oracle导出表dump,有趣的大表导出产生coredump文件问题

db

racle8163 +ops

os:aix433

昨天晚上对数据库参数进行过调整

db_file_mutliblock_read_count 从16更改为8

compatible 从8.1.0 更改为8.1.6

gc_releasable_locks 现在64000，要求大于db_block_buffers

log_buffer 3145728>262144

distributed_transactions = 10>30

并将回滚段重建，由于原来的回滚段是initial=4m,next=8m,initial=2

我将它改成initial=next=256k initial=20

重新启动DB

对数据库的表进行EXP发现，一旦表很大，超过300M，备份就无法继续，并产生coredump.

DB运行一天后，在对表进行EXP，并设置10046进行跟踪

$ ps -ef|grep exp

oracle 38732  6778  85 18:03:06  pts/7  0:03 exp userid=dll/daf file=pay

mentev.dmp tables=payment feedback=10000

oracle 51180 43044   0 18:03:13  pts/0  0:00 grep exp

$ ps -ef|grep 38732

oracle 38732  6778  95 18:03:06  pts/7  0:06 exp userid=dll/daf file=pay

mentev.dmp tables=payment feedback=10000

oracle 42416 38732  47 18:03:06      -  0:03 oracleora82 (DESCRIPTION=(LOCAL=Y

ES)(ADDRESS=(PROTOCOL=beq)))

oracle 51182 43044   0 18:03:19  pts/0  0:00 grep 38732

$ ./whoit.sh 42416

SQL> exec dbms_system.set_ev(42, 29468, 10046, 12, '');

EXP结果：

$ exp userid=dll/daf file=paymentev.dmp tables=payment feedback=10000

Export: Release 8.1.6.0.0 - Production on Wed Apr 28 18:13:20 2004

(c) Copyright 1999 Oracle Corporation.  All rights reserved.

Connected to: Oracle8i Enterprise Edition Release 8.1.6.3.0, 64 bit - Production

With the Partitioning and Parallel Server options

JServer Release 8.1.6.3.0 - Production

Export done in US7ASCII character set and US7ASCII NCHAR character set

About to export specified tables via Conventional Path ...

. . exporting table                        PAYMENT

...........................................................................

...........................................................................

...........................................................................

...........................................

EXP-00002: error in writing to export file

EXP-00002: error in writing to export fileerror closing export file

EXP-00000: Export terminated unsuccessfully