ID: 0339
Title: blade_bays blade_blades blade_blowers blade_health blade_mediatray blade_powerfan blade_powermod: fix scan function
Component: Checks & Agents
Level: 2
Class: Bug Fix
Version: 1.2.5i1
Due to a previous precision in the scan function API the scan functions of
the blade_* checks where never successful anymore. That way those checks
were not detected on new devices. This has been fixed. Existing checks
were not affected.
ID: 0360
Title: multipath: Inventory handles non loaded kernel module now
Component: Checks & Agents
Level: 1
Class: Bug Fix
Version: 1.2.5i1
The handling for not loaded multipath kernel module and not existing
multipath.conf was not working in all cases. This has been improved now.
ID: 0359
Title: Fixed exception in event simulator when one match group did not match
Component: Event Console
Level: 1
Class: Bug Fix
Version: 1.2.5i1
ID: 0358
Title: df: now ignores filesystems with a reported size of '-'
Component: Checks & Agents
Level: 1
Class: Bug Fix
Version: 1.2.5i1
Previous versions of the check raised an exception because the '-'
could not be converted via int().
ID: 0338
Title: Introduce time limit on availability queries
Component: Reporting & Availability
Level: 2
Class: Bug Fix
Version: 1.2.5i1
The availability view now imposes an upper limit on the execution of the
underlying livestatus query. The default is 30 seconds. This can be configured
in the reporting options. This limit keeps the system usable in situations
where a query over a long time range would need a very long time to execute.
ID: 0357
Title: megaraid_bbu: Fixed expected state checking
Component: Checks & Agents
Level: 1
Class: Bug Fix
Version: 1.2.5i1
The check was unable to check the hard coded expected states. This
made the check never raise expected alerts.
ID: 0337
Title: New header for limiting the execution time of a query
Component: Livestatus
Level: 1
Class: New Feature
Version: 1.2.5i1
The new Livestatus header Timelimit: sets an upper limit
for the execution time of a query. Specify the timeout as
a number of seconds. The following query will stop after
5000 lines of output. If that takes longer than 30 seconds,
then the query <b>fails with no output</b>:
F+:query.lql
GET log
Columns: message
Timelimit: 30
Limit: 5000
F-:
Please note, that <tt>Limit:</tt> and <tt>Timelimit:</tt> have
a different behaviour when the limit is reached:
<li class=list><tt>Limit:</tt> end the query, output all rows that
have been found so far.</li>
<li class=list><tt>Timelimit:</tt> abort the query with an error,
do not output any lines.</li>
ID: 0336
Title: Limit the number of lines read from a single logfile
Component: Livestatus
Level: 2
Class: Bug Fix
Version: 1.2.5i1
If for some reason you have very large history logfiles with millions of
entries then Livestatus would consume large amounts of memory while parsing
these when you access the <tt>log</tt> table. This could lead to a hanging
system or even a crash.
In order to avoid such situations the number of lines that are read from each
individual file are now limited. The default limit is one million lines. You
can change this default by defining the variable <tt>max_lines_per_logfile/tt>
in your <tt>nagios.cfg</tt>. The following example sets the maximum number
to 20 million lines:
F+:
broker_module=/some/path/livestatus.o max_lines_per_logfile=20000000 ...
F-:
When using the Check_MK Micro Core the configuration can be done
via WATO global setting <i>Maximum number of parsed lines per log file</i>.