| Moab Workload Manager® for Grids |
17.18 Grid Diagnostics and Validation
- 17.8.1 Peer Management Overview
- 17.8.2 Peer Diagnostic Overview
- 17.18.3 Peer-to-Peer Communication Issues
17.8.1 Peer Management Overview
- Use mdiag -R to view interface health and performance/usage statistics.
- Use mrmctl to enable/disable peer interfaces.
- Use mrmctl -m to dynamically modify/configure peer interfaces.
- Use mdiag -x to view general grid configuration and system diagnostics.
17.8.2 Peer Diagnostic Overview
- Use mdiag -R to diagnose general RM interfaces.
- Use mdiag -S to diagnose general scheduler health.
- Use mdiag -R -V job <RMID> to diagnose peer-to-peer job migration.
- Use mdiag -R -V data <RMID> to diagnose peer-to-peer data staging.
- Use mdiag -R -V cred <RMID> to diagnose peer-to-peer credential mapping.
17.18.3 Peer-to-Peer Communication Issues
If communication between peers is failing, generally the first step is run the command mdiag -R. This command will report general resource manager state, as well as indication of detected configuration issues and failures with specific commands. Among the failures reported can be the following:
Message | WARNING: Client not yet initialized. |
| Description | Before peer communication can be established, an initialize message must be exchanged. If this message is displayed, a peer has attempted to communicate without first sending this initialization message. |
| Recommended Actions | restart the peer |
Message | WARNING: client id 'XXX' is unknown' ERROR: command 'XXX' cannot be executed from host 'XXX' ERROR: user 'XXX' is not authorized to run command 'XXX' |
| Description | A peer has attempted to communicate whose identity cannot be authenticated. Either the peer is unknown or the peer is known and the secret key is invalid. |
| Recommended Actions | Verify the requesting client should be attempting to communicate and if so, verify that a correct corresponding entry exists in the moab-private.cfg file. |
If these commands do not provide adequate detail, additional communication diagnostics can be enabled by setting the DROPBADREQUEST client attribute to FALSE as in the following example: