Jeroen Reijn | 1 Jul 19:35 2012

Re: GC and performance

Hi,

see my comments inline:

On Sat, Jun 30, 2012 at 7:15 PM, Francisco Carriedo Scher
<fcarriedos <at> gmail.com> wrote:
> Hello!
>
> I am running JR in the following environment, a file is saved on a
> repository and then copied to another one through a GET method (the same
> way a FileInputStream can be provided, an InputStream comming from the HTTP
> connection). Memory, bandwitdh and CPU are enough and powerful, and it
> iniatilly completes about 15 / 16 transactions as described. But after an
> hour it comes down to 8-10 and after one hour more it keeps on 1-2.

How are you handling the file upload / copy? I've seen such slowdowns
mostly if a JCR session is reused and not refreshed or logged out.

>
> I am not running garbage collection as i assumed it does affect only to the
> space on disk, not the persistence manager, but looking about possible
> causes of such degradation on performance, misunderstanding the GC role
> could be an argument.

Yes the Garbage collector on the datastore is only for disk space
usage and not so much for performance.

>
> Could anyone please state if not running the GC does affect further than
> the space on disk?
(Continue reading)

Jeroen Reijn | 1 Jul 19:43 2012

Re: 'Invalid namespace index' exception which Apache Sling integration

Hi,

the error seems repository related. I personally have no experience
with Sling, but from the looks of it something is wrong while starting
up the repository.

What kind of configuration do you have for your repository. Perhaps
you can level up you loggin to see if the jackrabbit persistence
manager is connecting correctly to your database.

Also make sure that if you've used the embedded repository before that
the repository storage (which contains lucene indexes etc) is either
removed or pointing to a different location.

Perhaps you can share your repository.xml and your repository settings?

Regards,

Jeroen

On Mon, Jun 25, 2012 at 7:16 AM, Bhathiya Jayasekara
<tobhathiyaj <at> gmail.com> wrote:
> Hi all,
>
> I was trying to use Apache Sling to communicate with a Jacakrabbit
> repository. In Sling Launchpad[1], there is an embedded jackrabbit
> repository. But what I want is to plug an external jackrabbit repository.
> So I configured it through its admin console, according to this[2] thread.
>
> But then when I start Sling Launchpad, it gives following error in its
(Continue reading)

Bhathiya Jayasekara | 2 Jul 13:55 2012
Picon

Re: 'Invalid namespace index' exception which Apache Sling integration

Hi Jeroen,

Thank you for the reply. Since the issue was about namespace indices, I
compared ns_idx.properties file in embedded and external repositories and I
could find the indices were different in those 2. May be a jackrabbit
version problem. So few days ago I got a new version of Sling Launchpad
which has a new version of Jackrabbit and configured all again. Now that
namespace index issue does not occur. But, now there are some other
exceptions.

Please have a look at the last post of mine here[1]. I have posted the
current exceptions I get in error.log file. Do you have any idea why these
occur?

Thanks in advance.

--Bhathiya

[1]
http://apache-sling.73963.n3.nabble.com/Plugging-an-external-jackrabbit-repository-to-Sling-Launchpad-tc4018547.html

On Sun, Jul 1, 2012 at 11:13 PM, Jeroen Reijn <j.reijn <at> onehippo.com> wrote:

> Hi,
>
> the error seems repository related. I personally have no experience
> with Sling, but from the looks of it something is wrong while starting
> up the repository.
>
> What kind of configuration do you have for your repository. Perhaps
(Continue reading)

Jukka Zitting | 2 Jul 13:59 2012
Picon

Re: 'Invalid namespace index' exception which Apache Sling integration

Hi,

On Mon, Jul 2, 2012 at 1:55 PM, Bhathiya Jayasekara
<tobhathiyaj <at> gmail.com> wrote:
> Thank you for the reply. Since the issue was about namespace indices, I
> compared ns_idx.properties file in embedded and external repositories and I
> could find the indices were different in those 2. May be a jackrabbit
> version problem.

Sounds like a likely reason. You can upgrade to a more recent
Jackrabbit version, but using an existing Jackrabbit repository with
an older version of Jackrabbit is not guaranteed to work.

BR,

Jukka Zitting

Bhathiya Jayasekara | 2 Jul 16:07 2012
Picon

Re: 'Invalid namespace index' exception which Apache Sling integration

Hi Jukka,

I did the same with jackrabbit 2.4.2 with Sling Launchpad 7 which uses
jackrabbit 2.4.2. But I still get the following error in Sling error.log.
Any thoughts please?

02.07.2012 19:22:35.916 *INFO* [FelixStartLevel]
org.apache.jackrabbit.core.RepositoryImpl workspace 'security' initialized
02.07.2012 19:22:35.917 *INFO* [FelixStartLevel]
org.apache.jackrabbit.core.DefaultSecurityManager init: use Repository
Login-Configuration for Jackrabbit
02.07.2012 19:22:35.983 *INFO* [FelixStartLevel]
org.apache.jackrabbit.core.RepositoryImpl Repository started (6694ms)
02.07.2012 19:22:35.992 *INFO* [FelixStartLevel]
org.apache.sling.jcr.jackrabbit.server pingAndCheck; loginAdministrative
failed (javax.jcr.LoginException: LoginModule ignored Credentials)
javax.jcr.LoginException: LoginModule ignored Credentials
 at
org.apache.jackrabbit.core.RepositoryImpl.login(RepositoryImpl.java:1493)
at
org.apache.sling.jcr.base.AbstractSlingRepository.login(AbstractSlingRepository.java:205)
 at
org.apache.sling.jcr.base.AbstractSlingRepository.loginAdministrative(AbstractSlingRepository.java:171)
at
org.apache.sling.jcr.base.AbstractSlingRepository.pingAndCheck(AbstractSlingRepository.java:499)
 at
org.apache.sling.jcr.base.AbstractSlingRepository.startRepository(AbstractSlingRepository.java:755)
at
org.apache.sling.jcr.base.AbstractSlingRepository.activate(AbstractSlingRepository.java:581)
 at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
(Continue reading)

Luca Tagliani | 3 Jul 09:43 2012
Picon

Clustering local revision id getting very high

Hi all,
  we have a clustered production system with 4 nodes that are syncing every
2 seconds.
We are using DB Journaling using a central MSSQL Database.

We see that after a single day the local revision id reaches 434339.

At this rate of growth in a relative little time this field will reach its
highest value and, according to the source, after that jackrabbit will throw
an exception because it can't increment the field anymore.

Is there a way to reset this counter (perhaps stopping every node of the
cluster) to prevent this possible problem?

BR

Luca Tagliani

--
View this message in context: http://jackrabbit.510166.n4.nabble.com/Clustering-local-revision-id-getting-very-high-tp4655731.html
Sent from the Jackrabbit - Users mailing list archive at Nabble.com.

Alex Parvulescu | 3 Jul 11:19 2012
Picon

Re: jcr sql2 - contains() full text search not working

Hi Carl,

What version of jackrabbit are you on?

Next, are you sure you have the tika extractors in the classpath? maybe you
are seeing something along the lines of [0].

I would try to isolate the problem by taking tomcat out of the setup. Build
a simple test, see how it works then deploy on tomcat and verify.
A good place to start is the unit test collection available in jackrabbit
core [1].

best,
alex

[0] https://issues.apache.org/jira/browse/JCR-3287
[1]
http://svn.apache.org/viewvc/jackrabbit/trunk/jackrabbit-core/src/test/java/org/apache/jackrabbit/core/query/FulltextSQL2QueryTest.java?view=markup

On Wed, Jun 27, 2012 at 8:06 PM, Furst, Carl <Carl.Furst <at> mlb.com> wrote:

> So given the below I tried to use
>
> 'inclu*' and 'include*' and still no results so I'm going to start looking
> into perhaps maybe some of these reasons as why:
>
> https://wiki.apache.org/lucene-java/LuceneFAQ#Why_am_I_getting_no_hits_.2BA
> C8_incorrect_hits.3F
>
> Of course it could just be that the parser is not parsing the '*'.
(Continue reading)

KÖLL Claus | 3 Jul 13:08 2012
Picon

AW: Clustering local revision id getting very high

The Revision is a java long type inside jackrabbit.

The Max value of a long is 2^36 = 68 719 476 736
If you have 434339 changes on one day the max value will exceed in 158216 days.
These are 433 years ... i think you could run jackrabbit a long time.

Of course you can reset the revision. If all cluster members are up to date and stopped you can reset the
global revison and the local revisions in the db.
But be carefull and better test it before in a test environment...

greets
claus

hsp | 3 Jul 14:31 2012
Picon

Migrate Jackrabbit 1.4 to 2.x

Hi;
I would to know how to do the update in my project repository that is using
jackrabbit 1.4.x to the newest 2.x.
My repository is using datastore. I was thinking to migrate to 1.6 and after
migrate do 2.0, is it necessary?
And what about my nodetypes, will they remain as they are when in the
jsr283?
My doubt is because there is some legacy (a lot of...), and it is critical
to retain the same uuids in workspaces.
Hope can be helped, thanks!

--
View this message in context: http://jackrabbit.510166.n4.nabble.com/Migrate-Jackrabbit-1-4-to-2-x-tp4655746.html
Sent from the Jackrabbit - Users mailing list archive at Nabble.com.

Jukka Zitting | 3 Jul 15:10 2012
Picon

Re: Migrate Jackrabbit 1.4 to 2.x

Hi,

On Tue, Jul 3, 2012 at 2:31 PM, hsp <piccinatto <at> ibest.com.br> wrote:
> I would to know how to do the update in my project repository that is using
> jackrabbit 1.4.x to the newest 2.x.
> My repository is using datastore. I was thinking to migrate to 1.6 and after
> migrate do 2.0, is it necessary?
> And what about my nodetypes, will they remain as they are when in the
> jsr283?
> My doubt is because there is some legacy (a lot of...), and it is critical
> to retain the same uuids in workspaces.

You should be able to use the existing 1.4.x repository as-is also
with Jackrabbit 2.x. Existing content, node types, uuids, etc. should
all remain the same over the upgrade. Of course it's still a good idea
to backup your repository before the upgrade and do a full round of
testing after that to verify that everything's OK.

If you do encounter any problems, upgrading first to 1.6 and then to
2.x might help solve them or at least better pinpoint the issue.

Note that, depending on your deployment, you may also want to migrate
to a more recent repository configuration in order to take better
advantage of the performance and other improvements that we've added
over time. See http://jackrabbit.apache.org/standalone-server.html#StandaloneServer-Backupandmigration
for instructions on how to use the migration feature of the Jackrabbit
standalone jar.

BR,

(Continue reading)


Gmane