QA (public) » openQA Project (public)

I added the two related tickets regarding "external encoder" not assuming that any external encoder is used but that related code changes could have caused this.

retriggered jobs seem fine, e.g. https://openqa.suse.de/admin/workers/669 and https://openqa.suse.de/tests/4629502 are ok

I conducted the DB search that I added to the ticket description to find further occurrences and found

on o3 and osd (in this order):

   id    |     t_finished      | state |   result   | test  |                                             reason                                              |     host      
---------+---------------------+-------+------------+-------+-------------------------------------------------------------------------------------------------+---------------
 1409732 | 2020-09-28 01:12:01 | done  | incomplete | gnome | backend died: Encoder not accepting data at /usr/lib/os-autoinst/backend/baseclass.pm line 157. | openqaworker4
(1 row)

   id    |     t_finished      | state |   result   |                     test                      |                                             reason                                              |     host      
---------+---------------------+-------+------------+-----------------------------------------------+-------------------------------------------------------------------------------------------------+---------------
 4744792 | 2020-09-29 00:20:34 | done  | incomplete | mru-iscsi_client_normal_auth_backstore_fileio | backend died: Encoder not accepting data at /usr/lib/os-autoinst/backend/baseclass.pm line 157. | openqaworker6
 4742850 | 2020-09-28 00:29:22 | done  | incomplete | qam-rsync-client                              | backend died: Encoder not accepting data at /usr/lib/os-autoinst/backend/baseclass.pm line 157. | openqaworker5
 4628195 | 2020-09-02 02:31:20 | done  | incomplete | mau-qa_kernel_lynis                           | backend died: Encoder not accepting data at /usr/lib/os-autoinst/backend/baseclass.pm line 157. | openqaworker5
(3 rows)

One example is https://openqa.opensuse.org/tests/1409732 with log details:

[2020-09-28T03:11:04.065 CEST] [debug] tests/installation/bootloader.pm:57 called bootloader_setup::bootmenu_default_params -> lib/bootloader_setup.pm:478 called bootloader_setup::type_boot_parameters -> lib/bootloader_setup.pm:1330 called testapi::type_string
[2020-09-28T03:11:04.065 CEST] [debug] <<< testapi::type_string(string="  Y2DEBUG=1 vga=791 video=1024x768 plymouth.ignore-serial-consoles console=ttyS0 console=tty linuxrc.log=/dev/ttyS0 linuxrc.core=/dev/ttyS0 linuxrc.debug=4,trace reboot_timeout=0  ", max_interval=4, wait_screen_changes=0, wait_still_screen=0, timeout=30, similarity_level=47)
[2020-09-28T03:11:56.918 CEST] [debug] Backend process died, backend errors are reported below in the following lines:
Encoder not accepting data at /usr/lib/os-autoinst/backend/baseclass.pm line 157.

Notice the timestamp difference between mm:ss 11:04.1 and mm:ss 11:56.9 .
The relevant code in backend/baseclass.pm is:

sub _write_buffered_data_to_file_handle {
    my ($self, $program_name, $array_of_buffers, $fh) = @_;

    # write as much data as possible (this is called when $fh is ready to write)
    my $data         = shift @$array_of_buffers;
    my $data_written = $fh->syswrite($data);
    die "$program_name not accepting data" unless defined $data_written;

    # put remaining data it back into the queue
    unshift @$array_of_buffers, substr($data, $data_written) unless $data_written == length($data);

What we could do is check the status of the videoencoder process, e.g. output the PID and check if the process is still running.
Note that "Encoder" means the internal encoder, different than "External encoder" which we would see if any custom configurable video encoder would fail here.

I found out that syswrite sets the errno variable in case of errors, see https://perldoc.perl.org/functions/syswrite . So we should output this to have more details in the case of problems:
https://github.com/os-autoinst/os-autoinst/pull/1550

PR merged. I guess we should wait couple of weeks and check if we see the error recorded. Then we might have more information.

$ for i in o3 osd; do ssh $i "sudo -u geekotest psql openqa -c \"select jobs.id,t_finished,state,result,test,reason,host from jobs, comments, workers where t_finished >= '2020-01-01' and jobs.assigned_worker_id = workers.id and jobs.id = comments.job_id and comments.text ~ 'poo#70873' order by t_finished desc limit 20;\""; done

   id    |     t_finished      | state |   result   | test  |                                             reason                                              |     host      
---------+---------------------+-------+------------+-------+-------------------------------------------------------------------------------------------------+---------------
 1409732 | 2020-09-28 01:12:01 | done  | incomplete | gnome | backend died: Encoder not accepting data at /usr/lib/os-autoinst/backend/baseclass.pm line 157. | openqaworker4
(1 row)

   id    |     t_finished      | state |   result   |                     test                      |                                             reason                                              |     host      
---------+---------------------+-------+------------+-----------------------------------------------+-------------------------------------------------------------------------------------------------+---------------
 4744792 | 2020-09-29 00:20:34 | done  | incomplete | mru-iscsi_client_normal_auth_backstore_fileio | backend died: Encoder not accepting data at /usr/lib/os-autoinst/backend/baseclass.pm line 157. | openqaworker6
 4742850 | 2020-09-28 00:29:22 | done  | incomplete | qam-rsync-client                              | backend died: Encoder not accepting data at /usr/lib/os-autoinst/backend/baseclass.pm line 157. | openqaworker5
 4628195 | 2020-09-02 02:31:20 | done  | incomplete | mau-qa_kernel_lynis                           | backend died: Encoder not accepting data at /usr/lib/os-autoinst/backend/baseclass.pm line 157. | openqaworker5
(3 rows)

so no more occurences since more than a month