[
https://issues.apache.org/jira/browse/METRON-793?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15938439#comment-15938439
]
ASF GitHub Bot commented on METRON-793:
---------------------------------------
Github user cestella commented on a diff in the pull request:
https://github.com/apache/incubator-metron/pull/486#discussion_r107682535
--- Diff:
metron-platform/metron-pcap-backend/src/main/java/org/apache/metron/spout/pcap/HDFSWriterCallback.java
---
@@ -96,16 +109,18 @@ public HDFSWriterCallback withConfig(HDFSWriterConfig
config) {
this.config = config;
return this;
}
+
@Override
public List<Object> apply(List<Object> tuple, EmitContext context) {
-
- List<Object> keyValue = (List<Object>) tuple.get(0);
- LongWritable ts = (LongWritable) keyValue.get(0);
- BytesWritable rawPacket = (BytesWritable)keyValue.get(1);
+ byte[] key = (byte[]) tuple.get(0);
+ byte[] value = (byte[]) tuple.get(1);
+ if(!config.getDeserializer().deserializeKeyValue(key, value,
KeyValue.key.get(), KeyValue.value.get())) {
+ LOG.debug("Dropping malformed packet...");
--- End diff --
I could, but, of course, they're raw bytes and the value could be rather
big. How about I compromise and make this a bit more helpful by including:
* Their lengths
* The key itself in hex
> Migrate to storm-kafka-client kafka spout from storm-kafka
> ----------------------------------------------------------
>
> Key: METRON-793
> URL: https://issues.apache.org/jira/browse/METRON-793
> Project: Metron
> Issue Type: Improvement
> Reporter: Casey Stella
>
> In order to eventually support kerberos, the suggested path is to migrate to
> the new kafka spout (org.apache.storm:storm-kafka-client) which uses the new
> consumer API.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
