No posts for a few days but that's just because there has not been anything to say, not because I have gone off blogging :o)
Its taken a few days to move one of our applications to Manchester and it has taken a while to get over a sticky routing problem, in that my router at home which is IPSEC tunnelled to the office would not allow me to access the server at Manchester even though RIP was enabled. So in a nutshell from home I could get onto anything on the head office sub net but not access any further sub nets of which we have one each for the pubs and one for Manchester. After looking into static routes which didn't help and trying to investigate RIP I realised that this was a waste of time because of course the grand plan is to have Manchester as a self reliant application server in its own right and by routing through the head office I was therefore reliant upon head office to get to Manchester.
So the quest now became setting up a second IPSEC tunnel to Manchester alongside the existing tunnel to Head Office, a model which would eventually be rolled out to all the remote locations. Now on paper setting up a second tunnel should be really straight forward but if anyone else has tried it you will know that sometimes they work and sometimes they don't, the most frustrating part being that unless you go the whole hog and set up an event log server you get no feedback so when it doesn't connect you just have to keep trying. 10pm last night I was beginning to loose my rag, deleted the tunnel profile from my router, set up another identical profile and bingo - worked first time.
So now the model is as illustrated below which means that when the applications are mirrored properly and all the data is replicating we will have a nice fail over system without breaking the bank. The only trade off being that if Manchester went offline moving users to head office is simple but going back means a backup/restore of the data and a bit of a faff due to the choice of replicating data rather than clustering.