Shared Page Cache is coming

One of the biggest tasks for the long-awaited Firebird 3.0 release is the Shared Page Cache feature. With this feature, Firebird will finally be able to fully benefit of SMP machines without having to work with individual cache for every connection.

Vlad Khorsun had been busy working on this, and judging by recent commits to SoundForge, the task is almost finished. Part of the code came from Vulcan. He hopes that people will help doing lots of tests when the first alpha release appears.

Note: Up to now, Shared Cache was only available in the SuperServer model, but SS is not optimized for SMP.

Firebird on Linux and ICU > 4.2

Hi,

there was a bug in Firebird that did not handle well icu > 4.2 see (http://tracker.firebirdsql.org/browse/CORE-3447)

2.1.4 and 2.5.0 packages in distributions were affected (Debian, Fedora, Mandriva, Mageia) because they are using icu 4.4

Adriano fixed it and in these various distro updates are on the way

I know that some of you are using Centos 5.
The problem with Centos 5, is that you can’t have Firebird 2.5 from EPEL
repository. Centos 6 will give you the possibility to have Firebird 2.5 from EPEL6 repository, but seems that Centos6 take a long time to come.
https://www.centos.org/modules/newbb/viewtopic.php?topic_id=25878
http://wiki.centos.org/QaWiki/6/AuditStatus

You can try to use Scientific Linux instead (http://www.scientificlinux.org)
As Centos, the base SL distribution is basically Enterprise Linux
(RHEL), recompiled from source, and Scientific Linux 6.0 was released on
March 3, 2011

Philippe Makowski
http://www.ibphoenix.com

Firebird used in Citrix EdgeSight Database Agent Server

Quote from the Citrix blog also here is one KB article about Firebird configuration , of course there are more :
“I usually don’t get too excited about the release of a KB article, but this one took quite a bit of time and I’m very pleased we got this critical info back out there. If you’ve ever tried to implement EdgeSight or AppStreaming without excluding the Firebird database or RadeCache folders, you know how important these guidelines are.”

If you don’t know who Citrix is there is always wikipedia to help

Change made to make Faster restores with gbak

There is a simple change which makes restores much faster, tested with Linux 2.5 embedded restore
Two databases, with two different tables filled with 2,000,000 records:
create table t1 (n1 integer);
create table t10 (n1 integer, n2 integer, n3 integer, …, n10 integer);

----------------------
Restore timings:
T1: 14.5s
T10: 20.5s

Improved gbak timings:
T1: 9.7s
T10: 14.7s
----------------------

The good news is there is still room to improve the restore speeds , here is Adriano’s tweet:

I do have a slight different method, to insert 100K records in less than 1s over the same localhost method.

Also confirmed by the ending of the article:

PS: There is other restore improvement capable of take down time from
around 8.5s to 0.5s over TCP, but it still requires some analisys.

1 23 24 25 26 27 84