During VCS EMC SRDF Agent testing, pulling all fiber channel cables from a Solaris server causes total loss of access to a disk group(s).

The expectation was, the VCS EMC SRDF Agent would not monitor it's lock file, and then swap the EMC SRDF personalities (roles), making the primary site primary again.

Messages reported:

2008/10/29 12:58:55 VCS WARNING V-16-10001-1063 (bichette) DiskGroup:dg_BWPorabin:monitor:Disk Group: BWP_ora_bin_dg is disabled on system: bichette.
Not failing-over to protect data from potential corruption.

2008/10/29 12:58:56 VCS WARNING V-16-10001-1063 (bichette) DiskGroup:dg_BWPoraarch:monitor:Disk Group: BWP_ora_arch_dg is disabled on system: bichette.
Not failing-over to protect data from potential corruption


The issue has been fixed in 5.0 MP3.

In 5.0 MP3, when PanicSysOnDGLoss=1, the DiskGroup agent does not require fencing to be enabled in order to panic the system and force a failover, when the disk cable is removed.

With 5.0 MP1, the agent required fencing to be enabled.