Michael Eischer
Michael Eischer
What does this PR change? What problem does it solve? ----------------------------------------------------- Most restic commands are long running and would have to start over in case of a failed backend operation....
What does this PR change? What problem does it solve? ----------------------------------------------------- Since pkg/sftp 1.13.0 files were uploaded sequentially using 32kb chunks instead of sending 64 chunks in parallel. This PR...
Output of `restic version` -------------------------- 0.15.2 What should restic do differently? Which functionality do you think we should add? ---------------------------------------------------------------------------------- As requested in https://forum.restic.net/t/snapshot-copy-and-corrupted-data-recovery/6378, restic should allow a larger maximum...
What does this PR change? What problem does it solve? ----------------------------------------------------- Currently it is possible to accidentally remove all snapshots using e.g. `forget --keep-tag invalid` where `invalid` is a not...
What does this PR change? What problem does it solve? ----------------------------------------------------- Based on https://github.com/restic/restic/pull/4800 and includes all its changes. This PR removes the `Backend()` method from the repository. Thereby it...
What does this PR change? What problem does it solve? ----------------------------------------------------- The PR cleans up how data is loaded from the backend and ensures that all places include a fallback...
What does this PR change? What problem does it solve? ----------------------------------------------------- Improve dump performance for large files by loading their blobs concurrently using `repo.Connections()` goroutines. The largest part of the...
What does this PR change? What problem does it solve? ----------------------------------------------------- Restic currently retries attempts to access a non-existing file or out of bound reads for truncated files. These errors...
The generic Kubernetes cloud-controller-manager implementation ignores nodes that are marked with the `ToBeDeletedByClusterAutoscaler` taint (see https://github.com/kubernetes/cloud-provider/blob/81e4f58b4d1badd71d633d356faaaf69d971d874/controllers/service/controller.go#L64). This PR let's yawol also ignore those nodes. In addition, the `Terminating` condition set...
See the individual commits for a detailed description. The changes are the result of a code review, but are unfortunately untested so far, as I don't have a mac and...