headers
Hammond, Steve | 21 Nov 20:05
Favicon

RE: [Studio] Discussion about the release model

I wait until the final release and don't get the RC's.

I am not worried about the number of RC's.  Too man RC's and it
indicates insufficient unit testing, but 1, 2 or 3 seems reasonable.

On other projects I do get RC's, but only when I need the new features
and want an early preview.  Or projects that are central to my
development world that I will be lost if the features I use do not work
in the next release, or if I think I use the product in an unusual way
that I want to be sure still works in the next release.

-----Original Message-----
From: Stefan Seelmann [mailto:seelmann@...] 
Sent: Tuesday, November 18, 2008 9:52 AM
To: directory apache; users@...
Subject: [Studio] Discussion about the release model

Hi Users and Dev,

Back in January 2008 we discussed in [1] about the release model of
Apache Directory Studio, whether we should have release candidates or
not. At that time we decided to introduce release candidates.

Now, almost one year later I think we should reconsider this decision.
Pierre-Arnaud and I discussed about the release model.

In 2008 we released the following versions of Studio:
1.1.0 RC1
1.1.0 RC2
1.1.0
(Continue reading)

Permalink | Reply | 0 comments |
headers
Emmanuel Lecharny | 20 Nov 15:24
Gravatar

Backend recovery

Hi guys,

as I'm currently working on a recovery tool, I have reviewed the current 
configuration we have. Let's first analyse how a recovery tool is good 
for and how it should work.

1) Usage
We need to rebuild the indices if the database is corrupted. This can 
occur in many occasions, mainly when the server is brutally interupted 
(power shutdown, for instance). As indices are dynamically built while 
injecting new entries (or when updating those entries), it is important 
to have such a tool to rebuild them assuming the indices are not up to 
date, as otherwise we may have orphans entries, or even worst, indices 
pointing on non existing entries or wrong entries.

Another usage for such aa tool would be to create indices off line. This 
has the great advantage to allow a mass injection of entries into the 
master table (the table containing all the entries), then do a global 
re-indexing, potentially avoiding a lot of expensive controls, as we may 
run a pre-check on data before doing the injection. (of course, if we 
leverage the number of checks to do, we can go faster or slower, 
depending on how much we trust the data we inject into the server)

2) How it should work
This is what we have a real problem. Assuming that the backend storage 
can be totally FU, we can't really trust it to rebuild the indices. What 
are our options ? Let's see how the system works when we modify some 
entries :
- the database is in a S0 state at t0 (let's say when we have started 
the server the very first time)
(Continue reading)

Permalink | Reply | 0 comments |
headers
Felix Knecht | 18 Nov 10:52
Favicon

Re: Javadoc on web site

Juergen Weber schrieb:
> Thanks for the Javadocs links. What about linking to them from the DS web
> page?
>   
Thanks for the hint.
> The Xref pages are nice, too. But I did not see links from the API doc (i.e.
> from individual classes' docs) into the Xref.
>   

It's done now.
BTW: These links do not link to svn repository sources but just to their
own created html page showing the sources. I think that's also what
geronimo does.
>
> Felix Knecht-2 wrote:
>   
>> Juergen Weber schrieb:
>>     
>>> And really cool were links from Javadocs to the corresponding source file
>>> in
>>> the subversion trunk...
>>>
>>>       
>> Here you go:
>> API: http://vm094.oxylos.org/projects/apacheds/apidocs/index.html
>> Source Xref: http://vm094.oxylos.org/projects/apacheds/xref/index.html
>>
>> or all the generated docs from trunk (section Project Reports contains the
>> ones above).
>> http://vm094.oxylos.org/projects/apacheds/
(Continue reading)

Permalink | Reply | 0 comments |
headers
Felix Knecht | 18 Nov 10:51
Favicon

Re: [CONF] Apache Directory Server v1.5: Navigation (page edited)

Can please somebody with more rights (or the knwoledge how to do it)
update the changed navigation pages for ADSv1.5 [1] and Studio [2]?
I added in the documentation part a direct link to the site with the
generated maven reports.

Thanks
Felix

[1] http://cwiki.apache.org/confluence/display/DIRxSRVx11/Navigation
[2] http://cwiki.apache.org/confluence/display/DIRxSTUDIO/Navigation
Permalink | Reply | 1 comment |
headers
Emmanuel Lecharny (JIRA | 17 Nov 18:55
Favicon
Gravatar

[jira] Created: (DIRSERVER-1290) [Perf]Potential Speedup by returning a ref instead of a clone when asking for the RootDSE

[Perf]Potential Speedup by returning a ref instead of a clone when asking for the RootDSE
-----------------------------------------------------------------------------------------

                 Key: DIRSERVER-1290
                 URL: https://issues.apache.org/jira/browse/DIRSERVER-1290
             Project: Directory ApacheDS
          Issue Type: Improvement
    Affects Versions: 1.5.4
            Reporter: Emmanuel Lecharny
             Fix For: 2.0.0-RC1
         Attachments: rootDse.txt

We can speedup the server if we don't clone the rootDSE when we simply need to check for the presence of some
values, like in
protocol-ldap/src/main/java/org/apache/directory/server/ldap/handlers/SearchHandler.java :

         DirectoryService ds = session.getCoreSession().getDirectoryService();
         PartitionNexus nexus = ds.getPartitionNexus();
         Value<?> subschemaSubentry = nexus.getRootDSE( null ).get( SchemaConstants.SUBSCHEMA_SUBENTRY_AT ).get();
         LdapDN subschemaSubentryDn = new LdapDN( ( String ) ( subschemaSubentry.get() ) );
         subschemaSubentryDn.normalize(
ds.getRegistries().getAttributeTypeRegistry().getNormalizerMapping() );
         String subschemaSubentryDnNorm = subschemaSubentryDn.getNormName();

when we call the getRootDSE method, we get back a clone, and this is done once per search. We just need the
SubschemaSubentry attribute's value here, and we won't change it.

We could create a getRootDseRef() which returns a reference to the rootDSE object.

--

--
(Continue reading)

Permalink | Reply | 1 comment |
headers
Felix Knecht | 17 Nov 16:36
Favicon

Re: Javadoc on web site

Juergen Weber schrieb:
> And really cool were links from Javadocs to the corresponding source file in
> the subversion trunk...
> 

Here you go:
API: http://vm094.oxylos.org/projects/apacheds/apidocs/index.html
Source Xref: http://vm094.oxylos.org/projects/apacheds/xref/index.html

or all the generated docs from trunk (section Project Reports contains the ones above).
http://vm094.oxylos.org/projects/apacheds/

> Anybody knows a doclet that does this?

Maven does have a kind of such feature. The site above is generated by them. There exists more reports, see
http://maven.apache.org/plugins/index.html (project-info-reports e.g).

HTH
Felix

> 
> 
> Juergen Weber wrote:
>> Hi,
>>
>> I'd like to have the ds api docs on the web site, e.g. Documentation ->
>> api
>> like http://geronimo.apache.org/javadoc.html
>>
>> Would that be possible?
(Continue reading)

Permalink | Reply | 0 comments |
headers
Stefan Zoerner | 17 Nov 10:12

Re: Partition configurations [was Re: [CONF] Apache Directory SandBox: Draft - ... ]

Emmanuel Lecharny wrote:
> I'm afraid that I introduced some confusion with my previous mail. I was 
> talking about the partition from the user point of view, not from the 
> developper point of view. In this case, yes, you have to implement the 
> Partition interface.
> 
> What users want, mainly, is to define their own partition like, say, 
> "dc=my-company,dc=com", and inject data into it, not defining a new 
> partition with a SGBD storage (I guess).
> 
> I agree this is confusing, as we are mixing the partition we have in the 
> conf and the Partition interface, with all the magic needed to store 
> data and search into it.
> 
> Am I right about tis possible confusion ?

Yes, we talked about the way an advanced developer would implement his 
own backend storage, not configuration of a JDBM partition in server.xml

There are several mails on the lists about this topic. So I started to 
write something to have a starting point. In it's disclaimer, I have 
written the following:

--8<--
If you simply plan to add another suffix to ApacheDS (besides 
dc=example,dc=com, for instance) in order to store data, it is not 
necessary to write any code. You can simply add some lines to the 
configuration. The following is for developers who plan to implement 
another storage mechanism than the provided default.
-->8--
(Continue reading)

Permalink | Reply | 0 comments |
headers
Emmanuel Lecharny | 17 Nov 01:40
Gravatar

Re: Partition configurations [was Re: [CONF] Apache Directory SandBox: Draft - ... ]

Norval Hope wrote:
> Hi All,
>   
Hi Norval,

comment inline
> Just wanted to throw in some comments/observations from the
> prespective of a Virtual Directory implementation which sits on top of
> AD. In this kind of use-case it is completely necessary to discover
> connector implementations (aka custom partitions) dynamically at
> start-up (like a poor-man's OSGi implementation which requires a
> restart - as I too have had proper OSGi connector bundles on my list
> of TODOs for ages). For this purpose I have used Spring's
> FileSystemXmlApplicationContext's ability to discover files in jars on
> the the classpath, using a command line arguments like:
> "server_jcs.xml "classpath*:conf/connector.xml" (where each
> connector's .jar file has a specialized conf/connector.xml in it which
> configures only settings for the connector itself). Hence this
> approach is definitely possible, and desireable for cases like mine
> where connectors come and go regularly.
>   
Not sure I get a clear picture, but if it's just about separating 
partition's configuration in many different files to avoid having a big 
fat config file, then I agree : it's necessary. Tell me if I understood 
correctly.
> Secondly I get extremely nervous when I hear "use the DIT and ditch
> Spring" as has been mooted a number of times on the ML before, unless
> the DIT is merely a used as a storage place for XML orginally taken
> from "live" connector.xml files. 
Let's go back a bit backward. The very first version of ADS was using a
(Continue reading)

Permalink | Reply | 1 comment |
headers
Felix Knecht | 15 Nov 20:50
Favicon

Partition configurations [was Re: [CONF] Apache Directory SandBox: Draft - ... ]

Hi all

Following the ML I recently saw multiple questions about writing a custom partition so there seems to be an
interest in
this topic.

Now I'm just following Stefans (big thanks about this howto!) and hit on the needs to change the server.xml
when adding
a custom partition.

I'd like to raise a discussion about this.
Is it really needed to do this?!
Isn't a new partition just a module to plug into an existing server? Wouldn't it be nice just to drop in the jar
containing the custom partition and restart the server?
No need to change any configurations?

Why not let spring detecting all the existing partition modules for us?

It could be nicely done as all partitions implements the same interface. So get all beans implementing this interaface
and load them as partition. This way any partition module (=jar) contains its own definition. Adding the
jar to the ds
servers lib directory and restarting the server we be enough to get a new partition.

See also springs ListableBeanFactory [1] or [2]

[1] http://static.springframework.org/spring/docs/2.5.x/api/org/springframework/beans/factory/ListableBeanFactory.html
[2] http://forum.springframework.org/showthread.php?t=48066

Switching to such a module plug in way for partitions would it also make easy to create a maven archetype for
creating a
(Continue reading)

Permalink | Reply | 4 comments |
headers
Emmanuel Lecharny | 15 Nov 00:30
Gravatar

[ADS + MINA2] Very preliminary tests, take 2

Hi guys,

I did some more tests :

ADS + MINA 2.0 on my laptop : 3900 random search/s
ADS 1.5.4 + MINA 1.1.7 : 3490 random search/s

We still have a different JVM (1.6.0-10 / 1.6.0_10), but I'm not sure it 
makes a big difference. Also note that ADS-mina2 has been launched 
inside eclipse.

This is a 11% speedup.

More to come.

--

-- 
--
cordialement, regards,
Emmanuel Lécharny
www.iktek.com
directory.apache.org
Permalink | Reply | 0 comments |
headers
Emmanuel Lecharny | 14 Nov 19:48
Gravatar

[ADS + MINA2] Very preliminary tests

Hi guys,

just before I run to catch my train,

ADS + MINA 2.0 on my laptop : 3900 random search/s
ADS 1.5.4 + MINA 1.1.7 : 2000 random search/s

I have to validate those results, as I was running ADS/MINA2 from 
eclipse and ADS/MINA1 from studio, with a different JVM (1.6.0_10 and 
1.6.0_7).

Encouraging ...

--

-- 
--
cordialement, regards,
Emmanuel Lécharny
www.iktek.com
directory.apache.org
Permalink | Reply | 0 comments |

Gmane