Re: effectiveness of rsync and apt
Pierre Habouzit <madcoder@debian.org> writes:
> Le Lun 1 Mai 2006 15:31, Brian Eaton a écrit :
>> On 4/30/06, Goswin von Brederlow wrote:
>> > Look at zsync and help develope it far enough so it can look into
>> > debs. Without that the gain is practicaly 0 or less.
>>
>> It's entirely possible that the gain will be nothing no matter what
>> algorithm is used.
>>
>> The only time delta packages will be a win is for upgrades where the
>> client has the original package cached. If the client is installing
>> the package from scratch, delta packages are useless.
Or when the package is already installed. Delta packages should be
deltas to the contents of the package as oposed to a delta of the
compressed deb. Most people don't keep the debs in the cache and a
delta of the compressed data would be quite useless space wise.
> that's a good point, and I suppose that most of the stable traffic is
> due to first install of the packages. Though, I think it's not true
> for :
> - security mirrors (like pointed out in the thread already)
> - testing and unstable, where users do many upgrades per month or even
> per week (I think I do almost one per day â?? let's say 5 per week â??
> and I know a lot of people who do the same).
I agree. For stable delta package won't be worth it except for a short
while after the release.
> The real question is: do people clean their apt cache or not ? I do,
> because after a full X.org/kde/openoffice upgrade, it takes quite a lot
> of disk in /var (that is small on my computers). And with that cache
> cleaned, I fail to see how we could improve things much.
If the cache would be of use for the next upgrade then I'm sure more
people would not clean their cache.
> The mirrors replication could really benefit from that though.
MfG
Goswin
Reply to: