RSH login/connection slows down when one of DNS go down
Does anyone hit the situation when rsh/rlogin connection slows down when one of the DNS servers described in /etc/resolv.conf go down.
Is there any way to configure rshd daemon in CentOS 5.x linux systems to ignore the non existing DNS server?
The time measurements are the following:
1. both DNS are up
rsh from host A to host B takes approx 120-170ms
2. first DNS is down
rsh from host A to host B takes approx 2 min
3. second DNS is down
rsh from host A to host B takes approx 40s
I performed the tcpdump for DNS requests on host B and it shows that during the rsh login process check the in.rshd sends the same amount of DNS requests to the both DNS servers and not using only one of them and switching to second one if first one fails in 5 seconds like the host or dig.