Open Stack

Luca Milanesio <luca.milanesio at gmail.com> wrote: 
>>On 27 Mar 2019, at 15:46, Luca Milanesio <Luca.Milanesio at gmail.com> wrote: 
>>>On 27 Mar 2019, at 15:44, Thomas Dräbing <thomas.draebing at gmail.com <mailto:thomas.draebing at gmail.com>> wrote: 
>>>
>>>I suggested that some time ago, when reviewing this change: https://gerrit-review.googlesource.com/c/gerrit-ci-scripts/+/184730 <https://gerrit-review.googlesource.com/c/gerrit-ci-scripts/+/184730> . 
>>>The question just is, whether the openjdk 7 from the experimental repository is stable enough to be used by the CI. 
>>
>>Agreed. Best to just "tag" the current images for now than risking to have an unstable Java 7 build. 
>>
>>>
>>>Since we still have a year until jessy loses its LTS-status, I would rather try to use this year to try to move everybody to a newer Gerrit version, if possible. 
>>
>>That would be ideal, Java 7 is largely deprecated and unsupported. 
>
>$ curl https://review.openstack.org/config/server/version 
>)]}'
>"2.13.12-11-g1707fec" 
>
>Are the OpenStack guys happy to upgrade their Gerrit setup or move to GerritHub? 
>
>If we drop support for Gerrit v2.13, they'll lose any support from 
>the community for their setup. 

I can't "officially" represent the views of the folks in charge of the 
OpenStack infrastructure.  However I've spoken to them several times 
about Gerrit and to the best of my knowledge there is every intention 
to upgrade - I think it has been blocked by a lack of time, or (IIRC) 
perhaps by some difficulties encountered when testing the migration, 
or maybe a combination of both. 

I'll dare to cross-post this sub-thread to openstack-discuss, although 
I might be risking the wrath of the mailing list gods by doing so, so 
please don't take for granted that the cross-posting will work 
correctly :-)  I can also highly recommend joining #openstack-infra on 
Freenode IRC and chatting with the very helpful and friendly folks 
there. 

>OpenStack is a pretty big OpenSource project 

Yes, IIRC it's top 3 in terms of commit velocity these days, and 
possibly higher than the Linux kernel. 

>not supporting them anymore would be a problem I believe.

Good to hear - thanks a lot for thinking of us ;-)