Does anybody around here know any good Dtrace scripts that would be useful for troubleshooting a problem like what this guy is having? Maybe something from the Dtrace toolkit? I'm a Dtrace noob myself, so I personally would't know what to do.
It might be a defective network card if it only has problems when it's under a really heavy network load. I've had some similar issues with running Red Hat Enterprise / CentOS with some of these newer Intel Gigabit NICs that were defective. The most common explanation for this kind of occurrence (since it's in a remote data center) might be that your bandwidth is capped (say at a speed of 10 megabits per second, for example) and that when your server gets so many connections that it starts to exceed the bandwidth that you're capped at (maybe a DDOS attack?) everything will start to lag and your remote sessions will start to hang. Check out your MRTG bandwidth graphs if you have any from your host and see if you're hitting the speed that your bandwidth is capped at when the server starts to hang (usually your host will cap you at 10 megabits per second, sometimes 100 megabits per second, sometimes something really small like 2 megabits per second). Another possible problem is that the company that you are colocating your server with is having network issues. -- This message posted from opensolaris.org
