[GeoNetwork-users] Harvesting problem

Michael_Vanderhoydon · September 7, 2006, 7:53am

Hello Andrea,

Thanks very much for your reply. You're right, port number for node A
should be 8081. Changing this setting makes the ConnectException
disappear, but harvesting is still not working.

There is no error message generated, but there is no data harvested
either, see message below:

2006-09-07 07:31:40 | [geonetwork sync] Searching on : geonetwork
2006-09-07 07:31:41 | [geonetwork sync] Obtained:
<response xmlns:geonet="http://www.fao.org/geonetwork" from="0" to="0">
  <summary count="0" type="local">
    <keywords />
    <categories />
  </summary>
</response>
2006-09-07 07:31:41 | [geonetwork sync] Start of alignment for site-id : abc
2006-09-07 07:31:41 | [geonetwork sync] End of alignment for site-id : abc

In the meantime I have managed to get harvesting working, but only
between two SERVER installations. So I've installed a new node, which
is a server GeoNetwork installation pointing to a PostgreSQL database
and using Tomcat as servlet engine. The data for which I've set the
visible privilege to the internet group is harvested. Other data is
not.

Now I've got some questions concerning the harvesting mechanism
(hoping that someone can help me with it)

1/ Why is harvesting not working between a desktop installed node and
a server installed node? Someone experienced similar behaviour? (Both
the desktop and server installations have the same configuration
(using correct port numbers of course...)

2/ Why is it necessary to assign the visible privilege to the
internet group on a certain dataset for that dataset to be harvested?
Because this has some consequenses. For example: let's take again the
example of 2 nodes: node A and node B. I've got some private data on
node A (and there is a user/group which has access to these private
data, other users/groups on my node do not have), that I would like to
share with node B, but not with any other nodes. Is it possible?

(I know you can create the same user/group on node B holding the
private data and then revoking the visible and download privilege to
the internet group. This solves the problem of internet users
accessing my data on node B. But then my data is still accessible to
other nodes....

Hope that someone can shed some light on this issue,

Thank you very much,
Kind regards

Michael.

====================

Hi Michael,

the error is in config.xml for node B:
<site name="geonetwork">
    <general>
    <host>xxx.xx.xxx.xxx</host> (= IP of node A)
    <port>8080</port>
    ....

As you can see the port is 8080 but node A is installed into 8081.

Cheers,
Andrea

Andrea_Carboni1 · September 7, 2006, 2:31pm

Hi Michael,

here are some notes:

There is no error message generated, but there is no data harvested
either, see message below:

2006-09-07 07:31:40 | [geonetwork sync] Searching on : geonetwork
2006-09-07 07:31:41 | [geonetwork sync] Obtained:
<response xmlns:geonet="http://www.fao.org/geonetwork" from="0" to="0">
  <summary count="0" type="local">
    <keywords />
    <categories />
  </summary>
</response>
2006-09-07 07:31:41 | [geonetwork sync] Start of alignment for site-id : abc
2006-09-07 07:31:41 | [geonetwork sync] End of alignment for site-id : abc

Maybe the siteId is wrong or you imported the metadata in a way that bypassed
the lucene indexing. In general, the harvesting performs the same search as you
do in the main page: if you get results here you should get the same results with
harvesting. So, first try the same search using the main page.

In the meantime I have managed to get harvesting working, but only
between two SERVER installations. So I've installed a new node, which
is a server GeoNetwork installation pointing to a PostgreSQL database
and using Tomcat as servlet engine. The data for which I've set the
visible privilege to the internet group is harvested. Other data is
not.

Correct, if you do not provide an account you can see only public metadata.

1/ Why is harvesting not working between a desktop installed node and
a server installed node? Someone experienced similar behaviour? (Both
the desktop and server installations have the same configuration
(using correct port numbers of course...)

The desktop installation uses some defaults so maybe these are wrong.
Beside this, the 2 are identical.

2/ Why is it necessary to assign the visible privilege to the
internet group on a certain dataset for that dataset to be harvested?

'view' is the lower privilege so if a metadata can be viewed it can be harvested.
Maybe a separate 'harvest' privilege could be added. What do other users think
about this?

Because this has some consequenses. For example: let's take again the
example of 2 nodes: node A and node B. I've got some private data on
node A (and there is a user/group which has access to these private
data, other users/groups on my node do not have), that I would like to
share with node B, but not with any other nodes. Is it possible?

If I understand well, you can:

- create a group G in B
- setup B's harvesting using an A's account to get only your private data
- setup B's harvesting to assing 'view' privileges to group G

(I know you can create the same user/group on node B holding the
private data and then revoking the visible and download privilege to
the internet group. This solves the problem of internet users
accessing my data on node B. But then my data is still accessible to
other nodes....

Why? If other nodes want to harvest from B they need a B's account.

Thank you very much,
Kind regards

Michael.

Cheers,
Andrea