OE git servers
michael at yoctoproject.org
Tue Jun 18 21:10:37 UTC 2013
On 06/18/2013 01:47 PM, Khem Raj wrote:
> On Jun 18, 2013, at 12:41 PM, Jeff Osier-Mixon <jefro at jefro.net> wrote:
>> Hi all - it is one of my action items to investigate the possibility
>> of moving or changing the OE git servers to avoid the slow service
>> people experience sometimes. I believe I have a clue now about why
>> this is happening and how to avoid it.
>> Currently the OE git servers are available to anyone, and a lot of
>> normal users draw their nightly sync from master. This is different
>> from the way kernel.org, for example, handles things - they push
>> normal users off to mirrors and only let those with commit rights bang
>> on master.
> not true.
> I can still clone from git.ko.org even though I don't have push access to kernel.org
Actually it is true. Kernel committers do not use git.kernel.org as
their remote. They have a different address to push to. git.kernel.org
resolves to the DNS round-robin mirror list.
You'll notice that "ssh git.kernel.org" times out without a ssh handshake.
> and https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/
> published https/git/https ways to clone it.
>> Tom has told me we could solve that part of the issue by simply
>> disallowing http and git access, only allowing ssh, and by changing
>> the documentation to point users toward the mirrors (github). That
>> seems to me to be much easier to manage than changing servers.
>> I think it would be good for the TSC to discuss the issue of forcing
>> users to go to the mirrors. Meanwhile, I'll ask scottrif to make sure
>> the docs point to github instead of oe.org.
>> Jeff Osier-Mixon http://jefro.net/blog
>> Yocto Project Community Manager @Intel http://yoctoproject.org
>> tsc mailing list
>> tsc at lists.openembedded.org
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 4516 bytes
Desc: S/MIME Cryptographic Signature
More information about the tsc