Hi,
We(me and my coworker) don't know enough details about rrd file data structure.
Here is a problem we trying to resolve.
We have a pair of Xymon(Hobbit) servers, hb1 and hb2. They each receive one copy of hobbit message/metric data from each hobbit clients. ie hobbit client send message to hb1 and hb2.
Assuming hb1 is down and back online in 1 hour. now we have 20K of rrd files with 1 hour hole in them. But those missing 1 hour data is available on hb2 waiting for action ;).
We write a simple perl script to dump out one rrd file from hb1 and one from hb2 into xml files.
then we patch the missed data lines in hb1 from hb2 between the <database> </database> tags. after the patching we then restore the xml fle back into rrd format.
The question is , "Is that all we need to do to fill a hole from a good copy ?"
Looking at other tags in dumped XML file, is there a counter of how many lines within <database> ...</database> tags ?
by inserting the missing lines into hb1 rrd, will we corrupt/mess up the metrics in the hb1 rrd file ?
Thanks for your pointer
tj
