I think it was more on how docker-compose persist volumes and the lack of proper setup on my side. It was a pretty standard docker-file for development. Please find it attached.
On the good side, this incident raised my attention to some steps I should plan ahead for upgrades in production since we use the images in our ECS clusters. Cheers,. Em seg., 16 de mar. de 2020 às 11:51, Aldrin Piri <[email protected]> escreveu: > Thanks for following up. The issue you ran into, was it with the default > volumes as provided by our image or with additional volumes you specified? > Curious if there is anything we should rethink on this front to help > alleviate such issues and ease transitions. > > aldrin > > On Mon, Mar 16, 2020 at 9:46 AM Eric Chaves <[email protected]> wrote: > >> Hi Aldrin, >> >> Ok, I think I got my error. I'm using a docker-compose to run a local >> development cluster and it was first created using the 1.8 image. To >> upgrade to 1.11.3 I first removed the containers (using docking-compose >> down) but without informing the "-v" flag to remove volumes and then >> launched the containers again (docker-compose up) but since the volume was >> persisted it seems the nifi components was still at 1.8 (at least that was >> the version showing in the UI). Once I fully removed the containers and >> volumes (docker-compose down -v) and restarted it came with 1.11 version >> and the error was indeed gone. >> >> Sorry for the inconvenience. >> >> Cheers, >> >> >> >> Em dom., 15 de mar. de 2020 às 21:26, Aldrin Piri <[email protected]> >> escreveu: >> >>> Hi Eric, >>> >>> I can only speak to the image, but as it is tagged in 1.9.0, the >>> associated code will also be in what the community offers through Docker >>> Hub. We make use of the released convenience binaries for constructing the >>> image. >>> >>> Some things that would help us investigate further: >>> Could you please share the error you are encountering? >>> Have you verified the correct functionality for your setup with a >>> standalone NiFi instance? >>> How are you running your Docker containers? Do you have these on the >>> same Docker network? If you are using compose, this should be taken care >>> of by default, and if so, could you please share your compose file? >>> >>> >>> >>> >>> >>> On Sun, Mar 15, 2020 at 7:42 PM Eric Chaves <[email protected]> wrote: >>> >>>> Hi folks, >>>> >>>> I'm facing an error using Redis DMC in standalone mode with an external >>>> Redis server were the RedisConnectionPoolService is not able to connect to >>>> a non-localhost redis in this mode. >>>> >>>> This error was already reported ( >>>> https://issues.apache.org/jira/browse/NIFI-5830) and marked fixed in >>>> nifi 1.9 however I'm trying the nifi docker images from docker hub and all >>>> images that I've tried kept throwing the same error (I've tried the >>>> 1.9.0, 1.9.1 and the latest 1.11.3). >>>> >>>> In order to confirm if the bug I'm facing seems to be the same I did a >>>> port forward using socat between localhost (nifi container) and my redis >>>> container and the error was gone, >>>> >>>> Am I missing something? Why I'm getting this error if it's fixed? >>>> >>>> Best regards, >>>> >>>> Eric >>>> >>>
docker-compose.yaml
Description: application/yaml
