A new version of CLX was released on December 27, 1998 It was uploaded
to ftp.funet.fi:/pub/ham/incoming that day. For the very impatient,
the new files were also placed at ftp://ftp.muc.de/people/buettneb/clx.
I hope the archive maintainers will soon find the time to move
the files to the appropriate directories so you will find the files
at the "official" URL:
ftp://ftp.funet.fi/pub/ham/unix/Linux/cluster/clx
Here is the version 4.03 readme file:
-------------------------------------------------------------------------
CLX Version 4.03 released December 27, 1998
Version 4.03 is a patch over all of the previous 4.00, 4.01 and 4.02 versions
or intermediate fixes like 4.0{0,1,2}{a,b,c}. Only binaries
and some scripts have changed over the previous versions so you need *not*
re-install the whole software tree and re-read all the tables this time.
Take a look to the CLX WEB page for latest news and hot fixes:
http://www.lurpac.lancs.ac.uk/g0vgs/clx.html
*** APPLYING THE PATCH ***
If you are installing a new CLX, get the file clx_400.tgz, install
the software per instructions and when it comes to unpacking the
TAR archive, first unpack the clx_400.tgz file and then
unpack the upd_403.tgz file, hereby replacing some files.
If you are upgrading from 4.00{a,b}, 4.01{a,b,c} or 4.02, do the
following:
1. Shutdown your current CLX software but leave the postmaster
running. You may have to become user 'postgres' and then startup
postmaster with 'postmaster -i &'.
Take care that watchdog is deactivated if you are using it.
2. Remove some files from previous installations:
$ rm -rf ~/lib/libclxccl.so.4.*
$ rm ~/doc/sysop/README_* ~/doc/sysop/Errata*
$ rm -f ~/tools/clx_start ~/tools/clx_stop ~/tools/vacuum
3. As root do:
# cd /
# tar xvzf upd_403.tgz
# ldconfig
4. As clx_us do:
$ psql clx_db -c "drop table ar_band"
DROP
$ psql clx_db < ~/db/ar_band.cl
$ psql clx_db < ~/config/ar_band.cd
5. Now you can start the new CLX 4.03 without having to re-read
all the database tables.
$ clx -u
*** CHANGES ***
- clxd now also follows ~/config/adv_txt.default if it is there.
Otherwise ~/config/adv_txt is being used (tnx F6FGZ).
- Clarified the routes.tnt entries in the documentation.
- clean_log fixed to ``chown'' logs to clx_us:users when creating
a new one. A problem ocured with cba.log (tnx SV1AAW).
- Cleaned up the ~/doc/user directories. Now it's easy and
straightforward for the doc maintainers to keep the documentation
up to date and adding a new language (Spanish, Swedish, Slovenian,
any takers?) is easy. Thanks to many contributors, we have now
an English, French, Portuguese and German user manual for you
to print out. They can also be accessed from the CLX WEB page
as HTML documents.
- clx_watchdog redesigned and enhanced in that it now
(a) realizes when VACUUM is running.
(b) checks for a lock file which can be used to control
its activities. When the lockfile is there, watchdog will
never shut down CLX.
(c) has control commands like "SET/WATCHDOG", "SET/NOWATCHDOG"
or clx_watchdog {on|off} to create or remove the lockfile.
(tnx G0VGS for the ideas)
- modified clx slightly to allow running clx without database again
using "clx -u -nopostgres" to start it (tnx EA7WA). That was broken
in 4.02.
- Modified db_maint tool to run clean in batch mode. A new option
"batch" was impemented and the commands to be run in this mode can
be specified with the "batchcommands:" configuration command in
~/config/clx_par (tnx PE0MAR). Cleaned up the db_maint program a
little, added new options and partitioned the menus into sub-menus.
- A missing connect script made CLX crash (tnx PA3EZL, VE9PT and IK5PWJ).
Fixed.
- With 32+ users and frequent connects/disconnects by users and
nodes sometimes a situation would come up where icl_com would produce
a very long string of "PC16" messages and then crash. Fixed this
problem with the help of DK4VW and many users at DB0MDX.
- Some tools (clx_start, clx_stop, reboot and vacuum) were removed
from the ~/tools directory and documentation was cleaned up accordingly.
For regular database maintenance (including vacuum), you should
now use db_maint in batch mode.
- New user commands SHOW/LOCATOR, SHOW/COORDINATES and SHOW/SCATTER
(tnx DK0OG and OE5VRL).
- SET/NOANNOUNCE is now working. PC12 (announcements) are now also
honored with waz_ddd settings. TNX many).
- Fixed problem with clxd in that it was always getting the first
callsign that was entered wrong (ON1AFN, tnx NG7M and Eric for
fixing it).
- CLX now increases number of file descriptors to 3072 at startup time
by changing /proc/sys/kernel/file-max from the ~/tools/startup script.
This helps in occasional situations with many users (~40) and lots of
DB queries.
- Modified SET/LOCKOUT and SET/NOLOCKOUT that it now also locks out
remote nodes (however - not permanently).
- Fixed problem with occasional non-functionality of the SHOW/CBA
callbook lookup script. Increased timeout to wait for command
prompt to 15 seconds. Modified SHOW/CBA script to work with the
1999 Callbook CDROM and made it work with more than one callsign
(again), so you can call it like "SHOW/CBA DL6RAI DJ0ZY" and
get both addresses with a single command.
- Modified read_ak1a so that it now also accepts DL8EBW's nice collection
of VHF/UHF station data and allows a command SHOW/VHF to retrieve it.
- SHOW/DXSTATISTIC now looks a little friendlier carrying understandable
band labels (like "20 m" or "70 cm").
- BBS interface program for incoming BBS connections, calls must be specified
in bbs_lst in cls_par (tnx PE1NMB).
- PC24 is no longer transmitted triggered by Flexnet link check connects
(tnx DK4VW).
- The merge command at CLX initialization no longer affects the merging
node, the spots are no more sent out to other node connections. Also
DXDEDX, DX_LIM and SET/NODX setting are honored. This is for both
DX and WWV spots.
- When a user in an Pavillion node modifies his callsign with the SET/ALIAS
command (Pavillion feature - not implemented in CLX) and then spots
DX, spots may come out like this:
PC11^21202.0^KL2A/W7^13-Sep-1998^2041Z^WAE^EA1AK/7^EA7URA-5^H99
^
|
this slash was offending
Until now, CLX ignored these spots. Now the '/' is changed to '-'
internally (tnx EA7WA).
- repeated PC12's where the sender's call changed SSID are now recognized
(same policy as with PC11's where we had this problem earlier). These
repeats are generated by some faulty Cluster software not yet
identified.
- PC84 are now only forwarded to CLX nodes no longer to Pavillion nodes.
- An off-line logging feature was added to the log_monitor program to
allow analyzing log files after a crash.
*** KNOWN ISSUES ***
The following items are still "under construction" but did not make
it into this release:
- The reply command does still not forward the message to the correct
node. If the original message came in from a remote node, the
answer will not be delevered to that node.
- The bbs_lst can currently not be changed during runtime. This will
be changed in a future release.
- The grayline calculation is slightly wrong. SHOW/SUN and SHOW/GRAYLINE
show different results as the algorithms behind the calculations
are different.
- SHOW/VERSION is broken.
- DX spots with calls like "F/DL5MAE" are currently refused by CLX.
- external commands (SHOW/SUN, SHOW/HEADING) are currently not accepted
in UPPER CASE.
- Locking nodes with clx_adm does work only until cluster_par is
re-read.
Ben, DL6RAI
--
[] clx - PacketCluster Software for Linux
[] c/o Bernhard ("Ben") Buettner, DL6RAI
[] Internet: cpfvclth.iferozao@bessy.de
[] Packet: DL6RAI @ DB0AAB.#BAY.DEU.EU
--
[] clx - PacketCluster Software for Linux
[] c/o Bernhard ("Ben") Buettner, DL6RAI
[] Internet: terhi.victor@logonet.com
[] Packet: DL6RAI @ DB0AAB.#BAY.DEU.EU