Re: 2.4.14 prerelease call for testing

Jonathan Clarke wrote:

On 10.02.2009 13:52, Howard Chu wrote:

...

Works fine here...

Jonathan Clarke wrote:

...

I get this on two different boxes, and have observed similar behaviour with RE24 yesterday while testing on yet another box (not the test suite). If it works OK for you, let me know what data I can provide from my hanging slapd. GDB backtraces are below - dunno if they'll help.

In thread 3 frame 4 (syncrepl_updateCookie) print *si print *si->si_cookieState

Here it is: (also in text file attached, to avoid line wrapping)

Thanks. Looks like accessing a freed mutex, and valgrind confirms it. Will have to revisit this patch.