HP NetRAID-4M Configuration and Upgrade Guide (Release 5)

Chapter 9 Issues and Problem Resolution
66
Problem: If you initiate a manual move of resources in a Windows 2000 cluster,
some of the resources will not move the first time. You must attempt the move a
second time for them to move successfully. This only happens the first time that you
manually move a resource. Any subsequent times, the Move Group function
operates properly.
Workaround: Repeat the move.
57 System will not boot until RAID-5 finishes clear
Problem: If Clear is happening on a RAID-5 container that resides on the same
disks as your boot container, Windows 2000 can take up to 11 minutes to complete
the boot. This only occurs with certain model disk drives, and only when Clear is
initiated through the CCU on a RAID-5 container located on the same disks as the
boot container.
Workaround: Either wait for the Clear to finish, or else use Scrub instead of Clear.
Scrub offers immediate access to the file system.
58 NetWare 5.1 will not boot the driver with 4 GB of memory
Problem: The NetRAID-4M controller will not boot with 4 GB of memory installed
in the server.
Resolution: This problem is fixed in NetWare 5.1 Support Pack 2.
59 Novell NetWare event handling of drive pulls with no I/O
Problem: On a remotely managed operation, events occurring with the NetWare
server will not immediately sync up in a timely manner with the client. The time
delay is short but needs to be recognized.
Background: The FAST GUI utility can run on a Windows based client, and can be
used via TCP/IP as a remote management tool to access and manage an HP
NetRAID-4M controller installed in a HP Server running NetWare.
Drive controllers detect drive changes either directly (via SCSI command, read/
write, consistency check, rescan, etc) or indirectly (via the enclosure SAF-TE).
NetRAID-1Si and NetRAID-3Si controllers will poll the SAF-TE every 10 seconds
but do not poll the drives directly. SAF-TE is mostly used to detect insertions to
automatically start rebuilds. Removals do not become error conditions until the I/O
actually fails. The advantage is not to waste time rebuilding a perfectly good data
file. The disadvantage is that a drive could be bad (and the array vulnerable) for a
long time.