By using this site, you agree to our updated Privacy Policy and our Terms of Use. Manage your Cookies Settings.
443,610 Members | 2,155 Online
Bytes IT Community
+ Ask a Question
Need help? Post your question and get tips & solutions from a community of 443,610 IT Pros & Developers. It's quick & easy.

Is it bad to see "High water mark for database heap = 4291373408" when the server has 4G memory?

P: n/a
When I did db2batch to benchmark a query, it performance details came
back with this reading ...

"High water mark for database heap = 4291373408"

Does this mean it would have caused memory thrashing at the server?
Did I over-allocate resources.

We just increased BuffPage to 16384 from its default of 1000. Pagesize
of the bufferpool was of course set to -1 so it would use the DBM cfg
parameter. But that is just 16MB of memory. Not even 10% of all the
total 4G memory.

Thanks in advance.

Bill
Nov 12 '05 #1
Share this Question
Share on Google+
3 Replies


P: n/a
After running your db2batch; execute the following command:
get snapshot for db on database_name

Check the database heap high water mark and see if it matches what you
see from db2batch. Ypu'll also see additional information about the heap
- ie. maximum size.

Philip Sherman

sql-db2-dba wrote:
When I did db2batch to benchmark a query, it performance details came
back with this reading ...

"High water mark for database heap = 4291373408"

Does this mean it would have caused memory thrashing at the server?
Did I over-allocate resources.

We just increased BuffPage to 16384 from its default of 1000. Pagesize
of the bufferpool was of course set to -1 so it would use the DBM cfg
parameter. But that is just 16MB of memory. Not even 10% of all the
total 4G memory.

Thanks in advance.

Bill


Nov 12 '05 #2

P: n/a
le****@aptea.com (sql-db2-dba) wrote in message news:<c8**************************@posting.google. com>...
When I did db2batch to benchmark a query, it performance details came
back with this reading ...

"High water mark for database heap = 4291373408"

Does this mean it would have caused memory thrashing at the server?
Did I over-allocate resources.

We just increased BuffPage to 16384 from its default of 1000. Pagesize
of the bufferpool was of course set to -1 so it would use the DBM cfg
parameter. But that is just 16MB of memory. Not even 10% of all the
total 4G memory.

Thanks in advance.

Bill


Since the buffpage is in pages, then 16384 x 4096 (assuming 4K page
size) is about 65 MB. If you page size is larger than 4K, then adjust
accordingly.

IBM seems to be discouraging the use of buffpage in version 8 (it is
hardly mentioned in the manuals anymore), and it is probably better to
alter the bufferpools sizes explicitly (use the immediate keyword to
take effect immediately).
Nov 12 '05 #3

P: n/a

"Mark" <m0****@yahoo.com> wrote in message
news:a5**************************@posting.google.c om...
le****@aptea.com (sql-db2-dba) wrote in message

news:<c8**************************@posting.google. com>...
When I did db2batch to benchmark a query, it performance details came
back with this reading ...

"High water mark for database heap = 4291373408"

Does this mean it would have caused memory thrashing at the server?
Did I over-allocate resources.

We just increased BuffPage to 16384 from its default of 1000. Pagesize
of the bufferpool was of course set to -1 so it would use the DBM cfg
parameter. But that is just 16MB of memory. Not even 10% of all the
total 4G memory.

Thanks in advance.

Bill


Since the buffpage is in pages, then 16384 x 4096 (assuming 4K page
size) is about 65 MB. If you page size is larger than 4K, then adjust
accordingly.

IBM seems to be discouraging the use of buffpage in version 8 (it is
hardly mentioned in the manuals anymore), and it is probably better to
alter the bufferpools sizes explicitly (use the immediate keyword to
take effect immediately).


Another reason why IBM is discouraging the use of BUFFPAGE is that it's
nearly useless (and very confusing) on systems that have more than one
bufferpool, as you then have two places to check/change bufferpool sizes --
the system catalogs for bufferpools with explicit sizes, and then BUFFPAGE
for those with implicit sizes (catalogs have npages set to -1).

--
Matt Emmerton
Nov 12 '05 #4

This discussion thread is closed

Replies have been disabled for this discussion.