CHANGELOG 9.3 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267
  1. 26/02/2003 - 2.59
  2. - mbuf probe (Daniel Hartmeier; depends on 3.2)
  3. - example configuration is now installed in to symon/examples +
  4. better conflicts in port's PLIST (Christian Weisgerber)
  5. 24/01/2003 - 2.58
  6. Gustavo Chamone helped find a rare bug in symon/sm_io.c. strsep()
  7. would clobber stream information and make the first stream in the
  8. configuration file disappear.
  9. Daniel Hartmeier patched symon/sm_pf.c to deal with changes pf. His
  10. patch should also work for previous versions of pf.
  11. 10/01/2003 - 2.57
  12. Christian Weisgerber reported that netinet/ip_ipsp.h already defines
  13. htonq - fixed. Will Wainwright reported 1) a small bug in
  14. test_config.php - fixed, 2) that I left out some files
  15. (class_proc.inc and class_debug.inc) from the PLIST files.
  16. 29/12/2002 - 2.56
  17. Added proc and debug symon2web frontends. Debug has limited
  18. useability, you are better off defining one for yourself.
  19. 14/12/2002 -
  20. Added the proc module. This can be used to retrieve accumulated
  21. information about processes with the same name. A proc(httpd)
  22. statement in symon.conf, for instance, will result in the reporting
  23. of the number of processes called "httpd" and the amounts of cpu and
  24. memory that they take.
  25. Added em device to c_smrrds.sh - thanks to Jung. Fixed typo in
  26. SymuxClient.pm - thanks Tobias Gresch. c_config.sh did not deal with
  27. type mfs devices - thanks Dom De Vitto.
  28. 29/11/2002 - 2.55
  29. Markus Friedl reported that unresolvable ip addresses cannot be used
  30. in the configuration file - fixed.
  31. Overhauled the networking code. Everything is now ip6 aware. This
  32. does mean that the ':' is no longer a keyword in the lexer;
  33. i.e. '<host>:<port>' statements should now be written as '<host>
  34. <port>' or '<host> port <port>'.
  35. Added the debug module. This can be used to retrieve debug0
  36. ... debug19 variables from the kernel.
  37. 08/11/2002 - 2.54
  38. Added a perl module that allows ppl to connect to symux and do
  39. something with the measurements as they come in. (I plan to drive an
  40. LCD using this, but it could also be used to get, say, daily
  41. bandwidth usage)
  42. Clients of symux would not sleep after they had read and relayed data
  43. symux gave to them. This could lead to one child eating up multiple
  44. 'semaphore slots' and starvation in other clients - fixed.
  45. 25/10/2002 - 2.53
  46. Henning Brauer suggested a datadir statement for symux configuration
  47. - added.
  48. Henning also reported that c_smrrds.sh cannot create vlan10 and
  49. up. c_smrrds.sh now has a special case for pseudo-interfaces (bridge,
  50. enc, faith, gif, ppp, sppp, strip, tun, vlan)
  51. Nick Nauwelaerts helped improve dependency definitions in the port
  52. Makefile.
  53. 18/10/2002 - 2.52
  54. Julien Touche reported that symon/symux do not report initial cfg
  55. errors on the commandline - fixed.
  56. Reinhard Sammer, Julien and Henning Brauer all noticed that the php
  57. code needed globals - fixed. Passing by reference at call time
  58. resulted in php warnings - fixed.
  59. Henning thought it was nice to see a single machine at a time. Julien
  60. wanted clickable magnification for individual graphs. Both options
  61. can now be configured in datasources.inc.
  62. Resolved a bug reported by Vincent Kessler (who also provided the fix
  63. - cheers mate) in the rrdupdate call in symux - optind needs to be
  64. cleared ahead of the getopt call in rrdupdate.
  65. The Makefile.inc no longer overwrites CFLAGS. Configuration
  66. directories are now taken from the environment as expected by the
  67. openbsd ports system. (Thanks to Nick Nauwelaerts for pointing this
  68. out)
  69. Installation of symon now also generates a default symon.conf file if
  70. the system does not already have one.
  71. Someone at www.deadly.org pointed out that symon is no longer
  72. compatible with OpenBSD 3.0 - removed claim from the website.
  73. Added -f <file> switch to symon.
  74. The lexer started parsing from character position 1 instead of 0.
  75. symux would bomb out after a hup - fixed.
  76. 14/09/2002 - 2.51
  77. Steve Shockley reported that the port Makefile did not install rrd
  78. tool. (Which is required for compilation of symux)
  79. 13/09/2002 - 2.5 - kvm gone, renamed mon to symon
  80. Rewrote io.c to obtain information via sysctl only. All kvm code is
  81. now gone.
  82. Renamed mon to symon, monmux to symux and mon2web to symon2web. I
  83. liked mon better, but that name is already taken.
  84. Cleaned up webpages to be XHTML 1.0.
  85. 10/09/2002 - 2.4
  86. Added some example pf graphs. This data is hard to plot, ideas are
  87. welcome. Added better port support.
  88. 09/09/2002 - Towards a better port makefile
  89. Reworked port Makefile. It now generates a number of packages that
  90. can be used to just install mon, monmux or mon2web. The KVM stuff has
  91. become a special flavor of the monitor package.
  92. 06/09/2002 - Cosmetics again
  93. monmux/c_monrrds.sh:
  94. - Added RRD_ARGS support. All created rrds have
  95. start=epoch. Incoming data from hosts may be from an earlier time
  96. than rrd creation now.
  97. mon2web/class_[io|if|mem|cpu].inc, index.php:
  98. - Rewrote argument representation of the graphs to be the same as
  99. how it is configured in the configuration files.
  100. mon2web/class_pf.inc, graph_pf.php:
  101. - Initial pf support.
  102. 01/09/2002 - Internal cosmetics - not released
  103. mon/mon.c:
  104. - mon packetsize is still limited by size of struct monpacket in
  105. data.h:79. All other limits have been removed.
  106. lib/data.c mon/monnet.c monmux/muxnet.c:
  107. - header structure is now safely transported across the network.
  108. lib/data.c
  109. - added streamtypes b(yte) and s(hort) in preparation for new
  110. streams.
  111. 31/08/2002 - 2.3
  112. Makefiles, mon/mon.c:
  113. - Environment variable KVM will enable the building of a setgid
  114. kmem version of the monitor. If this environment variable is not
  115. present, a non priviledged monitor will be built. (Note that all
  116. kvm dependant modules will fail during operation)
  117. After today's changes, only the io probe needs kvm.
  118. mon/if.c:
  119. - obtain ifdata as per h1kari's suggestion.
  120. mon/pf.c:
  121. - global naming in sync with other probes
  122. - repeated initialisation would eat up fds
  123. Jan Sipke van der Veen
  124. - version number inconcistencies (main pkg/syslog)
  125. Daniel Hartmeier
  126. - shmat usage in monmux/share.c != 64 bit safe
  127. - some cpp macros missed precendence parenthesis
  128. 29/08/2002 - 2.2
  129. Merged contributions by Daniel Hartmeier:
  130. - cpu probe had a nasty va_arg bug that only showed up on
  131. big-endian machines.
  132. 29/08/2002 - 2.1
  133. Merged contributions by Daniel Hartmeier:
  134. - pf probe
  135. - changes to make mon compile on macppc and sparc64
  136. 28/08/2002 - released 2.0
  137. monitoring behaviour (mon) separated from storage/forwarding
  138. (monmux) and picture display (mon2web). Only mon has to run with
  139. kmem priviledges because of the if and io probes. Although the
  140. entire package was built with portability to other OpenBSD
  141. architectures in mind, it was never tested.
  142. ==== unreleased ====
  143. 25/07/2002 - rewrite stringptr functions in monmux to snprintf ones
  144. - package it for openbsd
  145. - added a crc32 algorithm in netmon and netmux
  146. - added DEBUG handling in Makefiles
  147. - removed a bug in monmux that made it reprocess already processed data.
  148. 22/06/2002 - added shared code, upgraded networking code
  149. 04/04/2002 - changed mon/monmux.conf format
  150. written manual pages
  151. look at mon/cpu.c/percentages, give appropriate people credits.
  152. 01/04/2002 - changed all memsets into bzero
  153. all output messages have the same format now
  154. write pids in /var/run
  155. SIGHUP = reload configuration
  156. 31/03/2002 - cleaned up source
  157. added error reporting routines.
  158. 29/03/2002 - mon and monmux no use monpacket to communicate with each
  159. other. This format defines: version, lenght and crc of the traffic
  160. exchanged.
  161. mon no longer connects. monmux and mon can be started and stopped
  162. independently.
  163. monmux now checks if incoming traffic is allowed from a certain host.
  164. added rrd write support to monmux
  165. made c_monrrds.sh determine what files to build from environment
  166. got mon2web up to date with mon1. changed interface layout to
  167. have output as a negative area - i'm not finished with this.
  168. 22/03/2002 - Added get/set preamble, still tinkering with those s(u)(n)pack
  169. functions. Added one that translates it into text.
  170. 06/01/2002 - Removed init and get functions from the monlib. (init functions
  171. were also required in non-mon applications, not like it is
  172. supposed to be :)
  173. Next step is to rewrite the get functions into using the pack
  174. function.
  175. 05/01/2002 - Wrote spack and sunpack; endianess agnostic values to bytestream
  176. (and back) functions. The packet format has almost been defined
  177. now.
  178. 24/11/2001 - Rewrote lib to KNF, OpCodes removed
  179. 03/11/2001 - Added datatypes and rewrote readconf to use SLISTs.
  180. 01/10/2001 - Mon configuration is read from a mon.conf.
  181. 29/09/2001 - Lexer had trouble dealing with ip-addresses. Cleaned up the number
  182. parsing code and removed a second comment reader.
  183. $Id: CHANGELOG,v 1.17 2003/01/26 20:40:46 dijkstra Exp $