December 15th, 2003

  • lisa


This morning we had a machine crash that was serving about 25% of our memcache farm. While it going down didn't cause any problems, having all the caches refill when it was back up did cause more load on our databases, making the site slow in the meantime. More information on memcache and how we're addressing the problem so it isn't as noticeable in the future is in lj_backend.
  • lisa

Maintenance tonight

We'll be doing some work tonight on our global database servers starting at 1am PST (-0800 GMT) for about an hour. All or parts of the site may be unavailable while we finish this as quickly as possible. Sorry for the late notice, but the site should be faster afterward.