cancel
Showing results for 
Search instead for 
Did you mean: 

Regarding resource online operations

allaboutunix
Level 6

Hi,

 

Suppose, a resouce got faulted in a SG and I need to make it online.

Shall I have to do in this way?

 

hagrp -clear <group> [-sys] <host> <sys>

and then online the service group

hagrp -online <group> -sys <sys>

        OR

 

by this way?

1.Flush the SG,

hagrp -flush <group> -sys <system>

2.Clear the faulted resource

hares -clear <resource> [-sys]

3.Online it,

hares -online <resource> [-sys]

Do flushing of the SG required?

Kindly assist how should i have to proceed for these cases.

 

 

3 ACCEPTED SOLUTIONS

Accepted Solutions

Sunil_Yadav
Level 4
Employee

Hi,

Corrective course of action depends upon the resources’ configuration, dependency, and current state of resources/SG.

1.    If the faulted resource is critical (Resource::Critical=1) resource, entire service group’s failover is initiated. Even if faulted resource is non-critical(Resource::Critical=0) resource, all the dependent parent resources are brought offline.  In these cases:
    a.    First we have to interrupt the failover/offline of the service group
# hagrp -flush [-force] <group> -sys <system>
    b.    Then we have to clear the fault of faulted resource(s)
# hagrp -clear <group> [-sys <system>]
OR
# hares -clear <res> [-sys <system>]

“hagrp –clear” will clear fault of all the faulted resources in service group. This can be handy if multiple resources in SG are faulted.
    c.    Thereafter we can initiate online of resource
# hagrp –online <group>
# hares –online <res>

Please refer to hagrp/hares man pages for all possible options for onlining. ‘hagrp –online’ takes care of resources’ dependency and onlines all the resources in SG. ‘hares –online’ is best suited for individual resources’ onlining.

 

2.    If the faulted resource is standalone (no dependency) and non-critical resource, SG’s offline/failover is not initiated. Thus, flushing is not required. Just clear the fault(hagrp/hares –clear) and initiate online(hagrp/hares –online).

 

Thanks & Regards,
Sunil Y

View solution in original post

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

You flush a Service group when you want to stop the service group from trying to go online or stop the offline process while VCS is busy taking it down.
You would normally do this when the online or offline of SG seems to be hanging.

See this explanation: 
https://sort.symantec.com/public/documents/sfha/6.0.1/aix/productguides/html/vcs_admin/ch06s12s09.ht... 

So - flushing is not required.

As per above excellent post, manual or automatic corrective action depends on Service Group and Resource attributes.

See all the links under this heading:

https://sort.symantec.com/public/documents/vcs/6.1/aix/productguides/html/vcs_admin/ch11.htm

Controlling VCS behavior

This section includes the following topics:

View solution in original post

starflyfly
Level 6
Employee Accredited Certified

Hi, allaboutunix

 

If resource in fault state, hagrp/hares clear is needed.

if sg in online/offline process, long time not finish, or in  waiting online/offline state,  flush sg is a good choice.

View solution in original post

3 REPLIES 3

Sunil_Yadav
Level 4
Employee

Hi,

Corrective course of action depends upon the resources’ configuration, dependency, and current state of resources/SG.

1.    If the faulted resource is critical (Resource::Critical=1) resource, entire service group’s failover is initiated. Even if faulted resource is non-critical(Resource::Critical=0) resource, all the dependent parent resources are brought offline.  In these cases:
    a.    First we have to interrupt the failover/offline of the service group
# hagrp -flush [-force] <group> -sys <system>
    b.    Then we have to clear the fault of faulted resource(s)
# hagrp -clear <group> [-sys <system>]
OR
# hares -clear <res> [-sys <system>]

“hagrp –clear” will clear fault of all the faulted resources in service group. This can be handy if multiple resources in SG are faulted.
    c.    Thereafter we can initiate online of resource
# hagrp –online <group>
# hares –online <res>

Please refer to hagrp/hares man pages for all possible options for onlining. ‘hagrp –online’ takes care of resources’ dependency and onlines all the resources in SG. ‘hares –online’ is best suited for individual resources’ onlining.

 

2.    If the faulted resource is standalone (no dependency) and non-critical resource, SG’s offline/failover is not initiated. Thus, flushing is not required. Just clear the fault(hagrp/hares –clear) and initiate online(hagrp/hares –online).

 

Thanks & Regards,
Sunil Y

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

You flush a Service group when you want to stop the service group from trying to go online or stop the offline process while VCS is busy taking it down.
You would normally do this when the online or offline of SG seems to be hanging.

See this explanation: 
https://sort.symantec.com/public/documents/sfha/6.0.1/aix/productguides/html/vcs_admin/ch06s12s09.ht... 

So - flushing is not required.

As per above excellent post, manual or automatic corrective action depends on Service Group and Resource attributes.

See all the links under this heading:

https://sort.symantec.com/public/documents/vcs/6.1/aix/productguides/html/vcs_admin/ch11.htm

Controlling VCS behavior

This section includes the following topics:

starflyfly
Level 6
Employee Accredited Certified

Hi, allaboutunix

 

If resource in fault state, hagrp/hares clear is needed.

if sg in online/offline process, long time not finish, or in  waiting online/offline state,  flush sg is a good choice.