Bulkwalk of OSPF neighbour table is broken on ProCurve
When check_ospf_nbr does an SNMP bulkwalk of OSPF-MIB::ospfNbrTable, it takes a very long time (many seconds), and the ospfNbrState column gives us some error, so it thinks the state is "None", even though it is full.
There is a workaround in place, doing normal SNMPv1 walk over the table if --vendor=procurve is specified on the command line, but we should figure out what the real problem is.
The snmpbulkwalk(1) command does not suffer from these problems, so it might be the re-implementation of that in trh_nagioslib.py that is broken in some subtle way.