Tips for monitoring performance on 4108GL

[Chris McFarling]

Users have been complaining lately about the server being "slow". This vague description is all I have to go on so I want to thoroughly test and monitor all aspects of the data path. I have a grasp on monitorig server performance but I'm not sure of the best method to monitor my switch (4108GL) to see if it's a bottleneck.

I'm looking for suggestions as to what commands to use, utilities to employ, etc. to see if my switch is holding things up. Is there a way to monitor, in real time, overall throughput on this switch so I can see how it's performing in real world situations?

For the record this switch has 6 modules installed:
4 - 100/1000 modules, all but a couple ports are in use regularly, all in-use ports run at 1Gb

2 - 10/100 modules, ~90% of the ports are in use at any given time, all but a couple run at 100Mb

Thanks for any feedback

[Jason F.]

To begin with, you need to make sure the switch is running clean; that is, without port errors and flapping. If errors and/or flapping is present, then performance will not be optimal.

From the CLI, enter the command "menu", then 4 (event log) and "E" for end of log for most recent events.

If you see multiple concurrent entries for "Port X offline", "Port X online, blocked by LACP" then, you can be assured that the clients on those ports are complaining of performance (slowness) issues. If the log shows flapping, starting with the uplink and server ports FIRST (if applicable), lock down BOTH PC and switch ports at 100-Full Duplex and monitor conditions.

Consider replacing cabling/trying other punch-downs for CRC & alignment errors.

Best of Luck-

JASON
ProCurve Networking