Re: apache-mpm-prefork debian perfomance tuning
Thanks a million Roberto for the Quick Reply :-)
On 3/28/07, Roberto C. Sánchez <roberto@connexer.com> wrote:
>
The hardware specs look good. How much bandwidth will the server have
going to it? I mean is your ISP connection a dedicated T1, T3, or
something else?
T1 burstable they say. That is more bandwidth will be automatically
alloted if needed.
> So I hope I should be Installing ( right? )
>
> kernel-image-2.6.8-12-amd64-k8-smp - Linux kernel image for version
> 2.6.8 on AMD64 SMP systems
>
> to harness the full power of the CPU as currently it shows only
>
You want to be using Etch.
May I Know why Is ther any problem using Sarge?
isn't that more stable.
I never used Etch for Production Servers.
I already upgraded one Server to the Sarge SMP kernel.
Should I be using Etch Instead?
=========================================================================
# uname -a
Linux websrv-1 2.6.8-12-amd64-k8-smp #1 SMP Thu Dec 7 18:44:52 UTC
2006 x86_64 GNU/Linux
# cat /proc/cpuinfo
processor : 0
vendor_id : AuthenticAMD
cpu family : 15
model : 65
model name : Dual-Core AMD Opteron(tm) Processor 2210
stepping : 2
cpu MHz : 1795.568
cache size : 1024 KB
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge
mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext lm
3dnowext 3dnow pni cmpxc
hg16b
bogomips : 3547.13
TLB size : 1088 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts fid vid ttp [4] [5]
processor : 1
vendor_id : AuthenticAMD
cpu family : 15
model : 65
model name : Dual-Core AMD Opteron(tm) Processor 2210
stepping : 2
cpu MHz : 1795.568
cache size : 1024 KB
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge
mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext lm
3dnowext 3dnow pni cmpxc
hg16b
bogomips : 3588.09
TLB size : 1088 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 40 bits physical, 48 bits virtual
power management: ts fid vid ttp [4] [5]
# mpstat -P ALL
Linux 2.6.8-12-amd64-k8-smp (tutorom-dbsrv-1) 03/28/2007
04:53:44 AM CPU %user %nice %system %iowait %irq %soft
%idle intr/s
04:53:44 AM all 0.78 0.00 1.96 3.31 0.00 0.02
93.93 1039.80
04:53:44 AM 0 0.90 0.00 3.40 5.88 0.01 0.03
89.77 1039.77
04:53:44 AM 1 0.65 0.00 0.52 0.75 0.00 0.00
97.99 0.00
top - 04:50:36 up 1 min, 1 user, load average: 0.23, 0.11, 0.04
Tasks: 53 total, 2 running, 51 sleeping, 0 stopped, 0 zombie
Cpu0 : 0.0% us, 0.0% sy, 0.0% ni, 100.0% id, 0.0% wa, 0.0% hi, 0.0% si
Cpu1 : 0.0% us, 0.0% sy, 0.0% ni, 100.0% id, 0.0% wa, 0.0% hi, 0.0% si
Mem: 4102044k total, 111748k used, 3990296k free, 10328k buffers
Swap: 2931852k total, 0k used, 2931852k free, 38700k cached
========================================================================
Does that look good enough?
I am a bit apprehensive about using Etch in Production :-(
Also the freequent kernel upgrades to Etch would mean freequent reboots right?
>
> Just wondering mdadm can monitor this hardware Raid Controller? is it
> supported?
>
I think you need the monitoring tools from 3ware. However, I am not
certain. In any case, mdadm is for the Linux software RAID
implementations.
yes yes forgot that :-)
> I hope to Install LVM and have seperate partitions for
> "/var/log" and "/var/www" so I can adjust them if there is any need in
> future.
>
Put everything on LVM. That allows you to create separate partitions as
necessary.
http://tldp.org/HOWTO/LVM-HOWTO/benefitsoflvmsmall.html
says
===========================================
"root on LV should be used by advanced users only
root on LVM requires an initrd image that activates the root LV. If a
kernel is upgraded without building the necessary initrd image, that
kernel will be unbootable. Newer distributions support lvm in their
mkinitrd scripts as well as their packaged initrd images, so this
becomes less of an issue over time. "
=======================================================================
Is it outdated info?
or shall I put
/boot on an Ext3
and put the rest in LVM on XFS
> What do you suggest use up the whole 500 GB disks or leave some free
> Space for later use?
>
I would say, start with the partitions as small as you can. Use XFS and
grow the partitions as necessary.
Thankyou so much :-)
hope Linux Can boot From XFS and kernel upgrades will go without probs.
and also handle kernel upgrades will go smoothly.
If you have any choice in the matter, use Python or Perl instead of PHP
so that you can use apache's worker MPM, which by all accounts should
perform lots better.
nope :-( the code is already in PHP.
At the risk of starting a flameware, please don't use MySQL. That is,
unless you are a super experienced expert. It is well known that
MySQL's performance drops very quickly as the number of connections
increases. This can be mitigated with clustering and performance tuning
(for example, slashdot uses a MySQL cluster and they have been
performance tuning it for like 8 years). However, since you have only
one server, you really want a database that will scale well in one
instance. For that, I can recommend PostgreSQL
Really?
I heard postgresql is slower than mysql. not true?
any way the code is written for php5, mysql5 so no choice :-(
Thanks a million :-))))))))))))))))))))))))))))
Kind Regards
Siju
Reply to: