Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
support ro rw degraded and skip_balance mount options #1728
The additions of the 'ro' (read only) and 'degraded' mount options will reduce (but not remove) the need to resource the command line in data recovery / drive failure scenarios. Ie the combination of these mount options can allow for a safe degraded pool mount which in turn would allow for Web-UI directed access to data on degraded pools. And the combination of rw and degraded options will allow mounting a pool that has missing devices in order that they might be replaced or deleted if btrfs raid levels allow, and if not then a raid level change could be initiated followed directly by a then legitimate device delete: given the btrfs raid level change.
It is understood that we currently do not report pool mount status #737 and #1531 and ideally this would come first however the ongoing enhancements to pool mount logic in #1547 and the planned enhancement of #1700 have these mount options are prerequisites. And situations such as those presented in #1416 and #1199 would also benefit.
But given the current state of btrfs rw,degraded mount having certain 'one shot' limitation in low drive count for raid level situations, we may, in adding this level of flexibility, allow for more foot shooting. However we have a chicken and egg scenario here and have to start somewhere. Plus our first responsibility is to report the state of the system and allow for mount options where they are useful. It is not proposed that we default to a degraded mount.