[IMPALA-212] Block metadata is showing up as -1 in the runtime profile

• Type: Bug
• Status: Resolved
• Priority: Blocker
• Resolution: Fixed
• Affects Version/s: Impala 0.7
• Fix Version/s: Impala 0.7
• Assignee: Alan Choi
• Reporter: Nong Li
• Created: 09/Apr/13 17:22
• Updated: 10/Apr/13 23:07
• Resolved: 10/Apr/13 23:07

Description

Alan, the queries running on the 10 node cluster having this in their runtime profiles:

HDFS_SCAN_NODE (id=0):(2s516ms 0.45%)
          File Formats: LZO_TEXT/NONE:241 
          Hdfs split stats (:<# splits>/): -1:241/13.96 GB                       <-------
           - AverageHdfsReadThreadConcurrency: 0.38 
             - HdfsReadThreadConcurrencyCountPercentage=0: 75.51 
             - HdfsReadThreadConcurrencyCountPercentage=1: 19.10 
             - HdfsReadThreadConcurrencyCountPercentage=10: 0.09 
             - HdfsReadThreadConcurrencyCountPercentage=11: 0.00 
             - HdfsReadThreadConcurrencyCountPercentage=12: 0.53 
             - HdfsReadThreadConcurrencyCountPercentage=2: 3.80 
             - HdfsReadThreadConcurrencyCountPercentage=3: 0.71 
             - HdfsReadThreadConcurrencyCountPercentage=4: 0.09 
             - HdfsReadThreadConcurrencyCountPercentage=5: 0.00 
             - HdfsReadThreadConcurrencyCountPercentage=6: 0.00 
             - HdfsReadThreadConcurrencyCountPercentage=7: 0.09 
             - HdfsReadThreadConcurrencyCountPercentage=8: 0.00 
             - HdfsReadThreadConcurrencyCountPercentage=9: 0.09 
           - AverageScannerThreadConcurrency: 34.15 

I think this is likely a reporting issue since the metrics look much reasonable
impala-server.scan-ranges.num-missing-volume-id:8
impala-server.scan-ranges.total:1936