Newbie question on what is transfered on a PUSH

Asked by Julian Knight

Sorry if this seems obvious to people here but I've not really found the answer in the documentation. Hope I didn't miss it.

I would like to understand how much information is transferred when I do a PUSH from a tree that has a single BINARY file that has some modification.

For example, if I were to have an image file with some embedded meta-data that had already been committed & pushed and then I made a small change to the wording of the meta-data, recommitted and then re-pushed. Would bzr push the equivalent of the whole file size again? More? Less?

This would be over a bzr+ssh connection.

The reason I am asking is that I am starting to use bzr to manage my working documents folders. These contain a mixture of file types, mainly Word, Excel, etc with a few images and smallish ZIP files.

I can, of course, then RSYNC the managed tree with a remote server but it would seem more sensible to be able to just PUSH it and cut out any possible RSYNC issues. But to do that, I need to understand whether it will be feasible over some of the slow, unreliable network connections I have to put up with on customer sites.

Many thanks for any help you can give.

--
Regards,
Julian Knight, http://it.knightnet.org.uk

Question information

Language:
English Edit question
Status:
Answered
For:
Bazaar Edit question
Assignee:
No assignee Edit question
Last query:
Last reply:
Revision history for this message
Launchpad Janitor (janitor) said :
#1

This question was expired because it remained in the 'Open' state without activity for the last 15 days.

Revision history for this message
Martin Pool (mbp) said :
#2

reopen

Revision history for this message
Martin Pool (mbp) said :
#3

Hi Julian, and sorry for the delay.

Precisely how much data is transferred will depend on the repository format and the version of bzr on both ends. It's improving all the time.

Generally speaking the worst we will ever do is to send a full copy of any modified files. We will often do better and send data proportional only to the changes within the modified file, ie converging on the speed of rsync.

A larger problem on very flakey connections may be that bzr will currently cancel the entire current "write group" transaction if the connection drops. We'd like to do better and resume it over a new ssh connection. In some cases you can eg run openvpn and get the illusion of a more reliable connection.

Can you help with this problem?

Provide an answer of your own, or ask Julian Knight for more information if necessary.

To post a message you must log in.