You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
While working on #2384, which has inevitably involved overhauling some of our outdated tests, it came to light that we block all disk removal from a raid0 pool. This is likely down to an early development-based fencing that has inadvertently remained. We have had the same for btrfs-single but this has since been removed (although it too had some legacy test content).
It is proposed that we move our absolute block on disk removal for btrfs-raid0 pools to have the same minimum disk restriction already imposed on btrfs-raid1 pools: i.e. don't go below 2 disks. Size/usage constrains for disk removal are already in place for all raid levels.
As this measure affects critical path functions it is proposed for testing branch (update channel) only.
The text was updated successfully, but these errors were encountered:
Noting this issue should only be attended too once we have #2384 resolved as we will then have a more robust testing setup to prove our proposed more flexible fencing on the raid0 front.
We previously blocked raid0 disk removal outright. Move to
allowing device removal, as per all other profiles, within
profile limits.
Abstract btrfs-raid profile minimum device count & maximum
device missing info and use the former through-out the fencing
code. Adapt all related tests to the newly uniform error messages
that result from hitting the remaining revised limitations.
These code changes are also intended to ease our adoption of
raid1c3 & raid1c4 profiles. And form the initial awareness of
these newer btrfs-raid profiles.
While working on #2384, which has inevitably involved overhauling some of our outdated tests, it came to light that we block all disk removal from a raid0 pool. This is likely down to an early development-based fencing that has inadvertently remained. We have had the same for btrfs-single but this has since been removed (although it too had some legacy test content).
It is proposed that we move our absolute block on disk removal for btrfs-raid0 pools to have the same minimum disk restriction already imposed on btrfs-raid1 pools: i.e. don't go below 2 disks. Size/usage constrains for disk removal are already in place for all raid levels.
As this measure affects critical path functions it is proposed for testing branch (update channel) only.
The text was updated successfully, but these errors were encountered: