https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=284437
[email protected] changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |[email protected] --- Comment #1 from [email protected] --- There has been a lot of confusion about this message. What we know: 1) The command can detect corruption, but 2) the command has internal limits to avoid excessive memory consumption. These limits probably seemed reasonable at the time but they are easily reached by a lot of people on perfectly healthy pools. Presumably the limits need to be adjusted. 3) The #2 false positive is scaring the heck out of people for no good reason. If you read the github bug https://github.com/openzfs/zfs/issues/15030 you'll see that people have been recovering pools. On linux they set zil_replay_disable (via procfs) to mount the damaged pool and then and zfs_recover (via procfs) and run a scrub. We have equivalent sysctls. YMMV of course. -- You are receiving this mail because: You are the assignee for the bug.
