HTTP protocol changed from HTTP2 to HTTP/1.0

Hi all,

I am facing this problem since yesterday :slight_smile:
I quiet often use web.dev to test and audit some applications/webpages/etc.

Since yesterday, all the webapps hosted with hestia tested with web.dev has the complain that cached resources or local assets are served with HTTP/1.0 protocol

image

It never happend before since HTTP2 should be by default in HESTIA

and here you can see a portion for webdomain access.log

image

So I am thinking now that it could be related to this downgrade/change

Maybe I am wrong, maybe it’s just on my hestia setup and I’ll investigate further, but can someone else advice if he has the same issue?

Thank you!

As far as I remember, HTTP/2 requires https. Make sure that you have a valid TLS certificate for the site you are testing.

1 Like

Yes, everything is under HTTPS and is working properly. Maybe the nginx update breaks something on my setup.

I use default templates (web, backend,proxy) except for some. All with this issue.

I ran a test on one of my sites, and indeed Lighthouse returns http/1.1 although the site is served via HTTP/2. It seems to be a bug on Lighthouse.

You can also verify HTTP/2:

Thank you for taking some time to check it.
Yes, it seems is only happening with local resources.
But on my logs it shows HTTP/1.0 or 1.1

What about yours? :slight_smile:

In the log file (Hestia Web UI > Web tab > Access Log), I see both http/1.1 and http/2 requests. I filtered the log to only show requests from my browser’s IP (that I know I’m using a browser that supports HTTP/2) and I only see HTTP/2.

To see only the requests you have made, run the following on the command line:
grep YOUR.IP.OVER.HERE /var/log/apache2/domains/yourdomain.here.log

HI @Felix on my logs i still see HTTP/1.0 from my IP.
It makes sense what you say, i will check on my installation what’s wrong.

Thank you.

If I use curl i got this message :

curl -I -k --http2 https://MYURL.COM
curl: (92) HTTP/2 stream 0 was not closed cleanly: PROTOCOL_ERROR (err 1)

So i check the error on google and nothing is working on my side.
(by adding proxy_request_buffering off; on nginx, adding protocol HTTP2 on apache, etc)

Has someone any tips here? :slight_smile:

PS
This is happening on 2 hestia servers in 2 different DCs. Both with Debian 10 and Apache (web), Nginx (Proxy) and multiphp (FPM backend)

Thanks

I can not think of any other possible fix for this :frowning: Since you said the problem was not there from the beginning, go though all the logs and changes, and try to figure out what changed (configs, software, etc).

If you can’t find anything to blame, then I guess you could try to fully replicate your setup to a brand new VPS (Hetzner has plans on a per hour basis so it should be very cost effective to run some tests). You can use curl to test. If the new VPS turns out to be OK, that would hint/suggest that something is off in ur current configuration.

Soz, no other aces up my sleeves :thinking:

1 Like

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.