I am managing a site that seems to have gone awol.  There is a category-based
sidebar menu (using the standard categories widget) but clicking on any of
these either gave the same result (ie it presented the same page/post) or
resulted in a 404.  Clicking on the Read More has the same effect.  The UK
cookie check at the top doesn't retain its setting, even though /tmp is
writable on the server etc.

This occurred today so I rolled the site (tree and database) back to Saturday
and there is still a problem, even though AFAIK the site was working fine
then.  I had added a .maintenance file into the docroot while restoring the
site and I have removed this.  But now instead of the 404 I get the
maintenance message as if $upgrading is still valid!

Even funnier, I am using Suffusion and for some reason after upgrading that,
the head background became black.  Now this should have gone back with the
restoring to the older version of the theme.

It isn't a local browser issue - I have seen the problem on 4 browsers.  But
there would appear to be a corruption or something.  .htaccess is set

You'll see what I mean at http://ayorkshirecook.com

BTW it has some great recipes when it's working!

--
Best wishes

John Logsdon

Trusting a sudden upgrade in Suffusion can cause a broken site.  Nice
theme.  Rock solid.  Try to find a way to test it before/after upgrade.
Google search for bugs in the new version.

The Suffusion upgrade didn't cause the problem as I did that after the event!  

When the problem was first brought to my attention, I upgraded a couple of
plugins that needed attention and then the theme.  

The problem would appear to be either a corruption in the tables (which all
analyse as OK) or a cookie issue but that doesn't happen on other sites on
the same server.

One solution would be to dump all the posts and rebuild the site but that's an
extreme option.  I may try an older regress (I have a month on tap).  And I
would like to know why it has happened anyway.

On Monday 06 August 2012 18:27:15 Richard Ibbotson wrote:

John Logsdon

One thing where I've seen inexplicable behaviour occurs where cache memory
(memcached) is either currupted or no longer matches what the code is
expecting (because something changed).

Similarly I've had opcode cached (in APC) give me problems when code gets
updated.

A restart of either memcached  (in the former case) or Apache (in the
latter)  sometimes fix the problem.

If you are on shared hosting, this is unlikely to be your problem.

Mike

On 6 August 2012 18:53, John Logsdon <j.logs...@quantex-research.co.uk>wrote:

I am using eaccelerator.  Either I can switch it off (comment out and restart
apache) or just clear all the cache files out with brute force and bash (!)
well rm  /var/cache/eaccelerator/*/*/*

I'll give that a whirl....

Some progress - at least I don't get the 404 or Maintenance message any longer
but it still always returns to the same post.  Continue reading now works but
it is still on the first post, now the one that is needed.

Why it should return the full post after read-more is a mystery.  And there is
still the UK cookie problem when /tmp is correctly permissioned.  I will
deactivate that plugin and reactivate it. You never know!

On Monday 06 August 2012 18:55:18 Mike Little wrote:

John Logsdon