I'm not sure, but I think you are correct, and you could change it to
only overwrite the first 16 bytes.
It would have security concerns, because it wouldn't erase old data.
But more importantly, it probably wouldn't buy you much in terms of IO
because the filesystem is probably writing in at least 2K chunks.
But it might be worth a try. If you have a filesystem formatted to 512
byte sectors, and you crank up the pagesize of your H2 database to 16K,
you may well gain some speed.
Try it by making the change and running the unit tests - if there is a
problem, the unit tests will catch it.
On 2012-04-05 16:01, Steve McLeod wrote:
Ah, I see.
However, the code for checkpoint() seems to zero not just the header
but the entire page:
byte[] empty = new byte[pageSize];
file.write(empty, 0, pageSize);
This is writing <pageSize> bytes, right?
Regards,
Steve
On Wednesday, 4 April 2012 20:06:57 UTC+2, Thomas Mueller wrote:
Hi,
Yes, a larger page size should help.
The header of the freed pages is zeroed out so that in case of a
crash, the free pages can be easily re-used.
It's unlikely the checkpoint can be easily improved I'm afraid.
I'm working on a new storage engine, but it will take a while
until this is available.
Regards,
Thomas
On Tuesday, April 3, 2012, Noel Grandin wrote:
I can't be sure, but it looks like it's zero-ing out the
header of a page, so I would assume that it's a little necessary.
you could try increasing the page size of your database, that
might speed up this operation.
On 2012-04-02 17:15, Steve McLeod wrote:
Hi Thomas,
When running TRUNCATE TABLE on a large table, it is
time-consuming. I've done some profiling of the H2 source,
and found that a step in PageStore.checkpoint() that
consumes one-third of the time needed for TRUNCATE TABLE.
In my specific case it uses 69 seconds out of seconds for
the whole TRUNCATE execution.
This line is the one consuming the time:
file.readFully(test, 0, 16);
which is org.h2.store.PageStore.java: line 451 in the
current SVN trunk.
It is part of a process of zeroing freed pages on disk.
I'm wondering if zeroing is an important part of the
workings of the PageStore. Or is it for security reasons?
If so, is this step something I could potentially make
optional by introducing a "INSECURE_FREE" database option?
I'm running on Mac OS X Lion, using an SSD drive.
Regards,
Steve McLeod
--
You received this message because you are subscribed to
the Google Groups "H2 Database" group.
To view this discussion on the web visit
https://groups.google.com/d/msg/h2-database/-/KpuvExBPO2sJ
<https://groups.google.com/d/msg/h2-database/-/KpuvExBPO2sJ>.
To post to this group, send email to
[email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/h2-database?hl=en
<http://groups.google.com/group/h2-database?hl=en>.
--
You received this message because you are subscribed to the
Google Groups "H2 Database" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/h2-database?hl=en
<http://groups.google.com/group/h2-database?hl=en>.
--
You received this message because you are subscribed to the Google
Groups "H2 Database" group.
To view this discussion on the web visit
https://groups.google.com/d/msg/h2-database/-/Z2ZMtLk5qWsJ.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/h2-database?hl=en.
--
You received this message because you are subscribed to the Google Groups "H2
Database" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/h2-database?hl=en.
