When the user issues a table space rollforward while another
table space rollforward is already in progress, error SQL4908N
will be returned, and the following message will be written into
the DB2 diagnostic log:
FUNCTION: DB2 UDB, recovery manager, sqlpInitTblRecovery,
probe:2770
DATA #1 : <preformatted>
A table space rollforward is already in progress.
One or more of the table spaces specified in the list to be
rolled forward is not in rollforward in progress state.
Use the LIST TABLESPACES SHOW DETAIL command to get a list
of table spaces in rollforward in progress or rollforward
pending state.
There are two ways to proceed.
1. Complete the table space rollforward that is already in
progress.
This will not be possible if rolling forward to a point in time
and
either a table space is not in rollforward in progress state on
all nodes or
the table spaces are interdependent and all table spaces are not
in
rollforward in progress state on all nodes.
Submit the rollforward command again with a list containing the
table spaces that are in rollforward in progress state. You can
submit
a different rollforward command for each node or subset of nodes
if
required, but any table space in rollforward in progress state
that is
not included in the list will be put into restore pending state.
Restore any table spaces place in restore pending state and then
submit the new table space rollforward command.
2. Cancel the table space rollforward that is in progress and
start another table space rollforward.
Submit the rollforward command again with the CANCEL option.
A list of table spaces that are in rollforward in progress state
and a node list are required only if rolling forward to end of
logs.
These table spaces will be put into restore pending state.
Restore the table spaces and submit the rollforward command
again.
The purpose of this APAR is to improve suggestion 2. The current
text is ambiguous, and some users may understand it in the way
that they can cancel the running rollforward and submit another
rollforward. However, this interpretation is incorrect.
Canceling a running rollforward means that the table space will
be put into restore pending state, which is a serious
consequence. Only after restoring the table space again, the
user will then be able to issue another rollforward.
The new text for suggestion 2 will be made consistent with our
manuals:
2. Cancel the table space rollforward that is in progress
(submit the rollforward command again with the CANCEL option) to
put one or more table spaces on all database partitions on which
forward recovery has been started in restore pending state. Use
this option with caution, and only if the rollforward operation
that is in progress cannot be completed because some of the
table spaces have been put in restore pending state. |
****************************************************************
* USERS AFFECTED: *
* See APAR Description *
****************************************************************
* PROBLEM DESCRIPTION: *
* See Error Description *
****************************************************************
* RECOMMENDATION: *
* Upgrade to DB2 for Linux, Unix, and Windows version 9.5 Fix *
* Pack 10 *
**************************************************************** |
