Read This Document Before Attempting To Install Or Use This...

LifeKeeper for Linux v4.1.0 Release Notes

Revision G

03/2002

Read This Document Before Attempting To Install Or Use This Product!

This document contains information about factors that must be

considered before, during, and after installation. This document and the information herein is the property of SteelEye Technology, Inc. and all unauthorized use and reproduction is prohibited. SteelEye Technology, Inc. makes no warranties with respect to the contents of this document. SteelEye Technology, Inc. reserves the right to revise this publication and make changes to the products described herein without prior notification.

LifeKeeper is a registered trademark and SteelEye is a trademark of SteelEye Technology, Inc. Caldera Systems is a registered trademark of Caldera Systems, Inc. DB2 Universal Database is a trademark of International Business Machines Corporation. Informix is a registered trademark of Informix Software, Inc. Java is a trademark of Sun Microsystems, Inc. Linux is a registered trademark of Linus Torvalds. Microsoft Internet Explorer is a registered trademark of Microsoft Corporation. Netscape and Netscape Navigator are registered trademarks of Netscape Communications Corporation in the United States and other countries. Oracle is a registered trademark of Oracle Corporation. Sendmail is a trademark of Sendmail, Inc. SuSE is a registered trademark of SuSE AG Turbolinux is a trademark of Turbolinux, Inc. Other brand and product names used herein are for identification purposes only and may be trademarks of their respective companies. It is the policy of SteelEye Technology, Inc. to improve products as new technology, components, software, and firmware become available. SteelEye Technology, Inc., therefore, reserves the right to change specifications without prior notice. To maintain the quality of our publications, we need your comments on the accuracy, clarity, organization, and value of this book. Address correspondence to: [email protected] Copyright © 2000-2002 By SteelEye Technology, Inc. Mountain View, CA U.S.A. All Rights Reserved

Table of Contents

Release Notes 3

Table of Contents Table of Contents .......................................................................................................................................................... 3 Introduction................................................................................................................................................................... 4 LifeKeeper Product Description ................................................................................................................................... 4

LifeKeeper Core .................................................................................................................................................... 5 LifeKeeper Optional Recovery Software............................................................................................................... 6

New Features in LifeKeeper v4.1.0 .............................................................................................................................. 7 LifeKeeper Product Requirements ................................................................................................................................ 8

LifeKeeper Optional Recovery Software............................................................................................................... 9 GUI Platforms and Browsers ............................................................................................................................... 10 Storage and Adapter Options * ............................................................................................................................ 11

Installation and Configuration .................................................................................................................................... 12 Upgrades ..................................................................................................................................................................... 13

Before Starting Your Upgrade ............................................................................................................................. 13 Upgrading from LifeKeeper v3.1 to v4.1.0.......................................................................................................... 13 Upgrading to Installation Support v4.1.0-9 or later ............................................................................................. 13

Technical Notes .......................................................................................................................................................... 14 LifeKeeper Features............................................................................................................................................. 14 Tuning.................................................................................................................................................................. 15 LifeKeeper Operations......................................................................................................................................... 15 Server Configuration............................................................................................................................................ 16 Storage and Adapter Configuration ..................................................................................................................... 16 Linux Configuration ............................................................................................................................................ 17 Network Configuration ........................................................................................................................................ 21 Application Configuration ................................................................................................................................... 22 GUI Configuration............................................................................................................................................... 23 GUI Limitations................................................................................................................................................... 25

Restrictions or Known Issues...................................................................................................................................... 26 Installation ........................................................................................................................................................... 26 LifeKeeper Core .................................................................................................................................................. 28 GUI ...................................................................................................................................................................... 28 Data Replication .................................................................................................................................................. 30 Oracle Recovery Kit ............................................................................................................................................ 30 Apache Recovery Kit........................................................................................................................................... 31 Print Services Recovery Kit................................................................................................................................. 31 NFS Server Recovery Kit .................................................................................................................................... 32 NAS Recovery Kit ............................................................................................................................................... 32

Documentation and Training ...................................................................................................................................... 33 LifeKeeper for Linux Documentation List .......................................................................................................... 33

LifeKeeper Documentation CD .................................................................................................................... 34 Man Pages..................................................................................................................................................... 34 Third Party Documentation .......................................................................................................................... 34

Training................................................................................................................................................................ 34 Support........................................................................................................................................................................ 35 Software Updates ........................................................................................................................................................ 35

Introduction

4 Release Notes

Introduction This release notes document is written for the person who installs, configures and/or administers the LifeKeeper for Linux product. The document contains important information not detailed in the formal LifeKeeper documentation set, such as package versions, last-minute changes to instructions and procedures, product restrictions, and troubleshooting hints and tips that were discovered through final product testing. It is important that you review this document before installing and configuring your LifeKeeper software.

LifeKeeper Product Description The LifeKeeper product includes fault detection and recovery software that provides high availability for file systems, network addresses, applications, and processes running on Linux. LifeKeeper supports the configuration and switchover of a given application across multiple servers. The servers on which the application is configured are assigned priorities to determine the sequence in which the application will move from server to server in the event of multiple failures.

LifeKeeper for Linux provides switchover protection for a range of system resources. Automatic recovery is supported for the following resource types.

• Processes and Applications • Shared Storage Devices • Network Attached Storage Devices • File Systems (ext2, ext3, reiserfs and nfs) • Communication Resources (TCP/IP) • Database Applications (Oracle, Informix Dynamic Server, MySQL, DB2) • Web Server Resources (Apache, Apache SSL) • Mail Server Resources (Sendmail, SAMS, SMMS) • Print Services (lpd)

LifeKeeper Product Description

LifeKeeper for Linux v4.1.0 5

LifeKeeper Core

LifeKeeper for Linux is bundled. The LifeKeeper Core Package Cluster includes the following installable packages:

Package Package Name Description

LifeKeeper steeleye-lk-4.1.0-4.i386.rpm The LifeKeeper package provides recovery software for failures associated with core system components such as memory, CPUs, the operating system, the SCSI disk subsystem, and file systems.

LifeKeeper GUI steeleye-lkGUI-4.1.0-4.i386.rpm The LifeKeeper GUI package provides a graphical user interface for LifeKeeper administration and status monitoring.

LifeKeeper IP Recovery Kit

steeleye-lkIP-4.1.0-4.i386.rpm The LifeKeeper IP Recovery Kit provides recovery software for automatic switchover of IP addresses.

LifeKeeper Raw I/O Recovery Kit

steeleye-lkRAW-4.1.0-4.i386.rpm Provides support for applications that use raw i/o to bypass kernel buffering.

LifeKeeper CCISS Recovery Kit

steeleye-lkCCISS-4.1.0-4.i386.rpm

Provides support for Compaq cciss devices.

LifeKeeper Man Pages

steeleye-lkMAN-4.1.0-4.i386.rpm The LifeKeeper Man Page package provides reference manual pages for the LifeKeeper product.

LifeKeeper Online Product Manual

steeleye-lkHLP-4.1.0-4.i386.rpm The LifeKeeper Online Product Manual (help) package provides HTML-based product. The manual includes product feature information and instructions for the LifeKeeper core administration tasks.

LifeKeeper Product Description

6 Release Notes

LifeKeeper Optional Recovery Software

The following optional software provides resource definition and recovery software for the application versions listed:

Package Package Name Protected Applications

LifeKeeper DB2 Recovery Kit

steeleye-lkDB2-4.0.0-2.i386.rpm IBM DB2 Universal Database (UDB) Enterprise Edition (EE) v7.1 and Enterprise Extended Edition (EEE) v7.1 and 7.2.

LifeKeeper Oracle Database Recovery Kit

steeleye-lkORA-4.1.0-1.i386.rpm Oracle 8.0.5 & 8i RDBMS products and Oracle 9i Standard Edition.

LifeKeeper Informix Database Recovery Kit

steeleye-lkINF-3.01-1.i386.rpm Informix Dynamic Server 9.2 products

LifeKeeper MySQL Recovery Kit

steeleye-lkSQL-3.01-1.i386.rpm MySQL 3.23

LifeKeeper Apache Web Server Recovery Kit

steeleye-lkAPA-3.01-2.i386.rpm Apache Web Server 1.3, and SecureWeb Server 3.12 for Red Hat 6.2

LifeKeeper Sendmail Recovery Kit

steeleye-lkSEN-3.1.4-2.i386.rpm Sendmail 8.11

LifeKeeper SAMS Recovery Kit

steeleye-lkSAM-4.1.0-1.i386.rpm Sendmail Advanced Message Server (SAMS) 1.2 or Sendmail Switch 2.2

LifeKeeper Print Services Recovery Kit

steeleye-lkLP-3.1.3-4.i386.rpm Linux print spoolers and daemons

LifeKeeper NFS Server Recovery Kit

steeleye-lkNFS-4.0.0-2.i386.rpm NFS exported filesystems

LifeKeeper Network Attached Storage Recovery Kit

steeleye-lkNAS-4.0.0-1.i386.rpm Mounted NFS filesystems from an NFS server or Network Attached Storage (NAS) device

LifeKeeper Data Replication

steeleye-lkDR-4.0.0-2.i386.rpm network RAID devices (synchronous mirrors)

LifeKeeper Software Development Kit (SDK)

steeleye-lkSDK-3.1.3-6.i386.rpm Provides sample code and tool for creating a basic application Recovery Kit. Documentation includes an overview of the Recovery Kit architecture, and a detailed description of Recovery Kit components and APIs.

New Features in LifeKeeper v4.1.0


New Features in LifeKeeper v4.1.0

New Feature Description

Installation Support Support for SuSE SLES 7 and Miracle Linux 1.1, 2.0. Includes drivers for the CCISS Recovery Kit, and kernel modules for the NFS Server Recovery Kit and LKDR.

Installation Support v4.1.0-9 or later includes support for Red Hat 7.2, kernel 2.4.9-21 support for Red Hat 7.1, and updated LKDR modular kernel patches for all 2.4 Linux kernels.

Licensing LifeKeeper v4.1.0 requires a unique runtime license key for each server. The installation support script installs a Licensing Utilities package which obtains and displays the host ID of your server. The host IDs and authorization code provided with your software are used to obtain permanent license keys from the SteelEye Technology Inc. website. You will also need to provide the host IDs of your servers to obtain 30 day evaluation licenses for demo or evaluation purposes. A license key is required to run the LifeKeeper software.

Enhanced LifeKeeper SCSI Reservation Daemon

The daemon process, lkscsid, that maintains SCSI reservations and performs health monitoring for shared SCSI devices has been multi-threaded for improved performance and error handling.

CCISS Support for support for Compaq cciss devices, implemented as CCISS_device resources, for use with the Compaq Proliant DL380 Generation 2 Packaged Cluster.

Oracle Recovery Kit Support for Oracle 9i Standard Edition.

LifeKeeper Product Requirements

8 Release Notes

LifeKeeper Product Requirements LifeKeeper for Linux is currently targeted for any Linux platform that satisfies the minimum requirements included in the table below.

Note: LifeKeeper on a Linux server will not inter-operate with LifeKeeper for Windows NT/2000, Solaris, or NCR SVR4 MP-RAS server.

Description Requirement

Linux Operating System Red Hat 6.2, 7.0, 7.1 and 7.2 *

Caldera eServer 2.3, 2.3.1 *

SuSE 7.1, 7.2, and SLES 7

Miracle Linux 1.1 and 2.0

TurboLinux Server 6.5

Perl perl-5.00503 (or later)

Note: If you have an older version of Perl, you must install the minimum version for LifeKeeper to work properly.

Korn Shell pdksh-5.2.14

Memory The minimum memory requirement for a system supporting LifeKeeper is 64 MB. Additional memory (beyond that required for LifeKeeper) is required to run user applications. Refer to “Application Configuration” under the Technical Notes section of this document.

Disk Space The LifeKeeper Core Package Cluster requires the following disk space:

/opt Approximately 9990 (1024-byte) disk blocks

/ Approximately 25 (1024-byte) disk blocks

* Refer to “Kernel updates” under the Technical Notes section of this document.



LifeKeeper Optional Recovery Software

The following table shows the software and disk space requirements for the optional LifeKeeper recovery software.

See “Application Configuration” under the Technical Notes section of this document for additional requirements and/or restrictions that may apply to applications under LifeKeeper protection.

Product Requirement(s) Disk Space Required

Apache Web Server

Apache Web Server 1.3.9 or later SecureWeb Server 3.12 or later

LifeKeeper v3.00 or later Core Package Cluster

Approximately 200 (1024-byte) disk blocks in /opt

DB2 IBM DB2 Universal Database (UDB) Enterprise Edition (EE) v7.1 or later

IBM DB2 Universal Database (UDB) Enterprise Extended Edition (EEE) v7.1 or later with Fix Pack #3 or later

LifeKeeper v3.1.3 or later Core Package Cluster

LifeKeeper NFS Server Recovery Kit v3.1.3-5 or later (for DB2 EEE only)

Approximately 175 (1024-byte) disk blocks in /opt.

Oracle ORACLE Version 8.0.5 or later



Informix Informix Dynamic Server 9.2 or later



MySQL MySQL 3.22 or later



Sendmail Sendmail 8.11 or later



Sendmail Advanced Message Server

Sendmail Advanced Message Server (SAMS) 1.2 or later

Sendmail Switch 2.2 or later




10 Release Notes

Product Requirement(s) Disk Space Required

Print Services Standard print services provided with the supported Red Hat Linux distributions.



NFS Server HANFS-RedHat-2.2.16 or later, nfs-utils-0.1.6-2 or later (not supported by Caldera)



Network Attached Storage



Data Replication

HADR-RedHat-2.2.16 or later or HADR-Caldera-2.2.14 or later, raidtools-0.90-6 or later



Software Development Kit (SDK)



GUI Platforms and Browsers

The LifeKeeper GUI is targeted for any platform that provides full JDK 1.3 support. Included in the following table is a list of the supported platforms and browsers for the LifeKeeper GUI.

Note: Other platforms and browsers may work with the LifeKeeper GUI (for example, Solaris x86/Netscape 4.5 or Windows NT/Netscape Navigator 3.0), but they have not been tested by SteelEye Technology, Inc.

Operating System

Internet Explorer 4.0

Internet Explorer 5.0

Netscape Navigator 4.7

Netscape Navigator 6.1

Windows 95/98 X X X X

Windows NT X X X X

Windows 2000 X X X

Linux X X



Included in the table below are the minimum system requirements for a LifeKeeper GUI client running Windows or Linux.

Windows Linux

Windows 95/98, Windows NT 4.0 SP5, Windows ME, Windows 2000

Red Hat 6.2, 7.0 and 7.1, Caldera eServer 2.3, 2.3.1, SuSE 7.1 and 7.2, TurboLinux Server 6.5

Linux kernel v2.2.12 or later, glibc v2.1.2-11 or later.

Note: Running the setup script on the LifeKeeper Installation Support CD will update your system to the minimum kernel and glibc revision required.

Java Plug-in 1.3.1 * Java Runtime Environment 1.3.1 **

16-Bit Color Mode 16-Bit Color Mode

Pentium 90 MHz or faster processor Pentium 90 MHz or faster processor

75 MB free hard disk space (additional 125 MB required for separate documentation download)

75 MB free hard disk space (additional 125 MB required for separate documentation download)

48 MB system RAM 48 MB system RAM

* Refer to “Java Plug-in software” under the Technical Notes section of this document. ** Refer to “GUI Java platform” under the Technical Notes section of this document.

Storage and Adapter Options *

The following table reflects disk array storage models and adapters that vendors state will support LifeKeeper for Linux minimum configurations.

Vendor Storage Model

Compaq RA 4100

MA/RA 8000 ***

Consan CRD5440 ***

CRD7220 (f/w 3.00)

Dell 650F (Clarion)

IBM FAStT200 ***

Winchester Systems FlashDisk RAID Disk Array

EMC Symmetrix 3000 Series **

Symmetrix 8000 Series

Hitachi Data Systems HDS7700

HDS5800

Installation and Configuration

12 Release Notes

Vendor Storage Model

MTI Gladiator 2550

Gladiator 3550

Gladiator 3600

MicroNet GenesisONE

ProCom Reliant 1000

Baydel Storage Arrays DAR3/5SE68C

DAR3/C/5SE68C

Radion Systems Rack U2W

Microdisk U2W

JetStor JetStor II

Adapter Type Adapter Model

Differential SCSI Adapter Adaptec 2944 W, Adaptec 2944 UW, or Adaptec 2940 U2W **

Compaq 64bit PCI Dual Channel Wide Ultra2 SCSI Adapter **

SCSI Host Based RAID IBM ServeRAID-4Mx Ultra160 SCSI controller ***

AMI Perc2/DC MegaRAID

Fibre Channel Qlogic QLA 2100, Qlogic QLA 2200

Compaq 64 bit /66MHz Fibre Channel Host Adapter 120186-B21 **

* Not required in LifeKeeper configurations involving non-shared storage with IP failover only, or for LifeKeeper Data Replication. ** Certified by SteelEye Technology, Inc.

*** Certified by SteelEye Technology, Inc. Refer to “Storage and Adapter Configuration” under the Technical Notes section of this document.

Installation and Configuration See the following documents for installation and configuration instructions:

• For detailed LifeKeeper configuration planning considerations and installation tasks, consult the LifeKeeper Planning and Installation Guide.

• The individual Recovery Kit Administration Guides provide configuration tips for their associated applications.

• For specific hardware configuration information, refer to www.steeleye.com/support/documentation for the Configuring LifeKeeper Clusters document and several Peripheral Configuration documents.

Upgrades


Upgrades LifeKeeper can be upgraded while preserving existing resource hierarchies. See the LifeKeeper Planning and Installation Guide for instructions on upgrading from a previous LifeKeeper version, and also for information about upgrading your Linux operating system.

Note: Customers who have developed customized remove and restore scripts for your applications should contact your SteelEye Systems Engineer before upgrading LifeKeeper.

Before Starting Your Upgrade

Run the /opt/LifeKeeper/bin/lcdstatus command on the server to be upgraded, and save the output to a file or print it for later reference. You may also want to save copies of various LifeKeeper configuration files (for instance, /etc/default/LifeKeeper) prior to the upgrade.

Upgrading from LifeKeeper v3.1 to v4.1.0

With LifeKeeper GUI v4.0.0 or later , the client administration is not compatible between earlier versions of LifeKeeper GUI software. Although status information is shown between the current and previous versions of the software, use of the administrative functions, such as in-service and out-of-service, with an earlier version will result in a java exception due to the incompatible versions.

To administer a server which has been updated with LifeKeeper GUI v4.0.0 or later, a LifeKeeper GUI v4.0.0 or later client should be used. Keep this issue in mind during rolling upgrades, where resources are moved in and out of service from one server to the next in order to minimize application down time. During rolling upgrades all the versions of the LifeKeeper GUI in a cluster are not the same until the upgrade of the each server in the cluster is complete. When in doubt, run the GUI client from the same server that is to be administered.

See the LifeKeeper Planning and Installation Guide for detailed instructions on upgrading from a previous version of LifeKeeper.

Upgrading to Installation Support v4.1.0-9 or later

For LifeKeeper Data Replication customers, it is recommended that you upgrade your kernel by running the setup script included on the Installation Support CD, in order to pick up updated LKDR modular kernel patches for all supported 2.4 Linux kernels. This should only affect existing LKDR customers who are upgrading from a previous version of v4.1.0.

Note: For LKDR customers who are running Miracle Linux 2.0, you will need to remove the HADR-<Distribution>-<Kernel Version>- 4.1-0-5 package using rpm –e before running the setup script included on the installation support CD. You can use the command rpm –qa | grep HADR to determine the HADR package which needs to be removed from the system.

For Red Hat 7.1 customers, if you have the kernel source installed, upgrading to the 2.4.9-21 kernel will require you to upgrade your compiler. Running the setup script on the LifeKeeper

Technical Notes

14 Release Notes

Installation Support CD will update your system to the minimum gcc version required. The dependency on gcc is in kernel-source-2.4.9-21 only. It requires gcc >= 2.96-85.

Red Hat 7.1 customers who are upgrading to the 2.4.9-21 kernel using the setup script on the LifeKeeper Installation Support CD, may see the following messages from the upgrade of the filesystem package which is required to satisfy a mkinitrd-3.2.6-1 dependency. Despite these messages, the installation of the kernel update does complete successfully.

Error: Cannot remove /var/lib/rpm – directory not empty

Error: Cannot remove /usr/local/man/man1 – directory not empty

Error: Cannot remove /usr/local/man – directory not empty

Error: rmdir of /mnt/cdrom failed: Device or resource

Technical Notes We strongly recommend that you read the following technical notes concerning configuration and operational issues related to your LifeKeeper environment.

LifeKeeper Features

Item Description

Large Cluster Support

LifeKeeper supports large cluster configurations, up to 16 servers. Due to storage interconnect solutions supported by the platforms, however, some cluster configurations in LifeKeeper for Linux may be limited to 2 servers per shared SCSI bus

International-ization

LifeKeeper for Linux v4.1 does support wide/multi-byte characters in resource and tag names, but does not include native language message support. The LifeKeeper GUI is internationalized; however, until the LifeKeeper core software is fully internationalized, internationalization of the GUI only provides a partial solution for users.

See also lklin00002100 under “Restrictions or Known Issues - LifeKeeper Core” for additional information.

Technical Notes


Tuning

Item Description

IPC semaphores and IPC shared memory

LifeKeeper requires Inter-Process Communication (IPC) semaphores and IPC shared memory. The default Red Hat values for the following Linux kernel compile time options are located in /usr/src/linux/include/linux/sem.h, and should be sufficient to support most LifeKeeper configurations.

Option Required Default Red Hat 6.1

SEMOPM 14 32

SEMUME 20 32

SEMMNU 60 32000

SEMMAP 25 32000

SEMMNI 25 128

System File Table

LifeKeeper requires that system resources be available in order to failover successfully to a backup system. For example, if the system file table is full, LifeKeeper may be unable to start new processes and perform a recovery. By default, MAX_INODE, the system file table limit in Red Hat 6.1 and 6.2 is 16384, which should be sufficient to support most LifeKeeper configurations.

LifeKeeper Operations

Item Description

Kernel Debugger (kdb)

init s

Before using the Kernel Debugger (kdb) or moving to init s on a LifeKeeper protected server, you should first either shutdown LifeKeeper on that server, or switch any LifeKeeper protected resources over to the backup server. Use of kdb with the LifeKeeper SCSI Reservation Daemon (lkscsid) enabled (it is enabled by default) can also lead to unexpected panics.

System Panic on Reserved Shared Devices

LifeKeeper uses SCSI reservations to protect shared data from being accessed by other servers on a shared SCSI Bus. If LifeKeeper cannot access a device as a result of another server reserving the device, then a critical error has occurred and data can be corrupted. When this condition is detected, LifeKeeper enables a feature that will cause the system to panic. A tunable called “FAILFASTTIMER” has been added to the LifeKeeper default file, /etc/default/LifeKeeper, to specify the interval for checking the reservation status of a device. By default it is set to 5 seconds. Setting the interval to 0 will disable the timer. Changes to this setting will not take effect until LifeKeeper is stopped and restarted.

If LifeKeeper is stopped by some means other than ‘lkstop,’ with shared devices still reserved (this could be caused by executing kdb or init s), then the LifeKeeper SCSI Reservation Daemon may trigger a kernel panic when the other server recovers the resource(s). All resources must be placed out of service before stopping LifeKeeper in this manner.

Technical Notes

16 Release Notes

Item Description

Recovering out-of-service hierarchies

As part of the recovery following the failure of a LifeKeeper server, resource hierarchies that were configured on the failed server, but which were not in service anywhere at the time of the server failure, are recovered on the highest priority alive server at the time of the failure. This is the case no matter where the out-of-service hierarchy was last in service, including the failed server, the recovering server, or some other server in the hierarchy.

Server Configuration

Item Description

BIOS updates The latest BIOS should be installed on Dell 2400s and Compaq ML350s. Dell 2400 should be at “V03” or higher. Compaq ML 350 should be 12/20/99 or newer.

Storage and Adapter Configuration

Item Description

Compaq MA/RA 8000

Certified by SteelEye Technology, Inc. for Qlogic 2200.

IBM FAStT200 Certified by SteelEye Technology, Inc. for single controller (i.e. single loop) configurations only.

Consan CRD5440

It is recommended that you disable command queuing in order to avoid a problem with the controller in this array that has been observed with the “old” aic7xxx driver. This can be done several ways, such as by rebuilding the driver with it turned off and setting options in modules.conf to disable queuing.

IBM ServeRAID-4Mx Controllers

Certified by SteelEye Technology, Inc. for use with LifeKeeper Data Replication only. Version 4.80 of the ips driver and BIOS is required for LifeKeeper Data Replication to function properly. The Red Hat 2.4.9-21 kernel includes this driver by default.

Redundant Controllers

The Compaq Smart Array Cluster Storage system, available as part of a packaged cluster with Compaq Proliant DL380 Generation 2 servers, is currently the only storage solution LifeKeeper supports in a redundant controller configuration. The ability to support redundant controllers in either servers and/or storage arrays, is not widely available for Linux today. SteelEye is actively working with server vendors, storage vendors, adapter vendors, and driver maintainers to provide support for these configurations.

Technical Notes


Linux Configuration

Item Description

Dynamic device addition

All devices should be powered on and ready during the loading of the Linux Operating System. Configuring devices after the system is loaded can cause devices to change their identity (e.g. move from being /dev/sdc to /dev/sdf) and Linux has a limit to the number of devices that can be dynamically or hot added. The 2.2 series of kernels has the limit set to 4 while the 2.4 series of kernel has the limit set to 40. The specific limit is defined in the kernel file /usr/src/linux/drivers/scsi/hosts.h in the define EXTRA_DEVS. To increase the limit on 2.2. kernels hosts.h must be manually edited to increase the define. To increase the limit on 2.4 kernels use make config to update a selection in the SCSI module called “Maximum number of SCSI disks that can be loaded as modules”.

Prior to LifeKeeper startup Linux must configure all devices. If a LifeKeeper protected device is configured after LifeKeeper is started then LifeKeeper must be stopped on each server that shares the device and restarted. This will enable the device detection and validation to confirm the configuration and enable LifeKeeper to access the device.

Refer to the item LUN support as this is a common reason that devices are not configured during boot up.

Refer to the item SCSI device nodes on issues that will occur when more than 8 SCSI devices are configured on a system.

Kernel updates In order to provide the highest level of availability for a LifeKeeper cluster, the kernel version used on a system is very important. The table below lists each supported distribution and version with the kernel that has passed LifeKeeper certification testing. As noted in the table, some kernels are provided with the LifeKeeper Installation Support CD to facilitate installation.

Distribution/Version Supported kernels

Red Hat 6.2 2.2.16-3 (LK I/S CD)

Red Hat 7.0 2.2.16-22 (default kernel)

Red Hat 7.1/7.2 2.4.9-21 (LK I/S CD)

Caldera eServer 2.3/2.3.1 2.2.14-3SteelEye (LK I/S CD)

TurboLinux 6.5 2.2.18-2 (default kernel)

SuSE 7.1 2.2.18 (default kernel)

SuSE 7.2 2.4.4 (default kernel)

SuSE SLES 7 2.4.7 (default kernel)

Miracle Linux 1.1 2.2.18-2.3 (default kernel)

Miracle Linux 2.0 2.4.7-2.24ml (default kernel)

Technical Notes

18 Release Notes

Item Description

LUN support The Linux SCSI driver has several parameters that control which devices will be probed for Logical Units (LUNs):

• List of devices that do not support LUNs – this list of devices are known to NOT support LUNs so the SCSI driver will not allow the probing of these devices for LUNs.

• List of devices that do support LUNs – this list of devices is known to support LUNs well so always probe for LUNs.

• Probe all LUNs on each SCSI device – if a device is not found on either list whether to probe or not. This parameter is configured by make config in the SCSI module section.

While most distributions have the Probe all LUNs setting on by default Red Hat has the setting “ disabled by default. External RAID controllers that are typically used in LifeKeeper configuration to protect data frequently are configured with multiple LUNs (Logical Units). To enable LUN support, this field must be selected and the kernel remade, or you may use the ”arrayscan” utility provided on the LifeKeeper for Linux Installation Support CD.

The arrayscan utility is run during the OS bring-up activities to configure LUNs on specific devices. In order for the arrayscan utility to work correctly LUN 0 for a device must be configured in the system. The arrayscan utility checks every LUN 0 that is configured and if it is a device listed in the “templates” file then the utility will issue commands to the SCSI driver to scan for LUN 1 through 7. Since this utility is run after the initial load of the SCSI modules any device configured will fall into the “hot add” or “dynamic device addition” category. Refer to “Dynamic Device Addition” above.

Note: If Probe all LUNS is enabled, some devices can cause the system to probe hundred of non-existent LUNS which will result in extremely slow reboots . This can be avoided by adding the following entry to the boot loader config file (e.g. etc/lilo.conf), which limits the maximum number of LUNs that will be scanned at boot time.

append= "max_scsi_luns=16"

Technical Notes


Item Description

SCSI device nodes

Most Linux distributions only configure 8 SCSI generic nodes (/dev/sg0 through /dev/sg7 linked to /dev/sda through /dev/sgh) and 16 SCSI disk nodes (/dev/sda through /dev/sdp). Virtually all LifeKeeper configurations will exceed these limits. Without the proper number of nodes created on a system, devices will not be accessible or configured.

SCSI disk nodes: There needs to be SCSI a disk node (/dev/sd) for each disk to be accessed by a system. Each partition on a disk (/dev/sdp) and each LUN will also need its own disk node. Use mknod or the “setup” script on the Installation Support CD to create the appropriate number of nodes to address each disk, partition, and LUN.

Generic nodes: LifeKeeper uses the Generic device nodes to configure and protect each disk/LUN. The generic driver addresses every SCSI device connected to a system whether it is a disk, tape, CDROM, disk array, etc. This means the SCSI generic driver must be compiled as a module or into the kernel and there should be enough generic nodes for every SCSI device connected to a system. Using the file /proc/scsi/scsi, verify that there are sufficient generic nodes. If there are a total of 20 unique devices listed in /proc/scsi/scsi then there should be at least 20 generic nodes. Use mknod or the “setup” script on the Installation Support CD to create the appropriate nodes.

Note: The “setup” script on the LifeKeeper Installation Support CD will automatically create SCSI disk nodes up to /dev/sdaz and generic nodes up to /dev/sg255.

Filesystem labels

The use of labels in /etc/fstab will slow down fsck since it has to scan all partitions to determine what device a label refers to. The mapping of labels to partitions is not stored in the OS so every time an fsck is done it must open every partition This also impacts LifeKeeper due to the fact that LifeKeeper has reserved the partitions that it has in-service. Every time fsck is performed from another server in the cluster, it will try to open the reserved partitions, and this will cause the OS/driver to perform error handling. The current workaround for this problem is to not use labels. Red Hat 7.0 provides the option of using labels for the root file system. User’s should not select this feature (and the monitor fstab feature) to make sure that the OS does not configure labels.

Reiser file system

A patch was applied to the reiser file system in the Red Hat 2.4.3-12 kernel that causes the unmount to fail on Red Hat 7.1. If the patch is backed out then the unmount will succeed. The patch is called linux-2.4.3-reiserfs.patch, and can be found in the 2.4.3-12 src rpm from Red Hat.

Technical Notes

20 Release Notes

Item Description

aic7xxx driver Version 6.1.13 of the "new" aic7xxx driver on SuSE 7.2 (default driver) will oops if an array (connected to an Adaptec controller) does not have all 8 LUNs configured. Note that this is not a problem on Red Hat 7.1 with the 2.4.3-12 kernel as the fixed driver is already provided.

Workaround:

1) The recommended workaround is to modify the ram disk using the steps described below so that the aic7xxx old driver is loaded at boot time.

Edit /etc/rc.config:

Change INITRD_MODULES="aic7xxx <other modules ...>" to INITRD_MODULES="aic7xxx_old <other modules ...>".

Edit /etc/modules.conf:

Change "alias scsi_hostadapter aic7xxx" to "alias scsi_hostadapter aic7xxx_old".

Note: If /etc/modules.conf has this setting as "off" ("alias scsi_hostadapter off"), then no change is necessary.

Save the default RAM disk to be used in an emergency:

cp /boot/initrd /boot/initrd.save

Create new RAM disk:

mk_initrd

Reload LILO:

lilo

Reboot server loading the label in lilo that uses the default RAM disk just created.

2) The fix to the aic7xxx driver is simple and has been included in version 6.2.0 and higher of the aic7xxx driver. For someone who wants to continue to use the new aic7xxx driver rather than using the aic7xxx_old.o driver as a workaround, the latest driver can be found at the maintainers web site http://people.freebsd.org/~gibbs/linux/. Follow the instructions from the maintainer to build and install the new driver.

qla2x00 driver The recommended Qlogic qla2x00 driver for use with LifeKeeper is version 4.27 or higher for both the 2.4 and 2.2 Linux kernels. The Qlogic driver, however, requires a SteelEye patch to function properly with LifeKeeper. The LifeKeeper for Linux v4.1.0-9 or later Installation Support CD will install the correct version of the qla2x00 driver for each supported distribution and kernel version. A source rpm is available for building custom kernels.

Note: The 5.x versions of the QLogic driver are not supported at this time.

Technical Notes


Network Configuration

Item Description

IP Recovery Kit impact on routing table

LifeKeeper-protected IP addresses are implemented on Linux as logical interfaces. When a logical interface is configured on Linux, a route to the subnet associated with the logical interface is automatically added to the routing table, even if a route to that subnet already exists (for example, through the physical interface). This additional route to the subnet could possibly result in multiple routing-table entries to the same subnet.

If an application is inspecting and attempting to verify the address from which incoming connections are made, the multiple routing-table entries could cause problems for such applications on other systems (non-LifeKeeper installed) to which the LifeKeeper system may be connecting. The multiple routing table entries can make it appear that the connection was made from the IP address associated with the logical interface - rather than the physical interface.

IP subnet mask For IP configurations under LifeKeeper protection, if the LifeKeeper-protected IP address is intended to be on the same subnet as the IP address of the physical interface on which it is aliased, the subnet mask of the two addresses must be the same. Incorrect settings of the subnet mask may result in connection delays and failures between the LifeKeeper GUI client and server.

EEpro100 driver initialization

The Intel e100 driver should be installed to resolve initialization problems with the eepro100 driver on systems with Intel Ethernet Interfaces. With the eepro100 driver, the following errors may occur when the interface is started at boot time, and repeat continuously until the interface is shut down.

eth0: card reports no Rx buffers

eth0: card reports no resources

Refer to http://www.scyld.com/pipermail/eepro100/2000-July/001135.html for additional information.

http://www.scyld.com/pipermail/eepro100/2000-July/001135.html

http://www.scyld.com/pipermail/eepro100/2000-July/001135.html

Technical Notes

22 Release Notes

Application Configuration

Item Description

Application memory requirements

On 2.4 kernels a server may hang if swap is in use frequently when the application is running. This is particularly a problem with Oracle 9i because of its large memory usage. To work around this problem, additional memory can be added to avoid continuous swapping.

Database memory requirements

Based on testing by SteelEye Technology Inc., the minimum memory recommended for Oracle 8i on Red Hat 7.1 and SuSE 7.2 is 256 MB. The minimum memory recommended for Oracle 9i on Red Hat 7.1 is 1 GB due to the application memory issue described above.

Database support with glibc 2.2

SuSE 7.1 uses the next generation glibc 2.2. The Oracle 8i and all Oracle products in the 8i series use glibc 2.1. To obtain the Oracle patch that will enable their products to work on glibc 2.2 systems go to http://technet.oracle.com/software/products/oracle8i/software_index.htm (This site requires your Oracle OTN login/password.) Informix Dynamic Server 9.2 also uses glibc 2.1. Informix Dynamic Server 9.21 or later is required on distributions that use glibc 2.2.

Database initialization files

The initialization files for databases need to be either on a shared device and symbolically linked to specified locations in the local file system or kept on separate systems and manually updated on both systems when changes need to be implemented.

Localized Oracle mount points not supported

Oracle mount points containing non-English characters cannot be put under LifeKeeper protection. A bug report has been submitted against Oracle 8.0.5. This problem does not occur when you use Oracle 8i.

Note: Localized Oracle 8i environments are different depending on whether you connect as internal or as sysdba. A database on a localized mount point must be created with “connect / as sysdba” if it is to be put under LifeKeeper protection.

Sendmail updates

Upgrading a LifeKeeper protected Sendmail application as part of upgrading the Linux operating system (e.g. Red Hat 6.1 to Red Hat 6.2) requires special instructions. LifeKeeper protected Sendmail hierarchies must be unextended before upgrading the Linux operating system. If symbolic links are used for the message queue (/var/spool/mqueue), and the receiving directory (/var/spool/mail), the operating system upgrade removes these links. They will need to be recreated after upgrade is completed.

Apache updates Upgrading a LifeKeeper protected Apache application as part of upgrading the Linux operating system (e.g. Red Hat 6.1 to Red Hat 6.2) requires that the default server instance be disabled on start up.

If your configuration file (httpd.conf) is in the default directory (/etc/httpd/conf), the Red Hat upgrade will overwrite the config file. Therefore, you should make a copy of the file before upgrading and restore the file after upgrading.

Also, see the Specific Configuration Considerations for Apache Web Server section in the Apache Web Server Recovery Kit Administration Guide.

Technical Notes


Item Description

Apache SecureWeb support

Versions of Red Hat 7.0 and greater have a different implementation of secure web support (SSL) in Apache than previous versions of Red Hat. The differences render the LifeKeeper Recovery Kit for Apache unable to protect instances running in the secure web configuration. Even if you have configured the web server to run secure web, LifeKeeper will only start and protect a non-secure instance.

DB2 EEE Support

LifeKeeper support for IBM DB2 Universal Database (UDB) Enterprise Extended Edition (EEE) v7.1 or later requires Fix Pack #3 or later which can be downloaded directly from ftp.software.ibm.com/ps/products/db2/english/db2linuxv71. Note: Fix Pack #3 or later is also required with v7.2.

GUI Configuration

Item Description

GUI client and server communication

The LifeKeeper GUI client and server use Java Remote Method Invocation (RMI) to communicate. For RMI to work correctly, the client and server must use resolvable hostnames or IP addresses. If DNS is not implemented (or names are not resolvable using other name lookup mechanisms), edit the /etc/hosts file on each client and server, to include the names and addresses of all other LifeKeeper servers.

GUI Java platform

The LifeKeeper GUI requires that the Java Software Development Kit (JDK) or Java Runtime Environment (JRE) be installed on each server. The JRE, which consists of the Java virtual machine, the Java platform core classes, and supporting files is the minimum required to run the LifeKeeper GUI.

The JRE 1.3.1 for Linux is available on the LifeKeeper for Linux Installation Support CD (See the LifeKeeper Planning and Installation Guide) or it can be downloaded directly from http://java.sun.com/j2se/1.3/jre/download-linux.html. This file is approximately 15 MB in size.

The JDK 1.3.1 for Linux is available from http://java.sun.com/j2se/1.3/download-linux.html. This file is approximately 26 MB in size.

To install use the rpm or tar command. If you are using the tar command extract the JRE into a directory named /usr/java/jre1.3.1, and the JDK into a directory named /usr/java/jdk1.3.1.

Note: By default, the LifeKeeper GUI expects the Java interpreter to be installed in the directory /usr/java/jdk1.3.1/bin or /usr/java/jre1.3.1. If you install the JRE or JDK in another directory location, you must edit the PATH in the LifeKeeper default file /etc/default/LifeKeeper to reflect the correct location. If LifeKeeper is running when you edit this file, you should stop and restart LifeKeeper to recognize the change. Otherwise, the LifeKeeper GUI will not be able to find the Java command.

ftp://ftp.software.ibm.com/ps/products/db2/english/db2linuxv71

http://java.sun.com/products/j2sedk/1.32/jre/download-linux.html

http://java.sun.com/products/j2sedk/1.32/download-linux.html

Technical Notes

24 Release Notes

Item Description

Linux support for Java

The Java Software Development Kit (JDK) or Java Runtime Environment (JRE) 1.3.1 is supported on Intel Pentium platforms running the Linux kernel v 2.2.12 and glibc v2.1.2-11 or later. Running the setup script on the LifeKeeper Installation Support CD will update your system to the minimum kernel and glibc revision required. See the LifeKeeper Planning and Installation Guide for details about running the setup script on the Installation Support CD.

NOTE: Sun recommends using Red Hat Linux 7.1 rather than 7.0, due to the known problems that are described on their web site. Running the setup script on the LifeKeeper Installation Support CD will update your system to a later glibc revision that SteelEye Technology Inc. has certified with JDK/JRE 1.3.1 and Red Hat 7.0.

Java Plug-in software

To ensure that the LifeKeeper GUI can be run as an applet reliably and consistently in both Internet Explorer and Netscape Navigator, the applet requires using the Java Runtime Environment (JRE) through Java Plug-in software. When versions of Internet Explorer or Netscape Navigator (which do not support JDK 1.3) are directed at the web page for the LifeKeeper GUI, Java Plug-in 1.3.1 software will be downloaded and installed. The compressed Java Plug-in download is approximately 32MB. After decompression and installation, the Java Plug-in uses approximately 70 MB of hard disk space.

Note: JPI 1.3.1 for Windows Platforms is available from http://java.sun.com/j2se//1.3//download-windows.html.

JPI 1.3.1. for Linux Platforms is available with the JRE for Linux from http://java.sun.com/j2se/1.3/jre/download-linux.html

Java security policy

The LifeKeeper GUI uses policy based access control. When the LifeKeeper GUI application or applet is loaded, it is assigned permissions based on the Java security policy currently in effect. There is by default a single system-wide policy file, and an optional user policy file. The system policy file is loaded first, and then the user policy file is added to it. The minimum permissions required in a user’s .java.policy file in order to run the LifeKeeper GUI are installed in the file /opt/LifeKeeper/htdoc/java.policy for reference on each server.

Java remote object registry port

The LifeKeeper GUI uses port 82 for the Java remote object registry on each server. This should allow servers to support RMI calls from clients behind typical firewalls.

LifeKeeper administration web server

The LifeKeeper GUI requires an administration web server for client browser communication. Currently, the LifeKeeper GUI is using a public domain web server, mhttpd, for its administration web server. The installation of the LifeKeeper GUI installs and configures this web server, using port 81, which should be different from any public web server.

GUI network access in a LifeKeeper cluster

The LifeKeeper GUI requires network access to all hosts in the LifeKeeper cluster. When running the LifeKeeper GUI in a browser, this is a security violation. Therefore, the administrator of the browser will have to globally enable network access for applets, and be careful not to visit other sites with security set to low values.

http://java.sun.com/j2seproducts/jdk/1.32/jre/download-windows.html

http://java.sun.com/products/j2sedk/1.32/jre/download-linux.html

Technical Notes


Item Description

Online Product Manual

The LifeKeeper Online Product Manual is included in the steeleye-lkHLP software package. The Product Manual is HTML-based, and accessed from a web browser. Once the Online Product Manual is installed, “help” topics are accessible from the Table of Contents in the Help menu of the LifeKeeper GUI. Context-sensitive Help is accessible from Help buttons within the GUI dialog boxes.

Online Product Manual access outside the GUI

The Online Product Manual can be accessed outside the GUI by opening the following URL for the GUI product manual web page directly from your browser: http://<server name>:81/help/lksstart.htm ,where <server name> is the name of the LifeKeeper server.

Note: The Product Manual requires a web server for client browser communication. Installation of the LifeKeeper GUI installs and configures a public domain web server, mhttpd, using port 81.

GUI Limitations

Item Description

GUI incompatibility with Java 1.4

Due to incompatible changes in the way rendering works in JDK 1.4, the current LifeKeeper GUI will have formatting problems with Java (JDK, JRE, or Plug-in) 1.4 or higher. It is recommended that you use Java 1.3.1 (JDK, JRE, or Plug-in), which has been certified by SteelEye Technology Inc. for use with LifeKeeper.

GUI client incompatibility during upgrade

With LifeKeeper GUI v4.0.0 or later, the client administration is not compatible with earlier versions of LifeKeeper GUI software. Although status information is shown between the current and previous versions of the software, use of the administrative functions, such as in-service and out-of-service, with an earlier version will result in a java exception due to the incompatible versions.

To administer a server which has been updated with LifeKeeper GUI v4.0.0 or later, a LifeKeeper GUI v4.0.0 or later client should be used. Keep this issue in mind during rolling upgrades, where resources are moved in and out of service from one server to the next in order to minimize application down time. During rolling upgrades all the versions of the LifeKeeper GUI in a cluster are not the same until the upgrade of the each server in the cluster is complete. When in doubt, run the GUI client from the same server that is to be administered.

GUI interoperability restriction

The LifeKeeper GUI may only be used to administer LifeKeeper on Linux servers running Red Hat, Caldera, SuSE or TurboLinux distributions. The LifeKeeper for Linux GUI will not interoperate with LifeKeeper for Windows NT/2000, Solaris or MP-RAS.

Internet Explorer GUI restriction

When using the LifeKeeper GUI from Internet Explorer, clicking in some areas of the GUI when it is partially covered by another window does not make it visible.

Note: This problem does not occur on Linux or when using Netscape on Windows.

Restrictions or Known Issues

26 Release Notes

Restrictions or Known Issues Included below are the restrictions or known issues open against LifeKeeper for Linux, broken down by functional area.

Installation

ID Description

lklin00001599 Unable to relocate package on Red Hat 7.1

There is a known bug in rpm version 4.0.2 which is shipped with RedHat 7.1 that causes any attempt to relocate a package to fail. This bug is being tracked on the RedHat bugzilla site as bug id 53272.

Workaround: Upgrade the rpm package to at least version 4.0.3-06 or later which can (currently) be downloaded from the rawhide distribution of RedHat, or downgrade the rpm package to version 4.0-4 which was shipped with Red Hat 7.0.

lklin000366 Uninstall using kpackage while LifeKeeper is running fails.

Uninstall using kpackage while LifeKeeper is running fails to remove LifeKeeper and leaves LifeKeeper processes running.

Workaround: Uncheck the “Use Scripts” checkbox (the default is checked). This will cause the preremove and postremove package scripts to execute, and LifeKeeper will uninstall cleanly. Note: With uninstall in kpackage 2.0.1 (the version shipped with SuSE 7.1), the “Use Scripts” option now functions properly, so it should remain checked (the default).

lklin000418 Installation error: lktest command not found

When using kpackage of GnoRPM, the following error occurs if a Recovery Kit installation script references the lktest command before it has been installed by LifeKeeper.

Error: /var/tmp/rpm-tmp.65502: lktest: command not found

Workaround: Install LifeKeeper first, and then go back and install any Recovery Kits (i.e. do not perform a “batch” install). You may also use the install tool option which orders package installation based dependencies.



lklin0000504 Errors on removal of a relocated LifeKeeper core package

With rpm versions later than rpm-3.0.3, removal of a relocated LifeKeeper core package and relocated distribution enabling packages fails to remove all directories. This is due to a problem with using the –prefix option to install with rpm-3.0.4 and 4.0.4 which causes an extra directory to be created one level above the relocation path.

Workaround: Relocate these packages using rpm –relocate with the– badreloc option, or remove the directories associated with these packages manually.

lklin00000783 Man page package does not install with “rpm –i *” on Caldera

The version of rpm in Caldera eServer 2.3 (rpm -2.5.5-3) is older than the version of rpm in RedHat 6.1 & 6.2 (rpm-3.0.3 and 3.0.4). When using “*” for the argument list, you pick up all files in the directory including those that are not rpm archives. When processing the file list, rpm knows to skip the non-rpm files, but sets something internal incorrectly that causes it to skip the last package in the list (which in this case is the man page package), if the non-rpm file shows up before the SteelEye packages. The same command executes successfully on RedHat 6.2, so somewhere between version 2.5.5 and 3.0.4 of rpm, the problem was fixed.

lklin00000917 Using kpackage to upgrade LifeKeeper on Caldera eServer 2.3 does not update the rpm database correctly.

When upgrading LifeKeeper packages on Caldera eServer 2.3 using kpackage, entries for both packages are left in the rpm database. Some files such as lkGUIapp, lkpasswd, and lkipbu are also not replaced. This only appears to be a problem with the version of kpackage that is installed with Caldera eServer 2.3.

Workaround: Perform the upgrade from the command line using rpm –Uhv –force <package>.rpm.

lklin00001458 Package check errors (rpm -V steeleye-lk) will occur on the core when installed on SuSE. The following errors will occur:

Because of the way SuSE runs shutdown scripts (vs other Linux distributions), the following scripts are moved to another location after installation, so LifeKeeper will be shutdown when changing run levels or rebooting. These should be the only errors that occur when verifying the steeleye-lk package.

Missing /etc/rc.d/rc0.d/K01lifekeeper




28 Release Notes

LifeKeeper Core

ID Description

lklin00000360 After failover, LifeKeeper shows resources as failed when they are not failed (intermittent problem).

After a failover when the failed node has returned to service, if the resource properties are set for Intelligent Switchback, LifeKeeper should show the resources on that node to be in Standby mode. In some infrequent situations, however, the resources on that node are shown as failed and Out of Service, Unavailable (OSU) even when they actually ARE in standby mode (available).

Workaround: To verify whether the returned node is fully available, you may bring that node back into service (a manual switchover).

lklin00002100 Language Environment Effects

Some LifeKeeper scripts parse the output of Linux system utilities, and rely on certain patterns in order to extract information. When some of these commands run under non-English locales, the expected patterns are altered, and LifeKeeper scripts fail to retrieve the needed information. For this reason, the language environment variable LANG has been set to en_US in /etc/default/LifeKeeper. This locale uses the ISO-8859-1 code set, so the extended characters of most European languages may be used in resource names and tag names. It is not necessary to install Linux with the language set to English (any language variant available with your installation media may be chosen); the setting of LANG in /etc/default/LifeKeeper will only influence LifeKeeper. If you change the value of LANG in /etc/default/LifeKeeper, be aware that it may adversely affect the way LifeKeeper operates. The side effects depend on whether or not message catalogs are installed for various languages and utilities, and if they produce text output that LifeKeeper doesn't expect. We plan to provide a fix in an upcoming release or patch that will allow LifeKeeper to work in any locale.

GUI

ID Description

lklin00001829 Certificate Exception when starting LifeKeeper GUI applet with Java Plug-in 1.3.1.

Beginning with Java Plug-in 1.3.0_01 and higher, you may see the following message displayed when the LifeKeeper GUI signed applet is loaded from within a web browser:

Java Plug-in Security Warning

Unable to verify the certificate - code will be treated as unsigned.

java.security.cert.CertificateException: Unable to verify the certificate with root CA

Workaround: If you know that you are loading this code from a known LifeKeeper server, then click OK to accept the java code.



ID Description

lklin00001698 Cannot access Online Product Manual from help menu or from quickstart on TurboLinux.

On TurboLinux, you are not able to access online Product Manual from the GUI help menu or from quickstart. The GUI status bar shows that it is trying to access the Online Product Manual file(s), but it does not succeed and no errors are given.

Workaround: Access the Online Product Manual from outside the GUI by opening the following URL directly from your browser:

http://<server name>:81/help/lksstart.htm ,where <server name> is the name of the LifeKeeper server.

lklin00000477 GUI does not immediately update IP resource state after network is disconnected and then reconnected.

When the primary network between servers in a cluster is disconnected and then reconnected, the IP resource state on a remote GUI client may take as long as 1 minute and 25 seconds to be updated, due to a problem in the RMI/TCP layer.

lklin00000327 Deleting hierarchy when one server is down does not update GUI properly.

When a hierarchy that exists on one or more servers is deleted while one of the servers is down, the GUI will still show that the hierarchy exists, with the “up” servers shown in the Unknown state.

Workaround: Once the “down” server is brought back up, delete the hierarchy on that server. This will update the GUI and remove the inconsistent view.

lklin00000019 GUI client never times out when it cannot connect.

Incorrect settings of the subnet mask and gateway address may result in connection delays and failures. If you are using DNS, long connection delays may also be the result of unresolved hostnames or addresses.

Workaround: To reduce the time it takes to make the first DNS lookup, specify all the hostnames/addresses involved in communication in the local file \winnt\system32\drivers\etc\hosts or \windows\hosts. Reference “Network-Related Troubleshooting” under the Preparing to Run the GUI section in the LifeKeeper Online Product Manual.

lklin00000035 GUI hangs when primary interface goes down

The LifeKeeper GUI uses Remote Method Invocation (RMI) to maintain contact between the GUI client and the GUI server. In nearly every case, contact is established over the primary network interface to the server. This means that if the server’s primary Ethernet interface goes down, contact is lost and the GUI client shows that server state as Unknown.

The only solution to this problem is to bring the server’s primary Ethernet interface up again. Additionally, due to limitations in RMI, this problem cannot be overcome by using a multi-homed server (server with multiple network interfaces). In a future LifeKeeper release, this problem will be solved by providing the ability to protect a primary Ethernet interface by failing it over locally to a redundant network card.


30 Release Notes

Data Replication

ID Description

lklin00001536 In symmetric active LKDR configurations with significant I/O traffic on both servers, the filesystem mounted on the netraid device (mirror) stops responding and eventually the whole system hangs.

Due to the single threaded nature of the Linux buffer cache, the buffer cache flushing daemon can hang trying to flush out a buffer which needs to be committed remotely. While the flushing daemon is hung, all activities in the Linux system with dirty buffers will stop if the number of dirty buffers goes over the system accepted limit (set in /proc/sys/kernel/vm/bdflush).

Usually this is not a serious problem unless something happens to prevent the remote system from clearing remote buffers (e.g. a network failure). LifeKeeper will detect a network failure and stop replication in that event, thus clearing a hang condition. However, if the remote system is also replicating to the local system (i.e. they are both symmetrically replicating to each other), they can deadlock forever if they both get into this flushing daemon hang situation.

The deadlock can be released by manually killing the nbd-client daemons on both systems (which will break the mirrors). To avoid this potential deadlock entirely, however, symmetric active replication is not recommended.

In LKDR v3.1.3, this problem has been observed with 2 symmetric active mirrors under heavy I/O load.

Oracle Recovery Kit

ID Description

lklin00000819 The Oracle Recovery Kit does not include support for Connection Manager and Oracle Names Net8 features..

The LifeKeeper Oracle Recovery Kit does not include support for the following Net8 features of Oracle 8i: Oracle Connection Manager, a routing process that manages a large number of connections that need to access the same service; and Oracle Names, the Oracle-specific name service that maintains a central store of service addresses.

The LifeKeeper Oracle Recovery Kit does protect the Net8 Oracle Listener process that listens for incoming client connection requests and manages traffic to the server. Refer to the LifeKeeper for Linux Oracle Recovery Kit Administration Guide for LifeKeeper configuration specific information regarding the Oracle Listener.



Apache Recovery Kit ID Description

lklin00001919 Support for Apache SSL not compatible with RH 7.x

Versions of Red Hat 7.0 and greater have a different implementation of secure web support (SSL) in Apache than previous versions of Red Hat. The differences render the LifeKeeper Recovery Kit for Apache unable to protect instances running in the secure web configuration. Even if you have configured the web server to run secure web, LifeKeeper will only start and protect a non-secure instance.

lklin00000193 In certain situations an Apache resource hierarchy may be brought In Service correctly, but appears as Failed in the GUI. This can occur when the Apache daemon is started so quickly after the LifeKeeper IP addresses are brought In Service that LifeKeeper doesn't detect immediately that the new daemon is responding to 'http' requests. The next time the LifeKeeper Resource Monitoring script for Apache runs, it will detect that the instance is already started and update the state to Active. Note that since the instance has actually been started correctly, clients will be able to access the protected web site(s) immediately.

If you bring an Apache resource hierarchy In Service manually and it appears as Failed in the GUI, try once again to bring the top-level Apache resource In Service, after making sure all the dependent resources are already In Service. If it fails to start correctly, check the LifeKeeper log using 'lk_log log' to determine the cause of the problem.

Workaround: To avoid having the problem re-occur, do the following:

a) Change to the 'action' directory for the Apache recovery kit (this assumes LKROOT=/opt/LifeKeeper):

cd $LKROOT/subsys/webserver/resources/apache/ actions

b) Make a backup copy of the 'restore' script.

c) Edit the 'restore' script and find the line 'sleep 1'.

d) Change the '1' to '2' (or a bigger number, if necessary for your environment). This increases the time between when the recovery kit starts the instance and when it checks to see if the instance is responding correctly to 'http' requests.

Print Services Recovery Kit

ID Description

lklin00000843 The lpc command doesn't work correctly on Caldera eServer 2.3 when used on a network printer.

Executing “lpc status shared” on Caldera eServer 2.3 hangs for 8-10 seconds, and doesn’t return any status. The lpc command is used fairly extensively in the recovery kit to test the status of a printer and perform various functions on a printer. Until this problem is resolved, the Print Services Recovery Kit is not supported on Caldera eServer 2.3.


32 Release Notes

NFS Server Recovery Kit

ID Description

lklin00000989 Removing an NFS resource hierarchy does not remove the dependent file system.

This is not consistent with other LifeKeeper Recovery Kits that remove the dependent resource when removing the application resource hierarchy.

lklin00001005 Export properties are not shown in the LifeKeeper resource properties.

Export properties are not currently included in the NFS Server RK info field, and customizable resource property sheets are not supported in the LifeKeeper GUI.

lklin00001123 Top level NFS resource hierarchy uses the switchback type of the hanfs resource.

The switchback type which dictates whether the NFS resource hierarchy will automatically switch back to the primary server when it comes back into service after a failure, is defined by the hanfs resource.

lklin00001310 On SuSE, the RPM install of the NFS Server RK must be forced because the nfs-utils dependency cannot be satisfied.

The NFS Server Recovery Kit rpm has a dependency on "nfs-utils >= 0.1.6-2" which causes the installation of the recovery kit to fail on SuSE because the package is named "nfsutils" (not “nfs-utils”). To work around the problem, use the "--nodeps" option from the rpm command line to install the NFS Server Recovery Kit.

NAS Recovery Kit

ID Description

lklin00001310 On SuSE, the RPM install of the NAS Recovery Kit must be forced because the nfs-utils dependency cannot be satisfied.

The NAS Recovery Kit rpm has a dependency on "nfs-utils >= 0.1.6-2" which causes the installation of the recovery kit to fail on SuSE because the package is named "nfsutils" (not “nfs-utils”). To work around the problem, use the "--nodeps" option from the rpm command line to install the NAS Recovery Kit.

Documentation and Training



LifeKeeper for Linux Documentation List

The following is a list of LifeKeeper documentation that provides instructions for installing, configuring, administering, and troubleshooting LifeKeeper Linux software:

Document Available Media*

Release Notes Print, Documentation CD, web

Online Product Manual** LifeKeeper Core CD, Documentation CD, web

Planning and Installation Guide Documentation CD, web

IP Recovery Kit Administration Guide Documentation CD, web

Configuring LifeKeeper Clusters Web only

LifeKeeper for Linux Software Development Kit (SDK)

LifeKeeper SDK CD, web

Apache Web Server Recovery Kit Administration Guide

Apache Recovery Kit CD, Documentation CD, web

Informix Recovery Kit Administration Guide Informix Recovery Kit CD, Documentation CD, web

Oracle Recovery Kit Administration Guide Oracle Recovery Kit CD, Documentation CD, web

Sendmail Recovery Kit Administration Guide Documentation CD, web

Sendmail Advanced Server Recovery Kit Administration Guide

Documentation CD, web

MySQL Recovery Kit Administration Guide MySQL Recovery Kit CD, Documentation CD, web

Print Services Recovery Kit Administration Guide


NFS Server Recovery Kit Administration Guide Documentation CD, web

DB2 Recovery Kit Administration Guide Documentation CD, web

Network Storage Recovery Kit Administration Guide


Data Replication Administration Guide Documentation CD, web

* Documents on the web are available at: www.steeleye.com/support/documentation.


34 Release Notes

** The HTML-based Online Product Manual is delivered in the steeleye-lkHLP package on the LifeKeeper Core CD, and installed in /opt/LifeKeeper/htdoc/help. The Online Product Manual may be accessed outside the LifeKeeper GUI by opening the URL for the product manual web page, http://<server name>:81/help/lksstart.htm, directly from your browser, where <server name> is the hostname of your LifeKeeper server.

LifeKeeper Documentation CD The LifeKeeper Documentation CD provides LifeKeeper documentation in PDF format (except for the Online Product Manual which is HTML-based webhelp), along with the Acrobat Reader. This CD is included with the LifeKeeper Core software. However, to access the latest versions, you should go to the SteelEye Technology, Inc. web site listed below. See the document revision and date to determine the most recent version; for example, Revision B is more recent than Revision A.

Man Pages The reference manual pages delivered with LifeKeeper are in the package steeleye-lkMAN, and are installed in /opt/LifeKeeper/man. The LifeKeeper man pages may be accessed by adding this directory to your MANPATH.

Third Party Documentation Most solutions involving a LifeKeeper environment will also include third party products, e.g. Linux, TCP/IP, servers and miscellaneous peripherals, etc. For specific technical information on these third party products, refer to their documentation for installation, configuration and administrative tasks and procedures.

Training

LifeKeeper training is available through SteelEye Technology, Inc or through your LifeKeeper provider. Contact your sales representative for more information. For an introduction to LifeKeeper, see the LifeKeeper for Linux Tutorial on the web at www.steeleye.com/support.

http://www.steeleye.com/support

Support


Support Support for LifeKeeper is provided by either SteelEye Technology, Inc. or your authorized LifeKeeper Reseller. You can contact SteelEye Support at:

1-877-457-5113 (Toll Free)

1-803-461-3970 (International)

Email: [email protected]

Software Updates Software updates will be made available through SteelEye’s support web page at the following URL:

www.steeleye.com/support

http://www.steeleye.com/support

Read This Document Before Attempting To Install Or Use This...

Documents

Transcript of Read This Document Before Attempting To Install Or Use This...