New server, load ..

R

Rprp

Verified User
Joined
Dec 25, 2006
Messages
28
Dear,

First of all, hi!

Recently i bought a new server, but the load is VERY high, and i am sure its a better one.
This server is running one website.

Old server specs:
- Pentium 4 D 3ghz
- 2GB of ram

New server specs:
- Intel(R) Xeon(R) CPU X3210 @ 2.13GHz
- 2GB of ram

At the old server, the load is 1-2 and the webpages are loading fast, at the new server the webpages are loading slow and the load is high (7-10)

The only change that made is, i upgraded from MySQL 4 to 5.

So, i am asking for advice..

Should i downgrade back to MySQL 4?

Apache requests p/s (Peak): 180
MySQL queries p/s (Peak): ~ 650.

HTTPD Settings:
Code: 
MinSpareServers 5
MaxSpareServers 20
StartServers 5
MaxClients 1024
MaxRequestsPerChild 0

KeepAlive is also off.

MySQL settings:
Code: 
[mysqld]
local-infile=0
max_connections=1000
thread_concurrency=8
server-id=1

Yes i know the MySQL settings suck, thats why i am asking for some advice..

Screenshot of top in peak hours: http://img513.imageshack.us/img513/8414/schermafdruk1ac3.png

Sorry for my bad English, hope u can understand me

Greetings.
 
Last edited:
Old server specs:
- Pentium 4 D 3ghz
- 2GB of ram

New server specs:
- Intel(R) Xeon(R) CPU X3210 @ 2.13GHz
- 2GB of ram
Click to expand...

The new server is less than the old one. The old one was 3Ghz and the new one is 2.13 Ghz. The RAM is the same.
 
floyd, frequency is not the only one parameter (yes, it is very important...), but there are more things.. As CPU cores, L3 (L2) cache (which is very important for webservers) etc. And in my opinion Intel(R) Xeon(R) CPU X3210 is a better choise.
 
We were only told what the frequency was so I had nothing else to compare.
 
Well, a cat /proc/cpuinfo here:

Old server (2 cores)
Code: 
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 15
model           : 6
model name      : Intel(R) Pentium(R) D CPU 3.00GHz
stepping        : 5
cpu MHz         : 3000.234
cache size      : 2048 KB
physical id     : 0
siblings        : 2
core id         : 0
cpu cores       : 2
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 6
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx lm constant_tsc pebs bts sync_rdtsc pni monitor ds_cpl est cid cx16 xtpr lahf_lm
bogomips        : 6005.75
clflush size    : 64

New server (4 cores)
Code: 
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 6
model           : 15
model name      : Intel(R) Xeon(R) CPU           X3210  @ 2.13GHz
stepping        : 11
cpu MHz         : 2133.513
cache size      : 4096 KB
physical id     : 0
siblings        : 4
core id         : 0
cpu cores       : 4
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 10
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe nx lm constant_tsc arch_perfmon pebs bts pni monitor ds_cpl vmx est tm2 ssse3 cx16 xtpr lahf_lm
bogomips        : 4270.04
clflush size    : 64

So i suppose server 2 is alot better.
 
What OS is that? It doesnt look like hyperthreading is enabled?
 
Both are Debian, and i just posted one core instead of them all, because its all the same info.
 
smtalk said:
Yes, I fully understand you :) And sorry if I offended you.
Click to expand...

No offense taken at all. I just don't have time to look stuff up and I don't have the specs memorized. I think the new server is better too since it has double the cache.

I think it is important to find out why one httpd process is using 85% and one mysqld is also using 85% of the cpu.

Maybe database needs to be optimized or something.
 
I suggest using "htop" which has a better understanding of multicore/multicpu systems to see if the CPU work is well balanced between cores.
Use the last kernel stable release available and make sure that all SMP features are turned on (if it's not the precompiled debian-patched SMP kernel, which is ok).

Check also with hdparm -tT on every disk that the bottleneck isn't the disk transfer and access speed, which is often the case.
 
The load is balanced between the cores.

hdparm:
Code: 
/dev/sda:
 Timing cached reads:   5818 MB in  2.00 seconds = 2912.66 MB/sec
 Timing buffered disk reads:  190 MB in  3.02 seconds =  62.93 MB/sec

And @ Floyd, it was oke at the old server.. (Database optimization etc)

And:
~/server-status statics
CPU Usage: u13.89 s2.77 cu0 cs0 - 8.17% CPU load
148 requests/sec - 1.1 MB/second - 7.8 kB/request
101 requests currently being processed, 0 idle servers
 
Last edited:
You must log in or register to reply here.
Back
Top