Bug #1721 | False duplicate key warnings on FULLTEXT columns | ||
---|---|---|---|
Submitted: | 31 Oct 2003 0:17 | Modified: | 23 Feb 2004 10:45 |
Reporter: | Tony Butcher | Email Updates: | |
Status: | No Feedback | Impact on me: | |
Category: | MySQL Server: MyISAM storage engine | Severity: | S2 (Serious) |
Version: | 4.0.16 | OS: | Linux (Linux SuSE 8.2) |
Assigned to: | Assigned Account | CPU Architecture: | Any |
[31 Oct 2003 0:17]
Tony Butcher
[31 Oct 2003 0:20]
Tony Butcher
Corrected version to 4.0.16
[31 Oct 2003 11:15]
Dean Ellis
I am unable to duplicate this behaviour using 4.0.16 or 4.0.17. I am able to duplicate as many binary values for `body` as I like, as one would expect with FULLTEXT indexes not enforcing uniqueness. If you can duplicate this with a new table and a specific 'binary' created for this test, please submit a test case with complete CREATE TABLE and INSERT statements for further testing. Thank you
[12 Nov 2003 1:13]
Sergei Golubchik
Looking at the code I can see that it is could be possible to get "duplicate entry" error for fulltext key. My guess is that it is indeed related to binary data and character set issues - which could result in some weird comparison problems if done incorrectly. In any case, I need a repeatable test case to fix the bug. Next time you'll get this error, please, try to save the row data that caused the error - I think that even inserting this row into an *empty* table (with identical structure) can cause "duplicate entry" error.
[12 Nov 2003 2:35]
Tony Butcher
This error has not recurred (although binaries are still being stored), and I've not had the opportunity to find the offending spam/virus. However, I think it's important to note that I am still getting table corruptions (about 1 per day) in my 'emails' and 'mail_cache' table, which have single FULLTEXT columns. (We even switched machines and this still happens). I will try to find a specific offender if I can, although when the corrupting INSERT happens it succeeds (with a corruption), while the duplicate key error actually fails (so is easier to trap).
[23 Feb 2004 10:45]
Sergei Golubchik
No feedback was provided. The bug is being suspended because we assume that you are no longer experiencing the problem. If this is not the case and you are able to provide the information that was requested earlier, please do so and change the status of the bug back to "Open". Thank you.