[WIP] malus-collator: implement initial version of malicious collator submitting same collation to all backing groups #6924
Conversation
…or submitting same collation to all backing groups
…o Collator implementation
…extend test scenario
polkadot/parachain/test-parachains/undying/collator/src/main.rs
Outdated
Show resolved
Hide resolved
| let scheduled_cores: Vec<CoreIndex> = claim_queue | ||
| .iter() | ||
| .filter_map(move |(core_index, paras)| { | ||
| Some((*core_index, *paras.get(claim_queue_offset.0 as usize)?)) | ||
| }) |
There was a problem hiding this comment.
Have a look at ClaimQueueSnapshot to reuse some of the code there with iter_claims_at_depth
There was a problem hiding this comment.
Wanted to use iter_claims_at_depth but by default there is no dependency and I can't use ClaimQueueSnapshot, so either I have to add some dependency or copy the code (which I did in this case).
| charlie: js-script ./assign-core.js with "2,2000,57600" return is 0 within 600 seconds | ||
|
|
||
| # Ensure parachains made progress. | ||
| alice: parachain 2000 block height is at least 10 within 300 seconds |
There was a problem hiding this comment.
we need to add some better assertions, including some log checks that indeed validators got invalid candidates.
Also, if you want this test being run in the CI, it needs to be added to the gitlab config in polkadot.yml
polkadot/zombienet_tests/functional/0020-same-collation-to-all-assigned-cores.zndsl
Outdated
Show resolved
Hide resolved
| @@ -0,0 +1,54 @@ | |||
| [settings] | |||
There was a problem hiding this comment.
DSL is deprecated, we have a goal to port tests to Zombienet SDK.
| @@ -86,6 +90,8 @@ pub struct GraveyardState { | |||
| pub zombies: u64, | |||
| // Grave seal. | |||
| pub seal: [u8; 32], | |||
| // Increasing sequence number for core selector. | |||
| pub core_selector_number: u8, | |||
There was a problem hiding this comment.
I was expecting we use the block number (LSB), is there anything preventing us to do so ?
There was a problem hiding this comment.
Since we use upward messages to place UMP signals and specify the core index, we needed this core_selector_number for the core selector. Do you mean doing something else?
There was a problem hiding this comment.
Andrei means that we typically use the block number for the core selector in real parachains.
The undying collator's state does not store the block number, so this core_selector_number is actually the block number
| } | ||
|
|
||
| // Wait before repeating the process. | ||
| sleep(Duration::from_secs(6 as u64)); |
There was a problem hiding this comment.
Why do we need to wait 6s ? If relay chain skips a slot, we build another candidate, which is different than what the honest collator is doing. We should be building on each relay parent we import.
There was a problem hiding this comment.
Do you mean no waiting at all, or should the waiting time somehow depend on the imports?
There was a problem hiding this comment.
The current code is very roughly mimicking what the slot based collator does (every slot, pick the best relay chain block)
What Andrei is suggesting is to simulate what the lookahead collator does and subscribe to relay chain block import notifications instead. This would indeed be closer to what the honest collator is doing. I think for a test collator either is fine
|
All GitHub workflows were cancelled due to failure one of the required jobs. |
| target: LOG_TARGET, | ||
| "Persisted validation data is None." | ||
| ); | ||
| Self::wait_for_next_slot(); |
There was a problem hiding this comment.
Instead of duplicating this on every continue, code it in a way so that you can move the sleep to the top of the loop.
Alternatively, to avoid these sleeps, we can do as Andrei suggested and subscribe to block import notifications and build in response to that (which is what the non-malicious undying collator does)
Issues
Description
Modified the undying collator to include a malus mode, in which it submits the same collation to all assigned backing groups.
TODO
same-collation-to-all-assigned-cores.tomlusing theassign-core.jsscript;same-collation-to-all-assigned-cores.zndslZombienet test;