Hi All,

We have set up IBM MQ 9.3 in AWS EC2 with EFS (NFS 4.1) as the shared file system. We have configured it as a multi-instance queue manager for high availability.

When we start the queue manager as a multi-instance queue manager using the "strmqm -x QMGRNAME" command, we observe a delay/latency when data is written to MQ. It takes 20 minutes for 60K messages to be placed in the MQ with the multi-instance queue manager.

However, when the same queue manager is started using "strmqm QMGRNAME" without the multi-instance setup, we do not observe this delay/latency while writing data to MQ. It takes less than 3 minutes for 60K messages to be placed in the MQ without the multi-instance queue manager.

Can anyone help us understand why there is such a significant delay when using the multi-instance setup and suggest any possible solutions to resolve this issue?

Hi All,

We have set up IBM MQ 9.3 in AWS EC2 with EFS (NFS 4.1) as the shared file system. We have configured it as a multi-instance queue manager for high availability.

When we start the queue manager as a multi-instance queue manager using the "strmqm -x QMGRNAME" command, we observe a delay/latency when data is written to MQ. It takes 20 minutes for 60K messages to be placed in the MQ with the multi-instance queue manager.

However, when the same queue manager is started using "strmqm QMGRNAME" without the multi-instance setup, we do not observe this delay/latency while writing data to MQ. It takes less than 3 minutes for 60K messages to be placed in the MQ without the multi-instance queue manager.

Can anyone help us understand why there is such a significant delay when using the multi-instance setup and suggest any possible solutions to resolve this issue?

Hi All,

We have set up IBM MQ 9.3 in AWS EC2 with EFS (NFS 4.1) as the shared file system. We have configured it as a multi-instance queue manager for high availability.

When we start the queue manager as a multi-instance queue manager using the "strmqm -x QMGRNAME" command, we observe a delay/latency when data is written to MQ. It takes 20 minutes for 60K messages to be placed in the MQ with the multi-instance queue manager.

However, when the same queue manager is started using "strmqm QMGRNAME" without the multi-instance setup, we do not observe this delay/latency while writing data to MQ. It takes less than 3 minutes for 60K messages to be placed in the MQ without the multi-instance queue manager.

Can anyone help us understand why there is such a significant delay when using the multi-instance setup and suggest any possible solutions to resolve this issue?

Hi Kishore,
we had this issue with long transactions. 
Feedback from IBM is that it is related to file locking. Transactions in multi-instance mode do more file locking. This was a surprise to us because we expected that multi-instance mode would only add locks on "active", "master" and "standby", and all other locks would be the same as with a normal start.
IBM told us that it is expected behavior and recommended shorter transactions. No fix is planned.
If your application or test is putting messages in transactions, especially long ones, this might be the explanation. I suggest testing without transacitions to confirm.

Hi All,

We have set up IBM MQ 9.3 in AWS EC2 with EFS (NFS 4.1) as the shared file system. We have configured it as a multi-instance queue manager for high availability.

When we start the queue manager as a multi-instance queue manager using the "strmqm -x QMGRNAME" command, we observe a delay/latency when data is written to MQ. It takes 20 minutes for 60K messages to be placed in the MQ with the multi-instance queue manager.

However, when the same queue manager is started using "strmqm QMGRNAME" without the multi-instance setup, we do not observe this delay/latency while writing data to MQ. It takes less than 3 minutes for 60K messages to be placed in the MQ without the multi-instance queue manager.

Can anyone help us understand why there is such a significant delay when using the multi-instance setup and suggest any possible solutions to resolve this issue?

Hi All,

We have set up IBM MQ 9.3 in AWS EC2 with EFS (NFS 4.1) as the shared file system. We have configured it as a multi-instance queue manager for high availability.

When we start the queue manager as a multi-instance queue manager using the "strmqm -x QMGRNAME" command, we observe a delay/latency when data is written to MQ. It takes 20 minutes for 60K messages to be placed in the MQ with the multi-instance queue manager.

However, when the same queue manager is started using "strmqm QMGRNAME" without the multi-instance setup, we do not observe this delay/latency while writing data to MQ. It takes less than 3 minutes for 60K messages to be placed in the MQ without the multi-instance queue manager.

Can anyone help us understand why there is such a significant delay when using the multi-instance setup and suggest any possible solutions to resolve this issue?