Solaris 5.6 init growing...

• Messages sorted by:[ date ][ thread ][ subject ][ author ]

Hi,

On Fri, Jan 22, 1999 at 09:31:28AM -0400, David L. Potter wrote:
> Thanks for your reply. We're interested in trying to solve the problem of
> the growing init process if possible. 

Good.

> (We like the fact that with mgetty
> we don't have to have the modems set to auto answer...)

So do I :)

> We have a second Solaris 2.6 machine that serves as a back-up and provides
> some secondary processing for our site (a community network). We're
> proposing to configure a single modem attached to a test line to the
> second machine and undertake some preliminary testing...
> 
> We thought...
> 
> 1) We'd start, by simply killing the mgetty and letting it respawn and
> see if that alone will cause the growth... 

I think it should, but maybe only very slowly.

> 2) The next step might be to test after a connection, before handing
> off to `login `

There should be no difference here - init can't tell the difference.  But
then, it might be worth a try.

> 3) ...testing after the hand-off
> 
> 4) ...ending the session and having init respawn mgetty.
> 
> By doing a `gcore ` we should be able to get a very accurate picture of
> the growth (if any) of init at any point....

Yep.

> Do you have any suggestions... we're using 1.0.0 on the production
> machine but I guess we should use the latest version...?

I don't think the mgetty version will make a difference.

gert
-- 
USENET is *not* the non-clickable part of WWW!
                                                           //www.muc.de/~gert/
Gert Doering - Munich, Germany                             gert@greenie.muc.de
fax: +49-89-35655025                        gert.doering@physik.tu-muenchen.de