NDMP backups for Large Volumes fails with Status 40 exactly after two hours . Detailed Job Status also shows Errors "Error bpbrm(pid=2004) could not write FILE ADDED message to stderr"

NDMP backups for Large Volumes fails with Status 40 exactly after two hours . Detailed Job Status also shows Errors "Error bpbrm(pid=2004) could not write FILE ADDED message to stderr"

Article: 100004901
Last Published: 2020-02-20
Ratings: 1 0
Product(s): NetBackup

Problem

NDMP backups for Large Volumes fails with Status 40 exactly after two hours . Detailed Job Status also shows Errors "Error bpbrm(pid=2004) could not write FILE ADDED message to stderr"

Error Message

Job Activity monitor Shows Error :

5:33:41 PM - connecting
5:33:41 PM - connected; connect time: 00:00:00
5:33:58 PM - begin writing
6:21:21 PM - Error bpbrm(pid=2004) could not write FILE ADDED message to stderr    
6:21:32 PM - Error bpbrm(pid=2004) could not write FILE ADDED message to stderr    
6:21:45 PM - Error bpbrm(pid=2004) could not write FILE ADDED message to stderr    
6:21:56 PM - Error bpbrm(pid=2004) could not write FILE ADDED message to stderr    
6:22:01 PM - Error bpbrm(pid=2004) could not write FILE ADDED message to stderr    
6:22:08 PM - Error bpbrm(pid=2004) could not write FILE ADDED message to stderr    
6:22:17 PM - Error bpbrm(pid=2004) could not write FILE ADDED message to stderr    
6:22:27 PM - Error bpbrm(pid=2004) could not write FILE ADDED message to stderr    
6:22:32 PM - Error bpbrm(pid=2004) could not write FILE ADDED message to stderr

 

Cause

The default TCP keep alive packets from the media server were being sent every two hours, which is insufficient.

Solution

Things to Check

1>  Make sure there is no firewall in between the Filers and the Netbackup Server.

If there is a Firewall in between, then increase the idle socket timeout on the firewall to an interval that is either long enough to allow the backup script to complete or to allow TCP keepalives to be sent, i.e. greater than 2 hours.

 

2> Decrease TCP keepalive settings on Master and Media Server  and confirm that the settings are less than any idle socket timeout configured within the network.

See:

TCP Keepalive Best Practices - detecting network drops and preventing idle socket timeout
https://www.veritas.com/content/support/en_US/article.100028680.html
 

3> We can minimize the load on the bpbrm process of the master server so it will update the file list less frequently,
Create the MAX_ENTRES_PER_ADD touch file on each media server that backs up the NDMP device.
UNIX:   echo 50000 > /usr/openv/netbackup/MAX_ENTRIES_PER_ADD
Windows: cd <Install_path\Veritas\NetBackup\
         echo > 50000 MAX_ENTRIES_PER_ADD

Note :  MAX_ENTRES_PER_ADD specifies how many files should be backed up before recording the information in the catalog database. The default is 5000. The maximum value is 100000.

 

 

 

Was this content helpful?