昨天在客户现场遇到STATUS CODE: 52错误,当时的配置是这样的:
GS1280 TruCluster 5.1b
Netbackup 5.1MP6
Oracle 9i
ADIC i2000
Oracle Archived log的备份失败出52号错误,备份作业一直处于mounting tape磁带状态,然后我们用相同的带库备份一般文件也出52号错误.这样能断定和Oracle没有关系.后询问客户知道换了一次光纤线,但zone配置没有变.
因此基本任务和光纤环境变化有关,机械手正常,但驱动器有问题,重新配置恢复正常.
下面是veritas support网上的文章,可供参考
STATUS CODE: 52 "vmoprcmd -d" and "tpconfig -d" do not show the same set of drives; also, robotic tape mounts may hang or time out.
Details:
Overview:
The two commands, tpconfig and vmoprcmd, can both be run with the -d flag to determine drive status. tpconfig is used to confirm drive configuration. vmoprcmd is used to help determine the drive state. Both these commands ultimately pull drive configuration from the same place, the /usr/openv/volmgr/database/ltidevs file (robotic configuration is pulled from the robotic_defs file in the same directory). However, vmoprcmd is a representation of what was read from ltidevs/robotic_defs when ltid was last started and read that information into running memory tables.
Since both commands refer to the same Veritas NetBackup (tm) configuration file, the thinking goes that the list of hardware you see in the output of the two commands should always be the same. It is possible, however, for the lists to differ. The first thing to check is that ltid has been restarted after any device configuration changes have been made. If ltid has been restarted to re-read the configuration and the two outputs still differ, the problem could be incomplete drive entries or even possible corruption of the ltidevs configuration file.
Troubleshooting:
When this type of situation occurs, it is common for backups to hang up or time out with a status code 52 during tape mounts. This may effect only some drives, or all of them, depending on the extent of the problem. If tpconfig has knowledge of a tape drive then NetBackup will try to mount a tape in it. However, if that drive is missing in the vmoprcmd output, then that means avrd will never scan the drive, and NetBackup will never flag the tape as successfully mounted.
To verify this is the problem, the command to run to check the ltidevs entries is tpconfig -tables. Similar in function to tpconfig -d, it will print out highly detailed configuration of the drives within NetBackup. Below is an example comparison of what to look for when there is an incomplete drive entry. Drive 9 is correctly configured, but Drive 10 is not. Notice the flags for Drive 10 are "0x4080." The "0x4000" bit is active when it shouldn't be, which indicates a problem (the 0x80 bit is normal however, and it appears in Drive 9's configuration as it should).
Drive 9 is robotic, ACS(1), mount_time = 18547, clean_freq = 0
Drive name = 9840a-lsw-00A4
RewDrivePath = /dev/rmt/10cbn
VHDriveName =
Flags = 0x80
Drive GUID = 00000000-0000-0000-0000-000000000000
Serial Number = 331002052183
Inquiry String = STK 9840 130.
*** End of Drive ***
Drive 10 is robotic, ACS(0), mount_time = 0, clean_freq = 0
Drive name = 9840a-hoard-0091
RewDrivePath = /dev/rmt/11cbn
VHDriveName =
Flags = 0x4080
Drive GUID = 00000000-0000-0000-0000-000000000000
Serial Number = 331002033508
Inquiry String = STK 9840 1.30
*** End of Drive ***
In this case, simply delete and re-add the drives, preferably through the wizard or tpautoconf.
Log Files: N/A
Resolution:
The solution in this case is to delete and re-add the drives within NetBackup. This can be done using the Device Discovery Wizard in the NetBackup Administration Console or by using the command line tpautoconf command.
阅读(2279) | 评论(0) | 转发(0) |
