Skip to content
This repository has been archived by the owner on Aug 2, 2022. It is now read-only.

Add conditional step in state history plugin how-to [docs] - 2.0 #9145

Merged
merged 3 commits into from
May 29, 2020
Merged
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Original file line number Diff line number Diff line change
Expand Up @@ -24,7 +24,7 @@ This procedure records the current chain state and future history, without previ

4. Look for `Placing initial state in block n` in the log, where n is the start block number.

5. Start a filler with `--fpg-create` (if PostgreSQL), `--fill-skip-to n`, and `--fill-trim`. Replace `n` with the value above.
5. If using a database filler, start the filler with `--fpg-create` (if PostgreSQL), `--fill-skip-to n`, and `--fill-trim`. Replace `n` with the value above.

6. Do not stop `nodeos` until it has received at least 1 block from the network, or it won't be able to restart.

Expand All @@ -36,7 +36,7 @@ If `nodeos` fails to receive blocks from the network, then try the above using `
| Either use a firewall to block access to your `http-server-address`, or change it to `localhost:8888` to disable remote access.

[[info]]
| Whenever you run a filler after this point, use the `--fill-trim` option. Only use `--fpg-create` and `--fill-skip-to` the first time.
| If you run a database filler after this point, use the `--fill-trim` option when restarting. Only use `--fpg-create` and `--fill-skip-to` the first time.

[[info]]
| On large chains, this procedure creates a delta record that is too large for javascript processes to handle. 64-bit C++ processes can handle the large record. `fill-pg` and `fill-lmdb` break up the large record into smaller records when filling databases.
| On large chains, this procedure creates a delta record that is too large for javascript processes to handle. 64-bit C++ processes can handle the large record. If using a database filler, `fill-pg` and `fill-lmdb` break up the large record into smaller records when filling databases.