S/390 PartnerWorld for Developers, ITSO/EFS Project EFS Systems on a Linux Base: Additional Topics

8 EFS Systems on a Linux Base: Additional Topics

S4WAS1 3390-3 WebSphere

S4WAS2 3390-3 more WebSphere

This provides us with an IPLable system (including our local work data sets) even if the

second disk is not installed.

We created two FLEX-ES definition files and two FLEX-ES startup shell scripts: One for

operation with only the internal hard disk and one for operation with both hard disks.

1.8 Cloning ThinkPad hard disks

Our operational ThinkPad hard disk often had dual boot, with both Microsoft Windows and

Linux available to boot. The Windows partition had a number of products installed and the

Linux partitions had FLEX-ES and z/OS installed. Our most common disk environment used a

60 GB ThinkPad disk with the following partitions:

򐂰 13 GB for Windows

򐂰 3 GB for Linux (root and all files except /boot)

򐂰 80 MB for /boot (part of Linux)

򐂰 512 MB for a Linux swap partition

򐂰 40 GB for /s390 (a Linux file system where we installed z/OS volumes)

We found it convenient to copy the complete disk, including the master boot record (with the

partition table). We placed a new hard disk, exactly the same size as the internal disk, in the

UltraBay slot. We then booted Linux, logged in as root, and used the command:

# dd if=/dev/hda of=/dev/hdc bs=1M

In this command, the internal ThinkPad hard disk is our input device (/dev/hda), the hard disk

in the Ultrabay slot is our output device (/dev/hdc), and we copy using 1 megabyte blocks. The

complete 60 GB disk copy took an hour. We could then move the new drive (from the

Ultrabay) to the internal disk position and boot from it. This was the easiest way we found to

obtain a complete backup of our system.

Using this technique, the source disk is a running Linux system (which is performing the dd

command.) The target system will have a “point in time” copy of a running system. When it is

booted, it will appear that it is a reboot of a system that crashed. On the first boot, it will use

fsck to recover disk metadata. If you then shut down Linux properly, subsequent boots will be

clean.

1.9 Multiple FLEX-ES instances

You can define multiple S/390s by defining multiple FLEX-ES system definitions. You can run

multiple S/390 systems, one at a time, by simply creating multiple shell scripts that name the

appropriate syscf file in the flexes command. You can run multiple S/390 systems at the

same time (multiple instances) if you have enough server memory and have defined your

FLEX-ES resources appropriately.

Only one resource definition (rescf file) can be active on a server, so it must define all the

resources needed by all the S/390 instances. Each S/390 instance must be started with a

unique system definition (syscf file). You would normally start a separate CLI window, with its

flexes prompt, for each instance. Some care is needed to use the right flexes windows when

controlling the S/390 instances. The CLI command set prompt can be used to change the

flexes prompt to something more meaningful for each instance.