We have noticed a problem regarding Cisco LS1010 Switches and are now asking
for advice to track it. Configuration: Netview 6.0.1 on two servers (AIX
4.3.2) and LS1010 Rel. 12.0 (several patch levels).
Most of the LS1010 have several interface cards showing up addresses of NMS
stations!
e.g. a switch with IP Address 10.1.1.1 (Interface Index 8) has also
interface cards for 10.3.1.119 and 10.12.12.12 (both are PC's that have been
connected to that device via telnet or have done "snmp get" operations). All
these interfaces refer to the same interface index.
If we delete one of the wrong interfaces and demand poll the node, it adds
the interface immediately. Deleting the node and rediscovering it leads to
the same result. In some cases we have noticed that the "wrong" addresses
show up in the MIB2-"IpAddrTable". Anyway, if we do an snmpwalk from the top
10.3.1.119 shows up only two times: in IpNetToMedia table and "at Table". I
guess netview does not use the "at" table anymore? There are a lot of other
entries in these tables, thus we don't know why only these two addresses are
found to be worth being an interface...
To make things even more complicated: discovering LS1010 devices with an
"old" netview 5.1.2 server shows...everything ok.
"call support" would of course be a good advice but right now I'm not sure
wether to call CISCO or Tivoli?
tia for any help
regards,
|