MDB_BAD_TXN error

[dawk0]

[dawk0]

having a i7-3770k and having the same issue with the skypool and sushipool miners....

[styppo]

Does this problem still occur in the lastest core version?

[vikneshvijayakumar]

Yes

[paberr]

Can you confirm that it only happens with Windows 10 after update 1803 using WSL?

[LTD-Dev]

I can confirm the same

[vikneshvijayakumar]

Yes it is happening of all of my devices on Windows WSL updated with 1803.

[paberr]

Apparently, Microsoft has broken cache coherency between read/write and mmap operations in its recent update.

Later today, I will release a temporary workaround that you can use when compiling Nimiq on WSL.

[paberr]

Please try using the code published on the branch paberr/wsl-workaround. Tell me if further issues exist.

[LTD-Dev]

{ Error: MDB_BAD_TXN: Transaction must abort, has a child, or is invalid
    at LMDBBackend._get (/home/iteonix/core/node_modules/@nimiq/jungle-db/dist/lmdb.js:7166:42)
    at LMDBBackend.getSync (/home/iteonix/core/node_modules/@nimiq/jungle-db/dist/lmdb.js:7368:28)
    at LMDBBackend.get (/home/iteonix/core/node_modules/@nimiq/jungle-db/dist/lmdb.js:7392:21)
    at ObjectStore.get (/home/iteonix/core/node_modules/@nimiq/jungle-db/dist/lmdb.js:3716:35)
    at PeerKeyStore.get (/home/iteonix/core/dist/node.js:34689:28)
    at DumbNetworkConfig._init (/home/iteonix/core/dist/node.js:33517:29)
    at DumbNetworkConfig.initPersistent (/home/iteonix/core/dist/node.js:33495:20)
    at <anonymous>
    at process._tickCallback (internal/process/next_tick.js:188:7)
    at Function.Module.runMain (module.js:684:11) code: -30782 }```

[sOm2y]

I'm getting mdb corruption error, on master branch and branch paberr/wsl-workaround as well, trying to run core nodejs client under wsl.

[sOm2y]

change node_modules/@nimiq/jungle-db/package.json line 4 to "main": "dist/leveldb.js", That will do temp fix

[vikneshvijayakumar]

It works!

[paberr]

Note that using levelDB instead can only be a temporary workaround, as it might result in "This method does not work reliably using LevelDB" errors when trying to retrieve transactions.

[hyc]

For completeness' sake, I did the work to diagnose this bug and notified Microsoft about it https://twitter.com/hyc_symas/status/1013800256088215552

[riptl]

Seems like this bug goes deep, its origin is in the Windows kernel: https://github.com/Microsoft/WSL/issues/3451#issuecomment-415251320

It's probably best to disable LMDB on WSL

[jeffesquivels]

Closing this stale issue.

For anyone experiencing this in the future, the fix is to update your system to WSL2, see https://github.com/Venemo/node-lmdb/issues/161#issuecomment-1118480146.