Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-11-12

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

All times shown according to UTC.

Time Nick Message
00:00 pipps joined #salt
00:07 mkoistinen Hi all. I'm using 0.17.1. When I attempt to use a iptables.append function, the result is that it appends --__env__ base --__sls__ [myslsname] to the end of the rule, and my linux (Ubuntu 12.04) is choking on those bits. I realize those bits are there to help manage states between runs, but its breaking stuff. What am I doing wrong?
00:08 pipps_ joined #salt
00:08 NV jcockhren: so about that pillar backed scheduler? :P
00:09 jcockhren https://gist.github.com/jcockhren/7d33ee39479e9a3ee6f4
00:10 jcockhren NV, cdcalef ^
00:10 jcockhren apparently I can't format text in the gist description so bare with me
00:11 jcockhren s/me/it b/c I'm right, gist is just wrong
00:11 zandy joined #salt
00:13 jkleckner joined #salt
00:15 jcockhren All of the posted gist is possible with yaml states except the datestamped file names. But probably someone better then me a jinja can figure that out. But I do have yaml versions of the ststes as well
00:15 jcockhren NV, cdcalef ^
00:16 jcockhren I'm thinking this can be a formula
00:16 jcockhren (once supporting multiple DBs and such)
00:17 jcockhren there's nothing saying formulas have to be about installing stuff
00:20 jcockhren pillar.clienta is what it is b/c it also will serve as the basis for salt assisted deployments
00:20 whiteinge ooh, good formula :)
00:21 mkoistinen So... is it me, or does managing IPTables rules just not work?
00:22 mkoistinen Opening a ticket...
00:23 UtahDave that's cool, jcockhren!
00:26 scott_w joined #salt
00:28 carlos_ joined #salt
00:30 mkoistinen https://github.com/saltstack/salt/issues/8419
00:30 UtahDave thanks, mkoistinen!
00:31 mkoistinen Thank you, Dave. And, I'm sorry for my poorly worded comment earlier =/
00:31 UtahDave ?  which one?
00:32 mkoistinen =) You're very gracious, thank you.
00:32 elfixit joined #salt
00:32 UtahDave :)
00:37 higgs001 joined #salt
00:45 pipps joined #salt
00:52 alunduil joined #salt
00:54 jkleckner joined #salt
00:54 mdasilva joined #salt
00:59 ipmb joined #salt
01:05 amckinley joined #salt
01:06 pipps_ joined #salt
01:09 jesusaurus hows 0.17.2 coming along?
01:11 malinoff joined #salt
01:16 UtahDave jesusaurus: pretty good!  I think they're wrapping up some final testing.
01:17 pipps joined #salt
01:17 jesusaurus awesome
01:18 UtahDave jesusaurus: anything in particular you're looking for in that release?
01:18 erasmas left #salt
01:18 heewa Anyone ever get an m2crypt error from their minion like this? RSAError: oaep decoding error
01:18 UtahDave what version of Salt are you on, heewa?
01:19 heewa v0.17.1, installed by pip from the github repo using that tag
01:19 carlos__ joined #salt
01:19 jesusaurus UtahDave: other than my rabbitmq changes, I just got bit by https://github.com/saltstack/salt/pull/8126
01:19 UtahDave heewa: are other commands working? Or do they all respond that way?
01:20 UtahDave jesusaurus: ah, yeah.
01:20 heewa UtahDave: Well, if I try to run salt again, it says the master's key is invalid. So I delete the key and it gives me that RSAError again.
01:22 amckinley joined #salt
01:22 zandy joined #salt
01:23 UtahDave is m2crypto installed?
01:24 heewa UtahDave: Yes, using apt-get
01:24 mdasilva joined #salt
01:25 maletor joined #salt
01:25 teebes joined #salt
01:26 UtahDave heewa: did you install all the deps with apt first?
01:26 higgs001 joined #salt
01:27 scott_w joined #salt
01:27 heewa UtahDave: exception stack trace: http://pastebin.com/ERmNfVRA
01:28 heewa UtahDave: apt-get install -y python-m2crypto python-pip gcc python-dev git && pip install pyzmq PyYAML pycrypto msgpack-python jinja2 psutil && pip install -e 'git+https://github.com/saltstack/salt.git@v0.17.1#egg=salt-minion'
01:28 mdasilva1 joined #salt
01:29 Gifflen joined #salt
01:31 bemehow joined #salt
01:32 mwillhite joined #salt
01:33 hvn joined #salt
01:34 zandy joined #salt
01:35 JulianGindi joined #salt
01:35 hvn left #salt
01:35 mkoistinen Attempting to modify a file via SLS with: https://gist.github.com/anonymous/155db7787072c0bdfa0f
01:36 UtahDave heewa: sorry, got pulled into a quick meeting
01:36 mkoistinen Results in: Comment:   Specified file config is not an absolute path
01:36 UtahDave heewa: did you add the salt ppa?
01:36 mkoistinen How do you make '/etc/mysql/my.cnf' more 'absolute'?
01:36 heewa UtahDave: Dude, I'm glad for the dredges of your attention, no apologies needed.
01:36 bemehow joined #salt
01:37 UtahDave mkoistinen: change "path"  to "name"
01:37 heewa UtahDave: I'm trying to install salt from pip, not from apt using the ppa. There were problems with the PPA, and now I'm trying to run it off pip. It works sometimes, but my new boxes are getting this.
01:37 mkoistinen ok
01:38 UtahDave heewa: Install from the ppa, then apt-get remove salt-minion
01:38 UtahDave heewa: then install with pip
01:38 bemehow joined #salt
01:38 UtahDave heewa: I think the ppa distributes a patched version of m2crypto
01:38 londo joined #salt
01:39 mkoistinen Works, thanks!
01:39 UtahDave you're welcome, mkoistinen!
01:39 ctdawe left #salt
01:40 heewa UtahDave: Jesus. Ok, wow, thanks, I'll try that.
01:40 ipmb joined #salt
01:41 UtahDave heewa: are you trying to skip a specific bug in 0.17.1?
01:43 amckinley joined #salt
01:43 heewa UtahDave: Trying to stop being dependent on other people's maintained repos. Burned us with other packages in the past. And at least once I've wanted to patch something in salt without waiting for a new release.
01:44 UtahDave heewa: have you tried building your own packages and hosting your own repo?
01:44 jkleckner joined #salt
01:45 heewa UtahDave: That's a great idea! Ok, I'll try that if I get some time after this mini-crisis (unrelated to salt) is over.
01:52 olliewalsh left #salt
01:53 mdasilva joined #salt
01:57 ipmb joined #salt
02:03 cowmix joined #salt
02:05 isomorphic joined #salt
02:06 melinath joined #salt
02:07 higgs001 joined #salt
02:08 heewa UtahDave: Still getting the mcrypto oaep error. I tried installing from ppa & uninstalling, same error. Then I tried installing from ppa and then installing from pip without uninstalling apt, same error.
02:12 kryptt joined #salt
02:12 ipmb joined #salt
02:12 kryptt How would I go about debugging render errors on a state file I have?
02:12 kryptt I'm getting:
02:12 kryptt render error: Undefined jinja variable; line 372 in template
02:13 kryptt but don't see anything obviously wrong
02:13 NV kryptt: log into the minion and use salt-call to execute the state
02:13 NV ie, salt-call state.sls mystate
02:13 NV (or state.highstate)
02:13 heewa kryptt: If it's a variable from grain, you can try 'salt call grains.items' on the box to see what vars are available
02:13 NV it prints far more debug info than using the salt command remotely
02:14 NV combine with grains.items and pillar.items if grains and/or pillars are used for further info as well as kermit suggested (although i assume he meant salt-call grains.items, or salt <host> grains.items)
02:14 kryptt hmm, I still get the same error
02:15 kryptt local:
02:15 kryptt Data failed to compile:
02:15 kryptt ----------
02:15 kryptt Rendering SLS artifactory failed, render error: Undefined jinja variable; line 372 in template
02:15 kryptt and then a copy of my state file
02:15 mkoistinen Anyone know what this means: State cron.set_job found in sls [my sls file] is unavailable
02:16 kryptt it should be a variable from pillar, but it's there, and I'm guessing get can supply a defualt value
02:16 teebes joined #salt
02:17 heewa kryptt: Yea, you can do default values {{ pillar.get('thing', 'oh no!') }}, but to debug why the pillar var isn't there, make sure you apply the pillar sls to the salt role, the way you'd apply a regular state file to a box, but in the pillar's top.sls file.
02:18 kryptt http://pastebin.com/AmR0s1nB
02:18 kryptt that's the state file
02:18 kryptt and pillar.item shows the artifactory entry.
02:19 kotulag joined #salt
02:20 kotulag Hello, I cannot seem to get non-verbose output, even with satet_verbose: False; state_output: mixed on both master and minion.
02:24 amckinley joined #salt
02:24 franky joined #salt
02:26 maj888 joined #salt
02:26 UtahDave heewa: could you open an issue on that?  I haven't seen that before.
02:27 UtahDave mkoistinen: you're probably using an execution module function and not a state module function
02:27 scott_w joined #salt
02:27 UtahDave kotulag: did you restart the master after making that change?
02:28 xl1 joined #salt
02:29 kryptt UtahDave do you see anything that is silly in my pastebin (http://pastebin.com/AmR0s1nB) ?
02:29 mkoistinen Hmmm, I'm just trying to set a cron task, wouldn't I use an execution module for that? Here's my sls: https://gist.github.com/anonymous/47fde6e6187da93174c8
02:30 UtahDave mkoistinen: in your sls files you normally will use states, not modules.   http://docs.saltstack.com/ref/states/all/salt.states.cron.html#module-salt.states.cron
02:31 UtahDave kryptt: what's the error you're seeing?
02:31 gamingrobot_ joined #salt
02:31 pcarrier joined #salt
02:31 mkoistinen UtahDave: Thanks. (again *sigh*)
02:32 akitada joined #salt
02:32 kryptt Rendering SLS artifactory failed, render error: Undefined jinja variable; line 372 in template
02:33 kryptt the file only has 42 lines though
02:33 UtahDave ah, yeah. I think we just pinpointed that error.  Let me check the mailing list
02:33 xerxas joined #salt
02:34 toastedpenguin joined #salt
02:35 UtahDave kryptt: yeah, here it is:  https://github.com/saltstack/salt/issues/8418
02:36 kryptt ok, that sorta helps
02:36 UtahDave kryptt: do dictionaries have the .get method in jinja?
02:36 kryptt ah
02:36 kryptt no
02:37 kryptt I sorta assumed it was a property of all dicts (not a python dev)
02:38 UtahDave those are jinja dicts. The confusing thing for me about jinja is that sometimes is follows python behavior and sometimes it doesn't.
02:39 kryptt hmm, cool, this helps, is there a way to specify a default if the pillar is underfined for example?
02:41 kryptt e.g. without using {% if version in art %} etc
02:45 whiteinge kryptt: there's a "piller" (should be "pillar") at the top. could that be it?
02:47 kryptt @whiteinge, no.. it's that the art dict that is retreived from the pillar doesn't have a get method associated with it
02:47 kotulag @UtahDave - Yes.  Multiple restarts of both minion and master.
02:48 kryptt hmmm, UtahDave, I might be able to just execute pillar.get('artifactory.home', 'opt/artifactory') right?
02:48 whiteinge you can access methods on python objects via jinja. that .get() syntax is valid
02:48 whiteinge kryptt: "piller" is a typo
02:48 fivethreeo joined #salt
02:48 kryptt lol, thanks!
02:49 whiteinge :)
02:49 honestly whiteinge!
02:50 whiteinge honestly!
02:50 honestly check this out: https://github.com/duk3luk3/template-formula https://github.com/duk3luk3/openvpn-client-formula
02:50 whiteinge finally a passing in IRC that spans less than three hours :)
02:51 whiteinge ooh, neat
02:51 honestly :D
02:52 whiteinge template-formula is a really good idea
02:52 honestly (:
02:53 honestly feel free to fork / import them into saltstack-formulas
02:53 whiteinge the openvpn-client formula is also killer
02:54 whiteinge honestly: i'll add you to the github org and you can transfer them yourself, if that works for you
02:54 honestly ok :D
02:55 whiteinge you are added, sir!
02:56 honestly so I just use the transfer feature in the danger zone?
02:56 whiteinge yeah
02:56 JulianGindi joined #salt
02:57 honestly why does it say repositories: 0 when I switch to the saltstack-formulas context?
02:58 honestly oh
02:58 honestly because I don't get any kind of access, lol
02:58 honestly well, template-formula made it, apparently
02:59 honestly gonna transfer openvpn-client-formula too
02:59 whiteinge it did. woot!
03:00 maletor joined #salt
03:00 honestly I don't know if you want https://github.com/duk3luk3/dirty-user-sync-formula
03:00 honestly it's dirty. real dirty.
03:00 malinoff Hi all
03:00 malinoff whiteinge, Do you profile salt?
03:01 whiteinge honestly: haha. that looks very useful
03:02 honestly whiteinge: imagine a node crashing hard while synching /etc/passwd
03:02 honestly whiteinge: you know have a corrupted, potentially unreadble /etc/passwd
03:02 whiteinge yeah...
03:02 whiteinge malinoff: we do from time to time, yes
03:02 honestly dirty (:
03:02 honestly don't tell anyone, but I'm using it in production right now
03:03 maletor left #salt
03:03 malinoff whiteinge, I got such thing: http://pastebin.com/69a5Y4YL
03:04 malinoff Looks like pyyaml slows salt as hell
03:04 lemao joined #salt
03:05 austin987 joined #salt
03:05 malinoff (reader.py and parser.py are pyyaml modules)
03:06 malinoff Should I open an issue?
03:07 kryptt Hey, the typo fix on pillar was enough for the rest of the file to render correctly, thanks again whiteinge
03:07 whiteinge honestly: https://github.com/saltstack/salt/commit/5476e7c0668d0ed6462a8fb39d93b030b1f8ca97
03:07 whiteinge honestly: thank you very much!
03:08 whiteinge malinoff: the pyyaml parser is very slow indeed. that pastebin is...wow
03:09 whiteinge malinoff: yes, please open a ticket
03:10 malinoff whiteinge, Yep. That's what I said when I saw that. And after a small investigation, I figured out that we have no alternative - I mean, there are no other bindings to libyaml
03:10 maj888 malinoff: Is your PyYAML linked to libyaml?
03:10 malinoff maj888, Of course, there is the libyaml-0-2 dependency in python-yaml package
03:10 * whiteinge has to run afk
03:11 whiteinge honestly: thanks again for transfering those repos. nice work!
03:11 zach Has anyone created an SVN backend yet?
03:12 zach We utilize SVN and not GIT (unfortunately) at work
03:12 * honestly bows
03:12 zach honestly: you've created one?
03:17 honestly no, sorry
03:21 ravibhure joined #salt
03:22 UtahDave zach: so far we have gitfs, hgfs and s3fs.
03:22 zach UtahDave: yeah. Is there any possibility to add svn support? Would be nice to be able to pull configs via svn as well
03:23 joehh UtahDave: just checked, ubuntu ppa does not appear to have patched m2crypto - may have had so for releases older than lucid, but does not appear to be there now
03:24 malinoff Opened https://github.com/saltstack/salt/issues/8424
03:24 UtahDave joehh: ah, ok. Thanks for letting me know!
03:24 UtahDave zach: yeah, we'd love to have an  svnfs
03:25 redondos joined #salt
03:25 redondos joined #salt
03:27 maj888 \quit
03:28 scott_w joined #salt
03:32 malinoff whiteinge, Well, it seems to be the bottleneck of salt... Each time I run 'salt ids test.ping', pyyaml takes 80%-99% of time
03:33 UtahDave malinoff: hm. That's great info.  Thanks!
03:33 malinoff UtahDave, check my issue: https://github.com/saltstack/salt/issues/8424
03:34 UtahDave yeah, I just read it
03:35 malinoff Strange thing that it takes so much time even on that simple test.ping
03:35 malinoff without transferring and rendering huge yaml files
03:35 steveoliver 'Undefined jinja variable' question — http://paste2.org/O4NNKXx2 I'm wondering why the /second/ {{ site }} variable is being marked as the undefined variable, and not the first.
03:35 UtahDave malinoff: do you get a stacktrace when running this from the cli?    https://github.com/saltstack/salt/pull/8422
03:39 malinoff UtahDave, I don't understand. That issue relates to windows, in my examples all minions are debian-like systems. But anyway, I think the answer is no
03:40 UtahDave malinoff: Oh, sorry. I didn't get the right clipboard
03:40 UtahDave python -c "from yaml import CLoader"
03:40 UtahDave try that.
03:41 Baribal_ joined #salt
03:41 malinoff UtahDave, it works, both on the master and on a minion
03:42 malinoff I mean, there is no 'ImportError', if you were asking about that
03:42 UtahDave ok, that's what I was looking for
03:42 * Gareth cleans up some code prior to a pull request
03:42 UtahDave that means libyaml is installed and working on your system.
03:42 UtahDave Gareth: nice!
03:43 malinoff UtahDave, Of course. I try to use package managers as much as possible. python-yaml has the libyaml-0-2 dependency, so yes, it is installed and it is binded to pyyaml
03:46 malinoff UtahDave, btw, there is no percentage difference between running test.ping on local VMs or AWS USA instances (I live in Russia, so the ping is significant)
03:46 malinoff So I think this problem is not related to network communications
03:47 UtahDave malinoff: I agree. It's definitley a pyyaml thing.  My initial googling seems to indicate that there's not a whole lot to be done since yaml is a somewhat complicated spec
03:47 UtahDave We'll have to have some engineers dig into that tomorrow
03:47 UtahDave and see what we can do about that.
03:47 austin987 joined #salt
03:48 malinoff UtahDave, nice to hear that
03:49 utahcon I think it is odd that the docs show the dev branch docs, and not stable. Could really confuse people.
03:49 malinoff utahcon, https://groups.google.com/forum/#!topic/salt-users/egbFq3hnSRw
03:50 utahcon and solved.
03:50 utahcon thanks!
03:50 utahcon I feel better now
03:50 malinoff UtahDave, if you will need any help, or more examples, feel free to ask about that in that issue. I'll do everything you need
03:56 austin987 joined #salt
03:57 kotulag sorry to be such a N00b, (I am a convert from Puppet), but I still cannot seem to get the output non-verbose, even with master/minion  state_verbose: False and state_output: terse
03:58 jhulten joined #salt
04:10 zach_ joined #salt
04:11 steveoliver 'Undefined jinja variable' question — http://paste2.org/O4NNKXx2 I'm wondering why the /second/ {{ site }} variable is being marked as the undefined variable, and not the first.
04:14 forrest joined #salt
04:26 malinoff kotulag, What do you want to achieve?
04:27 kotulag when I run sudo salt '*' state.highstate, I get multiline output.  I would like to get single-line output for success, and the multi-line output for Fails...
04:28 kotulag right now, I get multiline output for everything, including successes/no-ops.
04:28 forrest kotulag, as far as I'm aware that isn't currently supported. You could try running through a fail to see if the value returned in the shell is different than a success, and if so simply pipe the output to a file, then if the exit is an error, print to the terminal, otherwise trash the file.
04:29 scott_w joined #salt
04:29 kotulag hmm.  I thought that state_output would print a single line unless there was a failure...
04:30 kotulag (state_output: terse)
04:30 forrest that's per state I thought
04:30 anuvrat joined #salt
04:30 forrest if you're looking for single line if success, multi for failure just use the mixed option
04:31 kotulag I tried mixed with no difference in behavior
04:31 forrest oh ok
04:33 jacksontj joined #salt
04:36 jacksontj_ joined #salt
04:42 malinoff kotulag, write your own parsed or use linux pipes. Salt is not a silver bullet
04:42 malinoff parser*
04:47 malinoff kotulag, http://docs.saltstack.com/ref/output/all/index.html
04:49 junedm joined #salt
04:50 kotulag malinoff: thanks for the reference.
04:51 malinoff kotulag, https://github.com/saltstack/salt/tree/develop/salt/output - here the source code live
04:55 kotulag thanks for the reference. perhaps beyond my current skills, but it's time to roll up my sleeves... :-)
04:56 jcockhren forrest: yo
04:56 forrest jcockhren, hi
04:57 jcockhren https://gist.github.com/jcockhren/7d33ee39479e9a3ee6f4
04:57 jcockhren forrest: ^
04:58 forrest nice
05:02 jcockhren forrest: now on to salt assisted site deployments
05:02 forrest heh
05:04 jcockhren forrest: I'm thinking sites have, by definition, the state of deployed
05:04 jcockhren "deployed"
05:05 loo joined #salt
05:05 lemao joined #salt
05:06 jcockhren pillar.clienta in the gist contains most of the needed info for a deployment = code from a repo + a DB dump
05:06 jcockhren in the zero case, the DB only needs to exist
05:07 jcockhren oh. and it also needs webserver config file
05:07 loo hi anyone here runs minion in freebsd ?
05:08 loo seems like my service 'ipfw' always return false even if its there
05:08 pipps_ joined #salt
05:09 jhulten joined #salt
05:20 druonysus joined #salt
05:20 druonysus joined #salt
05:22 fllr joined #salt
05:22 anti joined #salt
05:22 fllr Is there a 'make' state for salt? I'm trying to install a library that doesn't distribute using apt.
05:28 pdayton joined #salt
05:28 anti any pointers for using salt.runner.RunnerClient?  My attempts have failed at querying job information on RunnerClient instantiation
05:30 scott_w joined #salt
05:44 jimallman joined #salt
05:57 Destro joined #salt
05:58 pipps joined #salt
06:02 sirtaj joined #salt
06:07 Pepper_ joined #salt
06:07 jalbretsen joined #salt
06:12 pdayton joined #salt
06:12 Pepper_ left #salt
06:15 krandun joined #salt
06:15 middleman_ joined #salt
06:21 Kyle_ joined #salt
06:26 middleman_ joined #salt
06:30 scott_w joined #salt
06:35 Katafalkas joined #salt
06:44 hvn joined #salt
06:44 hvn left #salt
06:49 dvogt joined #salt
06:53 Destro left #salt
07:06 lemao joined #salt
07:11 IJNX joined #salt
07:11 loo hello, is there any way to overwrite how service.running works ? I want to ensure running ipfw on freebsd, but its a kernel option and not a daemon. So you cant use SIG to search PS
07:11 vkurup joined #salt
07:15 carlos joined #salt
07:23 malinoff loo, is this really a service? I'm not very familiar with freebsd. How would you check by your hands?
07:26 hvn joined #salt
07:27 hvn left #salt
07:31 scott_w joined #salt
07:35 slav0nic joined #salt
07:35 slav0nic joined #salt
07:37 juasiepo joined #salt
07:42 giantlock_ joined #salt
07:43 thelorax123 joined #salt
07:49 balboah joined #salt
07:53 anitak joined #salt
07:54 IJNX why my requisites are not found but I have them defined in the same file?
07:57 harobed joined #salt
07:58 IJNX Here are the errors and the sls file: http://pastebin.com/K5wUG0KP
07:58 loo joined #salt
07:58 kadel joined #salt
07:58 loo hi malinoff
07:58 IJNX Are the lines 136-137 correct?
08:00 redondos joined #salt
08:00 redondos joined #salt
08:02 IJNX How do you require an id to be executed before another one?
08:03 IJNX it seems the problem is that my stateid doesn't contain any cmd.run tags and thus it cannot be 'watched' with cmd: <state>
08:03 malinoff IJNX, your 'couchdb-users' stateID is not related to 'cmd'
08:03 malinoff try to use group: couchdb-users
08:03 malinoff or user: couchdb-users
08:04 matanya joined #salt
08:04 IJNX what does that wait for then?
08:04 IJNX so there is no state: <stateid>
08:05 malinoff I mean, 'couchdb-users' state. "couchdb-users" is it's state ID
08:05 malinoff You have two states under that state ID: user and group
08:05 IJNX yes, but how do you wait for a state?
08:05 malinoff - require:
08:05 IJNX cmd waits for exec, file for file, …
08:05 malinoff - user: couchdb-users
08:06 malinoff - group: couchdb-users
08:06 malinoff I think that way
08:06 IJNX ok, thanks it seems to work like that. Not very intuitive tho...
08:07 IJNX it would make sense to offer keyword for stateid for watch
08:07 malinoff require syntax is: - <state>: <state-id>
08:08 malinoff cmd, pkg, git - are states
08:08 malinoff couchdb-users isn't a state actually
08:08 malinoff It's state-id that unite two states
08:08 IJNX uh
08:08 IJNX ok
08:09 malinoff Yeah, it's a bit complicated
08:10 IJNX even the debug log presents it as a single 'stateid' and doesn't split it into two copies
08:10 IJNX "rendered data" log
08:11 malinoff Well, that states can't run separately, so yes, this is the expected behavior
08:12 IJNX but is the order of the execution of sub-state-ids same as in file?
08:12 IJNX or do they run in parallel?
08:13 IJNX (user, group)
08:14 juasiepo joined #salt
08:14 malinoff IJNX, I'm not sure (haven't tested), but I think yes, they preserve the order under the same state-id
08:14 IJNX they are stored in OrderedDict but if there are many threads executing them on minion, then the order could be different. And thus I need to handle the ordering with the separate states and requires
08:15 malinoff What version of salt do you use?
08:16 IJNX 0.17.1
08:16 efixit joined #salt
08:17 malinoff You can set state_auto_order: True in /etc/salt/master, and your state-ids will be executed in the same order they appear in .sls files
08:18 IJNX but that's not the default/preferred option?
08:19 malinoff It's not default, but I found this very comfortable. I'm trying to keep my state files small as possible, and such ordering is quite enough for me
08:23 bezaban and another wonderful day in the world of salt dawns :)
08:24 IJNX how do I deny salt from trying to create directory for new user?
08:25 IJNX linux useradd doesn't create it at all, but salt seems to work differently
08:25 EugeneKay You mean when doing user.exists? Set home: /dev/null ;-)
08:25 EugeneKay s/exists/present/
08:26 IJNX user doesn't exist, but the home does
08:26 IJNX http://pastebin.com/rtyUcYPQ
08:26 EugeneKay Ah, you're trying to create a user with a preexisting homedir
08:27 IJNX yep, couchdb user
08:27 EugeneKay I've had similar adventures in lolwat when trying to create a user with conflicting UIDs, or set the UID of an existing one.
08:27 EugeneKay I think there's room for improvement in the state, but I'm too lazy to do it myself ;-)
08:28 EugeneKay You /could/ file.absent the dir, but only if the user does not exist.
08:28 IJNX linux man says 'The directory HOME_DIR does not have to exist but will not be created if it is missing.'
08:29 IJNX salt adds the -m to command
08:29 IJNX == --create-home
08:30 EugeneKay That would do it.
08:30 EugeneKay Sounds like issue time ;-)
08:30 IJNX hmm, so there isn't any option to disable this?
08:30 EugeneKay Not AFAIK, but I am not a Salt expert either
08:31 IJNX so instead I need to create another state to run this myself… *sigh*
08:33 EugeneKay I'd just file a bug and do it manually
08:34 EugeneKay (actually, what I did in my environment was toss out all of my pre-Salt machines)
08:34 malinoff EugeneKay, check this out: https://github.com/saltstack/salt/issues/8424
08:35 malinoff That's what I discovered today
08:35 EugeneKay What do you want me to do?
08:35 malinoff Just look and say 'wow' :)
08:35 EugeneKay I'll stick with "Duh?"
08:36 malinoff It's fine too
08:36 EugeneKay PyYAML is slow.
08:36 EugeneKay But the good news is Salt isn't a HPC app, so who cares
08:37 EugeneKay Anecdotally the Saltstack guys have customers with 4-digits worth of Minions, so it's not like it doesn't scale either
08:37 EugeneKay Like most modern software development, you can spend thousands of man-hours eking out the last 5% of performance.... or you can spend an extra $50 to get a quad-core CPU.
08:38 EugeneKay (or tell your users to do the same)
08:39 redondos joined #salt
08:39 redondos joined #salt
08:39 malinoff The problem is not related to performance issued, for me it influence on timeouts
08:39 malinoff issues*
08:39 malinoff And it's not very comfortable
08:41 malinoff Like "Hm, which timeout should I set this time... Okay, I'll try -t 15. Damn, I need to increase that" and so on
08:42 EugeneKay Ah, that.
08:42 ddv joined #salt
08:42 EugeneKay It takes even longer when you run pkg.installed states ;-)
08:44 malinoff Yes. We have instances in different amazon regions (of course, not 4k but around 100), and I tried to hand-pick that timeouts
08:44 malinoff And sadly that salt spends 80% of time to pyyaml calls
08:44 malinoff I'm 100% it can be way faster
08:44 malinoff 100% sure*
08:44 EugeneKay Patches welcome ;-)
08:45 malinoff As you seen, I opened the issue
08:45 IJNX how can I just execute something if user is not created?
08:45 EugeneKay That's a good question, and one I will dodge by going to sleep.
08:46 IJNX :)
08:47 malinoff {% if not __salt__['user.present']('your_user') %} # do smth {% endif %}
08:48 malinoff oops
08:48 malinoff not user.present
08:49 ajw0100 joined #salt
08:50 IJNX this could work also: getent passwd couchdb || adduser …
08:50 malinoff dunno, probably this http://docs.saltstack.com/ref/modules/all/salt.modules.useradd.html#module-salt.modules.useradd will help
08:53 chutz joined #salt
08:54 bezaban the signature lists a createhome= variable
08:54 bezaban not sure how you would go about setting it, as it's not detailed
08:54 bezaban (and I am just getting to know salt)
08:55 IJNX so createhome: false might work?
08:56 crashmag joined #salt
08:56 natewalck joined #salt
08:56 halfss joined #salt
08:57 flebel joined #salt
08:57 ajw0100 joined #salt
08:59 g3cko joined #salt
09:00 IJNX this has it: https://salt.readthedocs.org/en/latest/ref/states/all/salt.states.user.html?highlight=user.present#salt.states.user.present
09:04 IJNX testing cloud deployment scripts is really slow process...
09:05 shuveb joined #salt
09:06 IJNX ~10 minutes per round when running master-minion inside two virtualbox machine on my laptop.
09:07 shuveb What the best way to install something from source code using the regular untar/configure/make/make install commands on many machines with salt?
09:08 IJNX shuveb: that depends if those machines are identical or not
09:08 shuveb IJNX: they are
09:09 IJNX shuveb: why not then compile once and then just copy to other targets
09:09 shuveb IJNX: that's a good idea as well
09:10 shuveb IJNX: In general is there a way to change to a directory before executing a script like say, ./configure?
09:10 IJNX cwd: /<path
09:10 shuveb IJNX: ah, thanks. Didn't know it could remember states between calls, like current working directory. That's pretty cool.
09:11 IJNX I think that's not between calls, but inside one stateid you need to put cwd
09:12 shuveb IJNX: OK, let me confess. I'm new to Salt and haven't gotten around to learn about states
09:12 shuveb Now, I think is a good time to do that :)
09:13 IJNX shuveb: so are we all. see my node script: https://gist.github.com/ilkkanisula/7369980
09:13 shuveb IJNX: thanks, dude. I will!
09:14 IJNX but I replaced that with this:
09:14 sebgoa joined #salt
09:14 IJNX https://github.com/isaacs/nave/blob/master/nave.sh
09:15 halfss_ joined #salt
09:20 shuveb IJNX: thanks
09:25 yota joined #salt
09:28 shuss joined #salt
09:29 thelorax123 joined #salt
09:30 Damoun joined #salt
09:35 ggoZ joined #salt
09:39 scott_w joined #salt
09:42 ajw0100 joined #salt
09:56 elfixit joined #salt
10:02 crashmag joined #salt
10:02 scott_w joined #salt
10:04 lemao joined #salt
10:13 thelorax123 joined #salt
10:34 matanya joined #salt
10:36 scott_w joined #salt
10:47 scott_w joined #salt
10:48 hvn joined #salt
10:49 hvn left #salt
10:51 Katafalk_ joined #salt
10:52 scott_w joined #salt
10:52 whiskybar joined #salt
11:02 Destro joined #salt
11:04 logix812 joined #salt
11:04 matanya joined #salt
11:06 scott_w joined #salt
11:10 kyusan joined #salt
11:10 zandy joined #salt
11:14 e71c joined #salt
11:17 hhenkel Hi all, I'm currently playing around with salt and I'm trying to figure out to represent data of a weblogic domain with salt.
11:18 hhenkel I guess grains are the way to go and I was able to get going with some custom grains: https://www.refheap.com/20759
11:19 hhenkel Now my question is, if this is the way to go or if it is a better idea to have single entries for each value?
11:20 hhenkel My intention would be to create crontab entries from the information I got there.
11:32 hhenkel Is it possible to address a subsection of my wls grain (uk_1211) with grains.item ?
11:32 Destro {% set hostname = grains['id'] %}
11:32 Katafalkas joined #salt
11:33 Destro {{ hostname }}
11:33 Destro or you could do remote executions such as
11:33 Destro {% set gateway = salt['cmd.run']("hostname") %}
11:34 Destro Hope that answers the question
11:34 hhenkel Destro: I'll give it a try, thanks.
11:37 vkurupcld joined #salt
11:38 hhenkel Destro: hmm, not  sure if I made clear what I'm trying to achieve... is it possible to run something like 'salt -G "wls:uk_1211" cmd.run "uptime"'
11:38 hhenkel I think the stuff you mentioned is handy to be used within a sls file, right?
11:39 Destro sls states, which I was referring to
11:40 hhenkel okay, I will need that later on with the cronjob stuff then. So is it possible to filter on a datastructure like that?
11:40 bcc left #salt
11:41 viq Hello. Close to the end of https://github.com/saltstack/salt/wiki/Cheat-Sheet is an example of users. What would be a good way (and syntax) to pull the groups user is to belong to from somewhere else?
11:43 viq hhenkel: if you can get that information then you can filter on it, yes
11:43 viq hhenkel: https://salt.readthedocs.org/en/latest/topics/targeting/grains.html
11:44 scott_w joined #salt
11:49 cron0 joined #salt
11:54 hhenkel viq: Thanks for that link I allready had a look at it. My problem/question is, if I can use this with a nested datastructure.
11:55 hhenkel Something like: wls: {'uk_1211': {'server0': "test"}}
11:56 viq I don't think I follow. What is that structure, where does it come from?
11:57 AviMarcus joined #salt
11:59 backjlack joined #salt
12:05 hhenkel viq: it is provided by local grains (also see the link to refheap I provided).
12:06 hhenkel viq: I'm not sure if that is the way to go...I'm trying to find a good way to store data describing my weblogic domain and it seemed like a good idea to use grains here.
12:07 viq hhenkel: have you looked at pillars?
12:08 hhenkel viq: I read the page describing pillars, but since the info would be available on the minions I thought it would be easier to use pillars instead of grains.
12:09 viq hhenkel: also, in the examples on that grains targeting page you have: salt -G 'ec2_tags:environment:*production*'
12:09 juasiepo joined #salt
12:11 viq so in your case it would probably be -G 'uk_1211:server0:test'
12:11 hhenkel viq: If I try something like this: ' salt -G "wls:uk_1211:common:domainname:uk" test.ping ' I get: No minions matched the target. No command was sent, no jid was assigned
12:12 hhenkel I'll try with the simplified grains.
12:12 halfss_ joined #salt
12:12 viq hhenkel: try getting the grains off of the minion, see what it shows
12:12 tty-1 joined #salt
12:15 hhenkel viq: Seems like the master does not see the local grains when I run salt '*' grains.ls
12:16 viq IIRC there is some command to refresh the grains
12:17 viq hhenkel: or you may need to restart minon
12:17 viq minion
12:18 hhenkel The restart did help
12:18 viq yeah, grains are static, read/generated on minion start, AFAIK
12:19 viq That's why pillars can be better ;)
12:19 xl1 left #salt
12:19 hhenkel viq: But pillars are on the master if I recall it right?
12:20 viq hhenkel: "yes", but you may be misunderstanding things
12:20 viq hhenkel: what are you using grains for?
12:21 zandy joined #salt
12:22 hhenkel viq: I got a environment of around 400 servers, 150 of those run some kind of weblogic. Weblogic got a configuration unit, that is called weblogic domain.
12:23 viq hhenkel: yeah, but if you're setting everything from salt, salt knows what pillars machine has, so can set things accordingly
12:23 hhenkel I got a script that is able to provide me various information regarding those domains, but as it is a jython application it is rather slow. So the idea was to have this information available in salt.
12:24 hhenkel viq: Yes, it is kind of a chicken-and-egg problem I currently don't have all this information available. I would need to query each server to recieve it.
12:25 viq mhmm
12:25 shinylasers joined #salt
12:26 shinylasers joined #salt
12:29 hhenkel Also this config is quite static and it is currently not planned to have any configuration management system to alter anything, as this would involve calling generated jython scripts.
12:30 ppfeifer joined #salt
12:30 ppfeifer hello there
12:30 hhenkel But you're right for a initial setup it would be interesting to have this inforation available on the master.
12:31 ppfeifer i'm investigating different ccm's right now....and i've got a few... perhaps many questions...
12:31 hhenkel viq: But what is the benefit of having pillars for my specific question regarding the filtering of hosts?
12:33 viq hhenkel: not that much, just a different place to store data. Personally with that amount of machines I would prefer to have the info in a central location
12:36 hhenkel viq: okay, it may be a good idea to have it local you're right.
12:37 viq (or, well, even with 2+ machines for that matter, but that's me :P )
12:37 ppfeifer here is my use case first.... i've got arround 100 different packages available as sourcecode, then i have got many binary files as well, and a lot of *.deb packages which are dependencies on the bin files. and now i'd like to specify configurations where i can define which version of the sourcecode, bin files and dependeny in a certain version as one config-package. which i can deploy on different vm's
12:38 hhenkel viq: None the less is still struggling with addressing my datastructure: salt -G 'uk_1211:*' grains.item uk_1211 gives me something like: uk_1211: {'server0': 'test'}
12:38 ppfeifer furthermore i always want to know which configuration is currently running on every machine
12:38 ppfeifer can i do this with salt?
12:38 hhenkel But something like salt -G 'uk_1211:server*' grains.item uk_1211 fails.
12:41 viq ppfeifer: sounds somewhat like what eg pillars can be used for
12:43 viq ppfeifer: I am not sure I fully grasp what you're trying to do, but from what I understand, then yes, you can say "this machine is to have those packages and those configs installed"
12:44 viq hhenkel: and explicitly, -G 'uk_1211:server0:test' ?
12:45 scott_w joined #salt
12:46 hhenkel viq: No minions matched the target. No command was sent, no jid was assigned.
12:47 viq hhenkel: sorry, don't really know, haven't dealt much with grains yet, especially custom ones
12:49 hhenkel viq: nonetheless thanks for your afford!
12:50 ppfeifer viq: thx for the answer. what is pillars?
12:50 viq ppfeifer: http://docs.saltstack.com/topics/pillar/index.html
12:50 viq ppfeifer: "a central place to store variables"
12:52 ppfeifer viq: thanks. :) is there a ui available too, or do i have to edit lots of sls files?
12:52 viq ppfeifer: all of salt config is done via text files
12:52 abele joined #salt
12:52 viq So sure, you have a UI - your favourite text editor ;)
12:53 ppfeifer viq: lol
12:54 viq ppfeifer: there is something called halite, but it's not really relevant to your question how you asked it ;)
12:54 dangerousbeans joined #salt
12:54 dangerousbeans Hey salty people, I've got some noob debugging questions - can anyone help?
12:54 dangerousbeans my remote minions can't connect to my salt master, but a local minion can
12:55 dangerousbeans HOW TO DEBUG? :C
12:55 viq dangerousbeans: firewalls, listen address, hostname?
12:55 scott_w joined #salt
12:55 dangerousbeans have set listen address to 0.0.0.0
12:55 viq dangerousbeans: verify with 'netstat -pltun" (I like those flags ;)
12:55 ppfeifer_ joined #salt
12:55 dangerousbeans in the config, but it's not saying that it's binding to this inside the output of salt-master -l debug
12:56 dangerousbeans yeah netstat looks good, got some shizzle listening to 4505 and 4506
12:57 xt joined #salt
12:57 viq dangerousbeans: yeah, but on what address?
12:58 viq 127.0.0.1, or?
12:58 dangerousbeans 0.0.0.0
12:58 viq ok
12:58 dangerousbeans so all interfaces - right?
12:58 viq now, are you running a firewall?
12:58 viq yes
12:58 dangerousbeans I don't believe so, it's just vanilla ubuntu server
12:58 dangerousbeans on rackspace
12:58 viq iptables -vnL
12:59 blee joined #salt
12:59 dangerousbeans okay that's showing a whole bunch of stuff
12:59 viq well then :P
13:03 dangerousbeans OMG THAT WAS IT :D
13:03 dangerousbeans firewall => tunneled through
13:03 dangerousbeans mwohahaha
13:03 xt joined #salt
13:03 dangerousbeans viq: thanks
13:03 viq :)
13:05 anitak joined #salt
13:13 jeddi joined #salt
13:15 viq How would I write a state that would create all the groups machine sees from this pillar? https://gist.github.com/viq/7430604
13:22 scott_w joined #salt
13:23 Teknix joined #salt
13:24 premera joined #salt
13:28 bezaban logged in to the entire virtualization infrastructure, blissfully knowing that this will be the last time I need to do such a repetetive task
13:28 bezaban s/virtual hosts/virtualization infrastructure/
13:28 bezaban and switch them around..
13:29 bezaban now to get access to all the physical machines :s
13:30 bhosmer joined #salt
13:33 ravibhure joined #salt
13:36 mwillhite joined #salt
13:36 mike25ro joined #salt
13:37 mike25ro on centos ... how can i install the LATEST version... i have installed a version but is 0.15.1
13:37 mike25ro hi everyone
13:40 viq mike25ro: how did you install it?
13:40 mike25ro yum install ...
13:40 mike25ro # salt --version salt 0.15.1
13:41 viq mike25ro: and what repos do you have enabled?
13:42 mike25ro good question
13:42 viq Because in EPEL, which appears to be the recommended place, I see 0.17.1
13:42 viq mike25ro: http://docs.saltstack.com/topics/installation/index.html
13:43 mike25ro i followed that one
13:43 mike25ro centos-6-stable-base                                             CentOS 6 - stable - base                                             6,381 centos-6-stable-epel                                             CentOS 6 - stable - EPEL                                             9,004 centos-6-stable-extras                                           CentOS 6 - stable - Extras                                              19 centos-6-stable
13:43 ipmb joined #salt
13:44 Voziv Anyone have a good logwatch configuration?
13:44 mike25ro viq: thanks
13:46 sebgoa joined #salt
13:51 Teknix joined #salt
13:51 scott_w joined #salt
13:52 jslatts joined #salt
13:54 teebes joined #salt
13:57 Mua joined #salt
13:58 pdayton joined #salt
13:58 mike25ro guys i want to start creating a web app that handles some of the data from salt... for example a web dashboard with status of minions - that respond to ping. What is the best option to parse the data from salt... in php? did anyone do this ?
13:59 xt You don't need a backend for that
13:59 viq mike25ro: have a look at halite
13:59 xt just do JSON calls from a javascript client
14:00 elfixit joined #salt
14:00 viq mike25ro: also https://github.com/lincolnloop/salmon
14:00 mike25ro viq:  thanks a lot
14:01 Voziv what is the default file root for salt:// ?
14:01 pdayton joined #salt
14:02 viq Voziv: /srv/salt/ - or just look for file_roots in master config
14:02 shuss joined #salt
14:02 Voziv viq: I'm running in masterless mode
14:03 zandy joined #salt
14:03 viq I don't know what happens then. But you could run with -l debug and see what it shows
14:03 oz_akan_ joined #salt
14:04 scott_w joined #salt
14:05 Voziv I updated the paths assuming that /srv/salt/ was the root. It works now :)
14:05 oz_akan_ joined #salt
14:05 krandun left #salt
14:06 brianhicks joined #salt
14:06 ipmb_ joined #salt
14:09 Gifflen joined #salt
14:09 juicer2 joined #salt
14:12 Katafalkas joined #salt
14:13 shuss Hi all, I'd be grateful to anyone who'll point me to a state file that installs some software from source (tar, configure, make and make install)
14:13 Brew joined #salt
14:15 thelorax123 joined #salt
14:16 viq shuss: if you want to do it on many machines this may be a point where you look at fpm ;)
14:16 shuss viq: I'll look it up, thanks
14:17 viq shuss: https://github.com/jordansissel/fpm
14:17 shuss viq: thanks, mate
14:19 viq AKA "how to make your own packages in a simple(r) way"
14:19 racooper joined #salt
14:20 shuss viq: looks pretty cool
14:22 pkruithof joined #salt
14:22 viq shuss: and if you want this to happen regularly and automatically, it may be a point where you want to look at jenkins ;)
14:23 shuss viq: yeah, the CI stuff..
14:23 khodgson joined #salt
14:25 linjan_ joined #salt
14:26 thelorax123 joined #salt
14:26 pkruithof I just upgraded my Ubuntu version to 13.10 (from 13.04), and now I get errors that the salt-daily packages cannot be downloaded
14:27 pkruithof this is because the ppa is not imported. I tried adding ppa: saltstack/salt-daily to my managed pkgrepo's, but it does not work
14:27 pkruithof should this be done manually or something?
14:28 Iwirada joined #salt
14:29 jslatts joined #salt
14:29 Kholloway joined #salt
14:30 shuss pkruithof: did you check your /etc/sources.list.d directory? Can you see the list file there?
14:31 pkruithof I don't see it there
14:31 shuss pkruithof: You'll need to add the PPA list file again
14:31 techdragon joined #salt
14:31 sroegner joined #salt
14:31 pkruithof shouldn't the ppa be added automatically?
14:32 shuss pkruithof: yeah, it should. Also I don't understand how the upgrade can delete it. But if it's not there, adding it again is safe.
14:33 pkruithof It didn't delete it, I'm running a VM using Vagrant. So I upgraded the Ubuntu version there, and created a new box
14:33 bhosmer_ joined #salt
14:33 pkruithof but still, that shouldn't matter, I think
14:34 shuss yeah, try to manually add the repo as per: http://docs.saltstack.com/topics/installation/ubuntu.html
14:35 shuss don't forget to run "apt-get update" after you do.
14:35 mgw joined #salt
14:35 shuss You can verify if the list file exists and then try and install.
14:35 pkruithof Ah, so this is like a chicken/egg problem (that's how we call it here at least).
14:35 pkruithof ok got it, thanks
14:36 druonysus joined #salt
14:38 higgs001 joined #salt
14:39 deedubs joined #salt
14:40 deedubs Has anyone confirmed https://github.com/saltstack-formulas/users-formula works on latest?
14:40 deedubs I'm getting an error at https://github.com/saltstack-formulas/users-formula/blob/master/users/init.sls#L32
14:40 oraqol joined #salt
14:40 quickdry21 joined #salt
14:40 oraqol It works!
14:41 oraqol I left on Saturday and the webportal showed zero metrics for nodes, walked in today and they were all showing
14:41 oraqol Must have somthing to do with the checkin interval
14:41 oraqol in any case, it WORKS!
14:42 oraqol I may come back in the future, particularly to find out how to setup my own metrics to be monitored.  Thanks so much guys, your product is awesome!!
14:42 Katafalkas joined #salt
14:43 bhosmer joined #salt
14:45 mapu joined #salt
14:47 aptiko Can I include the contents of a file in the pillar?
14:48 aptiko Alternatively: Can I get an entire file from the pillar?
14:48 goodbytes joined #salt
14:51 DredTiger joined #salt
14:51 premera joined #salt
14:52 joeTSUNAMI joined #salt
14:53 Voziv How do I get salt to require additional packages that aren't sls files? (php5-cli for example)
14:54 MTecknology Write a state that includes that package and require that state.
14:54 amahon joined #salt
14:54 tyler-baker joined #salt
14:54 tyler-baker joined #salt
14:55 tty-1 joined #salt
14:57 utahcon salt returns mount: special device crwmc-test.pgx.local:/crc_nfs does not exist
14:58 utahcon but if I go to the server and run the mount by hand it works
14:58 utahcon ideas?
14:58 toastedpenguin joined #salt
14:58 Katafalkas joined #salt
14:59 micah_chatt joined #salt
14:59 premera_e joined #salt
15:00 tty-1__ joined #salt
15:00 cro utahcon: That's odd
15:00 Voziv Does salt run apt-get update everytime it sees "refresh: true" ? or will it skip if it knows it's already run?
15:01 cro If it was permissions, it would say 'access denied'
15:01 cro utahcon: and you're sure you're targeting the right minion?
15:02 networkpadawan joined #salt
15:02 premera joined #salt
15:04 utahcon cro: so if I run "salt 'crwmi-test*' mount.set_fstab /crc_nfs crwmc-test.pgx.local:/crc_nfs ext4" it works
15:04 higgs001 joined #salt
15:05 utahcon I lied
15:05 utahcon it makes the fstab entry, but the mount fails
15:06 cro right
15:06 cro huh
15:06 cro your minions run as root?
15:07 zandy_ joined #salt
15:07 gmoro joined #salt
15:08 deedubs How can I write to the log from a state file? Trying to debug :/
15:09 scott_w joined #salt
15:09 utahcon cro: yeah, as root
15:09 utahcon deedubs: you can use file.append
15:10 utahcon http://docs.saltstack.com/ref/states/all/salt.states.file.html#salt.states.file.append
15:11 shinylas_ joined #salt
15:11 khodgson joined #salt
15:12 forrest joined #salt
15:12 utahcon salt 'crwmi-test*' cmd.run 'mount crwmc-test.pgx.local:/crc_nfs /crc_nfs' works... what the heck
15:17 utahcon salt '*' mount.mount /crc_nfs crwmc-test.pgx.local:/crc_nfs works too
15:17 utahcon what the devil
15:18 kermit joined #salt
15:18 Katafalkas joined #salt
15:20 jkleckner joined #salt
15:21 cro what was the original command that didn't work?
15:23 cro bbl--switching from train to bus
15:24 alunduil joined #salt
15:24 deedubs debugging state files is the bane of my existence
15:24 Mef1 joined #salt
15:26 cedwards joined #salt
15:26 cedwards joined #salt
15:26 shuss joined #salt
15:27 deedubs is there a way to abort a state early if a condition isn't met?
15:31 pkruithof installing a package from a different pkgrepo does not work for me, but I don't see what I'm doing wrong
15:31 viq deedubs: uhm, state will not execute if you specified a requirement and it's not met
15:31 pkruithof for example: I want to install nginx using ppa:nginx/stable, the output shows this:
15:31 pkruithof [INFO    ] Executing command ['apt-get', '-q', '-y', '-o', 'DPkg::Options::=--force-confold', '-o', 'DPkg::Options::=--force-confdef', 'install', '"nginx=1.4.3-2~saucy0"'] in directory '/home/vagrant'
15:31 pkruithof [ERROR   ] Command '['apt-get', '-q', '-y', '-o', 'DPkg::Options::=--force-confold', '-o', 'DPkg::Options::=--force-confdef', 'install', '"nginx=1.4.3-2~saucy0"']' failed with return code: 100
15:32 deedubs viq: {% for name, user in pillar.get('users', []).items() %} I want to bail if name is not set
15:32 pkruithof when I enter this command manually however I get the usual prompt to install it, so that does work
15:32 viq deedubs: it will whine about unknown variable or some such
15:33 deedubs yes...
15:33 Teknix joined #salt
15:34 viq oh, there is some funny jinja syntax to deal with it, I think... Something like {%-
15:34 viq pkruithof: anything in /var/log/dpkg.log ?
15:34 viq pkruithof: als, pastebin of some kind
15:36 pkruithof these are the last 1000 lines: http://pastebin.com/raw.php?i=ZC03KxSK
15:36 pkruithof nothing to see, I think
15:37 pkruithof also, /etc/apt/sources.list.d/nginx-stable-saucy.list is there, with the deb line in it. So I think the repo is imported, but somehow salt cannot use it (wild guess)
15:37 Mef1 I'm having an odd problem moving my salt-minion from 16.4 to 17.1 on ubuntu.   I had a working sls file that when I moved to 17.1 broke when in rendering.  The pastebin is http://pastebin.com/JXSuQU6m
15:37 jetblack joined #salt
15:39 Mef1 After trying to debug this for a day (and essentially giving up) I decided to see if there was a later version of salt that actually fixed the issue.  i.e. maybe there was some regression in 17.1.
15:40 Mef1 The good news was when I got the "latest version" of salt stack by "curl -L http://bootstrap.saltstack.org | sudo sh -s -- git develop"   the rendering issue went away and everything worked
15:42 forrest Mef1, the plan was to release 0.17.2 on Friday, but there were issues integrating salt-cloud into the main repo. It should be out this week if all goes well.
15:42 Mef1 The bad news is or at least what I don't understand is the branching/revision number scheme of saltstack.  Specifically the salt-minion --version command after curl -L http://bootstrap.saltstack.org | sudo sh -s -- git develop returned salt-minion 0.17.0-4338-g44c0079
15:44 Mef1 Forrest is it odd that the version salt-minion 0.17.0-4338-g44c0079 is returned from getting out of dev branch and not something labeled 0.17.1.xxx?  Is 0.17.0-4338-g44c0079 later than 0.17.1?
15:44 forrest Yea someone else was having that issue as well Mef1, I don't think there was ever a resolution.
15:44 forrest but I have no idea why you're getting that git commit ID instead of 0.17.1-xxxx
15:45 Mef1 Is there some specific way to force getting something out 0.17.1.x?
15:46 Mef1 Obviously if the branch that I cloned is later than 0.17.1 and will be in 0.17.2 I'm happy otherwise I'm sort of xxxxed
15:46 Katafalk_ joined #salt
15:47 mpanetta joined #salt
15:47 ravibhure joined #salt
15:47 juasiepo joined #salt
15:48 forrest Mef1, you're doing it right as far as I know. maybe I just don't know the tag, I'd say take a look at a recently updated file on the develop branch, and compare that against a local file to see if they match up with the recent changes
15:48 forrest then you could confirm if you're getting the most recent release.
15:52 AdamSewell joined #salt
15:53 AdamSewell I have a minion that the master has accepted it's key but no command will run on int. If i run test.ping, the other minions will return as expected. Where do I start troubleshooting this?
15:53 forrest AdamSewell, start by adding the -l debug option
15:53 forrest to your test command
15:54 AdamSewell forrest, on the master or minion?
15:54 forrest when running your test.ping from the master
15:54 Mef1 forrest should I be submitting a bug for the issue regardless?
15:54 forrest for the pull issue?
15:55 AdamSewell forrest, interestingly, it doesn't even show the minion as being there
15:55 quickdry21 joined #salt
15:55 forrest Mef1, I don't think so, I'd say start a discussion on the mailing list.
15:55 forrest AdamSewell, and you accepted the key? Can you restart the master real quick?
15:56 AdamSewell forrest, just restarted and same result
15:56 AdamSewell I should also mention that the minion was setup by salt-cloud if that has anything to do with it
15:57 forrest hmm
15:57 quickdry21 joined #salt
15:57 forrest ok, what do you see in /etc/salt/pki/minions?
15:57 forrest is the box in there?
15:59 AdamSewell i'm guessing you're referring to /etc/salt/pki/master/minions?
15:59 forrest yea sorry
15:59 AdamSewell if so, yes it is
15:59 forrest hmm
15:59 forrest were your other machines built with salt-cloud?
15:59 forrest can you try to telnet from the minion to the maaster on 4506 and 4505?
16:00 mgw joined #salt
16:00 forrest *master
16:00 forrest You might also change your log level to debug, then restart the master to see if anything gets dropped into the logs.
16:02 AdamSewell forrest, one was, the other two were not
16:02 ipmb joined #salt
16:02 AdamSewell forrest, , interestingly the minion is saying it's waiting for the master to accept the key
16:02 forrest try salt-key -L
16:02 forrest on the master
16:03 TheRealBill joined #salt
16:03 utahcon duplicate keys?
16:03 utahcon I see that when I clone VMs a lot
16:03 AdamSewell also, i notice in the debug log here on the minion, that it says "setting up the salt minion "lixx-xx.members.linode.com". I'm not on linode.
16:03 jalbretsen joined #salt
16:03 AdamSewell forrest, salt-key -L says that it's acceppted
16:03 balboah [salt.config      ][WARNING ] Error parsing configuration file: /etc/salt/minion - 'str' object has no attribute 'copy'
16:03 forrest AdamSewell, hmm, maybe something triggered that makes it think you're on linode.
16:03 balboah wtf? after upgrading to 0.17.1 from 0.16.4
16:04 forrest AdamSewell, ok, like utahcon suggested then, see if there are dup keys maybe?
16:04 AdamSewell forrest, how do i do that?
16:04 forrest salt-key -L should show all keys
16:04 forrest are there any dups?
16:05 TheRealBill utahcon: yeah cloning an LXC container via LVM snapshot has problems as well. I have a script which cleans out the keys, the cache directory, and one other I dont' recall atm but it still doens't quite work right.
16:05 pdayton joined #salt
16:05 scott_w joined #salt
16:05 TheRealBill oh, it also removes /etx/salt/minion_id
16:05 TheRealBill etc*
16:06 AdamSewell forrest, oh no, there are no dups
16:06 forrest hmm
16:08 forrest AdamSewell, just to double check, can you confirm connectivity on 4505, and 4506 from the minion to the master?
16:08 UtahDave joined #salt
16:11 halfss_ joined #salt
16:12 AdamSewell forrest, sure, one sec
16:14 mgw joined #salt
16:14 viq OK, what am I doing wrong here, and how do I fix it? https://gist.github.com/viq/7430604
16:15 AdamSewell forrest, both work
16:16 thelorax123 joined #salt
16:17 higgs001 joined #salt
16:21 forrest AdamSewell, hmm that's really weird. Is there anything unique about this machine compared to the other one you built with salt-cloud
16:21 AdamSewell forrest, not that i'm aware of
16:22 AdamSewell forrest, it's on a cloud hosting provider
16:22 AdamSewell as are the others
16:23 viq And there are no keys awaiting approval on master? Maybe you could try purging all the keys on the minion and restarting salt-minion there, and see if new keys appear on master?
16:23 viq Also I guess tcpdump on master to verify the traffic is actually arriving...
16:24 jimallman joined #salt
16:24 danielbachhuber joined #salt
16:25 forrest viq, his minion still thinks the key hasn't been accepted for some reason
16:26 AdamSewell forrest, if i remove the key from the master, would it show back up later?
16:26 viq ah, I see it has been accepted and is on the master, sorry
16:26 forrest the minion would try to auth again most likely
16:26 forrest I mean the issue seems to be the master communicating to the minion
16:26 forrest can you try to turn off iptables if you have them enabled?
16:27 viq yeah, it should, though you may need to restart it (or maybe just to accelerate the process)
16:27 AdamSewell i can turn off ufw sure
16:27 Boohbah joined #salt
16:27 forrest let's try that, because so far everything you've shown seems to be communication problems to the minion from the master (can't ping, minion is waiting on master to say it has accepted the key, etc.)
16:27 balboah anyone knows how to re-publish a superseded package in launchpad ppa? I use to copy the salt ones because they usually break my states which it did now. And I want 0.16 back again. Deleting the package of 0.17 didn't help as I can't copy the old one again and it's not "published"
16:27 jslatts joined #salt
16:28 AdamSewell forrest, i take that back
16:28 JulianGindi joined #salt
16:28 AdamSewell telnet does not work from the minion to the master
16:28 forrest viq, on 4505/4506?
16:28 AdamSewell i was in the wrong terminal
16:29 Mef1 Forrest how to I find out what the latest candidate is for 0.17.2  I looked and could not find any build that were labeled .17.1.xxx other than the release.  Is the procedure to "promote" a 0.17.0.xxx version to become 0.17.2???  The real question is how to ensure I have the latest
16:29 swa joined #salt
16:29 viq forrest: hm?
16:29 forrest sorry meant for AdamSewell viq
16:29 viq ;)
16:30 forrest Mef1, I honestly don't know, I don't keep up with the version schemeing when it comes to that stuff very well :\
16:31 AdamSewell forrest, oh yea, it will not communicate on 4505 or 4506
16:31 AdamSewell minion to master
16:31 forrest to the master from the minion? Yea that's a problem
16:31 forrest you need to get that working
16:31 forrest ports aren't allowed out on the minion perhaps?
16:31 viq forrest: maybe you'd have some pointers on https://gist.github.com/viq/7430604 once you have a free moment? ;)
16:32 forrest I'm looking at it now viq
16:32 viq thanks :)
16:32 forrest having a lame brain here on the jinja handling
16:33 viq ;)
16:33 forrest should this be pillar['groups']('group1') ?
16:33 forrest I can't remember
16:33 viq don't think so, https://raw.github.com/jesusaurus/hpcs-salt-state/master/apt/repo/distro uses [] everywhere
16:34 forrest don't trust that jesusaurus guy, he's terrible :P
16:34 viq It would be salt['pillar.get']('groups:group1') where you would use round brackets
16:34 viq And maybe I should try that
16:35 forrest weird, it seems like what you have should work
16:35 viq Yeah...
16:36 forrest let's try this
16:36 viq and in a different place {% for user, args in pillar['group1'].iteritems() %} works just fine
16:36 AdamSewell forrest, no firewall is installed on this minion so no, i don't think that would be the issue
16:36 forrest are 4505/4506 open on the master? Seems like they would have to be.
16:36 * Gareth waves
16:36 forrest since your other machines work
16:36 AdamSewell other minions work, so yea i think so
16:37 viq AdamSewell: linux? You verified with 'iptables -vnL' ?
16:37 jergerber joined #salt
16:37 forrest viq: https://gist.github.com/gravyboat/7434139
16:37 forrest something like that
16:37 AdamSewell i'm using ufw and have allowed the salt ports
16:37 bitz joined #salt
16:37 forrest are they allowed outbound though?
16:37 AdamSewell i think something is screwy with this image
16:38 AdamSewell when i ping, it's saying from an IP address I've never seen before
16:38 viq forrest: though the machine in question does see the values... but let me see
16:39 viq forrest: nope
16:39 AdamSewell forrest, right now, the firewall is completely off
16:39 AdamSewell i can't seem to ping the master either
16:39 AdamSewell that's just weird
16:39 forrest yea that is weird
16:40 penguin_dan Anyone here using salt with vagrant?  Running into some trouble with the pillar data not being passed in from vagrantfile during provisioning
16:40 forrest viq, why'd you have to break it :P
16:40 zandy joined #salt
16:41 viq penguin_dan: I am, but I am using salt "normally", not putting much of stuff in vagrantfile
16:41 forrest I'm really not understanding why that doesn't work though viq, it looks like it should. syntax looks ok
16:41 viq penguin_dan: if you want to see - https://github.com/viq/cm-lab-salt
16:41 bhosmer joined #salt
16:41 scott_w joined #salt
16:41 ieldib joined #salt
16:41 bhosmer joined #salt
16:41 AdamSewell forrest, ha!
16:41 AdamSewell i see what the problem is
16:41 viq forrest: that's what I'm thinking too, unless I'm hitting some reserved keywords or something
16:41 AdamSewell i'm an idiot
16:41 forrest oh?
16:42 forrest viq, is groups a reserved keyword in pillar?
16:42 forrest maybe try changing it up?
16:42 forrest I don't see why it possibly would be though
16:42 AdamSewell forrest, so, all of this is working off the a private network, well this particular host does not have an interface with a private ip.
16:42 AdamSewell doh
16:42 forrest and also viq, is this going to an openBSD machine, or Debian box?
16:42 forrest AdamSewell, ahh that makes sense
16:43 viq forrest: http://pbot.rmdir.de/lDFxYRQ_vnBR3XRiXi2sRQ didn't help either
16:43 viq forrest: OpenBSD
16:43 penguin_dan viq: I just want to be able to make sure my vagrant user is a member of a certain group if the vagrant user exists... is there a convenient way to go about this?
16:43 viq forrest: fun thing is, that the users states get the groups properly from the same pillar
16:44 Drawsmcgraw joined #salt
16:44 penguin_dan thought i could just pass in some pillar data, but it doesn't appear to be working =/
16:44 viq penguin_dan: just run highstate upon machine creation?
16:44 viq penguin_dan: just put normal salt in there and run it in normal way? Admittedly I did not try to do much in the provisioning stage
16:44 AdamSewell well, that leads me to another question. in salt-cloud, how can I tell my digital ocean provider to include a private ip?
16:46 forrest wait a second viq
16:46 forrest I think I see it
16:47 forrest so you're trying to use iteritems on a single item. change pillar['groups']['group1'] to pillar['groups']
16:47 forrest because in jesusaurus' example, he's only going down to distros, and then he has the key value pairs, with how you're doing it, you're trying to go all the way down to the value, without the key, so there's nothing to iterate over.
16:47 viq forrest: then it will just create group1 and group3
16:48 viq I did try :P
16:48 viq Though not necessarily in the right way
16:48 forrest but gname should equal group1, and group3
16:48 forrest while glist equals the values
16:49 viq I'll paste you what I have right now, and the error
16:50 penguin_dan viq: I just didn't want to create the vagrant user on every machine if I didn't have to.
16:50 viq https://gist.github.com/viq/7434382
16:51 viq penguin_dan: are you connecting your vagrant machine to a central master?
16:51 penguin_dan viq: no, but i'm using the same configurations that I use on salt-master locally
16:52 fllr joined #salt
16:52 viq forrest: reloaded, with output for using both key and value in there
16:52 fllr Anyway to have salt install libraries using make? I'm trying to install a library that doesn't offer .deb files yet...
16:53 viq fllr: fpm to the rescue ;)
16:53 viq fllr: https://github.com/jordansissel/fpm/wiki
16:53 forrest as stupid as this sounds viq, can you add #!yaml|jinja to the top of groups.sls?
16:53 viq fllr: but you could, yes, with cmd
16:54 viq forrest: which, state or pillar?
16:54 forrest the state
16:54 forrest since it's saying that it expects document start, but is finding block mapping start, which is weird
16:54 fllr viq: Ah! Nice... this will do! :D
16:54 viq And literally #!yaml|jinja  ?
16:54 forrest I think that's the syntax
16:54 philipsd6 joined #salt
16:54 forrest so it parses as yaml then jinja
16:54 forrest I'd have to find the doc to be 100% sure
16:54 scott_w joined #salt
16:55 forrest actually, do #!jinja|yaml
16:55 forrest I found the doc
16:55 viq forrest: http://pbot.rmdir.de/B_pQe6HvSuEouw88qHUfXA
16:55 forrest yea switch to to jinja then yaml
16:56 viq http://pbot.rmdir.de/ZLXdQn-gX-J2QzKkoNkidg
16:57 forrest hmm, ok so it's almost as if it's not getting the group value...
16:57 forrest man, I don't understand, the iteritems should be taking each item in the groups section..
16:57 AdamSewell forrest, how do you rename a key with salt-key?
16:58 forrest rename a key?
16:58 viq hah, progress, or at least a difference ;)
16:58 viq http://pbot.rmdir.de/Y9zdusuNjcefj17thXOTtA
16:58 AdamSewell i recreated that server so that it would have a private ip, but now when i run salt-key -L is shows as that linode name istead of the hostname
16:58 forrest AdamSewell, maybe just move the file down in the /etc/salt/pki/masters/minions directory? I've never tried it before. I just remove the key and re-join.
16:59 forrest you are missing a quite on pillar.get viq
16:59 pdayton joined #salt
16:59 forrest or you added one
16:59 viq oh, duh
16:59 viq yah, same error then
17:00 forrest as the old one?
17:00 micah_chatt joined #salt
17:01 AdamSewell forrest, it's a unacceppted key right now so it doesn't show in the minions folder
17:02 ajw0100 joined #salt
17:02 forrest AdamSewell, oh ok, maybe try salt-key -d 'name_of_key'
17:02 forrest I don't know if it already has to be registered
17:02 forrest You could also try restarting the master/minion to see if it tries to push the key again
17:02 jslatts joined #salt
17:03 forrest I haven't messed with doing that much AdamSewell so I'm pretty useless there
17:03 AdamSewell forrest, got ya, thanks for the help though
17:04 carlos joined #salt
17:06 viq I did that, if you remove the key on master the minion will try to reauthenticate, new key should appear and you can accept it then
17:06 amckinley joined #salt
17:10 forrest AdamSewell, np
17:10 hazzadous joined #salt
17:11 viq forrest: yes, same as the old one, sorry, didn't notice your response
17:11 forrest np
17:12 forrest I really don't understand why it's not working
17:12 forrest it seems like it should
17:12 viq It does, doesn't it
17:12 Khazix joined #salt
17:12 forrest and you said the same thing is working elsewhere right?
17:12 viq yeah, let me see if I have that up
17:14 viq forrest: yes, that same pillar works fine in https://github.com/viq/cm-lab-salt/blob/users/salt/roots/salt/users/group1.sls
17:14 forrest there you have another section called group1?
17:14 forrest and then the users in your pillar?
17:15 forrest because in that one, you're only doing group1, not groups, group1
17:15 khodgson joined #salt
17:15 redondos joined #salt
17:15 redondos joined #salt
17:15 bemehow joined #salt
17:16 viq Yeah, because in pillar groups:group1 is info what groups users in group1 should belong to
17:16 forrest viq
17:16 forrest we are morons
17:16 forrest https://gist.github.com/viq/7430604#file-state-groups-sls-L2
17:16 forrest look at that line
17:17 IJNX joined #salt
17:17 forrest granted, it should stil be just 'groups'
17:17 forrest what is missing on that line?
17:17 patrek joined #salt
17:17 forrest a :
17:17 viq :D
17:17 forrest {{ glist }}:
17:17 forrest not {{ glist }}
17:17 jas joined #salt
17:17 forrest try that and iterate through the groups
17:18 higgs001 joined #salt
17:18 Guest31504 left #salt
17:20 ipmb joined #salt
17:21 ecjas joined #salt
17:22 viq forrest: ish. with http://pbot.rmdir.de/hI5S6pHin0Cqmq4304P0DA it creates group1 and group3... But that's not what I want.
17:22 viq I want to create groups that are contents of those sub-pillars
17:22 forrest yea you should have ba
17:22 forrest not ab
17:22 forrest ab is the group name
17:22 forrest or the key
17:22 forrest ba is the value of the key
17:23 forrest you can probably get rid of #!jinja|yaml at the top too
17:23 viq well, with ba I get http://pbot.rmdir.de/Qv_RGbyQUBU2AY6i4iibng
17:23 forrest weird, can you try it again with the way you originally had it ? ['groups']['group1'] maybe?
17:23 UtahDave https://hp.taleo.net/careersection/2/jobdetail.ftl?job=2834316
17:24 UtahDave ^^ cool job opening at HP working on OpenStack
17:24 forrest hey UtahDave
17:24 UtahDave Hey, forrest!
17:24 forrest You recover from your trip?
17:24 forrest Man HP must be throwing out the dollars on that job, 6-10 years of experience, jeez.
17:26 viq forrest: http://pbot.rmdir.de/8HDC7CRvztwVMDaZF0je8A
17:26 forrest yea because it can't iterate over that since it's not a key value pair
17:26 shinylasers joined #salt
17:27 viq well, http://pbot.rmdir.de/igRkuOCfT9Y6mHVfomCYBQ
17:28 UtahDave forrest: Yeah, I'm finally over the jet lag. It hit me pretty hard on Sunday.
17:28 forrest UtahDave, yea it's brutally. Last time I went overseas it took me a full week to be back 100%
17:28 forrest *brutal
17:30 forrest ok viq, so we need to do something after we iterate for the items
17:30 cdcalef joined #salt
17:31 forrest https://gist.github.com/gravyboat/7435127
17:31 forrest something like that
17:31 forrest not sure if that will work
17:32 forrest but something that looks like that is what needs to happen, I don't know if there's a cleaner way to do it
17:32 forrest except remove group1, so we just iterate over groups
17:33 viq http://pbot.rmdir.de/rIqz_QyHeVmkHpwbHOgY-Q
17:33 forrest yea get rid of ['group1']
17:33 forrest that's breaking it
17:33 forrest because there's nothing to iterate over in group1
17:34 forrest I modified my gist
17:35 viq ooh, that seems to have done it, let me look closer
17:36 viq forrest: yup, that generates all the groups, awesome, thank you!
17:36 forrest viq, yea np, sorry it took so long. You should make those variables less crappy though :P
17:37 viq Yeah, was just making sure I was not hitting some keywords, and didn't want to bother coming up with stuff ;P
17:37 forrest yea for sure
17:37 forrest It only took two of us 30 minutes to solve, wooo
17:37 viq hehe
17:38 viq ...that we count of today ;P
17:38 forrest hah
17:40 thelorax123 joined #salt
17:40 anuvrat joined #salt
17:41 zandy joined #salt
17:44 backjlack joined #salt
17:45 micah_chatt joined #salt
17:49 iMil joined #salt
17:49 iMil joined #salt
17:51 viq forrest: and it's in my repo for all to be seen ;)
17:51 viq Now to figure out ssh keys, and removal of users
17:51 forrest viq, haha nice.
17:52 forrest I'm sure someone will look at it and go 'what is with this terrible loop! Here is my better way to do it!'
17:52 viq forrest: hopefully they will follow it with a pull request or a comment so I know about it ;)
17:52 jslatts joined #salt
17:54 AviMarcus joined #salt
17:55 xmltok joined #salt
17:56 forrest yea
18:01 amckinley joined #salt
18:04 pipps_ joined #salt
18:07 khodgson joined #salt
18:13 ajw0100 joined #salt
18:14 rgarcia_ joined #salt
18:15 ipmb joined #salt
18:15 Drawsmcgraw1 joined #salt
18:19 HeadAIX joined #salt
18:19 jslatts joined #salt
18:22 micah_chatt joined #salt
18:22 bhosmer joined #salt
18:22 rgarcia_ joined #salt
18:23 Linz_ joined #salt
18:27 tty-1 joined #salt
18:29 Linz_ joined #salt
18:31 mwillhite joined #salt
18:33 amahon joined #salt
18:37 bhosmer_ joined #salt
18:37 Katafalkas joined #salt
18:40 MTecknology UtahDave: It's official. I'm now registered for SaltConf 2014! :D
18:40 forrest MTecknology, nice, are you gonna do one of the tutorials and take the eng test?
18:41 MTecknology the tutorials? pre-conf?
18:41 MTecknology I'm signed up for "Advanced SaltStack States for Configuration Management"
18:41 forrest yea that's one of them
18:41 zandy joined #salt
18:41 MTecknology I'd like to attend the third one too... :(
18:42 MTecknology Does that salt cert exam cost extra then?
18:42 forrest SSCE certification written exams will be available exclusively to SaltConf pre-conference training attendees and to individuals who have previously taken SaltStack regional or on-site training courses.
18:42 forrest doesn't look like it
18:43 oz_akan_ joined #salt
18:43 juicer2 joined #salt
18:43 MTecknology I figured that could be read as it's available only to them, but is an exam that costs money. I'm hoping it's meant as it sounds. That'd just be awesome.
18:43 forrest if you pass :P
18:43 UtahDave MTecknology: nice!
18:44 MTecknology I think I might possible know enough to maybe be able to pass... I'm hoping.
18:44 UtahDave I'm glad you're coming!
18:44 MTecknology Me too! :D
18:44 forrest UtahDave, are you guys gonna put up test objectives for the SSCE?
18:44 cron0_ joined #salt
18:44 UtahDave forrest: I'm not sure what the plans are for that.
18:44 MTecknology I'll get to meet all the awesome salt devs. Hopefully they can make me a smarter person too!
18:44 UtahDave all that is still being sorted out.
18:45 forrest gotcha
18:45 UtahDave MTecknology: of course!  :)
18:45 MTecknology UtahDave: Is it an extra cost, part of the pre-conf cost, or undecided?
18:45 Katafalkas joined #salt
18:45 Linz_ joined #salt
18:47 bhosmer joined #salt
18:48 UtahDave MTecknology: the actual test?
18:49 MTecknology UtahDave: ya
18:50 oz_akan_ joined #salt
18:51 UtahDave Just went and asked
18:52 UtahDave the pre-conf classes include the cost to take the exam
18:52 forrest oh nice
18:52 UtahDave yep!
18:52 forrest Can you just register for the pre-conference classes?
18:53 forrest Doesn't look like there is a way to do that through the site
18:53 MTecknology YAY!
18:53 sgviking joined #salt
18:53 MTecknology forrest: just the pre-conf and not the conf?
18:53 forrest Yea
18:53 MTecknology nope
18:53 MTecknology It's in the FAQ
18:54 MTecknology #5
18:54 forrest oh nice
18:54 forrest I don't remember that on the faq last time I checked, thanks
18:54 forrest probably missed it :P
18:54 MTecknology or they added it
18:54 MTecknology UtahDave: thanks for checking for us!
18:55 jhulten joined #salt
18:55 MTecknology My next question is... what's on the test? :P
18:55 forrest lol I already asked that!
18:55 UtahDave MTecknology: no problem!  I'm happy that it's included, too!
18:55 UtahDave MTecknology: docs.saltstack.com
18:55 UtahDave lol
18:55 forrest lol
18:56 UtahDave I'm sure we'll have some kind of synopsis up as we get closer
18:56 ieldib left #salt
18:56 MTecknology I'm actually hoping to eventually read all of the docs... before going to the conference.
18:57 MTecknology I'd like to write a few runners and ext_pillars and such too
18:59 jhulten joined #salt
19:00 MTecknology So... is this the first time SSCE is being offered?
19:00 forrest yea
19:00 ravibhure1 joined #salt
19:00 MTecknology nice
19:00 MTecknology Man... I'm excited.
19:01 forrest still got a few months to go :P
19:01 MTecknology but I'm going to start now
19:01 forrest working with salt you mean?
19:01 forrest I just meant to be excited
19:01 forrest that reminds me I need to go home and get my pycon registration done :\
19:01 MTecknology right now, I'm writing docs; when I get done I'll be reading the salt docs.
19:02 forrest writing salt docs?
19:02 MTecknology writing docs on all of the servers that I manage
19:02 forrest oh
19:02 chutz does a 0.17 master work with 0.16 minions?
19:03 forrest not really chutz, you should make sure they're all on the same version
19:03 MTecknology I thought it would work fine...
19:03 forrest uhh no I thought we discussed in here over the weekend that someone was trying it and it wasn't?
19:03 MTecknology oh..
19:03 forrest I mean some stuff might work, but they were encountering issues
19:03 forrest I could be wrong though
19:03 xerxas joined #salt
19:04 MTecknology Usually the master seems to keep enough backward compatibility for most stuff to work (it seems).
19:04 forrest Maybe it was because they were using zeromq 2 vs. 3....
19:04 troyready joined #salt
19:04 forrest I'm having a hard time remembering
19:04 chutz it makes migration rather difficult when you get to update thousands of machines at the same time...
19:04 pears joined #salt
19:04 forrest yea give it a shot chutz, or wait for one of the devs to weigh in
19:05 MTecknology Ya... that's the reason I thought they always pushed for backward compatibility. You ALWAYS update the master first, then trickle update the minions.
19:05 chutz UtahDave: ^
19:05 MTecknology but... maybe there were issues...
19:06 chutz I don't mind updating the master first
19:06 forrest like I said, it might have been related to zeromq 2 versus 3
19:06 chutz but having to do them all at once is rather nasty
19:06 forrest chutz, yea I understand.
19:06 MTecknology ah.. that could do that too
19:06 forrest where the master had 3, but the minions still had 2 or something
19:06 MTecknology Shall we wait for a dev? :)
19:06 forrest Yea of course
19:07 forrest don't want to risk anything
19:07 forrest and it's lunch time anyways
19:08 Chrisje joined #salt
19:09 ipmb_ joined #salt
19:10 KyleG joined #salt
19:10 KyleG joined #salt
19:11 AdamSewell Does anyone know a good video that will explain states and pillars?
19:12 khodgson joined #salt
19:13 Ryan_Lane joined #salt
19:14 anti_ joined #salt
19:14 anti_ Good afternoon - Anyone here have experience with the jobs runner?
19:16 anti_ running "salt-run jobs.lookup_jid 10000000000000" on the salt master as non-root user gives [CRITICAL] Salt configured to run as user "root" but unable to switch.
19:17 MTecknology I really really hate writing documentation
19:17 anti_ Client_acl is enabled for this user for .* in addition to permissive_pki_access set to true
19:19 Gareth AdamSewell: Look at the Salt-Air videos on Youtube.  I don't know which one off hand but you should find something helpful there.
19:19 Koma joined #salt
19:22 khodgson joined #salt
19:22 anti_ Anyone have any ideas?  My situation prevents salt-run from executing as root so I'm attempting to get a user whitelisted on the system to be able to lookup job history
19:23 mwillhite joined #salt
19:23 UtahDave chutz: no, once you move your master to 0.17.1 you have to upgrade your minions as well. There was a protocol change to fix a security issue
19:24 anti_ my other option was trying to use the Python API but I equally struck out finding any examples for using the RunnerClient.  Instantiation takes some object which is different than that of the salt.client class
19:24 UtahDave AdamSewell: I don't know of a video that explains that, but I'm going to add that to my todo list!
19:24 mr_chris Am I correct in saying that http://wsgiarea.pocoo.org/jinja/docs/loops.html#range does not work in the default Jinja renderer?
19:25 mr_chris I get "render error: unknown tag 'range'"
19:25 UtahDave anti_: have you checked to see if there's an open issue on that?
19:25 Drawsmcgraw joined #salt
19:26 AdamSewell UtahDave and Gareth, thanks guys
19:26 fys left #salt
19:26 anti_ UtahDave: I was looking around and saw something related to non-root permissioning but nothing specific for salt-run.  I'm open to any workaround at this point.
19:27 UtahDave anti_: have you tried giving that user sudo rights to just salt-run?
19:28 anti_ UtahDave: no but that's an excellent simple suggestion here.  Thanks! ;)
19:30 jhulten joined #salt
19:33 ajw0100 joined #salt
19:34 bhosmer joined #salt
19:36 tty-1_ joined #salt
19:37 DanGarthwaite joined #salt
19:37 Gifflen joined #salt
19:38 kermit joined #salt
19:41 ajw0100 joined #salt
19:41 zandy joined #salt
19:42 jpcw joined #salt
19:43 zandy_ joined #salt
19:48 jkleckner joined #salt
19:49 jpcw joined #salt
19:51 ipmb joined #salt
19:53 ckao joined #salt
19:55 dvogt joined #salt
19:59 redondos joined #salt
19:59 redondos joined #salt
20:00 vandemar joined #salt
20:02 vandemar why does the salt documentation encourage fetching the bootstrap script over http?  So it redirects to an https github url but that redirect response could be mitm'd
20:04 DanGarthwaite vandemar: good question.  I'm checking git log now.
20:05 mpanetta Anyone know when 0.17.2 will be out?
20:05 vandemar DanGarthwaite: it's not just the readme on github, it's also the main salt install doc: http://docs.saltstack.com/topics/installation/index.html
20:05 vandemar also bootstrap.saltstack.org is listening on port 443 but it's speaking plain http on that port without ssl...
20:06 DanGarthwaite I think that is generated from code in git, so the log should show something if it has flip-flopped over time.
20:06 IJNX joined #salt
20:06 DanGarthwaite I can easily see this happening because some base server installs will/won't have openssl, ssl libs, etc, installed.
20:06 DanGarthwaite but then it would fail when redirected, so I'm checking.
20:09 matanya joined #salt
20:09 jhulten_ joined #salt
20:09 DanGarthwaite This is all over the code.  Even salt-cloud uses non-ssl.
20:10 DredTiger joined #salt
20:11 DanGarthwaite There are ten instances of "http://bootstrap" in the git repo, and zero instances of "https://bootstrap".
20:13 forrest AdamSewell, you still around?
20:13 forrest MTecknology, what's wrong with writing docs?
20:13 bhosmer joined #salt
20:15 xmltok joined #salt
20:16 teebes joined #salt
20:16 pipps_ joined #salt
20:16 jkleckner joined #salt
20:17 forrest AdamSewell, check out this talk that Tom did in San Fran, http://www.youtube.com/watch?v=yphLKSjnSU8 . He explains some of those concepts pretty well
20:19 drags aa/win 21
20:22 jslatts joined #salt
20:22 MTecknology forrest: having to sit down and clearly explain everything and writi it out and think through everything anyone might need to know without going too in depth (the ls command does...)
20:22 MTecknology forrest: I just find it a pain.
20:22 forrest Fair enough
20:22 forrest I guess I like having it all written down, because I know I'll forget
20:22 forrest and plus then I don't have to spend time explaining it 6 times
20:23 blafountain joined #salt
20:26 anti_ joined #salt
20:26 ajw0100 joined #salt
20:33 fishpen0 joined #salt
20:36 jetblack joined #salt
20:37 steveoliver so i've got a mariadb state that calls for mariadb-server pkg (on Ubuntu) … During state.highstate I see it looking for a mysql service running state — but I haven't specific this as a salt state.  Does this " come from " the mariadb-server package?
20:37 forrest mpanetta, hopefully sometime this week if all goes well
20:37 mpanetta forrest: Awesome!  This is good news :)
20:39 Gifflen joined #salt
20:40 forrest mpanetta, yea the guys were trying to finish it Friday, but salt-cloud was being a pain
20:40 forrest granted it will probably be a bit longer before packages are ready for distros
20:41 forrest steveoliver, I'm not sure, can you confirm what it tries to do when you install mariadb-server from the box itself?
20:41 zandy joined #salt
20:42 steveoliver forrest: good call - thanks.
20:42 mpanetta forrest: I can probably build at least the rpm based packages for our purposes here based on the old specs...
20:42 steveoliver the indication i'm speaking from is line 105. of http://paste2.org/vVjDB2wE
20:42 forrest mpanetta, they will be built by the team mpanetta, it's just an issue because they have to go through EPEL testing first
20:42 forrest mpanetta, now if you have some EPEL credits you can use to push it through :P
20:43 forrest steveoliver, you're seeing that but you don't have a state for mysql??
20:43 forrest what about line 64
20:43 forrest which creates the mysql user
20:43 forrest same thing on 53, mysql group
20:43 steveoliver forest: user, both
20:43 mpanetta forrest: Ah, no, I personally do not.  I don't know if we as a company do or not.
20:44 steveoliver i have a mariadb-server state — that's it — pushing to GH
20:44 forrest I would check that first, are you running this via high state?
20:44 steveoliver forest: https://github.com/circatree/salt-hosting-states
20:44 steveoliver forrest: highstate, yes
20:45 steveoliver also tried via state.sls mariadb
20:45 mpanetta forrest: Yeah I think we may need it sooner rather then later here, It seems that salt-cloud has created a mixed environment for us...  Since it always pulls latest.
20:46 mpanetta And we cant use 17.1 because of some sort of issue with mine (I am not familiar with it as much as a colligue of mine is)
20:46 forrest mpanetta, yea for sure
20:46 forrest steveoliver, so when you run via state.sls mariadb, do you get the same error?
20:46 steveoliver yep - about to run now
20:47 cewood joined #salt
20:47 forrest does mariadb use the same conf file names as mysql? I'm not very familiar with it
20:47 steveoliver forrest: yes, i think so
20:47 forrest ok
20:47 steveoliver my.cnf for sure
20:48 steveoliver you can see i'm not using that yet, though
20:48 forrest yea
20:49 blafountain is anyone seeing a error that says __pilliar__ is a boolean?
20:49 blafountain File "/usr/lib/python2.7/dist-packages/salt/modules/config.py", line 119, in option     if value in __pillar__.get('master', {}):
20:49 forrest it looks like a lot of data is getting set from your pillar with the users and groups, and your mariadb/init.sls doesn't start the service, so I assume that's being done elsewhere?
20:49 blafountain AttributeError: 'bool' object has no attribute 'get'
20:51 steveoliver forrest: you're right - i'm not even enforcing a running state on mariadb-server yet
20:51 forrest steveoliver ahh ok
20:52 steveoliver the bigger issue i'm addressing here (in the paste) is that maria installs on the second highstate pass
20:52 forrest oh it takes two runs to install?
20:52 harobed joined #salt
20:52 cdcalef steveoliver just out of curiosity how are you getting the grain for 'roles' on your hosts?
20:53 steveoliver oh shit, i think i see my cmd.run apt-get update is commented out
20:53 harobed joined #salt
20:53 steveoliver grains:roles: in /etc/salt/minion s
20:53 jeddi joined #salt
20:53 steveoliver oh no, nevermind, I am apt-get updating
20:53 steveoliver ok, so ..
20:53 harobed joined #salt
20:54 forrest I mean something is obviously being required that is breaking it
20:54 forrest to require two runs
20:54 steveoliver ok, so i've just gotta clean up my mess? :)
20:54 forrest lol
20:54 steveoliver i think it's pretty clean so far
20:54 forrest Well, a good place to start might be here: https://github.com/circatree/salt-hosting-states/blob/master/mariadb/init.sls#L19
20:54 forrest yea if you know it's just the mariadb state that's requiring two runs
20:54 forrest start with that.
20:55 steveoliver ok we'll that's commented out
20:55 steveoliver should be, rather
20:55 forrest ok
20:55 steveoliver maybe that multiline comment needs a \n before {# and Todo
20:55 forrest so here's the question, if you have that commented out, how is the repo getting refreshed prior to installing mariadb-server?
20:56 forrest my ubuntu foo isn't great, is there anything similar to yum clean all to force a refresh on the data?
20:56 forrest I wonder if it installs that repo, then doesn't refresh, so it doesn't find mariadb until the second run
20:57 steveoliver sorry, long line
20:57 steveoliver side step to the right
20:57 steveoliver apt-get update
20:57 forrest right but that isn't happening currently
20:57 forrest if your comments are true
20:57 steveoliver look at http://paste2.org/vVjDB2wE
20:57 steveoliver those are my two runs
20:58 steveoliver and as you can see my mariadb.sls orders those states differently than they are run
20:58 steveoliver if salt.highstate's comments are true :)
20:58 steveoliver i
20:58 steveoliver i'm probably doing something wrong
20:58 forrest yea, change your cmd.run: to require_in: pkg: mariadb-server
20:59 forrest what release is this on by the way?
20:59 steveoliver 12.10
20:59 forrest lol
20:59 steveoliver no bueno?
20:59 forrest yea there's no ordered states in 0.12.10
20:59 steveoliver no
20:59 forrest so they don't execute in order
20:59 steveoliver ubuntu 12.10
20:59 steveoliver :)
20:59 steveoliver salt .17.0
20:59 forrest oh ok
20:59 steveoliver :)
20:59 forrest well, try the require_in for your cmd real quick
20:59 steveoliver k
21:00 steveoliver cmd.run\n  - require_in: mariadb-server ?
21:00 forrest oh 12.10, thought you were saying 12.1, there is no 12.10 even
21:00 steveoliver :)
21:00 forrest after the - name: apt-get update
21:00 forrest put the require_in
21:00 forrest try that
21:02 steveoliver https://github.com/circatree/salt-hosting-states/blob/master/mariadb/init.sls#L12
21:02 steveoliver (refresh)
21:03 steveoliver (i removed the old package.manged block also)
21:03 forrest oh nope, - require_in: \n - pkg: mariadb-server
21:03 steveoliver so i've got a fresh minion ready
21:03 forrest if you run it like that, it's going to screw up
21:03 steveoliver roles: mysqlserver (will get mariadb state)
21:03 steveoliver ok
21:03 forrest because it will require itself
21:03 steveoliver that's what looked weird to me
21:03 forrest yea
21:03 anti_ joined #salt
21:03 forrest sorry, I was being lazy
21:03 steveoliver is that not what you asked me to do?
21:03 steveoliver forrest++
21:03 forrest you need your require_in to require the package
21:04 forrest right now if you require_in mariadb-server, it's going to loop
21:04 forrest does that make sense?
21:04 steveoliver d'oh!
21:04 steveoliver package and state name are the same
21:04 forrest yea
21:04 steveoliver that a problem?
21:04 forrest no it's not the state name
21:04 forrest it's line 6
21:04 steveoliver right
21:04 steveoliver oh
21:04 forrest so when it runs mariadb-server, you're requireing that a command within that section, require the section
21:04 steveoliver sorry, we're only talking states here
21:04 steveoliver not pkg
21:04 forrest *requiring
21:05 steveoliver right
21:05 forrest yea we ONLY want to require_in, for the pkg
21:05 steveoliver that's what looked weird
21:05 forrest so that the pkg itself requires the command, not the mariadb-server 'section'
21:05 steveoliver oh, under pkg: installed?
21:05 forrest remember how you had it before?
21:05 forrest - require_in:
21:05 forrest - pkg: mariadb-server
21:05 forrest we want it to look like that
21:06 steveoliver cmd.run with that, yeahgot ya
21:06 forrest yea all you have right now is - require_in: mariadb-server
21:07 steveoliver refresh https://github.com/circatree/salt-hosting-states/blob/master/mariadb/init.sls#L13
21:07 forrest Yea that should work I think
21:07 steveoliver k, fresh minion
21:07 forrest cool
21:08 steveoliver highstate? or users.sls then mariadb via states.sls?
21:08 steveoliver high, prob
21:08 forrest when you ran the state.sls solo, did it require two runs?
21:08 steveoliver pretty sure
21:08 forrest It's up to you really, if you wanna do highstate to get a comparison, do highstate
21:08 steveoliver k
21:09 steveoliver bing
21:09 steveoliver that's it, man
21:09 jkleckner joined #salt
21:09 steveoliver you rock
21:09 steveoliver thank you so much
21:10 steveoliver that makes sense
21:10 forrest Awesome!
21:10 forrest Yea np
21:10 Gifflen joined #salt
21:10 steveoliver cuz the template processes the nodes and sets up the dep graph, that sets things straight
21:10 forrest yea usually you don't need requires that often, I'm interested to see what about your if statement causes that.
21:10 forrest exactly
21:10 forrest no updated database = no new packages
21:11 steveoliver but the apt-get update should have run since it was ordered first in the state?
21:11 steveoliver before pkg.installed tried for the new mariadb-server
21:11 forrest yea that's what I'm kind of confused on.
21:11 steveoliver which i why i reached out - i need to look into ordering now, i guess
21:11 forrest unless something about the if loop was causing it to get dropped in below
21:12 forrest yea, you might even open a discussion on the mailing list regarding that
21:12 forrest maybe I'm just missing it
21:12 steveoliver i'll test a few scenarios now
21:12 steveoliver thanks, forrest
21:12 forrest np
21:12 steveoliver i suspect that comment block, for one
21:12 forrest yea if you can try a few scenarios and let me know, that would be awesome
21:12 forrest maybe now it will work fine without the require_in
21:13 steveoliver that's what i'll try now
21:13 forrest cool
21:15 steveoliver well… that seems to work without the require_in
21:15 steveoliver guessing my comment ?
21:15 forrest well, then it's all your fault! :P
21:15 steveoliver always
21:15 forrest maybe yea
21:15 steveoliver trying that
21:16 Voziv Anyone know how I can set pecl to install a beta package? This is my sls for it but it doesn't seem to work; http://hastebin.com/ciripiveno.yml
21:16 DanGarthwaite joined #salt
21:17 Kareem joined #salt
21:18 forrest is it just not installing the beta version voziv?
21:18 forrest As in, can you install the stable?
21:18 davet joined #salt
21:18 Voziv there is no stable version. It keeps trying to install stable. It's not setting the "preferred stability" on pecl.
21:19 Voziv If I'm reading the docs right (http://docs.saltstack.com/ref/states/all/salt.states.pecl.html) I should be setting the key under "pecl.installed" as - preferred_state=beta
21:20 forrest Voziv, yea I'm trying to see what's going on over here: https://github.com/saltstack/salt/blob/develop/salt/states/pecl.py#L70
21:20 Nazca joined #salt
21:20 Nazca joined #salt
21:20 Kareem left #salt
21:21 Voziv what is __salt__ ?
21:22 thelorax123 joined #salt
21:23 anti_ joined #salt
21:23 UtahDave Voziv: __salt__ is a dictionary containing all the salt modules and functions available on the minion
21:25 Voziv gotcha
21:25 dvogt joined #salt
21:28 jhulten joined #salt
21:28 Voziv So am I placing preferred_state in the right place?
21:28 Linz joined #salt
21:29 UtahDave Voziv: looking
21:30 UtahDave Voziv: I'm thinking that what you have is correct. I haven't used the pecl state, but the docs indicate that what you have is correct.
21:31 ajw0100 joined #salt
21:32 Voziv I just added name and it did reflect in the command. Where are salt modules located on a minion?
21:32 Katafalkas joined #salt
21:34 UtahDave salt/modules
21:34 thelorax123 joined #salt
21:34 steveoliver forrest: update:: '- required_in:' /was/ required.
21:34 steveoliver without it, same fail
21:34 forrest really? Interesting
21:35 steveoliver yes
21:35 Voziv UtahDave: /salt?
21:35 jacksontj joined #salt
21:35 forrest I don't see why that would happen simply because of the loop, does something happen when the loop occurs that places anything inside the loop at the bottom of the pile?
21:36 steveoliver the highstate i called that "worked" without the require_in wasn't even targeting minion with the mariadb state
21:36 forrest gotcha
21:37 UtahDave Voziv: it depends on where your os puts site-packages
21:37 Voziv Well I got the pecl extension to install by specifying the name as exactly "xhprof-0.9.4". Not really ideal but until this gets figured out I suppose it'll have to do
21:37 Voziv It's a standalone minion on ubuntu. I couldn't find a salt directory anywhere that would have the default modules and states
21:38 Gareth salt-lint...you are a cruel mistress.
21:39 forrest Voziv, does it actually somehow label it as a beta package if you were to install it yourself?
21:39 adam_1 joined #salt
21:41 zandy joined #salt
21:41 adam_1 Anyone know of a top.sls in pydsl that is online somewhere I can see it?
21:43 forrest adam_1, I do not :\
21:44 JulianGindi joined #salt
21:44 forrest adam_1, have you reviewed the normal renderers documentation for the pydsl to try and swap it over?
21:44 Voziv forrest: nope
21:45 Voziv forrest: The command that should be run is "pecl -d preferred_state=beta install xhprof"  but it runs "pecl install xhprof"
21:45 forrest but is that because there is no 'beta' release available?
21:46 forrest as in it does't understand how to tag it?
21:46 Voziv Thankfully pecl will run "pecl install xhprof-0.9.4" and run it anyways
21:46 adam_1 forrest: I had a working top.sls in yaml that I tried porting over to pydsl.  I've been looking through the docs for pydsl and even the source to try to make sense of it
21:46 forrest gotcha
21:46 Voziv there are only beta releases available
21:46 forrest adam_1, you already checked out http://docs.saltstack.com/ref/renderers/all/salt.renderers.pydsl.html
21:46 forrest Voziv, ok
21:46 forrest I know you said that earlier, just wasn't sure how they are tagged
21:46 Voziv and this is according to the pecl state sls
21:46 ajw0100 joined #salt
21:46 Voziv I don't know why it's not getting the command, and I can't find the state sls on my minion to double check that it has the same as on github
21:47 adam_1 forrest: Yup, that's what I was reading
21:50 forrest adam_1, ok yea then I don't have much more, I haven't really messed with the renderer
21:50 Voziv forrest: I'm pretty sure whatever is on my minion isn't what's on github. github should always have the -d preferred_state=stable parameter
21:50 adam_1 Ok, well thanks anyways
21:51 forrest Voziv, just update your locate database then do locate pecl.py
21:51 khodgson joined #salt
21:51 forrest adam_1, yea, might be worth hanging around, someone else might know
21:51 adam_1 I guess most people just stick with yaml and jinja?
21:51 forrest that code is a month old, so it should be in there
21:51 scott_w joined #salt
21:52 forrest adam_1, that's the standard yea, there are a few guys using the renderer, but the yaml/jinja is a lot easier for people to get started with and up to speed with quickly
21:52 khodgson joined #salt
21:53 jslatts joined #salt
21:54 smccarthy joined #salt
21:54 Voziv forrest: Yeah my version doesn't match. Is there a way to update it? Or am I stuck with the ubuntu ppa?
21:54 forrest for salt itself?
21:54 Voziv yeah
21:55 forrest depends how dangerously you want to live. You could try just updating that file with the new one if this isn't prod, and you back up the file. Or you could try using the bootstrap (again back up first)
21:55 forrest what version did you get from the ppa?
21:55 forrest 0.17.1?
21:56 Voziv how do I check?
21:56 Voziv yeah 0.17.1 (salt-minion --version)
21:56 adam_1 left #salt
21:56 forrest ok, I'd say try to just update that one file then (after making a backup)
21:56 adam_1 joined #salt
21:56 forrest then restart the sale master
21:56 forrest might be worth a shot
21:57 forrest since 0.17.1 is the latest, your options at that point are pretty much to use the bootstrap to try and install the latest from develop, which is always risky
21:57 pipps__ joined #salt
21:58 Voziv If I were to make a folder called modules under /srv/salt would any modules there override whatever is in salt-minion?
21:58 oz_akan_ UtahDave: hi, I hope you are doing fine. we found that salt-cloud should have an option to install a specific salt-minion version
21:59 forrest Voziv, the problem there is that would be for the module, not the state
22:02 tty-1_ joined #salt
22:02 juicer2 joined #salt
22:02 UtahDave Voziv: make    /srv/salt/_modules       any modules you drop in there will overwrite the regular modules after running   salt '*' saltutil.sync_modules
22:02 UtahDave oz_akan_: Yeah, doing great!
22:02 juicer2 joined #salt
22:02 oz_akan_ UtahDave: good to hear that
22:02 oz_akan_ UtahDave: https://github.com/saltstack/salt-bootstrap
22:03 oz_akan_ I think we might modify this to install 0.16.4 for example?
22:03 ajw0100 joined #salt
22:03 UtahDave oz_akan_: Hm. you can do that.  you pass in a   script_args:      option
22:03 willist joined #salt
22:04 tty-1_ joined #salt
22:04 UtahDave so script_args: git v0.16.4
22:04 pipps joined #salt
22:04 oz_akan_ that sounds easy
22:05 oz_akan_ so do we pass script_args with salt-cloud or under provider section?
22:06 oz_akan_ UtahDave: ^^
22:06 UtahDave oz_akan_: Yeah, I'd put that with your profiles.  You might be able to put it in the providers
22:06 oz_akan_ lovely, thanks UtahDave we will test it asap
22:06 UtahDave cool!
22:06 snuffeluffegus joined #salt
22:08 willist hey all - somewhat new to salt and having trouble getting pillar data to update on the minion
22:08 willist i've run saltutil.refresh_pillar
22:08 willist but it doesn't seem to do anything
22:09 UtahDave willist: what's the output of   salt '<minion id>' pillar.items     ?
22:10 forrest also, did you run the saltutil.refresh_pillar from the master, or directly on the minion?
22:12 willist UtahDave: quite a bit, actually - top level items are disabledUsers, interns, master, and usersToAdd
22:12 willist possibly pulling pillar data from somewhere other than where i expect?
22:12 amckinley joined #salt
22:12 sroegner joined #salt
22:12 willist forrest: will try on minion
22:13 willist forrest: from minion: local: None
22:14 UtahDave willist: is the data you're expecting there in the output of pillar.items?
22:14 willist UtahDave: it is not - top.sls only has packages.sls from http://docs.saltstack.com/topics/pillar/index.html
22:15 UtahDave willist: can you pastebin your sanitized   /srv/pillar/top.sls  and any pillar sls files you're using?
22:15 mpanetta UtahDave: It looks like your suggestion is working :)
22:16 mpanetta '/tmp/deploy.sh -c /tmp/ git v0.16.4' is seen in the output of salt-cloud now
22:16 forrest Voziv, any luck?
22:16 UtahDave mpanetta: ah cool!  are you working with oz_akan_?
22:17 mpanetta UtahDave: Yes sir.
22:17 UtahDave cool.
22:17 UtahDave mpanetta: Are you coming out to SaltConf with him?
22:17 mpanetta Not sure on that. :)
22:17 mpanetta I would love to though.
22:17 UtahDave mpanetta: SaltConf will make your wildest dreams come true!!
22:18 UtahDave :)
22:18 forrest lol
22:18 mpanetta Nice sales pitch UtahDave :P
22:18 UtahDave he he.
22:19 gildegoma joined #salt
22:19 heewa Is SaltConf a thing? Or did I miss a joke or something? Link if it's real?
22:19 Gareth heewa: its real.  happening in January.
22:19 UtahDave It really is going to be great, though. They haven't made final decisions on the speakers, but it looks like Nasa is going to be presenting, oz_akan_ might be as well
22:20 UtahDave heewa: saltconf.com
22:20 icarus joined #salt
22:21 UtahDave We've got like twice as many talk proposals as we have slots, so it's been hard choosing.
22:21 heewa $900?!?!? Jesus.
22:23 willist UtahDave: http://pastebin.com/D7gbqKvz
22:23 alunduil joined #salt
22:23 oz_akan_ UtahDave: good to hear that
22:23 oz_akan_ I mean having so many proposals
22:24 oz_akan_ UtahDave: when can I expect to hear about the final decision?
22:24 Voziv forrest: Went home :P  I'm thinking I'm going to use the full package name for now until the next version comes out
22:25 UtahDave oz_akan_: Yeah.  I'm really looking forward to it.
22:25 forrest Voziv, ahh ok
22:26 UtahDave oz_akan_: I'm not sure exactly.  Let me go find out when they said they were going to announce the talks.
22:26 Gareth heewa: thats not badly priced, compared to a lot of training conferences.
22:26 oz_akan_ UtahDave: tks
22:26 willist UtahDave, forrest: tried clearing the cache and updating pillar data on minion again - didn't do anything
22:26 willist it's almost like the pillar top file is being ignored
22:27 forrest that's weird
22:27 UtahDave willist: run    salt '*' saltutil.refresh_pillar -l debug
22:27 UtahDave oz_akan_: Looks like they're shooting for the end of next week
22:28 oz_akan_ UtahDave: great, tis
22:28 oz_akan_ tks
22:28 scott_w joined #salt
22:29 willist UtahDave: anything in particular i should look for - appears to be using the correct master config
22:30 UtahDave np, oz_akan_!
22:30 UtahDave willist: are you seeing any errors or stacktraces?
22:30 willist UtahDave: would that show up as [ERROR] ?
22:31 willist UtahDave: no stacktraces
22:31 UtahDave possibly.
22:31 heewa Still can't figure out m2crypto problems. Any way of completely disabling crypto? We need to get moving on using salt, we can't just not configure our servers for several more days while we try to figure this out.
22:31 UtahDave can you pastebin the sanitized output of    salt '*' pillar.items   ?
22:33 UtahDave heewa: what OS are you on and how are you installing Salt?
22:33 jslatts joined #salt
22:35 heewa UtahDave: ubuntu 12.04, installed from apt ppa, then on top of that from pip from github on tag v0.17.1
22:36 mwillhite joined #salt
22:36 UtahDave heewa: Hm.  That's the same setup I use.  Let me test here.
22:36 UtahDave heewa: what's the exact pip command you're using?
22:38 heewa sudo pip install -e 'git+https://github.com/salt/salt.git@v0.17.1#egg=salt-minion'
22:39 UtahDave k, just a minute
22:40 patyx7 joined #salt
22:41 forrest willist, when you say you cleared the cache, do you mean you removed the cache dir and restart the master? Or minion? or what
22:41 Drawsmcgraw1 joined #salt
22:41 zandy joined #salt
22:41 willist forrest: salt-run cache.clear_all tgt='<boxname>'
22:41 willist from master
22:41 patyx7 left #salt
22:43 jhulten joined #salt
22:44 xmltok joined #salt
22:45 forrest can you check /var/cache/salt on the master to see if anything seems off in there willist? I can't remember if the pillar data gets pushed to that directory or not
22:45 forrest after looking through that output from before of course
22:47 Drawsmcgraw joined #salt
22:49 willist forrest: moved /var/cache/salt/master to master.old and recreated master - ran another refresh of pillar and unexplained data persists - data in my top.sls still not showing
22:49 halfss_ joined #salt
22:49 forrest did you restart the master after you moved the chache dir?
22:50 forrest *cache
22:50 forrest man I cannot type
22:50 willist forrest: i belive so - let me check bash history :)
22:50 forrest ok cool
22:52 halfss_ joined #salt
22:54 willist forrest: yes, restarted after clearing master
22:54 forrest ok something is odd, so is this all data stored in one pillar?
22:54 forrest that isn't getting updated on the minions?
22:56 willist forrest: i have no idea where all the data is
22:56 willist my top.sls has 1 entry
22:57 forrest ok, what about your pillar.sls
22:57 willist and that entry is the packages.sls from the walkthrough
22:57 forrest because you're saying the pillar data you have, isn't getting pushed to the minion?
22:57 xmltok is a release date set for 0.17.2? i had a bug fixed in a pull request today and im wondering when i may see it
22:57 forrest xmltok, they're trying to release it this week
22:57 willist forrest: should that be under /srv/pillar in a default config?
22:58 forrest the pillar file should be under /srv/pillar yes
22:58 higgs001 joined #salt
22:58 hazzadous joined #salt
23:00 kermit joined #salt
23:01 forrest This isn't perfect willist (doesn't use some cool new functionality) but it gives you an idea of the directory structure: https://github.com/gravyboat/hungryadmin-sls
23:01 forrest maybe that will help?
23:03 sgviking joined #salt
23:04 pentabular joined #salt
23:04 pentabular left #salt
23:05 willist forrest: well, i feel like an idiot
23:05 forrest lol
23:05 forrest did you just have to move the pillar.sls?
23:05 willist i've been editing a file in /srv/salt/pillar the whole time thinking I was in /srv/pillar
23:06 forrest haha
23:06 forrest Well, you've got it now :P
23:06 forrest is it working now that you've got the file in the right spot?
23:06 willist forrest: yep, and the unexplained data is magically explained now that i'm looking at the right top.sls
23:06 forrest Awesome!
23:07 pentabular joined #salt
23:08 pipps_ joined #salt
23:08 seb` joined #salt
23:09 jhulten joined #salt
23:09 dvogt joined #salt
23:16 forrest willist, were you following through the docs when you made that error?
23:16 forrest and if so, which docs
23:16 forrest the getting started?
23:17 cdcalef joined #salt
23:26 bemehow joined #salt
23:26 jhulten joined #salt
23:29 pentabular joined #salt
23:32 pentabular left #salt
23:39 Nazzy joined #salt
23:41 zandy joined #salt
23:42 jhulten joined #salt
23:43 mapu joined #salt
23:44 halfss_ joined #salt
23:45 shuss joined #salt
23:45 heewa UtahDave: Coworker found out this is the issue hitting us: https://github.com/saltstack/salt/issues/8402
23:49 bhosmer joined #salt
23:49 xt joined #salt
23:53 halfss_ joined #salt
23:56 jhulten joined #salt
23:56 thelorax123 joined #salt
23:58 pentabular joined #salt
23:59 jimallman joined #salt
23:59 scott_w joined #salt

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