0

I am trying to use a VPS as a SSH "proxy" to access my personal desktop that is behind multiple layers of NAT. Basically, I have a reverse SSH tunnel open from a TCP port on the VPS to my desktop. I am not running anything else (apart from what would be configured to run by default). But, no matter what I do, I eventually start triggering out of memory conditions as soon as minutes to a few days after a fresh boot.

This is the Vultr Console Output:

enter image description here

I tried the following to no avail:

  1. Switching from DigitalOcean to Vultr.
  2. Switching from Ubuntu 16.04 to Debian 8 (both x64).
  3. Increasing the allocated system memory from 512MB to 1GB.
  4. Adding a swap file of size 8GB. (It is persistent across boots).
  5. Changing the default SSH port to, successively, two different non-well-known ports. (Reason: security through obscurity, to some extent)

In addition to that, the OpenSSH server is configured to only allow public key access, and does not bind to port 22. Both my DigitalOcean and Vultr accounts as well as my domain registrar accounts are 2-factor protected, by the way, so I don't think any of those were compromised.

I looked at Why is my VPS running out of memory? but I don't see why adding a swap file won't help.

Can anyone explain what might be going on?

Improve this question
2
  • 1
    What does your memory and swap usage look like when this error happens?
    – Ramhound
    Commented Mar 22, 2017 at 19:07
  • @Ramhound So, the issue is that I'm not able to either SSH into the VPS or even log in via virtualization console once OOM is triggered, so I have no way of figuring out memory and swap usage at those moments.
    – Dhruv Singhal
    Commented Mar 22, 2017 at 22:37

0

Reset to default

You must log in to answer this question.

Browse other questions tagged .