ASM – basic things to look for

ASM is quite a big topic so the basic things to look for are pretty much the same – in that you still set the ‘. oraenv’ for environment variables. The log files are still under the $ORACLE_BASE/admin/+ASM/bdump etc. but there are other things to take into consideration

The DBA maybe called out because there needs to be a disk added or there maybe errors – ASM in 10g is a single point of failure so if ASM goes down on the node all instances on the node are gone as well… more over if ASM cannot mount a diskgroup then all instances and databases on the node will go down meaning a full outage!!! So understand the problems you can have and how to find your way around quickly…
What generally happens when adding disks is the following –

1.The SAN guy will present the disks

2. The unix person will then run config manager down the fiber nad this will present hdisks from lspv
3. The unix guy will then need to do a mknod c of the hdisks, which will create the ‘/dev/asm’ name

4. The unix guy will then initialise the header and remove the IBM header using the dd if=,something> of= command

5.Once this is all done then and only then will you be able to see the disks in ASM
once this is all done then and only then will you be able to see the disks in ASM

List disk (lspv and SQLPLUS asm views) and adding disks
lspv lists the disks at an OS point of view… the disks must be here before they can be added to ASM

p13504xxxx20-> lspv
hdisk0 00cfbdad335d716f rootvg active

hdisk1 00cfbdad41c6b4e5 rootvg active

hdisk2 00cfbdad57a03fc9 lo_oracvg_301 active

hdisk3 00cfbdad7bc7c818 lo_oracvg_301 active

hdisk4 none None

hdisk5 none None

hdisk6 none None

hdisk7 none None

hdisk8 none None

You can then query the v$asm_disk view in asm to see if any disks are ready to be added i.e. they are ‘candidates’

select group_number, state, name, path, header_status from v$asm_disk
SQL> /

GROUP_NUMBER STATE NAME PATH HEADER_STATU
------------ -------- --------------- ------------------------- ------------
0 NORMAL /dev/p2_asmdisk_20 CANDIDATE

0 NORMAL /dev/p2_asmdisk_19 CANDIDATE

0 NORMAL /dev/p3_asmdisk_14 CANDIDATE

0 NORMAL /dev/p3_asmdisk_13 CANDIDATE

0 NORMAL /dev/p3_asmdisk_12 CANDIDATE

0 NORMAL /dev/p3_asmdisk_11 CANDIDATE

0 NORMAL /dev/p3_asmdisk_21 CANDIDATE

0 NORMAL /dev/p3_asmdisk_20 CANDIDATE

0 NORMAL /dev/p3_asmdisk_19 CANDIDATE

0 NORMAL /dev/p3_asmdisk_18 CANDIDATE

0 NORMAL /dev/p3_asmdisk_17 CANDIDATE

0 NORMAL /dev/p3_asmdisk_16 CANDIDATE

0 NORMAL /dev/p3_asmdisk_15 CANDIDATE

0 NORMAL /dev/p2_asmdisk_11 CANDIDATE

0 NORMAL /dev/p2_asmdisk_12 CANDIDATE

0 NORMAL /dev/p2_asmdisk_13 CANDIDATE

0 NORMAL /dev/p2_asmdisk_14 CANDIDATE

0 NORMAL /dev/p2_asmdisk_21 CANDIDATE

0 NORMAL /dev/p2_asmdisk_15 CANDIDATE

0 NORMAL /dev/p2_asmdisk_16 CANDIDATE

0 NORMAL /dev/p2_asmdisk_17 CANDIDATE

0 NORMAL /dev/p2_asmdisk_18 CANDIDATE

1 NORMAL DATADG_0005 /dev/p2_asmdisk_1 MEMBER

1 NORMAL DATADG_0004 /dev/p2_asmdisk_2 MEMBER

2 NORMAL FLASHDG_0004 /dev/p2_asmdisk_3 MEMBER

2 NORMAL FLASHDG_0005 /dev/p2_asmdisk_4 MEMBER

1 NORMAL DATADG_0001 /dev/p2_asmdisk_5 MEMBER

1 NORMAL DATADG_0002 /dev/p2_asmdisk_6 MEMBER

1 NORMAL DATADG_0009 /dev/p2_asmdisk_7 MEMBER

1 NORMAL DATADG_0013 /dev/p2_asmdisk_10 MEMBER

1 NORMAL DATADG_0015 /dev/p2_asmdisk_9 MEMBER

1 NORMAL DATADG_0014 /dev/p2_asmdisk_8 MEMBER

1 NORMAL DATADG_0000 /dev/p3_asmdisk_1 MEMBER

1 NORMAL DATADG_0003 /dev/p3_asmdisk_2 MEMBER

2 NORMAL FLASHDG_0000 /dev/p3_asmdisk_3 MEMBER

2 NORMAL FLASHDG_0003 /dev/p3_asmdisk_4 MEMBER

1 NORMAL DATADG_0006 /dev/p3_asmdisk_5 MEMBER

1 NORMAL DATADG_0007 /dev/p3_asmdisk_6 MEMBER

1 NORMAL DATADG_0008 /dev/p3_asmdisk_7 MEMBER

1 NORMAL DATADG_0011 /dev/p3_asmdisk_8 MEMBER

1 NORMAL DATADG_0012 /dev/p3_asmdisk_9 MEMBER

1 NORMAL DATADG_0010 /dev/p3_asmdisk_10 MEMBER

42 rows selected.


Add disks into ASM – dependant upon disks being presented first

Once the this disks are there as shown in the above outputs from v$asm_disk…


Alter diskgroup add disk ‘’;


Then from the v$asm_operation view you can see how long the rebalance will take – this can be speeded up using the power supplement…


select * from v$asm_operation;