Found out recently that when reading a corrupt btrfs file using user shares the transfer doesn't abort with an IO error as it should, this is similar to what happened before with Samba AIO enable, but in this case it also affects local transfers, when using disk shares it works as it's supposed to, I never caught this before since I always tested with disk shares, it goes back to at least v6.6, which is the oldest release I tried.
At first I though it would be related to having direct IO enable, but I'm seeing the same with direct IO disable, I know of this issue occurring when O_DIRECT is used, does FUSE use that or anything similar?
This isn't a big deal since data corruption (on good hardware) is extremely rare and there other ways of monitoring for corruption, still it would be nice if it worked as it should.
-
1
Recommended Comments
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.