Taking a Back-End IT Nexus with High Latency Out of Service
If an I/O takes greater than 1 second to complete on an ITL (Initiator-Target-LUN on an IT Nexus), then the ITL and IT accrue a penalty where the allowed command limit for the ITL is reduced from five to one. If the accumulated penalty for an ITL exceeds 2 seconds, then the command limit on the ITL is reduced to zero indicating that no more commands are allowed on this ITL. Due to high latency, if greater than 20 ITLs on an IT Nexus are penalized, then the IT Nexus is marked degraded and metro node automatically stops using the IT Nexus for host-based I/O until performance improves.
NOTE: If the last available path to a Logical Unit is marked degraded, then it cannot be taken out of service, and a penalty is applied to allow a single I/O at a time to the LU. One ITL per Logical Unit per director continues to receive commands. Once the performance improves the metro node automatically restores the default outstanding I/O count to the Logical Unit.
Degraded back-end IT Nexuses can be monitored using the VPlexcli command
back-end degraded list. For more details, see
CLI Reference Guide for metro node. Due to continuous high latency, when an IT Nexus is marked degraded, then this command lists the Degradation Reason as
Degraded performance.
If a user finds that a degraded IT Nexus has been restored to health, it is also possible to manually restore its use through the VPlexcli command
back-end degraded recover.
Data is not available for the Topic
Please provide ratings (1-5 stars).
Please provide ratings (1-5 stars).
Please provide ratings (1-5 stars).
Please select whether the article was helpful or not.
Comments cannot contain these special characters: <>()\