Forum Discussion

NBU35's avatar
NBU35
Level 6
8 years ago

backint processes are stacking on client in SAP HANA D

Hi 

We have SAP Hana db backups setup in our environment.

Master, media, client are at NBU7.6.0.3

we are intermittently facing issuw iwth log backups of HANA DB.

whenever any backup fails, backint process related to HANA backups are getting stacked on client and not clearing by itself, i have to manually kill these processes.

please provide your suggestions on how to fix this. they should clear automatically once job finished.

Also Can we control log backups from Netbackup ? As per my knowledge it can only be configured using HANA COnsole.

 

  • Hi NBU35,

    It´s important to know which SAP HANA version you are running. SAP introduced a new behavior with the SPS updates through a backint response timeout of 10 minutes which applies to log backups only. That means if a log backup is not performed within 10 minutes than SAP will kill the backint process and spawn a new log backup request in a now shorter interval. The issue here is that this is a kill -9 and leaves child process that NetBackup is spawning on the backint in a defunct state (no longer able to communicate back to the parent process). Check your system for "defunct" processes. This happened e.g. on SPS 10.2.x and a fix is in SPS 10.4. The SPS 10/11/12 recent updates contain a change to signal to the backup vendors to abort jobs (child).

    As you mentioned log backups are controlled in SAP HANA. NetBackup provides the infrastructure to ingest and store them. You can increase the log size which will have an effect on the amount of logs being sent (there is a max log size). The good news is that there are some changes coming in SAP HANA which will allow to manage the log files more efficient.

    regards,

    Andreas, Regional Product Manager Veritas IPS EMEA

4 Replies

  • Hi NBU35,

    It´s important to know which SAP HANA version you are running. SAP introduced a new behavior with the SPS updates through a backint response timeout of 10 minutes which applies to log backups only. That means if a log backup is not performed within 10 minutes than SAP will kill the backint process and spawn a new log backup request in a now shorter interval. The issue here is that this is a kill -9 and leaves child process that NetBackup is spawning on the backint in a defunct state (no longer able to communicate back to the parent process). Check your system for "defunct" processes. This happened e.g. on SPS 10.2.x and a fix is in SPS 10.4. The SPS 10/11/12 recent updates contain a change to signal to the backup vendors to abort jobs (child).

    As you mentioned log backups are controlled in SAP HANA. NetBackup provides the infrastructure to ingest and store them. You can increase the log size which will have an effect on the amount of logs being sent (there is a max log size). The good news is that there are some changes coming in SAP HANA which will allow to manage the log files more efficient.

    regards,

    Andreas, Regional Product Manager Veritas IPS EMEA

    • NBU35's avatar
      NBU35
      Level 6

      Hi Andreas

      you have hit the nail at right place. I have investigated and came to same place. 

      Time out value is 10 Minutes for log backups and after that they get termnated by Hana end, but in NBU they still are either still active or queued and then fails by some error like 54, 23.

      I have to kill them manually using kill command.

      I beleive there is a possibility to avoid it by increasing time out value for log backups.

      But I would be more intrested to know if there is any other way to get rid of these orphaned BACKINT processes.

      PS: Hana version is greater than 10 and there is no Defunct process on system.

      • Marianne's avatar
        Marianne
        Level 6
        Not too much of a problem if it takes you 5 months to reply, right?