I’ve decided to commit to buying a vps so that I can better support the network by running an exit relay.
My main decision to make is do I migrate my current relay to the VPS so there is less bandwidth measuring time, and setup a new relay on my Pi, or do I:
Keep my current middle/guard on my Pi and setup a new relay on the VPS.
On a seperate unrelated note I have a 4TB monthly bandwidth allowance so what would be the most appropriate torrc option to stop me going over that? Implement a bandwidth limit and calculate it so that it doesnt exceed 4TB? Or apply accountingmax, which is less reliable, and set it to 2TB?
If you are able to keep both relays running, I suggest you just set up a new one on the VPS (more relays = better network).
If however you end up migrating the relay, check this out:
NOTE: make sure the ownership/permissions of the moved files are set correctly to work on the new machine.
Regarding the traffic cap, I also run a limited relay using the AccountingMax / AccountingStart options in the torrc; set it at about 90-95% of the available transfer to stay safe.
For details check out:
In 2014, the OpenSSL vulnerability named Heartbleed made the news and was a serious concern for relay operators. If I remember correctly, I think it was arma who recommended that relays be re-keyed and provided easy-to-follow instructions to do so.
I rebuild my relay every four years or so for a new version of Linux.
Before Heartbleed, I had been migrating the keys.
After Heartbleed, I stopped migrating the keys and simply copy the relevant lines to the new “torrc”. It takes a few weeks to build relay traffic; however, it is probably more secure.
Admittedly, it is disappointing to see the relay’s value “First Seen” drop to zero again.
