Difference between revisions of "LND Configuration Settings"
Jump to navigation
Jump to search
Line 67: | Line 67: | ||
== Channel Size limits == | == Channel Size limits == | ||
{| class="wikitable" | |||
! Setting | |||
! Description | |||
|- | |||
|minchansize=5000000 | |||
| Minimum Channel Size in Sats | |||
|- | |||
|maxchansize=50000000 | |||
| Maximum Channel Size in Sats | |||
|} | |||
== Database optimizations == | == Database optimizations == |
Revision as of 19:36, 16 July 2021
Intro
Ever since the Lightning Network has started and when people began opening channels with others they have been trying to get the most out their Node. People have been doing this ever since computers have gotten into peoples hands. Trying out different settings and finding out what works best.
Anyone who would love to go over a long list of config parameters which can be used could check out this Lnd GitHub page https://github.com/lightningnetwork/lnd/blob/master/sample-lnd.conf
There are more entries then I ever thought possible, but this Wiki page is how to make adjustments to the Lightning Node environment which helps ones node to perform better. Something which we'd all would like to happen.
The list which follows is the most common settings which Node Operators have come to agree which best suites a Node which wants to earn fees and route other Nodes Payments. These settings apply to LND 0.12.1 releases. The latest release still has many issues which need to be resolved.
Watch Tower Settings
Pretty good name for what this does. Watch Towers are communications between agreeing Nodes to keep on eye on each others Node. Node Operators pass their Watch Tower address between each other so when one of them is down for some reason, the other one watches for channel situations like a forced close. If this closure tries to use a non-current Channel state, then the Watch Tower will step in and penalize the Node trying to cheat. Currently, the fee for watching another Node is disabled, but at some point there will be a cost with this. Now if your Node does catch some foul play one will earn fees for being the good guy.
Setting | Description |
---|---|
wtclient.active=1 | Turns on the Watch Tower Client |
wtclient.sweep-fee-rate=10 | Sets the sweep fee in sats/vbyte for the justice transaction |
Routing Optimization
Setting | Description |
---|---|
routerrpc.apriorihopprob=0.5 | Probabilistic router hop bias |
routerrpc.aprioriweight=0.75 | Probabilistic router weighting bias |
routerrpc.attemptcost=10 | Probabilistic router cost base |
routerrpc.attemptcostppm=10 | Probabilistic router cost PPM |
routerrpc.maxmchistory=100000 | Maximum channels you keep in memory, increasing this will require more memory |
routerrpc.minrtprob=0.005 | Minimum probability to attempt a channel in a route path |
routerrpc.penaltyhalflife=2h0m0s | Probabilistic router penalty halflife period (how often it will 1/2 the penalty cost of associated with failed route attempt) |
routerrpc.attemptcost=10 | Probabilistic router attempt cost |
routerrpc.attemptcostppm=10 | Probabilistic router cost value |
max-cltv-expiry=5000 | Max CLTV value for route path (this is the sum of all the CLTVs from each hop, so it limits max hops as a result) |
max-commit-fee-rate-anchors=100 | Commit fee to use when opening a anchor style channel |
Channel Size limits
Setting | Description |
---|---|
minchansize=5000000 | Minimum Channel Size in Sats |
maxchansize=50000000 | Maximum Channel Size in Sats |
Database optimizations
- db.bolt.auto-compact=true
- gc-canceled-invoices-on-startup=1
- gc-canceled-invoices-on-the-fly=1
- caches.channel-cache-size=500000
Startup Optimizations
- sync-freelist=1
- stagger-initial-reconnect=1
Logging
- debuglevel=CNCT=debug,CRTR=debug,HSWC=debug,NTFN=debug,RPCS=debug
Pre 0.13 LND settings
- protocol.anchors=true