Hello all, I'm currently running FC4/64 on software RAID 5 setup running on 4 36GB SCSI drives. Each drive partition tables looks like this: /dev/sdX1: 100MB, RAID autostart, active. /dev/sdX2: 30GB, RAID autostart. /dev/sdX3: 4GB, empty. /dev/md0 is my boot, and it uses all the 100MB partitions on all the drives. /dev/md1 is my LVM, and it uses all the 32GB partitions on all drives. I decided I need more space on my LVM, so I slowly resized all the partitions, but failing each sdX2 partition, rebuilding to each all the remaining space (~35GB), and re-adding it to the array. After an hour, all the disk were rebuilt. Now, my question is simple. A. How do I instruct the RAID5 array (using mdadm?) to use all the available space on each drive. (Read: Use the extra 4GB on each drive). B. Once done, is FC5's pvresize support considered stable? Can I just issue pvresize /dev/md1 and it'll work? While the contents of the system are backed up... I rather do it on-line. Thanks. Gilboa
