I forgot to mention this concerns a Tru64 V4.0d box
After downsizing the file (within Oracle) just below its original size, and
after that enlargen it to just over 512K, the ls command came with the
correct size (approx. matching the du output); I'm still wondering about the
2 questions below...
regards,
Alex H
-----Original Message-----
From: Alex Harkema [mailto:HarkemaA_at_vertis.nl]
Sent: Friday, December 29, 2000 2:48 PM
To: 'tru64-unix-managers_at_ornl.gov'
Subject: differences in filesizes/diskusage
Hi all,
I noticed some weird size-differences in the output of various commands.
I know it's not very polite to ask more than one question at once, but those
two are very related, I don't think it can do any harm;-)
== Q.1 ==
After trying to resize a datafile, one filesystem complained about being
100% in-use
The original filesize was 512K. We tried to resize it to 768K,
[root]# du -k den0d03.dbf
524304 den0d03.dbf
[root]# ls -l den0*
786434048 Dec 29 13:47 den0d03.dbf
= checking for contents of the file (after the 512th block):
[root]# dd if=den0d03.dbf bs=1024 skip=512000 count=10 | od -x | more
returns with all zeroes, meaning there's no data after block 512000
Why does df -k show the filedomain being 100% used (showfdmn does, too)
while diskusage shows that the file is actually (almost) 512K of size?
Does df work with the actual diskusage?
== Q.2 ==
= Checking another file gives:
[root]# ls -l /xxx/oradata/xxx/system01.dbf
157288448 Dec 29 13:47 /xxx/oradata/xxx/system01.dbf
= diskusage is:
[root]# du -k /xxx/oradata/xxx/system01.dbf
153608 /xxx/oradata/xxx/system01.dbf
= 157286400 is the size Oracle says the file has.
[root]3# echo $((157288448 - 157286400))
= The difference between both values is 2048 (that's 1 block, right?)
Any ideas what causes this differences? (some meta-info perhaps?)
Thanx for your help/suggestions in advance.
Alex
harkemaa_at_vertis.nl
Received on Fri Dec 29 2000 - 18:44:27 NZDT