Search results

Looks like something wrong with LSAPI protocol, PHP size does not like the request received. What your setup? OS? 32bit 64bit? LSWS version? LSAPI version? PHP version?

Please do a force reinstall of 4.0.12 from web console or manually update it, it should have been fixed with latest build.

There are many ways to change umask, you have to find out what works for your system. Google, and try different methods, then run "umask" command from command line to verify that "0002" returns. then you will be good.

There should be one ruby process with parent PID of 1, that's the one started by LSWS, the other worker ruby processes are forked by that process. So, check that ruby process with ktrace/kdump, see if it is doing getppid(), and the result. The configuration change only affect newly started...

Try adding environment variable UMASK=002 to the lsphp5 external app configuration under LSWS web console.

Maybe you can try remove LSAPI_PPID_NO_CHECK, as latest LSAPI uses getppid(), instead of kill( ppid, 0 ) to test if parent process is alive or not, so the security setting in FreeBSD does affect kill(), but may not affect getppid() at all.

That's why those processes do not quit when parent process quit. Need to change the way how graceful shutdown works.

It is possible having something to do with zfs. If you do not experience any problem when switch cache directory to none-zfs partition, it indicates that.

It has something to do with freebsd 7 thread support, not sure relate to freebsd 6 compatible layer or not. we will give it a try in our lab later.

You need to change umask from 022 to 002, update /etc/login.defs

Please check those PHP and Rails process that refuse to quit with ktrace, LSAPI should check ppid every second, if ppid is -1, means that parent process is dead, LSAPI app should quit. I wonder if those LSAPI app stuck some where. I will check the lscgid issue.

Are you using LiteSpeed with Apache httpd.conf? or configure everything natively. If you use httpd.conf, you need to use "Options" directive. otherwise, you need to set the corresponding option at vhost level as well.

Please try a force reinstall from web console to apply the latest build of 4.0.12, add some code to deal with it.

Looks like some thing related to the AIO thread, if you turn off AIO, will it back to normal?

truss output is not helpful at all, too many unknown syscalls. Can you try ktrace/kump?

Please ktrace or truss the process see what is going on. Have you enabled AIO?

You should increase Memory Soft Limit (bytes): 500M Memory Hard Limit (bytes): 600M Process Soft Limit: 800 Process Hard Limit: 1000

http://www.litespeedtech.com/litespeed-web-server-editions.html Yes, you can change to 1-CPU, 2-CPU license without losing any configuration.

due to cPanel 11.25 graceful restart issue, we have changed graceful restart of 4.0.12 to finish all pending requests completely, my guess is that there are still requests in process. It takes longer for the lingering process to quit. Please strace and lsof the running process. if there are...

Seems it is a common problem with xcache if you google it, but no good solution. Maybe you can try other opcode cache.