Ceph

I'm still hunting this one, but I'm seeing high memory usage of my monitors (three in total).

My monitor configuration:

[mon]
    mon data = /var/lib/ceph/$name
    mon clock drift allowed = 1
    mon clock drift warn backoff = 30
    osd pool default size = 3
    osd pool default crush rule = 0
    osd min down reporters = 3
    osd min down reports = 5
    debug auth = 0
    debug mon = 0
    debug paxos = 0

[mon.pri]
    host = monitor
    mon addr = [2a00:f10:11b:cef0:230:48ff:fed3:b086]:6789

[mon.sec]
    host = monitor-sec
    mon addr = [2a00:0f10:011a:408::1]:6789

[mon.third]
    host = monitor-third
    mon addr = [2a00:0f10:011a:409::1]:6789

The memory usage I was seeing on 'sec' and 'third' was:

root       844  0.4 42.9 33748068 105804 ?     Dsl  Dec15   5:58 /usr/bin/ceph-mon -i third -c /etc/ceph/ceph.conf

In this case the VM where the monitor was running it had only 256MB of RAM.

However, when the monitor had been running for a while it would stay around 1 ~ 8% memory usage. At some point it would start eating more and more memory before crashing.

I do have core dumps, but the 256MB of RAM makes it hard to read them.

I don't know if you qualify this as a bug, but imho it is weird that a monitor jumps from 5% memory usage to 44% and not much later crashes (no OOM message btw).

Running version 0.39 btw