[
https://issues.apache.org/jira/browse/SOLR-6209?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14046729#comment-14046729
]
Arcadius Ahouansou commented on SOLR-6209:
------------------------------------------
Sanket, it seems the JDBC driver get stuck in the middle of a read.
Have you set readTimeout on the DB connection?
In our case, Solr DIH used to hang until we added readTimeout on the datasource.
We are connecting to Oracle and our datasoure definition looks like this:
{code}
<dataSource name="jdbc" driver="oracle...." url="..." autoCommit="false"
batchSize="1000" transactionIsolation="TRANSACTION_READ_COMMITTED"
oracle.net.READ_TIMEOUT="300000" oracle.jdbc.ReadTimeout="300000" />
{code}
So, try to find the equivalent of {code}oracle.jdbc.ReadTimeout{code} or
{code}oracle.net.READ_TIMEOUT{code} for MS SQL Server JDBC driver.
> Database connections lost during data import
> --------------------------------------------
>
> Key: SOLR-6209
> URL: https://issues.apache.org/jira/browse/SOLR-6209
> Project: Solr
> Issue Type: Bug
> Environment: OS: Windows 7
> RAM: 8 GB
> Reporter: SANKET
> Labels: solr
> Original Estimate: 30h
> Remaining Estimate: 30h
>
> Follow the steps to generate the error:
> 1. Configure the large amount of data (around 4 GB or more than 50 millions
> of records)
> 2. Give proper data-config.xml file for indexing the data from remote
> database server.
> 3. During indexing the data into solr from SQL SERVER 2010, at the half way
> unplug the network cable and see the status in solr.
> e.g.
> localhost:8083/solr/core1/dataimport?command=status
> or
> localhost:8083/solr/core1/dataimport
> 4. Pass few seconds then again plug back the cable.
> 5. You can clearly see that there is just only "Time Elapsed" parameter
> increase. "Total Rows Fetched" & "Total Documents Processed" remains same for
> infinite time.
> 6. You can regenerate this for small data also.
> 7. Work around is you need to restart the solr. (But this is not good
> solution)
> Note:
> This is very important issue because, so many organizations not using this
> valuable products just because of the this database infinite connection issue.
> Solution can be: Forcefully abort the data indexing or provide mechanism for
> forcefully abort the indexing.
> Hope you guys knows that abort command is also not working.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
