We define on the dependency tree the disk groups and we have dual paths to the SAN. Since there is usually a 2-min fail over for VxDMP ? How does this affect VCS when it probes for monitoring of the resource ? I understand during the duration Veritas detects and fails to the second path, the I/O's are essentially frozen ?
Multipath is managed by DMP, if any one of the HBA fails DMP will automatically to another HBA, which will not affect your cluster operation unless both the HBA goes down & SAN become inaccessible.
Thanks for the response. One more question, how should the bp.cong file ( for Veritas Netbackup) be configured on the cluster servers so that there is seamless backup/restore process. I do not want to setup NB_SERV parameters, but rather pick them up from the bp.conf file. Also should we be using the REQUIRED-INTERFACE parameter setup in bp.conf. Any thoughts , suggestions
