Date: Fri, 26 Nov 1999 13:23:34 +0100
From: vanliesh <[EMAIL PROTECTED]>
I have the following configuration:
Mainboard : Supermicro P6SBA
Memory : 128MB SDRAM
SCSI-controller : Symbios 875
RAID-controller : Mylex eXtreme RAID DAC1164P
Harddisk : 3 * Seagate Cheetah 9.1 Gb ST39102LW
The Symbios SCSI-controller is used for the CD-rom, the three harddrives are
configurated as RAID-5. I am using Redhat Linux 6.1. Everything works very
well, but I cannot rebuild a harddrive without rebooting the server.
This is an overview of the configuration messages:
***** DAC960 RAID Driver Version 2.2.4 of 23 August 1999 *****
Copyright 1998-1999 by Leonard N. Zubkoff <[EMAIL PROTECTED]>
Configuring Mylex DAC1164P PCI RAID Controller
Firmware Version: 5.07-0-79, Channels: 2, Memory Size: 16MB
PCI Bus: 2, Device: 8, Function: 0, I/O Address: Unassigned
PCI Address: 0xFEAFFF80 mapped at 0xC4809F80, IRQ Channel: 5
Controller Queue Depth: 128, Maximum Blocks per Command: 128
Driver Queue Depth: 127, Maximum Scatter/Gather Segments: 33
Stripe Size: 64KB, Segment Size: 8KB, BIOS Geometry: 255/63
Physical Devices:
0:1 Vendor: SEAGATE Model: ST39102LW Revision: 0006
Serial Number: LJG6871700001934JHGM
Disk Status: Online, 17782784 blocks
0:2 Vendor: SEAGATE Model: ST39102LW Revision: 0006
Serial Number: LJR02916000019460G8B
Disk Status: Online, 17782784 blocks
0:3 Vendor: SEAGATE Model: ST39102LW Revision: 0006
Serial Number: LJH0426100001934JLXW
Disk Status: Online, 17782784 blocks
0:7 Vendor: MYLEX Model: DAC1164P Revision: 0507
Serial Number:
1:7 Vendor: MYLEX Model: DAC1164P Revision: 0507
Serial Number:
Logical Drives:
/dev/rd/c0d0: RAID-5, Online, 35565568 blocks, Write Thru
I ran into problems when trying to rebuild a failed drive (it is
hot-swappable) .Then I waited for about 30 sec and put it back in. Then I
issued the rebuild command. The following message appears:
Rebuild of Physical Drive 0:2 Failed - Attempt to Rebuild Online or
Unresponsive Drive
There is a strange thing with the drive status: When I remove a harddisk, the
status becomes critical and then DEAD.
When I ask for the current status, it only shows
0:1 Vendor: SEAGATE Model: ST39102LW Revision: 0006
Serial Number: LJG6871700001934JHGM
Disk Status: Online, 17782784 blocks, 8 resets
0:2 Vendor: SEAGATE Model: ST39102LW Revision: 0006
Serial Number: LJR02916000019460G8B
0:3 Vendor: SEAGATE Model: ST39102LW Revision: 0006
Serial Number: LJH0426100001934JLXW
Disk Status: Online, 17782784 blocks, 8 resets
The lack of a Disk Status line for that drive indicates that the controller no
longer thinks that the drive is present in the configuration. I've never seen
this sort of behavior. Perhaps it is a bug in the controller firmware. Please
try the 5.07-0-07 firmware available from my web site and see if it exhibits
this same behavior.
As you can see, there is no diskstatus of drive 0:2 which is the drive I
removed. When I reboot the machine the status is there again and shows
0:1 Vendor: SEAGATE Model: ST39102LW Revision: 0006
Serial Number: LJG6871700001934JHGM
Disk Status: Online, 17782784 blocks,
0:2 Vendor: SEAGATE Model: ST39102LW Revision: 0006
Serial Number: LJR02916000019460G8B
Disk Status: Dead, 17782784 blocks
0:3 Vendor: SEAGATE Model: ST39102LW Revision: 0006
Serial Number: LJH0426100001934JLXW
Disk Status: Online, 17782784 blocks,
Now the status of drive 0:2 is DEAD and I can start a rebuild of drive 0:2.
Rebuild of Physical Drive 0:2 Initiated
Rebuild in Progress: Logical Drive 0 (/dev/rd/c0d0) 1% completed
....
Rebuild in Progress: Logical Drive 0 (/dev/rd/c0d0) 97% completed
Rebuild Completed Successfully
Physical Drive 0:2 is now ONLINE
Logical Drive 0 (/dev/rd/c0d0) is now ONLINE
What can be the cause of this?? I don't want to reboot the server when a
drive crashes, it should be possible to rebuild it online!
It seems to me that the RAID-controller should scan the SCSI bus for new
harddrives, even without doing a reboot.
Agreed. Are you by any chance also power-cycling the drive when you reboot the
system?
Leonard