blocks are written to the db in an incomplete state which halts indexing

[habibitcoin]

I just tried this using the latest archives, and then trying the 20240920 archives

Each time, I hit this issue:

Sep 20 17:27:19 ip-172-31-87-211 ordhook[3957448]: Sep 20 17:27:19.359 INFO Inscriptions data computation for block #860926 collected
Sep 20 17:27:19 ip-172-31-87-211 ordhook[3957448]: Sep 20 17:27:19.359 INFO Inscriptions data computation for block #860926 ended
Sep 20 17:27:19 ip-172-31-87-211 ordhook[3957448]: Sep 20 17:27:19.359 ERRO Unable to retrieve backward traversal result for inscription 0x8f24e1365658017220c603559f46f0b84748316dd8bf9261d3448cd8c2d61087

Any ideas?

Im wondering if perhaps there is a bug in the archives themselves? Curiously, the updated date when extracting the brc20.sqlite and hord.sqlite files are Sept 11 and Sept 12 respectively.

Related issues:
#334
#310

[inatendu29]

I have the same issue.
Recompiling ordhook = no go
Deleting the database = no go
Using other backup archives from Hiro = no go
It's exactly the same block and inscription that generates the problem
OS = Windows 10
I7 RAM 32gb m2 2to

[rafaelcr]

hi @inatendu29 thanks for reporting, I just confirmed that the archive does contain this flaw so even if you download our latest you'll stumble into this issue.

I haven't yet identified the cause for this issue, though, so I will continue to investigate.

what you can do to fix it for now is to run this command first

ordhook db drop 860926 860926 --config-path <path>

which will get rid of the corrupt block, and then go back to

ordhook service start --config-path <path>

this will resume indexing

[rafaelcr]

leaving here for reference, this line causes the error because some data for this block exists which causes the sequencing code to skip some previously-seen inscriptions