Hi Nicolás, The missing history was some mistake I made during the conversion.
Note that git-fast-import produces very badly compressed repositories. I > recommend running 'git gc --aggressive' after converting. I usually do 'git repack -a -f -d'. What I don't understand is, it may say 'nothing to do'. Then I push the code up to the server (Bit Bucket in my case). If I go to where the repo is stored on the server, and then do 'git gc' followed by 'git repack -a -d -f', it repacks the whole repository. I am not sure why it repacks it on the server, and says 'nothing to do' on the client. Do you know if there is a way to force it repack the whole repo? It really helps with performance, I can shave 1GB off the size of the repository if I can get it to repack the whole thing. Also, I have one repository which is 5GB in size and imports in 20 minutes. Then I have one which is 3.4GB in size but takes two hours to import. I think the reason here is the number of objects. The second one has over 1 million, as it has 17 years of history. John On 15 September 2017 at 17:26, Nicolás Alvarez <nicolas.alva...@gmail.com> wrote: > > El 15 sept 2017, a las 13:19, John Lawlor <jk.law...@gmail.com> escribió: > > Hi, > > I have converted a large SVN repo to git using svn2git. The repo is 5G on > disk: > > [johnl@pluto:/u01/conversion/tools/bin ]$ du -sh VoltComponents/ > 5.4G VoltComponents/ > [johnl@pluto:/u01/conversion/tools/bin ]$ > > > Note that git-fast-import produces very badly compressed repositories. I > recommend running 'git gc --aggressive' after converting. > > When I push it up to the server, it is only pushing 1.5GB of data, and the > repository is missing two years of commits. > > I am not sure what is happening... > > > Does your converted repository have multiple branches? Are you pushing > them all? A 'git push' might only push the current branch... > > -- > Nicolás > Award-winning git repo surgeon (?!) >