Situation: I periodically run backups from a system running Debian "Squeeze" under kernel 2.6.32-5-amd64. The backups are written to removable hard drives, swapped out during the process as necessary. AS each drive is unmounted, removed, and replaced with a fresh drive, I run a script designed to clear any RAID superblock, partition the drive if it has any more or less than a single partition, check to see if the partition is formatted with xfs, and if not, format the partition. I've tried two different methods, both of which have significant limitations. Is there a better way? 1. Method 1 xfs_check "$Dspec"1 || mkfs.xfs "$Dspec"1 If the file system does not exist, this works fairly quickly and very reliably. If the file system does exist, however, it brings the system to its knees for an uncomfortable amount of time, and on at least one occasion, caused a kernel panic. 2. Method 2 xfs_repair -n "$Dspec"1 || mkfs.xfs "$Dspec"1 This works extremely well and very quickly if the file system exists, but if it does not, then the xfs_repair utility goes off trying to find a secondary location for the xfs superblock. It evidently does this by reading every sector of the partition. These partitions are between 1 Terabyte and 3 Terabytes in size, and even at more than 80 MBps, that takes a long while. How else should I approach this? -- To unsubscribe from this list: send the line "unsubscribe linux-xfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
