Dear lazyweb
Apr. 25th, 2012 03:08 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
armbI want a network filesystem that continues to work when the network goes away, and magically reconciles stuff when it comes back. (For use on a Linux laptop which mostly lives in a docking station, but not always.)
I've found http://www.coda.cs.cmu.edu/ and http://code.google.com/p/tsumufs/
Both of them look as if they are trying to solve exactly this problem, but both look like interesting research projects not really ready for production use. Am I wrong? Have I missed something?
If I fall back on "have a local copy, sync regularly with the server", any recommendations?
("Make it all a DVCS repository" (bazaar looks a front runner, git another strong possibility), and Unison are possibilities.)
(And, related, I want to be able to login while disconnected, but keep account information in step with the stuff currently using NIS.
Use "ypcat passwd" (while connected) and compare with local entries?
http://manpages.ubuntu.com/manpages/oneiric/man8/sync-accounts.8.html ?
Something else?)
I've found http://www.coda.cs.cmu.edu/ and http://code.google.com/p/tsumufs/
Both of them look as if they are trying to solve exactly this problem, but both look like interesting research projects not really ready for production use. Am I wrong? Have I missed something?
If I fall back on "have a local copy, sync regularly with the server", any recommendations?
("Make it all a DVCS repository" (bazaar looks a front runner, git another strong possibility), and Unison are possibilities.)
(And, related, I want to be able to login while disconnected, but keep account information in step with the stuff currently using NIS.
Use "ypcat passwd" (while connected) and compare with local entries?
http://manpages.ubuntu.com/manpages/oneiric/man8/sync-accounts.8.html ?
Something else?)
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
(no subject)
Date: 2012-04-27 09:30 pm (UTC)Git and its tools are lightning fast, having been designed to cope with the kernel source tree. hg is OK but sometimes chugs a bit, particularly with large repos (work has around 58k commits), and tortoisehg can be slow at times because of this.
A few years ago, at a previous employer, hg and git were the two front runners in the race, considered about equally powerful - the choice came down to the windows tools being better for hg at that time, though I gather that git may have caught up a bit.