Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2016-03-31

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

All times shown according to UTC.

Time Nick Message
00:10 tampakrap joined #salt
00:30 ekristen joined #salt
00:36 rem5 joined #salt
00:37 cliluw joined #salt
00:45 flowstate joined #salt
00:55 ageorgop joined #salt
00:56 akhter joined #salt
01:00 pid1 joined #salt
01:02 pid1 joined #salt
01:02 djgerm joined #salt
01:03 digitalhero joined #salt
01:06 quasiben joined #salt
01:08 virtualguy joined #salt
01:08 xurong joined #salt
01:08 virtualguy does anyone have some pro-tips on how to reduce RAM usage on the salt-minion?
01:13 quasiben joined #salt
01:13 virtualguy im trying to run it on an embedded device and the minon process is using 122MB of ram across two processes, over half of that is swapped out as I only have 128MB of ram for the system
01:15 virtualguy I have just disabled all modules that are not in use though that doesn't appear to have made any difference
01:16 djgerm so… i triggerd an event to restart salt-minion and a lot of my minions just died. and now I have to ssh in and start minion back up. is this a common occurrence?
01:16 mavhq joined #salt
01:19 hemebond djgerm: They died? As in the service itself never started?
01:19 hemebond *didn't start?
01:20 econnell joined #salt
01:21 joe_n joined #salt
01:21 quasiben joined #salt
01:21 djgerm there was a restart but then the proc just seemed to die
01:21 djgerm no log entry
01:21 djgerm just gone
01:22 djgerm and some other minions all the sudden have multiple minion processes
01:22 hemebond Yeah...
01:22 djgerm really odd
01:22 hemebond I think the general process for restarting a minion (usually as part of updating it) is to schedule a restart using "at" or similar.
01:23 hemebond I think there's stuff about it in the docs.
01:23 djgerm heh. some good citizen should make sure the salt formula to salt yourself is updated to fix that….
01:29 djgerm that's one of those things that should be…. really obviously a gotcha somewhere documented right?
01:29 djgerm or am I expecting too much?
01:29 hemebond I think there's some inconsistency in the failures, so difficult to recommend stuff. Lemme try and find the info I saw once before.
01:30 catpigger joined #salt
01:30 hemebond https://docs.saltstack.com/en/latest/faq.html#what-is-the-best-way-to-restart-a-salt-daemon-using-salt
01:32 djgerm thanks!
01:33 hemebond Good luck :-)
01:34 djgerm Ugh I need it.
01:38 akhter joined #salt
01:38 quasiben joined #salt
01:39 digitalhero joined #salt
01:46 xurong joined #salt
01:47 ilbot3 joined #salt
01:47 Topic for #salt is now Welcome to #salt! | Latest Version: 2015.8.8 | SaltConf16: http://saltconf.com/register/ | Paid support available: https://www.saltstack.com/support/ | Logs: http://irclog.perlgeek.de/salt/ | Paste: https://gist.github.com/ (please don't multiline paste into channel) | See also: #salt-devel, #salt-offtopic | Ask with patience as we are volunteers and may not have immediate answers
01:48 nethershaw joined #salt
01:54 futuredale_ joined #salt
01:55 flowstate joined #salt
02:01 cyborg-one joined #salt
02:03 iggy I've never used at and never had problems with minions restarting (aside from self inflicted ones)
02:04 djgerm it's a pretty straight forward state… if I didnt have an emergency that just fell in my lap i'd show ya
02:04 djgerm but basically a watch service restart
02:04 digitalhero joined #salt
02:05 hemebond Is there really no difference between manage.alived, manage.allowed, manage.joined, manage.present and manage.reaped?
02:05 hemebond They all have the same description and they were all added in the same version.
02:06 iggy they probably have subtle differences that you can only find out from reading code (bonus points if you submit doc updates afterward)
02:08 hasues joined #salt
02:08 hasues left #salt
02:08 whytewolf the difference being what state the minions are in
02:09 whytewolf they all just call list_state with a state
02:11 hemebond Ah I see.
02:11 hemebond "Show minions being in specific state that is one of 'available', 'joined', 'allowed', 'alived' or 'reaped'."
02:13 hemebond Where could I find info about these "states" which are different to what Salt usually refers to as "states"?
02:14 hemebond Ah, presence.
02:25 kshlm joined #salt
02:25 fracklen joined #salt
02:25 barajasfab joined #salt
02:28 alexanderilyin joined #salt
02:28 subsignal joined #salt
02:29 kliquori joined #salt
02:35 ronrib joined #salt
02:36 CeBe joined #salt
02:37 norii joined #salt
02:42 joe_n joined #salt
02:44 Muchoz joined #salt
02:46 digitalhero joined #salt
02:48 ageorgop joined #salt
02:49 akhter joined #salt
02:54 fracklen_ joined #salt
02:54 hightekvagabond joined #salt
02:55 flowstate joined #salt
02:58 evle joined #salt
03:01 fracklen joined #salt
03:02 evle left #salt
03:07 flowstate joined #salt
03:15 digitalhero joined #salt
03:25 joe_n joined #salt
03:25 JonGretar_ joined #salt
03:25 akoumjian_ joined #salt
03:25 simonmcc_ joined #salt
03:26 InAnimaTe|whosto joined #salt
03:27 jeblair_ joined #salt
03:27 SteamWells_ joined #salt
03:28 chutz joined #salt
03:28 Ryan_Lane_ joined #salt
03:29 clouddale_ joined #salt
03:29 Patch_ joined #salt
03:29 Freek_ joined #salt
03:29 analogby1 joined #salt
03:29 kaushal_ joined #salt
03:30 arif-ali_ joined #salt
03:30 keekz_ joined #salt
03:30 JamieH_ joined #salt
03:30 czchen_ joined #salt
03:30 erjohnso joined #salt
03:30 FreeSpencer_ joined #salt
03:31 antonw_ joined #salt
03:31 georgemarshall_ joined #salt
03:31 munhitsu__ joined #salt
03:31 carmony_ joined #salt
03:31 renoirb_ joined #salt
03:31 lkannan_ joined #salt
03:31 foundatron_ joined #salt
03:32 gtaylor_ joined #salt
03:32 rodr1c_ joined #salt
03:32 colegatron_ joined #salt
03:32 jankmcjanker1 joined #salt
03:32 Antiarc_ joined #salt
03:33 lorengordon joined #salt
03:33 basepi joined #salt
03:33 dh_ joined #salt
03:33 alexlist joined #salt
03:34 KingJ joined #salt
03:34 paolo joined #salt
03:34 Zhen joined #salt
03:34 guerby joined #salt
03:35 davidbanham joined #salt
03:35 wm-bot4 joined #salt
03:35 peters-tx joined #salt
03:38 mihait joined #salt
03:38 serverascode joined #salt
03:38 Guest5213 joined #salt
03:39 Jimlad joined #salt
03:40 hacks joined #salt
03:44 ramteid joined #salt
03:46 racooper joined #salt
03:51 ronrib joined #salt
03:51 beardedeagle joined #salt
03:59 ageorgop joined #salt
03:59 ninjada joined #salt
04:00 akhter joined #salt
04:12 kshlm joined #salt
04:13 NightMonkey_ joined #salt
04:17 orion XenophonF: Hey, you awake?
04:23 Nazzy joined #salt
04:27 beardedeagle confused on why dawdaw isn't included in the full list of renderers
04:30 kliquori joined #salt
04:32 wendall911 left #salt
04:33 ageorgop joined #salt
04:38 rdas joined #salt
04:41 josuebrunel joined #salt
04:43 beardedeagle not mine, but a good read for any new people: http://bencane.com/2016/03/22/self-managing-servers-with-masterless-saltstack-minions/
04:46 XenophonF hey orion
04:50 iggy there is no such renderer as dawdaw
04:51 orion XenophonF: I had a question about one of your pillars, but I figured it out. :)
04:52 iggy one thing I've never really heard a good explanation for is how do you protect pillar data in masterless?
04:53 i90rr joined #salt
04:53 msn joined #salt
04:53 alexanderilyin joined #salt
04:53 quasiben joined #salt
04:55 XenophonF orion: awesome
04:55 XenophonF good night all
04:55 flowstate joined #salt
04:56 anmolb joined #salt
04:57 msn I am trying to create a state file for php-fpm and have to install one module from outside. Here is my sls file and the corresponding error in the end http://paste.debian.net/422892/
04:59 quasiben when running salt masterless why does runas fail ?
04:59 quasiben For example, salt-call --local cmd.run_all 'ls -al ' runas='admin'
04:59 quasiben assuming admin is on the box
05:00 quasiben Error running 'cmd.run_all': Failed to set supplemental groups to [80, 20]. Error: [Errno 1] Operation not permitted
05:03 darxmurf hi all
05:03 zer0def joined #salt
05:09 sauvin joined #salt
05:10 anmolb joined #salt
05:17 kliquori joined #salt
05:26 beardedeagle how the hell do you get pillars with mako, because this doesn't work (returns default which is nothing: ${salt['pillar.get']('makemaster-secrets:nova-credentials:os_tenant_name', '')}
05:27 impi joined #salt
05:49 slav0nic joined #salt
05:50 beardedeagle @iggy: good point, me either
05:56 flowstate joined #salt
05:56 msn I am trying to create a state file for php-fpm and have to install one module from outside. Here is my sls file and the corresponding error in the end http://paste.debian.net/422892/
05:57 rdas joined #salt
05:59 dh joined #salt
06:01 mavhq joined #salt
06:08 felskrone joined #salt
06:16 iggy msn: requisites take the form `<module>: <ID>`
06:17 iggy so line 35 would be `- pkg: php5-fpm`
06:17 msn ah well i actually toke that from salt documentation and copy pasted it
06:18 msn https://docs.saltstack.com/en/latest/ref/states/all/salt.states.pecl.html
06:18 iggy open an issue and/or PR to fix?
06:18 orion Does anyone know why when I run "salt-call --local config.get file_roots" I *always* get /srv/salt and /srv/spm/salt even though I have overridden this in the master config and restarted the salt-master?
06:19 iggy orion: --local (i.e. masterless) uses the minion config file
06:19 iggy not master config
06:19 orion ah
06:20 iggy so put file_roots in minion config or stop using --local with a master
06:22 orion Is using --local on a master deprecated?
06:22 beardedeagle does anyone know how to do conditional pillars in mako? The state.sls is completing but pillars are just returning None
06:22 iggy it's just antithetical
06:23 orion iggy: I can't use salt to bootstrap the salt-master?
06:23 iggy --local means masterless, which you don't want to do with a master
06:23 iggy you can have a minion on your master
06:23 iggy but you just treat it like any other minion
06:24 orion Do most production deployments have a minion on their master?
06:24 beardedeagle I view it as a smart choice
06:24 iggy I mean... there's not a technical reason you can't use --local on your master, it's just not common
06:24 beardedeagle especially if you are doing syndics
06:25 iggy the much more common method is to just run a minion on the master like everything else
06:25 orion I see.
06:26 msn iggy: but that's the format i am using
06:26 msn iggy: - require and then - pkg: php-pear
06:26 msn line 34/35
06:28 iggy msn: no you aren't
06:28 iggy msn: s/php-pear/php5-fpm/
06:29 mavhq joined #salt
06:29 msn but php-pear is what's needed, pecl install will work even if php5-fpm is not installed, but it will not work without php-pear
06:29 beardedeagle gist of the mako/pillar snafu I am dealing with: https://gist.github.com/beardedeagle/37c2429c52c1274c4260b97236c16706
06:29 msn php-pear provides the command "pecl"
06:30 iggy msn: but php-pear isn't a salt ID, it's a package name
06:31 xurong joined #salt
06:32 msn mmm that what i am checking, if package php-pear is installed
06:32 iggy that's not how requisites work
06:32 msn iggy: thanks i understood what you are saying
06:32 msn sorry my bad
06:33 iggy I _think_ if you used names instead of pkgs that would work
06:33 iggy but that has different semantics than pkgs (namely that it calls the pkg manager for each installation vs installing multiple packages in one go)
06:33 msn yes now i understand pkg target has to be a target defined in salt, not a an actual package
06:35 iggy the examples don't always help in this regard
06:35 iggy they try to make things easy to follow and end up obfuscating a certain amount of detail
06:36 beardedeagle nevermind id10t error, running against minion on master which does not match any test case in the pillar file.
06:37 freeaks joined #salt
06:39 fooma joined #salt
06:43 ageorgop joined #salt
06:46 KermitTheFragger joined #salt
06:47 teatime joined #salt
06:48 lorengordon joined #salt
06:49 ninjada joined #salt
06:50 babilen msn: You define requisites on salt *states* and, specifically, on their ID or name attribute. That means that you cannot, for example, "watch a directory" or "check if a package is installed" if you don't have a corresponding state for that directory or package.
06:50 fracklen joined #salt
06:50 msn babilen: thanks I figured that out. I am learning and implementing at same time so sometimes I overreach myself
06:51 babilen It comes up often :)
06:51 joe_n joined #salt
06:53 elsmo joined #salt
06:53 mavhq joined #salt
06:53 msn I have setup a pillar sls to list any packages missing from base install i want http://paste.debian.net/422898/
06:54 msn trying to read it in a state with http://paste.debian.net/422900/ throws error
06:55 flowstate joined #salt
06:56 babilen msn: pkgs takes a list of packages not a list of lists. The leading - on line 5, 6, 7 and eight is being parsed as "element in a list". The result is therefore pkgs: [{{ salt['pillar.get']('packages:utils') }}, {{ salt['pillar.get']('packages:performance') }}, ...]
06:56 om joined #salt
07:01 babilen msn: fwiw, you could simply iterate over utils, performance, ... and then the packages therein and create individual states for them.
07:02 Dr_Jazz joined #salt
07:03 eMBee joined #salt
07:04 msn k thanks
07:06 babilen msn: You tried http://paste.debian.net/422901/ ?
07:06 msn yes that worked :)
07:06 msn now i moved to adding multiple lists
07:07 msn u suggested that yesterday and it worked already
07:07 babilen Sure, but you would have to merge those lists into a single one that you then pass as "pkgs" argument. My feeling is that, at that point, you might just as well use dynamically constructed individual states.
07:08 babilen Yeah, I suggested that yesterday, but I wasn't sure if you tried it or if it worked.
07:08 msn time to learn something new
07:08 msn may be put in some python
07:08 teatime or loop over outside list, loop over loop arg
07:09 teatime TMTOWTDI
07:09 msn ??
07:09 teatime nm I'm just complicating things
07:10 babilen msn: http://paste.debian.net/422902/ is what I had in mind more or less
07:10 babilen Or the one-state-per-pkg solution I mentioned earlier :)
07:10 msn well i was thinking of finding out how to do that:)
07:10 msn i had a for loop in mind for google
07:11 msn salt for loop
07:11 msn :D
07:11 fracklen joined #salt
07:12 msn so I can dynamycally create new states :) awesome
07:13 fracklen joined #salt
07:13 joe_n joined #salt
07:15 babilen msn: SLS files are, essentially, templates written in jinja that is then being parsed as YAML by default (i.e. the default renderer is "#!jinja|yaml" (or "yaml_jinja" as it is also being called))
07:15 beardedeagle ok maybe I am going crazy so I will ask: you can have mako in your pillars right?
07:16 Dr_Jazz beardedeagle, i believe you can. I have never done it though
07:17 babilen msn: It doesn't really matter *how* you arrive at that point. In fact you can use any combination or pipeline of renderers as long as the end result is suitable. That means that you can write/generate states using all renderers discussed on https://docs.saltstack.com/en/latest/ref/renderers/
07:17 msn babilen: thanks
07:17 msn babilen: I am sure i will be needing them more and more as I make my current system more complex
07:17 babilen msn: It was a *very* unfortunate decision by saltstack to not require the shebang (i.e. "#!jinja|yaml") in SLS files, but to rely on the default.
07:18 beardedeagle error I am getting: https://gist.github.com/beardedeagle/e752d78083db10b8888662d3b014cad1
07:18 kliquori joined #salt
07:18 beardedeagle and the logs show nothing but the same message
07:18 babilen As it both doesn't make it obvious to people what they are writing there (jinja template that is then being parsed as YAML) and makes it hard to switch the default renderer to something else as you'd break a billion of SLS files that do not specify the shebang.
07:20 beardedeagle and added the pillar file
07:20 babilen msn: Jinja isn't very powerful and people often end up writing states in mako or Python right away. stateconf is also nice and I'd strongly encourage you to explore renderer options before you commit to a large jinja|yaml codebase.
07:20 beardedeagle @msn: I am making the change now, very painful
07:20 beardedeagle jinja -> mako
07:21 babilen beardedeagle: Be happy ... we have a large "legacy" jinja codebase and won't get rid of it anytime soon.
07:21 beardedeagle well I can't even get pillars to work
07:21 babilen hehe
07:21 babilen Why not?
07:22 kawa2014 joined #salt
07:22 beardedeagle I'd love to know myself, but the error is utterly useless and logs have nothing but the same error message. Even launched in -l debug and nothing
07:22 beardedeagle Pillar failed to render with the following messages:
07:22 babilen Could you paste what you have and the error? Maybe we can figure it out?
07:22 beardedeagle Rendering SLS 'cppv3' failed. Please see master log for details.
07:22 babilen \o/
07:22 beardedeagle yeah see my gist above
07:22 msn I am probably going to stick to default and use #!py where I want, to make it easier for someone not from python background to still make changes and maintain it
07:23 beardedeagle well...mako is great because I get the ease of a template engine, with the power of python
07:23 babilen msn: That is what we ended up doing. I would have *much* preferred mako as default renderer, but the point where you need additional power you might just as well go "full Python" rather than throwing in an intermediate step.
07:24 favadi joined #salt
07:24 msn babilen: I am setting up salt here today a yaer down the line I might not be around and if the next guy doesnt know python changing/adding things would be PIA
07:25 msn yaml/jinja is slightly easier to understand
07:25 iggy jinja is more comfortable to people who have used Django... Think that explains how I picked it up so quickly
07:25 beardedeagle problem is too, most people who can help you are only going to know jinja
07:25 iggy I'm usually the dumb one in the room
07:25 AndreasLutro jinja is fine for 95% of use cases if you ask me
07:25 beardedeagle mako is an outlier
07:25 beardedeagle as I am finding out
07:25 Dr_Jazz For our needs jinja is fine and I also find it pretty nice and easy to work with though
07:26 honestly jinja is fine unless you care about whitespace
07:26 honestly then it gets atrocious
07:26 AndreasLutro beardedeagle: try setting pillar_safe_render_error to false - remember setting it back afterwards
07:26 msn well that's yaml "2 space" thing but i am getting a hang of it
07:26 Ron11 Can anyone help me with install salt offline (package which contains all the packages)?
07:27 beardedeagle werd, doing so @AndreasLutro
07:30 danderson joined #salt
07:31 kshlm joined #salt
07:33 beardedeagle still gives me nothing
07:35 beardedeagle wondering if it is a mako filter gone wild...
07:37 Kurisutian joined #salt
07:37 akhter joined #salt
07:37 beardedeagle nope not it
07:38 CeBe joined #salt
07:39 Kurisutian Hi guys! Can anyone tell me if there is a way to use salt-ssh logging in as a user, using a key and then on the target system using sudo to switch to another user? I think ansible has something like that with the "become-user" option. I couldn't find this in salt-ssh. Does anythink like that exist?
07:40 beardedeagle su -u ?
07:40 beardedeagle is this in a state?
07:41 beardedeagle @Kurisutian ^
07:43 Kurisutian beardedeagle: I have to connect to the target server as my personal user and on the system I have to switch using "sudo -iu" to be able to edit certain files. I'd like to run this over salt-ssh but I could not find anything on how to do the "sudo -iu" part with salt-ssh. I'd like to run modules and not states with this...
07:44 AndreasLutro Kurisutian: check the docs on the roster file
07:44 beardedeagle well there is https://docs.saltstack.com/en/latest/ref/states/all/salt.states.cmd.html#salt.states.cmd.run
07:44 beardedeagle some of the cmd. methods allow to run as a user
07:45 beardedeagle of if that doesn't work you can cmd.run and something like this maybe? https://gist.github.com/beardedeagle/48eba530dd7a0e662b28bfc2fbd19770
07:45 Kurisutian beardedeagle: Mainly I need to run the file replace module (https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.file.html#salt.modules.file.replace)
07:46 beardedeagle why use that over file.managed?
07:46 beardedeagle we actually had a long talk about this yesterday I think lol
07:46 dariusjs joined #salt
07:47 beardedeagle well you could still go the route of cmd.* and run as a differnt user and just sed
07:47 beardedeagle or awk
07:47 ninjada joined #salt
07:47 babilen Do you guys know if it is possible to "consolidate" (as in pkgrepo.managed) repositories explicitly and not as part of a pkgrepo.managed definition?
07:48 babilen I really don't want that anywhere, but it would be nice if you could trigger it explicitly
07:48 dmaiocchi joined #salt
07:49 Kurisutian beardedeagle: Well, we are running a tool to deploy systems to the server which also creates apache config files. Unfortunately these files need to be modified in a certain way which I intended to do with the replace function as this seemed to be the best way to do this in that case. Unfortunately I'm also just a user on that system with limited access which I need the sudo switch for
07:49 flowstate joined #salt
07:51 beardedeagle hrm...there is the sudo module
07:51 beardedeagle but you would still have to do the search and replace via shell (sed, awk)
07:51 ToeSnacks joined #salt
07:51 AndreasLutro Kurisutian: did you check the salt docs for the roster file?
07:52 LondonAppDev joined #salt
07:53 dmaiocchi joined #salt
07:53 Kurisutian AndreasLutro: I did, but could not find anything in that regards... or at least it didn't look that way...
07:53 beardedeagle ha, sudo in the roster
07:53 AndreasLutro "sudo"?
07:53 beardedeagle https://docs.saltstack.com/en/latest/topics/ssh/roster.html#targets-data
07:53 flowstate joined #salt
07:54 eMBee /sb end
07:54 eMBee ooops
07:56 Kurisutian beardedeagle: I'm not sure this is what I need. This is just to run with sudo. What I need is to switch to another local user and just a command as that user. Simply running sudo with the command as the connecting user and not switching to the local user didn't work for me
07:56 beardedeagle then I refer you back to my previous two options
07:56 mavhq joined #salt
07:56 AndreasLutro if you set sudo: true, that means you can use the user kwarg of the various execution module functions
07:57 flowstat_ joined #salt
07:59 sfxandy joined #salt
08:04 Kurisutian AndreasLutro: So I actually would need to set the user in the module command, did I understand that correctly? If it's supported by the module at least.
08:05 kliquori joined #salt
08:05 AndreasLutro Kurisutian: if you need that module function to run as  the specific user, then yes
08:06 Kurisutian AndreasLutro: Awesome, sounds like this is exactly what I was looking for... :)
08:09 Kurisutian AndreasLutro: Unfortunately the file.replace module does not support kwargs from what I can tell reading the docs :(
08:10 om2 joined #salt
08:10 AndreasLutro that's not true... however, it does not have a user kwarg
08:11 AndreasLutro I'd just run it as root, and assume it doesn't mess with permissions/ownership
08:14 Kurisutian AndreasLutro: Running as root would work even if I am not allowed to be root at the target server? Wouldn't it deny any activities if I run them as root which I am not allowed to become?
08:15 AndreasLutro nevermind...
08:16 AndreasLutro salt is really not designed for non-root usage
08:16 AndreasLutro I assumed when you were talking about sudo that you had rights to sudo to root
08:19 niluje left #salt
08:25 elsmo joined #salt
08:27 felskrone joined #salt
08:28 GreatSnoopy joined #salt
08:28 punkoivan joined #salt
08:29 impi joined #salt
08:30 Kurisutian AndreasLutro: Sorry for not being clear on that. Most of the time I am not allowed to be root since another devision is handling the OS level while I usually start at an application level. So there are plenty of hacks here for granting certain rights without being root... ;)
08:30 dgutu joined #salt
08:36 N-Mi joined #salt
08:36 N-Mi joined #salt
08:44 rdas joined #salt
08:47 lero joined #salt
08:49 s_kunk joined #salt
08:49 damien__ joined #salt
08:54 flowstate joined #salt
08:59 beardedeagle welp I am no longer getting the non specific error
09:00 beardedeagle I am getting the good old 'no mako.exceptions' error
09:06 MadHatter42 joined #salt
09:13 josuebrunel joined #salt
09:13 xurong joined #salt
09:20 mavhq joined #salt
09:21 fracklen joined #salt
09:21 mavhq joined #salt
09:21 fracklen joined #salt
09:24 mavhq joined #salt
09:26 mavhq joined #salt
09:29 beardedeagle and now we start editing salt source
09:31 ninjada joined #salt
09:35 dgutu joined #salt
09:35 eMBee how do i wait for a state to complete before the next one is run
09:35 eMBee ?
09:37 dgutu joined #salt
09:37 catpig joined #salt
09:41 joe_n joined #salt
09:42 M0nte0 zero
09:45 babilen eMBee: That is normal. Which problem are you trying to fix?
09:47 xurong joined #salt
09:49 ronnix joined #salt
09:49 freeaks joined #salt
09:51 joe_n joined #salt
09:52 xurong joined #salt
09:54 flowstate joined #salt
09:56 ronnix joined #salt
09:57 mavhq joined #salt
09:59 beardedeagle @babilen: have you ever ran into import errors with salt?
10:00 babilen Mostly in the context of unavailable execution modules
10:06 kliquori joined #salt
10:07 eMBee babilen: i have one state that seemingly only works if i but a 10 second sleep before it
10:08 babilen eMBee: Could you elaborate? States are being executed in order and one after the other.
10:08 babilen But then, I'm off for lunch now ..
10:08 eMBee babilen: i am off to dinner too. i'll bring the issue up another time
10:09 babilen Enjoy!
10:09 mavhq joined #salt
10:09 favadi joined #salt
10:10 eMBee thanks and bye...
10:11 beardedeagle @babilen: getting this error now https://gist.github.com/beardedeagle/e752d78083db10b8888662d3b014cad1
10:12 beardedeagle mako.exceptions is deff their though, I just imported it through the python repl
10:13 mavhq joined #salt
10:17 capri joined #salt
10:18 capri hey, i just started to learn salt today and my first script uses the pkgrepo.managed
10:18 capri is it possible to add multiple sources at a time within one pkgrepo.manage query?
10:18 AndreasLutro beardedeagle: you sure salt runs the same python version as you? try to run the python repl as root instead
10:19 capri cause i want to add four different sources at one time and afterwards it should do apt-get update
10:19 beardedeagle I was root
10:19 mavhq joined #salt
10:19 beardedeagle 2.7.5 across the board
10:21 beardedeagle i fixed the pillar issue and they are passing to the state now
10:21 beardedeagle but this popped up the moment I got that fixed
10:24 sunbx joined #salt
10:26 garphy joined #salt
10:26 dgutu joined #salt
10:36 mavhq joined #salt
10:39 amcorreia joined #salt
10:39 msn joined #salt
10:39 mavhq joined #salt
10:45 mavhq joined #salt
10:47 PeterO joined #salt
10:50 beardedeagle I give up for now. been at this for 7 hours trying to figure out what is broken.
10:52 kliquori joined #salt
10:53 flowstate joined #salt
11:00 PeterO joined #salt
11:05 sfxandy ok, suspect this might be a "kick myself" moment but am struggling with the correct syntax for an if condition in jinja if a salt['pillar.get']('foo') call actually returns a structure or not
11:06 orion joined #salt
11:06 orion joined #salt
11:06 sfxandy currently have this... https://gist.github.com/phtx3/820325b3bcf9c76d309b21702b63e33f
11:07 sfxandy so i want to conditionally declare a Salt mine function only on the minions that have a particular Pillar structure ..... basically
11:09 Netwizard joined #salt
11:11 AndreasLutro that's correct, though I don't see why you need the != {}
11:14 sfxandy AndreasLutro, ok so that was my first thought.  however it seems like my condition is never actually being met.  although I can clearly see the Pillar structure present on certain minions
11:14 akhter joined #salt
11:15 AndreasLutro well, there's nothing wrong with your conditional statement
11:16 sfxandy i've modified the gist, is that what you meant AndreasLutro?
11:16 mavhq joined #salt
11:18 AndreasLutro sfxandy: yes
11:18 PeterO joined #salt
11:19 AndreasLutro sfxandy: what file is the gist of?
11:20 sfxandy its part of a Pillar Stack config file
11:21 AndreasLutro so you're trying to get pillar data inside a pillar
11:22 sfxandy yeah, using PillarStack, the structure I'm trying to query has already been evaluated and is available
11:22 graffic_ joined #salt
11:22 AndreasLutro mmkay
11:22 Ashald_ joined #salt
11:22 elektrix_ joined #salt
11:22 RandyT_ joined #salt
11:22 atmosx_alt joined #salt
11:22 jab416171 joined #salt
11:23 justanot1eruser joined #salt
11:23 terratom1 joined #salt
11:23 upb joined #salt
11:23 bastiandg joined #salt
11:23 roock joined #salt
11:23 dabb joined #salt
11:23 hax404 joined #salt
11:23 danemacmillan joined #salt
11:25 cro joined #salt
11:25 georgemarshall joined #salt
11:26 quasiben joined #salt
11:27 flowstate joined #salt
11:32 eMBee joined #salt
11:32 fredvd joined #salt
11:36 quasiben joined #salt
11:43 rem5 joined #salt
11:48 zer0def joined #salt
11:48 fracklen joined #salt
11:49 fracklen joined #salt
11:50 rory left #salt
11:58 quasiben joined #salt
12:05 cuonglm joined #salt
12:09 zer0def joined #salt
12:09 LondonAppDev__ joined #salt
12:09 eseyman fetchm
12:10 eseyman sorry, wrong tab
12:13 linjan joined #salt
12:13 flowstate joined #salt
12:13 LondonAppDev joined #salt
12:14 akhter joined #salt
12:20 DammitJim joined #salt
12:21 flowstate joined #salt
12:24 DammitJim guys, I watched  a youtube video of someone using salt with vmware
12:24 DammitJim any pointers on what to read?
12:26 TooLmaN joined #salt
12:27 metalseargolid joined #salt
12:33 dijit the official docs are great.
12:33 dijit try to make something?
12:33 philiaagape joined #salt
12:36 mavhq joined #salt
12:38 evle joined #salt
12:42 DammitJim is it salt cloud?
12:43 DammitJim https://docs.saltstack.com/en/latest/topics/cloud/vmware.html ? just to be clear 'cause there are other discussions about drivers and stuff
12:44 cro joined #salt
12:44 dyasny joined #salt
12:45 NightMonkey joined #salt
12:47 _Cyclone_ joined #salt
12:49 numkem joined #salt
12:54 akhter joined #salt
12:54 kliquori joined #salt
12:55 gtmanfred joined #salt
12:56 akhter joined #salt
12:59 ronnix joined #salt
13:01 subsignal joined #salt
13:02 dyasny joined #salt
13:04 morissette joined #salt
13:05 fracklen joined #salt
13:05 mpanetta joined #salt
13:06 subsignal joined #salt
13:07 kliquori joined #salt
13:09 racooper joined #salt
13:11 FreeSpencer joined #salt
13:11 FreeSpencer joined #salt
13:18 CampusD joined #salt
13:18 flowstate joined #salt
13:23 ekristen joined #salt
13:23 scoates joined #salt
13:24 mavhq joined #salt
13:26 Dr_Jazz Any of you using a salt master in AWS? and what kind of size are you running? If I might ask?
13:31 williamthekid joined #salt
13:32 flowstate I am,
13:32 flowstate running an m4.large
13:32 Dr_Jazz for how many minions?
13:33 Dr_Jazz Think i need to migrate to a bigger box, since it keeps getting hit by the OOM
13:33 flowstate but it's RId, so that's why. I'm at roughly 100
13:33 flowstate and this instance is way too strong
13:34 flowstate but if you're getting oom'd, that size may be the way to go
13:36 Dr_Jazz well clearly a t2.micro is _wayyyyy_ to small ;)
13:37 flowstate hahaha
13:37 flowstate yes.
13:37 flowstate that is probably a true statment
13:38 quix joined #salt
13:40 mavhq joined #salt
13:41 ninjada joined #salt
13:43 ronnix joined #salt
13:47 kbaikov joined #salt
13:54 ravenx joined #salt
13:54 ravenx anyone knwo why i'm getting this error:   State bower.bootstrap found in sls super-app is unavailable
13:55 ravenx for this state, it's nice and short, only have a -name  and -user
13:55 ravenx and i've made sure the dir is spelt correctly
13:55 MindfulMonk joined #salt
13:55 flowstate can you verify that bower is available to the machine in general?
13:56 ravenx yup
13:56 ravenx i've installed it via npm, globally
13:56 ravenx via a state.
13:56 ravenx and when i sudo as the user that i'm supposed to,
13:56 ravenx and type 'bower'
13:56 ravenx i can see a bunch of help commands
13:56 flowstate can you execute any bower commands via the salt cli?
13:57 ravenx when i do this:  salt '*' cmd.run 'cd /home/super-app/widgets-api && bower install'
13:58 ravenx i get this:  http://pastebin.com/K6uMLrgK
13:58 ravenx but when i go into my vagrant box, go to the dir, and as the user that my salt state is supposed to `bower install` as, it works.
13:58 ravenx just not via the state.
13:58 jerredbell joined #salt
13:59 flowstate can you execute a bower module command on the minion via salt-call --local?
14:00 ravenx do i do that from my miniopn machine or master mahine?
14:00 ravenx ah okay got it
14:00 ravenx i get the same error message as the pastebin.
14:01 flowstate then it looks like that's the issue
14:02 ravenx but i am not running it as sudo
14:02 flowstate do you specify a non-root user in your minion config?
14:02 ravenx in my salt state, the user is set to:  - user: app-launcher
14:02 ninjada joined #salt
14:02 ravenx flowstate: lemme see
14:03 ravenx it is in the /etc/salt/minion config?
14:04 flowstate yeah, that's where I believe it should go
14:04 hasues joined #salt
14:04 hasues left #salt
14:04 ravenx oaky i set it to a normal user
14:04 ravenx and all of a sudden, none of my commands go through
14:05 flowstate welp, sounds like it's working as expected
14:05 flowstate kidding. hmmm
14:05 flowstate this is odd
14:05 ravenx it is
14:05 ravenx so i just commented it out again
14:06 flowstate can you try converting the state to cmd.run, and adding an --allow-root to it?
14:08 flowstate sigh. has anyone else run into the 'object_pairs_hook' problem in boto_iam??
14:09 flowstate for god knows what reason, salt wants to use python2.6 on amaxon linux
14:11 numkem joined #salt
14:12 ravenx flowstate: sure i can try that.
14:12 flowstate lmk if it helps
14:13 ravenx yup, it works
14:14 flowstate k, I'd suggest rolling that way and trying to get to the root of your problem in parallel
14:14 ravenx but for the meant time
14:14 ravenx how can i do that in my state file
14:15 ravenx i've tried setting the -user to root, and that didn't help
14:15 ravenx unless i 'cheat' by doing cmd.run'
14:18 flowstate ype
14:18 flowstate yep*
14:18 flowstate cheat
14:18 ravenx :/
14:18 ravenx gah.
14:18 ravenx "IT DIDNT HAVE TO BE THIS WAY!!!"
14:18 flowstate well, you know what is happening
14:19 flowstate not why
14:19 ravenx fucking bower.
14:19 flowstate hahaha
14:19 ravenx first, i'm gonna try a local bower install first.
14:19 ravenx maybe that'll help
14:19 ravenx instead of doing a npm -g
14:19 ravenx if that still doesn't work, i will go back to npm -g and then cheat.
14:20 flowstate now if only I could actually cheat somehow and orchestrate my iam_profile, instead of salt crashing, I'd be a happy boy
14:20 ravenx iam_profile?
14:20 Brew joined #salt
14:20 flowstate AWS stuff
14:21 flowstate someone send out the Ryan D Lane signal. Apparently the problem I'm seeing has been around since April '15, and it's completely blocking me
14:21 OsakaFoo left #salt
14:24 andrew_v joined #salt
14:31 mavhq joined #salt
14:33 mavhq joined #salt
14:34 subsigna_ joined #salt
14:36 XenophonF flowstate: what error are you encountering?
14:36 dgutu joined #salt
14:36 XenophonF and with what versions of salt?
14:37 flowstate 2015.8.8
14:37 flowstate TypeError: __init__() got an unexpected keyword argument 'object_pairs_hook'
14:37 XenophonF i'm running a master plus minions in aws with salt-cloud, plus instance profiles (both masters and minions)
14:37 flowstate File "/usr/lib/python2.6/site-packages/salt/modules/boto_iam.py", line 900, in get_role_policy
14:37 XenophonF doing what? can you post the full trace to gist or something?
14:37 flowstate I'm not using salt-cloud, I'm using the boto library
14:38 flowstate http://pastebin.com/CGL5duL0
14:38 flowstate basically, on Amazon Linux, salt uses python 2.6
14:39 flowstate and so there's a version mismatch between the standard python on the box and the salt one, and the awscli versions are different
14:39 numkem joined #salt
14:39 flowstate and apparently the versions of json aren't compatible
14:40 flowstate here's the exact issue I'm describing in a github issue
14:40 flowstate https://github.com/saltstack/salt/issues/22290
14:40 saltstackbot [#22290]title: modules/boto_iam.py json parsing issue | Running Amazon Linux 2015.03 with Salt installed via bootstrap (yum packages)...
14:41 flowstate if I could somehow just switch salt to use my python 2.7, that would be great. I'm really hoping I don't have to set up a full venv to do this
14:41 CampusD joined #salt
14:42 spiette joined #salt
14:42 berserk joined #salt
14:43 LondonAppDev joined #salt
14:44 gtmanfred you would need to either repackage the rpms, or pip install
14:44 XenophonF my solution was to switch away from amazon linux :(
14:44 gtmanfred if you are using amazon linux based on rhel 6, i don't think you can get python2.7
14:45 flowstate so I have to switch operating systems because of an issue that's been known since April 2015?
14:45 flowstate ouch.
14:45 flowstate the python version pre-installed on new amazon linux boxes is already 2.7
14:45 flowstate so I know it def runs on the OS
14:46 XenophonF i migrated us off amazon linux (mostly) because of outdated packages compared to rhel7/centos7
14:46 onlyanegg joined #salt
14:46 XenophonF and even those are behind
14:46 gtmanfred XenophonF: meh, not always
14:46 gtmanfred XenophonF: https://ius.io
14:47 gtmanfred makes it pretty nice
14:47 gtmanfred plus for python 2.7.5 they backported the ssl module that was backported from 3.4 to 2.7.10
14:47 LondonAppDev joined #salt
14:47 gtmanfred so you aren't really missing out on anything
14:48 gtmanfred and ius.io has python 3.5.1 and you don't have to use the dumb SCL library (and can get the newer package versions on centos)
14:48 bharper XenophonF: how do you like amazon linux?
14:48 XenophonF i've abandoned it entirely except for two servers
14:48 orion XenophonF:
14:48 gtmanfred bharper: do you just get pinged everytime i say IUS?
14:49 orion XenophonF: Hey, what was your standard procedure for bootstrapping a salt master?
14:49 bharper gtmanfred: I have it set for all channel on freenode.  I might need to change that
14:49 gtmanfred IUS
14:49 gtmanfred <3
14:49 bharper hehe
14:49 mavhq joined #salt
14:49 XenophonF orion: https://gist.github.com/xenophonf/d8da7f47ea29d9ad46e7
14:49 XenophonF that's for centos7
14:50 orion XenophonF: Wow, thanks! That was an incredibly fast response.
14:50 XenophonF that script is how i deployed my production centos7 salt-master at work
14:51 XenophonF if you decide to run it yourself, run a search and replace "..." with meaningful stuff
14:51 flowstate joined #salt
14:52 XenophonF and "example.com"
14:52 XenophonF &c.
14:52 mavhq joined #salt
14:52 XenophonF orion: https://gist.github.com/xenophonf/95357d87b6e0b5e2b0e6
14:52 XenophonF that is how i setup my salt-states git repo
14:53 XenophonF where the master branch and the dev/test/stage/prod branches are separate from one another
14:54 punkoivan joined #salt
14:54 XenophonF gtmanfred: thanks for the link to IUS (hi bharper)
14:54 XenophonF i'll check it out
14:54 XenophonF we do a lot of bioinformatics work using python, so getting the right versions of everything is a big deal
14:54 XenophonF although mostly we just deploy biolinux on top of ubuntu 14 and call it done
14:54 flowstate joined #salt
14:55 bharper XenophonF: is you got questions about IUS feel free to hop into #iuscommunity
14:55 XenophonF i'll definitely check that out, thanks for the invite!
14:55 bharper anytime
14:57 flowstate okay, so I just monkey-patched my way around the issue. Just curious, is there a generic way to get more verbose failure comments from failed states?
14:58 flowstate because "Result: False" " Comment: Failed to create launch configuration." is just not super helpful
14:58 AndreasLutro flowstate: no, how much data is returned from a failed state is up to the state itself
14:58 flowstate k, figured so
14:59 flowstate you win this round, Ryan D Lane
14:59 flowstate (and the rest of the Lyft devops team)
14:59 AndreasLutro you could make a PR to include more info :)
14:59 flowstate I'm just about to, GET OUT OF MY HEAD!
14:59 flowstate :)
15:00 AndreasLutro :D
15:02 gcorey joined #salt
15:04 TooLmaN joined #salt
15:05 DammitJim my minions are running ubuntu and I am getting the salt-minion packages from the ppa
15:06 DammitJim should I be using something else to get a newer version than 2015.5.3 ?
15:07 AndreasLutro repo.saltstack.com
15:07 DammitJim is that the recommended repo?
15:07 gtmanfred yes
15:07 DammitJim I don't know why I ended up using the ppa.launchpad.net
15:07 gtmanfred that used to be the recommended way
15:07 DammitJim thanks... now I need to salt this upgrade ;)
15:08 oida joined #salt
15:09 flowstate okay, this is super funky
15:09 flowstate I just restarted salt-minion using -l debug
15:10 DammitJim lol... is there a state to do this?
15:10 flowstate and am executing code which has an except block defined as http://pastebin.com/V4hYrB5r
15:10 flowstate note 'log.debug(e)'
15:10 flowstate but the only thing I'm seeing in my output is the error output
15:11 _JZ_ joined #salt
15:11 RobSpectre joined #salt
15:12 slav0nic joined #salt
15:13 aw110f joined #salt
15:18 Kelsar joined #salt
15:19 rlatimore joined #salt
15:21 source47 joined #salt
15:22 source47 hi guys, i am having problems installing epel, when i install it installs it fine but still complains when it updates from null to 7-5 ? i am using centos 7 fyi
15:24 winsalt joined #salt
15:25 source47 fyi this is my sls https://gist.github.com/developerinlondon/aaacea3e0a5a2ceaad1148ef3da5acf3
15:26 bowhunter joined #salt
15:31 kshlm joined #salt
15:32 fracklen joined #salt
15:34 punkoivan joined #salt
15:38 edrocks joined #salt
15:38 quix_ joined #salt
15:38 mavhq joined #salt
15:41 quix_ joined #salt
15:43 punkoivan left #salt
15:50 mavhq joined #salt
15:53 spuder joined #salt
15:56 mavhq joined #salt
15:56 UtahDave joined #salt
15:56 UtahDave morning!
15:57 gtmanfred HI!
15:57 onlyanegg joined #salt
16:00 brianvdawson left #salt
16:12 ageorgop joined #salt
16:14 writtenoff joined #salt
16:15 XenophonF i like how the shell grain on freebsd defaults to /bin/csh
16:15 XenophonF like God intended it
16:15 XenophonF now to fix my state definition that assumes /bin/sh is the default ;)
16:16 overyander joined #salt
16:20 jeffspeff joined #salt
16:22 grumm_servire joined #salt
16:29 Ryan_Lane flowstate: fixed your issue?
16:29 hightekvagabond joined #salt
16:29 ninjada joined #salt
16:30 Ryan_Lane We have plans on fixing all of the modules for them to send back proper error messages
16:30 Ryan_Lane And to switch them all to boto3, but it'll be a long slow process :)
16:30 salty_solution joined #salt
16:32 beardedeagle joined #salt
16:34 djgerm left #salt
16:36 flowstate joined #salt
16:37 Netwizard joined #salt
16:43 VulcanRi1r joined #salt
16:44 flowstate joined #salt
16:44 beardedeagle Any chance I can get some eyeballs on this: https://github.com/saltstack/salt/issues/32267
16:44 saltstackbot [#32267]title: ImportError: No module named mako.exceptions | ### Description of Issue/Question...
16:45 mavhq joined #salt
16:53 barajasfab joined #salt
16:53 rm_jorge joined #salt
16:54 mavhq joined #salt
17:04 mavhq joined #salt
17:08 sfxandy joined #salt
17:09 bluenemo joined #salt
17:09 sjmh joined #salt
17:11 spuder_ joined #salt
17:16 drawsmcgraw There's a simple way of restarting the salt-minion after an update, isn't there?
17:16 drawsmcgraw I mean as an alternative to using `at` to schedule a restart ` minute from now?
17:18 AndreasLutro what's wrong with at?
17:19 drawsmcgraw AndreasLutro: not a thing. I just could've sworn I'd read some discussion that something like 'service.restart salt-minion' worked seamlessly now. The test I ran just now tells me that's not quite the case.
17:19 AndreasLutro what versions are you upgrading to/from?
17:20 drawsmcgraw 2015.5.3 -> 2015.8.8-2
17:20 drawsmcgraw of course, 'service.restart salt-minion' does, indeed, restart the minion. I just don't get the job response back (for obvious reasons)
17:20 PeterO joined #salt
17:21 drawsmcgraw AndreasLutro: ping
17:21 AndreasLutro that's a big upgrade
17:22 AndreasLutro but when you upgrade salt, the package upgrade takes care of the restart for you
17:23 drawsmcgraw wait, it does?
17:24 drawsmcgraw On CentOS (6|7) ?
17:25 AndreasLutro well it does on debian/ubuntu... if it doesn't on centos that'd be disappointing
17:26 drawsmcgraw I actually feel the other way around. I'm annoyed that Debian/Ubuntu start services upon installing :P
17:26 drawsmcgraw Though I can certainly see the utility for when I upgrade a package.
17:27 AndreasLutro agreed :)
17:34 whytewolf joined #salt
17:39 akhter joined #salt
17:39 akhter joined #salt
17:41 harkx joined #salt
17:42 mavhq joined #salt
17:43 GreatSnoopy joined #salt
17:47 mavhq joined #salt
17:49 drawsmcgraw AndreasLutro: Interesting. I have some Centos 6 machines. After importing the GPG key and configuring the new Yum repo, I ran "pkg.install pkgs=['salt-minion','python-crypto']".
17:49 drawsmcgraw The minion never returned
17:50 drawsmcgraw And examining the targeted minion, it looks like the service was, indeed, restarted...
17:59 Fabbe joined #salt
17:59 mavhq joined #salt
18:01 baweaver joined #salt
18:02 AndreasLutro beardedeagle: you said you ran the python repl and imported mako... you did that on the minion, right?
18:03 hightekvagabond joined #salt
18:05 thejrose1984 joined #salt
18:05 mavhq joined #salt
18:08 rmnuvg joined #salt
18:10 beardedeagle @AndreasLutro: good point
18:10 beardedeagle closed the issue
18:11 dariusjs joined #salt
18:12 mavhq joined #salt
18:13 baweaver joined #salt
18:13 spiette joined #salt
18:13 nickfrez joined #salt
18:18 dmaiocchi joined #salt
18:19 mavhq joined #salt
18:20 i90rr joined #salt
18:22 Edgan iggy: ping
18:23 amcorreia joined #salt
18:26 flowstate joined #salt
18:28 nickfrez joined #salt
18:30 chutzpah joined #salt
18:31 ninjada joined #salt
18:33 s_kunk joined #salt
18:35 flowstate joined #salt
18:39 TheBigNoob joined #salt
18:40 XenophonF when did salt-cloud start using "driver" instead of "provider"?
18:41 mavhq joined #salt
18:41 drawsmcgraw XenophonF: 2015.8.....something
18:41 drawsmcgraw May have been the first cut of 2015.8
18:43 XenophonF thanks
18:44 fracklen joined #salt
18:48 nickfrez Hey all.  I'm fairly new to Salt, and I've been getting some errors on the gpg.create_key example here: https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.gpg.html#salt.modules.gpg.create_key
18:48 nickfrez specifically a KeyError: 'USERNAME'  https://gist.github.com/jacktrades/5653fb2453e8c5accd8174dd67cb4771
18:49 nickfrez I see there's a pull request that references examples in gpg not working https://github.com/saltstack/salt/pull/31063
18:49 saltstackbot [#31063]title: Fixing typos,  | many examples in here still don't work  (ex. salt -t 15 '*' gpg.create_key throws errors), but some low hanging fruit is fixed.
18:49 GreatSnoopy joined #salt
18:50 nickfrez But I don't see a bug report for this.  Is this a known issue?
18:51 ahammond using python-requests-1.1.0 from repo.saltstack.com is completely non-functional: crashes even on simple requests.get("http://google.com/")
18:52 ahammond any reason not to use requests 2.6.0?
18:52 mavhq joined #salt
18:54 AndreasLutro ahammond: which distro is that?
18:55 ahammond AndreasLutro we're on a centos6
18:55 ajw0100 joined #salt
18:56 orion How does Salt handle conflicting configuration settings in the master.d/*.conf files?
18:57 UtahDave ahammond: I'm not sure. Let me ask
18:58 ahammond @UtahDave thanks man!
19:00 UtahDave ahammond: actually, I just tried installing from our repo on cent6 and it's giving me python-requests  2.6.0-3.el6
19:01 ahammond @UtahDave I wonder where I'm getting the 1.1.0?
19:01 UtahDave are you using epel, maybe?
19:01 ahammond nope: http://repo.saltstack.com/yum/rhel6/
19:01 UtahDave does an update upgrade it?
19:02 UtahDave interesting.  I'm getting requests from "base"   so it's not using the saltstack one.
19:02 ahammond that saltstack one looks broken.
19:03 ahammond we're currently stuck on cento6.2, but I can pull the base one from 6.7 into our compat repo.
19:03 ahammond yuck.
19:03 UtahDave Yeah, I'm following up on that.
19:06 mavhq joined #salt
19:08 baweaver joined #salt
19:10 UtahDave ahammond: do you have the main Centos repos enabled? It looks like that was providing python-requests 2.6.0 for me.
19:13 UtahDave Oh, my test vm was Cent 6.7.
19:14 jerredbell joined #salt
19:14 joshin joined #salt
19:21 GreatSnoopy joined #salt
19:24 iggy Edgan: pong?
19:24 iggy orion: last one wins would be my guess
19:25 tapoxi joined #salt
19:25 orion iggy: How is "last" defined?
19:26 iggy inter-file order -> intra-file order
19:26 tapoxi hey all, anyone using salt with the foreman? any idea where it wants me to put states?
19:26 orion Alright
19:26 orion iggy: You mean lexicographical order, right?
19:30 brotatochip joined #salt
19:30 mavhq joined #salt
19:32 ageorgop joined #salt
19:32 SirLagz joined #salt
19:34 jhauser joined #salt
19:35 fredvd joined #salt
19:37 flowstate joined #salt
19:42 baweaver joined #salt
19:42 onlyanegg joined #salt
19:44 ajw0100 joined #salt
19:44 patrek joined #salt
19:45 fracklen joined #salt
19:48 pwhack joined #salt
19:49 mavhq joined #salt
19:52 mavhq joined #salt
19:55 nethershaw joined #salt
19:56 onlyanegg joined #salt
20:05 flowstate Ryan_Lane: I'm actually going to do a PR against the create_launch_configuration code in boto_asg.py so that you can use security group groupNames if you're not in a default VPC
20:06 flowstate in terms of the json thing, I just added a monkey patch to "import simplejson as json" into my master bootstrap script
20:06 flowstate not sure how else to get around it
20:06 flowstate I'll probably submit a PR implementing your suggested change
20:07 Ryan_Lane We're not using the default vpc
20:07 aharvey joined #salt
20:07 Ryan_Lane What's the issue with the secgroup?
20:07 nethershaw joined #salt
20:07 flowstate You may not reference Amazon VPC security groups by name.
20:07 flowstate (the error message from AWS_^
20:07 Ryan_Lane I do
20:08 flowstate haha, I'm not saying it myself, hold on, I'll pastebin it
20:08 Ryan_Lane If it detects the vpc it'll translate the names into ids
20:08 aharvey Any users with major ruby-fu mind throwing a hand at this one?  salt ‘host1’ cmd.run ‘ruby —version’.  response = 2.0.  login to hosts, rvm profile gets parsed, proper version returns (not 2.0).  Obviously that is not being sourced by salt when I’m doing a state run, but the proper ruby version is a hard requirement.  Ideas on how to reconcile this?
20:09 Ryan_Lane But you need to tell it the vpc info
20:09 aharvey currently RVM is handling this old version of ruby
20:09 flowstate http://pastebin.com/tyH2Ys8u
20:09 flowstate oh, really?
20:09 flowstate I must have missed that
20:10 flowstate aaaaaaand I can remove that branch
20:10 flowstate how do you specify that? I'm not seeing it in the docs
20:11 flowstate "List of Names or security group id’s of the security groups with which to associate the EC2 instances or VPC instances, respectively."
20:11 flowstate the docs seem to track with that idea
20:11 Ryan_Lane Flowstate https://github.com/lyft/confidant/blob/master/salt/orchestration/confidant.sls
20:12 flowstate I can still submit a PR! clearly, that's what matters here
20:12 Ryan_Lane See the asg state there
20:12 Ryan_Lane It doesn't hurt to submit a pr, but it may waste your time :)
20:12 flowstate right, is it the vpc_zone_identifier?
20:12 Ryan_Lane Docs can definitely be improved
20:12 Ryan_Lane Yep
20:13 flowstate okay
20:13 flowstate so that's a property of the asg, not the lc, right?
20:13 Ryan_Lane If it has an identifier it'll translate sg names
20:13 flowstate that may be the missing piece
20:13 Ryan_Lane Note that the lc is inside of the asg
20:13 flowstate since I was doing a separate lc
20:13 flowstate yep, that's the key
20:13 flowstate I will be switching to doing it that way, then
20:13 flowstate cool! Always good to talk to the source himself
20:13 Ryan_Lane Its a bit of magic the states have :)
20:13 flowstate tyvm sir
20:14 Ryan_Lane And it's super cool because it'll automatically update the asg and fully manage the lcs
20:14 sroegner joined #salt
20:15 VulcanRi1r Greetings all. I'm a Salt N00b coming from Puppet for my home network of ~20 machines and containers. I'm migrating from Linux to PC-BSD, and since my puppetmaster broke, I figured salt is more BSD friendly. I'm working through the docs on the website, is this the best approach?
20:15 flowstate super sexy. I didn't see why you'd nest the lc, but I definitely get it now
20:15 flowstate thanks again
20:29 UtahDave VulcanRi1r: Yeah, definitely go through the Walkthroughs and tutorials
20:34 mavhq joined #salt
20:35 babilen VulcanRi1r: https://docs.saltstack.com/en/getstarted/
20:35 UtahDave ahammond: the next point release of Salt should have the updated python-requests
20:35 djgerm joined #salt
20:36 flowstate oh, Ryan_Lane, if you're still around, I'm suuuper curious how you handle the hostname part of your naming scheme
20:37 flowstate I get everything else about it, except the hostname / node number part
20:38 Ryan_Lane That's totally fake
20:38 ronnix joined #salt
20:38 Ryan_Lane We set it on the hosts, but there's no real dns for them
20:38 Ryan_Lane Elbs get real dns entries
20:38 Ryan_Lane Boto elb module will manage it's own records
20:39 ahammond @UtahDave thanks man!
20:39 UtahDave you're welcome!
20:43 jeffevans joined #salt
20:44 jeffevans hello, i'm getting started with orchestration.  i've read that states in an orchestration file happen sequentially or based on requires.  is there a way to run multiple orchestrations in parallel?  for example, an orchestration to bring up cassandra while running another orchestration to bring up zookeeper?
20:45 djgerm if want to be storing my user's hashes in a pillar, and using those to populate password: for the user.present state, how would I do that?
20:48 onlyanegg joined #salt
20:50 keisetsu I'm sure I'm just not getting it, but I need a certain user to exist before everything else runs. I tried require'ing the user in each state, but that doesn't work if I only run the state for a single service, for example, instead of the high state. Is there a good way to require a user be created? I guess another option could include the sls where the user is created, is that the way to go?
20:53 XenophonF can anyone offer any advice for templating xml files?
20:53 salty_solution XenophonF: what are you trying to do?
20:54 XenophonF the docs mention a non-existent xml_badgerfish serializer
20:54 babilen Just write them with Python in the run() function, XenophonF
20:54 babilen (i.e. return the string)
20:54 XenophonF i'm writing states that deploy/configure both the Shibboleth IdP and the Shibboleth SP
20:54 XenophonF so i have configs that look like this: https://raw.githubusercontent.com/irtnog/shibboleth-formula/master/shib/sp/files/shibboleth2.xml
20:55 sfxandy joined #salt
20:55 XenophonF and i've come up with yaml that looks something like this: https://raw.githubusercontent.com/irtnog/shibboleth-formula/master/shib/sp/files/shibboleth2.yaml
20:55 XenophonF but i don't like it
20:55 babilen I'd use et in Python
20:55 baweaver joined #salt
20:55 XenophonF et? elemtree?
20:55 babilen yeah
20:56 babilen or whatever generator you fabcy
20:57 rem5 joined #salt
20:57 XenophonF the generating part isn't what i'm struggling with, i guess
20:57 XenophonF it's how to represent the xml data in yaml
20:58 babilen Why do you have to?
20:58 onlyanegg joined #salt
20:59 XenophonF well, i don't _have_ to, i guess i could just template the xml file the way i might an html file
20:59 wych joined #salt
21:00 XenophonF the downside is that i'd have to account for every possible setting in shibboleth2.xml (among other config files)
21:00 XenophonF hm, i'll think about that
21:00 babilen suggestion is to use the py renderer for the file and to simply hand over suitable data. In there you use normal XML generators in the Python program. Naturally that data has to come from somewhere ..
21:02 chmod666org joined #salt
21:02 forrest joined #salt
21:04 tapoxi left #salt
21:06 UtahDave keisetsu: yeah, include that user sls file and require it.  Or include it and use  order: first
21:06 keisetsu UtahDave: thanks
21:07 UtahDave djgerm: use the shadow state to set the password hash
21:07 keisetsu UtahDave: Do you know if salt keeps track of previously run states like that, or is it going to call the method every time to check?
21:08 RobSpectre joined #salt
21:08 UtahDave keisetsu: Salt states will check the current state of your system before making any changes. So it will check for the correct hash each time you execute that state
21:09 keisetsu UtahDave: So, if I add my user check to a bunch of services and I do run highstate, will it check everytime? I'm guessing, compared to the other actions it is taking, a check like that is going to be relatively inexpensive, but just to be sure.
21:10 keisetsu UtahDave: Is that something salt itself is keeping track of, or the user module?
21:10 keisetsu (just wondering if I need to keep track of that stuff in my custom modules)
21:10 UtahDave Oh, no.  It will only check it one time, even if it's required by 1000's of other states
21:10 edrocks joined #salt
21:11 keisetsu Awesome
21:13 ninjada joined #salt
21:14 UtahDave Let me clarify. It will check the user password hash once per state call or highstate.  So if your highstate includes 1000's of states, the user password hash check will only happens once.
21:14 UtahDave But it will happen each time you call a highstate.
21:15 keisetsu Got it.
21:16 keisetsu Ok, I have another question: I've been using relative paths for includes, in the salt states it works from the same location as the salt state, but pillar seems to always need to state at the pillar base. Is there a reason for this?
21:17 keisetsu "need to start at"
21:21 UtahDave keisetsu: I'm not sure if that's just an oversight or if there's a security reason for that.
21:22 keisetsu UtahDave: Hmm, interesting, I guess I could see security might be a reason.
21:22 UtahDave keisetsu: likely it just hasn't been requested.  Feel free to open a github issue requesting that and  start a discussion
21:22 keisetsu I'll do that
21:23 UtahDave cool.
21:23 UtahDave I'll be back later.
21:23 UtahDave left #salt
21:24 amcorreia joined #salt
21:25 keisetsu UtahDave: I know you've left, but to follow up: Issue #8875 was raised a couple of years ago, still waiting for a solution.
21:25 onlyanegg joined #salt
21:27 aw110f joined #salt
21:40 TheBigNoob joined #salt
21:48 ninjada joined #salt
21:51 brotatochip joined #salt
21:52 flowstate joined #salt
22:03 andrew_v joined #salt
22:03 averell joined #salt
22:07 rmnuvg joined #salt
22:08 baweaver joined #salt
22:09 rem5 joined #salt
22:10 ninjada joined #salt
22:11 ajw0100 joined #salt
22:15 kliquori joined #salt
22:16 mavhq joined #salt
22:19 linjan joined #salt
22:23 dhscholb does anyone here use the mysql returner for master_job_cache? have you noticed that the salt_returns table grows very fast?
22:24 dhscholb the return and full_ret columns in salt_returns seem to be redundant and wasting a lot of space
22:25 baweaver joined #salt
22:28 justanotheruser joined #salt
22:28 jeddi joined #salt
22:35 mosen joined #salt
22:39 zenlot joined #salt
22:39 jagguli joined #salt
22:40 nickfrez left #salt
22:42 nickfrez joined #salt
22:43 evidence anyone using the pillar_cache ?
22:43 evidence i see the memory backend was marked expirimental, curious if anyone has used it in prod
22:46 brotatochip joined #salt
22:50 capricorn_1 joined #salt
22:59 aharvey joined #salt
23:00 bet0x joined #salt
23:07 TheBigNoob joined #salt
23:08 sprky joined #salt
23:13 ninjada joined #salt
23:16 rmnuvg joined #salt
23:23 berserk joined #salt
23:23 berserk joined #salt
23:25 maduro joined #salt
23:27 SubOracle joined #salt
23:27 North_ joined #salt
23:31 maduro hi folks, running debian jessie's salt (2014.1.13) and trying to use the docker module. it fails with 'State docker.build found in sls docker.latest_jessie_amd64 is unavailable'  even though i have docker-py installed via pip. sls contents and other relevant info at https://gist.github.com/anonymous/bf83650eedc9cfa9c220954e5bf7dee4
23:31 baweaver joined #salt
23:35 djinni` joined #salt
23:36 jerredbell joined #salt
23:40 mosu_ joined #salt
23:46 orion Hi. I'm running Salt 2015.8.8 on Ubuntu 14.04 with gitfs enabled (pygit2). I seem to be getting this error with every single command I type, such as salt-key: https://gist.github.com/centromere/9654ced2c3efee1c15d5450db0e6a629
23:47 iggy it's a warning
23:48 orion Does anyone know what could be going wrong?
23:48 orion Yes, it's technically a warning, but it's really annoying since it is printed, sometimes multiple times, every time I use any salt command.
23:50 ahammond I have a pillar which I can see when I do salt-call pillar.get. When I try to reference it in a state, I get "failed: Jinja variable 'dict' object has no attribute ...
23:54 spuder joined #salt
23:54 flowstate joined #salt
23:55 spuder_ joined #salt
23:56 orion Nevermind -- it's not a salt problem.

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