On Mon, Apr 22, 2024 at 2:31 PM Bertrand Drouvot
wrote:
>
> On Mon, Apr 22, 2024 at 11:32:14AM +0530, shveta malik wrote:
> > On Mon, Apr 22, 2024 at 5:57 AM Zhijie Hou (Fujitsu)
> > wrote:
> > > Attach the V3 patch which also addressed Shveta[1] and Bertrand[2]'s
> > > comments.
>
> Thanks!
>
Hi,
On Mon, Apr 22, 2024 at 11:32:14AM +0530, shveta malik wrote:
> On Mon, Apr 22, 2024 at 5:57 AM Zhijie Hou (Fujitsu)
> wrote:
> > Attach the V3 patch which also addressed Shveta[1] and Bertrand[2]'s
> > comments.
Thanks!
> Tested the patch, works well.
Same here.
Regards,
--
Bertrand
On Mon, Apr 22, 2024 at 5:57 AM Zhijie Hou (Fujitsu)
wrote:
>
> On Friday, April 19, 2024 10:54 AM Kuroda, Hayato/黒田 隼人
> wrote:
> > In your patch, the pg_dump.c was updated. IIUC the main reason was that
> > pg_restore executes some queries as single transactions so that ALTER
> > SUBSCRIPTION
On Friday, April 19, 2024 10:54 AM Kuroda, Hayato/黒田 隼人
wrote:
> In your patch, the pg_dump.c was updated. IIUC the main reason was that
> pg_restore executes some queries as single transactions so that ALTER
> SUBSCRIPTION cannot be done, right?
Yes, and please see below for other reasons.
>
On Fri, Apr 19, 2024 at 6:09 AM Zhijie Hou (Fujitsu)
wrote:
>
> Here is V2 patch which includes the changes for CREATE SUBSCRIPTION as
> suggested. Since we don't connect pub to alter slot when (create_slot=false)
> anymore, the restriction that disallows failover=true when connect=false is
>
Hi,
On Fri, Apr 19, 2024 at 12:39:40AM +, Zhijie Hou (Fujitsu) wrote:
> Here is V2 patch which includes the changes for CREATE SUBSCRIPTION as
> suggested. Since we don't connect pub to alter slot when (create_slot=false)
> anymore, the restriction that disallows failover=true when
Dear Hou,
Thanks for updating the patch! Let me confirm the content.
In your patch, the pg_dump.c was updated. IIUC the main reason was that
pg_restore executes some queries as single transactions so that ALTER
SUBSCRIPTION cannot be done, right?
Also, failover was synchronized only when we were
Dear Shveta,
> When I said that option 2 aligns with ALTER-SUB documented behaviour,
> I meant the doc described in [1] stating "When altering the slot_name,
> the failover and two_phase property values of the named slot may
> differ from the counterpart failover and two_phase parameters
>
On Thursday, April 18, 2024 1:52 PM Amit Kapila wrote:
>
> On Tue, Apr 16, 2024 at 5:06 PM shveta malik
> wrote:
> >
> > On Tue, Apr 16, 2024 at 1:45 PM Hayato Kuroda (Fujitsu)
> > wrote:
> > >
> > > Dear Hou,
> > >
> > > > Kuroda-San reported an issue off-list that:
> > > >
> > > > If user
On Thu, Apr 18, 2024 at 11:22:24AM +0530, Amit Kapila wrote:
> +1 for option 2 as it sounds logical to me and consistent with ALTER
> SUBSCRIPTION. BTW, IIUC, you are referring to: "When altering the
> slot_name, the failover and two_phase property values of the named
> slot may differ from the
On Thu, Apr 18, 2024 at 11:40 AM Hayato Kuroda (Fujitsu)
wrote:
>
> Dear Shveta,
>
> Sorry for delay response. I missed your post.
>
> > +1.
> >
> > Similar to ALTER SUB, CREATE SUB also needs to be fixed. This is
> > because we call alter_replication_slot in CREATE SUB as well, for the
> > case
Dear Shveta,
Sorry for delay response. I missed your post.
> +1.
>
> Similar to ALTER SUB, CREATE SUB also needs to be fixed. This is
> because we call alter_replication_slot in CREATE SUB as well, for the
> case when slot_name is provided and create_slot=false. But the tricky
> part is we call
On Thu, Apr 18, 2024 at 11:22 AM Amit Kapila wrote:
>
> On Tue, Apr 16, 2024 at 5:06 PM shveta malik wrote:
> >
> > On Tue, Apr 16, 2024 at 1:45 PM Hayato Kuroda (Fujitsu)
> > wrote:
> > >
> > > Dear Hou,
> > >
> > > > Kuroda-San reported an issue off-list that:
> > > >
> > > > If user execute
On Tue, Apr 16, 2024 at 5:06 PM shveta malik wrote:
>
> On Tue, Apr 16, 2024 at 1:45 PM Hayato Kuroda (Fujitsu)
> wrote:
> >
> > Dear Hou,
> >
> > > Kuroda-San reported an issue off-list that:
> > >
> > > If user execute ALTER SUBSCRIPTION SET (failover) command inside a txn
> > > block
> > >
On Tue, Apr 16, 2024 at 1:45 PM Hayato Kuroda (Fujitsu)
wrote:
>
> Dear Hou,
>
> > Kuroda-San reported an issue off-list that:
> >
> > If user execute ALTER SUBSCRIPTION SET (failover) command inside a txn block
> > and rollback, only the subscription option change can be rolled back, while
> >
Dear Hou,
> Kuroda-San reported an issue off-list that:
>
> If user execute ALTER SUBSCRIPTION SET (failover) command inside a txn block
> and rollback, only the subscription option change can be rolled back, while
> the
> replication slot's failover change is preserved.
>
> This is because
Hi,
On Tue, Apr 16, 2024 at 06:32:11AM +, Zhijie Hou (Fujitsu) wrote:
> Hi,
>
> Kuroda-San reported an issue off-list that:
>
> If user execute ALTER SUBSCRIPTION SET (failover) command inside a txn block
> and rollback, only the subscription option change can be rolled back, while
> the
>
Hi,
Kuroda-San reported an issue off-list that:
If user execute ALTER SUBSCRIPTION SET (failover) command inside a txn block
and rollback, only the subscription option change can be rolled back, while the
replication slot's failover change is preserved.
This is because ALTER SUBSCRIPTION SET
18 matches
Mail list logo
