This page contains the list of deprecations and important or breaking changes for Vault 0.8.0 compared to the most recent release. Please read it carefully.
If you are upgrading from Vault Enterprise, you should take one of the following upgrade paths. Please note that reindexing stops the ability of the node to process requests while the indexing is happening. As a result, you may want to plan the upgrade for a time when either no writes to the primary are expected and secondaries can handle traffic (if using replication), or when there is a maintenance window.
There are two reindex processes that need to happen during the upgrade, for two different indexes. One will happen automatically when the cluster is upgraded to 0.8.0. This happens even if you are not currently using replication. The other can be done in one of a few ways, as follows.
If not using replication, no further action needs to be taken.
The first option is to issue a write to
sys/replication/reindex on the
primary (it is not necessary on the secondaries). When the reindex on the
primary is finished, upgrade the secondaries, then upgrade the primary.
The second option is to upgrade all nodes to 0.8.0 at the same time. This removes the need to perform an explicit reindex but may equate to more down time since secondaries will not be able to service requests while the primary is performing an explicit reindex.
This path was meant to require
sudo capability but was not implemented this
way. It now requires
sudo capability to run.