[Avocado-devel] Make optional interfaces implementation for plugin system.

2017-06-06 Thread Andrei Stepanov
Hello. Currently Avocado Plugin System requires to implements all virtual methods in plugins. Can we remove this restriction? RFE: Allow to implement only necessary interfaces for plugins. Just skip their invocation of absent methods in plugins? 1. This will allow as easily add new methods to

Re: [Avocado-devel] [RFC v3]: Avocado maintainability and integration with avocado-vt

2017-05-03 Thread Andrei Stepanov
> But this actually brings a question: is EL7 + Avocado RPM from official > repo, still RHEL7? I mean, from a QE perspective, is this acceptable? > If not, the only real solution I see is also adding Avocado to EL7. SpiceQe currently sticks to python-virtualenv approach. Python virtual env holds

Re: [Avocado-devel] Subprocess termination

2017-02-20 Thread Andrei Stepanov
My point is: IMHO there's nothing broken with Avocado in this regard. > If you want automatic ("automagic") resources cleanup, something like > running tests on a VM with a snapshot is the way to go. But the most > correct way is indeed trying to keep track of the resources you crea

Re: [Avocado-devel] Subprocess termination

2017-02-17 Thread Andrei Stepanov
d recommend to finish the process in tearDown(). Not sure if >> we should make the runner to track the sub-processes created by the >> test... let me open a card to discuss that. >> >> >> > >> > Radek >> > >> > On Thu, Feb 16, 2017 at 5:09

[Avocado-devel] pre_command / post_command clarification

2017-02-10 Thread Andrei Stepanov
Hi. We need a reliable mechanism to notify Beaker server about start+result of a avocado-vt test. As you know, avocado-vt test is one of the many tests produced from cartesian config. We need to notify Beaker about start+results of _each_ test. Currently we use: pre_command / post_command I

Re: [Avocado-devel] Avocado ignores global config

2017-02-09 Thread Andrei Stepanov
Wow! Thank you. On Thu, Feb 9, 2017 at 9:31 AM, Lukáš Doktor <ldok...@redhat.com> wrote: > Dne 8.2.2017 v 17:54 Andrei Stepanov napsal(a): > >> Hi. >> >> I am trying to run avocado in virtualenv. Avocado ignores data_dir in >> global config. >> &g

[Avocado-devel] Avocado ignores global config

2017-02-08 Thread Andrei Stepanov
Hi. I am trying to run avocado in virtualenv. Avocado ignores data_dir in global config. (virtualenv) -bash-4.2# avocado --version Avocado 45.0 (virtualenv) -bash-4.2# cat /etc/avocado/avocado.conf [datadir.paths] data_dir = /mnt/tests/spice/qe-tests/avocado-data logs_dir =

Re: [Avocado-devel] Heads-up: dropping Python 2.6 support

2017-02-07 Thread Andrei Stepanov
avocado list How do you think, is this right approach to use avocado + avocado-vt on RHEL 6.x + RHEL 7.x? On Tue, Feb 7, 2017 at 2:41 PM, Cleber Rosa <cr...@redhat.com> wrote: > > > On 02/07/2017 08:34 AM, Andrei Stepanov wrote: > > > > > > On Tue,

Re: [Avocado-devel] Heads-up: dropping Python 2.6 support

2017-02-07 Thread Andrei Stepanov
On Tue, Feb 7, 2017 at 2:20 PM, Cleber Rosa wrote: > > > > Andrei, > > Running a command such as: > > pip install -r > https://raw.githubusercontent.com/avocado-framework/avocado/ > 36lts/requirements.txt > > Would have probably saved you a lot of this hassle. This is

Re: [Avocado-devel] Heads-up: dropping Python 2.6 support

2017-02-07 Thread Andrei Stepanov
Cleber Rosa <cr...@redhat.com> wrote: > On 02/07/2017 06:46 AM, Lucas Meneghel Rodrigues wrote: > > I'm not sure putting a file with references to a relatively independent > > plugin in the main avocado repo the correct thing to do. It's probably > > better to put a

Re: [Avocado-devel] Bug: Once created VM object stays for all tests.

2017-02-02 Thread Andrei Stepanov
to you? >> Lukáš >> >> PS: I don't say this is the optimal solution, but it exists for so long >> that no one sane would try to rewrite it with a different approach so >> I'd suggest just copy the solution already used in code rather >> than inventing something c

Re: [Avocado-devel] Bug: Once created VM object stays for all tests.

2017-02-02 Thread Andrei Stepanov
t just copy the solution already used in code rather than > inventing something clearer (like a better `VM.needs_restart` method). > > > Dne 2.2.2017 v 17:48 Andrei Stepanov napsal(a): > >> Ok. >> I do not agree with this approach. >> (Calling .create() on old VM o

Re: [Avocado-devel] Bug: Once created VM object stays for all tests.

2017-02-02 Thread Andrei Stepanov
key, value) self.spice_options[skey] = value What do you think about this solution? On Thu, Feb 2, 2017 at 5:35 PM, Lukáš Doktor <ldok...@redhat.com> wrote: > Dne 2.2.2017 v 15:07 Andrei Stepanov napsal(a): > >> 1. >> >> 2017-02-02 13:23:59,568 job L03

Re: [Avocado-devel] Bug: Once created VM object stays for all tests.

2017-02-02 Thread Andrei Stepanov
as > a backend). > > Would you please (publicly or in private) share more details I might be > able to identify why the machine is not being re-created. > > Regards, > Lukáš > > Dne 31.1.2017 v 18:15 Andrei Stepanov napsal(a): > >> Hi. >> >> It see

[Avocado-devel] avocado-vt: pull request 885, please merge it.

2017-02-02 Thread Andrei Stepanov
Hi. Could you please run your tests with applied: https://github.com/avocado-framework/avocado-vt/pull/855 If it doesn't broke your tests then I would ask you to merge it. Thank you.

[Avocado-devel] What is a right way to install avocado?

2017-02-01 Thread Andrei Stepanov
Hello. We are currently experiencing some issues with avocado / avocado-vt. Our automation can be described in next steps: 0. Install RHEL 6/7. 1. Clone "master" branches for avocado/avocado-vt from github. 2. In avocado dir: make requirements python setup.py install 3. In avocado-vt dir:

[Avocado-devel] avocado-vt broken travis check

2016-11-10 Thread Andrei Stepanov
Hello It seems that travis marks all pull reqs as bad. https://github.com/avocado-framework/avocado-vt/pulls Could someone confirm this? How much it takes to fix this? Thanks.

Re: [Avocado-devel] Forced values at /etc/avocado/conf.d/vt.conf

2016-11-02 Thread Andrei Stepanov
; whether this fix works for you https://github.com/avocado-fra > mework/avocado-vt/pull/742 > > (you need to set the config option net_type to "none". It won't work if > you leave the option without value) > > Regards, > Lukáš > > Dne 1.11.2016 v 12:21 Andrei S

[Avocado-devel] Forced values at /etc/avocado/conf.d/vt.conf

2016-11-01 Thread Andrei Stepanov
Hello. I noticed that values defined at /etc/avocado/conf.d/vt.conf have the highest priority. This approach is quite awkward, and inflexible. See: https://github.com/avocado-framework/avocado-vt/blob/master/shared/cfg/guest-hw.cfg#L368 variants: - @bridge: nettype = bridge - macvtap: nettype

[Avocado-devel] Broken sessions.

2016-10-20 Thread Andrei Stepanov
Hello I need to find a solution to one problem. Sometimes, ssh sessions to vms are terminated. Different sessions, at different places. I cannot find any explanation for this. Today in the morning I got ~20 times of this error. After a while it just disappeared. I just was running the test again

Re: [Avocado-devel] Broken cartesian_config by commit 81c6ce860b2f625e

2016-08-19 Thread Andrei Stepanov
Attached configs. On Fri, Aug 19, 2016 at 4:02 AM, Wei, Jiangang <weijg.f...@cn.fujitsu.com> wrote: > Hi > On Thu, 2016-08-18 at 12:06 +0200, Andrei Stepanov wrote: > > spice configs > > http://file.brq.redhat.com/astepano/tmp/spice-configs.tar.gz > > I can't fetc

Re: [Avocado-devel] Broken cartesian_config by commit 81c6ce860b2f625e

2016-08-18 Thread Andrei Stepanov
https://github.com/avocado-framework/avocado-vt/pull/668 On Thu, Aug 18, 2016 at 2:39 PM, Cleber Rosa <cr...@redhat.com> wrote: > On 08/18/2016 06:50 AM, Andrei Stepanov wrote: > > Hello. > > > > Hi Andrei, > > > We have now broken cartesian_config.py. >

Re: [Avocado-devel] Broken cartesian_config by commit 81c6ce860b2f625e

2016-08-18 Thread Andrei Stepanov
spice configs http://file.brq.redhat.com/astepano/tmp/spice-configs.tar.gz to test use: $ cartesian_config.py tests.cfg On Thu, Aug 18, 2016 at 11:50 AM, Andrei Stepanov <astep...@redhat.com> wrote: > Hello. > > We have now broken cartesian_config.py. > > It was

[Avocado-devel] Broken cartesian_config by commit 81c6ce860b2f625e

2016-08-18 Thread Andrei Stepanov
Hello. We have now broken cartesian_config.py. It was broken by: commit 81c6ce860b2f625ec31533779c479cf9bf14af38 Author: Xu Tian Date: Mon May 23 15:10:48 2016 +0800 virttest.cartesian_config: enable postfix_parse postfix string '_fixed', '_max' and '_min'

[Avocado-devel] Pull req 581 for avocado-vt

2016-07-20 Thread Andrei Stepanov
https://github.com/avocado-framework/avocado-vt/pull/581 1. Can I ask you to accept this pull request ? 2. Do we have policy for acceptance Pull requests for avocado-vt? This pull request is month old. 3. What do I need to do to accept pull requests more quickly? Sometimes, pull requests becomes

[Avocado-devel] Wrong subtests.cfg

2016-06-07 Thread Andrei Stepanov
Hi. Today I got wrong subtests.cfg This file is auto-generated. Now it holds snippets from another *.cfg files. Those are not supposed to be a part of subtests.cfg. Any ideas? ___ Avocado-devel mailing list Avocado-devel@redhat.com

[Avocado-devel] Processing "info_xxx" field for avocado-vt tests.

2016-05-05 Thread Andrei Stepanov
Hello. I want to propose to add to avocado-vt "info_xxx" field. As you know, avocado-vt tests use cartesian configs. Test is a dictionary with combination of entries from different variants. Test name is something like: Selection1.Selection2.Selection3 variants: -selection1 info

[Avocado-devel] remote_commander

2016-03-24 Thread Andrei Stepanov
Hello. I know that "remote_commander" is not widely used. There is only one example how to use it. But, anyway... I do not understand one approach. To be exact : why are pickable-commands encoded with Base64? What is wrong about sending raw Python strings? Thanks!