error messages from easyraid ex array

[Kenneth D. Merry]

On Thu, Mar 10, 2005 at 13:32:52 +0100, Fritz Heinrichmeyer wrote:
> Hello,
> 
> sometimes i see following messages with FreeBSD-5.3, ahc driver and an 
> easyraid ex system. We have two identical boxes one connected to another 
> system has no problem. The problematic error messages occur when 
> berkeley-db files are massaged by cyrus imap server.
> 
> I cannot believe there is a media error on an raid 5 system that is not 
> announced at the control panel of the raid box.

Well, firmware on the RAID box will do whatever the authors wrote it to do.

In theory, you're right.  On a non-critical RAID-5 array, they should be
able to recover from any single drive read error.  If your array is
critical, though (i.e. you've lost one disk), you won't be able to recover
from a read error.

> Driver:
> 
> ahc2: <Adaptec 29160 Ultra160 SCSI adapter> port 0x3800-0x38ff mem 
> 0xf7ed0000-0xf7ed0fff irq 11 at device 4.0 on pci1
> da1 at ahc2 bus 0 target 5 lun 0
> 
> Errors:
> 
> (da1:ahc2:0:5:0): READ(10). CDB: 28 0 0 71 6a 4f 0 0 10 0
> (da1:ahc2:0:5:0): CAM Status: SCSI Status Error
> (da1:ahc2:0:5:0): SCSI Status: Check Condition
> (da1:ahc2:0:5:0): MEDIUM ERROR asc:0,0
> (da1:ahc2:0:5:0): No additional sense information
> (da1:ahc2:0:5:0): Retrying Command (per Sense Data)

That's a little bit of a strange error, because they only set the sense key
but not the asc/ascq.  In any case, it looks like we retried it.

If the device was still returning errors, you'd eventually see a "Retries
Exhausted" message.

Ken
-- 
Kenneth Merry
ken at FreeBSD.org