Bug#406640: arcmsr driver, abort device command of id=0 lun=0
erich schrieb:
> If you got abort device command, it said that you had one or more
> scsi commands timeout from linux kernel. There were some reasones
> cause this problem. The Areca RAID firmware can not handle some of
> your physical sata disks. There were a lot of effect on this
> problem, such as power less and platform vibration. Areca have some
> linux users met this problem and they told me the problem is coming
> from their power less. When they upgrade their power supply and
> then the messages disappear forever. You can try to upgrade your
> power supply and see if there were any difference. If you need to
> doubt about Areca firmware compatible issue with your Western
> Digital Raptor disks. You can remove all your RAID5 containing 7
> Segate 750GB disks (need to unplug their power cables). And then
> run testing with your two Western Digital Raptor disks only. If the
> abort commands still there and we can make sure it is Areca
> firmware's compatible issue with your two WDC sata disks. If the
> abort command messages disappear, please upgrade your power supply.
I performed a new test with the latest driver from 2006-11-07.
ftp://ftp.areca.com.tw/RaidCards/AP_Drivers/Linux/DRIVER/SourceCode/arcmsr.1.20.0X.13-61107.zip
ftp://ftp.areca.com.tw/RaidCards/AP_Drivers/Linux/DRIVER/SourceCode/ReleaseNote.txt
11/07/2006 1.remove #include config.h and devfs_fs_kernel.h
2.enlarge the timeout duration of each scsi command.
it could aviod the vibration factor with sata disk on some bad chassis
This solves the problem, at least I didn't get any error message with ext3
again. I still can't understand how this could be a matter of chassis
vibration. Why don't I get this error with xfs filesystem? Why didn't I get
these errors with the 750GB Seagate disks in the Raid5 set?
Best Regards,
Ralf Gross
Reply to: