db-derby-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Bryan Pendleton <bpendleton.de...@gmail.com>
Subject Re: Where to look in the code to find out why XA transactions and database locks are not be cleaned up
Date Sat, 16 Mar 2013 23:47:15 GMT
 > when I turned on the tracing, the problem goes away so that pretty much is
 > a good indication that there is a multi-threading problem relating to timing.

Sounds like it. Bummer.

You've already opened a JIRA issue, right? If not, please do so.

The good thing is:
- you are able to (semi-)reliably reproduce the issue
- you can describe the symptoms quite clearly.

So you've got this bug in your cross-hairs :)

When I get a case like this, one approach I find productive is to
plant assertions in the code. Start adding code to validate
invariants, and keep moving those assertions back closer and closer
to the start of the processing, until you find the point
where an invariant is broken.

Since you know that the problem is a synchronization problem,
pay particular attention to the synchronization behaviors in
your instrumentation.

I hope this helps; I know it's rather a bit vague.

thanks,

bryan


Mime
View raw message