Sometime during normal operation, my node crashed (possibly due to this issue, but I do not know) and I had to perform a reindex. During this reindex, I noticed that it was consuming upwards of 26 GB of memory and constantly consuming more. Unfortunately this caused my computer to lock up and I had to reboot it to fix that. Starting Core again resulted in a reindex and the same behavior again.

I am using a build of the current master branch on commit 00d25e9

Machine info:
Ubuntu 17.10
32 GB RAM

I was going to run it in valgrind but it was taking too long to get anywhere remotely past initialization to the part where the memory leak was happening, so I killed that.

Now I am doing git bisect to find where the leak was introduced. AFAIK this leak is not present in 0.15.1.

I believe that @sipa experienced the same issue on his odroid but was not able to reproduce the problem.