java process loads the CPU to more than 100 percent
Results 1 to 7 of 7

Thread: java process loads the CPU to more than 100 percent

  1. #1
    Active Member
    Join Date
    Oct 2012
    Posts
    16

    java process loads the CPU to more than 100 percent

    Iīm running Zimbra Release 8.0.7.GA.6021.UBUNTU12.64 UBUNTU12_64 FOSS edition, with ZeXtras 1.10.1
    with 45 mail boxes and 14 domains.
    Normally Once or twice a week our Zimbra server is very slow, we cannot work, due to a java process that loads the CPU to more than 100%.
    If I restart the mailbox everything backs to normal and stays there for another couple of days.
    The server has 5082108k memory, 4767044k used, swap 598980k used.
    I follow this thread: Excessive NIO write buffering leads to mailboxd service outage
    But no luck.
    Any help?

  2. #2
    ZeXtras Community Manager ZeXtras Employee Cine's Avatar
    Join Date
    Apr 2011
    Posts
    2,364
    Hello Ferra!

    The NIO write buffer issue was solved in a Zimbra version older than 8.0.7 so I highly doubt this is the cause... Usually this kind of issue is caused by either a misconfiguration of Zimbra's JVM settings: please check Performance Tuning Guidelines for Large Deployments - Zimbra :: Wiki and make sure that all of your settings are coherent with your setup - especially the "-XX:PermSize=" and "-XX:MaxPermSize=" parts of the "mailboxd_java_options" setting and the "mailboxd_java_heap_size" setting.

    You'll likely find more informations about the cause of the issue in the mailbox.log file and/or in the logfiles generated by the "zmdiaglog" Zimbra tool (run it during a spike for the best chances possible).

    Have a nice day,
    Cine
    the ZeXtras Team

  3. #3
    Active Member
    Join Date
    Oct 2012
    Posts
    16
    Hi Cine,

    I don't know what exactly looking for in the logs, but I found this:

    sync.log:2014-12-09 13:43:01,052 WARN [cristina@dominio.pt Appl70044DG3A4S] [] sync - FolderSync[1156]: com.zextras.lib.Error.UnknownSyncStateError: Client 7b103cf0-a37d-4b76-8c47-d9c00400dbd8\Appl70044DG3A4S sent an unknown SyncState: 7b103cf0a37d4b768c47d9c00400dbd8/0/68/1
    sync.log:2014-12-09 13:56:07,197 WARN [fil@dominio.pt EHM6Q7C8D52CP4KC301T9R1FC8] [] sync - Ping[1201]: Folder hierarchy[1]: 157f3fc7-fad6-4146-b3f7-050be2dc5a85\EHM6Q7C8D52CP4KC301T9R1FC8

    Running this command zmlocalconfig mailboxd_java_options get:
    mailboxd_java_options = -server -Djava.awt.headless=true -Dsun.net.inetaddr.ttl=60 -XX:+UseConcMarkSweepGC -XX:PermSize=196m -XX:MaxPermSize=350m -XX:SoftRefLRUPolicyMSPerMB=1 -verbose:gc -XX:+PrintGCDetails -XX:+PrintGCTimeStamps -XX:+PrintGCApplicationStoppedTime -XX:-OmitStackTraceInFastThrow -Djava.net.preferIPv4Stack=true -Dorg.apache.jasper.compiler.disablejsr199=true

    and

    $ zmlocalconfig mailboxd_java_heap_size mailboxd_java_heap_memory_percent
    Warning: null valued key 'mailboxd_java_heap_memory_percent'
    mailboxd_java_heap_size = 1228



    Any help?

    Thanks

  4. #4
    ZeXtras Community Manager ZeXtras Employee Cine's Avatar
    Join Date
    Apr 2011
    Posts
    2,364
    Hello ferra!

    The sync.log error is not really helpful in this case, as the sync.log only includes ZxMobile logs while the issue here is with the mailboxd service - whose logs are written in the mailbox.log file.

    Heap Size and PermGen settings looks good, so I'd suggest to chech which process takes up all the CPU, search for errors in the mailbox.log (or in any other log based on the affected process) and look for pointers on the official Zimbra Community and Bugzilla.
    Unfortunately I can't help you further because investigating this kind of issues requires a knowledge of the infrastructure itself, but if you find any specific error feel free to post those here...

    Have a nice day,
    Cine
    the ZeXtras Team

  5. #5
    Active Member
    Join Date
    Oct 2012
    Posts
    16
    Canīt find any error and keep having the same problem. Maybe could be the virtualization platform!?

    Regards

  6. #6
    ZeXtras Community Manager ZeXtras Employee Cine's Avatar
    Join Date
    Apr 2011
    Posts
    2,364
    Quote Originally Posted by ferra View Post
    Canīt find any error and keep having the same problem. Maybe could be the virtualization platform!?

    Regards
    Hello Ferra!

    It's possible for the hypervisor to be the cause - either directly or because of a bad setup... One thing that comes to my mind is to make sure that the VM's RAM memory is not limited to a lower amount than the one available to the VM itself, but that's just one of the possibilities...

    Have a nice day,
    Cine
    the ZeXtras Team

  7. #7
    Active Member
    Join Date
    Oct 2012
    Posts
    16
    Yeahh it's an idea. At the moment the RAM guaranted is 2048MB and RAM size is 5120MB. I'll change it as soon as possible.

    Thanks very much.

Bookmarks

Posting Permissions

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