Several items - announcements, status, grid/db, etc

Announcements/status
For operational announcements it might make more sense to really start using a static /status page which is updated manually. The database issues interfere with the current announcements and with the more major problems over the last few days, it's really unproductive to go to a page that tells us there's a database error. It's handy for the community to know if there is an upgrade, internal problem or griefer so we can better handle our frustrations with any outages. Any information we can get helps to add to that "warm comfy feeling". (This would be perfect for one of those non-techy Lindens to do while the sysadmins, grid monkeys, DBAs and programmers do their thing)

DB/grid..
Regarding the more technical part about fixing DB issues.. When the DB was being worked on earlier, why wasn't the grid closed down the rest of the way? We still had a decent number of people on our land and they were plagued with problems from teleport to map to texture issues to money transfers not working. It's been my experience that databases are easier to fix when removed from normal usage as much as possible. Why limp along for hours when you could (hopefully) take the main DB down and do what's needed to reset things faster?

Reboots/unsits/disconnects
On the subject of resets/reboots/disconnects.. would it be possible to "unsit" everyone before rebooting a sim (or better yet, "try" to teleport them all home when possible)? For single sim or rolling reboots, the TP home could be quite nice 30 seconds before the sim goes down - meanwhile on full grid shutdowns, the stand up would be a nice addition.

Thanks in advance, fun place

First, announcements and status:

I think development does a good job of posting announcements about releases in advance of their deployment. Since 1.9.0, we've needed several rolling updates. I know that the timing hasn't been perfect for those, but effort has been made to keep them as minimally invasive as possible.

The status page, however, is not the optimal method of disseminating information in its current state. Our Web team has begun design on an improved page -- I look forward to its debut in the coming weeks. I know that it will be big improvement on the current implementation!

Second, database and grid issues:

It goes without saying that our goal is to keep the grid up. We did have to put the grid into restricted access mode during the database issues on Friday and Monday, as well as during this weekend's incident. We do this to prevent new users from logging in. We do try to avoid having to kick everyone off when we are dealing with an issue that has the potential of being resolved quickly. Often during our assesment of these types of situations, we analyze whether the situation requires total grid shut down or not.

Our goal this year has been to improve stability of the grid and all of our support systems. We have made great strides here and will continue to work hard at this goal.

Third, reboots, unsits and disconnects:

We have a new tool which gives an automated warning one minute prior to the region going down in a rolling update/restart. This is certainly an improvement over the old method where you would simply get kicked out when the region went down!

To do more than this, however, is problematic. Any action that we take which suddenly moves folks around without their knowledge or permission is likely to cause more issues that it solves. Consider new Residents making their way through Orientation Island, for example. Pulling them from their location suddenly, and without letting them go back to finish learning how to use Second Life would be detrimental to their initial experience. Alerting Residents to the downtime, and then letting them choose the proper course of action is the best decision, I think.