RAID5 expansion - Smartarray6i on DL380 G4

[coastertoaster]

Hi Guys,

Hope you’re all well.
Got a question about Smart Array 6i RAID5 expansion on one of our DL380 G4s. The research I’ve been doing leads me to believe that some people think this can be done, and others not. Hoping I can get a definitive Yes or No as regards it all. It’s a very crucial server so don’t want to just dive in headlong incase it just doesn’t work.

The system in question is a DL380 G4 (370596-421), with an onboard Smart Array 6i (fw RevB v2.36).

All 6 Hard drive slots are used (the reason we can’t just slap in more hard drives). The layout is:

Slots 0 to 3 - 4 x 72GB SCSI Hot swap in RAID5 (10K U320s)
Slots 4 & 5 - 2 x 300GB SCSI HS in RAID1+0 (not really a part of this question)

The RAID5 array contains everything of importance on the server (system & boot partitions, windows installation, AD logs, yada yada). The mirrored volumes are mounted as an NTFS folder.

So…have purchased 4 x 300Gb Hot swap SCSI drives with the thought of slowly replacing each of the 4x 72Gb drives. Once the data is replicated over all 4 again, will the smart array 6i detect we’ve got 4 bigger drives and allow us to increase the size of the partitions on the volume. It’s a Win2003 Standard server.

For example: Day 1, replace slot 0’s 72Gb drive with a 300Gb drive, then give data time to replicate.
Day 3, replace slot 1’s 72Gb drive with a 300Gb drive, then give data time to replicate.
Day 5, replace slot 2’s 72Gb drive with a 300Gb drive, then give data time to replicate.
Day 7, replace slot 3’s 72Gb drive with a 300Gb drive, then give data time to replicate.

So, maybe by day 9 or 10 we can just run DISKPART in Windows and the volume will grow to fill up the extra space.

Would this scenario work, or is it pie in the sky?

Big thanks, your thoughts and comments appreciated.
Coastertoaster :D

[Stoyanov Dmitriy]

Yes, your scenario will work but more faster solution is take a backup, replace all four drives, create a RAID5 on 4x300Gb and restore the backup.

Best Regards, Dmitriy Stoyanov