Adaptec RAID controller hanging machine since upgrade to 4.8-RELEASE-p22. Hardware or Driver changes?

[Scott Lambert]

I upgraded a machine from 4.7-STABLE to 4.8-RELEASE-p22 a couple of
weeks ago.

It has now stopped responding to anything other than ICMP requests
twice.  The previous uptime was over 80 days.  There is a web log rotate
script which runs at approximately that same time.

The error message on the console is (hand transcribed and third hand):

asr0:  Blink LED 0x7 resetting adapter

The web log rotate script does this a few times for various values of
httpd-access.log before issuing an apachectl restart:

gunzip /var/log/httpd-access.log.0.gz
cat /var/log/httpd-access.log >>/var/log/httpd-access.log.0
gzip /var/log/httpd-access.log.0
rm /var/log/httpd-access.log
touch /var/log/httpd-access.log

( Yes, I realize how inefficient and bad that script is.  It seems
  designed to abuse the disk subsystem.  It was written before I
  inherited the system and I hadn't actually looked at the script before
  tonight. )

I found a question about a similar message, s/0x7/0x3/, on a 2400S
in the FreeBSD-questions archives on July 2, 2003.  I don't see any
follow-ups to that message.  There is also an older three message thread
all from the same poster from October 14 and 16, 2002, about the 0x3
message on a 2100S saying that it had apparently been cured by replacing
the cache memory module on the controller.

So, it seems likely that I am suffering a hardware failure.  However,
beginning so soon after the upgrade of the OS and the error number being
slightly different scare me that it could be caused by a change to the
asr driver.  I decided to post this message tonight, just in case, and
talk to Adaptec tomorrow.

-- 
Scott Lambert                    KC5MLE                       Unix SysAdmin
lambert at lambertfam.org