-
Bug
-
Resolution: Unresolved
-
Major
-
rhos-17.1.8
-
None
-
5
-
False
-
-
False
-
?
-
openstack-tripleo-common-15.4.1-17.1.20250908140822.e5b18f2.el9ost
-
None
-
-
-
-
Proposed
-
Swift Sprint 3, Swift Sprint 4
-
2
-
Moderate
-
Customer Case Attached
swift swift_container_replicator container in unhealthy state due to slow replication.
29 hours ago Up 2 minutes (unhealthy) swift_container_replicator > >cat /var/cache/swift/container.recon > >{"container_auditor_pass_completed": 1220.1614661216736, "container_audits_failed": 0, "container_audits_passed": 488066, "container_audits_since": 1756560192.6049159, "container_updater_sweep": 4923.433475255966, "replication_last": 1756564267.295261, "replication_stats": {"attempted": 244033, "deferred": 0, "diff": 4118, "diff_capped": 0, "empty": 0, "failure": 0, "hashmatch": 0, "no_change": 239915, "remote_merge": 0, "remove": 0, "rsync": 0, "start": 1756559386.646013, "success": 244033, "ts_repl": 0}, "replication_time": 4880.64924788475}
This setup is 3 object storage nodes with one dedicated disk each.
Replication seems to work but it is significantly slower than the healthcheck. Its likely the disks is slow.
Tried to modify the default config but still having issues with unhealthy container:
[container-replicator] concurrency = 8 interval = 240
Looking for review and assistance with the following:
1. Config that will allow swift_container_replicator container to stay healthy.
2. How to modify container-replicator config in the OOO templates. It looks like the following should work but does not:
ObjectStorageExtraConfig:
swift::config::swift_container_config:
container-replicator/concurrency:
value: 2
container-replicator/interval:
value: 240
From /etc/puppet/modules/swift/README.md #### swift_container_config
Same as `swift_config`, but path is `/etc/swift/container-server.conf`
