Cisco 2950G, latest firmware (12.1(14)EA1) seeing our ports which are hardcoded to 100/full switching in the middle of the night to AUTO/AUTO. The HP servers connected to the switch are forced 100/full so this causes major problems.
Searching the Cisco release notes, I don't see a fix for this or anything mentioned. The only hardcoding is done to 100/full. No
10/full.Anyone else experience this?
Sorry, but are you saving the configs? Maybe you don't have to on that platform? I have this idea that some IOS Catalysts don't have a seperate running and startup config, however I hope that I am wrong. sh ver shows the uptime.
One workaround would be to see if having everything on Auto would work correctly.
If you have religious objections to this, another would be to set everything to 100/Half. Then if one end went to Auto you would still be OK.
If you are worried about the potential loss of throughput you could always do a test. In many situations I suspect that the loss of performance would be minimal or approximately zero in terms of 'user' performance if you changed to HD.
I now *absolutely* favour Auto/Auto until proved unsatisfactory, and it has not been so proved for years.
Show VER shows the switch has been up for months.
AUTO/AUTO is how most of our ports are configured. This switch handles only Proliant servers. The Compaq/HP NIC Drivers supplied do not negotiate correctly with Cisco switches. This is documented in various places.
The result is either one end goes to 100/half and the other 100/full when set AUTO/AUTO on both ends, or the NIC constantly re-negotiates. If we don't use the supplied HP Agents (Drivers) and use the Microsoft ones, this problem goes away. However there are features we need that these drivers supply.
This rack is full of these 1U/2U servers. And the NICs on some of them average 50% utilization, per MRTG. One misconfigured piece brings all servers in this rack to a crawl.
If AUTO/AUTO did work on these drivers, I'd use it. I use it everywhere else.
We've swapped in a different 2950G to see if the problem is with this unit. It's occured 3 times so far on this one switch. I am unaware of our other 2950G's having this issue. That is why I ask if others have seen this. I cannot find any Cisco document describing this problem on a 2950G.
Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.