This message was deleted.
# general
s
This message was deleted.
b
You can increase
druid.worker.capacity
in middle manager for more slots and in turn more peons
p
After adding this property it says 'middleManger configs (druid.indexer.runner.javaOptsArray or druid.worker.capacity) is present in middleManager/runtime.properties, add jvm.config for all other services'
b
To increase the slots, You need to update
druid.worker.capacity=new value
in all your middle manager
runtime.properties
file and then restart the middle managers.
generally i have seen 1/3rd or 1/2 for number of cores is used
First check if the property is present in runtime.properties, if yes then update else add new
p
Yes i have added the same in runtime.properties. So i have set it 'druid.worker.capacity=4' , then still the same error..and i have only 1 runtime.properties file
b
Share the config and the error log (may be log snippet from start failure )
p
This is the runtime.properties file from middle manager: druid.service=druid/middleManager druid.plaintextPort=8091 druid.worker.capacity=4 druid.worker.baseTaskDirs=[\"var/druid/task\"] druid.indexer.task.hadoopWorkingPath=var/druid/hadoop-tmp
b
Is this the complete file 😞
Could you attach the log here as well please
p
Yes sending
As i am very new to this i really dont have much knowledge
b
Thats ok, no worries, Just share the complete runtime.properties for the middle manager and the log and let me check, if i can find anything
p
middleManager.stdout.log
I have also made some changes in indexer, runtime.properties
Thanks a lot for helping me
b
I do not see the error that you were saying first file is empty and there the worker.capacity is set to 2 only. Could you please restart the middle manager and share the latest log
p
okay wil restart
b
You mentioned somewhere you updated indexer, are you using middle manager/indexer?
p
is it like we can use only one?
b
Seems like
The Indexer is an optional and experimental feature. Its memory management system is still under development and will be significantly enhanced in later releases. The Apache Druid Indexer process is an alternative to the MiddleManager + Peon task execution system. Instead of forking a separate JVM process per-task, the Indexer runs tasks as separate threads within a single JVM process. The Indexer is designed to be easier to configure and deploy compared to the MiddleManager + Peon system and to better enable resource sharing across tasks.
and for indexer : # Number of tasks (druid.worker.capacity) is automatically # determined based on available processor.
If you go to your console do you see middle managers running in the services tab?
If you are using indexer, then increasing the processor for the data node should be the solution i believe.
p
yes middle manager was runnning
b
Should we do a quick zoom session
p
can we do google meet, i will send the link
b
sure
p
yes just a minute
please join
b
Thank you for the call, We have done the changes to the config and the slots has been increased but as we discussed the processor xmx/xms/direct memory will have to be changed as per the resources available to run more jobs.
p
Thank you so much for resolving an issue..Now its working properly. 😊
👍 1