Changelog
sinfo (0.0.33-3) unstable; urgency=low
* Add myself as an uploader
* Remove unnecessary initrscipt dependency on ssh (Closes: #548737)
* Change default broadcast address to upstream default
sinfo (0.0.33-2) unstable; urgency=low
* Upload to Debian main (Closes: #547350)
* Add missing build dependencies
* Improve package description
* Link against libboost-signals instead of libboost-signals-mt
* Make debconf templates translateable
* Add missing information to debian/copyright
* Change cgi debconf question from critical to high
* Improve debconf template text
* Add ${misc:Depends} to sinfo binary package
* Increase debhelper compat level to 7 (no changes)
* Remove unneeded manual init script handling
* Remove --quiet option from sinfo.init
* Add versioned dependency on debhelper
* Correct purging and deinstall CGI link
* remove unneeded dh_* commands
* Remove shlibs file. This is not a shared library.
* Conflict with slurm-llnl which also contains /usr/bin/sinfo
* Replace dh_clean -k by dh_prep
* Simplify debian/rules install target
* Install addons as examples under /usr/share/doc/sinfo
* Register HTML documentation with doc-base
sinfo (0.0.33-1) unstable; urgency=low
* new release
sinfo (0.0.32-1) unstable; urgency=low
* new release
sinfo (0.0.31-1) unstable; urgency=low
* new release
sinfo (0.0.30-1) unstable; urgency=low
* new release
sinfo (0.0.29-1) unstable; urgency=low
* new release
sinfo (0.0.28-1) unstable; urgency=low
* new release
sinfo (0.0.27-1) unstable; urgency=low
* new release
sinfo (0.0.26-1) unstable; urgency=low
* new release
sinfo (0.0.25-1) unstable; urgency=low
* new release
sinfo (0.0.24-1) unstable; urgency=low
* new release
sinfo (0.0.23-1) unstable; urgency=low
* new release
sinfo (0.0.22-1) unstable; urgency=low
* new release
sinfo (0.0.21-1) unstable; urgency=low
* new release
sinfo (0.0.20-2) unstable; urgency=low
* tnx to some debian hints by T.Koschnick:
* added libncurses5-dev to Build-Depends
* added debconf to Depends of sinfo
* achanged debconf dependency back to 4 (for sarge users)
sinfo (0.0.20-1) unstable; urgency=low
* changed sinfo start/stop to rc2.d/S80sinfo
rc0,6.d/K10sinfo
(purge and install newly, since I don't force this change)
* switched to debhelper 5
* split up debian and source changelogs
* new release
sinfo (0.0.19) unstable; urgency=low
* fixed FreeBSD 5.3 compilation, tnx S.Buys
* added --help option to sindo and sinfod
sinfo (0.0.18) unstable; urgency=low
* debian package: added a RUN_SINFO option in /etc/default/sinfo
respectively /etc/init.d/sinfo (suggestion by M.Feuersaenger)
* Removed scripts sinfodrestartall sshallrup etc. from /sbin.
* install sshallsinfo in /bin and add a small manpage
sinfo (0.0.17) unstable; urgency=low
* changes to ensure 32, 64 bit compatibility, avoid long in exchange
structs and use int or float
sorry, but this makes the broadcasting incompatible again
restart your demons
* changed struct users.number from long to int
* changed struct cpustat from long to float (percentage for user,
nice, idle, ...)
* changed struct uptime from long to two ints for days and seconds
sinfo (0.0.16) unstable; urgency=low
* added a PPC970 patch to get the right CPU speed from F.Thomas
* added option -Ls so that the sinfo displays optionally the number of
CPUs and their speed beside the hostnames and the system_load. This
is made to be used by people using sinfo for rudimentary load-
balancing. Remark: you can suppress the headlines (normally going to
stderr) of this function by the command sinfo -qLs.
* changed fscanf format string in cpuinfo.cc to hopefully work with
some dual AMD Opterons
* Added information for the linux 2.6er states iowait + (soft/hard)irq
and broadcast them. The changes the format of the UDP packets, so
restart all your demons.
sinfo (0.0.15) unstable; urgency=low
* added a monitor for disk activity (linux 2.4.x only)
* in deb package;changed /etc/init.d/sinfo configuration to
file /etc/default/sinfo
* corrected swap and ram in html output
sinfo (0.0.14) unstable; urgency=low
* added output of the number of CPUs
sinfo (0.0.13-1) unstable; urgency=low
* updated to debian Standards-Version: 3.6.0
sinfo (0.0.13) unstable; urgency=low
* changed bar length from linear to logarithmic scaling based on the
minimum and maximum processor speeds
use 'c' to toggle between "log", "lin" and "full".
(I've added this function because of the big speed differences
in our local pc pool.)
* increased the buffer size for reading /proc/net/dev
* updated man-page
sinfo (0.0.12) unstable; urgency=low
* some adaptations for the FreeBSD 5.x series:
the layout of struct kinfo_proc changed from FreeBSD 4 to 5
sinfo (0.0.11) unstable; urgency=low
* some fixes for gcc-3.2 on FreeBSD
sinfo (0.0.10) unstable; urgency=low
* enable-SIMPLE_USER_CACHE for debian binaries
* renamed sinfod.1 to sinfod.8
* changed interface to get memory informations (linux system)
* Added a broadcast-only mode to sinfod. This may be useful, if you
have some slave/server computers where nobody works interactively.
This is a very robust mode because much functionality of sinfod is
deactivated (collection of UDP packets, server for connection of
sinfo). You may activate this with "--spymode". I've selected this
name to remind you not to uses this mode on computers that are used
interactively - probably your users will not like it.
* support for really big memories and swap space => changed broadcast
format
* changed default nice value of sinfod from 15 to 10
* changed rsh to ssh in sinfodrestartall
sinfo (0.0.9) unstable; urgency=low
* first debianized release
* removed rshall and rshallrup scripts because rsh normally is a
symbolic link to ssh on debian systems
* added long options for gnu/linux and FreeBSD
* added an option for --quiet startup of sinfod
* displaying IP number of a host in the case that there is no name
* some adjustments for gcc 3.2
* added a -B --bindaddress option to the sinfod to send UDP packets
with a fixed IP address. (useful on systems with many network
interfaces)
* added a small (fork, watchpid) watchdog that restarts the sinfod if
something goes wrong.
This can be activated at compile time with --enable-WATCHDOG
I've added this because some months ago there was a problem with the
LDAP library that made some programs (sinfod) to crash in the case
of LDAP-server failures.
* added --accessaddress, --accessnetmask and --accesslocalhost to
restrict the access to the sinfod
* fixed process percentage calculation on solaris
* changed memory variables from long to unsigned long so that a
maximum of 4GB can be represented (only tested on linux)
sinfo (0.0.8) unstable; urgency=low
* add a timeout within select() to reduce the system load of
"konsole" when displaying sinfo's ncurses output.
* do not install the sample cgi-script
but keep it in ./cgi
* increased value: drop hosts from being displayed if they were
not heard for 25 seconds (was: 10 seconds)
* fixed some possible security holes using found by
the code inspection tool "flawfinder"
http://www.dwheeler.com/flawfinder/
* added broadcasting of the name of the network interface
=> broadcasted struct changed
=> you have to restart all demons if you update
* fixed query of two separate networks with one "sinfo"
* security fix for the cgi-script:
"it was possible to start programs as cgi-user nobody"!
... see the cgi-script for details
sinfo (0.0.7) unstable; urgency=low
* minor bugs fix:
sinfo segfaults if there isn't an "export HOSTNAME"
* added HTML-output and a sample perl-cgi-script to read
sinfo's informations over the web.
sinfo (0.0.6) unstable; urgency=low
* added a parameter to set the nice value of the demon.
* added a simple cache for lookups of the names belonging to the
UIDs in lib/procinfo.cc
... activate with: ./configure --enable-SIMPLE_USER_CACHE
sinfo (0.0.5) unstable; urgency=low
* minor bugs fixed that prevented static linkage of sinfo...
(you can produce a static version of sinfo if
you "export CXXFLAGS=-static" before ./configure )
* minor bug fix for RedHat 7.1 (and other systems?)
added #include <string.h> in cpuinfo.cc
tnx "Eric Gravelle" <email address hidden>,
"Alan Wilter Sousa da Silva" <email address hidden>
* done a port to FreeBSD ( we got a new server at work :-) )
* changed format of the broadcast scheme... now: less overhead
(incompatible with older versions of sinfo!)
* broadcasting is done in a smarter way:
* If the CPU-Usage doesn't change (variation less 3%)....
the broadcasting frequency will be decreased
down to 6 Bcast per minute.
* If the CPU-Usage changes dramatically
(variation greater than 30%)
the TOP5 process list will be broadcasted right away.
sinfo (0.0.4) unstable; urgency=low
* Upper status line excluded from scrolling.
* Connection scheme from sinfo to sinfod changed.
A TCP connection will be kept for all transmissions.
* Sinfo can use connections to multiple sinfods and join the
data.
This may be useful if you want to monitor two different
networks with only one sinfo.
i.e.: sinfo -h host_on.first.computer.net -h host_on.second.computer.net
sinfo (0.0.3) unstable; urgency=low
* all internal errors of the information gathering functions
getCpuinfo, getNetload, ...
result in a "false" return value...
no more exit(1)s there!
* SIGPIPE ignored to make the demon more robust!
* System info: RAM and Swap - Space added
sinfo (0.0.2) unstable; urgency=low
* broadcast format changed....
=> update on all workstations necessary
* solaris port:
broadcast cpu-usage
* sinfo: highlight local workstation
* sinfo: all hostnames sorted alphabetically
* direct setting of the -h IP-address added
* broadcast CPU-Information
* scale bargraphs according to cpu-power
* first installation on our research cluster for
all users under /usr/local!
sinfo (0.0.1) unstable; urgency=low
* first local installation
-- Ubuntu Archive Auto-Sync <email address hidden> Thu, 19 Nov 2009 23:23:02 +0000
