Forum Discussion

mokkan's avatar
mokkan
Level 6
12 years ago
Solved

Veritas cluster basic quesiton about hastop -local -force

Hello,

I was reading a thread here and if I do hastop  -local -force on the node, it will leave service group and running. I have done it our VCS and service group went offline.

Am I doing some thing wrong here?

  • If you run "hastop  -local -force" it will stop the VCS processes on the node you run it on, but it will not run the offline entry points for the resources online on that node, so if for example, Oracle is running in a service group on that node, then VCS will not stop the database and umount the filesystems so if you checked on the node, then Oracle should still be running and the filesystems mounted.  However as VCS is stopped on that node, then the other node may report the service group is offline as it doesnt know the state.

    Mike

5 Replies

  • If you run "hastop  -local -force" it will stop the VCS processes on the node you run it on, but it will not run the offline entry points for the resources online on that node, so if for example, Oracle is running in a service group on that node, then VCS will not stop the database and umount the filesystems so if you checked on the node, then Oracle should still be running and the filesystems mounted.  However as VCS is stopped on that node, then the other node may report the service group is offline as it doesnt know the state.

    Mike

  • Hello,

    Agree with Mike, hastop -local -force will take all the VCS process offlline but the configured resources inside will be running.

    As there is no VCS process running, there is no one to manage VCS service groups as well hence service groups also go offline but the -force switch ensures that what ever was configured inside VCS like mount points, volumes, diskgroups, databases, applications etc, that will be kept running.

     

    G

  • Thank you very much Mike and Gaurav.

    If I undertand correctly, basically  other node can't find the status of this node. If I do hastatus  -sum on the other node it would show as unknow state, but if I have file system mounted using one of the resource, there is no affect, but when we do hastatus -sum it would show as offline or unknown state right?

     

    Also,  if I start the VCS on the node, would it restart the process? For an example if there is mount resource, would it remount ?

    When you say  "it will stop the VCS processes" It means it will stop  HAD daemon and HAShadow daemon?

    LLT, GAB, Application processes will run right?

     

    Thanks in advance.

  • See answers below:

    If I undertand correctly, basically  other node can't find the status of this node. Correct

    If I do hastatus  -sum on the other node it would show as unknow state, but if I have file system mounted using one of the resource, there is no affect, but when we do hastatus -sum it would show as offline or unknown state right? Correct

     

    Also,  if I start the VCS on the node, would it restart the process? For an example if there is mount resource, would it remount ? No - when you start VCS, VCS probes resources (runs monitor entry points) and if resource is online (example filesystem is mounted), then VCS marks resource as online - it will not restart the process so it would not remount a filesystem

    When you say  "it will stop the VCS processes" It means it will stop  HAD daemon and HAShadow daemon? Yes - and also the VCS agent processes.  VCS runs one VCS agent process per agent type so for example you will have a process for Mount and process for Diskgroup and these processes are stopped when you stop VCS

    LLT, GAB, Application processes will run right? Correct

  • You made my Day.

     

    Thank you very much Mikebounds.  You are bering very helpful.