Lots of you may have experienced problems when checking your heroes yesterday, as GV servers were not available or were returning error messages. First of all, we’re sorry that happened, and we want to let you know about the situation and what we plan to do in order to prevent it from happening again.
Our hosting provider informed us that due to some excavation work, a fiber optic cable that connects our datacenter (and 4 others) to the Internet had been cut, which caused everything to work really really slowly or sometimes not work at all. We tried to mitigate the issue by tuning up the servers, but that wasn’t enough to work around the underlying connectivity issue completely. After approximately 8 hours, the underlying connectivity issue was fixed by our hosting provider, which resolved the issues in accessing our servers.Unfortunately it’s almost impossible to manually restore or rollback all the bad things that happened to your heroes while they were wandering around without supervision from their gods, but here’s what we’re going to do to prevent it in the future:
- tune up timeouts in our mobile apps to better handle situations like this
- fix some newly found bugs in our Android app related to error messages when servers are not available (some of them were already fixed in v2.5 available in Android Market, but this version is still going through the Amazon Appstore review process and not yet available to those who downloaded the app from the Amazon Appstore)
- provision for additional server capacity to handle even more requests
- follow us on twitter @godvillegame to get the latest updates, especially at times when our site is not accessible
- if you’re using our Android app and for some reason you were logged out of your account, but had never set a password: please try to login with your godname and an empty password
- be aware that due to Amazon’s review process, app updates to Amazon Appstore are very likely to reach you several weeks later than the same update in Android Market
During yesterday’s non-stop endless patching & tuning session, we accidentally messed up one of the servers, which eventually backfired causing some of the accessibility issues observed today. As of now, this problem is resolved and everything should work fine. And to prevent this happening again, we promise to try get some sleep and be more careful even when our inbox is flooding and things are on fire.
If you don’t use twitter, here is the good old-fashioned RSS feed for blog posts.