[
https://issues.apache.org/jira/browse/HBASE-10292?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13865189#comment-13865189
]
Anoop Sam John edited comment on HBASE-10292 at 1/8/14 8:14 AM:
----------------------------------------------------------------
Oh sorry for the confusion. I was not saying about retry in this test .
What I was refering CoprocessorHost#handleCoprocessorThrowable
{code}
if (e instanceof IOException) {
throw (IOException)e;
}
// If we got here, e is not an IOException. A loaded coprocessor has a
// fatal bug, and the server (master or regionserver) should remove the
// faulty coprocessor from its set of active coprocessors. Setting
// 'hbase.coprocessor.abortonerror' to true will cause abortServer(),
// which may be useful in development and testing environments where
// 'failing fast' for error analysis is desired.
if
(env.getConfiguration().getBoolean("hbase.coprocessor.abortonerror",false)) {
// server is configured to abort.
abortServer(env, e);
} else {
LOG.error("Removing coprocessor '" + env.toString() + "' from " +
"environment because it threw: " + e,e);
coprocessors.remove(env);
try {
shutdown(env);
} catch (Exception x) {
LOG.error("Uncaught exception when shutting down coprocessor '"
+ env.toString() + "'", x);
}
throw new DoNotRetryIOException("Coprocessor: '" + env.toString() +
"' threw: '" + e + "' and has been removed from the active " +
"coprocessor set.", e);
}
{code}
Was wondering why we are not throwing back DNRIOE to client when RS also
aborts. May be that is correct and intended.
was (Author: anoop.hbase):
Oh sorry for the confusion. I was not saying about retry in this test .
What I was refering CoprocessorHost#handleCoprocessorThrowable
{code}
if (e instanceof IOException) {
throw (IOException)e;
}
// If we got here, e is not an IOException. A loaded coprocessor has a
// fatal bug, and the server (master or regionserver) should remove the
// faulty coprocessor from its set of active coprocessors. Setting
// 'hbase.coprocessor.abortonerror' to true will cause abortServer(),
// which may be useful in development and testing environments where
// 'failing fast' for error analysis is desired.
if
(env.getConfiguration().getBoolean("hbase.coprocessor.abortonerror",false)) {
// server is configured to abort.
abortServer(env, e);
} else {
LOG.error("Removing coprocessor '" + env.toString() + "' from " +
"environment because it threw: " + e,e);
coprocessors.remove(env);
try {
shutdown(env);
} catch (Exception x) {
LOG.error("Uncaught exception when shutting down coprocessor '"
+ env.toString() + "'", x);
}
throw new DoNotRetryIOException("Coprocessor: '" + env.toString() +
"' threw: '" + e + "' and has been removed from the active " +
"coprocessor set.", e);
}
{code}
Was wondering why we are not throwing back DNRIOE to client when RS also
aborts.
> TestRegionServerCoprocessorExceptionWithAbort fails occasionally
> ----------------------------------------------------------------
>
> Key: HBASE-10292
> URL: https://issues.apache.org/jira/browse/HBASE-10292
> Project: HBase
> Issue Type: Bug
> Affects Versions: 0.98.0
> Reporter: Andrew Purtell
> Assignee: Andrew Purtell
> Fix For: 0.98.0
>
> Attachments: 10292.patch, 10292.patch
>
>
> TestRegionServerCoprocessorExceptionWithAbort has occasionally failed for a
> very long time now. Fix or disable.
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
