You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@tomcat.apache.org by José Manuel Molina Pascual <ra...@gmail.com> on 2008/01/15 21:25:08 UTC

Slow Replication

Hello.

We use tomcat 5.5.23.

We have had some problems when replicating what we have called
"sleeping session", that is: sessions different from the one that you
get if you ask the container for a session (by using
httpRequest.getSession()). What we did is to use the session we get
(we call it carrier session) to replicate the changes to the other
instances using an attribute consisting on two hashmaps (changed/added
attributes and removed attributes) and a String (sleeping seesion Id).

I sent a mail with an strange problem (the instances that received the
replication received the past attributes, not the values of the latest
request served), I managed to solve it by killing the session each
time the node has to deal with a sleeping session.

Now, we have a less important problem, when we stress the cluster
about 10% of the requests seem to fail due to previous state not
having reached the other boards.

The question is: how can we improve the cluster performance when
replicating state?

Thanks in advance.

-- 
The complexity of software is an essential property, not an accidental one.

Our ignorance is God; what we know is science. (Robert Ingersoll).

Religion is regarded by the common people as true, by the wise as
false, and by the rulers as useful. (Edward Gibbon)

---------------------------------------------------------------------
To start a new topic, e-mail: users@tomcat.apache.org
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org


Re: Slow Replication

Posted by José Manuel Molina Pascual <ra...@gmail.com>.
We have solved the problem, in fact it seems that is not exactly slow
replication but slow deserialization of the sessiond data.

The way we solved this is not too good but deadlines..... you know, if
at the first try we don get the session data we try again up to three
times spaced by 100ms.
-- 
The complexity of software is an essential property, not an accidental one.

Our ignorance is God; what we know is science. (Robert Ingersoll).

Religion is regarded by the common people as true, by the wise as
false, and by the rulers as useful. (Edward Gibbon)

---------------------------------------------------------------------
To start a new topic, e-mail: users@tomcat.apache.org
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org