SA 5300A problems with ACU-XE config utility

[Richard Jordan]

I've got two DS10s with 5300A controllers being set up for VMS. They will run V7.3-2, fully updated, with the most recent Management Agents and ACU-XE for that version.

Due to the ORCA (ROM configurator) limitations and the lack of a spare drive, we created a single logical disk on a mirrored pair and built VMS on it with all patches, management agents, and ACU-XE, then used that to build up the remaining logical disks (smaller than the arrays they are on).

Then to rebuild the system disk logical disk, we backup/imaged to a local drive and after much struggle in tweaking to get ACU-XE running, deinstalled ACU-XE and the management agents, disabled SNMP, re-enabled SNMP, reinstalled the agents and ACU-XE, and were able to do the logical disk repartition we had to do to the array system disk.

Upon image copying back from the local disk to the new array system disk, we lost ACU-XE again. We also had to change the allocation class for disks on the system due to orders from above. Again tried deinstalling ACU-XE, management agents, and disabling/reconfiguring SNMP but now can't get ACU-XE to work no matter what. We even completely wiped the TCPIP configuration and rebuilt it from scratch, then reinstalled the layered products, still get the same problem. We can still run ACU-XE if we boot from the temporary local system disk. Everything else works regardless of which disk we boot from.

When we start ACU-XE with the command "@SYS$SYSROOT:[WBEM.ACUXE]CPQ$ACUXE.COM -R" we get the expected startup message and process ID, but the process dies immediately. The log is attached.

We don't have support for these systems yet, and C.O.V. apparently doesn't have any answers. Any thoughts would be appreciated... right now I'm leaning towards doing another from scratch VMS installation, since the second Alpha (whose array was configured initially with the local boot disk, and VMS built from scratch on) works fine. Sadly feels like a wintel solution... Somewhere in there is a hidden reference to a hardcoded disk name, or allocation class that is making things not work. I can't find it.

[Richard Jordan]

Sorry, missed this.

Versions:

OpenVMS Alpha V7.3-2, all class 1 and relevant class 2 ECOs installed

TCPIP V5.4 ECO 5

V7.3 Management Agents V3.3-1

- ECO V3.3-3 for above

ACU-XE V6.40-11P08

According to the current ADFU bootable disk, all drive firmware is up to spec. The 5300A is running the current firmware version as well (sorry don't have the number handy).

Thanks!

[Paul Dudson]

Hi Richard,
We had a similar problem but system was a DS25 with OpenVMS 8.2 and TCPIP 5.5 EC01.
Same access violation that you had after SA5302a was configured with multiple logical drives in the single Array.
Raised a software call and OpenVMS Engineering gave us a new copy of the the shareble images LIBCPQIMGR.EXE and LLPI.EXE with a new CPQ$ACUXE.COM. No rebuild was required and it all worked as expected.

Regards Paul

[Richard Jordan]

Paul,
thanks for the reply. Its good to know its at least a known problem.

I'll contact the hardware vendor to see if they can get the update for us since these systems are not under support yet so I can't go directly.

Rich

[Richard Jordan]

Appears there's nothing we can do until the systems go under support at the customer site. The vendor won't help us. Thanks for the info though; hopefully we'll be able to get the repair files and get this thing working again; in the meantime we're keeping a separate hot plug disk with a usable copy of VMS on it for array management.

[Richard Jordan]

See above