Forum Discussion

sugarboy's avatar
sugarboy
Level 4
13 years ago
Solved

Does a fresh backup don't fail when Highwater mark is reached?

Hi All,

I have an OST diskpool (Capacity : 500GB) which is configures with Highwater mark of 40% and low water mark eriof 30%.

Now i have ran backup of 600GB. As per the configuration this should fail at 200 GB (40% of 500 GB), however backup did not fail until LSU is completly full.

Is this expected?

Verified the other scernario,

where filled lsu up to 40% i.e. 200 GB, then ran the backup of 300 GB again which failed right away with Disk Storage Full or Highwater mark reached.

It works as expected in second scernario.

Please clarify me on the behavior of High and low water marks.

 

Thanks,

Sagar

  • When NBU runs a backup for a client it has not seen before it does actaully do a calculation of its size, but not based on that client or other clients, it does it based on the capacity and free disk space of the Storage Unit itself.

    Your strange combination of actual capacity, HWM and LWM setting may have actually led NBU to believe it could easily back up a new client and so continued with its backup.

    The disk filling up may have happened due to a combination of factors including NBU wanting to clear down space but not being able to.

    After this first run it would fail as it did then know the expected size of the client which it knew would exceed the disk capacity based on the HWM

    So your hunch in the title is about right - it did what it did because of it being the first time it had seen this client and combined with the setting you have just allowed it to keep on running.

    Hope some of this makes sense!

7 Replies

  • The backup will start, when it's initiated NetBackup does not calculate how much it's going to backup before running. It will just run. I'm guessing by your tests that it does check if the HWM is reached when a backup is ran. In your first case, it wasn't so it ran til it filled up.

    Which NBU version?

  • Sugarboy,

    It's just a guess at this stage, but I think it's working as intended. I don't think it's usual to have backups running that are the size or even half or quarter of the size of the storage units total size.

     

     

  • My question: WHY hwm of 40%? Just as a test?

    You are correct about your understanding of HWM - NBU is supposed to consider STU as full when HWM is reached.

    Just wonder how much time elapsed between you changing HWM and kicking off a job? Did NBU get a chance to re-read config? I have found that STU changes are not alway implemented/applied right away.

     

  • Yes, It is just a test.

    and i gave enough time for NBU to read the config. even restarted the service before i start the backup.

     

    Thanks,

    Sagar

  • Seems NBU has some 'built-in intelligence' that checks actual available space as well....

  • When NBU runs a backup for a client it has not seen before it does actaully do a calculation of its size, but not based on that client or other clients, it does it based on the capacity and free disk space of the Storage Unit itself.

    Your strange combination of actual capacity, HWM and LWM setting may have actually led NBU to believe it could easily back up a new client and so continued with its backup.

    The disk filling up may have happened due to a combination of factors including NBU wanting to clear down space but not being able to.

    After this first run it would fail as it did then know the expected size of the client which it knew would exceed the disk capacity based on the HWM

    So your hunch in the title is about right - it did what it did because of it being the first time it had seen this client and combined with the setting you have just allowed it to keep on running.

    Hope some of this makes sense!