How can we backup beagle data, including user_xattr?
What happens if we don't restore user_xattr correctly? Will beagle re-create it?
On Tue, 2006-03-07 at 10:59 -0500, Neal Becker wrote:
How can we backup beagle data, including user_xattr?
What happens if we don't restore user_xattr correctly? Will beagle re-create it?
Beagle does not use extended attributes in FC5 (at least not by default), it uses sqlite. So there should be no problem.
Matthias
Matthias Clasen wrote:
On Tue, 2006-03-07 at 10:59 -0500, Neal Becker wrote:
How can we backup beagle data, including user_xattr?
What happens if we don't restore user_xattr correctly? Will beagle re-create it?
Beagle does not use extended attributes in FC5 (at least not by default), it uses sqlite. So there should be no problem.
Is this documented? How can I tell?
If I look at beagle docs on the web, it tells me I really should mount with user_xattr, and that beagle will auto-detect this at runtime.
Beagle does not use extended attributes in FC5 (at least not by default), it uses sqlite. So there should be no problem.
Then, why do my files show attributes created by beagle?
# getfattr Desktop user.Beagle.AttrTime user.Beagle.Fingerprint user.Beagle.MTime user.Beagle.Uid
(BTW, the filesystem is being mounted with -o user_xattr).
Felipe Alfaro Solana wrote:
Beagle does not use extended attributes in FC5 (at least not by default), it uses sqlite. So there should be no problem.
Then, why do my files show attributes created by beagle?
# getfattr Desktop user.Beagle.AttrTime user.Beagle.Fingerprint user.Beagle.MTime user.Beagle.Uid
(BTW, the filesystem is being mounted with -o user_xattr).
thats the reason, by default filesystems are not mounted with user_xattr (why?)
On Tue, 2006-03-07 at 17:15 +0100, Felipe Alfaro Solana wrote:
(BTW, the filesystem is being mounted with -o user_xattr).
thats the reason, by default filesystems are not mounted with user_xattr (why?)
Yes... Any good reason for not doing so?
Beagle was added sort of late in the cycle, and changing the default mount options needs changes in anaconda etc, so we (the desktop team) really hasn't been pushing this. At some point we should sit down with the installer people and the kernel filesystem people and figure out if and when we should set user_xattr.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Alexander Larsson Red Hat, Inc alexl@redhat.com alla@lysator.liu.se He's a deeply religious Republican filmmaker from the Mississippi delta. She's a time-travelling motormouth safe cracker married to the Mob. They fight crime!
ons, 08 03 2006 kl. 09:21 +0100, skrev Alexander Larsson:
On Tue, 2006-03-07 at 17:15 +0100, Felipe Alfaro Solana wrote:
(BTW, the filesystem is being mounted with -o user_xattr).
thats the reason, by default filesystems are not mounted with user_xattr (why?)
Yes... Any good reason for not doing so?
Beagle was added sort of late in the cycle, and changing the default mount options needs changes in anaconda etc, so we (the desktop team) really hasn't been pushing this. At some point we should sit down with the installer people and the kernel filesystem people and figure out if and when we should set user_xattr.
For FC6 isn't one of the goals to get seperate partitions for /home - then it would be a great time to add the user_xattr to that partition at least.
- David
David Nielsen wrote:
For FC6 isn't one of the goals to get seperate partitions for /home - then it would be a great time to add the user_xattr to that partition at least.
/home by default is requested at https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=150670. Jeremy Katz had other ideas about doing this in a much better and flexible way though I dont remember the precise details now. Would you file a RFE against anaconda to add the user_xattr by default to track this?
On Wed, 2006-03-08 at 14:16 +0530, Rahul Sundaram wrote:
David Nielsen wrote:
For FC6 isn't one of the goals to get seperate partitions for /home - then it would be a great time to add the user_xattr to that partition at least.
/home by default is requested at https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=150670. Jeremy Katz had other ideas about doing this in a much better and flexible way though I dont remember the precise details now. Would you file a RFE against anaconda to add the user_xattr by default to track this?
The right way to get user_xattr by default is to change the default mount options for the filesystem in the kernel, not installer hacks.
Jeremy
On Tue, 2006-03-07 at 17:10 +0100, Felipe Alfaro Solana wrote:
Beagle does not use extended attributes in FC5 (at least not by default), it uses sqlite. So there should be no problem.
Then, why do my files show attributes created by beagle?
# getfattr Desktop user.Beagle.AttrTime user.Beagle.Fingerprint user.Beagle.MTime user.Beagle.Uid
(BTW, the filesystem is being mounted with -o user_xattr).
Then you turned it on yourself. user_xattr is not on by default.
On Tue, 2006-03-07 at 10:59 -0500, Neal Becker wrote:
How can we backup beagle data, including user_xattr?
What happens if we don't restore user_xattr correctly? Will beagle re-create it?
beagle uses the xattrs to store information about if/when/how the file was indexed. I think its fine if this information gets lost, although it will cause beagle to re-index the file.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Alexander Larsson Red Hat, Inc alexl@redhat.com alla@lysator.liu.se He's a scrappy chivalrous boxer She's a wealthy belly-dancing opera singer from a secret island of warrior women. They fight crime!
devel@lists.stg.fedoraproject.org
-
Alexander Larsson -
David Nielsen -
dragoran -
Felipe Alfaro Solana -
Jeremy Katz -
Matthias Clasen -
Neal Becker -
Rahul Sundaram