RepliStor is a file mirroring and replication package for Windows XP/2000 and Windows Server 2003. It integrates with Active Directory if you’re using it, but it’s equally happy on stand-alone systems. Alongside the core product installer on the CD is a pair of add-ons that deal with the two most common culprits for open files – Exchange and SQL Server – although rather oddly, the latter runs only in Windows Server 2003.
Once you’ve installed the application (which is a simple case of pointing it at its install directory and giving it the licence key), the package is oriented around the concept of “sites,” which basically means servers. When you fire up the package for the first time, it identifies itself and drops an icon in the site list (the icon is a traffic-light symbol, with red/yellow/green lights summarising the state of each entity).
To connect to another site, you simply right-click, select “Add Site”, then give it the address and login credentials of the other machine(s) with which it’s going to be synchronising its data.
To actually perform file synchronisations, you add what the package calls “specifications.” A specification has four main aspects: the source directory you want to synchronise; the destination directory for the replica to go; a set of options (such how to deal with deleted files, whether to exclude any particular files or subdirectories) and some control information for the synchronisation process itself. As well as having files copied you can, in fact, also replicate file shares (though the filesystem structures on the source and target systems need to be identical) and registry keys from one machine to another.
Synchronisation takes two forms. The simpler approach is to make synchronisation happen on a schedule, or even to kick it off by hand whenever you wish. The more advanced way is to turn on “mirroring” which, instead of relying on periodic synchronisation steps, watches the I/O buffers in the system and reacts when it sees changes to files that are part of a synchronisation specification.
RepliStor is really very straightforward to use. The list of sites (servers) is in a pane on the left, and the detail for whichever item you “attach to” (i.e. double-click in order to select) is shown on the right. There’s a set of traffic-light icons across the bottom of the screen showing an overview of the statuses of the site you’re attached to, broken down by category (so there’s one for disk space, one for log messages, and so on). If they’re all green, all is well; if one turns orange or red it means that there’s something amiss. You can tweak the specifics of individual specifications at will, and there’s also a vast wad of options relating to the overall settings of each site – disk space thresholds, email settings for alerts, logging options and so on.
The documentation and electronic help are a little disappointing in places (there’s at least one feature of the GUI that we came across which doesn’t even get a mention in any of the docs). This said, though, it does go to reasonable lengths to warn you about the perils of trying to replicate things you shouldn’t (machine-specific or system files) and/or replicating to places you shouldn’t (overwriting the Windows directory is probably a bad move, for example), and anyway, the usability of the package is such that you rarely need to look in the book anyway.
We were initially a bit underwhelmed, and thought the GUI was a bit clunky, but after an hour or so we’d got used to the GUI and the slightly weird terminology (why not call a server a “server” rather than a site?) it and the product grew on us, particularly when we got mirroring running happily and files we dropped onto machine A magically appeared on machine B.
RepliStor is an interesting product, then, and at £440 per server, it’s a reasonably priced way to achieve near-realtime file, share and registry replication between systems.
A file mirroring and replication product with a few idiosyncracies but a reasonable price and quite easy-to-use GUI.