(13) file read failed

[lovy1]

Hello,

master - linux vers. 7.7.2

Client - windows vers. 7.7.3

This is a big client in TB's and we taking backup of each drive in separate policy. And mostly full backup failing with below error. pls assist. pls let me know if you need any other info.

08/19/2018 10:00:03 - connecting
08/19/2018 10:00:04 - connected; connect time: 0:00:00
08/19/2018 10:00:07 - begin writing
08/19/2018 10:24:07 - Info bpbrm (pid=11814) TXSLM2MWFILE001-backup is the host to backup data from
08/19/2018 10:24:07 - Info bpbrm (pid=11814) reading file list for client
08/19/2018 10:24:07 - Info bpbrm (pid=11814) starting bpbkar on client
08/19/2018 10:24:08 - Info bpbkar (pid=8392) Backup started
08/19/2018 10:24:08 - Info bpbrm (pid=11814) bptm pid: 11816
08/19/2018 10:24:08 - Info bpbkar (pid=8392) change time comparison:<disabled>
08/19/2018 10:24:08 - Info bpbkar (pid=8392) archive bit processing:<enabled>
08/19/2018 10:24:08 - Info bptm (pid=11816) start
08/19/2018 10:24:08 - Info bptm (pid=11816) using 262144 data buffer size
08/19/2018 10:24:08 - Info bptm (pid=11816) using 30 data buffers
08/19/2018 10:24:09 - Info bpbkar (pid=8392) not using change journal data for <S:\>: not enabled
08/19/2018 10:24:10 - Info bptm (pid=11816) start backup
08/19/2018 10:24:10 - Info bptm (pid=11816) backup child process is pid 11818
08/20/2018 05:21:25 - end writing; write time: 19:21:18
08/20/2018 05:45:14 - Error bpbrm (pid=11814) socket read failed: errno = 104 - Connection reset by peer
08/20/2018 05:45:16 - Error bptm (pid=11816) media manager terminated by parent process
08/20/2018 05:45:26 - Info bpbkar (pid=8392) done. status: 13: file read failed
08/20/2018 08:45:27 - Info bpbrm (pid=59062) Starting delete snapshot processing
08/20/2018 08:45:28 - Info bpfis (pid=13532) Backup started
08/20/2018 08:45:29 - Error bpbrm (pid=59062) from client TXSLM2MWFILE001-backup: Delete bpfis state from txslnno_master_cluster failed. status = 71
08/20/2018 08:45:29 - Info bpfis (pid=13532) done. status: 0
08/20/2018 08:45:29 - Info bpfis (pid=13532) done. status: 0: the requested operation was successfully completed
file read failed (13)

Thanks

[Marianne]

You will need logs for all of the processes that we see in the job details:

On the media server:
bpbrm (pid=11814 and 59062)
bptm (pid=11816 and 11818)

On the client:
bpfis (pid=13532)
bpbkar  (pid=8392) 

I suggest level 3 logs.

bpbrm log will show connection info followed by periodic catalog updates from bpbkar.
bptm will show data being received from the client.

bpbkar will show each file being sent to media server.
bpfis will show snapshot details. 

[Amol_Nair · Accepted solution]

Why is the nbu version on your client higher than the nbu version on your master server.?

was it a typo.?
If not please downgrade the nbu version on the client from 7.7.3 to 7.7.2 to match the version on the master server and then attempt your backup..

[Lowell_Palecek]

I agree with preceding suggestions. My additions:

Since bptm and bpbkar both sent logs to bpbrm after it reported the lost connection, I suspect bpfis. In addition to the NetBackup logs, look for a crash in the Windows event log.

Don't mix 7.7.3 with other versions when the client is Windows. There was a misunderstanding in the Corba protocol from Discovery that would make bpfis crash. If this happens, the last thing in the bpfis log for that pid is Corba stuff. We made an EEB that makes either a 7.7.3 master compatible with earlier versions or a 7.7.3 client compatible with an 8.x master. An 8.x master with 7.7.2 or earlier client is fine.

The following is not a problem.

>Delete bpfis state from txslnno_master_cluster failed. status = 71

In Windows VSS, bpfis executes in pairs. A bpfis create process manages a transaction with VSS, creates a snapshot, and creates a state file describing the snapshot. A bpfis delete process communicates the backup status to the first process over a named pipe and cleans up. It uses the state file to get oriented.

The fact that the state file did not exist further suggests that the bpfis create process failed before it could create the file.

[Lowell_Palecek]

Funny thing. For a Windows file system backup, bpbrm doesn't usually communicate with bpfis. I expect bpbkar32 to execute bpfis on an as-needed basis. If you don't find the problem in the bpfis log, you may need to look upward from bpbrm. That is, check the nbjm log on the master. Never mind nbjm. If it lost contact with bpbrm you wouldn't be getting job details.

Two reasons for bpbkar32 to need a snapshot are Windows Open File Backup (WOFB) and System State backup.