Re: fsync-pgdata-on-recovery tries to write to more files than previously

Michael Paquier <michael(dot)paquier(at)gmail(dot)com> writes:
> On Fri, May 29, 2015 at 6:49 PM, Amit Kapila <amit(dot)kapila16(at)gmail(dot)com> wrote:
>> Test - 2 - Directory Symlink for pg_xlog
>> First 4 steps are same as Test-1
>> 5. mklink /D E:\ PGData\pg_xlog E:\TempLog\xlog_symlink_loc
>> 6. Restart Server - Error
>> - FATAL: required WAL directory "pg_xlog" does not exist
>> This error occurs in
>> ValidateXLOGDirectoryStructure()->stat(XLOGDIR, &stat_buf)
>> 7. If I manually step-through that line, it proceeds and in
>> SyncDataDirectory(), it detects the symlink;
>> 8. After that in SyncDataDirectory(), when it does walkdir for
>> datadir, for ./pg_xlog it reports the log message and then
>> proceeds normal and the server is started.

> Regarding this test, I just tested initdb -X and it works correctly
> after crashing the server, so there is at least a workaround for the
> error you are triggering here. Now, shouldn't we improve things as
> well with mklink? That's a rather narrow case and we have a workaround
> for it at initialization with initdb. I haven't looked much at the
> code so I don't have a patch at hand but thoughts on this matter are
> welcome.

My (vague) recollection is that what mklink creates is not the same
as a junction point and we intentionally only support the latter as
pseudo-symlinks. You'd need to trawl the archives for more detail,
unless someone else remembers.

So I think we're good here. Thanks for all the testing!

regards, tom lane