Hello Rick,
We were using similar RMAN scripts with TSM 5.4/5.5, TDPO 5.5 and TSM Client
5.x/6.2 for quite long time without any problems.
Each crosscheck creates just a few sessions without any performance problem.
By the way, I never use something like "setlimit channel tdp1 kbytes 1073741824
maxopenfiles 32 readrate 200;"
I think you need to ask your DBA to tune setlimit statement or try to remove it
at all.
This is my RMAN script which is working perfectly:
run {
#
# Allocate channel to Tivoli Storage Manager for database
allocate channel t1 type 'sbt_tape' parms
'ENV=(TDPO_OPTFILE=/oracle/base/admin/tdpo/$1.opt)';
#
# Set parameter for tablespace SYSTEM (block corruption is not allowed)
set maxcorrupt for datafile 1 to 0;
#
# Set date and time format
sql 'alter session set NLS_DATE_FORMAT = "YYYY-MM-DD:HH24:MI:SS"';
#
# Backup database to TSM
backup
incremental level $2
filesperset $FILES
tag='Database_$1.Level_$2.Date_$3'
format='/$HOST/$1.$3.$2.%s.%p.%t'
database
include current controlfile
;
#
# Switch and archive redo log files
sql 'ALTER SYSTEM SWITCH LOGFILE';
sql 'ALTER SYSTEM ARCHIVE LOG CURRENT';
#
# Backup archive redo logs
backup
filesperset=128
tag='Archive_$1.Level_$2.Date_$3'
format='/$HOST/Archive_$1.$3.%s.%p.%t'
(archivelog from logseq $START_SEQUENCE)
;
#
# Backup control file to trace and to disk
sql 'ALTER DATABASE BACKUP CONTROLFILE TO TRACE NORESETLOGS';
sql 'ALTER DATABASE BACKUP CONTROLFILE TO TRACE RESETLOGS';
sql "ALTER DATABASE BACKUP CONTROLFILE TO ''$4''";
#
# Release TSM channel
release channel t1;
}
#
# Allocate channel for maintenance
allocate channel for maintenance type 'sbt_tape' parms
'ENV=(TDPO_OPTFILE=/oracle/base//admin/tdpo/$1.opt)';
#
# Report unrecoverable data files
report unrecoverable database;
#
# Report data files required backup according to retention policy
report need backup database;
#
# Report number of redundant backups
report need backup redundancy 7 database;
#
# Report number of incremental backups
report need backup incremental=0 database;
#
# Delete obsolete backups
delete noprompt obsolete;
#
# Release TSM channel
release channel;
#
exit
EOF
-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:[email protected]] On Behalf Of Rick
Adamson
Sent: 29 05 2013 6:50 PM
To: [email protected]
Subject: [ADSM-L] TDP Oracle issue: thousands of near simultaneous sessions per
node
Looking for some direction...
I recently inherited a TSM server along with many TDP for Oracle nodes.
The TSM server is 5.4 (upgrade on the horizon).
The Oracle clients run on mostly AIX, but some Suse Linux, the TDP client is
ver.5.5.x and the BA clients are primarily 6.2.
All of the database and log backups are initiated by CRON and when they start a
single Oracle node will start and stop literally thousands of sessions on the
TSM server within a minute or two.
Oracle and RMAN are new to me and I have talked with our DBA team regarding the
problem but to no surprise they imply it is a TSM issue.
When this occurs the obvious impact to the TSM server is devastating,
performance is crushed and the console churn is a blur.
In reading the Oracle and TDP documentation I am suspicious of the RMAN scripts
as the cause, but before I go testing changes or opening a support case, I
thought I would ask the group.
The RMAN scripts do perform several "crosschecks" during the process and I
suspect that each backup piece it queries for starts and stops a client session
on the TSM server.
Has anyone seen this issue before? Is there any way to get the TDP client to
perform all operations during a single, or minimal, sessions?
I have included the RMAN script I suspect below.....
All feedback welcome,
~Rick
===================================================
rman target backup/sqbkup3 catalog rman10/sqbkup3@rcat1_catalog.world <<EOF #>
/opt/oracle/log/rman_full_csp1.log 2>&1 run { allocate channel tdp1 type
'sbt_tape'
parms 'ENV=(TDPO_OPTFILE=/opt/tivoli/tsm/client/oracle/bin64/tdpo.csp1.opt)';
setlimit channel tdp1 kbytes 1073741824 maxopenfiles 32 readrate 200;
sql 'alter system archive log current';
resync catalog;
change archivelog all validate;
backup database include current controlfile format 'df_%t_%s_%p'; backup
archivelog all delete input format 'al_%t_%s_%p'; release channel tdp1; } EOF #
rman <<EOF #> /opt/oracle/log/backup_delete_csp1.log 2>&1 connect target
backup/sqbkup3 connect catalog rman10/sqbkup3@rcat1_catalog.world
allocate channel for delete type 'sbt_tape'
parms 'ENV=(TDPO_OPTFILE=/opt/tivoli/tsm/client/oracle/bin64/tdpo.csp1.opt)';
configure retention policy to recovery window of 14 days; list backup of
database; list backup of database summary; list backup of archivelog all
summary; crosscheck backup; delete noprompt obsolete; delete noprompt expired
backup; crosscheck backup; release channel; allocate channel for delete type
disk; crosscheck backup; delete noprompt obsolete; delete noprompt expired
backup; crosscheck backup; list backup of database; EOF
Please consider the environment before printing this Email.
________________________________
CONFIDENTIALITY AND WAIVER: The information contained in this electronic mail
message and any attachments hereto may be legally privileged and confidential.
The information is intended only for the recipient(s) named in this message. If
you are not the intended recipient you are notified that any use, disclosure,
copying or distribution is prohibited. If you have received this in error
please contact the sender and delete this message and any attachments from your
computer system. We do not guarantee that this message or any attachment to it
is secure or free from errors, computer viruses or other conditions that may
damage or interfere with data, hardware or software.
