Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-10-20

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

All times shown according to UTC.

Time Nick Message
00:00 bagackiz joined #salt
00:01 bagackiz joined #salt
00:03 bagackiz joined #salt
00:05 bagackiz joined #salt
00:06 bagackiz joined #salt
00:07 cads joined #salt
00:08 bagackiz joined #salt
00:09 naegelin joined #salt
00:10 bagackiz joined #salt
00:11 bagackiz joined #salt
00:13 bagackiz joined #salt
00:15 bagackiz joined #salt
00:16 bagackiz joined #salt
00:18 bagackiz joined #salt
00:20 bagackiz joined #salt
00:21 bagackiz joined #salt
00:23 bagackiz joined #salt
00:23 mapu joined #salt
00:25 bagackiz joined #salt
00:25 bagackiz joined #salt
00:26 mapu joined #salt
00:27 bagackiz joined #salt
00:28 mapu joined #salt
00:29 bagackiz joined #salt
00:31 bagackiz joined #salt
00:31 mapu joined #salt
00:32 bagackiz joined #salt
00:32 shane_ joined #salt
00:34 mapu joined #salt
00:34 bagackiz joined #salt
00:34 TyrfingMjolnir joined #salt
00:35 mapu joined #salt
00:35 bagackiz joined #salt
00:37 bagackiz joined #salt
00:37 mapu joined #salt
00:39 bagackiz joined #salt
00:41 bagackiz joined #salt
00:41 mgw joined #salt
00:41 bagackiz joined #salt
00:42 mapu joined #salt
00:42 schimmy joined #salt
00:43 mapu joined #salt
00:44 bagackiz joined #salt
00:45 mapu joined #salt
00:45 bagackiz joined #salt
00:46 schimmy1 joined #salt
00:47 bagackiz joined #salt
00:48 mapu joined #salt
00:51 mapu joined #salt
00:56 mapu joined #salt
00:57 mapu joined #salt
01:00 mapu joined #salt
01:03 Outlander joined #salt
01:14 Mso150_c_r joined #salt
01:18 elfixit1 joined #salt
01:18 jsm joined #salt
01:18 Mso150_c_r joined #salt
01:21 jsm joined #salt
01:21 Outlander joined #salt
01:22 Mso150_c_r joined #salt
01:28 wnkz joined #salt
01:34 malinoff joined #salt
01:40 Mso150_c_r joined #salt
01:45 ph8 joined #salt
01:48 pdayton joined #salt
01:55 Nexpro joined #salt
02:00 druonysus joined #salt
02:00 druonysus joined #salt
02:13 pdayton joined #salt
02:14 jchen joined #salt
02:18 floWenoL_ joined #salt
02:28 ramishra joined #salt
02:33 belak What are possible reasons for salt '*' grains.ls hanging?
02:33 vejdmn joined #salt
02:33 shane_ Jinja error: cannot concatenate 'str' and 'NoneType' objects
02:33 shane_ {%- set nova_config = {    <======================
02:34 shane_ how is that possibly triggering that error?
02:34 shane_ belak: maybe try with verbose?
02:35 belak shane_: doesn't output anything. just sits there.
02:35 SheetiS joined #salt
02:39 mr_chris joined #salt
02:47 Jeromeyrome2 joined #salt
02:49 pdayton joined #salt
02:54 jalaziz joined #salt
03:00 mosen joined #salt
03:04 schimmy joined #salt
03:05 bhosmer joined #salt
03:10 schimmy joined #salt
03:11 pythopticon joined #salt
03:14 pythopticon joined #salt
03:24 naegelin joined #salt
03:36 TyrfingMjolnir joined #salt
03:38 speckle joined #salt
03:38 crocket joined #salt
03:38 crocket hello guys
03:38 crocket How do you use salt to control TFTP and DHCP to tag a physical server for provision?
03:40 TheThing joined #salt
03:46 pdayton joined #salt
03:48 schimmy joined #salt
03:52 schimmy1 joined #salt
03:55 shookees joined #salt
03:55 shookees joined #salt
03:58 ramishra joined #salt
04:00 gabriel joined #salt
04:07 gabriel Hi. I'm already using saltstack to manage some parts of windows workstations. Does anybody have a possible an approach on updating running applications, e.g. browsers, mail clients...? I'd be thankful if someone could point a way to follow in order to achieve this.
04:09 pdayton joined #salt
04:15 StDiluted joined #salt
04:16 ramishra joined #salt
04:22 crocket Can I provision a virtual machine image with salt?
04:24 basepi Setsuna666: No plans that I know of. Thanks for taking the initiative!
04:25 belak I know I'm missing something... I"m just getting used to it, but how do you choose which states are applied to which hosts?
04:32 speckle left #salt
04:47 belak Is it standard to run salt-master as root, or is it often run as another user?
04:51 scuwolf joined #salt
04:51 felskrone joined #salt
04:51 babilen belak: I started to run it as "salt" as it is easier that way to tailer access to salt independently of access to the root account on the salt master. Some security considerations and the ability to use a different set of dotfiles playes into that too.
04:51 belak Alright
04:52 belak babilen: I assume you leave the minions as root though?
04:52 babilen And target states to your minions in the top file -- http://docs.saltstack.com/en/latest/ref/states/top.html
04:52 babilen belak: Sure, they have to run as root
04:53 babilen (or rather: be able to acquire superuser privileges)
04:53 babilen http://docs.saltstack.com/en/latest/topics/targeting/
04:54 belak I guess the way I think about things is odd.. or wrong for salt... so I need to change something... right now I'm just imagining roles that each host plays (as this is more for some personal servers at the moment)... many will have nginx (and other common configs), but each will only have a few certain nginx hosts...
04:55 belak But it sounds like they're just matched in the main config?
04:55 belak in top.sls
04:58 babilen That is, in my opinion, the better approach. That being said: The "role" approach is quite common and you can definitely implement that based on pillars or grains. I don't like to do the latter because it makes it necessary to manage those grains *directly on the minions*, Preferably you want to manage that with salt as well which means that you can just use whatever scheme you use to target that information to the minions to target the actual states.
04:59 babilen You simply need *some* invariant that you can use to associate hosts with states. Salt is extremely powerful, but if there is nothing you can exploit (e.g. "all webserver end in "web$NUMBER") you will have to maintain lists of hosts and the states that should be matched.
04:59 belak babilen: so, in the master config, I could just match certain hosts and apply certain sls files there?
04:59 babilen yes, exactly
04:59 belak Ah, that's still fair
04:59 babilen Well, the top.sls not the master config, but yeah
04:59 belak Thanks for clearing that up... >_<
04:59 belak Yeah
05:00 vukcrni joined #salt
05:00 babilen In the easiest case you have "some-minion-id": - firststate \n - secondstate ...."
05:00 babilen s/state/sls/
05:01 belak I kinda wish you could apply tags to minions
05:01 belak as another way of matching
05:01 babilen You can, but how would you do that?
05:02 belak Not really sure, but it seems like you'd be able to have a "web" tag which would install nginx and some basic monitoring
05:02 babilen I mean, how would you manage *that* information? You have a catch22 there as you will have to manage which minion has which tag in some way and you end up listing "someminion": tag1, tag2, ...
05:02 belak Ah
05:02 belak Yeah, that's true
05:02 belak At that point you might as well just have a web.sls file... and use that
05:03 babilen I mean you can do that. You can, for example, use pillar (per minion specific dictionaries basically) in which you list roles/tags for that minion and each tag corresponds to a SLS file
05:03 belak Ah I see
05:03 Singularo you can use grains with roles.
05:03 belak Will salt automatically make changes when you change the sls files?
05:03 Singularo to target stuff.
05:03 babilen Some people also use grains (per-minion "static" information like its fqdn, memory, os, ...)
05:04 malinoff If i can add my 2 cents, roles approach is more flexible for autoscaling setup. For example, in AWS you cannot predict a minion ip/hostname thus you cannot predict a minion id. Using roles allows to abstract from concrete host and think in a cluster way
05:04 Singularo http://www.saltstat.es/posts/role-infrastructure.html
05:04 babilen I don't like to use grains as they are, in my opinion, the wrong area in which to keep that information (and you can just as well match on pillars as you can on grains)
05:04 Singularo Not my post, but got me started in using roles.
05:05 belak Ah, nifty
05:05 belak Thanks Singularo
05:05 malinoff And yes, grains are something wrong to store information about roles. Roles must be defined on the master side
05:05 babilen It is the wrong approach
05:05 malinoff Could you explain please?
05:06 babilen blog posts re-iterating that don't help and the more people use that the more ingrained (ba-dum-tsh) this approach will become
05:06 babilen malinoff: I just did, twice :)
05:06 malinoff Do you say that grains are bad to store roles, or that master is bad to store roles?
05:07 babilen grains are
05:07 _gabriel_ belak: your tag idea looks like node groups? :)
05:07 belak _gabriel_: what's that?
05:07 malinoff sure, sorry for misunderstanding
05:07 babilen http://docs.saltstack.com/en/latest/topics/targeting/nodegroups.html
05:07 _gabriel_ belak: u can group nodes in the master.conf
05:08 babilen _gabriel_: And ... how are you going to manage that information? ;)
05:09 _gabriel_ inside the top files? :)
05:10 babilen It always boils down to 1. Keep an exhaustive mapping between minions and tags/roles/nodegroups/... 2. Exploit *some* invariant to associate tags/roles/nodegroups/... based on that (e.g. "all linux boxes should get ...", "all servers whose name end in -web$NUM should get the ...", ...)
05:11 babilen And I'd argue that you should do 2. either in the top file (with the entire power of salt targeting) or keep the information in pillars (your exhaustive list/mapping that is)
05:12 _gabriel_ which one would you rather? top file or pillars?  when not dealing with too many instances.
05:13 babilen I mean if you manage your master's configuration with salt (with, for example, https://github.com/saltstack-formulas/salt-formula) you'd use pillars to get the information used by the formula to the minion running on the master.
05:14 Cidan joined #salt
05:15 babilen _gabriel_: It really depends. Most of our setups can exploit some invariant (e.g. "customer-project-role$NUM" or "xen software only on minions that run as Xen DomU/Dom0", ...), but I also keep manual mappings of minion → tag/role for setups in which we cannot do that.
05:16 babilen I *really* don't like the "manually log into minion, edit it's minion config to add grains, ..." approach. And once you start to dislike that you are back to square one in that you have to manage that information.
05:17 _gabriel_ I see.
05:17 belak Is there a good way to store all the sls files in a git repo?
05:17 belak And config files?
05:17 babilen That doesn't mean that tags/roles are not the right abstraction (and, mind you!, some cloud providers allow you to tag instances), but it simply means that you don't want to have to log into your minions to achieve that association.
05:18 babilen belak: There is: http://docs.saltstack.com/en/latest/topics/tutorials/gitfs.html
05:19 babilen belak: I'd recommend to store your top files in a separate repository if you ever plan to introduce different environments. (if you don't understand that yet, take a short look at what environments are or just start doing it and thank me later when you do)
05:19 babilen (or curse me for having to maintain a different repository even though you never started to use environments)
05:19 belak haha
05:20 belak babilen: ah, the different file_roots?
05:21 shookees joined #salt
05:21 babilen "base", "dev", "qa", ... If you use GitFS branches will be treated as environments and, as you likely want to have different top files and still want to be able to use "normal" git merging and branching workflows, you manage your top.sls in a different repo
05:22 babilen Otherwise the top.sls would be merged across all branches
05:22 crocket babilen, hey
05:22 babilen hmm?
05:22 crocket babilen, Does salt support provisioning a virtual machine image as packer does?
05:22 babilen No, it doesn't
05:23 crocket I want to provision a libvirt image.
05:23 crocket babilen, What does http://docs.saltstack.com/en/latest/ref/cli/salt-cloud.html do, then?
05:23 babilen You can use salt-virt to manage libvirt images
05:23 crocket "Provision virtual machines in the cloud with Salt"
05:23 bmcorser joined #salt
05:24 babilen crocket: salt-cloud is based on support from specific cloud providers and does not implement image creation, PXE booting, installation, ... for hyou
05:24 crocket babilen, How do you create libvirt domains with packer, then?
05:24 crocket packer still hasn't assimilated https://github.com/mitchellh/packer/pull/1330
05:24 ramishra_ joined #salt
05:25 babilen I am well aware of that (you might want to check who is discussing in there)
05:25 babilen But you don't need that to build libvirt images with packer. That is only support that allows packer to build vagrant boxes.
05:26 crocket babilen, How do you build a libvirt domain from a qemu image and a set of VM parameters?
05:26 malinoff Is it still a talk about salt?:)
05:26 babilen malinoff: crocket has been borderline off-topic for a while
05:27 babilen crocket: You can use virt-install (for example), but seriously: salt is, as of now, not doing that and we've been over all the tools already. Most companies who provide cloud service had to write software for their platform ...
05:28 ndrei joined #salt
05:28 crocket I'm sad that they're not sharing.
05:28 babilen I mean I'd love to fork GCE, rackspace, AWS, linode, gandi, ... code and just use it in salt, but it's not there yet
05:28 belak babilen: wait, so does it just pull the top.sls from master then?
05:29 babilen And these companies have a huge interest in not making *that* available.
05:29 belak By default, with one repo?
05:29 babilen belak: top.sls will be merged across all branches
05:29 belak ah, ok
05:30 crocket babilen, The best we can do is to use salt to orchestrate packer and other tools?
05:30 babilen belak: Yes, that is braindead. But then it is more sensible than having to maintain different top.sls files in each branch. That's even worse as you cannot use branch/merge workflows in git anymore. This is why you keep it in a separate repo (just do it, repos are cheap and you don't limit your options in the future)
05:31 babilen crocket: yes
05:31 crocket Or foreman
05:31 war2 joined #salt
05:31 crocket Foreman_salt adds salt support to foreman.
05:32 crocket Or cobbler.
05:34 linjan joined #salt
05:44 mgw joined #salt
05:45 _gabriel_ is there a way to close applications remotely on windows salt minions?
05:50 oyvjel joined #salt
05:52 crocket minion
05:53 crocket Is the master in charge of his dominion, then?
05:54 kermit joined #salt
06:00 crocket Are your minions ok?
06:00 colttt joined #salt
06:02 crocket It seems zeromq 4.x can bring security to salt.
06:02 crocket zeromq has NaCl.
06:04 techdragon joined #salt
06:08 notpeter_ joined #salt
06:13 crocket Is it beneficial to run salt as non-root?
06:14 Singularo babilen, you dont need to login to each minion to set grains, this works -   salt 'xxxx' grains.setval roles [ssh,zabbix,salt,logstash]
06:14 Singularo And updates the client from the central server.
06:16 ramishra joined #salt
06:17 crocket Is it ok to run salt-minion and salt-master on the same machine?
06:21 crocket The minion is under my command.
06:22 honestly yes.
06:26 mg__ joined #salt
06:26 crocket honestly yes?
06:26 cberndt joined #salt
06:26 crocket That's a good nickname.
06:30 elfixit1 joined #salt
06:30 * honestly bows
06:31 mechanicalduck_ joined #salt
06:38 flyboy joined #salt
06:42 crocket Salting my system
06:50 Mso150 joined #salt
06:51 trikke joined #salt
06:58 mgw dstufft: did you figure out the answer to your question?
06:58 mgw re ext_pillar
07:02 mgw dstufft: if so... you want to look at the "pillar_dirs" option in master config
07:02 mgw it takes a list of dirs that contain those .py files
07:03 che-arne joined #salt
07:09 oyvjel1 joined #salt
07:16 Mso150_c joined #salt
07:19 Mso150 joined #salt
07:22 trikke joined #salt
07:26 cberndt joined #salt
07:27 _gabriel_ left #salt
07:30 xsteadfastx joined #salt
07:30 chiui joined #salt
07:31 rattmuff joined #salt
07:43 TyrfingMjolnir joined #salt
07:43 ramishra joined #salt
07:45 superted joined #salt
07:47 slav0nic joined #salt
07:47 slav0nic joined #salt
07:49 cberndt joined #salt
07:49 spo0nman joined #salt
07:51 bahadir joined #salt
08:01 lcavassa joined #salt
08:04 crocket Well...
08:04 crocket I don't like salt file system at the first sight.
08:05 crocket A centralized file system where you pour all configurations?
08:05 crocket no good
08:05 jhauser joined #salt
08:07 goodwill you mean the git thing?
08:07 crocket goodwill, It's the opposite of git.
08:07 crocket Salt file system in /srv/salt
08:07 crocket salt://
08:07 goodwill ah
08:08 pviktori joined #salt
08:15 ramishra joined #salt
08:17 crocket Does vim refer to /etc/vimrc in sls files?
08:18 crocket Perhaps, I should put /srv/salt in git
08:18 rattmuff crocket: we have /srv under version control, works nicely with some hooks
08:19 crocket rattmuff, Why do you put /srv under one repository?
08:19 rattmuff crocket: to keep pillars, runners and other stuff in the same repos
08:19 crocket hmm....
08:19 crocket ok
08:19 crocket anyway
08:20 crocket rattmuff, I think /srv should exist for each machine.
08:20 xmj ??
08:20 rattmuff crocket: /srv is only on the master no?
08:21 marv__ well we used to put things to /var/lib/ or even /etc/pillarsomething
08:21 crocket xmj : I want each machine to have separate settings.
08:21 xmj crocket: why *wouldnt* you do things in /srv, in one centralized place?
08:21 xmj crocket: i'm not sure you understand salt just yet
08:21 crocket xmj : I'm reading about it.
08:21 xmj especially how top.sls assigns pillars/states to machines.
08:22 xmj the idea is that you don't need each machine to have different salt configuration *files*
08:23 oyvjel joined #salt
08:23 crocket xmj : That sounds too centralizing.
08:24 xmj yes, that's the point
08:24 xmj why *wouldn't* you do that
08:25 crocket xmj : That may complicate set up if the master goes off.
08:25 crocket I want minions to thrive independently.
08:25 crocket xmj : That's single point of failure.
08:26 superted Salt allows you to manage thousands of systems with minimal configuration. You can have different configs within that but if your configuring every server manually then i don't think a centralised config system is what you want.
08:26 superted And salt supports multi-master if your concerned about failure
08:26 crocket superted, It also supports masterless mode.
08:26 crocket hmm....
08:26 xmj yes but why would you do that.
08:27 superted good to know crocket :)
08:27 crocket xmj : If I want to build a virtual machine image via packer, masterless mode is beneficial.
08:27 xmj what's salt got to do with that ?
08:27 crocket Solo operations have their use cases.
08:27 crocket xmj : packer uses configuration management tools to configure OS.
08:28 xmj yes, well - by the time we're talking about managing several / dozens / hundreds of server you'll find that you woulnd't want to rely on your brain to remember who does what
08:28 crocket xmj : In that case, master would be beneficial.
08:29 superted ding ding
08:29 crocket For setting up and tearing apart development virtual machine imagees, solo opeartions are beneficial.
08:29 crocket Ephemeral virtual machine images.
08:30 xmj yes but why would you use salt for that
08:30 crocket xmj : Is ansible better for that?
08:31 crocket ansible has local mode.
08:31 xmj you don't need to install ansible on the thing you're running ansible operations on, do you?
08:31 crocket xmj : I'm exploring options.
08:31 crocket xmj : Why would I do that? to learn.
08:31 crocket I'm a beginner.
08:31 superted Chef has chef-zero or chef-solo for the same reasons. Perfect if you don't want centralised config. If you do, use a master based system
08:32 crocket xmj : Why does salt has materless mode?
08:32 Jellyfrog if you use salt-ssh and git you could clone the repo to any computer and run it from there
08:33 Jellyfrog with the right firewall-rules..
08:33 crocket Is salt inappropriate for solo operations?
08:33 crocket Or are there better tools?
08:34 rattmuff How much RAM would you recommend for handling about 100 minions?
08:38 superted Not sure on the official line but we have 4GB for 600 vm's
08:38 superted no swap used but it's using the available memory well enough
08:39 N-Mi joined #salt
08:41 crocket 4GB for 600.
08:42 crocket Quite a lot
08:42 crocket Each minion consume about 7MB in the master.
08:42 crocket The master is burdened.
08:43 babilen crocket: You are aware of GitFS aren't you?
08:44 xsteadfastx joined #salt
08:44 crocket babilen, I didn't know, but it sounds like a git file system.
08:45 crocket By the way
08:45 crocket How do I get indentation and syntax highlight of salt files in emacs?
08:46 babilen crocket: You might want to read up on it. May I also recommend to browse the entirety of the salt documentation (just chapter headers) to get an idea of what functionality you can use? It will be hugely beneficial in the long run for you if you come across certain problems
08:46 babilen crocket: Salt comes with vim settings, I'm sure there is somthing for emacs users too
08:47 babilen crocket: Did your search not yield any results?
08:47 crocket babilen, not yet
08:48 rattmuff crocket: thanks
08:48 crocket babilen, If salt files are yaml files, I should just turn on yaml mode in emacs.
08:48 crocket rattmuff, for what
08:48 babilen crocket: Sure, but you might want to have some extra syntax highlighting for salt specific elements
08:49 babilen And you can always learn vim ;)
08:51 kbyrne joined #salt
08:51 crocket babilen, Is it ok to use salt in masterless mode?
08:53 crocket I feel dizzy
08:53 babilen in .13 sure
08:53 babilen Why wouldn't it?
08:53 babilen You won't be able to take advantage of all features and you might want to consider looking into salt-ssh with 2014.7, but then there is noting necessarily wrong with that.
08:55 crocket Salt documentation is rather long, fragmented, and confusing.
08:56 giomandaz joined #salt
08:56 babilen Hnce my suggestoin to take a look at all chapter headers
08:56 che-arne joined #salt
08:56 babilen And it really isn't that bad
08:57 crocket babilen, The time it takes it bad.
08:58 giomandaz good morninng guys
08:58 giomandaz can someone propose some ideas on this?:
08:58 giomandaz http://justpaste.it/hn0v
08:59 giomandaz i know there is a typo erro on line 5 but yuo can ignore it
09:00 giomandaz question is, how do you tell a for loop, that if the return value of the pillar is not defined or empty...do nothing?
09:06 DolourousEdd joined #salt
09:07 colttt joined #salt
09:08 bhosmer joined #salt
09:09 ndrei joined #salt
09:11 ndrei joined #salt
09:12 ndrei joined #salt
09:12 bra1nout joined #salt
09:13 ndrei joined #salt
09:14 bra1nout Hey guys. Need some help with jinja :S
09:14 bra1nout I was reading manual when some mistery appeared.
09:14 bra1nout {% for mnt in salt['cmd.run']('ls /dev/data/moose*') %}
09:14 bra1nout - mount: /mnt/moose{{ mnt[-1] }}
09:14 bra1nout - file: /mnt/moose{{ mnt[-1] }}
09:15 bra1nout What is mnt[-1]? In Python it is the one from the end element of array
09:15 bra1nout but what is the reason to put the last element and only it?
09:16 babilen giomandaz: You can return an empty list from the pillar lookup. Something like "salt[pillar.get]('foo', [])"
09:18 CeBe joined #salt
09:19 giomandaz thanks babilen will try it now
09:20 giomandaz still get same error  Rendering SLS "base:state1" failed: Jinja error: 'NoneType' object is not iterable
09:20 giomandaz :-S
09:22 ramishra joined #salt
09:25 jrluis joined #salt
09:26 viq snmpd config, on some machines we do stick an 'extend' in there, on some we stick another 'extend', on some we don't put it in at all. What do you think would be a good way to do it? I was thinking file.accumulated, but the template listed there yells if there is no accumulator defined (the "nothing added" case)
09:27 k3sy3r joined #salt
09:31 CycloHex joined #salt
09:33 SheetiS viq: could do a file.managed on the config and then store the specific information in a pillar for whether a particular machine gets an extend or not.  shouldn't be too crazy to jinja it up if the base is fairly similar.
09:36 viq Well, pillar is ungainly because I can't for example say that this machine gets a pillar snmp.extend1 and that one gets snmp.extend1 AND snmp.extend2
09:37 viq Anwyay until we get 2014.7 available with pillar merging
09:42 SheetiS If you need dict merging in jinja right now, I saw this pasted about a month ago and is a pure jinja solution that could be imported as a macro for use in your state: https://gist.github.com/michaelkrupp/4e83ab08ffd32c8ca387.  I'm not sure if merging is how I would handle it (as I have no point-of-reference as to how your configs look right now), but it definitely gives an option if you needed one.
09:43 lcavassa joined #salt
09:47 martoss joined #salt
09:47 viq thanks, looking
09:48 martoss1 joined #salt
09:48 VSpike bra1nout: I'd guess the devices are called /dev/data/moose1 /dev/data/moose2 ...
09:48 VSpike bra1nout: so it's getting the last character of the string (i.e. the digit at the end) and appending it to /mnt/moose
09:50 bra1nout oh, god. Reallt, I am so stupid
09:50 bra1nout thanks VSpike!
09:50 SheetiS I need to head out for now.  Y'all have a good evening/morning/afternoon/whatever it is where you are.
09:52 Outlander joined #salt
09:54 DolourousEdd1 joined #salt
09:54 xsteadfastx joined #salt
10:13 linjan joined #salt
10:13 robinsmidsrod joined #salt
10:19 peters-tx joined #salt
10:22 ggoZ joined #salt
10:23 rjc joined #salt
10:25 xsteadfastx joined #salt
10:26 pythopticon joined #salt
10:26 fredvd joined #salt
10:31 linjan joined #salt
10:36 pythopticon left #salt
10:37 ndrei joined #salt
10:38 ndrei joined #salt
10:39 VSpike I want to run a cmd.run state that requires an XML file as input. Is there a way to copy the XML file to a temporary location?
10:39 VSpike This is Windows BTW, so I can't assume /tmp/
10:40 VSpike Sadly, %TMP% is not available as a grain on Windows machines
10:40 wvds-nl joined #salt
10:41 VSpike Hm, I wonder if could get away with cmd.script using the XML file as the script and the command I want to run as the shell?
10:42 rattmuff VSpike: I guess you could also make it "file.managed" and then run your command and then "file.absent"
10:43 wvds-nl hello folks. im totally new to salt. at the moment im figuring out how to manage configs of windows servers with salt. i've installed the windows minion but the strange part salt can only be installed in c:\salt is there e specific reason for installing in c:\salt instead of c:\program files?
10:45 ndrei joined #salt
10:46 ndrei joined #salt
10:47 VSpike rattmuff: looks like that's going to be the only way
10:47 ndrei joined #salt
10:52 ndrei joined #salt
10:54 ndrei joined #salt
10:55 simonmcc joined #salt
10:56 duncanmv joined #salt
10:56 duncanmv Hi, if I want to use the salt apis from outside python, is there something else than the net-api modules?
10:57 duncanmv (salt-api is fine, just double checking it is the most comfortable to use from Java)
11:01 ammbot joined #salt
11:02 wr3nch joined #salt
11:05 elfixit joined #salt
11:13 yomilk joined #salt
11:14 whiteinge duncanmv: that's probably your only bet. I haven't come across java bindings for salt.
11:15 whiteinge The rundeck project has salt-api integration written in java, if that's useful.
11:22 CycloHex If I run the following command salt'ID' grains.get ipv4, I get a list of the ips, I only want the first record. How can I achive this? I tried grains.get ipv4['1'], ipv4:1, ipv4{'1'}
11:22 duncanmv whiteinge: we can do them :-)
11:23 duncanmv whiteinge: is the only option for those to go via http?
11:23 duncanmv whiteinge: oh, rundeck, haven't seen it... looking
11:24 whiteinge duncanmv: that would be cool if you're up for the work. netapi modules are open-ended. Any transport that would be comfortable in java is fair game.
11:25 duncanmv whiteinge: we are at Hackweek this week at SUSE, I am trying to integrate salt with Spacewalk
11:26 whiteinge Hey, cool
11:26 tomspur joined #salt
11:26 whiteinge The only netapi modules written to date use http. But they were written with anything in mind.
11:27 duncanmv we are doing a java -> http right now...
11:27 duncanmv as we use xml-rpc extensively in Spacewalk, I was thinking going that way too
11:27 duncanmv right now the biggest questions are about design. Spacewalk operates on a database, mostly on cached data.
11:28 duncanmv I have to think how to cache the data from salt.. otherwise doing realtime queries (eg, list of systems, list of packages for a system) would not work for a dead system
11:29 duncanmv right now all this caching is initiated from the client (eg, package list is uploaded as a hook to a package install, if rpm db changed)
11:29 duncanmv but I could as well use cache.grains, not sure how durable is that, and that stil don't tell me the package list
11:34 whiteinge Yeah, the master only caches grains, mine, pillar. Anything else will need to be done yourself.
11:34 whiteinge The master's cache is reliable though.
11:35 diegows joined #salt
11:36 duncanmv whiteinge: we are thinking about connecting to the websocket, listen for events. If we can somehow get for example, an event of an accepted key, then we register that system with Spacewalk
11:37 duncanmv whiteinge: if we get an event that a pkg.install happened, we can call pkg.list, and store the current list in the db
11:37 whiteinge That would be very cool actually. An alternative to websockets is the SSE stream.
11:38 whiteinge If you don't need constant bidirectional communication
11:39 duncanmv with one websocket between spacewalk and the master is enough, we can run it as a thread or a separate service....
11:39 duncanmv omg, you just gave me an idea
11:39 duncanmv I can even do it in python... half of spacewalk is in python and registration is currently python
11:39 duncanmv I can have a simple service listening events and populating the db
11:40 whiteinge Ah, very nice.
11:40 xsteadfastx joined #salt
11:41 CycloHex I'd like to add my freshly deployed minions to my nagios server on which I run check_mk. Is this best accomplished with a bash-script (because it needs to create a file, append to it, make an invetory of the new file, reload check_mk. This has to happen on the monitoring server, but this server has no knowledge of my newly deployed server
11:41 whiteinge Then Spacewalk would access that pre-populated db and also initiate occasional commands?
11:42 duncanmv exactly, for the inventory, calculation of the state of patches for systems, it would operate with cached data
11:42 duncanmv for initiating actions, it would directly send to salt
11:42 viq duncanmv: https://github.com/felskrone/salt-eventsd
11:42 duncanmv as well as to query the state of actions
11:42 whiteinge Neat
11:43 whiteinge Gotta run afk. Back later.
11:43 whiteinge CycloHex: look at Mine or even the Reactor for doing that via Salt.
11:44 viq duncanmv: that projects ties into what you're saying, so might be useful to be used or for inspiration
11:44 duncanmv viq: looking, thanks
11:44 CycloHex whiteinge: The Reactor doesn't work well.. I had salt/cloud/*/created in my reactor conf, but eventhough I saw the tag with that line, it didn't trigger the sls I linked ot it :s
11:44 CycloHex But I'll look into Mine! thanks
11:45 zions joined #salt
11:45 zions Hi
11:45 zions I'm using salt-cloud to saltify minions.
11:45 zions For some reason I have a setup where /tmp is mounted using noexec option.
11:46 zions Salt-cloud doesn't handle that gracefully...
11:46 whiteinge Set your logging to debug to double-check for errors in the sls file.
11:46 whiteinge Sorry, back later.
11:47 VSpike Can anyone suggest how I can avoid the "Recursive requisite found" error here https://bpaste.net/show/371d4ef3a5e5 ?
11:47 CycloHex ofc not, zions.. the salt-cloud copies your script to /tmp/salt and runs it form there.
11:48 zions I saw it actually tries to run from /tmp/.saltcloud
11:48 viq Maybe you could set TMP to a different path beforehand?
11:49 CycloHex ye, kinda the path I ment :d
11:49 CycloHex I'm looking in the bootstrap-script now, but cannot immediatly locate where it says to copy everything to tmp
11:50 zions viq, cyclohex - Ideally I'd do that. However, I noticed that even though /tmp is mounted as no exec, one may still run the deploy.sh script using bash/sh
11:50 viq https://raw.githubusercontent.com/saltstack/salt-bootstrap/stable/bootstrap-salt.sh seems to have /tmp hardcoded
11:51 duncanmv viq: no releases? master is it...?
11:51 zions When I look at the debug log, it looks something like this:
11:51 zions [DEBUG   ] SSH command: 'ssh -t -t -oStrictHostKeyChecking=no -oUserKnownHostsFile=/dev/null -oControlPath=none -oPasswordAuthentication=no -oChallengeResponseAuthentication=no -oPubkeyAuthentication=yes -oKbdInteractiveAuthentication=no -i /srv/salt/custom/cloud-ssh root@XXX.XXX.XXX.XXX /tmp/.saltcloud/deploy.sh'
11:51 viq duncanmv: that salt-eventsd? I can't really tell you, I know of it, but haven't had a chance to look closely at it
11:51 ggoZ joined #salt
11:52 viq zions: maybe in advance prepare a symlink of /tmp/.saltcoud to somewhere else?
11:53 viq duncanmv: ask felskrone ;)
11:53 zions viq, that may work.
11:56 felskrone duncanmv: yet, its just master :-)
11:56 ginger_tone joined #salt
11:57 ginger_tone Hi :) Am getting "No minions matched the target. No command was sent, no jid was assigned " error.
11:57 ginger_tone when am using a single hostname like :
11:58 ginger_tone # salt 'my_hostname_1' cmd.run "ls -l"
11:58 felskrone duncanmv: i pushed it to github when it was almost done and since then there have only been minor changes. but i guess i could "release" a version some time soon :-)
11:59 viq ginger_tone: salt-key -L  - do you have in there a host named exactly 'my_hostname_1' ?
11:59 ginger_tone but for all minions its working. That is salt '*' cmd.run "ls -l"
11:59 flyboy82 ginger_tone: run "salt-key -L" and check if your minion is reported in the Accepted area. Your minion's minion_id may differ from its hostname
12:00 flyboy82 ginger_tone: salt-master communicates with minions by minion_id, not by hostname
12:00 k3sy3r is there a simple way to make minion_id == hostname ?
12:00 ginger_tone Ya my minion is accepted in accepted area..
12:01 ginger_tone So how to check ? if minion_id differs from its host name ?
12:01 viq k3sy3r: "echo `hostname -f' > /etc/salt/minion_id && service salt-minion restart" - of course then you have to remove on master the old ID and accept the new
12:02 flyboy82 ginger_tone: you don't have to check for that. Just issue a "salt '<minion_id>' test.ping" where minion_id : the id that's reported in salt-key -L
12:02 viq ginger_tone: 'salt * grains.item fqdn'
12:02 flyboy82 and if it's up, then run cmd.run
12:02 ginger_tone how to get the minion id ?
12:03 viq ginger_tone: 'salt * grains.item fqdn id'
12:03 ginger_tone ok
12:04 bhosmer joined #salt
12:05 ginger_tone @viq: its working
12:06 viq what is?
12:06 ginger_tone but what was wrong am not getting ?
12:06 ginger_tone salt 'fractal-dC5e.fractal.lan' cmd.run "ls -l /tmp"
12:06 ginger_tone am able to get output.
12:07 viq then maybe the node was not responding. I find it useful to pretty much always run salt with -v
12:07 ginger_tone before I issued - salt '*' grains.item fqdn
12:07 ginger_tone ok
12:07 viq or check with "salt-run manage.status"
12:12 intellix joined #salt
12:13 apergos hello folks, I've just updated our master to 2014.1.11, precise,  keeping old config and keys, and the master is taking minutes to look though the 3000 job cache entries; really really slow, in the meantime no minions can reach it of course because it hasn'tfinished starting up
12:13 apergos any hints?
12:14 AviMarcus joined #salt
12:14 apergos I"m running it with -l debug and so far it's produced exactly two lines of output, bout reading the config file.  no worker threads started or any of that.
12:25 johtso Does having a watch statement make a require statement redundant?
12:28 apergos it looks like this: https://github.com/saltstack/salt/issues/7319  but was the cause ever found?
12:40 viq johtso: yes
12:43 rattmuff Should I store minion specific configuration files in a pillar (srv/pillar) or only the path to the file which is placed under /srv/salt?
12:44 rattmuff Perhaps I should use templates with data only in pillar and map it to state :/
12:45 rattmuff Since I can't see a way to serve/include files from the pillar path
12:47 UtahDave joined #salt
12:48 jeddi joined #salt
12:51 viq rattmuff: what exactly are you trying to do?
12:52 viq rattmuff: also, you can have something like this: https://pbot.rmdir.de/NVbUx94oGn3-5_rviWNIGA
12:53 dude051 joined #salt
12:54 rattmuff I'm trying to add minion unique configuration files as pillar values rather than using salt://
12:54 rattmuff The files might have sensitive data in them
12:55 vejdmn joined #salt
12:55 rattmuff I was hoping I could simply include a file from /srv/pillar rather than /srv/salt
12:55 viq rattmuff: no, pillars are variables, not files
12:57 rattmuff viq: yes, I wish it was also files :P. Will have to try another solution :)
12:57 viq rattmuff: http://garthwaite.org/virtually-secure-with-openvpn-pillars-and-salt.html  ?
12:58 rattmuff viq: ooh, looks interesting. Thanks.
12:59 bhosmer joined #salt
13:00 jsm joined #salt
13:02 dude051 joined #salt
13:05 vbabiy joined #salt
13:05 jslatts joined #salt
13:06 miqui joined #salt
13:07 bhosmer joined #salt
13:07 bhosmer joined #salt
13:10 racooper joined #salt
13:11 bhosmer_ joined #salt
13:13 bahadir joined #salt
13:14 ozzzo1 joined #salt
13:15 mapu joined #salt
13:21 cpowell joined #salt
13:24 mpanetta joined #salt
13:28 mndo joined #salt
13:30 krak3n` joined #salt
13:32 nitti joined #salt
13:33 intellix joined #salt
13:35 ozzzo joined #salt
13:36 TyrfingMjolnir joined #salt
13:38 mpanetta_ joined #salt
13:46 jsm joined #salt
13:46 glyf joined #salt
13:52 ozzzo joined #salt
13:52 xsteadfastx joined #salt
13:53 mike25de left #salt
13:56 bhosmer joined #salt
14:01 thedodd joined #salt
14:01 giannello joined #salt
14:02 KyleB joined #salt
14:03 crocket Is it easy to run salt-minion as non-root?
14:03 mechanicalduck joined #salt
14:03 viq crocket: no
14:03 crocket viq, Is it easy to run any configuration management tool as non-root?
14:03 viq crocket: don't think so. They all need to make changes to system that requires root privileges
14:05 Jellyfrog salt-ssh + agent forward + pam sudo agent forwarding should work
14:05 viq You may try to give them privileges via sudo... But at that point I don't think it counts as "easy"
14:06 housl joined #salt
14:06 crocket sudo is not easy.
14:06 crocket It seems salt is our superpower now.
14:07 Jellyfrog you kinda want the best of all worlds, without any pain
14:08 viq chmod 777 everything!
14:08 crocket Jellyfrog, What's wrong?
14:08 shane_ is there a way to explicitly process things in order in top.sls ?
14:08 shane_ or should it be doing that by default?
14:08 crocket I installed salt-master and salt-minion on the same machine.
14:08 crocket But, salt-master can't see salt-minion.
14:08 crocket Why not?
14:09 Jellyfrog because you havent setup the dns
14:09 Jellyfrog i would guess
14:09 viq or didn't accept the key
14:09 racooper shane_,  http://docs.saltstack.com/en/latest/ref/states/ordering.html
14:10 shane_ that setting doesnt seem to change anything, maybe im on the wrong version or something
14:10 shane_ salt 2014.1.11
14:11 shane_ state_auto_order that is
14:12 racooper did you see the paragraph that says "the requisite system will overrride the ordering defined in the files"?
14:12 wnkz joined #salt
14:14 CycloHex Are all available modules available in reactor files? eg: I'd like to run a reactor when I have an event with tag salt/cloud/*/created, this event should touch a file on a tgt, then append some text to it
14:14 vejdmn joined #salt
14:16 _jslatts joined #salt
14:16 ekristen joined #salt
14:17 ekristen anyone run into issues installing pygit2 on ubuntu 14.04 for use with salt-stack?
14:17 ekristen I keep getting compilation errors
14:17 crocket If I watch httpd.conf, then are config files included in httpd.conf watched, too?
14:17 crocket no?
14:18 KennethWilke joined #salt
14:18 dude051 joined #salt
14:19 felskrone joined #salt
14:19 shane_ yeah state_auto_order definitely not working
14:20 shane_ racooper: yes i saw that, i dont have any requisites defined
14:20 CycloHex Can I tell a reactor to run a script which should run on the salt-master itself? the script contains salt '*' file.... commands
14:20 VSpike I still don't feel like I've quite grokked YAML. How can I do a list of dictionaries, where one of the dictionary values is itself a list?
14:21 racooper ok. if it's not something as simple as that, then afraid I don't have any additional assistance right now.  hopefully a dev can answer for you.
14:22 viq crocket: "config files included in httpd.conf" - what do you mean?
14:22 VSpike Salt states often seem to contain a list of single-item dictionaries, unless I'm mistaken. Which seems odd, but looks nicer in YAML
14:25 quantumriff joined #salt
14:27 ramishra joined #salt
14:29 crocket viq, httpd.conf can include other config files.
14:29 to_json joined #salt
14:29 crocket If I make salt watch httpd.conf, does it also watch files included by httpd.conf?
14:29 KyleB Hello everyone.  I'm a little confuse about the cmd state in regard to the creates option.  I'm using salt 2014.1.10 (Hydrogen), docs say the creates option was added to version 2014.7.0.  Latest version is 2014.1.11.  My command runs regardless of the  creates statement. Anyone run into this?
14:30 iggy KyleB: salt generally ignore extraneous data in states
14:31 viq crocket: "watch" in salt speak means "if salt changes this file, do those other things"
14:31 viq In other words: no.
14:32 KyleB I'm not sure I understand iggy.  I tell it  - creates: '/usr/local/lib/boost_1_47_0/libboost_filesystem.a' , which exists after the first run.  Should it then not run that command again?
14:33 iggy KyleB: if you're using 2014.1.x, it doesn't matter, "creates"
14:33 iggy isn't doing anything
14:33 schimmy joined #salt
14:34 crocket Do minions report to masters regularly?
14:34 KyleB so I need to find the 2014.7.0....that a beta line or something?  And thanks for clearing my understanding
14:34 viq crocket: what do you mean by "report"? They keep the connection alive constantly. Do they send some status updates, reports? No.
14:35 viq KyleB: release candidate, they say they expect a release "early this week"
14:35 crocket viq, puppet has modules for sending back reports.
14:35 crocket I guess salt also has it.
14:35 KyleB fantastic, thanks viq!
14:35 viq crocket: define "reports"
14:36 schimmy1 joined #salt
14:36 crocket viq : foreman reports
14:36 crocket viq, foreman receives reports about system information.
14:36 viq crocket: foreman has some support for salt, you could check what they do and don't do
14:36 crocket Each puppet slave sends reports per 30 minutes.
14:37 viq crocket: as in facts?
14:37 crocket viq, I don't know puppet well enough. I don't know puppet facts.
14:38 viq cpu version, IPs, RAM, serial number, etc
14:39 superted_ joined #salt
14:40 crocket yes
14:40 crocket like that
14:40 viq Then you have that in grains
14:40 micah_chatt joined #salt
14:40 crocket If salt-minion runs in masterless module, does it need top.sls as well?
14:40 viq For which you can query live, or cache in mine
14:40 viq Depends if you want to run highstate, or just specific states
14:42 DolourousEdd joined #salt
14:43 oyvjel joined #salt
14:46 gildegoma joined #salt
14:46 crocket state.highstate seems to require a master.
14:47 glyf joined #salt
14:47 ek6 joined #salt
14:48 viq no
14:49 hasues joined #salt
14:50 crocket I don't quite understand top.sls yet.
14:50 viq http://docs.saltstack.com/en/latest/ref/states/top.html
14:51 viq environment: 'match_something: - assigned.states
14:51 viq erm
14:51 viq environment: 'match_something': - assigned.states
14:51 viq crocket: any particular reason why you're running masterless?
14:52 scoates is there a programmatic way to get __salt__ (or the other dunders) instead of expecting magic injection? something like `import salt` `salt.get_dunder('__salt__')` would be great to avoid the "where did this dunder come from?" logical questions
14:55 _prime__ joined #salt
14:56 djstorm joined #salt
14:57 ajolo joined #salt
14:57 jalbretsen joined #salt
14:57 jaimed joined #salt
14:57 crocket Do I need a top file even in masterless mode?
14:58 viq Do you want to run highstate?
14:58 ggrieves joined #salt
14:58 viq If you do, then yes.
14:58 lcavassa joined #salt
14:59 Heartsbane joined #salt
14:59 Heartsbane joined #salt
15:00 crocket viq, What is highstate?
15:02 viq "make sure that the minion is in the state I defined by assigning states to it"
15:04 nahamu is Justin from SaltStack in here?
15:07 conan_the_destro joined #salt
15:07 conan_the_destro joined #salt
15:07 CycloHex I'm trying to watch the job with tag salt/cloud/*/destroyed in my reactor.conf, this won't respond however, the tag salt/minion/*/start does react. Is it possible that cloud and reactors just don't work well of is there someone here that has this running?
15:10 crocket left #salt
15:10 Ozack1 joined #salt
15:10 gmcwhistler joined #salt
15:11 eriko joined #salt
15:12 UtahDave joined #salt
15:13 prosper_ joined #salt
15:15 ggrieves Hi, can someone clear up some confusion I have from the documentation?   in one example they define paths in the file_roots, ('base', 'dev', 'qa' etc) then they put a topfile in each root, and each topfile defines which minions belong to which class. The base root has a top that defines which minions are in base and dev, but dev root has a top that also defines whats in base and dev?  isn't there be a "top" topfile that defines this
15:16 _prime__ is the the top.sls file for states, or for pillars?
15:16 viq ggrieves: they get merged, or something to that effect
15:16 viq _prime_: yes to both, you need two separate ones, one for each
15:17 _prime__ for states, you only need a top.sls file in base, but iirc for pillars each environment needs its own top.sls file.
15:18 _prime__ i'm not familiar with the example you're using, bet we opted to keep it simple for states and just have the one top.sls file in base (we have empty top.sls files in dev/qa/pilot/prod to appease git when we make changes to base, but we keep all our defs in base).
15:18 _prime__ But, for pillars we had to have a top.sls for each environment, defining the pillars in that particular environment.
15:19 Gareth joined #salt
15:20 ggrieves joined #salt
15:20 TheThing|24-7 joined #salt
15:20 _prime__ ggrieves: did that make sense to you?  It's a bit non-intuitive at first.
15:23 ggrieves yeah I'm trying to parse it, thanks, I just had to reconnect to irc.   So I can be in, lets say /dev root's topfile, but define some minions to be 'base' so that when I address all base, those are also included? or am I just assigning some .sls in the /dev root to those called base that's different from the ones actually in /base root?
15:24 UtahDave I'd actually recommend having just one top file for your states and one top file for your pillars.
15:26 ggrieves yeah, I'm just trying to make sense of this example from the docs (http://docs.saltstack.com/en/latest/ref/states/top.html)  but maybe I need to just start simple like you said with one topfile
15:26 UtahDave To simplify things, it helps to clarify a couple things about top files and evironments
15:27 _prime__ Utah Dave is right, one topfile is the way to go (keeps everything simple)
15:27 UtahDave generally, unless you set a minion config option, a minion will pull any sls file from any spot in your top file where the minion matches
15:28 UtahDave a Salt "environment" really just refers to the location of files on your master.
15:28 mgw joined #salt
15:28 UtahDave If the only difference between your different system environments is swapping out some variables, then I recommend just using one Salt environment (the base environment ) for your states
15:28 ggrieves ok, so what I think you're saying is that once file_roots are defines, all the topfiles are treated kinda like one big topfile?
15:29 UtahDave and then use pillar environments to collect the correct values for your system environment
15:29 UtahDave ggrieves: Yes, if you have multiple top files, they get compiled into one big python dictionary
15:29 ekristen UtahDave: I keep getting “GitError: This transport isn't implemented. Sorry” when trying to use the 2014.7 branch with pygit2 provider
15:29 perfectsine joined #salt
15:31 ggrieves ok, I will heed that advice.  I've just been tasked with selecting a config mgmt solution for my company, and I've decided salt was the best for our needs (primarily windows servers) so I'm figuring this out on my own here as fast as possible
15:32 FL1SK joined #salt
15:32 UtahDave ggrieves: cool!  Let me know if you need any help
15:33 ekristen anyone know why with pygit2 the public key is required?
15:33 vejdmn joined #salt
15:33 ggrieves thx Utah Dave, prime, viq, Im really glad you're here
15:33 UtahDave ekristen: I haven't worked on that but I know there are some limitations to each type of git provider
15:34 UtahDave ekristen: I think terminalmage has some pretty detailed docs on it.
15:36 CycloHex UtahDave: I'm trying to use reactors to fire an event when destroying a cloud-server. salt/cloud/*/destroyed does not trigger anything, salt/minion/*/start eg works.. Is it possible that the cloud reactor isn't working properly?
15:36 TyrfingMjolnir joined #salt
15:37 CycloHex https://github.com/saltstack/salt/issues/16755
15:37 wendall911 joined #salt
15:37 UtahDave CycloHex: Hm. The reactor doesn't really care what type of event comes through. If you can see the   salt/cloud/<minion id.>/destroyed come through the event system, then the reactor should be able to handle it just fine.
15:39 UtahDave CycloHex: Hm. just ready our issue.  That should work.  Have you tried running your salt-master in a terminal in debug mode?  That will often shed some light as to if there are errors or a stacktrace somewhere
15:39 CycloHex And I see the event come true, no reaction. When the event salt/minion/*/start comes through, it does fire a reaction, as I want it to, but my cloud/created is just the same but doesn't fire reactions
15:39 CycloHex you mean with -l debug?
15:39 UtahDave CycloHex: yeah.
15:40 tligda joined #salt
15:40 UtahDave CycloHex: Because, there really isn't a concept of "cloud" reactors.  You can even make your own events like    CycloHex/coolstuff/*/reallyawesome
15:40 oyvjel1 joined #salt
15:41 UtahDave so I'd check to see if your reactor sls is stacktracing or something
15:41 CycloHex Yes, I thought that was possible as well.. really weir dthat it doesn't work.. So you suggest I destroy my cloud with -l debug on?
15:41 CycloHex or do I need to -l debug somewhere else?
15:41 UtahDave No, just the master.     stop the salt-master daemon.  then run      sudo salt-master -l debug
15:42 UtahDave then destroy a minion and see if your salt master output sheds any light on to what's going on
15:42 * goodwill looks at UtahDave
15:42 * UtahDave highfives goodwill
15:43 goodwill ahem!!!
15:43 CycloHex ok, thanks UtahDave!
15:43 * goodwill bellybumps UtahDave
15:43 * UtahDave bellybumps goodwill!
15:43 jxFive joined #salt
15:45 UtahDave have you been behaving, goodwill?  ;)
15:46 goodwill no
15:46 Ahlee speaking of reactors, is there a reactor on host disconnect?
15:46 Ahlee er, sorry. Event
15:47 Ahlee having a hard time tracking minions dying that supervisord isn't restarting for $reasons
15:47 schristensen joined #salt
15:47 UtahDave goodwill: heh.
15:47 Ahlee oh, looks like salt/presence/change
15:47 UtahDave Ahlee: I think there's a minion shutdown event, but I'm not sure if there's one for when the minion just disappears.
15:48 Ahlee no addin tag here
15:48 Ahlee so i'm just gonna assume it's newer than what i have
15:48 Ahlee http://docs.saltstack.com/en/latest/topics/event/master_events.html#presence-events
15:48 Ahlee yeah, I don't think zmq supports end of stream/beginning of stream events
15:49 ekristen i’m starting to get really frustrated with the 2014.7 branch UtahDave
15:49 ekristen it seems like nothing wants to work properly
15:50 ekristen I’ve tried every combination git pygit2 and gitpython on this page http://docs.saltstack.com/en/latest/topics/tutorials/gitfs.html#tutorial-gitfs and I always get errors when trying to sync git repos for states
15:50 ekristen good news is that pillars seem to work fine regardless :/
15:50 ekristen http://docs.saltstack.com/en/latest/topics/tutorials/gitfs.html#gitfs-authentication
15:50 ekristen I meant that page
15:51 zions_ joined #salt
15:51 UtahDave ekristen: really? They've been testing the heck out of gitfs.
15:52 goodwill UtahDave: you got promoted to QA ;)
15:52 ekristen it is not out of the realm of possibility I’ve done something wrong
15:52 UtahDave I'm not doubting you, I'm just surprised they haven't ironed out any issues
15:52 ekristen but I’m having one hell of a time getting a clean 14.04 box setup as a master for testing right now on the 2014.7 branch
15:52 d3vz3r0 joined #salt
15:52 ekristen which has me really really concerned about upgrading my curent master right now
15:53 UtahDave ekristen: how are you setting up your master?
15:53 ekristen I'
15:54 ekristen I’m going to start over, perhaps I missed something fundamental in my git permissions or something and the error messages are just very not clear
15:54 ekristen UtahDave: I bootstrapped the master and the minion to a box, deleted the dist-package and did your trick to get onto the 2014.7 branch
15:55 UtahDave ekristen: ok. Let me know how it goes. I'd like to forward any issues you're having to the main dev for gitfs
15:55 StDiluted joined #salt
15:55 ekristen its probably my fault to be honest :) but I’ll let you know, at the very least we might be able to add a troubleshooting section that says if you see this error message then it probably means this
15:57 UtahDave ekristen: Yeah, definitely. If the docs don't help someone successfully get gitfs going, then we need to improve the docs. That's for sure.
15:57 ekristen UtahDave: I also found you cannot install pygit2 just using pip and apt on 14.04 you have to do it from source because of some bug in unicode/ascii decoding
15:58 ekristen basically have to do https://gist.github.com/olivier-m/5755638#file-pygit2-sh-L24-L29 and use tag 0.20.0 and then you can do a pip install pygit2==0.23.0 but if you try and install the deps for pygit2 using apt and then install pygit2 it’ll break ten ways to sunday and wpm
15:59 ekristen won’t install
16:01 UtahDave wow
16:03 dstufft ekristen: is there a pip error you can show me?
16:03 TyrfingMjolnir joined #salt
16:04 VSpike UtahDave: there's a pull request waiting for ya.. https://github.com/saltstack/salt-winrepo/pull/65
16:04 ekristen dstufft: um, I might have scrollback
16:04 SheetiS joined #salt
16:05 UtahDave VSpike: Thanks!  that's awesome. I really appreciate it!
16:05 VSpike Thanks UtahDave ... yw
16:05 to_json joined #salt
16:06 ekristen dstufft: https://gist.github.com/ekristen/92b8c0f50733677e026f
16:07 ekristen dstufft: all you need is a ubuntu 14.04 server box, apt-get install python python-dev python-pip libgit2-0 libgit2-dev libffi-dev and then do a pip install pygit2 and it pukes
16:08 dstufft ekristen: so the unicode thing is a false flag, it comes from trying to write out the debug log after the command has already failed
16:08 dstufft and it should be fixed in the develop version of pip
16:09 ekristen dstufft: ok, so its covering up another error then?
16:09 dstufft ekristen: the real error is up above, it failed to compile
16:10 VSpike I'm definitely getting this issue very repeatably where during a state.highstate call, file.recurse will fail saying "The directory 'salt://win/files/utils/sysinternals' does not exist on the salt fileserver in saltenv 'base' ...
16:10 VSpike but if i execute the state directly with salt-call state.sls it works fine
16:11 VSpike It's a bit of a showstopper really
16:11 ekristen woot! I finally have my test saltstack environment up again :)
16:12 ldlework left #salt
16:13 dstufft There's no ppa or anything for 2014.7 rc releases is there?
16:13 eriko joined #salt
16:13 iggy VSpike: when it fails, can you see the directory in cp.list_master?
16:13 ramishra joined #salt
16:14 eunuchsocket joined #salt
16:15 gildegoma joined #salt
16:15 ekristen UtahDave: so one of my devops guys disabled one of my private keys for testing on my salt states
16:15 bhosmer joined #salt
16:15 UtahDave doh!
16:16 KyleG joined #salt
16:16 KyleG joined #salt
16:16 ekristen UtahDave: but this was the error I was getting on gitpython was fatal could not read repo (that is what made me thing it was a private key issue)
16:16 ekristen but pygit2 I was getting transport method not supported
16:18 UtahDave is it supported in pygit2?
16:18 ekristen I tried every format listed on the documentation page
16:18 zions joined #salt
16:18 VSpike iggy: hm. Having run the highstate call, i then did salt-call state.sls on the two SLS files containing the file.recurse states that failed, and then did a highstate call again ...
16:19 VSpike iggy: while that was running in another console (and I saw the same file-recurse calls fail again on that run) I did your test in another box. The first result looked so wierd i issued the command again right away and got something totally different...
16:19 VSpike https://bpaste.net/show/cad1b4e077a7
16:19 thayne joined #salt
16:19 to_json joined #salt
16:20 iggy oh, windows... /me backs away slowly
16:21 VSpike iggy: the file.recurse calls which failed are on win/files/utils/sysinternals and win/files/fmpslib and yet a file.recurse on win/files/Carbon works every time
16:21 VSpike darn... the windows affect strikes again :)
16:22 VSpike Where the heck did that first list come from?
16:24 VSpike Wait, that first list is the gitfs_remote I have for the reverse-users-formula
16:25 prosper__ joined #salt
16:29 Mso150 joined #salt
16:30 mohae joined #salt
16:32 vejdmn1 joined #salt
16:33 mgw joined #salt
16:34 VSpike I disabled the git backend and i no longer get a random switch between showing the git backend or the roots with cp.list_master ... but the same file.recurse states still fail
16:36 UtahDave VSpike: which version of salt are you on for your master and minions?
16:41 vejdmn joined #salt
16:41 to_json joined #salt
16:42 shookees joined #salt
16:42 shookees joined #salt
16:44 aparsons joined #salt
16:46 to_json joined #salt
16:47 thedodd joined #salt
16:51 TheRealBill_here joined #salt
16:52 to_json joined #salt
16:53 ggrieves joined #salt
16:55 mgw joined #salt
16:55 _prime_ good morning.  I'm seeing salt-2014.1.13 show up in my latest Gentoo emerge sync.  Are there any release notes available to indicate what's changed between 2014.1.13 and 2014.1.12?
16:56 mpanetta Heh I thought it stopped at 11?
16:57 jhulten joined #salt
16:57 KyleG left #salt
16:58 UtahDave there should be some here http://docs.saltstack.com/en/latest/topics/releases/   but there's not.   I'll ping someone
16:59 Jellyfrog so .13 is out? latest release says .11
16:59 Jellyfrog .10 latest in the topic on irc :)
16:59 Jellyfrog how confusing :P
16:59 _prime_ .12 is definitely out and a nice fix from .11, but I don't want to install .13 until I know what the changes are :-)
17:00 TheThing joined #salt
17:01 dude051 joined #salt
17:01 _prime_ that said, to fix problems with core python libs getting clobbered I'm installing salt in a gentoo-prefix on various fedoras, centos, and sl6 environments, and in testing its working really well.  a bonus is they all get newer, dedicated python and zeromq libs, but this is why I saw .13 pop up.
17:04 Jellyfrog _prime_: https://github.com/saltstack/salt/compare/v2014.1.12...v2014.1.13
17:05 _prime_ nice one, thanks Jellyfrog
17:05 _prime_ looks like a very worthwhile version bump
17:08 druonysus joined #salt
17:09 Gareth joined #salt
17:11 Gareth morning morning
17:12 ggrieves morning (afternoon)
17:12 forrest hey Gareth, did you finish all your slides?
17:15 davet joined #salt
17:15 ipmb joined #salt
17:15 perfectsine joined #salt
17:16 Gareth forrest: Just about :) some last minute tweaking but just about done.  you?
17:17 schimmy joined #salt
17:17 prosper_ joined #salt
17:18 skyler_ joined #salt
17:19 skyler_ Hi all, I have an organization question.
17:19 skyler_ So I have a repo with a bunch of states and pillars.
17:20 skyler_ Now we are adding a new setup that is going to reuse a lot of the work we did on our current infrastructure.
17:20 skyler_ So the states will be mostly the same.
17:20 SheetiS skyler_: sharing the same states or replacing the old setup?
17:21 skyler_ However, the pillars will be mostly the same.
17:21 aparsons joined #salt
17:21 skyler_ I mean, different
17:21 skyler_ the pillars will be mostly different.
17:21 schimmy joined #salt
17:22 skyler_ SheetiS: So what I am wondering is, how I should deal with the two diverging setups?
17:22 skyler_ Should I just make another branch in my repo for the new setup? It would change almost every pillar, and leave almost every state the same.
17:22 SheetiS I'd try and standardize my states and deploy them out of the same repo.  If I have to make my states a little more intelligent so I don't have to maintain them in 2 places, I would do so
17:23 SheetiS then use separate repos or branches in my repo for the pillar data
17:23 SheetiS depending on how that looks best for your org.
17:24 skyler_ SheetiS: Yeah, that makes sense. I am just having a hard time finding an elegant way to store pillars in version control.
17:25 SheetiS I personally like using a different repo for pillars than states, but there are probably as many opinions as people here on that.
17:25 skyler_ Especially when there is some secure data in the pillars.
17:25 skyler_ SheetiS: Thanks for  the input. I like that too, I might have to switch to that.
17:25 murrdoc joined #salt
17:26 spookah joined #salt
17:26 SheetiS skyler_: You can always use gpg renderer for secure data in the pillar as well. http://docs.saltstack.com/en/latest/ref/renderers/all/salt.renderers.gpg.html
17:26 saurabhs joined #salt
17:27 SheetiS then handle your gpg keys outside of the state/pillar repos in some fashion.
17:29 conan_the_destro joined #salt
17:30 ekristen joined #salt
17:35 kickerdog joined #salt
17:36 kickerdog Does anyone have a good tutorial on using salt for building/deploying docker?
17:36 rap424 joined #salt
17:37 SheetiS https://www.youtube.com/watch?v=EVq6hn25lYM is right on the front page of saltstack.com.
17:37 SheetiS I think the states used in that demo are available on github as well
17:38 kickerdog awesome, thanks!
17:38 mjumelet joined #salt
17:42 aparsons joined #salt
17:42 jhulten joined #salt
17:43 jhulten joined #salt
17:45 rypeck joined #salt
17:48 ericof joined #salt
17:49 bezeee joined #salt
17:50 rawkode joined #salt
17:52 rawkode Before I slinging code, is anybody aware of a Salt module for reporting job status to HipChat/IRC? :)
18:02 toddnni joined #salt
18:02 aw110f joined #salt
18:03 Ryan_Lane joined #salt
18:04 ggoZ joined #salt
18:04 prosper_ joined #salt
18:05 jhulten joined #salt
18:06 ggrieves joined #salt
18:06 jhulten joined #salt
18:09 aw110f joined #salt
18:10 thedodd joined #salt
18:11 Ahlee I don't know of a returner, but it's been discussed before
18:12 eeffff joined #salt
18:13 Gareth rawkode: there was a new module that got added to the develop branch last week.
18:13 Gareth rawkode: for HipChat at least.
18:14 Ahlee huh.  I didn't see a _returner in develop just now
18:14 ggrieves I'm trying to shoehorn salt into an existing infrastructure in a minimally invasive way. I just want to retrieve config info for all my minions, including the contents of config files. What's the best way to monitor the contents of config files?  should I use cp.push to retrieve all files and diff on master? or make custom grains that the minions refresh values from the files?
18:14 ggrieves or something better?
18:14 rawkode Gareth: Thanks for the headsup - I'll take a look
18:14 UtahDave rawkode: there's a module for hipchat
18:14 notpeter_ joined #salt
18:15 Ahlee ggrieves: well, if you're file.managed: those configs, running that state with test=True will report differences (what would be changed)
18:15 viq joined #salt
18:15 Gareth UtahDave: what was that one you added last week, another chat looking service right?
18:15 thayne joined #salt
18:15 UtahDave Gareth: slack
18:15 Ahlee how does that work as a module?
18:15 eeffff I'm currently running into an issue with missing pillar data in a masterless setup -- there was talk about cutting a 2014.1.12 release that fixes the regression -- anyone know when that'll be available?
18:15 Gareth UtahDave: _the_ slack? :)
18:16 ggrieves ok thanks: i didnt know about that module, I'll read up
18:16 UtahDave Gareth: yeah, slack.com
18:16 Gareth Ahlee: the hipchat one?
18:16 mapu joined #salt
18:17 Gareth UtahDave: oh. nm. I was thinking of that other tool called slack.
18:17 Ahlee yeah.  My brain's not letting me understand how a module can be used for hipchat.  Do you define something like my_message: hipchat.send_message: - room_id: foo - message: bar within the state?
18:17 Ahlee well, within any state
18:17 Ahlee i guess that's how this module expects to be called.  Kind of cool way of doing it now that I think about it
18:18 SheetiS I'd run the state from the reactor triggered off of a returner myself or so.
18:18 SheetiS from what I see there.
18:18 ggrieves so, file.managed appears to push a file from master.  right now I don't want to risk changing anything mission critical. I just want to generate reports on files/differences?
18:18 Gareth Ahlee: the module you'd use to send a message from a minion, but that state would/could work in a similar fashion to the smtp or xmpp states, you define a state with some information that is called from another state.
18:19 Ahlee Gareth: gotcha.  The more fun explaining "no you use this sls state to call that execution state" to new users
18:20 Gareth would definitely interesting to see a HipChat returner too.
18:20 SheetiS ggrieves: You should be able to put test=True on your state call.  I always apply in a non-prod environment then test=True in prod before doing it for real myself.
18:21 Ahlee ggrieves: I feel you, though.  Salt managing files was a hard sell here, too.  People don't trust it (and don't want to be held accountable for not locally editing files). I just pushed it through slowly rather than trying to tippy-toe around
18:21 derelm joined #salt
18:21 ggrieves ok, thanks,  yeah I'm testing on some idle servers in dev for now, but I would like to read-only for now,  ok I will read more
18:21 ggrieves ok thanks Ahlee,
18:21 Ahlee ultimately I just ran through scenarios on new services and then back ported it to existing production services
18:22 ggrieves ok,  I'm new to the company, to CS (was a physicist), and to salt, so I am trepiditios as heck
18:23 Ahlee welcome to the party.
18:23 ggrieves thank you!
18:23 UtahDave Gareth: ah. No, not that one.  :)
18:25 bhosmer joined #salt
18:26 murrdoc1 joined #salt
18:26 rawkode I guess I should be able to copy hipchat.py to /srv/salt/_modules and restart Salt to use that?
18:26 Mohan__ joined #salt
18:26 rawkode without actually switching to develop?
18:26 murrdoc1 joined #salt
18:26 Ryan_Lane Gareth: there's a hipchat module?
18:26 Ryan_Lane I don't see one in the docs
18:27 Gareth Ryan_Lane: yeah.  got merged into develop last week.
18:27 Ryan_Lane neat
18:27 murrdoc1 hip chat is so last week, we on slack this week
18:27 murrdoc1 :|
18:28 Ryan_Lane oh, wow. a state, too
18:28 Mohan__ Hello all, Anyone know if there is any plan for salt w.r.t python 3 ? with the official End of Life for python2.7 being 2020 http://legacy.python.org/dev/peps/pep-0373/
18:28 rawkode Copying the module to _modules doesn't seem enough - damn
18:29 Ryan_Lane rawkode: do you also have the python module installed?
18:30 rawkode Copied module to /srv/salt/_modules and the state to /srv/salt/_states
18:30 Ryan_Lane oh. it only requires requests
18:30 Gareth Doesn't look like it uses a special hipchat module, just uses requests.
18:30 rawkode Hmm, sync_all seems to show it syncing at least
18:31 Ryan_Lane heh. it seems to only use the central api
18:31 fragamus joined #salt
18:31 Mohan__ I see that there is some python3 support - https://github.com/saltstack/salt/blob/develop/salt/_compat.py
18:32 Mohan__ not sure how pervasive this support is for the entire salt-stack eco system
18:32 Mohan__ does salt stack run on python 3?
18:32 gyre007_ joined #salt
18:33 cberndt joined #salt
18:33 holler_ joined #salt
18:33 holler_ hello, how can I figure out what is causing this in my vagrant + salt provisioner? http://dpaste.com/351H544
18:33 SheetiS Mohan__: I think it's mostly underlying dependencies right now that keep python3 from being the happy place.
18:34 UtahDave Mohan__: we're not quite running on python 3 yet. mostly because of a dep on M2Crypto
18:34 masterkorp Guys any formula for elastic searech ?
18:34 masterkorp *search even ?
18:35 Gareth https://github.com/saltstack-formulas/elasticsearch-logstash-kibana-formula
18:35 Mohan__ thanks SheetiS UtahDave - but is there any plan or roadmap that you know of?
18:36 masterkorp Gareth: lol all in one ?
18:36 masterkorp oh welll
18:36 masterkorp i will do an elastic search specific formula
18:37 rawkode What does "AttributeError: 'ReactWrap' object has no attribute 'hipchat'" mean?
18:37 rawkode That the HipChat module/state isn't available from within a reactor?
18:39 notpeter_ joined #salt
18:39 jngd joined #salt
18:42 snave joined #salt
18:42 lcavassa joined #salt
18:43 mohae joined #salt
18:43 VSpike UtahDave: 2014.1.11+ds-2trusty1 master ... 2014.1.10 minion
18:43 VSpike Hmm.. I didn't realise they were different versions
18:44 schimmy joined #salt
18:45 ggrieves in salt.modules.file how are the paths for windows minions formatted?    e.g  file.access (absolute path to file) (mode)    Also the modes are listed 'f' or 'x'  what are these modes?
18:45 lionel joined #salt
18:46 p2 joined #salt
18:46 ggrieves I've tried both "C:\file
18:46 ggrieves and c$\file
18:48 sschwartz_ee joined #salt
18:48 UtahDave ggrieves: 'c:\file' should work
18:48 sschwartz_ee So, has anyone run into an error message like: "TypeError: cmd_async() takes at least 3 non-keyword arguments (1 given)" when trying to run a command through salt-api? I am using cherrypy with webhooks -- I submit my query, get a "success": true status, and errors in the master log.
18:48 ggrieves in single quotes?  maybe that's why it didnt
18:49 hillna joined #salt
18:49 ggrieves yep, that was it. thx utah dave
18:51 fourmyle joined #salt
18:52 shookees joined #salt
18:54 fourmyle joined #salt
18:54 byron_ joined #salt
18:54 rawkode Could anybody assist me calling a state from a reactor, please?
18:55 SheetiS rawkode: have a paste of what you have now that you can link?
18:56 jngd joined #salt
18:57 fourmyle joined #salt
18:57 saurabhs1 joined #salt
18:57 rawkode SheetiS: Yep: http://pastebin.com/52wkFWYt
18:57 fourmyle joined #salt
18:57 whiteinge sschwartz_ee: what version are you on and what does your HTTP request look like?
18:57 rawkode SheetiS: The error is "AttributeError: 'ReactWrap' object has no attribute 'hipchat'"
18:58 rawkode SheetiS: So I believe that I need to somehow tell the reactor that it's a state somehow?
18:58 SheetiS rawkode: I think you want to have it appear under cmd.state.sls
18:58 SheetiS I'll find an example in the docs
18:59 sschwartz_ee_ joined #salt
18:59 sschwartz_ee_ whiteinge: 2014.1.10, and I'm going in with: curl -H "Accept: application/json" -d-d secretkey="secretkey" -k https://10.83.98.227:8888/hook/pingscript
19:00 rawkode Wouldn't that be too target minions and call a state on them - this is more of a global "This job is runnning" using the salt/job/*/new event
19:02 SheetiS so you don't want to run a hipchat state, but rather call the hipchat module
19:02 SheetiS try local.hipchat.send_message in that case
19:02 SheetiS similar to how that was local.state.sls it you used a state
19:03 rawkode Ahh, thanks SheetiS
19:04 whiteinge sschwartz_ee_: oh, that cmd_async. what version of salt-api are you running? do you have a reactor file running commands on minions when it sees that event? the reason I ask is that hook URL doesn't call LocalClient().cmd_async() at all.
19:04 SheetiS let me know if it works.  I've not tried the hipchat module itself yet
19:04 SheetiS but I may want to.
19:05 byron_ Hi all. Should the command "sudo salt.call --local state.highstate" read in Pillar data from /srv/pillar if it exists and be available in a jinja template?
19:05 forrest Gareth: Sorry was busy there for a few, slides are basically done.
19:06 jsm joined #salt
19:08 sschwartz_ee_ whiteinge: the salt-api is version 0.8.4.1.  The reactor I have set up is supposed to call a local sls in /srv/salt/netapi (or at least, that's how I think I've got it configured)
19:09 rawkode SheetiS: Doesn't seem to work, unfortunately. Nothing in log file and it's on debug as well
19:09 fourmyle joined #salt
19:10 whiteinge sschwartz_ee_: switch to debug logging on the master and fire that event. you should see the event in the logs, you can see any reactors that render in response to it, and hopefully you can see exactly where that traceback is coming from.
19:10 VSpike UtahDave: I was so hoping that would be the answer, but I just upgraded the minion to 2014.1.11 and still the same issue
19:12 sschwartz_ee_ whiteinge: weird -- I'm getting no aditional data, just the same thing from salt.utils.event.
19:12 SheetiS rawkode: I think there are a few other considerations.  I don't see a target in your jinja unless I am overlooking something
19:12 dstokes i have a bunch of salt minion jobs that are hanging, what's the proper way to kill them from the master? `salt-call saltutil.kill_job <jid>` isn't doing anything..
19:13 rawkode SheetiS: Heading home, back in ten. I'm using this hipchat module to notify when salt-master has run a certain job {% if data['fun'] == 'deploy' %}. So it's not actually on a minion
19:14 sschwartz_ee_ whiteinge: weird -- I'm getting no aditional data, just the same thing from salt.utils.event. Makes me wonder if the reactor is broken somehow, save that I know it works in other circumstances.
19:16 ekristen UtahDave: whats the timeline of the 2014.7 release? debating on if I should upgrade to the RC3 or wait
19:17 ekristen or rather rc4
19:17 whiteinge sschwartz_ee_: odd. do you see "gathering reactors for" and your event tag? maybe jinja is producing an empty result? is your reactor state wrapped in a big if-statement perchance?
19:17 UtahDave ekristen: they're about to cut another RC after lunch
19:18 ckao joined #salt
19:18 ekristen nice — maybe we can get https://github.com/saltstack/salt/pull/16710 in? :)
19:19 sschwartz_ee_ The script it's executing has one, but the reactor itself does not. This is what I'm getting: http://pastebin.com/d36J90WP
19:19 UtahDave ekristen: that pr is against develop, not 2014.7
19:19 ekristen ugh, I didn’t make it, I could fork and make the same PR against 2014.7
19:20 duncanmv joined #salt
19:20 ekristen UtahDave: let me know, more then willing to do that if it can get in, even if it isn’t in RC5, we definitely want to get it in so that docker works in the first release of 2014.7
19:21 UtahDave ekristen: I'd do it as soon as you can.
19:22 ekristen doing it right now
19:24 prosper_ joined #salt
19:24 aparsons joined #salt
19:24 dstufft Here's a question, is there any benefit to using a salt scheduler to run highstate automatically vs using a cron job?
19:26 UtahDave dstufft: nope. Just that your configs would be all in one place
19:26 ekristen UtahDave: https://github.com/saltstack/salt/pull/16761
19:27 UtahDave some people lose track of all the places they might have cron type things set up
19:27 dstufft UtahDave: ok cool :)
19:27 dstufft wnated to make sure I wasn't missing something
19:27 Ahlee in a state, the ret['changes'] is salt states applied?
19:27 * dstufft needs to figure out a good way to alert on when salt hasn't ran in some time
19:28 Ahlee nevermind, ret['changes']['foo'].  Was missing the foo.
19:29 * whiteinge hands Ahlee a spare foo.
19:30 Ahlee Much appreciated.  Always dropping those darn things all over my code.
19:30 SheetiS dstufft: one thing I've thought of doing was having a salt state touch a 'lastmodified' file and then you could see the mtime for a last run or monitor as needed.
19:30 SheetiS There is likely a better way
19:31 iggy the master log?
19:31 SheetiS but this way even if the logs have rotated away
19:31 SheetiS you'd have point-in-time of the last run at least
19:31 dstufft SheetiS: yea I'm thinking of doing similar, there's also the passthrough retcode thing and coupling with something that tees the output to log file and buffers it and only displays it if the retcode is non zero then the normal cron stuff will function
19:31 sschwartz_ee_ whiteinge: sorry, did you see my pastebin above?  http://pastebin.com/d36J90WP   That is all I get when I try and hit the webhook.
19:32 iggy I wrote an event listener that sends events to graphite
19:32 iggy so it shows up on our graphs anyway
19:32 dstufft (it's not really so much as did it run, but did it run successfully)
19:32 SheetiS iggy: that's not a bad way to look at it.
19:32 Ahlee whiteinge: started tinkering with expanding pepper.  Runners don't really have a default or set number of arguments.  In your opinion, would it be better to not support runners abstractly in pepper, allow a passed list or dict which gets passed through to the low state call, or other?
19:32 balltongu_ joined #salt
19:34 aparsons joined #salt
19:34 whiteinge Ahlee: is generally cleanest to stick to passing a dict of kwargs when dealing with runners.
19:34 rawkode joined #salt
19:34 whiteinge They need a good revamping...
19:34 Ahlee makes sense.
19:35 whiteinge Now that master-side execution modules are becoming a thing I'm wondering if runners will just go away in some future version.
19:35 aparsons_ joined #salt
19:35 whiteinge sschwartz_ee_: I did not see that. thanks for the link.
19:35 Ahlee master side execution modules? /me googles
19:35 sschwartz_ee_ whiteinge: thanks for the help. :)
19:37 whiteinge Ahlee: that's how the orchestrate runner is doing it's black magic.
19:38 whiteinge see the salt/states/saltmod.py and salt/modules/saltutil.py files
19:38 miqui joined #salt
19:38 whiteinge sschwartz_ee_: hm. can you pastebin your reactor sls file?
19:39 Ahlee whiteinge: thanks
19:39 dalibro joined #salt
19:39 dalibro hi guys! :)
19:39 sschwartz_ee_ whiteinge: http://pastebin.com/q3UhVPvN
19:40 whiteinge sschwartz_ee_: thanks. and also the (sanitized) /srv/salt/netapi/*sls files?
19:41 dalibro is there a possibility to randomly generate mysql root password and store it on master and then get it via sls files to set it on minion (.my.cnf and mysql> set password for root...)?
19:42 Gareth dalibro: You could store as a pillar item.
19:43 dalibro i need to create pillar manually right?
19:43 sschwartz_ee_ whiteinge: http://pastebin.com/waK2ECTx
19:44 Gareth dalibro: I believe so.
19:46 glyf joined #salt
19:46 dalibro Gareth: Ok, thanks, will try to come up with something
19:46 kedo39 left #salt
19:46 _prime_ joined #salt
19:47 rlarkin|2 is pillar data cached on the minion?  that question came up in a security meeting.  I did a find in /var/cache/salt, didn't find anything
19:47 rawkode How does one execute a salt module on the salt-master with no minion?
19:47 Gareth rlarkin|2: I do not believe it is.  It's my understanding that it's only ever on the master.
19:48 rlarkin|2 that's what I thought too
19:48 rlarkin|2 someone said in the meeting that it -is- cached, so now I have to prove it's not
19:48 whiteinge sschwartz_ee_: ah ha
19:49 SheetiS rawkode: you don't have a minion running on the master?
19:49 rawkode SheetiS: Nope - I never found the need?
19:49 sschwartz_ee_ whiteinge: I like the sound of that. What am I missing? (Still very new to salt, so I figured it might be something simple.
19:49 SheetiS I run one and target it just the same as any other minion
19:50 prosper_ joined #salt
19:50 rawkode I never expected it to be so difficult to write a reactor :P
19:50 SheetiS You could target it to run on any minion
19:50 SheetiS that cna reach hipchat
19:50 SheetiS doesn't really matter other than the minion has to be up
19:51 byron_ rlarkin|2: If I look in my cache I can see a file called highstate.p which contains lots of juicy data from the pillar
19:51 _jslatts joined #salt
19:51 SheetiS I manage a number of things on the master (connections to active directory, sudoers, other PAM stuff, etc), so I always planned on my salt master being another minion since it needed those same things as everyone else.
19:51 rlarkin|2 thanks byron
19:52 rawkode SheetiS: OK - I'll run a minion on my master and utilise it for one-off tasks
19:52 byron_ rlarkin|2: remember it'll only be items which the minion has access to in any case
19:53 rlarkin|2 Yeah
19:53 ekkelett left #salt
19:54 byron_ Do you know if the a masterless minion can access local pillar data stored in /srv/pillar?
19:54 whiteinge sschwartz_ee_: three things jump out. i commented here: http://pastebin.com/XUpJ8fKH
19:56 cmthornton byron_: masterless minions can have pillar data
19:57 prosper_ joined #salt
19:58 byron_ cmthornton: Thanks, When running debug I find that my Pillar data is read, but then not accessible during template rendering. Any ideas?
19:58 martoss joined #salt
19:59 sschwartz_ee_ whiteinge: Thank you! I'm not entirely sure I follow the syntax -- is the tgt a host, or the name of the script to be run?
19:59 martoss joined #salt
19:59 sschwartz_ee_ whiteinge: I edited the statements, and the error has gone away -- now I just need to get it to do something. ;)
20:00 whiteinge sschwartz_ee_: the Reactor is calling this function in order to send commands down to minions: http://docs.saltstack.com/en/latest/ref/clients/index.html#salt.client.LocalClient.cmd
20:00 whiteinge (actually it's calling the cmd_async function below that)
20:02 FL1SK joined #salt
20:02 dalibro Gareth: works like a charm :-)
20:02 cnelsonsic_ joined #salt
20:03 dalibro {% set LOOKUP = "mysql:lookup:config:" + salt['grains.get']('nodename') + ":password" %}
20:03 dalibro now i can setup password per nodename :)
20:03 cmthornton byron_: are you using `pillar['key']` inside an sls or jinja template to access the data?
20:03 prosper_ joined #salt
20:04 byron_ cmthornton: inside a jinja template. The config works master+minion, but not masterless
20:04 druonysuse joined #salt
20:04 druonysuse joined #salt
20:05 cmthornton byron_: ahh, ok, so it works master/minion setup. Are you running .11? IIRC there is a bug with masterless pillar data in that version
20:06 byron_ cmthornton: sure does, yep it's .11. That answer's my question. Thanks very much. I'll try an earlier version
20:07 sschwartz_ee_ whiteinge: Got it. This is what I tried next: http://pastebin.com/mBYZBb8P
20:09 CeBe1 joined #salt
20:09 schimmy joined #salt
20:09 cmthornton byron_: yep, np
20:11 whiteinge sschwartz_ee_: almost there, i think. try {% if postdata.secretkey == "<secret>" %}
20:11 dustinrc joined #salt
20:11 zions joined #salt
20:12 perfectsine joined #salt
20:19 sschwartz_ee_ whiteinge: still getting "success=true" and no output other than that that I can find.  the log shows: 2014-10-20 20:19:00,057 [cherrypy.access.49086352                 ][INFO    ] <sourceIP> - - [20/Oct/2014:20:19:00] "POST /hook/ping HTTP/1.1" 200 17 "" "curl/7.29.0"
20:20 dstokes anybody else seeing an OverflowError when trying to use saltutil.kill_job? https://github.com/saltstack/salt/issues/16762
20:22 schimmy joined #salt
20:22 whiteinge sschwartz_ee_: i'm 97% sure it's that if-statement. try pulling it entirely to see what is rendered. the result of the jinja rendering will be visible in the master log as a debug entry.
20:23 nitti_ joined #salt
20:25 sschwartz_ee_ whiteinge: I'm not sure it's getting there, because the master is set to debug, and all I get is that INFO line.
20:25 nitti__ joined #salt
20:26 murrdoc joined #salt
20:27 modafinil all of a sudden i'm getting a git error on my saltmaster -- "LookupError: unknown encoding: -----BEGIN PGP SIGNATURE-----"
20:28 modafinil trying to pull from a gitfs on github (worked fine last week, no changes to the repo)
20:28 Glen_ joined #salt
20:28 modafinil that sound familiar?
20:30 dstokes my minions are logging `[WARNING ] Ignoring re-auth delay because jobs are running`
20:30 dstokes no idea what that means :p
20:31 dalibro State mysql_user.present found in sls freebsd.mysqlDb is unavailable
20:31 dstokes ah, found it. https://github.com/saltstack/salt/blob/0c19ad80153b0f3517eba5049f83c8bf8ab17ba3/salt/minion.py#L897
20:31 dalibro do i need to include some modules for mysql management to work ?
20:32 ggoZ1 joined #salt
20:33 dstokes i guess authentication is failing for some minions, so i can't kill jobs b/c the handshake is busted..
20:34 whiteinge sschwartz_ee_: are you sure you have debug logging enabled? you should definitely see something like the following, regardless if there's an error or not: http://pastebin.com/4yfbQr4A
20:35 snave joined #salt
20:36 schristensen joined #salt
20:36 sschwartz_ee_ Where would it be enabled beyond in the master file?
20:37 hasues joined #salt
20:37 sschwartz_ee_ whiteinge: "log_level_logfile: debug" from the master.  And I've been seeing messages with debug scroll by.  Do I need to set a reactor debug setting as well?
20:38 jhulten joined #salt
20:40 druonysus joined #salt
20:40 druonysus joined #salt
20:40 jalaziz joined #salt
20:40 sschwartz_ee_ whiteinge: Ah-hah! Errors! We're getting somewhere! (watching salt-master dumping to console.
20:41 KaaK joined #salt
20:41 modafinil ah-ha GitPython can't handle signed commits
20:41 KaaK anyone here use salt with docker?
20:42 KaaK it appears to be a very powerful combination -- but I'm having a hard time seeing how to do configuration management inside the docker container
20:43 sschwartz_ee_ whiteinge: error messages and lack thereof saved here: http://pastebin.com/mBYZBb8P
20:43 KaaK i see the docker container being a wonderful way to deploy applications -- and encapsulating all the dependencies, but I have a much harder time seeing how to manage the configuration inside these contains once built
20:43 bezeee joined #salt
20:44 iggy KaaK: that's kind of the point of docker... you don't manage anything inside it
20:44 iggy you redeploy every time you make changes
20:46 Gareth dalibro: nice.
20:47 KaaK iggy, I think I see how that could be wrapped up and orchestrated by salt
20:47 dalibro Gareth: got it a little bit further
20:47 KaaK are there any great examples out there that accomplish that?
20:47 Ned_ joined #salt
20:47 dalibro but now I'm stuck with this message
20:48 dalibro "State mysql_user.present found in sls freebsd.mysqlDb is unavailable"
20:48 Ned_ I'm just cruising through the walkthrough, and noticed it refers to a command salt '*' disk.percent
20:48 Ned_ but that doesn't work in the latest version it seems, should be disk.usage ?
20:48 Ned_ anyone know the best way I could report that sensibly to someone ?
20:49 nitti joined #salt
20:49 iggy KaaK: there's a docker state (and module), I don't know of anything beyond the docs for them
20:49 forrest Ned_: you can open an issue on github, or if you want to find it in the docs and do a pull request that would be great as well!
20:50 Ned_ forrest: righto :-)
20:50 bhosmer joined #salt
20:51 TheThing joined #salt
20:51 glyf joined #salt
20:52 Ned_ forrest: done :-) ... thanks ...
20:53 martoss left #salt
20:53 dstokes Ned_: disk.percent works for me on rc2
20:54 Ned_ dstokes: interesting, I just installed a master and minion from the ppa
20:54 Ned_ and it didn't work for me :-9
20:54 Ned_ :-(
20:54 Ned_ both are Ubuntu 14.04
20:54 dstokes you may be looking at the latest docs. what's your salt --version?
20:54 Ned_ "disk.percent" is not available.
20:54 Ned_ dstokes: oh, I see
20:54 Ned_ you mean the docs might be ahead of the version ?
20:54 dstokes latest stable docs are http://docs.saltstack.com/en/latest/topics/releases/2014.1.11.html
20:54 Ned_ 2014.1.11+ds-2trusty1
20:55 cberndt joined #salt
20:55 dstokes bleeding edge docs are http://docs.saltstack.com/en/latest/topics/releases/2014.7.0.html
20:55 Ned_ yeah
20:55 iggy yeah, default docs are for "development" branch, not latest packaged version (which is silly imo)
20:55 dstokes iggy: seconded
20:55 Ned_ okay, http://docs.saltstack.com/en/latest/topics/tutorials/states_pt1.html
20:55 Ned_ I'm reading that
20:55 iggy dstokes: those are the release notes, not really the versioned docs
20:55 Ned_ gah,
20:55 Ned_ and someone just merged that pull request now :-(
20:55 Ned_ woops ...
20:55 Ned_ whoops*
20:59 dstokes iggy: just realized that, my bad
20:59 jsm joined #salt
21:02 masterkorp Hey is the .12 out on the distributions yet ?
21:03 TheRealBill_here joined #salt
21:05 druonysuse joined #salt
21:06 jhulten joined #salt
21:07 * JordanTesting looks around.
21:08 aparsons joined #salt
21:09 cpowell joined #salt
21:09 racooper joined #salt
21:09 vejdmn joined #salt
21:09 kballou joined #salt
21:16 masterkorp it seems noot
21:16 skyler_ Question for anyone who keeps pillars in a seperate repo from there states. Where do you put your salt/top.sls?
21:16 masterkorp *not
21:16 skyler_ common sense would say it shoudl go with the states.
21:16 masterkorp skyler_: i keep states in formulas separated repositories
21:16 skyler_ But if you think about it, it is kind of project specific, so that would indicate that it should go in the pillars repo.
21:16 forrest Ned_: Either way there should be a versionadd, just open another PR, but do a quick search in the repo for versionadd to see an example, just say it's correcting the error rom that PR, no big deal.
21:17 spookah joined #salt
21:18 viq skyler_: there is a separate top.sls for states and for pillars, so you need both
21:19 skyler_ viq: I know. What I am saying is that if you include all of your states in a standalone repository, and then have a repo with pillars for every project you do, then the salt/top.sls is also kind of project specific and doesn't really make sense to store in a project independent repo.
21:20 skyler_ masterkorp: Where do you store your salt/top.sls file?
21:20 Ned_ forrest: well, further comments on the PR seem to imply the command used to exist, but doesn't now?
21:21 masterkorp skyler_: i am main "salstack" repo
21:22 masterkorp that contains top.sls and pillar data
21:22 masterkorp and helper scripts
21:22 masterkorp formulas are per repository
21:22 pdayton joined #salt
21:23 teepark joined #salt
21:24 skyler_ masterkorp: That is cool. I was thinking the same thing for the pillars repo (pillars, repos, scripts). Do you use all your own formulas, or some combination with saltstack formulas?
21:25 masterkorp skyler_: a combo, if its public we look to a version, if we're not the maintainers of the formula
21:27 kermit joined #salt
21:29 ndrei joined #salt
21:29 skyler_ masterkorp: When you say "look to a version" do you mean that you clone a specific commit from the public repo then point gitfs to your copy? Or is there a way to tell gitfs to use a particular commit from a repository directly?
21:29 dalibro joined #salt
21:33 masterkorp skyler_: yeah private forks
21:33 masterkorp tags if the formula maintainer uses the
21:33 masterkorp altough very few use proper tagging much lesssemver
21:33 masterkorp *less semver
21:35 arctan joined #salt
21:35 arctan do any of the saltstack-formula guys hangout here?
21:36 iggy as in people that have commit access?
21:36 iggy people that have contributed?
21:36 iggy either way, yes
21:37 lionel joined #salt
21:38 masterkorp yes
21:38 lowkeyliesmith joined #salt
21:39 arctan iggy: well, i’m looking at the nginx formula and trying to figure out how it’s setup when using upstart because when i run highstate again, i throw an error
21:40 arctan it appears that the nginx-old-init file.rename needs an unless, but i’m not sure that i follow how nginx is being setup
21:42 arctan is there anyone that works with this formula that could take a look and tell me if skipping the rename if a symlink exists if the right approach?
21:45 Ozack1 joined #salt
21:45 jeddi joined #salt
21:45 admineric joined #salt
21:46 iggy we dumped the nginx formula because it was effectively unworkable for us
21:46 iggy I know others that had the same issues with it that we had
21:46 masterkorp yeah
21:46 masterkorp someone needs to write on that actually works
21:46 masterkorp of the refactor the current one to hell and back
21:46 arctan iggy: what wasn’t working?
21:47 arctan i put in a couple of PRs for it as we came across problems
21:47 arctan i was impressed by the 3 minute turn around time on the last one ;)
21:47 iggy I couldn't tell you off the top of my head, this was a while back
21:49 arctan we needed support for modern versions of nginx and the last PR took care of that
21:49 arctan i’m just confused about how upstart is being setup
21:54 iggy we also don't use ubuntu, so...
21:56 admineric Hello, after updating from Salt 2014.1.10 to 2014.1.11 on Ubuntu 14.04 pillar.item no longer gives results but the debug logs shows the keys as being there http://pastebin.com/ECUBE2TF
21:56 iggy admineric: using masterless?
21:56 admineric Does anyone have any ideas on this?
21:56 admineric yes masterless
21:57 iggy it's broke
21:57 iggy downgrade or upgrade
21:58 admineric thanks, I am guessing there is a bug filed for this do you happen to know what that bug# is?
21:58 iggy uhh
21:59 iggy there's like 4 of them, don't know off the top of my head
22:00 jsm joined #salt
22:00 schristensen joined #salt
22:01 snuffeluffegus joined #salt
22:02 iggy Fix regression in pillar in masterless (:issue:`16210`, :issue:`16416`, :issue:`16428`)
22:02 iggy from the 2014.1.12 release notes
22:07 UtahDave ekristen: did you pr get merged?
22:07 ekristen it did! :)
22:07 UtahDave nice!  thanks for putting that together
22:08 ekristen no worries, least I can do to help with small stuff like that
22:08 hasues UtahDave: How went the pumpkin search?
22:08 molaAMINE joined #salt
22:09 wnkz joined #salt
22:10 druonysuse joined #salt
22:12 UtahDave hasues: Great times were had!!  :)  My kids especially loved the little baby pumpkins.
22:12 hasues UtahDave: Cool beans.  We grew a few this year.
22:12 whiteinge sschwartz_ee_: get rid of the ``fun: ``. and add the missing colon to the ``arg: `` line
22:13 UtahDave nice!
22:15 vejdmn joined #salt
22:16 perfectsine joined #salt
22:19 avienu joined #salt
22:19 yomilk joined #salt
22:19 perfectsine joined #salt
22:19 dude051 joined #salt
22:20 Glen_ Relative newbie here.... We'd like to have a salt master with gitfs and any number of dynamically defined git branches.  The use case is that we'd spin up an environment of VMs and use Salt to configure them, but we'd want that particular environment based on a specific git committish.
22:20 derelm joined #salt
22:20 Glen_ Looks like using gitfs is limited to a smallish number of pre-defined branches.  Am I missing something?
22:21 sschwartz_ee_ whiteinge: Tried it, no dice. :(
22:22 Setsuna666 joined #salt
22:24 aquinas joined #salt
22:28 perfectsine joined #salt
22:34 VSpike UtahDave: it seems to depend on what comes before. If I edit out everything except for the three SLS with file.recurse in them, they all work when called from highstate
22:35 VSpike UtahDave: if I put back in the SLS files before the file.recurse SLS's, then the two that usually fail, fail
22:35 VSpike UtahDave: however, if I just leave one in, and the SLS before it, it works
22:35 VSpike So .. I dunno
22:35 * VSpike scratches head
22:36 __number5__ Glen_: gitfs in 2014.7.0x only support branches-environments mappings
22:37 ajolo joined #salt
22:37 hasues When Salt Cloud is used to provision a VM, what username and password/key config is used?  Does the VM needs to have a user account precreated with keys or some such?
22:39 __number5__ hasues: depends on the cloud provider you're using, normally they have a way to configure ssh public key before launch
22:40 hasues __number5__: This would be vSphere.
22:40 hasues __number5__: I think I follow what you are saying.  I'll look into it.  I presume in this setup, I would need to load a user account and key for this to work.
22:41 hasues __number5__: I'd rather it not be the root account, however, so if I create a user account with sudo access, would its post work still work via a directive that tells it to use sudo or?
22:42 __number5__ hasues: that account will be there afterward and salt-minion will run as root
22:43 __number5__ hasues: http://docs.saltstack.com/en/latest/ref/clouds/all/salt.cloud.clouds.vsphere.html
22:44 forrest Ned_: Sorry was working on some stuff, honestly I'm not familiar enough to say, sorry :\
22:44 hasues __number5__: So the user and password that it is referring to in that config is the username and password used to connect to VCenter.
22:44 Glen_ __number5__: that's what I thought.  Are there plans to extend this in the future or does this go against the general philosophies of Salt?  Thanks, by the way.
22:46 __number5__ Glen_: I'm not sure but I saw ticket to make gitfs mapping more flexible
22:46 StDiluted joined #salt
22:47 DaveQB joined #salt
22:47 Glen_ __number5__: Thanks, I'll see if I can chase that down.
22:49 sschwartz_ee_ whiteinge: new errors, though: http://pastebin.com/0uFALXsF
22:52 CheKoLyN joined #salt
22:58 bhosmer joined #salt
23:00 skyler_ UtahDave: The "Try the shiny new release" message in the docs still points rc3 rather than rc4 just so you know.
23:02 mgw joined #salt
23:05 whytewolf humm, just got asked an interesting question. how do you set a grain using setval. if the value has an =
23:06 snave left #salt
23:06 snave joined #salt
23:14 saurabhs1 left #salt
23:14 jsm joined #salt
23:14 UtahDave joined #salt
23:14 aquinas joined #salt
23:16 rgarcia_ joined #salt
23:16 UtahDave skyler_: :)   and I think they just cut rc5
23:17 rgarcia_ is there a good example repo on github that shows how to test modules/states? I'm looking at saltstack/salt and it seems close to what I want, albeit very complicated
23:20 admineric Does anyone know when Ubuntu PPA will get updated to 2014.1.12?
23:22 KyleG1 joined #salt
23:23 jsm joined #salt
23:23 Singularo joined #salt
23:26 elfixit joined #salt
23:26 nitti joined #salt
23:29 mechanicalduck joined #salt
23:30 KyleG joined #salt
23:30 KyleG joined #salt
23:32 jsm joined #salt
23:33 ipmb joined #salt
23:38 sschwartz_ee_ OK. If you're trying to pass an argument (via json) to state.sls, is it "args" or "kwargs", or what? I admit to now being completely confused. I've tried "fun":"state.sls","args":"netapi.ping" and "kwargs":"netapi.ping", but I get nowhere.
23:47 jsm joined #salt
23:49 pdayton joined #salt
23:49 StDiluted joined #salt
23:49 Ryan_Lane joined #salt
23:49 evidence has a multi-master for syndics discussion ever come up in the past?  a quick browse of issues on github doesn't show any hits
23:51 evidence i know the multi syndic stuff that fixes the duplicate commands being run when mutli-mastered to 2 syndics made it into 2014.7.. but what about 2 minions connected to 2 syndics connected to 2 masters?
23:51 evidence that or even a failover.. currently you can only guide the syndic with a single master, and if it fails you have to point to a new master in master.conf
23:52 skyler_ On the apache formula, there are state files for calling a2enmod on specific modules (aka proxy, proxy_http). It seems like we should just specify a list of modules to enable and then have a state that a2enmods everything in the list. Am I missing something?
23:52 dude051 joined #salt
23:54 sunkist I thought I had a problem with salty-vagrant but it looks like I'm getting the exact same error with salt-bootstrap.
23:54 sunkist my usage is: install_salt.sh git v2014.1.12
23:55 sunkist But it errors out after detecting that I've passed in a tag.
23:57 jsm joined #salt
23:59 gmcwhistler joined #salt
23:59 evidence sunkist: anything change if you use install_salt.sh -U ?

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