Very Computer

hosts=local4,bind4

and the AFTER line:

hosts=local,bind4

The OS was AIX 4.3.2, we patched from may 1999 PTF levels to August 1999
levels.

The most noticible failures was with rshd, the daemon would fail to
authenticate rsh/rcp commands in a kerberos 5 environment.  The failures
would occur even if we identified all hosts/interface in /etc/hosts.

To get the system working, the ONLY change we had to make was to change
the netsvc.conf file to reflect standard local lookup, the bind (ipv4)
remote DNS lookups.

?????

You can guess how long it took me, two other engineers, and IBM level 1
Poughkeepsie support to find this niggling little problem.

thanks in advance

Hi,

I recovered the same problem, but when in my case i got a core dump
file and an entry in the errpt which says that the rshd has core
dumped.
So waht does your errpt tell you ?

So finally I changed my netsvc.conf from
hosts=bind4,nis4.local
to
hosts=bind,nis,local
and rsh works fine again.

Hartwig

Quote:>Can anyone offer a good reason why this line in netsvc.conf would work
>BEFORE a recent PTF installation:

>hosts=local4,bind4

>and the AFTER line:

>hosts=local,bind4

>The OS was AIX 4.3.2, we patched from may 1999 PTF levels to August 1999
>levels.

>The most noticible failures was with rshd, the daemon would fail to
>authenticate rsh/rcp commands in a kerberos 5 environment.  The failures
>would occur even if we identified all hosts/interface in /etc/hosts.

>To get the system working, the ONLY change we had to make was to change
>the netsvc.conf file to reflect standard local lookup, the bind (ipv4)
>remote DNS lookups.

>?????

>You can guess how long it took me, two other engineers, and IBM level 1
>Poughkeepsie support to find this niggling little problem.

>thanks in advance

IBM has this as a sev1 in Austin and Poughkeepsie.  This is seither a
bug in bind on the latest set of PTFs or IBM has changed something
regarding name resolution/authentication without DOCUMENTING the change.