Don't put production data on a non-redundant array. If you are joining two disk slices into one big partition then there is no redundancy. Might think about hiring a sysadmin or two.
you have to mirror the whole disk... if the first disk fails, you want to be able to work with the second disk... won't you? and without swap and /var the second disk will not work!
if you have any sense, you will also mirror swap and /var. Most decent SAs would also put the user home accounts on a separate slice.
you should copy the disk label from one disk to the other and then mirror the partitions.
Also, assuming that your are using SVM (ex SDS, ex-ex-ODS), you should ensure that you have a enough state databases.
The command metadb should show you where they are. You should have then in at least two locations and convention when I worked at Sun, was to put three copies in each location.