pszilard
Apr-15-2011, 03:07 PM
THE FOLLOWING WAS SENT TO THE HELP DESK.
Dear SmugMug,
Account: REMEKTEK
I pay for a PRO level photo web site service, and normally I am delighted with what services I get for this. HOWEVER yesterday I wanted to show my portfolio to a client and the site was down for many HOURS! This is an unacceptable situation for a PRO site!
I am told that SmugMug backs up my photos to two different geographic locations (great!), so why don't you have a redundant server system? A business critical service should have a minimum of THREE servers: A and B should be redundant production servers (so if A fails, then B takes over), and C should be a test server.
When you want to roll out new enhancements you first test them on C. Then when all is working, assuming that Server A is working, you transfer to production the software from C to B and verify that it works still. THEN you switch production from A to B and transfer to production, the updates to A. When that was successful, you can switch back to A as primary and B as redundant backup. Then you can start working on the next cycle on C.
The above scenario is what should be implemented and is WELL within the scope of your IT department! However at the VERY LEAST if all your servers are down, then PLEASE put up a single page to say that the site is temporarily unavailable and what the expected time for return is (preferably in the time zone of the viewer, or else show the server's local time as well as people can figure it out).
I sincerely hope that you can implement this in the near future, and continue in your otherwise excellent service!
Dear SmugMug,
Account: REMEKTEK
I pay for a PRO level photo web site service, and normally I am delighted with what services I get for this. HOWEVER yesterday I wanted to show my portfolio to a client and the site was down for many HOURS! This is an unacceptable situation for a PRO site!
I am told that SmugMug backs up my photos to two different geographic locations (great!), so why don't you have a redundant server system? A business critical service should have a minimum of THREE servers: A and B should be redundant production servers (so if A fails, then B takes over), and C should be a test server.
When you want to roll out new enhancements you first test them on C. Then when all is working, assuming that Server A is working, you transfer to production the software from C to B and verify that it works still. THEN you switch production from A to B and transfer to production, the updates to A. When that was successful, you can switch back to A as primary and B as redundant backup. Then you can start working on the next cycle on C.
The above scenario is what should be implemented and is WELL within the scope of your IT department! However at the VERY LEAST if all your servers are down, then PLEASE put up a single page to say that the site is temporarily unavailable and what the expected time for return is (preferably in the time zone of the viewer, or else show the server's local time as well as people can figure it out).
I sincerely hope that you can implement this in the near future, and continue in your otherwise excellent service!