Rlinks are going into connect/disconnect cycle
book
Article ID: 100004938
calendar_today
Updated On:
Description
Error Message
Event ID 98: V-203-24583-98 RLINK disconnected to remote
Event ID 99: V-203-40967-99 RLINK connected to remote
Cause
The common issues for an Rlink are going into a disconnect/connect cycle are related to network connectivity:
- VVR kernel heartbeat between hosts time out (during periods of no primary data change)
- Primary does not receive a positive network ACK for an update sent to the secondary after multiple retries
- Primary dose not receive a positive data ACK for an update sent to secondary within timeout period.
For example, after a period of succesful replication, acknowledgements for packets sent either stop or return to the primary after their timeout period. VVR resends the packets up to 50 times before it stops trying to send the packet. These unacknowledged packets start filling up the VVR network window buffer space on the primary. Once the network window buffer space is completely full of unacknowledged packets, VVR cannot process other writes from the SRL on the primary and replication comes to a stop.
Resolution
In order to isolate the issue, please follow the recommendations below:
- Please update the NIC drivers and ensure configurations are for explicit network settings (for example 1000fdx) and not AUTO.
- Please ensure that Network Address Translation feature is disabled, if any
- Disable the HARDWARE Compression device.
- If possible use separate NICs for replication and the public network with different subnet connection.
- It would also be helpful to show ping output and vxprint -VPl during the issue.
Issue/Introduction
It may be seen in Windows system event log that vxio logs rlink disconnect and connect messages continuously.
Additional Information
UMI: V-203-24583-98
UMI: V-203-40967-99
Was this article helpful?
thumb_up
Yes
thumb_down
No