[jcifs] Possible jCIFS<->JDK1.4.2 problem

[Allen, Michael B (RSCH)]

I am not aware of any bad interaction of Java 1.4.2 with jCIFS. As far as I know jCIFS is
extreemly stable right now.

How long does it take you to get this corruption you're seeing? It would be very interesting if
you could get a packet capture of this happening.

Keep in mind that it is perfectly legitimate to open and write to the same file using multiple
threads. Unless the actions of these threads were closely coordinated this would certainly
lead to file corruption. You might try opening the file with a share access parameter of
FILE_SHARE_NONE. This will cause additional threads attempting to open a file for the
second time to get the "being accessed by another process" Exception.

If you can get a packet capture of the transition from no corruption to getting corruption I
might be able to determine if it's a problem with jCIFS or it's usage.

Also perhaps you can send me one of these corrupted files if it's not too big (e.g. 500K)?

Mike

> -----Original Message-----
> From:	barry.roberts at xactsites.com [SMTP:barry.roberts at xactsites.com]
> Sent:	Thursday, August 14, 2003 3:21 PM
> To:	jcifs at lists.samba.org
> Subject:	[jcifs] Possible jCIFS<->JDK1.4.2 problem
> 
> I'll admit I'm fishing here, but I'm having a problem with data corruption int a Tomcat app.  The app writes xml files and zip files to Samba server using jCIFS 0.7.11.  With JDK1.4.2 (Solaris 8) I
> can pretty consistently get corrupted files using a JMeter script with 6-10 threads.
> 
> Once it starts EVERY file written is corrupt until I re-start Tomcat.  Using JDK1.4.0, I can't re-produce the problem.
> 
> Are there any known problems with jCIFS and 1.4.2? 
>