Memory leak and high cpu ?

#1
Hi,

I am have a server running DirectAdmin + Cloudlinux + OpenLiteSpeed.
Installed 1 simple wordpress installation and load tested it. 50 concurrent users consumes more than 8GB PMemory.
Tried with and without cache. Same issue. tried another test installation. Same issue. In processes i can see LSPHP creating many threads and consume alot of cpu and memory.

Tried with 1.4.x and 1.5.3 - Same behaviour.
 
#4
Hi David,

indeed, i was in contact with DA - replacing the version to 1.5.3 did help alot, but there is still a very high load compared to normal usage.

Running this now:

LiteSpeed/1.5.3 Open
module versions:
modpagespeed 2.2-1.11.33.4
cache 1.61
modinspector 1.1
uploadprogress LSIAPI_VERSION_STRING
mod_security 1.1
(built: Mon Jul 15 18:57:15 UTC 2019)
module versions:
modpagespeed 2.2-1.11.33.4
cache 1.61
modinspector 1.1
uploadprogress LSIAPI_VERSION_STRING
mod_security 1.1



My setup is: CL7 + DA + OLS. The load is 10x more than normal.

E.x i migrated a wordpress site from cPanel running latest litespeed + CL (dont know if it was enterprise or OLS)
The memory usage for that site/user was normally 80-90MB in average

As soon as i migrated it to DA, i saw it raise to 1GB - 10x more than normal behaviour. Thats not isolated to that particular site, but in general all sites. I have tried with/without cache. Simple WP to advance WP. Same problem on all of them.

I suspect a bug in OLS/CL.

Can you give you access to server as well, so you can have a look.
 
Last edited:
#6
CL7 + DA + OLS same here, i also face this and when set enterprise trial it goes down, as well nginx, look like OLS has something to tune
 
#10
Yes, we know the reason.
You may set `httpdworkers 2` in the server conf file and set less php CHILDREN(such as 5) in the ENV setting, also you can set less `max connection` .
We will release 1.5.5-2 soon, which should save a lot resource.
 
Top