[jira] [Commented] (HADOOP-12403) Enable multiple writes in flight for HBase WAL writing backed by WASB

["stack (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HADOOP-12403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14741083#comment-14741083 ] 

stack commented on HADOOP-12403:
--------------------------------

bq. The latest HBase WAL write model (HBASE-8755) uses multiple AsyncSyncer threads to sync data to HDFS.

It would be preferable if we did not have to do this against HDFS Client. A single thread doing syncs back-to-back would be ideal but experiment had it that 5 threads each running a sync seems to be optimal (throughput-wise) for setting up a syncing pipeline. Need to dig in as to why 5 and why this is needed at all. Just FYI.

> Enable multiple writes in flight for HBase WAL writing backed by WASB
> ---------------------------------------------------------------------
>
>                 Key: HADOOP-12403
>                 URL: https://issues.apache.org/jira/browse/HADOOP-12403
>             Project: Hadoop Common
>          Issue Type: Improvement
>          Components: azure
>            Reporter: Duo Xu
>            Assignee: Duo Xu
>         Attachments: HADOOP-12403.01.patch, HADOOP-12403.02.patch, HADOOP-12403.03.patch
>
>
> Azure HDI HBase clusters use Azure blob storage as file system. We found that the bottle neck was during writing to write ahead log (WAL). The latest HBase WAL write model (HBASE-8755) uses multiple AsyncSyncer threads to sync data to HDFS. However, our WASB driver is still based on a single thread model. Thus when the sync threads call into WASB layer, every time only one thread will be allowed to send data to Azure storage.This jira is to introduce a new write model in WASB layer to allow multiple writes in parallel.
> 1. Since We use page blob for WAL, this will cause "holes" in the page blob as every write starts on a new page. We use the first two bytes of every page to record the actual data size of the current page.
> 2. When reading WAL, we need to know the actual size of the WAL. This should be the sum of the number represented by the first two bytes of every page. However looping over every page to get the size will be very slow, considering normal WAL size is 128MB and each page is 512 bytes. So during writing, every time a write succeeds, a metadata of the blob called "total_data_uploaded" will be updated.
> 3. Although we allow multiple writes in flight, we need to make sure the sync threads which call into WASB layers return in order. Reading HBase source code FSHLog.java, we find that every sync request is associated with a transaction id. If the sync succeeds, all the transactions prior to this transaction id are assumed to be in Azure Storage. We use a queue to store the sync requests and make sure they return to HBase layer in order.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)