| | | 1 | | // Copyright (c) Microsoft and contributors. All rights reserved. |
| | | 2 | | // |
| | | 3 | | // Licensed under the Apache License, Version 2.0 (the "License"); |
| | | 4 | | // you may not use this file except in compliance with the License. |
| | | 5 | | // You may obtain a copy of the License at |
| | | 6 | | // http://www.apache.org/licenses/LICENSE-2.0 |
| | | 7 | | // |
| | | 8 | | // Unless required by applicable law or agreed to in writing, software |
| | | 9 | | // distributed under the License is distributed on an "AS IS" BASIS, |
| | | 10 | | // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| | | 11 | | // |
| | | 12 | | // See the License for the specific language governing permissions and |
| | | 13 | | // limitations under the License. |
| | | 14 | | |
| | | 15 | | using Microsoft.Azure.Batch.Conventions.Files.Utilities; |
| | | 16 | | using Microsoft.WindowsAzure.Storage.Blob; |
| | | 17 | | using System; |
| | | 18 | | using System.Collections.Generic; |
| | | 19 | | using System.Diagnostics; |
| | | 20 | | using System.IO; |
| | | 21 | | using System.Linq; |
| | | 22 | | using System.Text; |
| | | 23 | | using System.Threading; |
| | | 24 | | using System.Threading.Tasks; |
| | | 25 | | |
| | | 26 | | namespace Microsoft.Azure.Batch.Conventions.Files |
| | | 27 | | { |
| | | 28 | | // Represents a storage path within Azure Storage - that is, either job storage |
| | | 29 | | // ({container}) or task storage ({container}/{taskId}). Private derived classes |
| | | 30 | | // represent the two kinds of path. |
| | | 31 | | internal abstract class StoragePath |
| | | 32 | | { |
| | | 33 | | private readonly CloudBlobContainer _jobOutputContainer; |
| | | 34 | | |
| | 35 | 35 | | protected StoragePath(CloudBlobContainer jobOutputContainer) |
| | | 36 | | { |
| | 35 | 37 | | if (jobOutputContainer == null) |
| | | 38 | | { |
| | 0 | 39 | | throw new ArgumentNullException(nameof(jobOutputContainer)); |
| | | 40 | | } |
| | | 41 | | |
| | 35 | 42 | | _jobOutputContainer = jobOutputContainer; |
| | 35 | 43 | | } |
| | | 44 | | |
| | | 45 | | // Uploads a file to blob storage. |
| | | 46 | | public async Task SaveAsync( |
| | | 47 | | IOutputKind kind, |
| | | 48 | | DirectoryInfo baseFolder, |
| | | 49 | | string relativePath, |
| | | 50 | | CancellationToken cancellationToken = default(CancellationToken) |
| | | 51 | | ) |
| | | 52 | | { |
| | | 53 | | Debug.Assert(baseFolder != null); |
| | | 54 | | |
| | 6 | 55 | | if (kind == null) |
| | | 56 | | { |
| | 2 | 57 | | throw new ArgumentNullException(nameof(kind)); |
| | | 58 | | } |
| | | 59 | | |
| | 4 | 60 | | Validate.IsNotNullOrEmpty(relativePath, nameof(relativePath)); |
| | | 61 | | |
| | 0 | 62 | | if (Path.IsPathRooted(relativePath)) |
| | | 63 | | { |
| | 0 | 64 | | throw new ArgumentException($"{nameof(relativePath)} must be a relative path", nameof(relativePath)); |
| | | 65 | | } |
| | | 66 | | |
| | 0 | 67 | | string sourcePath = Path.Combine(baseFolder.FullName, relativePath); |
| | 0 | 68 | | string destinationPath = GetDestinationBlobPath(relativePath); |
| | | 69 | | |
| | 0 | 70 | | await SaveAsync(kind, sourcePath, destinationPath, cancellationToken).ConfigureAwait(false); |
| | 0 | 71 | | } |
| | | 72 | | |
| | | 73 | | // Uploads a file to blob storage. |
| | | 74 | | public async Task SaveAsync( |
| | | 75 | | IOutputKind kind, |
| | | 76 | | string sourcePath, |
| | | 77 | | string destinationRelativePath, |
| | | 78 | | CancellationToken cancellationToken = default(CancellationToken) |
| | | 79 | | ) |
| | | 80 | | { |
| | 10 | 81 | | if (kind == null) |
| | | 82 | | { |
| | 2 | 83 | | throw new ArgumentNullException(nameof(kind)); |
| | | 84 | | } |
| | | 85 | | |
| | 8 | 86 | | Validate.IsNotNullOrEmpty(sourcePath, nameof(sourcePath)); |
| | 4 | 87 | | Validate.IsNotNullOrEmpty(destinationRelativePath, nameof(destinationRelativePath)); |
| | | 88 | | |
| | 0 | 89 | | var blobName = BlobName(kind, destinationRelativePath); |
| | 0 | 90 | | var blob = _jobOutputContainer.GetBlockBlobReference(blobName); |
| | 0 | 91 | | await blob.UploadFromFileAsync(sourcePath, null, null, null, cancellationToken).ConfigureAwait(false); |
| | 0 | 92 | | } |
| | | 93 | | |
| | | 94 | | // Uploads text to blob storage. |
| | | 95 | | public async Task SaveTextAsync( |
| | | 96 | | IOutputKind kind, |
| | | 97 | | string text, |
| | | 98 | | string destinationRelativePath, |
| | | 99 | | CancellationToken cancellationToken = default(CancellationToken) |
| | | 100 | | ) |
| | | 101 | | { |
| | 0 | 102 | | if (kind == null) |
| | | 103 | | { |
| | 0 | 104 | | throw new ArgumentNullException(nameof(kind)); |
| | | 105 | | } |
| | 0 | 106 | | if (text == null) |
| | | 107 | | { |
| | 0 | 108 | | throw new ArgumentNullException(nameof(text)); |
| | | 109 | | } |
| | | 110 | | |
| | 0 | 111 | | Validate.IsNotNullOrEmpty(destinationRelativePath, nameof(destinationRelativePath)); |
| | | 112 | | |
| | 0 | 113 | | var blobName = BlobName(kind, destinationRelativePath); |
| | 0 | 114 | | var blob = _jobOutputContainer.GetBlockBlobReference(blobName); |
| | 0 | 115 | | await blob.UploadTextAsync(text, null, null, null, null, cancellationToken).ConfigureAwait(false); |
| | 0 | 116 | | } |
| | | 117 | | |
| | | 118 | | // Uploads a file and tracks appends to that file. The implementation creates an append blob to |
| | | 119 | | // contain the file contents, then creates a file tracking object which runs a background operation |
| | | 120 | | // to upload the file to the append blob, then track appends to the file and append them to the blob. |
| | | 121 | | public async Task<TrackedFile> SaveTrackedAsync(IOutputKind kind, string relativePath, TimeSpan flushInterval) |
| | | 122 | | { |
| | 2 | 123 | | if (kind == null) |
| | | 124 | | { |
| | 0 | 125 | | throw new ArgumentNullException(nameof(kind)); |
| | | 126 | | } |
| | | 127 | | |
| | 2 | 128 | | Validate.IsNotNullOrEmpty(relativePath, nameof(relativePath)); |
| | | 129 | | |
| | 0 | 130 | | if (Path.IsPathRooted(relativePath)) |
| | | 131 | | { |
| | 0 | 132 | | throw new ArgumentException($"{nameof(relativePath)} must be a relative path", nameof(relativePath)); |
| | | 133 | | } |
| | | 134 | | |
| | 0 | 135 | | var destinationPath = GetDestinationBlobPath(relativePath); |
| | 0 | 136 | | return await SaveTrackedAsync(kind, relativePath, destinationPath, flushInterval).ConfigureAwait(false); |
| | 0 | 137 | | } |
| | | 138 | | |
| | | 139 | | // Uploads a file and tracks appends to that file. The implementation creates an append blob to |
| | | 140 | | // contain the file contents, then creates a file tracking object which runs a background operation |
| | | 141 | | // to upload the file to the append blob, then track appends to the file and append them to the blob. |
| | | 142 | | public async Task<TrackedFile> SaveTrackedAsync(IOutputKind kind, string sourcePath, string destinationRelativeP |
| | | 143 | | { |
| | 5 | 144 | | if (kind == null) |
| | | 145 | | { |
| | 1 | 146 | | throw new ArgumentNullException(nameof(kind)); |
| | | 147 | | } |
| | | 148 | | |
| | 4 | 149 | | Validate.IsNotNullOrEmpty(sourcePath, nameof(sourcePath)); |
| | 2 | 150 | | Validate.IsNotNullOrEmpty(destinationRelativePath, nameof(destinationRelativePath)); |
| | | 151 | | |
| | 0 | 152 | | var blobName = BlobName(kind, destinationRelativePath); |
| | 0 | 153 | | var blob = _jobOutputContainer.GetAppendBlobReference(blobName); |
| | 0 | 154 | | await blob.EnsureExistsAsync().ConfigureAwait(false); |
| | 0 | 155 | | return new TrackedFile(sourcePath, blob, flushInterval); |
| | 0 | 156 | | } |
| | | 157 | | |
| | | 158 | | public IEnumerable<OutputFileReference> List(IOutputKind kind) |
| | | 159 | | { |
| | 2 | 160 | | if (kind == null) |
| | | 161 | | { |
| | 2 | 162 | | throw new ArgumentNullException(nameof(kind)); |
| | | 163 | | } |
| | | 164 | | |
| | 0 | 165 | | return _jobOutputContainer.ListBlobs(BlobNamePrefix(kind), useFlatBlobListing: true) |
| | 0 | 166 | | .OfType<ICloudBlob>() |
| | 0 | 167 | | .Select(b => new OutputFileReference(b)); |
| | | 168 | | } |
| | | 169 | | |
| | | 170 | | public async Task<OutputFileReference> GetOutputAsync(IOutputKind kind, string filePath, CancellationToken cance |
| | | 171 | | { |
| | 6 | 172 | | if (kind == null) |
| | | 173 | | { |
| | 2 | 174 | | throw new ArgumentNullException(nameof(kind)); |
| | | 175 | | } |
| | | 176 | | |
| | 4 | 177 | | Validate.IsNotNullOrEmpty(filePath, nameof(filePath)); |
| | | 178 | | |
| | 0 | 179 | | var blob = await _jobOutputContainer.GetBlobReferenceFromServerAsync(BlobName(kind, filePath), null, null, n |
| | | 180 | | |
| | 0 | 181 | | return new OutputFileReference(blob); |
| | 0 | 182 | | } |
| | | 183 | | |
| | | 184 | | // Gets the string that should be prefixed to a blob name to locate it correctly |
| | | 185 | | // in the container. This is the equivalent of a file system directory path - for |
| | | 186 | | // example a prefix might be "$JobOutput/" or "MyTask/$TaskOutput" - but Azure |
| | | 187 | | // Storage does not have the notion of directories under a container, only of prefixes. |
| | | 188 | | internal abstract string BlobNamePrefix(IOutputKind kind); |
| | | 189 | | |
| | | 190 | | internal string BlobName(IOutputKind kind, string relativePath) |
| | 0 | 191 | | => $"{BlobNamePrefix(kind)}/{relativePath}"; |
| | | 192 | | |
| | | 193 | | private static string GetDestinationBlobPath(string relativeSourcePath) |
| | | 194 | | { |
| | | 195 | | const string up = "../"; |
| | | 196 | | |
| | 0 | 197 | | var destinationPath = relativeSourcePath.Replace('\\', '/'); |
| | | 198 | | |
| | | 199 | | // If we are given a path that traverses up from the working directory, |
| | | 200 | | // treat it as though it were rooted at the working directory for blob naming |
| | | 201 | | // purposes. This is intended to support files such as ..\stdout.txt, which |
| | | 202 | | // is stored above the task working directory. |
| | | 203 | | // |
| | | 204 | | // A user can intentionally try to defeat this simple flattening by using a path |
| | | 205 | | // such as "temp\..\..\stdout.txt" - this may result in the file being |
| | | 206 | | // stored in the 'wrong' part of the job container, but they can't write |
| | | 207 | | // outside the job container this way, so the only damage they can do is |
| | | 208 | | // to themselves. |
| | 0 | 209 | | while (destinationPath.StartsWith(up)) |
| | | 210 | | { |
| | 0 | 211 | | destinationPath = relativeSourcePath.Substring(up.Length); |
| | | 212 | | } |
| | | 213 | | |
| | 0 | 214 | | return destinationPath; |
| | | 215 | | } |
| | | 216 | | |
| | | 217 | | internal sealed class JobStoragePath : StoragePath |
| | | 218 | | { |
| | | 219 | | internal JobStoragePath(CloudBlobContainer jobOutputContainer) |
| | 14 | 220 | | : base(jobOutputContainer) |
| | | 221 | | { |
| | 14 | 222 | | } |
| | | 223 | | |
| | 3 | 224 | | internal override string BlobNamePrefix(IOutputKind kind) => BlobNamePrefixImpl(kind); |
| | | 225 | | |
| | 6 | 226 | | internal static string BlobNamePrefixImpl(IOutputKind kind) => $"${kind.Text}"; |
| | | 227 | | } |
| | | 228 | | |
| | | 229 | | internal sealed class TaskStoragePath : StoragePath |
| | | 230 | | { |
| | | 231 | | private readonly string _taskId; |
| | | 232 | | |
| | | 233 | | internal TaskStoragePath(CloudBlobContainer jobOutputContainer, string taskId) |
| | 21 | 234 | | : base(jobOutputContainer) |
| | | 235 | | { |
| | | 236 | | Debug.Assert(taskId != null); |
| | 21 | 237 | | _taskId = taskId; |
| | 21 | 238 | | } |
| | | 239 | | |
| | 4 | 240 | | internal override string BlobNamePrefix(IOutputKind kind) => BlobNamePrefixImpl(kind, _taskId); |
| | | 241 | | |
| | 8 | 242 | | internal static string BlobNamePrefixImpl(IOutputKind kind, string taskId) => $"{taskId}/${kind.Text}"; |
| | | 243 | | } |
| | | 244 | | } |
| | | 245 | | } |