Re: PATCH: logical_work_mem and logical streaming of largein-progress transactions - Mailing list pgsql-hackers

On Mon, Apr 27, 2020 at 4:13 PM Amit Kapila <amit.kapila16@gmail.com> wrote:
>
> On Mon, Apr 27, 2020 at 4:05 PM Dilip Kumar <dilipbalaut@gmail.com> wrote:
> >
> > I have also fixed a couple of bugs internally reported by my colleague
> > Neha Sharma.
> >
>
> I think it would be good if you can briefly explain what were the bugs
> and how you fixed those?

Issue1:  If the concurrent transaction was aborted then in CATCH block
we were not freeing the memory of the toast_has, and it was causing
the assert that after the stream is complete txn->size != 0.

Issue2: After streaming is complete we set the txn->final_lsn and we
remember that in the local variable,  But mistakenly it was remembered
in local TRY block variable so if there is a concurrent abort in the
CATCH block the variable value is always a zero.  So after streaming
the final_lsn were becoming 0 and that was asserting.

-- 
Regards,
Dilip Kumar
EnterpriseDB: http://www.enterprisedb.com