litespeed slower than apache?

furimedia

Well-Known Member
#21
Hey anewday

Basically Mistwang just went ahead to my system and optimized it himself out of his courtesy, so I cant really say much in terms of what he has changed.

But I did notice a bit of an improvement and am pretty satisfied with it.

One thing that stick to my head was the priority level was set to -19, i guess that is the highest cpu priority on the LS. I think that is what did the trick?
 

mistwang

LiteSpeed Staff
#22
What happens when binding is "Not Set"? I seem to have 4 or 5 lshttpd processes (4 cores) so the default is "bind all"?
Yes, that's correct when you create a listener from the web console.
For listener configured via Apache httpd.conf only one process is assigned, you need to override that by explicitly creating a listener from web console.
 

mistwang

LiteSpeed Staff
#23
Hey anewday

Basically Mistwang just went ahead to my system and optimized it himself out of his courtesy, so I cant really say much in terms of what he has changed.

But I did notice a bit of an improvement and am pretty satisfied with it.

One thing that stick to my head was the priority level was set to -19, i guess that is the highest cpu priority on the LS. I think that is what did the trick?
Yes, that's correct. his server is some what disk I/O bound, for linux kernel 2.6.13 with CFQ io scheduler, higher process priority will give higher disk I/O priority as well, please read the man page for more detail.

man ionice
 

mistwang

LiteSpeed Staff
#24
Hmm lets say that I have an 8-core cpu and license.

Does that mean that it'd be much better if I create 8 listeners with 1 cpu binded? rather than having a single listener bound with 8 CPU binded?
It is really up to you and how the traffic distributed among listeners, LSWS give you this flexibility to do either way.

Use 8 listeners is more flexible, like if you want to assign one large site to 5 processes, another site to other 3 processes. It is really up to you how to allocate processes to listeners and how many listeners should be used.
 

furimedia

Well-Known Member
#25
Hii Mistwang thanks for the reply.

Got a couple questions.

So by default, if i have no listeners, does that mean I have just one default listener running? If so, is it bound to 8-cpu by default?


And what is the benefit of having 8 listeners as opposed to 1 listener?
Is this for achieving much better concurrent connections?

If I have lots of virtual hosts, which do you think is recommended?
 

anewday

Well-Known Member
#26
Yes, that's correct. his server is some what disk I/O bound, for linux kernel 2.6.13 with CFQ io scheduler, higher process priority will give higher disk I/O priority as well, please read the man page for more detail.

man ionice
I think I may benefit from setting higher priority for the lshttpd process, what do you think? It's set to epoll for Event I/O Dispatcher since I'm using 2.6 kernel. :)
 

IrPr

Well-Known Member
#27
I have also tried to enable xcache on litespeed and notice that its not caching anything.
Only thing it caches is when I try to access the xcache admin page, only that page gets cached.

Is there some remedy to this?

Here is my xcache settings:

[xcache]
xcache.shm_scheme = "mmap"
xcache.size = 64M
xcache.count = 4
xcache.slots = 8K
xcache.ttl = 0
xcache.gc_interval = 0
xcache.var_size = 32M
xcache.var_count = 4
xcache.var_slots = 8K
xcache.var_ttl = 0
xcache.var_maxttl = 0
xcache.var_gc_interval = 300

I've tried xcache.count to 1 through 4 and they all do not seem to do the trick.

As for the softlimit/hardlimit on litespeed, i have set them to 512MB.

And also, I have placed my copy of php.ini to /usr/local/lsws/lsphp5/lib/php.ini , is this the correct standard way?

And I did make the change so that PHP suEXEC Max Conn is 200.

I was hoping for some performance gains once I move to litespeed but it seems I'm experiencing everything oposite of what litespeed advertises on performance.
Just want to note that Opcode cachers doent work properly with Suexec
 

mistwang

LiteSpeed Staff
#28
Yeah, some opcode cache is not suexec friendly as it requires all PHP processes forked off one parent process.

eaccelerator may be better as it uses a on disk cache which can be shared by all PHP processes.
 
Top