** Description changed:

  There was a code change in the NFS write code that went into the 6.11
  Kernel mainline. This bug was reported:
  https://bugzilla.kernel.org/show_bug.cgi?id=219508
  
  The symptoms of this are that when there are a lot of writes to a
  mounted NFS share, especially through a VM or Docker, a soft lockup will
  occur. Based on the reverted commit this is because the assumptions made
  in the newer write code didn't account for a failed or short write.
  Ultimately this would cause an infinite loop where the NFS write never
  could release.
  
  The offending commit was here:
  
https://github.com/torvalds/linux/commit/b571cfcb9dcac187c6d967987792d37cb0688610#diff-
  fc1b4de72479f94c3f3a9d6f626078f5392d88be93cfce08bb32883ac7a91a97L169
  
  This was recently reverted:
  
https://github.com/torvalds/linux/commit/66f9dac9077c9c063552e465212abeb8f97d28a7
  
  This has been applied to the latest mainline 6.11.11 patch as well. This
- reversion was not included in the recent kernel release.
+ reversion was not included in the recent ubunutu 6.11.0-12-generic
+ kernel release.
  
  Description:  Ubuntu 24.10
  Release:      24.10

** Description changed:

  There was a code change in the NFS write code that went into the 6.11
  Kernel mainline. This bug was reported:
  https://bugzilla.kernel.org/show_bug.cgi?id=219508
  
  The symptoms of this are that when there are a lot of writes to a
  mounted NFS share, especially through a VM or Docker, a soft lockup will
  occur. Based on the reverted commit this is because the assumptions made
  in the newer write code didn't account for a failed or short write.
  Ultimately this would cause an infinite loop where the NFS write never
  could release.
+ 
+ On two of my machines that run routine backups of my homelab, there were
+ multiple lockups per week. Re-compiling the 6.11.9 Kernel without the
+ offending code ahs resolved the soft lockup in one of the machines. The
+ second continues to lock-up at least once a week. It's hard to reproduce
+ consistently because you'd need to recreate the condition of a short
+ write or network failure after the first block completes.
  
  The offending commit was here:
  
https://github.com/torvalds/linux/commit/b571cfcb9dcac187c6d967987792d37cb0688610#diff-
  fc1b4de72479f94c3f3a9d6f626078f5392d88be93cfce08bb32883ac7a91a97L169
  
  This was recently reverted:
  
https://github.com/torvalds/linux/commit/66f9dac9077c9c063552e465212abeb8f97d28a7
  
  This has been applied to the latest mainline 6.11.11 patch as well. This
  reversion was not included in the recent ubunutu 6.11.0-12-generic
  kernel release.
  
  Description:  Ubuntu 24.10
  Release:      24.10

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2091789

Title:
  Soft Lock Ups in NFS Write in Linux Kernel

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2091789/+subscriptions


-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to