On 2026-09-12 12:12, Andy Smith wrote:
Hi,
On Sat, Sep 12, 2026 at 12:03:24PM -0400, Gary Dale wrote:
On 2026-09-12 11:00, Andy Smith wrote:
# wipefs -an /dev/sde1
say?
wipefs seems to do nothing. Nothing is reported and, after removing the "n",
the problem still remains.
Okay, nothing found on there that might be confusing things, then.
2026-09-12T11:47:08.931954-04:00 TheLibrarian kernel: I/O error, dev sde,
sector 2058 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 2
2026-09-12T11:47:08.931956-04:00 TheLibrarian kernel: Buffer I/O error on
dev sde1, logical block 1, async page read
This is a pretty low level error and I'd be a little worried.
Are you able to read the whole disk with dd without error?
# dd if=/dev/sde of=/dev/null
And smartctl long self-test comes out okay?
# smartctl -t long /dev/sde
Thanks,
Andy
OK, I tried a reboot. This time there was less drama as the array came
up properly. However, the drive letters switched - that probably
explains the SMART errors changing drives - I also rebooted a week ago.
I'm running the long test now but that will not complete until tomorrow.
I've been getting intermittent SMART errors for at least a couple of
years. And at one point a drive would sometimes vanish but a reboot
would bring it back. I've been suspicious of the drives for some time
but never enough to actually tracking down the fault and replacing
whichever physical drive had the problems.
This behaviour, of the drive being half there, is new though. The drive
identified as [7] shows up in one array but not the other.
I'll check in again tomorrow after the SMART test completes.
