not exactly what you're looking for, but advice from the school of
hard knocks is to make sure you beat the piss out of the box prior to
the real cutover.
gathering stats for several days straight sounds like a good start.
there's nothing quite like (finally, after what seems like years)
running the import of your 8.1.7 app schemas into a new 10g db only to
have a RAID controller decide that its not worthy.
I have to grant this one some respect, as it was IOs on the system
datafile that threw the error this time. It didn't mess around with
temp, undo, redo, archived logs - it went straight for the system
my track record with Dell PERC hardware/drivers ain't so hot lately.
Februrary and March just aren't good months for PERC cards.
yep. still got an old Dell 410 Precision holding down the fort for
what a PE 2800 can't (or doesn't want to) do.
AD, MC: looks like development is still back in 8.1.7 until further notice.
0 for 3 in migrations this weekend.
On Sat, 5 Feb 2005 18:51:50 -0800, Good DBA wrote:
I am in the process of upgrading 8i (220.127.116.11) and 9i (18.104.22.168) databases to 10g (10.1.0.3) on UNIX.
I assume many of us would have tried and successfully completed the process.
I was also able to migrate some databases.. but with every database there is a new challenge (I am not complaning about it.. It is fun and a good source to learn).
In some cases there are multiple databases up and running on the box during the upgrade.
Any tips on shared pool during the upgrade as well as OS Shared memory ? (NOTE : I have already set these parameters to more than or equal to the suggested values)
How about ULIMIT ? Do you set it to ulimited for all the values ?
If you can share your experiences on upgrades would be highly appreciated.
Thanks and have fun.