We are getting ready to set up a basic file server using a new HP ProLiant ML110 G7 with 3.3Ghz quad-core Xeon processor and 8 GB of RAM. We upgraded the RAID controller from the basic onboard B110i to an HP smart Array P410 with 256MB module so that we can run ESXi HyperVisor to run a single virtual machine with Windows 7 as the host OS for the file sharing duties. We are using all identical Western Digital Caviar Black 2 TB hard drives.
My concern is trying to determine which RAID configuration to use. My initial thought was to use two hard drives in RAID 1 array for the benefit of allowing always uptime, which would be the primary datastore for ESXi. Then we would install a third non-RAID member hard drive which would act to backup data from the RAID array in the event that the controller went down or something happened with the array. However, I'm not too sure how to set up so the ESXi server automatically backs up everything from one datastore onto a separate hard drive.
Next, I considered just doing three hard drives in a RAID 5 array in the server and not having a separate non-RAID member hard drive. This would eliminate needing to find a way to automatically backup the datastore onto a separate drive, but makes the entire storage system rely on the RAID array which I am hesitant about.
Either way we go, I have a separate fourth hard drive that will be put into an external enclosure and will be used simply for making a backup of the data directly from the Windows 7 host operating system instead of all the virtual hard drive files and everything in the full datastore. This backup can be run daily and taken off-site to store.
Any thoughts, experience, or ideas for what might be best in this situation would be incredibly appreciated!
Emerald, may I ask the benefit of the TLER in a small business server RAID? I've built numerous RAID 1 arrays using Western Digital Caviar Black drives and they actually work incredibly well. In one of these HP ProLiant ML110 G7 servers with a P410/256 MB card we installed Server 2008 R2 Standard and used HDTune to determine how the array was performing and it was quite impressive for two 7,200 rpm SATA drives.
In this case this is something I should definitely keep in mind and plan on in the future. I've never had issues yet with the WD Caviar Black drives in RAID 1, but again I'm just using a basic array and that sure doesn't mean that I won't have issues in the future. Thank you for the heads up!
If the drive itself is inherently reliable but has some bad sectors, then TLER and similar features prevent a disk from being unnecessarily marked as 'failed' by limiting the time spent on correcting detected errors before advising the array controller of a failed operation. The array controller can then handle the data recovery for the limited amount involved, rather than marking the entire drive as faulty.
Yes, I had odd symptoms like a raid 0 array that failed but could be rebuilt, and raid 1's that would have to be rebuilt once a week, the extra time before declaring failed made a lot of difference in my opinion.