Our Web Engineers use rdist/rsh from a RHEL3.8 server to push out their monitoring scripts to various Linux web servers. Their rsh/rdist process using a distfile always seems to hang after completion when the target server is a SLES9 server - the session always seem to complete/end successfully if the rdist target server is a RHEL server.

An strace of the RSH PID on both sides shows the process on the RHEL 3.8
source server as read(8, while on the target SLES server the PID shows as read(0,

A netstat -a|grep CLOSE_WAIT does not show any tcp session relating
to the rsh sessions.

Would appreciate any thoughts on this. Thank you.

What I see is that a password prompt is being initiated even though a valid .rhosts file exists and an entry in
/var/log/authlog would seem to indicate that the .rhosts is working:
Jan 11 12:22:41 xaqc1dsx pam_rhosts_auth[26511]: allowed to a320398@xwel1q.aetna.com as a320398

The rsh session just 'hangs' at the password promtp seemingly forever because days later I will have to logon and kill all of these rsh sessions to the SLES9 servers.

Seems to be related to Active Directory and/or VAS since rsh using a locally defined user account on the target server (with an .rhosts) works successfulyl with no prompts.

Resolved by adding 'auth sufficient pam_rhosts_auth.so' as the first non-comment line in /etc/pam.d/rsh.