9.8 Configuring the EVMS Remote Request Timeout

Novell Cluster Services for Linux uses the EVMS Cluster Segment Manager for shared devices. When load and unload scripts run, EVMS mounts or dismounts shared devices. The default timeout for remote requests for EVMS is 12 seconds. If you have many shared devices to be mounted or dismounted and the timeout is exceeded, you might see EVMS locking errors and some cluster resources can go comatose.

You must increase the EVMS engine’s remote_request_timeout value if you have many shared devices and it is taking longer than 12 seconds to process them as the load or unload scripts are run for a cluster migration. The time needed varies with the number of shared devices that are being processed. For example, for about 60 shared devices, set the timeout to 30 seconds.

To modify the EVMS timeout:

  1. Open the /etc/evms.conf file in a text editor.

  2. Edit the following section to increase the remote_request_timeout value:

    engine {
         # Timeout in seconds when waiting for a response from a remote node
         remote_request_timeout = 12
  3. Save the file.

  4. At the terminal console prompt, enter