| blob | 112fd904faba7d1e5bdea20c84a79823e9168a28 |
1 Technical notes on my code submission of 8 Dec 99
2 Kern Sibbald
4 General:
5 - Please delete apccontrol.sh and apccontrol.sh.in from the
6 scripts directory.
7 - Please delete apcnetlib.c from the cgi directory.
8 - With Carl Erhorn's help, we got apcupsd running on
9 Solaris, and we believe that we have resolved the
10 problems for the HP-UX as well.
11 - New html manual in doc/manual. It still needs more
12 work, but a lot of the sections are pretty well
13 documented.
14 - I added new directories for the Sun and HP-UX named
15 distributions/sun and distributions/hpux. They
16 contain skeletions of the necessary files.
17 - I added automatic detection of the Sun and HP systems
18 to the configure script.
19 - Due to the differences in the shutdown procedures
20 (calls to shutdown) between the Solaris operating
21 system and Linux, I had to move apccontrol.sh.in
22 from the scripts directory to each of the machine
23 dependent files (distributions/redhat, ...).
24 - Because of the problems with Solaris, I eliminated
25 the action process by combining it with the serial
26 port process.
27 - I corrected the CGI programs to properly recognize
28 different battery voltages (12, 24, and 48). I did
29 the same for the line voltages (100, 120, 220).
31 Changes submitted this submission:
32 - I corrected the compile and link flags in Makefile.in
33 for the Sun as suggested by Carl.
34 - I modified the installation part of Makefile.in to
35 pull apccontrol.sh from the appropriate distributions
36 subdirectory. This way, apccontrol.sh can be adapted for
37 the different operating systems.
38 - Removed an old le[] from apcaccess and aligned some
39 code.
40 - Modified apcupsd.c apcserial.c and apcaction.c so that
41 the previous action process is eliminated and instead
42 the action code is called immediately after the wait
43 on the serial port. This eliminates a system incompatibility
44 that arose because of the handling of alarm(), and it
45 ensures that any actions are attended to immediately
46 when they occur. In addition, in future releases, this
47 will permit sending of commands to the UPS as there
48 is now only a single process reading the serial port
49 and taking action.
50 - In writing the documentation, I realized that some of
51 the names of the new configuration directives that I
52 created are misleading. Consequently, I changed
54 NETSTATUS -> NETSERVER
55 STATUSPORT -> SERVERPORT
56 EVENTFILE -> EVENTSFILE
58 For this update, I left the old names in, but in a few
59 weeks, I will remove them.
60 - In order to make the STATUS code work for older style
61 UPSes, I had to move the "flags" variable into the shared
62 memory buffer. This was cleanup of the STATUS code that
63 I had forgotten to do.
64 - During truncating of the temporary EVENTS file, I forgot
65 to rewind() the file after truncating it. This caused the
66 subsequent write back of the old data to be lost.
67 - I changed the name "sun" to "apcsun" in apcipc.c to avoid
68 problems with the Sun compiler as reported by Carl.
69 - The file locking code that detected stale lock files was
70 not working correctly because it didn't take into account
71 the fact that the lock file was created by the parent of
72 the current process. This was observed by Carl and had
73 previously gone undetected because the details of these
74 failures were never logged until my last submission.
75 - In apclog.c, I was writing all events to the temporary
76 events file, but I should have excluded LOG_INFO, which
77 is the DATA logging. Bug observed by Carl.
78 - Carl noted a problem in apcnet.c pointed out by his compiler,
79 when I looked at the code, I realized that it apparently could
80 never work since it did a read and lock of the shared memory
81 area then entered an infinite loop from which it would never
82 exit. This had the effect of locking out all other processes
83 from reading the shared memory area. I restructure the
84 code. I also eliminate the alarm() and replaced it by a
85 sleep().
86 - Corrected several casts that caused compiler warnings on
87 Solaris and added INADDR_NONE in apcnetlib.c for Solaris.
88 - Carl reported that apcupsd.log was always created. This was
89 because vestiges of the old data logging remained in the
90 code. I removed them.
91 - In testing the shutdown code again, I found it did not send
92 the correct commands to the UPS for turning off the power.
93 Consequently, I reworked the logic a bit and made it
94 correctly detect all UPS responses as well as send the
95 correct codes.
96 - In my testing, I noticed that if I disconnected the serial port
97 cable, it took more than 2 minutes for apcupsd to complain.
98 The problem turned out to be the smart_poll() was not so smart
99 after all in that it ignored the error return status. I modified
100 the code to immediately call UPSlinkCheck() if it got an error.
101 This means that loss of serial port communications is detected
102 within 6 seconds and reported within 12 seconds. It also means
103 that only one bad data value is returned if the serial port
104 is disconnect. Previously, all data values after that point
105 were incorrectly reported.
106 - I moved the flow documentation from the bottom of apcupsd
107 to the top.
108 - I wrote new code to put apcupsd in daemon mode. This was added
109 to apcupsd.c at the end as subroutine daemon_start(). Although
110 the code is a good deal more complicated than the original code,
111 in principal, it now completely detaches from the control
112 terminal for all operating systems.
113 - I modified the cgi code to use the apcnetlib code in the
114 main source directory.
115 - The search string for the output voltage in upsfetch.c was
116 incorrect causing the output voltage bar graph not to display.
117 - I optimized upsimage.c reducing the volume of code significantly
118 by moving it to subroutines.
119 - I made the changed necessary to configure.in so that it will run
120 correctly on a Solaris system. These changes were provided by
121 Carl.
122 - I added sun and hpux to the system recognized by configure.
123 - The apccontrol script always executed any user script as well
124 as executing the default behavior. I modified the code so that
125 if the user script is executed and it returns a 99 as a return
126 code, the default action will not be executed.
127 - In the RedHat version of the apcupsd, I removed the killpower
128 option as it is not used, and it is rather dangerous.
129 - I noticed that on a full shutdown, the system would sometimes
130 automatically reboot (I'm not sure exactly why). As a consequence,
131 I modified the RedHat halt scripts to wait two days after a
132 power kill and then to exit.
133 - Updated the RedHat and SuSe apccontrol.sh.in script to have many
134 of the changes necessary for running on the Sun. These are minor
135 script changes that will make it easier for other systems to
136 take our scripts and modify them for their use.
137 - Eliminated the /sbin/powersc from the unknown directory and
138 replaced it with a call to apccontrol.
139 - Deleted apcstatus.man from the doc directory.
140 - Added a new doc/manual directory containing the new html documentation.
141 Lots of new files.
142 - Updated apcupsd.conf to have the new configuration directive
143 names and added more documentation for the TIMEOUT directive.
144 - Added example STATUS output files to the examples directory.
145 It would be nice to have a STATUS output example for nearly each
146 UPS type.
147 - Updated examples/apcupsd.master.conf and slave.conf.
148 - Added a safe.apccontrol to the examples directory for testing.
149 See the Testing section of the new manual for details.
150 - Added startapcupsd and stopapcupsd examples to the examples directory.
151 - Added the necessary changed to apc_config.h for the Sun compler --
152 thanks Carl.
154 Kern's ToDo List
156 To do:
157 - Complete documentation
158 - Add and test a bunch of events that email a message.
159 - Check and double check killpwr changes (one pass made).
160 - Fix the *** fixme ***s in apcupsd.c and apcconfig.c
161 - Expand Last UPS Self Test field in cgi program
162 - Automatic conversion of old apcupsd.config files
163 to the new format?
164 - Produce a RPM for RedHat
167 Wish list:
168 - Add remaining time before TIMEOUT to STATUS output.
169 - Add more commands (individual variables) to apcnetd
170 - Accumulate time on batteries and number of transfers
171 to batteries. Perhaps save history in file so that the
172 info can be recovered if apcupsd restarts.
173 - Fix apcupsd so that it respawns the server if it
174 dies (limit number of times). This will avoid the
175 possibility that someone can bring down our apcupsd
176 by connecting via Internet (denial of service attack,
177 or exploit possible buffer overflow).
178 - Make apcaccess use the network code as an option.
179 - Remember date and time when apcupsd started.
180 - Eliminate rest of character command codes using new
181 capabilities code in apcsetup.c (for setup stuff).
182 - Eliminate the rest of the printfs().
183 - Eliminate as many error_aborts as possible in making
184 new events.
185 - Possibly retab new cgi/net server code
186 - Apparently during self test, apcupsd thinks that the
187 power was lost. Distinguish this condition!
188 - Check out apmd and see if we should interface to it.
190 Done:
191 - Integrated changes for Solaris provided by Carl.
192 - Add TIMEOUT to STATUS output.
193 - Document TIMEOUT in apcupsd.conf
194 - Inhibit DATA records from being written to the
195 events file.
196 - Fix the battery voltage display in the cgi to handle
197 battery voltages of 12 and 48 volts as well as
198 24 volts.
199 - Document the new configuration options.
200 - Document installation.
201 - Update apcupsd.man and apcnetd.man, and create
202 new sgml documentation for EVENTS, DATA, LOGGING,
203 and STATUS output as well as documenting the
204 network "programming" interface.
205 - Document the CGI programs (partially).
206 - Document log_event for developers -- especially
207 the LOG_ values.
208 See info in developers/apcupsd.logging