You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@nifi.apache.org by Jo...@swisscom.com on 2022/07/18 08:35:08 UTC

NiFi SWAP Memory Usage Issue

Hi guys

We are running NiFi 1.15.3 on Centos 7.x with Java 1.8.0_322 / 1.8.0_332. We have a for us unexplainable behavior on some of the machines where NiFi is running. The servers have 256GB of RAM, but we limit NiFi to 31GB, that’s what you see below in the NiFi Cluster UI. Theres nothing else running except a few monitoring tools.


[A screenshot of a computer  Description automatically generated]


However, beside of the Max Heap Size we do see that the machine has a ton of “cached” RAM and it starts to eaten up as well our SWAP space – which we really don’t get. Why does the OS not release the cached stuff first?

[Timeline  Description automatically generated]
[A screenshot of a computer  Description automatically generated with medium confidence]

NiFi does use more than twice of the configured value, however this is still far below the installed memory. As far as I read the virtual memory shown here is anyway not the real used memory.

[A picture containing text  Description automatically generated]


Has anybody an idea where the cached stuff and the SWAP ram usage could come from? The SWAPing started just on the day where we have done the NiFi / Java Upgrade… Any help would be really appreciated.

Thanks a lot, cheers
Josef