Collecting diagnostic data through GUI
IBM® Support might ask you to collect logs, trace files, and dump files from the system to help them resolve a problem. You can perform this task from the management GUI or by using the gpfs.snap command. Use the page in the IBM Spectrum Scale™ GUI to collect details of the issues reported in the system.
The entire set of diagnostic data available in the system helps to analyze all kinds of IBM Spectrum Scale issues. Depending on the data selection criteria, these files can be large (gigabytes) and might take an hour to download. The diagnostic data is collected from each individual node in a cluster. In a cluster with hundreds of nodes, downloading the diagnostic data might take a long time and the downloaded file can be large in size.
- Include only affected functional areas
- Include only affected nodes
- Reduce the number of days for which the diagnostic data needs to be collected
- Standard diagnosticsThe data that is collected in the standard diagnostics consists of the configuration, status, log files, dumps, and traces in the following functional areas:
- Core IBM Spectrum Scale
- Network
- GUI
- NFS
- SMB
- Object
- Authentication
- Cluster export services (CES)
- Crash dumps
- All nodes
- Specific nodes
- All nodes within one or more node classes
- Deadlock diagnostics
The data that is collected in this category consists of the minimum amount of data that is needed to investigate a deadlock problem.
- Performance diagnostics
The data that is collected in this category consists of the system performance details collected from performance monitoring tools. You can only use this option if it is requested by the IBM Support.
- Traces from the GUI that contain the information about errors occurred inside GUI code
- Several configuration files of GUI and postgreSQL
- Dump of postgreSQL database that contains IBM Spectrum Scale configuration data and events
- Output of most mmls* commands
- Logs from the performance collector