handle larger volumes of data

[matti]

I opened #9 , which fixes the immediate problem, but I'm seeing

441e28f31fd7: Retrying in 1 second

and

blob upload unknown

when I push to docker registry running behind the proxy.

What else needs to be configured larger?

[mesudip]

I've been using registry:2 behind Nginx-proxy for 6months now. I never had any issues after setting client_max_body_size=2g.

blob upload unknown looks like an issue with the registry.

Can you try with docker run -d registry:2 instead and see if it works?

[matti]

it does work - I have used it like that for weeks.

My images are fairly large and I'm pushing them in parallel - I think there is some nginx setting which causes this.

On 25. Jan 2020, at 17.02, Sudip Bhattarai [email protected] wrote:

 I've been using registry behind Nginx-proxy for 6months now. I never had any issues.

blob upload unknown looks like an issue with the registry.

Can you try with docker run -d registry:2 instead and see if it works?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

[matti]

I also see this sometimes with eu.gcr.io - but I see it more when piping data through nginx-proxy now. I think the key is that I push 10 images in parallel which causes a lot of data being pushed and then some nginx (buffer?) runs low?

[mesudip]

Can you try modifying proxy_buffering off; inside http{ ...} in /etc/nginx/nginx.conf file?

There are other settings mentioned available in Nginx documentation

I don't know if it works, it's my random guess. The command I use is

docker cp <nginx-container>:/etc/nginx/nginx.conf nginx.conf
vim nginx.conf  # then edit the file
docker cp nginx.conf <container-name>:/etc/nginx/nginx.conf
docker exec  <container-name> nginx -s reload 

[mesudip]

Hey @matti I have found two configuration changes that might solve the issue.

proxy_buffering off; 
proxy_request_buffering off;

Furthermore, have you checked the file /var/log/nginx/error.log in the container? It might contain the detail of the issue.

I have a WIP status in mesudip/nginx-proxy:test that contains the above two configuration changes. Can you check it out?

[matti]

thanks, I try out soon!