Problem
NDMP backup jobs receiving a (636) error during the backup may not fail until a (227) error occurs during the same time frame as image cleanup.
Status 227
Status 636
Error Message
Error bpbrm (pid=##) db_FLISTsend failed: no entity was found (227)
read from input socket failed (636)
Cause
The overall cause is that nbjm is not receiving updates from bpbrm causing the (636) error. The backup image is then marked for deletion as a failed job. Bpbrm does not notice this interruption until image cleanup removes the image. Bpbrm is then no longer able to update bpdbm because the image can no longer be found (227).
The root cause is the same as TECH197144 where the TCP KeepAliveTime parameter on both master and media server were set higher than the KeepAliveTime set on the firewall.
Solution
Refer to Adjusting the TCP KeepAliveTime parameter to adjust the KeepAliveTime to match that of the firewall.
Applies To
NetBackup 7.5.x.x
Windows 2008 R2