[Search for users]
[Overall Top Noters]
[List of all Conferences]
[Download this site]
| Title: | DECmcc user notes file. Does not replace IPMT. |
| Notice: | Use IPMT for problems. Newsletter location in note 6187 |
| Moderator: | TAEC::BEROUD |
|
| Created: | Mon Aug 21 1989 |
| Last Modified: | Wed Jun 04 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 6497 |
| Total number of notes: | 27359 |
3152.0. "Notification stuck in BEING CREATED, BEING ENABLED state" by FOUR62::LICAUSE (Al Licause (338-5661)) Tue Jun 09 1992 20:35
VMS V5.5, T1.2.7 of DECmcc.
I recently reinstalled T1.2.7 along with the TSAM on a VAX4000/500....everything
ran w/o problems....I had several notification events defined and working from
both DECnet PHIV and Collectors.
I've also installed the AIRCO demo and have same running w/o problems.
Also had this combination running w/o problems on a VAX VLC 4000.
Now when I bring up DECmcc all comes up, but noticed that I don't get any
notification from my Phase IV DECnet events. Remote sinks are all in place,
the DNA4 startup procedure had been edited as suggested and ran w/o problems.
The DNA4 process is running.
When I list out the existing notifications, I see all of the events that I have
created, but the Phase IV events are now in BEING CREATED, BEING ENABLED state.
The are basically: DOMAIN name NODE4 * LINE/CIRCUIT(two different definitions) *
COUNTERS ZEROED.
This all worked w/o problems before.
I also note now that when MCC starts up in Iconic map format, the fileview window
opens and displays the following errors:
%Thread 11 terminating with exception
%CMA-F-EXCCOP, exception raised, VMS condition code follows
-SYSTEM-F-ACCVIO, Access violation, reason mask=04, virtual address=00000010,
PC=00D4DEC4, PSL=03C000005
I tried disabling and deleting the hung rules....when I did the state changed
to BEING CREATED, BEING DISABLED......then BEING DELETED.
I've also tried shuting down MCC, deleting the NOTIFICATION file and when
it comes up, recreate the notifications, but it goes into the same state.
I need to demo this over the next few days and would really like to show that
this feature works.
Any help resolving this would be greatly appreciated.
thanks,
Al
| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 3152.1 | QAR 3137 | TOOK::MINTZ | Erik Mintz, dtn 226-5033 | Wed Jun 10 1992 03:25 | 1 |
| Entered as QAR 3137
|
| 3152.2 | | FOUR62::LICAUSE | Al Licause (338-5661) | Wed Jun 10 1992 12:06 | 6 |
| Is there any way around it other than reinstalling DECmcc from scratch?
Is there a process that can be restarted or some sort of temporary workaround?
thanks,
Al
|
| 3152.3 | notification lockouts in pre SSB kit | TOOK::CALLANDER | MCC = My Constant Companion | Wed Jun 10 1992 18:36 | 35 |
| Hi Al.
Okay, here is the scoop... you will need a new set of CMA files
and a new notification FM; and this does mean waiting for the
next kit, sorry.
The problems you are experiencing are actually multiple problems.
Right now you probably have a lock stuck in the locked state
because of the accvio, do the following to check things out:
stop all mcc processes you know of
sho dev/file mcc_common >>> any open mcc files?
start mcc and get the notify commands outstanding again
analys/sys
sho summary
set proc /index=# (for mcc process)
sho proc /locks
is the mcc_event_pool lock outstanding, or being waited on
by the mcc process? see who has the lock and kill that
process.
Now the problems were as follows:
cma was not correctly handling deleting of threads and if you
did a terminate notify (delete a notify request; which
happens automatically when you change domains in the map)
which was screwing things in the system up.
Notification FM had a race condition in the code where if events
were coming in when a notify was terminated, it left a lock in
a locked state.
Hope this helps.
|
| 3152.4 | | FOUR62::LICAUSE | Al Licause (338-5661) | Fri Jun 12 1992 13:24 | 16 |
| Work around....
Thanks to Jill Callander, I was able to relieve this situation. Even a reboot
did not clear the situation, since the particular notification event was in
my events file and was loaded by default on startup.
Apparently the logical MCC_DNA4_AM_LOG had to be deassigned then reassigned
after startup of the DNA4 event logger. I also had some parameters in both
the UAF file and SYSGEN that needed to be increased.
However, there still appears to be a problem. Any notification event
defined in catagory NODE4 * will still cause a process lock. I did not
see this behavior on a VAXstation 3100, VLC4000 or VAXstation 4000/60, but
am seeing the problem on a VAX4000/500.
Al
|
| 3152.5 | QAR 3137 - answer | TOOK::MINTZ | Erik Mintz, dtn 226-5033 | Tue Jun 16 1992 20:08 | 5 |
| The answer for QAR 03137
at this point the hangs have been attributed to system problems
and cma problems which are being tracked independently.
|