From: fferrer10@vodafone.es
Date: Thu Jan 11 2007 - 19:48:24 ART
Hi guys:
We have a server farm with a lot of blade servers connected to several Catalyst 6500 switches through the respective blade-switches, and using channeling to give extra-bandwidth to the server connections.
Yerterday, a misconfiguration on a newly added fiber channel was traduced on "merging" 2 vlans in a layer 3 loop, making the 6500 giga interfaces going to very high level utilization and causing a general "outage" of 15 minutes on all the server environment, not only on the equipment of the 2 vlans affected by the misconfig.
Anyone knows effective solutions to minimize the impact of these misconfigs errors on critical environments? We are rate-limiting the % of bandwith allowed for broadcasts & multicasts on the core & distribution links of the farm, but we don't know if there are more efficient solutions to these problems.
Regards and TIA
Juan
This archive was generated by hypermail 2.1.4 : Thu Feb 08 2007 - 23:46:56 ART