《丢失的ASM磁盘组》由会员分享，可在线阅读，更多相关《丢失的ASM磁盘组（10页珍藏版）》请在金锄头文库上搜索。
1、 .近日遇到一个问题，ASM的磁盘组无法挂载，之前是正常的，由于一些其他的操作，数据库启动失败，当问题排除时候，发现在数据库整体启动时，挂载磁盘组的环节出现问题。环境介绍#硬件：Vmware ESX虚拟机OS: Red hat linux 5Oracle version: 11.2.0.2ASM disk是通过 asmlib挂载的这个磁盘组只有一个虚拟出的硬盘，是 /dev/sdb1.#下面是我整个分析的过程1. 首先通过ASM alert.log，发现如下错误,磁盘挂载失败，无法找到磁盘组1. SQLalterdiskgroupDATAmount2. NOTE:cacheregistered
2、groupDATAnumber=1incarn=0xc28a1e2d3. NOTE:cachebeganmount(first)ofgroupDATAnumber=1incarn=0xc28a1e2d4. TueDec1118:06:5520125. ERROR:noPSTquorumingroup:required2,found06. NOTE:cachedismounting(clean)group1/0xC28A1E2D(DATA)7. NOTE:dbwrnotbeingmsgdtodismount8. NOTE:lgwrnotbeingmsgdtodismount9. NOTE:cac
3、hedismountedgroup1/0xC28A1E2D(DATA)10. NOTE:cacheendingmount(fail)ofgroupDATAnumber=1incarn=0xc28a1e2d11. NOTE:cachedeletingcontextforgroupDATA1/0xc28a1e2d12. GMONdismountinggroup1at8forpid17,osid3216313. ERROR:diskgroupDATAwasnotmounted14. ORA-15032:notallalterationsperformed15. ORA-15017:diskgroup
4、DATAcannotbemounted16. ORA-15063:ASMdiscoveredaninsufficientnumberofdisksfordiskgroupDATA17. ERROR:alterdiskgroupDATAmount 2. 首先检查ASM pfile 文件，未发现异常1. asm_diskgroups=DATA2. instance_type=asm3. large_pool_size=12M4. remote_login_passwordfile=EXCLUSIVE3. 尝试通过以下命令检查磁盘是否物理存在，是如何对应物理设备的，发现查询不到ASM磁盘1. gri
5、dlgto_test$kfoddisks=all2. 3. -Nonoutput-4. 5. 6. 7. gridlgto_testpeer$cd/dev/oracleasm/disks/8. gridlgto_testdisks$ls9. 10. -Nonoutput-11. 12. 13. gridlgto_testdisks$/etc/init.d/oracleasmlistdisks14. 15. -Nonoutput- 4.但是直接检查物理设备，/dev/sdb1是存在的，说明OS已经识别该硬盘设备，只是ASMlib无法正常识别:1. 查询对应的物理硬盘2. oracleOEL $
6、/etc/init.d/oracleasm querydisk -d disk1 Disk DISK1 is a valid ASM disk on device 8,173. oracleOEL $ ls -l /dev/ |grep 8|grep 17 brw-r- 1 root disk 8, 17 Oct 16 14:01 sdb14. 5. rootlgto_test#ls-lst/dev/sd*6. 0brw-r-1rootdisk8,0Dec1119:29/dev/sda7. 0brw-r-1rootdisk8,2Dec1119:29/dev/sda28. 0brw-r-1roo
7、tdisk8,16Dec1119:29/dev/sdb9. 0brw-r-1rootdisk8,17Dec1119:29/dev/sdb1Thisisthemisseddiskgroup10. 0brw-r-1rootdisk8,1Dec1111:29/dev/sda15.起先是考虑是否是磁盘头损害，导致无法ASMlib识别该磁盘， dump磁盘头发现没有问题.1. #od-c/dev/sdb12. 3. 0000040ORCLDISKDATADG014. 5. 7760040ORCLDISKDATADG01这里补充下，如果磁盘头信息丢失，将会显示如下1. 0000040ORCLDISK000
8、00000如果显示这个结果，需要通过以下方式renamedisk，具体可以参考文档Oracleasm Listdisks Cannot See Disks (Doc ID 392527.1)1. Usetheoracleasmrenamediskutilitytoaddanasmliblabeltothedisk:2. /etc/init.d/oracleasmrenamedisk/dev/3. Ifitfails,usethe-fswitch:4. /etc/init.d/oracleasmrenamedisk-f/dev/ 6.重启ASMLib ，检查是否是ASMLib 问题1. root
9、lgto_test#/etc/init.d/oracleasmrestart2. DroppingOracleASMLibdisks:3. OK4. ShuttingdowntheOracleASMLibdriver:FAILED检查文件系统oracleasm文件系统已经成功挂载1. rootlgto_test#df-ha2. FilesystemSizeUsedAvailUse%Mountedon3. 4. oracleasmfs000-/dev/oracleasm7.检查 /dev/sdb1状态，查看是否已经marked为ASM disk，显示已经标记成功1. rootlgto_test#
10、oracleasmquerydisk/dev/sdb12. Device/dev/sdb1ismarkedanASMdiskwiththelabelDATADG013. 4. rootlgto_test#/sbin/serviceoracleasmscandisks5. ScanningthesystemforOracleASMLibdisks:6. OK7. 8. rootlgto_test#/etc/init.d/oracleasmlistdisks9. -Nonoutput-8.检查 rpm package也没有问题1. gridlgto_test$rpm-qa|greporacleas
11、m2. oracleasmlib-2.0.4-1.el53. oracleasm-support-2.1.7-1.el54. oracleasm-2.6.18-308.el5-2.0.5-1.el5 9. 收集 Kfed logs,没有检查到异常新信息。1. rootlgto_test#/oracle/ora11g/product/app/grid/bin/kfedread/dev/sdb12. 3. kfbh.endian:1;0x000:0x014. 5. kfbh.hard:130;0x001:0x826. 7. kfbh.type:1;0x002:KFBTYP_DISKHEAD8. 9. kfbh.datfmt:1;0x003:0x0110. 11. kfbh.block.blk:0;0x004:T=0NUMB=0x012. 13. kfbh.block.obj:2147483648;0x008:TYPE=0x8NUMB=0x014. 15. kfbh.check:3351358462;0x00c:0xc7c1abfe16. 17. kfbh.fcn.base:0;0x010:0x0000000018. 19. kfbh.fcn.wrap:0;0x014:0x0000000020. 21. kfbh.spare1:0;0
