In the process of debugging some seemingly random connectivity errors, I noticed the fast ethernet ports on the 2950 seemd to simply stop forwarding frames now and then. Replacing the switch removed the problem from the customer network, but I'm still curious if I can figure out what might be wrong with the switch. The network consists of only one switch.
Pulling it into my lab and tinkering a bit, I noticed weird patterns in the RTT when leaving it running overnight (second column is the time since last RTT > 2ms):
00:51:23 02:27:55 +2:36 02:49:02 +0:22 03:10:09 +0:19 03:31:15 +0:21 04:40:54 +1:09 06:02:04 +1:38 06:56:23 +0:54 08:11:47 +1:16 08:54:01 +0:43 09:12:20 +0:18 09:37:38 +0:25 09:48:18 +0:11 11:24:51 +1:36 13:01:22 +1:37 13:34:34 +0:33 13:55:42 +0:21 14:16:48 +0:21 15:11:06 +0:55 15:32:12 +0:21 16:26:31 +0:56 16:47:37 +0:21 18:03:02 +1:16As you can se, it's not very random. There's a high frequency of ~20m intervals, and it's also a noticable pattern at ~55m and
1h37m.I saw no debug information that seemd relevant, and posting it here would be spamming a lot...
Suggestions? The OS is c2950-i6q4l2-mz.121-22.EA5a.bin. I found spanning-tree running, but all ports are configured with the portfast feature (so I switched it off).