Topology snapshot explanation

classic Classic list List threaded Threaded
4 messages Options
rick_tem rick_tem
Reply | Threaded
Open this post in threaded view
|

Topology snapshot explanation

Hello,

I am wondering why offheap/heap memory continues to grow as the topology
shapshot evolves.  It must be client and server included.  Why is it being
reported for client?  Why would a client use offheap?  

Thanks,
Rick

event-processor_1        | [16:01:08] Topology snapshot [ver=5,
locNode=91131045, servers=1, clients=4, state=ACTIVE, CPUs=20,
offheap=11.0GB, heap=9.0GB]
event-processor_1        | [16:01:08] Topology snapshot [ver=6,
locNode=91131045, servers=1, clients=5, state=ACTIVE, CPUs=24,
offheap=13.0GB, heap=10.0GB]
event-processor_1        | [16:01:08] Topology snapshot [ver=7,
locNode=91131045, servers=1, clients=6, state=ACTIVE, CPUs=28,
offheap=15.0GB, heap=11.0GB]
event-processor_1        | [16:01:08] Topology snapshot [ver=8,
locNode=91131045, servers=1, clients=7, state=ACTIVE, CPUs=32,
offheap=17.0GB, heap=12.0GB]
event-processor_1        | [16:01:11] Performance suggestions for grid
'TemenosGrid' (fix if possible)
event-processor_1        | [16:01:11] To disable, set
-DIGNITE_PERFORMANCE_SUGGESTIONS_DISABLED=true
event-processor_1        | [16:01:11]   ^-- Disable grid events (remove
'includeEventTypes' from configuration)
event-processor_1        | [16:01:11]   ^-- Set max direct memory size if
getting 'OOME: Direct buffer memory' (add
'-XX:MaxDirectMemorySize=<size>[g|G|m|M|k|K]' to JVM options)
event-processor_1        | [16:01:11]   ^-- Disable processing of calls to
System.gc() (add '-XX:+DisableExplicitGC' to JVM options)
event-processor_1        | [16:01:11]   ^-- Decrease number of backups (set
'backups' to 0)
event-processor_1        | [16:01:11] Refer to this page for more
performance suggestions:
https://apacheignite.readme.io/docs/jvm-and-system-tuning
event-processor_1        | [16:01:11]
event-processor_1        | [16:01:11] To start Console Management &
Monitoring run ignitevisorcmd.{sh|bat}
event-processor_1        | [16:01:11] Data Regions Configured:
event-processor_1        | [16:01:11]   ^-- default [initSize=256.0 MiB,
maxSize=1.7 GiB, persistence=false]
event-processor_1        | [16:01:11]   ^-- 1G_Region [initSize=100.0 MiB,
maxSize=500.0 MiB, persistence=false]
event-processor_1        | [16:01:11]
event-processor_1        | [16:01:11] Ignite node started OK (id=91131045,
instance name=TemenosGrid)
event-processor_1        | [16:01:11] Topology snapshot [ver=4,
locNode=91131045, servers=1, clients=3, state=ACTIVE, CPUs=16,
offheap=8.7GB, heap=8.0GB]
event-processor_1        | [16:01:26] Topology snapshot [ver=9,
locNode=91131045, servers=1, clients=8, state=ACTIVE, CPUs=36,
offheap=20.0GB, heap=13.0GB]
event-processor_1        | [16:01:28] Topology snapshot [ver=10,
locNode=91131045, servers=1, clients=9, state=ACTIVE, CPUs=40,
offheap=22.0GB, heap=14.0GB]



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
Alexandr Shapkin Alexandr Shapkin
Reply | Threaded
Open this post in threaded view
|

RE: Topology snapshot explanation

Hello!

 

Client nodes still can keep some utility data that might be required for application.

I think the most suitable example is near cache [1].

 

[1] - https://apacheignite.readme.io/docs/near-caches

 

From: [hidden email]
Sent: Friday, September 13, 2019 7:10 PM
To: [hidden email]
Subject: Topology snapshot explanation

 

Hello,

 

I am wondering why offheap/heap memory continues to grow as the topology

shapshot evolves.  It must be client and server included.  Why is it being

reported for client?  Why would a client use offheap? 

 

Thanks,

Rick

 

event-processor_1        | [16:01:08] Topology snapshot [ver=5,

locNode=91131045, servers=1, clients=4, state=ACTIVE, CPUs=20,

offheap=11.0GB, heap=9.0GB]

event-processor_1        | [16:01:08] Topology snapshot [ver=6,

locNode=91131045, servers=1, clients=5, state=ACTIVE, CPUs=24,

offheap=13.0GB, heap=10.0GB]

event-processor_1        | [16:01:08] Topology snapshot [ver=7,

locNode=91131045, servers=1, clients=6, state=ACTIVE, CPUs=28,

offheap=15.0GB, heap=11.0GB]

event-processor_1        | [16:01:08] Topology snapshot [ver=8,

locNode=91131045, servers=1, clients=7, state=ACTIVE, CPUs=32,

offheap=17.0GB, heap=12.0GB]

event-processor_1        | [16:01:11] Performance suggestions for grid

'TemenosGrid' (fix if possible)

event-processor_1        | [16:01:11] To disable, set

-DIGNITE_PERFORMANCE_SUGGESTIONS_DISABLED=true

event-processor_1        | [16:01:11]   ^-- Disable grid events (remove

'includeEventTypes' from configuration)

event-processor_1        | [16:01:11]   ^-- Set max direct memory size if

getting 'OOME: Direct buffer memory' (add

'-XX:MaxDirectMemorySize=<size>[g|G|m|M|k|K]' to JVM options)

event-processor_1        | [16:01:11]   ^-- Disable processing of calls to

System.gc() (add '-XX:+DisableExplicitGC' to JVM options)

event-processor_1        | [16:01:11]   ^-- Decrease number of backups (set

'backups' to 0)

event-processor_1        | [16:01:11] Refer to this page for more

performance suggestions:

https://apacheignite.readme.io/docs/jvm-and-system-tuning

event-processor_1        | [16:01:11]

event-processor_1        | [16:01:11] To start Console Management &

Monitoring run ignitevisorcmd.{sh|bat}

event-processor_1        | [16:01:11] Data Regions Configured:

event-processor_1        | [16:01:11]   ^-- default [initSize=256.0 MiB,

maxSize=1.7 GiB, persistence=false]

event-processor_1        | [16:01:11]   ^-- 1G_Region [initSize=100.0 MiB,

maxSize=500.0 MiB, persistence=false]

event-processor_1        | [16:01:11]

event-processor_1        | [16:01:11] Ignite node started OK (id=91131045,

instance name=TemenosGrid)

event-processor_1        | [16:01:11] Topology snapshot [ver=4,

locNode=91131045, servers=1, clients=3, state=ACTIVE, CPUs=16,

offheap=8.7GB, heap=8.0GB]

event-processor_1        | [16:01:26] Topology snapshot [ver=9,

locNode=91131045, servers=1, clients=8, state=ACTIVE, CPUs=36,

offheap=20.0GB, heap=13.0GB]

event-processor_1        | [16:01:28] Topology snapshot [ver=10,

locNode=91131045, servers=1, clients=9, state=ACTIVE, CPUs=40,

offheap=22.0GB, heap=14.0GB]

 

 

 

--

Sent from: http://apache-ignite-users.70518.x6.nabble.com/

 

Alex Shapkin
rick_tem rick_tem
Reply | Threaded
Open this post in threaded view
|

RE: Topology snapshot explanation

Hi,

So you are saying clients too are using the data region settings and
reserving gigs of data for offheap (if gigs of data are configured) even if
I don't have near-cache configured?  I am using the same Spring
configuration for clients and servers, can I override that behavior at
runtime?  

Thanks,
Rick



--
Sent from: http://apache-ignite-users.70518.x6.nabble.com/
Ivan Pavlukhin Ivan Pavlukhin
Reply | Threaded
Open this post in threaded view
|

Re: Topology snapshot explanation

Hi Rick,

1. Clients should not "reserve" offheap memory.
2. Near caches do not use offheap.

Could you please check that a client process does not request offheap
memory from OS? I hope this is just a trouble with a "Topology
snapshot" misleading message.

сб, 14 сент. 2019 г. в 12:36, rick_tem <[hidden email]>:

>
> Hi,
>
> So you are saying clients too are using the data region settings and
> reserving gigs of data for offheap (if gigs of data are configured) even if
> I don't have near-cache configured?  I am using the same Spring
> configuration for clients and servers, can I override that behavior at
> runtime?
>
> Thanks,
> Rick
>
>
>
> --
> Sent from: http://apache-ignite-users.70518.x6.nabble.com/



--
Best regards,
Ivan Pavlukhin