Isilon: How to shut down a single node in a cluster

If you plan to shut down a single node for maintenance, we recommend performing an isi_flush of the specific node first. The isi_flush command writes all cached data from memory to the drives.
Without an isi_flush, the node may not shut down properly and may reboot instead. The reboot can take a prolonged period of time because the node flushes its cache and boots up again.
This in turn requires issuing the shutdown command a second time. To save time, flushing the node in advance allows the node to shut down quicker. 

1. Log in to the cluster and issue the isi_flush command.
 

# isi_for_array -n<specific-node-number> 'isi_flush'

example: (Flushes Node 5)

Isilon-3# isi_for_array -n5 'isi_flush'
Isilon-5: Cache flushing complete.
Isilon-3# 

Note: Depending on the size of the cluster and workflow, the isi_flush command may take several minutes to complete. Up to 30+ minutes to complete a flush is not unusual. Do not interrupt, kill, or cancel the isi_flush process as that may cause a temporary service interruption. 

2. Shut down the node.

# isi config
    >>> shutdown <specific-node-number>