rman ? nfs ? slow
lfree
edition ) ,I check to find
it is slow (vs Linux rhel 4.3 ). My nfs server use SFU for windows.
windows:
=> select * from v$version ;
BANNER
----------------------------------------------------------------
Oracle Database 10g Enterprise Edition Release 10.2.0.4.0 - 64bi
PL/SQL Release 10.2.0.4.0 - Production
CORE 10.2.0.4.0 Production
TNS for 64-bit Windows: Version 10.2.0.4.0 - Production
NLSRTL Version 10.2.0.4.0 - Production
SELECT start_time, end_time, status, input_type, elapsed_seconds,
compression_ratio, TRIM (input_bytes_display) inputb,
TRIM (output_bytes_display) outputb,
TRIM (input_bytes_per_sec_display) input,
TRIM (output_bytes_per_sec_display) output,
TRIM (time_taken_display) TIME
FROM v$rman_backup_job_details
WHERE start_time > TRUNC (SYSDATE) - 1;
START_TIME END_TIME STATUS INPUT_TYPE
ELAPSED_SECONDS COMPRESSION_RATIO INPUTB OUTPUTB INPUT OUTPUT TIME
------------------- ------------------- ---------------- ---------- --------
------- ----------------- -------- -------- -------- -------- ----------
2008-12-31 20:03:06 2008-12-31 20:19:37 COMPLETED ARCHIVELOG
991 2.59272 1.02G 402.88M 1.05M 416.29K 00:16:31
output_bytes_per_sec_display = 416.29K.
linux :
BANNER
----------------------------------------------------------------
Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 - 64bi
PL/SQL Release 10.2.0.3.0 - Production
CORE 10.2.0.3.0 Production
TNS for Linux: Version 10.2.0.3.0 - Production
NLSRTL Version 10.2.0.3.0 - Production
START_TIME END_TIME STATUS INPUT_TYPE
ELAPSED_SECONDS COMPRESSION_RATIO INPUTB OUTPUTB INPUT OUTPUT TIME
------------------- ------------------- ---------------- ---------- --------
------- ----------------- -------- -------- -------- -------- ----------
2009-01-01 02:10:17 2009-01-01 02:17:51 COMPLETED ARCHIVELOG
454 2.72805 2.91G 1.07G 6.56M 2.40M 00:07:34
2008-12-31 02:10:17 2008-12-31 02:16:25 COMPLETED ARCHIVELOG
368 2.59673 2.25G 888.76M 6.27M 2.42M 00:06:08
output_bytes_per_sec_display = 2.4M
in windows, I use to copy command , archive filesiez = 42M ,it takes only
2-3s。
copy arch... \\192.168.25.100\backup
why?
thank advanced.
Gerard H. Pille
You are taking a backup on system A (Linux or Whinedoze) while the data is in fact on system B
(your NFS)? Not such a good idea. In which case your problem could be network related.
I don't find it strange that Linux could beat Whinedoze on network throughput.
Gerard
Mark D Powell
>
> - Show quoted text -
On various versions of UNIX/Linux there are Oracle recommended mount
options for use with NFS file systems. (11g comes with an NFS driver
built into the Oracle Kernal). Check on your NFS configuration. It
could just be that the NFS device is not as fast on Windows as in a
UNIX environment or there may be some configuration changes you need
to make.
Consider using the parallel option in rman.
HTH -- Mark D Powell --
Michael Austin
You are using windows AND NFS and you are still wondering why it is
slow???
Steve Howard
Hi,
I see you are using compression. Are the CPU's saturated on the
Microsoft server?
Regards,
Steve
lfree
> edition ) ,I check to find
> it is slow (vs Linux rhel 4.3 ). My nfs server use SFU for windows.
>
> windows:
>
> => select * from v$version ;
> BANNER
> ----------------------------------------------------------------
> Oracle Database 10g Enterprise Edition Release 10.2.0.4.0 - 64bi
> PL/SQL Release 10.2.0.4.0 - Production
> CORE 10.2.0.4.0 Production
> TNS for 64-bit Windows: Version 10.2.0.4.0 - Production
> NLSRTL Version 10.2.0.4.0 - Production
>
> SELECT start_time, end_time, status, input_type, elapsed_seconds,
> compression_ratio, TRIM (input_bytes_display) inputb,
> TRIM (output_bytes_display) outputb,
> TRIM (input_bytes_per_sec_display) input,
> TRIM (output_bytes_per_sec_display) output,
> TRIM (time_taken_display) TIME
> FROM v$rman_backup_job_details
> WHERE start_time > TRUNC (SYSDATE) - 1;
>
> START_TIME END_TIME STATUS INPUT_TYPE
> ELAPSED_SECONDS COMPRESSION_RATIO INPUTB OUTPUTB INPUT OUTPUT TIME
> ------------------- ------------------- ---------------- ---------- ------
> ------- ----------------- -------- -------- -------- -------- ----------
> 2008-12-31 20:03:06 2008-12-31 20:19:37 COMPLETED ARCHIVELOG
> 991 2.59272 1.02G 402.88M 1.05M 416.29K 00:16:31
>
> output_bytes_per_sec_display = 416.29K.
>
>>Hi,
>>I see you are using compression. Are the CPU's saturated on the
>>Microsoft server?
YES. I use compression.
CPU is Intel Xenon X5460 3.16Ghz ,memory=8G.
lfree
> > ------- ----------------- -------- -------- -------- -------- ----------
> > 2008-12-31 20:03:06 2008-12-31 20:19:37 COMPLETED ARCHIVELOG
> > 991 2.59272 1.02G 402.88M 1.05M 416.29K 00:16:31
> >
> > output_bytes_per_sec_display = 416.29K.
> You are using windows AND NFS and you are still wondering why it is
> slow???
yes! My NFS server is a machine ,OS is windows 2003 .
Ronald Rood
(in article <gjjq40$gus$1...@news.cn99.com>):
>> I use rman backup product database on OS windows(2003 R2 standard 64
>> edition ) ,I check to find
>> it is slow (vs Linux rhel 4.3 ). My nfs server use SFU for windows.
>>
>> windows:
>>
Do you have resource manager enabled ?
I am working on an issue with a slow backup that happens when I have some
other plan then the internal_plan active. If you ask me, the cpu - or core -
count that resource manager is using is not the count we have in the system.
With a custom plan active we get 3MB/s, with the internal_plan we get 34MB/s
both cases are using 16 channel compressed backup sets to disk. The custom
plan does give unlimited cpu to the sys_group at level 1.
In my case running linux x86_64 (8 dual core Sun x4600) with 10.2.0.4
database.
If this is also your problem you might see a lot of scheduler wait events for
cpu quantum, while having plenty idle cpu left.
--
With kind regards / met vriendelijke groeten,
Ronald
lfree
begin backup xxx, running archive.rman : 2009-01-04 20:03:02.31
Recovery Manager: Release 10.2.0.4.0 - Production on Sun Jan 4 20:03:02 2009
Copyright (c) 1982, 2007, Oracle. All rights reserved.
connected to target database: ORCL (DBID=1196486461)
RMAN> run {
2> allocate channel c1 type disk ;
3> sql 'alter system archive log current ';
4>
5> backup as compressed backupset archivelog all
6> tag="archivelog"
7> not backed up 2 times
8> format '\\192.168.100.39\backup\archives\%d_%T_%s_archive' ;
9>
10> backup as compressed backupset current controlfile
11> tag="control_archive"
12> format '\\192.168.100.39\backup\controls\%d_%T_%s_control' ;
13> release channel c1 ;
14> }
15>
using target database control file instead of recovery catalog
allocated channel: c1
channel c1: sid=329 devtype=DISK
sql statement: alter system archive log current
Starting backup at 2009-01-04 20:03:06
current log archived
skipping archive log file
D:\ORACLE\PRODUCT\10.2.0\ARCHIVE\ARC00461_0668705725.001; already backed up
2 time(s)
skipping archive log file
D:\ORACLE\PRODUCT\10.2.0\ARCHIVE\ARC00462_0668705725.001; already backed up
2 time(s)
truncate ....
skipping archive log file
D:\ORACLE\PRODUCT\10.2.0\ARCHIVE\ARC00882_0668705725.001; already backed up
2 time(s)
channel c1: starting compressed archive log backupset
channel c1: specifying archive log(s) in backup set
input archive log thread=1 sequence=883 recid=1709 stamp=675122447
truncate ....
input archive log thread=1 sequence=901 recid=1745 stamp=675288189
channel c1: starting piece 1 at 2009-01-04 20:03:42
channel c1: finished piece 1 at 2009-01-04 20:15:57
piece handle=\\192.168.100.39\BACKUP\ARCHIVES\ORCL_20090104_253_ARCHIVE
tag=ARCHIVELOG comment=NONE
channel c1: backup set complete, elapsed time: 00:12:16
Finished backup at 2009-01-04 20:15:57
Starting backup at 2009-01-04 20:16:01
channel c1: starting compressed full datafile backupset
channel c1: specifying datafile(s) in backupset
including current control file in backupset
channel c1: starting piece 1 at 2009-01-04 20:16:02
channel c1: finished piece 1 at 2009-01-04 20:16:09
piece handle=\\192.168.100.39\BACKUP\CONTROLS\ORCL_20090104_254_CONTROL
tag=CONTROL_ARCHIVE comment=NONE
channel c1: backup set complete, elapsed time: 00:00:09
Finished backup at 2009-01-04 20:16:10
Starting Control File and SPFILE Autobackup at 2009-01-04 20:16:11
piece
handle=D:\ORACLE\PRODUCT\10.2.0\FLASH_RECOVERY_AREA\LIS\AUTOBACKUP\2009_01_0
4\O1_MF_S_675288971_4P1B4D3F_.BKP comment=NONE
Finished Control File and SPFILE Autobackup at 2009-01-04 20:16:15
released channel: c1
Recovery Manager complete.
end backup xxx, stoped archive.rman : 2009-01-04 20:16:19.95
I don't know why it is slow on windows + rman + nfs .
Gerard H. Pille
> My logfile :
>
> begin backup xxx, running archive.rman : 2009-01-04 20:03:02.31
>
> Finished backup at 2009-01-04 20:15:57
>
> Starting backup at 2009-01-04 20:16:01
...
> end backup xxx, stoped archive.rman : 2009-01-04 20:16:19.95
STOPED?
>
> I don't know why it is slow on windows + rman + nfs .
>
We know that you do not know, but a number of possible causes have been given.
Why don't you check them out, and come back with the results?
Gerard
lfree
>On various versions of UNIX/Linux there are Oracle recommended mount
>options for use with NFS file systems. (11g comes with an NFS driver
>built into the Oracle Kernal). Check on your NFS configuration. It
>could just be that the NFS device is not as fast on Windows as in a
>UNIX environment or there may be some configuration changes you need
>to make.
>Consider using the parallel option in rman.
>HTH -- Mark D Powell --
Thank your hint !!
I open parallel option .
rman > CONFIGURE DEVICE TYPE DISK PARALLELISM 2 BACKUP TYPE TO BACKUPSET;
and modify rman script .
run {
allocate channel c1 type disk ;
allocate channel c2 type disk ;
sql 'alter system archive log current ';
backup as compressed backupset archivelog all
tag="archivelog"
not backed up 2 times
format '\\192.168.100.39\backup\archives\%d_%T_%s_archive' ;
backup as compressed backupset current controlfile
tag="control_archive"
format '\\192.168.100.39\backup\controls\%d_%T_%s_control' ;
release channel c1 ;
release channel c2 ;
}
result diff:
START_TIME END_TIME STATUS INPUT_TYPE
ELAPSED_SECONDS COMPRESSION_RATIO INPUTB OUTPUTB INPUT OUTPUT TIME
------------------- ------------------- ---------------- ---------- --------
------- ----------------- -------- -------- -------- -------- ---------
2009-01-05 09:12:22 2009-01-05 09:13:53 COMPLETED ARCHIVELOG
91 2.55228 522.35M 204.66M 5.74M 2.25M 00:01:31
2009-01-04 20:03:06 2009-01-04 20:16:15 COMPLETED ARCHIVELOG
789 2.60098 708.50M 272.40M 919.52K 353.53K 00:13:09
it take only 1:31 minute. and now time 9:00-11:00 is peak load .
thank !
lfree
>STOPED
NO? this is my script bat .I think it must be replace 'finished' .
thank!
....
echo begin backup xxx, running %1 : %date:~0,10% %time% >> %logfile%
%ORACLE_HOME%\bin\rman target / log=%logfile% append @%1
echo end backup xxx, stoped %1 : %date:~0,10% %time% >> %logfile%
> > I don't know why it is slow on windows + rman + nfs .
> >
>
> We know that you do not know, but a number of possible causes have been
given.
>
> Why don't you check them out, and come back with the results?
>
> Gerard
In windows , using rman+nfs backup must be use format
'\\IP\backup\archives\%d_%T_%s_archive' ,
My test can not map driver .(example : format
'y:\backup\archives\%d_%T_%s_archive' )
I cat not input any nfs arguments.
I only test ,if mount (no options)
c:\>mount y: \\192.168.100.39\backup
c:\>mount
y: \\192.168.100.39\backup UID=500, GID=501
rsize=32768,
wsize=32768
mount=hard,
timeout=0.8
retry=1,
locking=yes
fileaccess=777,
lang=GB2312-80
casesensitive=no
I check those arguments , that is sameness to linux nfs .