We recently ran into an SP bug check on one of our VNX arrays and after it came back up several of the initiator paths to the VPLEX did not come back up. We were also seeing IO timeouts. This is a known bug that happens when there is an SP reboot and is fixed with Patch 1 for GeoSynchrony 5.3. EMC has released a script that provides a workaround until the patch can be applied: https://download.emc.com/downloads/DL56253_VPLEX_VNX_SCRIPT.zip.zip
The following pre-conditions need to happen during a VNX NDU to see this issue on VPLEX:
1] During a VNX NDU, SPA goes down.
2] At this point IO time-outs start happening on IT nexus’s pertaining to SPA.
3] The IO time-outs cause the VPLEX SCSI Layer to send LU Reset TMF’s. These LU Reset TMF’s get timed out as well.
You can review ETA 000193541 on EMC’s support site for more information. It’s a critical bug and I’d suggest patching as soon as possible.
