RE: [cobalt-users] Random sites not responding (since Apache update)

["Andy Brown" <andy.brown@xxxxxxxxxxxxx>]

<snip>
> On our RaQ3, since installing the Apache update, random sites 
> on our server
> stop responding on random ports. For example, yesterday one 
> domain name's
> email and web site became unaccessible, on all its subdomains. Today,
> another domain name's site became unaccessible.
> 
> The server provides its own DNS.
> 
> The way to cure it seems to be to visit another site on the 
> server, which
> appears to 'wake' the server up. This is getting very 
> annoying! Anyone have
> any ideas?
</snip>


This sounds very bizarre indeed! Do the sites have their own IPs or do they all share a single IP ?
I'm unsure why you find the email port closed at times, or the apache port closed at times. I'd suspect this is more due to bad DNS rather than apache/update.

When it fails, try telnetting to the IP address of the defective site on port 25 or on port 80. Check you get a connection, if you do on port 80 type:
get /

and press enter, see if you get back all the html source code, or if it just hangs/sits there.

Keep trying this with the IP addresses, as i suspect when you go via the IP it won't deliver this behavious, in which case it may be down to the DNS on that box.
Frequently on the RaQ's (and many other linux systems running ISC's bind) after a reboot the name server doesn't come back up 'correctly' in that its behavious is at best erratic.
Do a nameserver reload via ssh by typing:
/etc/rc.d/init.d/named restart

and see if this makes any difference. If not, post your findings back!

Regards,

Andy
andy@xxxxxxxxxxxxxxx
http://www.raqpak.com/ <-- Raq/Qube unofficial PKGs and support advice