在网络驱动器上存储 Svn Repositories 文件夹
是否有人将其存储库文件夹存储在已备份的网络驱动器上？与将其存储在运行 Subversion 的 C 驱动器的主机服务器上相比，是否有人遇到过任何问题？
我们本质上希望将存储库放在另一个网络驱动器上（与运行 VisualSVN Server 的服务器不同），并让 VisualSVN Server 从完全不同的计算机读取存储库。
我们每 5 分钟备份一次 VisualSVN 服务器指向的位置，以便在发生故障时随时备份 SVN 存储库文件夹。那么我们很可能不必担心将开发人员的任何最新更改合并回陈旧/过时的回购编号，因为我们经常备份它。
Does anyone store their Repositories folder on a network drive that's backed up? Is there any issues anyone's had with that as opposed to lets say keeping it stored on the host server that is running Subversion's C drive?
We essentially want to put our repo on another network drive (different than the server running VisualSVN Server) and have VisualSVN Server read the repository from a completely different machine.
We'd backup the location we point the VisualSVN Server to every 5 minutes so that the SVN repository folder would be backed up at any time a failure occurs. Then we most likely would not have to worry about then merging any latest changes from developers back into a stale/outdated repo number because we back it up so frequently.
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（5）
Subversion FAQ 有关于在 NFS 上存储存储库的讨论；基本上，在这种情况下您必须使用 fsfs 后端。使用fsfs对于存储库的热备份也是必要的；对于 bsdfs，您需要 hot-backup.py （或 svnadmin hotcopy）。使用常规备份程序时，“当前”文件可能会指向备份时尚未完全复制的修订版。如果您必须恢复发生这种情况的备份（或者确保在任何修订文件之前备份“当前”文件），“svnadmin recovery”可以解决这个问题。
The subversion FAQ has a discussion on storing repositories on NFS; basically, you must use the fsfs backend in this case. Using fsfs is also necessary for hot backups of the repository; for bsdfs, you would need hot-backup.py (or svnadmin hotcopy). With a regular backup program, it may happen that the "current" file points to a revision that is not yet fully copied at the time of backup. "svnadmin recover" can fix that should you have to restore a backup where this happened (or else make sure that "current" is backed-up before any of the revision files are).
希望您使用的是 file:// 协议之外的其他！
如果备份发生在非工作时间，一般来说应该没问题。
查看存储库备份< /a> SVN 书中的主题。
Hopefully, you are using something other than the file:// protocol!
It should be ok in general if the backups are occurring during off hours.
Check out the Repository Backup topic in the SVN Book.
我把我的仓库放在定期备份的网络服务器上，并为 apache 使用 mod-svn。
I put my repo on web server which is backed-up regularly and used mod-svn for apache.
我们曾经像这样设置过存储库，但将存储库移至服务器，并且在备份服务器之前，我们执行转储并备份转储文件。我们使用转储，因为文件格式将始终受到支持，并且存储库可以从无到有恢复，而热复制取决于存储库结构是否相同
We had our repositories set up like this at one point, but moved the repositories to a server, and before the server is backed up, we perform a dump and backup the dump file. We use dump, since the file format will always be supported and the repository can be recovered from nothing, where as hotcopy depends on the repository structure being the same
请注意，除非 svnserver 已停止，否则存储库的简单原始副本不能保证为您提供有效的备份。每 5 分钟停止一次服务器进行备份可能会很烦人。
确保您使用的是
svnadmin hotcopy，不仅仅是 SAN 快照或复制命令。此外，您还可以在
post commit挂钩中运行脚本，以便在实际发生更改时进行备份。Be careful, a simple naive copy of the repository is not guaranteed to give you a valid backup unless the svnserver has been stopped. It may be annoying to stop your server every 5 minutes for a backup.
Make sure you're using
svnadmin hotcopy, not just a SAN snapshot or copy command.Also, you can run scripts in the
post commithook to do backups when something's actually changed.