We've taken a step back with this release in order to consolidate and simplify the virtual memory interface. The REST API and
vmprobed packages weren't as flexible as we were hoping so they've been removed for now. We've got another API-like interface in the works — it's going to be really cool!
By intelligently managing the filesystem cache, we can speed up certain database operations dramatically. This white-paper demonstrates this as well as discussing how it works and some near-future plans for vmprobe.
One application of vmprobe technology that people have been interested in is using it to maintain hot-standby servers. By using vmprobe to duplicate the virtual memory state from the primary to the standby(s), fail-over events are much smoother than otherwise.
We're extremely excited about this release! We've split vmprobe into two components: the command-line tool you may already be using, and a daemon component with a REST API so you can integrate vmprobe into your own infrastructure.
This screencast demonstrates the features of vmprobe's web interface.
We've started screencasting on Youtube! In this screencast we walk through basic usage of the
vmprobe command-line utility.
In April 2009 I wrote up a simple utility over a weekend called
vmtouch. It helped me solve a few problems I was experiencing on the job and, most pressingly, let me run a series of experiments to satisfy my curiosity about how the filesystem page cache works. I have found
vmtouch very useful over the past 7 or so years, and apparently a few other people have too.