Peer-to-peer network - intermittent failure.

[Charles Prooth]

I have a peer-to-peer network. The server computer is
brand new, running W2K Professional with all the latest
patches and updates from Microsoft. It has an Asus A7N8X-
VM motherboard, Athlon XP2500+, 512K RAM and WD120GB hard
drives mirrored using a Highpoint-tech RocketRaid 133
controller (RAID 1). Rarely does the CPU show more than
10% and the network throughput looks really good, 8MB/sec
or more, at times, in short bursts.

There are 4 client PCs, 3 of which run XP Home and one
running XP Pro. The server runs Winfax Pro 10.02 and
provides a fax service to each of the client PCs, also
running Winfax Pro. That is the only application running
on the server other than PC Anywhere which is used for
remote support.

The clients use the server for almost all disk storage.
The applications are Paperport 9, Lotus Organizer
(installed to replace Daytimer 2000 to try to cure the
problem), WordPerfect, internet explorer and a few other
apps.

Once or twice a day there is a failure on the network. The
symptoms are as follows.

1) Suddenly, all of the client PCs are no longer able
to access the network drives on the W2K Pro computer but
the server can still access all of the clients' shared
drives. The clients can also access each other.

2) Winfax Pro still works in both directions.

3) The only apparent way to make the server drives
available again is to reboot the server. The clients don't
need to be rebooted.

4) There are no entries in the event logs that would
indicate that anything has gone wrong.

All of the clients are up to date with patches etc.

All of the devices use static IPs and have got Netbios
over TCP/IP enabled in the WINS tab.

Can anyone help me to identify the problem?

I plan to uninstall and reinstall TCP/IP and file sharing
on the W2K Pro computer.

I also plan to try a different computer in the place of
the new server just to see what happens.

Depending of the results, I am considering reinstalling
W2K.

 

[Dave]

does the server have a static IP? are you mapping the drives by name or IP?

 

[Charles Prooth]

The server has a static IP. I map by drive name.

Some more information has come to light:

Another similar but separate network has a similar
problem. In this case I am see error messages on both the
clients and server. The finger points to Paperport.

When one of the clients opens a networked folder in
Paperport the network drive becomes unavailable. On the
server several pages of open files appear all named
\pipe\spoolss and the other clients are unable to connect
to the network folders. They receive the following
message: Not enough server storage is available to process
this command. On the server in the log there is a message:
The server was unable to allocate from the non paged pool.

When the "offending" client exits from Paperport all of
the \pipe\spoolss disappear and the problem is not
immediately repeatable. The network drives/folders are
available again, no reboot required.

It looks as if Paperport is doing something "flaky" and
grabbing all of the memory for shared files/folders.

I am trying to get support from Scansoft but that is going
rather slowly.

~CP

-----Original Message-----
does the server have a static IP? are you mapping the drives by name or IP?

hard >> drives mirrored using a Highpoint-tech RocketRaid
133