Description of problem: In a CNS environment, host node does not have gluster installed and sosreport is generated from host. This causes many sos_commands to be omitted. If the sos report is obtained from pod then it does not give completed information of the host node. We need to analyze what is missing. The fix might be a guide to obtain sosreports from both pod and host node OR it might need some enhancements in sosreport tool.
Severity? Next step?
Severity is low now as the sosreports we have obtained in recent cases have provided required data. Next Steps: 1. verify if something is missing from sosreports even now 2. Make sure the correct method to run sosreport tool is documented. Keeping the needinfo on myself till I answer the two questions above.
(In reply to Raghavendra Talur from comment #3) > Severity is low now as the sosreports we have obtained in recent cases have > provided required data. > > Next Steps: > 1. verify if something is missing from sosreports even now > 2. Make sure the correct method to run sosreport tool is documented. > > Keeping the needinfo on myself till I answer the two questions above. Please update (1 month later, we really need some answers)?