HP3000-L Archives

November 1996, Week 3

HP3000-L@RAVEN.UTC.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Therm-O-Link <[log in to unmask]>
Reply To:
Therm-O-Link <[log in to unmask]>
Date:
Mon, 18 Nov 1996 08:18:58 EST
Content-Type:
text/plain
Parts/Attachments:
text/plain (43 lines)
On Wed, 13 Nov 1996 14:21:45 -0800 "Deborah L. McMahon" <[log in to unmask]>
writes:

>Homer Godwin ([log in to unmask]) wrote:
>:   Our very stable OMAR application has exhibited erratic behavior the last
>:   day of the month for the past two months.
>
>:   At the time of the failure, the operator gets a brief message that we
>:   have determined is a result of a failure in the MPE Mail sub-system.  The
>:   mail facility is used as means of process-to-process communications.
>:   These two processes exits for the purpose of allowing the operator to
>:   proceed to the next order while the program completes the update process
>:   for the previous order.  When this mail system fails, the background
>:   process retains DBLOCK that effectively brings the system to its knees.
>
>I just upgraded (Sept. 13th) to an HP939ks running CA Manman v 9.7
>We processed out end of year on Sept 30th and evidenced the same
>problem you are describing. CA passed the problem off to HP (I have
>support on both) and HP's response was that we were in a 'deadly
>embrace' [their official terminology - in mine it was a 'living
>nightmare'!] I wound up using DBUTIL to resolve the problems when
>they were occuring, but it was not clean or pretty.

Can you explain how you used DBUTIL to resolve a 'deadly embrace'?

My understanding of a 'deadly embrace' is where process 1 unconditionally
locks resource A and then unconditionally locks resource B, and
at the same time process 2 unconditionally locks resource B and
then unconditionally locks resource A, so process 1 waits to lock
resource B and can't because process 2 has it locked and is waiting
for resource A, which process 1 has locked.

For some years, there has been a request in the HP3K user community
for HP to provide a way of resolving 'deadly embraces'.  The only
way I know of resolving a true 'deadly embrace' is to re-boot the
system, so if we can use DBUTIL to resolve it, I would be interested
in knowing how.  Thanks.

Jim Phillips                            Manager of Information Systems
E-Mail: [log in to unmask]      Therm-O-Link, Inc.
Phone: (330) 527-2124                   P. O. Box 285
  Fax: (330) 527-2123                   Garrettsville, Ohio  44231

ATOM RSS1 RSS2