| Bug #13905 | OS limit on number of files breaks documented "raid_chunks" limits | ||
|---|---|---|---|
| Submitted: | 10 Oct 2005 18:21 | Modified: | 12 Oct 2005 16:18 |
| Reporter: | Joerg Bruehe | Email Updates: | |
| Status: | Won't fix | Impact on me: | |
| Category: | MySQL Server | Severity: | S3 (Non-critical) |
| Version: | 4.1.15 (likely: all 4.x) | OS: | Other (SCO 3.2 5.0.7 i386) |
| Assigned to: | CPU Architecture: | Any | |
Portions of this website are copyright © 2001, 2002 The PHP Group
Page generated in 0.053 sec. using MySQL 8.0.36-u4-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.
© 2024, Oracle Corporation and/or its affiliates
Description: Release build of 4.1.15, tested on machine "openserv". Test "raid" fails with this symptom: === cut === raid [ fail ] Errors are (from PATH/mysqltest-time) : mysqltest: At line NNN: query 'create table test_$1.r2 (i int) raid_type=1 raid_chunks=257' failed: 1: Can't create/write to file './test_$1/b0/r2.MYD' (Errcode: 24) (the last lines may be the most important ones) Ending Tests === cut === Testing with varying values for "raid_chunks" works for values up to and including 83, for 84 it fails in dropping the database, for 88 it fails as shown above in the original test; I did not check intermediate values. Error number 24 means "EMFILE", which means on "open()": OPEN_MAX descriptors are open in the current process. I could not find a setting for "OPEN_MAX", it seems to be configurable. As a consequence of this, the created database "test_$1" is not dropped, so it will be listed in the subsequent tests "rpl000009", "rpl_create_database", and "show_check" and so make them fail. How to repeat: Test suite failure. Suggested fix: 1) Change the manual to mention the (possible) OS limit. 2) Add a cleanup function after a failed test.