Video Screencast Help
Symantec to Separate Into Two Focused, Industry-Leading Technology Companies. Learn more.

mechanism

Created: 10 Jul 2014 • Updated: 11 Jul 2014 | 1 comment
This issue has been solved. See solution.

hi,

i like to verify things.

Over private network,public network(if is setup the llt traffic) or disk hearbeat go heart beat and gab traffic(when a resource is modifed,added).

Let`s say we have 2 nodes vcs and the service group failover from noda a to node b.What happened and what happens?The agent of faulty resource put the resource offline,HAD interogate this agent and saw the resource offline and is a critical resource,then had start to put the group offline on this node.So at every resource put offline,had will log in engine_A.log and will send out through gab the information that a certain resource is offline on node a.On node b the gab will receive this and will inform had about it.This had will modify it in RAM and will log in engine_A.log.right?

The same steps to all the resources within this group and with group itself.

On node b then had will start the group based on information from main.cf.So every agent will bring online the appropiate resource and had will write into engine_A.log.Example,the diskgroup agent will import the disk group and will start the volumes,then monitor this disk group.right?

Every time when an resource is added or modified on an node,had on that node will write this information on main.cf and log it,then will send out using gab to the other vcs nodes.right?

Please i like to verify also when i should to use gabconfig -c -x.Is this when i have 3 nodes cluster and 1 node goes down,or 2 nodes clusters and 1 node goes down?

thanks so much.

Operating Systems:

Comments 1 CommentJump to latest comment

Gaurav Sangamnerkar's picture

First thing, most of these understandings can be built from reading admin guides... I would suggest again to have a thorough reading of guides

Over private network,public network(if is setup the llt traffic) or disk hearbeat go heart beat and gab traffic(when a resource is modifed,added).

>> LLT is the first priority to send any traffic. Rest are all additional priorities.

Let`s say we have 2 nodes vcs and the service group failover from noda a to node b.What happened and what happens?The agent of faulty resource put the resource offline,HAD interogate this agent and saw the resource offline and is a critical resource,then had start to put the group offline on this node.So at every resource put offline,had will log in engine_A.log and will send out through gab the information that a certain resource is offline on node a.On node b the gab will receive this and will inform had about it.This had will modify it in RAM and will log in engine_A.log.right?

The same steps to all the resources within this group and with group itself.

>>>  pretty much right, HAD will keep logging in engine_A.log for each resource (not once for entire group). & engine_A.log will be updated at same time on both the nodes, so two nodes will keep talking via LLT constantly. So rather than sequential its a parallel process of updating logs at both nodes.

On node b then had will start the group based on information from main.cf.So every agent will bring online the appropiate resource and had will write into engine_A.log.Example,the diskgroup agent will import the disk group and will start the volumes,then monitor this disk group.right?

>> >Correct

Every time when an resource is added or modified on an node,had on that node will write this information on main.cf and log it,then will send out using gab to the other vcs nodes.right?

>> If resources are added/modified using ha commands, kernel copy is updated, once you run "haconf -dump" ,kernel copy is dumped to disk copy on all nodes in cluster

Please i like to verify also when i should to use gabconfig -c -x.Is this when i have 3 nodes cluster and 1 node goes down,or 2 nodes clusters and 1 node goes down?

>> No, you wil use gabconfig -c -x when you want to manually seed cluster (regardless of how many nodes in cluster). So for e.g you have 3 node cluster & all 3 nodes are rebooted. Once 1st node come up, it is not going to start VCS because gab will keep waiting for other nodes to join membership. In this case if you want VCS to start, you may use gabconfig -c -x  to force seed GAB & start VCS.

G

PS: If you are happy with the answer provided, please mark the post as solution. You can do so by clicking link "Mark as Solution" below the answer provided.
 

SOLUTION