Re: Load Distributed Checkpoints, revised patch

Heikki Linnakangas <heikki(at)enterprisedb(dot)com> wrote:

> Michael Paesold wrote:
> > Why do you think so? Is it too much risk to adapt the sorted writes? The
> > numbers shown by ITAGAKI Takahiro looked quite impressive, at least for
> > large shared_buffers configurations. The reactions where rather
> > positive, too.
>
> Well, it is a very recent idea, and it's not clear that it's a win under
> all circumstances. Adopting that would need more testing, and at this
> late stage I'd like to just wrap up what we have and come back to this
> idea for 8.4.

I'm planning it for 8.4. Through the discussion of LDC and the sorted
writes, we found that kernels are not so clever as we don't need to
schedule I/O. At least, we'd better to give them more hints.

Many self I/O management methods have been often proposed; read-ahead,
aio, bulk extension of files and so on. The sorted writes is one of them.
We should test them in various workloads and take what is good.

Regards,
---
ITAGAKI Takahiro
NTT Open Source Software Center