On 08/10/14 09:56, Jens Jensen wrote:
> On 08/10/2014 09:44, Christopher J.Walker wrote:
>> The Dirac instance at Imperial isn't production quality yet, but I
>> think that is the way we should be moving.
> Yes, although we might want to have a bit more experience with small VOs
> using DIRAC - I believe outside the UK there's quite a few but we might
> need a bit more experience with supporting small VOs on DIRAC ourselves.
>>> https://www.gridpp.ac.uk/wiki/Wider_VO_issues
> Yes, I am also beginning to think that the KeyDocs are, er, key
> documents. I mean having keydocs up to date would certainly be a good
> start and may even be sufficient.
>> QMUL turned on IPv6 on one of our two GridFTP nodes at 13:00 on Monday.
>>
> Interesting: so there is a way for it to communicate an IPv6 address in
> the protocol?
Apparently so.
There was an issue in logging - which is now fixed in epel-testing.
Duncan has discovered an issue with setting the appropriate flags to use
IPv6 if one end doesn't support IPv6 - I'll let him fill in the details.
> As I understood the half IPv4 socket was communicated as
> aaa,bbb,ccc,ddd,eee,fff, meaning aaa.bbb.ccc.ddd port ((eee << 8) | fff)
> Maybe I just need to read the documentation (aka source code).
>
> Thanks
> --jens
>
|