production/scrubber: checkers directory hashes are crash looping

Reasons are unclear. The equivalent processes in the staging environment are fine.

That might be related to the various warnings seen in the logs [1]:

• scrubber v3.0 deprecated the subcommand 'swh scrubber check storage' to the benefit of 'swh scrubber run' (probably unrelated)
• a missing postgresql storage configuration. It's just noise, because the checkers run on a cassandra storage!
• another warning about the Cluster(protocol_version) to set

[1]

check-scrubber-migration module: scrubber
check-scrubber-migration current code version: 7
check-scrubber-migration version: 7
check-storage-migration No postgresql configuration found!
check-storage-migration
strorage-checker Running swh command scrubber check storage storage-cassandra-hashes-directory
strorage-checker DeprecationWarning: The command 'storage' is deprecated.
strorage-checker WARNING:cassandra.cluster:Downgrading core protocol version from 66 to 65 for 192.168.100.183:9042. To avoid this, it is best practice to explicitly set Cluster(protocol_version) to the version supported by your cluster. h
strorage-checker WARNING:cassandra.cluster:Downgrading core protocol version from 65 to 5 for 192.168.100.183:9042. To avoid this, it is best practice to explicitly set Cluster(protocol_version) to the version supported by your cluster. ht
strorage-checker INFO:cassandra.policies:Using datacenter 'sesi_rocquencourt' for DCAwareRoundRobinPolicy (via host '192.168.100.183:9042'); if incorrect, please specify a local_dc to the constructor, or limit contact points to local clust
strorage-checker WARNING:cassandra.cluster:Downgrading core protocol version from 66 to 65 for 192.168.100.187:9042. To avoid this, it is best practice to explicitly set Cluster(protocol_version) to the version supported by your cluster. h
strorage-checker WARNING:cassandra.cluster:Downgrading core protocol version from 65 to 5 for 192.168.100.187:9042. To avoid this, it is best practice to explicitly set Cluster(protocol_version) to the version supported by your cluster. ht
strorage-checker INFO:cassandra.policies:Using datacenter 'sesi_rocquencourt' for DCAwareRoundRobinPolicy (via host '192.168.100.187:9042'); if incorrect, please specify a local_dc to the constructor, or limit contact points to local clust

changed the description

mentioned in commit swh/infra/ci-cd/swh-charts@1d6e74a2

scrubber v3.0 deprecated the subcommand 'swh scrubber check storage' to the benefit of 'swh scrubber run' (probably unrelated)

Adapted accordingly. It's still working for staging. As expected for production process though, this does not change the current behavior. It is still crash looping (but there isn't the deprecation warning about it any longer).

changed title from production/scrubber: checkers directory are crash looping to production/scrubber: checkers directory hashes are crash looping

changed the description

It must be related to the other issue on the scrubber [1] Mostly all storage-cassandra-hashes-* with * an object type are deactivated with the comment [2].

I noticed the following behavior. First, the process starts then it's marked as 'completed' then it crash loops...

Aligning this process with the other. No point in making the machines churning in crashloops.

[1] swh/devel/swh-scrubber#4699

[2]

# until https://gitlab.softwareheritage.org/swh/devel/swh-scrubber/-/issues/4699 fix

.

mentioned in commit swh/infra/ci-cd/swh-charts@4f99bd5c

changed the description