Hi, On 2022-08-04 18:05:25 -0400, Tom Lane wrote: > Andres Freund <and...@anarazel.de> writes: > > Yea. I think at the very least we need to start holding interrupts before > > the > > DropDatabaseBuffers() and do so until commit. That's probably best done by > > doing the transaction commit inside dropdb. > > We've talked before about ignoring interrupts across commit, but > I find the idea a bit scary.
I'm not actually suggesting to do so across commit, just until the commit. Maintaining that seems easiest if dropdb() does the commit internally. > In any case, DROP DATABASE is far from the only place with a problem. What other place has a database corrupting potential of this magnitude just because interrupts are accepted? We throw valid s_b contents away and then accept interrupts before committing - with predictable results. We also accept interrupts as part of deleting the db data dir (due to catalog access). Greetings, Andres Freund
