I have a brand new Dell md1000 disk array that is connected to two PE2950's. This implementation is simple: just a jbod configured as RAID5 with 6 and 4 disks respectively. Both servers are using the volumes to host MSSQL 2005 logs and data. One server runs MS Windows Server 2003 R2 64-bit standard and the other MS Windows Server 2008 64-bit standard. One server, the 2003 Server, frequently loses contact with the virtual disk (a RAID 5 volume). The disks suddenly appear in openmanage as down--one must reset the PERC and reassemble the volume from a foreign config --it's pretty simple to bring it back with no data lost. The 2008 2950 server never loses contact and has no problems thus far. The array is two months old, this has been happening for a month--pretty much as long as it has been heavily used. Dell had me update all drivers including the PERC 6 card, the server bios and the SAS disks firmware.
On the off chance that anyone has seen this problem or something similar, I thought I would put it out there for comment and thoughts. All equipment is Dell including a pair of PERC 6/E adapters purchased to connect to the array. The Dell tech support, while helpful, seems not to know a solution beyond just restoring the virtual disk. What might be the likely point of hardware or software problems in such a scenario?