I am having these problems while trying to backup files from a remote machine that I telnet into

I have shut down the machien and brought it back up again and I am still getting these errors.
The script I use can be found below as well

ERRORS
cpio: Read error at byte 0 in file /folder/oracle/testdata/system08.dbf, padding
with zeros
/folder/oracle/testdata/system08.dbf
cpio: Read error at byte 0 in file /folder/oracle/testdata/system09.dbf, padding
with zeros
/folder/oracle/testdata/system09.dbf
cpio: Read error at byte 0 in file /folder/oracle/testdata/system10.dbf, padding
with zeros
/folder/oracle/testdata/system10.dbf
cpio: Read error at byte 0 in file /folder/oracle/testdata/system11.dbf, padding
with zeros
/folder/oracle/testdata/system11.dbf
cpio: Read error at byte 0 in file /folder/oracle/testdata/temp01.dbf, padding w
ith zeros
/folder/oracle/testdata/temp01.dbf
cpio: Read error at byte 0 in file /folder/oracle/testdata/undo01.dbf, padding w
ith zeros
/folder/oracle/testdata/undo01.dbf
/folderora/oraInventory/install.platform
1366034 blocks


BACKUP SCRIPT
# the following is used to backup the TEST database files
# mount the drives from the HRIS test machine
service nfs restart
mount -t nfs -o rw,hard,intr 128.x.x.x:/u0 /folder
mount -t nfs -o rw,hard,intr 128.x.x.x:/etc /folderora
mount -t nfs -o rw,hard,intr 128.x.x.x:/u1 /folder1
-- database files are here
find /folder/oracle/testdata > tmp
-- parameter files are here
find /folderora/oraInventory >> tmp

# Weekly backup is included here for the TEST Instance
find /folder/oracle/testcomn >> tmp
find /folder/oracle/testdb >> tmp
find /folder/oracle/testora >> tmp
find /folder/oracle/*.* >> tmp

find /folder/oracle >> tmp

cat tmp| cpio -ocvB > /dev/st0


umount /folder
umount /folderora

Interesting.

I've seen this kind of message with Veritas Netbackup and it meant that the file had changed after it had been selected for backup. Did you make sure to shutdown your DB before you started the backup?

It may not be the same as Netbackup uses a customized version GNU Tar but the fact that they are db files suggests it is to me.

I'm positive that the DB is down and i ran cpio after it was down so it should not be a issue of size change.

any other suggestions?

After you shutdown the DB run "lsof" agains the directory it resides on (or even the specific file) to be sure nothing else is busying it out. Oracle for example run listeners, concurrent manager processes etc... that are not actually part of the base DB processes like pmon, lgwr etc...

lsof /folder/oracle/testdata/system08.dbf
Would show you any processes using that file.

lsof /folder/oracle/testdata
Would show you any processes using that directory

By the way - this isn't an ocfs filesystem is it? Standard tools don't work for those but there is a download for a tar that does.

Another thought is the file size. Is it greater than 2 GB? Many utilities aren't enabled for large file support. The man page (man cpio) should indicate whether it is. If not mentioned in the man page you should assume it isn't. Off the top of my head I don't think cpio is.