Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-11-17

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

All times shown according to UTC.

Time Nick Message
00:01 twork what's the right way to do an override of something from a formula?
00:03 hightekvagabond joined #salt
00:03 twork for example: the bind formula installs '/var/log/bind9/' root:bind 755, so bind dies when it can't write there. in the past i've hacked the formula, but that'll get lost when the formula upgrades.
00:04 twork i swear i've RTFLed.
00:09 * whytewolf shrugs. I don't use formulas so i never have that problem
00:10 whytewolf but if i was going to use formulas i would fork the repo for the formula, split a branch off for my own changes. and check that out into the master
00:10 twork i did that my first time around
00:10 twork but, i get the impression that's Not The Right Way.
00:11 Nazca__ joined #salt
00:11 whytewolf generally I have seen it recomended that you do not use the upstream repo directly.
00:11 fgimian joined #salt
00:12 twork what's the difference between that, vs. copying it?
00:13 twork i.e., download to /wherever, [adjust], copy to /srv/[wherever]?
00:13 Xevian joined #salt
00:14 whytewolf well, by keeping my changes in a repo i could make sure that upstream changes are filtered through. also all the hard work is handled in git.
00:14 whytewolf not to mention I don't lose them
00:15 twork ok, well, that is more or less what i do (as i struggle to learn git, but that's a different ball)
00:15 twork but,
00:16 whytewolf download a tool like sourcetree. or something to help you manage your repos.
00:16 twork sourcetree eh? i go look.
00:17 whytewolf there are a lot of tools that is just the first one i thought of. basicly find tools that make work easier
00:18 Karunamon joined #salt
00:18 fusionx86 joined #salt
00:19 whytewolf but in general to your question. find what works for you. the only truely wrong way to do work in salt is the one that doesn't work
00:20 twork well, yeah, the ones i've been trying... i'm awfully good at finding the obvious error the doc writer didn't think to point out
00:20 twork but anyhow
00:21 twork if changing the formula to put in place is more common that writing an overlay, i'm back on my way
00:22 ventris joined #salt
00:23 bfoxwell joined #salt
00:24 sjohnsen joined #salt
00:29 pbx joined #salt
00:34 ange joined #salt
00:44 otter768 joined #salt
00:51 kitplummer joined #salt
00:54 zmalone joined #salt
01:08 fsteinel_ joined #salt
01:09 Aleks3Y joined #salt
01:14 baweaver joined #salt
01:19 kitplumm_ joined #salt
01:20 HappySlappy joined #salt
01:42 Robert____ joined #salt
01:42 kitplummer joined #salt
01:45 Robert____ I use masterless salt-call, how can it watch file changes?  anyway remember the states ?
01:45 otter768 joined #salt
01:46 Robert____ I use http download the sls states files and then use salt-call to execute. But this situation, how salt-call watch the file changes?
01:57 Knuta Robert____: shouldn't make a difference? unless you're watching the state files themselves? As long as you download the files to a staging area somewhere and then have salt copy them from there, you should be able to watch for file changes.
01:58 HappySlappy joined #salt
02:01 Robert____ for example, salt-call is using SLS files for making sure the configurations. The httpd.conf is changed. Next time salt-call could not know it was changed.
02:02 Robert____ salt-call is executing and exit, so salt-call does not know httpd.conf was changed.
02:02 oida joined #salt
02:02 whytewolf Robert____: the functionality you are discribing doens't happen in salt with a master either.
02:03 whytewolf you can setup a schedule to check that the file stays the same
02:04 Robert____ @whytewolf,   I am sure there is a way, the watch keyword in salt
02:04 whytewolf the watch keyword watchs for state changes. not file changes
02:05 Vynce joined #salt
02:06 whytewolf the sls files normally come with a file that states how the file should look. and your sls files makes it happen. it doens't care how it looked before
02:07 baweaver joined #salt
02:08 whytewolf when you watch on a file. you are watching the file module in a state id with the name you give it. when that state runs and finds that a change is needed it flags the watch
02:08 MindDrive joined #salt
02:08 whytewolf master/masterless doesn't matter in that case
02:11 whytewolf now. there is a feature called beacons. that does require a master. but that is because it fires an event. and you have to setup reactors on that event.
02:12 Robert____ but watch watches file changes and restart the service ? confusing
02:14 whytewolf no, - watch ONLY watches state changes.
02:15 malinoff joined #salt
02:16 whytewolf the file state NEEDs to run in order for watch to know if there are changes
02:18 Robert____ if a depends on b and c, b and c are both changed, how can we make sure a is restarting only once ?
02:22 whytewolf Robert____: because you put both items under a single - watch in a single service.running state
02:22 whytewolf it will restart if either changes. or if both change.
02:22 whytewolf and only restart once
02:23 whytewolf Robert____: here is a quick example i tossed together https://gist.github.com/whytewolf/98e61640f39469e9539e
02:24 whytewolf notice how you are watching the state id, and not the file name. also how there are 2 files under the watch.
02:25 whytewolf now everytime a salt-call state.highstate is ran it will check both of the files. make changes as needed. and if either/both of the files change restart the service
02:26 HappySlappy joined #salt
02:28 hasues joined #salt
02:28 tristianc_ joined #salt
02:30 larsfronius joined #salt
02:36 Robert____ cannot open the https page
02:38 hasues left #salt
02:39 whytewolf you can't get to gist?
02:40 Robert____ too weak key,  ssl_error_weak_server_cert_key
02:40 Reinhart joined #salt
02:41 catpiggest joined #salt
02:41 whytewolf ...
02:41 ageorgop joined #salt
02:46 hightekvagabond joined #salt
02:51 TTimoT joined #salt
02:55 HappySlappy joined #salt
02:56 zmalone joined #salt
02:56 favadi joined #salt
02:58 quasiben joined #salt
03:03 sterfield joined #salt
03:04 sterfield left #salt
03:04 sterfield joined #salt
03:05 sterfield hi there
03:07 sterfield I'm coming from Ansible, and I'm testing salt for a few days now. I'm trying to do something: I want to extract the branch and the current git rev number from a repository located on the master and propagate those information in every minion's pillars
03:08 sterfield so basically, running a command on the master, extract the information and put it in pillar available to '*'
03:08 sterfield is it something that can be done in Salt ?
03:09 sterfield in Ansible, I was "delegating" to 127.0.0.1 for that, but there's no such things in Salt
03:09 phx sterfield, sure you can
03:09 phx sterfield, you can write pillars in quite a few ways, like using the python renderrer
03:10 phx that way, you can check that repo out, and return the data you need
03:11 sterfield hmmm I'm not sure to follow you on this...
03:12 phx basically i would do a python pillar, use some python module to connect to the repo, extract the needed info, then return it in a nicely structured way
03:15 sterfield okay, so let me rephrase this, to see if I understood it correctly: I create a pillar file, that contains some python code, that will do some actions locally (like subprocess)
03:15 sterfield and load this pillar file in the '*' target of the top file
03:16 phx yup
03:16 phx https://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.py.html
03:16 sterfield so each target will trigger the pillar, execute the pillar code that will be run locally (on the master so), and get the information
03:16 jhauser joined #salt
03:16 phx exactly
03:17 sterfield ok, that's tricky but I understand it
03:17 phx actually it's very easy to do
03:18 sterfield at first, I was trying to do it using salt['cmd.run'] but that would have been run on every target, not on the master
03:18 sterfield phx: yeah, the python code is fairly simple, I agree
03:18 ksalman joined #salt
03:19 sterfield it's just me, I've to move myself out of the 'Ansible' mindset
03:19 sterfield currently, I'm trying to replicate my Ansible playbooks
03:20 phx every and each config management requires a different approach, don't worry about it
03:20 phx focus on the functionality, and not on replicating the implementation
03:20 sterfield yes, I definitely have to do this
03:20 sterfield thanks for the help
03:20 sterfield much appreciated
03:21 sizeoftank joined #salt
03:22 phx you're welcome :)
03:23 oherrala joined #salt
03:24 clintberry joined #salt
03:26 AbyssOne joined #salt
03:26 mitsuhiko joined #salt
03:26 LotR joined #salt
03:26 NachoDuck joined #salt
03:27 bhosmer joined #salt
03:28 davedash joined #salt
03:30 KajiMaster joined #salt
03:31 nahamu joined #salt
03:32 schinken joined #salt
03:33 geekatcmu joined #salt
03:34 peters-tx joined #salt
03:34 kitplummer joined #salt
03:40 evle joined #salt
03:57 sterfield phx: done
03:57 sterfield that was indeed easy
04:03 phx told you :)
04:04 phx what i actually used the py renderrer in pilars was interfacing salt with my inventory database
04:04 phx so i generated the top.sls in a pillar, and extracted the highlevel configuration from the inventory, which then was the input for state renditions
04:06 sterfield ok
04:07 sterfield we are planning to do this through etcd, so apparently, there's already something for this
04:07 sterfield but the python renderer is really neat
04:11 otter768 joined #salt
04:13 cmclaughlin joined #salt
04:17 Reinhart joined #salt
04:18 clintber_ joined #salt
04:18 drig0r joined #salt
04:19 _JZ_ joined #salt
04:21 ashutoshn joined #salt
04:23 jaybocc2 joined #salt
04:26 ramteid joined #salt
04:30 clintberry joined #salt
04:34 kaji_ joined #salt
04:39 Kaji_ joined #salt
04:44 robertchen joined #salt
04:50 boargod2 joined #salt
04:52 RandyT joined #salt
04:53 bfoxwell joined #salt
04:53 sjorge_be joined #salt
04:54 mikepea_ joined #salt
04:54 joren joined #salt
04:55 yuhlw joined #salt
04:55 Arendtse1 joined #salt
04:55 mortis_ joined #salt
04:55 morsik_ joined #salt
04:55 baffle joined #salt
04:56 eofs_ joined #salt
04:56 davroman1ak joined #salt
04:56 pdx6_ joined #salt
04:56 Rebus_ joined #salt
04:56 jcristau_ joined #salt
04:56 shnguyen joined #salt
04:56 ablemann_ joined #salt
04:56 btorch_ joined #salt
04:56 Laogeodritt| joined #salt
04:56 adelcast1 joined #salt
04:57 rideh- joined #salt
04:57 gchao_ joined #salt
04:57 honestly_ joined #salt
04:57 nlb_ joined #salt
04:57 rdas joined #salt
04:58 TOoSmOotH_ joined #salt
04:58 dlloyd_ joined #salt
04:59 kidneb_ joined #salt
05:00 baoboa joined #salt
05:00 Ssquidly_ joined #salt
05:00 [vaelen] joined #salt
05:00 ronrib joined #salt
05:00 cromagnon_man joined #salt
05:00 JPaul joined #salt
05:00 techdragon joined #salt
05:01 moeyebus_ joined #salt
05:01 keekz joined #salt
05:01 FreeSpencer joined #salt
05:01 FreeSpencer joined #salt
05:01 kbyrne joined #salt
05:01 Craighton joined #salt
05:01 Craighton joined #salt
05:02 bharper joined #salt
05:02 mrtrosen joined #salt
05:02 douardda joined #salt
05:02 MK_FG joined #salt
05:02 virusuy joined #salt
05:02 Reinhart joined #salt
05:02 davisj joined #salt
05:02 GrueMaster joined #salt
05:03 kavakava joined #salt
05:03 serverascode joined #salt
05:04 hajhatten joined #salt
05:04 g3cko joined #salt
05:04 cwyse_ joined #salt
05:06 THE_BOULDER joined #salt
05:08 eliasp joined #salt
05:11 SWAT joined #salt
05:16 robertchen if I use masterless, agentless, just salt-call, can I do salt schedule ?
05:19 denys joined #salt
05:20 zemm_ joined #salt
05:20 morsik joined #salt
05:20 David_B55__ joined #salt
05:20 cswang_ joined #salt
05:20 ]V[_ joined #salt
05:21 charli joined #salt
05:21 oherrala1 joined #salt
05:21 cmclaughlin joined #salt
05:24 rodr1c_ joined #salt
05:24 ashutoshn robertchen: does the scheduler solve problem for you - https://docs.saltstack.com/en/latest/topics/jobs/schedule.html ?
05:25 Twiglet joined #salt
05:26 ashutoshn robertchen: and you can have schedule defined in minion's pillar or inside minion's config file
05:26 seblu joined #salt
05:26 gtmanfred joined #salt
05:26 robertchen I do not want to use minion, there is no agent.  Just salt-call
05:27 ageorgop joined #salt
05:27 akoumjian joined #salt
05:27 robertchen by the way, how to connect to this IRC  with irc client ?
05:27 dean joined #salt
05:28 bhosmer_ joined #salt
05:28 chutzpah joined #salt
05:28 Rockj joined #salt
05:29 czchen joined #salt
05:29 favadi joined #salt
05:29 Phtes joined #salt
05:30 kutenai joined #salt
05:30 kalessin joined #salt
05:30 TomJepp joined #salt
05:30 paolo joined #salt
05:30 davedash joined #salt
05:30 Aleks3Y joined #salt
05:30 hillna joined #salt
05:30 paolo joined #salt
05:31 Nazzy joined #salt
05:31 dijit joined #salt
05:31 Ryan_Lane joined #salt
05:33 mission712 joined #salt
05:35 kitplummer joined #salt
05:37 impi joined #salt
05:37 __number5__ robertchen: use your irc client connect to one of freenode's server https://freenode.net/irc_servers.shtml or just chat.freenode.net, then join #salt channel
05:39 Bryson joined #salt
05:40 terratoma joined #salt
05:43 Vynce joined #salt
05:45 robertchen my customer need to get the status for each step in real time, does salt support this?
05:47 Reinhart joined #salt
05:49 ashutoshn robertchen: what steps are you talking about ?
05:56 Reinhart joined #salt
06:03 indispeq joined #salt
06:04 calvinh joined #salt
06:04 jimklo joined #salt
06:12 otter768 joined #salt
06:21 indispeq joined #salt
06:30 iggy no and no
06:34 av_ joined #salt
06:37 RandyT joined #salt
06:56 nethershaw joined #salt
06:58 viq joined #salt
07:03 RobertChen joined #salt
07:03 KingJ joined #salt
07:04 RobertChen can the returner return result for each step in real time? like each cmd.run ?
07:05 davromaniak joined #salt
07:06 nebuchadnezzar joined #salt
07:07 ajw0100 joined #salt
07:16 RobertChen can the returner return result for each step in real time? like each cmd.run ?
07:20 Ssquidly joined #salt
07:26 jaybocc2 joined #salt
07:35 Vynce joined #salt
07:35 felskrone joined #salt
07:36 mohan_ joined #salt
07:36 felskrone joined #salt
07:36 felskrone joined #salt
07:36 mohan_ hi, how to create a database in ms-sql server using modules.mssql
07:37 felskrone joined #salt
07:39 boargod joined #salt
07:40 felskrone joined #salt
07:41 KermitTheFragger joined #salt
07:43 jaybocc2 joined #salt
07:52 jaybocc2 joined #salt
07:52 Ssquidly joined #salt
07:52 Ssquidly joined #salt
07:53 Ssquidly joined #salt
07:54 Ssquidly joined #salt
07:55 impi joined #salt
07:56 Ssquidly joined #salt
07:58 Ssquidly joined #salt
07:58 rotbeard joined #salt
08:00 voileux joined #salt
08:01 jaybocc2 joined #salt
08:02 rotbart joined #salt
08:08 booooo joined #salt
08:08 booooo Hello.
08:08 RandyT joined #salt
08:08 slav0nic joined #salt
08:10 iggy woo, finally did some formulas work
08:11 ramteid joined #salt
08:12 elsmo joined #salt
08:13 otter768 joined #salt
08:15 malinoff joined #salt
08:16 ITChap joined #salt
08:17 eseyman joined #salt
08:23 dkrae joined #salt
08:24 auzty joined #salt
08:25 Reinhart joined #salt
08:25 kawa2014 joined #salt
08:28 ramteid joined #salt
08:29 Andre_ks Hi all
08:29 Andre_ks just getting into Salt now.  have a couple questions
08:30 Andre_ks I use puppet and chef.  So typically the pull method happens on an interval
08:31 Andre_ks I just built master salt into /srv/salt with jenkins from git...
08:31 Andre_ks I assume it doesn't pull by default...
08:32 iggy if you use gitfs it does
08:32 Guest71482 joined #salt
08:32 iggy and there's a git ext_pillar for putting your pillar data into git as well
08:32 Andre_ks have some environments and states configured with pkg.installed's  - just asking for some guidance url's for where to read next..
08:33 larsfronius joined #salt
08:33 Andre_ks Not using gitfs
08:33 Andre_ks using jenkins to build
08:33 iggy are you talking about the minions pulling from the master?
08:33 Andre_ks git > jenkins > saltstack master > minions
08:33 Andre_ks yep
08:34 Andre_ks that
08:34 mehakkahlon joined #salt
08:34 iggy they don't do that automatically/periodically
08:34 Andre_ks ok, thought so
08:34 Andre_ks I don't actually want them too
08:34 iggy they run whatever you tell them to (i.e. a highstate, etc.)
08:34 Andre_ks I want to use another method
08:34 Andre_ks oh cool
08:35 Andre_ks iggy: so highstate is what I am looking for
08:35 Andre_ks thanks
08:35 Andre_ks pretty sure...
08:36 iggy a highstate tells the minion to put itself into the described state
08:36 iggy there are other things you can do (run single states, run orchestrates, etc)
08:37 mohan_ joined #salt
08:38 mohan_ hi, how to create a database in mssql using the modules.mssql
08:40 Andre_ks thanks iggy
08:40 Andre_ks will read up on states
08:40 Andre_ks does /srv/salt require specific ownerships on the master?
08:41 Andre_ks not sure if that's an issue...
08:41 Andre_ks I get this error:  No Top file or external nodes data matches found.
08:42 Andre_ks the top.sls has this    development:   'i3-d-salt-linux-test*':     - role_based_auth
08:42 Andre_ks which matches a node
08:42 Andre_ks I mean, minion :p
08:43 Andre_ks top.sls is inside \/srv\/salt\/top.sls
08:44 iggy that should work by default
08:46 informant1 joined #salt
08:46 Andre_ks I think it might be a directory structure problem?
08:46 Andre_ks I setup different environment directories
08:46 thalleralexander joined #salt
08:47 Andre_ks like this http://www.saltstat.es/posts/environment-based-clusters.html
08:47 iggy I'm honestly not terribly familiar with salt's environments... except that they suck and I will probably never use them
08:48 Andre_ks whoa
08:49 Andre_ks how do you do dev, staging, q/a and prod CI deployment ?
08:49 iggy separate masters
08:49 Andre_ks hmmm...
08:49 Andre_ks I was considering that
08:51 iggy I honestly would never want the same people that can do stuff in dev/qa/staging to be able to do those same things in prod (even if by accident)
08:51 Andre_ks I probably just have some config setting or dirs hierarchy wrong in there
08:52 s_kunk joined #salt
08:52 Andre_ks iggy: I agree in general with you
08:53 ramteid joined #salt
08:54 Andre_ks but small companies like the one I work for typically are limited in staff and the same people take on numerous or all roles
08:54 Andre_ks i would like to set it up so that as the company expands, it's isolated, but time is limited
08:54 iggy I've worked for small companies... still wouldn't let a dev anywhere near prod
08:54 keimlink joined #salt
08:59 Andre_ks righton
08:59 malinoff Andre_ks: another option is to secure the master and use jenkins project-based matrix security politics to disallow developers to even discover the production-deployment job
08:59 malinoff we had such setup like 2.5 years ago
09:00 Andre_ks malinoff: that sounds amazing
09:00 Andre_ks that's the plan in the longer run
09:01 Andre_ks right now I am doing POC on saltstack
09:01 sterfield joined #salt
09:02 malinoff (we couldn't upgrade a single salt master at the time, I can't imagine the maintenance burden if we had 5 masters * n projects)
09:02 jaybocc2 joined #salt
09:02 Andre_ks I am sure it will be fine, but would like to setup environments...?
09:04 Andre_ks do I have to assign the environment to the minion?
09:04 sterfield hi again
09:04 Andre_ks before running high.state?
09:05 malinoff Andre_ks: use pillars for that
09:05 Rumbles joined #salt
09:05 Andre_ks ah ok, will move on to read up on pillars then... thanks malinoff!
09:06 sterfield I have another small question: I figured out (with phx help, thanks!) how to retrieve information stored locally on the master. I've developed a small pillar file with the python renderer, that get to a git repository and extract the git rev and branch
09:07 sterfield so far so good. I put the pillar file in the top file, '*' target.
09:07 sterfield however, it'll call the same python script again and again, for each minion. It works, but it's not very pretty
09:08 sterfield so is it possible to create a "script" that generates the pillar file at the start of a run ? A sort of "start callback" in Salt ?
09:10 iggy it shouldn't be terribly difficult for salt to run a py rendered file a number of times
09:11 sterfield iggy: I think the same
09:12 sterfield however, as the branch and the rev will not move during a run, it's an information that will be called a lot of times, for nothing
09:13 sterfield I'm kind of new to Salt (figured out the renderer usage few hours ago) so I was wondering if I was not missing something here
09:13 iggy nope, that's pretty much the gist of things
09:14 sterfield ok
09:14 iggy maybe cache the value in a global var
09:14 sterfield hmm
09:15 sterfield I read the O'reilly book about salt, read a good chunk of the documentation, and I still discovering some stuff every day
09:15 sterfield I'll check the global variable / cache stuff, thanks for the pointer
09:15 felskrone joined #salt
09:15 GreatSnoopy joined #salt
09:15 iggy I've been using salt for >2 years and I still do
09:15 chiui joined #salt
09:16 sterfield yeah
09:16 sterfield for my defence, I'm coming from Ansible which is waaaaay easier to learn / handle
09:16 sterfield however, salt architecture is better, by an order of magnitude
09:16 iggy you might also just look at general python stuff
09:16 larsfronius joined #salt
09:17 iggy for the global var/cache thing
09:17 malinoff sterfield: define 'better'. We moved from salt to ansible mainly because salt's architecture is bad :)
09:17 felskrone joined #salt
09:17 sterfield malinoff: interesting
09:18 ziro` joined #salt
09:19 sterfield well proxies is a must have, communication between master and minion is more efficient, you have no issues regarding SSH / accounts (try to login with a specific user on a server, then sudo root, but another server, the login procedure is different)
09:19 sterfield to name a few
09:20 thefish joined #salt
09:20 sterfield also, we have hard time to scale a lot with Ansible. Launching hundreds of SSH connections on the Ansible master is quite difficult
09:20 sterfield tbh, I haven't reached my 2000+ servers using salt yet, so I'll compare the speed from that
09:21 sterfield but with Ansible, it's a linear trend
09:21 sterfield (requested time to reach all the nodes, I mean)
09:23 sterfield Also, the returners is a nice feature in Salt. We hacked a bit Ansible to get a lot of information from our servers, and store them as an inventory. We did this by getting Ansible inventory in JSON back to the master
09:23 sterfield It's working, but it's not pretty (big JSON to parse, in order to update the inventory database if needed)
09:24 iggy one of the guys I work with came up with a nifty trick... `salt --out=json | jq '<stuff>'` is very awesome
09:24 sterfield malinoff: your turn :) Why do you think Salt architecture is "worse" than Ansible ?
09:25 malinoff sterfield: 1) proxies exist because there are systems that you can't install python into. Ansible has raw module which allows you to do basically everything without remote python.
09:25 malinoff sterfield: 2) ansible with openssh-client transport supports ProxyCommand set in the ssh config file just fine
09:26 malinoff sterfield: also, I've never faced issues with SSH, but faced A LOT of issues with master-minion communication
09:26 malinoff sterfield: 3) I've never had a need to deploy all my servers at once, so I can't say much about scaling
09:27 sterfield agreed on ProxyCommand, we are doing it also
09:27 malinoff we use AWS, Jenkins, AutoScaling groups, ELB etc etc and everything *just works*(tm)
09:27 larsfronius joined #salt
09:28 malinoff sterfield: 4) ansible has callback plugin system which you can use to put the output into a database
09:28 malinoff sterfield: but you're saying about features, not architecture
09:29 sterfield yeah you're right, I drifted a bit out of the infrastructure discussion
09:29 malinoff our issues with salt were about presence of salt-minion on the target servers, authentication between master and minions (i know about predefined keys and other methods, but all of them are hard to maintain properly)
09:29 bhosmer joined #salt
09:29 CRM114 Hi Guys, I have a problem where salt is failing to enable statsite, but then marking the statsite service as enabled and running. This would appear to be a bug in salt-stack. Has anyone else seen something like this?
09:30 malinoff sterfield: we couldn't upgrade our master and minions separately because salt team have broke backwards compatibility for literally every release
09:30 malinoff sterfield: but at the same time we couldn't upgrade all of them at once because, well, you cannot upgrade 1500+ servers at once and believe that everything will work properly
09:31 VSpike yeah, upgrades are painful :/
09:32 iggy CRM114: I would be quicker to blame the init script, there's an issue about init scripts that don't return correctly
09:32 malinoff sterfield: salt-master is a SPOF which is hard to reproduce in case of a failure (just imagine that you've lost all accepted keys) and to make highly available
09:32 iggy why hang out in a channel for a product you seem to dislike so much
09:34 sterfield malinoff: however I disagree about the reasons on proxies. I have a distributed architecture with several DCs and it's not possible sometimes to reach ALL of your server from one point
09:34 sterfield but you can deal with that with SSH bastion
09:34 losh joined #salt
09:34 sterfield (for Ansible)
09:34 malinoff iggy: first of all, I don't "dislike" salt. I appreciate its help to deploy our services. But we have outgrown it by a number of reasons, and sterfield asked for them
09:35 sterfield Wait to have a script / command that will ask for a user prompt, that will stuck your ssh connection, then stuck your run forever
09:36 malinoff sterfield: what? salt won't handle this either
09:36 sterfield however, this may be the case also for salt
09:36 sterfield ninka
09:36 sterfield too fast
09:36 malinoff sterfield: just use expect in cases like that :)
09:38 sterfield as for the SPOF of master, I don't see it. The amount of data to backup is fairly small
09:38 EvaSDK iggy: PR updated :)
09:38 malinoff sterfield: yes, but with ansible you don't need to backup anything, everything is already in a git repository
09:40 gwmngilfen joined #salt
09:40 sterfield well, one folder full of keys and one folder in /etc is not a big deal for us, isn't it ? :)
09:41 malinoff sterfield: as I said, it's not a big deal :) but you have to think about it, you have to setup and verify your backups, while with ansible you simply don't care
09:42 Jimlad joined #salt
09:46 VSpike My biggest pain point with Salt is Windows. I wish we didn't have to deal with Windows, I absolutely hate it with a passion :/
09:49 malinoff windows is pain regardless of toolchain used
09:49 hemebond joined #salt
09:50 CRM114 iggy: thanks I will have a look.
09:51 AlberTUX1 joined #salt
09:52 VSpike malinoff: true dat
09:55 denys joined #salt
09:57 elsmo joined #salt
09:58 VSpike I saw Ansible has added a lot of Windows stuff since I last looked at it about a year and a half ago but I don't know how mature or widely used it is
09:59 ashutoshn left #salt
10:02 jaybocc2 joined #salt
10:04 fredvd joined #salt
10:07 Grokzen joined #salt
10:08 AlberTUX2 joined #salt
10:11 TTimoT joined #salt
10:14 otter768 joined #salt
10:18 seblu joined #salt
10:21 ggoZ joined #salt
10:28 ziro` joined #salt
10:32 hojgaard joined #salt
10:32 bluenemo joined #salt
10:32 hojgaard I am having trouble targeting minions using nodegroups
10:34 hojgaard My master conf looks like this:
10:34 hojgaard http://pastebin.com/QWxCCaPN
10:34 hojgaard No problem targeting serv and prod with salt -N prod/serv test.ping
10:35 hojgaard but for intl_web i just get: No minions matched the target. No command was sent, no jid was assigned.
10:35 hemebond hojgaard: Don't you need an "and" or an "or" in there?
10:35 hemebond e.g., 'N@serv or N@prod'
10:38 babilen definitely 'or'
10:39 hojgaard i've tried and
10:39 hojgaard i will try or
10:39 babilen Well, do you want the intersection or union?
10:40 hojgaard babilen i want them all. There will be no intersection
10:40 babilen Then you want "or" rather than "and"
10:41 babilen "A and B" is only true if both A and B are true ..
10:41 hojgaard Ahh ok, now i understand
10:41 hojgaard cool
10:42 hojgaard works perfect
10:42 hojgaard thank you
10:44 babilen One way to think about it is: What do I want to be true when I target my minions? The "I want minions from group A and *and* B" turn into the expression "A or B" rather than "A and B" as the latter would only be true for minions that are members of both A and B at the same time
10:47 sjorge joined #salt
10:47 sjorge joined #salt
10:48 moeyebus_ joined #salt
10:49 giantlock joined #salt
10:53 zerthimon joined #salt
10:57 N-Mi_ joined #salt
11:00 elsmo joined #salt
11:00 amcorreia joined #salt
11:03 jaybocc2 joined #salt
11:20 CRM114 joined #salt
11:23 CRM114 iggy: we aren't using init scripts... so it can't be that... we are using systemd. It seems to return correctly (error code 3 ).. see line 3 of https://gist.github.com/callumski/8509e63b00cd5287d8db
11:23 CRM114 iggy: could this be the problem? https://github.com/saltstack/salt/blob/develop/salt/modules/systemd.py#L524
11:30 bhosmer joined #salt
11:31 dgn joined #salt
11:41 eigart joined #salt
11:42 eigart hi !
11:42 eigart i'm trying to use gifs_remotes
11:42 eigart with pygit2
11:42 abednarik joined #salt
11:42 eigart and i don't understand why it requires to compile something, as I already installed the python-pygit2 package
11:43 eigart the output => http://pastebin.com/Z82SeyvX
11:45 quasiben joined #salt
11:45 Reinhart joined #salt
11:48 AlberTUX1 joined #salt
11:48 colegatron joined #salt
11:49 sardok joined #salt
11:50 CeBe joined #salt
11:52 sardok Hi everyone! I am using gem state to install particular gems, i need to find the directory of these gems, i tried to assign output of a finder command to a variable but this didn't work as jinja templates are rendered before sls files are executed. Any idea how to achieve that?
12:04 VSpike Does utahdave ever come by here any more?
12:04 jaybocc2 joined #salt
12:05 mehakkah_ joined #salt
12:07 Hydrosine Can anyone take a look at this? its probably an formatting error.. but can't see why
12:07 Hydrosine https://gist.github.com/hydrosine/b0e2da275e8562b5f353
12:08 Hydrosine it only fails when the if matches
12:08 kawa2014 joined #salt
12:11 Norrland Hydrosine: spaces on L16-20?
12:12 VSpike Hydrosine: you need to outdent line 16
12:12 Norrland Hydrosine: try: "opensips:" instead of "  opensips:"
12:12 VSpike or yeah, that block as Norrland says
12:14 Hydrosine that seems to do it :) thanks
12:14 abednarik joined #salt
12:15 otter768 joined #salt
12:17 abednarik joined #salt
12:22 eseyman joined #salt
12:23 favadi joined #salt
12:24 armyriad joined #salt
12:25 _JZ_ joined #salt
12:26 armyriad joined #salt
12:26 Erik_P joined #salt
12:28 Erik_P Hi all need help with chocolatey module to install a pkg from salt://XXXXX
12:28 _JZ_ joined #salt
12:29 shiriru joined #salt
12:31 colegatron joined #salt
12:36 MadHatter42 joined #salt
12:36 mbrgm joined #salt
12:37 mbrgm hi! should pillar data be put under source control? how do you handle credentials in the pillar?
12:38 bash1235123 joined #salt
12:39 tristianc joined #salt
12:41 rodio_ua joined #salt
12:41 rodio_ua left #salt
12:42 malinoff joined #salt
12:43 impi joined #salt
12:43 bash1235123 Can somebody tell me why debian jessie has version 2015.8 in their repos and wheezy 2015.3 ? :D
12:44 abednarik joined #salt
12:44 eliasp bash1235123: well, because the package maintainers didn't upgrade the package…
12:45 eliasp bash1235123: as long as you're not using repo.saltstack.com, that's a question to ask your distribution
12:45 bash1235123 I am using that
12:46 guest joined #salt
12:46 guest Hello. Is it possible to trigger a salt state on another minion?
12:46 eliasp bash1235123: well, then it's because repo.saltstack.com offers only packages for jessie, so your wheezy falls back to what the distribution/other repos offer(s)…
12:46 bash1235123 deb http://debian.saltstack.com/debian wheezy-saltstack main
12:46 guest So, I'll run a salt state on minion A, and have it trigger a (hopefully asynchronous) different state on minion B
12:47 guest minion A and minion B being identified by grains
12:47 eliasp guest: use event reactors for this… https://docs.saltstack.com/en/latest/topics/reactor/
12:47 guest ty eliasp
12:49 eliasp bash1235123: for the packages provided on repo.saltstack.com, I'm not sure what's the problem with other repositories… the PPA used to be maintained by Joe Healy, but I don't know what's the status now with repo.saltstack.com…
12:51 eliasp bash1235123: you might wanna subscribe to https://github.com/saltstack/salt/issues/28674
12:51 saltstackbot [#28674]title: Deprecate Ubuntu PPA mechanism in favor of repo.saltstack.com | Latest proposed version in https://launchpad.net/~saltstack/+archive/ubuntu/salt is 2015.5.3....
12:52 bash1235123 thats for ubuntu if I'm not mistaken
12:53 eliasp bash1235123: than maybe this one fits better: https://github.com/saltstack/salt/issues/27385
12:54 saltstackbot [#27385]title: repo.saltstack.com contains only current latest release, should contain previous releases, too | I see that only 2015.8.0 (the current latest release) is available in repo.saltstack.com. I would like to roll back to 2015.5.3, or at least 2015.5.5. The repo should contain at least a few prior releases.
12:54 eliasp some days, I don't even want to start debugging a Salt issue… "salt-call" produces different pillar merge results than "salt minion …"
12:55 larsfronius joined #salt
12:56 tristianc_ joined #salt
12:59 abednarik joined #salt
13:01 morissette joined #salt
13:05 jaybocc2 joined #salt
13:05 mbrgm left #salt
13:27 mik__R joined #salt
13:27 larsfron_ joined #salt
13:29 gladiatr joined #salt
13:31 TTimoT joined #salt
13:31 bhosmer joined #salt
13:39 amcorreia joined #salt
13:40 ziro` joined #salt
13:40 denys joined #salt
13:41 bluenemo joined #salt
13:43 Reinhart joined #salt
13:44 karlthane joined #salt
13:47 dean joined #salt
13:49 diegows joined #salt
13:49 subsignal joined #salt
13:50 diegows hi
13:50 diegows look like cmd.run doesn't report the exit status of the commands
13:50 diegows is there a workaround?
13:50 diegows something better than cmd.run "/bin/false || echo FAIL" :)
13:51 pbx_ joined #salt
13:51 eliasp diegows: cmd.retcode
13:52 diegows oh, good... I checked docs and I haven't see it :)
13:52 diegows thanks
13:53 ggoZ joined #salt
13:58 subsigna_ joined #salt
14:02 racooper joined #salt
14:05 jaybocc2 joined #salt
14:07 guest Can I view the results of a job knowing only its ID?
14:07 guest My reactor is failing somewhere and it's proving extremely difficult to debug
14:07 guest I call from a state with this: salt-call event.fire_master '{"target":"G@role:mytmoapp"}' 'update_hostsfile'
14:08 guest my reactor looks like this http://paste.rory.sh/r7l9
14:09 LotR guest: salt-run jobs.lookup_jid <jid> gets you the output it gave
14:10 guest ah fantastic
14:11 guest OK so it's returning just blank for what I assume is the reactor job
14:12 guest Well the minion id and a colon, in teal, but no actual content
14:12 mapu joined #salt
14:12 tristianc_ joined #salt
14:13 guest oh I'm an idiot, I need to put my reactor in reactor.d
14:13 guest not just upload the sls file
14:15 otter768 joined #salt
14:16 quasiben joined #salt
14:17 kitplummer joined #salt
14:18 Tyrm joined #salt
14:21 fe92 joined #salt
14:22 Tecnico1931 joined #salt
14:24 ssplatt joined #salt
14:25 ssplatt hey guys, got a couple of questions about salt-cloud.  first, when i’m provisioning a vm, why do I keep getting “[WARNING ] There's no master defined on the VM settings.” and how do i fix that.  I tried putting master: blah in the profile but that didn’t help
14:25 quix joined #salt
14:25 ssplatt second, when it’s provisioning, it goes through and says “salt installed” but then i get onto the new vm and there is no salt installed.
14:26 ssplatt i’m provisioning these right off my testing env salt master too
14:27 tristianc joined #salt
14:27 winsalt joined #salt
14:31 bhosmer joined #salt
14:32 numkem joined #salt
14:32 scoates joined #salt
14:38 cpowell joined #salt
14:39 ssplatt i was under the impression that using salt-cloud to create a new VM would immediately tie it into my salt env for provisioning. is that not true?
14:40 masterkorp Hello everyone
14:40 racooper does anyone have working examples using the Netscaler execution module?
14:41 Brew joined #salt
14:41 pgoetz_ joined #salt
14:41 masterkorp So I am runing a state and get this "Clonficting ID 'route'
14:43 larsfronius joined #salt
14:44 TomJepp joined #salt
14:47 racooper did you define 'route' twice in your state? or is it pulling in another state with 'route' defined?
14:47 masterkorp i have no idea
14:47 dlloyd left #salt
14:49 tanta_g joined #salt
14:50 racooper your debug logs should help
14:50 racooper https://docs.saltstack.com/en/develop/topics/troubleshooting/index.html
14:50 masterkorp that is all they tell me
14:53 bastion1704 joined #salt
14:54 Reinhart joined #salt
14:57 kitplummer joined #salt
15:01 techblaze joined #salt
15:03 zmalone joined #salt
15:03 Aleks3Y joined #salt
15:03 timoguin_ joined #salt
15:04 16WAASC6Z joined #salt
15:05 Xevian_ joined #salt
15:06 jaybocc2 joined #salt
15:07 abednarik joined #salt
15:07 mik__R joined #salt
15:09 ssplatt finally found that i can put “minion: master: blah” in my profile to get rid of that warning, but i still don’t see salt installed on the vm.  do i _have_ to make a host map to get this all working?
15:09 nethershaw joined #salt
15:09 subsignal joined #salt
15:11 mik__R_ joined #salt
15:12 hasues joined #salt
15:13 hasues left #salt
15:13 tristianc joined #salt
15:14 ssplatt “When Salt Cloud creates an instance, by default it will install the Salt Minion onto the instance, along with any specified minion configuration, and automatically accept that minion's keys on the master”  yeah so that isn’t happening.  where can i see what’s going wrong?
15:14 clintberry joined #salt
15:16 ssplatt could be the provider i’m using?
15:16 andrew_v joined #salt
15:18 clintber_ joined #salt
15:20 winsalt you could try putting deploy: True in the cloud profile
15:31 bhosmer joined #salt
15:33 subsignal joined #salt
15:41 twork joined #salt
15:42 Heartsbane So I know I can pass in pillar variables into a state via the command line but I just had a coworker ask if we do it without pillar='{"URLS": ["derp", "deh", "duh"]}' for quick setup so not so much typing.
15:42 Heartsbane Never thought to ask until now
15:44 Nazca joined #salt
15:44 bhosmer joined #salt
15:45 rmnuvg joined #salt
15:45 ssplatt winsalt: i see in the bootstrap-salt.log that it thikns it’s running config_salt(), install_debian_restart_daemons(), and daemons_running(). but none of those things are actually happeneing from what i can see on the Vm after the fact
15:46 winsalt im troubleshooting some issues with salt-cloud as well, have you added "-l all" to the salt-cloud command?
15:47 ssplatt just found teh -l debug option, running now
15:51 jose___ joined #salt
15:51 jose___ howdy
15:52 jose___ I have a question are will still able to execute jinja logic inside an orchestrator state?
15:53 jose___ I want to use a jinja logic to randomly select a minion for the tgt:
15:55 tanta_g is there a way to call another salt state from within a state file
15:56 mpanetta joined #salt
16:01 jose___ you can use include:
16:01 jose___ include: account.users
16:02 jose___ will include that state, is you just provide account it will look for the init.sls and execute whatever is in there
16:02 jose___ https://docs.saltstack.com/en/develop/ref/states/include.html
16:02 quix_ joined #salt
16:02 colegatron joined #salt
16:04 ashutoshn joined #salt
16:04 viq joined #salt
16:04 diegows joined #salt
16:04 ssplatt winsalt: salt-cloud -u, to update the bootsrap script seems ot have done the trick...
16:07 jaybocc2 joined #salt
16:11 sk_0 joined #salt
16:11 kitplummer joined #salt
16:15 doriftoshoes joined #salt
16:16 chadhs joined #salt
16:16 otter768 joined #salt
16:17 mohae_ joined #salt
16:17 scooby2 anyone using salt-cloud with vmware? For some reason it hates my datastore
16:17 tanta_g I figured it out, just call module.wait, pass in the right args, simple
16:18 tanta_g except in this case the - name: 'state.sls' is the first argument passed to module.wait:
16:18 scooby2 http://pastebin.com/22jBfcMS datastore: is set to datastore6 but it keeps showing datastore-81
16:19 zsoftich2 joined #salt
16:20 jimklo joined #salt
16:22 kitplummer joined #salt
16:24 pbx_ joined #salt
16:24 wendall911 joined #salt
16:26 jaybocc2 joined #salt
16:27 clintberry joined #salt
16:33 Reinhart joined #salt
16:38 sdm24 joined #salt
16:39 timoguin joined #salt
16:39 jalbretsen joined #salt
16:40 sunkist joined #salt
16:41 timoguin_ joined #salt
16:44 malinoff joined #salt
16:50 bhosmer joined #salt
16:51 kitplumm_ joined #salt
16:52 quix joined #salt
16:55 kbaikov joined #salt
16:56 perfectsine joined #salt
16:57 conan_the_destro joined #salt
16:58 kitplummer joined #salt
17:02 StolenToast can someone please explain "date > /tmp/crontest" and superscript? https://docs.saltstack.com/en/latest/ref/states/all/salt.states.cron.html
17:02 StolenToast there seem to be a bunch of conflicting examples here
17:02 zmalone date and superscript are examples
17:03 zmalone that entire line is what you want used as the cronjob action
17:03 zmalone so, run the "date" command, which prints the date, and dump it into a plain text file named /tmp/crontest
17:03 StolenToast ok so it's just piping the output of a script called date into /tmp/crontest?
17:03 StolenToast ok cool
17:03 zmalone date isn't a script, it's a unix command
17:03 StolenToast thanks
17:04 zmalone that's a redirect of output, not a pipe, but yeah.
17:04 Bryson joined #salt
17:05 wendall911 joined #salt
17:05 StolenToast details smetails
17:06 kawa2014 joined #salt
17:08 openfly_ joined #salt
17:09 geekatcmu "I don't understand what 'date > /tmp/crontest' means" kind of takes away your right to wave off "details".
17:09 geekatcmu Perhaps "attention to detail" is something you should consider.
17:09 StolenToast sorry I was just trying to be light about it
17:09 geekatcmu Sorry, I guess I took that wrong, then.
17:10 geekatcmu I guess I've just spent too much time recently helping people who are too smart to RTFM.
17:12 StolenToast well in this case it was the manual that confused me
17:12 StolenToast I didn't know the name in the top file could be either the task itself OR a path to a script
17:15 geekatcmu Yeah, one of the things about salt which occasionally irks me is that the id of the state is used as the name of the state (if the name is not explicitly specified) and that the *name* of the state is actually "the thing being changed/referenced/wHATEver"
17:15 Fiber^ joined #salt
17:15 whytewolf I wouldn't say it irks me.. only when i have to debug someones code that refuses to use - name
17:15 geekatcmu 8)
17:16 geekatcmu We have some code like that still because ... it took us a while to realize
17:16 geekatcmu a) what was going on, and
17:16 geekatcmu b) what that was bad.
17:16 geekatcmu s/what/why/
17:17 whytewolf the only time i use that "feature" is with file.managed. because it does making watches make a little more sense. although it confuses a lot of people that don't know watches watch states and not the files.
17:17 geekatcmu ^^
17:18 kickerdog joined #salt
17:18 geekatcmu Which is why our standards docs say to never use that feature.
17:18 kickerdog left #salt
17:18 geekatcmu *particularly* because the file name frequently changes with OS.
17:18 geekatcmu /etc/bacula/bacula-fd.conf vs. /usr/local/etc/bacula-fd.conf
17:18 whytewolf true. but I'm the only one reading my code. in a professional place I wouldn't do that
17:19 geekatcmu We have more FreeBSD hosts than we have Linux
17:19 geekatcmu So it's a very big deal to us.
17:20 twork so, for ages now, i've been trying to apply: https://docs.saltstack.com/en/latest/topics/tutorials/states_pt4.html#salt-fileserver-path-inheritance
17:21 twork ...and i *am* a moron, but it seems to me that the "first one wins" business isn't applying. stuff i put in the second-listed file root gets applied after the stuff above it.
17:21 twork what am i misunderstanding here?
17:22 whytewolf do they have the same name and root structure?
17:23 twork whytewolf: in the minion, yes (isn't that the point?); in the master, no, if i understand you correctly
17:23 twork i'll post, gimme a minute
17:24 twork (and you told me last night that nobody tries to do it this way, so point taken, but this is a grudge)
17:24 whytewolf I didn't say no one does. I said do what works for you
17:25 whytewolf I'm 1 person. not a very large sample size of the salt population
17:25 mpanetta whytewolf: I am pretty sure when you tell it to watch a file and not another state, that it actually watches the file
17:25 geekatcmu I'm one person, but I'm large.  DOes that make me a large sample size?
17:25 mpanetta Else this would not work:
17:26 mpanetta https://gist.github.com/bentwire/2a44542010fb78366ca7
17:26 mpanetta And I promise you that works :)
17:29 whytewolf mpanetta: strange. let me test that. I have a state that watches /etc/cinder/* [cause i do have some states that start eith /etc/cinder] and files that i don't manage in that directory. I just changed one of the files i don't manage lets see if it restarts the services
17:30 mpanetta Hmm maybe it can wildcard the state name then?
17:31 mpanetta Yeah the state name does start with /etc/rsyslog.d so maybe it is the state name and it just looks like it is the file?
17:33 whytewolf mpanetta: yes, unforchantly I apperently changed a file i do manage at some point so the states did restart
17:33 mpanetta Hmm
17:34 impi joined #salt
17:34 whytewolf but yes, globbing state names does work
17:34 mpanetta That is both powerful and kinda scary at the same time...
17:37 whytewolf yes. yes it is. which is the heart of the disscussion with geekatcmu. and his way is more correct. you shouldn't use the - name as a state in one place if you do it no where else. I do it cause {lazy} and I only deal with me and a single operating systems.
17:37 sunkist joined #salt
17:38 whytewolf s/in/id
17:38 hightekvagabond joined #salt
17:39 mpanetta Ah yeah that makes sense
17:45 troyready joined #salt
17:48 elsmo joined #salt
17:49 grumm_servire joined #salt
17:51 bhosmer joined #salt
17:52 perfectsine joined #salt
17:56 ziro` joined #salt
17:56 Xevian joined #salt
17:56 Tyrm_ joined #salt
17:58 cmclaughlin joined #salt
17:58 bhosmer joined #salt
17:59 jhauser_ joined #salt
17:59 Vynce joined #salt
17:59 wendall911 joined #salt
18:00 baffle joined #salt
18:00 davromaniak joined #salt
18:00 davromaniak joined #salt
18:01 tristianc_ joined #salt
18:01 shnguyen joined #salt
18:01 ablemann joined #salt
18:01 ]V[ joined #salt
18:03 writtenoff joined #salt
18:04 Deevolution Is there any way to have the salt cp module copy from Pillar rather then from Salt?
18:05 keekz joined #salt
18:05 nlb joined #salt
18:06 g3cko joined #salt
18:06 andrew_v_ joined #salt
18:06 beneggett joined #salt
18:07 iggy ssplatt: it'd be (profile|provider):minion:master and https://github.com/saltstack/salt/issues/26699 ?
18:07 saltstackbot [#26699]title: [salt-cloud] Minion doesn't install on some providers after upgrading to 2015.8.0 | Hi...
18:08 Bryson joined #salt
18:09 ssplatt iggy: yeah updating the salt-cloud bootstrap script and adding the minion:master config to my profile worked.
18:09 ssplatt now i’ll have to work on auto key accept and auto provisioning
18:09 winsalt any of the windows salt-cloud users here?
18:10 lemur joined #salt
18:10 iggy mpanetta: yes, there was support for globbing "name" at some point
18:11 iggy which imo makes it very strange as it only applies to that one case, but nobody ever listens to me
18:11 iggy ssplatt: if the bootstrap script runs correctly, the keys should be automatically accepted already
18:12 ssplatt ok then hmm.  i was thinking i had a config option wrong somewhere.  even if i have teh new vm connect to a syndic?
18:12 ssplatt but i’m calling salt-cloud off the master of masters
18:12 * iggy runs
18:13 losh joined #salt
18:17 otter768 joined #salt
18:18 Reinhart joined #salt
18:21 tawm04 joined #salt
18:31 colegatron joined #salt
18:31 lemur joined #salt
18:36 aw110f joined #salt
18:37 TooLmaN joined #salt
18:38 jaybocc2 joined #salt
18:39 DammitJim joined #salt
18:39 RandyT winsalt: I am here...
18:42 winsalt If you remember, I have been tripping over the bootstrap process for a couple days. But I think im getting closer to a solution.  When the bootstrap script failed to connect to the fresh vm, I could restart it through vcenter and suddenly its able to connect
18:43 kitplummer joined #salt
18:44 RandyT winsalt: I'm not going to be much help on the vmware side. Is it possible that the application of firewall changes is failing in some cases and a reboot is fixing that?
18:46 winsalt I dont think its a vmware problem, its definitely something wacky with the firewall.  I just dont get why it needs a reboot, or how I can get the image to stay in the "good" state
18:47 winsalt But I dont think there is a way to add anything to the bootstrap process is there? like a custom script
18:49 RandyT winsalt: iptables firewall changes?
18:49 kitplummer joined #salt
18:50 Craighton left #salt
18:51 RandyT winsalt: I guess to debug it, I would see if I could get on the image when it fails to see if the firewall is applied or not.
18:51 RandyT the boot process might need some way to validate that the firewall settings took effect and force a reload/restart of firewall rules if not.
18:51 RandyT doesn't really sound like salt is involved in this side of the issue.
18:52 winsalt yeah, ive definitely narrowed it down.
18:53 winsalt its fixable by running "winrm quickconfig" followed by a restart
18:53 winsalt but for some reason the windows image loses that configuration, i dunno
18:53 jeffpatton1971 joined #salt
18:54 chadhs joined #salt
18:54 kitplummer joined #salt
18:54 opensource_ninja joined #salt
18:54 RandyT winsalt: oh yeah, forgot for a second we were talking about windoze here...
18:54 winsalt i know :/
18:55 RandyT winsalt: sounds like you are still mixing winrm and winexe?  or are you just using the winrm tool to clear this?
18:56 hightekvagabond joined #salt
18:56 voileux joined #salt
18:56 liskl joined #salt
18:57 winsalt well I know winrm sets up some firewall rules, so whatever its doing lets winexe connect after the reboot
18:57 winsalt im just not sure what
18:57 ageorgop joined #salt
18:57 RandyT winsalt: I would expect unpredictable results if you have winrm and winexe configured in profile/provider
18:57 RandyT might be better in your case to just pull back to using winrm only?
18:57 abednarik joined #salt
18:58 winsalt only winexe is referenced in the provider.  I could try with winrm again, i guess
19:01 kitplummer joined #salt
19:02 sunkist joined #salt
19:03 perfectsine_ joined #salt
19:11 denys joined #salt
19:11 lemur joined #salt
19:11 kitplummer joined #salt
19:12 Diaoul joined #salt
19:13 ghostisic joined #salt
19:16 giantlock joined #salt
19:18 hightekvagabond joined #salt
19:19 Jester joined #salt
19:20 perfectsine joined #salt
19:22 larsfronius joined #salt
19:24 numkem joined #salt
19:27 jaybocc2 joined #salt
19:27 chadhs with salt-call rather than doing salt-call state.highstate can you just call one single state file?
19:28 chadhs `salt-call salt/test.sls` for example?
19:28 winsalt yeah, if you use state.sls statefile instead of state.highstate
19:29 chadhs i was so close lol, thnx
19:30 babilen chadhs: https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.state.html#salt.modules.state.sls + sls_id (directly underneath)
19:32 Guest71482 joined #salt
19:38 hackel joined #salt
19:38 ajw0100 joined #salt
19:39 chadhs thank ye
19:39 stomith joined #salt
19:39 Ryan_Lane joined #salt
19:40 iggy chadhs: state.apply is also handy (without args it's highstate, with args it's state.sls)
19:41 TyrfingMjolnir joined #salt
19:42 opensource_ninja joined #salt
19:45 toastedpenguin joined #salt
19:47 toastedpenguin joined #salt
19:51 winsalt is state.apply what we are supposed to use?
19:54 denys joined #salt
19:55 mik__R joined #salt
19:56 babilen it is, but i've had problems with some ivocations
19:57 numkem joined #salt
20:00 perfectsine joined #salt
20:00 stomith so what is the actual difference between a pillar and a state? don't they do the same thing basically?
20:01 sdm24_ joined #salt
20:02 hemebond stomith: Pillar is for data. State is for changes.
20:03 Andre_ks Hi.  I have an issue.  I created environments directories for dev and prod.  And placed one top.sls in the root dir.  But when running highstate it reports an error  No Top file or external nodes data matches found.
20:04 hemebond Andre_ks: top.sls has to go into base dir
20:04 hemebond You need to have a "base" environment unless you've changed the master config.
20:04 Andre_ks base of the environment?
20:04 hemebond What does your file_roots look like?
20:04 stomith hemebond, thanks
20:06 Andre_ks file_roots:   base:     - /srv/salt   development:     - /srv/salt/development   production:     - /srv/salt/production
20:07 Andre_ks previously, without base
20:07 Andre_ks but same results
20:07 hemebond Andre_ks: Hmm, not sure if it's bad practice to have them nested like that. I kept them adjacent.
20:07 Andre_ks the top.sls is in the base
20:07 hemebond And what does your top.sls contain? Use pastebin or something please.
20:08 babilen http://refheap.com, http://paste.debian.net, https://gist.github.com, http://sprunge.us, http://dpaste.de, …
20:08 Andre_ks http://pastebin.com/u1urS82a
20:08 Andre_ks top.sls
20:09 babilen I'd keep them separate
20:09 Andre_ks what do you mean by nested?
20:09 hemebond Your prod and dev environments are INSIDE your base environment.
20:09 babilen /srv/salt/base /srv/salt/development ...
20:09 hemebond ^ that's how I have it
20:09 toanju joined #salt
20:10 hemebond Andre_ks: and your top.sls is wrong.
20:10 hemebond Production has no targeting.
20:10 Andre_ks I only have one minion right now
20:10 Andre_ks in developmentg
20:11 Andre_ks what is the function of the base environment vs any other environment?
20:12 hemebond base is the default. usually production.
20:12 whytewolf base is a catch all
20:12 babilen It is, in a way, the base ;)
20:13 enigma99a joined #salt
20:14 enigma99a hey all.. Working with Jinja2 and want to manipulate IP addresses. Looks like there is no regex by default?
20:14 enigma99a is there a way to extend jinja filters
20:14 hemebond enigma99a: Custom filters.
20:15 babilen If only those would work in salt
20:15 hemebond Yeah, I was just wondering that.
20:16 babilen https://github.com/saltstack/salt/issues/28236
20:16 saltstackbot [#28236]title: Feature: Allow users to define custom jinja filters in _filters | It would be awesome to be able to easily extend the jinja renderer ourselves as detailed in http://jinja.pocoo.org/docs/dev/api/#custom-filters...
20:16 Andre_ks I also get ERROR: Minions returned with non-zero exit code
20:16 * whytewolf would love if jinja was jinja was more exposed so we could tweek all of its goodness
20:16 quix_ joined #salt
20:16 ajw0100 joined #salt
20:16 babilen salt is *extremely* limited due to the fact that they chose jinja and do not allow for the definition of custom jinja filters nor ship a plethora of in-house filters (like ansible does)
20:17 enigma99a hmm
20:17 enigma99a wow
20:17 babilen whytewolf: Well, comment on the bug. I *might* find time one day to implement that, but it is IMHO such a basic necessity that I wonder why it hasn't been implemented already (along with a huge library of useful filters)
20:18 babilen Which is why I write a bunch of things in pure Python
20:18 hemebond enigma99a: You could use the Python ... uh...
20:18 hemebond Thingy.
20:18 whytewolf will do
20:18 otter768 joined #salt
20:19 babilen enigma99a: I found that the best way to deal with this is to write your own custom execution module in which you collect all those functionality that is lacking. You can call other execution modules from in there too (like network.ip_addr)
20:19 hemebond Renderer
20:19 babilen There are also a bunch of Python based renderers
20:19 babilen .. thingies ;)
20:19 hemebond You could use a Python renderer instead of Jinja.
20:19 winsalt yeah, I dont like jinja either.  It's a lot more transparent and useful to use pure python imo
20:19 winsalt or pyobjects
20:20 babilen Indeed
20:20 stomith does the python renderer work on the client, or server side? or both?
20:20 babilen They should have chosen mako as it allows for the inclusion of literal Python blocks
20:20 whytewolf mako is support IIRC. question is. how much
20:21 quix joined #salt
20:21 babilen Yeah, it is supported .. it is just not the default which means that nobody uses it, there is very little docmentation, formulas aren't written in it and your colleagues will scream at you if you use three or four different renderers in your codebase
20:21 whytewolf lol. this is true
20:22 stomith job security
20:22 babilen And given that one doesn't even have to define the renderer pipeline shebang in *every* SLS one can't even change the default renderer locally as that breaks *all the things*
20:22 conan_the_destro joined #salt
20:22 babilen ... which is something I don't understand. It wouldn't be too hard to force users to write #!jinja|yaml in every SLS, but meh ...
20:23 timoguin joined #salt
20:23 winsalt there is too much emphasis on making config files "readable", and all it really does is make them dumber
20:24 hemebond But isn't that why you can just use pyobjects or something when needed?
20:24 perfectsine joined #salt
20:24 hemebond If your needs are simple, a simple config file is nice.
20:24 hemebond And you can keep complex processing out of your main config files.
20:25 babilen https://github.com/saltstack-formulas/mysql-formula/blob/master/mysql/files/my.cnf .. totally readable
20:25 winsalt yeah, but if the things we are trying to solve were so simple then they wouldnt be problems right?
20:25 hemebond babilen: That's not what I would call a simple config :-)
20:26 babilen No, that was my point
20:26 hemebond But... you could just use something other than Jinja to do that.
20:26 babilen The problem is one of scalability. Naturally you start writing your system using the basic tools that are available. You do this as you try to keep it *as simple* as you can.
20:27 babilen Which, in the context of salt, means that you start with YAML ... then you add some jinja and the logic gets harder. At a certain point the code is so complex and hard to maintain that you would, essentially, have to switch to a completely different tool (e.g. Python) which means that you have to implement everything from scratch.
20:27 Andre_ks hemebond: does the top.sls for multiple env go in /srv/salt or in what dir?
20:28 Bryson joined #salt
20:28 hemebond Andre_ks: You should create a /srv/salt/base directory and update your master config file_roots and then put top.sls in there.
20:28 babilen hemebond: It is grown .. Feel free to reimplement that in Python and replace it and risk breaking a bunch of production MySQL deployments ...
20:29 hemebond Don't nest (put inside) your environment directories; will only lead to problems.
20:29 babilen Didn't we already suggest that ten minutes ago?
20:29 hemebond babilen: Are formulas using tags?
20:29 hemebond Or does it just use the latest version of master if you use as a formula?
20:29 babilen No, there isn't really any versioning scheme or proper releases (yet)
20:29 hemebond Oh.
20:29 Andre_ks so top.sls inside /srv/salt/base ?
20:30 babilen It simply uses the latest version (which is why you should clone the repo and only push upstream changes once you are happy with them)
20:30 babilen Andre_ks: yeah
20:30 hemebond Andre_ks: Yes, and then change your file_roots to point your base environment to that directory.
20:30 Andre_ks yep
20:30 Andre_ks did all that
20:30 Andre_ks same errors
20:30 hemebond Andre_ks: Did you restart your master?
20:31 Andre_ks yes
20:31 hemebond babilen: Rewriting/reimplementing based on new complexity and requirements seems normal to me.
20:31 hemebond Andre_ks: Did you fix your top.sls?
20:31 hemebond Your production environment had no targeting rule.
20:31 Andre_ks yes, it only has development now, because I only have 1 minion and its in development
20:32 hemebond Can you repaste your top.sls and file_roots into a new paste please?
20:32 babilen hemebond: Yeah, I totally agree. It is just that you would want to do that in a context or using tools that allow you to refactor easily without switching to an entirely new system. I am completely on board if you argue that the code I showed should have been ported to Python (or something else) quite some time ago.
20:32 aw110f joined #salt
20:33 Andre_ks http://pastebin.com/EcvXMStG
20:33 Andre_ks http://pastebin.com/c2QcNWdE
20:33 babilen hemebond: I rather meant that you would want a language that allows you to add complexity in a natural way... I am quite a fan of domain specific languages and would love to see a LISP based renderer ;)
20:33 hemebond babilen: Ah I see. I don't like DSLs myself.
20:34 enigma99a babilen: How do you get pillar data into templates if they are pure python? They have to be passed in manually via the SLS?
20:34 hemebond enigma99a: Your pillar SLS is the Python code :-)
20:34 whytewolf humm, maybe they should have fully pluggable renderers
20:35 hemebond You return a dict
20:35 babilen enigma99a: You have __pillar__
20:35 hemebond I thought you could plug in renderers. No?
20:35 babilen Maybe .. haven't done that
20:35 ajw0100_ joined #salt
20:35 babilen Wouldn't be surprised though
20:36 hemebond Andre_ks: Can you paste the full output/error and command you're having trouble with please?
20:36 babilen hemebond: That's okay. I think that DSLs have a certain beauty ...
20:36 hemebond You can paste it all into the same pastebin if you like.
20:36 hemebond babilen: I have to use Puppet all day :-(
20:36 whytewolf there is _renderers
20:36 whytewolf but here is the extent of that documentation https://docs.saltstack.com/en/latest/ref/renderers/#writing-renderers
20:38 Andre_ks hemebond: http://pastebin.com/PCQEUMfg
20:38 babilen hemebond: I was thinking of *beautiful* DSLs not something that's horrible ...
20:38 hemebond babilen: ;-D
20:39 enigma99a thanks guys
20:39 enigma99a I guess I will just go pure python for some things for now
20:39 hemebond Andre_ks: Can you paste the role_based_auth state?
20:40 hemebond And where is that state?
20:41 hemebond Andre_ks: What if you do salt '*' state.show_top
20:41 Andre_ks its inside:  /srv/salt/development/states/role_based_auth/init.sls
20:41 Andre_ks i3-d-salt-linux-test:     ----------     development:         - role_based_auth
20:41 hemebond Okay, it can't be in a states subdirectory
20:42 andrew_v joined #salt
20:42 hemebond Your development env path just points to /srv/salt/development/
20:42 hemebond That is your state directory for the development environment.
20:42 TyrfingMjolnir joined #salt
20:43 Andre_ks there is a state dir, /srv/salt/development/states/
20:43 andrew_v joined #salt
20:44 hemebond Andre_ks: Yes, but Salt can't find it because your states should be in /srv/salt/development/
20:44 hemebond Not in a subdirectory.
20:44 Andre_ks oho
20:44 mpanetta iggy: What?  Sorry I missed your message.  I really ought to set up a bouncer...
20:44 hemebond file_roots is where you configure your states directory.
20:45 mpanetta I was just at our local saltstack meetup so I missed your message for a good reason :P
20:46 _Cyclone_ joined #salt
20:46 timoguin_ joined #salt
20:46 babilen (which you can, naturally, set to /srv/salt/development/states .. but why would you?)
20:46 Andre_ks so top.sls and states and pillars needs to be in file roots?
20:46 babilen No, pillars are somewhere else
20:46 babilen (/srv/pillar/{base,development,production} for example)
20:46 jeffpatton1971 joined #salt
20:47 babilen Why isn't that /srv/pillars btw?
20:47 hemebond Andre_ks: http://pastebin.com/aR3YsX2f
20:47 * babilen can never remember is a particular thing uses pillar or pillars
20:47 hemebond Not sure why pastebin broke my formatting...
20:48 babilen because it is pastebin.com and shit
20:48 babilen Use one of http://refheap.com, http://paste.debian.net, https://gist.github.com, http://sprunge.us, http://dpaste.de, …
20:48 hemebond *sniff*
20:48 babilen Seriously, that site should be banned
20:48 hemebond But it's so easy to remember.
20:48 hemebond How would I remember sprunge
20:49 hemebond Wait... what is sprunge.us
20:49 hemebond Oh
20:50 babilen (that one might not be too applicable in this channel)
20:50 enigma99a babilen: Any examples out there for a python file as a template? I see plenty for SLS files
20:50 baweaver joined #salt
20:51 babilen enigma99a: One second
20:51 enigma99a thanks :)
20:51 Andre_ks hemebond: the minion has nothing inside /srv/
20:51 Andre_ks not even a salt dir
20:52 hemebond Andre_ks: Are you not using a master?
20:52 Andre_ks yes I am
20:52 hemebond All this is done on the master.
20:52 Andre_ks ok
20:53 Andre_ks will reorganize the dir hierarchies
20:53 Andre_ks and see how it goes...
20:53 Andre_ks thanks
20:53 hemebond Good luck.
20:55 Andre_ks so if I want 1 top.sls file, where should it go?
20:55 Andre_ks for multiple environments?
20:55 Andre_ks in the root of the base state dir?
20:55 babilen enigma99a: https://www.refheap.com/111802 -- that one uses ConfigParser for writing a configuration file. In the end you just implement "run()" and return the string you want to be in the file.
20:55 enigma99a babilen: thanks!
20:56 babilen enigma99a: https://www.refheap.com/111803 -- this has the names right
20:56 bhosmer joined #salt
20:56 hemebond Andre_ks: put it into /base
20:57 sunkist joined #salt
20:57 babilen Andre_ks: You *might* also want to consider using GitFS (in which branches are environments). In that case I'd recommend to use a separate repository for your top.sls as top.sls files are merged across environments
20:58 perfectsine joined #salt
21:01 Vynce joined #salt
21:02 quix joined #salt
21:03 GreatSnoopy joined #salt
21:04 ajw0100 joined #salt
21:07 slav0nic joined #salt
21:07 basepi joined #salt
21:09 kitplumm_ joined #salt
21:10 Rumbles joined #salt
21:11 jaybocc2 joined #salt
21:16 baweaver joined #salt
21:20 Heartsbane Question is there a way to assign a IP with salt-cloud that is just spun up without modifying the profile?
21:20 aidalgol joined #salt
21:21 Heartsbane For example tell it what the IP is at bootup from the command line, since we don't what it will be yet
21:28 Rumbles joined #salt
21:32 guest left #salt
21:32 teemo joined #salt
21:43 tkharju joined #salt
21:45 notnotpe_ Hi folks, any word on when the 2015.8.2 release will be finalized?
21:46 jaybocc2 joined #salt
21:47 ssplatt does anyone have an example of a master sending a custom reaction to a syndic?
21:47 ssplatt i’m trying to figure out how to get my syndics to auto-accept keys when salt-cloud is run on my master of masters.  without having shared filesystems
21:48 ssplatt i see teh one example in the docs, but that only works with the local master
21:48 forrest joined #salt
21:48 ssplatt er, a single master.
21:49 notnotpeter I see there's a v2015.8.2 tag https://github.com/saltstack/salt/releases/tag/v2015.8.2 Are there pre-release packages somewhere for testing?
21:50 ssplatt idea is: run salt-cloud on master, master receives ‘salt/cloud/*/created’ and sends event to syndic1 to accept the key, then run highstate
21:51 openfly left #salt
21:53 hightekvagabond joined #salt
21:53 jaybocc2_ joined #salt
21:56 s_kunk joined #salt
21:56 forrest Serious question here, not trolling. How many people are using suse? I'm watching their closing keynote about the suse manager and salt's role in that.
21:56 stomith I've never used it. I'm told suse is popular in Europe.
21:57 ssplatt i think where i’m getting hung up on how exactly to use event.send/fire in a ‘tell-accept-key.sls’.   i see plenty of command line examples.
21:57 aurynn I've never used it either, and it's not common in my circles
21:57 geekatcmu I used it for 6-8 months in 2006, until I replaced it with RedHat (we had site licensing, so why use SuSE?).
21:57 forrest Interesting, I've only ever installed it once when I was playing with distros, I've never seen it used in a single prod environment I've worked at or interviewed at.
21:58 tristianc joined #salt
21:58 geekatcmu At a previous job where they were outsourcing to T-Systems, SuSE was *the* supported Linux, IIRC.
21:58 forrest geekatcmu, ugh outsourcing :(
21:59 geekatcmu That's the number one reason why I'm here.
21:59 forrest suse?
21:59 geekatcmu "Oh, we have a position for you here."  "Put that in writing and I'll consider staying." "..."
21:59 forrest ahh gotcha
22:00 geekatcmu The senior execs were mostly former site managers, very blue collar, and very much of the opinion that IT folk were just as replaceable as fork lift drivers.
22:01 geekatcmu Given their problems in hiring good forklift drivers, they may well have been correct in that thought, but still completely bass-ackwards.
22:01 tristianc joined #salt
22:02 eliasp whatever project is outsorced to T-Systems can be considered a total failure from the start on…
22:02 eliasp I've seen many major projects having been outsourced to them… and every single one simply failed or rotted away
22:02 forrest I've never even heard of that company
22:02 CheKoLyN joined #salt
22:02 eliasp I've never seen an IT company as incompetent as these guys
22:02 forrest also isn't it a bit late eliasp?
22:03 keimlink joined #salt
22:03 eliasp forrest: subsidiary of German Telekom (just as T-Mobile)
22:03 eliasp forrest: nah, never… have to work on my 2 side-jobs ;)
22:03 forrest Ahh okay, I did some research when I was considering a move to Germany and the telecom companies seemed shit
22:03 forrest 2 side jobs?? I think you need a new primary job... ;)
22:04 eliasp forrest: well, depends on how you look at it… one of the side-jobs might be a new primary one sooner or later ;)
22:04 forrest fair enough
22:04 eliasp and I'm quitting the other one by the end of the year…
22:04 forrest that's good, working too much as bad
22:05 eliasp exactly ;)
22:05 forrest If I've learned anything being unemployed recenty, it's that working on what you want is aaaaamazing.
22:06 eliasp forrest: it really is… I just want to be able to spend more time on it ;)
22:06 forrest Understandable
22:07 eliasp well… enough talked… got to continue, otherwise there'll be no results ;)
22:07 forrest Later
22:08 geekatcmu WOrking with T-Systems probably works just fine if you're in Europe.
22:09 geekatcmu In the US, though, the provisioning process for everything take easily 3x longer between offset in work shifts and general asshattery.
22:09 forrest geekatcmu, eliasp lives in Germany, so if he says they're shit, I trust his European opinion.
22:09 geekatcmu heh
22:10 eliasp geekatcmu: I'm at the very heart of T-Systems ass-hattery… I get to experience their incompetence currently on a daily basis as they're doing the base-IT for one of our customers…
22:10 geekatcmu "fun"
22:10 eliasp it's absolutely horrible… even my 3y old kid would do a better job than they are doing it
22:11 geekatcmu I cannot imagine surviving with them for long.
22:11 saltstackbot [reddit-saltstack] salt master firing events to syndics https://www.reddit.com/r/saltstack/comments/3t7qga/salt_master_firing_events_to_syndics/ - 2015-11-17 - 22:09:40
22:12 geekatcmu "Nope, sorry, it's been 40 hours, I'm out of here.  You planned on me working 60-80 hour weeks?  That seems like a pretty stupid idea to me."
22:12 geekatcmu Everyone who didn't bail during the outsourcing got forcibly converted to T-Systems and, yeah, that was the expectation.
22:13 geekatcmu Which was pretty clear from the first intro to the company.
22:13 baweaver joined #salt
22:15 MadHatter42 joined #salt
22:19 otter768 joined #salt
22:19 forrest lol
22:20 forrest geekatcmu, Yeah I've been having this discussion with a lot of places I've interviewed at. They want me to commute to an office (2 hours or more total each day), but don't want to pay a salary that would support losing all that time
22:20 forrest So you want 40+ hours, and on-call, and you want me to go from remote to commuting and wasting my life doing that... for how much?
22:20 forrest Oh no thank you...
22:21 geekatcmu yeah
22:21 geekatcmu WFH FTW
22:21 forrest The amount of people that think commuting is 'worth it' for the job is insane, like there's some sort of pride working in an office with a bunch of other people using chat and wearing headphones all day.
22:22 geekatcmu YOu know what I miss about working from the office?  Free food.  Then I go visit for a week and come back with an extra 5lbs and am happy again.
22:23 whytewolf humm. I actually gained weight by working from home. cause i had access to more tasty food
22:23 whytewolf and less distance to travel
22:23 geekatcmu We have only healthy food here.
22:24 geekatcmu The office has a full deli, only everything is free, including all the cookies, chips, ice cream, etc.
22:24 geekatcmu OK, that's not true, there is some unhealthy food here.
22:24 whytewolf lol
22:25 geekatcmu But there's a difference between "grab a single serving bag of chips, a cookie, and a soda" and "find the family-size bag of chips in the pantry, get out a bowl, serve, put away the bag, take the chips to my office, then have to take the bowl back to the kitchen when I'm done."
22:25 geekatcmu I'm *lazy*
22:25 basepi I just eat directly from the bad
22:25 basepi bag*
22:25 basepi which is why I don't buy chips
22:26 geekatcmu I do if I'm sitting at the dinner table having chips and salsa/queso, but not in my office.
22:26 whytewolf I do love work from home. and never understood going into an office for most of what i do anyway. the office isn't around the computers i typically support either. how is me being in an office vs being at home going to help me when the server is in another country
22:26 geekatcmu ^^
22:27 geekatcmu Even if I were at the office, the computers I work with are in a DC 30 miles away from the office.
22:27 whytewolf only 30 miles? when I worked at IGT, the servers were on the other side of the world.
22:27 geekatcmu Any distance beyond "get up and walk over to it" is effectively remote.
22:28 geekatcmu Oh, I have servers in APAC and EU, too.  But the majority of the ones I have to work with are near the office.
22:28 geekatcmu Well, when I've got my MySQL hat one.
22:28 geekatcmu s/one/on/
22:28 roock joined #salt
22:29 whytewolf the closest servers to me were a bunch in kamloops canada
22:29 geekatcmu For Cassandra, every place is equally remote
22:29 basepi geekatcmu: even if you were within get up and walk over to it distance, how many times in your career have you actually done that?
22:29 whytewolf I'm in vegas
22:29 forrest basepi, I don't know if I've ever poured chips into a bowl, onto my plate sure.
22:30 holler joined #salt
22:30 geekatcmu basepi: Uh, considing I personally physically deployed the CMUQ datacenter servers, and while at CMU I did all the physical deployment of the admin computing systems, lots
22:30 holler hello, how can I pass "-g" to npm.installed: -names: ember-cli ?
22:30 holler for e.g. "npm install -g ember-cli"
22:30 whytewolf basepi: when i was working at the paper. about once a week but that was cause we didn't have tiered IT. I was the begining and end of the unix support
22:30 geekatcmu In fact, this is the first job I've had where I wasn't required to lay hands on equipment on a regular basis
22:30 geekatcmu That's a good summary of my 21 years at CMUQ.
22:31 geekatcmu Er, CMU
22:31 forrest holler, What does -g even do? Their help page hilarious doesn't say
22:31 holler forrest: installs globally i think
22:31 holler im not even sure if its required but on ember-cli it says to do tha
22:31 forrest if it's a global install, wouldn't just running it as the root user satisfy that?
22:31 forrest I'm not very familiar with npm.
22:32 basepi geekatcmu: alright I'll give you that.
22:32 whytewolf forrest: https://docs.npmjs.com/cli/npm "Local mode is the default. Use -g or --global on any command to operate in global mode instead."
22:32 forrest thanks whytewolf I was looking at https://www.npmjs.com/package/cli-manpage
22:32 holler is there a way to pass that arg to npm.installed ?
22:32 forrest holler: dir
22:32 forrest The target directory in which to install the package, or None for global installation
22:33 forrest holler, Just exclude dir and the -g flag is automatically passed it looks like
22:33 holler ah ok nice
22:33 whytewolf lol. pssshhhh a manpage? we are talking about a system writen in javascript.
22:33 forrest whytewolf, I know, my expectations were low, but they weren't quite that low.
22:34 whytewolf oh wow, I never even looked at the npm manpage before. that i think is the shortest manpage i have ever seen
22:35 jhauser joined #salt
22:36 forrest whytewolf, Node doesn't even have packages supported by the actual node team.
22:36 forrest https://github.com/nodejs/node-v0.x-archive/wiki/Installing-Node.js-via-package-manager
22:36 forrest big disclaimer
22:36 whytewolf holler: looking at the docs. if you don't set dir: then it is globall
22:37 forrest whytewolf, I already said that
22:37 forrest whytewolf, YOU'RE TOO SLOW.
22:37 whytewolf oh i missed it
22:37 whytewolf lol
22:37 whytewolf trying to do work and talk in IRC at the same time
22:37 forrest ;)
22:37 holler forrest: whytewolf: thanks!
22:37 forrest Like I said, the 'joys' of unemployment
22:37 forrest holler, For sure.
22:37 Reinhart joined #salt
22:38 whytewolf eh, it is only complicated by the fact i am using two different computers for it. so i have to actually turn my head to see IRC
22:39 whytewolf seperation between the bank and my personal stuff.
22:42 whytewolf makes me miss the days when i would find the hole in ITsecs firewall and punch an ssh connection out to my own servers. so that i could IRC chat.
22:43 zmalone joined #salt
22:43 denys_ joined #salt
22:46 Vynce joined #salt
22:53 dkrae joined #salt
22:53 Reinhart joined #salt
22:57 bhosmer joined #salt
23:04 benegget_ joined #salt
23:09 conan_the_destro joined #salt
23:10 whytewolf humm. did not think of this. putting a filesystem on a new lvm created disk ... easy with orch. difficult with states
23:11 whytewolf unless I am missing a state :P
23:12 RandyT whytewolf: confirming that the data is passed under different name depending on event that is triggering as you demonstrated.
23:13 RandyT question: getting my head around pillars now. From cmdline, I can see pillar data for a specific minion with 'salt minionname pillar.items' and can view a specific key with 'salt minionname pillar.item keyname'.
23:13 whytewolf RandyT: I wouldn't call it different. it is kind of what the meaning of the values are. minion events repersents an id of a minion. but cloud events is giving the name of a instence they created. they should be the same. but it is possable they are not.
23:13 RandyT what do I need to do to see a value for a nested key?
23:14 whytewolf RandyT: pillar.get keyvalue1:subkeyvalue
23:14 RandyT whytewolf: it would make things more intuitive and enable sharing reactor states across events if these values were at least aliased...
23:15 geekatcmu whytewolf: it always comes down to this question: is there already data there that I need to not destroy?
23:16 * RandyT smack
23:16 RandyT tried everything but the :
23:16 RandyT whytewolf: tks
23:16 whytewolf geekatcmu: unforchantly this is going into highstates that will hopefully in the end of the day be in a schedule.
23:17 geekatcmu good luck storming the castle!
23:17 whytewolf lol
23:18 larsfronius joined #salt
23:19 teryx510 joined #salt
23:25 ssplatt joined #salt
23:27 KajiMaster joined #salt
23:30 mohae joined #salt
23:30 aw110f joined #salt
23:31 kitplumm_ joined #salt
23:32 jaybocc2 joined #salt
23:34 tiadobatima joined #salt
23:35 whytewolf so close. but it looks like the xfs module does not work right with ubuntu. issue fileing time
23:42 beneggett joined #salt
23:47 Tyrm joined #salt
23:50 jaybocc2 joined #salt
23:50 thefish joined #salt

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