On 01/05/2011 05:47 PM, derleader __ wrote:
>There is always chance to lose information. That is reason why we use
>database with journal.
I mean if Apache crashes the valuable data which will be stored in
shared memory will be lost and stored into RAM until system reboot. How
do you deal with this problem?
If receiving end will fail, we will repeat the communication. This is no
problem, if only one apache process will die. But if all apaches process
will crash, then something terrible will happen. Whole site will be
unavailable.
As far as know in shared memory communication method if one of the
daemons crashes the data will be lost and there will be memory leak.
>> Have you done any benchmark tests?
>Yes.
would you share some information how much calls with share memory can be
done per second and how much client servers can spacewalk handle.
We did not count share memory usage, as it is buried deep down in process.
But we do benchmark RHN Satellite, to catch possible regression in speed
of RHN Satellite.
I'm not allowed to speak about it in more deep as it is one of the
advantages RHN Satellite users get for paying for our work.
--
Miroslav Suchy
Red Hat Satellite Engineering
_______________________________________________
Spacewalk-devel mailing list
Spacewalk-devel@redhat.com
https://www.redhat.com/mailman/listinfo/spacewalk-devel
