-- new memory is due to arrive tomorrow. but in portland... we're in seattle. read more. we're trying to find somebody to drive it up tomorrow evening. new memory will help the db slaves out a lot.
-- we've been in contact with monty, the primary mysql author ... I noticed the slave dbs didn't seem to be doing something that the master was (replicating explicit locks). I had evan write to them (since livejournal.com has a support contract with them... good investment!) and he replied, saying that it does not, and asked if we'd benefit if it did. I explained excitedly why it would help us out a lot (bin/qbufferd). If he adds support for it we'll get much better performance on the slaves.
The new memory could help as early as 20 hours from now.
The MySQL change could happen in 10 minutes or never. The MySQL team never fails to impress me with how fast they can add or fix things, though, so I remain very optimistic.
I'm going to bed relaxed. :-)