This is cool

no community meeting here today, next week. but there is a #monero-events:monero.social meeting at 18:00UTC 👍️

If you're at FOSDEM feel free to ping or DM me

We're having a small meetup

here the monero community matrix room matrix.to/#/#monero-community:monero.social

* posted in the wrong room, sorry

it's true, HAL does have a bug

monerokon meeting in 5 minutes in #monero-events:monero.social

Todays loot an FOSDEM

Those matrix t-shirts are cool

<Stnby[m]> "Todays loot an FOSDEM" <- 50% of what you collected

Gamescom loot is on another level though

monerobull[m]1: I got 4 beers and a baked potato last time

monerobull[m]1: I really tried them to accept Monero, sadly aint working

<Stnby[m]> "PXL_20230204_193623183.MP.jpg" <- Curl stickers are pretty cool

I did not expect to see the Curl man there, they were advertising WolfSSL, just because how shit OpenSSL is

Wireguard

(or is this unrelated?)

I unmounted an external SSD with the block chain forgetting that monerod is running.

Now it says that DB is corrupted

--db-salvage doesn't seem to do anything

How can I forget last 100 blocks?

pop_blocks 100

I can't start the daemon because it fails to parse the DB and pop_blocks requires the daemon running

Are there any tools that operate directly on the DB?

Yes

./monero-blockchain-import --pop-blocks 100

you will likely have to resync

or restore from backup if you have one

It looks like block chain import tries to start the daemon too and fails with "can't parse a blob" error

But there is like 149GiB of the valid data and only the recent one is corrupted. There should be a way to truncate the DB without resyncing everything.

so we cant start monerod at all hm

Don't worry, i can't start bitcoind either

lol

Or at least my ord.exe can't

Will have to fix that

esac212[m]: the db isn't sequential where you can just take the top away to remove corruption

But it looks like I can fix it by just truncating this range

like a reverse pruning

It looks like I actually need to find the first corrupted node then discard it and then discard all nodes that contain data about blocks which are not present in the valid part of the DB, and there are no tools to manipulate LMDB so this must be coded from scratch 🥴

?

fyi never tried it

did you check the mdb_* tools? not sure if they are helpful for what you want to do

Yep, maybe with combination of mdb_dump + mdb_load I can achieve this but I'm not sure. I will play with the BlockchainDB class first.