Alexey Serbin has uploaded this change for review. ( http://gerrit.cloudera.org:8080/18059
Change subject: [scripts] add hard memory limit for mini-cluster processes ...................................................................... [scripts] add hard memory limit for mini-cluster processes Before this patch, the start_kudu.sh script ran all the Kudu server processes at the same node where every kudu-tserver and kudu-master assumed it was the only Kudu process running there, so it would automatically self-impose the hard limit to be 80% of all the memory available. However, since there may be many of those, the mini-cluster could end up in an OOM condition. This patch addresses the issue, dividing all the available memory at a node between the processes the script starts, so they all use up to 80% of all available memory. I verified that the OOM condition that I saw when running a mini-cluster in a GCP VM no longer happens with this patch. Change-Id: I4de6c96fb0227554edbd5e69b29840f84e408326 --- M src/kudu/scripts/start_kudu.sh 1 file changed, 56 insertions(+), 8 deletions(-) git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/59/18059/1 -- To view, visit http://gerrit.cloudera.org:8080/18059 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-Project: kudu Gerrit-Branch: master Gerrit-MessageType: newchange Gerrit-Change-Id: I4de6c96fb0227554edbd5e69b29840f84e408326 Gerrit-Change-Number: 18059 Gerrit-PatchSet: 1 Gerrit-Owner: Alexey Serbin <[email protected]>
