Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-3186

Setp5 of sharding process is very slow

    XMLWordPrintableJSON

Details

    • Icon: Question Question
    • Resolution: Incomplete
    • Icon: Major - P3 Major - P3
    • None
    • 1.8.1
    • Sharding
    • archlinux, soft raid 1, lvm, 7200rpm harddisk, xeon 3xxx, 4G ECC

    Description

      The collection being sharded has 25M records.
      First 4 steps only takes few seconds. The step5 (last one?) usually takes 20 minutes.
      It would take about 2 days to shard the entire collection.
      What does the `step5` mean? And how can I make the process faster?

      the output of db.currentOp() :

      {
      "opid" : 6903,
      "active" : true,
      "lockType" : "write",
      "waitingForLock" : false,
      "secs_running" : 940,
      "op" : "query",
      "ns" : "my_db.my_collection",
      "query" : {
      "moveChunk" : "my_db.my_collection",
      "from" : "rs1/mgrs1a:27017",
      "to" : "rs3/mgrs3a:27017",
      "min" :

      { "date" : "20100511", "place" : "02102010101" }

      ,
      "max" :

      { "date" : "20100520", "place" : "02102010101" }

      ,
      "maxChunkSizeBytes" : NumberLong(67108864),
      "shardId" : "my_db.my_collection-date_\"20100511\"place_\"02102010101\"",
      "configdb" : "mgconf1:27017,mgconf2:27017,mgconf3:27017"
      },
      "client" : "10.0.0.31:45831",
      "desc" : "conn",
      "msg" : "step5"
      }

      Attachments

        Activity

          People

            greg_10gen Greg Studer
            zir Senmiao Liu
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: