Comments on: Back Up (Some Of) That Local Data! http://scottbilas.com/blog/back-up-some-of-that-local-data/ Take what you want, and leave the rest (just like your salad bar). Tue, 09 Mar 2010 20:13:35 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: Simon Cooke http://scottbilas.com/blog/back-up-some-of-that-local-data/comment-page-1/#comment-75 Simon Cooke Mon, 12 Jan 2009 19:42:51 +0000 http://scottbilas.wordpress.com/2009/01/06/back-up-some-of-that-local-data/#comment-75 The way I designed a system like this to work for Surreal was to have a little C# app that would run daily (at about 2am). This would parse out a config file, which would tell the server which machines to backup, what file types it was allowed to backup, and where the backups should go. It would then go out to each machine (with admin privs) and access the LanMan C drive share on each machine (eg. \\dev-scooke\C$) - of course, having admin privs, it could poke around as much as it liked. On each machine in the root of C was another config file, specifying what the user wanted to copy (ie. which folders), and whether they wanted an email on success, on failure, and what their email address was. The server would parse this file, and then it would use Robocopy in backup mode to backup their files in low-network utilization mode. It would then RAR all those files up to a RAR file. It'd also keep historical backups for a few days. Worked quite well - except when people had immensely nested files and explorer would have issues reading folders with filename & path combos with more than 256 characters. (Which is weird, because NTFS supported it). The way I designed a system like this to work for Surreal was to have a little C# app that would run daily (at about 2am).

This would parse out a config file, which would tell the server which machines to backup, what file types it was allowed to backup, and where the backups should go.

It would then go out to each machine (with admin privs) and access the LanMan C drive share on each machine (eg. \\dev-scooke\C$) – of course, having admin privs, it could poke around as much as it liked.

On each machine in the root of C was another config file, specifying what the user wanted to copy (ie. which folders), and whether they wanted an email on success, on failure, and what their email address was.

The server would parse this file, and then it would use Robocopy in backup mode to backup their files in low-network utilization mode. It would then RAR all those files up to a RAR file.

It’d also keep historical backups for a few days. Worked quite well – except when people had immensely nested files and explorer would have issues reading folders with filename & path combos with more than 256 characters. (Which is weird, because NTFS supported it).

]]>