You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@servicemix.apache.org by Felipe Gaucho <fg...@gmail.com> on 2007/03/15 09:43:37 UTC

fail safe test

ok, I have 3 light-weight components.. it works - really slow, but works :)
and I will try to learn more about thread pooling.. but now I have this
priority: to check the fail-safe behaviour of my compoennts on ServiceMix.

---
The flow: 

c#1 reads a database data and pushes a normalized message into router (1 for
each new record) -> destination c#2

c#2 receives the message, do some business and push another message into
router -> destination c#3

c#3 reads the message, log and finish the flow -> done(message exchange)

OBS: 

* every component includes a timestamp into a database table, called
monitor: it means, I can check  what time each component received and sent
its messages.

** I include the new records into the database manually, a population insert
query. Each record has a unique identifier (primary key ID). This way I can
control how many records should be consumed by c#1 - ok, I have a
configurable Java class I can say how many records to insert and the period
to repeat - simulating a data importer from an external system.

---
The test:

keep the system working, and turn on/off the JBI container several times..
later, check the number of messages consumed by each component (the
throughput of the system) and also check if any message was lost.

question: how to configure the ServiceMix to support fail over mechanism ?

-- 
View this message in context: http://www.nabble.com/fail-safe-test-tf3407001s12049.html#a9490412
Sent from the ServiceMix - User mailing list archive at Nabble.com.