Quote:
Originally Posted by compurandom
If Davidfor is correct, it's technically not the reboot that fixes it, but the automatic filesystem repair that occurs during the reboot.
The implication here is that '31 is doing something wrong during the nickel->usb->nickel transition of the filesystem and corrupting it, and it needs a fsck to repair it. With luck, the damage is minor, and doesn't involve the database (or at least, not both databases).
|
I don't think so.. I've been able to reliably reproduce this behaviour when the only filesystem error that fsck can see/fix is the dirty bit.
I suspect that the database process gets stuck when the kobo switches to USB mode, (and hence, why the filesystem does not unmount cleanly in those times.) When the USB disconnects, the Kobo resumes, but no new data is getting recorded to the database.