Robot HA Helps Amphenol Migrate to POWER9 and Implement a Reliable Backup

Global manufacturer of connectors updates its infrastructure to protect business continuity

Image
Text

Amphenol designs and manufactures ruggedized, harsh environment connectors for the military and aerospace industries. You’ll find their connectors on every major platform in those industries, from commercial aircraft and military ground vehicles to soldier systems and satellites. They are part of the largest military and aerospace interconnect group in the world, with its largest division, Amphenol Aerospace, based in Sidney, New York.

Making Bold Moves for Business Continuity

Every so often, the technology stars align for IT professionals. For Bob Butcher, an iSeries Engineer at Amphenol, this happened when he learned IBM was ceasing support for his long-trusted POWER7 production server. At that time, Butcher was also weighing options for how to improve the company’s business continuity/disaster response infrastructure following two instances of hurricanes and flooding at their New York data center.

Recovery time and downtime were essential considerations for the team. “We looked at the cost of an hour of downtime,” Butcher said. “With the previous scenario, it would take the IBM i box two full days to get back up and running from tape backups.” This would be disastrous for the global user community that relied on 24/7 uptime and information access. The key to the project was in modernizing the back end with new hardware and high availability.

Text

A Three-Pronged Solution

Amphenol decided to build a new data center in Mesa, Arizona, and move to POWER9 servers. Butcher would establish a new production server in the Sidney, New York office with full replication to a second server in Arizona. Achieving this meant migrating data from the old POWER7 server to the new POWER9 server, performing a role swap, purchasing a CBU box from IBM, replicating side-by-side for a period of time, shipping the CBU box to the DR site in Mesa, and then establishing the Mesa server as the long-term failover option through daily replication.

Fortunately, the Robot HA high availability solution for IBM i from Fortra was up to the task. Having used Fortra solutions at a previous organization, Butcher knew his options. “I’m a big Fortra fan, and I don’t just say that,” he said. “If a new product is going to be touching IBM i somehow, I always start with Fortra.” Robot HA was also by far the most cost-effective option compared to the other solutions evaluated. “We couldn’t beat the deal Fortra had given us,” Butcher said.

Butcher decided to use Robot HA to replicate from the POWER7 box to the POWER9 box. “Rather than taking a 12-hour window to do a go/save/restore for 5 terabytes of data, I actually did a role swap to my POWER9 box. It worked awesome.” It happened so effectively that Butcher went from thinking the project would take 12 to 15 hours to completing it in 10 minutes.

After the data was moved over, Butcher worked with Fortra to determine how to approach the replication piece. “We didn’t want to be replicating all of the IBM libraries and IFS, so we sat down and identified our user libraries and what we needed to replicate on a daily basis,” he said. They decided 150 user libraries, 20 major IFS folders and directories, and 15 million documents would ultimately require daily replication. “I unplugged IBM i here on a Friday, flew to Arizona Monday morning, and met with IBM,” he said. “We plugged it in a new rack, turned on replication, and caught everything up.”

During this process, Butcher had a revelation about journaling. “In all the years I've been an administrator, I didn’t really understand journaling or remote journaling,” he said. Luckily, he received a presentation on the topic from Fortra and realized he could set everything up to be automatic. “We go through some pretty extensive audits, and I had my users go over and test things. They said the data looked exactly the same. So, I felt very, very comfortable.”

“The thing that impressed me the most about Robot HA was testing the role swap,” Butcher said. He knew the data was there but was concerned about how the configuration settings and entries for relational database would work. “All the ethernet lines, TCP/IP routes, and servers came over, and all of the ODBC traffic was pointing in the right direction. I was just amazed.”

Peace of Mind with a Future-Forward Solution

With the project complete, Butcher is happy with the outcome. “Users are logging onto the system, and we’re not missing a beat,” he said. “Robot HA is replicating all my data, making sure that all my objects are in sync.”

Robot HA has also opened up new possibilities for the team, such as the creation of a sandbox. This enables programmers to go on the DR box and test whatever they need to evaluate. “I don’t have to worry about them stepping on live data as Robot HA can replicate to different library names for testing,” Butcher said.

He also uses Robot HA’s email functionality to receive logs several times a day. “I open up the logs, and it’ll tell you what libraries or IFS directories you need to take a peek at,” he said. “It’ll show you which things are lagging, or if something didn’t run or didn’t update in the timing specified. With everything being in Arizona, I really look at those logs to make sure everything is okay.”

Throughout the project, Butcher has been especially impressed with Fortra customer support. “Tech support is the biggest thing,” he said. “I don’t know if there’s ever been a time when I’ve called and haven’t gotten an answer. That’s important, especially when it’s in the middle of the night.”

Protect Uptime for Global Users

Safeguard daily business operations with Robot HA high availability software for IBM i.

DEMO ROBOT HA