How can I verify the correctness of chaindata/mdbx.dat ?

[mookim-eth]

I'm running erigon on zfs above two 8TB Samsung SSD 870, and I found zpool status -v shows:

  pool: ssd
 state: ONLINE
status: One or more devices has experienced an error resulting in data
        corruption.  Applications may be affected.
action: Restore the file in question if possible.  Otherwise restore the
        entire pool from backup.
   see: https://openzfs.github.io/openzfs-docs/msg/ZFS-8000-8A
  scan: scrub repaired 0B in 08:03:12 with 2 errors on Thu Jul  7 19:43:54 2022
config:

        NAME        STATE     READ WRITE CKSUM
        ssd         ONLINE       0     0     0
          sdb       ONLINE       0     0 1.09K
          sda       ONLINE       4     0     2

errors: Permanent errors have been detected in the following files:

        /ssd/bsc/chaindata/mdbx.dat
        /ssd/eth/chaindata/mdbx.dat

It seems zfs thinks the chaindata/mdbx.dat is corrupted, is there a way to verify the correctness of the db?

Or should I just give up and re-sync from scratch?

[AskAlexSharov]

make db-tools
./build/bin/mdbx_chk --help
it may recover some kind of corruptions if they happen in last 1-2 transactions.
  -0|1|2	force using specific meta-page 0, or 2 for checking
  -t		turn to a specified meta-page on successful check

./build/bin/mdbx_chk -vv /ssd/bsc/chaindata

[mookim-eth]

# /ssd/bsc/erigon/build/bin/mdbx_chk -vv /ssd/eth/chaindata
mdbx_chk v0.11.7-35-g5d2eb580 (2022-05-19T13:11:25+03:00, T-43e6c3f611860f9ee55106fa7e5cd9cd8d2e6853)
Running for /ssd/eth/chaindata in 'read-only' mode...
   open-MADV_DONTNEED 541647259..542113792
   readahead OFF 0..541647259
 - monopolistic mode
 - current boot-id a63d543188b451e6-b085b1274f1f4de5
 - pagesize 4096 (4096 system), max keysize 1980..2022, max readers 118
 - mapsize 8796093022208 (8.00 Tb)
 - dynamic datafile: 12288 (12.00 Kb) .. 8796093022208 (8.00 Tb), +2147483648 (2.00 Gb), -4294967296 (4.00 Gb)
 - current datafile: 2220498092032 (2.02 Tb), 542113792 pages
 - meta-0: steady txn#1013753, stay
 - meta-1: steady txn#1013754, head
 - meta-2: steady txn#1013752, tail
 - performs check for meta-pages clashes
 - performs full check recent-txn-id with meta-pages
 - transactions: recent 1013754, latter reader 1013754, lag 0
Traversal b-tree by txn#1013754...
 - found 'AccountChangeSet' area
Bus error (core dumped)

Seems libmdbx cannot handle this error...

[AndreaLanfranchi]

@mookim-eth Bus error indicates an hw error: most likely RAM error. CPU cant' even try to access the address of the memory mapped file.
Highly suggested to run a memtest and possibly to use ECC RAM