vxfenmode - I/O fencing configuration file






The /etc/vxfenmode file contains setup information used by I/O fencing. The system administrator is responsible for maintaining this file correctly and consistently across all systems in the cluster. If the configuration file is not present, I/O fencing runs in SCSI3 mode. This default behavior is a relict from a prior release and should be abandoned in favor of explicit settings.

A pound sign (#) in the first column of a line means that the line is a comment. Blank lines are ignored.

The /etc/vxfenmode file supports the following parameters:
  - sets the main operation mode of I/O fencing

FORMAT: vxfen_mode=[scsi3|customized|disabled]

o scsi3: use scsi3 persistent reservation disks
o customized: use script based customized fencing (support for this is limited!)
o disabled: run the driver but don’t do any actual fencing
  - specifies which mechanism of customized fencing is to be used by I/O fencing. If the vxfen_mode is customized, a valid vxfen_mechanism is mandatory.
Note: The customized mechanism is only supported in certain configurations. It is not available for customers to write their own fencing mechanism. FORMAT: vxfen_mechanism=[cps|user_defined]
o cps: use coordination point server for cluster membership coordination
o user_defined: use a customized fencing mechanism created by the end-user (not supported at this time);
  - determines the time vxfend should wait for customized fencing scripts to finish execution before attempting to manually shut them down. If this option is omitted, the default value is 120 seconds. This parameter is only used when running vxfen_mode=customized

FORMAT: vxfen_script_timeout=[user_defined]

o user_defined: time to wait for customized fencing scripts before shutting them down
  - determines the way in which I/O fencing communicates with the coordination disks

FORMAT: scsi3_disk_policy=[raw|dmp]

o raw: connect to disks using the native interface
o dmp: use dynamic multipathing

Although vxfen supports raw for scsi3_disk_policy, it is now deprecated. It is recommended to use dmp as the scsi3_disk_policy for coordinator disks.

For clusters that use customized fencing, vxfen can cache the serial number or UUID of a coordination point and use it in case the coordination point is not accessible when vxfen starts up. The cluster starts up only if majority of the coordination points are up. The following two parameters control the behaviour related to this feature:
  - specifies the types of coordination points for which a node must not store the UUID/serial number. To specify multiple values, use a comma-separated list. By default, this feature is available only for coordination point servers. To turn it on for disks, you must set the value of the db_ignore_list to "none".

FORMAT: db_ignore_list=[disk|server|none]

o disk: do not cache or use from cache serial number of any disk based coordination point.
o server: do not cache or use from cache UUID of any CP server based coordination point.
o none: enable cache for both disk and server type of coordination points.
  - specifies the maximum number of UUIDs/serial numbers that a node can store in its cache. The default value for this attribute is 1000. If the default value is used, the node approximately requires 1MB of disk space to store the UUIDs/serial numbers.

FORMAT: db_entries_limit=integer


The following example shows a typical /etc/vxfenmode file configuring I/O fencing to run in scsi3 mode with a dmp disk connection:

# vxfen configuration file for SCSI3

The following example shows a typical /etc/vxfenmode file configuring I/O fencing to run in customized mode, using cps mechanism.

# vxfen configuration file for CPS


  I/O fencing configuration file


vxfenadm(1M), vxfentsthdw(1M), vxfendebug(1M), vxfenconfig(1M), vxfenclearpre(1M)


Copyright (c) 2011 Symantec.

All rights reserved.