I'm trying to run my own mempool.space. However, the default installation in Umbrel (and other node managers) isn't cutting it. Why not?
Well, these one-click self-hosted mempool installations are using Electrs (lightweight) backend which is fine, but it means my local mempool instance won't collect everything that the Official mempool.space is able to collect.
Looking at the Mempool source code, I see this hint which explains that only the Esplora backend can retrieve spent transaction outputs, which are needed to do forensic analysis on closed LN channels.
Has anyone gotten Mempool running with an Esplora backend? Are there any guides, or high-level steps you can point me towards?
Months ago, I successfully pulled the esplora docker image on an umbrel node and even synced the full esplora index. However, I couldn't figure out how to get mempool (in umbrel) to use the new esplora backend I had just installed, so I gave up.
Just thought I'd ask for tips before attempting this again on baremetal (no docker, no umbrel).
You should use mempool's fork of electrs: https://github.com/mempool/electrs
Also be aware that you should probably have 1.5TB+ available. The syncing process requires a lot, but then you should expect 1TB or so of data that stays after it is done syncing. I forget exact numbers.
I saw mempool/electrs and assumed it wouldn't work cause it wasn't named Esplora. But looking into it now, makes perfect sense.
The esplora index I made a few months ago had similar footprint to what you describe. Ballooning to 1.3TB then compacted to about 800GB.
This is a huge tip! Thanks Tony!
10 sats \ 1 reply \ @Zk2u 13 Jun
Yeah, I’m running a full node with mempool’s electrs fork. Didn’t want to run esplora because it’s wayyyy too heavy for most personal needs.
According to #272976 Seems like mempool's electrs is a fork of esplora, so it should be about as heavy as esplora is, right?
I'm using Start 9 StartOS and it's working very well.
when you go to this link on your mempool, do you see "Starting Balance"?