Hi Thomas! On Thu, Nov 21, 2024 at 2:38 PM Thomas Munro <thomas.mu...@gmail.com> wrote:
> On Thu, Nov 21, 2024 at 11:44 PM Jakub Wartak > <jakub.war...@enterprisedb.com> wrote: > > This literally looks like something like off_t/size_t would be limited > to 2^31 somewhere. > > off_t is 32 bits on Windows. I'd be quite suspicious of the > arithmetic involving 'currpos'. What happens if you change all the > off_t in walmethods.c/.h to pgoff_t? The lseek() is suspicious too, > and might need to be redirected to _lseeki64(). > Thank you for responding and suggesting this. First one was not enough, on REL_15_STABLE with just pgoff_t i still got the same error: C:\compiledpg\bin>pg_basebackup.exe -U postgres -D "c:\backup6" -F t -P -X stream -c fast --compress=none --create-slot --slot=slot10 Password: pg_basebackup: error: could not close file "0000000100000008000000AE": Invalid argument pg_basebackup: error: background process terminated unexpectedly but with attached _lseeki64 dirty patch I've got success and resulting tarball greater than 2^31 too: C:\compiledpg\bin>pg_basebackup.exe -U postgres -D "c:\backup7" -F t -P -X stream -c fast --compress=none --create-slot --slot=slot11 Password: 18134035/18134035 kB (100%), 1/1 tablespace C:\compiledpg\bin> C:\compiledpg\bin>dir c:\backup7\pg_wal.tar [..] 11/22/2024 10:37 AM 4,026,778,112 pg_wal.tar PoC patch is attached and I'll register CFentry to see how the tests will go (all of this MSVC/MinGW stuff is frightening to me, of course this will fail on non-Win32...). -J.
v1-0001-Bugfix-Windows-pg_basebackup-ability-to-create-2G.patch
Description: Binary data
