|
1. What is "kernel patch 2" about?
2. What do you mean by "I did NOT change the export entity in NO way"?
Please, try to do Show Exporting. If it works just submit the background
process.
It looks like 5 min delay between starting the background process and
invoking MCC is not enough. Could you, please, repeat the same experiment
with new database name?
Sam
|
| Sam,
I think I found the problem. During the VMS 5.4-2 to 5.4-3 upgrade the
system time was set back one year. I didn't reflect so much about it at the
time, but I did notice that the alarms wouldn't run with a time in the
past.
Then earlier this week when i did another VMS upgrade and agin the system
time was set back one year, the alarms stopped working. After fixing the
time it was possible to start the alarms again. But yet again the exporter
didn't work but this time I got a different error message very similar to
that one I got from the Alarm AM. The error message was some thing about
MIR not accessible.
This fact made me think that the MIR database in some way become corrupted
when the time was put back one year by the VMS upgrade procedure.
My action then was to copy all file documented on page 3-3 in Management
Module Programming V1.1 from another system with a similar installation.
This system was also upgraded to 5.4-3 but here was the system time controlled
by DTSS, so I didn't notice that the time maybe was set back one year.
After copying these files I cleaned up the old export files and then the
exporter started without any problems.
I have another question for you, to night the exporter background terminates
with error massage shown bellow, could that be an indication that I try to do
the export commands to soon after starting the background job ?
The failed startup that occured this night was the only error message I got
from the background process, I didn't get any from the batch job executing
the export commands. When I saw that the exporter had terminated I restarted
the background and then the exporter created the RDB files and started to
export data ! Any bright ideas ?
/Stefan
<------------------------ Backkground log file --------------------------->
$! SYLOGIN.COM
$ SET NOVERIFY
%MCC-F-NOATTREXIST, specified attribute record does not exist
%MCC-E-NOATTREXIST, specified attribute record does not exist
SSPMAN job terminated at 28-FEB-1992 00:09:51.01
Accounting information:
Buffered I/O count: 91 Peak working set size: 2871
Direct I/O count: 79 Peak page file size: 8346
Page faults: 1128 Mounted volumes: 0
Charged CPU time: 0 00:00:04.92 Elapsed time: 0 00:04:39.77
|
|
Stefan,
>I have another question for you, to night the exporter background terminates
>with error massage shown bellow, could that be an indication that I try to do
>the export commands to soon after starting the background job ?
>The failed startup that occured this night was the only error message I got
>from the background process, I didn't get any from the batch job executing
>the export commands. When I saw that the exporter had terminated I restarted
>the background and then the exporter created the RDB files and started to
>export data ! Any bright ideas ?
This is known behavior in Exporter V1.1. It happens sometimes only
when you create a new database file. After you restart the background process
everything should be OK.
Regards, Sam
|
| I have a customer with a similar problem, their DECmcc v1.1
exporter background process dies as soon as it starts up. With
the same error:
"%MCC-F-NOATTREXIST, specified attribute record doesn't exist"
One difference, it dies even when no export commands are issued.
It dies all by itself, or a least I think so. "Its extremelly
difficult to find if there is any forgotten export command still
active on the system."
Also, no sure its related but, once while issuing "delete export"
commands from a command file I got the following error:
"Request Queue is full, check background process status"
...
The exporter background process wasn't working even before this,
it was dying with:
"%MCC-E-ALERT_TERMREQ, thread termination requested"
It just took a little longer to do so, I must say that no-one
killed the process. And that no watch-dog program was running.
...
Be it as it may, the bottom line is that the exporter refuses to
work. Both the MCC account and the system have more then enough
resources, as taken fromn the docs and various notes here.
The system is a VAX 3800 with 32 MB of memory, running VMS 5.3-1
The normal export load is about 20 line, 20 circuit, and 20 SNMP
interface counters every hour, with two minutes intervals between
each 10 export commands. But i reduced this to nothing and the
exporter background process is still dying.
Any ideas ?
I even tought about re-installing, but I am not sure that would
fix anyrthing if its a MIR database corruption. We do want to keep
all the setup work that has been done so far...
Gil
|