Search results

I am out of ideas, looks like core dump is allowed. You can try making some changes, then sending SIGBUS signal to a child litespeed process, see if it can dump the core file. usually core file should be in /tmp/lshttpd.

The problem is completely in PHP side, It looks like a problem with PHP code, doing infinity looping somewhere. Have you made changes to your site recently? upgrade? It will cause some delay, but it should recover as LiteSpeed try another instance of PHP.

for core dump, please check sysctl kern.sugid_coredump kern.coredump kern.corefile and "ulimit -c" We will investigate the freebsd restart issue.

It should be removed from 4.0.x releases.

Currently, the Per VHost Output bandwidth statistics are not implemented. Should remove it for now, maybe add it later.

That's a issue between PHP and MySQL, not really something to do with LSWS. Check your MySQL query with "mysqladmin processlist", limit memory usage of PHP.

Are you using a Cloud Linux kernel?

Only 4.1RC2 or 4.1RC3 support AIO. check the 4.1RC discussion on the forum. You can try "strace -tt -T -p <pid>" litespeed processes, see what it is doing. I think it is stalled because high I/O wait. AIO will work better.

Maybe you can use persistent connections between LB and backend. It will be faster as well. I do not need the strace, it is difficult to catch the problem now I think.

Install it to /usr/local instead of overwrite the system default installed uner /usr, by default, autoconf will be installed to /usr/local

what edition of LSWS are yous using? standard, enterprise, vps?

You need to increase the memory limits of lsphp5 external app from LSWS web console as well.

For running PHP without suEXEC, you can set cache memory size high as one group PHP processes only use one copy of cache. It is up to you and the available memory of your server.

Please download 1.7 release and update again, I am not sure it is something wrong with interpreting getsockopt() results or not. Also strace need to be updated to the latest release to get correct result value for getsockopt() on 64bit Linux.

Yes, that's correct.

Are you using 32bit binary on 64bit OS? I wonder if it is an issue with the 32bit-64bit compatible layer.

should be your build tool chain, try installing autoconf 2.13

Please regenerate strace output with option "-tt -T", need timestamps to figure something out. The strace output shows some errors. Need timestamp in the output to further analyze it. Are you still get packet loss? it is a big problem if you get packet loss over LAN connection.

Those are not log entries produced by LiteSpeed, but Apache.

Have you notice anything abnormal when LiteSpeed is slow? high CPU utilization? high load? memory usage? If you run LiteSpeed on another port parallel to Apache, the page load is low on another port?