Marc Fournier [Fri, 14 Nov 2014 21:04:16 +0000 (22:04 +0100)]
write_redis: avoid passing a float/double to redisCommand()
... as it seems to not be well supported by hiredis 0.10.1 on Debian
7.0, leading to a segfault. Storing the string representation in a
variable instead is the compromise I found to make the plugin work on
this system.
Marc Fournier [Mon, 10 Nov 2014 06:58:13 +0000 (07:58 +0100)]
write_redis: fix format of commands sent to redis
The commands getting submitted to redis now look like this:
"ZADD" "collectd/hostname/entropy/entropy" "1415602051.335973024" "1415602051.335973024:823"
"SADD" "collectd/values" "hostname/entropy/entropy"
... which is the same as in the initial implementation, except for the
added decimals in the timestamp (the plugin was developped before
high-precision timestamps support was added to collectd).
When running f3706b0b87, the following command gets sent to redis:
"ZADD" "collectd/hostname/entropy/entropy" "1415487432.000000" "1415487432:932"
Meaning the value actually stored, and later returned by redis is:
"<timstamp>:<value>".
b7984797 accidentally dropped the comma separating the timestamp and the
value, which leads the plugin to store a somewhat confusing value in
redis:
"ZADD" "collectd/hostname/entropy/entropy" "1415487432.000000" "1415487432932"
Andrés J. Díaz [Thu, 7 Nov 2013 08:57:53 +0000 (09:57 +0100)]
Switch redis.c plugin from credis to hiredis.
Change the entire redis.c plugin to use libhiredis (tested with
libhiredis0.10) instead on credis. The libhiredis is supported in a number
of distributions like Debian or Ubuntu.
This patch keeps the same functionality that the old redis.c does.
Patrick Browne [Tue, 30 Sep 2014 08:31:05 +0000 (10:31 +0200)]
feat(python-plugin): prepend instead of append ModulePath to sys.path
When importing a module, you expect the plugin repertory to have higher priority
than your site-packages. It will prevent hard to debug clashing module names problems.
At least, if there is a problem, it will fail more loudly than "module did not register
any config callback"
When using ASCII protocols, such as the Graphite protocol, doubles are
converted to strings using sprintf(3) and parsed with strtod(3). If the
effective locale uses something different from a dot (".") as the
decimal separator (e.g. Portuguese, German), this breaks.
This patch sets the LC_NUMERIC locale to "C" on start-up, so that
strings are produced and parsed in a well-defined manner.
cpu plugin: Various, random fixes and improvements:
* Add Pierre-Yves to the list of copyright holders.
* Make sure the entire file is consistently indented.
* Renamed the "ReportActive" config option to "ReportByState"
(with negated logic).
* Renamed internal functions to better reflect their use.
* Let cpu_stage() (née submit()) work on individual counters / derives
rather than an array, removing logic from the cpu_read() function.
* Rename CPU_SUBMIT_* to CPU_STATE_*.
* Move legacy behavior it its own function, cpu_commit_without_aggregation().
swap plugin: Make SwapCached optional under Linux.
OpenVZ doesn't provide the line, which caused swap collection to be
skipped entirely on these systems. Instead, "cached" should simply be
ignored, as it is on other systems. Thanks to @zerkms for reporting this
problem!
openvpn plugin: Don't signal an error when no clients are connected.
In the multi1_read() function, an error (zero) was returned when no
clients were currently connected to the OpenVPN server, because the
"read" variable was initialized to zero and the while loop exited before
it was set to one. This is not the intended behavior.
John Ferlito [Sat, 13 Sep 2014 01:24:14 +0000 (11:24 +1000)]
Add ReconnectInterval option to Network plugin
The Network plugin only performs DNS resolution at initialization. This
can be problematic when trying to performs migrations of collectd
infrastructure or when trying to create HA solutions which are dependant
on DNS.
The ReconnectInterval options forces a reconnect of all the sockets at
the defined number of seconds. By default no re-connections are attempted
if this option is not set.
battery plugin: Import code to read metrics from sysfs.
This started as a simple import of Andy Parkins' sysfsbattery plugin into
the battery plugin. Since the battery plugin is ancient and hasn't been
touched in a while, this quickly escalated to a much bigger refactoring.
Sorry!
On the other hand, this fixes a couple of bugs. For example, all metrics
were always dispatched with plugin_instance "0". This is correct for the
majority of laptops, of course, but in theory this could be wrong.
Also ACPI charging / discharging rate is reported as "current", when
modern batteries actually report "power". The sysfs code does this
correctly, ACPI still needs to be patched.
src/configfile.c: Fix the default interval setting.
Also fixes a bug in global_option_get_time(): Values smaller than or
equal to zero are illegal and the default value should be returned.
Thanks to @anandkarthik for spotting this!
processes plugin: Limit the buffer size used for command line matching.
ARG_MAX is quite big on many systems, for example >100 kByte on
GNU/Linux. This is a problem for systems with tight memory constraints,
for example embedded devices.
This patch uses at most 4 kByte for this, which out to be enough for the
vast majority of users. Users with specific requirements can compile
with "CMDLINE_BUFFER_SIZE=${LOTS}" in their CPPFLAGS to override this
default.
This moves the socket creation logic so it's called from
networt_send_buffer_plain(). This allows us to recover after network
failures or when collectd was started before the network was available.
This is called from the curl and memcachec plugins. Previously, gauge
average, minimum and maximum values were calculated since the start of
the daemon, rather than per-interval. This is not the documented
functionality.
projects / thirdparty / collectd.git / log
