== Presto machines to be upgrades
[x] presto-worker-test 1 - `an-test-presto1001.eqiad.wmnet`
[x] an-presto1001
[x] an-presto1002
[] an-presto1003
[] an-presto1004
[] an-presto1005
These machines should be relatively easy to upgrade, owing to the following factors.
* There is no need for data retention - There is a `/srv/` volume mounted, but it does not contain any data of value.
* We have already shown that presto on bullseye works with an-presto10[06-15]
Therefore, the procedure to upgrade these hosts should be as simple as running the cookbook for each host as follows:
```
sudo cookbook sre.hosts.reimage --os bullseye -t 329361 an-presto1001
```
We could also add one (or more) of the currently disabled presto hosts into the cluster to pick up the load of one worker being down.
i.e. delete this file: https://phabricator.wikimedia.org/source/operations-puppet/browse/production/hieradata/hosts/an-presto1006.yaml