

0 -http /Users/andreas.schafferer/Library/Group Containers//http_proxy.json -dhcp /Users/andreas.schafferer/Library/Group Containers//dhcp.json -port-max-idle-time 300 -max-connections 2000 -gateway-ip 192.168. int ernal -vm-names docker- for -desktop -listen-backlog 32 -mtu 1500 -allowed-bind-addresses 0.0. mac.localhost -gateway-names gateway.docker. 0 /node_modules/dockerfile-language-server-nodejs/lib/server.js -node-ipc -node-ipc -clientProcessId= 693 2377 0 10 : 09 AM ? 0 : 00.12 /Applications/Docker.app/Contents/MacOS/ -watchdog fd: 4 53333 0 10 : 09 AM ? 0 : 00.06 serve -address fd: 3 -connect vms/ 0 /connect -control fd: 4 -log-destination aslĥ01 53335 53333 0 10 : 09 AM ? 0 : 03.38 -ethernet fd: 3 -port fd: 4 -diagnostics fd: 5 -pcap fd: 6 -vsock-path vms/ 0 /connect -host-names host.docker. 0 /node_modules/vscode-languageclient/lib/utils/electronForkStart /Users/andreas.schafferer/.vscode/extensions/peterjausovec.vscode-docker -0.3.

Output from ps -Af | grep docker from running the image locally withĭocker run -v tmp:/app/tmp -it 0 2393 1 0 Mon09AM ? 0 : 00.01 /Library/PrivilegedHelperTools/ĥ01 7020 6935 0 Mon01PM ? 0 : 04.47 /Applications/Visual Studio Code.app/Contents/Frameworks/Code Helper.app/Contents/MacOS/Code Helper /Users/andreas.schafferer/.vscode/extensions/peterjausovec.vscode-docker -0.3. After seeing this question, I tried storing the downloads and loading the model from an empty-dir volume, without any effect.Ĭould anybody shine a light on what’s going on here? Edit: I specifically rm the downloaded models after the download. When running in the cluster, the memory usage grows approximately 1gb per download, and never falls back down. When running the application without docker locally, it uses about 2gb of memory, burst to about 3 during the download, and drops back to 2gb. I have a pod which downloads a new machine learning model every 5 minutes, which causes the memory to grow indefinitely, and the pod is OOMkilled.
