Project

General

Profile

Actions

action #17594

closed

[tools]Missing characters in the middle of type_string/assert_script_run (Ninja Keys)

Added by szarate about 7 years ago. Updated almost 7 years ago.

Status:
Resolved
Priority:
Urgent
Assignee:
Category:
Infrastructure
Target version:
-
Start date:
2017-03-07
Due date:
% Done:

0%

Estimated time:
Difficulty:

Description

observation

While running jobs on certain machines, sometimes there are tests that are failing with characters missing in the middle of a string.

steps to reproduce

Review tests that are failing, most tests where type_string is used, can create this problem.

problem

H1 Asset loading through NFS might cause delays in response of the SUT.
H2 Key events are not reaching the qemu at all
H3 Key events are reaching the qemu, qemu processes the event, and it reaches the qeue, but not the SUT itself (they are not reaching X).

suggestion

Example of failing jobs

https://openqa.suse.de/tests/805748#step/gedit/4
https://openqa.suse.de/tests/795241#step/sshxterm/5
https://openqa.suse.de/tests/796232#step/gnome_terminal/7

workaround

Retrigger the job.


Related issues 6 (0 open6 closed)

Related to openQA Project - action #17574: [tools]Add caching/syncing of assetsResolvedszarate2017-04-29

Actions
Related to openQA Project - action #17442: [tools]Bad performance or too many failures because of typing too soon/mistyping (send_keys/type_string)ResolvedRBrownSUSE2017-03-02

Actions
Related to openQA Project - action #16438: [tools]System stall during test run not detectedResolved2017-02-03

Actions
Related to openQA Project - action #18132: [tools] test fails to assert command because a wrong string was written to serial - (anti-ninja-key)Closed2017-03-29

Actions
Related to openQA Tests - action #17456: [qam][x11regressions]test fails sporadically in gedit_saveClosedmitiao2017-03-02

Actions
Related to openQA Tests - action #14838: [sles][functional] test fails in change_passwordRejectedSLindoMansilla2016-11-16

Actions
Actions #1

Updated by szarate about 7 years ago

  • Related to action #17574: [tools]Add caching/syncing of assets added
Actions #2

Updated by szarate about 7 years ago

  • Related to action #17442: [tools]Bad performance or too many failures because of typing too soon/mistyping (send_keys/type_string) added
Actions #3

Updated by szarate about 7 years ago

  • Description updated (diff)
Actions #4

Updated by okurz about 7 years ago

  • Category set to Infrastructure
Actions #5

Updated by RBrownSUSE about 7 years ago

  • Subject changed from Missing characters in the middle of type_string (Ninja Keys) to [tools]Missing characters in the middle of type_string (Ninja Keys)
Actions #6

Updated by RBrownSUSE about 7 years ago

  • Related to action #16438: [tools]System stall during test run not detected added
Actions #7

Updated by nicksinger about 7 years ago

here is another example were keys are missed in the middle of typing: https://progress.opensuse.org/issues/17594 (happens at module "textinfo")

Actions #8

Updated by mgriessmeier about 7 years ago

Might this be related?
seems to stop typing completely
https://progress.opensuse.org/issues/17888

Actions #12

Updated by nicksinger about 7 years ago

Also another one here:
https://openqa.suse.de/tests/833485

Actions #14

Updated by okurz about 7 years ago

  • Related to action #18132: [tools] test fails to assert command because a wrong string was written to serial - (anti-ninja-key) added
Actions #15

Updated by RBrownSUSE about 7 years ago

  • Status changed from New to Resolved
Actions #16

Updated by okurz about 7 years ago

If I understood correctly the following two things are supposed to have fixed this issue

os-autoinst change https://github.com/os-autoinst/os-autoinst/pull/743

commit aebf404
Author: Stephan Kulow coolo@kde.org
Date: Mon Mar 27 13:08:44 2017 +0200

Make it possible to define slow typing machines (#743)

After a lot of forth and back, setting this to 10 is the only cure
for power and arm :(

with defining VNC_TYPING_LIMIT=10 in https://openqa.suse.de/admin/machines for aarch64 and ppc64le machines

plus also os-autoinst-distri-opensuse change https://github.com/os-autoinst/os-autoinst-distri-opensuse/pull/2577

commit 709c99c
Author: Stephan Kulow coolo@kde.org
Date: Wed Mar 22 15:53:26 2017 +0100

Add an explicit run of the cron jobs (#2577)

Our data shows pretty clear that the tests type much more unreliable
around the time the first cron.weekly is called (*/15) - and it's
believed that the btrfs maintenance is killing responsiveness
as described in boo#1017461 (and others). So better run these
cron jobs at a given time to avoid random tests becoming unreproducible
Actions #17

Updated by okurz about 7 years ago

  • Related to action #17456: [qam][x11regressions]test fails sporadically in gedit_save added
Actions #18

Updated by SLindoMansilla almost 7 years ago

  • Related to action #14838: [sles][functional] test fails in change_password added
Actions #19

Updated by riafarov almost 7 years ago

  • Subject changed from [tools]Missing characters in the middle of type_string (Ninja Keys) to [tools]Missing characters in the middle of assert_script_run (Ninja Keys)
  • Status changed from Resolved to In Progress
  • Target version deleted (Milestone 6)

Reappeared for aarch64 : https://openqa.suse.de/tests/942993#step/yast2_lan_restart/309
in assert_script_run method.
After assert_script_run 'dig suse.com|grep \'status: NOERROR\''; call, "." is missing in the console.

Autoinst log in the attachment

Actions #20

Updated by szarate almost 7 years ago

  • Status changed from In Progress to Resolved

Closing because it's not related, this needs another ticket specific for aarch64

Actions #21

Updated by okurz almost 7 years ago

  • Subject changed from [tools]Missing characters in the middle of assert_script_run (Ninja Keys) to [tools]Missing characters in the middle of type_string/assert_script_run (Ninja Keys)
Actions

Also available in: Atom PDF