Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2016-12-14

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

All times shown according to UTC.

Time Nick Message
00:00 bfrog joined #salt
00:00 bfrog I can't use grains in pillars can I?
00:00 sh123124213 I can only see max_event_size: 1048576
00:00 UtahDave bfrog: Yes, you can.
00:01 bfrog oh fantastic, I somehow was under the impression I couldn't :-)
00:01 UtahDave The master caches the minion's grains and the minion dictionary within the jinja environment has the specifiic minion's grains in it.
00:01 UtahDave sh123124213: I believe that is the correct config option
00:02 bfrog nice
00:04 ALLmightySPIFF joined #salt
00:05 sh123124213 UtahDave: so I need your option on a solution. I'm trying to download/upload files through http from the minion to the syndic/s and I need some way to authenticate the minion with the webserver
00:06 sh123124213 I though about auth module in nginx that would call another api to allow or dismiss a request for download upload
00:06 sh123124213 and this api would somehow authenticate and check if the minion is the correct one ><
00:07 UtahDave why through http?
00:07 sh123124213 I don't know how the api would do that though but I know the minion has its pub keys accepted so ..
00:07 joehh joined #salt
00:07 sh123124213 I was expecting that question :) cos its faster then salt
00:07 UtahDave are they large files or something?
00:07 sh123124213 yes
00:07 UtahDave gotcha.
00:08 UtahDave have you considered using rsync?
00:08 sh123124213 I don't want to involve ssh no
00:08 gtmanfred you could use rsyncd
00:08 gtmanfred which could be authenticated, but sshd is probably already running
00:08 UtahDave Yeah, I'm using rsyncd on an internal project.
00:09 gtmanfred and since you are already adding another service... nginx
00:09 sh123124213 yes but that service doesn't need a user on the server
00:09 sh123124213 when I say server I mean minion
00:09 gtmanfred neither does rsyncd
00:10 gtmanfred at least I am pretty sure you can put a password on a rsync:// endpoint, and it can take a password that is not in pam
00:11 gtmanfred yup
00:11 gtmanfred auth users
00:11 gtmanfred This parameter specifies a comma and/or space-separated list of authorization rules. In its simplest form, you list the usernames that will be allowed to connect to this module. The usernames do not need to exist on the local system.
00:11 sh123124213 I would do basic auth with http if I wanted to have passwords :P
00:11 gtmanfred or you could do it with rsyncd
00:11 sh123124213 reading docs on rsyncd
00:11 raspado joined #salt
00:11 gtmanfred yar, it can configure its own secrets file, similar to htpasswd
00:11 gtmanfred and doesn't require the user to be on the system
00:12 UtahDave Yep. that's exactly what I'm doing on this internal project.
00:12 gtmanfred :+1:
00:13 UtahDave I have to head out now, but you can ping m tomorrow if you'd like an example from what I'm doing, if that will help.
00:14 sh123124213 no encryption, allow per subnet/password.
00:15 sh123124213 don't see a reason to change from https to rsyncd
00:16 euidzero joined #salt
00:17 sh123124213 unless I would be able to use pub/priv keys from the minion
00:22 N-Mi_ joined #salt
00:25 jas02 joined #salt
00:26 ponyofdeath anyone can help me reverse engineer what the php.ng fomula expects for me to get a specific version from the php ppa? https://github.com/saltstack-formulas/php-formula/blob/master/php/ng/installed.jinja here is my pillar output: https://bpaste.net/show/75927ea22215
00:27 gtmanfred ponyofdeath: you should only have to match your stuff up to the pillar.example https://github.com/saltstack-formulas/php-formula/blob/master/pillar.example
00:29 ponyofdeath gtmanfred: there is no version there
00:29 ponyofdeath gtmanfred: and its older so it does not work on 16.04
00:29 brasticstack joined #salt
00:30 ponyofdeath gtmanfred: the ng one works but i cannot get it to install 5.6 instead of 7
00:30 brasticstack Hi! Is there a list of available tgt_types that can be used when using saltmod.state? I'd like to build a reactor state that uses the pcre matcher.
00:31 gtmanfred my guess would be to drop the package versions under here https://github.com/saltstack-formulas/php-formula/blob/master/pillar.example#L20 , but you could ask in the issues on that repository
00:31 gtmanfred or on ask on the salt-users list
00:31 gtmanfred brasticstack: any of the ones that you could set as match: in the top file
00:32 gtmanfred lemme see if I can find the list
00:32 brasticstack grazie!
00:33 gtmanfred if you want to use pcre, i believe it is just pcre
00:33 brasticstack looks like it would be called 'pcre'.
00:33 brasticstack Cool, thanks!
00:33 gtmanfred I think that it might use this https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.match.html
00:34 whytewolf I was just about to post that gtmanfred I think it is close if not exactly that.
00:35 gtmanfred ahh yeah it is, it is using salt.matcher.Matcher()
00:35 gtmanfred that should be documented on the top/targeting page
00:36 gtmanfred I am making a doc issue for that
00:37 brasticstack I searched all the docs for tgt_type and didn't find that, so yeah, makes sense
00:37 whytewolf it is normally easier just to look for targetting.
00:38 whytewolf instead of by tgt or tgt_type opr any of the other 15 ways salt refers to targeting
00:38 brasticstack Gotcha. Thanks for the info!
00:39 gtmanfred https://github.com/saltstack/salt/issues/38242
00:39 saltstackbot [#38242][OPEN] Document all the tgt_type|expr_form options in the top file and compound targeting pages | Description of Issue/Question...
00:39 gtmanfred brasticstack: you can use any of the functions on salt.minion.Matcher just remove the _match suffix, https://github.com/saltstack/salt/blob/develop/salt/minion.py#L2762
00:40 whytewolf one of these days someone should refactor everything in salt to a single name for some of these things... ;)
00:40 gtmanfred yup
00:40 gtmanfred whytewolf: so, i am starting over on creating my openstack environment
00:40 whytewolf really?
00:41 gtmanfred SOMEONE! logged into my icloud storage and deleted the images of my servers in the VMware directory
00:41 gtmanfred and that deletes them on the machines as well...
00:41 whytewolf wtf :(
00:41 gtmanfred that someone was me...
00:41 gtmanfred the code is all safe, it was just i have to make a new cluster
00:41 whytewolf ouch ...
00:41 whytewolf oh that won't be too difficult
00:41 gtmanfred yeah it is just going to take a day
00:42 whytewolf and you can use your already created code to do it right?? ;)
00:42 gtmanfred yeah, imagine my surprise when i come back to vmwarefusion, and find it saying all my vdisks are missing
00:42 gtmanfred heh, i am going to automate spinning up the cluster this time... i didn't last time, which is why I took a bunch of snapshots
00:42 whytewolf oh i can image we had something simalar happen at work
00:42 ponyofdeath gtmanfred: well I did a git pull it seems it was fixed 3 days ago
00:42 gtmanfred ponyofdeath: awesome
00:44 gtmanfred mmmm my salsa is very good
00:44 whytewolf mmmm salsa... i like salsa
00:45 mattp_ joined #salt
00:45 gtmanfred whytewolf: highly recommend making this one http://tideandthyme.com/rustic-roasted-tomato-salsa/
00:45 shbst joined #salt
00:46 gtmanfred bonus points if you roast some heirloom tomatos instaead of canned
00:46 gtmanfred i also used 4 serano peppers instead of 2 jalepenos
00:46 mattp_ is it possible to use environments for staging releases? should i be doing it another way?
00:46 mattp_ ie dev -> stage -> prod
00:46 gtmanfred you could do that
00:46 mattp_ i see these as common examples for environments but never an actual implementation that shows how it would work
00:47 mattp_ also some weirdness about merging of env's that confuses me
00:47 whytewolf ohh. book marking that recipe
00:47 gtmanfred mattp_:are you on 2016.11.0?
00:47 mattp_ gtmanfred: ya
00:47 gtmanfred mattp_: check out top_file_merge_strategy: same https://docs.saltstack.com/en/latest/ref/configuration/minion.html#std:conf_minion-top_file_merging_strategy
00:48 mattp_ I was looking at that
00:48 mattp_ same would be the desirable choice in this case?
00:48 gtmanfred i think so, it would only be able to find the state files in that specific environment
00:49 gtmanfred This file contains all the information there is on how the different merge strategies work https://docs.saltstack.com/en/latest/ref/states/top.html
00:50 gtmanfred and the 4 main scenarios that are designed for
00:51 mattp_ gtmanfred: so for a versioned release of a 'fooApp', it would start in dev
00:51 mattp_ to release into qa it would be a merge to the qa branch (assuming gitfs here), and once qa clears, the merge to prod branch?
00:53 mattp_ (just trying to understand how such a workflow would work with environments)
00:53 gtmanfred yes
00:53 gtmanfred that is what I would do
00:54 gtmanfred so you only ever make new commits to dev
00:54 gtmanfred and always merge to the other two
00:54 mattp_ gtmanfred: got it
00:54 mattp_ and the recommendation of a separate repo for your topfile with only a master is still true?
00:55 mattp_ learning salt was such a painful endeavor. but I really like it now that I'm here
00:55 gtmanfred yeah, you could still do that.  I don't use a top file for that kind of environment, cause ususally there is more than one step in the deployment, so I use a state.orchestrate from the runner and pin my servers to a specific minion
00:55 gtmanfred my minions to a specific environment
00:56 mattp_ hmm
00:56 mattp_ gtmanfred: you dont use a topfile *period* or you dont use a topfile specifically for deployment?
00:57 gtmanfred yeah, for deployment
00:57 gtmanfred i use a top file for the base setup
00:57 gtmanfred but you don't need to
00:57 gtmanfred you could do it either way
00:57 mattp_ because executing the orch.. will apply the states?
00:57 whytewolf for my i typically use orchestration for deployment and top for maint
00:58 mattp_ whytewolf: do you have a salt '*' state.apply on cron?
00:58 gtmanfred right, orchestrate just makes sure my states apply as just like top does, but it makes sure the database servers get updated first, and run in a batch, and rotated out of the cluster
00:58 j-- joined #salt
00:58 j-- joined #salt
00:58 whytewolf mattp_: kind of i use the built in schedule
00:58 gtmanfred mattp_: kill off cron :P https://docs.saltstack.com/en/latest/ref/states/all/salt.states.schedule.html
00:58 mattp_ right right
00:59 gtmanfred mattp_: https://docs.saltstack.com/en/latest/topics/jobs/#scheduling-jobs
00:59 mattp_ yes, thanks
00:59 gtmanfred :)
00:59 mattp_ gtmanfred: im struggling to differentiate your orch runs from your app formula
01:00 mattp_ so your db is updating, needs to be removed from the cluster. what state is dictating this change?
01:00 gtmanfred well, for my deployment and merging into prod from qa, i need my server pulled out of the load balancer, then the app updated, then restarting services, then plugged back into the load balancer
01:00 gtmanfred i don't need to make sure that vim is installed
01:01 mattp_ gtmanfred: so even though you merge the states from dev -> prod branch for example
01:01 gtmanfred dev -> qa -> prod, when the states get merged to prod, it kicks off the orchestration
01:01 mattp_ each branch has a orch.states.upgrade that you salt-orch
01:02 gtmanfred basically yeah
01:02 gtmanfred it is the same one, the branch to deploy the app from is controlled from pillars
01:03 gtmanfred I really need to sit down and come up with a concrete example of using environments and doing this deploymethod using my salt-docker-demo
01:03 gtmanfred whytewolf: i made this https://github.com/gtmanfred/salt-docker-demo/
01:03 mattp_ gtmanfred: er, what do you mean by that re; "it is the same one, the branch to deploy is controlled from pillars"
01:03 mattp_ was following up until then :)
01:03 gtmanfred so, they all have one, but it is the same one, because it has been merged all the way to prod
01:04 gtmanfred and then the branch of code for the app is controlled from pillars
01:04 mattp_ ahh ok
01:04 mattp_ branch, or tag?
01:04 mattp_ i guess it doesnt matter
01:05 gtmanfred right, either, it could also be versions from your pypi server
01:05 gtmanfred whatever you designate, I need to sit down and do a demo in docker of this cicd pipeline
01:05 gtmanfred and write a blog post about it
01:05 whytewolf gtmanfred: very nice, always nice getting more demos
01:06 gtmanfred i am using that docker demo to troubleshoot stuff for github issues now
01:07 gtmanfred aight, I am going to go get some dinner o/
01:07 mattp_ gtmanfred: thank you sir, this conversation has been much helpful
01:07 mattp_ and id be very curious to read said blogpost
01:08 madboxs joined #salt
01:08 madboxs joined #salt
01:08 mattp_ gtmanfred: one last one, are you automating an orch run from merge?
01:09 evle joined #salt
01:11 gtmanfred yes
01:12 gtmanfred mattp_: this is the masterless version https://blog.gtmanfred.com/masterless-reactor.html
01:12 mattp_ ah ok, so you are hooking the github webhook directly to a reactor. nice ok
01:12 mattp_ thank you!
01:13 gtmanfred yup
01:13 gtmanfred and if you grab the github stuff from nitrogen, you can verify the Signature from github too
01:14 mattp_ hmm
01:14 mattp_ so youd want the reactor to hit master in an orch run, yes?
01:15 gtmanfred yeah, you can use the salt-api + webhook to drop an event on the master reactor
01:16 mattp_ gtmanfred: thank you
01:16 gtmanfred though, in that case, i don't just do the deploy of the app, because it is only 19 states total, so not worth the optomization
01:17 mattp_ what do you mean?
01:19 gtmanfred i just use the top file on the masterless minion, since it doesn't need the extra state.orch treatment
01:20 gtmanfred in the blog post*
01:20 mattp_ ah gotcha, its only on one minion
01:20 gtmanfred right
01:23 gtmanfred maybe this weekend I will play with getting the travis ci stuff working with the salt api hook, and verifying the signature
01:23 gtmanfred aight, i am out
01:28 nZac joined #salt
01:35 scoates joined #salt
01:46 delpanto93 joined #salt
01:47 yetanotherzero joined #salt
01:55 djgerm1 joined #salt
01:56 jas02 joined #salt
01:57 jmickle joined #salt
02:05 catpigger joined #salt
02:10 snc joined #salt
02:24 Nahual joined #salt
02:30 nZac joined #salt
02:46 sh123124213 joined #salt
03:08 djgerm joined #salt
03:20 rpb joined #salt
03:26 jas02 joined #salt
03:30 debian112 joined #salt
03:33 bastiandg joined #salt
03:37 raspado joined #salt
03:40 huleboer joined #salt
03:43 netcho joined #salt
03:51 s_kunk joined #salt
04:06 nZac joined #salt
04:08 tvinson joined #salt
04:13 swills joined #salt
04:21 rdas joined #salt
04:48 sh123124213 joined #salt
04:54 cilkay joined #salt
04:55 bocaneri joined #salt
04:57 jas02 joined #salt
04:57 cilkay Hi. I'm upgrading from 2015.5.3 to 2016.11.0. In one of my states, I have {% set remote_version = salt['postgres.psql_query'](user=util_user, password=util_pw, port=pg_port, host=remote_host, maintenance_db=util_db, query=get_version) %}, which works fine in the older release. In the newer one, I get: Jinja variable 'salt.utils.templates.AliasedLoader object' has no attribute 'postgres.psql_query'
04:57 cilkay Any idea how to work around that?
04:59 cilkay This is in a masterless minion, if that makes a difference.
05:08 onlyanegg joined #salt
05:11 Deliant joined #salt
05:24 buu Why did I need to execute salt-call test.ping on the minion machine before the master could get a response from test.ping?
05:44 netcho joined #salt
06:02 justanotheruser joined #salt
06:06 onlyanegg joined #salt
06:18 netcho joined #salt
06:22 preludedrew joined #salt
06:28 jas02 joined #salt
06:31 felskrone joined #salt
06:32 netcho joined #salt
06:45 sh123124213 joined #salt
06:46 samodid joined #salt
06:49 bluenemo joined #salt
07:05 madboxs_ joined #salt
07:07 keimlink joined #salt
07:13 Miouge joined #salt
07:17 teclator joined #salt
07:17 madboxs joined #salt
07:23 svg joined #salt
07:25 Straphka because that way the master knew of the minions presense and auto-accept was turned on?
07:31 nidr0x joined #salt
07:34 inad922 joined #salt
07:36 ivanjaros joined #salt
07:37 sh123124213 joined #salt
07:39 buu It had already accepted the key =[
07:39 buu hrm
07:44 debian112 joined #salt
07:44 fracklen joined #salt
07:48 buu Would 'sudo salt-minion' use a different key than 'systemctl start salt-minion' ?
07:52 netcho_ joined #salt
07:54 hemebond buu: Shouldn't do.
07:59 AndreasLutro might be that your salt-call just re-initiated the authentication process which had failed previously
07:59 fracklen joined #salt
07:59 buu AndreasLutro: It did in fact say that in the log, but why? I had already accepted the key earlier and didn't re-accept it
08:03 AndreasLutro maybe you didn't give the minion enough time to re-try authentication on its own, dunno
08:07 jas02 joined #salt
08:08 samodid joined #salt
08:10 bayindirh joined #salt
08:10 buu AndreasLutro: Like, it all worked then I stopped touching it for 8 hours and the master couldn't find it any more
08:10 buu we'll see if it happens again
08:12 o1e9 joined #salt
08:20 netcho joined #salt
08:22 jas02_ joined #salt
08:24 fracklen joined #salt
08:45 swa_mobil joined #salt
08:45 PhilA joined #salt
08:48 fracklen joined #salt
08:53 fracklen joined #salt
08:54 darioleidi joined #salt
08:56 Spike_ joined #salt
09:07 mikecmpbll joined #salt
09:14 teclator joined #salt
09:19 s_kunk joined #salt
09:21 jesusaur joined #salt
09:22 geomacy joined #salt
09:36 rdas joined #salt
09:38 jhauser joined #salt
09:38 sh123124213 joined #salt
09:41 sebastian-w joined #salt
09:42 netcho i have wierd issue
09:43 jas02 joined #salt
09:43 netcho "Rendering SLS 'base:autoscaling.asg-gb-api-staging-blue' failed: Jinja error: unsupported operand type(s) for -: 'StrictUndefined' and 'StrictUndefined'
09:43 netcho says it's here
09:43 netcho Ensure elb-{{ app-name }}-{{ env }}-{{ deployment }} exists:    <======================
09:44 AndreasLutro {{ app-name }} is interpreted as "variable app minus variable name"
09:44 netcho used to work before
09:44 AndreasLutro definitely didn't
09:45 jas02 joined #salt
09:45 gwen joined #salt
09:47 netcho gotcha.. so app_name should work :)
09:47 AndreasLutro indeed
09:50 netcho tipfeller when changin variable order :)
09:50 netcho thanks
09:56 jas02 joined #salt
09:58 mikecmpbll joined #salt
10:02 AvengerMoJo joined #salt
10:03 swa_work joined #salt
10:08 netcho joined #salt
10:11 Reverend morning chappos
10:14 funabashi joined #salt
10:19 cyteen joined #salt
10:21 jesusaur joined #salt
10:26 gwen Hello
10:26 gwen I'm trying to call a salt command from a sub module
10:26 Reverend okay
10:27 gwen and __salt__ is not available in the sub module's functions
10:27 gwen https://paste.debian.net/902145/
10:29 gwen as far as I understand, the __salt__ variable is set just in time, at the function call
10:29 N-Mi_ joined #salt
10:30 AntiSlash joined #salt
10:31 cyborg-one joined #salt
10:33 AndreasLutro if you want to call salt execution modules you have to use __salt__['module.function'] afaik
10:33 AndreasLutro and not sure if it supports submodules like that
10:39 gwen yes, it's just that, we cannot acces __salt__['module.function'] from a submodule
10:40 amcorreia joined #salt
10:40 sh123124213 joined #salt
10:40 gwen the submodule's function lives in a submodule but it launched from the parent module, we thought that it could be enough
10:40 gwen but apparently not...
10:49 DanyC joined #salt
10:49 AntiSlash hello, I'm working with gwen
10:54 haam3r hey...does file.directory also support the new keep option?
10:58 netcho_ joined #salt
11:02 jas02_ joined #salt
11:08 netcho_ i have a state that creates load balancers and autoscaling group on aws
11:09 netcho_ i havent used it in a while and all i change was some jinja variables
11:09 netcho_ now i get errors  KeyError: 'boto_vpc.get_subnet_association'
11:09 netcho_ and: State 'boto_asg.present' was not found in SLS 'autoscaling.create_gbapi_stack_blue'
11:10 netcho_ Reason: 'boto_asg' __virtual__ returned False
11:18 jesusaur joined #salt
11:22 neilf__ joined #salt
11:24 vegardx Is there anything like test-kitchen (chef) for salt?
11:30 Nickelflac joined #salt
11:35 sh123124213 joined #salt
11:38 thealphaking01 joined #salt
11:39 thealphaking01 hey guys
11:39 thealphaking01 i have a requirement, but i am not sure if it works or not
11:39 thealphaking01 on salt
11:40 babilen vegardx: kitchen salt comes to mind
11:41 thealphaking01 so basically, i have a couple of servers that would go on/off at any time acting as minions, and one server that acts as a master. what i need to do is, the moment my minion connects to my master, my master should run some state based on the minion id
11:41 babilen Easy with a reactor
11:41 DanyC left #salt
11:42 thealphaking01 is that something like a listener? what does it do?
11:42 babilen It allows the master to react to certain events by, for example, executing states on minions
11:42 babilen https://docs.saltstack.com/en/latest/topics/reactor/
11:42 thealphaking01 great :) just saw it :) thanks a lot :)
11:43 thealphaking01 didn't know this functionality was present :)
11:43 sh123124213 joined #salt
11:43 babilen You can also send events to the master (from states, using 'beacons', ...)
11:44 babilen To which the master can then react again
11:54 thealphaking01 joined #salt
11:55 thealphaking01 also, what is the extension of the config file present in /etc/salt/master?
12:02 daha joined #salt
12:06 Reverend can one have more than one onlyif ?
12:06 Reverend ^ on a cmd.run obvs
12:07 netcho_ as i mentioned earlier i have a state that creates 2 load balancer and autoscaling groups. in sls file i first have states for load balancers and then for autoscaling. in stdout looks like they are ran in correct order but event bus says different. it rans last state first and it fails
12:07 netcho_ http://hastebin.com/lenigoyesa.vbs
12:08 netcho_ tried with require but then it says 1 of the req is not met
12:08 netcho_ and again ot rans boto_asg before boto_elb
12:09 babilen Reverend: No, you have to express that in the command itself
12:09 Reverend blerp okay. thanks babilen
12:09 babilen thealphaking01: That *is* the config file .. it has no extension
12:10 babilen thealphaking01: You can have additional configuration files in /etc/salt/master.d/ and they'd end in .conf
12:10 netcho_ tried with order: 1,2,3 but still the same
12:11 netcho_ if u need state file i can sanitize it
12:15 thealphaking01 is there a way to check for a regex? say i want to check {% if data['id'] == 'mysql1' %}, but this should match 'mysql' + any number
12:15 babilen No -- jinja is shite
12:16 babilen I typically parse that in a custom execution function
12:17 thealphaking01 lets say i have two different kinds of minions and i want certain commands to run when its of type 1 and others when its of type 2
12:17 babilen But you could easily solve this in targeting
12:17 babilen Not too fond of "{% if data['id'] == 'mysql1' %}" logic in states
12:17 thealphaking01 no no, this is in the reactor
12:18 rai_ joined #salt
12:18 thealphaking01 i have two kinds of minions that might join. i can write a custom python function, pass the id as an argument, store it in a file, take the data from there and then run the command based on the data, but that would be a very long and unneccessary process
12:19 babilen Does it have to match a regex?
12:19 babilen Or would it suffice if it simply starts with 'mysql' ?
12:19 thealphaking01 i want the reactor to handle two different kind of minions separately. note that i dont have only two, i have multiple, but they are of two kinds
12:19 thealphaking01 suffice if it simply starts
12:19 babilen Use .startswith('mysql') then
12:20 thealphaking01 any kind of basic matching without writing anything custom
12:20 babilen You asked specifically for regex for which you need the re module in Python
12:20 thealphaking01 data["id"].startswith('mysql') works?
12:21 thealphaking01 my apologies. anything more intelligent than a simple equal to should be sufficient for my cause.
12:26 hemebond joined #salt
12:30 keimlink joined #salt
12:35 netcho_ joined #salt
12:39 Reverend any reason my master log is spamming me with minionfs warnings? :/
12:39 Reverend 2016-12-14 12:38:26,118 [salt.loaded.int.fileserver.minionfs][WARNING ][29922] minionfs: could not find files directory under /var/cache/salt/master/minions/XXX.XXX.blerp
12:51 TomJepp joined #salt
12:54 rai_ joined #salt
12:58 Reverend Hi TomJepp
13:05 ciapatti joined #salt
13:09 numkem joined #salt
13:09 _Cyclone_ joined #salt
13:11 DammitJim joined #salt
13:25 impi joined #salt
13:36 hemebond joined #salt
13:44 sh123124213 joined #salt
13:45 johnkeates joined #salt
13:51 amontalban joined #salt
13:59 brontide joined #salt
14:04 thealphaking01 joined #salt
14:08 thealphaking01 anyway to ensure that a minion always has a unique id?
14:14 edrocks joined #salt
14:14 felskrone thealphaking01: define it in the minion config so the minion does not eval it by itself
14:16 babilen thealphaking01: Why wouldn't they have a unique id?
14:16 babilen (you can't have two minions with the same id)
14:17 thealphaking01 two minions on different servers cant have same id?
14:18 AndreasLutro you definitely can
14:19 pipps joined #salt
14:19 thealphaking01 lets say i want the id to start with a prefix. basically my motive is to have two group of minions
14:20 babilen AndreasLutro: The minion keys wouldn't match in that case
14:20 pipps joined #salt
14:20 thealphaking01 i have hardcoded the id right now, say two groups: web and app. a minion can have either one of these id's
14:20 netcho_ joined #salt
14:21 babilen You'd have to make those IDs unique (web1, web2, ... for example)
14:21 thealphaking01 yes, how can i do that?
14:22 babilen You could give them unique hostnames or simply set their minion ID to something unique
14:22 babilen (in the minion configuration)
14:22 gtmanfred Reverend: I know you asked this a while ago, but you can edit sls files in the enterprise ui, also in suse manager which is uses salt to do its stuff
14:22 thealphaking01 in the minion config, i can specify the id. but all minions will use the same config
14:23 babilen thealphaking01: Well, they can't use the same configuration (at least not for the id)
14:23 babilen What's the hostname of your minions at the moment?
14:24 babilen Would it, perhaps, make more sense to "tag" them with their roles in either pillars or grains rather than relying on the minion id?
14:24 thealphaking01 meaning, i have one config file. i am passing a parameter from pillar, which is then used to set the config for that particular minion. each minion is on a different VM
14:24 babilen How do you bootstrap your boxes? Can *that* system handle unique naming of boxes?
14:25 Reverend gtmanfred oh okay - that's pretty neat. Thanks dude.
14:25 thealphaking01 ohk, lets say i tag them. but when a minion comes up, will the master know what tag this minion is?
14:26 babilen How do you bootstrap your minions?
14:27 thealphaking01 that wont be taken care by me. probably a load balancer etc will come into play
14:27 babilen So who installs salt on the boxes and choses the hostname?
14:28 WKNiGHT joined #salt
14:29 thealphaking01 i got an idea, though. reactor can see what tag's are present in the fired event. so, if i can send that particular tag, that works. no need to worry about id. how to send a particular tag when the minion gets created so that it is fired onto the master
14:29 gtmanfred are you creating it via salt cloud?
14:29 thealphaking01 all of that process will be abstracted. i think thats also a way we can solve this, but i am trying to keep the control in salt as much as i want.
14:29 thealphaking01 nope
14:30 Trauma joined #salt
14:30 gtmanfred then you will need to use salt.utils.events https://docs.saltstack.com/en/2015.8/topics/event/index.html
14:30 Flying_Panda hi guys is there anyway to run a if statemanet against the output of a salt module ?>
14:30 Flying_Panda in a state
14:31 gtmanfred Flying_Panda: sure? {% if something in salt['module.function'](args) %}
14:31 gtmanfred or salt.module.function
14:31 gtmanfred Flying_Panda: https://docs.saltstack.com/en/latest/topics/tutorials/states_pt3.html
14:31 Flying_Panda might be me being silly using ( instead of [
14:32 gtmanfred you can also use salt.module.function and get rid of the brackets and the quote
14:32 Flying_Panda {% if salt[system.get_pending_reboot'] == True %}
14:32 Flying_Panda for example
14:32 gtmanfred salt.system.get_pending_reboot
14:32 DammitJim joined #salt
14:32 gtmanfred and add ()
14:32 gtmanfred {% if salt.system.get_pending_reboot() is True %}
14:33 Flying_Panda no [?
14:33 gtmanfred you have to call the function, otherwise you are asking if the function definition is true
14:33 gtmanfred yeah, you don't need [, the dot notation is equivalent in jinja
14:33 Flying_Panda ahhhhhh
14:33 gtmanfred salt is a dictionary of functions, and in jinja on dictionaries, you can just call it via the dot notation
14:34 Flying_Panda perfect thank you as always GT <3
14:34 gtmanfred but you do need the () after the function, whether you use salt[''] or the dot notation,
14:34 ALLmightySPIFF joined #salt
14:36 ALLmightySPIFF joined #salt
14:37 ivanjaros joined #salt
14:37 thealphaking01 cant find how to add a tag in the default fired event when a minion connects to a master
14:39 gtmanfred you cannot add that
14:39 gtmanfred the event is not configurable
14:39 gtmanfred that event*
14:39 gtmanfred you would have to fire an event that does a look up of the grains on the minion and fires another event with that info
14:40 thealphaking01 and this fired event will be captured by the master?
14:40 gtmanfred yes?
14:40 gtmanfred https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.event.html#salt.modules.event.send
14:41 gtmanfred have the connecting minion with a startup_states, that runs an sls, that fires an event with the roles you want it to fire
14:41 nickabbey joined #salt
14:41 mpanetta joined #salt
14:42 babilen thealphaking01: It *really* boils down to the way you bootstrap your minions. If you have absolutely no control nor knowledge about that  process it is very hard to help you or tackle this problem
14:42 gtmanfred ^^
14:43 Tanta joined #salt
14:44 mpanetta joined #salt
14:51 nZac joined #salt
14:54 armin joined #salt
15:03 nZac_ joined #salt
15:06 cscf joined #salt
15:08 jmickle joined #salt
15:08 impi joined #salt
15:15 gladia2r joined #salt
15:17 nickabbey joined #salt
15:18 racooper joined #salt
15:25 jmickle joined #salt
15:27 gableroux joined #salt
15:32 Flying_Panda with the BIGip salt states is there anyway to set the partition ? looks like it defaults to common with no way to change
15:38 anotherzero joined #salt
15:41 drawsmcgraw joined #salt
15:46 sh123124213 joined #salt
15:51 keltim joined #salt
15:56 anotherzero joined #salt
15:57 sarcasticadmin joined #salt
16:01 prpl joined #salt
16:01 Flying_Panda figured it
16:01 Brew joined #salt
16:01 prpl left #salt
16:01 sh123124213 joined #salt
16:01 user951 joined #salt
16:02 djgerm BigIP salt states? You're living my dream Flying_Panda!
16:02 fracklen joined #salt
16:03 Flying_Panda yup
16:03 Flying_Panda and if your in a different partition common
16:04 gtmanfred Flying_Panda: maybe? you could also use the cisconso proxy minion to control it
16:04 gtmanfred actually
16:04 gtmanfred I know the answer to this one
16:04 Flying_Panda just name: ~partition_{{name}}
16:04 Flying_Panda just name: ~partition~{{name}}
16:04 Flying_Panda and it will change the partition from common
16:04 gtmanfred Flying_Panda: right now it is not possible to set the partition for all calls in there, though we do have a customer request for it, so it will be in at some point
16:05 Flying_Panda got it working gt :D
16:05 gtmanfred yeah, we had that working, but there were still some problems with it
16:06 promorphus joined #salt
16:13 nZac joined #salt
16:15 _JZ_ joined #salt
16:18 user951 I've got a few Windows servers that each run 10-30 of our applications on them, set up in different directories with separate configurations and possibly separate versions. Would the best way to set this up be to run a separate minion on the server for every application?
16:18 teclator joined #salt
16:18 pipps joined #salt
16:18 sidx joined #salt
16:19 pipps joined #salt
16:19 drawsmcgraw user951: Separate Salt minions on the same box? I don't think you'd gain anything from that.
16:20 drawsmcgraw I'd rather look into managing the different versions/locations of the apps via some templating + Pillar data.
16:20 sidx Hi, I've successfully created minions. While doing a --query, I'm getting the details of minions. But when I run a salt '*' test.ping, I'm getting Minion did not return. [No response]
16:20 sidx Please help
16:21 user951_ joined #salt
16:23 user951_ Disconnected. I was thinking of trying some template stuff as well. Ideally, I'd like to be able to update applications by storing the configuration in a pillar and just doing 'salt 'app01' state.apply application.update'.
16:25 user951_ So I was thinking if I ran a minion for every application, it would let me target them like that.
16:25 thealphaking01 joined #salt
16:26 thealphaking01 event triggered by the minion will hit the master, right?
16:27 user951_ thealphaking01: not sure what you mean by that
16:27 sh123124213 joined #salt
16:31 amontalban joined #salt
16:31 fracklen joined #salt
16:31 user951_ Currently using Chef, and I'm targeting the individual applications with 'knife winrm "name:server1" "chef-client -o recipe[application:update] -E app01"', storing the application's configuration in an environment. It's a bit cumbersome though.
16:32 eprice joined #salt
16:37 sergei joined #salt
16:39 teclator joined #salt
16:47 Sayne joined #salt
16:49 Sayne Hello, I have a quick question regarding external_auth I'd like authorize jenkins pam use to be able to execute just one function myapp.deploy in every Salt environment
16:49 Sayne I tried many combination without success
16:50 djgerm does anybody know why the region grain for new EC2 instances would be blank?
16:52 djgerm sorry, let me TOTALYL rephrase that
16:52 djgerm Is there a good way to populate a grain that represents an instances region/AZ in EC2?
16:53 Sayne It's easy yes
16:54 codeape joined #salt
16:54 Sayne You can add any custom grains
16:54 djgerm heh.. yeah, but… i want it automatically :)
16:56 tercenya joined #salt
16:57 Sayne You can use the ec2 instance metatdata
16:57 Sayne to populate the grains at startup
16:58 djgerm ah…. yes… that makes sense. the instance does know somewhere that it's in us-west-1a for example
17:00 Sayne yes
17:00 Sayne run that on your ec2 instance: curl http://169.254.169.254/latest/meta-data/placement/availability-zone
17:01 djgerm WHOA! wizard
17:01 Sayne You should read about it, it's very powerful to retreive a lot of information :) http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ec2-instance-metadata.html#instancedata-data-retrieval
17:03 drawsmcgraw user951_: I can start to see why you'd want a separate minion for each application. But try thinking along the lines of data/role separation (instead of a minion->app mapping). Instead of separating the minions out, write the Salt states and organize Pillar data such that you can deploy/update any app anywhere. Then the specific minion targeting becomes unnecessary.
17:04 nicksloan joined #salt
17:05 djgerm that's good advice, helps you get dev/qa to deploy test environments with salt
17:07 user951_ drawsmcgraw: thanks, I'll give it a shot
17:08 drawsmcgraw user951_:  Yep! I know it's much easier said than done. Good luck!
17:09 lws joined #salt
17:09 amontalb1n joined #salt
17:11 pipps joined #salt
17:15 tduerr joined #salt
17:15 promorphus joined #salt
17:18 djhaskin987 joined #salt
17:21 racooper Howdy. I'm working on a set of apache.configfile states and they are failing.  I know the sls file is being read, because the user.present in the same state file is being processed.  the file is https://gist.github.com/racooper/817533bd9ec1548c30d4f4f3eebb482a
17:21 woodtablet joined #salt
17:22 woodtablet joined #salt
17:22 racooper and it's acting like the apache config is already present, but the file doesn't exist and isn't being created.
17:23 pipps99 joined #salt
17:24 racooper granted, this is an older version of Salt, 2015.5.10, and I don't know if that might have anything to do with it. I can't upgrade them though until salt centos/RHEL repo doesn't overwrite system packages.
17:31 promorphus joined #salt
17:35 Lionel_Debroux joined #salt
17:36 bemehow joined #salt
17:38 Salander27 joined #salt
17:41 codeape joined #salt
17:43 nickabbey joined #salt
17:46 debian112 joined #salt
17:49 pipps joined #salt
17:57 Sayne I can't make external_auth works with state.sls anyone have a working example? I always get 401 Unauthorized
18:00 nZac joined #salt
18:07 teclator joined #salt
18:07 drawsmcgraw joined #salt
18:08 onlyanegg joined #salt
18:09 Edgan joined #salt
18:14 theblazehen71 joined #salt
18:18 ajw0100 joined #salt
18:19 seanz joined #salt
18:21 one joined #salt
18:22 one joined #salt
18:22 samodid joined #salt
18:22 beardedeagle joined #salt
18:27 irctc403 joined #salt
18:27 johnkeates joined #salt
18:28 irctc403 hey got a question, maybe someone will know... in the pillar top file is it possible to target minions based on an ext_pillar?
18:28 nidr0x joined #salt
18:28 irctc403 in my case, I have ec2_tags, and would like to match based on tags
18:28 promorphus joined #salt
18:29 sarcasticadmin joined #salt
18:31 arif-ali joined #salt
18:33 Sketch hmm, service.dead fails when the service's unit file is not installed? that's annoying
18:36 Praematura joined #salt
18:37 iggy irctc403: if you set the master to process ext_pillar first
18:37 iggy but I wouldn't ever count on that functionality
18:38 irctc403 iggy: Via orchestration ?
18:38 iggy I don't understand the question
18:38 irctc403 how do you set precedence?
18:39 iggy it's a master config setting
18:39 irctc403 I see.
18:39 irctc403 I didn't realize order mattered
18:39 irctc403 I'll give it a shot
18:40 iggy I mean... you can't target based on data that hasn't been read yet
18:40 tercenya joined #salt
18:41 irctc403 yeah its a chicken and egg issue.. I guess I'll have to encrypt it and make it available to the whole environment
18:41 irctc403 I have dynamic instance ids, so I can't really target by minion id
18:41 iggy and I suppose you don't want to trust grains
18:42 irctc403 yeah
18:42 irctc403 thanks
18:43 iggy https://docs.saltstack.com/en/latest/ref/configuration/master.html#ext-pillar-first
18:43 iggy that's the only way I can think of... if that doesn't work, you're kind of boned I guess
18:57 Praematura_ joined #salt
18:57 promorphus joined #salt
19:00 nZac joined #salt
19:01 Flying_Panda can nodegroups just be a list of servers?
19:01 nicksloan joined #salt
19:03 sp0097 joined #salt
19:04 cscf Flying_Panda, yes
19:04 cscf https://docs.saltstack.com/en/latest/topics/targeting/compound.html
19:05 cscf Flying_Panda, Oh, even better: https://docs.saltstack.com/en/latest/topics/targeting/nodegroups.html#defining-nodegroups-as-lists-of-minion-ids
19:06 Flying_Panda when these are updated guessing just need to restart master and off we go
19:07 edrocks joined #salt
19:10 cscf Yep.  Btw, I usually put nodegroups separate in master.d/nodegroups.conf
19:10 raspado joined #salt
19:10 Flying_Panda yea seemed like the clever way to do it
19:12 tercenya joined #salt
19:13 KajiMaster joined #salt
19:18 stupidnic joined #salt
19:18 aw110f joined #salt
19:18 nZac joined #salt
19:21 ponyofdeath anyone know why apache apache.config apache.modules gets prioratized at the top of the salt highstate run even tho i have it much lower on the includes: https://bpaste.net/show/8e92218e6686
19:28 TheDeathRay joined #salt
19:36 jas02 joined #salt
19:37 jas02 joined #salt
19:40 iggy once you start putting in any requisite statements, the ordering goes to hell
19:40 iggy that said, how are you determining what the order is?
19:41 debian112 joined #salt
19:44 jor joined #salt
19:45 MTecknology I've made it a policy to *NEVER* trust top-down processing of sls files. If this state needs that state, it's explicitly defined.
19:46 tercenya joined #salt
19:46 MTecknology I've heard others argue that you should only ever depend on an sls and never on a state and only trust the top-down approach because that's all you need and anything else makes it complicated.
19:48 ivanjaros joined #salt
19:49 nicksloan joined #salt
19:51 lws joined #salt
19:52 cyborg-one joined #salt
19:54 promorphus joined #salt
19:56 jas02 joined #salt
19:56 cscf I can see an argument for using normal depends inside a file, and only sls depends on other files, that way you can refactor individual files all you like.
20:05 lws joined #salt
20:06 fracklen joined #salt
20:07 MTecknology cscf: Did you present at saltconf15?
20:08 cscf MTecknology, lol no
20:08 cscf Haven't gotten to attend one yet, in fact.
20:12 jas02 joined #salt
20:14 fracklen_ joined #salt
20:20 anotherzero joined #salt
20:20 Trauma joined #salt
20:22 jas02 joined #salt
20:25 jas02 joined #salt
20:26 codeape joined #salt
20:26 Aikar Hey, we are trying to use git.latest with a tag, we have the config like this, https://gist.github.com/aikar/cea9de0f54ce49a7588498e359e33f4c  - it shows no changes, though the tag points to a diff rev than where we currently are
20:27 Aikar is there anything we need to do different? 2016.11
20:29 Aikar ohh... server state shows "head detached at stg-current"
20:30 Aikar it seems something else is funny going on
20:30 Aikar it does appear its in the correct state, just did not run a post pull task
20:34 cilkay Hi. I'm upgrading from 2015.5.3 to 2016.11.0 in a masterless minion. In one of my states, I have {% set remote_version = salt['postgres.psql_query'](user=util_user, password=util_pw, port=pg_port, host=remote_host, maintenance_db=util_db, query=get_version) %}, which works fine in the older release. In the newer one, I get: Jinja variable 'salt.utils.templates.AliasedLoader object' has no attribute 'postgres.psql_query'
20:34 cilkay Any idea how to work around that?
20:35 tercenya joined #salt
20:38 frew joined #salt
20:48 jas02 joined #salt
20:51 SaucyElf joined #salt
20:53 SaucyElf joined #salt
20:53 iggy you've done the normal stuff right? (make sure you have deps installed and accessible, etc)
20:58 cilkay iggy: Yes. I can remove the new salt-common and salt-minion packages, install the old ones, and the state will work fine. All the other states work with the new packages. I'm using Debian so dependencies are handled via the package manager.
20:58 iggy not all dependencies
20:58 iggy but okay
20:59 iggy there's probably something extra that 2016.11 is looking for
20:59 cilkay Oh? What could/would be missing?
20:59 iggy sadly, you'll probably have to check the code
20:59 cilkay It's odd that every other one of the 80+ states works with 2016.11 but this particular line gives it heartburn.
21:00 cilkay I didn't see anything in the change notes.
21:06 cyteen joined #salt
21:08 bemehow joined #salt
21:08 jas02 joined #salt
21:08 edrocks joined #salt
21:14 cilkay It looks like this has something to do with cmd.run vs cmd.shell.
21:24 Sketch i...don't think there is a cmd.shell
21:25 cscf I don't see one in the docs.  And I'm not sure what it would do
21:26 cscf There's a - shell: option to cmd.run, though
21:26 johnkeates joined #salt
21:27 onlyanegg joined #salt
21:28 codeape joined #salt
21:28 rai_ joined #salt
21:34 buu Guys how do I get minion version from minions
21:35 buu Also better question
21:35 onlyanegg joined #salt
21:35 whytewolf salt-call --versions
21:36 whytewolf or salt 'minion' test.versions
21:36 buu I just ran sudo salt \* test.ping on the master
21:37 nixjdm joined #salt
21:37 buu The three machines on my lan all responded with alive
21:37 buu But the two machines outside my lan did not
21:37 buu Minion did not return. [No response]
21:37 buu oh
21:37 buu I run it again and it responds true?
21:37 buu Is this a salt master timeout thing?
21:38 whytewolf sounds like those minions have issues connecting to the master. most likely issues on a network level
21:40 buu Why do they respond fine the second time though?
21:40 buu Or are you trying to say there was a temporary network problem
21:41 whytewolf I'm saying intermittent. maybe packet loss.
21:41 whytewolf or a switch losing arp routes
21:41 sp0097_ joined #salt
21:41 whytewolf or they are compeating on a busy network
21:42 buu I can run it 100 times in a row now and get success every time..
21:42 buu And the first timme I run it after waiting 8 hours I get failure
21:42 buu It's more consistent than that.
21:45 whytewolf anything on the minions about them having problems connecting after that 8 hours?
21:45 buu It feels like the connection goes to sleep and the first command takes a long time to wake them up
21:46 buu Wait how do salt minions even work
21:46 whytewolf minions connect to the master
21:46 buu Oh, right
21:47 buu So they don't have any listening sockets right?
21:47 whytewolf no, they don't
21:48 buu right
21:48 saltnewb joined #salt
21:49 buu Anyways, no,, the last message in log/salt/minion is from like 18 hours ago from some debugging calls I made
21:49 debian112 joined #salt
21:49 buu 'test.versions' is not available. good sign
21:49 saltnewb Hello, I'm trying to use Salt to manage Jenkins plugins with this Package: https://github.com/tcpcloud/salt-formula-jenkins I'm struggling figuring out how to call the plugins information correctly in the states files. Would anyone be able to help give me some direction?
21:50 whytewolf buu: what version are you on? and you could also try updating your zmq there were issues on older versions like you are describing
21:51 buu whytewolf: where does zmq live?
21:51 whytewolf buu: everywhere
21:51 buu =O
21:51 buu Pretty sure these minions are on 2014-x-x
21:51 whytewolf ouch. that actually could be the problem then..
21:51 buu yeeeeeeeah
21:51 buu Thanks debian
21:52 whytewolf well look at the bright side. back then there was a issue with teh way that ubuntu had the packages. and if you didn't install some kind of repo you could en up with 0.17.5
21:52 nickabbey joined #salt
21:54 pipps joined #salt
21:56 buu Nifty.
21:56 buu Installing the backported version of 2016 now
21:57 buu 2016.3.4 yay
21:57 fracklen joined #salt
21:58 buu The other machine was actually on debian 6.4 or so
21:58 DammitJim joined #salt
22:00 whytewolf squeeze? wow that is EOL. not even LTS anymore
22:00 whytewolf or am i thinking of something else
22:01 buu Nope, it was released in 2014, lts until feb-2016
22:01 buu This was a surprise to me also.
22:02 whytewolf oh well lots of shops still run RH6.2
22:03 TimHarrison joined #salt
22:03 pipps joined #salt
22:04 TimHarrison Hello #Salt. I had some questions about the capability of Salt OSS
22:04 whytewolf i also know a shop or two that is running a redhat 5 box or two and maybe has a snuck away redhat 4
22:05 TimHarrison I'm looking for a solution to remotely manage client PCs that will be running Ubuntu 16.04 LTS. These will be employee laptops at multiple internet locations.
22:05 TimHarrison Would Salt OSS be an appropriate solution to silently push software to these client machines/
22:05 buu whytewolf: The worst part is "old" redhat comes with perl 5.8 which is like.. 16 years old now
22:06 whytewolf 5.8 was the standard for so long ... coarse so was 5.6 for a long time befor ethat
22:07 buu "standard'
22:07 buu It was like 5 years old when they release that rh version
22:07 buu eesh
22:07 buu Then they released a broken version
22:07 buu TimHarrison: Have you considered trained assault monkeys?
22:08 TimHarrison yes however, they have proven too careless with their handling of weaponry. Our insurance made us let them go.
22:08 buu oh
22:09 LotR buu: are you the same buu as the one on MAGnet?
22:09 TimHarrison So now I guess I'm stuck looking for software solution
22:09 buu probably
22:09 jas02 joined #salt
22:09 stooj joined #salt
22:09 whytewolf TimHarrison: in thoery. but it kind of gets tricky with machines that are not always connected to your network.
22:09 buu TimHarrison: My impression so far is salt is good at applying relatively static scripts to everything and low level cli bashing to individuals
22:10 buu But it doesn't have much in the way of reporting
22:10 buu LotR: I don't think I've been on magnet for a little while, is there an impostor?!
22:11 LotR buu: no
22:11 LotR buu: it's been quite a while since I've seen your nick, so seeing it here made me curious :)
22:12 buu That's because you abandoned #perl !
22:12 LotR I did that years before you left
22:13 buu =[
22:13 buu Traitor.
22:14 LotR I'm loyal to the language, not the subset of the community that's in #perl. if that makes me a traitor, so be it :)
22:15 buu LotR: I meant freenode
22:16 LotR oh. I've never been in freenode #perl
22:16 buu ?!
22:16 buu How did you manage that.
22:16 LotR well, not as a regular. I might have popped in once or twice
22:16 LotR I found magnet #perl first?
22:16 buu I'm sorry.
22:17 LotR it was ok when I started :)
22:17 buu eh
22:20 ecdhe I'm running masterless salt on vagrant boxes.  My /etc/salt/minion file says "file_client: local".  Everything is working as I expect, all my states are applied, but when I look in the logs it appears salt is at some point timeing out trying to find a master in "multiple-master" mode.
22:21 ecdhe Is there any way to stop this behavior?
22:21 stooj joined #salt
22:22 TimHarrison Thanks for the help. That's what I figured. It looks more like server infrastructure management rather than client PC management.
22:33 stooj joined #salt
22:38 codeape joined #salt
22:45 onlyanegg joined #salt
22:46 sp0097 joined #salt
22:48 saltnewb I'm struggling calling this as a formula in salt local mode. Anyone that could give me some pointers? https://github.com/tcpcloud/salt-formula-jenkins#id3
22:59 drew__ joined #salt
23:00 Klas joined #salt
23:03 drew__ Hi I am trying to write a Runner module that calls another Runner module. I am running into the following error:
23:03 drew__ Exception occurred in runner test_runner.run: Traceback (most recent call last):   File "/usr/lib/python2.7/site-packages/salt/client/mixins.py", line 399, in _low     data['return'] = self.functions[fun](*args, **kwargs)   File "/var/cache/salt/master/extmods/runners/test_runner.py", line 8, in run     return runner.cmd("manage.up")   File "/usr/lib/python2.7/site-packages/salt/runner.py", line 144, in cmd     full_return)
23:03 drew__ File "/usr/lib/python2.7/site-packages/salt/client/mixins.py", line 225, in cmd     self.functions[fun], arglist, pub_data   File "/usr/lib/python2.7/site-packages/salt/runner.py", line 50, in functions     self.utils = salt.loader.utils(self.opts)   File "/usr/lib/python2.7/site-packages/salt/loader.py", line 326, in utils     _module_dirs(opts, 'utils', ext_type_dirs='utils_dirs'),   File "/usr/lib/python2.7/site-packages/salt/load
23:04 drew__ line 133, in _module_dirs     ext_types = os.path.join(opts['extension_modules'], ext_type) TypeError: string indices must be integers, not str
23:06 drew__ my runner module is just simply calling manage.up and it still throws the same errors
23:08 drew__ actually never mind... I used the wrong opt for the RunnerClient args -_-
23:09 pipps joined #salt
23:22 jas02 joined #salt
23:25 Praematura joined #salt
23:31 debian112 joined #salt
23:34 bluenemo joined #salt
23:44 cyborg-one joined #salt
23:52 dxiri joined #salt
23:57 pipps joined #salt

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