Hi,
Am running into issues trying to get my ES cluster up and running when on non-local host settings. Getting bootstrap failures with
initial heap size [33554432] not equal to maximum heap size [524288000]; this can cause resize pauses and prevents mlockall from locking the entire heap
I understand the nature of these errors and have changed configuration (jvm.opts) to have equal sizes of 2g. That hasn't done any good. Still the same error with same mismatch values.
Out of curiosity, ran java -XX:+PrintFlagsFinal -version | grep HeapSize and got below output
uintx ErgoHeapSizeLimit = 0 {product}
uintx HeapSizePerGCThread = 87241520 {product}
uintx InitialHeapSize := 33554432 {product}
uintx LargePageHeapSizeThreshold = 134217728 {product}
uintx MaxHeapSize := 524288000 {product}
As you can see, these numbers match to the error I am getting. Which means ES server is ignoring jvm.opts and picking those values from somewhere else. So, thought explicitly setting JAVA_TOOL_OPTIONS might help. And, so I did by setting both (-Xmx and -Xms) to 2g.
However, when I try and start the server, still the same error as before (shows the same min/max values).
Curious as to where these values are being picked up from ?
This really is a blocker if the settings aren't read as expected from the relevant sources. Would really appreciate if someone could point me in a direction that I might've missed configuring.
Cheers.