{"thread":{"id":"8127","subject":"RFD Use git for off-site backups","startedAt":"2007-05-13T09:34:17Z","lastAt":"2007-05-13T23:42:54Z","messageCount":5,"participants":["Thomas Glanzmann","Johannes Schindelin","Jan Hudec","Shawn O. Pearce"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"42007","messageId":"20070513093417.GA18546@cip.informatik.uni-erlangen.de","threadId":"8127","inReplyTo":null,"subject":"RFD Use git for off-site backups","fromName":"Thomas Glanzmann","fromEmail":"thomas@glanzmann.de","sentAt":"2007-05-13T09:34:17Z","receivedAt":"2007-05-13T09:34:17Z","isPatch":false,"sender":{"key":"thomas@glanzmann.de","avatar":null},"body":"Hello,\nI am thinking about using git for doing differential backups of a whole\nmachines. And I would like to know what is necessary to obtain that. In\nsuch a scenario the machine to be backed up should have a local index\nfile but sends its objects directly to the remote machine via git-via-ssh\nor git protocol. Is it possible to store extra information like extended\nattributes / acls in git? I don't care about named pipes, sockets and\ndevice files for the moment.\n\n Thomas\n"},{"id":"42015","messageId":"Pine.LNX.4.64.0705131251370.6410@racer.site","threadId":"8127","inReplyTo":"20070513093417.GA18546@cip.informatik.uni-erlangen.de","subject":"Re: RFD Use git for off-site backups","fromName":"Johannes Schindelin","fromEmail":"johannes.schindelin@gmx.de","sentAt":"2007-05-13T11:16:51Z","receivedAt":"2007-05-13T11:16:51Z","isPatch":false,"sender":{"key":"johannes.schindelin@gmx.de","avatar":"https://avatars.githubusercontent.com/u/127790?v=4"},"body":"Hi,\n\nOn Sun, 13 May 2007, Thomas Glanzmann wrote:\n\n> I am thinking about using git for doing differential backups of a whole\n> machines. And I would like to know what is necessary to obtain that. In\n> such a scenario the machine to be backed up should have a local index\n> file but sends its objects directly to the remote machine via git-via-ssh\n> or git protocol.\n\nI think that this setup (have objects remotely) will kill performance \npretty quickly. You have to compare to other blobs in order to find \nreasonable deltas...\n\n> Is it possible to store extra information like extended attributes / \n> acls in git?\n\nThere has been talk about pre-commit scripts which put this information \ninto a file, and a post-checkout script to resurrect these metadata, but I \nhaven't seen functional versions of these scripts yet.\n\nCiao,\nDscho\n"},{"id":"42016","messageId":"20070513111841.GA24392@cip.informatik.uni-erlangen.de","threadId":"8127","inReplyTo":"Pine.LNX.4.64.0705131251370.6410@racer.site","subject":"Re: RFD Use git for off-site backups","fromName":"Thomas Glanzmann","fromEmail":"thomas@glanzmann.de","sentAt":"2007-05-13T11:18:41Z","receivedAt":"2007-05-13T11:18:41Z","isPatch":false,"sender":{"key":"thomas@glanzmann.de","avatar":null},"body":"Hello,\n\n> I think that this setup (have objects remotely) will kill performance\n> pretty quickly. You have to compare to other blobs in order to find\n> reasonable deltas...\n\nso true. I thought about that myself just after I sent the eMail.\n\n> There has been talk about pre-commit scripts which put this\n> information into a file, and a post-checkout script to resurrect these\n> metadata, but I haven't seen functional versions of these scripts yet.\n\nI thought there is a way to add information without loosing compability\nwith current git versions. But maybe I am just mistaken.\n\n Thomas\n"},{"id":"42019","messageId":"20070513123436.GE8983@efreet.light.src","threadId":"8127","inReplyTo":"20070513093417.GA18546@cip.informatik.uni-erlangen.de","subject":"Re: RFD Use git for off-site backups","fromName":"Jan Hudec","fromEmail":"bulb@ucw.cz","sentAt":"2007-05-13T12:34:36Z","receivedAt":"2007-05-13T12:34:36Z","isPatch":false,"sender":{"key":"bulb@ucw.cz","avatar":null},"body":"On Sun, May 13, 2007 at 11:34:17 +0200, Thomas Glanzmann wrote:\n> I am thinking about using git for doing differential backups of a whole\n> machines. And I would like to know what is necessary to obtain that. In\n> such a scenario the machine to be backed up should have a local index\n> file but sends its objects directly to the remote machine via git-via-ssh\n> or git protocol. Is it possible to store extra information like extended\n> attributes / acls in git? I don't care about named pipes, sockets and\n> device files for the moment.\n\nI don't actually think git is a good match for whole-machine backups. It's\nstrenghts are in distributedness and branching, but you don't need either of\nthose for large-scale backups (monitoring changes say in /etc is a different\nthing -- git is sensible match there -- see http://www.isisetup.ch/).\n\nI can suggest two solutions:\n\n - Use CVS on the client (machine to back up) and git-cvsserver on the\n server. You will probably be able to find some scripts that do backups to\n CVS. On server you will still have git for digging in history. Git is also\n very space efficient.\n\n CVS does not cache any data on the client. It only keeps metadata to\n detect which files were modified. Therefore it will have minimal overhead\n in storage, but needs to transfer the modified files in their entirety.\n\n - Use Subversion. There is summer of code project to create git-svnserver,\n so you might be able to use git on the server-side in future with\n subversion client too. But for backups you should not need any of the\n version control features subversion does not have, so subversion should be\n suitable.\n\n Subversion stores copy of the previous revision on the client, so all data\n under it's control will take twice as much space on client, but it will\n only transfer deltas over the network.\n\n-- \n\t\t\t\t\t\t Jan 'Bulb' Hudec \n"},{"id":"42070","messageId":"20070513234254.GN3141@spearce.org","threadId":"8127","inReplyTo":"20070513123436.GE8983@efreet.light.src","subject":"Re: RFD Use git for off-site backups","fromName":"Shawn O. Pearce","fromEmail":"spearce@spearce.org","sentAt":"2007-05-13T23:42:54Z","receivedAt":"2007-05-13T23:42:54Z","isPatch":false,"sender":{"key":"spearce@spearce.org","avatar":"https://avatars.githubusercontent.com/u/34844?v=4"},"body":"Jan Hudec wrote:\n> - Use Subversion. There is summer of code project to create git-svnserver,\n> so you might be able to use git on the server-side in future with\n> subversion client too. But for backups you should not need any of the\n> version control features subversion does not have, so subversion should be\n> suitable.\n\nUnfortunately Google is not sponsering the git-svnserver summer of\ncode project. We didn't get enough slots for us to get Google to\npay for that one. At least one of the students has still expressed\ninterest in working on it, but it won't be their full-time job this\nsummer. :-(\n\n-- \nShawn.\n"}]}