1 |
Neil Bothwick wrote: |
2 |
> On Tue, 05 Apr 2011 21:38:06 +1000, Jake Moe wrote: |
3 |
> |
4 |
> |
5 |
>>> for the distfiles http-replicator is great. Example: |
6 |
>>> Box one has the tarball. Box 2, 3 and 4 can download it as well. Box |
7 |
>>> one does not have the tarball. Box 2 may search for it, box one will |
8 |
>>> download it and give it to box 2. Then box 3 and 4 can download it |
9 |
>>> from box one as well. |
10 |
>>> |
11 |
> |
12 |
>> Well, I suppose I'll have to look into that then, won't I? Thanks for |
13 |
>> the info. |
14 |
>> |
15 |
> Alternatively, use an NFS mount as $DISTDIR and you don't end up with |
16 |
> copies of the tarballs on every machine on the network. |
17 |
> |
18 |
> |
19 |
|
20 |
I did that once and I don't recall having any problems. I had 4 rigs |
21 |
plus a really slow dial-up connection. When I found out about |
22 |
http-replicator and rsync, I started using it. I couldn't tell any |
23 |
difference tho. |
24 |
|
25 |
If you do use NFS, I think you have to run emerge --metadata on the |
26 |
secondary machines. That may have changed since it has been a while. |
27 |
|
28 |
Dale |
29 |
|
30 |
:-) :-) |