… I was too fast…
It now just took quite a bit longer, but influx got still killed by Ubunto because of an out-of-memory-panic.
The system now just seemed to use additional swap memory before crashing, so it took longer.
Right before its dead, influx had about 22GB in use (of which 6GB in swap)
I’ve seen already a lot of requests concerning how to avoid out-of-memory on influxdb, but never a definitive answer from InfluxData. My desire: Influxd lives with the memory that I assign it and adjusts its behavior accordingly.
