[ic] On restart of IC in rpc mode, 2 IC processes are not stopped

Ron Phipps interchange-users@icdevgroup.org
Mon Jan 13 18:12:00 2003


I'm running IC on Redhat 7.3 all errata up to date, Perl 5.6.1, mysql,
dual Xeon P4, 2GB ram.  The IC services is running in RPC mode, I have
tested with both MaxServers set to 0 and no MaxServers setting within
the RPC block in interchange.cfg.  Below are the relevant portions of my
interchange.cfg:

Variable  TRAFFIC  rpc

ifdef TRAFFIC =~ /rpc/i
Message RPC traffic settings.
PreFork             Yes
StartServers        5  
MaxRequestsPerChild 100
HouseKeeping        2
PIDcheck            120
MaxServers          0
Endif

When the service starts ps shows:

19452 ?        S      0:00 interchange
19462 ?        S      0:00 interchange
19464 ?        S      0:00 interchange
19466 ?        S      0:00 interchange
19468 ?        S      0:00 interchange
19470 ?        S      0:00 interchange
19472 ?        S      0:00 interchange

I then issue a restart with this command:

interchange -r -u

After the restart I see this in ps:

19464 ?        S      0:00 interchange
19466 ?        S      0:00 interchange
19470 ?        S      0:00 interchange
19472 ?        S      0:00 interchange
19728 ?        S      0:00 interchange
19736 ?        S      0:00 interchange
19738 ?        S      0:00 interchange
19740 ?        S      0:00 interchange
19742 ?        S      0:00 interchange
19744 ?        S      0:00 interchange
19746 ?        S      0:00 interchange

After another restart I see this in ps:

19464 ?        S      0:00 interchange
19466 ?        S      0:00 interchange
19470 ?        S      0:00 interchange
19472 ?        S      0:00 interchange
19738 ?        S      0:00 interchange
19740 ?        S      0:00 interchange
20043 ?        S      0:00 interchange
20051 ?        S      0:00 interchange
20053 ?        S      0:00 interchange
20055 ?        S      0:00 interchange
20057 ?        S      0:00 interchange
20059 ?        S      0:00 interchange
20061 ?        S      0:00 interchange

If you count the number of processes it increases by two each time.
This continues in this fashion unless I do a killall and then restart
IC.  You can probably figure out on a development server that IC would
be restarted a lot and this leaves a lot of processes which are not
doing anything.

Any ideas?

Thanks,
-Ron