Windows 2003 Server drops from network

Hello, I have a Dell Poweredge server that has been
running fine for about 2 months (since we bought it and
brought it online). In the last 2 weeks, the server has
started disappearing from the network. Mapped drives are
gone, Remote Desktop connections break, and most telling,
the machine fails to answer pings. We have switched it's
network connection to a jack that is on a different
switch, one which contains another server not reporting
these problems. The server with the problem still had the
problem, the server without the problem did not develop
the problem, so I don't believe the problem is with the
physical network.

These outages last from a few seconds to a couple minutes,
and so far have only been reported during the day when
useage would be higher. I have not been able to get to
the machine during an outage to try and run tests out from
it.

The machine has a Broadcom NetXTreme Gigabit NIC, we have
updated the driver for it. I see nothing wrong in the
ipconfig all nor the netstat. But again I haven't been
able to run these during an outage.

During the most recent outages we have started receiving
an error 5783:
The session setup to the Windows NT or Windows 2000 Domain
Controller \\iu-mssg-adsdc03.ads.iu.edu for the domain ADS
is not responsive. The current RPC call from Netlogon on
\\BL-SOC-OBERON to \\iu-mssg-adsdc03.ads.iu.edu has been
cancelled.

This error has not consistantly appeared and seems to me
to more indicate the server isn't talking on the network
than it is the problem. There really is nothing else in
the event log of interest. We have scanned the machine
with an updated Norton, both from the machine and from a
Norton AV server. It comes up clean (Norton has been
uninstalled temporarily as some articles pointed out
potential problems with it, but an episode happened after
the uninstall and reboot).

I was able to get to the server right after the last
outage, and could see the perfmon results. Nothing looked
strange, the network bandwidth didn't move which to me
indicates the server thought it was connected fine. It
still received and set packets, at least neither dropped
to 0 but they were reduced. Memory and processor looked
normal.

If there's more information anyone needs, please let me
know. This problem started occuring shortly after most of
the client machines moved to a new location. At first we
thought the problem was between the new location and old,
but now the problem is appearing on machines in the old
location as well. It happens about twice per day usually,
though it can be more often. The server has a static IP
and I see no conflicts with that.

Thanks in advance for any ideas.

Hi,

Id start by looking at some Network Monitor traces between the server and a
client machine.

--
Todd J Heron, MCSE
Windows 2003/2000/NT

"Rick" <(E-Mail Removed)> wrote in message
news:0d8e01c4a0cd$3ab0ed70$(E-Mail Removed)...
> Hello, I have a Dell Poweredge server that has been
> running fine for about 2 months (since we bought it and
> brought it online). In the last 2 weeks, the server has
> started disappearing from the network. Mapped drives are
> gone, Remote Desktop connections break, and most telling,
> the machine fails to answer pings. We have switched it's
> network connection to a jack that is on a different
> switch, one which contains another server not reporting
> these problems. The server with the problem still had the
> problem, the server without the problem did not develop
> the problem, so I don't believe the problem is with the
> physical network.
>
> These outages last from a few seconds to a couple minutes,
> and so far have only been reported during the day when
> useage would be higher. I have not been able to get to
> the machine during an outage to try and run tests out from
> it.
>
> The machine has a Broadcom NetXTreme Gigabit NIC, we have
> updated the driver for it. I see nothing wrong in the
> ipconfig all nor the netstat. But again I haven't been
> able to run these during an outage.
>
> During the most recent outages we have started receiving
> an error 5783:
> The session setup to the Windows NT or Windows 2000 Domain
> Controller \\iu-mssg-adsdc03.ads.iu.edu for the domain ADS
> is not responsive. The current RPC call from Netlogon on
> \\BL-SOC-OBERON to \\iu-mssg-adsdc03.ads.iu.edu has been
> cancelled.
>
> This error has not consistantly appeared and seems to me
> to more indicate the server isn't talking on the network
> than it is the problem. There really is nothing else in
> the event log of interest. We have scanned the machine
> with an updated Norton, both from the machine and from a
> Norton AV server. It comes up clean (Norton has been
> uninstalled temporarily as some articles pointed out
> potential problems with it, but an episode happened after
> the uninstall and reboot).
>
> I was able to get to the server right after the last
> outage, and could see the perfmon results. Nothing looked
> strange, the network bandwidth didn't move which to me
> indicates the server thought it was connected fine. It
> still received and set packets, at least neither dropped
> to 0 but they were reduced. Memory and processor looked
> normal.
>
> If there's more information anyone needs, please let me
> know. This problem started occuring shortly after most of
> the client machines moved to a new location. At first we
> thought the problem was between the new location and old,
> but now the problem is appearing on machines in the old
> location as well. It happens about twice per day usually,
> though it can be more often. The server has a static IP
> and I see no conflicts with that.
>
> Thanks in advance for any ideas.

I have had the same issue (same card, Dell 2650 under 2003
Enterprise). I only noticed the problem once I enabled teaming and/or
load balancing after loading the broadcom management tools. I also
noticed that if I ping -t some server at startup the machine doesn't
drop connections or fail to respond to pings so its workable just
ghetto. Are you using the Broadcom Management Software too?

-I

"Rick" <(E-Mail Removed)> wrote in message news:<0d8e01c4a0cd$3ab0ed70$(E-Mail Removed)>...
> Hello, I have a Dell Poweredge server that has been
> running fine for about 2 months (since we bought it and
> brought it online). In the last 2 weeks, the server has
> started disappearing from the network. Mapped drives are
> gone, Remote Desktop connections break, and most telling,
> the machine fails to answer pings. We have switched it's
> network connection to a jack that is on a different
> switch, one which contains another server not reporting
> these problems. The server with the problem still had the
> problem, the server without the problem did not develop
> the problem, so I don't believe the problem is with the
> physical network.
>
> These outages last from a few seconds to a couple minutes,
> and so far have only been reported during the day when
> useage would be higher. I have not been able to get to
> the machine during an outage to try and run tests out from
> it.
>
> The machine has a Broadcom NetXTreme Gigabit NIC, we have
> updated the driver for it. I see nothing wrong in the
> ipconfig all nor the netstat. But again I haven't been
> able to run these during an outage.
>
> During the most recent outages we have started receiving
> an error 5783:
> The session setup to the Windows NT or Windows 2000 Domain
> Controller \\iu-mssg-adsdc03.ads.iu.edu for the domain ADS
> is not responsive. The current RPC call from Netlogon on
> \\BL-SOC-OBERON to \\iu-mssg-adsdc03.ads.iu.edu has been
> cancelled.
>
> This error has not consistantly appeared and seems to me
> to more indicate the server isn't talking on the network
> than it is the problem. There really is nothing else in
> the event log of interest. We have scanned the machine
> with an updated Norton, both from the machine and from a
> Norton AV server. It comes up clean (Norton has been
> uninstalled temporarily as some articles pointed out
> potential problems with it, but an episode happened after
> the uninstall and reboot).
>
> I was able to get to the server right after the last
> outage, and could see the perfmon results. Nothing looked
> strange, the network bandwidth didn't move which to me
> indicates the server thought it was connected fine. It
> still received and set packets, at least neither dropped
> to 0 but they were reduced. Memory and processor looked
> normal.
>
> If there's more information anyone needs, please let me
> know. This problem started occuring shortly after most of
> the client machines moved to a new location. At first we
> thought the problem was between the new location and old,
> but now the problem is appearing on machines in the old
> location as well. It happens about twice per day usually,
> though it can be more often. The server has a static IP
> and I see no conflicts with that.
>
> Thanks in advance for any ideas.