[cas-user] Cas container generate a lot of java threads

[artur miś]

Hello,

  After trafic  only about 500 users  to cas 6.3.2 on docker container  i can see 6k threads  for java  on this container so it has taken  swap memory too, let's say  after18h i see problem  with threads. 

java for this container docker :

java -Xms512m -Xmx2025M

In dokcker container  memory is not limited but the quantity of threads doesnt seem to connected with this  unlimited access to memory by container   

What do you suppose that can be issue of consumption memory , this problem apears itself during  pasword change process ? .

TGT expire time and possibility password change after pm sent emial to user is configured for expire time 3h .

Regards and Thank you.

-- 

AM

--
- Website: https://apereo.github.io/cas
- Gitter Chatroom: https://gitter.im/apereo/cas
- List Guidelines: https://goo.gl/1VRrw7
- Contributions: https://goo.gl/mh7qDG
---
You received this message because you are subscribed to the Google Groups "CAS Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cas-user+u...@apereo.org.
To view this discussion on the web visit https://groups.google.com/a/apereo.org/d/msgid/cas-user/781cb2ff-806d-499e-8a21-3b13e2482cefn%40apereo.org.

[artur miś]

I suppose It is comming from   pm module and  possible problems with  Garbage Colllectors .   Is smb using  pm module in docerized  CAS  ? Do user activities with password  changing genarate  not closable threads ? Each action od changing  password  genarate 12 new threads .

Java : openjdk 11.X

To view this discussion on the web visit https://groups.google.com/a/apereo.org/d/msgid/cas-user/ba883598-828f-47bf-bf76-5142915b90f3n%40apereo.org.

[artur miś]

I find out  that:

 -inside cas docker container  i can see  that ldaptive  connection are establised while(in the same time)  on docker host i can not see such conection  with ldaps ( netstat -altnp) . Let's say   no user connection  now to cas.

- quanity of ldaptive threads  are equal numers of esabished connection inside container.

ps -T -p <pid_cas_container>  | wc -l

Question:

- Do i need in  cas.propierties  any ldap pool policy  to clear   passivate?

- If it is well that cas keep connection  why does cas have got such mechanism  (for acceterate speed connection)?

 - mayby it is problem od container , i'm not sure.

My cas is 6.3

To view this discussion on the web visit https://groups.google.com/a/apereo.org/d/msgid/cas-user/9dd86310-f45d-4f7a-9d22-1b57a97db868n%40apereo.org.

[Pablo Vidaurri]

Did you find a config that help with  your issue? I have cas 6.3.7.4 running on 4 nodes and having issue with JVM memory (6g) running low. Took a memory dump and it's all pointing to io.netty. Checked dependency tree and seems it is being used by ldaptive.

To view this discussion on the web visit https://groups.google.com/a/apereo.org/d/msgid/cas-user/4d9b280f-b422-43fb-b7de-9d9e37605130n%40apereo.org.