Comware

In Insight Manager 7 I keep getting Information events saying Link Up and Link Down, all from the same stack (ProCurve 4000M). I'm not receiving these from any of the other switches outside of this stack. I've viewed the event log of all the switches in the stack and all of them show the Link Up/Down on different ports. I was told that this message is normal when I device turns on/off, but I'm getting a constant flow of this trap every 2 to 10 minutes. Any ideas?

Hello Klaus -

What sort of devices are attached to the 4000M ports that are going up and down? Are those devices being powered on/off or reset?

Regards,
Ralph

Klaus??

Mostly workstations, servers, and printers. But nothing is being power cycled as often as I'm seeing the messages.

Might be printers that have "powersave" feature, when they enter the "powersave on" mode. HP printers do that, but it can be disabled by setting up the individual printers.

You can turn this kind of logging off on the 4000M switches, by entering "Switch Management Access Configuration - Trap Receivers", or typing 23 after logging on, and choose "Events Sent in Trap" = "Critical". Also see the Help screen associated to this menu option.

Ok, we decided to just change the trap message as suggested. When I went there, it was set to Not INFO, which I thought meant to send everything but the infromation stuff. I changed it to Critical and I'm still getting the linkup/down messages. I don't have to reboot the stack do I?

I also saw the community name on the commander is public and on the switches in which it controls the communtity name is public@sw1, public@sw2, etc. Is this correct?

You get one link up/down message each a device thinks link state went down and another each time the device thinks link state came back up. This will generate a lot of messages very quickly, and are normal when cables are physically moved at switch or end devices.

I have had this error message caused by a number of different situations:

1. Autonegotiation issues. I set the switch ports to operate at a specific rate and it went away. Believe this was actually a NIC driver issue on the workstations. Recommend trying this on one or two "problem" ports and see if it solves the issue for those ports.

2. Electromagnetic Interference (EMI). Once, cabling went directly behind a microwave oven and another case was laid across a subwoofer. Errors only when devices were in use.

3. Cabling problems. Had a site trying to run Fast Ethernet on long runs over CAT 4 cabling. Works....very poorly.