What are the problems caused by very large dentry slab? and How to use negative-dentry-limit sysctl parameter?
Issue
- Temporary unresponsiveness of container runtime.
- Performance issues of processes, commands and/or systemcalls working with files or directories.
- Slowness or even softlockups during unmount, reclaim and/or lookup routines.
- Large negative dentries can cause system unresponsiveness during systemd reloads
- OpenShift logs PLEG is not healty and NodeNotReady errors.
-
Use of the
curl
command in OpenShift liveness and/or readiness probes bloats the dentry cache. -
What is the
negative-dentry-limit
kernel parameter? - Need guidance to set the
negative-dentry-limit
kernel parameter.
Environment
- Red Hat Enterprise Linux 7
- OpenShift Container Platform
- SAP environments running on RHEL
crio
ordocker
container runtime
Note: Issues arising from overgrown dentry slab (namely due to negative dentries) can affect any Red Hat products or workloads utilizing RHEL kernel underneath. (i.e. not necessarily just RHOCP)
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.