|
In article <2242CF76-779F-4FF4-9F99-D1DE56811F17@xxxxxxxxxxx> you wrote:
> In my experience, I've had to remove the real server entirely when I
> have persistence active. Unless of course you're willing to wait an
> hour :). But I also have had new connections continue to trickle in
> to a zero-weight real server in the absence of persistence, so I
> think there's also a general issue where a zero weight isn't
> literally zero.
>
> I remove real servers within health monitoring in the general case,
> as opposed to zero-weighting them. If zero worked as expected, that
> slightly drastic measure (hard connection interruptions, orphans)
> wouldn't be necessary.
I belive that the problem you are seeing is the problem that is
addressed by /proc/sys/net/ipv4/vs/expire_quiescent_template
expire_quiescent_template - BOOLEAN
0 - disabled (default)
not 0 - enabled
When set to a non-zero value, the load balancer will expire
persistant templates when the destination server is quiescent.
This may be useful, when a user makes a destination server
quiescent by setting its weight to 0 and it is desired that
subsequent otherwise persistant connections are sent to a
different destination server. By default new persistant
connections are allowed to quiescent destination servers.
If this feature is enabled, the load balancer will expire the
persistance template if it is to be used to schedule a new
connection and the destination server is quiescent.
(I'm currently trying to work out why that and related documentation
is missing from the kernel tree)
http://archive.linuxvirtualserver.org/html/lvs-users/2004-02/msg00224.html
--
Horms
H: http://www.vergenet.net/~horms/ W: http://www.valinux.co.jp/en/
|
