W2012R2
I have a storage space set up with six SATA hard drives, RAID, (large and slow drives - 7TB and 5400). When someone tries to dump a large amount of files from a network share (D&D, normally from a 2008R2 server), the virtual disk will get a delayed write warning (data lost, etc.) and the volume will go offline. It is easy enough to place the volume back online and that seems to do it until the next large copy. The error message will have a UNC path like "\\BLAHSERVER\DATA\STUFF\SomeFile.xxx" and say that the copy failed, data lost, etc.). The delayed write message sometimes points to the logical volume number and sometimes it will pick a specific physical disk number.
The six drives are on a RAID controlled that is set up for JBOD. Whether this equates to pass though, I've no idea because there is nothing in the controller BIOS that specifically says "pass through".
I've scanned and rescanned the drives and there are no errors. There is no warning about a hardware failure; nothing is marked "unhealthy" and the storage pool never attempts a rebuild.
I turned off the delayed write feature of each of the disks as a dart throwing exercise. Doesn't let you do that for vdisks.
The vdisk was thinned provisioned. I redid the storage space and made the vdisk fixed to see if that would help. Jury is still out.
I suspect the RAID controller is not properly configured for pass through, but I thought I'd ask anyway.
Thanks