SGPanic for handling node failure

For internal disk rick storage (FPO clusters), unmountOnDiskFail must be configured as “meta”. If it is not, change the configuration by running the mmchconfig command.

With unmountOnDiskFail configured as meta, if you see file system SGPanic reported when nodes are non functional, there are more than three nodes with metadata disk down together or there are more than three disks with meta data down. Follow the steps in the section 8.1 to fix the issues. Run the mmfsck -n command to scan the file system to ensure that mmfsck displays the following message: File system is clean finally. If mmfsck -n does not report “File system is clean”, you need to open PMR to report the issues and fix this with guide from IBM Storage Scale.