How long does it take to reboot a few servers?
View Single Post
Join Date: Dec 2007
01-30-2010, 11:33 PM
Originally Posted by
Just wondering how long it takes to reboot the servers?
We don't know exactly what they're doing, but I'll pretty much bet it's not rebooting servers.
The only way I could imagine it could honestly be that it takes so long to reboot the servers is that they need to push new code onto all of them, and their servers all have to sync the new code from a limited number of "code synching" servers. But that seems pretty improbable to me. The bulk of the data is content, and that doesn't change rapidly. The actual "guts code" is likely very small, and that's where the bugs would be.
The only things I can think of, other than there not being enough people ready to get things done (which I doubt given the importance of this event to Cryptic) is that either:
1) They are freezing the universe and trying to find the cause of each problem and fix it so it never happens again, needing time to troubleshoot. IMO, this is a mistake. Better to copy out the data surrounding the crash, maybe quarantine the actual machine that faulted, and get the world back up. You can then work on the bug while people play and schedule downtime to fix it. Odds are, if the world crashes again, it will be a different bug, and you can find/fix more bugs in less time this way.
2) The bugs are requiring some time consuming process, such as updating/recovering/reindexing the entire database, to fix. This could be the case if they don't have good referential integrity in their back end storage. But having a large number of bugs of this kind would be extremely unusual.