Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2016-06-21

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

All times shown according to UTC.

Time Nick Message
00:12 dendazen joined #salt
00:13 freelock joined #salt
00:27 p3rror joined #salt
00:28 sjmh hm.  has anyone tried to install a minion from the official repo onto a RHEL7u0 box?
00:29 flowstate joined #salt
00:29 mosen joined #salt
00:36 fannet_ joined #salt
00:39 edrocks joined #salt
00:40 druonysus joined #salt
01:10 flowstate joined #salt
01:16 fxhp joined #salt
01:17 keldwud joined #salt
01:21 dfinn joined #salt
01:27 iceyao joined #salt
01:29 flowstate joined #salt
01:33 iggy Joe630: just change the word
01:34 iggy there was driver in 2 places and it was confusing
01:34 iggy repl1cant: I don't think there's much to document... just change the setting on the master and minions
01:38 flowstate joined #salt
01:38 k_sze[work] regarding using an alternate file_roots again.
01:40 k_sze[work] Is it ok if I make an alternate CONFIG_DIR (instead of /etc/salt), which contains an alternate master file, which is mostly identical to the production /etc/salt/master file, except it points to my alternate file_roots?
01:41 k_sze[work] And then I would call `salt -c <my alt config_dir>`
01:46 mapu joined #salt
01:46 SheetiS joined #salt
01:48 ilbot3 joined #salt
01:48 Topic for #salt is now Welcome to #salt! | Latest Versions: 2015.5.10, 2015.8.10, 2016.3.1 | Support: 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
01:50 SheetiS joined #salt
01:51 k_sze[work] It's still a bit confusing.
01:53 k_sze[work] I'm sure that the salt-master daemon reads from the CONFIG_DIR when it starts up because it at least needs to know where the minion keys are, so that it can accept connections from the minions. So why does the `salt` command have a `-c` option at all?
01:54 k_sze[work] What can I safely override by passing an alternate CONFIG_DIR via the `-c` option?
01:55 SheetiS joined #salt
01:56 neogenix joined #salt
02:00 SheetiS joined #salt
02:02 k_sze[work] Or is the `-c` option really not there for overriding settings in general, but to tell the salt command which master to communicate with (in cause some master was started on non-standard ports)?
02:03 misconfig joined #salt
02:05 SheetiS joined #salt
02:06 misconfig hello all, is there any way to force a salt state to not display changes? Something similar to ansible's changed_when => http://docs.ansible.com/ansible/playbooks_error_handling.html#overriding-the-changed-result
02:10 SheetiS joined #salt
02:15 SheetiS joined #salt
02:15 zero_shane joined #salt
02:20 SheetiS joined #salt
02:22 _JZ_ joined #salt
02:25 SheetiS joined #salt
02:25 DEger joined #salt
02:30 SheetiS joined #salt
02:34 teryx510 joined #salt
02:35 SheetiS joined #salt
02:35 teryx510 joined #salt
02:36 teryx510 joined #salt
02:36 teryx510 joined #salt
02:40 SheetiS joined #salt
02:42 edrocks joined #salt
02:45 SheetiS joined #salt
02:46 ramteid joined #salt
02:50 SheetiS joined #salt
02:55 SheetiS joined #salt
02:56 tuxx joined #salt
03:00 SheetiS joined #salt
03:05 SheetiS joined #salt
03:10 SheetiS joined #salt
03:15 SheetiS joined #salt
03:18 vaelen joined #salt
03:33 SpX joined #salt
03:41 k_sze[work] exit
03:41 k_sze[work] oops
03:46 rem5 joined #salt
03:51 ajw0100 joined #salt
03:58 iggy misconfig: there are a number of output options
03:59 iggy `state_output: changes` is my favorite
04:03 neogenix joined #salt
04:13 illern joined #salt
04:19 iceyao joined #salt
04:29 flowstate joined #salt
04:31 sjmh joined #salt
04:37 DEger joined #salt
04:38 fannet_ joined #salt
04:45 edrocks joined #salt
04:45 ageorgop joined #salt
04:50 hamsham joined #salt
04:58 keimlink joined #salt
05:01 bfig joined #salt
05:02 bfig hello, I'm having issues with vagrant+salt. vagrant up hangs at 'Calling state.highstate... (this might take a while)' this is my Vagrantfile https://gist.github.com/bfig/565c5eb97a58154410a3aa4c2f027121
05:03 iggy if you can't get some sort of output from Salt, it's going to be hard to track down what's going
05:05 bfig iggy: i put salt.verbose = true in the top level, shouldn't that give me the debug level output?
05:10 iggy I avoid vagrant at all costs... so no clue
05:13 bfig iggy: why do you avoid it
05:13 iggy it offers no value to me
05:14 bfig what do you do?
05:14 iggy a job
05:15 iggy one that isn't assisted by a program that is a frontend for virtualbox
05:17 bfig i don't really see your point, I'm really new to devops and the implications are not clear
05:17 sauvin joined #salt
05:18 iggy most of what I work on requires actual hardware (video cards, broadcast equipment, etc)
05:18 iggy stuff that isn't easily emulated in VMs
05:19 iggy if I was going to use something like that, it'd probably be lxd
05:19 bfig ahh ok, well it makes sense for you to not use it. Don't understand why you would specifically avoid it... it'd be like avoiding cobol for a web programmer
05:20 iggy there is no vagrant backend for lxd yet
05:20 bfig well, I don't see why there wouldn't eventually be one
05:20 iggy because hashicorp has a replacement for vagrant
05:21 iggy and ubuntu certainly aren't going to write the backend (they are all in on juju and openstack)
05:21 bfig i know almost zero about the devops landscape. Do you think those tools are better than vagrant?
05:22 bfig or at least, where would I get an overview of the most used tools out there?
05:22 iggy I think there is more than one tool that do similar things for a reason
05:22 iggy I don't know of a good overview of devops tools
05:23 iggy (aside from the fact that devops is such a broad term that those tools could cover a huge amount of territory)
05:24 KajiMaster joined #salt
05:24 iggy and yes, that is probably something that would be of good use, so if you find one, bookmark it for the next person you see asking
05:25 bfig i suppose provisioning, configuration, application lifecycle management and administration are the biggest topics, and they are super broad, sure
05:25 iggy monitoring is often lumped into devops
05:25 bfig monitoring < administration
05:25 iggy _that_ is a whole bunch of different tools
05:26 anmol joined #salt
05:27 bfig software stacks and reqs are different enough that no tool can reliably default to a monitoring/lifecycle configuration i think, but maybe a core 3-4 tools should be able to cover about 80% of cases
05:28 duncanmv_ joined #salt
05:28 flowstate joined #salt
05:29 rem5 joined #salt
05:29 bfig think I found my issue: '../...'
05:44 rdas joined #salt
05:51 neogenix joined #salt
05:52 iceyao joined #salt
05:56 kshlm joined #salt
06:10 impi joined #salt
06:16 evle1 joined #salt
06:19 kevinquinnyo joined #salt
06:21 Ayo joined #salt
06:21 DEger joined #salt
06:21 Elsmorian joined #salt
06:28 KermitTheFragger joined #salt
06:30 flowstate joined #salt
06:32 duncanmv_ joined #salt
06:39 fannet_ joined #salt
06:43 duncanmv_ joined #salt
06:46 iceyao joined #salt
06:47 fracklen joined #salt
06:47 edrocks joined #salt
06:51 fracklen joined #salt
06:56 iceyao_ joined #salt
07:10 CeBe joined #salt
07:14 voxxit joined #salt
07:16 keimlink joined #salt
07:17 dmaiocchi joined #salt
07:20 fracklen joined #salt
07:21 toanju joined #salt
07:22 tpaul joined #salt
07:22 kawa2014 joined #salt
07:23 sagerdearia joined #salt
07:25 manji joined #salt
07:26 ITChap joined #salt
07:28 flowstate joined #salt
07:38 dmaiocchi joined #salt
07:39 Electron^- joined #salt
07:42 linjan_ joined #salt
07:44 ageorgop joined #salt
07:46 deniszh joined #salt
07:48 Rumbles joined #salt
07:49 nmadhok joined #salt
07:50 iceyao joined #salt
07:51 iceyao_ joined #salt
07:55 fracklen Hey! Is there a way in the pillar top.sls to include other top.sls-files?
07:55 krymzon joined #salt
07:58 AndreasLutro not really. you can use jinja includes, though
07:59 neogenix joined #salt
08:00 fracklen AndreasLutro: thanks - do you know why? It makes sense to me, when your setup grows...
08:01 nmadhok joined #salt
08:02 AndreasLutro dunno
08:02 ronnix joined #salt
08:04 lero joined #salt
08:04 tuxick "Comment: State 'mysql_database.present' was not found in SLS"
08:05 tuxick what am i supposed to do about that?
08:07 babilen tuxick: The typicaly reason for that is that you are missing a (Python module) dependency of the corresponding execution module
08:07 tuxick on master?
08:07 babilen https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.mysql.html
08:08 babilen No, on the minion. You will also have to refresh modules. See the "depends:
08:08 babilen MySQLdb Python module
08:08 tuxick oh ok cheers
08:08 nmadhok joined #salt
08:08 babilen I'd just install that package with a pkg.installed and include "refresh_modules: True" and then require that state.
08:09 tuxick ack, was already at it :)
08:11 babilen Most modules have some tests like "check if module A, B and C is installed and if I'm running on a platform that has D, E and F available" and won't load if these requisites are not satisfied
08:12 s_kunk joined #salt
08:13 felskrone joined #salt
08:18 Edgan joined #salt
08:28 whitenoise joined #salt
08:29 flowstate joined #salt
08:31 Elsmorian joined #salt
08:45 joj[] joined #salt
08:47 joj[] What is the preferred git backend for gitfs? pygit2, gitpython, dulwich? what is the rationale based on which I would make a choice among them? https://docs.saltstack.com/en/latest/topics/tutorials/gitfs.html doesn't list any good reason, or why there are three options in the first place.
08:50 babilen I have been using gitpython for some time and didn't have a problem. Rationale for using it: It works
08:50 babilen Rationale for using pygit2: You want some of its features
08:50 edrocks joined #salt
08:50 permalac joined #salt
08:52 joj[] babilen: any differences in speed/performance/reliability in your experience? pygit2 seems to be a binding to libgit2 via C whereas gitpython is pure Python and calls out to git.
08:52 babilen joj[]: Do you have performance issues?
08:54 joj[] babilen: nope, building the core infrastructure right now and making choices. However I can imagine that using libgit might prevent bugs and mistakes when interfacing with git building command lines and such, to some extent. I think I'll go with pygit2 then.
08:57 AndreasLutro 90% of the time spent interfacing with git will be talking to the remote server anyway
08:58 AndreasLutro pygit2 has more features (ssh key authentication) but I had to compile it + libgit2 myself with very specific versions to get it to work
08:58 impi joined #salt
09:01 fredvd joined #salt
09:04 DEger joined #salt
09:04 joj[] AndreasLutro: well; using 16.04 lts here with pygit2 and SSH key auth works out of the box w/o problems.
09:06 glickind joined #salt
09:07 AndreasLutro neat
09:11 om joined #salt
09:11 babilen joj[]: If it works and is easy to deploy on most platforms then I'd use pygit2.
09:11 babilen s/most/all/ I guess :)
09:15 Favoretti joined #salt
09:16 Favoretti Ehlo salty people. I've got a rhethorical question that maybe someone out of the community can help me figure out. Coming from a puppet world and moving away to salt, we are kinda used to the workflow when one can test code changes out of a personal environment. Now, salt, having implemented merge strategies, kinda can do that as well, only if pillar merge strategy would work the same as state merge strategy. Is there anyone out there that can share their
09:17 babilen Favoretti: Your message got truncated at "hat can share their …"
09:19 west575 joined #salt
09:20 Favoretti truncated in IRC? :) alright
09:20 Favoretti > Is there anyone out there that can share their pipeline/change strategy around salt codebase?
09:22 AndreasLutro Favoretti: I have all the salt states/pillars git cloned to my local computer, where I use a small cluster of vagrant VMs to test them in development immediately. you can set up the same sort of pipeline for testing using things like serverspec
09:23 AndreasLutro salt does lack a bit in the testing department compared to puppet, not much to do about that
09:26 Favoretti AndreasLutro: well, I was talking to someone from saltstack, technically if pillars merging strategy would be adjusted to the same states use - it'd be golden. https://github.com/saltstack/salt/issues/28682 this one
09:26 saltstackbot [#28682][OPEN] Support top_file_merging_strategy=same equivalent for pillar | I have a similar use-case as what introduced top_file_merging_strategy=same (see https://github.com/saltstack/salt/issues/12483), and use it to ensure that only the top file of my current saltenv gets used. This allows me to keep identical topfiles in each environment, simply by using '{{ env }}:' instead of hardcoding the environment names in the topfiles. I can the
09:27 _mel_ joined #salt
09:27 Favoretti But so far everyone i'm talking to outside of salt - concur that testing is a bit of a pain.
09:28 AndreasLutro I don't see why you need to adjust pillar/state merging to test things
09:28 AndreasLutro I guess you want a separate top.sls for local dev? or testing?
09:29 Favoretti AndreasLutro: well, each release/change for us traverses multiple UAT/QA envs, loadtesting envs, etc, before it hits production. Those environments can share the same state files, but pillar data differs.
09:29 flowstate joined #salt
09:29 AndreasLutro well, you could just set up different matches for them in the single pillar top.sls
09:29 AndreasLutro assuming their hostnames are different or you have a grain you can match on
09:30 AndreasLutro or you could store the bulk of pillar data in some sort of database/api and use external pillars
09:30 Favoretti separate top file won't work, would it? It would mean a ton of duplication of the pillar files/sources. {{ env }} works great on states, not sure why pillars are still lagging behind that change :)
09:30 AndreasLutro duplication sure... I don't use salt envs at all
09:30 AndreasLutro I prefer duplication to unnecessary complexity
09:32 Favoretti envs would work great with different git branches. Let's say if you have a master branch that defaults to production, one could set up branch names the same as the env names and salt would automatically look in them once executed in a different environment
09:34 jhauser joined #salt
09:34 Favoretti but yeah, duplication is the only way I see for now, hence wondering if anyone has a different perspective :)
09:36 oida joined #salt
09:43 kbaikov joined #salt
09:45 SpX joined #salt
09:48 Rumbles joined #salt
09:50 Guest41482 joined #salt
09:52 om2 joined #salt
10:10 tuxick is it possible/wise to have roles include other roles? this to need to set only one in minion grains
10:14 neogenix joined #salt
10:17 babilen tuxick: What does that mean exactly? You can target states to multiple roles (regardless of where you set/maintain those (pillars, grains, ...)).
10:17 ronnix_ joined #salt
10:17 tuxick that sounds like doing it via top.sls no?
10:18 tuxick what i mean is that i'd like to define role: some-spefic-rol on minion, so i won't have to bother with specifying the actual roles required
10:21 Rumbles joined #salt
10:21 babilen You can define "role: some-specific-role" anywhere you like, but you will still have to maintain a "state <-> role" mapping. One way to do that is that target states by role in top.sls (based on pillar data or grain data or minion id or ..)
10:22 tuxick yes, and i was hoping i could do that via some 'include'
10:22 babilen Why would you want to save it on the minion?
10:22 tuxick no, i want to have to have just one role: on the minion
10:23 babilen Yes, I understand, but why do you want to save it on the minion in the first place?
10:23 tuxick uhm isn't that how roles work?
10:23 babilen Only if you design them that way
10:23 babilen You can easily set them in pillars and target by I@
10:24 tuxick well yes, and for now the idea here is to use the roles approach
10:24 tuxick might well turn out next week i'll convince them to do change that
10:24 tuxick investigating :)
10:24 babilen Either way: You can aggregate in states (with includes), simply target states to multiple roles or include pillars in pillar data.
10:25 tuxick ok
10:26 babilen https://www.refheap.com/120645 more or less
10:26 babilen (pillar data)
10:27 babilen s/include pillars in pillar data/include roles in pillar data/
10:29 flowstate joined #salt
10:32 Ayo joined #salt
10:35 irctc667 joined #salt
10:36 irctc667 Hi, is there anyway to get current unix timestamp in state files? without using extra dependencies
10:38 felskrone is anyone aware of an approach where i can have beta-minions for specific changes in modules? for example i only want minions 1-5 to be beta-minions for changeset-A, minions 5-10 beta minions for changeset-B, etc. and have that configured in pillar somehow? i have an idea in my head, but i wonder how others (would) do it
10:38 saltuser joined #salt
10:38 * buxy is looking for saltstack-formula admin to add https://github.com/rhertzog/letsencrypt-sh-formula to the official repository
10:40 saltuser Hi! I'm looking at users-formula and there are get() and append() functions. Where can I see docs what functions more are available?
10:40 saltuser jinja2 docs do not seem to include those
10:41 fannet_ joined #salt
10:41 saltuser For example {%- do used_user_files.append(1) %} - I can't find anything on jinja2 site about that function. Google is also kind of quiet about this
10:42 buxy forget this, I can fork my own repository into the saltstack-formula org
10:42 AndreasLutro saltuser: used_user_files is a python list, look up methods for python's list type
10:43 Ayo_ joined #salt
10:43 saltuser AndreasLutro: So it just calls python string functions?
10:43 AndreasLutro string?
10:43 saltuser Correction - python functions
10:44 AndreasLutro technically python object methods
10:44 AndreasLutro you can't call pure python functions in jinja
10:44 saltuser Mkay, it seems I have a lot to learn :)
10:44 saltuser But at least now I know where to look those docs :)
10:44 saltuser Thanks
10:47 kbaikov joined #salt
10:48 keimlink joined #salt
10:48 Rumbles joined #salt
10:52 kshlm joined #salt
10:53 edrocks joined #salt
10:54 flyboy joined #salt
10:57 flyboy82 Hi! Anyone knows of a way to utilize salt-mine data in a pillar? need to target specific minions and use their mine data to populate a pillar
10:57 babilen flyboy82: The way to do that is documented in mine.get
10:58 babilen flyboy82: Found it?
10:59 flyboy82 the {% set .....  %} substitute? yes, thanks.
11:00 babilen great
11:00 babilen You can call the runner directly
11:02 mikecmpbll joined #salt
11:03 neogenix joined #salt
11:16 thehaven joined #salt
11:17 amcorreia joined #salt
11:21 dmaiocchi joined #salt
11:22 kbaikov joined #salt
11:28 liskl joined #salt
11:28 flowstate joined #salt
11:31 dmaiocchi joined #salt
11:32 krymzon_ joined #salt
11:37 lompik joined #salt
11:42 fannet_ joined #salt
11:49 kbaikov joined #salt
11:51 west575 joined #salt
11:52 kbaikov joined #salt
11:53 dmaiocchi joined #salt
11:55 tpaul joined #salt
11:58 tuxick i need to rename a minion, is there a standard command/procedure for it?
12:01 babilen Remove old key and add new one
12:02 haaj joined #salt
12:03 tuxick hmm that's the bit i was wondering about yes
12:03 tuxick so renaming the key in pki/ is not enough?
12:03 haaj joined #salt
12:05 tuxick hm
12:07 tuxick to minimize effort i just gave it a shot the easiest way: changed on minion: hostname and minion_id, restarted salt-minion and on master renamed the key
12:07 tuxick seems to be enough
12:09 amcorreia joined #salt
12:10 abednarik joined #salt
12:11 flyboy82 ok, tougher one, is there a known way to declare in a jinja substitute a key name from a dict with a dot in its name? (eg. data.ip_interfaces.eth1.5[0]    where eth1.5 is the name of the vlan interface as it is gotten from grains)
12:13 rothsa joined #salt
12:16 dmaiocchi joined #salt
12:16 hlub joined #salt
12:22 AndreasLutro ip_interfaces['eth1.5'][0]
12:23 flyboy82 thanks, works like a charm!
12:24 deus_ex joined #salt
12:25 scoates joined #salt
12:29 flowstate joined #salt
12:31 anmol joined #salt
12:42 deus_ex joined #salt
12:46 edrocks joined #salt
12:47 netcho joined #salt
12:47 om2 joined #salt
12:48 dmaiocchi joined #salt
12:50 flowstate joined #salt
12:50 dmaiocchi_ joined #salt
12:50 gh34 joined #salt
12:54 rem5 joined #salt
13:01 edrocks joined #salt
13:03 ronnix joined #salt
13:03 duncanmv_ joined #salt
13:07 DEger joined #salt
13:08 duncanmv_ joined #salt
13:11 flowstate joined #salt
13:17 tuxick master log telling me to read master log for details?
13:18 manji yes, you get used to it after a while
13:18 komputes_ joined #salt
13:19 tuxick really?
13:19 kbaikov joined #salt
13:19 numkem joined #salt
13:19 * tuxick goes into assburger mode
13:22 tuxick thank bob log on the minion was slightly more informative
13:22 manji yep
13:24 tuxick meh, in pillar trying to set up mysql credentials, host: % or host: '%' both causing trouble
13:25 tuxick found character '%' that cannot start any token
13:26 tuxick leaving that for later
13:26 mapu joined #salt
13:26 manji ah
13:27 manji hangon
13:27 manji I did that recently
13:27 dfinn joined #salt
13:28 manji ah no I am sorry, I had an issue with *.* and db privileges
13:28 flowstate joined #salt
13:28 manji jinja didn't like it much
13:28 tuxick no worries, in fact i didn't need/want % anyway
13:28 tuxick just wanted to have something to start with
13:28 manji % is bad anyway :)
13:28 tuxick ack
13:28 evle joined #salt
13:28 AndreasLutro where do you set "host: '%'"? in your pillar? minion config?
13:29 AndreasLutro you might as well fix it now to learn about yaml/jinja rendering gotchas
13:29 tuxick ok, started friday, got as far setting up mysql accounts + grants, install postfix/mysql/dovecot etc
13:29 manji AndreasLutro, it is bad to use % either way, don't encourage :p
13:29 tuxick useraccounts of course
13:29 mpanetta joined #salt
13:30 tuxick AndreasLutro: well i wanted minimum amount of data in pillar, but host is part of that if you want to iterate through the mysql accounts
13:30 tuxick at least it makes it a bit easier
13:30 AndreasLutro let me just illustrate actually
13:32 AndreasLutro https://bpaste.net/show/386da7ab46ed tuxick
13:32 AndreasLutro best way to work around stuff like this is do pipe all your variables through the json filter (json is always valid yaml) -- {{ my_var | json }}
13:36 om2 joined #salt
13:37 _mel_ joined #salt
13:37 tuxick ah ok, will remember
13:41 lord2y joined #salt
13:42 lord2y hello boys
13:42 amcorreia joined #salt
13:42 lord2y I have a question
13:42 lord2y is it possibile to use jinja2 inside yaml sls?
13:42 lord2y I would like to skip the creation of one file
13:43 lord2y with
13:43 lord2y {% set envar=pillar['variable'] %}
13:43 lord2y {% if envar is equalto "goofy" %}
13:43 lord2y file: bla bla
13:44 lord2y {% endif %}
13:44 AndreasLutro sls are jinja2 + yaml by default
13:44 AndreasLutro I thought this was expressed in the documentation
13:44 lord2y it is
13:44 lord2y but it doesn't work
13:45 lord2y or I'm wrong
13:45 AndreasLutro it probably doesn't work for other reasons, share your full error message and sls file
13:45 jartsu joined #salt
13:46 lord2y {% set envvar = pillar['env'] %}
13:46 lord2y {% if envvar is even dw2 %}
13:46 lord2y ifcfg-team0.2003:
13:46 lord2y file.managed:
13:46 lord2y - name: /tmp/ifcfg-team0.2003
13:46 lord2y - source: salt://networks/ifcfg-team0.2003
13:46 lord2y - user: root
13:46 lord2y - group: root
13:46 lord2y - mode: 644
13:46 lord2y {% endif %}
13:46 AndreasLutro use a pastebin/gist.github.com next time please
13:47 AndreasLutro {% if envvar is even dw2 %} I don't think this is valid jinja
13:47 mpanetta shouldn't even be in ?
13:47 jartsu Hi, has anyone setup salt with git integration in the manner described here, https://clinta.github.io/salt-git-nogitfs/? I'm wondering if that's a good idea. From what I've read it sounds sensible, but I haven't seen any comments or very many references to that post or that way of setting things up.
13:47 mpanetta and quotes...
13:47 abednarik joined #salt
13:48 lord2y AndreasLutro: is my jinja syntax wrong?
13:48 lord2y anyway the error is this: TypeError: test_even() takes exactly 1 argument (2 given)
13:48 haaj joined #salt
13:48 AndreasLutro yes it's wrong
13:48 lord2y I guess the syntax is right
13:48 lord2y :-O
13:49 lord2y ok, I'm a newby of jinja
13:49 AndreasLutro just look up "is even" in the jinja docs
13:49 AndreasLutro I don't think it's what you want
13:49 AndreasLutro but I don't know what you want
13:49 lord2y I want this
13:50 lord2y if (envar == "dw2")
13:50 tuxick jinja-fu master!
13:50 AndreasLutro then why don't you just do that?
13:50 lord2y I did
13:50 AndreasLutro how did you even think to use "is even"
13:50 lord2y but it didn't work
13:50 lord2y sorry, wrong past
13:51 lord2y {% if envvar is equalto dw2 %}
13:51 lord2y this is the right one
13:51 lord2y but problem persist
13:51 lord2y and also doing {% if envvar == "dw2" %}
13:51 lord2y I get an error
13:51 mpanetta if dw2 isn't a variable it needs to be quoted
13:52 mpanetta try if 'dw2' is in envvar maybe
13:52 mpanetta But I think that only works for dicts.
13:52 lord2y mpanetta: let me try
13:53 mpanetta not sure why it would not take your if envvar == "dw2" tho...
13:53 AndreasLutro I think you should read and share your error messages instead of just trying random stuff :)
13:54 mpanetta probably would help haha
13:55 lord2y with dw2 is in envar the error is this Rendering SLS 'base:common.networks' failed: Jinja syntax error: no test named 'in'; line 40
13:55 av_ joined #salt
13:55 AndreasLutro just do {% if envvar == 'dw2' %}
13:55 AndreasLutro if that's what you want to do
13:56 impi joined #salt
13:56 lord2y with envar is equalto 'dw2' I get Rendering SLS 'base:common.networks' failed: Jinja syntax error: no test named 'equalto'; line 40
13:56 lord2y let me (re)try {% if envvar == 'dw2' %}
13:57 lord2y in this way
13:57 lord2y I don't get an error
13:57 lord2y but it doesn't work as expected
13:57 lord2y no wait
13:57 ajv joined #salt
13:58 mpanetta hrm
13:58 lord2y I confirmed
13:58 lord2y it doesn't work
13:58 lord2y file is there
13:59 lord2y even if pillar defined is different
13:59 lord2y let me check the pillar again
14:00 emaninpa joined #salt
14:00 lord2y uhm
14:01 lord2y it seems that the problem is elsewhere
14:02 lord2y I used this command to refresh the pillar
14:02 lord2y salt '*' saltutil.refresh_pillar
14:02 lord2y but they don't change :-(
14:03 mpanetta It would really help to see the error
14:03 dendazen joined #salt
14:03 lord2y Now I don't have any error
14:04 lord2y -------------
14:04 lord2y Succeeded: 30
14:04 lord2y Failed:     0
14:04 lord2y -------------
14:04 lord2y Total states run:     30
14:04 lord2y all is in the correct state
14:06 kawa2014 joined #salt
14:07 lord2y ok
14:07 lord2y I understand
14:07 lord2y wrond definition of the pillar
14:07 lord2y I applied the variable to all minions
14:07 lord2y what I cannot understand is why before the == didn't work
14:07 lord2y anyway thank you for help
14:11 keltim joined #salt
14:15 dyasny joined #salt
14:16 om2 joined #salt
14:17 anmol joined #salt
14:18 netcho can salt-cloud be used via salt-api
14:18 hasues joined #salt
14:18 netcho i.e. i wanna run salt-cloud command from jenkins
14:22 om joined #salt
14:22 om2 joined #salt
14:23 tuxick maybe painted myself in a corner: i defined some accounts/credentials in pillar, iterating through them to create the accounts
14:23 tuxick but elsewhere i need the password of one particular account only
14:24 om2 joined #salt
14:24 AndreasLutro change the list to a dict where the username is the key
14:24 tuxick and then?
14:24 AndreasLutro change your iteration to use python's dict items() method
14:26 tuxick and how to extract pass for this particular account then?
14:27 M-liberdiko joined #salt
14:27 AndreasLutro {{ pillar.foo.bar.baz }}
14:28 AndreasLutro or salt['pillar.get')('foo:bar:baz')
14:28 tuxick that's the syntax i don't seem to understand
14:28 salty_ joined #salt
14:28 AndreasLutro which one? pillar.foo.bar.baz?
14:28 dmaiocchi joined #salt
14:29 tuxick yes
14:29 flowstate joined #salt
14:30 tuxick it'd expect something like (user,'foo') or whatever
14:30 tuxick where foo is the name of the user i'm looking for
14:30 AndreasLutro not sure what you mean by (user, '
14:30 tuxick but then i need even more to indicate i need password from that entry
14:30 AndreasLutro not sure what you mean by (user, 'foo')
14:30 AndreasLutro but ok a more realistic example..
14:30 tuxick well what is foo, bar and baz??
14:30 AndreasLutro {{ pillar.users.some_username.password }}
14:31 tuxick right
14:31 AndreasLutro they're parts of your pillar structure
14:31 AndreasLutro do you know any programming languages?
14:31 tuxick ok, that'd been a bit cleared if 'some_username' were quoted or something
14:31 tuxick yes, except python and ruby :)
14:32 tuxick even done perl
14:32 * tuxick spits
14:32 hasues left #salt
14:32 AndreasLutro except? so you've done everything except python and ruby?
14:32 tuxick well basic, pascal, java, c, c++ perl in that order
14:32 tuxick not counting php :)
14:32 hlub joined #salt
14:33 AndreasLutro right, so the "pillar" variable is a hashmap (python dict), jinja lets you access keys of the hashmap with .
14:33 AndreasLutro my_hashmap.my_key
14:33 AndreasLutro alternatively you can access it like you would in python, my_hashmap['my_key']
14:33 tuxick right, got it
14:34 tuxick mind, started with saltstack last friday and took weekend off :)
14:36 duncanmv_ joined #salt
14:39 corichar joined #salt
14:40 netcho anyone used jenksins salt plugin?
14:40 _JZ_ joined #salt
14:41 duncanmv_ joined #salt
14:42 salty_ Hey, I have an issue with the minion installation using reactors in aws- sometimes the salt doesn't connect to the minion using ssh. most of the times its going without and issues but sometimes it shows the error" Failed to create VM i-c6***. Configuration value 'id' needs to be set"
14:47 flowstate joined #salt
14:48 sjmh joined #salt
14:48 M-MadsRC joined #salt
14:48 netcho can someon explain me salt modules for cloud
14:48 netcho i a m a bit confused
14:49 netcho salt myminion cloud.list_sizes my-cloud-provider
14:49 netcho why do i need to target a minion i'f i am creating a new one?
14:49 netcho or do i have to target one created with salt previpusly
14:53 rylnd joined #salt
14:55 thorrrr joined #salt
14:55 neogenix joined #salt
15:01 mdhas joined #salt
15:01 DEger joined #salt
15:02 housemouse139 joined #salt
15:03 housemouse139 Hello
15:04 numkem joined #salt
15:04 pmcg joined #salt
15:07 Brew joined #salt
15:07 irctc318 joined #salt
15:07 irctc318 hey guys, any idea why I don't have the "consul" module even though I'm on Salt 2016.3.1?
15:08 irctc318 https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.consul.html
15:11 irctc318 ah it is only meant to be used via "salt" command and not "salt-run"
15:11 om2 joined #salt
15:12 housemouse139 Anybody know if it's possible to set any kind of no-delete option in salt-cloud to avoid accidentally deleting a VM?
15:14 housemouse139 Wouldn't even be against hacking some python files to break the delete functionality.
15:14 mTeK joined #salt
15:14 jfindlay irctc318: salt modules won't load if their dependencies aren't installed
15:16 jfindlay netcho: execution and state modules are designed to be run by minions, which is also true of the cloud exec and state modules
15:17 jfindlay if you want to bypass the need for minion interaction for salt cloud, you can use the salt-cloud on the master directly
15:17 jfindlay but if you want to use the execution/state system, you can consider setting up a special cloud minion to do this work, possibly on the master itself, for example
15:19 jfindlay housemouse139: that sounds like a good idea.  I don't know of such a feature (doesn't mean it does not exist), but you may want to ask on the mailing list or even submit a feature request
15:21 netcho jfindlay:  sorry i dont understand
15:22 jfindlay netcho: you must target a minion if you are going to use the salt-cloud execution module
15:22 jfindlay if you don't want to target a minion, you may use the salt-cloud command directly
15:23 jfindlay netcho: `salt-cloud --list-sizes my-cloud-provider`
15:23 ry joined #salt
15:25 netcho oh i see
15:25 netcho iu was trying to run it from jenkins plugin
15:25 netcho i know how to do it with salt-cloud
15:26 netcho but jenkins plugin does not have that option
15:26 jfindlay netcho: you can run a minion daemon on your salt master and have jenkins target that
15:26 netcho true
15:28 mTeK joined #salt
15:32 kshlm joined #salt
15:36 zero_shane joined #salt
15:37 jeddi joined #salt
15:38 Rumbles joined #salt
15:38 repl1cant so w/tcp transport do you just set the transport: to tcp? Same ports as zeromq?
15:42 kshlm joined #salt
15:43 om2 joined #salt
15:43 edrocks joined #salt
15:48 neogenix joined #salt
15:51 Trauma joined #salt
15:54 lord2y joined #salt
15:59 bfig_ joined #salt
16:01 SpX joined #salt
16:02 mikecmpbll joined #salt
16:05 om2 joined #salt
16:09 kus joined #salt
16:13 __lord2y question: how may I instruct the minions to get the last state automagically?
16:15 Eugene https://docs.saltstack.com/en/latest/topics/jobs/#scheduling-jobs is the docs page on scheduling
16:15 iggy __lord2y: the minions pull states at the beginning of every highstate
16:16 Eugene Personally, I don't like to run states automagically. I use a Nagios plugin that calls highstate test=True
16:16 arnaud hello, is there a command taking an minion name as an argument an return back every nodegroup that match ? (for debug)
16:16 Eugene Then my normal alerting system tells me to go see what changed
16:16 iggy (if that doesn't answer your question, maybe rephrase it, because it's a little unclear what you're trying to do)
16:16 iggy arnaud: not that I've ever seen
16:16 arnaud it would be so useful!
16:17 arnaud the code must exists somewhere in the source I guess..
16:17 __lord2y iggy: sorry if it was unclear
16:17 __lord2y iggy: anyway you answer
16:17 __lord2y iggy: what is still unclear for me is the meaning of the highstate
16:18 babilen __lord2y: Highstate essentially means: Apply all states that have been targeted to the minion in the top file
16:19 iggy arnaud: I doubt it... that's not code that would need to exist to function properly
16:20 iggy nodegroups also aren't as popular as other targetting methods due to the whole having to restart the master to use them
16:20 haaj joined #salt
16:22 __lord2y babilen: so everytime I do a change in the top.sls file, every targeted minions start to get the new state?
16:24 toastedpenguin joined #salt
16:26 abednarik joined #salt
16:26 om joined #salt
16:27 alvinstarr joined #salt
16:28 flowstate joined #salt
16:32 woodtablet joined #salt
16:35 writtenoff joined #salt
16:37 Ahlee i just found 75 states that differ by a single line
16:37 zero_shane joined #salt
16:37 Ahlee short of doing a massive list and iterating over it, how would one reduce the amount of repetition?
16:37 Jogfat joined #salt
16:38 Ahlee that's about as clear as mud, ain't it
16:38 mpanetta_ joined #salt
16:42 ageorgop joined #salt
16:43 arnaud iggy: when salt master has to determine which group a minion belongs to, there should be such a code to do that
16:43 arnaud f(minion_name) = list_nodegroup
16:44 arnaud or maybe not..
16:44 arnaud just something like f(minion_name,nodegroup_name) = true|false
16:44 arnaud would be enough for master
16:45 arnaud i'm doing an audit and I will save a lot of time with such a tool
16:46 Ahlee why not just grep it out of the master config?
16:46 beardedeagle joined #salt
16:46 Ahlee since nodegroups are only defined in the master config file, it's the definitive source of the truth for node membership in the nodegroup
16:47 Ahlee and as iggy said, nodegroups are very seldom used due to needing to restart the master to modify membership, must easier to use grains or pillars
16:47 arnaud yeah but nodegroup definition are tricky
16:48 Ahlee really? grep <minion_name> /etc/saltmaster | awk '{print $1}' | sed -e s/://
16:48 Ahlee oh, i suppose if you defined them as a list
16:48 kshlm joined #salt
16:48 arnaud and regex
16:48 arnaud ^^
16:48 Ahlee and I just noticed you can do regex, he
16:48 Ahlee h
16:48 arnaud its not that easy
16:48 neogenix joined #salt
16:49 arnaud im trying to list group and members
16:49 arnaud thats all
16:49 Ahlee salt -N <nodegroupname> test.ping
16:49 rem5 joined #salt
16:49 arnaud thats a solution yes
16:49 stewart_ joined #salt
16:50 arnaud and loop on all group
16:50 Ahlee i'm assuming that isn't going to work for what you wanted, though
16:51 Ahlee there's probably something you could do with the mine
16:51 arnaud its a solution and its working but its not fast if i want to update
16:51 arnaud but hey I will do it like that
16:51 arnaud its just an audit
16:51 toastedpenguin joined #salt
16:51 Ahlee i don't get what you want.
16:52 Ahlee You want a way to target all your node groups, that isn't targeting all your node groups?
16:52 linjan_ joined #salt
16:52 arnaud I just want to output for each group, the member list
16:52 Ahlee there's probably a way to do it with a _runner, have the master spit back node membership
16:52 arnaud thats it
16:54 Ahlee arnaud: willing to dive into some python? I bet with import salt.runner from the master you could pull out the defined nodegroups and iterate over them quite quickly
16:55 Ahlee You've got my interest, one moment
16:56 arnaud thats where it becomes serious :D
16:56 slav0nic joined #salt
16:57 Ahlee gotta add a nodegroup to test
16:58 arnaud my python level is average, my knowlegde about salt source code is null
16:58 rem5 joined #salt
16:58 arnaud so it will takes time
16:58 kshlm joined #salt
16:58 arnaud maybe loop on all nodegroup name (grep minion.d/nodes.conf|sed ....) and do a ping will be faster for me
16:59 amcorreia joined #salt
17:00 aharvey joined #salt
17:03 kshlm joined #salt
17:03 patrickancillott joined #salt
17:06 Ahlee arnaud: https://gist.github.com/jalons/650f9686e855221751a38f03c3d0b302
17:06 mpanetta joined #salt
17:07 Ahlee not actually sure if that'll work or not
17:07 Ahlee well, it works with my simple group1: 'G@kernel:Linux'
17:08 Ahlee but if you have a list in there you'll probably need to isinstance(_target, list): and handle that different
17:08 Ahlee or something
17:09 arnaud hey thx a lot Ahlee
17:10 arnaud I'm testing it
17:10 arnaud so you do a test.ping
17:11 Ahlee yeah, gotta do something to target them
17:11 Ahlee then ret is a dictionary that returns group:[list of hosts that matched the expr form defined in the nodegroup]}
17:11 Ahlee happy extending
17:12 kevinquinnyo with file.replace is there a way to replace the line with literally nothing, equivalent of deleting the line?
17:12 kevinquinnyo or is there another function specifically for that?
17:13 arnaud thx again Ahlee :)
17:13 Ahlee kevinquinnyo: does repl: '' not get you what you want?
17:14 babilen kevinquinnyo: You can replace it with the empty string which removes it if you match the entire line and linebreak. Or use file.line with delete
17:14 abednarik joined #salt
17:17 kevinquinnyo babilen: thanks, the linebreak did the trick.. checking file.line though
17:17 patrickancillott joined #salt
17:18 impi joined #salt
17:18 kevinquinnyo babilen: why does the line function require the content paramater if it also allows a "match" parameter
17:20 teryx510 joined #salt
17:20 babilen kevinquinnyo: I have no idea why we need a line state to begin with
17:21 scarcry joined #salt
17:23 manji joined #salt
17:23 netcho joined #salt
17:26 keimlink joined #salt
17:26 patrickancillott joined #salt
17:28 TooLmaN joined #salt
17:42 cyborg-one joined #salt
17:43 toastedpenguin joined #salt
17:45 akio left #salt
17:51 ajw0100 joined #salt
17:51 ixeous joined #salt
17:54 fracklen joined #salt
17:56 s_kunk joined #salt
17:56 s_kunk joined #salt
18:00 ixeous I have created a state file for a service to run on a non-standard port.  The pkg install and service configuration work.  I have a cmd.run for semanage so that SELinux will allow the service on the non-standard port.  That "works", but whenever highstate is run after the initial setup, there is an error from the cmd.run semanage because the port is already defined.  What is the best way to fix that?
18:01 ixeous salt.states.selinux doesn't seem to have anything related to this.  Is a bash conditional the best option?
18:08 Ahlee That's how I would do it
18:09 mikecmpbll joined #salt
18:10 patricka1 joined #salt
18:12 aharvey joined #salt
18:12 Ayo joined #salt
18:15 netcho joined #salt
18:23 root____5 joined #salt
18:25 catpig joined #salt
18:25 Joe630 pywinrm is making me craaaazzzzyyyyy
18:26 Joe630 Is anyone using it?  SLL is totally broken.
18:26 Joe630 SS*
18:26 Joe630 SSL.
18:27 beardedeagle have used it, didn't have any issues with it when I did.
18:27 beardedeagle now I use winexe
18:27 om joined #salt
18:27 Joe630 requests bombs out with SSL errors, and there is nowhere for me to tell winrm to ignore them without patching deploy.py
18:27 Joe630 winexe is deprecated.
18:28 Joe630 2012r2 kills it.
18:28 beardedeagle I use it with 2012r2 exclusively so don't know why you think that
18:28 Joe630 buhhhhhhh what
18:28 flowstate joined #salt
18:29 beardedeagle I am about to put a PR into salt-pack with a updated winrm that works with 2012r2 so I guess look for it in the saltstack repo sometime in the near future. No idea how long it takes dmurphy to get those pushed out.
18:29 beardedeagle but it is only for centos
18:29 dmaiocchi joined #salt
18:30 Joe630 Now I need to figure out why our winexe is broken.  I was told by the other admin on this project it was deprecated.
18:30 beardedeagle what os is your master
18:30 Joe630 centos 7
18:30 Ahlee Joe630: 2012R2?
18:30 Joe630 Ahlee: yes
18:30 beardedeagle ok give me a moment
18:30 edrocks joined #salt
18:30 beardedeagle @Joe630: https://github.com/beardedeagle/winexe-rpm
18:31 beardedeagle that is mine
18:31 beardedeagle it will let you build a working winexe rpm
18:31 beardedeagle for centos 7 that works with win2012r2
18:31 anmol joined #salt
18:31 beardedeagle until I get the PR in that is the best I can do for you
18:31 Ahlee They deprecated the 1.0 version of their protocol, you have to use the new one
18:31 beardedeagle mine uses winexe 1.1
18:31 Ahlee i'm drawing a blank on what exactly
18:31 Ahlee it's on the windows side, was it smb1.0?
18:31 beardedeagle so yes ahlee you are correct
18:31 beardedeagle yes
18:32 beardedeagle you are correct again
18:32 Ahlee nice on the spec files beardedeagle
18:32 beardedeagle instructions on how to use for those who have never built a package before are in that repo too
18:32 Ahlee I just stopped using winexe in favor of, um.
18:34 Ahlee ah, i went with /usr/bin/net
18:34 Ahlee net rpc service
18:35 Ahlee and then added a decorator to each function that retries it three times
18:35 sagerdearia joined #salt
18:36 Joe630 thanks beardedeagle
18:36 beardedeagle npnp
18:37 flowstate joined #salt
18:39 perfectsine joined #salt
18:40 kevinquinnyo how can i call getattr() to get a function from a salt module
18:40 kevinquinnyo from another module
18:40 ronnix joined #salt
18:45 edrocks joined #salt
18:48 radhac joined #salt
18:51 dmaiocchi joined #salt
18:52 deniszh joined #salt
18:58 felskrone joined #salt
18:58 radhac Anyone familiar with ps beacon?  Following the guide I receive this error in salt-minion's log:  "Could not LazyLoad ps.beacon"
18:59 radhac Does this require some kind of dependency for cent7 to work correctly/
19:00 Ahlee https://github.com/saltstack/salt/blob/develop/salt/beacons/ps.py looks pretty simple, is your logging on debug?
19:00 Ahlee by my guess would be no psutil
19:01 radhac it's on debug.  psutils seems to be the culprit
19:02 heliocat joined #salt
19:03 DEger joined #salt
19:07 mikecmpbll joined #salt
19:10 fracklen joined #salt
19:11 radhac @Ahlee, python-psutil was it, just verified.  Thanks for your help
19:12 zenlot joined #salt
19:12 jfindlay looks like we still need to message those dependencies better
19:14 Ahlee get crackin' jfindlay
19:14 ajw0100 joined #salt
19:15 beardedeagle +1 for clearer all the things
19:17 jfindlay they should be returning to the command line, at least for modules that actually return a reason
19:18 iggy arnaud: but salt doesn't ever need to know that... it needs to know what minions are in a nodegroup... not what nodegroups a minion is in
19:19 haaj joined #salt
19:20 Ahlee I give up. This mess is going to stay a mess.
19:21 ageorgop joined #salt
19:25 harkx joined #salt
19:26 iggy logging on dependencies is a slippery slope
19:27 cyborg-one joined #salt
19:27 permalac_ joined #salt
19:27 ALLmightySPIFF joined #salt
19:27 iggy I personally get super annoyed when I see messages about vbox stuff 8 times every time I run a salt-cloud command
19:27 fannet__ joined #salt
19:28 iggy but I also understand it's usefullness at times :/
19:28 iggy if there was an agreed upon log level for that kind of stuff, that would be great
19:28 mpanetta_ joined #salt
19:28 beardedeagle you mean people picking a standard and sticking to it?
19:29 iggy then I could set mine slightly lower as to not want to kick puppies every time I see them
19:29 beardedeagle what is this sorcery?
19:29 iggy ikr
19:29 debian112 joined #salt
19:29 harkx joined #salt
19:31 AirOnSkin joined #salt
19:31 teryx510 joined #salt
19:31 esharpmajor joined #salt
19:31 charli joined #salt
19:32 cetex joined #salt
19:32 pfallenop joined #salt
19:32 eichiro joined #salt
19:32 babilen joined #salt
19:32 ramblinpeck joined #salt
19:32 hillna joined #salt
19:33 rem5 joined #salt
19:34 edrocks joined #salt
19:34 buxy joined #salt
19:35 garphy joined #salt
19:35 sjohnsen joined #salt
19:35 arif-ali joined #salt
19:36 Satyajit joined #salt
19:37 cro joined #salt
19:38 ajw0100 joined #salt
19:38 eseyman joined #salt
19:38 AbyssOne joined #salt
19:40 deniszh joined #salt
19:41 edrocks joined #salt
19:43 hoonetorg joined #salt
19:43 shalkie joined #salt
19:43 mavhq joined #salt
19:46 DammitJim joined #salt
19:47 tpaul joined #salt
19:48 aharvey salt-cloud : Nova driver :  Even when the provider.d/* is defined hard for one tenant, when I call salt-cloud to build a host, massive amounts of api calls to all available projects, flavors, servers, images are being collated every single time.  Is this expected behavior?
19:49 aharvey If this were kept in a cache I’d be fine with it but we’re talking 13+ minutes for one single small server to be kicked off vs. openstack provider/driver which is about 45 seconds
19:55 iggy aharvey: there's a ticket open for caching some of that stuff iirc
19:55 bltmiller joined #salt
19:55 iggy (may not be for all providers though)
19:56 aharvey hmmm
19:57 radhac Another beacon/reactor question...how would I have reactor look for a specific tag + event.  For instance, tag = "salt/beacon/*/ps/".  event id = * , $process running.  How would I be able to filter on this level of granularity on the reactor side?
20:02 Barbarossa left #salt
20:03 netcho joined #salt
20:04 ageorgop joined #salt
20:06 garphy joined #salt
20:08 rem5 joined #salt
20:09 irctc166 joined #salt
20:10 irctc166 hey guys, any ideas why half of the times when I restart the salt master only some of his workers are successfully restarted? some are left running and this causes things to break obviously
20:12 irctc166 the reason I'm restarting most of the time is due to new runners, modules, configuration. So maybe there's a better way for me to reload those?
20:12 jhauser joined #salt
20:12 babilen irctc166: https://github.com/saltstack/salt/pull/33792
20:12 saltstackbot [#33792][MERGED] Do not use KillMode=process for salt-master | What does this PR do?...
20:17 keimlink joined #salt
20:17 irctc166 babilen: it looks like that PR made it into 2016.3.1, yet that's the version I'm using as well
20:18 irctc166 or maybe it didn't
20:19 Ayo joined #salt
20:20 babilen Check please .. I'm very interested in it
20:20 jfindlay irctc166: that got merged into develop rather than 2016.3
20:20 Brew joined #salt
20:21 irctc166 Also I'm using upstart rather than systemd
20:22 jfindlay but, salt-pack has the final determination on what init script gets packaged and installed, so it may not matter
20:24 irctc510 joined #salt
20:26 patrickancillott joined #salt
20:32 patrek joined #salt
20:42 netcho joined #salt
20:49 keimlink joined #salt
20:50 rem5 joined #salt
20:51 netcho joined #salt
20:53 edrocks joined #salt
20:54 ageorgop joined #salt
21:04 lero joined #salt
21:09 manifold joined #salt
21:14 bltmiller joined #salt
21:15 netcho joined #salt
21:18 edrocks joined #salt
21:21 netcho joined #salt
21:22 aharvey joined #salt
21:23 cyborg-one joined #salt
21:24 deniszh joined #salt
21:24 cableninja_ joined #salt
21:29 flowstate joined #salt
21:31 tristianc joined #salt
21:34 cprior joined #salt
21:34 aharvey iggy: do you have a link to the ticket?  I’ve searched around to not much luck
21:35 aharvey I ended up reverting back to driver:openstack for the time being, would like to help provide feedback
21:36 jfindlay aharvey: gtmanfred has done some work on the nova driver recently
21:42 impi joined #salt
21:49 kevinquinnyo is there a way to extract a line from a file with the file module, or another one? or should i just 'shell out' to grep via the cmd module for that
21:51 DEger joined #salt
21:52 jfindlay kevinquinnyo: are you trying to get info out of the line?
21:52 kevinquinnyo yes i was to extract a section of a line based on a regular expression and parse it further from there
21:52 rem5 joined #salt
21:52 kevinquinnyo this is inside a custom salt module i'm working on
21:52 kevinquinnyo s/was/want
21:53 jfindlay there's `file.grep`
21:53 kevinquinnyo ohhh i missed that
21:53 kevinquinnyo thanks
21:53 jfindlay sure
21:53 DEger joined #salt
21:53 jfindlay it's an exec mod function, so you'd have to boilerplate wrap it to make it stateful if necessary
21:54 kevinquinnyo it looks like it just runs cmd.runall(grep ...) anyway
21:54 west575 joined #salt
21:55 jfindlay right
21:59 Favoretti jfindlay: hi there. I've been talking to colelagues here and after one of our PRs got merged today, we kinda noticed that the unit-testing on github is completely b0rked :) is that something that you guys are planning to address? :) Doesn't look too assuring when 4/5 test suites are failing :)
21:59 aharvey joined #salt
22:00 onlyanegg joined #salt
22:01 cableninja_ joined #salt
22:01 jfindlay Favoretti: there is a known test failure that has been fixed already, but takes a while to cycle out because pull requests are not always rebased on the freshest state of the branch
22:02 jfindlay also still needs a merge forward from what I last remember
22:03 Favoretti jfindlay: looking at the results of https://github.com/saltstack/salt/pull/34127 it looks more like an environmental issues to me tbh, or are we talking about the same?
22:03 saltstackbot [#34127][MERGED] Add timeout to ddns execution module and runner | What does this PR do?...
22:03 jfindlay Favoretti: is this the one: https://github.com/saltstack/salt/pull/34141?
22:03 saltstackbot [#34141][MERGED] Fixed boto_vpc_test failure | What does this PR do?...
22:03 Favoretti nope, the one I just linked :)
22:04 jfindlay yes, I have pasted the pull request that fixes one of those failures
22:05 Favoretti ah, cool. So generally the unit-testing should work, right?
22:05 jfindlay yes
22:05 jfindlay not sure about the other.  I am looking into it
22:05 Favoretti Thanks! :)
22:07 jfindlay the tests that are run against pull requests are quite extensive and can be brought down by the most ephemeral heterogeneous combinations of circumstances.  This is a continual work in progress for the automation and testing engineers who are trying to wrangle code from a dozen or more pull requests a day, but I do think we are making progress
22:09 Favoretti jfindlay: Understandable. There was a claim made today in a closed setting that unit-testing is generally broken and I wandered off to talk to you guys to see whether that is really the case. My personal view is that it can't be, otherwise each salt release would have a ton of breakage, whilst they actually don't ;)
22:10 Favoretti to give a bit of context we're in a semi-heated debate on whether we should continue our move from puppet to salt ;)
22:11 jfindlay there are a few known tests that we are working on resolving, but mainly the all of the tests pass on almost everything
22:11 jfindlay Favoretti: also, jenkins runs integration as well as unit tests
22:11 tmehlinger joined #salt
22:12 abednarik joined #salt
22:13 jfindlay we are working on revising that whole workflow to be more modular with containerized daemons for integration tests
22:13 Favoretti sounds great :)
22:13 jfindlay well, thanks for your pull request :)
22:13 Favoretti thanks for merging this quick :)
22:14 Favoretti caused us some headache with hanging jobs :)
22:19 mpanetta joined #salt
22:27 bltmiller joined #salt
22:28 flowstate joined #salt
22:29 haaj joined #salt
22:33 fannet_ joined #salt
22:46 manji joined #salt
22:51 Shirkdog joined #salt
22:51 Shirkdog joined #salt
22:56 amcorreia joined #salt
23:06 KajiMaster joined #salt
23:13 flowstate joined #salt
23:23 corichar joined #salt
23:49 ajw0100 joined #salt
23:53 abednarik joined #salt
23:55 edrocks joined #salt
23:57 bltmiller joined #salt

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