Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-10-12

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

All times shown according to UTC.

Time Nick Message
00:03 ageorgop joined #salt
00:07 teebes joined #salt
00:12 ageorgop joined #salt
00:34 quasiben joined #salt
00:34 cberndt joined #salt
00:38 sunkist joined #salt
00:40 ageorgop joined #salt
01:16 subsignal joined #salt
01:19 pravka joined #salt
01:35 ageorgop joined #salt
01:41 aidalgol joined #salt
01:42 StolenToast joined #salt
01:45 catpigger joined #salt
01:47 ilbot3 joined #salt
01:47 Topic for #salt is now Welcome to #salt | 2015.8.0 is the latest | Please use https://gist.github.com for code, don't paste directly into the channel | Please be patient when asking questions as we are volunteers and may not have immediate answers | Channel logs are available at http://irclog.perlgeek.de/salt/
01:48 ashirogl joined #salt
01:55 jodv joined #salt
01:58 cberndt joined #salt
02:05 yomilk joined #salt
02:05 pravka joined #salt
02:41 scoates joined #salt
02:53 malinoff joined #salt
02:56 sunkist joined #salt
02:57 stanchan joined #salt
03:15 evle joined #salt
03:25 favadi joined #salt
03:27 moogyver joined #salt
03:35 falenn joined #salt
03:40 ramteid joined #salt
03:58 ^C ....
03:59 ^C dont jinja templates get the env var?
04:06 scoates joined #salt
04:10 writtenoff joined #salt
04:10 ashirogl joined #salt
04:12 armguy seems pkg.installed does not pick up groups properly in Arch Linux. base-devel, multilib-devel are two examples
04:14 armguy :( seems a long going issue https://github.com/saltstack/salt/issues/15749
04:32 jab416171 joined #salt
04:45 malinoff joined #salt
04:50 subsignal joined #salt
04:50 viq joined #salt
04:53 armyriad joined #salt
04:54 armyriad joined #salt
04:55 armyriad joined #salt
04:56 armyriad joined #salt
04:57 cberndt joined #salt
05:27 iggy ^C: saltenv
05:41 rdas joined #salt
05:43 larsfronius joined #salt
05:46 cberndt joined #salt
05:48 felskrone joined #salt
05:52 tiski joined #salt
05:54 tiski Hello. I am experiencing a few troubles running the salt users formula. I ve correctly (i think) defined my pillar with 2 users, and called users in my top.sls file, but it doesn't fire.
05:55 Bryson joined #salt
06:03 tiski if anyone has 2 minutes to spare, it would be greatly appreciated :)
06:05 jeddi joined #salt
06:07 jbrnds joined #salt
06:20 tiski ok nevermind, found it, got 2 formulas defining the "users" state
06:23 doiknowu joined #salt
06:24 doiknowu left #salt
06:28 jodv joined #salt
06:29 jodv_ joined #salt
06:32 keimlink joined #salt
06:40 KermitTheFragger joined #salt
06:41 mattiasr joined #salt
06:49 sirex joined #salt
06:53 elsmo joined #salt
06:53 GreatSnoopy joined #salt
06:56 niels__ joined #salt
06:57 sirex joined #salt
07:00 ^C iggy: tah, ended up using salt['pillar_get'] since that was available
07:07 aboe joined #salt
07:08 Grokzen joined #salt
07:12 Ztyx joined #salt
07:13 malinoff_ joined #salt
07:18 eseyman joined #salt
07:19 auzty joined #salt
07:21 mehakkahlon joined #salt
07:21 Ztyx1 joined #salt
07:23 Ztyx joined #salt
07:24 Ztyx1 joined #salt
07:25 Ztyx2 joined #salt
07:26 favadi joined #salt
07:28 av___ joined #salt
07:30 nadley joined #salt
07:35 lb joined #salt
07:35 denys joined #salt
07:36 Ztyx joined #salt
07:42 mr-op5 joined #salt
07:49 rdas joined #salt
07:53 amontalban joined #salt
07:53 kbaikov joined #salt
07:54 rim-k joined #salt
07:54 rim-k morning
07:54 krymzon joined #salt
07:56 netcho1 joined #salt
07:56 netcho1 hi all
07:58 rdas joined #salt
07:58 s_kunk joined #salt
08:00 jhauser joined #salt
08:04 rodr1c joined #salt
08:04 rodr1c joined #salt
08:05 [vaelen] joined #salt
08:06 chutzpah joined #salt
08:06 esharpmajor joined #salt
08:06 packeteer joined #salt
08:06 impi joined #salt
08:08 thalleralexander joined #salt
08:09 traph joined #salt
08:11 pcdummy joined #salt
08:13 zerthimon joined #salt
08:13 linjan_ joined #salt
08:18 chiui joined #salt
08:23 onovy joined #salt
08:23 netcho1 windows minion takes couple of seconds to respond while my linux minions respond in a flash... is that normal?
08:25 elsmo joined #salt
08:34 babilen netcho1: I would hope not, but then I haven't touched Windows in more than a decade
08:35 babilen Do you see anything interesting in the minion and/or master log? What happens if you run "salt-call -ldebug test.ping" on the Windows box?
08:39 TyrfingMjolnir joined #salt
08:42 Nippy joined #salt
08:57 felskrone1 joined #salt
09:00 sgargan joined #salt
09:11 kawa2014 joined #salt
09:16 mattiasr joined #salt
09:23 denys_ joined #salt
09:33 bernieke when I run "salt-run --async state.orchestrate <sls>" I get a salt/run/... tag
09:33 bernieke apparantly this is not the same as a jobid
09:33 bernieke how I can follow up on that asynchronous orchestrate task?
09:40 shiriru joined #salt
09:42 SheetiS joined #salt
09:42 sgargan joined #salt
09:49 pprkut anyone know how I could figure out what's wrong when "salt-call -l debug test.ping" just gives me SaltReqTimeoutErrors?
09:50 pprkut I'm having trouble connecting my minions to the master, and from looking at the logs, the problem seems to be somewhere in the backchannel
09:50 pprkut the master reacts almost immediately to the minions requests, but the response never seems to make it back to the minion
09:55 Gareth joined #salt
09:57 slav0nic joined #salt
10:02 jbrnds pprkut
10:02 jbrnds pprkut did you open the correct ports _to_ the master?
10:03 cmek is there an 'official' way to get 2015.8.0 on centos/rhel? epel only has 2015.5.5
10:05 cmek oh, ok saltstack repo
10:06 pprkut jbrnds: ports 4505 and 4506 are open
10:07 linjan_ joined #salt
10:14 pprkut looking at wireshark I see most communication packets are marked as "TCP Retransmission"
10:17 netcho1 babilen: everything works, it's just so darn slow
10:18 babilen First class capitalist software!
10:20 netcho1 it takes 6sec for test.ping to return True
10:20 netcho1 linux minions return in a hype
10:22 netcho1 is there anything in minion conf that i can change?
10:27 giantlock joined #salt
10:30 larsfronius joined #salt
10:33 keimlink joined #salt
10:45 VSpike Hm. netcho1 has gone, but I just timed 10 pings to a windows minion (0m29.875s) and 10 to a linux minions (0m5.480s) on the same subnet, so I'd say it could be normal
10:46 netcho1 joined #salt
10:47 VSpike Do those that maintain a Windows repo generally fork the official one in order to get more control? I'd assume so, in line with the best advice on formulas
10:47 netcho joined #salt
10:48 VSpike I currently have some extra packages of my own, and they are in /srv/salt/win/repo .. I'm wondering if I should just put them into my fork of the upstream one, or keep them separate. Not sure. Any thoughts?
10:48 VSpike netcho: I just timed 10 pings to a windows minion (0m29.875s) and 10 to a linux minions (0m5.480s) on the same subnet, so I'd say it could be normal
10:49 falenn joined #salt
10:49 VSpike netcho: the factor of about 6 times slower makes your 6 seconds sounds related
10:49 netcho master and minions are on the same esxi host
10:50 netcho 6sec por ping test is ridicoulus
10:50 netcho ridiculous*
10:51 johtso joined #salt
10:51 ingslovak joined #salt
10:52 VSpike netcho: just tested to a remote subnet. Got 0m6.025s
10:52 VSpike for linux and 0m24.360s for Windows
10:52 netcho VSpike: i ping minions from master... time almost equals for win and linux
10:52 VSpike Those were Server 2012 R2 ... the previous test was 2008 .. wondered if it would make a difference
10:52 netcho mine is 2012 also
10:53 VSpike netcho: you mean network ping rather than salt ping?
10:53 netcho yes
10:53 netcho how can i time salt test.ping?
10:53 VSpike Yeah, should be no difference here either. Must be something specific to the Windows minion
10:53 VSpike What version of salt?
10:54 netcho lithium
10:54 netcho 2015.5.3
10:54 netcho ubuntu master
10:54 VSpike OK, mine are recently upgraded to 2015.8.. wondered if it was a regression, but I guess not
10:54 breakingmatter joined #salt
10:55 netcho how can i time salt test.ping?
10:55 VSpike It would explain why doing a highstate on the Windows hosts is so slow .. as if the problem of process fork time and powershell startup time was not already bad enough
10:55 VSpike I just did: time for i in `seq 0 10`; do echo $i; salt aws-dev-p-app-1 test.ping >&/dev/null; done
11:03 edulix joined #salt
11:05 netcho VSpike: when i time salt 'win2012' test.ping i get 0.32 user, 0.04 system, 0:08.00 elapsed
11:05 netcho 8 f*** sec
11:05 VoidVector0 joined #salt
11:06 VoidVector0 left #salt
11:06 netcho linux machine in 0.39 s
11:07 impi joined #salt
11:07 MadHatter42 joined #salt
11:09 MadHatter42 why do I get duplicate results when i invoke test.ping
11:09 MadHatter42 as in
11:09 MadHatter42 # salt 'minion' test
11:09 MadHatter42 minion:
11:09 MadHatter42 'test' is not available.
11:09 MadHatter42 minion:
11:09 MadHatter42 'test' is not available.
11:09 MadHatter42 i mean
11:09 MadHatter42 # salt 'minion' test.ping
11:09 MadHatter42 minion:
11:09 MadHatter42 True
11:09 MadHatter42 minion:
11:09 MadHatter42 True
11:12 Norrland MadHatter42: please use gist, hastebin or such for sharing long output.
11:12 MadHatter42 sorry
11:12 MadHatter42 i thought it wasnt that long
11:12 MadHatter42 anyway
11:12 MadHatter42 any ideas regarding the issue ?
11:15 netcho VSpike: is the reconnect delay in salt-call from the win minion an issue maybe? cuz its also huge
11:16 netcho in debufg mode
11:16 netcho debug
11:16 netcho it's always almost max (11s)
11:17 Roee joined #salt
11:17 Roee Hi All
11:17 Roee Have a question maybe you could help
11:17 Norrland Roee: dont ask about asking
11:17 MadHatter42 i did the vim.sls from the guide
11:18 MadHatter42 and when I do salt '*' state.sls vim
11:18 MadHatter42 i get a
11:18 MadHatter42 The function "state.sls" is running as PID 8306 and was started at 2015, Oct 12 13:14:00.687218 with jid 20151012131400687218
11:18 MadHatter42 ERROR: Minions returned with non-zero exit code
11:18 Roee i'm runing a cmd.run from salt master that running a install.sh script
11:18 MadHatter42 the ping goes OK
11:19 Roee from some reason the installer (at the minion) is failed with an error
11:19 denys joined #salt
11:19 Roee Norraland: :)
11:19 Roee Norrland: :)
11:20 Roee And the error is : org.gradle.api.GradleException: Could not list contents of '/opt/t2k/tomcat/webapps/cgs/WEB-INF/client/cgsHelp/fr_FR/chap_6_fig_166_sc??nario_de_r??ponses80.png'.
11:20 netcho VSpike: coul it be cuz the win minion is 2015.8.0 ? and the master is 215.5.3?
11:20 Roee and as you can see seems that there is a problem to parse france characters in salt
11:21 Roee if i'm running the same command (install.sh) in the minion the installation is working well
11:21 Roee did someone ever saw such error ?
11:22 VSpike netcho: ah, so it could be minion version then
11:23 VSpike netcho: are you able to install the 2015.5.x minion and see if that makes a difference?
11:23 netcho i'm gonna try to upgrade the master to .8.
11:23 VSpike netcho: I wouldn't right away.. I have master and minion at 2015.8 and it's slow, like yours
11:23 VSpike Since we're both running 2015.8 minion, that means it /could/ be a new bug in that version
11:23 netcho hm
11:24 VSpike If you can, I'd try rolling the minion back first and see if it's the same
11:24 VSpike You can get older versions at https://repo.saltstack.com/windows/
11:25 netcho i know
11:25 netcho is there a difference if i delete the key from the master before uninstalling the minion from win or after?
11:26 Ztyx1 joined #salt
11:27 Norrland Roee: https://github.com/saltstack/salt/issues/7190
11:27 thalleralexander joined #salt
11:29 tjuup joined #salt
11:31 tjuup installing 2015.8.0 salt-minion gives AttributeError: 'module' object has no attribute 'to_str' on Ubuntu 14.04 Trusty
11:31 cornfeedhobo joined #salt
11:33 briner joined #salt
11:34 netcho VSpike: looks like it's a bug in 2015.8
11:34 netcho i rolled back to 2015.5.3 and it returns in 0.33
11:35 quasiben joined #salt
11:37 netcho VSpike: in this version sal-call debug does not generate the response delay
11:37 netcho rmq reconnect delay
11:37 netcho zmq*
11:38 VSpike netcho: sounds like it's worth raising an issue about it if one doesn't already exist
11:39 netcho feel free to do that :)
11:42 Roee Norrland: you are great !!!
11:43 Roee thanks
11:43 Norrland Roee: I do the google.
11:43 Norrland np
11:43 Roee me too, but i didn't find your page
11:44 tjuup reran the install without errors, after fixing apt W: Failed to fetch
11:46 Ztyx joined #salt
11:48 linjan joined #salt
11:48 saltman joined #salt
11:50 saltman Hello, I have a python script which in the first stage upgade my salt from 2014.1.7 to 2015.5.2, and in the second run a command with: self.local = salt.client.Caller("/etc/salt/minion")
11:50 saltman and then: ret = self.local.function(func, args)
11:51 saltman It gives me the following error:   File "/usr/lib/python2.6/site-packages/salt/state.py", line 600, in load_modules     ret = pillar.compile_pillar()   File "/usr/lib/python2.6/site-packages/salt/loader.py", line 286, in render     opts, salt.exceptions.LoaderError: The renderer yaml_jinja is unavailable, this error is often because the needed software is unavailable
11:51 saltman When I run it again (and start the process from sakt 2015, and not 2014), it fixed
11:52 saltman What can be the problem?
11:53 slav0nic_ joined #salt
11:57 _mel_ joined #salt
12:03 denys joined #salt
12:08 tmclaugh[work] joined #salt
12:13 clintberry2 joined #salt
12:17 MadHatter42 how do I copy a file directly to the server
12:17 MadHatter42 i mean, i want a specific sources.list file
12:18 MadHatter42 i want to start my salt script by coping that into the server
12:18 MadHatter42 and then running an apt-get update
12:18 Norrland MadHatter42: read the documentation about file.managed ?
12:18 MadHatter42 let me check
12:18 MadHatter42 i've just started using this today
12:20 Norrland MadHatter42: check around on github, look at https://github.com/saltstack-formulas/ for examples.
12:20 MadHatter42 i got the docs but its like 2300 pages of docs
12:21 Norrland well. Don't read them all then.
12:23 MadHatter42 Norrland, thnx for the link , these formulas are AWESOME
12:27 larsfronius joined #salt
12:30 chrisw joined #salt
12:30 chrisw can anyone point me to an example of using salt.states.cron.file?
12:30 chrisw not following what I should be using as the key...
12:31 larsfron_ joined #salt
12:34 elfixit joined #salt
12:36 opal joined #salt
12:38 lothiraldan joined #salt
12:40 mehakkahlon joined #salt
12:40 D3lirium joined #salt
12:43 armyriad joined #salt
12:44 SubOracle joined #salt
12:45 opal left #salt
12:47 stephanbuys joined #salt
12:56 subsignal joined #salt
13:04 DammitJim joined #salt
13:04 dendazen joined #salt
13:07 jdesilet joined #salt
13:07 ferbla joined #salt
13:15 nexsja^ joined #salt
13:15 nexsja joined #salt
13:17 superseb joined #salt
13:21 pprkut is there a way to test render pillar files similar to cp.get_template for formulas?
13:23 Deevolution left #salt
13:24 s_kunk joined #salt
13:25 armyriad joined #salt
13:26 pravka joined #salt
13:27 MadHatter42 joined #salt
13:32 clintberry joined #salt
13:34 Deevolution joined #salt
13:35 ingslovak joined #salt
13:35 eliasp anyone with more than 1 git_pillar repo around? are you able to include from one repo a pillar SLS from the other repo? I just end up with: http://pastie.org/10476641
13:35 lothiraldan joined #salt
13:36 perfectsine joined #salt
13:37 SneakyPhil joined #salt
13:38 zmalone joined #salt
13:40 VSpike I see a lot of win repo SLS files with paths like "salt://win/repo/foo/bar.msi" in them. Is it possible to get the master's win repo path from grains or pillar so this can be done more elegantly?
13:40 teryx510 joined #salt
13:41 VSpike Also, was I imagining it or did the minion used to see the master's config in the pillar data?
13:42 I joined #salt
13:43 cpowell joined #salt
13:44 eliasp VSpike: yes, since 2015.8.0 you could use Pillar in win_repo definitions
13:44 eliasp VSpike: the master config is not visible in Pillar data anymore since 2015.5.0
13:44 jeddi joined #salt
13:45 VSpike Ah I wasn't imagining it https://docs.saltstack.com/en/latest/topics/pillar/#master-config-in-pillar
13:46 hasues joined #salt
13:46 hasues left #salt
13:48 VSpike eliasp: that path would have to be explicitly in the pillar though, right? I don't think you can access that configuration value from jinja templating
13:48 MadHatter69 joined #salt
13:49 MadHatter69 joined #salt
13:49 MadHatter42 joined #salt
13:50 eliasp VSpike: yes, you'd have to place that path in Pillar… but then you could just use the same Pillar value to also define on your Master where this stuff goes
13:51 mapu joined #salt
13:53 jeffpatton1971 joined #salt
13:54 kawa2014 joined #salt
13:55 lothiraldan joined #salt
13:56 irctc821 joined #salt
14:00 timoguin joined #salt
14:07 _JZ_ joined #salt
14:12 zie joined #salt
14:13 andrew_v joined #salt
14:18 racooper joined #salt
14:18 MadHatter42 joined #salt
14:19 sgargan joined #salt
14:21 dendazen joined #salt
14:22 thalleralexander joined #salt
14:23 sgargan joined #salt
14:24 Akhter joined #salt
14:26 scoates joined #salt
14:27 ferbla joined #salt
14:28 zwi joined #salt
14:33 tkharju joined #salt
14:34 ALLmightySPIFF joined #salt
14:37 briner joined #salt
14:39 MadHatter42 joined #salt
14:40 teebes joined #salt
14:45 WalterBlack joined #salt
14:47 polishdub joined #salt
14:49 BogdanR Hello.
14:49 BogdanR Is it possible to create a new glusterfs module with the salt state?
14:49 BogdanR This is the one I am taling about: https://docs.saltstack.com/en/develop/ref/states/all/salt.states.glusterfs.html
14:51 linjan joined #salt
14:52 ageorgop joined #salt
14:53 sgargan joined #salt
14:55 Akhter joined #salt
14:56 twork latest n00b query: i've been stung once or twice by the "-/_" issue -- some stuff (jinja? yaml? De Lawd?) prefers one over the other, and i think there are places where the converse is the case... is there a map to this or am i just confused?
14:57 anmolb joined #salt
15:03 scoates joined #salt
15:03 bernieke in case anyone is interested, I found solutions to my two problems.
15:05 sgargan joined #salt
15:05 bernieke I can get to the salt-run --async output by listening to the events (salt-run state.event salt/run/<id>/ret count=1), and then scrape the json out of the output, and pass the dict loaded from that to salt.output.out_format
15:07 debian112 joined #salt
15:08 thalleralexander joined #salt
15:10 CeBe joined #salt
15:10 impi joined #salt
15:10 jalbretsen joined #salt
15:12 whidbeywalker joined #salt
15:17 hasues joined #salt
15:21 freelock joined #salt
15:22 hasues left #salt
15:23 RedundancyD joined #salt
15:24 lothiraldan joined #salt
15:25 Brew joined #salt
15:27 Brew joined #salt
15:29 Akhter joined #salt
15:31 netcho joined #salt
15:32 VSpike Starting wish I had not been so quick to update to 2015.8
15:32 zmalone joined #salt
15:34 VSpike Seems like the code to download installers for the win repo from HTTP(S) links is broken. I couldn't work out why the exe file in the cache wouldn't run, and why its SHA1 hash changed every time.
15:34 VSpike I pulled it across to a Linux box and looked at it, and this is what you get https://bpaste.net/show/d917d8ff2502
15:34 VSpike It seems to be following the redirect, but including the redirect text at the start of the downloaded binary file
15:39 kawa2014 joined #salt
15:40 ashirogl joined #salt
15:52 Fiber^ joined #salt
15:53 fyb3r joined #salt
15:53 netcho joined #salt
16:00 armyriad joined #salt
16:01 anmolb joined #salt
16:02 giantlock joined #salt
16:06 scoates joined #salt
16:07 bluenemo joined #salt
16:08 melog1 joined #salt
16:10 guanophobic joined #salt
16:12 kinetic joined #salt
16:12 Ztyx1 joined #salt
16:13 kinetic anyone know of good windows examples using salt?
16:13 mattiasr joined #salt
16:13 salt265 joined #salt
16:13 salt265 left #salt
16:20 mehakkahlon joined #salt
16:23 kinetic okay I guess not...
16:24 Norrland kinetic: could help a bit more to know what you're trying to do.
16:24 Norrland kinetic: salt 'windowsminion' test.ping  should work
16:25 kinetic I just wanted some decent examples of salt and windows, as far as I know only really basic salt commands work
16:26 kinetic <Norrland> I was just wondering if I'm missing something or if salt is still not super mature at working with windows
16:27 Norrland not sure. using the pkg tools could be a start.
16:27 Norrland I haven't administered a Windows environment in a few years. So not sure if you're better off with their own tools or not.
16:28 thehaven joined #salt
16:28 whytewolf kinetic: IIRC windows support is actually really mature. however I personally don't run windows so can not attest to that personally. UtahDave is the one to talk to about Windows in salt
16:29 jodv joined #salt
16:29 kinetic Norrland: yeah I was looking at the pkg tools, even those were a bit flaky at times, i.e. salt does not detect the latest version of chocolatey, but it DID detect 0.9.8
16:30 darkhuy joined #salt
16:30 scout_ joined #salt
16:30 toastedpenguin joined #salt
16:30 kinetic whytewolf: okay, I'll try pinging and getting some thoughts, thanks!
16:31 whytewolf wait until he pops his head up. he is rather busy sometimes
16:31 numkem joined #salt
16:31 kinetic okay
16:31 darkhuy has anyone run into an issue where minions return from direct commands via `salt '*' blah` but fail to return with salt states?
16:31 jeffpatton1971 pretty sure salt 'host*' cmd.run 'get-wmiobject win32_class'
16:32 jeffpatton1971 might work
16:32 jeffpatton1971 you may have to put PowerShell in there...not sure
16:33 whytewolf darkhuy: "fail to return" you mean it comes up blank? with out printing output?
16:34 darkhuy whytewolf no response from minion
16:35 darkhuy whytewolf but executing salt '*' cmd.run or test.ping all work fine
16:35 kinetic jeffpatton1971: I can get some of the basic commands to work, but I just end up essentially writing bash files and doing cmd.run rather than leveraging many of the cool tools Salt provides because most don't seem to play nice with windows, even things like the git pkg
16:36 jeffpatton1971 @kinetic I've not yet worked with salt in windows, but I have worked extensively with windows, my last job was a sysadmin for a large university with about 500 windows servers, and I did A LOT with DSC
16:36 KyleG joined #salt
16:36 KyleG joined #salt
16:36 Bryson joined #salt
16:37 alexlist joined #salt
16:38 kinetic jeffpatton1971: what is DSC?
16:38 Akhter joined #salt
16:38 jeffpatton1971 @kinetic it's salt for windows...except pure powershell...on the windows side it's amazing https://technet.microsoft.com/en-us/library/dn249912.aspx
16:39 jeffpatton1971 @kinetic not sure your comfort level with PowerShell but if you just need to keep windows servers in line...this is about the easiest way in windows to do it
16:40 kinetic jeffpatton1971: might be a dumb question, but will this have about the same effects if they are not servers?
16:40 kinetic jeffpatton1971: Put more plainly, I am controlling a host of Windows 10 computers
16:40 kinetic jeffpatton1971: not servers, just regular desktops
16:40 jeffpatton1971 @kinetic not a dumb question at all...there are some things that obviously won't work...but for the most part it should go there are some minimum req's depending on the version of windows we're talking about
16:41 jeffpatton1971 @kinetic, are they running windows server, or windows client? host architecture doesn't matter at all
16:41 kinetic jeffpatton1971: thank you! I am going to look into this
16:42 jeffpatton1971 @kinetic happy to help, and good luck! if you get stuck or have questions, message me
16:42 kinetic jeffpatton1971: thank you for the offer, will do
16:46 mehakkah_ joined #salt
16:47 VSpike kinetic: salt on Windows is rather slow and a bit breaky, because it uses powershell for a lot of things, but spawns a powershell process for every state, then parses the output as text.
16:47 VSpike Also, as you and jeffpatton1971 suggested, you end up using your own powershell scripts for a lot of things, but the same applies .. salt has to spawn it each time, and then parse the output
16:48 VSpike There seem to be a lot more minor but annoying issues on Windows, presumably because of the much smaller user base
16:48 jeffpatton1971 @VSpike with salt I could see that, for straight dsc it's just powershell and there are many built-in's as well as a surprisingly large community
16:49 VSpike yeah.. it would be awesome if salt could use DSC directly for its states, but hard to see how. There's no way for python code to interface directly with the managed code.
16:50 VSpike AFAIK Chef uses DSC now for most Windows configuration.. makes me want to test it out again if I'm going to have to keep Windows servers around
16:51 VSpike Ideally Salt would start one powershell instance and interact with it directly via API calls, so it can get the results back as objects instead of parsing text
16:52 VSpike I did briefly look at the powershell remoting stuff to see if it could be integrated into salt, but oh dear god no.
16:54 thefish joined #salt
16:57 jeffpatton1971 one thing you could possibly do, which is a little hurty, is hand write mof files and let salt do the pushing somehow
16:57 jeffpatton1971 at the end of the day, the mof in DSC is what makes the magic happen
16:58 jeffpatton1971 those are what the powershell configurations are compiled into
16:58 aparsons joined #salt
16:59 kinetic hmm that may be a good way to do it then
16:59 jeffpatton1971 you may wind up losing the DSC magic you get from powershell...but the end result would probably be the same? IE I need this server to be a domain controller would result in the server being a domain controller...but DSC would not really know as it never went through the configuration process or was involved in it
16:59 jeffpatton1971 @kinetic using chef?
17:00 qx joined #salt
17:00 jeffpatton1971 I've not used it so I don't have any thoughts either way...my current boss used it for Linux and when I showed him the salt he wished he had used that in his last gig
17:00 kinetic jeffpatton1971: any thoughts on puppet?
17:01 kinetic with windows I mean
17:01 jeffpatton1971 @kinetic not really, I came from 7yrs of being a windows admin and using powershell/remoting, and finally DSC for everything to being a Linux admin and rolling out salt in order to reign in the crazy
17:02 jeffpatton1971 i had tossed it out there as one of the choices though
17:02 jeffpatton1971 that and ansible as i know one of the core dev guys
17:02 nate_c joined #salt
17:02 kinetic jeffpatton1971: I played with Ansible, it felt very clunky
17:03 jeffpatton1971 well...the one benefit is there is no agent
17:03 jeffpatton1971 so for a lot of folks that's a plus
17:03 Lionel_Debroux joined #salt
17:04 jodv joined #salt
17:04 ageorgop joined #salt
17:05 jodv joined #salt
17:06 jodv joined #salt
17:06 jodv joined #salt
17:08 zer0def joined #salt
17:08 malinoff joined #salt
17:12 breakingmatter joined #salt
17:13 conan_the_destro joined #salt
17:15 CeBe joined #salt
17:15 Norbell joined #salt
17:22 egil joined #salt
17:23 Ryan_Lane no need for an agent with salt either :)
17:23 Ryan_Lane I've been running masterless for like 2 years
17:23 jeffpatton1971 salt-ssh ?
17:23 Ryan_Lane nope
17:23 jeffpatton1971 ?
17:23 Ryan_Lane pull-based deployment
17:24 jeffpatton1971 tell me more...what is this?
17:24 Ryan_Lane artifacts in S3 that get pulled onto the nodes on changes. salt runs whenever a new artifact shows up
17:24 jeffpatton1971 where's the doc for this?
17:24 Ryan_Lane http://ryandlane.com/blog/2015/04/02/saltconf15-masterless-saltstack-at-scale-talk-and-slides/
17:25 Ryan_Lane no doc. this is something custom
17:25 Ryan_Lane it's pretty simple though
17:25 Ryan_Lane you generate tarball artifacts of your code (and salt) and push them into S3, then you have a file that lists the current artifact version (which for us is a git sha)
17:26 Ryan_Lane on each node, you have a cron that runs every minute that will see if the current SHA is different than the one listed in S3. if it is, it pulls down the SHA tarball, untars it and then runs salt
17:26 jeffpatton1971 how difficult would it be to not use S3 and instead azure?
17:26 Ryan_Lane I'm sure simple, since they have an S3 equivalent
17:27 Ryan_Lane our code is too specific to us to open source, but the basic version of this idea is like 100 lines of python
17:27 Ryan_Lane plus another 50 or so for the part that generates the artifacts, uploads them, and changes the "current" file
17:28 JDiPierro joined #salt
17:28 Ryan_Lane you do still need something like salt-ssh for emergencies
17:29 Ryan_Lane we're using fabric for that, but it's so infrequently used that we haven't switched to salt-ssh because we haven't felt it's worth the effort at this point
17:29 baweaver joined #salt
17:29 jeffpatton1971 ya...you don't need to leverage pillars or anything like that as your code IS the master?
17:29 Ryan_Lane using a deployment method like this means you do need to be pretty eventually consistent for deployments, too, but that's a /good/ thing
17:29 Ryan_Lane we use both pillars and grains
17:30 Ryan_Lane but they all live on the nodes
17:30 impi joined #salt
17:30 Ryan_Lane we use the pillars for handling different environments and sub-services
17:30 pravka joined #salt
17:30 baweaver joined #salt
17:30 jeffpatton1971 listening to your talk...while i bug you ;-) so...each node is like a master?
17:30 Ryan_Lane watch the talk I linked (or see some of the earlier blog posts the talk is based on)
17:30 Ryan_Lane basically, because all the files come from the node itself
17:31 Ryan_Lane we do use external pillars, but for secret management
17:31 Ryan_Lane since you're not in AWS I'd recommend vault or keywhiz for that
17:32 Ryan_Lane but the external pillars run on the nodes themselves, so they only pull the pillars for themselves
17:32 Akhter joined #salt
17:32 Ryan_Lane and access control to those pillars is restricted by our secret management system
17:35 pm90_ joined #salt
17:36 jeffpatton1971 how long did it take you to get to this state?
17:38 Ryan_Lane we switched from puppet to salt and this was our initial path
17:38 Ryan_Lane we wanted no master
17:39 Ryan_Lane it took us about 3 months to switch basically everything
17:39 jeffpatton1971 ok, so you already had some mgmt. in place
17:39 jeffpatton1971 that's not too bad
17:39 denys joined #salt
17:42 murrdoc joined #salt
17:42 jeffpatton1971 @Ryan_Lane so you wrote custom modules to handle all of your amazon stuff and then use those in your states? did i get that right?
17:42 Ryan_Lane all the AWS stuff we do is upstreamed
17:43 Ryan_Lane the deployment stuff we have uses salt, but the python api. it's just a basic python script that talks to S3
17:46 minaguib joined #salt
17:47 writteno1 joined #salt
17:48 aboe joined #salt
17:49 keimlink joined #salt
17:51 anotherZero joined #salt
17:55 rim-k joined #salt
17:55 CeBe joined #salt
17:59 elsmo joined #salt
18:00 DammitJim joined #salt
18:02 mattiasr joined #salt
18:02 murrdoc joined #salt
18:04 Ztyx joined #salt
18:05 fyb3r1 joined #salt
18:07 ashirogl joined #salt
18:08 Ryan_Lane basepi: something in 2015.8 is leaking memory when we use the salt python caller api
18:08 Ryan_Lane I'm trying to track it down, but not having a ton of luck, what kinds of things may eat memory when using the caller api?
18:08 giantlock joined #salt
18:09 GreatSnoopy joined #salt
18:10 Ryan_Lane basepi: this is basically all we're doing: https://gist.github.com/ryan-lane/ff8752b30af4c87d5118
18:10 stanchan joined #salt
18:10 Ryan_Lane salt_caller.opts and salt_caller.sminion.opts seem to be stable from call to call
18:17 baweaver joined #salt
18:18 rim-k joined #salt
18:19 wendall911 joined #salt
18:24 stanchan_ joined #salt
18:25 murrdoc joined #salt
18:27 pm90_ joined #salt
18:30 ferbla joined #salt
18:34 krymzon joined #salt
18:37 basepi Thanks, cachedout is seeing it as well and is investigating.
18:37 timoguin joined #salt
18:38 Ryan_Lane I'm trying to track it using pympler. we'll see how thayt goes
18:38 Akhter joined #salt
18:38 racooper joined #salt
18:39 Akhter joined #salt
18:48 pm90__ joined #salt
18:50 perfectsine joined #salt
18:51 rim-k joined #salt
18:53 Brew joined #salt
18:55 katyucha joined #salt
18:56 ferbla joined #salt
18:58 jeffspeff joined #salt
18:58 netcho basepi: what about th delay mentioned this morning on 2015.8 windows minion?
18:58 basepi ?
18:58 basepi I haven't heard anything about that.
18:58 Ztyx joined #salt
18:58 netcho windwos minion takes about 8 secs to return ping
18:59 iggy reboot
18:59 netcho in my case
18:59 netcho VSpike: also has that problem
18:59 netcho i rolled back on .5 and it's fine now
18:59 iggy you should open an issue
18:59 iggy so it can be properly tracked
19:00 basepi Yep, please open an issue.
19:00 iggy "it happened on IRC" bugs don't get tracked
19:00 netcho i know :)
19:00 KyleG they do in a way.
19:00 KyleG I've tracked bugs down due to the IRC logs online that show up in google results
19:00 KyleG lol
19:02 jeffspeff i'm having an issue with a custom module. it's a simple one that supposed to just create scheduled task in windows. however, even when running the module manually, it doesn't seem to be actually running the command to create the task. i'm not able to find any errors and the task is not being created on the machine. https://gist.github.com/jeffclay/81b54de29f9f3192988a
19:03 rlarkin joined #salt
19:05 larsfronius joined #salt
19:05 iggy jeffspeff: play around with shell arg?
19:05 mattiasr joined #salt
19:06 zmalone joined #salt
19:06 falenn joined #salt
19:07 jeffspeff iggy, i'm familiar with how to use the shell arg with cmd.run but not when using module.run. i don't have any shell arguments built into this module, it's just python code.
19:09 iggy I meant for the cmd.run
19:11 jeffspeff i'm using module.run in my state
19:11 jeffspeff oh wait
19:11 iggy but the module calls cmd.run
19:11 jeffspeff i see that
19:11 conan_the_destro joined #salt
19:12 jeffspeff sorry, it's been quite a while since i worked on this
19:12 iggy did it ever work?
19:12 Ryan_Lane so, it's something growing in the salt_caller and not the sminion
19:12 jeffspeff it did when i went from salt 2014.x.x to 2015.5.X
19:13 jeffspeff i'm now trying to go from 2015.5.x to 205.8.0
19:13 iggy jeffspeff: there are some bugs open about env changes
19:13 iggy (notably HOME not being set along with most of the other stuff that used to be set)
19:14 jeffspeff hmm
19:14 jeffspeff any idea how i can work around that in this case?
19:14 iggy figure out what bit of the ENV you need and add it to the cmd.run
19:15 iggy there's not a "add X arg to cmd.run and it fixes it"
19:20 zwi joined #salt
19:26 jeffspeff iggy, my master is already updated to 2015.8. are there any issues with leaving my minions on 2015.5 for now?
19:27 amcorreia joined #salt
19:29 baweaver joined #salt
19:34 hasues joined #salt
19:35 hasues left #salt
19:37 slav0nic joined #salt
19:37 iggy I haven't heard of any
19:38 iggy but that could be because I think most people don't do it
19:38 iggy ymmv (and other standard disclaimers)
19:39 jmreicha joined #salt
19:39 jeffspeff ok, i'll wait it out and see what happens. thanks for the help... as always
19:43 fusionx86 joined #salt
19:44 Ryan_Lane basepi: I'm pretty sure opts is being injected into sub-objects
19:44 Ryan_Lane I'm seeing opts multiple layers deep
19:45 Ryan_Lane though it may be whatever contains opts that's being injected into sub-objects that's growing down the chain
19:48 aidalgol joined #salt
19:52 baweaver joined #salt
19:52 flyx joined #salt
19:53 flyx left #salt
20:00 ajw0100 joined #salt
20:00 tpaul joined #salt
20:00 breakingmatter joined #salt
20:00 clintberry joined #salt
20:01 tpaul Is anyone using salt SmartOS minions? In particular with lx-branded zones? I'm wondering how I can target lx-branded zones specially so I don't apply certain states (such as iptables).
20:02 KyleG tpaul: There should be a grain item that says what type it is
20:02 KyleG List out all the grains for the host, I believe there is one
20:03 zmalone I bet the kernel or kernelrelease grain gives away that it is a lx brand zone.
20:03 zmalone if not one of the os ones, but I think kernel is a safer bet
20:04 tpaul All the suspect grains (kernel, os, etc) say that its a Linux CentOS machine
20:04 tpaul Which is kind of funny
20:05 zmalone I'd use https://docs.saltstack.com/en/latest/topics/targeting/grains.html#writing-grains then
20:05 zmalone maybe check for the presence of zfs, which you probably don't have on your "real" CentOS hosts
20:05 zmalone or dtrace.
20:06 tpaul zmalone: that might work
20:06 baweaver joined #salt
20:08 tpaul yeah, sounds like I need my own grain, and your suggestion of looking for dtrace or zfs should work well, thanks!
20:08 kunersdorf joined #salt
20:09 KyleG I know on my SmartOS KVM vm's, it shows KVM as the virtual tech in one of the grains
20:09 KyleG and my BSD jails show jail, and so on
20:09 KyleG No reason not to expect the same sort of thing for an LXC
20:09 kunersdorf I'm trying to use git pillar on 2015.8.0 and I'm getting "Unable to fetch SSH-based git_pillar remote  libgit2 must be compiled with libssh2 to support SSH authentication."
20:10 kunersdorf Im running latest libgit2 and libssh2 for centos7, do I need to run the FC pkgs?
20:11 aparsons joined #salt
20:12 linjan joined #salt
20:12 mattiasr joined #salt
20:15 Akhter joined #salt
20:20 ipmb joined #salt
20:20 pm90_ joined #salt
20:21 cliluw joined #salt
20:21 pm90__ joined #salt
20:22 cliluw joined #salt
20:24 scoates joined #salt
20:26 felskrone joined #salt
20:28 cliluw joined #salt
20:31 geomyidae_ joined #salt
20:32 cberndt joined #salt
20:34 DammitJim joined #salt
20:35 jeffpatton1971 i can do salt -G 'role:slave' test.ping to check all servers marked with role slave, is it possible to invert that? get a list of not role = slave?
20:40 moogyver you can exclude a minion id with 'not' but not sure about using it to exclude based on a grain/pillar/etc
20:42 melog1 joined #salt
20:43 CeBe joined #salt
20:43 jeffpatton1971 hrm
20:47 notnotpe_ joined #salt
20:47 whytewolf jeffpatton1971: it is rather a round about way but salt -c '* and not G@role:slave' test.ping
20:48 zmalone Huh, that's a cool approach.
20:48 whytewolf err -C
20:48 zmalone Salt matching is quirky.
20:48 zmalone https://docs.saltstack.com/en/develop/topics/targeting/compound.html#alternate-delimiters is a place where that's briefly mentioned.
20:49 moogyver in 2015.8 i think you don't need the '* and' anymore
20:49 pravka joined #salt
20:49 ashirogl joined #salt
20:50 whytewolf moogyver: I think you are right. tried it on my 2015.8 dev box and it worked with out the '* and' part and it worked. but with a mix of minions in 2015.5 and a master of 2015.8 might be better to use it
20:50 moogyver agreed
20:56 Akhter joined #salt
20:58 jeffpatton1971 @whytewold worked like a charm on 2015.5! Thanks!
21:02 whytewolf jeffpatton1971: no problem. honestly it is the same kind of logic that DBA's use on scripted SQL queries of having a 1=1 on the end so that if nothing gets added to the where clause sql doesnt dump an error from haveing where;
21:05 ViciousLove joined #salt
21:06 larsfronius joined #salt
21:06 kawa2014 joined #salt
21:08 moogyver hrm.  is there anyway of keeping the salt acl's external?  i know you can do the auth externally, but the actual targetting of what commands + what hosts
21:11 Norbell joined #salt
21:13 N-Mi joined #salt
21:14 Bryson joined #salt
21:16 Akhter joined #salt
21:17 baweaver joined #salt
21:19 teebes joined #salt
21:19 jmreicha_ joined #salt
21:27 teebes joined #salt
21:35 elsmo joined #salt
21:36 tmclaugh[work] joined #salt
21:38 quasiben joined #salt
21:39 cliluw joined #salt
21:41 pravka joined #salt
21:43 baweaver joined #salt
21:45 jmreicha joined #salt
21:46 baweaver joined #salt
21:48 jmreicha_ joined #salt
21:53 edulix joined #salt
21:54 sgargan joined #salt
21:56 GrueMaster joined #salt
21:58 moderation joined #salt
21:59 Ryan_Lane basepi: well, I give up. I can't track it down. something in salt.client is growing
21:59 mattiasr joined #salt
21:59 jmreicha joined #salt
22:00 wiqd joined #salt
22:00 Ryan_Lane so it's something that adds directly onto the module. as far as I can tell the caller object itself doesn't seem to be growing, but I could be wrong
22:00 OliverMT joined #salt
22:00 NachoDuck joined #salt
22:00 Ryan_Lane regenerating the object doesn't seem to decrease the process's memory, so it would also indicate that the leak is something on the module
22:04 bbhoss joined #salt
22:05 tmclaugh[work] joined #salt
22:09 stanchan joined #salt
22:11 sgargan joined #salt
22:11 simonmcc joined #salt
22:11 zie joined #salt
22:12 rim-k joined #salt
22:14 gazarsgo joined #salt
22:16 flyx joined #salt
22:18 codekobe joined #salt
22:23 LyndsySimon joined #salt
22:26 trave joined #salt
22:26 EWDurbin joined #salt
22:27 quasiben joined #salt
22:28 bfoxwell joined #salt
22:29 bfoxwell joined #salt
22:33 toastedpenguin joined #salt
22:35 tmkerr joined #salt
22:36 jay_d joined #salt
22:37 thalleralexander joined #salt
22:37 fxhp joined #salt
22:38 pm90_ joined #salt
22:39 ashirogl joined #salt
22:39 pm90__ joined #salt
22:45 r3tic3nc3 joined #salt
22:47 basepi Ryan_Lane: it's kicking cachedout's butt too. He found a place where we were leaking a singleton map via weakrefs, but after fixing it we're still not reclaiming the memory.
22:51 tmclaugh[work] joined #salt
22:53 tcolvin joined #salt
22:57 scoates joined #salt
22:59 ashb joined #salt
23:01 sjorge joined #salt
23:01 sjorge joined #salt
23:04 baweaver joined #salt
23:13 sunkist joined #salt
23:18 fusionx86 joined #salt
23:19 perfectsine joined #salt
23:20 pm90_ joined #salt
23:22 breakingmatter joined #salt
23:26 scoates joined #salt
23:28 CaptainMagnus joined #salt
23:29 CaptainMagnus_ joined #salt
23:29 CaptainMagnus_ left #salt
23:32 CaptainMagnus Hey all. In top.sls, is it possible to compound match on grains and nodegroups (something like this: 'G@ec2_tags:Roles:nginx and G@ec2_tags:Application:frontend and N@prod' where prod is a nodegroup)?
23:33 iggy CaptainMagnus: yes, is it not working?
23:34 iggy CaptainMagnus: oh wait, lists work, nodegroups don't
23:34 CaptainMagnus iggy,  No, it's not working. Just updated to 2015.8.0 to try it as well but no go
23:35 iggy I think because nodegroups already allow compound matches, it could end up with some sort of recursive compound matcher
23:35 CaptainMagnus Fair enough. Thanks iggy
23:45 tmclaugh[work] joined #salt
23:48 mattiasr joined #salt
23:48 falenn joined #salt
23:51 otter768 joined #salt
23:52 baweaver joined #salt

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