Users Guide

Fault handling by PowerStore import
This appendix contains the following information:
Topics:
• Import processing and fault handling overview
• Import network overview
• Import alerts and error codes
• Failures during non-disruptive import
• Failures during agentless import
• Consistency group handling
• Configuration changes during import
• Application integration software
Import processing and fault handling overview
Import copies the source volume content to PowerStore volumes in the back end while continuing host I/Os. For non-disruptive
import, it does this without any interruptions. For agentless import, a one time host application downtime is required prior to
the migration. Given the sensitive nature of this operation, a few run-time checks are made before starting the import process.
These checks help avoid failures in the import process. If any of these prechecks fail, import fails to start without host I/O
impact. You are advised to rectify the issues and retry the import. Sometimes a simple retry may also lead to a successful import
but depends on the nature of the failure encountered.
Non-disruptive import fault handling overview
It is still possible that the import runs into a failure during run-time, after the import process starts. If any such failure occurs
during import processing, internal cancel of import occurs and the I/Os are redirected back to the source. This action is taken to
ensure that host I/Os go on uninterrupted. After resolving the failure reason or reasons, you can retry the import.
You are also given an option to cancel the import and go back to the source. This option is also designed to work without any
impact to ongoing host I/O. However, here also, it is expected that all the network connections are live for cancel to succeed.
For example, if host loses the SAN path to source volumes midway during import, an ongoing import operation is not impacted.
There is no impact because the I/Os are going to the destination (PowerStore) volume. The paths from the host to source
volume are not used. However, if you attempt a cancel operation in this situation of the host not having live SAN paths to the
source volumes, it will fail.
Depending on the load, the timing of path failure, and configuration, cancel operations can take a long time before timing out.
Cancel operations can also cause stop responding like situations that impact host I/O. You are expected to ensure that these
connections are up before attempting the cancel. If the connections fail unexpectedly, manual intervention is required to ensure
the paths from host to source volumes are working. Based on the duration and nature of these network failures, either retry
canceling import or a cleanup operation may be required to fully recover.
Agentless import fault handling overview
If the import runs into a failure during run-time after the import process starts, the import state changes to Cancel Required.
Shutdown the host application and reconfigure it to access the source volume and cancel the import. After resolving the failure
reason or reasons, you can retry the import.
F
64 Fault handling by PowerStore import