-
Type:
Bug
-
Status: Done
-
Resolution: Done
-
Affects Version/s: 6.18.0, 6.18.1, 7.2.1
-
Fix Version/s: 6.21.0
-
Component/s: Docker, Docker Image, Replication, Smart Remote Repository
-
Labels:
-
Severity:High
With enabled OFFLINE mode of "releases-docker-registry-remote" (our own registry)
Artifactory.log:
2020-03-09 09:07:53,610 [http-nio-127.0.0.1-8081-exec-18] [INFO ] (o.a.a.d.r.v.r.v.DockerV2VirtualRepoHandler:117) - Fetching docker manifest for repo 'test/boom' and tag '1.4.1-1' 2020-03-09 09:07:53,629 [http-nio-127.0.0.1-8081-exec-18] [ERROR] (o.a.a.d.r.v.DockerV2RemoteRepoHandler:448) - Missing Manifest from releases-docker-registry-remote 'v2/test/boom/manifests/1.4.1-1' not found at releases-docker-registry-remote-cache:test/boom/1.4.1-1/list.manifest.json
On the client site you get:
docker pull releases-docker-registry-remote.our.domain/test/boom:1.4.1-1
Error response from daemon: manifest for releases-docker-registry-remote.our.domain/test/boom:1.4.1-1 not found: manifest unknown: The named manifest is not known to the registry.
After disabling offline mode:
2020-03-09 09:08:54,997 [http-nio-127.0.0.1-8081-exec-15] [INFO ] (o.a.u.r.s.a.c.r.r.ReplicationConfigService:118) - Updating remote replication config for repo releases-docker-registry-remote 2020-03-09 09:08:55,002 [http-nio-127.0.0.1-8081-exec-15] [INFO ] (o.a.c.CentralConfigServiceImpl:620) - Reloading configuration... old revision 408, new revision 409 2020-03-09 09:08:55,143 [http-nio-127.0.0.1-8081-exec-15] [INFO ] (o.a.c.CentralConfigServiceImpl:379) - New configuration with revision 409 saved. 2020-03-09 09:08:55,143 [http-nio-127.0.0.1-8081-exec-15] [INFO ] (o.a.s.ArtifactoryApplicationContext:520) - Artifactory application context set to NOT READY by reload 2020-03-09 09:08:55,294 [http-nio-127.0.0.1-8081-exec-15] [INFO ] (o.a.s.ArtifactoryApplicationContext:520) - Artifactory application context set to READY by reload 2020-03-09 09:08:55,295 [http-nio-127.0.0.1-8081-exec-15] [INFO ] (o.a.c.CentralConfigServiceImpl:638) - Configuration reloaded. 2020-03-09 09:09:01,303 [http-nio-127.0.0.1-8081-exec-18] [INFO ] (o.a.a.d.r.v.r.v.DockerV2VirtualRepoHandler:117) - Fetching docker manifest for repo 'test/boom' and tag '1.4.1-1' 2020-03-09 09:09:01,438 [http-nio-127.0.0.1-8081-exec-3] [INFO ] (o.a.a.d.r.v.DockerV2RemoteRepoHandler:196) - Fetching docker blob 'sha256:8cb108ce1232641068ed3c62e87df9645de221a7803c4dde0520f6690a64f83b' 2020-03-09 09:09:01,507 [http-nio-127.0.0.1-8081-exec-16] [INFO ] (o.a.a.d.r.v.DockerV2RemoteRepoHandler:196) - Fetching docker blob 'sha256:71995da3e5c0ea137021e624edc5de164c62d7ff7079149733bbdaa21a4b2945'
Same result when we do not use OFFLINE mode, but instead block traffic via iptables.
Thats really a porblem, because we have central artifactorys and caches on mutliple datacenters. At the moment they depend on the central instance and a outage of that instance will make the caches useless. That makes the artifactory docker remote repos also useless in terms of availability and reliability.
The issue is not resolved, also it is claimed to be in RTFACT-18779 , RTFACT-20905
- clones
-
- Done
