| From: | Greg Stark <gsstark(at)mit(dot)edu> |
|---|---|
| To: | Scott Mead <scott(dot)lists(at)enterprisedb(dot)com> |
| Cc: | David Fetter <david(at)fetter(dot)org>, "Prasad, Venkat" <venkat(dot)prasad(at)credit-suisse(dot)com>, PostgreSQL General <pgsql-general(at)postgresql(dot)org> |
| Subject: | Re: [BUGS] Integrity check |
| Date: | 2009-06-25 13:01:42 |
| Message-ID: | 407d949e0906250601j695d169frb1e7882d34392b73@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-bugs pgsql-general |
On Wed, Jun 24, 2009 at 11:45 PM, Scott
Mead<scott(dot)lists(at)enterprisedb(dot)com> wrote:
>
> If you're asking "Does the database have the ability to verify that
> whatever is in a block is what was put into that block", then Oracle has
> block check-summing (I'm not sure if this got into PG 8.4 or not...)
It didn't, due to technical issues.
However you should be aware that checksumming doesn't necessarily
provide a broad guarantee either. It can only protect against damage
that occurs between the time the checksum is generated and when it's
verified. If you have bad memory, for example, it can corrupt the data
before the checksum is calculated, so the block may be stored happily
with a valid checksum for the corrupt data.
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Jeremy Ford | 2009-06-25 13:32:27 | BUG #4880: gcc (4.4) + double precision differences |
| Previous Message | Fujii Masao | 2009-06-25 12:55:07 | BUG #4879: bgwriter fails to fsync the file in recovery mode |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Frank Heikens | 2009-06-25 13:01:52 | Re: create a table inside a function |
| Previous Message | Chris Barnes | 2009-06-25 13:00:29 | Problems with postgres online backup - restore |