You are browsing the archive for Design.

From the forums: End-user Recovery when file server is down

2:19 pm in Uncategorized by mikeresseler

Just read a very interesting topic on the forums (link)

A user reported that when the file server is down, the users in his company were still able to get to their files (read-only) through the DPM server.

So basically, when the file server is down, end-users were still able to contact the last backup data through the DPM server through a share named “protectedServerName_protectedServerShare".

He wondered if this was by design, or a bug.

The answer came quickly.  This is by design.  So if your file server is down, and the uses had access to the end-user recovery on the file server, then he or she will still be able to access the data through the DPM server.



System Center Data Protection Manager SP1: How to start a DPM project

1:43 am in Uncategorized by mikeresseler

// Note: For some reason, my tables are not shown.  I’ll try to fix this but I don’t seem to be able to figure out why… Sorry about that.

As promised, here is a guidance on how to start a DPM project, based on the IPD Guides of Microsoft.

First thing to know is when you implement a DPM project (and this counts for about ALL of the System Center Suite) you need to know and understand exactly the business requirements.  Without them, it is impossible to deliver a great implementation.  With them, and understanding them, then the implementation will be quick, easy and your backup worries will decrease a lot.

Oke, here goes

First, I will give the Decision Flow according to the IPD


All these steps will now be discussed

Step 1: Project Scope

In this Step, you will need to collect all the information necessary for the implementation.

  • AD domain & Forest information

You start with the AD Domain and Forest information.  The servers you need to protect need to be in the same domain or there have to be a two-way cross-domain or cross-forest trust between the domains where the protected servers will be located.

I normally use the following table to write down the information

Domain Name FQDN Netbiosname DC
  • Network Topology and Bandwidth

Make sure that you have an overview of the Network Topology and Bandwidth.  It will be difficult to protect a server every 15 minutes if it is located on a WAN connection that has high latency or doesn’t always have connectivity.  A drawing of the topology can be very handy when designing the solution.

  • Data Loss Tolerance

The business will need to give this input.  What is tolerated in case of a disaster.  This is the equivalent to the recovery point objective (RPO).  This is necessary to determine the load on servers, storage and tapes.  Don’t let the business tell you that there is no tolerance for data loss if they are not prepared to pay the price for the storage, servers and tapes.  If there is not enough budget, then you can’t get it all…

  • Retention Range

Q from IT: How long must data be kept for availability?

A from business:…

Mostly it is important to verify if all services / data / … have the same retention range.  From time to time it is not necessary to keep certain data for 6 months or longer.  You should always ask the business whether they need a 3 month copy or if the last week is ok.  The better you precise the question about the different applications, the better the answers will be and you will save storage and tapes that you can use for more important issues.

Q from IT: Are you under some regularity compliance? (HIPAA, SOX…)

A from business:…

If you are under compliance, the retention ranges and stuff are already defined for you.  Read them and implement them.  End of story.

  • Speed of data Recovery

This is similar to the Recovery Time Objective (RTO) and will determine when disk is used or when tape is used.  The quicker you need to be able to recover, the more disk you will use and vice versa.

  • End-User Recovery

Will end-users be able to recover their own deleted files without the intervention of IT? What’s the business requirement on this one?

  • BCP / DRP

Will this implementation be part of the Business Continuity Plan (BCP) or/and Disaster Recovery Plan (DRP).  In other words, if it is part of the BCP, this means that you need to be able to recover crashed items asap.  If it is only part of the DRP plan, then you need to have a good strategy to recover when things fail, but then it is not necessary to recover on the spot.

  • Future plans

Are there any business acquisitions or divestments planned in the near future?  Will the DPM solution be used for this?  Are there servers or applications that will be retired in the near future.  Do we need to calculate a new application in the design?

Step 2: Determine What Data Will Be Protected

In this step, you will need to figure out what kind of data you will be protecting.

  • Virtual Machines

As you know, DPM can protect entire guest VM’s.  Fill in the next tables to get an overview of all VM’s you need to protect (This includes Hyper-V and Virtual Server 2005 SP1 virtual machines)

Additional note: Pass-through disks are NOT protected with this method.  You will need to backup that disk with an agent inside the VM.


Host IP


Guest IP



  • Exchange Server

DPM can protect only mailbox servers.  So no edge servers or other roles.  Only the data is protected.

Additional information:

– Exchange Server 2003 and Exchange Server 2007 Single Copy Cluster (SCC); DPM agent on all nodes in the cluster

– Exchange Server 2007 local continuous replication (LCR): Install the DPM agent on both the active and passive node

– Exchange Server 2007 cluster continuous replication (CCR): The DPM agent must be installed on both nodes in the cluster

– Exchange Server 2007 SP1 standby continuous replication (SCR): Install the DPM agent on the active node and standby nodes

Again, I use a simple table and fill in the data



Server IP

Storage Group


  • Sharepoint services

What are we going to protect here?  Again my tables… :-)



Server IP




This can be for Sharepoint Services 3.0, MOSS 2007 or Sharepoint portal server 2003.

Please note that when you protect a sharepoint farm or a sharepoint services site that you don’t need to backup that database separately afterwards.  It will only cause you troubles.

Also note that for recovering your sharepoint sites, you will need a recovery server.  So keep that in mind when you need to ask for more servers.

  • Volumes, folders and shares

No explanation necessary here I think, only remember that we are talking about windows server 2003 SP1 or later.  No more windows server 2000!




  • System State

Things that are protected by the system state are listed in the following article:



Server IP

  • Exclusions

Write down the exclusions here.  This can be folder based, File based or File extension based.  Maybe interesting if you don’t want to backup the entire companies MP3 collection 😉

Think also about the following: If DFS-N is in place (Distributed File Share-Namespace) then map to the actual file locations, because shares through the DFS hierarchy cannot be selected for protection, only the target paths can be selected.  If DFS-R (Replication) is used, the map to all the replicas and then select one of them for protection.



Excluded folder





Excluded files




Server IP

Excluded file extension



Step 3: Create Logical Design for Protection

In this step, the protection requirements will be translated into a logical design.  And that logical design will be configured as one or more protection groups.  But before you start, stop for a moment and consider the following VSS limitations

  • File protection to disk is limited to 64 shadow copies
  • File protection can have a maximum of 8 scheduled recovery points for each protection group each day
  • Application protection to disk is limited to 512 shadow copies, incremental backups are not counted towards this limit

Keep those in mind while designing this step

To do this, you will now need to fill in the next table so you can determine recovery goals, protection media and how the replica’s will be created.

Here is the explanation for the various parameters

Server or workstation: Name of the Server and if it is a server or a workstation
Location: Location of the data
Data to be protected: Application data or File data.
Data Size: Current size of the data
Rate of Change: How fast does the data change?
Protected volume: The name of the protected volume (if applicable)
Synchronization Frequency: How many times do we need to apply the changes to the replica
Retention Range: How long must this data be kept available (online and/or offline)
Recovery Point Schedule: How many time between RPs
Media: Which media is used? (disk or tape or disk/tape)
Replica Creation Method: Automatic or manual (backup/restore)?
Protection group name: Choose a name
DPM Server: Choose the correct DPM server. (If more then 1 server will be in place)


SQL Production Databases

Server or workstation



Physical Location, eg Antwerp Office

Data to be Protected


Data Source Type


Data Size

in total, 600 GB, calculated to 1 TB in five years

Rate of Change


Protected Volume

SQL Store

Synchronization Frequency

15 min

Retention Range

7 days

Recovery Point Schedule

9.00, 12.00, 15.00, 18.00, 21.00



Replica Creation method


Protection Group name

SQL Production

DPM Server


If you make this map for each of the data you need to backup, you already designed your protection groups.

Step 4: Design the Storage

Here’s the tricky part.  It is almost impossible to correctly calculate how much storage you need.  There are a few helpfull hands on the internet, but most of the time I have seen that taking the complete storage and make it two times that size is good enough.  This is (of course) when you want to use the synchronization features at full.  If you are only interested in the traditional way of backing up, then you can go with less.

Anyway, here are a few links for storage calculation

  • Custom Volumes

Do we need to consider custom volumes?  Only if:

  1. Critical data must be manually separated onto a high performance LUN
  2. To meet regulatory requirements
  3. To separate IO-intensive workloads across multiple spindles
  • Choose the Disk Subsystem

If you have an option, decide what you are going to use as disk subsystem.  Will you be using DAS, SAN, iSCSI?  What RAID configuration?  Choose this based on the Peak IOps during backup or restore but in my humble opinion, a good iSCSI solution will do the trick without any problems (Think Dell MD3000i for example…)

  • Tape Storage

What tapedrive model or robotic library will you be using?  Is it compliant?

Check for compliancy

  • Placement of Disk and Tape Storage

What is the location of the disk and tape storage towards the DPM server?  Is it close?  Is it network connected, fiber? scsi connected?

Step 5: Design the DPM Server

Finally you are getting to the end of this process.  You can design the DPM server itself.

  • calculate how many DPM Servers are needed

These are the limitations of one DPM Server:

  1. Maximum 250 storage groups
  2. Maximum 10 Tb for 32-bit DPM servers
  3. Maximum 45 TB for 64-bit DPM servers
  4. Maximum 256 data sources per DPM server (64-bit) where each data source needs two volumes
  5. Maximum 128 data sources per DPM server (32-bit)
  6. Maximum 8000 VSS shadow copies
  7. VSS Addressing limits: Add a DPM server for each 5 TB (32-bit) or 22 TB (64-bit)
  8. Maximum 75 protected servers and 150 protected workstations per server
  9. Data sources in another domain / forest that is untrusted… Add a new DPM server
  • Map protection groups to servers and storage.

Well, as already said, if more then one DPM is in place, map the table to the correct server.  Few pointers here:

  1. Separate data that cannot coexist on the same server for legal or compliance reasons
  2. Group protection groups that have different synchronization frequencies
  3. Group protection groups with the same media requirements
  4. Group protection groups that comprise data sources that are within the same high-speed network.
  5. Group protection groups that will be backed up from or to VM’s.
  • Hardware requirements

According to Microsoft:





1 Ghz

2.33 Ghz quad-core CPUs


2 GB

4 GB ram


0.2 % the size of all recovery points + 1.5 times the RAM


Disk Space

Program Files: 410 MB

Database file drive: 900 MB

System Drive: 2650 MB

2-3 GB free on the PF volume

Disk Space for Storage Pool

1.5 times the size of the protected data

2-3 times the size of the protected data

Logical Unit Number (LUN)


Maximum of 17 TB for GPT dynamic disks

2 TB for MBR disks

  • Software Requirements

You need to know these 5 things before deciding to place DPM on a server.

  1. NO ia64-bit OS
  2. NO Microsoft System Center Operations Manager on same server.
  3. NO domain controller or application server
  4. Windows Server 2008 (Standard & Enterprise Edition)
  5. Windows Server 2003 with SP2 (R2)
  • Virtual or not?

Yes you can run DPM virtually when you use pass-through disks or iSCSI device.  Please note that you can’t connect to a tape library directly attached to that server at that time.

  • Database

Please keep in mind that you need to run the DPM database on a separate SQL instance!  You also need to plan for SSRS to be implemented on each DPM server.  It is necessary, you can’t without.

  • Dedicated Network

Will you be using a dedicated network?  If so, write it down.

  • Fault Tolerance and protection for DPM

Two components of DPM can be made fault tolerant: The DPM server and the DPM database.  However, keep this in mind for fault tolerance:

  1. Server cannot be run as an MSCS clustered application
  2. Server can run in a VM, which can be a part of a clustered environment
  3. Database is not supported in an MSCS cluster
  4. DPM Server can backup its own databases to tape.
  5. A DPM Server can be used to protect the data from other DPM Servers.


Oke, that’s it.  Before you even started to do something, you have gathered all the information necessary to deploy a good DPM implementation.

It will lower the changes of failure and even (if necessary) point out to the management that additional resources are needed or that you can not deliver the asked business requirements