Hello all!
Wednesday 14 October 1998, I wrote this:
> We have a Alpha Server 1000 which is running under Digital Unix v4.0B.
> It has 256Mo of RAM, a 400MHz processor...
> It often happens that the swap space is completly used, and that we hava
> the following messages:
> - no more processes
> - swap space below 10%
> - and sometimes: core dumped!!!
>
> The swap space is of 657Mo!
> The problem happens when we use tha database OpenIngres (10 users
> approximativly each time) and that we make some compilation of simple
programs.
> This is server used in an university! We have approximativly 50 X terminal
> connected on it! ANd it often arrives too, that the time of connection (while
it
> is loading the NCDware file) is anormally long (10 minutes and more). The file
> Xncdxpl is download on the nvram of each terminal in tftp mode.
>
I had lots of mails from all of you! I wont name you all, but I d like to thank
you a lot!
Many of you says me to do that:
- Change the mode of allocating swap space by deleting (or just
renaming) the file /etc/swapdefault and then to reboot the server to take
effect.
Since I ve made it, the problem doesnt happen again! When lots of
resources are required, it can take some times but it never makes any mistake!
Thank you again!
Florent VEJUX
IUT de Dijon
Received on Thu Oct 15 1998 - 09:14:38 NZDT