frassle/downtime


Sorry for the extended (and extended, and extended) downtime.

Frassle is back, and voo2do is on a different server now so Frassle's overly slow design won't impact that (much more popular) application.

Now to make with the blogging…

The new frassle server crashed last night, and was rebooted and restored this morning. The machine ran out of memory. I believe I have made a configuration change that should prevent the problem, but it's possible there is a memory leak somewhere on the system—probably not in frassle itself, but in one of the underlying components or in the operating system. We'll be keeping a close watch on it, but please let me know by email to shimon@rura.org or AIM shimonrura if things seem broken.

I've had scant internet access lately (experiencing severe withdrawal) and thus was unable to notice/fix the frassle server when it crashed—looks like it was down about 2 days. Clearly I need to automate this part of the sysadmin experience. For now, I am sitting in the Zurich airport at my gate and will be boarding my plane to Boston in a few minutes. If you're one of those Bostonish types, see you soon!

Frassle has been getting some more usage recently, both in terms of web hits and subscriptions being fetched by frassle's aggregator. The latter has been causing some trouble, as the aggregator's RSS fetcher is rather inefficient. I'm working behind the scenes to (1) keep things from blowing up and (2) actually fix things for the long term. In the mean time, frassle may have occassional spots of downtime. Feel free to IM me (shimonrura) if you notice one so I can fix it ASAP, or perhaps just tell you I'm working on it.

There have been a few outages in frassle service this afternoon. I hope things have straightened out. If you find frassle not working, contact me on IM at shimonrura or by phone or email.