Exchange 2013 – Replication problem when different disk configuration is used


My friend came with problem to add second copy of databese in Exchange 2013 DAG, because replication service failed to perform initial seed due to different configuration between source (first copy)  and target (second copy)disks. I wanted to test it, so here is case study:

LAB:

  • 2x DC, 2x Exchange 2013 – CAS and Mailbox role in DAG, HAProxy load balancer, Exchange 2010 SP3 multirole server
  • All running under VMWare player
  • TESTDB: First copy on one DAG member on mountpoint F: (GPT, SCSI, NTFS , 8kB / sector)

Getting info about NTFS:

I used the Powershell function presented in this article http://cornasdf.blogspot.cz/2010/03/using-powershell-to-get-ntfs-info-such.html?showComment=1362006532506#c4110837443209977959

Thanks to Conrad then gathering NTFS info is as easy as:

Get-NTFSInfo f

and result

Drive                           : f
NTFS_Volume_Serial_Number       : 0xfaf6f756f6f7121d
Version                         : 3.1
Number_Sectors                  : 20901887 (0x00000000013eefff)
Total_Clusters                  : 1306367 (0x000000000013eeff)
Free_Clusters                   : 1234271 (0x000000000012d55f)
Total_Reserved                  : 0 (0x0000000000000000)
Bytes_Per_Sector                : 512
Bytes_Per_Physical_Sector       : <Not Supported>
Bytes_Per_Cluster               : 8192
Bytes_Per_FileRecord_Segment    : 1024
Clusters_Per_FileRecord_Segment : 0
Mft_Valid_Data_Length           : 524288 (0x0000000000080000)
Mft_Start_Lcn                   : 393216 (0x0000000000060000)
Mft2_Start_Lcn                  : 1 (0x0000000000000001)
Mft_Zone_Start                  : 393248 (0x0000000000060020)
Mft_Zone_End                    : 418848 (0x0000000000066420)
RM_Identifier                   : 0E320AB6-7A27-11E2-B355-000C2940FA17

Plan:

Test replication issues while using the following setup:

  1. Create second copy of the database using SCSI, GPT, NTFS but larger size of the block (65kB)
  2. Create second copy of the database using SCSI, GPT, NTFS but lower size of the block (2kB)
  3. Create second copy of the database using IDE, MBR disk with the same NTFS config as the first copy of the database
  4. Create second copy of the database using IDE, MBR disk with the different NTFS config as the first copy of the database (65kB block)

Ad 1) Create second copy of the database using SCSI, GPT, NTFS but different size of the block (65kB)

I have preconfigured the disk with the same drive letter F: , GPT and now I will format the NTFS to 65kB block size. The info from NTFS:

[PS] C:\Windows\system32>get-ntfsinfo f

Drive                           : f
NTFS_Volume_Serial_Number       : 0x5ed0732ad0730793
Version                         : 3.1
Number_Sectors                  : 20901887 (0x00000000013eefff)
Total_Clusters                  : 163295 (0x0000000000027ddf)
Free_Clusters                   : 162095 (0x000000000002792f)
Total_Reserved                  : 0 (0x0000000000000000)
Bytes_Per_Sector                : 512
Bytes_Per_Physical_Sector       : 
Bytes_Per_Cluster               : 65536
Bytes_Per_FileRecord_Segment    : 1024
Clusters_Per_FileRecord_Segment : 0
Mft_Valid_Data_Length           : 65536 (0x0000000000010000)
Mft_Start_Lcn                   : 49152 (0x000000000000c000)
Mft2_Start_Lcn                  : 1 (0x0000000000000001)
Mft_Zone_Start                  : 49152 (0x000000000000c000)
Mft_Zone_End                    : 52384 (0x000000000000cca0)
RM_Identifier                   : 1749F9E1-7CB8-11E2-B6CB-000C29ECA938

Adding DB copy:

Add-MailboxDatabaseCopy -Identity TestDB -MailboxServer FrontEnd1 -ActivationPreference 2

Result:

Works normally. Database is seeded and also incremental seed works. Before point 2 I removed the mailbox database copy and formatted NTFS with lower block size.

[PS] C:\Windows\system32>Remove-MailboxDatabaseCopy testdb\frontend1

Confirm
Are you sure you want to perform this action?
Removing database copy for database "TESTDB" on server "FRONTEND1".
[Y] Yes  [A] Yes to All  [N] No  [L] No to All  [?] Help (default is "Y"): y
WARNING: The copy of mailbox database "TESTDB" on server "FRONTEND1" has been removed. If necessary, manually delete
the database copy's files located at "f:\TESTDB" and "F:\TESTDB\Testdb.edb" on that server.
[PS] C:\Windows\system32>

AD 2) Create second copy of the database using SCSI, GPT, NTFS but lower size of the block (2kB)

I have preconfigured the disk with the same drive letter F: , GPT and now I will format the NTFS to 2kB block size.

Works normally. Database is seeded and also incremental seed works. Before point 2 I removed the mailbox database copy and formatted NTFS with lower block size.

AD 3)Create second copy of the database using IDE, MBR disk with the same NTFS config as the first copy of the database

Seeding works normally as in other configurations.

PS C:\Users\administrator.SALONOVI> get-ntfsinfo f

Drive                           : f
NTFS_Volume_Serial_Number       : 0x42f4703af47031f1
Version                         : 3.1
Number_Sectors                  : 20965375 (0x00000000013fe7ff)
Total_Clusters                  : 1310335 (0x000000000013fe7f)
Free_Clusters                   : 1300725 (0x000000000013d8f5)
Total_Reserved                  : 0 (0x0000000000000000)
Bytes_Per_Sector                : 512
Bytes_Per_Physical_Sector       : 512
Bytes_Per_Cluster               : 8192
Bytes_Per_FileRecord_Segment    : 1024
Clusters_Per_FileRecord_Segment : 0
Mft_Valid_Data_Length           : 262144 (0x0000000000040000)
Mft_Start_Lcn                   : 393216 (0x0000000000060000)
Mft2_Start_Lcn                  : 1 (0x0000000000000001)
Mft_Zone_Start                  : 393216 (0x0000000000060000)
Mft_Zone_End                    : 418848 (0x0000000000066420)
RM_Identifier                   : 700D6323-8444-11E2-9E80-000C29ECA938

Ad 4) Create second copy of the database using IDE, MBR disk with the different NTFS config as the first copy of the database (4kB block)

Works as other configurations.

Drive                           : f
NTFS_Volume_Serial_Number       : 0x0a262c1b262c0a71
Version                         : 3.1
Number_Sectors                  : 20965375 (0x00000000013fe7ff)
Total_Clusters                  : 10482687 (0x00000000009ff3ff)
Free_Clusters                   : 9682285 (0x000000000093bd6d)
Total_Reserved                  : 0 (0x0000000000000000)
Bytes_Per_Sector                : 512
Bytes_Per_Physical_Sector       : 512
Bytes_Per_Cluster               : 1024
Bytes_Per_FileRecord_Segment    : 1024
Clusters_Per_FileRecord_Segment : 1
Mft_Valid_Data_Length           : 524288 (0x0000000000080000)
Mft_Start_Lcn                   : 3145728 (0x0000000000300000)
Mft2_Start_Lcn                  : 8 (0x0000000000000008)
Mft_Zone_Start                  : 3145728 (0x0000000000300000)
Mft_Zone_End                    : 3350560 (0x0000000000332020)
RM_Identifier                   : 700D6344-8444-11E2-9E80-000C29ECA938

Result:

I haven´t found error or problematic configuration, however, there might be some stuff useful for others. It took so much energy to test, that I would still like to post this article for future refference.

 

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s