abruciamonti_at_bci.it said:
| [alex_at_core2 alex]$ clog alex username: alex Password: WRONG PASSWORD
| Invalid login (RPC2_NOTAUTHENTICATED (F)). [alex_at_core2 alex]$ clog
| alex username: alex Password: CORRECT PASSWORD

| [17:48:54]SocketListener: "/usr/src/redhat/BUILD/coda-5.0.1/coda-src/
| rpc2/sl.c", line 436:    DecodePacket(RPC2_BUSY): bad bind seqno
| [17:48:55]SocketListener: "/usr/src/redhat/BUILD/coda-5.0.1/coda-src/
| rpc2/sl.c", line 436:    DecodePacket(RPC2_BUSY): bad bind seqno 

These errors are from the rpc2 layer. The auth2 daemon only has a single 
worker thread, and giving a wrong password gets this single thread stuck
in the first half of the binding sequence. So the second (valid) binding
is blocked in the `incoming queue', all retries will then get an RPC2_BUSY 
message back.

It automatically times out within 60 seconds, so restarting the auth2 
process is not really necessary. Changing the connection setup is on 
our list for some time now, but we haven't gotten around to cleaning it 
up yet.

Jan