iWeb Communities: Blog

• Blog
  • All
  • iWeb
  • Web Development
  • Web Hosting
• Forum
• Status

Comments

1. Jade
   November 13th, 2008 1:19 pm

   Hi again,

   Since the server’s still down, I though maybe you could post an explanation regarding the “Zero Downtime Architecture” and how this problem developed in the first place.

   The whole reason I came to iWeb was because of the “Zero Downtime” which is the most important thing to me.

   I don’t care about getting reimbursed for the downtime. I just don’t want any downtime at all.

2. iWeb-Op
   November 13th, 2008 1:37 pm

   Storage is still a single point of failure in the architecture. You can have details here : http://iweb.com/web-hosting/zero-downtime-architecture/ . While it is easy to separate the operation of the hosting environment from physical servers, it is still relying on the storage to serve files and the operating system from. In this case, we cannot restart the servers until the filesystem problem is solved.

3. Jade
   November 13th, 2008 1:43 pm

   Thanks for the answer. I appreciate your effort to provide an explanation.

   Although I was under the impression from the Reseller’s info (http://iweb.com/web-hosting/resellers/zero-downtime-architecture/) that there was redundant storage which should theoretically provide a duplicate filesystem to the servers while the original filesystem is being repaired.

4. iWeb-Op
   November 13th, 2008 3:33 pm

   The storage uses redundant arrays, if a hard disk fails it will not affect the storage. We replicate the storage to a backup storage system but that system can’t be used in production, it would be used to restore the data to the production file system. Switching to a backup file system, unless it is replicated in real-time would also be problematic if there is a replication delay. We will study the possibilities to protect the architecture from file system errors as what happened today and will post details if we decide to improve the storage architecture in the future.