Greg Ames wrote: >as I mentioned in another post, 2.0.29 also seems to have the high load average >problem. The run queue size numbers displayed by vmstat weren't quite as >extreme as what I saw yesterday pm, which may be because there's less traffic in >the evening. But then the load averages displayed by top gradually built up >until they were much higher than what we see with 2_0_28. > >anyway, there's a piece of the access log with the cpu counter at the end of >each record at http://apache.org/~gregames/log.29.sorted . This is sorted in >descending order of cpu usage, with the record for the physically biggest file >appended to the front, and everthing that used less than 3 ticks eliminated. >The biggest file looks a lot like last time, 9.8MB which took 26 cpu ticks to >serve. The biggest cpu hog was a 1283 byte html file which took 122 cpu ticks >to serve (nearly a second). I don't see any pattern to the requests that are >the big cpu users. >
If you pick one of the URLs with a high CPU utilization from this log like http://jakarta.apache.org/ant/manual/CoreTasks/echo.html , and send that server (or a copy of the same config on another port) a ton of requests for that same URL, one at a time, do they all use a similarly large amount of CPU time? If so, then we'll have a nice, repeatable case to debug. --Brian
