Review the following solutions to errors that you may encounter when you run
configure.sh
to configure monitoring.
Elasticsearch Errors
- could not determine matching interface
- Cause: The DNS is probably not setup correctly on this node.
- Solution: Contact your DNS administrator or verify that
/etc/hosts
and /etc/nssswitch.conf
are configured
correctly. The etc/hosts
file should list the host names and the hosts
parameter in /etc/nssswitch.conf
should be set to files dns.
- Failed to create <esdb directory name>
- Cause: There is not enough disk space,
configure.sh
was not
run as the root
user, or the index directory path is not valid.
- Solution: Complete the following steps and then re-run
configure.sh
:
- Verify that there is enough disk space.
- Verify that the file system where the index directory will be created is not
read-only. See Log Aggregation and Storage
for more information on the index directory location.
- If you included the
-ESDB
parameter, verify that the index
directory path uses a valid format:
/<existing_directory1>/<existing_directory2>/<new index
directory>
.
- Failed to resolve hostname
- Cause: The DNS is probably not set up correctly on this node.
- Solution: Contact your DNS administrator or verify that
/etc/hosts
and /etc/nssswitch.conf
are configured
correctly. /etc/hosts
should list the host names if the DNS database is
not updated, and the hosts parameter in /etc/nssswitch.conf
should be
set to files dns
OpenTSDB Errors
- Incompatible asynchbase jar found
- Cause:The version of asynchbase installed on this node is not compatible with
OpenTSDB.
- Solution:Install the correct asynchbase package on each OpenTSDB node. See the
EEP release notes to determine the
compatibility between package versions.
- Failed to install asynchbase Jar file
- Cause: There is not enough disk space, the file system is read-only, or
configure.sh
was not run as the root user.
- Solution: Complete the following steps and then re-run
configure.sh
:
- Verify that you are logged in as the root user or using sudo.
- Verify that there is enough disk space.
- Verify that the file system containing the
/opt/mapr/opentsdb
directory is mounted as read/write.
- Failed to create TSDB tables - need to rerun
configure.sh -R
or run
create_table.sh
as $MAPR_USER
- Cause: On a secure cluster, this issue usually occurs when you run
configure.sh
to configure the nodes to use monitoring without first creating user tickets for the root
user
and the $MAPR_USER
.
- Solution: Complete one or all of the following steps:
- Verify that you have a user ticket for both root and the $MAPR_USER before running
configure.sh
or create_table.sh
. Run
maprlogin print
to verify that you have a user ticket for the data-fabric user and the
root
user. If you need to generate a data-fabric user ticket, run maprlogin
password
.
- Re-run
configure.sh
. For example, configure.sh -R -OT
<comma-separate list of OpenTSDB nodes> -ES <comma-separate list of
Elasticsearch nodes>
- Run
${OTSDB_HOME}/share/opentsdb/tools/create_table.sh
to create
OpenTSDB tables in the mapr.monitoring
volume.
Fluentd Errors
- fluentd service not enabled - missing clusterid
- Cause: The data-fabric
cluster was not up and running before
configure.sh
was run with
the options to configure monitoring.
- Solution: Complete the following steps and then re-run
configure.sh
:
- Verify that the CLDB services is running. If not, start Warden with the following
command:
service mapr-warden start
.
- Verify that the
$MAPR_HOME/conf/clusterid
file exists.
Collectd Errors
- collectd service not enabled - missing clusterid
- Cause: The cluster was not up and running before
configure.sh
was run with the options to configure monitoring.
- Solution: Complete one or all of the following steps and then re-run
configure.sh
:
- Verify that the CLDB services is running. If not, start Warden with the following
command:
service mapr-warden start
.
- Verify that the
$MAPR_HOME/conf/clusterid
file exists.
Grafana Errors
- Failed to pick default data source host
- Cause: The OpenTSDB nodes list defined by the
-OT
parameter
was incorrect.
- Solution: Check the syntax and the validity of those nodes.
- Failed to create scratch config file
- Cause: There is not enough disk space,
configure.sh
was not
run as the root
user, or the file system was mounted as read-only.
- Solution: Complete one or all of the following steps and then re-run
configure.sh
:
- Verify that you are logged in as the
root
user or using
sudo.
- Verify that there is enough disk space.
- Verify that the file system containing
/opt/mapr/grafana
is
mounted as read/write.
- Failed to change the port
- Cause: The
sed
utility failed to edit a port value in the
temporary configuration file. This may occur due to issues with file system permissions,
disk space, or file corruption.
- Solution: Complete the following steps and then re-run
configure.sh
:
- Verify that the file system containing
/opt/mapr/grafana
is
mounted as read/write.
- Verify that there is enough disk space.
- Re-install the monitoring packages.
- Failed to configure ssl for grafana
- Cause: The
sed
utility failed to edit a port value in the
temporary configuration file. This can occur due to issues with file system permissions,
disk space, or file corruption.
- Solution: Complete the following steps and then re-run
configure.sh
:
- Verify that the file system containing
/opt/mapr/grafana
is
mounted as read/write.
- Verify that there is enough disk space.
- Re-install the monitoring packages.
- ERROR: Failed to install grafana warden config file
- Cause: There is not enough disk space,
configure.sh
was not
run as the root user, or the file system containing /opt/mapr/grafana
is mounted as read-only.
- Solution: Complete the following steps and then re-run
configure.sh
:
- Verify that you are logged in as the
root
user or using
sudo.
- Verify that there is enough disk space.
- Verify that the file system containing
/opt/mapr/grafana
is
mounted as read/write.
Kibana Errors
- Failed to configure elasticsearch server URL
- Cause: There is not enough disk space,
configure.sh
was not
run as the root user, or the file system containing /opt/mapr/kibana
is
mounted read-only.
- Solution: Complete the following steps and then re-run
configure.sh
:
- Verify that you are logged in as the
root
user or using
sudo.
- Verify that there is enough disk space.
- Verify that the file system containing
/opt/mapr/kibana
is
mounted as read/write.
- Failed to configure ssl for Kibana
- Cause: There is not enough disk space or the root user does not have the
required directory permissions to create the file in the
/opt/mapr/kibana/kibana-<version>/config
directory.
- Solution: Complete the following steps and then re-run
configure.sh
:
- Verify that there is enough disk space.
- Verify that the file system containing
/opt/mapr/kibana
is
mounted as read/write.
- Kibana logon unsuccessful because Searchguard "Service Unavailable"
- Cause: On a slower server, Kibana logons sometimes do not succeed because the
Searchguard configuration containing the user names and passwords has not finished
loading. You might notice an error like the following in the Elasticsearch
MapRMonitoring.log
:[2018-06-30T07:47:15,062][ERROR][c.f.s.a.BackendRegistry ] Not yet initialized (you may need to run sgadmin)
- Solution: Try using the following command to restart
Elasticsearch:
maprcli node services -name elasticsearch -action restart -nodes $(hostname -f)