swagentd don't fork correctly

[Matthias Haupt]

I have configured a HP720 with a standalone
kernel on the local disk. The contents of
/etc and /sbin I have copied from the
/export/private_roots/... of the server,
because the 720 was (is) a diskless client of
this server. After the /sbin/bcheckrc
a script configures the network devices and
nfs services, so that the 720 can nfs-mount
the directories /usr /var from /export/private_roots
from the server. Up to this moment I have a
/usr/bin and /usr/lib subset on the local disk
to start the network services, /usr is then moved to
/usr_boot to free the /usr mount point.

The boot process works properly, but the start of the
swagentd hangs because the swagent does'nt fork
correctly. Precisely: the command swagent forks
but the parent process will not finish,
the daemon is not given to the init process
and therefore the startup-script does'nt finish.

The same happens, if the startup-script is skipped
and swagentd is started manually.
No error message occurs and I can start some instances
of the swagentd, which should not be allowed.

Is anything wrong with my boot configuration
or with the swagentd configuration ?

PS:
The reason for this is, that our new server (C3000)
has only a 1000 Mbit/s network interface, and therefore
the rbootd for the old 720 fails during his startup.
This should be so for all network devices faster than
10 Mbit/s. With my approach I try to boot from local
disk, but still having a nearly diskless client
environment.

[Robert Gamble_3]

Idea: what if you try using 'nohup' or an '&' for the start-up of swagentd in
its rc script ?

Please post how this gets resolved. Thanks and Good Luck!