[Plone-docs] Where there is more than one var/Data.fs in Plone 3.x: how to plan for more than one var/blobstorage before an upgrade to 4.x?

[Graham Perrin]

Cc: ***@lists.plone.org (with apologies for cross-posting)

Reading
<http://manage.plone.org/documentation/kb/multiple-plone-sites-per-zope-instance-using-separate-data-fs-files-for-each-one>
alongside
<http://manage.plone.org/documentation/manual/upgrade-guide/version/upgrading-plone-3-x-to-4.0>
(some of which is not yet published):

If an administrator has designed their Plone 3.x installation to have more
than one Data.fs (say, one per site), then how should that person plan for
an upgrade to 4.x?

Would it be essential for their buildout for 4.x to create separate BLOB
storages *before* the upgrade?

Or could they upgrade first — to just one var/blobstorage for all sites
— then think about separation at some point in the future?

[Raphael Ritz]

Post by Graham Perrin
Reading
<http://manage.plone.org/documentation/kb/multiple-plone-sites-per-zope-instance-using-separate-data-fs-files-for-each-one>
alongside
<http://manage.plone.org/documentation/manual/upgrade-guide/version/upgrading-plone-3-x-to-4.0>
If an administrator has designed their Plone 3.x installation to have more
than one Data.fs (say, one per site), then how should that person plan for
an upgrade to 4.x?

By considering to use

http://pypi.python.org/pypi/collective.recipe.filestorage

Raphael

Post by Graham Perrin
Would it be essential for their buildout for 4.x to create separate BLOB
storages *before* the upgrade?
Or could they upgrade first — to just one var/blobstorage for all sites
— then think about separation at some point in the future?

[Graham Perrin]

Post by Raphael Ritz

… how should that person plan for an upgrade to 4.x?

By considering to use
http://pypi.python.org/pypi/collective.recipe.filestorage
Raphael

Would it be essential for their buildout for 4.x to create separate BLOB
storages *before* the upgrade?
Or could they upgrade first — to just one var/blobstorage for all sites
— then think about separation at some point in the future?

Raphael, thanks, that's an interesting recipe.

An underlying question: must a person with multiple Data.fs think about
multiple BLOB storages (with or without that recipe) *before* an upgrade
from 3.x to 4.x?

Put another way: if they upgrade, and over a period of weeks or months the
one /blobstorage/ is populated with BLOBs from all sites A, B and C, then
eventually they discover collective.recipe.filestorage and decide that
separate storages is the way to go, then will Plone:

i) intelligently work with the existing mixture of all three, to separate
the BLOBs for sites B and C, from the BLOBs for site A

or

ii) leave the *existing* BLOBs for A, B and C together; and separate only
the BLOBs that are added *after* separation has been specified?

[Raphael Ritz]

Post by Graham Perrin

Post by Raphael Ritz

… how should that person plan for an upgrade to 4.x?

By considering to use
http://pypi.python.org/pypi/collective.recipe.filestorage
Raphael

Would it be essential for their buildout for 4.x to create separate BLOB
storages *before* the upgrade?
Or could they upgrade first — to just one var/blobstorage for all sites
— then think about separation at some point in the future?

Raphael, thanks, that's an interesting recipe.
An underlying question: must a person with multiple Data.fs think about
multiple BLOB storages (with or without that recipe) *before* an upgrade
from 3.x to 4.x?

Good question.

I really don't know without trying this as I didn't come
across that myself yet.

@Andi (cced): any recommendations?

Raphael

Post by Graham Perrin
Put another way: if they upgrade, and over a period of weeks or months the
one /blobstorage/ is populated with BLOBs from all sites A, B and C, then
eventually they discover collective.recipe.filestorage and decide that
i) intelligently work with the existing mixture of all three, to separate
the BLOBs for sites B and C, from the BLOBs for site A
or
ii) leave the *existing* BLOBs for A, B and C together; and separate only
the BLOBs that are added *after* separation has been specified?