| From: | Claudio Freire <klaussfreire(at)gmail(dot)com> |
|---|---|
| To: | Vitaliy Garnashevich <vgarnashevich(at)gmail(dot)com> |
| Cc: | Gary Doades <gpd(at)gpdnet(dot)co(dot)uk>, Rick Otten <rottenwindfish(at)gmail(dot)com>, pgsql-performance(at)lists(dot)postgresql(dot)org |
| Subject: | Re: effective_io_concurrency on EBS/gp2 |
| Date: | 2018-01-31 19:34:18 |
| Message-ID: | CAGTBQpYCWStnzsjTrN+vBsEwZHH2PKwO7ybxp_JQv9CsEJj4xw@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-performance |
On Wed, Jan 31, 2018 at 1:57 PM, Vitaliy Garnashevich
<vgarnashevich(at)gmail(dot)com> wrote:
> More tests:
>
> io1, 100 GB:
>
> effective_io_concurrency=0
> Execution time: 40333.626 ms
> effective_io_concurrency=1
> Execution time: 163840.500 ms
In my experience playing with prefetch, e_i_c>0 interferes with kernel
read-ahead. What you've got there would make sense if what postgres
thinks will be random I/O ends up being sequential. With e_i_c=0, the
kernel will optimize the hell out of it, because it's a predictable
pattern. But with e_i_c=1, the kernel's optimization gets disabled but
postgres isn't reading much ahead, so you get the worst possible case.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Vitaliy Garnashevich | 2018-01-31 20:29:09 | Re: effective_io_concurrency on EBS/gp2 |
| Previous Message | Vitaliy Garnashevich | 2018-01-31 16:57:14 | Re: effective_io_concurrency on EBS/gp2 |