[ic] Intermittent PGP failures & Lost credit card info

Jerry interchange-users@interchange.redhat.com
Mon Feb 4 12:58:00 2002


> This is an interesting report, though I cannot see what we are
> doing different between those two versions. I wonder if the 
> traffic is greater on the newer machine.
> 
> bill:~% perl -le 'print 3072 >> 8'
> 12
> bill:~% grep ' 12 ' /usr/include/asm/errno.h
> #define	ENOMEM		12	/* Out of memory */
> 
> I can't think of what we would do that would cause this.

> 
> Just because something is obviously happening doesn't mean something
> obvious is happening. --Larry Wall

Well THAT prompted me to take a much closer look. I don't think you
did anyting to cause it. That server has much higher traffic and more
importantly does not have all the memory installed that it was 
suppose to have! Looking through all logs, it looks like during the
times this happened that it simply ran out of core during the
encrypt process. Need to watch those hardware people a little closer!

Perhaps trapping that error and giving the customer another shot 
wouldn't be a bad idea though. Off to find the hardware person.
(I think he's hiding).

Great tag line for this thread Mike.

Jerry Uphoff