Migrating a TSM server environment from z/VM to x86_64 platform

First order was to create a as-is analysis of the up and running environment based on z/VM. These TSM server were running fine, but there were some technical limitations and another reason to go away from z/VM was the high license fees for z/VM processor cards.

After presenting the as-is analysis we’ve got the order to build up a PoC-setup, based on HP ProLiant DL380 G8 servers. The PoC was working on two machines with RedHat Enterprise Server 6.4 and the HP ServiceGuard cluster software. At that time TSM 6.3.4 was the current choice.

The virtual tape appliances (Fujitsu Eternus CentricStor) in the backend was continued to use. Just the way of communication and the type of emulation had to change to reach a supported setup.

A couple of weeks later, we’ve had to enhance the PoC-setup with six additional DL380 servers. The resulting 8-node-cluster was the new hosting basis for all 19 TSM server instances. We did a lot of scripting to integrate the new servers to the given monitoring and ticket solutions before we went productive.

All TSM clients needed an update and new configuration entries. Everybody of our team (internals and externals) did an amazing job.

The project was concluded by delivering an Administrator’s handbook for the new environment and we also did some workshops to ensure a proper hand-over to the customer.

Dividing TSM backup and archive data from two companies

The department of a company became its own company and because of legal requirements there was a fixed date for completion. All network lines were divided by exact that fixed date.

Most of the applications and data were located on their own server boxes – so the dividing itself was an easy job to do. The regarding servers were identified and removed from the one data center and integrated in the new data center. But nobody thought about the single TSM server and its data repository including backups and long-term archives from both “companies”.

After setting up a new TSM server on new hardware, we had to wait for the new tape library and hundreds of tape cartridges. When the delivery took place, we were ready to go and our migration concept was finished at that time.

The migration was done via TSM Server-2-Server communication over LAN. We’ve got a dedicated link to speed up the process as much as possible, but we could finish the migration tasks before the network was finally divided.

Last steps were to create Admin’s handbook and to run the environment for a while, because the new company had no staff for running TSM at first, so we took over and ran this environment for nearly one year.

Introducing TSM to a 3rd data-center

A customer with two data center was enhancing high availability of their business critical applications by integrating additional servers located at a third data center.

This project included a partial project covering the integration of a TSM server infrastructure, which also had to be integrated in to the existing TSM environment. We were responsible for this partial project.

All TSM server are based on IBM pSeries 750 with AIX 7.1 and the existing tape infrastructure at the two existing data center are also enterprise-class IBM hardware: IBM TS3500 with 48 x IBM TS1140 tape drives

This setup was enhanced with the same hardware platform and an identical tape infrastructure at the third location. The current concept regarding the backup data flows was altered to integrated the third data center setup in to the overall backup concept for business critical identified server boxes and applications.

Before the overall project was finished, a final test was performed successfully while the connections to the original two locations were cut. During that test we performed several test cases to prove the ability to restore data plus the ability to backup new data at the third location.

We also integrated a archiving concept for keeping long-term data in the TSM repositories of all three locations.