Hi Shawn,
Thanks for reply. If we keep replication Async, Can error handling not work
same like replica down scenario?
-Original Message-
From: Shawn Heisey [mailto:apa...@elyograg.org]
Sent: 09 September 2015 19:40
To: solr-user@lucene.apache.org
Subject: Re: Replication Sync OR Async?
On 9/8/2015 11:16 PM, Maulin Rathod wrote:
> When replicas are running it took around 900 seconds for indexing.
> After stopping replicas it took around 500 seconds for indexing.
>
> Is the replication happens in Sync or Async? If it is Sync, can we
> make it Async so that it will not affect indexing performance.
Running things in that way results in problems like SOLR-3284, which was an
issue long before SolrCloud arrived on the scene. If you throw processing to
the background, it becomes very difficult to handle errors.
https://issues.apache.org/jira/browse/SOLR-3284
If you really don't care about knowing whether your indexing succeeded, then
something like this would be helpful. Most users *do* want to know whether
their indexing succeeded. Feel free to open an enhancement issue in Jira, but
please be aware of the disadvantages of what you're asking for.
Handling errors correctly is not impossible with async operation, but it is
very challenging, prone to bugs, and may end up erasing any speed advantage.
Thanks,
Shawn