--On Friday, May 27, 2005 7:58 AM -0500 Burton Strauss
<[EMAIL PROTECTED]> wrote:
I used to see that. Not frequently, but enough to be a problem. It was
due to a corrupted rrd file causing the rrdMainLoop thread to bomb. Once
that bombs, it doesn't restart and poof.
(1) Try enabling debug and - although you'll get a ton of messages -
buried in there should be either some errors or you will see the last one
that updates.
I'm assuming you mean tracing. I set it to 7 but don't see anything
reported from the RRD plugin after it says that it's started.
(2) Check the mtime (modification times) of the .rrd files. The one that
isn't being updated in a set is the suspect. i.e. if the last update was
to /a/b/c/d/xxxx.rrd, anything else in that directory that wasn't updated
on that cycle is suspect (but remember, only .rrd files which have data
points are updated).
I tried renaming the old rrd file and creating a new empty one. The plugin
recreates the 3 subdirectories at startup, but it doesn't create the "eth1"
interface subdirectory, and complains of its absence when I try to view
data. Seems like data collection just mysteriously won't start.
My system has eth0 shut down. Could that be an issue? (Just thinking that
the collector might fall over while enumerating interfaces.) I'll poke
through the sources and see what I can see.
_______________________________________________
Ntop mailing list
[email protected]
http://listgateway.unipi.it/mailman/listinfo/ntop
