Hello Thomas,
Implementing a buffering system would be good as well. I am not sure
what impact if any it would have on the reliability of the data, (i.e.
would it have more potential for data corruption in case of power
failures, etc?) Generally speaking flash is a lot more efficient at
writing a small number of large blocks than writing a large number of
small blocks. I did not look at the "fs" package yet, so I can not
comment on the proposed solution. Also the recovery step is
particularly problematic because it tends to have a large number of
write operations concentrated in one place. On top of that, for our
application the delete is an issue because we tend to delete the data
that is "timing out" i.e. "delete from table where timestamp < X" this
operation typically deletes a large number of contiguous blocks.
I'll try to poke around the "fs" package later in week when I have
some time.
Alex
On Oct 26, 2:31 pm, "Thomas Mueller" <[EMAIL PROTECTED]>
wrote:
> Hi,
>
> > This does not seem to be the case though...
>
> You are right, thanks for the test case! Obviously I forgot why the
> code is like that. I should have documented it better... I think I
> remember now why all delete entries need to be written first:>
> Because delete entries are always 1 block, while not-deleted entries
> can be many blocks. So for example:
> (A) block: 1 (delete)
> (B) block: 2 (delete)
> (C) block: 1 ('Hello', 2 blocks long - so actually block 1+2)
> Needs to be written in this order and not (A) (C) (B). I will document
> this description.
>
> I know my solution is not optimal, but I thought recovery is not a big
> performance problem. Looks like it is a problem for flash drives...
>
> But still I believe implementing a buffering file system is probably a
> better solution because it will improve performance for all
> operations, not just for recovery. What do you think?
>
> Regards,
> Thomas
--~--~---------~--~----~------------~-------~--~----~
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
-~----------~----~----~----~------~----~------~--~---