Description of problem:
Currently there is no option to turn off rollbackcopy sidecar container, which periodically takes snapshots of the etcd database and writes to disk. The action of snapshotting state itself is mostly non blocking but its a major I/O event.
When the database is huge (in Gbs), writing the huge data to the disk can impact other disk operations, and results in fsync delays in etcd metrics.
Version-Release number of selected component (if applicable):
All 4.4 and 4.5 have a sidecar implementation of periodic rollbackcopy
Steps to Reproduce:
1. Install 4.5
2. On leader node, notice snapshots being taken hourly into /etc/kubernetes/rollbackcopy directory
Snapshots are generated every hour on the leader node.
With option to turn off, the snapshots are no longer generated.
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory (Moderate: OpenShift Container Platform 4.5.21 bug fix and security update), and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.
*** Bug 1909586 has been marked as a duplicate of this bug. ***