feat(zot): keep only latest for the backstage repository
yaml / yaml (pull_request) Successful in 33s
yaml / yaml (pull_request) Successful in 33s
Backstage dev images are tracked by digest behind :latest, so the per-commit sha tags pushed before that change only accumulate. Retain every tag in all other repositories via an explicit catch-all policy. Co-Authored-By: Claude Opus 5.5 <[email protected]>
This commit is contained in:
+5
-2
@@ -26,8 +26,11 @@ SPIRE 认证链路已经验证,但当前没有常驻 publisher 或删除授权
|
||||
`/var/lib/registry` 是 `emptyDir`,仅用于运行时本地工作数据。
|
||||
|
||||
两个单副本实例共用同一 bucket 和前缀:`zot` 负责鉴权写入,`zot-reader` 负责匿名
|
||||
读取。关闭跨仓库 dedupe,不额外部署 Redis/DynamoDB 缓存。只有写入实例启用 GC,
|
||||
暂不配置自动删除已发布版本的 retention policy。增加副本、启用 dedupe 或搜索等
|
||||
读取。关闭跨仓库 dedupe,不额外部署 Redis/DynamoDB 缓存。只有写入实例启用 GC。
|
||||
retention policy 只针对 `panxiao81/backstage`:仅保留 `latest`,历史 sha tag 与失去
|
||||
tag 的 digest 在 24h 后回收(dev 镜像由 Flux 跟踪 `latest` 的 digest)。其余仓库由
|
||||
`**` 兜底策略保留全部 tag,行为与未配置 retention 时一致;新增按仓库的清理规则时,
|
||||
必须放在兜底策略之前,否则不会生效。增加副本、启用 dedupe 或搜索等
|
||||
扩展前,需要重新检查共享元数据与缓存的持久化要求。
|
||||
|
||||
凭据链路:
|
||||
|
||||
@@ -27,6 +27,21 @@ configFiles:
|
||||
"gc": true,
|
||||
"gcDelay": "24h",
|
||||
"gcInterval": "24h",
|
||||
"retention": {
|
||||
"delay": "24h",
|
||||
"policies": [
|
||||
{
|
||||
"repositories": ["panxiao81/backstage"],
|
||||
"deleteUntagged": true,
|
||||
"keepTags": [{ "patterns": ["^latest$"] }]
|
||||
},
|
||||
{
|
||||
"repositories": ["**"],
|
||||
"deleteUntagged": true,
|
||||
"keepTags": [{ "patterns": [".*"] }]
|
||||
}
|
||||
]
|
||||
},
|
||||
"storageDriver": {
|
||||
"name": "s3",
|
||||
"region": "us-east-1",
|
||||
|
||||
Reference in New Issue
Block a user