Block mode enabled for database copies

Question

Hi,

We have 3 servers in DAG:

1. Mbx01
2. Mbx02
3. mbx03

a. Mbx01 block mode is not enabled. Mbx02 and Mbx03 copies are enabled for block mode replication.

b. Backup does not work on Mbx02(Event: 8193- Volume Shadow Copy Service error: Unexpected error calling routine CoCreateInstance.  hr = 0x800401f0, CoInitialize has not been called. Operation:   Initialize For Backup, Event: 13- Volume Shadow Copy Service information: The COM Server with CLSID {XXXXXXXX } and name MSXML30 cannot be started. [0x800401f0, CoInitialize has not been called. Initialize For Backup.) It works on mbx01.

c. High replay queue length observed on Mbx02 and Mbx03 copies during DB failover and RQL takes long time to settle down. On MBx01 CQL settles down quickly. CQL is ok on all the copies.

d. No disk issues observed in event logs. Though it looks like disk issue.

e. AV exclusions are in place.

d. No high CPU/Memory issue observed.

e. Virtual memory updated correctly.

f. Disk sector size is verified and is same for all the DBs.

Any help towards the with CQL and backup is appreciated!

Thanks

Priya

Answer 1

Dealing with backup and replication challenges in a DAG setup, especially when block mode's involved, can get tricky. If Mbx02 and Mbx03 are set on block mode and you're hitting snags with backups not working and high replay queue lengths, it might not be just about block mode. Issues like these often point to deeper setup or environment factors - think network hiccups or initialization missteps for backup tools. Getting to the root of these issues could involve a closer look at network performance and ensuring all backup components are correctly initialized. But, if the internal troubleshooting maze gets too complex, consider a third-party backup solution tailored for Exchange environments. These tools are built to handle the peculiarities of Exchange backups and DAG configurations, offering a clearer, more manageable path to ensuring your data is safely backed up and recoverable, no matter the replication mode intricacies.

Answer 2

Hi @Priya Jayaraman ,

Block mode enabled for database copies

From your description, are you concerned about the "Block mode" enabled Mbx02 and Mbx03? From what I've learnt, block mode can be changed to File mode automatically if there are copy queue and reply queue length for databases, so personally I don't think it's the block mode that leads to the backup failure.

Does the backup work or not on Mbx03?
And how did you perform the backup, using Windows Server Backup or other tools?

Based on my experience, long RQL may also be caused by poor network, so it's also suggested to check the network connectivity.