This discussion is archived
1 2 Previous Next 22 Replies Latest reply: Oct 18, 2012 3:27 AM by 864103 RSS

rac issue

864103 Newbie
Currently Being Moderated
hi all ,

just for clear , if i have oracle 11g on linux my database run on two node with asm
and i have application server connected to database ,the first node it master and the second node it's slave
and actually this node on seperate server and this server have problem we can't use it from FVM and we can't access by ssh ,
so if we restart this server i think there is no problem on the node one ??? just confirm please ??? and correct me if i am wrong.

many thanks
  • 1. Re: rac issue
    sb92075 Guru
    Currently Being Moderated
    861100 wrote:
    hi all ,

    just for clear , if i have oracle 11g on linux my database run on two node with asm
    and i have application server connected to database ,the first node it master and the second node it's slave
    and actually this node on seperate server
    I don't know which node is "this" node.
    It is better to use actual hostnames to avoid confusion like above.
    and this server have problem
    my car has problem.
    tell me how to make my car go.

    It is better to use COPY & PASTE to show actual & complete ERROR MESSAGE & CODE!
    we can't use it from FVM
    what is FVM?
    and we can't access by ssh ,
    Why not? is COPY & PASTE broken for you?
  • 2. Re: rac issue
    phaeus Pro
    Currently Being Moderated
    Hello,
    you should give us more information, but here what i think that you mean.
    Normaly there is no problem when you restart one node in a rac cluster, if your application can handle this without problem. In your situation (if i understand this correct) one node is not responding anymore on the public lan. But it can have a network or disk heartbeat. So in your situation there is no clean restart possible, so there can be a reason that your split brain of the cluster can make problems to the hole system.

    You need to know, in which way your cluster knows in the moment from the second node. You can check this with commands like crsctl status res -t or srvctl status database and so on.
    Also your logfile under GRID_HOME/log can tell you more about your situation.
    And for the end ssh is nice. Did you also try ssh from one node to the other over the private interface?


    regards
    Peter
  • 3. Re: rac issue
    rukbat Guru Moderator
    Currently Being Moderated
    +... this thread has been relocated FROM the Database General forum+
    to the RAC forum (and your duplicate post in the RAC INSTALLATION forum has been removed) ...
    It seems you need to pay attention to where you should place your inquiries.
  • 4. Re: rac issue
    864103 Newbie
    Currently Being Moderated
    sory for this


    hostname:medi01 this is master node active and running ,sid medi1
    hostname:medi02 this is slave node now this node is down ,sid medi2
    the medi02 can't accessable by KVM monitor and also by ssh not reachable
    also we have app server 10g on win server 2008
    so we plan to restart this server ,is this affected on first node master node (is there is down time required on this scenario).


    i hope you get my idea .

    thanks
  • 5. Re: rac issue
    sb92075 Guru
    Currently Being Moderated
    861100 wrote:
    sory for this


    hostname:medi01 this is master node active and running ,sid medi1
    hostname:medi02 this is slave node now this node is down ,sid medi2
    the medi02 can't accessable by KVM monitor and also by ssh not reachable
    If medi02 is comatose & totally non-responsive, you can't make it worse by cycling power on it.
    also we have app server 10g on win server 2008
    so we plan to restart this server ,
    THIS? server is which server??????????????? App Server?
    use only hostname; not indefinite pronouns!
    is this affected on first node master node (is there is down time required on this scenario).
    use only hostname; not indefinite pronouns!
  • 6. Re: rac issue
    864103 Newbie
    Currently Being Moderated
    hostname:medi01 this is master node active and running ,sid medi1
    hostname:medi02 this is slave node now this node is down ,sid medi2
    the medi02 can't accessable by KVM monitor and also by ssh not reachable
    also we have app server 10g on win server 2008 (avaapps01)
    so we plan to restart this server (medi02),is this affected on first node master node(medi01) (is there is down time required on this scenario).
  • 7. Re: rac issue
    damorgan Oracle ACE Director
    Currently Being Moderated
    There is no concept like "master-slave" in Oracle RAC.
  • 8. Re: rac issue
    Levi-Pereira Guru
    Currently Being Moderated
    Do not misunderstand my answer, but I far as I can understand you are looking for approval here on forum if you what you are planning to do is the right thing to do.

    With all respect to my fellow Members who have tried to help you, I believe you need a professional who understands your problem (a qualified DBA to resolve this matter) we do not have the objective view of the issue, and you can not explain it coherently.

    So I can't help you or approve anything you plan to do. Since you do not know if what you are doing is right or wrong.
  • 9. Re: rac issue
    864103 Newbie
    Currently Being Moderated
    node 1 host 1 medi01 sid medi1

    node 2 host 2 medi02 sid medi2
  • 10. Re: rac issue
    864103 Newbie
    Currently Being Moderated
    so base on that please close this thread , due to there is some member here do not want to share
    his experance on this issue , i put these thread just for clear this issue ...

    respectfully
  • 11. Re: rac issue
    damorgan Oracle ACE Director
    Currently Being Moderated
    At the risk of repeating my self I will in fact repeat myself.

    In RAC there is NO concept of master and slave.

    If you are assuming Oracle RAC is in some manner similar to some other company's product you need to read the docs. If you think I am misunderstanding something please provide a link to the document you are reading that leads you to think I am incorrect.

    It seems to me your biggest RAC issue is not understanding what RAC is and how it works.
  • 12. Re: rac issue
    864103 Newbie
    Currently Being Moderated
    It seems to me your biggest RAC issue is not understanding what RAC is and how it works



    this is my issue i want to ask you if we have 2 node and one node is running and other node is down actually the server
    not accessible if there is any affect on node 1 if i restart the server belong to node 2 ???that's it .this is all the story
  • 13. Re: rac issue
    damorgan Oracle ACE Director
    Currently Being Moderated
    Of course there is an affect.

    When one node is offline the surviving node must accept 100% of the load. When the second node comes back online there will undoubtedly be remastering. But none of this has anything to do with "master" or "slave."

    RAC is a complex topic and, to be perfectly honest, I am still disturbed by your seeming lack of having read the docs. If you don't get the concepts well enough that you are throwing around keywords like "master" and "slave" your chance of having a successful experience are a very small number rapidly approaching zero.

    My recommendation is that you go to your management and have them bring in a consultant for 30-60 days to work with you and teach you what you need to know.
  • 14. Re: rac issue
    864103 Newbie
    Currently Being Moderated
    thanks for your reply ,i can read the document of rac , but now this is what i need to know
    of oracle real application cluster , due to we have an issue on the system ,we have server not reachable
    and not accessible from anywhere so we plan to restart this server , i think we don't have any option just restart
    i know what you say exactly concern the load but bye the way the second node is down , but what i understand from your answer
    it not effect on the first node and also on application server correct me if i am wrong???
1 2 Previous Next

Legend

  • Correct Answers - 10 points
  • Helpful Answers - 5 points