Video Screencast Help
Scheduled Maintenance: Symantec Connect is scheduled to be down Saturday, April 19 from 10am to 2pm Pacific Standard Time (GMT: 5pm to 9pm) for server migration and upgrades.
Please accept our apologies in advance for any inconvenience this might cause.

hastart

Created: 08 Jul 2012 • Updated: 01 Aug 2012 | 4 comments
This issue has been solved. See solution.

I have a 5 node cluster.The cluster service (HAD) was stopped on all nodes due to some maintaiance issues.Now i need to start VCS on all NOdes.

 

Whatz the best way to do this.What r the points i need to take into consideration.

 

regards,

 

siva

Discussion Filed Under:

Comments 4 CommentsJump to latest comment

arangari's picture

confirm that GAB is configured and have membership for all the 5 nodes. This can be done by running '/sbin/gabconfig -a'. 

run 'hastart' command on all the 5 nodes individually, starting from the node where you feel you have latest and correct configuration. 

 

 

 

 

Thanks and Warm Regards,

Amit Rangari

If this post helped you resolving the issue, please mark it as solution. _____________________________________________________________________________

mikebounds's picture

Just to clarify futher, gabconfig should show you:

GAB Port Memberships
====================================
Port a gen 4a1c0001 membership 01234

This is saying that gab (port a) is loaded on nodes 0, 1, 2, 3 & 4 which means communication is working between all 5 nodes.

There is a possiblity that the VCS configuration (main.cf) is not the same on all nodes if configuration was not closed when VCS was shutdown - if you are not sure about this, then you should verify this by comparing using something like "cksum /etc/VRTSvcs/conf/conf/main.cf" to check all nodes have the same main.cf  If any are different then you will have to compare these (manually, or copy different files to one node and use diff or sdiff) to check which you want to start VCS with.  So then run "hastart" on the node that has the config you want (which can be any node if all configs are the same which is the norm).

When you run hastart on the other nodes, they read main.cf from memory of the first node, so there is no need to copy main.cf to other nodes, if one node does have a different config.

Service groups may have been frozen when VCS was shutdown, so when VCS has been started on all nodes run "hastatus -sum" to verify all service groups are "ONLINE" on at least one node and if any are shown as frozen, then you can use "hagrp -unfreeze -persistent grp_name"

Mike

UK Symantec Consultant in VCS, GCO, SF, VVR, VxAT on Solaris, AIX, HP-ux, Linux & Windows

If this post has helped you, please vote or mark as solution

SOLUTION
shiv124's picture

HI Mike and arngari ,

Thanks for the replies.

 

An extension question to undertand how vcs reads main.cf

 

Once the cluster is up and running.For example in 5 node cluster .ABCDE

 

If i am doing any changes in A if we are not dumping the changes to the file main.cf file.Then it would be in System A memory.how the changes will be known to B node .Is gab propagates this changes ? if it propagates how this works it would propagate and will keep this in B system Memory ?

 

 

 

mikebounds's picture

When you make a change in VCS it is written to all nodes simulateouls via GAB and this is held in memory on all nodes.  When you dump config (haconf -dump), each node writes its in-memory copy to disk.

Mike

UK Symantec Consultant in VCS, GCO, SF, VVR, VxAT on Solaris, AIX, HP-ux, Linux & Windows

If this post has helped you, please vote or mark as solution