VMS Hang

[Kulothungan]

Hello, We have VMS 7.2 running on AlphaPC 164LX. It would run normally for several days; but for no apparent reason it will hang all of a sudden. It is clustered with a standby box of same kind and version and configured to fail over to it during failure. So we have to reboot the machine and fail the system over back to the duty box from the standby box manually.

The question may be vague; but that is the stage that I am in at the moment. Has anybody had this kind of situation or does anybody know a way to analyse to get to the bottom of the problem?

I tried the account utility, but could not find anything useful. I tried Analyse/error_log, but it comes up with the following message:

%ERF-F-CEHFND, New header format found. Install DECevent and run conversion utility

I cannot install anything on both the machines in the cluster as they are live.

Thanks in advance

[David B Sneddon]

How much physical memory is on the machine?
How much page/swap file space does it have?
It may be that it runs happily until it runs
out of physical memory but has no page/swap file
space.
I have seen this on systems running different versions.

Dave

[Wim Van den Wyngaert]

Did you check the file OPC$LOGFILE_NAME (operator.log) of both systems ? Previous version of the file after a reboot.

Were you able to do a "show sys/node=" for the hanging node on the remaining node ?

Next time, crash the hanging system (control-p and type command crash). At reboot an anal of the crash dump is made that can be found in sys$common:[syserr].

Wim

[Arch_Muthiah]

Hi Kulothungan,

OpenVMS Alpha is not supported on the AlphaPC 164LX and 164SX, though there are folks that have gotten certain of the LX series to load SRM and bootstrap OpenVMS.

Do you have IDE or SCSI cont?

This is the known problem reported earlier, that if you have IDE, the bootstrap fail occur, you need to have SCSI to try this.

Anyway, please go thru this thread discussion....

http://lyris.sunbelt-software.com/read/messages?id=231994

Archunan

[John Travell]

Before you do as Wim says, do check that you have a crash dump file. This can be in any one of 3 locations, sys$sysroot:[sysexe]sysdump.dmp, sys$sysroot:[sysexe]pagefile.sys, or on another disk if you have dump_off_system_disk (DOSD) enabled.
Sysdump.dmp is the default, and ideally SYSGEN parameter DUMPSTYLE should be 9. If you have neither sysdump.dmp nor DOSD enabled, the dump will be written to the pagefile. If so it will promptly be overwritten on reboot UNLESS you have SYSGEN parameter SAVEDUMP set to 1.
All this can be checked before the hang next occurs, and setup if needed.
Once you have a valid dumpfile you wait for the next time it hangs, you hit control_P and at the console prompt, >>> crash
After the reboot there should be a file sys$errorlog:clue$'nodename'_'date'_'time'.lis
This is a plain text file, post it here as 'clue.txt' and the experts among us will (try to) help you identify why the hang is occuring.
JT:

[Daniel Fernandez Illan]

Hi Kulothungan
You can also use
ANAL/ERROR/EVL
to check hardware failures on Alpha box.
Saludos.
Daniel.