Re: Most stable loglevel

["matthew sporleder" <msporleder@gmail.com>]

> On 10/25/07, matthew sporleder <msporleder@gmail.com> wrote:
> > On 10/25/07, Sumith Narayanan <sumith.narayanan@gmail.com> wrote:
> > > Hi All,
> > >
> > > I am running $OpenLDAP: slapd 2.3.27 and  Berkeley DB (Version:
> > > 4.4.20) on MacOSX 10.4.
> > >
> > > The process crashes with memory leak often. The maximum it can go with
> > > 32 bit processor is 2 GB of memory and then it crashes. However , I
> > > have found that the it starts leaking more and crashing often (1-2 hr
> > > ) when the loglevel is 256 or 512. When it is at 32 it crashes once in
> > > 2 days or so.
> > >
> > > Did anyone experience this before ? If so , what should be the
> > > loglevel when my intention is to make the process more stable.
> > >
> > > Please note that the server which crashes is getting updated 5-6 % of
> > > total transaction.
> > >
> > > Any help , suggestions will be appreciated.
> > >
> >
> > I have found that many people misinterpret an oversized cache with a
> > memory leak.  Tune down your system so it doesn't use so much memory.
> >
>


On 10/26/07, Sumith Narayanan <sumith.narayanan@gmail.com> wrote:
> Hi Mattew,
>
> We have 3 DBs and alltogether we are using less than 250 MB memory.
>
> When the process starts , it stabilzes with 250 MB and then it starts
> increasing.
>
> After putting the log level to 32 , it is slowly increasing , over a
> period of 2 days or so to 2 GB and then it crashes.
>
> If I increase the log level it goes to 2 GB in 1 hour or so.


It sounds like your syslog config is somehow messed up, or osx's
syslog is not very good and you should try installing something like
syslog-ng.