Refs. swh/infra/sysadm-environment#5512

It was misindented from 2 characters.

Otherwise, we schedule and fill-in queues which are not consumed. This then
escalates to spurious infra alerts.

Those loaders were stopped to avoid running too much loaders in the staging infra.

It's mostly the same way to trigger it. Only some configuration manipulation
must slightly be adapted to allow the extra service to deploy it.

The scheduler-runner already allows it but it was used only in the
`start-runner --with-priority` call. Now, we'll be able to run if for both.

This is just the list of task types so name it properly.

The kubernetes scheduler should do the right thing.

Refs. swh/infra/sysadm-environment#5593

That pvc/pv will be accessible from 2 pods objstorage-{ro,rw} on the same
host.

Refs. swh/infra/sysadm-environment#5593

Avoid using ceph storage. It should be used for programs that does a moderate
amount of I/O but still needs "singleton" persistence.

Objstorage can be I/O intensive.

This reverts commit 33eac1f7.

Refs. swh/infra/sysadm-environment#5593

The local-persistent does not support the accessMode ReadWriteMany

Refs. swh/infra/sysadm-environment#5593

PV does not declare namespace.

Refs. swh/infra/sysadm-environment#5593

Add one objstorage-pathslicing-pvc to manage the next-version objstorage.
That way, the data stored won't get decommissionned when the argocd
next-version is stopped.

Refs. swh/infra/sysadm-environment#5593

As in objstorage, alter, etc...

Refs. swh/infra/sysadm-environment#5593

As objstorage, alter, etc... does

Refs. swh/infra/sysadm-environment#5593

Refs. swh/infra/sysadm-environment#5593

Refs. swh/infra/sysadm-environment#5593

Refs. swh/infra/sysadm-environment#5593

Refs. swh/infra/sysadm-environment#5593

Refs. swh/infra/sysadm-environment#5593

Refs. swh/infra/sysadm-environment#5593

Refs. swh/infra/sysadm-environment#5593

This allows to share the pv/pvcs declaration instead of duplicating it inside
the various templates using it (e.g. graph, alter, cassandra-checks, ...)

Refs. swh/infra/sysadm-environment#5593

(with minReplicaCount == maxReplicaCount)

The svn request are usually taking time to be processed.

A few number of scn or mercurial requests can delay the processing
of a majority of git repositories

Related to swh/infra/sysadm-environment#5600

This reverts commit 936f2bc1.

Related to swh/infra/sysadm-environment#5594 (comment 198679)

This reverts commit 0dcb1f7d.

Related to swh/infra/sysadm-environment#5594 (comment 198663)

Vincent Sellier authored 3 weeks ago and Guillaume Samson committed 3 weeks ago

The volumes were not shared between the ro and the rw instance
so the ro instance didn't see any contents

Vincent Sellier authored 3 weeks ago and Guillaume Samson committed 3 weeks ago

It's actually a fake read-only storage as there
is no read-only user configured on the k8ssandra
deployment

Vincent Sellier authored 3 weeks ago and Guillaume Samson committed 3 weeks ago

To be able to ingest origins on demand on this environment