1 |
On Fri, Nov 30, 2012 at 10:57 AM, Richard Yao <ryao@g.o> wrote: |
2 |
> On 11/28/2012 11:08 AM, Matthew Thode wrote: |
3 |
>> On 11/28/2012 09:05 AM, Richard Yao wrote: |
4 |
>>> On 11/28/2012 09:17 AM, Maxim Kammerer wrote: |
5 |
>>>> On Wed, Nov 28, 2012 at 3:54 PM, Richard Yao <ryao@g.o> wrote: |
6 |
>>>>> We could slightly simplify the handbook installation procedure if we |
7 |
>>>>> told people to use emerge-webrsync to fetch the initial snapshot. |
8 |
>>>> |
9 |
>>>> Using emerge-webrsync also makes the installation process more robust, |
10 |
>>>> since it only requires HTTP access (whereas many firewalls restrict |
11 |
>>>> RSYNC). Besides, emerge-webrsync can check PGP signatures, so I think |
12 |
>>>> that it should be the primary recommended portage tree synchronization |
13 |
>>>> method. |
14 |
>>>> |
15 |
>>> |
16 |
>>> The only downside of which I am aware is increased network traffic. |
17 |
>>> However, we could redesign emerge-webrsync to take advantage of GNU |
18 |
>>> Tar's incremental archive functionality. |
19 |
>>> |
20 |
>>> That would permit us to mirror compressed diffs in addition to regular |
21 |
>>> portage snapshots. Doing that well could reduce bandwidth requirements. |
22 |
>>> |
23 |
>> weekly fulls and daily diffs? |
24 |
>> |
25 |
> |
26 |
> Determining what is right here probably requires calculus, but this |
27 |
> scheme does not seem like a bad choice to me. My main concern is that |
28 |
> maintaining weekly full snapshots would require too much space for the |
29 |
> mirrors. It might be better go monthly, with diffs on the following |
30 |
> intervals: |
31 |
> |
32 |
> 1 week |
33 |
> 1 day |
34 |
> 30 minutes |
35 |
> |
36 |
> Doing that would eliminate the benefit of rsync entirely, with the |
37 |
> caveat that we now need to mirror a ton of diffs. This would make it |
38 |
> easy for us to provide the ability to obtain historical snapshots, which |
39 |
> would be nice. |
40 |
|
41 |
Worth noting that all this moves us nicely in the direction of |
42 |
allowing HTTP proxies to cache data, reducing load on mirrors. And |
43 |
moves us in the direction of implementing mirrors themselves as a |
44 |
network of caching proxies. |
45 |
|
46 |
-- |
47 |
:wq |