[rabbitmq-discuss] Recovering from power failure

Mon Aug 3 21:32:24 BST 2009

If the message at
<http://lists.rabbitmq.com/pipermail/rabbitmq-discuss/2009-August/004719.html>
is representative of what people saw from my last email, then the
email got pretty badly mangled.  The rest of my email read like this:

>From the eof, I'm guessing that some file wasn't completely written,
so the persister is pretty angry.  So, I have a few questions :)
Should rabbit be robust to power failures?  i.e, is the persister a
durable file structure, or is it fragile to being killed when writing?
 Also, is there a better way to recover than by wiping out the files
in the rabbit queue dir and re-initializing the queues?

I also have the erl_crash.dump if this is a bug that could be fixed
with a crash.dump, and I have the files from the rabbit dir that are
causing the crash, if those would be useful to the developers.