ERR_CONNECTION_RESET when uploading large files on Azure
I'm using Blazor SfUploader Server-Side and I'm having trouble with uploading large files when deployed on Azure.
Locally I don't have any problems so most likely it is a setting/configuration on Azure but I can't find which one.
When uploading I get this error:
syncfusion-blazor.min.js:1 POST https://myWebApp.azurewebsites.net/upload/chunk net::ERR_CONNECTION_RESET
Because I've enabled Chunking and set a high RetryCount the file is eventually uploaded but it takes a long time.
This is my Blazor code:
< SfUploader ID="UploadFiles" AllowedExtensions=".xml" MaxFileSize="429496720" SequentialUpload="true" DropArea="#DropArea">
< UploaderAsyncSettings SaveUrl="upload/chunk" ChunkSize="25000000" RetryCount="10" RetryAfterDelay="3000">
The controller is the one from https://blazor.syncfusion.com/documentation/file-upload/chunk-upload/ and https://www.syncfusion.com/forums/150303/upload-large-file-to-azure-storage-using-chunk-upload.
Here's my controller code:
namespace FileUpload.Controllers
{
[DisableRequestSizeLimit]
public class UploadController : Controller
{
[HttpPost("upload/chunk")]
public async Task SaveChunked(IList chunkFile, IList UploadFiles)
{
try
{
if ((chunkFile == null || !chunkFile.Any()) && UploadFiles != null && UploadFiles.Any())
{
// File is smaller than chunk size:
// Upload to blob storage:
return;
}
if (chunkFile == null) throw new ArgumentException("No data uploaded");
foreach (var file in chunkFile)
{
var httpPostedChunkFile = HttpContext.Request.Form.Files["chunkFile"];
var chunkIndex = HttpContext.Request.Form["chunk-index"];
var totalChunk = HttpContext.Request.Form["total-chunk"];
using (var fileStream = file.OpenReadStream())
{
if (Convert.ToInt32(chunkIndex) <= Convert.ToInt32(totalChunk))
{
var streamReader = new MemoryStream();
fileStream.CopyTo(streamReader);
var byteArr = streamReader.ToArray();
var content = new byte[] { };
if (HttpContext.Session.Get("streamFile") != null)
{
content = HttpContext.Session.Get("streamFile").Concat(byteArr).ToArray();
}
else
{
content = byteArr;
}
HttpContext.Session.Set("streamFile", content);
}
if (Convert.ToInt32(chunkIndex) == Convert.ToInt32(totalChunk) - 1)
{
var fileArray = HttpContext.Session.Get("streamFile");
using (var fileStreams = new FileStream(httpPostedChunkFile.FileName, FileMode.Create))
{
foreach (var stream in fileArray)
{
fileStreams.WriteByte(stream);
}
fileStreams.Seek(0, SeekOrigin.Begin);
HttpContext.Session.Remove("streamFile");
// new MemoryStream(myByteArray);
// Save file: await blockBlob.UploadFromStreamAsync(fileStreams);
}
}
}
}
}
catch (Exception e)
{
Console.WriteLine(e.Message);
}
}
}
}
But with large files the controller isn't even triggered.
Smaller files are fine. Multiple small files are fine as well.
The error:
Please assist.
SIGN IN To post a reply.
7 Replies
Thanks for the reply and the sample.
I tried your code suggestions in my code, some of them I already had, but with the same problem.
Next, I tried your sample application and uploaded to my Azure account.
Although I still get connection errors, due to your retry mechanism I could eventually upload the files:
I will now compare your code with mine. I did see you are using some deprecated code.
I also submitted a support ticket to Microsoft Azure, hopefully they will give some advise on how to get rid of the connection aborts, because large files are now uploaded several times before their finished making the upload process take very long.
SIGN IN To post a reply.
