Link to home
Start Free TrialLog in
Avatar of izgoblin
izgoblin

asked on

Going to production with MS Exchange 2016 without a DAG - how dumb is it?

We are being pushed to migrate from a single Groupwise server (go ahead and laugh) to Exchange 2016 ASAP.    Trouble is, time is of the essence and additional hardware was not budgeted for.   Silly me thought that sizing the Exchange server would be similar to the Groupwise server, but I've since found out that this is far from being the case (I've seen the Exchange sizing spreadsheet and nearly cried afterwards).    My question is, if I have to go into production without a DAG for the remainder of the year, is this surely to be a mistake, or can I get by for a while?  

We are a 100-person shop without about 600 Gb of email on Groupwise which uses single instance storage.  I now know that Exchange no longer does this (though I've not heard the logic behind that glorious decision), which means I'm already expecting to chew up a ton more storage than I hoped and darn near all that I have left to spare on my SAN.    We virtualize with VMWare and are fully redundant hardware-wise.   Yes, our email and applications are expected to be up 24x7, but we don't go as far as to do SQL replication and for the most part have single virtual servers for each application rather than banks of 'em.   This usually works out just fine for us.

For you Exchange 2016 administrators:  How often do you encounter data corruption?   We've maintained SQL servers here for years and I've never had to perform a restore due to corruption.   And never once have I had to restore from backups with Groupwise - If it encounters data errors, I run maintenance on it and all is well.  Should I expect more likelihood of corruption from Exchange that will require me maintaining multiple database copies?

I am not worried about restoring a single message if a user deletes it.  I have a Retain archive server they can use for that.   I am not worried about an admin accidentally deleting a mailbox (I haven't done that in the 15 years I've worked here).    I am not worried about the server being down during scheduled maintenance for updates - that's what scheduled maintenance is for, and we have a non-MS perimeter server that will queue any mail received while Exchange is down.    What I am worried about is the stuff that I can't control like the application itself crashing and causing corruption.  Outside of scheduled maintenance hours, I do need the server to be stable and available.   What can you tell me about your experiences there?

I went through an online training session where the instructor said he couldn't imagine any situation where you would go into production with a single mailbox server, but that's precisely what I thought of doing until we can get a ton more disk space to create a DAG and a second mailbox server, which would be in early 2018.

Is my plan perfectly acceptable, or am I apt to regret this really quickly?
ASKER CERTIFIED SOLUTION
Avatar of Adam Brown
Adam Brown
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Ask your management these two things:

1) Recovery Time Objective: Time limit on recovery from a failure. What is your company looking?

2) Recovery Point Objective: Amount of data that can be afforded to be lost due to the failure. What is your company can afford?

Then you can further work on designing part and implementation will come at very end. Let me know, if you get the answer, I can tell you next question or step.
Avatar of arthurjb
arthurjb

Sounds like a RGE to me (Resume Generating Event).  If you let management push you into improperly doing this implementation, then you can cont on the fact that you will be blamed when it fails.

Armit's questions are good.  I would add, "Do you care about the data?' and "How long can the business function with the services down, before there is financial implications?"  

You need to get the answers to Armit's and these questions in writing.  Your email with these questions must include the statement that this course of action is dangerous to the company's production data, and that you believe that it should not occur until the extra hardware can be budgeted.  

This is a typical management move, because they fail to understand the severity of the business impact that can happen.

Print out the email before you start this insane move, so you can show them that they were the ones who initiated the move without proper data processing procedures.