AIX

AIX

Connect with fellow AIX users and experts to gain knowledge, share insights, and solve problems.

 View Only
  • 1.  kernel CPU usage great than 70%

    Posted Wed July 03, 2013 11:37 PM

    Originally posted by: yangwtao


    I have a AIX5300-12 sp6 syslog server is sufferring from CPU kernel usage is great than 70%. 

    Application is syslog and some shell script. No DB,no WAS etc...

    I am a new comer for performance. Anybody can teach me how to find out the reason.  Server setting as below listed.  Thanks!

     
    Topas Monitor for host:   xxxxxxxxxxxxxx     EVENTS/QUEUES    FILE/TTY
    Thu Jul  4 04:39:54 2013   Interval:  2         Cswitch    1815  Readch  4162.8K
                                                    Syscall   34609  Writech   15620
    CPU  User%  Kern%  Wait%  Idle%                 Reads      2616  Rawin         0
    ALL   16.0   84.0    0.0    0.0                 Writes      101  Ttyout      323
                                                    Forks       285  Igets         0
    Network  KBPS   I-Pack  O-Pack   KB-In  KB-Out  Execs       247  Namei      3665
    en0      13.1     59.0    16.0    11.4     1.6  Runqueue    3.5  Dirblk        0
    lo0       0.1      0.5     0.5     0.0     0.0  Waitqueue   0.0
     
    Disk    Busy%     KBPS     TPS KB-Read KB-Writ  PAGING           MEMORY
    hdisk1    0.0     0.0      0.0    0.0     0.0   Faults    29890  Real,MB    1024
    hdisk0    0.0     0.0      0.0    0.0     0.0   Steals        0  % Comp     39
    cd0       0.0     0.0      0.0    0.0     0.0   PgspIn        0  % Noncomp  45
                                                    PgspOut       0  % Client    2
    Name            PID  CPU%  PgSp Owner           PageIn        0
    sh            25996   6.1   0.5 syslog          PageOut       0  PAGING SPACE
    syslogd       11612   0.8   0.4 root            Sios          0  Size,MB    2560
    kulagent      14020   0.4   6.0 root                             % Used      5
    random         6098   0.3   0.0 root            NFS (calls/sec)  % Free     95
    topas         28874   0.2   1.2 cn210903        ServerV2      0
     

    $ vmstat -v
                   262144 memory pages
                   222549 lruable pages
                    39268 free pages
                        1 memory pools
                    77283 pinned pages
                     80.0 maxpin percentage
                     20.0 minperm percentage
                     80.0 maxperm percentage
                     53.2 numperm percentage
                   118495 file pages
                      0.0 compressed percentage
                        0 compressed pages
                      3.0 numclient percentage
                     80.0 maxclient percentage
                     6762 client pages
                        0 remote pageouts scheduled
                      145 pending disk I/Os blocked with no pbuf
                      104 paging space I/Os blocked with no psbuf
                    95545 filesystem I/Os blocked with no fsbuf
                        0 client filesystem I/Os blocked with no fsbuf
                       29 external pager filesystem I/Os blocked with no fsbuf
                        0 Virtualized Partition Memory Page Faults
                     0.00 Time resolving virtualized partition memory page faults
    $ no -a
                     arpqsize = 12
                   arpt_killc = 20
                  arptab_bsiz = 7
                    arptab_nb = 149
                    bcastping = 0
          clean_partial_conns = 0
                     delayack = 0
                delayackports = {}
             dgd_packets_lost = 3
                dgd_ping_time = 5
               dgd_retry_time = 5
           directed_broadcast = 0
             extendednetstats = 1
                     fasttimo = 200
            icmp6_errmsg_rate = 10
              icmpaddressmask = 0
    ie5_old_multicast_mapping = 0
                       ifsize = 256
              inet_stack_size = 16
                   ip6_defttl = 64
                    ip6_prune = 1
                ip6forwarding = 0
           ip6srcrouteforward = 1
           ip_ifdelete_notify = 0
                     ip_nfrag = 200
                 ipforwarding = 0
                    ipfragttl = 2
            ipignoreredirects = 0
                    ipqmaxlen = 100
              ipsendredirects = 1
            ipsrcrouteforward = 1
               ipsrcrouterecv = 0
               ipsrcroutesend = 1
              llsleep_timeout = 3
                      lo_perf = 1
                    lowthresh = 90
                     main_if6 = 0
                   main_site6 = 0
                     maxnip6q = 20
                       maxttl = 255
                    medthresh = 95
                   mpr_policy = 1
                  multi_homed = 1
                    nbc_limit = 131072
                nbc_max_cache = 131072
                nbc_min_cache = 1
             nbc_ofile_hashsz = 12841
                     nbc_pseg = 0
               nbc_pseg_limit = 262144
               ndd_event_name = {all}
            ndd_event_tracing = 0
                ndp_mmaxtries = 3
                ndp_umaxtries = 3
                     ndpqsize = 50
                    ndpt_down = 3
                    ndpt_keep = 120
                   ndpt_probe = 5
               ndpt_reachable = 30
                 ndpt_retrans = 1
                 net_buf_size = {all}
                 net_buf_type = {all}
            net_malloc_police = 0
               nonlocsrcroute = 0
                     nstrpush = 8
                  passive_dgd = 0
             pmtu_default_age = 10
                  pmtu_expire = 10
     pmtu_rediscover_interval = 30
                  psebufcalls = 20
                     psecache = 1
                 pseintrstack = 12288
                    psetimers = 20
               rfc1122addrchk = 0
                      rfc1323 = 0
                      rfc2414 = 1
                 route_expire = 1
              routerevalidate = 0
                     rto_high = 64
                   rto_length = 13
                    rto_limit = 7
                      rto_low = 1
                         sack = 0
                       sb_max = 1048576
           send_file_duration = 300
                  site6_index = 0
                   sockthresh = 85
                      sodebug = 0
                  sodebug_env = 0
                    somaxconn = 1024
                     strctlsz = 1024
                     strmsgsz = 0
                    strthresh = 85
                   strturncnt = 15
              subnetsarelocal = 1
           tcp_bad_port_limit = 0
                      tcp_ecn = 0
           tcp_ephemeral_high = 65535
            tcp_ephemeral_low = 32768
                 tcp_finwait2 = 1200
               tcp_icmpsecure = 0
              tcp_init_window = 0
        tcp_inpcb_hashtab_siz = 24499
                  tcp_keepcnt = 8
                 tcp_keepidle = 14400
                 tcp_keepinit = 150
                tcp_keepintvl = 150
         tcp_limited_transmit = 1
                  tcp_low_rto = 0
                 tcp_maxburst = 0
                  tcp_mssdflt = 1460
              tcp_nagle_limit = 65535
            tcp_nagleoverride = 0
                   tcp_ndebug = 100
                  tcp_newreno = 1
               tcp_nodelayack = 0
            tcp_pmtu_discover = 1
                tcp_recvspace = 16384
                tcp_sendspace = 16384
                tcp_tcpsecure = 0
                 tcp_timewait = 1
                      tcp_ttl = 60
               tcprexmtthresh = 3
                 tcptr_enable = 0
                      thewall = 524288
             timer_wheel_tick = 0
           udp_bad_port_limit = 0
           udp_ephemeral_high = 65535
            udp_ephemeral_low = 32768
        udp_inpcb_hashtab_siz = 24499
            udp_pmtu_discover = 1
                udp_recvspace = 42080
                udp_sendspace = 9216
                      udp_ttl = 30
                     udpcksum = 1
                     use_isno = 1
               use_sndbufpool = 1
    $  
    $ vmstat 1 10
     
    System configuration: lcpu=1 mem=1024MB
     
    kthr    memory              page              faults        cpu     
    ----- ----------- ------------------------ ------------ -----------
     r  b   avm   fre  re  pi  po  fr   sr  cy  in   sy  cs us sy id wa
     5  0 132019 38983   0   0   0   0    0   0 101 33461 1774  9 91  0  0
     4  0 131937 39065   0   0   0   0    0   0  62 34553 1818  9 91  0  0
     2  0 132022 38980   0   0   0   0    0   0 107 35060 1856 23 77  0  0
     5  0 131987 39015   0   0   0   0    0   0  67 35015 1837 14 86  0  0
     5  0 132003 38999   0   0   0   0    0   0  80 34604 1816 23 77  0  0
     4  0 132003 38999   0   0   0   0    0   0  59 34630 1828 16 84  0  0
     4  0 131986 39015   0   0   0   0    0   0  64 34603 1837 13 87  0  0
     4  0 131966 39034   0   0   0   0    0   0  75 34464 1848 16 84  0  0
     5  0 131981 39018   0   0   0   0    0   0  82 34496 1831 10 90  0  0
     5  0 132025 38973   0   0   0   0    0   0  82 34568 1840 10 90  0  0
    $

    $ oslevel -s
    5300-12-06-1216
    $

    XXXXXXXXXXX:/ # vmo -a
            cpu_scale_memp = 8
     data_stagger_interval = 161
                     defps = 1
       force_relalias_lite = 0
                 framesets = 2
                 htabscale = -1
         kernel_heap_psize = n/a
              kernel_psize = n/a
      large_page_heap_size = n/a
              lgpg_regions = 0
                 lgpg_size = 0
           low_ps_handling = 1
           lru_file_repage = 1
         lru_poll_interval = 10
                 lrubucket = 131072
                maxclient% = 80
                   maxfree = 1088
                   maxperm = 178039
                  maxperm% = 80
                    maxpin = 216811
                   maxpin% = 80
           mbuf_heap_psize = n/a
           memory_affinity = 1
             memory_frames = 262144
             memplace_data = 2
      memplace_mapped_file = 2
    memplace_shm_anonymous = 2
        memplace_shm_named = 2
            memplace_stack = 2
             memplace_text = 2
    memplace_unmapped_file = 2
                  mempools = 1
                   minfree = 960
                   minperm = 44509
                  minperm% = 20
                 nokilluid = 0
                   npskill = 5120
                 npsrpgmax = 40960
                 npsrpgmin = 30720
               npsscrubmax = 40960
               npsscrubmin = 30720
                   npswarn = 20480
          num_spec_dataseg = n/a
                 numpsblks = 655360
         page_steal_method = 0
              pagecoloring = 0
           pinnable_frames = 188697
      psm_timeout_interval = n/a
     pta_balance_threshold = 50
       relalias_percentage = 0
                  rpgclean = 0
                rpgcontrol = 2
                     scrub = 0
                scrubclean = 0
     soft_min_lgpgs_vmpool = 0
          spec_dataseg_int = n/a
          strict_maxclient = 1
            strict_maxperm = 0
                  v_pinshm = 0
      vm_modlist_threshold = -1
           vmm_fork_policy = 1
        vmm_mpsize_support = n/a
        wlm_memlimit_nonpg = 1
    sgsnidcwbzsyslog:/ # ioo -a
             j2_atimeUpdateSymlink = 0
     j2_dynamicBufferPreallocation = 16
                 j2_inodeCacheSize = 400
               j2_maxPageReadAhead = 128
                 j2_maxRandomWrite = 0
           j2_maxUsableMaxTransfer = 512
              j2_metadataCacheSize = 400
               j2_minPageReadAhead = 2
          j2_nBufferPerPagerDevice = 512
    j2_nPagesPerWriteBehindCluster = 32
                 j2_nRandomCluster = 0
          j2_nonFatalCrashesSystem = 0
             j2_syncModifiedMapped = 1
                  j2_syncPageCount = 0
                  j2_syncPageLimit = 256
           j2_syncdLogSyncInterval = 1
                     j2_unmarkComp = 0
                jfs_clread_enabled = 0
                 jfs_use_read_lock = 1
                        lvm_bufcnt = 9
                        maxpgahead = 8
                        maxrandwrt = 0
                     memory_frames = 262144
                        minpgahead = 2
                          numclust = 1
                         numfsbufs = 186
                         pd_npages = 65536
                pgahd_scale_thresh = 0
                       pv_min_pbuf = 256
                sync_release_ilock = 0
    XXXXXXXXXXXXX:/ # schedo -a
                  %usDelta = 100
              affinity_lim = 7
           allowMCMmigrate = 0
                 allow_vmx = 1
             big_tick_size = 1
     ded_cpu_donate_thresh = 80
                fast_locks = n/a
          fixed_pri_global = 0
                 force_grq = 0
           hotlocks_enable = n/a
    idle_migration_barrier = 4
        krlock_confer2self = n/a
      krlock_conferb4alloc = n/a
             krlock_enable = n/a
        krlock_spinb4alloc = n/a
       krlock_spinb4confer = n/a
                   maxspin = 16384
        n_idle_loop_vlopri = n/a
                  pacefork = 10
           proc_disk_stats = 1
                   sched_D = 16
                   sched_R = 16
     search_globalrq_mload = n/a
      search_smtrunq_mload = n/a
      setnewrq_sidle_mload = n/a
       shed_primrunq_mload = n/a
        sidle_S1runq_mload = n/a
        sidle_S2runq_mload = n/a
        sidle_S3runq_mload = n/a
        sidle_S4runq_mload = n/a
        slock_spinb4confer = n/a
          smt_snooze_delay = n/a
         smtrunq_load_diff = n/a
             tb_balance_S0 = 0
             tb_balance_S1 = 2
              tb_threshold = 100
                 timeslice = 1
            unboost_inflih = n/a
             v_exempt_secs = 2
             v_min_process = 2
               v_repage_hi = 0
             v_repage_proc = 4
                v_sec_wait = 1
           vpm_fold_policy = 1
                vpm_xvcpus = 0
    XXXXXXXXXXXX:/ #



  • 2.  Re: kernel CPU usage great than 70%

    Posted Thu July 04, 2013 05:43 AM

    Originally posted by: LaurentOliva


    i propose you to use "tprof" to see what the kernel is exactly doing :

    $ tprof -skeu -x sleep 10

    It will produce a file named sleep.prof. You will see what happen into the kernel.

    Then, i also propose you to turn on debug on syslog, to see if an application is sending too much message on /dev/log...

     

     



  • 3.  Re: kernel CPU usage great than 70%

    Posted Mon July 15, 2013 06:02 PM

    Originally posted by: GarlandJoseph


    Also suggest making sure you have the latest fixes on your system.