Veritas InfoScale™ 7.3.1 Troubleshooting Guide - Linux
- Introduction
- Section I. Troubleshooting Veritas File System
- Section II. Troubleshooting Veritas Volume Manager
- Recovering from hardware failure
- Failures on RAID-5 volumes
- Recovery from failure of a DCO volume
- Recovering from instant snapshot failure
- Recovering from failed vxresize operation
- Recovering from boot disk failure
- VxVM boot disk recovery
- Recovery by reinstallation
- Managing commands, tasks, and transactions
- Backing up and restoring disk group configurations
- Troubleshooting issues with importing disk groups
- Recovering from CDS errors
- Logging and error messages
- Troubleshooting Veritas Volume Replicator
- Recovery from configuration errors
- Errors during an RLINK attach
- Errors during modification of an RVG
- Recovery on the Primary or Secondary
- Recovering from Primary data volume error
- Primary SRL volume error cleanup and restart
- Primary SRL header error cleanup and recovery
- Secondary data volume error cleanup and recovery
- Troubleshooting issues in cloud deployments
- Recovering from hardware failure
- Section III. Troubleshooting Dynamic Multi-Pathing
- Section IV. Troubleshooting Storage Foundation Cluster File System High Availability
- Troubleshooting Storage Foundation Cluster File System High Availability
- Troubleshooting CFS
- Troubleshooting fenced configurations
- Troubleshooting Cluster Volume Manager in Veritas InfoScale products clusters
- Troubleshooting interconnects
- Troubleshooting Storage Foundation Cluster File System High Availability
- Section V. Troubleshooting Cluster Server
- Troubleshooting and recovery for VCS
- VCS message logging
- Gathering VCS information for support analysis
- Troubleshooting the VCS engine
- Troubleshooting Low Latency Transport (LLT)
- Troubleshooting Group Membership Services/Atomic Broadcast (GAB)
- Troubleshooting VCS startup
- Troubleshooting issues with systemd unit service files
- Troubleshooting service groups
- Troubleshooting resources
- Troubleshooting I/O fencing
- System panics to prevent potential data corruption
- Fencing startup reports preexisting split-brain
- Troubleshooting CP server
- Troubleshooting server-based fencing on the Veritas InfoScale products cluster nodes
- Issues during online migration of coordination points
- Troubleshooting notification
- Troubleshooting and recovery for global clusters
- Troubleshooting licensing
- Licensing error messages
- VCS message logging
- Troubleshooting and recovery for VCS
- Section VI. Troubleshooting SFDB
If a unit service has failed and the corresponding module is still loaded, systemd cannot unload it and so its package cannot be removed
A unit service may go into the Failed state during startup while the corresponding module is in the Loaded state due to some configuration issue. If this situation occurs, the module cannot be cleanly unloaded using systemd commands.
For example, you may encounter this message when you start the vxfen
service:
localhost]:~ # systemctl start vxfen Job for vxfen.service failed because the control process exited with error code. See "systemctl status vxfen.service" and "journalctl -xe" for details.
If you encounter this message, view the status of service:
localhost]:~ # systemctl status -l vxfen ? vxfen.service - VERITAS I/O Fencing (VxFEN) Loaded: loaded (/opt/VRTSvcs/vxfen/bin/vxfen; enabled; vendor preset: disabled) Active: failed (Result: exit-code) since Wed 2017-04-19 17:40:13 IST; 14s ago Process: 14721 ExecStart=/opt/VRTSvcs/vxfen/bin/vxfen start 2>&1 (code=exited, status=1/FAILURE) Apr 19 17:40:12 localhost] systemd[1]: Starting VERITAS I/O Fencing (VxFEN)... Apr 19 17:40:12 localhost] vxfen[14721]: Starting vxfen.. Apr 19 17:40:12 localhost] vxfen[14721]: Loaded 4.4.21-69-default on kernel 4.4.21-69-default Apr 19 17:40:13 localhost] vxfen[14721]: vxfen cannot generate vxfentab due to missing file /etc/vxfendg. Apr 19 17:40:13 localhost] systemd[1]: vxfen.service: Control process exited, code=exited status=1 Apr 19 17:40:13 localhost] systemd[1]: Failed to start VERITAS I/O Fencing (VxFEN). Apr 19 17:40:13 localhost] systemd[1]: vxfen.service: Unit entered failed state. Apr 19 17:40:13 localhost] systemd[1]: vxfen.service: Failed with result 'exit-code'.
Then, view the status of the module:
localhost]:~ # lsmod | grep vxfen vxfen 372992 0 veki 19214 4 gab,llt,vxfen localhost]:~ # /opt/VRTSvcs/vxfen/bin/vxfen status loaded
The systemd daemon does not accept a systemctl stop unitServiceFile command when the service is in the Failed state or the Inactive state.
For example, if the vxfen
service is in the Failed or the Inactive state, the following command fails to unload the module successfully:
systemctl stop vxfen
Additionally, if a module is in the Loaded state, the corresponding rpm package cannot be removed during uninstallation.
Recommended action
To cleanly stop and unload the module, run the stop command from the source script file as follows:
serviceSourceScript stop
For example:
localhost]:~ # /opt/VRTSvcs/vxfen/bin/vxfen stop VxFEN: Module already unconfigured. Only unloading localhost]:~ #/opt/VRTSvcs/vxfen/bin/vxfen status not loaded
To view the source path of any service, you can use the systemctl command as follows:
# systemctl show unitServiceFile -p SourcePath
For example:
# systemctl show vxfen -p SourcePath SourcePath=/opt/VRTSvcs/vxfen/bin/vxfen
After the module is cleanly unloaded, you can perform the next startup action or proceed to uninstall the associated package.