So my routing issues to my old box are still in force. Sporadic, inconsistent, the whole mess. Their NetOps team seems to think it's something on my box, but I don't see how it could be my box if certain routes always work, while the offending route fails sometimes. I've got a single NIC, a single network cable, a single IP and an IPTABLES configuration that is devoid of any rand() calls.
So I'm thinking my grand plan, which was to carefully migrate my apps from old server to new, one at a time, and clean up some of the cruft that has build up over the last couple years might go flying out the window. Now I'm thinking I'm going to turn off all the apps on the old box, do an rsync of everything, repoint DNS, and fight through anything that doesn't work as expected. Hardly elegant, but it'll at least get me to a box that is consistently accessable from the internet.
It also would alleviate the need deal with my WordPress quandry. After a weekend of reflection, the right solution seems to be patching WordPress so that it's well behaved (or at least better behaved), and then supply the patch back to Automattic and hope for the best. I'm a competent developer, and while I'd like to just use it out of the box, the benefit of open source is exactly this option, and as an open source proponent, I feel at least minimally obligated to leverage that option on principle alone.
That said, I don't know if I can deal with an only partially accessible box for much longer. I'm going to the symphony tonight, so I'm not going to be doing any admin work until at least tomorrow, but if the routing issues aren't resolved by then, I think I'm just going to punt. I hate punting.
Comments are closed.