on 05-19-2010 1:24 PM
Dear Team ,
Our Production Server Offline backup taking more and more time.But archive Log and Online backup is Ok.
My database size is 160 GB Online backup 90 minutes and Offline backup takes 4 hours 40 Minutes.
I am using IBM tape libarary and using Ultrim 4 tape.
our initPRD.sap setting as below , should we change some one in initPRD.sap
HP-UX sapprd B.11.23 U ia64 (ta)
login: root
Password:
sapprd :/#cd /oracle/PRD/102_64/dbs/
sapprd :/oracle/PRD/102_64/dbs#ll
########################################################################
backup mode [all | all_data | full | incr | sap_dir | ora_dir
| all_dir | <tablespace_name> | <file_id> | <file_id1>-<file_id2>
| <generic_path> | (<object_list>)]
default: all
backup_mode = all
restore mode [all | all_data | full | incr | incr_only | incr_full
| incr_all | <tablespace_name> | <file_id> | <file_id1>-<file_id2>
| <generic_path> | (<object_list>) | partial | non_db
redirection with '=' is not supported here - use option '-m' instead
default: all
restore_mode = all
backup type [offline | offline_force | offline_standby | offline_split
| offline_mirror | offline_stop | online | online_cons | online_split
| online_mirror | online_standby | offstby_split | offstby_mirror
default: offline
backup_type = online
backup device type
[tape | tape_auto | tape_box | pipe | pipe_auto | pipe_box | disk
| disk_copy | disk_standby | stage | stage_copy | stage_standby
| util_file | util_file_online | rman_util | rman_disk | rman_stage
| rman_prep]
default: tape
backup_dev_type = tape
backup root directory [<path_name> | (<path_name_list>)]
default: $SAPDATA_HOME/sapbackup
backup_root_dir = /oracle/PRD/sapbackup
stage root directory [<path_name> | (<path_name_list>)]
default: value of the backup_root_dir parameter
stage_root_dir = /oracle/PRD/sapbackup
compression flag [no | yes | hardware | only]
default: no
compress = hardware
compress command
first $-character is replaced by the source file name
second $-character is replaced by the target file name
<target_file_name> = <source_file_name>.Z
for compress command the -c option must be set
recommended setting for brbackup -k only run:
"compress -b 12 -c $ > $"
no default
compress_cmd = "compress -c $ > $"
uncompress command
first $-character is replaced by the source file name
second $-character is replaced by the target file name
<source_file_name> = <target_file_name>.Z
for uncompress command the -c option must be set
no default
uncompress_cmd = "uncompress -c $ > $"
directory for compression [<path_name> | (<path_name_list>)]
default: value of the backup_root_dir parameter
compress_dir = /oracle/PRD/sapreorg
brarchive function [save | second_copy | double_save | save_delete
| second_copy_delete | double_save_delete | copy_save
| copy_delete_save | delete_saved | delete_copied]
default: save
archive_function = save
directory for archive log copies to disk
default: first value of the backup_root_dir parameter
archive_copy_dir = /oracle/PRD/sapbackup
directory for archive log copies to stage
default: first value of the stage_root_dir parameter
archive_stage_dir = /oracle/PRD/sapbackup
delete archive logs from duplex destination [only | no | yes | check]
default: only
archive_dupl_del = only
new sapdata home directory for disk_copy | disk_standby
no default
new_db_home = /oracle/C11
stage sapdata home directory for stage_copy | stage_standby
default: value of the new_db_home parameter
stage_db_home = /oracle/C11
original sapdata home directory for split mirror disk backup
no default
orig_db_home = /oracle/C11
remote host name
no default
remote_host = <host_name>
remote user name
default: current operating system user
remote_user = <user_name>
tape copy command [cpio | cpio_gnu | dd | dd_gnu | rman | rman_gnu
rman_dd | rman_dd_gnu]
default: cpio
tape_copy_cmd = dd
disk copy command [copy | copy_gnu | dd | dd_gnu | rman | rman_gnu]
default: copy
disk_copy_cmd = copy
stage copy command [rcp | scp | ftp]
default: rcp
stage_copy_cmd = rcp
pipe copy command [rsh | ssh]
default: rsh
pipe_copy_cmd = rsh
flags for cpio output command
default: -ovB
cpio_flags = -ovB
flags for cpio input command
default: -iuvB
cpio_in_flags = -iuvB
flags for cpio command for copy of directories to disk
default: -pdcu
use flags -pdu for gnu tools
cpio_disk_flags = -pdcu
flags for dd output command
default: "obs=16k"
recommended setting:
Unix: "obs=nk bs=nk", example: "obs=64k bs=64k"
Windows: "bs=nk", example: "bs=64k"
dd_flags = "obs=128k bs=128k"
flags for dd input command
default: "ibs=16k"
recommended setting:
Unix: "ibs=nk bs=nk", example: "ibs=64k bs=64k"
Windows: "bs=nk", example: "bs=64k"
dd_in_flags = "ibs=128k bs=128k"
number of members in RMAN save sets [ 1 | 2 | 3 | 4 | tsp | all ]
default: 1
saveset_members = 1
additional parameters for RMAN
rman_channels and rman_filesperset are only used when rman_util,
rman_disk or rman_stage
rman_channels defines the number of parallel sbt channel allocations
rman_filesperset = 0 means:
one file per save set - for non-incremental backups
all files in one save set - for incremental backups
the others have the same meaning as for native RMAN
rman_channels = 1
rman_filesperset = 0
rman_maxpiecesize = 0 # in KB - former name rman_kbytes
rman_rate = 0 # in KB - former name rman_readrate
rman_maxopenfiles = 0
rman_maxsetsize = 0 # in KB - former name rman_setsize
additional parameters for RMAN version 8.1
the parameters have the same meaning as for native RMAN
rman_diskratio = 0 # deprecated in Oracle 10g
rman_pool = 0
rman_copies = 0 | 1 | 2 | 3 | 4 # former name rman_duplex
rman_proxy = no | yes | only
special parameters for an external backup library, example:
rman_parms = "BLKSIZE=65536 ENV=(BACKUP_SERVER=HOSTNAME)"
rman_send = "'<command>'"
rman_send = ("channel sbt_1 '<command1>' parms='<parameters1>'",
"channel sbt_2 '<command2>' parms='<parameters2>'")
rman_compress = no | yes
rman_maxcorrupt = (<dbf_name>|<dbf_id>:<corr_cnt>, ...)
remote copy-out command (backup_dev_type = pipe)
$-character is replaced by current device address
no default
copy_out_cmd = "dd ibs=8k obs=128k of=$"
remote copy-in command (backup_dev_type = pipe)
$-character is replaced by current device address
no default
copy_in_cmd = "dd ibs=128k obs=8k if=$"
rewind command
$-character is replaced by current device address
no default
operating system dependent, examples:
HP-UX: "mt -f $ rew"
TRU64: "mt -f $ rewind"
AIX: "tctl -f $ rewind"
Solaris: "mt -f $ rewind"
Windows: "mt -f $ rewind"
Linux: "mt -f $ rewind"
rewind = "mt -f $ rew"
rewind and set offline command
$-character is replaced by current device address
default: value of the rewind parameter
operating system dependent, examples:
operating system dependent, examples:
Solaris: "mt -f $ fsf $"
Windows: "mt -f $ fsf $"
Linux: "mt -f $ fsf $"
tape_pos_cmd = "mt -f $ fsf $"
mount backup volume command in auto loader / juke box
used if backup_dev_type = tape_box | pipe_box
split_cmd = "<split_cmd> [$]"
resynchronize mirror disks command
used if backup_type = offline_split | online_split | offline_mirror
| online_mirror
no default
resync_cmd = "<resync_cmd> [$]"
additional options for SPLITINT interface program
no default
split_options = "<split_options>"
90 m 4 mm DAT DDS-1 tape: 1800M
120 m 4 mm DAT DDS-2 tape: 3800M
125 m 4 mm DAT DDS-3 tape: 11000M
112 m 8 mm Video tape: 2000M
112 m 8 mm high density: 4500M
DLT 2000 10/20 GB: 10000M
DLT 2000XT 15/30 GB: 15000M
DLT 4000 20/40 GB: 20000M
level of parallel execution
default: 0 - set to number of backup devices
exec_parallel = 0
address of backup device without rewind
[<dev_address> | (<dev_address_list>)]
no default
AIX: /dev/rmt0.1
Solaris: /dev/rmt/0mn
Windows: /dev/nmt0
Linux: /dev/nst0
tape_address = /dev/rmt/1mn
Solaris: /dev/rmt/0m
used by brarchive
default: value of the tape_address_rew_arch parameter
operating system dependent
tape_address_ctl_arch = /dev/...
volumes for brarchive
[<volume_name> | (<volume_name_list>) | SCRATCH]
no default
volume_archive = (PRDA01, PRDA02, PRDA03, PRDA04, PRDA05,
PRDA06, PRDA07, PRDA08, PRDA09, PRDA10,
PRDA11, PRDA12, PRDA13, PRDA14, PRDA15,
PRDA16, PRDA17, PRDA18, PRDA19, PRDA20,
PRDA21, PRDA22, PRDA23, PRDA24, PRDA25,
PRDA26, PRDA27, PRDA28, PRDA29, PRDA30)
volumes for brbackup
[<volume_name> | (<volume_name_list>) | SCRATCH]
no default
volume_backup = (PRDB01, PRDB02, PRDB03, PRDB04, PRDB05,
PRDB06, PRDB07, PRDB08, PRDB09, PRDB10,
PRDB11, PRDB12, PRDB13, PRDB14, PRDB15,
PRDB16, PRDB17, PRDB18, PRDB19, PRDB20,
PRDB21, PRDB22, PRDB23, PRDB24, PRDB25,
PRDB26, PRDB27, PRDB28, PRDB29, PRDB30)
default: no parameter file
mount_par_file = initPRD.mnt
#
Should I change any parameter
Sunil
Sunil, check if there is any third party connection to this database, for example a monitoring tool running a script to report on tablespace freespace.
Often times, when there is a database connection that doesn't let go (or keep reconnecting in quick succession), the database takes a while to shutdown for the offline backup.
The brbackup log files capture time very accurately. If you follow the timestamps in the logs, you can see where it takes the most time. That will be a starting point.
Regards,
Shan
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
In addition:
You might find a hint in the log files of brbackup, in sapbackup directory. Where is time lost?
Just a guess: Sometimes it will take long to shut down the database.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
> our initPRD.sap setting as below , should we change some one in initPRD.sap
Very difficult to read. I suggest you post only the lines that have entries, not all comments like
grep -v "\#" initPRD.sap
Markus
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Dear sir,
I am senidng initPRD.sap log file that was without # mark
sapprd :/oracle/PRD/102_64/dbs#more initPRD.sap
backup_mode = all
restore_mode = all
backup_type = online
backup_dev_type = tape
backup_root_dir = /oracle/PRD/sapbackup
stage_root_dir = /oracle/PRD/sapbackup
compress = hardware
compress_cmd = "compress -c $ > $"
uncompress_cmd = "uncompress -c $ > $"
compress_dir = /oracle/PRD/sapreorg
archive_function = save
archive_copy_dir = /oracle/PRD/sapbackup
archive_stage_dir = /oracle/PRD/sapbackup
disk_copy_cmd = copy
stage_copy_cmd = rcp
pipe_copy_cmd = rsh
cpio_flags = -ovB
cpio_in_flags = -iuvB
cpio_disk_flags = -pdcu
dd_flags = "obs=128k bs=128k"
saveset_members = 1
copy_out_cmd = "dd ibs=8k obs=128k of=$"
copy_in_cmd = "dd ibs=128k obs=8k if=$"
tape_pos_cmd = "mt -f $ fsf $"
tape_size = 800G
tape_address = /dev/rmt/1mn
tape_address_rew = /dev/rmt/1m
expir_period = 0
tape_use_count = 100
sapprd :/oracle/PRD/102_64/dbs#
Hi Sunil,
Your init<SID>.sap file looks bit strange for me.
Mainly "tape_copy_cmd" and "util_par_file" parameters are not seems to be configured.
You might not using any thirdparty backup library.
If you are using thirdparty backup solution then check the below things.
We are also using IBM Tivoli as Backup solution.
and below additional parameters configured in init<SID>.sap
backup_dev_type = util_file
tape_copy_cmd = dd
util_par_file = <path/to/util_file>
Check with your configuration.
One thing I have observed in our setup. If another backup is using tape drive and there is no FREE tape drive available, SAP backup waits for that backup to be finished. In that case SAP backup duration get extended...
Regards.
Rajesh Narkhede
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.