Debian Codegraph @

contrib/exim-stats/README

raw
1 Here are the scripts I put together to gather EXIM processing information
2 for MRTG to display.
3
4 Please excuse the lack of elegance this whole thing has. It was thrown
5 together in about an hour and I haven't had time to go back and make it
6 pretty. There are a couple of ways I can easily think of to make this
7 better. One would be to not use 'finger' for the gathering of information.
8 The other would be to make one call every five minutes and use the output
9 from it for all three items being graphed (rather than three calls every 5
10 minutes). One of these days I'll get around to incorporating these
11 enhancements...
12
13 The scripts that are called from MRTG (email-cout1, email-queue1, and
14 email-totals1) get their information by 'fingering' the remote hosts. On
15 the remote hosts, the 'fingerd' process is replaced with the check-email
16 script. This is usually done by replacing the INETD entry on the mail host
17 from something that looks like this:
18
19 finger stream tcp nowait nobody /usr/sbin/in.fingerd in.fingerd
20
21 to something that looks like this:
22
23 finger stream tcp nowait nobody /usr/local/bin/check-email check-email
24
25 (Note the path is where ever you decide to put the check-email script)
26 Also, the path to GNU grep is hardcoded in the 'check-email' script. I use
27 GNU grep for the counts which is alot faster than the normal system grep's
28 out there.
29
30 The output of check-email is:
31
32 Working.
33 ======================================================================
34 E-mail status information for mercury.csc.com
35
36 Current Date/Time is: Monday, 01/04/99 07:07 EST
37
38 Number of messages currently in the queue: 54
39
40 Number of messages received so far today: 3465
41 Number of messages delivered so far today: 3945
42
43 END OF OUTPUT
44
45
46 Now then for what the scripts do. email-count1 takes the count of messages
47 received and delivered so far today and subtracts the prior values from
48 the current values obtained from 'check-email'. This gives you the number
49 of messages received and delivered in the last five minutes (since MRTG
50 is called every five minutes). email-count1 has two files it tries to read
51 without checking to see if they exist. Yes this is bad. To make it work,
52 create files called 'HOST-NAME.infile' and 'HOST-NAME.outfile' in a
53 directory (see the script) and put just the number 0 in there before you
54 run the scripts.
55
56 email-queue1 simply takes the queue information from 'check-email' and
57 returns something that MRTG wants.
58
59 Lastly email-totals1 takes the same count information and does a running
60 total of the amount received and delivered since the last time the log was
61 cleaned (for us that is every day at midnight).
62
63 The file 'email.cfg' is the MRTG config file for a single host to get all
64 three stats going. If you don't want a particular bit of information,
65 remove the section. You will have to globally replace the TAG 'MAIL-HOST'
66 with the name of the host you want to check. You will also probably want
67 to change the verbage around each bit of information. That is entirely up
68 to you :-)
69
70