[Fwd: Re: amr driver issues in 7.1-RELEASE]

[Simon]

On Fri, 03 Apr 2009 08:23:03 -0400, Steve Polyack wrote:


>> It appears this has not been fixed in 7.1-p4 I've been waiting to upgrade due
>> to this but was forced after latest root exploit was found recently. Upon MySQL
>> upgrade, running mysql_upgrade which upgrades and checks all the tables,
>> I got 2 of these messages while a large ,1.4G, mysql table was being dumped
>> to array so it could be repaired. I don't know what to make of this error and
>> are not sure if it could be safely ignored.
>>
>> amr0: Too many retries on command 0xc71ceaa8.  Controller is likely dead
>> amr0: Too many retries on command 0xc71cdce8.  Controller is likely dead
>>
>> Thank you in advance!
>>
>> -Simon
>>
>>   
>We still see these as well.  I've also been ensured that these are 
>harmless messages.  As far as I can tell from my own investigation, the 
>behavior has not changed other than the addition of tracking of how many 
>retries there have been when submitting commands to the controller.  
>When the controller mailbox is "busy", the driver must wait before 
>submitting a command.  If the driver has to retry 1000 times twice 
>within one second, it logs the messages you are seeing.  This logging 
>was simply not present before 7.1.

>As I've said and been told, the message is harmless if your machine is 
>still responding.


I just wanted to follow up and let the list know my instance of this error
was being caused by faulty drives. If you get this message, I would be
concerned. Backup your data and run every single drive connected to
this controller via manufacturer's diagnostic full test. A single faulty drive
in RAID setup can cause mediocre performance and this error to show
up.

-Simon