Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-11-09

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

All times shown according to UTC.

Time Nick Message
00:00 rocket joined #salt
00:06 moeyebus writing to /tmp/ and copying to the destination works fine. I don't have a reason to try more complicated stuff
00:16 zzzirk joined #salt
00:18 PeterO joined #salt
00:21 chris______ joined #salt
00:29 digitalwolf joined #salt
00:29 digitalwolf Good Evening!
00:32 cedwards /quit
00:34 chris______ left #salt
00:45 larsfronius joined #salt
00:58 dthom91 joined #salt
01:08 zzzirk joined #salt
01:12 amcorreia joined #salt
01:23 favadi joined #salt
01:24 sunkist joined #salt
01:35 dthom91 joined #salt
01:43 enigma99a joined #salt
01:43 enigma99a I got d/c, yeah I have 2015.8
01:43 enigma99a Salt States work fine through gitfs
01:43 enigma99a just having an issue with Ext Pillars
01:44 GreatSnoopy joined #salt
01:44 enigma99a it appears to be checkout the correct branch and the data is in the cache, but it doesn't seem to be available to the minion
01:46 enigma99a what is everyone using for their ext pillar data? DB or git?
01:48 favadi joined #salt
01:52 enigma99a do ext pillars work differently than regular ones? do you have to specify the namespace or type of ext pillar when calling it? ie.. mongo, git, mysql etc
01:54 enigma99a I guess I'm wondering if 'salt '*' pillar.items' will list everything regardless of the types of external pillars connected to it
01:55 enigma99a because all I get is nothing :(
01:56 aurynn you might have to force a refresh
01:56 enigma99a salt '*' saltutil.refresh_pillar right
01:57 enigma99a oh that works :)
01:57 enigma99a I'm sure I tried it
01:57 enigma99a who knows lol
01:57 enigma99a so it is working good through git
01:58 falenn joined #salt
01:58 enigma99a is it more popular to use a DB these days?
01:59 aurynn for...?
02:02 enigma99a pillar data
02:03 racooper joined #salt
02:05 nicksloan how do top files in git_pillar work? It's not very clear to me.
02:05 nicksloan I tried to have a single top file in base like I do for my states, but apparently I have to have a top file in each pillar environment for it to work.
02:06 aurynn enigma99a, I guess it's pretty much where you keep your data. Databases are good places for data.
02:06 nicksloan the question is, what happens when the environment is specified in multiple top files then. Is the top file in environment n authoritative for environment n? Does it get merged?
02:11 enigma99a nicsloan: that is specified in the docs
02:11 enigma99a you can merge or keep it separate
02:12 enigma99a a lot of different algorithms to handle it
02:12 edrocks joined #salt
02:12 nicksloan enigma99a: is the behavior the same for the state top file?
02:14 nicksloan I guess the git_pillar requires a top file for every environment thing threw everything else into questionable territory for me. I'm going to write a blog post next weekend about this setup, because I have hit a few snags that weren't obvious in the docs. Maybe I just didn't read carefully enough, but I bet I'm not the only one.
02:15 enigma99a i'm getting my kids cleaned up for bed, but if nobody else chimes in with the link to the docs, I'll look
02:17 enigma99a https://docs.saltstack.com/en/latest/ref/states/top.html
02:17 enigma99a see the bottom
02:17 enigma99a does this answer your question?
02:18 nicksloan enigma99a: I've seen that section. Was just looking for confirmation that that applies to pillar top files equally to state top files. Thanks for looking it up.
02:27 enigma99a See  PILLAR_SOURCE_MERGING_STRATEGY in the salt master config
02:28 JDiPierro joined #salt
02:30 nicksloan I guess the thing that I am struggling with is why git_pillar needs a top.sls file in every environment, when other pillars don't seem to need to (based on what I've read, admittedly, not experimentation). That special casing ruins the one top file recommendation established by the states documentation, which happens to be exactly what I needed to wrangle my states.
02:31 iggy nicksloan: that sounds like a bug... it's been common practice for years to have your top file outside of gitfs or in a separate repo
02:32 nicksloan iggy: mentioned here: http://drivingthestrugglebus.com/saltstack-environments/ (second paragraph (excluding lists) after "How to setup git external pillar" heading)
02:33 iggy I'm not saying that's not how it is, just that it shouldn't be that way
02:33 nicksloan that person suggests that it is documented that you need to have a top file in each environment. That seemed strange, so I looked up the docs, and I could only find the second sentence here: https://docs.saltstack.com/en/latest/ref/pillar/all/salt.pillar.git_pillar.html
02:34 nicksloan But I assume if it is documented, and not called out as a bug, it is intentional behavior.
02:34 PeterO joined #salt
02:34 iggy all the more reason to _not_ use salt's environments
02:34 dthom91 joined #salt
02:35 iggy if that is intentional, it's different from states
02:35 nicksloan yeah
02:35 nicksloan and special casing always scares me
02:35 nicksloan because then you have no idea what else is different.
02:36 enigma99a if you don't use different environments how do you test dev and stb before going to prod?
02:36 iggy separate masters
02:36 nicksloan enigma99a: that was my next question.
02:36 iggy which is a safer option imo anyway
02:37 nicksloan separate masters is where I came from, and that is a pretty disappointing solution.
02:37 nicksloan (not saying it is wrong, but it's certainly far from ideal)
02:39 nicksloan It's just frustrating to run into a heap of surprises every time I try to do something that salt is allegedly capable of. It's such a great project, but the curveballs are brutal.
02:41 enigma99a also, I wanted to have an env per az
02:41 favadi joined #salt
02:41 enigma99a so if something got messed up on one, there would be redundancy
02:42 enigma99a kinda sucks to make a whole different master for each, but I can understand it would be safer
02:45 iggy there's always MoM if multiple masters is too much to handle
02:45 iggy I know of installations with >10000 servers using multiple masters and it's workable
02:46 malinoff joined #salt
02:47 enigma99a iggy: in a large env, you'll have a lot of different teams so would naturally want a lot of envs
02:47 nicksloan iggy: the multiple masters is far less of a problem for me than pursuing a strategy suggested by the documentation, only to find out that it is untenable due to limitations that the docs basically whisper about.
02:47 enigma99a we've got it set up that way with Chef
02:48 enigma99a how do you separate out user teams all in the same env?
02:48 iggy we only have 1 env
02:48 nicksloan posted this in another channel: "I feel like if you have special cases, your documentation should probably scream about them, and there should be an open issue somewhere called "Resolve special case x?""
02:49 iggy dev in VMs/vagrant -> canary -> prod
02:50 nicksloan one env, multiple masters
02:50 catpigger joined #salt
02:50 nicksloan so you just keep your production master pinned at a commit until the states are ready to go.
02:50 iggy we don't use gitfs
02:50 nicksloan well, same idea though, right?
02:51 iggy yeah, effectively
02:51 iggy but we actually have multiple production masters (hence canarying)
02:51 nicksloan sure
02:51 iggy so you could look at each master as an "env"... they just all happen to have the same code
02:52 nicksloan well, code from the same tree
02:53 nicksloan the thing is, to me that just seems like a very expensive alternative to git branches.
02:53 zzzirk joined #salt
02:54 iggy I wish I didn't have to be so vague, but... secret sauwce
02:55 nicksloan and, I know branches can get fucked up much more easily than your solution, but in my organization, with a team of 4, and one product, branches are perfectly fine. Few people have the keys, and their accountability is very good.
02:55 iggy I've certainly worked places where we _would_ have used branches if they worked in a sane fashion
02:55 iggy but they don't... so I've moved on with my life
02:55 nicksloan welp, I had better get to work on this documentation Pull Request
02:56 nicksloan can't really go around saying  that docs should scream about special cases, if I don't ever do any of the shouting.
02:56 iggy I've got an issue open to make it easier to make docs PRs
02:56 jeddi joined #salt
02:56 iggy (i.e. a link directly to the source file for the docs)
02:56 nicksloan oh, fuck. It's all documented in the code, isn't it?
02:57 iggy mostly, yes
02:57 nicksloan ugh
02:57 iggy this might not be
02:57 nicksloan that seems like a great idea, until you have to vet doc changes as if they could break the build
02:58 enigma99a so iggy..  you have multiple master mirrors, and could go... salt '*az1*' state.highstate, update all the az1s on master1 and then do the same thing on az2 on master2 etc?
02:58 nicksloan pretty sure git_pillar is documented entirely in the file
02:58 iggy yeah, it is
02:58 nicksloan the top file stuff might not be
02:59 PeterO joined #salt
02:59 enigma99a I'm thinking since the syndic should send the request back the the original requesting master (if you config it that way)
02:59 enigma99a Sorry, i'm new to Salt
02:59 iggy enigma99a: you could also use MoM, so if you really want to do something across all the masters, you can
03:00 enigma99a yeah, I'll look into it thanks
03:00 iggy syndic is kind of a no go for some people... it has some of the same difficiencies as multi-master (grains matching is inconsistent, no mine, etc)
03:01 enigma99a Right now, I set up 2 masters (mirrored) and 2 syndics
03:01 enigma99a but yeah MoM would probably be better
03:04 bhosmer joined #salt
03:04 enigma99a If I had something like: https://sknjeri.files.wordpress.com/2014/03/33a14-multi-arch.png
03:05 enigma99a Would the jobs etc stay in sync?
03:05 enigma99a If you execute form the MoM it should work good yeah?
03:06 JDiPierro joined #salt
03:06 enigma99a I saw someone say here that you can also use a DB to store all of that info
03:07 enigma99a also what do you do if your master of master dies?
03:07 enigma99a just have a good backup? :)
03:09 falenn joined #salt
03:11 nicksloan is there a reason why fileserver modules and pillar modules aren't implemented in such a way that you configure them, and then you just drop a reference to that configuration in file_roots and pillar_roots, and the actual fileserver/pillar code works exactly the same way it does on every other file source or pillar.
03:12 nicksloan I mean, legacy at this point, obviously.
03:14 PeterO joined #salt
03:24 PeterO joined #salt
03:26 hemebond joined #salt
03:28 fyb3r joined #salt
03:28 fyb3r anyone available to help me with a minion auth issue for 2015.8.1?
03:29 ageorgop joined #salt
03:30 evle joined #salt
03:32 digitalwolf left #salt
03:33 digitalwolf joined #salt
03:34 digitalwolf left #salt
03:42 blast_hardcheese fyb3r, Just ask your question, if someone can help, they will
03:48 dthom91 joined #salt
03:51 Segfault_ joined #salt
04:01 PeterO joined #salt
04:02 otter768 joined #salt
04:03 Cruz4prez joined #salt
04:17 moogyver joined #salt
04:18 dthom91 left #salt
04:18 fyb3r eh dont really feel like waiting forever today. ill keep trying to figure it out mysef
04:21 PeterO joined #salt
04:27 Cruz4prez joined #salt
04:41 PeterO joined #salt
04:45 otter768 joined #salt
04:45 whytewolf fyb3r: http://sol.gfxile.net/dontask.html
04:46 moogyver joined #salt
04:49 ramteid joined #salt
04:51 ronrib joined #salt
04:52 anmolb joined #salt
04:55 anmolb joined #salt
04:57 mehakkahlon joined #salt
04:59 mehakkahlon joined #salt
05:00 hemebond fyb3r: I'm here. If you'd asked your question I could have started looking into it.
05:01 hemebond fyb3r: Still having a problem>?
05:07 hemebond I guess not.
05:12 malinoff joined #salt
05:19 fyb3r sry. yes i am. apparently the minion is having a ton of tcp retransmissions with the master
05:20 hemebond Is this a new master and minion?
05:20 fyb3r neither. existing minion
05:20 fyb3r it and a few others are having the same problem
05:20 fyb3r added a new minion without an issue
05:20 hemebond Any clues when you run the minion in the foreground with --debug?
05:20 fyb3r ive checked everything from the verions of packages compared to working minions and everything matches
05:21 fyb3r just that its not authenticating
05:21 fyb3r the firewalls are all totally open
05:21 fyb3r tcpdump shows that theres a bunch of retransmitting from the minion to the master
05:21 fyb3r thats about all ive got thus far
05:22 hemebond So you can telnet from the minion to the master on 4505 and 4506?
05:22 fyb3r SaltClientError: Attempt to authenticate with the salt master failed
05:23 fyb3r ill try. i believe so though. i know ncat worked
05:23 fyb3r yup telnet works
05:24 whytewolf does the master show any thing in the logs about regecting the minion keys?
05:24 fyb3r centos version: 3.10.0-229.20.1.el7.x86_64    is the other difference
05:24 fyb3r nope
05:24 hemebond Does salt-key --list-all list the minion that's having trouble?
05:25 fyb3r nope. its not there
05:25 hemebond Oh. Is there a file for it under the pki directory?
05:25 fyb3r nope. removed all that, any cached info, etc
05:26 hemebond Did you also remove the key from the minion?
05:26 fyb3r mhm
05:26 fyb3r totally ripped out the minion earlier, dependencies included
05:26 hemebond Has the minion re-created any keys?
05:26 fyb3r then tried different installation methods via yum with the latest repo, bootstrap, and nothing
05:26 fyb3r yum they are creating the keys
05:26 fyb3r yup*
05:27 fyb3r but it acts like they never get a proper response from the master.
05:27 hemebond And the master and minion are both 2015.8.1?
05:27 fyb3r yup
05:27 hemebond And you are telnetting to the same hostname/IP as the one in the minion config?
05:27 fyb3r and i have other 2015.8.1 minions working
05:27 fyb3r yes
05:28 hemebond Are you able to paste the output of running the minion in debug mode?
05:28 fyb3r i noticed that the port the master is trying to connect to on the minion is incrimenting as well
05:28 hemebond Are you able to run the master in debug mode?
05:28 fyb3r sure. just a sec.
05:28 whytewolf fyb3r: the master doesn't connect back to the minion.
05:28 whytewolf the minion only connect to the master
05:29 fyb3r meant when its sending its response
05:29 fyb3r not connecting but firing back the ack's
05:29 fyb3r preference in what paste method to use?
05:29 whytewolf that would be the minion picking a new exit port.
05:29 whytewolf gist
05:29 hemebond It's actually the outbound port.
05:30 hemebond Why have people moved away from pastebin, etc, to gist?
05:30 rdas joined #salt
05:30 whytewolf cause pastebin is blocked at a lot of work places. not to mention the ad software on it
05:30 hemebond Ah I see.
05:30 whytewolf also gist can be updated as well as have multiple files pasted to the same gist
05:32 whytewolf you can also git gist setups
05:32 fyb3r gonna be a minute. takes about 30s or so for the attempt to time out and the error to get thrown
05:32 fyb3r ;) and im on my home network so no issues in tearin gthings up if needed.
05:33 whytewolf I perfer to do that only in the most dire of circumstances. ;) good practice to not start from scratch everytime
05:34 whytewolf oh. you said it is centos. is selinux set to enforcing?
05:34 fyb3r minion : https://gist.github.com/fyb3r/fe1e5112d998ee2a354d    master: https://gist.github.com/fyb3r/d208c7a08f1d8ac60280
05:34 fyb3r nope i removed selinux
05:34 fyb3r er disabled
05:34 fyb3r running these in hyper-v for snapshotting and what not so ive been breaking them pretty bad trying to hack out the issue lol
05:35 whytewolf humm. strange tornado is throwing exceptions
05:36 fyb3r yeah its the same on all of the minions with the problem
05:36 fyb3r which is mainly where im confused lol
05:36 hemebond Is tornado installed?
05:37 whytewolf hemebond: yeah. other wise salt would be the one throwing exceptions :P
05:37 fyb3r yup. otherwise trying to start the minion fails saying tornado.ioloop is not there
05:40 timoguin_ joined #salt
05:40 whytewolf okay, strange way to handle a timeout. take a response. if it is a result return it, if it isn't a result check if exe_info is present and throw that error if it is
05:41 fyb3r yeah lol. and i can send the pcaps if you want
05:41 fyb3r im mainly stumped as to why it would timeout at all
05:42 fyb3r theres working minions on the same host machines, all centos7 and 2015.8.1
05:42 hemebond Have you checked syslog on master at all? (I'm just throwing out things I'd check)
05:43 fyb3r i havent, am now
05:44 whytewolf also check dmesg on master and minion
05:45 fyb3r bad load from minion?
05:46 malinoff joined #salt
05:47 whytewolf maybe
05:52 fyb3r i see eth0 falling in and out of promiscuous mode
05:52 fyb3r on both master and minion
05:52 fyb3r i think that was me redoing their profiles though
05:53 whytewolf most likely tcpdump action
05:55 fyb3r ah yeah that too
05:55 fyb3r man i got nothin lol
05:55 whytewolf strange one thats for sure
05:56 whytewolf I see no reason it shouldn't work. it is just timing out though. any network changes recently?
05:56 whytewolf hyper-v network going bonkers?
05:56 fyb3r yeah i rewired it. but thats not the issue since the same host machine that has the master has a minion on it that will connect and another that wont
05:57 fyb3r and another host machine has 2 minions, one connecting and one not
05:57 fyb3r dunno. cause ive got an esxi machine hosting 3 others that has 2 connecting and another not
05:57 fyb3r all minions are throwing that same error lol
05:57 fyb3r now the ones that can connect have this kernel version: 3.10.0-229.14.1.el7.x86_64
05:58 fyb3r nvm. i thought that was a link. but one of the minion failing just show me 3.10.0-229.14.1.el7.x86_64   :(
05:58 whytewolf :/ damn
05:58 mehakkahlon joined #salt
05:59 whytewolf tornado versions the same?
06:00 fyb3r nope i thought that was the issue but that seems to be negligible too. all are either 4.2.1 or 4.3. some of each are not connecting.
06:01 fyb3r one connecting doesnt even have m2crypto installed
06:03 fyb3r might just bite the bullet and scrap those minions and rebuild them lol
06:04 whytewolf lol. might be the simpler fix. since it seems to be minion side
06:05 fyb3r part of me has to know though
06:05 fyb3r lol
06:05 whytewolf lol. you tried the good old windows fix?
06:05 fyb3r whats that
06:06 whytewolf reboot
06:06 fyb3r lol yeah a few times
06:06 fyb3r but even the esxi machine has a minion going ape s on it
06:07 whytewolf I have no idea what is going on.
06:08 fyb3r it does seem like theres 2-3 extra minion processes running
06:09 fyb3r dunno if thats any indication
06:09 whytewolf interesting. should be 2 when idle
06:10 fyb3r every so often a 3rd will show up
06:10 Lionel_Debroux joined #salt
06:10 whytewolf hum. that might just be the connection attempt
06:11 fyb3r true
06:11 fyb3r well fml lol
06:12 whytewolf shut the minion down. delete the keys. start the minion backup?
06:14 fyb3r tried it. even went so far as uninstalling salt and then doing a find / -name salt and manually removing the folders and files
06:15 fyb3r including the python libs
06:15 whytewolf humm. that would imply it isn't in salt then.
06:15 fyb3r yeah thats what im assuming
06:15 fyb3r im thinking tornado is having a fit with something configured on the minions
06:15 keltim joined #salt
06:17 whytewolf possable. but i wouldn't know what.
06:19 fyb3r hm
06:19 fyb3r ill dig a bit deeper as i spin up new vms
06:24 whytewolf alright
06:26 Nazca joined #salt
06:33 felskrone joined #salt
06:34 hasues joined #salt
06:34 ubikite joined #salt
06:34 hasues left #salt
06:34 ubikite joined #salt
06:45 jhauser joined #salt
06:47 larsfronius joined #salt
06:54 ubikite joined #salt
07:00 rotbeard joined #salt
07:10 colttt joined #salt
07:10 otter768 joined #salt
07:23 yuhlw joined #salt
07:29 ubikite joined #salt
07:32 GreatSnoopy joined #salt
07:34 impi joined #salt
07:37 AndreasLutro joined #salt
07:39 rubendv joined #salt
07:40 KermitTheFragger joined #salt
07:51 cberndt joined #salt
07:59 NV joined #salt
08:05 fyb3r left #salt
08:09 eseyman joined #salt
08:09 CeBe joined #salt
08:10 golodhrim|work joined #salt
08:12 Guest55101 joined #salt
08:20 Guest1993 joined #salt
08:22 kawa2014 joined #salt
08:35 keimlink joined #salt
08:39 WildPikachu left #salt
08:43 Grokzen joined #salt
08:43 mehakkahlon joined #salt
08:43 thalleralexander joined #salt
08:54 opdude joined #salt
08:57 wych hi, I met a question with salt run in multi-enviroments, I am confused with states.highstate's behavior, this is my test, https://gist.github.com/MarloweW/76c0dbe2ee2de5ae4cd9 anyone can help ?
09:00 hemebond wych: You should just ask the question.
09:02 wych hemebond the question is states.highstate will run state files in all environment directories, I assume it will only run state files in base env only
09:03 AndreasLutro joined #salt
09:03 hemebond hmmm. Shouldn't your top.sls have targeting rules?
09:05 kawa2014 joined #salt
09:06 wych It has, check its content here https://gist.github.com/MarloweW/76c0dbe2ee2de5ae4cd9#files
09:07 stooj joined #salt
09:07 hemebond It would be best to test, but since you're putting all minions into both environments, you will get both states.
09:08 traph joined #salt
09:08 traph joined #salt
09:09 hemebond Oh, I see you're testing with environment parameter.
09:09 hemebond It applies both because you are targeting all minions in both environments.
09:10 hemebond There is a bit of an issue with the way states are applied and top.sls works. All top.sls files get merged together into one.
09:10 hemebond Then your targets are applied.
09:10 hemebond Or something like that :-)
09:11 otter768 joined #salt
09:12 wych Oh,so all states in different enviroments will be merged into one without saltenv parameter?
09:12 hemebond Yeah.
09:14 thefish joined #salt
09:15 wych It would be better if I could set default_env in master, since I reconfigure all minion's env config will be a huge work.
09:16 Grokzen joined #salt
09:16 hemebond What I've done is explicitly target them in the top.sls
09:17 hemebond My minion IDs are like FQDN, so I can target them quite easily.
09:19 opdude joined #salt
09:19 traph joined #salt
09:21 wych So do I,just I didn't put env in minion ID. I have got some hosts should be run in separate env even most of states are same with others. I thought configure different environments would be good choice.
09:23 Rumbles joined #salt
09:24 hemebond In your file_roots, you can have multiple directories.
09:25 hemebond For mine, I have the environment directory listed first, then the base or common directory. Allowing me to override states in a particular environment.
09:29 wych Most of my minions haven't configure environment, then states in some envs only but not in base environment will be applied into these minions.
09:29 wych That's not what I want, but I can still do something like this http://pastebin.com/5YwGpvB6
09:30 wych and use multiple directories.
09:30 hemebond Sorry, I had difficulty understanding your first comment there.
09:30 kbaikov joined #salt
09:31 hemebond http://pastebin.com/nU7aNan2
09:32 hemebond I have a top.sls in each environment directory except base.
09:32 hemebond When a highstate is run, the environment directory is checked first. If it's not there, it will look in base.
09:34 wych hemebond hmm... like I paste here https://gist.github.com/MarloweW/76c0dbe2ee2de5ae4cd9#files  dev.sls will be applied to all the hosts without environment specified, and It's not in base env directory.
09:35 ziro` joined #salt
09:35 hemebond Yes, your states have different names and your environments are being applied to all minions unless you specify the environment.
09:35 hemebond The environment system is odd at the moment and I believe there is some discussion about changes.
09:36 hemebond If you only want dev to apply (without having to specify an environment) then you need to change your targeting rules.
09:36 wych hemebond Yes, tha's what I am thinking, like this http://pastebin.com/5YwGpvB6 ?
09:37 hemebond Yes.
09:37 hemebond I thought you also wanted to fallback to the base env for some states (allowing environments to "override" a state) which is why I pasted by file_roots.
09:40 wych for most situation, yes. but this environment I am setting up requires no fallback ( I should have just setup a new salt-master and let it go...
09:41 dkrae joined #salt
09:44 denys joined #salt
09:46 elsmo joined #salt
09:58 rubendv joined #salt
09:58 rubendv joined #salt
09:58 rubendv #python
09:58 slav0nic joined #salt
10:02 Kurisutian joined #salt
10:03 s_kunk joined #salt
10:03 s_kunk joined #salt
10:04 Kurisutian Hi guys! Can anyone help me with how to compile salt as a user to the user directory. So far I compiled but when running salt-master I get "
10:04 Kurisutian File "./usr/local/bin/salt-master", line 6, in <module>
10:04 Kurisutian from salt.scripts import salt_master
10:04 Kurisutian ImportError: No module named salt.scripts
10:06 babilen Kurisutian: /usr/local is not the same as /home/$USER. How did you install salt and what are you trying to achieve?
10:06 ziro` joined #salt
10:08 ggoZ joined #salt
10:09 Kurisutian babilen: I did run "python2 setup.py --salt-pidfile-dir="/home/<user>/salt/run" install --root="/home/<user>/salt" --optimize=1" in order to compile salt as a user to his directory and adjusted the config files to keep it running strictly within the users home. Since our OPS team refuses to install packages and take care of the config files they suggested to install salt strictly as a user...#
10:10 babilen Sounds as if you have to talk to your OPS team as it is, IMHO, unacceptable to have a team that refuses to install packages
10:11 hemebond Couldn't you just create a virtualenv?
10:11 hemebond Does salt run in one easily?
10:11 babilen And then you might be looking for either "--home" or "--user" install flags for Python: https://docs.python.org/2/install/#alternate-installation-the-home-scheme
10:11 Kurisutian babilen: Trust me, there has been escalations, talking and whatever you can think of and they still get through with it. So that's why I thought "f*** you" and wanted to keep them out of this to not rely on them...
10:11 babilen Or install it in a virtualenv if you feel like it
10:12 babilen Kurisutian: That sounds horrible ... "Get off my lawn! You and your shiny toys!" :(
10:12 babilen Seriously, why are people this horrible?
10:12 babilen *some
10:14 babilen I would either install it with pip/bootstrap script into a virtualenv (also look into virtualenvwrapper in this context) or use the home scheme
10:14 Kurisutian babilen: it's simply stupid... they use puppet and for them it's the best tool according to them... well, they now refuse to install salt along with it since it would be to difficult to adapt puppet to that... so much for this... they regularly do this with all kind of tools and agents...
10:14 Kurisutian babilen: is there any documentation for this? Never done this before... ;-)
10:15 babilen Sounds as if they are afraid of change and want to prevent a switch to salt simply by making it harder to adopt
10:15 babilen Documentation for what exactly?
10:18 Kurisutian babilen: Yeah, I guess so... atrificially increasing their value seems to be the reason for that. Docu for the bootstrapping method (virtualenv)... what packages do I need for that to create this environment? What does it do and are there any drawbacks with such a virtual environment?
10:22 babilen https://virtualenv.readthedocs.org/en/latest/ + http://virtualenvwrapper.readthedocs.org/en/latest/ is what is meant by "virtualenv"
10:22 babilen The bootstrap script is https://github.com/saltstack/salt-bootstrap
10:23 Kurisutian Awesome, thank you very much! :)
10:25 Kurisutian babilen: One more question: Does the bootstrap script require root rights? I can only do all operations as a user with no root rights...
10:26 babilen I'm not really using it
10:27 babilen TIAS, the "--home" scheme or direct install in virtualenv should work. I guess you want to enable some "pip" stuff in the bootstrap script
10:27 Kurisutian Oh, OK... seems to me that it "just" handels to find out the method of choice accorting to the given OS and goes on there... as root... ;-)
10:29 Kurisutian I'll check with the virtual env and see what I can get running there...
10:29 babilen Just configure virtualenvwrapper, run "mkvirtuartualenv salt" and install via pip.
10:29 babilen no
10:29 babilen "mkvirtualenv salt" that is :)
10:29 babilen speling hälps
10:31 giantlock joined #salt
10:35 human01d joined #salt
10:37 human01d when I call #salt "minion" state.show_sls i1 saltenv=test
10:37 human01d - No matching salt environment for environment 'test' found
10:37 human01d - No matching sls found for 'i1' in env 'test'
10:38 babilen human01d: Why do you think that a "test" salt environment exists?
10:38 human01d but in file_roots: i target test: to directory
10:38 babilen do you?
10:38 human01d file_roots:     base:         - /srv/salt/base     test:         - /srv/salt/test
10:39 babilen Please use a pastebin and paste your configuration in its entirety (modulo comments), your command(s) and their output to one of http://refheap.com, http://paste.debian.net, https://gist.github.com, http://sprunge.us, http://dpaste.de, … please
10:39 ziro` joined #salt
10:40 babilen Also make sure to restart the salt-master before running your commends
10:42 human01d i restarted minion and master and right now it works )
10:43 babilen So you didn't restart your master after changing its configuration?
10:44 cal joined #salt
10:46 Guest25444 Hi all! Im struggling with file.uncomment. Im trying to uncomment two lines in a configuration file. however anything i do desn't lead to the desired result.
10:47 Guest25444 what i initially tried was to match both lines in this way:  regex: 'acl_(smtp_predata|not_smtp) = daily_limit_acl'
10:47 human01d babilen: i though i have done it. saltenv makes me crazy. I will try it in another environment and then write you back.
10:47 laidback joined #salt
10:48 skrobul Guest25444: how does actual line in file look like now?
10:48 Guest25444 which however resulted in thsi diff which is obviously not what i intended:
10:48 Guest25444 -# acl_smtp_predata = daily_limit_acl                   -# acl_not_smtp = daily_limit_acl                   +acl_(smtp_predata|not_smtp) = daily_limit_acl                   +acl_(smtp_predata|not_smtp) = daily_limit_acl
10:48 Guest25444 grml
10:48 ziro` joined #salt
10:48 amcorreia joined #salt
10:49 Guest25444 https://gist.github.com/mnaef/ce9bfca04da268f64bee
10:51 rocket_ joined #salt
10:57 human01d try to comment with {#   #}
10:59 human01d babilen: i removed file_roots with enries and wrote them again. after that saltenv works as a clock :)
10:59 AndreasLutro Guest25444: that seems like a bug
11:01 skrobul Guest55101: it works for me - https://gist.github.com/skrobul/97a992491a3efca6f09d
11:01 skrobul which version are you on?
11:01 cal2 joined #salt
11:01 cal2 rehi, sorry, my browser crashed
11:02 cal2 update on https://gist.github.com/mnaef/ce9bfca04da268f64bee:
11:02 skrobul cal2: it works for me - https://gist.github.com/skrobul/97a992491a3efca6f09d
11:03 cal2 the problem seens to be the fact, that the salt master ist running 2015.8.1 and this specific minion 2015.5.3
11:04 skrobul try to execute this module locally on the minion and see if it still happens
11:04 cal2 thanks skrobul, i will try
11:04 mage_ hello, any idea why with the following https://dpaste.de/DHyr when I state.show_sls the context is null ?
11:05 AndreasLutro your indentation is wrong mage_
11:05 mage_ ah.. my bad
11:05 mage_ thanks
11:05 babilen mage_: Where can I find the code of bbpf_apache.config ?
11:06 mage_ babilen: it's a custom function of mine
11:06 babilen It might return no data and you want extra spaces on line 19 I guess
11:06 mage_ babilen: https://dpaste.de/r1ys
11:07 bhosmer joined #salt
11:08 mage_ mmh with an extra space I have an error
11:08 mage_ https://dpaste.de/Y8CN
11:09 AndreasLutro your indentation is probably still wrong
11:09 mage_ yep (: but I don't find where
11:09 AndreasLutro well, share your updated sls
11:10 babilen mage_: And I misread it, you don't need an extra space on line 19
11:10 mage_ https://dpaste.de/B90a
11:10 mage_ ah
11:10 AndreasLutro er
11:10 AndreasLutro you do, but you need the same on line 17
11:10 mage_ why ? :)
11:11 AndreasLutro http://yaml-online-parser.appspot.com/?yaml=-+foo%3A%0A++bar%3A+baz&amp;type=json
11:11 babilen Because it would work then?
11:11 mage_ AndreasLutro: same error (:
11:11 AndreasLutro vs http://yaml-online-parser.appspot.com/?yaml=-+foo%3A%0A++++bar%3A+baz&amp;type=json
11:11 babilen With spaces on 17 and 19?
11:11 mage_ and .. 16 :)
11:11 babilen Context variables are indented by four spaces
11:11 AndreasLutro did I say 16?
11:11 AndreasLutro ;)
11:11 mage_ mmh no
11:12 mage_ ok it works now, thanks !
11:12 babilen https://docs.saltstack.com/en/latest/ref/states/all/salt.states.file.html (first example)
11:12 otter768 joined #salt
11:13 mage_ ok I missed the "The defaults and context arguments require extra indentation (four spaces instead of the normal two) in order to create a nested dictionary. More information." part
11:13 mage_ perhaps the "four spaces instead of the normal two" should be noted in bold :)
11:14 babilen The parser in Python has a number of idiosyncrasies
11:14 babilen I'm always surprised how many people use "context" in salt .. I have such a strong preference for performing the lookup in the template that I don't quite understand why you wouldn't always do that unless you depend on context information in the state itself
11:14 AndreasLutro it's not just the python parser
11:14 AndreasLutro this is the yaml standard
11:15 mage_ babilen: what do you mean by "the lookup in the template" ?
11:15 mage_ I'm doing this in my template
11:16 babilen mage_: It looks like a state to me ;)
11:17 mage_ this is what my template looks like https://dpaste.de/s5pa
11:17 babilen mage_: Yeah, and you perform https://dpaste.de/DHyr#L19 *outside* of it.
11:17 babilen https://dpaste.de/DHyr#L5,19 that is
11:18 AndreasLutro templates should be dumb is a developer rule of thumb
11:18 AndreasLutro I apply it to salt as well, to some extent
11:18 mage_ babilen: I try to keep my templates as clean as possible
11:18 mage_ babilen: and I'm doing like this because then I can iterate on it with my pillar data
11:18 babilen sure
11:19 mage_ :)
11:19 babilen It is fine in this example because you rely on context in the state itself (you generate multiple files automatically) and therefore need the context in which each file is being generated
11:33 babilen AndreasLutro: I can understand that if the backend is powerful, but in salt you use jinja in both cases. You are essentially deciding if you want users of the template to always have to specify the context or use more logic in the template to make them more versatile.
11:35 AndreasLutro it's not really related to the power of the backend, just separation of concerns
11:35 babilen If this would be a web framework in which you use templates you'd have all power outside it and would like to restrict templates to a few loops and *maybe* a conditional here and there, but this doesn't necessarily work in salt as the "backend" is not powerful at all.
11:37 kawa2014 joined #salt
11:38 babilen My preferred way of "templating" is to not use a template at all, but to simply hand suitable data to a generator and write the file directly.
11:52 MadHatter42 joined #salt
11:53 ponpanderer joined #salt
11:54 laidback Hello, could someone explain me the concept of envs in salt? Is it meant to have a separation of your servers in environments like having a dev/stg/prd server for e.g. a webapplication and treat it like that with different configs for e.g. logs of the server or is it meant to separate servers for salt development, like having a salt dev environment for formulas in dev/stg/prd?
11:54 ponpanderer @laidback: the former not the latter
11:57 mage_ laidback: it is usefull if, for example, you want to deploy a webapp in production or test mode. For example here we create a new jail when the webapp is deployed in "dev" mode, and everything is installed in
11:58 ponpanderer anyone seeing issues with local/default master job cache in 2015.8 spitting out "Could not write out jid file for job X" in a loop when running runner commands like jobs.print_job? I see this randomly on a 2015.8.1 salt-master. Adding more debugging the IOerror is actually that is can't find the file/directory for the job cache. Any idea what could be causing that?
11:59 ponpanderer running salt-master as not the root user, following standards for doing so and not seen this in any release until 2015.8
12:01 fredvd joined #salt
12:02 laidback Ok, thanks. Then my next question relates to your formula usage when you keep them in git. In that case branches/tags are mapped to the environments. But for me e.g. a different config for a specific env comes from using a dev/stg/prd file which is rendered dynamically. Do i really need dev/stg/prd branch of a formula having the same content?
12:03 babilen Use "base" for common things, but yes: You'll need a branch each
12:03 laidback I would rather have only one master branch for production formulas and having e.g. develop for formulas in development which get tested with serverspec/vagrant/jenkins before going to production/master
12:03 babilen (which is why the mapping in GitFS sucks)
12:04 laidback Good to hear, i was under the impression i am missing some internal knowledge how to get things configured properly. But in that case you are right.
12:06 babilen You definitely want to keep your top file(s) in a separate repository when you are using environments as they are being merged which would make "normal" git workflows rather tricky
12:13 mbrgm joined #salt
12:27 Vaelatern joined #salt
12:30 denys joined #salt
12:36 cpattonj joined #salt
12:38 cpattonj joined #salt
12:49 larsfronius joined #salt
12:51 _JZ_ joined #salt
12:58 TooLmaN joined #salt
12:59 ziro` joined #salt
13:01 nicksloan so, upon further testing, git_pillar just works nothing like normal pillars or gitfs where it concerns top files.
13:03 cpattonj_ joined #salt
13:04 nicksloan every git_pillar remote must have its own top file, and it appears that the pillar environment name assigned to a remote is only valid while processing that remote's top file. I assume other environments identified in pillar_roots are also recognized.
13:11 tracphil joined #salt
13:12 breakingmatter joined #salt
13:13 otter768 joined #salt
13:15 mbrgm left #salt
13:18 bluenemo joined #salt
13:26 nicksloan This means that if you use a development, staging, production branching strategy, you can keep all of your environments in your top file, and only the environment for the branch you're in will really matter.
13:26 PhatPete joined #salt
13:29 PhatPete Good morning! I am hoping to get some advice regarding docker + salt. I am using 2015.8.1 and the Docker CMD is ["/usr/bin/salt-minion"].  I am getting exactly two salt-minion zombies inside the container every time I do... anything.
13:31 subsignal joined #salt
13:33 deus_ex joined #salt
13:34 AndreasLutro PhatPete: maybe the salt install process starts a minion for you
13:36 DammitJim joined #salt
13:40 mik__R joined #salt
13:44 johtso joined #salt
13:50 lionel joined #salt
13:51 fredvd joined #salt
13:53 mehakkahlon joined #salt
13:54 antonw joined #salt
13:55 furrowedbrow joined #salt
13:56 tracphil joined #salt
14:01 tracphil Something along these lines is what I used to use https://gist.github.com/tracphil/8cef3ea97bd82ed4c2e1 but now I am looking into reclass for an ENC http://reclass.pantsfullofunix.net along with gitfs and 4 environments
14:01 deus_ex joined #salt
14:02 JDiPierro joined #salt
14:04 JDiPierro joined #salt
14:04 danielcb joined #salt
14:06 Kelsar joined #salt
14:10 numkem joined #salt
14:10 racooper joined #salt
14:13 mapu joined #salt
14:15 PhatPete joined #salt
14:16 pdayton joined #salt
14:22 hasues joined #salt
14:22 hasues left #salt
14:23 pyropoptrt joined #salt
14:26 bhosmer joined #salt
14:26 Deevolution joined #salt
14:31 techblaze joined #salt
14:31 scoates joined #salt
14:32 ubikite joined #salt
14:33 ubikite joined #salt
14:37 dthom91 joined #salt
14:37 analogbyte joined #salt
14:39 traph is there a way to use salt modules in states?
14:40 AndreasLutro traph: yes, how exactly do you want to use them?
14:40 traph writing a state using the mongodb module, for example
14:40 AndreasLutro using how
14:40 mik__R joined #salt
14:41 AndreasLutro the way you use salt modules in states varies depending on exactly what you want to do
14:42 traph basically adding users and roles
14:42 AndreasLutro well I don't know about roles but there is https://docs.saltstack.com/en/latest/ref/states/all/salt.states.mongodb_user.html
14:43 JDiPierro joined #salt
14:43 traph I'm looking at https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.mongodb.html
14:44 mik__R_ joined #salt
14:44 AndreasLutro I think translating those into states is going to be a bit difficult
14:45 AndreasLutro maybe poke this feature request issue https://github.com/saltstack/salt/issues/18757
14:45 zionsofer joined #salt
14:46 traph I found this https://gist.github.com/benhosmer/3749823
14:46 zionsofer I have a question regaring salt cloud: is it possible to make a spot request for an instance count larger than 1 with the ec2 driver?
14:47 AndreasLutro traph: yeah, that's possible, but there are very few scenarios where module.run in a state is actually a good idea
14:47 traph AndreasLutro, why is that?
14:47 AndreasLutro because module functions are not written to be part of an automated state run, so to speak
14:47 AndreasLutro like, there's a mongodb module function to add a user
14:48 AndreasLutro with a state you want to check if it exists and only if it doesn't, add it
14:48 Ph-x joined #salt
14:50 traph AndreasLutro, from what I get, every state has to be "aware" of its "state". as long as I can make it "aware", I don't see why there should be a problem
14:50 zer0def joined #salt
14:51 AndreasLutro in your words, I don't know a good way to make it "aware"
14:51 AndreasLutro without writing your own custom state module
14:51 AirOnSkin Heya, with salt.states.user.present you can set a password. It states, that it needs a password hash. What kind of hash are we talking about here?
14:52 quix joined #salt
14:52 AndreasLutro AirOnSkin: hash strings as found in /etc/shadow
14:52 AndreasLutro you can generate password hashes with https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.shadow.html
14:52 AirOnSkin AndreasLutro: I see, great, thank you :)
14:52 traph AndreasLutro, I see your point
14:53 AndreasLutro traph: for what it's worth, the mongodb state modules look very simple, making a pull request for it shouldn't be super hard if you want to take a shot at it
14:54 cpowell joined #salt
14:55 traph AndreasLutro, you mean this one https://github.com/saltstack-formulas/mongodb-formula ?
14:55 AndreasLutro traph: https://github.com/saltstack/salt/blob/develop/salt/states/mongodb_user.py https://github.com/saltstack/salt/blob/develop/salt/states/mongodb_database.py
14:55 mehakkah_ joined #salt
14:57 favadi joined #salt
14:59 traph AndreasLutro, this module doesn't have the option to add roles to users, so I'd have to add a custom salt state for that
14:59 AndreasLutro traph: maybe, yeah, or maybe just add a roles arg to the present function
15:00 babilen traph: My suggestion would be to "fork" states/mongodb_user.py, implement a suitable state, put it in _states and to submit a PR to saltstack
15:01 babilen Probably best to just add it to mongodb_user.present(..)
15:01 winsalt joined #salt
15:02 traph babilen, will try
15:02 traph thanks guys
15:02 mission712 joined #salt
15:02 zmalone joined #salt
15:02 babilen That's the cleanest solution I can think of and if you submit a PR the likelihood is high that it makes it into "proper" salt soon
15:02 babilen (which also helps everyone)
15:04 techblaze joined #salt
15:05 bhosmer joined #salt
15:06 Deevolution joined #salt
15:06 PhatPete joined #salt
15:07 AirOnSkin AndreasLutro: Yeah, we already do that. But a colleague of mine complained that it wouldn't work. I just checked it and it does work. Don't know what his problem was ;)
15:08 andrew_v joined #salt
15:11 PhatPete Aaah! I figured out the problem.
15:13 DammitJim joined #salt
15:13 favadi joined #salt
15:14 otter768 joined #salt
15:15 Heartsbane joined #salt
15:17 andrew_v joined #salt
15:17 Ph-x joined #salt
15:21 analogbyte joined #salt
15:28 Ph-x joined #salt
15:32 canci joined #salt
15:36 cro joined #salt
15:40 clintberry joined #salt
15:41 traph is there a valid example of preexisting state modules using python tuples as arguments?
15:51 bhosmer joined #salt
15:54 mr-op5 joined #salt
15:54 denys joined #salt
15:55 Brew joined #salt
15:58 PeterO_ joined #salt
15:59 Brew1 joined #salt
15:59 dthom91 joined #salt
16:10 zzzirk joined #salt
16:11 viq joined #salt
16:12 jjc001 joined #salt
16:12 ajw0100 joined #salt
16:13 josuebrunel joined #salt
16:13 akhter joined #salt
16:13 akhter Hey guys, having an issue with salt-cloud.
16:14 akhter I created a profile and when I try to map to it, it states that the profile doesn't exist.
16:14 armguy Wanted to make sure I am understanding proxy minions properly before I try and see if this will do what I need. I have some devices that could install a minion but have no method to get access to the actual salt-master in my infrastructure, There is one main box that does have access to the master and the ones that do not have access to this one that can reach the master. Would proxy minions be the right thing to look into for getting the
16:14 armguy ones that can't reach the master?
16:17 breakingmatter I can't seem to get pip.installed to work with `- bin_env: /usr/local/bin/pip3`. Gives me an error that says a pip module could not be found. Any ideas?
16:23 Brew joined #salt
16:26 kawa2014 joined #salt
16:27 Voziv joined #salt
16:27 ggoZ joined #salt
16:28 nethershaw joined #salt
16:29 yuhlw joined #salt
16:32 scoates joined #salt
16:34 ziro` joined #salt
16:35 Brew1 joined #salt
16:36 ageorgop joined #salt
16:36 larsfronius joined #salt
16:36 debian112 joined #salt
16:39 adelcast I am using 2015.8 and I see mine events all the time (_schedule.conf file scheduling mine events every 60 minutes is automatically created in my minion.d folder)
16:39 akhter joined #salt
16:40 adelcast I really don't want to use the mine, is there a way to disable it other than using disable_modules: -mine ?
16:40 keimlink joined #salt
16:41 dthom91 joined #salt
16:41 akhter adelcast: You can just comment the mine_interval on the minion config
16:43 adelcast the thing is that the minion on startup is automatically creating a _schedule.conf file with:
16:43 Brew joined #salt
16:43 adelcast schedule:
16:43 adelcast __mine_interval: {function: mine.update, jid_include: true, maxrunning: 2, minutes: 60}
16:43 adelcast as far as I know I am not setting the mine anywhere, so I believe it's a default (I might be wrong on this...)
16:45 jjc001 Hi I am running a salt minion which is transferring six files from master using file.managed state and the minion consumes all the memory on the VM as the file.managed is getting processed. Any ideas of where I could look?
16:45 whytewolf adelcast: mine does not have any default functions set. it has to be set somewhere. either on the minion or in a pillar
16:46 whytewolf oh wait. that is a scedule for mine. not a mine function
16:46 jfindlay I hate that pip bin_env argument.  It is too ambiguous and, clearly, is buggy
16:46 akhter whytewolf: https://docs.saltstack.com/en/latest/topics/mine/
16:46 akhter Yeah
16:47 akhter Still having issues with salt-cloud not locating my profile.....
16:47 akhter The syntax is correct.....
16:48 adelcast yeah, the schedule is the one that seems to be configured by default
16:48 whytewolf adelcast: so comment out the mine_interval settings and use https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.schedule.html#salt.modules.schedule.disable to disable it
16:49 stevednd basepi: you around?
16:49 winsalt joined #salt
16:51 ksj hi, I'm trying to manage user groups via group.present: members, and so when I create a user.present I don't have a groups option. however, salt seems to assume that if I don't pass a list of groups in user.present, then I'm passing an empty list, and subsequently it removes the user from all groups. so something like this (http://dpaste.com/0AR75E2) will add the user to the group deploy, then immediately
16:51 ksj remove him. this feels like a bug
16:51 aparsons joined #salt
16:51 adelcast so, that works, the problem is that on connect it will write the schedule again...it would be nice if there was a config that would disable the feature
16:51 adelcast I'll do some more testing with a vanilla install to make sure that is the case, then I might submit a bug report
16:52 troyready joined #salt
16:52 whytewolf akhter: some things to check that you might have already checked. permissions on the file, selinux blocking the reading of the file. profile in the correct location /etc/salt/cloud.profiles.d
16:52 akhter whytewolf: Everyone has read permissions, I've never ran into an SELinux issue before though.
16:52 akhter The profile is in the correct location with correct syntax and filename *.conf
16:53 basepi stevednd: yessir
16:53 nsenno joined #salt
16:55 nsenno Hi all
16:56 whytewolf akhter: if you use -l debug with the salt-cloud command does it say anything about "/etc/salt/cloud.profiles.d/your profile file"
16:56 akhter whytewolf: Checking
16:56 nsenno some of you has encountered this issue before:
16:56 nsenno File "msgpack/_packer.pyx", line 223, in msgpack._packer.Packer.pack (msgpack/_packer.cpp:223)
16:56 nsenno TypeError: can't serialize datetime.date(2007, 7, 6
16:56 stevednd basepi: I just pulled develop, and am trying to fix the rbenv.py module. recently there were some changes by terminalimage to use shlex.split in `do_with_ruby`. I'm not familiar with the function, but the `cmdline` arg is coming in as an array already ['gem', 'list', 'bundler'], but once shlex.split gets a hold of it it is now ['[gem,', 'list,', 'bundler]']
16:56 nsenno on centos ?
16:57 aparsons joined #salt
16:57 stevednd but wait, there's more! by the time it executes in the `do` function it looks like so: Executing command ['[[gem,,', 'list,,', 'bundler]]']
16:57 akhter whytewolf:
16:58 stevednd that gives a command execution error of unable to run command, "No such file or directory"
16:58 akhter https://gist.github.com/AkhterAli/d268b97ea2c1d6c544d2
16:58 akhter So it's reading the file.
16:58 whytewolf akhter: it is reading the profile.... but there isn't any providers
16:59 stevednd my question I guess, is is that looks like a disasterpiece, and I can't imagine that the command running is supposed to be looking like that. I don't know enough about the command running to know if that's true or not
16:59 akhter whytewolf: the provider is in /etc/salt/cloud
16:59 akhter I've updated the gist, full details there.
17:00 sunkist joined #salt
17:00 whytewolf humm, I take it test.conf has a profile for akhter?
17:00 akhter Yes.
17:00 skepner_ joined #salt
17:01 MadHatter42 joined #salt
17:02 kulty joined #salt
17:02 akhter whytewolf: The syntax is across the provider, profile, and the map file.  All 4 space indentation.
17:03 teryx510 joined #salt
17:03 teryx510 left #salt
17:03 nsenno better if I use a gist: https://gist.github.com/nicolasenno/6b4944f51f84214ae20b
17:03 whytewolf humm okay.
17:04 stevednd basepi: just confirmed. If I guard shlex.split with `if isinstance(cmdline, basestring)` the commands execute properly
17:04 basepi stevednd: that sounds like the correct fix
17:05 basepi if it's already a split command then you're good.
17:05 whytewolf akhter: it doesn't have any errors on the load. which defintly means the syntax is at least ok. however it isn't reconizing the name of the stanza which means that doesn't match. the files aer not utf-8 are they? or have extra chars in them?
17:06 Fiber^ joined #salt
17:06 stevednd basepi: shlex.split is used in 45 places currently. Maybe it would be worth writing a util function that has that check builtin so that it doesn't always have to be remembered to not try to shlex.split a list?
17:06 basepi That's actually not a bad idea at all.
17:07 keltim_ joined #salt
17:07 whytewolf akhter: try using salt-cloud --list-profiles=<provider that should match akhter>
17:07 basepi We sprinkled those in wherever user args were used in commands to prevent command-injection vulnerabilities.
17:08 whytewolf akhter: sorry i meant salt-cloud --list-profiles=all
17:08 whytewolf I am way to tired today
17:09 akhter whytewolf: They aren't UTF-8, same number of characters.
17:09 writtenoff joined #salt
17:10 akhter whytewolf: That's gotta be the issue.  https://gist.github.com/AkhterAli/d268b97ea2c1d6c544d2
17:10 akhter Err, wrong.
17:10 akhter I provided the provider, not the profile.
17:11 whytewolf akhter: yeah that was my bad. just use all should list all profiles
17:12 akhter Still no go.
17:12 whytewolf humm.
17:12 akhter whytewolf: With 'all' specified, it can't find my test profile.
17:13 akhter I don't know if there's a place to specify the cloud profiles location, I thought salt read all configs as one document.
17:13 whytewolf akhter: it is reading the file.
17:13 Gareth o/
17:13 akhter But it's not finding the particular profile name I have stated in the file.
17:14 whytewolf akhter: yes. something isn't correct about the contents of that file
17:14 whytewolf there isn't any whitespace before the name of the profile is there?
17:14 aparsons joined #salt
17:15 akhter whytewolf: I've updated the gist with the profile.
17:15 akhter And no whitespace.
17:15 otter768 joined #salt
17:15 whytewolf akhter: this shouldn't matter but it is worth a try. try using 2 space indentation instead of 4
17:16 akhter I tried that once before but I'll give it a go again.
17:16 akhter Still no go, I changed that syntax for provider, map, and profile.
17:17 whytewolf is the provider showing up when you salt-cloud --list-providers
17:18 akhter Strange, it isn't.
17:18 akhter I'll move that to cloud.providers.d
17:19 stevednd basepi: https://github.com/saltstack/salt/pull/28702 that's for my rbenv fix. When is the next release planned to go out?
17:19 malinoff joined #salt
17:19 akhter There we go.
17:19 basepi From develop? February, I think?
17:19 whytewolf akhter: provider is showing now?
17:20 akhter Ec2.
17:20 akhter Which is what I want.
17:20 akhter There we go.
17:20 akhter My profile finally works.
17:20 akhter Adding to the end of /etc/salt/cloud isn't good enough apparently.
17:21 akhter Perhaps there's something on the default /etc/salt/cloud that is preventing something.
17:21 whytewolf now that it can find the provider. could have sworn there were errors that popped up if a profile couldn't find a provider though
17:21 akhter Might need to create a github issue.
17:21 stevednd basepi: any chance the change could go out sooner on 2015.8, since it fixes a bug in the current rbenv behavior?
17:22 stevednd also, where's the best place to put the shlex.split util function?
17:22 whytewolf I wouldn't know about using /etc/salt/cloud. I know it was the old way of doing it. but have never used it.
17:22 akhter whytewolf: Thanks for your help, my instance is finally starting.  I'll re-create the issue and post it on their github page.
17:22 akhter According to their docs, I should be able to use /etc/salt/cloud and have my provider info there.
17:27 bhosmer joined #salt
17:33 tiadobatima joined #salt
17:36 impi joined #salt
17:36 Nikatjef joined #salt
17:38 akhter_1 joined #salt
17:41 stevednd basepi: https://github.com/saltstack/salt/pull/28704 that is a first pass at guarding shlex.split. Can you take a look at it, as I'm not entirely familiar with how it should operate.
17:42 londo joined #salt
17:50 dthom91 joined #salt
17:55 kawa2014 joined #salt
17:59 jalbretsen joined #salt
18:01 conan_the_destro joined #salt
18:02 tiadobatima joined #salt
18:04 tuxx joined #salt
18:06 markm joined #salt
18:07 pdayton joined #salt
18:08 jdubski_ joined #salt
18:08 basepi stevednd: definitely. Feel free to submit against 2015.8. And I'll take a look at your pr.
18:09 Destreyf joined #salt
18:09 colegatron joined #salt
18:11 nidr0x joined #salt
18:13 moogyver joined #salt
18:15 pdayton joined #salt
18:15 bhosmer joined #salt
18:15 fyb3r joined #salt
18:18 dthom91 joined #salt
18:20 tivvit joined #salt
18:21 stanchan joined #salt
18:21 mdupont joined #salt
18:21 Viper233 joined #salt
18:22 Viper233 left #salt
18:23 enigma99a joined #salt
18:24 cberndt joined #salt
18:25 tanta_g joined #salt
18:26 rm_jorge joined #salt
18:29 mitsuhiko joined #salt
18:31 dthom91 joined #salt
18:31 Voziv left #salt
18:33 Tween joined #salt
18:33 Tween Greetings
18:34 Antiarc left #salt
18:36 fyb3r hello
18:37 Tween I'm having a problem  on 2015.5.2 in regards to pkg management. I'm using the sources directive, but I seem to always get the error "Only one of "pkgs" and "sources" is permitted
18:37 larsfronius joined #salt
18:38 Tween This is on Ubuntu 14.04
18:38 enigma99a does anyone here have any config examples for setting up a MoM. All I could find is Syndic stuff in the docs
18:39 dthom91 joined #salt
18:41 ageorgop joined #salt
18:41 Tween My problem state looks similar to this: http://codepaste.net/azensz
18:41 ubikite joined #salt
18:43 iggy Tween: what happens if you take out the require?
18:43 big_area joined #salt
18:44 Tween pkgs contains ['update-docker'] and sources containers [OrderedDict([('lxc-docker-1.3.1'], 'http://......./lxc-docker.deb')])] getting sent to _find_install_targets
18:44 Tween One sec, I'll try
18:46 Tween I commented out the requires, verified in the log that the rendered yaml has # in front of the require, same error
18:48 giantlock joined #salt
18:51 ajw0100 joined #salt
18:51 enigma99a does every master connecting to another master have to be a syndic?
18:52 miked510 joined #salt
18:52 cliluw joined #salt
18:52 fyb3r no. masters can be a minion of another master
18:53 dthom91 joined #salt
18:54 Tween iggy: I took everything out of the sls file except that one state, and it fails with the error only one of ... etc
18:54 enigma99a fyb3r: Thanks. What about if you want to issue a command from the master of master to a node connected a master below -- will it be able to communicate without being a syndic?
18:55 fyb3r the only way ive found to do so is to issue a bash command through salt to the minion/master and have it execute a salt command to its minions
18:55 fyb3r kind of a pain but it works
18:56 iggy Tween: I'm pretty sure someone else has reported that before, but I can't find a ticket for it
18:57 iggy Tween: might check the irc logs and if you can't find anything, open a ticket
18:57 enigma99a hmm thanks.. I was thinking about Mine too, and if there was a way to get all of the data in one place
18:57 Tween iggy: will do thanks
18:58 iggy enigma99a: there are open ticekts about supporting mine in multi-master environments... short answer... not yet
19:00 enigma99a iggy thanks
19:02 TooLmaN joined #salt
19:02 LotR are there tools to turn pillar data into an xml config file?
19:03 iggy augeas?
19:06 opensource_ninja joined #salt
19:06 jaybocc2 joined #salt
19:08 LotR hmm. I wasn't thinking that sophisticated a tool.
19:08 LotR but maybe I should have
19:08 ziro` joined #salt
19:13 denys joined #salt
19:15 VR-Jack joined #salt
19:16 otter768 joined #salt
19:16 josuebrunel joined #salt
19:18 Ahlee pillar to xml? do you hate yourself that much
19:18 teryx510 joined #salt
19:19 LotR no, I just want to stick calendarserver config it the pillar
19:24 controversy187 joined #salt
19:25 tanta_ge joined #salt
19:27 sunkist joined #salt
19:35 deedubs LotR: why not just write a jinja template?
19:35 fyb3r or python. just import the yaml and dump it
19:36 LotR deedubs: I was thinking if there was something that actually understands xml, there'd be less risk of generating something invalid
19:38 ageorgop joined #salt
19:39 ggoZ joined #salt
19:44 MindDrive joined #salt
19:46 stevednd basepi: should I do a separate pull request from #28702 that is targeted at 2015.8, or should I modify 28702?
19:48 Cruz4prez joined #salt
19:49 hayer left #salt
19:51 basepi stevednd: actually, we can probably just backport it.
19:52 basepi Or you can resubmit against 2015.8. Your call.
19:57 miked510 joined #salt
19:57 jasondotstar joined #salt
19:59 londo joined #salt
20:01 londo joined #salt
20:03 stevednd if I just leave it how it is is there any special process to get it backported, or will it just magically find its way in?
20:03 Tween iggy: work around for the sources thing seems to be -pkgs: null
20:03 Tween in the state that uses sources
20:04 iggy Tween: interesting... definintely file a bug
20:04 stevednd basepi: also, did you see my question on the issue about using the util vs checking the type in the module?
20:07 keimlink joined #salt
20:08 pppingme joined #salt
20:09 basepi stevednd: the util function will be fine.
20:13 enigma99a hey iggy... without giving away too many details, are you guys running bash commands from the MoM to your other masters to execute commands?
20:13 forrest joined #salt
20:13 iggy enigma99a: yes
20:14 enigma99a alright, if that's how everyone is doing it, I will do so also :)
20:15 amcorreia joined #salt
20:16 bluenemo joined #salt
20:16 marmota_temivel joined #salt
20:17 alvinstarr joined #salt
20:17 forrest iggy, Have you run into errors with the salt-api where it throws back a 'method not allowed' error from the cherrypy server?
20:18 marmota_temivel left #salt
20:19 iggy forrest: no, but it sounds familiar... checked for bugs?
20:19 forrest ugh not yet I didn't
20:19 iggy (might also check the saltpad issues if you don't find it in salt)
20:19 dthom91 joined #salt
20:20 forrest Hmm, some guy commented 6 days ago it was working fine.
20:21 bhosmer_ joined #salt
20:21 pdayton joined #salt
20:24 racooper howdy. I'm looking for an example of using pkgrepo.managed with yum that uses an "exclude" line via kwargs.  I just need the syntax.
20:25 pcdummy joined #salt
20:25 pcdummy joined #salt
20:25 forrest racooper, Should just be - exclude: my_shitty_packages
20:26 racooper oh, that's too easy!
20:26 forrest racooper, I'd say just test it out, worst case you break the repo file, no biggie.
20:26 racooper I just wanted to make sure I wasn't missing something first. yeah, it's fine, thanks!
20:27 sunkist joined #salt
20:29 quix joined #salt
20:30 pgoetz joined #salt
20:32 racooper forrest, that worked exactly. thanks.
20:34 forrest np
20:37 bhosmer_ joined #salt
20:39 denys joined #salt
20:39 forrest What a great debug message: [INFO    ] 71.212.71.139 - - [09/Nov/2015:15:35:44] "GET /hook/travis/success-test HTTP/1.1" 405 751 "" "curl/7.35.0"
20:42 pppingme joined #salt
20:44 dthom91 joined #salt
20:47 s_kunk joined #salt
20:54 Drew__ joined #salt
20:55 Drew__ hey
20:55 Drew__ I am new to Saltstack.
20:56 Drew__ I was hired on a condition that I will have to work with saltstack
20:56 Drew__ I agreed and now looking forward to learn until I become an expert
20:57 Drew__ Please guide me though
20:57 hemebond Drew__: That was your condition?
20:57 Drew__ yeah
20:57 hemebond But you haven't used it?
20:57 Drew__ I mean I agreed to learn and work
20:57 robawt Drew__: start with the tutorial on the documentation page.
20:59 Drew__ okay
20:59 hemebond Jealous.
20:59 Drew__ what??
20:59 hemebond Here I can punching a rushed Puppet-based project in the face.
21:00 Drew__ Is there anything except the salt docs I need to go through?
21:00 hemebond Tutorial is a good start.
21:00 Drew__ I want to lean it from really basics
21:01 winsalt yep, go through the official docs, its really the only comprehensive place to learn salt
21:01 babilen https://docs.saltstack.com/en/latest/topics/tutorials/index.html -- just keep reading and playing
21:01 Drew__ Yeah.. I am on the same page rn
21:02 forrest Drew__, Go through the docs, it's just like when you learn anything else.
21:02 babilen http://shop.oreilly.com/product/9781784394608.do is worth a read too, followed by https://www.packtpub.com/networking-and-servers/mastering-saltstack
21:02 Drew__ So, while learning and working at company, if I have doubts, I can ask here.... right?
21:02 babilen Sure, that's what this channel is for
21:02 forrest Drew__, You can ask, but please don't abuse the IRC, none of us are paid to be here, so please search the docs first.
21:02 whytewolf Drew__: of coarse. we are a semi friendly bunch
21:02 Drew__ oh thanx, I will go through that too
21:02 hemebond Yes. That will allow me to live precariously through you.
21:03 forrest Unless you want to pay some of us to teach you, in which case then you can ask whatever questions you want.
21:03 whytewolf but yeah, don't get us to design your entire infrastructure :P most of us are paid to maintain our own.
21:04 forrest Exactly, there's been a few... abusers and it gets pretty frustrating that people don't use google to search for their solution, or expect us to figure out their entire saltstack infra. If I wanted to do that I'd tell your company to pay ME to do it since I know how.
21:04 Drew__ Nah... I will ask until I learn... I will ask to learn rather than to make you do my work
21:04 babilen Sounds like a plan :)
21:05 forrest I actually considered combing the logs to see how many questions I had answered, but they don't go back far enough sadly :(
21:05 MindDrive joined #salt
21:05 babilen forrest: I guess that I'll have pretty complete logs for the last two years at least
21:06 forrest babilen, I have between 10k-20k lines from prior to when the IRC was logged.
21:06 breakingmatter joined #salt
21:07 whytewolf forrest: not sure how compleate they are http://irclog.perlgeek.de/salt/
21:07 babilen The interesting problem (computational linguistic wise) is to determine when you answered a question
21:07 forrest whytewolf, They don't have all of them from when I was even more active in the IRC.
21:07 forrest babilen, Yes I would usually consider it a line, then a 'thanks' or a 'that worked' or something within X number of lines.
21:07 forrest Plus any 'thanks forrest' lines I would probably factor in for duplicates and consider that a question answered, or at least close to it.
21:08 forrest But I honestly don't care enough to go back through all of it, doesn't get me anything.
21:08 babilen Sure, but it starts with deciding which utterances belong to which discourse ... PhD colleague worked on IRC data and it was a little more tricky than one might expect at first. :)
21:08 forrest Does anyone have a working example of the reactor combined with the cherrypy portion of salt-api?
21:09 forrest babilen, Yeah exactly, I'm not interested in going down that rabbit hole.
21:09 babilen But then we could just start with a baseline "whenever forrest said something, it helped" and go from there :)
21:09 forrest That would be pretty worthless.
21:09 fyb3r apparently syndic doesnt create the default config files/folders under installation
21:09 babilen fyb3r: Please elaborate
21:10 fyb3r at least from what im reading, the syndic id is supposed to be configured inside /etc/salt/minion
21:10 fyb3r but after installing it, that file along with minion.d are not there.    not a huge deal but it could be confusing for someone
21:12 colegatron I currently have a multimachine vagrant file to have part of my servers up and provisioned (as salt-masterless minions). I am thinking to move to salt-cloud but I have a doubt. Vagrant keeps in sync the vm's up and running storing the aws instance ID in a hidden file in the filesystem.
21:13 colegatron does anyone knows how salt-cloud keeps track of the aws launched vm's and the configured ones in the salt-cloud files?
21:13 colegatron does it also use local files or does it make queries to aws api's?
21:16 mapu joined #salt
21:16 forrest basepi, Has anyone reported issues with salt-api on 2015.8.1? I don't see any issues around.
21:17 otter768 joined #salt
21:17 basepi Not off the top of my head, but I'm not as plugged in as I used to be when I was doing all the triage.
21:17 basepi jfindlay might know ^
21:18 forrest basepi, Alright thanks
21:19 RandyT greetings shakers
21:19 RandyT wondering if someone can help me with a winrepo issue
21:19 RandyT Outlined here: https://gist.github.com/rterbush/03dcdf90e41e9fe91618
21:20 RandyT this has worked in other environments... struggling to get it working in a soon to be production environment.
21:20 RandyT error messages don't give me much to go on
21:20 hemebond RandyT: There was someone in here recently with winrepo issues. Are you able to install software?
21:22 RandyT hemebond: yes, able to install software, run updates matching filters for "Critical", etc.
21:23 RandyT attempts to run the minion with -l debug don't seem to have the desired effect...
21:24 forrest If anyone is familiar with salt-api/reactor and has time to take a look at this I'd appreciate it. Not sure why it's returning a 405: https://gist.github.com/gravyboat/285e8b082715c4b0106e
21:29 drawsmcgraw forrest: I think you're missing 'webhook_url: /hook' in salt-api.conf
21:29 drawsmcgraw At least, my config has that (and works) and yours lacks it.
21:29 forrest drawsmcgraw, I already tried that actually, same result. In addition the default according to the docs is /hook
21:29 diegows joined #salt
21:29 forrest I'll add it back to double check.
21:30 drawsmcgraw Wasn't aware it had a default. Interesting.
21:30 forrest drawsmcgraw, What release of salt-api are you running?
21:30 drawsmcgraw I also have certs (albeit self-signed) but I don't think that's the issue
21:30 drawsmcgraw 2015.5.1
21:30 drawsmcgraw 2015.5.3, sorry
21:30 forrest yeah just double checked with the webhook_url value, still the same 405, this is on 2015.8.
21:31 DammitJim joined #salt
21:31 forrest 2015.8.1 that is
21:31 forrest Other than that your config looks very similar minus the SSL stuff though?
21:31 drawsmcgraw OH
21:32 Rumbles joined #salt
21:32 drawsmcgraw forrest: Try POST'ing instead of GET'ing ?
21:32 drawsmcgraw Yes, the configs are just about identical. Even using cherry_py
21:33 forrest drawsmcgraw, Nope, a POST gives a straight up 500 error
21:33 drawsmcgraw :/
21:34 drawsmcgraw very odd, but I have this in my reactor config -> - salt://reactor/riemann-update-configs.sls
21:34 drawsmcgraw I'm using the salt:// urn.
21:34 drawsmcgraw It's been a while since I stood this up - not sure why I went that route
21:34 drawsmcgraw And I don't understand how that difference would cause salt-api to throw a 405...
21:35 jalbretsen hmmmm.  I thought I read this a while ago, now I can't seem to find it.  Are pillar top.sls files merged together that are present in multiple environments?  (ie base/top.sls, test/top.sls, stage/top.sls)
21:35 enigma99a ugh I am starting to hate pygit2 -- Failed to checkout az1 from git_pillar remote 'az1 git@myhost.local:salt/az1-pillar.git': 4 conflicts prevent checkout
21:35 enigma99a all the sudden, no reason
21:36 drawsmcgraw forrest: er... is there any chance you can see the salt-api logs and try that POST again?
21:36 forrest enigma99a, Did you confirm that there weren't any local changes? That only happened to me when people had been messing around on the boxes.
21:36 drawsmcgraw I wonder if you *do* have to post, and doing that unearthed a potential config problem.
21:37 drawsmcgraw That it throws a 500 is interesting. My Github webhooks are POSTs to my salt-api machine.
21:37 enigma99a forrest: yeah no local changes
21:37 breakingmatter joined #salt
21:40 forrest drawsmcgraw, I mean there's https://gist.github.com/gravyboat/a391f211e6875f9eede2 but I think that's due to trying to post with no data since it's not expecting data for the request type.
21:40 drawsmcgraw forrest: Give it some bogus data? :)
21:40 drawsmcgraw I don't use any of the postdata from my Github web hooks
21:41 drawsmcgraw I vaguely remember this being an issue.... Seeing empty POSTs cause salt-api to have issues
21:41 forrest Hmm
21:41 whytewolf to be fair the docs do only show a POST method. no docs for a GET method
21:41 drawsmcgraw I think throwing a -d "foo=bar" (or however it looks in curl) may help
21:41 forrest drawsmcgraw, That was totally it :|
21:42 drawsmcgraw Nice. And yes, I remember this now
21:42 forrest That seems pretty stupid when you don't need to pass data through though
21:42 forrest lol
21:42 drawsmcgraw Yeah. That may be a cherrypy (or any other server) issue though?
21:42 forrest could be
21:42 forrest but passing through -X POST should allow the post to work you know?
21:43 drawsmcgraw No.... it's Salt code :/
21:43 drawsmcgraw https://gist.github.com/gravyboat/a391f211e6875f9eede2#file-gistfile1-txt-L6-L8
21:43 drawsmcgraw Yes. I agree
21:44 forrest anyways thanks drawsmcgraw and whytewolf
21:44 forrest I'll create an issue for this
21:45 forrest because it should work
21:45 drawsmcgraw forrest: good idea. Glancing at the code, I actually don't understand how that error happens. I'm clearly losing my edge
21:45 ziro` joined #salt
21:45 forrest hah
21:46 whytewolf most likely all testing was done using auth.
21:46 pdayton joined #salt
21:48 forrest drawsmcgraw, https://github.com/saltstack/salt/issues/28714
21:48 drawsmcgraw nice. Thanks! I'll follow that one
21:48 forrest whytewolf, The odd thing is though there are quite a few tutorials people wrote out there that don't use post at all
21:48 forrest so clearly something has changed at some point
21:48 forrest they just do a curl -sSk url and it's good to go.
21:49 whytewolf forrest: odd. maybe. cause looking at the code. it doens't have a GET method at all. so it much have changed
21:49 JDiPierro joined #salt
21:49 forrest whytewolf, Yeah, I figured the get method would work correctly specifically because it was a webhook, but apparently that isn't the case.
21:50 fyb3r whytewolf. i couldnt find the answer from my issue lastnight. as far as I can tell something came in via an update to some package other than salt and totally hosed it
21:51 fyb3r what that was I dont know. but im trying to track it down still just for the sake of letting you guys know lol
21:51 whytewolf fyb3r: lol. it is strange that even a full reinstall of salt on your minions left them broken
21:52 whytewolf only time i had something simalar happen to me was instaloing openvswitch and having a odd network setup after words
21:52 fyb3r funny you mention that. one of my minions failing is a neutron node lol
21:52 enigma99a forrest: Only thing I did was push some changes in the pillar, but if I clear /var/cache/salt/master/git_pillar/ the data comes back, so it's able to checkout even though but something isn't right. Maybe I should give up trying to use git for pillar data
21:52 fyb3r but that doesnt explain the one thats a couchbase node, and another thats running jenkins and thats all >_> lol
21:53 whytewolf yeah. that doesn't make sense
21:53 fyb3r the bad part, its ubuntu 14.04 and centos 7 minions i found.  the plot thickens!
21:53 forrest enigma99a, That is odd, did you see if there's an issue for it? That seems like a super annoying bug
21:54 bastion1_ joined #salt
21:54 enigma99a I think maybe I should be using a DB anyway, so I might just do that
21:55 enigma99a or just rsync them.. dunno little frustrated tho
21:55 bastion__ joined #salt
21:58 pdayton joined #salt
21:59 dthom91 joined #salt
22:00 ajw0100 joined #salt
22:01 colegatron joined #salt
22:03 RandyT so I found one of the problems with the winrepo issue...
22:03 RandyT https://github.com/saltstack/salt-winrepo/blob/master/putty.sls#L2
22:03 Rumbles joined #salt
22:04 forrest drawsmcgraw, Are you targeting via grains in your reactors?
22:04 RandyT pkg.available_version still failing though.
22:05 cpattonj_ joined #salt
22:05 forrest drawsmcgraw, Ahh nevermind I see my error, was expecting the syntax to be identical but it isn't
22:05 RandyT anyone know if there is a way to disable automagic pull of the default repos for winrepo?
22:06 enigma99a fixed... I think the lesson learned is not to use branches and just commit it to master for pillar git
22:07 nethershaw joined #salt
22:09 cberndt joined #salt
22:09 miked510 joined #salt
22:26 ntropy any word on when to expect 2015.8.2?
22:27 whytewolf last i heard was mid month
22:27 zmalone I think the prediction was sometime this week.
22:27 ntropy thanks, that's not too long to wait for some patches to ship :)
22:28 whytewolf oh, huh. looks like they are looking at tagging today. [well they were planning on today around oct 30th]
22:29 fyb3r the hell is virt-what
22:30 whytewolf lol, looks like a shell script that detects if it is running in a virtual machine
22:30 whytewolf https://people.redhat.com/~rjones/virt-what/
22:31 fyb3r lol right on. i just see it popping up in syndic. "Please install 'virt-what' to improve  results of the 'virtual' grain"
22:31 whytewolf lol. that would do it
22:32 fyb3r guess ubuntu doesnt come with it ;) haha
22:32 joyrida08 joined #salt
22:32 whytewolf shows that it is packaged for it. maybe it is just a simple install
22:33 techblaze joined #salt
22:36 jaybocc2 joined #salt
22:38 larsfronius joined #salt
22:40 murrdoc joined #salt
22:43 markm joined #salt
22:43 karlthane joined #salt
22:43 lynxman joined #salt
22:46 bhosmer joined #salt
22:56 lemur joined #salt
22:56 ahammond I'm having trouble reading the datastructure in front of my eyes. Help please? https://gist.github.com/ahammond/8ab765420b566ba74cd3
22:57 pppingme joined #salt
22:58 forrest what's under local on your iphone_certificates pillar? Seems like the data structure is screwed up
22:58 ahammond specifically, that extra layer of nesting is confusing me...
22:59 whytewolf that looks like something had a : that wasn't escaped
22:59 forrest agreed
22:59 ahammond hmmm, I think I got it. It's an empty key. :( Gross.
22:59 solidsnack joined #salt
23:00 ahammond this is a pillar from our custom pillar_files stuff. I'm not super in love with it, but it lets you pull a file-structure into a pillar and send it to a minion.
23:01 moogyver forrest: submitted a PR for your empty POST issue.
23:01 forrest moogyver, Nice! Did you tag it in the issue?
23:01 moogyver yeah
23:02 moogyver The code was assuming if something wasn't a list, then it had to be a dict, but that's not really true. :)
23:02 whytewolf ahammond: so you are saying that your team rewrote this https://docs.saltstack.com/en/latest/ref/pillar/all/salt.pillar.file_tree.html#module-salt.pillar.file_tree
23:02 forrest moogyver, Nice, looking at the code now, good fix.
23:03 ahammond whytewolf nope, we re-invented that wheel, but less round.
23:03 whytewolf lol. isn't that always how it is.
23:03 ahammond and... now I'm going to see if we can switch to the standard wheel. Sigh.
23:04 ahammond whytewolf ah, I see the difference. That one requires a hostname for access control.
23:05 ahammond ours is managed directly by pillar top.sls
23:05 whytewolf ahhh fun.
23:06 forrest https://gist.github.com/gravyboat/94973f4961b722d21ba7 Any ideas what is wrong with this pkgrepo.managed state?
23:07 whytewolf forrest: I have no idea. although i did hear that pkgrepo has gotten pretty buggy in 2015.8
23:08 alexhayes Is it possible to string format pillar data? ie.. I want to set a particular value to "foo/{name}" and then in my SLS format that, replacing {name} with something (ie.. python "foo/{name}".format(name='bar')
23:08 forrest whytewolf, I'm going to be quite pissed if pkgrepo is broken
23:08 forrest this is fundamental functionality
23:08 hemebond forrest: What is baseurl? Has it just been added?
23:08 hemebond forrest: In my pkgrepo state name is the URL to the repo.
23:09 forrest hemebond, baseurl is specifically for yum based distro
23:09 forrest *s
23:09 hemebond Ah, right.
23:09 forrest I already tried dropping that as well without luck
23:13 forrest I honestly don't even understand the failure, there are THREE different locations it could be getting the name from.
23:13 woodtablet joined #salt
23:14 woodtablet hello all
23:14 forrest jfindlay, Any issues on pkgrepo.managed for 2015.8.1 that aren't in the issues list already that you're aware of?
23:14 forrest hey
23:14 whytewolf forrest: this might not make sense but have you tried dropping the - name: since you already have - humanname
23:14 babilen forrest: hummanname != humanname
23:14 woodtablet anyone here know how to use salt-cloud to clone in vmware ? i am reading the tutorial, and i have my configs just like in the docs, but it doesnt say what the salt-cloud command to run. or maybe i am blind: https://docs.saltstack.com/en/develop/topics/cloud/vmware.html#cloning-a-vm
23:15 forrest babilen, Good catch
23:15 babilen Not sure if that's the issue though
23:16 Dev0n hey, is it me or is the latest debian saltstack repo completely messed up?
23:17 whytewolf woodtablet: since cloning is a profile setup. I would assume that you use it like any other profile.
23:17 forrest babilen, That did it, good catch> I must have looked over that half a dozen times
23:17 Dev0n https://github.com/saltstack/salt/issues/27820
23:18 otter768 joined #salt
23:18 woodtablet whytewolf -  hmm.. ok ill go intro page of the salt-cloud beginning and read up.
23:18 RandyT anyone here to lend a hand debugging a winrepo issue on 2015.8.1?
23:19 whytewolf I would RandyT but I'm like a bad maid. I don't do windows
23:19 RandyT whytewolf: yeah, wish I did not have to... :-)
23:19 moogyver @woodtablet - i think it just uses the 'create' command and if the profile is setup clone, then it creates it via cloning
23:20 moogyver @? ugh.  been writing too many comments on github.
23:20 forrest what an elitist over here...
23:20 forrest ;)
23:20 moogyver :P
23:20 forrest moogyver, whiteinge commented on your PR by the way
23:20 RandyT whytewolf: perhaps you can clear this up for me though...
23:20 RandyT where is most appropriate place to put minion configuration parameters?
23:20 iggy humman... I missed that too :/
23:21 moogyver ah, linting
23:21 moogyver I'll fix
23:21 iggy forrest: I'd still open an issue, that's a strange error message to get from that mistake
23:21 jfindlay forrest: not that I know of
23:21 jaybocc2 joined #salt
23:21 whytewolf RandyT: lately I have been moving more and more to pillar for minion configs.
23:21 whytewolf most of the configs can be done through pillar
23:21 forrest iggy, Yeah I'm considering it
23:21 forrest jfindlay, Alright I'll open an issue then
23:22 woodtablet whytewolf - ahhh ok, wonderful, thanks! i skipped straight to the vmware section.
23:22 jfindlay forrest: sounds good
23:23 RandyT whytewolf: for something that should be cross platform, like winrepo configuration, is there a more general place?
23:23 RandyT only reason I push back a bit is the added  complexity of the pillar right now to sort this problem out..
23:24 whytewolf RandyT: well then second place I put minion configs is in minion.d in the config directory.
23:24 moogyver whytewolf : you use the salt-cloud stuff right? but you use it with openstack?
23:25 lexter joined #salt
23:25 whytewolf moogyver: yeah. although right now my openstack cluster is gone
23:25 whytewolf [tore it down this weekend to start from scratch]
23:25 moogyver you played with any of the vmware stuff?  curious if there's a way to use some of the actions and specify which provider you want to do the action on.
23:26 whytewolf unforchantly not. I didn't get into salt until after i nuked my home VMware setup
23:30 whytewolf moogyver: I guess it depends on the action i know the -f takes a function and a provider
23:30 moogyver yeah, I was looking for ways to start/stop VMs, which seem to use the '-a' action
23:30 moogyver which doesn't let you specify a provider
23:30 moogyver ( that I could figure out )
23:31 subsignal joined #salt
23:32 LeProvokateur joined #salt
23:33 murrdoc1 joined #salt
23:33 whytewolf moogyver: try adding a provider to the command.
23:35 whytewolf although i think actions are based on profiles
23:35 moogyver whytewolf: i've tried.. it just interprets it as another vm
23:35 moogyver yeah
23:35 moogyver and we don't use profiles, was just hoping for some orchestration stuff
23:36 moogyver forrest : it got merged
23:36 forrest moogyver, sweet. Hopefully it will make it into the next release.
23:36 moogyver hopefully, i remember basepi saying they were tagging a new release today i think
23:37 basepi Shooting for tomorrow, for release next week
23:37 perfectsine joined #salt
23:40 whytewolf moogyver: kind of odd that salt-cloud is the only interaction with vmware that salt has. for openstack it also has modules and states that can be used for non salt-cloud stuff
23:41 Gareth hrm. salt through Azure endpoints is highly unreliable.
23:41 Bryson joined #salt
23:42 moogyver whytewolf : yeah :/ we use other tools for doing all the provisioning and such for vmware, but using salt for some orchestration pieces with vmware would be nice.
23:43 woodtablet moogyver  - what do you use ?
23:43 moogyver woodtablet - redhat's cloudforms
23:43 RandyT Just an FYI for anyone else lurking here that is dealing with hetero environments and salt: https://github.com/saltstack/salt/issues/28721
23:43 moogyver salt handles our baremetal provisioning, along w/ cobbler.
23:45 woodtablet moogyver - is that for vcloud or your local vmware esx cluster install ?
23:45 debian112 left #salt
23:45 jaybocc2 joined #salt
23:45 moogyver woodtablet - our internal ESX clsuters
23:45 woodtablet moogyver - hmm ok. i ll have to look at that. i am trying salt-cloud right now
23:45 otter768 joined #salt
23:45 moogyver woodtablet - it should work, we just had cloudforms before we had salt
23:45 moogyver so there was no reason to switch it
23:46 woodtablet moogyver - ok
23:48 forrest Has anyone tested out dockerng in 2015.8.1
23:48 forrest Getting inspect_image errors which seem to be related to the image not being pulled down, but it should be doing that automatically.
23:48 falenn joined #salt
23:50 cberndt joined #salt
23:50 fxhp joined #salt
23:51 fxhp joined #salt
23:54 LeProvokateur joined #salt
23:54 whytewolf interesting.... moogyver have you seen this? https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.cloud.html#salt.modules.cloud.action cloud.action with a provider=
23:54 moogyver ah i hadn't.  interesting way of doing it i guess
23:55 moogyver seems that it would make more sense as a runner.
23:56 whytewolf might want to try it at least. not 100% it will work. but it is the only way i have seen of providing action with a provider
23:57 bfoxwell joined #salt
23:58 fyb3r left #salt

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