Skotos came back up ~10.45 after we replaced the power supply in our user database machine. Today we also replaced a power supply in our mail machine and upgraded the memory in the machine hosting TEC and RPGnet's read-only MySQL. Whew!

Skotos is currently down due to the power supply in our user auth machine dying. We're working on it.

I'll be doing some work on RPGnet's mail server and one of its MySQL servers on Thursday, from ~1pm PT to ~3pm PT. The site should keep running in the interim, but it will likely be slow due to the loss of one MySQL server and mail will not be delivered during that time.

I am planning to take down the TEC machine starting around 1.45pm PT on Thursday afternoon and hopefully completing by 3pm PT or so. This is to upgrade its memory, which will hopefully improve how well the machine works. I'm not expecting any troubles, but you can never tell when you take a machine, actually powering it off (as I will to swap out the memory).

We should have a working security certificate again.

We are aware that the www.skotos.net certificate has expired. We've placed an order for a replacement, but the company has to check we are who we say we are and that good stuff. We suggest you tell your browser to visit the site despite the lack of valid certificate, and hope to have it back tomorrow.

GR should be back up now. Downtime was limited to 30 minutes. As planned, there was a little bit of rollback to the afternoon backup. We *hope* this'll help GR run more efficiently (fingers crossed).

If all goes well, we are planning to take Grendel's Revenge down this afternoon around 2.30pm PT for some maintenance. Exact downtime is unknown, but it shouldn't be more than a couple of hours.

We're rebuilding the post table for increased efficiency. Hoped it would finish over night, but not yet!

Marrach did see some extend lag today around 5pm, then after 11pm. This appears to be the direct result of some other work we were doing on the machine. I wouldn't *expect* it to lag Marrach (which is on a different virtual machine), but clearly it did. Tomorrow I'm going to add some monitoring software to make sure there aren't any hidden hardware issues that aren't showing up in our logs. If it's not that, I'm going to make sure to do the final required work in late hours ...