[jcifs] Re: Problems Upgrading From 0.7.15 to 1.1.9

[David Pattison]

Michael B Allen <mba2000 <at> ioplex.com> writes:

> 
> On Mon, 14 Mar 2005 12:24:41 +0000 (UTC)
> David Pattison <david.pattison <at> siemens.com> wrote:
> 
> > Hi Guys,
> > 
> > We have just upgraded the jcifs library in our web app and are having
> > problems with logging on. Basically we needed to take advantage of the
> > jcifs.netbios.wins parameter which allows multiple addresses in the newer
> > versions (which isnt in 0.7.15?), and to specify this and
> > jcifs.smb.client.domain instead of just pointing to
> > jcifs.http.domainController. So, with these parameters changed (and moved
> > from the command line to init-params in web.xml), I have been unable to
> > successfully logon.
> > 
> > I was originally getting a timeout error waiting for the DC to respond, so
> > I upped the timeout from 10 seconds to 60, and after about 20 seconds the
> > following error appeared:
> > 
> > jcifs.smb.SmbException: 0xC000018C
> 
> 0xC000018c is NT_STATUS_TRUSTED_DOMAIN_FAILURE "The trust relationship
> between the primary domain and the trusted domain failed.
> 
> > But on the command line the following appears when the timeoout is set to
> > 60 seconds (137.223.284.46 being the primary WINS server):
> 
> What timeout? What properties are you using?
> 
> > 137.223.284.46
> > java.net.UnknownHostException: 137.223.284.46: 137.223.284.46
> >         at java.net.InetAddress.getAllByName0(InetAddress.java:999)
> >         at java.net.InetAddress.getAllByName0(InetAddress.java:969)
> > 
> > But with the default timeout, this is the command line error:
> > 
> > Socket Closed: DOMAINNAME<1C>/137.223.234.128
> > java.net.SocketException: Socket Closed
> >         at java.net.SocketInputStream.socketRead0(Native Method)
> 
> This is just a low level socket error. Is that DC ok?
> 
> > I tried replacing jcifs.http.domainController with the domain and WINS
> > parameters back when we were using 0.7.15 and it worked fine. Has
> > something else changed that I should be configuring in the new versions?
> 
> A lot has changed since 0.7.
> 
> In general you do not provide enough information to decipher what the
> problem is. My instinct tells me your not a typical windows domain
> corporate intranet type environment.
> 
> Mike
> 



Mike,

As far as I know our intranet isn't anything special (what would make it
especially unique?). Its virtually all Windows, and I only know of Unix
machines, I havent even seen them in the building. The DC is also operating fine.

I'll try and give you some more details on the project and network:

Its a J2EE web app for IE, which naturally uses the NTLMAuthenticationFilter.
The only 2 jcifs properties I have explicitly set are jcifs.smb.client.domain
and jcifs.netbios.wins. The timeout increase I was referring to is
jcifs.smb.client.responseTimeout to 60000 which was your recommendation on a
post I found in the archives. As this only brought up another error I commented
it out as it didnt seem to be the source of the problem.

The DC is running Win2k Server and the web container we use is OC4J on
workstations and Oracle 9ias on the test server. The workstations are running
XP. However, the workstations and DC are on different domains and subnets, which
may explain the NT_STATUS_TRUSTED_DOMAIN_FAILURE error. Why would the domains be
trusted in 0.7.15 and not 1.1.9?

I do have a couple of Win2k test machines which sit next to my own workstation,
but are on the same domain as the DC, so I'm going to deploy the web-app on them
and see if anything is different. I'll be sure to let you kow how it goes.

Regards,
David