GitHub user zellerh opened a pull request:
https://github.com/apache/incubator-trafodion/pull/408
TRAFODION-1910 mxosrvr crashes on Hive query after reconnect (take 2)
NATableDB is caching a pointer to a HiveClient_JNI object
(HiveMetaData::client_), but that object gets deallocated when a JDBC
client disconnects. Fixing this by keeping the HiveClient_JNI around
across sessions.
Selva and Suresh commented on the first fix and suggested to treat both
HBaseClient_JNI and HiveClient_JNI the same and to remove the CLI
interface that's used to delete these objects.
Therefore, the new fix is to remove this CLI call. It gets called from
two places, one is when an ODBC/JDBC connection closes and the other
is from "initialize trafodion, drop". We believe that neither of them
is needed. Note that we have only one object of each type per CLI
context, and that we delete both objects when we delete the context
(ContextCli::deleteMe()), so there are no leaks.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/zellerh/incubator-trafodion bug/1910
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/incubator-trafodion/pull/408.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #408
----
commit 841321a57d03343faf7d894ff8987ea677bde05b
Author: Hans Zeller <[email protected]>
Date: 2016-03-31T00:08:50Z
TRAFODION-1910 mxosrvr crashes on Hive query after reconnect (take 2)
NATableDB is caching a pointer to a HiveClient_JNI object
(HiveMetaData::client_), but that object gets deallocated when a JDBC
client disconnects. Fixing this by keeping the HiveClient_JNI around
across sessions.
Selva and Suresh commented on the first fix and suggested to treat both
HBaseClient_JNI and HiveClient_JNI the same and to remove the CLI
interface that's used to delete these objects.
Therefore, the new fix is to remove this CLI call. It gets called from
two places, one is when an ODBC/JDBC connection closes and the other
is from "initialize trafodion, drop". We believe that neither of them
is needed. Note that we have only one object of each type per CLI
context, and that we delete both objects when we delete the context
(ContextCli::deleteMe()), so there are no leaks.
----
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
