Uploaded image for project: 'Couchbase Server'
  1. Couchbase Server
  2. MB-6605

unexecpted swap usage on node while loading data

    XMLWordPrintable

Details

    • Bug
    • Resolution: Fixed
    • Major
    • 2.0
    • 2.0-beta
    • performance
    • Security Level: Public
    • build-1708
      9 node cluster

      data loader command:
      python cbsystest.py run workload --create 100 --ops 20000 --postcondition "curr_items > 15000000"

    Description

      Loaded 15 Million items(sets only) on a 9 node cluster, causing it to go into dgm with 77% active-resident, 68% replica-resident
      No views, no rebalancing.

      Bucket Compaction kicks in and some nodes go into swap.

      Attaching atop history from node(10.3.3.59) as it has some swap usage, along with diags from master node (59) with screenshot.

      output from free on 10.3.3.59:
      root@plum-016:~# free
      total used free shared buffers cached
      Mem: 5958344 5905504 52840 0 28044 1497412
      -/+ buffers/cache: 4380048 1578296
      Swap: 4194296 236 4194060

      Recent excerpt from atop:
      ATOP - plum-016 2012/09/11 12:23:32 20 seconds elapsed
      PRC | sys 0.98s | user 2.10s | #proc 117 | #zombie 0 | #exit 3 |
      CPU | sys 2% | user 5% | irq 0% | idle 304% | wait 88% |
      CPU | steal 0% | stl/cpu 0% | | | |
      cpu | sys 1% | user 3% | irq 0% | idle 5% | cpu000 w 90% |
      cpu | sys 0% | user 1% | irq 0% | idle 98% | cpu001 w 1% |
      cpu | sys 0% | user 0% | irq 0% | idle 100% | cpu003 w 0% |
      cpu | sys 0% | user 0% | irq 0% | idle 98% | cpu002 w 2% |
      CPL | avg1 1.20 | avg5 1.60 | avg15 1.91 | csw 21116 | intr 13087 |
      MEM | tot 5.7G | free 41.2M | cache 1.4G | buff 28.2M | slab 73.3M |
      SWP | tot 4.0G | free 4.0G | | vmcom 4.4G | vmlim 6.8G |
      PAG | scan 8008 | stall 0 | | swin 0 | swout 0 |
      DSK | xvdb | busy 49% | read 1341 | write 1320 | avio 7 ms |
      DSK | xvda | busy 9% | read 136 | write 308 | avio 8 ms |
      NET | transport | tcpi 1632 | tcpo 1610 | udpi 0 | udpo 0 |
      NET | network | ipi 1692 | ipo 1610 | ipfrw 0 | deliv 1635 |
      NET | lo ---- | pcki 420 | pcko 420 | si 586 Kbps | so 586 Kbps |
      NET | eth0 ---- | pcki 1414 | pcko 1191 | si 108 Kbps | so 256 Kbps |

      PID SYSCPU USRCPU VGROW RGROW RDDSK WRDSK ST EXC S CPU CMD 1/1
      11914 0.57s 1.40s 13380K 5976K 38296K 41856K – - S 5% beam.smp
      15723 0.12s 0.27s 0K 0K - - NE 0 E 1% <couch_compac>
      15722 0.08s 0.20s 0K 0K - - -E 0 E 1% <couch_compac>
      12061 0.02s 0.16s 0K 0K 0K 144K – - S 0% memcached
      15726 0.11s 0.01s 18616K 1568K 7552K 5824K N- - D 0% couch_compact
      13482 0.02s 0.01s 0K 0K 0K 12K – - R 0% atop
      15724 0.00s 0.03s 50604K 2672K 2376K 0K N- - S 0% sshd
      398 0.02s 0.00s 0K 0K 0K 132K – - S 0% kjournald
      29196 0.01s 0.00s 0K 0K 0K 0K – - S 0% bash
      707 0.01s 0.00s 0K 36K 628K 0K – - S 0% sshd
      15727 0.00s 0.01s 50604K 1096K 0K 0K N- - S 0% sshd
      12074 0.00s 0.01s 0K 0K 0K 0K – - S 0% sigar_port
      51 0.01s 0.00s 0K 0K 0K 0K – - S 0% kswapd0
      189 0.01s 0.00s 0K 0K 0K 76K – - S 0% kjournald
      11949 0.00s 0.00s 0K 0K 0K 0K – - S 0% sh
      11950 0.00s 0.00s 0K 0K 0K 0K – - S 0% memsup
      15725 0.00s 0.00s 0K 0K - - NE 0 E 0% <df>

      Attachments

        No reviews matched the request. Check your Options in the drop-down menu of this sections header.

        Activity

          People

            tommie Tommie McAfee (Inactive)
            tommie Tommie McAfee (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved:

              Gerrit Reviews

                There are no open Gerrit changes

                PagerDuty