Perl 6 - the future is here, just unevenly distributed

IRC log for #salt, 2017-03-07

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

All times shown according to UTC.

Time Nick Message
00:01 PhilA_ joined #salt
00:08 sh123124213 joined #salt
00:39 DEger joined #salt
00:48 overyander joined #salt
00:52 dps joined #salt
00:55 raspado whats the best method to destroy a single VM instance with salt-cloud
00:55 hemebond raspado: --destroy?
00:55 raspado just --destroy <instance_name> ?
00:56 hemebond That's all I do.
00:56 filippos joined #salt
00:56 raspado ah k thx!
01:08 djgerm This jinja commenting (http://jinja.pocoo.org/docs/2.9/templates/#comments) should work in states right?
01:08 hemebond yes
01:23 edrocks joined #salt
01:30 DEger joined #salt
01:44 dyasny joined #salt
01:45 shoemonkey joined #salt
02:22 dankolbrs joined #salt
02:28 snarfy^ joined #salt
02:28 _JZ_ joined #salt
02:30 sp0097 joined #salt
02:33 systemexit joined #salt
02:33 snarfy^ joined #salt
02:47 ilbot3 joined #salt
02:47 Topic for #salt is now Welcome to #salt! <+> Latest Versions: 2016.3.5, 2016.11.3 <+> Support: https://www.saltstack.com/support/ <+> Logs: http://irclog.perlgeek.de/salt/ <+> Paste: https://gist.github.com/ <+> See also: #salt-devel, #salt-offtopic <+> Ask with patience as we are volunteers and may not have immediate answers <+> Meet the Maintainers Friday, March 10th 3:00 PM MST https://goo.gl/pLv8Fz
02:49 catpiggest joined #salt
02:54 dps joined #salt
02:59 DEger joined #salt
03:02 systemex1t joined #salt
03:06 Tanta joined #salt
03:17 evle joined #salt
03:20 bohr joined #salt
03:20 mdhas joined #salt
03:24 Praematura joined #salt
03:25 edrocks joined #salt
03:39 PatrolDoom joined #salt
03:57 mdhas joined #salt
03:59 puzzlingWeirdo joined #salt
04:00 ivanjaros joined #salt
04:05 mavhq joined #salt
04:09 leonkatz joined #salt
04:28 Praematura joined #salt
04:31 _JZ_ joined #salt
04:47 bohr joined #salt
04:48 DEger joined #salt
04:55 djgerm joined #salt
05:00 rdas joined #salt
05:07 impi joined #salt
05:09 mdhas joined #salt
05:23 preludedrew joined #salt
05:30 aarontc joined #salt
05:31 jagguli hi anyone using the azurearm cloud module ...
05:32 jagguli im getting an authentication error : AuthenticationError: , InvalidGrantError: (invalid_grant) AADSTS50034: To sign into this application the account must be added to the management.core.windows.net directory.
05:32 jagguli im following steps here https://docs.saltstack.com/en/latest/topics/cloud/azurearm.html
05:34 aarontc joined #salt
05:41 icebal_ joined #salt
05:45 SteamWells joined #salt
05:52 aarontc joined #salt
05:56 shoemonkey joined #salt
06:06 mavhq joined #salt
06:06 ronnix joined #salt
06:11 ProT-0-TypE joined #salt
06:14 bohr left #salt
06:15 bohr joined #salt
06:15 bohr left #salt
06:21 jimklo joined #salt
06:21 bocaneri joined #salt
06:24 aarontc joined #salt
06:25 jagguli ok got through that
06:27 edrocks joined #salt
06:30 jimklo joined #salt
06:38 jimklo joined #salt
06:51 angel_dark joined #salt
06:51 bocaneri joined #salt
06:52 bayindirh joined #salt
06:54 bayindirh Hello, I'm experiencing a problem with salt. I'm installing a set of local packages with "source" directive. Also I'm "hold"ing the packages & setting "update_holds" to true to keep the rule idempotent.
06:54 bayindirh However, salt reports an error after applying the state. The problem is error is false. Because in fact everything is installed and applied as it should be.
06:56 bayindirh I'm running 2016.11.3-1 on everything. Operating system is CentOS 7.3
06:57 DEger joined #salt
06:58 bayindirh Sorry, minions are on CentOS 6.8
07:00 jimklo joined #salt
07:09 darioleidi joined #salt
07:22 netcho_ joined #salt
07:38 duncanmv joined #salt
07:40 gmoro joined #salt
07:51 colttt joined #salt
07:56 armin_ joined #salt
07:56 jagguli has any one actually used the azurearm cloud module.. it seems so badly broken
07:57 shoemonkey joined #salt
07:57 nledez joined #salt
07:57 nledez joined #salt
07:58 colttt joined #salt
08:10 fracklen joined #salt
08:10 aldevar joined #salt
08:12 candyman88 joined #salt
08:13 rafaelcaricio joined #salt
08:16 colttt joined #salt
08:24 JohnnyRun joined #salt
08:32 Rumbles joined #salt
08:35 netcho_ joined #salt
08:40 mikecmpbll joined #salt
08:52 ronnix joined #salt
08:58 DEger joined #salt
09:01 geomacy joined #salt
09:08 colttt joined #salt
09:22 chowmein__ joined #salt
09:26 jacksontj joined #salt
09:29 edrocks joined #salt
09:29 netcho_ joined #salt
09:32 Mattch joined #salt
09:36 Rumbles joined #salt
09:36 rafaelcaricio joined #salt
09:37 jhauser joined #salt
09:37 bocaneri joined #salt
09:38 teclator joined #salt
09:40 GnuLxUsr joined #salt
09:41 Mattch joined #salt
09:43 Hybrid1 joined #salt
09:44 s_kunk joined #salt
09:47 cachedout joined #salt
09:52 impi joined #salt
09:58 shoemonkey joined #salt
10:21 citaret joined #salt
10:34 DanyC joined #salt
10:36 rafaelcaricio joined #salt
10:38 oms101_ joined #salt
10:50 o1e9 joined #salt
10:52 ivanjaros joined #salt
10:52 jas02 joined #salt
10:58 DEger joined #salt
11:09 N-Mi joined #salt
11:09 N-Mi joined #salt
11:19 vdt-mik joined #salt
11:19 vdt-mik hi
11:20 Xevian joined #salt
11:22 Alaniaris joined #salt
11:22 vdt-mik i want creat sls file for install openvpn client ib windows PC, but i don`t understand how i can run this file. Command salt 'minion' pkg.install /srv/salt/openvpn-client.sls
11:22 vdt-mik not working
11:23 LondonAppDev joined #salt
11:25 Norrland vdt-mik: https://docs.saltstack.com/en/latest/ref/modules/all/salt.modules.state.html#salt.modules.state.apply
11:26 Norrland https://docs.saltstack.com/en/latest/topics/tutorials/states_pt1.html
11:30 vdt-mik no this not work for sls file for pkg.install module
11:31 babilen vdt-mik: You can *either* use the "pkg.installed" execution module function to which you pass the name of the package you want to install or use "state.apply" to apply your /srv/salt/openvpn-client.sls state
11:32 babilen The latter would be done with "salt 'minion' state.apply openvpn-client"
11:32 babilen That is, I believe, what Norrland was hinting at
11:32 vdt-mik yes but how i cant set install package or uninstall packege ?
11:33 babilen I don't follow .. assuming your state is correct it should work (and so should a pkg.installed PKGNAME" run)
11:33 babilen If you encounter problems doing that paste your state(s), exact commands and the entire output to one of http://paste.debian.net, https://gist.github.com, http://sprunge.us, …
11:33 vdt-mik for example https://github.com/saltstack/salt-winrepo-ng/blob/master/openvpn.sls
11:34 vdt-mik this file install or uninstall openvpn
11:36 babilen I am in the lucky position where I don't ever have to deal with Windows, but I believe that those defintions allow you to use them as a Window Software Repository as defined on https://docs.saltstack.com/en/latest/topics/windows/windows-package-manager.html
11:36 mavhq joined #salt
11:39 entil https://github.com/saltstack-formulas/docker-formula/issues/49
11:39 saltstackbot [#49][MERGED] KeyError: 'pip.installed' -- after successful initial installation | Initial installation successful:...
11:39 entil I'm seeing that error even though I have a new version of pip and I'm calling it explicitly with `bin_env: /usr/local/bin/pip3` in my state
11:40 entil salt version 2016.11.2 and pip version 9.0.1 :/
11:43 babilen entil: Which error is that again?
11:43 entil babilen: "An importable pip module is required but could not be found on your system. This usually means that the system's pip package is not installed properly."
11:44 entil not even `use_vt: true` tells me which module it is
11:44 babilen Ah, so not the " KeyError: 'pip.installed'" the (initial) bug report is about?
11:44 entil err, sorry, that was misleading, it's not that, sorry
11:44 babilen That's why I ask :)
11:44 entil but it's kinda painful cuz doing salt-call on the minion installs it just fine and there's no real reason why it would fail
11:45 babilen So it works fine if you run it with "salt-call" on the minion?
11:45 entil sure
11:46 entil sudo salt-call cmd.run 'pip install docker-compose' (and also uninstall)
11:46 babilen But calling it from the master fails with the above error?
11:46 entil but pkg.installed in a state? no.
11:46 entil pip.installed sorry
11:46 entil having pip.installed: - name: docker-compose fails
11:47 babilen Wait .. running the *state* with salt-call still raises the error, but if you run an explicit command it doesn't?
11:47 entil babilen: http://paste.debian.net/918558/ my latest iteration of the state file, and applying it fails on the "could not be found" error
11:48 babilen Regardless of calling it on the minion with salt-call or from the master?
11:48 entil babilen: but running salt-call (on the minion), `sudo salt-call cmd.run 'pip install docker-compose'` installs just fine
11:48 entil I don't have salt-call on the master
11:48 babilen Sure
11:48 babilen salt-call == on the minion vs "from the master"
11:49 entil nowait, yes, but I couldn't get it working, "Could not access /etc/salt/pki/minion. Path does not exist."
11:49 babilen When do you run into that error?
11:50 entil invoking `sudo salt-call minionhostname cmd.run ls` on the master
11:50 babilen My impression is that your state (the one you pasted) does not work at all. It doesn't make a difference if you call it on the minion or from the master, you run into the "An importable pip module is required but could not be found on your system. This usually means that the system's pip package is not installed properly." error in either case
11:51 babilen You can, however, get pip to install the module by running a pip command explicitly with cmd.run
11:51 babilen Is that, so far, correct?
11:51 entil I didn't try to install it through a state with cmd.run
11:51 entil and I did not try to call the state, only cmd.run
11:51 babilen You'd do state.apply in that case
11:52 entil yeah, state.apply complains
11:52 entil that the module (whatever it is, doesn't say) cannot be found
11:52 babilen So you *did* call the state with state.apply :)
11:52 entil with that yes, but not salt-call
11:52 babilen k
11:52 babilen So my impression is correct
11:52 babilen I notice that this is pip 9.* vs pip 7.*
11:53 entil so would it be broken again with pip 9?
11:53 babilen I have no idea .. when/how did you install the python pip module and did you restart the minion afterwards?
11:53 Tanta joined #salt
11:53 babilen I just noticed *one* difference between the bug report and your setup
11:54 entil I ran `sudo /usr/bin/pip3 install --upgrade pip3` manually and I have restarted the minion
11:54 babilen I also wonder why you use pip3 in your state, but use pip in your cmd.run
11:54 entil does the minion track python paths or pip version numbers for the pip state thingie?
11:54 babilen How did you install pip in the first place?
11:54 entil babilen: that's because I'm tired and have been beating around this thing for too long :D
11:54 entil ubuntu 14.04's package
11:54 babilen Ah
11:55 entil but it does not matter, /usr/local/bin/pip is the first in PATH and /usr/local/bin/pip3 is the same file
11:55 babilen My guess would be that salt is using Ubuntu's package, while you performed the upgrade through pip. Salt might be picking up the packaged version still.
11:55 netcho_ joined #salt
11:56 entil if `- bin_env: /usr/local/bin/pip3` doesn't override, something's broken
11:56 babilen How do you know that salt uses that PATH?
11:56 entil "    Absolute path to a virtual environment directory or absolute path to a pip executable. The example below assumes a virtual environment has been created at /foo/.virtualenvs/bar."
11:56 entil I don't, I think
11:56 entil but if I give it explicitly, why wouldn't it work?
11:56 babilen But that would assume that the module can be imported from there
11:58 entil if it's a pip executable, there can't be a "from there" unless the pip state is somehow disregarding where this stuff is
11:59 entil I'll try one more thing, I seem to have forgotten `env_vars` is there
11:59 shoemonkey joined #salt
11:59 babilen I'm off for lunch, but my guess is that your underlying problem is that the system installation of pip clashes with your "manual" installation and that salt is *not* using the manual one via a suitable Python environment
12:00 entil yeah, something's broken with it
12:00 entil thanks for the help
12:02 babilen In particular "import pip" is not working on your minion
12:03 babilen (or "import pip.req")
12:04 babilen A failure in either would trigger the error you pasted above
12:04 babilen I'd import those manually and see what happens
12:04 babilen pip.req is, most likely, the underlying reason as you'll end up with the old Ubuntu version for that
12:06 rafaelcaricio left #salt
12:06 babilen And, I think, bin_env is more about the place where you install *to* not necessarily where "pip" is from
12:06 entil import pip :o where'd you get that?
12:07 babilen In the code
12:07 entil mmh
12:07 entil I tried setting env_vars: PYTHONPATH
12:07 babilen https://github.com/saltstack/salt/blob/develop/salt/states/pip_state.py#L35
12:07 Tanta just use cmd.run with a pip freeze | grep 'module' in the test
12:07 Tanta that's how I ended up dealing with the broken pip integration
12:08 babilen entil: I'm off for lunch .. good luck and keep me posted!
12:08 entil babilen: yeah thanks! I'm on the phone suddenly
12:08 bdrung_work joined #salt
12:19 GK___1wm____SU joined #salt
12:21 entil babilen: $ python3 -c 'import pip; print(pip)' and its __version__ points me to /usr/local/lib/python3.4/dist-packages/pip/__init__.py and import pip.req works too
12:22 entil babilen: this is in an interactive shell on the minion
12:23 entil on the minion `sudo salt-call cmd.run 'python -c "import pip"'` fails
12:25 toabi left #salt
12:26 entil I'll try making the state use cmd.run and pip install instead of pip.installed, because this came up in the phone call just now and it would be good if I could find a way to move on from this, no matter what the kludge
12:29 dps joined #salt
12:31 entil cmd.run seems to work just fine
12:32 edrocks joined #salt
12:46 mdhas joined #salt
12:54 puzzlingWeirdo joined #salt
12:54 jas02 joined #salt
12:55 filippos joined #salt
12:55 netcho_ joined #salt
12:58 gableroux joined #salt
13:00 johnkeates joined #salt
13:00 johnkeates using mount.mounted fails with "mount: only root can use "--options" option", how so?
13:03 tampakrap joined #salt
13:04 abednarik joined #salt
13:06 cowyn joined #salt
13:11 babilen entil: That's Python3 though
13:13 toastedpenguin joined #salt
13:14 johnkeates using mount.mounted fails with "mount: only root can use "--options" option", how so?
13:16 babilen johnkeates: "You are not running the command as root" comes to mind
13:16 johnkeates but salt-minion is running as root
13:16 johnkeates and so is the master
13:17 johnkeates sudo salt my_id cmd.run whoami > returns root
13:17 impi joined #salt
13:18 babilen Could you run the state on the minion with "salt-call -ldebug state.apply your.state.that.fails" and paste the output?
13:18 oms101_ joined #salt
13:19 rubenb_ joined #salt
13:20 JohnnyRun joined #salt
13:21 shoemonkey joined #salt
13:21 entil babilen: you mean python3 is expected to work worse?
13:21 ronnix joined #salt
13:22 babilen entil: I mean that salt is running in Python 2 and it is therefore unimportant if you can use pip in Python 3
13:22 johnkeates http://paste.debian.net/918577/
13:22 babilen Which raises the question of how to use Salt to install bits into a Python 3 virtualenv, but that is tangential to your issue
13:23 entil babilen: so salt doesn't support python3? am I not understanding?
13:23 johnkeates looks like by pasting i found a possible cause..
13:23 babilen johnkeates: Yay!
13:23 johnkeates the -user key is not he file owner user, but the user to run the command as
13:24 johnkeates but the docs state: The user to own the mount; this defaults to the user salt is running as on the minion
13:24 babilen entil: The salt-minion process is running in Python 2, so if *the minion* executes "import pip" it will import the Python 2 module from its PYTHONPATH.
13:24 entil babilen: yeah, that's pretty much the same as not supporting python3 :/
13:25 babilen entil: Salt does, not yet, support Python 3 completely (in that it can run in Python 3), but it should *still* offer you the ability to manage Python 3 virtualenvs. Those issues are somewhat related, but essentially separate.
13:26 babilen entil: I have not looked into managing py3 virtualenvs with salt -- Just trying to help you understand the issue
13:26 entil babilen: I don't like installing python stuff under /usr/local/ but always prefer virtualenvs, unfortunately that's not what I was asked for this time
13:26 entil babilen: so it's a nice comfort to know proper virtualenvs would work at least
13:27 babilen I wouldn't be sure about that .. I meant that it should™ not that it does support py3 virtualenvs
13:28 numkem joined #salt
13:29 entil myeah, well, cmd.run works ;P
13:29 edrocks joined #salt
13:31 babilen My assumption would be that it works for py3 virtualenvs defined with bin_env pointing to a suitable virtualenv
13:31 entil that, but not pip3
13:32 babilen Please note that there is a difference between the pip3 binary and the environment it runs in
13:32 babilen My suggestion would be to file a bug report with a minimal state that tries to install a package for py3
13:33 babilen fwiw, bin_env: /usr/bin/pip3 should work also
13:33 babilen Did you install the pip modules for py2 also?
13:34 babilen Just so that the salt process can use them
13:35 entil babilen: /usr/bin/pip3 is the old version, so if the version actually matters, it wouldn't work
13:35 entil and I did nothing for py2, which may have been a bad idea
13:35 babilen That wasn't specific to your issue, but more a generic "an absolute path to the pip3 binary should work"
13:35 entil unfortunately I had to move on from this, but in retrospect I should've stuck to py2 if I'd have known
13:35 babilen Well, why don't you try installing the Python module for pip so that salt itself can use them
13:36 entil because of time constraints on all the other things I need to get done today :(
13:36 babilen The fact that you did not install those modules is the reason for the error you get
13:39 twooster joined #salt
13:39 babilen I appreciate that it is annoying if bugs/problems get in the way of work, but sometimes it pays off to really solve them rather than using a workaround
13:40 babilen johnkeates: What was it?
13:40 jas02 joined #salt
13:42 rem5_ joined #salt
13:42 entil babilen: and it hurts my brain (and soul) to have to do it like this and the very least I'll do is document the findings, but this is the kind of project where I'm temporarily helping out in a hurry and need to do a million things, maybe my successor can fix this if I'm not on the case for long enough to return to it
13:43 babilen entil: I totally understand
13:43 babilen I'd recommend to 1. Install/Upgrade pip for Python 2 2. Restart minion 3. Revert to old/proper state and give it a try when you have a bit of time
13:44 babilen It might just work
13:44 babilen (with bin_env: /usr/local/bin/pip3 still in place)
13:44 johnkeates babilen: not sure, docs say user is for posix permissions, but code uses it to switch users a la sudo/su
13:44 johnkeates i've removed it for now and just do a chown every run, works as well, but not as clean
13:45 babilen johnkeates: Yeah, just looked into the code. Sounds like a bug that should be fixed!
13:46 johnkeates normally i'd poke at the issue tracker @ github, but i'm strapped for time :(
13:46 babilen .oO( Sounds familiar )
13:46 johnkeates it's been a bug for quite a while, just dug the database of all the state returns over the past 2 years, seems it failed every single time (but it didn't break anything and nobody bothered to check it out before me)
13:47 johnkeates i wonder if this channel could do with a salt-github-bot :p just !issue and have it appear on gh
13:47 babilen I think we have enough unattended bug reports already
13:47 johnkeates this is probably true
13:47 colegatron joined #salt
13:48 babilen (but then SaltStack is also not closing them nearly as aggressively as other projects)
13:48 johnkeates part of me 'dislikes' salt-enterpriser, but another part of me wants businesses to throw money at it and make more money available for developers
13:49 babilen I quite dislike all the "open core" projects, but .. apparently that's how things are done these days
13:49 ssplatt joined #salt
13:53 spicyJalapeno joined #salt
13:56 shoemonkey joined #salt
13:57 ksk hola
13:58 angel_dark joined #salt
14:01 cachedout joined #salt
14:17 cyborg-one joined #salt
14:18 racooper joined #salt
14:20 twooster Quick question -- is it possible to `include:` pillar or state files only if they exist? Perhaps using jinja templating? `file_exists` isn't working, and it may not be the best option.
14:22 brousch__ joined #salt
14:23 DexterF joined #salt
14:29 angel_dark1 joined #salt
14:30 shoemonkey joined #salt
14:33 tomatoschewps joined #salt
14:33 angel_dark2 joined #salt
14:34 angel_dark3 joined #salt
14:34 angel_dark4 joined #salt
14:44 tomatoschewps Hi there !
14:44 tomatoschewps I have to use special caractere in an SLS file (for vdata in reg.present)
14:45 tomatoschewps and  it won't work
14:45 tomatoschewps how can I specifie that I have special character ?
14:46 tomatoschewps - vdata: 'CheckCounter "Counter:traffic_out=\Interface réseau($ARG1$)\Octets envoyés/s" "Counter:traffic_in=\Interface réseau($ARG1$)\Octets reçus/s" ShowAll'
14:47 babilen What happens if you do that?
14:47 CeBe twooster: under which circumstances would a file not exist?
14:47 Rumbles joined #salt
14:49 nickabbey joined #salt
14:50 ssplatt joined #salt
14:51 tomatoschewps An exception occurred in this state: Traceback
14:51 tomatoschewps I can send all the error message
14:52 tomatoschewps http://pastebin.com/4piiRdnC
14:53 tomatoschewps it's a 'Windows' minion
14:54 Cadmus joined #salt
14:54 CeBe "'ascii' codec can't encode character" - does salt on windows not support utf8 encoding?
14:54 CeBe I assume that file is utf8?
14:57 babilen tomatoschewps, tomatoschewps: https://docs.saltstack.com/en/latest/ref/configuration/master.html#yaml-utf8
14:58 babilen Err, CeBe ^
14:58 tomatoschewps I have tried to specifie 'yaml_utf8: True' in salt master conf
14:58 babilen Oh, you didn't say
14:58 babilen What was the result of that?
14:58 tomatoschewps nothing ...
14:59 tomatoschewps but didn't do anything on the minion
14:59 tomatoschewps does I have to specify the same on the minion conf file ?
14:59 babilen No
15:00 CeBe tomatoschewps: did you reload the master after adding that to the config?
15:01 tomatoschewps is there some escape character I can use or replace special characters with the ascii corresponding code (if it exist)
15:01 CeBe also make sure the file is really utf8 and not something else
15:01 tomatoschewps service salt-master restart, yes
15:01 CeBe tomatoschewps: there is no ascii for those charaters
15:01 babilen Just to make sure: You set yaml_utf8: True in your master conf, that setting is not set *anywhere* else and you also start the SLS with "# -*- coding: utf-8 -*-" ?
15:02 scoates joined #salt
15:02 tomatoschewps # -*- coding: utf-8 -*- yes
15:02 twooster CeBe -- found an example to make it work (I needed the pillar_roots setting to tie it together). It's inclusion based upon a hostname. There's some external pillars that do it, but I wanted to avoid that.
15:02 tomatoschewps (copy/past just to be sure)
15:03 Xenophon1 tomatoschewps: you're using emacs, right?
15:03 tomatoschewps and yaml_utf8 in master conf anly
15:03 tomatoschewps nop, vim
15:03 Xenophon1 oh :(
15:03 Xenophon1 well make sure the file has the right BOM on it
15:03 tomatoschewps BOM ?
15:04 babilen byte order mark
15:04 PatrolDoom joined #salt
15:04 Xenophon1 in emacs you'd do something like C-x RET f utf-8-unix RET to convert the file from whatever encoding it's currently in to the correct UTF-8 format
15:05 tomatoschewps I'm searching what it is and next how to do it with vim
15:05 Xenophon1 it will put a (hidden) UTF-specific shebang at the beginning of the file to mark it as UTF-8 vs some other encoding
15:05 Cadmus Does salt care about self-signed certs when doing "source: https://self.signed.site/bork/bork/bork.foo" ?
15:05 tomatoschewps ok, thanks ! I will try it
15:06 Xenophon1 might be that your file is in another but similar-to-UTF encoding
15:06 babilen tomatoschewps: Could you paste your actual state, the configureation file and output of "salt --versions-report" ?
15:06 XenophonF joined #salt
15:07 tomatoschewps The state : http://pastebin.com/K0z5AYBC
15:07 CeBe encoding seems to match utf8: https://unicode-table.com/en/search/?q=e9 and the error message is complaining about 'ascii' parser not being able to read it
15:08 CeBe so I would search more in that direction, why does salt think it needs to use ascii instead of utf8?
15:08 Cadmus Because I have a problem with 'failed to list content's when doing an archive, extract, and I don't know if that's becuase it's not downloading it (it's not in the cache) or because I have a weirdly formed URL
15:08 tomatoschewps for the master conf, except the yaml_utf8, nothing change with the default conf
15:09 tomatoschewps (1069L and can copy/past anyway ...)
15:10 tomatoschewps http://pastebin.com/dW23EgYC
15:10 tomatoschewps salt version
15:10 tomatoschewps on CeBe, I will check it
15:10 jas02 joined #salt
15:10 tomatoschewps *ok
15:11 cachedout joined #salt
15:11 cachedou1 joined #salt
15:13 sarcasticadmin joined #salt
15:13 ssplatt joined #salt
15:14 CeBe tomatoschewps: I just tried this on my salt installation. yaml_utf8 seems to have no effect.
15:15 CeBe this case is linux -> linux
15:16 tomatoschewps so it could be that salt don't recognize UTF8 encoding and try to use it has ascii ?
15:16 Joy joined #salt
15:16 CeBe that is the case, yes. what is weird is that specifying # -*- coding: utf-8 -*-  and yaml_utf8: True  do not work
15:18 Joy newbie question - how do i pass along a specific module option via state? i want to create a filesystem with a label, but blockdev.formatted generic options don't seem to help
15:18 fracklen joined #salt
15:18 johnkeates check the sauce
15:18 CeBe tomatoschewps: https://github.com/saltstack/salt/issues/38776
15:18 saltstackbot [#38776][MERGED] State unable to handle data containing unicode | Description of Issue/Question...
15:19 CeBe there are some workarounds in there
15:19 abednarik joined #salt
15:19 CeBe saltstackbot: MERGED? that is an issue :)
15:20 Joy is kwargs used for this?
15:20 cachedout Heh, that bot needs some love.
15:21 Joy huh, am i really thinking of https://github.com/saltstack/salt/issues/33857 ?
15:21 saltstackbot [#33857][OPEN] Feature request: filesystem label support for blockdev.formatted | Description of Issue/Question...
15:24 AvengerMoJo joined #salt
15:25 tomatoschewps ok, lot to process and test ...
15:26 tomatoschewps I'm on it :)
15:26 tomatoschewps Thanks
15:29 mdhas joined #salt
15:33 mikecmpbll joined #salt
15:34 mdhas joined #salt
15:40 GK___1wm____SU joined #salt
15:42 MattP_ joined #salt
15:43 Brew joined #salt
15:43 _JZ_ joined #salt
15:44 MattP_ Hello everybody! Does anybody ever encountered a minion not being able to decipher a gpg message *only* at the first highstate ? It works fine after...
15:44 mikecmpbll joined #salt
15:47 impi joined #salt
15:51 GK___1wm____SU joined #salt
15:52 lompik joined #salt
15:52 jas02 Hello, how can I achieved following scenario in Salt. I need to run state1.sls on host1, then state2.sls on host2, then state2.sls on host1 and state4,sls on host2. I know that there are requisites and ordering.  Are they working between different hosts? Is there some example?
15:54 teclator joined #salt
15:54 cliluw joined #salt
15:54 GK___1wm____SU joined #salt
15:59 jimklo joined #salt
16:00 whytewolf jas02: https://docs.saltstack.com/en/latest/topics/orchestrate/orchestrate_runner.html#orchestrate-runner
16:02 TheoSLC joined #salt
16:02 cliluw joined #salt
16:02 duncanmv joined #salt
16:03 ssplatt joined #salt
16:03 TheoSLC Good morning.  So I want to get rid of the top file for states and move to a model where the minion defines it's own top.sls for states.  Is this a built-in feature?
16:04 jas02 whytewolf: Thanks.
16:05 sp0097 joined #salt
16:08 ssplatt joined #salt
16:12 CeBe TheoSLC: you can specify the states to apply on startup in the minion config
16:12 abednarik joined #salt
16:13 oms101_ joined #salt
16:18 GK___1wm____SU joined #salt
16:20 ivanjaros joined #salt
16:21 Shirkdog joined #salt
16:22 bantone is there something similar to the pkgrepo module that I can use for archlinux?
16:22 bantone any suggestions :)
16:24 cliluw joined #salt
16:25 XenophonF MattP_: I'm only using the gpg filter with Pillar, and I don't recall having that problem.
16:27 XenophonF TheoSLC: are you talking about a masterless Salt deployment?
16:27 XenophonF TheoSLC: maybe this will answer your question - https://docs.saltstack.com/en/latest/topics/tutorials/quickstart.html
16:35 GK___1wm____SU joined #salt
16:36 implicitnewt joined #salt
16:38 mdhas joined #salt
16:39 implicitnewt I'm testing deployment with salt cloud and have run into an issue.  My /tmp is noexec in this environment and the deploy script is failing.  My profile config is at http://pastebin.com/RKT70J9b, adding the bits from https://docs.saltstack.com/en/latest/topics/cloud/misc.html to try and get the deploy running
16:41 implicitnewt It continues to try and run from a dynamic /tmp/.salstack-UU-I-D/deploy.sh directly and doesn't use the deploy_cmd or tmp_dir in the profile
16:43 vegasq joined #salt
16:44 cliluw joined #salt
16:44 XenophonF pastebin, a wretched hive of tracking cookies and adware ;)
16:46 implicitnewt ghostery does block a lot of crap on that site, if you know another I would gladly use it :)
16:47 XenophonF paste.debian.net, ix.io, and gist.github.com are my go-to pastebins that aren't pastebin ;)
16:47 XenophonF anyway, i'm grepping the source and it looks like you're doing it right
16:47 XenophonF maybe there's something in salt-bootstrap that's insisting on that temp directory
16:49 XenophonF hm, i didn't realize the SLES AMIs mounted /tmp noexec
16:50 mikecmpbll joined #salt
16:50 XenophonF you could try setting tmp_dir to / or /home/my-user, and then adjusting deploy_command accordingly
16:51 vegasq joined #salt
16:52 TheoSLC XenophonF: I'm not talkinga bout masterless.  I'
16:53 implicitnewt there is a script arg for bootstrap that you can set the temp config dir that seems to be ignored.  I'll try moving everything to home and see what I get
16:54 implicitnewt and the SLES AMIs don't by default, this is a custom build for an environment that requires it
16:54 XenophonF ah
16:55 XenophonF maybe run salt-cloud with debug or trace logging enabled?
16:55 XenophonF you might also try running `salt-cloud -u` to update its copy of salt-bootstrap
16:56 XenophonF what version of salt are you running?
16:56 implicitnewt I did run the update piece, I haven't tried debugging quite yet, I'll see if that reveals anything
16:56 XenophonF oh, and salt-cloud has a --keep-tmp option, which might help
16:56 implicitnewt salt 2016.3.4 (Boron)
16:57 XenophonF if you can, you might want to upgrade to at least 2016.3.5 or all the way to 2016.11.3
16:57 rem5 joined #salt
16:58 XenophonF i realize that's kind of generic advice to be giving
16:58 XenophonF brb lunch time
16:59 shoemonkey joined #salt
17:00 Praematura joined #salt
17:01 Trauma joined #salt
17:03 heaje joined #salt
17:04 GK___1wm____SU joined #salt
17:05 tiwula joined #salt
17:08 mdhas joined #salt
17:09 ablemann joined #salt
17:11 nickabbey joined #salt
17:12 leonkatz joined #salt
17:13 mdhas joined #salt
17:15 GK___1wm____SU joined #salt
17:20 orionx joined #salt
17:20 implicitnewt So I set the tmp_dir to /home/.salt and removed the deploy_cmd and it clipped through.  Next challenge is to get it to set the hostname
17:23 edrocks joined #salt
17:30 alvinstarr joined #salt
17:30 GK___1wm____SU joined #salt
17:35 orionx joined #salt
17:39 aldevar left #salt
17:42 ssplatt joined #salt
17:43 iggy TheoSLC: no (aside from the minion running state.sls for every state you want)
17:54 Cottser joined #salt
17:55 cliluw joined #salt
17:55 XenophonF back
17:55 XenophonF implicitnewt: i figured out the hostname stuff
17:55 XenophonF hang on a sec, this should work for SLES
17:56 XenophonF implicitnewt: https://github.com/irtnog/salt-states/blob/development/hostname/init.sls
17:56 toastedpenguin there is a system load beacon, any CPU utilization beacons?
17:57 XenophonF i wrote that for RHEL/CentOS but think it will work on SLES/OpenSUSE, too.
17:57 Edgan joined #salt
17:59 implicitnewt nice... writing a state and including in my top was the direction I was heading so this will give me a start there to test out
17:59 implicitnewt thanks much
17:59 nickabbey joined #salt
17:59 whytewolf toastedpenguin: you mean like this? https://docs.saltstack.com/en/latest/ref/beacons/all/salt.beacons.status.html#module-salt.beacons.status
17:59 Cottser joined #salt
18:00 c-rad joined #salt
18:00 mkillebrew joined #salt
18:00 impi joined #salt
18:01 mdhas joined #salt
18:01 Cottser joined #salt
18:01 Cottser joined #salt
18:05 av_ joined #salt
18:14 mikecmpbll joined #salt
18:15 foundatron joined #salt
18:16 aldevar joined #salt
18:16 orionx joined #salt
18:18 DEger joined #salt
18:19 nixjdm joined #salt
18:20 jas02 joined #salt
18:21 GK___1wm____SU joined #salt
18:21 Trauma joined #salt
18:22 cyborg-one joined #salt
18:27 DEger joined #salt
18:34 nickabbey joined #salt
18:39 twooster joined #salt
18:43 fracklen joined #salt
18:46 tapoxi joined #salt
18:47 mdhas joined #salt
18:51 abednarik joined #salt
18:54 jhauser joined #salt
19:00 shoemonkey joined #salt
19:01 nickabbey joined #salt
19:03 nickabbey joined #salt
19:04 shanathon joined #salt
19:04 jas02 joined #salt
19:04 nickabbey joined #salt
19:04 ssplatt joined #salt
19:05 candyman88 joined #salt
19:05 shanathon good day. I am looking to see if anyone has seen or messed around with dynamic pillars yet. We had a gentleman come in yesterday from Salt who recommended looking at this for domain joins with the idea that the minions would only have temporary access for the domain join passwords and then the pillar would remove the password subsequently
19:05 shanathon has anyone ever heard or seen an implementation like this within saltstack?
19:06 shanathon * or i should say the password would be removed from the pillar subsequently
19:06 nickabbey joined #salt
19:07 theblazehen joined #salt
19:08 jas02 joined #salt
19:08 Praematura joined #salt
19:10 jhauser joined #salt
19:11 KyleG joined #salt
19:11 KyleG joined #salt
19:16 cmarzullo Do you mean like external pillar?
19:17 Edgan shanathon: dynamic pillars are external pillars, and you can have many sources
19:17 cmarzullo I could imagine a scenario where an external pillar looks up a one time password for a minion.
19:22 systemexit joined #salt
19:23 SaucyElf joined #salt
19:24 GK___1wm____SU joined #salt
19:24 theblazehen joined #salt
19:25 jas02 joined #salt
19:30 jas02_ joined #salt
19:35 mavhq joined #salt
19:40 shanathon yea not sure. I am familiary with external pillars. For example, storing the password in something like Hashicorp's vault but that doesn't neccessarily cover the temporary access peace of it
19:40 SaucyElf joined #salt
19:40 shanathon it just talks to the location of where the pillars are stored, right?
19:41 shanathon *piece :)
19:43 ChubYann joined #salt
19:43 * whytewolf wonders if they are talking about passing pillars in through the command line.
19:45 pcn A friend of mine talked about how google created a one-time time limited password for ssh access, and passed it in via node data.  So it has limited usefulness, but could be queried via the api and used to access a node without any other shared secret.
19:45 pcn My recollection could be missing a bit, but is that sort of the use case here?
19:48 jas02 joined #salt
19:52 cyborg-one joined #salt
19:57 nickabbey joined #salt
19:57 shanathon sort of.. the idea is that a minion should not need access to the password once it has been joined to the domain (which is the primary use of the password).
19:58 cliluw joined #salt
20:00 shanathon here is how I am calling the pillar via the state file... https://www.screencast.com/t/yvFgFsQ1GC
20:00 shanathon and im just applying the state by targeting the OS grain..i.e windows
20:04 shanathon just occurred to me.. maybe i could have a custom grain that indicated it is not joined to the domain, and once the it is domain joined, update that grain accordingly....
20:04 shanathon so it is no longer targeted
20:04 shanathon hmmm
20:10 jas02 joined #salt
20:11 leonkatz joined #salt
20:13 simondodsley joined #salt
20:22 simondodsley https://www.irccloud.com/pastebin/Zz8vHpGo/
20:25 fracklen joined #salt
20:28 cachedout We're pleased to announce that we're hiring a developer for the Salt core team. If you're interested, please have a look here: https://saltstack.bamboohr.com/jobs/view.php?id=50
20:31 hemebond Ooo, nice.
20:32 juntalis joined #salt
20:41 XmodeMonkX joined #salt
20:42 gsnewb joined #salt
20:42 XmodeMonkX Hi all. Can anyone spare a minute and help me try to resolve this problem: http://pastebin.com/wf6wZuWU
20:43 XmodeMonkX The unless: statement doesn't seem to work, however the named cmd.run does. By this, I mean, The unless statement always seems to return as true, and thus never runs the named cmd.
20:43 om2 joined #salt
20:43 XmodeMonkX If I remove the "unless" arguments, then the named cmd works fine, and as expected.
20:44 hemebond And the grep properly returns an exit code 1 when it fails to find 200?
20:44 gsnewb Hi there!  Looking for some guidance creating a state file for a windows scheduled task.  followed the example in salt.modules.win_task.create_task documentation, but getting a variety of errors.
20:44 hemebond Oh wait.
20:45 XmodeMonkX hemebond: Yes. Manually ran, it works as expected. Anything other then 200, then $? == 1 and on a 200 then $? == 0
20:45 gableroux joined #salt
20:45 hemebond Checked the minion log?
20:45 hemebond Tried running it locally using salt-call -l debug state.apply curlstate?
20:47 gsnewb example command and result here: http://pastebin.com/WDfvpVBE
20:47 hemebond gsnewb: Tried putting the arguments in quotes?
20:47 hemebond Each argument value I mean.
20:49 gsnewb tried a few individual args in quotes, but not all.  i will try that, thanks
20:50 XmodeMonkX hemebond: The debug doesn't present much in the way of helping me debug anything. Merely refers to it as "[INFO    ] Completed state ...."
20:50 XmodeMonkX So, it thinks that it returned perfectly fine.
20:50 hemebond XmodeMonkX: Can you see the "unless" request hit the API?
20:50 gsnewb hemebond: same result
20:51 cscf Is there a good way to provide a sudo password to salt-ssh?
20:52 hemebond gsnewb: Unfortunately I don't have a Windows machine to test with.
20:52 gsnewb no worries, thanks
20:53 hemebond Unless you can get a line number from the logs for the error.
20:53 hemebond Try using salt-call -l debug task.create_task ......
20:53 hemebond And look in the salt-minion log.
20:54 hemebond XmodeMonkX: Have you checked the curl output without grep?
20:54 hemebond Have you tested the request through salt itself rather than using curl directly?
20:54 XmodeMonkX hemebond: Yes. With grep directly, the returned int is exactly what I expect. The grep will return 0 if found the argument, and 1 if not found.
20:55 hemebond Even through Salt?
20:55 XmodeMonkX hemebond: I am trying to find the api logs, and not found the api logs yet.
20:55 XmodeMonkX hemebond: Not via salt, but I will do that now actually.
20:55 dyasny joined #salt
20:56 hemebond oh
20:56 hemebond You have - cmd: |-
20:56 hemebond But unless is a string.
20:56 hemebond er, takes a string.
20:56 hemebond Or list.
20:56 hemebond It wants a list.
20:56 hemebond So put your command directly into a list item, rather than under cmd:
20:57 XmodeMonkX hemebond: Oh really? Sweet. Lemme try
20:57 hemebond Or even as a string, e.g., unless: /usr/bin/curl .....
20:57 hemebond https://docs.saltstack.com/en/2016.3/ref/states/requisites.html#unless
20:57 XmodeMonkX hemebond: With all the "'s and ''s This method seemed to be the best method.
20:58 XmodeMonkX Also, the :'s etc :-)
20:58 hemebond XmodeMonkX: You can still use |- if you want to
20:58 hemebond But the cmd: shouldn'
20:58 hemebond t be there.
20:58 hemebond Unless is a list or a string, not a dict.
20:58 hemebond or
20:58 hemebond Takes a string or list of strings.
20:58 hemebond *er
20:59 hemebond Not a list of dicts, which is what your cmd: creates
20:59 whytewolf actually the stateful shouldn't be on that line either
21:00 whytewolf unless is what tests for statefullness.
21:00 XmodeMonkX whytewolf: That was something I added later, figured it couldn't hurt - but if it does, then I will remove it.
21:00 whytewolf but yeah your command is passing cmd: into your shell
21:01 XmodeMonkX aha!
21:01 XmodeMonkX Perfect, thanks!!
21:01 shoemonkey joined #salt
21:01 XmodeMonkX hemebond: Thanks for your time!!
21:01 hemebond ????  Happy to help
21:02 XmodeMonkX whytewolf: Whanks for the additional input!
21:02 XmodeMonkX *Thanks. Waw. that was a bad mistype!!
21:04 whytewolf no problem. i should have stepped out for lunch earlier and seen this convo from the start ;)
21:04 XmodeMonkX :-)
21:17 nZac joined #salt
21:17 jas02 joined #salt
21:23 citaret joined #salt
21:33 s_kunk joined #salt
21:34 dyasny joined #salt
21:57 leonkatz joined #salt
22:05 dyasny joined #salt
22:20 vegasq joined #salt
22:27 bltmiller joined #salt
22:28 prg3 Adding users to windows is driving me crazy.. it worked once, and now I'm getting python errors, even from a completely new system: "TypeError: chhome() takes exactly 2 arguments (3 given)". Using user.present, and there is no home defined.. but it doesn't matter if it is, same error.
22:29 jas02 joined #salt
22:30 nickabbey joined #salt
22:32 vegasq joined #salt
22:36 relidy Is there a "best practice" method to ensuring a specific system user account (that I'm not managing) exists before allowing a particular state to run? I'm allowing the pillar to control the username, but want to sanity check it before relying on it (e.g. setfacl).
22:36 Edgan relidy: you can set a require if you want
22:37 Edgan relidy: I am not a fan of requires, and prefer to just control the order of the formula
22:37 relidy That doesn't seem to work if I haven't previously managed the user in some way (user.present)
22:37 relidy And user.present tries to create the user if it doesn't exist.
22:37 Edgan relidy: yes, you have to create the user throw salt
22:37 Edgan relidy: Salt can't require things you haven't declared
22:38 Edgan relidy: if you do it right in salt, the user.present will just verify the existing user
22:38 whytewolf could use unless or onlyif
22:38 relidy I'm accomplishing it right now through a cmd.run (`id {{ username }}`), but I'm looking for a better way
22:38 Edgan relidy: What I said is the better way
22:38 relidy But that's the thing, it's a system account (apache). I just want to sanity check the pillar input before using it for other operations that wouldn't fail if the name is wrong
22:39 wryfi i think whytewolf has the right idea
22:39 Edgan relidy: Doesn't matter that it is an existing system account.
22:39 wryfi unless: getent passwd <username>
22:39 Edgan relidy: Just user.present it, and let it do the work
22:39 wryfi or something like that
22:40 relidy wryfi, whytewolf: I think that's probably the ticket.
22:40 Edgan relidy: The more salt knows about the better
22:40 relidy Edgan: It's a system account. I do *not* want Salt creating the account if someone typos the account name.
22:40 Edgan relidy: That is basically what user.present would do for you, why you trying to recreate the wheel?
22:41 Edgan relidy: Hard code the require for apache
22:41 Edgan relidy: don't rely on what is in the pillars
22:41 relidy And if my environment is mixed?
22:42 Edgan relidy: mixed as in apache sometimes and apache2 others?
22:43 relidy If the apache username isn't consistent. Whether that's due to platform differences, multiple processes running on the same machine, or whatever.
22:43 Edgan relidy: You can map the differences between distributions and releases with a map.jinja
22:43 whytewolf it is prodictable. so you could use a map.jinja for finding the name, setting the service name. ect.
22:44 Eugene joined #salt
22:44 Edgan relidy: https://paste.fedoraproject.org/paste/diGPOz9Exgw~uK~9G6ttF15M1UNdIGYhyRLivL9gydE=
22:44 relidy In simple/usual cases, yeah, it's predictable. But it isn't always and that's one of the issues I'm trying to plan for.
22:44 whytewolf ... if it is a service account setup by the operating system yes it is predictable
22:45 Edgan whytewolf: it is setup by the package, and he could be using a mix of packages even on the same distro/release
22:45 Edgan whytewolf: My solution to this problem is to short circuit the user the package creates by creating the user in salt before installing such packages
22:46 whytewolf agreed. also if you are using a mix of packages normally those packages follow the same set standards through out systems.
22:47 whytewolf the only time you run into unpredictable user names in packagees is when you have a chump in the back making your own packages
22:47 Edgan whytewolf: yes, or random mix of ppas/etc
22:47 relidy I think the unless approach is going to solve my current needs, but I do appreciate the discussion. It's easy to get stuck thinking there's only one approach (especially when you don't know all that much to start with)
22:48 Edgan whytewolf: I don't remember what it was, but in one case I ran into a package that didn't like me creating the user for it.
22:48 Edgan whytewolf: most check for the user already existing, and skip the step if it does
22:48 whytewolf ugh, ppas are one of the reasons i hate ubuntu. redhat still trys to follow standards with their repos which is why most people hrow fits with redhat repos that over right system files
22:49 overyander joined #salt
22:49 Edgan whytewolf: I agree with you, but in AWS and with developers, Ubuntu has pretty much taken over.
22:49 * relidy is a developer and doesn't use Ubuntu.
22:49 Edgan whytewolf: I am using Ubuntu here, but wish I was back on CentOS
22:50 Edgan relidy: I am talking in general, and in silicon valley where I am. But I have heard the same outside SV too.
22:50 Edgan relidy: AWS?
22:50 relidy Edgan: Nah, no AWS here. I'm a happy CentOS user, though.
22:51 whytewolf AWS is kind of a worse case example even their main distro is a basterdized redhat.
22:52 Edgan My employers in SV since 2006. CentOS, Ubuntu, Ubuntu(most)/CentOS(databases), was Ubuntu converted to CentOS, Amazon Linux and converted to CentOS, Ubuntu and converted to CentOS, Ubuntu, Ubuntu
22:52 hemebond I use Debian on AWS
22:52 Eugene I never understood why people complain about ubuntu PPAs as being evil - any idiot can set up a rpm repo just as easily and break all kidns of things
22:52 relidy Heh, too true
22:52 Eugene Its not a problem endemic to the packaging system, but to the individual packages
22:52 Eugene Related: installing fedora RPMs without doing a rebuild from srpm. Just say no.
22:54 Edgan Eugene: The hair one is when it is something like jenkins, and it is tied to the java packages, and there is no way to convert the package. I ended up a couple employers ago just running the Jenkins master on Fedora.
22:54 Eugene I've had zero problems with the official jenkins repo ;-)
22:55 Edgan Eugene: I prefer "proper" packages when I can get them.
22:55 Edgan Eugene: But in Ubuntu land yeah I am using the official debs
22:55 Eugene I find upstream packages, when provided, to be superior to the madness of volunteers that is EPEL
22:56 Edgan Eugene: Fedora standards are a head above any upstream package. Upstream will do crazy shit, but in general it works well enough.
22:56 Eugene I've been using Red Hat & friends for long enough that I think the opposite is true
22:56 Edgan Eugene: The latest form is docker images. Now they can even pick their own distro/release, and do even more craziness, because it is contained.
22:57 Eugene Oh god don't get me started about docker
22:57 Edgan Eugene: But that is exactly where upstream packages are going
22:57 Eugene I wish people would stop screwing-abotu with docker and just post VMDK or QEMU imgs
22:57 Edgan Eugene: Give it a few years and I expect Jenkins will probably consider the docker image the official way
22:58 Eugene IT is a cyclical industry, just like all the rest of them. QUick, what's the difference between a Mainframe and The Cloud
22:58 Edgan Eugene: the adoption is already high, it is a done deal as a standard
22:58 N-Mi joined #salt
22:58 N-Mi joined #salt
22:59 Edgan Eugene: yes, I am aware, but do you want to way another 20 years for it to come back around to sanity?
22:59 Eugene Nah, I'll just do it now and ignore all the haters.
22:59 Eugene Somebody has to set a trend, why not me?
22:59 Edgan what is it?
22:59 Edgan Eugene: Docker is win the trend setting by 10x.
23:00 Eugene They're both "somebody else's computer", but Mainframes come with a support contract
23:00 Edgan Eugene: AWS comes with a support contract too, and Canonical just started offering to support your Ubuntu in AWS under contract.
23:00 Eugene I've dealt with AWS "support" enough to not bother paying for it ever again
23:01 max42 joined #salt
23:01 Edgan Eugene: That is only because you had to deal with the first tier. I agree the bottom is ok at best. But you get big enough for them to care about and you will get a dedicated person.
23:02 shoemonkey joined #salt
23:02 Edgan Eugene: A major trend in the industry is management lets everything degrade to good enough level.
23:03 max42 I want to copy a file to the minion only if it exists on the master as salt://files/some_file_{{grains['host'}}. How do I check for existence of a file in the file backend?
23:06 zifnab Edgan: i can say (from experience) AWS support comes in two flavors: not enough zeros, and too many zeros
23:06 zifnab you don't pay them enough, they don't give a fuck to deal with anything
23:06 zifnab you pay too much, they don't give a fuck because it'd cost you too much to move
23:16 cliluw joined #salt
23:16 bltmiller joined #salt
23:20 vegasq joined #salt
23:21 roock joined #salt
23:24 cachedout Heh
23:26 guerby joined #salt
23:27 Horgix joined #salt
23:28 Qwazerty joined #salt
23:30 jas02 joined #salt
23:31 Edgan zifnab: I think they care about PR. Any of the household name companies that are on AWS give white glove service.
23:33 cachedout Speaking of AWS. Today I got them to agree to feature Salt in their QuickStart guide section. Any volunteers out there who might be interested in taking the lead on putting together a tutorial for Salt on AWS? https://aws.amazon.com/quickstart/
23:33 cachedout If so, contact me. :]
23:33 sarcasticadmin joined #salt
23:34 Edgan cachedout: cool
23:40 TheoSLC joined #salt
23:41 JPaul joined #salt
23:42 TheoSLC joined #salt
23:44 Roelt joined #salt
23:52 JPaul joined #salt
23:59 JPaul joined #salt

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