In 5.6, the process has been completely simplified so that all of the hosts are scanned automatically when the new storage wizard comes up. This fixes other issues that may be present as well. Engineering has provided the plugin for previous versions of xs to simplify the process there as well. With 5.5 and below, the process was not as apparent but still doable. Here is what you will want to do on pre-5.6 systems to avoid having to reboot:
a) Install the plugin under /etc/xapi.d/plugins on each node in the pool
b) Make it executable: chmod +x /etc/xapi.d/plugins/hostdevscan
c) Trigger it via the CLI for each host in the pool:
xe host-call-plugin host-uuid=<HOST_UUID> plugin=hostdevscan fn=scan
This will do the following:
Detect all FC, SAS and iSCSI adapters (Hardware and Software) on the host. For each adapter, detect the channels to be reprobed. It will then locate a LUN to issue a SCSI report LUNs command (Note that there must always be at least 1 active LUN for this to succeed). It will then loop through all LUNs known to the host, removing stale LUNs that the report LUNs command did not list, and probing for new ones that are not present. It will also validate that the SCSIid path matches the SCSIid returned by the active LUN, thereby detecting re-mapped LUNs. In the case a LUN has been remapped to the same LUNid it will re-probe it.