Hi all,
the version 1.1.0 is finally out. The main innovation is the new
event broker module "MK Livestatus" which allows direct access
to the status of all Nagios objects like hosts and services via
a UNIX socket. Check_MK Multiadmin now makes use of that interface
instead of status.dat.
MK Livestatus has also been made fit to be used as a backend
for NagVis. The new release of NagVis 1.4.5 has also been
published today - and is the first third party addon to officially
support MK Livestatus.
Check_MK 1.1.0 also has many improvement and bug fixes in other
areas. Here is the complete ChangeLog:
- First version with "Check_MK Livestatus Module"!
- Fixed problems in Windows agent (could lead
to crash of agent in case of unusal Eventlog
messages)
- Fixed problem sind 1.0.39: recompile waitmax for
32 Bit (also running on 64)
- Autoconfiscated MK Livestatus
- Fixed bug in cluster checks: No cache files
had been used. This can lead to missing logfile
messages.
- Check kernel: allow to set levels (e.g. on
pgmajfaults)
- Check ps now allows to check for processes owned
by a specific user (need update of Linux agent)
- New configuration option aggregate_check_mk: If
set to True, the summary hosts will show the
status auf check_mk (default: False)
- Check winperf.cpuusage now supports levels
for warning and critical. Default levels are
at 101 / 101
- New check df_netapp32 which must be used
for Netapps that do not support 64 bit
counters. Does the same as df_netapp
- Symlink PNP templates: df_netapp32 and
df_netapp use same template as df
- Fix bug: ifoperstatus does not produce performance
data but said so.
- Fixed bug in MK Livestatus: StatsGroupBy:
didn't work
- Fix bug in Multiadmin: Sorting according to
service states did not work
- Fix two bugs in df_netapp: use 64 bit counters
(32 counter wrap at 2TB filesystems) and exclude
snapshot filesystems with size 0 from inventory.
- Rudimentary support for monitoring ESX: monitor
virtual filesystems with 'vdf' (using normal df
check of check_mk) and monitor state of machines
with vcbVmName -s any (new check vmware_state).
- Fixed bug in MRPE: check failed on empty performance
data (e.g. from check_snmp: there is emptyness
after the pipe symbol sometimes)
- mk_logwatch -d now processes the complete logfile
if logwatch.state is missing or not including the
file (this is easier for testing)
setup.sh will compile, install and activate
Livestatus per default now. If you do not want
this, please disable it by entering <tt>no</tt>,
when asked by setup.
- Web pages (Multiadmin) now uses MK Livestatus
instead of parsing status.dat. This is much faster.
And that way you get *live* data.
- New Option --paths shows all installation, config
and data paths of Check_mk and Nagios
- New configuration variable define_hostgroups and
define service_groups allow you to automatically
create host- and service groups - even with aliases.
- Multiadmin has new filter for 'active checks enabled'.
- Multiadmin filter for check_command is now a drop down list.
- Dummy commands output error message when passive services
are actively checked (by accident)
- New configuration option service_descriptions allows to
define customized service descriptions for each check type
- New configuration options extra_host_conf,
extra_summary_host_conf
and extra_service_conf allow to define arbitrary Nagios options
in host and service defitions (notes, icon_image, custom
variables, etc)
- Fix bug: honor only_hosts also at option -C
The configuration of 1.1.0 is compatible to that of 1.0.39. The
setup now needs a C++ compiler installed (because of Livestatus).
It's possible to install check_mk without Livestatus - if you
do not want to use Multiadmin.
Greetings and thanks for using Check_MK!
Mathias Kettner
--
__ __ _ __
Mathias Kettner | \/ | |/ / M A T H I A S K E T T N E R
Preysingstr. 74 | |\/| | ' /
81667 München | | | | . \ Linux Beratung & Schulung
089 / 444 09 662 |_| |_|_|\_\ http://mathias-kettner.de
Hello to you all,
just a couple of minutes ago I've uploaded the version 1.1.0beta20
to the download pages:
http://mathias-kettner.de/check_mk_download.html
That version might be the last one before the final release
of the 1.1.0. I'd much appreciate your help, if you could
give the beta20 a try and help to find problems...
I've attached a log of the differences to 1.0.39.
Many greetings,
Mathias Kettner
--
__ __ _ __
Mathias Kettner | \/ | |/ / M A T H I A S K E T T N E R
Preysingstr. 74 | |\/| | ' /
81667 München | | | | . \ Linux Beratung & Schulung
089 / 444 09 662 |_| |_|_|\_\ http://mathias-kettner.de
1.1.0.beta20:
- Fixed problem sind 1.0.39: recompile waitmax for
32 Bit (also running on 64)
- Autoconfiscated MK Livestatus
- Fixed bug in cluster checks: No cache files
had been used. This can lead to missing logfile
messages.
- Check kernel: allow to set levels (e.g. on
pgmajfaults)
- Check ps now allows to check for processes owned
by a specific user (need update of Linux agent)
- New configuration option aggregate_check_mk: If
set to True, the summary hosts will show the
status auf check_mk (default: False)
- Check winperf.cpuusage now supports levels
for warning and critical. Default levels are
at 101 / 101
- New check df_netapp32 which must be used
for Netapps that do not support 64 bit
counters. Does the same as df_netapp
- Symlink PNP templates: df_netapp32 and
df_netapp use same template as df
- Fix bug: ifoperstatus does not produce performance
data but said so.
- Fixed bug in MK Livestatus: StatsGroupBy:
didn't work
- MK Livestatus supports keep alive
- Fix bug in Multiadmin: Sorting according to
service states did not work
- Fix two bugs in df_netapp: use 64 bit counters
(32 counter wrap at 2TB filesystems) and exclude
snapshot filesystems with size 0 from inventory.
- Rudimentary support for monitoring ESX: monitor
virtual filesystems with 'vdf' (using normal df
check of check_mk) and monitor state of machines
with vcbVmName -s any (new check vmware_state).
- Fixed bug in MRPE: check failed on empty performance
data (e.g. from check_snmp: there is emptyness
after the pipe symbol sometimes)
- MK Livestatus is now multithreaded an can
handle up to 10 parallel connections (might
be configurable in a future version).
- mk_logwatch -d now processes the complete logfile
if logwatch.state is missing or not including the
file (this is easier for testing)
- Added missing float columns to Livestatus.
- Livestatus: new header StatsGroupBy:
- First version with "Check_MK Livestatus Module"!
setup.sh will compile, install and activate
Livestatus per default now. If you do not want
this, please disable it by entering <tt>no</tt>,
when asked by setup.
- Web pages (Multiadmin) now uses MK Livestatus
instead of parsing status.dat. This is much faster.
And that way you get *live* data.
- New Option --paths shows all installation, config
and data paths of Check_mk and Nagios
- New configuration variable define_hostgroups and
define service_groups allow you to automatically
create host- and service groups - even with aliases.
- Multiadmin has new filter for 'active checks enabled'.
- Multiadmin filter for check_command is now a drop down list.
- Dummy commands output error message when passive services
are actively checked (by accident)
- New configuration option service_descriptions allows to
define customized service descriptions for each check type
- New configuration options extra_host_conf, extra_summary_host_conf
and extra_service_conf allow to define arbitrary Nagios options
in host and service defitions (notes, icon_image, custom variables,
etc)
- Fix bug: honor only_hosts also at option -C
