Revert to simple polling wait #6861
Conversation
|
Shame to throw the initial effort away for this code. I didn't anticipate this trace issue I'm looking at taking so long, otherwise would have got to this sooner. Could we wait another day and I'll see what I can find this evening/tomorrow before we revert? |
|
Think I've sorted this, a silly mistake on my behalf. Just running through the full hive suite now. Will push a PR tomorrow morning if all looks good. |
|
Can it be related to #6807 (comment) ? |
I'm not sure it's related, there's no mention of the stageLoopBusy call in those logs. Can't rule it out though. Be worth a try once the fix here is merged. Hive should finish shortly so will find out (computer went to sleep overnight). |
|
Created #6871 as an alternative |
|
Closing in favour of PR #6871. |
Unfortunately PR #6763 breaks a few Hive tests, for example "Invalid Ancestor Chain Sync, Invalid StateRoot, Invalid P9". Thus reverting to the simpler old logic, but with the maximum wait kept as 1 sec.