migrating databases from Solaris to Linux

Management has decided that Linux is the OS of the future, so I’m preparing to migrate our Tamino databases from Solaris/Sparc to Linux/x86. Has anyone done a migration like this recently? If so, and tips or advice?

I guess I’ve been spoiled by how easy it’s been over the past few years to move from old Solaris machines to new ones: just create the database from a backup on the new machine, make it a replication database, sync up with the old database, stop the old one, and make the new one the master. But you can’t create a database from a backup on another platform, so this is looking like a lot more work.

(I actually migrated Tamino to Solaris from HP-UX over a decade ago—the last time management decided on “the OS of the future”—but the databases were a lot smaller then, and I’ve forgotten just how I did it.)

Solaris/Sparc and Linux/x86 has different architecture (low/high order byte first).
For this type of migration all doctypes must be unloaded and reloaded into the new database.

Yes, that was my understanding. I was hoping someone had come up with a better way. (Like maybe a utility to convert a Solaris backup to Linux format?) I’ve had some unpleasant experiences with unload/reload, and it’s much slower than using a backup.

(About unload/reload experiences: early versions of Tamino—we’ve been using it since version 2—were not always good about enforcing correct namespace usage. We may still have some documents that are not valid XML but were accepted by earlier versions. When you unload them and reload them, the Tamino loader rejects them.)