On Wed, 2010-05-12 at 13:00 +0200, Mauro Vanetti wrote: > Hello! > I'm testing load balancing between four nodes placed in two distinct > machines (2 + 2). The load balancer is on the first machine, where nodes > 11 and 12 are. Nodes 21 and 22 are on the second machine. > When I reboot the second machine, the load balancer detects the error > and redirects all new requests to nodes 11 and 12 only, until the nodes > on the second machine are up again. > This is all very good, the only problem is what happens in the critical > seconds. Clearly, a few sessions will fail (those already started but > not yet finished on nodes 21 and 22 when the second machine is > rebooted). When this happens, the log shows that for 1-2 seconds no > requests are managed (including those on the first machine!). > This goes to the point that even if I launch an automated ping (10x a > second) on the first machine, the pinging is not logged during the > critical seconds. > > Real example (requests per second logged in access_log): > 100 > 112 > 102 > 92 > 32 <-- CRISIS > 0 <-- PEAK OF THE CRISIS > 3 <-- CRISIS > 25 <-- CRISIS > 55 <-- RECOVERY > 98 > 93 > 101 > > What can be the reason for this behaviour? > Thank you very much for your help. Best regards! What kind of Load Balancing? Apache wouldn't log a ping, since its not a valid HTTP request, so I don't know what kind of logging you're talking about. -- Mark Watts BSc RHCE MBCS Senior Systems Engineer, Managed Services Manpower www.QinetiQ.com QinetiQ - Delivering customer-focused solutions GPG Key: http://www.linux-corner.info/mwatts.gpg
Attachment:
signature.asc
Description: This is a digitally signed message part
|