Обсуждение: pgbasebackup necessary for master slave change?

Поиск
Список
Период
Сортировка

pgbasebackup necessary for master slave change?

От
Subhankar Chattopadhyay
Дата:
Hi,

Few days back I had asked if it is needed to to pg_basebackup for
every database update. From John I understood that it is unnecessary
and if the slave is syncing, even if it is catching up, it would be
able to sync without doing pg_basebackup. This is working also for me.

However, for a failover scenario, where a master goes down, and I make
the slave as master, and then when the old master comes back as a
slave again, if I don't take pg_basebackup from the new master, it
cannot follow the new master. This is kind of an overhead. Is there a
way I can make the old master follow the new master without having to
do full backup?



Subhankar Chattopadhyay
Bangalore, India


Re: pgbasebackup necessary for master slave change?

От
Chris Travers
Дата:


On Thu, Dec 8, 2016 at 10:33 AM, Subhankar Chattopadhyay <subho.atg@gmail.com> wrote:
Hi,

Few days back I had asked if it is needed to to pg_basebackup for
every database update. From John I understood that it is unnecessary
and if the slave is syncing, even if it is catching up, it would be
able to sync without doing pg_basebackup. This is working also for me.

However, for a failover scenario, where a master goes down, and I make
the slave as master, and then when the old master comes back as a
slave again, if I don't take pg_basebackup from the new master, it
cannot follow the new master. This is kind of an overhead. Is there a
way I can make the old master follow the new master without having to
do full backup?

Depending on your version and circumstance, pg_rewind may address your problem.
https://www.postgresql.org/docs/9.5/static/app-pgrewind.html 



Subhankar Chattopadhyay
Bangalore, India


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



--
Best Wishes,
Chris Travers

Efficito:  Hosted Accounting and ERP.  Robust and Flexible.  No vendor lock-in.

Re: pgbasebackup necessary for master slave change?

От
Subhankar Chattopadhyay
Дата:
I am on 9.4. Doesn't look like we have it available on 9.4.

Do we have any other option?

On Thu, Dec 8, 2016 at 3:15 PM, Chris Travers <chris.travers@gmail.com> wrote:
>
>
> On Thu, Dec 8, 2016 at 10:33 AM, Subhankar Chattopadhyay
> <subho.atg@gmail.com> wrote:
>>
>> Hi,
>>
>> Few days back I had asked if it is needed to to pg_basebackup for
>> every database update. From John I understood that it is unnecessary
>> and if the slave is syncing, even if it is catching up, it would be
>> able to sync without doing pg_basebackup. This is working also for me.
>>
>> However, for a failover scenario, where a master goes down, and I make
>> the slave as master, and then when the old master comes back as a
>> slave again, if I don't take pg_basebackup from the new master, it
>> cannot follow the new master. This is kind of an overhead. Is there a
>> way I can make the old master follow the new master without having to
>> do full backup?
>
>
> Depending on your version and circumstance, pg_rewind may address your
> problem.
> https://www.postgresql.org/docs/9.5/static/app-pgrewind.html
>>
>>
>>
>>
>> Subhankar Chattopadhyay
>> Bangalore, India
>>
>>
>> --
>> Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
>> To make changes to your subscription:
>> http://www.postgresql.org/mailpref/pgsql-general
>
>
>
>
> --
> Best Wishes,
> Chris Travers
>
> Efficito:  Hosted Accounting and ERP.  Robust and Flexible.  No vendor
> lock-in.
> http://www.efficito.com/learn_more



--




Subhankar Chattopadhyay
Bangalore, India


Re: pgbasebackup necessary for master slave change?

От
Achilleas Mantzios
Дата:
On 08/12/2016 11:33, Subhankar Chattopadhyay wrote:
> Hi,
>
> Few days back I had asked if it is needed to to pg_basebackup for
> every database update. From John I understood that it is unnecessary
> and if the slave is syncing, even if it is catching up, it would be
> able to sync without doing pg_basebackup. This is working also for me.
>
> However, for a failover scenario, where a master goes down, and I make
> the slave as master, and then when the old master comes back as a
> slave again, if I don't take pg_basebackup from the new master, it
> cannot follow the new master. This is kind of an overhead. Is there a
> way I can make the old master follow the new master without having to
> do full backup?
pg_rewind
>
>
>
> Subhankar Chattopadhyay
> Bangalore, India
>
>


--
Achilleas Mantzios
IT DEV Lead
IT DEPT
Dynacom Tankers Mgmt



Re: pgbasebackup necessary for master slave change?

От
Subhankar Chattopadhyay
Дата:
pg_rewind cannot be used as I am on 9.4.

Anything else?

On Thu, Dec 8, 2016 at 3:23 PM, Achilleas Mantzios
<achill@matrix.gatewaynet.com> wrote:
> On 08/12/2016 11:33, Subhankar Chattopadhyay wrote:
>>
>> Hi,
>>
>> Few days back I had asked if it is needed to to pg_basebackup for
>> every database update. From John I understood that it is unnecessary
>> and if the slave is syncing, even if it is catching up, it would be
>> able to sync without doing pg_basebackup. This is working also for me.
>>
>> However, for a failover scenario, where a master goes down, and I make
>> the slave as master, and then when the old master comes back as a
>> slave again, if I don't take pg_basebackup from the new master, it
>> cannot follow the new master. This is kind of an overhead. Is there a
>> way I can make the old master follow the new master without having to
>> do full backup?
>
> pg_rewind
>>
>>
>>
>>
>> Subhankar Chattopadhyay
>> Bangalore, India
>>
>>
>
>
> --
> Achilleas Mantzios
> IT DEV Lead
> IT DEPT
> Dynacom Tankers Mgmt
>
>
>
>
> --
> Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-general



--




Subhankar Chattopadhyay
Bangalore, India


Re: pgbasebackup necessary for master slave change?

От
Chris Travers
Дата:


On Thu, Dec 8, 2016 at 11:01 AM, Subhankar Chattopadhyay <subho.atg@gmail.com> wrote:
pg_rewind cannot be used as I am on 9.4.

Anything else?

Upgrade to 9.5 or 9.6? ;-)

This is a known limitation in Postgres that pg_rewind was written to address. 

On Thu, Dec 8, 2016 at 3:23 PM, Achilleas Mantzios
<achill@matrix.gatewaynet.com> wrote:
> On 08/12/2016 11:33, Subhankar Chattopadhyay wrote:
>>
>> Hi,
>>
>> Few days back I had asked if it is needed to to pg_basebackup for
>> every database update. From John I understood that it is unnecessary
>> and if the slave is syncing, even if it is catching up, it would be
>> able to sync without doing pg_basebackup. This is working also for me.
>>
>> However, for a failover scenario, where a master goes down, and I make
>> the slave as master, and then when the old master comes back as a
>> slave again, if I don't take pg_basebackup from the new master, it
>> cannot follow the new master. This is kind of an overhead. Is there a
>> way I can make the old master follow the new master without having to
>> do full backup?
>
> pg_rewind
>>
>>
>>
>>
>> Subhankar Chattopadhyay
>> Bangalore, India
>>
>>
>
>
> --
> Achilleas Mantzios
> IT DEV Lead
> IT DEPT
> Dynacom Tankers Mgmt
>
>
>
>
> --
> Sent via pgsql-general mailing list (pgsql-general@postgresql.org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-general



--




Subhankar Chattopadhyay
Bangalore, India


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



--
Best Wishes,
Chris Travers

Efficito:  Hosted Accounting and ERP.  Robust and Flexible.  No vendor lock-in.

Re: pgbasebackup necessary for master slave change?

От
Michael Paquier
Дата:
On Thu, Dec 8, 2016 at 7:08 PM, Chris Travers <chris.travers@gmail.com> wrote:
>
>
> On Thu, Dec 8, 2016 at 11:01 AM, Subhankar Chattopadhyay
> <subho.atg@gmail.com> wrote:
>>
>> pg_rewind cannot be used as I am on 9.4.
>>
>> Anything else?
>
>
> Upgrade to 9.5 or 9.6? ;-)
>
> This is a known limitation in Postgres that pg_rewind was written to
> address.

FWIW I maintain a version that works on 9.4:
https://github.com/vmware/pg_rewind
And that's battle-proven.
--
Michael