Of course, it's not going to be *that* simple... There seem to be two different symptoms here. One is that the server just stops responding, even though it does not have a huge number of thread going and nothing is visibly wrong. That seems to have gone away with Rob's LD_ASSUME_KERNEL idea.
The other is that the number of threads rises to the max, which is 100, and then people start queueing up and getting the "site busy" message. This seems to happen "all of a sudden". The site's not *that* busy; there are only 59,307 lines in the access log for today so far, so it seems unlikely to me that all 100 threads are actually in use at once. This one is still happening. So, any other ideas would be welcome! janine On 4/10/02 2:08 PM, "Janine Sisk" <[EMAIL PROTECTED]> wrote: > You know, I really think Rob is on to something here. Not only is the > number of threads not increasing so fast, I just saw it go *down* by one for > the first time. And the site has not hiccuped since I made this change, > though it is still being restarted every 30 minutes so that's not quite the > accomplishment it sounds like. > > I will report back again after I've given this a good trial. > > janine > > On 4/10/02 1:00 PM, "Rob Mayoff" <[EMAIL PROTECTED]> wrote: > >> +---------- On Apr 10, Lamar Owen said: >>> 7.2's glibc is so much newer than 6.2's that the linkage may be broken. >>> Binaries linked/compiled under 6.2 usually do _not_ work properly on 7.2. >>> Even with the compat libraries installed. The reason seems to be kernel >>> 2.4. >> >> To use some JVMs on 2.4, you have to set LD_ASSUME_KERNEL=2.2.5. Maybe >> setting it will help with AOLserver 2.3.3. >> > > -- > Janine Sisk > President/CEO > furfly.net, LLC > Mont Vernon, NH > Phone: 603-672-1122 > -- Janine Sisk President/CEO furfly.net, LLC Mont Vernon, NH Phone: 603-672-1122
