Forum Discussion

Jaywalker1979's avatar
Jaywalker1979
Level 3
8 years ago

network read failed(42)

Hello,

I'm currently experieincing problems trying to backup a physical Windows 2008 server (all local drives and system state) using Netbackup 7.6, the job backs up a random amount (usually about 8 -10 gb or so) then errors with the following.

11/10/2017 16:57:58 - Critical bpbrm(pid=5140) from client 192.168.100.90: FTL - socket write failed    
11/10/2017 16:58:54 - Info bptm(pid=8900) EXITING with status 42 <----------       
11/10/2017 16:58:54 - Error bpbrm(pid=5140) could not send server status message      
11/10/2017 16:59:16 - Info bpbkar32(pid=6724) done. status: 42: network read failed      
11/10/2017 16:59:16 - end writing; write time: 0:22:34
network read failed(42)

Does anyone have any ideas on what to troubleshoot?

Thanks in advance!

7.6
windows server 2012
  • Sujay24's avatar
    Sujay24
    8 years ago

    Kindly increase client read timout value on media and client server:-

    Admin console->Media server host properties->Select media server used for backups->Timeouts->Set Client read timeout to 3600

    Admin console->Media server host properties->Select media server used for backups->Timeouts->Uncheck use OS dependent timeout

    Admin console->Client server host properties->Select Client server whose backups is taken->Timeouts->Set Client read timeout to 3600

    Admin console->Client server host properties->Select Client server whose backups is taken->Timeouts->Uncheck use OS dependent timeout

    Monitor the backup job. If above changes made and still issue persists kindly set Client resliency by adding client name in Resilient Network:-

    Admin console->Master server host properties->Resilient Network->Add Client name whose backup is taken

    Initiate a fresh backup.

  • Sujay24's avatar
    Sujay24
    Level 4
    8 years ago

    Kindly increase client read timout value on media and client server:-

    Admin console->Media server host properties->Select media server used for backups->Timeouts->Set Client read timeout to 3600

    Admin console->Media server host properties->Select media server used for backups->Timeouts->Uncheck use OS dependent timeout

    Admin console->Client server host properties->Select Client server whose backups is taken->Timeouts->Set Client read timeout to 3600

    Admin console->Client server host properties->Select Client server whose backups is taken->Timeouts->Uncheck use OS dependent timeout

    Monitor the backup job. If above changes made and still issue persists kindly set Client resliency by adding client name in Resilient Network:-

    Admin console->Master server host properties->Resilient Network->Add Client name whose backup is taken

    Initiate a fresh backup.

  • Thiago_Ribeiro's avatar
    Thiago_Ribeiro
    Moderator
    8 years ago

    Hi Jaywalker1979,

    Did you check communication from media --> client and vice versa? Normally this error point to network communication problems. Something changed in your environment?

    Do you have another media server in your environment? If yes, do you already tried to run this backup job in another media server?

    Tell us more information about your environment. Do you know that NBU 7.6 is not supported by Veritas since Feb/17?

    NBU Master/Media version?
    O.S

    Take a look these TNs

    https://www.veritas.com/support/en_US/article.000028515

    https://www.veritas.com/support/en_US/article.000016227

    https://www.veritas.com/support/en_US/article.000082566

     

    Thiago

  • Marianne's avatar
    Marianne
    Level 6
    8 years ago
    Firstly, all local drives already includes system state. Having both in the policy means you are doing 2 system state backups in the same policy.

    Please select 'allow multiple data streams ' in policy attributes and remove system state from backup selection.

    Let us know how it goes.
  • manatee's avatar
    manatee
    Level 6
    8 years ago

    two phases of isolation:

    1. OS/network - check if you have firewall enabled; if you require firewall you have to let nbu services thru; ports 13782, 13724, and 1556
    2. if no firewall, from each end test for connection:
      • bptestbpcd -client <ip address or hostname> -verbose OR
      • bpclntcmd -hn <ip address or hostname>

     

    let us know the outcome.

    • Marianne's avatar
      Marianne
      Level 6
      8 years ago

      My guess is that the issue is caused by some sort of network interruption outside of NBU.

      The fact that it backs up a large amount of data before the failure means that there is no lookup or port connectivity issue.

      It could be a firewall timeout or TCP/IP at OS-level.

      The    contains a section for OS-related tuning factors for Windows. 

      Please have a look at network settings that need to be configured.

      Start with KeepAliveTime. Ensure this parameter is the same on master, media and client.

      Work through the rest of recommendations and see what is applicable to your OS. 
      Look at Kernel thread parameters as well.

      We have seen in the past that tuning of these parameters solved status 40 and 42 errors.

       

       

      • Jaywalker1979's avatar
        Jaywalker1979
        Level 3
        8 years ago

        Hi Everyone

        I've performed the troubleshooting tips you have all provided and I think adding the server to resilient network may have resolved the issue. I managed to run a complete backup after modifying that setting.

        I'll run another and let you know how I get on.

        Once again, thanks for all your time and responses, I really appreciate the help and support I receive in these forums.

        You guys are the best!