Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2013-09-09

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

All times shown according to UTC.

Time Nick Message
00:15 lineman60 joined #salt
00:22 pewpewarrows joined #salt
00:22 zzzirk joined #salt
00:32 dthom91 joined #salt
00:33 whit joined #salt
00:36 mwillhite joined #salt
00:39 sixninetynine joined #salt
00:42 oz_akan_ joined #salt
00:47 log0ymxm joined #salt
00:54 faldridge joined #salt
01:02 mgw joined #salt
01:15 mgw joined #salt
01:28 whit joined #salt
01:30 faldridge joined #salt
01:32 mgw joined #salt
01:35 liuyq joined #salt
01:47 liuyq joined #salt
01:47 sssslang joined #salt
01:48 sssslang hi all, how to sync a directory using salt? state.recurse seems not preserve symbol links and file modes.
01:50 bui_ joined #salt
01:53 mgw joined #salt
01:54 Nexpro joined #salt
01:57 m_george|away joined #salt
01:58 Ryan_Lane joined #salt
02:01 m_george left #salt
02:01 log0ymxm joined #salt
02:08 sgviking joined #salt
02:22 forrest sssslang, the file mode can be set with the file_mode value, but by default it's not going to copy down the permissions unless you specify them. Symbolic links aren't going to get copied at all (how would salt know what the target is, and if it exists).
02:24 sssslang forrest: But how to do this if I know my servers all run linux? I don't think manually assigned every file is a good option.
02:25 forrest are you trying to recurse the entire filesystem?
02:25 sssslang no, just a lot of dozens files and dirs.
02:26 dthom91 joined #salt
02:26 pdayton joined #salt
02:27 forrest so for the directories/subdirectories, you could use recurse within fiile.directory, that's equivalent to mkdir -p
02:27 forrest and you can set what the values are supposed to be on those subdirs
02:31 sssslang thus, I should set directories one by one?
02:32 forrest well you could use the recurse option in file.directory to recurse down through something like /srv/test/test1/test2/test3/test4
02:32 forrest so it will create all of those with specific ownership from a single command
02:32 Jahkeup joined #salt
02:33 forrest I'm not 100% sure, but what you might try to do is use the file.recurse to download the entire tree with certain defaults, then specify the other 'unique' values
02:33 forrest but I'm not sure if salt will consider that as a duplication
02:38 sssslang maybe making a tar is considerable at this point :-/
02:38 forrest booooooo
02:43 UtahDave joined #salt
02:50 sssslang forrest: anyway, thank you :-)
02:51 forrest sssslang, yea np, I wish I had a better answer. I haven't used that recurse function before to copy a ton of files over :\
02:51 UtahDave sssslang: are you referring to file.recurse?
02:52 sssslang forrest, UtahDave: I'm using cfengine before, I thought file.recurse could keep symbol links and file modes. but it doesn't.
02:56 UtahDave sssslang: file.recurse can set the mode
02:56 forrest I think he has different modes on specific files
02:59 sssslang yes, some files need special modes.
02:59 sssslang and some are links need to preserve.
02:59 Gwayne joined #salt
03:08 pass_by_value sssslang: Do you have the option to version those files and directories using git and doing a git pull --rebase during manipulation time?
03:09 forrest UtahDave, can you use file.recurse to pull down all the files, then declare a file and give it certain permissions again if it lives in that directory of files? Or will Salt not like that?
03:15 sssslang pass_by_value: some of the files are managed by svn.
03:17 pdayton joined #salt
03:22 lineman60 joined #salt
03:24 xl1 joined #salt
03:24 xl1 left #salt
03:25 xl1 joined #salt
03:26 pass_by_value sssslang: Can't say I have ever committed any repo managed by svn, or it's parent dirs to git. But thinking about doing so makes me feel like I am looking at Mordor with the intention of entering it.
03:27 robertkeizer joined #salt
03:28 UtahDave forrest: yeah, you can do that if you order it correctly
03:29 oz_akan_ joined #salt
03:31 liuyq joined #salt
03:33 stanchan joined #salt
03:36 xl1 joined #salt
03:42 oz_akan_ joined #salt
03:51 stanchan joined #salt
03:52 jalbretsen joined #salt
03:57 liuyq joined #salt
04:17 xpen joined #salt
04:34 bhosmer joined #salt
04:36 xpen_ joined #salt
04:45 baniir joined #salt
04:47 xpen joined #salt
05:18 sparc_ joined #salt
05:25 malinoff joined #salt
05:27 Rudd-O joined #salt
05:27 Rudd-O hello everybody.  I am trying to extend: a declaration in two different SLSes, to add two different groups in the different SLSes to a user
05:28 Rudd-O but for some reason only one of the SLSes successfully extends: its group, the other extension does not get merged properly
05:30 malinoff Rudd-O: Can you pastebin your states?
05:31 Rudd-O malinoff: yes
05:31 Rudd-O gimme a sec
05:34 Rudd-O malinoff:  https://gist.github.com/Rudd-O/6491847
05:34 Rudd-O there you go
05:36 Rudd-O the expected outcome is for groups to be [plone3, plone4, adm, wheel]
05:36 Rudd-O the actual outcome is just plone3
05:36 Rudd-O highly confusing
05:37 Rudd-O the three lists are not getting merged like I expected
05:39 forrest Rudd-O, can you remove plone3.sls from your top.sls and see if it adds plone4?
05:39 Rudd-O forrest: verified, yes it does that.
05:39 Rudd-O see, it's like one of the groups: list wins, whatever happens to be the "latest" or the "earliest" to get to extend rudd-o, is the winner
05:39 forrest yea
05:41 Rudd-O how can I get the same behavior then?
05:41 forrest I bet the extend is actually overwriting the groups
05:41 Rudd-O indeed and for sure
05:41 Rudd-O given the deep merging behavior it exhibits for dictionaries, I would have expected extend to merge lists too
05:42 malinoff Rudd-O: What do you want exactly? You can overwrite groups, you can store groups in pillars, may be something else
05:42 xpen joined #salt
05:42 Rudd-O well you see, this group in particular is present only on the machine running plone3, and that other group is only present on the machine running plone4
05:43 Rudd-O so far so good, it works in that case (except for the adm and wheel groups in the base users.sls)
05:43 forrest I'd say just add the adm and wheel groups to plone3/4
05:43 Rudd-O forrest: that won't allow rudd-o to sudo and read journalctl -a
05:43 Rudd-O malinoff: so now I am setting up a machine that is going to be a testbed for both plone4 and plone3 integrated with varnish together
05:44 Rudd-O when I run this thing on the machine that has plone3.sls and plone4.sls, I get this behavior
05:44 forrest I see
05:44 Rudd-O which is problematic because I need the rudd-o user to be member of BOTH plone3 and plone4 groups
05:45 malinoff Rudd-O: Try to specify an order in extending state
05:45 Rudd-O deets pls :-)
05:45 Rudd-O I am not sure how ordering will fix this issue
05:46 malinoff Me too, but it's not so difficult to try
05:47 Rudd-O ordering will simply change which SLS "wins" at wrongly extending the user
05:47 Rudd-O https://github.com/saltstack/salt/issues/7124
05:49 Rudd-O I guess I will wing it with groupadd and shit
05:49 malinoff Right. First extending will do : [1,2,3] -> [1,2,3,4], but the second one will: [1,2,3] -> [1,2,3,5] instead of [1,2,3,4] -> [1,2,3,4,5]
05:49 forrest yup
05:50 malinoff Tricky issue
05:50 forrest it's just that extend isn't really extending
05:50 malinoff forrest: Yeah, i'll raise an issue about that.
05:50 forrest Rudd-O just put in 7124
05:50 forrest which seems to address most of that
05:50 forrest might be better just to comment
05:51 malinoff forrest: Ok
05:51 forrest but yea, I mean if extend doesn't extend, it shouldn't be called extend
05:51 forrest I understand it extends the state, but it should extend content within the state too, or there needs to be a .append item
05:52 malinoff Rudd-O: I think, the best approach is to store all needed groups in pillars, and specify them especially for the minion
05:54 Furao joined #salt
05:54 Katafalkas joined #salt
05:55 forrest I agree malinoff
05:56 forrest but Rudd-O's method is something you could expect to see.
05:56 malinoff forrest: Yes, totally
05:57 linjan joined #salt
06:01 davidone joined #salt
06:07 Katafalkas joined #salt
06:08 xet7 joined #salt
06:20 arapaho joined #salt
06:21 middleman_ joined #salt
06:31 Rudd-O malinoff: in reply to: "First extending will do : [1,2,3] -> [1,2,3,4], but the second one will: [1,2,3] -> [1,2,3,5] instead of [1,2,3,4] -> [1,2,3,4,5]"
06:31 Rudd-O no, actually, if you have 123 and you extend with 4, what happens is 4, not 1234.
06:31 Rudd-O that's pretty bad lol
06:31 Rudd-O :-D
06:31 Rudd-O but that's okay at least as a temporary measure I used usermod to fix the issue
06:31 malinoff lol whut
06:32 gildegoma joined #salt
06:32 Rudd-O exactly
06:32 malinoff Rudd-O: i've been extending states a lot, but i've never met this
06:33 Rudd-O me neithr
06:33 Rudd-O i am equally baffled
06:33 malinoff Probably, it's just a bug right in *user* module
06:34 Rudd-O probably not, unless the module does something unique with respect to "extends", which seems like a global thing that modules don't care for
06:34 Rudd-O remember extend happens during rendering of the states
06:34 Rudd-O way before module functions are consulted about the actual parameters of each invocation
06:41 malinoff Rudd-O: Yes, 'extend' is something global, but it must use local data structures
06:53 balboah_ joined #salt
06:57 xpen joined #salt
07:01 Damoun joined #salt
07:03 ronc joined #salt
07:05 LucasCozy joined #salt
07:13 sfello joined #salt
07:38 vaxholm joined #salt
07:42 faldridge joined #salt
07:50 ml_1 joined #salt
07:53 felskrone joined #salt
07:53 xpen joined #salt
08:05 dthom91 joined #salt
08:10 bhosmer joined #salt
08:11 bejer joined #salt
08:12 xpen joined #salt
08:12 tomviner joined #salt
08:13 malinoff joined #salt
08:14 Katafalkas joined #salt
08:17 backjlack joined #salt
08:21 felixhummel joined #salt
08:33 vaxholm joined #salt
08:40 nocturn I'm trying to watch I directory recursively by using globbins
08:40 nocturn /somedir/*
08:41 nocturn but salt says: The following requisites were not found: /somedir/
08:41 nocturn I have an entry to copy the directory contents recursively from the master
08:41 nocturn What am I missing?
08:43 matanya joined #salt
08:48 NV joined #salt
08:51 ange hello
08:54 Furao joined #salt
08:55 ange I am looking into using salt-api  and salt-cloud, is the later usable from the former ?
09:06 xpen_ joined #salt
09:13 Furao ange: these are separate things they don't interact with each others
09:13 Furao but I used salt-api to build a webui on top of salt-cloud
09:13 s0undt3ch joined #salt
09:14 sfello joined #salt
09:23 az87c joined #salt
09:23 ezraw_away joined #salt
09:23 swa nocturn: you have to specify the full path e.g. : like - file: /etc/squid/ACL/*
09:24 monokrome joined #salt
09:29 nocturn swa, I found out how to do it
09:29 nocturn I have a resource called /somedir with file.recurse
09:30 nocturn if I just watch that, it works recursively
09:30 swa yeah you have to have a resource by the same name
09:30 nocturn Yes
09:30 nocturn that was my error
09:31 swa i did the same mistake back then :)
09:44 ggoZ joined #salt
09:45 Ryan_Lane joined #salt
09:47 ggoZ1 joined #salt
09:50 carlos joined #salt
09:54 xpen joined #salt
10:05 TheCodeAssassin joined #salt
10:07 NV joined #salt
10:16 carlos joined #salt
10:21 Ryan_Lane joined #salt
10:30 tomeff joined #salt
10:40 __number5__ joined #salt
10:51 Ryan_Lane joined #salt
10:55 swa grains.items won't always return the item "master" ?
11:01 swa hmm got it.. will appear only if specified outside of main config file.. hmm
11:01 swa nevermind :(§
11:09 carlos joined #salt
11:12 oz_akan_ joined #salt
11:22 Ryan_Lane joined #salt
11:24 linjan joined #salt
11:27 bhosmer joined #salt
11:33 ml_1 joined #salt
11:42 fredvd joined #salt
11:48 Koma joined #salt
11:52 efixit joined #salt
11:53 dthom91 joined #salt
11:53 LucasCozy joined #salt
11:53 anteaya joined #salt
11:54 Furao joined #salt
11:58 gildegoma joined #salt
11:59 carlos joined #salt
12:07 blee joined #salt
12:09 nocturn Can I create functions/templates like puppet defines in salt?
12:10 nocturn I have such a custom function to copy apache configs, enable them, create log dirs etc
12:10 LucasCozy joined #salt
12:12 efixit joined #salt
12:15 mwillhite joined #salt
12:16 unicoletti_ joined #salt
12:19 efixit joined #salt
12:20 matanya joined #salt
12:20 jslatts joined #salt
12:26 efixit joined #salt
12:38 kenbolton joined #salt
12:51 juicer2 joined #salt
12:52 linjan joined #salt
12:53 StDiluted joined #salt
12:53 Ryan_Lane joined #salt
12:54 pdayton joined #salt
12:57 djn how come 0.16.4 was released? what delays 0.17, halite?
13:00 ipmb joined #salt
13:00 copelco joined #salt
13:01 aurigus joined #salt
13:01 xt halite isn't part of salt
13:03 Gifflen joined #salt
13:03 sgviking joined #salt
13:05 brianhicks joined #salt
13:06 Jahkeup joined #salt
13:08 racooper joined #salt
13:09 joehh djn: 0.16.4 is a bugfix release for the 0.16 series
13:10 joehh I believe 0.17 is coming soon, but there were bugs they wanted to fix for those who wanted stability
13:10 xt joehh: there are people that don't want stability ? :-)
13:10 joehh true :)
13:11 joehh good point
13:11 joehh greater stability
13:11 joehh there are quite a few around here who recommend (rightly or wrongly) against 0.x.0 releases of salt
13:11 joehh for me they have been pretty good
13:11 xt rightly :-)
13:16 Ryan_Lane joined #salt
13:16 efixit joined #salt
13:19 xinkeT joined #salt
13:23 efixit joined #salt
13:23 oz_akan_ joined #salt
13:25 nocturn Anyone know if salt as an alternative for puppet's define?
13:25 oz_akan_ joined #salt
13:27 toastedpenguin joined #salt
13:27 toastedpenguin left #salt
13:29 Khollowa_ joined #salt
13:31 jeffasinger joined #salt
13:31 cron0 joined #salt
13:38 pdayton joined #salt
13:40 djn joehh: thanks for the info, just asking because the 0.17 release was actually due a while ago... not that I'd mind stability, I'm just looking forward to some of the features to be introduced with 0.17 ;)
13:41 Corey nocturn: How do you mean?
13:42 joehh djn: Last I heard on 0.17.0, they were aiming for this friday, but that was before the weekend and I'm sure tom and the
13:42 joehh others want to get it right before releasing.
13:42 nocturn Corey, like the puppet defines I have here: http://pastebin.com/9xFp7MeS
13:42 Corey Yeah, I'd rather it be released "when it's ready" than according to some marketing guy's schedule.
13:43 nocturn I can call them from a manifest like:  apache_config { '000-default': }
13:43 Khollowa_ joined #salt
13:43 joehh Corey: completely agree on that
13:47 mgw joined #salt
13:48 efixit joined #salt
13:51 mgw joined #salt
13:52 badluck joined #salt
13:54 badluck Hi, I'm trying to do the "getting started" guide. When I do "salt '*' state.sls vim" I go "Data failed to compile: The function "state.sls" is running as PID 2587 and was started at 2013, Sep 09 15:48:01.269224 with jid 20130909154801269224". I tried to reboot the client, but same probleme. Can I do something?
13:56 TheCodeAssassin joined #salt
13:57 bhosmer badluck: you could just kill the pid maybe?
13:57 bhosmer What does your vim state look like?
13:57 imaginarysteve joined #salt
13:58 Drawsmcgraw joined #salt
13:58 Drawsmcgraw Is there a way to force the Salt master to update the git_fs cache?
13:58 djn badluck: dunno what caused this, but if the job is not exiting it's hung up somewhere or the client does not respond... do a 'kill 2587' and try again. If the same thing happens, your state is faulty or your master-minion setup does not work
13:58 badluck I'm really now with linux. In fact, I'm just trying linux at the same time i'm truing salt. I'm not surhow to kill the PID.
13:58 Drawsmcgraw Right now I'm deleting the cache from /var/cache to force it.
13:58 badluck vim.sls => vim: pkg.installed
13:59 badluck <badluck> I'm really now with linux == I'm really "new"
13:59 djn badluck: as stated above, just do 'kill 2587' as the user salt-master is running (probably root, so use sudo)
14:00 badluck thanks
14:00 badluck I will try
14:00 djn where 2587 is the pid from your error
14:00 bhosmer But I think there is still something wrong with your vim state that caused it initially though.
14:00 djn yes probably
14:00 badluck bash: kill: (2587) - No such process in client and master
14:00 badluck humm PID change...
14:00 bhosmer Can you use a pastebin and paste your state? Spaces matter.
14:02 Drawsmcgraw Well.... clearing the cache *and* restarting the salt master. I wonder if just restarting it will cause it to refresh the git_fs stuff....
14:02 badluck super... each time I run "salt '*' state.sls vim" the PID change each time. If I trey "Kill pid" I got no such proccess...
14:03 badluck retarting the master fixed it
14:03 badluck thanks
14:03 bhosmer badluck: let's fix your vim state now.
14:04 badluck fix my vim state? I copy/paste the file from the "getting started". I restarted the master and Seem fine rith now. I still have something to fix?
14:05 bhosmer Your original error said that it failed to compile right?
14:06 badluck yes, but, now, it work...
14:06 KennethWilke joined #salt
14:06 badluck than I will continue the getting started. thanks
14:06 bhosmer Did you get a return that said vim was installed or updated?
14:06 badluck Comment:   Package vim is already installed
14:06 badluck yes, everything seams fine
14:07 bhosmer Was it already there before you ran state.sls? I just want to make sure you understand the YAML formatting.
14:08 badluck no
14:09 bhosmer Okay, just remember two spaces not tabs. That is very important.
14:09 badluck I restarted the master, reran "salt '*' state.sls vim" and I got " State: - pkg : Name:      vim :   Function:  installed : Result:    True :  Comment:   Package vim is already installed"
14:10 badluck sorry, was not able to add new line
14:10 badluck I add ":" instead of new line
14:11 nocturn I created a pilar called sites with two values, but {% for site in pillar['sites'].iteritems() %} gives me UndefinedError: 'list object' has no attribute 'iteritems', what did I miss?
14:11 bhosmer It looks like you have it then. You could try pkg.removed to make sure your state file looks good.
14:12 badluck ok, I will try this
14:12 nocturn I needed to remove iteritems
14:12 nocturn ok
14:12 bhosmer And then change it back to pkg.installed and run it again.
14:12 nocturn phew
14:14 kermit joined #salt
14:14 badluck <bhosmer> I tried to change "pkg.removed" and I got the same error. I rebooted the master and now it seems fine. Do I need to reboot the master each time I change a .sls?
14:14 bhosmer No, you shouldn't. Are you sure you don't have some mixed spaces, tabs etc in there?
14:15 bhosmer I wouldn't worry too much about killing the process. SALT will hopefully take care of all that for you.
14:16 badluck it's space no tabs. I retried "pkg.installed" after I "removed" it. Works the fist time "Comment:   The following packages were installed/updated: vim." , I did not need to restart the master... ok, then I proably did something, but seems to work. thanks
14:17 bhosmer But you still got a "data failed to compile"?
14:17 badluck no
14:17 bhosmer Okay, good. Enjoy.
14:22 teebes joined #salt
14:23 toastedpenguin joined #salt
14:24 tyler-baker joined #salt
14:24 Furao joined #salt
14:27 StDiluted joined #salt
14:34 baniir joined #salt
14:36 Lue_4911 joined #salt
14:41 danielbachhuber joined #salt
14:42 pakdel_ joined #salt
14:43 pakdel_ Hi all
14:43 zzzirk joined #salt
14:44 pakdel_ could the top.sls file be templated?
14:44 efixit joined #salt
14:44 pakdel_ Can I use Jinja in top.sls files?
14:44 KennethWilke pakdel_: yeah, i think you cna
14:44 KennethWilke can*
14:45 pakdel_ To my knowledge, top.sls files are arsed on the master. are they sent to  minions?
14:45 pakdel_ actually the question is:
14:45 pakdel_ Can I use Grains and Pillars in top.sls ?
14:46 KennethWilke hmm, i would imagine you could, but i have not tried it myself
14:46 mgw joined #salt
14:46 abe_music joined #salt
14:47 cedwards yes, you can target minions in the top.sls using grains or pillars
14:48 [diecast] joined #salt
14:49 pakdel_ Great! It's time to try my recipe :P
14:49 pakdel_ Thanks
14:51 jaequery joined #salt
14:54 lineman60 joined #salt
14:58 Katafalkas Hey, Is there some way to see print statements while developing modules ?
14:58 Katafalkas for debugging ..
14:59 jbunting joined #salt
14:59 KennethWilke Katafalkas: i've used the logger module for that, then ran the salt-minion manually with `salt-minion -l debug`
15:01 hhoover joined #salt
15:01 Jahkeup joined #salt
15:02 lineman60 joined #salt
15:02 Katafalkas cheers. it does work. KennethWilke. for some reason if u run it as daemon with debug mode on - it does not log the print. but running it manually - it does print. anyway. danke <3
15:07 faldridge joined #salt
15:09 ckao joined #salt
15:10 dthom91 joined #salt
15:12 gmoro joined #salt
15:14 UtahDave joined #salt
15:15 mgw joined #salt
15:17 sfello folks - is there a way to use rendering of files to add e.g. all the names of all the minions with a certain grain to a file that gets written to all of those minions (so they can have a reference to each other, for example)?
15:18 m_george|away joined #salt
15:19 badluck can I do something like this in a .sls "/tmp/grains['host']/testCopyfolder:"? Right now, it create a folder name "grains['host']" but I want to have a folder name with the minion's hostname
15:21 sfello badluck: something like: http://docs.saltstack.com/topics/tutorials/states_pt3.html#calling-salt-modules-from-templates ?
15:21 devinus joined #salt
15:21 badluck probably, thanks, I will read this right now
15:21 sfello like: {{ grains['host'] }}
15:21 m_george left #salt
15:21 hhoover left #salt
15:24 UtahDave sfello: Hey, you can do a file.managed   with template: jinja and then use the salt mine or the peer interface to fill that file with the list of minions
15:25 sfello UtahDave: ok, cool - I had a look at the salt mine stuff but it was pretty vague - do you know of any good example files/tutorials about? hard to google due to the overloading of names!
15:27 badluck sfello> like: {{ grains['host'] }}: Thanks this is exaclty what I wanted
15:27 sfello badluck: np!
15:28 UtahDave sfello: Have you found this doc yet? http://docs.saltstack.com/topics/mine/index.html
15:28 UtahDave and the execution module: http://docs.saltstack.com/ref/modules/all/salt.modules.mine.html#module-salt.modules.mine
15:28 jeff__ joined #salt
15:30 sfello UtahDave: Yes - looked at them both, but couldn't seem to get the minion to automatically send the mine functions to the master. Tried calling a few commands on the master like salt '*' mine.get '*' network.interfaces, but again, it was pretty confusing as to which out of send/get actually do the thing. I was trying to find out where it (locally) stores the received info.
15:31 sfello UtahDave: am incredibly novice with salt so probably i'm missing something obvious
15:31 UtahDave sfello: The salt mine is pretty new, so the docs do need some attention.
15:32 sfello hm.
15:32 UtahDave just a sec
15:32 dthom91 joined #salt
15:33 akio /names
15:34 UtahDave sfello: OK, so if you want to manually send data to the mine from the master run:   salt \* mine.send test.ping
15:34 SunSparc joined #salt
15:34 sfello so, is that sending the results of test.ping to *?
15:35 UtahDave then to access that data:   salt '*' mine.get '*' test.ping
15:35 UtahDave back ina minute.  conference call
15:35 sfello UtahDave:  okeydoke. I think the .get is the confusing one
15:36 teskew joined #salt
15:37 forrest joined #salt
15:38 unicoletti left #salt
15:38 faldridge joined #salt
15:39 baniir joined #salt
15:40 redbeard2 joined #salt
15:41 jalbretsen joined #salt
15:41 jergerber joined #salt
15:44 BrendanGilmore joined #salt
15:47 micah_chatt joined #salt
15:57 zwevans joined #salt
15:58 fxhp how do I perform a job run lookup on a masterless-minion (salt-run seems to need a salt-master daemon)
16:04 all_the_fisch joined #salt
16:05 mohae joined #salt
16:07 devinus joined #salt
16:08 forrest did you already try salt-call?
16:09 KennethWilke does the top.sls pillar file get cached?
16:10 KennethWilke i'm having this odd problem where it says it can't load sls files that are there when i do a state.sls run, but when i do a pillar.items call the pillar data is there from those files
16:10 mwillhite joined #salt
16:11 KennethWilke i tried renaming the files and changing the top.sls file to match, but it still complains about the old name
16:11 Koma joined #salt
16:11 forrest did you already do salt '*' state.clear_cache?
16:12 KennethWilke nope, wasn't aware that existed
16:12 forrest technically the cache is disabled by default
16:12 KennethWilke i did try stopping the master, rm -rf'ing the salt cache dir and starting the master back up
16:12 forrest so that shouldn't do anything
16:12 forrest hmm, yea that should have addressed it
16:12 forrest trashing the cache dir was gonna be my next suggestion
16:13 cmthornton joined #salt
16:13 KennethWilke hmm, ill make sure i didn't do anything funky with pillar in my master config
16:13 KennethWilke I probably did though, that's the only reasonable thing i can think of
16:14 KennethWilke this pebkac lives somewhere
16:14 jdenning joined #salt
16:15 KennethWilke damn,  nothing in there but pillar_roots set to where it should be
16:16 lineman60 joined #salt
16:22 TheCodeAssassin joined #salt
16:22 mmilano joined #salt
16:25 Furao joined #salt
16:26 druonysus joined #salt
16:26 lineman60__ joined #salt
16:29 forrest :\
16:37 troyready joined #salt
16:37 forrest KennethWilke, you are running the latest version right?
16:37 KennethWilke 0.16.3
16:37 forrest ok
16:37 chrisgilmerproj joined #salt
16:37 KennethWilke i got it fixed via UtahDave
16:37 forrest what was it?
16:37 KennethWilke i think i needed to saltutil.refresh_pillar
16:38 forrest oh the pillar doesn't refresh when you restart the service?
16:38 forrest or clear the cache?
16:38 KennethWilke but i inadvertantly fixed it by bouncing stopping the salt minions and running the minion manually
16:38 KennethWilke stopping* not bouncing
16:38 KennethWilke but i suspect bouncing woulda fixed it
16:38 chrisgilmerproj left #salt
16:38 KennethWilke but it was the minions i needed to restart, not the master
16:38 forrest ahh right because they had previously downloaded the pillar data
16:38 forrest makes sense
16:38 KennethWilke yeah
16:39 KennethWilke i tried the sync_all but forgot that was just to sync modules
16:39 KennethWilke doesn't touch pillar stuff
16:39 forrest is there a sync for pillar?
16:39 KennethWilke saltutil.refresh_pillar
16:39 forrest makes sense
16:39 KennethWilke yeah
16:40 KennethWilke i still find it odd that pillar.items showed that data
16:40 KennethWilke but that state.sls had problems with it
16:42 lineman61 joined #salt
16:45 mgw joined #salt
16:47 UtahDave left #salt
16:48 UtahDave joined #salt
16:48 scotticus left #salt
16:48 robertkeizer joined #salt
16:49 nickaz joined #salt
16:49 nickaz Hello everybody
16:49 nickaz I got a problem with eauth in salt
16:51 nickaz i execute this salt '*' -a pam test.ping, and i give my creditentials, then i got this error Failed to authenticate, is this user permitted to execute commands?
16:51 nickaz on the master log i got [INFO    ] Clear payload received with command publish [WARNING ] Authentication failure of type "eauth" occurred.
16:52 nickaz im working on this issue since 2 days
16:58 bitz joined #salt
17:03 faldridge joined #salt
17:05 Jahkeup_ joined #salt
17:09 baniir joined #salt
17:14 nickaz in /var/log/secure i got this error : pam_securetty(login:auth): cannot determine user's tty
17:14 nickaz when i execute salt '*' -a pam test.ping
17:14 UtahDave nickaz: what version of Salt are you running?
17:16 nickaz [root@localhost ~]# salt-master --versions-report            Salt: 0.16.3          Python: 2.6.6 (r266:84292, Jul 10 2013, 22:48:45)          Jinja2: unknown        M2Crypto: 0.20.2  msgpack-python: 0.1.13    msgpack-pure: Not Installed        pycrypto: 2.0.1          PyYAML: 3.10           PyZMQ: 2.2.0.1             ZMQ: 3.2.3
17:16 nickaz [root@localhost ~]# salt-minion --version salt-minion 0.16.3
17:17 UtahDave ok, can you pastebin your external_auth   section of your master config?
17:19 nickaz http://pastebin.com/x3jEvFuv
17:20 devinus joined #salt
17:20 jbunting joined #salt
17:24 carmony ping: UtahDave
17:25 UtahDave nickaz: ah, that's your problem.  Pam won't authenticate the root user
17:25 UtahDave carmony: pong!
17:25 carmony UtahDave: So I just got accepted to give a salt tutorial at True North PHP
17:25 UtahDave nice!
17:27 UtahDave when's True North?
17:28 carmony http://truenorthphp.ca/index.php -- Nov 7th - 9th
17:28 nickaz ok so i create a user salt with passwd salt, its gave me the same error  [INFO    ] Clear payload received with command publish [WARNING ] Authentication failure of type "eauth" occurred.
17:28 bhosmer joined #salt
17:29 nickaz but its didnt log anything in my var/log/secure
17:29 Ryan_Lane joined #salt
17:29 badluck joined #salt
17:29 badluck Hi, My master in on a VM. I created this VM with virtual box. I created a .sls file. This file just copy a folder. If I ran with "salt '*' state.sls copyFoler" it's working. Now, I try to mount a "shared folder" from Virtual box in the folder I try to copy. If I reran the command, the minion stop to respond and my master VM is not really slow. DO you know if salt do not support mounted folder?
17:30 UtahDave nickaz: did you restart the salt-master service
17:30 UtahDave ?
17:30 nickaz sure i have restart it, may times
17:31 UtahDave badluck: Salt doesn't care about whether a directory is mounted or not.
17:31 UtahDave nickaz: can you pastbin your new eauth section?
17:31 Jahkeup joined #salt
17:31 badluck then I should not have the result I have? ok, then I will continue to try
17:32 UtahDave badluck: yeah, just make sure your mounted directory sits within the file_roots
17:33 badluck my mounted directory is at the same place than my .sls file
17:35 nickaz http://pastebin.com/4SCpBLMP
17:36 UtahDave nickaz: Hm. That looks correct to me.
17:37 nickaz logged this on /var/log/secure : Sep  9 13:26:41 localhost unix_chkpwd[17599]: password check failed for user (salt) Sep  9 13:26:41 localhost python: pam_unix(login:auth): authentication failure; logname=root uid=0 euid=0 tty= ruser= rhost=  user=salt
17:39 xmj joy, those pam-failures.
17:40 faldridge joined #salt
17:40 UtahDave nickaz: can you verify that you can log in as user 'salt'  ?
17:41 _jslatts joined #salt
17:42 Furao joined #salt
17:46 bdf is there any way to use cmd.call in a standard yaml state file?
17:47 bdf like this.... https://gist.github.com/billfehring/3f15e050415e1615ccb7
17:49 UtahDave yep!
17:50 StDiluted joined #salt
17:51 bdf hmmm
17:51 bdf I'm getting TypeError: 'str' object is not callable
17:52 UtahDave bdf: why do you want to use cmd.call?
17:52 devinus joined #salt
17:54 bdf Here, I'll expand the gist
17:54 bdf https://gist.github.com/billfehring/3f15e050415e1615ccb7
17:55 bdf basically I'm working around popen issues similar to issue 5567
17:56 bdf with 'service postgresql-server onestart
17:56 bdf '
17:56 nickaz Thanks my problem is resolved
17:57 nickaz issue with pam
17:57 UtahDave nickaz: ah, good.  what fixed it?
17:57 zloidemon Hello
17:57 UtahDave bdf: can you pastebin the error you're getting?
17:58 bdf sure.
17:58 UtahDave hey, zloidemon!
17:58 bdf UtahDave: http://pastebin.com/3QJ6bVbn
17:59 zloidemon bdf: FreeBSD?
17:59 bdf zloidemon: yep
18:00 zloidemon bdf: This is TypeError: 'str' object is not callable
18:00 zloidemon You can't call string
18:01 bdf is yaml trying to convert the func name into a string or something?
18:01 UtahDave bdf: ok, you're really close,  I wouldn't use cmd.call, though
18:01 UtahDave https://gist.github.com/UtahDave/5853439
18:01 UtahDave drop that file in /srv/salt/_modules
18:01 UtahDave with a name of your choosing.
18:02 UtahDave then in your state   use  module.run  to execute that module.
18:02 bdf that's the code I borrowed ;)
18:02 UtahDave ah, ok.  Looked pretty similar. :)
18:02 bdf but it sounds like what I should be using is module.run
18:02 UtahDave yeah.
18:03 xmj don't you have to return proc ?
18:03 xmj returning the cmd_string doesn't exactly make as much sense.
18:03 bdf I want it to fire and forget.
18:03 bdf because otherwise it will hang.
18:04 faldridge joined #salt
18:04 xmj bdf: return proc
18:04 nickaz just changing passwd for the user, haha
18:04 mohae hi, how can I make sure a mysql user exists before grants are done?
18:05 UtahDave mohae: have your grants require the state where you create the mysql user
18:05 xmj bdf: does it work?
18:05 bdf testing it out now
18:05 zloidemon UtahDave: does grants for mysql user use a mysqlproxy?
18:06 mohae ah, so I should separate out into diff states. I was being lazy and iterating through my pillar data.
18:06 mohae thanks
18:10 bdf made it further this time :)
18:10 bdf TypeError: can't serialize <subprocess.Popen object at 0x806793f90>
18:10 bdf but that's with return proc
18:11 xmj wuut
18:12 Furao joined #salt
18:14 cjh joined #salt
18:22 tuxIO joined #salt
18:23 lemao joined #salt
18:25 ggoZ joined #salt
18:27 matanya joined #salt
18:28 honestly does salt in anyway account for packages having different names across different distributions?
18:29 UtahDave honestly: no, you have to take that into account yourself.
18:29 UtahDave honestly: some people will create pillar entries to differentiate or hard code them into the sls files.
18:30 honestly I think I will solve this by telling people to fix it themselves ^^
18:31 UtahDave We've had a lot of discussions about this, but it's a deep rabbit hole.  There's no way we could keep track of all the packages
18:32 UtahDave So it's just easier for the sysadmin to create the lists of the few packages he/she is managing in pillar
18:33 xmj someone should create a packagekit
18:33 xmj and unify things!
18:34 honestly http://xkcd.com/927/
18:34 UtahDave :)
18:38 xmj (before looking, i know it's the one with 15vs16 standards)
18:38 xmj 14vs15. damnit
18:39 jbunting joined #salt
18:42 LLckfan joined #salt
18:42 LLckfan Hello
18:42 LLckfan I have a fiend ho waas without power for three days and even though thier power is back on she thinks her under arms smell. Is there anything that someone would have in a house that she can use to get rid of the smell?
18:43 badluck Hi, I'm testing salt and everything seems to work for now. But I have a question. If I copy a folder on my computer I need less than 1 seconds. If I mount this folder in the Master (VM) and I copy the same folder a the same place but using a .sls file to copy the folder (mounted), it took more than 4 minutes. There is way to speed up this copy by removing some stuffs (example: diff)?
18:45 Jahkeup joined #salt
18:46 timoguin joined #salt
18:46 forrest LLckfan I think you're in the wrong IRC.
18:46 forrest this is for SaltStack the configuration management tool, not the chemical compound
18:48 timoguin so what formula is preferred for SSH? openssh-formula or ssh-formula. there are two in the saltstack-formulas org.
18:48 MTecknology forrest: He's been spamming like that in many other channels to the point that I have him on highlight.
18:48 LLckfan There is no right IRC channel for anything I ask
18:48 timoguin ssh-formula has more recent commits.
18:48 forrest I see MTecknology
18:49 MTecknology I have him banned in a few channels now.
18:49 honestly does this sls snippet look sane? I'm not solid on the require magic yet: https://gist.github.com/duk3luk3/6499836
18:49 forrest timoguin, either one is ok, ssh uses a newer style that whiteinge has been implementing in several locations, it's pretty slick.
18:50 forrest honestly, it comes down to person opinion, I like to have the file require the pkg.
18:51 forrest just because then I know the exact sequence, pkg is installed, directory is created, content is mounted. Wit how you have it now it could install the package, or create the directory first. Which in this case since it's owned by root is fine
18:51 mwillhite joined #salt
18:51 honestly *nod*
18:51 forrest if however this was something with a user unique to the package, it could fail
18:51 forrest because you could create the dir with a user that doesn't exist.
18:51 honestly ah, yeah
18:52 forrest so how you have it is A-OK, and will be especially once 0.17 is implemented, just something to keep in mind.
18:52 timoguin i gotta get git HEAD running on a master to try some of these newer formulas out.
18:52 forrest that bit me quite a few times with puppet
18:52 honestly will this also work if the nfs-common state is included from another sls file?
18:52 forrest is that the name of this sls?
18:52 honestly no
18:52 forrest or are you talking about the pkg
18:53 honestly ...yes?
18:53 forrest you'd need to include this entire sls
18:53 honestly yeah
18:53 forrest so what you might want to do is break it out, into a state that just installs nfs-common, and then another state, which includes your nfs-common state, and adds to it with these items.
18:54 forrest Does that make sense?
18:54 honestly yeah, that's exactly what I'm trying to do
18:54 forrest yea I'd break it out then, unless you know every machine is going to get this /mnt/nfs dir, and the mount you've specified there.
18:56 honestly ah, include doesn't execute the whole sls by default
18:56 forrest https://gist.github.com/gravyboat/6499905
19:00 jslatts joined #salt
19:01 tuxIO joined #salt
19:02 honestly okay so
19:02 Furao joined #salt
19:02 honestly I'm trying to develop this so it can be added to saltstack-formulas eventually
19:03 forrest ok
19:03 honestly but I'm very confused by the repo format there
19:03 forrest So that repo is kind of in disarray
19:03 honestly wait, nevermind
19:03 forrest take a look at this
19:03 forrest https://github.com/saltstack-formulas/ssh-formula/tree/master/ssh
19:03 forrest that's the most 'recent' method that whiteinge has been using, whether it will be the default, I don't know
19:03 forrest but it is nice
19:03 honestly yeah
19:07 timoguin I'm guessing the saltstack daily PPA isn't maintained much?
19:07 timoguin still has 16.0 looks like
19:08 UtahDave timoguin: I think it's updated every night.   The salt develop branch reports itself as 0.16.0 plus however many commits past 0.16.0 it is
19:09 timoguin ah okay
19:09 timoguin it does say last updated 18 hours ago
19:11 UtahDave lunchtime!!
19:13 honestly forrest: I have it like this now - https://github.com/duk3luk3/nfs-user-formula/tree/master/nfs-users
19:15 kermit joined #salt
19:16 mgw joined #salt
19:31 mohae joined #salt
19:32 whiteinge timoguin: i expect the openssh formula will get merged into the ssh formula
19:33 timoguin whiteinge, yea that's what i figured.
19:35 badluck Hi, I read "http://docs.saltstack.com/ref/states/all/salt.states.file.html" but I just to be sure. There is a way to ignore some file's type (example .lib) when I copy a folder in a .sls?
19:36 whiteinge honestly: the formula repos will be (and are in a few cases) the go-to for mapping package names between distros. going that route keeps the problem bite-size, so to speak :)
19:36 whiteinge honestly: e.g.: https://github.com/saltstack-formulas/mysql-formula/blob/master/mysql/map.jinja
19:37 honestly whiteinge: ah, that's a cool approach
19:38 p3rror joined #salt
19:38 honestly whiteinge: any comments on the structure of my repo?
19:39 * whiteinge looks
19:40 Viaken_ joined #salt
19:40 carlos joined #salt
19:40 honestly needing mako is obviously a slight problem...
19:42 whiteinge honestly: neat! that's a great start
19:43 Jahkeup joined #salt
19:43 whiteinge i love the pillar example. just throwing dictionaries at things is awesome
19:43 * timoguin nods
19:43 jacksontj joined #salt
19:45 whiteinge honestly: i'd suggest switching the init.sls and the nfs.sls so the base functionality is installing the requisites and the "outrageous" (haha) part requires that
19:45 whiteinge that way other formula can also take advantage of the base thing. (e.g., using a require to ensure the package is installed)
19:46 honestly hm.
19:46 whiteinge my other suggestion would be to wrap the user portion in an if-statement since it requires data from pillar so that it won't run if that data is missing
19:46 mgw joined #salt
19:47 honestly but this is a user-management formula that uses nfs, not a nfs-formula that has an optional user-management component.
19:47 honestly there's already a nfs-formula isn't there?
19:47 whiteinge i don't think so...
19:47 * whiteinge looks
19:47 honestly ah
19:47 honestly yeah, there isn't
19:48 whiteinge your point is still valid. it would make sense to have them be separate repos. i think it would also be fine to have them be one big repo
19:48 martoss joined #salt
19:48 whiteinge (case in point, the apache formula)
19:49 honestly whiteinge: I'm confused about how to deploy a formula using the current repo structure. clone the repo somewhere, then manually copy the sls directory to the salt root?
19:49 whiteinge so long as the additional functionality is opt-in, i think big formula repos are neat
19:50 martoss Hey folks, i have a pillar question: How do I (securely) distribute files like ssh or gpg keys via pillar? They are typically plaintext files.
19:50 whiteinge that workflow is fine. you can also add the repo URL as a gitfs remote. or you can clone the repo yourself and add that directory to the file_roots list in your master config
19:50 whiteinge honestly: that said, this will get a lot easier soon-ish
19:50 honestly whiteinge: adding the repo url as a gitfs remote would end up with a directory tree that won't run, won't it?
19:51 honestly sine you'll have the "x-formula" in the middle
19:51 honestly and the supporting files in the sls directory don't respect that
19:51 whiteinge it will run because gitfs (and file_roots) look at the contents of the directory specified. not the dir itself
19:51 honestly ah
19:51 mattmtl joined #salt
19:52 whiteinge martoss: there is no way to distribute files using pillar (for now). however yaml can handle multi-line strings so you can put those keys inline in your sls files
19:52 timoguin anyone know if gitfs is supported in pillar yet?
19:53 honestly is straightening out the pillar system so it can do that being worked on? I've seen it's a popular feature
19:54 martoss ah ok, so this means I just put e.g. the "backup key" inline in one file.
19:54 honestly martoss: only put public keys into salt, ideally
19:54 honestly there isn't yet a way to make minions only see things they're supposed to see
19:55 honestly actually, putting the keys themselves into pillar could work
19:56 whiteinge honestly: yeah, there's an open ticket to distribute files over pillar iirc
19:56 Jahkeup joined #salt
19:57 whiteinge pillar is specifically intended for secure distribution of minion-specific data, so although putting private keys inline in the sls file is a tad clunky it works fine
19:57 martoss honestly: ok - so it's better to distribute them by hand for now...
19:57 honestly martoss: what whiteinge said... you could do it but it's clunky
19:57 tblamer joined #salt
19:58 martoss ok, thx for the info.
19:58 honestly whiteinge: how would a pillar file look that has two different values for a specific key for two different minions?
19:59 honestly (key as in field generally, not key as in private key)
19:59 Katafalkas joined #salt
19:59 whiteinge it's file-based not key-based. so you use the pillar top file to determine which minion can see what sls file
20:00 honestly ah
20:00 whiteinge so you'd have two files with the same named key. one for each minion but with differing values
20:00 cbloss anyone know why setting the following variable as "on" in a pillar results in "True" instead of "on"  |       gzip {{ nginx.get('gzip', 'on') }};
20:00 timoguin_ joined #salt
20:01 whiteinge cbloss: http://docs.saltstack.com/topics/troubleshooting/yaml_idiosyncrasies.html#true-false-yes-no-on-off
20:01 martoss By the way, I was asking because of http://docs.saltstack.com/topics/tutorials/pillar.html "Information transferred via pillar is guaranteed to only be presented to the minions that are targeted, this makes pillar the engine to use in Salt for managing security information, such as cryptographic keys and passwords."
20:01 whiteinge cbloss: try this instead: ``gzip "{{ nginx.get('gzip', 'on') }}"``
20:02 martoss whiteinge: ah ok, now I got how this is suppose to work. And since plain text is not working, one can embed the information in yaml and write them out to file on the client?
20:03 whiteinge martoss: exactly! pillar is definitely the way to go for what you're trying to do. look at the example for the "contents" argument to the file.managed state
20:03 cbloss whiteinge: same result with "True". let me read this doc more
20:04 whiteinge martoss: hm. actually look at the "contents_pillar" argument instead. looks like it's been improved since i looked last
20:04 whiteinge martoss: http://docs.saltstack.com/ref/states/all/salt.states.file.html#salt.states.file.managed
20:05 cbloss whiteinge: gzip: "on"   fixed the issue in the pillar. thanks for pointing me in the right direction
20:05 martoss whiteinge: ok
20:07 honestly whiteinge: in the apache formula, how do the sub-states get away without including the init.sls?
20:08 honestly fcgid.sls at least does
20:08 honestly mod_wsgi.sls has the include...
20:08 Jahkeup_ joined #salt
20:08 whiteinge martoss: oh, looks like that may have gone in after all: https://github.com/saltstack/salt/pull/6508
20:09 whiteinge martoss: may not be in a current release yet though...
20:09 whiteinge honestly: the fcgid state probably doesn't work
20:09 honestly whiteinge: I am shocked and appalled!
20:10 whiteinge hehe
20:10 whiteinge we're in the process of getting the formula repos running through jenkins
20:12 Furao joined #salt
20:15 devinus joined #salt
20:15 valentt joined #salt
20:15 valentt Hi all!
20:15 valentt Has anybody tried compiling salt for openwrt ?
20:18 matanya joined #salt
20:21 honestly whiteinge: that sounds like a pain to setup.
20:21 isomorphic joined #salt
20:22 faulkner joined #salt
20:22 dstanek joined #salt
20:23 craig is there a list of tags the reactor system can id?
20:23 craig i thought i saw a list once... but i dont remember where...
20:23 craig UtahDave: whiteinge: ^
20:24 whiteinge not to my knowledge
20:24 whiteinge that should *really* be added to the docs though
20:24 craig :)
20:24 craig k. i need to add some metrics gathering
20:24 craig and we are going to use the reactor system
20:25 craig so, i can document what i find
20:25 whiteinge very cool
20:26 forrest Hey s0undt3ch, did you try to duplicate this? https://github.com/saltstack/salt-bootstrap/issues/222
20:26 chutzpah joined #salt
20:26 chutzpah joined #salt
20:28 UtahDave craig: there's not.   Right now the best thing to do is listen to everything on the event bus and look for the tags you want to match on when you execute various things
20:29 s0undt3ch forrest: nope, and it seems a mirror out of wack or keys not updated
20:29 s0undt3ch forrest: as you said
20:29 forrest Ok cool, I just wanted to double check as to whether you had messed with it at all.
20:30 martoss whiteinge: wow that looks even simpler than - blah: | at the beginning of the file ...
20:32 faulkner joined #salt
20:33 chutzpah joined #salt
20:33 chutzpah joined #salt
20:33 whiteinge yeah. it looks like a good change. i checked it's not in a current salt release
20:34 whiteinge martoss: if you don't want to wait for 0.17, the syntax to pull that off now looks like this:
20:34 whiteinge http://salt.readthedocs.org/en/v0.16.3-0/ref/states/all/salt.states.file.html#salt.states.file.managed
20:35 whiteinge (look at the example under "contents")
20:35 renoirb joined #salt
20:36 qba73 joined #salt
20:38 chutzpah joined #salt
20:44 mannyt joined #salt
20:47 honestly so I'm trying to test my users.sls, which is in salt://nfs-users/users.sls with salt-call, doing "salt-call -l debug nfs-users.users test=True"
20:47 honestly but it tells me "no matching sls found"
20:47 honestly I have "nfs-users.users" in my top.sls in base
20:48 cbloss i'm trying to use the salt-stack formula(https://github.com/saltstack-formulas/salt-formula)http://pastebin.com/8eHCLufZ   doesn't seem to be working. Any idea what i'm doing wrong here?
20:53 Furao joined #salt
20:55 forrest Does it throw an error?
20:55 cbloss no error
20:55 timoguin joined #salt
20:55 valentt I thought that nobody would try using something like puppet or salt on openwrt but when I saw that puppet got ported to openwrt I immediately wanted salt also ;)
20:56 cbloss 2013-09-09 20:51:37,288 [salt.state       ][ERROR   ] No matching sls found for 'salt' in env 'base' in the /var/log/salt/minion log file
20:57 honestly hah, that's the same error I'm getting right now
20:58 forrest did you guys try clearing the cache?
20:58 honestly which cache?
20:59 forrest the salt cache
20:59 honestly how?
20:59 timoguin delete /var/cache/salt
20:59 forrest and restart the service
20:59 forrest might have to restart the minion service
21:00 honestly huh
21:00 honestly nope, no dice
21:00 honestly I must be doing something wrong
21:01 forrest You broke it :(
21:01 honestly I have "base: '*': - nfs-users.users" in my top.sls
21:01 honestly with appropriate linefeeds and indentation, obviously
21:02 martoss whiteinge: thx
21:03 cbloss honestly: if you figure it out let me know, I think we are having the same issue
21:04 cbloss is it because it isn't located in "base" maybe?
21:08 oz_akan_ joined #salt
21:10 honestly I've no idea.
21:12 honestly state.show_top shows the top...
21:13 honestly hrm
21:14 honestly whiteinge: halp
21:14 honestly something must be wrong with my top.sls or with my file structure
21:14 fragamus joined #salt
21:14 honestly I put "nfs-users.users" into the top
21:14 honestly but it can't find it
21:15 forrest can you paste your sls?
21:15 forrest both your top, and the nfs one
21:15 forrest or did you already? Sorry been busy
21:15 honestly https://github.com/duk3luk3/nfs-user-formula
21:16 derelm joined #salt
21:17 honestly top: https://gist.github.com/duk3luk3/6501655
21:17 forrest yeaaaaaaa
21:18 krissaxton joined #salt
21:18 forrest you might want to name your users.sls, users.sls, and not users.nfs
21:18 honestly :|
21:18 forrest :P
21:18 honestly foiled by my brain
21:19 forrest This is why they pay me the big bucks
21:20 forrest Let me know if that works honestly
21:21 matanya joined #salt
21:22 honestly it does
21:22 honestly I'm fine
21:22 forrest cool
21:23 honestly no idea what cbloss' problem is
21:23 forrest cbloss did you already clear the cache, and restart the master/minion service?
21:24 kspath joined #salt
21:24 cbloss forrest: yes
21:25 forrest and you're still just getting nothing back?
21:25 cbloss nope, same problem.
21:25 forrest can we see your top.sls?
21:25 forrest and what does the directory structure look lik
21:25 forrest e*
21:26 cbloss let me make a new one
21:26 cbloss pastebin that is*
21:27 cbloss http://pastebin.com/usQ2Bjif
21:28 forrest oh duh this is what you showed earlier, sorry. The joys of working on 6 things at the same tim
21:28 forrest *e
21:28 forrest ok so it's still saying that it can't find salt in the 'base'
21:28 alunduil joined #salt
21:28 forrest and I assume your setup is /srv/salt/salt/master.sls ?
21:28 cbloss 2013-09-09 21:25:45,002 [salt.state       ][ERROR   ] No matching sls found for 'salt.master' in env 'base'
21:28 cbloss i'm trying to use
21:29 cbloss the gitfs
21:29 cbloss salt-formula
21:29 cbloss http://docs.saltstack.com/topics/conventions/formulas.html
21:29 cbloss mmm, i bet it is because I require a newer version of salt
21:29 forrest try salt-formula.salt.master
21:30 forrest oh what version are you on?
21:30 cbloss 0.16.3 , but in the top of that page the "note" references 0.16.4 and 0.17
21:30 forrest gitfs should work on 0.16.3
21:30 forrest I'm using it for my stuff.
21:31 forrest can you try that format I posted above?
21:31 cbloss 2013-09-09 21:31:02,140 [salt.state       ][ERROR   ] No matching sls found for 'salt-formula.salt.master' in env 'base'
21:31 forrest hmm
21:31 forrest maybe the one in the top is specific to 0.16.4?
21:31 forrest I don't know
21:31 forrest maybe this doc page is just in prep for 0.17
21:32 forrest and that feature isn't in yet
21:32 cbloss I'm thinking that is the case
21:32 cbloss maybe whiteinge can confirm?
21:32 KennethWilke yeah whiteinge was opting for the new grains.filter_by for some stuff
21:32 forrest He's away
21:32 KennethWilke i touched that formula a lil bit too though
21:32 forrest but looking at this note, you're probably right
21:33 cbloss I'll do a git submodule for now
21:33 KennethWilke cbloss
21:33 KennethWilke cbloss: it should be just salt.master
21:33 berto- joined #salt
21:33 forrest yea that's not working
21:33 cbloss KennethWilke: I tried that
21:33 KennethWilke what error did you get when you used that? no matching sls?
21:33 cbloss yep
21:34 KennethWilke did you load it up via gitfs?
21:34 forrest here's his pastebin
21:34 forrest http://pastebin.com/usQ2Bjif
21:34 cbloss did I setup the gitfs correctly?
21:34 KennethWilke i think so
21:35 jbunting joined #salt
21:35 dork joined #salt
21:35 KennethWilke if you have it set with that, and bouncing the master doesn't fix the problem i think it may be a gitfs version problem
21:35 KennethWilke or git-python rather
21:36 jacksontj joined #salt
21:36 cbloss I made sure to install the git-python package before I tried this
21:36 KennethWilke hmm, very odd
21:37 forrest cbloss brought up a good point that the formula stuff claims to only be support in 0.16.4/0.17
21:37 KennethWilke i tested that formula on 0.16.3 on debian, centos, fedora, gentoo and ubuntu
21:37 forrest and he's on 0.16.3
21:37 forrest and it worked?
21:37 forrest no the formula is fine
21:37 KennethWilke yep, at that time it did
21:37 forrest but were you using it like cbloss is?
21:37 forrest where he's trying to pull it in straight from github
21:38 KennethWilke yeah with git as a fileserver backend and gitfs to load it
21:38 KennethWilke i wrote what i used in the pillar.example
21:38 forrest Even after reviewing this note? http://docs.saltstack.com/topics/conventions/formulas.html
21:38 forrest hmm
21:38 cbloss mmmm
21:38 KennethWilke yeah that note true for a lot of formulas
21:38 KennethWilke as they use grains.filter_by
21:38 forrest but not this one
21:38 KennethWilke correct
21:39 cbloss do i need to do git+ssh?
21:39 KennethWilke shouldn't need to
21:39 forrest cbloss, the other items are getting applied right?
21:39 forrest packages, ssh, etc.
21:39 cbloss yeah all other items work fine
21:39 forrest ok
21:40 cbloss i must be doing something wrong, haha
21:41 cbloss is it looking in the incorrect environment "base" ?
21:41 KennethWilke at this point it might be prudent to run the master and minion manually with '-l debug' to see if anything funky comes up
21:43 cbloss KennethWilke: I was ( sudo salt-master --log-level=debug)…didn't notice anything abnormal…
21:44 forrest cbloss, can you try to just pull down the git repo into some tmp directory?
21:44 KennethWilke i'm still inclined to blame git-python
21:44 forrest what a hater
21:44 KennethWilke it's caused me this exact issue before
21:44 forrest I'm joking
21:44 KennethWilke when it was installed, but not a sufficient version
21:45 forrest just giving you a hard time
21:45 KennethWilke lol i took it that way
21:45 forrest cbloss what OS are you on?
21:45 forrest ok good
21:45 cbloss ubuntu 12.04
21:45 * KennethWilke rage!
21:45 forrest can you try an apt-get update, then try to update git-python
21:45 cbloss is there a gitfs package i need?
21:45 KennethWilke the ubuntu package might not be good enough
21:45 KennethWilke might need to pip install --upgradeit
21:45 cbloss updating now
21:45 forrest cool
21:46 forrest I keep having issues where salt 0.12 gets installed when building a new ubuntu vm
21:46 forrest have to update first, then reinstall to get up to 0.16.3
21:46 forrest so maybe something similar is happening with git-python
21:47 KennethWilke cbloss run `python -c 'import git; print git.__version__'`
21:48 cbloss 0.1.6
21:49 KennethWilke yeah that's the problem
21:49 cbloss so I need to remove the version of git-python i installed via apt-get
21:49 cbloss then install pip, and install git-python via pip?
21:50 avienu joined #salt
21:50 KennethWilke i think you can just install pip then `pip install --upgrade gitpython`
21:50 forrest agreed
21:50 KennethWilke then bounce that master
21:50 KennethWilke and enjoy the win
21:51 forrest lol
21:51 KennethWilke :p
21:51 forrest what a pain that the repo isn't up to date.
21:51 KennethWilke yeah
21:51 KennethWilke it's up to date on wheezy, and i think on 13.04
21:51 cbloss stupid question...do you have to sudo to use pip?
21:51 KennethWilke if you're not root yeah
21:52 forrest unless you're running salt in a venv
21:52 forrest s0undt3ch, do you have any sort of requires for https://github.com/saltstack/salt-bootstrap/issues/173 ? What do you want the RPM to accomplish, just dropping the bootstrap on?
21:53 forrest err requirements
21:54 KennethWilke forrest, see gitfs and I have some history with this issue
21:54 KennethWilke https://github.com/saltstack/salt/pull/6412
21:54 KennethWilke :p
21:54 KennethWilke it used to error if it wasn't installed, then fail silently if the version wasn't above 0.3.0
21:54 forrest heh
21:55 forrest cbloss, you should file an issue on that
21:55 forrest because it should return some kind of error
21:55 forrest if it isn't returning anything with the debug option, that's a problem
21:55 cbloss just finished the test…the new version worked
21:55 cbloss where do I file an issue?
21:55 forrest https://github.com/saltstack/salt/issues/new
21:56 KennethWilke it did at a time
21:57 KennethWilke it logged an error when i fixed it! :D
21:57 mwillhite joined #salt
21:57 forrest lol
21:57 KennethWilke but it's changed a lil bit since then
21:57 KennethWilke https://github.com/saltstack/salt/blob/develop/salt/fileserver/gitfs.py#L35
21:57 forrest yea I saw that in yours
21:57 forrest the error is in there, I don't get why it's not logging it.
21:57 forrest maybe some of the logic is having issues.
21:58 KennethWilke the gitfs_root may be causing it
21:58 KennethWilke well no, then it'd still be broken
22:00 chutzpah joined #salt
22:00 chutzpah joined #salt
22:01 KennethWilke this seems to me like it may not load the module unless gitfs_root is set: https://github.com/saltstack/salt/blob/develop/salt/fileserver/gitfs.py#L41
22:02 kermit joined #salt
22:06 KennethWilke yeah, i dunno if it's that exactly, but when i use the gitfs.py from git my gitfs_remotes do not work
22:07 mwillhite I'm having issues with a build…there are some states that I would like to consider optional…for example a git pull. Is there a way to not stop the entire build if I can't connect to github for some reason?
22:10 KennethWilke mwillhite: i'm not aware of a way to do that, but that'd certainly be handy sometimes, i think salt should normally continue on it's way as long as that resource isn't part of your require chain
22:10 UtahDave mwillhite: just don't require the git pull
22:10 HumanCell joined #salt
22:11 druonysus joined #salt
22:11 druonysus joined #salt
22:12 cbloss KennethWilke / forrest: I should put an issue in or not? It looks like it should be spitting an error out according to KennethWilke's link?
22:13 Furao joined #salt
22:13 KennethWilke cbloss: did you install a git-latest version of salt?
22:13 KennethWilke or how'd you go about installing it
22:13 cbloss KennethWilke: I'm using the stable release of salt
22:13 druonysus joined #salt
22:13 druonysus joined #salt
22:14 KennethWilke via the salt-bootstrap.sh or installing from ubuntu's packages?
22:14 KennethWilke either way, i think if you're running a stable version i'd submit a bug, if someone knows more about it they might chime in with some useful information
22:15 forrest Yes put in an issue
22:15 mwillhite joined #salt
22:15 forrest even if you're running an old version, it came from the repo, salt should kick back an error saying 'hey this failed to load properly, update it'
22:16 forrest I mean you wasted 30+ minutes trying to figure out what was wrong because there was no error, that's a problem :P
22:16 s0undt3ch forrest: just drop the script somewhere appropriate
22:16 forrest s0undt3ch, ok
22:16 s0undt3ch forrest: that was though to help with unit test deps while building packages
22:17 s0undt3ch forrest: which are usually executed in systems which can only install distro available packages
22:17 KennethWilke s0undt3ch: is the gitfs stuff covered by the tests?
22:17 KennethWilke just curious
22:17 forrest s0undt3ch, I gotcha.
22:19 Jahkeup joined #salt
22:19 forrest Has anyone used salt to set up oracle weblogic yet?
22:22 nocturn joined #salt
22:22 nocturn joined #salt
22:23 cewood joined #salt
22:23 forrest Interesting video regarding salt, talks about the enterprise model, and a bit about halite
22:23 forrest http://www.youtube.com/watch?v=xMl--TaP4YM
22:24 yo61 joined #salt
22:24 yo61 Anyone using salt on SmartOS?
22:24 yo61 I'm having trouble bootstrapping it
22:25 forrest no what's going on with it
22:26 KennethWilke i like the smartos background, very atari like
22:26 forrest lol
22:26 KennethWilke i know that's probably super helpful for you too :\
22:26 yo61 https://gist.github.com/robinbowes/6502338
22:26 forrest it reminds me of a fragmented disk
22:26 KennethWilke yo61: do you have pip installed?
22:26 yo61 And with -P https://gist.github.com/robinbowes/6502342
22:27 yo61 Mmm, not sure I do
22:27 yo61 Do now
22:28 yo61 Same error
22:28 yo61 Failed to run install_smartos_git_deps()!!!
22:28 KennethWilke kvm on solaris, that is interesting
22:28 forrest do you have git installed?
22:28 yo61 Mmm, nope
22:29 forrest the install function should be installing it.
22:29 forrest the git deps call it
22:29 forrest https://github.com/saltstack/salt-bootstrap/blob/develop/bootstrap-salt.sh#L2357
22:29 forrest not git though, hmm
22:30 forrest can you rerun it after installing git yo61?
22:30 yo61 Hmm, not sure I have git in my pkgin repo
22:30 KennethWilke forrest: https://github.com/saltstack/salt-bootstrap/blob/develop/bootstrap-salt.sh#L2404
22:31 KennethWilke tis the func from the error
22:31 forrest yea that's where I'm looking
22:31 KennethWilke ah okay, you linked the one above it :
22:31 KennethWilke :p
22:31 forrest first thing it does is call install_smartos_deps
22:31 yo61 Installing git manually
22:31 KennethWilke hey! just because you read the function doesn't mean i did
22:31 forrest lol
22:31 KennethWilke and now you know i didn't
22:31 forrest *shrug*
22:31 yo61 Wonder why it doesn't install with the bootstrap?
22:32 forrest it should
22:32 forrest line 2406 calls pkgin -y in scmgit
22:32 forrest which I assume is the installation procedure for installing packages in smartos?
22:32 forrest I'm not familiar with it
22:32 yo61 Well, it doesn't
22:32 whiteinge cbloss: did you get your formula working?
22:32 forrest I'm not disagreeing with you
22:32 yo61 Would be nice to have a pkgin package
22:32 forrest is scmgit the name of the git package yo61?
22:32 yo61 Yes
22:33 forrest ok
22:33 CheKoLyN joined #salt
22:33 forrest let me know what happens once you have git installed
22:33 yo61 Hmm, still failing
22:33 forrest same error?
22:33 cbloss whiteinge: yes, thanks for checking. With some help from KennethWilke and forrest…was a git-python version issue
22:33 yo61 * ERROR: Failed to run install_smartos_git_deps()!!!
22:33 forrest Pssh, don't implicate me in this, it was all KennethWilke :P
22:33 cbloss haha
22:33 forrest ok so something is going on in that function
22:34 whiteinge ah, yes. that's a non-obvious bug for sure :-/
22:34 forrest did you file an issue cbloss?
22:34 cbloss forrest: yes
22:34 forrest awesome
22:34 forrest the errors are in there, not sure why they aren't getting dumped out
22:34 forrest hey yo61, can you look here https://github.com/saltstack/salt-bootstrap/blob/develop/bootstrap-salt.sh#L2357
22:35 KennethWilke cbloss: you're welcome! glad that's all fixed for ya
22:35 KennethWilke i know the frustration there, too well!
22:35 forrest there we go, that's the check_pip_allowed function it bombs on
22:35 forrest https://github.com/saltstack/salt-bootstrap/blob/develop/bootstrap-salt.sh#L108
22:35 forrest hmm
22:35 yo61 No, it runs that OK
22:36 forrest * ERROR: pip based installations were not allowed. Retry using '-P'
22:36 forrest that's still happening right?
22:36 yo61 Yeah, I re-ran with -P
22:36 yo61 I know get echowarn "PyZMQ will be installed using pip"
22:36 yo61 <sigh>
22:36 forrest yea I see that now, sorry I was looking at the first gist
22:36 yo61 I now get: PyZMQ will be installed using pip
22:36 forrest right, so we've made itto line 2359, huzzah!
22:37 yo61 It's barfing somewhere else
22:37 forrest (pkg_info gcc-compiler > /dev/null 2>&1 && pkgin -y in gcc-compiler) || \
22:37 forrest (pkg_info gcc47 > /dev/null 2>&1 && pkgin -y in gcc47) || return 1
22:37 forrest can you check to see if those are installed?
22:37 forrest gcc-compiler, and gcc47?
22:37 yo61 There doesn't appear to be a gcc-compiler
22:37 forrest ok, what about gcc47
22:38 yo61 Installing that now
22:38 forrest hmm, why is that not getting installed though :\
22:38 forrest the bootstrap should pick it up
22:38 druonysus joined #salt
22:39 yo61 Bad logic
22:39 forrest as a shell statement can you try running (pkg_info gcc-compiler > /dev/null 2>&1 && pkgin -y in gcc-compiler) || (pkg_info gcc47 > /dev/null 2>&1 && pkgin -y in gcc47) || echo 'nope'
22:39 yo61 Should be: pkg_info gcc47 || pkgin -y in gcc47
22:40 forrest you're totally right, good catch
22:41 yo61 Re-running with that change
22:41 forrest hmm, does the pkg_info gcc-compiler output anything for you?
22:41 forrest you modified the script?
22:41 yo61 No
22:41 forrest ok, when you get a seck can you try pkg_info gcc-compiler > /dev/null 2>&1; $?
22:42 yo61 1
22:42 yo61 There is no gcc-compiler in the pkgin repos I'm using
22:42 forrest I know
22:43 forrest I just wanted to see what that output would be
22:43 forrest to see if somehow it was simply considering command output to be a successful completion
22:44 druonysus joined #salt
22:44 druonysus joined #salt
22:44 emocakes joined #salt
22:45 forrest Can you file an issue on that yo61? Unless you want to put in the pull request yourself, I just don't have a system to test it on after modifying the logic.
22:45 yo61 Odd, it's not working in the script
22:46 forrest when you change the && to || ?
22:46 yo61 This works OK:
22:46 yo61 (pkg_info gcc-compiler > /dev/null 2>&1 && pkgin -y in gcc-compiler) || \
22:46 yo61 >         (pkg_info gcc47 > /dev/null 2>&1 || pkgin -y in gcc47) || return 1
22:46 forrest that returns a 0?
22:46 yo61 in a shell
22:46 forrest yea
22:46 forrest but the script still fails with the same error?
22:46 jacksontj joined #salt
22:46 yo61 But, the same thing in the script just sits at: downloading gcc47-4.7.3nb1.tgz:   0%
22:47 forrest no error? Just hangs?
22:47 yo61 Yes
22:47 forrest if you already installed gcc47, what happens if you just run pkgin -y in gcc47 again?
22:47 forrest does it just say it's installed and return a 0?
22:47 yo61 I think so
22:47 yo61 It's not yet installed
22:48 forrest oh I thought you installed it manually
22:48 yo61 Not yet - it's 250MB and I'm on a slow connection
22:49 forrest ugh
22:49 forrest gcc
22:50 forrest ok, let me know how that goes, one of us should do a pull request to update that logic in the script
22:50 forrest so that smartos users aren't screwed.
22:50 yo61 k
22:51 yo61 There is actually a salt package
22:51 forrest right but you want develop
22:51 forrest the package is probably 0.16.3, not the latest develop branch
22:51 yo61 salt-0.15.3
22:51 forrest yea screw that
22:51 yo61 And it doesn't appear to have any smf
22:51 forrest Need to find who is building the salt packages for smartos and ask them to update
22:52 robertkeizer joined #salt
22:54 yo61 It's http://www.perkin.org.uk/pages/pkgsrc-binary-packages-for-illumos.html
22:55 yo61 I only want the minion
22:55 dgarstang joined #salt
22:55 dgarstang minions, server, syndic all suddenly frozen. Restarts didn't help...
22:56 lineman60 joined #salt
22:57 yo61 OK, once I installed gcc47 manually, the script continues
22:58 yo61 Hmm, no gcc found
22:58 forrest what
22:58 forrest doesn't pkg_info gcc47 return a 0?
22:59 yo61 It installs it, but it's not in the path
22:59 yo61 Put's it in /opt/local/gcc47/bin/gcc
23:00 forrest that's weird
23:00 tyler-baker joined #salt
23:00 yo61 So the script can't find it
23:01 redbeard2 joined #salt
23:01 bhosmer joined #salt
23:02 yo61 I put  /opt/local/gcc47/bin in $PATH and re-ran
23:03 Furao joined #salt
23:03 krissaxton joined #salt
23:03 yo61 make[2]: *** [libzmq_la-xpub.lo] Error 1
23:05 forrest is that happening oin the script?
23:05 forrest *in
23:05 yo61 Yes
23:06 forrest ok, I assume all this stuff got installed: https://github.com/saltstack/salt-bootstrap/blob/develop/bootstrap-salt.sh#L2367
23:06 forrest so we're trying to run the make on zeromq?
23:06 yo61 Yes
23:07 forrest Can you make it yourself if you download the content and run a ./configure in the directory?
23:07 yo61 No, same error
23:07 forrest :\
23:07 yo61 Time for bed
23:07 forrest ok, can you file an issue if you remember to tomorrow?
23:07 yo61 Thanks for the help, I'll maybe try again tomorrow
23:07 yo61 Will do
23:07 forrest Sounds good
23:08 nineteen1ightd joined #salt
23:09 oz_akan_ joined #salt
23:13 krissaxton joined #salt
23:13 Jahkeup joined #salt
23:20 oz_akan_ joined #salt
23:38 ninkotech joined #salt
23:38 cbloss http://docs.saltstack.com/ref/states/all/salt.states.git.html  <-- how do I specify a user/group who owns the files?
23:43 zzzirk joined #salt
23:46 alekibango joined #salt
23:48 danielbachhuber joined #salt
23:49 bhosmer joined #salt
23:56 oz_akan_ joined #salt
23:59 FluxD joined #salt

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