Servers to migrate:
- db1.internal.staging.swh.network
- belvedere.internal.softwareheritage.org
- somerset.internal.softwareheritage.org
**Plan**:
staging:
- stop the loaders and listers workers
- stop the indexers
- stop the scheduler runners *
- on db1:
- connect to the serial console => swh-ceph-osd1-adm.inria.fr pass in the credential store (infra/db1.staging/idrac)
- disable puppet
- stop and disable postgresql
- perform the last buster upgrade
- reboot
- perform the bullseye upgrade
- reboot
- restart and enable postgresql
- restart puppet
- restart the scheduler runners
- restart the indexers
- restart the lister / loaders
Production:
- upgrade somerset
- switch the webapp to belvedere
- on somerset
- disable puppet
- stop and disable postgresql
- perform the last buster upgrade
- reboot
- perform the bullseye upgrade
- reboot
- restart and enable postgresql
- check the replication with belvedere is ok
- switch back the webapp to somerset
- upgrade of belvedere
<< add a notification in the status.io page >>
- connect to the idrac: https://swh9-adm.inria.fr/
- stop the loaders and listers workers
- stop the indexers
- stop the scheduler runners + those in the tmux in saatchi
- **ensure the provenance experiment is stopped**
- on belvedere:
- stop puppet
- stop and disable postgresql**s ** (to avoid the restarts after the server reboots)
- perform the last upgrade of buster
- reboot
- upgrade to bullseye
- reboot
- check everything is going well after the reboot
- restart postgresql**s **
- start and enable the postgresql servers
- check the replication to somerset is ok
- reactivate puppet
- restart stopped services