zopistas, The 2.3.27+ version of the kernel supports large file sizes. I have 2.3.47 installed, and I've set the block size on ext2 to 4k, so with my kernel, i should be able to do files of 1TB. However, many of the userspace programs in linux don't recognize a file that's larger than 4GB: for example, ls reports that my 16GB test file is 64MB. What I'm wondering is whether Zope would have any problems with a large Data.fs on linux and/or any suggestions on how to test it. I expect my Data.fs to grow to 26GB in size... Thanks, Ethan Fremen http://imeme.net
Hi Peeps, Im just getting started with Zope and one of things I think would be extremely usefull would be a LocalUndo subclassable object. Essentially, subclassing from this LocalUndo object would add a LocalUndo property page which would enable a manager to undo operations related to this object and this object only. A variation of this feature would be to enable undo operations on this object and its subobjects. My Python is pretty good, and I'm not averse to delving into the source. If anyone has any tips or pointers on how to approach creating this feature, Id greatly appreciate it.
Damian Morton wrote:
Hi Peeps,
Im just getting started with Zope and one of things I think would be extremely usefull would be a LocalUndo subclassable object. Essentially, subclassing from this LocalUndo object would add a LocalUndo property page which would enable a manager to undo operations related to this object and this object only. A variation of this feature would be to enable undo operations on this object and its subobjects.
Good idea, not as easy as it sounds though. First, you don't undo object, you undo transactions, and any number of object may change in one transaction. This is just one possible difficulty to surmount. I'm not saying it's not possible, it just might require quite a bit of deep thinking about doing it the right way. This is definatly a Jim department, although I've been spending a bit of time researching how ZODB works down at the scientific level.
My Python is pretty good, and I'm not averse to delving into the source. If anyone has any tips or pointers on how to approach creating this feature, Id greatly appreciate it.
Well, the undoLog (where the list of undoable transactions come from) comes from the actualy storage system itself (since it's possible that hypothetical storages (and real ones, such as Ty Sarna's Berkeley storage) many not support undo...). Check out lib/python/ZODB/FileStorage.py and start looking at the undoLog method. This is agood place to start to see how undo works in Zope. -Michel
_______________________________________________ Zope-Dev maillist - Zope-Dev@zope.org http://lists.zope.org/mailman/listinfo/zope-dev ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope )
On Sun, Feb 27, 2000 at 08:30:17PM +0000, ethan mindlace fremen wrote:
The 2.3.27+ version of the kernel supports large file sizes. I have 2.3.47 installed, and I've set the block size on ext2 to 4k, so with my kernel, i should be able to do files of 1TB.
However, many of the userspace programs in linux don't recognize a file that's larger than 4GB: for example, ls reports that my 16GB test file is 64MB.
What I'm wondering is whether Zope would have any problems with a large Data.fs on linux and/or any suggestions on how to test it.
I did some test when LFS support first appeared in the 2.3 kernel. Zope's FileStorage died at 4Gb (didn't let me to add more objects, the rest was working fine). Didn't had time to see what is the real problem.
I expect my Data.fs to grow to 26GB in size...
-Petru
participants (4)
-
ethan mindlace fremen -
Michel Pelletier -
morton@dennisinter.com -
Petru Paler