What is the behavior of Alfresco Community v5.0 caching and how to control it

[mtgdavidchow]

We have a system running Alf5.0, you can think that we have customized field in metadata that storing the person id (unique).  When we search with the person id, if the no. of records is large (say 500 records), it needs more than few minutes to get the result.  However, after this completed, the next search of the same person id returns the result in no time. But may be few hours / days later,  the first search of it becomes slow again. And we guess this is the cache behavior (invalidated).

Can someone advise how to control the cache behavior such that we can cache person with many records and keep it not invalidated? Or any advice to make it user friendly on that? Thank you!

[mtgdavidchow]

is there anyone can help on the above?

[fedorow]

Check the java memmory allocation first. In my expiriense a lot's of v5 servers works with default 1GB RAM indepandantly how many phisical memory the host have.

[mtgdavidchow]

Can you advise which config. to be checked?

In tomcat

we already set to use 128GB ram

We used the following during tomcat service installation

set JAVA_OPTS=++JvmOptions="-XX:+DisableExplicitGC" ++JvmOptions="-XX:+UseConcMarkSweepGC" ++JvmOptions="-XX:+UseParNewGC" ++JvmOptions="-XX:ReservedCodeCacheSize=128m" ++JvmOptions="-Dalfresco.home=D:\Alfresco" ++JvmOptions="-Dcom.sun.management.jmxremote" --JvmMs 2560 --JvmMx 131072 

For solr4

we are using default setting, and I found that the followings are commented, should we adjust it?

In solrconfig.xml,

<!-- <ramBufferSizeMB>100</ramBufferSizeMB> -->
<!-- <maxBufferedDocs>1000</maxBufferedDocs> -->

[fedorow]

check this post:

https://hub.alfresco.com/t5/alfresco-content-services-hub/default-memory-allocation-issue-in-alfresc...

[mtgdavidchow]

I am already using correct JVM setting as mentioned in previous setting and I can see the tomcat can use around 6x GB memory

So, I am thinking other direction. Is the following related? and if yes, any way to calculate the best values?

For solr4

we are using default setting, and I found that the followings are commented, should we adjust it?

In solrconfig.xml,

<!-- <ramBufferSizeMB>100</ramBufferSizeMB> -->
<!-- <maxBufferedDocs>1000</maxBufferedDocs> -->

[mtgdavidchow]

Anyone can help to advise? Thank you.

[mtgdavidchow]

Anyone can help ? The problem is still here

[mtgdavidchow]

Anyone can help ?

[mtgdavidchow]

@All @fedorow 

Is there any solr4 setting for large size of data? I am not sure how to tune it. I have already tuned tomcat java memory one. Thank you.