Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-03-06

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

All times shown according to UTC.

Time Nick Message
00:00 nich0s What IRC client do you use on your phone?
00:00 dimeshake mangas: can you restate the question?
00:01 hal58th joined #salt
00:01 nich0s hal58th: o/
00:01 hal58th Hello nich0s
00:01 MTecknology pretty girl's flight is in 40min... i should talk to her more
00:01 mangas dimeshake, http://docs.saltstack.com/en/latest/topics/ssh/#define-cli-options-with-saltfile
00:02 mangas according to that document I should be able to create a file named Saltfile with some configurations inside
00:02 hal58th Of course you should talk to her more Mtecknology.
00:02 dimeshake yes - and as long as it's in the current directory, it should work
00:02 ccarney_ROCC left #salt
00:02 dimeshake according to the docs..
00:03 mangas dimeshake, it's the current dir
00:03 mangas but it's not working =/
00:03 Alan_S joined #salt
00:04 thayne joined #salt
00:04 dimeshake i've not really used salt-ssh much or ever Saltfile - not sure i can help more than that...
00:05 hal58th I've even heard of Saltfile. Literally a file called "Saltfile"? That seems odd
00:05 mangas dimeshake, do you know how I can override the cache dir ?
00:06 mangas hal58th, that is what the docs say
00:10 hal58th mangas did you try a lower case 'S'
00:10 baweaver joined #salt
00:13 hal58th There is a typo in there too. Let me know if you get it working. mangas. It should let you know if it loads that file if you use debug
00:18 ocdmw joined #salt
00:19 rgarcia_ joined #salt
00:23 XenophonF joined #salt
00:23 hal58th joined #salt
00:24 XenophonF howdy yall - does s3fs have any dependencies outside of what's already required by the salt-master/salt-minion packages?
00:24 XenophonF I can't seem to get s3fs working on my master, which is why I ask.
00:24 hal58th hmmmm, lets find out!
00:25 MTecknology Apparently I'm down to #77 in commits...
00:25 MTecknology I want to get back up to top 20 by the end of the year
00:26 XenophonF i went looking through salt/utils/s3.py and salt/utils/iam.py, but those look self-contained except for maybe a dependency on python-requests
00:26 zwi joined #salt
00:27 hal58th XenophonF take a look at this. https://github.com/saltstack/salt/blob/develop/salt/fileserver/s3fs.py
00:27 nich0s joined #salt
00:32 MTecknology There's not a lot of low hanging fruit in the issue queue, huh?
00:33 hal58th XenophonF you might need the pickle package "python-jsonpickle".
00:34 hal58th Haven't ever looked. I want to start fixing low hanging documentation problems though
00:35 baweaver joined #salt
00:43 MTecknology she's prettier up close. I'd move here for her
00:43 MatthewsFace joined #salt
00:45 hal58th I hope you are actually talking to her
00:45 jaloren joined #salt
00:45 jaloren hi all i have a question about dynamically passing pillar data to the salt-run orchestrate
00:45 jaloren like so
00:46 jaloren salt-run state.orch system.cluster  pillar="{'ips': [10.0.40.94,10.0.40.105,10.0.40.116,10.0.40.115]}"
00:46 iggy hal58th: did you leave alrwady?
00:46 MTecknology hal58th: I was, but she's now seeing if she can get onto a flight
00:46 jaloren so that works well enough in the actual SLS file that i am calling in the runner
00:46 jaloren BUT that pillar data is not available to any SLS state file called from system.cluster
00:46 jaloren is this by design?
00:47 hal58th iggy, nope chilling in my hotel room with a beer.
00:47 MTecknology hal58th: She used to live in San Fransisco. She's flying back to see some friends and right now seeing if she can get on an earlier flight.
00:47 hal58th Ah, good luck to you Mtecknology! Get dem digits
00:47 MTecknology Sounds like a couple other standby passengers are ahead of her (if they show up)
00:48 MTecknology hal58th: What would digits do? I'll never see her again.
00:48 iggy hal58th: come hang out with us
00:48 hal58th You'd be surprised Mtecknology.
00:48 hal58th Where you at iggy?
00:48 nich0s SaltConf16.
00:50 MTecknology looks like she made it on
00:51 iggy we are at blue iguana right now
00:51 MTecknology never will I see her again
00:51 MTecknology I lied.. funny
00:52 MTecknology she's just sitting at another gate now waiting standby on another flight. Which, to me, says she's not interested it talking further.
00:52 hal58th all right iggy, im on my way on foot. I'll send you my phone number over private chat in case you move.
00:53 MTecknology inside the airport or you guys out and about like I should be?
00:53 hal58th Out and about. We don't leave until tomorrow
00:53 bfoxwell joined #salt
00:53 MTecknology ah.. I knew that, huh?
00:54 bluenemo_ joined #salt
00:54 MTecknology I'm done drinking for the next 5mo, else I'd go pay the crazy premium on airport booze
00:56 hal58th well you tried Mteknology. That's better than nothing
00:59 MTecknology yup
00:59 MTecknology I really seem to have a thing for super skinny girls
01:00 Corey MTecknology: You're at the airport?
01:00 MTecknology yup
01:01 Corey I'm at the High West Saloon.
01:01 Corey (Airside)
01:01 MTecknology where is that?
01:02 otter768 joined #salt
01:02 MTecknology I'll see if I can find it.
01:02 nich0s The occasional roll of Utah skiing footage is getting a little old.
01:04 shaggy_surfer joined #salt
01:04 desposo joined #salt
01:09 MTecknology Corey: almost there
01:11 Corey Huzzah. I'm just holding down the end of the bar here.
01:16 Heartsbane Corey: in park city
01:16 Heartsbane ??
01:18 * Heartsbane thinks that reminds him that he needs some rye.
01:19 ocdmw joined #salt
01:21 nikogonzo woo salt conf! it was most excellent; good job -> all salt staff in this channel
01:26 subsignal joined #salt
01:27 dave_den joined #salt
01:29 Corey Salt Party at the High West Saloon in the airport.
01:29 nich0s Lol. How many folks have you accumulated?
01:30 Corey MTecknology and me so far.
01:30 Corey nich0s: You here?
01:30 nich0s I am.
01:30 Corey nich0s: When do you board?
01:30 nich0s 1935
01:30 Corey We're near the Delta gates. Get o'er here!
01:31 MTecknology nich0s: right after you get down the escalators heading toward D gates
01:31 nich0s Cool. I'll swing by for a bit.
01:32 otter768 joined #salt
01:34 ALLmightySPIFF joined #salt
01:37 bhosmer_ joined #salt
01:38 sfxandy joined #salt
01:39 bhosmer__ joined #salt
01:43 evilrob joined #salt
01:52 TyrfingMjolnir joined #salt
01:55 Heartsbane Corey: good seeing you... GL with your flight
01:55 viq joined #salt
02:00 Hex_Offender joined #salt
02:01 aquinas joined #salt
02:07 Corey Heartsbane: Thanks! Good seeing you, too.
02:11 mafro joined #salt
02:11 Gareth Corey: Looks like I just missed you guys at the High West, wandered off from the E gates but didn't see anyone there.
02:12 nich0s joined #salt
02:15 Corey Gareth: That's a shame. Boarding at B gates. Next time, Gadget!
02:15 Gareth Corey: Have a good flight :)
02:16 michelangelo joined #salt
02:17 dyasny joined #salt
02:19 iggy if anybody else leaves tomorrow, there's afew of us at green pig pub
02:20 Gareth i, hopefully, will be boarding in about 10-15 minutes :)
02:22 thayne joined #salt
02:23 ponpanderer joined #salt
02:24 nich0s Gareth: Same here. Fly safe!
02:24 ponpanderer anyone know when 2015.2 will shed it's RC status and replace 2014.7.x?
02:24 Gareth nich0s: you as well :)
02:24 ponpanderer or a better question....how semi-prod ready is 2015.2? :)
02:26 evle joined #salt
02:26 ocdmw joined #salt
02:27 malinoff joined #salt
02:30 cheus joined #salt
02:30 __number5__ ponpanderer: lesson learned last year, don't go straight to latest release...
02:31 Terminus- joined #salt
02:35 GrueMaster joined #salt
02:36 rgarcia_ joined #salt
02:36 helgim_ joined #salt
02:37 Sypher joined #salt
02:37 aparsons joined #salt
02:46 mosen joined #salt
03:11 ponpanderer __number5__: good point. i even noticed some surprises going from 2014.1 to 2014.7
03:16 schlueter joined #salt
03:17 StDiluted joined #salt
03:22 ALLmightySPIFF joined #salt
03:31 nich0s joined #salt
03:32 ocdmw joined #salt
03:34 pdayton joined #salt
03:36 CeBe1 joined #salt
03:41 TyrfingMjolnir joined #salt
03:42 smcquay_ joined #salt
03:45 intellix joined #salt
03:48 schlueter joined #salt
03:53 davet joined #salt
03:54 favadi joined #salt
03:57 monkey66 joined #salt
03:57 garthk Not sure from the docs whether onchanges points in the direction I think it does.
04:00 rgarcia_ joined #salt
04:01 malinoff We still use v0.16 :)
04:01 ALLmightySPIFF joined #salt
04:01 cheus joined #salt
04:06 jerematic joined #salt
04:11 Gareth joined #salt
04:11 favadi joined #salt
04:11 Brew joined #salt
04:14 huleboer joined #salt
04:14 garthk was not run because onchanges req did not change
04:14 garthk not sure I can get apt-get to run if either sources.list is updated or one of the keys in the list changes :/
04:16 garthk looks like multiple onchanges needs ALL the targets to change
04:16 hal58th yeah i think you are right
04:16 hal58th are you managing the sources.list?
04:17 garthk hal58th: yeah, and I also had some key cmd.runs that were trying to onchange_in to the same cmd.run for apt-get update
04:17 garthk the file.managed tripped, but because the cmd.runs didn’t the final cmd.run didn’t :)
04:18 hal58th you can just do a "watch_in" with your file.manged state to your cmd.wait to run apt-get update.
04:20 intellix joined #salt
04:20 rgarcia_ joined #salt
04:24 garthk cmd.wait, eh? cheers
04:24 hal58th cmd.run always runs. cmd.wait will actually act on requisites.
04:25 garthk … and not run unless they do something?
04:27 hal58th cmd.wait won't ever run unless called on by a requisite. most people use watch but some other requisites work as well
04:28 garthk thanks
04:28 hal58th woops, it's only watch. just read the document. http://docs.saltstack.com/en/latest/ref/states/all/salt.states.cmd.html#salt.states.cmd.wait
04:30 hal58th garthk, watch and the inverse watch_in. to be specific
04:30 StDiluted joined #salt
04:32 Furao joined #salt
04:32 favadi1 joined #salt
04:32 evle joined #salt
04:43 MTecknology i2
04:44 MTecknology I would have preferred a scrraming baby on that last leg.
04:47 MTecknology screaming*
04:47 nich0s Everyone on this flight is incredibly talkative. Loudly talkative.
04:49 favadi joined #salt
04:49 logix812 joined #salt
04:50 logix812 is there a way to make a state NOT run if another state HAS run?
04:50 logix812 have a a start and a restart state... if start ran, I don't need restart to run
04:51 logix812 in my case restart doesn't start the first time
04:51 logix812 so I can't just use restart for everytthing
04:54 MTecknology logix812: -unless
04:54 logix812 MTecknology: ya I got that going.. it's for circus... restart doesn't run "unless" circusd is in the process list
04:54 logix812 but the 1st time through
04:55 logix812 circusd isn't in the process list for circus starts.. after it starts it's in the process list, so restart will run
04:55 MTecknology you lost me, but the plane is leaving now
04:56 logix812 sorry.. 1st time though my circus.start will run "unless" circusd is in the process list... so it runs the first time.. the my circus.restart wants to run onlyif circusd is in the process list
04:57 logix812 on the 1st pass both end up running
04:57 logix812 second highstate only the restart runs as expected
04:58 logix812 does "unless" work with states? unless: cmd: <whatever> ?
04:58 logix812 cause that would do the trick
05:01 Corey MTecknology: I almost had the screaming baby next to me in Row 1.
05:01 Corey Gareth: Back in SFO. WHee.
05:04 ALLmightySPIFF joined #salt
05:07 jerematic joined #salt
05:10 johtso joined #salt
05:15 mafro joined #salt
05:15 bones050 joined #salt
05:27 Terminus- joined #salt
05:31 Terminus- hello. question, how do i check if the value of a key is a list or not? i have pillars that contain "class: foo" but i also want to allow for "class: [ foo, bar, baz ]" and render it in a managed file.
05:31 cheus Terminus, use 'in'
05:31 cheus wait, sorry
05:32 Terminus- cheus: but wouldn't that mean you already know at least one element in the list?
05:32 ALLmightySPIFF joined #salt
05:32 cheus Terminus-, Sorry late night at airport -- fuzzy headed... rereading
05:32 ocdmw joined #salt
05:32 cheus Terminus-, This is ugly... but there are ways... gotta go find a github file
05:33 cheus Terminus-, Seriously ugly -- https://github.com/saltstack-formulas/php-formula/blob/master/php/ng/files/php.ini#L12
05:34 cheus Note that dicts and strings are iterable as well so you have to self-select past them or
05:35 cheus do some other kind of magic
05:35 Terminus- cheus: thanks. i'm working on understanding it now.
05:36 Terminus- cheus: i'm thinking i can just short circuit it by checking if it's a string first.
05:36 cheus Terminus-, Basically it's easier to test for string than anything else so test for string or number then assume something else
05:36 cheus You said it just as I wrote it
05:37 Terminus- cheus: yup. just didn't know the foo is string syntax.
05:38 Terminus- at this point my conf files look more like a ton of jinja than actual config files. XD
05:42 __ale__ joined #salt
05:45 cheus Terminus-, welcome to salt
05:45 cheus ;-)
05:45 Terminus- =D
05:46 cheus The DRYer we make 'em the more impenetrably jinja-ish they get
05:47 cheus That example above? That's an entire php.ini in the form of a jinja template -- all declarations
05:47 Terminus- cheus: my other worry would be i'm really the only one here who writes code. too many sysadmins in general i think don't write code.
05:47 cheus Good keynote today about that at SaltConf
05:47 * Terminus- looks for vod
05:48 cheus The argument was that 'sysadmins are coders these days '
05:48 ajw0100 joined #salt
05:48 mosen cheus: I dont like the pattern of re-exposing an entire config grammar into another language
05:48 nich0s joined #salt
05:50 cheus mosen, Sure -- but the alternative is to copy every single possible option thought of and unthought-of into the formula which then sets up version dependencies (eg, a new config key would force a change in the formula and visa versa the other way around). It's unmaintanable and makes the states impossible to be reused
05:51 cheus It's also just really annoying and redundant
05:51 Terminus- cheus: btw, is it possible to get jinja in salt to not generate whitespace by default for control flow statements?
05:51 mosen cheus: ah I'm talking more along those lines
05:52 mosen cheus: not saltstack but, in puppetlabs-apache they re-exposed every vhost option as a parameter.. except for a few, which seemed like madness
05:52 cheus Terminus-, Yes see the - param eg, -%}
05:53 Terminus- cheus: i meant by default, without having to use '-'. =)
05:53 cheus Terminus-, Oh, no -- I hate the '-' but there's no way around it -- totally screws with indentation too
05:54 Terminus- cheus: yep. for, if, etc. making newlines is annoying. =|
05:55 cheus mosen, I agree. I had started developing a pattern around that ... the .ng formulas were prototypes of that... the idea being, use a pillar... for everything, and turn states into domain-specific data serializers with jinja templates for the config-style of the target such as the php.ini example above
05:56 mosen cheus: are you maintaining the php.ng formula?
05:57 mosen cheus: I definitely agree with that style of formula though
05:57 cheus The pillar holds all data but has sane defaults with deep merges of the specified pillar to the sane defaults. Not actively but I'd be interesting in working to adapt it.
05:57 hal58th joined #salt
05:57 cheus I've been out of salt land for a while... spring tends to be the time of year i get to focus on devops
05:57 cheus been meaning to fix it and the ng states
05:59 cheus If you're actually interested in formulas, sign-up for the Salt Formulas google group. I had a good roundtable today with passionate formula people that want to revamp the paradigm and we have some excellent bases and perspectives represented.
05:59 cheus We'll be coordinating further discussions through that google group
05:59 catpig joined #salt
05:59 mosen yeah for sure, I have just been writing modules and states for the moment
06:00 mosen but it would be good to publish some formulas
06:00 iromli joined #salt
06:03 cheus I'm generally happy with the ng states as proofs of concept but the states are too insane as-are, to some degree, the config files. I feel those should be jinja filters for the common types and that the states, which get really complex with this paradigm, might need to be shifted to pure python states. But THAT means we should probably have a state class to enforce some compliance, and a way to register jinja filters from formulas
06:04 cheus Basically, there's a fair amount of engineering to do
06:05 favadi1 joined #salt
06:08 neogenix joined #salt
06:10 mafro joined #salt
06:20 StDiluted joined #salt
06:22 hal58th joined #salt
06:26 I3olle joined #salt
06:29 otter768 joined #salt
06:31 ntropy joined #salt
06:31 germs_ joined #salt
06:32 Terminus- crap. i just realized this would be better if i rewrote everything as macros.
06:32 genediazjr joined #salt
06:33 stevednd joined #salt
06:38 clintberry joined #salt
06:51 felskrone joined #salt
06:52 mdupont joined #salt
06:59 toanju joined #salt
07:00 colttt joined #salt
07:02 clintber_ joined #salt
07:11 mafro joined #salt
07:13 JlRd joined #salt
07:16 AndreasLutro joined #salt
07:25 hustodemon joined #salt
07:27 favadi1 joined #salt
07:27 wnkz_ joined #salt
07:31 Nazca__ joined #salt
07:36 flyboy joined #salt
07:39 jhauser joined #salt
07:43 badon_ joined #salt
07:50 kawa2014 joined #salt
07:52 Auroch joined #salt
07:55 tomspur joined #salt
08:03 Auroch joined #salt
08:03 KermitTheFragger joined #salt
08:03 eseyman joined #salt
08:03 mikkn joined #salt
08:04 bash1245_ joined #salt
08:07 cwright joined #salt
08:08 Auroch joined #salt
08:10 evle1 joined #salt
08:10 evle1 joined #salt
08:12 cbezawad joined #salt
08:12 jri joined #salt
08:21 jchen joined #salt
08:22 jchen joined #salt
08:22 jchen left #salt
08:23 andygrunwald joined #salt
08:23 chiui joined #salt
08:24 devweasel joined #salt
08:25 MTecknology Corey: I was right next to loud obnoxious women nearly screaming. The very largest of them had no concept of personal space and kept her body friendly with mine.
08:26 MTecknology I've had screaming babies next to me and kicking me and I'd have preferred that hands down.
08:26 wincyj joined #salt
08:29 bluenemo joined #salt
08:29 bluenemo joined #salt
08:30 otter768 joined #salt
08:35 karimb joined #salt
08:35 jri_ joined #salt
08:39 bhosmer_ joined #salt
08:40 aurynn joined #salt
08:41 krelo joined #salt
08:43 lb1a joined #salt
08:44 jerematic joined #salt
08:51 lietu joined #salt
08:54 che-arne joined #salt
08:56 fredvd joined #salt
08:58 linjan joined #salt
09:00 netzmonster joined #salt
09:00 evilrob joined #salt
09:02 chandra joined #salt
09:03 yomilk joined #salt
09:04 yomilk joined #salt
09:06 KoFFiE joined #salt
09:07 andreyst joined #salt
09:16 sfxandy joined #salt
09:24 johanek joined #salt
09:38 I3olle joined #salt
09:44 markm joined #salt
09:46 intellix joined #salt
09:47 xsteadfastx joined #salt
09:53 yidhra joined #salt
09:57 cbezawad joined #salt
10:00 slav0nic joined #salt
10:00 slav0nic joined #salt
10:05 ksj hi, how do I get salt to ignore dict entries that don't exist?
10:05 ksj I have a series of nested dicts in yaml, and some of them are optional, but salt jinja rendering chokes on it every time saying they don't exist
10:09 VSpike ksj: probably something like salt['pillar.get'](key, dict())
10:10 VSpike ksj: IOW default to an empty dict, or use another default like '' or 0 or whatever makes sense in your case
10:10 mafro joined #salt
10:12 devweasel joined #salt
10:13 Terminus- ksj: i think i'd use a default filter or "if pillar.foo.bar is defined" which is more verbose.
10:14 Terminus- {{ pillar.foo.bar if pillar.foo.bar is defined else 'baz' }} is yet another way of doing it.
10:14 ksj VSpike, Terminus-, thanks, I'll give it a go
10:15 Terminus- ksj: pillar.foo.bar|default('baz') is probably the most compact.
10:15 ksj I wasn't aware of "is defined"
10:15 ksj Terminus-: yeah, I have |default
10:15 ksj the problem is that the higher level dicts may not be defined
10:16 ksj so I might have [x][y][z]|default(false), but if x or y don't exist, it will fail
10:16 Terminus- ksj: it's in the jinja docs -> http://jinja.pocoo.org/docs/dev/templates/#list-of-builtin-filters
10:16 Terminus- below that are tests.
10:17 VSpike Is the user_data mechanism for ec2 in salt-cloud in the latest production release?
10:17 chandra joined #salt
10:21 aquassaut joined #salt
10:24 xsteadfastx joined #salt
10:28 jcockhren joined #salt
10:29 pf_moore joined #salt
10:29 robinsmidsrod has anyone written a logstash returner?
10:31 otter768 joined #salt
10:31 N-Mi joined #salt
10:31 N-Mi joined #salt
10:33 linjan joined #salt
10:33 jerematic joined #salt
10:34 mike25de left #salt
10:36 devweasel joined #salt
10:37 ksj I'm trying to implement the pillar.get method. Can't figure out how to concatenate strings. I was expecting salt['pillar.get']('x:'+y+':z',default) to work, where y is a defined variable
10:38 markm joined #salt
10:38 phx ksj, is that in a jinja template?
10:38 matthew-parlette joined #salt
10:38 ksj phx: yes
10:39 phx ~ is strinct concatenation, not +
10:39 phx s/strinct/string/
10:39 ksj in a jinja execution (i.e. {%
10:39 ksj ok, thanks
10:40 Tinuva joined #salt
10:42 ksj still failing. the line is salt['pillar.get']('dotfiles:' ~ user ~ ':x:' ~ file ~ ':template', 'false')
10:42 ksj and yes, that's a valid part of the pillar
10:43 phx is there some kind of error message? logs and such?
10:43 ksj Encountered unknown tag 'x'. Jinja was looking for the following
10:43 ksj tags: 'endfor' or 'else'.
10:44 ksj sorry, I have to run out now. thanks for your help. will be back in a couple of hours
10:48 N-Mi joined #salt
10:48 giantlock joined #salt
10:52 matthew-parlette joined #salt
10:54 VSpike Wah! I don't want to have to deal with Windows any more
10:56 sfxandy joined #salt
10:58 zipkid joined #salt
11:00 badon_ joined #salt
11:01 zipkid Hello.
11:02 cbezawad joined #salt
11:04 N-Mi_ joined #salt
11:10 amcorreia joined #salt
11:11 mafro joined #salt
11:11 Singularo joined #salt
11:14 linjan joined #salt
11:27 N-Mi_ joined #salt
11:29 jtang joined #salt
11:31 kaiyou joined #salt
11:33 intellix joined #salt
11:44 kalinka joined #salt
11:45 kalinka hello
11:46 XenophonF left #salt
11:46 XenophonF joined #salt
11:49 kalinka maybe someone can help me to configurate my network via salt on a debian server. Actualy I want just to setup an network birdge for vms but I'm always getting "An exception occurred in this state: Traceback (most recent call last)"
11:50 kalinka do you have any idea how to setup an hypervisor on debian like http://docs.saltstack.com/en/latest/topics/tutorials/cloud_controller.html an red hat?
11:50 linjan joined #salt
11:52 simmel joined #salt
11:55 LotR kalinka: which hypervisor are you using?
12:02 kalinka kvm
12:03 Andre-B joined #salt
12:04 jerematic joined #salt
12:04 Auroch joined #salt
12:09 kalinka thats what im trying with the complete error: https://www.refheap.com/98148
12:10 mikkn joined #salt
12:11 mafro joined #salt
12:13 sfxandy joined #salt
12:21 linjan joined #salt
12:27 johtso joined #salt
12:32 otter768 joined #salt
12:33 XenophonF kalinka: network.managed doesn't take a "bridge" argument
12:33 XenophonF at least not according to http://docs.saltstack.com/en/latest/ref/states/all/salt.states.network.html#salt.states.network.managed
12:34 XenophonF then again, it's in the examples given
12:34 XenophonF so i dunno
12:35 XenophonF kalinka: what version of Salt do you have installed?
12:37 pdayton joined #salt
12:39 kalinka 2014.7.1
12:40 monkey66 left #salt
12:41 kalinka xenophoneF at the docs ragarding the network.managed state they have also type: bridge at there examples, so it should be fine, right?
12:41 MaliutaLap left #salt
12:41 bhosmer_ joined #salt
12:42 kalinka maybe tahts the reason: "When managing bridged interfaces on a Debian or Ubuntu based system, the ports argument is required" ?
12:44 JlRd joined #salt
12:45 giantlock joined #salt
12:45 rjc joined #salt
12:46 XenophonF kalinka: that sounds promising
12:47 kalinka ok thx i will try it with ports
13:02 elfixit joined #salt
13:14 karmatronic joined #salt
13:21 ponpanderer joined #salt
13:21 jcockhren joined #salt
13:27 QuisaZaderak joined #salt
13:28 sk_0 joined #salt
13:29 hojgaard joined #salt
13:30 dyasny joined #salt
13:31 ggoZ joined #salt
13:36 hustodemon hello everyone, is there some salt api expert around? I'm cooperating on saltstack-netapi-client-java on github (as the name suggests, it's a java client for saltstack api) and I've got a simple question about '/jobs' resource.
13:39 DorfOnGolf joined #salt
13:42 karimb joined #salt
13:42 ksj hi, is there a way to dump the interpreted jinja of a state file? e.g. some parameter to state.sls. I can't seem to find state.sls in the list of modules
13:42 ksj in the docs
13:46 subsignal joined #salt
13:47 cpowell joined #salt
13:48 I3olle joined #salt
13:50 XenophonF ksj: what problem are you trying to debug?
13:51 mattiasr joined #salt
13:53 jerematic joined #salt
13:53 ksj I'm just hitting errors and was wanting to see how the state file appeared after being rendered by jinja
13:54 ksj actually, I want to experiment by putting trying things out and seeing what the output is
13:54 ksj if I experiment now, all I get is a (for me) obscure python error
13:54 Nazca joined #salt
13:54 XenophonF what's the state? and what's the error
13:55 XenophonF post them on https://bpaste.net/ or something
13:55 ksj so there's no way of getting the rendered state file?
13:56 ksj I appreciate the help, it's just that I want to experiment with things
13:56 ksj and the way that I learn best is by trying different inputs and seeing what the output is
13:56 XenophonF of course
13:56 XenophonF see http://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.jinja.html#salt.renderers.jinja.render
13:56 ksj but the output I get from salt has already been through several layers of processing, so it becomes a bit of a guessing game
13:57 XenophonF i think you can run that with salt-call
13:57 XenophonF i'm trying to find some examples for you
13:57 ksj I actually think I may be overcomplicating things. I'm trying to differentiate between a file and a directory in a list in my pillar
13:58 ksj so that I can run either file.managed or file.recurse
13:58 ksj but maybe it's better to go the tried and tested route and just have two separate lists
13:59 Nazca__ joined #salt
14:00 XenophonF given a pathname, how would you know which was which?
14:00 XenophonF a trailing directory separator?
14:00 XenophonF looking at the file system?
14:00 jeremyr joined #salt
14:01 stylica joined #salt
14:01 Gareth joined #salt
14:02 ksj I was actually hoping to do it just by appending a / to the item name, but it appears jinja can't do any type of pattern matching
14:02 JDiPierro joined #salt
14:02 ksj unless I missed it in the docs
14:02 Deevolution joined #salt
14:02 SneakyPhil joined #salt
14:02 Nazca__ joined #salt
14:03 XenophonF posting your sls file would really help me understand what you're trying to do
14:03 kossy joined #salt
14:03 devweasel left #salt
14:03 oeuftete joined #salt
14:03 stylica joined #salt
14:04 XenophonF like, say {% set pathname = "foo/" %}
14:06 XenophonF can't you do a string match like {% if re.match(".*/", pathname) %}
14:06 ksj my sls file would probably confuse things, but I'll post it in a second once I've taken out all the garbage. Basically the pillar is a list of dotfiles, but some of them have attributes (e.g. only copy if os is linux). others are directories and therefore need file.recurses rather than file.managed
14:06 ksj ooh, I haven't seen re.match - will try now
14:06 ksj it's not in the jinja documentation
14:06 XenophonF i think jinja will execute that python code
14:07 ksj let me give it a try. thanks a lot
14:07 XenophonF alternatively, if it's a list of dotfiles with (optional) attributes, you can include an attribute like "directory: true"
14:08 ksj XenophonF: yeah, I could, but I was trying to be clever. I wanted the dotfiles list to be as simple and concise as possible
14:08 ksj so that option is a last resort if I can't get regex to work
14:09 Gareth joined #salt
14:09 kossy joined #salt
14:09 oeuftete joined #salt
14:10 kossy joined #salt
14:11 XenophonF yeah, if you go the optional attribute route, you'd have to add logic that can tell whether a given item in your list is a string or a hash
14:11 XenophonF makes iterating through the dotfile list a little more complicated
14:11 XenophonF doesn't file.recurse work on regular files, though?
14:12 ksj unfortunately, that bit of jinja regex doesn't work
14:12 ksj failed: Jinja variable 're' is undefined
14:13 XenophonF hm, i guess it comes down to the re namespace not being loaded
14:13 thayne joined #salt
14:18 XenophonF ksj: http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.state.html#salt.modules.state.show_sls
14:20 jerematic joined #salt
14:21 shannonh_ joined #salt
14:21 XenophonF but that only shows state data, while you want the SLS file post-render but before processing
14:23 nitti joined #salt
14:24 racooper joined #salt
14:24 Jouke joined #salt
14:26 Jouke what would be the way to fill configuration files with data out of pillars? For examle a config file with  'url=%url%' and automatically map that what is in between the %'s to a pillar config?
14:29 sfxandy joined #salt
14:31 jdesilet joined #salt
14:32 mikeywaites joined #salt
14:33 otter768 joined #salt
14:33 mikeywaites Hey guys,  Im trying to pdb in a reactor even handler i have written.  I manually run sudo salt-master -l debug and i see (Pdb) appear in the logs but i dont get a tty ?
14:33 mikeywaites anyone know how this should be done ?
14:34 smcquay_ joined #salt
14:34 __gotcha joined #salt
14:36 quist joined #salt
14:38 clintberry joined #salt
14:44 teall joined #salt
14:53 justyns joined #salt
14:54 kermit joined #salt
14:55 kaptk2 joined #salt
14:57 XenophonF Jouke: many (all?) of the saltstack-formulas use pillar data in sls files and configuration templates; look to them for examples - https://github.com/saltstack-formulas
14:59 yomilk joined #salt
14:59 perfectsine joined #salt
15:01 quist left #salt
15:04 out180 joined #salt
15:05 zwi joined #salt
15:06 XenophonF ksj: maybe this will do what you want - http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.cp.html#salt.modules.cp.get_template
15:07 hybridpollo joined #salt
15:08 _ale1_ joined #salt
15:11 debian112 joined #salt
15:12 justyns joined #salt
15:13 getling joined #salt
15:14 dude051 joined #salt
15:14 StDiluted joined #salt
15:15 andrew_v joined #salt
15:15 getling Hrm...burning in a new salt installation and I'm running into a situation where some runs over a large number of nodes hangs before exiting, requiring a ctrl-C to kill it...is this a common issue?
15:15 nich0s joined #salt
15:19 faliarin joined #salt
15:21 TomJepp joined #salt
15:22 murrdoc joined #salt
15:22 iggy ksj: {% if listentry[:-1] is '/' %} ?
15:23 iggy not sure if negative lookups like that work in jinja, but it's worth a shot
15:23 lumtnman joined #salt
15:24 perfectsine joined #salt
15:26 johnkeates joined #salt
15:26 johnkeates I have a "Error encountered while render pillar top file." with no further explanation
15:27 johnkeates debug and trace shows nothing, minion, salt-call and master logs give nothing, yaml validates and all files and paths and references exist
15:27 iggy johnkeates: paste
15:28 kiorky johnkeates: ensure top level pillar ids are unique.
15:28 johnkeates http://pastie.org/10005129
15:29 iggy bad quotes
15:29 johnkeates need double-quotes?
15:29 johnkeates err
15:29 johnkeates single
15:29 johnkeates i see it now
15:30 johnkeates i guess this is what happens when you stare at something too much.
15:30 iggy they all look like they are utf quotes (i.e. slightly turned)
15:30 johnkeates oh, yeah, but that's WebKit's fault
15:30 johnkeates it's actually plain-text in vim
15:31 johnkeates hmm
15:32 getling its timing out on connecting to eventfd
15:33 iggy after saltconf, we're all going to be inclined to blame zermq
15:33 johnkeates still not working
15:33 johanek joined #salt
15:34 johnkeates http://pastie.org/pastes/10005129/text (without webkit/pastie doing it's UTF thing)
15:35 iggy so how do you have the two different top.sls's working?
15:36 johnkeates like this: http://pastie.org/pastes/10005147/text
15:36 johnkeates in pillar_roots
15:36 iggy and you only have 2 top.sls files (not just in dev env)
15:37 johnkeates each env has a top yes
15:37 johnkeates i'll paste those too
15:37 iggy protip: gist.github.com allows multiple files per paste
15:38 johnkeates ah, that's nice yes, forgot about that
15:38 getling I'm wondering if this is bug 20639 that I'm hitting
15:38 getling https://github.com/saltstack/salt/issues/20639
15:38 QuisaZaderak joined #salt
15:38 Brew joined #salt
15:40 getling even after taking the services down, lots of TIME_WAIT sockets
15:40 johnkeates https://gist.github.com/johnkeates/801b62901830cb147d89
15:41 mikeywaites trying to debug a reactor state by using pdb; pdb.set_trace() and runing the salt master in the foreground but i keep getting a cannot render state error? this should be possible right?
15:41 iggy johnkeates: empty test:
15:41 iggy try "test: []"
15:42 johnkeates iggy: ah
15:42 CheKoLyN joined #salt
15:43 johnkeates excellent, that was the problem
15:44 iggy ^5
15:44 johnkeates i guess it's not as much a syntax problem as a 'trying to parse something that's not there' problem
15:44 johnkeates i didn't expect salt to try to read stuff that's simply not there
15:44 iggy it's pyyaml failing on that
15:45 iggy where did it succeed to parse?
15:45 johnkeates http://yaml-online-parser.appspot.com
15:45 johnkeates uses pyyaml :/
15:45 iggy that file passed too?
15:45 iggy or just the other 2?
15:45 XenophonF i'm trying to set up s3fs on my master, but as far as i can tell it isn't working - what's the trick?
15:45 johnkeates yeah, it passed there too
15:46 XenophonF i have s3.key and s3.keyid set in /etc/salt/master.d/s3.conf
15:46 johnkeates i actually catted the whole bunch together and it parsed to python just fine
15:46 iggy johnkeates: weird, it shouldn't... but it's fixed now
15:46 pdayton joined #salt
15:46 johnkeates iggy: yep :) thanks for spotting my mistake in there. I usually don't leave empty lines dangling in my files.. oh well
15:47 XenophonF i also set s3.buckets and s3.service_url
15:47 XenophonF but cp.list_master doesn't show anything from the bucket
15:47 iggy johnkeates: on the up side, I think 2015.2 would have given you a much better error message
15:47 iggy XenophonF: did you try salt-run -l debug fileserver.update?
15:48 XenophonF no i haven't - let me give that a try
15:48 johnkeates iggy: I wish 2015.2 was in the repo's already :)
15:48 murrdoc engines! https://github.com/saltstack/salt/blob/develop/salt/engines/sqs_events.py
15:48 iggy don't get me started
15:48 murrdoc iggy: I wish 2015.2 was in the repo's already
15:48 * murrdoc puts coin in machine
15:48 johnkeates haha
15:48 * johnkeates pushes all the buttons
15:48 iggy I'm pretty sure the pi of base and thatch are both tired of hearing about it too
15:49 murrdoc t to the Hatchet!
15:49 XenophonF iggy: this is what i get back - https://bpaste.net/show/444f437cce6b
15:49 XenophonF i can see it reading the s3fs.conf config file i made
15:50 iggy aside from the conf read (which you would see with any command btw), I don't see anythign about s3 in there
15:50 XenophonF yeah, i know
15:51 XenophonF let me make sure fileserver_backend is set correctly
15:52 jeremyr1 joined #salt
15:52 XenophonF it looks ok, as it's set to git, s3, roots
15:52 iggy can you salt-call config.get all the stuff involved?
15:53 XenophonF isn't that just the minion?
15:54 iggy config.get will pull minion conf -> pillar -> grains -> master conf
15:55 XenophonF i'm trying to get s3fs working on the master
15:55 iggy it's just an idea... I've never even used s3
15:55 murrdoc https://github.com/saltstack/salt/releases/tag/v2015.2.0rc1 "https://github.com/saltstack/salt/compare/v2015.2.0rc1...develop to develop since this tag "
15:55 XenophonF salt-call config.get s3 returns nothing
15:56 XenophonF salt-call config.get s3.keyid returns the right thing
15:56 iggy I get that... I'm just trying to see if it can see what you think it can
15:56 iggy murrdoc: you home already?
15:56 XenophonF same for s3.key, s3.buckets, and s3.service_url
15:56 XenophonF so at least the config's getting read on the master
15:56 johnkeates left #salt
15:56 murrdoc yeah late last night
15:57 murrdoc then up early to watch india squeak by west indies
15:57 iggy XenophonF: config.get fileserver_backend ?
15:57 murrdoc now we work
15:57 iggy hah! ...travel day
15:57 XenophonF iggy: that just shows roots
15:57 murrdoc does s3 show up in the roots
15:57 iggy XenophonF: so maybe we're on to something here
15:57 XenophonF that ought to show git, s3, and roots
15:58 iggy make sure it's not set in some other file somewhere and the master is restarted
15:58 XenophonF yeah i'm grepping now
15:59 nich0s joined #salt
15:59 XenophonF it looks right in the master config
15:59 iggy nothing in master.d or anything?
15:59 XenophonF nope
15:59 iggy restarted the master?
16:00 XenophonF only one occurrence of ^fileserver_roots
16:00 XenophonF let me do that
16:00 murrdoc xeno not that u asked, but i have reduced the master config in my setup to just have default_include, verify_env: False
16:00 murrdoc then put everything else in sensibly named files in the conf.d directory
16:00 murrdoc and use salt to manage that
16:00 murrdoc not that u asked
16:00 XenophonF murrdoc: thanks for the suggestion - i'm actually using salt-formula to configure the master
16:00 __gotcha joined #salt
16:01 murrdoc cool
16:01 XenophonF so everything's in _defaults.conf, generated by a jinja template
16:01 * murrdoc shuts up
16:01 XenophonF hah
16:01 XenophonF no worries!
16:01 iggy reminder: need to push all my changes back upstream for salt-formula
16:01 murrdoc heard that
16:01 yomilk joined #salt
16:02 XenophonF weird
16:02 XenophonF let me try a highstate on a minion, make sure it's pulling configs from git
16:02 murrdoc u know what i like about salt docs, they make sense in a file path sort of a way
16:02 iggy welcome to pythonj
16:03 iggy welcome to python
16:03 smcquay joined #salt
16:03 murrdoc thatch or anyone else who knows him, i would like to apologize for saying that using beacons for monitroing an <enter> at the command line was excessive, i am thankful for beacons
16:03 murrdoc iggy:  i done a ton of python
16:03 XenophonF brb
16:04 murrdoc i just appreciate being able to use the url as a directory
16:04 murrdoc for eg http://docs.saltstack.com/en/latest/topics/beacons/, http://docs.saltstack.com/en/latest/ref/beacons/all/
16:04 murrdoc that works for almost all, s/topics/ref and s/$/all/
16:05 mpanetta joined #salt
16:05 iggy we still need a bot
16:05 iggy I'm getting lazier and lazier
16:05 murrdoc MTecknology:  offered to bring one in
16:05 iggy plane, gotta run
16:06 murrdoc have fun
16:07 _ale1_ joined #salt
16:09 lothiraldan joined #salt
16:09 StDiluted joined #salt
16:11 micah_chatt joined #salt
16:11 MTecknology murrdoc: hm? what did I do?
16:12 murrdoc werent u mentioning bringing a bot in the chat
16:12 MTecknology oooh
16:12 murrdoc or was it someone else
16:12 MTecknology it was me
16:12 MTecknology However, I'm not going to do it without staff telling me to go for it.
16:13 MTecknology saltstack staff... not freenode
16:13 murrdoc ask em
16:13 murrdoc or bring it in
16:13 murrdoc and take it out later
16:13 murrdoc and ask for forgiveness
16:14 perfectsine joined #salt
16:14 MTecknology :P  I'd prefer not do that in this realm
16:14 smcquay joined #salt
16:15 murrdoc good call
16:15 smcquay joined #salt
16:16 intellix joined #salt
16:16 XenophonF back
16:17 XenophonF iggy: bon voyage
16:17 clintberry joined #salt
16:18 programmerq joined #salt
16:20 iggy i found it odd that most of the people at the conf looked like bums but you could probably pick any ten people out of the crowd and their combined salary would be 1M+
16:20 ccarney_ROCC joined #salt
16:21 WesleyTech joined #salt
16:23 jmccree It's not our job to look pretty.
16:23 ccarney_ROCC left #salt
16:24 XenophonF maybe i've got the wrong IAM API key?
16:25 Setsuna666 joined #salt
16:26 nich0s joined #salt
16:28 I3olle joined #salt
16:28 XenophonF alright, s3.get works
16:28 XenophonF now let's see if i can get s3fs to work
16:29 murrdoc \o/
16:29 Setsuna666 We are currently using salt version 2014.7.1 and are wondering if anyone has the same problem as this one. When running state.highstate on a batch of minions, it can happens that the master lose the connection with them, but the state are still running on the minions.
16:30 murrdoc thats by design
16:31 murrdoc read http://docs.saltstack.com/en/latest/topics/jobs/, http://docs.saltstack.com/en/latest/ref/cli/salt-run.html to understand how to get visibility into whats happening to your jobs
16:31 Setsuna666 We then lose the return output from the state execution from the minion, which we normally get
16:32 Setsuna666 But normally, when the master send a job to the minions, he wait a certain time to receive the states execution output ?
16:33 XenophonF see the "-t" option to the salt command
16:33 XenophonF you can have it wait longer for the response, if you want to receive it interactively
16:33 Setsuna666 @XenophonF Thanks, will try
16:34 otter768 joined #salt
16:34 johanek joined #salt
16:35 georgemarshall joined #salt
16:35 XenophonF but salt-run -l debug fileserver.update shows only git again
16:36 XenophonF and cp.list_master doesn't show the contents of the s3 bucket
16:36 tligda joined #salt
16:37 theologian joined #salt
16:37 XenophonF is there an s3 equivalent to gitfs_remotes?
16:37 murrdoc not sure, will have to google
16:37 rgarcia_ joined #salt
16:37 XenophonF it isn't mentioned on the s3fs page
16:38 XenophonF maybe i've got the directory structure of the bucket wrong
16:40 murrdoc what are u trying to do
16:40 murrdoc i missed the conversation
16:41 XenophonF ultimately, i'm trying to deploy windows packages
16:41 XenophonF the installers are too big for github, so i stashed them in s3
16:41 XenophonF this used to work - i'd just have the minion pull them down over https
16:42 XenophonF only now that broke, and windows minions throw sslerror when i try to use https:// URLs in states or package definitions
16:43 bhosmer joined #salt
16:43 XenophonF i also want to take advantage of the new "cache_dir" option
16:43 XenophonF which will let me distribute multi-file windows installer packages via salt
16:43 XenophonF anyway
16:43 XenophonF i want to stash files referenced by salt:// URLs in s3
16:44 XenophonF as a workaround i can put them on the master
16:44 XenophonF i can't put them in git because they're too big
16:44 SheetiS joined #salt
16:45 signull joined #salt
16:45 evidence when running multi master failover, is there any reason to use mm pki over just sharing the master keys between them?  aside from a security breach compromising one master rather than all
16:48 murrdoc XenophonF:  did u upgrade salt versions, when did they start breaking
16:49 XenophonF the https:// URLs? those used to work in 2014.1 but broke in 2014.7
16:49 XenophonF wasn't tracking it closely
16:49 XenophonF thought it was a bug in my configs and didn't really dive into it until recently
16:51 thayne joined #salt
16:51 murrdoc weird, that could be urllib2 issues also
16:51 XenophonF well, like i said, i need to switch to salt:// URLs anyway
16:51 murrdoc if it doesnt work use like cp.cache_file with module.run to get it working till they fix it upstream
16:51 __gotcha joined #salt
16:52 XenophonF cache_file doesn't work either
16:52 XenophonF at least not for https
16:52 XenophonF there's an error returned by requests somewhere, i think it has to do with certificate validation
16:52 murrdoc or use something like this https://github.com/saltstack/salt-contrib/blob/master/states/archive.py
16:52 murrdoc orly
16:55 shaggy_surfer joined #salt
16:55 shaggy_surfer joined #salt
16:56 _JZ_ joined #salt
16:58 MTecknology iggy: I would definitely injure that assumption (10 ppl >= $1M)
16:58 bluenemo joined #salt
16:58 bluenemo joined #salt
16:59 XenophonF ah forget it - i'm just going to put these files on the master
16:59 ajw0100 joined #salt
16:59 murrdoc :(
16:59 murrdoc try the custom state thing
16:59 murrdoc ?
16:59 XenophonF i just want to be able to push out my antivirus software, is that so wrong? :)
17:00 murrdoc (hug)
17:00 XenophonF hah thanks
17:01 pravka joined #salt
17:01 XenophonF maybe i'll try again after folks are back from the conference
17:01 germs_ joined #salt
17:01 desposo joined #salt
17:01 murrdoc ok
17:01 murrdoc http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.s3.html#salt.modules.s3.get
17:01 perfectsine joined #salt
17:01 murrdoc try that with module.run
17:01 XenophonF that works, actually
17:02 murrdoc oh then use that
17:02 * murrdoc waves flag, keep masters clean
17:02 murrdoc http://docs.saltstack.com/en/latest/ref/states/all/salt.states.module.html#salt.states.module.run
17:02 XenophonF i am totally with you
17:03 XenophonF however i'm also waving the "keep hacks out of shared state/package defs"
17:03 murrdoc damnit
17:03 XenophonF flag
17:03 XenophonF yeah i know
17:04 XenophonF it's a great idea and if everyone on my team were l33t haxxors, i'd be down with it
17:04 lz-dylan Howdy, folks!
17:04 XenophonF gotta keep things simple for now
17:05 lz-dylan Is there a current favorite way to store secure data in pillars that keeps 'em out of source control? I'm eyeing the S3 external pillar because it can be encrypted at rest, but maybe there's a new shiny that came up at SaltConf :)
17:05 Gareth mroning morning
17:05 XenophonF hah well i was just beating my head on s3fs
17:05 Gareth morning too.
17:06 XenophonF so lz-dylan, do you have s3fs working?
17:06 lz-dylan XenophonF: this is before actually setting anything up, to be honest
17:06 XenophonF bummer
17:06 lz-dylan Right? I've managed to get gitfs working before without issue. Not in the project I'm on now, though.
17:07 lz-dylan I haven't looked at the code -- unsure if the external pillar has much overlap with the s3 module
17:07 XenophonF well, here's the docs i'm reading - http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.s3.html and http://docs.saltstack.com/en/latest/ref/file_server/all/salt.fileserver.s3fs.html
17:07 XenophonF but i haven't been able to get s3fs to work
17:07 lz-dylan ...and I'm just over on http://docs.saltstack.com/en/latest/ref/pillar/all/salt.pillar.s3.html
17:08 lz-dylan (backscrolling a bit to see what you've been up to, onesec)
17:09 lz-dylan so, I know this is the equivalent of "is it plugged in" but you've done up your file_roots and all that in /etc/salt/master, yeah?
17:09 XenophonF yeah i have
17:10 XenophonF i'm using salt-formula so i have to do a little bit of manual tweaking
17:11 lz-dylan actually, I'd love to know how that works out for you
17:11 lz-dylan salt-formula is both tempting and intimidating.
17:11 XenophonF it lacks support for windows minions :(
17:11 XenophonF otherwise i like it
17:11 aparsons joined #salt
17:12 lz-dylan it's fascinating to me that so many people use saltstack for windows -- it definitely feels posixy to me. Are there just no other good windows provisioning options? or is the saltstack approach that darned good even on windows?
17:13 XenophonF lz-dylan: it's more like i lack the wherewithal to deploy SCCM
17:13 chiui joined #salt
17:13 lz-dylan (I'm not really that windows-literate anymore, excuse any ignorance)
17:13 XenophonF can't speak for others
17:13 shaggy_surfer joined #salt
17:13 lz-dylan I mean, salt probably _would_ be my approach, but you'd have to be crazy to hire me to admin windows machines
17:13 XenophonF i'd love to see better support for windows in salt, personally
17:14 lz-dylan how's the learning curve on salt-formula?
17:14 XenophonF pretty low
17:14 XenophonF bootstrapping's the hardest part
17:14 lz-dylan and I'm gonna grab some coffee &c and then probably fiddle with s3 a bit
17:14 XenophonF hang on and i'll post my bootstrapping script for centos7
17:15 lz-dylan sweet, thanks! I'm on deb/ubuntu, but it's worth a look
17:15 XenophonF https://bpaste.net/show/102cf74ba297
17:16 SheetiS joined #salt
17:16 XenophonF to bootstrap a new salt master using salt-formula, you have to checkout salt-formula to a temp directory
17:16 Ahlee we use salt on windows too so we have single entry poin
17:16 XenophonF in a nutshell
17:17 murrdoc is that your github pub key
17:17 XenophonF not any more it isn't!
17:17 murrdoc hahaha
17:17 XenophonF omfg what a newb mistake
17:17 Ahlee lol
17:18 murrdoc git clone xeno's job
17:18 murrdoc works for me
17:18 lz-dylan hahahaha, ruh-roh!
17:18 lz-dylan I'll be nice and leave it be
17:18 Gareth Everyone has pasted a password like something into IRC, anyone who says they have...is lying :)
17:18 XenophonF nah it's revoked already
17:19 murrdoc :)
17:19 murrdoc i should cloned something first
17:19 lz-dylan Gareth: IRC or worse
17:19 murrdoc then told u
17:19 XenophonF i noticed right before you said something
17:19 murrdoc riiiight
17:19 XenophonF was admiring my handiwork
17:19 XenophonF what a nice shell script, i thought to myself
17:19 XenophonF then i finished scrolling to the bottom
17:19 murrdoc hahaha
17:20 murrdoc fridays man
17:20 Gareth lz-dylan: I know a few people that accidentally sent stuff to twitter :)
17:20 jmccree relevant: http://www.infosecnews.org/want-to-know-the-wifi-password-for-the-brasil-world-cup-security-center/
17:20 murrdoc some good talks from this conf
17:20 murrdoc https://github.com/garethgreenaway/presentations/blob/master/SaltConf2015-Dr-SatlStack.pdf
17:20 murrdoc http://www.slideshare.net/stevendgonzales/saltconf2015-saltstack-at-scale-automating-your-automation
17:21 murrdoc http://www.slideshare.net/lovenyberg/using-saltstack-to-orchestrate-microservices-in-application-containers-at-saltconf15
17:21 murrdoc at least those are the good ones i have slides for
17:21 theologian joined #salt
17:21 lz-dylan Gareth: apparently there's a few scraper bots that just sit there, watching, waiting for you to push keys to public github repos
17:21 Gareth murrdoc: I heard good things about that DrSaltStack talk :)
17:21 murrdoc yeah no derailment about utc this time
17:21 lz-dylan murrdoc: awesome, thank you! I really want all the talks :D
17:22 murrdoc was a good talk Gareth
17:22 murrdoc i heard the cloudflare one was a good talk
17:22 murrdoc missed that
17:22 murrdoc linkedin gave a good talk about what they have done to optimise, no links yet
17:22 murrdoc the pyobjects talk was nice
17:23 murrdoc iggy's talk was most fun
17:23 lz-dylan speaking of, has anyone had lots of issues with cloudflare?
17:23 * murrdoc withdraws from conversation
17:23 lz-dylan my company sees quite a few errors; I'm still thinking it's problems with our underlying app, but with cf sitting in front it's hard to see.
17:24 Gareth murrdoc: Thanks :)
17:24 lz-dylan ooooor I could scare people off :P
17:24 XenophonF ah it's time for new ssh keys for that service account anyway
17:24 murrdoc lz-dylan:  I step away from the conversation cos cf is a competitor so you know :)
17:24 XenophonF thanks for the links murrdoc
17:24 giantlock joined #salt
17:25 murrdoc Ahlee:  stop slacking share links for slides please
17:25 lz-dylan murrdoc: gotcha. without further info, can I ask who you're working with?
17:25 pdayton joined #salt
17:25 murrdoc working with or for ?
17:26 murrdoc i work for a competitor
17:26 murrdoc one of the older ones
17:28 lz-dylan fair 'nuff. We might end up a customer of yours someday. :P
17:29 murrdoc do it
17:29 cwright joined #salt
17:30 murrdoc http://eventmobi.com/saltconf15/agenda/94463/505543 was an excellent talk too
17:30 johanek joined #salt
17:31 ndrei joined #salt
17:34 jab416171 joined #salt
17:35 MatthewsFace joined #salt
17:36 KyleG joined #salt
17:36 KyleG joined #salt
17:36 Morbus joined #salt
17:37 Morbus joined #salt
17:38 hal58th joined #salt
17:40 jalbretsen joined #salt
17:42 neogenix joined #salt
17:44 student1 joined #salt
17:45 student1 Hi, all, have a problem with ebuild - How to enforce it to recompile packages when portage keywords changed?
17:47 lumtnman joined #salt
17:48 student1 I think, that We could add into ebuild.py - emerge option --changed-use,
17:50 yomilk joined #salt
17:51 student1 I think that in function install add check portage keywords and if they changed then run emerge --changed-use... instead of emerge ...
17:51 hal58th joined #salt
17:52 student1 but I don't know, Who can help me, how to make it in right way
17:52 student1 but I don't know, Who can help me, to make it in right way
17:52 XenophonF student1: where's ebuild.py from?
17:53 student1 salt/modules/ebuild
17:54 XenophonF oh i see now
17:54 XenophonF OK
17:54 XenophonF so here's what you can do
17:54 XenophonF in the short term, create a directory called "_modules" in your state tree
17:55 XenophonF (e.g., "mkdir /usr/local/etc/salt/states/_modules" on FreeBSD, "mkdir /srv/salt/_modules" on RHEL, etc.)
17:55 hasues joined #salt
17:55 student1 then copy to this directory ebuild and prepare it
17:55 XenophonF then copy the patched version of ebuild.py there
17:56 hasues left #salt
17:56 XenophonF minions will pick up and load the patched file at the next highstate
17:56 XenophonF see http://docs.saltstack.com/en/latest/ref/modules/index.html for the gory details
17:57 XenophonF in the long term, you want to get your patches to saltstack, in which case see http://docs.saltstack.com/en/latest/topics/development/contributing.html for instructions
17:57 [7hunderbird] joined #salt
17:58 XenophonF basically, you file a pull request via GitHub
17:58 [7hunderbird] good morning.  I'm working on creating a basic Vagrantfile, etc. and would like to know if anyone had links to some examples I could read? I'm really looking for basics to get started with righting simple state files.
17:59 murrdoc https://github.com/timoguin/saltstack-dev-env
17:59 student1 I understand, thank you a lot, <XenophonF>
18:01 [7hunderbird] thanks murrdoc
18:01 murrdoc thank timoguin
18:01 murrdoc he did the work
18:05 lz-dylan so, not that I've really spent the time to use it much, but is halite still under development?
18:05 lz-dylan looks like nothing's been committed to it in ages
18:08 SheetiS I've heard good things about saltpad https://github.com/tinyclues/saltpad
18:08 murrdoc +1
18:09 aquinas joined #salt
18:09 lz-dylan cool, glad there's something out there :)
18:09 SheetiS I've been thinking about taking saltpad and extending it with my roles ext_pillar to also be a management place them.
18:10 lz-dylan I don't think I'll really be looking too closely until there's a GUI-ish thing in at least beta state, but it'd be nice to have a shiny thing to show folks outside of ops
18:11 mafro joined #salt
18:11 XenophonF wow saltpad looks nice from the screenshots
18:14 numkem is there issues with the exclude statement in an sls?
18:14 numkem I'm trying to resolve a conflict and no matter what I do with the exclude (either sls or id) it doesn't seem to affect the result, still getting an error message saying I have conflicting ids
18:16 SheetiS numkem: any reason you can't make the ids unique and live side-by-side?
18:16 numkem SheetiS: I guess I could change their names but in this case they refer to the same files and services
18:16 spookah joined #salt
18:17 numkem My case is an ntp server vs it's client, its the same files
18:17 numkem I guess i could just source a different file...
18:17 murrdoc i use ids and explicit names to get around this issue
18:17 numkem murrdoc: what do you mean with explicit names?
18:17 murrdoc name: ntp
18:17 murrdoc in the states
18:18 murrdoc instead of doing ntp: pkg.installed
18:18 numkem murrdoc: ok ok, I see
18:18 Corey MTecknology: Yeah, "I don't fit in the seat" is a common problem for some air travelers.
18:18 numkem thanks ill try that
18:19 TaiSHi m00
18:19 TaiSHi murrdoc: home already?
18:23 numkem murrdoc: that worked. So I understand that the ids must at all times be unique event if they should be excluded. Good to know
18:25 murrdoc numkem:  my understanding is all states basically become one giant dict
18:25 murrdoc so its easier to avoid conflicts by using unique-ish names
18:26 wincus joined #salt
18:27 __gotcha joined #salt
18:28 forrest joined #salt
18:28 forrest joined #salt
18:29 JDiPierro joined #salt
18:29 rap424 joined #salt
18:32 pdayton joined #salt
18:34 otter768 joined #salt
18:35 bmac2 joined #salt
18:36 micah_chatt joined #salt
18:36 teall joined #salt
18:38 Gareth Corey: An instance of when *Space* is an issue.
18:39 aparsons joined #salt
18:39 micah_chatt_ joined #salt
18:45 MTecknology Corey: She could have fit in /those/ seets. She was turning around and leaning back against me and grabbing me every time she wanted to say anything to me.
18:46 murrdoc what
18:46 murrdoc provide some context
18:47 Ahrotahntee is it weird to write 'uninstall' states that fully purge packages, configs, and logs?
18:47 MTecknology murrdoc: me?
18:47 TaiSHi murrdoc: wat
18:48 babilen Ahrotahntee: Yes, it is: http://docs.saltstack.com/en/latest/ref/states/all/salt.states.pkg.html#salt.states.pkg.purged
18:48 babilen (use that)
18:48 MTecknology babilen: why is it weird?
18:48 Ahrotahntee babilen: .purged doesn't eliminate logs, or non-default config files in debian
18:48 babilen MTecknology: Because you wouldn't want to write that yourself, but use the aforementioned state
18:49 Ahrotahntee babilen: is it supposed to?
18:49 MTecknology ah, I thought he meant, is it weird using pkg.purged
18:49 babilen Ahrotahntee: Well, I would not make that a single state, but multiple ones. But then: I would probably not have installed something I don't want on new boxes
18:50 arif-ali joined #salt
18:50 Ahrotahntee babilen / MTecknology: https://gist.github.com/Ahrotahntee/d70f7ee26a0c358f715f example
18:50 babilen always helps
18:50 baweaver joined #salt
18:50 murrdoc try 47 with state:nginx
18:51 Ahrotahntee you can watch a state?
18:51 murrdoc me thinks
18:52 baweaver joined #salt
18:52 babilen Ahrotahntee: I would also recommend to use the short-dec-style (pkg.purged: [[), but that isn't weird. I mean I don't quite understand why you need that in the first place (just don't install nginx on new boxes to begin with), but if this is meant for manual applications or transitions then fair enough
18:53 nich0s joined #salt
18:53 notnotpeter joined #salt
18:53 babilen Ahrotahntee: I understood you in the sense that you wrote the entire state (the actual code that deletes the files) yourself
18:54 Ahrotahntee I wrote it because while writing my nginx sls I needed to purge it a bunch of times to get the event/reactor relationship working right :V
18:55 Ahrotahntee it's not something I would apply via a top.sls file - I've engineered my states to be 'sticky', once you apply it, it gets a role: list value added that keeps the state up to date with highstate
18:57 Sypher joined #salt
18:58 Ahrotahntee I also wrote a zero-conf tinc state, but couldn't get past the race-condition that is assigning IPs to nodes.
18:58 Ahrotahntee (they all check to see what IPs are in use at the same time and all end up with 10.0.0.1)
18:58 MTecknology Ahrotahntee: service:nginx isn't really needed because pkg:nginx will kill that service
19:00 arif-ali joined #salt
19:00 Ahrotahntee huh, I can't remember why I put that in there
19:01 kermit joined #salt
19:02 ccarney_ROCC joined #salt
19:02 ccarney_ROCC left #salt
19:03 baweaver joined #salt
19:06 msciciel_ joined #salt
19:07 bluenemo joined #salt
19:08 triplecorn joined #salt
19:08 nitti_ joined #salt
19:10 triplecorn joined #salt
19:11 ajw0100 joined #salt
19:11 gargrag joined #salt
19:11 gargrag Hi all, !
19:11 gargrag how are you doing.
19:11 mafro joined #salt
19:12 murrdoc anyone here doing a state that requires a reboot, how are u guys handling that
19:12 gargrag Can anyone help me ? I am creating a state, that includes mysql and nginx formula, which is the correct way to override some default pillar variables.
19:13 gargrag so you can use the state as is.
19:13 TaiSHi gargrag: most of them should be on map.jinja inside formulas
19:13 murrdoc use the pllar.example from the formula and put it in your pillars directory/file_root
19:14 XenophonF murrdoc: i'm writing a state that runs dcpromo (Active Directory domain controller installation), and that will require a reboot
19:14 TaiSHi murrdoc: he's trying to avoid setting up a distinct pillar
19:14 gargrag yes TaiSHi, but suppose y want my own formula (the one who includes mysql) to install mysql server, on port 3390 by default
19:14 XenophonF i'm thinking of having a state that runs last, after the rest of the run
19:14 XenophonF and have it do a scheduled restart
19:14 murrdoc and use listen_in ?
19:14 gargrag it's ok to override in the map.jinja of the new state?
19:14 XenophonF yeah, using listen
19:14 murrdoc right
19:15 murrdoc that sounds good
19:15 murrdoc i am thinking of making it fire off an event
19:15 TaiSHi map.jinja will most likely have default values, you can override as murrdoc said using pillar/top.sls
19:15 murrdoc so highstate finishes
19:15 XenophonF exactly
19:15 XenophonF and 60s later, the computer reboots
19:15 gargrag but i can't distribute this satate as a formula for example,
19:16 murrdoc or better yet
19:16 murrdoc put a grain ?
19:16 Setsuna666 joined #salt
19:16 murrdoc wait it must be possible to check from the reactor that the minion is running jobs
19:16 murrdoc so maybe the reactor state should check for running jobs then reboot
19:16 gargrag the thing i want to do, is to create a new formula, that includes two other formulas.
19:17 gargrag and it can't depend on putting a pillar file,
19:17 perfectsine_ joined #salt
19:17 TaiSHi Formulas tend to have a default setting in map.jinja
19:17 TaiSHi Plus users can customize using pillars
19:18 Setsuna666_ joined #salt
19:18 gargrag so TaiSHi, the question is, can i override this from my own map.jinja
19:18 SheetiS I've used the reactor to check for changes on the return of a particular state inside of a highstate to trigger another state elsewhere.  looking at salt/job/*/ret/* or whatever
19:18 p0rkbelly joined #salt
19:18 forrest please don't store pillar data in map.jinja, they take longer to render
19:18 amcorreia joined #salt
19:18 TaiSHi gargrag: yeah
19:19 gargrag ok, i was in doubt if it is a good practice, because on the documentation i saw only os related variables on map.jinja
19:19 gargrag thanks
19:19 gargrag TaiSHi:
19:20 forrest You should NOT be putting non-os related stuff in the map.jinja gargrag
19:20 forrest that's bad practice
19:21 murrdoc totes should be documented forrest
19:21 murrdoc not disagreeing
19:21 murrdoc just saying
19:21 forrest if anything, create a settings.sls file like https://github.com/saltstack-formulas/hadoop-formula/blob/master/hadoop/settings.sls
19:21 forrest if you can't use pillar
19:22 forrest murrdoc: It's documented here regarding lookup tables: http://docs.saltstack.com/en/latest/topics/development/conventions/formulas.html#when-to-use-lookup-tables
19:22 forrest but you are correct, it doesn't say specifically not to put it into the map.jinja
19:22 mimiandi joined #salt
19:22 murrdoc nah its pretty clear on what to put in the map.jinja
19:23 * murrdoc redacts previous request
19:23 forrest murrdoc: Right, but it also doesn't specifically say NOT to, which would be useful, once in an example, once in writing :P
19:23 gargrag so forrest there's a way to do the thing i am willing?
19:23 forrest Well, I'd say use pillar, but if you can't use pillar, look at what I linked above with the settings.sls
19:23 murrdoc forrest:  aight then fix it !! :D
19:23 mimiandi jacksontj: ping
19:24 mimiandi jacksontj: enjoyed your talk very much the other day!
19:24 sjwoodr joined #salt
19:24 gargrag may be i don't have to make this formula including other two
19:24 baweaver joined #salt
19:24 gargrag forrest: ^
19:25 gargrag the thing i am looking for , is for example to have a "wordpress-formula" that , installs all the stack
19:25 gargrag if you include it , on your states.
19:25 williamthekid joined #salt
19:26 ckao joined #salt
19:26 forrest gargrag, just write a state for wordpress that uses include: http://docs.saltstack.com/en/latest/ref/states/include.html#include
19:26 SheetiS wordpress/init.sls could just be a bunch of includes for other states for things like apache/php-fpm/varnish/whatever and then
19:26 SheetiS have it's own stuff for like a wordpress source checkout
19:26 SheetiS forrest++
19:26 sjwoodr salt-cloud question for anyone who knows..... if my AMI has salt-minion baked into it already, the bootstrap of salt-minion via salt-cloud seems to fail.  That is, the /etc/salt/minion file is from the AMI, not populated with the master that initiated salt-cloud to bring up this minion.   Is there a workaround for this?
19:27 forrest someone already wrote it anyways SheetiS: https://github.com/saltstack-formulas/wordpress-formula :P
19:27 gargrag yes SheetiS and what about if in this state, i want to override the defailt mysql port?
19:27 forrest or at least a base
19:27 murrdoc it uses cmd.run
19:27 gargrag it;s not a wordpress formula
19:27 forrest gargrag: mysql port is set here: https://github.com/saltstack-formulas/mysql-formula/blob/master/pillar.example#L11
19:27 forrest just copy the example pillar to the correct pillar location, and put in your own values
19:28 SheetiS pillar=winning.
19:28 murrdoc thats what i said
19:29 forrest sjwoodr: try the -N option, that won't install the minion
19:29 SheetiS sjwoodr: I personally don't bake the salt-minion in for my ec2 minions.
19:29 sjwoodr yeah if we move to salt-cloud on a permanent basis, i'd want to remove salt-minion from our AMI
19:30 sjwoodr if -N skips minion install... that still leaves me with the wrong /etc/salt/minion file on the new instance
19:30 johanek joined #salt
19:30 forrest your ami installs the minion but doesn't drop on a conf?
19:30 sjwoodr it does -- but its the wrong one
19:31 gargrag thank you all for the help forrest SheetiS and TaiSHi
19:31 sjwoodr background: its pointing to our current salt master but i'm using a test salt-master for testing salt-cloud
19:31 forrest gargrag: np
19:31 TaiSHi No prob gargrag
19:31 forrest sjwoodr: Gotcha, yeah that's problematic then
19:31 forrest sjwoodr: I don't think there was ever logic planned to handle someone wanting to drop on the minion config with salt cloud, but not drop on the minion, manfred are you around?
19:32 SheetiS sjwoodr: salt-cloud does allow for custom bootstrap scripts to be configured.  that _might_ get you there.
19:32 sjwoodr im OK with the salt-minion pkg being overwritten by salt-cloud
19:32 sjwoodr im OK with it overwriting (allthethings)
19:32 forrest yeah but I don't think it does logic to upgrade the minion
19:32 forrest I think it only installs
19:33 SheetiS If you need to keep it baked in can look at http://docs.saltstack.com/en/latest/topics/cloud/misc.html#deploy-script-arguments and change the deploy script to one that does specifically what you need (custom script based on your specific needs)
19:34 druonysuse joined #salt
19:34 druonysuse joined #salt
19:34 lothiraldan joined #salt
19:34 manfred sjwoodr:  you can put them in pillar data
19:34 manfred but not for the salt-cloud command
19:35 sjwoodr so does the salt-minion install skip writing files into /etc/salt if they're already there?
19:35 manfred how are you installing it?
19:35 sjwoodr well a) its in the AMI  (an older version, btw) .... and then salt-cloud is bootstrapping it
19:35 sjwoodr and I was hoping for the salt-cloud bootstrap to at least have hte new minion pointing to the master that spawned it
19:36 TaiSHi iggy: let me know when you're around
19:37 manfred it can
19:37 manfred you need to put the master inside of your salt-cloud configu
19:37 manfred in /etc/salt/cloud,
19:37 manfred minion:
19:37 manfred master: x.x.x.x
19:38 manfred sjwoodr:  http://salt-cloud.readthedocs.org/en/latest/topics/config.html
19:38 sjwoodr ohh, i may have misunderstood how that was working
19:38 darien joined #salt
19:39 yomilk joined #salt
19:39 darien I can't find it in the manual, but maybe it's hiding: is there an easy way (besides using returners) to fetch files FROM minions TO the master?
19:39 cotton joined #salt
19:39 sjwoodr manfred/sheetis/forrest - thanks for the tips
19:39 manfred np
19:43 sjwoodr at SaltConf15 they said the salt community is very helpful... they weren't kidding.  :)
19:43 manfred :)
19:43 manfred i was there <3
19:43 sjwoodr same here - just got home at 1:30am this morning :|
19:43 manfred forrest:  you missed out on hanging out with me and ryan
19:44 manfred i need to write up a blog about my talk/experience, and then another one about how I can rewrite my talk using beacons
19:44 * SheetiS is thinking about how to handle partial reuse of configurations among minions for some checks with datadog.  Right now all of the checks that can be used are stored in a pillar, and then I have an ext_pillar with the checks that a particular minion uses tagged to it.  It then builds the config based upon what all it 'cherry-picks' based on the ext pillar.
19:45 murrdoc what was your talk manfred
19:45 forrest manfred: I didn't have a talk that I felt was good enough to present, really wish I could have been there though!
19:45 manfred :)
19:45 forrest Was too expensive to do on my own justifably.
19:45 SheetiS I don't like it because all of the checks are visible to all of the minions in this fashion, but a number of these checks are reused in multiple places, so I only want to change them in one spot
19:45 SheetiS if I have to update.
19:45 manfred murrdoc:  autoscaling rackspace cloud with rackconnectv3 and salt
19:45 murrdoc ah missed it, no rackspace in my work
19:45 murrdoc good reception ?
19:46 forrest manfred was there murrdoc, so it was probably awful
19:46 forrest heckling
19:46 manfred there wasn't a large group, but everyone seemed to like it, and it was my first talk anywhere
19:46 forrest complaints of not enough work on salt-cloud :D
19:46 murrdoc haha
19:46 murrdoc i only heckled iggy
19:47 forrest iggy needs to be heckled, he's such an angry angry man
19:47 murrdoc "iggy: my first talk "
19:47 SheetiS With 2015.2 I might be able to address the information bleed by loading the ext_pillar before the sls pillar and then doing some mojo w/ jinja iside the pillar, but I wonder if that will make my master scream bloody murder every time I run a highstate and rebuild that pillar.
19:47 manfred i don't know who iggy is :/
19:47 murrdoc "me; NO WAY!
19:47 manfred i will be back soonish I have a bunch of stuff to do on salt
19:48 manfred no that I have done one talk, i wanna do more
19:48 forrest I haven't worked on salt in a while either *shrug*
19:48 darien also, is there any reason my salt-master would be using 4 GB of RAM?
19:48 darien that seems excessive
19:49 SheetiS darien: there are many factors that can play into that.  Number of master threads, size of states/pillars, modules utilized, number of minions connected, data in the salt mine, etc.
19:50 darien SheetiS: I've got a fairly simple setup, not doing anything fancy other than states in /srv/salt/
19:50 darien SheetiS: and only about 50ish minions
19:50 peters-tx Anyone happen to know if the Cloudflare keynote (?) from yesterday is online?
19:50 manfred not yet
19:50 manfred it should be at some point
19:51 peters-tx Ok :)
19:51 manfred peters-tx:  watch for it on https://www.youtube.com/user/saltstack
19:51 manfred you can see the saltconf14 on there too
19:51 peters-tx Ok
19:51 manfred hopefully this time, they don't spend 5 months cutting the videos
19:52 manfred because last year apparently they recorded literally everything
19:52 manfred and had to cut out all the crowd shos
19:52 manfred shots*
19:52 murrdoc manfred throwing shade
19:52 SheetiS I'm using 1.5-2gb memory on my master.  75 minions or so.
19:52 SheetiS on salt 2014.7.1
19:52 peters-tx manfred, Thanks :)
19:52 SheetiS I have just over 100 states per minion on average.
19:52 sjwoodr hmm, so i added this to /etc/salt/cloud
19:53 sjwoodr minion:   master: 10.x.x.x  (with correct indents)
19:53 sjwoodr and it still did not update /etc/salt/minion on the new minion it created
19:53 toanju joined #salt
19:53 sjwoodr could it be that it won't touch that file if it already exists?
19:53 sjwoodr and maybe i need a deploy_script to run before bootstrap to delete that file
19:53 manfred that is correct
19:53 forrest manfred: Yes that is why it took so long last time, Rhett got tired of me bugging him about when the videos would be up
19:53 druonysuse joined #salt
19:53 druonysuse joined #salt
19:53 forrest manfred: I offered to help as well
19:54 sjwoodr if i specify deploy_script - does that get run *instead* of the default bootstrapping?
19:54 manfred sjwoodr:  i bet that is right
19:54 manfred yes
19:54 forrest Since I have video editing software, and cuts are pretty quick
19:54 fxhp joined #salt
19:55 manfred yar :)
19:55 manfred lol
19:56 forrest I also hope they are faster this time, keep the momentum going
19:56 darien SheetiS: that seems much better than mine, and I don't think I have nearly 100 states per minion
19:56 manfred sjwoodr:  you could have a script that removes /etc/salt/minion, then runs the bootstrap script
19:56 andrew_v joined #salt
19:56 sjwoodr yeah thats what i was thinking.
19:56 darien SheetiS: maybe I'll run some highstates and see if it grows
19:57 sjwoodr i take it you can only specify one deploy script ?
19:57 sjwoodr script: bootstrap-salt     script_args: -c /tmp/
19:57 SheetiS darien: could look at that.  Without seeing 'all-the-things', it can be hard to troubleshoot something like "my salt master processes are using a lot of memory"
19:57 darien SheetiS: yeah, that's kind of the problem. I was more wondering if that's normal, since 0.17.5 didn't use anywhere near that much
19:59 SheetiS sjwoodr: you can have different cloud profilles with their own bootstrap script
19:59 SheetiS http://docs.saltstack.com/en/latest/topics/cloud/profiles.html /etc/cloud.profiles.d/*conf or whatever
20:00 aparsons joined #salt
20:01 amcorreia joined #salt
20:01 smcquay_ joined #salt
20:02 aparsons joined #salt
20:03 ndrei joined #salt
20:03 aparsons joined #salt
20:05 tedski hrmm... looking to concatenate 2 multi line pillar values into a file.  file.managed only supports one value for contents_pillar and file.append doesn't take a contents_pillar arg
20:05 tedski any ideas how to do this?
20:07 SheetiS tedski, if you have multiline data in the pillar and want it to come across into the file.append properly, I think you could try {{ salt['pillar.get']('some_multiline_text', 'default')|json }} to make sure that no data is lost when passing on
20:08 SheetiS where some_multiline_text was the name of your pillar.
20:09 tedski and the json will get deserialized on write?
20:10 SheetiS I think so.  I will test and let you know momentarily
20:11 sjwoodr sheetis: yeah i have vm profiles setup, but what i wasn't sure - and cant tell from the docs - is if you can specify multiple scripts or if i'd have to have a script that did the cleanup in /etc/salt and then IT invoked salt-bootstrap...
20:11 mafro joined #salt
20:13 SheetiS sjwoodr: just one script, but you could do it like you just said
20:13 SheetiS tedski: the json was deserialized in this test: https://bpaste.net/show/23bb48cc771c
20:13 sjwoodr will salt-bootstrap be in the $PATH already if invoke from my script ?
20:14 tedski SheetiS: cool, thanks!
20:14 SheetiS your script you run might have to obtain it in some fashion.
20:14 SheetiS sjwoodr:
20:16 drawsmcgraw joined #salt
20:19 druonysus joined #salt
20:21 sjwoodr ok
20:23 aparsons joined #salt
20:27 neogenix joined #salt
20:30 johanek joined #salt
20:32 Aikar anyone able to help real quick with python/jinja stuff?
20:32 Aikar salt['mine.get'](pillar['role_content'] + '*', 'network.ip_addrs').items()|join(' ')    trying to get a space limited IP list
20:33 Aikar its close -- but giving me ('vag-content', ['ip'])
20:33 sjwoodr ok i just made a copy of bootstrap-salt and made my customizations in there.  that'll work until we update salt.  but this is just a test anyway... i wont have to do this down the road.
20:33 Aikar how can I filter this to just the IP?
20:33 sjwoodr tnx again sheetis
20:33 logix812 left #salt
20:35 otter768 joined #salt
20:36 SheetiS Aikar: .items gets the key and the value.  try .values() instead
20:36 SheetiS since you don't care about the hostnames (keys)
20:36 SheetiS sjwoodr: no problem.
20:36 darien is there a way in a jinja template to match against CIDR?
20:37 Aikar ok thats close! .values()|join(' ') I now have ['ip'] ['ip']
20:39 SheetiS Aikar: do you want all IPs from each minion, or just the first one?
20:39 SheetiS (I'd solve the 2 differently)
20:39 yomilk joined #salt
20:40 Aikar all
20:40 aparsons joined #salt
20:41 Aikar I need it for RPAF
20:41 Aikar so its RPAFproxy_ips 1.1.1.1 2.2.2.2 3.3.3.3
20:42 darien every time I try to solve a simple problem with salt, I discover some new functionality that would revolutionize how my infrastrucure works, if only I wasn't spending all my time on simple problems
20:42 SheetiS ok, since you cannot really do a list comprehension in jinja, I think a for loop with list.append() is the way to go.  I'll give an example.
20:43 darien oh, network.in_subnet
20:44 SheetiS Aikar: https://bpaste.net/show/d5eaa1707ab7
20:44 perfectsine joined #salt
20:45 SheetiS or you can just {{ ip_list|join(' ') }} where you need the output
20:46 SheetiS darien: yeah network.in_subnet looks good.
20:47 Aikar SheetiS, got it working as set content = salt['mine.get'](pillar['role_content']+'*', 'network.ip_addrs').values()|join(' ', attribute=0)
20:47 Aikar I needed the attribute=0 part
20:47 sjwoodr Configuration file '/etc/salt/minion'  ==> Deleted (by you or by a script) since installation.  ==> Package distributor has shipped an updated version.  ==> Using current old file as you requested.
20:48 sjwoodr hmm.  where do we override this 'Using current old file as you requested'  ?
20:49 giantlock joined #salt
20:49 SheetiS Aikar: attribute=0 would only grab the first IP from each minion from its network.ip_addrs from what I understand.  As long as that is the desired behavior, I think that should work great.
20:50 rvankleeck joined #salt
20:50 rvankleeck joined #salt
20:50 sjwoodr maybe i should have used --purge instead of -r
20:50 sjwoodr when the bootstrap cleaned up the ami's version of the minion
20:51 SheetiS sjwoodr: I haven't looked at the bootstrap script recently, so all I can say is try and report back :)
20:51 sjwoodr yep yep... i'll do that :)
20:52 johanek joined #salt
20:54 rvankleeck when you run a salt command (e.g. salt '*' state.highstate) on the salt master, is it sent directly to the minions or placed in a queue on the master and picked up later by the minion
20:56 SheetiS rvankleeck: the event is immedaitely fired out.  If you want to limit how much is going on at once, you can use --batch <int> to limit the number of concurrent highstates.  So if you wanted to limit to 5 at once you could run salt \* state.highstate --batch 5
20:56 SheetiS then it would run them in batches of 5 waiting for the return (or timeout) of the previous batch
20:56 sjwoodr ok with --purge, salt-minion installs fine and puts the default /etc/salt/minion into place.
20:57 SheetiS sjwoodr: success then? :D
20:58 rvankleeck SheetiS, worded another way: is it a connection TO the minion from the master, or is it a connection FROM the minion to the master?
20:58 Tyrel_ joined #salt
20:59 sjwoodr Only problem now is that my master isnt defined in it :)
20:59 thekeeper joined #salt
21:00 sjwoodr a minion: section in /etc/salt/cloud should have injected itself into /etc/salt/minion on the minion ?
21:00 lz-dylan ...is there a best practice for running highstate periodically?
21:00 SheetiS This is rough but a general idea: the minion makes a persistent connection to the master (zeromq by default).  The master fires an event over zeromq and the minion will see it and perform actions as needed.  The highstate itself runs on the minion with pillar data provided by master.  The minion will then fire back a return once it finishes that the master will then pick up and do whatever it needs.
21:01 SheetiS lz-dylan: there is a scheduler in salt
21:01 SheetiS http://docs.saltstack.com/en/latest/topics/jobs/schedule.html
21:01 lz-dylan Just generally, a cron.daily seems like it would do the trick, but it feels like I should use something Designed To Do That
21:01 lz-dylan awesome, thank you
21:01 rvankleeck SheetiS, okay. So the minion doesn't need any sort of ingress ports open
21:02 darien I think my salt installation is getting to the point where I can just schedule highstate to run every hour and if someone makes changes that get overwritten then F those guys
21:02 lz-dylan SheetiS: that'd let me manage and set up the scheduling in the same repo/codebase as the rest of salt. Excellent. :)
21:02 SheetiS rvankleeck: nope just open up 4505 and 4506 on the master for zeromq
21:02 rvankleeck SheetiS, thanks!
21:03 SheetiS lz-dylan: yep just use the schedule: section of the pillar and then just make sure that pillar is loaded for any minion that you want to run the schedule (via top file or whatever).
21:03 SheetiS rvankleeck: no problem
21:04 Guest89 joined #salt
21:05 amcorreia joined #salt
21:08 sjwoodr @SheetiS - the /etc/salt/minion file is written at the very end --- when should it get updated by salt-cloud ?
21:11 Fiber^ joined #salt
21:11 sjwoodr or do i need to update the minion file myself manually, in my custom deploy script
21:12 v0rtex can anyone offer any suggestions to help me troubleshoot why it is that whenever my mariadb states run on a new VM apt-get ends up in a zombie state and I have to kill my highstate run?
21:12 SheetiS could try and run salt-cloud with -l debug to see further.  Otherwise maybe update in your custom deploy script and trigger a minion restart to make sure the config is loaded.
21:12 zer0def joined #salt
21:13 v0rtex this is being run via salt-call in a masterless setup, btw
21:14 SheetiS v0rtex: does apt-get finish everything and the python subprocess just never returns?
21:14 sjwoodr updating in the vm profile seems to be the way its supposed to work .... but so far i dont see that happening.  i wonder if its because of my --purge stuff.
21:14 v0rtex SheetiS: that's my guess
21:14 sjwoodr perhaps its writing out a file, then doing the purge, then the minion install and we get the default template
21:16 v0rtex SheetiS: when I try to watch what's going on with htop I'm seeing the processes I would expect for it to install but yeah when it finishes I end up with an apt-get zombie and hung salt
21:16 druonysus joined #salt
21:16 jri joined #salt
21:17 SheetiS v0rtex: I've seen this with the startup of some sysvinit scripts in states (nginx in particular), but never found a good solution (replaced sysvinit script with upstart to work around) if that is what is going on.  what salt version btw?
21:18 v0rtex I just checked what I'm using and it's 2014.7.0
21:18 v0rtex I should probably upgrade that
21:18 debian112 I am trying to target servers by grains, to specify which accounts needs to be added to a server.
21:19 debian112 Maybe I am heading down the wrong path
21:19 debian112 http://paste.debian.net/160002/
21:19 SheetiS could try it just for science v0rtex, otherwise opening a github issue might be next step.
21:19 debian112 any idea?
21:19 sjwoodr i see that /tmp/.saltcloud-70a4c89a-8bfc-41c3-8d72-33ae55135340/minion is the correct file
21:20 v0rtex okay, will do. Thanks SheetiS
21:20 sjwoodr then i see /etc/salt/minion is gone
21:20 sjwoodr after my --purge
21:20 SheetiS debian112: I see you are setting username at the top of your paste
21:20 SheetiS then destroying it in your for loop
21:20 SheetiS by reusing username
21:21 SheetiS are you trying to just extract the usernames at the top from the pillar inside of the for loop?
21:22 debian112 SheetiS: yes
21:22 MTecknology Just want y'all to know I just had an ugly issue happening. I'm fixing it... (waiting for salt to finish)
21:22 debian112 I will post the pillar
21:24 SheetiS ok likely the easiest fix is to change lines 2 and 4 as set usernames = instead of username and then change your for loop to say "for username, details in pillar.get('user_accounts', {}).items() if username in usernames"
21:24 jacksontj mimiandi: thanks :)
21:24 SheetiS http://paste.debian.net/160003/ this cahnge
21:25 SheetiS debian112: I haven't tested that, but that is how I would start looking at it without checking your pillar or the validity of your if statements above to match.
21:26 debian112 pillar: http://paste.debian.net/160005/
21:26 MTecknology done fixing 180 servers.
21:26 aparsons joined #salt
21:28 Pixionus joined #salt
21:28 SheetiS debian112: ok try the changes I made above in 160003.  I think it should work based upon the pillar you provided.
21:29 debian112 SheetiS: ok
21:35 druonysus joined #salt
21:37 Gareth always...never...forget to check your salt master configuration.
21:39 debian112 SheetiS: I tried it, and it didn't execute
21:39 debian112 http://paste.debian.net/160008/
21:40 debian112 SheetiS: I see it here on the minion: /var/cache/salt/minion/files/base/users/people.sls
21:40 debian112 so atleast it's not crashing or erroring out!
21:40 yomilk joined #salt
21:43 SheetiS debian112: I just noticed where your endif.  one more change inc
21:44 Aikar SheetiS, the mine only looks for a single adapter, so its always 1 entry
21:44 Aikar its doing network.ip_addrs - eth0
21:44 SheetiS debian112: http://paste.debian.net/160009/
21:45 Aikar before i go to the php channel and get yelled at for off topic, anyone ran into apache mod_rpaf + php5-fpm and getting the real IP?
21:45 Aikar apache sees real ip in access log, but php doesn't get REMOTE_ADDR rebound, however it does see the XFF header
21:45 Aikar i'm thinking need to rebind the REMOTE_ADDR fcgi param?
21:46 markm joined #salt
21:46 SheetiS Aikar php-fpm doesn't get sent REMOTE_ADDR by default, so I think yes.
21:46 debian112 SheetiS: Works!
21:46 Aikar well it gets it for non proxied request just fine
21:46 debian112 Thank you!
21:47 Aikar its like it sends it to fpm before rpaf replaces it
21:48 chamunks joined #salt
21:48 SheetiS Aikar: that could be.
21:48 igorwidl joined #salt
21:49 SheetiS I haven't messed with apache + php_fpm in about a year now, so I don't remember how I takcled it, but I thought it had to do with setting the env so it would past fpm via fastcgi
21:49 drawsmcgraw left #salt
21:50 SheetiS ooh my engrish was strong in the excrement there.
21:50 iggy murrdoc: my talk was the most fun for you... sitting in the back making me laugh constantly
21:50 iggy TaiSHi: why?
21:50 murrdoc yes
21:50 murrdoc yes it was
21:50 SheetiS murrdoc is like a level 12 troll imo :D
21:51 murrdoc waaaa
21:51 SheetiS or maybe you just _always_ like to give iggy a hard time :D
21:52 TaiSHi iggy: whoopsies, my VPN failed
21:52 TaiSHi Mind if I PM you?
21:54 iggy no, as long as it isn't something tha would benefit everybody else
21:54 TaiSHi Oh, it was about world peace :(
21:54 TaiSHi :P
21:56 murrdoc need to start the chicago salt hangout
21:57 SheetiS <= salty from Kansas City :o
21:57 murrdoc where is that
21:57 * murrdoc looks at map
21:58 SheetiS Prolly 8 hours from Chicago by car =\
21:58 murrdoc not trolling i really dont know where kansas is
21:58 murrdoc i see it
21:58 iggy it's missouri!
21:58 murrdoc we can skype your ass in
21:58 murrdoc yours and iggys
21:58 peters-tx joined #salt
21:59 murrdoc drink beer share salt thoughts
21:59 murrdoc ahlee will be there to bring a decorum
21:59 murrdoc modicum of decorum
21:59 murrdoc whatever that is
21:59 iggy I'm interested to know how the mysterious "at lunch" -formulas meeting ended up
22:00 darien is there a way to see the list of SLS files a minion is processing (and not just the end result rules)?
22:00 iggy "we're going to meet at a lunch with 500 people packed in a room... hope you can find it suckers"
22:00 iggy darien: cp.list_master shows all files on the master (not exactly what you asked, but helpful at times)
22:01 darien hm
22:01 hg joined #salt
22:01 darien I'm merging two separate 'salt' environments (prod and devo) into one top file, but for some reason my devo server is getting the prod LDAP server IP, which means that my extend: isn't running despite my sls being specified
22:02 murrdoc uh hmm uh hmm
22:02 nitti joined #salt
22:03 Gareth darien: I'm betting one of the functions here will accomplish what you want: http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.state.html
22:03 pkirby joined #salt
22:03 Gareth maybe show_highstate or show_lowstate
22:04 Gareth maybe even show_sls
22:04 SheetiS show_top maybe to get the list of things called and then show_sls through them if you don't see what you want with show_highstate.
22:04 debian112 SheetiS: here's another question. What do you think I need to get every line read. Right now if any of the if statement are met then it uses that. For instance if I had a user I wanted to give access to allbox.greentoads.net. The if statement section will never be used because of the one above it that is checking with domain verses nodename.  I hope this makes sense!
22:04 darien show_top actually confirmed that the correct state file is being loaded
22:07 pkirby Hey all - is there a way to tell salt to apply only states from a certain environment? I know I can use state.sls to apply specific states from an environment, but I want to apply all states in an environment without needing to specify them
22:07 SheetiS debian112: your if statements at the top are determining what usernames are being applied.  You'd just need to set the conditionals to apply what your needs are.
22:08 triplecorn joined #salt
22:08 SheetiS pkirby: I think you'd need to specify them in your topfile and use a state.highstate, or have a state that you call with state.sls that includes specifically all the other states =\
22:08 debian112 SheetiS: ok thanks. I will play around with what I have right now.
22:09 pkirby Hm a nested state, that's a good idea, thanks. I'll try it
22:11 jeddi joined #salt
22:12 douglassssssssss joined #salt
22:12 aquassaut joined #salt
22:12 SheetiS debian112: you can also always start with an empty usernames = [] list and then use usernames.extend(['username', 'to', 'add']) to add more usernames to your list
22:12 SheetiS based on whatever conditionals you want to specify
22:13 aparsons joined #salt
22:13 triplecorn joined #salt
22:13 pkirby Nested states works perfectly, thanks SheetiS
22:14 SheetiS pkirby: no problem.
22:15 murrdoc we should totes get a tip jar
22:15 murrdoc hvae it goto a charity
22:16 murrdoc or an issue jar, where tiny issues need to be solved as a thank u or something
22:16 murrdoc or not
22:17 SheetiS hehe
22:21 shaggy_surfer1 joined #salt
22:22 nitti joined #salt
22:22 murrdoc http://www.commitstrip.com/wp-content/uploads/2015/03/Strip-Confession-650-finalenglish.jpg
22:23 MTecknology best part about getting back from the conference... having my new laptop
22:30 elfixit joined #salt
22:32 jaloren joined #salt
22:32 jaloren hi all i have a question about custom salt runners http://docs.saltstack.com/en/latest/ref/clients/#runnerclient
22:32 jaloren when i execute the runner's cmd function in a python script it generates additional output
22:32 jaloren even if i am assigning the result of the cmd to a variable
22:33 jaloren any idea what's going on?
22:34 murrdoc the peoples can help, if they get the gists
22:34 jaloren murrdoc: is that to me?
22:34 murrdoc heh yeah
22:34 rm_jorge joined #salt
22:34 murrdoc its late on friday sorry
22:35 murrdoc if u can paste i the runner u are using and the output that might help
22:35 XenophonF left #salt
22:35 jaloren https://gist.github.com/jaloren/60b75cd50a03d74ccf57
22:35 jaloren i am calling the runner in a custom script
22:35 darien what's the best way to use a salt state to push a salt grain to a minion from the master? just push a file into /etc/salt/minion.d/ and restart salt-minion?
22:36 jaloren murrdoc: here's the custom runner https://gist.github.com/jaloren/12912da5cc70cc7d186e
22:36 otter768 joined #salt
22:36 rm_jorge joined #salt
22:37 darien I guess I can put them in /etc/salt/grains
22:37 murrdoc no clue
22:37 jaloren murrdoc: the result of manage.up is being stored in a variable nodelist....sooo why would my script emit any data?
22:37 druonysus joined #salt
22:37 iggy darien: grains.present?
22:38 darien iggy: hmm, that seems pretty great
22:39 iggy protip: gist allows multiple files per paste
22:39 micah_chatt_ joined #salt
22:39 iggy so I don't have to open 400MB worth of chrome tabs
22:39 iggy <-- lazy ass
22:40 stwb joined #salt
22:41 yomilk joined #salt
22:41 jaloren murrdoc: its  salt.output.display_output(minion, '', __opts__)
22:42 jaloren had to use pdb to track it down
22:42 peters-tx I did a wget of "drac.py" from the 2015* saltstack github modules, into /srv/salt/_modules, then sync'ed out to a test minion... how do I call that module manually now?  I tried some obvious ways but it doesn't seem to find it... "'drac.getsysinfo' is not available."
22:42 murrdoc jaloren:  nice find
22:42 jaloren murrdoc: i wonder if this is one of those undocumented options things
22:42 murrdoc sounds like it
22:43 jaloren /usr/lib/python2.6/site-packages/salt/output/__init__.py(68)display_output() -> print_cli(display_data)
22:43 iggy peters-tx: do you have the dependencies for that module installed on the minion?
22:43 jaloren sooo...why does a custom salt runner always call the display output
22:43 jaloren interesting
22:43 jaloren i need to investigate this some more
22:44 peters-tx iggy, Probably not; although it is a fully functional latest 2014.7 minion
22:44 iggy peters-tx: i.e. do you have racadm installed on the minion?
22:45 peters-tx iggy, You think it may be failing at that point?  I'll check the logs, but it should have a racadm, full OMSA installed OK
22:45 andreyst joined #salt
22:45 bhosmer_ joined #salt
22:45 iggy well, that's the only thing the module itself looks for
22:45 iggy but!
22:45 iggy it's not going to work
22:45 peters-tx iggy, Looking at the minion logs I just don't think I'm getting to the module yet still
22:46 kermit joined #salt
22:47 peters-tx iggy, salt-call drac.getsysinfo on the minion returns "Function drac.getsysinfo is not available" ... hmm, looking into it
22:47 peters-tx iggy, and this minion has /var/cache/salt/minion/files/base/_modules/drac.py
22:47 iggy peters-tx: how did you install salt?
22:47 peters-tx iggy, RPM via EPEL
22:48 peters-tx iggy, salt-2014.7.1-1.el5
22:48 iggy does /usr/lib/python2.7/dist-packages/salt/ext/six.py exist?
22:48 peters-tx iggy, Hmm, let me check
22:49 sjwoodr Sheetis: i just took salt-minion out of my ami.  it was less work in the end.  :D   salt-cloud working smoothly now!
22:49 peters-tx I saw mention way up above of six.py, I'm guessing that's going to be a problem
22:50 peters-tx iggy, Only have /usr/lib/python2.4 and /usr/lib/python2.6
22:50 iggy peters-tx: yeah, so you're going to have to do some looking around for it
22:50 iggy I was going off a ubuntu minion
22:50 peters-tx iggy, /usr/lib/python2.6/site-packages/six.py
22:51 iggy no
22:51 iggy salt/ext/six.py
22:51 peters-tx Ahh, ok,  /usr/lib/python2.6/site-packages/salt/ext/six.py
22:52 iggy okay, there goes that idea
22:52 peters-tx iggy, site as opposed to dist
22:52 iggy seems like it should be working then
22:53 iggy maybe run "salt-call -l debug drac.system_info" and see if you see anything in the output
22:53 peters-tx iggy, Yah, I just tried that :)
22:55 peters-tx iggy, http://fpaste.org/194529/68251714/
22:55 ahammond anyone have a link for the SaltConf15 videos? There's a couple of sessions I couldn't catch that I'd love to see... :)
22:55 iggy not out yet
22:55 murrdoc u have to watch the brian jackson one
22:55 murrdoc good stuff
22:55 iggy ahammond: youtube.com/user/SaltStack or something
22:56 ahammond yeah, there was a lot of good stuff there.
22:56 iggy haha, mine didn't get recorded
22:56 murrdoc http://arstechnica.com/security/2015/03/in-major-goof-uber-stored-sensitive-database-key-on-public-github-page/ more about putting shit on github
22:57 peters-tx murrdoc, Thanks, checking that now O_o
22:57 peters-tx I used Uber (not Lyft) to-from SLC-saltconf
22:57 peters-tx Oh, Uber Drivers
22:57 jaloren what was the big themm at saltconf15?
22:58 ahammond peters-tx dude the light rail was $2.50 and went door to door
22:58 darien aw, no ganeti support in salt-cloud
22:58 ahammond jaloren beacons
22:58 ahammond jaloren they're a nice clean modular way to insert events into the saltstack event bus
22:58 iggy murrdoc: someone should have mentioned that to all the presenters this week
22:58 peters-tx ahammond, Well I was worried about timing a bit and also the hours that ran... and also lugging around this frickin ancient work laptop and my suitcase :/  Well I don't pay the bills either, so there's that
22:59 ahammond peters-tx yeah, I was there with my boss and a couple of co-workers. probably woulda been cheaper to uber it, but... whatever.
22:59 jaloren ahammond: is this a new addition? I remember that they had exposed way to put custom events on the bus which would then cause the reactor system to kick in and do something. But....it was hard to use and understand
22:59 jaloren so beacons is a userfriendly wa yto do that?
22:59 iggy saw so many twilio/etc keys
23:00 jaloren ahammond: ....anyone noticed that the documentation on beacons is terrible?
23:00 ahammond jaloren they're very, very, very new.
23:01 ahammond but yeah, nice clean modular architecture. gonna be a very sweet tool
23:01 ahammond needs some runtime with users to get a little cleaner though.
23:01 peters-tx iggy, So I thought I could just sync my modules and then call them, but I guess maybe there's more to it
23:02 Singularo joined #salt
23:02 peters-tx iggy, Well, that and I'm trying a 2015 module on 2014 :)
23:02 peters-tx I was hopeful
23:02 murrdoc step away from the salt
23:03 igorwidl wow, this channel is on fire
23:03 peters-tx iggy, Hmmm I think I may have figured it out, checking
23:04 Ryan_Lane joined #salt
23:04 peters-tx iggy, Yah.  Damn.  No such thing as getsysinfo.  It is system_info
23:04 jaloren ahammond: where's my built in beacon for docker containers? :)
23:04 peters-tx ty, drive through
23:04 shaggy_surfer1 how do you pass variables and templatize configuration files based on env or hostname for example?
23:04 shaggy_surfer1 I remember in puppet it was called hiera…. what is the equiv for salt?
23:05 Corey shaggy_surfer1: Pillar.
23:05 Corey shaggy_surfer1: Although in Puppet you're talking about facter for hostnames; the corresponding subsystem here is grains.
23:05 jaloren oh man anytime i have to deal with puppet manfiests
23:05 jaloren i thank the gods for salt
23:05 murrdoc plus the option to run facter too if needed
23:06 murrdoc if u are transitioning
23:06 jaloren shaggy_surfer1: bringing back terrible memories :)
23:06 shaggy_surfer1 yes, I have used grains.  I was just looking for the best way to example run multiple env's in amazon ec2, like prod1 prod2 prod3, etc.
23:06 murrdoc envs from top.sls ?
23:07 shaggy_surfer1 I want to be able to have the for example web config files have different server names based on the env as hostnames need to be unique and dns needs to be unique for outside access to each env
23:07 shaggy_surfer1 I use currently diff env's in the top.sls for dev,qa,staging,prod
23:08 shaggy_surfer1 the issue I am running into now is I am trying to fully automate the full turn up of different mirrored production env's with different hostnames.  There are lot of config files unique that point to diff hostnames and ports for diff services.
23:08 shaggy_surfer1 looking for the best way to do this yet only have to manage one subset of files with var's
23:09 shaggy_surfer1 @jaloren exactly why I moved away from puppet
23:09 iggy peters-tx: that will teach you not to copy/paste
23:11 peters-tx iggy, Actually I guess a bug in the docs?  I just didn't read carefully and anticipate that!  http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.drac.html
23:11 shaggy_surfer1 does anyone have an example or know of someone who has done what I am trying to achieve?  I want to fully automate amazon ec2 mult availablility zone and regional deploys with multiple prod env's so that I can just flip dns to move from one to the other when I want.  Kind of a rotational env thing.  This way I have the flexibility to roll back by changing dns.
23:11 peters-tx iggy, The very bottom-most entry, "CLI Example:" .. durp
23:11 murrdoc shaggy_surfer1:  master or masterless etup
23:11 shaggy_surfer1 my setup needs more parameterization i guess
23:11 shaggy_surfer1 master
23:12 shaggy_surfer1 I have a master in place right now w/ multiple env's in top.sls…. but I don't want to keep adding dup code for each env.  Looking to parmeterize it now
23:12 murrdoc interesting, i mean i have seen people use {{env}} in the top.sls to remove the need to add dupe code
23:13 murrdoc or is it {{ saltenv }}
23:13 murrdoc but to answer your question, might as well use grains
23:13 murrdoc and u can set a grain as a file on the minion
23:13 murrdoc so its easy
23:13 peters-tx iggy, Ok, now I can do salt-call on the box, but not from the Master... What am I doing wrong?  Still says "is not available" hmmmm, I'm looking
23:14 peters-tx iggy, salt-call drac.network_info works fine on Minion, but salt \* drac.network_info results in "...is not available."
23:15 peters-tx Seems like a really basic problem
23:15 mafro joined #salt
23:16 yomilk joined #salt
23:18 shaggy_surfer1 I don't understand the "set a grain as a file on the minion" part.
23:18 alexbst_ joined #salt
23:18 shaggy_surfer1 can you explain.
23:18 murrdoc u can set grains in the /etc/salt/grains file on the minion
23:18 murrdoc and then u can use those grains as decision points in your code
23:18 murrdoc 'can'
23:21 shaggy_surfer1 how to the grains get on the minion in /etc/salt/grains in the first place?  By pushing them with a state file?
23:23 eliasp I have quite a lot GitFS repositories in my fileserver config… when developing/testing something, I usually do so in a dedicated branch/environment like dev-feature-XYZ … but now salt fails to find SLS in the repos, which aren't affected by this…
23:23 eliasp as a workaround I just create the same branches in those repos too but just keep them there in sync to master/base
23:23 eliasp is there a better way to solve this, e.g. tell the fileserver/salt to fallback onto a certain branch/environment when a SLS isn't found?
23:24 peters-tx shaggy_surfer1, man 7 salt
23:24 peters-tx shaggy_surfer1, search for "Grains in /etc/salt/grains"
23:25 peters-tx shaggy_surfer1, also search for "Core grains can be overriden by custom grains."
23:26 murrdoc well you can do that with a state file or even a quick echo >> or a sed
23:26 peters-tx shaggy_surfer1, "4. Custom grain modules in _grains directory, synced to minions."
23:26 peters-tx sorry, I don't know :)  /me backs away slowly
23:36 baweaver joined #salt
23:44 SubOracle joined #salt
23:47 ipmb joined #salt
23:55 aurynn left #salt
23:57 baweaver joined #salt

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