Stream archives to Azure Blob Storage in bbs2gh migrate-repo, rathe…

…r than loading the whole archive into memory (#978)

When running a `bbs2gh migrate-repo` migration with Azure Blob Storage
where the archive is larger than 2GB, currently it fails with a
`System.IO.IOException` stating:

> The file is too long. This operation is currently limited to
supporting files less than 2 gigabytes in size.

This switches to streaming the file, rather than reading it all into
memory, mirroring the approach (and using the code!) from `gh gei`.

Fixes #974.

<!--
For the checkboxes below you must check each one to indicate that you
either did the relevant task, or considered it and decided there was
nothing that needed doing
-->

- [x] Did you write/update appropriate tests
- [x] Release notes updated (if appropriate)
- [x] Appropriate logging output
- [x] Issue linked
- [ ] Docs updated (or issue created)
- [ ] New package licenses are added to `ThirdPartyNotices.txt` (if
applicable)

<!--
For docs we should review the docs at:
https://docs.github.com/en/migrations/using-github-enterprise-importer
and the README.md in this repo

If a doc update is required based on the changes in this PR, it is
sufficient to create an issue and link to it here. The doc update can be
made later/separately.
-->

---------

Co-authored-by: Dylan Smith <dylanfromwinnipeg@gmail.com>

RELEASENOTES.md

@@ -1,4 +1,5 @@
 - Include the migration ID in the default output filename when running `download-logs`
 - Include the date with the timestamp when writing to the log
 - When blob storage credentials are provided to the CLI but will not be used for a GHES migration, log a clear warning, not an info message
-- Unhide the `--archive-download-host` argument in the documentation for `gh bbs2gh migrate-repo` and `gh bbs2gh generate-script`
+- Fix support for Azure Blob Storage in `bbs2gh` migrations when the archive is larger than 2GB
+- Unhide the `--archive-download-host` argument in the documentation for `gh bbs2gh migrate-repo` and `gh bbs2gh generate-script`

src/OctoshiftCLI.Tests/bbs2gh/Handlers/MigrateRepoCommandHandlerTests.cs

@@ -1,4 +1,5 @@
 using System;
+using System.IO;
 using System.Text;
 using System.Threading.Tasks;
 using FluentAssertions;
@@ -156,7 +157,7 @@ public async Task Happy_Path_Generate_Archive_Ssh_Download_Azure_Upload_And_Inge
             _mockBbsApi.Setup(x => x.GetExport(BBS_EXPORT_ID)).ReturnsAsync(("COMPLETED", "The export is complete", 100));
             _mockBbsArchiveDownloader.Setup(x => x.Download(BBS_EXPORT_ID, It.IsAny<string>())).ReturnsAsync(ARCHIVE_PATH);
             _mockFileSystemProvider.Setup(x => x.ReadAllBytesAsync(ARCHIVE_PATH)).ReturnsAsync(ARCHIVE_DATA);
-            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), ARCHIVE_DATA)).ReturnsAsync(new Uri(ARCHIVE_URL));
+            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), It.IsAny<FileStream>())).ReturnsAsync(new Uri(ARCHIVE_URL));
             _mockGithubApi.Setup(x => x.GetOrganizationId(GITHUB_ORG).Result).Returns(GITHUB_ORG_ID);
             _mockGithubApi.Setup(x => x.CreateBbsMigrationSource(GITHUB_ORG_ID).Result).Returns(MIGRATION_SOURCE_ID);
 
@@ -240,8 +241,7 @@ public async Task Happy_Path_Full_Flow_Running_On_Bbs_Server()
 
             _mockBbsApi.Setup(x => x.StartExport(BBS_PROJECT, BBS_REPO)).ReturnsAsync(BBS_EXPORT_ID);
             _mockBbsApi.Setup(x => x.GetExport(BBS_EXPORT_ID)).ReturnsAsync(("COMPLETED", "The export is complete", 100));
-            _mockFileSystemProvider.Setup(x => x.ReadAllBytesAsync(archivePath)).ReturnsAsync(ARCHIVE_DATA);
-            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), ARCHIVE_DATA)).ReturnsAsync(new Uri(ARCHIVE_URL));
+            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), It.IsAny<FileStream>())).ReturnsAsync(new Uri(ARCHIVE_URL));
             _mockGithubApi.Setup(x => x.GetOrganizationId(GITHUB_ORG).Result).Returns(GITHUB_ORG_ID);
             _mockGithubApi.Setup(x => x.CreateBbsMigrationSource(GITHUB_ORG_ID).Result).Returns(MIGRATION_SOURCE_ID);
 
@@ -296,7 +296,7 @@ public async Task Happy_Path_Full_Flow_Bbs_Credentials_Via_Environment()
             _mockBbsApi.Setup(x => x.GetExport(BBS_EXPORT_ID)).ReturnsAsync(("COMPLETED", "The export is complete", 100));
             _mockBbsArchiveDownloader.Setup(x => x.Download(BBS_EXPORT_ID, It.IsAny<string>())).ReturnsAsync(ARCHIVE_PATH);
             _mockFileSystemProvider.Setup(x => x.ReadAllBytesAsync(ARCHIVE_PATH)).ReturnsAsync(ARCHIVE_DATA);
-            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), ARCHIVE_DATA)).ReturnsAsync(new Uri(ARCHIVE_URL));
+            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), It.IsAny<FileStream>())).ReturnsAsync(new Uri(ARCHIVE_URL));
             _mockGithubApi.Setup(x => x.GetOrganizationId(GITHUB_ORG).Result).Returns(GITHUB_ORG_ID);
             _mockGithubApi.Setup(x => x.CreateBbsMigrationSource(GITHUB_ORG_ID).Result).Returns(MIGRATION_SOURCE_ID);
 
@@ -335,7 +335,7 @@ public async Task Happy_Path_Deletes_Downloaded_Archive()
             _mockBbsApi.Setup(x => x.GetExport(BBS_EXPORT_ID)).ReturnsAsync(("COMPLETED", "The export is complete", 100));
             _mockBbsArchiveDownloader.Setup(x => x.Download(BBS_EXPORT_ID, It.IsAny<string>())).ReturnsAsync(ARCHIVE_PATH);
             _mockFileSystemProvider.Setup(x => x.ReadAllBytesAsync(ARCHIVE_PATH)).ReturnsAsync(ARCHIVE_DATA);
-            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), ARCHIVE_DATA)).ReturnsAsync(new Uri(ARCHIVE_URL));
+            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), It.IsAny<FileStream>())).ReturnsAsync(new Uri(ARCHIVE_URL));
 
             // Act
             var args = new MigrateRepoCommandArgs
@@ -367,7 +367,7 @@ public async Task It_Deletes_Downloaded_Archive_Even_If_Upload_Fails()
             _mockBbsApi.Setup(x => x.GetExport(BBS_EXPORT_ID)).ReturnsAsync(("COMPLETED", "The export is complete", 100));
             _mockBbsArchiveDownloader.Setup(x => x.Download(BBS_EXPORT_ID, It.IsAny<string>())).ReturnsAsync(ARCHIVE_PATH);
             _mockFileSystemProvider.Setup(x => x.ReadAllBytesAsync(ARCHIVE_PATH)).ReturnsAsync(ARCHIVE_DATA);
-            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), ARCHIVE_DATA)).ThrowsAsync(new InvalidOperationException());
+            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), It.IsAny<FileStream>())).ThrowsAsync(new InvalidOperationException());
 
             // Act
             var args = new MigrateRepoCommandArgs
@@ -398,7 +398,7 @@ public async Task Happy_Path_Does_Not_Throw_If_Fails_To_Delete_Downloaded_Archiv
             _mockBbsApi.Setup(x => x.GetExport(BBS_EXPORT_ID)).ReturnsAsync(("COMPLETED", "The export is complete", 100));
             _mockBbsArchiveDownloader.Setup(x => x.Download(BBS_EXPORT_ID, It.IsAny<string>())).ReturnsAsync(ARCHIVE_PATH);
             _mockFileSystemProvider.Setup(x => x.ReadAllBytesAsync(ARCHIVE_PATH)).ReturnsAsync(ARCHIVE_DATA);
-            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), ARCHIVE_DATA)).ReturnsAsync(new Uri(ARCHIVE_URL));
+            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), It.IsAny<FileStream>())).ReturnsAsync(new Uri(ARCHIVE_URL));
             _mockFileSystemProvider.Setup(x => x.DeleteIfExists(It.IsAny<string>())).Throws(new UnauthorizedAccessException("Access Denied"));
 
             // Act
@@ -616,7 +616,7 @@ public async Task Uses_Archive_Path_If_Provided()
             var archiveBytes = Encoding.ASCII.GetBytes("here are some bytes");
             _mockFileSystemProvider.Setup(x => x.ReadAllBytesAsync(ARCHIVE_PATH)).ReturnsAsync(archiveBytes);
 
-            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), archiveBytes)).ReturnsAsync(new Uri(ARCHIVE_URL));
+            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), It.IsAny<FileStream>())).ReturnsAsync(new Uri(ARCHIVE_URL));
 
             _mockGithubApi.Setup(x => x.GetOrganizationId(GITHUB_ORG).Result).Returns(GITHUB_ORG_ID);
             _mockGithubApi.Setup(x => x.CreateBbsMigrationSource(GITHUB_ORG_ID).Result).Returns(MIGRATION_SOURCE_ID);
@@ -763,7 +763,7 @@ await _handler.Invoking(x => x.Handle(args))
         public async Task It_Does_Not_Set_The_Archive_Path_When_Archive_Path_Is_Provided()
         {
             // Arrange
-            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), It.IsAny<byte[]>())).ReturnsAsync(new Uri(ARCHIVE_URL));
+            _mockAzureApi.Setup(x => x.UploadToBlob(It.IsAny<string>(), It.IsAny<FileStream>())).ReturnsAsync(new Uri(ARCHIVE_URL));
 
             // Act
             var args = new MigrateRepoCommandArgs
@@ -778,7 +778,7 @@ public async Task It_Does_Not_Set_The_Archive_Path_When_Archive_Path_Is_Provided
 
             // Assert
             args.ArchivePath.Should().Be(ARCHIVE_PATH);
-            _mockFileSystemProvider.Verify(m => m.ReadAllBytesAsync(ARCHIVE_PATH));
+            _mockFileSystemProvider.Verify(m => m.OpenRead(ARCHIVE_PATH));
         }
 
         [Fact]

src/bbs2gh/Handlers/MigrateRepoCommandHandler.cs

@@ -188,11 +188,14 @@ private async Task<string> UploadArchiveToAzure(string archivePath)
     {
         _log.LogInformation("Uploading Archive to Azure...");
 
-        var archiveData = await _fileSystemProvider.ReadAllBytesAsync(archivePath);
-        var archiveName = GenerateArchiveName();
-        var archiveBlobUrl = await _azureApi.UploadToBlob(archiveName, archiveData);
-
-        return archiveBlobUrl.ToString();
+#pragma warning disable IDE0063
+        await using (var archiveData = _fileSystemProvider.OpenRead(archivePath))
+#pragma warning restore IDE0063
+        {
+            var archiveName = GenerateArchiveName();
+            var archiveBlobUrl = await _azureApi.UploadToBlob(archiveName, archiveData);
+            return archiveBlobUrl.ToString();
+        }
     }
 
     private string GenerateArchiveName() => $"{Guid.NewGuid()}.tar";