[OSM-dev] Handling 500 errors

[Tom Hughes]

In message <1188577145.28903.255.camel at localhost>
          Dave Hansen <dave at sr71.net> wrote:

> On Fri, 2007-08-31 at 12:26 +0100, Tom Hughes wrote:
> > 
> > I think this was just caused by lock timeouts, which seems to have
> > been down to MySQL suddenly deciding to use a very bad execution plan
> > for one of Potlatch's queries.
> > 
> > That query has been tweaked now and things seem to be running more
> > smoothly.
> 
> I'm still seeing quite a few of those as I speak, still at about the
> same rate.  Although, now tends to be a higher load time for the
> servers.

Because the ones you're seeing are down to deadlocks, as I explained
last night - the ones other people were seeing this morning with nodes
and segments were down to lock timeouts.

Tom

-- 
Tom Hughes (tom at compton.nu)
http://www.compton.nu/