Add cli-flag to keep db after changing assignments #280
Conversation
|
Master coverage: 70.97% |
|
One of the warp sync tests is failing because it seems the container chain db for |
That was because it was panicking on this line: I pushed a commit fixing that by using a different kind of channel to signal shutdown, now we can distinguish between "stopping because someone called .stop()" and "stopping because the node process is stopping". We will only delete the db in the first case. |
|
@tmpolaczyk thank you so much for the clarification! I was hesitating if that piece of code was in the correct place. |
node/src/container_chain_spawner.rs
Outdated
| .latest_block_number(orchestrator_chain_info.best_hash, container_chain_para_id) | ||
| .unwrap_or_default(); | ||
|
|
||
| let max_block_diff_allowed = 10u32; |
There was a problem hiding this comment.
| let max_block_diff_allowed = 10u32; | |
| let max_block_diff_allowed = 100u32; |
node/src/container_chain_spawner.rs
Outdated
| @@ -226,7 +254,7 @@ impl ContainerChainSpawner { | |||
| container_chain_cli_config.database.set_path(&db_path); | |||
|
|
|||
| // Delete existing database if running as collator | |||
| if validator { | |||
| if validator && !container_chain_cli.base.keep_db && warp_sync { | |||
There was a problem hiding this comment.
I am hesitating about this. What happens if:
- I have the keep_db flag to true
- but I detect that the genesis-hash of the chain-spec I am running is different (genesis-hash different)
- in that case I would be stuck (the node tells me to reboot and that I need to use warp sync, all the time, but I need all these 3 flags to be true to delete the db and hence be able to do warp sync)
There was a problem hiding this comment.
we should immediatly delete the container-chain-db if the spec is differnt I believe
There was a problem hiding this comment.
there is no reason to maintain a DB that does not match with the specs right?
|
I like the way it looks now, we can unitest the delete db cases quite easily! |
This PR adds a simple
keep_dbflag toContainerChainRunCmdstruct that allows to keep container-chain databases after changing collator assignments.