Hi, I have a source code license for version 1.6 of Mediachase file uploader which I'm using on one of my main Websites, however, customers sometimes complain about the file upload failing when uploading large files (approx 100MB and greater).

Locally, I can upload much more than this, but users in the West Coast US and in China connecting to my UK server seem to complain more than others.

I see the error:

Reqest was aborted
   at Mediachase.FileUploader.McHttpModule.Reqest was aborted
   at Mediachase.FileUploader.McHttpModule.694a7e1d01f817ae

Constantly in the Event log of my server, and the minority of users who are unable to use the service they are paying for because of this problem are obviously rather annoyed.

Do you have any possible explanations as to why this may be happening, and why I can't replicate the problem?

Mediachase FileIUploader.NET component reads and parses incoming request and throw
the "Reqest was aborted" exception if connection is broken between server and client.

May be clients use proxy? And the proxy closes long-time connection.

Thanks for the response. With that in mind, I've increased the Connection Timeout in IIS to see if that improves the situation.

If it helps you, please will send us your results.

After all this time, I've finally figured it out.

If I recycle the Application Pool in IIS during an upload, the "Request Reqest was aborted" error message is shown on screen.  This is more likely to occur for users during large uploads, since our application is set to recycle every 240 minutes (4 hours).  If a user in unlucky enough to span this period, then the user will receive this message. 

This has been happening to our users every single day and has done since we started using this product, but we've only just worked out what was causing it.  Obviously, this is no good for a production system which is aspiring for 100% uptime.

Once I had worked out this problem, I then decided to retest what happened with Load Balancing, since I thought that load balancing switchover could have an effect.

So, I disabled all but one Web server (server 1) in the hardware Load Balancer then started a large upload.  I then brought another server up (server 2) and disabled "server 1".  The progress bar then stopped functioning, but the upload was ultimately successful.

Obviously, now that I know the cause of the problem, I'd like a speedy solution to solving it!

What can we do to sort this out?

It would also tally with the fact that China users (who have high latency connections [often a ping time of > 500ms] to our servers) are going to spend the longest time uploading and are therefore more likely to suffer this problem.

Could it be that the product is storing information about the request in the Application which, when recycled, is being lost?

When I look at the source code it does appear that the product is storing values in the Application rather than persisting them to disk or (my preference) SQL Server.

adrian2:

If I recycle the Application Pool in IIS during an upload, the "Request Reqest was aborted" error message is shown on screen.  This is more likely to occur for users during large uploads, since our application is set to recycle every 240 minutes (4 hours).  If a user in unlucky enough to span this period, then the user will receive this message.

Unless I've mistaken, the recycling process restarts Asp.Net applications. So all open connections were closed and FileUploader throw Request aborted error.

Can you increase recycle application pool timeout? For example set to equal 1 day.

adrian2:

Could it be that the product is storing information about the request in the Application which, when recycled, is being lost?

When I look at the source code it does appear that the product is storing values in the Application rather than persisting them to disk or (my preference) SQL Server.

Version 2.0 allows to change Temp File and Progress Storage Provider, look

http://www.mediachase.com/FileUploader2/documentation/customprovider.html

Thanks for letting me try out v2.0 without an upload limit.  The upload progress is better across the load balancer with the SQL Backed Progress and actually, the UploadProgress table allows me to track upload speeds.

I can also add another table which will link the current logged in user to the upload to see which users are having problems uploading which will allow me to work out whether users in a particular region are getting poor uploads.

Do you have an upgrade path from 1.7 to 2.0?  We bought the source code license for v1.7, but we haven't done anything with the source code, so I'd like to upgrade to v2.0 Corporate.  Please let me know how much it would cost.

Thanks,
Adrian.

Please email sales@mediachase.com with your upgrade question and someone will be able to assist you.

Mediachase Software