[Watford] Watford Digest, Vol 136, Issue 1

[Yvan Seth]

On Mon, Aug 10, 2009 at 01:59:44PM +0100, Toby Deans wrote:
> Ok - So I am back in on ssh and I assume everyone else who has an
> account can now get back in. I have re-enabled the thttpd *just for
> now*  as we can reboot if it borks again without tying a piece of
> string to Magnus (thanks).

Perhaps we can either put a HTTP basic-auth password in front of the
thttpd UI?  Apache .htaccess style.  Or, lock down the UI to a whitelist
of IP addresses (either with thttpd config, or iptables.)

I've rm-ed the /unslung/rc.thttpd for now, this means the thttpd UI will
restart after a reboot too.

Nice to see the samba stuff didn't come back up after reboot at least!

> The Forum has broken - when I last looked at the logs it seemed that
> php just ran out of memory when we try to call the page from within
> the LightNEasy CMS - previously I had just linked it but even this
> does not seem to work properly now. Feel free to poke around if you
> get time as I'm stuck as to what the problem is here as it just about
> worked ok before.

Hm, I pushed the cherokee threads back to 3 (I reduced to 2 previously)
and the forum works again it seems.  Weird, but there you go!
 
> Ok.. so had a bit more of a poke around and there seems to be a bug in
> Simple Machines forum that calls for more php memory than is possible
> - looking at their forums this does not seem to be a problem with just
> the SLUG. I think I'm going to move to a lighter forum - do you have
> any ideas? Maybe phpbb?

I doubt phpbb is any lighter, probably quite the opposite.  Kat might
have time to do some research on lightweight forums.  

I don't suppose we can solder some more RAM into the SLUG? ;)

Another thought: we're running both cherokee and thttpd at the moment.
Would be worth researching whether or not we can just run our stuff
under the same thttpd as is running the Linksys UI.  It seems a bit
redundant to be running two separate HTTP daemons.

-Yvan