big-data/hadoop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

HDFS-11914. Add more diagnosis info for fsimage transfer failure. Contributed by Yongjun Zhang.

Browse Source
This commit is contained in:
Yongjun Zhang 2017-06-05 16:31:03 -07:00
parent 5fd9742c83
commit 6a28a2b914
2 changed files with 23 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/common/Util.java
View File

@ -225,6 +225,7 @@ public static MD5Hash receiveFile(String url, List<File> localPaths,
stream = new DigestInputStream(stream, digester); stream = new DigestInputStream(stream, digester);
} }
boolean finishedReceiving = false; boolean finishedReceiving = false;
int num = 1;
List<FileOutputStream> outputStreams = Lists.newArrayList(); List<FileOutputStream> outputStreams = Lists.newArrayList();
@ -256,7 +257,6 @@ public static MD5Hash receiveFile(String url, List<File> localPaths,
} }
} }
int num = 1;
byte[] buf = new byte[IO_FILE_BUFFER_SIZE]; byte[] buf = new byte[IO_FILE_BUFFER_SIZE];
while (num > 0) { while (num > 0) {
num = stream.read(buf); num = stream.read(buf);
@ -305,8 +305,8 @@ public static MD5Hash receiveFile(String url, List<File> localPaths,
// exception that makes it look like a server-side problem! // exception that makes it look like a server-side problem!
deleteTmpFiles(localPaths); deleteTmpFiles(localPaths);
throw new IOException("File " + url + " received length " + received + throw new IOException("File " + url + " received length " + received +
" is not of the advertised size " + " is not of the advertised size " + advertisedSize +
advertisedSize); ". Fsimage name: " + fsImageName + " lastReceived: " + num);
} }
} }
xferStats.insert(0, String.format("Combined time for file download and" + xferStats.insert(0, String.format("Combined time for file download and" +

21
hadoop-hdfs-project/hadoop-hdfs/src/main/java/org/apache/hadoop/hdfs/server/namenode/TransferFsImage.java
View File

@ -339,6 +339,11 @@ private static void copyFileToStream(OutputStream out, File localfile,
FileInputStream infile, DataTransferThrottler throttler, FileInputStream infile, DataTransferThrottler throttler,
Canceler canceler) throws IOException { Canceler canceler) throws IOException {
byte buf[] = new byte[IO_FILE_BUFFER_SIZE]; byte buf[] = new byte[IO_FILE_BUFFER_SIZE];
long total = 0;
int num = 1;
IOException ioe = null;
String reportStr = "Sending fileName: " + localfile.getAbsolutePath()
+ ", fileSize: " + localfile.length() + ".";
try { try {
CheckpointFaultInjector.getInstance() CheckpointFaultInjector.getInstance()
.aboutToSendFile(localfile); .aboutToSendFile(localfile);
@ -352,7 +357,6 @@ private static void copyFileToStream(OutputStream out, File localfile,
// and the rest of the image will be sent over the wire // and the rest of the image will be sent over the wire
infile.read(buf); infile.read(buf);
} }
int num = 1;
while (num > 0) { while (num > 0) {
if (canceler != null && canceler.isCancelled()) { if (canceler != null && canceler.isCancelled()) {
throw new SaveNamespaceCancelledException( throw new SaveNamespaceCancelledException(
@ -370,14 +374,27 @@ private static void copyFileToStream(OutputStream out, File localfile,
} }
out.write(buf, 0, num); out.write(buf, 0, num);
total += num;
if (throttler != null) { if (throttler != null) {
throttler.throttle(num, canceler); throttler.throttle(num, canceler);
} }
} }
} catch (EofException e) { } catch (EofException e) {
LOG.info("Connection closed by client"); reportStr += " Connection closed by client.";
ioe = e;
out = null; // so we don't close in the finally out = null; // so we don't close in the finally
} catch (IOException ie) {
ioe = ie;
throw ie;
} finally { } finally {
reportStr += " Sent total: " + total +
" bytes. Size of last segment intended to send: " + num
+ " bytes.";
if (ioe != null) {
LOG.info(reportStr, ioe);
} else {
LOG.info(reportStr);
}
if (out != null) { if (out != null) {
out.close(); out.close();
} }