Unsolved
This post is more than 5 years old
3 Posts
0
734
May 30th, 2008 09:00
How to Most Effectively Scale Up
This is a question about future scalability for budget purposes and whether our needs are best served by replacing size of existing disks within the DPE or adding a DAE shelf.
We currently have two RAID groups: 1) RAID 5 with seven 366.8 GB disks (four of which contain the FLARE) and 2) RAID 10 with four 133.7 GB disks. Then there is a 366.8 GB hotspare.
There are currently three host servers sharing the storage. Server #1 has a 2.2 TB RAID-5 LUN and a 267GB RAID-10 LUN; Servers #2 and 3 each have 100 GB RAID-5 Luns. That said, the available storage on the array is fully subscribed. All disks are SAS.
What are the considerations in scaling up the SAN to accommodate more servers? Would cost minimization lead me in one direction and minimizing the work effort and degree of complexity lead me in another? Are the options either adding a PAE shelf or replacing the existing disks with ones of greater capacity? I understand that if we have needs which can be met with SATA, that we would need a DAE because could not mix and match SAS/SATA on the DPE. If we replaced existing disks with larger disks and wanted to maintain existing LUNs but make significant additional storage available, what procedures would need to be followed? How much down time would be needed?
We currently have two RAID groups: 1) RAID 5 with seven 366.8 GB disks (four of which contain the FLARE) and 2) RAID 10 with four 133.7 GB disks. Then there is a 366.8 GB hotspare.
There are currently three host servers sharing the storage. Server #1 has a 2.2 TB RAID-5 LUN and a 267GB RAID-10 LUN; Servers #2 and 3 each have 100 GB RAID-5 Luns. That said, the available storage on the array is fully subscribed. All disks are SAS.
What are the considerations in scaling up the SAN to accommodate more servers? Would cost minimization lead me in one direction and minimizing the work effort and degree of complexity lead me in another? Are the options either adding a PAE shelf or replacing the existing disks with ones of greater capacity? I understand that if we have needs which can be met with SATA, that we would need a DAE because could not mix and match SAS/SATA on the DPE. If we replaced existing disks with larger disks and wanted to maintain existing LUNs but make significant additional storage available, what procedures would need to be followed? How much down time would be needed?
No Events found!
mpi2
2 Intern
•
215 Posts
1
June 4th, 2008 04:00
After the migration you can use the old disks from the exisiting DAE to use them for additional storage for any other purposes.
"LUN Migration" is a feature from the CX which allows you to online migrate from one LUN to another which the same or bigger size. (so no downtime needed ;-))
bertog
61 Posts
1
June 4th, 2008 13:00
It depends on what kind of system you have, what your capacity and performance needs are, and how much money you want to spend.
From your reference to SAS disks, I will infer that you have an AX4-5 with the expansion enabler. If that is the case, it would be fairly simple to add an additional DAE with as many disks as you think you will need. The AX4-5 does support mixing both SAS and SATA disks in the same enclosure, so you have a lot of flexibility. You could then expand your existing Virtual disks / LUNs with metaLUN technology.
Swapping out existing drives for larger capacity is a more complex procedure, and would require either replacing one drive in each RAID 5 RAID Group at a time (and leaving the RG in a degraded state and susceptible to a double fault, not advisable) or destroying the entire configuration and restoring from a backup.
Thanks,
Greg