The system of load balancing presents instabilities on cluster 003 005 and 015, causing access temporary interruptions on shared websites hosted on these clusters.
We will reboot the routing card to fix the problem.
An interruption of few minutes will take place during the reboot.
If it does not work, the card will be changed tonight.
Update(s):
Date: 2011-01-04 13:23:26 UTC We move the traffic via 2 \"purifiers\" of traffic which blocks the attacks.
Date: 2011-01-04 12:28:55 UTC There is a performance bug on the new card. Some IP's cannot join one of the clusters (the 003).
We will move to the old card while looking for the origin of the problem...
Date: 2011-01-04 02:15:42 UTC We started the intervention.
Date: 2011-01-03 20:34:33 UTC The situation seems to have stabilized, but we're going to replace the card this night as a safety measure.
Date: 2011-01-03 20:32:28 UTC The restart is completed, we monitor the dispatcher to see if the situation improves.