Fix sent to kernel team mailing list for review:
https://lists.ubuntu.com/archives/kernel-team/2019-May/100990.html
** Description changed:
SRU Bionic
== SRU Justification ==
Intensive I/O such as performed by ZFS send/receive can cause hangs because
of monopolization of the global system_taskq. The outcome is that z_zvol hung
tasks occur and I/O gets blocked.
== Fix ==
Upstream ZFS commit 77d8a0f1a4d0b2f59cee63088f7987cb38e66538 ("Fix hung
- z_zvol tasks during 'zfs receive'" fixes this issue, it adds a dedicated
- per-pool prefetch taskq that prevents the traverse code from
+ z_zvol tasks during 'zfs receive'") fixes this issue, it adds a
+ dedicated per-pool prefetch taskq that prevents the traverse code from
monopolizing the global (and limited) system_taskq by inappropriately
scheduling long running tasks on it. This fixes the z_zvol hung tasks. A
trivial backport is required for Bionic ZFS.
== Testcase ==
Perform large send/receives. Occasionally they lock up. With the fix,
this issue is addressed and no more lockups occur. Also must pass the
full ZFS ubuntu autotest tests to prove no regressions occur.
== Regression Potential ==
This fix adds more per-pool prefetch taskq's so we have more kernel
resources being used. There is therefore a very small risk that this may
impact ZFS running on memory and CPU constrained systems. However, the
fix is small, has been upstream for a while and is in Cosmic+ releases
and has not caused any regressions, so I think this is a relatively safe
fix.
------
I have experienced the problems fixed by this commit
https://github.com/zfsonlinux/zfs/pull/7343 a few times on my NAS. The
system hangs completely when it occurs. It looks like 0.7.9 brings other
interesting bug fixes that potentially freeze the system.
--
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to zfs-linux in Ubuntu.
https://bugs.launchpad.net/bugs/1772412
Title:
zfs 0.7.9 fixes a bug (https://github.com/zfsonlinux/zfs/pull/7343)
that hangs the system completely
Status in zfs-linux package in Ubuntu:
In Progress
Bug description:
SRU Bionic
== SRU Justification ==
Intensive I/O such as performed by ZFS send/receive can cause hangs because
of monopolization of the global system_taskq. The outcome is that z_zvol hung
tasks occur and I/O gets blocked.
== Fix ==
Upstream ZFS commit 77d8a0f1a4d0b2f59cee63088f7987cb38e66538 ("Fix
hung z_zvol tasks during 'zfs receive'") fixes this issue, it adds a
dedicated per-pool prefetch taskq that prevents the traverse code from
monopolizing the global (and limited) system_taskq by inappropriately
scheduling long running tasks on it. This fixes the z_zvol hung tasks.
A trivial backport is required for Bionic ZFS.
== Testcase ==
Perform large send/receives. Occasionally they lock up. With the fix,
this issue is addressed and no more lockups occur. Also must pass the
full ZFS ubuntu autotest tests to prove no regressions occur.
== Regression Potential ==
This fix adds more per-pool prefetch taskq's so we have more kernel
resources being used. There is therefore a very small risk that this
may impact ZFS running on memory and CPU constrained systems.
However, the fix is small, has been upstream for a while and is in
Cosmic+ releases and has not caused any regressions, so I think this
is a relatively safe fix.
------
I have experienced the problems fixed by this commit
https://github.com/zfsonlinux/zfs/pull/7343 a few times on my NAS. The
system hangs completely when it occurs. It looks like 0.7.9 brings
other interesting bug fixes that potentially freeze the system.
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/zfs-linux/+bug/1772412/+subscriptions
--
Mailing list: https://launchpad.net/~kernel-packages
Post to : kernel-packages@lists.launchpad.net
Unsubscribe : https://launchpad.net/~kernel-packages
More help : https://help.launchpad.net/ListHelp
