Linux Virtual Server

Install LVS

It involves with patching kernel with code related to ip-vs modules, compile ip-vs loadable modules (each scheduling is a separate modules) and ipvsdam program.

1. Upgrade kernel to 2.4.21

1.1. Put kernel source code in /usr/src/linux directory

I have downloaded linux-2.4.21.tar and related patach files ipvs-1.0.10.tar.gz to /root/lvs/ you can skip kernel source and ipvs patch download steps.
Those red mark commands are instructions to be executed. Make sure you understand what are their functions.
cd /root/lvs
- If you try to do it at home, download 2.4.21 kernel source from http://www.kernel.org/pub/linux/kernel/v2.4/linux-2.4.21.tar.bz2, 27.2MB
uncompress kernel source
bunzip2 linux-2.4.21.tar.bz2
The result is 159MB
tar xvf linux-2.4.21.tar
The following are the kernel installation steps in http://www.tldp.org/HOWTO/Kernel-HOWTO/quick_steps.html#2.4
mv linux-2.4.21 /usr/src
create soft link
cd /usr/src
ln -s linux-2.4.21 linux
- This is crucial step. If you create a soft link by type "ls -s linux-2.4.20-6 linux" you will compile a different older linux later.
- If you have compiled kernel before you should backup previous version kernel configuration files and reuse them.
  - mkdir /usr/src/kernelconfigs
  - cp /usr/src/linux/.config /usr/src/kernelconfigs/.config.save
    - Since our virtual machine was installed from redhat 9 CD and its kernel was never upgraded or recompiled, there is no .config file.Redhat 9 puts its kernel in directory linux-2.4 instead of linux (different from what the typical kernel how-to documents, maybe for a good reason).
    - The 2.4.21 kernel source code downloaded from kernel.org is a fresh copy, it was never upgraded or recompiled, therefore there is no .config
    - The above backup operation is for systems whose kernels have been upgraded or recompiled.
    - The configuration file for different version of kernels provided in redhat 9 are in /usr/src/linux-2.4/configs directory. They consists of
      - kernel-2.4.20-athlon.config kernel-2.4.20-i586-smp.config
        kernel-2.4.20-athlon-smp.config kernel-2.4.20-i686-bigmem.config
        kernel-2.4.20-i386-BOOT.config kernel-2.4.20-i686.config
        kernel-2.4.20-i386.config kernel-2.4.20-i686-smp.config
        kernel-2.4.20-i386-smp.config kernel-2.4.20-x86_64.config
        kernel-2.4.20-i586.config kernel-2.4.20-x86_64-smp.config
      - various of common architecture types, smp.

cp /usr/src/linux-2.4/.config /usr/src/kernelconfigs/.config.save

cd /usr/src/linux

1.2. Backup Existing Kernel Configurations and Reuse a Specific Existing Kernel Configuration

mkdir configs
cp /boot/config* configs
- This will copy /boot/config-2.4.20-6 and /boot/config-2.4.20-6smp to /usr/src/linux/configs directory as backup copies.
- The redhat 9 /boot directory contains the information about the kernels that are being used, config file, initrd, module-info, system.map, and the actual kernel (compressed) images to be loaded by bootloader at boot time.
- [root@lvs boot]# ls
  boot.b lost+found System.map-2.4.20-6
  chain.b message System.map-2.4.20-6smp
  config-2.4.20-6 message.ja vmlinux-2.4.20-6
  config-2.4.20-6smp module-info vmlinux-2.4.20-6smp
  grub module-info-2.4.20-6 vmlinuz
  initrd-2.4.20-6.img module-info-2.4.20-6smp vmlinuz-2.4.20-6
  initrd-2.4.20-6smp.img os2_d.b vmlinuz-2.4.20-6smp
  kernel.h System.map
- The redhat9 distribution is used grub loader. It uses /etc/grub.conf as configuration file. The /etc/grub.conf shows
  # 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/sda2
  # initrd /initrd-version.img
  #boot=/dev/sda
  default=0
  timeout=10
  splashimage=(hd0,0)/grub/splash.xpm.gz
  password --md5 $1$yp4y9ywr$GYbVD9RklLLfm6Vy6SkZe.
  title Red Hat Linux (2.4.20-6smp)
         root (hd0,0)
         kernel /vmlinuz-2.4.20-6smp ro root=LABEL=/
         initrd /initrd-2.4.20-6smp.img
  title Red Hat Linux-up (2.4.20-6)
         root (hd0,0)
         kernel /vmlinuz-2.4.20-6 ro root=LABEL=/
         initrd /initrd-2.4.20-6.img
- The title sections show there are two kernel versions are being configured to be used in boot time. One is Rehat linux 2.4.20-6 smp, the other is redhat linux 2.4.20-6. The first one with symmetric processor support and is the defulat one used (default =0). The password directive indicates that we are use md5 password to protect grub loader.
- An initrd image is needed for loading your SCSI module at boot time or if you are compiling the kernel with ext3 support as a module. In our virtual machine, both are true.
  - The vmware virtual machine uses virtual disk (a local file) and emulates it as scsi disk.
  - The redhat9 by default configures the file system as ext3.
  - The lsmod command lists all modules used. It shows indeed ext3 and scsi_mod are there in our lvs virtual machine.
  - ```
  [root@lvs boot]# lsmod
  Module                  Size  Used by    Not tainted
  parport_pc             19204   1  (autoclean)
  lp                      9188   0  (autoclean)
  parport                39072   1  (autoclean) [parport_pc lp]
  autofs                 13684   0  (autoclean) (unused)
  pcnet32                18592   3
  mii                     4124   0  [pcnet32]
  keybdev                 2976   0  (unused)
  mousedev                5656   0  (unused)
  hid                    22308   0  (unused)
  input                   6208   0  [keybdev mousedev hid]
  usb-uhci               27404   0  (unused)
  usbcore                82592   1  [hid usb-uhci]
  ext3                   73376   2
  jbd                    56336   2  [ext3]
  BusLogic              100988   3
  sd_mod                 13452   6
  scsi_mod              110488   2  [BusLogic sd_mod]
```
cp /boot/config-2.4.20-6smp /usr/src/linux/.config

We reuse the configuration file config-2.4.20-6smp for our kernel compilation.

1.3 Apply IP-VS patches

cd /root/lvs
tar xvzf ipvs-1.0.10.tar.gz
mv ipvs-1.0.10 /usr/src
- If you work at home, download the The IPVS Netfilter module for kernel 2.4 - Version 1.0.10,( ipvs-1.0.10.tar.gz) and
  linux-2.4.21-ipvs-1.0.10.patch.gz from http://www.linuxvirtualserver.org/software/ipvs.html#kernel-2.4
download gunzip linux-2.4.21-ipvs-1.0.10.patch.gz, you can use wget command
wget http://www.linuxvirtualserver.org/software/kernel-2.4/linux-2.4.21-ipvs-1.0.10.patch.gz
gunzip linux-2.4.21-ipvs-1.0.10.patch.gz
mv linux-2.4.21-ipvs-1.0.10.patch /usr/src
Apply patches with the following commands
cd /usr/src/linux
make sure you are in /usr/src/linux to apply these patches
cat /usr/src/ipvs-1.0.10/linux_kernel_ksyms_c.diff | patch -p1

you should see the following if you are sucessful in applying the patch
[root@rh9 linux]# cat /usr/src/ipvs-1.0.10/linux_kernel_ksyms_c.diff | patch -p1
patching file kernel/ksyms.c
Hunk #1 succeeded at 145 with fuzz 2 (offset 18 lines).
So long as your patch succeed, you can proceed. The offset line # may be different. Let me know what you got.

cat /usr/src/ipvs-1.0.10/linux_net_netsyms_c.diff | patch -p1

you should see the following if you are sucessful in applying the patch
[root@rh9 linux]# cat /usr/src/ipvs-1.0.10/linux_net_netsyms_c.diff | patch -p1
patching file net/netsyms.c
Hunk #1 succeeded at 263 with fuzz 2 (offset 15 lines).

cat /usr/src/linux-2.4.21-ipvs-1.0.10.patch | patch -p1
patching file Documentation/Configure.help
Hunk #1 succeeded at 3074 (offset -22 lines).
patching file include/net/ip_vs.h
patching file kernel/ksyms.c
Reversed (or previously applied) patch detected! Assume -R? [n] Enter
Apply anyway? [n] Enter
Skipping patch.
1 out of 1 hunk ignored -- saving rejects to file kernel/ksyms.c.rej
......
patching file net/netsyms.c
Reversed (or previously applied) patch detected! Assume -R? [n] Enter
Apply anyway? [n] Enter
......

1.4. Kernel Compilation

Next we will compile the Linux kernel. See http://www.tldp.org/HOWTO/Kernel-HOWTO/quick_steps.html for more detailed info.
make menuconfig
- This menuconfig will compile menuconfig program. The program allows you to pick and choose features to be hardcoded in kernels and those to be compiled as dynamic modules.
- You can also use xconfig, probably faster, but require x window)
- When kernel configuration menu appear, move cursor to "Networking Option --->" Hit enter.
- Move cursor to "IP: Netfilter Configuration --->" (about two pages down)

If you did not see IP:Netfilter, you may have forget to copy /boot/config-2.4.20-6smp file as .config in /usr/src/linux.

Move cursor to "ipchains (2.2-style) support" at the bottom of the list. Hit N to exclude the features.
Move cursor to "ipfwadm (2.0-style) support". Hit N to exclude the features.
- These two features are replaced by iptables.
Exit from the "IP: Netfilter Configuration" submenu by hit ESC key.
Move cursor to the next "IP: Virtual Server Configuration --->" menu. Hit enter to select it.
Move cursor to "IP virtual server debugging" and hit Y to include the feature.
Exit from the "IP: Virtual Server Configuration" submenu by hit ESC key.
Hit ESC key again to conclude the network option configuration.
Hit ESC key again to conclude the kernel configuration.
Hit Enter to accept "Yes" for saving the new kernel configuration.

make dep
- You should see a lot of genksyms (generate kernel symbol comamnd) like
  | /sbin/genksyms -p smp_ -k 2.4.21 > /usr/src/linux-2.4.21/include/linux/modules/ide-lib.ver.tmp
- This prepare makefiles and the source code (both kernel and module) for the kernel version 2.4.21.
- This will take a few minutes. In my virtual machine it is 6m28s.
Give a unique name to your new Kernel -
vi /usr/src/linux/Makefile
Set EXTRAVERSION =-lvs
- We should set it to -lvs (with '-') since with '-' the kernel version name become 2.4.21-lvs, without '-' it becomes 2.4.21lvs.
nohup make bzImage
- nohup - run a command immune to hangups and quits
- output will be redirect to nohup.out.
- You can still observe the compilation by "tail -f nohup.out" on another window.
- Making kernel image will take a long time.
- It takes about 7 min. on ardor virtual machine. On my slow 1.3Ghz toshiba tablet pc is 17m22s.
make modules
- It takes pretty long time 30-40min at least to compile all loadable modules (network driver, IOdriver, file systems).On my slow 1.3Ghz toshiba tablet pc is 86m16s.
- At the end, it may indicate error in compiling the ip_vs module. That is ok. Actually the ip_vs should be compiled after the new kernel is running.
- The following shows the end of "make modules"
- make[2]: Leaving directory `/usr/src/linux-2.4.21-lvs/net/ipv4'
  make -C ipv4/ipvs modules
  make[2]: Entering directory `/usr/src/linux-2.4.21-lvs/net/ipv4/ipvs'
  gcc -D__KERNEL__ -I/usr/src/linux-2.4.21-lvs/include -Wall -Wstrict-prototypes -Wno-trigraphs -O2 -fno-strict-aliasing -fno-common -fomit-frame-pointer -pipe -mpreferred-stack-boundary=2 -march=i686 -DMODULE -DMODVERSIONS -include /usr/src/linux-2.4.21-lvs/include/linux/modversions.h -nostdinc -iwithprefix include -DKBUILD_BASENAME=ip_vs_conn -c -o ip_vs_conn.o ip_vs_conn.c
  ip_vs_conn.c:46:25: linux/jhash.h: No such file or directory
  ip_vs_conn.c: In function `ip_vs_conn_hashkey':
  ip_vs_conn.c:132: warning: implicit declaration of function `jhash_3words'
  make[2]: *** [ip_vs_conn.o] Error 1
  make[2]: Leaving directory `/usr/src/linux-2.4.21-lvs/net/ipv4/ipvs'
  make[1]: *** [_modsubdir_ipv4/ipvs] Error 2
  make[1]: Leaving directory `/usr/src/linux-2.4.21-lvs/net'
  make: *** [_mod_net] Error 2
  real 86m15.963s
  user 44m29.660s
  sys 36m52.840s
make modules_install
- It moves the compiled modules to /lib/modules/2.4.21-lvs
cp /usr/src/linux/arch/i386/boot/bzImage /boot/vmlinuz-2.4.21-lvs
cp /usr/src/linux/System.map /boot/System.map-2.4.21-lvs
mkinitrd ./initrd-2.4.21-lvs.img 2.4.21-lvs
cp initrd-2.4.21-lvs.img /boot
cd /boot
mv System.map System.map.bak
ln -s System.map-2.4.21-lvs System.map
cp /usr/src/linux/.config /boot/config-2.4.21-lvs (back it up)

1.5. Add new compiled kernel to the list managed by GRUB loader

Configure GRUB with new kernel. Note that the following step is critical. Incorrect editing of /etc/grub.conf can render the virtual machine not bootable even to the older version of the kernel. Make sure you donot chagne other kernel version specifications in the grub.conf file. Just add the new entry. If something wrong with the new kernel images or configuration, we can still go back to the old kernels to fix the problem.
edit /etc/grub.conf
add the following entry right after "password" line (as the default entry)
title Red Hat Linux (2.4.2-1lvs)
       root (hd0,0)
       kernel /vmlinuz-2.4.21-lvs ro root=LABEL=/
       initrd /initrd-2.4.21-lvs.img
Reboot and check if the kernel is booting. (if not, you need to choose the old kernel during boot time)
After system boot, use "uname -a" to see if it shows the new kernel is running.
[root@lvs1 ipvs]# uname -a
Linux lvs1.csnet.uccs.edu 2.4.21-lvs #1 Tue Sep 23 23:07:21 MDT 2003 i686 i686 i386 GNU/Linux

1.6. Make the ipvs module

cd /usr/src/ipvs-1.0.10/ipvs
make
- This time the ipvs module will be successfully compiled.
make install

It indicates ip_vs has buffermem_pages not referenced.
I temporarily remove the reference of atomic_read(buffermem_pges) in ip_vs_ctl.c and allows it to compiled.
The following shows the make install result:
[root@lvs1 ipvs]# make install
rmmod ip_vs
rmmod: module ip_vs is not loaded
make: [install] Error 1 (ignored)
insmod ip_vs.o
lsmod
Module Size Used by Not tainted
ip_vs 67824 0 (unused)
parport_pc 19076 1 (autoclean)
lp 8996 0 (autoclean)
parport 37056 1 (autoclean) [parport_pc lp]
autofs 13364 0 (autoclean) (unused)
pcnet32 17920 2
mii 3976 0 [pcnet32]
ide-cd 35648 0 (autoclean)
cdrom 33696 0 (autoclean) [ide-cd]
ext3 70208 2
jbd 51924 2 [ext3]
BusLogic 100796 3
sd_mod 13068 6
scsi_mod 107188 2 [BusLogic sd_mod]
Now ip_vs is inserted to the kernel.
The README file in ipvs-1.0.10 indicates the step of insmod ip_vs_wlc.o
[root@lvs1 ipvs]# insmod ip_vs_wlc.o
[root@lvs1 ipvs]# lsmod
Module Size Used by Not tainted
ip_vs_wlc 1540 0 (unused)
ip_vs 67824 1 [ip_vs_wlc]
parport_pc 19076 1 (autoclean)
lp 8996 0 (autoclean)
parport 37056 1 (autoclean) [parport_pc lp]
autofs 13364 0 (autoclean) (unused)
pcnet32 17920 2
mii 3976 0 [pcnet32]
ide-cd 35648 0 (autoclean)
cdrom 33696 0 (autoclean) [ide-cd]
ext3 70208 2
jbd 51924 2 [ext3]
BusLogic 100796 3
sd_mod 13068 6
scsi_mod 107188 2 [BusLogic sd_mod]

make modules_install

This will put all modules (ip_vs.o ip_vs_wlc.o ...) in /lib/modules/2.4.21-1lvs/kernel/net/ipv4/ipvs/; the default place where insmod looks for modules for the current kernel version.

depmod -a

Note that here we insert weight least connection module. If we specifies -s rr (round robin scheduling algorithm), the virtual serve will not work (no table displayed). We can only specify -s wlc

make -C ipvsadm
- to compile ipvsadm command.
install ipvsadm
- cd ipvsadm (this is a directory underneath ipvs)
- make install (this will copy the ipvsadm to /sbin and install the man page. Check with makefiles in this homework exercise, learn about the how they are organized.)
install ipvs modules.

1.7. LVS-NAT configuration

I have configured a virtual machine on Ardor called lvsnat (in d:\vmware\cs526hw3\lvsnat) running Redhat 9.
- It has two network interfaces. Eth0 is assigned with 128.198.61.61. eth2 is assigned with 192.168.2.61.
- I followed the above procedure to install new 2.4.21 kernel with LVS patches.
- In /root/lvs, lvsnatStart.sh is provided to insert the ip_vs.o and ip_vs_wlc.o modules, configure the lvs nat configuration and weighted least connection discipline using ipvsadm. See the sample script below.
Two virtual machines rsnat1 and rsnat2 were setup with redhat 9.
- Their eth2 interface is assigned with 192.168.2.111 and 192.168.2.112 respectively.

Their eth0 interface is assigned with 128.198.61.111 and 128.198.61.112. When they are rebooted. Their default gateway is set to 128.198.61.1 and allows outside Internet access.
When they are configured as real servers for LVS-NAT cluster, their default gateway will be set to 192.168.2.61 using "route add default gw 192.168.2.61".
In /root/lvs, lvsnatStart.sh is provided to set up real server for lvs cluster operation. reset.sh is provided for resetting the NIC and routing entry as standalone machine.
The default web page is set to have "This is rsnat1 web server." as content and different background color.
These two machines can be reached indirectly via 128.198.61.61.

Note that those real servers after reboot will have eth0 configured as 128.198.61.111 and 128.198.61.112 respectively. During lvsnat test, eth0 interface should be shutdown so that return packet can be routed through lvsnat director using 192.168.2.0/24 subnet. If it is not shutdown, the return packets from real server goes directly back to client through the default 128.198.61.1 gateway or even to the client in local 128.198.61.0/25 subnet. The client will see packets with RIP instead of VIP, and therefore cannot deliver packets to the right web browser application. lvsnatStartup and reset shell scripts are provided.
Note that apache web servers on real servers do not start up properly without eth0 is down. It may try to resolve its domain name through DNS services and was not successful. We need to make sure the apache web server is running on both real server. When their eth0 is up and default gateway is set to 128.198.61.1, you can test them with normal browser on any machine. After you shutdown their eth0 to prepare for lvsnat cluster, you can login to your lvs director and type "lynx 192.168.2.53 or 192.168.2.55" to see if the default web page can be delivered.
On the lvs1, we create a director using the shell scripts in /root/lvs.
The nat.sh configuration script has the following content and is used to create lvsnat director with weighted least connection scheduling.
#!/bin/sh
echo "1" > /proc/sys/net/ipv4/ip_forward

# clear ipvsadm tables

/sbin/ipvsadm -C

#install LVS remove and reinstall ip_vs.o and ip_vs_wlc.o modules

/sbin/rmmod ip_vs
/sbin/rmmod ip_vs_wlc
/sbin/rmmod ip_vs_wrr
/sbin/insmod ip_vs.o
/sbin/insmod ip_vs_wlc.o

#add http to VIP with wlc sheduling

/sbin/ipvsadm -A -t 128.198.61.61:http -s wlc

#first realserver
#forward http to realserver 192.168.2.111 using LVS-NAT (-m), with weight=1

/sbin/ipvsadm -a -t 128.198.61.61:http -r 192.168.2.111:http -m -w 2

#check that realserver is reachable from director
ping -c 1 192.168.2.53

#second realserver
#forward http to realserver 192.168.1.12 using LVS-NAT (-m), with weight=1
/sbin/ipvsadm -a -t 128.198.61.50:http -r 192.168.2.55:http -m -w 1
#checking if realserver is reachable from director
ping -c 1 192.168.2.55

#list ipvsadm table
/sbin/ipvsadm
The ping commands are there to make sure the real server is up.

When ./natwlc.sh is executed in that directory, we got

[root@lvsnat lvs]# ./natwlc.sh

      rmmod: module ip_vs_wlc is not loaded

      rmmod: module ip_vs_wrr is not loaded

      PING 192.168.2.111 (192.168.2.111) 56(84) bytes of data.

      64 bytes from 192.168.2.111: icmp_seq=1 ttl=64 time=3.25 ms
    --- 192.168.2.111 ping statistics ---

      1 packets transmitted, 1 received, 0% packet loss, time 0ms

      rtt min/avg/max/mdev = 3.258/3.258/3.258/0.000 ms

      PING 192.168.2.112 (192.168.2.112) 56(84) bytes of data.

      64 bytes from 192.168.2.112: icmp_seq=1 ttl=64 time=3.02 ms
    
--- 192.168.2.112 ping statistics ---

      1 packets transmitted, 1 received, 0% packet loss, time 0ms

      rtt min/avg/max/mdev = 3.029/3.029/3.029/0.000 ms

      IP Virtual Server version 1.0.10 (size=65536)

      Prot LocalAddress:Port Scheduler Flags

      -> RemoteAddress:Port Forward Weight ActiveConn InActConn

      TCP 128.198.61.61:http wlc

      -> 192.168.2.112:http Masq 1 0 0

      -> 192.168.2.111:http Masq 2 0 0

If you did not see the similar result, you can execute the commands in the shell one at a time and identify the problem.
If you it complains the ip_vs.o or ip_vs_wlc.o module is missing, you need to check if it is installed in /lib/modules/2.4.21-lvs/kernel/net/ipv4/ipvs
[root@lvsnat ipvs]# ls
ip_vs_dh.o ip_vs_lblc.o ip_vs_lc.o ip_vs.o ip_vs_sed.o ip_vs_wlc.o
ip_vs_ftp.o ip_vs_lblcr.o ip_vs_nq.o ip_vs_rr.o ip_vs_sh.o ip_vs_wrr.o
[root@lvsnat ipvs]# pwd
/lib/modules/2.4.21-lvs/kernel/net/ipv4/ipvs
Once lvsnat is set up, you can go to any browser (except that on 192.168.2.0/24 subnet) and enter http://128.198.61.61/ as url.
You will observe that the default web pages of rsnat1 displayed twice followed by that of rsnat2 once in a round robin fashion..
Note that you need to turn off the eth0 interface and set up default gateway to point to your own lvs director 192.168.2.x (x is the last byte value of the IP address assigned to your lvs)

Set up round robin scheduling

rmmod ip-vs-wlc
cd /usr/src/ipvs-1.0.10/ipvs
then insmod ip_vs_rr.o
./natwrr.sh
You will now see the web page rotate between the default web pages of rs1 and rs2.

1.8. Configure LVS-DR.

Hint: The two real servers has eth0 assigned with 128.198.61.112 and 128.198.61.113 respectively.
I shutown eth0 down with "ifconfig eth0 down".. Since it interferes with the LVS-NAT operation.
You can turn it on by "ifconfig eth0 up" and set the default gw to 128.198.61.1.
Use 128.198.61.0/25 subnet address for your LVS-DR configuration.
Create equivalent scripts to setup yor LVS-DR cluster.

A dlvsnat cluster was set up on dilbert with 128.198.61.66 as IP address and use two real servers: rsnat3 and rsnat4.
You can login as root and use our normal password. When you finish your experiment, run the /root/lvs/reset.sh to reset the cluster back to round robin.

I also set up dilhose 128.198.61.22 for your hw3 exercise. You can run lvs using 128.198.61.23 (hose.csnet.uccs.edu), rsnat5 and rsnat6 real server there.

The lvsnat cluster was set up as a reference director. You can only login as cs526.

Linux Virtual Server

Install LVS

1. Upgrade kernel to 2.4.21

1.1. Put kernel source code in /usr/src/linux directory

cd /usr/src/linux

1.2. Backup Existing Kernel Configurations and Reuse a Specific Existing Kernel Configuration

1.3 Apply IP-VS patches

1.4. Kernel Compilation

1.5. Add new compiled kernel to the list managed by GRUB loader

1.6. Make the ipvs module

1.7. LVS-NAT configuration

Set up round robin scheduling

1.8. Configure LVS-DR.

A dlvsnat cluster was set up on dilbert with 128.198.61.66 as IP address and use two real servers: rsnat3 and rsnat4. You can login as root and use our normal password. When you finish your experiment, run the /root/lvs/reset.sh to reset the cluster back to round robin.

I also set up dilhose 128.198.61.22 for your hw3 exercise. You can run lvs using 128.198.61.23 (hose.csnet.uccs.edu), rsnat5 and rsnat6 real server there. The lvsnat cluster was set up as a reference director. You can only login as cs526.

When you finishes the exercises, email me or chat with me to set up demo.

A dlvsnat cluster was set up on dilbert with 128.198.61.66 as IP address and use two real servers: rsnat3 and rsnat4.
You can login as root and use our normal password. When you finish your experiment, run the /root/lvs/reset.sh to reset the cluster back to round robin.

I also set up dilhose 128.198.61.22 for your hw3 exercise. You can run lvs using 128.198.61.23 (hose.csnet.uccs.edu), rsnat5 and rsnat6 real server there.

The lvsnat cluster was set up as a reference director. You can only login as cs526.