Project

General

Profile

action #19544

[tools] Checksum/size comparison of autoinst.xml FAILED

Added by nicksinger over 4 years ago. Updated over 3 years ago.

Status:
Resolved
Priority:
High
Assignee:
Category:
Concrete Bugs
Target version:
-
Start date:
2017-06-02
Due date:
% Done:

0%

Estimated time:
Difficulty:

Description

https://openqa.suse.de/tests/975720 fails as incomplete. Everything inside the autoinst-log.txt looks fine except this line:

+++ worker notes +++
end time: 2017-06-01 22:48:25
result: done
uploading consoletest_setup-loadavg_consoletest_setup.txt
uploading consoletest_setup-psaxf.log
uploading install_and_reboot-y2logs.tar.bz2
uploading autoinst.xml
Checksum comparison (actual:expected) 4294967295:1179459072 with size (actual:expected) 0:51985
Checksum/size comparison of autoinst.xml FAILED
uploading vars.json
uploading serial0.txt
uploading autoinst-log.txt

First occasion I saw with this error and I'm not 100% certain if the failed log is really the reason for the incomplete failure.
Also a hard failure (with incomplete) just because of missing logs/assets seems exaggerated to me. Maybe an additional notice (which would have to be introduced) about failing "post test handling" makes such problems more obvious than a incomplete test - but that is just an idea of mine.


Related issues

Related to openQA Project - action #30352: Disk upload failuresResolved2018-01-15

History

#1 Updated by nicksinger over 4 years ago

  • Related to action #19194: [tools] figure out why QA-Power8-5-kvm fails more than malbec added

#2 Updated by okurz over 4 years ago

  • Status changed from New to In Progress
  • Priority changed from Normal to High

nicksinger: this was what rbrown told in before: ppc64le workers run by default in tmpfs. The actual size is 0 so it could not write the file

#3 Updated by okurz over 4 years ago

  • Assignee set to RBrownSUSE

rbrown wanted to move the image creation jobs to ppc64le-no-tmpfs

#4 Updated by nicksinger over 4 years ago

Hm… Now that you said it this seems to be the most likely cause of this.
But what I then cannot explain is, why this succeeds (EXIT 0 means successful to me):

22:40:42.7744 18211 running nice ionice qemu-img convert -O qcow2 raid/l1 assets_public/SLES-12-SP3-ppc64le-Build0409-gnome.qcow2 -c
18211: EXIT 0

while the only problem is with autoinst.xml.
IMHO we need better logging here.

#5 Updated by RBrownSUSE over 4 years ago

  • Related to deleted (action #19194: [tools] figure out why QA-Power8-5-kvm fails more than malbec)

#6 Updated by RBrownSUSE over 4 years ago

  • Assignee changed from RBrownSUSE to szarate

nothing to do with tmpfs, other tmpfs tests okay, unlinking and unassigning, santi will look at it

#7 Updated by szarate over 4 years ago

  • Project changed from openQA Tests to openQA Project
  • Category changed from Infrastructure to Concrete Bugs
  • Target version set to Milestone 8

#8 Updated by okurz over 4 years ago

sle-12-SP3-Server-DVD-x86_64-Build0410-sles12_textmode_create_hdd@64bit is another one. Ok, definitely not ppc64le tmpfs related :-)

#9 Updated by okurz over 4 years ago

https://openqa.suse.de/tests/990990 on a standard ppc64le worker

#10 Updated by okurz over 4 years ago

https://openqa.suse.de/tests/1019642 is another example on standard x86_64. I think cryptlvm is more prone to fail because likely the image is bigger?

#11 Updated by szarate over 4 years ago

  • Assignee deleted (szarate)
  • Target version deleted (Milestone 8)

It might be that somehow the chksum call is timing out due to nfs being too busy or so... But that's just a theory after talking to coolo when I came across this problem with the seattles (that basically had no nfs mount, and the checksums were failing all the time).

I guess splitting the checksum and removing the dependency with nfs (except from the repos), could be a wise thing to do.

Also, removing the milestone.

#12 Updated by szarate over 4 years ago

  • Status changed from In Progress to Feedback

#13 Updated by szarate almost 4 years ago

#14 Updated by coolo over 3 years ago

  • Status changed from Feedback to Resolved
  • Assignee set to EDiGiacinto

resolved by switching to multi chunk uploader

Also available in: Atom PDF