Re: COMMIT NOWAIT Performance Option
| От | Zeugswetter Andreas ADI SD |
|---|---|
| Тема | Re: COMMIT NOWAIT Performance Option |
| Дата | |
| Msg-id | E1539E0ED7043848906A8FF995BDA57901CAFD3F@m0143.s-mxs.net обсуждение исходный текст |
| Ответ на | Re: COMMIT NOWAIT Performance Option (Jeff Davis <pgsql@j-davis.com>) |
| Список | pgsql-hackers |
> > > > But we do don't we? fsync = off, full_page_writes = off? > > > > BTW, our testing seems to indicate that full_page_writes = off is safe > > on Solaris 10 on good hardware. At least, we haven't been able to break it yet. > > > > Is that an OS-dependent parameter? I always assumed it depended entirely > on hardware. I have no way to test it for myself though, so I just leave > full_page_writes=on to be safe. It also depends on the FS implementation. The OS/FS must guarantee, that it does not chunk single data page write calls. Usually that is the case, when OS/FS pagesize and pg pagesize are identical. And the HW needs to guarantee atomicity for single calls. e.g. on AIX you need to reduce the pg page size to 4k to be able to give those guarantees. Andreas
В списке pgsql-hackers по дате отправления: