mnesia disk 100% full

[tom_slason]

looking for a better understanding of best practices or alternatives - once our RabbitMQ Node is in a bad situation.

Consumers were off line, producers of persistent messages flooded RabbitMQ and the mnesia datastore is now 100% full.

don't want to loose any messages, so don't want to delete content off  mnesia datastore. 

Consuming application can not access virtual host while mnesia datastore is 100% full.

We can shutdown producers,  administratively we can add additional storage then consuming application can reconnect and process the backlog.

Is there a different approach we are not thinking of ? other options ?

It is understood that we will need to make adjustments to better utilize Memory and Disk Alarms to avoid this situation ,,

[Michael Klishin]

You can copy the database dir to a different machine, make sure its hostname/node name are the same, and the node might be able to start. I say might because when you run out of disk space, you never can be sure
what writes did not make it to disk.

If you can expand storage volume "in place", e.g. this is virtual hardware, that's probably the easiest option.