You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ambari.apache.org by "Prasanth Jayachandran (JIRA)" <ji...@apache.org> on 2017/06/15 09:11:00 UTC
[jira] [Updated] (AMBARI-21254) Alert if linux kernel changes are
not in place for LLAP
[ https://issues.apache.org/jira/browse/AMBARI-21254?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Prasanth Jayachandran updated AMBARI-21254:
-------------------------------------------
Description:
Following changes were made per [~rbalamohan]'s recommendation (count has been bumped up from 8K to 16K for somaxconn and backlog).
{code}
sudo sysctl -w net.core.somaxconn=16384
sudo sysctl -w net.core.netdev_max_backlog=16384
sudo sysctl -w net.ipv4.ip_local_port_range="4096 60999"
sudo sysctl -w net.ipv4.tcp_fin_timeout=10
sudo sysctl -w vm.swappiness=0
sudo sysctl -w vm.dirty_ratio=80
{code}
Since these configs has significant impact on shuffle, should this be enforced/alerted somehow to the user.
was:
Following changes were made per [~rbalamohan]'s recommendation (count has been bumped up from 8K to 16K for somaxconn and backlog).
{code}
sudo sysctl -w net.core.somaxconn=16384
sudo sysctl -w net.core.netdev_max_backlog=16384
sudo sysctl -w net.ipv4.ip_local_port_range="4096 60999"
sudo sysctl -w net.ipv4.tcp_fin_timeout=10
sudo sysctl -w vm.swappiness=0
sudo sysctl -w vm.dirty_ratio=80
{code}
Since these configs has significant impact on shuffle, should this be enforced/alerted somehow to the user. SmartSense has some rules for it. But Microsoft's cluster did not have SmartSense.
> Alert if linux kernel changes are not in place for LLAP
> -------------------------------------------------------
>
> Key: AMBARI-21254
> URL: https://issues.apache.org/jira/browse/AMBARI-21254
> Project: Ambari
> Issue Type: Bug
> Components: HiveServer2, Metastore, and Client Heap Sizes to Smart Configs
> Reporter: Prasanth Jayachandran
> Labels: llap
>
> Following changes were made per [~rbalamohan]'s recommendation (count has been bumped up from 8K to 16K for somaxconn and backlog).
> {code}
> sudo sysctl -w net.core.somaxconn=16384
> sudo sysctl -w net.core.netdev_max_backlog=16384
> sudo sysctl -w net.ipv4.ip_local_port_range="4096 60999"
> sudo sysctl -w net.ipv4.tcp_fin_timeout=10
> sudo sysctl -w vm.swappiness=0
> sudo sysctl -w vm.dirty_ratio=80
> {code}
> Since these configs has significant impact on shuffle, should this be enforced/alerted somehow to the user.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)