{"thread":{"id":"4452","subject":"Redhat stateless Linux and git","startedAt":"2006-06-09T22:59:44Z","lastAt":"2006-06-11T15:07:57Z","messageCount":3,"participants":["Jon Smirl","Geert Bosch"],"isPatch":false,"patchVersion":null,"patchTotal":null},"messages":[{"id":"21507","messageId":"9e4733910606091559m6a88e864m16f9d75a507ee684@mail.gmail.com","threadId":"4452","inReplyTo":null,"subject":"Redhat stateless Linux and git","fromName":"Jon Smirl","fromEmail":"jonsmirl@gmail.com","sentAt":"2006-06-09T22:59:44Z","receivedAt":"2006-06-09T22:59:44Z","isPatch":false,"sender":{"key":"jonsmirl@gmail.com","avatar":"https://gravatar.com/avatar/cff3bf5bfdfa6708b905712ff91f0f9b8aaca161659f38c02b787920d5d28b7e?d=mp&s=160"},"body":"Redhat is looking for a scheme to sync the disk system of their\nstateless Linux client. They were using rsync and now they are looking\nat doing it with LVM.\n\nWhat about using git?\n\n-- \nJon Smirl\njonsmirl@gmail.com\n"},{"id":"21593","messageId":"D5AC73C4-5A2F-482E-9B45-71A72C62D670@adacore.com","threadId":"4452","inReplyTo":"9e4733910606091559m6a88e864m16f9d75a507ee684@mail.gmail.com","subject":"Re: Redhat stateless Linux and git","fromName":"Geert Bosch","fromEmail":"bosch@adacore.com","sentAt":"2006-06-11T12:21:19Z","receivedAt":"2006-06-11T12:21:19Z","isPatch":false,"sender":{"key":"bosch@adacore.com","avatar":null},"body":"\nOn Jun 9, 2006, at 18:59, Jon Smirl wrote:\n> Redhat is looking for a scheme to sync the disk system of their\n> stateless Linux client. They were using rsync and now they are looking\n> at doing it with LVM.\n>\n> What about using git?\n\nThe data model is fine in principle, but git as-is isn't suitable\nfor general backup/sync-like schemes. Large (multi-GB) files\nare not really supported yet. Still, I think the underlying\ndata model, with some modifications to split large files on\ncontent-determined boundaries, would be really great for\ndistributed filesystems.\n\nMany people using laptops these days connect to different\nfilesystems on their office networks, home networks,\ndigital cameras and even their PDA, cellphone and MP3-player.\nWhat is commonly described as \"synching\", really is just a\nmerge between different branches. All arguments in favor\nof using a distributed SCM hold here too.\n\nRight now I'm using a hodge-podge of different manual and\nsemi-automated methods to keep my local filesystem with 1.5M\nfiles totalling 90GB somewhat in synch with various\nhomedirectories on different remote systems and backup disks.\nIMO, git is tantalizing close to be able to handle this, just\nneeds to get a bit more scalable. Probably you'd want to use\na different user interface as well, but all the underlying\ndata structures and merge strategies may be equally valid.\n\n   -Geert\n"},{"id":"21597","messageId":"9e4733910606110807m41fe82er9c4876a88336209c@mail.gmail.com","threadId":"4452","inReplyTo":"D5AC73C4-5A2F-482E-9B45-71A72C62D670@adacore.com","subject":"Re: Redhat stateless Linux and git","fromName":"Jon Smirl","fromEmail":"jonsmirl@gmail.com","sentAt":"2006-06-11T15:07:57Z","receivedAt":"2006-06-11T15:07:57Z","isPatch":false,"sender":{"key":"jonsmirl@gmail.com","avatar":"https://gravatar.com/avatar/cff3bf5bfdfa6708b905712ff91f0f9b8aaca161659f38c02b787920d5d28b7e?d=mp&s=160"},"body":"On 6/11/06, Geert Bosch <bosch@adacore.com> wrote:\n>\n> On Jun 9, 2006, at 18:59, Jon Smirl wrote:\n> > Redhat is looking for a scheme to sync the disk system of their\n> > stateless Linux client. They were using rsync and now they are looking\n> > at doing it with LVM.\n> >\n> > What about using git?\n>\n> The data model is fine in principle, but git as-is isn't suitable\n> for general backup/sync-like schemes. Large (multi-GB) files\n> are not really supported yet. Still, I think the underlying\n> data model, with some modifications to split large files on\n> content-determined boundaries, would be really great for\n> distributed filesystems.\n>\n> Many people using laptops these days connect to different\n> filesystems on their office networks, home networks,\n> digital cameras and even their PDA, cellphone and MP3-player.\n> What is commonly described as \"synching\", really is just a\n> merge between different branches. All arguments in favor\n> of using a distributed SCM hold here too.\n>\n> Right now I'm using a hodge-podge of different manual and\n> semi-automated methods to keep my local filesystem with 1.5M\n> files totalling 90GB somewhat in synch with various\n> homedirectories on different remote systems and backup disks.\n> IMO, git is tantalizing close to be able to handle this, just\n> needs to get a bit more scalable. Probably you'd want to use\n> a different user interface as well, but all the underlying\n> data structures and merge strategies may be equally valid.\n\nThat's why I though stateless Linux was a good place to start. The\nclient is read only so it is the simplest case to start with. I would\nmuch prefer a file orientated system for syncing over a block oriented\none, with the block one there is no easy way to tell what is being\ncopied to your machine.\n\nI added the stateless list to the cc, maybe they'll join in.\n\n-- \nJon Smirl\njonsmirl@gmail.com\n"}]}