Re: Moving Database Cluster to another drive

OMG! That is a reference I should have found and read before,
Many thanks Jan

-----Original Message-----
From: pgsql-novice-owner@postgresql.org
[mailto:pgsql-novice-owner@postgresql.org] On Behalf Of Jan Lentfer
Sent: May-18-15 12:58
To: Daniel Begin
Cc: Luca Ferrari; <pgsql-novice@postgresql.org>
Subject: Re: [NOVICE] Moving Database Cluster to another drive

Well, it points to page in a data file, so copying the data files does not
do any harm (when Postgres is NOT running, otherwise - disaster). We are
talking about files on a filesystem here, not raw devices as e.g. Informix
uses (can use).

You might find this helpful in general, especiall around ppage 50.

https://momjian.us/main/writings/pgsql/internalpics.pdf

Jan

Von meinem iPad gesendet

> Am 18.05.2015 um 18:28 schrieb Daniel Begin <jfd553@hotmail.com>:
>
> Thank Luca,
>
> You wrote that "each index has a pointer to the block on disk that
contains the tuple". If there is no mechanism that insure the tuples have
the same location on the new drive (the block address), I should then expect
a plane copy will corrupt some/all the indexes, am I right (linked to the
second question)?
>
> Best regards,
> Daniel
>
> -----Original Message-----
> From: pgsql-novice-owner@postgresql.org
> [mailto:pgsql-novice-owner@postgresql.org] On Behalf Of Luca Ferrari
> Sent: May-18-15 02:27
> To: Daniel Begin
> Cc: pgsql-novice@postgresql.org
> Subject: Re: [NOVICE] Moving Database Cluster to another drive
>
>> On Sat, May 16, 2015 at 7:39 PM, Daniel Begin <jfd553@hotmail.com> wrote:
>> - How indexes are actually implemented in PostgreSQL (how pgsql point
>> to a record from an index)?
>
> Not an easy answer. An index can have different implementation types, but
for short each index has a pointer to the block on disk that contain the
tuple. And indexes are, of course, on disk stuff.
>
>> - Could copying tables and indexes have had an effect on indexes?
>
> Should not, but it could be.
>
>> - How can I verify that some of the indexes were not corrupted?
>
> Use explain, see pg_stat_user_indexes and do a reindex if you believe an
index is corrupted.
>
> I would expect this being more likely an issue with the hard drive (e.g.,
different seek times from the previous one).
>
> Luca
>
>
> --
> Sent via pgsql-novice mailing list (pgsql-novice@postgresql.org) To make
changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-novice
>
>
>
> --
> Sent via pgsql-novice mailing list (pgsql-novice@postgresql.org) To
> make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-novice


--
Sent via pgsql-novice mailing list (pgsql-novice@postgresql.org) To make
changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-novice