Ulrich Eckhardt wrote:
> As for how to restore the repository to a working state, you simply
> reload revisions 0 to 126 from backup while using the live repository
> to dump the following revisions and then reload those as incremental
> changes. Alternatively, you simply load the backup up to where you
> have it and the rest from the live repository.
Thanks for the thoughts. Sadly the repository backups are already
corrupted: the problem arose in May 2006 and I first ran into it a few
Although I have 6 Months of backup archives, I did not dump the older
archives to DVDs. Stupid me.
I was quite confident having 6 Months of backups that went to an
external host, while the SVN repositories were on a local host...
Anyway, I started moving my repositories to an online host a few days
ago, and now I'll switch the configuration, have the backups go to a
local host, where I will be able to access and dump older archives to DVDs.
>> I'd like to restore this repository in a stable state, and restore a
>> fresh WC for a user (whose files I wiped away while doing multiple
>> concurrent operations in TortoiseSVN...)
> It should be possible to keep on using the same WCs as before.
Well, I am afraid not. This is another issue I didn't bother to raise,
as I did not search the list for an explanation.
If it is of interest, here is what happened: I cancelled Subversion's
operations (Cancel in TortoiseSVN), after I started accidentally
updating the same WC from two different levels: one of them, a root
folder, was importing other WCs thru the 'svn:externals' feature; while
one of its child folders, with the WC, was already updating.
When I canceled to top level importing folder 'SVN Update...', the local
WC of the 'lili' repository was left partially destructed, all '.svn'
folders had gone away, and only the latest unchecked files and folders
I thought of checking out a new fresh WC, copying over the files left,
and check them in. But then I ran into the checksum errors in the
Received on Thu Jan 11 20:37:47 2007