OK, there is at least one aspect of the configuration which I am absolutely sure of to be wrong:
> nodea # grep shared /etc/vfstab > /dev/did/dsk/d5s0 /dev/did/rdsk/d5s0 /shared ufs 5 yes global,logging For the block and special devices, /dev/global/... must be used, not /dev/did, otherwise the cluster services using /shared do not survive a panic of the node mastering the device: df: cannot statvfs /global/.devices/node at 1: I/O error df: cannot statvfs /shared: I/O error But the remaining question is if anyone thinks it is OK to use directly a disk which is used in a metaset. Thanks, Nils
