Try our conversational search powered by Generative AI!

AI OnAI Off

Custom ContentProvider - 404 on document blob, first load [11.9.2]

Coenraad Myburgh
Coenraad Myburgh
Vote:
 

Hi,

I am experiencing an issue with our custom ContentProvider (after upgrading to 11.9.2) where a document (pdf) is throwing a 404.

The weird thing is that it only does this the very first time it is loaded into the ContentProvider cache. When I refresh it works as expected but after +-5 min (ContentProvider cache duration) the same thing happens.

Sample url: http://localhost/globalassets/documents-repository/xxx/Latest/Latest.pdf

Execution:

The first time executing the request for the file it runs through the LoadContent method and populates the Blob:

protected override IContent LoadContent(ContentReference contentLink, ILanguageSelector languageSelector)
{
   var content = DocumentRepository.Service.Get(contentLink);

    //Other types handled here

	if (content is DocumentFile)
	{
		var documentContent = GetDefaultContent(ContentLoader.Service.Get(content.ParentLink),
			new string[] { ".JPG", ".JPEG", ".GIF", ".PNG" }.Contains(Path.GetExtension((content as DocumentFile).Href), StringComparer.OrdinalIgnoreCase) ? ContentTypeRepository.Load(typeof(ApiDocumentImage)).ID : ContentTypeRepository.Load(typeof(ApiDocument)).ID,
										languageSelector);

	documentContent.ContentGuid = content.ContentGuid;
	documentContent.ContentLink = contentLink;
	documentContent.Name = content.Name;
	(documentContent as MediaData).Changed = DateTime.Now;
	(documentContent as MediaData).StartPublish = DateTime.Now;
	(documentContent as MediaData).RouteSegment = content.Name;
	(documentContent as MediaData).Status = VersionStatus.Published;
	(documentContent as IApiMetadata).Href = (content as DocumentFile).Href;
	(documentContent as ISupportsDataDate).DataDate = (content as DocumentFile).DataDate;
	(documentContent as IApiMetadata).EntityId = (content as DocumentFile).EntityId;
	(documentContent as IVersionable).IsPendingPublish = false;

	try
	{
		var provider = ServiceLocator.Current.GetInstance().GetProvider(new Uri("epi.fx.blob://apidoc/"));
		var extension = Path.GetExtension((documentContent as IApiMetadata).Href);
		var blob = provider.CreateBlob((documentContent as MediaData).BinaryDataContainer, string.IsNullOrEmpty(extension) ? ".null" : extension) as ApiDocumentBlob;
		blob.Href = (documentContent as IApiMetadata).Href;
		(documentContent as MediaData).BinaryData = blob;
	}
	catch (Exception ex)
	{
		IApiMetadata apiMeta = (documentContent as IApiMetadata);
		var href = apiMeta == null ? "" : apiMeta.Href;
		Logger.Critical(href + ": " + ex.Message, ex);
		throw;
	}

	return documentContent;
}

Directly after this I get an error in the logs:

2018-11-03 18:23:23,268 [26] ERROR EPiServer.Global: Unhandled exception in ASP.NET
System.Web.HttpException (0x80004005): Not Found.
   at EPiServer.Web.BlobHttpHandler.System.Web.IHttpAsyncHandler.EndProcessRequest(IAsyncResult result)
   at System.Web.HttpApplication.CallHandlerExecutionStep.System.Web.HttpApplication.IExecutionStep.Execute()
   at System.Web.HttpApplication.ExecuteStepImpl(IExecutionStep step)
   at System.Web.HttpApplication.ExecuteStep(IExecutionStep step, Boolean& completedSynchronously)
System.Web.HttpException (0x80004005): Not Found.
   at EPiServer.Web.BlobHttpHandler.System.Web.IHttpAsyncHandler.EndProcessRequest(IAsyncResult result)
   at System.Web.HttpApplication.CallHandlerExecutionStep.System.Web.HttpApplication.IExecutionStep.Execute()
   at System.Web.HttpApplication.ExecuteStepImpl(IExecutionStep step)
   at System.Web.HttpApplication.ExecuteStep(IExecutionStep step, Boolean& completedSynchronously)

When I refresh, the code steps into the OpenRead method of the custom blob successfully and the file loads correctly:

public class ApiDocumentBlob : Blob
{
	internal string Href { get; set; }

	public ApiDocumentBlob(Uri id)
		: base(id)
	{
	}

	public override System.IO.Stream OpenRead()
	{
		var uriBuilder = new UriBuilder(ApiManager.Instance.ApiDocumentUrl);

		uriBuilder.Path = Href;
		var webClient = new WebClient();
		var url = uriBuilder.Uri.ToString();
		return webClient.OpenRead(url);
	}
}

Subsequent refreshes works as well, due to the ContentProvider cache I presume:

//Other SetCacheSettings overrides look the same

protected override void SetCacheSettings(IContent content, CacheSettings cacheSettings)
{
	var fiveMinutes = TimeSpan.FromMinutes(5);
	cacheSettings.AbsoluteExpiration = DateTime.Now.Add(fiveMinutes);
	cacheSettings.SlidingExpiration = Cache.NoSlidingExpiration;

	base.SetCacheSettings(content, cacheSettings);
}

If I wait 5+ minutes for the cache to expire I experience exactly the same symptoms as above.

The fact that it does work eventually makes me think that the segments etc. are setup correctly.

I am busy trying to setup my own BlobHttpHandler to try and see if I can figure out why exactly it is failing.

Any other ideas would be very much appreciated.

#198708
Nov 04, 2018 8:08
Coenraad Myburgh
Coenraad Myburgh
Vote:
 

So I setup a custom BlobHttpHandler and copied over some of the internal code.
I managed to track down the issue to this method:

private bool ProccessBlobRequest(HttpContextBase context, out Blob blob)
{
	blob = (Blob)null;
	IContent content = this.ContentRouteHelper.Service.Content;

    //this.RoutableEvaluator.Service.IsRoutable returns false when I get the 404, basically it is checking if the content is published via the PublishedStateAssessor.
	var isPublished = PublishedStateAssessor.Service.IsPublished(content);

	if (content == null || !this.RoutableEvaluator.Service.IsRoutable(content))
		return false;
	if (!content.QueryDistinctAccess(AccessLevel.Read))
	{
		this.AccessDeniedHandler.Service.AccessDenied(context);
		return true;
	}
	if (!this.HasAssetOwnerAccess(content))
	{
		this.AccessDeniedHandler.Service.AccessDenied(context);
		return true;
	}
	DateTime modifiedDate = DateTime.Today;
	if (content is IChangeTrackable)
		modifiedDate = ((IChangeTrackable)content).Changed;
	this.SetCachePolicy(context, modifiedDate.ToUniversalTime());
	if (this.NotModifiedHandling(context, modifiedDate))
		return true;
	blob = this.GetBlob(context);
	return blob != null;
}

It seems that for some reason it thinks the content is not published. The this.RoutableEvaluator.Service.IsRoutable method checks if published then returns false.
Upon refresh the content is recognized as published and the blob is processed.


#198710
Nov 04, 2018 19:25
Coenraad Myburgh
Coenraad Myburgh
Vote:
 

I might have found a workaround by setting the StartPublish date 10minutes in the past:

protected override IContent LoadContent(ContentReference contentLink, ILanguageSelector languageSelector)
{
   var content = DocumentRepository.Service.Get(contentLink);

    //Other types handled here

	if (content is DocumentFile)
	{
		var documentContent = GetDefaultContent(ContentLoader.Service.Get<ApiDocumentCategory>(content.ParentLink),
			new string[] { ".JPG", ".JPEG", ".GIF", ".PNG" }.Contains(Path.GetExtension((content as DocumentFile).Href), StringComparer.OrdinalIgnoreCase) ? ContentTypeRepository.Load(typeof(ApiDocumentImage)).ID : ContentTypeRepository.Load(typeof(ApiDocument)).ID,
										languageSelector);

	documentContent.ContentGuid = content.ContentGuid;
	documentContent.ContentLink = contentLink;
	documentContent.Name = content.Name;
	(documentContent as MediaData).Changed = DateTime.Now;
    (documentContent as MediaData).StartPublish = DateTime.Now.AddMinutes(-10);
	(documentContent as MediaData).RouteSegment = content.Name;
	(documentContent as MediaData).Status = VersionStatus.Published;
	(documentContent as IApiMetadata).Href = (content as DocumentFile).Href;
	(documentContent as ISupportsDataDate).DataDate = (content as DocumentFile).DataDate;
	(documentContent as IApiMetadata).EntityId = (content as DocumentFile).EntityId;
	(documentContent as IVersionable).IsPendingPublish = false;

	try
	{
		var provider = ServiceLocator.Current.GetInstance<IBlobProviderRegistry>().GetProvider(new Uri("epi.fx.blob://apidoc/"));
		var extension = Path.GetExtension((documentContent as IApiMetadata).Href);
		var blob = provider.CreateBlob((documentContent as MediaData).BinaryDataContainer, string.IsNullOrEmpty(extension) ? ".null" : extension) as ApiDocumentBlob;
		blob.Href = (documentContent as IApiMetadata).Href;
		(documentContent as MediaData).BinaryData = blob;
	}
	catch (Exception ex)
	{
		IApiMetadata apiMeta = (documentContent as IApiMetadata);
		var href = apiMeta == null ? "" : apiMeta.Href;
		Logger.Critical(href + ": " + ex.Message, ex);
		throw;
	}

	return documentContent;
}

Looks like there is a time mismatch internally with the ITimeProvider?

#198711
Nov 04, 2018 20:35
Coenraad Myburgh
Coenraad Myburgh
Vote:
 

Found the issue, seems that one has to set the StartPublish property slightly in the past:

protected override IContent LoadContent(ContentReference contentLink, ILanguageSelector languageSelector)
{
   var content = DocumentRepository.Service.Get(contentLink);

    //Other types handled here

	if (content is DocumentFile)
	{
		var documentContent = GetDefaultContent(ContentLoader.Service.Get<ApiDocumentCategory>(content.ParentLink),
			new string[] { ".JPG", ".JPEG", ".GIF", ".PNG" }.Contains(Path.GetExtension((content as DocumentFile).Href), StringComparer.OrdinalIgnoreCase) ? ContentTypeRepository.Load(typeof(ApiDocumentImage)).ID : ContentTypeRepository.Load(typeof(ApiDocument)).ID,
										languageSelector);

	documentContent.ContentGuid = content.ContentGuid;
	documentContent.ContentLink = contentLink;
	documentContent.Name = content.Name;
	(documentContent as MediaData).Changed = DateTime.Now;
    //Set the StartPublish slightly in the past
    (documentContent as MediaData).StartPublish = DateTime.Now.AddMinutes(-10);
	(documentContent as MediaData).RouteSegment = content.Name;
	(documentContent as MediaData).Status = VersionStatus.Published;
	(documentContent as IApiMetadata).Href = (content as DocumentFile).Href;
	(documentContent as ISupportsDataDate).DataDate = (content as DocumentFile).DataDate;
	(documentContent as IApiMetadata).EntityId = (content as DocumentFile).EntityId;
	(documentContent as IVersionable).IsPendingPublish = false;

	try
	{
		var provider = ServiceLocator.Current.GetInstance<IBlobProviderRegistry>().GetProvider(new Uri("epi.fx.blob://apidoc/"));
		var extension = Path.GetExtension((documentContent as IApiMetadata).Href);
		var blob = provider.CreateBlob((documentContent as MediaData).BinaryDataContainer, string.IsNullOrEmpty(extension) ? ".null" : extension) as ApiDocumentBlob;
		blob.Href = (documentContent as IApiMetadata).Href;
		(documentContent as MediaData).BinaryData = blob;
	}
	catch (Exception ex)
	{
		IApiMetadata apiMeta = (documentContent as IApiMetadata);
		var href = apiMeta == null ? "" : apiMeta.Href;
		Logger.Critical(href + ": " + ex.Message, ex);
		throw;
	}

	return documentContent;
}
#198716
Edited, Nov 05, 2018 8:12
This topic was created over six months ago and has been resolved. If you have a similar question, please create a new topic and refer to this one.
* You are NOT allowed to include any hyperlinks in the post because your account hasn't associated to your company. User profile should be updated.