1MD-MX-CTRL(8) System Manager's Manual MD-MX-CTRL(8)
2
3
4
6 md-mx-ctrl - Control mimedefang-multiplexor
7
9 md-mx-ctrl [options] command
10
11
13 md-mx-ctrl is a command-line tool for communicating with mimedefang-
14 multiplexor(8).
15
16
18 -h Displays usage information.
19
20
21 -s path
22 Specifies the path to the mimedefang-multiplexor socket. If not
23 specified, defaults to /var/spool/MIMEDefang/mimedefang-multi‐
24 plexor.sock.
25
26
27 -i This flag causes md-mx-ctrl to sit in a loop, reading commands
28 on standard input and printing results to standard output. It
29 is intended for use by a monitoring program such as watch-
30 mimedefang.
31
32
34 The following commands are available:
35
36
37 status Prints the status of all worker Perl processes in human-readable
38 format.
39
40
41 rawstatus
42 Prints the status of all worker Perl processes in a format easy
43 to parse by computer. The result is a single line with six
44 words on it. The words are separated by a single space charac‐
45 ter.
46
47 Each character in the first word corresponds to a worker, and is
48 "I" for an idle worker, "B" for a busy worker, "S" for a worker
49 which is not running, and "K" for a worker which has been
50 killed, but has not yet exited. A worker is "idle" if there is
51 a running Perl process waiting to do work. "Busy" means the
52 Perl process is currently filtering a message. "S" means there
53 is no associated Perl process with the worker, but one can be
54 started if the load warrants. Finally, "K" means the worker
55 Perl process has been killed, but has yet to terminate.
56
57 The second word is the total number of messages processed since
58 the multiplexor started up. The third word is the total number
59 of workers which have been activated since the multiplexor
60 started up. (That is, it's a count of the number of times the
61 multiplexor has forked and exec'd the Perl filter.)
62
63 The fourth word is the size of the queue for request queuing,
64 and the fifth word is the actual number of requests in the
65 queue. The sixth word is the number of seconds elapsed since
66 the multiplexor was started.
67
68
69 barstatus
70 Prints the status of busy workers and queued requests in a nice
71 "bar chart" format. This lets you keep an eye on things with a
72 script like this:
73
74 while true ; do
75 md-mx-ctrl barstatus
76 sleep 1
77 done
78
79
80 histo Prints a histogram showing the number of workers that were busy
81 each time a request was processed. A single line is printed for
82 the numbers from 1 up to the maximum number of workers. Each
83 line contains the count of busy workers (1, 2, 3 up to MX_MAXI‐
84 MUM), a space, and the number of times that many workers were
85 busy when a request was processed.
86
87
88 load Prints a table showing "load averages" for the last 10 seconds,
89 1 minute, 5 minutes and 10 minutes.
90
91 Each row in the table corresponds to a time interval, displayed
92 in the first column. The remaining columns in the table are:
93
94 Msgs: The number of messages scanned within the row's time
95 interval.
96
97 Msgs/Sec: The average number of messages scanned per second
98 within the row's time interval.
99
100 Avg Busy Workers: The average number of busy workers whenever a
101 message was scanned. (If you are processing any mail at all,
102 this number will be at least 1, because there is always 1 busy
103 worker when a message is scanned.)
104
105 If you have the watch(1) command on your system, you can keep an
106 eye on the load with this command:
107
108 watch -n 10 md-mx-ctrl load
109
110 If you do not have watch, the following shell script is a less
111 fancy equivalent:
112
113 #!/bin/sh
114 while true; do
115 clear
116 date
117 md-mx-ctrl load
118 sleep 10
119 done
120
121
122 rawload
123
124 Prints the load averages in computer-readable format. The for‐
125 mat consists of twenty-nine space-separated numbers:
126
127 The first four are integers representing the number of messages
128 scanned in the last 10 seconds, 1 minute, 5 minutes and 10 min‐
129 utes.
130
131 The second four are floating-point numbers representing the
132 average number of busy workers in the last 10 seconds, 1 minute,
133 5 minutes and 10 minutes.
134
135 The third four are floating-point numbers representing the aver‐
136 age time per scan in milliseconds over the last 10 seconds, 1
137 minute, 5 minutes and 10 minutes.
138
139 The fourth four are the number of worker activations (new work‐
140 ers started) over the last 10 seconds, 1 minute, 5 minutes and
141 10 minutes.
142
143 The fifth four are the number of workers reaped (workers that
144 have exited) over the last 10 seconds, 1 minute, 5 minutes and
145 10 minutes.
146
147 The sixth four are the number of busy, idle, stopped and killed
148 workers.
149
150 The seventh four are the number of messages processed, the num‐
151 ber of worker activations, the size of the request queue, and
152 the number of requests actually on the queue.
153
154 The final number is the number of seconds since the multiplexor
155 was started.
156
157
158 load-relayok
159 Similar to load, but shows timings for filter_relay calls.
160
161
162 load-senderok
163 Similar to load, but shows timings for filter_sender calls.
164
165
166 load-recipok
167 Similar to load, but shows timings for filter_recipient calls.
168
169
170 rawload-relayok
171 Similar to rawload, but shows timings for filter_relay calls.
172 Note that the worker activation and reap statistics are present,
173 but always 0. They are only valid in a rawload command.
174
175
176 rawload-senderok
177 Similar to rawload, but shows timings for filter_sender calls.
178 Note that the worker activation and reap statistics are present,
179 but always 0. They are only valid in a rawload command.
180
181
182 rawload-recipok
183 Similar to rawload, but shows timings for filter_recipient
184 calls. Note that the worker activation and reap statistics are
185 present, but always 0. They are only valid in a rawload com‐
186 mand.
187
188
189 load1 nsecs
190 The load1 command displays the load for various commands over
191 the last nsecs seconds, where nsecs is an integer from 10 to
192 600. The load1 command combines the output of load, load-
193 relayok, load-senderokf and load-recipok into one display.
194
195 You might use the command like this:
196
197 watch -n 10 md-mx-ctrl load1 60
198
199
200 rawload1 nsecs
201 Returns the load1 data in human-readable format. The result is
202 a line containing twenty-six space-separated numbers:
203
204 The first three numbers are the number of scans performed in the
205 last nsecs seconds, the average number of busy workers when a
206 scan was initiated and the average number of milliseconds per
207 scan.
208
209 The second three are the same measurements for filter_relay
210 calls.
211
212 The third three are the same measurements for filter_sender
213 calls.
214
215 The fourth three are the same measurements for filter_relay
216 calls.
217
218 The thirteenth through sixteenth numbers are the number of busy,
219 idle, stopped and killed workers, respectively.
220
221 The seventeenth number is the number of scans since mimedefang-
222 multiplexor was started.
223
224 The eighteenth number is the number of times a new worker has
225 been activated since program startup.
226
227 The nineteenth number is the size of the request queue and the
228 twentieth number is the actual number of queued requests.
229
230 The twenty-first number is the time since program startup and
231 the twenty-second number is a copy of nsecs for convenience.
232
233 The twenty-third through twenty-sixth numbers are the number of
234 workers currently executing a scan, relayok, senderok and
235 recipok command respectively.
236
237
238 workers
239 Displays a list of workers and their process IDs. Each line of
240 output consists of a worker number, a status (I, B, K, or S),
241 and for idle or busy workers, the process-ID of the worker. For
242 busy workers, the line may contain additional information about
243 what the worker is doing. The command slaves is a deprecated
244 synonym for this command.
245
246
247 busyworkers
248 Similar to workers, but only outputs a line for each busy
249 worker. The command busyslaves is a deprecated synonym for this
250 command.
251
252
253 workerinfo n
254 Displays information about worker number n. The command slave‐
255 info is a deprecated synonym for this command.
256
257
258 reread Forces mimedefang-multiplexor to kill all idle workers, and ter‐
259 minate and restart busy workers when they become idle. This
260 forces a reread of filter rules.
261
262
263 msgs Prints the total number of messages scanned since the multi‐
264 plexor started.
265
266
268 You can supply any other command and arguments to md-mx-ctrl. It per‐
269 cent-encodes each command-line argument, glues the encoded arguments
270 together with a single space between each, and sends the result to the
271 multiplexor as a command. This allows you to send arbitrary commands
272 to your Perl workers. See the section "EXTENDING MIMEDEFANG" in
273 mimedefang-filter(5) for additional details.
274
275
277 md-mx-ctrl uses the multiplexor's socket; therefore, it probably needs
278 to be run as root or the same user as mimedefang-multiplexor.
279
280
282 md-mx-ctrl was written by Dianne Skoll <dfs@roaringpenguin.com>. The
283 mimedefang home page is http://www.mimedefang.org/.
284
285
287 mimedefang.pl(8), mimedefang-filter(5), mimedefang(8), mimedefang-pro‐
288 tocol(7), watch-mimedefang(8)
289
290
291
2924th Berkeley Distribution 8 February 2005 MD-MX-CTRL(8)