Table of Contents
- You need to have an authority node to test this on, and you need to have at least 1 standby node.
- You should announce in #testnet-restarts when you are performing the swap/transfer, questions can be discussed in #testnet-development.
Brainswapping is the idea of having 2 nodes switch identities at the same time. We call is "brainswapping" because a node's identity dictates how it behaves.
Note: The procedure doesn't actually have to be a "swap"; a "brain-transfer" is also an alternative.
It was implemented as a way to update the network without having to bring it down, as federated servers can "brainswap" with standby nodes that have already been updated with the new code.
The network will not perceive this as a node going offline, as the identity (and thus the associated federated server) is still online.
After transferring the Authority identity the node operator can now shutdown their original node, perform necessary updates, bring it back online and finally brainswap the identity back into the original server.
The procedure can also be used for migrating the Authority identity to a new physical server, by not performing the brainswap a second time to reverse the first swap.
- Federated Node is the node that holds your authority identity, and needs to be updated.
- Standby Node is a follower node on the network that you control.
Prepping the Brain Swap
To perform the brain swap you will need 1 standby node that is ready. The standby node should be running the most recent Factomd software version.
Determining if you Standby Node is ready
If your standby node is not in sync with the network, performing the brainswap will result in your authority server going offline, so it is crucial to first check the health of the standby node.
- Check if the DBHeight matches that of the network
- Check if the minutes are following that of the network
- Check the process list for
<nil>, that indicates some network instability (Procedures for the above described at the bottom of this document)
Performing the Brain Swap (Read through before performing)
Once your standby node is ready, we can prep for the swap. The swap requires both config files (located on Federated Node and Standby Node) to be modified.
Open both config files in parallel in a text editor of your choice. (Procedures for editing is described at the bottom of this document)
Swap the following lines in the two config files:
IdentityChainID = FA1E000000000000000000000000000000000000000000000000000000000000 LocalServerPrivKey = 4c38c72fc5cdad68f13b74674d3ffb1f3d63a112710868c9b08946553448d26d LocalServerPublicKey = cc1985cdfae4e32b5a454dfda8ce5e1361558482684f3367649c3ad852c8e31a
(If you prefer to do a brain-transfer instead of a swap, you can just comment out the lines in your Federated node by placing a
; in front of these lines.)
Then add an additional line:
ChangeAcksHeight = 0
This additional line is the brainswap logic. You will want to set the
ChangeAcksHeight to some block height in the future (remember the block height on the control panel is the last saved height, not the current working height!).
The safe height is the one you see in the control panel + 4. (localhost:8090). If you know how to read the more detailed page, you can get away with a closer number.
Once you set the
ChangeAcksHeight to DBHeight+4, save both files.
At the block height
ChangeAcksHeight you should see both nodes change identities. If none of your nodes crash, and the identities change, the swap was successful.
Detailed instructions for some of the above aspects
Check if the DBHeight matches that of the network -- This is done by comparing DBHeight in your control panel (localhost:8090) with the dbheight of your federated server.
Check if the minutes are following that of the network
-- This is done by comparing the control panel of the Stanby node to that of the Federated. On the summary tab of the
More Detailed Node Information you will see this:
===SummaryStart=== FNode04[f0b7e3] L___vm01 0/ 0 0.0% 0.000 165[e0b9f8] 163/166/167 7/ 7 0/0/0/0 43400/0/0/0 0 0 2/40/100 0/0/0 0.07/0.00 0/0 - 309415
7/7 means you are on minute 7. You will want to make sure this number is the same on the Standby (
_/7) and the Federated (
Check the process list for
<nil>, that indicates some network instability
-- Process list is located in the control panel (localhost:8090) -> "more detailed node information". If any entries show
Modify config file -- You must edit the config file loaded by the docker container:
With the new docker swarm setup, the config files are located at
If you want to use vim
Edit the files per the instructions above.
Save & exit
If you want to use nano
Edit the files per the instructions above.
Save the file: