Monthly Archives: July 2008
Some trickery or resilience with Varnish
If you’ve got two or more backends, and under some condition can’t or won’t serve a request immediately or want to send it elsewhere depending on some circumstance, you can do this using HTTP return code or header with the not-so-well-documented feature ‘restart’ Continue reading
LVM with dmraid
hen adding a new disk for a customer running CentOS 4.7 on severely old hardware, I bumped into something I’ve never had happening to me before. Basically the system wouldn’t let me create the Physical Volume Continue reading
Resyncing slaves with slaves
When dealing with replicated setups with two or more slaves sharing a master, it appears as if a lot of people overlook the obvious. You don’t need to take your master down to resync a slave. I was hoping I wouldn’t need to post about this, but I see people taking down their masters when they have perfectly healthy slaves way too often to let it slip.
Continue reading