Short news from churro, aka i18n.debian.net.... The machine is back online. Indeed, the problem was a kernel crash with "xfs: attempt to write beyond end of device" errors. I suspect a problem that I had on my own servers : one of the disks being badly partitioned with an incorrect partition table leading the the system keeping track of a disk that is bigger than it really is. One of the disks (where we already have records of XFS errors) is used for local backups using rsnapshot and I suspect it is the culprit. If that's the case, solving the problem is easy : just repartition that disk and recreate the file system on it as the disk doesn't have vital data (only backups). Indeed, at the moment I write this, the disk is unmounted. --
Attachment:
signature.asc
Description: Digital signature