Re: [DISCUSS] Create a nightly docker image

2022-03-22 Thread Enrico Olivelli
Il giorno mar 22 mar 2022 alle ore 15:21 Ziyao Wei ha scritto: > > Cool, thanks for the ideas! I agree it makes sense to separate the > nightlies from the other versions. > > Just curious what's the advantage of using Artifactory over using Docker > Hub? I read it has better access control but I

Re: [DISCUSS] Create a nightly docker image

2022-03-22 Thread Ziyao Wei
Cool, thanks for the ideas! I agree it makes sense to separate the nightlies from the other versions. Just curious what's the advantage of using Artifactory over using Docker Hub? I read it has better access control but I don't have any experience with it. Also, I could send an email out to

Re: [DISCUSS] Create a nightly docker image

2022-03-19 Thread Enrico Olivelli
Il Sab 19 Mar 2022, 06:02 Dave Fisher ha scritto: > The ASF has a JFrog artifactory instance that could help. +1 It must be clear that they are official releases and they MUST not be used in Production We could also publish snapshots of Maven jars to repository.apache.org Enrico Ask on

Re: [DISCUSS] Create a nightly docker image

2022-03-18 Thread Dave Fisher
The ASF has a JFrog artifactory instance that could help. Ask on bui...@apache.org! Regards, Dave Sent from my iPhone > On Mar 18, 2022, at 9:53 PM, PengHui Li wrote: > > I think we can use a separate org to maintain the nightly builded snapshot > images? > So that the external repo can

Re: [DISCUSS] Create a nightly docker image

2022-03-18 Thread PengHui Li
I think we can use a separate org to maintain the nightly builded snapshot images? So that the external repo can pull the latest image from the new org. Regards, Penghui On Sat, Mar 19, 2022 at 12:33 PM Ziyao Wei wrote: > Hi Pulsar developers, > > > Recently on Docker Hub the `latest` image

[DISCUSS] Create a nightly docker image

2022-03-18 Thread Ziyao Wei
Hi Pulsar developers, Recently on Docker Hub the `latest` image has been changed from 2.8.2 to 2.9.1. As a result, the Golang client CI is failing. I have been debugging the issue, but to help timely detection of similar issues in the future and to ease debugging, I propose we should create a