2018-07-04 16:51:45 |
dann frazier |
bug |
|
|
added bug |
2018-07-04 17:00:05 |
Ubuntu Kernel Bot |
linux (Ubuntu): status |
New |
Incomplete |
|
2018-07-05 14:20:48 |
Joseph Salisbury |
linux (Ubuntu): importance |
Undecided |
Medium |
|
2018-07-05 14:21:02 |
Joseph Salisbury |
tags |
|
kernel-da-key |
|
2018-07-05 14:21:08 |
Joseph Salisbury |
linux (Ubuntu): status |
Incomplete |
Triaged |
|
2018-07-05 14:21:26 |
Joseph Salisbury |
nominated for series |
|
Ubuntu Bionic |
|
2018-07-05 14:21:26 |
Joseph Salisbury |
bug task added |
|
linux (Ubuntu Bionic) |
|
2018-07-05 14:21:33 |
Joseph Salisbury |
linux (Ubuntu Bionic): status |
New |
Triaged |
|
2018-07-05 14:21:36 |
Joseph Salisbury |
linux (Ubuntu Bionic): importance |
Undecided |
Medium |
|
2018-07-05 14:21:42 |
Joseph Salisbury |
tags |
kernel-da-key |
bionic kernel-da-key |
|
2018-07-05 21:53:58 |
dann frazier |
description |
We're seeing a very reproducible regression in the bionic kernel triggered by the stress-ng chdir test performed by the Ubuntu certification suite. Platform is a HiSilicon D05 arm64 server, but we don't have reason to believe it is platform specific at this time.
[Test Case]
$ sudo apt-add-repository -y ppa:hardware-certification/public
$ sudo apt install -y canonical-certification-server
$ sudo mkfs.ext4 /dev/sda1 (Obviously, this should not be your root disk!!)
$ sudo /usr/lib/plainbox-provider-checkbox/bin/disk_stress_ng sda --base-time 240 --really-run
This test runs a series of stress-ng tests against /dev/sda, and fails on the "chdir" test. To speed up reproduction, reduce the test list to just "chdir" in the disk_stress_ng script. Attempts to reproduce this directly with stress-ng have failed - presumably because of other environment setup that this script performs (e.g. setting aio-max-nr to 524288).
Our reproduction test is to use a non-root disk because it can lead to corruption, and mkfs.ext4'ing the partition just before running the test, to get to a pristine fs state.
I bisected this down to the following commit:
commit 555bc9b1421f10d94a1192c7eea4a59faca3e711
Author: Theodore Ts'o <tytso@mit.edu>
Date: Mon Feb 19 14:16:47 2018 -0500
ext4: don't update checksum of new initialized bitmaps
BugLink: http://bugs.launchpad.net/bugs/1773233
commit 044e6e3d74a3d7103a0c8a9305dfd94d64000660 upstream. |
We're seeing a very reproducible regression in the bionic kernel triggered by the stress-ng chdir test performed by the Ubuntu certification suite. We see this on both the HiSilicon D05 arm64 server and the HiSilicon D06 arm64 server. We have been unable to reproduce on other servers so far.
[Test Case]
$ sudo apt-add-repository -y ppa:hardware-certification/public
$ sudo apt install -y canonical-certification-server
$ sudo mkfs.ext4 /dev/sda1 (Obviously, this should not be your root disk!!)
$ sudo /usr/lib/plainbox-provider-checkbox/bin/disk_stress_ng sda --base-time 240 --really-run
This test runs a series of stress-ng tests against /dev/sda, and fails on the "chdir" test. To speed up reproduction, reduce the test list to just "chdir" in the disk_stress_ng script. Attempts to reproduce this directly with stress-ng have failed - presumably because of other environment setup that this script performs (e.g. setting aio-max-nr to 524288).
Our reproduction test is to use a non-root disk because it can lead to corruption, and mkfs.ext4'ing the partition just before running the test, to get to a pristine fs state.
I bisected this down to the following commit:
commit 555bc9b1421f10d94a1192c7eea4a59faca3e711
Author: Theodore Ts'o <tytso@mit.edu>
Date: Mon Feb 19 14:16:47 2018 -0500
ext4: don't update checksum of new initialized bitmaps
BugLink: http://bugs.launchpad.net/bugs/1773233
commit 044e6e3d74a3d7103a0c8a9305dfd94d64000660 upstream. |
|
2018-07-10 20:19:21 |
dann frazier |
bug |
|
|
added subscriber Ike Panhc |
2018-07-10 21:54:05 |
dann frazier |
description |
We're seeing a very reproducible regression in the bionic kernel triggered by the stress-ng chdir test performed by the Ubuntu certification suite. We see this on both the HiSilicon D05 arm64 server and the HiSilicon D06 arm64 server. We have been unable to reproduce on other servers so far.
[Test Case]
$ sudo apt-add-repository -y ppa:hardware-certification/public
$ sudo apt install -y canonical-certification-server
$ sudo mkfs.ext4 /dev/sda1 (Obviously, this should not be your root disk!!)
$ sudo /usr/lib/plainbox-provider-checkbox/bin/disk_stress_ng sda --base-time 240 --really-run
This test runs a series of stress-ng tests against /dev/sda, and fails on the "chdir" test. To speed up reproduction, reduce the test list to just "chdir" in the disk_stress_ng script. Attempts to reproduce this directly with stress-ng have failed - presumably because of other environment setup that this script performs (e.g. setting aio-max-nr to 524288).
Our reproduction test is to use a non-root disk because it can lead to corruption, and mkfs.ext4'ing the partition just before running the test, to get to a pristine fs state.
I bisected this down to the following commit:
commit 555bc9b1421f10d94a1192c7eea4a59faca3e711
Author: Theodore Ts'o <tytso@mit.edu>
Date: Mon Feb 19 14:16:47 2018 -0500
ext4: don't update checksum of new initialized bitmaps
BugLink: http://bugs.launchpad.net/bugs/1773233
commit 044e6e3d74a3d7103a0c8a9305dfd94d64000660 upstream. |
[Impact]
We're seeing a very reproducible regression in the bionic kernel triggered by the stress-ng chdir test performed by the Ubuntu certification suite. We see this on both the HiSilicon D05 arm64 server and the HiSilicon D06 arm64 server. We have been unable to reproduce on other servers so far.
[Test Case]
$ sudo apt-add-repository -y ppa:hardware-certification/public
$ sudo apt install -y canonical-certification-server
$ sudo mkfs.ext4 /dev/sda1 (Obviously, this should not be your root disk!!)
$ sudo /usr/lib/plainbox-provider-checkbox/bin/disk_stress_ng sda --base-time 240 --really-run
This test runs a series of stress-ng tests against /dev/sda, and fails on the "chdir" test. To speed up reproduction, reduce the test list to just "chdir" in the disk_stress_ng script. Attempts to reproduce this directly with stress-ng have failed - presumably because of other environment setup that this script performs (e.g. setting aio-max-nr to 524288).
Our reproduction test is to use a non-root disk because it can lead to corruption, and mkfs.ext4'ing the partition just before running the test, to get to a pristine fs state.
[Fix]
I bisected this down to the following commit:
commit 555bc9b1421f10d94a1192c7eea4a59faca3e711
Author: Theodore Ts'o <tytso@mit.edu>
Date: Mon Feb 19 14:16:47 2018 -0500
ext4: don't update checksum of new initialized bitmaps
BugLink: http://bugs.launchpad.net/bugs/1773233
commit 044e6e3d74a3d7103a0c8a9305dfd94d64000660 upstream.
Reverting that fixes the problem. Meanwhile, a proposed fix has been posted upstream:
https://www.spinics.net/lists/linux-ext4/msg61578.html
[Regression Risk] |
|
2018-07-11 08:03:49 |
Kleber Sacilotto de Souza |
linux (Ubuntu Bionic): status |
Triaged |
Fix Committed |
|
2018-07-11 08:30:19 |
Ike Panhc |
attachment added |
|
consolelog.txt https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1780137/+attachment/5162302/+files/consolelog.txt |
|
2018-07-11 18:46:34 |
Seth Forshee |
linux (Ubuntu): status |
Triaged |
Fix Committed |
|
2018-07-12 12:01:18 |
Brad Figg |
tags |
bionic kernel-da-key |
bionic kernel-da-key verification-needed-bionic |
|
2018-07-12 16:41:39 |
dann frazier |
tags |
bionic kernel-da-key verification-needed-bionic |
bionic kernel-da-key verification-done-bionic |
|
2018-07-20 15:39:45 |
Launchpad Janitor |
linux (Ubuntu Bionic): status |
Fix Committed |
Fix Released |
|
2018-07-26 05:13:52 |
Launchpad Janitor |
linux (Ubuntu): status |
Fix Committed |
Fix Released |
|