Dethdlr The latest from Rothgar on the merge:
Oasis should be unlocking now.
8 hours ahead of schedule isn’t bad!
And we have confirmation from zerigo that Oasis is back online now.
As for Guk, Rothgar had this to say:
Guk is up and unlocked! Have fun!
But wait, not so fast… After some reports of items being duplicated, Rothgar posted the following…
We are bringing Guk back down briefly to look into an item issue. As soon as I know more, I’ll post an update here.
I’ll try and keep you updated but I may be unavailable for the next few hours.
This just in from Zoltaroth at 6:21pm PST:
We have identified the issue with Guk. The fix is being applied. Current ETA for the fix to finish is about 2 hours, but that is just a ballpark estimate.
And this in from Rothgar at 7:43pm PST:
We’ve unlocked Guk again. The inventory issue should be resolved.
You may still have some items in overflow.
Keep your fingers crossed that this is it and the latest mergers are complete. 🙂
You missed the real gotcha however. Further down the same thread Rothgar admits the lag issue on the merged servers was because the HARDWARE wasn’t actually merged and will take up to 2 more days.
Rothgar has been blowing smoke all day long. What kind of merger of servers doesn’t merge the hardware? How can they announce the merger is complete when it’s NOT?
What were they doing with the hardware for the 24 hour downtime? Adding hardware to a server farm has been captive art for 15 years, and the “new” configuration should have been ready well before the database merge was done.
This is make believe.
@Glenolas: Here’s basically what they did:
Step 1: Merge the two sets of servers into one. This would be the process of running scripts to copy everything off the servers that are going away, onto the servers being merged into. This is a software process. This takes time. This is what the downtime was for, not hardware re-configurations.
Step 2: After QA, bring the servers back online and give it a day or two to make sure everything is still working and nobody finds any gotchas.
Step 3: Reconfigure the hardware/software of the servers that were turned off as additional hardware for the merged server. Again, this takes time.
Doing it any other way would have either been risky or required additional hardware that wasn’t already being used to be in place ready to go, post merge. They didn’t have extra hardware, only the hardware that was already being used.
It’s exactly how I would do it under the same circumstances. Leave the old servers alone for a few days so that if anything goes wrong, you have them available to compare/copy from immediately without having to restore a backup, etc. Otherwise, you’re risking additional downtime if anything goes wrong (like item duplication that happened with the Befallen/Oasis merge).
Also, it takes time to reconfigure the hardware/software of all the servers that got turned off to be used in the new merged server. Would you have preferred they leave the server off for a few days until they have all the hardware/software reconfigured?
Say it’s make believe all you want, but the way I see it, they did these just like they should have (other than the snafu about not letting people know chat was going down for the first one) and actually pulled it off better than most people expected.
Glenolas
“You missed the real gotcha however. Further down the same thread Rothgar admits the lag issue on the merged servers was because the HARDWARE wasn’t actually merged and will take up to 2 more days. ”
They are going to skip this bit entirely for Runnyeye/Splitpaw 🙁