11-STABLE vs 11.0-RELENG test

Andriy Gapon avg at FreeBSD.org
Tue Dec 6 17:14:24 UTC 2016


On 06/12/2016 18:17, Alex Tutubalin wrote:
> Followup, same box as in 1st message, but with different HBA and 5 HDD:
> 
> Summary:
>  1) read speed depends on 'on-disk' data, not by code used while read.
>  2) FreeBSD 11.0-releng (svn up today) creates 'fast' big files (420+ Mb/s read
> speed)
>  3)  FreeBSD 11-STABLE creates 'slow' big files. (~200  Mb/s)
>  4) zfs send slow-dataset | zfs recv   (under 11.0-releng) creates
> 'intermediate' files (320Mb/s)
>  5) file copy slow-file ... (under 11.0-releng) creates fast copy (400+Mb/s)
> 
> So, ZFS write code in 11-STABLE looks broken.
> 
> Here detailed log of my tests:
> 
> Box: Intel i5-2400, 16Gb of RAM, Adaptec 5805, 5x4Tb drives (HGST NAS drives) in
> JBOD, merged in RAIDZ2, enough free space:
> $ zpool list
> NAME      SIZE  ALLOC   FREE  EXPANDSZ   FRAG    CAP  DEDUP  HEALTH ALTROOT
> zdata    18,1T  5,11T  13,0T         -     9%    28%  1.00x  ONLINE -

That's an interesting observation.

Could you please try starting with a freshly created pool and writing some data?
I wonder if 'zpool list -v' will show any significant differences between
different FreeBSD versions?
Another thing to check would be 'zdb -mm' but it could be very large.  However,
a difference should be sufficiently small as most of the space would remain unused.

And, just in case, please compare 'zpool get all' and 'zfs get all' between
those two versions.

-- 
Andriy Gapon


More information about the freebsd-fs mailing list