Hello,
we've found a strange behavior on old Catalyst WS-C2950SX-48-SI with 12.1(22)EA6, Multiple STP running (MSTP), keepalive enabled and "errdisable recovery cause loopback" enabled. Uplink switch is 2960-X.
A non-managed switch was attached to "dynamic-auto (=access)" port of Catalyst switch (MSTP Type Edge P2p), forcing Catalyst to block port a few times with "err-disabled", then attempting to recover from loopback err-disable and again putting to err-disabled state.
I can see in LOG going to Err-disabled for 2 times (at 9:50:09, 10:13:39) and attempting to recover 2 times (at 9:54:50, 10:18:29).
After recovering for 2nd time the port stayed up (at least nothing else in the LOG). But - the loop was still there, and network gone mad.
09:36:43: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/36, changed state to up
09:50:09: %ETHCNTR-3-LOOP_BACK_DETECTED: Keepalive packet loop-back detected on FastEthernet0/36.
09:50:09: %PM-4-ERR_DISABLE: loopback error detected on Fa0/36, putting Fa0/36 in err-disable state
09:50:11: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/36, changed state to down
09:50:12: %LINK-3-UPDOWN: Interface FastEthernet0/36, changed state to down
09:54:50: %PM-4-ERR_RECOVER: Attempting to recover from loopback err-disable state on Fa0/36
09:54:54: %LINK-3-UPDOWN: Interface FastEthernet0/36, changed state to up
09:54:56: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/36, changed state to up
10:13:39: %ETHCNTR-3-LOOP_BACK_DETECTED: Keepalive packet loop-back detected on FastEthernet0/36.
10:13:39: %PM-4-ERR_DISABLE: loopback error detected on Fa0/36, putting Fa0/36 in err-disable state
10:13:40: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/36, changed state to down
10:13:42: %LINK-3-UPDOWN: Interface FastEthernet0/36, changed state to down
10:18:29: %PM-4-ERR_RECOVER: Attempting to recover from loopback err-disable state on Fa0/36
10:18:34: %LINK-3-UPDOWN: Interface FastEthernet0/36, changed state to up
10:18:36: %LINEPROTO-5-UPDOWN: Line protocol on Interface FastEthernet0/36, changed state to up
Would it be the reason not putting port down again? Too many traffic already taking place? Or some Cisco firmware bug / feature?
Gregor
