• I may have seen the same thing.

    Specifically I set up transactional replication on tables that see some usage, sent out a large batch update during the day (I don't leave the jobs running, I start and stop them so I can control when updates go out), and there was an instance where the user application locked up waiting.

    Since then we turned on read committed snapshot isolation for those replicated databases and I haven't seen the problem recur since. I'm not suggesting you turn it on, seeing as it could cause a host of issues, but it's an option.