You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tomcat.apache.org by Felix Schumacher <fe...@internetallee.de> on 2015/11/21 15:08:53 UTC
PersistenManager and StoreBase
Hi all,
while looking at the PersistentManager and StoreBase, I found a few more
inefficiencies other than bz 34319/47061.
If a Node is started, it will load all sessions from the Store, even if
those sessions are old enough, to be swapped out to the Store on the
next maintenance run. I think it would be nicer, to load only those
sessions, that stay in memory.
Another thing is the cluster awareness. At the moment a node will read
in all sessions from all nodes. I think it would be nicer to read in
only those sessions, that are meant for the node (jvmroute).
The same (slightly less problematic) is true for expiry of the sessions.
Every node will try to invalidate all sessions from all nodes. This is
not that much of a problem, when one node is fast enough to invalidate
all expired sessions, before the other nodes will try to do the same.
I would like to adress the first and probably the second point.
Any thoughts on this?
Regards,
Felix
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tomcat.apache.org
For additional commands, e-mail: dev-help@tomcat.apache.org
Re: PersistenManager and StoreBase
Posted by Mark Thomas <ma...@apache.org>.
On 21/11/2015 14:08, Felix Schumacher wrote:
> Hi all,
>
> while looking at the PersistentManager and StoreBase, I found a few more
> inefficiencies other than bz 34319/47061.
>
> If a Node is started, it will load all sessions from the Store, even if
> those sessions are old enough, to be swapped out to the Store on the
> next maintenance run. I think it would be nicer, to load only those
> sessions, that stay in memory.
>
> Another thing is the cluster awareness. At the moment a node will read
> in all sessions from all nodes. I think it would be nicer to read in
> only those sessions, that are meant for the node (jvmroute).
>
> The same (slightly less problematic) is true for expiry of the sessions.
> Every node will try to invalidate all sessions from all nodes. This is
> not that much of a problem, when one node is fast enough to invalidate
> all expired sessions, before the other nodes will try to do the same.
>
> I would like to adress the first and probably the second point.
>
> Any thoughts on this?
Go for it.
Mark
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tomcat.apache.org
For additional commands, e-mail: dev-help@tomcat.apache.org