Page 1
Parallels
Parallels Business
Automation Standard
Advanced Configuration guide
Release 4.5
(c) 1999-2013
Page 2
Copyright Notice
Parallels IP Holdings GmbH
Vordergasse 59
8200 Schaffhausen
Switzerland
Tel: + 41 52 632 0411
Fax: + 41 52 672 2010
Copyright © 1999-2013 Parallels IP Holdings GmbH. and its affiliates. All rights reserved.
This product is protected by United States and international copyright laws. The product's underlying
technology, patents, and trademarks are listed at http://www.parallels.com/trademarks
Microsoft, Windows, Windows Server, Windows NT, Windows Vista, and MS-DOS are registered
trademarks of Microsoft Corporation.
Linux is a registered trademark of Linus Torvalds.
Mac is a registered trademark of Apple, Inc.
All other marks and names mentioned herein may be trademarks of their respective owners.
Page 3
Contents
Preface 4
Typographi cal Convent ions .......................................................................................................................... 4
Feedback ....................................................................................................................................................... 5
Shell Prompts in Command Example s.......................................................................................................... 5
General Conventions..................................................................................................................................... 6
Parallels Business Automation - Standard Profiling 7
Turning Profiler On and Off ......................................................................................................................... 7
Viewing Profiler Statistics ............................................................................................................................ 8
Parallels Business Automation - Standard Cluster 9
Introduction to Clustering ........................................................................................................................... 10
Database Splitting And Cluste ring .............................................................................................................. 12
Backend Clustering ..................................................................................................................................... 13
Backing Up Management Node 14
Using Management Nod e Backup T ool ...................................................................................................... 15
Index 17
Page 4
4
Items you must select, such
as menu options, command
Titles of chapters, sections,
Used to emphasize the
importance of a point, to
introduce a term or to
designate a command line
is to be
The names of commands,
The license file is located in
C
HAPTER
1
Preface
In This Chapter
Typographical Conventions .................................................................................................. 4
Feedback ............................................................................................................................... 5
Shell Prompts in Command Examples .................................................................................. 5
General Conventions ............................................................................................................. 6
Typographical Conventions
Before you start using this guide, it is important to understand the documentation conventions
used in it.
The following kinds of formatting in the text identify special information.
Formatting convention Type of Information Example
Special Bold
buttons, or items in a list.
and subsect ions.
Italics
placeholder, which
replaced with a real name or
value.
Monospace
files, directories, and do main
names.
Go to the System tab.
Read the Basic Administration
chapter.
The system supports the so
called wildcard character
search.
the
http://docs/common/
licenses directory.
Page 5
Preface 5
Preformatted
On-screen computer output in
other programming
# ls –al /files
Preformatted Bold
What you type, contrasted
screen computer
# cd /root/rpms/php
Names of keys on the
Key combinations for which
the user must press and hold
your command-line sessions;
source code in XML, C++, or
languages.
with onoutput.
total 14470
CAPITALS
keyboard.
KEY+KEY
down one ke y and the n press
another.
SHIFT, C TRL, ALT
CTRL+P, ALT+F4
Feedback
If you have found a mistake in this guide, or if you have suggestions or ideas on how to improve
this guide, please send your feedback using the online form at
http://www.parallels.com/en/support/usersdoc/. Please include in your report the guide's title,
chapter and section titles, and the fragment of text in which you have found an error.
Shell Prompts in Command Examples
Command line examples throughout this guide presume that you are using the Bourne-again
shell (bash). Whenever a command can be run as a regular user, we will display it with a dollar
sign prompt. When a command is meant to be run as root, we will display it with a hash mark
prompt:
Bourne-again shell pr ompt
Bourne-again shell root prompt
$
#
Page 6
Preface 6
General Conventions
Be aware of the following conventions used in this book.
Chapters in this guide are divided into sections, which, in turn, are subdivided into
subsections. For example, Documentation Conventions is a section, and General Conventions
is a subsection.
When following steps or using examples, be sure to type double-quotes ("), left single-
quotes (`), and right single-quotes (') exactly as shown.
The key referred to as RETURN is labeled ENTER on some keyboards.
The root path usually includes the /bin, /sbin, /usr/bin and /usr/sbin directories, so
the steps in this book show the commands in these directories without absolute path names.
Steps that use commands in other, less common, directories show the absolute paths in the
examples.
Page 7
7
Viewing Profiler Statistics .................................................................................................... 8
C
HAPTER
2
Parallels Business Automation Standard Profiling
Profiling of web requests in Parallels Business Automation - Standard is very useful. This
allows understanding, which requests are slow (from some IP address, or to some specific URL,
etc.).
You can measure time needed to access the requested web page with the help of PBAS
profiling. This allows you to check the performance and spot slow pages needing
improvements.
You are not recommended to use profiler all the time as it is likely to put an additional load on
your system resources and disk space. When profiling is turned on, performance will reduce,
mostly due to operations with file system (for saving log files).
In This Chapter
Turning Profiler On and Off ................................................................................................. 7
Turning Profiler On and Off
Profiler cannot be turned on from Parallels Business Automation - Standard web interface. To
activate or deactivate profiler, it is necessary to edit the Parallels Business Automation Standard configuration file and restart the backend web server.
To turn profiler on:
1. Edit the Parallels Business Automation - Standard configuration file:
/etc/hspc/hspc.conf
Set the CORE_PROFILER parameter as follows :
CORE_PROFILER = 1
Page 8
Parallels Business Automation - Standard Profiling 8
2. Restart the backend web server:
/etc/init.d/hspcd restart
To turn profiler off:
1. Edit the Parallels Business Automation - Standard configuration file and set the
CORE_PROFILER to zero:
CORE_PROFILER = 0
2. Restart hspcd.
Viewing Profiler Statistics
When profiling is turned on, it collects all the necessary information and stores it in the Parallels
Business Automation - Standard database (logs are saved in files).
Debug/sql logs are saved into files in the /var/log/hspc/profiler directory.
In addition, all profiling statistics are is shown in the Provider Control Center web interface, on
the special page (Action Log > Profiler Statistics).
The profiler statistics columns include:
ID - Process ID;
Location - Request URL;
Wallclock - real time spent on the request processing from the very start to its finish. Wallclock
time may be not necessarily the time occupied only by the requested process. It may record the
time during which another interrupting process engaged the system and prevented the requested
process from execution.
User time - in UNIX terms, the user time of the process;
System time - in UNIX terms, the system time of the process;
Initiator's IP - the IP of the remote host initiating the process;
Started - Start date and time;
Finished - Finish date and time.
Page 9
9
Backend Clustering ............................................................................................................... 13
C
HAPTER
3
Parallels Business Automation Standard Cluster
Hosting provider business based on Parallels solutions grows rapidly day by day. Increasing
accounts number and resources volume used by them calls for loading of MySQL server. Rising
accounts-related activity and extending set of nodes (Parallels Virtuozzo Containers, Plesk,
SiteBuilder, etc.) increases Parallels Business Automation - Standard back-end and front-end
parts load. Hardware resources consumption exceeds a reasonable limit. The only one solution
of the problem is increasing hardware resources volume available both for Parallels Business
Automation - Standard management server and MySQL server. However, upgrading hardware
resources Parallels Business Automation - Standard is hosted on is not very efficient and
scalable solution. The real way out is to use clustering technologies.
Another problem is ensuring an uninterrupted and reliable access to service. In the situation
when a client base is massive, any minimal downtime of the management node provokes loss of
clients and therefore, loss of returns. Parallels Business Automation - Standard management
node and/or MySQL database server maintenance (upgrade, backup/restore operations,
hardware problems and so on) produces such undesirable downtimes. The cluster technology
will shorten or even allow avoiding downtimes and minimize related risks.
In This Chapter
Introduction to Clustering ..................................................................................................... 10
Database Splitting And Clustering ........................................................................................ 12
Page 10
Parallels Business Automation - Standard Cluster 10
Introduction to Clustering
The clustering solution is based on splitting Parallels Business Automation - Standard parts b y
different hardware nodes. There are three common parts of Parallels Business Automation Standard software that produce high load:
front-end (less loading)
back-end
database (most loading)
In case of front-end that produces less load, a node will be combined with load balancing
system and synchronization server. The MySQL database server can be clustered natively.
Parallels Business Automation - Standard back-end nodes can be replicated many t imes without
any limitations.
Parallels Business Automation - Standard can be installed and successfully used without any
clustering. After the node becomes overused the good solution is to move out MySQL server
from the management node to a separate server. In case a new MySQL server resources are
insufficient a database cluster can be created to prevent this. The time for next step comes when
the Parallels Business Automation - Standard management node becomes overloaded. The best
solution to solve the problem is to set a cluster for Parallels Business Automation - Standard
back-end and load balancing between them.
Here are a few words about load balancing system. The key purposes of the system are to tr ack
users’ requests and distribute them uniformly between all of available back-end systems. In
addition, the load balancing system monitors back-end nodes availability and latency. This
allows a balancing mechanism to provide more flexible and powerful loading results.
Another situation is when a hosting business is planned to be huge, so Parallels Business
Automation - Standard cluster should be set first to prevent any downtimes. Actually, this
situation does not call for any actions different from last stages of clustering described above, so
we are not focus on any individual use-cases.
Page 11
Parallels Business Automation - Standard Cluster 11
Figure 1: HSPcomplete Cluster
Here, client workstations are located in the “untrusted” Internet area. Parallels Business
Automation - Standard front-end node is only accessible through firewall. All other Parallels
Business Automation - Standard nodes are located in the “trusted” Ethernet area so there are no
any additional resources spending for security.
In a few words, the clustering process may consist of the following steps:
1. Move the database from the Parallels Business Automation - Standard Management Node to
a separate server.
2. Create Parallels Business Automation - Standard backend node(s). If Parallels Business
Automation - Standard is installed at a Virtuozzo Container (the most widely used
approach), this can be done by cloning the Parallels Business Automation - Standard
Container. Allocate IPs to backend nodes and assign hostnames. The original Parallels
Business Automation - Standard Container will later also be used as a backend node.
3. Create the frontend node. To this effect, create a 'light' Container and install Parallels
Business Automation - Standard at it. Then run hspc-cluster.pl --frontend at
the frontend node and specify all the backend nodes when prompted.
Page 12
Parallels Business Automation - Standard Cluster 12
Database Splitting And Clustering
Parallels Business Automation - Standard team has prepared a tool set that will help building the
cluster. The tool is shipped in the form of Perl script, its usage is quite simple and we describe it
step-by-step below.
First of all, it is necessary to prepare a clean Linux installation with installed MySQL server
version not lower than 4.1 with opened SSH access for user root.
To split database from the back-end node:
1. Start the hspc-cluster.pl tool in the back-end configuration mode ( with -b or --
backend flag).
2. Reply yes to the Do you want to move out database on another
server? question.
3. Enter a new server’s address. After this the script will generate (if needed) and put a DSA
SSH public key to a remote server.
4. The script will stop the MySQL service and start migrating its data file. For a bulky dataset
this can take some time.
5. After the database is migrated, the script will be automatically cloned at a remote server and
transparently started in the database configuration mode (with -d or --database flag).
The remote script will exit and its local copy will continue to work.
After the steps enlisted above are finished, the database node will be completely split from
Parallels Business Automation - Standard. Any unnecessary services will be stopped to prevent
parasitic loading of the hardware node. Do not forget to restart related services (will be
prompted by the configuration tool as well).
Page 13
Parallels Business Automation - Standard Cluster 13
Backend Clustering
Back-end clustering procedure is quite simple. It is possible to add a new back-end nodes ondemand. It is supposed that the database node is already moved out on a remote server. In
addition, it is required to have a number of clean Parallels Business Automation - Standard
installations which will be the next back-end nodes in a cluster plus another one (and only one!)
for the front-end server. It is recommended to use original Parallels Business Automation Standard installation as a future front-end node.
This procedure can be repeated many times:
1. Log in to the front-end node and start the hspc-cluster.pl tool in the front-end
configuration mode (with -f or --frontend flag)
2. Enter a list of back-end nodes available for the cluster. It is recommended but not required
to enter the most recently used (production) management node first.
3. Select whether you want to set up database access parameters generally or individually for
all the back-end nodes. Those settings will be cloned over all the Parallels Business
Automation - Standard cluster nodes.
4. The script will generate (if needed) a SSH DSA public key and set it for every back-end
server specified for future access, so the script will prompt for a root password for each of
back-end nodes.
5. Rsync synchronization server will be configured by scheduling synchronization every 15
minutes. In this case, the front-end server will be a primary storage node for data
synchronization.
6. Configuration for every back-end will be done by cloning the script for every node and
transparently starting it in an automated or interactive mode (depending on the choice made
on step 3).
7. Front-end node configuration itself starts. This will install a certain number of configuration
files and scripts for a load balancing system.
8. All the new services will be registered and configured. The script will prompt also for a
necessary services restart. The following services will be run on the front-end node: Task
Manager, Virtuozzo Collector and Power DNS services. The rest will be cloned amon g the
back ends.
Page 14
14
C
HAPTER
4
Backing Up Management Node
This chapter describes how to back up the server that runs Parallels Business Automation Standard (Management Node).
In This Chapter
Using Management Node Backup Tool ................................................................................ 15
Page 15
Backing Up Management Node 15
Using Managemen t No de Backup
Standard configuration files and
Parallels Virtuozzo Containers
Tool
Management Node can be backed up using the hspc-mn-backup utility shipped with
Parallels Business Automation - Standard distribution. After Parallels Business Automation Standard installation hspc-mn-backup is located at the Management Node in the
/usr/sbin directory.
The hspc-mn-backup utility allows you to:
Create Management Node backup.
Restore Management Node from backup.
Migrate Management Node from one server to another. In this case, hspc-mn-backup
creates a Management Node backup, moves it to a server specified, and restores
Management Node on a new place.
The following Management Node directories are backed up:
Directory Description
/var/lib/mysql/
/etc/my.cnf
/var/www/.ssh/
/etc/hspc/
/etc/httpd/
/etc/hspcd/
/var/opt/hspc-root
/var/opt/hspc-eventmanager/
/vz/
Parallels Business Automation Standard database
MySQL configuration file
Virtuozzo hardware nodes access
Parallels Business Automation -
license
Frontend configuration files
Application server configuration files
Customiza tion f i le s
Event manager templates
integration files
Management Node backup is a compressed tarball and by default, it is put into a current
directory. A backup file is named hspc-mn-backup_YYYY.MM.DD (as design), where
YYYY.MM.DD is a backup creation date in the Year.Month.Day format (for example, hspc-
mn-backup_2006.09.06.tar.bz2). A backup creation date is also used as a backup
identifier (ID) and you can use this ID in the hspc-mn-backup command line to specify a backup
file to deal with.
Page 16
Backing Up Management Node 16
Command line syntax:
Increase the verbose level. Single use of the
Decrease the verbose level. Single use of the
tory specified instead
Use a specified backup (ID ) instead of a dated of
actions consist in stopping all the services on a
Do not perform the check of the ability of
instead of
y database files copy. This option is
used when MySQL versions are different on the
hspc-mn-backup [options]
Options:
Option Description
-u or --usage
-? or -h or --help
-b or --backup
-r or --restore
-m or --migrate
-v or --verbose
-q or --quiet
-C or --directory
-i or --id
-p or --no-pre
Print a brief help and exit.
Print the manual page and exit.
Start Management Node backup process.
Start restore process.
Start migration process to the node specified.
option sets the output to debug level.
option sets the output only for serious problems.
Put backup files into a direc
of a current one.
today.
Do not perform pre-backup actions. Pre-backup
Management Node.
-P or --no-post
Do not perform post-backup actions. Post-backup
actions consist in starting all services.
-f or --force
restoring backup on a current node.
-d or --dbdump The script will perform mysqldump
the binar
source and the destination nodes.
Page 17
Index
B
Backend Clustering • 13
Backing Up Management Node • 14
C
Copyright Notice • 2
D
Database Splitting And Clustering • 12
F
Feedback • 5
G
General Conventions • 6
I
Introduction to Clustering • 10
P
Parallels Business Automation - Standard
Cluster • 9
Parallels Business Automation - Standard
Profiling • 7
Preface • 4
S
Shell Prompts in Command Examples • 5
T
Turning Profiler On and Off • 7
Typographical Conventions • 4
U
Using Management Node Backup Tool • 15
V
Viewing Profiler Statistics • 8
Loading...