zpool scrub makes zfs get slow?

[Matthew Seaman]

Part of our monitoring relies on running commands like:

/sbin/zfs get -rHp -o value usedbysnapshots zroot

through snmpd.  We noticed today that this was timing out and it
corresponded to a regular 'zpool scrub' starting up.  While scrubbing is
going on, that command can take in excess of a minute to return:

% time /sbin/zfs get -rHp -o value usedbysnapshots zroot > /dev/null
0.298u 0.826s 1:08.62 1.6%	104+166k 47618+0io 0pf+0w
% freebsd-version -ku
10.2-RELEASE-p9
10.2-RELEASE-p12

There's a large but not huge number of ZFSes in that pool: about 2500,
and the drives are geli encrypted.

Is this expected behaviour?  Is there anything we can do to make 'zfs
get' more snappy in these circumstances?

	Cheers,

	Matthew




-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 972 bytes
Desc: OpenPGP digital signature
URL: <http://lists.freebsd.org/pipermail/freebsd-fs/attachments/20160413/864ff939/attachment.sig>