it's hard to tell based on the test output, but this may be related to SOLR-959.

At the very least, if the root problem here is in the test itself (and not in the ReplicationHandler) and work is done to cchange the test, cleaning up some of the problems noted in SOLR-959 might be easy to deal with at the same time

Yonik, the exception in your test output is the same one which Noble found some time ago at SOLR-926. I still trying to figure out why execution of a Callable will be rejected. A commit has been called so the searchExecutor should be alive and the Executors.singleThreadedExecutor uses an unbounded LinkedBlockingQueue so queue.offer should not fail.

I haven't been able to reproduce the issue with Java6 on WinXP. I'll download java5 and try again.

Yonik has been known to use a much slower computer and both of these failures are what you would see if the 3 second sleep above both asserts was not long enough. I'm guessing we just need to add a bit of code to try waiting a bit longer if 0 docs are found.

Committed revision 769052.

Thanks Mark!

I'm marking this resolved for now. If needed, we can re-open the issue.

Bulk close for Solr 1.4