Hi All,
During our morning DB2 reorg window, I've been observing a table reorg of the
BACKUP_OBJECTS table using the command "db2pd -d TSMDB1 -reorg". It starts and
later pauses without appearing to do any work. I'm assuming this because
'CURCOUNT' is not incrementing, and remains at '0'. It tends to start and pause
twice during the 2 hr. window. There are no correlating entries in the server's
TSM actlog , db2diag.log, or dsmserv.err. In fact I see no errors logged about
this at all. This has been recurring for the past three days.
While we've had a variety of experiences with reorg's of other tables over the
past year or so, I've never seen this specific behavior. Is there
any recommended way to handle this scenario gracefully (without a server
restart), such as issuing a DB2 inplace stop/start or other method?
I realize it's not optimal to have client backups running during the reorg
window, however, it's unavoidable for us. Our load is typically light at
that time of day. Our reorgs have been running without serious problems for the
past few months, but I'd be willing to try cancelling all client
backup sessions before a future reorg window starts, to see if eliminating this
activity makes any difference in this case.
Does anyone have a solution for such a scenario they'd care to share?
Server details: TSM server v6.3.4.200 (not an upgrade from TSM v6.1) on AIX
6.1; DB2 9.7.6.
Thanks in advance!
Ruth Mitchell
U of I, Urbana, IL
|