System tablespace specifically though.
Jacob Miller, October 08, 2016 - 12:18 am UTC
Connor,
Thanks so much for your reply, it is greatly appreciated, but I have a little bit of followup.
I understand the FET$/UET$ bottleneck conceptually at least, but the counterargument I'm receiving from the previous DBA is that all the tablespaces in these databases are already locally managed, except SYSTEM. He's claiming that since the addition of SYSAUX, the SYSTEM tablespace is essentially static except during an upgrade so there wouldn't be enough benefit to justify the migration.
Which processes could I point out to an Oracle 8/9 era DBA and say, "Whenever the database does X, Y, and Z, we're running into the FET$/UET$ bottleneck in the SYSTEM tablespace." ?
With regards to the second question: in some of the older posts, people argued against using DBMS_SPACE_ADMIN because the extent size remains in 'USER' mode and thus the benefits are not fully realized. Is this no longer a relevant objection?
I would be happy to take exp/imp and transportable tablespaces off the table, as I agree it is a sledgehammer and not something I was looking forward to.
Thanks again for your time,
Jacob Miller
October 08, 2016 - 1:25 am UTC
I think that you cant have a read-only standby unless all the tablespaces are LMT. Its been so long since I've tinkered with DMT, so I cant be 100% sure on that.
"because the extent size remains in 'USER' mode and thus the benefits are not fully realized" is fiction.
If there is so much resistance to it, then just schedule it for your next major upgrade.
Most battles in IT are political not technical :-)