Server Roll Back?

[FadedSentry]

I had read on Facebook that there was a quick maintenance, however now that I've loaded the game up it's rolled back everything I had done since last night. I've lost about 9 hours of work and it even undid my arena, pet feeding and everything else.

[erebs]

i dont think you lost 9h of work if you see rested time then make it x2 to see how long you were offline for me it was 1h rollback

[FadedSentry]

It was offline for 360 minutes is what it said, so I guess I had a 3 hour rollback. I know it was a complete rollback because before I went to bed I made sure to update everything, do pet arena/soul orbs, etc and when I got up it had me redo all of that. I got less diamonds for expeditions, less XP for soul orbs, and had lost however much gold I had gotten in that time period due to the rollback.

[Aaron Storm]

From what I've seen if your computer loses connection with the servers and it doesn't kick you out of the game, usually it does, then you can still do a lot of things. I think usually it only happens if you don't fully lose connection but you have a really bad connection. Everything you do does not update on the server. I haven't had it happen to me in a few months.

There was like one hour I remember where I would lose connection and could then reconnect after like five minutes. But when I connected anything I did would not persist after I refreshed again. And there were certain things it would not let me do at all, such as my party showed as offline and I couldn't do any challenges.

With 360 minutes of rested time you would have been offline for at least 12 hours, 360 minutes times 2. I suspect when you did everything the night before you somehow weren't connected to the server. No way for me to know for sure though.

[jiecut]

They definitely have rollbacks also. It's because the DB crashes and then I guess a backup needs to be used.

One thing you might've noticed is that the event didn't have any rollbacks.

Quote:We will be aiming to release a patch ahead of time tomorrow around 7-8 PM EDT, to include some migration to the new database to alleviate the frequency of these maintenances.

Quote:The servers themselves are running incredibly stable - the database is what is crashing. We are already running a hybrid solution with MongoDB for the event that has been exceptionally stable (hence why the event never rolls back), so we are migrating to that. I was just hoping the frequency of these crashes wouldn't be so much that it would have to become a higher priority yet.