Florian Forster [Sat, 28 Feb 2009 09:29:08 +0000 (10:29 +0100)]
dns plugin: Don't pass a NULL pointer to `pcap_open_live'.
Although the documentation states clearly, that passing a NULL pointer
as device is okay and handled like "any", doing so will crash the daemon
on some systems, most notably *BSDs but Linux users have reported this
behavior, too.
This patch passes "any" when the pointer it NULL, which reportedly
resulted in a different behavior, but still crashing the daemon. We'll
keep trying ;)
Sebastian Harl [Tue, 24 Feb 2009 10:27:22 +0000 (11:27 +0100)]
src/common.c: Fixed a race condition in check_create_dir().
Between checking for the existence of a directory using stat() and creating
the directory using mkdir(), another thread might have already created the
directory thus causing mkdir() to fail with errno == EEXIST. This case is now
handled sanely, no longer causing the function (and thus some write callback)
to fail in this case.
Most likely, this only happens during startup when creating the data
directories - later, no two threads should try to create the same directory.
Interestingly enough, I frequently encountered this issue on a single core
machine.
Sebastian Harl [Thu, 19 Feb 2009 11:09:46 +0000 (12:09 +0100)]
contrib/cussh.pl: Fixed and improved command parsing.
The input line is now split into separate tokens which are either quoted or
unquoted strings. This simplifies e.g. the parsing of identifiers as the whole
token may be interpreted as just the id string. This allows for specifying a
somewhat greedy regex which before led to the whole remainder of the input
line ending up in the type or type instance.
Bruno Prémont [Tue, 10 Feb 2009 20:38:23 +0000 (21:38 +0100)]
contrib/php-collection: Save/load graph list and fix HTTPS handling.
Hi,
An update to the graphing interface. (incremental patch attached)
This fixes security complaint by a few browsers when page is accessed
via HTTPS and also makes sure the graphs and lists will load if HTTP
and HTTPS paths differ.
Thanks to Mullet- for spotting this issue.
In addition to the fix above this patch adds support for saving
current graph list to a cookie and appending graph list read from
cookie to the currently displayed list.
A future extension would be to allow saving/loading named graph lists
so multiple list can coexist (pretty useful when monitoring multiple
servers/services)
Bruno Prémont [Tue, 10 Feb 2009 20:35:44 +0000 (21:35 +0100)]
contrib/php-collection: Add a PHP frontend for graph generation.
Hi,
Attached is a patch with a set of PHP files for a complete graphing
environment for collectd-generated RRDs.
Before generating a graph with rrdtool it can tell collectd to flush
the RRDs that are about to be used.
The interface is built with dynamic HTML.
It provides following options:
- host selection
-> plugin selection
--> plugin instance selection
---> type selection
----> type instance selection (or meta graph)
- linear / logarithmic Y-scale
- verbose / minimal legend
- [Add Graph] [Remove all Graphs] [Refresh Graphs]
For each displayed graph:
- Move above previous graph
- Refresh graph
- Remove graph
- Move below following graph
I tested on following browsers:
- Firefox-3.0
- Safari-3.2 (Win32)
- Konqueror (KDE-4.1.3)
- Webkit (webkit-gtk-0_p40220)
- Internet Explorer (6, 7, 8rc - CSS layout issues with <8)
Dependencies:
- PHP-5 (might run with PHP-4)
> GD suport for error images
> Ability to execute rrdtool binary
> Unix socket for FLUSH support
- RRDTool (rrdtool graph, rrdtool info)
TODO: complete/improve graph definitions in definitions.php
though there is code to generate basic graph for any RRD of
unknown type, so definitions are rather a matter of color,
DS combination and stacking.
Bruno
File listing with short description:
- config.php
(configuration)
- functions.php
(common functions)
- definitions.php
(graph definitions for most? types from types.db
- based on collection.cgi)
- definitions.local.php
(place for site-local graph definitions, e.g. for
unixsock, tail, snmp generated RRDs)
- index.php
(main page)
- graph.php
(page returning the graph's PNG image)
- browser.js
(whole bunch of Javascript logic to show/hide/update graphs)
Not included are a few images:
- collectd-logo.png
(16x16, e.g. use collectd.org's favicon)
- favicon.png
(e.g. use the one in share/collection*)
- refresh.png
- move-up.png
- move-down.png
- delete.png
(16x16 take matching ones from your system's action-icons)
Florian Forster [Sat, 24 Jan 2009 10:00:30 +0000 (11:00 +0100)]
tcpconns plugin: Fix an endianness problem under *BSD.
The port numbers were assumed to be in the host's byte order, when in
fact they are in network byte order (big endian). The patch adds `ntohs'
where necessary to fix this problem.
Doug MacEachern [Tue, 20 Jan 2009 09:27:10 +0000 (10:27 +0100)]
contrib/cussh.pl: Add the PUTNOTIF command.
I was using this to test notifications for jcollectd like so:
echo "PUTNOTIF host=foo severity=warning message=my perl is rusty" |
perl -Mblib=bindings/perl contrib/cussh.pl
Florian Forster [Tue, 30 Dec 2008 13:44:39 +0000 (14:44 +0100)]
network plugin: Fix an invalid size of buffer being used.
When specifying the amount of data to copy, we used `sizeof (buffer)' where
`buffer' is a pointer, giving 4 or 8 bytes, depending on the architecture (and
not depending on the actual buffer size). This results in the `type' being sent
much more often than necessary and sometimes not sending a new type when is was
actually necessary. The only prominent case in the default configuration(s) was
`cpufreq' being used instead of `cpu', though.
While in the process, the global `type' buffer was replaced, because an
appropriate buffer is in `value_list_t' now.
Much thanks to Bruno Prémont for reporting and debugging this issue :)
Sebastian Harl [Fri, 19 Dec 2008 16:00:56 +0000 (17:00 +0100)]
collectd.conf(5): Updated the section about the "mysql" plugin.
* Do not talk about what RRD files things end up in - there is more than
"rrdtool" by now.
* Updated the section number of the "Server Status Variables" in the MySQL
reference manual.
* Mention that the MySQL user does not need any special privileges.
If a registered config key equals NULL, it is now ignored. Before, this would
cause a segfault. As this has happened a few times in the past, we not handle
it sanely.
Sebastian Harl [Fri, 5 Dec 2008 16:11:16 +0000 (17:11 +0100)]
processes plugin: Fixed a segfault when handling unknown config options.
When using STATIC_ARRAY_SIZE() to determine the number of config keys, the
list must not include NULL. Else, the configfile module will handle that as
another config key and dereference it in case no previous key matched, i.e.
when handling unknown config options.
Florian Forster [Wed, 3 Dec 2008 21:56:18 +0000 (22:56 +0100)]
configure.in: Check for `kvm_openfiles'.
Check for both, `kvm_nlist' and `kvm_openfiles' before enabling the
tcpconns plugin with libkvm: Solaris has a libkvm, too, but doesn't
provide the necessary functions to read the connections with it. At
least not in the same way.
Florian Forster [Wed, 3 Dec 2008 21:07:24 +0000 (22:07 +0100)]
configure.in: Some improvements for Solaris, expecially the swap plugin.
If Solaris is detected, "-D_POSIX_PTHREAD_SEMANTICS" will be added to
CPPFLAGS automatically now.
If sys/swap.h is not found, print a message suggesting to build a 64bit
binary. (sys/swap.h cannot be included when a 32bit binary is being built,
but it's most likely present under Solaris.)
Enable the swap plugin only if sys/swap.h has been found (when under
Solaris).
Sebastian Harl [Sat, 1 Nov 2008 10:28:30 +0000 (11:28 +0100)]
configure, mysql plugin: Check for mysql.h as well.
Up to now, only mysql/mysql.h has been checked for. However, mysql_config
--cflags usually adds the complete path to mysql.h to the include flags. In
most setups, mysql/mysql.h can be found in the search path as well (usually
the header is available in something like /usr/include/mysql/mysql.h) so this
issue has not been found so far. However, if that's not the case, the build
will fail.
Thanks to Dusty Doris <collectd@dusty.name> for reporting this.
Sebastian Harl [Thu, 30 Oct 2008 15:27:37 +0000 (16:27 +0100)]
filecount plugin: Fixed an off-by-one error in the configuration code.
The memory allocated to store the information about configured directories did
not take into account the newly added directory. This would cause a segfault
at some point of time (when specifying more than four directories in my case).
Thanks to Martin Janota <janota.m@cce.cz> for finding and reporting this bug.
Florian Forster [Thu, 16 Oct 2008 20:04:46 +0000 (22:04 +0200)]
src/collectd.h: Poison various insecure string functions only when debugging is enabled.
Because the `poison' pragma complains about macros (which are then never
used), this feature does more harm than good in everyday use. So we'll
only activate it when building in debug mode, so hopefully authors of
new plugins will use this.
Sebastian Harl [Tue, 14 Oct 2008 11:41:18 +0000 (13:41 +0200)]
postgresql plugin: Do not open new connections during reinitialization.
When reinitializing the plugin a new connection to all databases has been
opened. Now, we check if the connection already exists. In that case no new
connection will be opened but the existing connection will be pinged instead.
Thanks to Admin <collectd-info@internode.com.au> for finding and reporting
this.
Signed-off-by: Sebastian Harl <sh@tokkee.org> Signed-off-by: Florian Forster <octo@huhu.verplant.org>
Sebastian Harl [Wed, 8 Oct 2008 13:46:39 +0000 (15:46 +0200)]
types.db: Reintroduced the "spam_check" type.
This is used by the "email" plugin. Apparently, it got lost when converting
the plugin to the v4 interface: Instead of "spam_check", "email_check" had
been added to types.db. I did not remove "email_check" though as this would
introduce a regression - though very unlikely, somebody out there might be
using that type.
Signed-off-by: Sebastian Harl <sh@tokkee.org> Signed-off-by: Florian Forster <octo@huhu.verplant.org>
Bruno Prémont [Sun, 5 Oct 2008 12:09:08 +0000 (14:09 +0200)]
configure.in: Let configure bail out on missing dependencies
The patch below adds check in AC_PLUGIN() that verifies if the
dependencies are met for any enabled plugin.
In case some dependency is missing (default is disabled and enabled
is selected) configure will fail after dumping summary so all
dependency issues are visible in a single configure pass.
A later addition would be to list requirements (dependencies, OS/Kernel
restrictions) for all plugins in an easily accessible location (e.g.
INSTALL file)
Florian Forster [Mon, 6 Oct 2008 09:46:59 +0000 (11:46 +0200)]
src/collectd-nagios.c: Improve handling of lines returned from `GETVAL'.
Then selecting only certain data sources with the `-d' option, the program
would try to read more lines from the server than the server reports, resulting
in the communication never finishing.
This patch resolves that problem and introduces a couple other error handling
blocks.
Thanks to Fabian Linzberger for pointing this out.
Sebastian Harl [Thu, 2 Oct 2008 13:25:10 +0000 (15:25 +0200)]
collectdmon: Do not block SIGCHLD.
There is no reason to do so as we do not need to synchronize any waitpid()
calls. In fact, doing so was wrong because waitpid() should then return with
an error according to POSIX. This was still working so far since waitpid()
would not return until all children terminated and this is all we really care
about.
Signed-off-by: Sebastian Harl <sh@tokkee.org> Signed-off-by: Florian Forster <octo@huhu.verplant.org>
Sebastian Harl [Thu, 2 Oct 2008 12:53:06 +0000 (14:53 +0200)]
iptables plugin, utils_ignorelist: Fixed an off-by-one error each.
Those were introduced when unifying the string handling in commit 5f9ec13b in
cases where the exact length of the string to be copied is passed to sstrncpy
instead of the size of the destination buffer.
In case of the iptables plugin this prevented the table or chain name to match
correctly as the user configuration was truncated. In case of the ignorelist a
given regex was truncated.
Signed-off-by: Sebastian Harl <sh@tokkee.org> Signed-off-by: Florian Forster <octo@huhu.verplant.org>