L2 cache spanning multiple disks
Posted: Tue Feb 06, 2018 1:29 pm
Hello,
first of all thanks for this program, it looks great and makes almost everything I want to achive a possibility.
But there is 1 feature I would like very much to see (would turn this SW into instant buy for me and company):
I would like to be able to assign L2 cache on multiple disks - I know this can be done in a SW raid fashion, but this solution incurs following drawbacks:
*SW raid overhead
*Redundancy overhead - when using RAIDed disks for redundancy
*Total failure on disk failure - when using RAID0
My proposal is to be able to add multiple L2 locations for single cache task so that if one location fails the system still works utilizing the rest of the space available. I want to use many discarded older SSD to cache and those disks often fail very soon. Also there is much hassle in regard to raid rebuild and cache reconfiguration after disk failure (for maximizing storage RAID0 or simple span solution is used, redundancy raid drains too much space or significantly increases wear/reduces speed). I dont care about the RAID0 speed increase as much as the ability to preserve the nonfailed portion of disk space - in fact just simple join on the drive space is sufficient as it is still many times faster than spindle drive.
Thanks for your consideration.
first of all thanks for this program, it looks great and makes almost everything I want to achive a possibility.
But there is 1 feature I would like very much to see (would turn this SW into instant buy for me and company):
I would like to be able to assign L2 cache on multiple disks - I know this can be done in a SW raid fashion, but this solution incurs following drawbacks:
*SW raid overhead
*Redundancy overhead - when using RAIDed disks for redundancy
*Total failure on disk failure - when using RAID0
My proposal is to be able to add multiple L2 locations for single cache task so that if one location fails the system still works utilizing the rest of the space available. I want to use many discarded older SSD to cache and those disks often fail very soon. Also there is much hassle in regard to raid rebuild and cache reconfiguration after disk failure (for maximizing storage RAID0 or simple span solution is used, redundancy raid drains too much space or significantly increases wear/reduces speed). I dont care about the RAID0 speed increase as much as the ability to preserve the nonfailed portion of disk space - in fact just simple join on the drive space is sufficient as it is still many times faster than spindle drive.
Thanks for your consideration.