issue on write to tape via tmscp after OVMS upgrade from ovms v7.3-1 to vms v7.3-2

[FELIX WONG_1]

I've had an issue with the tape backup via tmscp on my alpha server 4100 after upgraded from OVMS v7.3-1 to v7.3-2.

It was working ok with OVMS v7.3-1 when I ran backup (own program, not vms backup or copy utilities) from local node using the 4mm tape drive on another node within a 2 nodes cluster environment with TMSCP enabled.

Unfortunately, after the OVMS upgrade to v7.3-2, it returns with error "error writing to tape" half the way during the backup or "volume not software enable".
However, the backup was ok when I ran the backup from that node, using the local tape drive there.
It's the same tape drive, same media.

Appears after the vms v7.3-2 upgrade, either the TMSCP has an issue or something broke.

Can someone advise !

[Duncan Morris]

Felix,

welcome to the ITRC OpenVMS forum!

You say that you have upgraded to 7.3-2, but do not indicate what your patch level is.

As 7.3-2 is quite old, there have been an awful lot of patches issued, including several significant changes to backup.

Please check against the patch database. There is a recent consolidated update (V8) available which covers many of the items.
File Name: ALPHA_V732_MASTER_ECO_LIST
Release Date: 2-NOV-2006

CURRENT UPDATE KIT RELEASE DATE
------------------ ------------
VMS732_UPDATE-V0800 14-SEP-2006

PROJECTED UPDATE KIT RELEASE SCHEDULE
-------------------------------------
December 4, 2006
March 4, 2007
June 4, 2007
September 11, 2006

KITS INCLUDED IN UPDATE KIT:
===========================
(note that kits included in the UPDATE kit are available for
individual download)

KIT NAME RELEASE DATE SUPERSEDED
----------------------------- ------------ ----------
VMS732_AMATHRTL-V0100 31-MAY-2005
VMS732_AUDSRV-V0200 16-AUG-2005
VMS732_BACKUP-V0500 14-APR-2006 Yes
VMS732_CLUSTER-V0100 6-JUN-2006
VMS732_COPY-V0100 6-APR-2005
VMS732_CMATIS-V0100 13-JUL-2005
VMS732_CPU2308-V0100 7-JUL-2006
VMS732_CPU270F-V0100 4-JAN-2005
VMS732_DCL-V0700 22-FEB-2006
VMS732_DDTM-V0100 8-NOV-2005
VMS732_DISMOU-V0100 24-JAN-2006
VMS732_DRIVER-V0200 25-APR-2005
VMS732_F11X-V0500 20-MAR-2006
VMS732_FIBRE_SCSI-V0900 9-MAY-2006
VMS732_GRAPHICS-V0400 15-AUG-2005
VMS732_HBMM-V0200 23-SEP-2004
VMS732_INSTAL-V0200 20-MAR-2006
VMS732_IOGEN-V0100 13-JUN-2005
VMS732_IPC-V0200 29-SEP-2004
VMS732_JOBCTL-V0200 13-SEP-2005
VMS732_LAN-V0400 24-AUG-2005
VMS732_LAT-V0100 21-DEC-2005
VMS732_LIBRTL-V0100 15-JUL-2004
VMS732_LINKER-V0100 22-FEB-2005
VMS732_LMF-V0200 24-FEB-2006
VMS732_LOADSS-V0200 24-MAY-2006
VMS732_MANAGE-V0500 21-NOV-2005
VMS732_MOUNT96-V0100 11-AUG-2004
VMS732_MP-V0100 17-JUL-2006
VMS732_MQ-V0400 3-FEB-2006
VMS732_MUP-V0100 27-OCT-2005
VMS732_OPCOM-V0100 16-JUN-2006
VMS732_PERFAPI-V0100 2-MAY-2006
VMS732_PRTSMB-V0100 6-APR-2005
VMS732_PTHREAD-V0500 25-APR-2006
VMS732_QMAN-V0100 6-JUL-2006
VMS732_RMS-V0200 4-JAN-2005 Yes
VMS732_RPC-V0400 6-APR-2005
VMS732_RTPAD-V0100 10-NOV-2005
VMS732_SHADOWING-V0500 1-SEP-2005
VMS732_SYS-V1000 22-MAR-2006 Yes
VMS732_SYSINI-V0100 25-JAN-2005
VMS732_SYSLOA-V0300 30-AUG-2005
VMS732_TDF-V0300 20-JAN-2006 Yes
VMS732_TRACE-V0200 1-NOV-2004
VMS732_TZ-V0200 26-JUN-2006
VMS732_UAF-V0100 20-OCT-2005
VMS732_VERIFY-V0100 25-APR-2005
VMS732_VMSMUP-V0100 23-FEB-2005
VMS732_XFC-V0200 20-DEC-2005


ECO KITS NOT INCLUDED IN UPDATE KIT: RELEASE DATE COMMENTS
------------------------------------ ------------ --------------------
VMS732_AUDSRV-V0300 2-NOV-2006
VMS732_SYS-V1100 1-NOV-2006
VMS732_RMS-V0300 4-OCT-2006
VMS732_SECSRV-V0100 28-SEP-2006
VMS732_XFC-V0300 27-SEP-2006
VMS732_TDF-V0400 22-AUG-2006
VMS732_BACKUP-V0600 2-AUG-2006
VMS732_ACRTL-V0300 17-APR-2006
VMS732_PCSI-V0300 15-MAR-2006


Regards,

Duncan

[Volker Halle]

Felix,

first consider to apply the current set of patches, if you've not already done so.

Then please check the error log on both systems at the time of the error. You need to use ANAL/ERR/ELV or even DECevent ($ DIAGNOSE) to be able to decode possible errlog entries.

Volker.

[FELIX WONG_1]

my patch update level is
VMS732_UPDATE-V0600--4

Following are the patches that applied
TCPIP_ECO-v0504-155-4
DNVOSIECO01-V0703-2-4 (Sep 28/04)
DWMOTIF_ECO2-V0103-1-4
VMS732_PCSI-V0100--4

Would you be able to tell which specific patch within the VMS732_update_xxx that
causing the problem.

[Jim_McKinney]

You should heed Volker's advice and rule out a hardware problem. Do you have a non-zero error count for the tape device? If so, you should examine the error log entries as "volume not software enable" is often indicative of failed hardware.

[Duncan Morris]

Felix,

it is unlikely that any patch which you have applied would be causing the problem you describe.

More likely, a later patch may address this issue, although I cannot sea any specific reference in the release notes.

Can you help by identifying the hardware involved here (SCSI controller, tape device)?

There is a note about the KZPCA in the release notes for VMS732_FIBRE_SCSI-V0800

5.2.6.1 Problem Description:

Customers running Ultra2 SCSI capable devices on a KZPCA
SCSI controller may see disk and tape operations (i.e.
COPY, BACKUP) fail with SS$_PARITY, SS$_CTRLERR, and
possibly other errors. These errors are accompanied by
DEVICE ATTENTION errorlog entries from the failing KZPCA
controller.

Your issue may not be directly related to TMSCP, but rather a timing issue which only surfaces because of the delays inherent in TMSCP operations.

Please make sure that you check the error logs as recommended by Volker.

Duncan

[FELIX WONG_1]

Hi Jim,

With the same hardware, I've had no problem when I ran the backup from the system with the local tape attached. Problem appears only when I ran the backup from another node
within the 2 nodes cluster environment.

I.e. tape drive is connected to node 2 as a local tape, when I ran backup from node 2, everything is fine.
Issue arose when I tried to run backup from node 1 (with tmscp enabled) using tape drive on node 2.

If there's an issue on hardware, the backup should have failed when I submitted from nod002.

[Volker Halle]

Felix,

you need 'something' to start diagnosis. It should be easy to rule out HW problems:

$ ANAL/ERR/ELV TRANSLATE/SUMM/SINCE=/BEFORE=

Run this on BOTH systems and report, whether any events have been found.

Then also check messages on the console terminal at the time backup is running, especially at the time your backup program has failed. This includes checking OPERATOR.LOG if OPA0: is not enabled as an OPCOM terminal.

Also please try to provide the full error message(s) text issued by your program. An 'error writing to tape' message is pretty generic. One would like to see the OpenVMS system services status or even the full IOSB.

Volker.