HA/High-Volume Deplyoments (planned on Linux)
Question asked by TK - 7/31/2024 at 12:52 AM
Unanswered
Anybody (including Smartermail) able to share how they have implemented (propose) a High Availability/High Volume setup for their own installations and also how that has worked out for them in terms of complexity or any challenges that have come along?

  • Have you offloaded domain data to shared storage?
  • Have you offloaded smartermail server settings to shared storage?
  • Have you implemented failover at all or are you relying on any other mechanisms?
    • Besides: documentation states failover is Windows only - any statements on this? Target is Linux.
  • Are you even spinning up servers/containers per customer?
  • Is it possible to distribute mailboxes of the same domain onto multiple smartermail instances and having coordination nodes (proxies) to keep track of the user<->node assignments?
  • Any other means of load distribution?
  • Any question missed here?
This discussion is not about spam processing which we have offloaded to front end gateway. it is solely a discussion around the mailbox servers. The documentation on failover suggests shared storage, buth the High Volume documentation is not talking about this at all. 

Matt Petty Replied
Employee Post
"Is it possible to distribute mailboxes of the same domain onto multiple smartermail instances and having coordination nodes (proxies) to keep track of the user<->node assignments?"

Regarding this stuff, Zach mentioned here some things were looking at in the background that revolve around very large user deployments across servers.
https://portal.smartertools.com/community/a96358/need-a-feature-that-force-ip-rotation-at-random-and-or-ever-x-times.aspx
Matt Petty Senior Software Developer SmarterTools Inc. www.smartertools.com

Reply to Thread

Enter the verification text