Hi,
On 08/23/2016 03:09 PM, Beard Lionel (BOSTON-STORAGE) wrote:
Hi,
I have noticed that when using a distribute volume, if a brick is not
accessible, volume is still accessible in read-write mode, but some
files can’t be created (depending on filename).
Is it possible to force a distribute volume to be put in read-only
mode during a brick failure, to avoid random access error?
You can either remount the volume as read-only. ( mount with "ro" option)
OR
You can set the volume read-only. (gluster volume set <volume name>
read-only on)
I’m afraid of application behavior in this case, when only some files
are not accessible. I prefer to have volume not writeable when a brick
failed.
It is better to configure a distribute replicate volume, so that volume
is accessible(and writable) even if one node is down.
Thanks,
Saravana
_______________________________________________
Gluster-users mailing list
[email protected]
http://www.gluster.org/mailman/listinfo/gluster-users
