Some networks with multiple hops may have a very high network latency and are not suitable for replication with default VVR packet size of 8400. Each VVR packet gets broken up into multiple network packets and transmitted over the network. During the network transmission some of the UDP packets gets lost and the VVR secondary can't assemble the original packet from the received fragments. This causes retransmission of the same packets multiple times. This causes replication times out and replication pauses. So, changing the packet_size from 8400 to 1400 would help to minimize retransmission rate and keep the rlink connected.
 

To workaround this issue the following commands can be run on the Primary:

 

# /etc/init.d/vxrsyncd.sh stop
 

# /etc/init.d/vras-vradmind.sh stop
 

# /etc/init.d/vxnm-vxnetd stop
 

Change the packet size to a smaller value.
 

# /usr/sbin/vxedit -g set packet_size=1400
 

# /etc/init.d/vxnm-vxnetd start
 

# /etc/init.d/vras-vradmind.sh start
 

# /etc/init.d/vxrsyncd.sh start
 

 

VxVM VVR vxrlink WARNING V-5-1-3359 Attaching rlink to non-empty rvg. Autosync will be performed.
VxVM VVR vxrlink INFO V-5-1-3614 Secondary data volumes detected with rvg loc1_rvg as parent:
VxVM VVR vxrlink INFO V-5-1-6183 vol0: len=1048576 primary_datavol=vol0
VxVM VVR vxrlink INFO V-5-1-3365 Autosync operation has started
....
VxVM VVR vradmin WARNING V-5-4-2402 Primary RLINK rlk_loc1_rvg not in CONNECT state.

The replication protocol was UDP and default packet size was 8400. As soon as replication is resumed the resync would pause and the RVG links would be disconnected.

# vxprint -Pl |grep flag