Date: Mon, 17 Nov 2014 11:33:59 -0800 From: John-Mark Gurney <jmg@funkthat.com> To: k simon <moremore2@outlook.com> Cc: freebsd-scsi@freebsd.org Subject: Re: HDD slow down in heavy load Message-ID: <20141117193359.GP24601@funkthat.com> In-Reply-To: <BLU436-SMTP193623C93FF6EEF15E339F2EE8B0@phx.gbl> References: <BLU436-SMTP193623C93FF6EEF15E339F2EE8B0@phx.gbl>
next in thread | previous in thread | raw e-mail | index | archive | help
k simon wrote this message on Mon, Nov 17, 2014 at 18:17 +0800: > I observed sometimes DragonflyBSD do much better than FB in heavy > load , eg. squid or zabbix's postgresql database. > The difference may related to DFBSD have enhanced CAM DA driver to > separate read and write streams, allowing concurrent write completion in > the face of many stalled read requests. > It's released within DFBSD V2.10 at 2011: > http://leaf.dragonflybsd.org/mailarchive/commits/2011-04/msg00062.html > http://leaf.dragonflybsd.org/mailarchive/commits/2011-04/msg00061.html > I guess I'm alone for notice it. Have you verified w/ camcontrol tags that you have no remaining tags available? camcontrol tags <dev> -v will inform you of the current counts... If you do see dev_active == dev_openinings, then yes, it might be interested to explore this... I have a feeling that if your server is so heavily loaded that all tags are busy, you should probably improve your system w/ more spindles or better disks or ssds or something of the like... -- John-Mark Gurney Voice: +1 415 225 5579 "All that I will do, has been done, All that I have, has not."
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20141117193359.GP24601>