JDBC Thin Client does not return

classic Classic list List threaded Threaded
5 messages Options
pg31 pg31
Reply | Threaded
Open this post in threaded view
|

JDBC Thin Client does not return

Hi

I am running a long query using JDBCThin Client.
It is a CREATE INDEX query and it takes upto 20 minutes.

After, that I see using DBeaver that the Index is created.
I login into my Ignite Cluster and see that all the operations have been
completed (i.e. there are NO WAL updates happening, Checkpoints are all done
(since I see the message skipping checkpoints as there are no dirty pages))

But the above Query does not exit i.e. it keeps on running for hours. Seems
like the above query is kind of stuck.

    val conn = DriverManager.getConnection(jdbcURL)
    val stmt = conn.createStatement()
    Logger.LogInfo(s"Executing:${sqlQuery}")
    stmt.execute(sqlQuery)


Running the same query on smaller tables succeeds.(i.e. it returns)
The table on which I am running this query has around 60Million records
(This is the query:
CREATE INDEX IDX_V1 ON MY_TABLE (MY_COL ASC) PARALLEL 16)

I have tried it both with and without parallel.
Please let me know if anything else is required.

Thank You



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
dmagda dmagda
Reply | Threaded
Open this post in threaded view
|

Re: JDBC Thin Client does not return

Hello, 

Could you please collect thread dumps from the servers and send them together with the logs while the query is hanging?

-
Denis


On Tue, Feb 11, 2020 at 5:22 AM pg31 <[hidden email]> wrote:
Hi

I am running a long query using JDBCThin Client.
It is a CREATE INDEX query and it takes upto 20 minutes.

After, that I see using DBeaver that the Index is created.
I login into my Ignite Cluster and see that all the operations have been
completed (i.e. there are NO WAL updates happening, Checkpoints are all done
(since I see the message skipping checkpoints as there are no dirty pages))

But the above Query does not exit i.e. it keeps on running for hours. Seems
like the above query is kind of stuck.

    val conn = DriverManager.getConnection(jdbcURL)
    val stmt = conn.createStatement()
    Logger.LogInfo(s"Executing:${sqlQuery}")
    stmt.execute(sqlQuery)


Running the same query on smaller tables succeeds.(i.e. it returns)
The table on which I am running this query has around 60Million records
(This is the query:
CREATE INDEX IDX_V1 ON MY_TABLE (MY_COL ASC) PARALLEL 16)

I have tried it both with and without parallel.
Please let me know if anything else is required.

Thank You



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
pg31 pg31
Reply | Threaded
Open this post in threaded view
|

Re: JDBC Thin Client does not return

Hi

Thank You for the reply Denis.

Please find the logs attached. The query was still running
Create_IDX_Stuck.zip
<http://apache-ignite-users.70518.x6.nabble.com/file/t2770/Create_IDX_Stuck.zip>  



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
pg31 pg31
Reply | Threaded
Open this post in threaded view
|

Re: JDBC Thin Client does not return

Hello

I know all folks are busy with 2.8 release.
It would be great, if someone can spare a little time to look at the above
issue.

Thanks



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
Taras Ledkov Taras Ledkov
Reply | Threaded
Open this post in threaded view
|

Re: JDBC Thin Client does not return

In reply to this post by pg31
Hi,

I've not seen the thread dump from the node with order 3:
'CAA9E3DB-266B-4040-B059-19BD281D5821'
(corresponds to log: ignite-caa9e3db.0.log) and from other nodes too.
It may hep a lot if the thread dumps are created when the query hangs.
Server node + client app thread dumps - the best info.

But JDBC was connected to this node (looks like the nodes' logs).

The described case is not clear for me.

But i see a lot of strange log messages at the node with order 3:
[...][WARNING][grid-nio-worker-client-listener-0-#30][ClientListenerProcessor]
Closing NIO session because of unhandled exception [cls=class
o.a.i.i.util.nio.GridNioException, msg=Operation timed out]

[...][SEVERE][grid-nio-worker-client-listener-0-#30][ClientListenerProcessor]
Closing NIO session because of unhandled exception.
class org.apache.ignite.IgniteCheckedException: Invalid handshake message

[...][WARNING][grid-nio-worker-tcp-comm-2-#26][TcpCommunicationSpi]
Unknown connection detected (is some other software connecting to this
Ignite port?) [rmtAddr=/x.x.x.x]

It looks like:
- some network clients/services are configured invalid.
- someone tried to hack you. Is not it?. %)

On 12.02.2020 8:31, pg31 wrote:

> Hi
>
> Thank You for the reply Denis.
>
> Please find the logs attached. The query was still running
> Create_IDX_Stuck.zip
> <http://apache-ignite-users.70518.x6.nabble.com/file/t2770/Create_IDX_Stuck.zip>
>
>
>
> --
> Sent from: http://apache-ignite-users.70518.x6.nabble.com/

--
Taras Ledkov
Mail-To: [hidden email]