[Info-vax] OpenVMS 7.3-2, Patch V26.0 -> sh mem/cache=(vol=*, topqio=2) -> EXE crashing

[IanD]

Just wondering if the brains trust can help me on this one

*Before anyone suggests raising a ticket with HP, these systems have no support contract - please don't shoot the messenger here, this is not my doing...

I have two systems, both running OpenVMS 7.3-2, with patch VMS732_UPDATE V26.0, that result in an EXE dump when the following command is executed

sh mem/cache=(vol=*,topqio=2)

It works with 

sh mem/cache=(vol=*) or sh mem/cache=(vol=DSA1:) but as soon as I introduce any TOPQIO or TOPHIT, it goes belly up

One system is in a cluster and is the only system running V26 in that cluster, the others are 2 nodes are running V20.0 and they do not exhibit the EXE crash

The other system that is experiencing the EXE crashing is a standalone system, which has been through multiple upgrades over the years (well before my time), it was running V20 before the V26 patch was applied some months ago

Things in common on the systems where the EXE's crash:
- Both are running V26
- Both are on Charon
- something else I am not aware of (obviously)

I do not know if this issue occurred before the V26 update because this command was not something I tried before patching with V26, so it's either something amiss in V26 or perhaps Charon is a factor here or there is some weird and wonderful thing introduced in the patches between V20 and V26 that are outside the scope of V26 or ...

Here is the EXE crash if this provides more of a clue as to where it's breaking

>From the clustered system:

sh mem/cache=(vol=dsa20:,topqio)
              System Memory Resources on  1-JUL-2015 00:55:01.03

Extended File Cache Top QIO File Statistics:

_DSA20: (DISK$QUORUMSYS), Caching mode is VIOC Compatible
%SYSTEM-F-ACCVIO, access violation, reason mask=38, virtual address=100000000000001B, PC=0000000000012300, PS=00021D80

  Improperly handled condition, image exit forced.
    Signal arguments:   Number = 0000000000000003
                        Name   = 000000000000000C
                                 0000000000046638
                                 100000000000001B

    Register dump:
    R0  = 000000000000000C  R1  = 0000000000000000  R2  = 0000000000012300
    R3  = 0000000000021D80  R4  = 00000000000231D0  R5  = 0000000000000001
    R6  = 0000000000000001  R7  = 000000007AEAF9A8  R8  = 00000000007810D2
    R9  = 0000000000000400  R10 = 0000000000000100  R11 = 000000007AEAF538
    R12 = 000000007AEAF438  R13 = 000000007AEAF9B8  R14 = 000000000001BB9C
    R15 = 0000000000024838  R16 = 000000000000000C  R17 = 0000000000000000
    R18 = 000000007AEAF44D  R19 = 0000000000000028  R20 = 0000000000000001
    R21 = 000000007AEAF560  R22 = 0000000000000000  R23 = 0000000000000000
    R24 = 00000000000000B1  R25 = 0000000000000001  R26 = 0000000000046638
    R27 = 000000007B616F60  R28 = 100000000000001B  R29 = 000000007AEAF410
    SP  = 000000007AEAF410  PC  = 0000000000046638  PS  = 100000000000001B

>From the standalone system:

sh mem/cache=(vol=dsa1:,topqio)
              System Memory Resources on  1-JUL-2015 01:54:23.74

Extended File Cache Top QIO File Statistics:

_DSA1: (DISK$ISAACSYS1), Caching mode is VIOC Compatible
%SYSTEM-F-ACCVIO, access violation, reason mask=38, virtual address=100000000000001B, PC=0000000000012300, PS=00021D80

  Improperly handled condition, image exit forced.
    Signal arguments:   Number = 0000000000000003
                        Name   = 000000000000000C
                                 0000000000046638
                                 100000000000001B

    Register dump:
    R0  = 000000000000000C  R1  = 0000000000000000  R2  = 0000000000012300
    R3  = 0000000000021D80  R4  = 00000000000231D0  R5  = 0000000000000001
    R6  = 0000000000000001  R7  = 000000007AE0D9A8  R8  = 00000000007810D2
    R9  = 0000000000000400  R10 = 0000000000000100  R11 = 000000007AE0D538
    R12 = 000000007AE0D438  R13 = 000000007AE0D9B8  R14 = 000000000001BB9C
    R15 = 0000000000024838  R16 = 000000000000000C  R17 = 0000000000000000
    R18 = 000000007AE0D44D  R19 = 0000000000000028  R20 = 0000000000000001
    R21 = 000000007AE0D560  R22 = 0000000000000000  R23 = 0000000000000000
    R24 = 00000000000000B1  R25 = 0000000000000001  R26 = 0000000000046638
    R27 = 000000007B606F60  R28 = 100000000000001B  R29 = 000000007AE0D410
    SP  = 000000007AE0D410  PC  = 0000000000046638  PS  = 100000000000001B


They appear to be crashing at exactly the same point in the EXE?  giving all the dumped information is identical

Rolling back the patch isn't going to happen nor be allowed in these environments

Ideas as to where to look, what further things to dig into, greatly welcomed :-)