The application uses -
Websphere 6.1 clusters with identical clones on nodes,
Websphere MQ thorugh datasources
Oracle database through datasources.
IHS thorugh Websphere plugin.
Could you please guide me to any primer tuning document (preferably from IBM) that suggests
how the various websphere pools (thread / datasource / plugin etc) should be arranged.
I faintly remember the concept of pool funnelling where the number of database pool should be less that server thread pool that in turn should be less than server thread pool. The concept was that keep more of the requests outside than bringing them in and making them wait.