Hi,
Sorry for the late update.
Currently I force to failover from one data center (PDC) to another data center (SDC) and failback to PDC, it resolve the issue but I am not able to failover within data center.
Thanks and regards,
Wiehan
This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
Hi,
I try to create storage replica between two data centre using cluster-to-cluster replication on premieses (async mode since Test-SRTopology give a warning for Network latency (8 millisecond). All other only information).
We recreate few times and still getting event ID: 5015 (successfully established a connection to a partner replica) and follow by event ID: 5014 (Connection lost to the share hosting the partner replica) -> based on this article we upgrade the RAM from 8 GB to 16 GB.
FYI: All servers are using Windows Server 2019 Standard with 2 vCPU, 16 GB, and the storage are using iSCSI (Data: 2TB, LOG: 2TB)
May I know how to resolve this issue.
Thanks and regards,
Wiehan
Hi,
Sorry for the late update.
Currently I force to failover from one data center (PDC) to another data center (SDC) and failback to PDC, it resolve the issue but I am not able to failover within data center.
Thanks and regards,
Wiehan
Hi,
I'm making the assumption that you're using DFS Replication here. If that is the case, I have seen this before in a system that I managed.
TCP off loading was causing my issue. Please follow the steps below:
Step 1)
Disable RSS, NetDMA and TCP Offload
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters
EnableTCPA
EnableTCPChimney
EnableRSS
Step 2) Turn off checksum offloading
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Tcpip\Parameters
I hope this answers your question.
-------------------------------------------------------------------------------------------------------------------------------------------------------
--If the reply is helpful, please Upvote and Accept as answer--