Bug #71479 sort_buffer_size is limited to 4GB-1 on 64-bit Windows
| Submitted: | 25 Jan 2014 15:32 | Modified: | 27 Jan 2014 15:14 |
|---|---|---|---|
| Reporter: | Valeriy Kravchuk | Email Updates: | |
| Status: | Verified | Impact on me: | |
| Category: | MySQL Server: Options | Severity: | S3 (Non-critical) |
| Version: | 5.6 | OS: | Windows (Win64) |
| Assigned to: | CPU Architecture: | Any | |
| Tags: | sort_buffer_size | ||
[25 Jan 2014 15:32]
Valeriy Kravchuk
[27 Jan 2014 10:10]
MySQL Verification Team
Hello Valeriy, Thank you for the bug report. Verified as described. Thanks, Umesh
[27 Jan 2014 11:57]
Shane Bester
but why does this limit apply to windows and not linux? shouldn't the limit rather be removed instead of document it?
[27 Jan 2014 15:14]
Valeriy Kravchuk
Good question, Shane... I suspect some (wrong?) data type is used for the related variable on Windows for some historical reason (like "nobody cared to ask"). I don't mind to see this re-classified as a server bug.
[27 Jan 2014 17:42]
Paul DuBois
Perhaps there are two issues here: 1) The off-by-one error in the description in the docs (4GB rather than 4GB-1). I can deal with this. 2) The capping of the value for Windows 64-bit, as opposed to other 64-bit systems. Similar capping applies to join_buffer_size, key_buffer_size, myisam_sort_buffer_size, where the limit was lifted for non-Windows 64-bit systems in 5.1.23. The limit on Windows was lifted for key_buffer_size in 5.1.31; not sure why it wouldn't have been lifted for the others. Maybe it could be, in which case this might better be considered a server bug, or feature request.
[30 Jan 2014 13:58]
Paul DuBois
Off-by-one aspect of the docs description has been fixed.
[30 Jan 2014 14:23]
Paul DuBois
Unassigning from myself; no longer a docs bug.
Portions of this website are copyright © 2001, 2002 The PHP Group
Page generated in 0.057 sec. using MySQL 8.4.7-u6-cloud
Timestamp references displayed by the system are UTC.
Content reproduced on this site is the property of the
respective copyright holders.
It is not reviewed in advance
by Oracle and does not necessarily represent the opinion of
Oracle or any other party.