[UPDATE 17:45 p.m. PST –teeple]
The primary upgrade process is finished. We’re still working on getting transaction histories re-enabled.
[UPDATE 15:49 p.m. PST –teeple]
The deployment team is re-applying the server code upgrade.
The sole culprit from the earlier attempt was a new database query introduced into the resident login sequence….one which performed charmingly in QA, but which showed serious antisocial tendencies once moved into production.
That query has been replaced and the code upgrade is underway again. Asset operations may be slow during the upgrade, but they should be reliable.
[UPDATE 14:16 p.m. PST –teeple]
Rez and transfer operations should be safe, and the Friends Online service will be turned back on momentarily. We’re still working on getting transaction history re-enabled; we’ll get that up ASAP.
[UPDATE 13:03 p.m. PST –teeple]
My abject apologies; I should’ve been clearer. “Partial rollback” is an unsettling term. We’re rolling back the server code. We’re not rolling back regions.
[UPDATE 12:46 p.m. PST –teeple]
Transaction history will be off line a bit longer than anticipated this afternoon. Transactions aren’t being lost; they’re just currently unavailable for viewing. In addition, we’re taking Friends Online out of service until we’ve stabilized the grid.
[UPDATE 12:35 p.m. PST –teeple]
Logins are slow or failing, inventory is problematic. We’re doing a partial rollback of the server code to stabilize as quickly as possible. Please do not rez or transfer valuable (no-copy) assets in world until we give an all clear. We’ll be broadcasting this message to the grid very shortly.
[UPDATE 11:56 a.m. PST –teeple]
As part of the backend upgrade, we’re about to temporarily disable account transaction history. We expect to have it back within fifteen minutes.
[UPDATE 11:55 a.m. PST –teeple]
Work proceeds apace. As part of the code upgrade, we’re also doing some tweaks to the login servers. If you’re using the Windlight viewer, you may notice that the login panel has temporarily relocated to the top of your screen. It’s functional in its temporary location, and it’ll be back in its normal location shortly.
On Tuesday, January 29, (10:30 AM – 12:00 PM) we will be upgrading the code running on our central second life servers. This will mainly add stability fixes and set the stage for upcoming new features. It will also include fixes for the following issues:
SVC-930: Prims set for sale – prices are incorrectly set when multiple prims taken to inventory and rezzed
SVC-1039: llResetScript() from a state without a touch_start handler does not re-enable touch in default state
We do not expect any service interruption during the upgrade, and we will be following it up with a rolling restart of most Second Life regions on Wednesday, January 30th. Restarts are expected to begin around 2:30 PM and take approximately 5 hours.
Watch this space for progress reports