Instructions for those struggling:
Please don't follow these steps without appropriately backing up your files and databases. Postgres db can be backed up using pg_dump -d dbname > file.sql, as well as pg_dumpall -f file.sql. (single db vs entire cluster)
The indexer has been fixed in latest release, so that if (but only if) you were indexed up to latest block pre-fork (237906), it will pick up nicely post-fork.
Those struggling are either due to 1) not being fully synced pre-fork or 2) running an older/custom version of the indexer or 3) both. solutions below:
If you have not synced fully prefork, you need to sync those blocks off an 31.9 rpc. We suggest if you followed due diligence in the migration on fork, you will have a backup you can use for this purpose. Basically if you are very careful(!) you can shut down your 32.1 rpc, switch folders (backup/move current, restore/move old backup in place, run 31.9 rpc, sync old blocks with indexer, shut down indexer, switch back to newer SE-folder, and pick up 32.1 where you left off). Don't do this if you don't know what you're doing, we take absolutely no responsibility if any of this goes horribly wrong or wrecks your system. This should solve the issue of not being synced up to 237906 if that's part of your problem.
Now that you are synced prefork, if you use newest indexer pulled from github, you should be good to run out of the box.
However, older indexer versions require a trick, we have prepared a little aid:
The following sql script is safe, but please verify before running! It inserts block 237907 into the indexer database and that's all it does. One row in blocks table. For your convenience:
https:// testnet.one / 237907.sql. (Leaving it for you to enter manually and download, as we want to make sure you know what you are doing. You can and should inspect the contents of the file)
when you are certain it's clean, you can run it off commandline as postgres user (su postgres), with: psql -d dbname -a -f 237907.sql, or alternately paste the sql code into psql cli.
This should insert the missing block into the db, and indexer should pick up fine after that.