测试一体机ASM failgroup的相关问题处理

环境：3台虚拟机 RHEL 7.3 + Oracle RAC 11.2.0.4

问题现象：RAC运行正常，ASM磁盘组Normal冗余，有failgroup整体故障，有failgroup配置错误。

温馨提示：本文并不是市场上任何一款商业的一体机产品，只是我为了学习这类分布式存储架构自己模拟的实验环境，为了区分我暂时称之为xData吧^_。

• 1.问题现象确认
• 2.重新加入CELL01的盘
• 3.修改failgroup的配置

1.问题现象确认

SQL> select group_number, name, total_mb, free_mb, USABLE_FILE_MB, offline_disks, state, type from v$asm_diskgroup;

GROUP_NUMBER NAME                             TOTAL_MB    FREE_MB USABLE_FILE_MB OFFLINE_DISKS STATE                  TYPE
------------ ------------------------------ ---------- ---------- -------------- ------------- ---------------------- ----------
           1 CRS                                  2000       1170            585             0 MOUNTED                NORMAL
           2 DATA                                40960      35652           7586             0 MOUNTED                NORMAL

SQL>  select group_number, disk_number, name, path, failgroup, mode_status, voting_file  from v$asm_disk order by 1, 2;

GROUP_NUMBER DISK_NUMBER NAME                           PATH                    FAILGROUP            MODE_STATUS    VO
------------ ----------- ------------------------------ ----------------------- -------------------- -------------- --
           0           0                                /dev/CELL01-data2                            ONLINE         N
           0           1                                /dev/CELL01-data1                            ONLINE         N
           0           2                                /dev/CELL01-crs1                             ONLINE         Y
           1           1 CRS_0001                       /dev/CELL02-crs2        CRS_0001             ONLINE         Y
           1           2 CRS_0002                       /dev/CELL03-crs3        CRS_0002             ONLINE         Y
           2           0 DATA_0000                      /dev/CELL03-data1       DATA_0000            ONLINE         N
           2           1 DATA_0001                      /dev/CELL03-data2       DATA_0001            ONLINE         N
           2           2 DATA_0002                      /dev/CELL02-data1       CELL02               ONLINE         N
           2           3 DATA_0003                      /dev/CELL02-data2       CELL02               ONLINE         N

9 rows selected.

可以看到不但CELL01节点的所有盘被删除，而且CELL03节点的数据盘，failgroup目前也配置不正确！

2.重新加入CELL01的盘

由于时间超过默认的3.6h，offline的盘已经被删除，只有重新加入CELL01的盘才可以。

alter diskgroup CRS add disk '/dev/CELL01-crs1';
alter diskgroup DATA ADD FAILGROUP CELL01 disk '/dev/CELL01-data1', '/dev/CELL01-data2' rebalance power 5;

直接这样加盘很可能会遇到下面这类错误，因为这些盘之前是被使用过的：

SQL> alter diskgroup CRS add disk '/dev/CELL01-crs1';
alter diskgroup CRS add disk '/dev/CELL01-crs1'
*
ERROR at line 1:
ORA-15032: not all alterations performed
ORA-15033: disk '/dev/CELL01-crs1' belongs to diskgroup "CRS"

这个问题可以通过dd盘头，也可以加盘尝试加force参数来解决，我这里选择dd盘头的方式：

[root@db01 ~]# dd if=/dev/zero of=/dev/CELL01-crs1 bs=8k count=1000
1000+0 records in
1000+0 records out
8192000 bytes (8.2 MB) copied, 0.0691801 s, 118 MB/s

dd盘头之后再次尝试添加就可以顺利完成：

SQL> alter diskgroup CRS add disk '/dev/CELL01-crs1';

Diskgroup altered.

同样的，将CELL01的数据盘也重新加入到DATA磁盘组中，failgroup名称为CELL01:

SQL> alter diskgroup DATA ADD FAILGROUP CELL01 disk '/dev/CELL01-data1', '/dev/CELL01-data2' rebalance power 5;

Diskgroup altered.

通过v$asm_operation视图可以查看磁盘重新平衡的进度，直到下面的查询不再返回结果说明重平衡完成：

SQL> select * from v$asm_operation;

GROUP_NUMBER OPERATION  STATE         POWER     ACTUAL      SOFAR   EST_WORK   EST_RATE EST_MINUTES ERROR_CODE
------------ ---------- -------- ---------- ---------- ---------- ---------- ---------- ----------- --------------------
           2 REBAL      RUN               5          5        366        529        348           0
SQL> select * from v$asm_operation;

no rows selected

3.修改failgroup的配置

CELL03的数据盘，failgroup目前配置还不正确。

SQL> alter diskgroup DATA drop disk DATA_0000, DATA_0001;

Diskgroup altered.

查询v$asm_operation视图可以查看磁盘重新平衡的进度，完成后再重新加回磁盘组，指定确切的failgroup(CELL03)：

SQL> alter diskgroup DATA ADD FAILGROUP CELL03 disk '/dev/CELL03-data1', '/dev/CELL03-data2' rebalance power 5;

Diskgroup altered.

再次关注重平衡进度，最后查询一切正常，结果如下：

SQL> col path for a50
SQL> select group_number, disk_number, name, path, failgroup, mode_status, voting_file  from v$asm_disk order by 1, 2;

GROUP_NUMBER DISK_NUMBER NAME                           PATH                    FAILGROUP            MODE_STATUS    VO
------------ ----------- ------------------------------ ----------------------- -------------------- -------------- --
           1           0 CRS_0000                       /dev/CELL01-crs1        CRS_0000             ONLINE         Y
           1           1 CRS_0001                       /dev/CELL02-crs2        CRS_0001             ONLINE         Y
           1           2 CRS_0002                       /dev/CELL03-crs3        CRS_0002             ONLINE         Y
           2           0 DATA_0000                      /dev/CELL03-data1       CELL03               ONLINE         N
           2           1 DATA_0001                      /dev/CELL03-data2       CELL03               ONLINE         N
           2           2 DATA_0002                      /dev/CELL02-data1       CELL02               ONLINE         N
           2           3 DATA_0003                      /dev/CELL02-data2       CELL02               ONLINE         N
           2           4 DATA_0004                      /dev/CELL01-data1       CELL01               ONLINE         N
           2           5 DATA_0005                      /dev/CELL01-data2       CELL01               ONLINE         N

9 rows selected.

SQL> select group_number, name, total_mb, free_mb, USABLE_FILE_MB, offline_disks, state, type from v$asm_diskgroup;

GROUP_NUMBER NAME                             TOTAL_MB    FREE_MB USABLE_FILE_MB OFFLINE_DISKS STATE                  TYPE
------------ ------------------------------ ---------- ---------- -------------- ------------- ---------------------- ----------
           1 CRS                                  3000       2033            516             0 MOUNTED                NORMAL
           2 DATA                                61440      56012          17766             0 MOUNTED                NORMAL

说明：一般我会将磁盘组的兼容性参数设置为11.2，如有特殊需求，还可以设置disk_repair_time（默认3.6h）。

SQL> col COMPATIBILITY for a30
SQL> col DATABASE_COMPATIBILITY for a30
SQL> select NAME, COMPATIBILITY, DATABASE_COMPATIBILITY from v$asm_diskgroup;

NAME                           COMPATIBILITY                  DATABASE_COMPATIBILITY
------------------------------ ------------------------------ ------------------------------
CRS                            11.2.0.0.0                     11.2.0.0.0
DATA                           11.2.0.0.0                     11.2.0.0.0

--设置DATA磁盘组disk_repair_time(可理解为磁盘离线删除的时间)属性为4.5h
SQL> ALTER DISKGROUP data SET ATTRIBUTE 'disk_repair_time' = '4.5h';
Diskgroup altered.