Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-09-17

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

All times shown according to UTC.

Time Nick Message
00:00 nate_c joined #salt
00:02 estahn joined #salt
00:02 Thiggy joined #salt
00:03 Thiggy Are there any examples of using a state arbitrary module redirect? The docs are a bit vague: https://docs.saltstack.com/en/latest/ref/states/providers.html#arbitrary-module-redirects
00:09 aqua^c joined #salt
00:15 alfborge joined #salt
00:16 aristedes joined #salt
00:24 jodv joined #salt
00:25 jodv_ joined #salt
00:27 jimklo joined #salt
00:27 murrdoc joined #salt
00:29 dthom91 joined #salt
00:30 seatan joined #salt
00:33 pravka joined #salt
00:35 pravka joined #salt
00:42 maduro joined #salt
00:42 dthom91 joined #salt
00:44 arc-impulse joined #salt
00:45 dexter91x joined #salt
00:46 otter768 joined #salt
00:46 dexter91x mongodb returner is not working on the latest version on saltstack 2015.8.0, however it works if i downgraded salt to 2015.5.5 , is there anyone have the same issue ?
00:48 anotherZero joined #salt
00:51 joe_n joined #salt
00:54 arc-impulse joined #salt
00:58 nofxrok joined #salt
01:04 arc-impulse joined #salt
01:08 pfallenop joined #salt
01:08 pfallenop joined #salt
01:12 fersur3 joined #salt
01:13 SheetiS joined #salt
01:14 arc-impulse joined #salt
01:14 zmalone joined #salt
01:16 alfborge joined #salt
01:17 aristedes joined #salt
01:24 arc-impulse joined #salt
01:25 estahn how do you use os.pathsep in saltstack?
01:25 estahn {% set deploy_dir = os.pathsep.join(pillar.app.deploy.dir, app_name) %}
01:25 estahn throws an jinja error that "os" is not defined
01:26 hasues joined #salt
01:26 breakingmatter joined #salt
01:26 aw110f joined #salt
01:26 hasues left #salt
01:28 cpattonj joined #salt
01:29 cpattonj I can't get salt.states.service.running to work at all despite confirming that "service myservice status" reports running... where should I start debugging?
01:30 __number5__ calebj: how do you use service.running, in your sls?
01:31 cpattonj according to https://docs.saltstack.com/en/latest/ref/states/all/salt.states.service.html with no sig
01:31 seatan joined #salt
01:32 cpattonj https://gist.github.com/buddyp450/381b1e3dce9e6123727a
01:33 arc-impulse joined #salt
01:36 cpattonj __number5__ did that help?
01:37 aristedes left #salt
01:38 timoguin joined #salt
01:39 __number5__ cpattonj: code looks ok, what errors do you get?
01:40 cpattonj No errors, it's just when I salt my highstate it says it's not running and it's set to be restarted with a result of "None"
01:40 Bryson joined #salt
01:40 cpattonj But if I "service elasticsearch status" on the machine it's clearly running
01:41 cpattonj I don't know what else to debug
01:41 catpigger joined #salt
01:41 cpattonj :/
01:41 __number5__ cpattonj: try run `salt-call` with `-l debug`
01:41 __number5__ also what OS are you using? Ubuntu?
01:42 cpattonj yes, Ubuntu
01:42 cpattonj can I salt-call with test=True?
01:45 __number5__ test=True won't give you any real result
01:45 aristedes joined #salt
01:45 aristedes left #salt
01:45 cpattonj What I meant was, I've been testing this from my master with test=True... I'd like to "test" the salt-call on the minion without actually pulling down my highstate
01:45 cpattonj I want to make sure things are fully functional before I do that
01:45 arc-impulse joined #salt
01:45 cpattonj and what do you mean won't give you any real result? like all the results are bogus? some are bogus?
01:45 __number5__ cpattonj: you better to do the test on a test environment, test=True is not very good for debugging
01:46 __number5__ cpattonj: test=True only runs the code won't make any changes, it only prove that you don't have errors in your salt code
01:47 cpattonj that's fine for my use case I think, afaik if it says it's going to restart my service because it's detected as not running that's proof enough that it couldn't find it as running
01:50 cpattonj to make matters even more interesting, running: "salt 'myhost' service.status elasticsearch" returns true!
01:51 cpattonj Maybe a bug in 2015.5.0?
01:53 womble Is there a way to detect whether a machine is running in a container (or, more generally, what sort of virtualisation is involved)?  Something akin to facter's 'virtual' fact.  Google is being surprisingly unhelpful.
01:53 arc-impulse joined #salt
02:00 dendazen joined #salt
02:01 dexter91x mongodb returner is not working on the latest version on saltstack 2015.8.0, however it works if i downgraded salt to 2015.5.5 , is there anyone have the same issue ?
02:03 arc-impulse joined #salt
02:06 Nazca__ joined #salt
02:09 __number5__ womble: check the `virtual` grains, but I'm not sure if it will come up with different values in GCE and AWS
02:10 womble __number5__: I only really care about LXC / !LXC at the present time, so if it'll do that, I'll be happy.  Thanks for the pointer.
02:13 arc-impulse joined #salt
02:13 womble Awwww... it detects "VirtualBox" for all the containers, rather than "LXC"
02:17 alfborge joined #salt
02:19 dexter91x anyone uses mongodb returner ?
02:19 zmalone1 joined #salt
02:20 __number5__ womble: that's weird, at least I get 'xen' for aws ec2
02:22 fshalor joined #salt
02:23 arc-impulse joined #salt
02:23 womble __number5__: I assume it's mostly accurate, it just doesn't seem to be real happy with LXC containers (at least the way that CentOS 6 is creating them)
02:23 cpattonj I want to be done with this so bad lol
02:23 cpattonj Considering upgrading to 2015.8.0 but can't find the backwards compatibility notes
02:26 womble Ah, I see, it's doing LXC detection a bit derpily, only looking for a cgroup path of `/lxc/`
02:28 breakingmatter joined #salt
02:29 DammitJim joined #salt
02:31 DammitJim how do I call a state multiple times?
02:31 DammitJim like I want to have a state that creates instances of services on a server
02:32 seatan joined #salt
02:33 arc-impulse joined #salt
02:36 womble DammitJim: Jinja looping.
02:36 DammitJim womble, yeah, that's what I am doing
02:36 sasa joined #salt
02:36 DammitJim I guess I just have to define the variables in pillar
02:37 __number5__ womble: better to write a custom grain or something
02:39 womble __number5__: Yeah, that's what I've ended up doing.  Does the job.
02:39 womble DammitJim: Yep.  I've got, for example, a list of containers I want to build on a machine in that machine's pillar, and a loop that creates the necessary states to do that for each one.
02:42 arc-impulse joined #salt
02:43 overyander joined #salt
02:43 DammitJim i still have to learn about custom grain vs pilars
02:44 dthom91 joined #salt
02:47 zmalone joined #salt
02:47 otter768 joined #salt
02:48 jimklo joined #salt
02:52 DammitJim do I need to tell the state that I am running a for loop?
02:52 arc-impulse joined #salt
02:55 zmalone joined #salt
02:56 joukard joined #salt
03:02 pcn DammitJim: You use jinja (mostly always jinja) to render out a template with all of the states
03:02 arc-impulse joined #salt
03:02 DammitJim I figured it out... spacing
03:02 DammitJim *sigh*
03:02 dexter91x joined #salt
03:02 DammitJim no trying to figure out why cmd.run is not doing whatit's supposed to
03:03 pcn Render the state, see if it looks right
03:03 pcn BTW, protip: don't use "[" use "test" if that's part of what you're doing
03:04 nofxrok joined #salt
03:06 jimklo joined #salt
03:10 jimklo joined #salt
03:12 arc-impulse joined #salt
03:12 sunkist joined #salt
03:18 alfborge joined #salt
03:18 kusams joined #salt
03:18 lampshades joined #salt
03:21 Sketch joined #salt
03:21 programmerq joined #salt
03:24 auzty joined #salt
03:27 favadi joined #salt
03:29 forrest joined #salt
03:29 evle joined #salt
03:30 arc-impulse joined #salt
03:32 writtenoff joined #salt
03:39 arc-impulse joined #salt
03:40 cyborg-one joined #salt
03:40 estahn joined #salt
03:48 larsfronius joined #salt
03:49 arc-impulse joined #salt
03:54 edrocks joined #salt
03:55 edrocks is there a way to get file.directory to work like file.managed? I need to sync all files in a directory
03:55 pcn Sort of?
03:56 pcn file.recurse
03:56 toastedpenguin joined #salt
03:56 edrocks pcn: thank you that works!
03:56 dexter91x anyone uses mongodb returner ?
03:57 pcn edrocks: glad to help
03:59 arc-impulse joined #salt
04:09 arc-impulse joined #salt
04:12 pratikmallya joined #salt
04:15 ITChap joined #salt
04:17 ageorgop joined #salt
04:18 alfborge joined #salt
04:20 ajw0100 joined #salt
04:23 arc-impulse joined #salt
04:27 jalbretsen joined #salt
04:31 Rebus joined #salt
04:33 arc-impulse joined #salt
04:35 seatan joined #salt
04:43 arc-impulse joined #salt
04:46 Furao joined #salt
04:46 katyucha joined #salt
04:48 arc-impulse joined #salt
04:58 arc-impulse joined #salt
04:59 orion203 joined #salt
05:00 favadi joined #salt
05:01 zer0def joined #salt
05:08 arc-impulse joined #salt
05:14 arc-impulse joined #salt
05:21 alfborge joined #salt
05:23 tzero joined #salt
05:23 arc-impulse joined #salt
05:31 Furao joined #salt
05:33 arc-impulse joined #salt
05:36 seatan joined #salt
05:43 arc-impulse joined #salt
05:48 sirex joined #salt
05:53 arc-impulse joined #salt
05:54 jodv joined #salt
05:55 jodv joined #salt
05:56 che-arne joined #salt
05:57 Bryson joined #salt
06:03 arc-impulse joined #salt
06:06 colttt joined #salt
06:13 arc-impulse joined #salt
06:22 alfborge joined #salt
06:22 arc-impulse joined #salt
06:23 dopesong joined #salt
06:23 aqua^c joined #salt
06:29 jhauser joined #salt
06:30 breakingmatter joined #salt
06:31 jhauser_ joined #salt
06:32 arc-impulse joined #salt
06:33 laax joined #salt
06:36 KermitTheFragger joined #salt
06:37 seatan joined #salt
06:42 arc-impulse joined #salt
06:48 GreatSnoopy joined #salt
06:49 otter768 joined #salt
06:52 arc-impulse joined #salt
07:00 cberndt joined #salt
07:01 seatan joined #salt
07:01 dopesong_ joined #salt
07:02 arc-impulse joined #salt
07:02 joe_n joined #salt
07:03 eseyman joined #salt
07:05 Ztyx joined #salt
07:06 zer0def joined #salt
07:08 lampshad_ joined #salt
07:11 arc-impulse joined #salt
07:12 Ztyx joined #salt
07:24 alfborge joined #salt
07:29 svinota joined #salt
07:32 Ztyx joined #salt
07:36 Ztyx joined #salt
07:43 joe_n joined #salt
07:48 aqua^c joined #salt
07:50 larsfronius joined #salt
07:52 Grokzen joined #salt
07:58 Ztyx joined #salt
07:58 thalleralexander joined #salt
08:02 favadi joined #salt
08:03 Rumbles joined #salt
08:03 kawa2014 joined #salt
08:07 jvv joined #salt
08:07 kbaikov joined #salt
08:08 s_kunk joined #salt
08:15 hemebond joined #salt
08:16 Xevian joined #salt
08:17 markm joined #salt
08:18 impi joined #salt
08:25 alfborge joined #salt
08:27 traph joined #salt
08:31 breakingmatter joined #salt
08:43 keimlink joined #salt
08:47 thefish joined #salt
08:53 lionel joined #salt
08:53 edulix joined #salt
09:05 Furao joined #salt
09:09 linjan_ joined #salt
09:10 estahn joined #salt
09:13 Furao joined #salt
09:13 vim-zz joined #salt
09:19 Ztyx joined #salt
09:22 losh joined #salt
09:22 Ztyx left #salt
09:24 Ludo- joined #salt
09:24 garthk joined #salt
09:27 slav0nic joined #salt
09:28 fredvd joined #salt
09:36 impi joined #salt
09:37 Rumbles joined #salt
09:37 trph joined #salt
09:47 spoksss joined #salt
09:48 spoksss Hi, could anyone check why documentation for the latest stable version is not present
09:48 spoksss https://docs.saltstack.com/en/pdf/Salt-2015.8.0.pdf
09:48 spoksss ?
09:48 spoksss There is a link from mainpage to that document, but is broken
10:06 joukard joined #salt
10:07 shiriru joined #salt
10:09 codehotter Can I put states/data directly into top.sls for states/pillar?
10:09 codehotter Or is top.sls special, referring only to other files?
10:13 estahn joined #salt
10:16 babilen codehotter: The latter
10:29 Ztyx joined #salt
10:32 breakingmatter joined #salt
10:34 favadi joined #salt
10:42 Furao joined #salt
10:49 giantlock joined #salt
10:50 ITChap joined #salt
10:53 estahn Is there some good docu somewhere how to setup the directory structure for environments?
10:54 SheetiS joined #salt
10:54 estahn I currently get "Unable to render top file: Unable to render top file. No targets found.".
10:55 estahn This is what i have in my config https://gist.github.com/estahn/88aec00ceca290eab575
10:59 elsmo joined #salt
10:59 jbrnds Good afternoon!
11:01 Ztyx joined #salt
11:01 jbrnds does {{ a_jinja_dict| yaml }} render to the yaml version of that dict?
11:01 quasiben joined #salt
11:02 estahn Are pillars from different environment get merged?
11:02 Phil` joined #salt
11:02 jbrnds .. i was wondering how to test this.
11:02 Phil` can a pillar read data from another pillar?
11:02 timbyr_ joined #salt
11:03 Ztyx left #salt
11:03 jbrnds Phil` as in actively read? I use jinja variable's for that. So I define {% set foo=bar %} and render this as {{ foo }} within the pillar
11:05 Phil` jbrnds, I have defined a list in one pillar that I wish to loop through in another pillar and create a larger structure from it
11:05 Phil` let me paste the example...
11:05 jbrnds Phil` otherwise. look at the mysql formula on github. There the map.jinja method is used to transport vars within a pillaar structure
11:07 jbrnds Phil` yeah - paste is pls somewher
11:07 Phil` thanks, I'll have a look at that
11:07 Phil` here's what I'm doing at the moment, which doesn't work: https://gist.github.com/anonymous/26065b22760ae377dbb3
11:07 Phil` these are separate pillars
11:09 jbrnds Phil` check. You need to actively load the other pillar from the nginx pillar. See https://github.com/saltstack-formulas/mysql-formula/blob/master/mysql/defaults.yaml for the first settings pillar and https://github.com/saltstack-formulas/mysql-formula/blob/master/mysql/server.sls how to load that pillar inside the other pillar and do more magic on it (and finally using it to parametrise states)
11:11 jbrnds in the 'first' settings pillar you need to expose the yaml usin {% load_yaml as varname %} in order to load it into the second pillar/state file.
11:11 jbrnds ... then you get a list structure you can loop through in your for loop
11:12 Phil` thanks, trying that now
11:13 estahn Do i have to duplicate the entire pillar file for a different environment?
11:14 jbrnds Phil` something like https://gist.github.com/anonymous/26065b22760ae377dbb3 (you mileage will vary as this is quick and dirty)
11:19 Phil` jbrnds, presumably the content that it imports is not related to the current state
11:19 Phil` in such that package/defaults.yaml will not be different between hosts
11:20 Phil` the idea here is to be able to define a few "package/defaults.yaml" files for the various configs and have the single nginx/nginx.sls file pull its data from the right one
11:21 jbrnds Phil` -- in that case you might want to look again at the mysql formula how they -based on the grain 'os' (or os_family, or even hostname)- alter the parameters for that state.
11:22 Phil` thanks, looking
11:22 jbrnds That is a very common method to work with a single defaults for various platforms and acutally filter_by a certain grain
11:23 Phil` ah ok, I see how they're doing that
11:23 jbrnds Phil` {%- set mysql = salt['grains.filter_by'](rawmap, grain='os', merge=salt['pillar.get']('mysql:lookup')) %} -- in this case the defaults come from (loaded from another file) and is filtered by the os. It even merges pillar settings.
11:23 jbrnds https://github.com/saltstack-formulas/mysql-formula/blob/master/mysql/server.sls
11:23 Xevian joined #salt
11:25 lietu joined #salt
11:27 lietu I'm getting an odd issue when running "salt 'server.name' state.highstate": http://pastebin.com/qcaEYXzR .. this used to work fine, and after looking at the python code I can't see why this error occurs .. apparently traverse_dict_and_list did not expect the argument "delimiter", but when I look at the code in the file it sure has that argument defined?
11:27 lietu updated from 2015.5.0 to 2015.5.5 via yum and that changed nothing
11:27 lothiraldan joined #salt
11:30 jbrnds lietu where are you looking at the python code?
11:30 lietu the files shown on the backtrace, on that server .. vim /usr/lib/python2.7/site-packages/salt/modules/config.py to see what it imported and called, which was /usr/lib/python2.7/site-packages/salt/utils/__init__.py:traverse_dict_and_list
11:31 lietu tried to delete the .pyc and .pyo files for that __init__.py just to be sure and that didn't change anything
11:31 jbrnds https://github.com/saltstack/salt/blob/634264f443be075e0ad007fe8e40f3bfa8432c1b/salt/utils/__init__.py#L1463
11:32 monkey661 left #salt
11:32 jbrnds hmmm strange indeed
11:32 lietu yup
11:33 lietu I'm wondering if I should do a find -name "*.pyc" -or -name "*.pyo" -delete in the salt directory
11:33 lietu but seems a bit pointless
11:33 Phil` jbrnds, is there any way to dump out a list for debugging purposes?
11:33 Phil` can't really tell if my grains.filter_by is working properly
11:34 Grokzen joined #salt
11:35 Ztyx joined #salt
11:35 jbrnds Phil` ha! Had the same question a while back. How I solve this is to make a state file.managed: - contents: {{ dict|yaml_dquote }}
11:35 jbrnds it renders the dict/list/structure into a file
11:35 Phil` ha - that works I guess
11:36 Ztyx left #salt
11:37 zer0def joined #salt
11:38 Phil` hm, it's not loading the yaml file correctly - do I have to put it in the top.sls or something?
11:38 jbrnds Phil` if you know how to render it as a prettyprinted yaml file that would be awesome.
11:40 lietu and yea, deleting all the .pyc and .pyo files solved nothing
11:40 lothiraldan joined #salt
11:40 jbrnds Phil` use the fill pathname the same as you would do with salt://<fullpath>. Remember that it difficult for a state to 'load' something directly from a pillar.
11:41 lietu there's something really odd going on, if I edit that file and add a "import pdb; pdb.set_trace()" above the call, not even the line number changes from 321 to 322
11:41 derxleben joined #salt
11:44 lietu hm, right .. it's running as a daemon and apparently was not restarted, if I restart it I guess it gets that update but then the breakpoint gets triggered on the daemon somewhere
11:45 jbrnds Phil` ok it seems that there are two things happening here. If you want a state to filter_by a grain. That varialbe has to come from somewhere. If that variable is provided to the minion as a pillar, that perform the filtering on that pillar. If you want to provide a 'clean' pillar to the state, you have to perform the filtering on the pillar side.
11:45 jbrnds (did not make much sense did it) :/
11:46 dthom91 joined #salt
11:47 lietu ok so I managed to figure out that it seems like that error is coming from the target server
11:48 jbrnds lietu you mean the minion. ssh into the minion and perform a salt-call state.highstate -ldebug?
11:48 lietu yea, minion
11:48 jbrnds Phil` i am going to update the gist for you to show what i mean
11:48 estahn does anyone  have an idea why this is not working?
11:48 estahn base:
11:48 estahn '*': &BASE_ALL
11:48 estahn - stack
11:48 estahn development:
11:48 estahn '*':
11:49 estahn <<: *BASE_ALL
11:49 estahn - stackmack
11:49 amcorreia joined #salt
11:50 breakingmatter joined #salt
11:50 lietu jbrnds, ran just fine
11:50 lietu running it from master still gets the same error
11:51 lietu minion seems to be running 2015.5.3
11:51 lietu and even on the minion that salt.utils contains a traverse_dict_and_list with the deilimiter argument
11:52 lietu -typos
11:52 lietu restarting the minion fixed this
11:52 lietu now the question is, why? why do I need to manually restart the minion?
11:53 lietu that is: systemctl restart salt-minion
11:53 jbrnds Phil` something like: https://gist.github.com/anonymous/26065b22760ae377dbb3
11:54 impi joined #salt
11:54 jbrnds lietu possibly to load and bytecompile the updated pythonfiles because of the update?
11:55 lietu yea but if the updated version is there on the server, I assume it's been updated by the salt bootstrap script or yum or something and it should probably know to restart the service
11:56 lietu I just ran a salt '*' cmd.run 'yum -y update salt-minion' and same for 'systemctl restart salt-minion' which should hopefully fix this for now, but it still feels wrong
12:00 jbrnds Phil` alternatively you could do the filtering on the pillar side in defaults.sls with the {% from defaults.yaml %} etc etc and than place that sls file in the top.sls. You have to put the actual defaults in another file -- or not at all. You define it right in that same default.sls file. Will update the gist
12:02 freelock joined #salt
12:04 Rumbles joined #salt
12:06 TooLmaN joined #salt
12:08 Nazzy joined #salt
12:12 zer0def joined #salt
12:13 tmclaugh[work] joined #salt
12:14 rawsik joined #salt
12:15 rawsik hi.    Is it available to change pillar interactive  from command line ?
12:17 rubendv joined #salt
12:18 twork joined #salt
12:20 dexter91x joined #salt
12:26 scoates joined #salt
12:30 pdayton joined #salt
12:31 rawsik I try '{ pillar='{"app_name": "new_archive_name.zip"}' } ,   but  it doesn't  work with orchestrate
12:32 theologian joined #salt
12:42 dendazen joined #salt
12:42 teebes joined #salt
12:43 giantlock joined #salt
12:43 sjorge joined #salt
12:43 sjorge joined #salt
12:44 pdayton1 joined #salt
12:45 Furao joined #salt
12:45 numkem joined #salt
12:47 numkem joined #salt
12:49 zwi joined #salt
12:50 geekatcmu joined #salt
12:51 subsignal joined #salt
12:52 shiriru joined #salt
12:52 dendazen joined #salt
12:53 racooper joined #salt
12:54 pcn salt 'foo' state.sls barstate pillar='{"key" : "val"}'
12:57 rubendv joined #salt
13:00 dthom91 joined #salt
13:04 jbrnds I want to concateanate 2 variables into a single 'string' within a file.managed: - content: <string> state declaration. However I everytime I run into problems saying: mapping values are not allowed here. This because the cotents of the variable is yaml?
13:04 cpowell joined #salt
13:05 losh joined #salt
13:09 pcn jbrnds: you should post an example
13:09 tkharju joined #salt
13:10 JDiPierro joined #salt
13:12 ferbla joined #salt
13:13 arc-impulse joined #salt
13:14 cpowell joined #salt
13:17 pdayton joined #salt
13:19 debian112 joined #salt
13:19 cpattonj joined #salt
13:20 jbrnds pcn https://dpaste.de/e9xb
13:20 pdayton joined #salt
13:21 cpattonj So I thought i was crazy but now I know I'm not, this is a bug right? https://gist.github.com/buddyp450/c935f536260d11a26243
13:21 jbrnds pcn - the list is properly 'filled', but the renderer seems to reinterprete it again. How can I ensure this becomes a string it should not interprete?
13:21 DammitJim joined #salt
13:24 mapu joined #salt
13:26 jbrnds made it work with extra "" around it and using filter |yaml
13:28 rubendv joined #salt
13:28 jbrnds now I use slack as remote debugger :) pushing contents of variables to my saltdebug pvt group using the state slack.post_message. The message is now formulated as: {{ '"%s"' % a_list_of_dict|yaml }} - works great
13:31 Nazzy joined #salt
13:32 pcn Cool
13:33 bhosmer joined #salt
13:34 pravka joined #salt
13:35 markizano joined #salt
13:35 markizano I have a job running in the scheduler that publishes data to graphite.
13:35 markizano When I run that job from the command line, it publishes to graphite with no issues.
13:36 markizano but when it runs in the scheduler, it doesn't send any data to graphite.
13:36 markizano I can see it connecting in graphite.
13:36 codehotter When using the gpg renderer for pillar data, do I need to create the key on the master or the minion?
13:36 codehotter Is decryption happening on the master or the minion? Who needs access to the keys?
13:36 markizano codehotter: the pillar is generated on the master - if that helps.
13:36 kusams joined #salt
13:37 codehotter markizano: well in https://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.gpg.html, it says to create the keys on the master, but then it also says decryption will happen on the minions
13:37 codehotter I'm confused
13:37 codehotter Don't you need the keys to be able to decrypt?
13:38 markizano hmm... based on how it's describing it, I would think the master is providing the key to the minion to perform the decryption.
13:40 timoguin joined #salt
13:40 Ztyx1 joined #salt
13:41 protoz joined #salt
13:41 codehotter Can I specify a different default renderer for pillar than I use for states?
13:43 codehotter I guess not: https://github.com/saltstack/salt/issues/13263
13:44 quasiben joined #salt
13:51 favadi joined #salt
13:51 Rumbles joined #salt
13:52 dendazen joined #salt
13:53 bhosmer joined #salt
13:53 cresten joined #salt
13:53 JDiPierro joined #salt
13:55 fractalizator joined #salt
13:58 johnkeates joined #salt
14:00 Ztyx joined #salt
14:00 jbrnds when I have a file.managed and that uses a template: jinja. How can I render a local defined variable (in the context of the state) correctly in the template.
14:00 zwi joined #salt
14:02 _JZ_ joined #salt
14:03 pratikmallya joined #salt
14:03 kevinquinnyo joined #salt
14:03 zmalone joined #salt
14:05 Akhter joined #salt
14:05 andrew_v joined #salt
14:06 Akhter joined #salt
14:07 jbrnds http://dpaste.com/0C8014R - the {{ a_var }} is not rendered to 'bar'.
14:08 anotherZero joined #salt
14:09 ageorgop joined #salt
14:09 _Kev_ hello - I run a number of states using 'salt --subset=1 'web0*' cmd.run 'command to run..'. Is there a way to pass the --subset=1 via the salt-api? I can get the command to run across all the hosts ok but not get --subset=1 passed to only select one of the available minions
14:11 impi joined #salt
14:12 Ztyx joined #salt
14:14 geekatcmu joined #salt
14:19 geekatcmu joined #salt
14:21 msciciel joined #salt
14:21 teryx510 joined #salt
14:22 cresten joined #salt
14:22 dexter91x jbrnds: you will need to import the variable first in order to call it
14:24 mapu_ joined #salt
14:25 dexter91x jbrnds: check this part https://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.jinja.html#jinja-in-files
14:25 Ztyx left #salt
14:26 jbrnds dexter91x found in the meantime a solution by using - context:
14:27 pratikma_ joined #salt
14:27 jbrnds dexter91x http://dpaste.com/1AW5N9C - this works
14:28 Tyrm joined #salt
14:29 SheetiS joined #salt
14:29 johnkeates left #salt
14:29 arc-impulse joined #salt
14:31 kaptk2 joined #salt
14:31 polishdub joined #salt
14:34 druonysus joined #salt
14:34 druonysus joined #salt
14:35 Rumbles joined #salt
14:40 arc-impulse joined #salt
14:44 cwyse joined #salt
14:44 Fiber^ joined #salt
14:45 rubendv joined #salt
14:45 rubendv joined #salt
14:46 kusams joined #salt
14:46 Nagyman_ joined #salt
14:46 rojem joined #salt
14:48 Nagyman_ is it possible to refer to pillar (or other shared variables) from within a pillar?  e.g. to construct a path.  this would be beyond the typical `include`
14:48 clintberry joined #salt
14:50 arc-impulse joined #salt
14:53 theologian joined #salt
14:54 dthom91 joined #salt
14:59 Grokzen joined #salt
15:00 arc-impulse joined #salt
15:01 pdayton I'm attempting to launch a in instance with salt-cloud in a VPC and I receive this error: There was a profile error: string indices must be integers, not str
15:02 pdayton I have a very simple profile config
15:02 pdayton profile-1:
15:02 pdayton provider: providername
15:02 pdayton image: ami-df6a8b9b
15:02 pdayton size: m3.medium
15:02 pdayton ssh_username: ubuntu
15:02 pdayton network_interfaces:
15:02 pdayton - DeviceIndex: 0
15:02 pdayton SubnetId: subnet-xxxxxxxx
15:02 pdayton SecurityGroupId:
15:02 pdayton - sg-xxxxxxxx
15:02 pdayton allocate_new_eips: True
15:02 pdayton What might I be missing?
15:02 lothiraldan joined #salt
15:02 elsmo try - DeviceIndex: '0' ?
15:03 debian112 left #salt
15:03 zer0def joined #salt
15:05 zwi joined #salt
15:06 pdayton That doesn't seem to make a difference
15:07 win_salt joined #salt
15:08 anotherZero joined #salt
15:10 cresten joined #salt
15:10 arc-impulse joined #salt
15:10 lothiraldan joined #salt
15:12 pdayton A similar config works outside of the vpc, where my provider is the same with the only difference being a security group.
15:12 Herle joined #salt
15:14 pdayton this is the 'classic' config that works without issue
15:14 pdayton profile-2:
15:14 pdayton provider: providername2
15:14 pdayton image: ami-df6a8b9b
15:14 pdayton size: m3.medium
15:14 pdayton ssh_username: ubuntu
15:15 jbrnds pdayton can you use gist or pastebin
15:15 JDiPierro I'm trying to use the salt mine to get info about a minion with a compound matcher on grains... it returns nothing.. but when I run a test.ping with the same compound matcher the minion responds. The minion has the function I need declared in it's minion file. Any idea what might be up?
15:16 Rumbles joined #salt
15:16 jbrnds pdayton SecurityGroudId should refer to the security group name that you defined in AWS. Let me check
15:18 jbrnds pdayton i have 'securitygroup: <groupname> in my ec2 providers configuration. not the securitygroupid
15:18 kwakhed joined #salt
15:18 JDiPierro Here's some console output to demonstrate what I'm seeing: http://pastebin.com/KYaQcskH
15:19 kwakhed hello all
15:19 pdayton jbrnds my complete config is here http://pastebin.com/RM9DR9Dh
15:20 Brew joined #salt
15:20 Rumbles joined #salt
15:21 jbrnds pdayton goes too far for my knowledge. only wondering if the securitygroupid needs a 'list'. unsure about that. Thought that in AWS you could only assign 1 security group (and not a list of them).
15:22 jalbretsen joined #salt
15:22 arc-impulse joined #salt
15:24 pdayton jbrnds I'll try it without a list, although the documentation here seems to include a list https://docs.saltstack.com/en/2015.5/topics/cloud/aws.html
15:24 tanta joined #salt
15:25 openfly bit of an oddball question, if i have a yaml file with a dict list in form  "  - value1: value2" per line... is there a way to import that list and address each set of variables as independent variables inside of a for loop?
15:25 cresten joined #salt
15:25 openfly {% for var1,var2 in line %}  did not work.
15:26 kwakhed Could someone help with an issue trying to create a custom runner?
15:26 JDiPierro Any ideas why I could use a compound target directly and it works, but using it through mine.get nothing is found?
15:27 jbrnds pdayton aha indeed. I see. I have defined a securitygroup (by name) in my providers, not the profile
15:29 pdayton jdbrnds I switched it around and I still seem to have the problem
15:29 jbrnds openfly {$ for item in your_list_of_dicts %}  ... do stuff and use {{ item }} {% endfor %}
15:29 kwakhed keep getting this error when trying to excute a custom runner: '[7617] Could not LazyLoad myrunner.test'
15:31 jbrnds openfly We just discussed it this morning. check out this gist for inspiration https://gist.github.com/anonymous/26065b22760ae377dbb3
15:31 oherrala joined #salt
15:31 openfly found it
15:31 openfly {% for var1,var2 in line.iteritems() %}
15:31 openfly did it
15:32 openfly http://blog.mattcrampton.com/post/31254835293/iterating-over-a-dict-in-a-jinja-template  <-- found there
15:32 jbrnds openfly :+1:
15:32 openfly jbrnds cool gist though
15:32 openfly very inspirational
15:33 jbrnds openfly thanks
15:33 arc-impulse joined #salt
15:35 harkx joined #salt
15:35 murrdoc joined #salt
15:38 zer0def joined #salt
15:39 dthom91 joined #salt
15:39 seatan joined #salt
15:39 Herle left #salt
15:42 cyborglone joined #salt
15:43 arc-impulse joined #salt
15:43 kwakhed anybody ever have 'Could not LazyLoad' issue when creating a custom runner?
15:48 anotherZero joined #salt
15:48 anotherZero joined #salt
15:51 iggy kwakhed: can you `python /path/to/your/runner/module.py` ?
15:53 Grokzen joined #salt
15:53 arc-impulse joined #salt
15:54 kwakhed yeah no errors when trying to call my runner
15:55 iggy does it have a __virtual__ function?
15:56 kwakhed yep
15:56 iggy is it failing?
15:56 kwakhed and I just put a print outside of a function and it printed fine
15:56 kwakhed no failes
15:56 rojem joined #salt
15:57 iggy can you gist the output maybe?
15:58 kwakhed this is the custom runner: https://gist.github.com/aryou/b1d8e595734622b562cc
15:59 kwakhed if have this in my master config: 'runners_dirs: ["/srv/salt/_runners/"]'
15:59 kwakhed oops runner_dirs: ["/srv/salt/_runners/"]
16:00 kwakhed derp i answered my own question
16:01 kwakhed had it as runners instead of runner
16:02 murrdoc also u have looked at manage.status
16:02 murrdoc yes /
16:03 dthom91 joined #salt
16:03 iggy off to work (maybe)
16:04 arc-impulse joined #salt
16:07 SeniorRoboto joined #salt
16:11 arc-impulse joined #salt
16:12 mpanetta_ joined #salt
16:12 zwi joined #salt
16:14 stephen144 joined #salt
16:14 spothanis joined #salt
16:15 wryfi left #salt
16:16 impi joined #salt
16:16 che-arne joined #salt
16:20 dthom91 joined #salt
16:22 sunkist joined #salt
16:23 arc-impulse joined #salt
16:23 xDamox joined #salt
16:23 fshalor joined #salt
16:24 cresten joined #salt
16:25 tardigrade joined #salt
16:26 tardigrade hello
16:26 tardigrade I was wondering if I could get a bit of clarification on the options discussed at the bottom of this page: https://docs.saltstack.com/en/latest/ref/states/top.html
16:26 tardigrade specifically, is `env_order` synonymous with "processing priority"?
16:27 zmalone joined #salt
16:27 tardigrade It says that it sets the order of processing, but it doesn't really explain the implications of that. Does the first or the last assignment win?
16:28 aparsons joined #salt
16:30 tanta I never had any luck with using multiple environments
16:30 SeniorRoboto My interpretation is it explicitly states an order you determine in a state file irrespective of the order in the file.
16:30 tanta I end up assigning my own environments via hostname state/pillar targeting in top.ls
16:31 tanta and globbing the entire set of minions into 'base'
16:31 iggy tardigrade: do you really need multiple top files (in each env)?
16:31 cpowell joined #salt
16:32 tardigrade iggy: not really.  I'm just experimenting at this point.
16:32 tardigrade and trying to make sense of the docs
16:33 arc-impulse joined #salt
16:33 tardigrade I'm more generally trying to figure out a strategy for managing testing/staging/production environments
16:33 iggy the docs should say "Avoid this situation if at all possible, but..."
16:33 stanchan joined #salt
16:34 iggy tardigrade: are you going to be using gitfs?
16:34 tardigrade iggy: what would be a sane alternative?
16:34 tardigrade not at the moment
16:34 iggy use 1 top file
16:34 tardigrade but perhaps in the future(?)
16:34 iggy that's sane
16:35 iggy and if you switch to gitfs, a common practice is to have a separate git repo just for the top file
16:35 tardigrade okay
16:35 iggy (so you don't have to worry about keeping them in sync between branches)
16:36 tardigrade so a single top file that breaks things down based on the minion environment?
16:37 dthom91 joined #salt
16:38 alemeno22 joined #salt
16:38 iggy if you really want to use environments, sure
16:38 iggy I don't personally suggest them, but some people have luck with them
16:39 tardigrade do you manage multiple environments with another method?
16:39 Lionel_Debroux joined #salt
16:41 openfly left #salt
16:42 tanta https://gist.github.com/anonymous/9ce4f7ce3a131aa367d7 here's an example from my salt/top.sls tardigrade
16:42 win_salt does anyone here have experience with cmd.run and powershell?  I cant figure out what salt is expecting false to be.
16:42 tardigrade tanta: thanks!  I'll take a look
16:44 bhosmer_ joined #salt
16:44 arc-impulse joined #salt
16:45 tardigrade ah, so just relying on naming conventions.  I was looking into doing that, but was trying to figure out if there were some better alternatives.
16:45 jodv joined #salt
16:45 kawa2014 joined #salt
16:45 tardigrade it definitely seems like the most straight forward route.  I very well might end up doing that.
16:46 prwilson joined #salt
16:47 writtenoff joined #salt
16:54 [dee] joined #salt
16:55 mrwboilers joined #salt
16:57 arc-impulse joined #salt
17:01 mrwboilers Is there a way to set cpu affinity for the minion? Can I pin the minion to a certain core?
17:02 dthom91 joined #salt
17:03 bhosmer_ joined #salt
17:04 fuzzy_id joined #salt
17:05 fuzzy_id can someone point me to an example where a service is stopped, then files moved around and afterwards the service is started again?
17:05 fuzzy_id i need this to change the root directory of mysql…
17:06 geekatcmu mrwboilers: yes, using the same affinity tools you use for everything else.
17:06 geekatcmu but there's nothing built in to salt to make it a config options
17:06 forrest joined #salt
17:07 forrest_ joined #salt
17:08 arc-impulse joined #salt
17:08 mrwboilers geekatcmu: Thanks, just thought I'd ask. I know of some applications where it's configurable -- "cpu=0" or whatever in the config file. Just wondering if Salt had anything like that. Guess not.
17:09 aw110f joined #salt
17:09 geekatcmu nope
17:09 geekatcmu sorry
17:13 murrdoc joined #salt
17:13 zmalone joined #salt
17:13 arc-impulse joined #salt
17:16 cresten joined #salt
17:17 pratikmallya joined #salt
17:19 pratikma_ joined #salt
17:20 ahammond how can I search job history for jobs that ran on a specific server?
17:21 Akhter joined #salt
17:21 ageorgop joined #salt
17:21 jodv ahammond: simplest way might be to just grep the minion log on the server
17:21 dendazen joined #salt
17:22 forrest joined #salt
17:22 viq joined #salt
17:23 Akhter joined #salt
17:25 arc-impulse joined #salt
17:25 timoguin joined #salt
17:28 bhosmer_ joined #salt
17:36 arc-impulse joined #salt
17:38 ThereIsNoYeti joined #salt
17:39 larsfronius joined #salt
17:40 dendazen_ joined #salt
17:40 amcorreia joined #salt
17:41 twork general question, sort of... vsftpd can create files named for a user, with setting tweaks for that user. given that there's some duplication of info, does it make more sense to define that stuff under the 'users' pillar, or the one for the app?
17:41 ThereIsNoYeti left #salt
17:43 Tyrm_ joined #salt
17:43 twork ...i should have said "can have account config specifications" rather thann "can create files for", but you get the gist
17:45 rodr1c joined #salt
17:46 ronT1V joined #salt
17:46 forrest twork: It's your pillar man, do whatever makes sense to you!
17:46 twork i was afriad you'd say that...
17:46 forrest Just keep in mind if it modifies the user, you'll modify the user via the user state, then that state will run and modify again
17:46 forrest so you'll always see changes on your runs
17:46 twork i'm trying to guess what future maintainability will look like
17:47 ronT1V is it possible to configure a master to function on a single port using zeromq?
17:47 arc-impulse joined #salt
17:47 tanta joined #salt
17:47 forrest ronT1V: Instead of two ports? Not as far as I know.
17:47 ronT1V many of our nodes only have port 443 open to internet and would like to avoid having to request each to have another port open
17:48 forrest uhhh, 443 is for SSL connections though
17:48 forrest So your web server/app is already listening on 443 right?
17:48 ronT1V this is on a server without 443 needed for web
17:48 Jahkeup joined #salt
17:49 ronT1V trying to find optimal solution for working across restricted networks
17:49 forrest wait wait wait
17:49 forrest You know you're only opening ports on the master right?
17:49 forrest or is outbound dictated by a firewall as well
17:49 ronT1V so I can set req_port to 443 and the minion will send auth request, it just is unable to pickup new jobs
17:50 ronT1V some of the clients are installed behind firewalls that restrict outbound from minion -> master
17:50 forrest Did you already read this? https://docs.saltstack.com/en/latest/topics/tutorials/firewall.html
17:51 ronT1V this is not a master configuration issue
17:51 ronT1V I can use any port on master and setup firewall
17:51 ronT1V trying to navigate restrictive firewalls
17:51 forrest Gotcha, so you want to specify the outbound minion connection
17:52 iggy you need 2 ports
17:52 ronT1V I have specified the req_port to 443 but the publish_port is a different port
17:52 forrest Yeah that's what I stated earlier Iggy :\ Have you ever tried it?
17:52 ronT1V so if you guys tell me it is definitive I need 2 I will leave it at that
17:52 iggy definitive
17:52 forrest I have never tested 1
17:52 whytewolf you need 2
17:52 forrest as far as I know there is no way it will work
17:53 iggy _however_
17:53 whytewolf proxy?
17:53 iggy 2015.8 has a new transport... not sure if it needs more than one port
17:53 oherrala How do I do state for package cron, which in reality doesn't do anything? So that I can require pkg cron in OpenBSD.
17:53 forrest for raet? I don't know.
17:53 iggy (or raet possibly)
17:53 forrest oherrala: Just treat it like a regular install state, it will ensure the package is installed.
17:54 oherrala forrest: OpenBSD doesn't have any package named cron.
17:55 oherrala so I need to have some kind of noop
17:55 forrest oherrala: Let's talk about what you are trying to achieve, because my only response to that is 'well use whatever package is the equivalent in bsd'
17:55 ronT1V whytewolf, any advice on specific proxies to try?
17:55 forrest Does cron not exist on openbsd?
17:55 whytewolf ronT1V: https://docs.saltstack.com/en/latest/topics/topology/proxyminion/index.html
17:55 forrest That seems insane to me if it doesn't
17:55 forrest whytewolf: Ugh, proxy minion for a real minion? Interestnig
17:56 oherrala forrest: cron is in base, but there's no package for it
17:56 iggy don't think a proxy minion will work... it still needs 2 ports
17:56 forrest oherrala: So why require it as a package then if it's always there?
17:56 iggy for zmq anyway
17:56 whytewolf humm, true.
17:56 kevin-wk joined #salt
17:57 whytewolf the other thing would be a syndic master .... but that is even worse
17:57 oherrala forrest; so that I can require pkg cron in places which require cron
17:57 arc-impulse joined #salt
17:57 oherrala forrest; and isolate the oddity of "missing package" in the cron state.
17:57 baweaver joined #salt
17:58 forrest oherrala: But why do you need to require the package if there is a 100% chance that cron is already on the system? Is missing package an error that you encounter in the cron state with openbsd?
17:58 oherrala forrest; to support multiple operating systems
17:58 petersson joined #salt
17:58 baweaver joined #salt
17:58 oherrala forrest; We run CentOS and OpenBSD, and probably some Ubuntu. And I'm trying to target them all.
17:59 dthom91 joined #salt
17:59 forrest oherrala: Ahh, in that case I'd either set a jinja flag based on os, or create a fake openbsd cron package and host it yourself.
17:59 whytewolf oherrala: use require_in on the package to the file. then suround the package in jina that doesn't install it on openbsd
17:59 JDiPierro With Salt 2014.7.6 I can do salt-run pillar.show_pillar <minion_name> and get the pillars... but calling pillar.items or pillar.raw on the minion returns nothing and the minion keeps saying "[ERROR   ] Got a bad pillar from master, type str, expecting dict: ".
17:59 forrest whytewolf: Only issue with that is if there are a lot of cron jobs, it could get lengthy, but that's a good idea otherwise
18:00 oherrala whytewolf: I would like to handle cron oddities in cron state and not in every other state requiring cron
18:00 oherrala so that every other state can just safely require pkg cron
18:00 TyrfingMjolnir joined #salt
18:00 whytewolf so you want cake and eat it to?
18:00 forrest Best solution, stop using openbsd /troll
18:01 forrest Only solution I can think of in that state is to make a fake openbsd package
18:01 oherrala forrest: heh :)
18:01 oherrala hmmh
18:02 whytewolf humm.  openbsd package that installs a txt file that calls openbsd stupid for installing cron in base
18:02 forrest whytewolf: Yep
18:03 whytewolf I would love to know the thought process behind that choice. oh security issue in cron lets update base
18:04 oherrala well, lets not go there :)
18:06 forrest whytewolf: I have no comment, never worked somewhere that used openbsd, and would never use it myself
18:06 forrest But at some point the choice was made
18:07 oherrala In case of cron, it's software which is pretty much required (for log rotation, etc.) to run the OS.
18:07 forrest They could have just made it a package though
18:08 forrest cent also needs cron for log rotation
18:08 oherrala OpenBSD (the base installation, not ports) doesn't have any packages.
18:08 arc-impulse joined #salt
18:08 oherrala and yes, you update the whole base or compile and install parts of it.
18:09 forrest That sounds very unfriendly to users
18:09 whytewolf well from what i understand. user friendly is not a goal in openbsd
18:09 teryx510 joined #salt
18:10 oherrala ah, I can do some Jinja and then pkg.installed: unless: ls /usr/bin/crontab for OpenBSD
18:10 iggy use jinja to turn the package state into a test.success (or whatever it is)
18:10 dthom91 joined #salt
18:10 iggy on obsd
18:10 oherrala iggy: how?
18:10 katyucha joined #salt
18:11 forrest lol, that's actually a good idea iggy, make the ID the same, but change the contents of the ID.
18:11 oherrala forrest, whytewolf; Well, part of investigating Salt is to have easily reinstallable systems :D
18:12 oherrala So that we can do OS reinstalls instead of OS updates
18:12 iggy crond:\n  {% if grains.os == 'OpenBSD' %}test.succeed_without_changes{% else %}pkg.installed{% endif %}
18:12 oherrala thanks, I'll try that
18:13 baweaver joined #salt
18:13 iggy or something like that... completely untested
18:14 forrest oherrala: https://gist.github.com/gravyboat/75845a6c6c7f6a59603b
18:15 forrest Bah iggy beat me to it
18:15 forrest damn you iggy
18:15 bhosmer_ joined #salt
18:16 oherrala :)
18:17 forrest oherrala: The grains key might be wrong on that, I can't remember if you need os or os_finger or whatever.
18:17 oherrala forrest: how does that satisfy the require: pkg: cron?
18:18 forrest oherrala: You might have to modify the requires based on the type :\
18:18 oherrala or can I do require: install_cron
18:18 forrest nah you need to specify the type of require.
18:18 oherrala yep
18:18 forrest That's just the best thing I can think of.
18:18 forrest I still think creating a cron package is the 'easiest' idea to avoid code repetition.
18:19 laax joined #salt
18:19 forrest Really though, I'd consider having a different state for bsd in general though that just deals with the bsd crons :)
18:19 Logaan joined #salt
18:20 oherrala I can do: pkg.install: unless: /bin/true
18:20 oherrala so it has the pkg type, but doesn't really do anything
18:20 forrest Hmm, yeah that could work as well, but then you need a managed file to drop that in
18:21 forrest specifically for bsd, that always executes prior to the cron ID :)
18:21 oherrala oh, well. Thanks for help. I'll continue writing stuff to learn Salt. This problem is now somehow solves =)
18:23 arc-impulse joined #salt
18:24 Logaan hello, using nssm to launch a java app on a windows server, which works fine. However, when I use nssm status, it has a return value of 0 regardless if the service is running or stopped.
18:25 Logaan Is there a way to capture the stdout of the nssm status and do an action based on the stdout?
18:25 Configio joined #salt
18:25 Logaan so, it is says "SERVICE_STOPPED", then I shoud start it
18:25 forrest Logaan: Hmm, I'm not super familiar with nssm, I did however see there are a couple open issues: https://github.com/saltstack/salt/issues?utf8=%E2%9C%93&amp;q=is%3Aissue+is%3Aopen+nssm
18:25 Logaan but stdout of SERVICE_RUNNING, I don't need to do anything
18:26 mrwboilers joined #salt
18:26 Logaan well, Salt uses nssm to run the minion on windows, I think that is what those issues are about
18:27 forrest Logaan: Gotcha, yeah my windows salt knowledge is pathetic ;)
18:27 Logaan ya, mine too. heh, Got a new job, and have to use Windows now
18:27 Logaan but I guess I want to do something like:
18:28 kevinquinnyo i need to create an ssh keypair on a minion and distribute the public key out to a group of other minions -- is there a best way to do this?
18:29 kevinquinnyo is it best to just generate it myself and store the priv key and pub key as pillar values
18:29 kevinquinnyo or is there a module that would help for this
18:29 Logaan javaapp: cmd.run: name: "nssm.exe start service" onlyif: {% if <output of nssm.exe status service> == "SERVICE_STOPPED" %>
18:30 Logaan I know that's not syntatically correct
18:32 forrest Logaan: how about this, use an unless in the cmd.run, so basically - unless: nssm.exe status service | pass to something that returns 1/0 such as powershell
18:32 forrest Logaan: That is how I'd do it
18:34 forrest I don't know enough about nssm to know what a return value on the command line looks like
18:34 arc-impulse joined #salt
18:37 zer0def joined #salt
18:39 zer0def quick question in regard to ec2_pillar - what do i need to provide for it to work?
18:40 trph joined #salt
18:40 trph joined #salt
18:40 zer0def getting a: "NoAuthHandlerFound: No handler was ready to authenticate. 1 handlers were checked. ['HmacAuthV4Handler'] Check your credentials"
18:41 Logaan hmm, the output of nssm LOOKS like a string, but its not really: c:\salt\nssm.exe status salt-minion | find "RUNNING"
18:41 Logaan c:\salt\nssm.exe status salt-minion | find "R"
18:41 trapha joined #salt
18:41 Logaan R
18:41 Logaan Just finds the letter, not the work
18:41 Logaan word
18:41 Logaan echo test | find "e"
18:41 cpowell joined #salt
18:41 Logaan returns test
18:44 zer0def well, nevermind, i'm asking in the wrong channel
18:45 arc-impulse joined #salt
18:46 Logaan >>> import subprocess >>> p = subprocess.Popen(['c:/salt/nssm.exe', 'status', 'salt-minion'], stdout=subprocess.PIPE) >>> p.stdout.read() 'S\x00E\x00R\x00V\x00I\x00C\x00E\x00_\x00R\x00U\x00N\x00N\x00I\x00N\x00G\x00\r\x00\n\x00'
18:46 Logaan from python shell
18:47 laax_ joined #salt
18:47 TomJepp it's utf-16
18:47 Logaan the output is like some binary format
18:47 TomJepp Logaan: nah, two byte unicode chars
18:47 TomJepp pretty universal on windows
18:48 Logaan ah, not too familiar with windows in all honesty
18:48 Logaan so, how to translate it then?
18:48 Guest91008 joined #salt
18:48 Guest91008 left #salt
18:49 TomJepp try something like <string var>.decode('utf-16-le').split()
18:49 ronT1V iggy: raet transport only requires 1 port, thank you for the direction
18:50 TomJepp also try "utf-16"
18:50 TomJepp I can't remember if windows uses little endian or big endian byte order for strings
18:50 TomJepp also I have no idea why that .split() is there
18:50 TomJepp ignore that :p
18:50 s_kunk joined #salt
18:51 bhosmer_ joined #salt
18:51 Logaan huh, cool
18:52 markm joined #salt
18:52 mohae_ joined #salt
18:52 TomJepp I do lots of windows... so I'm used to it :)
18:56 arc-impulse joined #salt
18:59 Configio joined #salt
18:59 bhosmer_ joined #salt
18:59 rocket joined #salt
19:00 rocket is there a way to conditionally include formula?  eg if a service is not installed only then do all of the steps to install that service?
19:00 rocket I am trying to think of ways to reduce the amount of time a state needs to go through all of its steps etc
19:01 dthom91 joined #salt
19:02 linjan_ joined #salt
19:04 Logaan forrest, TomJepp thanks for the tips. I'll play around with this
19:04 forrest Sounds good.
19:06 openfly joined #salt
19:06 openfly what's the right way to trigger an included sls state upon completion of another state.
19:06 openfly eXistenZNL: i have an action.restart state that i use to restart a service
19:06 openfly errr
19:06 openfly example
19:07 openfly and i call it several times here and there
19:07 arc-impulse joined #salt
19:07 openfly require seems to execute the sls state prior to completing it's own state
19:07 openfly so what's the ... require_after of salt?
19:09 iggy rocket: that's not really how salt works
19:09 iggy openfly: watch?
19:10 cpowell joined #salt
19:10 iggy rocket: f.ex. service.restart watches a config file
19:10 Nazzy joined #salt
19:10 openfly iggy i don't think so
19:10 openfly i'm not watching anything
19:11 openfly i kind of want to call the state after another state completes...
19:11 iggy I'm saying that's how you'd do something after something else
19:11 rocket iggy: as soon as I asked the question I realized I can just add an unless flag to my really slow step
19:11 protoz_ joined #salt
19:12 Nazzy joined #salt
19:12 openfly but does watch work on an sls... sub function?
19:12 cpowell joined #salt
19:12 openfly i mean... i guess i could break this thing into 4 more sls files... but that seems excessive
19:13 iggy check the docs for watch, I think it shows the use pretty well
19:13 iggy if not, you can look at the formulas for inspiration
19:13 openfly i think it's not what i want at all
19:13 openfly i've been reading the docs
19:13 iggy or https://github.com/SS-archive/salt-states
19:13 cpowell_ joined #salt
19:14 openfly it's kind of an ass backwards way to do this
19:14 Logaan TomJepp: Its strange that the first letter isn't in unicode?
19:14 TomJepp Logaan: I think you need to use 'utf-16' rather than 'utf-16-le', not sure though
19:14 Logaan well, just looking at the raw string
19:14 TomJepp byte order of a unicode char varies depending on if it's little endian of big endian
19:15 babilen rocket: I mean in a way salt already works like that (in that it checks if a state has already been achieved). You just want to add more "only do this if ..." logic to your definition of what you want to achieve
19:15 TomJepp the string is low-byte high-byte, low-byte, high-byte,
19:15 TomJepp etc
19:15 Logaan first letter is just a 'S', but the next is '\x00E'
19:15 TomJepp no
19:15 TomJepp first letter is 'S\x00'
19:15 TomJepp then 'E\x00'
19:15 Logaan ahhh, yes
19:15 iggy babilen: sounds like it's more of a "manual install" process (rather than salt just querying the package manager)
19:17 babilen iggy: Well, in that case there might be some programmatic check to determine if it has been installed previously. If that isn't possible then there's always the possibility to set a grain in a dependent state and check that
19:17 babilen But yeah, I'm sure you are right in that it is probably not properly packaged
19:18 openfly who is rocket?
19:18 denys joined #salt
19:18 openfly oh other person
19:18 openfly srry
19:18 babilen openfly: A salt user who asked for support in here
19:18 baweaver joined #salt
19:19 openfly yeah, i was just confused by the flow of conversation
19:19 Configio joined #salt
19:21 mrwboilers The doc page for salt.runners.drac says that the drac credentials need to be configured on the Master config file. But it doesn't say how to do that? Anyone have more info on that?
19:21 openfly btw here's how i got where i am... I have a service, tomcat, that i restart.  It needs a time delay to come up.  init_delay doesn't work, because... it just doesn't.  so i have a restart.sls for tomcat.  but i also have an operation that needs to execute that restart upon completion in another sls.  I want that to be visible in that operation, NOT in the restart script.
19:21 openfly require won't do that.
19:21 openfly watch... i think won't do that?
19:21 openfly so... now I am kinda confused.
19:24 win_salt joined #salt
19:25 Akhter joined #salt
19:26 baweaver joined #salt
19:27 denys damn, I can't get netapi to work (using rest_cherrypy): see https://gist.github.com/anonymous/b98f0ea150bd7d7a001a ; login works, but the second request gets a 401 ; any ideas?
19:27 arc-impulse joined #salt
19:28 mrwboilers Also, do some modules need to be installed/enabled? I just tried using the drac module but it just said "Module 'drac' is not available."  - https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.drac.html
19:30 Phtes pillars the only way to get "parameterized" data into a state?
19:30 babilen mrwboilers: Don't you need ipmitool for that?
19:31 mrwboilers babilen: I have no idea. I just discovered the drac module and runner exist.
19:31 babilen Or racadm ...
19:31 babilen Yeah, probably racadm ... let me check
19:32 babilen Indeed, you need to install racadm or the module won't load (it uses racadm to do things)
19:32 openfly Phtes you can import_yaml
19:32 openfly on a yaml
19:32 openfly or import_json as well
19:32 openfly etc
19:32 openfly and then iterate that as a list or dict
19:32 mrwboilers babilen: just on the master, or on each minion?
19:33 Phtes openfly:  I see, trying to come up with a decent way of a vhost generating script dealio where you just supply the server name and ssl true or false xD
19:33 dthom91 joined #salt
19:33 babilen mrwboilers: On the minions
19:33 Akhter joined #salt
19:33 babilen It is the actual command that is being run to achieve things
19:33 openfly yeah i do something similar for importing a plugin list to download / update
19:33 mrwboilers babilen: damn. That's a pain in the ass to install if you aren't using RHEL (it even doesn't like CentOS)
19:34 openfly Phtes some of this might be useful it was posted earlier by jbrnds :  https://gist.github.com/anonymous/26065b22760ae377dbb3
19:34 openfly one note .iteritems() works as well
19:35 babilen mrwboilers: http://linux.dell.com/repo/community/deb/latest/ (not too hard, have done it earlier this week)
19:35 Phtes openfly:  ty
19:36 Aleks3Y joined #salt
19:36 mrwboilers babilen: Really? I tried not that long ago and it checked the OS and if it was CentOS, it wouldn't install and just said RHEL only.
19:36 mrwboilers babilen: I'll revisit it
19:37 babilen mrwboilers: Well, the repository I linked is for Debian. I am sorry to learn that you are using CentOS and it might very well be a pain in the arse there.
19:37 mrwboilers My environment is a mix of both centos and ubuntu, but mostly centos.
19:37 TyrfingMjolnir joined #salt
19:38 iggy mrwboilers: paramiko and there's an example what to put in the config file (drac:\n  username: ...\n  password: ...)
19:38 TomJepp mrwboilers: can't you just edit the install script for it then?
19:38 Configio some of the execution modules support reporting back the version of the component its tied to, such as mysql.version, tomcat.version, and apache.version. Is there anyway to simply and dynamically link this info to a grain? Would the right way to go be to write a formula that updated the grain on a call to highstate, or is there an easier way to do it?
19:38 iggy mrwboilers: the runner needs racadm on the master, the drac module needs it on the minions
19:38 arc-impulse joined #salt
19:39 cresten Hey guys, so I want to use salt to handle my daily DB dumps
19:39 openfly https://gist.github.com/openfly/905963795159e7f791ca   <---- wierd behavior
19:39 openfly thoughts?
19:39 cresten ahh, is there like a fancy salt cron thing I can use? Or is it just the cron saltstate stuff
19:39 babilen iggy: So just SSH into the DRAC and work from there, yeah. But then that doesn't really work with the drac execution module, does it?
19:40 babilen cresten: You can schedule tasks in salt or configure cron jobs.
19:40 iggy babilen: the original question said runner
19:40 mrwboilers iggy: Where's that drac credential config example?
19:40 mrwboilers iggy: babilen: I'm interested in both the module and the runner
19:41 openfly as an include into a high state ... this one particular action doesn't import properly
19:41 babilen iggy: "drac module and runner" :)
19:41 openfly but it runs fine on its own
19:41 openfly so weird.
19:41 iggy mrwboilers: on the runner doc page... but I'm really not sure if you're talking about the runner or the module at this point
19:41 babilen cresten: https://docs.saltstack.com/en/latest/ref/states/all/salt.states.cron.html + https://docs.saltstack.com/en/latest/topics/jobs/schedule.html
19:41 cresten babilen: Thanks so much! Sorry, kind of a ‘let me google that for you question’, but I’m still learning the salt vocabulary
19:41 mrwboilers iggy: I'm interested in both. The doc page for the runner said the credentials need to be configured in the master config, but I didn't see any example
19:42 teryx510 joined #salt
19:42 babilen cresten: Don't worry. :)
19:42 iggy mrwboilers: that's because the docs are broken... check the source
19:42 fersur3 joined #salt
19:43 fersur3 I want to run multiple instances of a daemon.    Is there a way to specify something like a list of things in a Pillar variable and then have a state (Jinja?) that has a for loop, for "item" in list create three states?    Like, $item: file.managed    and create    item1: file.managed    item2: file.managed   and item3: file.managed   but have just the line $item: file.managed with the fore loop above it?
19:45 GreatSnoopy joined #salt
19:45 openfly https://github.com/saltstack/salt/issues/22852  mother fer
19:46 openfly i've been using saltstack for a week, i've already hit 2 bugs.  on basic operations.
19:47 iggy not a bug
19:47 Ztyx joined #salt
19:47 openfly looks like a bug to me
19:48 mapu joined #salt
19:48 opensource_ninja joined #salt
19:48 Logaan TomJepp, forrest: There we go, this should do the trick: c:\salt\bin>c:\salt\nssm.exe status salt-minion | c:\salt\bin\python -c "import sys;s=sys.stdin.read().decode('utf-16').split(); print s[0].find('SERVICE_RUNNING')"
19:48 Logaan returns 0,
19:48 forrest lol
19:48 Logaan -1 if the string isn't found
19:48 forrest awesome
19:49 Logaan Thanks again
19:49 elsmo joined #salt
19:49 Ztyx1 joined #salt
19:49 arc-impulse joined #salt
19:50 baweaver joined #salt
19:55 rm_jorge joined #salt
19:57 ajw0100 joined #salt
19:58 Ztyx joined #salt
19:58 Ztyx joined #salt
19:59 dthom91 joined #salt
20:00 arc-impulse joined #salt
20:01 cberndt joined #salt
20:02 _JZ_ joined #salt
20:04 Configio joined #salt
20:06 notnotpeter joined #salt
20:08 stanchan joined #salt
20:09 amcorreia joined #salt
20:11 arc-impulse joined #salt
20:11 baweaver joined #salt
20:14 openfly so... i am pretty sure this is a terrible way to do this...
20:14 zwi joined #salt
20:18 openfly https://gist.github.com/openfly/bfe1d647be702b9bd701
20:18 Ztyx joined #salt
20:19 Configio joined #salt
20:20 dthom91 joined #salt
20:20 monkey66 joined #salt
20:21 Ztyx joined #salt
20:21 pratikmallya joined #salt
20:22 pratikma_ joined #salt
20:22 knite joined #salt
20:24 knite if I have a salt state with, for example, pkg.managed and pkg.installed sequentially and both indented under the same label, will they always be run sequentially, or do I still need an explicit require?
20:24 dthom91 joined #salt
20:25 forrest knite: That won't work, you can only have one pkg.* declaration per state ID
20:26 KyleG joined #salt
20:26 KyleG joined #salt
20:26 arc-impulse joined #salt
20:29 Configio joined #salt
20:33 Ztyx1 joined #salt
20:34 denys joined #salt
20:34 zer0def joined #salt
20:37 trapha joined #salt
20:37 trapha joined #salt
20:37 arc-impulse joined #salt
20:43 cberndt joined #salt
20:43 baweaver joined #salt
20:43 Configio Does anyone here maintain a completely separate salt master server for their non-production environments than they do for their production environment?
20:48 arc-impulse joined #salt
20:49 iggy knite: what forrest said, but also everything should be top-down
20:49 knite forrest: typo, I meant pkgrepo.managed and pkg.installed. :-)
20:49 forrest ahh
20:50 knite iggy: hm, that runs counter to some docs I was reading before you replied. I concluded that the require statement was, well, required, and that salt isn't guaranteed to process top-down.
20:50 iggy knite: for pkgrepo, you should still use require_in (it's special for pkgrepo, it guarantees a pkg update before the require_in'ed state)
20:51 iggy knite: what docs say that?
20:51 knite iggy: so require_in on the repo is better than require on the package?
20:52 iggy there's a note in the pkgrepo docs about it
20:53 knite iggy: http://www.saltstat.es/posts/dependencygraph.html , bold text "order of appearance does not set a dependency graph". so blog post rather than docs.
20:54 openfly i think require may be a bit buggy
20:54 openfly but that's just me
20:55 openfly it's possible i don't understand it's execution chain
20:55 openfly creating the idea of unit tests in salt states... is kind of really really hard
20:57 iggy knite: that page is so wrong it's not funny... salt doesn't process states serially
20:57 knite sigh, learning things online is so hard sometimes.
20:58 tanta if you want to process in order
20:58 tanta just run a list of salt-call state.sls state
20:58 tanta I have done that before when order mattered, or I wanted to be surgical
20:59 knite I don't, I'm just trying to write minimal state files, so if require is unnecessary when I have a pkgrepo followed by a pkg.*, I want to whack the require. :-P
20:59 quantumriff joined #salt
20:59 iggy by default Salt operates top-down, using requisites can change the ordering... so you end up in a situation of doing everything with requisites or doing everything with ordering... newer versions aren't as bad in that regard
20:59 arc-impulse joined #salt
21:00 quantumriff I have a quick question about using shadow.set_password..
21:00 terratoma joined #salt
21:00 quantumriff is that something I can safely use in a state? or am I better off running that just once?
21:00 geekatcmu Last I checked, "doing everything with ordering" doesn't help if you need stuff in other states to happen first.
21:00 quantumriff I am trying to do this: http://pastebin.com/D5zEYMZQ and geting mapping values are not allowed here, for the name
21:00 geekatcmu Because you rapidly get into circular situations.
21:00 iggy knite: in general you don't need them, but in the case of pkgrepo, it's a special case that I mentioned before
21:04 quantumriff or is shadow.set_password mainly designed to run from the CLI (so you don't keep resetting it?)
21:04 Configio joined #salt
21:05 Configio salt '*' is too powerful with execution modules. I wish there was a way to force the equivalent of a dry-run by default. I can do that with highstate by setting test=True on the minions, but can't do it with general interactive usage via "salt '*'" at the CLI.
21:08 forrest Configio: You'd have to alias it to something else.
21:08 forrest Which would be meh
21:08 forrest or set up restricted users.
21:08 knite iggy: one more quick question. I have a salt sate for adding a private repo, I have multiple separate states installing packages from that repo (and configuring those packages, hence the multiple files). should each of those states do an include on the repo state and then require: sls: <repo state id>?
21:10 arc-impulse joined #salt
21:12 baweaver joined #salt
21:13 Configio forrest: It just feels to much like running as root all the time, but instead of doing that on one host, you're potentially doing it on all your hosts simultaneously.
21:13 openfly okay i gotta ask, if I want to run, then re-run a state...
21:13 openfly how am i supposed to do that?
21:13 Configio It's one thing to code review your various sls files and pillars, it's another thing entirely to worry about interactive use of salt '*' by any member of your operations team.
21:14 openfly especially if that is conditional
21:14 forrest Configio: Yep, one of the downsides of config management.
21:14 openfly you could jenkins your salt execution and throw some testing in front.
21:14 Configio forrest: well, this isn't as much config management, as I can build checks into highstate by setting test=True by default. It's more the command line execution capability of salt.
21:15 forrest Configio: Then you need to restrict access: https://docs.saltstack.com/en/latest/ref/clientacl.html
21:15 Configio config management via saltstack is achieved via writing files for states.
21:15 forrest If you can't trust your admins
21:15 kevin-wk knite: i do a require: pkgrepo: <name> for that
21:15 forrest or have some concerns over module execution
21:15 openfly forrest so do i
21:15 Configio It's not about not trusting them. It's about not having them do certain kind of actions without knowingly escalating to do them.
21:15 kevin-wk require: sls: would work too
21:15 openfly but different
21:15 openfly require: sls: doesn't work at all
21:16 openfly it just says did that sls get run at some point prior to now
21:16 Configio The most experienced sysadmin could make a terrible mistake with global CLI execution.
21:16 knite kevin-wk: so everything is loaded into a global namespace? gotcha.
21:16 openfly if you need to run at multiple points require is useless
21:16 forrest Configio: I agree, and it's why I run very few modules. I almost exclusively use states.
21:16 knite openfly: I pulled the require:sls option straight from docs: https://docs.saltstack.com/en/latest/ref/states/requisites.html
21:17 openfly knite and yes
21:17 openfly ive been working with it all day
21:17 openfly it's limited to the point of being worthless
21:17 openfly if you say... require: restart.sls  all you are saying is... at some point in this state.whatever did restart.sls get ran?
21:17 openfly yes then you are good
21:18 openfly but it won't help you if you want to know if restart.sls got ran after somethingelse.sls or something:
21:18 Configio I just wish you could tag certain modules used by the salt CLI to ask for confirmation before executing in your salt configuration. :-)
21:18 openfly which makes states functionally broken as far as I can tell
21:18 openfly as you can't verify states or build dependency states.
21:19 openfly which... makes salt very less than usable.
21:19 murrdoc damn lotsa hate
21:19 cornfeedhobo is there a popular repo for grains, or would those just be in the various formulas?
21:19 openfly but hell i'm still looking for a solution
21:19 murrdoc what u trying to do open fly
21:19 openfly in this case i am loading some plugins into a jenkins environment
21:19 openfly and i conditionally want to restart tomcat if a new plugin gets installed
21:20 openfly i have a sls for restarting tomcat, because the service flag delay_init doesn't work
21:20 openfly and now i can't seem to get that restart to occur
21:20 openfly because... there's no way to tell salt to run it again.
21:20 Configio openfly: do you want to script the installation/configuration of jenkins via saltstack so that jenkins will execute this? Jenkins has the most knowledge about what's happening to itself.
21:21 openfly at least as far as i can tell
21:21 openfly i have an sls for doing the plugin checks against another sls of plugins / versions
21:21 Configio We have to schedule our jenkins restarts around here, as we have build jobs running constantly for CI/CD.
21:21 openfly also a custom set of plugins that load manually
21:21 openfly i also eventually want to run jobs and look for responses to unit test a jenkins build
21:22 openfly before i migrate anything to it
21:22 openfly but ... i cannot for the life of me figure out how to conditionally restart tomcat with a delaty
21:22 openfly i can restart tomcat with a delay
21:23 openfly and i can include and require that restart... but i can't conditionally force that state to be applied at the time that i need it applied.
21:23 mrwboilers left #salt
21:23 cornfeedhobo :(
21:23 openfly now maybe orchestration runner can do this
21:23 openfly but that's not really what i want
21:23 openfly because i'm not orchestrating multiple systems here
21:23 monkey66 joined #salt
21:24 kevin-wk cmd.wait with watch, make the action include a check for some condition and a sleep?
21:25 openfly well i have an sls that does a test.sleep()
21:25 openfly that works great for the delay
21:25 dthom91 joined #salt
21:25 openfly but on a basic level
21:25 khaije1 joined #salt
21:26 openfly i need to be able to conditionall apply a state from an sls.
21:26 openfly and a big reason for that is legibility of the state files.
21:26 nitay joined #salt
21:26 openfly and conditional does include applying that state multiple times.
21:28 Ztyx joined #salt
21:31 arc-impulse joined #salt
21:31 iggy how about you read up on Salt a little more instead of just telling everyone you see that stuff that works as expected is broken and salt is useless, etc...
21:32 nitay is there a way to detect whenver salt is run on a particular machine
21:32 iggy nitay: there are various events that run for different things
21:32 nitay my use case is i have pagerduty hooked up and would like fire an alert anytime sth runs against production
21:33 pratikmallya joined #salt
21:33 jodv nitay: including salt-call or just commands run from the master?
21:33 nitay jodv: good q ideally both
21:34 pratikma_ joined #salt
21:35 nitay sth like this https://gist.github.com/nitay/08103def589daca93245 - although even that would only trigger for highstate not things like state.sls ?
21:35 nitay what’s the best way to do this?
21:35 openfly not the case iggy
21:35 openfly i'm partly in this hole because of 3 open bugs
21:35 openfly most of them high priority, and open in some cases for over a year.
21:37 iggy I told someone to blog this the other day
21:37 jodv nitay: a Reactor would solve for commands run from the master, but not for salt-call
21:37 jodv or scheduled jobs
21:39 kermit joined #salt
21:40 larsfronius joined #salt
21:41 arc-impulse joined #salt
21:44 bluenemo joined #salt
21:45 indispeq joined #salt
21:48 khaije1 when using salt-ssh to connect to Cent 5,7 I find that I can execute raw shell commands but not states... is there something I should be checking for in particular?
21:48 iggy khaije1: check the issue tracker?
21:48 iggy this sounds familiar
21:48 khaije1 It seems like I'm missing a pre-requesite somehow, just not sure what
21:48 khaije1 iggy: will do
21:49 * openfly steps away
21:49 Configio joined #salt
21:49 * khaije1 reads through https://github.com/saltstack/salt/issues/9790
21:51 whytewolf salt-ssh on cent 5.7? most likely a python issue?
21:51 khaije1 my experience seems to match the problem definition in issue 9790
21:52 iggy what version of salt?
21:53 khaije1 when I run $: salt-ssh '*' -r "python --version" ... the responses indicate version 2.6 or greater
21:54 khaije1 iggy: Lithium on Cent6, (updating from the saltstack repos may not be an option)
21:55 iggy version number... I don't have my periodic table handy
21:55 arc-impulse joined #salt
21:56 whytewolf 2015.5 was lithium
21:58 khaije1 pardon me iggy, 2015.5 is correct, I only just saw your message
21:59 murrdoc joined #salt
21:59 iggy that bug should have been fixed in 2014.7
22:00 khaije1 I'm aware
22:00 khaije1 the states work for Cent 6, just not for 5 or 7
22:01 arc-impulse joined #salt
22:03 khaije1 hold on ..
22:03 giantlock joined #salt
22:04 khaije1 I only updated from 2015.5.3 to 2015.5.5 in the last few minutes and it seems to have improved things
22:05 khaije1 there are now scattered errors for about half of the devices in a pattern that I might be able to attribute to insufficiently recent python versions
22:06 khaije1 probably, this can be regarded as fixed at least in so far as the initial operating contract now looks to be restored
22:09 trazomtg joined #salt
22:09 Configio joined #salt
22:10 trazomtg hello can i install a complete OS (Debain... windows...) on a clean ahd virgin machine with Saltstack
22:10 arc-impulse joined #salt
22:10 khaije1 thanks goodness ... I was really started to get worried there :)
22:11 robawt trazomtg: once you have installed a base operating system you can then configure it with Salt
22:11 protoz joined #salt
22:11 robawt trazomtg: look into PXE booting to get bits on disk
22:12 nitay joined #salt
22:13 trazomtg can i install an operating systeme on a clean machine with Saltslack?
22:13 corriera27 joined #salt
22:15 nitay jodv: how do I use that?
22:15 danlsgiga joined #salt
22:16 danlsgiga hi guys, anyone was able to do suffix all the items in a list? I have a server list and I want to suffix each item in this list with ':/folder'
22:16 corriera27 hola
22:16 danlsgiga was anyone able to do this using jinja?
22:16 trazomtg no response for me? help
22:17 whytewolf trazomtg: salt does not have PXE funcationality.
22:17 trazomtg so i can't?
22:18 corriera27 no hablo ingles chao
22:18 corriera27 left #salt
22:19 trazomtg if i can't which software use to install a remote operating system first then yse Saltslack
22:19 arc-impulse joined #salt
22:20 jodv nitay: https://docs.saltstack.com/en/latest/topics/reactor/index.html
22:21 whytewolf trazomtg: thats up to you. what ever tool you want to use. I like cobbler.
22:22 whytewolf but i also don't deal with windows at all
22:22 knite joined #salt
22:23 trazomtg thanks
22:24 trazomtg so i can onstall a complete operating system via cobbler then use Saltslack for instalmling applications?
22:25 whytewolf thats what i do with my current openstack setup. I use cobbler to install the host operating system. [even the salt minion] then use salt to install openstack on that host
22:25 whytewolf then use salt-cloud to launch instences in that cloud
22:26 opensource_ninja joined #salt
22:26 trazomtg thanks a lot!!! could you give me some web sites and informations about Cobbler?
22:26 whytewolf trazomtg: http://cobbler.github.io/
22:29 trazomtg is cobbler easy ti install and parameter? could you give me some of your personnal configuration files and parameters files?
22:29 arc-impulse joined #salt
22:29 whytewolf um, no. sorry. my install configs are my own stuff. and rather personal.
22:30 geekatcmu cobbler is easy in the same way salt is easy.
22:31 cresten joined #salt
22:31 trazomtg ok thanks very much !! i'm going to try
22:32 trazomtg my mail adress is trazomtg@free.fr Perhaps could you send me any help requests! bye!
22:33 trazomtg left #salt
22:37 kusams joined #salt
22:37 sunkist joined #salt
22:38 protoz joined #salt
22:39 bfoxwell joined #salt
22:42 arc-impulse joined #salt
22:46 aristedes joined #salt
22:52 arc-impulse joined #salt
22:52 Configio geekatcmu: Cobbler's getting pretty long in the tooth. Have you tried the Foreman with saltstack integration via foreman_salt and the smart salt proxy? It's an evolution of what cobbler was, and can sit on top of your config management tool. In this case, that would be salt.
22:52 geekatcmu nope
22:52 geekatcmu I haven't looked at Foreman in 4 years or so.
22:52 geekatcmu (and back then I ran a Puppet shop)
22:54 Configio I think it still gives you a little more functionality with puppet over chef or salt, but it does work pretty well with salt when you get it up and running. I believe there's someone from redhat who has a youtube video running down integration between Foreman and Salt. I've got it running in that configuration in test, and it seems pretty powerful.
22:54 Configio You can even use foreman as an external pillar source.
22:54 whytewolf you can use cobbler as a ext_pillar as well
22:55 Configio Who is maintaining cobbler these days?
22:55 Configio Foreman also has hooks into various provisioning services, and is easily extensible. The only downside, is that if you want to develop plugins for it, and you are not a ruby person, you might need to learn ruby first.
22:57 Configio though there are a bunch of plugins available. I think parts of the UI could use some polishing. It's too easy to destroy a VM without realizing it: when no compute resources are associated, deleting a host from the host list just deletes its information from foreman, but when a compute resource, like a VMWare environment is added, and allowed to associate it's VMs with your salt minions in the host list, deleting a host deletes the actual VM. That's a
22:57 Configio dumb UI design.
22:57 Configio yep
22:58 JDiPierro joined #salt
23:00 aristedes joined #salt
23:00 Configio bbiab
23:00 aristedes left #salt
23:02 arc-impulse joined #salt
23:03 jodv joined #salt
23:08 arc-impulse joined #salt
23:09 knite what's the difference between salt['pillar.get']('thing') and pillar['thing']? second form is easier on the eyes, but I see a lot of example states using the former.
23:21 Guest46 joined #salt
23:23 Configio joined #salt
23:25 zwi joined #salt
23:25 mosen joined #salt
23:25 kevinquinnyo joined #salt
23:25 forrest joined #salt
23:29 jbrnds joined #salt
23:37 Guest46 left #salt
23:39 kusams joined #salt
23:41 larsfronius joined #salt
23:43 notnotpe_ joined #salt
23:44 estahn joined #salt
23:46 JDiPierro joined #salt
23:47 jodv joined #salt
23:48 Ztyx joined #salt
23:49 nate_c joined #salt
23:49 Ztyx joined #salt
23:50 jodv_ joined #salt
23:53 jodv joined #salt
23:57 pratikmallya joined #salt
23:59 baweaver joined #salt

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