If you’ve read a couple of previous posts on this subject, or attempted to access the Columbia Missourian Web site recently, you might have noticed it’s been a bit quirky.
So, to sum up:
Monday, we got hit with a large load of traffic, mostly from the Drudge Report, that knocked both our Web servers offline. Thecampus load balancer (not controlled by us) apparently stopped searching for either of our web servers once both went down, so when the servers came back up and were ready to go, it wouldn’t allow either to serve up content. A frustrating day ensued … everything was running fine as far as we could tell, but we couldn’t track it down to the load balancer until close to 4 p.m. That’s now been fixed. Monday traffic was close to 29,000 visitors in 8 hours of uptime (a normal 24-hour day has maybe 10,000-11,000).
Tuesday, more traffic and more slowness. Our programmer spent most of the day restructuring our database to help with load times, tweaking caches, and the normal sort of things you do. Down to about 18,500 visitors.
Wednesday, we spent a couple hours in the late afternoon testing a new host location for the Missourian’s mySQL database, with help from a mySQL expert (neither my expertise nor our programmer’s). While the site was offline and the new database was being tested, our programmer made a few tweaks to firewall and cache settings and did some upgrades to help with speed. Only about 7,500 visitors made it through.
Today, we moved the SQL database to a new location that lets us basically store the whole database in RAM, which speeds up the site considerably. Hoping that will lead to visitor numbers coming back to what they used to be … and that we can get a couple of lingering Apache programming issues dealt with. More updates TK.
Recent Comments