Dude McDude

Red Hat Directory Server 8.2

Performance Tuning Guide
Optimizing performance on Red Hat Directory Server 8.2
Edition 8.2.1
Red Hat Directory Server 8.2 Performance Tuning Guide
Optimizing performance on Red Hat Directory Server 8.2
Edition 8.2.1
Dude McDude
My Org Best Div in the place
[email protected]
Legal Notice
Copyright 2010 Red Hat, Inc..
This document is licensed by Red Hat under the Creative Commons Attribution-ShareAlike 3.0 Unported
License. If you distribute this document, or a modified version of it, you must provide attribution to Red
Hat, Inc. and provide a link to the original. If the document is modified, all Red Hat trademarks must be
removed.
Red Hat, as the licensor of this document, waives the right to enforce, and agrees not to assert, Section
4d of CC-BY-SA to the fullest extent permitted by applicable law.
Red Hat, Red Hat Enterprise Linux, the Shadowman logo, JBoss, MetaMatrix, Fedora, the Infinity Logo,
and RHCE are trademarks of Red Hat, Inc., registered in the United States and other countries.
Linux is the registered trademark of Linus Torvalds in the United States and other countries.
Java is a registered trademark of Oracle and/or its affiliates.
XFS is a trademark of Silicon Graphics International Corp. or its subsidiaries in the United States
and/or other countries.
MySQL is a registered trademark of MySQL AB in the United States, the European Union and other
countries.
Node.js is an official trademark of Joyent. Red Hat Software Collections is not formally related to or
endorsed by the official Joyent Node.js open source or commercial project.
The OpenStack Word Mark and OpenStack Logo are either registered trademarks/service marks or
trademarks/service marks of the OpenStack Foundation, in the United States and other countries and
are used with the OpenStack Foundation's permission. We are not affiliated with, endorsed or
sponsored by the OpenStack Foundation, or the OpenStack community.
All other trademarks are the property of their respective owners.
Abstract
This article provides some procedures and options that administrators can use to optimize the
performance of their Red Hat Directory Server deployments. Performance tuning a Directory Server
instance is unique to each server because of differences for every server in its machine environment,
directory size and data type, load and network use, even the types of operations that users and clients
perform. The purpose of this article is to highlight the features that Red Hat Directory Server provides for
tracking and assessing server and database performance. There are also some procedures given to
help tune server performance. For more in-depth planning information, however, check out the Red Hat
Directory Server Deployment Guide, and for exhaustive command-line and UI-based administrative
instructions, see the Red Hat Directory Server Administrator's Guide.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table of Contents
1. Intro to Directory Server Performance Tuning
1.1. Setting Goals for Directory Server Performance
1.2. Giving Feedback
2. Tracking Server and Database Performance
2.1. Monitoring Server Activity
2.1.1. Monitoring the Server from the Directory Server Console
2.1.2. Monitoring the Directory Server from the Command Line
2.2. Monitoring Database Activity
2.3. Monitoring Database Link Activity
2.4. Viewing Log Files
2.5. Replacing Log Files with a Named Pipe
2.6. Improving Logging Performance
3. Optimizing System Performance
3.1. Tuning and Sizing System Memory
3.1.1. Setting the Minimum System Page Cache
3.1.2. Managing the Swap Space
3.1.3. Tuning for 32-Bit Systems
3.2. Using dsktune for Preliminary Tuning
3.3. Tuning File Descriptors
3.3.1. Resetting the Host Machine's File Descriptors
3.3.2. Setting Directory Server File Descriptor Values
3.4. Setting Shell Limits for the Directory Server User
3.4.1. Setting Shell Limits for File Descriptors
3.4.2. Setting Shell Limits for the Max Number of Processes
3.5. Tuning TCP Ports
3.6. Changing Firewall Settings for the Directory Server Console
3.7. DNS Requirements
4. Improving Search Performance (and Balancing Read Performance)
4.1. Using Indexes
4.2. Tuning Directory Server Resource Settings
4.3. Setting Index Scan Limits
4.4. Tuning the Database Cache for Searches
4.5. Tuning the Database Settings for Searches
4.6. Managing Special Entries
5. Tuning Transaction Logging
5.1. Changing the Location of the Database Transaction Log
5.2. Changing the Database Checkpoint Interval
5.3. Disabling Durable Transactions
5.4. Specifying Transaction Batching
6. Managing the Database Cache Settings
6.1. Tuning Entry Cache
6.2. Tuning Database Cache
7. Tuning Database Link Performance
7.1. Managing Connections to the Remote Server
7.1.1. Managing Connections to the Remote Server Using the Console
7.1.2. Managing Connections to the Remote Server from the Command Line
7.2. Detecting Errors During Normal Processing
3
3
4
4
5
5
9
11
16
17
18
19
20
21
21
21
21
22
22
23
23
24
24
25
25
25
26
26
26
28
29
29
30
31
31
32
32
33
33
34
34
35
36
36
36
38
39
Table of Contents
1
7.3. Managing Threaded Operations 40
Red Hat Directory Server 8.2 Performance Tuning Guide
2
1. Intro to Directory Server Performance Tuning
This article provides some procedures and options that administrators can use to optimize the
performance of their Red Hat Directory Server deployments. Performance tuning a Directory Server
instance is unique to each server because of differences for every server in its machine environment,
directory size and data type, load and network use, even the types of operations that users and clients
perform.
The purpose of this article is to highlight the features that Red Hat Directory Server provides for tracking
and assessing server and database performance. There are also some procedures given to help tune
server performance. For more in-depth planning information, however, check out the Red Hat Directory
Server Deployment Guide, and for exhaustive command-line and UI-based administrative instructions,
see the Red Hat Directory Server Administrator's Guide.
The performance tuning features here apply to Red Hat Directory Server 8.2.
1.1. Setting Goals for Directory Server Performance
Performance tuning is simply a way to identify potential (or real) bottlenecks in the normal operating
environment of the server and then taking steps to mitigate those bottlenecks.
The general plan for performance tuning is:
1. Assess the environment. Look at everything around the Directory Server: its usage, the load, the
network connection and reliability, most common operations, the physical machine its on, along
with any services competing for its resources.
2. Measure the current Directory Server performance and establish baselines.
3. Identify the server areas which can be improved.
4. Make any changes to the Directory Server settings and, potentially, to the host machine.
5. Measure the Directory Server performance again to see how the changes affected the
performance.
Directory Server provides some sort of monitoring in three areas:
The server process (counters and logs)
The databases (counters)
Any database links (counters)
In the Directory Server, most performance measurements are going to be how well the Directory Server
retrieves and delivers information to clients. With that in mind, these are the server areas that can be
tuned for the best Directory Server performance (and these are the areas covered in this article):
Search operations
Indexing performance (which affects both search and write operations)
Database transactions
Database and entry cache settings
Database links
Other changes can be made to the host machine's settings or hardware which can also affect Directory
Server performance:
1. Intro to Directory Server Performance Tuning
3
Available memory (based on directory size)
Other servers running on the same machine (which could compete for resources)
Distributing user databases across other Directory Server instances on other machines
Balancing server loads due to network performance
This changes relate much more to planning an effective Directory Server deployment than changes that
can be made to an instance. Reviewing the Deployment Guide can provide more detail about how to plan
an optimal enterprise deployment.
1.2. Giving Feedback
If there is any error in this Performance and Tuning Guide or there is any way to improve the
documentation, please let us know. Bugs can be filed against the documentation for Red Hat Directory
Server through Bugzilla, http://bugzilla.redhat.com/bugzilla. Make the bug report as specific as possible,
so we can be more effective in correcting any issues:
1. Select the Red Hat Directory Server product.
2. Set the component to Doc - performance-and-tuning.
3. Set the version number to 8.2.
4. For errors, give the page number (for the PDF) or URL (for the HTML), and give a succinct
description of the problem, such as incorrect procedure or typo.
For enhancements, put in what information needs to be added and why.
5. Give a clear title for the bug. For example, "Incorrect command example for setup
script options" is better than "Bad example".
We appreciate receiving any feedback requests for new sections, corrections, improvements,
enhancements, even new ways of delivering the documentation or new styles of docs. You are welcome
to contact Red Hat Content Services directly at [email protected].
2. Tracking Server and Database Performance
Red Hat Directory Server has two methods of recording and tracking performance data: performance
counters and logs. Counters are used to determine how well the Directory Server performing, particularly
in database performance; logs are used to diagnose any problem areas with server and LDAP
operations and configuration.
Performance counters focus on the operations and information of the Directory Server for the server, all
configured databases, and database links (chaining databases).
There are three types of logs: access (for client connections), errors (for errors, warnings, and details of
events), and audit (changes to Directory Server configuration). The access and error logs run by default
(and the errors log is required for the server to run). Audit logging, because of the overhead, must be
enabled manually.
NOTE
The access log is buffered. This allows full access logging even with highly loaded servers, but
there is a time lag between when the event occurs in the server and when the event is written to
the log.
Red Hat Directory Server 8.2 Performance Tuning Guide
4
2.1. Monitoring Server Activity
The Directory Server's current activities can be monitored from either the Directory Server Console or
the command line. It is also possible to monitor the activity of the caches for all of the database.
NOTE
Some of the counters for Directory Server database attributes monitored by server use 64-bit
integers, even on 32-bit systems (total connections, operations initiated, operations completed,
entries sent, and bytes sent). On high-volume systems, this keeps the counters from rolling too
quickly and skewing monitoring data.
2.1.1. Monitoring the Server from the Directory Server Console
1. In the Directory Server Console, select the Status tab.
2. In the navigation tree, select Performance Counters.
The Status tab in the right pane displays current information about server activity. If the server is
currently not running, this tab will not provide performance monitoring information.
The General Information table shows basic information about the server, which helps set a
baseline about the statistics that have been gathered.
Table 1. General Information (Server)
Field Description
Server Version Identifies the current server version.
Startup Time on Server The date and time the server was started.
Current Time on Server The current date and time on the server.
The Resource Summary table shows the totals of all operations performed by that instance.
2. Tracking Server and Database Performance
5
Table 2. Resource Summary
Resource Usage Since Startup Average Per Minute
Connections The total number of connections
to this server since server
startup.
Average number of connections
per minute since server startup.
Operations Initiated The total number of operations
initiated since server startup.
Operations include any client
requests for server action,such
as searches, adds, and
modifies. Often, multiple
operations are initiated for each
connection.
Average number of operations
per minute since server startup.
Operations Completed The total number of operations
completed by the server since
server startup.
Average number of operations
per minute since server startup.
Entries Sent to Clients The total number of entries sent
to clients since server startup.
Entries are sent to clients as the
result of search requests.
Average number of entries sent
to clients per minute since
server startup.
Bytes Sent to Clients The total number of bytes sent
to clients since server startup.
Average number of bytes sent
to clients per minute since
server startup.
The Current Resource Usage table shows the current demands on the server.
Red Hat Directory Server 8.2 Performance Tuning Guide
6
Table 3. Current Resource Usage
Resource Current Total
Active Threads The current number of active threads used for
handling requests. Additional threads may be
created by internal server tasks, such as
replication or chaining.
Open Connections The total number of open connections. Each
connection can account for multiple operations,
and therefore multiple threads.
Remaining Available Connections The total number of remaining connections that
the server can concurrently open. This number is
based on the number of currently open
connections and the total number of concurrent
connections that the server is allowed to open. In
most cases, the latter value is determined by the
operating system and is expressed as the
number of file descriptors available to a task.
Threads Waiting to Write to Client The total number of threads waiting to write to the
client. Threads may not be immediately written
when the server must pause while sending data
to a client. Reasons for a pause include a slow
network, a slow client, or an extremely large
amount of information being sent to the client.
Threads Waiting to Read from Client The total number of threads waiting to read from
the client. Threads may not be immediately read if
the server starts to receive a request from the
client, and then the transmission of that request
is halted for some reason. Generally, threads
waiting to read are an indication of a slow network
or client.
Databases in Use The total number of databases being serviced by
the server.
The Connection Status table simply lists the current active connections, with related connection
information.
2. Tracking Server and Database Performance
7
Table 4. Connection Status
Table Header Description
Time Opened The time on the server when the connection was
initially opened.
Started The number of operations initiated by this
connection.
Completed The number of operations completed by the
server for this connection.
Bound as The distinguished name used by the client to bind
to the server. If the client has not authenticated to
the server, the server displays not bound in
this field.
Read/Write Indicates whether the server is currently blocked
for read or write access to the client. There are
two possible values:
Not blocked means that the server is idle,actively
sending data to the client, or actively reading data
from the client.
Blocked means that the server is trying to send
data to the client or read data from the client but
cannot. The probable cause is a slow network or
client.
The Global Database Cache table lists the cache information for all databases within the Directory
Server instance.
NOTE
Although the performance counter for the global database cache is listed with the other server
performance counters in the Directory Server Console, the actual database cache entries are
located and monitored in cn=monitor,cn=database_instance, cn=ldbm database,
cn=plugins, cn=config, as are the other database activities.
Red Hat Directory Server 8.2 Performance Tuning Guide
8
Table 5. Global Database Cache Information
Table Header Description
Hits The number of times the server could process a
request by obtaining data from the cache rather
than by going to the disk.
Tries The total number of database accesses since
server startup.
Hit Ratio The ratio of cache tries to successful cache hits.
The closer this number is to 100%, the better.
Pages Read In The number of pages read from disk into the
cache.
Pages Written Out The number of pages written from the cache back
to disk.
Read-Only Page Evicts The number of read-only pages discarded from
the cache to make room for new pages. Pages
discarded from the cache have to be written to
disk, possibly affecting server performance. The
lower the number of page evicts the better.
Read-Write Page Evicts The number of read-write pages discarded from
the cache to make room for new pages. This
value differs from Pages Written Out in that
these are discarded read-write pages that have
not been modified. Pages discarded from the
cache have to be written to disk, possibly
affecting server performance. The lower the
number of page evicts, the better.
2.1.2. Monitoring the Directory Server from the Command Line
The Directory Server's current activities can be monitored using LDAP tools such as ldapsearch ,
with the following characteristics:
Search with the attribute filter objectClass=*.
Use the search base cn=monitor; the monitoring attributes for the server are found in the
cn=monitor entry.
Use the search scope base.
For example:
/usr/lib64/mozldap/ldapsearch -D "cn=directory manager" -w secret -p 389 -h
server.example.com -s base -b "cn=monitor" "(objectclass=*)"
The monitoring attributes for the Directory Server are found in the cn=monitor entry.
Monitoring the server's activities using ldapsearch shows the following information:
[1]
2. Tracking Server and Database Performance
9
Table 6. Server Monitoring Attributes
Attribute Description
version Identifies the directory's current version number.
threads The current number of active threads used for
handling requests. Additional threads may be
created by internal server tasks, such as
replication or chaining.
connection:fd:opentime:opsinitiated:opscompleted
:binddn:[rw]
Provides the following summary information for
each open connection (only available if you bind
to the directory as Directory Manager):
fd The file descriptor used for this connection.
opentime The time this connection was
opened.
opsinitiated The number of operations initiated
by this connection.
opscompleted The number of operations
completed.
binddn The distinguished name used by this
connection to connect to the directory.
rw The field shown if the connection is blocked
for read or write.
By default, this information is available to
Directory Manager. However, the ACI associated
with this information can be edited to allow others
to access the information.
currentconnections Identifies the number of connections currently in
service by the directory.
totalconnections Identifies the number of connections handled by
the directory since it started.
dtablesize Shows the number of file descriptors available to
the directory. Each connection requires one file
descriptor: one for every open index, one for log
file management, and one for ns-slapd itself.
Essentially, this value shows how many additional
concurrent connections can be serviced by the
directory. For more information on file descriptors,
refer to the operating system documentation.
readwaiters Identifies the number of threads waiting to read
data from a client.
opsinitiated Identifies the number of operations the server
has initiated since it started.
opscompleted Identifies the number of operations the server
has completed since it started.
entriessent Identifies the number of entries sent to clients
Red Hat Directory Server 8.2 Performance Tuning Guide
10
entriessent Identifies the number of entries sent to clients
since the server started.
bytessent Identifies the number of bytes sent to clients
since the server started.
currenttime Identifies the time when this snapshot of the
server was taken. The time is displayed in
Greenwich Mean Time (GMT) in UTC format.
starttime Identifies the time when the server started. The
time is displayed in Greenwich Mean Time (GMT)
in UTC format.
nbackends Identifies the number of back ends (databases)
the server services.
backendmonitordn Identifies the DN of each directory database.
2.2. Monitoring Database Activity
NOTE
Some of the counters for Directory Server database attributes monitored by server use 64-bit
integers, even on 32-bit systems (entry cache hits, entry cache tries, the current cache size, and
the maximum cache size). On high-volume systems, this keeps the counters from rolling too
quickly and skewing monitoring data.
1. In the Directory Server Console, select the Status tab.
2. In the navigation tree, expand the Performance Counters folder, and select the database to
monitor.
The tab displays current information about database activity. If the server is currently not running,
this tab will not provide performance monitoring information.
2. Tracking Server and Database Performance
11
The Summary Information section shows the cumulative information for all of the databases being
monitored and some cache-related configuration settings which are applied to all databases.
Red Hat Directory Server 8.2 Performance Tuning Guide
12
Table 7. Summary Information
Performance Metric Current Total
Read-Only Status Shows whether the database is currently in read-
only mode. The database is in read-only mode
when the nsslapd-readonly attribute is set to
on.
Entry Cache Hits The total number of successful entry cache
lookups. That is, the total number of times the
server could process a search request by
obtaining data from the cache rather than by
going to disk.
Entry Cache Tries The total number of entry cache lookups since
the directory was last started. That is, the total
number of entries requested since server startup.
Entry Cache Hit Ratio Ratio that indicates the number of entry cache
tries to successful entry cache lookups. This
number is based on the total lookups and hits
since the directory was last started. The closer
this value is to 100%, the better. Whenever an
operation attempts to find an entry that is not
present in the entry cache, the directory has to
perform a disk access to obtain the entry. Thus,
as this ratio drops towards zero, the number of
disk accesses increases, and directory search
performance drops.
To improve this ratio, increase the size of the
entry cache by increasing the value of the
nsslapd-cachememsize attribute in the
cn=database_name, cn=ldbm
database,cn=plugins,cn=config entry for
the database. In the Directory Server Console,
this is set in the Memory available for
cache field in the database settings.
Current Entry Cache Size (in Bytes) The total size of directory entries currently
present in the entry cache.
Maximum Entry Cache Size (in Bytes) The size of the entry cache maintained by the
directory.
This value is managed by the nsslapd-
cachememsize attribute in the
cn=database_name, cn=ldbm
database,cn=plugins,cn=config entry for
the database. This is set in the Memory
available for cache field in the database
settings in the Directory Server Console.
Current Entry Cache Size (in Entries) The number of directory entries currently present
in the entry cache.
Maximum Entry Cache Size (in Entries) DEPRECATED.
2. Tracking Server and Database Performance
13
The maximum number of directory entries that
can be maintained in the entry cache.
Do not attempt to manage the cache size by
setting a maximum number of allowed entries.
This can make it difficult for the host to allocate
RAM effectively. Manage the cache size by
setting the amount of RAM available to the cache,
using the nsslapd-cachememsize attribute.
There are many different databases listed for the database monitoring page, by default, because
databases are maintained for both entries and indexed attributes. All databases, though, have the same
kind of cache information monitored in the counters.
Red Hat Directory Server 8.2 Performance Tuning Guide
14
Table 8. Database Cache Information
Performance Metric Current Total
Hits The number of times the database cache
successfully supplied a requested page.
Tries The number of times the database cache was
asked for a page.
Hit Ratio The ratio of database cache hits to database
cache tries. The closer this value is to 100%, the
better. Whenever a directory operation attempts
to find a portion of the database that is not
present in the database cache, the directory has
to perform a disk access to obtain the appropriate
database page. Thus, as this ratio drops towards
zero, the number of disk accesses increases,
and directory performance drops.
To improve this ratio, increase the amount of data
that the directory maintains in the database cache
by increasing the value of the nsslapd-
dbcachesize attribute. This is the Maximum
Cache Size database setting in the Directory
Server Console.
Pages Read In The number of pages read from disk into the
database cache.
Pages Written Out The number of pages written from the cache back
to disk. A database page is written to disk
whenever a read-write page has been modified
and then subsequently deleted from the cache.
Pages are deleted from the database cache when
the cache is full and a directory operation
requires a database page that is not currently
stored in cache.
Read-Only Page Evicts The number of read-only pages discarded from
the cache to make room for new pages.
Read-Write Page Evicts The number of read-write pages discarded from
the cache to make room for new pages. This
value differs from Pages Written Out in that
these are discarded read-write pages that have
not been modified.
2. Tracking Server and Database Performance
15
Table 9. Database File-Specific
Performance Metric Current Total
Cache Hits The number of times that a search result resulted
in a cache hit on this specific file. That is, a client
performs a search that requires data from this
file, and the directory obtains the required data
from the cache.
Cache Misses The number of times that a search result failed to
hit the cache on this specific file. That is, a search
that required data from this file was performed,
and the required data could not be found in the
cache.
Pages Read In The number of pages brought to the cache from
this file.
Pages Written Out The number of pages for this file written from
cache to disk.
2.3. Monitoring Database Link Activity
The activity for database links (chained databases) can also be viewed, but only through the command
line using ldapsearch to return the monitoring attributes that are required. The monitoring attributes
are stored in the cn=monitor,cn=database_link_name,cn=chaining
database,cn=plugins,cn=config.
For example:
/usr/lib64/mozldap/ldapsearch -D "cn=directory manager" -w secret -p 389 -h
server.example.com -s sub -b "cn=monitor,cn=DBLink1,cn=chaining
database,cn=plugins,cn=config" "(objectclass=*)" nsAddCount
Red Hat Directory Server 8.2 Performance Tuning Guide
16
Table 10. Database Link Monitoring Attributes
Attribute Name Description
nsAddCount The number of add operations received.
nsDeleteCount The number of delete operations received.
nsModifyCount The number of modify operations received.
nsRenameCount The number of rename operations received.
nsSearchBaseCount The number of base-level searches received.
nsSearchOneLevelCount The number of one-level searches received.
nsSearchSubtreeCount The number of subtree searches received.
nsAbandonCount The number of abandon operations received.
nsBindCount The number of bind request received.
nsUnbindCount The number of unbinds received.
nsCompareCount The number of compare operations received.
nsOperationConnectionCount The number of open connections for normal
operations.
nsBindConnectionCount The number of open connections for bind
operations.
For more information about ldapsearch, see the Directory Server Configuration, Command, and File
Reference.
2.4. Viewing Log Files
NOTE
The access and error logs are enabled by default and can be viewed immediately. before the
audit log can be viewed, audit logging must be enabled for the directory, or the audit log will not be
kept.
1. In the Directory Server Console, select the Status tab.
2. In the navigation tree, expand the Log folder. There are three folders available, for the access,
error, and audit logs.
3. When you select the log type to view, a table displays a list of the last 25 entries in the selected
log.
4. Optionally, change the settings of the log display and click Refresh to update the display.
2. Tracking Server and Database Performance
17
The Select Log pull-down menu allows you to select an archived (rotated) log rather than
the currently active log.
The Lines to show text box changes the number of log entries to display in the window.
The Show only lines containing text box sets a filter, including regular expressions, to
use to display only certain matching log entries.
NOTE
Selecting the Continuous checkbox refreshes the log display automatically every ten seconds.=
Continuous log refresh does not work well with log files over 10 megabytes.
2.5. Replacing Log Files with a Named Pipe
Many administrators want to do some special configuration or operation with logging data, like
configuring an access log to record only certain events. This is not possible using the standard
Directory Server log file configuration attributes, but it is possible by sending the log data to a named
pipe, and then using another script or plug-in to process the data. Using a named pipe for the log
simplifies these special tasks, like:
Logging certain events, like failed bind attempts or connections from specific users or IP addresses
Logging entries which match a specific regular expression pattern
Keeping the log to a certain length (logging only the last number of lines)
Sending a notification, such as an email, when an event occurs
The basic format of the script is is:
ds-logpipe.py named_pipe [ --user pipe_user ] [ --maxlines number ] [[ --serverpidfile file.pid ] |
[ --serverpid PID ]] [ --servertimeout seconds ] [ --plugin=/path/to/plugin.py | [
pluginfile.arg=value ]]
More detailed usage information is in the Configuration, Command, and File Reference.
However, while that has the advantage of being simple to implement and not requiring any Directory
Server configuration changes, simply running the script has a big disadvantage: all of the log viewers in
Red Hat Directory Server 8.2 Performance Tuning Guide
18
the Directory Server Console and any script or tool (such as logconv.pl) that expect to access a real
file will fail.
If the Directory Server instance will permanently use the named pipe rather than a real file for logging,
then it is possible to reconfigure the instance to create the named pipe and use it for logging (as it does
by default for the log files). When the Directory Server instance is configured to use the named pipe then
all of the log analysis tools the Directory Server Console and any Directory Server scripts work
fine.
Three things need to be configured for the log configuration for the instance:
The log file to use has to be changed to the pipe (nsslapd-*log)
Buffering should be disabled because the script already buffers the log entries (nsslapd-*log-
logbuffering)
Log rotation should be disabled so that the server does not attempt to rotate the named pipe
(nsslapd-*log-maxlogsperdir, nsslapd-*log-logexpirationtime, and nsslapd-*log-
logrotationtime)
These configuration changes can be made in the Directory Server Console or using ldapmodify.
For example, this switches the access log to access.pipe:
/usr/lib64/mozldap/ldapmodify -D "cn=directory manager" -w secret -p 389 -h
server.example.com
dn: cn=config
changetype: modify
replace: nsslapd-accesslog
nsslapd-accesslog: /var/log/dirsrv/slapd-instance_name/access.pipe
-
replace: nsslapd-accesslog-logbuffering
nsslapd-accesslog-logbuffering: off
-
replace: nsslapd-accesslog-maxlogsperdir
nsslapd-accesslog-maxlogsperdir: 1
-
replace: nsslapd-accesslog-logexpirationtime
nsslapd-accesslog-logexpirationtime: -1
-
replace: nsslapd-accesslog-logrotationtime
nsslapd-accesslog-logexpirationtime: -1
TIP
Making these changes using the -f option will cause the server to close the current log file and
switch to the named pipe immediately. This can be very helpful for debugging a running server
and sifting the log ouput for specific messages.
2.6. Improving Logging Performance
Larger server deployments can generate several dozen of megabytes of logs per hour. Depending on
the resources available on the server host machine, reconfiguring or disabling access logging can
improve system and Directory Server performance.
2. Tracking Server and Database Performance
19
Before disabling access logging, first configure access log buffering. Buffering writes all log entries
directly to the disk, so that the Directory Server performance does not degrade even under a heavy load.
The access log is buffered by default, but make sure the log is using buffering for best performance.
/usr/lib64/mozldap/ldapmodify -D "cn=directory manager" -w secret -p 389
dn: cn=config
changetype: modify
replace: nsslapd-accesslog-logbuffering
nsslapd-accesslog-logbuffering: on
If that doesn't improve performance, then disable access logging entirely.
/usr/lib64/mozldap/ldapmodify -D "cn=directory manager" -w secret -p 389
dn: cn=config
changetype: modify
replace: nsslapd-accesslog-enabled
nsslapd-accesslog-enabled: off
WARNING
Access logging is extremely helpful for debugging issues in the server and monitoring client
connections and failed connection attempts. Don't disable access logging as the normal operating
environment.
For alternatives, see Section 2.5, Replacing Log Files with a Named Pipe, since using named
pipe log scripts can improve performance while still logging information on high performance
production servers.
3. Optimizing System Performance
There are a limited number of areas where settings on the machine on which the Directory Server
instance is running can be changed to improve Directory Server performance. Basically, the machine can
be configured to grant the Directory Server more system resources (TCP ports and file descriptors)
which helps it handle higher loads with better performance. The host machine can also be configured to
allow better client access to the server (firewalls and DNS), which doesn't have a direct impact on
Directory Server performance, but does affect using the server.
TIP
The Red Hat Enterprise Linux documentation includes a whitepaper on tuning the Red Hat
Enterprise Linux system settings to improve Oracle database performance and is available at
http://www.redhat.com/docs/manuals/enterprise/RHELTuningandOptimizationforOracleV11.pdf.
Similar adjustments can be made to any Red Hat Enterprise Linux system hosting a Red Hat
Directory Server instance, and that performance tuning whitepaper is a good resource to help
with Directory Server system tuning.
Red Hat Directory Server 8.2 Performance Tuning Guide
20
3.1. Tuning and Sizing System Memory
There are several different areas of the Red Hat Enterprise Linux machine that can be modified to
improve the performance of the Directory Server itself.
NOTE
The cache settings here relate to the cache allowed to the Directory Server process, not to any
cache settings in the Directory Server databases.
3.1.1. Setting the Minimum System Page Cache
Although Red Hat Enterprise Linux sets the page cache size dynamically, it is possible to require a
minimum number of free pages by setting the vm.min_free_kbytes parameter in the system
configuration. For example:
# echo vm.min_free_kbytes=1024 >> /etc/sysctl.conf
3.1.2. Managing the Swap Space
Long-running processes (and Directory Server is normally a long-running process) don't use all of the
pages that they obtain. Creating a swap space to partition the processes can improve the system's
performance because the system memory is used for disk cache rather than page cache.
The /proc/sys/vm/swappiness kernel parameter sets a threshold for when processes should be
swapped out in favor of I/O caching. The value of this is a percentage; the higher the percentage, the
higher the I/O cache and the faster that pages are swapped. The default value is 60. This can be
changed by editing the system configuration as root:
# echo vm.swappiness=10 >> /etc/sysctl.conf
Constant swapping must be avoided. If swap usage is high or increasing, then increase the RAM
(ideally) or the swap size.
3.1.3. Tuning for 32-Bit Systems
Ideally, Red Hat Directory Server will be run on 64-bit systems which naturally have the address space
to handle very large databases and operations. Still, 32-bit systems can be optimized to the best
possible performance.
First, use the hugemem kernel for databases that are 8GB or larger. This is the default kernel, anyway,
in Red Hat Enterprise Linux 5.
Next, configure huge pages on the system by setting how many large memory pages should be
allocated and kept in RAM for shared memory:
1. Check the amount of memory available for huge pages.
grep Hugepagesize /proc/meminfo
Hugepagesize: 4096 kB
This means that each huge page is 4GB.
2. Set the number of huge pages that can be used through the nr_hugepages kernel parameter.
3. Optimizing System Performance
21
# echo "vm.nr_hugepages=512" >> /etc/sysctl.conf
3. Check the new huge page settings.
The defined value for huge pages is viewable in Hugepagesize, while the total number of huge
pages (set in nr_hugepages) is shown in HugePages_Total, both in the /proc/meminfo:
HugePages_Total: 512
HugePages_Free: 512
HugePages_Rsvd: 0
Hugepagesize: 4096 kB
4. For the Directory Server databases to be able to use huge pages, add or set the memlock
parameter for the Directory Server user (nobody by default) to a high enough seting to allow them
to lock that much memory.
vim /etc/security/limits.conf
nobody soft memlock 1048576
nobody hard memlock 1048576
3.2. Using dsktune for Preliminary Tuning
When a Directory Server instance is first set up, the setup command automatically runs a tool called
dsktune which checks the current machine settings and compares them to the requirements for
Directory Server.
This tool can be run again at any point to check whether any system settings should be adjusted.
[jsmith@example ~]$ dsktune
Red Hat Directory Server system tuning analysis version 10-AUGUST-2007.
NOTICE : System is i686-unknown-linux2.6.18-164.2.1.el5 (2 processors).
NOTICE : The net.ipv4.tcp_keepalive_time is set to 7200000 milliseconds
(120 minutes). This may cause temporary server congestion from lost
client connections.
WARNING: There are only 1024 file descriptors (hard limit) available, which
limit the number of simultaneous connections.
WARNING: There are only 1024 file descriptors (soft limit) available, which
limit the number of simultaneous connections.
3.3. Tuning File Descriptors
The Directory Server is multi-threaded and efficiently handles many simultaneous requests and large
numbers of connections.
Most of the time, the only limit on Directory Server threads is whatever limit the operating system
imposes on the number of file descriptors available in the whole system and per process. The Directory
Server uses file descriptors in a range of directory operations logging, chaining, replication, client
connections, and indexes. Adjusting the number of file descriptors for the Directory Server, then, can
improve how well it handles large loads for a variety of operations.
Red Hat Directory Server 8.2 Performance Tuning Guide
22
3.3.1. Resetting the Host Machine's File Descriptors
1. First, check the current system limit for file descriptors:
cat /proc/sys/fs/file-max
2. If the setting is lower than 64000, edit the /etc/sysctl.conf file, and reset the fs.file-max
parameter:
fs.file-max = 64000
3. Then increase the maximum number of open files on the system by editing the
/etc/security/limits.conf configuration file. Add the following entry:
* - nofile 8192
4. Edit the /etc/pam.d/system-auth, and add this entry:
session required /lib/security/$ISA/pam_limits.so
5. Reboot the machine to apply the changes.
Additionally, the Directory Server keeps some file descriptors in reserve and is used by process like file
I/O (e.g. log files) and outgoing network connections. The number of simultaneous connections the
server can handle is set by the maximum number of server descriptors minus the assigned number of
reserve descriptors. The file descriptors parameters can be reset in the Directory Server configuration
to take advantage of the higher system settings.
3.3.2. Setting Directory Server File Descriptor Values
For performance, the most relevant attribute is the nsslapd-maxdescriptor attribute. This tells the
Directory Server how many file descriptors it has available. (This value cannot be larger than the system
setting, or the server ignores it.)
Reserve file descriptors rarely need to be reset. Decreasing the number of reserve file descriptors can
increase the number of LDAP clients that can connect to the server. However, too few reserve file
descriptors can still hirt directory performance by causing server errors. There are three instances that
can require raising the numbr of reserve file descriptors:
The server is replicating to a large number of consumer servers (more than 10), and/or the server is
maintaining a large number of index files (more than 30).
The server is servicing a large number of LDAP connections.
There are error messages reporting that the server is unable to open file descriptors (the actual
error message differs depending on the operation that the server is attempting to perform), but these
error messages are not related to managing client LDAP connections.
To calculate the number of reserved file descriptors to set, use the following formula:
nsslapd-reservedescriptor = 20 + (NldbmBackends * 4) + NglobalIndex +
ReplicationDescriptor + ChainingBackendDescriptors + PTADescriptors +
SSLDescriptors
NldbmBackends is the number of ldbm databases.
3. Optimizing System Performance
23
NglobalIndex is the total number of configured indexes for all databases including system indexes.
(By default 8 system indexes and 17 additional indexes per database).
ReplicationDescriptor is eight (8) plus the number of replicas in the server that can act as a supplier
or hub (NSupplierReplica).
ChainingBackendDescriptors is NchainingBackend times the nsOperationConnectionsLimit (a
chaining or database link configuration attribute; 10 by default).
PTADescriptors is 3 if PTA is configured and 0 if PTA is not configured.
SSLDescriptors is 5 (4 files + 1 listensocket) if SSL is configured and 0 if SSL is not configured.
To reset the file descriptors for Directory Server:
1. Edit the /etc/sysconfig/dirsrv file and add a new line setting the process ulimit.
ulimit -n 8192
2. Update the nsslapd-maxdescriptors and nsslapd-reservedescriptors values.
/usr/lib64/mozldap/ldapmodify -D "cn=directory manager" -w secret -p 389 -h
server.example.com
dn: cn=config
changetype: modify
replace: nsslapd-maxdescriptors
nsslapd-maxdescriptors: 8192
-
replace: nsslapd-reservedescriptors
nsslapd-reservedescriptors: 40
-
3. Restart the Directory Server.
service dirsrv restart
3.4. Setting Shell Limits for the Directory Server User
The Directory Server process runs as a regular user on the host system. (By default, that user is
nobody.) Once system settings have been changed, then the limits for that user can be updated.
TIP
The Red Hat Enterprise Linux documentation includes a whitepaper on tuning the Red Hat
Enterprise Linux system settings to improve Oracle database performance and is available at
http://www.redhat.com/docs/manuals/enterprise/RHELTuningandOptimizationforOracleV11.pdf.
Similar adjustments can be made to any Red Hat Enterprise Linux system hosting a Red Hat
Directory Server instance, and that performance tuning whitepaper is a good resource to help
with Directory Server system tuning.
3.4.1. Setting Shell Limits for File Descriptors
1. As root, open the system's /etc/security/limits.conf file.
2. Add two lines that set the hard and soft limits for the file descriptors (nofile) for the Directory
Red Hat Directory Server 8.2 Performance Tuning Guide
24
Server user. The soft limit sets how many file descriptors the user has available by default; the
user can manually adjust that setting until they hit the hard limit.
nobody soft nofile 4096
nobody hard nofile 63536
IMPORTANT
Do not set the hard limit for the Directory Server user equal to (or higher than) the
maximum number of file descriptors assigned to the system itself in
/proc/sys/fs/file-max. If the hard limit is too high and the user users all of those file
descriptors, then the entire system runs out of file descriptors.
3.4.2. Setting Shell Limits for the Max Number of Processes
1. As root, open the system's /etc/security/limits.conf file.
2. Add two lines that set the hard and soft limits for the number of processes (nproc) for the
Directory Server user. The soft limit sets how many processes the user has available by default;
the user can manually adjust that setting until they hit the hard limit.
nobody soft nproc 2047
nobody hard nproc 16384
IMPORTANT
Do not set the hard limit for the Directory Server user equal to (or higher than) the
maximum number of file descriptors assigned to the system itself in
/proc/sys/fs/file-max. If the hard limit is too high and the user users all of those file
descriptors, then the entire system runs out of file descriptors.
3.5. Tuning TCP Ports
If there is high traffic on a server like the same machine is running both a web server and a Directory
Server or there are a lot of clients who access the Directory Server directly then it may be necessary
to broaden the range of TCP ports available on the machine.
echo "1024 65000" > /proc/sys/net/ipv4/ip_local_port_range
3.6. Changing Firewall Settings for the Directory Server Console
When remote clients are access the Directory Server instance, particularly the Directory Server Console
or applications like Red Hat Certificate System, then make sure that the iptables configuration allows
those connections past the firewall.
For example, if the instance is using the default ports and TLS/SSL is configured, then there are three
ports that need to be opened in iptables.
1. Open the iptables file.
3. Optimizing System Performance
25
# vi /etc/sysconfig/iptables
2. Add a line to the configuration for each port that is in use for the Directory Server:
-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 389 -j
ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 636 -j
ACCEPT
-A RH-Firewall-1-INPUT -m state --state NEW -m tcp -p tcp --dport 9830 -j
ACCEPT
NOTE
Only add the TLS/SSL port (636) if TLS/SSL is actually enabled for the server.
3. Restart the firewall:
# service iptables restart
3.7. DNS Requirements
It is very important that DNS and reverse DNS be working correctly on the host machine, especially if you
are using TLS/SSL or Kerberos with Directory Server.
Configure the DNS resolver and the NIS domain name by the modifying the /etc/resolv.conf,
/etc/nsswitch.conf, and /etc/netconfig files, and set the DNS resolver for name resolution.
Edit the /etc/defaultdomain file to include the NIS domain name. This ensures that the fully-
qualified host and domain names used for the Directory Server resolve to a valid IP address and that
that IP address resolves back to the correct hostname.
Reboot the machine to apply these changes.
4. Improving Search Performance (and Balancing Read
Performance)
The most effective way to improve search operations against the directory is to configure thorough
indexes for entries, combined with reasonable limits on search results.
4.1. Using Indexes
An index (as it implies) is a tag that shows that a certain entry contains a certain attribute, without having
to contain any other detail about the entry (which saves space and makes returning search results
faster). Each index is organized around a Directory Server attribute and a certain way of matching that
attribute:
Presence index (pres) simply shows what entries contain an attribute.
Equality index (eq) shows which attribute values match a specific search string.
Approximate index (approx) is used for efficient sounds-like searches, which shows entries which
have a value that phonetically matches a string.
Substring index (sub) matches any substring of an attribute value to the given search string. (This
Red Hat Directory Server 8.2 Performance Tuning Guide
26
index if very expensive for the server to maintain.)
International index uses a matching rule to match strings in a directory which contains values in
languages other than English.
Browsing index, or virtual list view (VLV) index, sets an index to use to display entries in the Directory
Server Console.
NOTE
Indexing is described in much more detail in the Administrator's Guide.
However, just creating indexes is not ipso facto going to increase server performance. Maintaining
indexes puts a burden on the Directory Server for every modify, add, and delete operation by having to
verify every attribute in the change against every index maintained by the server:
1. The Directory Server receives an add or modify operation.
2. The Directory Server examines the indexing attributes to determine whether an index is
maintained for the attribute values.
3. If the created attribute values are indexed, then the Directory Server generates the new index
entries.
4. Once the server completes the indexing, the actual attribute values are created according to the
client request.
For example, the Directory Server adds the entry:
dn: cn=John Doe, ou=People,dc=example,dc=com
objectclass: top
objectClass: person
objectClass: orgperson
objectClass: inetorgperson
cn: John Doe
cn: John
sn: Doe
ou: Manufacturing
ou: people
telephoneNumber: 408 555 8834
description: Manufacturing lead for the Z238 line of widgets.
The Directory Server is maintaining the following indexes:
Equality, approximate, and substring indexes for cn (common name) and sn (surname) attributes.
Equality and substring indexes for the telephone number attribute.
Substring indexes for the description attribute.
When adding that entry to the directory, the Directory Server must perform these steps:
1. Create the cn equality index entry for John and John Doe.
2. Create the appropriate cn approximate index entries for John and John Doe.
3. Create the appropriate cn substring index entries for John and John Doe.
4. Create the sn equality index entry for Doe.
5. Create the appropriate sn approximate index entry for Doe.
4. Improving Search Performance (and Balancing Read Performance)
27
6. Create the appropriate sn substring index entries for Doe.
7. Create the telephone number equality index entry for 408 555 8834.
8. Create the appropriate telephone number substring index entries for 408 555 8834.
9. Create the appropriate description substring index entries for Manufacturing lead for the
Z238 line of widgets. A large number of substring entries are generated for this string.
Before creating new indexes, make sure to balance the overhead of maintaining the indexes against the
potential improvements in search performance. Especially important, match the types of indexes that you
maintain to the type of information stored in the directory and the type of information users routinely
search for.
Approximate indexes are not efficient for attributes commonly containing numbers, such as telephone
numbers.
Substring indexes do not work for binary attributes.
Equality indexes should be avoided if the value is big (such as attributes intended to contain
photographs or passwords containing encrypted data).
Maintaining indexes for attributes not commonly used in a search increases overhead without
improving global searching performance.
Attributes that are not indexed can still be specified in search requests, although the search
performance may be degraded significantly, depending on the type of search.
The more indexes you maintain, the more disk space you require.
TIP
Creating indexes is much more effective for directories which have a high search operation load
and low modify operation load.
4.2. Tuning Directory Server Resource Settings
The server's performance can be managed and improved by limiting the amount of resources the server
uses to process client search requests, which is done by defining four settings:
The maximum number of entries the server returns to the client in response to a search operation
(size limit attribute).
The maximum amount of real time (in seconds) for the server to spend performing a search request
(time limit attribute).
The time (in seconds) during which the server maintains an idle connection before terminating it (idle
timeout attribute).
The maximum number of file descriptors available to the Directory Server (max number of file
descriptors attribute).
To configure Directory Server to optimize performance:
1. In the Directory Server Console, select the Configuration tab, and then select the topmost
entry in the navigation tree in the left pane.
2. Select the Performance tab in the right pane.
Red Hat Directory Server 8.2 Performance Tuning Guide
28
3. Set the maximum number of entries the server will return to the client in response to a search
operation by entering a new value in the Size Limit text box.
To keep from setting a limit, type -1 in this text box.
4. Enter the maximum amount of real time (in seconds) for the server to spend performing a search
request in the Time Limit text box.
To keep from setting a limit, type -1 in this text box.
5. Enter the time (in seconds) for the server to maintain an idle connection before terminating it in the
Idle Timeout text box.
To keep from setting a limit, type zero (0) in this text box.
6. Set the maximum number of file descriptors available to the Directory Server in the Max Number
of File Descriptors text box. For more information on this parameter, see the Directory
Server Configuration, Command, and File Reference.
4.3. Setting Index Scan Limits
In large directories, the search results list can get huge. A directory with a million inetorgperson
entries would have a million entries that were returned with a filter like
(objectclass=inetorgperson), and an index for the sn attribute would have at least a million
entries in it.
Loading a long ID list from the database significantly reduces search performance. The configuration
parameter, nsslapd-idlistscanlimit, sets a limit on the number of IDs that are read before a key
is considered to match the entire primary index (meaning the search is treated as an unidexed search
with a different set of resource limits).
For large indexes, it is actually more efficient to treat any search which matches the index as an
unindexed search. The search operation only has to look in one place to process results (the entire
directory) rather than searching through an index that is nearly the size of a directory, plus the directory
itself.
The default value of the nsslapd-idlistscanlimit attribute is 4000, which is gives good
performance for a common range of database sizes and access patterns. It's usually not necessary to
change this value. If the database index is slightly larger than the 4000 entries, but still significantly
smaller than the overall directory, then raising the scan limit improves searches which would otherwise
hit the default limit of 4000.
On the other hand, lowering the limit can significantly speed up searches that would otherwise hit the
4000 entry limit, but where it's not necessary to scan every entry.
4.4. Tuning the Database Cache for Searches
4. Improving Search Performance (and Balancing Read Performance)
29
The database attributes that affect search performance mainly define the amount of memory available to
the server. The maximum values that can be set for the database's cache size attributes depends on
the amount of real memory on the machine. Roughly, the amount of available memory on the machine
should always be greater than sum total of the default database cache size and sum of each entry
cache size.
Use caution when changing these cache sizing attributes. The ability to improve server performance
with these attributes depends on the size of the database, the amount of physical memory available on
the machine, and whether directory searches are random (that is, if the directory clients are searching
for random and widely scattered directory data).
If the database does not fit into memory and if searches are random, attempting to increase the values
set on these attributes does not help directory performance. In fact, changing these attributes may harm
overall performance.
The attributes of each database used to store directory data, including the server configuration data in
the NetscapeRoot database, can be resized.
To improve the cache hit ratio on search operations, increase the amount of data that the Directory
Server maintains in the database cache, as described in Section 6.2, Tuning Database Cache, by
editing the values for the nsslapd-dbcachesize attribute.
4.5. Tuning the Database Settings for Searches
The Directory Server Console only shows the databases that contain the directory data and the
NetscapeRoot database. However, the server uses another database to manage these. On this
database, the following attributes can be changed to improve performance:
The amount of memory to make available for all databases (maximum cache size), which is
described in Section 6.1, Tuning Entry Cache.
The maximum number of entries for the server to verify in response to a search request (look-
through limit).
The amount of memory to make available for import (import cache size).
To configure the default database attributes that apply to all other database instances:
1. In the Directory Server Console, select the Configuration tab; then, in the navigation tree,
expand the Data icon, and highlight the Database Settings node.
2. Select the LDBM Plug-in Settings tab in the right pane.
This tab contains the database attributes for all databases stored on this server.
3. In the Maximum Cache Size field, enter a value corresponding to the amount of memory to
Red Hat Directory Server 8.2 Performance Tuning Guide
30
make available for all databases. This value is for the total of the entire backend, meaning all
databases cumulatively rather than the amount per single database instance.
4. In the Look-Through Limit field, enter the maximum number of entries for the server to check
in response to a search request.
5. There are two ways to set the amount of memory in bytes to make available for import. The default
is to have auto cache sizing, meaning 50% of the free memory is allocated for the import cache. It
is also possible to set the import cache size manually by deselecting the Use Cache Auto-
Size checkbox and then setting the value in the Import Cache Size field. For creating a very
large database from LDIF, set this attribute as large as possible, depending on the memory
available on the machine. The larger this parameter, the faster the database is created.
WARNING
Setting this value too high can cause import failures because of a lack of memory.
4.6. Managing Special Entries
The cn=config entry in the simple, flat dse.ldif configuration file is not stored in the same highly
scalable database as regular entries. As a result, if many entries, particularly entries that are likely to be
updated frequently, are stored under cn=config, performance will probably suffer.
Although Red Hat recommends that simple user entries not be stored under cn=config for
performance reasons, it can be useful to store special user entries such as the Directory Manager entry
or replication manager (supplier bind DN) entry under cn=config since this centralizes configuration
information.
5. Tuning Transaction Logging
Every Directory Server contains a transaction log which writes operations for all the databases it
manages. Whenever a directory database operation such as a modify is performed, the server creates a
single database transaction for all of the database operations invoked as a result of that LDAP
operation. This includes both updating the entry data in the entry index file and updating all of the
attribute indexes. If all of the operations succeed, the server commits the transaction, writes the
operations to the transaction log, and verifies that the entire transaction is written to disk. If any of these
operations fail, the server rolls back the transaction, and all of the operations are discarded. This all-or-
nothing approach in the server guarantees that an update operation is atomic. Either the entire
operation succeeds permanently and irrevocably, or it fails.
Periodically, the Directory Server (through internal housekeeping threads) flushes the contents of the
transaction logs to the actual database index files and checks if the transaction logs require trimming.
If the server experiences a failure, such as a power outage, and shuts down abnormally, the information
about recent directory changes is still saved by the transaction log. When the server restarts, the
directory automatically detects the error condition and uses the database transaction log to recover the
database.
Although database transaction logging and database recovery are automatic processes that require no
intervention, it can be advisable to tune some of the database transaction logging attributes to optimize
performance.
5. Tuning Transaction Logging
31
WARNING
The transaction logging attributes are provided only for system modifications and diagnostics.
These settings should be changed only with the guidance of Red Hat Technical Support. Setting
these attributes and other configuration attributes inconsistently may cause the directory to be
unstable.
5.1. Changing the Location of the Database Transaction Log
By default, the database transaction log file is stored in the
/var/lib/dirsrv/slapd-instance_name/db directory along with the database files themselves.
Because the purpose of the transaction log is to aid in the recovery of a directory database that was
shut down abnormally, it is a good idea to store the database transaction log on a different disk from the
one containing the directory database. Storing the database transaction log on a separate physical disk
may also improve directory performance.
To change the location of the database transaction log:
1. Stop Directory Server.
service dirsrv stop [instance_name]
2. Edit the dse.ldif file and change the nsslapd-db-logdirectory attribute to show the new log
file location. Provide the full path to the log directory in the attribute.
For information on the nsslapd-db-logdirectory attribute syntax, see the Directory Server
Configuration, Command, and File Reference.
3. Copy the existing transation logs and the DBVERSION file into the new transaction log directory.
cp /var/lib/dirsrv/slapd-instance_name/db/{log.*,DBVERSION}
/path/to/newdirectory
4. Restart Directory Server.
service dirsrv start [instance_name]
5.2. Changing the Database Checkpoint Interval
At regular intervals, the Directory Server writes operations logged in the transaction log to the database
index files and logs a checkpoint entry in the database transaction log. By indicating which changes
have already been written to the database indexes, checkpoint entries indicate where to begin recovery
from the transaction log, thus speeding up the recovery process.
By default, the Directory Server is set up to send a checkpoint entry to the database transaction log
every 60 seconds. Increasing the checkpoint interval may increase the performance of directory write
operations. However, increasing the checkpoint interval may also increase the amount of time required to
recover directory databases after a disorderly shutdown and require more disk space due to large
database transaction log files. Therefore, only modify this attribute if you are familiar with database
optimization and can fully assess the effect of the change.
To modify the checkpoint interval while the server is running, use the ldapmodify command-line utility
to add the nsslapd-db-checkpoint-interval attribute to the cn=config,cn=ldbm
Red Hat Directory Server 8.2 Performance Tuning Guide
32
database,cn=plugins,cn=config entry.
For more information on the syntax of the nsslapd-db-checkpoint-interval attribute, refer to the
Directory Server Configuration, Command, and File Reference.
5.3. Disabling Durable Transactions
Durable transaction logging means that each LDAP update operation, comprised of a sequence of
database operations in a transaction, is physically written to disk. Even though each LDAP operation can
be comprised of multiple database operations, each LDAP operation is treated as a single database
transation. Each LDAP operation is both atomic and durable.
WARNING
Turning off durable transactions can improve Directory Server write performance with the risk of
data loss.
When durable transaction logging is disabled, every directory database operation is written to the
database transaction log file but may not be physically written to disk immediately. If a directory change
was written to the logical database transaction log file but not physically written to disk at the time of a
system crash, the change cannot be recovered. When durable transactions are disabled, the recovered
database is consistent but does not reflect the results of any LDAP write operations that completed just
before the system crash.
By default, durable database transaction logging is enabled. To disable durable transaction logging:
1. Use the ldapmodify command-line utility to add the nsslapd-db-durable-transactions
attribute to the cn=config,cn=ldbm database,cn=plugins,cn=config entry, and set
the value of this attribute to off.
For information on the syntax of the nsslapd-db-durable-transactions attribute, see the
Directory Server Configuration, Command, and File Reference.
2. Restart the Directory Server.
service dirsrv restart
5.4. Specifying Transaction Batching
To improve update performance when full transaction durability is not required, use the nsslapd-db-
transaction-batch-val attribute to specify how many transactions will be batched before being
committed to the transaction log. Setting this attribute to a value of greater than 0 causes the server to
delay committing transactions until the number of queued transactions is equal to the attribute value.
This is similar to disabling durable transaction logging (in the nsslapd-db-durable-transaction
attribute), but setting the batch value gives more control over how many transactions can be potentially
lost.
To specify or modify transaction batching while the server is running, use the ldapmodify command-
line utility to add the nsslapd-db-transaction-batch-val attribute to the cn=config,cn=ldbm
database,cn=plugins,cn=config entry.
For more information on the syntax and values of the nsslapd-db-transaction-batch-val attribute,
refer to the Directory Server Configuration, Command, and File Reference.
5. Tuning Transaction Logging
33
6. Managing the Database Cache Settings
The best place to maintain data (despite some popular opinion) is the database cache, not the
filesystem cache. Pages the database cache are used directly from process memory, while storing
pages in the filesystem cache incurs a higher processing overhead, with a user/kernel transition and a
memory copy of the payload.
(Storing pages in the filesystem cache made some sense on older 32-bit systems, when the filesystem
cache was able to use 64-bit address spaces.)
Still, using the process cache provide much better performance, especially for larger directories.
The Directory Server directly uses two kinds of cache for its pages:
Entry cache, which contains individual directory entries
Database cache, which contains database index files (.db4 files)
The entry cache size impacts search and read performance for the server, so high-traffic systems
should have higher entry cache settings. The database cache size can affect searches against
database indexes.
Both the entry cache and database cache access the system's memory (RAM).
NOTE
The database cache is memory mapped. Some systems don't handle mmapping very well, so to
avoid any issues, put the database home directory in tmpfs.
6.1. Tuning Entry Cache
Entry cache is the memory (RAM) used to store directory entries in the internal representation. This
cache is referenced for search and read operations by the Directory Server. Good settings for the entry
cache can improve search performance in the Directory Server. As a rule of thumb, more memory
assigned to the entry cache, the better. At a minimum, the Directory Server entry cache must have
enough memory to handle its regular amount of use with additional memory to cover most spikes in use
or other events.
For the best search performance, make the entry cache large enough to contain all entries in the
database.
If entry caching is not set, then the Directory Server reads the entry straight from the entry index for the
database (id2entry.db4). This operation involves disk I/O to find the entry and then converts the
entry from the on-disk format (LDIF) to the in-memory representation (Slapi_Entry). Keeping an entry
cache allows the server to skip the disk I/O and conversion steps in directory searches.
When setting this value, monitor the entry cache usage.
To set the entry cache size from the Directory Server Console:
1. Select the Configuration tab and, in the navigation tree, expand the Data icon.
2. Expand the suffix associated with the database, such as dc=example,dc=com, and then select
the database.
Red Hat Directory Server 8.2 Performance Tuning Guide
34
3. In the Database Settings tab in the main window, the Memory available for cache
field sets the amount of memory available, in bytes, that is assigned for the entry cache.
This value can be modified from the command line by editing the nsslapd-cachememsize attribute
value for the LDBM plug-in instance for the database. For example:
/usr/lib64/mozldap/ldapmodify -D "cn=directory manager" -w secret -p 389
dn: cn=database_name,cn=ldbm database,cn=plugins,cn=config
changetype: modify
replace: nsslapd-cachememsize
nsslapd-cachememsize: 20971520
6.2. Tuning Database Cache
The database cache contains the Berkeley database index files for the database, meaning all of the
*.db4 and other files used for attribute indexing by the database. This value is passed to the Berkeley
DB API function set_cachesize.
This cache size has less of an impact on Directory Server performance than the entry cache size, but if
there is available RAM after the entry cache size is set, then increase the amount of memory allotted to
the database cache.
NOTE
The operating system also has a file system cache which may compete with the database cache
for RAM usage. Refer to the operating system documentation to find information on file system
cache settings and monitoring the file system cache.
The Berkeley database documentation for the set_cachesize function gives a good, solid description of
the functions and behavior of the database cache parameter, nsslapd-dbcachesize.
When setting the database cache value, monitor the cache usage.
To set the database cache size from the Directory Server Console:
1. Select the Configuration tab and, in the navigation tree, expand the Data icon.
2. Select the Database Settings node, and then open the LDBM Plug-in Settings tab.
3. The database cache size is set in the Maximum cache size field. This sets the allotted cache
size for the entire database and all its index files in bytes.
6. Managing the Database Cache Settings
35
Attempting to set a value that is not a number or is too big for a 32-bit signed integer returns an
LDAP_UNWILLING_TO_PERFORM error message with additional error information explaining the
problem.
4. Restart the server to apply the changes.
service dirsrv restart
This value can be modified from the command line by editing the nsslapd-dbcachesize attribute value
for the LDBM plug-in configuration. For example:
/usr/lib64/mozldap/ldapmodify -D "cn=directory manager" -w secret -p 389
dn: cn=config,cn=ldbm database,cn=plugins,cn=config
changetype: modify
replace: nsslapd-dbcachesize
nsslapd-dbcachesize: 20971520
Be sure to restart the server to apply the changes.
7. Tuning Database Link Performance
Database link performance can be improved through changes to the Directory Server's connection and
thread management.
7.1. Managing Connections to the Remote Server
Each database link maintains a pool of connections to a remote server. The connections to optimize
resources can be configured for the directory.
7.1.1. Managing Connections to the Remote Server Using the Console
1. Select the Configuration tab, expand the Data folder in the left pane, and select the suffix,
then the database link to change.
Red Hat Directory Server 8.2 Performance Tuning Guide
36
2. Click the Limits and Controls tab in the right navigation pane.
3. In the Connection Management section, make changes to any of the following fields:
Maximum TCP connection(s). The maximum number of TCP connections that the database
link establishes with the remote server. The default value is 3 connections.
Bind timeout. Amount of time, in seconds, before the database link's bind attempt times out.
The default value is 15 seconds.
Maximum binds per connection. Maximum number of outstanding bind operations per TCP
connection. The default value is 10 outstanding bind operations per connection.
Time out before abandon (sec). Number of seconds before the server checks to see if a
timed-out connection should be abandoned. The default value is 1 second.
Maximum LDAP connection(s). Maximum number of LDAP connections that the database link
establishes with the remote server. The default value is 10 connections.
Maximum bind retries. Number of times a database link attempts to bind to the remote server.
A value of 0 indicates that the database link will try to bind only once. The default value is 3
attempts.
Maximum operations per connection. Maximum number of outstanding operations per LDAP
connection. The default value is 2 operations per connection.
Connection lifetime (sec). How long a connection made between the database link and remote
server remains open. Connections between the database link and the remote server can be
kept open for an unspecified time or closed after a specific period of time. It is faster to keep
7. Tuning Database Link Performance
37
the connections open, but it uses more resources. For slow connections, it may be desirable to
limit the connection time. A value of 0 indicates there is no limit. By default, the value is set to 0.
7.1.2. Managing Connections to the Remote Server from the Command Line
Use ldapmodify to add connection attributes to the database link entry.
The default connection management attributes are stored in the following entry:
cn=default instance config,cn=chaining database,cn=plugins,cn=config
The connection management attributes for a specific database link are stored in the following entry:
cn=database_link,cn=chaining database,cn=plugins,cn=config
The connection management attributes specified in this entry take precedence over the attributes
specified in the cn=default instance config entry.
Red Hat Directory Server 8.2 Performance Tuning Guide
38
Table 11. Database Link Connection Management Attributes
Attribute Name Description
nsOperationConnectionsLimit Maximum number of LDAP connections that the
database link establishes with the remote server.
The default value is 20 connections per
database link instance.
nsBindConnectionsLimit Maximum number of TCP connections that the
database link establishes with the remote server.
The default value is 3 connections.
nsConcurrentOperationsLimit Maximum number of outstanding operations per
LDAP connection. The default value is 2
operations per connection.
nsConcurrentBindLimit Maximum number of outstanding bind operations
per TCP connection. The default value is 10
outstanding bind operations.
nsBindRetryLimit Number of times a database link attempts to bind
to the remote server. A value of zero (0) indicates
that the database link will try to bind only once.
The default value is 3 attempts.
nsConnectionLife Connection lifetime, in seconds. Connections
between the database link and the remote server
can be kept open for an unspecified time or
closed after a specific period of time. It is faster to
keep the connections open, but it uses more
resources. For example, it may be wise to limit the
connection time for a slow connection. A value of
0 indicates there is no limit. By default, the value
is set to 0. When the value is 0 and there is a list
of failover servers in the nsFarmServerURL
attribute, the first server is never contacted after
failover to the alternate server. The default value
is 0 seconds.
nsBindTimeout Amount of time, in seconds, before the bind
attempt times out. The default value is 15
seconds.
nsAbandonedSearchCheckInterval Number of seconds that pass before the server
checks for abandoned operations. The default
value is 1 second.
7.2. Detecting Errors During Normal Processing
Protect server performance by detecting errors during the normal chaining operation between the
database link and the remote server. The database link has two attributes nsMaxResponseDelay and
nsMaxTestResponseDelay which work together to determine if the remote server is no longer
responding.
The first attribute, nsMaxResponseDelay, sets a maximum duration for an LDAP operation to complete. If
the operation takes more than the amount of time specified in this attribute, the database link's server
7. Tuning Database Link Performance
39
suspects that the remote server is no longer online.
Once the nsMaxResponseDelay period has been met, the database link pings the remote server. During
the ping, the database link issues another LDAP request, a simple search request for an object that
does not exist in the remote server. The duration of the ping is set using the
nsMaxTestResponseDelay.
If the remote server does not respond before the nsMaxTestResponseDelay period has passed, then
an error is returned, and the connection is flagged as down. All connections between the database link
and remote server will be blocked for 30 seconds, protecting the server from a performance degradation.
After 30 seconds, operation requests made by the database link to the remote server continue as
normal.
Both attributes are stored in the cn=config,cn=chaining database,cn=plugins,cn=config
entry. The following table describes the attributes in more detail:
Table 12. Database Link Processing Error Detection Parameters
Attribute Name Description
nsMaxResponseDelay Maximum amount of time it can take a remote
server to respond to an LDAP operation request
made by a database link before an error is
suspected. This period is given in seconds. The
default delay period is 60 seconds. Once this
delay period has been met, the database link
tests the connection with the remote server.
nsMaxTestResponseDelay Duration of the test issued by the database link to
check whether the remote server is responding. If
a response from the remote server is not
returned before this period has passed, the
database link assumes the remote server is
down, and the connection is not used for
subsequent operations. This period is given in
seconds. The default test response delay period
is 15 seconds.
7.3. Managing Threaded Operations
Generally, Directory Server performs best using a limited number of threads for processing operations. A
limited number of threads can generally process operations very quickly, preventing the queue of
operations waiting for a free thread from growing too long.
On multi-core or multi-CPU machines, a good starting point is to use two threads per core or CPU. This
gives a good balance between performance and concurrency. If performance becomes an issue, or
clients are not being served as quickly as possible, increase the number of threads.
However, the database link forwards operations to remote servers for processing. The database link
contacts the remote server, forwards the operation, waits for the result, and then sends the result back
to the client application. The entire operation can take much longer than a local operation.
While the database link waits for results from the remote server, it can process additional operations.
When using database links, performance can be improved by increasing the number of threads available
Red Hat Directory Server 8.2 Performance Tuning Guide
40
for processing operations. While the local CPU waits for a response from a remote server, it can
process other operations rather than stand idle.
To change the number of threads used for processing operations, change the nsslapd-threadnumber
global configuration attribute in the cn=config entry. The default thread number is 30. Restart the
server after changing the thread count to apply the changes.
[1] The LDAP tool s referenced i n thi s gui de are Mozi l l a LDAP, i nstal l ed wi th Di rectory Server i n the /usr/lib64/mozldap
di rectory on Red Hat Enterpri se Li nux 5 (64-bi t). However, Red Hat Enterpri se Li nux systems al so i ncl ude LDAP tool s from
OpenLDAP. It i s possi bl e to use the OpenLDAP commands as shown i n the exampl es, but you must use the -x argument to
di sabl e SASL and al l ow si mpl e authenti cati on.
7. Tuning Database Link Performance
41