For a Linux ECS with a SCSI disk attached, if automatic SCSI disk attaching upon ECS startup is enabled in /etc/fstab and the disk drive letter (for example, /dev/sdb) is used, restarting the ECS may fail.
SCSI disk allocation is determined based on the ID of the slot accommodating the disk as well as the available drive letter in the ECS. Each time when a disk is attached to the ECS, an idle drive letter is automatically allocated in sequence. When the ECS starts, the disks are loaded in slot sequence. Therefore, a slot ID corresponds to a drive letter.
After the SCSI disk is detached from the running ECS, the slot sequence for disks may be changed, leading to the disk drive letter change after the ECS is restarted. As a result, the slot IDs do not correspond to the drive letters, and restarting the ECS failed.
sudo su -
ll /dev/disk/by-id/|grep Disk drive letter
For example, if the drive letter of the SCSI disk is /dev/sdb, run the following command:
ll /dev/disk/by-id/|grep sdb
CNA64_22:/opt/galax/eucalyptus/ecs_scripts # ll /dev/disk/by-id/|grep sdb lrwxrwxrwx 1 root root 9 Dec 6 11:26 scsi-3688860300001436b005014f890338280 -> ../../sdb lrwxrwxrwx 1 root root 9 Dec 6 11:26 wwn-0x688860300001436b005014f890338280 -> ../../sdb
For example, if the SCSI ID obtained in step 3 is scsi-3688860300001436b005014f890338280, use the following data to replace /dev/sdb: