[Date Prev][Date Next]
Re: slapd 2.0.7 and solaris 8
I appears that the leak is coming from the regcomp() function in libc.
So, this is an OS bug. I have a Solaris 7 installation and I do not see
this problem. Sun has apparently integrated the fix in the s28u4_06 and
s81_20 releases. I am not sure what these releases mean. It is possible
that these are media kit releases. Does anyone know?
On Thu, 1 Feb 2001, Dr. Ian C. Smith wrote:
> I have *exactly* the same problem. I did a soak test overnight and got upto
> about 512 MB RAM (i.e. nearly all of it) + approx. 1 GB of virtual memory
> taken by slapd. The server seems fairly stable on a solaris 6 machine
> though - the memory requirements stay constant at around 10 MB. The solaris
> 6 version does not have TCP wrapper support enabled whereas the solaris 8
> version does. I will look into whether this is the cause although I think
> it more likely to be an OS issue.
> Ian C. Smith,
> e-mail: email@example.com
> Address: Computer Sevices Department,
> Brownlow Hill Building - room G11,
> The University of Liverpool,
> Liverpool L69 3BX.
> --On Wednesday, January 31, 2001 11:39 PM -0500 Igor Brezac
> <firstname.lastname@example.org> wrote:
> > Hello,
> > I am using openldap 2.0.7 on solaris 8 and the bundled nss and pam
> > services. Both ldap server and the nss service are on the same machine.
> > I have about 10 entries in the directory and these entries should remain
> > in the cache even with the default slapd settings. I noticed that the
> > slapd process keeps growing. I used NS directory server in the past and I
> > did not see such behaviour. I tried to truss the slapd process and run it
> > the debug mode in an attempt to find the leak, but I was not successful.
> > I was wondering if anyone else has seem this situation? Or, if you would
> > share other debugging tips, I would appreciate it.
> > Your help will be appreciated.
> > -Igor