Hi Jan,
Thank you for your reply.
The scavenging defaults are fine with me. I was just thinking the
log lines indicated I forgot something.
The session ids are a but clumsy for me since our system optionally
has it's own JGroups-based clustering logic and therefore already
prepends its own cluster prefixes (and a separating $-sign) to
session ids. The current combination of prefixes looks very strange.
Perhaps I could somehow set the cluster name to an empty String?
I already subclass the DefaultSessionIdManager and have something
like:
override def newSessionId(request : HttpServletRequest,seed : Long)
= mapping(super.newSessionId(request,seed))
where mapping is a String => String function that optionally
prepends a cluster id prefix.
Can I get the raw session id directly somehow or do I need to trim
off the length of the current cluster id?
Thanks again,
Cheers,
Silvio
On 12/15/2016 11:19 AM, Jan Bartel
wrote:
Hi,
The node id has been used for quite a while
now in jetty session ids to increase likelihood that they
are unique across nodes in the cluster. If you'd like to
change that then I suggest that you subclass the
Defaultsessionidmanager and override the I'd generation
method.
The log lines you are seeing about the
scavenger is just letting you know what the default settings
are. If you want to change it then make a Housekeeper
instance, set its scavenge interval, and set the
Housekeeper onto the Defaultsessionidmanager.
Jan
_______________________________________________
jetty-users mailing list
jetty-users@xxxxxxxxxxx
To change your delivery options, retrieve your password, or unsubscribe from this list, visit
https://dev.eclipse.org/mailman/listinfo/jetty-users
|