Due to underneath SAN LUN issue, after manually dismounting problem disk group, the queries on v$asm_* views hang forever.
1) Check blocking sessions, and find blocking session is SID=426, and the final blocking session SID=392.
SQL> select SID,LAST_CALL_ET,EVENT,BLOCKING_SESSION,FINAL_BLOCKING_SESSION from v$session where BLOCKING_SESSION is not null; SID LAST_CALL_ET EVENT BLOCKING_SESSION FINAL_BLOCKING_SESSION 12 2313 enq: DD - contention 426 392 17 9526 enq: DD - contention 426 392 23 7726 enq: DD - contention 426 392 25 527598 enq: DD - contention 426 392 40 384765 enq: DD - contention 426 392 418 132675 enq: DD - contention 426 392 425 3397 enq: DD - contention 426 392 426 530330 rdbms ipc reply 392 392 444 15982 enq: DD - contention 426 392 ... .. .
2) Check what is session SID=392, which is ASM RBAL background process.
SQL> select s.sid,p.SPID,p.PNAME,p.BACKGROUND
from v$process p, v$session s
where s.PADDR=p.ADDR and s.sid in ( 426, 392);
SID SPID PNAME B
---------- ------------------------ ----- -
392 11670 RBAL 1
426 24638
$ ps -eaf | grep 11670
grid 11670 1 0 Mar22 ? 00:11:26 asm_rbal_+ASM1
3) Check which processes are still holding dismounted disk group.
$ lsof | grep ASMD_TEST oracleasm 1246 grid 4r BLK 253,10 0t0 18446612242469980264 /dev/oracleasm/disks/ASMD_TEST1 oracle_20 2087 grid 14r BLK 253,10 0t0 18446612242469980264 /dev/oracleasm/disks/ASMD_TEST1 oracle_33 3371 grid 14r BLK 253,10 0t0 18446612242469980264 /dev/oracleasm/disks/ASMD_TEST1 oracle_47 4754 grid 17r BLK 253,10 0t0 18446612242469980264 /dev/oracleasm/disks/ASMD_TEST1 ... .. . $ fuser /dev/oracleasm/disks/ASMD_TEST1 /dev/oracleasm/disks/ASMD_TEST1: 1246 2087 3371 4754 ... ... ..
In this case, ASM instance bounce is required for resolving this issue.
