We are trying to perform TSM backups with DB2 V8 FP4
on AIX 5.x.
We have tried on 2 different machines, one running AIX
5.1.0
and TSM client version 5.2.2; the other with AIX 5.2.0
and
TSM client 5.1.6.
We are able to backup to TSM successfully as long as
the DB2
instance is defined to use a single partition, but if
we change
it to use 2 or more partitions, the backup command
fails as 
shown below with an SQL2062N error.
The partitions are defined as logical partitions that
all reside
on the same host machine.  The db2nodes.cfg for 2
partitions 
looks like:
   0 machineA 0
   1 machineA 1
and the ".rhosts" file is set up correctly with the
correct
permissions.  The test database works OK except when
trying to backup using TSM.  Backups to disk work
fine.
The error looks like:
$ db2 backup db sample use tsm
SQL2062N  An error occurred while accessing media 
"/home/db2inst8/sqllib/adsm/libtsm.a".  Reason code:
"406".
We have tried adding the option "open 2 sessions" and
also
the "parallelism" option, but this had no effect.
The 406 reason code identicates that the "dsm.opt"
file 
cannot be found.  The DSM variables are defined in the
DB2 instance user profile as shown below:
export DSMI_CONFIG=/home/db2inst8/tsm/dsm.opt
export DSMI_LOG=/home/db2inst8/tsm
export DSM_DIR=/usr/tivoli/tsm/client/ba/bin
export DSM_CONFIG=/home/db2inst8/tsm/dsm.opt
export DSMI_DIR=/usr/tivoli/tsm/client/api/bin
The "dsm.opt" file exists and is definitely found when

doing a TSM backup with a single partition database.
We have tried configuring the "traceflag" and
"tracefile" options in dsm.opt and they do create
and populate the trace file when doing a backup of
a single partition database.
When the instance is changed to use 2 partitions,
however,
the trace file does not get created.  This suggests
either
that the file isn't found (even though the file hasn't
moved
and the variables still have the same definitions as
for
the single partition case) , or else something went
wrong
during the processing before the file was located
or at least before the trace functionality was
activated,
leaving us with an erroneous error message.
When switching the number of partitions, we drop the 
current database, stop the instance (db2stop), change
the db2nodes.cfg file for the new environment, then 
restart the instance (db2start).  The instance starts
OK 
and we can set database parameters (logretain,
userexit) 
on each partition.  A script ensures that we aren't 
attached or connected to anything, then sets both the
DB2NODE variable and the DB2 client parameters 
(attach_node and connect_node) to the desired
partition
number before attempting to perform the backup
command; e.g.:
   export DB2NODE=$N
   db2 set client attach_node $N connect_node $N
We've tried doing an instance attachement, or a
database
connect, before issuing the backup, but this doesn't
change the symptoms.
We do not currently have any of the TSM parameters
defined
in the database configuration; e.g.:
$ get db cfg for <database> |grep TSM
 TSM management class                   
(TSM_MGMTCLASS) = 
 TSM node name                           
(TSM_NODENAME) = 
 TSM owner                                  
(TSM_OWNER) = 
 TSM password                            
(TSM_PASSWORD) = 
These aren't necessary for the single partition case;
it doesn't seem that they would be required for
multiple
partitions, but ... (?)
Has anybody seen this problem before?  Any ideas for
what else we 
can look for?  Any hints will be most appreciated,
thanks!



__________________________________
Do you Yahoo!?
Yahoo! Mail SpamGuard - Read only the mail you want.
http://antispam.yahoo.com/tools
-
:::  When replying to the list, please use 'Reply-All' and make sure
:::  a copy goes to the list ([EMAIL PROTECTED]).
***  To unsubscribe, send 'unsubscribe' to [EMAIL PROTECTED]
***  For more information, check http://www.db2eug.uni.cc

Reply via email to