RAID5 fails to recognize & recreate a substituted disk.

[Hernani Rodrigues]

RAID5 fails to recognize & recreate a substituted disk.

We have a Bull Escala running AIX 4, recently one of its raid5 disks
failed, sounding an alarm. Because we didn't know what was the problem
we proceeded to shutdown the server and called the assistance. They
instructed us to turn on the server while they came with a new disk
that theoretically would by recognized and reconstructed automatically
by the RAID5 with only same degradation to the servers performance.

When they made the substitution nothing happened. After tow days, they
reach the conclusion that the disk wasn't being recognized by the
RAID5 has belonging to the read group and therefore wasn't being
reconstructed. They also say that this was because we shutdown the
server when this disk had failed. And they also say that the only way
to solve the problem is to reconstruct all the RAID5 disks and their
related information manually as if we where configuring a new server.

Same relevant information:
# lspv
hdisk0 00600598987ecb22 rootvg
hdisk1 00600598e183b218 xxxxxvg
------------------------//-------------------
# lsdev -Cc disk
hdisk0 Available 10-10-00-0,0 Bull 9.1 GB 16 Bit SCSI Disk Drive.
hdisk1 Available 17-20-00-0,0 BULL PCI RAID 5
------------------------//-------------------
# lsdev -Cc adapter
sa0 Available 01-S1 Standard I/O Serial Port
sa1 Available 01-S2 Standard I/O Serial Port
sa2 Available 01-S3 Standard I/O Serial Port
ppa0 Available 01-R1 Standard I/O Parallel Port Adapter
fda0 Available 01-D1 Standard I/O Diskette Adapter
siokma0 Available 01-K1 Keyboard/Mouse Adapter
scsi0 Available 10-10 Wide/Ultra-2 SCSI I/O Controller
scsi1 Available 10-18 Wide/Fast-20 SCSI I/O Controller
dom0 Defined 17-18 BULL PCI LVD RAID adapter
srai0 Available 17-20 BULL PCI LVD RAID adapter
ent0 Available 20-40 DEC 21143 PCI Fast Ethernet Controller
(9f110208)
mg20 Available 20-48 GXT130P Graphics Adapter
sioka0 Available 01-K1-00 Keyboard Adapter
sioma0 Available 01-K1-01 Mouse Adapter
------------------------//-------------------
# lsdev -Pc pdisk
pdisk 1000mbC ssar 1GB SSA C Physical Disk Drive
pdisk 2000mbC ssar 2GB SSA C Physical Disk Drive
pdisk 4000mbC ssar 4GB SSA C Physical Disk Drive
pdisk 1000mbF ssar 1GB SSA F Physical Disk Drive
pdisk 2000mbF ssar 2GB SSA F Physical Disk Drive
pdisk 4000mbF ssar 4GB SSA F Physical Disk Drive
pdisk 4500mbC ssar 4.5GB SSA C Physical Disk Drive
pdisk 9100mbC ssar 9.1GB SSA C Physical Disk Drive
pdisk 4500mbF ssar 4.5GB SSA F Physical Disk Drive
pdisk 9100mbF ssar 9.1GB SSA F Physical Disk Drive
pdisk ossadisk ssar Other SSA Disk Drive
pdisk scsd ssar SSA Physical Disk Drive


Does anybody have another, more simple and less arduous, solution for
this problem.

[Hans-Joachim Ehlers]

"Hernani Rodrigues" <hernaniluis@yahoo.co.uk> schrieb im Newsbeitrag
news:3edd8120.0309250934.1340672b@posting.google.c om...

> RAID5 fails to recognize & recreate a substituted disk.
>
> We have a Bull Escala running AIX 4, recently one of its raid5 disks
> failed, sounding an alarm. Because we didn't know what was the problem
> we proceeded to shutdown the server and called the assistance. They
> instructed us to turn on the server while they came with a new disk
> that theoretically would by recognized and reconstructed automatically
> by the RAID5 with only same degradation to the servers performance.
>
> When they made the substitution nothing happened. After tow days, they
> reach the conclusion that the disk wasn't being recognized by the
> RAID5 has belonging to the read group and therefore wasn't being
> reconstructed. They also say that this was because we shutdown the
> server when this disk had failed. And they also say that the only way
> to solve the problem is to reconstruct all the RAID5 disks and their
> related information manually as if we where configuring a new server.
>
> Same relevant information:

Normaly a raid system just DO NOT take a new disk and add it automaticly to
its configuration. Sounds silly. For that you have to have a "hot-spare
disk"

So lets do some trouble-shooting:
1 ) Does the new disk has exactly the same size as the failed disk ?
2) Does the new disk can be seen - aka: lscfg -l pdisk#
if yes for 1 and 2
3) use "smitty ssaraid" to set the new pdisk as "hot-spare" disk
4) use "smitty ssaraid" to add this "hot-spare" disk to your raid5 system
( Hopefully this step is possible - could not be testet right now )
5) Now the system should use the hot-spare disk and rebuild the raid system
6) remove the failed disk from the raid configuration

All from memory

HTH
Hajo

CATE AIX/RS6000

P.S
if critical i can come from germany ;-)