Home/ Questions/Q 3991094
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-05-20T06:34:25+00:00

Is there any way (server or client side) to force the browser to pull

  • 0

Is there any way (server or client side) to force the browser to pull a new version of a file (image) from the server. The image in question is otherwise cached for a long time. I know I can append a random number, for instance, to the URL of the image but this is not acceptable in this situation. I need for the image to be refresh from the exact same URL.

What I’m doing: a YouTube like portal where users upload videos. Each video has a thumbnail which is shown on various pages on the portal. User can, at any time, change the thumbnail (he can select from three generated thumbnails). So when this happens (a new image overwrites the ‘original’ image), I wan’t to refresh the video’s thumbnail so that the owner (I don’t care if other users see the old thumbnail) will see the new thumbnail no matter where the thumbnail is shown.

I’m afraid this can’t be done but I’m asking here just to be sure.

update: I’m using nginx and PHP on the server side

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    You could use ETAGs on your thumbnails. This would prevent the transmission of the actual thumbnail data if it hasn’t changed (i.e. still has the same hash). However, you would still face the clients HTTP requests to check if the ETAG has changed (normly to be answered by HTTP 304.

    But combined with a rather short freshness threshold (say a couple of minutes), you could achieve tradeoff between caching and freshness while still conserving resources. If you need absolute freshness, you might have to stick to ETAGs though. If you create a clever hash function, you could handle the ETAG requests on your frontend loadbalancer (or at least near it) which could thus be rather cheap.

    Edit: Add alternative from my other comment.

    An alternative could be to use added request parameters to force a re-fetch when the resource changed as suggested in another answer. A variation of that schema (which is used by many Rails applications) is to append the timestamp of the last change (or some kind of hash) as a parameter to the file which only changes when the file actually does change. Something like this, or one of the above methods, is actually the only way to be really sure to not have unnecessary cache validation requests while at the same time having always the freshest resource.

    • 0