Issue File download breaks when browser sends HTTP/1.0 request?

[Ksech]

We have a large file that we sometimes need customers to download from our server. We send them a link, and most download it fine. But sometimes they get a 404 error. We asked a customer for a screenshot today, and they took a photo instead so part of it is cropped out, but the link is correct, the browser appears to be Chrome (not sure because it's cut off in the photo) and it's showing the website's 404 page (not a generic browser error.)

I found it in the logs, and I can't make heads or tails out of this, but I think it has something to do with the server's response to the HTTP REQUEST. This is most of the attempts the customer made this morning - there are 3 HTTP/2.0 requests that appeared to have been somewhat successful (except that the file is actually over 200mb) so there may be other issues on the customer's end.

I've tested from multiple browsers on multiple computers and can always download the file successfully. Looking through the last few weeks of logs, I can see that the 404s on this file are ALWAYS an HTTP/1.0 request, and are often followed up by a successful download with an HTTP/2.0 request but not always.

My concern is that the server somehow isn't communicating clearly with some browsers (or at certain times?) regarding the HTTP protocols. I also realize it could be an ancient browser, but I think that's highly unlikely, and definitely not the case this morning.

Any ideas?

 

[Kaspar]

I also realize it could be an ancient browser, but I think that's highly unlikely, and definitely not the case this morning.

I wish I had explanation for the issue you're facing, but unfortunately I don't. However, for whats it's worth, I want to mention that if you hover over the little "gentlemen" icon with your mouse a tooltip shows with the User-Agent info of the log entry. Which can be useful to determine if an old browser is used.

 

[ChristophRo]

Hmm, I think that HTTP 1.0 does not use SNI and thus such a client is unable to connect to most https:// sites nowadays. (as we don't use dedicated IP addresse per site anymore)

So this client would land on the default site of your server and not this particular website, so the 404 is explainable

 

[Ksech]

if you hover over the little "gentlemen" icon with your mouse a tooltip shows with the User-Agent info of the log entry

Thank you Rasp - I didn't know that! But I did know that this one specific customer was using a modern browser, not an ancient one.

 

[Ksech]

So this client would land on the default site of your server and not this particular website, so the 404 is explainable

Thank you ChristophRo, but I really don't understand? The customer did NOT land on the server's default site, they landed on the 404 page for the correct website.

Also, I thought if the browser requested HTTP/1.0 that the server responds with "no, you need to request HTTP/2.0" or something like that. That's where I suspect the problem is - some sort of miscommunication happening here regarding the HTTP protocol, and I fear the problem's on my server, not on the customer's browser.