The table below lists and explains names and meanings of the DRBD parameters. These values are tuned by editing the /etc/default/LifeKeeper configuration file.
| Parameter Name | Meaning | Setting Value | Default Value | When to Apply | Notes |
|---|---|---|---|---|---|
| DRBD_ALLOW_FAILOVER_SYNC_TARGET | Allow failover to a target being synchronized. | 0/1 | 1 | Failover of a resource | 1 (default) allows failover to a sync target. 0 does not allow failover to a sync target |
| DRBD_BASE_PORT | Port used for DRBD connections. | Integers | 7788 | Creating a resource | Each resource uses a unique TCP port for its network connections. This value is the first port that will be used and incremented as resources are added. |
| DRBD_ALLOW_DEVNAME | Allow using a device that does not have a GUID. | 0/1 | 0 | Creating and extending a resource | A GUID device node is recommended to assure that DRBD uses the correct device when devices are renamed by the OS. This setting allows using a non-GUID device node when a GUID device node is not available. The required internal metadata will prevent the resource to connect and cause data corruption when a non-GUID device node is used and the OS renames devices. |
| DRBD_DEBUG_ENABLED | Enable debug in the DRBD API. | 0/1 | 0 | As required (takes effect immediately) | The DRBD debug is logged when LifeKeeper debug is enabled (debug_drbd LifeKeeper flag). |
| DRBD_WAIT_FOR_USABLE_TIMEOUT | How long to wait for a DRBD resource to be ready. | Integers | 30 | As required (takes effect immediately) | DRBD resources are not immediately usable when a resource is brought “up”. In most cases the DRBD resource is available in less than 10 seconds. Setting this value too low may cause a transient failure. |
| DRBD_TAKE_DOWN_ON_REMOVE | Determines if the resource is taken ‘down’ when a resource is taken out-of-service. | 0/1 | 0 | As required (takes effect immediately) | 0: Leaving the resource ‘up’ when out-of-service allows synchronization to continue during a switchover and allows for a faster switchover. 1: Taking the resource down stops all activity, which can be useful in certain maintenance or troubleshooting scenarios. |
| DRBD_RECOVER_TIMEOUT | Number of seconds to wait before aborting a recover operation that is hung. | Integers | 300 | As required (takes effect immediately) | |
| DRBD_ALLOW_DISKLESS | Allow a resource to be ‘primary’ when the backing device fails. | 0/1 | 1 | As required (takes effect immediately) | When DRBD detects an IO error its default behavior is to “detach” the backing device (see section 4.13 in the DRBD User’s Guide), continue in “diskless” mode (on-io-error setting) and allow remote reads (allow-remote-reads setting). When set to ‘0’ LifeKeeper will immediately initiate a switchover when the ‘primary’ device is ‘diskless’. While a ‘primary’ device is ‘diskless’ all IO’s occur over the network to a mirror target. |
Post your comment on this topic.