D
Deleted member 174149
I am working with a vendor on some firmware updates they did to some POE switches we have custom-made.
I am noticing some odd behavior that I'm not really able to explain. Our topology for the installation of these switches is a loop: each switch connects to the one above it, and then a line is dropped from the top switch to the bottom. This provides cheap redundancy. RSTP is left alone to elect its own root bridge, since it is a big loop there is little to gain in manually configuring the root.
The issue I am having is, when I install up to about 24 or 25 switches, RSTP converges just fine. No loops, data flows normally across all of the switches.
However, I see odd behavior with anything more than 21 switches. If I run a constant ping to all switches and break a link, I see several go offline for 2-6 seconds while RSTP reconverges. However, one switch will go offline and never respond to ping again while the link is broken, despite the rest of the network coming back online. If I plug directly into that switch, I can ping it, but I can't ping it from anywhere else on the network.
If I then reconnect the broken link, the switch will come online happily.
I am at a loss to describe this. If RSTP were not converging, I should see problems while the loop is connected, not after it is broken. If I serial into the lost switch, I see that it appears normal - it has the correct root bridge (same as the others), it has one root port and one designated port...
Any thoughts?
I am noticing some odd behavior that I'm not really able to explain. Our topology for the installation of these switches is a loop: each switch connects to the one above it, and then a line is dropped from the top switch to the bottom. This provides cheap redundancy. RSTP is left alone to elect its own root bridge, since it is a big loop there is little to gain in manually configuring the root.
The issue I am having is, when I install up to about 24 or 25 switches, RSTP converges just fine. No loops, data flows normally across all of the switches.
However, I see odd behavior with anything more than 21 switches. If I run a constant ping to all switches and break a link, I see several go offline for 2-6 seconds while RSTP reconverges. However, one switch will go offline and never respond to ping again while the link is broken, despite the rest of the network coming back online. If I plug directly into that switch, I can ping it, but I can't ping it from anywhere else on the network.
If I then reconnect the broken link, the switch will come online happily.
I am at a loss to describe this. If RSTP were not converging, I should see problems while the loop is connected, not after it is broken. If I serial into the lost switch, I see that it appears normal - it has the correct root bridge (same as the others), it has one root port and one designated port...
Any thoughts?