Hi gurus,
A user of mine has a system running in a machine with OSF1/4.0D. His
system comprises a group of RPC servers and clients. He tells me that
the portmap daemon is losing the registration for a particular server
every other day. We have a cron running rpcinfo every five minutes so
that we can notice the phenomenon and retart the system, but so far we
couldn't understand the problem.
The server is running and its socket port is available (in the LISTEN
state) even after the portmap loses the registration. My user assures
me that the server isn't talking to the portmap asking for the
deregistration.
What I'd like to know is if there is a way to debug or monitor the
execution of the portmap daemon. The manpage doesn't mention it but
running 'strings /usr/sbin/portmap' I can see the following lines
which hint that there must be a way to put portmap under debug mode:
Usage: portmap [-v] [-d] [-t seconds]
portmap debugging enabled -- will abort on errors!
The other thing that concerns me is if there's a way to perform a
"warm-start" in the daemon. Under Solaris I can kill the rpcbind
daemon (portmap's alter ego) with a SIGINT or a SIGTERM so that it
dumps its table of registrations in a file. Then I can restart it
with the -w option to perform a warm-start, i.e., so that it reads the
previously dumped information. Without I lose all the information
about the currently running RPC servers.
Is there a way to do that under OSF1?
Thank you. I'll summarize.
---
Gustavo Leite de Mendonça Chaves
fone:55.19.705.7003
mail:gustavo_at_cpqd.com.br
http://www.cpqd.com.br/~gustavo
Received on Thu Oct 21 1999 - 11:26:20 NZDT