Optimization question

[aziernest]

Hi,

I have a vps with litespeed enterprise license for vps'es over 2 gb of ram(my vps have 3 gb of ram)

When I am running a software to test for broken links on my own website, after sometimes I am getting time out messages

Following are the results from action/server/tuning

PHP Code:

Connection       Edit
Max Connections    




2000
Max SSL Connections    




500
Connection Timeout (secs)    




200
Max Keep-Alive Requests    




1000
Smart Keep-Alive    




Yes
Keep-Alive Timeout (secs)    




10
Send Buffer Size (bytes)    




0
Receive Buffer Size (bytes)    




0
I/O Event Dispatcher    




best (All platforms)
Max Cached Small File Size (bytes)    




8192
Total Small File Cache Size (bytes)    




64M
Max MMAP File Size (bytes)    




512K
Total MMAP Cache Size (bytes)    




64M
Use sendfile()    




Yes
Use Asynchronized disk I/O (AIO)    




Not Set
AIO Block Size    




128K
No I-Node in ETag    




Yes
SSL Hardware Accelerator    




(built-in) OpenSSL internal engine
Request/Response       Edit
Max Request URL Length (bytes)    




8192
Max Request Header Size (bytes)    




16380
Max Request Body Size (bytes)    




500M
Max Dynamic Response Header Size (bytes)    




8K
Max Dynamic Response Body Size (bytes)    




500M
GZIP Compression       Edit
Enable Compression    




Yes
Enable Dynamic Compression    




Yes
Compression Level (Dynamic Content)    




1
Compressible Types    




text/*, application/x-javascript, application/xml, application/javascript
Auto Update Static    




Yes
Static GZIP Cache Directory    




Not Set
Compression Level (Static File)    




1
Max Static File Size (bytes)    




1M
Min Static File Size (bytes)    




300 


my vps have 3 gb of ram and only 1 gb is utilized, any way to allocate more memory to litespeed with more child / worker threads or perhaps caching, I hope memory cache is much faster than disk/page cache so in case of memory cache disk i/o restriction would be lifted and memory i/o would be utilized.

any thought/suggestions for tuning vps/litespeed to entertain maximum requests without giving any timeout error etc???

[webizen]

LSWS caching (page cache) is only available for 2-CPU Enterprise license or above and not for 1 cpu license.

you can use opcode cache to speedup php (compile) and application cache option like Super Cache for WP, etc with shm.

what kind of timeout did you get? any error message in error log?

[aziernest]

No error message in error log, but it appears in client site.

How can I use LSWS to use memory for caching in order to avoid or reduce disk i/o ?

How can I see historical graphs of my litespeed concurrent connection in past to know what's the max concurrent connections that I may have ever reached ?

Awaiting detailed insight / answer.

Thanks

Originally Posted by webizen

LSWS caching (page cache) is only available for 2-CPU Enterprise license or above and not for 1 cpu license.

you can use opcode cache to speedup php (compile) and application cache option like Super Cache for WP, etc with shm.

what kind of timeout did you get? any error message in error log?

[webizen]

Originally Posted by aziernest

No error message in error log, but it appears in client site.

How can I use LSWS to use memory for caching in order to avoid or reduce disk i/o ?

How can I see historical graphs of my litespeed concurrent connection in past to know what's the max concurrent connections that I may have ever reached ?

...

to use memory for caching, as pointed out previously, you can use opcode cache (i.e., APC, xCache, etc for compiled php code) AND application cache like WP super cache, etc with shared memory (/dev/shm) as storage.

for history data, use tools like cacti, rrdtool, awstats

http://www.litespeedtech.com/support...ku.php#add-ons