Large repositores (hg or git)

Question

I'd like to use hg or git for repo with size in about 5gb. Are they work proper on repo with such a big size?

score 3 · Accepted Answer · answered Mar 18 '12 at 16:42

A repository with 5 GB is large, but not outrageously large. You don't say if the 5 GB is for the history alone, the working copy, or a combined figure?

For comparison, I can tell you that the OpenOffice repository has has a history of 2.0 GB (about 270,000 changesets) and a working copy of 2.3 GB. Mercurial runs okay on that size:

$ time hg status
hg status  0.63s user 0.26s system 99% cpu 0.886 total

With a cold cache is takes 2.4 sec — a little longer, but not too bad. There are 69,000 files in the working copy.

In general, you can expect both Git and Mercurial to slow down as the repository grows. The complexity of the operations differ: hg status is obviously O(number of files in the working copy), hg commit has the complexity of hg status plus O(number of changed files). A simple command like hg cat has O(1) complexity in the number of files and number of changesets — Mercurial can reconstruct any version of a file in constant time.

score 2 · Answer 2 · answered Mar 16 '12 at 14:50

2

I have never used hg but facebook recently sent a post about their git repo slowing down. I think their issue was all their "apps" are in one repo

http://thread.gmane.org/gmane.comp.version-control.git/189776

It's worth a read through for people trying to help them out

answered Mar 16 '12 at 14:50

Mike

22,310
7
56
79

1

Should this be a comment instead of an answer? – Bart Silverstrim Mar 16 '12 at 15:14

Large repositores (hg or git)

2 Answers2