Cluster abend when running backup

  • 7001810
  • 06-Nov-2008
  • 27-Apr-2012

Environment

Novell NetWare 6.5 Support Pack 6
Novell NetWare 6.5 Support Pack 7
Novell NetWare 6.5 Support Pack 5
Novell Open Enterprise Server (NetWare based)
Syncsort Backup Express

Situation

Cluster master node abends a few minutes after a backup has started.

You will see the following in the abend.log:
Server SERVER1 halted Friday, 29 August 2008   3.28.49,471
Abend 1 on P00: Server-5.70.07-0: This node in the Minority partition and the node in Majority partition is Alive.

Packet Receive Buffers have reached its maximum.

Resolution

By default Syncsort Backup Express distributes the catalog information traffic via the node running the Master_IP_Address_Resource. This will cause excessive traffic and also will cause the Packet Receive Buffers to increase. Due to the fact that the server will not receive enough or timely Packet Receive Buffers, the server will stop communicating via the NIC.

Backup Express now has the ability to distribute the catalog information traffic across cluster nodes hosting each volume, thus reducing load on the Master IP node. To enable this functionality, add --enableehdproxy to the SYS:\ETC\SSEVTHND file on all cluster nodes.

This is documented in Syncsort article A002790