17.10 Forming a Cluster without Enough Votes


17.10 Forming a Cluster without Enough Votes

What if your cluster is down and you are unable to boot enough members to form a cluster? If you have insufficient votes to form a cluster, you can boot the systems interactively and set cluster_expected_votes to a value that will allow the cluster to form.

If we consider our six-member cluster configuration again, let's assume that all the members came down but only three can be rebooted. Since we can boot three systems, we should set the cluster_expected_votes value to 3.

Boot each member as follows:

 >>> boot –fl ai (boot dkb100.1.0.3.2 -flags ai) block 0 of dke100.1.0.3.2 is a valid boot block reading 19 blocks from dke100.1.0.3.2 bootstrap code read in Building FRU table base = 200000, image_start = 0, image_bytes = 2600 initializing HWRPB at 2000 initializing page table at 1f2000 initializing machine state setting affinity to the primary CPU jumping to bootstrap code UNIX boot - Wednesday August 01, 2001 Enter: <kernel_name> [option_1 ... option_n]  or: ls [name]['help'] or: 'quit' to return to console Press Return to boot 'vmunix' # 

At the prompt enter the following:

 # vmunix clubase:cluster_expected_votes=3 

Once two of the members configure the Connection Manager, the cluster will have 2 votes, which is sufficient for quorum and the cluster will form. As with the example in section 17.8.5, the change will not survive a reboot.

To make a more permanent change, once you can login, use the clu_quorum command to adjust the votes on the members that will be unavailable for an extended period of time using the "-m" option or adjust the cluster's expected votes value using the "-e" option. And remember, do not edit a member's sysconfigtab file manually!

17.10.1 Forming a Cluster without Enough Votes and/or a Failed Quorum Disk

If we modify our previous scenario to include a quorum disk that failed, can we boot interactively by simply using the cluster_expected_votes value alone? Yes and no. While you can boot a cluster with a failed quorum disk, the quorum disk thread will attempt to retry the disk for some time before finally giving up. Prior to V5.1A, this would delay the boot process. As an alternative, at the interactive boot prompt, add a second attribute, cluster_qdisk_votes.

 # vmunix clubase:cluster_expected_votes=3 clubase:cluster_qdisk_votes=0 

Another option would be to add the cluster_qdisk_major and cluster_qdisk_minor attributes to the interactive boot prompt so that the cnx will not load the qdisk attributes at all.

 # vmunix clubase:cluster_expected_votes=2 clubase:cluster_qdisk_votes=0 clubase: cluster_qdisk_major=0 clubase:cluster_qdisk_minor=0 

Caution

The interactive boot prompt line must be one continuous line. Do not put in a "\" to continue onto the next line; it will not work. Also, be extremely careful to type everything correctly. Typos are generally ignored by the subsystem. This, unfortunately, will not give you the desired effect since you will not have successfully overridden the attribute value(s).




TruCluster Server Handbook
TruCluster Server Handbook (HP Technologies)
ISBN: 1555582591
EAN: 2147483647
Year: 2005
Pages: 273

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net