Robot/contrib-munin
2
0
Fork 0
mirror of https://github.com/munin-monitoring/contrib.git synced 2018-11-08 00:59:34 +01:00
Code Issues Releases Wiki Activity

Fix for https://github.com/munin-monitoring/contrib/pull/826

Browse Source
This commit is contained in:
K.Cima 2017-03-27 16:23:33 +09:00
parent 2590132015
commit fe097689f9
1 changed files with 27 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

41
plugins/solaris/io_disk
View File

@ -9,10 +9,11 @@
These functions are implemented:
ops : similar to iostat r/s, w/s
bytes : similar to iostat kr/s, kw/s
busy : similar to iostat %b, %w
busy : similar to iostat %b, %w (%w usually indicates 0 in Sol10,11)
queue : similar to iostat actv, wait
latency : similar to iostat asvc_t, wsvc_t
This plugin is improved from Solaris io_busy_, io_ops_, io_bytes_ plugins.
Any device found in /usr/bin/kstat can be monitored.
Tested with Solaris 10 and 11, and it should work with Solaris 8, 9.
@ -30,19 +31,26 @@
By default, this plugin monitors disk devices. And also it can monitor
NFS and Tape devices as much as io_* plugins with setting environments.
Note that instance names of nfs (e.g. nfs1) can be changed after reboot or
remount, this plugin may not monitor nfs properly. (same as original ones)
=head1 ENVIRONMENT VARIABLES
env.class - kstat class. See man kstat -c option.
example: env.class /disk|nfs|tape/
default: disk
env.module - Module name. Only used in internal graph name. If you would not
to like to take over RRDs generated by io_*_sd plugins, please change it
to another name.
env.module - Module name. Only used in internal graph name.
example: env.module something
default: sd
If you would not like to take over RRDs generated by io_*_sd plugins,
please change it to another name.
And also, if you would like to take over RRDs of nfs, tape devices,
please set it 'nfs' or 'tape' (and set env.class, env.module_regex).
It may be good to link like ln -s /path/to/io_disk io_nfs if necessary.
env.module_regex - kstat module. See man kstat -m option.
example: env.module_regex /^(s?sd|vdc|zvblk|dad|md|nfs|tape)$/
default: /^(s?sd|vdc|zvblk|dad|md)$/
@ -187,16 +195,15 @@ do_config() {
echo "graph_width $graph_width"
fi
# Get device instance names by kstat
kstat -p -c "$class" -m "$module_regex" -s "/^${conf_in}\$/" | sed 's/:/ /g' | awk '{ print $3 }' | \
while read -r dev
do
if is_excluded "$dev" ; then
continue
fi
is_excluded "$dev" && continue
# replace instance name to logical device name if exists
# Replace instance name to logical device name if exists
ref="name_$dev"
devname=${!ref:-}
devname=${!ref:-} # ${!varname} means indirect evaluation (similar to eval)
# reads and writes are necessary to calculate latency
if [ "$func" = 'latency' ]; then
@ -244,12 +251,18 @@ do_getvalue() {
echo "multigraph ${plugin_name}_${func}_${module}"
# Get device instance names, stat names and values by kstat
# kstat output example:
# $ kstat -p -c disk -m '/^sd$/' -s '/^reads$/'
# sd:0:sd0:reads 51432610
# sd:1:sd1:reads 52671435
# ...
kstat -p -c "$class" -m "$module_regex" -s "$stat_regex" | sed 's/:/ /g' | awk '{ print $3,$4,$5 }' | \
while read -r dev stat value
do
if is_excluded "$dev" ; then
continue
fi
is_excluded "$dev" && continue
echo "${dev}_${stat}.value ${value%.*}"
done
@ -285,7 +298,7 @@ getvalue() {
done
}
# main
# Main
case ${1:-} in
autoconf)
autoconf