by Anybody noticed that rsh occasionally hangs without doing anything?
System: 4 CPU SS1000; Solaris 2.3, most patches, including 101318-45.
truss(1) shows it sitting in getmsg(2);
myhost# truss -p 10505
*** SUID: ruid/euid/suid = 143 / 0 / 0 ***
getmsg(4, 0xDFFFF3D8, 0xDFFFF3CC, 0xDFFFF3E4) (sleeping...)
lsof(1) shows it connected to the peer host;
myhost% lsof -p 10505
...
rsh 10505 iand 3u inet 0xf73c6330 0x4 TCP myhost:998->peerhost:shell
rsh 10505 iand 4u inet 0xf71e4e30 0x0 TCP *:992
netstat(1) for the ports that lsof(1) mentions shows this:
myhost% netstat -an |grep 998
*.998 *.* 0 0 9112 0 BOUND
myhost% netstat -an |grep 992
*.992 *.* 0 0 8576 0 LISTEN
Note that there is nothing actually in ESTABLISHED state and the
there appears to be no connection to the peer as far as netstat is
concerned. Strange that lsof thinks otherwise...
This problem occurs probably about 1 in 20 rsh's that I do. With other
non-Solaris 2 hosts to the same peerhost, there is no problem.
Ian D