Date: Tue, 24 Sep 2019 12:27:19 -0400 From: John Fleming <john@spikefishsolutions.com> To: Pete Wright <pete@nomadlogic.org> Cc: freebsd-stable@freebsd.org Subject: Re: Question about bottle neck in storage Message-ID: <CABy3cGx5_HMNjaL=%2B9ZvCqHVK9ZkdGs%2BHdznymTX8ODZ7Gk=zA@mail.gmail.com> In-Reply-To: <9415ff89-a36a-86ee-3f3f-47e9b807059e@nomadlogic.org> References: <CABy3cGxjhTcg%2Bpg2FiCc4OqG4Z1Qy1vFdvo7zU_t0ahC3mb%2BYw@mail.gmail.com> <9415ff89-a36a-86ee-3f3f-47e9b807059e@nomadlogic.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Sep 24, 2019 at 12:05 PM Pete Wright <pete@nomadlogic.org> wrote: > > > > On 9/24/19 8:45 AM, John Fleming wrote: > > Is there anyway to see how busy a SAS/Sata controller is vs disks? I > > have a R720 with 14 Samsung 860 EVOs in it (its a lab server) in raid > > 10 ZFS. > > > > When firing off a dd I (bs=1G count=10) seems like the disks never go > > above %50 busy. I'm trying to figure out if i'm maxing out SATA 3 BW > > or if its something else (like terrible dd options). > > > > my setup is Dell R720 with 2 x LSI 9361 cards. Each card is going to a > > dedicated 8 drive board inside the front of the R720. Basically i'm > > just saying its not a single SAS cable to 14 drives. > > > > Don't have cpu info hand.. zeon something. DDR3-1600 (128GB) > > > > Both controllers are in 8x slots running PCIe gen 3. > > might want to take a look at sysutils/intel-pcm > (https://github.com/opcm/pcm). I *think* this should give you metrics > on PCIe bus utilization among other useful status. > ok I can check that, but to be clear what i meant was like have i maxed out SATA3 bw on the card not so much PCIe bw. > Also, lookup the bandwidth for the PCIe bus and see if your aggregate > disk throughput on one of the PCIe lanes is saturating the bus (pcm > should also help here). You can also run "zpool iostat -v 2" to see per > disk i/o metrics to help determine if this is an issue. > I think i've looked at that before but i'll check again. > > > BTW i'm sure this has been asked a million times but what would be > > some decent benchmark tests while i'm at it? > > I generally run several tests and then compare results, for example > bonnie++, iozone, iperf (writing over the wire and to disk) as well as > some more realistic scripts based on the use-case i'm building a > solution for. hope that helps. > I've done iperf just across network IO. I'm getting 4x25 Gb/sec so i'm good there i think (Mellonox Connectx4 ethernet mode). I'll poke around with bonnie++ and iozone. I used bonnie++ once but didn't really get what it was telling me but I really didn't put much into it. Thanks! > -pete > > -- > Pete Wright > pete@nomadlogic.org > @nomadlogicLA >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CABy3cGx5_HMNjaL=%2B9ZvCqHVK9ZkdGs%2BHdznymTX8ODZ7Gk=zA>