Buffer I/O error on dev x, logical block 0, lost sync page write

kalingolas · 11-14-2019, 01:13 PM

Hello,

this is my first topic on this forum, so dont be mad with me if anything is wrong with it. Even with grammar ;-) (not native speaker)

I'm running an ESXI Server with virtualized Debian 10 Linux.
On this virtual machine a USB3 device (HW Raid device) is attached running a raid5 with 5x4TB.

Irregular the device falls back into readonly mode.

Code:

[Mo Okt 28 19:41:35 2019] Aborting journal on device dm-3-8.
[Mo Okt 28 19:41:35 2019] Buffer I/O error on dev dm-3, logical block 1953005568, lost sync page write
[Mo Okt 28 19:41:35 2019] JBD2: Error -5 detected when updating journal superblock for dm-3-8.
[Mo Okt 28 19:41:35 2019] Buffer I/O error on dev dm-3, logical block 0, lost sync page write
[Mo Okt 28 19:41:35 2019] EXT4-fs error (device dm-3): ext4_journal_check_start:56: Detected aborted journal
[Mo Okt 28 19:41:35 2019] EXT4-fs (dm-3): Remounting filesystem read-only
[Mo Okt 28 19:41:35 2019] EXT4-fs (dm-3): previous I/O error to superblock detected

All harddisk are fine and are checked with their smart status.
After remounting them, all is working fine again for time x.

As Raid Cage i use a Sharkoon Raid Station. I allready changed the USB3 cable.

I guess it could be the chipset on the raid cage. it is a JMicron Chip JMB 393.

Anyone got an idea what i can do?

Kind regards
Michael

business_kid · 11-16-2019, 03:24 AM

Welcome to LQ

There's a large mismatch between disk speed & usb speed, even usb3. Some errors are normal. Does the disk work? Does it work internally in a pc? Does a usb key work? Divide & conquer

syg00 · 11-16-2019, 04:26 AM

I too was concerned with USB write rates and buffering. But if it's a true hardware RAID card it will present as a single device. However I also see device manager involvement - that's another level of (block layer) emulation on top of the VMWare virtualization.
What I/O scheduler does the USB use ?. Also let's see a "lsblk -f" as a start.