FreeNAS 11.1 ZFS panic as replication target with encryption

Status
Not open for further replies.

bittenoff

Dabbler
Joined
Nov 26, 2017
Messages
20
When doing a replication from FreeNAS 9.3 to 11.1 and where the target volume has been enabled for encryption, ZFS reliably panics:

panic()
dblfault_handler()
Xdblfault()
--- trap 0x17
vdev_queue_io_to_issue()
vdev_queue_io_done()
zio_vdev_io_done()
zio_execute()
zio_vdev_io_start()
zio_execute()
...
[repeats of zio_execute, zio_vdev_io_start, zio_execute, vdev_queue_io_done]
...
taskqueue_run_locked()
taskqueue_thread_loop()
fork_exit()
fork_trampoline()

There is a massive amount of recursion of zio_execute ... vdev_queue_io_done. Maybe 4 or 5 console screens worth.

See:
https://forums.freenas.org/index.ph...ed-encryption-with-freenas.59367/#post-423520
 
D

dlavigne

Guest
Please create a report at bugs.freenas.org and post the issue number here.
 

bittenoff

Dabbler
Joined
Nov 26, 2017
Messages
20
https://redmine.ixsystems.com/issues/27755

There are two possibilities that I can think of here:
* there's something in the ZFS filesystem on the sending server that the receiver doesn't like
* the format of the ZFS send is activating a bug in the ZFS recv on the destination system.

I say this because it appears that the receiver seems to always panic after it receives X amount of data.
 
Status
Not open for further replies.
Top