forked from docs/doc-exports
Yang, Tong
6182f91ba8
Reviewed-by: Hasko, Vladimir <vladimir.hasko@t-systems.com> Co-authored-by: Yang, Tong <yangtong2@huawei.com> Co-committed-by: Yang, Tong <yangtong2@huawei.com>
15 lines
1.6 KiB
HTML
15 lines
1.6 KiB
HTML
<a name="mrs_01_1640"></a><a name="mrs_01_1640"></a>
|
|
|
|
<h1 class="topictitle1">Operation Failures Occur in Stopping BulkLoad On the Client</h1>
|
|
<div id="body1596003893196"><div class="section" id="mrs_01_1640__sedb9130033b948eebe961984204b43df"><h4 class="sectiontitle">Question</h4><p id="mrs_01_1640__add4ff60959b143ed8dca0968a6991f26">Why submitted operations fail by stopping BulkLoad on the client during BulkLoad data importing?</p>
|
|
</div>
|
|
<div class="section" id="mrs_01_1640__s61ac3b4524f241fcb6a47ca4e153f2a5"><h4 class="sectiontitle">Answer</h4><p id="mrs_01_1640__a526634d68881455594c3c16a8040a072">When BulkLoad is enabled on the client, a partitioner file is generated and used to demarcate the range of Map task data inputting. The file is automatically deleted when BulkLoad exists on the client. In general, if all map tasks are enabled and running, the termination of BulkLoad on the client does not cause the failure of submitted operations. However, due to the retry and speculative execution mechanism of Map tasks, a Map task is performed again if failures of the Reduce task to download the data of the completed Map task exceed the limit. In this case, if BulkLoad already exists on the client, the retry Map task fails and the operation failure occurs because the partitioner file is missing. Therefore, it is recommended not to stop BulkLoad on the client during BulkLoad data importing.</p>
|
|
</div>
|
|
</div>
|
|
<div>
|
|
<div class="familylinks">
|
|
<div class="parentlink"><strong>Parent topic:</strong> <a href="mrs_01_1638.html">Common Issues About HBase</a></div>
|
|
</div>
|
|
</div>
|
|
|