We took out all of those after isolating a network-wide problem to a single Netgear 4 port switch flooding with bad packets... took bloody ages to find the thing!
Originally Posted by dave.81
This is an old thread but I recognised the problem from another recent experience myself so thought it would be worth a mention.
Your problem was caused by the redundant links.
Originally Posted by link470
These must be configured for LAG (Trunked) connections using LACP or for fail over with STP assigned to these two links specifically.
I suspect your switch had already detected a problem with this link and had disabled the unwanted one with STP or broadcast controls.
Something happened to reset this state and both links became live again.
Had someone reset or power cycled the switch? this could have forced the disabled port to come back on line again. Had somebody changed the STP settings on the switches?
Common mistake is for somebody to make a specific change to a switch like enable STP plug in the dual uplinks. The switch takes off for a few minutes then STP kicks in and blocks off one of the ports.
Everything works fine and everyone goes home.
Oh pooh! We forgot to save the changes into NV Ram!
The power recycles and the switch reverts back to a previous state!
Now we have two uplinks and no configuration to stop them from going loopy!
If your really unlucky it could be months before this could happen and you would have long forgotten that a switch setting had been changed anyway.
In my case this happened after a UPS change over.
A switch that had obviously had been changed was running fine until the power was reset.
It reverted to a config over a year old on reboot!
If you make a switch setting change, backup the config before and after. Always commit the changes when leaving the CLI.
interestingly we were told by rm not to use spanning tree at one of their technical seminars (we do use it) never had an issue with it.
It should be used only where it is needed.
"The whole idea behind spanning tree is that you can have a link fail, because you have a pair of bridges connected via two paths. Spanning tree will leave a port blocked until it is needed. Then we should be able to unplug redundant links and connect new ones without interruptions? Sorry, but no, it doesn't work that way."
Read the Full Networking 101 Article on STP it's a must read for switch admins.
As most schools are built from the core outwards in a star topology there is no real need to use STP unless multiple links are intentionally deployed for redundancy purposes.
STP can be enabled per switch or per port where needed.
In most cases I find that STP has been enabled by people that either do not understand it's purpose or has been done so in an attempt to fix another issue that the in house team cannot identify.
One of the most common side effects of badly configured STP is the loss or extremely slow propogation of DHCP packets. So if your having difficulty in getting DHCP in parts of your network theres a good chance that STP is enabled and/or is incorrectly configured.
We only use it to stop little pests from causing loops in the network and taking it down. Where we have multiple link we have channel bonded/agragated/trunked them, so do not need STP for that.