
The European ones and Asian ones maybe - US servers are still down.jellygoose wrote:ralph and plev: servers aren't down there is a problem with their authentication server. i played this morning and there are report of people still playing now.
Now I'm no IT guy but my thought was this;
Why cant they use the Tuesdays low player population to push all players onto one half of the servers, shutdown and update the empty servers. Then once the first half the servers have been updated have a restart boot everyone off for 5mins so they call all log into the updated servers while they update the other half?
I'm sure this approach would have hurdles to jump before it could be done, But i know there are amazing things happening server side such as combining realms for the low pop zones in the next wow expansion.


jellygoose wrote:ralph and plev: servers aren't down there is a problem with their authentication server. i played this morning and there are report of people still playing now.
Now I'm no IT guy but my thought was this;
Why cant they use the Tuesdays low player population to push all players onto one half of the servers, shutdown and update the empty servers. Then once the first half the servers have been updated have a restart boot everyone off for 5mins so they call all log into the updated servers while they update the other half?
I'm sure this approach would have hurdles to jump before it could be done, But i know there are amazing things happening server side such as combining realms for the low pop zones in the next wow expansion.
Ralph Wiggum wrote:My issue isn't so much the maintenance, as opposed to the fact that it is on during peak time in Australia. I can't think of many other AAA titles which screw over Oz gamers so much. I've learned a big lesson from the way Blizzard treats its customers and will think twice about buying one of their games in future which use a similar process. And they sure as hell wont be getting a cent from me using the RMAH.


cdkit3 wrote:I personally don't understand how in this day we need to have an 8 hour downtime for maintenance. What should be done is the servers should be ran as a load balancing cluster. This would allow them to pull out individual nodes, install the updated code, however still remain running the previous code. Put the server back into the cluster, continue until all nodes have the update code, and then have a ~30 minute downtime where everything is offline to switch over to the new code. Then bam, no extremely upset customers.
Sure it may mean that there is an increase in the number of people who are stuck receiving Error 37, but surely that is a better outcome than preventing the whole customer base access...

Users browsing this forum: No registered users and 1 guest