Re: [Proposal] Fully WAL logged CREATE DATABASE - No Checkpoints
| От | Dilip Kumar |
|---|---|
| Тема | Re: [Proposal] Fully WAL logged CREATE DATABASE - No Checkpoints |
| Дата | |
| Msg-id | CAFiTN-u4JMq2=XkgmNhPLgcZMZODBzp8vPrJ2ff1A1MP-hkU=g@mail.gmail.com обсуждение исходный текст |
| Ответ на | Re: [Proposal] Fully WAL logged CREATE DATABASE - No Checkpoints (Robert Haas <robertmhaas@gmail.com>) |
| Список | pgsql-hackers |
On Tue, Mar 22, 2022 at 8:53 PM Robert Haas <robertmhaas@gmail.com> wrote: > > On Tue, Mar 22, 2022 at 5:00 AM Dilip Kumar <dilipbalaut@gmail.com> wrote: > > In my previous patch mistakenly I used src_dboid instead of > > dest_dboid. Fixed in this version. For destination db I have used > > lock mode as AccessSharedLock. Logically if we see access wise we > > don't want anyone else to be accessing that db but that is anyway > > protected because it is not visible to anyone else. So I think > > AccessSharedLock should be correct here because we are just taking > > this lock because we are accessing pages in shared buffers from this > > database's relations. > > Here's my worked-over version of your previous patch. I haven't tried > to incorporate your incremental patch that you just posted. Thanks for working on the comments. Please find the updated version which include below changes - Worked on the XXX comments added by you. - Added database level lock for the target database as well. - Used fake relcache and removed direct access to the smgr, I think it was not really necessary in ScanSourceDatabasePgClass() because we are using it for a very short period of time but still I have changed it, let me know if you think that it is unneccessary to create the fake relcache here. - Removed extra space in createdb.c and fixed test case. -- Regards, Dilip Kumar EnterpriseDB: http://www.enterprisedb.com
Вложения
В списке pgsql-hackers по дате отправления: