iTera Goes 'High Availability Lite' With GuardianSave

by Alex Woodie

iTera furthered its strategy of developing cheaper, simpler high availability software for iSeries servers last week, when it announced the general availability of GuardianSave, a new disaster recovery tool also intended to reduce downtime associated with performing data backups and full-system saves. iTera designed GuardianSave to protect users from losing transactional data between regular backups by using OS/400's journaling feature to send that transactional data offsite, to a second OS/400 server or even a garden variety PC running Linux.

iTera cofounder and president Dan Neville says there's a large gap between full-blown iSeries clustering and mirroring applications, such as iTera's own Echo2 product, and the traditional tools that OS/400 shops use to back up their data on a daily, weekly, or monthly basis. "We are convinced that GuardianSave provides a 'missing link' in the spectrum of backup and recovery solutions on the market," he says. "In some ways, GuardianSave provides 'high availability lite' for iSeries shops that cannot cost-justify a true high availability solution."

iTera thinks GuardianSave improves backup and disaster recovery in two main areas, and the company uses different technologies and approaches to accomplish each of them. First, GuardianSave reduces or eliminates the user downtime normally required to get full locks on objects in order to back them up to tape or disk and create a known recovery point, or a synchronized checkpoint. GuardianSave uses the Save While Active functionality, which IBM includes in OS/400, and which iTera uses in other products, such as Reorganize While Active and Convert While Active.

iTera says it has a proprietary method of enhancing and speeding OS/400's basic Save While Active functionality. Although Save While Active requires all users to be logged off for about 15 minutes to create a synchronized checkpoint, it claims GuardianSave can accomplish the same task in just a minute or two. iTera says GuardianSave can run full system saves that might normally take up to six hours in only half an hour or less.

Besides cutting down on user downtime by using a variation on Save While Active, iTera thinks GuardianSave makes improvements in the way disaster recovery products move data. Like all OS/400 high availability applications, GuardianSave uses the journaling function that IBM has built into OS/400 to capture changes made to application data. However, this is where GuardianSave differs from traditional high availability.

Instead of replicating this journal data to another OS/400 server, via IBM's remote journaling functionality or a proprietary journal-scrape method, GuardianSave uses good old FTP to send the journaled data--otherwise known as the journal receiver--to an offsite location. This data is important because it was not included in the last save or synchronized checkpoint, which can be recovered from tape. When the production machine goes down, the journal receiver data is FTPed back to the production OS/400 server, and the user loses less data than if he recovered from the last save or synchronized checkpoint. Ideally, the user loses no data at all.

Using FTP as a transport method has advantages. You don't need another expensive AS/400 or iSeries server to house the journal receiver, and any computer capable of running an FTP server will work as GuardianSave's offsite companion. However, iTera favors running the Linux operating system over Windows because of its higher security and stability. Users can even create RAIDs of their journal receiver data if they want.

"The good thing about this is, at least you're getting the day's transactions off the box," Neville says. "Whether it's a small PC or a small '400, it doesn't matter to us." By offloading transactional data, Neville says, companies will also be able to perform some data analysis through tools such as iTera's IT Director.

GuardianSave also recovers a fallen production OS/400 server. Many of the apply processes that iTera uses in its Echo2 high availability and clustering software are included in GuardianSave. After recovering from tape, the menu-driven application guides the user through restoring the data objects, or a single object, from the journal receivers, to provide a complete recovery, right up to the point of failure.

If it's not important for a company to have synchronized checkpoints, GuardianSave can perform backups without requiring any downtime from users, not even for a couple of minutes. GuardianSave's backup capabilities provide other safeguards against data loss, and companies may prefer to run this way, while others will want the insurance that comes with creating a known point-of-failure where the integrity of data objects is ensured, says Jeff Ashman, iTera's director of research and development.

Ashman says that although benchmarks have shown OS/400's journaling function decreases an application's performance by about 2 to 3 percent, GuardianSave runs very fast and doesn't require much in the way of system resources.

For most AS/400 or iSeries operators, minimal training is required to use GuardianSave, Ashman says. The menu-driven, green-screen interface guides users through the configuration, and users don't have to know a thing about setting up journaling or journal receivers. The setup will take a day or two, but operators won't have to monkey around with the remote FTP server that houses the journal receiver, because they can control it from the OS/400 console.

GuardianSave functions can be automated using standard job scheduling software available on the market, Ashman says, and the product integrates with tape management products, like IBM's BRMS and others on the market. One company that beta-tested GuardianSave successfully integrated the product with its existing backup and recovery software, Ashman says.

For future releases, iTera will add additional tape management functions to GuardianSave, the company says. The company is also working to add remote journaling capabilities to augment the current version's reliance on the FTP data transport method. This will provide even higher levels of data integrity and system reliance.

GuardianSave 2.1 is available now. iTera says it is including a Linux PC with 200 GB to 300 GB hard drive space with the cost of the product. License fees for GuardianSave are tier-based and range from $5,000 at the P05 level to $25,000 at the P50 level. For more information, go to www.iterainc.com.

Sponsored By BOSANOVA

Disaster Recovery Assistance for Twinax Users The Best Way to Keep Twinax Users Working is with a Pure TCP/IP Controller In most disaster recovery scenarios these days TCP/IP is the protocol of choice for switching between machines. Most solutions to switch users from the primary AS/400 to the back-up AS/400 require that the IP address under which the AS/400 appears be changed. For example, say that the production AS/400 has IP address 10.1.1.10, while the back-up AS/400 normally uses 10.1.1.20. In the event of a disaster, the back-up AS/400 switches to 10.1.1.10. All the PC users using TN5250 do not have to change their settings; they simply restart their sessions and sign back on. What Happens to Twinax Users? In the same scenario, Twinax users suffer a harsher plight. During disaster recovery the local Twinax users can be left without any access to the back-up AS/400. Remote Twinax users usually can be switched over using some scripts on the AS/400, however is often a challenge. Is There a Better Way? IP-based Twinax controllers like the e-Twin@x Controller can easily switch between AS/400's because they treat Twinax users the same way they treat the PC users, once the session is reset. The e-Twin@x Controller, as a matter of fact, does this automatically; the users can sign on to the back-up AS/400 and continue to work, without any operator intervention! What About Locally Attached Twinax Users? Even locally attached Twinax users can work in this fashion, using the 47xx local e-Twin@x Controller series. These very aggressively priced controllers are designed to work in a local Twinax environment. All the local Twinax cables connect into the 47xx Controller in place of the local Twinax controller resident in the AS/400. The 47xx connects to the AS/400 over the built-in 10/100 MHz Ethernet card, ensuring fast response times. The 47xx Controller will switch over to the back-up AS/400, provided it is on the same IP subnet as the production machine. Special Features for Switch-Over The e-Twin@x Controller includes the following special features, which are especially relevant for disaster recovery situations: DNS Look Up The e-Twin@x Controller has the ability to configure the AS/400 via a DNS name instead of IP address. A timer in the e-Twin@x Controller refreshes the DNS look up every time it expires. This way, the System Administrator changes the internal DNS server, which lists the AS/400 domain name, and after the timer expires, all the devices are automatically changed to the new IP address. Auto Reconnect If the connection to the AS/400 is lost, the e-Twin@x Controller tries to re-establish the connection, without user intervention, after a predetermined time-out period. Default Rollover Two AS/400's can be configured to act as default and back-up, respectively. When the primary (default) AS/400 can no longer be accessed, the e-Twin@x Controller automatically attempts to connect to the secondary (back-up) AS/400. After successful connection to the back-up AS/400, it will stay connected to that AS/400 until the e-Twin@x Controller is reset, which can be done by using a Web browser. By the way, you can define that the process be reversed if the first host is also defined as a backup system (for the secondary host). In this case, if the backup system fails (or if the TCP/IP interface is terminated), all sessions will be re-connected to the primary system. www.e-twinaxcontroller.com BOSaNOVA, Inc. - BOScom Master Distributor