• Shard revert:
    The one thing a shard administrator never wants to announce is a shard revert. This one was a full day, so it's especially disappointing.
    Details: Yesterday, we had lots of freeze-up issues. These freeze-ups were caused by a race condition involving our login database. Essentially, multiple servers were 'racing' to access it, and when there is a collision, bad things happen.
    To eliminate the 'race condition', I suspended shard Saves until a fix could be put in place. It was my intention to manually save the shard before I put the fix in.
    I have my server configured to only install updates when I say so. Unfortunately, I had the dialog box open that shows the pending update and somehow Server 2019 viewed this as an assent to go forward with the update installation. Before I could stop it, It had already shut down the server, and the last day's progress was lost.
    I feel bad about this, and I am sorry for all your lost accomplishments.
  • New shared database model:
    In order to mitigate yesterday's freeze-ups, we reworked the logic for database access and believe this solves the race condition.
    A side benefit of this work is that the account saves are an order of magnitude faster.
    In addition, we decoupled account saves from world saves. This gives us developers a much clearer view into overall shard performance.
  • Reduce magic weapon and armor drops in treasure chests – should help crafters and bolster the economy.
    The effect won't be noticed overnight, but will slowly improve the economy as the surplus of dropped magic weapons and armor wane.
  • [internal] new jail gumps, and a gump for GMs to view player tags/notes.
  • [internal] World saves and ocean cleanup profiling.
  • [internal] Z-order exploit improvement.