Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-03-09

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

All times shown according to UTC.

Time Nick Message
00:05 bdf joined #salt
00:12 dooshtuRabbit joined #salt
00:15 helderco joined #salt
00:20 saltpepper joined #salt
00:23 saltpepper help, my minions are always disconnected from the master, test.ping randomly works or not. I have upgraded to ZMQ4 (4.1) on minions and master and still see the same behaviour. unfortunately salt is absolutely unusable with these disconnects, its baffling to me how a remote execution system can have such a fundamental flaw which completely defeats its purpose? :(
00:23 saltpepper im talking about these issues here, I have tried all fixes described and none are working, minions are still randomly disconnecting from the master, a restart of salt-minion will connect them again, but only for a while (hours, sometimes even less). https://github.com/saltstack/salt/issues/15415
00:23 saltpepper https://github.com/saltstack/salt/issues/6231
00:24 saltpepper these issues are closed in github, but im still experiencing same thing with the latest versions: Salt: 2014.7.1 PyZMQ: 14.5.0 ZMQ: 4.1.0
00:25 johtso joined #salt
00:28 bones050_ joined #salt
00:31 saltpepper so what im really hoping for, is there an alternative to ZMQ for salt? since ZMQ relies on TCP keepalive, which both AWS and AZURE clouds dont respect, making ZMQ the wrong choice (imho completely broken for any scenario involving connections over WAN), is there no alternative?
00:31 saltpepper i have tried salt-ssh, but only very basic formulas work there, more complex ones have severe limitations and issues :(
00:41 felskrone joined #salt
00:55 lnr joined #salt
00:55 cotton joined #salt
01:02 iggy saltpepper: in newer versions, there's raet
01:03 iggy i suggest you read up on it and decide if it's for you
01:03 aqua^mac joined #salt
01:04 saltpepper thanks iggy, i will look into it!
01:06 saltpepper so from the salt documentation it says this about RAET: The concern posed by some over RAET reliability is based on the fact that RAET uses UDP instead of TCP and UDP does not have built in reliability. RAET itself implements the needed reliability layers that are not natively present in UDP, this allows RAET to dynamically optimize packet delivery in a way that keeps it both reliable and asynchronous.
01:06 saltpepper so it sounds exactly like what I need :) thanks, i'll go and try it now
01:09 dennisharper_ii joined #salt
01:13 Hell_FireW joined #salt
01:21 mafro joined #salt
01:25 monk-ee joined #salt
01:29 yomilk joined #salt
01:42 shnguyen joined #salt
01:44 khris joined #salt
01:46 thayne joined #salt
01:47 TyrfingMjolnir joined #salt
01:48 viq joined #salt
01:48 viq joined #salt
01:49 _JZ_ joined #salt
02:12 hal58th joined #salt
02:15 Georgyo joined #salt
02:16 CheKoLyN joined #salt
02:18 donmichelangelo joined #salt
02:18 fllr joined #salt
02:28 dwfreed joined #salt
02:29 aphorise joined #salt
02:31 catpigger joined #salt
02:38 I3olle joined #salt
02:49 enarciso joined #salt
02:51 aphoriser joined #salt
02:54 * MTecknology hands basepi a sticker.
02:57 beau_zeaux joined #salt
02:59 beau_zeaux left #salt
03:08 nich0s joined #salt
03:09 KFDM joined #salt
03:10 KFDM left #salt
03:12 Ryan_Lane joined #salt
03:13 gnord joined #salt
03:15 kevinforbes joined #salt
03:16 kevinforbes Can I check the length of an array in a Jinja template file?
03:16 kevinforbes Trying to do this:
03:16 kevinforbes {% if len(pillar['roles']) == 1 %}
03:16 kevinforbes actually pillar[‘roles’] is a dictionary
03:17 peters-tx2 kevinforbes, http://stackoverflow.com/questions/1465249/jinja2-get-lengths-of-list perhaps
03:18 chitown joined #salt
03:18 iggy usually roles is a list
03:18 iggy easier to work with
03:19 kevinforbes lemme try it out
03:19 iggy i.e. {% if 'web' in pillar['roles'] %}
03:19 bash1245_ joined #salt
03:23 kevinforbes @peters-tx2 that worked thx
03:23 tmh1999 joined #salt
03:24 kevinforbes @iggy, I’m trying to do a thing where different pillars can registar a role, and the output is a combined list of all defined roles. My solution was to do it like this: http://pastebin.com/4McY0wKU
03:25 kevinforbes Would love to know if you can think of a better way
03:25 iggy that's not a lot to go on
03:26 favadi joined #salt
03:26 ajw0100 joined #salt
03:26 TyrfingMjolnir joined #salt
03:26 kevinforbes Ok, so if multiple pillars define the same variable, the output is non-deterministically one of the values
03:27 kevinforbes e.g. [pillar1.sls] foo: bar  [pillar2.sls] foo: baz  —> the pillar variable foo will be bar or baz
03:27 kevinforbes What if I wanted a list of both?
03:27 MTecknology "wanted a list of both" ... both variables?
03:27 kevinforbes yeah, so foo: [bar, baz]
03:27 MTecknology can't be done
03:27 shaggy_surfer joined #salt
03:28 kevinforbes I thought so
03:28 hal58th joined #salt
03:28 kevinforbes But it can be done if you use dictionaries
03:28 MTecknology You only get one. The master builds the dictionary that gets sent to the minion. It's a dictionary.
03:28 kevinforbes [pillar1.sls] foo: bar: {} [pillar2.sls] foo: baz: {}
03:29 kevinforbes rendered output: foo: {bar: {}, baz: {}}
03:30 MTecknology Hm?.. did you try that and it actually worked that way?
03:30 kevinforbes Yep
03:30 MTecknology Weird. Apparently I don't understand the pillar render system enough.
03:31 kevinforbes The pillar docs describe as much. It’s a great way to build up a dictionary across pillars
03:31 iggy that's why it should be roles:\n - web and roles:\n - db
03:31 kevinforbes http://docs.saltstack.com/en/latest/topics/pillar/#pillar-namespace-merges
03:31 iggy that should merge
03:32 kevinforbes oh cool, that looks better
03:32 kevinforbes I’ll try
03:32 CeBe2 joined #salt
03:32 CeBe3 joined #salt
03:32 MTecknology Reading through this, I now know some of the questions I got wrong on the cert exam.
03:32 iggy lol
03:33 iggy you passed, just be happy about that
03:33 iggy I noticed I still can't lookup/login to the cert site... you MTecknology?
03:34 MTecknology I can log in, but can't see the result yet. I imagine it's taking them some time to get it filled in. I'm waiting for the email that they're supposed to eventually send us.
03:36 MTecknology Was it 49 that passed?
03:40 TyrfingMjolnir joined #salt
03:43 subsignal joined #salt
03:46 jchen joined #salt
03:46 jchen left #salt
03:49 iggy never heard
03:49 iggy but that would be what 10%?
03:49 iggy well, 10% of total attendees, don't know how many were just Tuesday
03:51 dimeshake MTecknology: the ssc.saltstack.com site is kinda messed up.
03:53 hal58th iggy, mtecknology: it was 49 new people that passed. I still can't login either.
03:57 MTecknology I forgot what the total attendance was. I think it was just over 400, though. I think it was something like just under 200 that attended pre-conference. I don't remember for sure, though.
03:58 hal58th I believe that's what Colton said Mtecknology
04:00 iggy I thought it was closer to 500 total
04:01 MTecknology that was the first number that we were told.
04:02 MTecknology I actually hope they can make it the same place next year. I love that area.
04:03 hal58th I wouldnt mind san francisco because it's even closer. but it was nice taking the tram straight to my hotel
04:06 MTecknology That would put it farther away from me
04:06 MTecknology and clearly that isn't gonna be cool by anyone's standards of acceptable
04:10 iggy lol
04:10 iggy it'll be bay area
04:11 iggy think about the users...
04:11 dimeshake yeah, screw east coasters like me ;)
04:12 MTecknology iggy: but what about the most important user?...
04:12 iggy I'd just prefer it weren't somewhere freezing cold... I've been sick since the 3rd day I was there
04:12 MTecknology It should be in SF
04:12 MTecknology (sufu, the real SF)
04:14 hal58th With all three tech companies there.
04:15 MTecknology three?
04:15 MTecknology What three are there?
04:15 hal58th Daktronics, 3M and I'm sure I can find one mroe
04:16 hal58th 3M doesn't really count though
04:16 hal58th Wasn't gateway there?
04:16 MTecknology daktronics and 3M aren't in sufu
04:16 hal58th eh, close enough
04:16 MTecknology gateway was here, ya, but closed
04:16 Ahlee It took us weeks last year to be able to log in
04:17 MTecknology There's also no venue here that could support it
04:17 hal58th1 joined #salt
04:17 Ahlee those fucking cow boxes
04:18 MTecknology It really doesn't matter to me where it's at if it's in the US. It'll pretty much be the same amount of time flying.
04:18 iggy I'm always torn between "let's make it easy for as many people as we can to get here" and "let's screw over everyone equally"
04:19 iggy in the latter case, make it in oklahoma city or something
04:21 mafro joined #salt
04:22 tkharju joined #salt
04:26 hal58th joined #salt
04:36 Brew joined #salt
04:38 basepi MTecknology: 49/110 passed.
04:38 basepi hal58th ^
04:38 * basepi goes to bed
04:39 MTecknology good night!
04:39 hal58th oh wow. I didn't realize so many people took that test basepi. When will our certs be up on the website?
04:39 MTecknology thanks for the info :D
04:39 basepi hal58th: unknown, I'm not involved with that at all. I'll check tomorrow. Ping me sometime tomorrow if I don't remember to get back to you. =)
04:39 hal58th Thanks basepi. Have a good night sleep
04:39 basepi Also, I don't know if that 110 is total tests taken or if it accounts for retakes
04:39 * basepi shrugs
04:40 * basepi head off to battle against stupid daylight savings changes
04:40 basepi heads*
04:43 iggy I think I have pillar:// paths working
04:43 iggy needs lots of testing
04:45 aqua^mac joined #salt
04:46 TyrfingMjolnir joined #salt
04:50 fintler joined #salt
04:50 fintler joined #salt
04:52 MTecknology I try to avoid doing "fancy" things with salt, but I am attempting to make salt do 100% of server configuration and maintenance within my network.
04:52 MTecknology someday... maybe...
05:02 TyrfingMjolnir joined #salt
05:04 enarciso joined #salt
05:07 nethershaw joined #salt
05:07 garthk MTecknology: I’m not sure I can get there: both the ops and devs on either side of me are reluctant to learn enough salt to make it do everything.
05:07 chandra joined #salt
05:07 garthk MTecknology: they’re just too hooked on connecting right in, doing something, and not worrying about whether it’s durable, consistent, whatever meh yolo
05:09 linjan joined #salt
05:11 MTecknology garthk: in that case, they're probably just as well off not using salt
05:15 TyrfingMjolnir joined #salt
05:15 garthk MTecknology: well, yeah, but they want the operational benefits, so tough luck :)
05:15 MTecknology thankfully, I'm "the guy" so what I say goes.
05:15 garthk MTecknology: I met ‘em half way. Puppet scared them to death. Salt’s much less scary. It’s just configuration, right?
05:16 MTecknology puppet is horrible in soooo many ways. It's like the best of the worst
05:17 MTecknology I have other things that need to be priorities and because I'm "the guy" I have a lot that I need to worry about. However, when I say things will happen in a specific way, that's the way it'll happen. I like that.
05:17 garthk Hmm. Anyone using ipsec to glue together clusters that don’t make it easy to secure their interconnect ports, e.g. Elasticsearch?
05:20 fintler joined #salt
05:20 fintler joined #salt
05:24 garthk Aah, I’m about to wish for the one Puppet feature I want in Salt: parameterised states. Might be able to get away with heavy use of pillar.
05:25 garthk {% set is_selinux_enabled = salt.cmd.run('sestatus') == '1' %} # GENIUS
05:26 garthk There’s a lot of intermediate Salt best practice stuff hidden in this Formulas guide. http://docs.saltstack.com/en/latest/topics/development/conventions/formulas.html
05:26 thayne joined #salt
05:28 iggy {% set is_selinux_enabled = salt['selinux.getenforce']() in ['enforcing', 'permissive'] %}
05:29 garthk iggy: heh; I’m still delighted I can wait on a salt command
05:31 nene joined #salt
05:31 garthk ok, this is sinking in: use racoon to set up an ipsec mesh, which lets my cluster nodes talk to each other… then set up ipip tunnel interfaces to which I can bind the cluster interconnect ports instead of having everything exposed to 0.0.0.0
05:32 garthk or, use the firewall to block access to the ports by anyone else
05:39 otter768 joined #salt
05:45 MTecknology iggy: If you present next year, plan for 40min and make it about things like that. I would so attend that.
05:45 MTecknology not that I give a crap about selinux, but concepts like that are neat shit
05:48 MTecknology g'night!
05:54 ramteid joined #salt
06:06 Exploit joined #salt
06:11 Exploit whats salt?
06:12 hal58th NaCL
06:16 kevinforbes joined #salt
06:21 mafro joined #salt
06:21 asaladin joined #salt
06:32 malinoff joined #salt
06:41 Kevalin joined #salt
06:44 monkey66 joined #salt
06:46 zz_whytewolf joined #salt
06:54 krelo joined #salt
06:57 zer0def joined #salt
07:03 bhosmer joined #salt
07:04 colttt joined #salt
07:05 enarciso joined #salt
07:14 monkey66 joined #salt
07:19 jhauser joined #salt
07:20 schristensen joined #salt
07:20 KFDM joined #salt
07:20 KFDM left #salt
07:22 AndreasLutro joined #salt
07:26 evle1 joined #salt
07:26 MaliutaLap joined #salt
07:26 MaliutaLap left #salt
07:31 felskrone joined #salt
07:32 johanek joined #salt
07:32 toanju joined #salt
07:36 felskrone joined #salt
07:38 KermitTheFragger joined #salt
07:39 bash124512 joined #salt
07:40 otter768 joined #salt
07:41 illern joined #salt
07:41 tomspur joined #salt
07:43 mattiasr joined #salt
07:44 flyboy joined #salt
07:48 mike25de joined #salt
07:48 monkey66 joined #salt
07:52 markm joined #salt
07:54 linjan joined #salt
07:54 mfournier joined #salt
08:01 trikke joined #salt
08:01 monkey66 joined #salt
08:02 kawa2014 joined #salt
08:03 mikkn joined #salt
08:04 enarciso joined #salt
08:04 jri joined #salt
08:04 yuhl_work_ joined #salt
08:05 Auroch joined #salt
08:07 mekstrem left #salt
08:07 mekstrem joined #salt
08:07 I3olle joined #salt
08:07 mekstrem left #salt
08:18 chiui joined #salt
08:19 eseyman joined #salt
08:19 JlRd joined #salt
08:20 monkey66 joined #salt
08:20 intellix joined #salt
08:25 Auroch joined #salt
08:26 DenkBrettl joined #salt
08:28 bdols joined #salt
08:34 jtang joined #salt
08:34 TyrfingMjolnir joined #salt
08:40 thehaven joined #salt
08:46 wincyj joined #salt
08:48 lb1a joined #salt
08:52 DenkBrettl joined #salt
08:53 robinsmidsrod joined #salt
08:56 denys joined #salt
08:58 _monokrome joined #salt
08:58 jri joined #salt
09:02 thehaven joined #salt
09:03 dcmorton_ joined #salt
09:03 intellix joined #salt
09:04 djinni`_ joined #salt
09:04 karimb joined #salt
09:05 losh joined #salt
09:06 kossy joined #salt
09:06 InAnimaTe joined #salt
09:06 Deevolution joined #salt
09:06 pipeep joined #salt
09:07 kossy joined #salt
09:12 Jouke is it allready possible for a pillar to reference itself?
09:15 monkey66 joined #salt
09:16 Xevian joined #salt
09:21 monkey66 left #salt
09:22 mafro joined #salt
09:23 N-Mi_ joined #salt
09:27 karimb joined #salt
09:28 lynxman joined #salt
09:33 Roee joined #salt
09:33 Roee Hi All
09:34 Roee I'm running salt version "2014.7.0 (Helium)"
09:34 mikkn joined #salt
09:34 Roee and would like to start work with salt cloud
09:34 Roee can someone explain how to active salt cloud
09:35 Roee should i install it differently or it's built in in my version
09:40 lynxman joined #salt
09:41 otter768 joined #salt
09:54 bhosmer joined #salt
09:55 fllr joined #salt
10:06 aquinas joined #salt
10:09 che-arne joined #salt
10:12 aquassaut joined #salt
10:13 matthew-parlette joined #salt
10:15 evle joined #salt
10:32 linjan joined #salt
10:35 I3olle joined #salt
10:36 tmh1999 joined #salt
10:40 monkey66 joined #salt
10:53 huddy joined #salt
10:54 dyasny joined #salt
10:56 chandra Hi
10:59 chandra left #salt
11:10 andreyst joined #salt
11:10 drawsmcgraw joined #salt
11:12 felskrone joined #salt
11:14 jespada joined #salt
11:15 drawsmcgraw joined #salt
11:15 felskrone joined #salt
11:17 amcorreia_ joined #salt
11:20 fredvd joined #salt
11:22 mortis_ any vids from saltconf 2015 out yet? :)
11:32 drawsmcgraw mortis_: I'm told subscribing to the youtube channel (https://www.youtube.com/user/SaltStack) is the best way to be notified of when they're available
11:37 patrek joined #salt
11:40 mortis_ drawsmcgraw: thanks, yeah, checked that, but only old stuff :)
11:40 mortis_ wonder if they recorded the talks
11:41 mortis_ would be great
11:42 otter768 joined #salt
11:44 fllr joined #salt
11:44 lothiraldan joined #salt
11:45 jespada_ joined #salt
11:46 diegows joined #salt
11:51 denys joined #salt
11:53 nocturn joined #salt
11:54 nocturn Hi, I would like to use a pillar inside a jinja template, but it doesn't work when I do {% for domain in pillar['dnsdomains'] %}
11:54 arnoldB joined #salt
11:54 nocturn Can I do that and what am I missing?
11:55 nich0s joined #salt
11:59 TyrfingMjolnir joined #salt
12:01 Guest15 joined #salt
12:14 intellix joined #salt
12:17 genediazjr joined #salt
12:22 drawsmcgraw mortis_: They did. I imagine it'll just be a bit before they're online.
12:22 mortis_ sweet drawsmcgraw :)
12:25 mafro joined #salt
12:26 Nazca joined #salt
12:27 johtso joined #salt
12:33 lothiraldan joined #salt
12:33 I3olle joined #salt
12:37 wincyj joined #salt
12:42 clintberry joined #salt
12:45 JlRd joined #salt
12:45 cpowell joined #salt
12:46 nocturn nm, I made it work :-)
12:52 subsignal joined #salt
12:53 the_lalelu grml ... found a bug in my commits.
12:53 the_lalelu sorry about that. ;)
12:53 cpowell joined #salt
13:00 seb` joined #salt
13:00 pdayton joined #salt
13:02 saltymoli joined #salt
13:05 perfectsine joined #salt
13:06 peters-tx2 Hmm, I need to write some outputters....
13:20 teall joined #salt
13:27 numkem joined #salt
13:30 thayne joined #salt
13:33 fllr joined #salt
13:34 jerematic joined #salt
13:35 bhosmer joined #salt
13:35 lietu joined #salt
13:38 amcorreia_ joined #salt
13:43 otter768 joined #salt
13:46 cmcmacken joined #salt
13:48 bluenemo joined #salt
13:48 bluenemo joined #salt
13:50 murrdoc joined #salt
13:54 debian112 joined #salt
13:57 douglassssssssss joined #salt
13:58 elfixit joined #salt
13:59 andrew_v joined #salt
14:02 drawsmcgraw If I'm using `user.present` and I specify a 'maxdays' of 0, will that disable password expiration?
14:02 drawsmcgraw I ask because we're using SSH keys instead of passwords and I don't want non-existant password expiries locking us out in the future.
14:04 murrdoc probably better to use expire
14:04 murrdoc or empty_password
14:04 murrdoc http://docs.saltstack.com/en/latest/ref/states/all/salt.states.user.html#salt.states.user.present
14:04 drawsmcgraw As in, just set it to some far-distant date in the future?
14:05 drawsmcgraw oh! That's probably much more elegant than what I had in mind
14:05 drawsmcgraw Thanks murrdoc!
14:09 pdayton left #salt
14:13 pdayton joined #salt
14:19 faliarin joined #salt
14:21 amcorreia joined #salt
14:26 andreyst1 joined #salt
14:28 ja-s joined #salt
14:28 tenobi joined #salt
14:32 clmsy joined #salt
14:32 conan_the_destro joined #salt
14:33 WesleyTech joined #salt
14:36 WesleyTech_ joined #salt
14:36 dn5 joined #salt
14:37 SneakyPhil joined #salt
14:38 jansauer joined #salt
14:39 raygunsix joined #salt
14:39 rn joined #salt
14:41 jansauer left #salt
14:42 brick__ joined #salt
14:42 thayne joined #salt
14:42 jansauer joined #salt
14:43 jansauer left #salt
14:44 brick__ joined #salt
14:44 jaloren joined #salt
14:45 jaloren how i use the python api to execute a command asychronously http://docs.saltstack.com/en/latest/ref/clients/#localclient
14:45 stwb joined #salt
14:45 murrdoc its natively async
14:46 jaloren murrdoc: the messaging is but not the client command
14:46 jaloren by default it is synchrnoous
14:46 jaloren you have to pass the --async to make he client asychronously as well
14:46 jaloren otherwise it has a timeout to wait for all the clients to respond
14:47 jaloren i want to simulate the asychronous part via the salt utilities --async command option
14:48 Exploit joined #salt
14:51 jaloren murrdoc: i am a moron....cmd_async
14:51 jaloren boo
14:52 clintberry joined #salt
14:52 Brew joined #salt
14:52 murrdoc :)
14:53 ALLmight_ joined #salt
14:55 hasues joined #salt
14:56 techdragon joined #salt
14:56 jalbretsen joined #salt
14:58 Andre-B joined #salt
15:01 losh joined #salt
15:01 shaggy_surfer joined #salt
15:02 adelcast joined #salt
15:02 lytchi_ joined #salt
15:04 nich0s joined #salt
15:08 sjwoodr joined #salt
15:08 bryguy joined #salt
15:08 conan_the_destro joined #salt
15:10 jdesilet joined #salt
15:11 monkey66 joined #salt
15:16 monkey66 joined #salt
15:19 lytchi joined #salt
15:21 schristensen joined #salt
15:21 iggy Jouke: nope
15:21 nicolerenee joined #salt
15:22 fllr joined #salt
15:23 johanek joined #salt
15:23 iggy mortis_: they recorded everything (except mine)
15:23 timoguin joined #salt
15:24 monkey66 joined #salt
15:24 basepi iggy: still mad about that, I wanted to see yours!
15:25 iggy the only thing you missed was me talking like the micro-machine man apparently
15:25 murrdoc video out ?
15:25 basepi And missing that was a travesty.
15:25 basepi murrdoc: nah, will come out over the coming weeks.
15:25 murrdoc i should be in iggy's video, got jokes
15:25 basepi Since Rhett didn't have to do that editing himself (since we outsourced the video) it should come out more consistently this year.
15:25 murrdoc nice basepi
15:25 nich0s basepi: I was there. I saw the whole thing. It was glorious.
15:26 murrdoc i was hoping to say wassup to you too, but u werent around on irc
15:26 murrdoc so u know almost not existant
15:26 basepi nich0s: You must have clones, there were 5 or 6 concurrent tracks! ;)
15:26 basepi But yes, the conference was awesome
15:26 basepi If you missed it, come next year!
15:26 basepi nich0s: oh, you were saying iggy's talk
15:26 * basepi is jealous
15:27 nich0s Yup. Just iggy's talk.
15:27 rgarcia_ joined #salt
15:27 iggy it was my first talk... I expect to do better next year
15:28 timoguin The conf was great. Didn't get to meet as many people as I wanted though.
15:28 basepi Same. It seems like I only met about 10% of the people that I wish I had met. =P
15:28 nich0s timoguin: Agreed. I met quite a few people, though. Next year, I'm hoping to connect to a few more.
15:29 murrdoc that nichos dude was straight stalking
15:29 iggy I never got to say hi to utah_dave
15:29 murrdoc i mean sup nich0s
15:29 iggy I think that's the only one I really missed missing
15:29 basepi Ooh, that is a travesty, UtahDave is awesome.
15:29 nich0s iggy: You did fine, though, really. There were worse talks.
15:29 murrdoc no there werent
15:29 murrdoc :D
15:29 iggy there really weren't
15:29 murrdoc there might have been
15:29 nich0s murrdoc: The shady dude in the back of the room speaks. :P
15:30 monkey66 joined #salt
15:30 basepi I didn't hit any bad talks, but I only hit about 25% of talks
15:30 basepi Or maybe 20%
15:30 murrdoc basepi:  one thing i will say though, the talks didnt show enough use cases/examples
15:30 basepi I'm sure they were all great
15:30 murrdoc not that u are asking for recommendations
15:31 iggy I'll always remember murrdoc as the brown guy with the yellow linen shirt
15:31 basepi murrdoc: Interesting. I was actually worried there were *too man* "case study" talks
15:31 fllr joined #salt
15:31 basepi too many*
15:31 jtang hmm, i'd love to go to saltconf, but my employers probably wont approve :P
15:31 murrdoc no the case study talks didnt present enough case study
15:31 jtang salt lake city is nice
15:31 basepi murrdoc: good feedback
15:31 iggy what murrdoc said
15:31 StDiluted joined #salt
15:32 bluenemo_ joined #salt
15:32 iggy (I was guilty too)
15:32 murrdoc for eg, the first talk i attended the salt.util.http talk, great talk, walked through the code and all that
15:32 murrdoc i really wanted to know what problems it solved
15:32 iggy they all started off with 15 minutes of jibber jabber about people, companies, places, etc.
15:32 nich0s (and bad jokes)
15:32 iggy and that
15:33 basepi You're right, I did see a lot of that.
15:33 murrdoc good talks were gareth cron one, the emc talk , the cloudflare one and almost everytime thatch spoke
15:33 timoguin There were a lot of great case studies. They were my favorite.
15:33 basepi The most intense case study talk I saw was probably ryan lane's
15:33 murrdoc that dude … much respect
15:33 murrdoc yeah ryans stuff was +1
15:34 basepi It's fun to see a system that was architected from the ground up to use Salt
15:34 timoguin indeed. EMC's got my head spinning about formulas....
15:34 murrdoc i will basically take his talk about ordered execution and make my whole team sit through it
15:34 iggy yeah, even though he hates on the formulas too much
15:34 basepi Rather than being bound down in legacy code
15:34 overyander joined #salt
15:34 timoguin iggy: I wouldn't say too much. It's deserved hate at this point.
15:34 basepi You must not have heard me hate on the formulas, iggy.
15:34 murrdoc formulas is modularization
15:35 iggy that's what I wanted to talk more about (fresh start with salt), but my notes got left on the table (was expecting a podium)
15:35 timoguin basepi: Did you see what EMC is doing? They're building .deb packages for their formulas.
15:35 murrdoc i hate the part where it pollutes the pillars
15:35 basepi timoguin: nice.
15:35 murrdoc but otherwise <3 formulas
15:35 basepi Formulas are good in theory. But only a few are being maintained, none are being versioned, etc.
15:35 basepi When we have time to make them good, then I may stop hating.
15:35 timoguin We had a nice roundtable at the conf about fixing some of the issues.
15:35 iggy oh, that reminds me
15:35 murrdoc u havent seen the work nmadhok is putting in
15:35 basepi timoguin: oh?
15:35 iggy I forgot I took notes last week
15:36 timoguin yea, seth was there
15:36 murrdoc oh snap missed that
15:36 overyander what is the worker count (on master) to minion ratio?
15:36 basepi Ah, good, he's been our main contact point for formulas.
15:36 murrdoc when was that
15:36 nich0s timoguin: Yeah, it was like 90% emc folks. :P
15:36 murrdoc timoguin:  share noteses
15:36 timoguin I didn't take any. I just happened to sit down at the right table for lunch.
15:36 murrdoc as a thank you , heres emc's talk
15:36 murrdoc http://www.slideshare.net/stevendgonzales/saltconf2015-saltstack-at-scale-automating-your-automation
15:36 nich0s https://groups.google.com/forum/#!topic/salt-formulas/
15:37 timoguin There's going to be follow-up on the salt-formulas mailing list.
15:37 iggy yeah, I wanted to go to that, but the most info I could get out of anyone was "at lunch" like that narrows things down at all in a 500+ person conference
15:37 basepi overyander: worker procs doesn't need to rise much with a rise in minion count. Kind of a logarithmic scale or something. I know of people who are running 8000 minions on a single master with <30 worker procs.
15:37 murrdoc i think settings.sls pulled into map.jinja is the right way and, all love to forrest
15:37 iggy eww... I hate the formulas with settings.sls
15:37 basepi I was sad forrest couldn't come this year. =(
15:38 basepi I like that guy
15:38 murrdoc i hate having to include pillars for every formula i want to use
15:38 murrdoc along with every formula i dont want to use
15:38 murrdoc pillar.get is useless for me
15:38 murrdoc or pillar.items
15:38 iggy then how do you expect to configure them?
15:38 Ahlee boo, https://github.com/saltstack/salt/issues/2896
15:38 iggy magic?
15:38 timoguin The other option is to just automatically include all formulas?
15:38 murrdoc you can still configure them using pillars
15:39 iggy psychic paper?
15:39 murrdoc no smarty pants
15:39 murrdoc setup defaults in settings.sls, setup os based stuff in map.jina, then pull in settings.sls and then to the filter_by to pull in formula:lookup
15:39 hasues joined #salt
15:40 murrdoc after you include the map.jina, u have one contextual variable with all settings for your formual
15:40 murrdoc it makes sense
15:40 timoguin I just set my defaults in map.jinja and merge with pillar at the end.
15:40 iggy I _might_ be okay with that in theory
15:40 murrdoc so now when u do salt 'minion' pillar.items, u get all relevant confgs
15:40 basepi Ahlee: haha, that one was filed before I even worked at Salt! =P
15:41 amcorreia joined #salt
15:41 murrdoc Ahlee:  our resident hipster/greybeard
15:41 murrdoc Ahlee:  chicago meetup!
15:41 jri left #salt
15:41 iggy I just don't want it to end up all ugly as fuck like the zookeeper formula
15:42 murrdoc timoguin's is clean too
15:42 Ahlee basepi: sounds like a great issue to cut your teeth on today then, right?
15:42 Ahlee ;)
15:42 bhosmer_ joined #salt
15:42 iggy murrdoc: make an example? (or do I just go look at emc's talk?)
15:42 basepi Ha! If I weren't 350 conversations behind on salt-users, I might have a snowball's chance in hell of coding today. =P
15:43 murrdoc emc's talk and then look at the hadoop example in saltstack-firmulas
15:43 Ahlee murrdoc: yeah yeah
15:43 murrdoc here i was going to support your basepi naggings Ahlee
15:43 * murrdoc shushes up
15:43 basepi hehehe
15:43 Ahlee murrdoc: going to do an introduction to salt talk at the geekon meetup
15:44 murrdoc geekon ?
15:44 basepi Seriously though, that particular feature has just fallen by the wayside. I don't feel like it would be very hard....
15:44 otter768 joined #salt
15:44 Ahlee groupon hosts a weekly meetup for all things geekish
15:44 iggy nope
15:44 Ahlee i can get to it "later"
15:44 iggy that is ugly as all hell
15:44 monkey66 joined #salt
15:44 Ahlee i'm currently neck deep in trying to make schedule async so i can pass a timeout so i can upper bound a running scheduled task
15:45 iggy isn't it already async in 2015.2?
15:45 murrdoc use two ATs
15:45 murrdoc one to fire it off
15:45 murrdoc one to kill it if its still running
15:46 murrdoc why yes i did do lotsa webdev, why do u ask
15:47 Ahlee iggy: https://github.com/saltstack/salt/blob/develop/salt/utils/__init__.py#L324
15:47 denys joined #salt
15:49 murrdoc whats the salt-formula group ?
15:49 nich0s https://groups.google.com/forum/#!topic/salt-formulas/
15:50 murrdoc :|
15:50 iggy it's the group that only 2 people with ridiculous ideas post to
15:51 timoguin Because you haven't posted your good ideas yet!
15:51 murrdoc there is this brian jackson dude snitching about irc convos
15:51 timoguin haha
15:51 iggy yeah, that guy is not to be trusted
15:52 nich0s Yea- wait a minute...
15:52 kawa2014 joined #salt
15:53 Ahlee i never did find you timoguin
15:53 Ahlee or maybe i did, and just didn't know it'
15:54 timoguin I didn't turn IRC on until the last day for some reason...
15:55 Ahlee lol, low hanging fruit label
15:55 Ahlee thanks basepi
15:55 basepi hehehe
15:55 timoguin Those are the things you work on if you aren't very tall.
15:55 basepi It won't be a 20 minute fix, but is something that someone new to salt could probably figure out. =)
15:55 iggy I spent most of the "who am I? Where am I?" time at the beginning of talks on irc
15:55 VSpike Anyone know ... this documentation http://docs.saltstack.com/en/latest/topics/cloud/windows.html#firewall-settings suggest that passing user data to AWS instances is now supported in dev version, but I couldn't find any other corroboration of that. Is it, or not?
15:55 ggoZ joined #salt
15:56 Ahlee k, work beckons
15:56 Ahlee stay classy folks
15:56 basepi iggy: sounds like a good use of time. =)
15:56 basepi Ahlee: NEVER!
15:56 murrdoc bam, /me starts shit
15:56 diegows joined #salt
15:56 tligda joined #salt
15:56 timoguin VSpike: yea that'd be nice to know! it's something i've been needing as well...
15:57 timoguin To the code!
15:57 timoguin also, beacons and engines are very exciting
15:57 thayne joined #salt
15:58 VSpike timoguin: i had a look, but I'm not hugely familiar with the codebase yet
15:58 timoguin https://github.com/saltstack/salt/blob/develop/salt/cloud/clouds/ec2.py#L1417
15:58 timoguin Looks like it's possible now!
15:59 VSpike that would be useful. I can't get the existing windows bootstrap to work, but if I could pass something in userdata, I could download and run my own bootstrap script at least
15:59 timoguin It's in 2015.2
15:59 timoguin Well, will be....
16:00 timoguin Yea, userdata is hugely helpful.
16:00 timoguin I need to be able to pass instance hostnames via userdata to be able to automatically configure hostnames on EC2
16:00 murrdoc iggy:  2 cents it up https://groups.google.com/forum/#!topic/salt-formulas/tRXex3FKVHo
16:00 murrdoc timoguin:  u too if u want
16:00 hasues left #salt
16:00 ja-s joined #salt
16:00 beneggett joined #salt
16:01 VSpike timoguin: I was going to try and read them from the EC2 name tag
16:01 monkey66 joined #salt
16:01 VSpike Although it's surprisingly hard for an instance to read its own tags. On Windows at least, havent tried on Linux
16:02 timoguin VSpike: I'm wanting to do this: http://www.ducea.com/2009/06/01/howto-update-dns-hostnames-automatically-for-your-amazon-ec2-instances/
16:02 iggy done
16:02 timoguin @murrdoc: so, there are formulas with pillars in them that are supposed to get included into the pillar tre?
16:03 iggy lol, 2015.2... thatch caught so much crap about that the first day
16:03 timoguin and the last day
16:03 murrdoc yeah cause they all give u pillar.example
16:03 murrdoc which the code assumes
16:03 VSpike timoguin: oh yeah, that looks handy
16:03 timoguin murrdoc: that's gross. pillar.example is just an example. i don't ever actually use that.
16:03 beneggett joined #salt
16:03 iggy I still don't understand the problem really
16:03 timoguin and if formulas are importing that or something... that's yuck
16:04 iggy you want to use a reusable, configurable "thing" without configuring it... seems backwards
16:05 malinoff joined #salt
16:05 timoguin Formulas should be able to be installed with no config to get a sane baseline default.
16:05 timoguin I think at least...
16:06 murrdoc the formulas in saltstack-formulas just work ? without a base pillar ?
16:06 VSpike Next time could y'all have the saltconf in the UK, e.g. Old Salt Quay, London (if "salt" in the hame is requirement)? thx
16:06 timoguin murrdoc: yes, many of them
16:06 iggy I would not expect that to work
16:06 iggy I don't think that's even a good idea
16:07 Matt__ joined #salt
16:07 timoguin Why not? If I want to install something with all the defaults settings, why would I want to have to configure it?
16:07 iggy what's default settings?
16:07 desposo joined #salt
16:07 jespada_ joined #salt
16:08 timoguin iggy: something like the OS installer for the package might put in place.
16:08 iggy let's take the aptly formula for example... what is the default place to look for pkgs to put in a repo?
16:08 timoguin If you want to test something out, it should be easy and with no config work required.
16:08 iggy some things have defaults that make sense... a lot of things don't
16:08 basepi I don't know that it makes sense for all formulas, but I think it makes sense for many of them. Like timoguin said, defaults like a package manager might make.
16:09 timoguin I'm sure there are certain apps wehre that might not make sense
16:09 iggy I think there are quite a lot (at least the stuff I use)
16:09 murrdoc for example the nginx formula
16:09 VSpike Ah, 2015.2 not in Ubuntu PPA yet.
16:10 pdayton1 joined #salt
16:10 iggy 2015.2 not out yet
16:10 schristensen joined #salt
16:10 murrdoc the os defaults are what packages to install
16:11 VSpike Ah yes, that probably explains it. Seeing "2015.2 release notes" in a google search, combined with the version number suggest Feb made me think I'd missed a release and not noticed
16:11 timoguin Well the packages will put config files in place too. So if I were to install salt with a formula, I'd expect to get the default config files if I didn't configure pillar.
16:11 teskew joined #salt
16:12 iggy I'm waiting for an example where that isn't what happens today
16:12 timoguin But then there's also the issue of keeping the formula / config file templates up to date with the software, which is an issue I've ran into.
16:12 iggy to go back to the aptly formula, it has defaults in the pillar lookups for things that make sense
16:13 iggy that's what pillar.get is for
16:13 timoguin iggy: yea looking at it. not too familiar with aptly.
16:13 timoguin are there some parts of it that just don't make sense to have a default?
16:13 iggy where your packages are, what gpg key to use, etc.
16:14 hal58th joined #salt
16:14 timoguin The aptly:rootdir pillar key?
16:15 iggy that one has a default
16:15 timoguin I guess that makes sense.
16:16 iggy there's something to be said for centralizing those lookups to ease maintainability, but settings.sls is way overboard (and wrong in that it looks up config from grains)
16:16 timoguin Keys are probably the clearest example of things you might not want to default.
16:16 timoguin I tend to give my internal formulas insecure defaults just to make local testing easier.
16:16 andygrunwald joined #salt
16:17 ja-s joined #salt
16:17 timoguin But I can see how that'd be bad for a community formula.
16:18 triplecorn joined #salt
16:18 andygrunwald Hey, a short question: If i configure a job_cache / returner on the master. Why does the minions need access to the job cache target system. E.g. I configure redis or kafka. Why does every minion need access to the redis or kafka directly?
16:19 beardo joined #salt
16:20 timoguin andygrunwald: because each minion handles the return itself.
16:20 jab416171 joined #salt
16:20 timoguin andygrunwald: there's a new setting that lets the master handle it though....
16:21 triplecorn joined #salt
16:21 MatthewsFace joined #salt
16:22 andygrunwald timoguin, do you know how the new setting is called?
16:22 timoguin andygrunwald: it's called master_job_cache, in 2014.7+
16:22 timoguin had to look it up
16:22 andygrunwald Thx!
16:22 timoguin just learned about it at the conference
16:22 andygrunwald I will try it
16:22 andygrunwald Thanks :)
16:23 felskrone andygrunwald: if the master_job_cache does not suit your requirements, have a look at salt-eventsd, that might be a better fit https://github.com/felskrone/salt-eventsd/
16:25 mafro joined #salt
16:28 theologian joined #salt
16:30 johanek joined #salt
16:30 shaggy_surfer joined #salt
16:33 gladiatr joined #salt
16:33 codysoyland joined #salt
16:36 hal58th basepi, just a reminder to talk to whomever about when the ssc.saltstack.com will be updated with test results.
16:36 monkey66 joined #salt
16:37 aparsons joined #salt
16:38 devweasel joined #salt
16:42 andygrunwald thanks felskrone for this idea. PS Felskrone is a german beer afaik ;)
16:43 felskrone you're correct :-)
16:45 basepi hal58th: sounds like in the next day or two. We're working on the e-mails now.
16:45 hal58th ok thanks basepi
16:47 kawa2014 joined #salt
16:48 Ahrotahntee joined #salt
16:49 Ahrotahntee huh, I left?
16:49 Ahrotahntee that doesn't seem like me
16:49 monkey66 joined #salt
16:50 apergos joined #salt
16:50 mimianddaniel joined #salt
16:51 mimianddaniel anyone here has zmq_monitor set and using it?
16:52 iggy seems like it's not something you'd want generally set
16:53 cpowell joined #salt
16:53 iggy Mike talked about it a little in his troubleshooting/optimizing talk (iirc)
16:53 iggy which generally means to me there's probably about 5 people in the world that use it
16:54 kevinforbes joined #salt
16:54 apergos good evening or morning as the case may be... I'm wondering (hoping) that the fix for https://github.com/saltstack/salt/issues/18317 (see https://github.com/saltstack/salt/commit/cec5ba34c62605d654be11d6c27c61c346e78cd0) will be backported to the 2014.7.x branch.
16:54 zwi joined #salt
16:54 apergos Or, alternatively, if there is a workaround that would permit us to send multiple arguments in with some other formatting
16:55 JPT joined #salt
16:56 timoguin joined #salt
16:58 hal58th apergos, Salt doesn't really backport. They just release a new version. Sometimes minor fixes get their own version but I wouldn't hope for it. Looks like the milestone for that is Lithium which is the next realse
16:58 apergos argh
16:58 monkey66 joined #salt
16:58 apergos really don't want to maintain my own set of packages or 4 platforms for this bug
16:58 apergos *for 4
16:59 Ryan_Lane joined #salt
17:00 iggy look at the PR
17:00 iggy it was merged into the 2014.7 branch
17:00 iggy that means it'll be forward ported to 2015.2 and devel
17:00 iggy so yes it should be in the next 2014.7 release (if there is one)
17:01 hal58th Where is this PR? I don't know where you are looking iggy
17:01 iggy https://github.com/saltstack/salt/pull/20992
17:01 iggy linked from the original bug
17:01 felskrone has anyone written tabs to a file with file.managed->contents:… ? can seem to get it right, i get compile errors or i end up with actual \t strings in my file
17:02 Aikar felskrone, paste a tab?
17:02 Aikar inside if " "
17:02 felskrone like ctrl+v+tab?
17:02 Aikar likely cant write a raw tab outside of " "
17:03 Aikar or just press tab inside of " "
17:03 Aikar or look at simply using a ninja template instead
17:03 felskrone sorry but that sounds really wrong as a "solution" :-)
17:03 Aikar jinja....*
17:03 Aikar how is using a tab character wrong when thats what your wanting to do
17:04 iggy if it's a leading tab, I think you're boned
17:04 felskrone because putting a literal tab in the string and being able to see it is dependant on your editor
17:05 denys joined #salt
17:05 Aikar why is that an issue? if you depend on a tab, then use it. or use a source file vs contents
17:06 amcorreia joined #salt
17:06 ndrei joined #salt
17:08 sym-jp joined #salt
17:09 felskrone Aikar: never mind, thx anyway :-)
17:10 toanju joined #salt
17:15 mimianddaniel iggy: haha
17:15 ajw0100 joined #salt
17:15 mimianddaniel iggy: thats very assuring
17:15 mimianddaniel iggy: i was in that talk as well and which is the reason why i wanted to test it
17:16 mimianddaniel iggy: its a bit too premature for us to move to raet, even tho we are seein some issues on zmq
17:16 iggy oh cool, yeah there was some good stuff in that talk
17:16 iggy do you actually have an issue you are trying to solve?
17:16 mimianddaniel iggy: yes ...
17:16 mimianddaniel specifically related to missing return events
17:17 rsimpkins left #salt
17:17 mimianddaniel i am suspecting that keep-alive set in the ret that mike mentioned to be the culprit
17:17 mimianddaniel thats only avail in devel irght now
17:17 mimianddaniel which i am going to push ahead and merge into our repo and test it
17:17 mimianddaniel but i would like to get more visibility to zmq tranport at the same time
17:18 iggy yeah, I'm probably going to be testing raet in QA this week
17:18 iggy we have issues galore in GCE with networking
17:18 mimianddaniel iggy: do you work for saltstack as well ?
17:18 iggy no, it's too cold there
17:19 mimianddaniel iggy: it wasnt too bad
17:19 mimianddaniel iggy: but i am from Canada
17:19 bantone ha
17:19 bantone saltconf was some of the best conference food I've had
17:19 mimianddaniel bantone: ++
17:19 iggy I'm from TX... it was balls cold and I've been sick since like my second day there
17:19 bantone and some good speakers
17:19 bantone iggy: what part of tx?
17:19 bantone <-- san antonio
17:19 iggy H-town
17:19 bantone ah not too far
17:20 murrdoc its where the WHATABURGER is
17:20 mimianddaniel bantone: yes indeed i really appreciated some of info shared by the speakers
17:20 bantone yeah I passed by you and didn't put two and two together until last minute heh
17:20 iggy I'm actually going to SA this weekend (tattoo artist is there)
17:20 bantone I just wish I didn't look at the salt-cloud stuff
17:20 bantone nice
17:20 bantone ya a few of us from rackspace were there
17:20 * mimianddaniel going to bother Mike again with zmq_monitor thingy
17:21 mimianddaniel i need some answers!!!
17:21 iggy oh, yeah, I saw some rackspace shirts/jackets wandering around
17:21 iggy I'm terrible with names though, so I wasn't really sure who all to say hi to
17:21 bantone haha yea
17:22 bantone well myself and manfred (daniel) were the only two of us from the chan that was there probably
17:23 iggy Next year I'm going to wear a sky blue velour track suit so everybody can find me easily
17:23 murrdoc u dont stick out at all
17:24 iggy that doesn't mean much from the guy that probably got 18 "random" searches on his way home ;)
17:24 bantone lol
17:24 mafro joined #salt
17:24 murrdoc 3
17:25 itru joined #salt
17:26 forrest joined #salt
17:27 iggy You know, the sheer number of times thatch mentioned the deps for raet, you'd think I'd remember at least some of them
17:27 I3olle joined #salt
17:27 iggy and why doesn't salt-bootstrap have raet support!
17:28 marcel joined #salt
17:28 murrdoc its a flag
17:29 iggy I don't think so
17:29 iggy I did actually go look before I made such a bold statement
17:30 aron_kexp joined #salt
17:31 baweaver joined #salt
17:32 manfred iggy:  i was the tall ginger with a rackspace shirt on on thursday
17:32 hal58th was it you I met manfred at the end of the keynote on wednesday?
17:33 iggy I'd say you tried to steal my soul if I hadn't sold it years ago
17:34 hybridpollo joined #salt
17:34 drawsmcgraw oh hey it's manfred! Are you done with the blog post version of your talk from Thursday?
17:34 teall joined #salt
17:35 drawsmcgraw btw I'm the guy that accosted you after said talk, which was excellent
17:35 manfred :)
17:35 manfred i have not
17:35 manfred i mostly slept yesterday
17:35 manfred it will probably get done next weekend
17:35 drawsmcgraw I can understand that
17:36 manfred hal58th:  probably? what did we talk about?
17:36 hal58th manfred: nothing really, just said hi since I was sitting a couple rows behind you
17:36 murrdoc d'aawww
17:36 murrdoc u guys
17:37 manfred ahh then maybe
17:38 theologian joined #salt
17:39 timoguin Is the #salt-dev channel no more?
17:39 manfred it is #salt-devel
17:39 timoguin thanks
17:40 manfred np
17:40 iggy if I start haning out there, I'll stop answering noob questions in here
17:40 manfred that channel is mostly for watching github
17:43 racooper joined #salt
17:43 shaggy_surfer joined #salt
17:43 ajw0100 joined #salt
17:43 amcorreia joined #salt
17:44 linjan joined #salt
17:44 spookah joined #salt
17:45 felskrone joined #salt
17:45 otter768 joined #salt
17:46 apergos iggy: thanks for the heads up, that's good to know. now I just need to nag about a new 2014.7 :-D
17:47 andrew_v joined #salt
17:51 baweaver joined #salt
17:54 enarciso joined #salt
17:54 peters-tx Got an email from saltstack re: cert.  "We just wanted to let you know that we will be sending you an update email when your scores are loaded into https://ssc.saltstack.com.  We are also aware of a few issues that previous test-takers have had when trying to update their information there and we are working on it."
17:56 iggy Sweet... Yeah, I'm pretty sure my cert ended up with "Iggy Jackson" as the name (despite me telling the guy my actual first name)
17:56 drawsmcgraw I think having a handle instead of a name on your cert is pretty cool, actually
17:56 timoguin Except employeers knowing your handle is a little weird...
17:57 iggy everybody at work calls me Iggy
17:57 drawsmcgraw Potentially. Unless your portfolio is your Github account
17:57 iggy it's even on my name tag on my cubicle
17:57 timoguin Mine's not weird because it's my name... but yea
17:57 kawa2014 joined #salt
17:58 Ahrotahntee a custom module that runs only on the master is a runner, right?
17:59 iggy generally
18:05 toanju joined #salt
18:05 KyleG joined #salt
18:05 KyleG joined #salt
18:08 wincyj joined #salt
18:08 intellix joined #salt
18:15 devweasel left #salt
18:16 ndrei joined #salt
18:24 mafro joined #salt
18:30 bash1245_ joined #salt
18:33 Andre-B joined #salt
18:36 chitown anyone up to speed on the apt packing?   specifically, i only see 2014.7.1
18:37 murrdoc yeah thats the only one up there
18:38 jespada_ joined #salt
18:39 jri joined #salt
18:39 iggy who knows what's going on...
18:39 desposo joined #salt
18:42 chitown that brings up another question: who exactly is responsible for the various pkgs?
18:42 chitown is it "salt" or someone from yum/RPM, apt/deb, ....?
18:42 iggy depends
18:42 chitown iiuc, usually there is a "pkg maintainer"
18:43 chitown depends... nice... that is my answer for almost everything :)
18:43 iggy ubuntu ppa is maintained by someone (outside of the company)
18:44 iggy debian I think is by someone inside the company
18:44 jespada__ joined #salt
18:44 iggy everything else, nfc
18:44 chitown good point: http://ppa.launchpad.net/saltstack/salt/ubuntu/
18:44 chitown that one specifically :)
18:45 jtanner joined #salt
18:45 murrdoc pi base can help u with, or maybe he cant, its just fun to ask him
18:46 iggy I think there was a problem
18:46 iggy but generally speaking joehh iirc
18:46 ahammond I'm looking at https://copr.fedoraproject.org/coprs/saltstack/zeromq4/ and see a clever "restart the minion when I'm done" trick using at
18:47 ahammond is there an alternative solution that doesn't require at?
18:47 ahammond I'm asking because apparently at isn't part of the default install and I'd rather not add it
18:47 iggy salt 'whatever' service.restart salt-minion?
18:48 toanju joined #salt
18:48 ahammond iggy er... won't that bonk the minion while it's running?
18:48 ahammond iggy the trick is a cmd.wait that does echo service salt-minion restart | at now + 1 minute
18:48 iggy depends
18:49 ahammond iggy so, it should let the minion finish the current job (since it's waiting on a pkg.latest that has order: last)
18:49 iggy yeah, that's just really not needed most of the time
18:50 iggy so yeah, in that case, you've got to do weird stuff
18:50 ahammond iggy ok, I'll go with the simple / clean approach and Not Worry
18:51 iggy use listen/listen_in... I think ryan said that waited till the end to do the restarts/etc
18:51 denys joined #salt
18:52 jri joined #salt
18:53 murrdoc https://github.com/saltstack/salt/pull/21443 oooh
18:53 iggy I can't figure out why they are using order: there at all
18:53 murrdoc listen_in ftw
18:54 Gareth joined #salt
18:54 manfred iggy:  chitown debian and ubuntu are both maintained by joehh
18:54 thedodd joined #salt
18:57 gargrag joined #salt
18:57 _JZ_ joined #salt
18:59 peters-tx ahale, Yah, I put in a Salt git req to update it in fact
18:59 KyleG1 joined #salt
18:59 peters-tx sorry
18:59 peters-tx ahammond, Yah, I put in a Salt git req to update it in fact
18:59 gargrag Hello all! : What do you think is the correct way, to do this: (I am using the saltstack/nginx-formula) I need to create a nginx vhost from other state, that inherits nginx-formula, so which is the best way to create the vhost, by reusing the code on the nginx-formula, i was thinking about reuse a macro, but the formula doesn't have a jinja macro for that.
19:00 peters-tx ahammond, https://github.com/saltstack/salt/blob/develop/doc/faq.rst#id61
19:02 wincyj joined #salt
19:02 peters-tx ahammond, Fwiw, it works very well in my environment, I tested it quite a bit
19:02 linjan joined #salt
19:03 ajw0100 joined #salt
19:05 baweaver joined #salt
19:08 gargrag smdby?
19:08 iggy gargrag: that seems like a big job for a single jinja macro anyways, but if you can figure out a way to refactor that formula that makes sense for you and continues to function (and retains maintainability), go for it
19:09 monkey66 joined #salt
19:09 gargrag iggy: which approach you suggest me to do that, The thing is: is correct to create my own vhost definition while I have the nginx-formula that create vhosts from pillars?
19:10 theologian joined #salt
19:10 TaiSHi ping
19:11 TaiSHi Whoa, my first chat text in 3 days and I just write ping
19:11 TaiSHi I should be hanged.
19:11 zwi1 joined #salt
19:12 iggy gargrag: use whatever makes the most sense for your setup... I would try to make sure your files won't conflict with anything the formula might lay down in the future
19:13 gargrag thanks iggy, I am not clear at all, but I will look for a solution, that seems clear.
19:14 iggy gargrag: it's not unheard of to copy code from formulas
19:15 iggy that's part of why they are out there
19:15 Ahlee indeed.
19:15 Ahlee I'd never run code directly from the formulas
19:15 Ahlee as I do'nt want to be beholden to somebody else's upgrade schedule
19:15 iggy that's why you fork to your own org
19:15 iggy (or user)
19:16 Ahlee ah, good point
19:16 iggy (as is mentioned in the formulas docs...)
19:16 smcquay joined #salt
19:17 Ahlee who reads docs?
19:17 Ahlee I have you to read the docs for me
19:17 Ahlee I just make outlandish claims like I invented the question mark, and fine people such as yourself correct me where applicible
19:17 iggy people that are too lazy to read code (but not so lazy that they can't be arsed to at least put some pillar data down)
19:17 MorbusIff joined #salt
19:18 MatthewsFace joined #salt
19:18 igorwidl left #salt
19:19 gargrag joined #salt
19:20 gargrag iok iggy and Ahlee, i have read the docs
19:20 gargrag i'm just looking for the better way.
19:21 rap424 joined #salt
19:21 hybridpollo joined #salt
19:22 iggy we wrote our own nginx formula, but that's mostly because we have fairly simplistic requirements
19:22 ndrei joined #salt
19:23 hasues joined #salt
19:23 jcsp joined #salt
19:23 hasues left #salt
19:25 aron_kexp joined #salt
19:25 johanek joined #salt
19:25 monkey66 joined #salt
19:26 Fiber^ joined #salt
19:28 Nazca__ joined #salt
19:28 enarciso joined #salt
19:29 ckao joined #salt
19:30 plameniv joined #salt
19:35 ndrei joined #salt
19:38 baweaver joined #salt
19:38 desposo joined #salt
19:42 plameniv joined #salt
19:43 monkey66 joined #salt
19:43 bhosmer__ joined #salt
19:44 wormwood joined #salt
19:45 Ahrotahntee has anyone used exclude_pat recently, does it work with file.recurse now?
19:46 otter768 joined #salt
19:46 wormwood greetings all, new to salt and trying to use salt-ssh but keep getting the following stdout: /bin/sh: /home/tmp-user/.2014b841d85d: No such file or directory
19:46 wormwood running the command: salt-ssh -i -c . '*' cmd.run "ifconfig -a"
19:47 iggy Ahrotahntee: should work in devel (I read code for it last night when I was working on pillar:// uris)
19:47 iggy Ahrotahntee: I think it's actually expected to work in 2014.7+
19:48 Ahrotahntee I'll run a test
19:49 theologian joined #salt
19:49 thedodd joined #salt
19:51 hal58th wormwood what os are you trying to hit? Does it have python installed on target?
19:53 desposo joined #salt
19:53 wormwood hal58th: an ec2 instance, and yep, it has python installed
19:53 hal58th what os?
19:53 wormwood centos 7
19:54 wormwood tmp-user is a local user, so that stdout error looks like something on my end
19:54 zer0def joined #salt
19:54 jri joined #salt
19:54 wormwood looks like some tmp dir that salt uses
19:54 wormwood I don't have master or minion running, which according to the docs is fine
19:54 monkey66 joined #salt
19:55 hal58th yeah, salt-ssh will deploy a thin salt deployment on the target box in the temp directory.
19:55 Ahrotahntee iggy: I imagine you write salt (state) formulas from time to time, how often do you find yourself writing them in pure python?
19:55 linjan joined #salt
19:55 hal58th wormwood so it's odd that it tells you the file name but then it says it can't find it.
19:55 wormwood yeah, thats what I was thinking so I ran it through strace
19:56 wormwood but don't know enough about salt yet to comment on what I found
19:57 hal58th I would suggest running on debug but I don't know how much help it will be. just add "-l debug" to the end of the command
19:57 hal58th Also add "-W" to use a random thin dir. Maybe the old one got messed up.
19:58 overyander what is the proper way to do a pkg.install but only if a previous version of the same pkg is already installed?
19:58 wormwood hal58th: thanks for the help, I'll look at those
20:01 aboe joined #salt
20:01 jcsp joined #salt
20:03 jcsp joined #salt
20:03 iggy Ahrotahntee: haven't yet
20:03 monkey66 joined #salt
20:03 iggy which is odd because I have a decent python background
20:05 iggy overyander: pkg.latest\n - onlyif: <some command that would be installed by the pkg>
20:05 jcsp joined #salt
20:06 overyander iggy, what about for a winrepo?
20:06 darvon joined #salt
20:06 ksalman does onchanges not work similar to watch? onchanges work with one - file: but when I am watching two files (I have two - file: ) it doesn't see then changes in either of those files
20:07 iggy overyander: nfc, don't use Windows
20:08 overyander lol, ok
20:08 ksalman ha
20:09 jcsp joined #salt
20:09 iggy ksalman: do they both return true?
20:10 ksalman iggy: when i have two "- file:" and the change is only in one file, the state watching those files does not see any changes
20:11 ksalman i didn't try changes in both files
20:12 zwi joined #salt
20:12 mfournier joined #salt
20:12 ksalman iggy: it looks like when both files are changed, it does work
20:12 ksalman as expected
20:13 darvon joined #salt
20:14 iggy that's what the docs say
20:15 iggy only if the required states generate changes
20:15 ksalman iggy: one file does generage change, is that not enough?
20:16 madduck joined #salt
20:16 ksalman that's how I understand the doc
20:16 iggy both have to
20:16 ksalman huh
20:17 iggy it doesnt' say "if any of the required states generate changes"
20:17 hal58th hmmm, " The unless requisite operates as NOR and is useful in giving more granular control over when a state should execute." But it does not specifically say that about onchanges or onfail.
20:17 hal58th Guess I would have to look at the code to really know for sure
20:17 ksalman well poops
20:18 ksalman so there's no way to watch two files for changes
20:18 iggy I think the docs are fairly clear
20:18 Jimlad joined #salt
20:19 darvon joined #salt
20:19 soren_ joined #salt
20:19 eichiro joined #salt
20:19 MWheelz joined #salt
20:20 iggy ksalman: do you need the state to _only_ run if there are changes? if not, just use watch
20:20 stoogenmeyer joined #salt
20:20 ksalman iggy: yea, thats exactly why i was trying to use onchanges
20:20 stoogenmeyer hey does anyone know about bash completion for salt ?
20:21 hal58th eh, I err on the side of being more specific in my documentation. But you don't want to overwhelm the user
20:21 hal58th stoogenmeyer it doesn't work
20:21 stoogenmeyer also, silly question.. is it not overly complicated to get zsh + ohmyzsh on a minion using salt?
20:21 stoogenmeyer hal58th: oh thats a bummer.. im getting tired of writing sys.list_functions, etc
20:21 iggy zsh:\n  pkg.installed
20:22 soren_ With Puppet, I've been able to use its exported resources to automatically populate my /etc/hosts across the entire infrastructure with the info of every node. How can I do the same with Salt?
20:22 iggy ksalman: listen?
20:22 ksalman ahh..
20:22 ksalman i haven't really looked at that
20:22 ksalman hm
20:22 shaggy_surfer joined #salt
20:23 hal58th soren_ probably want to use "salt mine"
20:23 iggy although the docs say it behaves like watch, it also says succeed and result in changes
20:23 hal58th soren_http://docs.saltstack.com/en/latest/topics/mine/
20:23 stoogenmeyer iggy: thats the easy part.. what about ohmyzsh? thats usually installed using a script
20:23 iggy stoogenmeyer: never heard of it
20:24 Ryan_Lane soren_: if your architecture grows a lot, that's going to be really slow
20:24 soren_ Ryan_Lane: It won't.
20:24 Ryan_Lane gotcha :)
20:24 murrdoc Ryan_Lane:  did you put your presentations slides up anywhere ?
20:24 stoogenmeyer iggy: its a common occurrence where something is installed by curl'ing a script and piping it to sh
20:24 Ryan_Lane murrdoc: not yet
20:24 Ryan_Lane I need to
20:24 Ryan_Lane my slides by themselves aren't amazingly helpful
20:24 hal58th stoogenmeyer: cmd.wait state and then watch the pkg.install of zsh
20:25 murrdoc true true, but i consider them appetizers for the crew at work
20:25 Ryan_Lane heh
20:25 mafro joined #salt
20:25 murrdoc set them up for when the videos come out
20:25 iggy they are useful if you were at the talk
20:25 Ryan_Lane I'll try to post them today or tomorrow :)
20:25 stoogenmeyer hal58th: will the cmd.wait be run everytime i set the highstate.. ?
20:25 murrdoc Ryan_Lane:  i appreciate it
20:25 Ryan_Lane yw
20:26 soren_ Ryan_Lane: There's 8 nodes. It's unlikely to even double in size. It just happens that I spin up test environments where a populated /etc/hosts would be very handy.
20:26 hal58th cmd.run will run every time, but cmd.wait will only run if the state it is watching is changed. Look up the "watch" requisite.
20:26 iggy stoogenmeyer: there are various requisites (some that only cause states to run when the requisite has changes)
20:26 Ryan_Lane soren_: the thing I dislike about populating hosts that way is that it takes multiple runs for things to look correct
20:26 Jose4183 joined #salt
20:27 ndrei joined #salt
20:29 _JZ_ joined #salt
20:30 ksalman iggy: that did not work either. I converted my at.present to cmd.wait and i am just using watch
20:31 iggy I didn't say watch
20:32 ksalman I did try listen
20:33 iggy if you've tried them all and none of them work for you, then you seem to have found a good reason to go open an issue in github
20:33 Gareth ahoy hoy.
20:33 murrdoc what about unless
20:33 murrdoc or onlyif
20:34 iggy onlyif is "all of the commands"
20:35 murrdoc http://docs.saltstack.com/en/latest/ref/states/requisites.html#unless
20:35 murrdoc no ?
20:36 iggy those are also commands
20:36 iggy OP wants a state to only run if 2 others are true and change
20:36 iggy wait
20:37 theologian joined #salt
20:37 ksalman I want a state to only run of either or all of the files change
20:37 murrdoc http://docs.saltstack.com/en/latest/ref/states/requisites.html#onchanges
20:37 iggy OP wants a state to only run if 1 of 2 others are true and change
20:37 iggy onchanges is if all of them change
20:37 murrdoc op needs to simplify their life
20:38 iggy ^
20:38 murrdoc what are u trying to do OP
20:38 murrdoc the requisite states
20:38 murrdoc what are they doing
20:40 tomh- joined #salt
20:40 theBarGuy joined #salt
20:41 wincyj joined #salt
20:41 mgw joined #salt
20:42 theBarGuy i'm trying to ssh into a salt minion from my salt master via reverse tunnel
20:42 johanek joined #salt
20:42 theBarGuy and I can't quite figure out what commands to run
20:42 theBarGuy i've tried using cmd.run 'ssh user@remotehost' but it spits out an error
20:43 murrdoc hahah
20:43 murrdoc uh wrong window
20:44 theBarGuy balls
20:44 theBarGuy first time here
20:44 theBarGuy can you direct me to the right place?
20:44 murrdoc no no
20:44 murrdoc i was in the wrong window
20:44 murrdoc u are in the right place for salt stuff
20:45 racooper theBarGuy,  what you're trying would be to start an interactive session from a non-interactive salt session...I'm not sure how you would accomplish what you're describing
20:45 theBarGuy racooper: does that mean i'm just up shit creek?
20:47 hal58th theBarGuy, to do an interactive session, yes. To run just one ssh command, you would just need to change some ssh flags.
20:47 racooper you would need to do something with the -R optoin of ssh
20:47 iggy have you thought of trying to accomplish what you need to do a different way?
20:47 racooper http://www.thegeekstuff.com/2013/11/reverse-ssh-tunnel/
20:48 hal58th theBarGuy, all salt commands are for the most part asynchronus.
20:48 iggy Do you really need an interactive shell back to the minion? Can't you run anything you would run through ssh via salt anyway?
20:50 fxhp joined #salt
20:50 gladiatr joined #salt
20:52 jmccree I would love an interactive shell via salt.
20:53 teskew1 joined #salt
20:53 toanju joined #salt
20:53 theBarGuy i'm just starting off with salt
20:53 iggy I have only logged into my salt masters in the past month... one of the other guys here still ssh's straight to instances, but I'm trying to get him out of that habit
20:54 enarciso joined #salt
20:55 theBarGuy my naive solution was to see if i can ssh from my minion to my master so i can reverse tunnel through the port it binds to. i'm pretty much lost once it gives me the error you describe of not being able to start an interactive session
20:55 jmccree I have a few uses where opening ssh port is a hassle but salt's outgoing connection is no issue.
20:56 theBarGuy iggy: for my purposes i do need an interactive session for my minion
20:56 hal58th theBarGuy yes you need to learn some of the other flags that ssh uses for a proxy forward. No need for interactive session
20:57 sym-jp joined #salt
20:57 iggy that just seems backward... I got into salt so I wouldn't have to log into individual boxes anymore
20:58 iggy but hey, you might have a legitimate reason behind all the cloak and dagger secrecy
20:58 murrdoc pron
20:59 hasues joined #salt
21:00 hasues left #salt
21:05 theBarGuy ditto
21:09 murrdoc or w4rez
21:11 baweaver joined #salt
21:12 thedodd joined #salt
21:14 aron_kexp joined #salt
21:15 jeddi joined #salt
21:15 bash1245_ joined #salt
21:18 ja-s joined #salt
21:18 mpanetta joined #salt
21:19 mpanetta joined #salt
21:23 peters-tx joined #salt
21:27 jaloren joined #salt
21:27 aw110f joined #salt
21:27 jaloren i am try to target based on a compound match for nodegroups
21:27 jaloren salt -C 'N@slavenodes or N@masternode' test.ping
21:27 jaloren this isn't find anything
21:28 jaloren here are the nodegroups https://gist.github.com/jaloren/86fc9a9a84914fba448a
21:28 jri joined #salt
21:28 iggy I don't think nodegroups are supported in compound matching
21:28 jaloren iggy: this says it is http://docs.saltstack.com/en/latest/topics/targeting/compound.html
21:28 jaloren doesn't it?
21:29 iggy I don't see it anywhere on that page
21:29 lothiraldan joined #salt
21:30 iggy at the very least there's no "N@" anywhere on that page
21:30 pfallenop joined #salt
21:30 jaloren iggy: your right...what a bummer
21:30 jaloren oh well
21:30 jaloren i probably just need to write a custom salt client and call it a day
21:30 iggy you're and yeah, I've seen a number of people try to do it
21:31 jaloren pardon?
21:31 iggy you aren't the first person to try to use groups in compound matches
21:31 jaloren :)
21:31 LtLefse jaloren: could you make another nodegroup: 'slaveandmasternodes' or 'allnodes' etc
21:31 jaloren is there an issue rerquest for it? I'll upvote!!
21:32 jaloren LtLefse: yea i could. My root issue is trying to work around the fact that because of DNS and virtualization environment, i cannot know until deploy time what hostnames and ips are for the VMS
21:32 jaloren so i am having to jump through an enormous number of hoops to make the targetting work
21:32 jaloren its very maddening
21:33 LtLefse yeah
21:33 jaloren the amount of custom state, execution modules and wrapper scripts to autogenerate yaml configuration filies....
21:33 iggy how are you specifying the nodegroups?
21:33 jaloren its getting so ugly
21:33 LtLefse if it help, remember that while you can't use nodegroups in compound matches, you can use compound matches in nodegroups
21:34 jaloren iggy: the user  runs a python script and provides the the node hostnames on the command line
21:34 jaloren from there the nodegroups is generated and the salt master is restarted
21:34 iggy so it's a list?
21:34 jaloren iggy: for one nodegroup
21:35 jaloren iggy: there are others
21:35 iggy group3:
21:35 iggy group3: 'not L@node1,node2,node3'
21:35 iggy ?
21:35 jaloren opposite
21:35 jaloren group 1 or group 2 or group 3
21:36 jaloren iggy: i am leaning towards scrapping any attempt to use the built in sub systems
21:36 jaloren and just write a custom salt client and runner
21:36 jaloren then i can just dynamically target on whatever critieria i define in a python function
21:36 iggy what is the eventual goal?
21:36 jaloren well
21:37 jaloren i am writing a test harness to deploy tests on a distributed system
21:37 jaloren this involves using salt to create said distributed system
21:37 Singularo joined #salt
21:37 jaloren the "state" then requires complex dependencies ACROSS nodes
21:37 iggy could you set a grain and target based on that?
21:37 jaloren so of course i want look at the salt-run orchestrate and tyring ot make that work
21:38 jaloren iggy: no roles are defined solely within the distributed system itself....there's nothing about the OS itself that would tell me what role it should be
21:38 iggy roles:\n  - slavenode  .... roles:\n  - masternode
21:38 jaloren iggy: orchestrate almost works but trying to target dynamically based on every changing hostnames is a nightmare
21:38 iggy salt -C 'G@roles:slavenode or G@roles:masternode' test.ping
21:39 jaloren iggy: right but that takes me one step back which is how to "define" thos roles
21:39 iggy okay, that sucks
21:39 iggy maybe ansible/chef/puppet would work better
21:39 jaloren yuck
21:39 jaloren i've used puppet
21:39 jaloren that would be worse
21:39 jaloren i don't care for chef
21:40 jaloren i could play around with ansible though
21:40 jaloren not sure how much that would help me -- not terribly familiar with wit
21:40 jaloren the problem is the jinja templating
21:40 jaloren if i could define variables when the command is executed
21:40 jaloren which are passed in via the command line
21:40 jaloren then i could work something out
21:41 iggy you can pass pillar data on the command line if that's what you mean
21:41 jaloren iggy: yes.....i can and that took me only so far....
21:41 iggy well, for some states
21:41 baweaver joined #salt
21:41 jaloren if the file run by orhcestrate in turn calls another sls, that sls does not have access to that data
21:41 jaloren and that is in fact exactly what i need to do
21:42 jaloren whereas if i follow the zen of python....i'd just write a python script using the available python apis
21:42 jaloren and simulate the orchestrate functionality
21:42 jaloren cause that would be much simpler
21:42 jaloren i just hate reinventing the wheel
21:43 murrdoc what are u trying to do
21:43 murrdoc dont hate, participate
21:44 iggy I want to say there's an example of passing pillar data through an orchestrate job
21:44 sym-jp joined #salt
21:45 jaloren iggy: there is
21:45 johanek joined #salt
21:45 jaloren iggy: this works salt-run state.orch esg.cluster  pillar="{'masternode':'esg-1.obt.data'}"
21:46 jaloren the problem is that when "esg.cluster" is called it in turn calls a different state which will NOT have access to that pillar vairasble
21:46 jaloren variable
21:46 iggy pillar-ception
21:46 jaloren iggy: pardon?
21:46 otter768 joined #salt
21:47 murrdoc u dont need it to be a pillar
21:47 iggy seen the movie inception?
21:47 murrdoc use conf.get instead of pillar.get in your state
21:47 jaloren murrdoc: i am trying to use salt to construct a distributed system in which the state one one node is dependent on another AND hostnames and IPs will be dynamic and change
21:47 murrdoc conf.get will use __opts__
21:47 jaloren murrdoc: conf.get?>
21:47 murrdoc yeah man
21:47 murrdoc let me show u the light
21:47 murrdoc (i could be totes wrong)
21:48 iggy totes
21:48 murrdoc there is a module that iterates over opt, pillar, grain
21:48 murrdoc and now i have forgotten the module name
21:49 iggy config.get
21:49 murrdoc config!
21:49 murrdoc that might work for u jaloren
21:50 nich0s joined #salt
21:51 jaloren http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.config.html#module-salt.modules.config
21:51 jaloren this module?
21:52 jaloren Attempt to retrieve the named value from the minion config file, pillar, grains or the master config.
21:52 murrdoc damnit i was wrong
21:52 murrdoc sorry
21:52 jaloren how does that help me with being able to pass this in via the command line?
21:52 murrdoc try it out tho code might be different from docs
21:52 jaloren if i could just pass an arbitgrary dict on the command line and then have every state etc pick that up
21:52 jaloren i'd be set
21:52 jaloren is there anyway to do that?
21:53 enarciso joined #salt
21:55 murrdoc jaloren:  i think the config.get might work, cos __opts__ (http://docs.saltstack.com/en/latest/topics/development/dunder_dictionaries.html#opts) might pull in cli too
21:55 murrdoc but that might be based on the daemon u use at cli
21:55 ndrei joined #salt
21:56 jaloren murrdoc: so ...insert dict items into the __opts__ dictionary?
21:56 jaloren ohhh
21:56 jaloren i see what you are saying
21:56 jaloren hmmm
21:56 jaloren let me test that
22:12 aquassaut joined #salt
22:16 jespada joined #salt
22:23 fxhp joined #salt
22:26 bhosmer joined #salt
22:26 yomilk_ joined #salt
22:28 johanek joined #salt
22:30 JPaul trying to use salt-ssh, setup the roster file, quadruple checked the syntax I used in it to the file provided by the salt-ssh package and online, but it still comes back saying it find any hosts, even when I use a straight '*' glob
22:30 JPaul \any ideas what might be the issue?
22:31 iggy if "salt '*' test.ping" doesn't return anything, then you don't have any keys accepted
22:31 JPaul salt-ssh
22:31 JPaul not using salt for these servers in particular, yet
22:31 iggy oh right
22:32 * iggy backs away slowly
22:33 linjan joined #salt
22:34 bhosmer_ joined #salt
22:34 JPaul I've used salt-ssh before just fine, but that was early last fall, but I haven't used salt-ssh on this current master until just today
22:35 zwi joined #salt
22:37 JPaul hum, I just remembered this server doesn have ssh access to anything. I wonder if that is what is causing it...
22:39 dooshtuRabbit joined #salt
22:40 baweaver joined #salt
22:42 zwi1 joined #salt
22:42 jespada joined #salt
22:47 jcsp joined #salt
22:47 jcsp joined #salt
22:48 jcsp joined #salt
22:50 Vynce joined #salt
22:55 zwi joined #salt
22:57 zwi1 joined #salt
22:59 KyleG joined #salt
22:59 KyleG joined #salt
23:00 jmastron joined #salt
23:00 teall joined #salt
23:03 bfoxwell joined #salt
23:05 cloudyhead joined #salt
23:06 jmastron Hi all, got a bit of a weird question:  Is there an elegant way to rebuild the minion cache on the master?  I've got a bunch of minions that I've changed the minion id on (from hostname to fqdn), and they're now showing up as both names in 'salt-run manage.present'.
23:06 cloudyhead joined #salt
23:09 ajw0100 joined #salt
23:10 iggy jmastron: do the old names show up in manage.down?
23:11 jmastron Nope.  They're not in manage.up either.
23:11 iggy so what are you trying to clean up?
23:13 jmastron The output of manage.present, ideally.  I'd like to be able to have the cached version of the connected clients for scripting that doesn't have to wait on a client response every time.
23:13 dude051 joined #salt
23:15 iggy I can't find any masters that show anything weird there
23:15 iggy guess that means I'm reading code :/
23:15 jmastron From code-diving, it appears manage.present is checking each cache file for its cached IP and seeing if that IP is active, which would explain why it gets confused by stale caches.  I'm sure this is an edge case, just wondering if there's a clean way to rebuild those caches.
23:15 jmastron Haha
23:15 markm joined #salt
23:15 dimeshake what the heck does manage.present do differently than manage.up?
23:15 dimeshake it's not clear from the docs
23:16 dimeshake also, strangely, managed.present only shows mininos that are NOT hosted at AWS, while the majority of my minions *are* aws, as is the saltmaster
23:17 jmastron It's supposed to be a cached version that accesses the presence engine on the master rather than actually running a ping, which means it returns much faster.
23:17 iggy it would appear one goes by open connections and the other actually sends a test.ping or something
23:17 iggy I suppose you've restarted the master?
23:19 jmastron Yep.  Also tried deleting the /var/caches/salt/master/minions dir and it doesn't seem to rebuild.
23:20 _JZ_ joined #salt
23:23 jespada joined #salt
23:26 Diaoul joined #salt
23:30 ninkotech joined #salt
23:32 baweaver joined #salt
23:32 zwi joined #salt
23:35 \ask joined #salt
23:38 emostar joined #salt
23:39 baweaver joined #salt
23:40 yomilk joined #salt
23:42 yomilk_ joined #salt
23:42 zwi1 joined #salt
23:42 drawsmcgraw joined #salt
23:47 otter768 joined #salt

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