Re: Problems with "ALTER FRAGMENT ON TABLE DETACH..."

---- you wrote: > Save me, please ! > > We are looking for options to problems we've had detaching fragments on > very large tables, where the dbspace has a lot of rows in it.. > > Background: hpux 10.20 HP t500 with Informix IDS 7.30.uc6, table has 24 > fragments each 6-7gb of data in 8gb dbspace. > > On a previous release (7.30.uc2) the detach would hang and never > finish. While > hanging it was not accumulating any cpu use. We are able to kill the > session > with onmode -z without any problems. > > Informix says, that's a bug in uc2, fixed in uc6. I still see the same > hang. > We don't want to go go to another release wight away (e.g. 7.31) for > fear of > something else breaking... We are in serious 24x7 production. > > We ARE able to delete the rows from the dbspace/fragment and issue > "alter fragment blabla drop frag.." and that works almost instantly. > > Q. Can I use the alter fragment DROP with rows in the dbspace ? > Q. Will a special update statistics be necessary before or after ? > Q. Are there any know tricks to getting DETACH to work ? These are such > > huge tables (8gb fragments) and we peel off one a month we have to > figure out what to do.... Dude I suspect you need to look at your index fragment strategy. It probably needs to be the same as the table for an in-place alter to work on a detach. Not sure why it would hang, but I have seen a lot of space related issues with detach. AB ---------------------------------------------------------------- Get your free email from AltaVista at http://altavista.iname.com