You are viewing a plain text version of this content. The canonical link for it is here.
Posted to oak-dev@jackrabbit.apache.org by Marcel Reutegger <mr...@adobe.com> on 2014/03/26 17:12:29 UTC

RE: svn commit: r1581376 - /jackrabbit/oak/trunk/oak-jcr/src/test/java/org/apache/jackrabbit/oak/jcr/NodeStoreFixture.java

can we revert this change? this increases the build time quite a bit on
my machine. alternatively we could slightly increase the heap size...

regards
 marcel

> -----Original Message-----
> From: reschke@apache.org [mailto:reschke@apache.org]
> Sent: Dienstag, 25. März 2014 16:36
> To: oak-commits@jackrabbit.apache.org
> Subject: svn commit: r1581376 - /jackrabbit/oak/trunk/oak-
> jcr/src/test/java/org/apache/jackrabbit/oak/jcr/NodeStoreFixture.java
> 
> Author: reschke
> Date: Tue Mar 25 15:35:34 2014
> New Revision: 1581376
> 
> URL: http://svn.apache.org/r1581376
> Log:
> OAK-1478 - use in-file instead of in-mem H2 database (because of OOMs)
> 
> Modified:
>     jackrabbit/oak/trunk/oak-
> jcr/src/test/java/org/apache/jackrabbit/oak/jcr/NodeStoreFixture.java
> 
> Modified: jackrabbit/oak/trunk/oak-
> jcr/src/test/java/org/apache/jackrabbit/oak/jcr/NodeStoreFixture.java
> URL: http://svn.apache.org/viewvc/jackrabbit/oak/trunk/oak-
> jcr/src/test/java/org/apache/jackrabbit/oak/jcr/NodeStoreFixture.java?rev=
> 1581376&r1=1581375&r2=1581376&view=diff
> ==========================================================
> ====================
> --- jackrabbit/oak/trunk/oak-
> jcr/src/test/java/org/apache/jackrabbit/oak/jcr/NodeStoreFixture.java
> (original)
> +++ jackrabbit/oak/trunk/oak-
> jcr/src/test/java/org/apache/jackrabbit/oak/jcr/NodeStoreFixture.java Tue
> Mar 25 15:35:34 2014
> @@ -19,10 +19,12 @@
>  package org.apache.jackrabbit.oak.jcr;
> 
>  import java.io.Closeable;
> +import java.io.File;
>  import java.io.IOException;
>  import java.util.UUID;
> 
>  import com.mongodb.DB;
> +
>  import org.apache.jackrabbit.mk.core.MicroKernelImpl;
>  import org.apache.jackrabbit.oak.kernel.KernelNodeStore;
>  import org.apache.jackrabbit.oak.plugins.document.DocumentMK;
> @@ -78,13 +80,15 @@ public abstract class NodeStoreFixture {
>          @Override
>          public NodeStore createNodeStore() {
>              String id = UUID.randomUUID().toString();
> -            return new
> DocumentMK.Builder().setRDBConnection("jdbc:h2:mem:" + id +
> ";MVCC=true", "sa", "").getNodeStore();
> +            String folder = (new File("target")).isDirectory() ? "target/" : "";
> +            return new DocumentMK.Builder().setRDBConnection("jdbc:h2:file:"
> + folder + id + ";MVCC=true", "sa", "").getNodeStore();
>          }
> 
>          @Override
>          public NodeStore createNodeStore(int clusterNodeId) {
>              try {
> -                return new
> DocumentMK.Builder().setRDBConnection("jdbc:h2:mem:oaknodes-" +
> clusterNodeId, "sa", "").getNodeStore();
> +                String folder = (new File("target")).isDirectory() ? "target/" : "";
> +                return new
> DocumentMK.Builder().setRDBConnection("jdbc:h2:file:" + folder +
> "oaknodes-" + clusterNodeId, "sa", "").getNodeStore();
>              } catch (Exception e) {
>                  return null;
>              }
> 


Re: svn commit: r1581376 - /jackrabbit/oak/trunk/oak-jcr/src/test/java/org/apache/jackrabbit/oak/jcr/NodeStoreFixture.java

Posted by Julian Reschke <ju...@gmx.de>.
On 2014-03-27 10:53, Marcel Reutegger wrote:
> OK, now I got the OOME as well with my local in-memory H2 changes.
>
> it failed in ObservationRefreshTest. I analyzed the heap dump and
> it looks like the memory usage of this particular test is quite high
> because it uses orderable child nodes and adds child nodes one
> by one.
>
> I think that can be easily changed...
>
> Regards
>   Marcel

That's of course a much better idea :-)


RE: svn commit: r1581376 - /jackrabbit/oak/trunk/oak-jcr/src/test/java/org/apache/jackrabbit/oak/jcr/NodeStoreFixture.java

Posted by Marcel Reutegger <mr...@adobe.com>.
OK, now I got the OOME as well with my local in-memory H2 changes.

it failed in ObservationRefreshTest. I analyzed the heap dump and
it looks like the memory usage of this particular test is quite high
because it uses orderable child nodes and adds child nodes one
by one.

I think that can be easily changed...

Regards
 Marcel

> -----Original Message-----
> From: Julian Reschke [mailto:julian.reschke@gmx.de]
> Sent: Mittwoch, 26. März 2014 21:23
> To: oak-dev@jackrabbit.apache.org
> Subject: Re: svn commit: r1581376 - /jackrabbit/oak/trunk/oak-
> jcr/src/test/java/org/apache/jackrabbit/oak/jcr/NodeStoreFixture.java
> 
> On 2014-03-26 21:08, Marcel Reutegger wrote:
> > Hi,
> >
> > locally, I reverted your change and the build runs fine for me.
> > that's why I thought it might work for you as well if we increase
> > the heap slightly.
> >
> > or does it only happen when you run the build with some specific
> > profile?
> > ...
> 
> No, it was failing for me with the default profile.
> 
> Best regards, Julian

Re: svn commit: r1581376 - /jackrabbit/oak/trunk/oak-jcr/src/test/java/org/apache/jackrabbit/oak/jcr/NodeStoreFixture.java

Posted by Julian Reschke <ju...@gmx.de>.
On 2014-03-26 21:08, Marcel Reutegger wrote:
> Hi,
>
> locally, I reverted your change and the build runs fine for me.
> that's why I thought it might work for you as well if we increase
> the heap slightly.
>
> or does it only happen when you run the build with some specific
> profile?
> ...

No, it was failing for me with the default profile.

Best regards, Julian

RE: svn commit: r1581376 - /jackrabbit/oak/trunk/oak-jcr/src/test/java/org/apache/jackrabbit/oak/jcr/NodeStoreFixture.java

Posted by Marcel Reutegger <mr...@adobe.com>.
Hi,

locally, I reverted your change and the build runs fine for me.
that's why I thought it might work for you as well if we increase
the heap slightly.

or does it only happen when you run the build with some specific
profile?

regards
 marcel

> -----Original Message-----
> From: Julian Reschke [mailto:julian.reschke@gmx.de]
> Sent: Mittwoch, 26. März 2014 20:47
> To: oak-dev@jackrabbit.apache.org
> Subject: Re: svn commit: r1581376 - /jackrabbit/oak/trunk/oak-
> jcr/src/test/java/org/apache/jackrabbit/oak/jcr/NodeStoreFixture.java
> 
> On 2014-03-26 17:12, Marcel Reutegger wrote:
> > can we revert this change? this increases the build time quite a bit on
> > my machine. alternatively we could slightly increase the heap size...
> >
> > regards
> >   marcel
> > ...
> 
> I'm not sure that increasing slightly is going to help; do you want to
> give it a try?
> 
> Alternatively we should work on having more control over what
> persistences get tested by default.
> 
> Best regards, Julian

Re: svn commit: r1581376 - /jackrabbit/oak/trunk/oak-jcr/src/test/java/org/apache/jackrabbit/oak/jcr/NodeStoreFixture.java

Posted by Julian Reschke <ju...@gmx.de>.
On 2014-03-26 17:12, Marcel Reutegger wrote:
> can we revert this change? this increases the build time quite a bit on
> my machine. alternatively we could slightly increase the heap size...
>
> regards
>   marcel
> ...

I'm not sure that increasing slightly is going to help; do you want to 
give it a try?

Alternatively we should work on having more control over what 
persistences get tested by default.

Best regards, Julian