We have installed Web 8 and have object caching configured on both Discovery and Content service. We have also set the defaultStorageId
to defaultDb
and set cache
equal to true
. We have setup a logger for caching to make sure cache lists are being created and to confirm there are no errors about memory size or other things.
However, when I load my CD webpage with all that caching enabled I get the same performance from the Web services as I do when all the caching is disabled. The performance is not good and enabling caching does not help it looks like. Any ideas?
FYI. When I enabled longlived cached on the CD web app performance is great. The page loads in < 500 ms, but when I disable the CD web app cache it takes 10 seconds to load because it has to talk to the Content and Discovery service.
Thoughts?
...
<ObjectCache Enabled="true">
<Policy Class="com.tridion.cache.LRUPolicy" Type="LRU">
<Param Name="MemSize" Value="512mb"/>
</Policy>
<Features>
<Feature Class="com.tridion.cache.DependencyTracker" Type="DependencyTracker"/>
</Features>
<RemoteSynchronization FlushCacheDuringDisconnectInterval="20000" Queuesize="512" ServiceMonitorInterval="10000">
<Connector Class="com.tridion.cache.RMICacheChannelConnector" Host="...." Port="1099"/>
</RemoteSynchronization>
</ObjectCache>
<Storages>
<Storage Class="com.tridion.storage.persistence.JPADAOFactory" Id="defaultdb" Type="persistence" dialect="MSSQL">
<Pool CheckoutTimeout="120" IdleTimeout="120" MonitorInterval="60" Size="10" Type="jdbc"/>
<DataSource Class="com.microsoft.sqlserver.jdbc.SQLServerDataSource">
...
</DataSource>
</Storage>
</Storages>
</Global>
<ItemTypes cached="true" defaultStorageId="defaultdb"/>
...
I can't remove the comment at this point but as noted below I also tried excluding the taxonomy but that did not change anything.
<ItemTypes cached="true" defaultStorageId="defaultdb">
<Item cached="false" storageId="defaultdb" itemType="Taxonomy" />
</ItemTypes/>
Also in regards to other comments:
1- I removed the FlushCacheDuringDisconnectInterval
and ServiceMonitorInterval
nodes so they would take on their default value but no change.
FlushCacheDuringDisconnectInterval
andServiceMonitorInterval
- in which case - what was the rationale for the figures used above?