Forum: Other

Post: VoltDB hangs

VoltDB hangs
energyd
Feb 9, 2016
Hi there,
I was trying to test an adHoc query in the web monitor center. In the log VoltDB keeps saying "HOST: Procedure AdHoc_RO_MP is taking a long time to execute" and the whole cluster refuse to take any other queries except system procedure call(just hang after I type any other query).

I then stopped one of the three host in the cluster and rejoined it, hoping it would somehow stop the query execution and put the cluster back to normal. But the rejoin itself seemed to have failed. On the host that is rejoining, the log keeps saying "SNAPSHOT: Partition detection was unable to submit a snapshot requestbecause one already existed. Retrying." What should I do?

How do I stop the execution of this query?
jpiekos
Feb 9, 2016
You can set a query timeout for the database. In version 6.0 we set a default query timeout of 10 seconds, which will abort any long running read after that amount of time.

For details on this setting, see: https://docs.voltdb.com/AdminGuide/HostConfigDBOpts.php
Section A.3.9.

John
energyd
Feb 9, 2016
You can set a query timeout for the database. In version 6.0 we set a default query timeout of 10 seconds, which will abort any long running read after that amount of time.

For details on this setting, see: https://docs.voltdb.com/AdminGuide/HostConfigDBOpts.php
Section A.3.9.

John


Is there a way to set the default before 6.0? (I'm on 5.8.0)
jpiekos
Feb 9, 2016
Yes, the same setting/directions work for v5.8. v5.8 did not have a default timeout setting enabled (now v6.0, our latest release, does).

John