Table of Contents
NFS clients can maintain use of a file over a halt or reboot, although the node will fail to respond during that time. CIFS, FCP, and iSCSI clients cannot safely maintain use of a file over a halt or reboot. If the node is running CIFS, FCP or iSCSI, you may use the -t option to specify the time before shutdown. If halt is invoked without -t, it displays the number of CIFS users, the number of open CIFS files, the number of mapped LUNs and the number of connected FCP and iSCSI clients. Then it prompts you for the number of minutes to delay. cifs terminate automatically notifies all CIFS clients that a CIFS shut-down is scheduled in mins minutes, and asks them to close their open files. CIFS files that are still open at the time the node halts will lose writes that had been cached but not written. FCP and iSCSI will not notify clients, but will allow administrators to confirm that the mapped LUNs are not in use. LUNs that are in use at the time the node halts will result in client failures.
halt logs a message in /etc/messages to indicate that the node was halted on purpose.
The halt command is not available in partner mode. That is, you cannot enter the partner halt command on the live node after it takes over the failed partner. This is because a node that has been taken over is no longer running and cannot be halted.
When the -d option is used, cached data is not flushed to disk. All data not yet on disk is stored in the NVRAM. The node will automatically replay NVRAM during the next boot, bringing the disks up to date with the most recent operation. However, if NVRAM loses charge, some of the most recently modified data may be lost. Because of this, the -d option should be used only to produce coredumps requested by technical support. NVRAM retains charge for three days, so all data will be intact if NVRAM is replayed within three days of the dirty shutdown.
Table of Contents