I've recently built a DB2 system to run Content Manager 8.3 under some pressure from unscrupulous project manager and crazy customers, some of your may know the feeling...
It's currently using
DB2 18.104.22.1684 (I plan to upgrade the whole solution soon so it is supported). I'm concerned about DB2 performance as volumes grow.
Windows 2003 SP2 Ent.
Content Manager 8.3a.
Is it safe to depend on the Maintenance Wizards / Health Monitor recommendation to cover all requirements for typical smooth running?
If I need to implement a bunch of standard maintenance jobs maybe someone can point me in the right direction? The wealth of information is a bit overwhelming and i have the added problem of figuring out if it's safe to a Content Manager system.
First of all DB2 V8 ran out of support at Apr 30th 2009.
In Version 9.5 you can define Policies to Backup, reorg and runstats the database. Also most of the Memory Parameters could be automated. In that case DB2 is changing the Parameterss dynamically. Search for Self Tuning Memory Maneger for details.
I had enabled RunStats but not ReOrg as an example of my inexperience. I'm now worried about applying these setting retrospectively on a system which is now live (perhaps the first one with take an especially long time, etc). I can read up on the subjects in the link.
I think 8.x is missing a lot of the more recent tools in 9.5. The Memory Visualiser is interesting but a step short for helping me automatically manage the memory, for now I'll try and keep the system up and tackle the memory related error when they occur.
usually you don't need any reorg at the productive start, so don't worry. Reorgs are needed if you have a lot of inserts, updates and deletes on your tables. There are two methods for the reorgs online and offline. The online reorgs are done with reorg ... inplace... (Caution! If you are using archive logging it will generate a lot of Transactionlogs. Roughly three times the size of the table), but the people can work with the tables. During an offline reorg the table is blocked (Not all the time, but in general).
You can check your bufferpool and Memory settings with a
get snapshot for database on <DBNAME>
It shows eg. teh number of total sorts and sort overflows and the hitratio on the bufferpools
Buffer pool data logical reads = 39926246415
Buffer pool data physical reads = 27984338
(Many Hits on Bufferpool; only few physical reads)
Same from Indexes and Temp
(The Information is the sum over all bufferpools. You can check single BP with get snapshot for bufferpools on <DBNAME>)