Hi, I found the "bug": the problem was the "wait" option. in inetd.c se_wait is definde short but it take the pid (int) values. So (sep->se_wait == pid) never matchs and the server never restarts.
Here is the very small patch to inetd.c (short => int) Regards Enrico Enrico Bernardini wrote: > > Hi all, > I'm porting amanda2.4.2p2 client on cygnus (win98se cygwin1.dll 1.3.6). > Inetd starts amandad first time, amandad completes the job and exits. > When the server side recall amandad to continue working, inetd does not > restart amandad. > > You can see the same behaviour if you start and close a talk session: > > 1) talk foo@bar > inetd -d says > > someone wants ntalk > + Closing from 7 > 234265 execl /usr/sbin/in.talkd > > 2) talk from foo@bar and exit > 3) wait for timeout: > inetd -d says > > 234265 reaped, status 0 > > 4) try to "talk foo@bar" > and you will see nothing from inetd (I'll wait for another execl...) > > The tcp connections work correctly. > > inetd.conf: > ntalk dgram udp wait root /usr/sbin/in.talkd in.talkd > amanda dgram udp wait enrico /usr/local/libexec/amandad > amandad > > Thanks in advance > Enrico
inetd.c-patch
Description: Binary data
-- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
