Optimizing Linux Servers

Optimizing Linux Servers

Davor Guttierrez
dguttierrez@me.com
3 Gen d.o.o.

Agenda

● What is optimization?
● Performace (software / hardware)
● Server optimization
● Performace monitoring
● System Monitoring Tools
● Benchmark Tools

What is optimization?

● Our server is slow
● We have new very expensive server but ...
● We have new Linux distribution but ...

● What is slow in your server?
● Too many services running, ...
● Disk, I/O, ...

Performance

To boost performance of a server, we need
just both its hardware and software
components to make it operate efficiently.

Server optimization

Optimization can include fine tuning of
● web servers (Apache, lighttpd, nginx etc),
● disk I/O, block devices, RAID or different filesystems (including SCSI and SSD devices),
● kernel,
● network I/O, TCP/IP network stack,
● firewall etc.

as well as
● databases optimization – benchmarking and profiling, finding bottlenecks, settings optimization;
● data storages tuning;
● disk and memory usage optimization

Start here … with installation

● Always make custom installation of server, don't use default
settings and default install
● Custom partitioning (your life will be easier)
● Install only needed packages, never install whole group of
packages – make minimal installation and then add
packages
● You don't need X Window and GNOME on your server
● Console is 40x25 characters long and not 1024x768

Performance monitoring

● Linux system administrators should be proficient in Linux
performance monitoring and tuning.To identify system
bottlenecks and come up with solutions to fix it, you should
understand how various components of Linux works.
● On a very high level, following are the four subsystems that
needs to be monitored:
– CPU
– Network
– I/O
– Memory

CPU

● You should understand the four critical performance
metrics for CPU

– context switch,
– run queue,
– cpu utilization,
– load average.

Context Switch

● When CPU switches from one process (or thread) to another, it
is called as context switch
● When a process switch happens, kernel stores the current
state of the CPU (of a process or thread) in the memory
● Kernel also retrieves the previously stored state (of a process
or thread) from the memory and puts it in the CPU
● Context switching is very essential for multitasking of the CPU
● A higher level of context switching can cause performance
issues

Run queue

● Run queue indicates the total number of active
processes in the current queue for CPU
● When CPU is ready to execute a process, it picks it up
from the run queue based on the priority of the process
● Processes that are in sleep state, or i/o wait state are
not in the run queue
● A higher number of processes in the run queue can
cause performance issues

CPU Utilization

● This indicates how much of the CPU is currently
getting used
● This is fairly straight forward, and you can view the
CPU utilization from the top command
● 100% CPU utilization means the system is fully
loaded
● A higher % of CPU utilization will cause performance
issues

LOAD average
● This indicates the average CPU load over a specific time period.
● On Linux, load average is displayed for the last 1 minute, 5 minutes,
and 15 minutes. This is helpful to see whether the overall load on the
system is going up or down.
● Load average of “0.25 1.20 1.90″ indicates that the load on the
system is coming down. 0.25 is the load average in the last 1
minute. 1.20 is the load average in the last 5 minutes. 1.90 is the
load average in the last 15 minutes.
● This load average is calculated by combining both the total number
of process in the queue, and the total number of processes in the
uninterruptable task status.

Disk I/O optimization
● Linux currently ships with four different I/O schedulers. They are: deadline, noop,
anticipatory, and cfq. There are many differences between these scheduling
algorithms:
● CFQ: This is the default algorithm in most Linux distributions. It attempts to distribute
all I/O bandwidth evenly among all processes requesting I/O. It is ideal for most
purposes.
● NOOP: The noop algorithm attempts to use as little cpu as possible. It acts as a basic
FIFO queue expecting the hardware controller to handle the performance operations
of the requests.
● Anticipatory: This algorithm attempts to reorder all disk I/O operations to optimize disk
seeks. It is designed to increase performance on systems that have slow disks.
● Deadline: This scheduling algorithm places I/O requests in a priority queue so each is
guaranteed to be ran within a certain time. It is often used in real-time operating
systems.

System scheduler … how to change

● cat /sys/block/sda/queue/scheduler

● change it with:

● echo noop > /sys/block/sda/queue/scheduler

More about System scheduler

● Changing schedulers on the fly allows you to test and
benchmark the algorithms for your specific application
● Once the change is issued, any current I/O operations will
be executed before the new scheduler goes into effect, so
the change will not be instant
● Also remember that once one is set and performs to your
liking, be sure to set the change to be applied on
subsequent reboots

And a little bit more ...

● It is often recommend to use noop or deadline on any
SSD drive
● There is usually no definitive answer to which algorithm to
use
● Benchmarking each one will be your best option
● There are cases where cfq may not be the best scheduler
for your system or application
● An example is if you are running a raid disk array with a
caching raid controller

I/O optimization

● I/O wait is the amount of time CPU is waiting for I/O. If you
see consistent high i/o wait on you system, it indicates a
problem in the disk subsystem.
● You should also monitor reads/second, and writes/second.
This is measured in blocks. i.e number of blocks read/write
per second. These are also referred as bi and bo (block in
and block out).
● tps indicates total transactions per seconds, which is sum
of rtps (read transactions per second) and wtps (write
transactions per seconds).


● Filesystem to use:
– EXT2
– EXT3
– ReiserFS
– EXT4
– BTRFS

Use FS options in fstab (noatime, ...)


● RAID (software or hardware)

– RAID0
– RAID1
– RAID5
– RAID10

Disk optimization

● Use benchmark programs like Bonnie++
● Use hdparm
● Upgrade BIOS of your server and Firmware of your disk

Network Tuning

● A good understanding of TCP/IP concepts is helpful
while analyzing any network issues.
● For network interfaces, you should monitor total
number of packets (and bytes) received/sent through
the interface, number of packets dropped, etc.,

TCP Tunning
For servers that are serving up huge numbers of concurent sessions, there are some tcp options that
should probabaly be enabled. With a large # of clients doing their best to kill the server, its probabaly
not uncommon for the server to have 20000 or more open sockets.
Allows more local ports to be available
echo 1024 65000 > /proc/sys/net/ipv4/ip_local_port_range

Increasing the amount of memory associated with socket buffers can often improve performance

echo 262143 > /proc/sys/net/core/rmem_max

echo 262143 > /proc/sys/net/core/rmem_default

These reduce the amount of work the TCP stack has to do, so is often helpful in this situation

echo 0 > /proc/sys/net/ipv4/tcp_sack

echo 0 > /proc/sys/net/ipv4/tcp_timestamps

Memory Optimization

● If you have 16 GB RAM installed on your system, you have 16
GB of physical memory
● Virtual memory = Swap space available on the disk + Physical
memory. The virtual memory contains both user space and
kernel space
● Using either 32-bit or 64-bit system makes a big difference in
how much memory a process can utilize
● On a 32-bit system a process can only access a maximum of
4GB virtual memory
● On a 64-bit system there is no such limitation

More about Memory optimization ...
● Unused RAM will be used as file system cache by the kernel
● Linux system will swap when it needs more memory. i.e when it needs
more memory than the physical memory
● When it swaps, it writes the least used memory pages from the physical
memory to the swap space on the disk
● Lot of swapping can cause performance issues, as the disk is much
slower than the physical memory, and it takes time to swap the memory
pages from RAM to disk

Memory optimization :)

● Use ECC RAM … yes it's expensive but it's faster
● Use more swap space … application ...
● Don't use swap partition, use swap file, …
● Is everything OK with your swap?
– try memeat script ...

80/20

● Remember the 80/20 rule

80% of the performance improvement comes from
tuning the application, and the rest 20% comes from
tuning the infrastructure components.

System Monitoring Tools

● vmstat
● netstat
● ps
● top
● atop
● mtop
● Iostat
● xosview

Kernel Tuning

● Recompile your kernel
● Exclude unneeded modules
● Use RealTime kernel
●

Samba tuning

● Rebuild it with mmap support. In cases where you
are serving up a large amount of small files, this
seems to be particularly useful.
● You just need to add a "--with-mmap" to the
configure line.

Database optimization

● MySQL – use MySQL tuning App
● Optimize table
● Do you use InnoDB or MyISSAM DB?
● Change database if possible

OpenLDAP Tuning
● The most important tuning aspect for OpenLDAP is deciding what attributes
you want to build indexes on.

Cachesize 10000
dbcachesize 100000
sizelimit 10000
loglevel 0
dbcacheNoWsync

index cn,uid
index uidnumber
index gid
index gidnumber
index mail

● If you add the following parameters to /etc/openldap/slapd.conf before entering the info
into the database, they will all get indexed and performance will increase.

Apache Tuning
Make sure you starting a ton of initial daemons if you want good benchmark scores.
Something like:
MinSpareServers 20
MaxSpareServers 80
StartServers 32

this can be higher if apache is recompiled
MaxClients 256
MaxRequestsPerChild 10000

Note: Starting a massive amount of httpd processes is really a benchmark hack. In most real world cases, setting
a high number for max servers, and a sane spare server setting will be more than adequate.
It's just the instant on load that benchmarks typically generate that the StartServers helps with.

Slow websites

● Use optimizers
● Use memcache
● Tune your apache
● Minimize number of Apache modules
● Change Apache for Nginx

Benchmark

● A good set of benchmarking utilities are often very
helpful in doing system tuning work. It is impossible
to duplicate "real world" situations, but that isnt really
the goal of a good benchmark.
● A good benchmark typically tries to measure the
performance of one particular thing very accurately.
● If you understand what the benchmarks are doing,
they can be very useful tools.

Benchmark Tools
● bonnie++ - is a free file system benchmarking tool for Unix-like operating
systems
● DBench - is a tool to generate I/O workloads to either a filesystem or to a
networked CIFS or NFS server
● http_load - runs multiple http fetches in parallel, to test the throughput of a
web server
● dkftpbench - measuring how many simultaneous dialup users can be down
loading from an FTP site at the same time
● tiobench - is a multi-threaded I/O benchmark
● ttcp - is a utility program for measuring network throughput
● netperf – network performance tester

Identify and solve performace issue

● Understand the problem
– Half of the problem is solved when you clearly understand what
the problem is.
● Monitor and collect data
– After defining the problem clearly, monitor the system and try to
collect as much data as possible on various subsystems
● Eliminate and narrow down issues
– After having a list of potential issues, dive into each one of them
and eliminate any non issues
● Make one change at a time
– don’t try to make multiple changes at one time

TNX ...

● E-mail: dguttierrez@me.com
● Blog: www.d-mashina.net
● CV: www.guttierrez.org

Optimizing Linux Servers

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Optimizing Linux Servers (20)

Recently uploaded (20)

Optimizing Linux Servers