Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-02-11

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

All times shown according to UTC.

Time Nick Message
00:00 ninkotech__ joined #salt
00:01 frasergraham joined #salt
00:01 Guest__ joined #salt
00:01 krow joined #salt
00:02 frasergr_ joined #salt
00:03 frasergr_ joined #salt
00:04 krow1 joined #salt
00:09 frasergr_ i'm using service.running and it's claiming it has started the service, but it's not running
00:09 frasergr_ minion is OSX and I see it executing a launchctl load call
00:09 frasergr_ but not a start call
00:13 ninkotech__ joined #salt
00:13 __number5__ frasergr_: load will run the service, have you check your service xml?
00:14 frasergr_ RunAtLoad is True
00:15 frasergr_ i had manually stopped the service to test this
00:15 frasergr_ so it was loaded but not running
00:15 frasergr_ when I call load in that state is says "Already Loaded" but does not start it
00:17 frasergr_ it works fine if I unload the service to test this.
00:18 oz_akan_ joined #salt
00:18 frasergr_ Seems like a problem though, if a service is loaded but stopped then highstate claims it started the service but it didn't
00:18 __number5__ how do you stop the the service, launchctl unload?
00:20 frasergr_ launchctl stop
00:20 frasergr_ lauunchctl stop <service name>
00:20 frasergr_ as opposed to the path to the plist
00:21 __number5__ so after that launchctl status <service name> show it stopped?
00:23 frasergr_ there is no launchctl status
00:23 frasergr_ but when it is stopped it has no PID listed
00:23 frasergr_ launchctl list -x <service>
00:23 __number5__ sorry, launchctl list
00:26 frasergr_ the service start call could always issue a "launchctl start" in addition to the load call. If the service is loaded and running start does nothing
00:26 __number5__ just test on Moutain Lion, launchctl stop <service name> won't stop a service
00:26 __number5__ better to just use load and unload
00:26 ninkotech__ joined #salt
00:27 __number5__ s/test/tested/
00:27 frasergr_ it depends on how your service is configured. For my services stop does stop the service
00:27 frasergr_ i'm running on Mavericks
00:28 gadams999 joined #salt
00:28 kermit joined #salt
00:29 __number5__ AFAIK, salt service state/osx using launchctl load/unload
00:29 __number5__ https://github.com/saltstack/salt/blob/08be9bd47d9a98568a16b68b454ce571c635a743/salt/modules/launchctl.py#L181
00:29 __number5__ if you want to use stop/start, you need to create your own state/module
00:30 frasergr_ or I may submit a patch to salt to handle those commands too.
00:30 frasergr_ the fact that highstate tells me the service was started when it was not is a problem
00:30 Corey frasergr_: Yeah, unfortunately that's a bit hairy to work with.
00:31 Corey I really fail to see the launchd stuff as "valid," but you're right.
00:31 frasergr_ yes it's a different paradigm to Linux services but it's all we have on OSX
00:31 Corey frasergr_: I'm tempted to grab in something like https://github.com/jimbojsb/launchrocket
00:32 Corey frasergr_: (Hey, I'm on Mavericks too; I agree with you.)
00:32 frasergr_ even if the salt launchctl module called start after load, it would be an improvement without negative side effects
00:33 frasergr_ anyway, I need to run. I'll figure something out later :) Thanks for your help everyone
00:33 frasergr_ left #salt
00:34 frasergraham joined #salt
00:35 __number5__ I'm not sure launchd behave the same between different OSX releases...
00:35 ninkotech__ joined #salt
00:40 rojem joined #salt
00:45 ninkotech_ joined #salt
00:56 blast_hardcheese joined #salt
00:56 ninkotech_ joined #salt
00:56 juicer2 joined #salt
00:57 ninkotech_ joined #salt
00:58 MaZ- joined #salt
01:03 rgbkrk joined #salt
01:05 ofhhgf joined #salt
01:05 TheRealBill_here joined #salt
01:05 ofhhgf Hi
01:05 tresor joined #salt
01:06 ofhhgf Is there a way to bootstrap a server with salt
01:06 ofhhgf e.g using salt-ssh?
01:07 rojem joined #salt
01:07 ninkotech_ joined #salt
01:08 rojem joined #salt
01:09 rojem joined #salt
01:09 rojem joined #salt
01:09 david_a joined #salt
01:10 taion809 joined #salt
01:12 ofhhgf I am looking to take a rackspace cloud server from a bare state to a fully working server with salt, and I would like some points.
01:13 cedwards ofhhgf: there is salt-bootstrap
01:13 ccase joined #salt
01:14 ofhhgf cedwards: That requires that I SSH into the server first
01:14 ofhhgf I want to do everything in one commandf
01:14 ofhhgf command*
01:14 quickdry21 joined #salt
01:15 ofhhgf Right now I am using fabric to do everything
01:17 quickdry21 joined #salt
01:18 jhenry82_ joined #salt
01:18 Linz joined #salt
01:20 ioni joined #salt
01:22 ioni joined #salt
01:24 dancat joined #salt
01:25 Damoun joined #salt
01:28 iMil joined #salt
01:28 iMil joined #salt
01:29 LLKCKfan joined #salt
01:30 lionel joined #salt
01:31 Bosch[] joined #salt
01:32 fllr joined #salt
01:33 Veticus__ joined #salt
01:36 majoh joined #salt
01:37 Veticus__ left #salt
01:42 hummus3 joined #salt
01:43 n8n joined #salt
01:45 btp joined #salt
01:45 rojem joined #salt
01:46 rojem joined #salt
01:47 zz_Cidan joined #salt
01:48 Gifflen joined #salt
01:49 sroegner joined #salt
01:50 majoh joined #salt
01:56 halfss_ joined #salt
01:57 halfss_ Hi, how to be a contirbutor of salt? :)
01:57 austin987 joined #salt
02:00 mbarrien halfss_: Make a pull request. They've been pretty good about looking at and accepting code from random contributors.
02:01 abe_music joined #salt
02:01 halfss_ mbarrien: i have a few pull request, and the code have been merged. :)
02:02 krow joined #salt
02:08 diegows joined #salt
02:10 Katafalkas joined #salt
02:13 quickdry21 joined #salt
02:20 Ryan_Lane joined #salt
02:20 rcsheets joined #salt
02:20 krow joined #salt
02:23 yomilk joined #salt
02:23 drags joined #salt
02:26 krow joined #salt
02:28 ravibhure joined #salt
02:31 krow joined #salt
02:32 ajw0100 joined #salt
02:34 zain is there a way i can put an entire directory of files on the minion from the master without having to make a file.managed directive for each individual file?
02:35 abe_music joined #salt
02:36 EWDurbin zain: http://docs.saltstack.com/ref/states/all/salt.states.file.html#salt.states.file.recurse
02:36 zain ah, exactly what i'm looking for. thanks
02:36 xl1 joined #salt
02:37 krow joined #salt
02:40 alexthegraham_ joined #salt
02:40 alexthegraham_ does "test=True" *not* work for salt-call ?
02:43 krow joined #salt
02:45 alexthegraham *crickets*
02:45 EWDurbin alexthegraham: i'm taking a look
02:45 alexthegraham Thanks!~
02:46 EWDurbin but from what i see, there's no documentation for test=True, and the CLI for salt-call has no exposed support for it
02:47 EWDurbin http://docs.saltstack.com/ref/cli/salt-call.html
02:47 EWDurbin alexthegraham: i only see test parameter for the state modules http://docs.saltstack.com/ref/states/testing.html
02:47 alexthegraham I couldn't find anything in the docs for it either. Assuming undocumented = absent, should I assume that  running cmd.run "salt-call [command] test=True" actually doesn't *test*, it just runs the salt-call command?
02:48 EWDurbin you can `salt-call state.highstate test=True`
02:48 EWDurbin or salt-call state.sls test=True
02:48 krow joined #salt
02:48 EWDurbin or salt-call state.single test=True
02:48 EWDurbin but salt-call does not accept that parameter, the state modules do
02:49 alexthegraham OK, good to know.
02:49 EWDurbin alexthegraham: that make sense?
02:50 krow1 joined #salt
02:51 kalloc joined #salt
02:51 alexthegraham Yeah. I've been having some trouble getting state files to include a way to use pkg.group_install. The workaround I'd come up with was using cmd.run for it, but testing the command using salt-call ... test=True kept actually performing the command.
02:52 EWDurbin yes, because cmd.run does not accept a test parameter
02:52 EWDurbin nor does salt-call
02:52 alexthegraham ^ that. Any chance that's on the dev to-do list?
02:52 EWDurbin (shrug) i'm just a user :)
02:53 alexthegraham Well thanks for the info. You found a good way to get states to include group package installs, by any chance? :)
02:53 vejdmn joined #salt
02:54 EWDurbin alexthegraham, do you mean ala `yum groupinstall` or some other thing?
02:54 zain how can I debug "salt://myfile does not exist on the salt fileserver" errors? as far as I can tell, /srv/salt/myfile exists...
02:55 alexthegraham Yeah. Basically. salt.module.pkg.group_install does it, but I haven't figured out how to make a state actually use that.
02:55 alexthegraham http://docs.saltstack.com/ref/modules/all/salt.modules.yumpkg.html#module-salt.modules.yumpkg
02:55 krow joined #salt
02:56 alexthegraham zain, where are you referencing salt://myfile from? the root of file_roots ?
02:57 zain alexthegraham: file_roots is set to base: /srv/salt. my statefile is /srv/salt/foo/state.sls and myfile is at /srv/salt/foo/myfile
02:57 ipmb joined #salt
02:57 zain and the path i'm trying to reference in state.sls is salt://foo/myfile
02:57 clintberry joined #salt
03:05 ilbot3 joined #salt
03:05 Topic for #salt is now Welcome to #salt | Thanks to everyone who made SaltConf 2014 amazing! | 0.17.5 is the latest | Please be patient when asking questions as we are volunteers and may not have immediate answers | Channel logs are available at http://irclog.perlgeek.de/salt/
03:07 EWDurbin alexthegraham: do you have a gist?
03:07 EWDurbin and your goal is to test the state that's using cmd.run for the pkg.groupinstall?
03:08 basepi joined #salt
03:09 alexthegraham Well, my goal is to figure out the best way to get state.highstate to verify that the packages included in a yum group are installed - essentially to do pkg.group_installed and have it work.
03:09 bdf joined #salt
03:09 alexthegraham (that doesn't exist yet)
03:09 alexthegraham Gist coming up.
03:09 arnoldB joined #salt
03:09 sirtaj joined #salt
03:10 lionel joined #salt
03:11 Sypher joined #salt
03:11 EWDurbin alexthegraham: http://docs.saltstack.com/ref/states/all/salt.states.module.html is what you're looking for
03:12 EWDurbin sorry, i missed that earlier... getting a little late in the evening for me
03:12 EWDurbin so that's a state, that calls a module
03:13 EWDurbin alexthegraham: for instance... https://github.com/python/pypi-salt/blob/33458bdfa9f239136560e6c97a8e32bfa4cb3fae/provisioning/salt/roots/salt/nginx/init.sls#L91
03:13 LLKCKfan joined #salt
03:14 MedicalJaneParis joined #salt
03:14 ajw0100 joined #salt
03:15 andyshinn` joined #salt
03:15 bdf joined #salt
03:19 alexthegraham EWDurbin, thanks! I think I figured it out. I had it right hours ago (damnit), but the output from adding test=True doesn't indicate anything of value, so I thought I had it wrong.
03:20 alexthegraham using test=True I get back one "Comment:   Module function pkg.group_install is set to execute" for each pkg.group_install, but no indication of what the group package is.
03:20 alexthegraham Running it, though, seems to work great.
03:22 EWDurbin just ran a bunch of my states under test=True alexthegraham
03:22 EWDurbin very mixed results
03:22 rostam joined #salt
03:23 alexthegraham Yeah. That's what I keep seeing.
03:23 EWDurbin :-/
03:23 alexthegraham test=True seems to prevent changes being made, but doesn't actually provide much useful information.
03:23 alexthegraham Or it's useful, just inconsistent.
03:23 alexthegraham Thanks for your help EW.
03:24 jalbretsen joined #salt
03:30 Linz joined #salt
03:30 ravibhure1 joined #salt
03:31 Ryan_Lane joined #salt
03:35 joehoyle joined #salt
03:37 dren joined #salt
03:38 Linz joined #salt
03:42 Furao_ joined #salt
03:52 clintberry joined #salt
03:54 Linz joined #salt
03:59 joehillen joined #salt
04:01 krow joined #salt
04:03 krow joined #salt
04:04 shadowsun joined #salt
04:11 schimmy joined #salt
04:18 dancat joined #salt
04:21 thayne joined #salt
04:34 middleman_ joined #salt
04:34 mbarrien joined #salt
04:45 rojem joined #salt
04:46 rojem joined #salt
04:46 vejdmn joined #salt
04:46 rojem joined #salt
04:53 krow joined #salt
04:57 clintberry joined #salt
05:02 torrancew joined #salt
05:02 yomilk joined #salt
05:03 zz_cro joined #salt
05:08 rojem joined #salt
05:10 darless Am I missing something? I have git.latest (foce, force_checkout, and always_fetch) all set to True, but I still get the "The following untracked working tree files would be overwritten by merge
05:13 anuvrat joined #salt
05:15 rhand joined #salt
05:21 TheSojourner joined #salt
05:21 TheSojourner joined #salt
05:24 mgw joined #salt
05:47 jeremyfelt joined #salt
05:50 passenger joined #salt
05:50 passenger my passenger pkg fails to install the passenger
05:50 passenger Phusion Passenger4.0.37:   pkg:     - installed   service:     - running     - require:       - pkg: Phusion Passenger4.0.37
05:51 passenger the above is my passenger.sls
05:55 dirjax joined #salt
05:58 passenger let me know how to kill the nginx process using pkg. thanks in advance
05:59 kermit joined #salt
05:59 LLKCKfan joined #salt
06:01 fllr joined #salt
06:08 joehoyle joined #salt
06:13 ajw0100 joined #salt
06:17 Psyke joined #salt
06:18 n8n joined #salt
06:18 abique joined #salt
06:20 joehoyle joined #salt
06:22 passenger how to install passenger in salt minion using pkg
06:34 thayne joined #salt
06:37 scott_w joined #salt
06:46 joehillen joined #salt
06:55 kalloc_ joined #salt
07:13 sinh__ joined #salt
07:15 abique Hi, I wonder how could I have something like that: http://pastebin.com/Dj9gZSPK ? What is the right syntax? Thanks!
07:17 scott_w joined #salt
07:19 troyready joined #salt
07:21 tomspur joined #salt
07:21 tomspur joined #salt
07:22 nebuchadnezzar joined #salt
07:25 fllr joined #salt
07:25 ninkotech__ joined #salt
07:40 asdff joined #salt
07:53 clintberry anyone in here tonight?
08:03 balboah joined #salt
08:04 mike25ro @my side is morning :)
08:08 harobed joined #salt
08:11 joehoyle joined #salt
08:13 Shenril joined #salt
08:16 Iwirada joined #salt
08:17 zooz joined #salt
08:17 tulu joined #salt
08:19 Kenzor joined #salt
08:26 LLKCKfan joined #salt
08:26 fllr joined #salt
08:27 mikkn_ Does anyone have a good tip on how to automatically test sls files for "development"?
08:27 Furao mikkn_: I wrote a test framework for salt states
08:27 Furao it's quite complicated
08:27 Furao I even test monitoring and backups
08:27 Furao a single test run for everything takes 16.5 hours
08:27 Furao on SSD
08:28 mikkn_ Furao: Yeah, that's my impression as well. Seems to be rather complicated to set up if you don't want to do "full" runs every time you push
08:28 Furao I have a wrapper around my framework that let developers to only run partial tests
08:28 mikkn_ I would like to be able to test states on only the affected servers
08:29 Furao on specific servers? well use gitfs + environment and test=True
08:30 Furao for some client i have a testing branch that if jenkins build pass it goes into staging, and jenkins run test=True on it. and when client is ready they can state.highstate
08:30 mbarrien joined #salt
08:31 mikkn_ Hmm, that does sound "ok"
08:31 Furao sometimes they need to update pillars, I'm not comfortable yet to have everything 100% automated
08:32 Furao in term of automatic deployments
08:36 slav0nic joined #salt
08:36 slav0nic joined #salt
08:38 mike25ro hey guys in  a pillar file .. can i comment some lines using #my comment     or   should i use {#  my comment  #}  ?
08:40 Furao mike25ro: use {# #}
08:40 Furao it's safer
08:41 mike25ro Furao: thanks
08:41 passenger my pkg fails to install the passenger. do let me know what should i need to do
08:41 gw joined #salt
08:41 gw joined #salt
08:42 Furao I can't do something, please help me. but I won't tell what the error is nor what I'm trying to do :)
08:45 giantlock joined #salt
08:45 mike25ro haha Furao :)
08:50 yomilk joined #salt
08:53 miguitas joined #salt
08:56 passenger my pkg fails to install the passenger. do let me know what should i need to do
08:56 Furao passenger: check your minion logs
08:56 mgw joined #salt
08:57 Furao you don't help yourself, you don't show any error message, no stack trace, no .sls
08:58 kadel joined #salt
08:58 passenger thanks i will paste the pastebin link here
09:01 passenger furao: please take a note of this link http://pastebin.com/XmYWdpnH
09:03 Furao check minion logs
09:03 Furao run it @ debug level
09:04 passenger let me know how to check the minion logs
09:04 Furao http://salt.readthedocs.org/en/latest/topics/troubleshooting/minion.html
09:06 dimens joined #salt
09:08 dimens left #salt
09:09 Zuru joined #salt
09:13 mgw1 joined #salt
09:15 Psi-Jack joined #salt
09:18 kadel joined #salt
09:26 fllr joined #salt
09:27 MohShami joined #salt
09:28 MohShami huy guys, I have 2 mysql commands watching another state, they are still being run even if the watched state doesn't change, am I missing something?
09:30 Furao MohShami: use cmd.wait instead of cmd.run
09:30 Furao and specify -watch
09:31 MohShami Thanks Furao, that's what I usually do, but this time I'm using module.run with -watch, does module.run support watch to begin with? I can't find any documentation
09:31 Furao use module.wait
09:32 Poooogles joined #salt
09:32 MohShami oh, didn't find that either, thanks for the tip mate :)
09:32 MohShami btw, can I use module.run with a select query and send the result to another state?
09:33 Poooogles joined #salt
09:33 Furao nope
09:33 MohShami that would have been nice :(
09:33 Furao well, maybe
09:33 ChrU joined #salt
09:33 Furao with {% set output = salt['modname.funcname'](someargs…) %}
09:33 Furao and pass output to something that can consume the output
09:33 Furao in such case you should write your own state
09:33 MohShami I'm building an sls file that would read my DNS information from a database then write NSD configuration files
09:34 Furao use ext_pillars for that
09:34 MohShami I already used that, thought there would be a cleaner way, I'll check ext_pillars
09:34 MohShami thanks mate :)
09:34 MohShami my current solution is kinda slow, looking for ways to improve speed
09:35 MohShami module.wait fixed it for me, thanks Furao :)
09:36 Furao no problem
09:40 MohShami Furao, Is it possible to run a state only if something changes?
09:40 mike25ro MohShami: i think that only by running a state... you know that changes are applicable... no?
09:41 Furao exact
09:41 MohShami let me show you my sls file
09:41 MohShami to better explain my point
09:41 MohShami http://pastebin.ca/2638574
09:42 MohShami in this config, if no domain had the modified field set to 1, I don't want to go through the hassle of creating nsd.conf
09:43 MohShami trying to scale the process as much as possible
09:43 Furao not sure watch_in for a file module have any effect
09:43 Psi-Jack joined #salt
09:43 MohShami thought as much, but wanted to give it a shot
09:44 MohShami the only way I could think of it is to do a "select count" query
09:44 MohShami inside an if
09:44 mikkn_ Furao: Do you use the integration/unit test framework that ships with salt to do any of your testing? :)
09:44 Furao mikkn_: no as I do a lot more than that
09:44 MohShami nope, I'm rather new to salt and still do manual testing :@
09:44 MohShami :$*
09:44 Furao I shown it to utahdave
09:45 MohShami just thought of something
09:46 MohShami I could set a variable to 0, and inside the loop set it to 0, then when I generate nsd.conf check the variable first
09:46 MohShami that should do it, right?
09:46 DaveQB joined #salt
09:46 fragamus joined #salt
09:47 mikkn_ We're thinking about running integration/unit-tests on the developer machines for custom states and modules, and doing full env-based tests on pushing to the development branch.
09:49 Furao I don't know what you deploy exactly, but in my case, as we manage everything (that include the OS low-level stuff), we can only test in isolated VM
09:50 Furao as each test need to run on a "blank" target
09:50 Furao to make sure nothing had been forgot in term of requirements
09:51 Furao each sls are tested state.sls individually, to be sure that no include are missing
09:52 Furao and there is an integration test that also check if all process are running, ports open, http get return expect results, backup script run and there is content into the archive
09:52 Furao etc
09:52 mikkn_ Furao: Yeah, that's also something I'd like to do, but I can't really figure out a good way to do "automatic" targetting. What I would ideally like to do is to test single states into isolated environments, since usually a commit/feature branch will only involve a single state/stategroup which touches a single server type.
09:53 jeddi joined #salt
09:53 mikkn_ But I don't want to run the 2y old states that have just worked unless I'm doing a full run
09:56 clintberry joined #salt
09:57 Furao mikkn_: I let my dev target which set of states to test
09:57 Ztyx joined #salt
09:57 Furao but I still keep a complete test job as they sometimes don't think of the implication of a change
10:02 mikkn_ Furao: Thank you :)
10:04 Ztyx Hey, I'm evaluating Salt for provisioning our cloud VMs. There will be a group (role) of VMs that I'll ship a secret/password to. Seems like most targeting in Salt is done using client-side information (platform, pillars etc.). Is there any way I can be sure no minions can claim to be part of the group/role (and thus get access to the password)?
10:05 Ztyx I'm thinking of targetting based on the public key of a minion, or similar.
10:05 Furao Ztyx: pillars is rendered in the master. if you design your pillar right it's relatively safe
10:06 scott_w joined #salt
10:06 Ztyx Furao: But pillars targeting is based on minion information, right?
10:06 Furao for password that used only within a VM (no need to be shared) I used a module that I created that randomly create password
10:07 Furao Ztyx: it's based on the id
10:07 Furao a minion can change it's id
10:07 Furao if an attacker can control the VM
10:07 Furao but the master will reject the minion if it try to use an existing id
10:07 Ztyx Great. This is exactly the answer I was looking for. Thanks!
10:07 Furao and the master won't automatically accept a new id, unless you configure the master to do that
10:08 Furao salt-cloud create key and accept it for you in the master
10:08 Ztyx Yeah. Sweet. This means Salt is still an option ;)
10:10 jeddi More a distro question ... I've got a stack of states for my debian boxen, and now trying to generalise them to cope with ubuntu machines.   the git pkg is giving me grief - even though apt-cache show git (on ubuntu) shows that it's there, with 12.10 - my state that works fine on debian is failing on ubunt with a 'requisite not found' error.  i note 'git-core' pkg is also available on ubuntu, but can't work out why it's choking on 'git
10:10 jeddi .. which is actually installed anyway.
10:10 rojem joined #salt
10:10 krow joined #salt
10:11 Furao git-core is the old name of git client
10:11 Furao maybe it's just a virtual pkg
10:12 Furao old name of pkg, I mean
10:12 Furao you should run salt-minion @ debug level and check what happens exactly
10:15 ggoZ joined #salt
10:17 jeddi Furao: argh .. debug level .. yeah .. okay ..  ;|
10:20 MrTango joined #salt
10:25 viq or salt-call, that spits out a lot of messages as well
10:26 viq even without the full debug
10:26 fllr joined #salt
10:28 MohShami joined #salt
10:31 gildegoma joined #salt
10:34 middleman_ joined #salt
10:37 babilen rgr: Hi all -- I am encountering a TypeError when running highstate and am not sure how to debug it further. You can see details in http://paste.debian.net/81335/ -- What is going on? How would I debug this? What did I do wrong and how do I mitigate the problem?
10:37 babilen s/rgr://
10:38 Johbe joined #salt
10:39 babilen Furao: btw, thank you for the basic idea and hints on getting the pillars running
10:40 Furao maybe you'll prefer that top.sls for roles https://gist.github.com/bclermont/18d278439041726ee1b2
10:40 viq babilen: random guess of that .1 messing you up
10:41 babilen viq: Let me try that .. I can easily remove that
10:41 Furao I think it's because top.sls render as a an empty string
10:41 Furao it render as {'base': ''}
10:41 Furao as this minion don't have any active role
10:41 Furao oh no wait
10:42 Furao the base structure is wrong
10:42 Furao it must be {'base': {'*': ['role1', 'role2']}}
10:42 MohShami so why not add the ability to send the save the result array of a state?
10:42 MohShami just a question that popped to my mind
10:42 Furao base can't be just a list
10:43 Furao and you don't need .iteritems()
10:43 babilen *sigh*
10:43 babilen Furao: Okay, so what do I have to change?
10:44 Furao sorry
10:44 Furao I didn't noticed the ``role, active``
10:45 babilen Furao: Yeah, the problem is that I want roles to be merged (i.e. I want to add additional roles later) and using that structure was the only way to make that happen
10:45 babilen (as of now as salt doesn't yet support proper merging strategies)
10:45 Furao babilen: https://gist.github.com/bclermont/44c771eb344d9d7f77ff
10:46 Furao but if you plan to have a top.sls like this look at https://gist.github.com/bclermont/18d278439041726ee1b2
10:47 babilen Furao: I will, I tried to use a simple one first and then add more logic.
10:50 babilen Furao: Unfortunately I still get that error with the top.sls that includes the '*': (and which you pasted)
10:51 londo joined #salt
10:52 babilen Ah, ... "-l debug" provides some important hints. "No matching sls found for 'misc.abaddon.1' in env 'base'" (and so on for most roles) -- should have thought about that. Can i somehow make this conditionally? That is: Include if role is defined, active and the state is actually available?
10:54 ninkotech__ joined #salt
10:54 diegows joined #salt
10:55 babilen Furao: And I can't claim that I understand your top.sls completely -- What are, in your opinion, the major advantages over mine? I mean I can't simply use the list of roles so that has to be ruled out, but I don't quite understand the importance of "local mode" and would therefore be grateful for an explanation
10:55 qba73 joined #salt
10:56 Furao you use gitfs?
10:56 Furao local mode is when you run salt-call --local (masterless)
10:57 Furao it's true that I probably cover more than your actual requirements
10:57 clintberry joined #salt
10:58 miguitas joined #salt
10:58 babilen No, I do use GitFS, is that because calling salt-call on the minion requires the branch to be renamed? (master -> base)
10:58 MohShami I'm running the following on the VM I use to build and test new state information "salt-call state.sls dns.nsd" and it is taking a LONG time before it even starts, anything I should look at? It's a masterless setup
10:58 Furao the master -> base thing is for gitfs only
10:59 babilen Furao: yes, I do use GitFS already, but my understanding was that the "master" branch in Git is already mapped to the "base" environment and that I don't have to do that explicitly.
11:00 Furao yes and no, in my case i have a pillar key to let admin specify which branch they want to deploy
11:01 Furao so if they say master it need to be renamed to base
11:02 babilen But in my case I don't have to do that yet and the renaming is done already?
11:03 tru_tru joined #salt
11:03 Furao no you don't need this
11:08 babilen Can I test if a state is available in top.sls? I don't necessarily want to create a lot of empty init.sls files, but would prefer to create only those that are actually in use.
11:11 david_a joined #salt
11:11 Furao babilen: http://docs.saltstack.com/ref/modules/all/salt.modules.state.html#salt.modules.state.show_top
11:12 Furao oups
11:12 Furao sorry
11:12 Furao I misread your question :)
11:13 ggoZ joined #salt
11:13 qybl joined #salt
11:17 babilen Furao: No, more like: Would there be a foo.bar state or has that not been defined?
11:18 xinkeT joined #salt
11:19 Dinde joined #salt
11:21 ninkotech__ joined #salt
11:22 EvaSDK joined #salt
11:22 EvaSDK joined #salt
11:22 passenger furao:
11:23 passenger furao: i didn't anyway to install passenger, i have installed  apache nginx by this method
11:23 joehoyle joined #salt
11:23 passenger and let me know tp upgrade the installed package
11:24 beardo_ joined #salt
11:25 kolaman joined #salt
11:26 fllr joined #salt
11:28 joehoyle_ joined #salt
11:28 elfixit joined #salt
11:28 ololo joined #salt
11:28 qybl joined #salt
11:29 scott_w joined #salt
11:32 LucasCozy joined #salt
11:36 LucasCozy How i can add an environnement variables in cmd_async via the python Api.
11:37 LucasCozy I use client_salt.cmd_async(minion, module, [arg], env={"PATH": "/mycompletpath"}).
11:37 LucasCozy Not working at all ;-)
11:38 LucasCozy Is it only possible?
11:39 rofl____ joined #salt
11:42 linjan joined #salt
11:43 Poooogle_ joined #salt
11:45 pwistrand joined #salt
11:49 Poooogles joined #salt
11:52 Katafalkas joined #salt
11:53 qybl joined #salt
11:57 ninkotech__ joined #salt
11:58 clintberry joined #salt
11:59 zzzirk joined #salt
12:05 gadams999 left #salt
12:05 carlos joined #salt
12:16 pwistrand I'm trying to setup a masterless minion running out of a local folder but get "No Top file or external nodes data matches found" when issuing "salt-call --local state.highstate". can anyone help?
12:18 backjlack joined #salt
12:20 vejdmn joined #salt
12:29 sroegner joined #salt
12:30 scott_w joined #salt
12:36 n8n joined #salt
12:36 brianhicks joined #salt
12:38 gildegoma joined #salt
12:39 MohShami joined #salt
12:40 debaser joined #salt
12:45 dsumsky joined #salt
12:46 mkoistinen joined #salt
12:47 mkoistinen left #salt
12:48 favadi joined #salt
12:50 dsumsky help
12:50 dsumsky ?
12:50 ekristen joined #salt
12:56 dhananjay_ joined #salt
12:58 dsmk80 joined #salt
12:59 clintberry joined #salt
12:59 dsmk80 btw, are somewhere presentations from SaltConf2014 published?
13:00 zz_cro joined #salt
13:02 blee joined #salt
13:06 gadams999 joined #salt
13:09 passenger furao: please do open this link http://pastebin.com/D87cshG8. i tried to install jboss. but it failed
13:18 mgw1 joined #salt
13:21 tessellare joined #salt
13:26 fllr joined #salt
13:27 halfss joined #salt
13:30 sroegner joined #salt
13:31 sroegner joined #salt
13:32 blee_ joined #salt
13:33 BRYANT__ joined #salt
13:34 jpcw joined #salt
13:34 n8n joined #salt
13:35 kadel joined #salt
13:36 sroegner joined #salt
13:37 kalloc joined #salt
13:37 ipmb joined #salt
13:38 kalloc joined #salt
13:38 halfss joined #salt
13:39 halfss joined #salt
13:39 blee joined #salt
13:41 scott_w joined #salt
13:42 AdamSewell joined #salt
13:42 AdamSewell joined #salt
13:46 ninkotech__ joined #salt
13:49 mgw1 joined #salt
13:50 jaimed joined #salt
13:55 Mua joined #salt
13:56 mike25ro hi guys .. is it possible to allow a normal user on salt-master to run only some states? if yes can you give me a hint please?
13:58 Mua_ joined #salt
13:58 bronsen joined #salt
14:00 clintberry joined #salt
14:01 bezaban mike25ro: you could regulate that in sudo, as for a salt way I don't know :)
14:01 jaimed joined #salt
14:05 debaser mike25ro: have you tried external_auth system?
14:05 timoguin joined #salt
14:07 Gifflen joined #salt
14:08 scott_w joined #salt
14:08 Ahlee external_auth is the correct way
14:08 Ahlee mike25ro: http://docs.saltstack.com/topics/eauth/index.html
14:10 kalloc_ joined #salt
14:10 Gifflen joined #salt
14:14 kalloc joined #salt
14:15 pass_by_value left #salt
14:20 racooper joined #salt
14:24 brianhicks joined #salt
14:24 juicer2 joined #salt
14:26 fllr joined #salt
14:27 premera joined #salt
14:27 cweisel joined #salt
14:30 andresr joined #salt
14:31 ipmb joined #salt
14:33 ccase joined #salt
14:33 elfixit joined #salt
14:34 mike25ro Ahlee: thanks...
14:34 mike25ro debaser: i haven't .. i am looking into it now
14:39 fllr joined #salt
14:40 gadams999 left #salt
14:41 fllr joined #salt
14:41 tyler-baker joined #salt
14:41 tyler-baker joined #salt
14:42 swa joined #salt
14:46 babilen What happens if I declare the same state using different globs? (e.g. "'*': - foo" and "'specific-host: - foo")
14:47 ninkotech_ joined #salt
14:48 TJ1980 joined #salt
14:49 TJ1980 #Curious, when the Salt 2014.1.0 release is likely to occur?
14:49 jslatts joined #salt
14:51 mjulian joined #salt
14:51 mjulian joined #salt
15:00 quickdry21 joined #salt
15:01 SkyRocknRoll__ joined #salt
15:04 mpanetta joined #salt
15:05 GradysGhost joined #salt
15:05 mpanetta joined #salt
15:06 kedo39 joined #salt
15:08 ninkotech_ joined #salt
15:09 rojem joined #salt
15:12 joehoyle joined #salt
15:13 joehoyle_ joined #salt
15:14 pass_by_value joined #salt
15:14 joehoyle joined #salt
15:15 brianhicks joined #salt
15:15 joehoyle joined #salt
15:15 Ztyx Are all files from the file server always accessible from the minions all the time? That is, should I make sure that sensitive files are not exposed using the file server?
15:15 d10n joined #salt
15:16 joehoyle joined #salt
15:21 kermit joined #salt
15:23 rostam joined #salt
15:23 vejdmn joined #salt
15:25 debaser left #salt
15:31 Gordonz joined #salt
15:31 abe_music joined #salt
15:32 Gordonz joined #salt
15:32 babilen Ztyx: Use pillars for that and target only those minions that should have access to that data
15:34 ggoZ joined #salt
15:37 forrest joined #salt
15:37 cnelsonsic joined #salt
15:37 mbarrien joined #salt
15:38 Ztyx babilen: can I put file content in pillars?
15:42 fllr joined #salt
15:43 tedski Ztyx: you can use pillar data in jinja templates in files
15:43 [diecast] joined #salt
15:43 abique joined #salt
15:45 alunduil joined #salt
15:46 Ztyx tedski: I see, in my case I'm thinking of a larger file that I'd like to not expose to all machines.
15:47 babilen Ztyx: https://github.com/saltstack/salt/issues/1543
15:48 tedski Ztyx: yeah
15:49 Ztyx babilen: +1 Thanks
15:50 babilen Ztyx: Also take a look at related issues -- I'm sure that there is more appropriate documentation .. somehwere
15:51 Ztyx babilen: Yeah, https://github.com/saltstack/salt/issues/9569 seems related for example.
15:51 babilen Ztyx: Sorry, pretty new to salt so not entirely sure, but targetting data on minions was a pretty large concern for us and I made sure that it is somehow possible
15:51 babilen Ztyx: Yeah, that looks even better
15:52 fxdgear joined #salt
15:53 Ztyx babilen: I see. Appreciate your input anyhow! I come from an Ansible background. Since you are new I'm curious to hear, did you consider Ansible as an alternative?
15:54 babilen Ztyx: I did, yes. A colleague was already looking into Salt and a few articles I read gave me the feeling as if Salt would be the better solution so I started with that.
15:55 babilen Ztyx: Once I have a halfway decent setup in salt I will investigate for a week/two weeks how the same would look in Ansible and then decide what is better suited
15:55 babilen Ztyx: What is your impression?
15:56 babilen I do not really like to make decisions without knowing the options, but I had to start somewhere and, well, salt somehow "felt" better
15:57 forrest In short, if Chef and Puppet are dinosaurs roaming around, Salt is the meteor about the enter orbit and wipe them off the face of the earth.
15:57 Shish XD
15:57 forrest had to use that to describe Salt to someone after they were confused about it's 'production readiness'.
15:58 Ztyx babilen: Aight. I'd love to hear your thoughts. My impressions so far is that Ansible is slightly easier to learn and get productive in, while Salt is more powerful and has more thorough documentation (which is slightly less organized).
15:58 Ztyx forrest: Thanks for your constructive input! :-P
15:58 forrest Ztyx, I have no idea what you guys are talking about actually, just jumped in
15:58 forrest looks like it was good timing though
15:59 Ztyx babilen: Also, I wasn't entirely happy about the Ansible community. So far, I feel Salt is a more active and welcoming community.
15:59 babilen Ztyx: I agree with you ont eh less organised part -- Right now a common thing I run into is: Hmm, I've read that somewhere, but where? It must be in ... no, not here. Then it must be .... no, not there either.
15:59 forrest babilen, Ztyx, it's actually on my 'list', as well as several of the developers to restructure how the docs look
15:59 babilen Ztyx: It is slightly annoying when I try to show Salt to colleagues and they make fun of me because I can't navigate the documentation :)
16:00 Ztyx babilen: Heh, I think the index in the left column here sums it up pretty well: http://salt.readthedocs.org/en/latest/
16:00 forrest scalability-junk, https://github.com/gravyboat/salt/blob/develop/doc/topics/best_practices.rst is where I'm at right now
16:00 babilen Ztyx: Yeah, my gut feeling was based a lot on how I perceived the different communities to act and I had the impression as if the salt one was more open (also in the code sense), welcoming and more active
16:00 Ztyx forrest: Good to hear it's on someone's list!
16:01 forrest babilen, Ztyx, here is the issue: https://github.com/saltstack/salt/issues/9912
16:01 forrest if you guys have input, or create another issue and it relates, feel free to link it there.
16:01 Poooogle_ joined #salt
16:01 clintberry joined #salt
16:01 babilen forrest: Would you appreciate it if I took notes whenever I think "I am looking for ... would expect a ... doc" ?
16:01 forrest babilen yes
16:01 forrest that would be fantastic
16:01 Ztyx forrest: I subscribed to the issue now.
16:01 Poooogles joined #salt
16:02 forrest I've already done a LOT of that as I've gone through, but the more unique perspectives we can get, the better we can make the docs.
16:02 forrest Ztyx, great!
16:02 Ztyx forrest: That "best practices" page was great! Thanks!
16:02 forrest Ztyx, it's nowhere near done :(
16:02 forrest I didn't finish it up this weekend like I was planning
16:02 ninkotech__ joined #salt
16:03 babilen forrest: Okay, I will start doing that. It becomes quite apparent when i like to show other developers/admins how to use what we have so far ... just the other day I wanted to explain grains and searched for 5 minutes for a simple "This is all the data in the grain", "this is how you call it on the CLI", "this is the output" and "this is how you use it for targetting" doc.
16:03 Ztyx forrest: I'll see if I can submit a couple of smaller issues, too.
16:03 babilen Ztyx: And I agree on that sidebar ...
16:03 forrest babilen, ahh yea, I don't think there is actually an awesome page for 'grains in depth' or even getting started with them aside from the walkthroughs, if you open an issue on that it would be great
16:04 forrest Ztyx, awesome! I primarily work the docs as my contributions, so I'm always happy to see more issues.
16:04 scalability-junk forrest: issue no vim subdirectories
16:04 forrest scalability-junk, ?
16:04 scalability-junk wanted to fork, but github seems to don't like that
16:04 forrest you should just be able to fork my repo
16:04 babilen forrest: Will do - but that was just one example. I worked through the initial walkthroughs (which are great, but should be factored out) and then started "free reading" and now always run into the "where did I see that?" question
16:04 Psyke joined #salt
16:04 fragamus joined #salt
16:04 forrest Corey said he had no problem
16:04 JasonSwindle joined #salt
16:04 dave_den joined #salt
16:05 forrest babilen, yea that is totally understandable, and I even run into it still. During Saltconf the main maintener of the docs and I talked, and that's where that issue came from. We might try to do something similar to the django docs: https://docs.djangoproject.com/en/1.6/
16:05 forrest I really like how they are laid out. It always comes back to time though :P
16:06 cachedout joined #salt
16:06 forrest Even if you don't want to contribute, the issues are still great because it helps me see how people are navigating the docs or being confused. Sometimes it's as easy as linking between sections that are related.
16:06 forrest contribute to the docs that is, the issues are great contributions to me!
16:07 Shish is there a way of seeing if a module is available from inside a sls? Specifically the selinux module registers itself as unavailable if selinux is installed but disabled, so I'd like my "selinux.mode: permissive" to be only attempted if the selinux module is there
16:07 scalability-junk can't edit and fork seems strange
16:07 forrest you can't fork?
16:07 scalability-junk anyway when you give the vim directory tree I think on each line except the first the vim subdir is missing
16:07 forrest that's weird, I don't have it locked...
16:08 forrest which line?
16:08 forrest sorry I'm at work right now so a bit distracted
16:08 babilen forrest: In a way it would be great if it could be organised a bit more like the Python documentation. There are Sections "Tutorials & Getting Started" (articles that tell one story and hint at what can be done), "Modules" - *One* article for grains/pillars/states with all the gory and less gory details written like the official Python module documentation and *also* including pydoc like information, "How Tos" - Articles describing how to do ...
16:08 SkyRocknRoll__ joined #salt
16:08 babilen ... specific tasks, ...
16:08 forrest oh gotcha
16:08 forrest thanks
16:08 scalability-junk forrest: 48-53
16:08 babilen forrest: Sorry, just thinking ... but the main problem was/is that I never know what is official documentation, what type of information can be found where and so on.
16:09 forrest scalability-junk, yea I see it now
16:09 Linz joined #salt
16:09 Gifflen_ joined #salt
16:09 scalability-junk great will bookmark for a free moment ;)
16:09 forrest babilen, In what way? Is it due to a lack of navigation within the docs themselves, or via confusing titles when doing google searches and such?
16:09 forrest scalability-junk, sounds good!
16:09 babilen forrest: I am quite aware that organising documentation and coming up with a sensible and maintainable structure is hard, but it is worth it.
16:10 forrest babilen, oh yea totally, that's where we want to get
16:10 forrest babilen, when you say you don't know what is official documentation though, is that because there is the readthedocs content, and the docs.saltstack content? Or another reason?
16:12 Shish "if 'selinux.getenforce' in __salt__" seems to be used in the python file, but __salt__ seems to not be defined in the sls context :(
16:12 scristian joined #salt
16:13 mgw joined #salt
16:13 forrest Shish, I'm trying to think of a good way to do that..
16:14 dave_den Shish: use 'salt' in sls files
16:14 babilen forrest: Hmm, well. When i was reading the walkthorughs I had the feeling that I am somehow following the "official path", then read the referenced documents (Pillars Walkthrough, Targeting) ... With Pillars there is http://docs.saltstack.com/topics/tutorials/pillar.html (which I read) and then http://salt.readthedocs.org/en/latest/topics/pillar/ -- I was a bit confused why the latter exists and why, generally speaking, the documentation was split ...
16:14 babilen ... that way.
16:14 forrest there are work arounds, where you could run a check to see if the output of a selinux check command was a certain value
16:15 forrest babilen, ahh gotcha, yea a lot of that stuff in the pillar topic page should be in the walkthrough.
16:15 babilen forrest: But I am in no way prepared for this discussion, let me take some notes and come up with sensible suggestions
16:16 dave_den forrest: I think shish means that __salt__ dunder does not exist in sls state jinja files like it does in custom modules/grains.
16:16 dave_den ?
16:16 forrest cool, if you have time real quick can you open an issue on the pillar topic versus walkthrough?
16:16 forrest dave_den, oh yea, his original goal was to check if the module was available, and ONLY if so run his command to modify the selinux context.
16:17 babilen forrest: It was really the "Okay, will there always be the split into walkthrough and module documentation? Then I discovered http://docs.saltstack.com/ref/modules/all/salt.modules.pillar.html and there was a third thing.
16:17 forrest yea, it's not explained very well that there are docs for the modules, and docs for the states
16:17 Shish "{% if 'selinux.getenforce' in salt %}" around the selinux bits seems to work :)  though something does seem weird about that approach and I can't quite put my finger on it...
16:17 forrest Shish, that's surprisingly elegant
16:18 babilen forrest: In a way I would have expected the "Pillar of Salt" one and the API one to be merged (in the sense of, say, the re module in Python has http://docs.python.org/3.3/library/re.html and http://docs.python.org/2/howto/regex.html
16:19 dave_den Shish: for the purpose of checking that a module can run on the minion, that is a god approach. modules should have chacks in their __virtual__() functions that do conditional checks and if the minion does not have the required dependencies satisfied, then the module is not loaded.
16:19 dave_den good
16:20 babilen forrest: Look at http://docs.python.org/2/index.html -- it is fairly obvious where to find which information (yes, you are probably quite familiar with that already) -- The documentation landing page really shouldn't be larger (I find the Django one to not be easily navigatable)
16:21 forrest babilen, gotcha, yea the way the salt.modules.X.html files work is they are pulled straight from the python files themselves.
16:21 dave_den Shish: in the case of the selinumx module: https://github.com/saltstack/salt/blob/develop/salt/modules/selinux.py#L24
16:21 Shish selinux seems an unusual case though -- it is installed and all the commands exist (if it's set to "enabled" or "permissive", salt works fine), it's just when it's set to "disabled" that the module claims dependencies are missing
16:21 dave_den man, i cannot type today
16:22 gadams999 joined #salt
16:22 Ztyx forrest: Could you pinpoint me to where the order of the chapters is defined?
16:23 babilen forrest: Yeah, but Python has sphinx to write documentation which could be combined fairly easily with autodoc (i.e. http://sphinx-doc.org/ext/autodoc.html )
16:23 jeremyBass1 joined #salt
16:25 dave_den salt already compiles documentation from docstrings
16:26 MedicalJaneParis joined #salt
16:27 faldridge joined #salt
16:27 babilen forrest: That way it is integrated into the source and also provides code documentation http://hg.python.org/cpython/file/400a8e4599d9/Lib/re.py for example is pretty obvious
16:27 babilen forrest: Anyway, give me a bit of time and I shall keep it in the back of my head
16:28 forrest Ztyx, It really isn't defined anywhere other than the main http://docs.saltstack.com/ page
16:29 forrest babilen, Right and Salt is using a lot of that style of stuff. Sounds good! Feel free to create issues because then I'll actually remember it instead of documenting it on existing ones, or emailing myself :P
16:29 Psyke joined #salt
16:33 mbarrien joined #salt
16:33 Ztyx forrest: I see. Expect an issue ;)
16:34 forrest Ztyx, cool!
16:38 teskew joined #salt
16:40 ipmb_ joined #salt
16:41 bemehow joined #salt
16:42 ninkotech__ joined #salt
16:42 babilen forrest: http://hg.python.org/cpython/file/400a8e4599d9/Doc/library/re.rst is the other part of the documentation ... What is salt using right now to create the documentation?
16:43 forrest sphinx
16:43 babilen Ah, well ;)
16:43 babilen forrest: I ask because I *hate* it when the first solution someone proposes is: Use a different tool! All will be good!
16:43 babilen Anyway, /me has to run now.
16:44 forrest yea the docs are all built from the doc strings and RST files, so it's pretty simple
16:44 forrest ok later
16:44 forrest aka autodoc.
16:45 [diecast] joined #salt
16:48 bemehow joined #salt
16:48 Dale12 joined #salt
16:48 halfss joined #salt
16:51 jslatts joined #salt
16:51 s0undt3ch joined #salt
16:51 amckinley joined #salt
16:53 kaptk2 joined #salt
16:56 gmtech joined #salt
16:56 gmtech Hey all
16:56 gmtech how do I watch a pg_hba.present ?
16:56 frasergraham joined #salt
16:56 gmtech Is it:  watch:  - pg_hba: present ?
16:57 vbabiy_ joined #salt
16:57 dave_den gmtech: it is int he form of "state_module:  ID declaration".
16:58 teskew joined #salt
16:58 gmtech so… pg_hba.present:  ID ?? or pg_hba:  ID ?
16:58 dave_den i don't think there's a state module called pg_hba, but if there was it would be - pg_hba: ID
16:59 pwistrand joined #salt
17:03 m87carlson joined #salt
17:05 Ztyx left #salt
17:06 d10n joined #salt
17:08 schimmy joined #salt
17:08 gmtech lovely thanks
17:09 jslatts joined #salt
17:09 jeremyfelt joined #salt
17:11 kalloc_ joined #salt
17:16 kalloc joined #salt
17:17 gnugnu1 joined #salt
17:22 vbabiy_ joined #salt
17:23 bemehow_ joined #salt
17:24 scott_w joined #salt
17:24 jalbretsen joined #salt
17:25 KyleG joined #salt
17:25 KyleG joined #salt
17:27 HeadAIX joined #salt
17:35 MedicalJaneParis joined #salt
17:37 alekibango joined #salt
17:37 Iwirada left #salt
17:41 yurei joined #salt
17:47 Ryan_Lane1 joined #salt
17:47 gnugnu1 joined #salt
17:50 thayne joined #salt
17:50 Katafalkas joined #salt
17:50 Ryan_Lane1 joined #salt
17:51 rhand joined #salt
17:54 puppet joined #salt
18:01 rojem joined #salt
18:02 mgw1 joined #salt
18:10 ggoZ joined #salt
18:14 elfixit joined #salt
18:14 gnugnu1 joined #salt
18:16 JasonSwindle joined #salt
18:19 Ryan_Lane1 joined #salt
18:24 frasergraham when I try to run the salt master as non-root user the commands just hang, what am I doing wrong?
18:24 oleg joined #salt
18:24 frasergraham i set the user in the config file and all the folders have the right permissions
18:34 craque joined #salt
18:34 pwistran_ joined #salt
18:35 nkuttler frasergraham: increase loglevel?
18:35 frasergraham ok, i'll try that
18:41 nliadm I'm getting this: msgpack.exceptions.UnpackValueError: Unpack failed: error = 0
18:41 nliadm the only issue I can find is regarding esky builds on windows
18:41 nliadm this is esky on linux
18:42 Gareth morning
18:42 forrest hey Gareth
18:43 forrest frasergraham, can you try the command again with -l debug on the end/
18:43 forrest *?
18:43 forrest that should give some debug output
18:43 Gareth forrest: hey. hows it going?
18:43 forrest oh nkuttler already said so, huzzah for skimming
18:43 forrest Gareth, another day! Getting excited for SCALE?
18:44 craque hi, is there a way for me to force a remote execution to return output from a script run remotely?
18:44 craque i have a script out on some servers and i'd like to see its output as i batch run a cmd.run from the master
18:45 craque but when i do that, i can't see output from the remote script
18:45 frasergraham with debug I get Failed to authenticate, is this user permitted to execute commands?
18:45 Gareth forrest: of course :)
18:45 frasergraham i'm running "salt-run -l debug manage.status"
18:46 clearclaw joined #salt
18:47 jcristau joined #salt
18:47 dave_den craque: if the script prints to stdout or stderr, you will see it in the job return info. are you getting any return data when you publish the command? if not, try increasing your timeout ( -t 300 for 5 minute timeout )
18:48 n8n joined #salt
18:48 forrest So Gareth, does the full access pass not provide Friday access?
18:49 thayne joined #salt
18:49 TheRealBill joined #salt
18:55 mgw joined #salt
19:00 jero joined #salt
19:00 nliadm odd, it seems to be a cache issue
19:01 craque left #salt
19:09 SEJeff_work joined #salt
19:22 ipmb joined #salt
19:25 joehillen joined #salt
19:26 jrdx joined #salt
19:28 Gareth forrest: Full access is access to everything.
19:28 Gareth forrest: All events on Friday are included.
19:28 joehillen joined #salt
19:29 Psyke joined #salt
19:31 BrendanGilmore joined #salt
19:31 d10n joined #salt
19:32 Gareth forrest: going to make it out for the show?
19:32 troyready joined #salt
19:33 jankowiak joined #salt
19:35 HeadAIX joined #salt
19:35 gnugnu1 joined #salt
19:35 diegows joined #salt
19:41 ajw0100 joined #salt
19:44 themadcanudist joined #salt
19:46 Ryan_Lane1 joined #salt
19:48 darless left #salt
19:48 themadcanudist hey guys… what's the difference between modules and states? for instance postgres has a salt.states.postgres_user and there is a salt.modules.postgres.user_* (series of methods)?  Is it that the state leverages all the various postgres module methods to control a user?
19:48 d10n joined #salt
19:51 hunter joined #salt
19:51 zooz joined #salt
19:51 ninkotech__ joined #salt
19:51 hunter Can someone help me with some jinja foo?
19:52 hunter I'm looking at existing salt formulas to learn and keep hitting this pattern:
19:52 hunter https://github.com/saltstack-formulas/ssh-formula/blob/master/ssh/map.jinja
19:52 hunter I know _WHAT_ that does, but I don't quite get whats going on.
19:52 herlo joined #salt
19:52 herlo joined #salt
19:53 hunter Line 1 I understand we're using a grain to figure out which part of the table to us (but not why we use filter_by - why not just say the os grain directly?
19:53 hunter Line 12 is a total mystery as to what is happening.
19:55 ninkotech__ joined #salt
19:56 pwistrand joined #salt
19:56 herlo joined #salt
19:56 herlo joined #salt
19:56 herlo left #salt
19:58 Katafalkas joined #salt
20:02 ninkotech__ joined #salt
20:06 pwistrand joined #salt
20:07 pwistrand joined #salt
20:09 ninkotech__ joined #salt
20:11 sroegner Has anyone ever run an overstate.sls that was in a gitfs_remote?
20:11 sroegner looks like that doesn't work
20:11 DaveQB joined #salt
20:12 n8n joined #salt
20:18 ajw0100 joined #salt
20:20 hunter Any help on jinja - if there's a doc (or better commented formula) I'd love that too
20:21 hunter whiteinge: This one is yours... :) https://github.com/saltstack-formulas/pip-formula/blob/master/pip/map.jinja
20:24 hunter Found partial docs: http://docs.saltstack.com/topics/conventions/formulas.html
20:25 hunter People should be pointed at that (at least as a start) when that question(s) comes up.
20:26 yomilk joined #salt
20:27 ninkotech joined #salt
20:29 ggoZ joined #salt
20:31 forrest Gareth, I was thinking about it, the drive is a bit too much for me, and flights are kinda pricey so I'll probably have to pass this year.
20:31 forrest I only have enough funds left in my personal 'conference budget' for one more, and I'd like to go to pycon
20:31 JasonSwindle Gareth:  I want to go to SCALE…..but I think SaltConf used up my away-from-desk time
20:32 forrest JasonSwindle, just use vacation time
20:32 forrest :P
20:32 ninkotech joined #salt
20:34 JasonSwindle Ha…… that is for Germany
20:38 ninkotech joined #salt
20:40 LLKCKfan joined #salt
20:44 jslatts joined #salt
20:44 Ryan_Lane1 joined #salt
20:45 forrest Ahh nice, beautiful country, I want to go back
20:45 d1gital joined #salt
20:48 blee joined #salt
20:49 krow joined #salt
20:52 ipmb what are the valid hash_types for the file module?
20:53 ipmb nm, https://github.com/saltstack/salt/blob/develop/salt/modules/file.py#L49-L56
20:53 joehoyle joined #salt
20:53 ninkotech joined #salt
20:54 forrest ipmb, I was just gonna link that.
20:54 forrest that should probably be stated in the docs.
20:56 forrest ipmb, https://github.com/saltstack/salt/issues/10363
20:56 forrest feel free to make a pull request :P
20:57 basepi Corey: I didn't recognize you at all from that blurry slide.  Once you said it was you, I could kind of see it.....
21:00 Poooogles joined #salt
21:00 quickdry21 joined #salt
21:02 hunter joined #salt
21:04 hunter_ joined #salt
21:07 ninkotech_ joined #salt
21:09 hunter_ joined #salt
21:14 joehoyle joined #salt
21:14 n8n joined #salt
21:15 cewood joined #salt
21:16 cewood joined #salt
21:16 rojem joined #salt
21:17 ninkotech__ joined #salt
21:17 asnark joined #salt
21:19 krow joined #salt
21:21 TJ1980 joined #salt
21:21 gnugnu1_ joined #salt
21:25 ninkotech joined #salt
21:28 xmltok joined #salt
21:35 nicksloan joined #salt
21:35 nicksloan if you stop using something, do you generally put contrary states in your configuration?
21:35 nicksloan ie, stop supevisor, remove it, etc.
21:36 mpanetta In the cloud I find that it is often more secure to just delete the instance and regenerate it with the formula then to rely on things being removed...
21:37 mpanetta If it is bare metal I guess removing it and all the config bits or users or whatever would be the best way.
21:38 mpanetta And I suppose dependencies as well?
21:38 nicksloan mpanetta: yeah, it just seems like a nighmare.
21:39 mpanetta That is another reason it is easier to just destroy the system and rebuild it ;)
21:40 nicksloan mpanetta: good point.
21:42 mpanetta Basically I look at it this way, it would take hours to write an uninstall formula and use it (trying to find all the little bts that need to be undone, etc) then it takes to rebuild a cloud server and initialize it with the formula you already have (sans the removed bits of course)
21:42 krow joined #salt
21:43 schimmy joined #salt
21:47 ninkotech_ joined #salt
21:48 JasonSwindle mpanetta:  Amen.  Treat a server like live stock and not a pet.
21:48 JasonSwindle That is the way I do it...
21:49 mpanetta JasonSwindle: exactly!
21:50 JasonSwindle and with Vagrant and docker, you can create and destroy really really fast
21:50 Ahlee Such a horrible analogy as it assumes you should treat animals differently :P
21:50 forrest lol
21:50 forrest quote of the year from JasonSwindle 'Treat livestock like they are totally replaceable and worth nothing!'
21:50 kiorky :B 1
21:51 kiorky oups,
21:51 kiorky :) touché
21:51 JasonSwindle Only for their tasty tasty flesh
21:52 thayne joined #salt
21:52 jY is there a way to display on the master what the set file roots would be for a minion?
21:54 gnugnu1 joined #salt
21:54 yomilk joined #salt
21:55 n8n joined #salt
21:56 Ahlee by default, that should show up if you query the pillar.items
21:57 jY ok
21:57 Ahlee although it does not appear to show gitfs remotes
21:57 themadcanudist left #salt
22:01 ajw0100 joined #salt
22:02 jab416171_ joined #salt
22:03 jab416171_ I have a managed file (/etc/foo/bar) but the directory (/etc/foo) doesn't exist. How can I tell salt to make the directory?
22:03 jacksid joined #salt
22:03 jab416171_ nevermind
22:03 vbabiy_ makedirs=True
22:03 vbabiy_ jab416171_beat me to it
22:04 jab416171_ it looks like I didn't even need that
22:05 krow joined #salt
22:05 jab416171_ salt made /etc/foo for me
22:05 baffle What is the correct way to put grains into the mine? I have a pillar-entry that looks like this: http://pastebin.com/zK50tP4c .. network.interfaces and disk.usage is in the db, but not my grains.
22:07 oeuftete do Jinja raw blocks work in sls files?  AFAICT, they don't, but I might be doing something wrong
22:07 baffle Uhh. Actually. I think it might have worked.
22:08 kermit joined #salt
22:13 hunter joined #salt
22:13 xxx123 joined #salt
22:15 Kenzor joined #salt
22:15 yomilk joined #salt
22:17 xxx123 I would have a question regarding pillars, let's have a pillar pillar['file'] storing some file
22:18 xxx123 Now, I would like to run a command defined with cmd.run state but only if that file exists
22:18 xxx123 so I would like to use something like "onlyif: {{ pillar['file'] }}
22:19 xxx123 onlyif: test -e {{ pillar['file'] }
22:19 halfss joined #salt
22:19 xxx123 but it doesn't work, it ends with python error
22:21 ndrei joined #salt
22:23 ninkotech joined #salt
22:25 forrest xxx123, take a look at using jinja if statements: http://jinja.pocoo.org/docs/templates/#if
22:25 baffle xxx123: What do you mean "file"? You mean if it is defined in a minions pillar or not?
22:25 gnugnu1 joined #salt
22:27 felixhummel joined #salt
22:28 xxx123 please have a look on simple dummy state here: http://pastebin.com/C1cmiZPD
22:30 xxx123 it seems that a pillar can't be passed as a value of the onlyif argument
22:31 kalloc joined #salt
22:31 halfss joined #salt
22:31 xxx123 the problem with jinja {% if ... %} statement is that I need to evaluate it at a minion and not on the master
22:32 kalloc joined #salt
22:32 joehoyle joined #salt
22:33 baffle The pillar is only evaluated at the master, that's kindof the point..?
22:34 Gordonz_ joined #salt
22:34 baffle xxx123: But what have you set some_file: to in the pillar?
22:34 ninkotech_ joined #salt
22:35 baffle xxx123: Pillar only declares variables/secrets and export it to the minion..
22:36 xxx123 the pillar['some_file'] contains a file path I want to check if it exists and then, to run the command witn cmd.run state
22:36 gadams999 left #salt
22:36 xxx123 yea, exactly
22:36 _fllr_ joined #salt
22:36 forrest xxx123, https://gist.github.com/gravyboat/8945670
22:36 forrest did you try pillar.get?
22:37 _fllr_ joined #salt
22:37 xxx123 not yet :-)
22:38 xxx123 does it work?
22:38 forrest I don't know, I'm not at a machine I can test it on
22:38 forrest seems like the first example should have worked
22:38 xxx123 same with me, I just have it in my head and I'm thinking about it now :-)
22:38 forrest I added another comment as well
22:38 forrest did you run this state through using -l debug yet?
22:40 xxx123 yep
22:40 ckao joined #salt
22:40 xxx123 it produced the following error:
22:40 xxx123 TypeError: unsupported operand type(s) for -: 'Strict Undefined' and 'Strict Undefined'
22:42 faldridge joined #salt
22:42 forrest does your path perhaps need to be wrapped in quotes? so '{{ pillar['some_file'] }}'
22:43 ninkotech_ joined #salt
22:43 forrest seems like something is being returned weird
22:43 forrest or there is something going on with a variable name or something
22:44 forrest is there no more data in the error? There should be some lines for where it is finding the error, that should point you in the right direction
22:44 xxx123 I will try to investigate it more and will return here once I find som new details
22:45 jrdx joined #salt
22:45 mjulian joined #salt
22:45 mjulian joined #salt
22:50 david_a joined #salt
22:52 fllr joined #salt
22:52 gnugnu1 joined #salt
22:57 jacksid left #salt
22:57 hunter joined #salt
22:58 terminalmage 0.17.5 packages are being pushed to stable, they should make their way out to the mirrors within 2 days or so
22:58 sroegner joined #salt
22:59 krow joined #salt
23:00 hunter_ joined #salt
23:00 halfss joined #salt
23:03 Corey terminalmage: ...for what?
23:04 KyleG joined #salt
23:04 KyleG joined #salt
23:04 Ryan_Lane1 joined #salt
23:04 krow joined #salt
23:07 ninkotech_ joined #salt
23:08 fllr joined #salt
23:12 ninkotech__ joined #salt
23:12 faef joined #salt
23:12 forrest Corey ?
23:12 dave_den joined #salt
23:12 forrest for salt
23:13 mbarrien joined #salt
23:13 faef Is there any way to add a pillar env without restarting salt-master ?
23:13 Corey forrest: No kidding. Which packages specifically? We support more than one OS.
23:13 Corey A number of them don't have "mirrors" for instance. :-)
23:13 forrest epel = centos/rhel, YEESH
23:13 Corey He never mentioned EPEL.
23:13 forrest he's just talking about the epel shennanigans
23:13 forrest Well, what other system takes 2 days to sync to the mirrors? :P
23:14 scalability-junk +1 for the new approach on networking and crypto
23:17 dave_den for anyone in denver, i'm getting a Salt Users Group meetup going. http://www.meetup.com/Denver-SaltStack-Users-Group/
23:17 dave_den join up and help salt denver
23:18 Corey dave_den: Hmm.
23:18 Corey I could fly in for a meeting if that'd be useful at some point.
23:18 fllr joined #salt
23:19 MindDrive joined #salt
23:19 MindDrive left #salt
23:19 dave_den Corey: sweet - will definitely keep that in mind. its super early - the group just opened for members a few hours ago and i need to shedule the first meetup and figure out what talk/lightnings might be good for the members
23:19 MindDrive joined #salt
23:19 Corey dave_den: I gave a few good salt talks lately, and I'm giving a lightning talk at SCaLE entitled "You Are All Idiots." So either way. :-)
23:20 dave_den i'll post to the salt-users mailing list in the next day or two. i'd like to be added to the salt community resources list, too
23:20 dave_den haha
23:20 forrest don't let Corey lie to you, he is terrible at talks
23:20 ninkotech__ joined #salt
23:20 Corey :-( And after I bothered to show up to support forrest as he tripped all over himself...
23:20 forrest lol
23:20 forrest it was the suit
23:20 forrest too dressed up
23:20 forrest throwing everyone off
23:21 forrest 'who is this marketing guy?', 'what is with the suit? Is that some business analyst?'
23:21 forrest no one actually said that though
23:21 forrest I'm just pretending they did
23:21 forrest oh Corey, did you work on the best practices stuff?
23:22 forrest just hoping to avoid merge requests.
23:22 Corey forrest: I passed out blissfully on the plane.
23:22 forrest err merge conflicts
23:22 forrest man
23:22 Corey forrest: Go ahead and update it, I'll re-pull.
23:22 forrest no screaming children?
23:22 forrest Corey, I haven't done much since last time
23:22 Corey forrest: Not in first/business class!
23:22 forrest lol
23:22 Corey forrest: I have a list of topics but haven't expounded upon them. Want a pasteflood in your PM?
23:22 bemehow joined #salt
23:22 forrest no I'm at work
23:23 forrest so I won't have it when I go home
23:23 halfss joined #salt
23:25 Corey Bah.
23:25 Corey forrest: needs moar persistent connections.
23:25 forrest yea I'm not gonna tunnel out to a VM while at work
23:25 gnugnu1_ joined #salt
23:25 forrest securitah
23:25 terminalmage Corey: EPEL
23:25 terminalmage my bad
23:26 Corey terminalmage: Well that's a bit more clear, you redhat centric jerk.
23:26 terminalmage haha
23:26 Corey I DEMAND MORE PLAN9 SUPPOART!
23:26 terminalmage yeah, I'm a total redhat fanboi
23:26 terminalmage as evidenced by my Arch Linux stickers :P
23:26 kalloc joined #salt
23:27 halfss joined #salt
23:27 forrest don't let him lie to you, he's running gnome desktop
23:28 forrest maybe some fedora 20
23:29 terminalmage lol
23:29 terminalmage SHHHHH
23:29 terminalmage gotta keep up my geek cred
23:33 halfss joined #salt
23:36 jankowiak joined #salt
23:36 pirateduck81 joined #salt
23:39 d1gital if a state "is unavailable", does that mean the execution failed, or that salt is unaware of the state?  (I'm getting this message trying to use composer.installed)
23:40 bemehow_ joined #salt
23:45 KyleG1 joined #salt
23:47 gw joined #salt
23:47 gw joined #salt
23:47 alunduil joined #salt
23:48 faef Is there any way to add a pillar env without restarting salt-master ?
23:48 Ryan_Lane1 faef: nope. every config change requires a restart
23:48 danielbachhuber joined #salt
23:48 gw joined #salt
23:48 gw joined #salt
23:49 faef Is there any plan to allow for adding pillars env without restarting salt, like you can file_roots via gitfs without restarting?
23:49 Ryan_Lane1 well, the env is a config variable, but the top file and the pillars themselves are not
23:50 Ryan_Lane1 you can change those as much as wanted without a restart
23:50 Corey faef: What's the problem with restarting the master?
23:51 faef Corey: it's not restarting the master that's annoying, it's that someone has to go in and modify the file for every new pillar env
23:51 faef Corey: we create a bunch of throwaway envs  for various reasons, and it would be nice to have it decentralized (like creating a branch on a repo)
23:53 Ryan_Lane1 faef: you should consider an external top
23:53 Ryan_Lane1 those work with pillars, right?
23:55 faef Ryan_Lane1: not sure how that would let people instantiate a new environment with separate pillar_roots and file_roots without restarting salt master or change master  config
23:56 Ryan_Lane1 I thought external top files allow you to use something other than the top.sls to specify that information?
23:56 Ryan_Lane1 so that you could store that info in a database or such
23:57 * Ryan_Lane1 looks
23:58 schimmy There is an issue that has been bugging me: https://groups.google.com/forum/#!msg/salt-users/QoL8W1qHGbQ/-YZdDWuk3nEJ, however it has been fixed: https://github.com/saltstack/salt/commit/7ad966b1fa767ba6fffdb3897c80a4b23ce905fc
23:58 Ryan_Lane1 faef: http://docs.saltstack.com/ref/configuration/master.html#master-tops ?
23:58 faef why not just have pillar roots support gitfs like file_roots does?
23:58 schimmy yet there is no github issue for it
23:58 Ryan_Lane1 well, yeah, that would be nice too. heh
23:58 schimmy I will create one, if I'm not missing anything here
23:58 schimmy and thanks to @jasonrm
23:59 schimmy I've tested briefly with _modules and it looks good
23:59 faef Ryan_Lane1: that seems like overkill
23:59 faef guess I'll stick to preallocating a bunch of envs for each dev
23:59 elfixit joined #salt

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