Red Hat Advanced Server

And The

Logical Volume Manager

Cookbook

Chad Kerner

TABLE OF CONTENTS

Business Requirements.......................................................................................................................... 3

Cluster Requirements............................................................................................................................................................................................... 3

Hardware Requirements........................................................................................................................................................................................... 3

Software Requirements............................................................................................................................................................................................. 3

Installation Procedures...................................................................................................................... 4

Kernel Modifications................................................................................................................................................................................................. 4

Physical Disk Layout................................................................................................................................................................................................ 5

LVM Configuration................................................................................................................................................................................................... 6

Cluster Configuration........................................................................................................................... 7

Node 1: CMIL012....................................................................................................................................................................................................... 7

Node 2: CMIL013....................................................................................................................................................................................................... 7

Floating IP Addresses............................................................................................................................................................................................... 7

Service Configuration............................................................................................................................ 8

Appendix A – Cluster Diagram........................................................................................................... 10

Appendix B – Cluster Definition...................................................................................................... 11

Appendix C – Startup Script Modifications.............................................................................. 12

Original File: /etc/rc.d/rc.sysinit......................................................................................................................................................................... 12

Modified File: /etc/rc.d/rc.sysinit......................................................................................................................................................................... 12

Appendix D – Cluster Scripts........................................................................................................... 13

Appendix E – /etc/cluster.conf........................................................................................................ 18

Appendix F – /etc/grub.conf............................................................................................................... 20

Appendix G – Rotating The Cluster Logfile............................................................................. 22

Business Requirements

Cluster Requirements

· The first requirement for this project was that it had to be able to replace a single IBM RS/6000 server with a clustered Linux solution.

· The “services” that were currently running on the RS/6000 needed to be distributed between both nodes in the cluster.

Hardware Requirements

· Compaq 4100 Storage Array – 10 10K RPM disks @ 72Gb each in a RAID 5, and 2 10K RPM disks @ 36Gb in a RAID 1+0. These disks were then carved into 30 logical volumes on the Compaq array. The Compaq 4100 supports 32 devices on the SCSI bus, so you have to leave 2 open id’s for the 2 controllers.

· 2 – Compaq DL580 servers with 4 CPU’s @ 700Mhz each, 2.5Gb RAM each, 2 single port network cards each, and one Compaq Fiber Card for each server.

· Compaq Fiber Hub

· 4 Fiber cables to connect the disks and servers.

· 1 crossover network cable, for the dedicated network heartbeat.

Software Requirements

· Red Hat Advanced Server 2.1 – Patched to the latest level.

· LVM 1.3 RPM

Installation Procedures

Kernel Modifications

1. Install Red Hat Advanced Server(RHAS) on both of the Compaq servers.

2. Download and install all of the latest patches for RHAS.

3. Recompile the kernel to be LVM aware. I used the following steps:

a. cd /usr/src/linux-2.4

b. make mrproper

c. make clean

d. cp configs/kernel-2.4.9-i686-smp.config .config

e. make xconfig

f. make dep

g. depmod

h. make modules

i. make bzImage

j. make modules_install

k. make install

l. cp vmlinux /boot/ vmlinux-2.4.9-e.12custom

m. cd /boot

n. mkinitrd /boot/ initrd-2.4.9-e.12custom.img 2.4.9-e.12custom

4. Edit the /etc/grub.conf file to contain the stanza for the new kernel.

5. Install the latest LVM RPM.

6. Edit /etc/rc.d/rc.sysinit and comment out the LVM startup. You want to let the clustering scripts take care of the LVM disks. NOTE: THIS IS EXTREMELY IMPORTANT. IF YOU DO NOT DO THIS, YOU WILL CORRUPT YOUR DATA!!

7. Reboot the machine with the LVM aware kernel.

Physical Disk Layout

Device	Size(Gb)	Volume Group
/dev/sda	0.64	None (Raw)
/dev/sdb	8.00
/dev/sdc	8.00
/dev/sdd	8.00	/dev/vgfti
/dev/sde	8.00	/dev/vgdba
/dev/sdf	1.90
/dev/sdg	40.00	/dev/vgcdr
/dev/sdh	40.00	/dev/vgcdr
/dev/sdi	40.00	/dev/vgcdr
/dev/sdj	40.00	/dev/vgcdr
/dev/sdk	40.00	/dev/vgcdr
/dev/sdl	40.00	/dev/vgfti
/dev/sdm	40.00	/dev/vgfti
/dev/sdn	40.00	/dev/vgfti
/dev/sdo	40.00	/dev/vgfti
/dev/sdp	40.00	/dev/vgfti
/dev/sdq	40.00
/dev/sdr	20.00	/dev/vgcdr
/dev/sds	20.00	/dev/vgfti
/dev/sdt	20.00	/dev/vgfti
/dev/sdu	20.00	/dev/vgfti
/dev/sdv	20.00
/dev/sdw	10.00	/dev/vgdba
/dev/sdx	10.00	/dev/vgdba
/dev/sdy	10.00	/dev/vgdba
/dev/sdz	10.00
/dev/sdaa	10.00
/dev/sdab	10.00
/dev/sdac	4.00
/dev/sdad	6.53

Total Disk Space	645.07

LVM Configuration

1. Create physical volumes on the new disks, except for the disk that is to contain the raw devices for the clustering software(/dev/sda). For this, I did a simple for loop: “for i in b c d e f g h i j k l m n o p q r s t u v w x y z aa ab ac ad; do pvcreate /dev/sd$i; done”

2. vgcreate -s 8M vgcdr /dev/sdg /dev/sdh /dev/sdi /dev/sdj /dev/sdk /dev/sdr

3. lvcreate -L 10G -n cdr vgcdr

4. lvcreate -L 200G -n cdr01 vgcdr

5. vgcreate -s 8M vgfti /dev/sdd /dev/sdl /dev/sdm /dev/sdn /dev/sdo /dev/sdp /dev/sds /dev/sdt /dev/sdu

6. lvcreate -L 10G -n fti vgfti

7. lvcreate -L 150G -n fti01 vgfti

8. lvcreate -L 20G -n vad01 vgfti

9. lvcreate -L 5G -n fulf vgfti

10. lvcreate -L 75G -n work vgfti

11. vgcreate vgdba /dev/sde /dev/sdw /dev/sdx /dev/sdy

12. lvcreate -L 30G -n dba01 vgdba

13. mke2fs -j /dev/vgcdr/cdr

14. mke2fs -j /dev/vgcdr/cdr01

15. mke2fs -j /dev/vgfti/fti

16. mke2fs -j /dev/vgfti/fti01

17. mke2fs -j /dev/vgfti/vad01

18. mke2fs -j /dev/vgfti/fulf

19. mke2fs -j /dev/vgfti/work

20. mke2fs -j /dev/vgdba/dba01

21. Create the mount points: “for a in cdr cdr01 fti fti01 vad01 fulf work dba01; do mkdir /$a; done”

Things To Note:

· Don’t put the mounts into /etc/fstab. Let the cluster script take care of it.

· Don’t activate the same volume group on 2 different machines at the same time or you will corrupt the Volume Group Descriptor Area(VGDA).

Cluster Configuration

Node 1: CMIL012

Interface	IP Address	Netmask	Broadcast	Network
Eth0	10.120.9.131	255.255.255.0	10.120.9.255	10.120.9.0
Eth1	192.168.0.1	255.255.255.0	192.168.0.255	192.168.0.0

Node 2: CMIL013

Interface	IP Address	Netmask	Broadcast	Network
Eth0	10.120.9.132	255.255.255.0	10.120.9.255	10.120.9.0
Eth1	192.168.0.2	255.255.255.0	192.168.0.255	192.168.0.0

Floating IP Addresses

DNS Alias (Package)	IP Address	Primary Node
CMILC01 (Cluster Alias)	10.120.9.130	CMILC01
TRANS_CDR (cluster_cdr)	10.120.9.133	CMIL012
TRANS_FTI (cluster_fti)	10.120.9.134	CMIL013
TRANS_DBA (cluster_dba)	10.120.9.135	CMIL012

Service Configuration

For this cluster of machines, there are 3 services that can be failed back and forth. Each of the services is named cluster_<service> and has an associated volume group called vg<service>.

The following is the output from the “cluadmin – service show config cluster_<service>” commands.

preferred node: cmil012

relocate: yes

user script: /usr/local/bin/cluster_cdr

monitor interval: 300

IP address 0: 10.120.9.133

netmask 0: 255.255.255.0

broadcast 0: 10.120.9.255

preferred node: cmil012

relocate: yes

user script: /usr/local/bin/cluster_dba

monitor interval: 300

IP address 0: 10.120.9.135

netmask 0: 255.255.255.0

broadcast 0: 10.120.9.255

preferred node: cmil013

relocate: yes

user script: /usr/local/bin/cluster_fti

monitor interval: 300

IP address 0: 10.120.9.134

netmask 0: 255.255.255.0

broadcast 0: 10.120.9.255

Summary Of Critical Elements

· Build all of the services on a single node to begin with. Get the scripts working there before trying to move it to a different node. It will just make debugging that much easier.

· Make sure that you have adequate logging in your script so that you can determine what went wrong. If you look at the script below, it does a tremendous amount of logging. However, start and stop are calling the cluster logging utility with a loglevel of 6(informational) and the status script calls the cluster logging utility with a loglevel of 7(debug). You can control the level of logging in the logfile with the “cluadmin – cluster loglevel” command.

· If you are going to do any logical volume and/or volume group work, verify that the volume group is only active on one system at a time. If it is active in both places, it can potentially cause data loss of the volume group.

· The Java gui, while a nicety, consumes too much CPU. I recommend informing people about the clustat command instead.

· Take backups weekly of the quorum data if you are changing it often. You can use it to rebuild the quorum partitions if they get corrupted for some reason. You can use a simple shell script like the following to accomplish this via cron:

#!/bin/bash

DATE=`date +"%Y%m%d"`

CLUADMIN=`type -p cluadmin`

FILENAME="/root/cluster.conf.${DATE}"

${CLUADMIN} -- cluster saveas ${FILENAME}

Appendix A – Cluster Diagram

Appendix B – Cluster Definition

· Configure WU-FTP on both nodes in the cluster.

· Make sure the customized cluster scripts are in sync on both servers.

· Make sure all of the mount points exist on both servers.

Appendix C – Startup Script Modifications

You do not want the system to automatically try and mount the LVM controlled disks. This WILL cause corruption in the VGDA and you will have to reboot the box to correct the corruption. To correct this, /etc/rc.d/rc.sysint was modified to bypass the LVM startup on each of the nodes in the cluster.

Original File: /etc/rc.d/rc.sysinit

# Remount the root filesystem read-write.

state=`awk '/(^\/dev\/root| \/ )/ { print $4 }' /proc/mounts`

[ "$state" != "rw" ] && \

action $"Remounting root filesystem in read-write mode: " mount -n -o remount,rw /

# LVM initialization

if [ -e /proc/lvm -a -x /sbin/vgchange -a -f /etc/lvmtab ]; then

action $"Setting up Logical Volume Management:" /sbin/vgscan && \ /sbin/vgchange -a y

# Clear mtab

>/etc/mtab

Modified File: /etc/rc.d/rc.sysinit

# Remount the root filesystem read-write.

state=`awk '/(^\/dev\/root| \/ )/ { print $4 }' /proc/mounts`

[ "$state" != "rw" ] && \

action $"Remounting root filesystem in read-write mode: " mount -n -o remount,rw /

# LVM initialization

#if [ -e /proc/lvm -a -x /sbin/vgchange -a -f /etc/lvmtab ]; then

# action $"Setting up Logical Volume Management:" /sbin/vgscan && \ #/sbin/vgchange -a y

#fi

# Clear mtab

>/etc/mtab

Appendix D – Cluster Scripts

The script below is used by the cluster package to start, stop and query one of the services that are executing under the control of the cluster.

#!/bin/bash

#----------------------------------------------------------------------------#

# cluster_cdr: This script will handle the failover of the the packages on #

# the linux cluster. #

#----------------------------------------------------------------------------#

# 20030310 Chad Kerner - Initial Coding #

# #

#----------------------------------------------------------------------------#

# Set the log level variables for the clulog command. #

#----------------------------------------------------------------------------#

LOG_EMERG=0 # system is unusable

LOG_ALERT=1 # action must be taken immediately

LOG_CRIT=2 # critical conditions

LOG_ERR=3 # error conditions

LOG_WARNING=4 # warning conditions

LOG_NOTICE=5 # normal but significant condition

LOG_INFO=6 # informational

LOG_DEBUG=7 # debug-level messages

PACKAGE="CLUSTER_CDR" # Name of the package referenced by this script.

PKG_VG="vgcdr" # Volume Group For The Package

PKG_LVS="cdr cdr01" # Logical Volumes In The Volume Group

TXMIT_START="/usr/local/bin/trans_cdr" # Start/Stop script for the application

PERFORM_FSCK=1 # 1=Do an FSCK, 0 = don't do an fsck

#----------------------------------------------------------------------------#

# Find the executables required for this script. #

#----------------------------------------------------------------------------#

CLULOG=`type -p clulog`

VGSCAN=`type -p vgscan`

VGCHANGE=`type -p vgchange`

VGDISPLAY=`type -p vgdisplay`

LVCHANGE=`type -p lvchange`

LVDISPLAY=`type -p lvdisplay`

DFCMD=`type -p df`

MOUNT=`type -p mount`

UMOUNT=`type -p umount`

E2FSCK=`type -p e2fsck`

FSCKOPTS="-p -y"

SYNCCMD=`type -p sync`

TCH_CMD=`type -p touch`

RM_CMD=`type -p rm`

#----------------------------------------------------------------------------#

# This routine is to format "standardized" messages to the logfile. #

#----------------------------------------------------------------------------#

log_msg () {

MSG=$1

${CLULOG} -s ${MSG_LEVEL} "${PACKAGE} ==> ${MSG}"

return 0;

}

#----------------------------------------------------------------------------#

# This routine will scan the disks and rebuild the lvm configuration. It #

# will then activate the volume group. Then it will mount the logical #

# volumes under the correct mountpoints. #

#----------------------------------------------------------------------------#

start () {

MSG_LEVEL=6

log_msg "Starting the package"

log_msg "Creating the lock file"

${TCH_CMD} /var/run/${PACKAGE}

RC=$?

log_msg "Return Code: ${RC} From ${TCH_CMD} /var/run/${PACKAGE}"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

log_msg "Executing: ${VGSCAN}"

${VGSCAN}

RC=$?

log_msg "Return Code: ${RC} From ${VGSCAN}"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

log_msg "Executing: ${VGCHANGE} -a y ${PKG_VG}"

${VGCHANGE} -a y ${PKG_VG}

RC=$?

log_msg "Return Code: ${RC} From ${VGCHANGE} -a y ${PKG_VG}"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

for DISKS in `echo ${PKG_LVS}`

do log_msg "Calling: mount_vols $DISKS"

mount_vols $DISKS

RC=$?

log_msg "Return Code: ${RC} From mount_vols $DISKS"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

done

log_msg "Calling: ${TXMIT_START} start script"

${TXMIT_START} start

RC=$?

log_msg "Return Code: ${RC} From ${TXMIT_START} start"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

log_msg "Executing: ${RM_CMD} -f /var/run/${PACKAGE}"

${RM_CMD} -f /var/run/${PACKAGE}

RC=$?

log_msg "Return Code: ${RC} From ${RM_CMD} -f /var/run/${PACKAGE}"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

log_msg "Lock File Has Been Removed"

log_msg "Package Startup Complete"

return 0

}

#----------------------------------------------------------------------------#

# This routine will perform a file system integrity check and then mount the #

# specified logical volume to the designated mount point. #

#----------------------------------------------------------------------------#

mount_vols () {

lvol=$1

if [ "${PERFORM_FSCK}" -eq "1" ] ; then

log_msg "Executing: ${E2FSCK} ${FSCKOPTS} ${lvol}"

${E2FSCK} ${FSCKOPTS} /dev/${PKG_VG}/${lvol}

RC=$?

log_msg "Return Code: ${RC} From ${E2FSCK} ${FSCKOPTS} ${lvol}"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

log_msg "Executing: ${MOUNT} /dev/${PKG_VG}/${lvol} /${lvol}"

${MOUNT} /dev/${PKG_VG}/${lvol} /${lvol}

RC=$?

log_msg "Return Code: ${RC} From ${MOUNT} /dev/${PKG_VG}/${lvol} /${lvol}"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

return 0;

}

#----------------------------------------------------------------------------#

# This routine will unmount the specified mount point. #

#----------------------------------------------------------------------------#

do_umount () {

lvol=$1

log_msg "Executing: ${UMOUNT} /${lvol}"

${UMOUNT} /${lvol}

RC=$?

log_msg "Return Code: ${RC} From ${UMOUNT} /${lvol}"

if [ ${RC} -gt 1 ] ; then return ${RC}; fi

log_msg "Executing: ${LVCHANGE} -a n /dev/${PKG_VG}/${lvol}"

${LVCHANGE} -a n /dev/${PKG_VG}/${>lvol}

RC=$?

log_msg "Return Code: ${RC} From ${LVCHANGE} -a n /dev/${PKG_VG}/${lvol}"

if [ ${RC} -gt 1 ] ; then return ${RC}; fi

return 0

}

#----------------------------------------------------------------------------#

# This routine will unmount the filesystems contained in the volume #

# group and then mark the volume group inactive so that it can be activated #

# on the new node of the cluster. #

#----------------------------------------------------------------------------#

stop () {

MSG_LEVEL=6

log_msg "Stopping The Package"

log_msg "Calling ${TXMIT_START} stop"

${TXMIT_START} stop

RC=$?

log_msg "Return Code: ${RC} From ${TXMIT_START} stop"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

log_msg "Executing: ${SYNCCMD} to flush buffer cache"

for i in $(seq 1 3)

do ${SYNCCMD}

RC=$?

done

log_msg "Return Code: ${RC} From ${SYNCCMD}"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

for DISKS in `echo ${PKG_LVS}`

do log_msg "Calling: do_umount $DISKS"

do_umount $DISKS

RC=$?

log_msg "Return Code: ${RC} From do_umount $DISKS"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

done

log_msg "Executing: ${VGCHANGE} -a n ${PKG_VG}"

${VGCHANGE} -a n ${PKG_VG}

RC=$?

log_msg "Return Code: ${RC} From ${VGCHANGE} -a n ${PKG_VG}"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

log_msg "Package Shutdown Complete"

return 0

}

#----------------------------------------------------------------------------#

# A status command is a required command for clustering, probably so it can #

# attempt to check the status of a node. Therefore, we'll just issue a df #

# for the filesystems and a vgdisplay for the volume group. #

#----------------------------------------------------------------------------#

status () {

MSG_LEVEL=7

log_msg "Check Package Status"

if [ -e /var/run/${PACKAGE} ] ; then

log_msg "Package Has Not Completely Started Yet."

return 0

log_msg "Executing: ${VGDISPLAY} -D ${PKG_VG}"

${VGDISPLAY} -D ${PKG_VG}

RC=$?

log_msg "Return Code: ${RC} From ${VGDISPLAY} -D ${PKG_VG}"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

for DISKS in `echo ${PKG_LVS}`

do log_msg "Executing: ${LVDISPLAY} /dev/${PKG_VG}/$DISKS"

${LVDISPLAY} /dev/${PKG_VG}/$DISKS

RC=$?

log_msg "Return Code: ${RC} From ${LVDISPLAY} /dev/${PKG_VG}/$DISKS"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

log_msg "Executing: ${DFCMD} -h /$DISKS"

${DFCMD} -h /$DISKS

RC=$?

log_msg "Return Code: ${RC} From ${DFCMD} -h /$DISKS"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

done

log_msg "Calling: ${TXMIT_START} status"

${TXMIT_START} status

RC=$?

log_msg "Return Code: ${RC} From ${TXMIT_START} status"

if [ ${RC} -gt 0 ] ; then return ${RC}; fi

log_msg "Package Status Successful"

return 0

}

#----------------------------------------------------------------------------#

# This routine will stop and restart the process. #

#----------------------------------------------------------------------------#

restart () {

stop

start

RETVAL=$?

return $RETVAL

}

case "$1" in

start)

start

;;

stop)

stop

;;

status)

status

;;

restart)

restart

;;

echo $"Usage: $0 { start | stop | status | restart }"

RETVAL=1

esac

exit $RETVAL

Appendix E – /etc/cluster.conf

[cluhbd]

logLevel = 6

[clupowerd]

logLevel = 6

[cluquorumd]

logLevel = 6

[clurmtabd]

logLevel = 6

[cluster]

alias_ip = 10.120.9.130

name = cmilc01

timestamp = 1047472053

[clusvcmgrd]

logLevel = 6

[database]

version = 2.0

[members]

start member0

start chan0

name = cmil012h

type = net

end chan0

id = 0

name = cmil012

powerSwitchIPaddr = cmil012

powerSwitchPortName = unused

quorumPartitionPrimary = /dev/raw/raw1

quorumPartitionShadow = /dev/raw/raw2

end member0

start member1

start chan0

name = cmil013h

type = net

end chan0

id = 1

name = cmil013

powerSwitchIPaddr = cmil013

powerSwitchPortName = unused

quorumPartitionPrimary = /dev/raw/raw1

quorumPartitionShadow = /dev/raw/raw2

end member1

[powercontrollers]

start powercontroller0

IPaddr = cmil012

passwd = unused

type = null

end powercontroller0

start powercontroller1

IPaddr = cmil013

passwd = unused

type = null

end powercontroller1

[services]

start service0

checkInterval = 300

name = cluster_cdr

start network0

broadcast = 10.120.9.255

ipAddress = 10.120.9.133

netmask = 255.255.255.0

end network0

preferredNode = cmil012

relocateOnPreferredNodeBoot = yes

userScript = /usr/local/bin/cluster_cdr

end service0

start service1

checkInterval = 300

name = cluster_dba

start network0

broadcast = 10.120.9.255

ipAddress = 10.120.9.135

netmask = 255.255.255.0

end network0

preferredNode = cmil012

relocateOnPreferredNodeBoot = yes

userScript = /usr/local/bin/cluster_dba

end service1

start service2

checkInterval = 300

name = cluster_fti

start network0

broadcast = 10.120.9.255

ipAddress = 10.120.9.134

netmask = 255.255.255.0

end network0

preferredNode = cmil013

relocateOnPreferredNodeBoot = yes

userScript = /usr/local/bin/cluster_fti

end service2

Appendix F – /etc/grub.conf

# grub.conf generated by anaconda

# Note that you do not have to rerun grub after making changes to this file

# NOTICE: You have a /boot partition. This means that

# all kernel and initrd paths are relative to /boot/, eg.

# root (hd0,0)

# kernel /vmlinuz-version ro root=/dev/ida/c0d0p7

# initrd /initrd-version.img

#boot=/dev/ida/c0d0

default=0

fallback=1

timeout=10

splashimage=(hd0,0)/grub/splash.xpm.gz

title Red Hat Advanced Server - LVM Kernel (2.4.9-e.12custom)

root (hd0,0)

kernel /vmlinuz-2.4.9-e.12custom ro root=/dev/ida/c0d0p7

initrd /initrd-2.4.9-e.12custom.img

title Red Hat Linux (2.4.9-e.12smp)

root (hd0,0)

kernel /vmlinuz-2.4.9-e.12smp ro root=/dev/ida/c0d0p7

initrd /initrd-2.4.9-e.12smp.img

title Red Hat Linux (2.4.9-e.12debug)

root (hd0,0)

kernel /vmlinuz-2.4.9-e.12debug ro root=/dev/ida/c0d0p7

initrd /initrd-2.4.9-e.12debug.img

title Red Hat Linux (2.4.9-e.12)

root (hd0,0)

kernel /vmlinuz-2.4.9-e.12 ro root=/dev/ida/c0d0p7

initrd /initrd-2.4.9-e.12.img

title Red Hat Linux Advanced Server (2.4.9-e.3smp)

root (hd0,0)

kernel /vmlinuz-2.4.9-e.3smp ro root=/dev/ida/c0d0p7

initrd /initrd-2.4.9-e.3smp.img

title Red Hat Linux Advanced Server-up (2.4.9-e.3)

root (hd0,0)

kernel /vmlinuz-2.4.9-e.3 ro root=/dev/ida/c0d0p7

initrd /initrd-2.4.9-e.3.img

Appendix G – Rotating The Cluster Log File

If you want the cluster log to be rotated correctly, the best place to do it is with the rest of the syslog files. The file to edit is: /etc/logrotate.d/syslog . Here is the syslog file before it was edited.

/var/log/messages /var/log/secure /var/log/maillog \ /var/log/spooler /var/log/boot.log /var/log/cron {

sharedscripts

postrotate

/bin/kill -HUP `cat /var/run/syslogd.pid 2> /dev/null` \ 2> /dev/null || true

endscript

}

Here is the syslog file after it was edited.

/var/log/messages /var/log/secure /var/log/maillog \ /var/log/spooler /var/log/boot.log /var/log/cron \ /var/log/cluster {

sharedscripts

postrotate

/bin/kill -HUP `cat /var/run/syslogd.pid 2> /dev/null` \ 2> /dev/null || true

endscript

}