On Thu, Jan 01, 2004 at 04:25:17AM +0800, Matt Massie wrote:
> if i hear that everything is ok.. i'll officially release 2.5.6.
I just committed a change to configure.in which should fix FreeBSD ia64
support (there was a bogus catchall ia64 case in the match against the
os, vendor, machine triplet
guys-
i fixed the negative TN bug. you can download 2.5.6 from
http://matt-massie.com/ganglia/
i'd like to hear from the people who are having problems that their
problem is now fixed. the fix is very simple.
in gmond...
if(tn<0)
tn=0;
i'd be shocked if the problem is not fixed but i'm ne
Ah ha! You hit the problem on the head.
gmetad is assigning a negative value to tn, which is a uint32. This
messes up the comparison.
So. We can abs() tn when it gets assigned, in process_xml.
Or you can make tn and tmax regular int32s.
Though another solution is to make sure TN is never ne
On Thu, 2004-01-01 at 00:55, Jason A. Smith wrote:
> I have noticed the same exact problem here, occasionally some nodes
> would get marked as down even though they are still up, and it appears
> to be the same timing issue. Based on what you discovered below it
> appears that gmetad is using an u
I have noticed the same exact problem here, occasionally some nodes
would get marked as down even though they are still up, and it appears
to be the same timing issue. Based on what you discovered below it
appears that gmetad is using an unsigned int to store TN and gmond is
using a signed int.
I
