Clients attached to old 10mbs Cabletron hub access Unix servers on same hub ok. When client is moved to a Cisco 5000 switch with 100mbs FD connection, which connects to another Cabletron switch before coming back to the old Cabletron hub, the amount of time required to save drawing files from the 100 mbs pc to the 10 mbs Unix boxes goes up drastically (takes minutes instead of seconds). We moved the Unix boxes to the same 5000 switch (100mbs FD), and the users on the same 5000 switch now enjoy much improved performance. However, the users who are still on the old 10mbs Cabletron now have drastically increased save times to the Unix boxes that are attached to the Cisco 5000. I understand how this can increase save times (because of traffic contention), but not to the huge extent that is happening. We are not running VLANS (only 1), and have a flat IP structure (/20). The 5000 is set mostly for defaults, except portfast is enabled on all ports connecting to computers. Not using trunking or FEC. Not seeing errors on any ports. I have considered changing PAGP (etherchannel) to off. Any other config (or other) ideas?
Are you forcing 100MB F/D??? We have a slew of 2900XL's that we are using about 10 VLAN's on. When we had the FA's set to force F/D as well as 100MB I ended up in the same boat you're in (with a ton of latency for file transfers). Set all the ports to auto-detect:
int fa0/X (replace X with port #)
Do this for all your interfaces and see if that affects performance any... It did in my case!
We are pleased to announce availability of Beta software for 16.6.3. 16.6.3 will be the second rebuild on the 16.6 release train targeted towards Catalyst 9500/9400/9300/3850/3650 switching platforms. We are looking for early feedback from custome...