When doing a design for Exchange 2010 it is takes a lot of effort to look at all aspects of a system before choosing a product. Choosing the right size load balancer for exchange entails looking at the systems as a whole and how it will be used. A few of the things I look at before making a decision are
1. Client Count
2. Protocols used
3. Client connection speeds
4. Encryption Used
5. Device type
6. Message size
These numbers will tell a person a lot about Exchange and how it is being used so that one can make an educated decision about the size and capabilities of a load balancer. Unfortunately time constraints or a number of other factors do not always allow one the time to analyze and Exchange environment. In other cases a general idea is need for general discussion or budgetary estimates are needed prior to a deep analysis is done. Luckily Kemp Technologies has come up with a handy tool that can be used to get a broad look at what products they offer and a general idea of how to size them, given certain parameters. Take a look at the tool and give it a whirl.
http://www.kemptechnologies.com/fileadmin/templates/sizingDoc/lme_calc_2k/lme_calc_2k.htm
Category Archives: Technology
BlackBerry Delivery Delays and Exchange 2010: UPDATE
Recently I reposted an article written by a
fellow MCM about BlackBerry and Exchange 2010 Delivery Delays. There has been a
lot of work on this issue and I thought it wise to post an update. Microsoft and
BlackBerry have all but fixed this problem up until last week. Microsoft even
went as far as posting a guidance related to the fix. However, you will not
have much luck in implanting the fix. Microsoft last week pulled Exchange 2010
SP1 RU3 due to a new issue that popped up with BlackBerry Enterprise Server
after installing RU3. The new issue no longer delays messages to BES it
duplicates them. Microsoft is hard at work trying to nail down the problem and
is asking for people to hold off installing RU3. If RU3 is installed and you are
experiencing this issue the update can be removed from the mailbox servers and
functionality should return to normal. For some customers this news is extremely
unfortunate due to the fix in RU3 that enabled UPD notifications for outlook
2003. As information is made available, I will post another update.
Lab Update!
So all of my server equipment arrived on Thursday and is now setup and running
in full affect. wasting no time breaking it in, I moved this blog site from an
old LAMP box to a new and improved VM hosted on the new equipment. After
talking to a friend and giving some thought to my lab setup, I have decided to
drop some of the single site HA and go with a new and improved lab setup. I
will also be looking for fellow techie people to Federate environments with.
have a lot of work to do and not much time to get it all done. I think I am
going to need some more RAM….
New Server Coming Tomorrow:)
In preparation for my training class I ordered a new server for my Exchange 2010 Lab. I started off with a system that is highly expandable so I won’t have to buy another server any time soon. for all the geeks the part list is below and a picture of the Chassis. I have also include a draft of the lab setup..
1 x CHASSIS NORCO|RPC-4220
1 x SERVER_MB TYAN|S8230GM4NR
1 x CPU AMD|OPT X8 6128 2.0G G34
2 x MEM 8G|WINTEC 3SH13339R5-8GR
1 x RAID CARD LSI|LSI00200 9240-8I SGL
4 x HD 1T|SEAGATE 7K ST31000528AS
1 x PSU COOLMAX| ZP-1000B 1000W
1 x CABLE ATHENA|CABLE-YEPS828
1 x CPU COOL DYNATRON| A6
1 x CABLE ARECA|CB-8787-75M
Exchange 2010 Recover Server Setup
A little help for anyone who has had the joy of or in the future plan on
recovering an Exchange 2010 server from scratch. If you read This TechNet Article it seems so simple. A few
pointers can help in a pain from recovery. if you can’t restore the OS from
backup and you are going to just build a new server follow these steps
- Build OS
- Patch OS to same version
- Give Server same name and IP address
- Join to domain
- Install all required Exchange 2010 prerequisites
- Add the computer back to all related Exchange 2010 groups
- Exchange Install Domain Servers
- Exchange Servers
- Exchange Trusted Subsystems
- run Exchange setup.com /m:RecoverServer
Life is good!
TweetExchange 2010 Mailbox Moves and Database Availability
Not every mailbox move is a one off. Most of the time in my line of work I am trying to move a significant number of mailboxes in any given evening. This can be complicated if you target Mailbox happens to be sitting on a Database Availability Group (DAG) database. An issue that can rear it ugly head is the log replay depth of the secondary or ternary copy of the database can get significantly long. This is usually attributed to one of the following: speed of links, disk performance, a database(s) offline, or slew of other possibilities. Once this log depth gets to high the database is flagged as being to far out of sync the mailbox move will be stalled. Now if you are lucky and have a fast link the queues will empty and life goes on. However, if you are like most of the world in and suffer from inadequate resource the mailbox move will stall out and fail
Microsoft has 3 fixes for this problem
1. Remove move the request
2. Resolve the issue with the database copies. (Bring them up to date and healthy)
3. Modify the database with the set-mailboxdatabase command to update the DataMoveReplicationConstraint to allow the moves to continue regardless of copy status.
However I have another idea that work very well and save a lot of bandwidth and time. with that being said this is not always an option depending on your service level agreements and database copy requirements.
Here are the steps from a high level
1. Remove all remote copies of the database.
2. Perform mailbox moves
3. Backup database to removable storage.
4. Walk,Bus,Drive,Fly, UPS or FedEx removable drive to remote location
5. Copy database to proper location
6. Add a mailbox copy using powershell with the SeedingPostponed parameter set
7. Database plays logs and catches up with active copy*
8. Done.
*All logs must be present since the database backup was made in order to re-sync with active copy
Tweet


