Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 4 Jan 2017 10:30:37 -0800
From:      Freddie Cash <fjwcash@gmail.com>
To:        FreeBSD Filesystems <freebsd-fs@freebsd.org>
Subject:   Check status of background destroy
Message-ID:  <CAOjFWZ4S42mebL15iT8=TLcfvabBWeUdLDzi_d3WfhKTpuGdOw@mail.gmail.com>

next in thread | raw e-mail | index | archive | help
Is there a way to check the progress of a background destroy process?

I did a very bad thing in the middle of December.  "zfs destroy -r
poolname/dataset", where there's approx 40 TB of data in that dataset,
spread across about 400 snapshots, sitting on a dedupe-enabled pool.  :(
 (Usually, I delete individual snapshots, but thought I would try a
recursive destroy instead.)

Needless to say, the 128 GB of RAM in the box, with 128 GB of L2ARC on SSD,
is not enough to hold the DDT, and the box deadlocks after about 8 hours of
disk churning.  Once Wired memory goes above 120 GB, the system is
unresponsive to everything, including pings.

During that 8 hour period, the pool is usable; I can read and write data to
the pool, I can create and destroy snapshots, and I can receive snapshots
from other systems.  But the background destroy continues to run after
every boot, continues to churn through RAM, and locks everything up after
about 8 hours.

This has been going on for several weeks now.

Is there any way to check the progress of the background destroy to see how
much longer this will continue, how much data has been freed, how much of
the DDT has been updated, etc?  "zfs list" shows very little change in free
space.

Or am I just going to have to live with resetting the box every 8-10 hours
for the next few weeks?  Unfortunately, adding more RAM to the box isn't an
option right now.  :(

-- 
Freddie Cash
fjwcash@gmail.com



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAOjFWZ4S42mebL15iT8=TLcfvabBWeUdLDzi_d3WfhKTpuGdOw>