You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Greetings! We are having a scylla cluster deployed using scylla operator on our self hosted kubernetes cluster and it works fine, but I am facing issues when creating backups using scylla manager, we are using Minio as a backup target with the below agent configurations
Looks like the scylla manager assumes that it's on AWS even after specifying provider: Minio flag in the agent config.
bash-4.2# scylla-manager-agent check-location -L s3:scylla-backups --debug
{"L":"DEBUG","T":"2020-11-23T05:39:39.920Z","N":"rclone","M":"AWS failed to fetch instance identity: Get http://169.254.169.254/latest/dynamic/instance-identity/document: net/http: request canceled while waiting for connection (Client.Timeout exceeded while awaiting headers)"}
{"L":"INFO","T":"2020-11-23T05:39:39.920Z","N":"rclone","M":"registered s3 provider [name=s3, upload_concurrency=2, provider=AWS, chunk_size=50M]"}
{"L":"INFO","T":"2020-11-23T05:39:39.920Z","N":"rclone","M":"registered gcs provider [name=gcs, chunk_size=50M]"}
{"L":"ERROR","T":"2020-11-23T05:40:39.922Z","N":"rclone","M":": error reading destination directory: NoCredentialProviders: no valid providers in chain. Deprecated.\n\tFor verbose messaging see aws.Config.CredentialsChainVerboseErrors"}
{"L":"INFO","T":"2020-11-23T05:40:39.923Z","N":"rclone","M":"S3 bucket cluster-backups: Waiting for checks to finish"}
{"L":"INFO","T":"2020-11-23T05:40:39.923Z","N":"rclone","M":"S3 bucket cluster-backups: Waiting for transfers to finish"}
FAILED: no providers - attach IAM Role to EC2 instance or put your access keys to s3 section of /etc/scylla-manager-agent/scylla-manager-agent.yaml and restart agent
The scylla pods are able to contact the minio cluster and the backup seems to work fine when uploading small system/default key-spaces like system_schema, system_traces etc, but fails when uploading non-system keyspaces having bigger datasize.
Below are the scylla manager errors produced during a backup task (sctool backup -c bda13859-e979-4d5f-852b-671cac448106 -L s3:scylla-backups --retention 5 --interval '24h'), as seen in the logs the backup fails for user_ks keyspace in which we have about 75G data.
Thank you @zimnx 👍🏿 As discussed, The backup task works fine when tried after increasing the resources allocated to the scylla-manager-agent sidecar container.
It would have been great if there is a way to specify the scylla-manager-agent resources in the cluster CRD, I will open an enhancement request for that. Thanks!
Greetings! We are having a scylla cluster deployed using scylla operator on our self hosted kubernetes cluster and it works fine, but I am facing issues when creating backups using scylla manager, we are using Minio as a backup target with the below agent configurations
Looks like the scylla manager assumes that it's on AWS even after specifying
provider: Minio
flag in the agent config.The scylla pods are able to contact the minio cluster and the backup seems to work fine when uploading small system/default key-spaces like
system_schema
,system_traces
etc, but fails when uploading non-system keyspaces having bigger datasize.Below are the scylla manager errors produced during a backup task (
sctool backup -c bda13859-e979-4d5f-852b-671cac448106 -L s3:scylla-backups --retention 5 --interval '24h'
), as seen in the logs the backup fails foruser_ks
keyspace in which we have about 75G data.Environment:
Any recommendations for fixing this would be much appreciated! 🙂 🙏🏿
The text was updated successfully, but these errors were encountered: