Start a Conversation

Solved!

Go to Solution

3414

December 24th, 2019 11:00

SC5020, Red Hat: How to SCSI Unmap to reclaim space?

Storage: SC5020

OS: Redhat 6.9 Clustered. 

I did 750GB  db restoration, during the job, I had to write and delete backup files on 1.5TB volume. Now the volume is pretty much empty (99%). However, SC5020 shows the volume is 97% full. 

It seems as the space emptied out in Red Hat doesn't reflect to SC5020. What is the proper way and command to send space reclamation from Red Hat 6.9 to SC5020? 

 

Thank you in advance. 

December 26th, 2019 09:00

To share with anyone. 

 

1. from RHEL, # fstrim -v /dir1 (this will send SCSI UNMAP command to SC5020)

2. Expire all snapshots and coalesce last snapshot to Active Space. This will remove all space occupied by snapshots and truncate the whitespace. 

 

 

 

4 Operator

 • 

1.5K Posts

January 9th, 2020 10:00

Hello, 

 Just a  quick note.  It's up to the filesystem to manage this.  Not all filesystems support UNMAP. 

 What filesystem are you using for your cluster? 

  Common filesytems like EXT4 or XFS should have the option 'discard' in the /etc/fstab file.   Enabling UNMAP on each mounted volume you want UNMAP enabled on.  Or for local SSD drives as well. 

 Regards, 

Don 

 

 

No Events found!

Top