[jcifs] Broken Pipes

Allen, Michael B (RSCH) Michael_B_Allen at ml.com
Thu Jan 3 11:16:09 EST 2002

Yes, connections are "cached". The broken pipe is obviously the pipe/lanman pipe we issue
RAP calls on. I suppose jcifs could catch that and reissue the request. I'll try to reproduce
the scenario and make it a little more robust in this respect. There's probably somewhere
down lower I can trap these.


> -----Original Message-----
> From:	Rob Wygand [SMTP:rob at filefish.com]
> Sent:	Wednesday, January 02, 2002 7:00 PM
> To:	jcifs at samba.org
> Subject:	[jcifs] Broken Pipes
> Mike,
> I'm running Samba 2.0.10 as my file server. (I've not tested against 
> Windows proper yet). But when I do the following, jCIFS keeps throwing 
> an exception complaining of a broken pipe... here are the steps:
> (This all needs to occur in one VM)
> create an SmbFile to list shares in a server (smb://servername/)
> call list()
> shut the server down
> create a new SmbFile using the same path
> call list() (The right thing happens, here)
> start the server
> create a new SmbFile using the same path
> call list()
> On the last list() I keep seeing the broken pipe error... my guess is 
> that something is being cached as being down by jCIFS even though the 
> server has returned to normal operations. Seperate threads in the same 
> VM have this problem... it goes away by stopping and starting the VM though.

More information about the jcifs mailing list