Re: Slow catchup of 2PC (twophase) transactions on replica in LR

On Thu, Jul 18, 2024 at 9:42 PM Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> wrote:
>
...
> I agree and have done that in the attached. I have made some
> additional changes: (a) removed the unrelated change of two_phase in
> protocol.sgml, (b) tried to make the two_phase change before failover
> option wherever it makes sense to keep the code consistent, (c)
> changed/added comments and doc changes at various places.
>
> I'll continue my review and testing of the patch but I thought of
> sharing what I have done till now.
>

Here some minor comments for patch v21

======
You wrote "tried to make the two_phase change before failover option
wherever it makes sense to keep the code consistent". But, still
failover is coded first in lots of places:
- libpqrcv_alter_slot
- ReplicationSlotAlter
- AlterReplicationSlot
etc.

Q. Why not change those ones?

======
src/backend/access/transam/twophase.c

IsTwoPhaseTransactionGidForSubid:
nitpick - nicer to rename the temporary gid variable: /gid_generated/gid_tmp/

======
src/backend/commands/subscriptioncmds.c

CheckAlterSubOption:
nitpick = function comment period/plural.
nitpick - typo /Samilar/Similar/

======
src/include/replication/slot.h

1.
-extern void ReplicationSlotAlter(const char *name, bool failover);
+extern void ReplicationSlotAlter(const char *name, bool *failover,
+ bool *two_phase);

Use const?

======
99.
Please see attached diffs implementing the nitpicks mentioned above

======
Kind Regards,
Peter Smith.
Fujitsu Australia