Many thanks everyone.
The consensus is that this was probably a sparse file,
which was made sparse because the disk filled up during
an NFS write; but because the application probably ploughed
on regardless, NFS file-end pointers were incremented
even tho the file didnt grow.
Unfortunately the user deleted the file before I had
a chance to investigate further.
Cheers,
T.
>Original msg:
>Can anyone explain this filesize inconsistency please?
>
>alf1_tsh> ls -l /pnac0/di/Golm/Sphere180302_2/farmjob707.log
>-rw-r--r-- 1 di di 3089268736 Mar 19 10:34 /pnac0/di/Golm/Sphere180302_2/farmjob707.log
>
>alf1_tsh> du -k /pnac0/di/Golm/Sphere180302_2/farmjob707.log
>661256 /pnac0/di/Golm/Sphere180302_2/farmjob707.log
>
>
>/pnac0 was NFS-exported to the machine from which the file was being
>written (hostname fm01).
>
>The disk was mounted NFS V3:
>fm01_tsh> /sbin/mount | grep pnac0
>alf1:/pnac0 on /pnac0 type nfs (v3, rw, udp, hard, intr, timeo=100)
>
>The file is a plain ASCII file, and the disk filled up whilst
>it was being written.
>
>Both machines are running TU64 5.0A PK3.
Received on Wed Mar 20 2002 - 11:04:42 NZST