Linux PXE boot Step by step guide

I wish I could post the screen shots of provisioning/cloning etc using RHN satellite server.
hmm everything is confidential (corporate companies). provisioning using RHN satellite server
is easy and more controllable.




1)Install dhcp and tftp-server
yum install dhcp tftp-server


2)Configure DHCP for PXE boot
Add following entries

#cat /etc/dhcpd.conf

ddns-update-style interim;
allow booting;
allow bootp;

# Standard configuration directives...
subnet 192.168.95.0 netmask 255.255.255.0 {

# option domain-name "nwd.ne.gov";
option subnet-mask 255.255.255.0;
option broadcast-address 192.168.95.255;
option domain-name-servers 10.41.10.105;
# option routers default_router;
next-server 192.168.95.3;


}

# Group the PXE bootable hosts together
group {
# PXE-specific configuration directives...
#next-server 192.168.95.3;
filename "pxelinux.0";

# You need an entry like this for every host
# unless you're using dynamic addresses
host lrhdol1 {
hardware ethernet 00:0C:29:09:AA:E8;
fixed-address 192.168.95.5;
option host-name "lrhdol1";
}
}




3)copy pxelinux.cfg, menu.c32 on to tftpboot directory

#For pxelinux image
cp /usr/lib/syslinux/pxelinux.0 /tftpboot


#For menu (Blue screen with lables)
cp /usr/lib/syslinux/menu.c32 /tftpboot


if you want to use other /tftpboot dir, change in /etc/xinetd.d/tftp file


grep server /etc/xinetd.d/tftp
# description: The tftp server serves files using the trivial file transfer \
server = /usr/sbin/in.tftpd
server_args = -s /tftpboot

and
Edit the file /etc/xinetd.d/tftp and change the line that says “disable = yes” to “disable = no “.


3) create pxelinux.cfg directory for menu options
cd /tftpboot
mkdir pxelinux.cfg
touch 000C2909AAE8
touch C0A85F05
touch C0A85F0
touch C0A85F
....
..
touch C
touch default

4) create Menu file
This I have created for CENT OS and RHEL 5

#cat default
PROMPT 0
TIMEOUT 100
MENU TITLE PXE INSTALL press Esc for boot menu
F1 msgdir/boot.msg
F2 msgdir/options.msg

LABEL localdisk
localboot 1

LABEL RHEL5
kernel linux/RHEL5/vmlinuz
append initrd=linux/RHEL5/initrd.img netboot=nfs nfsroot=192.168.95.3:/media/RHEL5

LABEL CENTOS5
kernel linux/CENTOS5/vmlinuz
append initrd=linux/CENTOS5/initrd.img netboot=nfs nfsroot=192.168.95.3:/media/CENTOS5 ks=nfs:192.168.95.3:/nfsroot/CENTOS5.cfg

LABEL RESCUE
kernel linux/CENTOS5/vmlinuz
append rescue initrd=linux/CENTOS5/initrd.img netboot=nfs nfsroot=192.168.95.3:/media/CENTOS5 ks=nfs:192.168.95.3:/nfsroot/rescue.cfg


4a )

Add *.msg files are text,edit/create according to your needs
# pwd;ls -m
/tftpboot/msgdir
boot.msg, general.msg, options.msg, param.msg, rescue.msg


4b) CENTOS5.cfg,RHEL5.cfg created using system-config-kickstart utility.
cat rescue.cfg
# System keyboard
keyboard us
# System language
lang en_US
# Use NFS installation media
nfs --server=192.168.95.3 --dir=/media/CENTOS5
# Network information
network --bootproto=static --device=eth0 --gateway=192.168.95.2 --ip=192.168.95.5 --nameserver=10.41.10.105 --netmask=255.255.255.0 --onboot=on



6)Mount respective media and do NFS export.
EX:
mount -o loop /dev/cdrom /media/RHEL5/
cat /etc/exports
/media/RHEL5 *(insecure,rw,no_root_squash,async,no_subtree_check)
Run exportfs

7)start dhcpd,xinetd and tftp

service dhcpd start
service xinetd restart
/sbin/chkconfig tftp on

#/sbin/chkconfig --list tftp
tftp on

for DHCP,tftp errors check /var/log/messages file


8) PXE boot server and enjoy.



Directory listing of /tftpboot

#ls -R
.:
linux memdisk menu.c32 msgdir pxelinux.0 pxelinux.cfg

./linux:
CENTOS5 RHEL5

./linux/CENTOS5:
initrd.img README TRANS.TBL vmlinuz

./linux/RHEL5:
initrd.img README TRANS.TBL vmlinuz

./msgdir:
boot.msg general.msg options.msg param.msg rescue.msg

./pxelinux.cfg:
000C2909AAE8 C000025B default

AIX Linux Solaris Cheat sheet

Click Here

Secure and non secure item page

In our environment,
few pages were giving error "

"This page contains both secure and nonsecure items. Do you want to display the nonsecure items?"

after removing this tag <html:base /> the error went away.

configuring devices

AIX

Like panacea, just one command cfgmgr.




Solaris:
Either of the following commands

1 and 2 for newbies :)

1. In Open Boot PROM, use: boot -r
2. With your OS online, reboot your system with this flag:
   reboot -- -r
   (The system boots with the -r flag as it does with boot -r.)
3. Execute these commands: drvconfig, devlinks, disks

view devices and status

 cfgadm -la
prtdiag

network management

I'll try to explain how to create link aggregation a.ka. Ether channe(AIX), Net-bond(Linux).

Before we create one, we have to know, how many interfaces are present in that system.

List network interfaces.

AIX

# lsparent -C -k ent
ent0 Available 03-08 10/100/1000 Base-TX PCI-X Adapter (14106902)
ent1 Available 03-09 10/100/1000 Base-TX PCI-X Adapter (14106902)



Linux

systool -c net
Class = "net"

Class Device = "bond0"
Device = "0000:02:01.0"

Class Device = "eth1"
Device = "0000:02:05.0"

Class Device = "lo"

Class Device = "sit0"


ip link show
1: lo: mtu 16436 qdisc noqueue


link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
2: eth0: mtu 1500 qdisc pfifo_fast qlen 1000
link/ether 00:0c:29:81:c0:13 brd ff:ff:ff:ff:ff:ff
3: sit0: mtu 1480 qdisc noop
link/sit 0.0.0.0 brd 0.0.0.0

Solaris

# dladm show-dev
e1000g0 link: up speed: 1000 Mbps duplex: full
e1000g1 link: up speed: 1000 Mbps duplex: full

change NIC settings to full duplex/Autoneg

AIX
chdev -l 'ent0' -a media_speed='1000_Full_Duplex'

Linux
ethtool -s speed 1000 duplex full autoneg on
ethtool --change eth0 autoneg off speed 1000 duplex full

ethtool eth0 (same as entstat ent0 on AIX)

Solaris
http://www.sun.com/bigadmin/jsp/descFile.jsp?url=descAll/how_to_set_your_nic
http://sysunconfig.net/unixtips/set_nic_settings.txt


Etherchannel configuration




AIX

smitty etherchannel

very easy huh :)

Linux
# ls "$PWD/ifcfg-eth0"
/etc/sysconfig/network-scripts/ifcfg-eth0
# cat ifcfg-eth0

MASTER=bond0
DEVICE=eth0
SLAVE=yes
ONBOOT=yes
USERCTL=no
BOOTPROTO=none

#cat ifcfg-eth1

MASTER=bond0
DEVICE=eth1
SLAVE=yes
ONBOOT=yes
USERCTL=no
BOOTPROTO=none

add/append following lines to /etc/modprode.conf
alias bond0 bonding
options bond0 mode=balance-alb miimon=100

modprobe bonding (for loading bond module)

service network restart
less /proc/net/bonding/bond0
Output
Ethernet Channel Bonding Driver: v3.2.4 (January 28, 2008)

Bonding Mode: load balancing (round-robin)
MII Status: up
MII Polling Interval (ms): 0
Up Delay (ms): 0
Down Delay (ms): 0

Slave Interface: eth0
MII Status: up
Link Failure Count: 0
Permanent HW addr: 00:0c:29:81:c0:13

Slave Interface: eth1
MII Status: up
Link Failure Count: 0
Permanent HW addr: 00:0c:29:81:c0:1d


Refer: http://www.linuxfoundation.org/en/Net:Bonding



Solaris




the last step changes the hostname on to aggr1 .
A reboot is required .

For setting jumboframes
edit MaxFrameSize /kernel/drv/e1000g.conf/hme.conf
edit AutoNegAdvertised for setting full duplex etc


# dladm show-aggr -s
key: 1 ipackets rbytes opackets obytes %ipkts %opkts
Total 3930 408207 2337 222124
e1000g0 3084 309475 1601 158419 78.5 68.5
e1000g1 846 98732 736 63705 21.5 31.5

Check below link for SOlaris dladm and ethtool comparison

http://hub.opensolaris.org/bin/view/Project+brussels/ethtool-dladm-comparison

db2 update

run following command after db2fix pack installation
# ./db2_08_01/instance/db2iupdt db2inst1
DBI1070I Program db2iupdt completed successfully.

Solaris VXVM prerequisites

pkgadd -d /mnt/Solaris_10/Product SUNWscpr SUNWscpu SUNWtoo

pkginfo -d /mnt/Solaris_10/Product SUNWtoo ( required for od)
Above packages are required for Veritas volume-manager installation.

# which od
/usr/bin/od
# pkgchk -l -p /usr/bin/od
Pathname: /usr/bin/od
Type: regular file
Expected mode: 0555
Expected owner: root
Expected group: bin
Expected file size (bytes): 19276
Expected sum(1) of contents: 34966
Expected last modification: Jan 22 21:13:49 2005
Referenced by the following packages:
SUNWtoo
Current status: installed

Solaris Disk management

List disks

printf "\0x1A" | format | grep DEFAULT

or

perl -e "print \"\\x1A\";" | format

0x1A is CTRL \D in ASCII
or
format < /dev/null

Size of a disk and partition/layout info

# prtvtoc /dev/dsk/c1t0d0s2
* /dev/dsk/c1t0d0s2 partition map
*
* Dimensions:
* 512 bytes/sector
* 32 sectors/track
* 128 tracks/cylinder
* 4096 sectors/cylinder
* 4094 cylinders
* 4092 accessible cylinders
*
* Flags:
* 1: unmountable
* 10: read-only
*
* First Sector Last
* Partition Tag Flags Sector Count Sector Mount Directory
0 2 00 1052672 9863168 10915839 /
1 3 01 4096 1048576 1052671
2 5 00 0 16760832 16760831
7 8 00 10915840 5844992 16760831 /export/home
8 1 01 0 4096 4095
# bc
4096*128*32*512 (cyclinders*tracks*sectors*512) -->for total number of bytes of disk space.
8589934592 BYTES
scale=4
8589934592/1024
8388608.0000 KB
8388608.0000/1024
8192.0000 MB
8192.0000/1024
8.0000 GB

scan for new luns /disks (devfsadm -c disk)

# printf "\0x1A" | format | grep DEFAULT
0. c1t0d0
# devfsadm -c disk
# printf "\0x1A" | format | grep DEFAULT
0. c1t0d0
1. c1t1d0

for SAN

devfsadm -C
cfgadm -al     (to get a list of controllers)
cfgadm -c configure c2::500610003412EAB23141
cfgadm -c configure c4::500610003412EAB23142

vxdctl enable     (to tell veritas to look for new disks)
THEN
vea                   (to work with them)

Remove a disk drive (cfgadm)
# printf "\0x1A" | format | grep DEFAULT
0. c1t0d0
1. c1t1d0
# cfgadm -c unconfigure c1::dsk/c1t1d0
# printf "\0x1A" | format | grep DEFAULT
0. c1t0d0

http://www.utahsysadmin.com/2009/05/15/new-lun-solaris9-vxfs-powerpath/

update time using ntpdate

On Linux
# ntpdate -u 10.0.250.10
23 Sep 09:10:21 ntpdate[21154]: step time server 10.0.250.10 offset -68.557099 sec


on AIX
# /usr/sbin/ntpdate 10.0.250.10
23 Sep 09:10:59 ntpdate[442560]: step time server 10.0.250.10 offset 7.100248

filter selective IPs in http access_log

F5 (BIG IP) was filling up http acess logs while doing health check , I have decided not to write this IP in access_log.
avoided images reuest messages as-well.

included following settings in httpd.conf


SetEnvIf Remote_Addr "10\.41\.15\.249" nolog
SetEnvIf Request_URI .gif$ images nolog
#CustomLog logs/access_log common

CustomLog logs/images_log common env=images
CustomLog logs/access_log common env=!nolog

Day to Day

Find and replace

find ./ -name "*.xml" | xargs perl -pi -w -e 's/ port="3268"/ port="389"/g;'

find size of files and sort them in order.

OSR5:
#!/bin/ksh
find $1 -type f -size +2048 -xdev -exec ls -s {} | /bin/sort -brut " "

UW7:

#!/bin/ksh
find $1 -type f -size +2048 -xdev -exec ls -s {} + | /bin/sort -nr

Not sure where I found these one liners.Thanks to the author.

gzip and untar a file in to separate directory , by keeping source as it is.

gzip -dc /tmp/SxRT-5.0MP3-2008-09-11a.dvd4_sfb-sol_x64.tar.gz | tar -xvf -

if you don't want to keep source as it is

just use

tar zxvf /tmp/SxRT-5.0MP3-2008-09-11a.dvd4_sfb-sol_x64.tar.gz

Creating a service in AIX

Wanted to control dmgr with startsrc and stopsrc commands

First create a subsystem using mkssys

#!/bin/ksh
#-R \
mkssys -s dmgr -G was61 -p "/usr/IBM/WebSphere/AppServer/profiles/Dmgr01/bin/dmgrService.sh" -u $(id -u) -S -n 15 -f 9 -Q


#!/bin/ksh
trap "" HUP INT QUIT TSTP STOP
trap "stopService" TERM
stopService ()
{
#cleanup stuff

echo " Cleaning now `date` " >>/tmp/test1.log

#/usr/IBM/WebSphere/AppServer/profiles/Dmgr01/bin/stopManager.sh -username <>
ath.Varanasi -password <>
exit 0

}

#launchScript=/usr/IBM/WebSphere/AppServer/profiles/Dmgr01/bin
numRetries=3

#binDir=`dirname $0`
# Set the ulimit
LIMIT=`ulimit -n`
if [ "${LIMIT}" != "unlimited" ]
then
if [ $LIMIT -lt 1024 ]
then
ulimit -n 1024
fi
fi

RETRY=0
while [ $RETRY -lt $numRetries ]
do
echo launching server using start_dmgr.sh >>/tmp/test1.log
#/usr/IBM/WebSphere/AppServer/profiles/Dmgr01/bin/start_dmgr.sh
echo pid of dmgr is $$ >>/tmp/test1.log
#/bin/ksh exec sleep 3600
/usr/IBM/WebSphere/AppServer/profiles/Dmgr01/bin/sleep.sh

while true
do
ps -aef | grep sleep | grep -v grep
if [ $? -eq 0 ]
then
sleep 1
fi
done
# $binDir/$launchScript
rc=$?
echo exit code: $rc

# Increment retry count on anything other than a normal exit code
if [ $rc -gt 0 ]
then
RETRY=`expr $RETRY + 1`
fi

case $rc in
0) break ;;
esac
done

exit 0

Tomcat Automation

I am maintaining multiple tomcat servers and started automating environment as much as I can, like following.

1)Deploying applications to all servers
2)starting all tomcat servers
3) stoping all tomcat servers
4) checking the status of all tomcat servers.

As you know, Deploying war files across multiple servers consistently, is as much as difficult as
maintaining an unruly class room. Tomcat doesn't provide any means of distributed mechanism like websphere and weblogic application servers. So I am forced to use all unix commands.

One way to deploy application to all servers is, to deploy the application on one server and push same to all servers using rsync.

#!/bin/ksh
###
##
# rsync between host1 and host2 for only
##
###
echo "-------------------`date`----------------" >>/tmp/rsync.log
rsync -gloprtuv -e ssh --stats /usr/tomcat/webapps/ host2:/usr/tomcat/webapps 2>&1 >>/tmp/rsync.log

don't forget to enable password less logins between host1 and host2 by exchanging authorized keys.


2)starting/stoping all tomcat servers

to start and stop either you could do regular scripting "for host in host1,host2....do stopserver.sh
done" or dsh

dsh makes your script simple and easy.

export DSH_NODE_RSH=/usr/bin/ssh
export DSH_NODE_LIST=/usr/tomcatadm/nodes
dsh /usr/tomcat/bin/startup.sh
dsh /usr/tomcat/bin/shutdown.sh

Funny blog name

Well, I wanted to give some meaningful name to my blog related to UNIX.
Most of the names were already taken, So I coined the word halsinx.(not NIX)

H(HP-UX)A(AIX)L(Linux)S(Solaris)nix.

Here I'll be blogging my daily findings, work, cracks etc.