PowerStore: IO Errors on NAS File Systems On Node Reboot

When one of the nodes of the appliance is rebooted, a NAS file system might experience IO errors. When the node comes up from the reboot, the IO errors will no longer occur, and the file systems should automatically recover. However, this may indicate that there is an issue with the internal access to the NAS file system on the surviving node.

 

---

Cause

It is possible that the internal multipath access to an SDNAS volume has lost its path to the volume on the local (surviving) node, and only has access to the volume on the peer node. When the peer node reboots, all access is lost. The error messages for this issue appear on the host. They are from the specific application that is being used so will vary depending on the tool or application. 

The following example is the error messages that are produced for IO using Infiload (the output will differ for other tools and applications):

 

---

Resolution

The volume will be available once the peer node boots. However, if the peer node were to reboot again, IO errors would occur again. To correct the issue on the surviving node, the surviving node should be rebooted at the customer’s earliest convenience. See Powering down and Reboot Procedures Guide for instructions on performing the reboot procedure.

 

---