[
https://issues.apache.org/jira/browse/KUDU-808?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Todd Lipcon resolved KUDU-808.
------------------------------
Resolution: Fixed
Assignee: Adar Dembo
Fix Version/s: n/a
This umbrella is no longer useful hanging around. We can file new work in this
area as new separate JIRAs.
> Throttle/reject writes to avoid running out of memory
> -----------------------------------------------------
>
> Key: KUDU-808
> URL: https://issues.apache.org/jira/browse/KUDU-808
> Project: Kudu
> Issue Type: Bug
> Components: client, tserver
> Affects Versions: Private Beta
> Reporter: Adar Dembo
> Assignee: Adar Dembo
> Priority: Blocker
> Fix For: n/a
>
>
> When faced with a write-heavy workload, Kudu should take action to prevent
> itself from running out of memory. This includes the following:
> # Providing operators with a mechanism to set a predefined memory limit, or
> choosing one based on a reasonable heuristic.
> # When the limit is reached, taking actions to reduce memory consumption, or
> to at least prevent it from rising further. These actions may include:
> ** throttling incoming writes
> ** rejecting writes
> ** performing maintenance operations that reduce memory consumption (e.g.
> flushing an MRS)
> A design document for this problem can be found
> [here|https://docs.google.com/a/cloudera.com/document/d/14PgHstfbU1qV2oKpKlsP4Z9YV8ueECA876Uvw52wbFg/edit?usp=sharing_eid].
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
