You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@stanbol.apache.org by Bradley Falzon <br...@teambrad.net> on 2015/03/25 23:14:44 UTC
Recommended Backup Strategy
Hi All,
We're beginning initial testing of Stanbol, and wanted to know the
recommended approach to backups. We've tried checking the mailing list and
stanbol.apache.org for best practices, to no avail.
Is there any features built in? We're really unsure on the internal
workings and best strategy here.
Thoughts?
--
Bradley Falzon
brad@teambrad.net
Re: Recommended Backup Strategy
Posted by Rupert Westenthaler <ru...@gmail.com>.
Hi all,
I agree with Rafa.
For configuration you have three strategies:
(1) build everything into a custom launcher
(2) manage the configuration via the Sling Installer Framework [1] (e.g.
the FileInstaller).
(3) manage your configuration directly via the OSGI Configuration Admin [2]
As (1) and (2) are also based on the OSGI Configuration Admin you can also
mix those strategies.
In the simplest case you can put static configurations directly in the
launcher and use the Apache Sling Fileinstaller or if you like the OSGI
FileInstaller for managing custom configurations.
For binary data Stanbol uses the DataFileProvider [3] infrastructure. In
the simplest case just put all your files in the "stanbol/datafiles" folder.
Finally the Stanbol Solr integration manages an embedded Solr server. The
index data of this server are stored under "stanbol/indexes". So this
folder will hold your index data. If you use Entityhub ManagedSites with a
Solr backend your data will be stored under this folder.
Finally you will also find the default Apache Sling and Apache Felix folders
> stanbol/conf/
> stanbol/config/
This is the runtime configuration of the OSGI components. Never change
those by hand and keep a backup
> stanbol/startup/
> stanbol/felix/
Copies of the bundles. Created on the first startup. Updated only when you
install, update, remove bundles in the OSGI environment.
> stanbol/fileinstall/
If you have the Sling Fileinstall in use this contains configuration files.
If you also enable syncing back configurations created by other means you
will also have those files in this folder. I would strongly recommend to
deactivate this feature for production. NOTE: the actual runtime
configuration is in "stanbol/config" this is an 'external' view.
> stanbol/installer/
This is used by the Sling Installer framework to check if configurations
are updated. Needed to recover the status of components configured via the
Sling Installer framework.
> stanbol/sling.properties
The Sling configuration file
I hope this helps
best
Rupert
[1] http://sling.apache.org/documentation/bundles/osgi-installer.html
[2]
http://felix.apache.org/documentation/subprojects/apache-felix-config-admin.html
[3] http://stanbol.apache.org/docs/trunk/utils/datafileprovider
On Thu, Mar 26, 2015 at 10:58 AM, Rafa Haro <rh...@apache.org> wrote:
> Hi Aree and Bradley,
>
> About the backup, it really depends on your uses cases. Most of the
> Stanbol services are completely stateless and that (initially) means you
> don’t need a backup. If you are using a custom dataset, or some custom
> configurations you probably would like to backup them. For the datasets, if
> you have used the generic indexer to build them, the outcomes of that tool
> could be your backup: at the end you will have a Solr index that you can
> install in a clean Stanbol whenever you want and a bundle for the EntityHub
> site that you can also install in any moment.
>
> Regarding configuration, probably the best solution is to configure your
> own Stanbol launcher with your final configuration. With all these
> resources you can easily replicate your instance.
>
> Hope that helps. Cheers,
> Rafa
>
>
> En 25 de marzo de 2015 en 23:31:18, Aree Cohen (aree.cohen@sbs.com.au)
> escrito:
>
> Perfect timing. We are in the same boat. We're about to deliver a critical
> project using Stanbol and need help with setting up our production
> environment.
> Moreover, does anyone know the best strategy for replication between two
> Stanbol servers. Is rsyncing a feasible strategy?
>
> Any help with this would be extremely appreciated.
>
> Thanks,
> Aree
>
> ________________________________________
> From: Bradley Falzon <br...@teambrad.net>
> Sent: Thursday, 26 March 2015 9:14 AM
> To: dev@stanbol.apache.org
> Subject: Recommended Backup Strategy
>
> Hi All,
>
> We're beginning initial testing of Stanbol, and wanted to know the
> recommended approach to backups. We've tried checking the mailing list and
> stanbol.apache.org for best practices, to no avail.
>
> Is there any features built in? We're really unsure on the internal
> workings and best strategy here.
>
> Thoughts?
>
> --
> Bradley Falzon
> brad@teambrad.net
>
>
> [Vikings. Thursdays, 8.30pm, SBS ONE.]<http://sbs.com.au/vikings>
>
--
| Rupert Westenthaler rupert.westenthaler@gmail.com
| Bodenlehenstraße 11 ++43-699-11108907
| A-5500 Bischofshofen
| REDLINK.CO
..........................................................................
| http://redlink.co/
Re: Recommended Backup Strategy
Posted by Rafa Haro <rh...@apache.org>.
Hi Aree and Bradley,
About the backup, it really depends on your uses cases. Most of the Stanbol services are completely stateless and that (initially) means you don’t need a backup. If you are using a custom dataset, or some custom configurations you probably would like to backup them. For the datasets, if you have used the generic indexer to build them, the outcomes of that tool could be your backup: at the end you will have a Solr index that you can install in a clean Stanbol whenever you want and a bundle for the EntityHub site that you can also install in any moment.
Regarding configuration, probably the best solution is to configure your own Stanbol launcher with your final configuration. With all these resources you can easily replicate your instance.
Hope that helps. Cheers,
Rafa
En 25 de marzo de 2015 en 23:31:18, Aree Cohen (aree.cohen@sbs.com.au) escrito:
Perfect timing. We are in the same boat. We're about to deliver a critical project using Stanbol and need help with setting up our production environment.
Moreover, does anyone know the best strategy for replication between two Stanbol servers. Is rsyncing a feasible strategy?
Any help with this would be extremely appreciated.
Thanks,
Aree
________________________________________
From: Bradley Falzon <br...@teambrad.net>
Sent: Thursday, 26 March 2015 9:14 AM
To: dev@stanbol.apache.org
Subject: Recommended Backup Strategy
Hi All,
We're beginning initial testing of Stanbol, and wanted to know the
recommended approach to backups. We've tried checking the mailing list and
stanbol.apache.org for best practices, to no avail.
Is there any features built in? We're really unsure on the internal
workings and best strategy here.
Thoughts?
--
Bradley Falzon
brad@teambrad.net
[Vikings. Thursdays, 8.30pm, SBS ONE.]<http://sbs.com.au/vikings>
Re: Recommended Backup Strategy
Posted by Aree Cohen <Ar...@sbs.com.au>.
Perfect timing. We are in the same boat. We're about to deliver a critical project using Stanbol and need help with setting up our production environment.
Moreover, does anyone know the best strategy for replication between two Stanbol servers. Is rsyncing a feasible strategy?
Any help with this would be extremely appreciated.
Thanks,
Aree
________________________________________
From: Bradley Falzon <br...@teambrad.net>
Sent: Thursday, 26 March 2015 9:14 AM
To: dev@stanbol.apache.org
Subject: Recommended Backup Strategy
Hi All,
We're beginning initial testing of Stanbol, and wanted to know the
recommended approach to backups. We've tried checking the mailing list and
stanbol.apache.org for best practices, to no avail.
Is there any features built in? We're really unsure on the internal
workings and best strategy here.
Thoughts?
--
Bradley Falzon
brad@teambrad.net
[Vikings. Thursdays, 8.30pm, SBS ONE.]<http://sbs.com.au/vikings>