8 messages in net.java.dev.glassfish.usersloadbalancer - active-healthcheck-ena...| From | Sent On | Attachments |
|---|---|---|
| glas...@javadesktop.org | Jul 31, 2008 7:34 pm | |
| glas...@javadesktop.org | Jul 31, 2008 8:03 pm | |
| glas...@javadesktop.org | Jul 31, 2008 11:08 pm | |
| glas...@javadesktop.org | Aug 3, 2008 5:09 pm | |
| glas...@javadesktop.org | Aug 4, 2008 4:18 am | |
| glas...@javadesktop.org | Aug 4, 2008 3:57 pm | |
| glas...@javadesktop.org | Aug 5, 2008 3:21 am | |
| glas...@javadesktop.org | Aug 5, 2008 6:18 am |
| Subject: | loadbalancer - active-healthcheck-enabled=true ?? | |
|---|---|---|
| From: | glas...@javadesktop.org (glas...@javadesktop.org) | |
| Date: | Jul 31, 2008 7:34:06 pm | |
| List: | net.java.dev.glassfish.users | |
AS 9.0, WS 6.1 - the versions are not in my control. Can't upgrade. auto-deploy
turned on.
I had clusterjsp deployed. Cluster has 2 instances. one instance was turned off.
failover works - all is well. However, when the instance was brought backup up,
the loadbalancer kept sending the requests over to the one that was running,
seemingly oblivious of the fact that the other one came back up. In short, the
failvoer is ok, but load-balancing was not happening. So, I set
active-healthcheck-enabled = true, restarted the webserver and the loadbalancer
then started working as expected (sticky round-robin)
Now the questions I have are:
1). Did setting the flag =true make it work?
2) I tried it on another installation of glassfish, but setting that flag to
true and stop/start doesn't do the trick. Only one server is getting hit all the
time..
3) per the documentation the flag is to be set to monitor 'healthy' instances.
now, why would we want to do that?
driving me nuts, this load-balancing..please [Message sent by forum member 'chilak' (chilak)]