You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Jan 21, 2022. It is now read-only.
Hi folks -
My colleagues and I are running a few bosh-lite VM's in AWS to do some manifest validation before passing changes on to a "real" microbosh environment. Therefore, we have a few central bosh-lites running (one for each set of environments) that run multiple deployments. Currently, we're running into a problem where the Garden containers consume so much memory (running in a 32GB bosh-lite btw) that other Garden containers become unable to consume memory, and we have to restart the VM. In researching this problem, I was able to query Garden in bosh-lite (logged in as vcap) with the following command to get the memory limits - according to Garden, this limit is not being set - even if I specifically set it in the resource pools in any of the manifests:
Hi folks -
My colleagues and I are running a few bosh-lite VM's in AWS to do some manifest validation before passing changes on to a "real" microbosh environment. Therefore, we have a few central bosh-lites running (one for each set of environments) that run multiple deployments. Currently, we're running into a problem where the Garden containers consume so much memory (running in a 32GB bosh-lite btw) that other Garden containers become unable to consume memory, and we have to restart the VM. In researching this problem, I was able to query Garden in bosh-lite (logged in as vcap) with the following command to get the memory limits - according to Garden, this limit is not being set - even if I specifically set it in the resource pools in any of the manifests:
The command:
for i in $(curl -s http://127.0.0.1:7777/containers | jq -r '.handles[]'); do echo $(curl -s "http://127.0.0.1:7777/containers/$i/limits/memory" | jq -r '.limit_in_bytes'); done;
The output:
9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000 9223372036854772000
It has been confirmed in the cf-bosh forum that this is expected behavior. Would it be possible to add manifest-driven memory caps in the future?
So for example:
cloud_properties:
ram: 2048
This could help us diagnose a memory-hogging release, without it crashing all of bosh-lite.
Let me know if you need further details - Thanks!
Jeremy
The text was updated successfully, but these errors were encountered: