Kjetil Torgrim Homme wrote:
> Jesper Krogh <jes...@krogh.cc> writes:
> 
>> I have a feature request, that is more a "usabillity issue" than
>> anything else. Should be trivial to implement if people think it is
>> equally important as I do.
>>
>> Item ?:      On restore, check space on target volume
> 
> no, it's not trivial to do, at least not accurately, since the restore
> job can span filesystems.  an approximation may be good enough, though.
> e.g., take the sum of all free space in filesystems at or below the
> restore location.  even if this will be optimistic at times, it will
> probably help in many cases.

It could, simply just do it at "write to target-filesystem time". A
small configuration option that would force it to bail out when there
was some threshold is met and not when the filesystem went full. So:

1) Check stats for filesystem
2) if threshold is met, bail out
3) write file.

statfs calls are not that expensive, are they? Otherwise it can be
optimized a bit to reduce the amount of calls.

I do admit that a "really nice" solutions would be very hard to
implement in all cases. But I'm just looking for "not to render my
systems broken by a restore", and above would do that.

Jesper
-- 
Jesper

------------------------------------------------------------------------------
Return on Information:
Google Enterprise Search pays you back
Get the facts.
http://p.sf.net/sfu/google-dev2dev
_______________________________________________
Bacula-devel mailing list
Bacula-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-devel

Reply via email to