If a Connection is interrupted, while implicitly activating the
Database in a Purescale environment, a table space can be marked
as Locally Offline in SD mode. (This isn't true for the system
catalog table space.)
Startup does not fail altogether and continues to start other
table spaces that this member is able to access.
The administrator will need to use the ALTER TABLESPACE SWITCH
ONLINE command to bring the table space back online.
The following messages can be seen in the db2diag.log :
FUNCTION: DB2 UDB, buffer pool services, sqlbStartPools,
probe:2542
MESSAGE : ADM6023I The table space "USERSPACE1" (ID "2") is in
state "0x00000000".
The table space cannot be accessed. Refer to the documentation
for
SQLCODE -290.
FUNCTION: DB2 UDB, buffer pool services,
sqlbStartPoolsErrorHandling, probe:4151
MESSAGE : ADM6081W The table space "USERSPACE1" (ID "2") is in
the OFFLINE state
and is not accessible. The table space state is "0x00004000".
Refer
to the documentation for SQLCODE -293.
However db2pd -tablespaces shows that the table space in
question is in normal state (0x00000000) which can be confusing.
This is because db2pd -tablespaces only shows the Global table
space State not the Local State.
MON_GET_TABLESPACE will report the Local State of the Table
space but the output does not make it clear whether it is
reporting Global or Local table space States.
A connection interrupt should not prevent a table space start
up.
The specific error code could be handled better (i.e. retry this
tablespace start up if interrupt occurs).
db2pd and MON_GET_TABLESPACE need to be more clear about
local/global tablespace states. |
