Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2016-05-16

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

All times shown according to UTC.

Time Nick Message
00:06 ninjada joined #salt
00:27 autofsckk joined #salt
00:31 rem5 joined #salt
00:34 estahn joined #salt
00:35 ronrib joined #salt
00:35 futuredale joined #salt
00:38 aqua^c joined #salt
00:41 penguinpowernz joined #salt
00:42 Edgan joined #salt
00:42 brianfeister joined #salt
00:45 amcorreia joined #salt
00:45 rem5 joined #salt
00:46 Edgan_ joined #salt
00:49 Edgan__ joined #salt
00:56 Edgan__ joined #salt
00:58 quasiben joined #salt
01:00 iceyao_ joined #salt
01:07 west575 joined #salt
01:32 catpigger joined #salt
01:33 Vaelatern joined #salt
01:34 hasues joined #salt
01:34 hasues left #salt
01:37 rovar joined #salt
01:38 hal58th joined #salt
01:46 thehaven joined #salt
01:48 ilbot3 joined #salt
01:48 Topic for #salt is now Welcome to #salt! | Latest Versions: 2015.5.10, 2015.8.8, 2016.3.0rc3 | Support: 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
02:03 AdamSewell joined #salt
02:15 quasiben joined #salt
02:16 Vishvendra joined #salt
02:22 quasiben joined #salt
02:23 k_art joined #salt
02:25 k_art I took a minion out of our live cluster and am currently trying to connect it to a different (development) master. I have changed the IP in /etc/hosts, and have removed /etc/salt/pki/minion/minion_master.pub. When restarting the minion it still connects to the old master. What am I missing?
02:36 k_art Previous master was still in /etc/salt/minion.d/minion.conf. Very confusing to have the master in two different places.
02:37 k_art (/etc/salt/minion.d/minion.conf and /etc/salt/minion)
02:43 favadi joined #salt
02:52 brianfeister joined #salt
02:57 josuebrunel joined #salt
03:07 karlthane joined #salt
03:08 kshlm joined #salt
03:09 dezertol joined #salt
03:10 izaki joined #salt
03:11 ninjada joined #salt
03:12 ninjada joined #salt
03:19 dezertol joined #salt
03:20 pipps joined #salt
03:32 estahn joined #salt
03:39 hal58th joined #salt
03:41 zenlot joined #salt
03:51 hoonetorg joined #salt
04:03 rem5 joined #salt
04:05 cswang_ joined #salt
04:07 estahn joined #salt
04:13 sarlalian joined #salt
04:18 mowntan joined #salt
04:22 mowntan /quit
04:22 mowntan joined #salt
04:34 mpanetta_ joined #salt
04:40 brianfeister joined #salt
04:41 k_art Where can I actually see which command is executed when Salt reports: "[ERROR   ] Container '****' cannot be started"?
04:46 teatime dunno, but it's probably in either the minion or master log, so check both.
04:47 teatime you can stop both services, and run both in the foreground with -l info or even -l debug, also
04:55 k_art running with -l debug, neither logfile nor console show the command :-(
05:03 brianfeister joined #salt
05:11 papertigers joined #salt
05:13 papertigers Trying to write a custom grain and im importing salt.module.foo, however foo.something is giving me NameError: global name '__salt__' is not defined
05:14 papertigers can anyone point me in the right direction
05:15 mavhq joined #salt
05:15 nidr0x joined #salt
05:18 keimlink joined #salt
05:24 hoonetorg joined #salt
05:28 mohae joined #salt
05:40 hal58th joined #salt
05:49 favadi joined #salt
05:56 Vishvendra joined #salt
05:56 sarlalian joined #salt
05:56 rdas joined #salt
06:05 AndreasLutro papertigers: you can't use salt modules in grains, because many modules depend on grains
06:07 papertigers AndreasLutro: hmm this is a small module with no grains
06:07 teatime papertigers: could it be an execution module instead?
06:08 papertigers i wanted to use the module to populate a grain
06:09 papertigers guess i could make it a state and set the grain
06:10 papertigers basically i need to pull a lists of hosts out of salt to generate a DNS config
06:10 teatime oh, you probably want mine for that
06:10 teatime I've been working on the same hting
06:10 teatime you're gonna generate the zone files?
06:10 teatime it's kind of a pita
06:10 teatime one of those things that isn't hard, but it's hard to decide the least crappy ways to do things
06:11 AndreasLutro papertigers: doesn't matter, because *some* (many) modules rely on grains, all modules have to be loaded after grains
06:12 teatime pillar is for master giving data to minion.  grains are minion giving data to master.  minions *can* query each other during state/module run, but mine was created to make that not suck.  mine is for when you want minions to deposit data into a central location, that other minions can query later at their leisure.
06:12 teatime you can control which minions have access to which mine data, if necessary.
06:14 impi joined #salt
06:15 papertigers right, in this case the minions can have multiple ips. i can pull the network uuid out for each nic which is what i wanted in a grain uuid:ip
06:15 papertigers ill rethink this based on what you just said
06:31 teatime eh, tbh I'm a n00b and I've been awake for two days
06:31 teatime so beware my advice.
06:33 ageorgop joined #salt
06:35 mavhq joined #salt
06:40 rihannon joined #salt
06:41 rihannon1 joined #salt
06:51 EulerWiles joined #salt
06:52 Miouge joined #salt
06:53 ageorgop joined #salt
06:53 hlub wtf, I just read that it is possible to require a whole sls file by including it and then adding a requisite 'sls: some.sls.file' but this is definitely not working.
06:54 EulerWiles left #salt
06:54 Ayo joined #salt
07:05 punkoivan joined #salt
07:05 hlub "As of Salt 0.16.0, it is possible to require an entire sls file."
07:11 hlub ah, it seems that it can be used only with require, not with onfail.
07:12 hlub :(
07:15 kawa2014 joined #salt
07:29 josuebrunel joined #salt
07:33 John_Kang joined #salt
07:37 ninjada_ joined #salt
07:41 hal58th joined #salt
07:44 John_Kang left #salt
07:51 manji joined #salt
07:53 manji joined #salt
07:54 manji joined #salt
07:55 XenophonF joined #salt
07:56 k_sze[work] joined #salt
07:57 Vishvendra joined #salt
08:00 iceyao joined #salt
08:04 impi joined #salt
08:04 dgutu joined #salt
08:05 Xenophon1 joined #salt
08:16 linjan__ joined #salt
08:17 rim-k joined #salt
08:20 ivanjaros joined #salt
08:24 tuxx joined #salt
08:26 rim-k ls
08:26 dmaiocchi joined #salt
08:36 SpX joined #salt
08:39 rrei joined #salt
08:41 GreatSnoopy joined #salt
08:45 rrei joined #salt
08:47 pipps joined #salt
08:47 XenophonF joined #salt
08:51 JohnnyRun joined #salt
08:58 ninjada joined #salt
09:16 impi joined #salt
09:21 pipps joined #salt
09:30 Rumbles joined #salt
09:32 catpig joined #salt
09:41 hal58th joined #salt
09:49 impi joined #salt
10:01 kshlm joined #salt
10:03 Rumbles joined #salt
10:12 evle joined #salt
10:21 Rumbles joined #salt
10:24 Vishvendra1 joined #salt
10:31 quasiben joined #salt
10:31 Vishvendra joined #salt
10:45 AdamSewell joined #salt
10:48 toastedpenguin1 joined #salt
10:50 AdamSewell joined #salt
10:50 yidhra joined #salt
10:51 pipps joined #salt
10:57 ninjada joined #salt
10:58 Biopandemic joined #salt
11:04 ivanjaros3916 joined #salt
11:13 Rumbles I'm trying to install mongo-hacker via npm using salt, but I get a weird error unless I give a user. Any suggestions? https://paste.fedoraproject.org/367055/
11:19 kshlm joined #salt
11:20 ninjada joined #salt
11:20 Kruge left #salt
11:26 AdamSewell joined #salt
11:27 Fiber^ joined #salt
11:31 amcorreia joined #salt
11:42 hal58th joined #salt
11:44 aljosa joined #salt
11:45 ninjada joined #salt
11:55 CeBe joined #salt
12:01 TooLmaN joined #salt
12:07 izaki joined #salt
12:16 debian112 joined #salt
12:21 west575 joined #salt
12:21 ninjada joined #salt
12:26 WildPikachu joined #salt
12:36 DammitJim joined #salt
12:37 martoss joined #salt
12:38 dmaiocchi joined #salt
12:42 tracphil joined #salt
12:42 garphy joined #salt
12:47 XenophonF joined #salt
12:47 TooLmaN joined #salt
12:49 komputes joined #salt
12:53 garphy` joined #salt
12:55 kus joined #salt
12:57 mohae joined #salt
12:59 ssplatt joined #salt
13:00 Calpico joined #salt
13:02 wendall911 joined #salt
13:03 pipps joined #salt
13:03 dijit 1; does the homedir exist, 2; can you use a full path. ~ isn't expanded automatically probably.
13:04 Calpico joined #salt
13:04 rem5 joined #salt
13:08 Calpico joined #salt
13:10 bhosmer_ joined #salt
13:11 toastedpenguin joined #salt
13:13 dyasny joined #salt
13:15 Calpico joined #salt
13:15 martoss joined #salt
13:18 Ayo joined #salt
13:25 mapu joined #salt
13:25 edrocks joined #salt
13:26 evle joined #salt
13:30 racooper joined #salt
13:30 haam3r joined #salt
13:32 andrew_v joined #salt
13:34 ivanjaros joined #salt
13:37 toastedpenguin joined #salt
13:40 andrew-l joined #salt
13:40 _JZ_ joined #salt
13:43 hal58th joined #salt
13:43 toastedpenguin joined #salt
13:49 dfinn joined #salt
13:52 wangofett joined #salt
13:53 elsmo joined #salt
13:53 squishypebble joined #salt
13:53 wangofet1 joined #salt
13:54 Tanta joined #salt
13:58 elsmo joined #salt
14:04 jerredbell joined #salt
14:05 mpanetta_ joined #salt
14:05 Elsmorian joined #salt
14:06 mpanetta_ joined #salt
14:20 esharpmajor_ joined #salt
14:21 edrocks joined #salt
14:24 manji joined #salt
14:24 ntropy are saltconf16 videos going to be published any time soon?
14:25 kaptk2 joined #salt
14:30 hasues joined #salt
14:30 hasues left #salt
14:33 manji joined #salt
14:33 scaryclam joined #salt
14:34 ssplatt joined #salt
14:36 keimlink joined #salt
14:37 farblue Hi all :) Is there an easy way in jinja to truncate a string before a token. for instance, if I have the string “172.17.0.0/29” and I want to strip the “/29”?
14:38 Heartsbane joined #salt
14:38 tvinson farblue: .split('/')[0] should work
14:39 farblue ah, right, is that a python thing rather than a jinja thing?
14:41 tvinson python -c "import jinja2 ; print jinja2.Template(\"172.17.0.0/29\".split('/')[0]).render()"
14:41 tvinson looks like jinja's got those moves
14:42 AndreasLutro you can call python object methods in jinja, and a string is a python object
14:43 scaryclam Does any know the best way to get an environment variable via an sls?
14:43 farblue tvinson: I’ve heard jinja can use python but I’m not a python developer so I know what I want to do but just sometimes can’t find the syntax :)
14:44 farblue scaryclam: you mean to use an environment variable in an sls?
14:44 scaryclam Yes
14:44 farblue https://docs.saltstack.com/en/latest/topics/tutorials/states_pt3.html
14:44 farblue I’ve just been learning the same thing :)
14:46 scaryclam Yeah, this is what I've used in the past, though the environ state/module doesn't appear to be available :/
14:47 scaryclam I'm just getting "Jinja variable 'dict object' has no attribute 'environ.get'"
14:48 adelcast joined #salt
14:49 stooj joined #salt
14:53 Rumbles hmmmm I'm written a state which creates a symlink for a list of users in pillar. However, I wanted to make sure that salt only tries to create the symlink when the user home folder exists... I thouhgt the following would work but I get the error in the paste... https://paste.fedoraproject.org/367146/
14:54 knine joined #salt
14:54 Rumbles can anyone explain why I'm getting requiste not found when immediately after the state is shown as resulting in True?
14:55 AndreasLutro scaryclam: are you sure you're doing salt['environ.get'] ?
14:56 AndreasLutro Rumbles: your home-check state isn't a "module" state, it's a "file" state
14:56 Rumbles gah
14:56 Rumbles thanks AndreasLutro :/
14:57 scaryclam Pretty sure, yeah
14:57 Rumbles so obvious, couldn't se eit :)
14:57 scaryclam For example, I have an sls file with a single line:
14:57 scaryclam {{ salt['environ.get']('FOO') }}
14:58 scaryclam FOO="world" salt-call --local state.highstate
14:58 scaryclam is the example call
14:58 scaryclam Unless I'm just not seeing a typo :/
14:58 AndreasLutro state.highstate calls all your state files, not just the one you're testing
14:58 scaryclam Yeah, sorry, I have a top.sls with just that test sls :)
14:58 AndreasLutro do a grep for environ.get in your state tree
14:59 AndreasLutro your error message would be different if you only did salt['environ.get']
15:02 pipps joined #salt
15:03 tpaul joined #salt
15:04 mhug joined #salt
15:05 autofsckk joined #salt
15:06 scaryclam Oddly, {{ salt|pprint }} shows that environ.get is missing
15:09 MadHatter42 joined #salt
15:13 jerredbell1 joined #salt
15:14 hal58th joined #salt
15:15 majikman joined #salt
15:16 rihannon joined #salt
15:18 tristianc joined #salt
15:19 karlthane joined #salt
15:19 pipps joined #salt
15:20 gh34 joined #salt
15:21 mschiff joined #salt
15:21 mschiff joined #salt
15:22 zmalone joined #salt
15:27 dfinn1 joined #salt
15:29 dfinn joined #salt
15:30 Cadmus joined #salt
15:30 iceyao joined #salt
15:30 dfinn1 joined #salt
15:31 Cadmus Hello, I'm trying to use archive.extracted and I'm getting an error "ReadError: file could not be opened successfully". It happens very swiftly (5ms) and I can't figure out why, I can extract it using tar on my salt master.
15:31 brotatochip joined #salt
15:32 zzzirk joined #salt
15:32 Cadmus Ah, I've narrowed it down, python's tarfile module on the master can't , but my local python can.
15:36 Cadmus So the next question is... why?
15:38 gtmanfred cedwards: whats up?
15:39 Eugene Is the tarball xz/bzip2/gzip'ed?
15:41 Lee- joined #salt
15:41 Cadmus Eugene: tar.gz
15:43 Eugene Dunno, then. I've had dumb problems with xz in the past is why I asked.
15:44 Cadmus Heh, might be a bug in python's tarfile module, as it works in Python 3 on my machine but not Python 2 on the salt master.
15:44 brianfeister joined #salt
15:44 mapu joined #salt
15:45 pipps joined #salt
15:46 UtahDave joined #salt
15:47 Cadmus "Invalid header" if I coerce it with 'r:gz', but again not on Python 3, so probably just a poorly made tar, or a bug in the module, I'll repack it and see if that helps
15:51 edrocks joined #salt
15:53 Eugene That dos sound pretty suspicious
15:55 Cadmus I think the tar is a bit badly packed, and the older tarfile module can't handle it. Super annoying as this has taken a while to get to.
15:56 dezertol joined #salt
16:01 martoss joined #salt
16:02 bltmiller joined #salt
16:03 manji joined #salt
16:05 snergster joined #salt
16:05 pipps joined #salt
16:07 snergster joined #salt
16:10 brianfeister joined #salt
16:11 pipps joined #salt
16:14 rem5 joined #salt
16:14 totte joined #salt
16:15 brianfeister joined #salt
16:16 dgutu joined #salt
16:17 hasues joined #salt
16:18 dgutu joined #salt
16:19 ivanjaros joined #salt
16:19 coval3nce joined #salt
16:20 hasues left #salt
16:20 xenoxaos joined #salt
16:24 grainyhead joined #salt
16:30 writtenoff joined #salt
16:31 murrdoc joined #salt
16:33 averell joined #salt
16:34 devster31 joined #salt
16:35 sdemura joined #salt
16:35 devster31 hi guys, just discovered salt looking for an alternative to ansible, salt seems to fit quite nicely and is similar enough, one thing I don't understand though is how can I preseed a minion with a key and make the master accept it, besides salt-cloud I didn't find a clear solution
16:35 sdemura I'm playing with the vmware provider for salt cloud. When running "salt-cloud --list-images vcenter", it returns several templates I had stored in the base directory, but not in any folders. Is this by design?
16:36 devster31 in this page https://docs.saltstack.com/en/latest/topics/tutorials/preseed_key.html it tells me I can do it with an api call, but I can't understand why
16:36 Eugene I don't think the vmware driver has any concept of vsphere/vcenter folder objects
16:36 devster31 *how
16:37 Eugene Even in vcenter those mostly just provide organization anyway. Resource Pools are where the cool stuff happens.
16:37 rihannon joined #salt
16:37 sdemura I see an optional argument on for where to store the cloned template, but nothing for where to clone it from: https://docs.saltstack.com/en/latest/topics/cloud/vmware.html
16:38 sdemura I guess 'clonefrom:' doesn't support a path, just at template name.
16:38 Garo_ joined #salt
16:39 Eugene IIRC the VM name has to be unique anyway? It's been a while since I worked with that API
16:41 rob_benton joined #salt
16:41 rem5 joined #salt
16:43 druonysus joined #salt
16:46 sdemura i feel dumb. would help if my template was actually a template... ;-)
16:47 dfinn joined #salt
16:48 ninjada joined #salt
16:49 ageorgop joined #salt
16:51 scottk_ joined #salt
16:51 scottk_ Does anyone know of a salt formula for installing VmWare tools on Debian based machines?
16:51 * scaryclam facepalms as she realises the version of salt-call is *really* old
16:53 scaryclam AndreasLutro, thanks for trying to help earlier. It was the version of salt-call being too old  to contain the environ feature
17:00 UtahDave devster31: what kind of environment are you in?
17:02 edrocks joined #salt
17:03 devster31 UtahDave: I'm just trying salt out on vagrant machines
17:03 SpX joined #salt
17:03 devster31 so no specific requirements for no
17:04 UtahDave devster31: OK, I have a vagrant file example for that.
17:04 UtahDave Just a sec
17:04 UtahDave devster31: Here you go: https://github.com/UtahDave/salt-vagrant-demo
17:05 UtahDave You can see that the minion keys are already deployed in this example.
17:05 scoates joined #salt
17:05 UtahDave Basically, if a minion starts up and doesn't have a key, the minion will create its own key and then present them to the master.  If the minion already has a key, it will use that key to talk to the mater
17:05 UtahDave master
17:06 pipps joined #salt
17:06 quasiben joined #salt
17:06 UtahDave The master just has a directory with minions' public keys, so if a minion's key already exists in that directory, then it will use that for communication
17:07 pipps99 joined #salt
17:09 quasiben joined #salt
17:11 devster31 thanks, so basically you tell each machine the IPs of minion/master but how does the minion create the keys automatically? in the tutorials it says to create them on the master, move them to the clients and then accept on the master
17:11 supertom joined #salt
17:15 hal58th joined #salt
17:15 UtahDave devster31: generally I suggest not fiddling with Salt's keys unless you really need to.
17:15 UtahDave devster31: When you start the salt-minion daemon it will create it's own key if it doesn't exist
17:16 UtahDave Then on the salt master you'll see the key appear in the unaccepted list when you run       salt-key -L
17:16 UtahDave then you can accept the key using     salt-key -a <minion name>
17:16 UtahDave or      salt-key -A     if you want to accept all unaccepted minion keys.
17:18 devster31 ok, thanks, I'll start simple, it just seemed important
17:18 zmalone joined #salt
17:20 UtahDave Yeah, there's not a real need to preseed keys and get all fancy, generally
17:26 AndreasLutro odd statement, preseeding was one of our requirements when evaluating configuration management solutions
17:26 zzzirk_ joined #salt
17:26 rojem joined #salt
17:27 AndreasLutro if you don't pre-seed you could theoretically end up accepting the key of a rogue minion presenting the same ID as the one you're provisioning
17:28 UtahDave AndreasLutro: you can definitely preseed the keys if you want or need to.  It's just not terribly necessary and often new users get tripped up dealing with that.
17:28 quasiben joined #salt
17:28 UtahDave AndreasLutro: That could happen. You can manually verify key fingerprints or you can preseed keys if you need to make sure.
17:30 snergster or sign your masters keys or or
17:30 ageorgop joined #salt
17:32 mohae_ joined #salt
17:32 UtahDave Yeah, there's lots of options.
17:33 josuebrunel joined #salt
17:33 cnk joined #salt
17:39 hasues joined #salt
17:40 bltmiller I've always manually accepted keys from my master because at the scale I've operated at hasn't really required that much on-demand infrastructure change (e.g elasticity).
17:41 copelco joined #salt
17:42 kalessin joined #salt
17:46 hal58th joined #salt
17:47 DammitJim joined #salt
17:48 rojem joined #salt
17:48 scaryclam What's the best way of copying a file from one place in the file system to another?
17:49 manji joined #salt
17:50 rojem joined #salt
17:50 hasues left #salt
17:54 aw110f joined #salt
17:56 pfallenop joined #salt
17:56 zzzirk joined #salt
18:00 pipps joined #salt
18:02 cableninja joined #salt
18:03 cableninja morning guys. I'm having a strange issue, I cant seem to sort out. I keep getting 'No Top file or external nodes data matches found' on some systems. Some have randomly started working again, but most are still throwing that error. I've been digging on google for over an hour with no luck. Any one have any suggestions?
18:07 pipps joined #salt
18:07 ozborn joined #salt
18:11 Trauma joined #salt
18:12 snergster copying a file from one place to another? you want something other then file.copy ?
18:14 martoss joined #salt
18:14 Bryson joined #salt
18:14 Lionel_Debroux joined #salt
18:18 brianfeister joined #salt
18:18 dfinn joined #salt
18:27 devster31 what's the -P do in salt.bootstrap_options = "-P -c /tmp" UtahDave ?
18:28 jfindlay devster31: that means that dependencies can be installed with pip
18:30 jfindlay that is required for some platforms
18:30 jfindlay depending on the current state of support for those dependencies
18:33 totte joined #salt
18:34 devster31 thanks
18:34 devster31 does the script also install salt-cloud by default?
18:37 jfindlay I'm not sure
18:37 jfindlay if not, it might if you specify `-M`
18:37 teryx510 joined #salt
18:38 jfindlay the salt-* commands are really only simple wrappers around internal API points
18:38 cyborg-one joined #salt
18:39 cableninja I figured out my issue, it was strangely a version issue. the latest deb version on  http://repo.saltstack.com/apt/debian/8/amd64/latest were not working properly. After going back to the debian.saltstack url instead, everything worked fine again.
18:39 jfindlay cableninja: what deb version?
18:40 TyrfingMjolnir joined #salt
18:40 pjs joined #salt
18:42 cableninja 2015.8.8+ds-2 was the problematic version (even if I had matching minion versions) but when I downgraded the minions to 2015.5.3+ds-1~bpo8+1 (and eventually the master due to local salt minion) everything worked fine again
18:42 cableninja jfindlay, ^
18:42 jfindlay hmm
18:43 jfindlay cableninja: that was the version from the official debian repo and it wasn't working?
18:44 cableninja nope
18:44 cableninja I kept getting the 'No Top File' message
18:44 cableninja but on minions with the 5.5.3 version they were updating fine
18:44 quasiben joined #salt
18:51 GreatSnoopy joined #salt
18:58 alvinstarr joined #salt
18:59 skrobul joined #salt
19:01 DammitJim joined #salt
19:01 bltmille_ joined #salt
19:04 salty_solution joined #salt
19:05 tvinson UtahDave: I've been looking at the various wtmp/btmp beacon bugs (29468, 29451, 26294, 21203) and i'm wondering if it wouldn't make more sense for the binary format parsing portion to be in python-psutil. It's already most of the way there for psutil.users() and it's reading in the c headers that describe the format for a variety of os's.
19:07 UtahDave tvinson: nice. it would be great to squash those related bugs.
19:07 dmaiocchi joined #salt
19:08 SpX joined #salt
19:13 pipps joined #salt
19:16 rim-k joined #salt
19:20 fracklen joined #salt
19:25 jmedinar joined #salt
19:29 teryx510 joined #salt
19:29 jmedinar Is possible to execute a initial state and if it fails execute another state and if it succeed execute a third state all in a single SLS file ?
19:30 jmedinar like a jinja IF - ELSE using an exist status or something?
19:31 win_salt joined #salt
19:33 rylnd joined #salt
19:34 subsignal joined #salt
19:35 UtahDave jmedinar: there's an   onfail  requisite now that will allow you to do that.   you could also try using the orchestrate runner
19:35 rylnd hey guys, i was wondering if anyone has found a workaround for this bug so far: https://github.com/saltstack/salt/issues/28114. in general, should i think about switching to a RHEL derivative if i want to use saltstack? any recommendations for a saltstack newbie?
19:35 saltstackbot [#28114]title: debian_ip: IPv6 only interfaces not possible (again) | This is the same as issue #14530. It seems to have been re-introduced at some point....
19:36 jmedinar Thanks UtahDAve I am aware of the onfail requisite but what about do something when succeed ? will that just be a require ?
19:36 UtahDave jmedinar: yep. just a regular require
19:38 UtahDave rylnd: would you mind adding a comment providing the version of Salt you're seeing that issue on?
19:38 keimlink joined #salt
19:39 jmedinar thanks!
19:40 felskrone joined #salt
19:40 jmedinar To launch a reactor from a state will the cmd.run be the best option ?
19:41 mapu joined #salt
19:42 iggy !salt states.event
19:42 saltstackbot https://docs.saltstack.com/en/latest/ref/states/all/salt.states.event.html
19:42 iggy jmedinar: ^
19:43 punkoivan joined #salt
19:49 rylnd UtahDave: Done. I mentioned other issues as well. It looks like I really should use Centos 7 for now.
19:53 bltmiller joined #salt
19:54 dgutu joined #salt
19:54 rylnd UtahDave: I spend the last 3 or 4 hours building VMs, applying states, tearing down VMs and looking in the code what could cause the issue with the broken IPv6 network config. So far no luck unfortunately. I guess I'll take a break and try again later, maybe I'll find something.
19:55 DammitJim joined #salt
19:57 UtahDave sorry about that, rylnd.  Thanks for adding your comment. That helps triage
19:58 rylnd UtahDave: thanks! If i come up with anything I'll update the thread on Github
19:58 pcn Is anyone here using the redis returner?  I'm running into a surprising situation where the redis event bus doesn't seem to have all 3 events coming across for 3 separate async jobs I've thrown at it via libpepper
20:02 felskrone1 joined #salt
20:03 Muchoz joined #salt
20:04 pcn Hmm... looks like the redis-cli sees all the events, while the python library doesn't.  Time to go do some research, I guess.
20:04 felskrone joined #salt
20:07 zer0def joined #salt
20:07 brianfeister joined #salt
20:11 punkoivan joined #salt
20:12 tristianc joined #salt
20:13 punkoivan joined #salt
20:13 UtahDave pcn: not sure on that one.
20:14 berserk joined #salt
20:21 brianfeister joined #salt
20:24 wendall911 joined #salt
20:25 pcn I think I got it
20:26 pcn I need to already have started the listening function (I'm trying to yield async) otherwise all but the most recent event may have already passed me by
20:27 DammitJim joined #salt
20:27 sjmh joined #salt
20:28 TooLmaN joined #salt
20:29 edrocks joined #salt
20:29 dgutu joined #salt
20:29 pipps joined #salt
20:29 KingJ joined #salt
20:32 pcn So I need to call the first next() before submitting my job.
20:33 Edgan joined #salt
20:33 bltmiller joined #salt
20:40 rim-k joined #salt
20:44 Elsmorian joined #salt
20:51 hal58th joined #salt
20:51 punkoivan joined #salt
20:53 punkoivan joined #salt
20:55 punkoivan joined #salt
20:55 mpanett__ joined #salt
20:55 XenophonF joined #salt
20:56 mpanett__ joined #salt
20:57 punkoivan joined #salt
20:59 punkoivan joined #salt
21:00 nidr0x joined #salt
21:00 brianfeister joined #salt
21:00 pipps joined #salt
21:00 punkoivan joined #salt
21:02 punkoivan joined #salt
21:03 josuebrunel joined #salt
21:05 nidr0x joined #salt
21:07 bantone joined #salt
21:10 rem5 joined #salt
21:11 kaisero joined #salt
21:13 cableninja_ joined #salt
21:20 AdamSewell joined #salt
21:24 pipps joined #salt
21:26 tracphil joined #salt
21:31 stickmack joined #salt
21:32 bltmiller joined #salt
21:32 bltmiller joined #salt
21:33 pipps joined #salt
21:33 robawt joined #salt
21:37 cableninja__ joined #salt
21:38 nidr0x joined #salt
21:43 edrocks joined #salt
21:44 mattbillenstein joined #salt
21:46 mattbillenstein hi all -- can anyone recommend where I start for debugging slow master / minion communication?
21:46 mattbillenstein test.ping on a single host is taking 15s
21:49 pcn Off the top of my head, things to check: OOM killer messages in logs, PMTU-D is functioning properly, and other general network issues like older ciscos breaking on speed/duplex autodetect
21:53 * Heartsbane blames UtahDave.
21:54 sqwishy left #salt
21:54 UtahDave haha  :)
21:55 mattbillenstein nothing is crashing
21:55 mattbillenstein it always works
21:55 mattbillenstein just very slow
21:56 mattbillenstein master/minion on a test vm seems to work properly
21:56 UtahDave mattbillenstein: what kind of infrastructure are you on?
21:56 mattbillenstein Ubuntu / aws
21:56 mattbillenstein 4 minions connect to a single master
21:58 mattbillenstein looking at the master debug log
21:58 mattbillenstein seems to do nothing for most of this time
21:58 mattbillenstein then it publishes the command
21:58 mattbillenstein which immediately returns
21:58 mattbillenstein as I would expect I guess
21:58 mattbillenstein tried an strace running salt
21:59 mattbillenstein hard to decipher what's going on with that
22:00 zmalone joined #salt
22:00 mattbillenstein https://gist.github.com/mattbillenstein/a50abc3dc570c3844a40acdacf759889
22:01 mattbillenstein looking at _stamp -- that looks like < 300ms to me
22:03 UtahDave do you have any minions accepted by the master that are not currently on?
22:04 hasues joined #salt
22:06 mattbillenstein nope, they're all alive
22:06 mattbillenstein I'm pinging just one here, but if I ping them all, they all return successfully
22:08 hasues left #salt
22:13 nidr0x joined #salt
22:19 rem5 joined #salt
22:24 rem5 joined #salt
22:31 hal58th_ joined #salt
22:37 zenlot6 joined #salt
22:40 pipps joined #salt
22:40 jgelens joined #salt
22:48 crashmag joined #salt
22:48 brianfeister joined #salt
22:50 Muchoz joined #salt
22:51 UtahDave mattbillenstein: are your VM overloaded?
22:51 mattbillenstein no, mostly idle
22:52 mattbillenstein any idea why the master is trying to generate a minion id?
22:53 mattbillenstein it's calling generate_minion_id() from the logs
22:53 iggy what version of master?
22:56 mattbillenstein 2015.8.8.2
23:00 UtahDave mattbillenstein: how did you install Salt on both master and minions?
23:01 mattbillenstein pip into a virtualenv
23:01 fracklen joined #salt
23:02 keimlink_ joined #salt
23:07 zmalone joined #salt
23:09 bltmiller joined #salt
23:11 ok_ joined #salt
23:16 ninjada joined #salt
23:20 ninjada joined #salt
23:30 fxhp joined #salt
23:30 quasiben joined #salt
23:30 keimlink joined #salt
23:34 nidr0x joined #salt
23:36 quasiben joined #salt
23:37 * iggy backs away slowly
23:37 ageorgop joined #salt
23:41 majikman joined #salt
23:41 mattbillenstein is that bad?
23:42 iggy just not familiar with it... I generally just go with packages
23:42 ok__ joined #salt
23:42 estahn joined #salt
23:42 majikman i've configured my salt mine to cache network.ip_addrs. when i retrieve it using jinja, how do i retrieve just the ip address and not the name of the host? curently, it gives me { "hostname": ["192.168.1.1"] }
23:43 sjmh sigh.  i feel like i ask this like once every 6 months, but what the heck is the difference between the master config options extension_dir and module_dirs ( besides the fact that one is a single directory, and one takes multiple.. ) ?
23:44 iggy majikman: loop over the dict? dict.keys() ? something else?
23:44 west575 joined #salt
23:45 majikman oh... there isn't someway to specify it like salt['mine.get']('*', 'network.ip_addrs[0]')?
23:45 zmalone left #salt
23:46 majikman does that also mean it's not possible from the command line?
23:46 mapu joined #salt
23:46 iggy not that I know of
23:47 majikman ok, thanks
23:48 DammitJim joined #salt
23:55 sjmh mattbillenstein - i'd suggest trying to figure out where the slowness is by adding some logging to some of the modules.  I had a similar issue a few weeks ago that turned out to be a DNS-related issue
23:55 sjmh was causing salt-run jobs.lookup_jid to be super slow
23:56 mattbillenstein yeah, sorta seems like I'm hitting some 15s timeout or something somewhere
23:56 mattbillenstein do you remember what specific issue you had?
23:57 sjmh salt-run jobs.lookup_jid was taking ~52 seconds to run.  Turns out it was caused by 'myhostname' nss switch module and the fact that we disable ipv6 on our systems.
23:57 ninjada joined #salt
23:57 sjmh the master tries to create a 'master minion' and it runs grains when it does so
23:57 sjmh and it was loading the fqdn_ipv6 module, which was trying to do a AAAA lookup on our FQDN
23:58 mattbillenstein hmm
23:58 mattbillenstein I see
23:58 sjmh but I finally traced the issue down by just putting real simple log statements everywhere in the path
23:58 sjmh and then seeing where the delay was

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