Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2014-02-16

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

All times shown according to UTC.

Time Nick Message
00:03 gnugnu5_ joined #salt
00:04 mgw1 joined #salt
00:05 scott_w_ joined #salt
00:08 scoates kiorky: ? was that for s0undt3ch ?
00:13 bhosmer joined #salt
00:14 forrest ok guys, look for some feedback regarding the sections of this doc, and if you can think of stuff that should go in here: https://github.com/gravyboat/salt/blob/develop/doc/topics/best_practices.rst
00:16 thayne joined #salt
00:17 sieve joined #salt
00:19 arnoldB forrest: did you know that salt (at least develop) is able to read automatically from defaults.jinja / defaults.json now?
00:19 arnoldB forrest: might be useful to updating this doc/ formular doc concerning this new feature
00:20 forrest arnoldB, you mean regarding the usage of the from x import y? instead just putting that inside defaults.jinja?
00:20 arnoldB forrest: yes
00:20 arnoldB ( https://github.com/saltstack/salt/issues/10300 )
00:20 forrest arnoldB, I did not, I don't know what the intended usage of defaults.jinja is going to be, seems like that would be different than the map file
00:21 forrest that's good to know though
00:22 arnoldB forrest: as far as I understand it correcty, it's nearly the same thing as the import map.jinja feature
00:22 warmwaffles arnoldB: do you know of any decent examples of state and pillar configurations? I'm sure I am doing this terribly wrong
00:22 arnoldB forrest: warmwaffles take a look at the links I've posted some minutes ago
00:23 arnoldB argh sry
00:23 arnoldB warmwaffles: ^
00:23 forrest arnoldB, right but the map is specifically intended to be used for ensuring that your formulas are os agnostic.
00:23 warmwaffles arnoldB: yea man I'm staring at those right now
00:24 warmwaffles https://gist.github.com/warmwaffles/9027343
00:24 warmwaffles does that look right to you?
00:25 arnoldB forrest: hm right
00:25 sieve left #salt
00:26 arnoldB warmwaffles: use /srv/salt instead of /srv/states if you haven't configured file_roots manuelly in your config
00:28 warmwaffles when setting up the gitfs_remotes is it better to have a separate repo for salt and one for pillar? or can I shove both in the same dir
00:28 vu I don't get the effort behind salt-ssh use case. It uses ssh concurrency and so?
00:31 forrest vu, it's for when you can't install a minion, maybe a server doesn't have the right version of python, or you don't want connections established all the time.
00:32 acabrera joined #salt
00:32 acabrera left #salt
00:32 acabrera joined #salt
00:33 vu forrest: I see
00:33 acabrera I'm wondering what the size my EC2 instance should be for a salt-master, any suggestions?
00:33 jtang1 joined #salt
00:33 forrest acabrera, how many minions?
00:33 acabrera atm < 100
00:34 forrest acabrera, are you going to use salt for remote execution only, or also for configuration management
00:34 acabrera CM
00:35 forrest Are you looking at EC2? I think the small is what? 1 core (with no thread details), and 1.75GB of RAM?
00:35 acabrera yeah, something like that
00:35 forrest I'd just start with small
00:35 scoates I have fewer nodes than that, currently, but my m1.small doesn't even flinch at ~20 nodes.
00:35 acabrera m1.small is what I was thinking
00:35 scoates don't use t1.micro
00:35 forrest most of the work is gonna be done on the minions anyways
00:35 forrest yea no not micro
00:35 forrest small
00:35 forrest lol
00:36 forrest jeez, that's like asking anything to fail
00:36 acabrera tis guys
00:36 forrest scoates, :P
00:36 acabrera s/tis/tks/
00:36 forrest yea np
00:37 forrest scoates, I feel like you could go around trolling people, 'yea the micro is fine! You will never get screwed by Amazon when they need resources!'
00:37 scoates t1.micro : EC2 instances :: vermouth : churchill martini
00:37 scoates (it's ok to look, but NEVER actually use it)
00:38 xfactor973 joined #salt
00:38 acabrera micro is great for something like qmail/ezmlm
00:38 xfactor973 can i use grains to match a hostname regex in a jinja template?
00:38 scoates only if you never load it
00:38 acabrera For those old-time things that don't need memeory
00:38 forrest xfactor973, of course!
00:38 xfactor973 hey forrest how's it going
00:38 forrest another fantastic day
00:38 scoates it's not about memory. it's the annoying burst + punish cycle
00:39 scoates the one where ssh stops returning keystrokes because the instance is at >99% CPU steal
00:39 forrest xfactor973, you can do something like {% if 'asdf' in grains['hostname'] %} \n my stuff here {% endif %}
00:39 xfactor973 forrest: i have 3 servers called dlceph01,02,03 and i have this in my template {% if grains['host'] == 'dlceph*' %}.  When i run highstate though nothing happens
00:39 xfactor973 oh that's what i'm missing
00:39 forrest no endif?
00:39 xfactor973 yeah i have an endif also
00:39 forrest that reminds me I need to write a friggin jinja tutorial for salt
00:39 xfactor973 haha
00:40 xfactor973 that'd be useful
00:40 xfactor973 so can i match things like that?
00:40 forrest xfactor973, I don't know if * works correctly in that context, I'm trying to remember.
00:40 xfactor973 from what i'm seeing it doesn't
00:40 scoates no * there
00:40 scoates you might be able to use startswith ?
00:40 xfactor973 so i think i should switch it to if 'dlceph' in grains['hostname']
00:41 scoates or (annoyingly) write a module function that matches with regex… which is what I did…
00:41 forrest host is fine if that's returning the hostname
00:41 forrest xfactor973, you can use normal python stuff in there
00:41 xfactor973 oh?
00:41 forrest yea
00:41 xfactor973 interesting
00:41 forrest that's pretty much what I'm doing with if in
00:41 forrest try to keep it simple if you can't, don't get crazy like scoates
00:41 acabrera What ports need to be open for the salt master?
00:41 xfactor973 right
00:42 arnoldB xfactor973: are you working on a public ceph formula?
00:42 scoates yeah. don't (-:
00:42 xfactor973 yeah simple is good
00:42 forrest acabrera, 4505 and 4506 on the master, the tutorial has some stuff on that as well I believe if you're going through that.
00:42 xfactor973 arnoldB: it's not public but i suppose i could make it public.
00:42 forrest Hey has anyone seen yogi around today?
00:42 xfactor973 so far it only installs packages and sets up the directories
00:42 xfactor973 i haven't gotten further than that
00:43 arnoldB alright
00:43 xfactor973 arnoldB: do you have a ceph cluster as well?
00:44 xfactor973 forrest: that jinja line you wrote works great :)
00:44 forrest sweet!
00:44 arnoldB xfactor973: 3 x dell r320 systems are waiting for it
00:44 xfactor973 awesome :)
00:45 forrest scoates, did you put that regex module in contrib?
00:45 xfactor973 arnoldB: do you need some help getting the salt setup to work for ceph?
00:45 scoates forrest: no; it does weird matching on some other pillar data, so it's not reusable.
00:45 xfactor973 i can get you part way there
00:45 scoates I'm going to rewrite those states in #!py
00:45 forrest scoates, gotcha
00:45 forrest let me know if you guys want a ceph formula repo made
00:46 forrest actually I'll just make it
00:46 scoates got to run.
00:46 forrest https://github.com/saltstack-formulas/ceph-formula there you go
00:46 forrest fork and contribute whenever you're ready and one of us will approve the pull req.
00:47 schimmy joined #salt
00:47 xfactor973 sweet
00:47 xfactor973 haha there's nothing there
00:47 xfactor973 easy fork
00:47 arnoldB xfactor973: I'm going to build a feature-rich ceph formula and looking for code to start with. but starting for scratch shouldn't be a problem
00:47 forrest yea I just initialized it with a blank repo minus the readme :P
00:47 xfactor973 hehe
00:47 xfactor973 arnoldB and forrest: lemme fork this repo and add my 2 cents.
00:48 xfactor973 i'll have something in a little bit
00:48 forrest cool
00:48 arnoldB cool
00:51 felixhummel joined #salt
01:00 xfactor973 arnoldB: so here's my rough example: https://github.com/cholcombe973/ceph-formula
01:00 xfactor973 all it does is setup the packages and make the directories
01:00 xfactor973 i haven't gotten to using the ceph-deploy tool to actually setup the cluster or anything
01:02 xfactor973 there's a lot of hard coded stuff in there that needs to be removed
01:03 arnoldB xfactor973: cool. that looks like the best code so far on github (there aren't a lot :D )
01:03 xfactor973 lol
01:03 xfactor973 terrible
01:05 xfactor973 maybe you or forrest can help me get rid of the hard coded crap
01:05 forrest I don't know, that sounds like work
01:06 xfactor973 lol
01:06 forrest it's not that bad looking at it
01:06 xfactor973 well i'm only intermediate at salt.
01:06 xfactor973 i learned pillar last week :)
01:06 forrest heh
01:06 arnoldB xfactor973: I will start to setup ceph within a few days. this includes the creation of a ceph formula. I can fork your code and refactor it
01:06 xfactor973 ok
01:07 forrest yea I'll help out if arnoldB doesn't feel like it, tonight though I'm squarely focused on some documentation I want to finish up, been letting the doc issues get out of hand because I've been too lazy to fork another copy :P
01:07 xfactor973 i think salt could probably build the ceph.conf but it's just a theory at the moment
01:07 forrest it probably can
01:07 forrest and I know nothing of ceph
01:07 xfactor973 forrest: any idea if there will be a ceph backend in addition to s3, git, hg?
01:08 forrest xfactor973, you know I'm not sure, terminalmage loves learning new stuff like that, so we should ask him
01:08 forrest whenever he is around that is
01:08 xfactor973 ok
01:09 xfactor973 arnoldB: word of caution.  ceph only posts the latest point release of the packages on their site.  this formula works for awhile and then they push another point release and it breaks
01:09 xfactor973 they pull down previous releases off of their site
01:10 arnoldB should be no problem for me
01:10 dancat joined #salt
01:10 xfactor973 i set it up this way so that it doesn't roll the cluster whenever a new release comes out
01:10 xfactor973 that'd be annoying
01:14 s0undt3ch left #salt
01:16 acabrera I have the salt master and minion installed
01:16 acabrera it doesn't seem that the minion sent a key to the master
01:17 acabrera not sure what to do once I installed the software, is there a tutorial page?
01:17 forrest acabrera,
01:17 forrest http://docs.saltstack.com/topics/tutorials/walkthrough.html
01:17 forrest ctrl v fail!
01:17 ajw0100 joined #salt
01:18 arnoldB acabrera: logs or it didn't happen
01:21 arnoldB argh I loose more time installing salt with my patches and the desired version instead of using salt itself
01:21 xzarth joined #salt
01:26 acabrera So does the minion send the key over to the master for it to be accepted?
01:26 arnoldB xfactor973: if you have useful feature requests, add them to the README.md or file a ticket on https://github.com/bechtoldt/ceph-formula or https://github.com/saltstack-formulas/ceph-formula . maybe I'll handle it
01:27 forrest arapaho, xfactor973 make the readme.md a readme.rst please
01:27 jtang1 joined #salt
01:28 forrest we have a structure for the readmes if you look at some of the other formulas
01:28 forrest or I can fix it later, doesn't really matter.
01:28 arnoldB forrest: why prefer rst over md?
01:28 forrest because rst is used for all the documentation files
01:29 forrest it's just consistency
01:29 arnoldB hm ok
01:29 forrest You can literally just go in and copy paste one of the other readme's 'raw' contents
01:29 forrest I believe I've fixed most, if not all of them
01:31 joehoyle joined #salt
01:33 krow joined #salt
01:33 xfactor973 forrest: ok sounds good
01:33 acabrera Ahh, I see that it does, well, that was too easy.
01:37 acabrera What should i see if I run salt '*' cmd.run 'ls -l /etc'?
01:37 forrest acabrera, the output of that command
01:38 forrest acabrera, is it just dropping you back to the console?
01:38 acabrera yes, after about 10-20s
01:38 forrest start with salt '*' test.ping
01:38 acabrera Lots lot log messages on master
01:38 acabrera no messages on minion logs
01:39 forrest did you connect the minion up and accept the key?
01:39 forrest test.ping should return 'true'
01:40 acabrera [root@salt-master ~]# salt '*' test.ping
01:40 acabrera [root@salt-master ~]#
01:40 forrest try salt '*' test.ping -t 10
01:41 forrest if you still don't see anything, confirm you see the minion is accepted with salt-key -L, then if that still looks good, do netstat -aln | grep 4505, and netstat -aln | grep 4506 to ensure that you see the IP of the minion connected.
01:42 acabrera [root@salt-master ~]# salt '*' test.ping -t 10
01:42 acabrera [root@salt-master ~]#
01:43 acabrera I know that the minion can talk to the master since it sent a key to the master
01:43 forrest which you have already accepted?
01:43 acabrera I do see it w/ salt-key -L
01:43 acabrera yes
01:43 forrest ok, what does netstat say?
01:43 acabrera [root@salt-master ~]# netstat -aln | grep 4505
01:43 acabrera tcp        0      0 10.160.47.53:4505           0.0.0.0:*                   LISTEN
01:44 acabrera [root@salt-master ~]# netstat -aln | grep 4506
01:44 acabrera tcp        0      0 10.160.47.53:4506           0.0.0.0:*                   LISTEN
01:44 forrest ok yea, so the minion isn't connected.
01:44 forrest the minion's IP should be in there
01:44 acabrera In minion log [DEBUG   ] Attempting to authenticate with the Salt Master at 10.160.47.53
01:44 forrest is the minion service running?
01:44 forrest any iptables that might be blocking?
01:45 terminalmage xfactor973: I'm focusing on making the git, hg, and svn backends awesome right now, but I'd be happy to look at it
01:45 xfactor973 sweet
01:45 terminalmage I'd just need to have something against which to test
01:45 xfactor973 which lang do you prefer, python or c?
01:45 acabrera [INFO    ] Authentication request from nexus
01:45 acabrera [INFO    ] Authentication accepted from nexus
01:45 forrest acabrera, can you telnet from the minion to the master on 4505 and 4506?
01:46 terminalmage xfactor973: python... I'm lazy ;)
01:46 xfactor973 haha
01:46 xfactor973 alright lemme get the docs for python
01:46 terminalmage xfactor973: this is my latest pull req for gitfs: https://github.com/saltstack/salt/pull/10468
01:46 xfactor973 terminalImage: http://ceph.com/docs/next/rbd/librbdpy/
01:46 xfactor973 ok lemme have a look
01:46 terminalmage xfactor973: nice, bookmarked
01:47 xfactor973 terminalImage: librdb is the low level ceph interface.  it should be easier to work with than the ceph filesystem interface
01:48 terminalmage ok
01:48 forrest xfactor973, he's terminalmage, not terminalimage :P though that is more comical
01:48 terminalmage I know nothing about it at the moment
01:48 xfactor973 :D
01:48 terminalmage so I'd need to RTFM
01:48 xfactor973 ok
01:48 terminalmage but I did that for mercurial
01:48 forrest WORK HARDER ON SATURDAY terminalmage!
01:48 xfactor973 thanks forrest.  i guess my font size is too small :D
01:48 terminalmage ha
01:48 forrest xfactor973, no worries.
01:48 terminalmage I'm gonna go get dinner
01:48 acabrera forrest: AWS security group misconfiguration.  thanks for your help
01:49 forrest later
01:49 terminalmage we just had our furnace go out so I had to pay $600 to have a guy come out on a Saturday and repair it
01:49 forrest acabrera, yea np, I think about 90% of people who have come in here unable to connect on AWS have had issues with the security group stuff.
01:49 forrest terminalmage, psssh
01:49 forrest you should have just flown out here, would have been cheaper
01:49 forrest I'm running my AC right now
01:50 terminalmage haha, you love mentioning that to me don't you
01:50 terminalmage jerk
01:50 forrest yes
01:50 forrest because you hate the cold so much
01:50 forrest it always result in some sort of funny comeback
01:50 xfactor973 man this gitfs.py file is long as hell
01:51 terminalmage xfactor973: yeah it didn't used to be, but it now supports three different python backends
01:51 xfactor973 heh
01:51 terminalmage so, a lot of stuff had to be written 3 times
01:51 xfactor973 that'll certainly do it
01:51 bhosmer joined #salt
01:51 terminalmage s/backends/frontends/
01:51 terminalmage the new support is all in develop
01:51 terminalmage won't be officially released for a while
01:52 terminalmage it's really new
01:52 terminalmage anyway, I'm out. have a good night, and forrest I hope your A/C dies :P
01:52 xfactor973 haha
01:52 forrest terminalmage, lol thanks
01:52 terminalmage :)
01:53 xfactor973 forrest: where are you hailing from?
01:53 forrest it's almost into the 70s now, so I will open the windows soon :P
01:53 forrest xfactor973, Arizona in the US
01:53 xfactor973 oh cool
01:53 xfactor973 i'm in CA
01:53 forrest STEALIN MAH WATER
01:53 xfactor973 nah i'm on a well
01:53 xfactor973 san jose though... they're stealing your shit :D
01:53 forrest ahh ok
01:53 forrest yea san jose always does though
01:53 forrest friggin power and water thieves
01:53 xfactor973 yep
01:54 xfactor973 without that the place would be a desert
01:54 forrest it should be, no one should have a grassy lawn out here :\
01:54 xfactor973 nope
01:54 xfactor973 i don't get the whole lawn thing out here
01:54 forrest Maybe you just don't like to waste money
01:54 xfactor973 i moved from PA.  lawns happen there because it rains all damn year
01:54 forrest yea
01:55 xfactor973 but out here it's silly
01:55 forrest I'd just be too lazy to maintain it
01:56 xfactor973 heh
01:56 xfactor973 it's a lot of work
01:59 ndrei joined #salt
02:01 arnoldB hm I need to set a high timeout now (long cmd.run command running) when executing state.highstate now. the messages "Execution is still running on host" doesn't appear. what could be wrong?
02:02 bhosmer joined #salt
02:02 forrest arnoldB, well, salt by default only has a 5 second 'timeout' basically it waits 5 seconds, then goes 'hey minion what are you doing' then waits something like 2 seconds for a response, if it doesn't hear back it drops you to the console
02:02 joehoyle joined #salt
02:02 forrest the job still runs
02:02 forrest and you can query the status via the jobid and such
02:03 forrest but you can increase that timeout value in either the master config, or straight from the commandline with the -t option
02:03 arnoldB forrest: already done
02:03 arnoldB the master doesn't get the reply from the minion that the job is still running
02:03 forrest ok, I'd run it through with -l debug and see if that helps, if not hop over to the minion and see what is going on there, or restart the minion with debug logging turned on for more output
02:03 forrest did the job actually finish?
02:04 forrest what happens if you query the jid?
02:04 joehoyle joined #salt
02:04 arnoldB the job finished fine. I also got the expected results
02:04 forrest oh
02:04 forrest so what is the problem then?
02:05 arnoldB the master doesn't get the reply from the minion that the job is still running
02:05 forrest but you just asid the job finished
02:05 forrest *said
02:05 forrest or do you mean during the run you don't want to be dropped to the console?
02:07 arnoldB I don't want the process of state.highstate I've started to be killed automatically. the master doesn't get the minion's reply. that's the reason why I executed state.highstate with a high timeout
02:07 arnoldB killed in favor of gracefully exited due to timeout
02:07 forrest is it exiting prior to the timeout you set being met?
02:07 jidar man, I'm really getting frustrated here
02:08 forrest jidar, it is Saturday, break time perhaps?
02:08 arnoldB forrest: it's exiting when using the default timeout only
02:08 jidar arnoldB: I really can't make sense of what you're doing with the /srv/salt/states stuff you posted to me earlier
02:08 jankowiak joined #salt
02:08 arnoldB jidar: /srv/salt/states ?
02:08 jidar https://np.adm.sx/view/9cf1b7c6
02:09 arnoldB hm with "move" I meant "rename". sorry lol ...
02:09 jidar I guess I just don't understand the relationship pillars have in terms of file location and use of top.sls
02:10 jidar arnoldB: https://np.adm.sx/view/ad405902 this is the pillars/top.sls file right?
02:10 arnoldB jidar: yes
02:10 jidar why would you be using host names here?
02:10 arnoldB jidar: because I can
02:10 jidar but.. isn't that not what a pillar is supposed to be?
02:11 arnoldB you can use whatever filenames you want
02:11 arnoldB think of having host specific pillar data
02:11 jidar arnoldB: that's another issue I'm having here, let me explain
02:12 jidar http://docs.saltstack.com/topics/tutorials/pillar.html lists that you can make a /srv/pillar/top.sls that has a /srv/pillar/data.sls, and then there are vairables in there
02:13 jidar I put `foo: some data` in my /srv/pillar/data.sls file, and when I run salt-call --local pillar.get foo I get nothing
02:13 krow joined #salt
02:13 jidar I poked around in the pillar.items, and didn't find it there either
02:14 arnoldB cat /srv/salt/pillar/top.sls
02:14 jidar base: '*' - data
02:14 arnoldB https://np.adm.sx/
02:14 arnoldB hm
02:15 arnoldB are you using a masterless setup for a specific reason?
02:15 jidar https://np.adm.sx/view/924a9d2c
02:15 jidar arnoldB: I was wondering if that's going to be an issue
02:15 arnoldB I guess it would be easier to start with a usal master minion setup
02:16 jidar arnoldB: I was really trying to setup standalong services for the time being
02:16 jidar arnoldB: just so I could reduce the inter-dependency needs
02:16 arnoldB install master; mv /etc/salt/minion /etc/salt/master; echo 'master: 127.0.0.1' > /etc/salt/minion;
02:17 acabrera joined #salt
02:17 arnoldB restart minion+master; salt-key -A; salt '*'
02:18 arnoldB state.highstate test=True
02:22 jtang1 joined #salt
02:32 krow joined #salt
02:33 dleedyna joined #salt
02:46 MTecknology gitlab seems pretty awesome except for the ruby part
02:46 forrest well, github is ruby
02:46 joehoyle joined #salt
02:46 MTecknology ya... and that makes me surprised that it runs at all
02:47 arnoldB MTecknology: *sign*
02:48 arnoldB MTecknology: we're using gitlab at project. it does it's job pretty well
02:48 arnoldB *a
02:49 MTecknology I'll believe you when you say that it gets the job done, not when you say get it done well
02:50 arnoldB what are your requirements?
02:51 terminalmage joined #salt
02:51 MTecknology not ruby, java, .net, mono, or jcl
02:51 forrest PHP IT IS THEN CAPTAIN!
02:52 MTecknology I'd take php over ruby
02:52 forrest DECLARE THE VARIABLES THAT MAKE NO SENSE! SCREW THE STANDARDS BECAUSE THERE ARE NONE ARGHHHHHHH
02:52 forrest I don't think I'd go that far
02:52 forrest I've seen some very very poorly written php
02:52 MTecknology I've seen some very very poorly written cobol
02:52 MTecknology I've seen some very very poorly written python
02:52 MTecknology I've seen some very very poorly written C
02:53 forrest yea
02:53 MTecknology poor devs can make a mess of anything
02:53 arnoldB ha, you bloody newbies. I have to write puppet code 8h x 5d
02:53 funzo_ joined #salt
02:53 forrest I did that at my last job arnoldB
02:53 forrest why do you think I contribute to salt
02:53 arnoldB the same reason as I do :D
02:54 forrest exactly
02:54 forrest so I can hopefully one day see the charred corpse of a puppet server
02:54 MTecknology I remember way back when salt was unusable; it happened to be right when I was first looking into automation and config management
02:54 MTecknology I decided if puppet really was the best option, I wanted nothing to do with any of it
02:54 MTecknology 2yr later, I started looking again and looked at what I liked best back then
02:55 MTecknology all of a sudden, salt looked pretty dang awesome
02:55 forrest yup
02:55 MTecknology I kinda wanna go into work and get a git server up and running
02:56 MTecknology unfortunately, cgit won't be enough for them and neither will <can't remember the name>
02:57 forrest don't go into work man...
02:57 forrest Has your hatred already fizzled?
02:57 MTecknology nah, but... I love what I do
02:57 MTecknology and I'm hourly and anything I do today is time and a half
02:58 forrest *shrug*
02:58 forrest yea I am not
02:58 forrest so unless I am on-call
02:58 forrest they aren't getting shit from me on a weekend.
02:58 forrest I have actual work to do which I enjoy
02:59 MTecknology I basically own the entire network. I can go into just about every single server
02:59 forrest yea
02:59 MTecknology all but about... 50 out of the 700+; when I tell the networking guys to do something, I'm not questioned
02:59 MTecknology it's nice
03:00 MTecknology but... ya... I've ranted enough
03:02 MTecknology forrest: pretty much - I've made the environment so that I can do fun things and keep them directly work related - things like adding salt to everything
03:02 forrest fair enough
03:03 MTecknology right now I want to put up a git server...
03:03 forrest use gitlab man
03:03 forrest it's good
03:03 herzi joined #salt
03:03 xfactor973 forrest: i have a friend that uses it at drexel and he loves it also
03:03 forrest I can tell you this, almost all, if not all the git projects like that are ruby
03:04 forrest so you might as well use gitlab
03:04 forrest xfactor973, gitlab?
03:04 xfactor973 yup
03:04 forrest yea gitlab is the best
03:04 forrest we had some SHIT one at my last company, was the worst
03:04 forrest still written in ruby
03:04 forrest so you should just use gitlab, they have a freebie version you can use for a trial or whatever I think too.
03:04 xfactor973 nice
03:07 forrest MTecknology, really though, just go gitlab
03:07 forrest we messed with 3-4 other solutions, they were all awful
03:07 forrest or just half baked
03:07 arnoldB ack
03:07 forrest so they didn't have basic functionality you needed
03:08 forrest and eventually (as I was leaving) they decided to go with gitlab
03:16 jab416171_ joined #salt
03:16 jeddi joined #salt
03:16 jtang1 joined #salt
03:17 DaveQB joined #salt
03:18 MTecknology forrest: that's my intention... the painful part will be updating viq's formula for debian
03:18 forrest I thought his formula was for debian
03:18 MTecknology for rhel
03:19 herzi joined #salt
03:19 forrest oh gotcha
03:19 forrest yea
03:20 MTecknology DANGIT!
03:21 MTecknology It's time to change my password again...
03:21 rostam joined #salt
03:22 pcarrier joined #salt
03:28 patrek joined #salt
03:31 vbabiy_ joined #salt
03:32 arnoldB 4:32 am. gn8
03:33 forrest later
03:42 MTecknology g'nate?
03:42 vbabiy_ joined #salt
03:44 vbabiy_ joined #salt
03:50 bhosmer joined #salt
03:50 favadi joined #salt
03:53 vbabiy_ joined #salt
03:55 MTecknology I'm starting to wonder if it's not perhaps better to install gitlab manually
03:56 vbabiy__ joined #salt
03:58 vbabiy_ joined #salt
04:00 justBob joined #salt
04:02 vbabiy_ joined #salt
04:06 vbabiy_ joined #salt
04:07 forrest ok who wants to review https://github.com/gravyboat/salt/blob/develop/doc/topics/best_practices.rst for me?
04:08 forrest it's only like 525 lines :P
04:09 vbabiy_ joined #salt
04:10 jtang1 joined #salt
04:11 thayne joined #salt
04:12 vbabiy_ joined #salt
04:13 vbabiy_ joined #salt
04:13 mgw1 joined #salt
04:13 vbabiy_ joined #salt
04:18 fragamus joined #salt
04:18 vbabiy_ joined #salt
04:22 vbabiy_ joined #salt
04:23 joehoyle joined #salt
04:26 vbabiy_ joined #salt
04:28 psyl0n joined #salt
04:29 vbabiy_ joined #salt
04:31 vbabiy_ joined #salt
04:32 schristensen joined #salt
04:32 joehoyle joined #salt
04:35 elfixit joined #salt
04:37 vbabiy_ joined #salt
04:37 forrest vbabiy, you're really spamming hard here man
04:37 forrest damn didn't catch him before the disconnect
04:38 vbabiy_ joined #salt
04:39 MTecknology he wouldn't have seen it anyway
04:39 forrest yea I just tweeted at him
04:39 forrest hopefully he'll see it
04:41 MTecknology I'm trying to adapt that gitlab formula to debian and my environment - the packages seem to be less than fun to match up
04:43 dwfreed heh
04:44 schimmy joined #salt
04:44 MTecknology dwfreed: hm?
04:45 dwfreed MTecknology: just chuckling at the fact that every distro names their packages different
04:46 MTecknology ya- rather frustrating
04:46 MTecknology and they all have their own special reason to name it different
04:46 joehoyle joined #salt
04:59 krow joined #salt
05:04 joehoyle joined #salt
05:04 jtang1 joined #salt
05:07 MTecknology dang...
05:11 MTecknology I somehow accidentally installed some things from unstable... now I need to do a lot of damage control or keep things from unstable installed and keep installing some things from there
05:13 schimmy joined #salt
05:26 Vye joined #salt
05:27 psyl0n joined #salt
05:34 dave_den mgw; you around?
05:35 manfred is there a way to pass a list to an argument on the command line for a module? (i am trying to write my own module, and want to pass a list of peers)
05:35 mgw dave_den, what's up?
05:35 herzi joined #salt
05:35 manfred nm, just thought to look at the pkgutil.py one
05:35 dave_den mgw: do you have nacl renderer working? i have tried on 0.17.5 and two commits from develop
05:36 scristian joined #salt
05:36 mgw On develop, yes.
05:36 mgw Haven't tried 0.17
05:37 dave_den the latest develop commit?  519ceec
05:37 mgw1 joined #salt
05:37 manfred can anyone point me to where I can see how in salt/modules/pkgutil.py .... you switch to using pkg. for things like pkg.install ?
05:38 mgw1 dave_den, no my develop is a few days old
05:38 bhosmer joined #salt
05:38 dave_den mgw: i am getting stack traces when trying to do saltutil.sync_all, but it throws exception on returners, which is strange. if i remove /srv/salt/base/_renderes/nacl_renderer.py and do sync_all it works fine all the time
05:39 xerxas joined #salt
05:39 mgw1 can you gist the stacktrace?
05:39 mgw1 dave_den
05:39 dave_den yeah, but there's not info about nacl in it
05:39 dave_den sec
05:40 mgw1 ValueError: unknown url type: top.sls?
05:41 dave_den mgw: https://gist.github.com/dlanderson/85787ee7a21f55873f2e
05:41 dave_den yep
05:41 mgw1 i hadn't tried to sync
05:41 krow joined #salt
05:41 mgw1 I had it symlinked into the cache
05:41 dave_den are you just directly dropping the renderer
05:41 dave_den ah, ok
05:42 dave_den maybe i will just do that to test
05:42 mgw1 it's not even in sync_renderers
05:43 dave_den you just linked /var/cache/salt/minion/extmods/renderers/nacl_renderer.py ?
05:44 mgw1 yes
05:44 mgw1 dave_den: I think I may know why, and I was already working on a fix
05:44 mgw1 stand by
05:44 dave_den ah, sweet
05:47 dave_den even symlinking it is causing the breaking all states on the minion
05:47 dave_den will wait for your update
05:54 MTecknology I just got to do a bunch of work on our water cooler. Yay.
05:55 MTecknology mgw1: If you chose mwg, you would be the name of one of the most important applications I manage.
05:56 mgw1 MTecknology: I'll change my name right away.
05:56 mgw1 I promise
05:56 MTecknology :P
05:58 jtang1 joined #salt
05:59 schimmy joined #salt
06:02 dave_den mgw1: and does evan ever show up in #salt ?  what's his nick?
06:04 mgw1 I think he's fatbox
06:04 mgw1 I've not spoken with him here though
06:04 mgw1 My "fix" didn't fix that
06:04 dave_den ah, no worries
06:05 dave_den thanks for checking
06:05 dave_den i will see if i can look at it on monday if it's not already fixed by then
06:05 mgw1 dave_den: It's something in nacl.auto
06:06 mgw1 oh, maybe not
06:06 mgw1 yeah, it is
06:09 mgw1 it is, in particular line 16
06:11 krow1 joined #salt
06:17 dave_den hrm
06:17 dave_den it would probably be easier to implement directly into the salt codebase instead of trying to load libs
06:19 psyl0n joined #salt
06:28 mgw1 dave_den: that's likely
06:29 mgw1 dave_den: i'll look at this more Monday, the problem is SMinion is being loaded within the minion, and either that's not supposed to happen, or it just doesn't work
06:29 mgw1 It does seem to work other than the sync
06:30 krow joined #salt
06:30 mgw1 But I'd take the sync as a warning that something may be getting muddled by nacl that hasn't yet become apparent
06:31 mgw1 Evan originally developed it to be executed externally by the #py renderer, so it probably would not be an issue
06:32 mgw1 but now that it's being loaded by a proper renderer, it's a bit different and needs some testing
06:34 dave_den mgw1: cool, i'll look again monday too. i really like this vs py/pydsl and i think it'd be worth it to just integrate into the salt repo to get it working there
06:34 dave_den thanks for poking at it tonight.
06:34 mgw1 Agreed, as I mentioned in my email to the list.
06:35 dave_den i'm off to zzzz
06:35 mgw1 Hopefully Tom will agree too
06:35 mgw1 Have a good weekend
06:35 dave_den yeah, i'm gonna send an email to the list tomorrow
06:35 dave_den you too
06:35 mgw1 good idea
06:35 mgw1 keep the interest up :-)
06:37 forrest hey mgw1, dave_den if you guys have time take a look at: https://github.com/saltstack/salt/pull/10481
06:38 forrest let me know if you see something missing.
06:38 mgw1 forrest: will look later… getting late here
06:38 forrest no worries
06:39 dave_den forrest: saw your PR a bit ago. will look monday :)
06:39 forrest cool
06:53 jtang1 joined #salt
06:59 pmcg_ joined #salt
06:59 n8n joined #salt
07:01 linjan__ joined #salt
07:04 krow joined #salt
07:08 MTecknology woohooo! only 14 states fail!
07:10 MTecknology /usr/bin/c++: 2: /usr/bin/c++: clang++: not found
07:10 MTecknology *grumble*
07:11 xmj joined #salt
07:16 bhosmer joined #salt
07:20 krow joined #salt
07:26 bhosmer joined #salt
07:28 MTecknology I hate ruby more the more I use it.
07:28 MTecknology I love python more the more I use it.
07:42 n8n joined #salt
07:47 jtang1 joined #salt
07:48 JordanRinke joined #salt
07:57 n8n joined #salt
07:58 pdayton joined #salt
08:02 krow joined #salt
08:07 krow joined #salt
08:12 waverider joined #salt
08:13 krow joined #salt
08:14 jimallman joined #salt
08:14 ndrei joined #salt
08:21 ndrei joined #salt
08:21 krow1 joined #salt
08:26 thayne joined #salt
08:41 jtang1 joined #salt
08:53 krow joined #salt
08:55 yomilk joined #salt
09:08 jtang1 joined #salt
09:14 bhosmer joined #salt
09:28 JordanRinke joined #salt
09:31 krow joined #salt
09:33 scarcry joined #salt
09:40 zooz joined #salt
10:14 scott_w_ joined #salt
10:23 scott_w_ joined #salt
10:28 nn0101 joined #salt
10:36 scott_w_ joined #salt
10:37 viq MTecknology: actually what I have now works on debian and not on rhel, I hope to clean it up a bit and push to github today
10:37 sroegner joined #salt
11:03 bhosmer joined #salt
11:41 scalability-junk joined #salt
12:06 scott_w_ joined #salt
12:18 arnoldB salt (2014.1rc3) wants to chmod (file.managed) a file on every state.highstate run altough mode is already set correctly. known bug?
12:20 n8n joined #salt
12:21 m joined #salt
12:23 bhosmer joined #salt
12:25 scott_w_ joined #salt
12:26 sroegner joined #salt
12:51 bhosmer_ joined #salt
12:51 scott_w_ joined #salt
12:54 taion809 joined #salt
12:54 Mua joined #salt
12:55 mrud joined #salt
13:22 justBob joined #salt
13:25 pdayton joined #salt
13:26 ndrei joined #salt
13:37 gnugnu5 joined #salt
13:45 jtang1 joined #salt
13:54 justBob joined #salt
14:01 sroegner joined #salt
14:03 gnugnu5 joined #salt
14:03 Wingfold joined #salt
14:04 mgw1 joined #salt
14:09 mgw1 joined #salt
14:19 quickdry21 joined #salt
14:20 quickdry21_ joined #salt
14:25 linjan__ joined #salt
14:27 psyl0n joined #salt
14:32 zzzirk joined #salt
14:39 bhosmer_ joined #salt
14:42 mirko joined #salt
14:42 favadi joined #salt
14:48 nineteeneightd joined #salt
14:50 anuvrat joined #salt
14:59 Johnyang joined #salt
15:00 Johnyang hello,do you meet the problem : exec the highstate but not any response return, run again, all return ok
15:02 mrud left #salt
15:05 nebuchadnezzar Johnyang: I think it's explained in the FAQ: http://docs.saltstack.com/faq.html#when-i-run-test-ping-why-don-t-the-minions-that-aren-t-responding-return-anything-returning-false-would-be-helpful
15:13 ndrei joined #salt
15:14 pdayton joined #salt
15:21 rostam joined #salt
15:27 justBob joined #salt
15:30 analogbyte joined #salt
15:48 mwmnj joined #salt
15:52 jankowiak joined #salt
15:56 jankowiak joined #salt
15:58 Linz joined #salt
16:01 gadams999 joined #salt
16:20 bhosmer joined #salt
16:24 Mua joined #salt
16:28 bhosmer joined #salt
16:43 bhosmer joined #salt
16:49 jtang1 joined #salt
16:52 cachedout joined #salt
16:56 johtso joined #salt
16:59 mlen joined #salt
16:59 jtang2 joined #salt
17:02 longdays_ joined #salt
17:02 JesseC-Work joined #salt
17:02 aurigus_ joined #salt
17:03 pdayton joined #salt
17:03 philipsd7 joined #salt
17:04 mirko_ joined #salt
17:04 marcinkuzminski joined #salt
17:04 Corey_ joined #salt
17:04 babilen_ joined #salt
17:05 Guest36654 joined #salt
17:05 oz_akan_ joined #salt
17:07 vipuls joined #salt
17:07 twinshadow_ joined #salt
17:10 insatsu joined #salt
17:10 Vye joined #salt
17:13 ndrei joined #salt
17:16 jnials joined #salt
17:16 fxhp joined #salt
17:17 kermit joined #salt
17:21 ndrei joined #salt
17:21 terminalmage joined #salt
17:21 Tekni joined #salt
17:21 johtso joined #salt
17:31 psyl0n joined #salt
17:32 oz_akan joined #salt
17:38 krow joined #salt
17:39 mwmnj joined #salt
17:41 Linz joined #salt
17:41 xmj left #salt
17:43 babilen joined #salt
17:47 pdayton joined #salt
17:48 res0nat0r joined #salt
17:52 faldridge joined #salt
17:54 77CABHZI3 joined #salt
17:54 djaykay joined #salt
18:16 bhosmer joined #salt
18:18 terminalmage joined #salt
18:21 terminalmage joined #salt
18:22 n8n joined #salt
18:24 seblu joined #salt
18:25 seblu helo
18:26 seblu is it possible to use a list from yaml inside a jinga2 {% for loop?
18:26 acabrera joined #salt
18:28 __number5__ joined #salt
18:32 terminalmage seblu: yes
18:33 terminalmage {% for foo in list %}
18:33 terminalmage just like in python
18:33 seblu ok, but, the list can be define earlier in the sls file?
18:33 scoates seblu: no
18:34 seblu make sense
18:34 seblu thx
18:34 terminalmage scoates: yes it can
18:34 terminalmage seblu: no, scoates is wrong
18:34 ndrei joined #salt
18:34 seblu terminalmage: how ?
18:34 terminalmage {% set mylist = ['foo', 'bar', 'baz'] %}
18:34 scoates you can't define something in the yaml and expect jinja in the same file to somehow read the yaml
18:34 terminalmage just can't be from yaml
18:34 scoates yes. that's in the jinja. I was about to suggest that.
18:35 terminalmage seblu: if you want to define it in YAML, make it a pillar variable
18:35 terminalmage {% for foo in pillar['mylist'] %}
18:35 ajw0100 joined #salt
18:36 seblu http://pastebin.com/RhGYUNRH
18:36 seblu something like that is not possible, right?
18:36 terminalmage seblu: correct
18:36 dvogt joined #salt
18:36 terminalmage seblu: have you used pillar before?
18:36 terminalmage http://docs.saltstack.com/topics/pillar/index.html
18:36 seblu nop
18:36 seblu yep, that's the next step :)
18:36 seblu thx
18:36 terminalmage you can add user-defined variables
18:37 terminalmage they are assigned using a separate fileserver root
18:37 terminalmage with their own top.sls
18:37 terminalmage and the SLS files can just be YAML
18:38 terminalmage so you could drop the "plugins" bit there into a pillar SLS file, and then reference it using pillar['plugins'] in an SLS file
18:40 pdayton joined #salt
18:40 seblu I see now. I will need to think my salt hierarchy again.
18:40 seblu :)
18:42 terminalmage yeah. FYI, the reason your solution won't work is because the jinja template is evaluated first, to produce valid YAML that can then be loaded
18:42 seblu yep
18:43 seblu I am a sweet dreamer
18:59 scalability-junk joined #salt
19:01 nextdoorwarren joined #salt
19:02 hoodow joined #salt
19:09 n8n joined #salt
19:12 ndrei joined #salt
19:12 mgw1 joined #salt
19:15 terminalmage joined #salt
19:15 arnoldB https://github.com/bechtoldt/binddns-formula - feedback is welcome :)
19:17 terminalmage joined #salt
19:28 schimmy joined #salt
19:29 krow joined #salt
19:33 quickdry21_ joined #salt
19:45 scooby2 joined #salt
19:48 thayne joined #salt
19:50 elfixit joined #salt
19:51 jeremyBass joined #salt
19:52 jeremyBass left #salt
19:52 jimallman joined #salt
19:57 viq MTecknology: https://github.com/viq/gitlab-formula has now what works for me on debian
20:00 schimmy joined #salt
20:02 ndrei joined #salt
20:04 bhosmer joined #salt
20:05 schimmy joined #salt
20:06 bhosmer_ joined #salt
20:13 viq And pull request submitted ;)
20:13 terminalmage joined #salt
20:20 qasket joined #salt
20:21 schimmy joined #salt
20:23 qasket i'm using salt-ssh and needed to know how to pass variables, here is my command salt-ssh "$1" -r netstat -r | grep default | ping -c 1 `awk '{print $2}'`'   <-- i want $2 to take the value of the grep
20:24 eliasp joined #salt
20:24 qasket can anyone tell me how i can prevent salt-ssh from interpreting $2 on the host, and just pass it to the destination
20:25 oz_akan_ joined #salt
20:25 viq qasket: use quoting
20:27 schimmy joined #salt
20:27 viq qasket: also what you wrote seems either mis-formated or incomplete
20:27 qasket viq: should i encapsulate the whole awk in " " or just $2 in quotes
20:28 qasket viq: run netstat as sudo, grep for default line, awk the gateway, ping the gateway
20:28 manfred you don't want to incapsulate the whole thing in "" because then bash will expand the $2
20:28 viq I would do something like salt-ssh "$1" -r 'netstat -r | grep default | ping -c 1 $(awk \'{print $2}\')'
20:28 schimmy joined #salt
20:28 qasket viq: i run it on the destination server and it works, i may have typed it here wrong
20:29 manfred why is your awk at the end?
20:29 manfred why not awk '/default/ {print $2}' ?
20:29 qasket viq: sudo netstat -r | grep default | ping -c 1 `awk '{print $2}'`
20:30 qasket manfred: let me try it
20:31 manfred what are you trying to get results back for
20:31 viq qasket: you need to escape ' in salt-ssh, also instead of `` you can use $()
20:31 manfred netstat -r | awk '/default/ {print $2}' | xargs ping -c 1
20:32 qasket ah ok
20:32 qasket manfred: that works also
20:32 qasket viq: let me try the escapes
20:32 manfred ip route  | awk '/default/ {print $3}' | xargs ping -c 1
20:33 acabrera joined #salt
20:48 samkottler joined #salt
20:50 Led_Zeppelin joined #salt
20:51 schimmy joined #salt
21:01 n8n joined #salt
21:05 cewood joined #salt
21:11 andrej How do I specify a match in a pillar as a regex?  I want to create a pillar that matches only two similarily named hosts
21:13 jtang1 joined #salt
21:13 dave_den andrej: if you can't to the match using globbing match, you can use a compound matcher using PCRE on the minion IDs. http://docs.saltstack.com/topics/targeting/compound.html
21:18 andrej I suppose I could use a plain glob ... the two hosts share the first 4 characters of their host-name, and no other hosts do.
21:18 andrej I just like the fine-grained control regex give
21:18 andrej thanks for the pointer, dave_den
21:19 dave_den yep, no prob
21:20 andrej Oh ... no, that compound thing is a run-time commandline thing; nice, powerful, but not what I need.  I need to specify matching inside a jinja template that configures any of the hosts on our network
21:21 dave_den andrej: are you talking about assigning pillar data with that matcher?
21:21 viq andrej: you can use PCRE matching in top.sls from which you match pillars
21:21 viq I meant from which you assign pillars
21:21 dave_den andrej: for top files, look at the "That same example expressed in a http://docs.saltstack.com/ref/states/highstate.html#term-top-file looks like the following:" section
21:21 david_a joined #salt
21:22 andrej Thanks dave_den
21:22 dave_den andrej: and if you're looking to use the matcher in a state file, you can use the match module: http://docs.saltstack.com/ref/modules/all/salt.modules.match.html
21:23 harobed joined #salt
21:23 ndrei joined #salt
21:25 andrej I guess I am not expressing myself very well :)
21:25 andrej What I am trying to achieve is to have a jinja.template of a config file for nagios-nrpe
21:25 oz_akan_ joined #salt
21:26 andrej Inside that template I'm trying to implement if statements that determine whether or not a machine needs the given line.
21:26 andrej I need to be able to have a bunch of lines only matched by two hosts with similar names
21:26 andrej all others at this stage are fine with just an os:Ubuntu and os:Centos match
21:28 andrej in other words, if I run a state.highstate I want all machines to do the right thing w/o my using any qualifiers on the command-line
21:28 andrej s/my/me/
21:34 andrej http://pastebin.com/kJmr31Vq
21:35 andrej Where it says grains['fqdn'] I'd like to be able to target two machines
21:36 ajw0100 joined #salt
21:37 dave_den andrej: if you want to do that in your state file, use the match module
21:38 dave_den {%- if salt['match.pcre']('your_regex_here') -%}
21:41 andrej Oh, cool. I didn't realist that was possible
21:42 ckao joined #salt
21:42 forrest joined #salt
21:44 Networkn3rd joined #salt
21:45 yomilk joined #salt
21:50 ndrei joined #salt
21:52 bhosmer joined #salt
21:57 krow joined #salt
22:00 rawzone joined #salt
22:01 fragamus joined #salt
22:02 jtang1 joined #salt
22:04 ndrei joined #salt
22:09 elfixit joined #salt
22:16 mgw1 joined #salt
22:26 oz_akan_ joined #salt
22:28 jtang1 joined #salt
22:36 Trauma_Reaper joined #salt
22:36 joehoyle joined #salt
22:38 scott_w_ joined #salt
22:38 jeremyfelt joined #salt
22:45 MTecknology viq: You're mean!!!! I have it working almost 100% except that when I connect to nginx it'll hang for 30sec and then return a 502
22:45 acabrera joined #salt
22:46 MTecknology I may just blow the box away completely and start from scratch and not try to adapt this formula to my environment
22:50 MTecknology viq: It looks you have gitlab:use_rvm used to decide whether or not rvm will be installed, that's not in the example pillar, and if true, you have nothing for debian in there
22:51 MTecknology also- it doesn't handle if you want to use a remote database, only if you want pgsql installed locally; that much should be an easy adjustment, though
22:59 mgw1 joined #salt
23:04 ndrei joined #salt
23:05 jacksontj joined #salt
23:07 rawtaz joined #salt
23:08 quickdry21_ joined #salt
23:09 quickdry21_ joined #salt
23:09 thayne joined #salt
23:13 res0_ joined #salt
23:15 valgrind joined #salt
23:18 frasergraham joined #salt
23:19 jtang1 joined #salt
23:27 oz_akan_ joined #salt
23:31 n8n joined #salt
23:31 scott_w_ joined #salt
23:35 scott_w__ joined #salt
23:37 diegows joined #salt
23:41 bhosmer joined #salt
23:42 bhosmer_ joined #salt
23:50 acabrera I want to setup DNS lookup of salt to point to my salt master, is there any documentation on how to do that?
23:58 alice|wl acabrera: put it in /etc/hosts

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