Forum Discussion

shiv124's avatar
shiv124
Level 4
13 years ago

hagrp flush and probe

Hi ,

I was reading abt hagrp -flush when troubleshooting problems on service group and rsources and was not able to understand when we would be using this hagrp -flush  and use of probe?

I t would be great if somebody could  explain with some example

I tried to read abt Flushing a service group but was not able to understand perfectly.

Occasionally, agents for the resources in a service group can appear to become
suspended waiting for resources to be brought online or be taken offline.------------->in this scenario  what would be the state of resource and service group.

Generally, this condition occurs during initial configuration and testing because
the required attributes for a resource are not defined properly or the underlying
operating system resources are not prepared correctly. If it appears that a resource
or group has become suspended while being brought online, you can flush the
service group to enable corrective action.------------------------------------->what they mean by suspended ? Is it s state of a resource group.

Flushing a service group stops VCS from attempting to bring resources online and
clears any internal wait states. You can then check resources for configuration
problems or underlying operating system configuration problems, and then attempt
to bring resources back online.----------------------------------------> when we configure any resource with wrong attibutes and tries to bring online

 

  • Answer to questions below:

     

    -->in this scenario  what would be the state of resource and service group.

    The resource state would be in a "waiting to online" state

    -->what they mean by suspended ? Is it s state of a resource group.

    Suspended means it appears the resources and service groups are doing nothing (i.e suspended) as the resource just sits in the same state "waiting to online" state until resource times out (or you flush service group).  Suspended it not a state.

    --> when we configure any resource with wrong attibutes and tries to bring online

    Yes - this is when it is normally used in in initial configuration where a resource has wrong attribute in VCS or the resource is configured wrong in the O/S and therefore you realise it is not going to online - if you don't flush, then you have to wait for the OnlineTimeOut period (by default 5 mins) before you can try online again and if you do wait the 5 mins and the resource is critical then the resource will fault and service group will fail to other system(s), so flushing group means you don't have to wait for resource to timeout and the resource will not fault.

    You use probe to force a monitor of a resource - by default VCS monitors online resources every 60 seconds and offline resources every 5 mins - so probing resource does a one-off monitor.  You can use this when testing killing a resource so you don't have to wait up to 60 seconds for VCS to take action or you can use this if you bring resource up outside of VCS control and you don't want to wait up to 5 mins for VCS to realise it is online.

    Mike

     

  • Thanks for the reply mike.That cleared things abt hagrp -flush.

    when we are setting attributes to resource  there would be few madatory attributes to be set  to bring the resource online.

    In GUI this can been known easily as we see the mandatory attributes would be BOLD.Accordingly we would be give values .if somebody  body  is configuring through CLI  how he knows abt  the mandatry attributes for any resource ?

    Thanks and regards,

    Siva

     

     

  • Answer to questions below:

     

    -->in this scenario  what would be the state of resource and service group.

    The resource state would be in a "waiting to online" state

    -->what they mean by suspended ? Is it s state of a resource group.

    Suspended means it appears the resources and service groups are doing nothing (i.e suspended) as the resource just sits in the same state "waiting to online" state until resource times out (or you flush service group).  Suspended it not a state.

    --> when we configure any resource with wrong attibutes and tries to bring online

    Yes - this is when it is normally used in in initial configuration where a resource has wrong attribute in VCS or the resource is configured wrong in the O/S and therefore you realise it is not going to online - if you don't flush, then you have to wait for the OnlineTimeOut period (by default 5 mins) before you can try online again and if you do wait the 5 mins and the resource is critical then the resource will fault and service group will fail to other system(s), so flushing group means you don't have to wait for resource to timeout and the resource will not fault.

    You use probe to force a monitor of a resource - by default VCS monitors online resources every 60 seconds and offline resources every 5 mins - so probing resource does a one-off monitor.  You can use this when testing killing a resource so you don't have to wait up to 60 seconds for VCS to take action or you can use this if you bring resource up outside of VCS control and you don't want to wait up to 5 mins for VCS to realise it is online.

    Mike