asm 磁盘分区丢失恢复----惜分飞

Posted 2021-01-22 xifenfei

tags:

篇首语：本文由小常识网(cha138.com)小编为大家整理，主要介绍了asm 磁盘分区丢失恢复----惜分飞相关的知识，希望对你有一定的参考价值。

有朋友反馈，他们做了xx存储的双活之后，重启主机发现gi无法正常启动，分析发现所有该存储的磁盘分区信息丢失，导致asmlib无法发现磁盘(使用分区做asm disk)
类似如下错误(磁盘分区丢失)

--fdisk -l 显示部分结果
Disk /dev/mapper/datahds1: 1099.5 GB, 1099511627776 bytes
255 heads, 63 sectors/track, 133674 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x00000000
 
--ls -l /dev/mapper/   显示结果无分区信息
lrwxrwxrwx 1 root root      7 May  6 03:44 datahds1 -> ../dm-1
lrwxrwxrwx 1 root root      7 May  6 03:26 datahds2 -> ../dm-3
lrwxrwxrwx 1 root root      7 May  6 03:26 datahds3 -> ../dm-8
lrwxrwxrwx 1 root root      7 May  6 03:26 ocrhds1 -> ../dm-0
lrwxrwxrwx 1 root root      7 May  6 03:26 ocrhds2 -> ../dm-2
lrwxrwxrwx 1 root root      7 May  6 03:26 ocrhds3 -> ../dm-4

asm日志显示

SUCCESS: diskgroup DATADG was mounted
NOTE: Instance updated compatible.asm to 11.2.0.0.0 for grp 3
SUCCESS: diskgroup OCRHDS was mounted
ORA-15032: not all alterations performed
ORA-15017: diskgroup "DATA" cannot be mounted
ORA-15063: ASM discovered an insufficient number of disks for diskgroup "DATA"

分析系统日志

May  6 02:23:27 db2 kernel: sdb: unknown partition table
May  6 02:23:27 db2 kernel: sde: unknown partition table
May  6 02:23:27 db2 kernel: sdc: unknown partition table
May  6 02:23:27 db2 kernel: sdf: unknown partition table
May  6 02:23:27 db2 kernel: sdd: unknown partition table
May  6 02:23:27 db2 kernel: sdj:Dev sdj: unable to read RDB block 0
May  6 02:23:27 db2 kernel: unable to read partition table
May  6 02:23:27 db2 kernel: sdi: sdi1
May  6 02:23:27 db2 kernel: sdk: sdk1
May  6 02:23:27 db2 kernel: sdg: unknown partition table
May  6 02:23:27 db2 kernel: sdl: sdl1
May  6 02:23:27 db2 kernel: sdm:Dev sdm: unable to read RDB block 0
May  6 02:23:27 db2 kernel: unable to read partition table
May  6 02:23:27 db2 kernel: sdo:Dev sdo: unable to read RDB block 0
May  6 02:23:27 db2 kernel: unable to read partition table
May  6 02:23:27 db2 kernel: sdn:Dev sdn: unable to read RDB block 0
May  6 02:23:27 db2 kernel: unable to read partition table
May  6 02:23:27 db2 kernel: sdp:Dev sdp: unable to read RDB block 0
May  6 02:23:27 db2 kernel: unable to read partition table
May  6 02:23:27 db2 kernel: sds:Dev sds: unable to read RDB block 0
May  6 02:23:27 db2 kernel: unable to read partition table
May  6 02:23:27 db2 kernel: sdh:
May  6 02:23:27 db2 kernel: sdt: sdt1
May  6 02:23:27 db2 kernel: sdv:Dev sdv: unable to read RDB block 0
May  6 02:23:27 db2 kernel: unable to read partition table
May  6 02:23:27 db2 kernel: sdq:Dev sdq: unable to read RDB block 0
May  6 02:23:27 db2 kernel: unable to read partition table
May  6 02:23:27 db2 kernel: sd 1:0:1:9: [sdr] Very big device. Trying to use READ CAPACITY(16).
May  6 02:23:27 db2 kernel: sdr:Dev sdr: unable to read RDB block 0
May  6 02:23:27 db2 kernel: unable to read partition table
May  6 02:23:27 db2 kernel: sd 2:0:0:9: [sdab] Very big device. Trying to use READ CAPACITY(16).
May  6 02:23:27 db2 kernel: sdab: unknown partition table
May  6 02:23:27 db2 kernel: sdac: unknown partition table
May  6 02:23:27 db2 kernel: sdw: sdw1
May  6 02:23:27 db2 kernel: sdu:Dev sdu: unable to read RDB block 0
May  6 02:23:27 db2 kernel: unable to read partition table
May  6 02:23:27 db2 kernel: sdx: sdx1
May  6 02:23:27 db2 kernel: sdy: sdy1
May  6 02:23:27 db2 kernel: sdaa: sdaa1
May  6 02:23:27 db2 kernel: sdz: sdz1
May  6 02:23:27 db2 kernel: sdae: unknown partition table
May  6 02:23:27 db2 kernel: sdaf: unknown partition table
May  6 02:23:27 db2 kernel: sdag: unknown partition table
May  6 02:23:27 db2 kernel: sdai:
May  6 02:23:27 db2 kernel: sdah: unknown partition table
May  6 02:23:27 db2 kernel: sdad: unknown partition table
May  6 02:23:28 db2 mcelog: failed to prefill DIMM database from DMI data

这里错误比较明显unknown partition table，磁盘的分区信息损坏.使用fdisk无法发现分区

partprobe也无效

[[email protected] oracle]# partprobe /dev/mapper/ocrhds3
[[email protected] oracle]#
[[email protected] oracle]# ls -l /dev/mapper/ocrhds3*
lrwxrwxrwx 1 root root 7 May  6 07:30 /dev/mapper/ocrhds3 -> ../dm-4

从尚需信息看，磁盘的分区表信息应该已经损坏，现在能够做的，就是希望运气好，磁盘的分区的实际数据没有损坏

分析磁盘实际分区数据

[[email protected] ~]$ dd if=/dev/mapper/datahds1 of=/tmp/datahds1.dd bs=1024k count=50

[[email protected] ~]$ dd if=/tmp/datahds1.dd of=/tmp/xff01.dd bs=3225 skip=1

[[email protected] ~]$ kfed read /tmp/xff01.dd |more

kfbh.endian: 1 ; 0x000: 0x01

kfbh.hard: 130 ; 0x001: 0x82

kfbh.type: 1 ; 0x002: KFBTYP_DISKHEAD

kfbh.datfmt: 1 ; 0x003: 0x01

kfbh.block.blk: 0 ; 0x004: blk=0

kfbh.block.obj: 2147483648 ; 0x008: disk=0

kfbh.check: 3110278718 ; 0x00c: 0xb963163e

kfbh.fcn.base: 0 ; 0x010: 0x00000000

kfbh.fcn.wrap: 0 ; 0x014: 0x00000000

kfbh.spare1: 0 ; 0x018: 0x00000000

kfbh.spare2: 0 ; 0x01c: 0x00000000

kfdhdb.driver.provstr: ORCLDISKHDSDATA1 ; 0x000: length=16

kfdhdb.driver.reserved[0]: 1146307656 ; 0x008: 0x44534448

kfdhdb.driver.reserved[1]: 826364993 ; 0x00c: 0x31415441

kfdhdb.driver.reserved[2]: 0 ; 0x010: 0x00000000

kfdhdb.driver.reserved[3]: 0 ; 0x014: 0x00000000

kfdhdb.driver.reserved[4]: 0 ; 0x018: 0x00000000

kfdhdb.driver.reserved[5]: 0 ; 0x01c: 0x00000000

kfdhdb.compat: 186646528 ; 0x020: 0x0b200000

kfdhdb.dsknum: 0 ; 0x024: 0x0000

kfdhdb.grptyp: 1 ; 0x026: KFDGTP_EXTERNAL

kfdhdb.hdrsts: 3 ; 0x027: KFDHDR_MEMBER

kfdhdb.dskname: DATADG_0000 ; 0x028: length=11

kfdhdb.grpname: DATADG ; 0x048: length=6

kfdhdb.fgname: DATADG_0000 ; 0x068: length=11

kfdhdb.capname: ; 0x088: length=0

kfdhdb.crestmp.hi: 33050696 ; 0x0a8: HOUR=0x8 DAYS=0x2 MNTH=0x4 YEAR=0x7e1

kfdhdb.crestmp.lo: 3813740544 ; 0x0ac: USEC=0x0 MSEC=0x44 SECS=0x35 MINS=0x38

kfdhdb.mntstmp.hi: 33050701 ; 0x0b0: HOUR=0xd DAYS=0x2 MNTH=0x4 YEAR=0x7e1

kfdhdb.mntstmp.lo: 411385856 ; 0x0b4: USEC=0x0 MSEC=0x150 SECS=0x8 MINS=0x6

通过上述分析，我们可以初步判断，分区磁盘的信息很可能是好的(因为asm disk header是好的，根据一般的规则从前往后覆盖，既然header是好的，后面的block被覆盖的概率非常小)

通过准备新磁盘直接把磁盘分区dd到新设备上

dd if=/dev/mapper/ocrhds1 of=/dev/mapper/ocrhdsnew1 skip=1 bs=3225
dd if=/dev/mapper/ocrhds2 of=/dev/mapper/ocrhdsnew2 skip=1 bs=3225
dd if=/dev/mapper/ocrhds3 of=/dev/mapper/ocrhdsnew3 skip=1 bs=3225
dd if=/dev/mapper/datahds1 of=/dev/mapper/datahdsnew1 skip=1 bs=3225
dd if=/dev/mapper/datahds2 of=/dev/mapper/datahdsnew2 skip=1 bs=3225
dd if=/dev/mapper/datahds3 of=/dev/mapper/datahdsnew3 skip=1 bs=3225

asmlib重新扫描磁盘

[[email protected] disks]# oracleasm scandisks

Reloading disk partitions: done

Cleaning any stale ASM disks...

Scanning system for ASM disks...

Instantiating disk "HDSOCR3"

Instantiating disk "HDSDATA2"

Instantiating disk "HDSDATA1"

Instantiating disk "HDSDATA3"

Instantiating disk "HDSOCR1"

Instantiating disk "HDSOCR2"

[[email protected] disks]# ls -ltr

total 0

brw-rw---- 1 grid asmadmin 8, 160 May 6 13:49 HDSOCR3

brw-rw---- 1 grid asmadmin 8, 192 May 6 13:49 HDSDATA2

brw-rw---- 1 grid asmadmin 8, 176 May 6 13:49 HDSDATA1

brw-rw---- 1 grid asmadmin 8, 208 May 6 13:49 HDSDATA3

brw-rw---- 1 grid asmadmin 8, 128 May 6 13:49 HDSOCR1

brw-rw---- 1 grid asmadmin 8, 144 May 6 13:49 HDSOCR2

kfed验证拷贝的分区

[[email protected] tmp]# /oracle/app/11.2.0/grid_1/bin/kfed read /dev/oracleasm/disks/HDSDATA1