Russ Allbery [Tue, 19 Oct 2010 16:48:49 +0000 (09:48 -0700)]
More correctly check and report syntax errors
Print an UNKNOWN status on standard output on syntax errors in all
scripts, rather than reporting the problem only to standard error.
Check that the host to check was specified and report a syntax error
if it wasn't. Thanks, Tobias Wolter.
Russ Allbery [Tue, 19 Oct 2010 16:32:41 +0000 (09:32 -0700)]
Report a warning if the salvager is running
If the salvager is running (such as when started manually with bos
salvage), check_bos now reports a warning stating that, rather than a
critical error showing the auxiliary status line. Reported by Steve
Rader.
Russ Allbery [Sat, 16 Oct 2010 18:29:25 +0000 (11:29 -0700)]
Initial version of distribution Makefile
The Makefile for this package only automates make dist and isn't included
in the package, since the scripts don't require compilation or further
modification and we don't provide an install target.
Russ Allbery [Sat, 16 Oct 2010 18:28:32 +0000 (11:28 -0700)]
Remove unneeded parentheses around the date in version output
The date in the $VERSION string is going to include parens, so don't
add additional parens when outputing the version string (in response
to -v, for example).
Russ Allbery [Fri, 15 Oct 2010 20:21:05 +0000 (13:21 -0700)]
Update coding style, version method, documentation
Update the coding style in a few places where my Perl style has changed
and clean up the scripts a bit. Switch the versioning system from the
CVS $Id$ tags to a variable that will be replaced during make dist (not
yet in place). Proofread the documentation and update it as necessary.
Refer to the new afs-monitor distribution, which these scripts will now
be part of.
Russ Allbery [Wed, 9 Nov 2005 20:43:04 +0000 (20:43 +0000)]
rxdebug -noconn tells us how many waiting connections there are; there's
no need to parse the whole output and study it. Also, it's helpful to
look for waiting connections to the file server, rather than waiting
connections to the local cache manager (doh). Lower the warning threshold
to two from four while we're at it.
Russ Allbery [Thu, 25 Mar 2004 00:55:18 +0000 (00:55 +0000)]
Completely rewritten to avoid keeping any state. The monitor now checks
all of the output from bos status against a set of known-okay regexes and
throws an alert if there's any line in the bos output that isn't okay.
This means that this check will no longer catch a server restart that
successfully completed before the probe ran, but on the plus side it also
won't throw additional errors when the file server has come back up (since
the correct output is still different than the old incorrect output).
Also redid the coding style, added real option parsing, required the
standard Nagios -H option, and added full documentation.
Russ Allbery [Wed, 24 Mar 2004 23:56:49 +0000 (23:56 +0000)]
Extensively reworked to do regular option parsing, support -h and -V
options, support configuration of the critical and warning levels, support
a timeout value, provide a bit of information for okay results, and add
complete documentation.
Russ Allbery [Wed, 24 Mar 2004 20:10:12 +0000 (20:10 +0000)]
Reorganized extensively, simplified the code a little bit, simplified and
shortened the output since Nagios will add the relevant host information
itself on errors, added full documentation, added better options parsing,
and made the output and options comply better with the Nagios plugin
standards. Added a timeout option.
Russ Allbery [Fri, 13 Feb 2004 01:08:27 +0000 (01:08 +0000)]
Don't use -allconn -rxstats. We aren't paying any attention to the
statistics, just the mode of the connection, and -allconn just adds in the
(thousands of) idle connections. Let's assume that an idle connection
can't also be blocked.