Agnello George wrote:
> #!/bin/sh -x
>
> if [ $(tac /var/log |grep -e "error: syswrite()" | wc -l ) = 0 ] ; then
This can be improved. Let's walk through it.
On my system /var/log is a directory of log files and not an actual
log file. The typical log file is /var/log/syslog one many systems
and /var/log/messages on others and yet slightly different paths on
different systems. I will assume that you are not getting an error
message from that invocation above and on your system it really is a
file at /var/log but it would be good to verify this. Didn't you say
you wanted to check /var/log/spamd.log?
'tac' is 'cat' in reverse. 'tac' concatenates and prints files in
reverse. It does this by reading the file into memory and then
walking through the memory image in reverse. If the size of the image
is large enough then the file is copied to a temporary file. The
/var/log file is usually large enough to need a temporary file. This
is very inefficient.
In fact since the grep is going to walk through the entire file it
doesn't need to do so in reverse.
if [ $(tac /var/log |grep -e "error: syswrite()" | wc -l ) = 0 ] ; then
Is the same as the following. This removes an large inefficiency in
the check.
if [ $(grep -e "error: syswrite()" /var/log | wc -l ) = 0 ] ; then
Next let's look at grep itself. 'grep something file | wc -l' will
count the number of lines. But here you only care if it is zero or
nonzero. That can be done more efficiently with 'grep -q'. In the
case of -q since grep knows that it is a zero or nonzero occurrence it
can optimize and stop as soon as it knows the answer. Therefore the
above can be improved by using this:
if ! grep -q "error: syswrite()" /var/log; then
Let's now move to the collection of lines in the script using this new test.
if ! grep -q "error: syswrite()" /var/log; then
exit 1
else
echo "your mailserver is down" |mail -s " pls check server ip
216.185.xxx.xxx " [EMAIL PROTECTED]
fi
Using 'exit 1' indicates an error. But actually there is no error in
that case. It should be 'exit 0' instead. Also since this is an exit
the following parts past that in the script will not be executed.
It is considered good programming style to exit in the early part of
the 'if' statement and then not use an 'else' part in that case. But
usually this is done for error conditions and not for okay
conditions. Lets reverse the condition and try that here. Here is
a potentially final version of the script.
if grep -q "error: syswrite()" /var/log/spamd.log; then
echo "your mailserver is down" | mail -s " pls check server ip
216.185.xxx.xxx " [EMAIL PROTECTED]
exit 1
fi
exit 0
This way the script exits 0 (success) when there are no errors and
exits 1 (error) when errors exist in the logfile. This is a typical
way to program these checks.
You might also consider using 'mailq' to check your mta status.
> then i add a crontab to run for ever 10 min
>
> crontab -e
> 10 * * * * /your/location/of/script
As Eddy Beliveau pointed out this should be */10 using Vixie cron
syntax to run every ten minutes. But if your cron is not a Vixie cron
but is instead a traditional cron then you would need to list out all
of the minutes. 0,10,20,30,40,50 would run every ten minutes using
the older cron syntax.
> if there is an easier way kindly tell me !!!
Hope this helps,
Bob
