These were short outages, and ultimately leading to loss of 25% of network performance.
We suspected first the transit link or the SFP+ transceiver, but issue was not there.
Finally it was found on one of the Brocade FESX448-2G interior distribution switches, which admittedly has been causing some grief in the past too.
We have reverted back 1/4 of nodes from that switch to the old IDS switch, and rebooted the switch.
We will continue to monitor the switch, and ultimately if it does this again we will swap it for couple older generation switches.
Çərşənbə, Oktyabr 7, 2015
