Each service maintains its own set of logs. By default, the logs are maintained in the install_path/hci/log folder on each instance in the system. During installation, you can configure each service to store its logs in a different, non-default location.
Log management
You can manage any of the HCI log files yourself, deleting or archiving them as necessary.
System logs are managed automatically in these ways:
- All log files are periodically added to a compressed file and moved to install_path/hci/retired/. This occurs at least one time a day, but can also occur:
- Whenever you run the log_download script.
- Hourly, if the system instance's disk space is more than 60% full.
- After a log file grows larger than 10MB in size, the system stops writing to that file, renames it, and begins writing to a new file. For example, if exampleService.log.0 grows too large, it is renamed to exampleService.log.1 and the system creates a new exampleService.log.0 to write to.
Retrieving logs and diagnostic information
The log_download tool lets you easily retrieve logs and diagnostic information from all instances in the system. This tool is located at this path on each instance:
install_path/hci/bin/log_download
For information about running the tool, use this command:
install_path/hci/bin/log_download -h
- When using the log_download tool, if you specify the --output option, do not specify an output path that contains colons, spaces, or symbolic links. If you omit the --output option, you cannot run the script from within a folder path that contains colons, spaces, or symbolic links.
- When you run the log_download script, all log files are automatically compressed and moved to the folder install_path/retired/.
- If an instance is down, you need to specify the option --offline to collect the logs from that instance. If your whole system is down, you need to run the script log_download with the option --offline on each instance.
Default log locations
By default, each service stores its logs in its own folder at this path:
install_path/hci/log
This table shows the default log folder names for each service. Depending on how your system was configured when first deployed, your system's logs might not be stored in these folders.
Related service | Default log directory name | Contains information about |
---|---|---|
Admin-App | com.hds.ensemble.plugins.service.adminApp | The System Management application. |
Cluster-Coordination | com.hds.ensemble.plugins.service.mesosMaster | Hardware resource allocation. |
Cluster-Worker | com.hds.ensemble.plugins.service.mesosAgent | Work ordered by the Cluster-Coordination service. |
Clustered-File-System | com.hds.ensemble.plugins.service.hdfs | Usage of the Clustered-File-System service by other services for storing data. |
Dashboard | com.hds.ensemble.plugins.service.kibana | The advanced Dashboard Management service. |
Database | com.hds.ensemble.plugins.service.cassandra |
|
Index | com.hds.ensemble.plugins.service.solr | Index collections and search indexes. |
Logging | com.hds.ensemble.plugins.service.logstash | The transport of system events and workflow task metrics to the Metrics service. |
Metrics | com.hds.ensemble.plugins.service.elasticsearch | The storage and indexing of:
|
Message Queue | com.hds.ensemble.plugins.service.kafka | Transmission of data between instances. |
Monitor-App | com.hds.ensemble.plugins.service.mapApp | The HCM application. |
Network-Proxy | com.hds.ensemble.plugins.service.haproxy | Network requests between instances. |
Scheduling | com.hds.ensemble.plugins.service.chronos | Workflow task scheduling. |
Search-App | com.hds.ensemble.plugins.service.searchApp | The Search App |
Sentinel | com.hds.ensemble.plugins.service.sentinel | Internal system processes. |
Service-Deployment | com.hds.ensemble.plugins.service.marathon | The deployment of high-level services across system instances. High-level services are the ones that you can move and configure (such as Index), not the services grouped under System Services. |
Synchronization | com.hds.ensemble.plugins.service.zookeeper | Coordination of actions and database operations across instances. |
Watchdog | com.hds.ensemble.plugins.service.remoteAction | Internal system processes |
Watchdog | com.hds.ensemble.plugins.service.watchdog | General diagnostic information. |
Workflow-Agent jobs | com.hds.ensemble.plugins.job.workflow | Workflows |