#3325 【代码仓】项目分支较多时,加载项目主页较慢

Closed
created 1 year ago by avadesian · 5 comments
avadesian commented 1 year ago
参考项目: https://openi.pcl.ac.cn/OpenI/Paddle 参考Gitee加载,似乎是分批次加载分支信息
avadesian added the
enhancement
label 1 year ago
tanglj was assigned by avadesian 1 year ago
tanglj added this to the V20221228 milestone 1 year ago
tanglj commented 1 year ago
Collaborator
#360 平台首页推荐项目里的Paddle好像点击没反应 https://openi.pcl.ac.cn/zeizei/OpenI_Learning/issues/360
ychao_1983 was assigned by tanglj 1 year ago
tanglj removed their assignment 1 year ago
ychao_1983 was unassigned by tanglj 1 year ago
wangj modified the milestone from V20221228 to V20230215 1 year ago
ychao_1983 self-assigned this 1 year ago
zhoupzh self-assigned this 1 year ago
ychao_1983 commented 1 year ago
Owner
前端无变化,后端把项目页的分支查询接口的命令从git show-ref --heads改为了git for-each-ref --format='%(refname:short)' refs/heads/ ,在生产环境paddle项目下测试这个接口比原来的show-ref明显快一些. 另外生产环境慢是因为在paddle项目下 refs/remotes下面有很多无效的目录,之前手工删除并执行gc后,生产环境已经比原来快了不少,从8每秒左右到3秒左右后来发现镜像同步后删除的目录还会同步回来,所以手工删除的方法无效(git show-ref --heads会受这些目录影响,git for-each-ref --format='%(refname:short)' refs/heads/测试时秒回,感觉不受影响,应该可以优化项目打开时间)。 验证时在开发和测试环境 项目分支显示和原来一样即验证通过。上生产环境再测试paddle项目看下实际效果。
wangj added the
test
label 1 year ago
tanglj removed the
test
label 1 year ago
ychao_1983 added the
test
label 1 year ago
wangj was assigned by ychao_1983 1 year ago
ychao_1983 commented 1 year ago
Owner
如果分支数很多,到了上万的级别,加载还是会慢,例如pytorch有1万多个分支,这种目前还没有解决。 gitea官方有一个issue(https://github.com/go-gitea/gitea/issues/14180)也是反馈相同问题目前还是open状态。
ychao_1983 removed the
test
label 1 year ago
wangj was unassigned by ychao_1983 1 year ago
ychao_1983 added the
test
label 1 year ago
wangj was assigned by ychao_1983 1 year ago
wangj commented 1 year ago
Owner
开发环境挑选了2个项目验证,项目分支显示和原来一样。上线了再观察。
wangj commented 1 year ago
Owner
生产环境 OpenI/aiforge项目510毫秒, OpenI/Paddle项目第一次加载耗时6秒(比之前快了),第二次加载884毫秒。 关闭此单。
wangj closed this issue 1 year ago
Sign in to join this conversation.
No Milestone
4 Participants
Notifications
Due Date

No due date set.

Dependencies

This issue currently doesn't have any dependencies.

Loading…
There is no content yet.