That happened to me as well a couple weeks ago, not sure what caused it. I fixed it by looking up documentation on the RRDTOOLS that create the graphs, finding the graph data files at /var/mailcleaner/spool/rrdtools/127.0.0.1 and using the 'rrdtool' utility to dump the CPU data to a text file, editing the text file to remove the huge numbers -- I replaced numbers with e+04 or larger with 'NaN' (Not a Number) -- and then using the rrdtools to convert the text file back to a binary rrd file. There were a few dozen numbers to change and it was a bit of a pain, but it worked for me. Here's documentation on the rrd tools:
http://oss.oetiker.ch/rrdtool/doc/rrdtool.en.html
As an example, I did it something like this:
Code: Select all
cd /var/mailcleaner/spool/rrdtools/127.0.0.1
/usr/rrdtools/bin/rrdtool dump cpu.rrd >cpu.xml
Edit cpu.xml using your favorite text editor and replace huge numbers, numbers ending in e+04 through e+07 in my case, with NaN or zero. Then make a backup and convert the text file back to binary:
Code: Select all
cp cpu.rrd cpu.rrd.backup
/usr/rrdtools/bin/rrdtool restore cpu.xml cpu.rrd
I don't know if it would be better to shut down or disable the rrd process or something before doing this; I did it live while everything was running. Next RRD update (every 5 minutes?) the graph was back to normal.