You are right, we added more than one
coordinator nodes for high availability and to avoid single point of failure.
Thanks
Swathi
On Tue, Jun 1, 2021, 3:54 PM Bharath Rupireddy <bharath.rupireddyforpostgres@xxxxxxxxx> wrote:
On Tue, Jun 1, 2021 at 3:31 PM Etsuro Fujita <etsuro.fujita@xxxxxxxxx> wrote:
>
> Hi,
>
> On Fri, May 14, 2021 at 6:08 PM Swathi P <swathi.bluepearl@xxxxxxxxx> wrote:
> > In our sharding solution, we have multiple coodinator nodes. If we declare the table column as serial data type, we might end up having duplicate values for id column in the table_a in host_b (data node) as cconnections come from multiple coordinatoor nodes and might end up in duplicate key violations.
> >
> > Hence we decided to have the coordinator nodes as stateless and hence declared the column with no serial/sequence. Let me know if this makes sense.
>
> It seems reasonable to me to make coodinator nodes stateless, but may
> I ask the reason you use multiple coordinator nodes?
Perhaps, as a redundant node to avoid single point of failures? It's
just a guess as I'm not the right one to answer that question though.
With Regards,
Bharath Rupireddy.
