Why do HP server fans go crazy when a network loop is created?

[talksr]

I have been deliberately creating some loops on the network to check that STP and loop-protect are working on our HP Procurve switches.

 

Interestingly, if I turn off STP and loop-protect and cause a nasty loop on the network, within seconds of the loop taking hold, both of our new HP servers begin to make a very loud noise from their fans - identical to when they first boot up. The fans will continue to run at full spin until the loop is disabled.

 

What would cause this? It is actually quite a useful indicator as if I am sat in the server room and the fans spin up, I know something is going on.

[glen_j]

The servers are having to process the looping packets making them work harder possibly?

[DGardiner]

I have been deliberately creating some loops on the network to check that STP and loop-protect are working on our HP Procurve switches.

 

Interestingly, if I turn off STP and loop-protect and cause a nasty loop on the network, within seconds of the loop taking hold, both of our new HP servers begin to make a very loud noise from their fans - identical to when they first boot up. The fans will continue to run at full spin until the loop is disabled.

 

What would cause this? It is actually quite a useful indicator as if I am sat in the server room and the fans spin up, I know something is going on.

 

 

I've had this too at one of our schools, the server failed a few months later after what I'm assuming was the ILO controller started to malfunction... not certain though but the device driver failed and would work - after attempting a reboot it never turned on again....

[talksr]

That was what I was thinking? But didn't expect them to make such a noise over a few loop packets. Could understand it if you were transferring 1tb worth of data or something.

[DGardiner]

That was what I was thinking? But didn't expect them to make such a noise over a few loop packets. Could understand it if you were transferring 1tb worth of data or something.

 

 

No idea how it happened, i just got a phonecall from the school saying they could hear the servers from the other end of the block, and had been like that for 2-3 days... probably related to the failure

[ITGURU]

never had this issue with our servers, all are HP. When a loopback is detected the switch is intelligent enough to disable the 2 looped ports and send a syslog email to me to notify switch location giving a rough idea where the loop is to go and sort before it attempts to re-make the connection. When it does this, if the loop is still present it will disable it again for that duration and keep trying to release the ports until it has been sorted.