Solved

File System Subclient File Scan take longer

  • 1 June 2023
  • 7 replies
  • 285 views

Userlevel 2
Badge +7

Hi Team,

I have this Scan phase issue where the scan took longer for one particular subclient (NFS volume) after Synthetic full job. Is there any setting / need add any additional setting to make this scan phase finish faster?  From the logs and behavior, I can see the mounting not happen and the scan result from the logs is 0. after many hours, it will act to mount the snap and scan the folders. what may cause this issue?

I only can suspect the subclient as the other subclient working as normal. The CIFS subclient for this same SVM have no issue. 

 


 

icon

Best answer by Damian Andre 1 June 2023, 23:16

View original

7 replies

Userlevel 6
Badge +14

Hi @Raj Balaraj ,

This is expected behavior when different access nodes is used on scanphase for the first job run after trueup job, can you confirm?

If not, we can use a key, but would prefer we check first.

Are you able to log a case?

Best Regards,

Sebastien

Userlevel 2
Badge +7

@Sebastien Merluzzi 
If the first job means the First Full job? yes. i did use another node for that backup and in middle of incremental job, i change it to current node. means it cant read back the index in current node? 

Userlevel 6
Badge +14

@Raj Balaraj ,

The Incremental after the Synthetic Full.

TrueUp is enabled via subclient Option [0], On-Demand TrueUp [0], Incremental after SynthFull [1], State file set to run scan [0], State file set to run index query [0]

Any reasons why you changed the Access Node?

Userlevel 2
Badge +7

@Sebastien Merluzzi 
Yes. i want to troubleshoot the 2nd node disk space (the exact node). so i change it to 1st node for the backup job. i believe the job result should still exist in that node yet i didn't do any changes so far on that node disk. 

Userlevel 6
Badge +14

@Raj Balaraj ok, run another backup afterwards.

Userlevel 7
Badge +23

Some extra info here on true up.

 

True-up is a special function that validates that the synthetic full is not missing data. Consider that file share backups use modified time to detect changed files, if there is a blimp in the system and that file is unavailable during scan, then it will never be picked up by an incremental or synthetic full ever again. So true-up runs on the first incremental after a synthetic full as extra protection to make sure all files are properly captured. The time it takes is dependent on the number of objects / files to scan across the original source.

Userlevel 2
Badge +7

@Sebastien Merluzzi @Damian Andre 
Thanks. I have also like to increase the Linux Media Agent Disk space and seems the 60GB is not enough. planned to upgrade another 60GB for each node. my NDMP data total size about 150TB which i converting to FS. 

Reply