Bug 778560 (SOA-1032)

Summary: Performance tuning details for ESB configuration
Product: [JBoss] JBoss Enterprise SOA Platform 4 Reporter: Dana Mison <dmison>
Component: DocumentationAssignee: Mark Little <mark.little>
Status: CLOSED NEXTRELEASE QA Contact:
Severity: high Docs Contact:
Priority: high    
Version: unspecified   
Target Milestone: ---   
Target Release: 4.3 CP02   
Hardware: Unspecified   
OS: Unspecified   
URL: http://jira.jboss.org/jira/browse/SOA-1032
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2009-08-28 09:41:10 UTC Type: Feature Request
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:

Description Dana Mison 2008-11-17 03:33:34 UTC
Affects: Documentation (Ref Guide, User Guide, etc.)
Date of First Response: 2008-11-17 08:38:13
project_key: SOA

Forwarded email:

I was recently on-site at a customer in *********************  who's
creating a PoC whether to use JBoss SOA-P or ProgressESB in their
environment. Opportunitiy size: 32-64 CPUs premium SOA-P subscription.
The PoC was about mainly two different use cases:

1. Proxy a WebService 
2. Consuming messages from an external JMS provider (Websphere MQ)

And then of course about documentation, scalability etc.


Problem here was:
------------------

1. Default configuration
By default they had on case 1) a very poor performance, because of the
default maxThread property. Why do we use 1 as a maxthread pool size?
In regards to our support team, it is because we want to make developers
aware to think about settings like this. -- It's true, but then we need
some kind of documentation about this. For example a warning.txt or
readme.txt saying

"If you want to have proper throughput, tune this and that, because the
default is 1".


2. Message consumer
Here we had very good performance (10.000 messages in 2.5 minutes), but
it was unclear why. They tried to configure the same maxThreads settings
as on the jbr-listener for case 1. -- Without success.

Support says, it must be the connectionPool size for the JMS provider.

How to solve:
-------------
We need a section in the configuration guide which belongs only to
tuning the ESB.
I am thinking about having an example (like trailblazer), which has all
you need in real live (webservices, jms, ftp, etc.). And then you talk
about tuning. So that it becomes clear what to configure and why.

Comment 1 Mark Little 2008-11-17 13:38:13 UTC
We already have a tuning section. Plus, which version of SOA-P is this affecting?

Comment 3 Jiri Pechanec 2009-08-28 09:41:10 UTC
Verified in the Admin Guide