Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-12-18

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

All times shown according to UTC.

Time Nick Message
00:00 wt I just saw 3 or 4
00:00 wt 2
00:01 iggy yeah, I've never installed anything special for git (after 2014.1 which only did git-python)
00:02 wt yeah
00:02 rikair left #salt
00:03 wt I am occasionally seeing 3 git fetches
00:03 wt I bet it would get worse if I turn up worker_threads.
00:03 wt there's 2
00:04 wt 2 git fetches for pillars
00:04 iggy pillars use a different backend than the normal state gitfs backend
00:05 wt there was an fs fetch
00:05 iggy so I wouldn't be surprised if you saw some overlap between pillars and states
00:05 wt Some are pillars...some are files.
00:05 sunkist joined #salt
00:05 wt I just was  seeing pillars when I said that.
00:06 wt 3 pillars
00:06 hasues joined #salt
00:08 wt Just saw an instance of 4
00:09 Whissi joined #salt
00:10 wt haven't seen more than 4 yet
00:12 sunkist joined #salt
00:13 superted666_____ joined #salt
00:15 maxd_ joined #salt
00:16 bytemask_ joined #salt
00:16 goal joined #salt
00:17 gmoro joined #salt
00:17 jeddi joined #salt
00:18 DaveQB joined #salt
00:18 forze joined #salt
00:19 kermit joined #salt
00:20 jorge_lo left #salt
00:22 hasues left #salt
00:23 Psi-Jack left #salt
00:25 yomilk joined #salt
00:27 spookah joined #salt
00:40 Ryan_Lane joined #salt
00:44 bfoxwell joined #salt
00:45 budman joined #salt
00:50 SheetiS joined #salt
00:51 sunkist joined #salt
00:58 wt I have evidence of 6 simultaneous git fetches. I don't know which repo they were for.
00:58 ponpanderer joined #salt
00:59 ponpanderer Hello, is there a configuration option on the master that does not accept any key as "pending", basically requiring all keys are pre-seeded and accepted?
00:59 iggy you could replace the git binary with a wrapper that logged the time and full command line every time git was called
01:00 aqua^mac joined #salt
01:00 wt Maybe when I have more time to mess with this.
01:03 sunkist joined #salt
01:04 otter768 joined #salt
01:07 mosen joined #salt
01:14 aurynn so if I wanted to run a state based on the output of a commandline command... should I just write a pillar to do that?
01:15 aurynn do that == provide the data
01:16 aparsons joined #salt
01:16 joehh1 hal58th: that is a feature of launchpad.
01:17 joehh1 hal58th: if you are after a particular version, I should have the original source package
01:17 aurynn I'm currently considering doing something absurd, like an onlyif hanging off an 'echo 1' cmd
01:18 joehh1 if it is the last of the series, then you should be able to find it in one of the other salt ppas at https://launchpad.net/~saltstack
01:30 aparsons_ joined #salt
01:33 murrdoc aurynn:  dont do it
01:33 aurynn murrdoc, I know, it's awful and made of technical debt
01:34 murrdoc :)
01:41 xenoxaos joined #salt
01:49 sunkist joined #salt
01:58 mordonez joined #salt
02:00 sunkist joined #salt
02:06 peters-tx joined #salt
02:08 malinoff joined #salt
02:32 RainDoctor joined #salt
02:34 sunkist joined #salt
02:39 bhosmer_ joined #salt
02:45 Ryan_Lane joined #salt
02:46 nitti joined #salt
02:56 smcquay joined #salt
03:05 peters-tx joined #salt
03:05 otter768 joined #salt
03:12 cberndt joined #salt
03:16 Mso150 joined #salt
03:26 xenoxaos joined #salt
03:28 mosen joined #salt
03:38 chandi joined #salt
03:46 evidence joined #salt
03:46 chandi Hi, can we use templating logic in pillar sls files?
03:47 notnotpeter joined #salt
03:52 hasues joined #salt
03:53 hasues left #salt
03:57 dimeshake chandi: sorta. you can do ifs on grains and things. what do you want to do?
03:57 dimeshake you can't really refer to pillar data within pillars
03:58 chandi just sent an email to the mailing list which explains further....
03:59 chandi I was just confused as to why state sls processesing was different to pillar sls processing, the difference is not clearly explained in the docs I think
04:03 agend joined #salt
04:08 dimeshake i totally forgot there was a mailing list that i'm subscribed to
04:08 dimeshake damn you gmail filters
04:08 chandi oh
04:14 rojem joined #salt
04:16 dimeshake chandi: if you declare the same dictionary id twice, the latter will overwrite the former
04:16 dimeshake they do not merge in pillar like that
04:17 dimeshake you are redeclaring it
04:17 scott2b joined #salt
04:18 ajolo joined #salt
04:19 budman joined #salt
04:23 chandi they do seem to merge if I am including into the same namespace without using the key option
04:23 chandi I think that's what happens on the root of the namespace..that part is explained in the docs
04:25 budman_ joined #salt
04:26 aquinas joined #salt
04:28 scott2b Can anyone explain this statement in the comments of the mongodb formula? "This setup for mongodb assumes that the replica set can be determined from the id of the minion" What does it mean to determine the rs from the id?
04:28 dimeshake hmm.. any way to specify the --user option to pip when using pip.installed?
04:28 dimeshake the -user state option runs it as the user, but can't write to /usr/lib/python obviously, but i don't want it to.
04:30 dimeshake ah. --install_options: - --user
04:32 voxxit joined #salt
04:35 dimeshake hmmm
04:35 dimeshake why would i be getting 'cp.push_dir' is not available
04:35 Silvenga joined #salt
04:35 dimeshake i have enabled file_recv on master
04:36 budman joined #salt
04:39 scooby2 joined #salt
04:39 dimeshake ah. new in 2014.7 and minion is running old version :)
04:40 viq joined #salt
04:40 viq joined #salt
04:49 FineTralfazz joined #salt
04:51 Leonw Hi... any trick on using pillars through a syndic server?
04:53 FineTralfazz joined #salt
04:54 FineTralfazz joined #salt
04:58 FineTralfazz joined #salt
05:03 atbell joined #salt
05:05 FineTralfazz joined #salt
05:06 otter768 joined #salt
05:09 cnelsonsic joined #salt
05:12 TheThing joined #salt
05:31 atbell joined #salt
05:48 ramteid joined #salt
05:49 nnion joined #salt
05:57 hasues joined #salt
05:57 hasues left #salt
06:03 agend joined #salt
06:07 Thugal joined #salt
06:13 TheThing joined #salt
06:15 Furao joined #salt
06:20 boredatwork joined #salt
06:28 dynamicudpate joined #salt
06:30 thehaven joined #salt
06:31 davidone joined #salt
06:31 evidence joined #salt
06:32 jonasbjork joined #salt
06:42 wt joined #salt
06:44 Rockj joined #salt
06:46 catpigger joined #salt
06:49 Rockj joined #salt
06:52 Rockj joined #salt
06:54 Ryan_Lane joined #salt
06:57 atbell joined #salt
06:59 Rockj joined #salt
07:06 otter768 joined #salt
07:08 hebz0rl joined #salt
07:09 ttrumm_ joined #salt
07:13 felskrone joined #salt
07:18 Edgan joined #salt
07:18 Edgan How do I make the output of salt one line per host?
07:19 Furao i guess with a custom outputter
07:19 Edgan Furao: How hard is that?
07:21 Furao if you know python probably not much
07:22 Furao but you can run salt with —output=yaml (or an other easily parsable output)
07:22 Furao and load that file and print it in an other format
07:22 Furao https://github.com/saltstack/salt/tree/develop/salt/output
07:22 Furao some example of outputers if you know python it’s easy
07:22 Edgan raw seems to do close enough to what I want
07:23 Edgan I want to be able to grep the output
07:24 OnTheRock joined #salt
07:25 covox Edgan: salt --out-indent=-1 --output=json "*" test.ping ?
07:26 Edgan covox: not bad, but raw is better
07:26 OnTheRock joined #salt
07:26 harkx joined #salt
07:27 jdmf joined #salt
07:27 covox it's pretty similar, one is single-line JSON and the other is single-line python repr()
07:28 Edgan covox: I am getting { and } on their own lines with indent json
07:28 Edgan covox: Not that they really matter
07:28 covox Edgan: --out-indent=-1 should take care of that
07:29 dotz joined #salt
07:29 covox at least it does here
07:29 Edgan It takes out the leading space but not the {}
07:29 pdayton joined #salt
07:30 covox wfm on 2014.7.0
07:31 Edgan salt-2014.1.13-1.fc21.noarch
07:31 Edgan Looks like I need to upgrade
07:31 Edgan I am just playing with salt. I am impressed with how simple it is to setup.
07:32 slafs joined #salt
07:32 slafs left #salt
07:37 Edgan covox: Found the newer version in updates-testing. You are right. The newer version does it in one line that way
07:39 yomilk joined #salt
07:40 schlueter joined #salt
07:41 covox Edgan: \o/
07:46 Edgan Does salt have a equivalent to facter?
07:47 covox Edgan: quite a bit of free info is populated in the grains
07:47 covox try salt-call grains.items
07:50 Edgan covox: Thanks :)
07:50 covox np
07:50 harkx joined #salt
07:51 Edgan covox: Lots of interesting new facts but missing a lot too
07:53 Edgan Looks like there is a contrib to add facter facts as grains
07:54 covox anything in particular you're looking for?
07:58 trikke joined #salt
07:59 babilen grains are pretty easy to write yourself if you know some Python. You essentially just return a dictionary with the data you want to see in grains -- http://docs.saltstack.com/en/latest/topics/targeting/grains.html#writing-grains
07:59 Edgan covox: Haven't tried it in ec2 yet, but definitely want grains like ec2 facter facts
08:00 dotz joined #salt
08:00 dotz Guys, I need to copy some files once with salt-ssh from master to minion... my question, does cp.cache_local_file work with salt-ssh? Because, either I do something wrong or it doesn't..
08:01 schlueter joined #salt
08:03 schlueter1 joined #salt
08:09 Auroch joined #salt
08:10 Edgan Looks like the ec2 stuff I want is also in salt-contrib
08:13 claytron joined #salt
08:14 ktosiek I've played with Salt a bit yesterday, and I think I know the basic words now. Where to go next? Are there any high-level articles describing using Salt in production? I'm mostly interested in the flow from changing some code and testing it locally through review, some staging/preprod etc. to safely rolling out the change in production
08:16 desposo joined #salt
08:16 babilen I'm not aware of such an article. If you were to have more specific questions we could, however, try to answer them.
08:16 phx ktosiek, depending on the size of the infrastructure and various other requirements (like validation and stuff), for that I would involve other components, like distributed filesystems (like AFS) which can propagate code and stuff between various servers centrally, and also providing other neat features. salt is basically a CM and parallel execution system, for distributing code, and supporting the devops workflow for app development adding other stuff is somet
08:17 lb1a joined #salt
08:23 hotsnow joined #salt
08:23 karimb joined #salt
08:23 robot_hands joined #salt
08:25 dotz_ joined #salt
08:26 ktosiek hmm, ok. So what I want to have: a way to push new code to staging (I'm keeping configuration in git, if that matters) without touching production (just having separate branches should be ok). A way to update some of minions to the new version of code. A way to report which minions are still running old version.
08:28 jhauser joined #salt
08:32 ktosiek My dream workflow (so far) is: hack and test locally, push to Gerrit, have CI run some tests, merge from Gerrit to devel, have CI apply the new config (possibly in some staggered way where needed), check that monitoring is still green and services still work, click some big "promote to prod" button that applies this to production (with all the staggering/orchestration needed)
08:33 noway_ joined #salt
08:33 bezaban ktosiek: you can look at targeting. http://docs.saltstack.com/en/latest/topics/targeting/ also you can run with test=True to see what changes would have been set for seeing what ones are still old
08:33 bezaban note I am a novice myself, I run it but don't have very extensive state files
08:33 babilen You'd keep your salt configuration in GitFS (cf. http://docs.saltstack.com/en/latest/topics/tutorials/gitfs.html ) -- branches there are treated as different environments, but that gets a bit tricky if you want to support proper git branching workflows (you essentially have to keep the top.sls in its own repo)
08:34 babilen My workflow right now is: I develop salt in a local branch and sync that with a few scripts to a local /srv/ directory (along with pillars). I then fire up a couple of vagrant boxes (cf. https://gist.github.com/babilen/e9479fdfbcca431db208) and run that until I am happy.
08:35 bezaban oh right. I misunderstood the bigger picture of the question
08:35 ktosiek but targeting is still half the story - can I have one code used by minion for managing configuration drift, and then only run test=True with the new version?
08:37 babilen So, once that is done I merge that into a dev branch locally which I push to a central repository. My top.sls targets smaller "dev" deployments specifically and we run a highstate against them. People then test it and once I get the go ahead from the developers/qa I merge the changes into my local master branch, push that and provision production.
08:38 babilen As of now we don't run any automated tests with, say, jenkins that ensures that specific functionality is given (e.g. website foo is up, backend bar responds to query quux with data ....)
08:38 mikkn joined #salt
08:39 babilen ktosiek: What do you mean by "one code used by minion for managing configuration drift" ?
08:40 ktosiek babilen: I should've said "one code used by minion for scheduled runs"
08:40 babilen What is this "one code" you are referring to?
08:40 ktosiek some previous version (so a minion will have his "current" version, and I'd run test=True of the "next" version without changing "current")
08:41 bhosmer joined #salt
08:42 hejjahejja joined #salt
08:43 babilen Sure, provision a minion with the previous version and then run a highstate with test=True against it with HEAD
08:43 babilen s/previous version/HEAD^/
08:47 CeBe joined #salt
08:47 ktosiek cool, I'll go check out gitfs now and come back with more questions in the evening :-)
08:48 ktosiek thank you for help!
08:48 Mindfart joined #salt
08:48 dotz_ ktosiek: gifts?
08:48 dotz_ is it christmas yet?
08:48 dotz_ Ah, GITfs.. NVM.
08:48 ktosiek dotz_: how did you read that? :-D
08:49 babilen ktosiek: have fun!
08:51 ktosiek oh, gifts (apperently /me is blind too :-P)
08:53 kawa2014 joined #salt
08:56 VSpike Someone here gave me the link to http://devop.ninja/configuration%20management/2014/08/25/the-grains-conundrum/ (I think it was the author). I've finally come back to it to try and use it....
08:56 VSpike I was wondering about the last bit on the page, "The target and tgt_type are pieces of information that are parsed through a macro I wrote that allows for the Orchestration run to target different groups of servers."
08:56 VSpike I wasn't quite sure how such a macro would work... can anyone help?
09:00 s51itxsyc joined #salt
09:01 fredvd joined #salt
09:02 VSpike Ah, it was SheetiS :)
09:03 ktosiek can I force gitfs to only look at one branch? (Layering of data from different branches is pretty incompatible with every other git-based tool >_>)
09:03 babilen And it is *still* wrong to keep this data on the minion
09:03 phx ktosiek, you can utilize salt's parallel execution for that stuff. also, if you read up a bit on AFS, you can push stuff to an AFS cell, mark various versions as prod/uat/qa/dev/whatever, have various servers executing code with given tags, then whatever is on the AFS stuff is seen the same on all boxes
09:03 phx ktosiek, also, gerrit/jenkins/whatever can work from that. it's just a hint, however requires some considerable effort :)
09:04 akafred joined #salt
09:04 N-Mi joined #salt
09:05 babilen VSpike: grains are the biggest problem salt has right now. What salt needs is a way to define static data about minions that is available from the beginning. Grains are only used for things like roles, environments, ... and so on because they are the data that is being made available first. You do, however, *not* want any information *on* the minion (i.e. an insecure datastore) that you use for deciding if the same minion should be able to see some ..
09:05 s51itxsyc guys, wired missing pyyaml pkg in epel6? http://fpaste.org/160881/41889343/
09:05 babilen ... sensitive data or states.
09:05 babilen salt really needs a "pre-pillar" that you can reference and use for storing static data.
09:07 otter768 joined #salt
09:10 VSpike babilen: hm, I see your point. But how would you solve the problem with what's currently available?
09:11 iwishiwerearobot joined #salt
09:12 pttc joined #salt
09:13 analogbyte joined #salt
09:17 aanriot joined #salt
09:17 lothiraldan joined #salt
09:18 babilen VSpike: Well, you can get pretty far with pillars for targeting states, but you cannot use them for targeting themselves. Arguably the targeting of pillars is even more sensitive than states and this poses a real problem.
09:19 babilen So: I don't really know, but relying on data provided by a minion to decide which sensitive data you make available to it (i.e. not by consulting a secure and authoritative data source) is a *huge* no no
09:28 VSpike babilen: I'm currently using specially crafted host-names to decide what config the machines get. (Although because I have a corpus of existing machines that don't fit the naming scheme, the actual logic is full of special cases).
09:28 VSpike babilen: But in practice what's difference security-wise between using a host-name and custom grains? Both are effectively stored on the minion
09:29 VSpike And even if you use the pillar to assign roles to boxes, surely at some point you have to choose the pillar data based on some property of the minion, which leads you back to the same problem?
09:29 babilen VSpike: You accept the minion *once* with its ID and it cannot change its ID without you having to reaccept it. A minion can, however, change grains as it sees fit.
09:30 VSpike Hm, a very good point :)
09:31 babilen The only reason why grains are being used for it is because there isn't another datasource that is available *everywhere* and supported for targeting in both pillar and states.
09:32 aanriot joined #salt
09:33 pttc how do you get more detailed feed back on why a package failed to install?
09:34 pttc i get "Comment: The following packages failed to install/update: nginx." and "Duration: 30717.313 ms" which suggests it's timing out but not sure how to get more detail
09:36 felskrone joined #salt
09:37 babilen pttc: Check the minion log
09:46 pttc makes sense :)
09:48 lothiraldan joined #salt
09:55 lb1a joined #salt
09:55 goudale joined #salt
09:55 goudale hi all
09:56 goudale once again, all my minions stopped answering
09:56 goudale what can i investigate ?
09:58 babilen define "stopped"
10:00 goudale no one respond under 'manage.up' nor 'test.ping' nor any regular state
10:02 babilen goudale: Could you log into one of the minions and check its log file? Please also run the eventlisten script on the master and check what kind of events you are seeing from your minions. Which version of salt is that? (OS/Deployed where/...)
10:02 goudale salt-minion service seems to have a pid
10:02 babilen Is the network you use under your control or could this also be a routing issue?
10:03 goudale I am exclusively using Ubuntu
10:03 goudale If i log to one minion, and restart the salt-minion service, the minions shows on the master
10:03 babilen yes, I figured as much, but would have been interested in the state before that
10:04 goudale I find nothing usefull in my log, running "grep salt -R /var/log". No trace of failure nor stopping
10:05 babilen What *do* you see in the minion log? Any events coming in from the stopped ones?
10:07 goudale absolutely nothing
10:07 goudale last line is one week old
10:09 goudale I pasted it here : http://sebsauvage.net/paste/?883b6a085f00dde7#c1CxD6UdMolGYoK7yVD1eYqFr0BytJESUgeaQh2nHyg=
10:09 colttt joined #salt
10:10 babilen goudale: So /var/log/salt/minion is empty?
10:11 goudale It is like the fourth time I experience this behaviour, still have no clue why
10:11 goudale babilen: yep
10:11 goudale 0 byte
10:11 babilen Has it been rotated?
10:12 Nils- joined #salt
10:14 goudale I guess it have been, thus not explicitely
10:14 babilen So what does /var/log/salt/ contain? (on the minion(s))
10:14 mordonez joined #salt
10:15 goudale -rw-------  1 root root    1732 Dec 18 09:33 minion
10:15 goudale -rw-------  1 root root     326 Dec 11 17:02 minion.1.gz
10:15 goudale -rw-------  1 root root    1830 Dec  3 11:29 minion.2.gz
10:15 goudale -rw-------  1 root root     161 Nov 28 17:52 minion.3.gz
10:16 goudale I downloaded the minion.1.gz, opened it, and there is a few line related to what I done, but no trace of failure
10:17 Furao joined #salt
10:17 babilen I'd also like to point out that 1732 != 0
10:18 babilen Mind pasting the output of "cat /var/log/salt/minion" and "zcat /var/log/salt/minion.1.gz" to http://refheap.com ?
10:19 goudale Indeed, I mistaken myself, this one have a size because I did some command this morning
10:20 goudale babilen: https://www.refheap.com/95091
10:21 goudale the traceback come from some stuff I did this morning and is not present on other minions
10:22 babilen So, could you pick a minion that is not responding and one on which you haven't done "stuff". Also elaborate on the exact nature of "stuff" if you can't find a single minion that meets these requirements
10:22 goudale have a look at this one : https://www.refheap.com/95092
10:28 lothiraldan joined #salt
10:29 vectra joined #salt
10:30 budman joined #salt
10:31 budman joined #salt
10:36 goudale I think I am experiencing this one : https://github.com/saltstack/salt/issues/13558
10:40 colttt joined #salt
10:45 lendar joined #salt
10:46 lendar left #salt
10:47 lendar joined #salt
10:48 VSpike babilen: so I guess the best you can do is assign pillar data by minion id, and perhaps include a role pillar item in there, then use the pillar role in states... IOW you can do states by role securely but not pillar since you can't use the pillar in pillars
10:48 VSpike Sorry, badly expressed
10:49 babilen That's what I meant, yeah
10:50 lendar hi, where should I send PR with fixes: to 2014.7 or develop? I want them released in 2014.7.1 :)
11:00 workingcats joined #salt
11:01 RainDoctor joined #salt
11:07 babilen goudale: So, do you have a master in /etc/salt/minion ?
11:08 otter768 joined #salt
11:11 babilen lendar: You should develop against develop and fixes will get backported (you might ask for that in your PR)
11:11 smoothify joined #salt
11:11 dnai23 joined #salt
11:12 babilen http://docs.saltstack.com/en/latest/topics/development/contributing.html has more information on that
11:13 babilen lendar: And what is "fixes" exactly. If it is strictly a bugfix then work on 2014.7
11:13 bernieke joined #salt
11:14 lendar babilen: broken module and incorrect doc, but just typos
11:15 babilen Did it ever work?
11:15 babilen But I guess it is fair enough to work on the 2014.7 branch in that case (even if "finally working module foo" could be seen as a new feature)
11:23 giantlock joined #salt
11:27 mortis__ running 2014.7.0, some minions seem to use like 260MB of mem while others use much less .....even tho the minion is doing basically nothing
11:27 mortis__ any ideas why this may happen?
11:28 Furao joined #salt
11:30 lendar babilen: sent 2 PRs (through only 1 to 2014.7), thanks for guidance
11:31 dRiN joined #salt
11:33 lothiraldan joined #salt
11:34 z3uS joined #salt
11:39 kormoc joined #salt
11:41 joehh1 ~.
11:48 bhosmer joined #salt
11:48 Furao joined #salt
11:50 pdayton joined #salt
11:51 dotz joined #salt
11:51 aqua^mac joined #salt
11:51 jonatas_oliveira joined #salt
12:00 Furao joined #salt
12:07 Twiglet joined #salt
12:08 Jahkeup joined #salt
12:12 TheThing joined #salt
12:20 diegows joined #salt
12:24 zadock joined #salt
12:25 sfritz joined #salt
12:26 Jimlad joined #salt
12:26 sfritz hi, I guess I encountered a bug/feature, is there a public bugtracker somewhere?
12:27 sfritz When I declare a "cmd.wait" that contains a "unless" and a "watch" the cmd is executed twice.
12:28 sfritz so, when watch always kicks in, ignoring any return from the unless
12:30 sfritz You find an example in the gitlab-formula "gitlab-initialize", see https://github.com/saltstack-formulas/gitlab-formula/blob/master/gitlab/gitlab.sls (row #90)
12:31 viq sfritz: https://github.com/saltstack/salt/issues
12:33 jtanner_ joined #salt
12:37 VSpike I'm browsing salt formulas to see what's available and what they can do (so I don't re-invent the wheel) and to try and learn from their style. I'm confused by https://github.com/saltstack-formulas/salt-formula ... how *can* salt salt itself?
12:37 CeBe1 joined #salt
12:38 VSpike I can see that you might have a box with a running minion, and you want to turn it into a configured master. But how can you salt a salt minion?
12:39 smoothify joined #salt
12:40 viq VSpike: there's also masterless, local mode
12:40 jtanner__ joined #salt
12:42 ericof joined #salt
12:42 VSpike ah, I wondered if it was some clever trick using salt-ssh (which I haven't looked at yet, I confess) to bootstrap a system from nothing
12:44 VSpike I was just wondering, because on my to-do list is to create bootstrap script for windows boxes and ubuntu boxes that will do the bare minimum requred to get a minion working and talking to the master from a totally vanilla install
12:44 VSpike Again, wanted to make sure I wasn't re-inventing the wheel
12:44 viq salt-ssh is another way
12:45 ZombieTwiglet joined #salt
12:46 VSpike salt-ssh still needs a package installed on the remote though, right? And ssh enabled and working (which means networking configured and working)
12:46 viq no, just python and couple other python libs
12:47 viq And yes, networking, ssh & co
12:47 VSpike Oh, really? That's clever.
12:47 viq salt-ssh is very similiar to ansible in that respect
12:49 jtanner_ joined #salt
12:52 Morbus joined #salt
12:56 sfritz viq: That's strange ... just looked on github and it said issues is deactivated. However, it works now.
12:57 jtanner__ joined #salt
12:58 iwishiwerearobot joined #salt
12:59 sfritz the issue already exists: https://github.com/saltstack/salt/issues/17177
13:00 Kelsar are grains cached on the master?
13:00 viq Kelsar: yes
13:00 Kelsar damn
13:00 viq why?
13:00 Kelsar viq: who can i force a update?
13:01 Kelsar viq: i can reach a minion with salt, but not via ssh, so i guess the ip has changed
13:01 Kelsar ok, cmd.run ip proved this
13:03 babilen Kelsar: You can use http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.saltutil.html#salt.modules.saltutil.sync_grains
13:04 Kelsar babilen: ty
13:04 Kelsar next problem... why is my dns not updating...
13:05 babilen Kelsar: I'd argue that that is a question you might want to discuss in a different channel (whatever is appropriate for your DNS software)
13:05 Kelsar babilen: sure, didn't plan to do it here
13:06 thawes joined #salt
13:09 otter768 joined #salt
13:09 forcer joined #salt
13:11 lothiraldan joined #salt
13:17 calvinh joined #salt
13:18 TheThing joined #salt
13:19 goudale joined #salt
13:21 elfixit joined #salt
13:25 TheThing joined #salt
13:32 TheThing joined #salt
13:32 iwishiwerearobot joined #salt
13:38 scott2b joined #salt
13:41 ericof joined #salt
13:42 toastedpenguin joined #salt
13:43 toastedpenguin joined #salt
13:43 cpowell joined #salt
13:45 beeno joined #salt
13:45 toastedpenguin joined #salt
13:46 TheThing joined #salt
13:47 ml_1 joined #salt
13:51 Slimmons joined #salt
13:51 sfritz left #salt
13:53 Slimmons Trying to set up a new salt-minion (i've got 6 successfully set up), and I ran these commands (ubuntu 14.04 server)    sudo add-apt-repository ppa:saltstack/salt, sudo apt-get install python-software-properties, sudo apt-get install software-properties-common, sudo apt-get update, sudo apt-get install salt-minion, I get "The following packages have unmet dependencies: salt-minion: Depends: salt-common (.......and 5 other depends for p
13:54 Slimmons python-msgpack, python-zmq, dctrl-tools, Unable to correct problems, you have held broken packages.  Any ideas?
13:55 Slimmons Found this, and it appears to be pretty old https://github.com/saltstack/salt-bootstrap/issues/71
13:58 dimeshake http://askubuntu.com/questions/223237/unable-to-correct-problems-you-have-held-broken-packages
13:59 CeBe1 joined #salt
14:00 jeddi joined #salt
14:01 maxd joined #salt
14:04 Slimmons using aptitude didn't help, and I have no output from dpkg --get-selections | grep hold
14:04 Slimmons aptitude just says the same thing.
14:05 aquinas joined #salt
14:05 giantlock joined #salt
14:06 pdayton joined #salt
14:08 nitti joined #salt
14:10 gngsk joined #salt
14:12 dimeshake try the bootstrap script instead Slimmons
14:12 FRANK_T joined #salt
14:12 dimeshake remove the ppa first
14:12 dimeshake and do a full apt-get update / apt-get upgrade
14:13 Slimmons sry, kind of new.  How would I remove the ppa?
14:14 dimeshake sudo add-apt-repository --remove ppa:whatever/ppa
14:14 Slimmons thanks
14:15 Slimmons sorry again, but where would the bootstrap script be?
14:15 racooper joined #salt
14:15 dimeshake use google
14:15 dimeshake salt bootstrap
14:16 jtanner joined #salt
14:18 cvvs joined #salt
14:18 Slimmons same errors,
14:18 Slimmons this is crazy.  I installed salt-minion on 5 other servers that were mirrors of this one.
14:19 dimeshake put up a pastebin of the full output including the command you used - pastie.org
14:23 CeBe1 joined #salt
14:23 blaffoy joined #salt
14:24 blaffoy Hey all. Question: if minion ID is not explicitly set in a config file, how does the ID automatically get assigned?
14:25 dimeshake http://docs.saltstack.com/en/latest/faq.html#how-does-salt-determine-the-minion-s-id
14:25 _JZ_ joined #salt
14:25 cvvs uses the hostname
14:25 CeBe2 joined #salt
14:26 viq and/or revdns
14:27 blaffoy I had thought that it was from socket.getfqdn(), but this doesn't seem to be the case. On a minion (Ubuntu 14.04 on AWS EC2) the command `python -c "from socket import getfqdn; print getfqdn()"` returns aws-ops-001.my-org.local
14:27 blaffoy from bash, `hostname --fqdn` returns the same
14:27 goal I am trying to use cmdmod.run in mine_functions pillar to run a command and get the output back into the salt mine. but the minion logs '[salt.loaded.int.module.mine][ERROR   ] Function cmdmod.run in mine_functions not available'
14:27 dimeshake what'd the minion id end up as?
14:27 goal any idea?
14:28 cvvs Dimeshake's link has a link which describes how the hostname is determined in detail.
14:28 blaffoy But if I don't set "id: aws-ops-001.my-org.local" in the /etc/salt/minion, then the minion id appears to be ip-10-3-10-44
14:28 ggoZ joined #salt
14:28 blaffoy Which is the default name format issued by EC2
14:28 dimeshake blaffoy: check /etc/hosts too - i've seen that when it isn't set there, too
14:30 CeBe1 joined #salt
14:33 blaffoy dimeshake: I just tried adding the line "10.3.10.44 aws-ops-001.my-org.local aws-ops-001 " to /etc/hosts
14:33 blaffoy No luck
14:33 thawes joined #salt
14:33 CeBe2 joined #salt
14:33 rogst joined #salt
14:34 goal in fact, even when I remove cmd.run from the mine_functions, the minion is still trying to execute it. Why isn't the minion seeing the updated pillar mine_functions
14:34 dimeshake blaffoy: you're removing /etc/salt/minion_id and restarting salt-minion right?
14:35 dotz joined #salt
14:36 blaffoy dimeshake: ...
14:36 blaffoy I was not removing minion_id
14:36 blaffoy Right
14:36 blaffoy That did it.
14:37 dimeshake :D
14:37 blaffoy Thanks!
14:38 scott2b joined #salt
14:39 blaffoy That was not clear from the docs. Maybe I missed something.
14:40 lothiraldan joined #salt
14:42 ecdhe joined #salt
14:42 perfectsine joined #salt
14:43 dotz joined #salt
14:45 Furao joined #salt
14:49 si joined #salt
14:49 si left #salt
14:49 jngd joined #salt
14:50 dimeshake blaffoy: i think i discovered that myself as well
14:51 saru11 joined #salt
14:54 cvvs The minion_id cache is fairly new, correct?
14:54 ttrumm joined #salt
14:54 rojem joined #salt
14:55 cvvs Or not… over a year ago.
14:55 ttrumm_ joined #salt
15:04 glyf joined #salt
15:04 kermit joined #salt
15:07 housl joined #salt
15:08 dotz joined #salt
15:08 CeBe joined #salt
15:10 otter768 joined #salt
15:11 CeBe1 joined #salt
15:11 kaptk2 joined #salt
15:13 DaveQB joined #salt
15:14 saru11 when you call state.highstate with salt-call command from CLI  the output is nicely formatted and you can verify immediately what states were executed successfully and what not
15:14 vbabiy joined #salt
15:15 iggy blaffoy: open an issue (or better yet, send a PR)
15:15 saru11 if I schedule state.highstate execution with salt built-in scheduler I can check only minion log file and according the log file level set I can see more or less details. but this is quite clumsy and it's not directly visible what failed and what not.
15:16 saru11 is it somehow possible how to get something like salt-call command output to minion log file?
15:18 thedodd joined #salt
15:20 dimeshake saru11: you can use a --return option to get the output pretty much anywhere you want
15:20 dimeshake syslog might be good for you
15:20 saru11 wth! how could I overlook returners!
15:20 saru11 thanks!
15:22 __TheDodd__ joined #salt
15:22 saru11 ok, but returner is also valid ofr salt or salt-call command only
15:24 rickh563 joined #salt
15:24 dimeshake valid for both
15:24 cvvs joined #salt
15:24 dimeshake remember that the return comes from the minion so it'll be in minion syslog (if that's the returner you use) not master
15:24 saru11 I found it in scheduler docu: http://docs.saltstack.com/en/latest/topics/jobs/schedule.html
15:24 saru11 going to give it a try
15:27 dotz joined #salt
15:28 ksk joined #salt
15:28 ksk hey guys
15:29 thawes joined #salt
15:32 CeBe1 joined #salt
15:34 VSpike Is there a way in the top.sls format to match multiple globs to one set of sls files?
15:34 dimeshake compound matching
15:35 lothiraldan joined #salt
15:35 VSpike dimeshake: Ooh, I missed that one. Thanks!
15:36 overyander does anyone know the correct way to check if a package is installed when using onlyif? here's my state and the error from the master cli as well as debug from the minion log.  http://pastebin.com/tQfaHWYr
15:37 Twiglet joined #salt
15:38 perfectsine joined #salt
15:38 iggy I would think it would be - pkg.installed: net_framework
15:38 theologian joined #salt
15:39 iggy and for the sake of our eyes and not being tracked by every ad network in existence, use something other than pastebin.com
15:42 overyander thanks iggy. i corrected that, going to reboot and see if the highstate applies properly now. do you have a suggested pastebin?
15:42 iggy gist refheap or one of the distro ones are usually a good bet
15:43 cvvs left #salt
15:45 cvvs joined #salt
15:48 dimeshake i rather like pastie.org - because it's easy to remember :D
15:55 schlueter joined #salt
15:56 iggy really, any of them that don't run a shit ton of obtrusive ads are fine
15:56 mpanetta joined #salt
15:57 viq bonus points if they work without javascript and/or via TOR ;)
15:58 iggy let's get rid of the ads first
15:58 dimeshake https://ptpb.pw/
15:58 dimeshake someone posted that one - some interesting features
16:00 mpanetta joined #salt
16:02 Auroch joined #salt
16:14 budman joined #salt
16:16 scott2b joined #salt
16:17 conan_the_destro joined #salt
16:18 scott2b When using the mongodb formula, how are folks bootstrapping their replica set configurations? It seems like, in addition to getting a replSet parameter into the config file, mongo requires an rs.initiate and for each node an rs.add in the shell. I'm not sure of the best way to handle that with salt.
16:22 jonbrefe joined #salt
16:23 dotz joined #salt
16:28 forrest joined #salt
16:30 smcquay joined #salt
16:30 dotz joined #salt
16:32 cotton is anyone using docker with salt?  is there any get the actual docker command it runs?
16:32 alexbst joined #salt
16:37 davesque joined #salt
16:38 iggy cotton: run with debug? it'll be in there
16:38 iggy although I wouldn't do that for long, you'll fill up logs quite fast
16:40 bhosmer joined #salt
16:40 _JZ_ joined #salt
16:42 Furao joined #salt
16:42 Ozack joined #salt
16:51 Furao joined #salt
16:51 pdayton joined #salt
16:52 mapu joined #salt
16:53 jngd joined #salt
16:55 bhosmer joined #salt
16:55 StDiluted joined #salt
16:58 rlima joined #salt
16:58 KyleG joined #salt
16:59 StDiluted joined #salt
17:01 zadock joined #salt
17:02 wendall911 joined #salt
17:02 ingwaem joined #salt
17:03 rlima please, I need help with /key API call....I 've set up the rest_cherrypy in a conf file, and the /login is working, but the /keys not working. Anyone know what I should do ?
17:07 perfectsine joined #salt
17:08 dotz joined #salt
17:11 otter768 joined #salt
17:12 ggrieves joined #salt
17:14 thedodd joined #salt
17:15 housl joined #salt
17:15 micah_chatt joined #salt
17:16 dotz joined #salt
17:17 zlhgo joined #salt
17:17 diegows joined #salt
17:18 jngd joined #salt
17:18 lendar joined #salt
17:19 thedodd joined #salt
17:22 ajolo joined #salt
17:34 teebes joined #salt
17:35 desposo joined #salt
17:37 ericof joined #salt
17:39 dotz joined #salt
17:43 scott2b joined #salt
17:43 nitti_ joined #salt
17:45 smcquay joined #salt
17:47 smcquay joined #salt
17:49 aparsons joined #salt
17:50 funzo joined #salt
17:51 theologian joined #salt
17:51 frvge joined #salt
17:55 goudale joined #salt
17:55 rojem joined #salt
17:56 frvge Hi, is it possible to re-load modules during a highstate? I'm using salt to install git while provisioning a vagrant box, but subsequent packages (which include my git.sls file) can't use git.latest, because at the time of the loading of salt git was not installed.
17:56 bhosmer joined #salt
17:57 frvge If I run vagrant up --provision, it works like a charm, because then the git module gets activated.
17:57 frvge (so the second try it works)
17:58 jonbrefe joined #salt
17:58 frvge Version is: salt-call 2014.7.0 (Helium)
17:58 cvvs Will saltutils.sync_all() fix it?
17:58 bhosmer joined #salt
17:59 desposo joined #salt
18:00 frvge I'm using a masterless minion set-up for Vagrant. I don't think it's possible to call that from an sls file, but I might be wrong.
18:00 cvvs Oh, never mind. I have little experience with git module, and even less with masterless.
18:00 frvge the only control I have is vagrant up... and then it runs highstate on my sls files.
18:01 rap424 joined #salt
18:01 frvge My own experience is also still low :)
18:01 viq frvge: saltmod ?
18:01 viq http://docs.saltstack.com/en/latest/ref/states/all/salt.states.saltmod.html
18:02 frvge Seems like it's supposed to work on masters (which I don't have) but I can give it a try
18:02 viq ah, no, I was thinking of http://docs.saltstack.com/en/latest/ref/states/all/salt.states.module.html
18:03 frvge Maybe module.run will work
18:04 thedodd joined #salt
18:05 smcquay joined #salt
18:07 meylor joined #salt
18:09 iggy frvge: reload_modules: True
18:10 iggy (although a pkg.install should do that automatically)
18:12 thedodd joined #salt
18:13 frvge I'm building from source
18:13 RedundancyD joined #salt
18:13 viq any particular reason why?
18:14 frvge Gitlab needs a higher git version than Ubuntu's packages provides.
18:14 viq Wouldn't there be a ppa or something with a newer git?
18:14 viq Or maybe their omnibus packges would do?
18:15 frvge AFAIK we don't use omnibus on production. It has to match. :-(
18:15 viq Ah. In that case I'd look for a repo with proper version of packages. Or use FPM once, and use those packages
18:16 ericof joined #salt
18:17 frvge reload_modules seems to have done the trick
18:17 overyander i'm still running into issues trying to use onlyif  i've tried using pkg.installed and pkg.version both result in the same error http://pastie.org/9788593
18:17 perfectsine joined #salt
18:18 jaimed joined #salt
18:18 ggrieves joined #salt
18:19 bhosmer joined #salt
18:20 frvge overyander, as far as I know, you only need to specify the module. So pkg: net_framework
18:20 overyander iggy, any other ideas  ^^^
18:20 ttrumm_ joined #salt
18:20 overyander ?
18:20 smcquay joined #salt
18:20 frvge I could be wrong though :)
18:20 iggy overyander: what version of salt
18:21 overyander 2014.7.0-1-win32 is running on minion 2014.7.0-3 is running on master
18:22 ggrieves Hi,  So I'm just now getting to the point where I can use file.managed but I'm not the only person that can alter files on servers.  I'm wondering what are some good ways of automating and monitoring file.get_diff?  Do you call state.highstate from a script that parses the returns for important changes or? what's best practice?
18:22 Furao joined #salt
18:22 iggy overyander: onlyif wants commands to run
18:23 overyander ok, so this takes me back to my original question from yesterday. how can i run chocolatey.bootstrap only when .net framework 4.5.2 is already installed?
18:23 neogenix joined #salt
18:23 neogenix left #salt
18:24 neogenix joined #salt
18:24 neogenix Hello!
18:24 iggy is there some command that you can run that only returns true if .net 4.5 is installed?
18:24 iggy I'm not a Windows guy... so nfc really
18:25 frvge unless: framework --version | grep 4.5 ? But then the windows equivalent?
18:25 alexbst a powershell call I think
18:25 overyander no, the ms recommended way to check .net version is to look at a hex value in the registry
18:25 alexbst take a look at the dotnet cookbooks for chef, they might hold somethingto do it.
18:25 neogenix I’m installing RVM with Salt, and for some reason salt isn’t picking up /etc/profile.d/rvm.sh environment variables/changes, and just keeps trying to reinstall RVM. How do I make sure that salt actually reads environment variables from /etc/profile*
18:26 iggy uhhh... well... good luck then
18:26 overyander ok, thanks
18:26 FRANK_T I have a syntax question https://www.hobapolis.com/paste/?c0b8cb6ff6b1f560#OK4tx68PZvihsfOMgyOCt5gjt2dyL2noW/r31iNb1EI=
18:26 FRANK_T this will stop the httpd services
18:26 iggy neogenix: try reload_modules: True in whatever state you are installing rvm with, otherwise, you'll have to restart the daemon
18:26 stephas joined #salt
18:27 FRANK_T How about if I want to Stop a list of services Is an way to put a list of services without  copy and paste the same syntax with different service name
18:27 smoothify joined #salt
18:28 mindfab Hi, whom do I have to contact if I want to share something on https://github.com/saltstack-formulas ?
18:28 iggy FRANK_T: {% for svc in ['list', 'of', 'services'] %} ?
18:28 iggy mindfab: completely new formula? or changes to an existing one?
18:29 mindfab completly new...
18:29 FRANK_T iggy let me test! thanks
18:30 mindfab I have already created a github repo...
18:31 iggy mindfab: make it under your user, then contact one of the people (I suggest babilen) to make a new formula, then you pull request yours to it like normal (iirc)
18:31 mindfab Okay, thx
18:32 micah_chatt joined #salt
18:32 FRANK_T iggy do I have to close something @ the end
18:32 desposo joined #salt
18:32 FRANK_T Its giving me an error
18:32 conan_the_destro joined #salt
18:32 iggy FRANK_T: yes, there's more to it than what I put... that was more the beginning of an idea
18:33 FRANK_T Rendering SLS "base:service" failed: Jinja syntax error: unexpected end of template, expected 'block_end'.; line 1
18:33 FRANK_T This is what I have https://www.hobapolis.com/paste/?95c3e53d9f7cc40b#pTYE71PgAcT5fXrZB1HtKUolwali/vK9EK9qKTFD1H0=
18:34 micah_chatt_ joined #salt
18:34 davesque joined #salt
18:34 iggy more like https://www.hobapolis.com/paste/?038d012532b3ecea#YGMmJ+BJo1/TdUIJBtPs9dbO0aval/92YetjKikmgTI=
18:34 smcquay joined #salt
18:36 neogenix iggy: thanks, I’ll give it a whirl.
18:37 FRANK_T Good! iggy thanks
18:38 bhosmer_ joined #salt
18:40 Ryan_Lane joined #salt
18:44 pdayton joined #salt
18:46 CryptoMer Hey, I'm running into a problem where a list of packages is failing to install, because one of the packages is already installed. Is there a good workaround for this? http://pastebin.com/8tjFa0qC
18:56 ggoZ joined #salt
18:57 scott2b joined #salt
18:58 dotz joined #salt
19:00 pdayton joined #salt
19:00 zadock joined #salt
19:04 monkey66 joined #salt
19:04 FRANK_T CryptoMer
19:04 FRANK_T https://www.hobapolis.com/paste/?1f54b96fb0c70e20#uBu7Tm2G0Zx+9w9drB7Veq+BFKZNG1364/Gc6ngQ2+0=
19:04 FRANK_T try this..
19:05 CryptoMer thanks FRANK_T, I'll give it a go.
19:05 FRANK_T ok
19:06 ttrumm joined #salt
19:06 zadock joined #salt
19:06 smcquay joined #salt
19:07 frvge iggy: thanks for the reload_modules hint, my vagrant up is now completing successfully :)
19:08 desposo joined #salt
19:12 otter768 joined #salt
19:15 * iggy adds another checkbox in the problems solved column
19:17 dthorman joined #salt
19:17 FRANK_T iggy I was able to resolve my problem thank you.
19:19 davesque joined #salt
19:21 CryptoMer FRANK_T: Thanks for the assist. That resolved it.
19:22 CryptoMer I now have a different problem. I'm storing some data in a pillar, and using jinja templating in a script to make sure the data gets imported into the script. However, the pillar data is being encapsulated in [' ']. Is tehre a way around this? relevant code found here: http://pastebin.com/HiBYy4qk
19:25 druonysus joined #salt
19:26 thawes joined #salt
19:28 jalbretsen joined #salt
19:30 rojem joined #salt
19:31 CryptoMer Ahh, I figured it out, thanks to a posting from 2012: https://groups.google.com/forum/#!msg/salt-users/-pV2KA_Sdfs/m0bNvETZDnEJ. End result, my pillar data was being parsed as a list. Removing the preceding '-' in the pillar data resolves the issue.
19:33 neogenix I’m getting a ‘Unkown Yaml Render’ and I’ve no idea why. What looks wrong with this: http://pastebin.com/MpZNs9h6
19:34 CryptoMer no colon on line 2.
19:34 neogenix CryptoMer: halleluja! thank you!
19:34 CryptoMer heh, it's *always* something dumb. hah
19:36 smcquay joined #salt
19:41 cotton joined #salt
19:42 StDiluted_ joined #salt
19:42 monkey66 left #salt
19:43 victor- joined #salt
19:43 victor- hi all.. trying to saltstack today. i have a directory of files, each containing a public key, that i'd like SS to turn into users. can someone point me to documentation or an example of how to do this?
19:43 neogenix CryptoMer: now, if you could tell me why salt isn’t seeing that rvm is installed, but it’s totally there… that’d be great ;)
19:43 victor- is there a way to iterate through files in a directory within an SLS file?  or would i need to have a pillar yml that defines user names that match the public key names where i can use a jinja template?
19:44 theo joined #salt
19:45 nitti joined #salt
19:46 neogenix iggy: reload_modules didn’t seem to do the trick, however in looking at the logs I’m not sure why it’s not seeing rvm installed. From the cmd line I can do a rvm version/install/etc… : http://pastebin.com/KM0jJUNt
19:47 Ahlee victor-: {% for username in salt['cmd.run']('ls /home/').split() %}
19:48 victor- oh, salt['cmd.run']. duh.
19:48 iggy neogenix: most system daemons don't get a full environment (i.e. from sourcing profile) do you _need_ to rely on those env variables
19:49 neogenix iggy: rvm.install seems to use the rvm variables to be able to install the ruby environments.
19:49 neogenix iggy: salt seems to *want* them, although tbh, I haven’t looked int he code to see how it tries to find rvm.
19:50 Ahlee That's why I just write bash scripts for salt to execute.
19:50 Ahlee environment variables, login/interactive vs noninteractive, etc
19:50 cberndt joined #salt
19:51 iggy neogenix: is rvm installed to a system location? or under a user?
19:52 SneakyPhil joined #salt
19:52 neogenix iggy: system location (ran as root). Although I’ve tried to install it as user rvm, and then user root.
19:52 neogenix iggy: rvm.is_installed returns True, rvm.version returns ‘unavailable’
19:52 SneakyPhil hey everybody
19:52 iggy oh good lord, that module is pulling something from some random user's github account to install with
19:52 * iggy runs away quickly
19:53 SneakyPhil when I restart a salt minion server, the salt master never realizes that that server comes back online
19:53 iggy test.ping it a few times
19:53 SneakyPhil I have confirmed that the salt-minion service is running and I have a connect from the minion to the master
19:54 SneakyPhil hey thanks, that is odd
19:55 SneakyPhil I just rolled it out into my environment today and was able to patch x86_64 kernels due to http://seclists.org/oss-sec/2014/q4/1052 . tl;dr Salt is awesome
19:55 iggy I think we're all hoping raet will solve things like that
19:58 spookah joined #salt
20:01 neogenix wow, this rvm module just behaves crazy…
20:01 linuxdaemon joined #salt
20:01 dimeshake blame ruby devs
20:02 neogenix oh,I totally have no love for ruby, but this module behaves nutty
20:05 iggy if I saw that and had to do something with ruby, I'd roll my own
20:05 iggy just saying..
20:07 CeBe1 joined #salt
20:08 neogenix at this rate I’ll just rebuild the whole thing as cmd.run’s.
20:09 jonbrefe joined #salt
20:09 neogenix iggy: this totally looks like module pokery: rvm.is_installed = True, rvm.is_installed runas=root = False
20:09 neogenix *blink*
20:10 pdayton joined #salt
20:10 davesque joined #salt
20:10 iggy when you set runas, it looks at $RUNAS_HOME/.rvm/bin/rvm
20:10 iggy when you don't set it at all, it is hardcoded to look at /usr/local/rvm/bin/rvm
20:11 iggy I don't even use rvm (and hope I never have to), but I'm about to file some issues on this module
20:11 iggy or just one with a bunch of complaints in it
20:12 neogenix iggy: I’d be more than happy to help out on that :)
20:13 neogenix also, bah! can’t get the damned thing to import my HTTP_PROXY environment variable
20:22 theo__ joined #salt
20:32 Ozack1 joined #salt
20:35 kballou joined #salt
20:35 overyander wha tis the correct way to express this?  {% if salt.cmd.run('for /f "tokens=1,2,3" %i in ('reg query "HKLM\SOFTWARE\Microsoft\NET Framework Setup\NDP\v4\full" ^| findstr Release') do echo %k') = '0x5cbf5') %}
20:36 bhosmer_ joined #salt
20:36 overyander i know the command works. I keep getting Jinja syntax errors when testing the state
20:39 londo joined #salt
20:45 stan_k overyander: you've got an extra close paren, at least, '0x5cbf5')
20:45 overyander thanks stan_k
20:46 iggy and I think it's salt['cmd.run'](....)
20:46 iggy at least most of the other modules are that way
20:47 overyander i found this http://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.jinja.html#calling-salt-functions
20:47 overyander trying {% if salt['cmd.run']('for /f "tokens=1,2,3" %i in (''reg query "HKLM\SOFTWARE\Microsoft\NET Framework Setup\NDP\v4\full" ^| findstr Release'') do echo %k') == '0x5cbf5' %}        gives same results of syntax error
20:48 schlueter joined #salt
20:52 dotz joined #salt
20:52 stan_k perhaps a quoting problem?
20:53 stan_k sorry, jinja does some wizardry that i constantly have trouble following.
20:53 stan_k er, consistently
21:01 overyander stan_k, tell me about it. i think voodoo would be easier to practice than using jinja in salt
21:04 iggy it's not that bad, but yeah, I would try to escape those single quotes with a backslash
21:05 overyander iggy, same error    {% if salt['cmd.run']('for /f "tokens=1,2,3" %i in (\'reg query \"HKLM\SOFTWARE\Microsoft\NET Framework Setup\NDP\v4\full" ^| findstr Release\') do echo %k') == '0x5cbf5' %}
21:06 rojem joined #salt
21:06 overyander just removed that backslash from behind the doublequote
21:06 schlueter joined #salt
21:06 overyander {% if salt['cmd.run']('for /f "tokens=1,2,3" %i in (\'reg query "HKLM\SOFTWARE\Microsoft\NET Framework Setup\NDP\v4\full" ^| findstr Release\') do echo %k') == '0x5cbf5' %}
21:06 overyander same error
21:06 skyler_ joined #salt
21:07 skyler_ Anyone here using the swift module? I am getting an error when I use it, but the swift client works from the command line (not using salt).
21:08 wt joined #salt
21:09 iggy overyander: why don't you try pasting the full error
21:10 iggy neogenix: https://github.com/saltstack/salt/issues/19105 feel free to add anything else you see
21:13 otter768 joined #salt
21:14 smcquay joined #salt
21:15 overyander iggy, http://pastie.org/9788899
21:16 skyler_ Here is the full swift error (with some internal details omitted): https://gist.github.com/skylerberg/f952f24ce240cd5a04a8
21:16 zadock joined #salt
21:16 iggy overyander: \\ the reg query (KKLM\\SOFTWARE\\....)
21:18 iggy the "malformed \N character escape" was kind of a dead giveaway
21:21 rbjorklin joined #salt
21:28 kickerdog joined #salt
21:28 lionel joined #salt
21:28 kickerdog does anyone remember the boot option to start linux cd in ram?
21:29 bmonty joined #salt
21:31 overyander iggy, now it just skips that if statement entirely. doesn't show in cli output from "salt-call state.highstate test=True" nor does it show up in the debug logs
21:32 iggy try running with salt-call -l debug
21:32 iggy I don't really know, this seems like some Windows oddness
21:34 raner joined #salt
21:34 active8 joined #salt
21:38 overyander iggy, what about {% if salt['pkg.installed']('net_framework') %}   ?? could that work or can you not use pkg.installed like that?
21:38 iggy no
21:38 iggy pkg.installed is for installing packages, not see if something is installed
21:38 overyander pkg.version ? that returns false if the package isn't installed
21:41 skyler_ If you make a bugfix that you need for production, how do you integrate that into your setup? I mean, I am running 2014.7.0, and I don't want to wait for another release to get my fix.
21:42 aurynn this is where running your own apt/yum server is useful
21:43 cpowell joined #salt
21:44 giantlock joined #salt
21:44 ckao joined #salt
21:45 Cidan joined #salt
21:46 Edgan We run in AWS. So I put my yum repo in S3, and serve it via Cloudfront.
21:47 skyler_ Do you guys usually just have the packages that you have made/changed in your apt/yum servers or do you mirror all the packages for your distro?
21:47 Edgan skyler_: I do both
21:47 vectra joined #salt
21:47 aurynn I haven't needed to patch master yet, so I'm running stock packages, but I will need to set this up
21:47 aurynn eventually
21:47 Edgan skyler_: The advantage of mirror all the packages is speed and control.
21:48 Edgan skyler_: I also set a priority in yum so that my packages always override stock. So if I update say grub2, even if an update comes out, it stays with my patched grub2.
21:49 kormoc joined #salt
21:50 Edgan skyler_: If you go down this road, fpm is your friend.
21:50 skyler_ Edgan: Do you have a mirror and then a separate server for custom packages? Or do you run a single server providing a superset of the distro's packages?
21:50 spookah joined #salt
21:50 Edgan skyler_: I mirror and create the repo on one server, and then push a copy to S3 to be served by Cloudfront. So I hava high availability.
21:51 Edgan skyler_: The stock stuff I leave the repos intact
21:51 thawes joined #salt
21:51 Edgan skyler_: My stuff goes into my separate repo
21:52 * aurynn takes notes
21:52 skyler_ Edgan: Yeah, now I need to learn the details of how packaging actuallly works. I have a mirror, but I actually don't know that much about the internals of a deb or anything like that.
21:53 kickerdog1 joined #salt
21:54 davesque joined #salt
21:54 Edgan skyler_: you can apt-get source packagename to get source for stock stuff, if you are using stock repos
21:54 Edgan skyler_: That gives you what you need to rebuild a deb
21:55 Edgan skyler_: I don't care for dpkg, because most of the definitions for a package our auto-magic. It makes it harder to learn. Everything is a override of defaults.
21:55 Edgan s/our/are/g
21:56 Edgan skyler_: Last time I was using apt I was looking at using freight.
21:58 aurynn fpm is a great tool for generating dpkg packages
21:59 skyler_ Edgan: This is should be a good learning experience. Looks like I have a lot more to do than I originally bargained for with my current task :)
22:00 Edgan aurynn: yeah, it can't be used 100% of the time, but it is great for 80% of packages
22:00 iggy skyler_: there's also the possibility of just using the updated module if it's not in core code
22:01 skyler_ iggy: Patching it manually?
22:01 yomilk joined #salt
22:01 cpowell joined #salt
22:02 iggy no, say the bug you're worried about is in the aptpkg module, you can just shove that in _modules in your state tree and it'll get used
22:02 smoothify joined #salt
22:03 skyler_ iggy: Oh, I was not aware that was possible, that is awesome! I think the problem is in the utils folder, though.
22:04 kickerdog joined #salt
22:04 iggy maybe next time...
22:04 skyler_ I can actually fix the problem by using cmd.run instead of the module, but eventually I would like to use the module directly. So I can work around the issue for now and implement a good solution later.
22:05 skyler_ iggy: thanks for letting me know about that. That is a seriously cool feature I had no idea about.
22:07 cberndt joined #salt
22:07 JPaul joined #salt
22:08 Gareth_ joined #salt
22:18 teebes joined #salt
22:20 victor- how would i create a user and upload a specific ssh private key?
22:21 __number5__ victor-: look at user state and ssh_auth state
22:22 victor- __number5__, i see how i can include a public key in ssh_auth for authorized_keys.. i'd like to upload a pre-made private key for that user
22:23 victor- how would one copy a file from the pillar into a location?
22:24 __number5__ victor-: use file.managed with contents_pillar
22:24 iggy look at the users-formula, I think it does some of this stuff for you
22:26 __number5__ victor-, iggy: yep, https://github.com/saltstack-formulas/users-formula/blob/master/users/init.sls#L87
22:27 victor- iggy, __number5__ just found it. thanks
22:32 badon joined #salt
22:33 thawes joined #salt
22:33 funzo joined #salt
22:43 StDiluted joined #salt
22:46 CeBe1 joined #salt
22:47 hal58th1 joined #salt
22:52 RainDoctor joined #salt
22:53 jonbrefe joined #salt
22:57 kermit joined #salt
22:57 thehaven_ joined #salt
22:58 glyf_ joined #salt
22:59 CryptoMer Is anyone here practiced in using the GPG renderer?
22:59 CryptoMer I'
23:00 CryptoMer I've been following the docs on 'http://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.gpg.html', but salt is failing to render on this portion of the PGP message:
23:00 ninkotech__ joined #salt
23:00 CryptoMer Version: GnuPG v2.0.14 (GNU/Linux)    <======================
23:00 ckao1 joined #salt
23:00 garthk_ joined #salt
23:01 druonysuse joined #salt
23:01 druonysuse joined #salt
23:02 wedgie_ joined #salt
23:03 muebel_ joined #salt
23:03 sk_0 joined #salt
23:04 KyleG joined #salt
23:04 KyleG joined #salt
23:06 jpaetzel joined #salt
23:06 __number5__ joined #salt
23:06 samed joined #salt
23:06 jab416171 joined #salt
23:06 DaveQB joined #salt
23:06 bernieke joined #salt
23:06 peters-tx joined #salt
23:06 racooper joined #salt
23:06 jacksontj joined #salt
23:07 flebel joined #salt
23:07 analogbyte joined #salt
23:08 gmoro joined #salt
23:08 aquinas joined #salt
23:09 conan_the_destro joined #salt
23:11 davesque joined #salt
23:12 overyander joined #salt
23:14 otter768 joined #salt
23:15 aquinas joined #salt
23:17 moos3 anyone have a network formula that is configured by pillar data ?
23:19 fintler joined #salt
23:21 aquinas joined #salt
23:21 scott2b joined #salt
23:22 scott2b Is pillar data available to a custom state python module, or would I need to explicitly pass in the pillar data values as parameters to the custom function?
23:26 dimeshake scott2b: it's available
23:26 dimeshake via __pillar__
23:26 aquinas joined #salt
23:26 scott2b ah, excellent. Thanks!
23:29 perfectsine joined #salt
23:30 smcquay joined #salt
23:32 aquinas joined #salt
23:35 druonysus joined #salt
23:35 druonysus joined #salt
23:38 CeBe1 joined #salt
23:41 mosen joined #salt
23:44 mapu joined #salt
23:46 iggy https://github.com/saltstack-formulas/salt-formula/blob/master/salt/cloud.sls#L2 lol, overwriting global jinja variables... smart
23:51 moos3 salts network module, how does that work with highstate ? I can't seem to get it to apply
23:52 yomilk joined #salt
23:54 KyleG Any idea when I can expect https://github.com/saltstack/salt/pull/18376/commits and https://github.com/saltstack/salt/issues/16139 to be fixed on a release?
23:55 smoothify joined #salt

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