The orginal question can be found below.
I finally did find a related bug report that gave
a suggested patch.
Bugs:
4034454 multihomed NIS+ client specifies wrong interface
in callback
4034454 adding a third interface at Solaris 2.5.1, niscat hangs
the rpc.nisd in getmsg
Patch:
103612-24 SunOS 5.5.1: libc, libnsl, libucb, nis_cachemgr and
rpc.nisd patch
( the machine in question had rev 23 of the patch)
Original question:
Sun-Managers,
I am experiencing a problem with my NIS+ setup. I
currently have a root master with two root replica's.
root master sparc 10 2.5
root replica sparc 20 2.5
root replica Ultra2/300 2.5.1
All machines have the latest patches including 103187-25
and 103612-23 depending on the OS rev, and all are on
different subnets. The command 'nisshowcache' shows
correct NIS+ info on all clients and servers. The NIS_PATH
is also correct.
When I run 'niscat networks' on a number of clients, the
command hangs and comes back with the message:
can't list table: NIS+ servers unreachable
All NIS+ services then freeze for 5 minutes, and the following
messages appear on all the NIS+ servers:
nisd[123]: nis_make_rpchandle: RPC: Miscellaneous tli error - An
event requires attentionInvalid argument
nisd[123]: Unable to create callback
Once NIS+ is 'unfrozen', normal NIS lookups from the clients work.
I thought this may be the result of not having a replica on the same
subnet as the client, but my attempts to create a replica on that
subnet failed. I got the same messages as above when the master
attempted to update the new replica.
A search of Sunsolve Online gave lots of bugs with similar messages,
but no solutions. Has anyone seen this message and freezing problem
before.
Thanks.
==================================================
Noel Fardy
System Administrator
Trout Trading Management Co.
==================================================
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:01 CDT