MySQL Bugs: #32733: ndb_binlog_format.test fails randomly

Bug #32733	ndb_binlog_format.test fails randomly
Submitted:	26 Nov 2007 18:50	Modified:	30 Nov 2007 19:48
Reporter:	Ingo Strüwing	Email Updates:
Status:	Closed	Impact on me:	None
Category:	Tests: Cluster	Severity:	S7 (Test Cases)
Version:	6.0.4	OS:	Linux (Debian)
Assigned to:		CPU Architecture:	Any
Tags:	pbfail

Description:
ndb.ndb_binlog_format          [ fail ]

mysqltest: In included file "./include/have_multi_ndb.inc": At line 21: query 'drop table if exists t1, t2' failed: 2013: Lost connection to MySQL server during query

Stopping All Servers

Disabling the test case. Please re-enable after fix.

How to repeat:
See pushbuild, for example https://intranet.mysql.com/secure/pushbuild/showdir.pl?dir=mysql-6.0-engines
log https://intranet.mysql.com/secure/pushbuild/getlog.pl?dir=mysql-6.0-engines&entry=istruewi...

Pushed into 6.0.4-alpha

I don't think that this is fair. Investigation, if the bug is reported already, and filling the form takes some minutes already. Alone downloading of the archive would take much more than 5 minutes. Analyzing the problem in so far that one knows which files to add, producing a stack backtrace, and so on, belongs more to bug fixing than to bug reporting. You cannot publish a list of files to save for all kinds of bugs?

If we want to put that burden on the bug reporter, those bugs won't be reported at all. One would just hope that it won't happen in the next run. I believe this is what most developers do anyway, as even a terse report like the above is sufficiently inconvenient to make.

A different story is the fact that the archive has gone already (2.5 days after the push). This shows that the development resources in terms of disk space are much too small. (Didn't I mention that somewhere before?) I wanted to measure the time required to provide you with the requested information, but now I cannot do that. If the given information does not help, then this bug has escaped. The log file is still present, though the link is broken by BugDB. But it doesn't have more information than pasted above anyway.

In my point of view, test suite failures have to be fixed with the topmost priority as they block other developers. But I agree that having less than 2 days, before the archive vanishes, is impossible.

Regards
Ingo