This appears to be a timing-related issue with the non-sticky sessions.
4.0.20 will have a fix.
-- Scott
On 07/07/2011 04:17 PM, Eric Kreiser wrote:
On 07/07/2011 07:02 PM, Scott Ferguson wrote:
On 07/07/2011 03:51 PM, Eric Kreiser wrote:
I am having trouble getting resin-admin running inside my servers... is
there another way I can verify that the servers can see each other?
JMX or something?
The resin-admin is the easiest.
If you look at the logs (at the "finer" level), you should see something
like:
[11-07-07 15:30:14.378] {main} JvmDeadlockHealthCheck[OK:OK]
[11-07-07 15:30:14.381] {main} MemoryTenuredHealthCheck[OK:87.63%
free (627.312M free, 715.850M max, 88.538M used)]
...
[11-07-07 15:30:14.383] {main} HeartbeatHealthCheck[OK:OK]
You can look at JMX under the ClusterServer bean and look for the
isHeartbeatActive() and LastHeartbeatTime().
-- Scott
using JMX to look at both servers... the HeartbeatActive=true and the
LastHeartbeatTime is getting updated as well
On 07/07/2011 05:01 PM, Scott Ferguson wrote:
On 07/07/2011 01:37 PM, Eric Kreiser wrote:
I have my resin conf all setup... my servers starting/running, but if I
load balance... and don't use stickiness to servers... the session is
not kept in sync between the servers. I am trying to test this all...
so I am only running 2 servers - but both servers are configured as
triads so that they should know about each other. They are running on
the same machine(just with different ip/ports) - so I am not having
network issues or anything.
snippet below from my config
<cluster id="app-tier">
<persistent-store type="cluster"/>
<server id="a" address="10.10.1.145" port="6803">
<bind-ports-after-start>false</bind-ports-after-start>
<watchdog-port>6701</watchdog-port>
<watchdog-address>10.10.1.145</watchdog-address>
</server>
<server id="b" address="10.10.1.141" port="6803">
<bind-ports-after-start>false</bind-ports-after-start>
<watchdog-port>6701</watchdog-port>
<watchdog-address>10.10.1.141</watchdog-address>
</server>
<web-app-default>
<session-config cookie-length="100" use-persistent-store="true">
<session-timeout>200</session-timeout>
<always-load-session/>
<save-mode>before-headers</save-mode>
<reuse-session-id>false</reuse-session-id>
<serialization-type>java</serialization-type>
<always-save-session>true</always-save-session>
</session-config>
</web-app-default>
</cluster>
what am I missing?
That looks fine. In the /resin-admin, do you see the servers able to
connect to each other?
(The health tab will show connection failures in the Resin|Heartbeat
check and also the Heartbeat tab.)
-- Scott
thanks
Eric
_______________________________________________
resin-interest mailing list
resin-interest@caucho.com
http://maillist.caucho.com/mailman/listinfo/resin-interest
_______________________________________________
resin-interest mailing list
resin-interest@caucho.com
http://maillist.caucho.com/mailman/listinfo/resin-interest
_______________________________________________
resin-interest mailing list
resin-interest@caucho.com
http://maillist.caucho.com/mailman/listinfo/resin-interest
_______________________________________________
resin-interest mailing list
resin-interest@caucho.com
http://maillist.caucho.com/mailman/listinfo/resin-interest
--
*Eric S. Kreiser* *
* Senior Software Architect
*M**z**inga**
* 5095 Ritter Road . Mechanicsburg, PA 17055
---------------------------------------------------
*Call me:* 717.458.9804
*Fax me:* 717.790.0401
*Email me:* ekrei...@mzinga.com <mailto:ekrei...@mzinga.com>
*Learn more:* http://mzinga.com/v/ekreiser/
*Follow Mzinga:* http://twitter.com/mzinga
*Our Customer Experience Community:*
http://community.omnisocial.mzinga.com/app/signon/
_______________________________________________
resin-interest mailing list
resin-interest@caucho.com
http://maillist.caucho.com/mailman/listinfo/resin-interest
_______________________________________________
resin-interest mailing list
resin-interest@caucho.com
http://maillist.caucho.com/mailman/listinfo/resin-interest
