Yea i have search for this and found the Bug report and there is only a solution for the 48h statistic Bug.
The only thing that says something to avoid the warnings was this:
Code:
By putting an if around the variable the warnings seem to be stopped. I'm
really not sure if this is in any way correct but FYI, this is how I modified
my zmstat_allproc delta_stat function:
sub delta_stat($$$$) {
my $prev_stats = shift;
my $curr_stats = shift;
my $pid = shift;
my $stat = shift;
if($prev_stats->{$pid}->{$stat}) {
$curr_stats->{$pid}->{$stat} - $prev_stats->{$pid}->{$stat};
} else {
0;
}
} The author says he has no idea if this is correct. Maybe this self made patch is in some way or in some environments harmful and i get other problems. There are also no comments about it if this is a good or bad solution.
The only thing that i can say is that i have no idea about the Zimbra code and i don't would test such patches (where the author is not sure about it) on working environments.
Can someone say something about this patch?
I have also found out that this problem only comes up in 64 bit environments. I have some 32 bit Zimbra servers with the same version where i don't get this warnings.
By the side the "sub delta_stat($$$$)" functions in "/opt/zimbra/libexec/zmstat-allprocs" are on 64 and 32 bit identically (line 111 - 119).
It seems for me that the above posted patch only avoid the warning and does not solve the problem itself.
How can it be that "$prev_stats->{$pid}->{$stat}" is uninitialized?
And sorry i have to say that but i think the only one who should mark my posts as solved was me
Yogg
Edit:
Is there a way to fix the tile of the post?