Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2015-04-09

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

All times shown according to UTC.

Time Nick Message
00:01 coval3nce Is there a best practice for setting up salt environments to use git based branches/tags for setting up an entire “pillar environment” including formulas similar to the concept of R10k in puppet land?
00:02 coval3nce e.g - I define which formulas and versions of them to pull in to setup my master states?
00:03 honestly joined #salt
00:03 StDiluted joined #salt
00:03 coval3nce or is it simply just git_remotes in the master config?  Seems like  that may actually just cover that use case.
00:04 Andre-B joined #salt
00:08 GabLeRoux joined #salt
00:10 coval3nce left #salt
00:20 StDiluted joined #salt
00:26 amcorreia joined #salt
00:31 subsignal joined #salt
00:32 ipmb joined #salt
00:36 hasues joined #salt
00:36 krobin joined #salt
00:37 hasues left #salt
00:38 jerematic joined #salt
00:45 pipeep joined #salt
00:48 baweaver joined #salt
00:51 markm joined #salt
00:55 subsignal joined #salt
00:55 ITChap joined #salt
00:57 zwi joined #salt
00:59 subsignal joined #salt
01:04 MatthewsFace joined #salt
01:07 sunkist joined #salt
01:08 glyf joined #salt
01:11 pipeep joined #salt
01:13 kaiqi-admin joined #salt
01:14 kaiqi-admin who is online?
01:15 catpig joined #salt
01:16 llua the internet
01:19 bhosmer_ joined #salt
01:27 donmichelangelo joined #salt
01:28 smcquay joined #salt
01:48 ilbot3 joined #salt
01:48 Topic for #salt is now Welcome to #salt | 2014.7.2 is the latest | Please use https://gist.github.com for code, don't paste directly into the channel | 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/
01:49 otter768 joined #salt
01:49 scbunn joined #salt
01:49 kevc is there a way to get virtualenv with salt to use requirements.txt in an ordered fashion?
01:49 kevc stupid virtualenv
01:51 randomuser joined #salt
01:58 amcorreia joined #salt
02:02 desposo joined #salt
02:13 smcquay joined #salt
02:17 mapu joined #salt
02:18 evle joined #salt
02:22 scbunn joined #salt
02:32 iromli joined #salt
02:34 HexOffender joined #salt
02:34 michelangelo joined #salt
02:44 ajw0100 joined #salt
02:49 writtenoff joined #salt
03:02 Furao joined #salt
03:05 ipmb joined #salt
03:06 beneggett joined #salt
03:09 zwi joined #salt
03:11 rhodgin_ joined #salt
03:11 jonlangemak joined #salt
03:12 jonlangemak Im newer to salt and Jinja.  Is it possible to access a grain inside of a jinja pillar.get?
03:12 seev no, there are separate callers for grains
03:12 jonlangemak Ideally, I’d like to dynamically pull data from a pillar based on the hostname of the device.
03:13 jonlangemak So its not possible then?
03:13 seev salt['grains.get']('key', 'defaultvalue')
03:14 jonlangemak I’d like to have a pillar that had a server defined as well as a series of attributes with it.  Then I’d like to template a config file based on the name of the device
03:14 Furao joined #salt
03:14 seev grains exist locally on the minion
03:14 jonlangemak Does that make sense?  So lookup the hostname through a grain, and then pull data out of apillar based on that
03:14 jonlangemak yep
03:14 seev pillar exists on the master
03:14 jonlangemak Im pushing the pillar to the minions
03:17 seev example?
03:18 seev you can do something in Jinja like: {% set the_node_type = salt['grains.get']('node_type', '') %}
03:18 seev and then call that variable elsewhere in the template
03:18 seev not sure if using that in a pillar call would work though
03:18 seev you would have to test that
03:19 iggy {{ salt['pillar.get']('hosts:' ~ grains['nodename']) }}
03:19 scbunn joined #salt
03:19 favadi joined #salt
03:19 jonlangemak Yeah, I was trying something like this….
03:19 iggy you can actually see some examples in some of the formulas
03:20 jonlangemak {{ salt['pillar.get']('kube_minions:grains[‘host’]:docker0_bip') }}
03:20 iggy nope
03:21 iggy that's like putting function calls inside a string literal
03:21 jonlangemak ok
03:21 jonlangemak darn
03:22 jonlangemak worth a shot
03:22 iggy ('kube_minions:' ~ grains['nodename'] ~ ':docker0_bip')
03:22 iggy use string concatenation
03:23 jonlangemak Ah, awesome.  I’ve been trying to figure out how to do that
03:24 jonlangemak You’re officially my hero for tonight
03:24 jonlangemak Thank you!
03:29 aCodinMan joined #salt
03:31 malinoff joined #salt
03:39 jonlangemak joined #salt
03:45 favadi joined #salt
03:49 otter768 joined #salt
03:52 ndrei joined #salt
03:55 dude051 joined #salt
03:55 autonomy joined #salt
03:56 dude051 joined #salt
03:56 autonomy heyas
04:01 MatthewsFace joined #salt
04:11 dude051 joined #salt
04:12 subsignal joined #salt
04:13 emid joined #salt
04:16 iggy joined #salt
04:24 ajolo joined #salt
04:25 mosen joined #salt
04:26 SeeDickCode joined #salt
04:27 emid joined #salt
04:32 rdas joined #salt
04:34 jerematic joined #salt
04:37 tkharju joined #salt
04:40 I3olle joined #salt
04:41 ChesFTC left #salt
04:52 theologian joined #salt
04:55 clintberry joined #salt
05:01 favadi left #salt
05:03 Zachary_DuBois joined #salt
05:08 kormoc joined #salt
05:09 kormoc joined #salt
05:13 kormoc joined #salt
05:15 kormoc joined #salt
05:19 Guest31811 joined #salt
05:20 Guest31811 left #salt
05:21 bhosmer_ joined #salt
05:31 favadi joined #salt
05:33 kormoc joined #salt
05:33 kormoc joined #salt
05:35 kormoc joined #salt
05:49 zer0def joined #salt
05:49 iwishiwerearobot joined #salt
05:49 Norbell_ joined #salt
05:50 otter768 joined #salt
05:57 colttt joined #salt
06:07 rsanchez joined #salt
06:10 colttt joined #salt
06:12 MatthewsFace joined #salt
06:12 autonomy hey guys, anybody have any good guides/tutes for running salt-cloud? having a few minor niggly issues with minion_id's..
06:12 srini_ joined #salt
06:13 srini_ hi have a question on state.highstate
06:13 srini_ If i run state.highstate from master to minion
06:14 srini_ how can i see from minion whether the highstate is running or not
06:14 srini_ can you please help me on this
06:19 freelock joined #salt
06:25 harkx joined #salt
06:26 pravka joined #salt
06:27 kormoc joined #salt
06:27 JayFK joined #salt
06:29 kormoc joined #salt
06:30 kormoc joined #salt
06:30 kormoc joined #salt
06:31 kormoc joined #salt
06:33 AndreasLutro joined #salt
06:33 KermitTheFragger joined #salt
06:34 MatthewsFace joined #salt
06:35 kormoc joined #salt
06:37 flyboy joined #salt
06:39 kormoc joined #salt
06:43 malinoff joined #salt
06:45 malinoff joined #salt
06:46 kormoc joined #salt
06:46 dopesong_ joined #salt
06:47 itru joined #salt
06:49 kormoc joined #salt
06:51 kormoc joined #salt
06:52 kormoc joined #salt
06:52 kormoc joined #salt
06:53 hebz0rl joined #salt
06:53 kawa2014 joined #salt
06:53 kormoc joined #salt
06:57 nkuttler srini_: check the log?
06:57 nkuttler autonomy: what seems to be the problem?
06:57 Furao joined #salt
06:58 Romlok joined #salt
06:59 srini_ no in the log
06:59 srini_ only highstate completed the log gets updates
07:00 Furao_ joined #salt
07:01 srini_ As on the minion i want to run an script that should not collide with highstate running from salt master
07:01 srini_ So, i have to wait till the high state compltes to run the script
07:02 srini_ for this i need to know whether highstate is running or not
07:03 dRiN joined #salt
07:04 nkuttler hm. couldn't you run the script from salt? other than that there's http://docs.saltstack.com/en/latest/topics/jobs/
07:04 eseyman joined #salt
07:05 seev check the output from ps
07:05 nkuttler minion process != job running
07:06 autonomy nkutler: deploying to aws using salt-cloud, instance gets bootstrapped fine but the minion_id is not being set to the hostname i specify
07:07 autonomy nkuttler: also not setting correct dns for master. defaulint to 'salt'
07:07 srini_ No i am running highstate from master
07:07 autonomy defaulting*
07:07 srini_ So, in minion i do not see any processes related to this
07:08 srini_ The jobs functions can call only from master
07:08 nkuttler autonomy: you specify the id on the command line?
07:08 srini_ But my requirement is i will run the script only if highstate is not running
07:08 nkuttler autonomy: are you maybe overwriting the minion config files in your highstate? or maybe setting it from your cloud profiles, though not sure if you can do that
07:08 iwishiwerearobot joined #salt
07:09 srini_ so, how can i get the check from minion only
07:09 srini_ Is there any setting or piece of python code
07:09 srini_ ?
07:10 autonomy nkuttler: no real highstate setup. basic. and yeh im running salt-cloud -p minion1.server.com <my-profile>.
07:10 autonomy profile and provider config are v.basic
07:13 autonomy correction, im running salt-cloud -p <myprofile> <server.hostname>
07:14 pravka joined #salt
07:15 fxhp joined #salt
07:16 TinuvaMac joined #salt
07:18 hojgaard joined #salt
07:27 Grokzen joined #salt
07:30 Andre-B joined #salt
07:31 dopesong joined #salt
07:31 dopesong joined #salt
07:32 dopesong_ joined #salt
07:35 MatthewsFace joined #salt
07:39 bluenemo joined #salt
07:41 sinh joined #salt
07:51 otter768 joined #salt
07:53 rdas joined #salt
07:57 lb1a joined #salt
07:59 vstoniest joined #salt
08:00 denys joined #salt
08:03 IOMonster joined #salt
08:10 dkrae joined #salt
08:11 faust joined #salt
08:12 pipeep joined #salt
08:15 djn joined #salt
08:16 ktosiek joined #salt
08:16 slav0nic joined #salt
08:18 Xevian joined #salt
08:18 glyf joined #salt
08:20 Negher joined #salt
08:21 jhauser joined #salt
08:21 jrluis joined #salt
08:26 ckao joined #salt
08:29 ITChap joined #salt
08:30 gdm85 joined #salt
08:33 N-Mi_ joined #salt
08:33 iwishiwerearobot joined #salt
08:34 slav0nic joined #salt
08:35 JlRd joined #salt
08:37 gdm85 hi there. is it possible to use the hostname (different than minion id) in a jinja template?
08:37 analogbyte gdm85: {{ grains['host'] }}
08:38 gdm85 :o
08:38 analogbyte gdm85: also see salt-call grains.items for a list of all grains :)
08:39 gdm85 analogbyte: thanks. is there a documentation page that also lists the default grains? or this was left out on purpose because it's in high development flux?
08:39 gdm85 I was searching for "hostname" and obviously missed "host"
08:40 analogbyte gdm85: there is none I am aware of... no need to, imo, as grains.items is close at hand most of the times
08:40 gdm85 ok, I understand - yeah self-documenting is always best
08:41 malinoff gdm85, the default list of grains may be very different for different hosts. It mostly depends on virtualization/hardware type, available lsb functions, cloud engine, etc
08:41 gdm85 malinoff: so also on the OS type/features. makes sense.
08:44 ITChap joined #salt
08:55 jhauser joined #salt
08:57 echo joined #salt
08:57 dkrae joined #salt
09:04 rumbala joined #salt
09:05 agend joined #salt
09:09 yawniek im getting a ] {'Errors': {'Error': {'Message': 'Invalid device name /dev/sda', 'Code': 'InvalidBlockDeviceMapping'}}  when trying to set the ebs root vol size for an profile with ec2 ebs.
09:10 yawniek how should this be done?
09:11 douardda I'm trying to play with (and update debian packages) raet, but it seems the current version of raet (0.6.3) is incompatible with salt 2017.7.4
09:11 yawniek i tried this: https://gist.github.com/yannick/4c2427b13390296f5978
09:11 douardda it would be nice to have properly defined version dependencies in these kind of situation
09:12 douardda btw, the way  ioflo deal with .flo mission files is at least weird
09:13 douardda since it does open these files with 'r+' flag for reading them
09:14 toanju joined #salt
09:21 Negher joined #salt
09:22 yawniek no one using block_device_mappings  to set the ebs root vol size?
09:24 MatthewsFace joined #salt
09:25 yawniek ok apparently it needs to be called /dev/xvda for my profile
09:27 mortis_ did anyone find a nice proper way of sending diffs between the two last highstate to a mail-account?
09:28 babilen mortis_: http://docs.saltstack.com/en/latest/ref/returners/all/salt.returners.smtp_return.html ?
09:29 analogbyte mortis_: what do you mean by diffs? else, use the smtp returner babilen linked
09:29 babilen I guess only those states (and their output) that contained changes
09:30 viq joined #salt
09:33 mortis_ i'm aware of the returner, but what we want is an email whenever something has changed
09:33 mortis_ say a new package was installed
09:33 mortis_ so the diff in the output
09:33 mortis_ but only when something happened
09:33 mortis_ not every time highstate runs
09:36 analogbyte mortis_: I don't know for sure, but I think these settings might affect the returners too: http://docs.saltstack.com/en/latest/ref/output/all/salt.output.highstate.html
09:37 mortis_ analogbyte: maybe, but whatever you chose it will still output the summary afaik
09:37 analogbyte so, set verbose to false, and output to changes?
09:37 mortis_ hmm, i'll try that
09:38 analogbyte mortis_: yeah, propably the summary will show, but you could just filter your mail for that... if the report has only 6 lines or something, mark as read :)
09:38 analogbyte the output for every state module executed will be supressed, hopefully
09:39 randomuser joined #salt
09:39 mortis_ will be a lot of spam tho :)
09:39 mortis_ was hoping to only send a mail if there was something in the output
09:39 mortis_ should be a quick hack tho
09:39 mortis_ just make a wrapper or whatever
09:40 analogbyte mortis_: yes, you could implement this option in salt.output.highstate, pullreqs are accepted very fast in my expierience :)
09:40 jalaziz joined #salt
09:41 ITChap joined #salt
09:42 analogbyte mortis_: you'd have to add an option and wrap this line in an if: https://github.com/saltstack/salt/blob/develop/salt/output/highstate.py#L280
09:43 analogbyte wait, not that line ofc, but starting with that line downwards
09:43 mortis_ hmm yeah
09:43 mortis_ thanks for the input, im gonna play a bit
09:44 analogbyte mortis_: np, have fun
09:44 lynxman joined #salt
09:46 favadi joined #salt
09:48 jhauser joined #salt
09:52 otter768 joined #salt
10:00 plindgren joined #salt
10:00 plindgren hello, has any body used mysql as pillar data before?
10:01 plindgren i have trouble troubleshooting it :(
10:02 babilen so trouble. much meta.
10:02 babilen What is the issue?
10:02 plindgren well
10:03 plindgren so i have usernames in a database
10:03 plindgren which i'd like to use as pillar data for a looping state i ahve
10:03 plindgren so i have done some pillaring with ext_pillar
10:03 plindgren which has mysql support
10:03 plindgren but when i run state.highstate i get this
10:03 plindgren endering SLS 'base:users.create' failed: Jinja variable 'list' object has no attribute 'items'
10:04 plindgren it feels like the query does not execute against mysql
10:04 autonomy joined #salt
10:04 plindgren and if i do pillar.get on the pillar i only get the stuff from the pillar text file
10:04 plindgren no mysql data
10:04 plindgren no mysql errors or anything
10:05 plindgren i have been trying to follow this guide
10:05 plindgren http://docs.saltstack.com/en/latest/ref/pillar/all/salt.pillar.mysql.html
10:09 ITChap joined #salt
10:09 plindgren so basically
10:09 plindgren i need to now how to pillar.get the acutal mysql data
10:09 plindgren and not the pillar which holds the mysql queries
10:14 denys joined #salt
10:14 gdm85 what I am doing wrong here? http://pastebin.com/raw.php?i=hrdqrUxd
10:14 bluenemo joined #salt
10:14 AndreasLutro plindgren: you should feel less and know more, it sounds like you're doing .items or .items() on a thing from a pillar that's not a dictionary
10:15 AndreasLutro plindgren: `salt my-minion pillar.items users` if you want to show the pillar data
10:15 faust joined #salt
10:16 plindgren http://pastebin.com/e8DSk0Y8
10:16 plindgren could you take a look at that?
10:16 AndreasLutro oh ok pillar.items doesn't get the actual data
10:16 AndreasLutro hm
10:16 plindgren so i should do salt '*' pillar.tiems users ??
10:16 plindgren yeah
10:17 AndreasLutro if you can install 2015.2 you could try this http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.pillar.html#salt.modules.pillar.ext
10:18 calve joined #salt
10:18 calve hi salt
10:18 AndreasLutro if not I'd just try finding calls to `.items` in your state sls and experiment from there
10:18 AndreasLutro plindgren: personally I think it sounds like you used to have your users stored as a dictionary in pillar sls files, but now you're storing them as a list in mysql, and you haven't updated your state sls to reflect that
10:18 AndreasLutro but that's just a guess
10:19 calve I just executed ``saltutil.refresh_pillar`` on all my minions; they all responded True. Just after that I runned ``manage.status``, and now there are all down !
10:19 plindgren yeah finding a pillar.something to get the actual data would be nice
10:19 AndreasLutro it looks like that is available in 2015.2
10:20 calve was it a know bug ? a misconfiguration ?
10:21 calve now i have to ssh on all minions to restart them manually ; not fun at all
10:21 plindgren but how would you use data with salt.pillar.mysql if you cannot use pillar.items or pillar.get to get the data into a state?
10:21 AndreasLutro plindgren: the pillar data is still available in your states, it's just the modules that don't let you show them from the command line
10:21 AndreasLutro (as far as I can tell)
10:21 plindgren aha
10:22 AndreasLutro I don't use external pillars so I can't test this for you
10:24 brayn joined #salt
10:26 plindgren hmm
10:26 plindgren quick question
10:26 plindgren whats the difference between pillar.get and pillar.items ?
10:28 AndreasLutro I use pillar.get in states, pillar.items on the command line
10:28 AndreasLutro not sure what either would do in the other context tbh
10:29 plindgren alright
10:29 plindgren so in my state i should still use pillar.get
10:29 AndreasLutro yeah, I just mentioned pillar.items as it's a useful way to debug pillars on the command line
10:30 plindgren alright
10:30 plindgren yeah seems to upgrade will be the easiest way to troubleshoot this
10:31 plindgren could you help me understand how i can check the actual data from the external pillar with the new command lines?
10:31 AndreasLutro are you using a master/minion setup?
10:32 plindgren yes
10:32 AndreasLutro then you might not want to for stability reasons
10:32 AndreasLutro but `pip install salt==2015.2.rc2`
10:32 plindgren oh its only available in an rc?
10:33 AndreasLutro yes, it's not a stable release yet
10:33 plindgren aww :(
10:34 alexbst I've got a salt master with a few servers - is there some way I can dynamically put that list of hosts, by hostname, in a file created by a salt stat e?
10:34 alexbst or ip (:)
10:36 babilen alexbst: fill the mine and use that information in a file.managed file
10:37 alexbst "fill the mine" ?
10:38 alexbst (goes to docs)
10:38 babilen http://docs.saltstack.com/en/latest/topics/mine/
10:47 aquassaut joined #salt
10:47 Romlok I love how the salt docs use parameters and values in examples, but never mention why
10:49 AndreasLutro Romlok: example?
10:50 Romlok on the afore-linked mine page
10:50 Romlok {% for server, addrs in salt['mine.get']('roles:web', 'network.ip_addrs', expr_form='grain').items() %}
10:50 Romlok more specifically, "expr_form"
10:51 AndreasLutro well, they're just examples
10:51 AndreasLutro if you know what to look for you know that mine is a salt module, get is a function, so http://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.mine.html#salt.modules.mine.get
10:51 Romlok "if you know what to look for"
10:51 Romlok if you're reading the documentation, that's not a guarantee
10:52 CeBe joined #salt
10:52 Romlok if the parameter isn't relevant to the documentation, it shouldn't be in the example
10:52 Romlok if it *is* relevant, then it should be documented
10:52 AndreasLutro I agree the salt documentation is bad
10:53 ndrei joined #salt
10:53 Romlok I've spent a lot of time with Django documentation, so I think I'm somewhat spoiled
10:53 Romlok or rather, had my expectations raised
10:54 AndreasLutro nah I think it's objectively bad. on the other hand I know the salt team handles pull requests extremely quickly so if you have improvements to the documentation you shouldn't shy away from submitting them
10:55 AndreasLutro and once you understand how modules/states/functions are organized it's easier to search your way to what you're looking for
10:58 bhosmer_ joined #salt
11:01 babilen AndreasLutro: What's your main issue with it?
11:03 AndreasLutro babilen: mainly that it's poorly organized for newcomers. I had an easier time just setting up a basic project from looking at examples and then looking up specific state/module documentation as I went along, than I did following the tutorial
11:15 I3olle joined #salt
11:16 AndreasLutro salt could do well with changing the current tutorial to regular standalone documentation pages and then writing a new, more coherent, shorter tutorial (quickstart?) for people new to salt
11:16 AndreasLutro in my opinion anyway
11:17 dkrae joined #salt
11:20 johnkeates joined #salt
11:29 Twiglet I don't even bother trying to navigate the docs from the docs, just google the things i need/think i need
11:29 Twiglet and hope for the best
11:29 Twiglet for me where the docs do fail is anything other than states/modules
11:30 Twiglet and to some extent (after babilen's reply to my github issue) the fact they default to the dev version
11:34 Romlok which wouldn't be so bad, if the docs were consistent in mentioning when features were added or changed
11:36 Andre-B joined #salt
11:37 johnkeates i just grep the python files. best docs ever. never out-of-date or too little/too few features.
11:37 AndreasLutro me too, but it takes a certain level of knowledge just to do that
11:37 * johnkeates points out that means the docs need some serious spanking
11:38 AndreasLutro like knowing that salt['foo.bar'] calls the module foo's function bar, not the state, and that modules can be dynamically loaded and merged...
11:38 johnkeates but no matter salt's shortcomings, it's still better than puppet/chef/etc
11:38 johnkeates :P
11:39 AndreasLutro I'll take other's words for it! salt is the only one I've used
11:39 ltsampros good work on salt-cloud
11:39 ltsampros pretty amazing
11:39 johnkeates i'm sure you'd hate the other tools, they are the worst.
11:40 AndreasLutro ansible was the only one I really considered other than salt
11:40 johnkeates unless you are a ruby nut, then you might like puppet
11:40 johnkeates ansible is like bash+multi-ssh stapled together
11:40 AndreasLutro I'd be prefectly ok with that... I use salt-ssh and we only run debian/ubuntu :)
11:41 AndreasLutro but jinja + yaml was easier to grok, and salt gives you more freedom in how to organize things
11:42 johnkeates well, salt is nice and salty, just the way i like my popcorn.
11:42 johnkeates (so much for useless information :D )
11:43 plindgren joined #salt
11:44 dopesong joined #salt
11:47 Romlok salt.modules.win_system.set_system_date(newdate)
11:47 Romlok Set the Windows system date. Use <mm-dd-yy> format for the date.
11:47 Romlok mm-dd-yy
11:47 Romlok wat
11:47 Romlok oh, and setting the time is a separate call
11:47 Romlok glorious
11:48 babilen wtf?
11:49 babilen is that a windows default?
11:49 johnkeates windows sucks
11:49 johnkeates it doesn't even work properly with UTC clocks
11:49 Romlok I don't know about Windows, but that's the interface exposed by the salt module
11:49 johnkeates not even with the UTC registry hack
11:49 babilen Romlok: Yeah, that date format is ludicrous, but it might be due to Windows rather than the salt API
11:49 johnkeates windows isn't built on logic, but on personal preconceptions, and then they tried to layer logic on top of that
11:50 johnkeates which gives this sort of weird behaviour
11:50 johnkeates for example, de data and time api's are first built with illogical defaults en assumptions
11:50 AndreasLutro Romlok: that module function just passes through to the "date" command on the windows command line
11:50 johnkeates then they layer other API's and GUI's on top of that
11:50 johnkeates so from the user perspective it looks like it is correct
11:50 bbradley joined #salt
11:50 Romlok but a decision must have been made to not wrap the functionality with something saner, and just pass through the value :(
11:51 giantlock joined #salt
11:51 bbradley hello
11:51 johnkeates but when using the API's or commands, you get the old weird stuff
11:51 johnkeates and salt passes that directly to you
11:51 johnkeates so yes, it could have been wrapped
11:51 johnkeates so you can enter normal dates
11:51 johnkeates or localised dates
11:51 johnkeates but no, that didn't happen, so you're stuck with windows's default
11:52 Romlok but anyway, what I really want to do is ensure, or at least check, that the time on a server is accurate
11:52 Romlok as a state
11:52 johnkeates that's gonna be hard, better use NTP
11:52 AndreasLutro Romlok: win_ntp ?
11:53 Romlok I can start the time-sync service, but that refuses to sync if the time is *too* wrong
11:53 Romlok yeah, that's what I'm doing
11:53 johnkeates then set the time first, and then run NTP
11:53 johnkeates setting the time will set it close enough, NTP can take over and fix the few minutes/seconds it might be off
11:53 otter768 joined #salt
11:53 johnkeates one problem; windows handles time in weird non-standard ways
11:53 johnkeates so you'll need to pull the localisation state first
11:54 johnkeates then translate timezones
11:54 johnkeates and then you can "set" the time
11:54 Romlok nah, all our systems are UTC \o/
11:54 johnkeates windows doesn't have UTC
11:54 johnkeates RTC and windows clock are all localised and there is no 'standard' time
11:54 Romlok it's UTC if we say it's UTC
11:54 johnkeates it's cheese if you say it's cheese
11:55 johnkeates does that mean it's cheese 'o clock?
11:55 johnkeates :p
11:55 Romlok it says the time zone is UTC
11:55 johnkeates okay
11:55 johnkeates in that case you are very lucky to live in the UTC zone
11:55 johnkeates i wish i did
11:55 johnkeates no more funky clock messups
11:56 Romlok this is only for servers, and we migrated everything to UTC a while ago
11:56 johnkeates anyway, in that case, you can just 'push' the server (salt) clock onto the windows machine and NTP afterwards
11:56 Romlok that's what I thought
11:56 Romlok is there a standard module to do that?
11:56 johnkeates not sure, but w32time is what i use to manage time on windows
11:56 Romlok or do I need to futz around with that bassackwards windows-specific module
11:56 johnkeates windows is very weird abut it's date and time
11:57 johnkeates best bet is shell execution
11:57 Romlok which is precisely why salt should provide a standard wrapper -_-
11:57 johnkeates well, what's preventing you from writing that wrapper? :p
11:57 Romlok I know next to diddly about Windows :P
11:57 johnkeates :p
11:57 johnkeates the wrapper is in python as well i think
11:58 johnkeates and it wouldn't wrap an API, it would just do a shell execution
11:58 johnkeates but "hide" it from the user
11:58 johnkeates might as well do the shell exec directly
11:58 AndreasLutro "standard wrapper"?
11:59 AndreasLutro https://github.com/saltstack/salt/blob/develop/salt/modules/win_system.py#L426-L427
11:59 AndreasLutro as you can see it just calls the date command
11:59 Romlok AndreasLutro: that doesn't work on anything but Windows
11:59 johnkeates if there is a wrapper for net time or net server you'd be halfway there
11:59 johnkeates otherwise just run a windows shell command
11:59 AndreasLutro well, blame windows!
11:59 * johnkeates blames windows
11:59 johnkeates also: use a jinja.map
11:59 johnkeates in this case
11:59 AndreasLutro I'm sure that maintaining a module that needs to work on linux and windows is way more work than this
11:59 johnkeates because you can then make windows-exceptions
12:00 johnkeates the problem with windows is that it does everything in a non-compatible non-consistent way
12:00 johnkeates and the ROI for FOSS devs is really low for windows-work
12:00 johnkeates doesn't mean it should be dropped
12:00 johnkeates just means that in this case you're better off doing a switch for POSIX *nix systems vs. Windows using some jinja magic :)
12:02 elfixit joined #salt
12:07 Romlok okay, so the time-setting stuff in win_system is actually a non-standard extension to the usual salt.modules.system interface
12:07 Romlok somehow that is not surprising
12:08 Romlok however, there doesn't appear to be any module for explicit setting of date/time for any other OS
12:09 johnkeates weird
12:09 amcorreia joined #salt
12:09 johnkeates maybe the guys doing salt windows know windows sucks at timekeeping :p
12:10 johnkeates on linux is pretty normal to have NTP setup correctly by default
12:10 Negher left #salt
12:10 johnkeates and it doesn't whine if the clock is off too far
12:11 Romlok bah, all these salt-related names make googling things annoying
12:12 johnkeates did you find a great recipe for making a steak?
12:12 Romlok right now, I'm seeing a lot of results for "current local time in Salt Lake City"
12:12 johnkeates well, that's important too
12:13 johnkeates imagine living in salt lake city and not being sure if time is passing at the correct speed
12:13 johnkeates the answer would be only a google search away
12:13 Romlok right now, my third search result is "Sodom and Gomorrah - Wikipedia, the free encyclopedia"
12:13 phx what's so challanging in a steak that you need a recipe for it? you just have to spank the cow's ass on the ember, and eat it while still hot :)
12:13 johnkeates are you sure it's not a secret hobby?
12:14 johnkeates well, putting rock salt on the grill helps
12:16 robothands salt on the grill? why not just put it direct on the steak?
12:18 Romlok okay, so any tips how I can get the master's current time into a pillar (or wherever)?
12:19 AndreasLutro {{ salt['win_system.get_system_date'] ~ ' ' ~ salt['win_system.get_system_time'] }}
12:20 Romlok the master is/will be Linux
12:21 Romlok so I guess something similar with cmd.run?
12:21 johnkeates yeah
12:21 Grokzen joined #salt
12:21 gdm85 johnkeates: windows keeps time just fine as well. it has the equivalent of ntpd running
12:21 johnkeates that will work
12:21 johnkeates gdm85: well, not if the clock drifts too far or of the BIOS clock is shared with an actual UTC clock when windows isn't in a UTC zone
12:22 Romlok I would've thought there would be something that stayed within Python though, rather than calling out to a shell
12:22 gdm85 johnkeates: if the clock drifts then you'll see the same on Linux.
12:22 gdm85 and you can tell windows to behave with UTC too, it's just the default
12:22 johnkeates gdm85: true, but NTP will fix it on linux, on windows it stops syncing at some point if it's too far out
12:23 johnkeates gdm85: also, UTC hardware clocks are not officially supported and not cross-API available, that's what i meant with microsoft's semi-approved registry hack
12:23 Romlok hmm, IIRC ntpd will only fix it when first started
12:24 gdm85 if it's not supported, I am afraid that will never fix. legacy wins there.
12:24 johnkeates gdm85: you can force the RTC to UTC in Windows, but that doesn't mean all Windows API's wil still work the way it should
12:24 gdm85 but the new APIs as well?
12:24 gdm85 johnkeates: I remember there was some cleanup in last years
12:26 gdm85 on topic: https://www.cl.cam.ac.uk/~mgk25/mswish/ut-rtc.html
12:26 johnkeates gdm85: there has been a lot of cleanup yeah :) mostly in the .net departments (and VC11 i think, not sure)
12:26 gdm85 although, really, it's 14 years old.. :P
12:26 gdm85 johnkeates: yeah, that's what I was referring to. there's hope for future :P
12:26 johnkeates gdm85: but most low-level and referenced API's still crappy and old and behave like it's 1985 :p
12:27 gdm85 if you dig deep now, it *IS* still 1985 in many layers :P
12:28 johnkeates yep :p
12:35 ndrei joined #salt
12:35 tkharju joined #salt
12:43 mortis_ whats a good way of checking when a highstate was last run?
12:43 mortis_ on a minion
12:43 mortis_ or ... i mean ...from the master :)
12:44 paolo hello, i have a Vagrantfile using shell scripts as provisioner, i'm translating the provisioning scropts to salt states. in the meantime i wonder if it'd possible to use salt-cloud to deploy my app to digitalocean using the existing bash scripts. any idea?
12:45 mortis_ paolo: you can use scripts as bootstrap with salt-cloud if thats what you mean
12:45 zwi joined #salt
12:46 paolo i'll try that, thanks
12:46 dendazen joined #salt
12:46 mortis_ np
12:46 paolo so i'd need a salt state which will call existing scripts
12:47 mortis_ no, its a param you can set in the config
12:47 mortis_ in the profile config that is
12:47 mortis_ script: yourscript
12:48 paolo right, ok
12:51 jerematic joined #salt
12:52 mortis_ paolo: i pasted some notes i made earlier here : http://pastie.org/10082424
12:52 mortis_ maybe you find it useful
12:52 evle joined #salt
12:55 seshan joined #salt
12:55 paolo mortis_: useful, i missed the /etc/salt/cloud.deploy.d/ part
12:55 JDiPierro joined #salt
13:00 jonlangemak joined #salt
13:00 seshan joined #salt
13:01 MatthewsFace joined #salt
13:02 johngrasty joined #salt
13:02 cmcmacken joined #salt
13:02 bhosmer joined #salt
13:04 subsignal joined #salt
13:04 racooper joined #salt
13:05 tmh1999 joined #salt
13:09 fredvd joined #salt
13:13 andreyst joined #salt
13:17 sunkist joined #salt
13:18 jdesilet joined #salt
13:23 glyf joined #salt
13:25 mpanetta joined #salt
13:27 babilen joehh1: Do you have an ETA for 2014.7.4 packages?
13:27 Tecnico1931 joined #salt
13:28 Deevolution left #salt
13:28 douardda humm about 2014.7.4, why http://docs.saltstack.com say the latest release is 2014.7.2? (2014.7.4 tarball is available on pypi but not linked from the salt pypi page)
13:29 Deevolution joined #salt
13:30 douardda I've tried to build packages for this later version and make it work with raet, but the comaptible version of reat is not declared somewhere
13:30 babilen douardda: Because it has not been "released" yet. Even though saltstack released that version by tagging it in git and uploading it to the cheeseshop it has not been released at the same time because packages aren't available yet.
13:30 jcristau it's released but not released?
13:31 douardda I really don't get the reelease workflow there
13:33 douardda and about raet? (targeting 2014.7)
13:33 babilen jcristau: Well, it has been released. It just hasn't been announced nor have many pointers/links been changes because saltstack wants to wait for their packagers to make packages available before they do that.
13:33 debian112 joined #salt
13:33 douardda I'm in the process of packaging libnacl and ioflo
13:34 douardda and I'd like to be sur which versions shoulb be packaged
13:34 babilen jcristau: It's this sort of interesting limbo you get when upstream is also involved in the packaging
13:35 rdas joined #salt
13:38 renoirb_ joined #salt
13:38 mapu joined #salt
13:39 AndreasLutro wtf, does anyone import macros in state files? regardless of what I do I end up with a TemplateNotFound error from jinja
13:40 rojem joined #salt
13:40 andreyst joined #salt
13:43 AndreasLutro I can import the macro from a .sls state file - but not from a jinja template? interesting...
13:44 mpanetta joined #salt
13:45 douardda for the record, I confirm I cannot sat a salt-master with raet in a fresh venv
13:46 douardda start
13:47 ndrei joined #salt
13:47 Tyrm joined #salt
13:50 JDiPierro joined #salt
13:50 timoguin joined #salt
13:51 jeremyalons is there a way yet to use pillars with salt-call --local?
13:54 otter768 joined #salt
13:55 zwi joined #salt
13:56 niseak joined #salt
13:56 LtLefse salt-call --local --file-root=/srv/salt --pillar-root=/srv/pillar ...
13:58 andrew_v joined #salt
13:58 LtLefse works for me on 2014.7.1
13:59 seshan joined #salt
14:00 jeremyalons doesn't get honored for me in 2014.7.4
14:01 jeremyalons thanks, i'll try downgrading for testing
14:01 LtLefse that's odd
14:02 jeremyalons i actually wonder if it's a pillar top.sls issue for me
14:03 jeremyalons nope, base: '*'
14:03 jeremyalons strange
14:03 jeremyalons must be something else I have broken, thanks for confirminglt
14:04 ndrei joined #salt
14:04 I3olle_ joined #salt
14:05 LtLefse permissions? just grasping at straws
14:06 jeremyalons brought out sudo /path/to/venv/bin/salt-call, so doubtful
14:07 perfectsine joined #salt
14:09 thehaven joined #salt
14:10 iggy jeremyalons: I may have seen an issue open about pillars and --local recently
14:13 niseak_ joined #salt
14:19 sectionme joined #salt
14:20 andreyst joined #salt
14:23 zmalone joined #salt
14:25 StDiluted joined #salt
14:30 perfectsine_ joined #salt
14:36 dude051 joined #salt
14:38 johnchalekson joined #salt
14:40 redzaku joined #salt
14:44 sunkist joined #salt
14:49 fusionx86 joined #salt
14:50 aparsons joined #salt
14:50 rhodgin joined #salt
14:53 perfectsine joined #salt
14:53 ggoZ joined #salt
14:54 aparsons_ joined #salt
14:55 murrdoc joined #salt
14:56 jalbretsen joined #salt
14:58 _JZ_ joined #salt
15:01 gdm85 is there any guarantee that templates will be applied in the order I set them in top.sls?
15:02 manfred gdm85:  it renders to an ordered dictionary, so it should be
15:03 manfred unless you have requires on later states
15:03 gdm85 manfred: ok, thanks for clearing my doubt
15:03 manfred or if you have watch, watch executes immediately, if you want it to only be added at the end, use listen
15:04 gdm85 I am not yet into the more advanced features :)
15:04 manfred gdm85:  https://www.youtube.com/watch?v=yBF055Ye3kg
15:04 manfred that talk at saltconf actually covers what I think you want
15:04 gdm85 ah, remembers me of await in C#
15:04 clintberry joined #salt
15:05 tdjb_ joined #salt
15:07 davisj joined #salt
15:08 scbunn joined #salt
15:10 hasues joined #salt
15:10 gdm85 "Sequentially ordered states are more repeatable" (from the video)
15:10 hasues left #salt
15:11 gdm85 frankly I am surprised why this is not the default/best practice :)
15:13 murrdoc it wasnt doable
15:13 murrdoc till listen_in was made available
15:14 murrdoc all other requisites will reorder state
15:15 gdm85 I find that implicit behavior a bit user-unfriendly. I mean, when using a specific feature order can change in a completely unexpected way
15:15 gdm85 oh, I see Ryan Lane is also talking about docker
15:16 MatthewsFace joined #salt
15:16 gdm85 actually I am planning to use more salt inside Dockerfile's and less of the Dockerfile-syntax. reason is that with salt I can do secrets sharing, at least.
15:23 beneggett joined #salt
15:25 sandah joined #salt
15:28 desposo joined #salt
15:28 johnchalekson joined #salt
15:29 Ahrotahntee joined #salt
15:30 ccarney_ROCC joined #salt
15:30 rumbala joined #salt
15:30 Ahrotahntee joined #salt
15:31 Aikar does salt cloud work with hivelocity sparknode? like is one of the things it supports some generic protocol that SparkNode uses? or is it all very specific integrations?
15:31 StDiluted_ joined #salt
15:34 semarj joined #salt
15:35 kaptk2 joined #salt
15:35 bersace joined #salt
15:36 bersace Hi, i missed something. Where do you specify the saltenv value ?
15:37 bersace ok, got it !
15:37 bersace -_−
15:37 bersace left #salt
15:37 markc joined #salt
15:39 bbradley howdy
15:39 bbradley is there a problem with salt-ssh and the fileserver?
15:39 mark__ joined #salt
15:42 mark__ I am having serious issues with the periodic disappearance of my custom grains on my minions. Per the documentation I set up a reactor to address the initial startup of the minion daemon and verified the custom grains were usually but not always available between minion daemon restarts. Debug output from the master show the reactor is run every time the minon daemon is restarted.
15:42 mark__ Another issue appears when I manually run hightstate from either the master or the target minion. I first validated the existence of the custom grains before each highstate run. During the run, salt happily executed my top.sls file (custom grain dependent) and performed the expected actions.
15:42 mark__ After each highstate run, all my custom grains disappear and I am forced to manually run stateutil.sync_grains. Since all my states files (and formulas) depend heavily on custom_grains, my entire environment seizes when the custom grains disappear. This problem is prevalent in salt  2014.7.0 and 2014.7.2
15:42 mark__ My master and minions are running RHEL6 and are all running the same salt versions. Any help on the matter would be really appreciated.
15:45 manfred i haven't heard of that happening on RHEL, only on windows minions
15:45 babilen mark__: Where do you define your custom grains?
15:48 timoguin joined #salt
15:48 StDiluted joined #salt
15:49 dalexander joined #salt
15:50 murrdoc in /etc/salt/grains as a key value
15:55 otter768 joined #salt
15:56 mark__ in /srv/salt/environments/base/_grains/custom_grains.py
16:02 mark__ babilen: in /srv/salt/environments/base/_grains/custom_grains.py
16:03 slav0nic joined #salt
16:03 slav0nic joined #salt
16:04 SeeDickCode joined #salt
16:05 desposo joined #salt
16:07 ccarney_ROCC left #salt
16:09 five04tluv joined #salt
16:10 desposo1 joined #salt
16:16 MatthewsFace joined #salt
16:18 dalexander joined #salt
16:19 primechuck joined #salt
16:19 catpig joined #salt
16:19 rm_jorge joined #salt
16:21 writtenoff joined #salt
16:22 Grokzen joined #salt
16:22 evilrob I'm digging into salt-cloud today.  I'm trying to configure it to point to generic openstack.  the doc examples list HP or rackspace variants.  how do these vary from standard openstack?
16:24 JDiPierro joined #salt
16:25 enarciso joined #salt
16:26 hernan604 joined #salt
16:28 raygunsix joined #salt
16:29 jalbretsen1 joined #salt
16:30 murrdoc listen_address: [::]:69
16:30 murrdoc thoughts on why that fails in pillar
16:30 murrdoc and what i should be using
16:31 stanchan joined #salt
16:32 pullphinger joined #salt
16:33 KyleG joined #salt
16:33 KyleG joined #salt
16:33 theologian joined #salt
16:33 babilen listen_address: "[::]:69"
16:33 murrdoc dont work
16:34 babilen g&t on the couch again?
16:34 murrdoc g&t ?
16:34 babilen gin and tonic
16:34 murrdoc ehehe
16:34 murrdoc doesnt work via a load_yaml call
16:34 babilen iow: why doesn't it work?
16:34 murrdoc i aint lying
16:38 glyf joined #salt
16:39 rojem joined #salt
16:41 andreyst1 joined #salt
16:43 giantlock joined #salt
16:45 Brew joined #salt
16:45 toastedpenguin joined #salt
16:45 elfixit joined #salt
16:46 mark__ babilen:  any ideas on what I need to do next?  thanks
16:46 bash124512 joined #salt
16:55 dopesong joined #salt
16:57 evilrob noting on salt-cloud?  anyone?
16:57 evilrob Bueller?
16:59 ajw0100 joined #salt
17:01 rojem joined #salt
17:02 smcquay joined #salt
17:04 FRANK_T joined #salt
17:07 tmh1999 joined #salt
17:07 semarj joined #salt
17:08 manfred evilrob: i do not believe the generic openstack supports api keys, otherwise it should be about the same
17:08 manfred just need to configure your endpoint
17:08 manfred and make sure to tell it what the compute name is
17:12 itru joined #salt
17:15 evilrob manfred: yeah, I'm trying it with plain password auth.  getting a MalformedResponseError
17:16 hasues joined #salt
17:16 hasues left #salt
17:18 evilrob hmmm... "message": "get_version_v2() got an unexpected keyword argument..."  and my internal openstack provider has noted that while the URL says v2.0, they're using 1.0 (or maybe it's 1.1) protocol
17:18 manfred evilrob:  http://docs.saltstack.com/en/latest/topics/cloud/openstack.html
17:18 manfred that should be the generic documentation for the openstack provider
17:19 manfred i do not know if it can work with v1
17:19 manfred i have never tried it
17:21 ajw0100 joined #salt
17:29 semarj joined #salt
17:29 FRANK_T joined #salt
17:32 smcquay joined #salt
17:33 evilrob thanks manfred, that's the doc I'm working from already.  I think it has something to do with a deprecated novaclient warnign I'm getting.  either that or the fact that my provider is running v1 protocol on a v2.0 URL
17:36 iggy Aikar: if it's supported by libcloud it's probably already supported or easy to support
17:36 iggy bbradley: salt-ssh has various limitations depending on which version you are using... best to mention what version you use along with salt-ssh questions
17:37 bbradley iggy: i think i found a bug. already put up an issue.
17:37 bbradley but thank you!
17:37 Aikar iggy, they said they use solusvm and doesn't support salt cloud :/
17:37 bbradley i always check develop branch with salt if something is troubling me :)
17:40 iggy Aikar: might be a good addition for libcloud (then salt-cloud)... open an issue?
17:40 denys joined #salt
17:41 Aikar it sounded like they dont have any form of external provision api
17:45 Andre-B joined #salt
17:46 tomh- joined #salt
17:53 iggy oh... that's pretty bad for a "cloud provider" these days
17:54 rojem joined #salt
17:55 aCodinMan joined #salt
17:55 ndrei joined #salt
17:56 otter768 joined #salt
17:57 sunkist joined #salt
18:07 bhosmer joined #salt
18:10 aCodinMan joined #salt
18:11 jeremyalons dangit, can't pass pillar on the CLI to salt-call
18:11 druonysus joined #salt
18:11 druonysus joined #salt
18:14 bhosmer_ joined #salt
18:16 shaiguitar joined #salt
18:17 tmh1999 joined #salt
18:18 aparsons joined #salt
18:19 baweaver joined #salt
18:24 TyrfingMjolnir joined #salt
18:25 Norbell_ joined #salt
18:26 shaiguitar joined #salt
18:29 thehaven joined #salt
18:30 perfectsine joined #salt
18:31 _JZ__ joined #salt
18:31 shaiguitar joined #salt
18:33 timoguin joined #salt
18:33 adelcast1 joined #salt
18:39 pdayton joined #salt
18:46 ProT-0-TypE joined #salt
18:47 kunit joined #salt
18:47 catpig joined #salt
18:48 techdragon joined #salt
18:48 aberdine left #salt
18:48 semarj joined #salt
18:51 shaiguitar joined #salt
18:52 semarj left #salt
18:55 toanju joined #salt
18:55 rosenfs joined #salt
18:58 renoirb_ joined #salt
19:00 XenophonF joined #salt
19:01 ajw0100 joined #salt
19:09 Brew joined #salt
19:19 FRANK_T joined #salt
19:19 jerematic joined #salt
19:25 niseak joined #salt
19:27 druonysuse joined #salt
19:31 hemphill joined #salt
19:35 dalexander is salt able to do gitfs_remotes over SSH on ports other than 22?
19:36 ajw0100 joined #salt
19:36 baweaver joined #salt
19:40 manfred dalexander:  i think it can, if you configure your /root/.ssh/config
19:43 dalexander I'm attempting to set up a gitfs_remote against a stash server that listens on 7999 for ssh rather than 22 and it seems to be totally ignoring the port in the SSH URL.  When we run it in debug and accept the host key it lists it without the port definition in known_hosts, indicating to us that it's making the connection on 22, not 7999.
19:43 dalexander then I found this: https://github.com/saltstack/salt/issues/21979
19:46 sunkist joined #salt
19:49 XenophonF dalexander: add an entry for your remote to ~root/.ssh/config
19:49 XenophonF assuming you're using GitPython
19:49 iggy dalexander: what backend are you using? what uri are you using?
19:49 florinandrei joined #salt
19:50 XenophonF dalexander: https://bpaste.net/show/518d705f87df
19:50 XenophonF dalexander: if you're using GitPython, it's running git/ssh, so you can override the destination port in the ssh client
19:52 dalexander we're using python_pygit2 per recommendation.   URL looks like: git@stash.ommitted.com:7999/scm/salt/salt.git
19:52 dthorman joined #salt
19:52 dalexander python-pygit2 rather
19:53 aCodinMan joined #salt
19:53 overyander joined #salt
19:53 dalexander XenophonF: it hadn't occurred to me to try that override per host approach, will that work with pygit2?
19:54 XenophonF dalexander: not sure
19:54 XenophonF i know it uses libgit2, but i don't know what libgit2 uses for its ssh client
19:55 iggy libssh
19:55 XenophonF looks like libgit2 uses libssh2
19:56 manfred dalexander: theoretically it should
19:56 manfred because they should just be making calls to the git libraries
19:56 manfred which should know to check .ssh/config for ssh configurations
19:57 XenophonF manfred: pygit2 doesn't call out to ssh
19:57 XenophonF gitpython does
19:57 otter768 joined #salt
19:57 manfred pygit2 calls to libgit2, which should use ssh somewhere
19:57 manfred if you specify an ssh url
19:57 ajw0100 joined #salt
19:57 XenophonF it doesn't use the ssh binary
19:57 XenophonF it uses libssh2
19:58 manfred sure, but why wouldn't libssh2 read the regular ssh library
19:58 manfred that should be expected behavior
19:58 XenophonF libssh2 is its own SSH2 implementation
19:59 manfred XenophonF:  http://api.libssh.org/master/group__libssh__session.html#ga82371e723260c7572ea061edecc2e9f1
19:59 manfred it still parses the .ssh/config
20:02 dalexander sounds like a good place to start, thanks!
20:02 manfred I have no idea if it uses it correctly, but it is probably a pretty good chance that if you use that same ssh/config, it will work
20:02 XenophonF i think libssh and libssh2 are different
20:02 XenophonF but it's worth a shot, as you said
20:02 manfred XenophonF:  i am pretty sure you are right
20:03 manfred but, imo, if the people who are writting libssh2 are smart at all, they should make the configuration and basic usage of libssh2 as close as possible to ease friction of adoption
20:03 XenophonF man, _that's_ confusing
20:03 manfred so idk, /shrug, but i hope that it does work :)
20:03 XenophonF it's like the adafruit.cc-vs-adafruit.org business
20:06 rosenfs joined #salt
20:10 rojem joined #salt
20:13 dopesong joined #salt
20:15 dopesong joined #salt
20:16 niseak joined #salt
20:17 I3olle joined #salt
20:21 cberndt joined #salt
20:22 rojem joined #salt
20:23 XenophonF left #salt
20:24 baweaver joined #salt
20:27 iggy I still can't believe how hard it was to add support for .sls files to github
20:28 techdragon joined #salt
20:29 dalexander joined #salt
20:30 dopesong joined #salt
20:30 rumbala joined #salt
20:42 nickdew joined #salt
20:44 murrdoc joined #salt
20:45 bhosmer joined #salt
20:46 fantnn joined #salt
20:46 fantnn whats the github repo that has common salt scripts for wordpress, drupal et al?
20:46 nickdew saltstack-forumlas is the organization .
20:46 fantnn nickdew ty sir
20:47 nickdew https://github.com/saltstack-formulas
20:47 nickdew fantnn: you're welcome
20:47 iggy it's worthwhile just searching github usually too
20:48 iggy formulas isn't appropriate for everything (and indeed some people never even know about formulas)
20:48 nickdew I usually reference formulas, but do my own thing. I really don't like that much Jinja.
20:53 iggy yeah, I'd personally like to see some of the more complex ones *cough*salt-formula*cough* get some of the complex bits converted to _states
20:55 baweaver joined #salt
20:55 zmalone I hadn't spotted the drupal one before.  Is https://github.com/saltstack-formulas/drupal-formula/blob/master/drupal/files/fix-permissions.sh a common design pattern throughout?
20:55 zmalone It seems odd to be to be shelling out to enforce file permissions instead of doing it with salt itself.
20:55 zmalone I'd love an explanation on that design choice.
20:56 iggy that looks like something someone grabbed from a drupal tarball or something
20:56 iggy yeah, look at the commit message
20:57 tvinson is there a way to get more granular with permissions in the external authentication system than @wheel? i'd like to be able to restrict a user to a particular module under wheel.
20:58 zmalone https://www.drupal.org/node/244924 yeah, still quirky to embed a script that would be better handled in salt
20:58 iggy if I had seen that commit, I would have made that person break that commit up into different commits
20:58 iggy but I don't follow all the formulas
21:03 fusionx86 joined #salt
21:04 bmac2 joined #salt
21:08 bhosmer joined #salt
21:11 TaiSHi Do I need to set up separate states for multiple git.config statements? Or is there a way to set up multiple name/value combinations ?
21:13 murrdoc you could jinja it up
21:13 jonher937 TaiSHi: I did not really follow your question, but you should look at jinja and pillars
21:13 CeBe joined #salt
21:13 murrdoc but its probably more readable with different states
21:13 andreyst1 joined #salt
21:13 TaiSHi Hmm, the thing is I'm trying to do a git.latest but it failts with "tell me who you are, git config --global blah bleh"
21:13 subsigna_ joined #salt
21:14 mpanetta Is there a doc somewhere that describes updating to salt 2015.2 from 2014.7?
21:15 manfred mpanetta:  http://docs.saltstack.com/en/latest/topics/releases/2015.2.0.html
21:15 mpanetta manfred: Thank you very much
21:16 manfred np
21:16 spookah joined #salt
21:16 TaiSHi Just read up about it, got to do multiple states
21:17 baweaver joined #salt
21:17 murrdoc :)
21:19 zmalone Can someone explain releases to me?  2014.7.2 is listed as current stable, but http://docs.saltstack.com/en/latest/topics/releases/2015.2.0.html exists, as well as a 2015.2.0rc2 which is listed as the most recent release candidate.
21:19 iggy no
21:19 iggy we can not
21:19 iggy they don't make any sense
21:19 murrdoc zmalone:  stable means there is a package for it
21:19 iggy they change every release
21:19 murrdoc in the repos
21:20 murrdoc u know deb or so on
21:20 * iggy ends rant
21:20 mpanetta manfred: Hmm that page only seems to discuss the changes in 2015.2 not how to upgrade to it :(
21:21 iggy mpanetta: well, it's not out yet, so...
21:21 mpanetta Oh it isn't?
21:21 TaiSHi ah, git.config will drive me nuts
21:21 murrdoc nope
21:21 mpanetta For some reason I thought it was hah
21:21 murrdoc its rc2
21:21 mpanetta Oops
21:21 iggy but once it is, you just install the packages... then proceed to fix your states
21:22 mpanetta Ah ok
21:22 murrdoc or
21:22 murrdoc salt manage.up
21:22 murrdoc and then step out for coffee
21:22 murrdoc its the ruby way
21:22 mpanetta ooo
21:23 iggy of note (and not really mentioned in the release notes for whatever reason)... cmd.run is basically completely different than what it used to be
21:23 murrdoc oh nice
21:23 murrdoc how
21:23 iggy if you use any kind of shell functionality it's broken
21:24 TaiSHi Is there a way to make this ( https://gist.github.com/TaiSHiNet/b406d50c32d536854084 ) a bit more... OCD friendly ?
21:24 iggy |'s, echo, etc.
21:24 manfred TaiSHi:  yes, one second
21:24 TaiSHi manfred: You're the man...fred
21:24 murrdoc iggy:  please splain
21:25 iggy I use git.latest and don't remember it complaining about user.*
21:25 mpanetta iggy: Um so you can't execute arbitrary shell scripts?  Or you can't do them on the command line?
21:25 manfred TaiSHi:  http://docs.saltstack.com/en/latest/ref/states/highstate.html#names-declaration
21:25 iggy murrdoc: I opened an issue
21:25 murrdoc i cant possibly search your issues
21:25 murrdoc u have one a day
21:25 mpanetta haha
21:25 iggy lol
21:26 mpanetta murrdoc: Open an RSS feed :P
21:26 iggy https://github.com/saltstack/salt/issues/22052
21:26 manfred TaiSHi:  commented
21:28 TaiSHi manfred: I knew about the names declaration, but didn't know I could use 'sub values' with it...
21:29 manfred yeah
21:29 iggy mpanetta: the biggest killer for me was stuff like this: salt '*api*' cmd.run "grep foo /var/log/somelog | tail"
21:29 manfred i added it for 2014.7
21:29 mpanetta iggy: YEah we do stuff like that too. :(
21:29 Diaoul joined #salt
21:29 nickdew joined #salt
21:29 manfred TaiSHi:  so you have to be on atleast 2014.7
21:29 TaiSHi Thanks manfred for pointing me out
21:29 manfred np :)
21:29 TaiSHi Yeah, I am on 2014.7.2 and use 'names' on several other states!
21:29 TaiSHi :D
21:29 murrdoc iggy:  seems backwards
21:30 murrdoc but jfindlay must know something i dont
21:30 murrdoc totes possible
21:30 iggy mpanetta: check that issue I linked
21:30 mpanetta iggy: I saw that, thank you
21:30 iggy murrdoc: no offense to anyone, but from what I've seen (on that and other issues), I don't think so
21:31 coval3nce joined #salt
21:31 murrdoc just saying, could have been a config on the master or a minion
21:31 * iggy hopes that person isn't hiding out in here somewhere
21:31 murrdoc with a disclaimer that, turn this on and shits insecure
21:31 coval3nce anyone around that could help answer some salt diretory structure questions regarding envirnoments with states/pillars?
21:32 murrdoc iggy:  like srsly
21:32 murrdoc dont change a function to break backwards compatibilty
21:32 iggy murrdoc: yeah, there was more discussion somewhere else that I can'd find right now where it was mentioned that it should have been a config option or had a deprecation period, etc...
21:32 coval3nce left #salt
21:33 murrdoc weak sauce
21:33 coval3nce joined #salt
21:33 timoguin iggy: dang...
21:33 timoguin That's definitely something that should be way more upfront.
21:33 iggy some https://github.com/saltstack/salt/pull/19914
21:34 iggy https://github.com/saltstack/salt/issues/21698
21:34 murrdoc at least iwasnt sexy base pi
21:35 iggy yeah, I don't think he can take any more tarring and feathering
21:36 itru joined #salt
21:37 itru joined #salt
21:37 Jarus joined #salt
21:39 Diaoul joined #salt
21:41 hemphill joined #salt
21:41 iggy coval3nce: sorry, we were busy pissing and moaning over things that the salt devs obviously don't think are important... go ahead with your questions
21:42 coval3nce ha no worries
21:43 itru joined #salt
21:43 coval3nce Was hoping for some RTFM links on how multiple stages/environments can be setup for both file_roots and pillar_roots.  I thought I could have just one “base” file_root in order to re-use stats accross different “pillar” environments, but maybe there must be a 1:1 mapping of file_roots to pillar_roots environments?
21:43 dstufft left #salt
21:43 bhosmer_ joined #salt
21:43 babilen Does any of you guys know what the P{1,2,3,4} labels are for on GH?
21:44 coval3nce Basically, looking for the best practices directory and top layout for using one set of state/formulas that are completely driveen by different environment pillar data.
21:44 murrdoc p1 is AWWW SHIT
21:44 murrdoc p2 is aww shit
21:44 murrdoc p4 is lol we aint fixing
21:44 murrdoc p3 ?
21:44 babilen ack
21:44 coval3nce profit? hehe
21:44 babilen Is that documented somewhere? I suspected as much, but couldn't really find it
21:45 itru left #salt
21:50 five04tluv is there a config option for a salt minion to bind to a certain interface if it has multiple
21:50 iggy five04tluv: it doesn't bind() it connect()s
21:51 five04tluv I need a minion to connect to a different interface, ie eth1 vs eth0
21:51 five04tluv not finding an example anywhere
21:52 iggy coval3nce: file_roots and pillar_roots work a little differently... I suggest starting small and building... Salt's environments can get complex quickly (we don't even use them in favor of different masters per env)
21:52 iggy five04tluv: change your route?
21:53 coval3nce iggy: yeah think i’ll gut out the concept of pillars from my  environment all together until I figure out how the file_root is working.  I’m coming from role/profiles puppet design pattern and trying to do the same in salt.
21:54 coval3nce was hoping i can ultimately end up with data driven states - all capabilites seem to be there from docs, just figuring out how to bring them together.
21:54 iggy coval3nce: yeah, your best bet is just giving it a shot and see how things work and asking questions
21:55 notnotpeter Afternoon all.  Since upgrading to 2014.7 I've noticed that output from salt commands is sometimes "chunky."
21:55 murrdoc its what we all do
21:55 iggy there's not much in the way of "general guidance" in that area
21:55 iggy (and in salt in general)
21:55 notnotpeter Both local salt-call and output on the master display as chunks, like 4K characters, 0.5 pause, another 4K. It'll even happen mid state output.
21:55 coval3nce yeah, hence showing up in this room ;)
21:55 notnotpeter Has anyone else seen this and is there a way to disable this buffered output?
21:56 dalexand_ joined #salt
21:57 iggy I still don't see how you could get shell injection vuln's... do you just let anybody run commands on your infrastructure?
21:58 timoguin babilen: priority levels with 1 being the highest
21:58 otter768 joined #salt
22:00 iggy oh, I didn't notice someone added tags to 22052 yesterday... good
22:01 baweaver joined #salt
22:03 zmalone I'm working out a workflow for integrating public salt formulas (and the dependencies for them), internal stuff, and testing together for my environment.  Does anything like Berkshelf exist for salt?  I found proposed features for this, but I have not yet found anything already existing.
22:04 iggy kitchen something or other
22:05 iggy https://github.com/simonmcc/kitchen-salt ?
22:06 eliasp zmalone: I outlined here how integrating public formulas is IMHO handled best: https://github.com/saltstack/salt/issues/21413#issue-60262701 - but unfortunately, it doesn't work currently due to GitFS bugs ;-/
22:06 eliasp zmalone: for the testing stuff: environments + what iggy says
22:07 simonmcc zmalone: we're using gitshelf for Berksfile style collecting of formula etc
22:07 iggy I personally don't have a problem with restarting the master to just shove the formulas in the master config
22:07 iggy but that's just me
22:07 eliasp iggy: this implies, that those who are managing the SLS/Formulas have access to the master…
22:07 simonmcc It doesn't support dependency resolution, but then we don't have any metadata about dependencies yet anyway
22:08 aparsons joined #salt
22:08 iggy eliasp: or that your master is managed by salt-formula and they have access to the git ext_pillar where that config is...
22:08 eliasp iggy: well, yeah… true ;)
22:09 zmalone sounds like a decent distance left to go, but it's a start.  Thanks!
22:10 simonmcc zmalone: https://github.com/gitshelf
22:10 iggy yeah, keep in mind, puppet has what... a decade on Salt?
22:10 iggy we gettin there
22:10 zmalone chef in this case.
22:10 zmalone The test driven development approach in chef cookbooks over the last few years is cool too, although it is overhead.
22:10 iggy still... couple years difference
22:10 simonmcc Warning, we've ended up pretty much re-inventing AOSP's repo tool when building gitshelf
22:11 simonmcc That might be a better starting point for managing what is essentially a set of git clones
22:12 simonmcc TDD with kitchen-salt is pretty neat, but I am biased.
22:12 simonmcc And also not using it on a daily basis at the moment either due to $work changing direction
22:13 ajw0100 joined #salt
22:13 theologian joined #salt
22:14 murrdoc kitchen salt is good stuff
22:14 murrdoc u should come work with us simoncc
22:14 murrdoc and figure out how to make tdd a thing for salt
22:14 murrdoc :D
22:17 coval3nce tops.sls is merged in some way between different environments right?
22:18 ajw0100 joined #salt
22:19 zmalone simoncc: https://github.com/simonmcc/kitchen-salt/blob/master/INTRODUCTION.md is really helpful, how do you flatten out, merge, or symlink your /var/salt directory into something sane?  Chef has the cookbooks directory that lends itself to that kind of thing.
22:20 zmalone no dep resolution again though, I suppose.
22:20 iggy coval3nce: it's best not to (I think a common way of handling top.sls is putting in a repo all on it's own)
22:21 coval3nce http://docs.saltstack.com/en/latest/ref/states/top.html#how-top-files-are-compiled
22:21 coval3nce just found that….
22:22 simonmcc zmalone: we don't, there are a few different approaches, you can specify dependencies in your .kitchen.yml, which means we'll copy them down to the guest
22:22 simonmcc But rely on you making them available
22:22 aparsons joined #salt
22:23 itru joined #salt
22:24 simonmcc Or have an integration repo which is just a top.sls, kitchen.yml & gitshelf.yml
22:25 coval3nce iggy: guess i can literally just throw a grain down on a box with “environment” and map that to different top.sls files in each of my environment directories
22:26 zmalone yeah, that sounds a little hard to test single formulas, but it's a lot easier then doing EVERYTHING by scratch.
22:26 iggy coval3nce: there's an environment setting in the minion config
22:27 coval3nce ah nice, environment: stage in /etc/salt/grains worked too
22:28 simonmcc zmalone: single formula are easy to test, ping me via GitHub if something is unclear, I need to step away.
22:28 zmalone No problem, thanks.
22:30 baweaver joined #salt
22:33 shoemonkey joined #salt
22:36 scoates joined #salt
22:36 Dev0n joined #salt
22:38 litwol joined #salt
22:38 litwol Hello
22:38 litwol After almost a week "battling" salt (that's what i call learning :-p) i'm finally feeling the joy of it
22:38 litwol that aside
22:39 litwol I am backing up mysql using salt. wonderful stuff.
22:39 litwol incredibly simple too.
22:39 litwol however, i am stuck on the final step.
22:39 shoemonkey joined #salt
22:39 iggy I really need to document my postgres backup process
22:39 litwol i would like to cp.push from minion to master if my cmd.run ran successfully.
22:39 litwol http://dpaste.com/3Z2DZWB
22:40 litwol Don't shoot me if this is horrible way of doing backups. i'm learning..
22:40 litwol point is that it works great.
22:40 Diaoul joined #salt
22:40 iggy we don't use a dump or anything... we use volume snapshots
22:40 litwol what i'd like to do next is an equivalent to cp.push, but only if mysql_backup_{{dbname}} ran
22:41 litwol i have snapshots on FS level (using zfs).
22:41 iggy require the mysql_backup_{{dbname}} bit and it won't run if the cmd.run returns !0
22:41 zwi joined #salt
22:42 autonomy joined #salt
22:46 litwol This fails http://dpaste.com/1KPJN62
22:46 iggy http://dpaste.com/1XWJC1N
22:47 iggy cp is a module (not a state)
22:48 aquassaut joined #salt
22:50 nickdew left #salt
22:50 nickdew joined #salt
22:51 litwol Module function cp.push ....  is not available
22:51 litwol ... being file path that does exist
22:52 iggy can you cp.push from the command line?
22:53 iggy i.e. salt-call cp.push /etc/hosts
22:53 litwol from master i can: salt 'www*' cp.push [existing file]
22:53 iggy and it works?
22:53 litwol yes
22:54 iggy i.e. the files end up in /var/cache somewhere?
22:54 litwol yes
22:55 iggy oh
22:55 iggy - name: cp.push\n- path: /path/to/somefile
22:58 iggy http://dpaste.com/0J0C66T
23:00 litwol osm
23:02 cwyse joined #salt
23:02 ajw0100 joined #salt
23:02 litwol nowi just need to add /correct/ killswitches to avoid allow transmission only once.
23:03 dalexander joined #salt
23:03 autonomy exit
23:04 baweaver joined #salt
23:05 tdjb_ joined #salt
23:06 litwol i suspect my find command always returns true
23:18 bluenemo_ joined #salt
23:22 rhodgin joined #salt
23:28 Kazzzzz joined #salt
23:32 aphor joined #salt
23:32 __aphor__ joined #salt
23:34 Singularo joined #salt
23:39 sunkist joined #salt
23:40 zwi joined #salt
23:44 nickdew joined #salt
23:48 nitenq joined #salt
23:49 bfoxwell joined #salt
23:49 nickdew left #salt
23:49 nickdew joined #salt
23:51 nitenq Hi there, I am looking to add a parameter to one of my file from my salt config. I mean, I have a file client.json who have a line:  name: "" and I want to be able to put the hostname of my client on it. This file is cp on the machine by typing : salt-call --local state.sls depository.file
23:52 iggy make it a jinja template and put {{ salt['grains.get']('nodename') }} in the quotes
23:53 nickdew left #salt
23:55 _JZ_ joined #salt
23:56 ek6 any suggestions out there on how to form  salt-call --master altmaster cp.push_dir /foo     into a schedule format?
23:58 iggy cmd.run?
23:58 iggy there are perfectly valid uses for cmd.run... this sounds like one
23:58 itru_ joined #salt
23:59 otter768 joined #salt
23:59 ek6 was hoping for something a little less duct tape...but sure why not if there wasnt a cleaner way to get that --master option in there
23:59 baweaver joined #salt

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