rssLink RSS for all categories
 
icon_red
icon_green
icon_red
icon_red
icon_blue
icon_green
icon_green
icon_red
icon_red
icon_red
icon_orange
icon_green
icon_green
icon_green
icon_green
icon_blue
icon_green
icon_orange
icon_red
icon_green
icon_red
icon_red
icon_green
icon_red
icon_red
icon_red
icon_red
icon_orange
icon_green
 

FS#2893 — FS#6854 — 46.105.104

Attached to Project— Network
Incident
Roubaix 4
CLOSED
100%
We have perturbations on the network 46.105.104.0/24

This provokes sporadic connection losses.

We are investigating.
Date:  Friday, 20 July 2012, 16:33PM
Reason for closing:  Done
Comment by OVH - Thursday, 07 June 2012, 17:33PM

It's a bug from Cisco. The Mac addresses are not well learnt in hardware. That's why the trafic is flooded.

In the meanwhile, yo avoid this problem we do a "clear" of the Mac table. This have for purpose to allow to the switch to learn again the Macs and re-install them in hardware.

The bug was escalated to the constructor.


Comment by OVH - Thursday, 14 June 2012, 19:57PM

The bug was reproduced by the manufacturer.

They're trying to solve them as soon as possible.


Comment by OVH - Thursday, 05 July 2012, 15:05PM

5.104.249# sh plat fwm info error stats | i clone
[Wed Jun 20 12:25:32 2012 636088 usecs] msg 1 - fwm_api_mac_update(269): Mac Cannot learn mac 0007.b400.0102 on vlan 1.2 from if_index 0x1f670b80 : rc can't learn pvlan internal cloned mac address 0x41e80062 caller ra: 0x8420d65 0xb7cdd8be 0xb7ce0bd 0x8146d76 0xb36c7da3 0xb36dd6e8 0x814b459 0x414735c5
[Wed Jun 20 12:25:32 2012 636030 usecs] msg 0 - fwm_api_mac_update(269): Mac Cannot learn mac 0007.b400.0103 on vlan 1.2 from if_index 0x1f670b80 : rc can't learn pvlan internal cloned mac address 0x41e80062 caller ra: 0x8420d65 0xb7cdd8be 0xb7ce0bd 0x8146d76 0xb36c7da3 0xb36dd6e8 0x814b459 0x414735c5


The mac address of the gateway is not registered correctly in the forwarding system of the switch. In this case the trafic is flooded.


Thanks to this error message we could identify the server which was provoking this kind of problems.


The switches are currently stable and the infrastructure is operational.

In the meantime we escalated the problem to Cisco in order to set up a system allowing to protect the switch.