[rabbitmq-discuss] procedure for dealing with timeout_waiting_for_tables

David Wragg david.wragg at lshift.net
Tue Sep 8 16:31:16 BST 2009


Hi Tsuraan,

tsuraan <tsuraan at gmail.com> writes:
> I had a machine that had the timeout_waiting_for_tables (not in a
> cluster, just a single node).  I'm guessing that its hostname had
> changed, and that was making mnesia angry.  I have two questions about
> this.  Is there a good way to figure out exactly why mnesia isn't
> starting, i.e. is it a hostname change, a corrupt mnesia file, etc?

The answer to this part is no, not really.  Searching for reports of
similar symptoms on this list is currently your best bet.

> Second, to fix this, I thought I could copy out my persister logs,
> wipe the rabbit dir from /var/lib/.../mnesia, restart rabbit, and move
> the persisters back in.  That didn't seem to work, since the queues
> weren't defined anymore, so the log replay apparently just threw
> everything away.  Would remaking my exchanges, bindings, and queues,
> then moving in my persister logs do the trick?

Yes, the likely cause is your hostname change. If you recreate all the
resources, stop rabbit, copy the saved persister logs back in, and
start rabbit, it should do the trick.

David

-- 
 [][][] David Wragg       | mail: david.wragg at lshift.net
   [][] Senior Developer  | tel: +44 (0)20 7729 7060
 []  [] LShift Ltd        | web: www.lshift.net




More information about the rabbitmq-discuss mailing list