关于fsck_vxfs

wanfei

有一个Veritas做的数据盘镜像，两块磁盘。其中一块磁盘物理有坏道，造成在它上面的volume被disable了，而这个块上有Disk1－01（文件块）Disk1－02（log块）。另一个磁盘的块是disk2-01。故障前，Disk1－02，Disk1－02的状态为stale。Disk1－01为IOfault。更换了一个新磁盘后，volume都正常了。\r\n但用mount时，出错。\r\nmount -F vxfs -o rw /dev/vx/dsk/MYSQLdg/backup_index /mysql\r\nvxfs mount: /dev/vx/dsk/MYSQLdg/backup_index is corrupted. needs checking\r\n在用命令\r\nfsck -F vxfs /dev/vx/dsk/MYSQLdg/backup_index\r\nvxfs fsck: file system had I/O error(s) on user data.\r\n\r\n再用命令format检查磁盘，没有坏块\r\n\r\n再用命令\r\n\r\nfsck -F vxfs -o full,nolog /dev/vx/dsk/MYSQLdg/backup_index\r\n\r\nvxfs fsck: file system had I/O error(s) on user data.\r\npass0 - checking structural files\r\npass1 - checking inode sanity and blocks\r\nsubr.c       301:       ASSERT(HELD_BP(bp) == 0) failed\r\nsubr.c       301:       ASSERT(HELD_BP(bp) == 0) failed\r\nfileset 999 primary inode 248 has invalid number of blocks (0)\r\nfileset 999 primary inode 248 has invalid block map\r\nfileset 999 primary inode 248 failed validation clear? (ynq)n\r\nsubr.c       301:       ASSERT(HELD_BP(bp) == 0) failed\r\nsubr.c       301:       ASSERT(HELD_BP(bp) == 0) failed\r\nfileset 999 primary inode 327 has invalid number of blocks (3632)\r\nfileset 999 primary inode 327 has invalid block map\r\nfileset 999 primary inode 327 failed validation clear? (ynq)y\r\nreadi.c      280:       ASSERT(ic->bp == bp) failed\r\nvxfs fsck: fsck write failure bno = -1, off = 0, len = 0\r\nfile system check failure, aborting ...\r\n\r\n\r\n只要选y，就会退出，文件系统不能修复。\r\n\r\n\r\n怀疑故障前volume的信息，没有被删除掉，或者记录不完整。\r\n是不是必须重新做配置，生成一个新卷，再把原来的数据到回去\r\n\r\n看其它大哥还有什么好的建议。\r\n\r\n软件版本Veritas FS. Veritas Volume Manager. 3.1

wanfei

要不最坏只能试试\r\nnewfs dev/vx/rdsk/MYSQLdg/backup_index \r\n倒回原数据。看来没有定期做数据备份，只有死路一条啊