|
Hello mike,
Thursday, November 16, 2000, 10:49:10 PM, you wrote:
mjo> Hello,
mjo> It sounds to me like you're thinking of a cluster more like the Veritas
mjo> or Microsoft hot-standby products. Lvs is not like that at all. The
mjo> former keeps a second machine in 'ready' mode, and swaps it into use when
mjo> the primary machine fails (thereby creating a window where requests are
mjo> not being answered). Lvs, on the other hand, balances requests between a
mjo> group of live servers. All servers in the cluster are up all the time, so
mjo> if a single server fails, the only disruption is the delay between the
mjo> failure and your monitor removing that server from the pool. Does that
mjo> explaination help at all?
What I really wanted to say is about directors' failover, not
realservers', but I didn't make myself very clear, sorry!
Thank you mike!
(I cannot send email to this address mike@xxxxxxxxxx, sorry!)
--
Best regards,
huatao mailto:thua@xxxxxxxxxxxxxxx
|
