We have removed LUNs from the SAN storage, but processes on the system still have them in use.

Issue

• We have removed LUNs from the SAN storage, but processes on the system still have them in use.
• Recently we decommissioned a few LUNs from servers (removed the zoning, so they are no longer visible to the fiber HBA). There are still some lingering processes attached to the decomussioned LUNs, it seems that those processes are still reading off of those LUNs:

$ lsof /dev/oracle/s3116_0E91
COMMAND   PID    USER   FD   TYPE DEVICE SIZE   NODE NAME
oracle   1862 oragrid  265u   BLK 253,90      175933 /dev/oracle/s3116_0E91
oracle  15078  dbacdb  295u   BLK 253,90      175933 /dev/oracle/s3116_0E91
$ ll /dev/mapper/mpath75
brw-rw---- 1 root disk 253, 90 Sep 26  2012 /dev/mapper/mpath75
$ dd if=/dev/oracle/s3116_0E91 of=/dev/null count=10 bs=1024
10+0 records in
10+0 records out
10240 bytes (10 kB) copied, 2.3e-05 seconds, 445 MB/s
$ dd if=/dev/oracle/s3116_0E91 iflag=direct of=/dev/null count=10 bs=1024
dd: reading `/dev/oracle/s3536_0E96': Input/output error
0+0 records in
0+0 records out
0 bytes (0 B) copied, 0.000136 seconds, 0.0 kB/s

The processes still stuck to these devices are all Oracle ASM and they are causing our alert_ASM log files to be filled with WARNING messages like this:

WARNING: Read Failed. group:0 disk:11 AU:0 offset:0 size:4096
Errors in file /opt/oraBase/diag/asm/+asm/+ASM2/trace/+ASM2_ora_26.trc:
ORA-27061: waiting for async I/Os failed
Linux-x86_64 Error: 5: Input/output error
Additional information: -1
Additional information: 4096
WARNING: Read Failed. group:0 disk:9 AU:0 offset:0 size:4096
Errors in file /opt/oraBase/diag/asm/+asm/+ASM2/trace/+ASM2_ora_26.trc:
ORA-27061: waiting for async I/Os failed
Linux-x86_64 Error: 5: Input/output error
Additional information: -1
Additional information: 4096
WARNING: Read Failed. group:0 disk:8 AU:0 offset:0 size:4096

• Is dropping the caches with echo 3 > /proc/sys/vm/drop_caches going to help?