WS26 with Bad Management

[adairw]

We swapped out a couple WS12's for a single WS26 on friday and it's started to exhibit some strange behaviors. Out of the hundreds of wisp switches we have I'm pretty sure this is the first time I've ever seen anything like this. It seems to pass traffic just fine but pinging the switch is unstable and mgmt access drops in and out due to the lack of connectivity.
The site this switch is on has two WS26's and due to the way it's hooked up I can mange the switch in a couple of different ways by coming in different router ports and from the other switch. I've been able to rule out bad ports, cables, IP conflict etc because of this. The only issue is managing the switch itself. All transit traffic seems fine. Even rebooted and upgraded (scarily) to the latest RC without luck
Will plan on replacing the switch unless someone has any other ideas.

CPU and memory usage look ok.

This is what the switch looks like when you ping it's IP.

[sirhc]

Please post up the complete Device/Configuration TAB

[adairw]

Here is the config. You can't see the "options" section but nothing is enabled there.

While typing my initial post the thought occurred to me, "maybe there is an ip conflict?". I changed the switch IP from .250 to .252 and even at layer 2 could not ping anything on .250. Changed the switch back to .250 and now pings to the switch are normal. No packet loss or high latency. Plus the interface is responding normally. WTH?
It's been clean for about 2 hours now. Will keep an eye on it.

[lligetfa]

PCWT (Problem Cleared While Testing) is every tech's nightmare. Since the root cause could not be determined, the fear is it will return.

[adairw]

Exactly. Changes had been made to the config, turning stuff off and on to see if it was something specific. Reboots and firmware upgrades. Why in the would would changing the IP matter?
A possible thought, odd arp entry in the site router since the switch did change from one IP to another after we were doing swapping things around before this morning...

[lligetfa]

Bad ARP tables is a plausible explanation. I chased just such an issue years ago where one switch was misreporting a MAC in its ARP table. Two switches fought back and forth, both claiming to own the route to the MAC.

[adairw]

I think that's what I am going to go with. A WS12 had the .250 IP, when we installed the WS26 in place of the two WS12's we changed the IP to one the old WS12's IP's..
very odd

[lligetfa]

A WS12 had the .250 IP, when we installed the WS26 in place of the two WS12's we changed the IP to one the old WS12's IP's..
very odd

Whenever I swap hardware, I always change the IP of the outgoing kit JIC it were to get connected to the network and cause a conflict. Been bitten that way before when co-workers would make a hardware swap and not change the IP on the outgoing kit.

Even DHCP reservations can bite you if/when they wake up before their lease half-life so I would always edit the reservation and force a DHCP renew on the outgoing kit.