[Bug 225337] z_teardown_inactive_lock held inordinately long

[bugzilla-noreply at freebsd.org]

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=225337

--- Comment #4 from Garrett Wollman <wollman at FreeBSD.org> ---
I changed the schedule of the replication cron job (which runs every 15
minutes) to *not* run during the usual backup window (which I don't have any
control over but is fairly consistent) and thus far have not seen the hang, so
I think I can reasonably conclude that the combination of "zfs recv -F" and
active backup traversal (or any other heavy use of the filesystem?) is toxic.

The filesystems in question are a special case for us, because the active write
site is on another group's server; we're doing the replication for backup/DR
purposes.

-- 
You are receiving this mail because:
You are the assignee for the bug.