Posted: Mon Oct 28, 2013 3:01 pm Post subject: All neooffice.org websites moved to new server
If you had any problems connecting to any neooffice.org websites or downloading NeoOffice 3.4 over the weekend, it was because we were moving all of our websites to a new server.
We completed the move earlier today so all websites should be up and running normally. Although most sites remained up during the move, we did encounter unanticipated problems that forced us to shut down each of the websites for short periods of time.
While we did not wish to move our servers right after a new NeoOffice release, unfortunately we had no choice. Our webhosting provider notified us that they were moving all their customers to a new data center and they only gave us a few weeks to complete the move.
After having to move to a new server at a new webhosting company, it is now apparent to Ed and I that we cannot assume that any webhosting company will give us more than 30 days notice if they decide to drop support for our servers.
So, instead of setting up a second redundant server like we had before, Ed and I have upgraded our local infrastructure in California. Previously, we only had backup machines connected to our local business internet provider but since hardware is relatively inexpensive, we decided to setup 2 new machines locally that will be used as failover machines should our main server go offline for an extended period for any reason.
One of the limitations of our local business internet provider is that maximum bandwidth throughput is much more lower than the throughput provide by our webhosting provider. To prevent limited bandwidth throughput being an issue if we need to switch over to our failover machines, we made the following changes:
1. We now serve NeoOffice binaries using Amazon S3 and only serve binaries from our servers when Amazon S3 is unavailable
2. The 2 new machines are connected to 2 separate connections from our local business internet provider so that we can double are effective throughput
Although this rearrangement of our infrastructure was timeconsuming (and certainly not planned), we believe that we will be able to handle any future unplanned changes by our webhosting provider with a minimal amount of effort and minimal disruption of our various websites.
As an added bonus, we are finding that downloads from Amazon S3 are faster for users outside of North America. Amazon S3 apparently has optimized downloads for users in Europe, Eastern Asia, and Australia.
Although this rearrangement of our infrastructure was timeconsuming (and certainly not planned), we believe that we will be able to handle any future unplanned changes by our webhosting provider with a minimal amount of effort and minimal disruption of our various websites.
Thanks for the update, Patrick. It's a shame you had to do all of this unplanned server work, but glad to see you feel everything is in a better situation afterwards.
Smokey _________________ "[...] whether the duck drinks hot chocolate or coffee is irrelevant." -- ovvldc and sardisson in the NeoWiki
Thanks for the update, Patrick. It's a shame you had to do all of this unplanned server work, but glad to see you feel everything is in a better situation afterwards.
Well, things could have been much worse. We always had this feeling that whomever we were using could just go out of business without any warning and leave us with no servers.
Fortunately, Ed and I assume all machines and disk will fail at some point so we are very big on redundancy. All of our servers run on mirrored disks so that when one disk fails we can replace the bad disk within a few hours after the disk failure and the replacement disk will get automatically restored from the remaining good disk. Disks fail and mirroring has already saved us from a lot of pain over the years.
We also make database dumps and push those and backups of all files to a remote server every 6 hours so that we won't lose much data if there is a more serious hardware failure.
On top of all that, because there is a huge amount of irreplaceable information in our code repository and in our forums and NeoWiki databases, we do a manual backup of all such data and store such backup disks remotely in a secure location. Essentially, these manual backups constitute our low-cost, low-tech "disaster recovery" plan.
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum You cannot attach files in this forum You cannot download files in this forum
