Hello, I''m trying to figure out why I''m getting about 10MB/s
scrubs, on a pool where I can easily get 100MB/s. It''s 4x 1TB SATA2
(nv_sata), raidz. Athlon64 with 8GB RAM.
Here''s the output while I "cat" an 8GB file to /dev/null
root at solaris:~# zpool iostat 20
capacity operations bandwidth
pool alloc free read write read write
---------- ----- ----- ----- ----- ----- -----
rpool 123G 26.0G 2 5 206K 38.8K
tera 3.04T 598G 19 43 813K 655K
---------- ----- ----- ----- ----- ----- -----
rpool 123G 26.0G 1 0 199K 0
tera 3.04T 598G 966 0 121M 0
---------- ----- ----- ----- ----- ----- -----
rpool 123G 26.0G 1 8 212K 60.7K
tera 3.04T 598G 1.53K 7 195M 20.9K
---------- ----- ----- ----- ----- ----- -----
and here''s what happens when I''m scrubbing the pool:
---------- ----- ----- ----- ----- ----- -----
rpool 123G 26.0G 1 7 106K 78.2K
tera 3.04T 598G 87 8 10.5M 20.7K
---------- ----- ----- ----- ----- ----- -----
rpool 123G 26.0G 0 7 90.3K 81.8K
tera 3.04T 598G 87 7 10.3M 18.1K
---------- ----- ----- ----- ----- ----- -----
rpool 123G 26.0G 1 0 130K 0
tera 3.04T 598G 88 0 10.5M 0
---------- ----- ----- ----- ----- ----- -----
I''d be glad to provide any info you might need.
Thanks.
--
This message posted from opensolaris.org
> From: zfs-discuss-bounces at opensolaris.org [mailto:zfs-discuss- > bounces at opensolaris.org] On Behalf Of Hernan F > Subject: [zfs-discuss] Scrub extremely slow?Perhaps this is related? http://hub.opensolaris.org/bin/view/Community+Group+zfs/11 Zpool version 11, introduced "improved scrub performance"> Hello, I''m trying to figure out why I''m getting about 10MB/s scrubs, on > a pool where I can easily get 100MB/s. It''s 4x 1TB SATA2 (nv_sata), > raidz. Athlon64 with 8GB RAM.With sata disks, I expect approx 500Mbit/sec per disk for sustainable sequential operations. So I would agree you should easily be able to do 100MB/s = 800Mbit/sec. You should actually be able to do something around 2x that. I don''t know why your scrub is slow. Perhaps if you sit and watch it longer, you''ll see sometimes it''s high and sometimes it''s low?
I tested with Bonnie++ and it reports about 200MB/s. The pool version is 22 (SunOS solaris 5.11 snv_134 i86pc i386 i86pc Solaris) I let the scrub run for hours and it was still at around 10MB/s. I tried to access an iSCSI target on that pool and it was really really slow (about 600KB/s!) while the scrub is running. -- This message posted from opensolaris.org
----- Original Message -----> I tested with Bonnie++ and it reports about 200MB/s. > > The pool version is 22 (SunOS solaris 5.11 snv_134 i86pc i386 i86pc > Solaris) > > I let the scrub run for hours and it was still at around 10MB/s. I > tried to access an iSCSI target on that pool and it was really really > slow (about 600KB/s!) while the scrub is running.iSCSI access (and ZFS) during scrub is quite bad. If you add an SLOG (on SSD), this will help out. Also, adding an L2ARC SSD (or two) will help further. The scrub performance issues have been worked on in later ON versions, but then, Oracle doesn''t seem to want to release those as distros, which is a bitch. Vennlige hilsener / Best regards roy -- Roy Sigurd Karlsbakk (+47) 97542685 roy at karlsbakk.net http://blogg.karlsbakk.net/ -- I all pedagogikk er det essensielt at pensum presenteres intelligibelt. Det er et element?rt imperativ for alle pedagoger ? unng? eksessiv anvendelse av idiomer med fremmed opprinnelse. I de fleste tilfeller eksisterer adekvate og relevante synonymer p? norsk.
Too bad then, I can''t afford a couple of SSDs for this machine as it''s just a home file server. I''m surprised about the scrub speed though... This used to be a 4x500GB machine, to which I replaced the disks one by one. Resilver (about 80% full) took about 6 hours to complete - now it''s twice the size and it''s taking 10x. Guess I''ll just have to wait for a fix. Thanks! -- This message posted from opensolaris.org