Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2016-03-23

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

All times shown according to UTC.

Time Nick Message
00:10 evaryont joined #salt
00:10 devops left #salt
00:15 nobrak joined #salt
00:15 nobrak joined #salt
00:15 devops joined #salt
00:16 devops joined #salt
00:19 beardedeagle I'm assuming you can't mix and match a state in terms of yaml and pure python. it's either got to be one or the other?
00:27 akhter joined #salt
00:27 mosu_ joined #salt
00:29 berserk joined #salt
00:30 antpa joined #salt
00:33 rhodgin joined #salt
00:34 hemebond Oh, ew, nodegroups are only defined in the master config?
00:35 justanotheruser joined #salt
00:41 hemebond I suppose I can just use a Pillar to group minions instead.
00:45 bhosmer joined #salt
00:45 keimlink_ joined #salt
01:02 DaveQB joined #salt
01:02 edrocks joined #salt
01:07 overyander joined #salt
01:09 RobertChen117 joined #salt
01:11 sarlalian joined #salt
01:11 keimlink joined #salt
01:11 XenophonF joined #salt
01:15 akhter joined #salt
01:21 aqua^c joined #salt
01:21 anmol joined #salt
01:29 iggy devops: unfortunately no
01:29 om joined #salt
01:30 devops :( ok thanks
01:30 joe_n joined #salt
01:31 k_sze[work] joined #salt
01:32 devops left #salt
01:32 mavhq joined #salt
01:39 bhosmer joined #salt
01:41 brianfeister joined #salt
01:46 tristianc joined #salt
01:48 aqua^c joined #salt
01:51 ninjada joined #salt
01:52 Glasswalker Hey, so how do I go about clearing the cache in svnfs from the master? I had to make changes to the svn repo, and now I'm getting errors starting the master...
01:54 kliquori joined #salt
02:01 racooper joined #salt
02:02 berserk joined #salt
02:06 hightekvagabond joined #salt
02:17 edrocks joined #salt
02:19 rem5 joined #salt
02:19 XenophonF absent an actual answer, my approach would be to shut down the master and nuke the appropriate directory under /var/cache/salt/master
02:19 XenophonF after a backup
02:19 XenophonF maybe two backups ;)
02:22 RobertChen117 joined #salt
02:24 iceyao joined #salt
02:31 iceyao joined #salt
02:31 antpa joined #salt
02:32 antpa joined #salt
02:33 guanophobic joined #salt
02:36 catpigger joined #salt
02:37 aharvey joined #salt
02:40 kliquori joined #salt
02:41 evle joined #salt
02:46 zmalone joined #salt
02:46 nZac joined #salt
02:47 ilbot3 joined #salt
02:47 Topic for #salt is now Welcome to #salt! | Latest Version: 2015.8.7 | SaltConf16: http://saltconf.com/register/ | Paid support available: https://www.saltstack.com/support/ | Logs: http://irclog.perlgeek.de/salt/ | Paste: https://gist.github.com/ (please don't multiline paste into channel) | See also: #salt-devel, #salt-offtopic | Ask with patience as we are volunteers and may not have immediate answers
02:51 edrocks joined #salt
03:00 mavhq joined #salt
03:07 iceyao joined #salt
03:16 favadi joined #salt
03:17 anmol joined #salt
03:20 mapu joined #salt
03:26 SheetiS joined #salt
03:26 joe_n joined #salt
03:27 bhosmer joined #salt
03:29 beardedeagle *sigh*, love it when something just stops working with no changes made in the code.
03:30 brianfeister joined #salt
03:32 bhosmer joined #salt
03:33 evaryont joined #salt
03:35 beardedeagle So I see this: https://github.com/saltstack/salt-pack/issues/13, was just wondering if any of the salt dev's could comment on when that is comming. messing with salt-pack is like taking a stab in the dark with a sewing needle.
03:35 saltstackbot [#13]title: add minimum documentation | Would it be possible to add minimum documentation to the readme to explain how to use salt-pack to generate the salt packages?
03:36 hemebond Is that to build masters?
03:37 beardedeagle packages for saltstacks repo
03:37 hemebond Ah
03:38 beardedeagle yeah, writting the cent 7 package for winexe
03:38 beardedeagle but no docs makes it difficult
03:38 beardedeagle and I don't want to do it the way cent 6 is set up since that just seems...wrong
03:38 beardedeagle even dmurphy isn't exactly happy with the way cent 6 winexe is in there
03:40 iceyao_ joined #salt
03:44 ageorgop joined #salt
03:47 wendall911 joined #salt
03:47 rhodgin joined #salt
03:49 timoguin joined #salt
03:51 ramteid joined #salt
03:53 brianfeister joined #salt
03:53 nZac joined #salt
04:01 timoguin joined #salt
04:04 hightekvagabond joined #salt
04:05 antpa joined #salt
04:10 beardedeagle so this is a bit odd. this was working yesterday and this morning. I could spin up vm's and add ad users to the box without issue. now it says it completes with no errors but never runs the state that adds the ad users even though I see the event for it kick off.
04:10 beardedeagle https://gist.github.com/beardedeagle/437bffb5bf057292668a
04:11 berserk joined #salt
04:11 beardedeagle I even set the timeout to something absurd. Like 600 seconds
04:12 beardedeagle not an issue with the state because if I run it manually from command line it works
04:15 beardedeagle the state when ran alone takes about 60 seconds to complete. provisioning a server between 3 to 5 minutes.
04:15 beardedeagle 600 seconds = 10 minutes. so shouldnt be that either
04:18 zmalone joined #salt
04:18 ageorgop joined #salt
04:31 edrocks joined #salt
04:35 kshlm joined #salt
04:38 hightekvagabond joined #salt
04:43 kshlm joined #salt
04:50 onlyanegg joined #salt
04:52 rem5 joined #salt
04:54 RobertChen117 joined #salt
04:54 om joined #salt
04:55 beardedeagle I think it is a race condition that I am hitting
05:01 beardedeagle yup. race condition. eff
05:02 kshlm joined #salt
05:03 sauvin joined #salt
05:07 nZac joined #salt
05:10 edrocks joined #salt
05:18 hemebond Even though you're using a reactor?
05:20 beardedeagle yeah, if I throw a runner.test.sleep in there for like 30 - 60 second between the runner.cloud.profile and the salt.state it works fine
05:21 beardedeagle I think the salt.state tries to run before minion and syndic services are started and that is causing the issue
05:23 antpa joined #salt
05:25 antpa joined #salt
05:26 alemeno22 joined #salt
05:27 kliquori joined #salt
05:28 RobertChen117 joined #salt
05:30 ageorgop joined #salt
05:31 cberndt joined #salt
05:33 bhosmer_ joined #salt
05:41 edrocks_ joined #salt
05:49 favadi joined #salt
05:55 nZac joined #salt
05:59 cberndt joined #salt
06:04 LostSoul Hi
06:04 LostSoul I've got stupid question
06:04 LostSoul Can I use 2 variables of same name?
06:04 LostSoul Like when I have list of dict for crontabs
06:04 LostSoul And I want to have one for each machine and one global
06:04 LostSoul with same crontab list
06:05 hemebond "Can I use 2 variables of same name?" not in the same scope, no.
06:11 LostSoul I see, thanks
06:11 LostSoul I used crontab outside the other states and added global state cron that get each line from pillar
06:11 LostSoul Do you think it's good solution?
06:16 hemebond I don't really understand, sorry.
06:18 edrocks joined #salt
06:22 jfelchner joined #salt
06:22 jfelchner Has anyone else had an issue with gitfs files being fetched *very* slowly?
06:23 jfelchner I'm talking like one per minute.
06:23 jfelchner I'm using pygit2
06:23 antpa joined #salt
06:23 jfelchner Because I need the HTTPS authentication
06:30 om joined #salt
06:35 LostSoul hemebond: Then I'm lost :P
06:37 jfelchner It almost seems like it's making a new connection for EVERY file.  Instead of just cloning the repo down once and grabbing the files off.
06:41 Vaelatern Is there any reason v2015.8.8 isn't marked as a release?
06:48 fxhp joined #salt
06:51 rdas joined #salt
06:54 kshlm joined #salt
06:56 nZac joined #salt
06:58 edrocks joined #salt
07:05 pwalsh joined #salt
07:09 elsmo joined #salt
07:13 mavhq joined #salt
07:13 babilen Vaelatern: But it is (tagged and available in the cheeseshop)
07:17 catpig joined #salt
07:18 illern joined #salt
07:20 aqua^c joined #salt
07:21 bhosmer_ joined #salt
07:24 Rumbles joined #salt
07:29 kliquori joined #salt
07:29 edrocks_ joined #salt
07:32 jaybocc2 joined #salt
07:37 impi joined #salt
07:37 cberndt joined #salt
07:37 keimlink joined #salt
07:40 slav0nic joined #salt
07:43 fooma joined #salt
07:44 KermitTheFragger joined #salt
07:48 cyborglone joined #salt
07:48 jaybocc2 joined #salt
07:51 cyborg-one joined #salt
07:52 cyborgIone joined #salt
07:58 punkoivan joined #salt
07:59 punkoivan joined #salt
08:01 edrocks joined #salt
08:02 kawa2014 joined #salt
08:09 ninjada joined #salt
08:12 favadi joined #salt
08:15 kliquori joined #salt
08:15 txmoose joined #salt
08:15 josuebrunel joined #salt
08:20 TyrfingMjolnir joined #salt
08:25 colegatron joined #salt
08:27 RobertChen117 joined #salt
08:28 mackripeum joined #salt
08:29 dariusjs joined #salt
08:30 ninjada joined #salt
08:31 oida joined #salt
08:33 kawa2014 joined #salt
08:33 jhauser joined #salt
08:38 dmaiocchi joined #salt
08:39 kawa2014 joined #salt
08:43 felskrone joined #salt
08:46 edrocks joined #salt
08:48 keimlink joined #salt
08:53 Acidix joined #salt
08:53 Acidix Hi everybody
08:55 Acidix i've an error to call the salt.scripts module
08:56 Acidix Traceback (most recent call last):   File "/usr/bin/salt-master", line 6, in <module>     from salt.scripts import salt_master ImportError: No module named salt.scripts
08:56 Rumbles joined #salt
08:59 mavhq joined #salt
09:00 keisetsu joined #salt
09:03 VR-Jack3 joined #salt
09:04 dunz0r Acidix: What's the command you're running?
09:07 antpa joined #salt
09:13 Acidix I've this error for "~$ sudo salt-* #master, key, minion, etc ..." I think it's a bad install
09:16 fredvd joined #salt
09:16 Trauma joined #salt
09:17 babilen Acidix: What is the *exact* command you are running and what are you trying to achieve?
09:18 babilen Could you paste it and its entire output to one of http://refheap.com, http://paste.debian.net, https://gist.github.com, http://sprunge.us, … ?
09:21 AndreasLutro Acidix: sounds like you haven't installed salt properly
09:23 GreatSnoopy joined #salt
09:27 bhosmer joined #salt
09:28 babilen Which raises the question: "How did you install it?", "On which platform are you running salt?", ...
09:28 Acidix My OS is Debian 7.9, but I don't have access to the internet to install packages from repository. I tooke from someone rpm packages. I converted extention with Alien to have a ".deb" package. ( I only have dpkg). Indeed, I haven't installed salt properly
09:29 babilen You are aware that there are packages for Debian, aren't you?
09:30 elsmo joined #salt
09:30 babilen http://debian.saltstack.com/ has packages for wheezy
09:32 babilen This repository is, however, no longer maintained. We'll try to get recent versions of salt* into wheezy-backports-sloppy soon, but that'll probably take some weeks.
09:33 babilen http://repo.saltstack.com/#debian has automatically build packages for wheezy and I'd use them for now.
09:34 thallera1 uh is there a way to tell the saltmaster that it should write to a new logfile (after logrotation)? I tried a sighup but with no effect.
09:35 attawn joined #salt
09:35 mavhq joined #salt
09:37 AndreasLutro good question
09:40 Acidix @AndreasLustro @babilen yes, the person directly give me the rpm package .. well, I'll manage all dependences .. thanks !
09:41 lero joined #salt
09:44 hemebond Is there anything I can do about the check errors https://github.com/saltstack/salt/pull/32042 ?
09:44 saltstackbot [#32042]title: Discard duplicate YAML nodes when merging anchors to avoid duplicate … | ### What does this PR do?...
09:44 hemebond I don't actually know what they were testing or what is wrong.
09:44 hemebond Though a couple of failures appear to be because of the test minion failing to bootstrap.
09:45 AndreasLutro jenkins is hard to read... http://166.78.178.63:8080/job/PR/job/salt-pr-lint-n/386/violations/ this is something you can do something with at least
09:45 hemebond Oh, nevermind. Just had a look through other pull requests that the checks are failing on most.
09:45 AndreasLutro minion failed to bootstrap is out of your hands
09:46 hemebond Aha, thanks.
09:46 hemebond Some of that was copied straight out of the YAML module so I didn't want to play with it.
09:47 babilen Acidix: "the person" ?
09:50 hemebond AndreasLutro: So Salt wants me to use .format instead of %s, would that be correct?
09:50 hemebond That appears to be a couple of the lint errors.
09:50 hemebond "String substitution used instead of string formatting"
09:53 hemebond I'll just change it to match other Salt files.
09:55 s_kunk joined #salt
09:56 Acidix babilen : the person = My boss
09:56 AndreasLutro thallera1: grepping for 'signal.signal' in salt's codebase, it doesn't look like there is anything
09:59 marnom salt-cloud complains about InsecureRequestWarning: Unverified HTTPS request is being made. Adding certificate verification is strongly advised. - anyone else seeing this?
09:59 jaybocc2 joined #salt
10:00 babilen Acidix: You might make your boss aware of the fact that saltstack provides Debian packages
10:01 thallera1 AndreasLutro: are there plans to add something like this?
10:04 AndreasLutro thallera1: found a couple of issues for it: https://github.com/saltstack/salt/issues/13558 https://github.com/saltstack/salt/issues/11716
10:04 saltstackbot [#11716]title: Deb: log rotation issue | I just noted that my Debian installed salt-minion still has its deleted log file open:...
10:06 thallera1 AndreasLutro: ah you added a comment on the issues I will monitor the issues then. thx for your time
10:08 hemebond lol, my lint rating is going _down_ with each edit.
10:09 hemebond Even though I'm resolving issues it lists.
10:13 attawn Hi everybody
10:13 attawn somebody here to help a noob with managing lxc containers ?
10:14 attawn i've read the doc but still cant understand a few things
10:15 jaybocc2 joined #salt
10:16 kliquori joined #salt
10:17 ninjada joined #salt
10:18 MadHatter42 joined #salt
10:27 bhosmer joined #salt
10:27 CeBe1 joined #salt
10:30 ryan__ joined #salt
10:32 edrocks joined #salt
10:42 lionel joined #salt
10:45 fredvd joined #salt
10:52 babilen attawn: Anything in particular that you are curious about?
10:54 ninjada joined #salt
10:55 attawn babilen: thanks for answering. I don't really understand how to use the conf. profile from the doc
10:55 attawn i put it in the master.d directory but don't know how to use it
10:56 RobertChen117 joined #salt
10:56 spankalish joined #salt
10:56 babilen "conf. profile from the doc" -- which documentation are you referring to?
10:57 attawn this one : https://docs.saltstack.com/en/latest/topics/tutorials/lxc.html#profiles
10:59 _Cyclone_ joined #salt
11:02 spankalish Hi can anyone help me with querying the salt mine?
11:02 kliquori joined #salt
11:02 spankalish When I do sudo salt '*' mine.get '*' network.ipaddrs
11:02 amcorreia joined #salt
11:02 spankalish The Minions return with a blank
11:03 spankalish minion1:     ----------
11:05 fir3wall hello, can I require other state to be finished before applying my current one ? something like require: other_state (instead of using include on the top of my state)
11:06 edrocks joined #salt
11:08 N-Mi joined #salt
11:10 hemebond spankalish: network.ip_addrs
11:10 hemebond fir3wall: Yes.
11:11 hemebond fir3wall: https://docs.saltstack.com/en/latest/ref/states/requisites.html
11:11 hemebond fir3wall: It's covered in the tutorial.
11:12 DammitJim joined #salt
11:13 spankalish hemebond: yes
11:13 fir3wall hemebond: thanks, missed it as only look on top of file ! thanks
11:13 spankalish hemebond: is that wrong?
11:14 hemebond spankalish: You missed the underscore "_"
11:15 spankalish hemebond: oh yea I missed it on here, but not in the terminal. It's still returning blanks
11:15 hemebond spankalish: Have you set up the minion to return that info?
11:16 hemebond How are you configuring that mine function?
11:17 spankalish hemebond: ok now we're getting somewhere........ how do you get the minions to return the info? I tried sudo salt '*' mine.update
11:17 hemebond spankalish: You have to first tell the minion to return that information. I'll paste my pillar.
11:17 spankalish hemebond: but I guess it's not that simple?
11:17 spankalish hemebond: cool thanks
11:18 hemebond http://paste.debian.net/418328/
11:18 hemebond I have that in a pillar and apply that pillar to minions.
11:18 hemebond Then I can do, e.g.: salt-run mine.get '*' network.ip_addrs
11:19 hemebond or: salt '*' mine.get '*' network.ip_addrs
11:20 spankalish hemebond: does a sudo salt '*' mine.send network.ip_addrs eth1 not do the same?
11:20 hemebond Yeap. If you want to do that I think you change eth0 in the pillar to []
11:22 yomilk joined #salt
11:22 spankalish hemebond: so a you have to set up the pillar first before you can query the salt mine? Can you put as many inputs as you want into a single pillar?
11:23 hemebond I don't know what you mean by "inputs"
11:23 hemebond You have to tell the minion to send this data via Salt Mine.
11:23 hemebond Minions don't automatically send anything back via the Mine.
11:23 hemebond I'm using Pillars to do it.
11:23 hemebond You can also configure it on the minion itself.
11:24 spankalish hemebond: Ok. So you have it sending the network address for eth0 so can I also send like "group: group1" in the same pillar?
11:25 spankalish hemebond: or do they have to be seperate?
11:25 hemebond spankalish: I think you have the wrong idea about what Salt Mine is.
11:25 spankalish hemebond: Yea maybe I do
11:25 yomilk joined #salt
11:26 hemebond Pillar is any data you want available to your minion. Configuration data if you will.
11:26 hemebond Salt Mine is used to have the minion send back various information, e.g., IP addresses, disk space, etc.
11:27 hemebond mine_functions is a specific value that I am sending to the minion.
11:27 hemebond The minion goes through those functions and runs those functions or fetches that data and sends it back to the master.
11:27 spankalish hemebond: So you use a pillar to configure the minion with what data you want sent to the salt mine?
11:28 hemebond btw, testing now I'm not sure how to get network.ip_addrs to return a specific interface via Mine.
11:28 hemebond Correct. Not just that (mine_functions) but yes, I use Pillars to do it.
11:28 hemebond You can also configure mine_functions directly on the minion I believe.
11:28 spankalish hemebond: ok
11:29 spankalish hemebond: Still got a lot of learning to do :)
11:30 hemebond Okay, putting eth0 after network.ip_addrs does nothing. It just needs to be []
11:30 hemebond That'll return all IP addresses.
11:30 hemebond Salt Mine is more advanced than Pillars and States.
11:32 spankalish hemebond: Thanks, just reading up about the salt mine now, and configuring the minions
11:32 Rahul joined #salt
11:33 Rahul hi
11:33 Rahul how can i create a new VM with salt minion on it
11:33 Guest99646 how can i create a new VM with salt minion on it
11:33 spankalish hemebond: why I didn't do this in the first place, I don't know. I guess I thought the minions populated the salt mines with data automatically
11:33 hemebond Guest99646: Not really a Salt question/function.
11:34 hemebond spankalish: Yeah, Salt Mine is probably not something you need when you're first learning.
11:34 Guest99646 i am provisioning new vm from master and i want to have minion on this vm
11:35 punkoivan joined #salt
11:36 punkoivan joined #salt
11:36 spankalish hemebond: it's just something I want to get done, it's a haproxy load balancer, I'm setting it up for zero downtime deployment for 6 webservers
11:37 punkoivan joined #salt
11:38 punkoivan joined #salt
11:39 spankalish hemebond: I'm enjoying it, but it's heavy going, especially since I am in college and have a lot of other stuff on the go. I don't have much time for this.
11:39 spankalish hemebond: Anyway thanks a million for the help
11:39 hemebond It takes a long time to grok it.
11:39 hemebond You're welcome. Good luck.
11:39 punkoivan joined #salt
11:45 RobertChen117 joined #salt
11:46 hemebond Passing an argument to network.ip_addrs in Mine doesn't seem to do anything.
12:13 hemebond In fact none of the Mine examples seem to work as expected.
12:14 RobertChen117 joined #salt
12:16 keimlink joined #salt
12:20 akhter joined #salt
12:22 Travis joined #salt
12:23 DammitJim is there no salt state for logrotate?
12:24 edrocks joined #salt
12:26 Guest8241 New to Salt. Trying to provision a Win2k12 into an ESX environment. Getting a NetBIOSError (ERRCLASS_OS) from file nmb.py
12:27 jaybocc2 joined #salt
12:28 bhosmer joined #salt
12:29 favadi joined #salt
12:33 rem5 joined #salt
12:34 blaman joined #salt
12:34 metalseargolid joined #salt
12:35 blaman is there any way to tell salt that the desired state of a file is only whatever has happened as a result of running other salt states? i.e I want to discard outside modifications
12:37 blaman basically I have a state that sets sane values in sshd_config, and then some other states run and add their own stuff
12:37 blaman I meant sane defaults
12:37 keimlink joined #salt
12:38 blaman I'm using augeas for this. now if I put rm * at the top of the set-sane-defaults state, so as to remove external changes, that wouldn't work because it would wipe out the changes made by the other states
12:39 blaman so the other states would add their stuff again, and the cycle repeats everytime i do highstate
12:42 jaybocc2 joined #salt
12:46 kliquori joined #salt
12:47 felskrone1 joined #salt
12:48 evle1 joined #salt
12:49 felskrone2 joined #salt
12:53 AlbertoKPN joined #salt
12:53 babilen blaman: I typically write templates for that case and compose pillar data with a defined order (cf. pillarstack and https://docs.saltstack.com/en/latest/ref/configuration/master.html#pillar-source-merging-strategy )
12:54 babilen blaman: That way you could, for example, use https://github.com/saltstack-formulas/openssh-formula with a pillar such as https://github.com/saltstack-formulas/openssh-formula/blob/master/pillar.example
12:54 AlbertoKPN Hello all, I have an issue for what I can not find information in google....
12:55 edrocks_ joined #salt
12:56 AlbertoKPN when running my state the minion is updating the information about packages from all the repositories
12:56 blaman babilen: hmmmm actually I think what would be perfect is something like require_in for augeas.
12:56 blaman babilen: I'll check out the links, thank you
12:56 AlbertoKPN this activity uses almost all the available memory and cpu
12:57 babilen blaman: I haven't worked much with augeas, but my feeling is that you crossed the line in which it makes sense to use it when you run into the problems you describe
12:59 AlbertoKPN by "running salt-call state.highstate -l debug" the minion stays in the message   [INFO    ] Executing command "repoquery --plugins --queryformat '%{NAME}_|-%{VERSION}_|-%{RELEASE}_|-%{ARCH}_|-%{REPOID}'   --all --quiet --whatprovides " in directory '/root'  for several minutes,   is there a way to avoid this other than commenting out the pkg.installed declaration?
12:59 babilen Is that really such a costly action?
13:00 AlbertoKPN the minion has 1GB ram
13:01 subsignal joined #salt
13:02 edrocks joined #salt
13:02 AlbertoKPN Started: 12:40:27.846628     Duration: 1171255.908 ms
13:02 quix joined #salt
13:04 AlbertoKPN first run of highstate seems ok, but every time after that is slow....
13:07 numkem joined #salt
13:10 babilen AlbertoKPN: Could you try passing "- refresh: False" to the pkg.installed states?
13:10 blaman babilen: basically what I want is to be able to have states adding to a config file and then if there is an external change to the parts they care about, it'll get reverted when running highstate
13:10 babilen blaman: Yes, exactly
13:11 babilen That's why you let salt manage the entire file rather than tell it that there are specific changes
13:11 blaman a pillar would help me with this as you said earlier?
13:11 blaman yeah but the changes will come from states within salt
13:12 blaman basically I wanted to put the changes that state X needs to make to sshd_config into the state X sls file
13:12 babilen Well, in a way. The basic idea is that salt is able to render the file completely. That way salt can check if the file it finds on the minion is identical to the one it expects. If it is not it would make the necessary changes and display a diff.
13:12 AlbertoKPN thank you babilen I am recreating the minion, will test your option it in few minutes
13:14 babilen blaman: Pillars allow you to generate that file in a data driven way based on a template (or other generator). You would therefore not include "changes" in your states, but rather in the data that's used to generate the file (i.e. pillars)
13:14 blaman babilen: oh that sounds like what I want I think. thank
13:14 blaman s
13:14 babilen blaman: It is a different approach in that the responsibility for rendering that file lies solely with a single state rather than a multitude of them.
13:15 blaman babilen: oh then it isn't.
13:15 babilen It also makes it more obvious when you have conflicting settings or one of them overrides another as there will be, in the end, a single pillar after merging.
13:15 slav0nic joined #salt
13:15 babilen It's just that you compose the data that is being used, rather than states that say "oh, edit line 4 also"
13:16 blaman oh yes yes. so states can add to the pillar.
13:16 blaman ok then that could work
13:16 blaman I'll go read the docs
13:18 s_kunk joined #salt
13:18 s_kunk joined #salt
13:19 s_kunk_ joined #salt
13:20 hasues joined #salt
13:20 hasues left #salt
13:20 s_kunk joined #salt
13:22 bhosmer joined #salt
13:24 uictamale morning all
13:25 uictamale I successfully created a gce instance with salt cloud, and I can see it when I run salt-cloud --query, but when I try to do salt <instance> test.ping it says No mions matched the target.  Any ideas?
13:25 AlbertoKPN babilen: The option "- refresh: False" should do the trick according to documentation however it is not working
13:26 AlbertoKPN I tried with "postgresql95-server.x86_64:   pkg.installed:     - refresh: False" and still same effect
13:27 bluenemo joined #salt
13:31 nZac joined #salt
13:31 darvon joined #salt
13:31 linjan__ joined #salt
13:32 AlbertoKPN minion version is "salt-minion.noarch               2015.8.1-1.el6"
13:33 AndreasLutro AlbertoKPN: that documentation was for service states, not pkg states
13:33 AndreasLutro what exactly are you trying to refresh?
13:34 AlbertoKPN I am not trying to refresh
13:34 AndreasLutro oh wait there is a refresh arg to pkg.installed, I was sure that was refresh_db
13:34 AlbertoKPN by "running salt-call state.highstate -l debug" the minion stays in the message   [INFO    ] Executing command "repoquery --plugins --queryformat '%{NAME}_|-%{VERSION}_|-%{RELEASE}_|-%{ARCH}_|-%{REPOID}'   --all --quiet --whatprovides " in directory '/root'  for several minutes,   is there a way to avoid this other than commenting out the pkg.installed declaration?
13:35 AndreasLutro I doubt it
13:36 AlbertoKPN :(
13:36 AndreasLutro I'm not familiar with yum though
13:36 AndreasLutro your salt version seems a bit old, you could try upgrading, there could be performance fixes
13:37 AlbertoKPN well at the moment we just comment out the pkg.installed declaration for every highstate in our vagrant development boxes
13:37 babilen AndreasLutro: refresh_db is the module function :)
13:37 AlbertoKPN but it is a pain that I would like to solve
13:38 AlbertoKPN upgrading/update version is something that is on our roadmap but not yet
13:38 AndreasLutro do you have the entire root drive as nfs/vbsf or something? I don't see why it should be that slow
13:39 AlbertoKPN no, but we have our own copy of the repositories, I can assume that the network from my localhost to our repositories is veeeeeery slow
13:40 cpowell joined #salt
13:40 AlbertoKPN but it is not slow when installing packages
13:40 AndreasLutro mm
13:41 iceyao joined #salt
13:41 AlbertoKPN anyway it is not a critial issue
13:41 AndreasLutro interestingly enough, in 2016.3 yumpkg has been rewritten to not use repoquery at all
13:42 AlbertoKPN yes that is interesting
13:43 kawa2014 joined #salt
13:43 AlbertoKPN I think next week we will add to the sprint the upgrade of salt to latest newer/est version
13:43 AndreasLutro the call to repoquery is wrapped in an "if not skip_suggestions:"
13:43 AndreasLutro so if you set skip_suggestions to true....
13:44 AndreasLutro might break other things though
13:44 _JZ_ joined #salt
13:45 ronnix joined #salt
13:47 jschoolcraft joined #salt
13:48 ronnix joined #salt
13:49 andrew_v joined #salt
13:49 timoguin joined #salt
13:49 mapu joined #salt
13:53 kaptk2 joined #salt
13:55 uictamale I successfully created a gce instance with salt cloud, and I can see it when I run salt-cloud --query, but when I try to do salt <instance> test.ping it says No mions matched the target.  I checked that the minion process is running.  Any ideas?
13:55 AndreasLutro uictamale: salt-key -L on the master? check both master and minion logs
13:56 rburkholder joined #salt
13:57 uictamale I see this instance under accepted keys
13:57 kshlm joined #salt
13:57 AndreasLutro time to check the logs then
13:58 uictamale Master log is pretty boring - just says "Created Cloud VM '<instance>'
13:58 racooper joined #salt
13:59 Ashald hi guys, is there is a way to register 'one off' reactions in Salt?
13:59 uictamale Hmm minion log says "Minion failed to start" over and over
13:59 Ashald That's so I don't have to include reaction configuration into master config file
13:59 AndreasLutro uictamale: if you're on a systemd system, try journalctl -u salt-minion
13:59 quix_ joined #salt
13:59 AndreasLutro if not, try starting the minion manually
13:59 zmalone joined #salt
14:00 uictamale I am on systemd :)
14:00 uictamale Yeah that just spits out the same stuff.. Minion failed to start, Restarting minion
14:01 AndreasLutro odd... well, `service salt-minion stop; salt-minion -l debug`
14:01 tvinson joined #salt
14:03 jerredbell joined #salt
14:04 uictamale [DEBUG   ] SaltReqTimeoutError, retrying. (1/7)
14:04 uictamale Looks like my minions can't reach my master.
14:04 uictamale What port are they trying to connect over?
14:05 AndreasLutro 4505 and 4506
14:05 babilen https://docs.saltstack.com/en/latest/topics/tutorials/firewall.html
14:06 uictamale ok thanks, looks like I have some network debugging to do.
14:06 uictamale The master and minions are in separate google projects so this is not going to be fun, ha.
14:07 uictamale I think ssh is the only thing we have open between them now.
14:07 uictamale It's not unusual for masters and minions to be in very different networks, right?
14:08 bowhunter joined #salt
14:08 cpowell joined #salt
14:09 honestly uictamale: make a VPN then
14:09 honestly (or an ssh tunnel if you feel especially ghetto)
14:09 mavhq what's the best way to specify an http proxy for my minions to use to download internet files?
14:10 mavhq on debian
14:10 Shirkdog_ joined #salt
14:11 Ludo- joined #salt
14:11 ze- joined #salt
14:13 Ashald mavhq, use http_proxy or https_proxy environment variables
14:13 jfelchner joined #salt
14:13 Ashald depending on your proxy type
14:13 quix joined #salt
14:13 LotR joined #salt
14:14 diegows joined #salt
14:14 Twiglet joined #salt
14:15 jfred joined #salt
14:16 patarr joined #salt
14:16 JPT joined #salt
14:17 davidbanham joined #salt
14:17 beardo_ joined #salt
14:18 blu_ joined #salt
14:18 mavhq I've set those in /etc/environment, should it be picking that up?
14:18 Qwazerty joined #salt
14:18 lynxman_ joined #salt
14:18 linjan_ joined #salt
14:18 bluenemo joined #salt
14:18 al joined #salt
14:20 bdrung_work joined #salt
14:21 adelcast joined #salt
14:21 impi joined #salt
14:21 ronp_usa joined #salt
14:22 dmaiocchi joined #salt
14:22 ronp_usa which GUIs are avaialble for saltstack? Is there a tool to help plan dependencies in state files?
14:24 honestly mavhq: set them in the minion service startup file
14:26 zmalone ronp_usa: saltpad, halite, and the commercial web gui
14:26 zmalone halite is deprecated
14:26 zmalone I don't know of any tool for planning dependencies
14:27 jerredbell joined #salt
14:34 cpowell joined #salt
14:34 fir3wall Is that correct way to set variable to default value : {% set instance_ssl = instance.manage_ssl|default(true) -%}
14:34 ronp_usa thanks z
14:35 fir3wall I'm gettting instance_ssl is undefined ... I want it to be true when somebody will not include that in his pillar
14:35 rhodgin joined #salt
14:35 cpowell_ joined #salt
14:43 keltim joined #salt
14:52 Brew joined #salt
14:54 subsignal joined #salt
14:55 hightekvagabond joined #salt
14:55 hightekvagabond joined #salt
14:56 nZac_ joined #salt
15:02 mavhq for some reason I have a /etc/default/salt-minon.environment file, no idea why. Made a /etc/default/salt-minon file, added export http_proxy="http://proxy:80/" to it, restarted salt-minion, ran salt '*' cmd.run "env", proxy wasn't listed
15:03 subsigna_ joined #salt
15:07 mavhq joined #salt
15:10 rem5 joined #salt
15:14 kshlm joined #salt
15:15 ageorgop joined #salt
15:20 gh34 joined #salt
15:22 teryx510 joined #salt
15:22 mohae joined #salt
15:25 oida joined #salt
15:27 beardedeagle joined #salt
15:28 rem5 joined #salt
15:28 bhosmer_ joined #salt
15:28 berserk joined #salt
15:30 ronnix joined #salt
15:32 josuebrunel joined #salt
15:37 rm_jorge joined #salt
15:37 berserk joined #salt
15:38 zmalone joined #salt
15:40 ageorgop joined #salt
15:41 sfxandy joined #salt
15:42 sfxandy quick question re orchestration...
15:42 sfxandy if I want to target a specific minion when i call the orchestration via state.orchestrate, whats the best way to do that?
15:42 mavhq set http_proxy via environ.setenv, which seems to work, still getting Unable to manage file: Error: HTTP 599: Timeout though. wget works for same file manually
15:43 nZac joined #salt
15:43 colegatron hi
15:44 AndreasLutro sfxandy: pass it in as a pillar value
15:44 saltpxf joined #salt
15:45 sfxandy AndreasLustro, that did occur to me ... just wondered if there was another way
15:45 elsmo Hey all. Trying to setup salt to use a gpg-agent but unsure how to proceed - doesn't seem to be a whole lot of information out there on how to tell salt where the gpg-agent is running
15:45 mavhq salt '*' cmd.run "wget  http://ftp.uk.debian.org/debian/dists/jessie/main/installer-amd64/current/images/netboot/netboot.tar.gz" works too
15:45 kevinqui1nyo a state module i wrote contains a function that can not be lazy loaded.  I assume there's a syntax error but i can't see it.  How can i troubleshoot this to see where the problem is?  State 'wordpress.user_present' was not found in SLS 'web.wordpress'
15:46 saltpxf left #salt
15:52 kevinqui1nyo nm on my question -- the reason debug logging on the minion is not showing anything is because i had the function indented incorrectly so it appeared to be part of the function above
15:53 fs5237n32 joined #salt
15:53 cyborg-one joined #salt
15:54 ageorgop joined #salt
15:54 kshlm joined #salt
15:59 heaje joined #salt
16:03 cyborglone joined #salt
16:05 onlyanegg joined #salt
16:06 amcorreia joined #salt
16:07 quix_ joined #salt
16:08 felskrone joined #salt
16:09 nZac joined #salt
16:12 mrwboilers joined #salt
16:12 bgulla joined #salt
16:13 mrwboilers Has anyone else ran into issues using os.setreuid or os.seteuid in a custom module?
16:13 mrwboilers When I do that, the module just seems to hang
16:14 bgulla i need to compile from source. whenever I try to build the spec file in /pkg/rpm/salt.spec, i got the following error: error: File /root/rpmbuild/SOURCES/salt-2014.7.2.tar.gz: No such file or directory
16:14 bgulla that seems like an old build, can anyone help me get past this?
16:16 nZac joined #salt
16:17 alemeno22 joined #salt
16:18 mrwboilers When I run the same python locally (not via salt) it works fine.
16:19 dijit bgulla: that's a really old version.
16:19 dijit bgulla: why are you running 'sudo ' when building the rpm?
16:19 dijit rpmbuild can run without root.
16:19 dijit (and should be)
16:20 kochary joined #salt
16:21 kochary Hey guys, I installed salt using the bootstrap script, but now I want to completly remove it, how can I do?
16:21 mavhq joined #salt
16:26 TooLmaN joined #salt
16:31 bhosmer_ joined #salt
16:32 flowstate joined #salt
16:33 flowstate Does anyone have a good resource for practical examples of setting up a saltstack infrastructure in AWS?
16:33 flowstate I've been referring to Ryan D Lane's masterless stuff, but I want to look at what a masterful system would look like as a comparison point
16:33 quix joined #salt
16:34 AndreasLutro flowstate: anything specific you're wondering about? mostly it's just getting the amazon instance to boot up and connect to the salt master
16:35 dhscholb for the returner system, what exactly is the difference between the returner function and the save_load function?
16:35 flowstate well, I guess I'm kinda starting from zero, as I jumped on Lane's stuff and started setting up all my stuff to be masterless
16:35 dhscholb i'm using the mysql returner for master_job_cache, and it seems to be using both functions
16:36 dhscholb but i can't tell why it would use one function over the other. they seem redundant
16:36 flowstate so basically, it's a standard salt setup, with the only real difficulty being setting up communication between master and minion?
16:36 ronnix joined #salt
16:37 AndreasLutro flowstate: well, there's not much difference between AWS and other cloud providers. use whatever provisioner you use (salt-cloud, terraform, cloud-init, packer...) to set up a salt minion that connects to the correct master and you're done
16:37 flowstate welp, I guess that's pretty simple, haha
16:37 flowstate it would appear that doing it masterless comes with more complexity. I'm really curious as to the reasons you might choose to go masterless
16:38 flowstate I understand key and secret management could be a bit difficult, but apart from that...
16:38 AndreasLutro having a self-contained system, not having to manage salt keys
16:38 flowstate yeah, that's fair enough
16:38 AndreasLutro not having to set up NATs/VPNs between VPCs
16:38 flowstate is there a lot of complexity to managing the salt keys?
16:38 AndreasLutro if you want full automation there may be
16:39 flowstate oh yeah, that's right, you'd have to somehow automatically accept the keys for the new minions
16:39 AndreasLutro preferably pre-generate the keys on the master and transfer them to the minion in a secure way
16:39 flowstate ahhh, interesting
16:40 AndreasLutro also distributing the accepted key between masters
16:40 flowstate oh wow, yeah, generating keys for minions in an autoscaling group could be really interesting
16:40 flowstate okay, that makes sense
16:40 flowstate thanks for the help!
16:40 grumm_servire joined #salt
16:40 flowstate oh, btw, your blog has been super helpful to me, thanks for that as well
16:41 AndreasLutro nice :)
16:42 angvp Hey guys, I am checking my salt-master is consuming lot of RAM (4.5GB~), it has a basic configuration no big deal, so may be is something on my env that might causing this, any ideas how to collect info that help us determine why it's happening this?
16:49 mavhq joined #salt
16:53 beardedeagle mmm that errbot integration with salt...
16:54 sagerdearia joined #salt
16:55 writtenoff joined #salt
16:55 ageorgop joined #salt
16:55 punkoivan joined #salt
16:56 zmalone joined #salt
16:56 punkoivan joined #salt
16:58 hightekvagabond joined #salt
16:58 punkoivan joined #salt
16:59 digitalhero joined #salt
17:00 iggy https://github.com/saltstack-formulas/ntp-formula/pull/26 <-- somebody who can, should
17:00 saltstackbot [#26]title: Revert "fix service name for SLE11" | Reverts saltstack-formulas/ntp-formula#23
17:00 punkoivan joined #salt
17:02 punkoivan joined #salt
17:03 AndreasLutro iggy: I'm in no way invested or involved with this but if I were responsible for that repo I'd ask you for an explanation for the pull request
17:05 Trauma joined #salt
17:06 punkoivan joined #salt
17:07 zmalone joined #salt
17:07 peter-xyz joined #salt
17:07 moy joined #salt
17:07 punkoivan joined #salt
17:09 peter-xyz does anybody know how I can use the iptables save features without storing the fail2ban rules?
17:10 digitalhero joined #salt
17:11 akhter joined #salt
17:11 Kelsar joined #salt
17:11 peter-xyz I had this hack working where I would remove the fail2ban rules afterwards with file.lines, but now I can't get the ordering right
17:13 peter-xyz never mind, order: last is working for me
17:14 _mel_ joined #salt
17:14 djgerm joined #salt
17:14 djgerm What's an easy way to define the relationship between a state and a module?
17:15 punkoivan joined #salt
17:15 kevinqui1nyo djgerm: what do you mean by 'define the relationship'?  Are you asking what the relationship is in english?
17:15 kevinqui1nyo or do you mean something more technical
17:16 punkoivan joined #salt
17:16 djgerm in english
17:16 bhosmer joined #salt
17:16 djgerm yeah not like a run down of any specific state to a module
17:17 punkoivan joined #salt
17:18 AndreasLutro custom states/modules or built-in ones?
17:18 hightekvagabond joined #salt
17:18 djgerm built in
17:19 AndreasLutro they're not really related, it just happens that states with name abc often tend to use module abc as well
17:19 punkoivan joined #salt
17:20 punkoivan joined #salt
17:21 kevinqui1nyo if there's a module like "unicorn", it often does all the dirty work, and the corresponding state module, "unicorn" merly contains idempotent functions, like unicorn_present() for instance
17:21 kevinqui1nyo but there's no requirement for them to have the same name
17:21 kevinqui1nyo it just seems like a common practice if you need the idempotent state module, and then something else to go and do it's work
17:22 punkoivan joined #salt
17:23 kevinqui1nyo like unicorn_present() behind the scenes might call some module functions like __salt__['unicorn.get'](id=12), and if it doesn't exist, __salt__['unicorn.create']() djgerm if that helps at all
17:23 punkoivan joined #salt
17:26 digitalhero joined #salt
17:26 impi joined #salt
17:27 jfelchner All, am I doing something wrong?  Because the gitfs stuff takes forever to download.  It's almost like it's making a new connection to the git repo for every single file instead of just downloading the entire repo first and then getting all the files out of it afterwards.
17:27 jfelchner It takes almost 30 minutes to get all the files.
17:28 punkoivan left #salt
17:28 jfelchner It renders that feature basically unusable for me. :(
17:28 AndreasLutro sounds odd
17:29 baweaver joined #salt
17:29 Fiber^ joined #salt
17:29 AndreasLutro I used gitfs with pygit2 over ssh and never had any problems like that
17:31 kevinqui1nyo jfelchner: if you go to your salt file_roots path and do:  strace -etrace=network -f git remote show origin >/dev/null
17:31 kevinqui1nyo do you see anything weird like attempt to connect over ipv6 instead of ipv4 and you dont have an ipv6 address there?
17:31 jfelchner AndreasLutro I'm using HTTPS connections if it makes a difference
17:31 jfelchner kevinqui1nyo lemme check
17:32 zmalone1 joined #salt
17:33 kevinqui1nyo jfelchner: really i was just wondering if running git commands to your remote are slow across the board and maybe network related
17:34 kevinqui1nyo or if it was just an issue with gitfs
17:34 armyriad joined #salt
17:35 bhosmer_ joined #salt
17:36 shawnbutts i see 2015.8.8 is available on repo.saltstack.com.  but, i can't find the release notes.  is this in the middle of being released or something?
17:36 baweaver joined #salt
17:36 jfelchner kevinqui1nyo well I guess my question is, does it make a brand new connection for *every* file, or does it download the repo and then first check to see if the file is in the repo it has already downloaded.
17:37 jfelchner kevinqui1nyo AndreasLutro here is what I'm seeing: https://www.dropbox.com/s/6djjfiqtvwlezfg/slow-gitfs.mp4?dl=0
17:38 zmalone shawnbutts: I think it's mid release
17:38 zmalone earlier, there were a bunch of empty directories on repo.saltstack.com
17:39 cpowell joined #salt
17:40 jfindlay yes, new releases on 2015.5 and 2015.8
17:41 jfelchner kevinqui1nyo I haven't specified any file_roots so I'd just run that in /srv/salt?
17:42 shawnbutts thanks
17:43 shawnbutts looks like https://github.com/saltstack/salt/pull/32023 was missed.  i'm getting the failed import error
17:43 saltstackbot [#32023]title: Move constant declaration into member variable to avoid issues when m… | ### What does this PR do?...
17:43 jfelchner I'm using pygit2 as well.
17:45 baweaver joined #salt
17:46 jfelchner In masterless mode
17:56 digitalhero joined #salt
17:58 BretFisher joined #salt
18:00 digitalhero joined #salt
18:08 sfz- joined #salt
18:10 digitalhero joined #salt
18:10 gcorey joined #salt
18:10 digitalhero joined #salt
18:11 baweaver joined #salt
18:12 nidr0x joined #salt
18:13 sjorge joined #salt
18:13 sjorge joined #salt
18:16 bhosmer__ joined #salt
18:19 jfelchner I had about 30 git remotes in my remote list.  I'm attempting to consolidate them down to one to see if it changes things.
18:20 nidr0x joined #salt
18:20 colegatron_ joined #salt
18:23 punkoivan joined #salt
18:23 punkoivan joined #salt
18:28 jfelchner DRASTIC difference.
18:29 jfindlay :)
18:30 jfelchner Basically went back to normal.  With a large number of remotes (I had 27), loading the files took about 30-40 minutes.  When I consolidated them into one repo, it took about 1.
18:31 jfelchner So there must be some sort of issue with the way that the gitfs lookup is working that is causing it to be slow with a lot of remotes.
18:32 jfelchner Is this a thing that needs an issue filed for it?
18:33 hightekvagabond joined #salt
18:33 jfelchner Again, if you missed it, see my screencast here for just how slow it actually is: https://www.dropbox.com/s/6djjfiqtvwlezfg/slow-gitfs.mp4?dl=1
18:34 zmalone https://github.com/saltstack/salt/issues/32004 :/
18:34 saltstackbot [#32004]title: 2016.3 now complains about win_dacl on Solarish platforms | ### Description of Issue/Question...
18:34 Netwizard joined #salt
18:35 jfindlay jfelchner: I'm not sure.  I can see the case that performance might be better, but it is also likely that this is caused by the git backend you're using
18:35 zmalone looks like that made it into the current 2015.8.8 release that just was pushed
18:35 jfelchner @jfindlay you mean that I'm using pygit2?
18:36 flowstate joined #salt
18:37 jfelchner shawnbutts I literally just got that message https://gist.github.com/shawnbutts/00305a3c83193f8ae465
18:37 jfelchner How are you working around it?
18:37 jfindlay jfelchner: yeah, there are one or two other python git libs also, but I'm not very familiar with all that
18:37 GreatSnoopy joined #salt
18:37 jfelchner jfindlay ok.  I have to use pygit2 due to it supporting HTTPS auth, but I'll just leave it alone for now then.
18:38 nidr0x joined #salt
18:39 antpa joined #salt
18:39 punkoivan hi.
18:39 lroe joined #salt
18:39 lroe how do I cmd.ping all hosts whose osfullname grain matches Debian?
18:39 akhter joined #salt
18:40 punkoivan I try to use cmd.run to create keyspace in cassandra with command cqlsh -f file.cql
18:40 punkoivan but get error connection refused. So, When i try to do it manual - all ok.
18:40 punkoivan Why with salt I got this error, any reason?
18:41 jfindlay jfelchner: also, 27 git remotes may be scaling gitfs and pygit2 beyond its space of common use cases
18:41 jfindlay s/its/their/
18:41 forrest joined #salt
18:42 jfindlay lroe: `salt -G 'os:Debian' test.ping`
18:42 jfelchner jfindlay ok cool.  I was just trying to keep things modular by putting everything relating to one formula in a repo: nginx, logwatch, tripwire, etc.  Having everything in one repo isn't the end of the world, I just wanted to make sure to bring it up and let you all know what my solution was. :)
18:42 jfindlay lroe: `salt -G 'osfullname:Debian' test.ping`
18:43 Edgan Can someone point me at some advanced salt cloud examples? I have finding the documentation very fragmented, and a mix of old and new syntaxes.
18:43 lroe jfindlay, thanks
18:44 jfindlay Edgan: that documentation could generally be updated.  What cloud are you trying to work with?
18:44 Edgan jfindlay: multiple EBS volumes
18:44 Edgan jfindlay: It seems to be block_device_mappings, but when I give it a second one, it says it is invalid.
18:45 jfindlay Edgan: can you paste your config somewhere with the private info removed?
18:45 Edgan jfindlay: yes
18:46 jfindlay not guaranteeing that I'm going to know how to fix it, but I'll try :-)
18:47 zmalone jfindlay: it sounds like that win_dacl issue breaks on all state runs, I've got a coworker running into it now
18:47 zmalone I'm about to try replicating it, but you guys may need to pull 2015.8.9 out quickly
18:48 jfindlay thanks, zmalone
18:48 zmalone (or not, I don't have a support contract, unfortunately)
18:48 Edgan jfindlay: http://fpaste.org/344322/75889714/
18:49 Edgan jfindlay: 'Error': {'Message': 'Invalid device name /dev/sdb1'
18:50 ajw0100 joined #salt
18:51 jfindlay Edgan: the best I can say is that that is looking like a bug.  Do you see anything revealing in the debug log of salt-cloud?
18:52 akhter joined #salt
18:52 kliquori_ joined #salt
18:53 quasiben1 joined #salt
18:53 zmalone great, it throws errors, but doesn't exit, and runs states fine
18:53 zmalone it just spams the logs, which is livable
18:54 zmalone it doesn't prevent you from them using salt to upgrade to the next version, it's just really chatty
18:54 zmalone (the win_dacl loading bug, that is)
18:54 punkoivan1 joined #salt
18:55 punkoivan joined #salt
18:55 quix_ joined #salt
18:55 saltstackbot joined #salt
18:55 evilrob joined #salt
18:55 vaelen joined #salt
18:55 punkoivan joined #salt
18:57 Edgan joined #salt
18:57 punkoivan joined #salt
18:57 nahkiss joined #salt
18:57 tyler-baker joined #salt
18:58 cliluw joined #salt
18:58 punkoivan joined #salt
18:58 johtso joined #salt
19:01 arif-ali joined #salt
19:01 jfelchner_ joined #salt
19:01 punkoivan joined #salt
19:01 feliks joined #salt
19:02 Bucciarati joined #salt
19:02 quix joined #salt
19:04 quix__ joined #salt
19:06 tkharju joined #salt
19:06 Fiber^ joined #salt
19:07 Edgan jfindlay: how do I disable bootstraping in salt cloud?
19:08 BretFisher joined #salt
19:08 ALLmight_ joined #salt
19:08 jfindlay Edgan: `salt-cloud --no-deploy`
19:10 teryx510 joined #salt
19:10 bhosmer joined #salt
19:12 Edgan_ joined #salt
19:12 twodayslate_ joined #salt
19:12 zmalone looks like the develop branch docs have broken, https://docs.saltstack.com/en/develop/index.html
19:12 zmalone left #salt
19:12 dmaiocchi joined #salt
19:12 Phtes_ joined #salt
19:13 serverascode_ joined #salt
19:13 esharpmajor_ joined #salt
19:13 mavhq joined #salt
19:13 gazarsgo_ joined #salt
19:14 bbhoss_ joined #salt
19:14 Edgan_ jfindlay: not sure if you saw it, since I keep getting kicked/split
19:14 tawm04_ joined #salt
19:14 Laogeodritt| joined #salt
19:14 hillna_ joined #salt
19:14 Edgan_ jfindlay: that --no-deploy helps, but may have run into an actual bug
19:14 trent__ joined #salt
19:14 Edgan_ jfindlay: I tried volumes:\n- { size: 1024, device: /dev/sdb1, type: gp2 }   and then get Errors': {'Error': {'Message': "The zone 'us-west-2a' does not exist."
19:15 Edgan_ jfindlay: Which is odd, because I don't exactly specify 2a anywhere, because with VPCs you have to use subnets
19:15 jfindlay Edgan_: it's fine, I really can't feel shame about salt bugs anymore :-)
19:15 Edgan_ jfindlay: haha
19:16 jfindlay Edgan_: that is strange.  EC2 should default to US East
19:16 hemebond1 joined #salt
19:17 c4rc4s_ joined #salt
19:17 georgemarshall_ joined #salt
19:17 basepi_ joined #salt
19:18 smkelly_ joined #salt
19:18 josue joined #salt
19:19 relidy_ joined #salt
19:19 quix joined #salt
19:19 metalseargolid joined #salt
19:19 armyriad joined #salt
19:21 erjohnso joined #salt
19:23 mavhq joined #salt
19:24 arapaho joined #salt
19:26 baweaver joined #salt
19:28 teatime joined #salt
19:31 cyborg-one joined #salt
19:31 nZac joined #salt
19:36 jfindlay develop docs are back online
19:37 baweaver joined #salt
19:39 mapu_ joined #salt
19:45 drawsmcgraw joined #salt
19:46 Edgan_ jfindlay: I am telling it us-west-2, but I never refer to it as us-west-2a. I have to refer to it as the subnet which implies 2a
19:47 berserk joined #salt
19:48 Edgan_ jfindlay: ah, and that points me to the cause. it is actually trying to hit us-east-1 for the volume creation even though it creates the instance in us-west-2
19:50 zer0def joined #salt
20:00 jfelchner_ shawnbutts calling the bootstrap script with `-P git v2015.8.7` solved the issue for me.
20:03 inire joined #salt
20:03 felskrone joined #salt
20:04 CeBe joined #salt
20:05 cliluw joined #salt
20:14 djgerm hello! I am having trouble the DDNS module. I can't see how to specify the RR type in add_host. I want to update the PTR of a zone and I am getting "SyntaxError" https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.ddns.html
20:15 rem5 joined #salt
20:15 djgerm it's strange, because the doc says "Add, replace, or update the A and PTR (reverse) records for a host." almost indicating that both forward and reverse zones would be updated…. but that's just not sense making (how would ddns client know what reverse zone to update?)
20:15 Fiber^ joined #salt
20:17 teryx5101 joined #salt
20:17 ktosiek joined #salt
20:18 slav0nic joined #salt
20:18 gcorey joined #salt
20:21 berserk joined #salt
20:21 Edgan_ jfindlay: The workaround seems to be to specify the location in the provider, not just in the profile.
20:22 Edgan_ jfindlay: Now I am hitting another bug where it claims terminated instances already exist with 2015.8.8. It wasn't an issue in 2015.8.7 from what I saw before. I am going to try downgrading.
20:23 jfelchner joined #salt
20:23 sagerdearia joined #salt
20:25 rem5 joined #salt
20:25 jaybocc2 joined #salt
20:25 keimlink joined #salt
20:26 quix joined #salt
20:27 teatime joined #salt
20:27 flowstate is anyone here experienced with having a master pull state files from git before running them?
20:27 metalseargolid joined #salt
20:27 gcorey left #salt
20:28 elsmo joined #salt
20:28 flowstate I'm trying to figure out the process and any gotchas, but the only search results I find are about the git states and module
20:29 MTecknology flowstate: the only gotcha I'm aware of is... not everything can exist in git (reactors and such) and there's a periodic poll, it's not instant, but salt-run fileserver.update lets you trigger a manual sync
20:30 flowstate okay, the manual sync isn't an issue, as I'm using Jenkins for deployment and provisioning events
20:31 N-Mi joined #salt
20:31 N-Mi joined #salt
20:31 flowstate but I'm confused as to why reactors couldn't live in git
20:32 MTecknology maybe they can now, but they couldn't when I moved everything into git, ext_pillars in git were still new
20:32 flowstate ah, gotcha
20:33 baweaver joined #salt
20:34 digitalhero joined #salt
20:35 flowstate ty for the help!
20:35 MTecknology flowstate: I'm sure they'd be much more than happy for you to build that feature!!! :D
20:35 MTecknology (as would I..)
20:38 jfindlay Edgan_: you are welcome to submit a bug (or two) for that
20:39 jfindlay flowstate: gitfs should cache files local to the master
20:39 Netwizard joined #salt
20:39 jfindlay I'm not sure if that helps your use case
20:41 Edgan_ jfindlay: I found the already exists is actually defined as a feature. Though I am tempted to file a bug saying I consider it a bug.
20:41 jfindlay hm
20:42 AndreasLutro flowstate MTecknology: I've used reactors in git, works fine
20:42 AndreasLutro the reactor *config* still has to be managed separately, though (as far as I know)
20:43 quix joined #salt
20:45 doompatrol joined #salt
20:45 MTecknology AndreasLutro: what do you mean? .. the sls files no longer need to be on the file system?
20:46 AndreasLutro MTecknology: yeah, I just refer to them via salt://whatever/file.sls
20:46 MTecknology AndreasLutro: When I started using reactors, git wasn't an option so I wrote salt states for the master that dropped the reactor files onto the file system
20:46 doompatrol howdy, y'all - having a weird issue where from the salt master I can't do a pillar.get 'site:function' and get the minions pillar data. however doing a pillar.items shows the data and doing a salt-call pillar.get 'site:function' works from the minion
20:46 rem5 joined #salt
20:46 AndreasLutro doompatrol: the two use different caches I think... try a saltutil.refresh_pillar
20:47 doompatrol AndreasLutro: yeah i've done that and even cleared the salt git cache as well to ensure it wasn't using old data
20:48 AndreasLutro doompatrol: hmm, check the master log for errors?
20:49 doompatrol hrm, nothing, let me try w/ -v flag (idk why i didn't think of that till now ...)
20:49 AndreasLutro -v?
20:49 doompatrol verbose
20:49 doompatrol well that wasn't very helpful heh
20:49 AndreasLutro on what? salt? that won't work
20:50 AndreasLutro you can try salt-call pillar.get whatever -l debug
20:50 AndreasLutro on the minion
20:50 AndreasLutro or salt 'minion' pillar.get whatever -l debug
20:50 AndreasLutro on the master
20:50 doompatrol ah yeah just saw that in man page
20:50 doompatrol ty
20:52 ronnix joined #salt
20:55 felskrone joined #salt
20:55 berserk joined #salt
20:58 bhosmer_ joined #salt
20:59 MTecknology AndreasLutro: I'm not finding any documentation for it. :(
21:00 doompatrol ill have to debug this more later. in the meantime, it appears to at least apply states to the minions but it is annoying not being able to call minions via pillar.
21:01 AndreasLutro MTecknology: https://docs.saltstack.com/en/latest/topics/reactor/#mapping-events-to-reactor-sls-files ?
21:02 MTecknology AndreasLutro: ah, nice! Thanks! :D
21:02 MTecknology I wonder how long I haven't known about that sexyness
21:05 fredvd joined #salt
21:08 flowstate joined #salt
21:08 ronnix joined #salt
21:10 subsignal joined #salt
21:12 sfxandy joined #salt
21:16 ninjada joined #salt
21:18 XenophonF left #salt
21:18 Netwizard joined #salt
21:22 aw110f joined #salt
21:23 dmaiocchi joined #salt
21:23 sagerdearia joined #salt
21:24 knite joined #salt
21:25 knite Having some trouble with pkg.installed. I've bumped the version number, but the package doesn't upgrade on the minion. adding refresh: true didn't help. I've verified that the minion's apt cache has the correct (newer) version number, it does.
21:25 knite I've tried changing the salt state to an invalid version number and that still passes, so it looks like the state is being skipped because the package was already installed?
21:25 AndreasLutro I don't think pkg.installed ever changes an already installed package
21:26 capricorn_1 joined #salt
21:27 knite What's the right way to do it? The whole point is to be declarative - I want package X at version Y, I don't want to special case based on whether package is already installed.
21:27 onlyanegg joined #salt
21:28 AndreasLutro let's see...
21:28 AndreasLutro pkg.installed has an argument allow_updates=False
21:28 AndreasLutro try that maybe
21:29 knite I'm not using sources, FYI - there were some bugs about using sources + version number in the same state. and I'm on Salt 2015.5.3
21:30 fir3wall how can I check if user specified in pillar exist - if yes do nothing , if not create it . I've just want to check is user exists, without what is his home dir, group, password etc
21:30 knite allow_updates docs say that option relates to allowing packages to auto update themselves, but I'll give it a shot.
21:30 AndreasLutro fir3wall: so... a user.present state?
21:30 fir3wall AndreasLutro: thanks just found it , was too much looking on exist state!
21:31 edrocks joined #salt
21:31 knite no dice with allow_updates: true
21:31 djgerm hey i figured out that DDNS thing! The module DOES attempt to iterate through available in.addr-arpa zones based on the A record, and automagically adds it. (but fails silently if it can't…)
21:31 fir3wall ahh sorry yeah ok got it
21:32 AndreasLutro hmm dunno then knite. you'd probably have to read through the source code a bit
21:32 jaybocc2 knite: use pkg.latest
21:33 knite jaybocc2: I don't necessarily want latest? if I do a rolling deploy, some servers might have package version X and others X+1, with the version number given in a pillar file.
21:33 jaybocc2 but be careful it will always upgrade to the latest package when available
21:34 fracklen joined #salt
21:34 BretFisher joined #salt
21:34 knite isn't this, well, server management 101? I'm not trolling, I've worked with salt on a few occasions, and have pored over the docs. I'm really surprised that this would be broken, I'm assuming I'm doing something wrong...
21:34 fir3wall AndreasLutro: yeah but user.present if user exist but with different for example home path or password is that will not overwritte it ?
21:34 AndreasLutro fir3wall: not unless you actually specify the "home" arg
21:35 jfindlay knite: it seems that `allow_updates` should do what you're expecting, although there may be some issues with a source package as you mention
21:35 fir3wall AndreasLutro: yeah but I want to also create this user with specific home if not exist (from my defaults file) so want to avoid every salt-call to overwrite something that is already existing on system
21:35 aphor fir3wall: are you able to try this out?
21:36 jaybocc2 allow_updates is not what he wants: allow_updates (bool) --
21:36 jaybocc2 Allow the package to be updated outside Salt's control (e.g. auto updates on Windows). This means a package on the Minion can have a newer version than the latest available in the repository without enforcing a re-installation of the package.
21:36 knite jfindlay: I tried allow_updates per AndreasLutro earlier. the docs say that allows packages to auto-update. I tried it anyway and it didn't help. to be clear, I'm *not* using an extra sources arg, I just wanted to be pre-emptive and point that out from the get-go.
21:36 aphor fir3wall: salt is designed to be idempotent.
21:36 jaybocc2 he's probably looking for param bool only_upgrade: Only upgrade the packages, if they are already installed. Default is False. Currently only works with APT-based systems.
21:36 AndreasLutro fir3wall: in that case you'll have to wrap the home arg in a jinja if statement checking if the user exists
21:37 AndreasLutro knite: what distro is this on?
21:37 knite jaybocc2: that's not quite right either. I want, well, pkg.installed @ version X. if package is not present, it should be installed. if version too low, upgrade. if version too high, downgrade.
21:37 knite Ubuntu 14.04
21:37 fir3wall AndreasLutro: yeah so that was my original intention can i use state in if ? so {% if fred.user.exist %} something {% else %} something {% endif %} or I need to do something more complicated ?
21:38 AndreasLutro fir3wall: no, you have to use an execution module. {% if salt['module.function')(args) %}
21:38 fir3wall AndreasLutro: thanks
21:38 jfindlay knite: can you paste the state you're using and the output you're getting?
21:38 jaybocc2 knite: can you pastebin how your setting the version
21:38 jaybocc2 ^
21:38 AndreasLutro knite: is that even possible though? apt is not really designed to have multiple versions of the same package
21:38 spaceSub left #salt
21:39 jfindlay there's also that
21:39 AndreasLutro unless you have multiple apt repos configured and one of them just happens to have the version you need
21:40 AndreasLutro not sure that'd help, I think even if you specify "fromrepo" it might not force an up/downgrade
21:40 jfindlay AndreasLutro: but could be nice to have the exact version verified even if you know there will only be one version available
21:40 knite AndreasLutro: some repo software has support for multiple versions of the same package. alternately, you can do separate repos for, eg, master vs stage, providing different package versions.
21:40 AndreasLutro jfindlay: agreed
21:40 knite so I already have the repos configured elsewhere via a different salt state.
21:41 AndreasLutro I'd test this myself knite if the situation wasn't so hard/annoying to reproduce ;)
21:42 ajw0100 joined #salt
21:42 knite https://gist.github.com/ariscn/b1682fde90cc1e8b7f3e
21:42 knite super simple state.
21:42 AndreasLutro foo?
21:42 knite real package name redacted. :-)
21:43 AndreasLutro is the package name "version"?
21:43 AndreasLutro because if not you're using the wrong arg to specify the version...
21:43 lero joined #salt
21:43 jaybocc2 have you tried `foo: '==1'` or whatever the version # is
21:43 knite eh? foo: 1.2.3.4 installs foo @ 1.2.3.4
21:43 knite line 4 in gist.
21:44 jfindlay there's a version parameter to package.installed
21:44 AndreasLutro pretty sure that doesn't work!
21:44 knite doh, I think you guys are right!
21:44 knite I can do it this way if I have a sublist called pkgs.
21:44 jaybocc2 foo: version only works for pkgs not pkg
21:44 knite for a single package I need version as an arg instead.
21:45 knite if this fixes it, I'll happily put on my dunce cap...
21:45 BretFisher joined #salt
21:45 knite kinda surprised it accepted an invalid parameter, though.
21:46 AndreasLutro that's what happens when you have **kwargs in python
21:47 knite well, you could do something like: for arg in expected_args: kwargs.pop(arg); if lwargs: raise Exception('unexpected!')
21:47 AndreasLutro easier said than done when you have 3 layers of function calls with **kwargs
21:48 AndreasLutro in the pkg.installed function, **kwargs is passed on to 3 different other functions
21:51 teryx510 joined #salt
21:51 jfindlay I'm putting together crowdsourcing efforts for the salt sprint at saltconf to work on argument typing, which may be able to be plugged into the state system at some point
21:52 teryx510 joined #salt
21:52 bhosmer_ joined #salt
21:53 AndreasLutro how would that be implemented?
21:54 josue joined #salt
21:58 fir3wall AndreasLutro: is that looks more or less fine : {% if not {{ jasper_settings.user}} in salt['user.list_users']('') %}
21:59 AndreasLutro whatever works I guess :P
22:00 fir3wall does not work ;p expected token colon got }
22:00 AndreasLutro you don't use {{ }} inside {% blocks %}
22:00 AndreasLutro {{ }} is for printing
22:01 kliquori joined #salt
22:01 baweaver joined #salt
22:03 fir3wall ahh yeah still not work ... argument of type function is not iterable, ok will solve it tomorrow if not in next 5 min need some sleep. thanks for your help all
22:04 ninjada joined #salt
22:05 ninjada joined #salt
22:06 onlyanegg joined #salt
22:07 elsmo joined #salt
22:12 jab416171 can I dynamically create branches in my pillar repo, and then have that map to the environment I have set on the minion with the same name?
22:12 jab416171 or do you have to enumerate each branch name in the config?
22:20 huddy joined #salt
22:21 kliquori joined #salt
22:22 hoonetorg rallytime and cachedout around here (don't know their irc names, only github account names)
22:22 forrest hoonetorg, Nope they go by the same name and are not here.
22:22 mavhq joined #salt
22:22 hoonetorg k
22:22 hoonetorg forrest thx
22:23 forrest np
22:23 hoonetorg had some questions to a pr and thought it's easier and faster to discuss here
22:23 forrest Yeah I would just post on the PR.
22:24 hoonetorg i did
22:24 forrest Cool
22:24 hoonetorg but they don't seem to be around this week generally or busy or whatever
22:25 hoonetorg so i wait and continue to work on the pr as best as i can :)
22:26 amcorreia joined #salt
22:27 rmnuvg joined #salt
22:28 jhauser joined #salt
22:32 jfindlay AndreasLutro: I'm not sure yet, the arg types can be rst encoded in the function doc string
22:32 jfindlay hoonetorg: what questions do you have?
22:32 baweaver joined #salt
22:32 Rumbles joined #salt
22:33 hoonetorg https://github.com/saltstack/salt/pull/31859
22:33 saltstackbot [#31859]title: initial commit for a pcs module/state | This PR adds a new module and state for managing pacemaker/corosync-clusters via pcs....
22:33 hoonetorg still many tests fail
22:33 hoonetorg *checks
22:34 hoonetorg is it only because i have no tests implemented yet?
22:34 jfindlay nope, Mike and Nicole are very busy this week
22:34 jfindlay I've got to go to the train, but exec module fcns only need to return something serializable by msgpack
22:35 hoonetorg jfindlay: ah, ok that means build of development branch is generally broken?
22:38 brianfeister joined #salt
22:39 rmnuvg joined #salt
22:40 zenlot joined #salt
22:43 edrocks joined #salt
22:46 akhter joined #salt
22:48 drawsmcgraw joined #salt
22:53 cliluw joined #salt
22:56 onlyanegg is there a command to see everything the salt file server can serve? for debugging purposes?
22:57 Ch3LL_ onlyanegg: salt-run fileserver.file_list
22:57 brianfeister joined #salt
23:00 onlyanegg Ch3LL_: Thanks!
23:01 zer0def joined #salt
23:06 relidy_ joined #salt
23:09 Ch3LL_ onlyanegg: np
23:10 capricorn_1 joined #salt
23:10 ninjada joined #salt
23:12 onlyanegg is that possible from a minion? I am running a masterless configuration.
23:16 onlyanegg I don't have salt-run installed... I assume it comes in the master installation
23:17 AndreasLutro onlyanegg: check out the cp module - list_files or something may be what you want
23:22 onlyanegg AndreasLutro: thanks, that works
23:23 flowstate joined #salt
23:23 jfelchner joined #salt
23:23 noraatepernos joined #salt
23:25 noraatepernos I am still completely unclear on something regarding grains.  Shouldn’t sls files on master reference only abstractions of grains?  For instance, doesn’t using ec2_grains.py couple my sls files to amazon ec2?  How is this addressed with salt?  Or are all of our sls files supposed to be deployment environment specific?
23:26 nethershaw joined #salt
23:27 AndreasLutro noraatepernos: you'd check if the appropriate grains were set if they aren't set on every host
23:27 AndreasLutro you can also target .sls files with grains
23:29 colegatron_ joined #salt
23:31 noraatepernos AndreasLutro: Thanks.  In my case, I have (for now) amazon ec2 tag “role” which can be worker, webserver, database.  When an ec2 instance is launched, it is done so with this tag set…this determines the minion’s role and as such, in top.sls, is referenced with ‘worker:*’ — does that make sense?  How would I accomplish this?
23:31 noraatepernos I found this: https://github.com/saltstack/salt-contrib/blob/master/grains/ec2_tags.py
23:32 AndreasLutro noraatepernos: https://docs.saltstack.com/en/latest/ref/states/top.html#advanced-minion-targeting
23:36 jaybocc2_ joined #salt
23:37 djgerm left #salt
23:37 baweaver joined #salt
23:38 toanctruong joined #salt
23:39 rhodgin joined #salt
23:40 bhosmer_ joined #salt
23:43 flowstate joined #salt
23:43 rogst joined #salt
23:44 akhter joined #salt
23:45 ajw0100 joined #salt
23:56 mowntan joined #salt
23:56 mowntan joined #salt
23:56 rem5 joined #salt
23:56 mowntan joined #salt

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