On a deployment site, an S2600i is installed and a 6 TB RAID group is created on it. Initially, three LUNs (1 TB/2 TB/3 TB) are allocated on the RAID group and separately mapped to three servers. Later a server is offline, and its occupied storage space is planned to be released and used by the rest two servers, that is, the previous three LUNs need to be changed to two 3 TB LUNs. The plan is as follows: Remain the second LUN. Delete the first and third LUNs. Create a new LUN by using the released space. However, the new LUN fails to be created by using the released space. An alarm is generated, indicating that the LUN capacity exceeds the threshold and fails to be created.
Delete all the previously created LUNs. Create new LUNs. The fault is rectified.
The first and third LUNs are deleted and the second LUN remains; therefore, the remaining storage space is incontinuous and cannot be combined. A new LUN fails to be created by using the remaining space.
Data blocks with incontinuous addresses cannot be combined. Furthermore, disk space needs to be logically allocated before LUN creation; otherwise, the capacity change of a LUN (not the last LUN) causes data loss.