White Papers

BP1013 Best Practices for Enhancing Microsoft Exchange Server 2010 Data Protection and Availability 25
Remote host the task runs as part of a schedule on a remote server, which has all the
ASM/ME and Exchange software components installed (Management tools)
In any of these cases after the checksum runs, the backup document of the Smart Copy is updated
accordingly.
We tested the impact of this activity when running on an ‘on demand’ basis directly on the Exchange
host running in the DAG and overhauled the host resources during the running of the checksum
verification against the mailbox databases residing in the Smart Copy snapshots. An outline of the
operations that were automated by ASM is reported in the following list:
• Submit/schedule the checksum verification
• The snapshot is turned online and presented to the host, mounted as a mount point under the
TEMP folder of the user account impersonating the job
• A scripted set of ESEUTIL commands are executed against the files in the volume in order to
validate their integrity (checkpoint, logs, database)
• An execution log file (EqlExVerifier.log) is generated under the EqualLogic folder and the
Smart Copy backup document is updated with the output of each run of ESEUTIL
• The volume is dismounted, disconnected and the snapshot is set back offline
While most of these steps did not considerably impact the local host resources of the Exchange server,
the execution of the set of ESEUTIL commands did. We recorded an average processor utilization of
16% for the ESEUTIL process alone (split across up to 15 threads) when it was running against the files
within the Smart Copy snapshot.
The time taken to execute the checksum verification was as reported in Table 5.
Table 5 Checksum verification time taken
File type
Size
Time taken
Bandwidth
Checkpoint (*.chk)
8KB
< 1 sec
N/A
Logs (*.log)
1MB each
< 1 sec each
N/A
Database (*.edb)
420GB
32 minutes
225 MB/second
The entire duration of this task would be variable depending on the amount of logs present in the
volume at the time of the snapshot (it can be thousands). The conservative forecast for this duration is
evaluated using the formula reported below, which is consistent with the performances of our
simulated infrastructure:
𝐶ℎ𝑒𝑐𝑘𝑠𝑢𝑚 𝐷𝑢𝑟𝑎𝑡𝑖𝑜𝑛
(
𝑖𝑛 𝑠𝑒𝑐𝑜𝑛𝑑𝑠
)
= 1 +
(
𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑙𝑜𝑔𝑠 ∗ 1
)
+
𝑆𝑖𝑧𝑒 𝑜𝑓 𝑑𝑎𝑡𝑎𝑏𝑎𝑠𝑒 (𝑖𝑛 𝑀𝐵)
𝐵𝑎𝑛𝑑𝑤𝑖𝑑𝑡ℎ (𝑖𝑛 𝑀𝐵/𝑠𝑒𝑐𝑜𝑛𝑑)
The checksum burden must be evaluated and should be taken into consideration carefully. The
amount of CPU cycles taken by the Exchange command line utility can be added to the regular
operation of the reference mailbox server, but, in the case of a critical situation, when the entire user
workload of the database is owned by one single host, the additional impact of a running checksum