large indexes

When building an index, is there a way to following its progress in detai? Periodically we must build indexes on very large tables (+3.5 gigs, fragmented). This takes hours. Sometimes it takes so long, that we become suspicious and break the process before the index is completed. However, when we do this we are in the dark. We don't know if something is "wrong", or if the index is 90% done. oncheck -pe gives us some information about the dbspace in qestion (oh, I forget to mention that the indexes are always built in their own dedicated dbspaces), but its obviouly not up-to-date or complete. Got any suggestions? Also, it appears that while an index is being built on a table the table is exclusively locked and therefore unavailable to the users. Is this just a fact of life or does anyone (a) know a way aroundt this or (b) know how to speed up the buidling of indexes. I'll throw in a last question. Is it worth fragmenting indexes, and if so, under what circumstances? By fragmenting I mean the same as fragmenting data. I don't mean letting the index be fragmented with the data, but instead into its own dedicated fragments. Oh some background information. The tables in questions have a row size of appromixately 650 bytes (please don't ask me why the table is so large), and 35 million rows. Each table has four indexes ranging from 8 bytes (excluing overhead) to 29 bytes (excluding overhead). Thanks for any help!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! -------------------==== Posted via Deja News ====----------------------- http://www.dejanews.com/ Search, Read, Post to Usenet