-
alMalsamo
2022-02-13 01:56:27.674 I Synced 735208/2558182 (28%, 1822974 left, 3% of total synced, estimated 3.0 months left)
-
alMalsamo
3 months good grief! >_<
-
alMalsamo
Can someone explain the difference between the two percentages, why does it say both 29% and 3%?
-
Steven_M
Hi all :-)
-
rbrunner
alMalsamo: Your blockchain file currently contains 28% of all blocks.
-
rbrunner
From what is left for you to sync beyond those 28%, 1822974 blocks, you synced about 3% of them (thus about 55,000) in your current syncing run.
-
rbrunner
This regularly confuses people, but I think it's difficult to find a decidedly clearer message that is still reasonably short ... (I am the one to blame, my code)
-
moneromooo
alMalsamo: not too late.
-
alMalsamo
moneromooo: Really? I just kill the process and restart it with the additional switch and it will resume fine?
-
moneromooo
Yes.
-
alMalsamo
OKay I killed monerod and tried to restart it after remounting my /mnt partition and it totally fails now
-
alMalsamo
2022-02-13 09:28:02.474 W Failed to query m_blocks: MDB_BAD_TXN: Transaction must abort, has a child, or is invalid
-
alMalsamo
2022-02-13 09:28:02.478 E Error opening database: Failed to query m_blocks: MDB_BAD_TXN: Transaction must abort, has a child, or is invalid
-
alMalsamo
2022-02-13 09:28:02.480 I Stopping cryptonote protocol...
-
alMalsamo
2022-02-13 09:28:02.480 I Cryptonote protocol stopped successfully
-
alMalsamo
2022-02-13 09:28:02.481 E Exception in main! Failed to initialize core
-
moneromooo
You can try running once with --db-salvage and hope you get lucky.
-
alMalsamo
moneromooo: its giving me the same errors :(
-
alMalsamo
You said I could kill the process and restart it! >_<
-
moneromooo
Yes, you can.
-
moneromooo
Now, you get to delete the chain though.
-
alMalsamo
Okay so why am I getting these DB errors? There is no way to resume I really have to delete and start from scratch?
-
alMalsamo
If I can kill the process I would hope it can resume okay like was suggested here...
-
rbrunner
Well, if everything is still fully under control, it's much better to use the daemon interactive command "exit" to leave because that just closes everything in a clean way.
-
rbrunner
This gives a very reliable resuming.
-
alMalsamo
rbrunner: How do I type exit when there is no CLI prompt when I run the monerod process it just shows progress of sync like I was pasting, there is no prompt to enter commands...?
-
rbrunner
How do you start it?
-
rbrunner
If if shows something on the console usually it just is ready for you to type something as well
-
rbrunner
It just does not show any special prompt to welcome you to do so, if you understand what I mean
-
rbrunner
Even if it's busy outputting lines you should be able to type "exit" and enter
-
alMalsamo
monerod --data-dir /mnt/monero --prune-blockchain --db-salvage
-
alMalsamo
Oh I didn't have db-salvage initially
-
alMalsamo
I am trying with db-salvage now and it still is broken :(
-
rbrunner
Looks like some bad luck then. Yes, if you start that way and see its output in the console, you can also type the "exit" command.
-
moneromooo
Why do you have a car accident ? You don't set out to have one, and most rides don't end in an accident. It's just bad luck, and it happens from time to time, unknown cause so far.
-
moneromooo
But you could try killing it again after restarting and see if it dies the same way on restart. This would hint that your particular setup is prone to those and might help debug things.
-
rbrunner
You mean cases where the daemon terminates / is terminated where it should be able to commit properly but mysteriously isn't?
-
moneromooo
I don't know whether the problem is inability to commit properly.
-
moneromooo
alMalsamo: what filesystem's on /mnt/monero btw ?
-
alMalsamo
moneromooo: NTFS lol whoops
-
alMalsamo
FUSE mounted
-
rbrunner
That setup may be a tad optimistic :)
-
moneromooo
And it'll also explain the extreme slowness.
-
moneromooo
Has anyone seen any difficulty desync between nodes in the last, say, year ?
-
moneromooo
That is, nodes on the same top height, but with slighly different difficulty values.
-
moneromooo
s/height/block/
-
selsta
-
alMalsamo
selsta: Do you know a tool to benchmark FUSE filesystems vs. in-kernel filesystems?
-
selsta
no I don't and monero sync is extremely random io heavy which isn't well captured by benchmark tools
-
selsta
moneromooo: as far as I know there haven't been any difficulty desync reports since you have fixed this bug
-
moneromooo
OK. I was looking at sync performance again a bit yesterday, and this would be an easy re-win. It was such a rare and devious bug though. Would be a shame to miss a repro.
-
selsta
moneromooo: can't you just run this code after the initial sync as a compromise?
-
selsta
or otherwise I would just remove it again, it seems fixed at this point
-
moneromooo
The code calculates the diff with (1) the incremental cache system and (2) the brute force lookups, and compares. Running it only after initial sync will only catch an issue long after it happened.
-
garth
Whatever happened to that guy who was going to do a better statistical approach to the decoy selection algorithm?
-
garth
Did that ever come to anything?
-
moneromooo
I think that's Rucknium[m] - still around.
-
Rucknium[m]
garth: Yes I'm working on it
-
Rucknium[m]
-
Rucknium[m]
Do you want to see a draft of the technical specification? (above is just some testing code)
-
garth
Sure! I was just curious. The talk you gave (I think on Monero talk) was just super interesting so I’m eager to see how things progress
-
Rucknium[m]
I haven't given any talks like that, so you may be thinking of someone else. I will PM you.
-
dell
hi
-
narf
hi