[ 
https://issues.apache.org/jira/browse/IGNITE-4374?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15857821#comment-15857821
 ] 

Vyacheslav Daradur commented on IGNITE-4374:
--------------------------------------------

[~dmagda]

Done.

Current output:
{code}
[11:54:40] 
[11:54:41] Message queue limit is set to 0 which may lead to potential OOMEs 
when running cache operations in FULL_ASYNC or PRIMARY_SYNC modes due to 
message queues growth on sender and receiver sides.
[11:54:41] Security status [authentication=off, tls/ssl=off]
[11:54:43] Performance suggestions for grid  (fix if possible)
[11:54:43] To disable, set -DIGNITE_PERFORMANCE_SUGGESTIONS_DISABLED=true
[11:54:43]   ^-- JVM heap size is greater than 30.5Gb, JVM will not use 
compressed oops
[11:54:43]   ^-- Enable server mode for JVM (add '-server' to JVM options)
[11:54:43]   ^-- Enable G1 Garbage Collector (add '-XX:+UseG1GC' to JVM options)
[11:54:43]   ^-- Specify JVM heap max size (add '-Xmx<size>[g|G|m|M|k|K]' to 
JVM options)"
[11:54:43]   ^-- Specify New I/O max total size (add 
'-XX:MaxDirectMemorySize=<size>[g|G|m|M|k|K]' to JVM options)
[11:54:43]   ^-- Enable the thread-local allocation blocks (add '-XX:+UseTLAB' 
to JVM options)
[11:54:43]   ^-- Disable processing of calls to System.gc() (add 
'-XX:+DisableExplicitGC' to JVM options)
[11:54:43]   ^-- Speed up flushing of dirty pages by OS (alter 
vm.dirty_writeback_centisecs and vm.dirty_expire_centisecs parameters by 
setting to 500)
[11:54:43]   ^-- Reduce pages swapping ratio (set vm.swappiness=10)
[11:54:43]   ^-- Disable NUMA memory reclaim (set vm.zone_reclaim_mode=0)
[11:54:43]   ^-- Avoid direct reclaim and page allocation failures (set 
vm.extra_free_kbytes=1240000)
[11:54:43] Refer to this page for more performance suggestions: 
https://apacheignite.readme.io/docs/jvm-and-system-tuning
[11:54:43] 
[11:54:43] To start Console Management & Monitoring run ignitevisorcmd.{sh|bat}
[11:54:43] 
[11:54:43] Ignite node started OK (id=8bf18ef1)
[11:54:43] Topology snapshot [ver=1, servers=1, clients=0, CPUs=1, heap=1.0GB]
{code}

These lines are optionaly:
{code}
[11:59:07]   ^-- Speed up flushing of dirty pages by OS (alter 
vm.dirty_expire_centisecs parameter by setting to 500)
or 
[12:34:01]   ^-- Speed up flushing of dirty pages by OS (alter 
vm.dirty_writeback_centisecs parameter by setting to 500)
or
[12:33:01]   ^-- Speed up flushing of dirty pages by OS (alter 
vm.dirty_writeback_centisecs and vm.dirty_expire_centisecs parameters by 
setting to 500)
{code}

bq. Avoid direct reclaim and page allocation failures (set vm.extra_free_kbytes 
to 1240000 or other value).
We check equality to 1240000. If we write "or other value", we should change 
the check.

[Latest ci.tests|http://ci.ignite.apache.org/viewQueued.html?itemId=449798]

> Ignite should validate JVM and OS configuration and output warning in log
> -------------------------------------------------------------------------
>
>                 Key: IGNITE-4374
>                 URL: https://issues.apache.org/jira/browse/IGNITE-4374
>             Project: Ignite
>          Issue Type: Improvement
>          Components: documentation
>            Reporter: Yakov Zhdanov
>            Assignee: Vyacheslav Daradur
>         Attachments: x32_not_optimized.png, x32_optimized.png, 
> x64_not_optimized.png, x64_optimized.png
>
>
> Currently we have GridPerformanceSuggestions that output suggestions to logs 
> on Ignite start on how Ignite can be improved.
> I suggest to go a little bit deeper and validate more configuration options 
> and add validation for JVM and OS settings.
> Ignite should output warning if:
> * GC logging is not enabled
> * MaxDirectMemorySize is not set (-XX:MaxDirectMemorySize)
> * Heap size is greater than 30,5G and JVM cannot use compressed oops
> * Any of the recommended OS setting described here 
> https://apacheignite.readme.io/docs/jvm-and-system-tuning are not properly 
> set 



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Reply via email to