Log entry time 13:24:18 on November 07, 2010

Entry number 338993

This entry is a followup to: 338981

This is in regards to the "pingpong" script that chooses the disk
for the data file.  

I fixed a bug which had caused it to not skip /adaql2/data3 which
at present is 100% full.  

Although I had tested this script thoroughly, it was a last minute bug
where I tuned the parameter to 95% -- i.e. skip if disk is > 95% full.
This seemed reasonable, but now I remember that the "bytes available"
are NOT really available (because of RAID, I suppose), so I've backed 
this off to 90% so that it skips if there's not enough space to 
write 2 GB.