[Prev][Next][Index][Thread]
NIS problems under OSF/1 v2.0
I just upgraded some of my DEC 3000/400's from OSF/1 v1.3 to
v2.0. The upgrade was pretty straightforward, but I am noting
a problem with the NIS/YP system.
In the past, I've distributed the passwd and host databases
from a master server to a group of five slave servers. I have
things set up in the "standard" way, recommended by DEC, with
ypbind -S host1,host2,host3,host4 (where I specify host1 as
the local host and the others as some other members of the yp domain).
The problem is that occasionally with requests like "ypcat passwd",
something goes flaky and nothing is received. Typing ypwhich shows
no response. After about a minute, the machine binds to another host
in the domain and stays there. Why does it lose binding with the
local host to begin with? This never happened in previous versions
of OSF/1 or ULTRIX.
I found some error messages in my /usr/adm/syslog.dated/xxx/user.log file
that appear to be related to the problem:
Apr 7 13:14:55 eos1 syslog: yp_all - RPC clnt_call (TCP) failure: RPC: Unable t
o receive; errno = Connection reset by peer
Apr 7 13:15:01 eos1 syslog: yp_all - TCP channel create failure: RPC: Remote sy
stem error - Connection refused
Apr 7 13:20:46 eos1 syslog: yp_all - RPC clnt_call (TCP) failure: RPC: Unable t
o receive; errno = Connection reset by peer
Apr 7 13:23:10 eos1 syslog: yp_all - RPC clnt_call (TCP) failure: RPC: Unable t
o receive; errno = Connection reset by peer
Apr 7 13:40:10 eos1 syslog: yp_all - RPC clnt_call (TCP) failure: RPC: Unable t
o receive; errno = Connection reset by peer
Apr 7 13:50:04 eos1 syslog: yp_all - RPC clnt_call (TCP) failure: RPC: Unable t
o receive; errno = Connection reset by peer
Apr 7 13:50:05 eos1 syslog: yp_all - TCP channel create failure: RPC: Remote sy
stem error - Connection refused
Has anyone noticed this problem? Any solutions?
Richard Eckman
NASA Langley
eckman@xxxxxxxxxxxxxxxxxx