[Date Prev][Date Next]
[Chronological]
[Thread]
[Top]
Re: Fwd: 2.4.40 memory leak?
On Fri, Apr 17, 2015 at 05:43:45PM +0300, Sergey Esin wrote:
It's still happening, see http://i.imgur.com/NL8ztmp.png. The only solution
for us now is to reboot slapd on a regular basis.
What information can I provide to help to find the reason and fix it?
If you can provide a reproducible test case, and verify that it still
happens with recent code from git (branch OPENLDAP_REL_ENG_2_4), you
should file an ITS.
http://openldap.org/its
As glibc's allocator is known to suffer from fragmentation issues, I'd
suggest mentioning which alternate allocators you've tried and what the
results were. (Although it may still be closed with the rationale that
memory fragmentation is not slapd's problem...)
The symptoms of ITS#7299 are similar. It might be the same problem, or
at least related.
In my experience, this particular problem does not affect delta-syncrepl
(changelog-based replication). Have you tried setting that up?
http://www.openldap.org/doc/admin24/replication.html#Delta-syncrepl