Solaris Cluster SVM reconfiguration fails in svmstep1, ucmmd died, cluster paniced (Doc ID 1333117.1)

Last updated on APRIL 19, 2013

Applies to:

Solaris Cluster - Version 3.1 to 3.3 [Release 3.1 to 3.3]
Information in this document applies to any platform.

Symptoms

Solaris Cluster reconfiguration fails in svmstep1. SVM (Solaris Volume Manager) with multi-owner sets has to reconfigure when Solaris Cluster reconfigures.  You may see the following error in /var/adm/messages

Jun  8 16:39:47 host1 metaclust: [ID 288242 user.info] Starting Step1:    0:000:000
Jun  8 16:39:47 host1 metaclust: [ID 913294 user.info] Step1 aborted:clear locks failed    0:020:412
Jun  8 16:39:47 host1 metaclust: [ID 410844 user.error] exiting with 1
Jun  8 16:39:47 host1 Cluster.RGM.global.rgmd: [ID 784560 daemon.notice] resource rac_svm_rs status on node kronwall change to R_FM_FAULTED
Jun  8 16:39:47 host1 Cluster.RGM.global.rgmd: [ID 922363 daemon.notice] resource rac_svm_rs status msg on node kronwall change to <Error in step svmstep1>

After this metaclust will abort:

Jun  8 16:39:51 host1 metaclust: [ID 355579 user.info] Starting Abort step:    0:000:000
Jun  8 16:39:51 host1 metaclust: [ID 708556 user.info] FATAL, can not create message ID


Metaclust abort causes ucmmd to exit which will panic the cluster:

Jun  8 16:39:54 host1 cl_runtime: [ID 367077 kern.warning] WARNING: Failfast: timeout - unit "failfast_now"global".
Jun  8 16:39:56 host1 cl_dlpitrans: [ID 624622 kern.notice] Notifying cluster that this node is panicking



Cause

Sign In with your My Oracle Support account

Don't have a My Oracle Support account? Click to get started

My Oracle Support provides customers with access to over a
Million Knowledge Articles and hundreds of Community platforms