[linux-lvm] Problems while trying to crash Oracle database

Fred Ruffet fred.ruffet at free.fr
Sun May 11 08:33:02 UTC 2003


You need to be sure Oracle reads where your dd owerwrote its data. While Oracle doesn't read, it won't crash.

Even if it reads, it depends on what have been recovered. You might just have errors in alert.log file telling you you have corrupted blocks, but the databse may not crash.

The more efficient would be to make sure dd writes over the system tablespace.

Regards,

On Thu, 01 May 2003 12:02:20 -0700, Steven Dake <sdake at mvista.com> told us :

> you could try kill -9 ? :)
> 
> HERUR,CHANNABASAPPA (HP-India,ex2) wrote:
> 
> >Hi ,
> >	I am basically trying to crash the oracle database which has been
> >created using raw logical volumes . When I use 'dd' command to crash the
> >oracle database I find that the database is actually not crashing . 
> >
> >	I used the following steps to create Oracle database on raw logical
> >volumes .
> >
> >
> >1. Created Physical Volumes's using pvcreate command 
> ># pvcreate /dev/sdp
> >pvcreate -- physical volume "/dev/sdp" successfully created
> >
> ># pvcreate /dev/sdq
> >pvcreate -- physical volume "/dev/sdq" successfully created
> >
> >2. Created Volume Groups's
> ># vgcreate vg06 /dev/sdp
> >vgcreate -- INFO: using default physical extent size 4.00 MB
> >vgcreate -- INFO: maximum logical volume size is 255.99 Gigabyte
> >vgcreate -- doing automatic backup of volume group "vg06"
> >vgcreate -- volume group "vg06" successfully created and activated
> >
> ># vgcreate vg07 /dev/sdq
> >vgcreate -- INFO: using default physical extent size 4.00 MB
> >vgcreate -- INFO: maximum logical volume size is 255.99 Gigabyte
> >vgcreate -- doing automatic backup of volume group "vg07"
> >vgcreate -- volume group "vg07" successfully created and activated
> >
> >3. Created Logical Volume's
> >
> ># lvcreate -l 25 -n control01.ctl /dev/vg06
> >lvcreate -- doing automatic backup of "vg06"
> >lvcreate -- logical volume "/dev/vg06/control01.ctl" successfully created
> >
> ># lvcreate -l 25 -n control02.ctl /dev/vg06
> >lvcreate -- doing automatic backup of "vg06"
> >lvcreate -- logical volume "/dev/vg06/control02.ctl" successfully created
> >
> ># lvcreate -l 100 -n system01.dbf /dev/vg06
> >lvcreate -- doing automatic backup of "vg06"
> >lvcreate -- logical volume "/dev/vg06/system01.dbf" successfully created
> >
> ># lvcreate -l 100 -n log01.log /dev/vg07
> >lvcreate -- doing automatic backup of "vg07"
> >lvcreate -- logical volume "/dev/vg07/log01.log" successfully created
> >
> ># lvcreate -l 100 -n log02.log /dev/vg07
> >lvcreate -- doing automatic backup of "vg07"
> >lvcreate -- logical volume "/dev/vg07/log02.log" successfully created
> >
> >4. Removed the existing raw devices
> >
> >[root at LNXSRVZ /dev]# rm /dev/raw/raw41
> >rm: remove `/dev/raw/raw41'? y
> >[root at LNXSRVZ /dev]# rm /dev/raw/raw42
> >rm: remove `/dev/raw/raw42'? y
> >[root at LNXSRVZ /dev]# rm /dev/raw/raw43
> >rm: remove `/dev/raw/raw43'? y
> >[root at LNXSRVZ /dev]# rm /dev/raw/raw44
> >rm: remove `/dev/raw/raw44'? y
> >[root at LNXSRVZ /dev]# rm /dev/raw/raw45
> >rm: remove `/dev/raw/raw45'? y
> >
> >5.Recreate raw devices using 'mknod' command
> >
> ># mknod /dev/vg06/rsystem01.dbf c 162 41
> ># mknod /dev/vg06/rcontrol01.ctl c 162 42
> ># mknod /dev/vg06/rcontrol02.ctl c 162 43
> ># mknod /dev/vg07/rlog01.log c 162 44
> ># mknod /dev/vg07/rlog02.log c 162 45
> >
> >
> >6. Used 'raw' command to bind raw device to block device
> >
> ># raw /dev/vg06/rsystem01.dbf /dev/vg06/system01.dbf
> >/dev/raw/raw41: bound to major 58, minor 19
> ># raw /dev/vg06/rcontrol01.ctl /dev/vg06/control01.ctl
> >/dev/raw/raw42: bound to major 58, minor 17
> ># raw /dev/vg06/rcontrol02.ctl /dev/vg06/control02.ctl
> >/dev/raw/raw43: bound to major 58, minor 18
> ># raw /dev/vg07/rlog01.log /dev/vg07/log01.log
> >/dev/raw/raw44: bound to major 58, minor 20
> ># raw /dev/vg07/rlog02.log /dev/vg07/log02.log
> >/dev/raw/raw45: bound to major 58, minor 21
> >
> >
> >7. CHANGED THE FILE PERMISSIONS
> ># chmod 766 /dev/vg06
> ># chmod 766 /dev/vg06/*
> ># chmod 766 -R /dev/vg07         
> ># chmod 766 -R /dev/vg07/*            
> ># chown oracle:oinstall -R /dev/vg07  
> >
> ># ll /dev/vg06/*
> >brwxrw-rw-    1 oracle   oinstall  58,  17 Apr 29 14:06
> >/dev/vg06/control01.ctl
> >brwxrw-rw-    1 oracle   oinstall  58,  18 Apr 29 14:06
> >/dev/vg06/control02.ctl
> >crwxrw-rw-    1 oracle   oinstall 109,   6 Apr 29 14:05 /dev/vg06/group
> >crwxrw-rw-    1 oracle   oinstall 162,  42 Apr 29 14:10
> >/dev/vg06/rcontrol01.ctl
> >crwxrw-rw-    1 oracle   oinstall 162,  43 Apr 29 14:10
> >/dev/vg06/rcontrol02.ctl
> >crwxrw-rw-    1 oracle   oinstall 162,  41 Apr 29 14:10
> >/dev/vg06/rsystem01.dbf
> >brwxrw-rw-    1 oracle   oinstall  58,  19 Apr 29 14:06
> >/dev/vg06/system01.dbf
> >
> ># ll /dev/vg07/*
> >crwxrw-rw-    1 oracle   oinstall 109,   7 Apr 29 14:05 /dev/vg07/group
> >brwxrw-rw-    1 oracle   oinstall  58,  20 Apr 29 14:06 /dev/vg07/log01.log
> >brwxrw-rw-    1 oracle   oinstall  58,  21 Apr 29 14:07 /dev/vg07/log02.log
> >crwxrw-rw-    1 oracle   oinstall 162,  44 Apr 29 14:10 /dev/vg07/rlog01.log
> >crwxrw-rw-    1 oracle   oinstall 162,  45 Apr 29 14:10 /dev/vg07/rlog02.log
> >
> >9. init$ORACLE_SID.ora file had the following contents
> >
> >db_name                         = rawlvm1
> >
> >db_files                        = 400
> >
> >db_file_multiblock_read_count   = 16
> >
> >db_block_buffers                = 550
> >
> >shared_pool_size                = 5000000
> >
> >log_checkpoint_interval         = 10000
> >
> >processes                       = 100
> >
> >parallel_max_servers            = 8
> >
> >log_buffer                      = 32768
> >
> >global_names                    = TRUE
> >
> >control_files                   = (/dev/raw/raw42, /dev/raw/raw43)
> >
> >db_block_checksum               = true
> >
> >db_block_size                   = 4096
> >
> >background_dump_dest            = /u00/app/oracle/admin/hard/rawlvm/bdump
> >
> >core_dump_dest                  = /u00/app/oracle/admin/hard/rawlvm/cdump
> >
> >user_dump_dest                  = /u00/app/oracle/admin/hard/rawlvm/udump
> >
> >10. Created oracle database
> >
> >SQL> startup nomount;
> >ORACLE instance started.
> >
> >Total System Global Area   53178448 bytes
> >Fixed Size                   450640 bytes
> >Variable Size              50331648 bytes
> >Database Buffers            2252800 bytes
> >Redo Buffers                 143360 bytes
> >SQL>
> >SQL> create database "rawlvm1"
> >  2    controlfile reuse
> >  3    maxinstances 8
> >  4    maxlogfiles 32
> >  5    datafile
> >  6         '/dev/raw/raw41' size 40M reuse
> >  7      logfile
> >  8         '/dev/raw/raw44'  size 20M ,
> >  9         '/dev/raw/raw45'  size 20M ;
> >
> >Database created.
> >
> >	Please let me know if there is something fundamentally wrong in the
> >above steps used for creating Oracle database . After creating the database
> >I used the following 'dd' command to crash the database 
> >
> >	dd if=/boot/vmlinux-2.4.2-2 of=/dev/sdp
> >
> >	Strangely this command does'nt seem to crash the database . Can
> >anybody help me out to crash the Oracle database ?
> >
> >
> >
> >_______________________________________________
> >linux-lvm mailing list
> >linux-lvm at sistina.com
> >http://lists.sistina.com/mailman/listinfo/linux-lvm
> >read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/
> >
> >
> >
> >  
> >
> 
> 
> _______________________________________________
> linux-lvm mailing list
> linux-lvm at sistina.com
> http://lists.sistina.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://tldp.org/HOWTO/LVM-HOWTO/


-- 
Fred Ruffet - fred.ruffet at free.fr

"When people say nothing, they don't necessarily mean nothing."




More information about the linux-lvm mailing list