I am not sure why VMWare would say that. It may be that they do not understand the RAID Group defragment process on the CLARiiON arrays.
Running a defrag on a RAID Group is safe for the data on the LUNs and the applications accessing those LUNs. Just set the defragmentation priority to Low or Medium. I have run defragmentation on RAID Groups containing SQL databases, Exchange databases, file shares etc, and have never had an issue on the application side.
The key is ensuring the appropriate defragmentation priority is selected to ensure you minimize the performance impact to the array and the hosts.
My guess is that VMware support thinks it's at block level on the host instead of on the array. IMHO they don't know how a clariion works and "sees" LUNs. A Clariion cannot look into LUNs and won't see file systems.
The defrag is on block level, but the host won't know the difference, since pointers make sure all data blocks - as seen by the host - are still pointing to the same data.
Also keep in mind the "defrag" that is performed internal to the array doesn't change how the LUN appears to any host (in this case, VMware). The defrag performed by the array only serves to move LUNs "closer" together within a RAID Group. For example if you created LUNs 1,2,3,4,5,6,7,8 within a specific RAID Group, and then deleted LUNs 2,3,6,7 - running the array-internal defrag will move the remaining LUNs 1,4,5,8 together within the RAID Group, in order to maximize the contiguous free space available to create new LUNs. The data logically contained _within_ any specific LUN is not altered, at the block or file level. A byproduct of the above is; if you've never deleted any LUNs within a particular RAID Group, there is no need/benefit to performing an array-internal defrag operation on that RAID Group.
AranH1
2.2K Posts
0
December 3rd, 2008 14:00
Running a defrag on a RAID Group is safe for the data on the LUNs and the applications accessing those LUNs. Just set the defragmentation priority to Low or Medium. I have run defragmentation on RAID Groups containing SQL databases, Exchange databases, file shares etc, and have never had an issue on the application side.
The key is ensuring the appropriate defragmentation priority is selected to ensure you minimize the performance impact to the array and the hosts.
RRR
4 Operator
•
5.7K Posts
0
December 4th, 2008 03:00
The defrag is on block level, but the host won't know the difference, since pointers make sure all data blocks - as seen by the host - are still pointing to the same data.
DGM3
238 Posts
0
December 7th, 2008 20:00
Regards,
DGM
RRR
4 Operator
•
5.7K Posts
0
December 16th, 2008 12:00
did we answer your question ? I think the answer is right there: don't worry about a defrag !
Please mark the question as answered and reward some points for the best 3.