Re: (ITS#4108) LWL causing core dumps in HEAD

[hyc@symas.com]

quanah@stanford.edu wrote:
> --On Friday, October 28, 2005 3:09 PM -0700 Howard Chu <hyc@symas.com> 
> wrote:
>
>   
>> Howard Chu wrote:
>>     
>>> Using this slight modification of Jason's bindstress patch in ITS#3855
>>> I can reproduce the crash quite rapidly. On my system (Linux x86_64)
>>> it tends to crash much sooner using epoll than using select, but in
>>> both cases it is the same assert failing.
>>>
>>>       
>> It appears that the problem existed before. I was able to reproduce the
>> same crash against 2.3.11 after running the bindstress script for several
>> minutes. I.e., it doesn't happen as quickly in 2.3.11, but it eventually
>> hits the same assert.
>>     
>
> You are filling me with cheer... :P
>
> I think we need a 2.3.12 release soon anyway. ;)

So it turns out there were a few race conditions and cases of trying to 
close the same connection twice. (Which was bad news, because the 
descriptor could get reused immediately after the first close.)
I believe this is now fixed in HEAD. Please test. Heavily.

-- 
  -- Howard Chu
  Chief Architect, Symas Corp.  http://www.symas.com
  Director, Highland Sun        http://highlandsun.com/hyc
  OpenLDAP Core Team            http://www.openldap.org/project/