[Bug 264141] nvme(4): Heavy load to SSD wedges 13.1 system: Controller in fatal status, resetting ... Resetting controller due to a timeout and possible hot unplug.

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=264141

--- Comment #24 from dgilbert@eicat.ca ---
(In reply to Warner Losh from comment #23)

... so my messages file back from yesterday says:

Jun 15 22:46:17 vr kernel: nvme0: Allocated 200MB host memory buffer

... the trouble starts with:

Jun 17 04:28:00 vr kernel: nvme0: RECOVERY_START 459377299057992 vs
459376982792873
Jun 17 04:28:00 vr kernel: nvme0: RECOVERY_START 459377569149703 vs
459376982792873
Jun 17 04:28:00 vr kernel: nvme0: RECOVERY_START 459378788907460 vs
459376987085573
Jun 17 04:28:00 vr kernel: nvme0: RECOVERY_START 459378788907460 vs
459376995708726
Jun 17 04:28:00 vr kernel: nvme0: RECOVERY_START 459379058646386 vs
459376995708726
Jun 17 04:28:00 vr kernel: nvme0: Controller in fatal status, resetting
Jun 17 04:28:00 vr kernel: nvme0: Resetting controller due to a timeout and
possible hot unplug.
Jun 17 04:28:00 vr kernel: nvme0: RECOVERY_WAITING
Jun 17 04:28:00 vr kernel: nvme0: resetting controller
Jun 17 04:28:00 vr kernel: nvme0: waiting
Jun 17 04:28:00 vr kernel: nvme0: failing outstanding i/o
Jun 17 04:28:00 vr kernel: nvme0: READ sqid:1 cid:118 nsid:1 lba:2985560513
len:256
Jun 17 04:28:00 vr kernel: nvme0: ABORTED - BY REQUEST (00/07) sqid:1 cid:118
cdw0:0
Jun 17 04:28:00 vr kernel: nvme0: failing outstanding i/o

... which repeats ad-nausium.

At least in my case, this was a zfs-cache drive ... so I have a log.

... but sadly, I returned the drive already.  I can't afford a $300 non-working
bit of hardware.  Had to return and order a new one.

I can tell you that another NVMe in it's place does not _ever_ do this.  And I
can detail that I RMA'd the first one that did this, and the replacement did
the same thing.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
You are the assignee for the bug.