RMAN恢復(fù)NBU磁帶庫的備份經(jīng)常報(bào)RMAN-10038和RMAN-03002錯(cuò)誤

Samdy_Chan

各位好：
      之前我們在Oracle RAC主庫做了NBU全備進(jìn)磁帶的備份，現(xiàn)在將其 restore database 還原數(shù)據(jù)文件到備庫時(shí)，經(jīng)常會(huì)報(bào)如下的錯(cuò)誤：
channel d2: restoring datafile 00178 to /share/mydb/datafile/xxt108.dbf
channel d2: restoring datafile 00260 to /share/mydb/datafile/xxtidx70.dbf
channel d2: restoring datafile 00343 to /share/mydb/datafile/xxtidx83.dbf
channel d2: restoring datafile 00408 to /share/mydb/datafile/xxtidx103.dbf
channel d2: reading from backup piece bk_10932_1_910658292
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================
RMAN-03002: failure of restore command at 05/06/2016 02:11:16
RMAN-10038: database session for channel d2 terminated unexpectedly


----
有時(shí)重試多一兩次又可以了，但到從別的 backup piece 恢復(fù)的時(shí)候，又報(bào)如上的錯(cuò)誤了。請問這是怎么回事？
恢復(fù)腳本如下：
$ cat rman_restore.sh
#!/bin/sh


export NLS_LANG="AMERICAN_AMERICA.ZHS16GBK"


rman target / log /home/oracle/myrman.log <<EOF
run {
  allocate channel d1 type 'SBT_TAPE';
  allocate channel d2 type 'SBT_TAPE';
  send channel d1 'NB_ORA_SERV=SCLOUD-GD-NBU,NB_ORA_CLIENT=mydb1,NB_ORA_POLICY=oracle_db_mydb1_full';
  send channel d2 'NB_ORA_SERV=SCLOUD-GD-NBU,NB_ORA_CLIENT=mydb1,NB_ORA_POLICY=oracle_db_mydb1_full';
  restore database;
  release channel d1;
  release channel d2;
}
EOF

------------
環(huán)境是，主備機(jī)都是 AIX 6.1 操作系統(tǒng)，數(shù)據(jù)庫版本是  11.2.0.4，NBU 服務(wù)器版本是 7.6.0.3 for Windows，NBU客戶端版本是 7.5 for AIX
麻煩哪位高人幫忙看看，感謝！