1 |
On Thu, 8 Apr 2021 20:22:25 +0300 Joonas Niilola wrote: |
2 |
> |
3 |
> |
4 |
> On 4/8/21 7:37 PM, Alec Warner wrote: |
5 |
> > |
6 |
> > It's admittedly a grey area here. We use CDNs for various web |
7 |
> > components (packages.gentoo.org for example) and we use a CDN for |
8 |
> > distfiles.gentoo.org. Is Github simply a CDN for gentoo.git? Its an |
9 |
> > open question we have discussed on the gentoo-nfp list. |
10 |
> > |
11 |
> > In general I'm not really sold on the benefits of git as a replication |
12 |
> > protocol; while I dislike running a global rsync network the |
13 |
> > maintenance of the network is basically nil from infra's end and so I |
14 |
> > don't feel significant pressure to move to git. Could you perhaps |
15 |
> > articulate why you think it's important for clients to move to git? |
16 |
> > |
17 |
> > -A |
18 |
> > |
19 |
> |
20 |
> It provides much better user experience with continuously faster |
21 |
> sync-times. Also there's error posts frequently in the forums when using |
22 |
> rsync, even recently. |
23 |
|
24 |
Only if you have fast and reliable uplink. Try to terminate sync |
25 |
connection and resume in cycle. Rsync will do the job and recover, |
26 |
so users will be able to sync on poor connection; git will start |
27 |
over again and again, and again… Git can't continue failed |
28 |
transfer, it will restart it from scratch. Second try will be |
29 |
likely faster due to server-side cache, but a bulk transfer itself |
30 |
can't be resumed, only repeated. |
31 |
|
32 |
Try to fetch historical git using git clone by the way. |
33 |
|
34 |
Best regards, |
35 |
Andrew Savchenko |