Count number of bytes in multiple directories

I have been busy moveing (scp-ing/reorganising) large numer of databases over the network, this little awk script makes me able to count the number of bytes and thus check if my nighlty scp’s didn’t miss a byte. (or two)

/cygdrive/c/oracle/oradata/DB01/DB01
$ find . -name "*DBF" -exec ls -l {} \;
-rwxrwx--- 1 Administrators SYSTEM 10493952 Oct  9 16:49 ./DB01/DATAFILE/O1_MF_TEST_3HSWQYNH_.DBF
-rwxrwx--- 1 Administrators SYSTEM 104865792 Oct  9 16:49 ./EXAMPLE01.DBF
-rwxrwx--- 1 Administrators SYSTEM 293412864 Oct  9 16:49 ./P1PERG001.DBF
-rwxrwx--- 1 Administrators SYSTEM 104865792 Oct  9 16:49 ./P1PERG002.DBF
-rwxrwx--- 1 Administrators SYSTEM 272637952 Oct 13 08:34 ./SYSAUX01.DBF
-rwxrwx--- 1 Administrators SYSTEM 513810432 Oct  9 16:49 ./SYSTEM01.DBF
-rwxrwx--- 1 Administrators SYSTEM 20979712 Oct 13 08:36 ./TEMP01.DBF
-rwxrwx--- 1 Administrators SYSTEM 47194112 Oct  9 16:49 ./UNDOTBS01.DBF
-rwxrwx--- 1 Administrators SYSTEM 5251072 Oct  9 16:49 ./USERS01.DBF
 
/cygdrive/c/oracle/oradata/DB01/DB01
$ find . -name "*DBF" -exec ls -l {} \; | awk '
> BEGIN { BYTES = 0 }
> { BYTES = BYTES + $5 }
> END { print "Total number of bytes = " BYTES}'
Total number of bytes = 1373511680
 
/cygdrive/c/oracle/oradata/DB01/DB01
$
 

Even shorter.

find /m*/oradata/ai018001/ -name "*dbf" -exec ls -l {} \; | awk '{B+=$5} END {print B}'
 

Display only directories in current directory

/cygdrive/c/temp $ ls -ld */
drwxr-xr-x+ 2 rz4tlj mkgroup-l-d 0 Sep 18 07:50 _backup_18-SEP-2007/
 
/cygdrive/c/temp $ 

send file as attachment

Use uuencode to encode a file and pipe it to mailx to be send as attachment (when you quickly need a file on your desktop it’s faster then ftp…)

uuencode file to mailx

$ uuencode file.txt file.txt | mailx -s file.txt <your-email-address>
$ 

In my environment the -r flag (return address) is needed since my mailserver is blocking mail without a known email address reply header.

uuencode file to mailx with -r flag

$ uuencode file.txt file.txt | mailx -s file.txt -r <your-email-address> <your-email-address>
$ 

nohup and SQL*Plus

Nohup and SQL*Plus can be easily used together:

$ nohup sqlplus -s remi/remivisser @update.sql &
[4]     14433
$ Sending output to nohup.out
 
$
$

Prevent password from being displayed in the ps list

$ export password=remivisser
$ nohup sqlplus -s remi/${password} @update.sql &
[8]     16768
$ Sending output to nohup.out
$
$
 

import from compressed export file

Example below demonstrates how to import ‘directly’ from a compressed file. Handy when you have not enough space on your mountpoint to unpack the file.

mknod pipe_db01 p
 
zcat exp.dmp.Z > pipe_db01 &
 
sleep 3
 
imp / file=pipe_db01 ... 
 

Same as above but now for the other compression utilities.

(win)zip

unzip -p exp.dmp.zip > pipe_db01 &

gzip

gunzip < exp.dmp.gz > pipe_db01 &
 
gzip -cNf < pipe_db01 > exp.dmp.Z &
gunzip -c exp.dmp.Z > pipe_db01 &

cpio

cpio -icvdu < exp.dmp.cpio > pipe_db01 &

The other way around

Export to pipe and compress.

$ mknod c.pipe p
$ nohup compress < c.pipe > exp_DB01_2007-10-31.dmp.Z & 
$ nohup exp system/$(cat .pwd) file=c.pipe owner=inotime log=exp_DB01_2007-10-31.log & 
 

Find/Remove old files

Find files modified more then 365 days ago;

find . -mtime +365 -exec ls -l {} \;

Use touch to test

$ touch -t 197801130300 oldfile
$ find . -mtime +365 -exec rm {} \;
$ ls oldfile
ls: oldfile: No such file or directory