Date: Sun, 16 Nov 2014 04:26:20 +0200 From: Alexander Motin <mav@FreeBSD.org> To: Garrett Cooper <yaneurabeya@gmail.com>, "Ranjan1018 ." <214748mv@gmail.com> Cc: FreeBSD CURRENT <freebsd-current@freebsd.org> Subject: Re: [PATCH] gstat on SSD Message-ID: <54680B4C.9020207@FreeBSD.org> In-Reply-To: <1263A474-E3B8-4942-A0AA-6692F509138F@gmail.com> References: <CACyC=qbssehwZrnrLoZZNTEWaTMTsp0S-zj_VkLdAP_-pd7Z=w@mail.gmail.com> <1263A474-E3B8-4942-A0AA-6692F509138F@gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On 15.11.2014 23:35, Garrett Cooper wrote: > On Nov 15, 2014, at 8:04, Ranjan1018 . <214748mv@gmail.com> wrote: >> I am running a pair of servers with SSD. In gstat the default >> display of operation/ms is only one digit after the comma and >> there are a lot of R/W operation displayed with the 0.1 value. >> This simple patch displays the operation/ms with two decimal >> digit. >> >> What’s next ? After testing this patch for a day, recording the >> values in a log file, I have found that some write operation are >> only 0.02 ms short, on a OCZ Vertex SSD. Probably the next patch >> is to display micro seconds instead of milli seconds. >> >> Maurizio >> >> FreeBSD 11 gstat.c.patch <http://pastebin.com/B7B8SEy9>; > > LGTM, but I don’t have hardware capable of that precision so I > can’t verify the patch other than build test it and run it for what > existing cases I have. > > What do you think Alexander? SSD in my laptop can to 20K simple case IOPS in one stream, that means ~50us time. So I think this patch should not harm. From the other side these numbers are mostly informational, since any real performance investigation would require a histogram rather the a single average number. -- Alexander Motin
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?54680B4C.9020207>