change or two i think might have caused it.  if so, was it likely a oversight, or was support for that method dropped purposely?
23:52 desposo joined #salt
23:52 iggy if it was dropped purposely, it should be documented
23:53 bluenemo_ joined #salt
23:53 iggy seems like a very valid use case to have it available for custom modules/etc
23:54 evidence thx for the verification, i'll open an issue once i have more info on the changes that affected it
23:55 otter768 joined #salt
23:59 GabLeRoux joined #salt

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