Looking at the dockerhub page know, all 1.13.1 images are listed as
amd64, with the last push being 5 days ago.
Perhaps some error happened during the rebuild, which has since been
corrected?
On 10/08/2021 00:43, Oran Shuster wrote:
Our 1.13.1 k8s deployments started failing with
tandard_init_linux.go:228: exec user process caused: exec format error
Not understanding how an old image with an old depoyment starts failing like
that i looked at the apache/flink dockerhub page.
Although 1.13.1 was released more than a month ago, all the images (apart from
1.13.1-scala_2.11-java11) were pushed 6 days ago and their architecture changed
from linux/amd64 to
linux/arm64
Luckily the "official" flink images (that were also pushed 5-6 days ago) are
still amd64 and can be used
