Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2016-08-02

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

All times shown according to UTC.

Time Nick Message
00:05 pipps joined #salt
00:07 voxxit joined #salt
00:07 flowstate joined #salt
00:12 fannet joined #salt
00:15 MTecknology This is gonna be fun... I have a state that /should/ run on every system but has only run on about 20 of them. I need to figure out which 20 it ran on and then force only them to re-run those states
00:15 MTecknology heheh... fun
00:15 ajw0100 joined #salt
00:16 jesusaur joined #salt
00:16 hasues joined #salt
00:20 MTecknology which is a funny chicken/egg thing because, without salt, I would probably have already given up for the day. Instead... no problem, just gonna take a lot longer.
00:21 hasues left #salt
00:21 ZachLanich iggy: I have a question regarding targeting. You around?
00:21 timoguin joined #salt
00:22 iggy sure
00:27 bluepaladin joined #salt
00:28 ZachLanich I'm trying to figure out how to set up my pillar files, etc for different types of data regarding the setup I explained to you last night. On one hand, I have more straight forward data like API Keys, Plugin License Keys, etc that can be distributed to almost all my webservers, but I have other data like Site-specific data, etc to send down the pipe. I know you can target by minion id, grains, etc, but I'm not 100% sure what the
00:28 ZachLanich best practice would be for targeting in this scenario. One option would be to have a minion id naming convention like so: <purpose>-<environment>-<cms>-<numeric_id>-<website_domain> (cms = content mgmt system like Wordpress or Drupal) - Ex. web-prod-wp-1-somesite_com. Is that a normal thing to do? Or should I be using some other method to define properties of my minions and target them. I know you set custom grains on the CL, but
00:28 ZachLanich it didn't look like that was a good thing to do.
00:30 onlyanegg joined #salt
00:30 ZachLanich iggy: Then again, it shows in the docs creating custom grains in the minion config, so would it make sense to create custom grains using pillar files or something? I'm not 100% clear on what's taboo for grains.
00:30 iggy I would try to avoid targeting pillar data (that could be sensitive) based on grains
00:31 iggy since minions could fake grains data to get access to stuff they shouldn't be able to
00:33 ZachLanich Could they fake custom grains any easier than they can fake pillar data? Or is pillar data only stored on the master (seemed to be the case)?
00:33 scsinutz joined #salt
00:34 iggy that is the case
00:34 iggy the master renders the pillar data and only gives the data to each minion
00:36 racooper joined #salt
00:36 ZachLanich Here's an example of what made sense to be for setting custom grains though (non-sensitive stuff): https://docs.saltstack.com/en/latest/topics/grains/#grains-in-the-minion-config - Would you recommend doing something like that for any reason or have you? Wouldn't you have to set those grains on a per provision basis rather than a config file on the master? I might be missing something.
00:37 ZachLanich Instead of using a bulky naming convention, wouldn't it make sense to set grains like purpose: webserver, environment: prod, etc as grains?
00:37 iggy when I've heavily used grains before it's been via some cloud provider metadata (i.e. a custom grain module that queried aws tags, or gce metadata, or local CMDB)
00:37 ZachLanich And can you do that in your cloud profiles or something?
00:38 ZachLanich ^ That makes sense too.
00:39 iggy you might also check out things like reclass, etc
00:40 scsinutz1 joined #salt
00:45 ZachLanich Interesting. That seems like a valid replacement candidate for what I had in mind when moving to a GUI to generate my config files. Am I on the right track here?
00:46 iggy https://github.com/digitalocean/netbox with an ext_pillar that talked to it maybe?
00:47 orion joined #salt
00:47 orion Hi. When will there be an official package for Xenial?
00:47 mohae_ joined #salt
00:48 iggy there already are?
00:50 orion Ah, so there is. Thank you.
00:50 ZachLanich iggy: Seems like it might be overkill since my interface doesn't have to care too much about racks, networking equip, etc
00:50 orion left #salt
00:50 Nahual joined #salt
00:50 jesusaur joined #salt
00:51 iggy ZachLanich: to each his own
00:51 iggy their own
00:51 ZachLanich Might prove to be useful way down the road when my infrastructure gets more complicated thogh
01:00 jesusaur joined #salt
01:01 bltmiller joined #salt
01:03 krymzon joined #salt
01:08 flowstate joined #salt
01:10 schemanic joined #salt
01:10 scsinutz joined #salt
01:11 ZachLanich iggy: reclass seems like it might help a lot, but the docs aren't super complete, so it's gonna take ma while to wrap my head around where it starts and ends in relation to salt and finding examples.
01:12 schemanic Hi. Im setting up a cloud provider for ec2 and I'm trying to understand what the key settings mean. It says to make sure the key is owned by root with permissions at 0400. What is this key? is this a pem key generated by amazon, or am I making one right there?
01:13 schemanic Is salt cloud assuming that I'm supposed to create a keypair in aws specifically *for* my salt master?
01:14 Edgan whytewolf: I actually get the same error from salt master mode
01:14 iceyao joined #salt
01:14 onlyanegg joined #salt
01:15 Edgan whytewolf: nm, wrong directory name
01:15 jenastar joined #salt
01:16 iceyao_ joined #salt
01:17 PerilousApricot joined #salt
01:19 schemanic anyone here use salt-cloud?
01:19 nicksloan joined #salt
01:20 catpigger joined #salt
01:29 edrocks joined #salt
01:36 hemebond schemanic: I believe id and key are the role access key.
01:37 hemebond The access id+key allows salt-cloud to make the changes.
01:38 schemanic hemebond, i noticed that one, and I have those values. but the salt-cloud quickstart seems to indicate an ec2-keypair
01:38 schemanic I've generated one for now
01:39 hemebond Is an "ec2-keypair" different to an Access Key?
01:39 schemanic hemebond, Im sorry, I should have said 'ssh-key
01:39 hemebond Ah, that's to allow salt-cloud onto the instance in order to bootstrap it.
01:39 schemanic AWS will allow you to create a keypair, and I guess salt wants the master to own the key
01:40 schemanic Oh I see
01:40 schemanic then from there the communication is all along ZeroMQ?
01:40 hemebond Are you talking about the "private_key" and "keyname" parameters?
01:40 schemanic Yeah I am
01:40 hemebond Once the instance is bootstrapped it will be talking to the master and be acccessible like any other minion.
01:41 nicksloan joined #salt
01:41 hemebond Yeah, "private_key" and "keyname", in my provider config, are the SSH pem that the instance will be created with and salt-cloud uses it to SSH onto the instance.
01:42 schemanic hemebond thanks, that provides some context I wasn't picking up from the doc
01:46 hemebond It took me a bit of trial-and-error too I remember.
01:46 nicksloan forrest
01:46 nicksloan whoops
01:47 ilbot3 joined #salt
01:47 Topic for #salt is now Welcome to #salt! | Latest Versions: 2015.5.10, 2015.8.10, 2016.3.1 | Support: https://www.saltstack.com/support/ | Logs: http://irclog.perlgeek.de/salt/ | Paste: https://gist.github.com/ (please don't multiline paste into channel) | See also: #salt-devel, #salt-offtopic | Ask with patience as we are volunteers and may not have immediate answers
01:49 bbbryson joined #salt
01:51 Bryson joined #salt
01:53 iceyao joined #salt
01:59 scsinutz joined #salt
02:05 nicksloan joined #salt
02:06 mpanetta joined #salt
02:06 flowstate joined #salt
02:08 anotherZero joined #salt
02:08 badon_ joined #salt
02:12 spuder joined #salt
02:13 fannet joined #salt
02:25 evle joined #salt
02:38 Andrew joined #salt
02:39 ajw0100_ joined #salt
02:41 bastiand1 joined #salt
02:41 ZachLanich iggy or hemebond: I have a more specific task I'm working through right now and I'd like your advice plz :) - I have a staging server with multiple vhosts, each vhost representing a "Site", each "Site" having it's own production VPS. I have Wordpress being automatically installed and a plugin called WP Migrate DB installed automatically. This plugin is very slick and allows one-click push/pulls of databases between environments, and
02:41 ZachLanich I can configure and run this plugin using the command line through WP CLI. My issue is that each environment has its own generated "Secret Key" in the WP database. I'm already in the process of being able to configure this secret key via the command line, but the issue lies in me not knowing how to make the specific "Staging" environment/site/vhost aware of it's respective production VPS so I can pass these Secret Keys back and
02:41 ZachLanich forth and configure the environments to link to each other. How do I do this with Salt?
02:43 coval3nce joined #salt
02:51 ninjada joined #salt
02:52 hemebond ZachLanich: What about an event containing the secret key that a reactor then pushes to the other environment?
02:54 ZachLanich hemebond, that is definitely a possibility. How would I solve this scenario though: someone accidentally resets the secret key on one of the environments. I'd like my infrastructure to automatically be able to resync them in scenarios like that.
02:55 hemebond Uh. Pass.
02:59 mapu joined #salt
03:03 scsinutz joined #salt
03:03 ZachLanich lol
03:04 PerilousApricot joined #salt
03:07 nethershaw joined #salt
03:08 flowstate joined #salt
03:10 ZachLanich hemebond: So, excluding that last scenario, should I just use reactor to trigger a state that specifically inserts those secret keys?
03:10 hemebond I think it would work.
03:10 hemebond I haven't used reactors myself.
03:11 krymzon joined #salt
03:12 hemebond Ooh, there is an Elastic beat for Salt events. Neat.
03:14 tonybaloney joined #salt
03:14 tonybaloney got salt-minion working on an intel edison
03:14 tonybaloney sharing a video of it now..
03:15 tonybaloney pretty impressive, grains.items shows CPU as 500Mhz
03:16 flowstate joined #salt
03:17 hemebond Neat.
03:17 hemebond Although... Edison? Really, Intel?
03:18 tonybaloney I had one on my desk. so why not :-)
03:19 tonybaloney https://software.intel.com/en-us/iot/hardware/edison
03:20 tonybaloney http://www.youtube.com/watch?v=R4ShthNPpSQ
03:23 tonybaloney ok now its ready
03:31 edrocks joined #salt
03:38 rdas joined #salt
03:46 spuder joined #salt
03:50 XenophonF joined #salt
03:51 racooper joined #salt
03:56 racooper joined #salt
04:09 flowstate joined #salt
04:10 iggy ZachLanich: your original question was cutoff... normally that would be pillar data that you'd pass around, but I'm not sure how you'd get it from whatever you are running to generate it to pillar
04:11 tonybaloney joined #salt
04:11 fxhp joined #salt
04:14 fannet joined #salt
04:17 ZachLanich iggy Yea, that was the question lol. Interesting issue to solve.
04:17 Rebus joined #salt
04:17 ZachLanich iggy How much do we have available to us for minion -> minion communication?
04:18 hemebond ZachLanich: iggy means they're not sure how you generate or get that secret key.
04:18 rdas joined #salt
04:19 racooper joined #salt
04:20 ZachLanich It's automatically generated when the plugin is installed on the first highstate. I can retrieve it and retrieve/reset it with $ wp wcutils get-wpmdb-secret-key
04:20 ZachLanich So anytime after the plugin is installed and the WP install is fully functional, I can retrieve that key and do w/e with it
04:20 iggy but that value exists on certain minions only, right?
04:20 hemebond Oh, you could also use Salt mine to retrieve it; though that's not secure from other minions.
04:21 iggy i.e. it's created on the host that has the "prod" instance (or something)
04:21 ZachLanich Just not sure how to keep an ongoing synchronization wherein I periodically re-check the keys for a match and reset them if they don't match anymore
04:21 ZachLanich iggy: yes
04:21 iggy there's sdb
04:21 ZachLanich The secret key must be unique to each minion
04:21 iggy I've never used it
04:21 hemebond sdb?
04:22 JPT joined #salt
04:22 iggy I might consider putting that data outside of salt and just referring to it with an ext_pillar or some other method
04:23 ZachLanich iggy: There's a key on every WP install, so there's a secret on the slot on the staging server for the site, as well as a different key on the production site. Which one you use depends on whether you're creating a Push or a Pull configuration for DB migration.
04:23 iggy I feel like I've said all the buzzwords I can think of about something I really know nothing about
04:24 ZachLanich sdb might be a good contender
04:24 ZachLanich :P
04:24 scsinutz joined #salt
04:25 ZachLanich sdb:// would only be available in pillar and state files I'd imagine, so I don't see it being a security concern. Thoughts?
04:27 iggy haven't used it, not sure on it's specifics
04:29 sagerdearia joined #salt
04:31 ZachLanich iggy & hemebond: If you will, look this over and let me know if you see any reasons why this isn't a good start to my conf setup: https://gist.github.com/zlanich/1d759994751b0d73714caa7434ee6dcb
04:31 ZachLanich I'm looking to loop through all the domains, etc to create the vhosts on the relevant servers, configure domains wherever else needed, etc
04:32 ZachLanich_ joined #salt
04:32 onlyanegg joined #salt
04:33 ZachLanich_ :)
04:34 scsinutz joined #salt
04:34 racooper joined #salt
04:37 racooper joined #salt
04:39 josuebrunel joined #salt
04:39 iggy doesn't seem too bad... there's some bad indentation in the profiles
04:40 iggy and I never had luck with extends
04:40 WebDude joined #salt
04:42 badon_ joined #salt
04:42 WebDude I got disconnected, so if anyone answered, please repost.
04:43 WebDude joined #salt
04:44 iggy 00:40 < iggy> doesn't seem too bad... there's some bad indentation in the profiles
04:44 iggy 00:40 < iggy> and I never had luck with extends
04:49 WebDude I'm not worried about indentation. Might have been a bad paste, so I'll work the bugs out. Extend is very limited but seems to work for me so far.
04:50 WebDude I'll keep working on this and see where I get :)
05:00 POJO joined #salt
05:01 ajw0100 joined #salt
05:07 flowstate joined #salt
05:25 scsinutz joined #salt
05:31 impi joined #salt
05:33 edrocks joined #salt
05:56 zer0def joined #salt
05:58 ivanjaros joined #salt
05:59 felskrone joined #salt
06:01 kshlm joined #salt
06:02 alexhayes joined #salt
06:06 flowstate joined #salt
06:07 kshlm joined #salt
06:14 fannet joined #salt
06:15 bocaneri joined #salt
06:26 Miouge joined #salt
06:59 kshlm joined #salt
07:06 CeBe joined #salt
07:07 ninjada_ joined #salt
07:08 flowstate joined #salt
07:12 ivanjaros joined #salt
07:20 ninjada joined #salt
07:20 fracklen joined #salt
07:23 ravenx joined #salt
07:23 ribx joined #salt
07:34 CeBe1 joined #salt
07:36 edrocks joined #salt
07:36 josuebrunel joined #salt
07:42 fracklen joined #salt
07:45 fracklen joined #salt
07:50 antpa joined #salt
07:51 impi joined #salt
08:03 ivanjaros joined #salt
08:05 dh__ joined #salt
08:07 flowstate joined #salt
08:08 dh joined #salt
08:09 Deliant joined #salt
08:11 lero joined #salt
08:14 fixit26 joined #salt
08:15 fannet joined #salt
08:16 deniszh joined #salt
08:16 watersoul_ joined #salt
08:17 fixit26 joined #salt
08:18 IdoKaplan joined #salt
08:19 tcolvin joined #salt
08:19 blueelvis joined #salt
08:20 abele joined #salt
08:20 hacks joined #salt
08:20 Freek joined #salt
08:20 clouddale joined #salt
08:20 gazarsgo joined #salt
08:20 renoirb joined #salt
08:20 NightMonkey joined #salt
08:20 jeddi joined #salt
08:20 TOoSmOotH joined #salt
08:21 darvon_ joined #salt
08:24 brent_ joined #salt
08:25 IdoKaplan Hi, I'm using state.sls in order to execute a specific state. I cannot use state.sls if there is a sls require. Is there a way to ignore this require so I will be able to use state.sls even if there is a sls require?
08:25 fracklen joined #salt
08:26 babilen Don't think so .. you could use state.sls_id to execute a specific 'state' in that 'state'  ;)
08:26 armin just did a grep -r zeromq through my irc clients log file of this channel.
08:26 armin seems like "the zeromq version mismatch" is already an established static term.
08:26 fracklen IdoKaplan: You could use "state.sls foo.bar baz.bax" if foo.bar depends on baz.bax
08:31 Rumbles joined #salt
08:33 IdoKaplan fraclen: I'm getting this error - No matching salt environment for environment 'baz.bax' found
08:37 jhauser joined #salt
08:37 IdoKaplan fraclen: i am trying to execute a state that depands on another state, the sls called "spray" and the depand sls called javajdk, i tried to run the command that you meantion as following:
08:37 IdoKaplan "salt-call state.sls spray javajdk "
08:38 IdoKaplan and i got the error
08:38 IdoKaplan local:     Data failed to compile: ----------     No matching salt environment for environment 'javajdk' found ----------     No matching sls found for 'spray' in env 'javajdk'
08:38 IdoKaplan can you please advice?
08:39 s_kunk joined #salt
08:40 Mate 1. use include in the dependent state 2. state.sls expects a comma separated list of states
08:41 IdoKaplan Thanks the comma separate did the job :)
08:43 GreatSnoopy joined #salt
08:43 kbaikov joined #salt
08:45 jab416171 joined #salt
08:47 M-liberdiko joined #salt
08:47 ribx joined #salt
08:49 kbaikov joined #salt
08:50 alrayyes Hi there
08:50 alrayyes Are there any best practices for syncing large directories from master to minions?
08:51 alrayyes cp.get_dir takes a long time
09:04 jhauser joined #salt
09:05 babilen alrayyes: file.recurse would be more common
09:05 flowstate joined #salt
09:06 alrayyes babilen thanks!
09:07 Brijesh1 joined #salt
09:09 babilen https://docs.saltstack.com/en/latest/ref/states/all/salt.states.rsync.html maybe
09:10 babilen Hmm, that doesn't work between boxes
09:10 alrayyes yeah i already checked that
09:10 alrayyes i was surprised there wasn't a built in command
09:11 alrayyes as this is something that often needs to be done
09:11 alrayyes as a one time thing
09:12 babilen Well, file.recurse is that built-in command
09:13 alrayyes true, but i mean without having to first put the files in /srv/salt/blahblahblah/files
09:14 babilen You can put whatever you want in file_roots
09:15 babilen And all data pertaining to salt that should be available to the minions is expected to be in file_roots in a way
09:16 alrayyes well there is the salt-cp you can run from the command line
09:16 alrayyes it just needs recursive functionality
09:16 babilen I wouldn't consider that a proper way of using salt
09:17 alrayyes it isn't
09:17 babilen If you just want to run a command on the command line then use rsync
09:17 alrayyes but sometimes it does the trick
09:17 alrayyes for example now i'm moving a wordpress install to a new server
09:17 alrayyes only need to move the files once
09:18 alrayyes never going to really touch them again
09:18 alrayyes for situations like this a command is handy
09:19 alrayyes that's an option i suppose
09:20 alrayyes but then you need to run ssh
09:21 alrayyes suppose you could just run ssh server on internal provisioning network
09:21 necronian joined #salt
09:21 M-MadsRC joined #salt
09:22 Brijesh2 joined #salt
09:25 ange joined #salt
09:25 ange hello!
09:26 mortis is it me youre looking for?
09:26 ange mortis: 10 points for the good cultural reference
09:26 mortis lol
09:28 ange I remember vaguely reading in the doc that it was possible to define that reactors files should be pulled from within the states ; I am particularly interested to find a way to have them pulled from something kept in sync through gitfs; does that sound familiar to one of you ?
09:28 hasues joined #salt
09:28 babilen alrayyes: In a way such "one off" actions are better modelled with explicit cmd.run states .. you don't really want these to be "states", but rather "sequences of commands" in that you don't expect to run them over and over again
09:29 babilen alrayyes: The way this would work would be to configure SSH, keys and users and then run rsync .. That's all a one-off and you'd be done.
09:29 alrayyes ok
09:29 alrayyes so install and uninstall ssh as well?
09:29 ribx joined #salt
09:30 babilen All my boxes have SSH installed anyway
09:30 babilen But yeah, if you want it gone afterwards
09:31 babilen Just think about how you would do it without salt and then automate that
09:31 alrayyes Of course ssh is handy, but isn't the whole point of provisioning not having to do stuff manually anymore?
09:31 babilen But you want to do stuff manually
09:32 babilen The salt way would be to not care about the "old server" at all, but to simply describe the setup of the new server in toto so that it can be bootstrapped independently
09:32 alrayyes if you really need to login for whatever reason because something is really wrong there's the web console
09:32 fredvd joined #salt
09:33 antpa joined #salt
09:33 ange ah maybe simply through https://docs.saltstack.com/en/latest/topics/reactor/index.html#mapping-events-to-reactor-sls-files
09:33 babilen ange: Just keep them in salt://reactors/
09:35 infrmnt joined #salt
09:35 ange babilen: and by that you mean in my "gitfs" pulled repository in which I store the states ?
09:35 babilen ange: Exactly
09:36 babilen Or in a different repository in which you only configure the reactors
09:36 ange so a ls of that repo would look : . .. top.sls reactor/ ?
09:36 babilen Not even top.sls
09:36 babilen (you assign them in the master config)
09:36 ange babilen: just an example :p
09:37 ange babilen: ok, good point on the separate repo; any "good" practices on that ? are they usually stored in separate repo by most ?
09:37 Rumbles joined #salt
09:38 edrocks joined #salt
09:41 ange thanks!
09:41 ninjada joined #salt
09:42 Brijesh1 joined #salt
09:44 babilen ange: Well, the organisation is .. well .. not easy to generalise. It might make sense to keep them in a separate repository or together with a small set of states (the states that are being triggered by the events)
09:44 babilen You could organise it by service or customer or ...
09:49 mrBen2k2k2k joined #salt
09:51 ange babilen: ok, fair enough; thanks !
10:05 flowstate joined #salt
10:11 ninjada joined #salt
10:12 Ryndaxon joined #salt
10:15 Ryndaxon Hi. Does it have a specific meaning if a certain parameter is not described in the reference guide? See for example https://docs.saltstack.com/en/latest/ref/states/all/salt.states.virtualenv_mod.html. The purpose of system_site_packages is not explained.
10:16 fannet joined #salt
10:18 AndreasLutro Ryndaxon: check --system-site-packages for the virtualenv command
10:19 Ryndaxon I do know the meaning of the parameter. I just wonder if we should avoid using it since it's not in the reference guide. Perhaps it is subject to change or something else. Or am I reading into this too much?
10:19 antpa joined #salt
10:20 AndreasLutro the latter, salt often has undocumented arguments
10:20 AndreasLutro simply test if it does what you want/expect
10:21 AndreasLutro in this case, the state docs say "Also accepts any kwargs that the virtualenv module will."
10:21 AndreasLutro and you'll find system_site_packages mentioned here https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.virtualenv_mod.html
10:22 Ryndaxon Good point. Hadn't caught that. Thanks for clearing that up for me.
10:36 jhauser_ joined #salt
10:39 POJO joined #salt
10:43 fannet joined #salt
10:45 N-Mi joined #salt
10:45 N-Mi joined #salt
10:48 Pulp joined #salt
10:51 Brijesh1 joined #salt
10:52 flowstate joined #salt
10:54 Ryndaxon joined #salt
11:00 ingslovak joined #salt
11:03 adulteratedjedi joined #salt
11:06 CeBe1 joined #salt
11:13 blueelvis joined #salt
11:17 Ryndaxon joined #salt
11:21 amcorreia joined #salt
11:26 Ryndaxon left #salt
11:26 msn joined #salt
11:26 ninjada joined #salt
11:26 msn what's the command to execute only a single state component
11:35 B1nny joined #salt
11:35 AndreasLutro state.sls or state.single, depending on what you mean
11:39 nicksloan joined #salt
11:39 edrocks joined #salt
11:40 msn i got it thanks
11:41 msn i am tring to use saltstack-forumla mysql seems the debconf part is failing and root password is not being set at all
11:42 jhauser joined #salt
11:42 toanju joined #salt
11:44 msn this is the pillar I am using http://paste.debian.net/786541/
11:44 av_ joined #salt
11:45 oida joined #salt
11:47 msn with no changes to default mysql-formula
11:50 ravenx what is the difference between state.single and state.sls anyways
11:50 ravenx reading it doesn't seem to tell me muhc.
11:52 benjiale joined #salt
11:55 evle1 joined #salt
12:02 ravenx joined #salt
12:10 POJO joined #salt
12:10 antpa joined #salt
12:12 B1nny joined #salt
12:15 numkem joined #salt
12:15 B1nny joined #salt
12:18 B1nny left #salt
12:21 blueelvis joined #salt
12:24 antpa joined #salt
12:25 ravenx hi, i am using local.cmd from the python client api like so:  local.cmd(['one', 'two'], 'test.ping', expr_form='list')
12:26 ravenx and it returns to me something like:  {'two': True, 'one': True}   which is great and stuff.
12:26 ravenx i wish to output this in a highstate fashion:     salt.output.display_output(ret, "highstate", _opts_)
12:26 edrocks joined #salt
12:26 ravenx however, it only prints the first pair:  'two: True,   and completely ignores the 'one': True
12:26 whitenoise joined #salt
12:26 ravenx any ideas?
12:34 jhauser joined #salt
12:35 lovecraftian joined #salt
12:40 gh34 joined #salt
12:43 TooLmaN joined #salt
12:47 TooLmaN joined #salt
12:48 hasues left #salt
12:55 XenophonF ravenx: state.single allows you to run a single function from one of the salt state modules (salt.states.foo.bar())
12:55 XenophonF state.sls allows you to run all of the states listed in one sls file
12:56 goal joined #salt
12:56 XenophonF i'm not sure about your second question
12:57 goal is there a reason why the EL5 latest packages don't exist in the yum repo path used on repo.saltstack.com (yet exist inside /archive). eg. https://repo.saltstack.com/yum/redhat/5/i386/latest does not exist, but 2016.3.1 exists inside https://repo.saltstack.com/yum/redhat/5/i386/archive
12:58 TooLmaN joined #salt
12:59 ravenx aah i see
12:59 ravenx thanks
13:00 dyasny joined #salt
13:00 jhauser joined #salt
13:08 ravenx any ideas of the question i had early up above though?
13:09 ravenx that one is a real head scracther for me
13:10 Tyrm joined #salt
13:10 racooper joined #salt
13:13 ribx joined #salt
13:14 amcorreia joined #salt
13:14 CeBe1 joined #salt
13:15 jhauser joined #salt
13:17 XenophonF sorry ravenx which question?
13:17 ravenx i am using a local.cmd call:  local.cmd(['one', 'two'], 'test.ping', expr_form='list')
13:17 ravenx and it returns to me something like:  {'two': True, 'one': True}   which is great and stuff.
13:17 ravenx i wish to output this in a highstate fashion:     salt.output.display_output(ret, "highstate", _opts_)
13:18 ravenx however, it only prints the first pair:  'two: True,   and completely ignores the 'one': True
13:18 ravenx :/
13:18 netcho joined #salt
13:19 flowstate joined #salt
13:22 west575 joined #salt
13:29 deus_ex joined #salt
13:30 antpa joined #salt
13:31 LessSneaky joined #salt
13:34 deus_ex joined #salt
13:35 subsignal joined #salt
13:36 ze- if we have a minion & syndic running on the same host, can the minion connect to the local master, and the syndic connect upstream ?
13:37 ze- I have /etc/salt/minion with master: local-ip, and syndic_master: master-ip, but both connects to master-ip, and the minion would not connect localy anymore...
13:39 flowstate joined #salt
13:40 LessSneaky joined #salt
13:40 mapu joined #salt
13:42 protoz joined #salt
13:49 lovecraftian joined #salt
13:49 lovecraftian joined #salt
13:53 CTaylor_ joined #salt
13:55 benjiale joined #salt
13:57 iceyao joined #salt
13:57 schemanic joined #salt
14:00 PerilousApricot joined #salt
14:00 timoguin joined #salt
14:04 mpanetta joined #salt
14:06 netcho joined #salt
14:08 ravenx how can i use the python client api to get the returned job status
14:08 ravenx like, whether or not it has failed or pass.  or even an exit code woould be nice
14:10 gtmanfred you can run the runner to do a jobs.lookup_jid
14:10 gtmanfred or i think you can use the saltutil module to doo a find_job https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.saltutil.html#salt.modules.saltutil.find_job
14:11 gtmanfred from the job id returned from the first call
14:11 ravenx i dont thnk the local.cmd returns a jid though.
14:12 netcho joined #salt
14:12 hasues joined #salt
14:13 hasues left #salt
14:15 wangofett joined #salt
14:16 perfectsine joined #salt
14:16 jhauser joined #salt
14:16 timoguin joined #salt
14:18 gtmanfred ahh the local.cmd returns the whole output, what are you trying to run?
14:18 ravenx ultimately what i want is to be able to run a state.sls command for an app with local.cmd
14:18 ravenx then check if the job succeeded or not
14:19 ravenx origianlyl i wanted the jid for checking success, i dont know if there is another way to do it, besides parsing the big nasty salt highstate output
14:19 ZachLanich joined #salt
14:20 ravenx looking at this though:  https://docs.saltstack.com/en/2015.8/ref/clients/index.html#localclient   it looks as if i can pass in my own jid.
14:21 gtmanfred yeah, i think you can pass one in so that you can reference it later if you don't use local.cmd_async
14:21 gtmanfred i am not sure that there is a way to do it without parsing each one of the things and making sure the result is True
14:21 ravenx how will this differ if i used cmd_async?
14:22 anotherZero joined #salt
14:22 _JZ_ joined #salt
14:22 gtmanfred if you use cmd_async, it immediately returns a jid, which you would need to lookup
14:22 ravenx oh la la
14:22 gtmanfred and you would need to use a runnerclient to look it up with jobs.find_jid
14:22 ravenx hang on though....it returns the jid
14:23 gtmanfred but you could use saltutil.is_running with the localclient to see if that jid is still running
14:23 ravenx but where/when does it return the highstate output?
14:23 gtmanfred it never returns the highstate output to you, you have to look it up
14:23 ravenx o_o
14:23 ravenx is there a way to do that?
14:23 gtmanfred it sends the output to the master, and it is stored in /var/cache
14:23 gtmanfred yes, salt-run jobs.find_jid 'returnedjid'
14:24 gtmanfred run.cmd('jobs.lookup_jid', ['20160802141451086078'])
14:24 blueelvis joined #salt
14:24 gtmanfred that ^^ lookup_jid instead of find_jid
14:24 ravenx i see
14:24 ravenx but then in that case i will need to input a ticker to see when the job finishes
14:24 ravenx cuz right now it just exits from printing the jid to my screen.
14:25 gtmanfred correct, you would need to do that if you used cmd_async
14:25 ravenx T_T
14:25 ravenx i suppose i will just parse the returned data like a chump.
14:26 ravenx except, i can pass it a jid, do you by anychance know what format the jid takes after?
14:26 ravenx like year + unixepoch?
14:27 knine joined #salt
14:27 flowstate joined #salt
14:27 anotherZero joined #salt
14:27 gtmanfred the date
14:27 ravenx well i do see the year, month and date in the thing
14:28 ravenx for exampel:  20160802162405467361
14:28 gtmanfred i think it is date + epoch
14:28 ravenx but i would like to know if there is something special about the 162405467361
14:28 gtmanfred hrm, not the epoch
14:28 ravenx hmm, might not be as this is ours:  1470148121
14:28 gtmanfred [root@f2adb9b4d80d salt]# curl icanhazepoch.com
14:28 gtmanfred 1470148128
14:29 gtmanfred might be the whole date with utc
14:29 gtmanfred hrm, that should be 14 too...
14:29 gtmanfred lemme ask
14:30 ravenx please!
14:30 gtmanfred salt.utils.jid:gen_jid
14:30 gtmanfred return '{0:%Y%m%d%H%M%S%f}'.format(datetime.datetime.now())
14:32 Hazelesque joined #salt
14:33 ravenx sweet, thansk.
14:33 gtmanfred %f is Microsecond as a decimal number, zero-padded on the left.
14:33 gtmanfred yup, np
14:33 gtmanfred found it right after I asked internally
14:34 ravenx thanks! :)
14:35 Andrew joined #salt
14:35 ravenx https://docs.saltstack.com/en/2015.8/ref/clients/index.html#salt.client.LocalClient.get_cli_returns  i have managed to use the get_cli_returns method succesfully with my jid
14:35 ravenx however, it returns a generator object, how can i get attributes such as a return code and stuff
14:36 ravenx i have tried going print dir(nameofobject)
14:36 ravenx to no avail.
14:37 gtmanfred that i do not know, i have never used that
14:37 gtmanfred it looks like it is listening on the event bus to see returns from that jid for the specified minions
14:38 ravenx gonna yolo it by putting it random stuff, like "Failed"
14:38 ravenx maybe that will return 0 to me
14:38 ravenx AttributeError: 'generator' object has no attribute 'Failed'   :D
14:38 PerilousApricot joined #salt
14:41 bowhunter joined #salt
14:44 SpX joined #salt
14:44 syndikate joined #salt
14:44 flowstate joined #salt
14:47 syndikate joined #salt
14:52 deniszh joined #salt
14:52 msn joined #salt
14:58 jcalero joined #salt
15:01 Brew joined #salt
15:02 onlyanegg joined #salt
15:04 deniszh1 joined #salt
15:08 spuder joined #salt
15:11 fredvd joined #salt
15:14 nicksloan joined #salt
15:14 jhauser joined #salt
15:17 jenastar joined #salt
15:19 Anat joined #salt
15:19 jenastar joined #salt
15:20 Anat Hey. I'm trying to install SQL server via salt on my windows 2008 R2 machine. Can anyone give me a hint of how to do it? I couldn't find something like that on the codebase
15:23 jenastar joined #salt
15:24 edrocks joined #salt
15:24 tapoxi joined #salt
15:27 rc joined #salt
15:29 cyborg-one joined #salt
15:31 Guest98123 Looking for some advice.  I want to use salt to start some windows services in a particular order, but one of my services takes a long time to start.   How would I write the state to wait for the service to start before moving on to the next cmd?
15:32 spuder_ joined #salt
15:34 MikkiWeesenaar joined #salt
15:34 MikkiWeesenaar Hello
15:34 cmarzullo o/
15:34 MikkiWeesenaar Im in need of some help with SaltCloud and AWS! :)
15:36 cmarzullo I do salt-cloud but not aws. just ask your question
15:37 benjiale joined #salt
15:38 scsinutz joined #salt
15:39 MikkiWeesenaar I am spinning up a couple of AWS (medium) Ubuntu machines, and use Salt to install Kubernetes and such. However, quite frequently, the Salt deploy script fails due to 'Unable to lock the administration directory (/var/lib/dpkg/), is another process using it?'...
15:40 Fredrick joined #salt
15:40 MikkiWeesenaar I am not manually installing other kinds of packages or such, so I guess there is some interference - but I have no clue how to debug/verify this - since Salt is arranging the installs etc.
15:40 Fredrick [salt.engines     ][CRITICAL][4890] Engine <salt.loader.LazyLoader object at 0x7f1c0ca58a50> could not be started! Error: get_messages() got an unexpected keyword argument 'wait_time_seconds'
15:40 Fredrick after enabling sqs for autoscale
15:41 cmarzullo MikkiWeesenaar: sounds like there might be an automated process that is locking apt. Maybe patching or something?
15:41 scsinutz MikkiWeesenaar: Are you running salt in a  bootstrap script?
15:43 MikkiWeesenaar @scsinutz: Yes, if Im not mistaken >_<'
15:43 MikkiWeesenaar The file containing 'FILE: bootstrap-salt.sh'
15:43 MikkiWeesenaar Yep, I am sure now, bootstrapping :)
15:45 scsinutz it seems that the bootstrap / provisioning script(s) is invoking apt before salt gets do it's work
15:46 scsinutz are you running the script in user-data?
15:50 MikkiWeesenaar (I'm handling some of my colleague's work since he's on holidays..) - it seems that the script is being called from `./cloud.profiles.d/amazon.conf'. There's a line: `script: install_salt.sh` --> which refers to the bootstrap script
15:50 XenophonF Guest98123: states don't run in parallel, so just make one service.running state dependent on the other
15:51 MikkiWeesenaar scsinutz: So, Im not entirely sure what you mean with 'user-data'.. :)
15:52 mpanetta joined #salt
15:52 XenophonF e.g., `sqlserver: service.running: [enable: True]` and `w3svc: service.running: [enable: True, require: [service: sqlserver]]`
15:53 XenophonF hm, that's not quite valid yaml but hopefully you get the idea
15:54 scsinutz MikkiWeesenaar is the bootstrap-salt.sh the official salt script or your own?
15:54 XenophonF alternatively you could use the "names" kwarg to a service.running state, and just list the services you want running, and salt should start them in the listed order
15:54 MikkiWeesenaar scsinutz: It seems very much like the official, but I will do a diff, if I get my hands on the original
15:54 scsinutz apt is getting fired somewhere
15:55 XenophonF something like: `start services: {service.running: [{names: [sqlserver, w3svc]}, {enable: True}]}`
15:56 XenophonF again, might not be a valid YAML 1-liner but hopefully it makes sense
15:56 benjiale joined #salt
15:56 XenophonF oh, anat left
15:56 MikkiWeesenaar scsinutz: It's exactly the copy from https://github.com/saltstack/salt-bootstrap/releases/tag/v2015.11.09, apart from one newline somewhere in a comment - but that shouldn't be a problem :p
15:56 Guest98123 Thank you XenophonF
15:57 adulteratedjedi left #salt
15:57 XenophonF well, Anat, if you come back later looking through the IRC log, you can (ab)use cmd.run to execute the sql server installer
15:57 XenophonF or you can use salt's windows package manager
15:57 XenophonF Guest98123: np hope that helps
15:57 syndikate joined #salt
15:59 XenophonF Anat: hopefully, this will give you an idea of how to create a package definition usable with salt - https://github.com/irtnog/salt-winrepo-private/blob/master/ssms.sls
15:59 XenophonF Anat: there are other examples in https://github.com/saltstack/salt-winrepo-ng/
16:04 flowstate joined #salt
16:06 ALLmightySPIFF joined #salt
16:07 flowstate joined #salt
16:07 west575 joined #salt
16:10 Bryson joined #salt
16:11 iggy MikkiWeesenaar: I've seen that when salt tries to connect and do the bootstrap while cloud-init is still running... dirty hack was to add a sleep to the bootstrap script
16:13 Llmiseyhaa hrrm, so network.get_hostname seems to get the short hostname; I don't see one that obviously grabs the fqdn?
16:14 timoguin joined #salt
16:14 flowstate joined #salt
16:14 Guest98123 XenophonF:  Another scenario.  I have to reboot 2 servers, but I need to wait until my long starting service is up and running on the first server before rebooting the 2nd server.  Suggestions?
16:15 iggy orchestrate
16:15 XenophonF what iggy said
16:15 MikkiWeesenaar iggy: Let me try to find a suitable place for the sleep. Thanks for that advice!
16:15 Guest98123 Can you show me how the orchestrate will look?   I have not done much with orchestration yet
16:17 Guest98123 In the orchestration, how do I wait to run the 2nd reboot until the service is running on the 1st server?
16:18 Llmiseyhaa Guest98123: Give me a moment and I'll get you an example
16:18 XenophonF Guest98123: there's also the docs here - https://docs.saltstack.com/en/latest/topics/orchestrate/orchestrate_runner.html
16:18 Llmiseyhaa short form is you run a state that 1) does whatever, then 2) reboots the server, then 3) starts the service, and you tell orchestrate to run no more than one copy of that at a time and to quit if even one fails
16:19 Llmiseyhaa though with the reboot, the minion will go inaccessible for a bit so that is a complication
16:20 Llmiseyhaa ugh... there's got to be a way to get the fqdn of a minion
16:20 Guest98123 And if I tell it to reboot, won't it move on as soon as the reboot is done, even though my service is taking 5 minutes more to start.   I don't want it to fail, I just want it to wait until the service actually starts after the reboot before it moves on to reboot the next server
16:21 Miouge joined #salt
16:21 Llmiseyhaa Guest98123: You write the state that starts the service and requires the reboot, which requires whatever else
16:21 Llmiseyhaa then from orchestrate you call the state that starts the service, it requires the state that does the reboot so the reboot gets fired
16:21 * Llmiseyhaa pokes around salt.modules.network and tries to figure out why get_hostname() only seems to give her the shortname. ) =
16:22 darix joined #salt
16:23 flowstate joined #salt
16:23 iggy Llmiseyhaa: just cmd.run it
16:23 * Llmiseyhaa facepalms
16:23 Llmiseyhaa Thanks.  I'm an idiot.  Should have thought of that.
16:23 iggy and file a ticket to add a fqdn arg to network.get_hostname
16:24 ALLmightySPIFF joined #salt
16:25 iggy Llmiseyhaa: actually, duh... fqdn is available in grains (along with the short name)
16:25 * Llmiseyhaa facepalms double time.  Feels like an idiot.
16:25 Llmiseyhaa thanks
16:26 iggy yeah, I did a facepalm this time too ;)
16:26 flowstate joined #salt
16:26 Llmiseyhaa well that saves me from having to write a mine function for this.  Thanks
16:28 scsinutz joined #salt
16:30 armin joined #salt
16:34 Bryson joined #salt
16:37 writtenoff joined #salt
16:37 my10c joined #salt
16:38 Xopher joined #salt
16:41 ivanjaros joined #salt
16:44 anotherZero joined #salt
16:44 Andrew left #salt
16:45 tpaul joined #salt
16:54 lovecraftian joined #salt
16:54 Brijesh1 joined #salt
16:54 impi joined #salt
16:57 nicksloan joined #salt
16:58 scsinutz joined #salt
16:59 ZachLanich joined #salt
16:59 Fredrick [DEBUG   ] Initializing new AsyncZeroMQReqChannel for ('/etc/salt/pki/minion', 'server.minion.test', 'tcp://10.18.4.89:4506', 'clear') then just [DEBUG   ] SaltReqTimeoutError, retrying. (1/7)
16:59 Fredrick any one know how to fix?
16:59 sjorge joined #salt
16:59 MTecknology can you telnet to that address and see the magic character?
16:59 MTecknology if not .. firewall
17:00 pipps joined #salt
17:01 Fredrick Yes
17:01 pipps joined #salt
17:01 Fredrick I upgraded to the new salt-minion worked prior
17:02 ageorgop joined #salt
17:06 error_ joined #salt
17:06 error_ left #salt
17:06 cyborg-one joined #salt
17:07 error_ joined #salt
17:08 error_ left #salt
17:10 MTecknology Fredrick: I see debug logging, but maybe try with trace and see if it catches anything.   salt-call -l trace
17:12 MTecknology This is interesting. I have a minion connected to a syndic which is connected to a master. The minion is running a state and generating an error 'Rendering SLS 'users.all_groups' failed. Please see master log for details.' Okay, sounds easy enough. I look at the syndic logs .. nothing at all. Wondering if maybe it made its way up to the master?... nope, not there either. :S
17:13 netcho joined #salt
17:13 Brijesh1 joined #salt
17:22 mpanetta joined #salt
17:30 Fredrick salt-call -l trace just prints the help for salt-call
17:30 Fredrick Not sure what it should do?
17:31 netcho joined #salt
17:31 iggy salt-call -l trace <rest of the options you used before>
17:32 Fredrick I was running just a salt-minion -l debug
17:32 cyborg-one joined #salt
17:33 Edgan whytewolf: Your jinja renderer doesn't work for me in master mode either. It does find the jinja file, but the output is the hostname: and empty lines. It seems like salt.utils.templates.TEMPLATE_REGISTRY is empty from trying to debug it.
17:34 Edgan Fredrick: I have repeatedly found in some cases trace is a must, debug isn't enough.
17:34 whytewolf Edgan: it renders the jinja, but if you are trying to render a map.jinja there is no output
17:34 Fredrick Cool looking at it now.
17:35 Edgan whytewolf: ok, what is the point of this is not to see the results?
17:36 whytewolf Edgan: map.jinja keeps everything in jinja. and only creates variables. point it at somethign that outputs variables
17:36 Edgan whytewolf: you are saying rewrite debug.py to do what I want?
17:37 whytewolf no
17:37 Edgan whytewolf: I need some background
17:37 Fredrick https://gist.github.com/anonymous/fc46437abdce3017dcbf1e6be4de7496
17:37 Fredrick Still just hangs
17:37 whytewolf that would require a lot more then rewriting debug.py that would require rewriting jinja
17:38 whytewolf you just need a template that calls the map.jinja and {{ variable }}
17:38 edrocks joined #salt
17:38 whytewolf and to call that template
17:38 Edgan whytewolf: ok, let me give that a try
17:38 iggy Fredrick: the master is newer than the minion right?
17:39 Fredrick Yes
17:39 iggy then I got nothing
17:39 tuxx joined #salt
17:39 Fredrick Actually both on 2016.3.1
17:39 iggy if you can connect to 4505/4506 on the master from the minion, you shouldn't have any issues
17:40 Fredrick That is what I thought as well and it did work before I updated the minion to the new version.
17:41 bluenemo joined #salt
17:41 * Llmiseyhaa pokes at jinja... headdesks.  Feels like she should just say screw it and go to a python interpreter file, because trying to figure out what datatype any given item is in jinja for applying python methods is making her head hurt.
17:41 iggy oh... no other minion processes running?
17:42 whytewolf Llmiseyhaa: there is always pyobjects
17:42 Llmiseyhaa hrm?
17:42 whytewolf Llmiseyhaa: https://docs.saltstack.com/en/2015.8/ref/renderers/all/salt.renderers.pyobjects.html
17:43 bluepaladin joined #salt
17:43 whytewolf jinja is not the only render engine in salt. it is just the most documented one
17:43 Fredrick root      7765  4448  0 11:43 pts/1    00:00:00 grep --color=auto minion
17:43 Fredrick Nope
17:43 Llmiseyhaa In this case, it's to replace my map.jinja in a state that has gotten... ugly.
17:43 * Llmiseyhaa reads through that
17:43 Edgan whytewolf: I tried creating saltstack/test.jinja. It imports saltstack/map.jinja, and map.jinja sets saltstack. The test.jinja file also says {{ saltstack }}. Response from debug, Minion did not return. [No response]
17:44 Llmiseyhaa oooh there is specific support for that, nice
17:44 whytewolf Edgan: wtf. I don't know what is going on in your enviroment.
17:45 amcorreia joined #salt
17:46 Llmiseyhaa Actually, hrm, that shows how to use a pyobjects map with other pyobjects states... if I wrote a pyobjects map would it work with the standard {% import ... %} in a yaml/jinja state?
17:46 Llmiseyhaa doesn't say either way; I'm hoping the answer is yes but
17:46 whytewolf I do not belive so.
17:46 zer0def joined #salt
17:47 Llmiseyhaa ugh.
17:48 Edgan whytewolf: https://paste.fedoraproject.org/399993/60039147/
17:49 Edgan whytewolf: Here is the minion error, https://paste.fedoraproject.org/399996/60176147/
17:51 Fredrick Got something different finally. [TRACE   ] Client disconnected from IPC /var/run/salt/minion/minion_event_1be82bdb7c_pull.ipc
17:53 bluepaladin joined #salt
17:53 impi joined #salt
17:56 whytewolf Edgan: msgpack blowing up? while trying to pack render_tmpl but it shouldn't be trying to do that at all.
17:56 ajw0100 joined #salt
17:57 whytewolf unless salt.utils.templates.TEMPLATE_REGISTRY[template] is doing it, but that is an internal salt function.
17:57 Edgan whytewolf: simplifying the test
17:57 whytewolf Edgan: this is the test I recently run it on while trying to help Llmiseyhaa yesterday
17:58 whytewolf https://gist.github.com/whytewolf/ff4e7acd721577a9dfc57236245365e6
17:58 edrocks anyone run into "TypeError: 'NoneType' object is not iterable" with docker.running?
17:58 LessSneaky joined #salt
17:58 edrocks full error http://pastebin.com/7cHQjebs
17:59 Edgan whytewolf: https://paste.fedoraproject.org/400008/47016076/  Still blows up
18:02 Edgan whytewolf: going to try your example
18:03 whytewolf Edgan: I think there might be something actually different about your setup. I have no idea why it is blowing up on msgpack.
18:03 whytewolf I return the content of the outputted template not a render function
18:03 iggy edrocks: what does the state look like? ... also, file a ticket, tracebacks are bad mkay
18:05 IdoKaplan joined #salt
18:05 IdoKaplan Hi, I'm using a lot with pillars, package version for example. Every time that i'm running highstate in order to install a new package, I need to run before "saltutil.refresh_pillar". Is there a solution that the pillar will be refresh faster?
18:05 iggy highstate should refresh pillar automatically
18:06 scsinutz joined #salt
18:06 Edgan whytewolf: I am going to try a few things, but your example blows up the same way
18:06 edrocks iggy: sorry had to live patch the docker module
18:06 edrocks iggy: whole site went down
18:07 whytewolf Edgan: your example outputs this for me.  {'state': {'name': 'systemd'}, 'config': {'dir': '/etc/systemd/system'}}
18:07 Edgan whytewolf: My salt is patched with various bug fixes. I doubt it is anything that affects this, but maybe. Also it might be a slightly different combination of python module versions.
18:07 emaninpa joined #salt
18:07 Edgan whytewolf: distro/version?
18:08 whytewolf Edgan: currently centos7. however i have used this code on ubuntu 14.04 as well. and have used it since salt version 2014.7 i think. it has been a long time
18:09 Edgan whytewolf: version of salt?
18:09 whytewolf current version is 2016.8.1
18:09 timoguin joined #salt
18:09 whytewolf err 2016.3.1
18:09 Edgan whytewolf: gitfs or filesystem?
18:09 whytewolf filesystem.
18:09 whytewolf but that shouldn't matter
18:10 Edgan whytewolf: going to try unpatched salt 2016.3.1 on Fedora 24 with gitfs and see how it goes
18:11 edrocks isn't 2016.3.2 out?
18:11 Llmiseyhaa huh, that's odd... what's the difference between {%- %} and {% %}?
18:11 whytewolf Llmiseyhaa: {%- %} removes the whitespace before it
18:11 Llmiseyhaa Oh!  So that's how I avoid having globs of whitespace where there were {% %}s?
18:11 whytewolf yeah, but they can be touchy
18:12 Tanta joined #salt
18:12 Edgan edrocks: yes as of 4 days ago. I look regularly and hadn't noticed yet
18:12 whytewolf Llmiseyhaa: http://jinja.pocoo.org/docs/dev/templates/#whitespace-control
18:12 Llmiseyhaa thanks, that's new since the last time I worked with this stuff
18:13 edrocks Edgan: I just updated as my main app state was broken
18:14 whytewolf 2016.3.2 is not out.
18:15 Miouge joined #salt
18:15 whytewolf it is still in testing. 2015.8.11 was released 4 days ago
18:16 edrocks why would pip install it then?
18:16 Edgan whytewolf: it is on the github page hiding in ...
18:16 edrocks iggy: I made an issue. It's something to do with getting a list of tags for your image https://github.com/saltstack/salt/issues/35139
18:16 saltstackbot [#35139][OPEN] dockerng: RepoTags not iterable with docker 1.12 on ubuntu 14.04 | Description of Issue/Question...
18:16 whytewolf they have it tagged
18:16 whytewolf that doesn't mean it has been officially released
18:17 blueelvis joined #salt
18:17 edrocks anyone else using dockerng and run into above issue?
18:17 Edgan whytewolf: I see no debs, but in my book they shouldn't tag if it isn't official
18:18 numkem Where do I need to put grains .py files in? I've tried in the grains folder of the folder that is set in modules_dir and while the file does get a .pyc it doesn't show up in grains.items
18:18 whytewolf numkem: you put it in a directory that the system can see called _grains off of the root file_dir. then run saltutil.sync_all
18:19 numkem whytewolf: I've tried that and it doesn't seems to do anything, it doesn't get a .pyc
18:20 whytewolf numkem: it puts the file where it can be seen by salt. normally somewhere in /var/cache/salt/minion
18:21 pipps joined #salt
18:21 whytewolf numkem: as for a .pyc don't worry about those so much it will compile when it runs
18:22 benjiale joined #salt
18:22 numkem whytewolf: somehow while I did all the things you said, doing the saltutil sync from the master instead of the minion seems to have done the trick. Thank you
18:26 scsinutz joined #salt
18:26 Edgan whytewolf:  won't even run on Fedora 24, https://paste.fedoraproject.org/400024/70162381/
18:27 whytewolf Edgan: did you sync_modules first?
18:28 netcho joined #salt
18:28 MTecknology I'm reading about something I've never used before and trying to understand it's logic "New in version 0.17.0."  ... and it feels new.
18:28 patarr joined #salt
18:28 MTecknology Love that the source actually made it clear as day what's up, though.
18:29 ajw0100 joined #salt
18:29 Edgan whytewolf: no, now it works :)
18:30 Miouge joined #salt
18:33 Edgan whytewolf: gitfs works, patched salt works
18:35 babilen MTecknology: What was it?
18:37 MTecknology babilen: using grains.filter_by in a pillar. Every time I read the sls, I was thingking "filter by... WHAT?!"
18:38 MTecknology Right there in the docs (and function definition) ''' salt.modules.grains.filter_by(lookup_dict, grain='os_family', '''
18:42 IdoKaplan Hi, I'm using a lot with pillars, package version for example. Every time that i'm running highstate in order to install a new package, I need to run before "saltutil.refresh_pillar". Is there a solution that the pillar will be refresh faster?
18:44 Edgan whytewolf: works on ubuntu 14.04
18:45 babilen MTecknology: Wonderful :)
18:46 LessSneaky joined #salt
18:48 Edgan whytewolf: trying to narrow down what it is about 16.04
18:50 Fredrick https://gist.github.com/anonymous/b29ea9c5fef227d68f6edb30c5539d95 Anything in this that would help explain why it can no longer connect to the master
18:53 whytewolf Edgan: python version?
18:54 dyasny_ joined #salt
18:55 whytewolf it might not be python3 ready
18:56 Edgan whytewolf: same version of python 2
18:56 Edgan whytewolf: using the debs based on the official saltstack debs
18:56 ribx joined #salt
18:57 Edgan whytewolf: this isn't the first 16.04 bug I have run into
18:57 babilen Edgan: Could you try the official packages in xenial?
18:57 Edgan whytewolf: But I have to go to 16.04 to get pygit2 for gitfs
18:57 Edgan babilen: doubt it, my patched version works on Fedora 24
18:57 babilen Edgan: That should be 2015.8.8+ds-1
18:58 Edgan babilen: no, they have 2016.3.1 packages
18:58 babilen Edgan: I packaged that and xenial has 2015.8.8+ds-1
18:58 babilen Saltstack might provide 2016.3.1 packages in their third-party repositories
18:58 Edgan babilen: oh, you mean Canonical/Ubuntu packages, not saltstack.
18:59 babilen yes .. I admit that "official" might have been misleading
18:59 edrocks joined #salt
19:00 Edgan babilen: will give it a quick try
19:00 babilen Ah, guess I should package .2 soon :)
19:01 pipps joined #salt
19:04 Miouge joined #salt
19:04 Lionel_Debroux joined #salt
19:06 timoguin joined #salt
19:07 toanju joined #salt
19:07 Asido joined #salt
19:08 numkem is it possible to add the --pid option to docker.running?
19:08 Asido I've been successfully using https://github.com/saltstack-formulas/postgres-formula
19:08 Asido until a couple days ago when state apply all the time says "Comment: Failed to create database"
19:08 Asido I have a couple questions regarding this:
19:09 Asido 1. can I specify in gitfs_remotes a revision it should use? I see the formula received some commits in the last couple days
19:09 Edgan babilen: still "Minion did not return. [No response]", but no backtrace in /etc/salt/minion
19:10 Miouge joined #salt
19:10 whytewolf Asido: common practice is to fork something you are going to be using so that you don't have to worry about changes upstream and can determine if those changes apply to you on your own terms
19:11 Asido whytewolf, makes sense. thanks
19:13 IdoKaplan Hi, I'm using a lot with pillars, package version for example. Every time that i'm running highstate in order to install a new package, I need to run before "saltutil.refresh_pillar". Is there a solution that pillars will be refreshed immediately?
19:14 cmarzullo are you masterless?
19:15 IdoKaplan cmarzullo: Are you asking me? if so, no - i'm using salt master
19:16 babilen Edgan: sry afk
19:18 POJO joined #salt
19:19 cmarzullo I was asking you. Seems odd. My pillar updates happen right away.
19:20 cmarzullo show a pillar.show (or whatever it is) to see if your generating the right pillar.
19:20 jhauser joined #salt
19:21 nicksloan why might I see data from salt-run pillar.show_pillar some.minion.id but not salt some.minion.id pillar.items?
19:21 badon joined #salt
19:26 GreatSnoopy joined #salt
19:27 IdoKaplan cmarzullo: AFAIK, this is by design - https://docs.saltstack.com/en/latest/topics/pillar/#refreshing-pillar-data
19:27 iggy salt-run runs on the master (which always has the latest pillar data) whereas salt ... runs on the minion (which has cached pillar data)
19:27 iggy nicksloan: ^
19:28 nicksloan iggy: I tried salt some.minion.id saltutil.refresh_pillar
19:28 nicksloan and got the same behavior
19:28 nicksloan does that not do what I think it does?
19:29 iggy maybe some jinja rendering differences?
19:29 cmarzullo hmmm cached pillar data? Cached on the minion? Isn't the point of pillar is that the data doesn't live on the minion?
19:30 iggy cmarzullo: it doesn't "live" there, but a minion caches the pillar data that is targeted at it (otherwise you'd have 100s or 1000s of calls to the master for ever highstate)
19:32 nicksloan there is a small bit of jinja involved in one file, but it's all self contained in that file. Just assigns a dictionary to a jinja variable, and then formats it into a string
19:35 Fredrick joined #salt
19:39 toanju joined #salt
19:40 fracklen joined #salt
19:41 peters-tx Anyone else getting an error from the Saltstack Yum repository for RHEL7?
19:41 nicksloan duplicate key in the state top file
19:41 whytewolf they might be right in the middle of putting 2016.3.2 out
19:42 Sketch peters-tx: zeromq-4.1 from epel conflicts with python-zmq in salt?
19:42 nicksloan not reported to the master, and not relevant to the runner. Pain in the ass to debug
19:42 gtmanfred if you have a duplicate key, it should print out what they key that is a duplicate in the message
19:42 Edgan whytewolf: https://github.com/saltstack/salt/issues/13486  This traceback looks like mine
19:42 nicksloan gtmanfred: did not in this case
19:42 saltstackbot [#13486][OPEN] Grains that cause errors are caught in msgpack serialization, the error is not helpful. | If you deploy a grain that fails to catch it's own errors it will result in the following:...
19:43 peters-tx skeezix-hf, No, getting a 404 Not Found error w/ a RepoData file at the Repository itself
19:43 nicksloan though it did log it in the minion
19:43 peters-tx Err Sketch rather
19:43 gtmanfred peters-tx: 2016.3.2 is being released right now it looks like, so it may be the repo updating that data
19:43 gtmanfred give it a bit
19:43 peters-tx gtmanfred, Aha, yes that would explain it.  Thanks
19:43 gtmanfred also note this https://groups.google.com/forum/#!topic/salt-announce/y31sz9_AX-g
19:44 cmarzullo iggy: how can it not live on the minion but be cached on the minion? Seems odd. We are very heavily pillar dependant and I do see the master generate pillar responses when highstating.
19:44 whytewolf Edgan: it is very close :/
19:45 peters-tx gtmanfred, Ok, thank you for the info
19:45 BretFisher joined #salt
19:45 whytewolf Edgan: except that is about catching errors. which is why it can't serialize exceptions.ValueError
19:45 BretFisher joined #salt
19:46 iggy cmarzullo: how many do you see?
19:47 whytewolf Edgan: yours is about unable to serialize render_tmpl which should not be output anywhere
19:47 Edgan whytewolf: yeah, not exactly the same, but very like backtrace
19:47 Edgan whytewolf: and your module is not that different than python code in a grain
19:47 CeBe1 joined #salt
19:48 cmarzullo iggy: looking into it. I do see /var/cache/salt/minion/highstate.cache.p peeking inside there.
19:48 whytewolf Edgan: grains are written very simalar to modules
19:49 whytewolf Edgan: but the problem is the output is showing that it is outputing a render_tmpl function which is WAY wrong
19:50 whytewolf that isn't even an exception
19:50 cmarzullo ok that's targeted pillar data. so not so bad.
19:50 Sandman joined #salt
19:51 iggy "caches the pillar data that is targeted at it"
19:51 whytewolf ... unless
19:52 cmarzullo yeah think I brain farted there for a sec.
19:52 whytewolf Edgan: I'm wondering if it is line 105. which it shouldn't be unless the template you are passing in doesn't exist
19:52 cmarzullo my bad
19:53 ajw0100 joined #salt
19:53 whytewolf no that isn't right, template is just the tenplate name
19:54 whytewolf Gah. I see no way this code could be leaking a render_tmpl
19:55 whytewolf the only render_tmpl it loads is set into data
19:55 whytewolf which is never returned directly
19:56 Edgan whytewolf: haha, here is one for you
19:57 Edgan whytewolf: double checking the result
19:58 Miouge joined #salt
19:58 Edgan whytewolf: when does salt normally do a sync_modules on it's own?
19:59 whytewolf um, as far as i know it doesn't
19:59 Edgan whytewolf: That is a serious oversight. :)
19:59 whytewolf i could be wrong about that
20:00 jesusaur joined #salt
20:00 Edgan whytewolf: Well I replaced the 105 return with a return error where error = 'Error', and it started working after running sync_modules, but I am wondering if it was just the sync_modules.
20:01 whytewolf humm they do say that it is applyed when state.apply is used in the docs. but i have never had that actually be the case
20:01 whytewolf Edgan: put it back and see?
20:01 Edgan whytewolf: I did, keeps working
20:01 Edgan whytewolf: So maybe I was fighting a lack of sync_modules the whole time
20:02 whytewolf lol. okay... that makes me wonder what debug you were calling then :P
20:03 Edgan whytewolf: Thanks for the help
20:03 whytewolf no problem. was racking my brain on what should have been very simple code
20:03 Edgan whytewolf: I think I will add a debug.jinja to my formulas make debugging easy
20:04 Edgan whytewolf: and now to see if I can figure out how to get it working with salt-ssh
20:05 whytewolf good luck with that. salt-ssh is a forgien beast to me. I can't get anything to run right on it.
20:05 cyrus_mc joined #salt
20:05 whytewolf if you do get it working feel free to send me a PR
20:05 Edgan whytewolf: ok
20:06 Edgan whytewolf: I have fixed a few bugs in salt-ssh. It is really just a wrapper around salt-call.
20:06 cyrus_mc reading through deploying zip archive module. I created the zip, stored it in the file root base but when I do saltutil.sync_modules call it doesn't sync
20:07 whytewolf cyrus_mc: it needs to be in a directory called _modules off of the file root base
20:08 onlyanegg joined #salt
20:09 Fredrick So found the fix to my problem.
20:10 Fredrick Added the hash_type to the minion file and it is happy again
20:12 IdoKaplan Hi, I'm using a lot with pillars, package version for example. Every time that i'm running highstate in order to install a new package, I need to run before "saltutil.refresh_pillar". Is there a solution that pillars will be refreshed immediately?
20:14 whytewolf alias saltit='salt '*' saltutil.pillar_refresh && salt '*'' state.apply'
20:15 whytewolf well, basic idea anyway. with changed quoting
20:16 fredvd joined #salt
20:16 Edgan whytewolf: salt-ssh doesn't seem to be pushing any of the sls/jinja code to the remote host
20:16 Llmiseyhaa IdoKaplan: And for bonus points, add in git where you have to wait up to ~1min for salt-master to fetch the most up to date copy of the states and pillars from git...
20:17 Edgan whytewolf: it is copying grains and modules
20:17 whytewolf Edgan: weird
20:18 IdoKaplan L1miseyhaa: i'm running "salt-run fileserver.update" before.
20:19 whytewolf sooo, alias salt-the-earth="salt-run fileserver.update && salt '*' saltutil.sync_all refresh=true && salt '*' state.apply"
20:20 Llmiseyhaa It's Ll, not L1. (=
20:20 Llmiseyhaa ok zookeeper is working, now for solr.
20:20 whytewolf yay!! progress is always good
20:21 Llmiseyhaa I just wish I had more time to do this in and I didn't keep running into stuff that takes me a solid day or more of fighting just for one piece... but, eh, learning something while using it.  Especially with the fact that I'm figuring out new personal best practices as I go so I'll need to go back, document them, then clean up all the states I've written to match them and...
20:24 IdoKaplan whytewolf: It means that i will refresh 100 minions every time
20:24 IdoKaplan Do you have maybe another idea?
20:26 pipps joined #salt
20:27 whytewolf IdoKaplan: there isn't really another way. you are going to be calling pillar_refresh before a highstate. if you are changing pillars that often. [only really need to refresh when something changes].
20:28 whytewolf i would put it into an alias anyway. or a bash/tsch/ksh/zsh/fish function
20:31 IdoKaplan whytewolf: I don't want to use "pillar_refresh" inside the state because it means that if i execute highstate I always get a message that there was a changed.
20:32 IdoKaplan whytewolf: but if there is no other option, I will live with that :( thanks you very much!
20:33 whytewolf I didn't say put it in a state
20:33 whytewolf i said make it an alias in your default shell that runs before salt
20:33 IdoKaplan that will refresh all the minions?
20:34 whytewolf it will just be like running it from the command line [cause it is] just turns it into a single command
20:35 cmarzullo something must be off IdoKaplan I update pillar all the time and see the changes immediately on the minion when I highstate.
20:36 ajw0100 joined #salt
20:36 whytewolf cmarzullo: are you masterless?
20:36 IdoKaplan I think that i understand what do you mean, but it means that this command will refresh all pillars, unless I will write a simple bash with args
20:36 whytewolf IdoKaplan: yes
20:37 cmarzullo no. running master and syndics.
20:37 whytewolf IdoKaplan: doesn't have to be a full script can just be a bash function
20:38 IdoKaplan ok, i see, thanks! can you please explain what are the diff between highstate and state.apply?
20:38 IdoKaplan Can you please help me with the function? I understand if not, I guess that I can managed it by myself.
20:39 whytewolf IdoKaplan: state.apply is basicly just a wrapper for calling either state.sls or state.highstate
20:39 fracklen joined #salt
20:40 PerilousApricot joined #salt
20:41 IdoKaplan Sorry, not sure that I understand. Can you please try to explain again?
20:41 jhauser joined #salt
20:42 pipps joined #salt
20:43 whytewolf !salt state.apply
20:43 saltstackbot https://docs.saltstack.com/en/latest/ref/states/all/salt.states.apply.html
20:43 sagerdearia joined #salt
20:44 scsinutz joined #salt
20:47 whytewolf IdoKaplan: here is a $0.02 version of the bash function. https://gist.github.com/whytewolf/dba51dfe0821fbb9a232844c0aaf511c
20:48 fracklen_ joined #salt
20:54 pipps99 joined #salt
20:54 ajv joined #salt
20:56 dfinn joined #salt
20:59 IdoKaplan whytewolf: 10x, thank you. I have another question please :)
20:59 IdoKaplan Hi, i'm trying to find a solution to debug states and pillar locally. For example, I'm editing the state "/var/cache/salt/minion/files/dev/foo.sls" and running state.template and it's working. I'm trying to do the same on pillar, i.e, editing the file "/var/cache/salt/minion/pillar/cache.sls", but after I execute state.template or "salt-call pillar.items", the pillar that I changed was rollback to the previous value. Is there a way t
21:00 IdoKaplan Is there a way that the cache pillar will not be changed?
21:02 pipps joined #salt
21:03 cmarzullo So are so far away from normal I'm not sure what to say.
21:03 jcalero joined #salt
21:03 perfectsine joined #salt
21:04 whytewolf cmarzullo: actually he isn't. pillar cache is a thing. it is why the pillar_refresh command exists.
21:04 cmarzullo editing the cache files directly is a thing?
21:05 whytewolf oh didn't see he was editing cache files directly
21:05 whytewolf thats.... wtf
21:06 scsinutz1 joined #salt
21:06 cmarzullo Just feel there's something else going on. Like I litterally just updated pillar and highstated a minion and it took the changes. no cache clearing or anything.
21:07 coval3nce joined #salt
21:07 whytewolf cmarzullo: unless that is a recent change that pillar is refreshed every state.apply it didn't used to be that way.
21:10 whytewolf cause pillar.get always pulls from cache.
21:10 cmarzullo really? Admittdly I haven't been using salt for long. (last sept. before that it was ansible / puppet/ chef )
21:10 cmarzullo but I've always had it update regularly.
21:10 cmarzullo when highstating.
21:10 whytewolf are you using pillar.get or pillar.item?
21:10 coval3nce Anyone have recommendations on pillar enviornment strategies?  I haven’t been able to find a clean way to trutly segregate pillar data between environments.  All attempts seem to merge environment pillar files.
21:10 cmarzullo state.highstate
21:11 whytewolf in your templates
21:11 josuebrunel joined #salt
21:11 cmarzullo In templates I either do the whole import map.jinja business or more recently just pass in the top of the hash i care about on the file.managed line
21:12 cmarzullo again with the map.jinja business
21:12 hop left #salt
21:12 whytewolf ok, so you don't call pillar.get directly in any templates.
21:13 whytewolf which would be why you never see the issue
21:13 Tyrm_ joined #salt
21:14 cmarzullo I see. Thanks for helping me understand that.
21:14 whytewolf wait. how are you calling data into your merge in map.jinja? that should still be a pillar.get
21:14 Fredrick How to update salt-run manage.versions?
21:15 whytewolf Fredrick: that should be updated by the minion when they connect. if it is still showing an old version then an old version of the minion is still running
21:15 Fredrick nevermind, ran into a glitch with a local repo and did not get minions updated yesterday.
21:16 cmarzullo hehe yeah that's exactly what I was just checking. :)
21:16 cmarzullo http://pastebin.com/t2Khv4fC
21:16 Fredrick Did not catch they pulled a new version today.
21:16 cmarzullo ^^ my map.jinja basically ripped from the guides
21:16 whytewolf yeah you are pulling in using pillar.get
21:16 oida joined #salt
21:17 scsinutz joined #salt
21:18 scsinutz1 joined #salt
21:19 ajw0100 joined #salt
21:20 tkharju joined #salt
21:22 cmarzullo oh well hope he solves it. . . .
21:26 ajw0100 joined #salt
21:32 ajw0100 joined #salt
21:34 coval3nce whytewolf: Know of anywhere i can read up on recommendations for pillar environment segregation?
21:34 whytewolf coval3nce: sorry not really. I avoid enviroments like the plague
21:35 coval3nce I’m having so many issues with it.
21:36 coval3nce I’ve commented on a few github issues about it, but crickets so far.
21:36 onlyanegg joined #salt
21:36 coval3nce Crickets on the mailing list too.
21:37 babilen just don't
21:37 coval3nce If i don’t how are folks segregating dev/test/stage/prod environments?
21:37 whytewolf seperate salt-masters
21:37 coval3nce All physically copied different masters and git repos?
21:38 coval3nce So, folks don’t use a workflow where you push from stage branch to master branch?
21:38 coval3nce They update completely different repos?
21:38 seblu joined #salt
21:39 flowstate joined #salt
21:41 hemebond I would/do use branches but I don't use gitfs
21:41 Edgan coval3nce: technically you could use the flow with different salt masters, you just have to control the branch checked out on different masters
21:41 hemebond ^
21:41 babilen You can whitelist branches
21:41 coval3nce Ah crap, maybe all i’ve been missing alla along is “pillarenv” minion option
21:42 coval3nce babilen: @Edgan yeah i was thinking of using a different whitelist config on stage salt master vs prod ones to get the desired effect
21:42 coval3nce BUT, i just tested setting “saltenv” on a minion, and maybe im working
21:42 babilen https://docs.saltstack.com/en/latest/ref/configuration/master.html#gitfs-env-whitelist (and so on)
21:48 coval3nce @babilen @Edgan it appears if `pillarenv` is set to match the `environment` tag in the minion config, then pinnin works.  Probably not recommended to go that approach tho?
21:57 mpanetta Hey all, good afternoon.  I have a question... Is it possible to do something similar to a glob on grains in the pillar top.sls but with an ext_pillar instead?
21:58 mpanetta Basically I have my server roles stored in consul, but the only way I have to access that data from salt is through the consul ext_pillar module.
21:59 mpanetta So not quite sure how to set up my top file so the various roles get their own pillar files.
21:59 MTecknology How can I figure out what version of salt a PR was released in? In the case of PR #33268, it was pulled into develop, but I don't know if it made it into 2016.3.2.
22:00 Edgan MTecknology: What is the date?
22:02 Edgan MTecknology: very likely not in 2016.3.2
22:02 MTecknology Edgan: merged may 16
22:02 Edgan MTecknology: on the other hand rumor has it there may be an new August release
22:03 pipps joined #salt
22:03 MTecknology Do these need a PR to cherry pick into the next release update?
22:03 pipps joined #salt
22:04 Edgan MTecknology: I am about to rebuild my patches 2016.3.2
22:04 MTecknology I don't know what that means :(
22:05 Edgan MTecknology: https://docs.saltstack.com/en/latest/topics/releases/2016.3.2.html  official list of the PRs in .2
22:06 Edgan MTecknology: I have various patches that I had to patch into 2016.3.1. I think all are merged fixes, some mine, some other people's. When I built 2016.3.1 it didn't have a merged change from April.
22:06 flowstate joined #salt
22:09 jcalero joined #salt
22:09 MTecknology Edgan: so... I have two PRs that have been merged and should really find their way into 2016.3.2 ... how do I make this happen?
22:10 MTecknology this is one of them
22:10 ribx joined #salt
22:12 Edgan MTecknology: Are the PRs in the list in the release notes?
22:13 MTecknology nope
22:13 Edgan Then they are most likely not in .2
22:13 MTecknology ...
22:13 MTecknology that doesn't answer my question...
22:14 Edgan MTecknology: I don't think you do
22:14 MTecknology What do I do to make these PR's get cherry picked?
22:14 Edgan MTecknology: as in I don't think it is in your control
22:14 Edgan MTecknology: ask jfindlay
22:15 MTecknology jfindlay: Any chance you could help me get a couple patches sucked into the next 2016.3.x?
22:16 Tyrm joined #salt
22:16 Edgan MTecknology: My guess to here answer is we decide internally, and if you really want to help pick PRs in builds, pay for Salt Enterprise.
22:18 MTecknology Edgan: Obviously, you guys would decide internally, that's well above obvious. These PR's address functionality/usabliity bugs in the current 2016.3.x release and at least one of these /should/ be picked. I remember having done this about four years ago, but I can't remember how to do this anymore.
22:19 MTecknology Edgan: Without the request, the powers that be have no ablity to review the request.
22:20 Edgan MTecknology: Why don't they just suck in all PRs that pass tests?
22:20 * MTecknology sighs
22:20 MTecknology Edgan: because of the obvious implications?
22:20 MTecknology because... that's the reason we have to request cherry picking in the first place
22:21 protoz joined #salt
22:21 Edgan MTecknology: but what makes your PRs special over all the other PRs, other than you asked them nicely?
22:21 whytewolf MTecknology: normally you should build your patches agenst the develop branch for earlist version that it affects. then they push into develop from there. however if something goes into develop that needs to be back ported. you just let someone there know. like jfindlay
22:22 vook Is there a saltstack repo (archive or otherwise) that has freebsd i386 packages?
22:22 vook just for the minion
22:24 MTecknology whytewolf: so, I should done git checkout v2016.3.1 (or the correct tag), built the patch there, submitted a PR, and it would have been identified as needing to go into the next 2016.3.x?
22:25 MTecknology whytewolf: I thought it used to be as simple as adding a cherry-pick tag to a PR that had been sucked into develop.
22:25 whytewolf branch 2016.3
22:25 MTecknology If I built it against that branch, would it find it's way back into develop?
22:26 whytewolf most of the time yes
22:26 whytewolf ... there have been failures when it was done between when they tag and when they release
22:28 whytewolf https://docs.saltstack.com/en/latest/topics/development/contributing.html#sending-a-github-pull-request
22:28 MTecknology okay... so future reference! If I'm going to build a patch againsta currently released version of salt, build the patch on that branch and do the PR into develop and let magic ensue.
22:28 onlyanegg joined #salt
22:28 MTecknology a PR from my develop to your develop is new features or things that I don't expect to be backported?
22:29 whytewolf yeah
22:29 whytewolf [not my develop, saltstacks]
22:29 whytewolf hehe
22:29 MTecknology :P
22:29 ajw0100 joined #salt
22:29 MTecknology Sweet! Now that I know that, I can do better in the future.
22:30 MTecknology whytewolf: https://github.com/saltstack/salt/pull/33268  --  This is the big one I'm looking at now. It
22:30 saltstackbot [#33268][MERGED] Fix two pure python errors causing UnboundLocalErrors | What does this PR do?...
22:31 Edgan MTecknology: Well they did better than expected. All but one of my patches made it into .2.
22:32 whytewolf ohh that would be a nice fix.
22:32 MTecknology ;)
22:32 MTecknology That's one I feel should definitely absolutely be reviewed for picking
22:33 whytewolf well hopefully someone from saltstack is paying attention :P
22:33 whytewolf I miss the old days when they manned this channel all the time
22:33 MTecknology ya... :(
22:34 MTecknology the days of kuffs
22:35 MTecknology whytewolf: This is what I'm trying to figure out, though... they don't hang out in here like they used to, so how do I flag that request for cherry picking? I swear they used to have a magical little git tag for it...
22:35 MTecknology err... github tag*
22:36 MTecknology Maybe it's just one of the pains of not doing it correctly in the first place.
22:36 anotherZero joined #salt
22:37 MTecknology The other PR includes an added feature so it's highly unlikely to make it into 2016.3.x, but this one definitely should.
22:38 tristianc joined #salt
22:41 raspy_ joined #salt
22:41 raspy_ hi has anyone seen salt-cloud stop at  "[INFO    ] Salt interface set to: 178.12.12.17" and not go any further?
22:47 Edgan raspy_: try -l trace  ??
22:48 raspy_ -l trace, let me try
22:54 raspy_ Edgan: http://pastebin.com/jbjFLTL2
22:54 raspy_ just stops at line 20
22:55 raspy_ this only happens when creating hosts in eu_west*
22:55 raspy_ but I can create the hosts no problem in us_west*
22:55 Edgan raspy_: telnet 174.12.12.101 22   see if you can connect, if you can't, probably a security group issue
22:56 raspy_ connected fine
22:56 raspy_ so it does build the host
22:56 raspy_ i see it in aws but it doesnt go through the entire bootstrap process
22:58 Edgan raspy_: why does your pastebin show two different keys?
22:59 raspy_ good question
22:59 raspy_ ill check thx Edgan
23:01 pipps joined #salt
23:02 mpanetta joined #salt
23:02 bluepaladin joined #salt
23:06 flowstate joined #salt
23:10 ninjada joined #salt
23:13 raspy_ Edgan: is there a way to use the latest version of bootstrap
23:13 raspy_ without running a salt-cloud -u?
23:16 jenastar joined #salt
23:24 rsys joined #salt
23:28 Edgan raspy_: don't know
23:38 raspy_ is there a way to downgrade the salt-bootstrap.sh if needed?
23:38 onlyanegg joined #salt
23:40 raspy_ or is it possible to bootstrap a host that is already running on aws?
23:45 filippos joined #salt
23:52 amcorreia joined #salt
23:59 scsinutz joined #salt
23:59 benjiale joined #salt

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