Hi,

Thanks. I will try this approach.

Could you also please state the reason why is it happening in case of large objects? Because concurrent transactions are very well handled for other data types, but the same is not happening for lobs. Is it because the fomer are stored in toast table and there is no support for concurrent txns in pg_largeobject table?

On 13-Dec-2019 5:25 PM, Daniel Verite <daniel@manitou-mail.org> wrote:

Shalini wrote:

> > Is there a workaround to this concurrency issue without creating a
> > new large object?

The transaction failing with the "Tuple concurrently updated"
error could be resubmitted by the client, as if it was a
serialization failure.
Or the failure could be prevented by using advisory locks:
https://www.postgresql.org/docs/current/explicit-locking.html

Best regards,
--
Daniel Vérité
PostgreSQL-powered mailer: http://www.manitou-mail.org
Twitter: @DanielVerite