advice on Raid5

[Dissolved]

Does anyone here run raid5?
if so,
have you ran it on an oc'd system? pci bus mhz?
Any data loss?
What drives did you use? how many?
What raid card?
And how was the performance?
Thanks

 

[Xaotic]

I run RAID-5, but do not run an out of spec PCI bus. The system is an older board and OC'd only enough to bring the bus back into spec. The adapter is a Serveraid 4M running a 3 drives on a 1E array internally and 5 drives on a RAID-5E array in an external enclosure. I haven't had any data loss with this array, though I have seen data loss in another array due to technician error. Normally, the data can be recovered without any problems by reading the RAID configuration back onto the controller when a controller problem has occurred or rebuilding the data on a single drive when a drive goes bad and no hot spare is available. Performance on the home system is only fair since the MB is older and I haven't upgraded to a new board with the higher bandwidth slots. Speed is not my primary concern on this machine since it's used as a file server.

 

[Dissolved]

Are they scsi or ide?
And if so what model drives are they?
and can you bench them?
even tho there older, id just like to see how they would perform compaired to 1 of the drives in its regular mode.

 

[litghost]

i run a Raid5 with 5 scsi 40gb
i have linux installed on the system and i dont know how to run benchmarks
all of the drives are seagate's.
all and all i am impressed with the speed. I notice a large difference in data transfer speed inside the raid5, than compared to other drives

oh and the drives are 3 years old

just my .02

 

[Xaotic]

SCSI on both internal and external arrays. External array is composed of DMVS IBM 10K SCA drives, while the internal array is Seagate 68 pin drives. Benchmarking a single drive would be a bit problematic, since the RAID-5 array is SCA and in an external enclosure. Read speed is very good, with write a bit slower, but my bottleneck is the system(and 33Mhz PCI bus) it's attached to. I didn't bench the system when it was assembled and to disassemble it to bench a single drive would involve a risk to the data, since the RAID controller's configuration would have to be changed twice to remove and replace the spare. When I upgrade the system, I'll benchmark it fully, but till then I'd rather not risk it.

 

[Dissolved]

SCSI on both internal and external arrays. External array is composed of DMVS IBM 10K SCA drives, while the internal array is Seagate 68 pin drives. Benchmarking a single drive would be a bit problematic, since the RAID-5 array is SCA and in an external enclosure. Read speed is very good, with write a bit slower, but my bottleneck is the system(and 33Mhz PCI bus) it's attached to. I didn't bench the system when it was assembled and to disassemble it to bench a single drive would involve a risk to the data, since the RAID controller's configuration would have to be changed twice to remove and replace the spare. When I upgrade the system, I'll benchmark it fully, but till then I'd rather not risk it.

well i bet i could find a benchone for single drive online, but can u bench the raid array?

 

[Xaotic]

It'll probably be this weekend before I have a chance to bench the array, due to work, commute and current projects, but I'll get some results as soon as I get a chance.