Im still nagging them about it astral.
They assigned more connections to the server, but missed the point I'd been making about how sometimes it was a nightmare to be able to connect and also the fact that more often than not I seemed to be shunted towards ccgi02.
Since putting the server script up, Ive noticed that if I hit any other the other servers, then the site will run fine, and can be ok.. but ccgi02 seems to be the one running slow and has a much higher load average.
They have changed something on the loadbalancer today, but as colin pointed out, it seemed like the load balancer was working in reverse and trying to shunt everyone on ccgi02.
PN are still saying they cant seem to replicate this as it all seems to be running ok from there end.. but point being.. yeah hit a server thats fine and things will run sweet. Im sure that the times when I cant connect are those when Im hitting ccgi02 which is too busy to accept anymore connections.
----------
ok.. just as I was typing this, just checked and have a reply to a question directed at one of the more senior engineers.
With regard to CCGI - we have signoff for more hardware and indeed it's on the Development workstack to buy and build additional servers for the platform.
With regard to CCGI02, we have recognised it to be a problem and over the last few days one of my engineers has been working to identify what's going on with it - he has seen the slowdowns and can see a pattern to them and he's currently investigating what's going on with it. He's also been doing some tweaking to the way the load balancer works to try and make the platform more... well, balanced!
He and I will continue to work on the open CCGI slowness problem and we'll report back when anything conclusive has been identified.
So it seems they have finally acknowledged a problem with ccgi02
I knew that there had been mention of new hardware for the ccgi platform a couple of months ago, and iirc it is/was next in line after the new mail servers.