Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-02-01

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

All times shown according to UTC.

Time Nick Message
23:00 macorsen joined #salt
23:01 dlam joined #salt
23:01 dlam hmm how do I stop a minion thats currently running?
23:02 dlam is it ok to just 'kill' its process id?
23:03 bemehow joined #salt
23:05 jeddi dlam: /etc/init.d/salt-minion stop   ?  (or equiv for your distro)
23:06 Nazca__ bretep: but yeah ... it looks right to me, so other than trying a different character just in case ... salt['pillar.get']() maybe? sorry
23:07 bretep I'm debugging the code now...
23:08 Nazca__ good option :)
23:08 ajw0100_ joined #salt
23:09 fllr joined #salt
23:11 kalloc joined #salt
23:16 Mua joined #salt
23:18 ndrei joined #salt
23:20 Katafalkas joined #salt
23:28 dlindquist joined #salt
23:33 dfinn joined #salt
23:34 dfinn i'm getting some strange and inconsistent results when trying to run a highstate either on it's own or against a single server.  a little while ago this worked fine when run against the single server.  now I'm seeing this: http://pastebin.com/DtgcnPLm
23:38 taion809 joined #salt
23:40 taion809_ joined #salt
23:45 TronPaul_ joined #salt
23:47 jnials joined #salt
23:48 Nazca__ it's not gonna happen for at least a week ... but at some point I think I shall have to write a salt-api netapi module to have tornado power
23:49 dlindquist joined #salt
23:49 cedwards dfinn: can you share the state file in question?
23:49 DaveQB joined #salt
23:49 dfinn i think I actually solved that error, it was a typo in the top.sls however now I'm running into a different issue
23:49 taion809 joined #salt
23:50 dfinn if I run state.highstate directly against a server it works as expected.  if I run state.highstate against '*' it doesn't make any changes on the server.  I'm just testing this out and I only have 1 server in my top.sls
23:51 cedwards dfinn: by directly against a server do you mean salt 'servername' state.highstate?
23:51 dfinn yes
23:51 cedwards do you have only the one key in salt-key -L?
23:52 dfinn no, I have a couple of hundred
23:52 dfinn to test this, I added some bogus text into a file that salt manages.  running directly against the server detects it and removes it.  running against '*' does nothing
23:55 cedwards are all of the other minions up? makes me wonder if it's timing out or something
23:55 dfinn yeah, pretty sure they are all up
23:56 macorsen left #salt
23:57 cedwards is that server targetted for that state in the top.sls?
23:57 cedwards ..just thinking of what else to double check..
23:58 dfinn it is, I think if it wasn't then it wouldn't work when I ran state.highstate against it
23:58 xfactor973 joined #salt
23:59 xfactor973 salt: after upgrading from 0.17.1 to 0.17.5 i'm getting an error saying jinja variable release is not defined.  has it changed names?
23:59 rchrd joined #salt
00:00 JoeK joined #salt
00:02 jbub joined #salt
00:03 renoirb Hey guys, where did you record last SaltConf's videos?
00:03 renoirb are they available?
00:03 cedwards renoirb: They were recorded, but I don't think they've available anywhere yet.
00:03 fragamus joined #salt
00:04 cedwards xfactor973: i'd be surprised if it changed names between those releases
00:04 xfactor973 cedwards: i did a grains.items.  looks like release went away and oscodename replaced it
00:05 renoirb thanks cedwards, guys from an open source project are asking me about salt. Do you know where they will be?  E.g. on vimeo, youtube... will it be salt stack who will publish them directly?
00:05 renoirb (just so I can make regular checks whenever they are published... where to search for)
00:08 dlindquist joined #salt
00:08 cedwards renoirb: I'm really not sure.
00:08 cedwards renoirb: if I were to guess it'd be their website or youtube
00:10 Nazca__ renoirb: you have one of those nicks I automatically want to read backward ¬_¬
00:10 xfactor973 anyone know how i can include the grain: network.ip_addrs cidr="192.168.1.0/24" into my template file?  I want to get the ip address but don't want to go by eth# because some servers have different interface numbers
00:11 kalloc joined #salt
00:12 ndrei joined #salt
00:14 alunduil joined #salt
00:15 jbub joined #salt
00:15 tedski ugh, this is frustrating... the docs don't seem to line up with the actual versions
00:16 tedski running 0.17.5 minion and master... looking here: http://salt.readthedocs.org/en/latest/ref/states/all/salt.states.cron.html?highlight=cron#salt.states.cron.env_present
00:16 tedski cron.env_present seems to be available
00:16 tedski but, then i get this: Comment:   State cron.env_present found in sls <statename> is unavailable
00:17 Katafalkas joined #salt
00:20 n8n joined #salt
00:21 rchrd tedski: the latest documentation is on docs.saltstack.com
00:22 rchrd tedski: the error seems to indicate a problem with the file_roots config, do other states work?
00:22 xfactor973 nvm i found my answer by using the grain: fqdn_ip4
00:23 tedski rchrd: all other states work
00:24 tedski rchrd: there are many states in that sls that run fine, also
00:24 tedski rchrd: docs.saltstack.com say they are for the development branch and latest is on rtfd
00:24 tedski These docs are for Salt's development version: 85cbec9.
00:25 tedski Docs for previous releases are available on salt.rtfd.org.
00:25 tedski from docs.saltstack.com ^
00:25 sroegner salt-call sys.doc cron doesn't list env_present on my 0.17.5 minion
00:26 tedski that's my point
00:26 tedski i see a set_env though
00:26 tedski but, that's in the module, not the state, right?
00:27 babilen Sigh, /me is so unhappy that #9184 didn't make it into the last release. If only I had been more vocal a few days ago. Is anybody aware of plans to merge that?
00:30 Katafalkas joined #salt
00:36 fllr joined #salt
00:38 dlam joined #salt
00:50 n8n joined #salt
00:50 Thiggy joined #salt
00:50 Persaud joined #salt
00:51 Persaud Alguien Me Explica SOy NUevo
00:51 x5257 joined #salt
00:52 Persaud left #salt
00:59 Thiggy joined #salt
00:59 BenCoinanke joined #salt
00:59 Nazca__ babilen: if you ask basepi really nicely, he might cherry pick it over :p
01:01 babilen Nazca__: I have no idea about policy regarding minor releases, but that is potentially breaking behaviour. Are 0.17.* meant to be maintenance releases that do not change behaviour?
01:01 babilen I mean it is a fairly simple patch (couldn't be much simpler in fact), but it changes behaviour rather dramatically.
01:02 babilen Nazca__: I would, naturally, appreciate it a lot if basepi could cherry pick it into 0.17.*
01:02 Nazca__ this is also true
01:02 babilen basepi: #9184 that is (for context)
01:05 babilen I guess I'll have to just fork it locally and hope that it becomes the "official" behaviour soon.
01:11 kalloc joined #salt
01:14 anitak joined #salt
01:23 mgw joined #salt
01:23 dlindquist joined #salt
01:26 cedwards I'm playing with salt-ssh and the only output I'm getting is '/tmp/.salt/salt-call: Permission denied'
01:28 MTecknology viq: that wasn't so much a shot. You've done a LOT of work and it's awesome. You're trying to make 100% of the logic be pillar driven, right?
01:28 MTecknology does salt-ssh need shared keys?
01:28 cedwards MTecknology: I've got keys configured
01:32 cedwards hmm. I can run raw commands (-r), but modules give me permission denied
01:32 MTecknology I need to play around with salt-ssh and salt-cloud sometime...
01:33 cedwards I should probably be testing this on 2014.1 anyway..
01:34 druonysuse joined #salt
01:36 justBob_ joined #salt
01:37 MTecknology I need to update my linkedin profile picture... I think that one is ~7yr old
01:37 cedwards lol i need to do the same
01:37 cedwards but i so rarely log in to linkedin
01:38 MTecknology I imagine you get your fill of linkedin. :P
01:39 mgw2 joined #salt
01:39 MTecknology It seems like a fun place to work,
01:50 sroegner joined #salt
01:58 VladK joined #salt
02:11 munderwo joined #salt
02:11 kalloc joined #salt
02:14 bejer joined #salt
02:15 lzhang joined #salt
02:20 n8n_ joined #salt
02:25 bejer joined #salt
02:25 tinuva joined #salt
02:36 lzhang joined #salt
02:38 wenzowski joined #salt
02:42 n8n joined #salt
02:45 cachedout_home joined #salt
02:46 jeremyBass joined #salt
02:46 zz_cro joined #salt
02:51 jeremyBass hello, question.. i have a file local.xml that exist.. i need to overwrite the file.  I use file.managed and replaces the file but it also says it errors.. http://pastebin.com/vE0LnZvh it does what is needed but is and error is what i can't seem to get why
02:52 xmltok joined #salt
02:56 vmdsch left #salt
02:58 mpanetta joined #salt
02:58 zz_cro joined #salt
02:59 mpanetta joined #salt
03:03 mgw joined #salt
03:08 jimallman joined #salt
03:09 cedwards erg. why can't i get salt-ssh to work
03:11 kalloc joined #salt
03:16 basepi babilen: we're actually not planning on a 0.17.6. 2014.1.0 should be released next week. Just FYI.
03:16 Corey basepi: Doesn't that make it 2014.2.0?
03:17 basepi It might, except it was branched as 2014.1 so that's the release version. =)
03:17 basepi Still gives a good estimate of release time. It got slightly delayed because of SaltConf.
03:18 basepi But too hard to fix existing tags to change it now.
03:19 Corey Quite.
03:20 Corey Finishing up a 25 minute Salt presentation that I'll be giving later today.
03:20 Corey ANnoying, not enough time for a demo.
03:21 Eugene "today" ? It's drinkng hour.
03:22 Corey Eugene: It's half past 4AM. :-)
03:27 Katafalkas joined #salt
03:30 Ahlee breakfast stout's a thing, ain't it?
03:32 Eugene Fourthmeal!
03:34 druonysus joined #salt
03:34 druonysus joined #salt
03:35 StDiluted joined #salt
03:47 ravibhure joined #salt
03:48 Katafalk_ joined #salt
03:54 n8n joined #salt
03:54 ravibhure joined #salt
04:01 Hipikat hi. i've got a master and two minions, in virtualbox, via vagrant. they're talking through a private network. `test.ping`ing the minions returns True for both. my top.sls pillar includes 'miner' for base: '*' which does the standard mine_functions: network.interfaces: [].
04:02 Hipikat doing `salt '*' mine.get '*' network.interfaces` from the master returns what you'd expect for one minion, and False from the other. running salt-minion in non-daemon mode from the failing minion with '-l debug' shows that it's getting the request and thinks it's returning something…
04:02 Hipikat so my question is: how/where do i even start debugging this?
04:03 Eugene Alcohol is always a good start.
04:03 Hipikat i guess i am due at a bar in two hours.
04:03 Eugene I honestly have no idea, though.
04:03 Hipikat :)
04:05 Ahlee that's a really interesting one
04:05 Ahlee test.ping proves the communication layer is there
04:05 xmltok joined #salt
04:05 wenzowski left #salt
04:06 Ahlee well, the mine_interval appears to default to 60 minutes - possible it's just negatively cached?
04:08 Hipikat hmm, i was calling mine.update on all minions before anything, which i think should refresh everything?
04:08 Hipikat i'll set mine_interval to 1 for a start…
04:11 kalloc joined #salt
04:11 Katafalkas joined #salt
04:13 Hipikat looking at http://docs.saltstack.com/topics/mine/#mine-interval i'm not 100% sure whether the mine_interval should be a minion config setting, master config setting or pillar data.. o.o
04:14 Ahlee that's an excellent point
04:14 Ahlee i'm trying to get mine_functions working now to play along
04:15 Hipikat this is all me trying to use the mine functions for the first time, and only started working with Vagrant/Virtualbox in the last week instead of Digital Ocean, so it's possible i'm doing something stupid on one of those layers
04:15 Hipikat and the salt provisioning in Vagrant has seemed slightly shonky with the latest version…
04:16 xmltok joined #salt
04:16 jeremyBass hello, question.. i have a file local.xml that exist.. i need to overwrite the file.  I use file.managed and replaces the file but it also says it errors.. http://pastebin.com/vE0LnZvh it does what is needed but it as says the state is false, why it does what it's suppost to and errors is what i can't seem to get why
04:17 Ahlee Where's the error?
04:17 jeremyBass it the pastbin
04:17 jeremyBass it's in*
04:18 Ahlee the diff?
04:18 jeremyBass well the top part is the state in the .sls .. the bottom is the Result:    False debug output
04:19 Ahlee I can't tell from what's provided, sorry
04:19 jeremyBass well it there any issue with the state?
04:20 Ahlee well, is -context: on valid yaml?  doesn't look like it
04:20 Ahlee but
04:20 Ahlee nevermind, that would blow up
04:21 jeremyBass yes it is http://docs.saltstack.com/ref/states/all/salt.states.file.html
04:21 jeremyBass it works, but it errors.. it's only cause there was a file before hand..
04:22 jeremyBass i can reproduce it with other file with or with out the templating action
04:24 Ahlee Then i'd open an issue detailing as such
04:24 jeremyBass hmm
04:25 jeremyBass well shoot i was hoping for a trick or something with replacing a file..
04:25 jeremyBass i don't get why it's doing a dif though
04:25 jeremyBass i would figure it'd just delete the old file and add the new
04:25 Ahlee Hipikat: looking at apply_minion_config mine_interval is a minion config option
04:26 Ahlee moving an entire file isn't effecient
04:26 Ryan_Lane joined #salt
04:26 Hipikat yup, found it in /etc/salt/minion. currently re-bootstrapping salt on failing minion to see if it just starts working. :P
04:27 Ahlee i'm still trying to get mine working, heh
04:27 Ahlee i really hate pillars.
04:28 Hipikat i'm quite fond of them, but mine are turning out to be much more jinja than raw yaml…
04:28 Ahlee jeremyBass: well, in the mean time you can always remove it, then lay down a new file
04:28 Ahlee i just hate the multiple places.  I never think to look in the pillar config for options
04:30 jeremyBass @Ahlee that is true, but well the issue does need to be solved for sure
04:33 mpanetta_ joined #salt
04:34 Hipikat Ahlee: aaaaand now it works. okay all the 'slightly unpredictable' behaviour i've seen in the last couple of days has been from trusting vagrant-provisioned salt, so I'm just going to go ahead and stop doing that for now. :)
04:34 Ahlee hehe
04:34 Ahlee well, awesome
04:34 Ahlee and curse you as now I can't log off until _I_ get it working ;)
04:35 Mua_ joined #salt
04:36 fllr joined #salt
04:36 Hipikat my sincerest apologies :)
04:36 mpanetta joined #salt
04:37 Ahlee No worries. It's well beyond time I actually figure all this out
04:39 Katafalkas joined #salt
04:39 jalbretsen joined #salt
04:42 MTecknology viq: puias is a distro and you're pulling from their repos?
04:46 yomilk joined #salt
04:46 jacksontj joined #salt
04:47 Ahlee lol. now it's working here
04:47 Ahlee well shit.
04:47 blah joined #salt
04:51 blah Hi guys, just wondering, why salt exists at all. What puppet or cheeff features doesn't do the thing, so some people decided to create another cms?
04:51 Ahlee well, puppet's ruby, so it's out immediately
04:52 Ahlee i don't have enough chef experience
04:52 yomilk joined #salt
04:52 jeremyBass ok figured it out .. i put     - show_diff: False on it and the real error should up.. but it's back to a permission error i keep having
04:52 Ahlee but, salt's philosophy of being fast and offering both remote execution as a first class citizen and config management sets it apart
04:53 blah Ahlee: you mean, they created salt, mainly because of python? Does it seem rediculous to you?
04:53 Ahlee permission issues? on a minion? strange. Trying to run it as non-root?
04:53 Ahlee blah: No more rediculous than having gentoo and arch and fedora ans slackware and red hat and debian and ubuntu
04:53 MTecknology daaaang, gitlab has a LOT of dependencies
04:54 Ahlee blah: personally salt being written in python is what sold it for us.  We use python already
04:54 blah Ahlee: But python is an ugly, compare to ruby. I'm using both, so I can say.
04:54 Ahlee *shrug* to each their own
04:55 jacksontj joined #salt
04:56 cast salt and puppet don't really compare
04:56 Ahlee I'd agree, but people will always compare them
04:56 cast its more reasonable to compare salt and puppet+mcollective
04:56 blah cast: So what is a major difference. Besids of python, ofcause.
04:56 Ahlee indeed
04:57 blah Well, imaging, it already has mcollective, why you shouldn't.
04:57 Ahlee no DSL required, though if you really want one you can use one
04:57 Ahlee salt's just dictionaries.  Lots and lots of dictionaries :)
04:57 blah Yes, but it nailed to yaml instead, which puppet is not.
04:58 MTecknology you actually think ruby is pretty?
04:58 MTecknology or sane?
04:58 blah Yes.
04:58 Ahlee it is easy to read
04:58 MTecknology scary
04:58 Ahlee the element of least surprise is nice
04:58 cast blah: i don't understand, imagining what?
04:58 Ahlee but, puppet's DSL is terrible
04:58 blah Ruby is much better than python in any way.
04:58 Ahlee ok.
04:59 Ahlee this isn't going anywhere
04:59 blah Yes, puppet's DSL is not Ruby, so it kind uly too.
04:59 Ahlee i've laid out why I replaced puppet with salt
04:59 blah *ugly
04:59 logix812 joined #salt
04:59 Ahlee if you really want to use a ruby based CSM, salt's not for you
05:00 blah It is look's much familiar to python to me.
05:00 linjan joined #salt
05:00 cast one could use ruby dsl in puppet
05:01 Ahlee it's true, you could
05:01 cast i don't know of anyone that did though, not sure if that says anything
05:01 Ahlee you can also use heira and reclass
05:01 Ahlee sky's the limit for how crazy you want to be :)
05:01 cast blah: aside from the language difference, i think salt is much easier to setup
05:02 blah MTecknology: Do you think python is better than ruby. Did you use both? Or you just learned it somewhere?
05:03 cast setting up puppet+mcollective is more work than salt in my experience, thats almost inevitable as they're two seperate frameworks
05:03 tessellare joined #salt
05:03 MTecknology I've done development in java, js, python, php, ruby, vb, c#, c, cobol, jcl, basic, qbasic, perl, and many many many others.
05:04 BenCoinanke joined #salt
05:05 cast this certainly isn't the place to debate language choice, but i don't think there can be much argument around that some people find salt attractive because of the python component [i imagine the same may be true of puppet and ruby].
05:05 blah cast: Do you really thins this http://docs.saltstack.com/topics/tutorials/states_pt3.html is easier to understand.
05:05 Ryan_Lane joined #salt
05:06 Ahlee Given we've been using it for so long, and you link to the third part of a tutorial? yes, we do
05:06 jacksontj joined #salt
05:06 Furao joined #salt
05:06 Ahlee It's jinja templates, if you're not a python user, you likely would struggle with that
05:07 blah MTecknology: This could not be true. Unless you were writting hello words or you are 150 years old.
05:07 MTecknology huh?
05:07 Ahlee this troll is getting old
05:07 MTecknology indeed
05:07 Ahlee still trying to figure out if it's non-native english or just a teenager
05:08 MTecknology I'm going with intentional troll. I'm done with it.
05:08 Ahlee yeah, same
05:08 Ahlee on the plus side, i have the mine working at last
05:08 cast blah: i've done more than a bit of templating in puppet, it is about the same difficulty [its the same thing after all, just in .erb files]
05:09 MTecknology Ahlee: nice, I've not had any reason to play with it
05:09 MTecknology I'm trying to install, configure, and deploy gitlab through salt
05:10 blah So in conclusion. Salt is better because of python. Tool for the tool. Pretty understandable. So long, guys.
05:10 Ahlee you missed so much troll
05:10 Ahlee oh well
05:11 Ahlee MTecknology: yeah, I didn't until Hipikat asked about it, it's been on my to look at list for a while
05:11 Ahlee now i don't know what i'd do with it
05:11 kalloc joined #salt
05:11 ravibhure joined #salt
05:11 Ahlee and I also can't figure out how to query deep into it
05:11 Ahlee but, i can pull network interfaces from minion2 on minion1
05:11 Ahlee so i guess i have that going for me, which is nice
05:12 MTecknology honestly, python is why I first took salt seriously
05:12 Ahlee same
05:12 MTecknology I have a few hundred servers I manage at work, and I still don't know of any reason I'd want the mine.
05:12 Ahlee python + 0mq sold us
05:13 Ahlee i want an easy way to query what interface is active in a bond
05:13 Ahlee that seems like the right place to put it
05:13 MTecknology If I were working with clustered software, I think it would be a good fit
05:13 Ahlee yeah
05:14 fllr joined #salt
05:14 Ahlee oh well, will poke this some more tomorrow
05:17 dlindquist joined #salt
05:19 ravibhure joined #salt
05:23 ravibhure1 joined #salt
05:25 Katafalkas joined #salt
05:26 brianhicks joined #salt
05:27 justBob joined #salt
05:33 justBob_ joined #salt
05:37 Guest34802 joined #salt
05:38 MTecknology Why am I drawing a blank?... pillar.data for pillar, *.item for.. * ?
05:38 MTecknology *grumble*
05:38 MTecknology grains! how could I not figure that out?
05:44 fllr joined #salt
05:46 jsm joined #salt
05:46 _fllr_ joined #salt
05:48 mgw2 joined #salt
05:56 jimallman joined #salt
06:00 jeddi i'm watching a file.managed to restart a service (init.d/networking, debian) and the file it's watching doesn't change on every run *but* it wants to restart the service on every run of the highstate.   where do i start looking?   http://rn0.ru/show/m5fFPzEhPdXOD0BL0oWu/http://rn0.ru/show/m5fFPzEhPdXOD0BL0oWu/ (networking state file - don't mind the verbosity, i've been migrating host-specific data out of states into pillar)
06:02 mgw2 joined #salt
06:03 alunduil joined #salt
06:10 n8n joined #salt
06:11 MTecknology jeddi: - watch: \n\t - file: /etc/network/interfaces  <-- you need that on any service you want restarted
06:11 kalloc joined #salt
06:12 mpanetta_ joined #salt
06:16 elithrar joined #salt
06:17 jeddi MTecknology: taa - just fiddling with that pararaph now .. I liekto have sensible paragraph names, and use the name: parameter to point at the specific path ..  will try your suggestion.
06:19 yomilk joined #salt
06:19 jeddi MTecknology:  this is what I have now http://rn0.ru/show/puthWphwYtS81sRNRsli/
06:21 jeddi Actually, this is what I have right now, including jinja template ..  nothing complex http://rn0.ru/show/ywgpbaX7IgH2KoYzOb87/
06:23 MTecknology - file: /etc/network/interfaces  <-- this no longer exists; salt will tell you that in red text when it fails
06:26 jeddi actually, no error when i run that !  i'll change it to match the stanza heading.
06:29 anuvrat joined #salt
06:40 jeddi MTecknology: thanks for your help - think I've cracked it (I'm moving a bunch of hardcoded vars in my state files into pillar - still not sure on *best practice* for this, but I want to provision a new machine just by adding a single pillar file to describe a host).  I think I had a duplicate watch or network/interfaces  file.managed reference for one host, there, while transitioning - it's clean on subsequent runs on my other test box
06:41 MTecknology jeddi: best practice... no specifics in states, anything about the machine comes from pillars
06:42 zzzirk joined #salt
06:46 fllr joined #salt
06:49 micko joined #salt
07:02 jeddi MTecknology: do you define an entire machine in one pillar file - or do you have bits of it all over the place?  I suspect breaking things down into *types* and *roles* is the key here ..  still need to sort out a good way of doing the initial provisioning - currently using a base image I build up, then clone (esxi), modify grains.conf, reboot, and pick up from that point.
07:03 Corey jeddi: Ooh, you're reimplementing Puppet's Roles and Profiles?
07:04 MTecknology I'm not sure what the best way to use pillars is. Everything in one pile is probably not best
07:04 MTecknology everything all over probably is not either
07:04 jeddi Corey: god I hope not ;)
07:06 jeddi MTecknology:  I'll just want 'base' information about a box (network interfaces configuration is the first bit I'm migrating) defined in /srv/pillar/host/{hostname}.sls ... and from top.sls I can start to play with roles and groupings.
07:06 jeddi you know, for the more sophisticated configurations.
07:07 Corey jeddi: It's not a bad approach really.
07:07 Corey (R&P)
07:09 jeddi Corey: I never really dug into Puppet heavily - we started implementing it where I worked years ago, but I departed soon into that process.
07:10 jeddi i gather historically the pillar stuff came from puppet / hiera refugees (I may be misremembering this)
07:10 jeddi s/came from/was inspired by demands from people with a background in/
07:11 kalloc joined #salt
07:11 Corey jeddi: Actually I was one of the people demanding pillar, if you'd like a quick stroll down history lane; my reasoning was a bit different.
07:11 Corey s/history/memory/
07:12 jeddi Corey: Always happy to hear -- I started dipping the electronic toe into puppet soon after pillars were first introduced.  can't imagine doing large organisational rollouts without them.
07:13 Corey I wanted to store SSL certificates on the master so I could provision my web frontends-- but getting access to those certs via compromising, say, a monitoring server was inexcusable. Since any minion can walk the entire fileserver, that was a non-starter.
07:14 jeddi This makes sense.  :)  I seemto remember an early solution to this was to have hosts check out git (or similar) repos locally, and shuffle certs, ssh private keys, etc around on the minion.
07:14 MTecknology doncha hate it when you figured something out once and can't figure it out again?
07:15 jeddi MTecknology: personal wikis!
07:22 MTecknology Comment:   State mysql_database.present found in sls sys.pkgs.mysql is unavailable
07:22 MTecknology but... what?... yes it is :(
07:22 MTecknology http://docs.saltstack.com/ref/states/all/salt.states.mysql_user.html  <-- it's here
07:22 MTecknology errr... that one, http://docs.saltstack.com/ref/states/all/salt.states.mysql_database.html
07:22 Corey jeddi: Yes, and that was obnoxious enough that "salt didn't manage my sensitive data" was the solution.
07:23 jeddi Corey: obnoxious enough :)
07:23 Corey MTecknology: Versioning issues?
07:23 jeddi Corey: the driving force behind almost everything in IT.
07:23 Corey jeddi: And once upon a time I was a driving force behind Salt, specifically because I was "obnoxious enough." :-D
07:23 MTecknology Corey: I'm on 0.17.4 and it seems it's been available for a loooong time now - http://dpaste.com/1580004/
07:24 jeddi Corey: I hope that's on your CV.
07:24 MTecknology He doesn't need a CV or a resume. His name is all he needs.
07:24 jeddi MTecknology: Oh .. one of *those* people.
07:25 MTecknology oooooooh!!!!!
07:25 MTecknology Corey: reading helps!!! docs.saltstack.com/ref/modules/all/salt.modules.mysql.html
07:31 Corey jeddi: You didn't hear?
07:31 jeddi Corey:  no one tells me anything.
07:31 Corey jeddi: I passed the SSCE! That's my LinkedIn profile headline!
07:31 Corey <-- Senior Big-Data Cloud Devops, SSCE
07:31 jeddi Corey: aha - nicely done!
07:32 fragamus joined #salt
07:33 jeddi Oh, I remember the 'except it doesn't suck' post ... and the flurry of activity on the HN discussion page. :)
07:33 Corey Yeah, I'm trying to outlive that. :-)
07:34 jeddi I thought it was nicely done.  Always uncomfortable when you make throwaway comments about other people's products .. and then have that person pop up on the discussion, however.  :)
07:34 jeddi But, hey, you recovered nicely from that from memory.
07:35 Corey I'm a rock star. We know this. Everybody knows this. ;-)
07:36 * jeddi shudders at rock star reference .. braces for ninja usage.
07:37 MTecknology ssce
07:37 MTecknology spiffy
07:39 MTecknology and now this... http://dpaste.com/1580030/
07:39 ckao joined #salt
07:40 timc3 joined #salt
07:40 MTecknology that looks more like a bug than something wrong on my part
07:41 Corey MTecknology: That looks like you might have dropped a colon or whatnot in your state file.
07:43 MTecknology Corey: anything obvious that I'm too blind to see? http://dpaste.com/1580032/
07:44 Corey MTecknology: I think so.
07:44 Corey MTecknology: Your grant line. Where did you get that from?
07:45 MTecknology http://docs.saltstack.com/ref/states/all/salt.states.mysql_grants.html#module-salt.states.mysql_grants
07:45 Corey MTecknology: That LOOKS wrong. Can you single quote it?
07:46 Corey I think that space is what's hosing you.
07:46 jeddi MTecknology: will you be publishing your set of state files for gitlab?  I looked at doing this about 6 months ago, but rapidly lost interest (it's *really* painful, from memory).  I note it's getting easier.  I also note news last week of a major security issue with gitlab that, presumably, they've now fixed.
07:46 fllr joined #salt
07:46 MTecknology Corey: same result, looking in the source isn't helping me
07:47 Corey MTecknology: "all privileges" just seems like a very strange string there.
07:47 Corey MTecknology: Any luck if you set it to merely "all?"
07:48 MTecknology same thing
07:49 Corey MTecknology: Yeah, I'm wrong. Hmmm.
07:49 Corey If it helps anything, UtahDave is about to walk into my hotel room.
07:49 jeddi MTecknology: you trying to use REQUIRE SSL on your grants btw?
07:49 MTecknology jeddi: nope, not at home
07:50 MTecknology Corey: two brilliant minds together? You'll have it in no time!
07:50 jeddi MTecknology: fairy nuff.  I got bitten by another aspect of this a while back ( https://github.com/saltstack/salt/issues/5778 )
07:50 jeddi apropos almost nothing at all.
07:52 timc3 I am having problems with getting salt stack to restart daemons when a config file is changed. I thought that I just had to put a - watch in?
07:52 n8n joined #salt
07:55 MTecknology jeddi: if you want redhate and postgresql, I can give you something that should work like magic right out of the box
07:56 emostar_ joined #salt
07:58 MTecknology jeddi: https://github.com/viq/gitlab-formula <-- sounds like it works perfect if you use redhat and postgresql. I'm trying to do this on debian with mysql and not from a formula but in the structure that I prefer
07:58 emostar_ joined #salt
08:01 emostar_ joined #salt
08:03 emostar_ joined #salt
08:03 emostar_ joined #salt
08:04 MTecknology I suck at python... wowo
08:06 MTecknology Corey: I found part of my problem, though!  https://github.com/saltstack/salt/blob/develop/salt/modules/mysql.py#L1518
08:07 MTecknology user_grants() returns False
08:08 MTecknology BAM!
08:09 MTecknology damnit... Now I need to do the whole PR thing
08:09 MTecknology heh.... again, damnit...
08:09 MTecknology In typical salt fashion, this has already been fixed in head.
08:11 kalloc joined #salt
08:12 emostar_ joined #salt
08:15 jeddi MTecknology: sadly I'm a Debian bigot .. happily I don't work at that place anymore, so the need for require-ssl on my grants is gone now.
08:15 jeddi MTecknology: oh - for gitlab?  yeah - again I'd want that built and running on a debian vm here .. you're working entirely on rhel/centos?
08:16 jeddi timc3: can you share the fragments of your state file?
08:16 MTecknology I absolutely hate rhel
08:16 timc3 jeddi: sure i will put it up on paste bin
08:16 jeddi timc3: and also mention what distro you're using :)
08:16 jeddi MTecknology: I think we all do.
08:17 MTecknology some people actually like it... :S
08:17 jeddi I've met people who voluntarily used to watch Friends, too.
08:19 timc3 jeddi: It's salt  0.17.5 on Ubuntu 12.04 LTD Server - http://pastebin.com/Xa9nTYhK
08:19 jeddi wow.  someone that uses more.  :)
08:20 timc3 jeddi: yes, bad habit ;-) - forgot to show the top.sls (which for some reason I used cat) http://pastebin.com/1KBtqpmJ
08:20 jeddi timc3: doest a change to rsyslog.conf trigger a restart?
08:21 timc3 No, it doesn't seem to
08:21 jeddi right.  wasn't sure if it was the wildcard on the /* entry
08:22 timc3 jeddi: I was changing the configuration to stop listening on external port, and I noticed that it never restarted rsyslog because it was still bringing in logs
08:22 jeddi timc3: how is the rsyslog.sls pulled in?   i can see you've got includes for for the other two files.
08:22 timc3 jeddi: Mmm. It isn't
08:23 jeddi timc3: how new to salt are you?  (helps working out whether to start with the 'have you tried turning it off and then on again' type questions :)
08:23 timc3 jeddi: extremely new, this is first day.
08:23 jeddi with these kinds of greenfield type sls files, that's the first thing i look for - that i'm actually pulling in the file i'm working on :)
08:23 jeddi you can check that by making a delightfully simple syntax or spelling mistake and rerunning hightstate test=True (of course)
08:24 anuvrat joined #salt
08:24 jeddi timc3: I'd be *tempted* to put all the rsyslog specific stuff into one state file... well, all the stuff that can logically be bundled together of course.  this stuff all seems to be server-side.
08:24 timc3 jeddi: I think it might be because I have missed rsyslog.sls in my top file
08:24 jeddi timc3: i think you might be right :)
08:25 timc3 jeddi: The reason that I am breaking it up is I want rsyslog on many servers, but with different configurations
08:25 jeddi timc3: gotcha.
08:25 jeddi timc3: a sensible approach on day one.  :)
08:26 timc3 jeddi: One master server which listens on external port, then logs to fluentd etc.. All the other machines will log to that.
08:27 jeddi timc3: cool.
08:30 jeddi never even heard of fluentd before - just read up on it now.  shame it's ruby ;)
08:31 timc3 jeddi: Yes, I have tried to get rid of most ruby/perl stuff but this looked really useful, uses C for collection and doesn't need Java unlike logstash.
08:32 timc3 I might run logstash anyway in the end.
08:32 jeddi timc3: i used to wear a sysadmin hat - looked at a few suites to do this on the cheap, but never really settled on one.  came from a cassandra shop, so they were less than ecstatic about a mongo-backed system (which ruled many of them out)
08:34 timc3 Does logstash need mongo? I actually want to send everything through to elasticsearch + kibana, so that I can monitor the software that we build
08:35 jeddi oh god knows ... all that cruft has been pushed out of my brain. :)
08:37 timc3 jeddi: my changes to the files are now restarting the daemon as needed - thanks for the help!
08:37 MTecknology gitlab is sooo much like typical ruby crap
08:38 MTecknology it's really neat, but comes with 70,000,000,000 different dependencies and weird crap that you need to configure in some ugly config language
08:38 MTecknology still don't have it running yet, but I think I'm close
08:39 jeddi timc3: glad i was able to help.  by the time you get to day two you'll have to talk to the smarter guys.
08:39 timc3 MTecknology: Have a look at rhodecode, it's written in python
08:39 jeddi MTecknology: that's pretty much why I gave up after an afternoon of battling it .. I think I ended up concluding I'd need to pretty much build everything, stick itinto a gitolite repo, and then just push itout with salt to one box.
08:40 timc3 MTecknology: free up to 20 users might be the only catch for you.
08:40 jeddi which means it's $-free, not actual free.
08:41 timc3 I am willing to put up with that if it means supporting a python project ;-)
08:41 MTecknology jeddi: does it have teh same features as gitlab?
08:41 jeddi MTecknology: I have no idea.
08:41 jeddi timc3:  ?
08:42 MTecknology err.. timc3? **
08:42 jeddi MTecknology: gitolite meets my needs, but I don't collaborate, or need an issue tracker, code review, etc.
08:43 timc3 MTecknology: I think it has more features than gitlab, but whether it has the right features for you I don't know
08:43 MTecknology I don't need that at home, but I need it at work, and I want it at home
08:43 MTecknology I'll have to take a look...
08:44 timc3 MTecknology: The only problem I have come across is that developers/support are really bad at explaining things.
08:46 martoss joined #salt
08:46 fllr joined #salt
08:47 Simon____ joined #salt
08:47 joehillen joined #salt
08:48 MTecknology I think I might almost have this working....
08:50 MTecknology sooooooo close
08:53 akoumjian joined #salt
08:53 z3uS| joined #salt
08:58 MTecknology missing library
09:06 MTecknology jeddi: so... rhodecode lets you do pull requests and such?
09:08 MTecknology timc3: **  dangit
09:08 timc3 MTecknology: What's up?
09:09 MTecknology hilighted the wrong person again
09:11 timc3 MTecknology: Yes it let's you do pull requests, think that you might have to set up the repository in a certain way. I must admit I looked into it, installed it, but haven't had time to really get working with it
09:11 kalloc joined #salt
09:12 MTecknology the installation seems to be about sixty thousand times easier
09:13 retr0h joined #salt
09:13 retr0h joined #salt
09:23 ampex_ joined #salt
09:23 MTecknology timc3: I ran through the installer and it seems like all it did was download stuff, nothing about launching
09:24 timc3 MTecknology: I think that you have to launch it as a daemon manually.
09:24 scassidy_ joined #salt
09:24 MTecknology where's the file that I need to launch?
09:24 timc3 timc3: I would probably use something like supervisord if it doesn't come with something (of course you should wrap this up in salt).
09:25 timc3 MTecknology: /etc/init.d/rhodecode
09:25 MTecknology there's no file there
09:26 insatsu_ joined #salt
09:26 MTecknology All I seem to have:  data  repos  rhodecode-installer.py  system    under ~/rhodecode
09:27 harobed joined #salt
09:28 harobed joined #salt
09:28 MTecknology I feel like that installer finished way too fast
09:29 MTecknology ya... something made it bomb
09:34 Simon____ joined #salt
09:34 micko joined #salt
09:34 cast joined #salt
09:34 dave_den joined #salt
09:34 mlen joined #salt
09:34 Heartsbane joined #salt
09:34 chitown joined #salt
09:34 smkelly joined #salt
09:34 patrek joined #salt
09:34 oc joined #salt
09:34 whyzgeek joined #salt
09:34 goodwill joined #salt
09:34 gmtech joined #salt
09:34 Troic_ joined #salt
09:35 elithrar joined #salt
09:35 DaveQB joined #salt
09:36 MTecknology timc3: ran out of ram... thanks for the tip, I'm excited to see how this goes
09:37 timc3 MTecknology: no problem
09:38 EnTeQuAk joined #salt
09:41 djaykay joined #salt
09:41 justBob joined #salt
09:42 elithrar joined #salt
09:43 mlen joined #salt
09:45 MTecknology timc3: How do you do a pull request?
09:46 fllr joined #salt
09:46 MTecknology ah... makes sense
09:47 MTecknology this is really really fast and actually does what I wanted
09:52 elithrar joined #salt
09:52 MTecknology hrm... I created the pull request, I'm now reviewing it as the admin user. I'm not finding the buttons to actually pull it. Is that a manual process?
09:55 babilen MTecknology: No, https://help.github.com/articles/merging-a-pull-request .. just click "Merge pull request"
09:55 MTecknology Ah... server side merges aren't a thing yet
09:56 MTecknology babilen: not github
09:56 MTecknology I'm playing with rhodecode, it's pretty snazzy
09:56 babilen MTecknology: Sorry, only ever encountered that terminology in the context of GH. /me goes back to his coffee
09:57 MTecknology same concept
09:59 MTecknology k... BED TIME! Before the sun comes up.
10:02 ajw0100 joined #salt
10:08 Corey Sun's up here!
10:11 kalloc joined #salt
10:19 elithrar joined #salt
10:28 gasbakid joined #salt
10:42 elithrar joined #salt
10:46 fllr joined #salt
10:48 gasbakid joined #salt
10:50 elithrar joined #salt
10:50 elfixit joined #salt
10:53 emostar joined #salt
10:59 elithrar joined #salt
11:00 kalloc joined #salt
11:05 dlindquist joined #salt
11:10 elithrar joined #salt
11:12 linjan joined #salt
11:16 giantlock joined #salt
11:44 martoss joined #salt
11:46 fllr joined #salt
11:46 dren joined #salt
11:48 ravibhure joined #salt
11:51 ravibhure1 joined #salt
11:58 kalloc joined #salt
12:03 taion809 joined #salt
12:10 dlindquist joined #salt
12:19 diegows joined #salt
12:26 joehh babilen: I need to do a new release to fix http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=736783
12:27 joehh Should I include  #9184?
12:27 joehh are there any others?
12:34 yomilk joined #salt
12:45 favadi joined #salt
12:46 fllr joined #salt
12:49 yomilk_ joined #salt
12:59 ekristen joined #salt
13:00 joehh AdamSewell: as far as I am aware, the packages are in the ppa (https://launchpad.net/~saltstack/+archive/salt/+packages)
13:05 zzzirk joined #salt
13:14 dlindquist joined #salt
13:23 patrek_ joined #salt
13:36 ndrei joined #salt
13:37 lzhang joined #salt
13:46 fllr joined #salt
13:52 jinnko joined #salt
13:53 tinyhippo joined #salt
14:06 scalability-junk what's the difference in publish.publish and the reactor system? are they just different options to do something or is there any real difference?
14:08 jslatts joined #salt
14:19 dlindquist joined #salt
14:28 ChoHag joined #salt
14:40 vic0 joined #salt
14:43 bwghughes joined #salt
14:46 fllr joined #salt
14:50 jkleckner joined #salt
14:52 bejer joined #salt
14:55 jsm joined #salt
15:01 quickdry21 joined #salt
15:04 teebes joined #salt
15:15 dlindquist joined #salt
15:15 kalloc joined #salt
15:17 kalloc_ joined #salt
15:27 Corey What a day.
15:28 Corey scalability-junk: Good question! Publish.publish is for asking questions to everything; event reactor is for triggering on events.
15:28 Corey joehh: You got a shout-out in my SaltConf talk.
15:28 jeremiah joined #salt
15:28 jacksontj joined #salt
15:28 zach I'm kind of curious how difficult it would be to make an 'expect' module/state
15:29 zach Something you can use to pass info to an interactive command, from cmd.run
15:30 Corey_FOSDEM zach: Not horribly, but there's already something there that does that.
15:30 zach Oh?
15:30 zach I need to send \r, 1, \r
15:31 jeddi I miss being at FOSDEM.  {sigh}
15:31 Corey_FOSDEM Digging it out.
15:31 zach I *might* get to go to the RSA conf, we'll see ;-)
15:31 jimallman joined #salt
15:32 Corey_FOSDEM zach:  salt.modules.cmdmod.exec_code
15:32 Corey_FOSDEM zach: See http://docs.saltstack.com/ref/modules/all/salt.modules.cmdmod.html
15:32 zach gasp
15:33 Corey_FOSDEM zach: Took me a second to dig out the name.
15:33 zach let me try it!
15:33 zach thanks Corey_FOSDEM
15:33 ndrei joined #salt
15:33 Corey_FOSDEM zach: Yay!
15:34 zach i'll tell you in a second if it works
15:34 zach have to spin up a new cloud server first though
15:35 Corey_FOSDEM Quote of the day from my talk at FOSDEM: "Welcome to the cloud. It's where we devops."
15:35 zach Now that I have 'salt-cloud' configured, life has been <3
15:35 zach Yeah, I'm getting this new company up on the devopsy stuff
15:35 zach I told them some buzz words, like "cloudify"
15:35 zach So that's what I am implmenting
15:37 scalability-junk Corey_FOSDEM: alright so publish is for running echo "hello I'm the minion" and publish that to all other minions targeted.
15:38 scalability-junk and the reactor system is a master implementation to react to these published messages and even more events occuring. new target webserver coming up etc. pp
15:38 Corey_FOSDEM scalability-junk: Not quite. I've used it more for "Everyone, tell me how much disk space you're using,"
15:38 Corey_FOSDEM And event-reactor is when an event happens on the minion, it fires off a reactor on the master.
15:39 scalability-junk Corey_FOSDEM: alright reactor is understood
15:39 scalability-junk Corey_FOSDEM: mhh alright, but you are using publish from one minion to get the disk space of other minions?
15:40 SEJeff_work joined #salt
15:40 zach Corey_FOSDEM: GENIUS! It worked! /me bows
15:40 Corey_FOSDEM zach: I'm awesome, but Salt is even better. :-)
15:40 zach It is now fully automated and I have achieved godmode
15:41 Corey_FOSDEM SEJeff_work: Hello.
15:41 Gareth zach: Corey is *extremely* modest too.
15:41 Corey_FOSDEM Gareth: Corey is also not slated to speak at SCaLE. *glare*
15:41 scalability-junk does saltstack have any "offices" in europe?
15:41 zach nothing better than knowing your shit :P
15:41 Gareth Corey_FOSDEM: Corey is.  Corey is giving an upSCALE talk.
15:41 Corey_FOSDEM I'm drunk on the power of presentations that have gone very, very well this week.
15:42 Corey_FOSDEM Gareth: I've received no emails about this of any sort, and it's ~3 weeks away. :-)
15:42 Gareth Corey_FOSDEM: I sent you an IRC pm weeks ago.
15:42 Corey_FOSDEM Gareth: Yes, but I was hoping for something a bit more official than UnidentifiedGareth. :-)
15:42 kalloc joined #salt
15:43 Gareth there is nothing more official than UnidentifiedGareth.
15:43 Corey_FOSDEM Gareth: Is it my You Are All Idiots talk?
15:43 Gareth Corey_FOSDEM: Yup
15:43 Corey_FOSDEM Amazing.
15:43 Corey_FOSDEM upSCALE is 5 minutes, slides advance every 15s automatically?
15:44 Gareth 20 slides.  5 minutes.  15secs per slide.
15:44 Corey_FOSDEM Yoink. I've got some prep work to do.
15:44 ChoHag joined #salt
15:46 fllr joined #salt
15:46 jimallman joined #salt
15:49 mgw joined #salt
15:55 zach Hmm, I can't remember...does salt follow order in a state?
15:55 zach i have a set of things I need to run in a specific order
15:59 zach ah it does, I just need to rewrite my 600 line state now :)
16:06 scalability-junk zach: it does, if you want it too. But afaik it's recommended not to rely on it.
16:07 scalability-junk It's better to rely on require, watch etc. as it makes the states much more interchangable as far as I understand
16:07 Ahlee That's as official as a statement as I've heard RE: saltstack
16:09 scalability-junk Ahlee: ?
16:10 Ahlee better to utilize require/watch/etc
16:10 scalability-junk ah alright didn't know if your comment was related to my message
16:11 scalability-junk zach: it get's way more complicated if you rely on ordering later on.
16:12 munderwo joined #salt
16:17 lzhang joined #salt
16:20 xfactor973 joined #salt
16:20 xfactor973 left #salt
16:24 cedwards anyone have any idea how the salt-ssh "scanner" roster works?
16:24 Ahlee first i've heard of scanner, interesting
16:27 mgw joined #salt
16:28 zach scalability-junk: the ordering is only required once (at install), after that, it never gets used again
16:28 zach scalability-junk: I'm using it to setup a cloud server in a very particular and required order to deploy a java application
16:29 Ahlee cedwards: not seeing much at all in the source either
16:29 zach once it's deployed, wont need to do it again :)
16:34 cedwards Ahlee: the more I dig into salt-ssh the more I realize it needs attention
16:35 Ahlee it is young
16:35 zach Just need to get salt to generate me a java keystore and I'm fully automated
16:53 zach ugh. You can only have file.replace defined once on a file?
16:56 zach This is what I mean: http://pastebin.com/nbBNs5eU
16:56 zach I need something like that
16:57 zach Rendering SLS "base:Moo.Deploy.Moo" failed: Conflicting ID "file.replace"
16:57 zach I guess using it 12 times on 1 file is frowned upon
16:58 zach wow that paste is all sorts of broken too
17:03 mgw joined #salt
17:06 munderwo joined #salt
17:11 StDiluted joined #salt
17:12 jslatts joined #salt
17:13 patrek joined #salt
17:18 ndrei joined #salt
17:24 jslatts joined #salt
17:26 jeremyBass there is no ` salt.states.mysql_table ` is there?
17:27 btorch jeremyBass: don't see it in http://docs.saltstack.com/ref/states/all/
17:27 btorch there is a module though
17:32 jeremyBass well i suppose i can always do a cmd.run .. just hoped there was a little better db support was all
17:36 matanya joined #salt
17:39 Ahlee is it possible to call custom states from the command line?
17:40 zach salt blahdomain state.sls State.File ?
17:42 vic0 left #salt
17:45 matanya joined #salt
17:49 Ahlee not an sls, a state module
17:50 teebes joined #salt
17:53 jeremyBass question, i need to set an env in one set cmd.run and be able to get it in another.. i did a quick test i thought should have worked .. http://pastebin.com/wDf0GH65
17:54 heewa joined #salt
17:54 jeremyBass for some odd reason it doesn't though.. i have tried a few ways .. using ` $ export var=foo` and testing with `$ env `  which also does work
17:55 heewa You know how the 'cmd' state has 'unless'? Is there a general technique for doing that with other states, like pip?
17:56 heewa Or, like can I have a state that's not normally run, but another state triggers it?
17:56 heewa Hmmm, maybe I can run a module fn to check something in a template if statement around the whole state?
18:00 diegows joined #salt
18:03 kalloc joined #salt
18:10 dlindquist joined #salt
18:13 ilbot3 joined #salt
18:13 Topic for #salt is now Welcome to #salt | Thanks to everyone who made SaltConf 2014 amazing! | 0.17.5 is the latest | Please be patient when asking questions as we are volunteers and may not have immediate answers | Channel logs are available at http://irclog.perlgeek.de/salt/
18:16 bharper joined #salt
18:22 ndrei joined #salt
18:25 n8n joined #salt
18:26 MindfulMonk joined #salt
18:29 bharper I am running into some issue with salt-call on freebsd, am I doing something wrong? http://pastebin.com/qQQL3qm0
18:29 ajw0100 joined #salt
18:30 AdamSewell joined #salt
18:31 Ryan_Lane joined #salt
18:35 lzhang joined #salt
18:39 heewa left #salt
18:40 ajw0100 joined #salt
18:44 opapo joined #salt
18:49 taion809 joined #salt
18:56 jslatts joined #salt
18:59 jfzhu_us joined #salt
19:01 Ahlee ah, you have to write a sls state file to use custom _states
19:01 elfixit joined #salt
19:05 lzhang joined #salt
19:07 mikef2 joined #salt
19:08 DaveQB joined #salt
19:08 MrTango joined #salt
19:09 bharper where does to sls files live on freebsd as there is no /srv?
19:09 Ahlee bharper: whereever file_roots is set to, defaults to /srv/salt if not set
19:09 bharper should I be concerned about the zfs stderr stuff?
19:10 Ahlee only if you plan to use zfs
19:10 Ahlee or rather
19:10 bharper I am
19:10 Ahlee manage zfs pools via salt
19:11 Ahlee looks like it doesn't like /sbin/zfs help
19:11 bharper yea....
19:12 Ahlee https://github.com/saltstack/salt/blob/develop/salt/modules/zfs.py#L44
19:12 Ahlee looks like it's just trying to build a list of commands
19:12 Ahlee i remember reading somehwere that salt's thus far only been run on 9.x, 8.x and 7.x - are you on 10.x?
19:13 Ahlee ah, yeah
19:13 Ahlee you're trying to run 10.0-RELEASE
19:14 bharper http://docs.saltstack.com/topics/installation/freebsd.html maybe here?
19:14 Ahlee likely where i pulled it from
19:15 Ahlee i think the last time i ran freebsd 4 was the stable version
19:15 bharper should I submit a bug or is it a know issue on 10?
19:15 bharper hehe
19:16 Ahlee https://github.com/saltstack/salt/issues/10087?source=cc
19:16 Ahlee looks known
19:17 Ahlee way too lazy to set up both a 9.x and 10.x system to compare the output
19:18 bharper let me know if you need some help
19:19 _mrtango_ joined #salt
19:20 teebes joined #salt
19:23 jeremyBass does anyone know why i can't seem to set an environment in one state via ` cmd.run: ` and pick it up to use in another?  heck any way would be good..
19:23 opapo joined #salt
19:24 Ahlee i'm relatively sure you can only use one environment per
19:24 mikef2 jeremyBass: they are not persistent,  (like calling export in one shell )
19:25 Ahlee oh, you mean shell environment variables?
19:25 Ahlee not salt environments?
19:26 jeremyBass yes
19:26 tedski that's a good question, Ahlee... i think we all assumed shell env since he said cmd.run :)
19:26 tedski but, that's a poor assumption
19:26 jeremyBass the shell environment this is one of the many tests http://pastebin.com/wDf0GH65
19:27 jeremyBass but Im not limiting to shell env.. anything will do as long as i can pass a set value from one state to read in another
19:27 Ahlee drop it into a .sh file and source it
19:28 tedski or bunch both of those states into a cmd.script
19:28 tedski and template the source
19:30 jeremyBass well um.. in the test that is in the same state so it would already be rendered out right
19:30 tedski maybe i'm not understanding... you want the result of a cmd run on the target minion to be parsed into the next state?
19:33 joehillen joined #salt
19:35 jeremyBass sort of .. i need to know if the last cmd.run had a value in another cmd.run that a simple true state of salt will not work for.. i need a value of script i can't controll the output, so i want to assign it to a system env  like ` cmd.run: - name: FOO=$(nginx -v) ` and then call it in another state like ` cmd.run: - name: echo $foo `
19:35 Ahlee bharper: loading up some freebsd vms now, as i'm curious.
19:36 Ahlee then drop the output into a file and read it, save yourself the headache
19:37 jeremyBass oh well the echo would be $FOO but yeah
19:37 jeremyBass lol
19:37 Ahlee i have a feeling 'boot only' freebsd images arne't going to let me install, are they
19:38 Ahlee nope, looks liek it will
19:38 Ahlee neat
19:38 bharper Ahlee: still running into issues using a state http://pastebin.com/3pdm3wnL (I am not worried about the zfs stuff)
19:40 bharper looks like I might need to change root_dir??
19:40 jfzhu_us Can you have your entire file_root (namely /srv/salt) on a gitfs?
19:41 Ahlee jfzhu_us: yes
19:41 Ahlee master maps to base, branches map to environments
19:41 Ahlee be careful with top.sls in your branches
19:43 joehillen joined #salt
19:45 gasbakid joined #salt
19:47 mgw joined #salt
19:48 MTecknology Ahlee: You up for a challenge?
19:49 Ahlee MTecknology: mayhaps
19:49 Ahlee got a while before freebsd9/10 finish installing
19:49 MTecknology If I give you a file, can you turn it into a formula?
19:49 crooton joined #salt
19:50 bharper Ahlee: I think this 2nd issue is another differance between freebsd 9 and 10
19:50 MTecknology https://rhodecode.com/dl/rhodecode-installer.py
19:51 MTecknology That installer is actually so easy that turning it into salt states almost seems silly
19:52 Ahlee a formula?
19:52 Ahlee let me read through it
19:53 anuvrat joined #salt
19:53 tedski anyone know of a module that throws a deprecation warning?
19:53 Ahlee tedski: i believe file has some depecated functions
19:53 Ahlee like file.sed
19:53 tedski i know this isn't #salt-dev, but i'm just looking for an example
19:53 tedski okay
19:53 tedski is ee
19:53 tedski err.. i see
19:53 tedski thanks, Ahlee
19:54 Ahlee bharper: let me look into that second one, i remember hearing terminalmage mention somethings are broken on freebsd, but i don't remember if it was pkg
19:54 Ahlee np tedski
19:54 bharper Ahlee: in freebsd 10 pkgng is now pkg and is installed by default.
19:54 Ahlee bharper: ah, yeah.  That's it.
19:54 Ahlee pretty sure pkgng support isn't baked in until hydrogen
19:55 bharper and freebsdpkg.py check the old location for it's configuration file
19:55 bharper hydrogen?
19:55 Ahlee hydrogen is hte next release after 0.17.x
19:55 bharper ahhh thanks
19:56 Ahlee was gonna be 2014.1.0, but, well, tehy missed Jan
19:56 Ahlee so dunno what's happening with it
19:56 bharper then I'll wait till the next major release
19:56 bharper thanks for looking
19:56 Ahlee i'm looking to see if i can find the differences, maybe you can backport what you need from pkgng to 0.17
19:56 Ahlee oh, fair enough
19:56 bharper if zfs and pkg is not ready in current release
19:57 bharper I would prefer to wait
19:57 Ahlee makes sense
19:57 MTecknology Ahlee: https://rhodecode.com/dl/rhodecode-installer.py
19:57 bharper I have had good luck on centos and fedora
19:58 Ahlee MTecknology: yeah, looking through it - you want a state to wrap that?
19:58 bharper Ahlee: I appreciate your time
19:58 MTecknology yup! I'm looking at it now
19:58 Ahlee bharper: no problem
19:58 MTecknology Ahlee: probably not wrap it, but actually disect it and recreate what it does using salt
19:59 Ahlee yeah
19:59 bharper left #salt
20:00 Ahlee from a cursory skim through, it's like 90% of the way there
20:01 terminalmage Ahlee: just FYI, freebsdpkg no longer does any pkgng stuff
20:01 Ahlee he lives!
20:01 Ahlee i knew you mentioned something about that
20:01 terminalmage freebsdpkg used to try to do both, and attempted to use pkgng if it found the pkgng sqlite databases
20:01 terminalmage *database
20:02 terminalmage Ahlee: yeah I got back just in time...
20:02 Ahlee get in this morning?
20:02 terminalmage landed at 8pm, and overnight we got a ton more snow
20:02 Ahlee oh wow
20:02 Ahlee yeah, good timing
20:02 Ahlee it's snowing like crazy here now
20:02 terminalmage yep
20:02 Ahlee snow crew's been through three times already this morning/early afternoon
20:02 n8n joined #salt
20:02 crooton Can I define which machine a state will run on by defining it in a pillar? I think I'm missing something simple to be honest.
20:03 terminalmage wow... what's it like having people that actually remove the snow?
20:03 Ahlee terminalmage: expensive.
20:03 terminalmage crooton: what is your use case? there are a number of different ways
20:03 Ahlee but, not having to own a shovel in chicago makes it worth it ;)
20:03 terminalmage Ahlee: ha, touché
20:03 crooton Here is a pastebin link: http://pastebin.com/LePVJRXt
20:04 Ahlee terminalmage: i need to chat with you about passing in jinja from the command line to a state file calling a custom state
20:04 crooton I can make the states I've made run on fixed machines by directing them to the command line
20:05 terminalmage crooton: so, given your pillar data, you can wrap the user.present state in an {% if grains['host'] == args['host'] %}
20:06 crooton let me give that a try.
20:06 Ahlee er, sorry, not jinja.  passing in pillar data
20:07 Ahlee holy crap it's 2
20:07 rofl____ joined #salt
20:11 Guest2503 joined #salt
20:14 crooton Ahlee, many thanks for your help.
20:15 fllr joined #salt
20:16 jeremyBass what would `  stdout: ERROR: "//app/etc/" must exist for proper tool work.  ` mean when using cmd.script: ?
20:17 EWDurbin_ joined #salt
20:19 pythopticon joined #salt
20:19 terminalmage Ahlee: feel free to ping me directly
20:21 honestly joined #salt
20:21 honestly joined #salt
20:24 rofl____ joined #salt
20:25 jeremyfelt joined #salt
20:27 MindfulMonk left #salt
20:28 BrendanGilmore joined #salt
20:28 jeremyBass oh never mind on the stderr it was from the shell script it's self .. that was hard to tell
20:35 jeremyBass ok now all i need to do is figure out how to set a system env var from one state cmd.run and use it in another state cmd.run ..
20:36 jeremyBass I'm un sure why it's not working.. if i do it from the command line, set the var and then echo it on the next I see it set..
20:36 rofl____ joined #salt
20:36 jeremyBass so it's not like i don't have the command right to do it, but when the next state happens if i do a cmd.run: - name: set  that var is not there..
20:37 dlindquist joined #salt
20:40 jimallman joined #salt
20:42 jeremyBass it's odd .. in this test http://pastebin.com/WVKWrCfG you can see i set the env var, then echo it in the first cmd.run but then the second state cmd.run it's not there anymore
20:43 a_sk joined #salt
20:45 jalbretsen joined #salt
20:55 diegows joined #salt
20:55 honestly_ joined #salt
21:01 Ahlee cmd.run forks a process, when that forked process exits, any environment variables will be removed
21:03 jeremyBass yeah was starting to believe that was it.. so i need to correctly add it to the user profile then i'd guess
21:09 Ahlee either that, or create a temporary file to read the variable from
21:09 mgw joined #salt
21:10 jeremyBass this works .. cmd.run: -name: echo "export ADDED{{ ext_key|replace("-","") }}=True" >> /etc/profile and then in the next cmd.run: -name: set , i see the var now.. so yips
21:10 jeremyBass :D
21:15 a_sk joined #salt
21:17 ashtonian joined #salt
21:18 jankowiak joined #salt
21:29 BenCoinanke joined #salt
21:31 fejjerai joined #salt
21:31 forrest joined #salt
21:34 lzhang joined #salt
21:39 david_a joined #salt
21:39 fejjerai joined #salt
21:41 dlindquist joined #salt
21:44 fragamus joined #salt
21:46 fllr joined #salt
21:50 lzhang joined #salt
21:51 yomilk joined #salt
21:55 scalability-junk anyone from halite here?
21:55 scalability-junk just want to run some ideas against the inofficial roadmap and if I should try to built it into halite or separate.
21:55 coolj joined #salt
22:04 TronPaul_ joined #salt
22:04 TronPaul_ is there a way to get a masterless minion to output it's id?
22:04 TronPaul_ I'm trying to use state.highstate, but it's not matching the hostname
22:04 TronPaul_ nvm I think
22:05 markizano TronPaul_: salt['grains.get']('fqdn', socket.gethostname())
22:05 TronPaul_ cached minion_id
22:05 TronPaul_ is there a salt built in to clear that cached id?
22:05 TronPaul_ rather than removing /etc/salt/minion_id
22:05 TronPaul_ directly
22:06 markizano clear it ?
22:06 markizano wat ?
22:06 markizano TronPaul_: What are you trying to do ?? O.o
22:06 TronPaul_ salt-call state.highstate
22:07 TronPaul_ on a linode
22:07 TronPaul_ had run it before properly setting the hostname
22:07 TronPaul_ and it cached the linode default hostname
22:07 forrest scalability-junk, pass_by_value is your man, but I don't see him today, he'll probably be around monday though
22:07 scalability-junk forrest: alright I'll keep an eye out
22:08 ashtonian joined #salt
22:08 markizano TronPaul_: Try chaning the hostname on the machine, as root # hostname <name-you-want>.<domain>.<tld>
22:08 markizano unless it's a shared linode ? :x
22:08 TronPaul_ nah, I've got it working now
22:08 anuvrat joined #salt
22:09 TronPaul_ what happend was: 1 I ran salt-call w/ default linode hostname, 2 updated hostname, 3 tried to rerun salt-call w/o killing cached minion_id
22:09 markizano oh - i'm thinking in some automated fashion in salt itself, derp - yeah, you could just edit the file and restart the minion.
22:11 tvle joined #salt
22:18 scalability-junk is there a way to say require: database and then define the actual database within pillars or the top.sls file or so
22:18 scalability-junk this would make states a lot more reusable.
22:32 forrest what kind of DB scalability-junk? You could try just require: \n  - mysql_database: name
22:32 forrest or postgres_database
22:33 scalability-junk forrest: the thing is that it would be more reusable when mysql | maria | $db would be specified within a pillar instead of reusable state logic. or am i just wrong in thinking so?
22:34 forrest scalability-junk, you could always do some logic
22:34 forrest where you base it on pillar data
22:35 scalability-junk mhh true
22:35 forrest so it would be require: \n  - {{ my_db_type }}: {{ my_db_name }}
22:35 forrest obviously pulling from pillar though
22:35 scalability-junk sounds legit.
22:35 scalability-junk would this something to be good to add to the state guides?
22:36 forrest hmm, I don't really know. I don't about using two variables like that as a 'normal' occurrence
22:38 scalability-junk mhh alright probably makes reuse too important and not taking ease of access etc. into account.
22:38 forrest scalability-junk, a lot of the time I'd think it should be in different states, but it's up to the user
22:38 forrest both are good, just seems like you should clearly have a postgres state, versus a mysql state
22:39 forrest unless your app uses two types of databases, in which case you should tell your devs they suck
22:39 scalability-junk yeah you have postgres and mysql state. but looking at wordpress statest you have the wordpress state which requires a database, a webserver and php.
22:39 scalability-junk instead of requiring mysql
22:40 scalability-junk the state for the specific database will still be separated
22:40 scalability-junk it's more about reusability.
22:40 * scalability-junk is more brainstorming about reusable states than any real production use right now.
22:40 forrest right, so that kind of scenario seems like you should break it out, so you have a wordpress state, a database (mysql/postgres) state, a few webserver states, and a few php states. Then you can include whichever ones you need in your app
22:41 forrest yea I'm all about the reuse, but I'll take clarity in two states, over cleverness in a single state any day
22:41 forrest learned that the hard way with puppet
22:42 scalability-junk yeah yeah you don't see my point I think
22:42 scalability-junk I am suggesting a mysql, maria, postgres, wordpress, php, apache state etc.
22:42 scalability-junk and the wordpress state requires a database, a webserver and php
22:42 scalability-junk but wordpress shouldn't specify apache and mysql for example.
22:43 scalability-junk if they do for each specific setup reuse you need to edit it.
22:43 forrest yea gotcha
22:43 scalability-junk would be much easier to rely on pillars.
22:43 scalability-junk cause pillars are always "new"
22:43 forrest well, that would be up to the end user to just modify the state to include that I think
22:43 forrest yea fair enoug
22:43 forrest *h
22:44 scalability-junk I'm not sure if salt formulas should recommend reuse over ease of usage
22:45 forrest honestly I just don't want to put that much data into the pillar
22:45 forrest takes longer to generate that way, and then we run into the issue of 'where does this end'
22:45 forrest so now our includes go into the pillar? And our requires for the TYPE of require?
22:46 forrest at some point someone is gonna put everything in pillar
22:46 forrest and it will make me sad
22:46 scalability-junk that's what I wanted to understand
22:46 scalability-junk so salt tries to be reusable, but for specifc setups.
22:47 scalability-junk the thing is it would be fun in the future to make: salt deploy wordpress-apache-mariadb-varnish-whatever
22:48 Ahlee why not have them all as independent states?
22:48 Ahlee or am i missing something here
22:49 scalability-junk Ahlee: you mean wordpress-apache-mariadb, wordpress-nginx-mariadb etc. within a new state?
22:49 scalability-junk oh man that makes them sooo not predictable.
22:49 Ahlee no, i mean individual states for apache, mariadb, wordpress, etc
22:50 scalability-junk Ahlee: yeah we are at that state ;)
22:50 scalability-junk sure each tier has it's own state
22:50 forrest honestly if the user can't figure out to change include from mariadb to mysql
22:50 forrest then they probably should be reviewing docs.
22:51 scalability-junk forrest: it's not the he can't figure out, but the should it be specified in the state
22:51 forrest in the name of the state?
22:51 forrest I would not do that
22:51 forrest but that's me *shrug*
22:51 scalability-junk nope in the require statement
22:51 Ahlee if it's that important, why not just make a pillar for db type? i mean, the only difference i know off between what you'd do with salt with those is install, so just hte package name changes
22:51 scalability-junk am I talking so horrible english here :D
22:51 forrest Ahlee, that was what I suggested above
22:51 Ahlee yeah
22:51 Ahlee but i'm not scrolling back
22:51 Ahlee :P
22:52 scalability-junk Ahlee: ^
22:52 scalability-junk Ahlee: ^^
22:52 forrest heh
22:53 forrest tedski, Why did you fork my salt fork?
22:53 forrest tedski, or did you already fix that?
22:53 yomilk joined #salt
22:54 scalability-junk Ahlee, forrest: the main point (brainstorming) I wanted to make is. should we encourage the state formulas (reusable ones) have specific databases within require statements or should we edit the guidance to say use pillars for the types.
22:55 forrest scalability-junk, the formulas should provide functional support without having to implement pillars, so forcing a user to set that value is not preferred. We should have a sane default, which if they want to change, or set up via pillar, they always can
22:55 jfzhu_us1 joined #salt
22:55 Ahlee I'd second that.
22:56 forrest it's bad enough we have to use jinja 'if' statements to support some freebsd stuff, or even include files in a few formulas (which we should not be doing really)
22:56 tedski forrest: i fixed that... why? because dum.
22:56 scalability-junk forrest: I'm not interfering with the idea of having a default pillar ;)
22:56 tedski forrest: clicky clicky clicky and i don't get along well... heh
22:56 forrest tedski, yea I looked at your pulls and realized you did, just wanted to make sure since I wouldn't approve a pull request back onto my fork, then do a pull on the main repo :D
22:57 munderwo joined #salt
22:57 Ahlee Pillars are a bitch when you're first starting out, and the duplication of targetting and a completely seperate (yet named the same) config file
22:57 scalability-junk forrest: alright so the stand is reusable states are examples and not production ready "packages"
22:57 tedski forrest: sorry about that
22:57 tedski forrest: i'm new here and getting better, though :)
22:57 forrest tedski, No need at all to apologize man! It doesn't negatively impact me at all, I just wanted to make sure it wasn't going to cause you pain later! Good work so far it looks like here as well
22:58 forrest scalability-junk, they are the 'basics', so you could download them, and get going with a basic configuration, so in a sense they are production ready, but only for basic stuff.
22:59 tedski forrest: yeah, i was looking at your pull req to an issue, and said, "wait, i can take care of the other todo!!" and got too excited.  haha... then i forked your repo... then i realized as i submitted a pull req what the fork was from... doh!  rm -rf; git clone
22:59 forrest tedski, haha, which issue were you looking at?
23:00 tedski iirc, it was a rabbitmq one?
23:00 scalability-junk forrest: alright. I thought perhaps we wanted them to be some sort of production ready state, which gets improved by everyone using it. but that's probably too idealistic to be working.
23:00 forrest tedski, ahh alright
23:01 forrest scalability-junk, some of them are pretty advanced, and totally able to be used that way. But I assume almost all, if not all users would require SOME changes for their specifi environment.
23:01 forrest *specific
23:01 forrest so far we've really been focused on making the formulas work for every distro (which still has quite a ways to go), and from there we can start adding more states for more advanced stuff.
23:03 scalability-junk forrest: as I said it's mostly my saturday night open source brainstorming where should I put my work into time.
23:03 forrest tedski, my bad if that first statement came off a bit 'harsh', I always forget there isn't an IRC inflection detector.
23:03 tedski it didn't
23:03 tedski i just hate being the noob
23:03 tedski haha
23:04 tedski don't worry, no offense taken or anything
23:04 forrest scalability-junk, Yea for sure, gotta keep thinking to make it awesome
23:04 forrest tedski, cool
23:04 tedski i've got a thick skin
23:04 scalability-junk But it depends on what we/saltstack community want. It seems to be more some even advanced examples > really one maintained state used my 80% of the community with some additional overwrite states.
23:04 tedski you'll have to try much harder to offend me, sir.
23:04 forrest tedski, Yea as do I, which sometimes makes my comments a bit harsh since they wouldn't bother me :P
23:04 zzzirk joined #salt
23:04 scalability-junk tedski: being the noob is awesome. the learning curve is so much more steep :D
23:04 forrest scalability-junk, and it totally could evolve in that way, I have no idea where formulas might go thus far.
23:05 forrest err into the future I gues
23:05 forrest *guess
23:05 jeremyBass i need to echo {{ "#salt-set REMOVE" }} on the cmd.run .. but seems that the # gets read.. who do you excape it? \# just ends up with \\# in the file but it does not error
23:05 forrest man I can't type today
23:05 jeremyBass any idea?
23:05 tedski jeremyBass: have you tried hard quotes?
23:05 scalability-junk forrest: yeah but there are no real guidances here. if there was declared that the way would be reusable states with most stuff in pillars and overwriting takes place within overwriting-states it would give another impression as here are some examples.
23:06 tedski the soft quotes are removed by jinja, then bash parses the # as a comment is what i'm thinking
23:06 jeremyBass @tedski, yes i did
23:06 tedski s/bash/$SHELL/
23:06 kalloc joined #salt
23:08 forrest scalability-junk, I'm actuall working on writing a best practices, but I don't know if I'll get that deep into that stuff. It's already something like 250 lines long with the examples I've done so far, and it's not even 1/4 done.
23:10 scalability-junk forrest: give me a repo/gist/versioned_editable_document and I'm not too unmotivated to get my hands dirty or at least give a few ideas
23:10 forrest scalability-junk, Yea I need to get the structure done a bit more before I push it back to even my own repo, once I do though I'll let you know
23:10 munderwo joined #salt
23:11 scalability-junk forrest: alright (just as a side note. I'm quite familiar with unstructured writeup and don't mind getting into that)
23:11 forrest Yea for sure, I'm just not ready to put it up yet. I assume it's gonna need a lot of revisions.
23:13 jeremyBass i did  {{ "\#salt-set REMOVE" }}  and   {{ '\#salt-set REMOVE' }}  both pass but.. when it's echoed to the file, i get the \# because jinja renders \\# and that is what leads to \# being entered into the file
23:13 scalability-junk forrest: then let me know when it's "ready"
23:14 forrest will do, I want to try and get something pushed back to my own repo today/tonight, so hopefully something will be there :P
23:15 scalability-junk ^
23:15 scalability-junk ^^
23:15 scalability-junk keep my suggestions in mind. We have to decide what we want. easy/advanced examples or getting really reusable base states, which only get overwritten in 80-90% of the time.
23:15 phyde joined #salt
23:16 forrest yea, formula wise I'd rather have easier to get into stuff since it's targeting newer users
23:16 forrest and it gives them a way to improve their skills as they expand on them
23:16 forrest but we'll see what happens
23:18 fllr joined #salt
23:18 smcquay joined #salt
23:23 ajw0100 joined #salt
23:25 scalability-junk forrest: not sure if formulas should be the highschool for new users.
23:26 scalability-junk I would suggest them having their play in the docs and example repos linked from there. not "production" states as they are sometimes linked from the example deploys.
23:26 scalability-junk if you look into the django deploy even multinode deploys with the formulas as base are suggested.
23:27 forrest the django one is special
23:27 forrest as it is intended to be a full deploy of the application.
23:27 forrest you can see that sroegner has done some similarly complex stuff with the hadoop formula
23:27 forrest just depends on the needs of that specific application
23:28 Vivi-Idle joined #salt
23:29 scalability-junk yeah that's what I'm trying to get clear on. Do we want formulas to be the go to basis and recommend to only overwrite them and perhaps contributed great overwrite back. or do we want them as a playing field. right now there is no clear sign for any of these possibilites
23:29 scalability-junk *possibilities
23:29 scalability-junk *overwrites
23:30 forrest scalability-junk, I'm really not sure.
23:32 rstone joined #salt
23:34 diegows joined #salt
23:44 teebes joined #salt
23:46 jeremyBass ok anyone know how i can get `   echo "export ADDED=True {{ "\#salt-set REMOVE" }}" >> /etc/profile   ` to work right?  I need `   export ADDED=True #salt-set REMOVE   ` added to that file, but i get `   export ADDED=True \#salt-set REMOVE   ` .. that is because jnjna renders `  echo "export ADDED=True \#salt-set REMOVE" >> /etc/profile   ` .. i need to drop that \ but not have the # be rendered as a comment in jnjna
23:54 munderwo joined #salt
23:59 jeremyBass ok figured it out .. `   'echo "export ADDED=True {% raw %}#salt-set REMOVE{% endraw %}" >> /etc/profile'   `

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