Troubleshooting and Basic Remediation
Troubleshooting questions are scenario-based: a host cannot see a LUN, I/O is slow, or replication has stopped. Marks go to answers that scope the problem before changing anything and that use the right diagnostic tool.
A repeatable method
- Define the symptom precisely: one host or all hosts, one volume or all volumes, read or write, intermittent or constant.
- Scope by layer: host (driver, multipath, rescan), fabric (zone, name server, ISL), array front end (port group, masking view, SG membership), back end and SRP (device ready, capacity), replication (SRDF or SnapVX state).
- Change one thing at a time and record it; capture evidence before remediation so you can explain what happened.
Commands worth knowing
- symcfg list discovers arrays; symcfg -sid list -fa all shows front-end ports.
- symaccess list view -sid shows masking views with IG, PG and SG; symsg show -sid -sg shows membership.
- symdev list -sid -noport lists devices that are not mapped; sympd list shows physical drives.
- symrdf -g query, symsnapvx -sid -sg list and symevent show replication and event state.
- symdiag collects logs and runs checks; it is the standard first artifact for a support case.
Worked example
A host that previously had eight paths now shows two, and only that host is affected. Because other hosts are fine, the fabric and array are probably healthy: check the host HBA state, then the zones containing that host's WWPNs. A common cause is a zone removed during switch maintenance, or an HBA replaced without updating the initiator group. Fix the zone or IG, rescan, and confirm path count and LUN visibility before closing.
Traps
Rebooting the array or deleting a masking view as a first step, both destructive and rarely needed; assuming every host is affected without checking scope; diagnosing from memory instead of reading symaccess and symdev output; and escalating without symdiag.