在Suse 9下安装的Oracle 10.2老是出问题,隔三差五就出现oracle对其数据文件readonly.请各位能人帮忙解决一下!
具体环境: suse 9 enterprise server
oracle 10.2
SCSI
其中oracle的实例安装在SCSI存储设备上.
我怀疑linux对SCSI设备的支持有问题,或者是I/O太频繁造成的,不知正确与否.
这个是suse的message日志中出错信息:
May 16 15:54:08 linux kernel: mptbase: ioc0: IOCStatus(0x0043): SCSI Device Not There
May 16 15:54:08 linux kernel: SCSI error :return code = 0x10000
May 16 15:54:08 linux kernel: end_request: I/O error, dev sdb, sector 34415
May 16 15:54:08 linux kernel: Buffer I/O error on device sdb1, logical block 4294
May 16 15:54:08 linux kernel: lost page write due to I/O error on sdb1
May 16 15:54:08 linux kernel: mptbase: ioc0: IOCStatus(0x0043): SCSI Device Not There
May 16 15:54:08 linux kernel: SCSI error :return code = 0x10000
May 16 15:54:08 linux kernel: end_request: I/O error, dev sdb, sector 1480081575
May 16 15:54:08 linux kernel: Buffer I/O error on device sdb1, logical block 185010189
May 16 15:54:08 linux kernel: lost page write due to I/O error on sdb1
May 16 15:54:08 linux kernel: mptbase: ioc0: IOCStatus(0x0043): SCSI Device Not There
May 16 15:54:08 linux kernel: SCSI error :return code = 0x10000
May 16 15:54:08 linux kernel: end_request: I/O error, dev sdb, sector 1481668919
May 16 15:54:08 linux kernel: mptbase: ioc0: IOCStatus(0x0043): SCSI Device Not There
May 16 15:54:08 linux kernel: SCSI error :return code = 0x10000
May 16 15:54:08 linux kernel: end_request: I/O error, dev sdb, sector 1506416111
May 16 15:54:09 linux kernel: mptbase: ioc0: IOCStatus(0x0043): SCSI Device Not There
May 16 15:54:09 linux kernel: SCSI error :return code = 0x10000
May 16 15:54:09 linux kernel: end_request: I/O error, dev sdb, sector 34423
May 16 15:54:09 linux kernel: Buffer I/O error on device sdb1, logical block 4295
May 16 15:54:09 linux kernel: lost page write due to I/O error on sdb1
May 16 15:54:09 linux kernel: mptbase: ioc0: IOCStatus(0x0043): SCSI Device Not There
May 16 15:54:09 linux kernel: SCSI error :return code = 0x10000
May 16 15:54:09 linux kernel: end_request: I/O error, dev sdb, sector 1480081583
May 16 15:54:09 linux kernel: Buffer I/O error on device sdb1, logical block 185010190
May 16 15:54:09 linux kernel: lost page write due to I/O error on sdb1
May 16 15:54:09 linux kernel: mptbase: ioc0: IOCStatus(0x0043): SCSI Device Not There
May 16 15:54:09 linux kernel: SCSI error :return code = 0x10000
May 16 15:54:09 linux kernel: end_request: I/O error, dev sdb, sector 1480095375
May 16 15:54:09 linux kernel: Buffer I/O error on device sdb1, logical block 185011914
May 16 15:54:09 linux kernel: lost page write due to I/O error on sdb1
May 16 15:54:09 linux kernel: mptbase: ioc0: IOCStatus(0x0043): SCSI Device Not There
May 16 15:54:09 linux kernel: SCSI error :return code = 0x10000
May 16 15:54:09 linux kernel: end_request: I/O error, dev sdb, sector 1481668927
May 16 15:54:10 linux kernel: mptbase: ioc0: IOCStatus(0x0043): SCSI Device Not There
May 16 15:54:10 linux kernel: SCSI error :return code = 0x10000
May 16 15:54:10 linux kernel: end_request: I/O error, dev sdb, sector 1506416119
May 16 15:54:10 linux kernel: mptbase: ioc0: IOCStatus(0x0002): Busy
May 16 15:54:10 linux kernel: REISERFS: abort (device sdb1): Journal write error in flush_commit_list
May 16 15:54:10 linux kernel: REISERFS: Aborting journal for filesystem on sdb1
具体在oracle的bdump中有如下alert:
ORA-1653: unable to extend table MYUSER.MYTABLE by 128 in
tablespace MYDATA
ORA-1653: unable to extend table MYUSER.MYTABLE by 8192 in
tablespace MYDATA
事实上,我在linux中使用df -h 可以看到sdb1有900多G的free space,同时查看tablespaces的各数据文件,读写属性正常,可就是oracle罢工.
需要强调的是,tablespace中各数据文件设置了自动增长,且不限制最大空间,单个文件可以有32G左右.
虽然提示SCSI Error,但是可以进入挂载SCSI的目录并查看各文件,只是不能write.我只有remount该SCSI 设备,才可以再次正常工作,但两三天后就又出现上述问题了.
|
