2

Can someone please educate me, when i run a full dump on our repo using the following command

svnadmin dump /repo > backup.dump

This dump file is around 90GB

if i do a dump file with the following command on the same repo

svnadmin dump /repo -r 21782:HEAD > backup.dump

This dump file comes out as 152GB

Even though i have missed out the first 20,000 revisions this is almost twice as big, how is this even possible.

I would of thought it would have been smaller, is this something to do with compression?

If you dump using revisions its uncompressed?

bahrep
  • 29,961
  • 12
  • 103
  • 150
Brian Mitchell
  • 151
  • 3

2 Answers2

2

the first revision is always dumped out 'fully expanded', subsequent revisions are dumped as deltas. This means the data in rev 21782 is quite large that, if you have dumped rev 21781:HEAD you might see it smaller.

If the first revision was dumped as a delta then you wouldn't be able to recreate it from your dumps! (obviously)

gbjbaanb
  • 51,617
  • 12
  • 104
  • 148
0

This is basically a duplicate of Why is a SVN dump of a single revision larger than a full dump? which explains that the first revision of svn dump, by default, contains a full representation of the entire repository to ensure that the revisions exported (in your case, r21782 to HEAD) can each be committed successfully.

That post also mentions the way around this (the --incremental option.)

Community
  • 1
  • 1
Johann
  • 4,107
  • 3
  • 40
  • 39