Saturday 7 April 2018

Possible network disconnect with primary database | Oracle Standby

Possible network disconnect with primary database in Oracle Standby Alert Log


Issue:

After starting the log shipping in Standby from Primary, the below mentioned warnings/errors started popping up in the Standby Alert log:

RFS[418]: Possible network disconnect with primary database
RFS[421]: Possible network disconnect with primary database
Thu Sep 21 11:07:50 2017
RFS[400]: Possible network disconnect with primary database
Thu Sep 21 11:07:53 2017
RFS[411]: Possible network disconnect with primary database


Resolution:

These errors in primary/standby indicate there is network/firewall hang or timeout causing ARCx process to hang on network on primary. Please note that these errors/warnings do not indicate any permanent network/firewall issue (as per to me). These indicate transient network/firewall issue like intermittent network slowness/hang etc.

There is mechanism where if primary detects that ARCx process is stuck on network, it will terminate and respawn that. Similarly there is mechanism on standby side to detect hang for RFS on standby side.

Please refer below the Oracle Doc which discusses more details regarding this.

Logs are not shipped to the physical standby database ( Doc ID 1130523.1 )

Also you can try the below mentioned steps to fix your issue:

1) Stop the log shipping by deferring the log_archive_dest_state belonging to the DR.
2) Stop the recovery at DR Site.
3) Restart the DR Database.
4) Start the MRP process to begin recovery.
5) Enable the log shipping again from Primary.

If these warning are not causing any issue in redo transfer, then we can safely ignore them. If redo transport is affected by these frequent kills/respawn, then please troubleshoot issue at network/firewall level. 

No comments:

Post a Comment