Closed jonathanknowles closed 1 year ago
There can't really be any race-conditions here — I think the wallet simply is overloaded and can't keep up.
Logs of past failures: 2565.zip
Looking for db-busy warnings: didn't find that many.
Although before the logging rework, they wouldn't have been visible.
for f in *.log; echo $f (cat $f | grep "No more logs until it finishes" | wc -l); end; 13919.log 0
13941.log 0
14006.log 15
5803003.log 0
5831764.log 0
5873202.log 0
6163056.log 25
6235418.log 0
6298068.log 25
Using ./scripts/bors-stats.rb download-logs --tag "#2565" --after "25 Mar"
to fetch logs from after the rework: I think it looks like there are "SQLITE_BUSY" errors right before each failure.
Not a prominent problem. Unclear whether a problem at all still. Closing.
Context
https://hydra.iohk.io/build/5803003/nixlog/1/tail
Test Case
BYRON_RESTORE_01
Failure / Counter-example
Resolution
QA