Extend storage disk space for staging environment

As the current storage [1] is running out of space, we have planned a workaround to recycle db1.staging as a new storage server.

The high level plan could be:

• #5251 (closed): Recycle beaubourg (old proxmox hypervisor) as a kubernetes(archive-staging-rke2) compute node with "quite fast" ssd disks (in progress #5251 (closed))
• swh/infra/puppet/puppet-swh-site!698 (merged): Add db1 as a kubernetes node in the staging cluster (like storage1)
• Deploy a r/w objstorage service on db1 and make r/o objstorage service on storage1
  • Prepare the objstorage volume on db1
  • db1: Add label swh/objstorage=true
  • Adapt charts
    • prepare plan
    • db1: add r/o objstorage
    • node: Deploy multiplexer objstorage to use db1 and storage1's r/o objstorage
    • Check read-only clients (web, indexer) use the new read-only rpc with success
    • db1: add r/w objstorage
    • swh/infra/ci-cd/swh-charts!357 (merged): node: Deploy multiplexer objstorage to use db1 and storage1's r/w objstorage
    • swh/infra/ci-cd/swh-charts!359 (merged): Make writer use the new multiplexer r/w objstorage
    • Check contents are written in the db1 objstorage
    • Ensure no one is writing on storage1 objstorage
    • storage1: Decommission r/w objstorage
• Migrate contents from storage1 objstorage to db1 objstorage (until the warning threshold alert from zfs disk usage in storage1 subside)
  • swh/devel/snippets!22 (merged): Develop script
  • Move contents from storage1 to db1
  • Space used down to 90%
• swh/infra/puppet/puppet-swh-site!702 (merged): Adapt icinga alerts to the new threshold so current alert subside (we no longer write in that objstorage)
• #5311 (closed): Migrate the staging database to a kubernetes service managed by a postgresql operator (poc/validation/...)
• Rename db1 as storage2 (if poc is validated and the db is migrated from db1)

Several advantages:

• Add around 26To of disk space to staging objstorage
• Migrate the staging database to a faster server (no really important because cassandra is the main backend)
• Start to play with an operator managed postgresql

[1] #5150 (closed)