Troubleshooting
Problem
DB connection pooling in Yantra 7x agent/app server
Symptom
In the Yantra 7x web application deployed on an application server, the number
of DB connections in the pool (min/max) etc. are controlled via exposed
properties in the application server.
The Yantra 7x agent/integration
servers however manage their own DB connection pooling. Is there an
optimization done based on the number of services etc? Or is this is a fixed
min/max number that are not exposed? Wanted to understand this better, since
this might have an impact on how many time-triggered transactions we might want
to have in a single agent/integration server.
Cause
Resolving The Problem
The logic/process followed for connection pool managing in the
agent/integration servers is as follows:
1) On startup, the
agent/integration server establishes # of connections = (sum of # of threads
across all criteria/services within a given instance/JVM of that
service/agent). So, if you started the schedule agent with 3 threads, it would
establish 3 DB connections right off the bat.
2) An additional connection
"may" be established for the trigger.
3) An additional connection can be
established for the statistics persistence.
In the yfs.properties, we
expose a couple of parameters for the connection pool reap time etc. These come
into play only if a connection in the pool within the agent JVM has been idle
for more than a certain period of time.
So, in the worst case scenario,
for the above example, you could briefly have 5 DB connections but, for the
most part you will have 3.
Historical Number
PRI49567
Product Synonym
[<p><b>]Fact[</b><p>];
Was this topic helpful?
Document Information
Modified date:
16 June 2018
UID
swg21531315