With infrastructure going increasingly virtualized, it is routine to live-move entire servers from chassis to chassis (or datacenter to datacenter) with zero downtime. Similarly, we can live-move VM storage from one disk volume to another with zero downtime.
However, with SmarterMail, when a domain's storage grows to unexpected levels and we need to move it to another volume, we have to take the entire
SmarterMail server offline, affecting all
customers for the entire duration of the move-- which can take substantial time depending on size of the domain's storage.
Example: Customer started small but business prospered and domain storage exploded from around 10 GB to nearly 2 TB, causing space constraints on the server's on-board storage. Need to move this domain to an iSCSI drive with huge storage available. Per KB https://portal.smartertools.com/kb/a2727/move-a-domain-to-a-different-hard-drive-on-the-same-server.aspx
-- we have to:
1.) Stop the SmarterMail service, which affects ALL customers.
2.) Copy these 2 TB of data across the network to an iSCSI chassis, which in real-world terms even on a 10-GigE network, means downtime exceeding 30 minutes... affecting ALL customers on that server.
3.) Manually edit the .json files in a text editor for the domain, to point to the new storage location.
4.) And only then, can we bring the service back online.
Ideally, moving SM domains to other storage would be a live-move, just as with most other virtualized services/servers. But even if we can't get that far "today", at least give us the ability to specify the new storage location in the web GUI. Then the SM service can drop just that domain
offline, move its storage, then bring it back online-- without disturbing any other customers on that server.