I’ve been running a stable Renegade (ROC-RK3328-CC) for 1.5 years now with Debian Bullseye 11. About a week or so ago a service running java started to restart every few minutes. I thought there was a corrupted SD card which I replaced but a second card behaves the same way. I looked in the dmesg logs and the service restart corresponds to high numbers of BTRFS errors. Is the built in memory or another part of the hardware starting to fail? Is there a way to isolate where the problem is?
May 21 08:09:39 mylibre kernel: [ 189.822008] BTRFS warning (device mmcblk1p2): csum failed root 5 ino 100901 off 81920 csum 0xd0ceeeb2 expected csum 0xe317e021 mirror 1
May 21 08:09:39 mylibre kernel: [ 189.822039] BTRFS error (device mmcblk1p2): bdev /dev/mmcblk1p2 errs: wr 0, rd 0, flush 0, corrupt 13471, gen 0
May 21 08:09:39 mylibre kernel: [ 189.828030] BTRFS warning (device mmcblk1p2): csum failed root 5 ino 100901 off 0 csum 0xf532379a expected csum 0x3510f919 mirror 1
May 21 08:09:39 mylibre kernel: [ 189.828064] BTRFS error (device mmcblk1p2): bdev /dev/mmcblk1p2 errs: wr 0, rd 0, flush 0, corrupt 13472, gen 0
May 21 08:09:43 mylibre kernel: [ 194.452236] btrfs_print_data_csum_error: 636 callbacks suppressed
May 21 08:09:43 mylibre kernel: [ 194.452258] BTRFS warning (device mmcblk1p2): csum failed root 5 ino 100746 off 0 csum 0x020c4103 expected csum 0x83e211dd mirror 1
May 21 08:09:43 mylibre kernel: [ 194.452278] btrfs_dev_stat_inc_and_print: 636 callbacks suppressed
May 21 08:09:43 mylibre kernel: [ 194.452282] BTRFS error (device mmcblk1p2): bdev /dev/mmcblk1p2 errs: wr 0, rd 0, flush 0, corrupt 14109, gen 0
May 21 08:09:43 mylibre kernel: [ 194.452328] BTRFS warning (device mmcblk1p2): csum failed root 5 ino 100746 off 4096 csum 0xdccb7f45 expected csum 0x136f90cd mirror 1