6509 / eigrp : network problems / random servers lost connectivity to network?

[JMCNEL]

Over the past week or so we are experiencing problems with several servers that gets disconnected from the network - the logs on the servers show "While transmitting or receiving data, the server encountered a network error"

Please see the below summary of our problems - perhaps someone had the same problem before. Any ideas what might cause this would be greatly appreciated.

servers log that they are disconnected from the network at odd times – sometimes numerous times a day

Depending on the software – it either just logs an error or it crashes the application

Whatever it is – it may have compromised two hardware ports – I believe.

server 1.   Dell is tracking continuous disconnection errors – it appears to have take port 1 out – the card was chattering large packets into the network – changed to new port on 6509 – freezing stopped but errors are still incremented on the server side. interface is clean and does not show any errors.    This server is connected to another on the cat 6509.  Right now I am collecting performance logs for the server – the event id 2012 – is logged in spurts every few hours.  "While transmitting or receiving data, the server encountered a network error”.   Have update the firmware and nic drivers with no
improvement.

server 2. This serves : same problem – a reboot of the server fixes this problem I suppose because it reconnects the san.  The san is connected to 2/45 on the 6509 and the server to 0/34 on another switch (directly connected to the 6509)   The server logged that it lost its connection to the default at 3:46 pm Monday afternoon.  We did not have any power outages at that time.  There was no reboot.   There is no reason it should have lost it connection to the default gateway at that moment.   Ihave increased the logging interval on the san.

server 3.    The server application crashes every day.   The archivalware application log indicates it is being disconnected from the network and the event logs confirm the problem.  The server is not rebooting.    Last night we decided to change to the other network port – I will do this at lunch.  This is connected to 0/46 on the switch directly connected to the 6509

There are no errors on the switch ports – changing switchports - didn’t help.   The vendors are screaming “Network Problem

any help would be greatly appreciated

[Fabienne Stephanoff]

Hello,

You will probably need to work with a TAC engineer to resolve this one, using sniffer traces among other things to see what goes on with the port connected to the server when issue happens. You can open a TAC case from the community under the action menu.

Fabienne

[Florin Barhala]

Is there any firewall involved in the configuration? Are these problems related to other outages in the network: EIGRP, STP? Did you check cables?

[JMCNEL]

no firewall and no errors on interfaces on the 6509 and nothing in the logs or show tech.

[Brent Rockburn]

Are all your servers bare steel or are you supplying VM as well? Have issues with the SAN been looked at? "disconnected from the network" is a generic server/application error so it might be a problem with the storage aspect of things as well. At least it's something to look into while you're getting TAC on the line

[JMCNEL]

Are all your servers bare steel or are you supplying VM as well?

Bare steel

IF we were having problems with the sans and not the servers I would look at the sans - but the servers complain about losing the sans AFTER they have lost connection to the network.   The sans do not log any errors but the servers log loss of san.  

Also not all affected servers have a san.