Hello,

I was running some tests with near-full disks (>80%) and noticed that I do get plenty of mails (thanks!), but the disk usage for that period is not available in the disk usage graphs.

If you want this (IMHO critical) period to be visible from your server logs as well, apply the following patch to /opt/zimbra/libexec/zmdisklog:

Code:
--- zmdisklog--	2008-10-27 13:41:57.000000000 +0100
+++ zmdisklog	2008-10-27 13:45:20.000000000 +0100
@@ -89,13 +89,12 @@
     my $pct=int(($used/$total)*100) if ($total > 0);
     #print "TOTAL: $total USED: $used AVAIL: $avail PCT: $pct\n";
     my $lvl = "info";
     $lvl = "err" if ($pct > $DISK_WARN_THRESHOLD);
     $lvl = "crit" if ($pct > $DISK_CRIT_THRESHOLD);
-    if ($lvl eq "info") {
-		  Zimbra::Mon::Logger::Log( "$lvl", "$dt, DISK: ${hostname}: dev: $dev, mp: $mp, tot: $total, avail: $avail" );
-    } else {
+    Zimbra::Mon::Logger::Log( "info", "$dt, DISK: ${hostname}: dev: $dev, mp: $mp, tot: $total, avail: $avail" );
+    if ($lvl ne "info") {
 		  Zimbra::Mon::Logger::Log( "$lvl", "Disk warning: ${hostname} $dev at $pct%" );
     }
 	}
   alarm(0);
 }
This will make sure that the current disk usage is always logged (and thus later rendered into the statistics), even if a disk full warning is also generated.

-Marcel