当用户加载mount一个之前可用的diskgroup 时，将在ASM的告警日志中看到下面的错误：
如果自己搞不定可以找ASKMACLEAN专业ORACLE数据库修复团队成员帮您恢复!
SQL> ALTER DISKGROUP ALL MOUNT Tue Jul 19 09:31:09 2005 Loaded ASM Library - Generic Linux, version 1.0.0 library for asmlib interface Tue Jul 19 09:31:09 2005 NOTE: cache registered group DBFILE_GRP number=1 incarn=0xc3fd9b7d NOTE: cache registered group FLASHBACK_GRP number=2 incarn=0xc40d9b7e NOTE: cache dismounting group 1/0xC3FD9B7D (DBFILE_GRP) NOTE: dbwr not being msg'd to dismount ERROR: diskgroup DBFILE_GRP was not mounted NOTE: cache dismounting group 2/0xC40D9B7E (FLASHBACK_GRP) NOTE: dbwr not being msg'd to dismount ERROR: diskgroup FLASHBACK_GRP was not mounted [oracle@vrh8 ~]$ oerr ora 15032 15032, 00000, "not all alterations performed" // *Cause: At least one ALTER DISKGROUP action failed. // *Action: Check the other messages issued along with this summary error. // [oracle@vrh8 ~]$ oerr ora 15063 15063, 00000, "ASM discovered an insufficient number of disks for diskgroup \"%s\"" // *Cause: ASM was unable to find a sufficient number of disks belonging to the // diskgroup to continue the operation. // *Action: Check that the disks in the diskgroup are present and functioning, // that the owner of the ORACLE binary has read/write permission to // the disks, and that the ASM_DISKSTRING initialization parameter // has been set correctly. Verify that ASM discovers the appropriate // disks by querying V$ASM_DISK from the ASM instance. //
主要是出现三个错误：
ORA-15032: not all alterations performed
ORA-15063: diskgroup “FLASHBACK_GRP” lacks quorum of 2 PST disks; 0 found
ORA-15063: diskgroup “DBFILE_GRP” lacks quorum of 2 PST disks; 0 found
检查V$ASM_DISK 的HEADER_STATUS 显示为PROVISIONED，甚至当这个disk没有被asmlib label过：
SQL> select path, MOUNT_STATUS, HEADER_STATUS, MODE_STATUS, STATE from v$asm_disk; PATH MOUNT_S HEADER_STATUS MODE_ST STATE ------------- ------- ------------- ------- -------- /dev/raw/raw1 CLOSED PROVISIONED ONLINE NORMAL
导致该问题的原因一般是硬件故障，或者由于存储的固件升级
使用kfed来检查disk header:
kfed read /dev/raw/raw1
kfbh.type: 1 ; 0x002: KFBTYP_DISKHEAD
kfdhdb.driver.provstr: ORCLDISKASM1 ; 0x000: length=12
kfdhdb.grptyp: 1 ; 0x026: KFDGTP_EXTERNAL
kfdhdb.hdrsts: 3 ; 0x027: KFDHDR_MEMBER
kfdhdb.dskname: ASM1 ; 0x028: length=4
kfdhdb.grpname: DBFILE_GRP ; 0x048: length=10
kfdhdb.fgname: ASM1 ; 0x068: length=4
kfdhdb.capname: ; 0x088: length=0
kfdhdb.dskname: ASM1 ; 0x028: length=4
kfdhdb.grpname: DBFILE_GRP ; 0x048: length=10
kfdhdb.fgname: ASM1 ; 0x068: length=4
kfdhdb.capname: ; 0x088: length=0
观察kfdhdb.hdrsts: 3 ; 0x027: KFDHDR_MEMBER， KFDHDR_MEMBER说明其header状态实际是MEMBER，而V$ASM_DISK.HEADER_STATUS 则显示为PROVISIONED，2者并不匹配。
在disk header中kfdhdb.hdrsts 标记了本disk的状态， 下表显示了几个状态的描述：
| kfdhdb.hdrsts | 描述 |
| MEMBER | 属于当前diskgroup的disk |
| FORMER | 这个disk以前属于一个diskgroup,现在这个diskgroup被删除了 |
| CANDIDATE | 当使用裸设备，一个新的可以被diskgroup所用的disk |
| PROVISIONED | 当使用asmlib，一个新的可以被diskgroup所用的disk |
如果kfed read发现其状态为0x027: KFDHDR_MEMBER，则V$ASM_DISK不应当显示PROVISIONED。
我们发现当checksum不正确时可能导致V$ASM_DISK.HEADER_STATUS显示为PROVISIONED。
例如当发生硬件故障导致header的部分显得正常而仅仅checksum不正确。那么我们可以通过kfed来修复该问题的， kfed不愧和 amdu 、adhu合称ASM 三神器。
Comment