Upgrading from <GD.CN> 3.0.0 to 3.1.0, we're seein...
# gooddata-cn
Upgrading from GD.CN 3.0.0 to 3.1.0, we're seeing the following error from the helm upgrade:
Copy code
time="2023-12-15T22:52:53Z" level=fatal msg="etcdserver: request is too large"
But when we revert the chart back to 3.0.0, the error goes away. Any suggestions on what we can do to get around this? Or please let us know if additional logs would be helpful in debugging this. Thanks! 😀
@Pete Lorenz I've never seen such error before. Would it be possible to tell us if this error comes from GD.CN etcd (etcd which is part of GD.CN installation, you should see it running in some pods) or this comes directly from kubernetes API/etcd (etcd which is part of kubernetes control plane)
Good question. It seems the issue would be during installation because the artifacts are never deployed. Note that we are using Rancher to deploy the gooddata helm chart. The message we're getting, apparently from helm, is "etcdserver: request is too large". I'm looking for more error messages ...
This was a Rancher problem caused by the size of the git repo that contains our GD.CN helm chart, which was too large because we were keeping previous (unneeded) versions of the chart in the repo. Removing one of the older versions resolved the issue.
👍 1