HBase
  1. HBase
  2. HBASE-11236

Last flushed sequence id is ignored by ServerManager

    Details

    • Type: Bug Bug
    • Status: Reopened
    • Priority: Major Major
    • Resolution: Unresolved
    • Affects Version/s: None
    • Fix Version/s: None
    • Component/s: None
    • Labels:
      None

      Description

      I got lots of error messages like this:

      2014-05-22 08:58:59,793 DEBUG [RpcServer.handler=1,port=20020] master.ServerManager: RegionServer a2428.halxg.cloudera.com,20020,1400742071109 indicates a last flushed sequence id (numberOfStores=9, numberOfStorefiles=2, storefileUncompressedSizeMB=517, storefileSizeMB=517, compressionRatio=1.0000, memstoreSizeMB=0, storefileIndexSizeMB=0, readRequestsCount=0, writeRequestsCount=0, rootIndexSizeKB=34, totalStaticIndexSizeKB=381, totalStaticBloomSizeKB=0, totalCompactingKVs=0, currentCompactedKVs=0, compactionProgressPct=NaN) that is less than the previous last flushed sequence id (605446) for region IntegrationTestBigLinkedList, �A��*t�^FU�2��0,1400740489477.a44d3e309b5a7e29355f6faa0d3a4095. Ignoring.

      RegionLoad.toString doesn't print out the last flushed sequence id passed in. Why is it less than the previous one?

        Activity

        Hide
        Pankaj Kumar added a comment -

        Duo Zhang Looks like root cause is different, we have HBASE-13811 in our version. I am also trying to figure out the relevant info from logs.

        Show
        Pankaj Kumar added a comment - Duo Zhang Looks like root cause is different, we have HBASE-13811 in our version. I am also trying to figure out the relevant info from logs.
        Hide
        Duo Zhang added a comment -

        I think this could happen without HBASE-13811 where stack fix a issue that we may report a wrong last flushed sequence id if a flush is aborted.

        Elliott Clark Any more informations? What's happened to the region before this log(reassignment or flush?)? Maybe there are other issues.

        Thanks.

        Show
        Duo Zhang added a comment - I think this could happen without HBASE-13811 where stack fix a issue that we may report a wrong last flushed sequence id if a flush is aborted. Elliott Clark Any more informations? What's happened to the region before this log(reassignment or flush?)? Maybe there are other issues. Thanks.
        Hide
        Pankaj Kumar added a comment -

        I also observed same log in product environment and regions were not opened.

        2015-08-31 19:04:25,448 | WARN  | PriorityRpcServer.handler=13,queue=1,port=21300 | RegionServer *.*.*.*,21302,1441017551749 indicates a last flushed sequence id (53128) that is less than the previous last flushed sequence id (53131) for region hbase:meta,,1 Ignoring. | org.apache.hadoop.hbase.master.ServerManager.updateLastFlushedSequenceIds(ServerManager.java:299)
        
        Show
        Pankaj Kumar added a comment - I also observed same log in product environment and regions were not opened. 2015-08-31 19:04:25,448 | WARN | PriorityRpcServer.handler=13,queue=1,port=21300 | RegionServer *.*.*.*,21302,1441017551749 indicates a last flushed sequence id (53128) that is less than the previous last flushed sequence id (53131) for region hbase:meta,,1 Ignoring. | org.apache.hadoop.hbase.master.ServerManager.updateLastFlushedSequenceIds(ServerManager.java:299)
        Hide
        Elliott Clark added a comment -

        I've seen this on 1.2 a decent ammount

        Show
        Elliott Clark added a comment - I've seen this on 1.2 a decent ammount
        Hide
        Jimmy Xiang added a comment -

        Don't see it any more. Close it.

        Show
        Jimmy Xiang added a comment - Don't see it any more. Close it.

          People

          • Assignee:
            Unassigned
            Reporter:
            Jimmy Xiang
          • Votes:
            0 Vote for this issue
            Watchers:
            8 Start watching this issue

            Dates

            • Created:
              Updated:

              Development