| From: | "Robert Haas" <robertmhaas(at)gmail(dot)com> |
|---|---|
| To: | "Gregory Stark" <stark(at)enterprisedb(dot)com> |
| Cc: | Postgres <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Updated posix fadvise patch v19 |
| Date: | 2008-11-18 17:07:31 |
| Message-ID: | 603c8f070811180907m55e41593v71243605585009a0@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
> One thing which is bothering me is that the guc assign hook is throwing an
> error if you set effective_io_concurrency when your system's posix_fadvise is
> deemed inadequate (either unavailable or from an old version of glibc). I'm
> starting to think it shouldn't throw an error, just not set the internal
> variable and possible output a warning. We do have some GUC variables which
> throw errors if you use them and support isn't compiled in, but I'm not sure
> it's such a hot idea even for those.
I can't see why this would be a good idea. Warnings are easy to
overlook, and then you have completely different behavior without
knowing it. If I wanted a database that silently did something
completely different from what I asked it to do, I'd use... well,
let's just say products like that are available.
...Robert
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Sam Mason | 2008-11-18 17:09:56 | Re: is any reason why only one columns subselect are allowed in array()? |
| Previous Message | KaiGai Kohei | 2008-11-18 16:42:38 | Re: Updates of SE-PostgreSQL 8.4devel patches (r1197) |