[HBASE-22928] ScanMetrics counter update may not happen in case of exception in TableRecordReaderImpl

• Type: Bug
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 3.0.0-alpha-1, 2.3.0, 2.2.1, 2.1.7
• Component/s: mapreduce
• Hadoop Flags: Reviewed
• Assignee: Pankaj Kumar
• Reporter: Pankaj Kumar
• Created: 27/Aug/19 08:08
• Updated: 02/Sep/19 17:36
• Resolved: 02/Sep/19 02:51

Description

ScanMetrics counter update may not happen in case of exception in TableRecordReaderImpl.nextKeyValue(),

There are multiple cases as below,

1. In case of DoNotRetryIOException

 // do not retry if the exception tells us not to do so
 if (e instanceof DoNotRetryIOException) {
 throw e;
 }

2. While creating a new scanner,

 if (lastSuccessfulRow == null) {
 restart(scan.getStartRow());
 } else {
 restart(lastSuccessfulRow);
 scanner.next(); // skip presumed already mapped row
 }

3. In case of IOException

 catch (IOException ioe) {
 if (logScannerActivity) {
 long now = System.currentTimeMillis();
 LOG.info("Mapper took " + (now-timestamp)
 + "ms to process " + rowcount + " rows");
 LOG.info(ioe.toString(), ioe);
 String lastRow = lastSuccessfulRow == null ?
 "null" : Bytes.toStringBinary(lastSuccessfulRow);
 LOG.info("lastSuccessfulRow=" + lastRow);
 }
 throw ioe;
 }

Issue Links

• links to: GitHub Pull Request #546