highperfocused/lnd
1
0
Fork 0
mirror of https://github.com/lightningnetwork/lnd.git synced 2025-06-23 07:11:11 +02:00
Code Issues Packages Projects Releases Wiki Activity

docs: update leader election docs [skip ci]

Browse Source
This commit is contained in:
Andras Banki-Horvath 2021-11-01 16:21:16 +01:00
parent 217019aa8a
commit c9533bf6b9
No known key found for this signature in database
GPG Key ID: 80E5375C094198D8
1 changed files with 12 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
docs/leader_election.md
View File

@ -94,19 +94,18 @@ readinessProbe:
periodSeconds: 1 periodSeconds: 1
``` ```
## Replication of non-critical data ## What data is written to the replicated remote database?
All critical data is written to the replicated database, including LND's wallet Beginning with LND 0.14.0 when using a remote database (etcd or PostgreSQL) all
data which contains the key material and node identity. Some less critical data LND data will be written to the replicated database, including the wallet data
however is currently not written to that same database for performance reasons which contains the key material and node identity, the graph, the channel state,
and is instead still kept in local `bbolt` files. the macaroon and the watchtower client databases. This means that when using
leader election there's no need to copy anything between instances of the LND
cluster.
For example the graph data is kept locally to improve path finding. Other examples ## Is leader election supported for Postgres?
are the macaroon database or watchtower client database. To make sure a node can
become active and take over quickly if the leader fails, it is therefore still
recommended to have the LND data directory on a shared volume that all active and
passive nodes can access. Otherwise the node that is taking over might first need
to sync its graph.
As we evolve our cluster support we'll provide more solutions to make replication No, leader election is not supported by Postgres itself since it doesn't have a
and clustering even more seamless. mechanism to reliably **determine a leading node**. It is, however, possible to
use Postgres **as the LND database backend** while using an etcd cluster purely
for the leader election functionality.