-
Notifications
You must be signed in to change notification settings - Fork 1.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
build: backport changes for v1.30.0-rc2 #12598
Conversation
--------- Co-authored-by: Rod Vagg <[email protected]> Co-authored-by: Mikers <[email protected]>
…ts check (#12572) fix(migrations): update actor code IDs to Version15 for NV24 invariants check
chore: set calibration epoch and F3-activation
…12574) Default `address.CurrentNetwork` was changed in an upstream dependency. Update the build constants for non-mainnet to explicitly set it to `Testnet`. Add tests to assert it.
* fix(f3): fix hot loop in F3 participation * fix(f3): Correct lease expiration log message * feat(f3): log not-before and not-after for new leases * test(f3): make the f3-enabled test wait for at one lease to expire --------- Signed-off-by: Jakub Sztandera <[email protected]> Co-authored-by: Steven Allen <[email protected]>
1. Change Restarting -> Renewing. Restarting sounds like something is going wrong. 2. Fix spelling of rechecking.
Otherwise, we can get into a situation where everyone is waiting for the next round to participate, but we'll never get there because not enough participants acted in the current round.
Default address.CurrentNetwork was changed to mainnet. As a result if `LOTUS_USE_TEST_ADDRESSES` is set we need to explicitly use testnet address.
Apply suggestions from code review Co-authored-by: Rod Vagg <[email protected]>
* subscribe-type calls always wants future reverts but only historical reverts when a tipset is specified * get-type calls only want reverts when a tipset is specified Add a new Fill() method to avoid the unnecessary install + remove step for get- type calls. Closes: #12584
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
ETH RPC changes lgtm but we should get another pair of 👀 on the F3 changes here as well.
In _theory_ a lotus node can return no manifest from `F3Manifest`. I don't think that happens in practice but we should handle the case and start over at the top (try to get another lease). If we really have no manifest, lotus will tell us that F3 isn't ready yet and we'll backoff for a bit.
1. Instead of hard-coding the renewal to be within 2 instances of the end of the lease, specify the renewal to be within 1/2th of the end of the lease, rounded up (towards the end of the lease). 2. Instead of hard-coding the lease check interval to be 10s (problematic for tests where the block time is 100ms), set it to be the minimum amount of time we expect to have to wait for the remaining instances to complete.
…2589) 1. Move it next to all the F3 code. 2. Make it look more like a "service" so I won't have to duplicate the start/stop code in the itests. 3. Rename participator to participant. Not perfect, but it's a word. 4. Make all the constants public. 5. Only depend on the parts of the API we need (will help with unit testing).
…12592) Otherwise, the node's `stop` function returns an error and the splitstore and window post dispute tests randomly fail.
All backport label items in the rc2-section has been pushed in this PR. Will merge once CI is happy |
I will bypass this failing f3-test, given that we know this has been unreliable. |
Related Issues
#12480
Proposed Changes
Backports the list outlined in the rc2-section in
🪢 Dependencies for releases
in this issue: #12480Edit: All backport label items in the rc2-section pushed in this PR.
Checklist
Before you mark the PR ready for review, please make sure that: