jsdsupport wrote:Andy,
Cluster hosting means you shouldn't get down time if one of the servers in the cluster fails, however if badly set up it could be the reason why you are getting random response times, for example if you are not being allocated enough RAM. I have used both One and One (www.1and1.co.uk) and Web Wiz - my first nopCommerce site is with WebWiz currently - this is www.bgpaynecrafts.co.uk.
Thanks for this. Happy with clustering (Just so you know I am a technical business analyst for a russian software firm for my day job!) It would amaze me if someone the size of 123-reg had set up their systems wrong. I'm not saying they haven't though. Saying that not being allocated enough RAM would do it, especially if they are using VM instead!!! They have 'suggested' that I consider renting a dedicated box from them so I think I see their true motivation coming through... As a side note their dedicated boxes only come with 2Gb RAM so what does the shared hosting get with '00's if not '000's sharing a single box???
jsdsupport wrote:
If you are 64 bit in my view you need 2GB at least to run SQL/Server and IIS and more is better (One and One recommened 6GB - I am using 3GB with Webwiz at present for my Webwiz Hyper V server (effectively a cluster from a simple point of view, i.e. while not strictly speaking a SQL cluster with quorum for example the entire server is on shared hardware).
As a side note I was actually looking at possibly moving to Webwiz, what are they like? Any good?
jsdsupport wrote:
My other thought while reading your post is that NT is OLD, what year are we in, 2000? The only experience I have had recently of NT and 2000 is the sys admins didn't have a clue to be still using it in 2010.
NT 6.06.001 is Windows Server 2008 Enterprise Edition, NT just refers to the NT kernel underlying it that Microsoft developed IIRC for 'higher business reliability use'.
jsdsupport wrote:
I have done two things to improve my site (based on 1.80) performance wise. Firstly not all the tables have indexes (they have constraints but this hasn't been backed up with an index, so quite easy to find these and fix). I have also searched for all calls to the stored proc outlined in this thread and while the stored proc is still there, it isn't called in half as many places, in fact I am only calling it when paging of the output is required, in all other cases I have created new middle tier routines using Linq.
Hope this helps you in some way. I'll be posting my indexes on shortly.
Thanks for this, I will have a look into this and if you could post your findings it would be massively helpful. The part to me that is strange is that there are so many sites using this fantastic product and very few of them seem to suffer slow response times. Hence my first thought is that the host is just cr*p...
Either way thanks again and I hope everyone has a fantastic Christmas and a great new year.
Kind regards,
Andy