rlogin & rcp problem

[BAUKnight]

dears,
we have to HP-UX server running B.11.23, hostnames are dbclust1 , dbclust2 respectively.
when i do rlogin or rcp from dbclust2 to dbclust1, the command pass with no probelms, but when i try to do rlogin from dbclust1 to dbclust2 system is asking for password.
taking in mind that root user has the same password on both nodes.
also i ensure that the follwoing files are the same on both nodes, and with the same rwx modes

/etc/hosts
/etc/hosts.equiv
~root/.rhosts

i tried removing the old files and create them again from scratch.
i attached the files from both nodes on a .rar format.
please advice what might be the problem, and how i can resolve it.

[Laurent Menase]

usual causes are:
rights of .rhosts which must be 644
client system have more than one ip address, and the address used to go out is not the one expected.
$so do a nettl trace , you may get some hints.
also look at syslog.log

[Jeeshan]

use IP address instead of hostname.and try not to use host.equiv file.

change accordingly and check the result.

[Johnson Punniyalingam]

Hi,

>>I Agree with ashan, << You can give try with ip address,

Example:-

root Server1
root 192.168.1.1
root Server2
root 192.168.1.1


Thx,
Johnson

[BAUKnight]

i checked the access rights for the files, they are correct, and tried to use the IP to do rlogin, but still face the same issue,, and don't know what't up.

my machines are part of oracle RAC -not yet inistalled-, and the installation won't pass with this error exist, my machines are using 3 IPs,
one for the corporate LAN
one for the heart beat
one for the oracle private IP

that's why i tries to use the IP, but still same issue.

any ideas.

[Asif Sharif]

Hi BAUKnight,

If /etc/hosts has entires of your servers then no need to check with IP's (if they are resolving it properly).

rlogin from dbclust1 to dbclust2 and provide password. execute who -R to see what your login/host is resolving as. Then do the reverse from dbclust2 to dbclust1 .

You need to put plus"+" sign in /.rhosts or give it as,

dbclust1 root
dbclust2 root

and check again.


Regards,
Asif Sharif

[BAUKnight]

dear,
first of all let me thank everyone who suggested any help.

i solved the problem by re-creating the /etc/hosts on each machine, i used SAM to configure the netowrk card on each node, then copy the contents from each node and adding them to the other node.

i tested the new files using both rlogin & rcp commands

attached the new /etc/hosts from one ofthe nodes