What I meant by exit code is the value of $?.
Under normal circumstances, like a successfully run command, the exit code will be 0. If there's an error, it will be something else. The sample script above checks to make sure that $? = 0, and if not, it will send an email.
The concern I have is that something in whatever script is run may run successfully after the failed command... I don't know if I'm explaining very well, so I'll try to show you:
Code:
$ cat tmp1.sh
#!/bin/sh
echo
wuzzle
$ ./tmp1.sh
./tmp1.sh: wuzzle: command not found
$ echo $?
127
$ cat tmp2.sh
#!/bin/sh
wuzzle
echo
$ ./tmp2.sh
./tmp2.sh: wuzzle: command not found
$ echo $?
0
So, you see, unless you told the fullback script to "exit 1" (or exit anything not equal to 0) after any component failed, it might not email you if it does fail.
It depends completely on your script (the fullback.ksh one), is should be easy to make sure it does this correctly...
On the other hand, if you're going to be rewriting to do this, why not put them all together into one? Here's a very simple example:
Code:
#!/usr/bin/ksh
/bin/tar vcf /dev/rmt/0 /home/admin >/var/adm/backup_log 2>&1
if [[ $? -ne 0 ]]; then
/bin/mail admin <<EndOfMail
Subject: Your backup had errors
See the contents of /var/adm/backup_log
for more details.
.
EndOfMail
fi
I guess it all depends on what you're trying to do.
And please let me know if I don't make any sense... I'm kind of tired here
Getting ready for Sarbanes Oxley and I was told that I need to create a cron.allow file for the box so that only specified users can create cron jobs. That was easy enough, and I put in all the accounts that had scheduled cron jobs such as:
root
lp
My oracle account
A couple of programmer... (1 Reply)
I have two different cron jobs that run on the same days. The jobs are 7 hours apart. Both jobs are set to send notification emails when they start running.
Both jobs always run successfully, but I only receive an email from the first job. I never get the email from the second job.
... (0 Replies)
Hi
Im very new at working with unix and this problem I simply can not understand. I know there are a lot of threads about problems with shell scripts behaving differently when run from a terminal and from a cronjob. I have tried everything(almost) but I still havent cracked this problem.
Im... (15 Replies)
I have a very basic script that essentially sends a log file, via FTP, to a backup server. My cron entry to run this every night is:
55 23 * * * /usr/bin/archive_logs
The script runs perfectly when executed manually, and actually worked via cron for about three weeks. However, it mysteriously... (3 Replies)
Hello,
I have added the following entry in my cron file:
* * * * * . /home/texprd/.profile>/dev/null 2>&1; echo "test" >> /home/texprd/test.log 2>&1
The cron fails ie the added command fails to append "test" in the test.log file. However if I run the command manually from the shell... (3 Replies)
Im having a problem with my ftp CRON jobs. I have another related (but not the same question in shell scripting).
My FTP CRON jobs connect to a server and drop some files off, if for some reason it cannot connect to the server (timesout/password wrong etc) instead of stopping, it tries again... (3 Replies)
I added a shell script to cron that executes every 5-15 minutes. It's used to monitor multiple environments, so it has to ssh to each machine to check the status of that machine. Sometimes, I may have more than one instance of this script running at the same time. So, here's my problem. I've... (2 Replies)
Hi Guys...
I have a bulk email program that sends out an email to a subscribed mail list.
The list contains around 2000 addresses.
My server is Centos 5 64bit
PHP 5.2
Mysql 5.1
The program is set to send emails in batches of 100.
Cron is set to run every minute
Now for the problem:... (2 Replies)
Hi,
When run the script directly...it executes as expected. But when put it in crontab the job fails with this error:
Connection not open
08003: Connection not open
Unable to connect to the database...
how come it is not able to connect when cron job fires?... Do I need to... (7 Replies)
Hi,
Below is my script which sends out emails.
more test.sh
1 #!/bin/ksh
2 {
3 print "From: Prod@mycomp.com"
4 # print "From: Prod@`hostname`.com"
5 # print "From: Prod@${HOSTNAME}.com"
6 print "To: me@mycomp.com"
7 print "Subject:... (5 Replies)
Discussion started by: mohtashims
5 Replies
LEARN ABOUT DEBIAN
mailping-cron
2004-04-15
mailping
0.0.4
MAILPING-CRON(1) Mailping MAILPING-CRON(1)NAME
mailping-cron - run periodic processing to test email service availability and functioning
SYNOPSIS
mailping-cron
DESCRIPTION
mailping-cron processes incoming emails, updates status and sends out probe messages.
The idea is to configure multiple email "circuits", send probe messages regularly, and see whether they complete the circuit and how long
it took.
Setting up a circuit
1. Give a name to the circuit, hereafter referred to as circuit.
2. Arrange for a local email address to be delivered to maildir /var/lib/mailping/state/circuit/incoming/. See mailping-store.
3. Create a configuration for the circuit, by creating directory /etc/mailping/circuit.
If you need non-default sender and/or recipient addresses, create files from and to there, containing the sender and recipient
addresses suitable for the circuit.
4. You're done! Munin should now see the data.
Testing multiple servers
To test functioning of more than one email server, arrange an email alias at a remote site pointing to an address on your server, and set
the address of that alias here.
Here's an example of testing a system consisting of two email servers and everything in that path (smarthosts, primary MXs, virus checkers,
etc.):
Local address <mailping+that@this.example.com> is delivered with mailping-store to /var/lib/mailping/state/that/incoming/.
Remote address <echo@that.example.com> is an alias that redirects all email to <mailping+that@this.example.com>.
/etc/mailping/that/to is set to <echo@that.example.com>.
FILES
/etc/mailping/circuit/from
Sender address for the probe emails. Default: <currentuserid@fullyqualifiedhostname>.
/etc/mailping/circuit/to
Recipient address for the probe emails. You must arrange for the email to eventually get delivered to the maildir
/var/lib/mailping/state/circuit/incoming/. mailping-store will probably be useful in that. Default:
<currentuserid+circuit@fullyqualifiedhostname>
/etc/mailping/circuit/admin
Admistrative address, set as Reply-To in probe messages. Default: do not add Reply-To.
/etc/mailping/circuit/interval
How often a probe message is sent, in seconds. Default: 600 seconds.
/var/lib/mailping/state/circuit/
Stored state for the probing.
/var/lib/mailping/state/circuit/junk/
Maildir used to store all messages in incoming that do not look like probe messages. Read and delete them regularly.
/var/lib/mailping/state/circuit/broken/
Maildir used to store all messages in incoming that do look like probe messages, but a corresponding pending entry cannot be found.
Duplicated probe messages cause these. Read and delete them regularly.
ENVIRONMENT
MAILPING_CONFIGDIR
Override the location of the configuration directory. Default: /etc/mailping
MAILPING_STATEDIR
Override the location of the state directory. Circuit states are stored in the state subdirectory of this directory, in subdirectories
named after the circuit name. Default: /var/lib/mailping
SEE ALSO mailping-store(1), mailping-success(1), mailping-latency(1)AUTHOR
Tommi Virtanen <tv@havoc.fi>
Havoc Consulting
Author.
COPYRIGHT
Copyright (C) 2004 Havoc Consulting
mailping 0. 2004-04-15 MAILPING-CRON(1)
03-15-2002
