Exchange - where email goes to die - The Something Awful Forums

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > Exchange - where email goes to die

Studebaker Hawk: May 22, 2004

Anyone here running a multi-site DAG scenario and has tested/experience HA? All my EXCH2010 deployments have been single server and am contemplating a mutiple dag scenario (CAS/MB onsite, onsite MB (archival)/CAS, offsite cas/mb all having HT roles).

# ¿ Jun 24, 2011 04:26

Adbot: ADBOT LOVES YOU

# ¿ Apr 26, 2024 21:26

Studebaker Hawk: May 22, 2004

citywok posted:

This is because the DAG role requires Failover Clustering, and the CAS role requires Network Load Balancing, and you can't have both roles on the same server.

The ideal solution is two enterprise servers running DAG (enterprise exchange and enterprise server), and two additional standard edition servers running CAS. This is the model we're using. On top of this, at another location we store another DAG copy, and a CAS for DR.

Don't forget you need server 2008 r2 enterprise to get the failover clustering role, and you need exchange enterprise to go beyond something like 5 mailbox databases. That said I'm not sure if you need enterprise to get the DAG feature itself.

You don't- standard will get you dags as long as your OS is enterprise, but there is a db limit.

# ¿ Aug 18, 2011 19:32

Studebaker Hawk: May 22, 2004

FYI even though this states that Online Archives are supported with Office 2007 it neglects to mention that only certain versions of office 2007 ACTUALLY support the feature.

Just added 20K to an already sold/halfway completed project

# ¿ Aug 23, 2011 16:25

Studebaker Hawk: May 22, 2004

Corvettefisher posted:

all windows 7 <75 users, good to know any other opinions?

Assuming it has to be hosted for budget reasons, look at intermedia or any other large hosted exchange provider? Intermedia is massive and while I rarely experienced issues with users they would do poo poo like change their IP range and not update their resellers which would then end up in mail being blocked if you use a 3rd party spam product.

Agree w/ linux nazi, O365 sucks.

# ¿ Sep 9, 2012 15:04

Studebaker Hawk: May 22, 2004

Corvettefisher posted:

So a mid-large client(600 users) that we will be developing a solution for decided that instead of an onsite and offsite colo site for exchange they want office 365. My question is really, is there any way to use that in conjunction with an onsite exchange? Such as mail can be sent or cached to an exchange server from 365, and in the event of a 365 outage the local or colo exchange would respond to requests?

Yes/No. Not in the traditional HA/DAG model but you can setup hybrid deployments for 0365. I would go through this first http://technet.microsoft.com/en-us/exdeploy2010/default.aspx#Index

AFAIK this only gives you the ability to have co-existance between the two environments, a given mailbox will ultimately still reside either locally or in the cloud.

I would probably suggest colo+dag if you are looking for HA.

# ¿ Oct 10, 2012 17:33

Studebaker Hawk: May 22, 2004

platypusmalone posted:

I'm just finishing up a 2003-2010 upgrade/migration. Everything seems pretty perfect and went very smoothly. The last few mailboxes are finishing the move to the new server now.

One thing I am a bit confused about is replication of my public folders... on the 2003 server they show as "Local Modified" in replication status. On 2010 I can't find where I would see if they've properly been replicated. Clients seem to be able to see them though and they aren't high volume anyway so I am assuming this might just be a reporting issue.

I plan on leaving the old server up for a couple weeks but want to get it out of the loops sooner than later - it is ancient and a bottleneck as it is still in a DC role. Is there anything else I should be concerned about provided the mailboxes and public folders are all on the new server and mail is flowing properly?

get-publicfolderstatistics -server name

# ¿ Oct 16, 2012 13:43

Studebaker Hawk: May 22, 2004

Gyshall posted:

What are people using now that Postini is going to be moving into Google Apps? My company wants to push MX Logic but I'd rather not have to support a McAfee product.

We're looking at Forefront Security for Exchange, but our three needs are:

- Mail spooling (ie. firewall or server goes down locally mail is held until they come back up)
- Spam/antivirus
- Ability to access and send mail while servers are down/mail is being spooled.

FWIW McAfee's hosted spam is actually a great product, and well supported.

# ¿ Oct 17, 2012 14:44

Studebaker Hawk: May 22, 2004

Nitr0 posted:

aaaaaahahahaha

Guess you missed the great mxlogic outage of Aug 8 2012? Mcafee is poo poo and anyone that uses them deserves what they get.

Yep, was on the other side of the world at that time

Put a call into mimecast and liked what I see, but they are a bit pricing and there partner margins aren't great

# ¿ Oct 18, 2012 23:45

Studebaker Hawk: May 22, 2004

Drighton posted:

Once again I'm out of ideas and need help. We have four 2010 SP2 RU4 servers in two locations All the hardware and configurations were tested at the same location, including the routers, and the databases remained at 0/0 for both copy and replay queues. But now that the none of the databases in our DR site can keep up with log generation at the main site, except at night. Thing is, when we've had failovers and the logs are flowing the other direction, everything is fine and working normally. This last part proves to me the configuration on all four servers is good.

Now add to all this the information I've read online, and things appear to be pointing to bandwidth at our DR site. Looking at the utilization charts at that site doesn't seem to support this, however. I'm not quite ready to give up on bandwidth as the culprit. So I'd like to know if anyone has any other suggestions? Are we perhaps dismissing the hardware and configuration prematurely?

Without any real details, this sounds like a bandwidth issue. What type of connectivity/latency do you have between sites? Have you enabled compression on your DAG traffic (set-databaseavailabilitygroup networkcompression)?

# ¿ Nov 6, 2012 15:21

Adbot: ADBOT LOVES YOU

# ¿ Apr 26, 2024 21:26

Studebaker Hawk: May 22, 2004

TjyvTompa posted:

I only came in here because I was interested as to why my company recently did a really bad Exchange upgrade for a customer.

From some posts here I get the idea that all their (our) problems was because this customer ran Exchange 2003 which we upgraded to Exchange 2013. I guess this explains why the new email system is in a whole new domain and users must now login to outlook using a separate username and password then to their computers.

Needless to say, the users are experiencing many weird problems. Some can't access their calendar through Outlook, but it works in OWA, and in Outlook if you right-click the calendar and select "Open in new window" then it works.
Some users receive the login window every ~10 seconds even though they enter their password and it is accepted (users can send/receive with this login prompt active).
Some users can't receive external emails.
Some resources (Meeting rooms et.c.) couldn't be migrated at all.
All users who had previously changed their name received their old name in the new email system (I guess the new mail system took the first added smtp adress from the AD user as the username+mail adress instead of the primary smtp adress on the ad account).

Add to this a horrible web-based administration interface made by these clowns: http://www.ceondemand.com and you have a receipe for sleepless nights.

I...just...WHY? I am sorry, but your company got hosed. I can't imagine that a new domain without proper trust/authentication was part of the scope and approved plan.

edit: oh, your company PERFORMED the upgrade. Yikes.

# ¿ Dec 10, 2012 15:09

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > Exchange - where email goes to die