Dirty page queues during a checkpoint (WAS: The old raw devic

Dirty page queues during a checkpoint (WAS: The old raw devic

Post by Bill Dar » Sat, 17 Apr 2004 22:22:26



http://www.yqcomputer.com/

Page 7-44 for chunk writes, which are the type of writes that are done
during page flushing.

Regards,
Bill


sending to informix-list
 
 
 

1. Page splits/ Dirty pages/ Checkpoint

2. VARCHAR column index -- dirty pages -- checkpoint

Hi,
I have a problem with index pages on a varchar column in my database.

The varchar field is an Identifier which is not unique. The field length is 15
DataId varchar (15)

I have an index on that column. When we are inserting data in the table, we
do lot of transactions per second (~2000), after a while since we begin
inserting the data, the dirty pages for the index on this varchar column is
much more(~2000) than anything else. That is driving the checkpoint to take
longer and therefore reducing our systems capacity(tps).

What is the best way to handle this situation? I need to reduce the dirty
pages so that the check point takes less time flushing the index pages(along
with other pages). The data length we typically insert averages about 3-5
chars. But sometimes we do insert data with 15 characters. Is varchar the
right data type or the char or some other? Tried with fixed length char(15)
also but the dirty pages remain the same.

Looking for some suggestions/opinions. Your input is greatly appreciated.

Thanks in advance.

3. committing dirty pages, recovery interval, checkpoint, etc.

4. [RFC] ext4: Don't send extra barrier during fsync if there are no dirty pages.

5. Why does Word say I am on Page 11 when I am only on Page 1?

6. How do I remove old, old, old stock prices from investment summary on page 1

7. How do I remove old, old, old stock prices from investment summary on page 1??

8. new page won't overwrite old one...what am I doing wrong?

9. question about dirty buffers in page cache pages

10. tx queue start entry x dirty entry y (was 8139too PCI IRQ issues)

11. [PATCH 6/6] vmscan: Kick flusher threads to clean pages when reclaim is encountering dirty pages

12. [PATCH 0/4] [RFC][PATCH] fs-writeback: redefining the dirty inode queues

13. Initializing log/checkpoint information....-> fails during oninit -isvy

14. I love swallowing Cum, i am believably dirty, so I smell you, Greek Large Horndog.

15. Feel me, for I am Screech --- Master of the Dirty Sanchez