Ryan Schmidt wrote:
> On Mar 20, 2006, at 16:27, Ralph Navarro wrote:
>
>>> The far likelier cause of file loss is operator error, not disk
>>> failure in my experience, and it's a lot easier to recover the files
>>> from a read-only repository.
>>
>> Interesting. Are you saying it is possible for an operator to
>> corrupt a
>> checked in version such that you can't use the main repository to get
>> the file? If so, is it more likely to happen with bdb or FSFS?
>
> All you need is one careless "rm -rf" by an administrator and it
> doesn't matter what kind of data you're talking about, it's gone, and
> your RAID won't help you there.
Especially with cron jobs: I've seen a core server get wiped out by a
carelessly written clean-up script that ran at 3:00 AM on Saturday night.
Fortunately for me, that happened two days *before* I started that job. But
having the on-line mirror has saved my butt on a number of occasions, such
as when I accidentally trashed a repository by bringing in new material with
the "svn ci file:///path-to-repository localdir" command.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Tue Mar 21 02:49:14 2006