Evenly distribuite the data amount many directories!

[gilbertoferreira]

Hi there..

I have two servers with the follow disk layout:

/dev/sdc1 1,8T 1,6T 178G 91% /data1
/dev/sdd1 1,8T 1,6T 175G 91% /data2
/dev/sdb1 1,8T 225G 1,6T 13% /data3
/dev/sdf1 1,8T 108G 1,7T 7% /data4
/dev/sdg1 1,8T 108G 1,7T 7% /data5
/dev/sdh1 1,8T 105G 1,7T 6% /data6

All this directory data1....data6 are used to create a gluster distribuite-replicate volume.

As you can see, the /data1 and /data2 are almost full.
Question is:
Can I run gluster vol rebalance VOL start to fix this?
I meant, to distribute the data between all the other directoris evenly?
Initially I have had 3 directories in both servers, like:
server1|server2
/data1
/data2
/data3

Sometime after that, like 6 months after create the glusterfs volumes, I added more 3 disks in to both servers.
Now that's how it is:
server1|server2
/data1
/data2
/data3
/data4
/data5
/date6

I never perform any rebalance or fix layout, because never needed.
Now that 2 directories are nearly full, this situation comes up!

Is there any impact, if I took too long to perform a rebalance?
Honestly, I never know if a rebalance is needed when adding new bricks.
This should be in the gluster add-brick command output, something like this: You'll need to rebalance of fix the layout of you volume, after add-brick.

My worry is if the vm disks used will be read-only during the rebalance procedure.
I have more than 20 VM with Linux and Windows, using Proxmox VE qemu.

Best regards.

[sanjurakonde]

You can continue without doing rebalance as well. If you do rebalance, there might be a performance impact during the rebalance window. The disks won't be read only.
@pranithk please suggest.

[gilbertoferreira]

Hi

Thanks for your help.

Best regards