VNX1 Series : SP Panic (bugcheck code 0000007E or 000000D1)

 Symptoms found in the logs:

From /sr/143377/SPA/SPA_CKMxxxxxxxxxxx_38b007_05-21-2017_00-07-15.analysis.txt:

Stack
=====

STACK_TEXT:
fffff881`8aa77618 fffff881`8c013368 : 00000000`00000000 00000000`00037234 00000000`00000001 00000000`0000af28 : RPSplitterDriver!ConnectionBase_io_send
fffff881`8aa77620 fffff881`8c015ef1 : fffffa80`3f438fe0 00000000`00000200 00000000`00000005 00000000`00000000 : RPSplitterDriver!RPSplitterVolume::io_send+0x94
fffff881`8aa776a0 fffff881`8c016353 : fffffa80`53bb92c0 fffffa80`3f438fe0 00000000`00000034 fffff880`0a9e4e80 : RPSplitterDriver!RPSplitterVolume::spl_sendToKbox+0x1a1
fffff881`8aa77740 fffff800`01ed28d1 : fffff880`0a9e5343 fffff884`00000007 00000000`00000000 fffffa80`0fc78480 : RPSplitterDriver!RPSplitterVolume::spl_IoDoneCasToStorage+0x22b

==> Access violation is done inside ConnectionBase_io_send.
This access violation can also happen from other functions. For example: ConnectionBase_io_cancel (bug 141763), ConnectionBase_uscsi_send, ...

FAULTING_IP:
RPSplitterDriver!ConnectionBase_io_send+0
fffff881`8c00d010 488b4108        mov     rax,qword ptr [rcx+8]

Issue :SP Panic (bugcheck code 0000007E or 000000D1)

Impacted Configuration & Settings [Replication Mode,Splitter Type,Compression,CDP/CLR/Multi,FC/IP/iSCSI, Policies etc.]:
VNX

Impact on RP:
Single disaster. Splitter will be down until SP is back up.

More details: Crash while trying to use a path to the RPA that was already released. This is done because pointer to the the path saved in each IO regardless of the path's state\health.

Affected versions:
All CX R32 Splitters.
Bug is more exposed in versions where BZ-42061 is fixed (for example in R33 - BowRiver and above)

Workaround:
Handle network issues to minimize communication issues between array and RPA

ResolutionFixed at: 

• VNX 05.33.009.5.217
• VNXe/Unity 4.2 SP2 (4.2.2.9632250)