I have a Procurve 2626 on a small network (~200 users); about 100 of them are one port, which every now and then sets itself to "off" in the running configuration.

There is no indication of why this happens in the logs.

From searching the net and HP support forums, it seems others have this problem too, but no solution seems to have been found.

As it is extremely difficult to rewire the network, using multiple links for the users is not possible. (The oldest parts of the buildings are from the ~1625, so just thinking about rewiring is virtually a no-go...)

  1. How many switches do you have attached on this network segment, just this one? If so, read about the spanning tree protocol.

  2. Next, I think you might have exhausted some assumption made by the Procurve developers - 100 MAC addresses will now all be mapped to the 1 switch port - which is fine, unless the vendors had some other limit.

    • One thing that typically occurs with low-end/mid-range switches when its MAC-Address-Table is exhausted (flooded) is that it simply turns into a hub, but anything can happen really.
    • The specification sheet for the HP Procurve 2626 states that its MAC-Address-Table size is 8K - can you see how many entries you have before the time of the crash?
    • I somehow doubt you've reached the global maximum, but maybe some kind of local (per-port) limit? That's just an "I have no idea what I'm talking about" type of a guess though :)
  3. Can you try another switch instead of that one and see if does the exact same thing so you can rule out any problem specifically with that port?

  4. Can you turn up debugging? (That switch supports syslog according to the docs)

  5. Having at the datasheet now.

  6. And here's the troubleshooting guide.

    • See the section STP Blocks a Link in a VLAN Even Though There Are No Redundant Links in that VLAN. It might be relevant.
| improve this answer | |

I'm not familiar with ProCurves, but with that many users on one port, is it possible that the switchport is being disabled due to too many discards? What speed is the port the users are hanging off?

| improve this answer | |

Are you running STP on that port? If not, you could have a switching loop which is causing the port to down itself.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.