Conference orarep::nomahs::dbms

2612.0. "6.0-11 monitor problem?" by M5::LWILCOX (Chocolate in January!!) Thu Mar 20 1997 12:46

6.0-11.  I have a customer reporting that starting just about a week ago they
get into a situation where their DBMS monitor "dies", or at least they think
it dies.  That is, they have never actually checked to see if the VMS process
goes away.  Their symptom is that they see in the monitor log file "remote 
monitor nn is not responding".  This is a single node system.  I have explained
that it is perfectly normal to see this message.

After they see this, however, then they get hundreds of rollbacks, on all
databases.  They usually reboot the system at this point.

In one instance of this they received numerous dbr bugcheck dumps with 
exception DBR$RECOVER and FLK, referencing the RUJ file as still being
locked.  I know that was a problem in the past, but all release notes, etc.
appear to indicate that was fixed.

They believe this situation will occur again today so I have asked him to
check to see if the monitor process is really still there or not.  They
have not seen the NOMONITOR message, just the "not responding" message.

This is on an alpha system.  They have had this version of DBMS installed
for quite a long time and it's been running smoothly until a week ago.
Naturally "nothing has changed" :-).

Does any of this seem familiar?

Thanks,

Liz

Turns out he was getting a SYSTEM-F-NOSLOT so he'll increase max process count.