You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Are you running your instance in front of CloudFlare or a similar service? If so, please check that you are not proxying. Many of these proxying services, like CloudFlare, have an upload limit. For example, I'm using CloudFlare free DNS; I disabled the CloudFlare proxy and used raw DNS forward because the free CloudFlare DNS proxying service has an upload limit of 100MB. When I updated my DNS record for raw export and turned off the proxying, the upload went through without a hitch. I hope this helps you. To find your issue, please draw an exact network path of your setup, including services like CloudFlare.
CloudFlare is known to cause issues like this; however, it's hard to find because only a few people know about the 100MB upload size limit when using the CloudFlare free DNS upload limit.
The docker push failed when the layer is larger than 1 GB.
Storage is set to S3 (MinIO). The chunk size has been increased.
REGISTRY_STORAGE_S3_CHUNKSIZE: 514288000
I see that during upload the size of the root disk under GitLab is reduced. Probably the tmp file is placed in /var/lib/nginx/body.
Is there a way to set the proxy_request_buffering variable in the gitlab-registry nginx conf file?
GitLab version: 16.3.0 (sameersbn/gitlab:16.3.0)
Registry version: registry:2.8.2
The text was updated successfully, but these errors were encountered: