Hello,
I've got three makes of IBM servers, all running 2003 server standard. Each of the three servers has embedded NICs- 2 are AMD based (Netfinity 4500R and eServer x342) and and the other Intel (eServer x345).
These servers have been running in our production environment for several months, and are now "randomly" exhibiting a behavior of falling off, in terms of network connectivity. The servers are all actually still running when it fails, but the NIC shows as being disconnected. Pings and Term sessions obviously stop at this time also. Link also ceases, but there are no other symptoms, i.e. everything else on the console indicates all is well, nothing screwy I noticed in Event viewer etc.
These servers are all connected to blades (2 different) in a Cisco 5500- and I've verified all as being hardcoded to 100/full on each end, no power settings are shutting the NIC down ,etc.
I've not applied any new patches that I know of to the servers, other than the latest IBM rec'd drivers for that NIC (I'm one of three Admins, but the guys said they didn't do anything either) and the other 2000 servers in our mixed domain appear fine. This issue sucks because I can't find a pattern. And we've been so busy taking systems down for maintenance every month there is no longevity to allow the units to fail after x days. They seem to work most of the time, then WHAM, out of the blue we get the support calls... A repair of the NIC status sometimes reinstates link, but a reboot always seems to help/ ;-) IBM just has us send in logs...then says it the motherboard - on three servers?!?!? FRUSTRATING.
I've got three makes of IBM servers, all running 2003 server standard. Each of the three servers has embedded NICs- 2 are AMD based (Netfinity 4500R and eServer x342) and and the other Intel (eServer x345).
These servers have been running in our production environment for several months, and are now "randomly" exhibiting a behavior of falling off, in terms of network connectivity. The servers are all actually still running when it fails, but the NIC shows as being disconnected. Pings and Term sessions obviously stop at this time also. Link also ceases, but there are no other symptoms, i.e. everything else on the console indicates all is well, nothing screwy I noticed in Event viewer etc.
These servers are all connected to blades (2 different) in a Cisco 5500- and I've verified all as being hardcoded to 100/full on each end, no power settings are shutting the NIC down ,etc.
I've not applied any new patches that I know of to the servers, other than the latest IBM rec'd drivers for that NIC (I'm one of three Admins, but the guys said they didn't do anything either) and the other 2000 servers in our mixed domain appear fine. This issue sucks because I can't find a pattern. And we've been so busy taking systems down for maintenance every month there is no longevity to allow the units to fail after x days. They seem to work most of the time, then WHAM, out of the blue we get the support calls... A repair of the NIC status sometimes reinstates link, but a reboot always seems to help/ ;-) IBM just has us send in logs...then says it the motherboard - on three servers?!?!? FRUSTRATING.
