<div dir="ltr"><div>Hi.</div><div><br></div><div>Gluster 5.13 is running directly on VM. Got mount failed, endpoint is not connected in K8s. Sometimes connection refused of 1 of 3 Gluster nodes also. Reboot helps sometimes.<br></div><div><br></div><div>My idea is to optimize performance.</div><div><br></div><div>I have no idea.</div><div><br></div><div>These are the settings I found (from older versions?).</div><div><br></div><div>client.event-threads 8<br>cluster.lookup-optimize on<br>cluster.lookup-unhashed auto<br>cluster.quorum-type auto<br>cluster.read-hash-mode 0<br>cluster.readdir-optimize on<br>cluster.self-heal-daemon enable<br>cluster.server-quorum-type server<br>cluster.shd-max-threads 12<br>cluster.use-compound-fops on<br>dht.force-readdirp off<br>features.cache-invalidation on<br>features.cache-invalidation-timeout 600<br>network.inode-lru-limit 1000000<br>nfs.disable on<br>performance.cache-samba-metadata on<br>performance.cache-invalidation on<br>performance.cache-max-file-size 1GB<br>performance.cache-refresh-timeout 60<br>performance.cache-samba-metadata on<br>performance.cache-size 1GB<br>performance.client-io-threads on<br>performance.enable-least-priority off<br>performance.force-readdirp off<br>performance.io-cache off<br>performance.io-thread-count 16<br>performance.md-cache-timeout 600<br>performance.nl-cache on<br>performance.nl-cache-timeout 600<br>performance.open-behind off<br>performance.parallel-readdir on<br>performance.quick-read off<br>performance.rda-cache-limit 25MB<br>performance.read-ahead off<br>performance.readdir-ahead off<br>performance.stat-prefetch off<br>performance.strict-o-direct on<br>performance.write-behind off<br>performance.write-behind-window-size 16MB<br>server.allow-insecure on<br>server.event-threads 8<br>server.outstanding-rpc-limit 128<br>storage.fips-mode-rchecksum on<br>transport.address-family inet</div><div><br></div><div>I can add them like this to my storage class:<br><br>kind: StorageClass<br>apiVersion: <a href="http://storage.k8s.io/v1beta1">storage.k8s.io/v1beta1</a><br>metadata:<br>    name: gluster-heketi-external<br>provisioner: <a href="http://kubernetes.io/glusterfs">kubernetes.io/glusterfs</a><br>parameters:<br>    resturl: &quot;http://&lt;HOSTNAME&gt;:8080&quot;<br>    restuser: &quot;admin&quot;<br>    secretName: &quot;heketi-secret&quot;<br>    secretNamespace: &quot;kube-system&quot;<br>    # volumeoptions: &quot;ctime off&quot;<br>    # volumetype: replicate:2</div><div>    volumeoptions: &quot;key value key value key value&quot;</div><div><br></div><div>See last line.</div><div><br></div><div>Please help.</div><div><br></div><div>Best regards<br></div></div>