[cvsnt] Sudden degrade in checkout performance

[Ian Long]

On my Linux test machine (Red Hat) the time taken to checkout a module has 
recently changed from around 1minute to ~12 hours (it has never run to 
completion yet).

On the Linux machine, I am running cvs client 1.11.17, the CVS server is a 
Windows SmallBusinessServer running CVSNT2.5.03 build 2383.

The network between the two machines is a single gigabit switch, pings 
between then show no loss and return in ~0.5mSec - they are on the same 
subnet as each other. Other Windows and Linux machines on the same subnet 
are still able to check out the module in around a minute.

The linux machine is running nothing else at the moment, top show it sitting 
idle. There is masses of disk space available. Coming from the Windows 
world, I have even restarted the Linux box just in case.

The output from -ttt looks like this, the first few entries take a few 
seconds, but large files take 10s of minutes.
[ian at test checkout]$ cvs -ttt checkout mymodule
 -> main loop with CVSROOT=:pserver:user at myserver:/CVS/Repository
 -> safe_location( where=(null) )
 -> Connecting to myserver(192.168.30.1):2401
 -> Sending file `asm.jar' to server
 -> Sending file `mysql-connector-java-3.1.12-bin.jar' to server
? noael/WEB-INF/lib/.new.struts2-c
 -> unlink_file(CVS/Entries.Static)
cvs server: Updating noael
 -> unlink_file(CVS/Tag)
 -> unlink_file(CVS/Entries.Static)
cvs server: Updating noael/ERs
 -> unlink_file(CVS/Tag)
 -> unlink_file(CVS/Entries.Static)
cvs server: Updating noael/WEB-INF
 -> unlink_file(CVS/Tag)
 -> unlink_file(CVS/Entries.Static)
cvs server: Updating noael/WEB-INF/lib
 -> unlink_file(CVS/Tag)
 -> Register(asm.jar, 1.1, Tue Jun 19 06:16:54 2007, -kb,  )
 -> rename(CVS/Entries.Backup,CVS/Entries)
 -> unlink_file(CVS/Entries.Log)
 -> Register(mysql-connector-java-3.1.12-bin.jar, 1.3, Tue Jun 19 07:00:34 
2007, -kb,  )
 -> rename(CVS/Entries.Backup,CVS/Entries)
 -> unlink_file(CVS/Entries.Log)

I have tried checking out the module to a new dirrectory, but encounter the 
same performance.

Please can someone give me some suggestions for the cause or how I may go 
about looking for the cause.

Many thanks,
Ian.