Results 1 to 4 of 4

Thread: Zimbra High Load Average with high iowait

  1. #1
    Join Date
    Jan 2013
    Posts
    4
    Rep Power
    2

    Unhappy Zimbra High Load Average with high iowait

    Hi all


    For one week i identified high load average and high iowait in my Zimbra server. When it happens i ran top, iotop, iostat and hdparm (
    as can be seen in attached screenshot)


    email_top_3_edit.png



    -> iostat

    avg-cpu: %user %nice %system %iowait %steal %idle
    8,26 3,00 3,34 75,73 0,00 9,67

    Device: rrqm/s wrqm/s r/s w/s rsec/s wsec/s avgrq-sz avgqu-sz await svctm %util
    scd0 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00
    sdb 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00
    sda 0,00 746,00 24,00 403,00 232,00 17296,00 41,05 123,80 57886,14 2,34 100,00
    vg1-tmp 0,00 0,00 0,00 11,00 0,00 88,00 8,00 1,77 33719,82 84,91 93,40
    vg1-optzimbraredolog 0,00 0,00 0,00 21,00 0,00 168,00 8,00 1,96 14884,29 45,14 94,80
    vg1-optzimbralogger 0,00 0,00 0,00 43,00 0,00 344,00 8,00 45,00 238246,33 21,49 92,40
    vg1-optzimbradb 0,00 0,00 1,00 107,00 8,00 8224,00 76,22 4,79 3900,33 9,13 98,60
    vg1-optzimbra 0,00 0,00 6,00 286,00 48,00 2288,00 8,00 126,71 90326,38 3,29 96,00
    vg1-optzimbralog 0,00 0,00 0,00 19,00 0,00 152,00 8,00 11,75 126996,53 47,74 90,70
    vg1-usr 0,00 0,00 1,00 0,00 8,00 0,00 8,00 0,80 64819,00 800,00 80,00
    vg1-var 0,00 0,00 0,00 34,00 0,00 272,00 8,00 8,05 48605,09 27,21 92,50
    vg1-optzimbrastore 0,00 0,00 11,00 463,00 88,00 3704,00 8,00 6,18 739,12 2,09 99,30
    vg1-optzimbraindex 0,00 0,00 0,00 25,00 0,00 200,00 8,00 3,30 6950,48 36,72 91,80
    vg1-backup 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00 0,00

    HDPARM

    -> With low load:
    hdparm -T -t /dev/sda

    /dev/sda:
    Timing cached reads: 13986 MB in 2.00 seconds = 6999.67 MB/sec
    Timing buffered disk reads: 928 MB in 3.01 seconds = 308.78 MB/sec


    -> With high load:
    hdparm -T -t /dev/sda

    /dev/sda:
    Timing cached reads: 13958 MB in 2.00 seconds = 6985.69 MB/sec
    Timing buffered disk reads: 874 MB in 13.83 seconds = 63.19 MB/sec


    The system serves 2.500 mailboxes, it runs on CentOS 6.3 x86_64 with Release 8.0.2_GA_5569.RHEL6_64_20121210115059 CentOS6_64 FOSS edition. The server runs in VMWare environment with 12 CPUs and 16 GB of RAM and current vmware tool. I already checked the mail queues during peak and was normal status. This event has been happening for a week without specific times (as can be seen in the graphs ) however had no change in the environment these last days.

    21-07-2014-Load Average-1.png24-07-2014-Load Average.png

    Any idea why this is happening?

  2. #2
    phoenix is offline Zimbra Consultant & Moderator
    Join Date
    Sep 2005
    Location
    Vannes, France
    Posts
    23,587
    Rep Power
    58

    Default

    The first thing you need to do is upgrade (I mean immediately) your ZCS version because of this: http://www.zimbra.com/forums/adminis...tml#post328213

    I'd also suggest you might have too many vCPUs on that server, have you also looked at the wiki article on VMware performance tuning?
    Regards


    Bill


    Acompli: A new adventure for Co-Founder KevinH.

  3. #3
    Join Date
    Jan 2013
    Posts
    4
    Rep Power
    2

    Default

    Quote Originally Posted by phoenix View Post
    The first thing you need to do is upgrade (I mean immediately) your ZCS version because of this: http://www.zimbra.com/forums/adminis...tml#post328213

    I'd also suggest you might have too many vCPUs on that server, have you also looked at the wiki article on VMware performance tuning?
    Thanks for the response Bill. I running through the doc you mentioned and didn't detect any vulnerabilities

    [root@mail tmp]# find /opt/zimbra/zimlets-deployed -mtime -35 -o -ctime -35 -> nothing returns

    [root@mail tmp]# find /opt/zimbra -iname \*.jsp -exec grep -i getRuntime {} \; -> nothing returns

    I will reduce the allocated vCPUs to 4 as shown in wiki
    and planning to migrate to the latest version ASAP.

    Is there any documents that can assist me in the migration process ? steps?

  4. #4
    phoenix is offline Zimbra Consultant & Moderator
    Join Date
    Sep 2005
    Location
    Vannes, France
    Posts
    23,587
    Rep Power
    58

    Default

    Quote Originally Posted by sarah.milani View Post
    Thanks for the response Bill. I running through the doc you mentioned and didn't detect any vulnerabilities
    It wasn't the thread I meant you to read (but it's good you don't have those problems) it was specifically the last post by Quanha about LDAP corruption that's the reason for my comment about upgrading.

    Quote Originally Posted by sarah.milani View Post
    Is there any documents that can assist me in the migration process ? steps?
    The Release Notes always carry details of the ipgrade process, it's basically a backup then run the installer script but do check the notes for any possibly problems/bugs in the release notes.
    Regards


    Bill


    Acompli: A new adventure for Co-Founder KevinH.

Similar Threads

  1. high cpu load with zimbra + centos
    By anandi in forum Installation
    Replies: 0
    Last Post: 02-05-2013, 07:04 AM
  2. Load average too high.
    By soscode in forum Administrators
    Replies: 2
    Last Post: 01-31-2013, 06:34 PM
  3. High CPU Load on zimbra 8.0.0
    By Paul Csiki in forum Administrators
    Replies: 0
    Last Post: 12-18-2012, 03:51 AM
  4. zimbra 8 load average high
    By neilkcl in forum Administrators
    Replies: 0
    Last Post: 10-29-2012, 03:29 AM
  5. High iowait from logger after upgrade to 4.5.9
    By pi13137 in forum Installation
    Replies: 9
    Last Post: 11-09-2007, 04:29 AM

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •