Hi,
I just upgraded our cluster from the CDH 5 beta to the newly available version 5.0.0-1, using parcels.
When restarting the cluster after upgrading, the Impala daemons fail to start. The error message in the log is as follows:
E0403 09:25:58.360676 59211 impala-server.cc:208] ERROR: block location tracking is not properly enabled because
- dfs.client.file-block-storage-locations.timeout.millis is too low. It should be at least 10 seconds.
E0403 09:25:58.361166 59211 impala-server.cc:210] Aborting Impala Server startup due to improper configuration
When searching for a configuration called 'dfs.client.file-block-storage-locations.timeout.millis' nothing shows up. What I can find is instead the setting 'dfs.client.file-block-storage-locations.timeout', which is set to 3000 seconds.
I am guessing that this configuration setting has changed name at some point in time, and that my software versions are somehow out of sync. Not really sure which one I should be looking at though, so any suggestion on how to resolve are appreciated.
Our current package versions are:
| Bigtop-Tomcat (CDH 5 only) | CDH5 | 0.7.0+cdh5.0.0+0 |
| Cloudera Manager Agent | Not applicable | 5.0.0-beta-2 |
| Cloudera Manager Management Daemons | Not applicable | 5.0.0-beta-2 |
| Crunch (CDH 5 only) | CDH5 | 0.9.0+cdh5.0.0+21 |
| Flume NG | CDH5 | 1.4.0+cdh5.0.0+111 |
| Hadoop | CDH5 | 2.3.0+cdh5.0.0+548 |
| MapReduce 1 | CDH5 | 2.3.0+cdh5.0.0+548 |
| HDFS | CDH5 | 2.3.0+cdh5.0.0+548 |
| HttpFS | CDH5 | 2.3.0+cdh5.0.0+548 |
| MapReduce 2 | CDH5 | 2.3.0+cdh5.0.0+548 |
| YARN | CDH5 | 2.3.0+cdh5.0.0+548 |
| HBase | CDH5 | 0.96.1.1+cdh5.0.0+60 |
| Lily HBase Indexer | CDH5 | 1.3+cdh5.0.0+38 |
| Hive | CDH5 | 0.12.0+cdh5.0.0+308 |
| HCatalog | CDH5 | 0.12.0+cdh5.0.0+308 |
| Hue | CDH5 | 3.5.0+cdh5.0.0+365 |
| Impala | CDH5 | 1.3.0+cdh5.0.0+0 |
Regards
\Knut Nordin
