Hi Team,
We are seeing too much variation in available heap size. Let me explain to you our architecture in brief.
Our complete application is divided into 6 PUs. among 3 different applications
1) Core
2) Outbound
3) Inbound
Core app has 2 PUs
a) Core Pu -> This is an embedded space pu, whose purpose is to process the incoming space documents.
b) Mirror PU
Outbound app has 3 Pu
a) DS0 -> Takes data from core pu, converts into some internal format writes back to core space
b) DS1 -> Takes converted data from space and persists to DB
c) DS2 -> Takes data from core space and dumps in to flat file
Inbound app has 1 Pu
a) Welcome -> Takes data from another space and write to the core Pu
We have 4 GSCs with 4 GB heap size. Requirement is that we get data at around 5pm everyday we process and maintain 2 days data in our space.
Now what we have observed is that once the completed data is processed, no more input is coming, no more output is generated. We can see that there are no transactions running but if i see the heap of each GSC, it varies a lot. It climbs uptio 3GB then after a minute it will drop to 1GB and then again it will rise to 3GB and come back to 1 GB.
Since the space is ideal, only thing thats running is our polling containers nothing else, so i am not sure what is causing the huge variation. I dont know if GS does something internal on its own or is it something we are doing, i checked the GSC logs our app logs, nothing is in process but still GSC heap size varies a lot.
It would be great if you could provide some directions or any suggestions where to proceed ??
Thanks a lot :)
↧