Hey all!, First post so hello to the community! I have a couple of questions on the hub transport role. First relates to backup. From what I can gather (from documentation etc) is that backup of the hub servers is not required due to the transient nature of the queue database. Fair enough. However, I'm trying to measure the impact should the queue database become corrupt and cannot be fixed with ESEUTIL. If some messages are in the queue when I loose the database on the hub, are those messages lost? If the queue is large this could be 1000's of messages. While I accept that I could ask users to resend from their sent items, this seems impractical for a large amount of messages. So I guess my question is how do I protect myself from message loss while messages are being queued? My second question is relating to hub transport volumes and sizes. How should I distribute the database logs etc? I assume seperate disks for logs and databases? I'm also not sure how I should tackle sizing? I know the message queue uses circular logging but how do I control the database size and the number of logs it uses before it starts it's circular process? What are practical sizes for partitions? The Exchange system will be for around 10,000 users sending around 100,000 messages per day. Any help will be gratefully received. Regards

WHWMan wrote: If some messages are in the queue when I loose the database on the hub, are those messages lost? ... So I guess my question is how do I protect myself from message loss while messages are being queued? YES - you can backup the hub transport database just like you would the mailbox servers. Using an Exchange aware backup application to backup the Information Stores. However to be effective, you'll need to backup every 15 minutes or something like that, due to, like you said, "the transient nature". WHWMan wrote: My second question is relating to hub transport volumes and sizes. How should I distribute the database logs etc? Leave the transaction logs on the OS drive. they are set to circular by default. Move the que.edb to another drive by using -TransportDatabase.ps1. WHWMan wrote: I'm also not sure how I should tackle sizing? This is just something you'll have to watch. It totally depends on the amount of mail your company sends, and how many servers you have. I realize you say 10k, but you should just monitor it over a weeks time - especially since you didn't say how big those messages are, and if they are sent evenly across 24 hours or in spurts. System Center Operations Manager can help automate this. If I had to guess, you'd be fine with a pair of standard mirrored 146GB drives. and always use as fast as possible, which means scsi 15k in today's tech. By the way, you have at least two hub transport servers, right? You should.

Apologies Mike, I totally forgot to thank you for your response. It was very helpful. And yes, 3 transport hubs.