You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@nifi.apache.org by "Paul Yang (Jira)" <ji...@apache.org> on 2020/01/10 01:24:00 UTC
[jira] [Updated] (NIFI-7005) nifi process maybe crash while call
controller-services api DBCPConnectionPoolLookup associate a great many
jdbc pool
[ https://issues.apache.org/jira/browse/NIFI-7005?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Paul Yang updated NIFI-7005:
----------------------------
Description:
have a DBCPConnectionPoolLookup associate a great many (for our environment 170+) DBCPConnectionPool configurations.
nifi.properties:
nifi.cluster.node.read.timeout=600 sec
nifi.cluster.node.connection.timeout=600 sec
nifi.ui.autorefresh.interval=300 sec
While call controller-services rest api from anyway included clicking ProgressGroup's GUI configure menu, nifi node that called one's CPU and memory surge even process crash.
Fortunately, the jdbc lookup service's function is unaffected. But for the amazing function the behavior should be perfect.
The root cause is: 'controller-services' api return json object is so large, the json string size is even more than 1GB or more then cannot be observed.and the nifi process cannot handle so large size response. the DBCPConnectionPool's information referenced DBCPConnectionPoolLookup's information and vice versa.
was:
have a DBCPConnectionPoolLookup associate a great many (for our environment 170+) DBCPConnectionPool configurations.
nifi.properties:
nifi.cluster.node.read.timeout=600 sec
nifi.cluster.node.connection.timeout=600 sec
nifi.ui.autorefresh.interval=300 sec
While call controller-services rest api from anyway included clicking ProgressGroup's GUI configure menu, nifi node that called one's CPU and memory surge even process crash.
The root cause is: 'controller-services' api return json object is so large, the json string size is even more than 1GB or more then cannot be observed. the DBCPConnectionPool's information referenced DBCPConnectionPoolLookup's information and vice versa.
> nifi process maybe crash while call controller-services api DBCPConnectionPoolLookup associate a great many jdbc pool
> -----------------------------------------------------------------------------------------------------------------------
>
> Key: NIFI-7005
> URL: https://issues.apache.org/jira/browse/NIFI-7005
> Project: Apache NiFi
> Issue Type: Bug
> Components: Core Framework, Core UI
> Affects Versions: 1.9.0
> Environment: cenots 7
> cpu 16cores
> memory 64G
> 3 nifi nodes java heap 32G
>
> Reporter: Paul Yang
> Priority: Major
>
> have a DBCPConnectionPoolLookup associate a great many (for our environment 170+) DBCPConnectionPool configurations.
> nifi.properties:
> nifi.cluster.node.read.timeout=600 sec
> nifi.cluster.node.connection.timeout=600 sec
> nifi.ui.autorefresh.interval=300 sec
> While call controller-services rest api from anyway included clicking ProgressGroup's GUI configure menu, nifi node that called one's CPU and memory surge even process crash.
> Fortunately, the jdbc lookup service's function is unaffected. But for the amazing function the behavior should be perfect.
> The root cause is: 'controller-services' api return json object is so large, the json string size is even more than 1GB or more then cannot be observed.and the nifi process cannot handle so large size response. the DBCPConnectionPool's information referenced DBCPConnectionPoolLookup's information and vice versa.
>
--
This message was sent by Atlassian Jira
(v8.3.4#803005)