-
winter
<selsta> didn't they just have a security release a couple days ago? <-- you may be thinking of the pre-advisory for this
-
reminder
hi
-
reminder
Why -30 here use_2021_scaling = use_fork_rules(HF_VERSION_2021_SCALING, -30 * 1);
-
nassau
hello
-
nassau
I have a error Unexpected hard fork version v16 at height Make sure your wallet is up to date. Blocks received:
-
moneromooo
Did you make sure your wallet is up to date ?
-
moneromooo
(also daemon, always use matching versions if possible)
-
nassau
Yes i did
-
nassau
All updated with latest version
-
moneromooo
Is your wallet software monero-wallet-cli ?
-
nassau
yes
-
moneromooo
What does "status" report in monero-wallet-cli ?
-
nassau
Refreshed 1/1, synced, daemon RPC v11.11, SSL
-
moneromooo
What does "status" report in monerod ?
-
nassau
Height: 1/1 (100.0%) on mainnet, not mining, v16, 11(out)+11(in) connections, uptime 0d 0h 5m 50s
-
moneromooo
What does "version" say in both monero-wallet-cli and monerod ?
-
nassau
0.18.1.0-c9cfa251
-
moneromooo
Same for both ?
-
nassau
Yes
-
moneromooo
Did you change the code ?
-
nassau
Not
-
moneromooo
Did you run with --regtest ?
-
nassau
Not
-
moneromooo
Then... it's bizarre...
-
moneromooo
The daemon is not syncing at all, and it should be v1 at that point.
-
moneromooo
I do not understand why it'd be v16 without code changes nor regtest :/
-
moneromooo
Can you paste the output of "sync_info" to paste.debian.net ?
-
moneromooo
(monerod)
-
nassau
This is the output from monero-wallet
-
nassau
"Unexpected hard fork version v" + std::to_string(hf_version) + " at height " + std::to_string(height) + ". " +
-
nassau
If i install old monero version i have no problems
-
moneromooo
Both old and new will use the same db, so if the new monerod reports height 1, the old also should.
-
selsta
nassau: try to compile the latest version
-
selsta
(I know they left already)
-
linuxperia
Hi all. I am a monero newbie. after i synced the monero blockchain db and run p2pool in parallel somehow i crashed the monero db. i am getting always this error message now here when i try to start the monerod => "W Failed to find txpool tx blob to match metadata" "Exception in main! Failed to find txpool tx blob to match metadata" I checked the lmdb database using the "mdb_stat -fff lmdb/" command and it segfault with this message here =>
-
linuxperia
"Freelist Status
-
linuxperia
Tree depth: 2
-
linuxperia
Branch pages: 1
-
linuxperia
Leaf pages: 12
-
linuxperia
Overflow pages: 0
-
linuxperia
Entries: 336
-
linuxperia
Transaction 23, 5 pages, maxspan 4
-
linuxperia
36554021
-
linuxperia
36554027[4]
-
linuxperia
Transaction 24, 6 pages, maxspan 4
-
linuxperia
36554022
-
linuxperia
36554024
-
linuxperia
36554033[4]
-
linuxperia
Transaction 6720514773099737224, 3 pages, maxspan 1
-
linuxperia
35700948
-
linuxperia
35708585
-
linuxperia
35728040
-
linuxperia
Segmentation fault
-
linuxperia
"
-
linuxperia
need to get my monero blockchain synced again without that i download the whole blockchain again as it will wear down my mlc ssd. How can this problem be fixed. Monero DB looks very instable compared to other much more advanced key/value blockchain databases
-
moneromooo
You can try running monerod once with --db-salvage.
-
moneromooo
If it does not work, I don't think you can avoid a full resync.
-
moneromooo
hyc: mdb_stat crash above, if there's info you'd like to ask linuxperia.
-
moneromooo
linuxperia: did your computer crash/wedge ?
-
linuxperia
i just restarted p2pool severaltimes in parallel
-
moneromooo
Windows is known to be shit at preserving filesystem data when it does.
-
moneromooo
If it happens again, you may want to run monerod with --db-sync-mode safe (IIRC).
-
moneromooo
I assume this means "no" ?
-
linuxperia
it does not work with --db-sync-mode
-
linuxperia
i will try again
-
linuxperia
moneromoo: tryed both options --db-sync-mode safe --db-salvage it does not help :-(
-
linuxperia
looks like need to resync :-( very bad
-
linuxperia
thanks still for the tips moneromooo
-
moneromooo
hyc: if there's a txn bug in monerod, can it lead to mdb_stat crash ? I'd think it'd just lead to semantically inconsistent data, but still with a synctactically valid db.
-
linuxperia
i was able to do a more verbose log output. it provides much more info now what happens before the crash. its quite long. will try upload it to a paste website. 1 moment
-
linuxperia
its 1.8 MB big and most pastebins refuse such a long log. need split it up. 1 minute
-
linuxperia
moneromoo: hyc: here is the crash log of monerod =>
pastebin.com/raw/NfDyKgNp
-
hyc
moneromooo: correct, a bug in monerod cannot cause an invalid DB.
-
moneromooo
So it looks like a possible lmdb bug if this happened without an OS/HW crash.
-
moneromooo
Or OS bug I suppose.
-
hyc
nah. from discussion in other channel, he's using a weak SSD
-
moneromooo
Ah.
-
linuxperia
moneromooo: hyc: i think p2pool destroyed the database as it uses the database of monerod too. overall think main problem happens when several programms at the same time in parallel try to use the monero database happens such problems. other have reported also problems with the database. see here. =>
reddit.com/r/monerosupport/comments…now_a_solution_for_the_error_failed yeah it happened without a hw Crash. I just
-
linuxperia
Started and Stoped p2pool very fast in parallel
-
hyc
yeah that's false.
-
hyc
p2pool doesn't touch the monerod database
-
hyc
and LMDB is rock solid with multiple process use. but that's not the situation here.
-
hyc
as for the error message you linked above - that happens routinely, doesn't mean anything
-
hyc
in particular doesn't mean the DB is broken
-
linuxperia
hyc: okey just checked the which programm i started in parallel very fast and used the lmdb databse of monero and it is xmrblocks it require direct access to the lmdb data abse for this i created a symlink to the database as follow on the ssd "xmrblocks/lmdb/data.mdb -> /ssd-data/monero/lmdb/data.mdb" so it was not p2pool but xmrblocks
-
hyc
I dunno what xmrblocks is
-
linuxperia
its a monero explorer that you can run local
-
linuxperia
-
hyc
is monerod always active when xmrblocks starts? why does xmrblocks start/stop "very fast in parallel" ?
-
linuxperia
i changed the html template so it looks much better and needed to restart xmrblocks. yes monerod runs always in background. it could be that however i stoped monerod too then restarted it and in parallel run after this xmrblocks too. it may xmrblocks blocked monerod. i rember i had to restart xmrblocks as i did not see my html changes in xmrblocks and thinked its becouse of some cach so when i restar it may go away.
-
linuxperia
but then the monerod database crash happened after this restarts
-
linuxperia
and nothing worked anymore
-
linuxperia
now ofcourse i only try to start monerod alone without p2pool and xmrblocks are running but it fails
-
hyc
do the mdb_dump and mdb_load.
-
linuxperia
aww okey let me look into it. need to lookup the snytax of this two command. use it for the first time
-
linuxperia
hyc: it does the dump right now. used this command here => "mdb_dump -a -f 20220929.backup lmdb"
-
hyc
does it succeed?
-
linuxperia
it still works. its about 130 GB. will write as soon the dump finish
-
linuxperia
hyc: i see lot of io on the ssd so something is going on but it still not finished see picture =>
pasteboard.co/NUWFdp0XfcUk.png
-
linuxperia
hyc: it dumped about 75GB till now so about 50% of the work is done
-
linuxperia
hyc: the monero lmdb has a size of 149725786112 Sep 29 14:50 data.mdb and the dump that is still not finished is allready over 160GB => 160902246400 Sep 29 16:41 20220929.backup hope it finish soon :-(
-
hyc
it dumps in hex, so it will be at least 2x larger than original file
-
hyc
on a fresh build of xmrblocks, it always SEGVs on exit
paste.debian.net/1255451
-
hyc
this is built with monero masster
-
linuxperia
hyc: thank you very much for the information. good to know the size of the dump will be 2x asked me allready how big it will be
-
linuxperia
hyc: ahh the segfault of xmrblock yeah it could be yes. i installed xmrblocks so i can check the blockchain localy direct with my monerod blockchain wihout to use the internet or other websites. xmrblocks is really great but i think it bites with monero especially when you restart it a lot. it could be even that i used two ssh sessesions to stop and run it so something with parallel access to the blockchain lmdb database of monero got wrecked.
-
linuxperia
-
rbrunner
That sounds strange somehow, because you would expect that a block explorer only reads anyway, which *should* be quite harmless
-
linuxperia
rbrunner: yes think also like this but i experienced a problem with the lock file i think in lmdb they both use the lock file if i am not wrong to write
-
revuoxmr
Revuo Monero. Issue 139: September 22 - 29, 2022.
revuo-xmr.com/issue-139.html
-
linuxperia
hyc: YEAAYYYY it Finished :happy:
-
linuxperia
now trying to load the dump then. aww lets hope it will work
-
linuxperia
hyc: you was absolute righ with the 2 time size of the dump. so whenever someone need to do a backup dump he need to have two time free space on disk than the lmdb file size. okey going to do the dump load now
-
hyc
yeah they all have to use the lockfile to coordinate
-
hyc
anyway it makes no sense to start xmrblocks multiple times in parallel. it acquires a single listener port, so only one can succeed.
-
linuxperia
yes did not start it multiple times becouse i was not able to access it but becouse after i changed the html template files ( it come with own webserver ) it did not delivered the changes but showed the old version so thinked it my be a cache problem. restarted it 1 time checked the browser nothing. then repeated it several times and things got broke
-
linuxperia
i am getting this output here right now with mdb_load: asking me if mdb_dump really worked good. checking the files size of monero lmdb looks like its growing instead of staying same. looks like mdb_dump did dumped some wrong data mdb_load: line 7: unrecognized keyword ignored: db_pagesize
-
linuxperia
mdb_load: line 16: unrecognized keyword ignored: duplicates
-
linuxperia
mdb_load: line 20: unrecognized keyword ignored: db_pagesize
-
linuxperia
mdb_load: line 5420013: unrecognized keyword ignored: duplicates
-
linuxperia
mdb_load: line 5420017: unrecognized keyword ignored: db_pagesize
-
hyc
"ignored" - doesn't matter.
-
linuxperia
okey thanks a lot for the info. if mdb_dumps works it is really the prefered solution than resynching everything from scratch.will write again when mdb_load finish
-
hyc
did you alsu use -a flag on mdb_load?
-
hyc
should speed things up a little
-
hyc
and you must delete the original DB before running mdb_load
-
linuxperia
no it does not has the -a flag from what i saw in the docs =>
systutorials.com/docs/linux/man/1-mdb_load but i will check right now. ohh thinked i can write direct over it without to delete it
-
linuxperia
looks like it append the data then
-
linuxperia
okey aborting
-
hyc
you should only trust the documentation bundled in the monero source tree
-
hyc
looks like the doc you linked is for an older version
-
linuxperia
hyc: thanks a lot for the helpful tips. doing a new dump_load now. asked me allready why the lmdb grows in size instead that it should overwrite the data and stay same size.
-
linuxperia
good
-
linuxperia
okey all good now
-
linuxperia
thumbs up :-)
-
linuxperia
lmdb is absolute new to me. will need do more research about it in the future. downloaded allready some php librarys to access lmdb and it works. could do direct read access to lmdb over php as in recent versions lmdb was integrated into lmdb. See the last row in the table about the lmdb in php here =>
php.net/manual/en/dba.requirements.php
-
hyc
be careful about using other LMDB modules. must always link against the same liblmdb.a that monerod uses.
-
hyc
any distro-provided liblmdb.so is nearly always several versions out of date
-
linuxperia
yeah true. especially on lts long term support version they quite outdated. thanks again for the tip.
-
linuxperia
i will write back after a few hours back or maybe even come back tomorrow. it looks like mdb_load works slower then the mdb_dump. load progress at the moment passed 1% right now so it will take some time. thank you all for helping me as a monero newbie out. you really all amazing here. big compliments to all of you especially hyc moneromooo!