[jcifs] jCIFS NTLM in clustered WebSphere
Michael B Allen
mba2000 at ioplex.com
Wed Jul 28 22:15:30 GMT 2004
Eric Glass said:
> This is likely caused by the fact that the NTLM challenge is specific
> to the SMB connection for which it was generated. Basically, if the
> client hits Server A in the cluster first, and gets a challenge, it
> needs to respond to Server A; if the response is routed to Server B,
> that server will create its own SMB connection to the domain
> controller, and the response will not be valid for the new connection.
> This results in an authentication failure (and the prompt to log in).
> Can you use sticky sessions (i.e., a client is always sent to the same
> server in the cluster)? That might fix the issue.
I don't recall if 0.9.2 had the 'req.getSession();' line before
negotiation the first token. I believe I added that somewhere around that
release. Upgrading might do the trick if they already have "sticky
More information about the jcifs